0266R_HalfTitle 2/12/04 9:44 AM Page 1

Mechanical
Testing of Bone
and the
BoneImplant
Interface

0266/ fm/frame Page 2 Thursday, October 21, 1999 1:47 PM
0266R_Title 2/12/04 9:43 AM Page 1
edited by
Yuehuei H. An, M.D.
Robert A. Draughn, D.Sc.
Mechanical
Testing of Bone
and the
BoneImplant
Interface
CRC PRESS
Boca Raton London New York Washington, D.C.

This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with
permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts have been made to publish
reliable data and information, but the authors and the publisher cannot assume responsibility for the validity of all materials
or for the consequences of their use.
Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical,
including photocopying, microlming, and recording, or by any information storage or retrieval system, without prior
permission in writing from the publisher.
All rights reserved. Authorization to photocopy items for internal or personal use, or the personal or internal use of specic
clients, may be granted by CRC Press LLC, provided that $.50 per page photocopied is paid directly to Copyright Clearance
Center, 222 Rosewood Drive, Danvers, MA 01923 USA. The fee code for users of the Transactional Reporting Service is
ISBN 0-8493-0266-8/00/$0.00+$.50. The fee is subject to change without notice. For organizations that have been granted
a photocopy license by the CCC, a separate system of payment has been arranged.
The consent of CRC Press LLC does not extend to copying for general distribution, for promotion, for creating new works,
or for resale. Specic permission must be obtained in writing from CRC Press LLC for such copying.
Direct all inquiries to CRC Press LLC, 2000 N.W. Corporate Blvd., Boca Raton, Florida 33431.

Trademark Notice:

Product or corporate names may be trademarks or registered trademarks, and are used only for
identication and explanation, without intent to infringe.

Visit the CRC Press Web site at www.crcpress.com

2000 CRC Press LLC
No claim to original U.S. Government works
International Standard Book Number 0-8493-0266-8
Library of Congress Card Number 99-36531
Printed in the United States of America 2 3 4 5 6 7 8 9 0
Printed on acid-free paper

Library of Congress Cataloging-in-Publication Data

Mechanical testing of bone and the bone-implant interface / edited by
Yuehuei H. An and Robert A. Draughn.
p. cm.
Includes bibliographical references and index.
ISBN 0-8493-0266-8 (alk. paper)
1. BonesMechanical properties. 2. Orthopedic implants.
3. BonesEffect of implants on. 4. Biomechanics. I. An, Yuehuei H.
II. Draughn, Robert A.
[DNLM: 1. Bone and Bonesphysiology. 2. Biocompatible Materials.
3. Biomechanics. 4. Materials Testingmethods. 5. Prostheses and
Implants. 6. Stress, Mechanical. 7. Surface Properties.
8. Tensile Strength. WE 200 M4856 1999]
QP88.2.M424 1999
617.4

71dc21 99-36531
DNLM/DLC

0266.fm Page 1 Tuesday, February 17, 2004 12:43 PM

Foreword

The skeletal system gives the body its form, facilitates movement, and protects internal organs
from traumatic forces. Any disease, drug, or biological process that inuences bone directly
inuences the fundamental mechanical function of the skeleton. Skeletal imaging data and quan-
titative assessments of bone composition and histomorphology are often important only because
they reect something about mechanical competence. To truly assess the mechanical competence
of the skeleton, however, it is imperative that we assess the mechanical characteristics of the bone
and boneimplant constructs.
Early investigations of the mechanical properties of bone in the last half of the twentieth century
by Evans, Yamada, Katz, Ascenzi, Currey, Burstein, Boneld and Lanyon helped form the modern
foundation for how bones should be tested and viewed from a mechanical and material perspective.
From these initial pioneers, new investigators began to build a research literature and a series of
laboratory approaches for the mechanical testing of bone and boneimplant systems. The mechan-
ical testing of bone and the boneimplant interface has progressively become an important aspect
of a wide variety of research projects on bone growth, adaptation, regeneration, and aging. This
work has resulted in a better understanding of the material mechanical properties of bone and
greater standardization of testing protocols.
This collection by editors Yuehuei H. An and Robert A. Draughn marks a transition in the role
of mechanical testing of bone and boneimplant systems. It addresses the eld of bone mechanics
not so much as an arena of basic science but rather as a compendium of practical research tools
and mechanical assay techniques. It is a wonderful addition to the literature in that it summarizes
much of the data generated in recent years. The greatest value of the book, however, is that it
provides a synopsis of laboratory approaches that a broad range of investigators have used and
continue to use in their research.
The treatise begins at a very basic level and can therefore serve as an effective introduction to
those without signicant previous experience in bone mechanics. The 39 chapters of the book are
separated into three sections that progress from concise descriptions of techniques used to charac-
terize and test bone to more specialized laboratory studies. The perspectives of the many contributors
to the book who work in many different laboratories contribute to the richness and breadth of the
book. The coherence of the book is maintained under the clear direction of the editors. The emphasis
is on teaching the reader how to do it. This theme is maintained throughout and is surely enhanced
by the involvement of the editors as coauthors in many chapters.
This book serves as a comprehensive primer for the mechanical testing of bone. For new
investigators in this area, it is an invaluable tool. For the more experience investigators, it serves
as a touchstone for evaluating new testing protocols and data. This text has a great deal to offer
all of us.

Dennis R. Carter, Ph.D.

Director, Palo Alto VA Rehabilitation R&D Center
Professor, Biomechanical Engineering Division
Mechanical Engineering Department
Stanford University

0266/ fm/frame Page 5 Thursday, October 21, 1999 1:47 PM

0266/ fm/frame Page 6 Thursday, October 21, 1999 1:47 PM

Preface

Biomechanics is an integral part of the study of bone as an organ or tissue. Mechanical testing of
bone specimens is a basic method in bone-related research. The mechanical properties of whole
bones or bone tissues and boneimplant interfaces are equally important as their morphological or
structural aspects. The former is evaluated by mechanical testing and the latter is mostly studied
using histological techniques.
This book is an outgrowth of the editors own quest for information on mechanical testing of
bone and, more importantly, a response to signicant needs in the orthopaedic research community.
Most researchers are not well trained in biomechanics and assistance from an expert is not always
readily available. What many researchers really need to know are basic mechanical principles in
bone-related research and most importantly how to conduct mechanical testing of bone specimens.
This book is designed to be an experimental guide for orthopaedic or dental residents, bioengi-
neering graduate students, orthopaedic or dental researchers, biomaterials scientists, laboratory
technicians, and anyone not well trained in biomechanics who plans to conduct mechanical testing
of bone specimens. Most readers belong to societies in the elds of orthopaedic or dental research,
biomechanics, or biomaterials, such as the Orthopaedic Research Society, American Society of
Biomechanics, American Society of Mechanical Engineers, American Society for Bone and Mineral
Research, Society for Biomaterials, or Materials Research Society. This text is intended to be a
beginners guide, and no prior training in biomechanics is required to understand the contents.
It should also serve as a useful handbook for biomechanical and bioengineering researchers and
students at all levels.
This is the rst inclusive and organized reference book on how to perform mechanical testing
of bone. The topic has not been adequately covered by any existing textbook on bone biomechanics.
The 39 chapters of this book are divided into three major parts: Section I mechanical properties
of bone and general considerations and basic facilities for mechanical testing; Section II specic
mechanical testing procedures on bone tissues; and Section III mechanical testing procedures
on the boneimplant interface.
The book is designed to be concise as well as inclusive and more practical than theoretical.
The text is simple and straightforward. Numerous diagrams (~150), tables (~150), line drawings
(~150), and photographs (~150) are included to help readers better understand the main principles.
Full bibliographies at the end of each chapter guide readers to more detailed information. A book
of this length cannot discuss every method in biomechanical testing of bone that has been conducted
over the years, but it is hoped that major methods and their applications have been included.

Yuehuei H. An, M.D.

Charleston, South Carolina

0266/ fm/frame Page 7 Thursday, October 21, 1999 1:47 PM

0266/ fm/frame Page 8 Thursday, October 21, 1999 1:47 PM

The Editors

Yuehuei H. (Huey) An, M.D.,

graduated from the Harbin Medical University, Harbin, Northeast
China in 1983. He completed residency training in orthopaedic surgery at Ji Shui Tan Hospital in
Beijing, China and went on to a fellowship in hand surgery at Sydney Hospital in Sydney, Australia.
In 1991, Dr. An joined Dr. Richard J. Friedman in the Department of Orthopaedic Surgery at the
Medical University of South Carolina to establish the MUSC Orthopaedic Research Laboratory,
which is now a multifunctional orthopaedic research center.
Soon after beginning his career in orthopaedic research, Dr. An developed an interest in bone
mechanics and he learned much about practical mechanical testing by trial and error. His under-
standing of the importance of biomechanical principles and mechanical testing techniques to
researchers in bone-related elds of work led to his desire to organize this effort.
Dr. An has published more than 70 scientic papers and book chapters and more than 60
research abstracts. He has edited three reference books. His rst book,

Animal Models in
Orthopaedic Research

, a major contribution to orthopaedic research, was published by CRC
Press in 1998. His third book

, Handbook of Bacterial Adhesion Principles, Methods, and
Applications

, will be published by Humana Press in late 1999. He created many of the line
drawings used in his books and papers. He is an active member of eight academic societies
including the American Society of Biomechanics, Orthopaedic Research Society, Society for
Biomaterials, and the Tissue Engineering Society. Dr. Ans current research interests include
bone and cartilage repair using tissue engineering techniques, bone or soft tissue ingrowth to
implant surfaces, bone structure and biomechanics, bacterial adhesion and prosthetic infection,
and animal models in orthopaedic research.

Robert A. Draughn, D.Sc.,

is Professor and Chairman of the Department of Materials Science at
the Medical University of South Carolina, Charleston, South Carolina. He is also an Adjunct Associate
Professor of Bioengineering at Clemson University, Clemson, South Carolina. Dr. Draughn earned
his Doctor of Science degree in Materials Science from the University of Virginia in 1968. He has
been on the faculty of the College of Dental Medicine at the Medical University of South Carolina
since 1973. His principal research interests have been in the general area of biomedical applications
of composite materials. The emphasis of much of his work has been the mechanical properties of
particle-reinforced polymers and hard tissues as well as studies of adhesive bonding processes.
Dr. Draughn has over 70 publications and over 90 research presentations and published
abstracts. He is active in several professional organizations. His activities have included Chairperson
of the Biomaterials Section of the American Association of Dental Schools, Chairperson of the
Gordon Research Conference on the Science of Adhesion, and membership on the executive
committee of the Adhesion Society.

0266/ fm/frame Page 9 Thursday, October 21, 1999 1:47 PM

0266/ fm/frame Page 10 Thursday, October 21, 1999 1:47 PM

Contributors

C. Mauli Agrawal, Ph.D.

Associate Professor
Department of Orthopaedics
University of Texas Health Science Center
at San Antonio
San Antonio, Texas

Yuehuei H. An, M.D., M.Sc.

Associate Professor and Director
Orthopaedic Research Laboratory
Department of Orthopaedic Surgery
Medical University of South Carolina
Charleston, South Carolina
and
Adjunct Assistant Professor
Department of Bioengineering
Clemson University
Clemson, South Carolina

Antonio Ascenzi, Ph.D.

Professor Emeritus
Department of Experimental Medicine
and Pathology
University of Rome La Sapienza
Rome, Italy

Maria-Grazia Ascenzi, Ph.D.

Mathematical Researcher
Department of Sciences
University of California Extension
Los Angeles, California

Kyriacos Athanasiou, Ph.D., P.E.

Associate Professor of Orthopaedics
and Engineering
Director of Musculoskeletal
Bioengineering Center
Director of Orthopaedic Biomechanics
The University of Texas Health Science Center
at San Antonio
San Antonio, Texas

William R. Bareld, Ph.D.

Assistant Professor
Orthopaedic Research Laboratory
Department of Orthopaedic Surgery
Medical University of South Carolina
Charleston, South Carolina

Christopher V. Bensen, M.D.

Orthopaedic Surgery Resident
Department of Orthopaedic Surgery
Medical University of South Carolina
Charleston, South Carolina

Alessandro Benvenuti, Ph.D.

Research Associate
Department of Experimental Medicine
and Pathology
University of Rome La Sapienza
Rome, Italy

Aivars Berzins, M.D.

Assistant Professor
Department of Orthopaedic Surgery
Rush Medical College
Rush-Presbyterian-St. Lukes Medical Center
Chicago, Illinois

Ermanno Bonucci, Ph.D.

Professor and Chairman
Department of Experimental Medicine
and Pathology
University of Rome La Sapienza
Rome, Italy

Matthew S. Crum, B.Sc.

Undergraduate Student
Department of Mechanical Engineering
Clemson University
Clemson, South Carolina

0266/ fm/frame Page 11 Thursday, October 21, 1999 1:47 PM

John M. Cuckler, M.D.

Professor and Chairman
Department of Orthopaedic Surgery
University of Alabama
Birmingham, Alabama

A. U. (Dan) Daniels, Ph.D.

George Thomas Wilhelm Endowed Professor
Department of Orthopaedic Surgery
University of Tennessee
Memphis, Tennessee

James R. Davis, B.Sc., FRCS

Research Fellow
Department of Orthopaedic Surgery
University of Maryland
Baltimore, Maryland

Wouter J. A. Dhert, M.D., Ph.D.

Professor
Department of Orthopaedics
Utrecht University Hospital
Utrecht, the Netherlands

Robert A. Draughn, D.Sc.

Professor of Materials Science
Department of Materials Science
College of Dental Medicine
Medical University of South Carolina
Charleston, South Carolina
and
Adjunct Associate Professor of Bioengineering
Department of Bioengineering
Clemson University
Clemson, South Carolina

Lisa A. Ferrara, M.Sc.

Director
Spine Research Laboratory
Department of Neurosurgery
Cleveland Clinic Foundation
Cleveland, Ohio

Jos Luis Ferretti, M.D., Ph.D.

Director
Centro de Estudios de Metabolismo
Fosfocalcico (CEMFoC)
Hospital del Centenario
National University of Rosario
and
Instituto/Fundacion de Investigaciones
Metabolicas (IDIM)
Buenos Aires, Argentina

Richard J. Friedman, M.D., FRCSC

Professor
Department of Orthopaedic Surgery
Medical University of South Carolina
Charleston, South Carolina
and
Adjunct Professor of Bioengineering
Department of Bioengineering
Clemson University
Clemson, South Carolina

Benjamin R. Furman, M.S.

Research Associate
Division of Biomaterials
Department of Restorative Dentistry
University of Texas Health Science Center at
San Antonio
San Antonio, Texas

Vasanti M. Gharpuray, Ph.D.

Associate Professor
Department of Bioengineering
Clemson University
Clemson, South Carolina

Steven A. Goldstein, Ph.D.

Henry Ruppenthal Family Professor of
Orthopaedic Surgery and Bioengineering
Professor of Mechanical Engineering and
Applied Mechanics
Director of Orthopaedic Research
Interim Associate Dean for Research and
Graduate Studies
University of Michigan
Ann Arbor, Michigan

0266/ fm/frame Page 12 Thursday, October 21, 1999 1:47 PM

C. Edward Hofer, M.S.

M.D./Ph.D. Candidate
Orthopaedic Research Laboratories
The University of Michigan
Ann Arbor, Michigan

Sarandeep S. Huja, Ph.D., B.D.S.,
M.D.S., M.Sc.

Graduate Dental Student
Section of Orthodontics
Indiana University School of Dentistry
Indianapolis, Indiana

Ivan Hvid, M.D., D.M.Sc.

Professor
Department of Orthopaedics
Aarhus University Hospital
Aarhus, Denmark

Kenneth S. James, Ph.D.

Associate Director of Orthopaedic Programs
Tissue Engineering, Inc.
Boston, Massachusetts

John A. Jansen, D.D.S., Ph.D.

Professor and Head
Department of Biomaterials
University of Nijmegen Dental School
Nijmegen, the Netherlands

Riyaz H. Jinnah, M.D., FRCS

Professor of Orthopaedic Surgery
The University of Maryland School of
Medicine
Baltimore, Maryland

Qian Kang, M.D.

Associate Chief Surgeon
Department of Orthopaedic Surgery
Beijing Ji Shui Tan Hospital
Beijing, China
Former Research Fellow (19951997)
Orthopaedic Research Laboratory
Department of Orthopaedic Surgery
Medical University of South Carolina
Charleston, South Carolina

Thomas R. Katona, Ph.D., DM.D.

Associate Professor of Orthodontics
Indiana University School of Dentistry
Associate Professor of Mechanical Engineering
Purdue University School of Engineering
and Technology
IUPUI Biomechanics and Biomaterials
Research Center
Indianapolis, Indiana

J. Lawrence Katz, Ph.D.

Professor, Department of Biomedical
Engineering
Case Western Reserve University
Cleveland, Ohio

Tony S. Keller, Ph.D.

Associate Professor
Department of Mechanical Engineering
University of Vermont
Burlington, Vermont

Fadi M. Khoury, M.Sc.

School of Engineering and Applied Science
University of Pennsylvania
Philadelphia, Pennsylvania

Ivars Knets, Ph.D.

Professor of Biomechanics
Department of Mechanical Engineering
Riga Technical University
Riga, Latvia

David H. Kohn, Ph.D.

Associate Professor
Department of Biologic and Materials Sciences
School of Dentistry
University of Michigan
Ann Arbor, Michigan

Michael A. K. Liebschner, Ph.D.

Orthopaedic Research Laboratory
University of California at Berkeley
Berkeley, California

Frank Linde, M.D., DM.Sc.

Consultant Orthopaedic Surgeon
Department of Orthopaedics
Aarhus University Hospital
Aarhus, Denmark

0266/ fm/frame Page 13 Thursday, October 21, 1999 1:47 PM

Robert A. Lofthouse, M.A., FRCS

Research Fellow
Department of Orthopaedic Surgery
University of Maryland School of Medicine
Baltimore, Maryland

Mandi J. Lopez, D.V.M., M.S.

Postdoctoral Fellow
Comparative Orthopaedic Research Laboratory
School of Veterinary Medicine
University of Wisconsin-Madison
Madison, Wisconsin

Mark D. Markel, D.V.M.

Professor
Comparative Orthopaedic Research Laboratory
School of Veterinary Medicine
University of Wisconsin-Madison
Madison, Wisconsin

Barbara R. McCreadie, M.S.

Ph.D. Candidate
Orthopaedic Research Laboratories
University of Michigan
Ann Arbor, Michigan

Brodie E. McKoy, M.D.

Orthopaedic Surgery Resident
Department of Orthopaedic Surgery
Medical University of South Carolina
Charleston, South Carolina

Peter L. Mente, Ph.D.

Assistant Professor
Department of Biological and Agricultural
Engineering
North Carolina State University
Raleigh, North Carolina

Sanjiv H. Naidu, M.D., Ph.D.

Assistant Professor
Department of Orthopaedic Surgery
Pennsylvania State University
Hershey, Pennsylvania

Takashi Nakamura, M.D., Ph.D.

Professor
Department of Orthopaedic Surgery
Kyoto University
Kyoto, Japan

Shigeru Nishiguchi, M.D.

Research Assistant
Department of Orthopaedic Surgery
Graduate School of Medicine
Kyoto University
Kyoto, Japan

George M. Pharr, Ph.D.

Professor
Department of Mechanical Engineering and
Materials Science
Rice University
Houston, Texas

William S. Pietrzak, Ph.D.

Director
Resorbable Technology
Biomet Co.
Warsaw, Indiana

Jae-Young Rho, Ph.D.

Assistant Professor
Department of Biomedical Engineering
University of Memphis
Memphis, Tennessee

W. Eugene Roberts, D.D.S., Ph.D.

Professor of Orthodontics
Indiana University School of Dentistry
Professor of Physiology and Biophysics
Indiana University School of Medicine
and
Professor of Mechanical Engineering
Purdue University School of Engineering
and Technology
Indianapolis, Indiana

Timothy C. Ryken, M.D.

Assistant Professor
Division of Neurosurgery
University of Iowa School of Medicine
Iowa City, Iowa

Subrata Saha, Ph.D.

Professor
Department of Bioengineering
Clemson University
and
Director
Bioengineering Alliance of South Carolina
Clemson, South Carolina

0266/ fm/frame Page 14 Thursday, October 21, 1999 1:47 PM

David R. Sarver, B.Sc.

Product Development Engineer
Biomet Co.
Warsaw, Indiana

Naoki Sasaki, D.Sc.

Associate Professor
Division of Biological Sciences
Graduate School of Science
Hokkaido University
Sapporo, Japan

Rakesh Saxena, Ph.D.

Department of Mechanical Engineering
The University of Vermont
Burlington, Vermont

Chris W. Smith, Ph.D.

Research Fellow
Department of Engineering
University of Exeter
Exeter, United Kingdom

Erica A. Smith, M.S.

Ph.D. Candidate
Orthopaedic Research Laboratories
University of Michigan
Ann Arbor, Michigan

Dale R. Sumner, Ph.D.

Professor and Chairman
Department of Anatomy
and
Professor
Department of Orthopaedic Surgery
Rush Medical College
Rush-Presbyterian-St. Lukes Medical Center
Chicago, Illinois

John A. Szivek, Ph.D.

Professor
Orthopaedic Research Laboratory
Department of Surgery
University of Arizona School of Medicine
Tucson, Arizona

Charles H. Turner, Ph.D.

Associate Professor and Director of
Orthopaedic Research
Department of Orthopaedic Surgery and
Biomechanics and Biomaterials
Research Center
Indiana University Medical Center
Indianapolis, Indiana

Rong-Ming Wang, Ph.D.

Associate Professor
Head of Metal Physics and
Failure Analysis Laboratory
Beijing Institute of Aeronautical Materials
Beijing, China

Xiaodu Wang, Ph.D.

Assistant Professor
Department of Orthopaedics
University of Texas Health Science Center
at San Antonio
San Antonio, Texas

Keith R. Williams, B.Sc., Ph.D.

Reader
Department of Basic Dental Science
Dental School
University of Wales College of Medicine
Cardiff, Wales, United Kingdom

Franklin A. Young, Jr., D.Sc.

Professor
Department of Materials Science
College of Dental Medicine
Medical University of South Carolina
Charleston, South Carolina

Peter Zioupos, Ph.D., MIPEM

Lecturer
Department of Materials and Medical Sciences
Craneld University
Shrivenham, United Kingdom

0266/ fm/frame Page 15 Thursday, October 21, 1999 1:47 PM

0266/ fm/frame Page 16 Thursday, October 21, 1999 1:47 PM

Acknowledgments

The editors would like to acknowledge Kylie Martin for her tireless assistance in communication
with contributors, manuscript review, and editorial assistance. We would also like to thank
Drs. Christopher Bensen and Brodie McKoy for revising manuscripts and preparing gures. We
are also grateful to Drs. Richard Friedman and Angus McBryde, Jr. and all members of the
Departments of Orthopaedic Surgery and Materials Science of the Medical University of South
Carolina for their continuous administrative support of our work. Finally, we wish to thank Liz
Covello, Acquiring Editor at CRC Press, for her help on this and the previous text,

Animal Models
in Orthopaedic Research

.

0266/ fm/frame Page 17 Thursday, October 21, 1999 1:47 PM

0266/ fm/frame Page 18 Thursday, October 21, 1999 1:47 PM

To Kay Q. Kang, M.D.
Without her love, inspiration, and support,
this book would not have been possible

Yuehuei H. An, M.D.

To Donna, Sally, and Margaret

Robert A. Draughn, D.Sc.

0266/ fm/frame Page 19 Thursday, October 21, 1999 1:47 PM

0266/ fm/frame Page 20 Thursday, October 21, 1999 1:47 PM

Contents

Section I General Considerations

Chapter 1

Basic Composition and Structure of Bone........................................................................................3

Ermanno Bonucci

Chapter 2

Basic Concepts of Mechanical Property Measurement and Bone Biomechanics..........................23

Yuehuei H. An, William R. Bareld, and Robert A. Draughn

Chapter 3

Mechanical Properties of Bone........................................................................................................41

Yuehuei H. An

Chapter 4

Factors Affecting Mechanical Properties of Bone ..........................................................................65

Peter Zioupos, Chris W. Smith, and Yuehuei H. An

Chapter 5

Basic Facilities and Instruments for Mechanical Testing of Bone.................................................87

Christopher V. Bensen and Yuehuei H. An

Chapter 6

Methods of Evaluation for Bone Dimensions, Densities, Contents, Morphology,
and Structures ...............................................................................................................................103

Yuehuei H. An, William R. Bareld, and Ivars Knets

Chapter 7

General Considerations of Mechanical Testing.............................................................................119

Yuehuei H. An and Christopher V. Bensen

Chapter 8

A Hierarchical Approach to Exploring Bone Mechanical Properties...........................................133

C. Edward Hofer, Barbara R. McCreadie, Erica A. Smith, and Steven A. Goldstein

Chapter 9

Nondestructive Mechanical Testing of Cancellous Bone..............................................................151

Frank Linde and Ivan Hvid

Chapter 10

Synthetic Materials and Structures Used as Models for Bone .....................................................159

John A. Szivek

0266/ fm/frame Page 21 Thursday, October 21, 1999 1:47 PM

Section II Methods of Mechanical Testing of Bone

Chapter 11

Tensile and Compression Testing of Bone....................................................................................175

Tony S. Keller and Michael A. K. Liebschner

Chapter 12

Bending Tests of Bone...................................................................................................................207

Mandi J. Lopez and Mark D. Markel

Chapter 13

Torsional Testing of Bone..............................................................................................................219

Benjamin R. Furman and Subrata Saha

Chapter 14

Indentation Testing of Bone...........................................................................................................233

Brodie E. McKoy, Qian Kang, and Yuehuei H. An

Chapter 15

Penetration Testing of Bone Using an Osteopenetrometer ...........................................................241

Ivan Hvid and Frank Linde

Chapter 16

Microhardness Testing of Bone.....................................................................................................247

Sarandeep S. Huja, Thomas R. Katona, and W. Eugene Roberts

Chapter 17

Nanoindentation Testing of Bone ..................................................................................................257

Jae-Young Rho and George M. Pharr

Chapter 18
Single Osteon Micromechanical Testing.......................................................................................271
Maria-Grazia Ascenzi, Alessandro Benvenuti, and Antonio Ascenzi
Chapter 19
Micromechanical Testing of Single Trabeculae ............................................................................291
Peter L. Mente
Chapter 20
Strain Gauge Measurements from Bone Surfaces ........................................................................305
John A. Szivek and Vasanti M. Gharpuray
Chapter 21
Screw Pullout Test for Evaluating Mechanical Properties of Bone .............................................321
Matthew S. Crum, Franklin A. Young, Jr., and Yuehuei H. An
Chapter 22
Viscoelastic Properties of Bone and Testing Methods..................................................................329
Naoki Sasaki
0266/ fm/frame Page 22 Thursday, October 21, 1999 1:47 PM
Chapter 23
Observation of Material Failure Mode Using a SEM with a Built-In Mechanical
Testing Device................................................................................................................................349
Rong-Ming Wang and Yuehuei H. An
Chapter 24
Ultrasonic Methods for Evaluating Mechanical Properties of Bone............................................357
Jae-Young Rho
Chapter 25
Evaluating Mechanical Properties of Bone Using Scanning Acoustic Microscopy.....................371
Charles H. Turner and J. Lawrence Katz
Chapter 26
Peripheral Quantitative Computed Tomography for Evaluating Structural
and Mechanical Properties of Small Bone....................................................................................385
Jos Luis Ferretti
Chapter 27
Computer Modeling for Evaluating Trabecular Bone Mechanics ...............................................407
Rakesh Saxena and Tony S. Keller
Section III Methods of Mechanical Testing of the BoneImplant Interface
Chapter 28
Factors Affecting the Strength of the BoneImplant Interface.....................................................439
Brodie E. McKoy, Yuehuei H. An, and Richard J. Friedman
Chapter 29
Implant Pushout and Pullout Test..................................................................................................463
Aivars Berzins and Dale R. Sumner
Chapter 30
The Validity of a Single Pushout Test ...........................................................................................477
Wouter J. A. Dhert and John A. Jansen
Chapter 31
Tensile Testing of BoneImplant Interface ...................................................................................489
Takashi Nakamura and Shigeru Nishiguchi
Chapter 32
Fracture Toughness Tests of the BoneImplant Interface.............................................................499
Xiaodu Wang, Kyriacos A. Athanasiou, and C. Mauli Agrawal
Chapter 33
In Vitro Measurements of Implant Stability..................................................................................515
Aivars Berzins and Dale R. Sumner
0266/ fm/frame Page 23 Thursday, October 21, 1999 1:47 PM
Chapter 34
In Vitro Testing of the Stability of Acetabular Components.........................................................527
James R. Davis, Robert A. Lofthouse, and Riyaz H. Jinnah
Chapter 35
In Vitro Testing of the Stability of Femoral Components ............................................................541
Sanjiv H. Naidu, Fadi M. Khoury, and John M. Cuckler
Chapter 36
Screw Pullout Test .........................................................................................................................551
Lisa A. Ferrara and Timothy C. Ryken
Chapter 37
Finite Element Analysis for Evaluating Mechanical Properties of the
BoneImplant Interface..................................................................................................................567
Keith R. Williams
Chapter 38
Fatigue Testing of Bioabsorbable Screws in a Synthetic Bone Substrate....................................581
William S. Pietrzak, David R. Sarver, and David H. Kohn
Chapter 39
Testing Intervertebral Stability after Spinal Fixation....................................................................593
Kenneth S. James and A. U. Daniels
Appendix 1....................................................................................................................................607
Appendix 2....................................................................................................................................609
Index ..............................................................................................................................................611
0266/ fm/frame Page 24 Thursday, October 21, 1999 1:47 PM

Section I

General Considerations

0266/ ch01/frame Page 1 Thursday, October 21, 1999 1:52 PM

0266/ ch01/frame Page 2 Thursday, October 21, 1999 1:52 PM

3

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

1

Basic Composition and
Structure of Bone

Ermanno Bonucci

CONTENTS

I. Introduction ..............................................................................................................................3
II. Basic Components of Bone Matrix .........................................................................................4
A. Organic Matrix ...................................................................................................................4
1. Collagen Fibrils .............................................................................................................4
2. Noncollagenous Components........................................................................................4
B. Mineral Substance..............................................................................................................5
III. Whole Bone..............................................................................................................................6
IV. Types of Bone ..........................................................................................................................9
A. Compact Bone....................................................................................................................9
B. Cancellous Bone...............................................................................................................10
C. Woven and Parallel-Fibered Bone ...................................................................................10
1. Woven Bone ................................................................................................................10
2. Parallel-Fibered Bone..................................................................................................11
D. Primary and Secondary Bone ..........................................................................................11
V. Osteons and Trabeculae .........................................................................................................11
A. Osteons .............................................................................................................................11
1. Microradiography of Osteons .....................................................................................13
2. Osteon Remodeling.....................................................................................................13
3. Lamellae ......................................................................................................................13
B. Trabeculae.........................................................................................................................15
VI. Conclusions ............................................................................................................................16
Acknowledgments ............................................................................................................................16
References ........................................................................................................................................17

I. INTRODUCTION

Bone is a specialized tissue which, although apparently immobilized in a petried state, has
fundamental physiological functions. First, together with the intestine and kidney, it contributes to
the regulation of calcemia. This can, in fact, be decreased by the diversion of calcium ions from
the serum into the bone matrix during its mineralization, or be increased by the passage of calcium
ions from osteoclasts into the bloodstream during bone resorption. Besides this fundamental met-
abolic activity, which requires the participation of bone cells and systemic and/or local factors,

1

bone is devoted to vital mechanical functions.

2

Its hardness, moderate elasticity, and very limited
plasticity and brittleness make it an ideal tissue for standing and moving, i.e., for the insertion of
muscles, the formation of levers able to make muscles respond, and the protection of soft tissues

0266/ ch01/frame Page 3 Thursday, October 21, 1999 1:52 PM

4

Mechanical Testing of Bone and the BoneImplant Interface

and organs, including bone marrow. Moreover, the arrangement of its microstructures and the
presence of cavities in its interior give bone an optimum mass-to-strength ratio. These well-known
properties seem to be common to all skeletal segments, as the structure and composition of bone
appear at rst glance to be the same in all cases. On closer examination, bone turns out to be a
highly heterogeneous tissue; its composition and structure both vary in a way that depends on
skeletal site, physiological function, the age and sex of subjects, and the type of vertebrate species.
In contrast with this heterogeneity, the basic components of the tissue are remarkably consistent.

II. BASIC COMPONENTS OF BONE MATRIX

Calcied bone matrix contains two components: the organic matrix and the mineral substance.

A. O

RGANIC

M

ATRIX

The organic matrix of bone mainly consists of type I collagen brils, which account for over 90%
of the whole matrix, with the remaining 10% corresponding to noncollagenous proteins, proteogly-
cans, and phospholipids.

3

1. Collagen Fibrils

Several surveys

4-6

deal with the structure and composition of collagen brils, which are only summa-
rized here. Collagen brils are formed by the assemblage of lamentous molecules which are them-
selves made up of three polypeptide chains arranged in a helical conguration. These chains may
include a variety of amino acid sequences, so that the molecules may show diversity. Type I collagen
consists of relatively thick brils (mean diameter 78 nm) resulting from the assemblage of molecules
consisting of two

!

1

(I) chains and one

!

1

(II) chain. The molecules are assembled in such a way as
to give the brils a characteristic periodic banding (repetitive axial period of about 66.8 nm) when
examined under the electron microscope (Figures 1.1A,B). According to the classical bidimensional
model of Hodge and Petruska,

7

parallel collagen molecules overlap in such a way that they are
staggered by approximately a quarter of their length; the aligned chemical groups give rise to periodic
bands, and the quarter staggering to gaps or holes (Figure 1.1A). The molecules are stabilized by
intra- and intermolecular cross-links, which are essential for the tensile strength of the brils and their
mineralization.

8

The three-dimensional assemblage of collagen molecules in the bril is not known
precisely; the subject has been discussed in several surveys.

4,6,9

2. Noncollagenous Components

The noncollagenous components of bone include noncollagenous proteins, proteoglycans, phospho-
lipids, glycoproteins, and phosphoproteins. Detailed information can be found in several surveys.

10-12

Moreover, the calcied matrix contains growth factors

13

and enzymes such as alkaline
phosphatase

14,15

and metalloproteinases.

16,17

The distribution and amounts of noncollagenous pro-
teins (osteopontin, osteonectin, bone sialoprotein) are variable according to types of bone

18,19

and
zones of bone matrix where they are located: substantial amounts of bone sialoprotein and osteopon-
tin are present in the cement lines and in discrete interbrillary patches;

20-23

osteocalcin is found
in lamellar bone

19

; woven bone is unique in containing the bone acidic glycoprotein-75.

19

The
function of noncollagenous proteins seems to vary, too.

10,11,24-27

The degree of calcication may be
a contributing factor in this connection,

28

while calcication may itself be inuenced by noncol-
lagenous proteins.

29,30

Proteoglycans may have a regulatory effect, although their role as inhibitors
or promoters of the calcication process is still under discussion.

11,31

Phospholipids, which are
present in the calcifying matrix,

32

are also considered to have a signicant role in calcication.

33

0266/ ch01/frame Page 4 Thursday, October 21, 1999 1:52 PM

Basic Composition and Structure of Bone

5

Several growth factors are involved in bone cell differentiation and recruitment. One substance of
special interest is bone morphogenetic protein, which has osteoinductive properties and is found
in the calcied bone matrix.

34

B. M

INERAL

S

UBSTANCE



The mineral substance of bone is a calcium phosphate hydroxyapatite.

35

The problems arising
from its molecular structure, mechanism of formation, and deposition in bone matrix have been
discussed in several surveys.

3,35-37

Morphologically, the inorganic substance of bone, which appears
homogeneous under the light microscope, can be resolved into characteristic crystallites under
the electron microscope (Figure 1.1C). Surprisingly, there is no complete agreement about their
shape and dimensions, or about their relationship with the organic components.

38

The results of
electron microscope and X-ray diffraction studies have led to their being described as tablets,

39

or as tablet-, rod-, and needlelike crystals.

40

Needles and tablets may occur simultaneously
(Figure 1.1C), the former apparently deriving from the latter.

41

Investigation of the problem of the crystal shape and location in bone may contribute much to
an understanding of the biophysical properties of the tissue. The ultrastructural evidence that the
mineral substance forms bands corresponding to the collagen period

3,40

(Figures 1.1C and 1.2A)
and, more exactly, to the holes zones,

3

suggests that the crystals t exactly into the holes. This
would allow the crystals to be located inside the brils without disrupting their structure. However,
this possibility is called into question by the observation that with the progression of the calcication

FIGURE 1.1

(A) Diagram showing the arrangement of the collagen molecules according to the model of
Hodge and Petruska; note that the shifting of molecules gives rise to zones of holes (

h

) and zones of overlapping
(

o

). (B) Under the electron microscope, decalcied ultrathin sections show that the period of collagen brils
is due to the repetition of two main bands, which correspond to the

h

and

o

zones. Uranyl acetate and lead
citrate,

100,000. (C) During early calcication, the mineral substance forms bands corresponding to the

h

zones of the collagen brils. Note the presence of needlelike crystals. Unstained,

100,000.

0266/ ch01/frame Page 5 Thursday, October 21, 1999 1:52 PM

6

Mechanical Testing of Bone and the BoneImplant Interface

process the mineral periodic bands are obscured by needlelike and lament-like crystals
(Figure 1.2B) whose length may be greater than that of the holes:

40,42

if located inside the collagen
brils, they would have to perforate the walls of the holes and the bril molecular organization to
attain such a length.

43,44

As a result, although collagen solubility should increase with the degree
of calcication, bone collagen from fully calcied chick metatarsals has been found to be insoluble
in reagents which solubilize the collagen from soft tissues.

45

This discrepancy has been explained
by assuming that (1) the intrabrillary space is increased by the connection of holes with pores;

3

(2) crystal elongation can occur not only through crystal growth but also through the multiplication
of mineral particles which ll all the available space;

3

(3) only short, platelike crystals are located
in the bril holes, whereas the long needle-like crystals are located in the interbrillar spaces.

38

In
reality, the organicinorganic relationships in bone are still incompletely known. On the other hand,
they may vary according to the type of bone,

19

and may change during the calcication process.

46

This lack of information is not without consequences, because the strength, elasticity, and other
biophysical characteristics of bone largely depend on the amount of inorganic substance, the relative
loss of water and organic material during calcication, the integrity of collagen brils, and above
all the relationship between collagen brils and crystals. In any case, all the biophysical properties
of bone may be altered by treatments (dehydration, xation, embalming) which modify the organic
components of the bone matrix

47

or lead to the removal of the inorganic substance (decalcication).

48

III. WHOLE BONE

The variable assemblage of the basic components reported above leads to the formation of bone
as a tissue. Whole bone, i.e., bone as an organ, consists not only of calcied bone matrix and bone
cells, but also of nonosseous cells, blood vessels, nerve bres, and bone marrow, whose relative
proportions change with the type and age of the bone. Obviously, the physiological properties of
bone are closely dependent on the variable presence of these soft structures. However, they can be
neglected if bone is considered strictly as a calcied tissue.
Depending on the skeletal sites, bones may appear as long, tubular segments (long bones),
bilaminar plates (at bones), or short, irregularly prismatic structures (short bones). In long bones
three different regions can be distinguished (Figure 1.3): the diaphysis, that is, the central shaft

FIGURE 1.2

Electron microscope pictures of a low calcified (A) and fully calcified (B) bone matrix: at low
degree of calcication, the mineral substance gives rise to bands corresponding to the period of collagen
brils; as calcication increases, the periodic pattern is masked by lament- and needlelike crystals. Unstained,
(A)

36,000; (B)

80,000.

0266/ ch01/frame Page 6 Thursday, October 21, 1999 1:52 PM

Basic Composition and Structure of Bone

7

which represents the longest part; the epiphyses, which are present at the two extremities; and the
metaphyses, which lie between. A cartilaginous layer, the so-called growth cartilage, separates the
metaphysis from the epiphysis in the growing skeleton, but tends to disappear as the skeleton
matures. The long bones are present in the peripheral or appendicular skeleton, i.e., the limbs, and
in ribs and clavicles; the at bones are typically found in the skull, scapula, and pelvis; and the
short bones in the axial skeleton (vertebrae, sternum), carpus, and tarsus. With the exclusion of the
membranous bones (skull, clavicle, part of the mandible and of the facial skeleton) which develop
on the basis of a brous anlage, all other bones are formed by the ossication of a cartilaginous
model (endochondral ossication). Articular cartilage covers the part of the external epiphyseal
surface that has an articular function; the periosteum covers the outer bone surface.
Independently of their macroscopic anatomy, all skeletal segments consist of an outer layer of
compact bone (also called the compacta, or cortical bone) and an inner zone (the medulla), which
contains bone marrow (see Figure 1.3). The relative proportion between the compacta and the
medulla varies with the skeletal segments and their function (Figures 1.3, 1.4, 1.5), but the maximum
strength-to-weight ratio retains its validity. The diaphysis of long bones shows a thick compacta,
in which about 90% of the volume is calcied, and a medulla, which corresponds to an axial, more
or less eccentric cylinder containing red, hemopoietic bone marrow in youth, and yellow, fat-
repleted, nonhemopoietic marrow in adults (see Figure 1.3). The thickness of the compacta dimin-
ishes as the ared metaphyseal zone is approached (see Figure 1.5). The at and short bones (see
Figure 1.4), as well as the epiphysis and metaphysis of the long bones (see Figure 1.5), have a thin
outer compacta. The medulla is prevalent and consists of a frame of interlacing laminar or rodlike
osseous trabeculae (see Figures 1.4, 1.5) which delimit irregular, intercommunicating spaces con-
taining bone marrow, blood vessels, nerve bres, and other soft structures. For this reason, the area
covered by the medulla is much greater than that covered by the compacta, although the calcied
matrix covers only 20 to 25% of the total volume in the former, but as much as 95% in the latter.

FIGURE 1.3

Schematic representation of the upper third of the tibia, an example of long bone; i.c.s. = inner
circumferential system; o.c.s. = outer circumferential system.

0266/ ch01/frame Page 7 Thursday, October 21, 1999 1:52 PM

8

Mechanical Testing of Bone and the BoneImplant Interface

FIGURE 1.4

Section of the body of a lumbar vertebra, showing the thin outer compacta (left and right side of
the vertebral section), and the vertical and horizontal trabeculae which form the spongy bone of the medulla. The
upper and lower surfaces correspond to articular cartilage, which is in continuity with the intervertebral disk.

FIGURE 1.5

(A) Section of one half of the upper third of the tibia and (B) its microradiograph. The spongy
bone of the metaphysis consists of comparatively thick vertical trabeculae connected by thin transverse
trabeculae; the thickness of the compacta decreases from the diaphysis to the metaphysis. The degree of
calcication is rather uniform.

0266/ ch01/frame Page 8 Thursday, October 21, 1999 1:52 PM

Basic Composition and Structure of Bone

9

IV. TYPES OF BONE

The outer compacta of the skeletal segments consists of compact bone (see Figures 1.3 through
1.5); the inner medulla corresponds to the bone marrow cylinder in long bones, to interlacing
osseous trabeculae in short and at bones, and in long bone epiphyses (see Figures 1.4 and 1.5).
The osseous trabeculae form the spongy, cancellous, or trabecular bone. These denominations do
not refer to a basic difference in the aggregation state of the bone tissue. Although the true density
of fully calcied cancellous bone is a little lower (3%), and its proteoglycan content a little greater,
than those of the fully calcied compact bone and in spite of other minor differences, the two types
of bone (compact and trabecular) have a very similar basic composition and degree of aggregation
(however, see below for the differences between lamellar and woven bone). The real difference
between compact and spongy bone depends on its porosity: that of compact bone, mainly due to
the voids provided by osteon canals, Volkmanns canals, osteocytes and their canaliculi, and
resorption lacunae (see below), varies from 5 to 30% (apparent density about 1.8 g/cm

3

); the porosity
of cancellous bone, chiey due to the wide vascular and bone marrow intertrabecular spaces, ranges
from 30 to more than 90% (apparent density 0.1 to 0.9 g/cm

3

).

49

A. C

OMPACT

B

ONE

Starting as soon as a few months after birth (for bone structure and osteogenic activity during
embryonic and fetal life see the survey by Gardner

50

), a cross section cut through the middiaphysis
of a long bone shows that the cortical, compact bone is structured in three main concentric systems:
an outer circumferential system, an intermediate osteonic area, and an inner circumferential system
(see Figure 1.3). The names of these systems are due to their microscopic conguration. Both
circumferential systems consist of concentric layers of bone trabeculae running parallel to the
periosteal (outer system) or endosteal (inner system) surface (Figure 1.6). Both the inner and outer
systems are conspicuous at birth (see primary bone) and are then partly substituted by osteonic, or
secondary, bone (Figure 1.7), so that their width progressively decreases; in particular, the inner
system may become inconspicuous in the elderly.

49

The intermediate area, which enlarges with age
at the expense of the circumferential systems, contains Haversian systems or osteons (see below,
and Figure 1.8). The thin compacta of the at and short bones shows both trabeculae and osteons,
but the typical three-system arrangement of the long bones is not apparent.

FIGURE 1.6

Microradiography of primary bone in the shaft of an embryonal long bone. Note the parallel
trabeculae, and the interposed vascular canals, of the inner circumferential system (bone marrow canal partly
visible above),

15.

0266/ ch01/frame Page 9 Thursday, October 21, 1999 1:52 PM

10

Mechanical Testing of Bone and the BoneImplant Interface

B. C

ANCELLOUS

B

ONE

Cancellous, or spongy, or trabecular bone is present in the medulla of at and, above all, short
bones, and in the epiphysis and metaphysis of long bones. It is almost absent in the central part of
the diaphysis, whereas increasing numbers of trabeculae protrude from the inner, endosteal surface
into the marrow cavity as the metaphyses are approached. The name of this type of bone derives
from its architecture; it consists of trabeculae, that is, osseous structures having a sheetlike, tubular,
or barlike conguration, which interlace and anastomose to form a cancellous (lattice-like) or
spongy structure (see Figures 1.4 and 1.5). As shown by the variability of the apparent density (see
above), the dimensions of the holes in this lattice are extremely variable, the variation increasing
with age.

51

C. W

OVEN



AND

P

ARALLEL

-F

IBERED

B

ONE

According to the arrangements of the collagen brils, both compact and spongy bones can be of
woven or parallel-bered type.

52

1. Woven Bone

Woven bone, also called coarse-ber bone, is characterized by the presence in the matrix of coarse,
irregularly oriented collagen brils which, under the polarization microscope, appear as uneven

FIGURE 1.7

Cross section of the diaphysis of neonatal long bone as seen under the light (A) and polarization
(B) microscope: the primary, parallel-bered bone (which crosses the picture obliquely) is partly substituted
by secondary osteonic bone. The arrow points to still unrepaired resorption lacuna; arrowheads point to two
osteons, which are still under construction as shown by their wide vascular canals. Unstained section prepared
by grinding,

35.

0266/ ch01/frame Page 10 Thursday, October 21, 1999 1:52 PM

Basic Composition and Structure of Bone

11

anisotropic structures. They encircle osteocyte lacunae of globular shape which are irregularly
distributed in the matrix (osteocyte morphology and function have been reviewed elsewhere

1,53-55

).
Moreover, they surround relatively large canal-like structures which are penetrated by capillary
vessels. When these canals are small, they correspond to primary osteons (see below) and the
appearance of bone is compact; when they are conspicuous, their appearance is that of spongy
bone and the intertrabecular spaces contain bone marrow elements. The matrix of woven bone is
characterized by the presence of wide interbrillary spaces which contain abundant interbrillar
noncollagenous material

18

and relatively abundant proteoglycans, as shown by its metachromasia
after toluidine blue staining.

56

Moreover, woven bone is unique in containing BAG-75 (bone acidic
glycoprotein-75).

19

Woven bone is the bone formed rst during skeletal embryogenesis. After birth,
it is gradually removed by the process of bone remodeling, and is substituted by lamellar bone.

57

It can, however, be formed again in pathological conditions, such as callus formation, bone tumors,
and ectopic ossication. This is in line with the suggestion that woven and lamellar bone are the
result of a rapid and a slow osteogenic process, respectively.

58

2. Parallel-Fibered Bone

Parallel-bered bone consists of relatively thin, parallel-oriented collagen brils. In reality, their
orientation is only prevalently parallel, because many of them interlace during their course

59

;
moreover,



they may be regularly organized into unit layers called lamellae (see below). Because
collagen brils in parallel-bered bone are rather uniformly oriented, they appear anisotropic when
their axis is perpendicular to the optical axis of the polarization microscope (see Figures 1.7B). In
this type of bone, as well as in lamellar bone, osteocytes have an elongated, ovoidal shape; however,
they do not show the regular distribution they have in lamellar bone.

55

D. P

RIMARY



AND

S

ECONDARY

B

ONE

Because of its appearance early in life, woven bone is often called primary, or immature, bone.
However, primary and woven bone are not synonymous, because the former may consist of collagen
brils that run parallel to each other, with an orderly arrangement (see parallel-bered bone, and
Figures 1.6 and 1.7). On the other hand, lamellar bone is also called secondary or mature bone,
because it replaces bone of woven or primary type as it is resorbed after birth, and because it is
formed at the end of the life span of the basic multicellular units during the process of bone
remodeling.

60,61

As already mentioned, lamellar bone can be found in primary bone. Both primary
and secondary bone may be structured as osteons or trabeculae.

V. OSTEONS AND TRABECULAE
A. O

STEONS

As mentioned, osteons, or Haversian systems, can be found both in primary and secondary bone.
Primary osteons are formed as the primitive vascular spaces are lled up by growing bone. As a
result, they merge into the surrounding bone matrix without any precise delimitation, and repeat
its structural characteristics. Secondary osteons are formed during, and because of, bone remodeling.
They are new entities which substitute previously formed structures removed by osteoclastic
resorption. For this reason, they are separated from the surrounding matrix by a reversal, or cement
line (see below, osteon remodeling).
Osteons (see Figure 1.8) are more or less regular cylindrical structures, whose length, although
hardly measurable and varying according to animal species and age, ranges from 3 to 6 mm, and
can reach 12 mm in branched osteons.

62

Their mean cross-sectional area varies too, and changes
with the animal species and bone site;

63

it has been reported to be 5.16

5.12

"

10

4

cm

2

in human
tibial cortical bone,

64

or 17.45

m
2
in males and 17.62 m
2
in females.
65
In horse metacarpal bone
0266/ ch01/frame Page 11 Thursday, October 21, 1999 1:52 PM
12 Mechanical Testing of Bone and the BoneImplant Interface
the osteon diameter shows a range comprising 156 m in the dorsal region, 179 m in the medial
region, and 182 m in the lateral region.
63
The osteons are oriented parallel to the axis of bone,
66
but with an inclination varying from 5 to 15,
66
so that they have a spiral diaphyseal distribution.
68-70
Their major axis consists of a vascular canal (or Haversian canal). Because osteons are built in a
centripetal way, at the beginning of their formation the canal is wide (see Figures 1.7 and 1.8),
with the inner wall lined by a layer of plump, active osteoblasts (for osteoblast morphology and
function see Marks and Popoff,
1
Scherft and Groot,
71
and Marks
72
); subsequently, the canal gradually
narrows, and the osteoblasts become at,
73
so that at the end of the process the canal is bordered
FIGURE 1.8 Cross-sectioned osteons of compact bone as seen (A) under the light microscope; (B) in a
microradiograph; and (C) under the polarization microscope (arrows point to the same osteon). Note in (B)
that osteons have different degrees of microradiographic density, i.e., of calcication, that the less calcied
ones have the largest Haversian canals (arrowheads), that the highest degree of calcication is that of the
interstitial bone (asterisks), and that the border of the Haversian canals is apparently hypermineralized. Note
in (C) that most of the osteons are of intermediate type. Unstained section prepared by grinding, 35.
0266/ ch01/frame Page 12 Thursday, October 21, 1999 1:52 PM
Basic Composition and Structure of Bone 13
by so-called lining cells.
74
Adjacent osteons may anastomose; moreover, they can be connected by
other vascular canals (Volkmanns canals) which arise from the periosteum or the endosteum and
run obliquely or transversally through the bone.
1. Microradiography of Osteons
The discovery of microradiography
75
cast new light on the physiology of bone, by allowing the
demonstration that osteon calcication is variable: it is at a maximum in primary osteons, whose
calcication is rapid and occurs soon after their formation (the same is true of woven bone), while
it can range from a minimum to a maximum in secondary osteons, where a sudden but partial
initial calcication of the organic matrix is followed by slow mineral deposition which lasts until
completeness is reached
75
(see Figure 1.8B). Due to the process of bone remodeling, osteons are
renewed continuously. As a result, osteons at different degrees of calcication are always present
in adult compact bone: those at the initial stage of formation are less calcied and more transparent
to X rays than those at the nal stage of formation, which are fully calcied; intermediate patterns
are easily recognizable (see Figure 1.8B). Because of their different degrees of calcication, osteons
and other bone structures can be distinguished on the basis of their different microhardness,
76
and
can be separated by means of density gradient fractionation.
28,77,78
The degree of microradiographic density may not be uniform in single osteons. The borders
of their vascular canals often appear to be hypermineralized (see Figure 1.8B), a condition that has
been thought to be an early sign of bone matrix destruction or delitescence.
79
Moreover, adjacent
lamellae may display different degrees of X-ray density.
80,81
Both these microradiographic patterns
may depend on X-ray scattering rather than on differences in mineral content.
82
2. Osteon Remodeling
Although secondary osteons are already present in bone during the last period of intrauterine life,
most of them develop after birth.
57
This means that they replace the already present primary bone,
while they in their turn will be replaced by other secondary osteons during the continuous process
of bone remodeling. Because resorption is usually partial and asymmetric, osteon remains persist
between the newly formed ones (see Figures 1.8A,B). These remains appear as irregularly polygonal
fragments of fully calcied lamellar matrix, which connect osteons at various degrees of calcica-
tion. They represent the so-called interstitial bone. The new osteons are surrounded, with separation
from the interstitial bone, by a continuous, laminar structure, called the cement or reversal line.
This provides proof that the osteon formation has been preceded by bone resorption, because the
matrix of the cement line is the rst to be laid down on the wall of the Howships lacunae at the
end of osteoclast activity (the structure, morphology, and function of osteoclasts have recently been
reviewed
1,83-87
). Cement lines are reported to be highly mineralized;
88
however, they seem to contain
signicantly less Ca and P and more S, and to have a lower Ca/P ratio, than the surrounding bone
matrix.
89
Moreover, they contain substantial amounts of bone sialoprotein
23
and osteopontin.
21
This
may act as an interfacial adhesion promoter and may help to maintain the integrity of the tissue
and regulate cell dynamics.
20,90
3. Lamellae
The arrangement and orientation of collagen brils in secondary bone may give rise to lamellae.
For this reason, it is often called lamellar bone, although lamellae may be found in primary bone
(see above). Lamellae can be seen under the light microscope if thin sections are examined
(Figure 1.9A), and are best recognized under the polarization (Figures 1.8C and 1.9B) and electron
microscopes (Figure 1.9C). However, the organization of collagen brils in them, and above all in
osteons, is still controversial. When examined under a polarization microscope, lamellar and
osteonic bone may show an alternation of isotropic and anisotropic lamellae which implies a
0266/ ch01/frame Page 13 Thursday, October 21, 1999 1:52 PM
14 Mechanical Testing of Bone and the BoneImplant Interface
different orientation of the collagen brils in successive lamellae (Figures 1.8C and 1.9B). This
optical pattern was explained by Gebhardt
91
by assuming that the collagen brils have a helical
course around the osteon axis, and that the brils which have the same spiral orientation form a
lamella; the bril course in one lamella is opposite and more or less perpendicular to the course
in the adjacent lamellae. On the basis of this classic model, and of the ndings obtained by
polarization microscopy, three types of osteons have been described in cross sections of diaphyseal
bone: bright, i.e., anisotropic; dark, i.e., isotropic; and intermediate, i.e., with alternate bright and
dark lamellae.
92
These optical patterns have been considered to depend on the arrangement of
collagen brils in successive lamellae, corresponding to a transversal spiral course in all lamellae
(bright osteons), a longitudinal spiral course in all lamellae (dark osteons), and a transversal spiral
course in one lamella and a longitudinal spiral course in the next (alternate osteons).
92
Obviously,
these types constitute paradigms of the different lamellar arrangements the osteon can have and
refer to extreme, sharply differentiated models, without consideration of the intermediate arrange-
ments the lamellae may have. If these are examined in depth, other osteon models can be described
on the basis of their birefringence.
63,93
On the other hand, even the early studies (reviewed by Frank
et al.
94
and Smith
95
) slightly modied Gebhardts model. By considering both staining and optical
properties, Smith
95
described three types of lamellar arrangement in osteons: type I, characterized
FIGURE 1.9 (A) Cross-section of part of an intermediate osteon (vascular canal below), showing the
lamellar arrangement of its matrix. Thin section (about 1 m thick) stained with Azure IIMethylene blue,
450. (B) Cross section of intermediate osteons, which show alternating bright and dark lamellae under the
polarization microscope, 120. (C) Electron microscope picture of an intermediate osteon, showing alter-
nating thin and thick lamellae. Unstained, 6000.
0266/ ch01/frame Page 14 Thursday, October 21, 1999 1:52 PM
Basic Composition and Structure of Bone 15
by alternate lamellae of similar density with longitudinal or transversal brils; type II, the same as
type I, but with different lamellar density; type III, without any evident lamellar arrangement. On
the basis of transmission electron microscope ndings, two lamellar arrangements were considered
possible, one consisting of dense lamellae alternating with arched lamellae, another showing a
herringbone pattern.
94
A goniometric control of the ultrastructural ndings has prompted the sug-
gestion that the various lamellar patterns partly depend on the direction of the section plane, and
that the lamellae derive from the arrangement of the brils according to either a cylindrical twisted
plywood or a cylindrical orthogonal plywood, which may coexist in the same osteon.
96
The
twisted plywood pattern could account for the alternate appearance the osteons can have under
the polarization microscope. Five arrays of parallel collagen brils, each offset by 30, have been
proposed as components of individual lamellae in a plywoodlike structure.
97
On the basis of
scanning electron microscope studies, several models have been proposed: the classic alternate
osteon pattern and the twisted plywood pattern;
98
alternation of bril orientation in adjacent lamel-
lae, with domains of differently oriented brils in the same lamella, and with brils connecting
adjacent lamellae;
99
thin and thick lamellae with a structure like rotated plywood;
100
variable
brillar architectures in the same osteon;
101
interlamellar transitional zones, in which both the brils
and the crystals have an intermediate orientation.
102
Again, on the basis of SEM results, lamellar
bone has been considered to consist of two types of lamellae, one dense, or collagen rich, another
loose, or collagen poor, the former thinner and less calcied than the latter, the only one containing
osteocyte lacunae, both showing a highly interwoven arrangement of collagen brils.
59,103-105
Whatever the ultrastructural pattern of the lamellae may be, the polarization microscope permits
the selection of differently structured osteons whose mechanical properties are closely related to
their optical characteristics and lamellar conformation. In agreement with theoretically predictable
results, dark, isotropic osteons best resist tensile forces,
106
whereas bright, anisotropic osteons best
resist compressive ones.
92
Other mechanical properties of lamellar bone are consistent with this
structurefunction relationship.
107
Also the distribution of different osteon types within the skeleton
reects the forces which predominate in a specic skeletal segment.
70,107,108
B. TRABECULAE
Trabeculae are the unit components of the cancellous bone. Usually, they are described as rod- or
sheetlike structures. In the calcaneus, both types of trabeculae are present; however, many (up to
83%) of those of rodlike conformation are tubular, due to the vascular canal running through them,
so that they are similar to Haversian systems.
109
Microradiography shows different levels of calci-
cation in these trabeculae; the polarization microscope shows that in their outer portion the collagen
brils mostly run parallel to the long axis, whereas in the inner, osteonic portion they run perpen-
dicular. This basic structure is probably common to the tubular trabeculae of all spongy bones,
because it has been found in the cancellous bone of the mastoid, and the epiphysis of the femur
and tibia.
110
The microarchitecture of spongy bone appears random; however, the connections and orienta-
tion of the trabeculae are found to have precise patterns which are believed to be related to the
specic mechanical properies. The structure of spongy bone in the head and neck of the femur is
classically put forward as an example of the correlation between the orientation of the trabeculae
and the linear distribution of the principal forces during load bearing so-called stress trajectoral
theory.
111
The possibility that the distribution of trabeculae depends on the prevailing direction of
the mechanical forces has been viewed by some authors
112
as being in line with the mathematical
calculations, but it has been evaluated cautiously by others, especially because of the complicated
effects the traction of muscles may have on the overall load.
113
In any case, there is a close
relationship between the numbers and arrangement of trabeculae and the strength of spongy
bone.
114,115
This is conrmed by the fractures which can follow the age-induced loss of trabeculae,
whose total volume can fall, at least in the iliac crest, from about 25% in youth to about 12% in
0266/ ch01/frame Page 15 Thursday, October 21, 1999 1:52 PM
16 Mechanical Testing of Bone and the BoneImplant Interface
the elderly.
116-121
The loss is rather selective, as is shown by the falling frequency of transverse
trabeculae and the persistence of vertical ones in the central zone of the osteoporotic vertebral bodies,
122
and by the total disappearance of individual trabeculae in elderly women, and the generalized, sharp
fall in their numbers in elderly men.
51,123
This selective effect may be very dangerous because it causes
not only a fall in bone volume, but also a breakdown in bones connectivity,
114
that is, the trabecular
frame which greatly contributes to the strength of spongy bone.
124,125
The possibility that trabeculae may be lost in a selective way introduces an important concept,
which is that spongy bone contains some bundles of trabeculae whose main function is that of
resisting mechanical forces, while others have chiey metabolic functions. These last must not be
confused with those of the spongy bone which is classically considered metabolic, on account
of its being completely devoid of mechanical function, for example, the medullary bone of egg-
laying birds. This bone is formed under estrogen stimulation, represents a reserve of calcium for
eggshell formation, and is characterized by high amounts of acid proteoglycans and glycoproteins
in its matrix.
126
The possibility that the trabeculae of the metaphyseal spongy bone may be pre-
ponderantly metabolic or mechanic is in line with their behavior under different stimulations.
Thus, in growing animals the osteogenic activity is higher in the peripheral (or tubular) than in
the central (or lamellar) spongiosa, whereas in adult animals, in which osteogenesis mainly
depends on bone metabolism, the results are just the opposite.
127
The already mentioned loss of
transverse trabeculae in the central zone of aged vertebral bodies
122
points in the same direction.
The same is true of the osteoporosis which follows ovariectomy in rats: the fall in the concentration
of estrogens stimulates bone resorption in the central metaphyseal trabeculae, whereas the peripheral
trabeculae are to a large extent preserved.
128-130
Thus, not only is spongy bone more active meta-
bolically, and more susceptible to variations in mechanical conditions, than compact bone,
19
but
inside it single trabecular zones may have metabolic rather than mechanical functions.
VI. CONCLUSIONS
Bone is a heterogeneous tissue because its basic components are assembled in different ways, the
main structural determinants being the type of bone, age, loads, and metabolic activity. The type
of bone essentially depends on the density of its structures, which ranges from a very compact
state in cortical bone to a spongelike appearance in cancellous bone, and from a compact aggregation
of collagen brils in lamellar bone to their comparatively loose state in woven bone. Age chiey
induces a transformation of woven bone into lamellar bone soon after birth, and a late, almost
physiological loss in bone volume with alterations to connectivity.
131
The mechanical functions are
responsible for the maintenance of the architecture of bone,
132,133
and the metabolic ones for its
renewal
1
; both are mediated by the various aspects of bone remodeling, which induces the erosion
and reconstruction of trabecular segments in spongy bone,
131
and of osteons and other lamellar
structures in compact bone.
134
The heterogeneity of the tissue fully justies the study of single
components dissected and isolated from the bone context, such as osteons, trabeculae, and lamel-
lae.
135
Clearly, structural heterogeneity by itself greatly contributes to osseous properties such as
stiffness, elasticity, hardness, and strength, which are necessarily those of bone as a whole. The
study of a tissue as complex as bone calls for careful analysis of all its variables.
ACKNOWLEDGMENTS
The preparation of this chapter and the personal research mentioned supported by grants from the
Italian Ministry of University and Scientic and Technological Research (MURST), the Italian
National Research Council (CNR), and the University of Rome La Sapienza. The author is grateful
to Paola Ballanti, Silvia Berni, Carlo Della Rocca, and, above all, Giuliana Silvestrini for their
suggestions, discussions, and technical assistance.
0266/ ch01/frame Page 16 Thursday, October 21, 1999 1:52 PM
Basic Composition and Structure of Bone 17
REFERENCES
1. Marks, S.C. and Popoff, S.N., Bone cell biology: the regulation of development, structure, and function
in the skeleton, Am. J. Anat., 183, 1, 1988.
2. Keaveny, T.M. and Hayes, W.C., Mechanical properties of cortical and trabecular bone, in Bone, Hall,
B.K., Ed., CRC Press, Boca Raton, FL, 1993, 285.
3. Glimcher, M.J., Composition, structure, and organization of bone and other mineralized tissues and
the mechanism of calcication, in Handbook of Physiology: Endocrinology, Greep, R.O. and Astwood,
E.B., Eds., American Physiological Society, Washington, D.C., 1976, 25.
4. Chapman, J.A. and Hulmes, D.J., Electron microscopy of the collagen bril, in Ultrastructure of the
Connective Tissue Matrix, Ruggeri, A. and Motta, P.M., Eds., Martinus Nijhoff Publishers, Boston,
1984, 1.
5. van der Rest, M., The collagens of bone, in Bone, Vol. 3: Bone Matrix and Bone Specic Products,
Hall, B.K., Ed., CRC Press, Boca Raton, FL, 1991, 187.
6. Veis, A. and Sabsay, B., The collagen of mineralized matrices, in Bone and Mineral Research, Vol.
5, Peck, W.A., Ed., Elsevier Science Publishers, Amsterdam, 1987, 1.
7. Hodge, A.J. and Petruska, J.A., Recent studies with the electron microscope on ordered aggregates
of the tropocollagen molecules, in Aspects of Protein Structure, Ramachandran, G.N., Ed., Academic
Press, London, 1963, 289.
8. Yamauchi, M., Chandler, G.S., and Katz, E.P., Collagen cross-linking and mineralization, in Chemistry
and Biology of Mineralized Tissues, Slavkin, H. and Price, P., Eds., Elsevier Science Publishers,
Amsterdam, 1992, 39.
9. Chapman, J.A., Molecular organization in the collagen bril, in Connective Tissue Matrix, Hukins,
D.W.C., Ed., Macmillan, London, 1984, 89.
10. Boskey, A.L., Noncollagenous matrix proteins and their role in mineralization, Bone Miner., 6, 111,
1989.
11. Butler, W.T., Matrix macromolecules of bone and dentin, Collagen Relat. Res., 4, 297, 1984.
12. Seyedin, S.M. and Rosen, D.M., Matrix proteins of the skeleton, Curr. Opinion Cell Biol., 2, 914, 1990.
13. Seyedin, S.M. and Rosen, D.M., Unique bone-derived cytokines, in Cytokines and Bone Metabolism,
Gowen, M., Ed., CRC Press, Boca Raton, FL, 1992, 109.
14. Bonucci, E., Silvestrini, G., and Bianco, P., Extracellular alkaline phosphatase activity in mineralizing
matrices of cartilage and bone: ultrastructural localization using a cerium-based method, Histochem-
istry, 97, 323, 1992.
15. de Bernard, B., Bianco, P., Bonucci, E., et al., Biochemical and immunohistochemical evidence that
in cartilage an alkaline phosphatase is a Ca
2+
-binding glycoprotein, J. Cell Biol., 103, 1615, 1986.
16. Eeckhout, Y. and Delaisse, J.M., The role of collagenase in bone resorption. An overview, Pathol.
Biol., 36, 1139, 1988.
17. Jilka, R.L., Procollagenase associated with the noncalcied matrix of bone and its regulation by
parathyroid hormone, Bone, 10, 353, 1989.
18. Bonucci, E. and Silvestrini, G., Ultrastructure of the organic matrix of embryonic avian bone after
en bloc reaction with various electron-dense stains, Acta Anat., 156, 22, 1996.
19. Gorski, J.P., Is all bone the same? Distinctive distributions and properties of non-collagenous matrix
proteins in lamellar vs. woven bone imply the existence of different underlying osteogenic mecha-
nisms, Crit. Rev. Oral Biol. Med., 9, 201, 1998.
20. McKee, M.D., Farach-Carson, M.C., Butler, W.T., et al., Ultrastructural immunolocalization of non-
collagenous (osteopontin and osteocalcin) and plasma (albumin and
!2
HS-glycoprotein) proteins in
rat bone, J. Bone Miner. Res., 8, 485, 1993.
21. McKee, M.D., Glimcher, M.J., and Nanci, A., High-resolution immunolocalization of osteopontin and
osteocalcin in bone and cartilage during endochondral ossication in the chicken tibia, Anat. Rec.,
234, 479, 1992.
22. McKee, M.D. and Nanci, A., Postembedding colloidal-gold immunocytochemistry of noncollagenous
extracellular proteins in mineralized tissues, Micros. Res. Tech., 31, 44, 1995.
23. Riminucci, M., Silvestrini, G., Bonucci, E., et al., The anatomy of bone sialoprotein immunoreactive
sites in bone as revealed by combined ultrastructural histochemistry and immunohistochemistry,
Calcif. Tissue Int., 57, 277, 1995.
0266/ ch01/frame Page 17 Thursday, October 21, 1999 1:52 PM
18 Mechanical Testing of Bone and the BoneImplant Interface
24. Bianco, P., Ultrastructural immunohistochemistry of noncollagenous proteins in calcied tissues, in
Ultrastructure of Skeletal Tissues, Bonucci, E. and Motta, P.M., Eds., Kluwer Academic Publishers,
Boston, 1990, 63.
25. Ingram, R.T., Clarke, B.L., Fisher, L.W., and Fitzpatrick, L.A., Distribution of noncollagenous proteins
in the matrix of adult human bone: evidence of anatomic and functional heterogeneity, J. Bone Miner.
Res., 8, 1019, 1993.
26. Silbermann, M., Von der Mark, H., and Heinegard, D., An immunohistochemical study of the distri-
bution of matrical proteins in the mandibular condyle of neonatal mice. II. Non-collagenous proteins,
J. Anat., 170, 23, 1990.
27. Sommer, B., Bickel, M., Hofstetter, W., and Wetterwald, A., Expression of matrix proteins during the
development of mineralized tissues, Bone, 19, 371, 1996.
28. Mbuyi-Muamba, J.M., Dequeker, J., and Gevers, G., Collagen and non-collagenous proteins in dif-
ferent mineralization stages of human femur, Acta Anat., 134, 265, 1989.
29. Cowles, E.A., DeRome, M.E., Pastizzo, G., et al., Mineralization and the expression of matrix proteins
during in vivo bone development, Calcif. Tissue Int., 62, 74, 1998.
30. Fisher, L.W. and Termine, J.D., Noncollagenous proteins inuencing the local mechanisms of calci-
cation, Clin. Orthop., 200, 362, 1985.
31. Buckwalter, J.A., Proteoglycan structure in calcifying cartilage, Clin. Orthop., 172, 207, 1983.
32. Silvestrini, G., Zini, N., Sabatelli, P., et al., Combined use of malachite green xation and PLA
2
-gold
complex technique to localize phospholipids in areas of early calcication of rat epiphyseal cartilage
and bone, Bone, 18, 559, 1996.
33. Boyan, B.D., Schwartz, Z., Swain, L.D., and Khare, A., Role of lipids in calcication of cartilage,
Anat. Rec., 224, 211, 1989.
34. Sakou, T., Bone morphogenetic proteins: from basic studies to clinical approaches, Bone, 22, 591,
1998.
35. Posner, A.S., Bone mineral and the mineralization process, in Bone and Mineral Research, Vol. 5,
Peck, W.A., Ed., Elsevier Science Publishers, Amsterdam, 1987, 65.
36. Eanes, E.D., Dynamics of calcium phosphate precipitation, in Calcication in Biological Systems,
Bonucci, E., Ed., CRC Press, Boca Raton, FL, 1992, 1.
37. Glimcher, M.J., Mechanism of calcication: role of collagen brils and collagenphosphoprotein
complexes in vitro and in vivo, Anat. Rec., 224, 139, 1989.
38. Bonucci, E., Role of collagen brils in calcication, in Calcication in Biological Systems, Bonucci,
E., Ed., CRC Press, Boca Raton, FL, 1992, 19.
39. Robinson, R.A., An electron-microscopic study of the crystalline inorganic component of bone and
its relationship to the organic matrix, J. Bone Joint Surg., 34-A, 389, 1952.
40. Ascenzi, A., Bonucci, E., and Steve Bocciarelli, D., An electron microscope study of osteon calci-
cation, J. Ultrastruct. Res., 12, 287, 1965.
41. Hhling, H.J., Kreilos, R., Neubauer, G., and Boyde, A., Electron microscopy and electron micro-
scopical measurements of collagen mineralization in hard tissues, Z. Zellforsch., 122, 36, 1971.
42. Ascenzi, A., Bonucci, E., Ripamonti, A., and Roveri, N., X-ray diffraction and electron microscope
study of osteons during calcication, Calcif. Tissue Res., 25, 133, 1978.
43. Bonucci, E., The locus of initial calcication in cartilage and bone, Clin. Orthop., 78, 108, 1971.
44. Bonucci, E., The structural basis of calcication, in Ultrastructure of the Connective Tissue Matrix,
Ruggeri, A. and Motta, P.M., Eds., Martinus Nijhoff, Boston, 1984, 165.
45. Glimcher, M.J. and Katz, E.P., The organization of collagen in bone: the role of noncovalent bonds
in the relative insolubility of bone collagen, J. Ultrastruct. Res., 12, 705, 1965.
46. Pugliarello, M.C., Vittur, F., de Bernard, B., et al., Chemical modications in osteones during calci-
cation, Calcif. Tissue Res., 5, 108, 1970.
47. Evans, F.G., Factors affecting the mechanical properties of bone, Bull. N.Y. Acad. Med., 49, 751, 1973.
48. Bonucci, E. and Reurink, J., The ne structure of decalcied cartilage and bone: a comparison between
decalcication procedures performed before and after embedding, Calcif. Tissue Res., 25, 179, 1978.
49. Carter, D.R. and Spengler, D.M., Mechanical properties and composition of cortical bone, Clin.
Orthop., 135, 192, 1978.
50. Gardner, E., Osteogenesis in human embryo and fetus, in The Biochemistry and Physiology of Bone,
Bourne, G.H., Ed., Academic Press, New York, 1971, 77.
0266/ ch01/frame Page 18 Thursday, October 21, 1999 1:52 PM
Basic Composition and Structure of Bone 19
51. Aaron, J.E., Makins, N.B., and Sagreiya, K., The microanatomy of trabecular bone loss in normal
aging men and women, Clin. Orthop., 215, 260, 1987.
52. Smith, J.W., Collagen ber patterns in mammalian bone, J. Anat., 94, 329, 1960.
53. Aarden, E.M., Burger, E.H., and Nijweide, P.J., Function of osteocytes in bone, J. Cell. Biochem.,
55, 287, 1994.
54. Bonucci, E., The ultrastructure of the osteocyte, in Ultrastructure of Skeletal Tissues, Bonucci, E. and
Motta, P.M., Eds., Kluwer Academic, Boston, 1990, 223.
55. Marotti, G., Can, V., Palazzini, S., and Palumbo, C., Structurefunction relationships in the osteocyte,
Ital. J. Miner. Electrol. Metab., 4, 93, 1990.
56. Hancox, N.M., Biology of Bone, Cambridge University Press, Cambridge, U.K., 1972, Chapter 3.
57. Amprino, R., and Bairati, A., Processi di ricostruzione e di riassorbimento nella sostanza compatta
delle ossa delluomo. Ricerche su cento soggetti dalla nascita sino a tarda et, Z. Zellforsch., 24, 439,
1936.
58. Amprino, R. La structure du tissu osseux envisage comme expression de diffrences dans la vitesse
de laccroissement, Arch. Biol., 58, 315, 1947.
59. Marotti, G., The structure of bone tissues and the cellular control of their deposition, Ital. J. Anat.
Embryol., 101, 25, 1996.
60. Bonucci, E., The basic multicellular unit of bone, Ital. J. Miner. Electrol. Metab., 4, 115, 1990.
61. Partt, A.M., Bone remodeling: relationship to the amount and structure of bone, and the pathogenesis
and prevention of fractures, in Osteoporosis: Etiology, Diagnosis, and Management, Riggs, B.L. and
Melton, L.J.I., Eds., Raven Press, New York, 1988, 45.
62. Filogamo, G., La forme et la taille des ostones chez quelques mammifres, Arch. Biol., 57, 137, 1946.
63. Martin, R.B., Gibson, V.A., Stover, S.M., et al., Osteonal structure in the equine third metacarpus,
Bone, 19, 165, 1996.
64. Black, J., Mattson, R., and Korostoff, E., Haversian osteons: size, distribution, internal structure, and
orientation, J. Biomed. Mater. Res., 8, 299, 1974.
65. Broulik, P., Kragstrup, J., Mosekilde, L., and Melsen, F., Osteon cross-sectional size in the iliac crest,
Acta Pathol. Microbiol. Immunol. Scand. Sect. A, 90, 339, 1982.
66. Benninghoff, H., Spaltlinien am Knochen, eine Methode zur Ermittlung der Architektur platter
Knochen, Verh. Anat. Ges., 34, 189, 1925.
67. Hert, J., Fiala, P., and Petrtyl, M., Osteon orientation of the diaphysis of the long bones in man, Bone,
15, 269, 1994.
68. Cohen, J. and Harris, W.H., The three-dimensional anatomy of Haversian systems, J. Bone Joint Surg.,
40-A, 419, 1958.
69. Petrtyl, M., Hert, J., and Fiala, P., Spatial organization of the Haversian bone in man, J. Biomech.,
29, 161, 1996.
70. Portigliatti Barbos, M., Bianco, P., Ascenzi, A., and Boyde, A., Collagen orientation in compact bone:
II. Distribution of lamellae in the whole of the human femoral shaft with reference to its mechanical
properties, Metab. Bone Dis. Relat. Res., 5, 309, 1984.
71. Scherft, J.P. and Groot, C.G., The electron microscopic structure of the osteoblast, in Ultrastructure
of Skeletal Tissues, Bonucci, E. and Motta, P.M., Eds., Kluwer Academic, Boston, 1990, 209.
72. Marks, S.C.J., The structural basis for bone cell biology. A review, Acta Med. Dent. Helv., 2, 141, 1997.
73. Marotti, G., Decrement of volume of osteoblasts during osteon formation and its effect on the size
of the corresponding osteocytes, in Bone Histomorphometry, Meunier, P.J., Ed., Armour Montagu,
Paris, 1976, 385.
74. Miller, S.C., de Saint-Georges, L., Bowman, B.M., and Jee, W.S.S., Bone lining cells: structure and
function, Scanning Microsc., 3, 953, 1989.
75. Amprino, R. and Engstrm, A., Studies on X-ray absorption and diffraction of bone tissue, Acta Anat.,
15, 1, 1952.
76. Marotti, G., Favia, A., and Zambonin Zallone, A., Quantitative analysis of the rate of secondary bone
mineralization, Calcif. Tissue Res., 10, 67, 1972.
77. Engfeldt, B. and Hjerpe, A., Density gradient fractionation of dentine and bone powder, Calcif. Tissue
Res., 16, 261, 1974.
78. Engfeldt, B. and Hjerpe, A., Glycosaminoglycans and proteoglycans of human bone tissue at different
stages of mineralization, Acta Pathol. Microbiol. Scand. Sect. A, 84, 95, 1976.
0266/ ch01/frame Page 19 Thursday, October 21, 1999 1:52 PM
20 Mechanical Testing of Bone and the BoneImplant Interface
79. Dhem, A., Etude histologique et microradiographique des manifestations biologiques propres au tissu
osseux compact, Bull. Acad. Med. Bel., 135, 368, 1980.
80. Engstrm, A. and Engfeldt, B., Lamellar structure of osteons demonstrated by microradiography,
Experientia, 15, 19, 1953.
81. Vincent, J., Microradiographie des lamelles de los haversien, Arch. Biol., 69, 561, 1958.
82. Bohatirchuk, F.P., Campbell, J.S., and Jeletzky, T.F., Bone lamellae, Acta Anat., 83, 321, 1972.
83. Baron, R., Molecular mechanisms of bone resorption. An update, Acta Orthop. Scand., 66, 66, 1995.
84. Gthlin, G. and Ericsson, J.L.E., The osteoclast. Review of ultrastructure, origin, and structure-function
relationship, Clin. Orthop., 120, 201, 1976.
85. Hall, T.J. and Chambers, T.J., Molecular aspects of osteoclast function, Inamm. Res., 45, 1, 1996.
86. Sasaki, T., Recent advances in the ultrastructural assessment of osteoclastic resorptive functions,
Microsc. Res. Tech., 33, 182, 1996.
87. Suda, T., Nakamura, I., Jimi, E., and Takahashi, N., Regulation of osteoclast function, J. Bone Miner.
Res., 12, 869, 1997.
88. Philipson, B., Composition of cement lines in bone, J. Histochem. Cytochem., 13, 270, 1965.
89. Schafer, M.B., Burr, D.B., and Frederickson, R.G., Morphology of the osteonal cement line in human
bone, Anat. Rec., 217, 223, 1987.
90. McKee, M.D. and Nanci, A., Osteopontin at mineralized tissue interfaces in bone, teeth, and osseoin-
tegrated implants: ultrastructural distribution, and implications for mineralized tissue formation, turn-
over, and repair, Microsc. Res. Tech., 33, 141, 1996.
91. Gebhardt, W., Ueber funktionell wichtige Anordnungsweisen der grsseren und feineren Bauelemente
der Wilbertierknochens. II. Spezieller Teil. Der Bau der Haversschen Lamellensysteme und seine
funktionelle Bedeutung, Arch. Entwicklungsmech. Org., 20, 187, 1906.
92. Ascenzi, A. and Bonucci, E., The compressive properties of single osteons, Anat. Rec., 161, 377, 1968.
93. Vincentelli, R., Relation between collagen ber orientation and age of osteon formation in human
tibial compact bone, Acta Anat., 100, 120, 1978.
94. Frank, R., Frank, P., Klein, M., and Fontaine, R., Los compact humain normal au microscope
lectronique, Arch. Anat. Microsc. Morphol. Exp., 44, 191, 1955.
95. Smith, J. W., The arrangement of collagen bres in human secondary osteones, J. Bone Joint Surg.,
42-B, 588, 1960.
96. Giraud-Guille, M.-M., Twisted plywood architecture of collagen brils in human compact bone
osteons, Calcif. Tissue Int., 42, 167, 1988.
97. Weiner, S., Arad, T., Sabanay, I., and Traub, W., Rotated plywood structure of primary lamellar bone
in the rat: orientations of the collagen bril arrays, Bone, 20, 509, 1997.
98. Raspanti, M., Guizzardi, S., Strocchi, R., and Ruggeri, A., Collagen bril patterns in compact bone:
preliminary ultrastructural observations, Acta Anat., 155, 249, 1996.
99. Pannarale, L., Braidotti, P., dAlba, L., and Gaudio, E., Scanning electron microscopy of collagen ber
orientation in the bone lamellar system in non-decalcied human samples, Acta Anat., 151, 36, 1994.
100. Weiner, S., Arad, T., and Traub, W., Crystal organization in rat bone lamellae, Fed. Eur. Biochem.
Soc., 285, 49, 1991.
101. Raspanti, M., Guizzardi, S., Strocchi, R., and Ruggeri, A., Different brillar architectures coexisting
in Haversian bone, Ital. J. Anat. Embryol., 100, 103, 1995.
102. Ziv, V., Sabanay, I., Arad, T., et al., Transitional structures in lamellar bone, Microsc. Res. Tech., 33,
203, 1996.
103. Marotti, G., The original contribution of the scanning electron microscope to the knowledge of bone
structure, in Ultrastructure of Skeletal Tissues, Bonucci, E. and Motta, P.M., Eds., Kluwer Academic,
Boston, 1990, 19.
104. Marotti, G., A new theory of bone lamellation, Calcif. Tissue Int., 53, S47, 1993.
105. Marotti, G. and Muglia, M.A., A scanning electron microscope study of human bony lamellae.
Proposal for a new model of collagen lamellar organization, Arch. Ital. Anat. Embriol., 93, 163, 1988.
106. Ascenzi, A. and Bonucci, E., The ultimate tensile strength of single osteons, Acta Anat., 58, 160, 1964.
107. Ascenzi, A., The micromechanics vs. the macromechanics of cortical bone A comprehensive
presentation, J. Biomech. Eng., 110, 357, 1988.
108. Carando, S., Portigliatti Barbos, M., Ascenzi, A., and Boyde, A., Orientation of collagen in human
tibial and bular shaft and possible correlation with mechanical properties, Bone, 10, 139, 1989.
0266/ ch01/frame Page 20 Thursday, October 21, 1999 1:52 PM
Basic Composition and Structure of Bone 21
109. Lozupone, E., The structure of the trabeculae of cancellous bone 1. The calcaneus, Anat. Anz., 159,
211, 1985.
110. Lozupone, E. and Favia, A., The structure of the trabeculae of cancellous bone. 2. Long bones and
mastoid, Calcif. Tissue Int., 46, 367, 1990.
111. Bell, G.H., Bone as a mechanical engineering problem, in The Biochemistry and Physiology of Bone,
Bourne, G.H., Ed., Academic Press, New York, 1956, 27.
112. Koch, J.C., The laws of bone architecture, Am. J. Anat., 21, 177, 1917.
113. Rybicki, E.F., Simonen, F.A., and Weis, E.B., Jr., On the mathematical analysis of stress in the human
femur, J. Biomech., 5, 203, 1972.
114. Kleerekoper, M., Villanueva, A.R., Stanciu, J., et al., The role of three-dimensional trabecular micro-
structure in the pathogenesis of vertebral compression fractures, Calcif. Tissue Int., 37, 594, 1985.
115. Partt, A.M., Trabecular bone architecture in the pathogenesis and prevention of fracture, Am. J. Med.,
82, 68, 1987.
116. Ballanti, P., Bonucci, E., Della Rocca, C., et al., Bone histomorphometric reference values in 88
normal Italian subjects, Bone Miner., 11, 187, 1990.
117. Birkenhger-Frenkel, D.H., Courpron, P., Hpscher, E.A., et al., Age-related changes in cancellous
bone structure. A two-dimensional study in the transiliac and iliac crest biopsy sites, Bone Miner., 4,
197, 1988.
118. Hoikka, V. and Arnala, I., Histomorphometric normal values of the iliac crest cancellous bone in a
Finnish autopsy series, Ann. Clin. Res., 13, 383, 1981.
119. Malluche, H.H., Sherman, D., Meyer, W., and Massry, S.G., Quantitative bone histology in 84 normal
American subjects, Calcif. Tissue Int., 34, 449, 1982.
120. Melsen, F., Melsen, B., Mosekilde, L., and Bergmann, S., Histomorphometric analysis of normal bone
from the iliac crest, Acta Pathol. Microbiol. Scand. Sect. A, 86, 70, 1978.
121. Schnitzler, C.M., Pettifor, J.M., Mesquita, J.M., et al., Histomorphometry of iliac crest bone in 346
normal black and white South African adults, Bone Miner., 10, 183, 1990.
122. Atkinson, P.J., Variation in trabecular structure of vertebrae with age, Calcif. Tissue Res., 1, 24, 1967.
123. Compston, J.E., Mellish, R.W.E., and Garrahan, N.J., Age-related changes in iliac crest trabecular
microanatomic bone structure in man, Bone, 8, 289, 1987.
124. Goldstein, S.A., Goulet, R., and McCubbrey, D., Measurement and signicance of three-dimensional
architecture to the mechanical integrity of trabecular bone, Calcif. Tissue Int., 53, S127, 1993.
125. Partt, A.M., Mathews, C.H.E., Villanueva, A.R., et al., Relationship between surface, volume, and
thickness of iliac trabecular bone in aging and in osteoporosis, J. Clin. Invest., 72, 1396, 1983.
126. Bonucci, E. and Gherardi, G., Histochemical and electron microscope investigations on medullary
bone, Cell Tissue Res., 163, 81, 1975.
127. Lozupone, E., A quantitative analysis of bone tissue formation in different regions of the spongiosa
in the dog skeleton, Anat. Anz., 145, 425, 1979.
128. Ballanti, P., Martelli, A., Mereto, E., and Bonucci, E., Ovariectomized rats as experimental model of
postmenopausal osteoporosis: critical considerations, Ital. J. Miner. Electrol. Metab., 7, 243, 1993.
129. Miller, S.C., Bowman, B.M., Miller, M.A., and Bagi, C.M., Calcium absorption and osseous organ-,
tissue-, and envelope-specic changes following ovariectomy in rats, Bone, 12, 439, 1991.
130. Yoshida, S., Yamamuro, T., Okumura, H., and Takahashi, H., Microstructural changes of osteopenic
trabeculae in the rat, Bone, 12, 185, 1991.
131. Partt, A.M., Age-related structural changes in trabecular and cortical bone: cellular mechanisms and
biomechanical consequences, Calcif. Tissue Int., 36, S123, 1984.
132. Frost, H.M., Changing concepts in skeletal physiology: Wolffs law, the mechanostat, and the Utah
Paradigm, Am. J. Hum. Biol., 10, 599, 1998.
133. Frost, H.M., Ferretti, J.L., and Jee, W.S.S., Perspectives: some role of mechanical usage, muscle
strength, and the mechanostat in skeletal physiology, disease, and research, Calcif. Tissue Int., 62, 1,
1998.
134. Partt, A.M., Osteonal and hemi-osteonal remodeling: the spatial and temporal framework for signal
trafc in adult human bone, J. Cell. Biochem., 55, 273, 1994.
135. Ascenzi, A., Microscopic dissection and isolation of bone constituents, in Skeletal Research: An
Experimental Approach, Kunin, A.S. and Simmons, D.J., Eds., Academic Press, New York, 1983, 185.
0266/ ch01/frame Page 21 Thursday, October 21, 1999 1:52 PM
0266/ ch01/frame Page 22 Thursday, October 21, 1999 1:52 PM

23

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

2

Basic Concepts of Mechanical
Property Measurement and
Bone Biomechanics

Yuehuei H. An, William R. Bareld, and Robert A. Draughn

CONTENTS

I. Introduction ............................................................................................................................23
II. Concepts Relevant to Mechanical Testing.............................................................................24
A. Statics and Dynamics .......................................................................................................24
B. Force and Displacement...................................................................................................25
C. Stress and Strain...............................................................................................................26
D. Concepts Related to Material Dimensions ......................................................................29
E. Loading Modes.................................................................................................................30
1. Compressive or Tensile Loading.................................................................................30
2. Bending........................................................................................................................30
3. Torsional Loading........................................................................................................31
4. Multiaxial Loading......................................................................................................32
5. Static or Cyclic............................................................................................................32
F. Strength.............................................................................................................................33
G. Elasticity (Stiffness and Elastic Modulus) and Compliance...........................................33
H. Fracture Energy and Toughness.......................................................................................35
I. Viscoelasticity.................................................................................................................235
J. Failure, Fracture, Fatigue, Fatigue Fracture, Buckling, and Cracking ...........................36
III. Mechanical Modeling and Simulation...................................................................................38
IV. Summary ................................................................................................................................39
References ........................................................................................................................................39

I. INTRODUCTION

Mechanics is a physical science that assesses the effects of force on objects. Mechanical properties
of bone are basic parameters which reect the structure and function of bone and can be measured
by testing whole anatomical units or specimens prepared to isolate particular structural components.
Within this context the fracture of bone can represent failure of whole bone at the structural level
and bone tissue at the material level. The mechanical behavior of bone in normal physiological
situations is similar to that of an elastic material with no visible change in external appearance.
Bone, however, can be degraded and still retain its morphological features for an indenite period
of time. Unlike inorganic materials, bone has adaptive mechanisms which give the tissue the ability
to repair itself, altering its mechanical properties and morphology in response to increased or
decreased function. This chapter covers common mechanical concepts and terminology used in the
eld of bone biomechanics and mechanical property measurement.

0266/ ch02/frame Page 23 Thursday, October 21, 1999 2:00 PM

24

Mechanical Testing of Bone and the BoneImplant Interface

II. CONCEPTS RELEVANT TO MECHANICAL TESTING

There are two main unit systems for measurements, the British system and the SI system. The
British system uses pound, foot, and second as its basic units and the SI system uses kilogram,
meter, newton, second, and Celsius (

C). Most journals or publishers prefer the SI system due to
its simplicity. Measurement involves two groups of parameters, scalars and vectors. A scalar is a
quantity, such as temperature, length, or mass, which provides information regarding magnitude
only. Vector quantities possess both magnitude and direction. Measures such as force, moment,
and torque are typical examples. Symbols commonly used in the eld of bone biomechanics are
listed in Table 2.1.

A. S

TATICS



AND

D

YNAMICS

Mechanics deals with the effects of forces on the form or motion of bodies or subjects, and can
be divided into two categories, statics and dynamics. Statics studies bodies at rest or when there
is equilibrium of forces. The state of equilibrium means that the sum of the forces and the sum of

TABLE 2.1
Common Symbols Used in the Field of Bone Biomechanics

Symbol Meaning Common SI Units Used

A

Surface area m

2

E

Elastic modulus or Youngs modulus Pa (N/m

2

), MPa, GPa

F

Force N

I

Area moment of inertia of cross section m

4

J (joule) SI unit of energy (work) Nm

J

Polar moment of inertia m

4

L

Length or span m

M

Mass kg

P

Load N

r

Radius m

S

Strength Pa (N/m

2

), MPa, GPa

S

b

Bending strength Pa (N/m

2

), MPa, GPa

S

c

Compressive strength Pa (N/m

2

), MPa, GPa

S

t

Tensile strength Pa (N/m

2

), MPa, GPa

S

s

Shear strength Pa (N/m

2

), MPa, GPa

T

Torque or torsional moment Nm

W

Weight N

!

Strain mm/mm, %

!

y

Yield strain mm/mm, %

!

ult

Ultimate strain mm/mm, %

"

Viscosity Pa

#

sec

$

Poissons ratio length/length

%

a

Apparent density kg/m

3

, g/cm

3

, mg/mm

3

%

ash

Ash density kg/m

3

, g/cm

3

, mg/mm

3

&

Stress Pa (N/m

2

), MPa, GPa

&

y

Yield stress Pa (N/m

2

), MPa, GPa

&

ult

Ultimate stress Pa (N/m

2

), MPa, GPa

'

Shear stress Pa (N/m

2

), MPa, GPa

Source

: Low, J. and Reed, A.,

Basic Biomechanics Explained,

Butterworth Heine-
mann, London, 1996. With permission.

0266/ ch02/frame Page 24 Thursday, October 21, 1999 2:00 PM

Basic Concepts of Mechanical Property Measurement and Bone Biomechanics

25

the moments is equal to zero. There must be no overlapping resultants. Dynamics studies moving
bodies and is subdivided into kinematics and kinetics. Kinematics describes the relations among
displacements, velocities, and accelerations in all kinds of motion without regard to the forces
involved and has been described as geometry of motion. Kinetics deals with the forces causing
movement. Most of the concepts, studies, methods of evaluation, and ndings presented in this
book fall into the category of statics.

B. F

ORCE



AND

D

ISPLACEMENT

Force (

F

) or load, the primary physical entity in mechanics, is a measurable vector, which has a
magnitude, direction, and point of application. Forces act on a body and tend to change the velocity
of the body (external effect) or shape of the body (internal effect). Changes in shape are determined
by changes of the relative positions of the structural elements within a body. The changes in shape,
structure, or morphology of bodies or objects (such as bone) result from the effects of the forces.
For instance, movements of an implant relative to the bone in which it resides is likely the result
of external forces, although micromovement that occurs at the boneimplant interface is also impacted
by concentric and eccentric muscle forces. Force is a vector quantity, meaning that it has both intensity
and direction. There are basically three types of fo2525rces: tensile, compressive, and shear force,
which are determined by the direction and effect of the force(s) acting on the body (Figure 2.1).
The magnitude of a force is expressed in the SI system of units as newtons. A newton is the
force required to give 1 kilogram mass an acceleration of 1 meter per second per second (m/sec

2

).
Small forces can be expressed in units of dynes. A dyne is the force that gives 1 gram mass an
acceleration of 1 centimeter per second per second (cm/sec

2

).
In most measurements of mechanical properties the applied force is measured by a load cell and
changes in specimen dimensions are indicated by the motion of the load application system. In a
mechanically actuated test machine (a screw machine) the motion of the crosshead denes the total
displacement of the specimen and the test xture. In a hydraulically actuated test system, the movement
of the actuator piston proscribes the displacement. While a test object is being loaded, a loaddis-
placement curve is recorded by a chart recorder or computer (Figure 2.2). The loaddisplacement
(PD) curve denes the total deformation of the specimen in the direction of force application.
Actually, the displacement includes the deformation of the specimen and the deformation of the testing
system, such as in the case of an unstable or soft interface between the specimen and the machine.
PD curves are particularly useful for measuring the strength and stiffness of whole structures;
however, to compare behavior of different materials, stressstrain curves are needed for standardiza-
tion. When transforming the loaddisplacement curve to a stressstrain curve, force and deformation
are normalized as stress and strain by the dimensions of the sample.

FIGURE 2.1

Three types of pure forces.

0266/ ch02/frame Page 25 Thursday, October 21, 1999 2:00 PM

26

Mechanical Testing of Bone and the BoneImplant Interface

C. S

TRESS



AND

S

TRAIN

Stress is the internal resistance of a material body to a force acting upon it (Figure 2.3). In bone,
stress (

&

) arises from the forces or bonds between molecules, between collagen bers, and the
bonding between collagen and hydroxyapatite crystals. In solid mechanics, stress is a normalized
force. It is a ratio (i.e., force per unit area) which is calculated by the magnitude of the force (

F

)
divided by the surface area (

A

) over which the force acts:

&

=

F

/

A

(2.1)
Axial or normal stresses are categorized as either compressive or tensile stresses. Pressure is
a normal stress acting on the surface of an object. Shear stress (

'

) exists when the force is applied
parallel to the surface of a material body. For example, when long bones are subjected to a torsional
load, shear stresses are developed in the bone.

FIGURE 2.2

A typical loaddisplacement curve.

FIGURE 2.3

Illustration of (A, B) normal stress

(&)

and shear (C) stress (

'

).

0266/ ch02/frame Page 26 Thursday, October 21, 1999 2:00 PM

Basic Concepts of Mechanical Property Measurement and Bone Biomechanics

27

The standard unit of stress (SI system) is the pascal (Pa), which is 1 N force distributed over
one square meter (1 N/m

2

). The pascal is a small unit and typically stress is expressed as multiples
of pascals including kilopascal (kPa or 10

3

N/m

2

), megapascal (MPa or 10

6

N/m

2

), and gigapascal
(GPa or 10

9

N/m

2

). The physiological stress levels for bone are generally below the megapascal
range. The unit of stress in English units is pounds/in.

2

(psi).
Stress concentration is the increase in stress around a defect in a material, such as a screw hole
or a defect in bone, and is discussed in detail in other chapters in this text.
Strain represents the dimensional changes of a subject or body under the action of a force or
several forces. When force is applied, the object changes its dimensions (Figure 2.4). This change
in dimension is termed deformation (

(

L). Deformation per unit length (

L

) is strain (

!

):

!

=

(

L/

L

(2.2)
Strain is a dimensionless measure since it is the ratio of two quantities, both in units of length.
Strain is dened as the geometric change in a material in response to force application and is also
known as a normalized displacement. The types of strains in a body are the same as the types of
force producing them: normal strain (compressive or tensile) and shear strain. The latter is dened
as the angular deformation of the material measured in radians.
Normal and shearing strains are concepts analogous to normal and shear stress. Normal strain is
a measure of the change in length per unit length and shear strain is half the change of an original 90

angle and usually is measured in radians.

1

Shear strain can also be dened as the angle between original
and deected locations on the edge of a material (

)

) (Figure 2.4).

2

Ultimate strain (

!

ult

) is the strain that

FIGURE 2.4

Illustration of deformation and strain. (A) Deformation (

(

L) and normal strain (

!

) under a
compressive load (

F

). (B) shows that a shearing force (

F

) acting parallel to the surface (A) of the cube produces
shear strain (

)

).

0266/ ch02/frame Page 27 Thursday, October 21, 1999 2:00 PM

28

Mechanical Testing of Bone and the BoneImplant Interface

occurs at fracture. Strain can be measured directly and is recorded as length per unit length (such as
cm/cm or mm/mm) or can also be reported as a percentage. Positive strain occurs when material is
subjected to tensile stress in the direction of the applied load, reected as an increase on the length of
the material. If the material is compressed, the length is decreased and the strain is negative. Strain
rate is the deformation per unit time, which is an important parameter for viscoelastic materials such
as bone. Strain rate is discussed in more detail in the section on viscoelasticity.
Strain can be measured by bonding strain gauges directly to bone or with extensometers, brittle
lacquer coatings, or birefringent plastic coatings.

3

Uniaxial gauges allow for single directional
measures, while triaxial gauges or rosettes permit direction and magnitude of principal strains to
be determined. Several technical manuscripts offer detailed information with regard to how these
procedures are carried out.

4

A stressstrain curve (Figure 2.5) displays the relationships between applied stress and resulting
strain in a mechanical property measurement test. The curves are generated by conversion of applied
force data into units of stress or plotting against directly measured values of strain. Modern
computer-based data acquisition systems signicantly facilitate the generation of stressstrain
curves. From the stressstrain curve, the following properties can be determined:
1. The beginning of the elastic portion, which represents the engagement of the machine
specimen;
2. The proportional limit or the limit to which stress and strain are proportional (PL);
3. The elastic limit or the limit at which the greatest stress can be applied without leaving
permanent deformation upon removal of the load;
4. The elastic range, that part of the curve where strain is directly proportional to stress;
5. The yield point, or point at which permanent deformation occurs;
6. The range of plastic deformation, the part of the curve from the yield point to the failure point;
7. The breaking point or ultimate stress or ultimate strength; and
8. The amount of energy absorbed by the specimen prior to failure.
The stressstrain plot also allows calculation of the elastic modulus of the material by measuring
the slope of the curve in the elastic region.

FIGURE 2.5

A typical stressstrain curve.

0266/ ch02/frame Page 28 Thursday, October 21, 1999 2:00 PM

Basic Concepts of Mechanical Property Measurement and Bone Biomechanics

29

D. C

ONCEPTS

R

ELATED



TO

M

ATERIAL

D

IMENSIONS

When a material specimen is subjected to a compressive uniaxial force, its dimension decreases in
the axial (loading) direction and increases in the transverse direction (Figure 2.6). The relationship
between these two strains is given by Poissons ratio (

$

):

$

=

!

t

/

!

a

(2.3)
where

!

t



is transverse strain and

!

a



is axial strain. Poissons effect applies as well in the opposite
sense, tensile loading. In general, Poissons ratio is less than 0.5 which means the volume of the
material under simple tensile load cannot diminish and that compression cannot increase a volume
of the material. The Poissons ratio of human cancellous bone is 0.2 to 0.3.

5,6

The clinical impli-
cations of Poissons ratio may be minor, although there may be technical implications with regard
to the accuracy of mechanical testing, especially in trabecular bone.
7
A moment is the tendency to produce rotation, and is the product of force magnitude and
perpendicular moment arm length. Like forces, moments have both internal and external effects
upon their objects. The external effect of a moment is to change, or attempt to change, the angular
or rotational velocity of the body. The internal effects of a moment is to cause a state of strain.
Inertia, a fundamental characteristic of all matter, is the property that causes a body to remain
at rest or in uniform motion in one direction unless acted upon by an external force that changes
the body with respect to the state of rest, velocity, or direction (Newtons rst law). Areal moment
of inertia (I) is the structural feature governing stiffness in bending. If a bone is considered to be
a hollow cylinder, the areal moment of inertia is calculated by
I = * (a
3
b a+
3
b+)/64 (2.4)
FIGURE 2.6 Illustration of Poissons effect: (A) axial direction; (B) transverse direction.
0266/ ch02/frame Page 29 Thursday, October 21, 1999 2:00 PM
30 Mechanical Testing of Bone and the BoneImplant Interface
where a, a+, b, and b+ are the mean external and internal anteroposteral and side-to-side diameters
for the cross sections at the loading points of the bone (Figure 2.7). The external diameters (a and
b) are measured before testing by use of a digital caliper. After testing, the pieces are glued together
and cut transversely at the break point. The dimensions of the medullary canal (a+ and b+) are then
measured. I, as a geometric parameter, expresses the characteristics of the cross-sectional distribu-
tion relative to the transverse axis. Simply put, this explains the large differences in bending
resistance of a meterstick held at compared with when it is held on the edge and bent.
8
Polar
moment of inertia is a structural feature governing stiffness in torsion and is a function of the
distance the individual masses of the object are located from the axis of rotation. Basically, the
further the individual masses of bone are from the axis of rotation, the greater will be the moment
and the more resistant the material will be to deformation.
9
E. LOADING MODES
Depending upon the direction of application, a force applied to an object or body can be axial
(compressive and tensile), bending, torsional, or multiaxial and can be static or cyclic (repeated).
1. Compressive or Tensile Loading
In compressive or tensile loading, the force is applied perpendicular to the surface of the body or
object. Typical examples are the in vitro compressive and tensile tests. Compressive and tensile
loading can be thought of as numerous small loads directed toward or away from the surface of
the structure, with maximal compressive or tensile stress occurring in a plane perpendicular to the
applied load. Compressive loading in vivo creates a shortening and widening of the bone and is
commonly seen in human vertebra. Clinically, fractures that result from tensile loading are commonly
seen in cancellous bone, as in the calcaneus fracture when a strong contraction from the triceps surae
occurs. Another similar condition is the avulsion fracture of the medial epicondyle of the humerus.
2. Bending
Although bones are not beams, frequently they are modeled as such, especially bones of the
appendicular skeleton. Due to the general curvature of long bones, they are subjected to axial and
FIGURE 2.7 Illustration of the external and internal anteroposteral and side-to-side diameters for the cross
sections at the loading points of the bone.
0266/ ch02/frame Page 30 Thursday, October 21, 1999 2:00 PM
Basic Concepts of Mechanical Property Measurement and Bone Biomechanics 31
bending forces in vivo. Bending loads on bone may occur with three-point bending such as that
seen in the tibia during a boot-top fracture which occurs when skiers fall forward over their skis.
Bending causes tensile forces and lengthening on the convex side of the bone and compressive
forces and shortening on the concave side. In bending, stress and strains are maximal at the surfaces
of the beam and are zero at the neutral axis (Figure 2.8). Frequently, bending forces are coupled
with axial and transverse loading. Because bone is best able to resist compressive loads, muscle
contraction of the triceps surae, in the ski-boot-top fracture, may work to attenuate the tensile forces
seen on the posterior (convex) side of the tibia, thereby reducing the chance of fracture.
3,8
3. Torsional Loading
In torsional loading a cylinder of long bone is twisted under a torsional test, and shear stress
develops (Figure 2.9). Shear stress due to torsion is expressed as:
2

' = Tr/J (2.5)
FIGURE 2.8 Illustration of bending of diaphyseal bone.
FIGURE 2.9 Illustration of torsional loading. ' = shear stress, T = torque, J = polar moment of inertia,
R = the distance from the center of the cylinder to any point.
0266/ ch02/frame Page 31 Thursday, October 21, 1999 2:00 PM
32 Mechanical Testing of Bone and the BoneImplant Interface
where T is the torque or torsional moment of the applied force, r is the radial distance from the
neutral axis to the point in the cross section at which the shear stress acts, and J is the polar moment
of inertia.
4. Multiaxial Loading
While considerable progress has been made in the study of uniaxial compressive and tensile properties
of bone tissue in the last several decades, multiaxial loading of bone to measure material mechanics
has lagged due to the complexity and difculty of the processes of loading and recording. Living
bone is seldom loaded in one direction, but is mostly multidirectional and therefore multiaxial. For
this reason, the study of multidimensional loading and the effect on bone is essential. Basically,
multiaxial loading uses the same concepts as in uniaxial loading, but extends the concepts to two-
and three-dimensional space. Consideration in multiaxial loading must be given for normal stresses
and strains as well as shear stresses and strains. Three-dimensional analysis generally provides a more
realistic assessment of the material loading complexity and may involve as many as nine material
property characteristics to describe the anisotropic sample completely. Analysis of the mechanical
properties of the material can be simplied by assuming symmetry and/or isotropy in two or more
of the principal planes.
2
In Figure 2.10, biaxial and three-dimensional tensile loading is shown. The
concepts are equally applicable to compressive loading when the force vectors would be reversed.
10
In vivo loading of bone is difcult to assess because of the irregular geometric properties of
bone and the fact that bone is commonly exposed to multiple indeterminate loads. Recent in vivo
measures have been calculated for walking and jogging and these convincingly demonstrate the
complexity of multiaxial loading. As might be expected, compression occurs at heel strike and toe-
off and are tensile during the stance phase. Shear forces normally occur during the late phases of
gait, and torsional forces are seen during external rotation of the tibia, as part of the screw home
mechanism, as stance approaches. Running, as might be anticipated, creates a totally different
loading pattern.
11
5. Static or Cyclic
In seeking the answers to tolerable levels of force that a bone can withstand prior to fracture, the
static ultimate properties of bone are of interest; however, repetitive, submaximal trauma is fre-
quently the more signicant area of concern.
12
FIGURE 2.10 Multiaxial loading with applied forces in the x, y, and z directions.
0266/ ch02/frame Page 32 Thursday, October 21, 1999 2:00 PM
Basic Concepts of Mechanical Property Measurement and Bone Biomechanics 33
Loading a bone specimen cyclically with progressively higher forces produces a highly nonlinear
stressstrain curve. For example, in Figure 2.11, the bone specimen is loaded up to point A. When the
force is removed, the specimen exhibits reversible behavior and returns to its original unloaded length.
The load that causes the stressstrain curve up to point B also permits the material to return to its
original length, although the time required to return to normal is longer. Stressstrain up to point C
produces a permanent change in the original length of the material and is not reversible. The loading
and unloading curves do not overlap, but rather create a closed loop, known as a hysteresis loop, which
is indicative of the inefciency of storing and releasing of strain energy. The area under the unloading
curve represents strain energy release during unloading. The area enclosed by the hysteresis loop
represents energy dissipated within the material through mechanical damage and internal friction.
12
Test-type inuences the mechanical properties that bone will exhibit. Cyclic loading produces
microdamage that accumulates with each cycle and the damage increases as the intensity of testing
increases. Intensity can be varied through change in the load magnitude and through the number
of cycles to which the specimen is exposed. Lessons learned from cyclic loading include that, once
a crack occurs, the number and cyclical load required for propagation of the fracture decreases
rapidly and that there is a strong negative correlation between load intensity and number of cycles
needed for failure.
12
F. STRENGTH
Strength can be dened as the internal resistance of a material to deformation and ultimately failure
or fracture. Proportional limit is the point on the loaddisplacement curve where the load is no
longer proportional to the deformation. At this point there begins a brief region of relatively large
strain for little increase in stress. It indicates that a portion of the bone structure starts to fail or
crack and the structure becomes plastic. Yield strength (&
y
) is dened as the stress corresponding
to a specic amount of permanent deformation. Ultimate strength denes the stress required to
fracture the bone or the boneimplant interface. It is called the strength, or ultimate stress. The
magnitude of strength is calculated from a mechanical test, based on the load, deformation, and
the dimensions of the specimen.
G. ELASTICITY (STIFFNESS AND ELASTIC MODULUS) AND COMPLIANCE
Elasticity is the ability of a material to return to its original shape when an applied stress is removed.
Elasticity is quantied by a simple stiffness value or elastic modulus. Stiffness is the ability of a
FIGURE 2.11 A typical hysteresis loop.
0266/ ch02/frame Page 33 Thursday, October 21, 1999 2:00 PM
34 Mechanical Testing of Bone and the BoneImplant Interface
material to resist being deformed when a force is applied to it. A simple stiffness value is any force
divided by its corresponding deformation within the elastic range of the loaddisplacement curve.
S = F/d (2.6)
Elastic modulus is a standardized stiffness value and is the ratio between stress (&) and strain
(!) or any (&/(! within the linear portion of the stressstrain curve (Equation 2.7). The elastic
modulus determined in a tensile test is also called Youngs modulus.
13
The modulus of elasticity
is computed in terms of force per unit area from the slope of the elastic region of the stressstrain
curve. The slope is usually obtained by drawing a line tangent to the stressstrain curve. The ratio
of stress to strain is not a function of the size and shape of the material being tested, but rather is
a measure of the ability of the material to maintain shape under application of external loads and
is therefore material dependent (Figure 2.12). The property of elasticity is time independent and
completely reversible.
E = &/! = (&/(! (2.7)
Linearly elastic materials obey Hookes law, which proposes that the stress and strain are
linearly related and mathematically expressed as
& = E! (2.8)
The mechanical properties of biological tissues, such as bone, typically are not linear throughout
their physiological range due to the nonlinear characteristics of their uid component.
FIGURE 2.12 Illustration of the elasticities of several different materials.
0266/ ch02/frame Page 34 Thursday, October 21, 1999 2:00 PM
Basic Concepts of Mechanical Property Measurement and Bone Biomechanics 35
Compliance is the inverse of stiffness or modulus and is dened as the ratio of deformation to
load or strain to stress.
14
Compliance = !/& (2.9)
Relative to other biological materials, bone has relatively high E values (high slope of
stressstrain curve) and low compliance values. Compliant materials, such as cartilage or skin,
have low E values and high compliance values. Mechanical testing machines are constructed of
high-modulus materials (E of steels , 200 GPa), so their compliance is considered as zero when
testing softer materials such as cortical bone (E = 5 to 21 GPa) or cancellous bone (E < 1 GPa).
If several xture parts are used in the testing assembly, a signicant machine compliance can exist.
A routine testing for machine compliance is recommended by testing the xture column without
the specimen in place.
H. FRACTURE ENERGY AND TOUGHNESS
The area under the stressstrain curve represents the energy absorbed when the object is loaded.
This energy is stored as elastic strain energy and can be dissipated as heat. Work of fracture
(toughness) is the energy required per unit volume of a material to produce fracture. The unit for
work and toughness is the joule (Nm).
I. VISCOELASTICITY
Viscoelasticity describes the time-dependent mechanical characteristics of materials. Bone is a
viscoelastic material, which means that the stress developed within bone is dependent on the rate
at which the bone specimen is strained (strain rate). With increasing strain rates, the material appears
stiffer and stronger, with smaller deformations (Figure 2.13). Most biological materials exhibit
some degree of viscoelasticity. Two behaviors of viscoelastic materials which are important in
quantifying the mechanical properties of bone are (1) stress relaxation and (2) creep. Stress
relaxation is the decay of stress within a material subjected to a constant strain (Figure 2.14). Stress
relaxation rate is the slope of the stresstime curve determined under a constant strain.
13
Creep is
the gradual increase in strain of a material subjected to a constant load (or deformation under
constant load). A creep curve displays strain vs. time under a constant force or stress (Figure 2.15).
In a linearly viscoelastic material energy is dissipated by plastic or viscous ow within the material.
So, the loading and unloading curves do not overlap, instead forming a closed hysteresis loop.
FIGURE 2.13 Effect of strain rate on elastic properties of a viscoelastic material.
0266/ ch02/frame Page 35 Thursday, October 21, 1999 2:00 PM
36 Mechanical Testing of Bone and the BoneImplant Interface
J. FAILURE, FRACTURE, FATIGUE, FATIGUE FRACTURE, BUCKLING, AND CRACKING
Failure is the degradation of a material property beyond a set limit,
13
or loss of material continuity.
Fatigue is the damage due to repetitive stresses below the ultimate stress. Fatigue is a slow
progressive process, as opposed to an acute, catastrophic event which results when the ultimate
strength of a material is surpassed. Typically, repetitive cyclical loading (smaller than the ultimate
strength) causes a crack through a material with subsequent separation of the object into pieces.
Fatigue fracture in whole bone is a common nding that frequently results from the stresses imposed
on the skeletal system by the muscular system during locomotion. The failure is referred to as
stress fracture and is more common in females than males. The mechanism of the injury commonly
occurs when there are acute bouts of strenuous exercise over an extended period of time which
FIGURE 2.14 A typical stress relaxation curve of a viscoelastic material.
FIGURE 2.15 A typical creep curve of a viscoelastic material.
0266/ ch02/frame Page 36 Thursday, October 21, 1999 2:00 PM
Basic Concepts of Mechanical Property Measurement and Bone Biomechanics 37
was immediately preceded by a relatively sedentary period. Typically, there are no prodromal
signs.
15
Fatigue fractures in vivo result when the remodeling process is outpaced by the microdamage
which accumulates with repetitive loading.
9,11
From a biomechanical point of view stress fracture
is most likely to occur when the bone is repeatedly loaded at the elastic limit region.
16
Endurance (fatigue) limit is the maximum stress that a material can sustain repeatedly without
failure. Some materials, especially steels, have an endurance limit below which the material will
withstand an innite number of cycles without failure. An SN curve is the stressnumber curve
representing the relationship between stress and minimum number of loadings at the stress required
to produce fracture (Figure 2.16). The SN curve is illustrated with stress plotted on the ordinate and
the logarithm of the corresponding number of cycles to cause failure plotted on the abscissa. The
fatigue life of a material can be affected microscopically and macroscopically by stress concentration,
specimen surface geometry, and the surrounding environment.
17,18
Two distinct features of the SN
curve are (1) the lower the stress magnitude the greater the number of cycles prior to failure, and (2)
the material will either demonstrate an endurance limit at or below where the material can be stressed
an innite number of times without fracture or the SN curve will curve monotonically downward
and the specimen eventually fracture no matter how low the applied load.
15
According to a report by Choi and Goldstein,
19
trabecular specimens had signicantly lower
moduli and lower fatigue strength than cortical specimens, despite their higher mineral density
values. Fracture surface and microdamage analyses illustrated different fracture and damage patterns
between trabecular and cortical bone tissue, depending upon their microstructural characteristics.
9
When long, slender columns are considered, similar to long bones in the body, collapse of the wall
of the tube laterally can produce local buckling failure, which leads to subsequent complete buckling
and material failure.
2
Cracking is incomplete loss of material continuity with near absence of unrecoverable strain,
dependent somewhat on the ductility of the specimen material.
13
When stress is applied through
plastic deformation, the material will ultimately break with complete molecular separation. Brittle
bone cracks more readily than bone that is healthier.
20
The crack may be initiated as a stress riser,
such as an indentation, scratch, or hole. Stresses near these areas can be concentrated, thereby
creating local material failure with ultimate propagation of the crack.
18
Cracks will spread when
energy release from the material as the crack spreads exceeds energy necessary to extend the crack
and/or when stress at the crack tip reaches a value that exceeds the cohesive strength of the atoms
just ahead of the crack. The latter parameter is known as the critical stress intensity factor.
15
FIGURE 2.16 A typical SN curve shows the fatigue characteristics of materials.
0266/ ch02/frame Page 37 Thursday, October 21, 1999 2:00 PM
38 Mechanical Testing of Bone and the BoneImplant Interface
III. MECHANICAL MODELING AND SIMULATION
Finite-element analysis (FEA), originally developed in the 1950s, is the most popular methodology
for modeling bone structure and its mechanical properties. FEA has had a profound impact on the
eld of orthopaedics and the modeling of bone by assisting researchers and physicians in the
quantitative analysis of bone and other complex biological structures. FEA is especially useful for
the modeling of irregularly shaped bones, such as vertebra and for identication of high-stress
areas on bone.
21
Basically, FEA makes use of simple shapes, known as elements (building blocks), which are
assembled to form complex geometric structures which are used to solve complex problems, most
of which can be represented with one or more partial differential equations.
22
The elements are
connected at points known as nodes. In a model, a nite number of elements (or shapes) are
connected at nodes to form a mathematical representation of a structure such as bone. The process
of dividing the structure into elements of interest to create the nite-element mesh is known as
discretization. The governing equations are only approximately satised at nodal points; therefore
the nite-element technique involves replacement of partial differential equations having an innite
number of degrees of freedom with a more discrete system with a nite, albeit sometimes large,
number of degrees of freedom.
22
As forces are applied to the model or as the model is deformed,
elaborate equations predict the stress and strain responses of the structure to loading. The complexity
of a nite-element model is determined by the imagination of its creator, the level of mathematical
sophistication, and computing power.
The nite-element method (FEM) originated as a tool to assist engineers in the design of
structures. The FEM approach often requires lengthy and complex calculations and therefore
became tractable only with the advances in computer technology. Originally, its use was restricted
to experts specializing in FEM who used large, mainframe computers to solve problems. As
computer technology advanced, FEM became more accessible to nonspecialists via commercially
available FE program packages. In recent years biomechanists have found nite-element modeling
to be a valuable tool for investigating a wide range of biological problems, such as designing better
articial hip joints.
The three basic phases of FEM are (1) creation of the model; (2) solution; and (3) validation
and interpretation of the results.
The goal of modeling is the creation of a mathematical nite description of nodes, elements,
material properties, boundary and interface conditions, and load application. This is the most labor-
intensive phase of FEA and can takes weeks or months, although computer technology has hastened
the modeling phase. One of the initial steps is determining whether the model will be one, two, or
three dimensional. Because bone is a three-dimensional structure, the model will include linear
(two nodes per side) and quadratic elements (three nodes per side). Common three-dimensional
elements include the linear eight-node brick and the quadratic 20-node brick, although tetrahedral
and linear wedge models are also available. The costs for a three-dimensional FEM are higher than
for a two-dimensional model, although limitations in the two-dimensional model frequently dictate
three-dimensional use. There are two primary questions that inuence whether a two-dimensional
or a three-dimensional model is used: What mechanical characteristic does the third element
provide? Is it necessary to solve the problem at hand?
22
The material property specications for FEA can be difcult when modeling bone due to the
inhomogeneity and anisotropic behavior of the material. In the biomechanics literature, bone is
mostly modeled as linear elastic material in a nonimpact loaded state.
Load specications and boundary and interface conditions are the nal tasks in modeling. Loads
include joint reaction forces, musculotendinous forces, and ligament forces such as exist during
gait, rising from a chair, or stairclimbing. Many of these loads (bending and torsion) have been
identied in the recent past in vivo as being out-of-plane for traditional two-dimensional FEA.
Symmetrical and asymmetrical boundary conditions consist of displacement constraints which are
0266/ ch02/frame Page 38 Thursday, October 21, 1999 2:00 PM
Basic Concepts of Mechanical Property Measurement and Bone Biomechanics 39
required to prevent rigid-body motion. For structures that are constructed of different materials the
interface can affect the system response and needs to be accounted for. For example, in a prosthetic
joint, cartilage, metal, cement, bone, and polyethylene regions need to be accounted for in the FEM.
22
The solution requires an FEA computer program that is based on the input data from the model.
Computer time, which is usually through a batch process, rather than interactive, is a function of
the model nodes and numbers of degrees of freedom per node. The solution phase is completed
when nodal displacements and derived quantities (stresses, strains) have been calculated and stored
in digital form.
FEA validation is assessed with two distinct issues, validity and model accuracy. That is, does
the model represent the true system of interest based on the modeling phase? Convergence can be
used to measure the accuracy of the model. Convergence testing involves repeated renement of
the nite-element mesh and subsequent reanalysis to determine how the changes affect the predic-
tions of the model. If possible, the validation should include comparisons of strain gauge data, in
vivo load-sensing implants, and other analytical solution techniques.
Interpretation of results involves interactive graphics based on a postprocessing program.
Contour plots and color fringe plots are common methods of stress and strain display. A common
pitfall, especially in orthopaedics, is overemphasis on stress and strain variables, because despite
the most accurate models the values are idealized at some level. The results of FEA need to be
interpreted based on meaningful measures: Will the bone fracture? Is the prosthesis appropriate for
the bone stock available? Will the prosthesis loosen as a result of activity?
Due to the variability and uncertainties with use of FEA one should view FEA as more of a
qualitative tool than as a quantitative measurement device, although when the FEA of today is
compared with the same technique from 20 years ago it is clear that the technique is becoming
more relevant, accessible, and valuable.
22
IV. SUMMARY
Bone is a nonlinear, viscoelastic, anisotropic, and heterogeneous material and therefore can be
complex to analyze mechanically. The fact that bone has the inherent ability to adapt continually
to metabolic and environmental changes in vivo provides even more complexity, further limiting
our understanding of specic bony mechanisms. Many of the basic mechanical concepts presented
in this chapter are currently in use and will continue to be utilized to explore the mechanical
behavior of bone and other biological tissues. Despite the fact that the major points in this chapter
were divided into sections, it is important to note that all of the concepts are interrelated in
challenging and fascinating ways. Analytical models, including FEA, which account for bone
geometry and its heterogeneous properties, depend on appropriate material and mechanical esti-
mates. While sophisticated, complex testing is demanded in some circumstances to answer relevant
questions, it is wise to begin testing with simple ideas and hypotheses. The long-term objective for
a majority of bone studies is to characterize and better understand the relationship between the
structure and mechanical behavior of bone. The central issue, however, for all researchers in
bioengineering is to identify the most appropriate variables for their area of interest and then to
blend them into understandable and useful contexts.
REFERENCES
1. Cowin, S.C., Mechanics of materials, in Bone Mechanics, Cowin, S.C., Ed., CRC Press, Boca Raton,
FL, 1989, 16.
2. Biewener, A.A., Overview of structural mechanics, in Biomechanics, Structures and System. A Prac-
tical Approach, Biewener, A.A., Ed., IRL Press, Oxford, 1992.
3. Frankel, V.H. and Burstein, A.H., Orthopaedic Biomechanics, Lea & Febiger, Philadelphia, 1970.
0266/ ch02/frame Page 39 Thursday, October 21, 1999 2:00 PM
40 Mechanical Testing of Bone and the BoneImplant Interface
4. Ashman, R.B., Experimental techniques, in Bone Mechanics, Cowin, S.C., Ed., CRC Press, Boca
Raton, FL, 1989, 91.
5. Vasu, R., Carter, D.R., and Harris, W.H., Stress distributions in the acetabular region I. Before and
after total joint replacement, J. Biomech., 15, 155, 1982.
6. Brodt, M.D., Swan, C.C., and Brown, T.D., Mechanical behavior of human morselized cancellous
bone in triaxial compression testing, J. Orthop. Res., 16, 43, 1998.
7. Keaveny, T.M. and Hayes, W.C., A 20-year perspective on the mechanical properties of trabecular
bone, J. Biomech. Eng., 115, 534, 1993.
8. Hayes, W.C. and Bouxsein, M.L., Biomechanics of cortical and trabecular bone: implications for
assessment of fracture risk, in Basic Orthopaedic Biomechanics, Mow, V.C. and Hayes, W.C., Eds.,
Lippincott-Raven, Philadelphia, 1997, Chapter 3.
9. Einhorn, T.A., Biomechanics of bone, in Principles of Bone Biology, Bilezikian, J.P., Raisz, L.G., and
Rodan, G.A., Eds., Academic Press, San Diego, CA, 1996, 25.
10. Whiting, W.C. and Zernicke, R.F., Biomechanical concepts, in Biomechanics of Musculoskeletal
Injury, Whiting, W.C. and Zernicke, R.F., Eds., Human Kinetics, Champaign, IL, 1998, 41.
11. Nordin, M. and Frankel, V.H., Biomechanics of whole bones and bone tissue, in Basic Biomechanics
of Skeletal System, Frankel, V.H. and Nordin, M., Eds., Lea & Febiger, Philadelphia, 1980, 15.
12. Burstein, A.H. and Wright, T.M., Fundamentals of Orthopaedic Biomechanics, Williams & Wilkins,
Baltimore, MD, 1994.
13. Black, J., Orthopaedic Biomaterials in Research and Practice, Churchill Livingstone, New York, 1988.
14. An, Y.H., Kang, Q., and Friedman, R.J., Mechanical symmetry of rabbit bones studied by bending
and indentation testing, Am. J. Vet. Res., 57, 1786, 1996.
15. Currey, J., The Mechanical Adaptations of Bones, Princeton University Press, Princeton, NJ, 1984.
16. Chamay, A., Mechanical and morphological aspects of experimental overload and fatigue in bone,
J. Biomech., 3, 263, 1970.
17. Caputo, A.A. and Standlee, J.P., Eds., Biomechanics in Clinical Dentistry, Quintessence Publishing,
Chicago, 1987.
18. Tencer, A.F. and Johnson, K.D., Biomechanics in Orthopedic Trauma. Bone Fracture and Fixation,
M. Dunitz, London, 1994.
19. Choi, K. and Goldstein, S.A., A comparison of the fatigue behavior of human trabecular and cortical
bone tissue, J. Biomech., 25, 1371, 1992.
20. Low, J. and Reed, A., Basic Biomechanics Explained, Butterworth Heinemann, London, 1996.
21. Ranu, H.S., The role of nite-element modeling in biomechanics, in Material Properties and Stress
Analysis in Biomechanics, Yettram, A.L., Ed., Manchester University Press, Manchester, 1989, 164.
22. Beaupr, G.S. and Carter, D.R., Finite element analysis in biomechanics, in Biomechanics Structure
and Systems, Biewener, A.A., Ed., Oxford University Press, Oxford, 1992, 150.
0266/ ch02/frame Page 40 Thursday, October 21, 1999 2:00 PM

41

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

3

Mechanical Properties of Bone

Yuehuei H. An

CONTENTS

I. Introduction ............................................................................................................................41
II. Mechanical Properties of Cortical Bone ...............................................................................42
A. General Mechanical Properties ........................................................................................42
B. Bone Density....................................................................................................................43
C. Porosity.............................................................................................................................47
D. Anisotropy and Heterogeneity .........................................................................................47
E. Single Osteons and Microspecimens ...............................................................................48
III. Mechanical Properties of Cancellous Bone...........................................................................50
A. Structural Properties.........................................................................................................50
B. Bone Density....................................................................................................................50
C. Microstructure ..................................................................................................................52
D. Anisotropy and Heterogeneity .........................................................................................55
E. Material Properties ...........................................................................................................56
IV. Continuum Assumption..........................................................................................................57
V. Summary ................................................................................................................................58
References ........................................................................................................................................58

I. INTRODUCTION

Before and during the process of testing bone mechanical properties, the researcher has to learn
the basic structural and mechanical properties of bone. These properties include cortical and
cancellous bone at the level of whole bones, bone tissues, osteons or trabeculae, bone lamellae,
and ideally the nano- or ultrastructure such as collagen bers, brils, molecules, and mineral
components.

1,2

These basic structural and mechanical data are searchable in the journal literature
and also have been written about in numerous textbooks.
Bone has a hierarchical structure and coherent mechanical properties as proposed initially by
Katz

3

in 1970s and further developed recently by Rho et al.

4

and Hofer et al. (see Chapter 8). In
general, this hierarchical structure and the related mechanical properties can be investigated and
considered at the ve levels shown in Table 3.1.
The author believes that one of the reasons to separate whole bone and bone tissue blocks
is that the mechanical determinants of these two levels of structures are different. For example,
the bending mechanical properties of a long bone are determined by its tubular shape and bone
densities, while that of a cortical cut beam are by bone densities and osteonal direction (see
Section IIB). As pointed out by Katz

3

nearly 20 years ago, it is essential to understand the
formfunction (structuremechanics) relationship of the bone specimen to be tested. The bone
hierarchical composite modeling is based on both the structural evaluations and mechanical
measurements at different levels.

0266/ ch03/frame Page 41 Thursday, October 21, 1999 2:06 PM

42

Mechanical Testing of Bone and the BoneImplant Interface

However, due to the scope of this text, this chapter covers only the basic mechanical properties
of cortical and cancellous bone at whole bone, bone tissue, and microstructural levels (osteon and
trabeculae), the effects of porosity and densities on mechanical properties, the anisotropic and
heterogeneous characteristics of bone mechanical properties, and some basic considerations of the
validity of the continuum assumption commonly used in mechanical testing.
For more detailed descriptions of basic and theoretical mechanics of bone, one can refer to
several books, including

Strength of Biological Materials

by Yamada (1970),

5



Mechanical Prop-
erties of Bone

by Evans (1973),

6



The Mechanical Adaptations of Bones

by Currey (1984),

7



Bone
Mechanics,

edited by Cowin (1999),

8

and

Skeletal Tissue Mechanics

by Martin et al.,

9

and several
book chapters or journal articles by Nordin and Frankel (1980),

10

Albright (1987),

11

Einhorn
(1996),

12

Hayes and Bouxsein (1997),

13

Whiting and Zernicke (1998),

14

and Rho et al.

4

II. MECHANICAL PROPERTIES OF CORTICAL BONE
A. G

ENERAL

M

ECHANICAL

P

ROPERTIES

For mechanical testing, cortical bones are often used as a whole bone or tailored into beams or
rods. A whole diaphyseal bone is commonly tested using bending and torsional tests. A beam is a
bar or rod with constant cross-sectional shape and area, which can be spherical, square, or rectan-
gular. A variable beam is a beam with inconsistent cross-sectional shape and area, such as long
bones. A cantilever beam is a beam that is xed at one end and usually used for cantilever bending
tests. A dumbbell sample is a dumbbell-shaped bone specimen made specically for mechanical
testing, such as tensile or torsional tests. The dense nature of cortical bone determines its strong
and stiff mechanical properties compared with cancellous bone. For comparison, Figure 3.1 gives
the elastic modulus and strength of bone (cortical bone) and several other common tissues and
biomaterials.

TABLE 3.1
The Hierarchical Levels of Bone

Level Elements (Specimens) Main Factors Determining Bone Strength

Macrostructure
(whole bone)
Femur, humerus, vertebrae,
frontal bone, phalangeal
bones, calcaneous, etc.
Macrostructure such as tubular shape, cross-sectional area, and
porosity of long bone, cortical bone-covered vertebrae, or the
irregular pelvic bone
Architecture
(tissue level)
Compact bone or cancellous
bone blocks, cylinders,
cubes, or beams
Densities, porosity, the orientations of osteon, collagen bers, or
trabeculae
Microstructure
(osteonal or
trabecular level)
Osteons, trabeculae Loading direction, with maximum strength along their long axis
Submicrostructure
(lamellar level)
Lamella, large collagen
bers
Collagen-HA brils are formed into large collagen bers or lamellar
sheets with preferred directions. The orientations of the brils
dene directions of maximum and minimum strengths for a
primary loading direction
Ultrastructure
(nanostructure)
Collagen bril and molecule,
mineral components
HA crystals are embedded between the ends of adjoining collagen
molecules; this composite of rigid HA and exible collagen
provides a material that is superior in mechanical properties to
either of them alone, more ductile than hydroxyapatite, allowing
the absorption of more energy, and more rigid than collagen,
permitting greater load bearing
Adapted from the work by Rho et al.

4

and Hofer et al. in Chapter 8.

0266/ ch03/frame Page 42 Thursday, October 21, 1999 2:06 PM

Mechanical Properties of Bone

43

The mechanical properties of cortical bone depend on the type of mechanical testing. According
to the data collected in Table 3.2, the strength and elastic modulus by compression tests range from
133 to 295 MPa (200

36 MPa) and from 14.7 to 34.3 GPa (average 23

4.8 GPa), respectively.
The strength and elastic modulus by tensile tests range from 92 to 188 MPa (average 141

28
MPa) and from 7.1 to 28.2 GPa (average 19.6

6.2 GPa), respectively. The strength and elastic
modulus by torsional tests range from 53 to 76 MPa (average 65

9 MPa) and from 3.1 to 3.7
GPa, respectively. The tensile strength is about

that of compression strength. The torsional
(shear) strength is approximately

to

of the values of the longitudinal strength (tested by
bending, tensile, or compressive tests) (Table 3.3). And the torsional (shear) modulus is only about

to

of the longitudinal modulus. Although the tensile test is the standard method for testing
mechanical properties of cortical bone, bending tests are used the most often.
The bending strength and elastic modulus of cortical bone ranges from 35 to 283 MPa and
from 5 to 23 GPa, respectively (excluding the questionable values marked with

c

, Tables 3.3 and
3.4). Note the signicant differences between the two levels.
As mentioned earlier, the bending mechanical properties of a long bone are determined by its
tubular shape and bone densities, while that of cortical cut beams are by bone densities and osteonal
direction. Table 3.5 contains the average values of the individual reports listed in Tables 3.3 and
3.4. It clearly shows that the values of strength and elastic modulus of the two levels of bone tissue
are different. Both the strength and elastic modulus of whole bone are about 60% of that of cortical
bone beams. One should be aware of the fact that besides the true difference between the two levels
of structures, there are several other factors possibly playing important roles, such as the equations
used for calculations, the specimen aspect ratio (larger for beams), or the size of the specimens
tested. These speculations have already been partially addressed by Sedlin and Hirsch.

25

Further
studies are still needed to determine the relevance of each of these factors.

B. B

ONE

D

ENSITY

The material density of cortical bone is the wet weight divided by the specimen volume. It is a
function of both the porosity and mineralization of the bone materials. Cortical bone has an average

FIGURE 3.1

Illustrations of the elastic modulus (A) and strength (B) of bone (cortical bone) and other
common tissues and biomaterials.

0266/ ch03/frame Page 43 Thursday, October 21, 1999 2:06 PM

44

Mechanical Testing of Bone and the BoneImplant Interface

apparent density of approximately 1.9 g/cm

3

.

1,2

For cortical bone, apparent density and material
density are basically the same, as there is no marrow space in compact bone. Therefore, cortical
bone density is commonly used to describe the density of cortical bone. There is a positive
correlation between apparent density of cortical bone and its mechanical properties.

45

The true
meaning of bone mineral density (BMD) is bone mineral mass per unit bone volume, or ash

TABLE 3.2
Mechanical Properties of Human and Bovine Cortical Bones Tested by Compression, Tensile,
and Torsional Testing (all at the tissue level)

Strength Elastic
Species Bone Specimen Dimensions (MPa)
Modulus
(GPa) Reference
Compression Test

Human Femur 2

6 mm dumbbell 167215

a

14.719.7

a

Reilly 1974

15

2

6 mm dumbbell 179209

a

15.418.6

a

Burstein 1976

16

3 mm diam. cylindrical dumbbell 205206

a

Cezayirlioglu 1985

17

Tibia 2

6 mm dumbbell 183213

a

24.534.3

a

Burstein 1976

16

3 mm diam. cylindrical dumbbell 192213

a

Cezayirlioglu 1985

17

Bovine Femur 3.8

2.3

76 mm dumbbell 133 24.127.6

a

McElhaney 1964

18

2

6 mm dumbbell 240295

a

21.931.4

a

Reilly 1974

15

Tibia 4

5 mm rectangular 165 23.8

2.2 Simkin 1973

19

2

6 mm dumbbell 228

31 20.9

3.26 Reilly 1974

15

3 mm diam. cylindrical dumbbell 217

27 Cezayirlioglu 1985

17

Means

SD (

n

b

) 200

36 (10) 23.0

4.8 (7)

Tensile Test

Human Femur 3.8

2.3

76 mm dumbbell 66107

a

10.920.6

a

Evans 1951

20

2

6 mm dumbbell 107140

a

11.419.7

a

Reilly 1974

15

2

6 mm dumbbell 120140

a

15.617.7

a

Burstein 1976

16

3 mm diam. cylindrical dumbbell 133136

a

Cezayirlioglu 1985

17

Tibia 2

6 mm dumbbell 145170

a

18.929.2

a

Burstein 1976

16

1.7

1.8

25 mm beam 162

15 19.7

2.4 Vincetelli 1985

21

3 mm diam. cylindrical dumbbell 154158

a

Cezayirlioglu 1985

17

Bovine Femur 3.8

2.3

76 mm dumbbell 92 20.5 McElhaney 1964

18

2

6 mm dumbbell 129182

a

23.130.4

a

Reilly 1974

15

3 mm diam. cylindrical dumbbell 162

14

a

Cezayirlioglu 1985

17

Tibia 4

30 mm dumbbell 136 7.1

1.1 Simkin 1973

19

2

2 6 mm dumbbell 152 17 21.6 5.3 Reilly 1974
15
2 2 6 mm dumbbell 188 9 28.2 6.4 Burstein 1975
22
Means SD (n
b
) 141 28 (13) 19.6 6.2 (10)
Torsional Test
Human Femur ? 53 Hazama 1964
23
? 54 0.6 3.2 Yamada 1970
5
2 2 6 mm dumbbell 3.13.7
a
Reilly 1974
15
3 3 6 mm dumbbell 6571
a
Reilly 1975
24
3 mm diam. cylindrical dumbbell 6871
a
Cezayirlioglu 1985
17
Tibia 3 mm diam. cylindrical dumbbell 6671
a
Cezayirlioglu 1985
17
Bovine Femur 3 3 6 mm dumbbell 6267
a
Reilly 1975
24
3 mm diam. cylindrical dumbbell 76 6 Cezayirlioglu 1985
17
Means SD (n
b
) 65 9 (7) 3.3 0.1 (2)
a
Range of average values from different subjects.
b
Number of data sets from the literature.
0266/ ch03/frame Page 44 Thursday, October 21, 1999 2:06 PM
Mechanical Properties of Bone 45
density if an ashing (or burning) method is used.
26,46
Similarly, the true meaning of bone mineral
content (BMC) describes the ratio of unit weight of the mineral portion to dry bone unit weight
and is frequently reported as a percentage.
BMD and BMC are positively correlated with the strength and stiffness of various bones, such
as human ulna,
47
human femur and tibia,
45,48
bovine femur and tibia,
32,46
feline femur,
49
and a wide
variety of animal bones.
50
Using tension testing of wet bovine Haversian cortical bone, Burstein
et al.
22
demonstrated the role of mineral content on mechanical strength. Progressive surface decal-
cication of this bone with dilute hydrochloric acid resulted in progressive decreases in the tension
yield point and the ultimate stress with no change in the yield strain or ultimate strain unless
decalcication was complete. Their ndings are consistent with an elastic-plastic model for the
mineral phase of bone tissue in which the mineral contributes the major portion of the tension yield
strength. Currey
51,52
and Schafer and Burr
46
found that small changes in the amount or mineral
density of cortical bone exert a more pronounced inuence on its elastic property than would similar
changes in trabecular bone. Currey studied the relationship between the bending and tensile elastic
modulus of cortical bone from 17 vertebrate species to porosity and mineralization. He found that
these two factors together accounted for 84% of the stiffness variation.
30
Recently, the inuence
of wet and dry apparent density, percentage of mineral on the tension and shear fracture toughness
of tubular bone have been studied by Yeni et al.
45
They found that compositional parameters
altogether can explain 35 to 59% of the variation in fracture toughness of the cortical bone.
Many reports have shown linear or exponential increases in bone stiffness with increasing
mineralization, such as the one proposed by Schafer and Burr:
46
E = 89.1M3.91 (3.1)
where E is compressive elastic modulus and M is mineralization of bovine cortical bone.
TABLE 3.3
Bending Properties of Cortical Bones at the Bone Tissue Level
Species Bone Specimen
Strength
(MPa)
Elastic Modulus
(GPa) Reference
Human Femur 2 5 50 mm beam 181 15.5 Sedlin 1966
25
3 3 30 mm beam 103238
a
9.8215.7
a
Keller 1990
26
0.4 5 7 mm beam 225 28 12.5 2.1 Lotz 1991
27
2.0 3.4 40 mm 142170
a
9.114.4
a
Curry 1997
28
Cattle Femur 2 3.5 30 mm beam 18.5 2.8 Curry 1988
29
2 4 35 mm beam 228 5 19.4 0.7 Curry 1988
30
2 30.4 mm beam 209 13 18.1 0.5 Curry 1995
31
Tibia 4 4 35 mm beam 14.1 Simkin 1973
19
4 10 80 mm beam 230 18 21.0 1.9 Martin 1993
32
Horse Femur 2 2 40 mm beam 204247
a
17.119.9
a
Schryver 1978
33
2 3.5 30 mm beam 21.2 1.9 Curry 1988
29
Radius 2 2 40 mm beam 217249
a
16.220.2
a
Schryver 1979
33
Metacarpus 2 2 40 mm beam 226240
a
17.018.4
a
Schryver 1978
33
3MT, 3MC
b
1.8 4.5 70 mm 195226
a
1416
a
Bigot 1996
34
Sheep Metatarsus 2 3.5 30 mm beam 18.9 2.2 Curry 1988
29
Donkey Radius 2 3.5 30 mm beam 17.6 2.0 Curry 1988
29
Goose Femur 0.75 0.75 25 mm beam 232283
a
16.920.7
a
McAlister 1983
35
a
Range of average values.
b
Third metatarsus and third metacarpus.
0266/ ch03/frame Page 45 Thursday, October 21, 1999 2:06 PM
46 Mechanical Testing of Bone and the BoneImplant Interface
With the development of modern absorptiometric techniques, BMD and BMC can be measured
noninvasively. Such methods include radiographic absorptiometry (RA), single photon absorpti-
ometry (SPA), dual energy absorptiometry (DEA), or dual-energy X-ray absorptiometry (DEXA),
quantitative computed tomography (QCT), micro-CT (CT), peripheral CT (pCT), magnetic res-
onance imaging (MRI), and ultrasound methods.
53
Extensive studies have been done on the corre-
lation between cortical bone densities measured by above-mentioned methods, such as BMD and
BMC, and the mechanical properties of bone.
48,53,54
These methods are commonly used for predict-
ing fracture risk and for the diagnosis of osteoporosis.
TABLE 3.4
Bending Properties of Cortical Bones at the Whole Bone Level
Species Bone Specimen
Strength
(MPa)
Elastic Modulus
(GPa) Reference
Monkey Tibia Whole bone 9.0 1.3 Kasra 1994
36
Dog Humerus Whole bone 193 35 2.7 0.6
c
Kaneps 1997
37
Pig Femur Whole bone 39.9 0.37
c
Crenshaw 1981
38
Rib Whole bone 35.6 2.24
c
Crenshaw 1981
38
3MC
b
Whole bone 37.2 0.22
c
Crenshaw 1981
38
Cat Femur Whole bone 36 9.47 7.1 0.9 Ayers 1996
39
Tibia Whole bone 60.5 12 11.4 3.2 Ayers 1996
39
Rabbit Femur Whole bone 130 5 13.6 0.4 An 1996
40
Whole bone 88 20 10.7 2.5 Ayers 1996
39
Tibia Whole bone 195 6 21.3 0.7 An 1996
40
Whole bone 192 47 23.3 7.0 Ayers 1996
39
Humerus Whole bone 167 5 13.3 0.6 An 1996
40
Rat Femur Whole bone 180 6 6.9 0.3 Jrgensen 1991
41
Whole bone 134 4 8.0 0.4 Barengolts 1993
42
Whole bone 153 45 4.9 4 Ejersted 1993
43
Mouse Femur Whole bone 104173
a
8.811.4
a
Simske 1992
44
Whole bone 40 13 5.3 1.8 Ayers 1996
39
Tibia Whole bone 78 12 8.9 0.2 Ayers 1996
39
a
Range of average values.
b
Third metatarsus and third metacarpus.
c
Value is questionable.
TABLE 3.5
Comparison of Mechanical Properties of Whole Long Bones
and Cortical Bone Rods or Beams
Whole Bones Rods or Beams Difference Ratio, % P Value
Strength (MPa) 125 58 (n = 14
a
) 202 40 (n = 13
b
) 77 61.8 <0.05
Elastic modulus (GPa) 10.3 5.7 (n = 15
a
) 16.5 3.6 (n = 16
b
) 6.2 62.4 <0.05
a
n = the number of strength or elastic modulus values from 14 to 15 studies or experiments chosen from
Table 3.4 (the three sets of questionable values are not included).
b
Number of values taken from Table 3.3.
0266/ ch03/frame Page 46 Thursday, October 21, 1999 2:06 PM
Mechanical Properties of Bone 47
C. POROSITY
The strong effects of porosity of cortical bone on mechanical properties have been well studied.
9
It is easy to understand that a more porous bone has a weaker mechanical strength. Porosity (p) is
dened as the ratio of void volume to total volume, which is commonly measured on two dimen-
sional histologic sections (traditionally point counting)
29,46
or X rays.
55
In cortical bone, the mechan-
ical properties are affected by Haversion canals and related resorption cavities and vascular chan-
nels. There are reports on the correlations of porosity and mechanical properties, such as the
equation proposed by Schafer and Burr
46
on bovine cortical bone using tensile tests:
E = 33.9 (1 p)
10.9
(3.2)
where E is the elastic modulus and (1 p) is the bone volume fraction, and the equation by Currey
for cortical bone of a wide variety of species under tension is
E = 23.4 (1 p)
5.74
(3.3)
McElhaney et al.
56
found that
E = 12.4 (1 p)
3
(3.4)
for compression of human skull bones.
Figure 3.2 shows signicant porosity of rabbit femoral cortical bone, a localized osteopenia
caused by an experimental inammatory knee arthritis.
57
The bone loss leads to subendosteal
cavitation and conversion of the inner portion of the cortex to a trabecular-like structure. These
changes reduced the bending strength of the femur from 97 21 MPa to 80 16 MPa and the
elastic modulus from 8.3 1.5 to 7.1 1.4 GPa.
D. ANISOTROPY AND HETEROGENEITY
The meaning of anisotropic is nonuniform or unevenly distributed, and is the opposite of isotropic.
The structural anisotropy determines the mechanical anisotropy. The mechanical properties of
cortical bone depend on loading directions of the testing method. The longitudinal (0 normally
FIGURE 3.2 Illustrations of the significant porosity of rabbit femoral cortical bone, a local osteopenia caused
by an experimental inammatory knee arthritis.
0266/ ch03/frame Page 47 Thursday, October 21, 1999 2:06 PM
48 Mechanical Testing of Bone and the BoneImplant Interface
the weight-bearing direction) elastic modulus is the highest, the transverse (90 lateral directions)
elastic modulus is the lowest, and the moduli of the specimens taken at any angles in between 0
and 90 have intermittent magnitudes (Figure 3.3).
2,4,58-60
The reason for the anisotropic phenomenon
is the longitudinally oriented collagen bers and osteons.
9,32
Sasaki et al.
61
studied the orientation
of hydroxyapatite (HA) crystals in bovine femoral mineral using X-ray pole gure analysis. It was
found that the c-axis of HA generally orients parallel to the longitudinal axis of bone (bone axis)
and a signicant amount of c-axis was oriented in other directions, in particular, perpendicular to
the bone axis. They concluded that the anisotropy in mechanical properties of bone can be well
explained by taking account of the nonlongitudinal (off-bone) axial distribution of orientation of
bone mineral. Similar ndings were also reported by Turner et al.
62
Cortical bone is mechanically heterogeneous, which had already been documented nearly 50
years ago by Evans and Lebow in 1951.
20
They found that the middle third of the femoral shaft
has the highest ultimate strength and elastic modulus and the lower third has the lowest average
strength and modulus. The lateral quadrant of the shaft has the highest ultimate tensile strength
while the anterior quadrant has the lowest. McAlister and Moyle
35
studied the ultimate compressive
strength and modulus of elasticity of femoral cortical bone (femur) from adult geese by sex and
by quadrant by compressing small right circular cylinders and the bending strength and elastic
modulus by a three-point bending test on small rectangular prisms. They found that both the females
and males had a signicantly lower modulus and strength in the anterior quadrant as compared
with other quadrants by both compression and bending tests. Their data are in accordance with
that of Evans and Lebow.
20
Using ultrasonic techniques, Rho
122
measured the elastic properties of
eight human tibiae to determine and map the elastic properties of cortical and cancellous bone.
The study showed cortical bone to be at least orthotropic in its material symmetry. The mechanical
properties of cortical bone are more homogeneous along the length than around the circumference.
The variations in the properties around the quadrant of cortical bone are small, less than 10%.
E. SINGLE OSTEONS AND MICROSPECIMENS
The mechanical properties of single osteons or cortical bone microspecimens have been tested by
several groups using bending, compression, tensile, and torsional methods (Table 3.6). Based on
the angles of the ber directions of two successive lamellae, Ascenzi et al.
63,64
classied three osteon
types: (1) transversal (T) osteon having marked transversal spiral course of ber bundles in
successive lamellae; (2) alternate (A) osteon having ber bundles in one lamella making an angle
FIGURE 3.3 Cortical bone cylindrical specimens taken from different directions from the cortex, showing
the anisotropic characteristics of cortical bone.
0266/ ch03/frame Page 48 Thursday, October 21, 1999 2:06 PM
Mechanical Properties of Bone 49
of nearly 90 with the ber bundles of the next one; and (3) longitudinal (L) osteon having marked
longitudinal spiral course of ber bundles in successive lamellae (Figure 3.4). They found that the
A osteons are more resistant to bending stress. The Type L osteons are stronger in tension and
Type A osteons are weaker in tension. Under compressive loading, Type T osteons are stronger
than Type A and L osteons. Based on the data included in Table 3.6, most osteonal or microcortical
bone specimens have their elastic moduli falling in the range of 2 to 12 GPa. Variations in the
mechanical properties of osteons with different collagen ber directions suggest that they are
individually adapted to enhance locally the ability of bone to support a particular type of stress.
Microanisotropy can be used to describe the uneven distribution of structures at the osteonal
level.
62
Based on the study by Turner et al.,
62
the anisotropic elastic symmetry of osteonal bone
reects the ultrastructural organization of collagen brils and mineral crystals within the osteons
as well as the lamellar microstructure. Turner et al. reported measurements of bone anisotropy using
high-precision acoustic microscopy. The elastic properties of canine femoral bone specimens were
TABLE 3.6
Mechanical Properties of Single Osteons or Microcortical Bone Specimens
Specimen Test Method
Elastic Modulus
(GPa) Reference
Human tibia Acoustic speaker (alternate tension and compression) 11.7 Fraska 1981
65
Three-point-bending (100 170 1500 m beam) 3.077.63 Choi 1990
67

Human femur
L osteon Three-point-bending 2.32 1.20 Ascenzi 1990
67
A osteon Three-point-bending 2.69 0.93
L osteon Tension 11.7 5.8 Ascenzi 1968
66
A osteon Tension 5.5 2.6
L osteon Compressive 6.3 1.8 Ascenzi 1968
63
A osteon Compressive 7.4 1.6
T osteon Compressive 9.3 1.6
Goose femur Compressive (0.8 2.5 mm cylinder) 13.2 34 McAlister 1983
35
FIGURE 3.4 Illustrations of the three types of osteons transversal (T) osteon, alternate (A) osteron, and
longitudinal (L) osteron found by Ascenzi et al. (Adapted from Ascenzi, A. and Bonuci, E., The compressive
properties of single osteons, Anat. Rec., 161, 377, 1968.)
0266/ ch03/frame Page 49 Thursday, October 21, 1999 2:06 PM
50 Mechanical Testing of Bone and the BoneImplant Interface
measured at 10 increments from the long axis of the bone. Half of the bone specimens subsequently
were demineralized in EDTA solution, the other half were decollagenized in sodium hypochlorite
solution, and the acoustic measurements were repeated. It was found the elastic symmetry of
osteonal bone deviates signicantly from orthotropic theory, supporting the hypothesis that the
lamellar microstructure forms a rotated plywood.
68
The principal orientation of bone mineral was
along the long axis of the bone, while bone collagen appeared to be aligned at a 30 angle to the
long axis. The misalignment between the mineral and the collagen suggests that (1) a substantial
percentage of the mineral is extrabrillar, and (2) the alignment of extrabrillar mineral is governed
by external inuences, e.g., mechanical stresses.
III. MECHANICAL PROPERTIES OF CANCELLOUS BONE
The porous nature of cancellous bone, with bony trabecular columns and struts and marrow-lled
pores or cavities (a two-phase structure)
69-72
lends itself to a mechanical description by both
structural and material properties. The mechanical properties of cancellous bone are determined
by several major factors, including apparent density and ash density, trabecular connectivity, and
location and function.
A. STRUCTURAL PROPERTIES
The structural properties of cancellous bone are commonly measured by compression, tensile, or
bending tests. The common phrase mechanical properties of cancellous bone means the structural
properties. It is known that the strength and elastic modulus by tensile tests are smaller than that
by compression tests. For example, the strength by tensile test is approximately 60% of the value
by compression test reported by Kaplan et al.,
73
and the elastic modulus by tensile test is approx-
imately 70% of the value by compression test reported by Keaveny et al.
74,75
According to the selected data from the literature (Table 3.7), the values of strength and elastic
moduli of cancellous bone are 1.5 to 38 MPa and 10 to 1570 MPa, respectively. The structural properties
of cancellous bone are much smaller than those of cortical bone. The average values of elastic modulus
are several hundred mega pascal for cancellous bone,
76
compared with 5 to 21 GPa for cortical bone.
29
B. BONE DENSITY
There is a strong correlation between the mechanical properties of cancellous bone, both for strength
and stiffness, and its apparent density and mineral (or ash) density. The apparent density of
cancellous bone ranges from 0.14 to 1.10 g/cm
3
(average: 0.62 g/cm
3
, n = 16; see Table 3.7). The
compressive strength (! in MPa) of cancellous bone is related to its apparent density (" in g/cm
3
)
by a power law of the form:
13
! = 60"
2
(3.5)
The compressive modulus (E in MPa) of cancellous bone is related to the apparent density
(" in g/cm
3
) by:
13
E = 2915"
2
(3.6)
Selected data of ash densities of human and animal cancellous bones are also listed in Table 3.2;
they range from 0.19 to 0.56 g/cm
3
with an average of 0.37 0.10 g/cm
3
(n = 12), which is about
60% of the value of apparent density as shown in the following equation:
"
Ash
# 0.6 "
Apparent
(3.7)
0266/ ch03/frame Page 50 Thursday, October 21, 1999 2:06 PM
Mechanical Properties of Bone 51
TABLE 3.7
Mechanical Properties and Densities of Cancellous Bone Tissues
Bone Specimen
Ultimate
Strength
(MPa)
Elastic
Modulus
(MPa)
Apparent
Density
(g/cm
3
)
Ash
Density
(g/cm
3
) Reference
Human
Femoral head 8 mm diam. cylinder 9.3 4.5 900 710 Martens 1983
77
Proximal femur 8 mm diam. cylinder 6.6 6.3 616 707 Martens 1983
77
Distal femur 8 mm cube 5.6 3.8 298 224 0.43 0.15 0.26 0.08 Kuhn 1989
78
10.3 mm diam., 5 mm cylinder 1.545
a
10500
a
0.24 0.09 Carter 1977
69
5 mm diam./7.5 mm cylinder 5.96 1031058
a
0.46 Odgaard 1989
79
Proximal tibia 7.5:7.5 mm cylinder 5.3 2.9 445 257 Linde 1989
80
Vertebral body Dimensions ? cylinders 165 110 0.14 0.06 Keaveny 1997
81
Monkey
Femoral head 5 mm diam./6 mm cylinder 23.1 5.4 372 54 Kasra 1994
36
Cattle
Distal femur 5.5 mm diam./8 mm cylinder 8.5 4.2 117 61 Poumarat 1993
82
Proximal tibia 15 mm cube, ultrasonic method 648 430 0.41 0.16 Rho 1997
83
Proximal humerus Dimensions ? cylinders 1570 628 0.71 0.22 Keaveny 1997
81
Vertebral body 6 mm diam./7.5 mm cylinder 7.1 3.0 173 97 0.45 0.09 0.19 0.06 Swartz 1991
84
Dog
Femoral head 5 mm cube 12 5.8 435 Vahey 1987
85
Distal femur 8 mm cube 7.1 4.6 209 140 0.44 0.16 0.26 0.08 Kuhn 1989
78
4 mm diam./5 mm cylinder 1328
b
210394
b
0.690.98 0.400.56
b
Kuhn 1998
86
Proximal tibia 4 mm diam./5 mm cylinder 524
b
106426
b
0.410.83b 0.220.44
b
Kuhn 1998
86
12.5 mm diam./10 mm cylinder 301850 Sumner 1994
87
5 mm cube 3441278 Sumner 1994
87
Humeral head 4 mm diam./5 mm cylinder 18 6 350 171 0.84 0.17 0.43 0.06 Kuhn 1998
86
Distal humerus 6 mm diam./15 mm cylinder 13 3 1490 300 Kaneps 1997
37
Vertebral body 5 mm diam./8 mm cylinder 10.1 2.6 530 40 Acito 1994
88
Goat
Femoral head 4 mm diam./5 mm cylinder 19.2 6.9 502 268 0.91 0.04 0.48 0.03 An 1998
89
Distal femur 4 mm diam./5 mm cylinder 14.123.5
b
399429
b
0.540.66
b
0.320.40
b
An 1998
89
Proximal tibia 4 mm diam./5 mm cylinder 24.726.1
b
532566
b
0.931.1b 0.500.56b An 1998
89
Humeral head 4 mm diam./5 mm cylinder 10.0 1.0 247 20 0.75 0.03 0.36 0.01 An 1998
89
Sheep
Femoral neck 8 mm diam./10 mm cylinder 3.2 0.3 2.0 0.2
c
Geusens 1996
90
Vertebral body 7 mm diam./9 mm cylinder 23.6 4.4 Deloffre 1995
91
.5 mm diam./9 mm cylinder 22.3 7.1 1510 784 0.60 0.16 0.37 0.11 Mitton 1997
92
Pig
Vertebral body 7 mm diam./5 mm cylinder 27.5 3.4 1080 470 0.46 0.04 Mosekilde 1987
93
Rabbit
Epiphyseal Ground bone surfaces 3581 An 1996
40
long bones Indentation test
Rat
Epiphyseal Ground bone surfaces 3871 An 1997
94
long bones Indentation test
a
Range of values.
b
Range of average values from different parts.
c
Value is questionable (too low).
0266/ ch03/frame Page 51 Thursday, October 21, 1999 2:06 PM
52 Mechanical Testing of Bone and the BoneImplant Interface
In the authors laboratory, a set of data was generated on the correlations between the mechanical
properties and bone densities of canine trabecular bone.
86
Cancellous bone specimens (n = 72) were
taken from different locations including humeral head, femoral head, femoral condyle, and upper
tibia. Figure 3.5 shows the strong correlation between the mechanical properties and their apparent
density. Similarly, the mechanical properties of cancellous bone correlate well with their ash density
(Figure 3.6).
57
C. MICROSTRUCTURE
The more commonly used parameters for trabecular bone structures or architecture include (1) BV
(bone volume) or TBA (trabecular bone area, which is the trabecular surface area divided by the
total area in m
2
); (2) Tb.Th (trabecular thickness, the average thickness of trabeculae in m); and
(3) Tb.Sp (trabecular separation, the average distance between trabeculae, representing the amount
of marrow space in m).
95
FIGURE 3.5 The correlation between apparent density and properties of cancellous bone.
FIGURE 3.6 The correlation between ash density and properties of cancellous bone.
0266/ ch03/frame Page 52 Thursday, October 21, 1999 2:06 PM
Mechanical Properties of Bone 53
Common parameters for trabecular bone spatial connectivity include Tb.N (trabecular number,
the average number of continuous trabecular elements encountered per unit area), Ho.N (hole
number, the average number of holes per unit area), N.Nd (trabecular node number; nodes: trabe-
cular branch points), N.Tm (trabecular terminus number; termini: trabecular end points), and Nd/Tm
ratio. Most of the parameters can be measured using specialized imaging software based on a single
histological section (two dimensional), serial sections, or serial image scanning (three dimensional).
In a local osteopenic model reported from the authors laboratory,
57
the signicant reduction
of cancellous bone strength (26 8 MPa compared with the control side 68 15 MPa using an
indentation test) could be explained by the reduction of trabecular BV or TBA and increased
perforation and disconnectivity of the trabecular tissue (Figure 3.7). According to morphometrical
analysis, the cancellous bones showed obvious perforation and disconnectivity in a very short period
of time, as indicated by TBV, Nd, free end (Tm), continuous CTE, Ho.N, Nd/Tm ratio, and Euler
number (calculated by deducting Ho.N. from Th.N
96
). These changes may represent a rapid bone
loss
97
featuring perforation and disconnection of the trabecular network and increased size of
marrow cavities. Signicant correlations have been found between the mechanical and morpholog-
ical parameters (Table 3.8).
Recent development in three-dimensional imaging of cancellous bone has made possible true
three-dimensional quantication of trabecular architecture. This provides a signicant improvement
of the tools available for studying and understanding the mechanical functions of cancellous bone.
Goldstein et al.
98
utilized a three-dimensional, microcomputed tomography (CT) system to meas-
ure trabecular plate thickness, trabecular plate separation, trabecular plate number, surface-to-
volume ratio, bone volume fraction, anisotropy, and connectivity in isolated specimens of trabecular
bone. The results of these studies demonstrate that in normal bone, more than 80% of the variance
in its mechanical behavior can be explained by measures of density and orientation.
98
Odgaard
et al.
99,100
argued that connectivity and architectural anisotropy (fabric) are of special interest in
mechanicsarchitecture relations. They addressed the possible signicance of trabecular connec-
tivity for the mechanical quality of cancellous bone. By using the detailed three-dimensional
reconstructions as input for microstructural nite-element models, the complete elastic properties
of the trabecular architecture were obtained and maximum and mean stiffness could be calculated.
Volume fraction and true three-dimensional architectural measurements of connectivity density and
surface density were determined. Connectivity density was determined in an unbiased manner by
the Euler number. By using multiple regression analysis it was found that volume fraction explained
by far the greatest part (84 to 94%) of the variation in both mean and maximum stiffness. When
connectivity density and surface density were included, the correlations increased marginally to 89
to 95%. Recently, Kinney and Ladd
101
used a nite-element model to explore the relationship
FIGURE 3.7 A scanning electron microscope (SEM) image showing the osteopenic changes of the upper
tibial trabecular bone in a rabbit knee inammatory arthritis model.
0266/ ch03/frame Page 53 Thursday, October 21, 1999 2:06 PM
54 Mechanical Testing of Bone and the BoneImplant Interface
FIGURE 3.8 Illustration of a test for screw-holding powers of bovine cancellous bone from different directions,
showing the anisotropic behavior of cancellous bone. (A) The sampling site and orientation. (B) The directions
of screw insertion and the average values of pullout strength (n = 16).
TABLE 3.8
Correlation Analysis (Pearson Correlation Coefcient)
between Mechanical and Morphometric Parameters
a
Ultimate Load Stiffness Ultimate Strength
TBV (%) 0.796
b
0.888
c
0.796
Tb.Th (m) 0.918
c
0.869
b
0.918
c
Tb.sp 0.459 0.562 0.462
Ho.N/mm
2
0.823
b
0.880
b
0.823
b
Euler number/mm
2
0.832
b
0.892
b
0.883
c
Nd/mm
2
0.868
b
0.950
d
0.868
Tm/mm
2
0.879
b
0.925 0.881
Nd/Tm ratio 0.940
d
0.831
c
0.940
d
Note: A means a negative correlation.
a
n = 6; degrees of freedom n$ = n 2 = 4; one-way (for positive r) or
two-way (negative r) analysis.
b
P < 0.05.
c
P < 0.01.
d
P < 0.005.
Adapted from Kang et al., J. Mater. Sci. Mater. Med., 9, 463, 1998.
0266/ ch03/frame Page 54 Thursday, October 21, 1999 2:06 PM
Mechanical Properties of Bone 55
between connectivity density and the elastic modulus of trabecular bone. Although no functional
relationship was found between connectivity and elastic modulus, there was a linear relationship,
after a full cycle of atrophy and recovery, between the loss of elastic modulus and the overall loss
of connectivity. The results indicate that recovery of mechanical function depends on preserving
or restoring trabecular connectivity.
D. ANISOTROPY AND HETEROGENEITY
Cancellous bone is anisotropic based on its trabecular morphology.
102
Several investigations have
addressed the orthogonal or anisotropic mechanical properties of cancellous bone of both human
and animals.
85,103-107
The strength and elastic modulus of cancellous bone depend on the direction
of the load employed, as normally measured at SI (superior-interior), AP (anterior-posterior), or
ML (medial-lateral) directions. Ciarelli et al.
106
found the highest overall mean of elastic moduli
of human long bone metaphyseal locations to be in the SI direction, which is about 2.5 times the
value at the AP direction. The AP direction is higher than the ML direction. An earlier study using
vertebral cancellous bone specimens by Galante et al.
103
also showed a similar pattern. In a recent
study in the authors laboratory, it was found that the screw pullout strength of bovine cancellous
bone also depends on the direction of the screw insertion (loading direction). The strength was the
strongest (55 5 MPa) at the SI direction (0), the weakest (37 5 MPa) at the lateral direction
(90), and intermediate (43 4 MPa) at a direction of 45.
108
This phenomenon may be explained
by a columnstrut model proposed by the authors group (see Chapter 21).
109
Figure 3.8 shows the
effects of the directions of screw insertion on the pullout strength of cancellous bone in bovine
distal femoral cancellous bone.
As stated by Goldstein,
76
nearly 40 years ago Evans and King documented the mechanical
properties of trabecular bone from multiple locations in the proximal human femur. Since that time,
many investigators have cataloged the distribution of trabecular bone material properties from
multiple locations within the human skeleton to include the femur, tibia, humerus, radius, vertebral
bodies, calcaneus, and iliac crest. According to the data list summarized by Goldstein (21 sets of
data generated using compression tests), the average values of strength and elastic modulus of
human cancellous bone from different locations are 6.6 to 36.2 MPa and 130 to 1080 Mpa,
respectively.
76
Both linear and power functions have been found to explain the relationship between
trabecular bone density and material properties.
76
In the authors laboratory, the strength and elastic
modulus of epiphysometaphyseal bones of animals, such as rats,
94
rabbits,
39
dogs,
86,110
and goats,
110
have been investigated using compression and indentation tests. Generally, for both humans and
animals, the cancellous bones of lower limbs (hind limbs) are stronger and stiffer than those of
upper limbs (front limbs). This heterogeneous mechanical characteristic of cancellous bone is
determined by functional adaptation.
76,111
Cancellous bone is also heterogeneous at a given location. At the metaphysioepiphyseal area
of long bones, more trabecular bone material is located at the subchondral bone plate and gradually
becomes less concentrated toward the diaphysis (Figure 3.9). This structural pattern determines a
decreasing mechanical strength from the subchondral bone plate toward the medullary canal. In
our recent study, cylinders (4 mm diam. 5 mm length) from two levels of cancellous bone from
canine medial femoral condyles were measured for strength and elastic modulus using compression
tests.
86
The results showed that the distal level (the one close to the joint surface) has a much higher
strength and modulus (28 7 vs. 19 5 MPa). A similar nding was also demonstrated in a study
on the glenoid.
112
If a cut is made below the upper tibial joint surface, the mechanical property was found to be very
inhomogeneous.
113
Figure 3.10 also shows that the high-strength areas, tested by an indentation test,
114
are where the intimate joint contact occurs. At the peripheral area, the bone is stronger than the
immediate neighboring areas, which is caused by the strengthening effect of cortical bone shell.
115
In
one of Rhos studies,
122
he found that variations in the properties around the quadrant of tibial cortical
0266/ ch03/frame Page 55 Thursday, October 21, 1999 2:06 PM
56 Mechanical Testing of Bone and the BoneImplant Interface
bone are small, less than 10%, while the differences in the properties around the circumference of
cancellous bone are more apparent, approximately ve times those of cortical bone. The elastic
properties of cancellous bone exhibited inhomogeneity and some consistency pattern along both the
length and the circumference. Similar morphology is also seen in the vertebral body. Very inhomoge-
neous bone structure and mechanical properties are also found in almost every possible location, such
as the human glenoid,
112
human proximal and distal femur,
116-118
human upper tibia,
118
canine upper
tibia,
87
human patella,
104
vertebrae,
117
calcaneus,
117
and even the pig mandibular condyle.
119
Table 3.7 lists articles on the mechanical properties of animal cancellous bones, which include
studies of bovine, canine, or goat distal femur, proximal tibia, and vertebrae determined by com-
pression test, or of canine, rabbit, or rat epiphysometaphyseal bones examined using indentation
test.
57,86,94
E. MATERIAL PROPERTIES
The material properties of cancellous bone are dened by the intrinsic properties of individual trabe-
culae, which have been measured by mechanical testing of single trabeculae using methods such as
buckling analysis,
120
compression test,
121
microtensile test,
122
cantilever test plus nite-element model-
ing,
123
nite-element modeling,
124
or ultrasound methods.
122
The elastic modulus of trabecular bone
material (individual trabeculae) is about 10 to 30% less than that of cortical bone (Table 3.9). For
FIGURE 3.9 Rabbit upper tibial cross section at coronal plane (SEM), showing the heterogeneous trabecular
structure from the subchondral bone plate toward the medullary canal. Note the dense structure at the
subchondral plate (above Line 1) and the loose trabeculae below Line 3.
0266/ ch03/frame Page 56 Thursday, October 21, 1999 2:06 PM
Mechanical Properties of Bone 57
example, the elastic modulus is 14.8 GPa for trabeculae and 20.7 GPa for cortical bone measured by
an ultrasonic technique and 10.4 and 18.6 GPa, respectively, using a microtensile test.
122
IV. CONTINUUM ASSUMPTION
Most existing analyses of strength and stiffness of both cortical and cancellous bone assume that
it can be modeled as a continuum.
127
One basic requirement of using the continuum assumption is
that the minimum dimension of the sample must be signicantly larger than the dimension of its
structural subunits. Harrigan et al.
128
developed a criterion for the validity of this assumption. The
limitations of the continuum assumption appear in two areas: near biologic interfaces and in areas
of large stress gradients, such as subchondral bone, the transitional area between cortical shell and
FIGURE 3.10 Illustrations of the heterogeneous distribution of bone strength values. Note the high-strength
areas are in the medial and lateral condyles where the tibia and femur articular surfaces contact.
TABLE 3.9
Mechanical Properties of Cancellous Bone Material
Bone specimen Test Method Elastic modulus (GPa) Ref.
Human distal femur Buckling 8.69 3.17 (dry) Runkle 1975
120
Human proximal tibia Buckling 11.38 (wet) Townsend 1975
121
Human femur Ultrasound 12.7 2.0 (wet) Ashman 1988
125
Bovine femur Tensile 10.9 1.6 (wet) Ashman 1988
125
Human femur, tibia Cantilever 8.7 (6.211.2) Menta 1989
123
Human iliac crest Tensile 0.8 0.4 Ryan 1989
71
Human upper tibia Three-point bending 4.59 Choi 1990
67
Human tibia Four-point bending 5.7 1.3 Choi 1990
126
Tensile 10.4 3.5 Rho 1993
122
0266/ ch03/frame Page 57 Thursday, October 21, 1999 2:06 PM
58 Mechanical Testing of Bone and the BoneImplant Interface
cancellous bone, and bone under intimate joint contact areas. These limitations are explored using
a probabilistic line-scanning model for density measurements, resulting in an estimate of density
accuracy as a function of line length which is experimentally veried. For cancellous bone, within
a distance of three to ve trabeculae (about 300 to 1500 m in length), a continuum model is not
valid. For compression testing on cancellous bone, Linde
129
suggested that a specimen dimension
larger than 5 mm can fulll the requirement of continuum assumption. For cortical bone, because
of its dense structure the limit for continuum assumption can be smaller.
V. SUMMARY
Bone is an elastic, anisotropic, heterogeneous, and composite material. The determinants of bone
mechanical properties include (1) its density (apparent density and mineral density); (2) porosity
(vascular canals in cortical bone and marrow space in cancellous bone); and (3) microscopic
structure, such as cortical bone architecture (primary and secondary osteons), osteonal structure
(compositions of lamellae with different collagen ber arrangement), trabecular structure (trabecular
orientation, trabecular bone volume, and trabecular connectivity), and collagen ber orientation.
127
Due to the limited scope of this text, only selected mechanical properties of bone obtained
using traditional mechanical testing methods and their determinants at macro- or microstructural
levels are included in this chapter. This is not enough to understand fully the biomechanics of bone
tissues. As mentioned by Rho et al.,
4
detailed descriptions of the structural features of bone abound
in the literature; however, the mechanical properties of bone, in particular those at the micro- and
nanostructural level (material level), remain poorly understood. Therefore, further investigations of
mechanical properties at the materials level,
68
in addition to the studies at the structural level
are needed to ll the gap in our present knowledge and to achieve a complete understanding of the
mechanical properties of bone.
Although one should know the limitations of traditional mechanical testing methods (which
often underestimate the true values of strength and stiffness of the materials) and also that of
noninvasive techniques, such as ultrasound, 3-D imaging, or nite-element analysis, most of the
mechanical data of bone obtained by traditional testing methods remains valid and serves as basic
data for validating any noninvasive techniques.
Bone is a heterogenous and anisotropic material and cannot be treated simply as a homogeneous
material like Daro foam (formed by mixing isocyanate and a resin, polyol) or saw bone. However,
certain assumptions can be applied to bone specimens within reason in order to achieve useful data
of mechanical properties of bone.
REFERENCES
1. Spatz, H.C., OLeary, E.J., and Vincent, J.F., Youngs moduli and shear moduli in cortical bone, Proc.
R. Soc. Lond. B. Biol. Sci., 263, 287, 1996.
2. Ashman, R.B., Experimental techniques, in Bone Mechanics, Cowin, S.C., Ed., CRC Press, Boca
Raton, FL, 1989, 91.
3. Katz, J.L., The structure and biomechanics of bone, Symp. Soc. Exp. Biol., 34, 137, 1980.
4. Rho, J.Y., Kuhn-Spearing, L., and Zioupos, P., Mechanical properties and the hierarchical structure
of bone, Med. Eng. Phys., 20, 92, 1998.
5. Yamada, H., Ed., Strength of Biological Materials, Williams & Wilkins, Baltimore, MD, 1970.
6. Evans, F.G., Mechanical Properties of Bone, Charles C Thomas, Springeld, IL, 1973.
7. Currey, J., The Mechanical Adaptations of Bones, Princeton University Press, Princeton, NJ, 1984.
8. Cowin, S.C., Ed., Bone Mechanics, CRC Press, Boca Raton, FL, 1999.
9. Martin, R.B., Burr, D.B., and Sharkey, N.A., Skeletal Tissue Mechanics, Springer, New York, 1998.
10. Nordin, M. and Frankel, V.H., Biomechanics of whole bones and bone tissue, in Basic Biomechanics
of Skeletal System, Frankel, V.H. and Nordin, M., Eds., Lea & Febiger, Philadelphia, 1980, 15.
0266/ ch03/frame Page 58 Thursday, October 21, 1999 2:06 PM
Mechanical Properties of Bone 59
11. Albright, J.A., Bone: physical properties, in The Scientic Basis of Orthopaedics, Albright, J.A. and
Brand, R.A., Eds., Appleton & Lange, Norwalk, CT, 1987, 213.
12. Einhorn, T.A., Biomechanics of bone, in Principles of Bone Biology, Bilezikian, J.P., Raisz, L.G., and
Rodan, G.A., Eds., Academic Press, San Diego, CA, 1996, 25.
13. Hayes, W.C. and Bouxsein, M.L., Biomechanics of cortical and trabecular bone: implications for
assessment of fracture risk, in Basic Orthopaedic Biomechanics, Mow, V.C. and Hayes, W.C., Eds.,
Lippincott-Raven, Philadelphia, 1997, Chapter 3.
14. Whiting, W.C. and Zernicke, R.F., Biomechanical concepts, in Biomechanics of Musculoskeletal
Injury, Whiting, W.C. and Zernicke, R.F., Eds., Human Kinetics, Champaign, IL, 1998, 41.
15. Reilly, D.T., Burstein, A.H., and Frankel, V.H., The elastic modulus for bone, J. Biomech., 7, 271, 1974.
16. Burstein, A.H., Reilly, D.T., and Martens, M., Aging of bone tissue: mechanical properties, J. Bone
Joint Surg. Am., Vol. 58, 82, 1976.
17. Cezayirlioglu, H., Bahniuk, E., Davy, D.T., and Heiple, K.G., Anisotropic yield behavior of bone
under combined axial force and torque, J. Biomech., 18, 61, 1985.
18. McElhaney, J.H., Fogle, J., Byars, E., and Weaver, G., Effect of embalming on the mechanical
properties of beef bone, J. Appl. Physiol., 19, 1234, 1964.
19. Simkin, A. and Robin, G., The mechanical testing of bone in bending, J. Biomech., 6, 31, 1973.
20. Evans, F.G. and Lebow, M., Regional differences in some of the physical properties of the human
femur, J. Appl. Physiol., 3, 563, 1951.
21. Vincentelli, R. and Grigorov, M., The effect of Haversian remodeling on the tensile properties of
human cortical bone, J. Biomech., 18, 201, 1985.
22. Burstein, A.H., Zika, J.M., Heiple, K.G., and Klein, L., Contribution of collagen and mineral to the
elastic-plastic properties of bone, J. Bone Joint Surg. Am., Vol. 57, 956, 1975.
23. Hazama, H., Study on the torsional strength of the compact substance of human beings, J. Kyoto Pref.
Med. Univ., 60, 167, 1956.
24. Reilly, D.T. and Burstein, A.H., The elastic and ultimate properties of compact bone tissue, J. Biomech.,
8, 393, 1975.
25. Sedlin, E.D. and Hirsch, C., Factors affecting the determination of the physical properties of femoral
cortical bone, Acta Orthop. Scand., 37, 29, 1966.
26. Keller, T.S., Mao, Z., and Spengler, D.M., Youngs modulus, bending strength, and tissue physical
properties of human compact bone [see comments], J. Orthop. Res., 8, 592, 1990.
27. Lotz, J.C., Gerhart, T.N., and Hayes, W.C., Mechanical properties of metaphyseal bone in the proximal
femur, J. Biomech., 24, 317, 1991.
28. Currey, J.D., Foreman, J., Laketic, I. et al., Effects of ionizing radiation on the mechanical properties
of human bone, J. Orthop. Res., 15, 111, 1997.
29. Currey, J.D., The effect of porosity and mineral content on the Youngs modulus of elasticity of
compact bone, J. Biomech., 21, 131, 1988.
30. Currey, J.D., The effects of drying and re-wetting on some mechanical properties of cortical bone,
J. Biomech., 21, 439, 1988.
31. Currey, J.D., Brear, K., Zioupos, P., and Reilly, G.C., Effect of formaldehyde xation on some
mechanical properties of bovine bone, Biomaterials, 16, 1267, 1995.
32. Martin, R.B. and Boardman, D.L., The effects of collagen ber orientation, porosity, density, and
mineralization on bovine cortical bone bending properties, J. Biomech., 26, 1047, 1993.
33. Schryver, H.F., Bending properties of cortical bone of the horse, Am. J. Vet. Res., 39, 25, 1978.
34. Bigot, G., Bouzidi, A., Rumelhart, C., and Martin-Rosset, W., Evolution during growth of the mechan-
ical properties of the cortical bone in equine cannon-bones, Med. Eng. Phys., 18, 79, 1996.
35. McAlister, G.B. and Moyle, D.D., Some mechanical properties of goose femoral cortical bone,
J. Biomech., 16, 577, 1983.
36. Kasra, M. and Grynpas, M.D., Effect of long-term ovariectomy on bone mechanical properties in
young female cynomolgus monkeys, Bone, 15, 557, 1994.
37. Kaneps, A.J., Stover, S.M., and Lane, N.E., Changes in canine cortical and cancellous bone mechanical
properties following immobilization and remobilization with exercise, Bone, 21, 419, 1997.
38. Crenshaw, T.D., Peo, E.R., Jr., Lewis, A.J., et al., Inuence of age, sex and calcium and phosphorus
levels on the mechanical properties of various bones in swine, J. Anim. Sci., 52, 1319, 1981.
0266/ ch03/frame Page 59 Thursday, October 21, 1999 2:06 PM
60 Mechanical Testing of Bone and the BoneImplant Interface
39. Ayers, R.A., Miller, M.R., Simske, S.J., and Norrdin, R.W., Correlation of exural structural properties
with bone physical properties: a four species survey, Biomed. Sci. Instrum., 32, 251, 1996.
40. An, Y.H., Kang, Q., and Friedman, R.J., Mechanical symmetry of rabbit bones studied by bending
and indentation testing, Am. J. Vet. Res., 57, 1786, 1996.
41. Jorgensen, P.H., Bak, B., and Andreassen, T.T., Mechanical properties and biochemical composition
of rat cortical femur and tibia after long-term treatment with biosynthetic human growth hormone,
Bone, 12, 353, 1991.
42. Barengolts, E.I., Lathon, P.V., Curry, D.J., and Kukreja, S.C., Effects of endurance exercise on bone
histomorphometric parameters in intact and ovariectomized rats, Bone Miner., 26, 133, 1994.
43. Ejersted, C., Andreassen, T.T., Oxlund, H., et al., Human parathyroid hormone (1-34) and (1-84)
increase the mechanical strength and thickness of cortical bone in rats, J. Bone Miner. Res., 8, 1097,
1993.
44. Simske, S.J., Guerra, K.M., Greenberg, A.R., and Luttges, M.W., The physical and mechanical effects
of suspension-induced osteopenia on mouse long bones, J. Biomech., 25, 489, 1992.
45. Yeni, Y.N., Brown, C.U., and Norman, T.L., Inuence of bone composition and apparent density on
fracture toughness of the human femur and tibia, Bone, 22, 79, 1998.
46. Schafer, M.B. and Burr, D.B., Stiffness of compact bone: effects of porosity and density, J. Biomech.,
21, 13, 1988.
47. Jurist, J.M. and Foltz, A.S., Human ulnar bending stiffness, mineral content, geometry and strength,
J. Biomech., 10, 455, 1977.
48. Stromsoe, K., Hoiseth, A., Alho, A., and Kok, W.L., Bending strength of the femur in relation to
noninvasive bone mineral assessment, J. Biomech., 28, 857, 1995.
49. Shah, K.M., Goh, J.C., Karunanithy, R., et al., Effect of decalcication on bone mineral content and
bending strength of feline femur, Calcif. Tissue Int., 56, 78, 1995.
50. Currey, J.D., Physical characteristics affecting the tensile failure properties of compact bone,
J. Biomech., 23, 837, 1990.
51. Currey, J.D., The mechanical consequences of variation in the mineral content of bone, J. Biomech.,
2, 1, 1969.
52. Currey, J.D., The relationship between the stiffness and the mineral content of bone, J. Biomech., 2,
477, 1969.
53. Wong, D.M. and Sartoris, D.J., Noninvasive methods for assessment of bone density, architecture,
and biomechanical properties: fundamental concepts, in Osteoporosis: Diagnosis and Treatment,
Sartoris, D.J., Ed., Marcel Dekker, New York, 1996, 201.
54. Van der Perre, G. and Lowet, G., Physical meaning of bone mineral content parameters and their
relation to mechanical properties, Clin. Rheumatol., 13 (Suppl. 1), 33, 1994.
55. Lazenby, R., Porositygeometry interaction in the conservation of bone strength, J. Biomech., 19,
257, 1986.
56. McElhaney, J.H., Alem, N., and Roberts, V., A porous block model for cancellous bones, ASME Publ.,
70-WA1BHF-2, 1970.
57. Kang, Q., An, Y.H., Butehorn, H.F., and Friedman, R.J., Morphological and mechanical study of the
effects of experimentally induced inammatory knee arthritis on rabbit long bones, J. Mater. Sci.
Mater. Med., 9, 463, 1998.
58. Hirsch, C. and Silva, O.D., The effect of orientation on some mechanical properties of femoral cortical
specimens, Acta Orthop. Scand., 38, 45, 1967.
59. Boneld, W. and Grynpas, M.D., Anisotropy of the Youngs modulus of bone, Nature., 270, 453, 1977.
60. Katz, J.L., Anisotropy of Youngs modulus of bone, Nature, 283, 106, 1980.
61. Sasaki, N., Matsushima, N., Ikawa, T., et al., Orientation of bone mineral and its role in the anisotropic
mechanical properties of bone transverse anisotropy, J. Biomech., 22, 157, 1989.
62. Turner, C.H., Chandran, A., and Pidaparti, R.M., The anisotropy of osteonal bone and its ultrastructural
implications, Bone, 17, 85, 1995.
63. Ascenzi, A. and Bonucci, E., The compressive properties of single osteons, Anat. Rec., 161, 377, 1968.
64. Ascenzi, A., Baschieri, P., and Benvenuti, A., The bending properties of single osteons, J. Biomech.,
23, 763, 1990.
65. Frasca, P., Jacyna, G., Harper, R., and Katz, J.L., Strain dependence of dynamic Youngs modulus for
human single osteons, J. Biomech., 14, 691, 1981.
0266/ ch03/frame Page 60 Thursday, October 21, 1999 2:06 PM
Mechanical Properties of Bone 61
66. Ascenzi, A. and Bonucci, E., The tensile properties of single osteons, Anat. Rec., 158, 375, 1967.
67. Choi, K., Kuhn, J.L., Ciarelli, M.J., and Goldstein, S.A., The elastic moduli of human subchondral,
trabecular, and cortical bone tissue and the size-dependency of cortical bone modulus, J. Biomech.,
23, 1103, 1990.
68. Weiner, S. and Traub, W., Bone structure: from angstroms to microns, FASEB J., 6, 879, 1992.
69. Carter, D.R. and Hayes, W.C., The compressive behavior of bone as a two-phase porous structure,
J. Bone Joint Surg. Am., Vol. 59, 954, 1977.
70. Carter, D.R. and Hayes, W.C., Bone compressive strength: the inuence of density and strain rate,
Science, 194, 1174, 1976.
71. Ryan, S.D. and Williams, J.L., Tensile testing of rodlike trabeculae excised from bovine femoral bone,
J. Biomech., 22, 351, 1989.
72. Gibson, L.J., The mechanical behavior of cancellous bone, J. Biomech., 18, 317, 1985.
73. Kaplan, S.J., Hayes, W.C., Stone, J.L., and Beaupre, G.S., Tensile strength of bovine trabecular bone,
J. Biomech., 18, 723, 1985.
74. Keaveny, T.M., Wachtel, E.F., Ford, C.M., and Hayes, W.C., Differences between the tensile and
compressive strengths of bovine tibial trabecular bone depend on modulus [see comments], J. Bio-
mech., 27, 1137, 1994.
75. Keaveny, T.M., Guo, X.E., Wachtel, E.F., et al., Trabecular bone exhibits fully linear elastic behavior
and yields at low strains, J. Biomech., 27, 1127, 1994.
76. Goldstein, S.A., The mechanical properties of trabecular bone: dependence on anatomic location and
function, J. Biomech., 20, 1055, 1987.
77. Martens, M., Van Audekercke, R., Delport, P., et al., The mechanical characteristics of cancellous
bone at the upper femoral region, J. Biomech., 16, 971, 1983.
78. Kuhn, J.L., Goldstein, S.A., Ciarelli, M.J., and Matthews, L.S., The limitations of canine trabecular
bone as a model for human: a biomechanical study, J. Biomech., 22, 95, 1989.
79. Odgaard, A., Hvid, I., and Linde, F., Compressive axial strain distributions in cancellous bone
specimens, J. Biomech., 22, 829, 1989.
80. Linde, F., Hvid, I., and Pongsoipetch, B., Energy absorptive properties of human trabecular bone
specimens during axial compression, J. Orthop. Res., 7, 432, 1989.
81. Keaveny, T.M., Pinilla, T.P., Crawford, R.P., et al., Systematic and random errors in compression
testing of trabecular bone, J. Orthop. Res., 15, 101, 1997.
82. Poumarat, G. and Squire, P., Comparison of mechanical properties of human, bovine bone and a new
processed bone xenograft, Biomaterials, 14, 337, 1993.
83. Rho, J.Y., Flaitz, D., Swarnakar, V., and Acharya, R.S., The characterization of broadband ultrasound
attenuation and fractal analysis by biomechanical properties, Bone, 20, 497, 1997.
84. Swartz, D.E., Wittenberg, R.H., Shea, M., et al., Physical and mechanical properties of calf lumbo-
sacral trabecular bone, J. Biomech., 24, 1059, 1991.
85. Vahey, J.W., Lewis, J.L., and Vanderby, R., Jr., Elastic moduli, yield stress, and ultimate stress of
cancellous bone in the canine proximal femur, J. Biomech., 20, 29, 1987.
86. Kang, Q., An, Y.H., and Friedman, R.J., The mechanical properties and bone densities of canine
cancellous bones, J. Mater. Sci. Mater. Med., 9, 1998.
87. Sumner, D.R., Willke, T.L., Berzins, A., and Turner, T.M., Distribution of Youngs modulus in the
cancellous bone of the proximal canine tibia, J. Biomech., 27, 1095, 1994.
88. Acito, A.J., Kasra, M., Lee, J.M., and Grynpas, M.D., Effects of intermittent administration of
pamidronate on the mechanical properties of canine cortical and trabecular bone, J. Orthop. Res., 12,
742, 1994.
89. An, Y.H., Kang, Q., and Friedman, R.J., The mechanical properties and bone densities of goat
cancellous bones, unpublished data, 1998.
90. Geusens, P., Boonen, S., Nijs, J., et al., Effect of salmon calcitonin on femoral bone quality in adult
ovariectomized ewes, Calcif. Tissue Int., 59, 315, 1996.
91. Deloffre, P., Hans, D., Rumelhart, C., et al., Comparison between bone density and bone strength in
glucocorticoid-treated aged ewes, Bone, 17, 409S, 1995.
92. Mitton, D., Rumelhart, C., Hans, D., and Meunier, P.J., The effects of density and test conditions on
measured compression and shear strength of cancellous bone from the lumbar vertebrae of ewes, Med.
Eng. Phys., 19, 464, 1997.
0266/ ch03/frame Page 61 Thursday, October 21, 1999 2:06 PM
62 Mechanical Testing of Bone and the BoneImplant Interface
93. Mosekilde, L., Kragstrup, J., and Richards, A., Compressive strength, ash weight, and volume of
vertebral trabecular bone in experimental uorosis in pigs, Calcif. Tissue Int., 40, 318, 1987.
94. An, Y.H., Zhang, J.H., Kang, Q., and Friedman, R.J., Mechanical properties of rat epiphyseal cancel-
lous bones studied by indentation test, J. Mater. Sci. Mater. Med., 8, 493, 1997.
95. Partt, A.M., Drezner, M.K., Glorieux, F.H., et al., Bone histomorphometry: standardization of nomen-
clature, symbols, and units. Report of the ASBMR Histomorphometry Nomenclature Committee,
J. Bone Miner. Res., 2, 595, 1987.
96. Compston, J.E., Connectivity of cancellous bone: assessment and mechanical implications [editorial],
Bone, 15, 463, 1994.
97. Partt, A.M., Age-related structural changes in trabecular and cortical bone: cellular mechanisms and
biomechanical consequences, Calcif. Tissue Int., 36, S123, 1984.
98. Goldstein, S.A., Goulet, R., and McCubbrey, D., Measurement and signicance of three-dimensional
architecture to the mechanical integrity of trabecular bone, Calcif. Tissue Int., 53, S127, 1993.
99. Odgaard, A., Three-dimensional methods for quantication of cancellous bone architecture, Bone, 20,
315, 1997.
100. Kabel, J., Odgaard, A., van Rietbergen, B., and Huiskes, R., Connectivity and the elastic properties
of cancellous bone, Bone, 24, 115, 1999.
101. Kinney, J.H. and Ladd, A.J., The relationship between three-dimensional connectivity and the elastic
properties of trabecular bone, J. Bone Miner. Res., 13, 839, 1998.
102. Whitehouse, W.J., The quantitative morphology of anisotropic trabecular bone, J. Microsc., 101(2),
153, 1974.
103. Galante, J., Rostoker, W., and Ray, R.D., Physical properties of trabecular bone, Calcif. Tissue Res.,
5, 236, 1970.
104. Townsend, P.R., Raux, P., Rose, R.M., et al., The distribution and anisotropy of the stiffness of
cancellous bone in the human patella, J. Biomech., 8, 363, 1975.
105. Williams, J.L. and Lewis, J.L., Properties and an anisotropic model of cancellous bone from the
proximal tibial epiphysis, J. Biomech. Eng., 104, 50, 1982.
106. Ciarelli, M.J., Goldstein, S.A., Kuhn, J.L., et al., Evaluation of orthogonal mechanical properties and
density of human trabecular bone from the major metaphyseal regions with materials testing and
computed tomography, J. Orthop. Res., 9, 674, 1991.
107. Njeh, C.F., Hodgskinson, R., Currey, J.D., and Langton, C.M., Orthogonal relationships between
ultrasonic velocity and material properties of bovine cancellous bone, Med. Eng. Phys., 18, 373, 1996.
108. An, H.Y., Kang, Q., Friedman, R.J., and Young, F.A., The effect of microstructure of cancellous bone
on screw pullout strength, Trans. Soc. Biomater., 20, 385, 1997.
109. An, H.Y. and Draughn, R.A., Mechanical properties and testing methods of bone, in Animal Models
in Orthopaedic Research, An, Y.H. and Friedman, R.J., Eds., CRC Press, Boca Raton, FL, 1999,
Chapter 8.
110. An, Y.H., Kang, Q., and Friedman, R.J., et al., Do mechanical properties of epiphyseal cancellous
bones vary?, J. Invest. Surg., 10, 221, 1997.
111. Goldstein, S.A., Wilson, D.L., Sonstegard, D.A., and Matthews, L.S., The mechanical properties of
human tibial trabecular bone as a function of metaphyseal location, J. Biomech., 16, 965, 1983.
112. Mansat, P., Barea, C., Hobatho, M.C., et al., Anatomic variation of the mechanical properties of the
glenoid, J. Shoulder Elbow Surg., 7, 109, 1998.
113. Ashman, R.B., Rho, J.Y., and Turner, C.H., Anatomical variation of orthotropic elastic moduli of the
proximal human tibia, J. Biomech., 22, 895, 1989.
114. Kang, Q., An, Y.H., and Friedman, R.J., Effects of multiple freezingthawing cycles on ultimate
indentation load and stiffness of bovine cancellous bone, Am. J. Vet. Res., 58, 1171, 1997.
115. Hvid, I., Jensen, J., and Nielsen, S., Contribution of the cortex to epiphyseal strength. The upper tibia
studied in cadavers, Acta Orthop. Scand., 56, 256, 1985.
116. Brown, T.D. and Ferguson, A.B., Jr., Mechanical property distributions in the cancellous bone of the
human proximal femur, Acta Orthop. Scand., 51, 429, 1980.
117. Augat, P., Link, T., Lang, T.F., et al., Anisotropy of the elastic modulus of trabecular bone specimens
from different anatomical locations, Med. Eng. Phys., 20, 124, 1998.
118. Behrens, J.C., Walker, P.S., and Shoji, H., Variations in strength and structure of cancellous bone at
the knee, J. Biomech., 7, 201, 1974.
0266/ ch03/frame Page 62 Thursday, October 21, 1999 2:06 PM
Mechanical Properties of Bone 63
119. Teng, S. and Herring, S.W., Anatomic and directional variation in the mechanical properties of the
mandibular condyle in pigs, J. Dent. Res., 75, 1842, 1996.
120. Runkle, J.C. and Pugh, J., The micromechanics of cancellous bone. II. Determination of the elastic
modulus of individual trabeculae by a buckling analysis, Bull. Hosp. Joint Dis., 36, 2, 1975.
121. Townsend, P.R., Rose, R.M., and Radin, E.L., Buckling studies of single human trabeculae,
J. Biomech., 8, 199, 1975.
122. Rho, J.Y., Ashman, R.B., and Turner, C.H., Youngs modulus of trabecular and cortical bone material:
ultrasonic and microtensile measurements, J. Biomech., 26, 111, 1993.
123. Mente, P.L. and Lewis, J.L., Experimental method for the measurement of the elastic modulus of
trabecular bone tissue, J. Orthop. Res., 7, 456, 1989.
124. van Rietbergen, B., Weinans, H., Huiskes, R., and Odgaard, A., A new method to determine trabecular
bone elastic properties and loading using micromechanical nite-element models, J. Biomech., 28,
69, 1995.
125. Ashman, R.B. and Rho, J.Y., Elastic modulus of trabecular bone material, J. Biomech., 21, 177, 1988.
126. Choi, K. and Goldstein, S.A., A comparison of the fatigue behavior of human trabecular and cortical
bone tissue, J. Biomech., 25, 1371, 1992.
127. Martin, R.B., Determinants of the mechanical properties of bones [published erratum appears in
J. Biomech. 25, 1251, 1992], J. Biomech., 24, 79, 1991.
128. Harrigan, T.P., Jasty, M., Mann, R.W., and Harris, W.H., Limitations of the continuum assumption in
cancellous bone, J. Biomech., 21, 269, 1988.
129. Linde, F., Elastic and viscoelastic properties of trabecular bone by a compression testing approach,
Dan. Med. Bull., 41, 119, 1994.
0266/ ch03/frame Page 63 Thursday, October 21, 1999 2:06 PM
0266/ ch03/frame Page 64 Thursday, October 21, 1999 2:06 PM

65

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

4

Factors Affecting Mechanical
Properties of Bone

Peter Zioupos, Chris W. Smith, and Yuehuei H. An

CONTENTS

I. Introduction ............................................................................................................................65
II. Systemic or

in Vivo

Factors Affecting the Mechanical Properties of Bone .........................66
A. Age ...................................................................................................................................66
B. Sex ....................................................................................................................................68
C. Species..............................................................................................................................68
D. Composition Porosity, Density, Mineralization..........................................................68
1. Cortical Bone...............................................................................................................68
2. Cancellous Bone..........................................................................................................70
E. Function............................................................................................................................72
1. Microstructure vs. Function........................................................................................72
2. Activity Levels ............................................................................................................74
3. Weightlessness.............................................................................................................74
F. Hormones .........................................................................................................................75
1. Sex Hormones .............................................................................................................75
2. Growth Hormone, Insulin-Like Growth Factor, and Thyroid Hormone....................75
3. Parathyroid Hormone and Calcitonin .........................................................................75
4. Steroid Administration ................................................................................................75
G. Arthritis.............................................................................................................................75
H. Other Systemic Factors ....................................................................................................76
III.

In Vitro

Factors Affecting the Mechanical Properties of Bone.............................................76
A. Embalming or Fixation ....................................................................................................76
B. Boiling and Autoclaving ..................................................................................................77
C. Storage..............................................................................................................................77
D Drying and Freeze-Drying ...............................................................................................78
E. Sterilization.......................................................................................................................79
F. Holes in Bone...................................................................................................................79
G. Sample and Machine........................................................................................................79
IV. Summary ................................................................................................................................79
References ........................................................................................................................................80

I. INTRODUCTION

The usual purpose of mechanical testing of bone is to characterize the range of normal mechanical
properties and to dene abnormalities according to those normal values. In this section and the
next, the factors affecting the mechanical properties of bone are discussed, including (1) systemic
or

in vivo

factors, such as age, sex, species, function, composition, weightlessness and hypergravity,

0266/ ch04/frame Page 65 Thursday, October 21, 1999 2:12 PM

66

Mechanical Testing of Bone and the BoneImplant Interface

hormones, steroids, and arthritis, and (2)

in vitro

factors, such as embalming or xation, boiling
and autoclaving, freezing storage, drying and freeze-drying, sterilization methods, holes in bone,
sample preparation, and the testing machine.

Bone

in this chapter, in agreement with the rest of the book, refers to the structure and properties
of the bonetissue material and not those of the bone as a structure (as in a whole femur or tibia).
The distinction is important in this chapter because much of what is dened as

factors

or

determi-
nants

of the mechanical properties of bone are more effective when coupled with changes in the
shape, size, and thickness of the whole bone than when simply acting on the bone tissue level. In
the same vein,

normal

refers to the majority of bone tissue characteristics while

abnormal

or

pathophysiological

refers to changes caused by the so-called factor.
To begin, one needs a basic appreciation of the general nature of the mechanical characteristics
of bone and its intricate microstructure. These two elements may allow one, in most circumstances,
to anticipate the likely deviations from normality. For instance, in bone mechanics terms, bone is
a (1) viscoelastic, (2) elastic/brittle, (3) anisotropic material. The viscoelastic property means that
the applied strain rate during testing is in itself a determinant/factor of the recorded behavior

in
vivo

or

in vitro

; the elastic/brittle property describes when the elastic behavior ends and bone
develops internal damage when some threshold limits for stress or strain are reached; and anisotropic
properties depend on the direction of loading or the direction of preparation of a sample. On the
other hand, bone at the nanomicroscale is composed of collagen impregnated by mineral crystallites.
The behavior of this bonetissue matrix will be a function of the individual properties of the two
constituents, their degree of association and cohesion, the size and shape of the enclosed crystallites,
and their orientation. It is conceivable, therefore, that a disease state which will preferentially affect
any of these determinants will directly reect on the average material properties of the bone itself,
often in a somewhat predictable manner.
Provided that due attention is given to these and similar considerations, readers are invited to
exercise their own critical faculties to discern whether the conditions of some bone tests reported
in the past allow for a great degree of condence in the results. In general, anything other than
physiological temperature, ambient conditions, and loading rates similar to those

in vivo

should be
viewed with skepticism. However, it must be stressed that most research on deriving material
properties for bone tissue is carried out almost inevitably in quasi-static conditions which are easier
to control in the lab and allow easier standardization between labs.
One last word of caution: the factors referred to here can interplay and affect each other. For
instance, it is entirely feasible that a certain increase in the degree of porosity caused by osteonal
remodeling may cause a change in the degree of anisotropy of the tissue. Similarly, a certain
sterilization method or chemical treatment, which is meant to affect, for instance, only the collagen
component of bone, may also inadvertently cause demineralization or structural microdamage in
the material. The possibility that most factors can, therefore, act in a somewhat composite way
should not be ignored.

II. SYSTEMIC OR

IN VIVO

FACTORS AFFECTING
THE MECHANICAL PROPERTIES OF BONE
A. A

GE

Age affects the properties of bone in animals and humans. However, the so-called aging effect
is of primary importance to humans who suffer the effects of senescence as they grow and survive
into old age. Most animals by comparison do not live long after the end of their child-bearing
and child-rearing age. It is questionable, therefore, whether animal models (like rats and baboons)
can be used to study aging effects, although they may be perfectly acceptable for studies on
age alone.

1,2

0266/ ch04/frame Page 66 Thursday, October 21, 1999 2:12 PM

Factors Affecting Mechanical Properties of Bone

67

In general, with age there is an increase in the mineral content of the bone tissue, which achieves
its best strength and stiffness at maturity. Maturity is set at a nominal age of about 35 years for
humans (and varies from animal to animal). Thereafter, the elastic, ultimate, and fracture properties

3

of bone tissue deteriorate in both men and women.

4-7

Early results are summarized in Tables 4.1
and 4.2.
Lindahl and Lindgren

4

tested mixed male/female samples at room temperature and 65% humid-
ity; they observed that the ultimate tensile strength (UTS) fell from about 147 MPa at 30 years of
age to about 125 MPa at 90 years; the strain at failure fell from 2.3% at 30 years to about 1.6%
at 90 years. Wall et al.,

6

testing samples at body temperature which were sprayed with Hanks
solution, found that UTS fell from 102 MPa at 40 years to 73.5 MPa at 90 years. McCalden et al.

7

conducted tensile tests at room temperature on wet specimens at a strain rate of 0.03 s

1

; they found

TABLE 4.1
Data from Wet Specimens at Room Temperature (Femur/Tibia)

5

Age (years)
2030 3040 4050 5060 6070 7080 8090
Elastic Modulus (GPa)

Tension 17.0/18.9 17.6/27.0 17.7/28.8 16.6/23.1 17.1/19.9 16.3/19.9 15.6/29.2
Compression 18.1/ 18.6/35.3 18.7/30.6 18.2/24.5 15.9/25.1 18.0/26.7 15.4/25.9

Ultimate Strength (MPa)

Tension 140/161 136/154 139/170 131/164 129/147 129/145 120/156
Compression 209/ 209/213 200/204 192/192 179/183 190/183 180/197

Ultimate Strain (%)

Tension 3.4/4.0 3.2/3.9 3.0/2.9 2.8/3.1 2.5/2.7 2.5/2.7 2.4/2.3
Compression

Source:

Burstein, A. H. et al.,

J. Bone Joint Surg.,

58A, 82, 1976.

TABLE 4.2
Data from Relatively Wet Specimens at Room Temperature
(Femur)

8

Age (years)
1020 2030 3040 4050 5060 6070 7080
Ultimate Strength (MPa)

Tension 114 123 120 112 93 86 86
Compression 167 167 161 155 145
Bending 151 173 173 162 154 139 139
Torsion 57 57 52 52 49 49

Ultimate Strain (%)

Tension 1.5 1.4 1.4 1.3 1.3 1.3 1.3
Compression 1.9 1.8 1.8 1.8 1.8
Torsion 2.8 2.8 2.5 2.5 2.7 2.7

Source:

Martin, R.B. and Burr, D.B.,

Structure, Function, and Adaptation of Compact
Bone,

Raven Press, New York, 1989, 214.

0266/ ch04/frame Page 67 Thursday, October 21, 1999 2:12 PM

68

Mechanical Testing of Bone and the BoneImplant Interface

UTS fell from 120 MPa at 30 years to 70 MPa at 100 years of age, ultimate strain fell from 3.3%
at 30 years to 1% at 100 years, fracture energy reduced by 90% between 30 and 100 years of age
similar to the reduction in the plastic energy stored, and the elastic component was mostly unaffected
by age. Zioupos and Currey,

3

testing in three-point bending at 37

C and using Ringers solution,
found that the elastic modulus, 15.2 GPa at 35 years of age, fell by 2.3% of its value per decade
in later life; the bending strength fell similarly from 170 MPa by 3.7%; the transverse fracture
toughness

K

C

from 6.4 MPa m

1/2

by 4.1%; the

J

-integral from 1.2 kJ m

2

by 3% and the work to
transverse fracture from 3.4 kJ m

2

by 8.7%.
Cancellous bone aging effects are masked by the gross changes in architecture and density of
this tissue with age. There is a consensus today that strength and stiffness of bulk cancellous bone
change (roughly) as a quadratic power of the apparent density. Two recent review articles also
claim that, after consideration of all recent experiments, the material that makes up the trabeculae
in cancellous bone is in terms of stiffness and strength within 10 to 20% of that of neighboring
cortical bone. Hence, the age effects of cancellous bone are primarily reduced into describing its
architecture, connectivity, and level of porosity and, following that, in order to complete the picture
it may be assumed that the trabecular tissue material degrades similarly to its cortical neighbor.

B. S

EX

Lindahl and Lindgren

4

found that there was no difference between the sexes with respect to the
mechanical properties of cortical bone. In general, the differences between male and female bone
(Table 4.3) are caused by differences in mass, that is, the quantity of bone not the quality of it.
Males have on average bigger and heavier skeletons, not necessarily comprised of denser bone.
From birth until menopause, female bone material properties shadow those of the males. However,
after menopause female bones show accelerated resorption rates which cause increased porosity
levels and consequently produce a weaker bone matrix material (internal porosity) as well as a
thinning of the structure of the bones (breakdown of bone matrix).

C. S

PECIES

The properties of the bone material of the long bones of mammalian animals differ considerably
in absolute values. Table 4.4 shows indicative values for a number of mechanical properties from
four species.

D. C

OMPOSITION

P

OROSITY

, D

ENSITY

, M

INERALIZATION

1. Cortical Bone

Cortical bone is compact material with a small degree of internal porosity. The porosity of bone
(measured in a small bone specimen) is the fraction of the

actual bone material volume

(usually

TABLE 4.3
Some Indicative Values for Both Sexes in Humans and Animals

Subject Bone Mechanical Property Male Female

P

Value Ref.

Human Femur cortical Tensile strength (MPa) 138

2 131

3 >0.05 4
Elastic modulus (GPa) 14.9

0.0 14.7

0.0 >0.05
Vertebral cubes Compressive yield force (lb) 123

62 78

57 0.064 9
Consolidation force (lb) 165

93 100

51 0.041
Pigs Diaphyseal bones Bending strength (kg cm

2

) 340 428 <0.01 1
Bending modulus (kg cm

2

) 5453 6401 <0.01

0266/ ch04/frame Page 68 Thursday, October 21, 1999 2:12 PM

Factors Affecting Mechanical Properties of Bone

69

measured by immersion in water and use of the Archimedes principle) over

the total apparent
specimen volume

, which can be measured externally by using calipers (provided that the specimen
has a regular shape such as a cylinder or cube). Porosity affects the modulus of elasticity of compact
bone as a power that ranges between 3 and 5 depending on whether results come from one species
or across a range of species.

11,12

The material density of cortical bone is the wet weight divided by the actual bone material
volume and ranges between 1.7 and 2.1 g cm

3

. The most commonly used

wet bone density

is
simply the wet weight divided by the externally measured apparent specimen volume and this
includes a small degree of porosity, which for cortical bone varies between 0 and 5%. It is more
or less generally accepted today that apparent density inuences

11

the elastic stiffness as a power
of 2 and the strength as a power of 3.
The mineral content of bone tissue varies with the species, age, the mechanical function of the
bone, the health or pathophysiological condition of the individual, etc. All these factors are inter-
dependent and each one makes bone what it is. The level of mineral mass ranges between 40 and
70% of the total mass, but in some extreme cases it is as high as 80% (for the n whale,

Tympanic
bulla

) or even 98% (for the rostral bone of the whale,

Mesoplodon densirostris

). Figure 4.1 shows
in a ternary diagram some variations between species and how the three main components of bone
water, mineral, and collagen complement each other.

13

Table 4.5 shows the combined species mineral content effect on the mechanical properties of
various bone tissues.

13,14

The Youngs modulus of elasticity can range from 4 to 32 GPa, bending strength from 50 to
300 MPa, and the work of fracture from 200 to 7000 J m

2

. It is not possible for any one type of
bone to have high values for all three properties. Very high values of mineralization produce high
values of Youngs modulus but low values of work of fracture (which is a measure of fracture
toughness). Rather low values of mineralization are associated with high values of work of fracture
but low values of Youngs modulus and intermediate values of bending strength. The reason for
the high value for Youngs modulus associated with high mineralization is intuitively obvious, but
has not yet been rigorously modeled. The low fracture toughness associated with high mineralization
may be caused by the failure of various crack-stopping mechanisms that act at low mineral contents
when the cohesion between the various building blocks of bone (i.e., mineralized bone brils,
lamellae, osteons, etc.) is low. When the mineral content of the bone is high, an advancing crack
sees in front of it a rather uniform (it could be said, single-phase) material and advances
unhindered, causing rapid failure. The adoption of different degrees of mineralization by different
bones, leading to different sets of mechanical properties, is shown to be adaptive in most cases
studied, but some puzzles still remain.

15

TABLE 4.4
Mechanical Property Values

10

for Human, Equine, Bovine, and Porcine Bone Tissue

Mechanical Property Human Horses Cattle Pigs

Ultimate tensile strength (MPa) 124/174/125/152 121/113/102/120 113/132/101/135 88/108/88/100
Ultimate extension (%) 1.41/1.50/1.43/1.50 0.75/0.70/0.65/0.71 0.88/0.78/0.76/0.79 0.68/0.76/0.70/0.73
Elastic modulus in tension (GPa) 17.6/18.4/17.5/18.9 25.5/23.8/17.8/22.8 25.0/24.5/18.3/25.9 14.9/17.2/14.6/15.8
Ultimate compressive strength (MPa) 170/// 145/163/154/156 147/159/144/152 100/106/102/107
Ultimate contraction (%) 1.85/// 2.4/2.2/2.0/2.3 1.7/1.8/1.8/1.8 1.9/1.9/1.9/1.9
Elastic modulus in compression (GPa) /// 9.4/8.5/9.0/8.4 8.7/// 4.9/5.1/5.0/5.3
Ultimate shear strength (MPa) 54/// 99/89/90/94 91/95/86/93 65/71/59/64
Elastic modulus in torsion (GPa) 3.2/// 16.3/19.1/23.5/15.8 16.8/17.1/14.9/14.3 13.5/15.7/15.0/8.4

Note:

In each cell the four values are in order for femur/tibia/humerus/radius.

0266/ ch04/frame Page 69 Thursday, October 21, 1999 2:12 PM

70

Mechanical Testing of Bone and the BoneImplant Interface

Table 4.6 shows the combined effects of species, methods, and direction of testing (anisotropy
factor) on the measured toughness of various bones. The antler analogue included in the list helps
to illustrate that above and beyond the considerations described in the previous paragraph (which
are solely focused on the degree of mineralization), there are other factors such as intricate internal
architecture that help to produce some very tough bone tissue examples.

16

Some researchers

33-35

have attempted to examine the effects of mineralization

in vitro

by
controllably dissolving the mineral and demineralizing bone by means of acids and chelating agents
(Table 4.7). This approach has one inherent and inevitable disadvantage in that the dissolving
medium acts by diffusion from the free surfaces that are in contact with the solution and hence it
cannot guarantee a uniform result throughout the whole volume of the test sample at all times.
Nevertheless, the results are consistent with the previous

in vivo

patterns.
Walsh et al.

36-38

studied the effects of phosphate and uoride ions on the compressive properties
of cortical bone.



Bone tissue is an anisotropic nonhomogeneous composite material composed of
inorganic bone mineral bers (hydroxyapatite) embedded in an organic matrix (type I collagen and
noncollagenous proteins). One factor contributing to the overall mechanical behavior is the inter-
facial bonding interactions between mineral and organic matrix. This interfacial bonding is based,
in part, on electrostatic interactions between negatively charged organic domains and the positively
charged mineral surface. Phosphate and uoride ions have been demonstrated to alter
mineralorganic interactions, thereby inuencing the mechanical properties of bone in tension. It
is now accepted that uoride increases the resistance of bone tissue to chemical attack and degra-
dation, but it impairs the mechanical strength of the material as a result of the previously mentioned
altered interactions at the mineralcollagen interface.

2. Cancellous Bone

There are three main contributors to the material properties of cancellous bone: (1) the material
characteristics of the trabeculae; (2) the architecture of the trabeculae; and (3) the quantity of the

FIGURE 4.1

The drawing shows in a ternary diagram some variations between species and how the three
main components of bone water, mineral, and collagen complement each other.

0266/ ch04/frame Page 70 Thursday, October 21, 1999 2:12 PM

Factors Affecting Mechanical Properties of Bone

71

bone material.

39-41

These features are reected in the previously mentioned apparent density,
expressing mostly the quantity of bone present, and real material density, which may reect the
quality of the trabecular material, (Table 4.8). Theoretical expectations suggest that apparent density
should inuence stiffness as a power of 2 and strength as a power of 3. However, the most recent
studies show both these properties to change roughly as a quadratic power of the apparent density.

TABLE 4.5
Mechanical Properties Youngs Modulus (E) Ultimate Tensile Stress
(UTS), Ultimate Tensile Strain (

!

f

), and Mineral Content (mineral
weight/wet bone weight) of Various Bones

13,14

Species and Tissue Mineral Content E (GPa) UTS (MPa)

!

f

Red deer, immature antler 0.385 10 250 0.109
Red deer, mature antler 0.393 7.2 158 0.114
Reindeer, antler 0.411 8.1 95 0.051
Polar bear (3 months), femur 0.441 6.7 85 0.044
Narwhal, tusk cement 0.454 5.3 84 0.060
Narwhal, tusk dentine 0.466 10.3 120 0.037
Sarus crane, tarsometatarsus 0.467 23.1 218 0.018
Walrus, humerus 0.482 14.2 105 0.026
Fallow deer, radius 0.493 25.5 213 0.019
Human adult, femur 0.496 16.7 166 0.029
Bovine, tibia 0.499 19.7 146 0.018
Polar bear (9 months), femur 0.501 11.2 137 0.042
Leopard, femur 0.514 21.5 215 0.034
Brown bear, femur 0.517 16.9 152 0.032
Donkey, radius 0.522 15.3 114 0.020
Sarus crane, tibiotarsus 0.523 23.5 254 0.031
Flamingo, tibiotarsus 0.523 28.2 212 0.013
Roe deer, femur 0.525 18.4 150 0.011
Polar bear (3.5 years), femur 0.529 18.5 154 0.022
King penguin, radius 0.540 22.1 195 0.010
Horse, femur 0.541 24.5 152 0.008
Wallaby, tibia 0.551 25.4 184 0.010
Bovine femur 0.562 26.1 148 0.004
King penguin, ulna 0.577 22.9 193 0.011
Axis deer, femur 0.586 31.6 221 0.019
Fallow deer, tibia 0.589 26.8 131 0.006
Wallaby, femur 0.599 21.8 183 0.009
King penguin, humerus 0.621 22.8 175 0.008
Fin whale,

T. bulla

0.768 34.1, 31.3

a

, 33

a

0.002, 0.0011

a

Mesoplodon rostrum 0.960 41

a

60

a

0.0015

a

Note:

The mineral content was derived from calcium content measurements assuming that the tissues
have the same element stoichiometry. Note that the two most mineralized tissues at the bottom of
the list are extremely brittle compared to all the rest.

a

Values are in three-point bending

13

for the last two most brittle bones the ultimate tensile stress is
also the stress at yield. The authors believe that the reason for this is that in three-point bending the
region of maximum stress is very conned to the outmost bers and, therefore, when the tissue yields
( = microcracks) the failure is catastrophic because the generated macrocrack cannot be stopped
effectively.

0266/ ch04/frame Page 71 Thursday, October 21, 1999 2:12 PM

72

Mechanical Testing of Bone and the BoneImplant Interface

Material density of cancellous bone is measured using the weight of bone material (only
trabeculae) divided by the volume of only trabeculae, which is in the 1.6 to 1.9 g cm

3

range and a
little smaller than that of cortical bone.

E. F

UNCTION

1. Microstructure vs. Function

One extreme example demonstrating the different mechanical properties of bone tissues with greatly
differing functions was on the mechanical property differences

43

between deer antler, cow femur, and
n whale,

T. bulla

(Table 4.9). The femur of a cow has to be stiff and relatively strong and shows the

TABLE 4.6
Fracture Mechanics Properties and Values from the Literature (in chronological order)

Bone Orientation
K

c


(MPa m

1/2

) Energy Requirements Type of Test Speed Ref.

Bovine femur Long 3.21

G

IC

= 1.42.6 SENT Slow 17
Long 5.05 SENT Fast 17
Trans 5.6

G

IC

= 3.15.5 SENT Slow 17
Trans 7.7 SENT Fast 17
Trans 2.24.6

G

IC

= 0.781.12 SENT Slow 18
Long 3.62 CT Slow 19
Trans 5.7 3pb Slow 20
Long 2.45.2

G

IC

= 0.92.8 CT Slow 21
Bovine tibia Long 2.8

G

IC

= 0.63 CT Slow 22
Long 6.3

G

IC

= 2.88 CT Fast 22
Human tibia Long 2.45.3 CT Slow 23
Bovine tibia Trans 11.2

W
f
= 7.96 SENB Slow 24
Long 3.2 CT (v) Very slow 25
Trans 6.5 CT (v) Very slow 25
Human tibia Long 3.7 G
IC
= 0.36 CT Slow 26
Bovine tibia Long 7.2 CT Slow 26
Long 8.0 G
IC
= 0.94 CT Very slow 27
Antler Trans 5.4 SENT Slow 28
Human tibia Long 4.04.3 G
IC
= 0.590.83 CT Slow 29
Bovine tibia Long 6.26.7 G
IC
= 0.91.0 CT Slow 29
Human (75 years),
porosity (5%)
Femur Long G
IC
= 0.70, G
IIC
= 3.00 CT Slow 3
Tibia Long K
IC
= 2.12,
K
IIC
= 8.32
G
IC
= 0.40, G
IIC
= 5.50 31
Human femur (35 years) Trans 6.5 W
f
= 3.5, J
int
= 1.2 SENB Slow 3
Baboons femurs Long 2.3 CT Slow 32
Bovine femur Trans 5.0 W
f
= 3.00 SENB Slow 13
Mesoplodon rostrum Trans 1.3 W
f
= 0.091 SENB Slow 13
Note: Orientation is the direction with respect to the bone axis, either longitudinal or transverse. Type of test indicates the
test conguration and specimen geometry. CT: compact tension; (v): the specimen has been grooved to force the crack travel
in a particular direction; SENT: single-edge notch tension; SENB: single-edge notch bending; W
f
: work of fracture; 3-pb:
three-point bending with a single notch. The energy requirements are either for W
f
, or for the critical energy release rate G
C
,
or for the J-integral (all in comparable values for kJ m
2
). Speed: is the speed of crack growth, in most cases this was for
slow, stable crack growth, only two studies reported on the unstable rapid propagation of cracks.
0266/ ch04/frame Page 72 Thursday, October 21, 1999 2:12 PM
Factors Affecting Mechanical Properties of Bone 73
characteristic behavior of normal compact bone tissue. The deer antlers are relatively softer, but
extremely tough because in life they are expected to survive high impact loading. The ear bone of T.
bulla is by comparison very brittle. The mechanical requirements placed upon this bone (like the
transfer of loads and moments) are minimal; the crucial factor is the acoustic impedance and the loss
of vibration quality through viscoelastic phenomena. Natures answer to this is a very high mineral
content which makes the earbone of T. bulla very stiff and very brittle. Another example is the
differences between weight-bearing and non-weight-bearing bones (Table 4.10).
44
TABLE 4.7
Effects of Mineral Content (Examined by Demineralization)
on Mechanical Properties of Bone
Species Bone Treatment Decalcication (%) Mechanical Property Change (%) Ref.
Cats Whole bone EDTA* 20 Bending strength "35 35
40 "51
60 "68
80 "84
Cows Tibial samples EDTA 3 Bending strength "30 33
5 "39
9 "52
24 "76
* Ethylenediaminetetra-acetic acid, disodium salt
TABLE 4.8
Correlation Analysis
42
between the Mechanical Parameters
by Compression Test and Canine Cancellous Bone Densities
Mechanical Parameter Test Method Apparent Density Ash Density
Elastic modulus (GPa) Compression 0.778 0.737
Stiffness (N/mm) Compression 0.868 0.853
Indentation 0.944 0.923
Ultimate load (N) Compression 0.966 0.966
Indentation 0.940 0.954
Ultimate strength (MPa) Compression 0.934 0.912
Indentation 0.939 0.954
Values are correlation coefcients; n = 10, n# = n 2, one-way analysis.
P values are at least <0.05.
TABLE 4.9
Mechanical Property Differences
43
of Bones Due
to Different Functions and Species (n = 3 10)
Parameter Deer Antler Cow Femur Whale T. bulla
Bending strength (MPa) 179.4 6.3 246.7 11.6 33.0 6.6
Bending modulus (GPa) 7.4 0.3 13.5 1.0 31.3 1.0
0266/ ch04/frame Page 73 Thursday, October 21, 1999 2:12 PM
74 Mechanical Testing of Bone and the BoneImplant Interface
2. Activity Levels
Activity levels have an effect upon the bone mass and the bone material itself.
45-52
The mass of a
normal bone is determined essentially by the balance between the two remodeling processes,
resorption and deposition of the periosteal and endosteal surfaces.
53
The daily loading/straining
pattern experienced by an individual bone strongly inuences the two processes; increased loading
leading to an increase in bone mass and decreased loading leading to a reduction in bone mass.
This has become widely known as Wolffs law.
54
The usual example given of remodeling in response
to exercise is that of the greater bone mass in the racket arm of tennis players compared with their
contralateral arm. This response is also seen in rat femurs in response to an activity regime
(Table 4.11) and in response to hypergravity.
55
The magnitude of the effect of activity levels is not clearly understood although it seems that
any effects are greater in adolescents than in mature individuals. In adults any increase in bone
mass brought about by increased activity lasts only as long as that activity level is retained.
45-50
3. Weightlessness
Weightlessness and reduced gravity are known to have very marked effects upon bone both in
humans and animals.
59-61
Both situations drastically reduce the daily loading patterns experienced
by the bones and thus affect the remodeling process. Space ight, with zero or near zero gravity,
inevitably leads to marked net bone loss, something which even modern exercise regimes cannot
completely prevent.
59
Examples of the converse situation, extreme increased loading, may be seen in people who
have suddenly changed their activity patterns, e.g., recent widows/widowers, athletes, ballet dancers,
or new army recruits.
62-64
In these cases, the sudden large increase in loading (carrying shopping
bags or performing housework, heavy training prior to a championship or a premiere, forced
marches carrying heavy packs) may stimulate increased deposition, but this is usually outpaced by
TABLE 4.10
Mechanical Properties of Rabbit Bone
44
Due to Different Functions
(all from the right side, n = 17 for each bone or location)
Test Bone
Ultimate Load
(N)
Stiffness
(N/mm)
Ultimate Strength
(MPa)
Elastic Modulus
(GPa)
Flexure Humerus 284 8 367 17 165 5 13.6 0.7
Femur 353 13 413 19 137 6 15.1 0.7
Tibia 320 12 301 15 198 8 21.5 1.1
Indentation Humeral head 136 5 683 66 32 1
Femoral head 362 22 1531 133 86 5
Medial tibial plateau 244 20 1257 139 55

5
TABLE 4.11
Bending Properties of Rat Femur Due to Different Levels of Activity
Parameter Normal Immobilization Exercise Ref.
Strength (ultimate load) 100% (79.1 N) +23.9% (97.2 N) 56, 57
Elasticity (stiffness) 100% (18.4 N/mm) 8.7% (16.8 N/mm) +7.1% (19.7 N/mm) 58
0266/ ch04/frame Page 74 Thursday, October 21, 1999 2:12 PM
Factors Affecting Mechanical Properties of Bone 75
the increased rate of fatigue damage accumulation. The normal bone repair processes are not able
to cope with the effects of this fatigue loading either, and the result is what is clinically termed a
stress fracture, which is a fracture resulting from a buildup of fatigue damage.
F. HORMONES
1. Sex Hormones
The cessation of estrogen production in the female, either through natural or surgical means,
signicantly affects bone metabolism, reducing bone mass and thus affecting bone quality.
65-68
Androgen production is also lowered following menopause, which also reduces bone mass.
69
The gradual reduction of endogenous production of male sex hormones with aging also leads to
reduced bone mass, and is increasingly being recognized as a clinical problem.
70-74
Hormone
replacement therapy (oral administration of estrogen) is the usual treatment option for bone loss
in women
65-67
but can result in increased risk of artherosclerosis.
75
During pregnancy and lactation
the normal calcium homeostasis is altered to allow for the calcium demands made by the fetus
while preventing dangerous bone loss from the mother.
76-79
This process is governed mainly by
parathyroid hormone.
76,79
2. Growth Hormone, Insulin-Like Growth Factor, and Thyroid Hormone
Growth hormone deciency in adult humans has been shown to lead to reduced bone mass and
mineral density.
80-82
Growth hormone replacement is one of the treatment options. The action of
growth hormone is mediated by insulin-like growth factors (IGF).
83-86
Thyroxine treatment, e.g.,
for hypothyroidism, is thought to reduce bone mass and density.
87-90
3. Parathyroid Hormone and Calcitonin
Parathyroid hormone is the principal regulator of the remodeling process in the skeleton, mostly
promoting resorption.
67
Oral administration is one of the treatment options for bone loss.
67,91
Calcitonin has been shown to be successful in increasing bone density.
92,93
Its use has been suggested
for reversal of the side effects of long-term treatment with glucocorticoids,
93,94
also in combination
with dietary calcium supplements.
66
4. Steroid Administration
Some anabolic steroids, e.g., nandrolone, are recommended for cases where bone mass is lost and
does not respond to other bone-promoting drugs, for instance, in hypogonadal men or those
receiving treatment with corticosteroids.
72,75
However, they are thought to be ineffective for eugo-
nadal men.
71
Long-term use of glucocorticoids is common for treatment of noninfectious inam-
matory disease but leads to loss of bone mass.
93-96
Glucocorticoids are known to suppress osteoblast
activity and thus increase resorption.
72,92-94,96
Hence, glucocorticoids suppress deposition of new
bone but their action can be somewhat reversed by calcitonin
92,95
and by vitamin D.
72
G. ARTHRITIS
Rheumatoid or inammatory arthritis (RA) is either caused by an infection or by an autoimmune
reaction and is seen as a disease of the cartilage, whereas osteoarthritis (OA) is a disease of the
bone tissue and occurs due to mechanical factors, e.g., wear or impact injury, to the bone or
cartilage.
97-99
Patients with RA are at a much higher risk of femoral neck fractures than patients
with OA, and it seems RA acts to reduce bone mass,
97-100
but does not alter the mineral content of
the bone tissue. OA, on the other hand, tends to increase the bone mass and apparent density by
thickening the trabeculae while also reducing the mineral content of the bone tissue. OA is rarely
0266/ ch04/frame Page 75 Thursday, October 21, 1999 2:12 PM
76 Mechanical Testing of Bone and the BoneImplant Interface
seen in patients with osteoporosis.
97-99
In short, RA leads to thinning of the bone and reduction in
bone strength and stiffness, whereas OA leads to thickening of bone and an increase in bone strength
and stiffness, yet both are pathological.
H. OTHER SYSTEMIC FACTORS
Fluoride has a marked toxic effect upon bone tissue, impairing material properties, especially at a
high dosage;
101,102
yet it is also capable, paradoxically, of stimulating an increase in bone mass and
improving bone material qualities, and has been suggested as a possible treatment for glucocorti-
coid-induced bone loss.
95,102-104
This paradox is not well understood.
Biphosphophonates are suggested as possible treatment options for bone loss,
71,72,92,103,104
acting
to suppress osteoclast activity.
100
Insufcient Ca intake can result from either low levels in the diet
or from absorption problems. In its extreme form it results in osteomalacia where the collagen
matrix remains unmineralized. Vitamin D, calcifediol, and calcitriol help to increase absorption of
Ca.
95,101
The absence of both selenium and vitamin E in the diet can lead to KashinBeck, a type of
degenerative osteoarthritis. Experiments with rabbits have shown that selenium- and vitamin-E-
decient diets reduce both bone strength and elastic modulus.
105
It is well established that chronic alcohol consumption leads to reduction in bone mass and
increase in fracture risk,
92,104,106-112
although its exact pathology is unclear.
Zinc deciency has been linked to bone disorders and hypogonadism, although these two are
in themselves linked. Zinc is important for normal growth hormone and IGF production.
113
III. IN VITRO FACTORS AFFECTING
THE MECHANICAL PROPERTIES OF BONE
A. EMBALMING OR FIXATION (TABLE 4.12)
Bone from formalin-embalmed bodies is not appropriate for testing of mechanical properties
because of the cross-linking of the collagen protein which alters the mechanical properties.
114,115
It
is known that the cross-links between the collagen molecules have a signicant inuence upon the
mechanical properties of bone tissue.
116
It has been reported
115
that in quasistatic loading, mechanical properties of bovine bone were
almost unaffected by a certain formalin xation protocol, but at the same time a signicant decrease
in impact strength was found. These results indicated that there may be some interaction between
xation- and strain-rate-dependent effects, and, therefore, some caution is needed when using
common biomechanical measurement methods on xed bone material.
TABLE 4.12
Effects of Embalming or Fixation on Mechanical Properties of Bone (MPa)
Species Bone Treatment Mechanical Property Change (%) Ref.
Bovine cortical 10% formalin mixture
b
Tensile modulus No change 115
Impact strength "
a
Compressive strength "12% 117
Compressive modulus "6%
a
Statistically signicant.
b
A mixture of ethanol, phenol, glycerine, formalin, and water.
0266/ ch04/frame Page 76 Thursday, October 21, 1999 2:12 PM
Factors Affecting Mechanical Properties of Bone 77
B. BOILING AND AUTOCLAVING (TABLE 4.13)
Borchers et al.
118
studied the effects of boiling and autoclaving on the compressive modulus and
strength of bovine trabecular bone. The results showed 26 and 58% reductions in modulus and
strength, respectively. Autoclaving on its own also signicantly reduced the compressive modulus
(by 59%). Another experiment showed that heating in saline for 2 h at temperatures up to 95C
had no effect upon the strength in three-point bending, but did have a signicant effect upon the
elastic modulus (a 12% drop after 2 h at 95C).
119
Viceconti et al.
120
also demonstrated a large
reduction in mechanical performance following exposure to high temperatures; only 30% remained
of the compressive strength following autoclaving at 132C for 1 h. It is clear that placing bone in
high temperatures leads to changes in its mechanical properties.
C. STORAGE (TABLE 4.14)
Owing to complexity of an experiment or unforeseen circumstances, sometimes a specimen may
be thawed and frozen a number of times. The question arises whether multiple freezing and thawing
is harmful to the mechanical properties. This question has been partially answered by Linde and
Srensen
123
who found that freezing and thawing up to ve times did not alter the compressive
properties of cancellous bone. In other recent experiments the proximal portion of the tibia of adult
cows was sectioned to produce bone slices. The slices were then subjected to four freezingthawing
cycles: freezing with and without saline solution, then thawing in saline solution or exposed to the
air. The mechanical properties of the bone before and after the treatments (ve cycles of freezing
and thawing) were measured using an indentation test. No signicant effect on the ultimate load
and stiffness of the bone was found; only a trend of difference was noticed for the specimens frozen
without saline soaking and thawed in air.
124
This work supports the widely used practice of freezing
and thawing bone specimens in saline solution.
The common method for storing bone specimens is freezing at 20C. The effects of storage
on mechanical properties of bone at 20C for short periods of time are minor. The maximum
TABLE 4.13
Effects of Boiling and Autoclaving on Mechanical Properties of Bone
Bone Species Treatment Temp. (C) Time (min) Mechanical Property Change (%) Ref.
Diaphyseal Rabbit Autoclave 110 255 Torsional strength "35 121
Torsional stiffness "27
121 20 Torsional strength "23
Torsional stiffness "20
131 2 Torsional strength " 9
Torsional stiffness "10
Cancellous Porcine Boiling 60 60 Compressive strength 0 122
80 60 Compressive strength " ?
100 60 Compressive strength "40
Cortical Bovine Hot saline 95 120 Flexural strength 0 119
Flexural modulus "12
Boiling 100 30 Compressive strength "32 118
Compressive modulus "25
Autoclave 127 10 Compressive strength "48
Compressive modulus "47
132 20 Compressive strength "70 120
0266/ ch04/frame Page 77 Thursday, October 21, 1999 2:12 PM
78 Mechanical Testing of Bone and the BoneImplant Interface
effect reported is a 4.6% reduction in torsional strength of canine long bones.
125
However, after
thawing, enzymes such as collagenase and protease may become active and degrade the tissue.
Also, enzymatic degradation is not completely arrested at 20C.
126
With concerns about the effects
of enzyme degradation
127
and evaporation,
128
a question arises as to whether there are signicant
effects of long-term storage at 20C. Panjabi et al.
129
found no signicant effects of freezing for
7 to 8 months on the mechanical properties of human vertebral bone. Roe et al.
130
found that bones
frozen at 20C for 8 months did not become signicantly weaker. Because time periods longer
than 8 months have not been reported for frozen storage at 20C, storage at this temperature for
more than 8 months is not recommended. Alternatively, 70C, 80C, or even lower temperatures
or liquid nitrogen are suggested for long-term bone storage, since these temperatures may minimize
evaporation
128
and markedly reduce enzyme activity.
127
However, the remote possibility may then
arise that at these lower temperatures other kinds of alterations may be inicted upon bone tissue,
either by microcracking or by damaging the collagen moiety of bone.
D. DRYING AND FREEZE-DRYING
It is not uncommon that specimens are sometimes, probably unintentionally, allowed to dry out in
air prior to or during a mechanical test. Some specimens may also only be available in the dried
state. Early work
133
has shown that the mechanical properties did not change following drying in
air and rewetting. Nothing exists in the literature concerning the changes in properties if specimens
remain in the dry state for a period of time, perhaps because it is obvious that they are very different
from those in the wet state.
Freeze-drying is a very common preservation technique. During this process, water changes
from the solid state into the gaseous state without entering the liquid state (sublimation). It is
usually felt that this process is better at preserving ne structure. Indeed, there is usually no obvious
change with many tissues following freeze-drying and rehydration. However, most studies have
shown that there is a reduction in mechanical properties of bone following this process.
119,132,134-136
It is known that freeze-drying has little or no effect in purely collagenous tissues such as tendon,
137-139
TABLE 4.14
The Effects of Freezing on the Mechanical Properties of Bones
Subject/Bone Temp (C) Time Saline Saturation Testing Mode Change in Strength? Ref.
Human
Femur 20 34 wk Yes Bending/tension No
a
131
Long bones ? ? ? Bending No 128
Tibia 20 Five times in 15 days ? Compression No 123
Cattle
Trabecular 20, 70 8 days No Compression No 118
20 Eight times in 8 days No Compression No 118
20 Five times in 5 days No Indentation "(trend) 124
Dog
Femur/tibia 40 2 days Yes Torsion " 4.6% 125
Femur 20 1 wk No/sealed Compression $ 130
20 16, 32 wks No/sealed Compression No 130
20 1, 16, 32 wks No/sealed Screw pullout No 130
Rat
Femur 20 2 wks ? Torsion No 132
20 2 wks ? Compression No 132
a
No = no statistical difference.
0266/ ch04/frame Page 78 Thursday, October 21, 1999 2:12 PM
Factors Affecting Mechanical Properties of Bone 79
so it would appear that the observed effects in bone result from its composite nature, whereby the
mineralized collagen matrix contracts differentially and distorts and hence experiences damage
(microcracks), which affects its properties.
E. STERILIZATION
Common techniques for sterilization of bone are %-irradiation and exposure to ethylene oxide
gas. It is normal to dry bone prior to exposure to ethylene oxide as toxic residues are produced
with water present. Common dosages used for %-irradiation by Co
60
are between 1 and 3 MRad
(10 to 30 kGy) while the bone is frozen in water or saline. It is not clear from the literature
whether bone material suffers signicant reductions in its mechanical competence following
exposure to irradiation in such circumstances.
140,141
It is probable that the conditions during
irradiation determine the changes in properties. It is known that irradiation has two contradictory
effects in purely collagenous materials like tendons. On the one hand, it causes direct molecular
chain scission and, on the other hand, it promotes crosslinks between molecular chains.
142
The
mobility of the molecules inuences the cross-linking process, while scission is controlled by
the dose. The greater the amount of cross-linking relative to chain scission, the smaller the
reduction in mechanical competence. The presence of water, especially as a liquid, during
irradiation will most likely be benecial for mechanical competence.
F. HOLES IN BONE
Holes are introduced in bone during restorative applications, when inserting screws or pins for
xation, etc. These have undoubtedly a detrimental effect
143-145
on the strength of a whole bone as
shown in Table 4.15. Bone has two ways of coping with this adversity. In the long term, if the
insertion is removed, the bone will ll in the hole and (by repairing) alleviate some of the danger.
In the short term, bone benets from its ability (which is also a material property) to yield around
stress concentrating defects (as by microcracking) and thus blunting their deleterious effect.
146
G. SAMPLE AND MACHINE
Sample preparation, sample size and shape, test conditions, samplemachine interface, and machine
compliance are common determining factors for the scattering results of mechanical testing of
bone.
147,148
Detailed discussion follows in Chapter 7.
IV. SUMMARY
There are many factors that can potentially inuence the mechanical properties of bone. However,
unlike anthropogenic composites, bones are able to adapt and change in life (and in disease), and
TABLE 4.15
Effect of Transcortical Drill Hole on the Mechanical Strength of Diaphyseal Bone
Subject Bone
Bone Diam.
(mm)
Hole Diam./
Outer Diam.
of Bone Hole Type
Mechanical
Test
Strength Reduction
(%) Ref.
Sheep Femur ? 20 Unicortical Torsion 34 143
19.6 50 Unicortical Torsion 60 144
Dog Femur 13.76 44 Unicortical 4-pt bend 38 145
13.76 44 Bicortical 4-pt bend 51 145
0266/ ch04/frame Page 79 Thursday, October 21, 1999 2:12 PM
80 Mechanical Testing of Bone and the BoneImplant Interface
that adds a few extra complications (unknown factors). Composites are what they are as a function
of (1) the mixture/combination of a few primary elements; (2) the properties of these elements;
and (3) their interaction. In bones mere considerations of mineral content, hydroxyapatite stoichi-
ometry and the collagen condition do not apply simply and in a straightforward manner across
species, ages, disease, treatments, etc. To make matters worse, these factors interplay and interde-
pend on each other. It becomes obvious that examining the mechanical properties of bone is an art
in its own right. It requires due care and attention and, it could be said, some reection on what
is actually happening, what the knowledge is that is needed to be acquired and how deeply the
process/situation at hand is comprehended. None of these should deter ambitious workers from
engaging in bone biomechanics research and making their own mark in this eld.
REFERENCES
1. Crenshaw, T.D., Peo, E.R., Jr., Lewis, A.J., et al., Inuence of age, sex and calcium and phosphorus
levels on the mechanical properties of various bones in swine, J. Anim. Sci., 52, 1319, 1981.
2. Lawrence, L.A., Ott, E.A., Miller, G.J., et al., The mechanical properties of equine third metacarpals
as affected by age, J. Anim. Sci., 72, 2617, 1994.
3. Zioupos, P. and Currey, J.D., Changes in the stiffness, strength and toughness of human cortical bone
with age, Bone, 22, 56, 1998.
4. Lindahl, O. and Lindgren, A.G., Cortical bone in man. II. Variation in tensile strength with age and
sex, Acta Orthop. Scand., 38, 141, 1967.
5. Burstein, A.H., Reilly, D.T., and Martens, M., Aging of bone tissue: mechanical properties, J. Bone
Joint Surg., 58A, 82, 1976.
6. Wall, J.C., Chatterji, S.K., and Jeffery, J.W., Age related changes in the density and tensile strength
of human femoral cortical bone, Calcif. Tissue Int., 27, 105, 1979.
7. McCalden, R.W., McGeouch, J.A., Barker, M.B., and Court-Brown, C.M., Age-related changes in the
tensile properties of cortical bone, J. Bone Joint Surg., 75A, 1193, 1993.
8. Martin, R.B. and Burr, D.B., Aging effects, in Structure Function and Adaptation of Compact Bone,
Raven Press, New York, 1989, 214.
9. Oyster, N., Sex differences in cancellous and cortical bone strength, bone mineral content and bone
density, Age Ageing, 21, 353, 1992.
10. Fung, Y.C., Biomechanics Mechanical Properties of Living Tissues, Springer-Verlag, New York, 1981.
11. Martin, R.B., Determinants of the mechanical properties of bones, J. Biomech., 24, 79, 1991.
12. Currey, J.D., Physical characteristics affecting the tensile failure properties of compact bone,
J. Biomech., 23, 837, 1990.
13. Zioupos, P., personal communication, 1999.
14. Currey, J.D., Mechanical properties of vertebrate hard tissues, Proc. Inst. Mech. Eng., 212, 399, 1998.
15. Currey, J.D., Effects of differences in mineralization on the mechanical properties of bone, Philos.
Trans. R. Soc. London B, 304, 509, 1984.
16. Zioupos, P., Currey, J.D., and Sedman, A.J., An examination of the micromechanics of failure of bone
and antler by acoustic emission tests and laser scanning confocal microscopy, Med. Eng. Phys., 16,
203, 1994.
17. Melvin, J.W. and Evans, F.G., Crack propagation in bone, in Biomech. Symp. ASME, American Society
of Mechanical Engineers, New York, 1973, 87.
18. Boneld, W. and Datta, P.K., Fracture toughness of compact bone, J. Biomech., 9, 131, 1976.
19. Wright, T.M. and Hayes, W.C., Fracture mechanics parameters for compact bone effects of density
and specimen thickness, J. Biomech., 10, 419, 1977.
20. Robertson, D.M., Robertson, D., and Barret, C.R., Fracture toughness, critical crack length and plastic
zone size in bone, J. Biomech., 11, 359, 1978.
21. Boneld, W., Grynpas, M.D., and Young, R.J., Crack velocity and the fracture of bone, J. Biomech.,
11, 473, 1978.
22. Behiri, J.C. and Boneld, W., Fracture mechanics of bone the effects of density, specimen thickness
and crack velocity on longitudinal fracture, J. Biomech., 17, 25, 1984.
0266/ ch04/frame Page 80 Thursday, October 21, 1999 2:12 PM
Factors Affecting Mechanical Properties of Bone 81
23. Boneld, W., Behiri, J.C., and Charalambides, B., Orientation and age-related dependence of the
fracture toughness of cortical bone, in Biomechanics: Current Interdisciplinary Research, Perrin, S.M.
and Schneider, E., Eds., Martinus Nijhoff, Dordrecht, the Netherlands, 1984, 185.
24. Moyle, D.D. and Gavens, A.J., Fracture properties of bovine tibial bone, J. Biomech., 19, 919, 1986.
25. Behiri, J.C. and Boneld, W., Crack velocity dependence of longitudinal fracture in bone, J. Mater.
Sci., 15, 1841, 1989.
26. Norman, T.L., Vashishth, D., and Burr, D.B., Mode I fracture toughness of human bone, Adv. Bioeng.,
20, 361, 1991.
27. Norman, T.L., Vashishth, D., and Burr, D.B., Effect of groove on bone fracture toughness, J. Biomech.,
25, 1849, 1992.
28. Sedman, A.J., Mechanical Failure of Bone and Antler: The Accumulation of Damage, D.Phil. thesis,
University of York, U.K., 1993.
29. Norman, T.L., Vashishth, D., and Burr, D.B., Fracture toughness of human bone under tension,
J. Biomech., 28, 309, 1995.
30. Norman, T.L., Nivargikar, S.V., and Burr, D.B., Resistance to crack growth in human cortical bone
is greater in shear than in tension, J. Biomech., 29, 1023, 1996.
31. Yeni, Y.N., Brown, C.U., Wang, Z., and Norman, T.L., The inuence of bone morphology on fracture
toughness of human femur and tibia, Bone, 21, 453, 1997.
32. Wang, X.D., Masilamani, N.S., Mabrey, J.D., et al., Changes in the fracture toughness of bone may
not be reected in its mineral density, porosity, and tensile reports, Bone, 23, 67, 1998.
33. Broz, J.J., Simske, S.J., and Greenberg, A.R., Material and compositional properties of selectively
demineralized cortical bone, J. Biomech., 28, 1357, 1995.
34. Guo, M.Z., Xia, Z.S., and Lin, L.B., The mechanical and biological properties of demineralised cortical
bone allografts in animals, J. Bone Joint Surg., 73B, 791, 1991.
35. Shah, K.M., Goh, J.C., Karunanithy, R., et al., Effect of decalcication on bone mineral content and
bending strength of feline femur, Calcif. Tissue Int., 56, 78, 1995.
36. Walsh, W.R. and Guzelsu, N., The role of ions and mineralorganic interfacial bonding on the
compressive properties of cortical bone, Biomed. Mater. Eng., 3, 75, 1993.
37. Walsh, W.R. and Guzelsu, N., Compressive properties of cortical bone: mineralorganic interfacial
bonding, Biomaterial, 15, 137, 1994.
38. Walsh, W.R., Labrador, D.P., Kim, H.D., and Guzelsu, N., The effect of in vitro uoride ion treatment
on the ultrasonic properties of cortical bone, Ann. Biomed. Eng., 22, 404, 1994.
39. Carter, D.R. and Hayes, W.C., The compressive behavior of bone as a two-phase porous structure,
J. Bone Joint Surg., 59A, 954, 1977.
40. Ashman, R.B. and Rho, J.Y., Elastic modulus of trabecular bone material, J. Biomech., 21, 177, 1988.
41. Rice, J.C., Cowin, S.C., and Bowman, J.A., On the dependence of the elasticity and strength of
cancellous bone on apparent density, J. Biomech., 21, 155, 1988.
42. Kang, Q., An, Y.H., and Friedman, R.J., The mechanical properties and bone densities of canine
cancellous bones, J. Mater. Sci. Mater. Med., 9, 463, 1998.
43. Currey, J.D., Mechanical properties of bone tissues with greatly differing functions, J. Biomech., 12,
313, 1979.
44. An, Y.H., Kang, Q., and Friedman, R.J., Mechanical symmetry of rabbit bones studied by bending
and indentation testing, Am. J. Vet. Res., 57, 1786, 1996.
45. Chen, M.M., Yeh, J.K., Aloia, J.F., et al., Effect of treadmill exercise in tibial cortical bone in aged
female rats a histomorphometry and dual energy X-ray absorptiometry study, Bone, 15, 313, 1994.
46. Lin, B.Y., Jee, W.S., Chen, M.M., et al., Mechanical loading modies ovariectomy-induced cancellous
bone loss, Bone Miner., 25, 199, 1994.
47. Bagi, C.M. and Miller, S.C., Comparison of osteoporotic changes in cancellous bone induced by
ovariectomy and/or immobilization in adult rats, Anat. Rec., 239, 243, 1994.
48. Palle, S., Vico, L., Bourrin, S., and Alexandre, C., Bone tissue response to 4 month anti-orthostatic
bedrest a bone histomorphometric study, Calcif. Tissue Int., 51, 189, 1992.
49. Li, K.C., Zernicke, R.F., Barnard, R.J., and Li, A.F., Differential response of rat limb bones to strenuous
exercise, J. Appl. Phys., 70, 554, 1991.
50. Forwood, M.R. and Burr, D.B., Physical activity and bone mass exercises in futility, Bone Miner.,
21, 89, 1993.
0266/ ch04/frame Page 81 Thursday, October 21, 1999 2:12 PM
82 Mechanical Testing of Bone and the BoneImplant Interface
51. Gordon, K.R., Levy, C., Perl, M., and Weeks, O.I., Adaptive modelling in a mammalian skeletal model
system, Growth Dev. Aging, 57, 101, 1993.
52. Reilly, G.C., Currey, J.D., and Goodship, A.E., Exercise of young thoroughbred horses increases
impact strength of the third metacarpal bone, J. Orthop. Res., 15, 862, 1997.
53. Currey, J.D., The Mechanical Adaptations of Bone, Princeton University Press, Princeton, NJ, 1984.
54. Cowin, S.C., Bone Mechanics, 1986, CRC Press, Boca Raton, FL, 1986.
55. Kimura, T., Amtmann, E., Doden, E., and Oyama, J., Compressive strength of the rat femur as
inuenced by hypergravity, J. Biomech., 12, 361, 1979.
56. Barengolts, E.I., Curry, D.J., Bapna, M.S., and Kukreja, S.C., Effects of endurance exercise on bone
mass and mechanical properties in intact and ovariectomized rats, J. Bone Miner. Res., 8, 937, 1993.
57. Umemura, Y., Ishiko, T., Yamauchi, T., et al., Five jumps per day increase bone mass and breaking
force in rats, J. Bone Miner. Res., 12, 1480, 1997.
58. Nordsletten, L., Kaastad, T.S., Skjeldal, S., et al., Training increases the in vivo strength of the lower
leg: an experimental study in the rat, J. Bone Miner. Res., 8, 1089, 1993.
59. Johnson, R.B., The bearable lightness of being: bone, muscles and spaceight, Anat. Rec., 253, 24,
1998.
60. Bikle, D.D., Halloran, B.P., and Morey-Holton, E., Space ight and the skeleton: lessons for the
earthbound, Endocrinologist, 7, 10, 1997.
61. Sulzman, F.M., Life sciences space missions overview, J. Appl. Physiol., 81, 3, 1996.
62. Norman, E., Walter, M.D., and Wolf, M.D., Stress fractures in young athletes, Am. J. Sports Med., 5,
165, 1977.
63. Orava, S. and Hulkko, A., Stress fractures of the mid-tibial shaft, Acta Orthop. Scand., 55, 35, 1984.
64. Schneider, H.J., King, A.Y., Bronson, J.L., and Miller, E.H., Stress injuries and developmental change
of lower extremities in ballet dancers, Radiology, 113, 627, 1974.
65. Notelovitz, M., Estrogen therapy and osteoporosis: principles & practice, Am. J. Med. Sci., 313, 2,
1997.
66. Nieves, J.W., Komar, L., Cosman, F., and Lindsay, R., Calcium potentiates the effect of estrogen and
calcitonin on bone mass: review and analysis, Am. J. Clin. Nutr., 67, 18, 1998.
67. Masiukiewicz, U.S. and Insogna, K.L., The role of parathyroid hormone in the pathogenesis, preven-
tion and treatment of postmenopausal osteoporosis, Aging (Milan), 10, 232, 1998.
68. Oursler, M.J., Estrogen regulation of gene expression in osteoblasts and osteoclasts, Crit. Rev. Eukary-
otic Gene Expression, 8, 125, 1998.
69. Abraham, D. and Carpenter, P.C., Issues concerning androgen replacement therapy in postmenopausal
women, Mayo Clin. Proc., 72, 1051, 1997.
70. Bennell, K.L., Brukner, P.D., and Malcolm, S.A., Effect of altered reproductive function and lowered
testosterone levels on bone density in male endurance athletes, Br. J. Sports Med., 30, 205, 1996.
71. Burns-Cox, N. and Gingell, C., The andropause: fact or ction? Postgrad. Med. J., 73, 553, 1997.
72. Reid, I.R., Glucocorticoid osteoporosis mechanisms and management, Eur. J. Endocrinol., 137,
209, 1997.
73. Maas, D., Jochen, A., and Lalande, B., Age-related changes in male gonadal function implications
for therapy, Drugs Aging, 11, 45, 1997.
74. Ebeling, P.R., Osteoporosis in men new insights into aetiology, pathogenesis, prevention and
management, Drugs Aging, 13, 421, 1998.
75. St. Clair, R.W., Estrogens and atherosclerosis: phytoestrogens and selective estrogen receptor modu-
lators, Curr. Opinion Lipidol., 9, 457, 1998.
76. Hosking, D.J., Calcium homeostasis in pregnancy, Clin. Endocrinol., 45, 1, 1996.
77. Khovidhunkit, W. and Epstein, S., Osteoporosis in pregnancy, Osteoporosis Int., 6, 345, 1996.
78. Koeger, A.C., Timsit, M.A., and Oberlin, F., Calcium metabolism in normal pregnancy and lactation,
Rev. Med. Intern., 18, 533, 1997.
79. Koeger, A.C. and Oberlin, F., Calcium and bone disorders during pregnancy and lactation, Rev. Med.
Intern., 18, 546, 1997.
80. Meling, T.R. and Nylen, E.S., Growth hormone deciency in adults: a review, Am. J. Med. Sci., 311,
153, 1996.
81. Russell-Jones, D.L. and Weissberger, A.J., The role of growth hormone in the regulation of body
composition in the adult, Growth Regul., 6, 247, 1996.
0266/ ch04/frame Page 82 Thursday, October 21, 1999 2:12 PM
Factors Affecting Mechanical Properties of Bone 83
82. Marcus, R. and Hoffman, A.R., Growth hormone as therapy for older men and women, Annu. Rev.
Pharmacol. Toxicol., 38, 45, 1998.
83. Bonjour, J.P., Chevalley, T., Ammann, P., et al., Growth hormone, IGF and bone metabolism in the
adult, Ann. Endocrinol., 57, 147, 1996.
84. Eriksen, E.F., Kassem, M., and Langdahl, B., Growth hormone, insulin-like growth factors and bone
remodelling, Eur. J. Clin. Invest., 26, 525, 1996.
85. Rosen, C.J. and Donahue, L.R., Insulinlike growth factors: potential therapeutic options for osteoporo-
sis, Trends Endocrine Metab., 6, 235, 1995.
86. Inzucchi, S.E. and Robbins, R.J., Growth hormone and the maintenance of adult bone mineral density,
Clin. Endocrinol., 45, 665, 1996.
87. Bartalena, L., Bognzzi, F., and Martino, E., Adverse effects of thyroid hormone preparations and
antithyroid drugs, Drug Saf., 15, 53, 1996.
88. Lauwers, A. and Alexandre, C., Impact on bone of thyroid hormone therapy, Rev. Rhum., 64, 112, 1997.
89. Williams, J.B., Adverse effects of thyroid hormones, Drugs Aging, 11, 460, 1997.
90. Williams, G.R., Robson, H., and Shalet, S.M., Thyroid hormone actions on cartilage and bone:
interactions with other hormones at the epiphyseal plate and effects on linear growth, J. Endocrinol.,
157, 391, 1998.
91. Cosman, F. and Lindsay, R., Is parathyroid hormone a therapeutic option for osteoporosis? A review
of the clinical evidence, Calcif. Tissue Int., 62, 475, 1998.
92. Picado, C. and Luengo, M., Corticosteroid-induced bone loss prevention and management, Drug
Saf., 15, 347, 1996.
93. Moe, S.M., The treatment of steroid-induced bone loss in transplantation, Curr. Opinion Nephrol.
Hypertension, 6, 544, 1997.
94. Keenan, G.F., Management of complications of glucocorticoid therapy, Clin. Chest Med., 18, 507,
1997.
95. Schatz, M. and Hamilos, D., Osteoporosis in glucocorticoid-dependent asthmatic patients; literature
review and recommendations, Clin. Immunother., 4, 180, 1995.
96. Lane, N.E. and Lukert, B., The science and therapy of glucocorticoid-induced bone loss, Endocrine
Metab. Clin. North Am., 27, 465, 1998.
97. Li, B.H. and Aspden, R.M., Mechanical and material properties of the subchondral bone plate from
the femoral head of patients with osteoarthritis or osteoporosis, Ann. Rheum. Dis., 56, 247, 1997.
98. Li, B.H. and Aspden, R.M., Composition and mechanical properties of cancellous bone from the
femoral head of patients with osteoporosis or osteoarthritis, J. Bone Miner. Res., 12, 641, 1997.
99. Li, B.H. and Aspden, R.M., Material properties of bone from the femoral neck and calcar femorale
of patients with osteoporosis or osteoarthritis, Osteopororis Int., 7, 450, 1997.
100. Bellingham, C.M., Lee, J.M., Moran, E.L., and Bogoch, E.R., Bisphosphonate (Pamidronate APD)
prevents arthritis-induced loss of fracture toughness in the rabbit femoral diaphysis, J. Orthop. Res.,
13, 876, 1995.
101. Wolinskyfriedland, M., Drug induced metabolic bone disease, Endocrine Metab. Clin. North Am., 24,
395, 1995.
102. Kleerekoper, M., Fluoride and the skeleton, Crit. Rev. Clin. Lab. Sci., 33, 139, 1996.
103. Eastell, R., Boyle, I.T., Compston, J., et al., Management of male osteoporosis: report of the U.K.
Consensus Group, Q. J. Med., 91, 71, 1998.
104. Anderson, F.H., Osteoporosis in men, Int. J. Clin. Pract., 52, 176, 1998.
105. Turan, B., Balcik, C., and Akkas, N., Effect of dietary selenium and vitamin E on the biomechanical
properties of rabbit bones, Clin. Rheum., 16, 441, 1997.
106. Kelepouris, N., Harper, K.D., Gannon, F., et al., Severe osteoporosis in men, Ann. Intern. Med., 123,
452, 1995.
107. Wark, J.D., Osteoporotic fractures: background and prevention strategies, Maturitas, 23, 193, 1996.
108. Auquier, P., Manuel, C., and Molines, C., Risk factors of post menopausal osteoporosis. Analysis of
the literature 19901995, Rev. Epidem. Sante Publique, 45, 328, 1997.
109. Preedy, V.R., Peters, T.J., and Why, H., Metabolic consequences of alcohol dependency, Adverse Drug
React. Toxic. Rev., 16, 235, 1997.
110. Hachulla, E. and Cortet, B., Prevention of glucorticoid-induced osteoporosis, Rev. Med. Intern, 19,
492, 1998.
0266/ ch04/frame Page 83 Thursday, October 21, 1999 2:12 PM
84 Mechanical Testing of Bone and the BoneImplant Interface
111. Gennari, C., Martini, G., and Nuti, R., Secondary osteoporosis, Aging Clin. Exp. Res., 10, 214, 1998.
112. Treves, R., Louer, V., Bonnet, C., et al., Osteoporosis in men, Presse Med., 27, 1647, 1998.
113. Nishi, Y., Zinc and growth, J. Am. Coll. Nutr., 15, 340, 1996.
114. Wilke, H.J., Krischak, S., and Claes, L.E., Formalin xation strongly inuences biomechanical
properties of the spine, J. Biomech., 29, 1629, 1996.
115. Currey, J.D., Brear, K., Zioupos, P., and Reilly, G. C., Effect of formaldehyde xation on some
mechanical properties of bovine bone, Biomaterials, 16, 1267, 1995.
116. Oxlund, H., Barchman, M., Rotoft, G., and Andreassen, T.T., Reduced concentrations of collagen
cross-links are associated with reduced strength of bone, Bone, 17, S365, 1995.
117. McElhaney, J.H., Fogle, J., Byars, E., and Weaver, G., Effect of embalming on the mechanical
properties of beef bone, J. Appl. Physiol., 19, 1234, 1964.
118. Borchers, R.E., Gibson, L.J., Burchardt, H., and Hayes, W.C., Effects of selected thermal variables
on the mechanical properties of trabecular bone, Biomaterials, 16, 545, 1995.
119. Smith, C.W., The Mechanical Properties of Biological Materials Stored by Tissue Banks, Ph.D. thesis,
University of Leeds, U.K., 1995.
120. Viceconti, M., Toni, A., Brizio, L., et al., The effect of autoclaving on the mechanical properties of
bank bovine bone, Chir. Organi. Mov., 81, 63, 1996.
121. Kohler, P., Kreicbergs, A., and Stromberg, L., Physical properties of autoclaved bone. Torsion test of
rabbit diaphyseal bone, Acta Orthop. Scand., 57, 141, 1986.
122. Knaepler, H., Haas, H., and Puschel, H.U., [Biomechanical properties of heat and irradiation treated
spongiosa], Unfallchirurgie, 17, 194, 1991.
123. Linde, F. and Srensen, H.C., The effect of different storage methods on the mechanical properties
of trabecular bone, J. Biomech., 26, 1249, 1993.
124. Kang, Q., An, Y.H., and Friedman, R.J., Effects of multiple freezingthawing cycles on ultimate
indentation load and stiffness of bovine cancellous bone, Am. J. Vet. Res., 58, 1171, 1997.
125. Strmberg, L. and Daln, N., The inuence of freezing on the maximum torque capacity of long
bones. An experimental study on dog, Acta Orthop. Scand., 47, 254, 1976.
126. Tomford, W.W., Doppelt, S.H., Mankin, H.J., and Friedlaender, G.E., 1983 bone bank procedures,
Clin. Orthop., 174, 15, 1983.
127. Sumner, D.R., Willke, T.L., Berzins, A., and Turner, T.M., Distribution of Youngs modulus in the
cancellous bone of the proximal canine tibia, J. Biomech., 27, 1095, 1994.
128. Malinin, T.I., Martinez, O.V., and Brown, M.D., Banking of massive osteoarticular and intercalary
bone allografts-12 years experience, Clin. Orthop., 197, 44, 1985.
129. Panjabi, M.M., Krag, M., Summers, D., and Videman, T., Biomechanical time-tolerance of fresh
cadaveric human spine specimens, J. Orthop. Res., 3, 292, 1985.
130. Roe, S.C., Pijanowski, G.J., and Johnson, A.L., Biomechanical properties of canine cortical bone
allografts: effects of preparation and storage, Am. J. Vet. Res., 49, 873, 1988.
131. Sedlin, E.D. and Hirsch, C., Factors affecting the determination of the physical properties of femoral
cortical bone, Acta Orthop. Scand., 37, 29, 1966.
132. Pelker, R.R., Friedlaender, G.E., and Markham, T.C., et al., Effects of freezing and freeze-drying on
the biomechanical properties of rat bone, J. Orthop. Res., 1, 405, 1984.
133. Currey, J.D., The effects of drying and re-wetting on some mechanical properties of cortical bone,
J. Biomech., 21, 439, 1988.
134. Triantafyllou, N., Sotiropoulos, E., and Triantafyllou, J.N., The mechanical properties of lyophilized
and irradiated bone grafts, Acta Orthop. Belg., 41, 35, 1975.
135. Bright, R.W. and Burstein, A.H. Material properties of preserved cortical bone, Trans. Orthop. Res.
Soc., 24, 210, 1978.
136. Simonian, P.T., Conrad, E.U., Chapman, J.R., et al., Effect of sterilization and storage treatments on
screw pullout strength in human allograft bone, Clin. Orthop., 302, 290, 1994.
137. Webster, D.A. and Werner, F.W., Mechanical and functional properties of implanted freeze-dried exor
tendons, Clin. Orthop., 180, 302, 1983.
138. Jackson, D.W., Grood, E.S., Arnoczky, S.P., et al., Freeze-dried anterior cruciate ligament allografts,
Am. J. Sports Med., 15, 295, 1988.
139. Smith, C.W., Young, I.S., and Kearney, J.N., Mechanical properties of tendons: changes with steril-
ization and preservation, J. Biomech. Eng., 118, 56, 1996.
0266/ ch04/frame Page 84 Thursday, October 21, 1999 2:12 PM
Factors Affecting Mechanical Properties of Bone 85
140. Sugimoto, M., Takahashi, S., Toguchida, J., et al., Changes in bone after high dose irradiation
biomechanics and histomorphology, J. Bone Joint Surg., 73B, 492, 1991.
141. Anderson, M.J., Keyak, J.H., and Skinner, H.B., Compressive mechanical properties of human can-
cellous bone after gamma irradiation, J. Bone Joint Surg., 74A, 747, 1992.
142. Smith, C.W. and Kearney, J.N., The effects of irradiation and hydration upon the mechanical properties
of tendon, J. Mater. Sci. Mater. Med., 7, 645, 1996.
143. Edgerton, B.C., An, K.N., and Morrey, B.F., Torsional strength reduction due to cortical defects in
bone, J. Orthop. Res., 8, 851, 1990.
144. Hipp, J.A., Edgerton, B.C., An, K.N., and Hayes, W.C., Structural consequences of transcortical holes
in long bones loaded in torsion, J. Biomech., 23, 1261, 1990.
145. An, Y.H. and Bell, T.D., Experimental design, evaluation methods, data analysis, publication, and
research ethics, in Animal Models in Orthopaedic Research, An, Y.H. and Friedman, R.J., Eds., CRC
Press, Boca Raton, FL, 1999.
146. Zioupos, P., Currey, J.D., Mirza, M.S., and Barton, D.C., Experimentally determined microcracking
around a circular hole in a at plate of bone: comparison with predicted stresses, Philos. Trans. R.
Soc. London B, 347, 383, 1995.
147. Ashman, R.B., Experimental techniques, in Bone Mechanics, Cowin, S.C., Ed., CRC Press, Boca
Raton, FL, 1989, 91.
148. Wall, J.C., Chatterji, S., and Jeffery, J.W., On the origin of scatter in results of human bone strength
tests, Med. Biol. Eng., 8, 171, 1970.
0266/ ch04/frame Page 85 Thursday, October 21, 1999 2:12 PM
0266/ ch04/frame Page 86 Thursday, October 21, 1999 2:12 PM

87

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

5

Basic Facilities and Instruments
for Mechanical Testing of Bone

Christopher V. Bensen and Yuehuei H. An

CONTENTS

I. Introduction ............................................................................................................................87
II. Mechanical Testing Machines................................................................................................88
A. Components of Mechanical Testing Machines................................................................88
1. Hydraulic Power Supply Units ...................................................................................88
2. Load Units ...................................................................................................................88
3. Force Transducers........................................................................................................89
4. Controllers ...................................................................................................................89
5. Sample-Holding Devices.............................................................................................89
6. Data Management Systems.........................................................................................91
B. Common Types of Mechanical Testing Machines or Devices........................................92
1. Single-Axis Machines .................................................................................................92
2. Multiaxial Machines....................................................................................................92
3. Micromechanical Testing Machines ...........................................................................93
4. Hardness/Indentation Testing Devices ........................................................................93
III. Sample Potting Materials.......................................................................................................95
IV. Other Basic Items...................................................................................................................97
V. Summary ................................................................................................................................98
References ......................................................................................................................................100

I. INTRODUCTION

The biomechanical evaluation of bone, bone implants, and the boneimplant interface has been
carried out for many years. Such investigations nearly always employ the use of mechanical testing
systems to generate information on the physical properties of these materials. From simple com-
pression and tension failure testing to fatigue analysis of new total joint prostheses, modern
computer-driven machines are commonly used to provide analysis and information.
Increased prevalence of debilitating conditions such as degenerative joint disease as well as
rising popularity of internal devices for fracture xation has led to rapid growth of the orthopaedic
and biomechanical research communities. Along with this growth has been a commensurate rise
in the diversity and production of commercially available testing systems to meet the ever-increasing
demand for better and less expensive implants and the specic needs of the modern investigator.
Implementation of a materials testing laboratory is neither an easy nor inexpensive endeavor.
However, the utility and potential capabilities of even the most basic laboratory can provide the
opportunity to perform numerous experiments and far outweigh the initial difculties or expenses
encountered. In addition to being a useful research platform, a mechanical testing laboratory is a

0266/ ch05/frame Page 87 Thursday, October 21, 1999 2:17 PM

88

Mechanical Testing of Bone and the BoneImplant Interface

valuable and practical hands-on teaching tool for orthopaedic surgery residents, graduate students,
and technicians alike.
This chapter reviews the scope of modern mechanical testing systems, both commercial and
otherwise, and examines the requirements for implementation of a materials testing laboratory.
Finally, the chapter provides information on necessary ancillary materials which facilitate the
execution of biomechanical investigations. It is the authors hope that this information will be useful
to both the established and the new investigator attempting to set up a materials testing laboratory.

II. MECHANICAL TESTING MACHINES

Although numerous types of mechanical testing systems are currently in production, the vast
majority can be divided into two types based on the methods by which load is applied to the
specimen: servohydraulic and electromechanical. Servohydraulic testing machines comprise the
majority of systems in use for examination of orthopaedic materials. In these systems, a servovalve
is used to transform electrical energy into hydraulic uid pressure. This pressure is then applied
as load to the specimen. Even smaller servohydraulic machines are capable of delivering relatively
large loads to test specimens, often exceeding 25 kN. Electromechanical machines provide relatively
small loads, usually 1 kN or less. They are suitable for evaluating smaller specimens, such as
sutures, bonetendon complexes, and screw pullout models. There are numerous types of servo-
hydraulic systems currently available and their components and major types along with their
potential applications are described below.

A. C

OMPONENTS



OF

M

ECHANICAL

T

ESTING

M

ACHINES

The typical mechanical testing machine has several key components, including the hydraulic power
supply (HPS), actuator, controller, load unit, force transducer, xture devices, and data generator.
Each of these integral components is briey described below.

1. Hydraulic Power Supply Units

The HPS provides the uid pressure necessary to drive the actuator piston. Hydraulic uid is
circulated at ow rates between 3 gallons per minute (gpm) to over 200 gpm on larger models
providing continuous pressures of 1000 to 3000 psi. The regulation of hydraulic uid ow and
pressure is controlled by a servovalve. The servovalve is usually mounted to an actuator and uses
changes in voltage to govern direction and ow of hydraulic uid, which consequently moves the
actuator rod. Figure 5.1 illustrates a typical HPS unit. HPS units can be either air cooled or water
cooled depending on the size and maximum pressures attained by the system. HPS units are typically
protected from overheating and low uid levels by electronic interlocks that automatically turn off
the unit should hazardous conditions exist.

2. Load Units

The load unit of the system employs actuator rods to apply load to the specimen. These actuators
can be mounted in linear, rotary, or angular alignment depending on the system. In servohydraulic
systems, the actuator operates under control of a servovalve and contains a linear variable differential
transducer (LVDT), which provides rod displacement information to the controller. In axial-only
load units, there is a single linear mounted actuator rod which can apply compression or tensile
forces to a specimen. For torsional forces, a rotary actuator is added to the linear one allowing the
application of biaxial forces to the specimen. More complex and specialized load units are also
available for a variety of applications.

0266/ ch05/frame Page 88 Thursday, October 21, 1999 2:17 PM

Basic Facilities and Instruments for Mechanical Testing of Bone

89

3. Force Transducers

Load applied to specimens during testing procedures is measured with a force transducer, commonly
called a load cell, which is mounted on the xed side opposite the load unit. Transducers are
available in a variety of load ranges and should be selected based on both the operating limitations
of the actuator as well as the intended use of the testing system. Another type of transducer
commonly used in mechanical testing of bone is the extensometer. This device measures the
displacement and/or strain on a specimen which is typically being tested to loads below the yield
point.

4. Controllers

Controllers provide several functions including control of the HPS and servovalve, transducer
conditioning, function generation, and data output. The most basic controllers are usually single-
channel devices which are suited to uniaxial testing. They typically include function generators
which provide waveforms such as sine and square which are commonly used in fatigue testing.
The controller is used to set testing variables and parameters including the rate of load application
to the specimen, peak load, and maximum displacement of the actuator rod. Tests are based either
on displacement control in which displacement is carried out at a constant rate while the load is
measured or on displacement control in which displacement is measured during a constant loading
rate. Figure 5.2 illustrates a typical controller for basic mechanical testing systems. More powerful
controllers have the capability of receiving externally generated command signals and even allow
the design and programming of specic testing protocols. Controllers for joint simulation systems
are multichannel and are capable of controlling multiaxial motion in a temperature-controlled
environment for several testing stations.

5. Sample-Holding Devices

A critical aspect of any testing procedure is mounting the specimen on the mechanical testing
system. It is imperative that the specimen be held rmly without slippage and, more importantly,
without damage to the specimen itself. Incomplete or irregular gripping of the specimen can cause
premature failure at the specimenclamp interface and can lead to erroneous testing values.

FIGURE 5.1

Typical hydraulic power supply unit for a small mechanical testing machine (MTS Model 512).
(Courtesy of MTS Systems, Inc., Eden Prairie, MN.)

0266/ ch05/frame Page 89 Thursday, October 21, 1999 2:17 PM

90

Mechanical Testing of Bone and the BoneImplant Interface

Numerous holding devices and grips are commercially available for this purpose. For bone and
bonetendon complexes, simple screw clamps often will sufce. For testing wires, sutures, small
tendons, and similar specimens, specialized grips can facilitate the testing and minimize risk of
damage to the specimen (Figure 5.3).
Over the past several years, pneumatically driven holding devices have become the standard
method of specimen xation. These units have interchangeable grips for a variety of testing
requirements and have signicantly reduced the problems associated with specimen xation.

FIGURE 5.2

Single-channel controller useful for all types of single-axis research (MTS Model 407).
(Courtesy of MTS Systems, Inc., Eden Prairie, MN.)

FIGURE 5.3

Typical pneumatic grips used for holding small specimens. (Courtesy of MTS Systems, Inc.,
Eden Prairie, MN.)

0266/ ch05/frame Page 90 Thursday, October 21, 1999 2:17 PM

Basic Facilities and Instruments for Mechanical Testing of Bone

91

An alternative to these grips and holding devices is available if there is access to a machine
shop. Simple, yet useful clamps and grips with screw and wing nut locking mechanisms can be
machined from stock aluminum or stainless steel at a fraction of the cost of commercially manu-
factured ones and will often sufce, especially with durable specimens. In fact, most investigations
are carried out with custom-designed grips and xtures. An example of such a xture used for
testing screw pullout strength is shown in Figure 5.4. This xture was designed and produced in
the authors laboratory.

6. Data Management Systems

Historically, data acquisition from most mechanical testing systems was achieved by means of a
graphic chart recorder. Figure 5.5 shows an older single-axis testing system with a chart recorder
in one of the laboratories at the authors institution.
Loaddisplacement curves can be easily plotted using such a device; however, meticulous
manual measurements and calculations are required to determine biomechanical parameters of the
specimen. Consequently, there is additional risk of obtaining errors in data analysis.
In modern testing machines, data are transferred from the controller to a computer where data
points can be stored and parameters can be automatically calculated. There are numerous software
packages available for mechanical testing systemlinked PCs such as LabView (National Instru-
ments Corporation, Austin, TX). These programs can greatly facilitate data acquisition, manage-
ment, and analysis. Alternatively, the major mechanical testing machine companies also offer
software support both for the controller and PC-based systems.

FIGURE 5.4

Sample holding device used for screw pullout testing.

0266/ ch05/frame Page 91 Thursday, October 21, 1999 2:17 PM

92

Mechanical Testing of Bone and the BoneImplant Interface

B. C

OMMON

T

YPES



OF

M

ECHANICAL

T

ESTING

M

ACHINES



OR

D

EVICES

1. Single-Axis Machines

Mechanical testing systems with a single linear actuator are suitable for a very large variety of
testing procedures and are the systems most widely used today (Figure 5.6). Compression, tensile,
bending, and indentation or hardness testing can all be accomplished with single-axis machines.
Compression testing is commonly used to determine the mechanical properties of both cortical and
cancellous bones.

1

Tensile testing is also very commonly used to measure the elasticity of bones,
tendons, ligaments, as well as bonetendon and boneligamentbone complexes. Three-point bend-
ing tests may also be performed on either standard-cut cortical specimens or intact bones. Finally,
single-axis machines can be used for screw pullout tests to compare various types of screws or
screw xation. Pushout testing is also one of the most common and simplest ways to evaluate the
boneimplant interface.

2-4

2. Multiaxial Machines

The addition of a rotary actuator allows simultaneous testing of two mechanical properties of a
specimen. This type of machine is essential for numerous types of testing, including that of the
spine (Figure 5.7). Additionally, torsional testing of whole long bones and other materials can be
evaluated while an axial load is being applied. This load simulates the body weight of an animal
and more closely resembles

in vivo

physiological forces.

FIGURE 5.5

Older model mechanical testing machine with chart recorder.

0266/ ch05/frame Page 92 Thursday, October 21, 1999 2:17 PM

Basic Facilities and Instruments for Mechanical Testing of Bone

93

3. Micromechanical Testing Machines

Evaluation of the microstructure of bone requires mechanical testing of single osteons or trabeculae.
Tension, compression, bending, and torsional tests of single osteons have all been undertaken over
the past several decades.

5-9

Highly specialized equipment is required in these investigations, as
sample isolation, preparation, and mounting all require a higher degree of precision. Numerous
devices have been designed and assembled by investigators themselves and one example of such
a machine developed by Mente is shown in Figure 5.8.

10

Another example is the tensile micro-
mechanical testing device developed by Rho (Figure 5.9). Additionally, several of the major
mechanical testing machine companies have the capability to custom-design testing platforms to
specications required by the investigator. A few machines are also commercially available
(Figure 5.10). Micromechanical testing is covered further in Chapters 18 and 19.

4. Hardness/Indentation Testing Devices

Another specialized mechanical testing system is the hardness or indentation testing machine.
Hardness is a common parameter used to compare resistance of materials (including bone) to
indentation or abrasion. There are a variety of testing systems based on the size and geometry of
the specimens and the loads applied. First, macroindentation testing such as the Brinell method is
useful for evaluating the hardness of specimens such as cortical or cancellous bone surfaces using
loads of 1 kg or greater (Figure 5.11).

11,12

Microhardness testing using the Knoop and Vickers
indenters allows evaluation of resistance of bone to indentation which can reect biochemical

FIGURE 5.6

Tabletop single-axis mechanical testing machine (MTS Mini Bionix Model 858). (Courtesy of
MTS Systems, Inc., Eden Prairie, MN.)

0266/ ch05/frame Page 93 Thursday, October 21, 1999 2:17 PM

94

Mechanical Testing of Bone and the BoneImplant Interface

properties such as mineral content.

13-15

Finally, over the past decade, the development of nano-
indentation testing machines has allowed the testing of mechanical properties of specimens on the
order of nanometers, such as individual trabeculae and osteonic lamellae.

16-18

Figure 5.12 shows a
typical nanoindentation testing device. There are also combination imagingtesting devices such
as the TriboScope

!

(Figure 5.13, Hysitron, Inc., Minneapolis, MN). This device combines atomic
force microscopy with a nanoindenter which allows the investigator to image the sample, choose

FIGURE 5.7

Multiaxis mechanical testing machine capable of torsional testing. (Courtesy of Instron
Corporation, Canton, MA.)

FIGURE 5.8

Micromechanical testing machine. (Courtesy of Dr. Peter Mente.)

0266/ ch05/frame Page 94 Thursday, October 21, 1999 2:17 PM

Basic Facilities and Instruments for Mechanical Testing of Bone

95

the test location, indent, scratch, and wear surfaces with a single device. Nanoindentation testing
is covered in further detail in Chapter 17.

III. SAMPLE POTTING MATERIALS

Many specimens cannot be mounted directly to the mechanical testing system and must be potted
in another medium in order to be tested. These may include small and/or irregularly shaped
specimens which do not afford adequate purchase to which standard grips can be attached, as well

FIGURE 5.9

Tensile micromechanical testing machine. (Courtesy of Dr. Jae-Rong Rho.)

FIGURE 5.10

Latour-Black dynamic micromechanical tester. (Courtesy of Dynatek Dalta Scientific Instru-
ments, Galena, MO.)

0266/ ch05/frame Page 95 Thursday, October 21, 1999 2:17 PM

96

Mechanical Testing of Bone and the BoneImplant Interface

as complex constructs involving several bones. In addition, potting a specimen allows precise
alignment and allows testing with minimal risk of damaging the integrity of the specimen.
A variety of materials are available for this purpose including polymethylmethacrylate
(PMMA), plaster of paris,

19

epoxy resins,

20

and calcium sulfate-based materials such as dental
stone.

21

One such material typically used for creating dental casts from alginate impressions is
called Labstone (Heraeus Kulzer, Inc., Dental Products Division, South Bend, IN). This material
comes in powder form, is mixed with water in either a disposable or exible, reuseable container,
and has a curing time of approximately 8 to 10 min. This provides ample working time to set the
specimen, yet hardens quickly enough such that it can be tested within an hour or so. When dry,
it has a compression strength of 8000 psi and volumetric expansion of only 0.12%. The authors
have used this material in several investigations without complications. An example of the use of
this material in the authors laboratory is shown in Figure 5.14. In this experiment, the biomechan-
ical properties of various methods of scapulothoracic arthrodesis was examined. Three cadaver ribs

FIGURE 5.11

Brinell-Rockwell-Vickers optical hardness tester. (Courtesy of Nanotek, Inc., Opelika, AL.)

0266/ ch05/frame Page 96 Thursday, October 21, 1999 2:17 PM

Basic Facilities and Instruments for Mechanical Testing of Bone

97

were potted in anatomic position prior to xation of the scapula. The entire construct could then
be mounted on the mechanical testing system using C-clamps at the base of the construct, providing
a solid link between the specimen and system.

22

IV. OTHER BASIC ITEMS

Proper preparation of specimens for mechanical testing requires several key tools. A band saw with
a

-in. ne-tooth blade is an invaluable general resource for preparing gross bone samples for
testing (Figure 5.15). For making small, cylindrical bone specimens for compression testing, a
tabletop drill press with a trephine bit is sufcient (Figure 5.16). However, a lathe or milling machine
is more appropriate for larger specimens. A wheel grinder/polisher (Figure 5.17) is also a useful
tool in nal preparation of bone specimens in which specic dimensions must be identical. Elec-
tronic calipers also aid in this procedure.

FIGURE 5.12

Nanoindentation machine (Nanindenter XP). (Courtesy of MTS Systems, Inc., Eden Prairie, MN.)

FIGURE 5.13

Triboscope. (Courtesy of Hysitron, Inc., Minneapolis, MN.)

0266/ ch05/frame Page 97 Thursday, October 21, 1999 2:17 PM

98

Mechanical Testing of Bone and the BoneImplant Interface

One common method of storing bones is freezing at 20

C; however, 70

C is ideal. Numerous
investigators have shown that there is little or no detrimental effect on the mechanical properties
of bone that is stored at 20

C for periods up to 8 months.

23-25

It is important to freeze bone
specimens in an airtight plastic bag containing normal saline. Because it may be impractical to
collect all specimens for a single experiment at once, a large-volume freezer is an essential
component to a mechanical testing laboratory.
Finally, an operating table and some basic surgical instruments (such as a scalpel, forceps, and
scissors), as well as a supply of surgical towels, gloves, plastic bags, saline, etc. should also be
readily available in or near the laboratory for proper specimen preparation.

V. SUMMARY

In conclusion, mechanical testing systems offer the orthopaedic researcher the ability to measure
numerous properties of a bone specimen or construct. A large variety of machines are commer-
cially available from several companies; it is up to the individual researcher or team to decide
which model is appropriate for the research being carried out in the respective laboratory. It is
the authors opinion that a single-axis system affords a relatively inexpensive, yet versatile tool
with which mechanical testing of bone can be performed. The vast majority of testing procedures
commonly in use, including compression, indentation, and three-point bending, can be achieved

FIGURE 5.14

Illustration of the use of Labstone (Denstone) to embed ribs for testing various methods of
scapulothoracic arthrodesis.

0266/ ch05/frame Page 98 Thursday, October 21, 1999 2:17 PM

Basic Facilities and Instruments for Mechanical Testing of Bone

99
FIGURE 5.15

Illustration of an 8-in. tabletop band saw.

FIGURE 5.16

Tabletop drill press.

0266/ ch05/frame Page 99 Thursday, October 21, 1999 2:17 PM

100

Mechanical Testing of Bone and the BoneImplant Interface

with a single-axis machine. Alternatively, should specialized testing situations demand, multiaxis
and even custom-built machines are also available from the major manufacturers.
As the demand for improved implants and devices in the treatment of disorders of the muscu-
loskeletal system continues to rise, it is the authors hope that the technology of the testing systems
designed for their evaluation will continue to improve.

REFERENCES

1. Linde, F., Elastic and viscoelastic properties of trabecular bone by a compression testing approach,

Dan. Med. Bull.,

41, 119, 1994.
2. Dhert, W.J., Verheyen, C.C., Braak, L.H., et al., A nite-element analysis of the pushout test: inuence
of test conditions,

J. Biomed. Mater. Res.,

26, 119, 1992.
3. Friedman, R.J., Bauer, T.W., Garg, K., et al., Histological and mechanical comparison of hydroxya-
patite-coated cobalt-chrome and titanium implants in the rabbit femur,

J. Appl. Biomater.,

6, 231, 1995.
4. An, Y.H., Friedman, R.J., Jiang, M., et al., Bone ingrowth to implant surfaces in an inammatory
arthritis model,

J. Orthop. Res.,

16, 576, 1998.
5. Ascenzi, A., Baschieri, P., and Benvenuti, A., The bending properties of single osteons,

J. Biomech.,

23, 763, 1990.
6. Ascenzi, A., Baschieri, P., and Benvenuti, A., The torsional properties of single selected osteons [see
comments],

J. Biomech.,

27, 875, 1994.
7. Ascenzi, A. and Bonucci, E., The compressive properties of single osteons,

Anat. Rec.,

161, 377, 1968.
8. Frasca, P., Harper, R.A., and Katz, J.L., Micromechanical oscillators and techniques for determining
the dynamic moduli of microsamples of human cortical bone at microstrains,

J. Biomech. Eng.,

103,
146, 1981.
9. Frasca, P., Jacyna, G., Harper, R., and Katz, J.L., Strain dependence of dynamic Youngs modulus for
human single osteons,

J. Biomech.,

14, 691, 1981.
10. Mente, P.L. and Lewis, J.L., Experimental method for the measurement of the elastic modulus of
trabecular bone tissue,

J. Orthop. Res.,

7, 456, 1989.
11. Bjrkstrom, S. and Goldie, I.F., Hardness of the subchondral bone of the patella in the normal state,
in chondromalacia, and in osteoarthrosis,

Acta Orthop. Scand.,

53, 451, 1982.
12. Markel, M.D., Wikenheiser, M.A., and Chao, E.Y., A study of fracture callus material properties:
relationship to the torsional strength of bone,

J. Orthop. Res.,

8, 843, 1990.
13. Knoop, P., Peters, C.G., and Emerson, W.B., A sensitive pyramidal-diamond tool for indentation
measurements,

J. Res. Natl. Bur. Stand.,

23, 39, 1939.
14. Huja, S.S., Katona, T.R., Moore, B.K., and Roberts, W.E., Microhardness and anisotropy of the vital
osseous interface and endosseous implant supporting bone,

J. Orthop. Res.,

16, 54, 1998.

FIGURE 5.17

Wheel grinder/polisher. (

Note:

Tropical plant foliage improves overall working conditions in
the mechanical testing laboratory.)

0266/ ch05/frame Page 100 Thursday, October 21, 1999 2:17 PM

Basic Facilities and Instruments for Mechanical Testing of Bone

101

15. Amprino, R., Investigations on some physical properties of bone tissue,

Acta Anat.,

34, 161, 1958.
16. Lee, F.Y., Rho, J.Y., Harten, R., Jr., et al., Micromechanical properties of epiphyseal trabecular bone
and primary spongiosa around the physis: an

in situ

nanoindentation study,

J. Pediatr. Orthop.,

18,
582, 1998.
17. Rho, J.Y., Tsui, T.Y., and Pharr, G.M., Elastic properties of human cortical and trabecular lamellar
bone measured by nanoindentation,

Biomaterials,

18, 1325, 1997.
18. Zysset, P.K., Guo, X.E., Hofer, C.E., et al., Mechanical properties of human trabecular bone lamellae
quantied by nanoindentation,

Technol. Health Care,

6, 429, 1998.
19. Davis, P.K., Mazur, J.M., and Coleman, G.N., A torsional strength comparison of vascularized and
nonvascularized bone grafts,

J. Biomech.,

15, 875, 1982.
20. Paavolainen, P., Studies on mechanical strength of bone. I. Torsional strength of normal rabbit tibio-
bular bone,

Acta Orthop. Scand.,

49, 497, 1978.
21. Lepola, V., Vaananen, K., and Jalovaara, P., The effect of immobilization on the torsional strength of
the rat tibia,

Clin. Orthop.

, 55, 1993.
22. Bensen, C.V., Bareld, W.R., Draughn, R.A., and Thompson, J.D., Biomechanical evaluation of
methods of scapulothoracic fusion for treatment of FSH muscular dystrophy,

Trans. Orthop. Res.
Soc.,

23, 1147, 1998.
23. Strmberg, L. and Daln, N., The inuence of freezing on the maximum torque capacity of long
bones. An experimental study on dogs,

Acta Orthop. Scand.,

47, 254, 1976.
24. Roe, S.C., Pijanowski, G.J., and Johnson, A.L., Biomechanical properties of canine cortical bone
allografts: effects of preparation and storage,

Am. J. Vet. Res.,

49, 873, 1988.
25. Panjabi, M.M., Krag, M., Summers, D., and Videman, T., Biomechanical time-tolerance of fresh
cadaveric human spine specimens,

J. Orthop. Res.,

3, 292, 1985.

0266/ ch05/frame Page 101 Thursday, October 21, 1999 2:17 PM

0266/ ch05/frame Page 102 Thursday, October 21, 1999 2:17 PM

103

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

6

Methods of Evaluation for
Bone Dimensions, Densities,
Contents, Morphology,
and Structures

Yuehuei H. An, William R. Bareld, and Ivars Knets

CONTENTS

I. Introduction ..........................................................................................................................103
II. Macromeasurements of Bone Dimensions ..........................................................................104
A. Direct Measurements......................................................................................................104
B. Measurements Based on X-ray Images .........................................................................104
C. Measurements of Microspecimens.................................................................................105
D.

In Vivo

Measurements....................................................................................................105
E. Orientator........................................................................................................................105
III. Micromeasurement of Bone Structures ...............................................................................105
A. Histomorphometry..........................................................................................................105
B. Scanning Electron Microscopy ......................................................................................107
C. Confocal Microscopy .....................................................................................................107
D. Ridge Number Density...................................................................................................107
IV. Bone Apparent Density, Material Density, Mineral Content, and Mineral Density...........107
A. Denitions and Methods of Evaluation.........................................................................107
B. Bone Density Values Measured by

in Vitro

Methods ...................................................109
V. Special Techniques...............................................................................................................110
A. Computed Tomography..................................................................................................110
B. Magnetic Resonance Imaging........................................................................................111
C. Single-Photon, Single X-ray, and Dual-Energy Absorptiometry ..................................111
D. Ultrasound ......................................................................................................................111
VI. Summary ..............................................................................................................................113
References ......................................................................................................................................113

I. INTRODUCTION

Bone dimensions (such as length, thickness, area, and volume), densities, mineral contents, and
structure at both macro- and microlevels, even at the nanolevel, are basic parameters for bone
mechanical testing and data analysis. Bone strength and modulus are assessed through determining
the macroscopic and microscopic characteristics, which include the size, shape, density, and archi-
tecture of the bone, as well as the tested mechanical values. Although simple to execute, the accuracy
of these measurements is frequently related to the validity of the methods being used.

0266/ ch06/frame Page 103 Thursday, October 21, 1999 2:22 PM

104

Mechanical Testing of Bone and the BoneImplant Interface

II. MACROMEASUREMENTS OF BONE DIMENSIONS

Many methods have been used for measuring length, area, and volume of bone tissues. These
methods include direct measurement using a ruler, micrometer, or caliper, measurement based on
radiographs, or the use of specially designed devices.

1,2

There may be signicant intermethod or
interobserver variability.

3,4

Therefore, several observers may be needed to perform the same pro-
cedure independently, or more than one method may be used. Caution should be taken when
interstudy comparison is made, especially if different measurement techniques are employed. Bone
mineral content and the structural properties of femoral cross sections have been calculated from
simplifying assumptions based on geometric properties, moduli, and whole-bone strength indexes.

5

A. D

IRECT

M

EASUREMENTS

Most length and area measurements such as the measurements of long bone dimensions (length,
internal and external width) are accomplished reliably by such traditional techniques as a ruler or
sliding digital caliper.

3,4

Intraobserver error can result from the use of calipers, but can be lessened
with use of an electronic digitizer.

6

A specially designed electronic caliper has also been reported.

7

A magnifying glass to measure radiographs has also been demonstrated as useful. This technique
has been reported to provide greater reliability and precision than caliper measurement when
measuring bone width and cortical thickness.

7,8

According to a recent report, a digital coordinatometer-goniometer connected to an optical
collimator allows angles and bone lengths to be measured. This new technique allows for location
of the bone axis and torsion values using a reference plane and symmetry and tangency criteria.
Findings are that angular measures are more precise than through use of traditional instruments.

9

The archaeological collection of Pecos Pueblo femora and tibiae were examined through use
of an electronic digitizer and computer program. Geometric cross sections were determined using
a FORTRAN computer program which divided the total bone area of interest into a series of
small geometric shapes which were subsequently summed to determine the total composite
properties. Cross sections were scaled and electronically digitized with subperiosteal and
endosteal boundaries manually traced with a stylus. The geometric data generated from these
bone samples pointed to specic

in vivo

loading of the lower limb which served as a natural
protective mechanism to manage effectively the stressstrain loadings sustained by the Pecos
Pueblo people during this period of history.

10

B. M

EASUREMENTS

B

ASED



ON

X-

RAY

I

MAGES

Two-dimensional (2D) measurements are often made from X-ray images using a ruler or cali-
per.

6,8,11,12

The percentage magnication should be considered when using X-ray images for meas-
urements. The amount of magnication depends on the distance between the specimen and the
lm. A metal bar or strip with known length can be used as a reference. It should be placed at the
same distance from the lm as the subject. A standard goniometer is effective for measuring angles
based on radiographic images. A custom computer program based on X-ray images has been
developed in the authors laboratory. It is capable of evaluating the periosteal and endosteal
dimensions of the upper humerus and glenoid. Parameters which may be evaluated include humeral
canal width, shaft width, tuberosity offset, head offset, radius of curvature of the head and glenoid,
head diameter, canal are index, glenoid height and depth, arc of enclosure, radius of curvature,
and depth of cancellous bone.

13-15

The length or perimeter of irregular lines or the area of irregular bone or tissue specimens can
be measured using computer image analysis. Most image software has the capacity to measure
length and area. Images of interest (photographs, radiographs, or prints) can be scanned into the

0266/ ch06/frame Page 104 Thursday, October 21, 1999 2:22 PM

Methods of Evaluation for Bone Dimensions, Densities, Contents, Morphology, and Structures

105

computer, displayed on the screen, outlined, and measured. Careful calibration of the software is
necessary before accurate measurements can be made.
Measurement of the torsion angle of long bones (in humans) and the femoral angle of inclination
(in dogs) has been reported with the use of computed tomography (CT),

16

a digital coordinator-
goniometer,

9

and a symmetric axis-based method.

17

C. M

EASUREMENTS



OF

M

ICROSPECIMENS

Implicitly, when bone microspecimens are tested, the assumption is that the tissue of interest has
homogeneous physical properties microscopically, and thereby represents the entire specimen.

18-21

A dissecting microscope can be used for simple distance measurement, such as thickness of
trabeculae or microcortical bone beam.

22

Dissecting microscopy with magnication up to 4

has
been used for examining and photographically documenting the surface morphology of bone or
implanttissue interface. Wet specimens should be used, which is an advantage over regular, low-
magnication scanning electron microscopy (SEM), because during specimen preparation, using
critical point drying, the morphology of the specimen may be changed.

D.

I

N

V

IVO

M

EASUREMENTS

In vivo

measurement of bone length or limb length is a challenge. Often, soft tissue landmarks are
drawn on the skin and the distance between the marks is measured with a tape measure. A technique
called kyniklometry has been reported for measuring the distance between soft tissue landmarks
on the lower legs of conscious rabbits. This technique has been demonstrated as comparable to
X-ray stereophotogrammetry.

23

A specially designed goniometer has been reported for the meas-
urement of joint angles in clinical practice.

24

Limb circumference measurements are useful for
monitoring the progress of a swollen limb or joint, or the growth of a limb.



For the quantication
of limb circumference, a tape measure is effective. Methods using an electronic digitizer and a
mathematical formula for an ellipse (for fetal head and body circumferences) have also been
reported.

25

E. O

RIENTATOR

The orientator is a technique developed approximately 10 years ago for the estimation of length,
surface density, and other stereological parameters using isotropic sections.

26

No special equipment
is required. Knowledge of the axis of anisotropy optimizes the efciency. Random points having
uniform probability in space are selected using combinations of simple, stratied, and systematic
random sampling. A mapping algorithm ensures isotropic planes, thereby allowing unbiased esti-
mates of surface and length density measurements through classical stereological formulae. The
stereological approach randomizes a three-dimensional (3D) polar coordinate system whereby every
direction can be dened by a pair of angles and each direction can be graphically represented by
a point on the unit sphere.

26

III. MICROMEASUREMENT OF BONE STRUCTURES
A. H

ISTOMORPHOMETRY

Parafn embedding and sectioning remains the most common method for histological study of soft
tissues (subcutaneous tissue, muscle, tendon, ligament), cartilage, and also decalcied bone spec-
imens. Undecalcied preparation and sectioning are specialized procedures for the evaluation of
osseous tissues (bone, calcied tissues), dental tissues, and especially specimens containing metal

0266/ ch06/frame Page 105 Thursday, October 21, 1999 2:22 PM

106

Mechanical Testing of Bone and the BoneImplant Interface

implants. In this technique specimens are embedded in plastic media such as methylmethacrylate
or Spurrs resin. There are three major sectioning methods for plastic embedded specimens:
(1) direct sectioning using heavy-duty microtomes; (2) sawing-grinding; and (3) sawing only.

27-29

Observation and characterization are normally conducted using a light microscope. Descriptive
histology and histomorphometry are the two main types of histological study. Depending on the
particular situation, either or both may be used. Descriptive histology is used to give a general
picture of the tissue of interest, including the morphology, structure, and arrangement of cells,
matrix, trabeculae, or marrow space. Scoring systems are often designed in order to semiquantify
the components of interest. Care must be observed when comparing control data from other sources,
particularly if other techniques are employed. Intermethod and interobserver variation on bone area
measurement, osteoid perimeter, and width have been shown to occur; some bone area measure-
ments differ due to inherent sampling variation.

30

Histomorphometric analysis has been performed using histological sections, microradiographs,
and backscattered electron microscopic (BSEM) images (on plastic-embedded surfaces).



Standard
SEM images of a specimen surface are less favorable for histomorphometric analysis due to the
fact that overlying components from adjacent areas are not well demonstrated.

27

Histomorphometry is a methodology for quantitatively analyzing (1) length (perimeter or
boundary), such as the surface perimeter of an implant; (2) distance between points, such as the
clearance at the implanttissue interface or distance between the central lines of two trabeculae;
(3) area, such as trabecular bone area or repair tissue area; and (4) the number of components of
interest, such as trabecular number.

31

These parameters are the four types of primary measurements
which can be made based on 2D images. Three-dimensional parameters or structures can be
calculated or reconstructed from 2D measurements according to carefully considered assumptions.
Although accurate 3D data are necessary for proper comparison between different specimens (such
as treated and control bone structure), it is often very difcult to reconstruct a 3D structure based
on a single 2D image because the structures of most biological tissues (such as bone tissue) are
anisotropic. This problem has been partially conquered by the introduction of quantitative CT

32,33

and magnetic resonance imaging (MRI),

33

which can easily section and reconstruct the specimen.
In spite of its limitations, 2D histomorphometric analysis remains a common and useful
method for analyzing the structural changes in trabecular bone,

33,34

the callus composition in
healing fracture sites, the repair tissues of bone or defects, and the bone apposition and ingrowth
into implant surfaces.
Standard nomenclature, symbols, and units for bone histomorphometry can be found in the
review by Partt et al.

31

The more commonly used terms for trabecular bone structures include
BV (bone volume) or TBA (trabecular bone area, which is the trabecular surface area divided
by the total area in mm

2

); Tb.Th (trabecular thickness, the average thickness of trabeculae in

m);
and Tb.Sp (trabecular separation, the average distance between trabeculae, representing the
amount of marrow space in

m). Common parameters for trabecular bone spatial connectivity
include Tb.N (trabecular number, the average number of continuous trabecular elements encoun-
tered per unit area); Ho.N (hole number, the average number of holes per unit area); N.Nd
(trabecular node number; nodes: trabecular branch points); N.Tm (trabecular terminus number;
termini: trabecular end points); and Nd/Tm ratio. Most of the parameters can be measured using
specialized imaging software.
In the histomorphometric analysis of implantbone interfaces, the useful parameters are
(1) bone apposition (or ongrowth), which is the fractional linear extent of bone apposed to implant
surface divided by the total surface perimeter of the implant (i.e., the surface potentially available
for apposition)

35,36

and (2) bone ingrowth, which represents the amount of ingrown bone per unit
of available surface area, porous space, and ingrowth depth.

35,37,38

In the case of bone ingrowth
within an osteopenic bone bed, the structure of the bone, represented by TBA, Tb.Th, Tb.N, and
Tb.Sp, should be also analyzed.

35,38

0266/ ch06/frame Page 106 Thursday, October 21, 1999 2:22 PM

Methods of Evaluation for Bone Dimensions, Densities, Contents, Morphology, and Structures

107

B. S

CANNING

E

LECTRON

M

ICROSCOPY

SEM and BSEM are important methods for evaluation of the structure and morphology of bone
structures

34,39

and boneimplant interfaces.

40,41

The shortcomings of SEM are that the specimen
must be dried before observation, causing distortion of the original spatial structure and morphol-
ogy,

42

and in some instruments specimen size is limited. The rst problem seems to have been
solved by the new low-temperature or cryo-SEM system.

43

BSEM provides better resolution than
microradiography with demonstrated consistency and is not affected by projection-effect errors.
BSEM also accurately images despite bone and mineral variations.

44

C. C

ONFOCAL

M

ICROSCOPY

Confocal laser scanning microscopy (CLSM) utilizes a laser beam that can penetrate tissue to a
depth of 300 to 500

m and thus reects images beneath the surface of a specimen. CLSM records
the intensity of light from a very narrow aperture, excluding light from out-of-focal planes. Stepwise
movement permits an artifact-free, topographic image to be obtained in a nondestructive approach
without the use of special staining.

45

Stored multilayer 2D images can then be reorganized to show
3D or cross-sectional pictures. CLSM has been used for viewing the structures at the implanttissue
interface, such as unmineralized bone matrix or mineralized bone.

46,47

Using CLSM, Piattelli et al.

46

found that a layer of unmineralized bone matrix lies between mineralized bone and the titanium
screw interface in a rabbit tibial model. Their study revealed that while 40% of the titanium surface
contained bone apposition, only 10% of the bone was in direct contact with the screw surface while
the other 30% was separated from the surface by an unmineralized tissue layer.

D. R

IDGE

N

UMBER

D

ENSITY

Quantitative CT (QCT) and MRI have been shown to be effective methods for measurement of the
microarchitecture of cancellous bone in addition to its density. Both, however, lack the spatial
resolution to image the individual trabeculae with real precision. Ridge number density (RND) is
a process whereby number of trabeculae can be determined from high-resolution QCT 3D images.
In the data process step, ridges that correspond to the center points of the trabeculae are extracted
from the 3D image. High spatial resolution and low doses of radiation lead to image noise which
is managed with a 3D algorithm based on directional derivatives of approximated t functions. The
advantages of RND measurement of bone strength are (1) reproducibility and (2) low dosages of
radiation allowing for longitudinal and cross-sectional studies.

48

IV. BONE APPARENT DENSITY, MATERIAL DENSITY,
MINERAL CONTENT, AND MINERAL DENSITY
A. D

EFINITIONS



AND

M

ETHODS



OF

E

VALUATION

The meanings of bone apparent density, material density, bone mineral content (BMC), and bone
mineral density (BMD) should be clearly dened (Table 6.1). Although it looks simple and straight-
forward, it is a difcult task to dene them, especially for BMC and BMD. Based on the methods
of evaluation and investigators preferences, the expressions of BMC and BMD can be mass/unit
length, area, or volume, percentage, image gray level, or a number.
For cortical bone, apparent density and material density are basically the same as there is no
marrow space in compact bone. Therefore, cortical bone density is commonly used to describe
the density of cortical bone. For cancellous bone there are different material characteristics arising
from the two-phase structure (trabeculae and marrow).

52

Based on their structural (apparent) density
and material density, respectively, two mechanical properties are generally considered, the structural
and material properties.

0266/ ch06/frame Page 107 Thursday, October 21, 1999 2:22 PM

108

Mechanical Testing of Bone and the BoneImplant Interface

Cortical bone material density is between 1.7 and 2.0 g/cm

3

and cancellous bone material
density is between 1.6 and 1.9 g/cm

3

. One of the simplest methods for measurement of cortical
bone density is based on Archimedes principle. Alternative approaches include (1) if the specimen
is a simple geometric shape, the volume can be calculated via direct measurement; (2) use of
preparations of methylene iodine and xylene based on different specic gravity which cause bone
specimens to exhibit neutral buoyancy in the solution closest to its own density.
The measurement of structural (apparent) density

(!

a

) is achieved by weighing the cancellous
structure without free water in its marrow cavities (wet weight, w

b

) and dividing the wet weight
by the structural volume (including both trabeculae and marrow cavities):

!

a

=

w

b

/("

d

2

h

/4) (6.1)
where

d

and

h

represent diameter and height of a cylindrical specimen. Other specimen shapes, such
as cubic, can be used, but they are technically more demanding and have more sharp corners than
cylinders, which may cause bone materials to fracture from the specimen during the processes of
defatting or marrow removing. Just as in cortical bone measurement an accurate method for cancellous
bone volume is through use of a gravity bottle, based on Archimedes principle (before marrow removal).
Many methods have been reported for removing bone marrow, including boiling in water with
detergent, high-pressure water jet, or chemical solvent. Depending on the size and shape of the
specimen, an individualized combination of the above-mentioned methods is appropriate. In the
authors laboratory the following procedure has been used for small specimens (e.g., 4-mm-diam.,
5-mm-length cylinder): (1) defatting in 50/50 acetone/ethanol mixture with agitation for 24 h;
(2) removing marrow in low concentration bleach (1.0 to 1.5% sodium hypochlorite) with agitation
for 12 h; and (3) removing marrow residues with a high-pressure water jet (using a syringe).
For

in vitro

measuring of bone mineral density, the traditional method is burning bone
specimens in air in a 500

C furnace and weighing the ash.

53

Instead of ash weight or ash
fraction, the authors prefer to use ash density, which is dened as ash weight per unit bone
volume (including trabeculae and marrow space for cancellous bone). It is suggested that the
crucibles be dried at 500

C overnight, weighed, loaded with the bone specimen, and heated at

TABLE 6.1
Denitions of BMC and BMD (only selected references are included)

Method of
Evaluation Unit Denition

Apparent density Weight g/cm

3

Wet weight per unit structural volume including bone (such as
trabeculae) and marrow space, not including marrow
Material density Weight g/cm

3

Wet weight per unit material volume
BMD Weight g/cm

3

Bone mineral mass/unit bone volume, or named ash density if
an ashing (or burning) method is used.
2D imaging Gray scale The intensity of the image portion due to the mineral content to
a dened sample thickness; example: X-ray images
RA Gray scale Percentage difference to a normal or standard number

49

SPA g/cm

2

Bone mass/(unit of length

width)

50

DEXA g/cm

2

Bone mass/measured area

50

QCT g/cm

3

Bone mass/measured volume

50

BMC Weight % The ratio of unit weight of the mineral portion to dry bone unit
weight and is frequently reported as a percentage

51

SPA g/cm

2

Bone mass

50

DXA g/cm

2

Bone mass/unit length

50

QCT g/cm

3

Bone mass/measured length

50

0266/ ch06/frame Page 108 Thursday, October 21, 1999 2:22 PM

Methods of Evaluation for Bone Dimensions, Densities, Contents, Morphology, and Structures

109

500

C for 18 h to remove the organic phase. Then, the crucible containing the ash is weighed
to determine the weight of ash.
Image-based BMDs using radiographic absorptiometry (RA), single-photon absorptiometry
(SPA), dual-energy X-ray absorptiometry (DEXA), or QCT are used as predictors of the breaking
strength of bone.

49

Caution must be exercised, however, since inaccuracies in BMD, which is
proportional to the square of the apparent density, can cause large errors in predicted bone strength.
When several methods for determination of bone strength, including bone mineral content, areal
BMD, volumetric BMD, and bone apparent density were assessed

in vitro, in spite of measurement
errors, bone mass, areal, and volumetric bone density were found to be equally accurate, sensitive
and specic surrogates of the breaking strength of bone strength in vitro.
11
Less frequently used methods for determining bone mineral content include the use of
decalcifying solution or measuring the radiographic density of whole bone or bone sections. The
latter is more suitable for in vivo conditions. Other indirect methods for bone mineral content
include radiographic and spectrographic methods.
51
One method for the assessment of bone
density utilizes cutting resistance measures at low speeds for the identication of various bone
densities. By using porcine rib specimens, the outcome of cutting resistance measures were
compared with that of radiographic technique. The two procedures demonstrated agreement in
the ability to identify bone density.
54
B. BONE DENSITY VALUES MEASURED BY IN VITRO METHODS
The material density of cortical bone is the wet weight divided by the specimen volume. Cortical
bone has an average density of approximately 1.9 g/cm
3
.
51,55
The common ways to measure the
volume of a cortical bone specimen include the use of a gravity bottle based on Archimedes
principle, and directly measuring the dimensions of the specimen. The latter requires that the
specimen have a regular shape, such as cylindrical.
Selected reports on apparent densities of human and animal cancellous bones are listed in
Table 6.2. The apparent density of cancellous bone ranges from 0.14 to 1.10 g/cm
3
(average: 0.62
g/cm
3
, n = 16). The compressive strength (# in MPa) of cancellous bone is related to its apparent
density (! in g/cm
3
) by a power law of the form:
# = 60!
2
(6.2)
Similarly, the compressive modulus (E in MPa) of cancellous bone is related to the apparent density
(! in g/cm
3
) by
E = 2915!
2
(6.3)
Material density of cancellous bone is measured using the weight of bone material (only trabeculae)
divided by the volume of only trabeculae. The density is a little lower than that of cortical bone, being
1.6 to 1.9 g/cm
3
.
51
The principle is again that the marrow needs to be cleaned thoroughly before the
measurements of weight and volume. Using a gravity bottle based on Archimedes principle is the
common way to measure both the weight and volume of the bone specimen. To make the measurement,
the marrow is removed so no air bubbles or water will be trapped inside the marrow cavities.
Selected data of ash densities of human and animal cancellous bones are also listed in Table 6.2,
ranging from 0.19 to 0.56 g/cm
3
with an average of 0.370.10 g/cm
3
(n = 12), which is about 60%
of the value of apparent density as shown in the following equation:
!
Ash
$ 0.6 !
Apparent
(6.4)
The latter is calculated from the 11 data sets containing both values of apparent density and ash density.
0266/ ch06/frame Page 109 Thursday, October 21, 1999 2:22 PM
110 Mechanical Testing of Bone and the BoneImplant Interface
V. SPECIAL TECHNIQUES
A. COMPUTED TOMOGRAPHY
CT has been used for examining bone structure and geometry, bone destruction, new bone formation
during fracture healing, bone lengthening in animal models, and dimensions of human bone with
regard to the femoral component cortical bone ingrowth.
16,49,64-67
CT is a nondestructive method for
determining the cross-sectional contours of various anatomical structures, including bone. Based
on a 2D array, Feldkamp et al.
68
developed what has become known as the -CT scanner for 3D
reconstruction of bone. -CT, initially developed for detection of ceramic material defects, operates
similarly to commercial CT scanners except with -CT the specimen is rotated rather than rotating
the source. The specimen is limited in size, and a 2D detector instead of a linear array is used to
create a 3D image. The ndings are that -CT images are not signicantly different from sections
measured histologically.
18
Sumner and colleagues
69
have reported that separate CT thresholds need to be used to distinguish
endosteal and periosteal surfaces. The absence of such can result in errors up to 30% for cortical
area estimates.

QCT is capable of analyzing bone structure, even in small rat bones, and is believed
to be more sensitive than DEXA.
70
The spatial resolution of CT on cancellous specimens can reach
8 to 80 m.
71,72
The recent development of QCT has resulted in images with high 3D resolution,
which may be used for 3D reconstruction of cancellous bone.
33,73
Another CT method for 3D
reconstruction is the X-ray tomographic microscope (XTM), which allows in vivo evaluation of
cancellous bone.
32
QCT has also been used for evaluating the density and mechanical properties of bone. It can
be applied in vivo or on excised bone specimens.
74
CT numbers, image intensity, or CT density
values (such as Hounseld units: HU) are measured in the areas of interest. CT density is based
on relative attenuation of X rays by a scanned body as compared with attenuation by water. In
general, zero HU equals the density of water and 1000 HU corresponds to the relative density of
TABLE 6.2
Apparent and Ash Densities of Cancellous Bones (selected data from
the literature)
Species Bone Apparent density (g/cm
3
) Ash density (g/cm
3
) Ref.
Human Distal femur 0.43 0.15 0.26 0.08 57
0.46 58
Vertebral body 0.14 0.06 59
Cattle Vertebral body 0.45 0.09 0.19 0.06 60
Dog Distal femur 0.44 0.16 0.26 0.08 57
0.690.98 0.400.56
a
53
Proximal tibia 0.410.83
a
0.220.44
a
53
Humeral 0.84 0.17 0.43 0.06 53
Goat Femoral head 0.91 0.04 0.48 0.03 61
Distal femur 0.540.66
a
0.320.40
a
61
Proximal tibia 0.931.1
a
0.500.56
a
61
Humeral 0.75 0.03 0.36 0.01 61
Sheep Vertebral body 0.60 0.16 0.37 0.11 62
Pig Vertebral body 0.46 0.04 63
Mean SEM 0.62 0.37
a
Range of average values from different locations.
0266/ ch06/frame Page 110 Thursday, October 21, 1999 2:22 PM
Methods of Evaluation for Bone Dimensions, Densities, Contents, Morphology, and Structures 111
air. Cortical bone has CT density greater than +1000 HU and cancellous bone has values ranging
from 25 to 714 HU. An average CT value of water is determined for each scan to adjust the
systematic error of the machine.
74
By correlation analysis, power functions between CT density
and mechanical values (such as strength or elastic modulus), apparent density and ash density of
bone can be formulated. Therefore, mechanical values and densities of bone can be predicted by
CT values.
64,74,75
The advantage of QCT is that it can be applied noninvasively and in vivo.
B. MAGNETIC RESONANCE IMAGING
Recent studies have shown that MRI may also provide high-resolution 3D images of trabecular
architecture.
33,49,71,76,77
The technique can add to the quantication of trabecular architecture,
anisotropy, and connectivity, factors which provide important contributions to the biomechanical
properties of trabecular bone.
78
MRI is believed to be superior to CT and ultrasound methods
for this purpose due to its ability to distinguish the boundary between muscle and bone and even
between the cortical and cancellous regions within the bone.
71
MRI has also been used for
evaluating BMD
79-81
and predicting bone elastic modulus (needs further study),
82
which is very
signicant for the diagnosis and monitoring of osteoporosis. MR-derived measures can replicate
trends that have been previously established and may have future uses to resolve prior issues in
in vitro and in vivo studies.
78
C. SINGLE-PHOTON, SINGLE X-RAY, AND DUAL-ENERGY ABSORPTIOMETRY
SPA and dual energy absorptiometry (DEA) are two noninvasive methods for measuring BMC,
BMD, and cross-sectional geometry.
49,67,83
They are most commonly applied to the appendicular
skeleton.
64
DEXA measures the mineral mass rather than bone mass. Therefore, absorptiometry
measures the physical denition and material density of bones differently, which may account for
observed differences.
84
The radioactive sources used for SPA are
125
I and
241
Am. SPA has commonly been used for
measuring the mineral density of the distal radius, ulna, calcaneus and femoral neck. By using
formulas generated by regression analysis, SPA can also be used to estimate the mechanical
properties of healing bone.
85-87
A new method for measurement of bone mass, reported by Borg
et al.,
88
is single X-ray absorptiometry (SXA).

The SXA device has an X-ray tube which emits
X rays at an energy level of 40 kVp and 0.2 mA. It has been used to measure the BMC and
BMD of forearm bones, and the results have shown a positive correlation with the more traditional
SPA method.
88
DEA can be performed with either radioisotopes or X rays. When the dual-energy source
is derived from X-rays, the technique is termed DEXA. A high correlation has been found
between DEXA and traditional methods for measuring bone density.
88
DEXA has been dem-
onstrated to measure accurately the BMC and BMD of very small areas of interest,
89
such as
in rat bone.
90-92
Like SPA, DEXA has been commonly used to evaluate BMC/BMD (even for
small bones), and mechanical properties of normal bone,
84
healing bone,
64,85,93
and osteoporotic
bone
92,94,95
in animal models.
DEXA is also an accurate and precise method to measure the dimensions of human long bones.
Sievanen et al.
83
found that the standard DEXA technique provides a reliable measurement of the
width and length in human humerus and femur in vivo, and thus may be useful in evaluating the
properties of these bones in conjunction with the standard bone mineral measurements.
D. ULTRASOUND
Quantitative ultrasound (QUS) parameters, such as broadband ultrasound attenuation, ultrasound
velocity, and ultrasound attenuation, have been demonstrated to measure densitometric and geo-
metric properties of human bones effectively.
49,88,96,97
In an in vitro study on trabecular bone cubes,
0266/ ch06/frame Page 111 Thursday, October 21, 1999 2:22 PM
112 Mechanical Testing of Bone and the BoneImplant Interface
T
A
B
L
E

6
.
3
M
e
t
h
o
d

S
e
l
e
c
t
i
o
n

f
o
r

M
e
a
s
u
r
i
n
g

B
o
n
e

D
i
m
e
n
s
i
o
n
s
,

S
t
r
u
c
t
u
r
e
,

B
M
C
,

B
M
D
,

a
n
d

M
e
c
h
a
n
i
c
a
l

P
r
o
p
e
r
t
i
e
s

(
o
n
l
y

s
e
l
e
c
t
e
d

n
e
w

o
r

r
a
r
e

r
e
f
e
r
e
n
c
e
s

a
r
e

c
i
t
e
d
)
L
e
v
e
l
E
l
e
m
e
n
t
s
M
e
t
h
o
d
s

f
o
r

S
p
e
c
i
m
e
n

D
i
m
e
n
s
i
o
n
s
M
e
t
h
o
d
s

f
o
r

B
o
n
e

S
t
r
u
c
t
u
r
e
s
M
e
t
h
o
d
s

f
o
r

M
a
t
e
r
i
a
l

D
e
n
s
i
t
i
e
s

o
r

B
M
D
M
e
t
h
o
d
s

f
o
r

B
M
C
M
e
t
h
o
d
s

f
o
r

B
o
n
e

M
e
c
h
a
n
i
c
a
l

P
r
o
p
e
r
t
i
e
s
M
a
c
r
o
s
t
r
u
c
t
u
r
e

(
w
h
o
l
e

b
o
n
e
)
F
e
m
u
r
,

h
u
m
e
r
u
s
,

v
e
r
t
e
b
r
a
e
,

f
r
o
n
t
a
l

b
o
n
e
,

p
h
a
l
a
n
g
e
a
l

b
o
n
e
s
,

c
a
l
c
a
n
e
u
s
R
u
l
e
r
,

c
a
l
i
p
e
r
,

X
-
r
a
y
,

C
T
,

D
E
X
A
,

M
R
I
M
a
c
r
o
p
h
o
t
o
g
r
a
p
h
y
,

X
-
r
a
y
,

C
T
,

D
E
X
A
,

M
R
I
R
A
,

S
P
A
,

D
E
X
A
,

Q
C
T
,

M
R
I
,

u
l
t
r
a
s
o
u
n
d
S
P
A
,

D
E
X
A
R
e
g
u
l
a
r

m
a
t
e
r
i
a
l

t
e
s
t
i
n
g

s
y
s
t
e
m
s
,

u
l
t
r
a
s
o
u
n
d
A
r
c
h
i
t
e
c
t
u
r
e

(
t
i
s
s
u
e

l
e
v
e
l
)
C
o
m
p
a
c
t

b
o
n
e

o
r

c
a
n
c
e
l
l
o
u
s

b
o
n
e

b
l
o
c
k
s
,

c
y
l
i
n
d
e
r
s
,

c
u
b
e
s
,

o
r

b
e
a
m
s
R
u
l
e
r
,

c
a
l
i
p
e
r
,

X
-
r
a
y
,

C
T
,

M
R
I
M
i
c
r
o
p
h
o
t
o
g
r
a
p
h
y
,

h
i
s
t
o
l
o
g
y
,

h
i
s
t
o
m
o
r
p
h
o
m
e
t
r
y
,

S
E
M
,

C
L
S
M
,
1
0
6
-
1
0
9

X
-
r
a
y
,

C
T
,

M
R
I
,

A
F
M
#
1
1
0
A
r
c
h
i
m
e
d
e
s

m
e
t
h
o
d
,

a
s
h
i
n
g
,

u
l
t
r
a
s
o
u
n
d
,

-
C
T
A
s
h
i
n
g
,

S
P
A
,

D
E
X
A
R
e
g
u
l
a
r

m
a
t
e
r
i
a
l

t
e
s
t
i
n
g

s
y
s
t
e
m
s
,

u
l
t
r
a
s
o
u
n
d
,

m
a
c
r
o
i
n
d
e
n
t
a
t
i
o
n
,

l
a
s
e
r

s
p
e
c
k
l
e

s
t
r
a
i
n

m
e
a
s
u
r
e
m
e
n
t
1
1
1
M
i
c
r
o
s
t
r
u
c
t
u
r
e

(
o
s
t
e
o
n
a
l

o
r

t
r
a
b
e
c
u
l
a
r

l
e
v
e
l
)
O
s
t
e
o
n
s
,

t
r
a
b
e
c
u
l
a
e
,

o
r

m
i
c
r
o
b
e
a
m
s

o
r

c
y
l
i
n
d
e
r
s
M
i
c
r
o
s
c
o
p
y
,

S
E
M
,

C
L
S
M
M
i
c
r
o
p
h
o
t
o
g
r
a
p
h
y
,

h
i
s
t
o
l
o
g
y
,

h
i
s
t
o
m
o
r
p
h
o
m
e
t
r
y
,

S
E
M
,

T
E
M
,

C
L
S
M
1
1
2
A
r
c
h
i
m
e
d
e
s

m
e
t
h
o
d
,

a
s
h
i
n
g
,

m
i
c
r
o
r
a
d
i
o
g
r
a
p
h
y
1
1
3
D
e
c
a
l
c
i

c
a
t
i
o
n

+

a
t
o
m
i
c

a
b
s
o
r
p
t
i
o
n

s
p
e
c
t
r
o
p
h
o
t
o
m
e
t
r
y
1
1
4
M
i
c
r
o
m
e
c
h
a
n
i
c
a
l

t
e
s
t
s

u
s
i
n
g

i
n
v
e
s
t
i
g
a
t
o
r

d
e
s
i
g
n
e
d

m
i
c
r
o
t
e
s
t
e
r
s

o
r

s
m
a
l
l
-
s
c
a
l
e

m
a
t
e
r
i
a
l

t
e
s
t
i
n
g

s
y
s
t
e
m
s
,

m
i
c
r
o
-

o
r

n
a
n
o
i
n
d
e
n
t
a
t
i
o
n
,

F
E
A
,
1
1
5

a
c
o
u
s
t
i
c

m
i
c
r
o
s
c
o
p
y
1
1
6
,
1
1
7
S
u
b
m
i
c
r
o
s
t
r
u
c
t
u
r
e

(
l
a
m
e
l
l
a
r

l
e
v
e
l
)
L
a
m
e
l
l
a
,

l
a
r
g
e

c
o
l
l
a
g
e
n

b
e
r
s
M
i
c
r
o
s
c
o
p
y
,

S
E
M
,

T
E
M
,

C
L
S
M
M
i
c
r
o
p
h
o
t
o
g
r
a
p
h
y
,

h
i
s
t
o
l
o
g
y
,

h
i
s
t
o
m
o
r
p
h
o
m
e
t
r
y
,

S
E
M
,

T
E
M
,

X
-
r
a
y

d
i
f
f
r
a
c
t
i
o
n
,
1
1
8

C
L
S
M
,

A
F
M
N
A
N
A
N
a
n
o
i
n
d
e
n
t
a
t
i
o
n
,
1
1
9

m
i
c
r
o
w
a
v
e

e
x
t
e
n
s
o
m
e
t
e
r
,
1
2
0

a
c
o
u
s
t
i
c

m
i
c
r
o
s
c
o
p
y
1
1
6
,
1
1
7
U
l
t
r
a
s
t
r
u
c
t
u
r
e

(
n
a
n
o
s
t
r
u
c
t
u
r
e
)
C
o
l
l
a
g
e
n

b
r
i
l

a
n
d

m
o
l
e
c
u
l
e
,

m
i
n
e
r
a
l

c
o
m
p
o
n
e
n
t
s
S
E
M
,

T
E
M
,

A
F
M
A
F
M
,
1
2
1

S
E
M
,

T
E
M
,

X
-
r
a
y

d
i
f
f
r
a
c
t
i
o
n
1
1
8
,
1
2
2
N
A
N
A
A
F
M
,
1
2
3

a
c
o
u
s
t
i
c

m
i
c
r
o
s
c
o
p
y
1
2
4
A
F
M

=

a
t
o
m
i
c

f
o
r
c
e

m
i
c
r
o
s
c
o
p
y
;

F
E
A

=

n
i
t
e
-
e
l
e
m
e
n
t

a
n
a
l
y
s
i
s
;

N
A

=

u
n
k
n
o
w
n

t
o

t
h
e

a
u
t
h
o
r
s
.
0266/ ch06/frame Page 112 Thursday, October 21, 1999 2:22 PM
Methods of Evaluation for Bone Dimensions, Densities, Contents, Morphology, and Structures 113
ultrasound parameters were shown to be signicantly associated with bone structural indexes, such
as Tb.Sp or trabecular connectivity.
88
However, when ultrasound was used as an independent
predictor of femoral strength when combined with femoral or calcaneal BMD, the ndings did not
improve the prediction of femoral strength. Contributing factors to this weak relationship are likely
based on differences in the material properties of the calcaneus and the femur.
98
QUS is becoming
an alternative to photon absorptiometry in assessing bone density. This has been useful in the
diagnosis and management of osteoporosis.
99-101
The diagnostic sensitivity of QUS on BMD is
similar to that of DEXA, even on small rat bones.
99
Ultrasound is also a very important tool for measuring mechanical properties of bone. Ultrasonic
techniques offer some advantages over direct mechanical tests for measuring the elastic modulus
of bone.
51
Specically, the specimens can be smaller, with less-complicated shapes (cylinder or
cube), and several anisotropic properties can be tested using one specimen.
102
Recently, with the
combination of vibration analysis and ultrasound velocity measurements, whole bone mechanical
characteristics have been be assessed in vivo.
103
VI. SUMMARY
Bone has a sophisticated hierarchical structure ranging from macro- to nanoscales,
104
or from whole
bone to ultrastructural level (see Chapter 8). Investigations of mechanical properties of bone at all
levels are essential for complete understanding of the mechanical properties of bone.
104,105
One
should choose the appropriate methods for measuring bone dimensions and evaluating bone structure
based on the scale of the specimen (Table 6.3). Attention has been paid to several new technologies,
including nanoindentation, acoustic microscopy, CLSM, and atomic force microscopy.
REFERENCES
1. Pensler, J. and McCarthy, J.G., The calvarial donor site: an anatomic study in cadavers, Plast. Reconstr.
Surg., 75, 648, 1985.
2. Kim, Y.J. and Kim, C.H., A survey on thickness of the Korean calvarium, J. Korean Soc. Plast.
Reconstr. Surg., 13, 147, 1986.
3. Horsman, A. and Leach, A.E., The estimation of the cross-sectional area of ulna and radius, Am. J.
Phys. Anthropol., 40, 173, 1963.
4. An, Y.H., Kang, Q., and Friedman, R.J., Mechanical symmetry of rabbit bones studied by bending
and indentation testing, Am. J. Vet. Res., 57, 1786, 1996.
5. van der Meulen, M.C., Ashford, M.W., Jr., Kiratli, B.J., et al., Determinants of femoral geometry and
structure during adolescent growth, J. Orthop. Res., 14, 22, 1996.
6. Chumlea, W.C., Mukherjee, D., and Roche, A.F., A comparison of methods for measuring cortical
bone thickness, Am. J. Phys. Anthropol., 65, 83, 1984.
7. Ross, W.D., Rempel, R.D., Quibell, R.W., et al., Technical note: an electronic caliper designed for
measuring bone breadths in living subjects, Am. J. Phys. Anthropol., 90, 373, 1993.
8. Rico, H. and Hernandez, E.R., Bone radiogrametry: caliper vs. magnifying glass, Calcif. Tissue Int.,
45, 285, 1989.
9. Gualdi-Russo, E. and Russo, P., A new technique for measurements on long bones: development of
a new instrument and techniques comparison, Anthropol. Anz., 53, 153, 1995.
10. Ruff, C.B. and Hayes, W.C., Cross-sectional geometry of Pecos Pueblo femora and tibiae a
biomechanical investigation: I. Method and general patterns of variation, Am. J. Phys. Anthropol., 60,
359, 1983.
11. Tabensky, A.D., Williams, J., DeLuca, V., et al., Bone mass, areal, and volumetric bone density are
equally accurate, sensitive, and specic surrogates of the breaking strength of the vertebral body: an
in vitro study, J. Bone Miner. Res., 11, 1981, 1996.
12. Horsman, A. and Simpson, M., The measurement of sequential changes in cortical bone geometry,
Br. J. Radiol., 48, 471, 1975.
0266/ ch06/frame Page 113 Thursday, October 21, 1999 2:22 PM
114 Mechanical Testing of Bone and the BoneImplant Interface
13. McPherson, E.J., Friedman, R.J., An, Y.H., et al., Anthropometric study of normal glenohumeral
relationships, J. Shoulder Elbow Surg., 6, 105, 1997.
14. Fox, K.M., Kimura, S., Powell-Threets, K., and Plato, C.C., Radial and ulnar cortical thickness of
the second metacarpal, J. Bone Miner. Res., 10, 1930, 1995.
15. Hanson, P.D. and Markel, M.D., Radiographic geometric variation of equine long bones, Am. J. Vet.
Res., 55, 1220, 1994.
16. Pfeifer, T., Mahlo, R., Franzreb, M., et al., Computed tomography in the determination of leg geometry,
In Vivo, 9, 257, 1995.
17. Rumph, P.F. and Hathcock, J.T., A symmetric axis-based method for measuring the projected femoral
angle of inclination in dogs, Vet. Surg., 19, 328, 1990.
18. Kuhn, J.L., Goldstein, S.A., Feldkamp, L.A., et al., Evaluation of a microcomputed tomography system
to study trabecular bone structure, J. Orthop. Res., 8, 833, 1990.
19. Ashman, R.B. and Rho, J.Y., Elastic modulus of trabecular bone material, J. Biomech., 21, 177, 1988.
20. Pugh, J.W., Rose, R.M., and Radin, E.L., A structural model for the mechanical behavior of trabecular
bone, J. Biomech., 6, 657, 1973.
21. Townsend, P.R., Raux, P., Rose, R.M., et al., The distribution and anisotropy of the stiffness of
cancellous bone in the human patella, J. Biomech., 8, 363, 1975.
22. Kuhn, J.L., Goldstein, S.A., Choi, K., et al., Comparison of the trabecular and cortical tissue moduli
from human iliac crests, J. Orthop. Res., 7, 876, 1989.
23. Hermanussen, M., Bugiel, S., Aronson, S., and Moell, C., A noninvasive technique for the accurate
measurement of leg length in animals, Growth Dev. Aging, 56, 129, 1992.
24. Yang, R.S., A new goniometer, Orthop. Rev., 21, 877, 1992.
25. Hadlock, F.P., Kent, W.R., Loyd, J.L., et al., An evaluation of two methods for measuring fetal head
and body circumferences, J. Ultrasound Med., 1, 359, 1982.
26. Mattfeldt, T., Mall, G., Gharehbaghi, H., and Moller, P., Estimation of surface area and length with
the orientator, J. Microsc., 159, 301, 1990.
27. An, Y.H., Methods of evaluation in orthopaedic animal research, in Animal Models in Orthopaedic
Research, An, Y.H. and Friedman, R.J., Eds., CRC Press, Boca Raton, FL, 1999, 85.
28. Gruber, H.E. and Stasky, A.A., Histological study in orthopaedic animal research, in Animal Models
in Orthopaedic Research, An, Y.H. and Friedman, R.J., Eds., CRC Press, Boca Raton, FL, 1999, 115.
29. Jansen, J.A., Animal models for studying soft tissue biocompatibility of biomaterials, in Animal Models
in Orthopaedic Research, An, Y.H. and Friedman, R.J., Eds., CRC Press, Boca Raton, FL, 1999, 393.
30. Wright, C.D., Vedi, S., Garrahan, N.J., et al., Combined inter-observer and inter-method variation in
bone histomorphometry, Bone, 13, 205, 1992.
31. Partt, A.M., Drezner, M.K., Glorieux, F.H., et al., Bone histomorphometry: standardization of nomen-
clature, symbols, and units. Report of the ASBMR Histomorphometry Nomenclature Committee,
J. Bone Miner. Res., 2, 595, 1987.
32. Kinney, J.H., Lane, N.E., and Haupt, D.L., In vivo, three-dimensional microscopy of trabecular bone,
J. Bone Miner. Res., 10, 264, 1995.
33. Odgaard, A., Three-dimensional methods for quantication of cancellous bone architecture, Bone, 20,
315, 1997.
34. Kang, Q., An, Y.H., Butehorn, H.F., and Friedman, R.J., Morphological and mechanical study of the
effects of experimentally induced inammatory knee arthritis on rabbit long bones, J. Mater. Sci.
Mater. Med., 9, 463, 1998.
35. An, Y.H., Friedman, R.J., Jiang, M., et al., Bone ingrowth to implant surfaces in an inammatory
arthritis model, J. Orthop. Res., 16, 576, 1998.
36. Friedman, R.J., An, Y.H., Ming, J., et al., Inuence of biomaterial surface texture on bone ingrowth
in the rabbit femur, J. Orthop. Res., 14, 455, 1996.
37. Vigorita, V.J., Minkowitz, B., Dichiara, J.F., and Higham, P.A., A histomorphometric and histologic
analysis of the implant interface in ve successful, autopsy-retrieved, noncemented porous-coated
knee arthroplasties, Clin. Orthop., 293, 211, 1993.
38. Moroni, A., Caja, V.L., Egger, E.L., et al., Histomorphometry of hydroxyapatite coated and uncoated
porous titanium bone implants, Biomaterials, 15, 926, 1994.
39. Whitehouse, W.J., Dyson, E.D., and Jackson, C.K., The scanning electron microscope in studies of
trabecular bone from a human vertebral body, J. Anat., 108, 481, 1971.
0266/ ch06/frame Page 114 Thursday, October 21, 1999 2:22 PM
Methods of Evaluation for Bone Dimensions, Densities, Contents, Morphology, and Structures 115
40. McNamara, A. and Williams, D.F., Scanning electron microscopy of the metaltissue interface.
II. Observations with lead, copper, nickel, aluminium, and cobalt, Biomaterials, 3, 165, 1982.
41. Orr, R.D., de Bruijn, J.D., and Davies, J.E., Scanning electron microscopy of the bone interface with
titanium, titanium alloy and hydroxyapatite, Cells Mater., 2, 241, 1992.
42. Kobayashi, S., Yonekubo, S., and Kurogouchi, Y., Cryoscanning electron microscopic study of the
surface amorphous layer of articular cartilage, J. Anat., 187, 429, 1995.
43. Kobayashi, S., Yonekubo, S., and Kurogouchi, Y., Cryoscanning electron microscopy of loaded
articular cartilage with special reference to the surface amorphous layer, J. Anat., 188, 311, 1996.
44. Bloebaum, R.D., Skedros, J.G., Vajda, E.G., et al., Determining mineral content variations in bone
using backscattered electron imaging, Bone, 20, 485, 1997.
45. Grotz, K.A., Piepkorn, B., Bittinger, F., et al. [Confocal laser scanning microscopy (CLSM) for validation
of nondestructive histotomography of healthy bone tissue], Mund. Kiefer Gesichtschir., 2, 141, 1998.
46. Piattelli, A., Trisi, P., Passi, P., et al., Histochemical and confocal laser scanning microscopy study of
the bonetitanium interface: an experimental study in rabbits, Biomaterials, 15, 194, 1994.
47. Takeshita, F., Iyama, S., Ayukawa, Y., et al., Study of bone formation around dense hydroxyapatite
implants using light microscopy, image processing and confocal laser scanning microscopy,
Biomaterials, 18, 317, 1997.
48. Laib, A., Hildebrand, T., Hauselmann, H.J., and Ruegsegger, P., Ridge number density: a new param-
eter for in vivo bone structure analysis, Bone, 21, 541, 1997.
49. Wong, D.M. and Sartoris, D.J., Noninvasive methods for assesment of bone density, architecture, and
biomechanical properties: fundamental concepts, in Osteoporosis: Diagnosis and Treatment, Sartoris,
D.J., Ed., Marcel Dekker, New York, 1996, 201.
50. Van der Perre, G. and Lowet, G., Physical meaning of bone mineral content parameters and their
relation to mechanical properties, Clin. Rheumatol., 3 (Suppl. 1), 33, 1994.
51. Ashman, R.B., Experimental techniques, in Bone Mechanics, Cowin, S.C., Ed., CRC Press, Boca
Raton, FL, 1989, 91.
52. Carter, D.R. and Hayes, W.C., The compressive behavior of bone as a two-phase porous structure,
J. Bone Joint Surg., 59A, 954, 1977.
53. Kang, Q., An, Y.H., and Friedman, R.J., The mechanical properties and bone densities of canine
cancellous bones, J. Mater. Sci. Mater. Med., 9, 263, 1998.
54. Friberg, B., Sennerby, L., Roos, J., et al., Evaluation of bone density using cutting resistance meas-
urements and microradiography: an in vitro study in pig ribs, Clin. Oral Implants Res., 6, 164, 1995.
55. Spatz, H.C., OLeary, E.J., and Vincent, J.F., Youngs moduli and shear moduli in cortical bone, Proc.
R. Soc. London B Biol. Sci., 263, 287, 1996.
56. Hayes, W.C. and Bouxsein, M.L., Biomechanics of cortical and trabecular bone: implications for
assessment of fracture risk, in Basic Orthopaedic Biomechanics, Mow, V.C. and Hayes, W.C., Eds.,
Lippincott-Raven, Philadelphia, PA, 1997, Chap. 3.
57. Kuhn, J.L., Goldstein, S.A., Ciarelli, M.J., and Matthews, L.S., The limitations of canine trabecular
bone as a model for human: a biomechanical study, J. Biomech., 22, 95, 1989.
58. Odgaard, A., Hvid, I., and Linde, F., Compressive axial strain distributions in cancellous bone
specimens, J. Biomech., 22, 829, 1989.
59. Keaveny, T.M., Pinilla, T.P., Crawford, R.P., et al., Systematic and random errors in compression
testing of trabecular bone, J. Orthop. Res., 15, 101, 1997.
60. Swartz, D.E., Wittenberg, R.H., Shea, M., et al., Physical and mechanical properties of calf lumbo-
sacral trabecular bone, J. Biomech., 24, 1059, 1991.
61. An, Y.H., Kang, Q., and Friedman, R.J., The mechanical properties and bone densities of goat
cancellous bones, unpublished data, 1998.
62. Mitton, D., Rumelhart, C., Hans, D., and Meunier, P.J., The effects of density and test conditions on
measured compression and shear strength of cancellous bone from the lumbar vertebrae of ewes, Med.
Eng. Phys., 19, 464, 1997.
63. Mosekilde, L., Kragstrup, J., and Richards, A., Compressive strength, ash weight, and volume of
vertebral trabecular bone in experimental uorosis in pigs, Calcif. Tissue Int., 40, 318, 1987.
64. Markel, M.D., Wikenheiser, M.A., Morin, R.L., et al., The determination of bone fracture properties
by dual-energy X-ray absorptiometry and single-photon absorptiometry: a comparative study, Calcif.
Tissue Int., 48, 392, 1991.
0266/ ch06/frame Page 115 Thursday, October 21, 1999 2:22 PM
116 Mechanical Testing of Bone and the BoneImplant Interface
65. Schumacher, B., Albrechtsen, J., Keller, J., et al., Periosteal insulin-like growth factor I and bone
formation. Changes during tibial lengthening in rabbits, Acta Orthop. Scand., 67, 237, 1996.
66. Feng, Z., Ziv, I., and Rho, J., The accuracy of computed tomography-based linear measurements of
human femora and titanium stem, Invest. Radiol., 31, 333, 1996.
67. Bouxsein, M.L., Myburgh, K.H., van der Meulen, M.C., et al., Age-related differences in cross-
sectional geometry of the forearm bones in healthy women, Calcif. Tissue Int., 54, 113, 1994.
68. Feldkamp, L.A., Goldstein, S.A., Partt, A.M., et al., The direct examination of three-dimensional
bone architecture in vitro by computed tomography, J. Bone Miner. Res., 4, 3, 1989.
69. Sumner, D.R., Olson, C.L., Freeman, P.M., et al., Computed tomographic measurement of cortical
bone geometry, J. Biomech., 22, 649, 1989.
70. Gasser, J.A., Assessing bone quantity by pQCT, Bone, 17, 145S, 1995.
71. Mehta, B.V., Rajani, S., and Sinha, G., Comparison of image processing techniques (magnetic reso-
nance imaging, computed tomography scan and ultrasound) for 3D modeling and analysis of the
human bones, J. Digit. Imaging, 10, 203, 1997.
72. Gluer, C.C., Wu, C.Y., Jergas, M., et al., Three quantitative ultrasound parameters reect bone struc-
ture, Calcif. Tissue Int., 55, 46, 1994.
73. Muller, R., Hildebrand, T., Hauselmann, H.J., and Ruegsegger, P., in vivo reproducibility of three-
dimensional structural properties of noninvasive bone biopsies using 3D-pQCT, J. Bone Miner. Res.,
11, 1745, 1996.
74. Ciarelli, M.J., Goldstein, S.A., Kuhn, J.L., et al., Evaluation of orthogonal mechanical properties and
density of human trabecular bone from the major metaphyseal regions with materials testing and
computed tomography, J. Orthop. Res., 9, 674, 1991.
75. Ferretti, J.L., Capozza, R.F., and Zanchetta, J.R., Mechanical validation of a tomographic (pQCT)
index for noninvasive estimation of rat femur bending strength, Bone, 18, 97, 1996.
76. Jara, H., Wehrli, F.W., Chung, H., and Ford, J.C., High-resolution variable ip angle 3D MR imaging
of trabecular microstructure in vivo, Magn. Reson. Med., 29, 528, 1993.
77. Chung, H.W., Wehrli, F.W., Williams, J.L., and Wehrli, S.L., Three-dimensional nuclear magnetic
resonance microimaging of trabecular bone, J. Bone Miner. Res., 10, 1452, 1995.
78. Majumdar, S., Kothari, M., Augat, P., et al., High-resolution magnetic resonance imaging: three-
dimensional trabecular bone architecture and biomechanical properties, Bone, 22, 445, 1998.
79. Ito, M., Hayashi, K., Uetani, M., et al., Bone mineral and other bone components in vertebrae evaluated
by QCT and MRI, Skeletal Radiol., 22, 109, 1993.
80. Kroger, H., Vainio, P., Nieminen, J., and Kotaniemi, A., Comparison of different models for interpreting
bone mineral density measurements using DXA and MRI technology, Bone, 17, 157, 1995.
81. Bradbeer, J.N., Kapadia, R.D., Sarkar, S.K., et al., Disease-modifying activity of SK&F 106615 in
rat adjuvant-induced arthritis. Multiparameter analysis of disease magnetic resonance imaging and
bone mineral density measurements, Arthritis Rheum., 39, 504, 1996.
82. Jergas, M.D., Majumdar, S., Keyak, J.H., et al., Relationships between Youngs modulus of elasticity,
ash density, and MRI derived effective transverse relaxation T2* in tibial specimens, J. Comput. Assist.
Tomogr., 19, 472, 1995.
83. Sievanen, H., Kannus, P., Oja, P., and Vuori, I., Dual energy X-ray absorptiometry is also an accurate
and precise method to measure the dimensions of human long bones, Calcif. Tissue Int., 54, 101, 1994.
84. Sievanen, H., Kannus, P., Nieminen, V., et al., Estimation of various mechanical characteristics of
human bones using dual energy X-ray absorptiometry: methodology and precision, Bone, 18, 17S,
1996.
85. Markel, M.D., Wikenheiser, M.A., Morin, R.L., et al., Quantication of bone healing. Comparison of
QCT, SPA, MRI, and DEXA in dog osteotomies, Acta Orthop. Scand., 61, 487, 1990.
86. Aro, H.T., Wippermann, B.W., Hodgson, S.F., et al., Prediction of properties of fracture callus by
measurement of mineral density using micro-bone densitometry, J. Bone Joint Surg., 71A, 1020, 1989.
87. Nordsletten, L., Kaastad, T.S., Skjeldal, S., et al., Fracture strength prediction in rat femoral shaft and
neck by single photon absorptiometry of the femoral shaft, Bone Miner., 25, 39, 1994.
88. Borg, J., Mollgaard, A., and Riis, B.J., Single X-ray absorptiometry: performance characteristics and
comparison with single photon absorptiometry, Osteoporosis Int., 5, 377, 1995.
89. Markel, M.D., Sielman, E., and Bodganske, J.J., Densitometric properties of long bones in dogs, as
determined by use of dual-energy X-ray absorptiometry, Am. J. Vet. Res., 55, 1750, 1994.
0266/ ch06/frame Page 116 Thursday, October 21, 1999 2:22 PM
Methods of Evaluation for Bone Dimensions, Densities, Contents, Morphology, and Structures 117
90. Mosheiff, R., Klein, B.Y., Leichter, I., et al., Use of dual-energy X-ray absorptiometry (DEXA) to
follow mineral content changes in small ceramic implants in rats, Biomaterials, 13, 462, 1992.
91. Lu, P.W., Briody, J.N., Howman-Giles, R., et al., DXA for bone density measurement in small rats
weighing 150250 grams, Bone, 15, 199, 1994.
92. Bagi, C.M., Ammann, P., Rizzoli, R., and Miller, S.C., Effect of estrogen deciency on cancellous
and cortical bone structure and strength of the femoral neck in rats, Calcif. Tissue Int., 61, 336, 1997.
93. Markel, M.D., Bogdanske, J.J., Xiang, Z., and Klohnen, A., Atrophic nonunion can be predicted with
dual energy X-ray absorptiometry in a canine ostectomy model, J. Orthop. Res., 13, 869, 1995.
94. Vanderschueren, D., Van Herck, E., Schot, P., et al., The aged male rat as a model for human
osteoporosis: evaluation by nondestructive measurements and biomechanical testing, Calcif. Tissue
Int., 53, 342, 1993.
95. Turner, A.S., Alvis, M., Myers, W., et al., Changes in bone mineral density and bone-specic alkaline
phosphatase in ovariectomized ewes, Bone, 17, 395S, 1995.
96. Augat, P., Reeb, H., and Claes, L.E., Prediction of fracture load at different skeletal sites by geometric
properties of the cortical shell, J. Bone Miner. Res., 11, 1356, 1996.
97. Louis, O., Soykens, S., Willnecker, J., et al., Cortical and total bone mineral content of the radius:
accuracy of peripheral computed tomography, Bone, 18, 467, 1996.
98. Nicholson, P.H., Lowet, G., Cheng, X.G., et al., Assessment of the strength of the proximal femur in
vitro: relationship with ultrasonic measurements of the calcaneus, Bone, 20, 219, 1997.
99. Amo, C., Revilla, M., Hernandez, E.R., et al., Correlation of ultrasound bone velocity with dual-
energy X-ray bone absorptiometry in rat bone specimens, Invest. Radiol., 31, 114, 1996.
100. Njeh, C.F., Boivin, C.M., and Langton, C.M., The role of ultrasound in the assessment of osteoporosis:
a review, Osteoporosis Int., 7, 7, 1997.
101. Gregg, E.W., Kriska, A.M., Salamone, L.M., et al., The epidemiology of quantitative ultrasound: a review
of the relationships with bone mass, osteoporosis and fracture risk, Osteoporosis Int., 7, 89, 1997.
102. Ashman, R.B., Rho, J.Y., and Turner, C.H., Anatomical variation of orthotropic elastic moduli of the
proximal human tibia, J. Biomech., 22, 895, 1989.
103. Van der Perre, G. and Lowet, G., in vivo assessment of bone mechanical properties by vibration and
ultrasonic wave propagation analysis, Bone, 18, 29S, 1996.
104. Rho, J.Y., Kuhn-Spearing, L., and Zioupos, P., Mechanical properties and the hierarchical structure
of bone, Med. Eng. Phys., 20, 92, 1998.
105. Ascenzi, A., The micromechanics vs. the macromechanics of cortical bone a comprehensive
presentation, J. Biomech. Eng., 110, 357, 1988.
106. Boyde, A., Hendel, P., Hendel, R., et al., Human cranial bone structure and the healing of cranial
bone grafts: a study using backscattered electron imaging and confocal microscopy, Anat. Embryol.,
181, 235, 1990.
107. Zioupos, P., Currey, J.D., and Sedman, A.J., An examination of the micromechanics of failure of bone
and antler by acoustic emission tests and laser scanning confocal microscopy, Med. Eng. Phys., 16,
203, 1994.
108. Hein, H.J., Czurratis, P., Schroth, D., and Bernstein, A., A comparative study of the application of
scanning acoustic microscopy and confocal laser scanning microscopy to the structural assessment
of human bones, Anat. Anz., 177, 427, 1995.
109. Boyde, A., Wolfe, L.A., Maly, M., and Jones, S.J., Vital confocal microscopy in bone, Scanning, 17,
72, 1995.
110. Lekka, M., Lekki, J., Shoulyarenko, A.P., et al., Scanning force microscopy of biological samples,
Pol. J. Pathol., 47, 51, 1996.
111. Kirkpatrick, S.J. and Brooks, B.W., Micromechanical behavior of cortical bone as inferred from laser
speckle data, J. Biomed. Mater. Res., 39, 373, 1998.
112. Kabasawa, M., Ejiri, S., Hanada, K., and Ozawa, H., Histological observations of dental tissues using
the confocal laser scanning microscope, Biotech. Histochem., 70, 66, 1995.
113. Feldkamp, L.A. and Jesion, G., 3D X-ray computed tomography, in Review of Progress in Quantitative
Non-destructive Evaluation, Thompson, D.O. and Chimenti, D.E., Eds., Plenum Press, New York,
1986.
114. McAlister, G.B. and Moyle, D.D., Some mechanical properties of goose femoral cortical bone,
J. Biomech., 16, 577, 1983.
0266/ ch06/frame Page 117 Thursday, October 21, 1999 2:22 PM
118 Mechanical Testing of Bone and the BoneImplant Interface
115. Hogan, H.A., Micromechanics modeling of Haversian cortical bone properties, J. Biomech., 25, 549,
1992.
116. Turner, C.H., Chandran, A., and Pidaparti, R.M., The anisotropy of osteonal bone and its ultrastructural
implications, Bone, 17, 85, 1995.
117. Katz, J.L. and Meunier, A., Scanning acoustic microscope studies of the elastic properties of osteons
and osteon lamellae, J. Biomech. Eng., 115, 543, 1993.
118. Ascenzi, A., Bigi, A., Ripamonti, A., and Roveri, N., X-ray diffraction analysis of transversal osteonic
lamellae, Calcif. Tissue Int., 35, 279, 1983.
119. Rho, J.Y., Tsui, T.Y., and Pharr, G.M., Elastic properties of human cortical and trabecular lamellar
bone measured by nanoindentation, Biomaterials, 18, 1325, 1997.
120. Ascenzi, A., Benvenuti, A., and Bonucci, E., The tensile properties of single osteonic lamellae:
technical problems and preliminary results, J. Biomech., 15, 29, 1982.
121. Wiesmann, H.P., Chi, L., Stratmann, U., et al., Sutural mineralization of rat calvaria characterized by
atomic-force microscopy and transmission electron microscopy, Cell Tissue Res., 294, 93, 1998.
122. Ascenzi, A., Benvenuti, A., Bigi, A., et al., X-ray diffraction on cyclically loaded osteons, Calcif.
Tissue Int., 62, 266, 1998.
123. Tao, N.J., Lindsay, S.M., and Lees, S., Measuring the microelastic properties of biological material,
Biophys. J., 63, 1165, 1992.
124. Gardner, T.N., Elliott, J.C., Sklar, Z., and Briggs, G.A., Acoustic microscope study of the elastic
properties of uorapatite and hydroxyapatite, tooth enamel and bone, J. Biomech., 25, 1265, 1992.
0266/ ch06/frame Page 118 Thursday, October 21, 1999 2:22 PM

119

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

7

General Considerations
of Mechanical Testing

Yuehuei H. An and Christopher V. Bensen

CONTENTS

I. Introduction ..........................................................................................................................119
II. Sources of Bone ...................................................................................................................119
III. Harvesting of Bone ..............................................................................................................120
IV. Preservation of Bone............................................................................................................120
V. Preparation of Bone Specimens for Mechanical Testing....................................................122
VI. Testing Procedures ...............................................................................................................123
VII. Factors Affecting the Test Results .......................................................................................125
A. Machine Compliance......................................................................................................125
B. SpecimenFixture Interface ...........................................................................................126
C. Specimen Size and Geometry........................................................................................127
D. Specimen End Effects ....................................................................................................129
E. Strain Rates ....................................................................................................................129
F. Testing Conditions..........................................................................................................130
VII. Summary ..............................................................................................................................130
References ......................................................................................................................................131

I. INTRODUCTION

Before the start of testing, each researcher has to understand clearly what kind of bone material will
be tested and what mechanical properties are to be determined. The best results can only be obtained
if the researcher plans his or her research project carefully, with a detailed protocol. The protocol should
include the sources of bone specimens, the harvesting procedures, methods of storage, preparation of
bone specimens for testing, testing procedures (testing conditions, testing, data collection, and analysis),
and potential factors which may affect the test results.

II. SOURCES OF BONE

Depending on the source of the bone specimens, considerations should be paid to the timing,
harvesting technique, and the methods to prevent the bone from postmortem autolysis.
There are several major sources of bone which can be used in mechanical testing. First, bone
specimens can be obtained from a patient during surgery. Because of the functional need and limited
volume in the human body, precision of the procedure is essential. It is essential that a consent
form be lled out when a bone specimen is taken from or an experimental procedure is performed
on a volunteer patient.

1,2

Bone specimens can also be obtained at necropsy. In order to avoid artifacts caused by autolysis
(especially obvious in small animals),

2

a necropsy should be done immediately after the subject,

0266/ ch07/frame Page 119 Thursday, October 21, 1999 2:27 PM

120

Mechanical Testing of Bone and the BoneImplant Interface

human or animal, has died. If this is not possible, the body should be placed in a leakproof plastic
bag and put in a refrigerator until the necropsy is performed. If bone specimens are taken within
several days (up to 3 to 4 days) of death or euthanasia (for animals), there will be no signicant
change on the mechanical properties of the bone.

3

The body should never be frozen if histopatho-
logical examination of the tissues is also needed, because ice crystals form inside the cells, and
when thawing ocurrs, the cells rupture making histological evaluation difcult or impossible.
When harvesting bones from human cadavers, bone specimens should be taken immediately
after death. However, most available cadavers are fresh-frozen, which is acceptable for most
mechanical testing purposes. Formalin-embalmed bodies are not appropriate for testing for mechan-
ical properties of bone, because the bone is also xed by formalin and the mechanical properties
have been partially or completely changed.

4-6

Another source of bone is the animal slaughterhouse. Most animals are euthanized and stored
in a cold room for several days before the meat is removed. Arrangement should be made with the
owner on the day of harvesting, so adequate procedures can be observed. Ideally, bone should be
harvested immediately after the animal is euthanized.

III. HARVESTING OF BONE

When dissecting soft tissues, caution is needed to avoid cutting the bone surface, which will cause
stress concentrations. This is especially important when the bone is harvested for whole-bone testing
and especially signicant for small animal bones. Bone specimens to be used for mechanical testing
should be harvested with sufcient extra tissues around the area of interest (not applicable for sampling
in patient surgery). A handsaw or wire saw is efcient for cutting bone. Keeping the surrounding soft
tissue (muscle, fascia, or skin) intact is very helpful for protecting the bone from drying.
When harvesting large bone specimens, which requires an electric autopsy saw, wire saw, or
bow saw, it is important to keep the site irrigated with saline. If only a portion of the bone is
needed, a surgical marking pen is helpful for maintaining the size of the desired specimen.
Subdivision or trimming with a small bench saw or other saw unit is sometimes necessary before
mechanical testing or storage in a freezer.
Before testing or storage, it is good practice to obtain radiographs and photographs of the bone
specimens which are useful as documentation or as a guide for further preparation in the future.
For harvesting pathological bone specimens, attention should be paid to the size of fracture callus,
the alignment of the diaphysis, the positions of the implants or xation devices, and the surrounding
tissues. For conditions involving joints, the morphology of the articular cartilage is another impor-
tant consideration. Pathological ndings include roughened areas caused by arthritis, cartilage
defects, fracture lines, or osteophytes. Also, the size of the joint, joint capsule and synovium, the
amount and characteristics of the joint uid, and the appearance of ligaments, menisci, and the soft
tissues around the joint should be observed and recorded.
Another good practice is proper labeling of the harvested specimens. Laboratory personnel
should be ensured that the patients information (or animal number), the date of harvesting, the
name of the bone, right or left side, and biohazard status are properly labeled on the plastic bag
and recorded in the laboratory logbook. Often, some bone specimens in the freezer are found useless
for certain projects (which need certain sex, age, and pathological conditions of the bone) simply
because the specimens were not properly labeled when they were harvested.

IV. PRESERVATION OF BONE

Regarding the storage conditions for bone specimens, several factors should be considered including
the temperature, moisture, use of preservation solutions (such as normal saline or xatives), and
the pretreatment (such as sterilization methods).

0266/ ch07/frame Page 120 Thursday, October 21, 1999 2:27 PM

General Considerations of Mechanical Testing

121

Melnis and Knets

3

found that different conditions for the storage of bone tissue play an important
role in their mechanical properties. They evaluated ve different storage conditions:
1. Stored in room conditions (18

C, moisture: 65%) and also tested in these conditions;
2. Stored for 30 days in polyethelene packages at 4

to 7

C and tested at 37

C and
moisture 90%;
3. Stored for 30 days in 0.9% NaCl at 4 to 7

C and tested at 37

C and moisture 90%;
4. Same as 3, plus during testing the samples were wrapped in soft material saturated with
saline; and
5. Stored at 18

C and moisture 65% and two days before testing they were kept in saline.
Testing was carried out at 37

C and moisture 90%. During testing the samples were wrapped
in soft material saturated with saline. It was found that the largest tensile creep strain arose from
those specimens that were stored and tested at room temperature (18

C) and moisture (65%) and
the overall results demonstrated the effects of temperature and moisture on mechanical properties
of bone.
Roe et al.

4

studied the effects of various preparation and storage procedures and of different
storage times on structural properties of canine cortical bone. Preparation and storage procedures
evaluated were (1) sterile collection and storage at 20

C; (2) ethylene oxide sterilization and
storage at 22

C; (3) chemical sterilization (methanol and chloroform, then iodoacetic acid) and
storage at 20

C; and (4) chemical sterilization, partial decalcication, and storage at 20

C. The
results revealed that chemically sterilized bone had not changed after 1 week of storage, whereas
chemically sterilized and partially decalcied bone had a 40 to 60% decrease in compressive load
to failure, pullout load, and screw-stripping torque. Chemically sterilized and partially decalcied
bone remained weak after 16 and 32 weeks of storage. Signicant structural alterations were not
detected in aseptically collected bone after 16 or 32 weeks of storage. Ethylene oxidesterilized
bone had a reduced pullout load after 32 weeks of storage. Chemically sterilized and partially
decalcied bone specimens had signicantly reduced mechanical strength.
Refrigeration is appropriate for short periods of time, i.e., several days, which has been partially
veried by Kaab et al.

5

The only adequate method for the long-term storing of bone specimens for
mechanical testing is freezing. Because of the evidence of damage to soft tissues, such as cartilage,

6

tendon,

7

or skin,

8

due to the freezing procedure, there are some questions on the appropriateness
of freezing preservation of bone specimens. Owing to the complexity of an experiment or unforeseen
circumstances, sometimes a specimen must be thawed and frozen multiple times. The question
arises whether multiple freezing and thawing is harmful to the mechanical properties. This question
has been partially answered by Linde and Srensen,

9

who found that freezing and thawing ve
times did not alter the compressive properties of cancellous bones. In our recent study, the proximal
portion of the tibia of adult cows was sectioned to produce bone slices. They were then subjected
to four freezingthawing conditions: freezing with and without saline solution, then thawing in
saline solution or exposed to air. The mechanical properties of the bone before and after the
treatments (ve cycles of freezing and thawing) were measured using an indentation test. It was
found that there is no signicant effect on the ultimate load and stiffness of the bone. Only a trend
of difference was noticed for the specimens frozen without saline soaking and thawed in air.

10

This
work supports the practice of freezing and thawing bone specimens in saline solution.
The common method for storing bone specimens is freezing at 20

C. The bones should always
be frozen in saline to prevent dehydration.



The effects of storage on mechanical properties of bone
at 20

C for short periods of time are minor.

10

The maximum effect reported is a 4.6% reduction
of torsional strength of canine long bones.

11

However, after thawing, enzymes such as collagenases
and proteases may become active and degrade the tissue. Also, enzymatic degradation is not
completely arrested at 20

C.

12

With concerns about the effects of enzyme degradation

13

and
evaporation,

14

a question arises as to whether there are signicant effects of long-term storage at

0266/ ch07/frame Page 121 Thursday, October 21, 1999 2:27 PM

122

Mechanical Testing of Bone and the BoneImplant Interface

20

C. Panjabi et al.

15

found no signicant effects of freezing for 7 to 8 months on the mechanical
properties of human vertebral bone. Roe et al.

4

found that bones frozen at 20

C for 8 months did
not become signicantly weaker. Because time periods longer than 8 months have not been reported
for frozen storage at 20

C, storage at this temperature for more than 8 months is not recommended.
Alternatively, 70

C, 80

C, or even lower temperatures or liquid nitrogen are suggested for long-
term bone storage, since these temperatures may minimize evaporation

14

and markedly reduce
enzyme activity.

13

Based on previous investigations, bone specimens should be soaked in saline or phosphate
buffered saline (PBS) and frozen at 20

C (a 70

C freezer is ideal if available) in an airtight
plastic bag until testing.

10

Before testing, bone specimens should be thawed in saline at room
temperature for at least 3 h. If testing cannot be conducted after samples have been prepared and
the procedure is expected to be done within 1 or 2 days, the samples should be stored in a 4

C
refrigerator until testing. Otherwise, the samples should be put back into saline and frozen. One
may encounter a situation where specimens consisting of boneimplant interface have been har-
vested according to the protocol and there is an unexpected problem with the mechanical testing
machine. The specimens should be stored in a refrigerator and effort made to conduct the testing
as soon as possible, because the potential effects of freezing on the boneimplant interface has not
been documented.

V. PREPARATION OF BONE SPECIMENS FOR MECHANICAL TESTING

Before preparation of specimens, it has to be clear what structural level of the bone is going to be
tested. If the purpose is to determine the properties of bone tissue or organ (whole bone), there
exists the minimum dimension of the cross section of specimen.

16

This dimension is determined
by analysis of the structural levels existing in the composite bone structure and is recommended
to be no less than 2 mm. In the case of testing thinner specimens (micromechanical testing), one
would expect to get the properties of structural elements of the bone tissue, such as a single osteon,
lamellae, or individual trabeculae. Their mechanical behaviors, certainly, may not represent that of
bone tissue in general, or whole-bone structural properties.
Rough cuts can be made with a regular bandsaw equipped with a

-in. ne-tooth saw blade.
For parallel cuts, a bandsaw installed with a customized guide is sufcient for most purposes. To
prevent burning, a relatively low speed should be used with sufcient saline irrigation. This kind
of cutting may only affect a 1 mm depth of bone at the surface, which can be ground off using a
polishing wheel.
Fine cuts can be made using a diamond wafering saw (such as the Buehler Isomet 1000, Leco
VC-50, or Struers Accutome-5) (Figure 7.1) or a diamond wire saw (Histosaw, Delaware Diamond
Knives, Wilmington, DE) (Figure 7.2) which is particularly good for making smooth, parallel cuts.
For fabrication of cylindrical samples, a tabletop drill press is sufcient for relatively large
samples (>5 mm diam.). To prevent the adverse effects of vibration, C-clamps may be used to
secure the bone to the machine platform. For 4 to 5 mm diam. samples or less, a lathe or milling
machine is recommended. Although electric hand drills can be used, they are not ideal for making
cylindrical samples. When drilling holes in the bone for screw pullout tests, a drill press also
functions better than a hand drill.
Grinding or polishing is often used to adjust uneven cut surfaces. This is especially important
in fracture tests because the imperfections on the surface of the specimen may serve as the initiators
of cracks. The commonly used commercially available grinding machines include the Buehler
Ecomet 3, Struers Dap-V, or Leco VP-160.
By using a coring bit made from a hypodermic needle on a miniature drill press, small cylindrical
specimens, measuring 2 to 3 mm in length and less than 1 mm in diameter, can be obtained.

17

With
a low-speed diamond wafering saw and a specially designed miniature milling machine, 100-

m-
thick and less than 200-

m-rectangular beams can be made.

18,19

0266/ ch07/frame Page 122 Thursday, October 21, 1999 2:27 PM

General Considerations of Mechanical Testing

123

To harvest single osteons for micromechanical testing, two methods have been reported. The
rst one was reported by Ascenzi and Bonucci

20

using a turning needle method. Briey, the
device consists of a ne and well-sharpened steel needle which is eccentrically inserted on a dental
drill. While the drill is turning, the tip of the needle describes a circle whose diameter corresponds
to the average diameter of osteons. By using this method, cylindrical samples measuring 200

m
diam.

500

m length can be made. The second method for harvesting single osteons was reported
by Frasca et al.

21

using a splitting and scraping method.

21-23

Under a 30

stereoscopic microscope,
single osteons are isolated by propagating fractures along their natural boundaries using a pair of
ne tweezers and a scalpel. They stated that osteons up to 1 or 2 cm can be obtained using their
method. Also of note, a more technically demanding method for isolating single osteonic lamellae
is also available in the literature reported by Ascenzis group.

24,25

Also under a stereoscopic microscope and using microsurgical instruments, single trabeculae
can be obtained.

26,27

A common method to mount a trabeculae to the microtester is embedding in
polymethylmethacrylate (PMMA)

28

or epoxy.

27,29

A method using cyanoacrylate glue with tube-
shaped grips has also been reported.

30

These methods have been reported to study individual
trabeculae include the buckling,

26,27

compressive,

28

tensile,

28,30

torsional,

22

cantilever,

29

and bending
tests.

31

Also, the tests can be nondestructive.

28

For a macroindentation test, a cut surface should be further polished using SiC papers (a nal
400 grit is good enough). Small bone specimens need to be potted in dental stone or PMMA to
facilitate grinding and testing. For micro- and nanoindentation tests, the bone specimens must be
mounted in a resin or PMMA block to provide support for polishing and testing. For nanoindentation
tests, bone specimens can to be dehydrated and embedded in PMMA to provide support for the
porous network.

32,33

Dry bone can also be used without inltration and embedding.

34

After polishing
with successive 400-, 800-, and 1200-grit SiC papers, at least 1.0

m aluminum oxide paste (Buehler
Micropolish C alpha Alumina, Lake Bluff, IL) should be used to nalize the surface for micro-
indentation testing. For nanoindentation, 0.3 to 0.05

m particle size aluminum oxide paste should
be used after the polishing with 1200-grit SiC paper or 1.0-

m polishing paste.

33,34

VI. TESTING PROCEDURES

The temperature and moisture of the laboratory should be controlled as much as possible. If
there are no specic requirements, most experiments can be done at room temperature (24

C) and
relative moisture (40 to 90%). The specimens should be kept moist with periodic application of
normal saline.

FIGURE 7.1

Fine cuts can be made using a diamond wafering saw (Buehler Isomet 1000, Buehler,
Lake Bluff, IL).

0266/ ch07/frame Page 123 Thursday, October 21, 1999 2:27 PM

124

Mechanical Testing of Bone and the BoneImplant Interface

Specimens are directly placed on the platform in the case of compression tests or on the two
supporting fulcra in the case of bending tests. Bone ends need to be potted in dental stone or resin
in tensile tests or torsional tests. The use of screws, wires, or rods for the xation of specimens to
the testing machine or the use of potting procedures for stronger xation has also been reported.
The mechanical testing machine is operated in displacement control for most tests. The machine
linear variable displacement transducer (LVDT) should be periodically calibrated using an exten-
someter. Loading is commonly conducted at a constant slow rate (1 mm/min is the rate used in
the authors laboratory). Load at the peak point of the loaddisplacement curve is taken as the
ultimate load (Figure 7.3). A stiffness measure is obtained by measuring the slope of the linear
portion of the curve. If the test machine is controlled with a linear displacement rate and the
specimen xture is very rigid, the time base of the recorder can be converted to specimen

FIGURE 7.2

The Histosaw, by Delaware Diamond Knives, Wilmington, DE, is particularly good for making
smooth, parallel cuts.

0266/ ch07/frame Page 124 Thursday, October 21, 1999 2:27 PM

General Considerations of Mechanical Testing

125

deformation. An extensometer is recommended when a tensile test is conducted or for other tests
when a complicated or less rigid specimen xture is employed. The deformation measured by a
built-in LVDT includes the deformation of the specimen and potential displacements within the
specimen xture or at the specimenxture interface. An extensometer attached to the specimen
provides a direct measurement of specimen strain without the complications of machine or xture
deformation. Load control mode is often used for fatigue tests. Normally, a small or physiological
load or stress (much smaller than the ultimate load or strength of the material) is applied to the
specimen repeatedly at certain frequencies until it fails. Both macro- and microspecimens

31,35

can
be tested using cyclic loading.
Mechanical testing of materials involves the application of measurable loads to specimens of
uniform dimensions. The applied stress is calculated by dividing the applied force by the area over
which the force acts. Change in a specimen dimension divided by the original specimen dimension
denes strain. Dependent upon the direction in which the force is applied, the test may be tensile,
compression, or bending. A simple way to record the data is a loaddisplacement curve from which
the ultimate load, stiffness, and displacement can be obtained. A stressstrain curve is not always
plotted. Ultimate strength and elastic modulus (if applicable) are often calculated using the recorded
loads, ultimate load and displacements, and dimensions of the specimen.

VII. FACTORS AFFECTING THE TEST RESULTS
A. M

ACHINE

C

OMPLIANCE

If several xture parts are used in the testing assembly, signicant machine compliance can exist.
Routine testing for machine compliance is recommended by testing the xture column without
the specimen in place.

1,2

If the machine and xture deformation is found to be signicant (more
than 10% of the specimen value), it should be accounted for in the data analysis.

1

Because the
testing gear should be the same for each specimen, only the mean machine compliance (or
machine stiffness,

S

m

) is needed. The stiffness of the bone sample (

S

b

) is calculated using the
following equations:

FIGURE 7.3

A typical loaddisplacement curve.

0266/ ch07/frame Page 125 Thursday, October 21, 1999 2:27 PM

126

Mechanical Testing of Bone and the BoneImplant Interface

S

b

=

P

/(

d

b+m

d

m

) (7.1)
=

P

/(

P

/

S

b+m

P/S

m

) (7.2)
=

S

b+m

S

m

/(

S

m

S

b+m

) (7.3)
where

P

is the load where deformation of the testing machine (

d

m

) or deformation of the machine
plus bone specimen (

d

b+m

) are taken.

S

b+m

is the tested stiffness value (the stiffness of the machine
plus bone specimen).
One does not have to worry about the rigidity of the machine frame, which is very rigid such
that minimal or no deformation within the allowed loading range will exist. If a signicant
compliance is observed, the sources are more likely the specimen xture (made of weak materials)
and/or the xturemachine interface (loose connection, such as loose screw connection). To reduce
the effects of machine compliance, specimen xtures should be rmly machined using rigid
materials, should be as simple as possible, and should be rmly connected to the machine. The
more connections within the specimenxture assembly, the more machine compliance.
B. SPECIMENFIXTURE INTERFACE
When a compression test is used, friction at the specimen and platens should be considered. There
can be two completely different forms of fracture depending upon conditions of specimen-machine
interface. If this interface is dry leading to strong friction between the bone specimen and loading
plate, then fracture of specimen will be caused by maximum shear stresses. The fracture surfaces
in the specimen in this case will be oriented at the angle of 45 to the axes of loading. If the
specimenmachine interface is oily or wet allowing some sliding along the interface in the transverse
direction, then a fracture will be caused by transverse strains arising during loading. The fracture
lines in the specimen will be oriented parallel to the axes of loading. Figure 7.4 illustrates a
loaddisplacement curve with microfractures occurring during specimen loading. A similar pattern
can also be observed with slippage at the specimenxture interface.
An uneven specimen surface causes a triaxial stress eld, leading to overestimation of the
specimen stiffness. This effect can be limited by using a more accurate procedure for specimen
fabrication to achieve parallel end surfaces. An overestimation of specimen stiffness can also be
caused by the horizontal friction between the surfaces of the specimen and the platens. It is known
that both the axial and lateral deformations of a specimen between the upper and lower platen are
larger at the ends of the specimen than in the central part of the specimen (end phenomenon or
end effect). Any restrictions to the lateral expansion, such as a rough platen surface, will cause an
overestimation of the true specimen stiffness. Common methods for reducing this kind of friction
include the use of grease at the interface and using low-friction stainless steel platen surfaces
(polished mirror surfaces).
When a tensile test is employed, the effect of the specimenxture interface should be consid-
ered. Any loosening or low rigidity at the interface will lead to an underestimation of the true
specimen values. Therefore, a rigid connection between the specimen and the xture is essential.
Using a dumbbell-shaped specimen or a PMMA end-coated specimen are two common strategies
to achieve good bonding between the specimen and xture.
36,37
An external extensometer should
be used in these situations to measure the specimen deformation accurately. Slipping out of the
ends of the specimen from the grips of the testing machine may cause the most undesirable effect
in a tensile test. In this case, when the displacement between the ends of grips is measured
automatically by the test machine during testing, then this aforementioned slipping will give an
incorrect result. This, consequently, will lead to a distorted stressstrain curve and, further, to
incorrectly calculated material parameters, such as strength or modulus of elasticity. Therefore, it
is recommended to measure the displacements by special strain gauges attached to the surface of
0266/ ch07/frame Page 126 Thursday, October 21, 1999 2:27 PM
General Considerations of Mechanical Testing 127
specimen in the middle fth of the length of the specimen. For very short tensile specimens, there
could also be the negative inuence of the nonuniform stress distribution at the gripspecimen
contact line.
For compressive and tensile testing and most other mechanical testing procedures, preloading
with a small load is useful for tightening the specimenxture interface to further limit the effect
of the interface.
C. SPECIMEN SIZE AND GEOMETRY
The size and geometric dimensions of a specimen have signicant inuence on the outcome of
mechanical testing. In the case of whole-bone testing, it is imperative that bones be of uniform
size. If procedures or treatments on whole bones are being compared against one another, different
sized specimens must be equally distributed among the different groups. This can be achieved by
using right and left bones of the same individual and/or by measuring and weighing the specimens.
It must be emphasized that signicant differences in bone mineral density may exist between
individuals and must be considered in these investigations.
FIGURE 7.4 An actual loaddisplacement curve illustrating microfractures during specimen loading.
0266/ ch07/frame Page 127 Thursday, October 21, 1999 2:27 PM
128 Mechanical Testing of Bone and the BoneImplant Interface
When smaller specimens, i.e., cut bone sections or blocks, are being tested, the question arises
in which shape the specimen should be prepared. The most common specimen geometry is that of
a cube or cylinder. These specimens are the easiest to prepare and/or machine and are the most
reproducible. One important concept in the production of these samples for testing is the length to
diameter (L/D) ratio of the specimen. Various ratios have been reported in the literature ranging
from 2
38
to 0.25
39
with diameters between 5
40,41
and 20 mm.
39
Wixson et al.
42
compared the strength
and stiffness of human bone samples of various length obtained at total knee arthroplasty. They
found these parameters to be directly proportional to the length of the specimen. Longer specimens
(L/D ratio > 5) have a tendency to buckle during testing and should be avoided. Conversely,
shorter specimens (L/D < 1) will exhibit a signicant friction effect between the specimen and the
platens leading to overestimation of stiffness. Consequently, most investigators have recommended
an L/D ratio of between 1 and 2 for typical compression testing. Keaveny et al.
43
compared
parameters of specimens using an accurate nondestructive method and the platens compression test
and found a signicant inuence of aspect ratio. Specimens with an L/D ratio of 2 had the least
differences between the two methods.
Numerous investigators have also employed a dumbell-shaped specimen for compression,
tensile, and torsional testing of human and animal bones. These specimens allow improved grip
holding and rigidity at the specimenmachine interface. However, they are more difcult to machine
than standard cylindrical or beam specimens.
The surfaces of specimens should be smooth and without indentations or defects. Such imper-
fections cause stress concentrations which are especially signicant in bending and torsional testing.
The two surfaces of cylindrical or cubic specimens to be tested should be parallel to each other.
Unparallel specimens make the loading uneven or off-center, which generates a bending moment
to the cylinder and leads to inaccurate data (Figure 7.5).
When preparing a specimen containing a cylindrical implant for a pushout or pullout test, it is
important that the two cut surfaces be made perpendicular to the implant followed by precise ne
cuts and grinding. If this is not done properly, it will not be possible to align the implant accurately,
FIGURE 7.5 Off-center loading on cylindrical bone specimen. (A) Pre-loading; (B) Loading.
0266/ ch07/frame Page 128 Thursday, October 21, 1999 2:27 PM
General Considerations of Mechanical Testing 129
resulting in off-axis load application. If the load applied to the implant is oblique, then erroneous
data will be obtained as a result in catching effects (Figure 7.6B).
D. SPECIMEN END EFFECTS
In compression testing of bone specimens, attention must be given to the effects of the interface
between the specimen and the test platen. One such effect is the structural end phenomenon.
First described by Linde and Hvid in 1989,
50
this occurs when vertically oriented and unsupported
trabeculae in a cut-out specimen slide along the surface of the test platen. Consequently, strain
inhomogeneity occurs in the specimen, leading to underestimation of stiffness and overestimation
of ultimate strain. The magnitude of these errors is dependent on the length and L/D ratio of the
specimen.
38
One method of correcting for this effect is embedding the ends of the specimen in
PMMA bone cement. One study showed a 40% increase in stiffness with the use of embedded
specimens.
50
E. STRAIN RATES
When performing a mechanical testing procedure, one of the most important considerations in
testing parameters is that of strain rate. One of the rst investigations of the effect of loading rate
on the mechanical properties of bone during compression testing was that of McElhaney and
Byars.
44
They studied cortical bone from human and bovine femurs and determined that both
ultimate load and modulus of elasticity increased with increased strain rate. For example, a bovine
FIGURE 7.6 (A) Force transmitted in line with the implant yields reproducible results with a smooth
loaddisplacement curve. (B) If the force is applied obliquely, an irregular curve (catching effects) results,
making the data reduction difcult.
0266/ ch07/frame Page 129 Thursday, October 21, 1999 2:27 PM
130 Mechanical Testing of Bone and the BoneImplant Interface
femur specimen loaded at a rate of 0.001 in./in./s yielded an ultimate compressive strength of 17.93
kgf/mm
2
and an elastic modulus of 1898 kgf/mm
2
. However, when the strain rate was increased to
1500 in./in./s, those values increased to 37.26 and 4288 kgf/mm
2
, respectively. Conversely, the
energy absorption capacity (kg-cm/cm
3
), maximum strain to failure (%), and Poissons ratio all
decreased with increasing rate of loading. Several other investigators have also demonstrated load
rate sensitivity with human calvarial bone plugs.
45-47
Compression testing studies on human and bovine trabecular bone have also shown that both
stiffness and ultimate strength are directly proportional to strain rate.
39,48,49
Linde also found a
positive correlation between strain rate and ultimate strength, a nding opposite that of McElhanys
results using cortical bone.
These data are of particular interest in the evaluation of fracture models insofar as the properties
of specimens in a fatigue test subjected to repetitive, low strain loading may be quite different from
those of a traumatic fracture model.
F. TESTING CONDITIONS
There has been considerable difference of opinion in the literature whether to conduct testing
procedures on wet or dry bone specimens. Several investigations comparing mechanical parameters
of both wet and dry specimens suggest that dry specimens will exhibit a higher ultimate tensile
strength.
54,55
The specimens should be kept moist with periodic application of normal saline.
The effect of temperature on variations in results of mechanical testing has been well studied.
Brear et al.
51
demonstrated that strength, ultimate strain, and Youngs modulus parameters of
trabecular bone specimens were all lower at body temperature (37C) than at room temperature
(20 to 24C). Recently, Mitton et al.
2
conducted compression and shear tests on ewe vertebral
trabecular bone under two different conditions: room temperature in air (standard test conditions)
and in a physiological saline bath regulated at 37C. Testing of the specimens in a 37C physio-
logical saline bath induced a decrease in the shear strength from 32.5% (p = 0.0005) to 37.3% (p =
0.0001) of those measured under standard test conditions. Similar studies have conrmed this
effect in cortical bone as well.
52,53
Although modern investigations of the boneimplant interface,
such as total joint replacement simulation, are conducted at 37C, most mechanical testing of bone
is performed at room temperature in air. If there are no specic testing requirements, it is the
authors opinion that most experiments could be done at room temperature (20 to 24C) at a relative
humidity of 40 to 90%.
VIII. SUMMARY
In summary, obtaining accurate and reproducible data in the mechanical testing laboratory requires
attention to detail at several important key points in the study. First, specimen collection must be
performed carefully in order to avoid damage to the specimen at the time of harvest. This should
be carried out immediately prior to the testing procedures if at all possible. However, soaking the
specimens in saline and freezing at 20C is adequate for relatively long-term storage up to 8
months. Preparation of the specimens should be carried out methodically, keeping in mind the
effects of heat generation from sawing and/or drilling as well as drying from prolonged exposure
to air. Most researchers conduct mechanical testing at room temperature and humidity. However,
the specimens should be kept moist during the periods of both preparation and testing. During
testing procedures, the investigator must consider the possibility of machine compliance and,
perhaps most importantly, the stability and rigidity of the specimenxture interface. It is the
authors opinion that more errors are generated in the latter than in any other components of
mechanical testing procedures.
0266/ ch07/frame Page 130 Thursday, October 21, 1999 2:27 PM
General Considerations of Mechanical Testing 131
REFERENCES
1. An, Y.H., Kang, Q., and Friedman, R.J., Mechanical symmetry of rabbit bones studied by bending
and indentation testing, Am. J. Vet. Res., 57, 1786, 1996.
2. Mitton, D., Rumelhart, C., Hans, D., and Meunier, P.J., The effects of density and test conditions on
measured compression and shear strength of cancellous bone from the lumbar vertebrae of ewes, Med.
Eng. Phys., 19, 464, 1997.
3. Melnis, A. and Knets, I., Viscoelastic properties of compact bone tissue [in Russian], in Modern
Problems of Biomechanics, Vol. 2, Knets, I., Ed., Zinatne, Riga, 1985, 38.
4. Roe, S.C., Pijanowski, G.J., and Johnson, A.L., Biomechanical properties of canine cortical bone
allografts: effects of preparation and storage, Am. J. Vet. Res., 49, 873, 1988.
5. Kaab, M.J., Putz, R., Gebauer, D., and Plitz, W., Changes in cadaveric cancellous vertebral bone
strength in relation to time. A biomechanical investigation, Spine, 23, 1215, 1998.
6. Bass, E.C., Duncan, N.A., Hariharan, J.S., et al., Frozen storage affects the compressive creep behavior
of the porcine intervertebral disc, Spine, 22, 2867, 1997.
7. Smith, C.W., Young, I.S., and Kearney, J.N., Mechanical properties of tendons: changes with steril-
ization and preservation, J. Biomech. Eng., 118, 56, 1996.
8. Quirinia, A. and Viidik, A., Freezing for postmortal storage inuences the biomechanical properties
of linear skin wounds, J. Biomech., 24, 819, 1991.
9. Linde, F. and Srensen, H.C., The effect of different storage methods on the mechanical properties
of trabecular bone, J. Biomech., 26, 1249, 1993.
10. Kang, Q., An, Y.H., and Friedman, R.J., Effects of multiple freezingthawing cycles on ultimate
indentation load and stiffness of bovine cancellous bone, Am. J. Vet. Res., 58, 1171, 1997.
11. Strmberg, L. and Daln, N., The inuence of freezing on the maximum torque capacity of long
bones. An experimental study on dogs, Acta Orthop. Scand., 47, 254, 1976.
12. Tomford, W.W., Doppelt, S.H., Mankin, H.J., and Friedlaender, G.E., 1983 bone bank procedures,
Clin. Orthop., 15, 1983.
13. Sumner, D.R., Willke, T.L., Berzins, A., and Turner, T.M., Distribution of Youngs modulus in the
cancellous bone of the proximal canine tibia, J. Biomech., 27, 1095, 1994.
14. Malinin, T.I., Martinez, O.V., and Brown, M.D., Banking of massive osteoarticular and intercalary
bone allografts 12 years experience, Clin. Orthop., 44, 1985.
15. Panjabi, M.M., Krag, M., Summers, D., and Videman, T., Biomechanical time-tolerance of fresh
cadaveric human spine specimens, J. Orthop. Res., 3, 292, 1985.
16. Knets, I., Pfafrods, G., and Saulgozis, J., Deformation and Fracture of Hard Biological Tissue [in
Russian], Zinatne, Riga, 1980.
17. McAlister, G.B. and Moyle, D.D., Some mechanical properties of goose femoral cortical bone,
J. Biomech., 16, 577, 1983.
18. Choi, K., Kuhn, J.L., Ciarelli, M.J., and Goldstein, S.A., The elastic moduli of human subchondral,
trabecular, and cortical bone tissue and the size-dependency of cortical bone modulus, J. Biomech.,
23, 1103, 1990.
19. Kuhn, J.L., Goldstein, S.A., Choi, K., et al., Comparison of the trabecular and cortical tissue moduli
from human iliac crests, J. Orthop. Res., 7, 876, 1989.
20. Ascenzi, A. and Bonucci, E., The compressive properties of single osteons, Anat. Rec., 161, 377, 1968.
21. Frasca, P., Harper, R.A., and Katz, J.L., Isolation of single osteons and osteon lamellae, Acta Anat.,
95, 122, 1976.
22. Ascenzi, A., Baschieri, P., and Benvenuti, A., The torsional properties of single selected osteons [see
comments], J. Biomech., 27, 875, 1994.
23. Ascenzi, A., Baschieri, P., and Benvenuti, A., The bending properties of single osteons, J. Biomech.,
23, 763, 1990.
24. Ascenzi, A., Bonucci, E., and Simkin, A., An approach to the mechanical properties of single osteonic
lamellae, J. Biomech., 6, 227, 1973.
25. Ascenzi, A. and Benvenuti, A., Evidence of a state of initial stress in osteonic lamellae, J. Biomech.,
10, 447, 1977.
26. Runkle, J.C. and Pugh, J., The micromechanics of cancellous bone. II. Determination of the elastic
modulus of individual trabeculae by a buckling analysis, Bull. Hosp. Joint Dis., 36, 2, 1975.
0266/ ch07/frame Page 131 Thursday, October 21, 1999 2:27 PM
132 Mechanical Testing of Bone and the BoneImplant Interface
27. Townsend, P.R., Rose, R.M., and Radin, E.L., Buckling studies of single human trabeculae,
J. Biomech., 8, 199, 1975.
28. Samelin, N., Koller, W., Ascherl, R., and Gradinger, R., [A method for determining the biomechanical
properties of trabecular and spongiosa bone tissue], Biomed. Tech (Berlin), 41, 203, 1996.
29. Mente, P.L. and Lewis, J.L., Experimental method for the measurement of the elastic modulus of
trabecular bone tissue, J. Orthop. Res., 7, 456, 1989.
30. Rho, J.Y., Ashman, R.B., and Turner, C.H., Youngs modulus of trabecular and cortical bone material:
ultrasonic and microtensile measurements, J. Biomech., 26, 111, 1993.
31. Choi, K. and Goldstein, S.A., A comparison of the fatigue behavior of human trabecular and cortical
bone tissue, J. Biomech., 25, 1371, 1992.
32. Hodgskinson, R., Currey, J.D., and Evans, G.P., Hardness, an indicator of the mechanical competence
of cancellous bone, J. Orthop. Res., 7, 754, 1989.
33. Lee, F.Y., Rho, J.Y., Harten, R., Jr., et al., Micromechanical properties of epiphyseal trabecular bone and
primary spongiosa around the physis: an in situ nanoindentation study, J. Pediatr. Orthop., 18, 582, 1998.
34. Rho, J.Y., Tsui, T.Y., and Pharr, G.M., Elastic properties of human cortical and trabecular lamellar
bone measured by nanoindentation, Biomaterials, 18, 1325, 1997.
35. Ascenzi, A., Ascenzi, M.G., Benvenuti, A., and Mango, F., Pinching in longitudinal and alternate
osteons during cyclic loading, J. Biomech., 30, 689, 1997.
36. Reilly, D.T., Burstein, A.H., and Frankel, V.H., The elastic modulus for bone, J. Biomech., 7, 271, 1974.
37. Ashman, R.B. and Rho, J.Y., Elastic modulus of trabecular bone material, J. Biomech., 21, 177, 1988.
38. Linde, F., Hvid, I., and Madsen, F., The effect of specimen geometry on the mechanical behavior of
trabecular bone specimens, J. Biomech., 25, 359, 1992.
39. Carter, D.R. and Hayes, W.C., The compressive behavior of bone as a two-phase porous structure,
J. Bone Joint Surg. Am., Vol. 59, 954, 1977.
40. Ducheyne, P., Heymans, L., Martens, M., et al., The mechanical behavior of intracondylar cancellous
bone of the femur at different loading rates, J. Biomech., 10, 747, 1977.
41. Odgaard, A., Hvid, I., and Linde, F., Compressive axial strain distributions in cancellous bone
specimens, J. Biomech., 22, 829, 1989.
42. Wixson, R.L., Elasky, N., and Lewis, J., Cancellous bone material properties in osteoarthritic and
rheumatoid total knee patients, J. Orthop. Res., 7, 885, 1989.
43. Keaveny, T.M., Borchers, R.E., Gibson, L.J., and Hayes, W.C., Theoretical analysis of the experimental
artifact in trabecular bone compressive modulus [published erratum appears in J. Biomech. 26(9),
1143, 1993], J. Biomech., 26, 599, 1993.
44. McElhaney, J.H. and Byars, E.F., Dynamic Response of Biological Materials, ASME Publ.,
65-WA/HUF-9, 1, 1965.
45. Roberts, V.L. and Melvin, J.W., The measurement of the dynamic mechanical properties of human
skull bone, Appl. Polym. Symp., 12, 235, 1969.
46. Carter, D.R. and Hayes, W.C., Bone compressive strength: the inuence of density and strain rate,
Science, 194, 1174, 1976.
47. Wood, J.L., Dynamic response of human cranial bone, J. Biomech., 4, 1, 1971.
48. Galante, J., Rostoker, W., and Ray, R.D., Physical properties of trabecular bone, Calcif. Tissue Res.,
5, 236, 1970.
49. Linde, F., Norgaard, P., Hvid, I., et al., Mechanical properties of trabecular bone. Dependency on
strain rate, J. Biomech., 24, 803, 1991.
50. Linde, F. and Hvid, I., The effect of constraint on the mechanical behavior of trabecular bone specimens
[see comments], J. Biomech., 22, 485, 1989.
51. Brear, K., Currey, J.D., Raines, S., and Smith, K.J., Density and temperature effects on some mechan-
ical properties of cancellous bone, Eng. Med., 17, 163, 1988.
52. Sedlin, E.D. and Hirsch, C., Factors affecting the determination of the physical properties of femoral
cortical bone, Acta Orthop. Scand., 37, 29, 1966.
53. Smith, J.W. and Walmsley, R., Factors affecting the elasticity of bone, J. Anat., 93, 503, 1959.
54. Evans, F.G. and Lebow, M., Regional differences in some of the physical properties of the human
femur, J. Appl. Physiol., 3, 563, 1951.
55. Dempster, W.T. and Coleman, R.F., Tensile strength of bone along and across the grain, J. Appl.
Physiol., 16, 355, 1960.
0266/ ch07/frame Page 132 Thursday, October 21, 1999 2:27 PM

133

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

8

A Hierarchical Approach
to Exploring Bone
Mechanical Properties

C. Edward Hofer, Barbara R. McCreadie, Erica A. Smith,
and Steven A. Goldstein

CONTENTS

I. Introduction ..........................................................................................................................133
II. Hierarchy of Bone Architecture...........................................................................................135
III. Whole-Bone Level ...............................................................................................................136
IV. Architectural Level...............................................................................................................137
V. Tissue Level..........................................................................................................................138
VI. Lamellar Level .....................................................................................................................140
VII. Ultrastructural Level.............................................................................................................141
VIII. Examples of Multilevel Investigations ................................................................................142
A. Trabecular Bone .............................................................................................................142
B. Cortical Bone..................................................................................................................143
IX. Computational Methods .......................................................................................................144
X. Concluding Remarks............................................................................................................146
Acknowledgments ..........................................................................................................................146
References ......................................................................................................................................147

I. INTRODUCTION

The skeleton, as an organ system, dually functions to perform a variety of mechanical and metabolic
activities. It provides the supportive framework enabling locomotion, protects other organs, and par-
ticipates in mineral homeostasis. All of these functions require the maintenance of mechanical integrity,
and compromises in this integrity may impact substantially on an individuals quality of life.
Like any other organ system, functional evaluations of its components are useful for clinical
diagnoses and for addressing hypotheses concerning fundamental operational mechanisms. In bone,
there are at least four primary motives for mechanical testing. First, structurefunction studies may
detail specic structural characteristics of bone to resolve the functional consequences of their
variation. The second reason is to determine the etiology and pathogenesis of disease to guide
clinical therapy and prevention. Typically, the functional consequences (i.e., bone fragility) are
clearer than the mechanisms that compromise mechanical integrity. The rst and second motives
are really complementary approaches to investigating bone mechanical properties. One recognizes
a functional decit and searches for the responsible characteristics. The other approach details
variations in bone mechanical properties to understand their functional implications. A third objec-
tive is to evaluate bone responses functionally to arthroplasty, fracture xation techniques, or other

0266/ ch08/frame Page 133 Thursday, October 21, 1999 2:33 PM

134

Mechanical Testing of Bone and the BoneImplant Interface

FIGURE 8.1

The hierarchical organization of bone is illustrated. Both the measurement of mechanical
properties and the interpretation of their values need to be considered within the context of structural scale.
(A) A radiograph of a whole rat femur. (B) Haversian bone light micrograph and (C) trabecular bone

-CT
image at the architectural level. (Courtesy of Nancy J. Caldwell.) (D) Haversian bone and (E) trabecular bone
light micrographs at the tissue level. (F) Light micrograph of several bone lamellae. (G) Scanning electron
microscopy images of collagen bers. (Modied from Marotti, G.,

Calcif. Tissue Int.

, 53S1, S47, 1993. With
permission.) (H) bone mineral at the ultrastructural level. (Modied from Mackie, I.G., Green, M., Clarke,
H. and Isaac, D.H.,

J. Bone Joint Surg.

, 71B, 509, 1989. With permission.)

0266/ ch08/frame Page 134 Thursday, October 21, 1999 2:33 PM

A Hierarchical Approach to Exploring Bone Mechanical Properties

135

reconstruction procedures. Last, mechanical tests provide input for computational models of bone
mechanics, adaptation, and repair. These categories are not mutually exclusive or exhaustive, but
provide a framework for the overwhelming diversity of investigations that rely on bone mechanical
measurements.

II. HIERARCHY OF BONE ARCHITECTURE

From a mechanical perspective, the most striking feature of bone is the hierarchical organization
of its architecture (Figure 8.1). Bone structural heterogeneity varies precisely with the scale of
magnication employed in its study. Based on the theory of continuous materials, bone hierarchy
can be arranged into (1) a whole-bone level (Figure 8.1A); (2) an architectural level, referring to
large volumes of cortical or trabecular bone tissue (Figure 8.1B and C); (3) a tissue level, largely
containing single trabeculae, single osteons, and cortical microbeams (Figure 8.1D and E); (4) a
lamellar level (Figure 8.1F); and (5) an ultrastructural level containing isolated molecular and
mineral components of bone (Figure 8.1G). The authors believe that bone mechanical properties
at any scale of organization are prescribed by the inherent properties of more microscopic scales.
Most importantly, this architectural paradigm shapes the way one thinks about bone mechanical
testing and properties.
The hierarchical structure of bone must be considered whenever one decides to evaluate bone
mechanically. The different hierarchical, structural elements contribute distinct characteristics to
mechanical properties measured at the more global level. These relationships must be considered
when an investigation employs bone mechanical testing.
Separation of the various levels of organization is a precarious endeavor and can best be
approached through continuous materials theory. Without resorting to mathematical rigor, a con-
tinuum is dened to have continuously distributed material properties without discrete local vari-
ations. This denition can be consistently applied to resolve levels of bone architecture. Two scales
can be distinguished if the global scale properties are independent of local variations in the
properties of the more microscopic scale. The structural difference must allow the microscopic
scale to approximate continuum properties when observed from the more global level. Disconti-
nuities in structure or material properties are not permitted. Specically, the microscopic scale must
appear to ll completely the volume of the global scale and to have continuously distributed
properties.
Compression testing of trabecular bone cubes illustrates an application of these continuum
principles. One may measure the stiffness and use it to compute an apparent (or effective) modulus
of the cube. By omitting the effect of trabecular architecture in the apparent modulus calculation,
the cube is modeled as a single solid continuous material. This is appropriate when the cube scale
is much larger than the scale of structural variations. Alternatively, one may use the detailed
trabecular structure to generate a nite-element model (assuming homogeneous tissue properties)
and calculate the elastic modulus of the trabecular tissue. This approach accounts for the trabecular
architecture, but assumes that microstructural tissue features like trabecular packets are sufciently
small that the tissue approximates a continuous material. Note that in each case, the results will
be very different because the characteristics which were estimated as continuous are different.
Clearly, the continuum concept has practical implications for bone mechanical testing. For
example, as increasingly smaller volumes of cortical bone beams are tested in four-point bending,
a threshold can be crossed where the measured modulus begins to vary with the size of the sample.

1

Hence, a practical limit can be dened where cortical bone no longer behaves as a continuum,
which allows a more microscopic level of organization to be explored. At a specic hierarchical
level, specimens are not typically tested at the upper threshold, but rather using a structural unit
that allows one to isolate specic properties of that level. For example, in cortical bone tissue,
single osteons and microbeams devoid of Haversian canals are often used; yet larger specimens
including few Haversian systems are still considered to be at this level.

0266/ ch08/frame Page 135 Thursday, October 21, 1999 2:33 PM

136

Mechanical Testing of Bone and the BoneImplant Interface

Most importantly, the specic hypothesis in question dictates the appropriate hierarchical levels
at which to evaluate bone mechanical integrity. Consider two osteogenesis imperfecta (OI) variants:
(1) a quantitative defect where normal collagen is produced, but in reduced amounts; (2) a qualitative
defect where the triple helix is kinked due to an amino acid substitution in one alpha chain.

2

Bending
tests of whole femurs, along with geometric measures, would reveal that the material properties
were compromised in both variants, possibly leading to increased fragility. Ultrastructural tensile
testing of individual collagen molecules would distinguish the mechanisms responsible for the
decreased properties. Therefore, the two scales of testing are appropriate for addressing different
hypotheses.
The purpose of this chapter is to emphasize the importance and relevance of a hierarchical
approach to the mechanical evaluation of bone. While several investigators have taken a hierarchical
approach to testing bone mechanically, the authors have selected a sample of studies from their
laboratory which have broadened our understanding of bone structural hierarchy and mechanical
integrity. The illustrations are organized as a function of structural scale.

III. WHOLE-BONE LEVEL

Mechanical testing at the whole-bone level measures properties of the entire bone as a structure,
which incorporates the properties of the materials that compose the whole bone, as well as its
internal and external geometry. At this hierarchical level, most specimens include both cortical and
trabecular bone, and therefore contain multiple architectures. These specimens can be either entire
excised bones (i.e., a whole femur) or large portions of an entire bone (i.e., a proximal femur). The
mechanical behavior of whole-bone specimens most closely approximates the behavior of these
structures

in vivo

. In testing whole-bone specimens, it is assumed that the various architectural
features are insignicant as individual entities. Hypotheses regarding the structural mechanical
behavior of an intact bone, as well as how this may be altered due to aging, various therapies, or
genetic mutations, may be addressed at the whole-bone level of hierarchy.
Mechanical testing of whole vertebral bodies has been used in conjunction with quantitative
computed tomography (QCT) images in an effort to understand mechanisms of localized trabecular
failure and ultimately vertebral fracture. This approach is potentially useful for improving prediction
of vertebral fracture risk. Cody et al.

3

found strong correlations between regional bone mineral
density (rBMD) measurements from QCT and whole vertebrae failure load. McCubbrey et al.

4

determined a relationship between rBMD measurements and whole vertebral static and fatigue
properties.
Mechanical testing of whole femora has proved to be a valuable technique for quantifying
alterations in mechanical integrity of cortical bone caused by genetic mutations or drug treatments.
These experiments can provide insight regarding how the components of bone extracellular matrix
contribute to its functional capacity. Evaluation of whole-bone mechanical properties should be
accompanied by characterization of the bone geometry, so deductions regarding potential mecha-
nisms can be made. In the authors laboratory, mechanical testing of whole femora from transgenic
mice is typically performed in four-point bending, and three-dimensional specimen geometry is
quantied using digital three-dimensional micro-CT (

-CT) images. For example, Tseng



et al.

5

determined that local production of human growth hormone in transgenic mice stimulated increased
femoral bone formation, as indicated by increased cross-sectional geometry (cross-sectional area
and moment of inertia), without an associated increase in whole bone mechanical properties. This
suggested that the material properties of the bone matrix were decreased due to the local growth
hormone production. In addition, interleukin-4 (IL-4) overexpression was demonstrated to cause
signicant decrease in both mechanical parameters and cortical thickness, as well as an overall
shape change, in the middiaphysis of femurs from lck-IL-4 transgenic mice.

6

Ovariectomy was
found to have a signicant effect on structural properties of whole femurs from osteocalcin-decient
mice, but not wild-type controls.

7

0266/ ch08/frame Page 136 Thursday, October 21, 1999 2:33 PM

A Hierarchical Approach to Exploring Bone Mechanical Properties

137

Mechanical testing on the whole-bone level is also a useful tool to analyze the performance of
various joint replacement or fracture xation techniques, hardware, or cements. The xation integ-
rity of ve commercially available cannulated screw systems were evaluated by Rouleau et al.

8

through mechanical testing of the screws in a femoral head. Moore et al.

9

analyzed the mechanical
integrity of a novel

in situ

-setting calcium phosphate cement system to reinforce compression screw
xation of unstable intertrochanteric fractures. In a similar study, Norian SRS

cement (Norian
Corp., Cupertino, CA) was evaluated for improvement in strength and stability of femoral neck
fracture xation.

10

In this study, experimental fractures were xed with cannulated cancellous screws
with or without Norian SRS, and were then mechanically tested in fatigue.
Mechanical testing on the whole-bone level of bone hierarchy can be used to determine
properties of the complete structural component, which are most directly related to the functional
capacity of the specic bone

in vivo

. This technique has been successful in determining the effects
of various factors such as aging, pharmacological treatments, genetic mutations, or surgical tech-
niques on the mechanical behavior of whole bones. Due to the complex geometric and material
characteristics of whole bones, however, it is not reasonable to calculate directly material-level
parameters such as modulus, strength, or stresses in the bone matrix. Mechanical testing solely at
the whole-bone level cannot identify particular alterations of bone architecture or extracellular
matrix; therefore, specic mechanisms of alterations in mechanical properties must be addressed
at more microscopic levels. Similarly, hypotheses regarding damage or fracture initiation, mechan-
ical stimuli to bone cells, or integrity of bone matrix must be investigated at lower levels of bone
hierarchy. On the other hand, adaptations to disorders at the cellular or extracellular matrix level
may best be evaluated by whole-bone tests. This rationale is based on the possibility that the
objective function of the adaptation is to restore whole-bone functional properties.

IV. ARCHITECTURAL LEVEL

The architectural level of bone hierarchy represents specimens dominated by a single type of internal
architecture (e.g., trabecular, Haversian/interstitial, circumferential, etc.). In this case, the mechan-
ical properties of a test specimen depend on the arrangement of the dominant type of architecture,
as well as the properties of the bone material. The upper limit on specimen size is dictated by the
geometry of the bone region from which it is extracted. The lower limit on specimen size is more
difcult to dene, and is likely different for cortical and trabecular bone. This threshold is dened
as the minimum specimen size where the local variations in microstructural properties do not create
large variations in properties at the more macroscopic level, as stated by continuum theory. In other
words, the specimen must be sufciently large that the microstructural component does not cause
large, discrete uctuations in mechanical properties. This question was addressed analytically for
trabecular bone by Harrigan et al.

11

who concluded that in three dimensions a specimen must contain
ve intertrabecular lengths to be considered a continuum.
Mechanical testing of trabecular bone specimens on the architectural level is commonly per-
formed using cubes or cylinders of an appropriate size to be considered a continuum. Analogous
specimens of cortical bone at the architectural level would be large portions of cortical bone
containing many Haversian systems and interstitial regions. In this case, the lower limit on cortical
bone specimens is derived from local uctuations in properties due to the Haversian and interstitial
systems. As specimens become smaller, the size of Haversian and interstitial structures relative to
specimen size increases, resulting in discontinuities which are too large to be neglected. The
continuum assumption, therefore, is only valid if a sufcient number of osteons or interstitial
systems are included in the specimen, although this threshold has not been precisely dened.
Architectural-level testing of Haversian bone has not been the focus of the authors laboratory,
although others have studied these properties.

12-14

The primary advantage of testing bone specimens at the architectural level, as compared with the
whole-bone level, is the capability of isolating effective properties of a single type of architecture.

0266/ ch08/frame Page 137 Thursday, October 21, 1999 2:33 PM

138

Mechanical Testing of Bone and the BoneImplant Interface

Mechanical characterization of these specimens allows one to estimate of the contribution of each
architectural type to whole-bone behavior, and a description of how this relationship may be altered
due to aging and disease.
At the architectural level of bone hierarchy, many fundamental questions regarding variations
in mechanical behavior of trabecular bone due to age, disease, gender, and anatomic location have
been addressed. For example, Goldstein et al.

15

tested cylinders of human tibial metaphyseal
trabecular bone in compression, and demonstrated strengthening and stiffening of bone in areas of
maximum load bearing. Ciarelli et al.

16

tested trabecular bone cubes from various metaphyseal
locations in compression in three orthogonal directions. These results showed great variability in
properties within each metaphysis, as well as between metaphyseal regions. These data were
compared to specimens from canine distal tibiae in a study by Kuhn et al.

17

Results indicated that
canine bone displays a lower modulus but higher ultimate strains when compared with human bone,
and denes limitations in using canine bone to model the mechanical behavior of human tissue.
There has also been a large research effort in the authors laboratory to dene a strategy to
predict fracture risk using various imaging modalities. This has involved searching for relationships
between mechanical properties of trabecular bone and measures of density or microstructure.
Ciarelli et al.

16

found varying correlations between QCT data from whole metaphyseal regions and
orthogonal mechanical properties of trabecular bone cubes. Cody et al.

18

investigated the usefulness
of dual-energy X-ray absorptiometry (DEXA) and QCT in predicting local mechanical properties
of trabecular bone cubes. This study found that either density measurement could explain 30 to
40% of variance in modulus, and 50 to 60% of the variance in ultimate strength. Goulet et al.

19

tested trabecular bone cubes from human metaphyses in three orthogonal directions and compared
these results with morphological measurements obtained from

-CT. This study determined that
bone volume fraction or mean-intercept length measurements alone could only predict about 50 to
60% of variation in mechanical properties, while combining these parameters improved predictions
to explain about 90% of mechanical property variations. Taken together, these studies demonstrate
that scalar measurements of density alone have some capability in predicting the mechanical
behavior of trabecular bone, but indicate that directional measurements of architecture are necessary
for more complete characterizations.
Mechanical testing of trabecular bone specimens at the architectural level has provided valuable
insight into the behavior of trabecular bone as a function of age, gender, disease, and location. This
information can be applied to results from the whole-bone level, and assists in understanding of
the contributions of each particular type of architecture to behaviors and pathologies observed

in
vivo

. Data derived from studies at the architectural level also indicate that the organization of
trabecular bone signicantly impacts its mechanical properties. Unfortunately, these studies cannot
address the contributions of bone tissue properties to mechanical integrity at either this level or the
whole-bone level. However, it is conceivable that these properties may be altered in specic disease
states or throughout life. It is therefore necessary to quantify bone properties at lower levels of
hierarchy, in order to provide a more complete characterization of its behavior.

V. TISSUE LEVEL

The tissue level describes the properties of bone matrix independent of cortical and trabecular
architecture. Tissue-level specimens always include lacunae, and may also include lamellae, cement
lines, or Haversian canals, depending on how the specimen is selected. Creating equivalent descrip-
tions of cortical and trabecular tissue-level specimens is difcult, although analogous denitions
can be formulated. Trabecular bone tissue includes a number of trabecular packets, but does
incorporate the distinctive porous architecture generally associated with trabecular bone. The
geometry of trabecular bone places a practical upper limit on the size of specimens, which should
not be larger than a single trabecular strut. Once the specimen becomes any larger, its structural
organization plays a role, and isolating tissue-level properties becomes mathematically complex.

0266/ ch08/frame Page 138 Thursday, October 21, 1999 2:33 PM

A Hierarchical Approach to Exploring Bone Mechanical Properties

139

In cortical bone, a specimen at the tissue level generally incorporates several portions of interstitial
regions or osteons, or maybe a single osteon. Inclusion of more than one Haversian canal makes
direct evaluation of tissue properties difcult without complicated calculations.
All tissue-level specimens include lacunae and extracellular matrix, which may or may not be
in the form of lamellae. Cement lines are generally included in trabecular specimens. Depending on
the methods used, Haversian canals and/or cement lines may be contained in cortical test samples.
When Haversian canals are included, the calculations should account for the canal and therefore
measure the properties of the matrix excluding the canal. The cement lines, if included, are generally
the feature which determines the minimum size of specimens. In this case, specimens must contain
several cement lines, and therefore several trabecular packets, interstitial packets, or osteons. How-
ever, in the case of woven bone tissue or specimens which are obtained without cement lines, the
lacunae and their associated cells are the features which will rst contradict the continuum assump-
tion. In these cases, it is reasonable to assume that a specimen must only be large enough to contain
several lacunae in each direction. When designing experiments or evaluating results, it is particularly
important at this level to understand exactly which features are included in the test specimens. In
addition, conclusions at this level must be made with the understanding that there are many com-
ponents of the tissue, such as lacunae geometry and number, lamellar thickness, collagen bril
organization, mineral crystal aggregates, etc., which inuence the tissue-level properties.
Several research questions have been addressed at this level. It has been used to explain
differences in observed properties at the architectural or whole-bone levels. Specically, these
experiments were designed to determine whether the tissue properties contribute to differences seen
in trabecular bone cube (architectural level) or whole-bone mechanical tests. Experimental studies
have also been conducted to examine the microstructural level specically for differences resulting
from age, gender, and disease.
Most of the work in the authors laboratory at this level has involved testing microbeams, bone
tissue specimens cut into approximately 120

120

1200

m parallelopipeds. Custom-made
micromilling and micro-testing machines were developed specically for these experiments. Early
testing relied on three-point bending,

1,17

while later experiments employed four-point bending.

5,20-25

Several studies investigated various properties using monotonic tests,

1,5,17,23

while Jepsen et al.

22

included yield, failure strength, and measures of fatigue. Trabecular and cortical tissues from
humans, animals, and transgenic animals have been tested. Limitations of this particular method
of testing bone tissue include machining artifacts with both trabecular and cortical specimens, and
the required removal of the outer layers of tissue (likely less mineralized) from trabeculae to obtain
regularly shaped test specimens.
Several signicant results were found for a wide range of research questions. Kuhn et al.

17

found that cortical tissue had a higher modulus than trabecular tissue in humans. Choi and
Goldstein

21

conrmed that human cortical tissue had a higher modulus than trabecular tissue,
and found that the trabecular tissue had a lower fatigue strength. Choi et al.

1

found a correlation
between modulus and specimen size. Wong et al.,

20

Tseng et al.,

5

Jepsen et al.,

22

and Tseng and
Goldstein

23

have shown that alterations in gene expression affect the mechanical properties of
bone at the tissue level.
There are several other methods which have been utilized to determine properties at the tissue
level in both trabecular and cortical bone. Townsend et al.

26

conducted buckling tests of single
whole trabeculae from human subchondral bone, without machining a regularly shaped specimen.
Rho et al.

27

used single trabeculae and machined cortical bone of similar size, tested them mechan-
ically in tension, and incorporated ultrasonic measurement of properties. Mente and Lewis

28

sub-
jected single trabeculae to cantilever tests, and with the aid of nite-element models of the trabe-
culae, calculated the modulus of the tissue. Ascenzi and co-workers

29,30

utilized single osteons
machined from cortical bone to investigate cortical tissue properties in a variety of testing modal-
ities, including tensile

29

and bending.

30

To the authors knowledge, these single-osteon studies are
the only tissue-level experiments that have not included cement lines in the test specimens.

0266/ ch08/frame Page 139 Thursday, October 21, 1999 2:33 PM

140

Mechanical Testing of Bone and the BoneImplant Interface

Regardless of the methods used, specimen preparation and testing is extremely difcult at the
tissue level due to the size and delicacy of the specimens. It is important to understand thoroughly
the limitations imposed by the methods, the accuracy of any measuring techniques, and any artifacts
that may result from creating the specimens. Specimen size has been found to have a signicant
impact on the calculated modulus, so it is important to use similarly sized specimens throughout
the study and clearly state the sizes used.
Despite the limitations, many conclusions have been drawn from the wide range of studies
conducted at this level. Many studies have been able to deduce the contribution of specic extra-
cellular matrix proteins to the tissue properties of bone, separating the effects on the tissue properties
from compensatory modications at the architecture or whole-bone level. Several studies suggest
that trabecular bone tissue is less stiff than cortical bone tissue. Throughout testing at this level,
the properties of bone tissue and reasons for their variation have been claried.
The specic features that inuence tissue properties are still unknown. These may include the
geometry and properties of the lamellae, geometry of the lacunae, and the geometry and properties
of cement lines. In addition, the structure and properties of the components of the lamellae, including
the collagen bril stiffness and orientation, the mineral aggregate size and location, and the number
and orientation of the canaliculi are likely to inuence properties obtained at the tissue level. In
particular, the observed difference between trabecular and cortical bone tissue suggests that there
is a fundamental difference in one or more of these features. Explanations for the differences found
at the tissue level are, therefore, sought at more microstructural levels.

VI. LAMELLAR LEVEL

The lamellar level of bone resolves mechanical properties independent of osteocyte lacunae and
all microstructural features of remodeling except the lamellae themselves. The upper limit of this
scale is dened by the maximum tissue interval that does not incorporate one of the aforementioned
structural discontinuities. The lower threshold depends on the scale of bone mineral crystal aggre-
gates, collagen brils, and canaliculi. These limits were created for secondary tissue, but can also
describe other bone tissues if the upper boundaries are adjusted for the absence of secondary
microstructures.
As with other levels, a wealth of research questions can be approached through the lamellar
scale. The most fundamental concerns are the property values at this level as a function of age,
gender, anatomical location, and disease. One can explore the matrix mechanical consequences of
a bone protein deletion or alteration in a transgenic animal. One can also compare primary and
lamellar bone properties to determine if remodeling may alter the physical environment of the cell.
The strength of lamellar-level measured mechanical properties is their independence from the
contributions of bone microstructure. Importantly, lamellar-level studies have direct implications
for cell behavior because the mechanical properties are evaluated in the neighborhood of the cell.
These extracellular matrix properties can be used for predictive models of tissue behavior in
adaptation and remodeling.
The authors laboratory has been characterizing bone tissue as the medium through which cells
receive mechanical signals. The desire is to understand how structure and material properties
prescribe bone mechanical integrity and transduce mechanical signals to elicit a bone cell response.
To begin addressing these issues, a unique tool called nanoindentation has been used to test
hypotheses of bone structurefunction at the lamellar level. Nanoindentation uses depth-sensing
technology to measure elastic modulus with submicron precision. Recently, the technique has been
validated for

in vitro

bone lamellae measurements at 5

m resolution.

31

Equipment constraints and
the scale of the ultrastructure limit one to measuring the properties of several lamellae at once.
Nanoindentation has permitted testing of hypotheses that were previously unapproachable with
direct experiments. Initially, the authors investigated the relationship between microstructural het-
erogeneity and lamellar mechanical properties.

32

Next, they quantied elastic modulus differences

0266/ ch08/frame Page 140 Thursday, October 21, 1999 2:33 PM

A Hierarchical Approach to Exploring Bone Mechanical Properties

141

between osteonal, interstitial, and trabecular tissue in four areas of high fracture incidence.

33

Finally, they measured the elastic modulus variation with age and gender at the lamellar level
in order to identify microstructural properties possibly responsible for age- and gender-related
reductions in mechanical integrity.

34

The results of these investigations have provided a host of new insights into bone mechanical
integrity. Differences in properties between microstructures are consistent with known patterns of
homeostatic tissue turnover and greater mineralization, reecting increased tissue age.

32

Elastic
modulus was found to vary with location, suggesting that the local response to the mechanical or
metabolic environment determines the elastic properties.

33

The measured trabecular lamellar prop-
erties were also inconsistent with whole-bone fragility fracture patterns,

35

suggesting that other
mechanical parameters are more important in these maladies. Elastic modulus and hardness did
not correlate with age or gender. If lamellar post-yield, fracture, and fatigue properties are also
found to be independent of age and gender, these data would suggest that age- and gender-related
fragility increases involve the regulation of tissue mass and organization and not the inherent
properties of the extracellular matrix.

34

Whereas nanoindentation has recently emerged as a technique to test human bone mechani-
cally,

36-39

microhardness

40,41

and scanning acoustic microscopy (SAM)

42

have traditionally been
used to explore properties at the lamellar level. Microhardness integrates several complex material
behaviors and is not a true material property. Therefore, it is qualitatively distinct from elastic
modulus and difcult to interpret. SAM measures acoustic reectivity which is related to the
acoustic impedance and the elasticity coefcient of the same direction. Hence, tissue anisotropy
can be characterized more completely. SAM also provides a continuous array of data across a
surface compared with the discrete sampling required by indentation techniques. Unfortunately,
theoretically based elastic modulus measures have been elusive.

42

Empirically based SAM and
nanoindentation elastic modulus measurements do not correlate strongly at the lamellar level.

43

Ascenzi and co-workers

44

have admirably attempted to isolate single lamellae and test them in
tension, but the experiment has proved technically challenging.
Lamellar-level properties have been shown not to vary consistently with age and gender,
suggesting that age- and gender-associated fragilities are more heavily determined by bone mass
and organization. However, lamellar properties do vary with anatomical location, possibly indicating
different mechanical or metabolic demands. Equally important, bone lamellar properties may reect
mineral content and tissue maturity and are likely inuenced by remodeling rate.
As bone is characterized at increasing microscopic levels, one must begin to understand the
contributions of matrix protein and mineral to anatomical variations in lamellar properties. One
may also explore whether there are components of the extracellular matrix that dene a cellular
mechanical environment which predispose the bone cell to specic remodeling behaviors. Do
collagen bril properties change with anatomical location and increase cell sensitivity to cyclic
loading? Do variations in proteoglycan (PG) exibility ultimately change the uid shear environ-
ment within canaliculi and lacunae? These are the questions for the next hierarchical level. One
must begin to understand the relationship between chemical and mechanical properties of bone
matrix molecules and the resulting mechanotransduction consequences for the cell.

VII. ULTRASTRUCTURAL LEVEL

Ultrastructural level describes the molecular network of proteins, glycoproteins, and minerals. It
is important to emphasize that the extracellular matrix is no longer characterized as continuous,
but rather as a diverse group of mechanical elements interacting with each other and the surface
of the cell. At this level, it is appropriate to test the mechanical properties of matrix constituents
like collagen brils and bone mineral spheroids. Additionally, the interaction between components
can be measured. The upper threshold of the ultrastructural level is practically limited by the scales
of mineral aggregates, molecules, and macromolecules like a type I collagen bril. The lower limit

0266/ ch08/frame Page 141 Thursday, October 21, 1999 2:33 PM

142

Mechanical Testing of Bone and the BoneImplant Interface

is dened by the scale at which these macromolecules and mineral clusters fail to behave as
continuous elements. It is difcult to specify a lower threshold for this level as much of the
information about architectural and mechanical property variations remains unknown. Moreover,
the structural heterogeneity of matrix constituents implies that different components will no longer
approximate continuum behavior at different scales of testing. Currently, all molecular studies are
grouped as ultrastructural, and attempts to distinguish more microscopic scales should await more
detailed characterizations of this level.
Bone ultrastructural components are subcellular in scale and provide the nal interface between
the external mechanical environment and the cell. Characteristics of more global levels are derived
from the molecular properties intrinsic to this scale. However, the ultrastructural level concerns
mechanical characteristic of molecules, which are by denition chemical characteristics. Therefore,
the hypotheses generated at this scale must describe molecular mechanical behavior as a function
of fundamental chemical phenomena. Hydration, salt bridging, homophilic association, and covalent
modication are just some of the interactions that will inuence molecular mechanics.
Ultrastructural evaluation refers to the direct mechanical characterization of bone components
specic to this scale. As dened in this architectural framework, ultrastructural mechanical testing
does not include investigations that quantify the contributions of ultrastructural features (i.e.,
collagen ber orientation, mineral content, canaliculi volume) to higher-order mechanical proper-
ties. Reconsider the OI variant with reduced collagen synthesis. Whole-bone testing is a higher-
order test which may reveal that OI bones have weaker material properties when collagen content is
reduced. Tensile testing of collagen brils would reveal that ultrastructural mechanical properties were
unaffected. Again, the two scales of testing are appropriate for addressing different hypotheses. None-
theless, higher-order studies have provided valuable insights into the effect of collagen ber orientation
on osteonal elastic properties,

45

cross-linking concentration on whole-bone strength

46

and the contri-
butions of collagen and mineral to both elastic-plastic behavior

47

and tissue anisotropy.

48,49

Basic investigations may quantify the elastic properties of mineral aggregates and type I
collagen brils. Additionally, one may begin to understand the relative effects of covalent cross-
links and collagen-binding proteins on collagen bril stiffness. Other questions involve noncollag-
enous matrix constituents. Can hydration alter the ability of PGs to transmit hydrostatic or dilata-
tional strains? Ultimately, the hope is to understand the cell mechanical environment as a function
of these molecular properties and determine their inuence on metabolism and gene expression.
Unfortunately, technology cannot maintain the pace of investigator creativity, and the ability
to test hypotheses is often limited by the inadequacy of available experimental techniques. As a
result, mechanical testing at the ultrastructural level has been rudimentary at best. To date, the most
compelling example of ultrastructural mechanical experiments is single-collagen-molecule tensile
testing performed by Luo and colleagues

50

using optical tweezers. Measuring the stiffness of a
single collagen molecule is a technical tour de force, and represents the promise of technological
advances that will aid understanding of the biophysical issues critical at this level of magnication.

VIII. EXAMPLES OF MULTILEVEL INVESTIGATIONS
A. T

RABECULAR

B

ONE

In some studies, the combination of several levels of experimental analysis can explain relationships
among the properties at various levels. Often, an investigator will nd a difference in properties at
one level, and examine a more microscopic level to explain the difference. Other times, one study
will be designed

a priori

to evaluate the properties at various levels of hierarchy and discover many
relationships between the levels.
One of the authors current research programs has used both of these methods. It is a series of
studies focused on the mechanical properties of human trabecular bone at various hierarchical
levels, emphasizing how the properties change with age or differ between males and females. The

0266/ ch08/frame Page 142 Thursday, October 21, 1999 2:33 PM

A Hierarchical Approach to Exploring Bone Mechanical Properties

143

following summary will describe a subset of these experiments, looking at mechanical tests from
the architectural, tissue, and lamellar levels.
Coordinated experiments at the architectural and tissue levels have been reported in part.

24,25

Using the same specimens for both levels allowed comparisons between age and gender groups,
as well as analysis of the contribution of the tissue-level properties to the architectural level. Instead
of simply comparing the trends observed at various levels as one could do with any set of studies,
the authors were able to calculate correlations in the properties between the two levels and directly
calculate the contribution of the tissue properties to the architecture level properties. However, it
also placed additional requirements on the testing, such as the need for nondestructive testing so
the same tissue could be used at multiple levels.
The more macroscopic testing was conducted on trabecular bone cubes, loaded in compres-
sion to nondestructive strain levels. Then microbeams were created from trabecular struts obtained
from the same cubes, and tested to failure in four-point bending. Although the trends showed
that the tissue modulus increased from the 55- to 65- to the 75- to 85-year-old groups, the trend
in the cube modulus was to decrease with age. Further analysis showed no correlation between
the tissue modulus and the cube modulus, when including males and females in the same two
age groups (unpublished analysis). These results led to the conclusion that the fragility seen in
osteoporosis and aging is more related to changes in bone mass and organization than it is to
the fundamental properties of the bone tissue.
Following testing at the architectural and tissue levels, a large unexplained variation in the
properties of the trabecular bone tissue remained. Therefore, the authors attempted to explain these
variations by exploring the lamellar level. This encouraged development of nanoindentation tech-
niques for bone tissue, using specimens from the proximal femur. These studies revealed that the
lamellar elastic modulus and hardness did not differ with age or gender.

34

Unfortunately, the authors
were unable to use the same specimens as those used in the studies previously described, and
therefore, could not make direct correlations between the various levels of microstructure.
This research program demonstrates a series of experiments at increasingly microscopic levels
designed to explore the etiology of bone fragility associated with osteoporosis and aging. The
strength of evaluating mechanical properties on two levels in succession is that direct relationships
can be established between properties of the two levels. The clarity of these relationships are further
enhanced by conducting these bi-level experiments on the same tissue samples. Combined results
of testing at two levels demonstrate that tissue modulus does not signicantly inuence architectural-
level modulus. Since lamellar-level properties do not differ with age and gender, while the whole-
bone fracture incidence does, the authors also suggest that the lamellar-level changes may not play
a primary role in age- and gender-related bone fragility. Interestingly, material properties at tissue
and lamellar levels were unable to explain whole-bone fragility increases. The same specimens for
both the tissue and lamellar levels were not used, so the specic relationship between tissue- and
lamellar-level properties remains unknown. However, the results as a whole suggest that the
distribution and organization of trabecular bone architecture are responsible for increases in bone
fragility associated with age and gender. Using a multilevel approach, the authors have rened
understanding of fragility increases associated with age and gender by learning that bone mechanical
integrity is compromised at a specic hierarchical level. Future investigations into mechanisms of
increased bone fragility should focus on trabecular architecture.

B. C

ORTICAL

B

ONE

Analysis of cortical bone on multiple hierarchical levels is capable of providing specic information
on how bone properties are altered by disease, and how this translates into decreased functional
capacity at the whole-bone level. An example of this type of analysis was performed using Mov13
mice, which serve as a model of human osteogenesis imperfecta type I. Heterozygous Mov13 mice
produce as much as 50% less type I collagen than littermate controls.

51

Although less collagen is

0266/ ch08/frame Page 143 Thursday, October 21, 1999 2:33 PM

144

Mechanical Testing of Bone and the BoneImplant Interface

produced by these mice, the collagen is still normal in structure. Characterization of mechanical
properties of bone from these animals was performed at the whole-bone and tissue levels.
Initially, whole-bone four-point bending tests were performed to failure using femora from
Mov13 mice and littermate controls at 8 and 15 weeks of age.

52

The bones were also scanned on
the

-CT system, and diaphyseal cross-sectional geometry was measured. Results indicated that
the genetic mutation associated with Mov13 caused signicant decreases in stiffness and postyield
deformation, and signicant increases in failure load and cross-sectional geometry at 15 weeks of
age. It is important to note the progressive decrease in stiffness and postyield deformation despite
an increase in cross-sectional geometry, which indicates changes in tissue-level properties. This
study identied a potential adaptive response to these tissue-level alterations, in which an increase
in cross-sectional geometry substantially improved failure properties.
Jepsen et al.

53

investigated potential mechanisms responsible for alterations in postyield behav-
ior of Mov13 whole bones. Whole femora were again tested in four-point bending, and ash density,
tissue porosity, collagen ber orientation, and tissue structure were measured. Although whole-
bone mechanical test results showed no signicant differences in yield and failure load at 8 weeks
of age, postyield deformation was signicantly decreased in specimens from Mov13 animals.
Visualization of fracture surfaces demonstrated that control bones appeared to fail by cleavage
between lamellae, while this mechanism was not effective in Mov13 tissue. Mov13 animals also
exhibited a 22% decrease in total collagen content, a two-fold increase in tissue porosity, and an
80% increase in area fractions of woven bone when compared with littermate controls. Taken
together, these data indicate that energy absorption mechanisms of various components of the
architecture, and maybe even the extracellular matrix, were not functional, resulting in decreased
resistance to crack propagation. These ndings demonstrate important relationships between whole-
bone mechanical integrity and features at the tissue and lamellar levels.
Hypotheses regarding alterations in mechanical properties at the tissue level due to the Mov13
genetic mutation were then addressed by Jepsen et al.

22

In this study, microbeams were milled from
the femoral middiaphysis, and then tested either monotonically to failure or cyclically in four-point
bending. Monotonic testing revealed that the mechanical integrity of Mov13 tissue was signicantly
decreased when compared with controls. Similar reductions in fatigue properties were exhibited
by Mov13 tissue, but this was possibly related to the decrease in strength of these specimens. In
addition, fatigue testing data showed that Mov13 bone tissue began accumulating damage earlier
than control tissue. These alterations in tissue properties from Mov13 specimens may be due, in
part, to increased woven bone fractions. In general, these ndings at the tissue level corroborate
suggestions of decreased resistance to fracture made by previous results on the whole-bone level.
These studies, taken together, begin to address mechanisms of bone quality alteration due to a
genetic alteration that mimics one form of the disease osteogenesis imperfecta, and may lead to
improved strategies for treatment. Use of the Mov13 mouse model also provides valuable insight
into the contributions and function of type I collagen in long bones. This series of experiments
emphasizes a complementary approach to understanding bone structurefunction relationships, as
well as the etiology and pathogenesis of disease. These investigations explore a type I collagen
mutation that translates into alterations in tissue-level properties, which leads to whole-bone func-
tional decits. Results of studies conducted at the whole-bone level directed hypotheses for more
microscopic levels. This systematic progression through the scales of bone hierarchy allowed a
more complete characterization of the role of type I collagen in bone mechanical integrity.

IX. COMPUTATIONAL METHODS

Computational methods play an important role in the analysis of bone properties. Investigations
with a large number of parameters can rst be conducted using a computer model, and the analytical
results used to determine an experimental design including only the most promising treatment groups.
In other instances, experiments are difcult or impossible to conduct, or the cost is prohibitive.

0266/ ch08/frame Page 144 Thursday, October 21, 1999 2:33 PM

A Hierarchical Approach to Exploring Bone Mechanical Properties

145

Computational analyses can allow greater consistency, since an identical model, free from experimental
and interindividual variation, can test various experimental treatments. If feasible, experimental results
are generally preferred to computer-generated analyses because fewer assumptions and simplica-
tions are required. However, analytical models are critical tools in the study of bone mechanics.
This section will focus on the use of computational methods which have improved understanding
of structurefunction relationship in bone.
Material properties of bone can be estimated at many hierarchical levels using a variety of
computational techniques. There are two basic methods that employ a combination of computational
and experimental analyses to evaluate these properties. The rst method evaluates material properties
at a single hierarchical level based on a mechanical test and a mathematical model of the experiment.
In this case, the mechanical test by itself is not enough to determine material properties directly
because the test specimen is nonuniform in shape and/or contains inherent architecture. By assuming
that the material properties of the specimen are homogeneous and uniform, and by obtaining the
geometry of the test specimen by some imaging means, a nite-element model is produced with a
preliminary set of material properties. The model is generally linear, resulting in a linear relationship
between the property to be measured and the experimental result. The nite-element model is solved
using experimental loads to obtain deformation, or vice versa. The preliminary properties are then
scaled so the computational model results match the experimental results. Although some experimental
artifacts are avoided (such as those induced in the machining process) using this approach, there are
additional imaging errors in acquiring the geometry needed to build the model.
Mente and Lewis

28

used this method to calculate tissue properties from mechanical tests of
single trabeculae. They dissected individual trabecular struts from human bone, which were tested
as cantilever beams. Geometry of the specimen was obtained by a serial grinding and imaging
technique. A nite-element model was then created for each trabecular specimen. The experimental
displacement results were compared with the results of the nite-element analysis using the exper-
imentally applied loads, and the properties of the trabecular bone tissue were calculated.
Second, it is also possible to calculate mechanical properties at a given hierarchical level from
the properties at another level and known geometry. In this case, which will be called the bi-level
method, the properties at either the more macroscopic or more microscopic level may be known,
but the geometry must be from the more macroscopic level. For example, if one wanted to calculate
the properties of trabecular bone tissue based on a trabecular bone cube test (architectural level),
the geometry of the trabecular bone cube must be available. In analyses using this method, there
are many assumptions that must be acknowledged. First, one assumes that the properties of the
more microscopic level are homogeneous and uniform throughout the region of the more macro-
scopic level analyzed. Second, the analysis is generally linear, so errors associated with the exper-
imental determination of properties at one level will propagate to the other level. Finally, inaccurate
geometry will result in inaccurate material properties.
One example of this bi-level method is a study by Guldberg et al.

54

using biopsied tissue from
a hydraulic bone chamber in dogs. Finite-element models were constructed based on geometries
obtained from

-CT. They then compared the experimental results from mechanical tests of trabe-
cular bone cubes with the nite-element results in an effort to calculate tissue modulus. Another
example is Ladd et al.,

55

who calculated the tissue modulus of human vertebral trabecular bone.
Finite-element models were created from synchrotron tomography images. Experimental results
from mechanical testing of the trabecular bone cubes were compared with the nite-element results,
which again allowed calculation of the tissue properties.
A variation of the bi-level method is to develop mathematical expressions of the properties at
one level based on measures of the structure and knowledge of the properties at the next most
microscopic level. An example of this method is a study by Zysset et al.

56

Mean intercept length
and average bone length, measures of the structure at the architectural level, were obtained from

-CT images of trabecular bone cubes. They calculated an elastic tensor for the bone specimen
using the homogenization method, and then the closest orthotropic tensor using an optimization

0266/ ch08/frame Page 145 Thursday, October 21, 1999 2:33 PM

146

Mechanical Testing of Bone and the BoneImplant Interface

procedure. There was a strong relationship between the structural measures (mean intercept length
and average bone length) and the orthotropic elastic tensor. These results suggest that the properties
of bone at the architectural level can be estimated based on measures of the structure, values of
the tissue properties, and the relations they reported.
A second variation of the bi-level approach is to assume a simple geometric structure for the
architecture of trabecular bone. Finite-element models can be developed based on the idealized
structure and used to estimate the properties of the bone at the architectural level. Again, properties
of the bone at the tissue level must be assumed. This method is often used to determine how changes
in the structural organization of bone trabeculae will affect the properties at the architectural level.
One example is a study by Gibson

57

in which trabecular bone is modeled as a cellular material
similar to foam. Using different models for the unit cell, which represent the different possible
trabecular architectures, Gibson was able to simulate the compressive behavior of the bone. In fact,
the asymmetric models were able to match the experimental results from previous studies. Hollister
et al.

58

investigated the use of regularly shaped unit cells to model trabecular architecture. Using a
strut model and a spherical void model, they found that the tissue strain results were very dependent
on the geometry of the model, particularly at low tissue volume fractions. Therefore, they suggested
that a single model cannot be used to estimate the properties of all regions of trabecular bone.
Instead, separate microstructural models may be required for each location. Trabecular bone archi-
tecture has also been modeled using tetrakaidecahedral cells, with struts or plates to create open
or closed cells, respectively. This method has been used to model failure of trabecular bone, showing
that localized plastic collapse is a probable failure mode, while elastic buckling is unlikely.

59

A
similar model, restricted to two dimensions, examined the effects of damage accumulation. Guo
et al.

60

modeled the trabecular architecture as repeating hexagonal cells and removed elements when
they were determined to be damaged. By applying cyclic compressive loading, they were able to
demonstrate that fracture of a few trabecular struts can result in a signicant decrease in overall
stiffness of the structure. They conclude that the model accurately predicts fatigue behavior of
trabecular bone under low-stress, high-cycle loading conditions.
Although a wide variety of computational studies have been described, this section has not
come close to describing the breadth of possible uses of these techniques. The main limitation of
these methods is that one cannot model the true complexity of bone. Instead, one must apply
simplifying assumptions for the structure and material properties. Nonetheless, computational
analyses are critically important in advancing understanding of bone properties and function.
Obviously, only computational analyses can be conclusive. However, computational approaches
provide the advantages of repeatability and inexpensive parametric testing which are required to
address a wide variety of research questions.

X. CONCLUDING REMARKS

The purpose of this chapter was to outline methods of mechanically testing bone within the context
of a hierarchically based paradigm. Specic methods or approaches chosen by investigators should be
applied at a scale consistent with the hypotheses being tested. The interpretation of (or inferences from)
the results, however, may transcend more than one hierarchical level. The remaining chapters in this
text provide detailed methodological descriptions of biomechanical test methods. Choosing a protocol
for any study should be done only after careful consideration of the scale of the research question.

ACKNOWLEDGMENTS

The authors would like to acknowledge the help of Peggy Piech in the preparation of this chapter,
and Nancy Caldwell for the trabecular bone architecture image. Bone structurefunction studies
were primarily funded by the National Institutes of Health (AR 34399).

0266/ ch08/frame Page 146 Thursday, October 21, 1999 2:33 PM

A Hierarchical Approach to Exploring Bone Mechanical Properties

147

REFERENCES
1. Choi, K., Kuhn, J.L., Ciarelli, M.J., and Goldstein, S.A., The elastic moduli of human subchondral,
trabecular, and cortical bone tissue and the size-dependency of cortical bone modulus, J. Biomech.,
23, 1103, 1990.
2. Prockop, D.J., Kuivaniemi, H., and Tromp, G., Molecular basis of osteogenesis imperfecta and related
disorders of bone, Clin. Plast. Surg., 21, 407, 1994.
3. Cody, D.D., Goldstein, S.A., Flynn, M.J., and Brown, E.B., Correlations between vertebral regional
bone mineral density (rBMD) and whole bone fracture load, Spine, 16, 146, 1991.
4. McCubbrey, D.A., Cody, D.D., Peterson, E.L., et al., Static and fatigue failure properties of thoracic
and lumbar vertebral bodies and their relation to regional density, J. Biomech., 28, 891, 1995.
5. Tseng, K.F., Bonadio, J.F., Stewart, T.A., et al., Local expression of human growth hormone in bone
results in impaired mechanical integrity in the skeletal tissue of transgenic mice, J. Orthop. Res., 14,
598, 1996.
6. Lewis, D.B., Liggitt, H.D., Effmann, E.L., et al., Osteoporosis induced in mice by overproduction of
interleukin 4, Proc. Natl. Acad. Sci. U.S.A., 90, 11618, 1993.
7. Ducy, P., Desbois, C., Boyce, B., et al., Increased bone formation in osteocalcin-decient mice, Nature,
382, 448, 1996.
8. Rouleau, J.P., Blasier, R.B., Tsai, E., and Goldstein, S. A., Cannulated hip screws: a study of xation
integrity, cut-out resistance, and high-cycle bending fatigue performance, J. Orthop. Trauma, 8, 293,
1994.
9. Moore, D.C., Frankenburg, E.P., Goulet, J.A., and Goldstein, S.A., Hip screw augmentation with an
in situ-setting calcium phosphate cement: an in vitro biomechanical analysis, J. Orthop. Trauma, 11,
577, 1997.
10. Goodman, S.B., Bauer, T.W., Carter, D., et al., Norian SRS cement augmentation in hip fracture
treatment. Laboratory and initial clinical results, Clin. Orthop., 348, 42, 1998.
11. Harrigan, T.P., Jasty, M., Mann, R.W., and Harris, W.H., Limitations of the continuum assumption in
cancellous bone, J. Biomech., 21, 269, 1988.
12. Burstein, A.H., Reilly, D.T., and Martens, M., Aging of bone tissue: mechanical properties, J. Bone
Joint Surg., 58A, 82, 1976.
13. Carter, D.R., Hayes, W.C., and Schurman, D.J., Fatigue life of compact bone II. Effects of
microstructure and density, J. Biomech., 9, 211, 1976.
14. Schafer, M.B., Radin, E.L., and Burr, D.B., Mechanical and morphological effects of strain rate on
fatigue of compact bone, Bone, 10, 207, 1989.
15. Goldstein, S.A., Wilson, D.L., Sonstegard, D.A., and Matthews, L.S., The mechanical properties of
human tibial trabecular bone as a function of metaphyseal location, J. Biomech., 16, 965, 1983.
16. Ciarelli, M.J., Goldstein, S.A., Kuhn, J.L., et al., Evaluation of orthogonal mechanical properties and
density of human trabecular bone from the major metaphyseal regions with materials testing and
computed tomography, J. Orthop. Res., 9, 674, 1991.
17. Kuhn, J.L., Goldstein, S.A., Ciarelli, M.J., and Matthews, L.S., The limitations of canine trabecular
bone as a model for human: a biomechanical study, J. Biomech., 22, 95, 1989.
18. Cody, D.D., McCubbrey, D.A., Divine, G.W., et al., Predictive value of proximal femoral bone
densitometry in determining local orthogonal material properties, J. Biomech., 29, 753, 1996.
19. Goulet, R.W., Goldstein, S.A., Ciarelli, M.J., et al., The relationship between the structural and
orthogonal compressive properties of trabecular bone, J. Biomech., 27, 375, 1994.
20. Wong, M., Lawton, T., Goetinck, P.F., et al., Aggrecan core protein is expressed in membranous bone
of the chick embryo. Molecular and biomechanical studies of normal and nanomelia embryos, J. Biol.
Chem., 267, 5592, 1992.
21. Choi, K. and Goldstein, S.A., A comparison of the fatigue behavior of human trabecular and cortical
bone tissue, J. Biomech., 25, 1371, 1992.
22. Jepsen, K.J., Schafer, M.B., Kuhn, J.L., et al., Type I collagen mutation alters the strength and fatigue
behavior of Mov13 cortical tissue, J. Biomech., 30, 1141, 1997.
23. Tseng, K.F. and Goldstein, S.A., Systemic over-secretion of growth hormone in transgenic mice results
in a specic pattern of skeletal modeling and adaptation, J. Bone Miner. Res., 13, 706, 1998.
0266/ ch08/frame Page 147 Thursday, October 21, 1999 2:33 PM
148 Mechanical Testing of Bone and the BoneImplant Interface
24. Riemer, B.A., Eadie, J.S., Weissman, D.E., et al., Characterization of the architecture, tissue properties,
and continuum behavior of aging trabecular bone, Trans. Orthop. Res. Soc., 19, 189, 1994.
25. Riemer, B.A., Eadie, J.S., Wenzel, T.E., et al., Microstructure and material property variations in
compact and trabecular vertebral bone tissue, Trans. Orthop. Res. Soc., 20, 529, 1995.
26. Townsend, P.R., Rose, R.M., and Radin, E.L., Buckling studies of single human trabeculae,
J. Biomech., 8, 199, 1975.
27. Rho, J.Y., Ashman, R.B., and Turner, C.H., Youngs modulus of trabecular and cortical bone material:
ultrasonic and microtensile measurements, J. Biomech., 26, 111, 1993.
28. Mente, P.L. and Lewis, J.L., Experimental method for the measurement of the elastic modulus of
trabecular bone tissue, J. Orthop. Res., 7, 456, 1989.
29. Ascenzi, A. and Bonucci, E., The tensile properties of single osteons, Anat. Rec., 158, 375, 1967.
30. Ascenzi, A., Baschieri, P., and Benvenuti, A., The bending properties of single osteons, J. Biomech.,
23, 763, 1990.
31. Hofer, C.E., Guo, X.E., Zysset, P.K., et al., Evaluation of bone microstructural properties: effect of
testing conditions, depth, repetition, time delay and displacement rate, in Proceedings of the 1997
Bioengineering Conference, Chandran, K.B., et al., Eds., ASME, New York, 1997, 567.
32. Zysset, P.K., Guo, X.E., Hofer, C.E., and Goldstein, S.A., Elastic modulus of human cortical and
trabecular tissue lamellae, Trans. Orthop. Res. Soc., 22, 798, 1997.
33. Moore, K.E., Hofer, C.E., Zysset, P.K., and Goldstein, S.A., Effect of anatomical location on elastic
modulus of human bone tissue lamellae, Trans. Orthop. Res. Soc., 23, 560, 1998.
34. Hofer, C.E., Moore, K.E., Kozloff, K.M., et al., Femoral neck bone lamellae elastic properties are
independent of age and gender, Trans. Orthop. Res. Soc., 24, 753, 1999.
35. Lips, P., Epidemiology and predictors of fractures associated with osteoporosis, Am. J. Med., 103,
3S, 1997.
36. Ko, C.C., Douglas, W.H., and Cheng, Y.S., Intrinsic mechanical competence of cortical and trabecular
bone measured by nanoindentation and microindentation probes, in Proceedings of the 1995 Bioengi-
neering Conference, Hochmuth, R.M., et al., Eds., ASME, New York, 1995, 415.
37. Roy, M., Rho, J.Y., Tsui, T.Y., and Pharr, G.M., Variations of Youngs modulus and hardness in human
lumbar vertebrae measured by nanoindentation, in Advances in Bioengineering, Rastegar, S., Ed.,
ASME, New York, 385, 1996.
38. Rho, J.Y., Tsui, T.Y., and Pharr, G.M., Elastic properties of human cortical and trabecular lamellar
bone measured by nanoindentation, Biomaterials, 18, 1325, 1997.
39. Su, X.W., Feng, Q.L., Cui, F.Z., and Zhu, X.D., Microstructure and micromechanical properties of
the mid-diaphyses of human fetal femurs, Connect. Tissue Res., 36, 271, 1997.
40. Amprino, R., Investigations on some physical properties of bone tissue, Acta Anat., 34, 161, 1958.
41. Weaver, J.K., The microscopic hardness of bone, J. Bone Joint Surg., 48A, 273, 1966.
42. Katz, J.L. and Meunier, A., Scanning acoustic microscope studies of the elastic properties of osteons
and osteon lamellae, J. Biomech. Eng., 115, 543, 1993.
43. Hofer, C.E., Zhang, N., Kozloff, K.M., et al., Comparison of scanning acoustic microscopy and
nanoindentation measures of the elastic properties of human bone lamellae, in Proceedings of the
1999 Bioengineering Conference, Goel, V.K., et al., Eds., ASME, New York, 1999, 315.
44. Ascenzi, A., Benvenuti, A., and Bonucci, E., The tensile properties of single osteonic lamellae:
technical problems and preliminary results, J. Biomech., 15, 29, 1982.
45. Ascenzi, A. and Bonucci, E., The compressive properties of single osteons, Anat. Rec., 161, 377, 1968.
46. Oxlund, H., Barckman, M., Ortoft, G., and Andreassen, T.T., Reduced concentrations of collagen
cross-links are associated with reduced strength of bone, Bone, 17, 365S, 1995.
47. Burstein, A.H., Zika, J.M., Heiple, K.G., and Klein, L., Contribution of collagen and mineral to the
elastic-plastic properties of bone, J. Bone Joint Surg., 57A, 956, 1975.
48. Hasegawa, K., Turner, C.H., and Burr, D.B., Contribution of collagen and mineral to the elastic
anisotropy of bone, Calcif. Tissue Int., 55, 381, 1994.
49. Turner, C.H., Chandran, A., and Pidaparti, R.M., The anisotropy of osteonal bone and its ultrastructural
implications, Bone, 17, 85, 1995.
50. Luo, Z.P., Bolander, M.E., and An, K.N., A method for determination of stiffness of collagen mole-
cules, Biochem. Biophys. Res. Commun., 232, 251, 1997.
0266/ ch08/frame Page 148 Thursday, October 21, 1999 2:33 PM
A Hierarchical Approach to Exploring Bone Mechanical Properties 149
51. Bonadio, J., Saunders, T.L., Tsai, E., et al., Transgenic mouse model of the mild dominant form of
osteogenesis imperfecta, Proc. Natl. Acad. Sci. U.S.A., 87, 7145, 1990.
52. Bonadio, J., Jepsen, K.J., Mansoura, M.K., et al., A murine skeletal adaptation that signicantly
increases cortical bone mechanical properties. Implications for human skeletal fragility, J. Clin. Invest.,
92, 1697, 1993.
53. Jepsen, K.J., Goldstein, S.A., Kuhn, J.L., et al., Type-I collagen mutation compromises the post-yield
behavior of Mov13 long bone, J. Orthop. Res., 14, 493, 1996.
54. Guldberg, R.E., Caldwell, N.J., Goulet, R.W., et al., Mechanical stimulation of bone repair: an
evaluation of in vivo tissue strains in hydraulic bone chamber model, Trans. Orthop. Res. Soc., 22,
114, 1997.
55. Ladd, A.J., Kinney, J.H., Haupt, D.L., and Goldstein, S.A., Finite-element modeling of trabecular
bone: comparison with mechanical testing and determination of tissue modulus, J. Orthop. Res., 16,
622, 1998.
56. Zysset, P.K., Goulet, R.W., and Hollister, S.J., Prediction of the elastic behavior of human trabecular
bone from morphology and tissue properties, Trans. Orthop. Res. Soc., 22, 64, 1997.
57. Gibson, L.J., The mechanical behavior of cancellous bone, J. Biomech., 18, 317, 1985.
58. Hollister, S.J., Fyhrie, D.P., Jepsen, K.J., and Goldstein, S.A., Application of homogenization theory
to the study of trabecular bone mechanics, J. Biomech., 24, 825, 1991.
59. Guo, X.E., Zysset, P.K., and Goldstein, S.A., Study of post-yield behavior of trabecular bone using
a 3-D microstructural model, in Advances in Bioengineering, Hull, M.L., Ed., ASME, New York,
1995, 165.
60. Guo, X.E., McMahon, T.A., Keaveny, T.M., et al., Finite element modeling of damage accumulation
in trabecular bone under cyclic loading, J. Biomech., 27, 145, 1994.
0266/ ch08/frame Page 149 Thursday, October 21, 1999 2:33 PM
0266/ ch08/frame Page 150 Thursday, October 21, 1999 2:33 PM

151

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

9

Nondestructive Mechanical
Testing of Cancellous Bone

Frank Linde and Ivan Hvid

CONTENTS

I. Introduction ..........................................................................................................................151
II. Guidelines for Nondestructive Testing in Compression......................................................152
A. Type of Test Machine and Strain Measurement Devices..............................................152
B. Handling the Test Specimens.........................................................................................153
C. Preload............................................................................................................................153
D. Upper Strain Limit .........................................................................................................153
E. Strain Rate and Testing Frequency................................................................................154
F. Mechanical Conditioning ...............................................................................................154
G. Determination of Normalized Stiffness .........................................................................154
H. Determination of Energy Absorption.............................................................................154
III. Variation in Test Technique .................................................................................................155
References ......................................................................................................................................156

I. INTRODUCTION

Cancellous bone is a tricky material to test mechanically. First of all it is not homogeneous. There
can be an enormous variation of the strength in an anatomical site.

1

Second, cancellous bone is
anisotropic, exhibiting differences up to tenfold in the stiffness of different directions within the
same anatomical location.

2-4

Third, it is a structure, which becomes apparent when it is cut into
test specimens since the surface of this structure behaves differently from the rest of the test
specimen.

5-7

Furthermore, cancellous bone is viscoelastic.
A standard geometry of a test specimen in order to obtain even stress distribution is a specimen
with a reduced diameter of the central part and a length to diameter ratio of 2:1. Furthermore, the
axial deformation has to be measured at the part of the specimen with reduced diameter. Most
cancellous bone is too fragile and too inhomogeneous to machine into such a specimen geometry.
Therefore, the most popular type of testing has become a compression test using test specimens
with cylindrical or cubic geometry, and with a strain measurement device attached to the test
columns close to the test specimen (Figure 9.1).
This deviation from the standard engineering setup introduces systematic errors which tend to
underestimate the stiffness of the test specimen (as the testing axis is noncoincident with a major
structural axis

8

and the structural surface phenomena),

6-7

while other systematic errors tend to
overestimate the stiffness (uneven stress distribution due to friction at the surface)

9

and viscoelas-
ticity. The net results of these errors have been estimated to be an underestimation of the stiffness
between 20 and 40%.

10

0266/ ch09/frame Page 151 Thursday, October 21, 1999 2:37 PM

152

Mechanical Testing of Bone and the BoneImplant Interface

Why should one perform such an inaccurate test? It is obvious, when accurate data are
imperative such as for providing bone data to put into nite-element models analyzing the rela-
tionship between implants and bone, that the data should be produced more accurately.

11

The trouble
is that this is extremely difcult and maybe even impossible using weak bone. The advantages of
a nondestructive test such as that described below are related to assessment of precision, the ability
to determine mechanical properties in different directions, and the ability to determine viscoelastic
properties.
This technique will be superior in studies where a test specimen can be its own control, such
as in studies of the effect of different preservation methods,

12

temperature, water saturation, and
extraction of different connective tissue elements (collagen, elastin) on the mechanical properties.
By using cubic geometry the test specimen can also be its own control in analyzing anisotropy by
nondestructive testing in different directions.

3

Since the precision can be determined, the technique
also has a great advantage in studies where it is important to calculate the number of patients/samples
which have to be included in a specic study in order to be able to detect a predetermined possible
difference.

II. GUIDELINES FOR NONDESTRUCTIVE TESTING IN COMPRESSION
A. T

YPE



OF

T

EST

M

ACHINE



AND

S

TRAIN

M

EASUREMENT

D

EVICES

The change in length during nondestructive testing of cancellous bone is so small that the quality
of the test machine and deformation measurement device becomes critical for the outcome of the
test.
It has been observed that machines driven by two screws occasionally produce an extra loop
on the top of the hysteresis loop during nondestructive testing. The reason for this extra loop is
probably a small tilting of the crosshead of the testing machine when it turns from loading to
unloading. This can occur because the two screws are driven by one motor. In order to avoid this
phenomenon a hydraulic machine is superior.

FIGURE 9.1

(A) The standard test setup for a mechanical test with an extensometer attached directly to the
reduced section of the test sample. (B) The test setup used for compression testing of a cancellous bone
specimen with the extensometer attached to the testing columns.

0266/ ch09/frame Page 152 Thursday, October 21, 1999 2:37 PM

Nondestructive Mechanical Testing of Cancellous Bone

153

Sometimes a machine runs a little farther than the actual set limits. This overshoot is related
to the speed of transmission and transformation of data. It is therefore obligatory for the research
worker to be familiar with the specic test machine and data-collecting system, so that adjustments
can be made.
Because of the small deformations involved in cancellous bone testing, the deformation of the
machine and testing column are not insignicant, and the built-in deformation measurement device
is not sufcient. It is imperative to use an extensometer or a similar accurate device placed on the
test columns close to the test specimen or directly on the specimen in order to measure changes
in the length accurately.
The following description of nondestructive testing implies the use of a strain measurement
device attached to the test columns close to the test sample.

B. H

ANDLING



THE

T

EST

S

PECIMENS

A test specimen that has been stored frozen should be kept in saline at room temperature for a few
hours before testing. The authors put a droplet of mineral oil on the ends of the testing column in
order to reduce friction. Drying of a specimen is not a problem during the short time of nonde-
structive testing including the mechanical conditioning procedure. If the specimen is going to be
retested later the same day, it should be kept in saline. It can be stored frozen for later retesting
without damage.

12

C. P

RELOAD

When the test is started it is necessary to have contact between the specimen and test columns in
order to avoid artifacts from uid on the surfaces and in order to dene zero strain. This is best
done by dening the zero strain to be at a specic loading stress. The authors usually use a stress
of 0.1 MPa. This stress corresponds to a load of a few newtons (preload) when the diameter/side
length of the testing specimen is between 5 and 7 mm (cylindrical/cubic geometry).

D. U

PPER

S

TRAIN

L

IMIT

The options for setting the upper test limit in nondestructive testing are a xed load, a percentage
of the ultimate load determined from the density of the test specimen (obtained by quantitative
computed tomography or photon absorptiometry) or a xed strain. Whereas ultimate stress (the
strength) varies considerably, ultimate strain is nearly constant. Accordingly, a strain limit has been
found to be best in producing a stiffness which correlates strongly with the stiffness of a test to
failure.

13

The test is conducted between a lower stress limit (preload) and an upper strain limit. This
upper strain limit should be chosen so that the test runs into the linear part of the loaddisplacement
curve, but not so far that the slope of the curve is decreasing. By using the cutting/sawing technique
(EXAKT Apparaturbau, Hamburg, Germany) for specimen preparation, it was found that for test
specimens between 5 and 7.5 mm, the maximum strain limit in order to avoid beginning failure
during testing was 0.8%; 0.6% strain was found to be the optimum safe upper strain limit. However,
if the machining technique is more precise in producing parallel end plates, or less traumatizing
to the trabecular struts at the surface, or if specimens longer than 7.5 mm are chosen, then the
upper strain limit should be smaller.
It is advisable to run a pilot study by testing a number of specimens to failure. The average
strain corresponding to the middle of the linear part of the loaddisplacement curve should be
determined. The middle of the linear part can be determined by tting a third-order (or higher
order) polynomial to the stressstrain data set from zero strain to a point between yield strain and
ultimate strain. The middle of the linear part can be dened and found as the strain corresponding
to the point where the tangent changes from increasing to decreasing. A simpler method, and fairly

0266/ ch09/frame Page 153 Thursday, October 21, 1999 2:37 PM

154

Mechanical Testing of Bone and the BoneImplant Interface

accurate for that purpose, would be to draw the tangent to the point with maximum slope on a
printout and determine the strain corresponding to the middle of that segment where the testing
curve does not differ signicantly from a straight line. The upper strain limit should be selected
as 75% of that strain.
If the pilot study reveals an average strain corresponding to the middle of the linear part of the
loaddisplacement curve larger than 1%, one should seriously consider if the technique of machin-
ing test specimens is sufciently good or if there is a major error in displacement measurement.

E. S

TRAIN

R

ATE



AND

T

ESTING

F

REQUENCY

Since cancellous bone has viscoelastic properties, the results of testing are time dependent. As a
consequence, the strain rate and testing frequency have to be xed. A sequence of test cycles can
be made by loading and unloading using the same strain rate without a pause between cycles, or
a predetermined pause between cycles can be used. The authors nd a strain rate of 0.01 s

1

and
a testing frequency of 0.2 Hz to be suitable. This method provides a pause of about 4 s between
cycles, which is sufcient time for resetting the strain channel during conditioning (see below).

F. M

ECHANICAL

C

ONDITIONING

The reproducibility of the rst test cycle is not particularly good. The end point of the testing curve
will not coincide with the starting point. This is partly due to the viscoelastic properties of cancellous
bone and probably partly due to smoothing of small surface irregularities during the rst cycle.
The end point of each cycle moves a little to the right (a small residual strain) during repetitive
testing, but the difference between the starting point and the end point of a cycle becomes smaller
and smaller by continued testing. A steady state (the end point of the curve coincides with the
starting point) is usually reached by 5 to 15 cycles. The real test can be performed immediately
after conditioning either as a single nondestructive test or as the average of a number of nonde-
structive tests.
Since the starting point of a new cycle moves to the right until a steady state has been reached,
the strain at start is not zero and the strain interval will be smaller than selected (Figure 9.2). The
magnitude of this shift to the right of the stating point may vary considerably between specimens.
The authors found that the reproducibility of the nal test was best, when the strain channel was
reset to zero at the end of each test cycle, so that the nal test was conducted within the zero strain
limit and the selected upper strain limit for that particular cycle. The authors laboratory uses a
xed number of conditioning cycles (15 or 20) or denes the steady state to be reached when the
starting point was reproduced

m within three consecutive conditioning cycles.

G. D

ETERMINATION



OF

N

ORMALIZED

S

TIFFNESS

The stiffness is dened as the steepest part of the loading curve. Since data are collected and stored
by a computer, the maximum slope of the loading curve is easily calculated by tting a third-order
polynomial to the data set of the loading curve (Figure 9.3).

H. D

ETERMINATION



OF

E

NERGY

A

BSORPTION

The area underneath the loading curve represents energy stored or absorbed by the specimen and
the interface between specimen and test machine (loading energy). The area underneath the unload-
ing curve represents energy regained during unloading (elastic energy). The area between the
loading and the unloading curve represents energy lost during a cycle (hysteresis). This energy loss
is the sum of viscoelastic energy absorbed by the test specimen and energy lost by friction at the
interface between test specimen and test the columns (see Figure 9.3).

0266/ ch09/frame Page 154 Thursday, October 21, 1999 2:37 PM

Nondestructive Mechanical Testing of Cancellous Bone

155

III. VARIATION IN TEST TECHNIQUE

If cancellous bone from large mammals and from anatomical sites with strong bone is the subject
of study, it may be possible to machine specimens with a 2 : 1 length-to-diameter ratio strong
enough to attach a miniextensometer directly to the specimen without causing damage to it.

11

In
that case a great part of the systematic error arising from the interface between specimen and test
column will be eliminated. The upper strain limit should in that case be selected considerably
smaller. An upper strain limit as small as 0.2% should be selected

14

in order to avoid signs of
beginning failure (decreasing stiffness on repeated testing).
The technique described including the above-mentioned variation works with physiological
strains. Nondestructive testing by small strain amplitudes (ultrasound) has been used for determi-
nation of both elastic and viscoelastic properties. Strong correlations have been found between
properties obtained by ultrasound techniques and properties obtained by large strain mechanical
testing technique. Since the stiffness obtained by the ultrasound technique usually is high, it may
look less affected by those systematic errors listed in the beginning of this chapter. The ultrasound
technique is also useful for determination of properties in different directions, and it is even possible
to obtain the shear stiffness.

15,16

FIGURE 9.2

The first four cycles of the mechanical conditioning procedure are show. (A) Conditioning
performed to a xed load. Note that the cycles move to the right. The strain at the end of the loading curve
is comparable to the changes in strain in a creep test. (B) The four cycles from (A) are put on top of each
other with the same starting point. Note that the stiffness is increasing from test to test and the area of the
hysteresis loop is decreasing. (C) Conditioning to a xed strain by resetting the strain channel between each
cycle. Note that the stiffness and the stress at the end of the loading curve increase from test to test.

0266/ ch09/frame Page 155 Thursday, October 21, 1999 2:37 PM

156

Mechanical Testing of Bone and the BoneImplant Interface

REFERENCES

1. Hvid, I., Mechanical strength of trabecular bone at the knee,

Dan. Med. Bull.

, 35, 345, 1988.
2. Evans, F.G. and King, A.I., Regional differences in some physical properties of human spongy bone,
in

Biomechanical Studies of the Musculo-Skeletal System

, Evans, F.G., Ed., Charles C Thomas,
Springeld, 1961, 19.
3. Linde, F., Pongsoipetch, B., Frich, L.H., and Hvid, I., Three-axial strain controlled testing applied to
bone specimens from the proximal tibial epiphysis,

J. Biomech

., 23, 1167, 1990.
4. Martens, M., Van Audekercke, R., Delport, P., et al., The mechanical characteristics of cancellous
bone at the upper femoral region,

J. Biomech

., 16, 971, 1983.
5. Allard, R.N. and Ashman, R.B., A comparison between cancellous bone compressive moduli deter-
mined from surface strain and total specimen deection,

Trans. Orthop. Res. Soc.

, 16, 151, 1991.
6. Linde, F. and Hvid, I., The effect of constraint on the mechanical behavior of trabecular bone
specimens,

J. Biomech

., 22, 485, 1989.
7. Odgaard, A. and Linde, F., The underestimation of Youngs modulus in compression testing of
cancellous bone specimens,

J. Biomech

. 24, 691, 1991.
8. Turner, C.H. and Cowin, S.C., Errors induced by off-axis measurement of the elastic properties of
bone,

J. Biomech. Eng

., 110, 213, 1988.
9. Dieter, G.E.,

Mechanical Metallurgy

. McGraw-Hill, New York, 1961, 479.
10. Linde, F., Elastic and viscoelastic properties of trabecular bone by a compression testing approach,

Dan. Med. Bull

., 41, 119, 1994

FIGURE 9.3

A typical loaddisplacement curve. The stiffness is the slope of the steepest part of the loading
curve. The stiffness is normalized using the cross-sectional area and the original length of the specimen.
Elastic energy is the area underneath the unloading curve. The energy loss (viscoelastic energy) is the area
enclosed by the loading/unloading cycle.

0266/ ch09/frame Page 156 Thursday, October 21, 1999 2:37 PM

Nondestructive Mechanical Testing of Cancellous Bone

157

11. Keaveny, T.M., Guo, X.E., Wachtel, E.F., et al., Trabecular bone exhibits fully linear elastic behavior
and yields at low strains,

J. Biomech.

, 27, 1127, 1994.
12. Linde, F. and Srensen, H.C., The effect of different storage methods on the mechanical properties
of trabecular bone,

J. Biomech

., 26, 1249, 1993.
13. Linde, F., Gthgen, C.B., Hvid, I., et al., Mechanical properties of trabecular bone by a nondestructive
compression testing approach,

Eng. Med

., 17, 23, 1988.
14. Rhl, L., Larsen, E., Linde, F., et al., Tensile and compressive properties of cancellous bone,

J. Biomech.

, 24, 1143, 1991.
15. Ashman, R.B., Experimental techniques, in

Bone Mechanics

, Cowin, S.C., Ed., CRC Press, Boca
Raton, FL, 1989, 75.
16. Ashman, R.B., Rho, J.Y., and Turner, C.H., Anatomical variation of orthotropic elastic moduli of the
proximal human tibia,

J. Biomech

., 22, 895, 1989.

0266/ ch09/frame Page 157 Thursday, October 21, 1999 2:37 PM

0266/ ch09/frame Page 158 Thursday, October 21, 1999 2:37 PM

159

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

10

Synthetic Materials and
Structures Used as Models
for Bone

John A. Szivek

CONTENTS

I. Cortical Bone Substitutes.....................................................................................................159
A. Introduction ....................................................................................................................159
B. Preparation of Synthetic Bones for Testing and Test Selection....................................161
C. Evaluation of Mechanical Properties Prior to Use........................................................161
D. Testing Implants with Synthetic Bone Models..............................................................163
II. Trabecular Bone Modeling ..................................................................................................164
A. Introduction ....................................................................................................................164
B. Foam Preparation ...........................................................................................................166
C. Foam Testing ..................................................................................................................169
III. Summary ..............................................................................................................................169
References ......................................................................................................................................169

I. CORTICAL BONE SUBSTITUTES
A. I

NTRODUCTION

Cadaver bones have routinely been used as support structures during bench-top testing to evaluate
biomechanical changes caused by articial joints,

1-3

orthopaedic fracture xation devices,

4-6

and
new orthopaedic procedures. Obtaining fresh disease-free cadaver bones to be used during mechan-
ical testing of orthopaedic implants and xation systems is becoming difcult and extremely
expensive. Large interbone shape and materials property variations with standard deviations in
excess of 100% have been noted.

1,7,8

Variations of this size imply that several hundred bones are
required to approach statistical signicance. Sample sizes reported in the literature are typically
less than 50 specimens and are often as few as three specimens. Interbone variations have prompted
investigations to carry out comparative testing within the same bone to assess biomechanical
parameters prior to and following implant placement. This approach, in which the bone is used as
its own control, limits experimental design considerably. Moreover, the handling and storing of
human cadaver specimens prior to and during testing can also cause problems leading to changes
in properties during the course of the test.
These factors have prompted the development of synthetic support structures for use during
mechanical testing. In general, accurately simulating the stiffness of bone is most important when
stress shielding near articial joints or fracture xation devices are being evaluated, while

0266/ ch10/frame Page 159 Thursday, October 21, 1999 2:42 PM

160

Mechanical Testing of Bone and the BoneImplant Interface

accurately simulating the strength of bone is most important when screws or other orthopaedic
attachment devices are being evaluated. Early studies that simulated the properties of bone as a
support structure used tubular metal,

9

wood, and constructs such as laminated linen.

10-12

Although
these constructs demonstrated consistent materials properties, they did not accurately simulate the
strength, stiffness, and time-dependent properties of bone. In addition, the shape of these constructs
did not allow accurate modeling of physiological loads. One of the earliest reports of the preparation
and use of a ber-reinforced proximal femur

13

described a structure which simulated bone strength
and stiffness properties better and attempted to model the shape of the proximal femur. Fiber-
reinforced femora of this kind were further developed and used to compare fracture xation
techniques

14

and evaluate stability of articial hip designs.

15

Uta

16

suggested the use of polyurethane
model bones to standardize testing of fracture xation devices.
Fiber-reinforced epoxy bones became commercially available in the late 1980s (Pacic
Research Labs, Inc., Vashon Island, WA). The rst generation (F-type) amber-colored left femora
were carbon berreinforced epoxy. The geometric measurements of these bones fell within the
spectrum of sizes corresponding to cadaver femora.

17,18

There were mixed reports on the useful-
ness of this type of bone. Beals

19

reported favorable results but Szivek et al.

20

concluded, from
torsional and anatomical bending tests, that smaller variations in materials properties were needed
for these to serve as valuable bone models and support structures for mechanical testing. Neither
investigator had carried out a quantitative comparison with cadaver specimens of the same size.
However, both authors agreed that the consistency in size of these bones facilitated the reuse
of xtures to run multiple comparisons using the synthetic bones. This offered advantages over
the use of cadaver materials.
More recently, a second-generation glass berreinforced bone was introduced by the same
manufacturer. Examination of the mechanical properties

21

indicated that the cortical component
of these synthetic bones demonstrated consistent mechanical properties and a similar deformation
response to that of a size-matched cadaver bone. This study suggested that the polyurethane foam
used to model the trabecular bone was weak,

21

although a recent report

22

indicated that the stiffness
of this foam specied at 69 MPa falls within the range of properties reported for trabecular bone.

23

Other investigators have noted that these bones have a similar deformation response to cadaver
bone in four-point bending,

22

axial loading,

22

and during simulated physiological loading,

21

but not
during torsion testing.

20,22

Otani et al.

25

have observed that these models are not reliable for circumferential strain meas-
urement under axial load or for any type of strain measurement under torsion load. They indicated
that differences in comparison to bone were likely a result of the cross-ply ber orientation in the
synthetic bones. They did not observe these discrepancies when micromotion measurements were
collected for implant stability studies. Examination of femoral head deection and strain distribu-
tions across proximal bone sections during axial testing

22

have also shown that while overall
mechanical properties are quite consistent from one synthetic bone to another, local strains can
vary. This suggests that testing a bone intact and then with an implant will provide the best
understanding of the effect of the implant. It has been noted that while time constants on the
viscoelastic response of these bones are short, it is benecial to leave loads in place during static
testing for up to 4 min.

22

In addition to two sizes of composite femora, a synthetic tibia is now also available. There are
no published comparisons of the deformation response of the synthetic tibia with that of the cadaver
bone. These tibia models do offer the advantage of providing consistency in size, allowing the
reuse of xtures and making it easy to run multiple comparisons. The lack of mechanical properties
information on these bones suggest that studies in which these are used should be carefully planned
to allow adequate characterization of intact properties before using them to evaluate the performance
of implants.

0266/ ch10/frame Page 160 Thursday, October 21, 1999 2:42 PM

Synthetic Materials and Structures Used as Models for Bone

161

B. P

REPARATION



OF

S

YNTHETIC

B

ONES



FOR

T

ESTING



AND

T

EST

S

ELECTION

Variations in the fabrication process occasionally lead to some anomalies in the location of the
berglass-reinforcing sleeve within the bones which can locally affect mechanical properties. X-
raying the bones prior to testing provides a means of detecting gross differences, although it cannot
be used as a predictive tool to assess the way in which mechanical properties are affected. The
author has used a high-resolution Faxitron cabinet X-ray system (Hewlett-Packard, McMinniville,
OR) with X-OMAT AR scientic imaging lm (Eastman Kodak Co., Rochester, NY) as a screening
tool. In order to reduce interbone variability during testing several specimens can be purchased
and bones which pass radiographic scrutiny should be compared intact. Selection of a subgroup
from this larger group provides test bones with very similar deformation responses. In preparation
for testing, which requires accurate cortical and trabecular support, removal of the foam is recom-
mended. It can be replaced with a polyurethane foam called Daro foam (see Section II on trabecular
bone models for details) which can be produced with a range of stiffness from 63 to 104 MPa,

24

simulating the properties of the trabecular bone from a range of patient populations.
Test and measurement system selection should be based on the fact that synthetic femora have
a similar deformation response to cadaver femora in four-point bending, axial loading, and during
simulated physiological loading but not torsional loading. Deection of components of the femur
such as the head have also been noted to be consistent. Test design should include testing the femur
intact and then with an implant in place to provide the best understanding of the effect of the
implant. The viscoelastic response of these bones dictate that tests be separated by several minutes
to allow viscoelastic effects to dissipate. Although these synthetic bones are more consistent during
retesting than cadaver bones, retesting of the same construct at least three times is mandatory to
assess reproducibility.

21,22

C. E

VALUATION



OF

M

ECHANICAL

P

ROPERTIES

P

RIOR



TO

U

SE

The simplest testing procedure that allows comparison of whole-bone properties is four-point
bending (Figure 10.1). It has been used to assess stiffness characteristics of fracture xation systems.
Four-point bending is preferable to three-point bending since it assures a uniform bending moment
between the inner two supports and potentially more accurate deection measurements from this
region of the specimen. The separation of the roller supports is dependent on the size of the specimen
being tested. Ideally, the supports should be as far apart as possible and located within the diaphysis
so that the metaphyseal ares do not affect testing. A displacement transducer (commonly an
extensometer) attached to the centerline of the middiaphysis of the bone and to a reference bar is
used to evaluate deection accurately. A small pin or K wire placed in the bone can facilitate
attachment of the extensometer to the bone.

FIGURE 10.1

Diagram showing a four-point bend testing configuration which can be used to characterize
overall bending stiffness of a synthetic bone with or without an implant.

0266/ ch10/frame Page 161 Thursday, October 21, 1999 2:42 PM

162

Mechanical Testing of Bone and the BoneImplant Interface

Alternatively, strain gauges can be attached to the lateral, medial, anterior, and posterior bone
surfaces. Ideally, gauges used on the surface of these composite bones should have a self-temper-
ature compensation factor eliminating errors due to thermal expansion of the gauge. They should
also have a relatively large surface area and high resistance to minimize the amount of heating.
For greatest accuracy a Micro-Measurements EA-00-250BK-10C uniaxial gauge (Measurements
Group, Raleigh, NC) can be used. Since measurements are often taken over relatively short periods
of time thermal factors generally have a small effect on accuracy, and it is common to use more
readily available CEA-06-125UW-120 (Measurements Group, Raleigh, NC) uniaxial-type gauges.
For biaxial and rosette strain measurements CEA-06-062WT-120 and WA-06-060WR-120, respec-
tively, are satisfactory. These gauges can be attached using a suitable cyanoacrylate adhesive such
as M-Bond 200 (Measurements Group, Raleigh, NC). Additional choices are available in the M-
Line strain gauge accessories catalog (A-110-6, Measurements Group, Raleigh, NC). European
investigators

22,26

have successfully used H.B.M. 6/120 LY 11 supplied by Hottinger Baldwin
Messtechnik, Stuttgart, Germany.
Bones should be tested in ML (medial to lateral) and AP (anterior to posterior) bending. It is
valuable to clamp or in some way x one end of the femur to prevent rotation during the loading
procedure. The triangular shape of the tibia may preclude the need to do this. Three to six repetitions
of a test will provide an indication of the variation in response to a particular test setup. No
conditioning effects should be observed; i.e., no progressive change in properties should be noted
over the course of the repeated testing.
In cases where synthetic bones are to be used to evaluate implants which must support torsion
loading, a simple whole-bone torque test can be used to characterize the response of the model
bone prior to testing with an implant in place. The bone should be gripped securely at the ends
and positioned with the center of the diaphysis aligned along the center of the torsion axis.
For testing of the femur, this alignment will place the intercondylar space and the neck of the
femur on this centerline (Figure 10.2A). The xture holding the proximal femur should be a stiff,
channel-shaped part with a deep section so that the surface of the head, neck, and greater trochanter
are all rotated simultaneously by the test machine actuator. The distal femur should be placed in

FIGURE 10.2

(A) Diagram showing the fixture and test configuration used to test an intact or implanted synthetic
femur. (B) Diagram showing the xture and test conguration used to test an intact or implanted synthetic tibia.

0266/ ch10/frame Page 162 Thursday, October 21, 1999 2:42 PM

Synthetic Materials and Structures Used as Models for Bone

163

a rectangular steel channel the height of the condyles or, alternatively, can be embedded to a level
above the condyles in a low-melting-point compound such as Cerrobend (Scottsdale Tool, Phoenix,
AZ) or a polymethylmethacrylate (PMMA) such as a dental repair resin (Hygienic Repair Resin,
Hygienic Corp., Akron, OH). The xture ends should be observed closely during loading to ensure
solid gripping so that torsional displacements measured by the actuator will be accurate. The
direction of rotation will affect the response and should be chosen carefully to simulate the response
of the bone accurately, e.g., internal rotation of the femoral head can be used to simulate the
physiological loading during rising from a chair. A slight amount of initial torque load (up to 5
Nm) can be applied to ensure the system is tight. Specimens should not be tested beyond 30 Nm
unless nonlinear responses are of interest in the study.
Torsion testing of a tibia is slightly more complex. A transcortical pin or screw embedded in
a potting compound such as PMMA can be used to hold the distal tibia (Figure 10.2B). The proximal
tibia can be xtured in a similar way. This setup will work if testing of plates or external xation
devices is to be undertaken, although it may interfere with the ability to place intramedullary
devices. In general, this type of setup will also allow the use of combined loading such as combined
axial and torsion loading. This is best accomplished with a test machine that can independently
apply torsion and axial loads, but for situations in which only an axial actuator is available a xture
with a loading ramp (Figure 10.3) can provide combined loading.

27

Axial loading has also been used to characterize the axial deformation response of femora but
is not commonly used to evaluate implants. For axial compression or tension testing, the same
xture design used for torsion testing can be employed. However, the deformations will be small
and the most sensitive deformation measurement approach available is recommended. Strain gauges
placed around the circumference of the bone are sufciently sensitive for this task, but the ram
excursion measurement from a servohydraulic test machine is generally not.

D. T

ESTING

I

MPLANTS



WITH

S

YNTHETIC

B

ONE

M

ODELS

Using synthetic femurs for studies in which articial joints or fracture xation devices are tested
is best done by simulating physiological loading. At the very least, body weight and the effects of

FIGURE 10.3

Illustrations of the way in which a loading ramp can be used to generate a torsion load on a
bone if only an axial-loading xture is available. (From Szivek, J.A., and Yapp, R.A.,

J. Biomed. Mater. Res.
Appl. Biomater.

, 23 (A1 Suppl.), 105, 1989. With permission.)

0266/ ch10/frame Page 163 Thursday, October 21, 1999 2:42 PM

164

Mechanical Testing of Bone and the BoneImplant Interface

the abductors should be modeled. Head and abductor loading angle can be altered to simulate
several stance positions.

28

Prior to testing a model femur with implants in place, a control test of
the intact femur alone is imperative. Figure 10.4 shows one design of a head and abductor loading
arrangement used to apply a simulated stance load to a model femur. In this arrangement the
abductor loading xture can be strain-gauged and calibrated with a load cell or weights. The gauge
circuit arrangement uses four gauges to provide uniform uniaxial strain measurements.

29

Alterna-
tively, a small load cell can be purchased and incorporated into the xture to measure load.
Testing using this arrangement can be carried out while monitoring strain gauges attached to
the bone surface or while monitoring linear variable differential transformers placed near locations
of interest on the bone or implant. It can also be used in conjunction with motion analysis systems
to monitor motion or gross deformation of the bone. Markers such as K wires or pins allow
monitoring of the motion of selected surface locations of interest.
Recently, there has been an interest in modeling additional muscle loads on this type of model,
and Figure 10.5 shows one conguration of this test setup in which the addition of lateral muscle
loads has been incorporated into the loading procedure. There is no consensus on the ideal loading
arrangement for this type of testing, and a detailed discussion of the advantages of various loading
arrangements is beyond the scope of this chapter but can be found in reviews of this topic such as
Colgan et al.

30

II. TRABECULAR BONE MODELING
A. I

NTRODUCTION

The reproducibility of biomechanical test results for orthopaedic implants tested in cadaver trabe-
cular bone has often been hampered by the wide variation in the mechanical properties of the

FIGURE 10.4

Diagram illustrating the way in which a head load and trochanter load can be applied to a
femur.

0266/ ch10/frame Page 164 Thursday, October 21, 1999 2:42 PM

Synthetic Materials and Structures Used as Models for Bone

165

substrate material.

31-39

Differences between samples have been attributed to variations in specimen
selection site,

32,34,40

and the age, sex, and metabolic conditions

35,36,39

of the donor. Materials with
consistent and controllable mechanical properties similar to those of human cancellous bone would
provide a valuable alternative to cadaver bone as a test substrate. In addition, the ability to produce
unlimited numbers of these substrates in a variety of shapes offers a major advantage over cadaver
specimens whose shapes are dened by the tissue bed from which they were harvested.
The structure of trabecular bone is a network of connecting rods and plates which form columns
and struts.

41

This results in an interconnected pore structure within the bone. There have been no
reports of synthetic porous structures with interconnecting pores having the same stiffness and
strength characteristics of human trabecular bone. Implantable coralline hydroxyapatite (Interpore
International Irvine, CA) or porous lters made of metals or ceramics (Astromet, Cincinnati, OH)
could be used if interconnecting porosity is of primary importance and strength characteristics are
of secondary consideration. If strength or stiffness need to be more accurately modeled, porous
polymers are better candidates even though they have a closed pore structure.
One popular synthetic closed-cell polyurethane foam (Daro, Butler, WI) rst used as a test bed
for fracture xation devices

42

and evaluation of the stability of articial joint components

43,44

has
been noted to have a structure with some similarities to human trabecular bone

24,45

(Figure 10.6).
The two parts provided to form Daro foam are an isocyanate (methylene diisocyanate) and a resin
(polyol). The foam is formed during a process when liquid polymercoated gas bubbles impinge
and solidify.

24,45

Varying the density (and bubble size) affects the mechanical properties of the foam,
allowing some tailoring of the properties to model bone from a range of types of patients.
The stressstrain curve of this closed-cell foam is similar to that of trabecular bone when
compression-loaded.

24,31,36,37

In both cases, the rst phase of loading these materials leads to a linear
elastic response as the components of the material are subjected to compression or tension loads.
At higher loads, they yield as the cell walls begin to collapse. The resistance to load increases,
causing a nal increasingly steep slope to the stressstrain curve.

31,33

FIGURE 10.5

Diagram illustrating the way in which additional lateral muscle loads can be applied to a femur.

0266/ ch10/frame Page 165 Thursday, October 21, 1999 2:42 PM

166

Mechanical Testing of Bone and the BoneImplant Interface

Initial characterization of the mechanical properties of this foam were reported for a formulation
representing normal trabecular bone.

24

Recent studies

45

have shown that by varying the ratio of
resin to isocyanate the mechanical properties of the foam can be altered so that it can model the
trabecular bone in a range of patient populations. This allows its use for the evaluation of implants
or procedures specic to a particular patient group. Some examples include young healthy trauma
patients, patients with osteoporosis, or those with rheumatoid arthritis.

B. F

OAM

P

REPARATION

Although the properties of Daro foam can be very consistent, factors that cause variations in its
properties are the ratio of isocyanate and resin, temperature, humidity, mixing protocol, pouring
protocol, container shape, container size, and thermal conductivity and whether the container is
open or sealed during setting. Since the foam can form with different properties, care should be
taken to control as many of these parameters as possible and to test small samples separately from
each fabrication procedure to characterize property variations.
The rst step in the preparation of the foam for evaluating orthopaedic devices requires the
selection of a container to provide an appropriate shape or test bed. Although the implant to be
tested will to some extent dictate a preferred shape, cube-shaped containers provide relatively more
uniform properties compared with nonsymmetric or oddly shaped containers. To avoid edge effects
(unless one is intentionally attempting to model bone property variations near surfaces) the foam
blocks should be relatively large compared with the implant being tested. Analytical modeling of

FIGURE 10.6

Scanning electron micrographs showing the structure of two solidified foams with different
bubble densities.

0266/ ch10/frame Page 166 Thursday, October 21, 1999 2:42 PM

Synthetic Materials and Structures Used as Models for Bone

167

screw pullout has shown that the damage area is equivalent to two screw diameters.

46

Blocks with
a width greater than twice the diameter of the screws should be made, and if multiple screws are
pulled out of the same block they should be separated by this distance.
A skin of relatively denser material will form at all surfaces when foam is made. It can be used
to simulate a cortical shell formed around trabecular surfaces of some bones, or this skin can be
removed in order to model the trabecular bed only. In either case, it is imperative that a sample of
material (in the form of a small cube or cylinder if possible) from the interior of the foam be tested
in either uniaxial tension, compression, or shear to conrm its mechanical properties and to ensure
interbatch consistency. If the skin is to be used to evaluate implants, the mechanical properties of
samples of the skin (possibly in the form of a thin sheet) should also be characterized.
Ideally, the loading of the cubes, cylinders, or the skin samples should be similar to the way
in which the implants inserted into the foam blocks will load the foam during implant evaluation.
While the results of recent studies

27,45

were obtained using samples compressed without constraining
specimen ends or sides, studies on trabecular bone have suggested that unconstrained testing
procedures produce failure strength values that are low in comparison with those predicted by
nite-element models which are constrained by surrounding bone.

47,48

However, most literature
values for trabecular bone reported to date have been measured using unconstrained experimental
testing. As such, it has been possible to compare synthetic material properties with these values.
When contemplating the use of these foams, future characterization will benet from testing with
end and side constraints. Specimen ends can be constrained by using PMMA or epoxy to attach
specimens to the platens of the test machine. Accurate side constraint is harder to achieve. A close-
tting sleeve around a loading piston with the exact cross-sectional area of the face of the test
specimen is one option. Pushout testing through a block larger than the sample being compressed
may provide more accurate side constraint modeling. Few test results from trabecular bone evaluated
using these approaches are currently available.
Daro foam properties vary depending on the molding parameters. Figure 10.7 provides an
example of the range of strength and stiffness properties noted when the ratio of resin to isocyanate
is altered from 10.0 : 10.0 to 10 : 7.9 and 10.0 : 5.0.

24,45

Over this range of ratios, the bubble
diameters increase by approximately 13%. The foam density changes by approximately the same
percentage, decreasing from 269 g/cc



to 235 g/cc. Although this is a fairly narrow range of foam
density, published evidence over a wider range for other foams suggests a linear relationship
between density and strength or stiffness.

49

To ensure uniform properties throughout each foam, resin should be mixed prior to use and
added to the isocyanate and then mixed with a blender or paint mixer at 1500 rpm for at least 20 s.
Although properties as a function of mixing time have not been evaluated quantitatively, shorter
mixing times do not allow thorough mixing and result in uneven bubble sizes and properties through
the section of the foam. Some bubble size variation is to be expected and cannot be avoided even
with much longer mixing times (Figure 10.8).
Mixtures must immediately be poured into a container and allowed to cure for 24 h. In cases
in which a at, dense skin is desirable to model a at, cortical shell, the container should be
enclosed. Properties of foams made using this procedure are slightly different and warrant exam-
ination since they could be used to study mandibular ridge or skull modeling applications.
In cases in which specimens are intended to simulate a trabecular bed only, a band saw should
be used to remove the surface of the foam. It is advisable to mark the specimen orientation and to
mark the orientation of any samples cut from it. Some variation in properties within specimens
resulting from the foaming process and direction of bubble migration during solidication have
been observed. Variations across specimens smaller than 5 cm in thickness are minor. When
measuring sample shapes prior to testing, several measurements along each edge are recommended
since cut tolerances for this material even when prepared by machining are not tight. However,
since large specimens are easily prepared, this material characteristic will not adversely affect the
ability to calculate failure stresses and stiffness values accurately.

0266/ ch10/frame Page 167 Thursday, October 21, 1999 2:42 PM

168

Mechanical Testing of Bone and the BoneImplant Interface

FIGURE 10.7

A graph summarizing the strength and stiffness of the foams as a function of isocyanate to
resin mixtures. Values were drawn from published information.

24,45

FIGURE 10.8

An example of a block taken from a foam prepared using a resin-to-isocyanate ratio of 10.0 : 5.0.

0266/ ch10/frame Page 168 Thursday, October 21, 1999 2:42 PM

Synthetic Materials and Structures Used as Models for Bone

169

C. F

OAM

T

ESTING

The most facile specimen characterization technique is compressive loading. A load rate comparable
to that to be used when testing the orthopaedic implant in the model material should be used in
this test. Implants to be used in older, total-joint patients should be tested relatively slowly in
comparison with tests of devices used in young athletes. A rate of 0.5 mm/s on a servohydraulic
test frame provides a moderate test speed to evaluate the foam for tests involving study of implants
during the postoperative period when patients are expected to have activities with moderate load
rates. Each sample must be conditioned by axial loading three times to approximately 70% of the
compressive yield strength, since surface damage caused by cutting will initially produce inaccurate
stiffness readings. After conditioning, each sample should be compressed axially in the same
direction until permanent damage is apparent. Deformation of test specimens can most easily be
measured with the LVDT in the test frame. An extensometer can also be used for greater accuracy
or if deformation in other than the loading direction is of interest. For greatest accuracy optical
extensometry of the middle portion of a sample is ideal.

47

Loaddeformation values should be
plotted to establish the stiffness of the material and its failure strength. Failure stresses should be
calculated using stress = force/area, and failure strain using strain = (change in height from failure
load original height)/original height. Stiffness is calculated using stiffness = stress/strain.
Shear testing of this material is more difcult and has not been described in the literature.
While a rectangular specimen that is attached to the surfaces of a lap shear xture with a high-
strength epoxy could be used, a pushout-type test may be more appropriate for the type of
specimen that can easily be created from these foams. Similarly, tensile testing would require
epoxying specimens to platens which could be used to pull the specimens apart. Calculation of
cross-sectional area and consequently failure strength values from this type of test are expected
to be difcult. Since few implant interfaces are loaded in pure tension, this test may offer
information of limited usefulness.

III. SUMMARY

The cortical and trabecular bone models developed to date offer the advantages of consistency
in shape, strength, and stiffness. These advantages alone make them a better choice than cadaver
bone as support materials for testing orthopaedic and dental implants in almost all applications.
Continued development of ber-reinforced epoxy to create a model with more similar torsional
deformation to cortical bone would be valuable. Testing of a range of types of trabecular bone
from patients with various diseases and additional development of polyurethane foams with an
even greater range of properties would provide foams simulating the properties of trabecular
bone in a wider range of patients. Recently developed standards

50

for polyurethane foam materials
for use during mechanical testing should also be considered as a guide during the development
of new trabecular bone substitutes.

REFERENCES

1. Diegel, P.D., Daniels, A.U., and Dunn, H.K., Initial effect of collarless stem stiffness on femoral bone
strain,

J. Arthroplasty

, 4, 173, 1989.
2. Englehardt, J.A. and Saha, S., Effects of femoral component section modulus on the stress distribution
in the proximal human femur,

Med. Biol. Eng. Comput.

, 26, 38, 1988.
3. Finlay, J.B., Chess, D.G., Hardie, W.R., et al., An evaluation of three loading congurations for the

in vitro

testing of femoral strains in total hip arthroplasty,

J. Orthop. Res.

, 9, 749, 1991.
4. Anderson, J.T., Erickson, J.M., Thompson, R.C., and Chao, E.Y., Pathologic femoral shaft fractures
comparing xation techniques using cement,

Clin. Orthop.,

131, 273, 1978.

0266/ ch10/frame Page 169 Thursday, October 21, 1999 2:42 PM

170

Mechanical Testing of Bone and the BoneImplant Interface

5. Laurence, M., Freeman, M.A., and Swanson, S.A.V., Engineering consideration in the internal xation
of fractures of the tibial shaft,

J. Bone Joint Surg.,

51B, 754, 1969.
6. Cordey, J. and Perren, S.M., Stress protection in femora plated by carbon ber and metallic plates;
mathematical analysis and experimental verication, in

Biomaterials and Biomechanics

, Ducheyne,
P., Van der Perre, G., and Aubert, A. E., Eds., Elsevier, Amsterdam, the Netherlands, 1984.
7. Crowninshield, R.D., Pederson, D.R., and Brand, R.A., A physiologically based criterion of muscle
force prediction in locomotion,

J. Biomech. Eng.,

102, 230, 1990.
8. Shybutt, G.T., Askew, M.J., Hori, R.Y., and Stulberg, S.D., Theoretical and experimental studies on
femoral stresses following surface replacement hip arthroplasty, in Hip,

Proceedings of the 8th Open
Scientic Meeting of the the Hip Society,

Frank Stincheld Award Paper 192, The Hip Society, C. V.
Mosby Co., St. Louis, MO, 1980.
9. Simon, B.R., Woo, S.L.-Y., Stanley, G.M., et al., Evaluation of one, two, and three dimensional nite
element and experimental models of internal xation plates,

J. Biomech.

, 10, 79, 1977.
10. Briggs, B.T. and Chao, E.Y.S., The mechanical performance of the standard HoffmanVidal external
xation apparatus,

J. Bone Joint Surg.

, 64A(4), 566, 1982.
11. Chao, E.Y. and An, K.N., Biomechanical analysis of external xation devices for the treatment of
open bone fractures in nite elements, in

Biomechanics

, Gallagher, R.H., Simon, B.R., Johnson, P.C.,
and Gross, J.F., Eds., John Wiley & Sons, New York, 1982.
12. Behrens, F., Johnson, W.D., Koch, T.W., and Kovacevic, N., Bending stiffness of unilateral and bilateral
xator frames,

Clin. Orthop.

, 178, 103, 1983.
13. Niederer, P.G. and Chiquet, C., Articial proximal femur of ber reinforced polyester for the study
of load transmission of cemented hip prostheses: the prosthesis cement interface,

Trans. Annu. Soc.
Biomater.,

2, 88, 1978.
14. McKellop, H., Ebramzadeh, E., Matta, J., et al., Stability of femoral fractures with interlocking
intramedullary rods,

Trans. Annu. Orthop. Res. Soc.,

319, 1986.
15. McKellop, H., Ebramzadeh, E., Niederer, P.O., and Sarmiento, A., Comparison of the stability of
press-t prosthesis femoral stems using a synthetic femur model,

J. Orthop. Res.

, 9, 297, 1991.
16. Uta, S., Development of synthetic bone models for the evaluation of fracture xation devices, Nippon
Seikegeka Gakai Zasshi, 66, 1156, 1992.
17. Noble, P.C., Alexander, J.W., Lindahl, L.J., et al., The anatomical basis of femoral component design,
Clin. Orthop., 235, 148, 1988.
18. Rubin, P.J., Leyvraz, P.F., Aubaniac J.M., et al., The morphology of the proximal femur, J. Bone Joint
Surg., 74B, 28, 1992.
19. Beals, N., Evaluation of a Composite Sawbones Femur Model, Research Report ML-87-25, Richards
Medical Company, Memphis, TN, 1987.
20. Szivek, J.A., Weng, M., and Karpman, R., Variability in the torsional and bending response of a
commercially available composite femur, J. Appl. Biomater., 1, 183, 1990.
21. Szivek, J.A. and Gealer, R.L., Comparison of the deformation response of synthetic and cadaveric
femora during simulated one-legged stance, J. Appl. Biomater., 2, 277, 1991.
22. Cristofolini, L., Viceconti, M., Cappello, A., and Toni, A., Mechanical validation of whole bone
composite femur models, J. Biomech., 29, 525, 1996.
23. Martens, M., Van Audekercke, R., Delport, P., et al., The mechanical characteristics of cancellous
bone at the upper femoral region, J. Biomech., 16, 971, 1983.
24. Szivek, J.A., Thomas, M., and Benjamin, J.B., Characterization of a synthetic foam as a model for
human cancellous bone, J. Appl. Biomater., 4, 269, 1993.
25. Otani, T., Whiteside, L.A., and White, S.E., Strain distribution in the proximal femur with exible
composite and metallic femoral components under axial and torsional loads, J. Biomed. Mater. Res.,
27, 575, 1993.
26. McNamara, B.P., Cristofolini, L., Toni, A., and Taylor, D., Evaluation of experimental and nite-
element models of synthetic and cadaveric femora for pre-clinical design-analysis, J. Clin. Mater.,
17, 131, 1995.
27. Szivek, J.A. and Yapp, R.A., A testing technique allowing cyclic application of axial, bending, and
torque loads to fracture plates to examine screw loosening, J. Biomed. Mater. Res. Appl. Biomater.,
23-A1, 105, 1989.
28. McLeish, R.D. and Charnley, J., Abduction forces in the one legged stance, J. Biomech., 3, 191, 1970.
0266/ ch10/frame Page 170 Thursday, October 21, 1999 2:42 PM
Synthetic Materials and Structures Used as Models for Bone 171
29. Measurements Group, Publication No. TN-514, 5, Raleigh, NC, 1988.
30. Colgan, D., Trench, P., Slemon, D., et al., A review of joint and muscle load simulation relevant to
in vitro stress analysis of the hip, Strain, 30, 47, 1994.
31. Carter, D.R. and Hayes, W.C., The compressive behavior of bone as a two-phase porous structure,
J. Bone Joint Surg., 7, 954, 1977.
32. Galante, J., Rostoker, W., and Ray, R.D., Physical properties of trabecular bone, Calcif. Tissue Res.,
5, 236, 1970.
33. Gibson, L.J., The mechanical behavior of cancellous bone, J. Biomech., 18, 317, 1985.
34. Goldstein, S.A., Wilson, D.L., Sonstigard, D.A., and Mathews, L., The mechanical properties of human
tibial trabecular bone as a function of metaphyseal location, J. Biomech., 16, 965, 1985.
35. Koeneman, J.B., Norman, J.P., and Szivek, J.A., The mechanical properties of cancellous bone in the
femoral head: correlation with CT measurements, Trans. 20th International Society for Biomaterials,
Kyoto, Japan, 11, 267, 1988.
36. Lindahl, O., Mechanical properties of dried defatted spongy bone, Acta Orthop. Scand., 47, 19, 1976.
37. Linde, F., Hvid, I., and Pongsoipetch, B., Energy absorptive properties of human trabecular bone
specimens during axial compression, J. Orthop. Res., 7, 432, 1989.
38. Ashman, R.B., Rho, J.Y., and Turner, C.H., Anatomical variation of orthotropic elastic moduli of the
proximal human tibia, J. Biomech., 22, 895, 1989.
39. Wixson, R.L., Elasky, N., and Lewis, J., Cancellous bone material properties in osteoarthritic and
rheumatoid total knee patients, J. Orthop. Res., 7, 885, 1989.
40. Brown, T.D. and Ferguson, A.B., Mechanical property distributions in the cancellous bone of the
human proximal femur, Acta Orthop. Scand., 51, 429, 1980.
41. Mosekilde, L., Consequences of the remodeling process for vertebral trabecular structure: a scanning
electron microscopy study (uncoupling of unloaded structures), Bone Miner., 10, 13, 1990.
42. Hein, T.J., Hotchkiss, R., Perissinotto, A., and Chao, E.Y., Analysis of bone model material for external
fracture xation experiments, J. Biomech. Instr., 22, 43, 1987.
43. Volz, R.G. and Lee, R.W., The effect of the stem and stem length on the mechanical stability of tibial
knee components, Trans. Orthop. Res. Soc., 13, 1988.
44. Lee, R.W., Volz, R.G., and Schroder, D.Q., Laboratory analysis of threaded acetabular cup stability,
Trans. Am. Acad. Orthop. Surg., New Orleans, LA, 1990.
45. Szivek, J.A., Thompson, J., and Benjamin, J.B., Three synthetic foams used to model a range of
properties of human cancellous bone, J. Appl. Biomater., 6, 125, 1995.
46. Thompson, J., Benjamin, J.B., and Szivek, J.A., Pullout strengths of cannulated and non-cannulated
cancellous bone screws: a comparative study, Clin. Orthop., 341, 241, 1997.
47. Odgaard, A., Hvid, I., and Linde, F., Compressive axial strain distributions in cancellous bone
specimens, J. Biomech., 22, 829, 1989.
48. Linde, F. and Hvid, I., The effect of constraint on the mechanical behavior of trabecular bone
specimens, J. Biomech., 22, 485, 1989.
49. Gibson, L.J. and Ashby, M.F., Material properties of cellular solids, in Cellular Solids: Structure and
Properties, Cambridge University Press, Cambridge, U.K., 1997.
50. ASTM Subcommittee F04.21 of committee F-04 on Medical and Surgical Materials and Devices,
Standard Specication for Rigid Polyurethane Foam for Use as a Standard Material for Testing
Orthopaedic Devices and Instruments, Designation F 1839-97, American Society for Testing and
Materials, West Conshohocken, PA, 1998.
0266/ ch10/frame Page 171 Thursday, October 21, 1999 2:42 PM
0266/ ch10/frame Page 172 Thursday, October 21, 1999 2:42 PM
Section II
Methods of Mechanical Testing
of Bone
0266/ ch10/frame Page 173 Thursday, October 21, 1999 2:42 PM
0266/ ch10/frame Page 174 Thursday, October 21, 1999 2:42 PM

175

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

11

Tensile and Compression Testing
of Bone

Tony S. Keller and Michael A. K. Liebschner

CONTENTS

I. Introduction ..........................................................................................................................175
II. Practical Considerations.......................................................................................................176
III. Specimen Preparation...........................................................................................................178
A. Preservation ....................................................................................................................178
B. Cutting and Machining of Bone ....................................................................................179
C. Specimen Geometry .......................................................................................................179
D. Potting of Specimens in Bone Cement..........................................................................181
IV. Standard Materials Testing Methods ...................................................................................181
A. General Considerations ..................................................................................................181
B. Tensile Testing................................................................................................................183
C. Compression Testing ......................................................................................................186
D. Fatigue Testing ...............................................................................................................188
E. Testing Whole-Bone Specimens ....................................................................................189
V. Equipment ............................................................................................................................190
VI. Measurement Procedures .....................................................................................................192
A. Compliance and Validation ............................................................................................192
B. Precision and Accuracy..................................................................................................193
C. Data Acquisition and Analysis.......................................................................................196
1. Determination of Mechanical and Structural Properties ..........................................196
2. Toe Region Compensation ........................................................................................198
VII. Animal Models.....................................................................................................................198
VIII. Summary ..............................................................................................................................199
References ......................................................................................................................................200

I. INTRODUCTION

Mechanical testing studies of bone have been directed at determining the mechanical properties of
whole bone and bone tissue under different loading conditions. In general, determination of mechan-
ical properties of bone is done by the same methods used to study similar properties in metal,
woods, and other structural materials and composites. These methods are based on fundamental
principles of mechanics.

1-4

Consequently, some basic knowledge of mechanics and the terminology
employed is essential in order to apply these principles.
Bone is a viscoelastic, composite material. The organization of the composite varies from
animal to animal and is strongly inuenced by aging, activity, and disease. Unlike engineering
composite materials, however, bone has a brous structural component (collagen) as its matrix and
exhibits a composite behavior microscopically as well as macroscopically. The main constituents

0266/ ch11/frame Page 175 Thursday, October 21, 1999 4:27 PM

176

Mechanical Testing of Bone and the BoneImplant Interface

of bone are mineral (hydroxyapatite,

dry weight,

volume), collagen (

dry weight,

volume), and water. At the whole bone or organ level, bone consists of a dense tissue (cortical
bone) which forms a stiff, hollow shaft coupled to a porous, less dense tissue (trabecular bone)
that is located adjacent to joint articulations and which acts to dissipate loads and absorb energy.
Thus, from a histological point of view, bone can be considered a composite material at both the
tissue and organ levels. Furthermore, since bone tissue is a living composite material, methods of
preservation, sectioning, and mechanical xation must also be considered in order to ensure that
reliable test results are obtained.
In this chapter, basic specimen preparation, standard materials testing, and stressstrain mea-
surement procedures are reviewed.

II. PRACTICAL CONSIDERATIONS

Bone tissue is part of a biological structure and its mechanical properties can only be fully
appreciated if one understands how the structure functions as a whole. The functional behavior
of bone is one of the most practical aspects that the investigator must consider prior to testing.
When one starts to investigate the mechanical properties of bone, therefore, it is rst necessary to
determine what structural level or levels of organization should be investigated. Starting at the
highest level of organization, forces acting on the whole body can be considered. At lower levels,
whole-bone, macromechanical specimens machined from whole-bone, micromechanical test spec-
imens, or single bone cells can be investigated.

5-11

To some extent, the specic interests of the
investigator may determine the level of organization to be studied. Some investigations may require
examination of the bone structure at several levels of organization.
Aging and disease processes produce signicant changes in the composition, geometry, and
architecture of bone, each of which is associated with alterations in the mechanical properties of
bone and consequently the response of bone to loading. If the goal is to study the fracture risk of
old bones compared with young bones, one could begin by testing the whole bone. Subsequently,
one may nd that the load required to break old bones is less than that required to fracture young
bones, but at this stage of the investigation intrinsic differences in the geometry, material quality,
and/or structural organization of the young vs. old bone tissue is unknown. Quantication of geometry
(histomorphometry) such as cortex thickness and cortical area are easily performed on the bone
specimens at this point, and subsequent statistical analyses may reveal a signicant correlation, for
example, between bone cortex thickness and age. Whole-bone structural properties, however, are
dependent upon both geometric and material properties. Thus, consideration of geometry alone may
not account for all the observed differences in the whole bone or structural strength of the bone
samples. Additional tests on machined bone samples may become necessary to determine how aging
and changing composition inuence bone mechanical properties. Tests conducted on machined bone
specimens reveal bone material properties (at least at the macrostructural level). On the other hand,
if the objective of the study is to determine how bone tissue histology inuences bone mechanical
behavior, one could begin by preparing machined specimens from whole bones, and then quantify
how the material properties of the specimens change with histology. At this point one may nd a
clear relationship between bone material properties and histology, but it may be difcult to predict
how bone material comprised of different histological types behaves as a composite. An investigator
studying bone must therefore consider the possibility that the answer to a specic research question
might be found only by investigating the properties of bone at several organization levels.
In principle, mechanical testing of bone is straightforward. Experimental results, however, can
be affected by specimen preparation and test methods used and by environmental conditions. In
particular, loading rate, deformation rate, specimen size, specimen shape, mode of loading, and the
method of gripping test specimens can inuence the mechanical response of bone and engineering
materials. Recognizing this, strict standards for testing engineering materials have been well
established. In the case of bone, however, standardized engineering materials testing methods cannot

0266/ ch11/frame Page 176 Thursday, October 21, 1999 4:27 PM

Tensile and Compression Testing of Bone

177

always be utilized due to restrictions imposed by the nite size of the bone specimens, difculties
in gripping the specimens, and/or relatively low loads that can be applied to bone. Nevertheless,
it is important to implement standards developed for testing engineering materials whenever pos-
sible. American Society for Testing and Materials (ASTM) designations for compressive testing
(ASTM C469, D1621), tensile testing (ASTM C565, D1623, D3039, D3044, E8, and E132), and
shear testing (ASTM D143) provide a source of mechanical testing techniques.

12-15

These can
generally be applied to bone, although modications to specimen size and method of gripping the
test specimens may be necessary.
Of the ASTM mechanical test methods available, the tension test is the easiest to apply
accurately both to cortical and to trabecular bone specimens. Trabecular bone specimens, with
dimensions implemented by Linde,

16

Keller et al.,

17

Keaveny et al.,

18

have become standard. Tra-
becular bone specimens are more difcult to grip in tensile testing. Two methods that have been
used successfully to grip dry trabecular bone tensile test specimens involve bonding the trabecular
bone specimen to at plates or inside brass tubing using epoxy or cyanoacrylate adhesive.

18,19

Stresses up to 500 kPa can be applied using cyanoacrylate.
The relationship between load applied to a structure and deformation in response to the load
is called a loaddeformation curve. The loaddeformation curve can be divided into two regions:
the elastic deformation region and the plastic deformation region. Within the elastic deformation
region the structure imitates a spring the geometric deformation in the structure increases linearly
with increasing load and, after the load is released, the structure returns to its original shape. The
slope of the elastic region of the loaddeformation curve represents the extrinsic stiffness or rigidity
of the structure. Larger structures will have greater rigidity than smaller structures of similar
composition. Load and deformation can be converted to stress and strain by engineering formulae.
The slope of the resulting stressstrain curve within the elastic region is called the modulus of
elasticity or Youngs modulus. Youngs modulus is independent of specimen size and is therefore
a measure of the intrinsic stiffness of the material.
stiffness = force/deformation (11.1)
elastic modulus = stress/strain (11.2)
The denition of stiffness for trabecular bone is more difcult. Trabecular bone is a two-phase,
porous, composite structure consisting of individual trabeculae organized in a lattice structure and
marrow made up of cells, fat, and vessels. Whereas individual trabeculae are relatively uniform,
the lattice structure may exhibit a large degree of variability in terms of porosity, structural
orientation, and connectivity.

20-24

Indeed, the bony elements of trabecular bone can be organized
as open- and closed-cell rodlike or platelike lattice structures. Thus, trabecular bone forms a complex
structure that has its own unique stiffness. Consequently, trabecular bone exhibits both an intrinsic
or material stiffness, which is the stiffness of an individual trabeculae and an extrinsic or
structural stiffness, which is the stiffness of the trabecular structure.

25-27

Most biomechanical
studies of trabecular bone concentrate on structural properties because material properties of
individual trabeculae are difcult to measure. Structural properties, however, can vary appreciably
(several orders of magnitude) for different anatomical regions, and are closely dependent upon the
density, distribution, and orientation of the trabeculae.
Analysis of loaddeformation and stressstrain behavior in bone is further complicated by the
fact that bones and other biological materials do not behave as a perfect spring. Rather, most
biological materials exhibit nonlinear loaddeformation and stressstrain behavior, which are fur-
ther inuenced by loading rate and temperature. Such behavior is termed

viscoelastic

and is the
result of internal energy losses due to friction in the structure (intrinsic viscoelasticity) or uid ow
(uid-dependent viscoelasticity) during deformation. Bone exhibits only a slight degree of
viscoelasticity, and it is therefore reasonable to treat bone as a linear-elastic or springlike (Hookean)

0266/ ch11/frame Page 177 Thursday, October 21, 1999 4:27 PM

178

Mechanical Testing of Bone and the BoneImplant Interface

material. Alternatively, nonlinear stress analyses may be implemented and the stressstrain depen-
dency on the applied stress can be quantied.
These simple examples and basic mechanical testing denitions are intended to point out that
neither testing the whole bone nor testing of standardized bone specimens alone may prove a
complete or practical answer to questions of how aging affects bone fracture risk. Summarizing
these points, practical evaluation of whole or machined bone mechanical functionality should
include considerations of the following:
Specimen geometry;
Effect of loading type;
Effect of test conditions (hydration) and preparation;
Directional properties of the bone tissue;
Composite nature of bone tissue;
Composition of bone tissue;
Lattice structure of trabecular bone;
Viscoelastic properties of bone tissue;
Nonlinear loaddeformation and stressstrain behavior.
These variables inuence and govern the overall mechanical properties of bone.

III. SPECIMEN PREPARATION
A. P

RESERVATION

Water (matrix) accounts for approximately 6% of the weight and 11% of the volume of hydrated
bone. Thus, changes in water content have a signicant effect on mechanical properties. Mechanical
properties have been shown to vary signicantly depending upon the storage and handling proce-
dures used following removal of the tissue from the body. Changes on the order of 10% are not
uncommon.

28-31

Any treatment of bone, which changes the nature or relative composition of these
components, can inuence mechanical behavior. Thus, drying, freezing, storage in saline or alcohol
solutions, and embalming affect the properties of bone.

32,33

The properties of fresh tissue can vary in
a short period of time if bone is allowed to dry. For example, bone specimens maintained at room
temperature for 24 h without preservation will demonstrate about a 3% decline in Youngs modulus.

34

For optimal preservation of bone physical and mechanical properties, the following storage
methods are recommended. For long-term storage, bone should be frozen and kept as moist and
hydrated as possible. In order to minimize freeze-drying of the bone tissue, the surrounding muscu-
lature should be left intact, and a plastic wrap and bag should cover the musculature to further minimize
freeze-drying and freezer burn. If the musculature and surrounding soft tissues must be removed
before freezing, the bone tissue should be wrapped in gauze, soaked in normal saline, wrapped with
plastic wrap, and placed in sealed, airtight plastic bags. The bone tissue should be placed in the freezer
within 1 h after it has been harvested and stored at 20

C. Upon removal from the freezer and during
all stages of tissue preparation, the bone should be kept hydrated in saline.
For time periods of up to 3 months, small specimens (including machined bone) may be
preserved at room temperature in a solution of 50% saline and 50% alcohol, or in biostatic saline.
Preservation of bone in an ethanol/saline solution results in minimal changes in the mechanical
properties of bone. Ashman

35

found that keeping samples in 50% ethanol and 50% saline solutions
for up to 90 days resulted in less than a 2% decline in Youngs modulus. Sedlin and Hirsch

34

found
ethanol to be somewhat less effective as a preservative. Bone samples stored in 40% ethanol for 5
to 10 days had a 2.5 to 4% decrease in Youngs modulus. Bone specimens preserved in ethanol
solution will lose some residual water, so it is important to soak them in an isotonic saline solution
for several hours prior to testing, during which time the specimens should be refrigerated.

0266/ ch11/frame Page 178 Thursday, October 21, 1999 4:27 PM

Tensile and Compression Testing of Bone

179

Bone test specimens can also be xed in formalin or glutaraldehyde. Fixation in this manner
increases collagen cross-linking and will, therefore, alter the properties of the bone tissue more
signicantly than alcohol preservation.

34,36

Sedlin and Hirsch

34

reported that embalmed bone exhibits
different values of strength and elasticity compared with fresh tissue. Evans

37

reported a 68%
increase in Youngs modulus and ultimate tensile strength. McElhaney et al.

36

found a 1 to 9%
decrease in tensile strength and a 12 to 18% decrease in the compressive strength of bovine bone.
Although the results of these two studies are equivocal, both indicate that embalming dramatically
alters the mechanical behavior of bone. Mechanical testing of formalin-xed samples only provides
data relative to other xed samples, and results from individual samples will not provide an accurate
measure of the true properties of bone. Whenever possible, therefore, bone tissue should be tested
in an unembalmed state. Unquestionably, the best method of long-term preservation is to store
saline-soaked, gauze-wrapped specimens in airtight bags or containers at 20

C. Embalmed tissue
should be stored in a manner similar to unembalmed tissue.

B. C

UTTING



AND

M

ACHINING



OF

B

ONE

Cutting and machining of bone samples can be one of the most time-consuming steps in preparation
of bone specimens for mechanical testing (Figure 11.1). Rough cuts can be made through bone
with a band saw, hacksaw, or a jigsaw, but it can be difcult to keep the bone moist while cutting
and when using a band saw it is easy to overheat and even burn the bone tissue. Damage due to
overheating during rough-cutting procedures generally affects an area of bone only 1 to 2 mm from
the cut. The affected area can be removed using wet sandpaper. Alternatively, ner cuts can be
made using a diamond-impregnated wire saw or a low-speed diamond-impregnated wafering saw.
Diamond wafering saws produce smooth parallel cuts and are particularly well suited for preparing
rectangular test specimens or for creating coplanar surfaces in cylindrical specimens. Cylindrical
specimens can be cored using a diamond-coring tool. When using a coring tool, the specimen and
tool should be completely immersed in a water or saline bath. For more intricate machining a
vertical end mill or lathe can be used. When milling or lathing cortical or trabecular bone, cutting
rates similar to those suggested for aluminum should be used. Irrigation with water or saline is
necessary to prevent overheating of the bone samples during the machining process. Specimen
surfaces should be examined microscopically for cracks and other defects caused by machining,
cutting, or sawing. Radiographs of machined and sectioned bone specimens can also be used to
identify cracks and voids within the specimens.

C. S

PECIMEN

G

EOMETRY

A key factor that needs to be considered when preparing bone samples for mechanical testing is
specimen geometry. The most common specimen geometries are cubes with a side length between
6 and 8 mm or cylinders with a length-to-diameter (L/D) ratio between 1 and 2 and a diameter
between 6 and 8 mm (Figure 11.2). The extremes of dimensions for cubic specimens reported in
the literature are 4.5 mm

38

and 10 mm.

39

Cylindrical specimens with a L/D ratio between 2

40

and

41

have been used with diameters ranging from 5 mm

42

to 20 mm,

41

and lengths ranging from
2.75 mm

40

to 12 mm.

43

Considerations of specimen geometry are particularly important for tests
conducted on trabecular bone, which has been shown to be very susceptible to experimental artifacts
during compression testing.

40,44-48

In a recent study, Keaveny et al.

46

performed a theoretical analysis of the effect of friction
(between the bone specimen and compression platen) and the damage artifact (structural end phe-
nomenon associated with cut surfaces) on the experimental determination of Youngs modulus. These
investigators found that Youngs modulus determinations were signicantly underestimated for
certain bone specimen geometries. They noted that a specimen with an aspect ratio of 2:1 was least
sensitive to the combined effects of friction and the damage artifact on modulus underestimation.

0266/ ch11/frame Page 179 Thursday, October 21, 1999 4:27 PM

180

Mechanical Testing of Bone and the BoneImplant Interface

They also argued against the use of cubic specimen geometry as a standard in biomechanical testing
because cylinders can be made more easily and accurately than cubes, and because the surface-to-
volume ratio is lower for cylinders than for cubes with the same aspect ratio (equal width). Both of
these factors permit more accurate density measurements to be performed when using cylinders with
a 2 : 1 aspect ratio. Furthermore, they found evidence that the correlation coefcient in the modu-
lusdensity regression increases as the aspect ratio increases, and noted that the 2 : 1 aspect ratio
cylinder was superior to a cube in that respect. They also found more accurate (lower standard error
of the estimate) predictions of modulus and strength could be made using 2 : 1 aspect ratio cylinders
than with cubes of the same width. Experimental support for the theoretical analysis performed by
Keaveny et al.

46

was provided from a study by Choi et al.

49

Zhu and associates

48

performed a
comprehensive series of experiments on open-cell foams and human trabecular bone specimens and
developed a surface damage theory to explain the modulus underestimation associated with the cut
surface structural end phenomenon in porous materials such as trabecular bone. They recommended
that the overall height of trabecular bone compression test specimens should be at least 10 mm.
Linde et al.

40

also pointed out that the diameter of trabecular bone specimens should be large enough
to satisfy continuum scale assumptions,

45,50,51

but at the same time should be small enough to ensure
that specimen homogeneity was preserved.

52-54

Based on these and the aforementioned nding by

FIGURE 11.1

Trabecular bone specimen preparation for tensile or compressive mechanical testing. Steps
include slicing of the bone section, X-raying of slices to avoid voids and precracks of specimens, cutting of
cubic specimens with band saw and coring cylindrical specimens, density analysis using QCT, alignment of
specimen in sockets, and nal preparation of reduced diameter gauge section using a low-speed lathe.

FIGURE 11.2

Geometric properties of test specimens commonly used in biomechanical testing.

0266/ ch11/frame Page 180 Thursday, October 21, 1999 4:27 PM

Tensile and Compression Testing of Bone

181

Keaveny and associates, a 5-mm-diameter, 2 : 1 aspect ratio cylindrical specimen with a specimen
height of at least 10 mm would appear to be optimal for compression testing of human and bovine
trabecular bone.

D. P

OTTING



OF

S

PECIMENS



IN

B

ONE

C

EMENT

Bone specimens are often potted in cement (polymethylmethacrylate or PMMA) in order to obtain
a reliable grip interface for the testing apparatus. However, some precautions should be taken when
using potting procedures. First, removal of bone marrow and fat is essential for adequate bonding
of PMMA to bone surfaces. Bone marrow and fat can be removed mechanically using a water jet
or air jet or the fat can be removed chemically using a fat-dissolving detergent or alcohol or
trichloroethylene. Usually, a combination of a mechanical and a chemical method provides the best
results. The following is the suggested method for preparing bone specimens for potting:
Clean the specimen mechanically with a high pressure water jet at the region to be cast.
Defat the specimen at that particular region in trichlorethylene or 10% bleach + 90%
water solution in an ultrasound bath for about 10 min.
Air-dry the region to be cast using an air jet.
For small specimens and for low-force mechanical tests (e.g., nondestructive testing) the
specimen can be potted immediately. For high-force mechanical tests (e.g., destructive testing) or
testing of cortical bone samples with smooth surfaces, several layers of cyanoacrylate cement should
be applied on the bone surface. Cyanoacrylate cement provides a stronger bond between the bone
and the PMMA potting material. Specimens should be rehydrated immediately following the surface
preparation and potting procedures in order to prevent any strength loss associated with preparation-
related specimen dehydration.

55-57

In addition, the potting mold volume should not exceed three
times the volume of the bone region that is embedded in bone cement. By using this rule of thumb,
overheating of the specimen during the heat curing process of the PMMA (

60

C) will be mini-
mized. This also minimizes shrinkage of the potting mold. For bone tissue tests that include a
bonePMMA interface it may also be necessary to account for the material properties of PMMA.
However, the authors are aware of only one study specically conducted on bonePMMA composite
properties

58

and a similar study on planar reinforced plastic resin.

59

IV. STANDARD MATERIALS TESTING METHODS

The strength of human cortical bone varies depending upon the kind of stress applied to the bone.

60

The ultimate tensile strength of femoral bone in the longitudinal direction is 135 MPa, the ultimate
compressive strength is 205 MPa, and the shear strength is 67 MPa.

61

Like the Youngs modulus,
the strength of cortical bone also varies with direction. The tensile strength of the femur in the
transverse direction is only 53 MPa, compared with 135 MPa in the longitudinal direction.

61

Tensile
strength in trabecular bone can vary from 1 to over 20 MPa, and is strongly dependent upon apparent
density and trabecular orientation.

39,48,62,63

The mechanical behavior of bone is also loading-rate
and strain-rate dependent.

64-70

A. G

ENERAL

C

ONSIDERATIONS

In bone, like wood and many other biological structures, there is a grain or preferred direction
associated with the structure. Consequently, the mechanical behavior of bone and other directional
composites is dependent upon the direction of the applied load. Materials that have different
properties in different directions are termed

anisotropic

, and as many as 21 independent elastic
constants are required to characterize their mechanical behavior completely. Most materials have

0266/ ch11/frame Page 181 Thursday, October 21, 1999 4:27 PM

182

Mechanical Testing of Bone and the BoneImplant Interface

planes of symmetry that reduce the number of material constants. For example, materials having
properties that differ in each of three mutually perpendicular directions are termed

orthotropic

, and
nine elastic constants are required to characterize their mechanical behavior fully. Plexiform bone
(e.g., bovine femur) is an example of a tissue with orthotropic material symmetry. Materials that
have properties that are constant within a given plane are termed

transversely isotropic

. Human
osteonal bone is an example of a transversely isotropic material because it has the same Youngs
modulus in all transverse directions, but has a higher Youngs modulus in the longitudinal direc-
tion.

61,71-76

Materials that have the same elastic properties in all directions have the highest order
of symmetry and are termed

isotropic

.
Complete characterization of the mechanical behavior of anisotropic material properties
requires mechanical testing to be performed in several different orientations. For example, to
determine the nine independent elastic coefcients of an orthotropic material the following mechan-
ical tests are required:

73,77-79

1. Tensile or compressive tests in each of three mutually perpendicular material directions;
2. Three lateral deection tests to obtain Poissons ratios; and
3. Three torsion tests to obtain shear moduli.
Ideally, mechanical test specimens should be oriented relative to the axes of material symmetry.
In the case of cortical bone, the mutually perpendicular axes of material symmetry are generally one
axis oriented parallel to the long axis of the bone, another oriented radially outward from the center,
and the third oriented in a circumferential direction. In order to assure that specimens are cut in the
proper orientation, the axes of material symmetry must be determined prior to testing, and are typically
dened based upon histology. Loads applied in tension, compression, torsion, shear, and in combined
modes on specimens cut in many different orientations are necessary to describe the anisotropic failure
surface completely.

80-82

Table 11.1 summarizes the orthotropic and transversely isotropic material
properties that have been reported for the human femur and tibia, respectively.

TABLE 11.1
Engineering Elastic Constants for
Human Bone

Femur

a

Tibia

b

E

1

(GPa) 11.5 6.91
E

2

(GPa) 11.5 8.51
E

3

(GPa) 17.0 18.4
G

12

(GPa) 3.6 2.41
G

31

(GPa) 3.28 3.56
G

23

(GPa) 3.28 4.91
?

12

0.58 0.49
?

13

0.31 0.12
?

23

0.31 0.14
?

21

0.58 0.62
?

31

0.46 0.32
?

32

0.46 0.31

a

Reilly, S.B., Burstein, A.H., Frankel, V.H., The elastic
modulus of bone,

J. Biomech

., 7, 271, 1974.

b

Obrazcov, I., Adamovich, I., Burer, A., Knets, I. et al.,

The Problems of Strength in Biomechanics

, Vishaya
Shkola Publ. House, Moscow, 1988 (in Russian).

0266/ ch11/frame Page 182 Thursday, October 21, 1999 4:27 PM

Tensile and Compression Testing of Bone

183

Several standards from the ASTM have been adapted for mechanical testing procedures on
biological tissue.

84-90

The specimen shape most widely used in testing bone tissue is the so-called dog
bone specimen, and the end portions of these specimens are enlarged diameter regions where the
specimen will be attached to the testing apparatus. The test region or gauge section consists of a
turned-down section of decreased cross-sectional area, and ideally it is in this region that the specimen
should fail. If the specimen fails outside of the gauge region or near the edge of the gauge region,
this is an indication that alignment errors or other interface artifacts caused premature failure of the
specimen. The ultimate strength of the material being tested is underestimated when this occurs.
During mechanical testing of biological tissues, the test specimen should be saturated with
saline, particularly during long-term fatigue tests.

B. T

ENSILE

T

ESTING

Tensile testing can be one of the most accurate methods for measuring bone properties, but bone
specimens must be relatively large and should be carefully machined. Tensile test specimens for
cortical bone and cancellous bone are illustrated in Figures 11.3 and 11.4.
Dimensions are derived from ASTM standards. In Figure 11.3, the ratio

d/D

should be around

and the parallel length of the narrow section should be at least three times the size of the gauge
diameter

d

. The radius of curvature

R

should be very large to avoid stress concentrations and should
have the same dimensions as the parallel length

A

. The grip length

M

is one quarter of the whole
specimen length

L

. Because of the relatively homogeneous microstructure of cortical bone, cortical
bone specimens can be made comparatively small in size (gauge diameter =

d

3 mm).
In principle, the same geometry used for cortical bone tensile test specimens applies to trabe-
cular bone.

91

However, because of the intrinsic lattice and inhomogeneous structure of trabecular
bone, a minimum gauge diameter of 5 mm is required to ensure that continuum scale criteria are met.
Tensile test specimens are designed so that the highest strains will occur in the central portion or
gauge region of the specimen. Strain measurements can be obtained by attaching a clip-on extensom-
eter to the gauge section of the specimen (refer to Figure 11.1). Stress is calculated as the applied
force divided by the bone cross-sectional area measured in the specimen midsection. Assuming that
FIGURE 11.3 Tensile test specimen geometry for cortical bone tests. A = parallel length, GL = gauge length,
M = grip length, E = neck length, R = curvature radius, D = specimen outer diameter, d = specimen gauge
diameter.
0266/ ch11/frame Page 183 Thursday, October 21, 1999 4:27 PM
184 Mechanical Testing of Bone and the BoneImplant Interface
the force is applied without inducing a coupled bending moment, a tensile test will provide a very
accurate measurement of the bone mechanical properties. A technical concern associated with tension
testing is bending imposed on the specimen. Incorporating pivoting elements in the loading chain
typically reduces bending. Attaching universal joints onto the sockets or self-aligning socket holders
can accommodate eccentricities that may be present during specimen mounting and will therefore
minimize bending artifacts. To correct for residual bending, four stressstrain tests should be run with
the extensometer placed on each of the four sides of the specimen. The four moduli are then averaged.
If strain is only measured in a single plane, bending can signicantly affect the measured modulus.
92
The elastic portion of the stressstrain curve is characterized by a straight line (Hookes law)
and the slope of this line, or the ratio of stressstrain within the elastic range, is dened as the
modulus of elasticity E (Youngs modulus). As the stress is increased, a point is reached where a
further increase in stress will show a departure of the curve from the straight line. The greatest
stress intensity for which stress is still proportional to strain is called the proportional-elastic limit
(indicated by PEL in Figure 11.5). This is not strictly the same as the elastic limit. The elastic limit
is dened as the greatest stress that can be applied without leaving a permanent deformation upon
complete release of the load. To determine the elastic limit, it is necessary to load and unload the
specimen with increasing values of the load until a permanent set is found after complete unloading.
Since this procedure is time-consuming and since the elastic limit differs little from the proportional
limit, the true elastic limit is seldom obtained in actual practice.
If the stress is increased further from the proportional limit, the stressstrain curve departs more
and more from the straight line. Unloading the specimen at point A (Figure 11.5), the portion AB is
linear and is essentially parallel to the original line OC. The horizontal distance OB is the plastic
deformation corresponding to the stress at A. This is the basis for the construction of the arbitrary
(offset) yield strength. To determine the offset yield strength, a straight line AB is drawn parallel to
the initial elastic line OC but is displaced from it by an arbitrary value of permanent strain. The
permanent strain most commonly used is 0.2% of the original gauge length. When reporting the yield
strength, the permanent strain value should be specied. The arbitrary yield strength is typically used
for those materials that do not exhibit a natural yield point, but is not necessarily limited to such
materials. For analysis of brittle materials like cortical bone, one may need to choose a permanent
strain value lower than 0.2%, since cortical bone typically has a relatively low failure strain (<3%).
FIGURE 11.4 Tensile test specimen geometry for trabecular bone tests.
0266/ ch11/frame Page 184 Thursday, October 21, 1999 4:27 PM
Tensile and Compression Testing of Bone 185
The actual yield point in a stressstrain curve has been dened as the stress for which a marked
increase in strain occurs without a corresponding increase in stress. In such materials both an upper
and a lower yield point are usually identied. Following initial yielding, the stress drops and the
curve remains approximately horizontal for a period of deformation before it begins to rise again.
The upper yield point (
ypH
) is the stress level at which the initial drop occurs. The lower yield
point (
ypL
) is taken as the lowest value of stress after the initial drop-off and before the load begins
to rise continuously or, more properly, as the average stress during this interval (Figure 11.6).
ASTM standards E6-36 specify that the term yield point should not be used in connection with
FIGURE 11.5 Typical stressstrain curve for tensile test.
FIGURE 11.6 Typical stressstrain curve with visible yielding.
0266/ ch11/frame Page 185 Thursday, October 21, 1999 4:27 PM
186 Mechanical Testing of Bone and the BoneImplant Interface
materials where the stressstrain diagram does not become horizontal or does not show an actual
drop of stress with increasing strain in the yield region.
Upon further deformation, the load reaches a maximum value and then drops somewhat before
fracture occurs. The tensile strength (ultimate tensile stress) is obtained by dividing the maximum
load during the test by the original cross-sectional area. A measure of the ductility of a material
after fracture is given by the percent elongation and also by the reduction of cross-sectional area.
Percent elongation after fracture is determined by dividing the change in the original gauge length
by the original gauge length (multiplied by 100%). The original gauge length should always be
stated in reporting the percent elongation values. The percent reduction of area after fracture is the
ratio of the change in the original area determined at the smallest cross section divided by the
original area of cross section (multiplied by 100%).
As discussed earlier, there is a difference between intrinsic stiffness and extrinsic stiffness
(rigidity). For a tensile test of bone the intrinsic stiffness is equal to the Youngs modulus (E), while
the extrinsic stiffness is equal to EAL, where A is the cross-sectional area of the specimen and L
is the gauge length of the specimen. The extrinsic stiffness is dependent not only upon elasticity,
but also on specimen size. In the case of trabecular bone the specimen must be large enough that
the trabecular structure can be treated as a continuum. A specimen width of at least 4 to 8 mm is
recommended for trabecular bone specimens.
51,93-96
C. COMPRESSION TESTING
Compression testing of bone specimens is a popular technique, especially for cortical bone because
relatively small specimens can be used. Compressive tests, however, tend to be less accurate than
tensile tests due to friction and compression-platen end effects imposed on the bone specimen
during the test. Friction at the load platenbone surface interface can be minimized using polished
stainless steel platens lubricated with a coating of lightweight machine oil. A surface roughness of
2 m cm
1
is recommended.
48
If the load faces of the bone specimen are slightly misaligned with
respect to the compression loading platen, then large stress concentrations can occur, resulting in
18618an underestimation of both Youngs modulus and compressive strength. Placement of a
pivoting platen in the load train reduces misalignment error (Figure 11.7). By using a micrometer,
the parallelism of the load-contacting surfaces of each specimen can be assessed by measuring the
height differences between each of the four sides and a central point of the load-contacting surfaces.
Four height differences, recorded by assigning a value of zero to the lowest point among the ve
points, can be utilized to determine the parallelism index.
48
I = G (D
max
+ D
max
D
1
+ D
max
D
2
+ D
max
D
3
)
I = D
max
G (D
1
+ D
2
+ D
3
) (11.3)
where D
max
is the maximum height difference, D
1
, D
2
, and D
3
are the other height differences
ranging from 0 to D
max
, and the factor is used to obtain an averaged nominal height difference.
This index considers both the absolute differences between each point and zero reference point,
and the relative difference between individual points. The higher the index, the greater the irregu-
larity of the surface plane. This index may be normalized by the linear dimension of a cross-
sectional area of a sample (width or diameter) to indicate a nominal atness of a contacting surface.
Another problem associated with compressive testing of trabecular bone is the previously noted
end effect created by cutting or machining the faces of trabecular bone test specimens.
53
At the
boundary where the specimen contacts the loading platen, the cut surfaces of the trabeculae lattice
are unsupported, and the strain tends to be much greater in the boundary region than in the middle
of the specimen.
48
Elevated strains at the ends of the specimen result in an overestimation of the
average specimen strain and concomitant underestimation of modulus. Thus, simple strain calcu-
0266/ ch11/frame Page 186 Thursday, October 21, 1999 4:27 PM
Tensile and Compression Testing of Bone 187
lations tend to be inaccurate. More accurate specimen strain measurements can be obtained by
directly measuring the local strain at the midsection of the specimen. Mechanical or optical
extensometers can be used for local strain measurements in trabecular bone.
Although it is considerably more difcult to achieve accurate results using a compressive test
compared with a tensile test, the compressive test has several advantages. First, compression test
specimens need not be as large as tensile specimens, which is a major advantage when testing
trabecular bone. Second, fabrication of compressive specimens is not as difcult as fabrication of
tensile test specimens. Finally, in some regions of the skeleton (e.g., the vertebrae) compressive
tests may more closely mimic the in vivo loading conditions to which the bone is exposed. Even
with measurement error, compression tests are often very precise, particularly if one is simply
interested in comparing data from experimental and control groups (assuming the measurement
error did not change as a result of the treatment).
In a compression test, most ductile materials have stressstrain responses that are very similar to
tensile stressstrain responses during the initial phases of testing. As the area of cross section increases
(due to the Poisson effect), however, the stressstrain curve usually shows a gradual increase in slope,
and does not exhibit the plastic deformation which is characteristic of the tension test. In addition, the
compressive stressstrain curve does not necessarily reach an analytic maximum as in the case of
tension, and unless shearing, splitting, or crumbling occurs there may not be any overt fracture.
Consequently, no denite compressive strength may be noted and compressive strength has no
real meaning in such cases. Rather, porous composites and porous bone specimens will exhibit a
crushing phenomenon associated with progressive pore collapse and stabilization (Figure 11.8). In
this case failure is generally dened as the point at which pore collapse is rst observed (typically
the point at which the initial drop in load is observed). Fracture under compression does not occur
in ductile materials since the material merely ows laterally as the height is decreasing. Thus, the
denition of compressive strength depends upon the degree of distortion that is regarded as
indicating failure of the material. In fact, many plastic materials will continue to deform in
compression until a at disk is produced. In such cases, the compressive stress (nominal) rises
steadily without any well-dened fracture occurring. Fracture is usually of the shear type with
sliding along inclined planes starting at the surface on which the pressure is applied. Low ductility
or brittle materials may not necessarily have well-dened yield points, but do exhibit denite fail
points, since they fail in compression by a shattering type of fracture. Fracture of brittle material
depends upon
1. the ratio of height to lateral dimension (aspect ratio),
2. friction between the compression platens and the specimen, and
3. the shape of the specimen.
FIGURE 11.7 Spherical socket used to compensation for nonparallel load-bearing surfaces during compres-
sion testing.
0266/ ch11/frame Page 187 Thursday, October 21, 1999 4:27 PM
188 Mechanical Testing of Bone and the BoneImplant Interface
D. FATIGUE TESTING
Fatigue refers to the failure of materials under the action of repeated stresses. Fatigue is the result
of slip occurring along certain crystallographic directions accompanied by local crystal fragmen-
tation rupturing the atomic bonds, culminating in the formation of submicroscopic cracks, which
soon become visible cracks.
98-104
Standard testing methods are important since fatigue susceptibility of a material is dependent on
specimen shape, size, and cyclic rate. Standard fatigue testing techniques have been established for
engineering materials (ASTM C394, D671, D3166, and E206), but because of the size limitations,
standard fatigue testing methods must be modied for bone material. In fatigue testing, the test specimen
is subjected to periodically varying stresses by means of mechanical devices. The applied stresses may
alternate between equal positive and negative values (fully reversed cyclic fatigue tests), from zero to
maximum positive or negative values, or between unequal positive and negative values. Cyclic fatigue
testing has been applied to both cortical and trabecular bone in uctuating axial tension (0 to Tension),
completely reversed axial loading (Compression to Tension), uctuating bending (0 to Moment), and
completely reversed bending (Moment to +Moment).
41,62,101,105,106
A series of fatigue tests is usually made on a number of specimens of the material at different
stress levels.
107
The stress level endured is then plotted against the number of cycles sustained, and
the resulting diagram is called the stress-cycle diagram or SN diagram (Figure 11.9). From SN
curves, it is possible to predict fatigue failure at a particular stress level. By choosing lower and
lower stresses, a value may be found which will not produce failure, regardless of the number of
applied cycles. This stress value is called the endurance limit. The endurance limit may be estab-
lished for most materials between 2 and 10 million cycles. Surface defects such as scratches or
notches as well as surface roughness will reduce the fatigue strength of a specimen. In general,
fatigue specimens are prepared with a turned-down or gauge section to ensure fracture will occur
in the gauge region. Fatigue testing, more so than static mechanical testing, is dependent on
specimen shape. The radius of curvature of the neck portion of the specimen and surface nish are
critical factors inuencing the fatigue life of the specimen.
As in single cycle, uniaxial, tensile, and compression mechanical testing, fatigue test results
are strain-rate dependent. Cyclic loading frequencies from 2 to 125 Hz have been used to charac-
terize the fatigue behavior of cortical bone.
41,62,101,106,108
Lafferty and Raju
106
presented corrections
for the effects of cyclic rates on fatigue life. Load or deections can be applied to the fatigue
specimens, although the preferred method is to apply constant loads. Failure under load-control
FIGURE 11.8 Fracture line on a 10-mm thick vertebral body section.
0266/ ch11/frame Page 188 Thursday, October 21, 1999 4:27 PM
Tensile and Compression Testing of Bone 189
fatigue testing occurs as a distinct fracture. Specimens subjected to controlled deection fail in a
more gradual manner due to relief of stress (stress relaxation) and decreased stiffness. Failure under
controlled deection or strain is often dened as the point at which stress has been reduced to 70%
of its original value.
64,101

E. TESTING WHOLE-BONE SPECIMENS
Performing tensile, compressive, or torsion tests on whole-bone specimens has the added difculty of
attaching the specimen to the testing machine. Whole-bone mechanical test specimens do not have a
nice prismatic or symmetrical shape like the machined test specimens used for the standardized tensile,
compression, and shear tests discussed in the previous section. Thus, special xtures and casting
procedures are required (Figure 11.10).
109
The cast should be symmetrical to allow easy attachment to
the testing apparatus. Bone cement or epoxy resin is generally preferred as the casting material, and
commonly used materials include methylmethacrylate and plastic padding (Figure 11.11).
In some cases an intact bone structure such as the vertebral motion segment (two vertebral bodies
and the intervertebral disk) are placed between two compression platens and no casting material is
used.
112-117
In such cases, a rough surface nish on the platens is used to prevent slipping during the
compression test. However, such procedures may lead to asymmetrical loading of the specimen and can
contribute to measurement error, typically seen as underestimation of the stiffness of the vertebral body.
For whole-bone mechanical test specimens the most common method of strain determination
is measurement of the crosshead displacement. However, it is also possible to attach strain-
measuring devices (strain gauges or extensometers) directly to the specimen,
111
or one can perform
noncontact strain measurements using an optical motion analysis system.
118-120
Optical systems,
based on measuring the translation of points on the surface by image analysis, have shown increasing
success in accurately determining the mechanical properties of bone. By using CCD-cameras, the
resolution of the optical system can be better than 0.01% of the eld of view (FOV) provided that
there is good contrast between the specimen and targets. The targets captured by the camera can
be passive reective markers, active markers (light emitting diodes or LEDs), or ink lines drawn
by the test specimen. By using reective markers it is possible to create an array of targets, which
can be used to measure surface deections in the direction of loading as well as in directions normal
to the loading axis. With two cameras, laser tracking systems or magnetic eld-based devices,
three-dimensional data of an area of interest can be recorded and analyzed. Bonded strain gauges
and acoustic measurement devices can also be used to quantify whole-bone deformation,
121,122
and
are described later in this chapter.
FIGURE 11.9 Comparison of stress amplitude vs. cycles to failure in fatigue study. Both axes are
logarithmic scale.
0266/ ch11/frame Page 189 Thursday, October 21, 1999 4:27 PM
190 Mechanical Testing of Bone and the BoneImplant Interface
V. EQUIPMENT
Machines and apparatus for the mechanical testing of materials usually contain the following elements:
(1) specimen gripping mechanism; (2) specimen loading mechanism; and (3) load measurement
transducer. Some machines include an integrated apparatus for measurement of both load and defor-
mation, while others may rely on an auxiliary apparatus to record the specimen deformation (e.g.,
extensometer, strain gauge). Testing machines can be screw driven, pneumatic, or hydraulic. More
sophisticated testing apparatus tends to be servo-controlled. Servo-based testing machines can apply
loads or deections to bone specimens at a variety of different rates and magnitudes under feedback
FIGURE 11.10 Mounted rodent whole-bone test specimen.
110,111
FL = free length, GL = arbitrary chosen
gauge length.
FIGURE 11.11 Whole vertebral body prepared for compression testing.
0266/ ch11/frame Page 190 Thursday, October 21, 1999 4:27 PM

Tensile and Compression Testing of Bone

191

control. In most general-purpose mechanical testing systems, deformation can be controlled as the
independent variable and the resulting load measured. In other systems, particularly those intended
for use with low loads, the load is controlled and the resulting deformation is measured. Special
features of general-purpose test systems are capabilities for constant rate of loading, constant strain
rate, constant load, constant deformation, and cyclical loading (fatigue).
One of the simplest ways to control loading during mechanical testing is by crosshead position
or stroke-control feedback. In stroke-control feedback, the crosshead motion is controlled at a xed
speed or constant displacement (strain) rate. For high-stiffness test specimens, however, motion of
the crosshead is very small and it is very easy to break the specimen prematurely. Use of load-
control feedback from a load cell attached in the load chain provides more precise control. In the
load-control method, a load time history prole is programmed into the load-feedback loop and
crosshead position is adjusted to provide the programmed load prole. Stress rate (Pa/s), rather
than strain rate (

/s), is controlled in a load-feedback test setup, and the actual strain rate is dependent
on the modulus (Pa) of the test material:
strain rate = stress rate/modulus (11.4)
The choice of strain rate depends on the nature of the investigation, but for studies of normal
bone activity the strain rate should be selected to lie in the physiological range, i.e., 0.002 to 0.01/s
(0.2 to 1%

/s).

106

To simulate trauma and impact failure, strain rates of 0.1 to 1.0/s (10 to 100%

/s) can be employed,

41,62

but higher strain rates may be limited by the functional capacity of the
mechanical test apparatus. For most materials, an increased strain rate will result in an increase in
the stress necessary to produce a given strain. Ordinarily, variations in strain rate of 10 to 1 or lower
produce quite small changes in the measured stress. However, very large changes in the measured
stress may occur when testing at impact-level strain rates of 100% or higher. This effect is known
as hydraulic strengthening (HS) and is a uid-ow phenomenon associated with viscoelastic mate-
rials such as bone tissue.

69

Strain rate effects on the apparent elastic modulus

E

app

can be accounted
for by multiplying the true or specic modulus

E

s

by the apparent density/specic density ratio
expressed as a power function of strain rate :

69


(11.5)
For small ranges of strain rate, the strain rate power function exponent, , obtained
by Carter and Hayes

41,62

can be used to predict the apparent modulus:



(11.6)
Most mechanical testing machines are equipped with a load cell for detecting the applied load
and a transducer for measuring the displacement of the crosshead. The crosshead displacement
provides a measurement from which strain in the specimen can be calculated. Crosshead displace-
ment measurements, however, are often inaccurate because inhomogeneous strains may be present
within the specimen. In addition, crosshead displacement measurements include the deformation
associated with the crosshead xture, load cell xture, and specimen grip xtures. Consequently,
more accurate strain measurement methods are obtained by attaching resistance strain gauges or
extensometers directly to the specimen. Extensometers and strain gauges change electrical resis-
tance when deformed. A Wheatstone bridge type of amplier is generally used with these trans-
ducers to provide a precise voltage output for very small displacement and strain changes. Stress
and strain can be recorded by connecting the readout channel from the testing machine, load cell,
or strain measurement device to an

xy

plotter, strip chart recorder, or a data acquisition system
attached to a computer. Load and displacement output must be converted to stress and strain using
f ( )
E
app
E
s

app

s
--------


f ( )

f ( ) 0.06
E
app

0.06

0266/ ch11/frame Page 191 Thursday, October 21, 1999 4:25 PM

192

Mechanical Testing of Bone and the BoneImplant Interface

appropriate formulae. Once the data are stored on computer, data analysis software is used to
determine mechanical properties (Youngs modulus, strength, yield stress, toughness, etc.).
Machine test grips are used to hold the specimen. Ideally, the test grips should not only hold
the test specimen without slipping, but should also apply the load in the desired manner (pure
tension, compression, and shear). In compression, centering of the load is important, and should
not be neglected in tension testing if the material is brittle. Errors of up to 25% in strength prediction
can result from misalignment during compression testing.

79

Swiveling (ball-and-socket, or pivoting)
holders or compression blocks should be used for compression testing of all but very ductile
materials, and rough surfaces should be smoothed or capped. In tensile tests, serrated wedge-type
grips may be used to hold the shanks of ductile materials. A tapered specimen in a conical holder
provides good protection against slipping and protects the specimen from grip damage. A taper of
1 in 6 on the wedge faces gives a self-tightening action without excessive jamming (Figure 11.12).
As noted previously, environmental testing conditions play a signicant role during mechanical
testing. Differences of up to 10% in material properties can result from changes in temperature
and humidity alone (see previous section and References 123 and 124).

VI. MEASUREMENT PROCEDURES
A. C

OMPLIANCE



AND

V

ALIDATION

If a direct measurement of strain is too difcult (e.g., the specimens are too short for the application
of extensometers or unsuitable for strain gauges), then the strain must be determined from the

Figure 11.12

(a) Specimen holder with universal joints on both ends; (b) same as (a), however, with tapered
specimen in a conic holder. The conic holder is made out of two halves connected with screws or clamps
perpendicular to the loading direction.

0266/ ch11/frame Page 192 Thursday, October 21, 1999 4:25 PM

Tensile and Compression Testing of Bone

193

crosshead position measurements. When crosshead position measurements are used to determine
strain, the compliance (1/stiffness) of the load frame, including the load cell and loading platens,
must be considered. Standard specimens of steel, aluminum, and acrylic can be used to calibrate
the stiffness of the system. Load frame stiffness can also be determined directly by loading the
system without a specimen. The compliance-corrected (deection/load) stiffness response of the
bone test specimen,

k

bone

, is obtained from
1/

k

bone

= 1/

k

total

1/

k

machine

(11.7)
where

k

total

is the measured stiffness of the bone and machine, and 1/

k

machine

is the compliance of
the machine (load cell, loading platens, load frame). Note that Equation 11.7 assumes that the load
frame and bone specimen are a system of serially connected springs. With small testing machines,
however, the stiffness of the bone specimen may be greater than that of the load frame, whereas
for large testing machines the opposite is the usual case. In general, the compliance of the ideal
testing apparatus should be 10

5

mm/N or better.
Standard materials should be used whenever possible to ensure experimental measurements are
reliable and accurate. After a particular test procedure has been selected, the procedure should be
veried using several standardized specimens. Plastics (e.g., Plexiglas) have a similar modulus to
that of cortical bone and therefore make good materials standards for cortical bone. By using
materials standards, the measurements derived from a given experimental method can be easily
checked since the mechanical properties of a specic plastic are readily available from the manu-
facturer. Plastic standards are not only useful for machined bone specimens, but can also be used
to calibrate whole-bone tests. In the case of whole bone, specimens that approximate the bone
dimensions are useful to verify that there are no potential errors associated with the test apparatus
or data reduction procedures. For example, when validating a simple compression test experiment
on Plexiglas, Turner and Burr

125

found that the test procedure resulted in an underestimation of the
Youngs modulus of Plexiglas by 30% compared with the manufacturers data sheet. They found
that the addition of a pivoting platen to the load train corrected this error.

B. P

RECISION



AND

A

CCURACY

Precision (or reproducibility) and accuracy are two very different terms.

Precision

describes the
variation in the determination of a property associated with a specic method, whereas

accuracy

describes the ability of the method to estimate the true value of a measurement procedure. Precision
and accuracy are not necessarily related. For example, the precision of a method can be very good,
but the accuracy of the same method may be very poor, or vice versa. Increasing the number of
repeated measurements generally increases the precision, and is a simple addendum to nondestruc-
tive testing techniques.

126

Precision derived from the yield point and the failure point during mechanical testing to failure
cannot be assessed due to the singular nature of such test. Precision measurements derived from
nondestructive tests, however, can easily be assessed by repeated measurements. For example, the
precision of the stiffness and nondestructive energy absorption properties of bone have been
determined in a number of studies.

40,127-131

These studies found that the precision of a series of
repeated measurements was best when the specimens were not removed from the test machine
between measurements. In addition, the precision of stiffness measurements was improved (smaller
standard deviation) when the stiffness measurements were determined as the average stiffness of
ve consecutive test cycles as opposed to a single test. Furthermore, the stiffness precision was
improved when measurements were performed after a number of conditioning cycles in comparison
with tests conducted without conditioning.

73,132

Moreover, precision of stiffness and elastic energy
storage was also better for larger specimens than for smaller bone specimens, whereas energy
dissipation did not exhibit such a dependency. Viscoelastic energy dissipation measurements,

0266/ ch11/frame Page 193 Thursday, October 21, 1999 4:25 PM

194

Mechanical Testing of Bone and the BoneImplant Interface

however, are generally less precise than the stiffness and elastic energy storage.

133

A number of
investigators have also examined the effects of the testing order in orthogonal tests on the precision
of mechanical property measurements. Test order was not found to affect the precision.

40,127,129,131

The measurement precision of load cell and deformation measurement devices is usually very
good (0.3% of full scale). However, there are a number of factors that may reduce the precision
of load and displacement measurements, including
Misalignment of test columns;
Placement of the specimen away from the centroidal axis;
Nonparallel specimen ends;
Nonhomogeneity of the specimen.
These factors tend to push the test columns farther away from the central axis of the test
specimen. A compliant (sensitive) load cell and/or long testing column will amplify such a tendency.
Other factors that can reduce measurement precision include changes in specimen hydration and
temperature, the state of mechanical conditioning, or structural changes produced during a previous
testing session.

123

Factors affecting the accuracy of mechanical property measurements are usually related to the
test machine, the bone specimen, and/or the interface between the grip or platen and the specimen.
Although the stiffness of commercial test machines (>10

5

N/mm) is usually much larger than the
stiffness of the bone sample, some deformation occurs in the load cell. Consequently, when using
crosshead displacement measurements, it is easy to overestimate both the axial deformation of the
bone specimen as well as the strain rate of the test. If necessary, mechanical properties (strength,
stiffness) can be adjusted using power function relationships established between mechanical
properties and strain rate.

41,62,134

Errors in mechanical property measurements that are due to an
overestimation of strain rate, however, are generally very small in comparison to errors caused by
over-estimation of axial deformation. As noted previously, specimen preparation can have a sig-
nicant effect on the accuracy in determining the mechanical properties of bone. In particular,
embalming is known to signicantly and variably effect bone mechanical properties and should be
avoided if possible.
Another problem that occurs primarily in screw driven test machines is a variable tendency to
produce small loops at the bottom and top of the loaddeformation curve during cyclic loading.
This phenomenon would seem to indicate that the system is producing energy rather than dissipating
energy. However, these small loaddeformation loops are most likely caused by a slight tilting of
the crosshead produced by asynchronous activation of the screw drives responsible for moving the
crosshead. This generally occurs because the up/down (tension/compression) screw mechanism is
from a single belt attached to the motor of the test machine. This phenomenon is often a major
factor affecting the reproducibility of the test data, because repositioning of specimens that are not
absolutely homogeneous or that have slightly nonparallel end plates may create slight changes in
the mechanical test axis.
Retaining marrow

in situ

may be important in preventing water loss from the trabecular tissue
lattice during storage and testing.

28,62,77,135

If marrow must be extracted, the specimens should be
rehydrated prior to mechanical testing. Nevertheless, the properties of rehydrated trabecular bone
may differ from trabecular bone tested with marrow

in situ.

Carter and Hayes

62

analyzed the
dependence of mechanical properties on apparent density and strain rate using specimens allocated
to either testing with marrow or testing without marrow. They found a variable effect of marrow
on the strength and stiffness of the trabecular bone samples depending on the strain rate used to
test the specimens. Trabecular bone strength and stiffness were higher in specimens tested with
marrow in comparison with specimens tested without marrow at strain rates of 10 s

1

, but not at
strain rates of 1 s

1

and lower. Experimental differences associated with the high strain rate tests
may be due to the hydraulic strengthening effect.

68,69,136-138

0266/ ch11/frame Page 194 Thursday, October 21, 1999 4:25 PM

Tensile and Compression Testing of Bone

195

In uniaxial compression tests it is generally assumed that grip stresses are constant and that there
is no friction between the surface of the test grip and the specimen. This is only true if the specimen
is homogeneous and if the test column grips produce a constant stress on the specimen during loading.
Ordinarily, however, the grip stresses will vary as the specimen dimensions change. When this occurs,
the general effect is an uneven stress distribution at the specimenplaten interface, which in turn causes
a triaxial stress eld in the specimen.

139

The effect of a triaxial stress eld was studied by Filon,

140

who
analyzed the stress distribution in a cylindrical homogeneous specimen subjected to compression,
tension, and shear. This investigator found that the central regions of the specimen were subjected to
lower strains in comparison with regions near the specimen end. Such axial strain nonhomogeneity
leads to an overestimation of Youngs modulus. To investigate this phenomenon, Brown and Ferguson

141

performed a nite-element analysis of a cubic specimen, and found that Youngs modulus was over-
estimated by about 5%. Odgaard and Linde

142

performed a similar nite-element analysis on a homo-
geneous, isotropic cube where the interface was held rigid, and no deformation in the test column was
allowed. Using a Poissons ratio of 0.22, which was considered realistic for trabecular bone, these
authors found a 3% overestimation of Youngs modulus. A recent study by Keaveny et al.

46

indicates
that such modulus overestimations are highly dependent on the Poissons ratio of the specimen.
The effect of friction at the interface between specimen and test column was investigated
experimentally by Linde and Hvid.

130

These authors compared compressive mechanical properties
derived from nondestructive testing of trabecular bone specimens between ordinary steel columns,
steel columns with polished ends, and steel columns with polished ends lubricated with mineral
oil. The compressive stiffness was reduced by 5% when polished steel columns were used in
comparison with unpolished columns. When polished steel columns lubricated with mineral oil
were used, the compressive stiffness was reduced 7% compared with the stiffness obtained using
ordinary steel columns without oil as lubricant. Corresponding reductions were also found for
viscoelastic energy dissipation. Their differences are of the same order of magnitude as the results
obtained by nite-element analysis.

141,142

Thus, frictional effects on the mechanical behavior of
bone can be appreciable. Indeed, if the bone surface is fully constrained from sliding by gluing the
specimen with PMMA, then there is a 40% increase in stiffness compared with testing between
ordinary steel columns without oil.

130

Differences in stiffness obtained between a system with a
rigid interface and a system with minimal interface friction is even greater (nearly 50%)!
In another recent study, Allard and Ashman

143

made simultaneous compressive strain measure-
ment using both crosshead deection and extensometer measurements at the central region of cubic
bone specimens. They found that the compressive stiffness derived from the central region of the
bone specimen was considerably larger than that derived from the total deection. This nding is
well explained by the cut surface or structural end phenomenon of the specimen as described by
Zhu et al.

48

As a result of the structural end phenomenon, axial compressive strain measurements
based upon crosshead deection measurements will always be overestimated to a certain extent,
and the magnitude of this overestimation is inversely related to the specimen length.

40,48

Specimens
with embedded ends for tensile testing, however, are not expected to be affected by the structural
end phenomenon and will accordingly be expected to have a larger stiffness and smaller ultimate
strain than specimens that are not so constrained at the grip interface.

144-146

One of the most important factors inuencing mechanical test measurement accuracy is related
to errors in the strain measurement. The net underestimation of stiffness resulting from interface
phenomena (including overestimation caused by friction and structural end effects) has been
estimated to be 20 to 40% in specimens with a length of 7 to 7.5 mm,

131,142

and about 40% in
15-mm-long specimens with a 3:1 aspect ratio.

46

Stiffness overestimation is expected to be larger
in shorter specimens and smaller in longer specimens.

40,48

In addition, since the axial stress distri-
bution is heterogeneous in bone specimens with or without embedded ends, tests based upon local
strain measurements (extensometers, strain gauges, and optical targets) may be the only reliable
way to obtain accurate results. For local strain measurements, it is advisable to obtain specimen
strain measurements from several sides of the specimen in order to take into account any bending

0266/ ch11/frame Page 195 Thursday, October 21, 1999 4:25 PM

196

Mechanical Testing of Bone and the BoneImplant Interface

that is imposed on the specimen.

35

Ignoring structural end phenomena, it should also be noted that
the stiffness determined is not necessarily the true or intrinsic material stiffness, since testing is
generally not performed at an innitely low strain rate.
The accuracy of strength measurements is generally better than that of stiffness measurements
and measurements of energy absorptive properties, since no strain measurement is involved. In the
case of ultimate strain determinations, the accuracy is only affected by the general inaccuracy of
the strain measurements. Ultimate strain is generally slightly overestimated, and the most accurate
values are obtained from testing embedded specimens as is the case for stiffness and energy
absorptive properties.

C. D

ATA

A

CQUISITION



AND

A

NALYSIS

Whether tension or compression loads are used, it is advisable to apply the load to the specimen
several times before recording the load and deection, since some settling occurs between the
specimen and grips during the rst cycles. When presenting results, bone density, mineral content,
age, sex, and health of the bone source should be included since such information serves as a
benchmark for comparison with other studies. The following sections describe key stressstrain
measurements and methods for compensating for artifacts in stressstrain measurements.

1. Determination of Mechanical and Structural Properties

Yield point

In a mechanical sense, the yield point is the point where structural changes begin. For bone
specimens this point is presumed to be near the point where the slope of the stressstrain curve is
seen to decrease. Such a denition, however, is not very precise, and methods for determination
of the yield point vary accordingly. In analyzing a stressstrain curve wherein there is no distinct
yield point the ASTM Committee on Mechanical Testing suggests that the 0.2% offset criteria for
engineering material be used to determine the yield stress. The yield point by the offset criterion
is dened as the intersection between the compression curve and a line parallel with the maximum
slope and displaced 0.2% strain (ASTM E8). Although this denition is precise, it has some
disadvantages. Signicant structural changes have probably occurred before that point, and due to
the slightly different shape of stressstrain curves of weak specimens compared with stronger
specimens, a yield strain close to and even larger than the ultimate strain is sometimes found in
weak specimens. A smaller strain offset will eliminate this problem, and for biomechanical testing
a strain offset of 0.1% is preferred and may become standardized for future testing.

63

A more physiological denition of the yield point would be the point where the slope has
reached maximum in its mathematical sense. This point is easily determined whenever test data
are stored directly on computers, and can be derived from the second-order differential equation
of a third-order (or higher) polynomial t to the stressstrain data. This point has been found to
be about 0.8% strain in unconned compression testing of trabecular bone specimens.

128

Ultimate strength

The maximum stress the bone can sustain is called the ultimate strength, and the breaking strength
is the stress at which the bone actually breaks. In bone the ultimate strength and the breaking
strength usually have the same value, but this is not necessarily true for all materials. For example,
a specimen of ductile steel will stretch considerably before it breaks, and due to this stretching,
the stress sustained at fracture (breaking strength) may actually be less than the maximum stress
attained (ultimate strength). It should be noted that strength, as it is dened above (e.g., stress), is
an intrinsic property of bone. That is, strength values are independent of the size and shape of the
bone. The force required to break the bone differs from the intrinsic strength, because the breaking
load or fracture load will vary with bone or specimen size. One must keep this distinction in mind
because intrinsic strength and breaking load can show very different trends between control and

0266/ ch11/frame Page 196 Thursday, October 21, 1999 4:25 PM

Tensile and Compression Testing of Bone

197

experimental groups in studies wherein the treatment affects the size of the bone. For example,
uoride treatment decreases the intrinsic strength of bone in young rats, but it also increases the
bone size such that breaking load remains unchanged.

Tensile strength

The tensile strength is calculated by dividing the maximum load recorded (from the tensile test to
failure) in newtons by the original minimum cross-sectional area of the specimen in square meters.
This result is generally expressed in pascals (N/m

2

) and should be reported to three signicant
gures as tensile strength at yield or as the tensile strength at break, whichever term is applicable.
When a nominal yield or breaking load is less than the maximum, it may be desirable to calculate
the corresponding tensile stress at yield or tensile stress at break and report these using at least
three signicant gures.

Percent elongation

If the test results in a yield load that is larger than the load at break, one should calculate the
percent elongation at yield. Otherwise, the percent elongation at break should be calculated. Percent
elongation is determined by reading the deformation (change in gauge length) at the moment the
applicable load is reached, and dividing that value by the original gauge length (multiply by 100).
Report percent elongation at yield or percent elongation at break to two signicant gures. When
a yield or breaking load is less than the maximum, both percent elongation at yield and percent
elongation at break should be reported.

Modulus of elasticity

Calculate the modulus of elasticity by extending the initial linear portion of the loaddeformation
curve and dividing the difference in stress corresponding to any segment of the section on this
straight line by the corresponding difference in strain. Compute all elastic modulus values using
the average initial cross-sectional area of the test specimens in the calculations. Express the result
in pascals and report to three signicant gures.
The stressstrain relations of many biological tissues do not conform to Hookes law throughout
the elastic range but may deviate from this idealized linear stressstrain behavior at stresses well
below the elastic limit. For such materials the slope of the tangent to the stressstrain curve at a
low stress is usually taken as the modulus of elasticity. Since the existence of a true proportional
limit in biological tissue is debatable, the appropriateness of applying the term

modulus of elasticity

to describe the stiffness of rigidity of a biological specimen has been seriously questioned. The
exact stressstrain characteristics of biological materials are dependent on such factors as strain
rate, temperature, previous specimen stress history, etc. Thus, when applied to biological tissue,
the precise meaning of mechanical property is most useful if this dependency is understood.

Energy absorption

The area under the compression or tension curve represents the work put into the material by
compression/tension (loading energy or strain energy). Unloading the specimen prior to reaching
the elastic limit of an ideal elastic material results in the same amount of energy being released
(unloading energy). In other words, the unloading curve coincides with the loading curve. However,
since bone is a viscoelastic material, it dissipates energy. The loading energy within the elastic
range is composed of elastic energy that is released during unloading and viscoelastic energy that
is absorbed by the material and converted to other energy forms such as frictional heat. The
unloading curve of a viscoelastic material will follow a lower course than the loading curve, and
the two types of energy can be determined from such a loadingunloading cycle (the so-called
hysteresis loop). The area underneath the unloading curve represents the energy released during
unloading (elastic energy), and the area enclosed by the hysteresis loop represents energy dissipation
during the loading cycle (viscoelastic energy). For each series of tests, calculate the arithmetic
mean of all values obtained and report it as the average values for the particular property in question.

0266/ ch11/frame Page 197 Thursday, October 21, 1999 4:25 PM

198

Mechanical Testing of Bone and the BoneImplant Interface

2. Toe Region Compensation

In a typical stressstrain curve (Figure 11.13) there is a toe region, AC, that does not represent a
property of the material.

147

The toe region is an artifact caused by a take-up of slack, alignment
and/or seating of the test specimen. In order to obtain correct values of mechanical parameters such
as modulus, yield, and ultimate strain, the corrected zero point on the strain or extension axis must
be determined. In case of a material exhibiting a linear or Hookean stressstrain behavior, a
continuation of the linear (CD) region of the curve is constructed through the zero-stress axis. This
intersection (B) is the corrected zero-strain point from which all displacements or strains must be
measured, including the yield offset point (BE), if applicable. The elastic modulus can be determined
by dividing the stress at any point along the line CD (or its extension) by the strain at the same
point (measured from point B, dened as zero strain).
In the case of a material that does not exhibit any linear region (refer to Figure 11.13), the
same kind of toe correction for the zero-strain point can be made by constructing a tangent to the
maximum slope at the inection point (H

). This is extended to intersect the strain axis at point B

,
which is the corrected zero-strain point. By using point B

as the zero strain reference, the stress
at any point (G

) on the curve can be divided by the strain at that point to obtain a secant modulus
(slope of line B

). For those materials with no linear region, any attempt to use the tangent
through the inection point as a basis for determination of an offset yield point may result in
unacceptable error.

VII. ANIMAL MODELS

The animal model chosen determines, to a large extent, the specic biomechanical tests that can
be performed. Any of the aforementioned tests can be performed for bones from larger animals,
but the test choice is more restrictive for smaller animals. With rats, for instance, tests are usually
limited to bending or torsion of long bones and compression of vertebral bodies. It is difcult, but
not impossible, to accurately measure the mechanical properties of trabecular bone in rodents.
Trabecular bone cores can be prepared from rat vertebral bodies for purposes of compression
testing, but more commonly compression tests of intact vertebral bodies are conducted since core
preparations can be exceedingly difcult. When testing intact structures, such as an entire vertebral
body, the contributions of both the cortical shell and trabecular bone core will have a signicant
inuence on the mechanical test results. Another important consideration is the type of bone being

FIGURE 11.13

Typical stressstrain curve obtained from tensile testing: (a) with linear region; (b) without
linear region.

0266/ ch11/frame Page 198 Thursday, October 21, 1999 4:25 PM

Tensile and Compression Testing of Bone

199

studied. Many larger species (e.g., sheep and cows) have predominantly plexiform cortical bone,
which differs mechanically from osteonal cortical bone, especially in fatigue.

148

Canine trabecular
bone also differs appreciably from human trabecular bone.

149

Because many mechanical tests are prone to experimental bias, a good experimental design
should always include a control group. Fortunately, most biomechanical tests (e.g., compression,
tension, bending, and torsion) are very precise and reasonably accurate, so that comparison between
treatment groups and controls can be done very effectively.

VIII. SUMMARY

Knowledge of the mechanical properties of bone is of signicant importance for understanding
diseases such as osteoporosis and osteoarthritis. Bone loss associated with osteoporosis reduces
the mechanical strength of trabecular bone, increases the risk of fracture, and produces painful and
spontaneous collapse of bones, especially vertebral bodies. During the slow and progressive devel-
opment of osteoarthritis, subchondral trabecular bone undergoes characteristic changes, and it has
been hypothesized that the initial changes in osteoarthritis are the result of changes in the shock-
absorptive properties of trabecular bone.
Knowledge of mechanical properties of bone is also of major importance for the design of
fracture xation devices, as well as the design, xation, and survival of articial joints. Articial
joints, such as the acetabular component in hip arthroplasty and the femoral and tibial components
in knee arthroplasty are intimately associated with cortical and trabecular bone. An understanding
of the mechanical properties of the bone is thus very important for the design of joint prostheses.
A precise and accurate knowledge of bone mechanical properties is also a prerequisite for numerical
design and optimization (nite-element analysis) of implants and for simulation of load-induced
stress-morphology processes (remodeling) that take place in bone. The validity of such numerical
analyses draws upon the accuracy of the mechanical property data used to describe bone, implant,
and boneimplant interface.
Because of the wide variety of bone shapes and sizes, and the fact that there are no established
standards for bone biomechanical testing, there are a large number of variables to consider when
establishing testing procedures. Unfortunately, there are no well-established industry standards
for biomechanical testing of bone. Consequently, a lot of published biomechanical test data are
inaccurate because of poor testing techniques or inattention to confounding variables. This chapter
was not intended to provide an exhaustive review of experimental techniques used by researchers;
rather, the chapter focused on several common test protocols, which are currently used in bio-
mechanical testing of bone tissue.
The following summarizes several important factors and guidelines that should be considered
when performing mechanical tests of bone:
1. Specimen geometry has a highly signicant inuence on mechanical properties such as
stiffness, ultimate strain, and energy absorption. A cube with a side length of 8 mm and
a cylindrical specimen with a length of 16 mm and a diameter of 8 mm are suggested
as standard geometries providing comparable results.
2. Standard testing of small trabecular bone specimens is associated with systematic
errors. The most signicant of these errors are believed to be related to the integrity
of trabeculae at the cut or machined surfaces of the test specimen and friction at the
specimenplaten interface.
3. The best method of long-term preservation prior to testing is to freeze the specimens at
20

C in saline-soaked gauze. Small specimens can be preserved up to 90 days if kept
in a solution of 50% ethanol and 50% saline. During testing, care should be taken to
ensure that the test specimens are kept hydrated.

0266/ ch11/frame Page 199 Thursday, October 21, 1999 4:25 PM

200

Mechanical Testing of Bone and the BoneImplant Interface

4. Reproducibility can be improved by preconditioning the test specimens using a number
of conditioning cycles to achieve a viscoelastic steady state.
5. The stiffness derived from nondestructive tests will generally be lower than that obtained
from a destructive test because of inherent nonlinearity in the loaddeformation curve.
6. For each experiment, a series of tests should be conducted so that calculations of the
arithmetic mean of all values can be obtained.

REFERENCES

1. American Society for Metals International, Mechanical testing, in

Metals Handbook

, 8, ASM, Mate-
rials Park, OH, 1989, 589.
2. Evans, F.G. and Lebow, M., The strength of human bone as revealed by engineering techniques,

Am.
J. Surg

., 83, 326, 1952.
3. Gerard, G.,

Introduction to Structural Stability Theory

, McGraw-Hill, New York, 1962, 19.
4. Roak, R.J. and Young, W.C.,

Formulas for Stress and Strain

, McGraw-Hill, New York, 1975.
5. Banse, X., Delloye, C., Cornu, O., and Bourgois, R., Comparative left-right mechanical testing of
cancellous bone from normal femoral heads,

J. Biomech.

, 29, 1249, 1996.
6. Bartley, M.H., Arnold, J.S., Haslam, R.K., and Jee, W.S., The relationship of bone strength and bone
quantity in health, disease and aging,

J. Gerontol., 21, 517, 1966.
7. Evans, F.G. and Band, S., Differences and relationships between the physical properties and the
relationships between the physical properties and the microscopic structure of human femoral, tibial
and bular cortical bone, Am. J. Anat., 120, 79, 1967.
8. Gibson, L.J., The mechanical behavior of cancellous bone, J. Biomech., 18, 317, 1985.
9. Goldstein, S.A., The mechanical properties of trabecular bone: dependence on anatomic location and
function, J. Biomech., 20, 1055, 1987.
10. Martin, R.B. and Atkinson, P.J., Age and sex-related changes in the structure and strength of the
human femoral shaft, J. Biomech., 10, 223, 1977.
11. Welch, D.O., The composite structure of bone and its response to mechanical stress, Recent Adv. Eng.
Sci., 5, 245, 1970.
12. ASTM, Standard: Standard test methods of compression testing of metallic materials at room tem-
perature, Annual Book of ASTM Standards, ASTM, Philadelphia, E9-89a, 1983.
13. ASTM, Standard: Standard test method for compressive properties of rigid cellular plastics, D1621,
ASTM, Philadelphia, 1991, 11.
14. ASTM, Standard: Standard test method for compressive properties of rigid plastics, D695, ASTM,
Philadelphia, 1991, 276.
15. ASTM, Standard: Standard testing for compressive strength of cylindrical concrete specimens, C39,
ASTM, Philadelphia, 1991, 86.
16. Linde, F., Elastic and viscoelastic properties of trabecular bone by a compression testing approach,
Dan. Med. Bull., 41, 119, 1994.
17. Keller, T.S., Hansson, T.H., Abram, A.C., et al., Regional variations in the compressive properties of
lumbar vertebral trabeculae. Effects of disc degeneration, Spine, 14, 1012, 1989.
18. Keaveny, T.M., Borchers, R.D., Gibson, L.J., and Hayes, W.C., Trabecular bone modulus and strength
can depend on specimen geometry, J. Biomech., 26, 991, 1993.
19. Carter, D.R. and Spengler, D.M., Mechanical properties and composition of cortical bone, Clin.
Orthop., 135, 192, 1978.
20. Cowin, S.C., Mechanics of materials, in Bone Mechanics, Cowin, S.C., Ed., CRC Press, Boca Raton,
FL, 1989, 15.
21. Currey, J.D., The mechanical properties of bone, Clin. Orthop., 73, 210, 1970.
22. Hollister, S.J., Fyhrie, D.P., Jepsen, K.J., and Goldstein, S.A., Application of homogenization theory
to the study of trabecular bone mechanics, J. Biomech., 24, 825, 1991.
23. Mosekilde, Li., Viidik, A., and Mosekilde, Le., Correlation between the compressive strength of iliac
and vertebral trabecular bone in normal individuals, Bone, 6, 291, 1985.
24. Mosekilde, Li., Mosekilde, Le., and Danielsen, C.C., Biomechanical competence of vertebral trabe-
cular bone in relation to ash density and age in normal individuals, Bone, 8, 79, 1987.
0266/ ch11/frame Page 200 Thursday, October 21, 1999 4:25 PM
Tensile and Compression Testing of Bone 201
25. Behrens, J.C., Walker, P.S., and Shoje, H., Variations in strength and structure of cancellous bone at
the knee, J. Biomech., 7, 201, 1974.
26. Cowin, S.C., The mechanical properties of cancellous bone, in Bone Mechanics, Cowin, S.C., Ed.,
CRC Press, Boca Raton, FL, 1989, 129.
27. Goldstein, S.A., Wilson, D.L., Sonstegard, D.A., and Matthews, L.S., The mechanical properties of
human tibial trabecular bone as a function of metaphyseal location, J. Biomech., 16, 965, 1983.
28. Lindahl, O., Mechanical properties of dried defatted spongy bone, Acta Orthop. Scand., 47, 11, 1976.
29. Panjabi, M.M., Krag, M.K., Summers, M.K., and Videman, T., Biomechanical time-tolerance of fresh
cadaveric human tissue specimens, J. Orthop. Res., 3, 292, 1985.
30. Pelker, R.R., Friedlaender, G.E., Markham, T.C., et al., Effects of freezing and freeze-drying on the
biomechanical properties of rat bone, J. Orthop. Res., 1, 405, 1984.
31. Sonstegard, D.A. and Matthews, L.S., Mechanical property dependence on storage technique and
local of knee joint trabecular bone, Trans. Orthop. Res. Soc., 2, 283, 1977.
32. Komender, A., Inuences of preservation on some mechanical properties of human Haversian bone,
Mater. Med. Pol., 8, 13, 1976.
33. Linde, F. and Srensen, H.C., The effect of different storage methods on mechanical properties of
trabecular bone, J. Biomech., 26, 1249, 1993.
34. Sedlin, E.D. and Hirsch, C., Factors affecting the determination of the physical properties of femoral
cortical bone, Acta Orthop. Scand., 37, 29, 1966.
35. Ashman, R.B., Experimental techniques, in Bone Mechanics, Cowin, S.C., CRC Press, Boca Raton,
FL, 1989, 75.
36. McElhaney, J.H., Fogle, J., Byars, E., and Weaver, G., Effect of embalming on the mechanical
properties of beef bone, J. Appl. Physiol., 19, 12334, 1964.
37. Evans, F.G., Mechanical Properties of Bone, Charles C Thomas, Springeld, IL, 1973.
38. Hvid, I. and Hansen, S.L., Trabecular bone strength patterns at the proximal tibial epiphysis, J. Orthop.
Res., 3, 464, 1985.
39. Keller, T.S., Predicting the mechanical behavior of bone, J. Biomech., 27, 1159, 1994.
40. Linde, F., Hvid, I., and Madsen, F., The effect of specimen size and geometry on the mechanical
behavior of trabecular bone, J. Biomech., 25, 359, 1992.
41. Carter, D.R. and Hayes, W.C., Compact bone fatigue damage, Clin. Orthop., 127, 265, 1977.
42. Odgaard, A., Hvid, I., and Linde, F., Compressive axial strain distribution in cancellous bone specimen,
J. Biomech., 22, 829, 1989.
43. Rohlmann, A., Zilch, H., Bergmann, G., and Koebel, R., Material properties of femoral cancellous
bone in axial loading. Part I: Time independent properties, Arch. Orthop. Trauma Surg., 97, 95, 1980.
44. Hansson, T.H., Keller, T.S., and Panjabi, M.M., A study of the compressive properties of lumbar
vertebral trabeculae: effects of tissue characteristics, Spine, 12, 56, 1987.
45. Hvid, I., Mechanical strength of trabecular bone at the knee, Dan. Med. Bull., 35, 345, 1988.
46. Keaveny, T.M., Borchers, R.E., Gibson, L.J., and Hayes, W.C., Theoretical analysis of the experimental
artifact in trabecular bone compressive modulus, J. Biomech., 26, 599, 1993.
47. Tanner, K.E., Sharp, D.J., Turner, S., and Boneld, W., Specimen length and the measured mechanical
properties of trabecular bone, Trans. World Congr. Biomech., 1, 131, 1990.
48. Zhu, M., Keller, T.S., and Spengler, D.M., Effects of specimen load-bearing and free surface layers
on the compressive mechanical properties of cellular materials, J. Biomech., 27, 57, 1994.
49. Choi, K., Kuhn, J.L., Ciarelli, M.J., and Goldstein, S.A., The elastic moduli of subchondral, trabecular,
and cortical bone tissue and the size-dependency of cortical bone modulus, J. Biomech., 23, 1103, 1990.
50. Cowin, S.C. and Mehrabidi, M.M., Identication of the elastic symmetry of bone and other materials,
J. Biomech., 22, 503, 1989.
51. Harrigan, T.P., Jasty, M., Mann, R.W., and Harris, W.H., Limitations of the continuum assumptions
in cancellous bone, J. Biomech., 21, 269, 1988.
52. Bachus, K.N., Bloebaum, R.D., and Hofmann, A.A., Minimum trabecular width: the biomechanical
limit of load bearing cancellous bone, Trans. Orthop. Res. Soc., 15, 54, 1990.
53. Sumner, D.R., Turner, T.M., and Galante, J.O., Symmetry of the canine femur: implications for
experimental sample size requirements, J. Orthop. Res., 6, 758, 1988.
54. Tanner, K.E., Harris, J.R., Evans, G.P., and Boneld, W., Assessment of parameters which affect the
measured mechanical properties of human cancellous bone, Proc. Meet. Eur. Soc. Biomech., 8, 66, 1992.
0266/ ch11/frame Page 201 Thursday, October 21, 1999 4:25 PM
202 Mechanical Testing of Bone and the BoneImplant Interface
55. Currey, J.D., The effects of drying and re-wetting on some mechanical properties of cortical bone, J.
Biomech., 21, 439, 1988.
56. Currey, J.D., Physical characteristics affecting the tensile failure properties of cortical bone, J. Bio-
mech., 23, 837, 1990.
57. Jameson, M.W., Hood, J.A., and Tidmarsh, B.G., The effects of dehydration and rehydration on some
mechanical properties of human dentine, J. Biomech., 26, 1055, 1993.
58. Williams, J.L. and Johnson, W.J., Elastic constants of composites formed from PMMA bone cement
and anisotropic bovine tibial cancellous bone, J. Biomech., 22, 673, 1989.
59. Padawer, G.E. and Beecher, N., On the strength and stiffness of planar reinforced plastic resins, Polym.
Eng. Sci., 10, 185, 1994.
60. Gray, R.J. and Korbacher, G.K., Compressive fatigue behavior of bovine compact bone, J. Biomech.,
7, 287, 292.
61. Reilly, S.B. and Burstein, A.H., The elastic and ultimate properties of compact bone tissue, J. Biomech.,
8, 393, 1975.
62. Carter, D.R. and Hayes, W.C., The compressive behavior of bone as a two phase porous structure, J.
Bone Joint Surg., 59A, 954, 1977.
63. Turner, C.H., Yield behavior of bovine cancellous bone, J. Biomech. Eng., 111, 256, 1989.
64. Carter, D.R., Caler, W.E., Spengler, D.M., and Frankel, V.H., Fatigue behavior of adult cortical bone:
the inuence of mean strain and strain range, Acta Orthop. Scand., 52, 481, 1981.
65. Crowninshield, R.D. and Pope, M.H., The response of compact bone in tension at various strain rate,
Ann. Biomed. Eng., 2, 217, 1974.
66. Currey, J.D., The effects of strain rate, reconstruction, and mineral content on some mechanical
properties of bovine bone, J. Biomech., 8, 81, 1975.
67. Ducheyne, P., Heymans, L., Martens, M., et al., The mechanical behavior of intracondylar cancellous
bone of the femur at different loading rates, J. Biomech., 10, 747, 1977.
68. Kafka, V., On hydraulic strengthening of bones, Biorheology, 20, 789, 1983.
69. Liebschner, M.A.K. and Keller, T.S., Cortical Bone Exhibits Hydraulic Strengthening, 44th Annual
Meeting of the Orthopaedic Research Society, New Orleans, LA, 23, 559, 1998.
70. Panjabi, M.M., White, A.A., III, and Southwick, W.O., Mechanical properties of bone as a function
of rate of deformation, J. Bone Joint Surg., 55A, 322, 1973.
71. Ciarelli, M.J., Goldstein, S.A., Kuhn, J.L., et al., Evaluation of the orthogonal mechanical properties
and density of human trabecular bone from the major metaphyseal regions with materials testing and
computed tomography, J. Orthop. Res., 9, 674, 1991.
72. Dempster, W.T. and Liddicoat, R., Compact bone as a non-isotropic material, Am. J. Anat., 91, 331,
1952.
73. Fondrik, M., Bahniuk, E., Davy, D.T., and Michaels, C., Some viscoelastic characteristics of bovine
and human cortical bone, J. Biomech., 21, 623, 1988.
74. Rice, J.C., Cowin, S.C., and Bowman, J.A., On the dependence of the elastic and strength of cancellous
bone on apparent density, J. Biomech., 21, 155, 1988.
75. Rhl, L., Larsen, E., Linde, F., et al., Tensile and compressive properties of cancellous bone, J.
Biomech., 24, 1143, 1991.
76. Stone, J.L., Beaupre, G.S., and Hayes, W.C., Multiaxial strength characteristics of trabecular bone,
J. Biomech., 16, 743, 1983.
77. Ashman, R.B. and Rho, J.Y., Elastic moduli of trabecular bone material, J. Biomech., 21, 177, 1988.
78. Bird, F., Becker, H., Healer, J., and Messer, M., Experimental determination of the mechanical
properties of bone, Aerospace Med., 39, 44, 1968.
79. Reilly, S.B., Burstein, A.H., and Frankel, V.H., The elastic modulus of bone, J. Biomech., 7, 271, 1974.
80. Katz, J.L., Anisotropy of Youngs modulus of bone, Nature (London), 283, 106, 1980.
81. Klever, F.J., Klumpert, R., Horenberg, J., et al., Global mechanical properties of trabecular bone:
experimental determination and prediction from a structural model, in Biomechanics: Current Inter-
disciplinary Research, Perren, S.M. and Schneider, E., Eds., Martinus Nijhoff, Dordrecht, the Neth-
erlands, 1985, 167.
82. Tsai, S.W. and Wu, E.M., A general theory of strength for anisotropic material, J. Compos. Mater.,
5, 58, 1971.
0266/ ch11/frame Page 202 Thursday, October 21, 1999 4:25 PM
Tensile and Compression Testing of Bone 203
83. Burstein, A.H., Currey, J.D., Frankel, V.H., and Reilly, D.T., The ultimate properties of bone tissue:
the effects of yielding, J. Biomech., 5, 35, 1973.
84. Jonas, J., Burns, J., Able E.W., et al., A technique for the tensile testing of demineralized bone,
J. Biomech., 26, 271, 1993.
85. Katsamanis, F. and Raftopoulos, D.D., Determination of mechanical properties of human femoral
cortical bone by the Hopkinson bar stress technique, J. Biomech., 23, 1173, 1990.
86. Knauss, P., Material properties and strength behavior of spongy bone tissue at the proximal human
femur (Part I), Biomed. Tech., 26, 200, 1981 [translated from German].
87. Lewis, J.L. and Goldsmith, W., The dynamic fracture and prefracture response of compact bone by
split Hopkinson bar methods, J. Biomech., 8, 27, 1975.
88. Mente, P.L. and Lewis, J.L., Experimental method for measurement of the elastic modulus of trabecular
bone tissue, J. Orthop. Res., 7, 456, 1989.
89. Tarnopolskii, I.M. and Kincis, T., Static Test Method for Composites, Van Nostrand Reinhold, New
York, 1985.
90. Yahia, L.H., Drouin, G., and Duval, P., A methodology for mechanical measurements of technical
constants of trabecular bone, Eng. Med., 17, 169, 1988.
91. Carter, D.R., Schwab, G.H., and Spengler, D.M., Tensile fracture of cancellous bone, Acta Orthop.
Scand., 51, 733, 1980.
92. Turner, C.H. and Cowin, S.C., Errors induced by off-axis measurement of the elastic properties of
bone, J. Biomech. Eng., 110, 213, 1988.
93. Hayes, W.C. and Black, D., Post-yield energy absorption characteristics of trabecular bone. In 1979
Biomechanics Symposium, Van Buskirk, W.C., Ed., ASME, New York, 1979, 177.
94. Hodgkinson, R. and Currey, J.D., The effect of variation in structure on the Youngs modulus of
cancellous bone: a comparison of human and non-human material, Proc. Inst. Mech. Eng. H: J. Eng.
Med., 204, 115, 1990.
95. Inoue, N., Sakakida, K., Yamashita, F., et al., The elasto-plastic behavior of bovine cancellous bone
under tensile loading, Proc. Meet. Eur. Soc. Biomech., 6, 8A, 1988.
96. Keaveny, T.M., Guo, X.E., Wachtel, E.F., et al., Trabecular bone exhibits fully linear elastic behavior
and yields at low strain, J. Biomech., 27, 1127, 1994.
97. Kopperdahl, D.L. and Keaveny, T.M., Damage behavior of vertebral sections is dependent on applied
strains and dominated by trabecular bone, in Trans. Orthop. Res. Soc., 34, 1033, 1999.
98. Caler, W.E. and Carter, D.R., Bone creep-fatigue damage accumulation, J. Biomech., 22, 625, 1989.
99. Carter, D.R. and Hayes, W.C., Fatigue life of compact bone I. Effects of stress amplitude, temper-
ature and density, J. Biomech, 9, 27, 1976.
100. Chamay, A., Mechanical and morphological aspects of experimental overload and fatigue in bone,
J. Biomech., 3, 263, 1970.
101. Keller, T.S., Lovin, J.D., Spengler, D.M., and, Carter, D.R., Fatigue of immature baboon cortical bone,
J. Biomech., 18, 297, 1985.
102. Michel, M.C., Xiang-Dong, E.G., Gibson, L.J., et al., Compressive fatigue behavior of bovine trabe-
cular bone, J. Biomech., 26, 453, 1993.
103. Wachtel, E.F. and Keaveny, T.M., The effects of unloading conditions and strain-rate on the residual
properties of damaged trabecular bone, ASME Adv. Bioeng., BED, 31, 159, 1995.
104. Wagner, H.D. and Weiner, S., On the relationship between the microstructure of bone and its mechan-
ical stiffness, J. Biomech., 25, 1311, 1992.
105. Carter, D.R. and Caler, W.E., Cycle dependent and time dependent bone fracture with repeated loading,
J. Biomech. Eng., 105, 166, 1983.
106. Lafferty, J.F. and Raju, P.V., The inuence of stress frequency on the fatigue strength of cortical bone,
J. Biomech Eng., 101, 112, 1979.
107. Choi, K. and Goldstein, S.A., A comparison of the fatigue behavior of human trabecular and cortical
bone tissue, J. Biomech., 25, 1371, 1992.
108. Schafer, M.B., Radin, E.L., and Burr, D.B., Long-term fatigue behavior of compact bone at low
strain magnitude and rate, Bone, 11, 321, 1990.
109. McBroom, R.J., Hayes, W.C., Edwards, W.T., et al., Prediction of the vertebral body compressive
fracture using quantitative computed tomography, J. Bone Joint Surg., 67A, 1206, 1985.
0266/ ch11/frame Page 203 Thursday, October 21, 1999 4:25 PM
204 Mechanical Testing of Bone and the BoneImplant Interface
110. Keller, T.S., Spengler, D.M., and Carter, D.R., The inuence of exercise on maturing rat long bone
structural and material properties, Trans. Orthop. Res. Soc., 29, 60, 1983.
111. Keller, T.S., Spengler, D.M., and Carter, D.R., Geometric, elastic, and structural properties of maturing
rat femora, J. Orthop. Res., 4, 57, 1986.
112. Bell, G.H., Dunbar, O., Beck, J.S., and Gibb, A., Variations in strength of vertebrae with age and their
relations to osteoporosis, Calcif. Tissue Res., 1, 75, 1967.
113. Hansson, T.H., Keller, T.S., and Spengler, D.M., Mechanical behavior of the human lumbar spine.
II. Fatigue strength during dynamic compressive loading, J. Orthop. Res., 5, 479, 1987.
114. Hansson, T., Roos, B., and Nachemson, A., The bone mineral content and ultimate compressive
strength of lumbar vertebrae, Spine, 5, 46, 1980.
115. Hardy, W.G., Lissner, H.R., Webster, J.E., and Gurdijian, E.S., Repeated loading tests of the lumbar,
Spine Surg. Forum, 9, 690, 1958.
116. Keller, T.S., Spengler, D.M., and Hansson, T.H., Mechanical behavior of the human lumbar spine.
I. Creep analysis during static compressive loading, J. Orthop. Res., 5, 467, 1987.
117. McCubbrey, D.A., Cody, D.D., Kuhn, J.L., et al., Static and fatigue failure properties of thoracic and
lumbar vertebral bodies and their regional density, Trans. Orthop. Res. Soc., 15, 178, 1990.
118. Derwin, K.A., Soslowsky, L.J., Green, W.D., and Elder, S.H., A new optical system for the determi-
nation of deformations and strains: calibration characteristics and experimental results, J. Biomech.,
27, 1277, 1994.
119. Woo, S.L., Gomez, M.A., Seguchi, Y., et al., Measurement of mechanical properties of ligament
substance from a bone-ligament-bone preparation, J. Orthop. Res., 1, 22, 1983.
120. Woo, S.L., Danto, M.I., and Ohland, K.J., The use of a laser micrometer system to determine the
cross-sectional shape and area of ligaments: a comparative study with two existing methods,
J. Biomech. Eng., 112, 426, 1990.
121. Wright, T.M. and Hayes, W.C., Strain gauge application on compact bone, J. Biomech., 12, 471, 1979.
122. Wright, T.M., Vosburgh, F., and Burstein, A.H., Permanent deformation of compact bone monitored
by acoustic emission, J. Biomech., 14, 405, 1981.
123. Boneld, W., and Li, C.H., The temperature dependence of the deformation of bone, J. Biomech., 1,
323, 1968.
124. Brear, K., Currey, J.D., Rainers, S., and Smith, K.J., Density and temperature effects on some
mechanical properties of cancellous bone, Eng. Med., 17, 163, 1988.
125. Turner, C.H. and Burr, D.B., Basic biomechanical measurements of bone: a tutorial, Bone, 14, 595,
1993.
126. Wall, J.C., Chatterji, S., and Jeffery, J.W., On the origin of scatter in results of human bone strength
tests, Med. Biol. Eng., 8, 171, 1970.
127. Linde, F., Hvid, I., and Jensen, N.C., Material properties of cancellous bone in repetitive axial loading,
Eng. Med., 14, 173, 1985.
128. Linde, F. and Hvid, I., Stiffness behavior of trabecular bone specimens, J. Biomech., 20, 82, 1987.
129. Linde, F., Gthgen, C.B., Hvid, I., et al., Mechanical properties of trabecular bone by a nondestructive
compression testing approach, Eng. Med., 17, 23, 1988.
130. Linde, F. and Hvid, I., The effect of constraint on the mechanical behavior of trabecular bone
specimens, J. Biomech., 22, 485, 1989.
131. Linde, F., Hvid, I., and Pongsoipetch, B., Energy absorptive properties of human trabecular bone
specimens during axial compression, J. Orthop. Res., 7, 432, 1989.
132. Burstein, A.H. and Frankel, V.H., The visco-elastic properties of some biological materials, Ann. N.Y.
Acad. Sci., 146, 158, 1968.
133. Madsen, F., Odgaard, A., and Linde, F., The consequences of compression strain level on energy
absorption in trabecular bone specimens, Proc. Meet. Eur. Soc. Biomech., 6, 7A, 1988.
134. Lakes, R.S., Katz, J.L., and Sternstein, S.S., Viscoelastic properties of wet cortical bone I. Torsional
and biaxial studies, J. Biomech., 12, 657, 1979.
135. Kaplan, S.J., Hayes, W.C., Stone, J.L., and Beaupre, G.S., Tensile strength of bovine trabecular bone,
J. Biomech., 18, 732, 1985.
136. Linde, F., Nrgaard, P., Hvid, I., et al., Mechanical properties of trabecular bone. Dependency on stain
rate, J. Biomech., 24, 803, 1991.
137. Saha, S. and Hayes, W.C., Tensile impact properties of human compact bone, J. Biomech., 9, 243, 1976.
0266/ ch11/frame Page 204 Thursday, October 21, 1999 4:25 PM
Tensile and Compression Testing of Bone 205
138. Schafer, M.B., Radin, E.L., and Burr, D.B., Mechanical and morphological effects of strain-rate on
fatigue of compact bone, Bone, 10, 207, 1989.
139. Vrijhoef, M.M. and Driessens, F.C., On the interaction between specimen and testing machine in the
mechanical testing procedures, J. Biomech., 4, 233, 1971.
140. Filon, L.N., On the elastic equilibrium of circular cylinders under certain practical systems of load,
Philos. Trans. R. Soc. London A, 198, 147, 1992.
141. Brown, T.D. and Ferguson, A.B., Mechanical property distributions in the cancellous bone of the
human proximal femur, Acta Orthop. Scand., 51, 429, 1980.
142. Odgaard, A., and Linde, F., The underestimation of Youngs modulus in compression testing of
cancellous bone specimens, J. Biomech., 24, 691, 1991.
143. Allard, R.N. and Ashman, R.B., A comparison between cancellous bone compressive moduli deter-
mined from surface strain and total specimen deection, Trans. Orthop. Res. Soc., 16, 151, 1991.
144. Aspden, R.M., Constraining the lateral dimensions of uniaxially loaded materials increases the cal-
culated strength and stiffness: application to muscle and bone, J. Mater. Sci. Mater. Med., 1, 100, 1990.
145. Guo, X.E., Gibson, L.J., and McMahon, T.A., Fatigue of trabecular bone: avoiding end-crushing
artifacts, Trans. Orthop. Res. Soc., 18, 584, 1993.
146. Keaveny, T.M., Pinilla, T.P., Crawford, R.P., et al., Systematic and random errors in compression
testing of trabecular bone, J. Orthop. Res., 15, 101, 1997.
147. Keaveny, T.M., Wachtel, E.F., Ford, C.M., and Hayes, W.C., Differences between the tensile and
compressive strength of bovine tibial trabecular bone depend on modulus, J. Biomech., 27, 1137, 1994.
148. Carter, D.R. and Hayes, W.C., Bone compressive strength: the inuence of density and strain-rate,
Science, 194, 1174, 1976.
149. Kuhn, J.L., Goldstein, S.A., Ciarelli, M.J., and Matthews, L.S., The limitation of canine trabecular
bone as a model for human: a biomechanical study, J. Biomech., 22, 95, 1989.
0266/ ch11/frame Page 205 Thursday, October 21, 1999 4:25 PM
0266/ ch11/frame Page 206 Thursday, October 21, 1999 4:25 PM

207

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

12

Bending Tests of Bone

Mandi J. Lopez and Mark D. Markel

CONTENTS

I. Introduction ..........................................................................................................................207
II. Three-Point Bending ............................................................................................................209
A. Whole Bone....................................................................................................................209
1. Sample Preparation ...................................................................................................209
2. Testing Methods ........................................................................................................209
3. Data Collection and Calculation...............................................................................210
B. Cancellous Bone.............................................................................................................212
C. Cortical Bone..................................................................................................................212
III. Four-Point Bending..............................................................................................................213
A. Whole Bone....................................................................................................................213
B. Cancellous Bone.............................................................................................................214
C. Cortical Bone..................................................................................................................214
IV. Cantilever Test......................................................................................................................214
A. Whole Bone....................................................................................................................214
B. Cancellous Bone.............................................................................................................216
C. Cortical Bone..................................................................................................................216
References ......................................................................................................................................216

I. INTRODUCTION

The purpose of bending tests of bone is to establish the relative strength of the bone when loads
are applied in a manner that causes it to bend about an axis. The tests may be applied to bone
alone or with xation devices such as interlocking nails, plates and screws, and external xators.
The bone may be intact, ostectomized, osteotomized, or otherwise modied. For most experimental
models, the contralateral limb is used as the control condition. Regardless of the bone or alterations
made to its structure, the same basic principles for bending tests apply.
A bone is subjected to a combination of tension and compression when it is loaded in bending.
Compressive stresses and strains act on one side of the neutral axis while tensile stresses and strains
act on the other.

1,2

There are no normal stresses or strains acting along the neutral axis (Figure 12.1).
The magnitude of the stresses is proportional to their distance from the neutral axis. The farther
the stresses are from the neutral axis, the higher their magnitude. Due to asymmetry of the bone,
the tensile and compressive stresses may not be equal. Since bone is weakest in tension, fractures
propagate from the tensile surface of the bone to the compressive surface transversely until shear
forces acting on a 45

plane become high enough to result in a buttery component on the
compressive side of the bone (Figure 12.2).

3,4

During a bending test, the load applied can be controlled by load control with feedback from
the load cell, or through displacement control with feedback from the crosshead. Both methods
have been employed in bending testing of bone.

5-7

Load is typically applied over a single-cycle

0266/ ch12/frame Page 207 Thursday, October 21, 1999 3:07 PM

208

Mechanical Testing of Bone and the BoneImplant Interface

FIGURE 12.1

Typical fracture morphology in bending. Initially the bone fails in tension and the fracture
propagates toward the compression surface of the bone, resulting in a large buttery fragment.

FIGURE 12.2

Illustration of a cross section of a tibia subjected to bending, showing the distribution of
stresses around the neutral axis (solid line). Tensile stresses act on the anterior surface of the bone and
compressive stresses act on the posterior surface. The stresses are highest on the periosteal surface of the bone
and lower near the neutral axis. The tensile and compressive stresses are unequal because the bone is
asymmetrical.

0266/ ch12/frame Page 208 Thursday, October 21, 1999 3:07 PM

Bending Tests of Bone

209

ramp function. The rate of load application depends on the nature of the investigation, but should
be in the physiological range for studies of normal bone activity, and higher for trauma fracture
studies.

5

For nondestructive testing, preliminary trials may be required to determine the maximum
load or deformation the bone can withstand without sustaining plastic deformation.

6,8

Similarly, if
specic loading rate values are not available for fracture studies, trials may be needed to determine
a loading rate that consistently fractures the bone in a predetermined time period.

7

The specics
of load and displacement control are covered in depth elsewhere in this text.

II. THREE-POINT BENDING

Three-point bending occurs when three forces acting on a bone produce two equal moments. Each
moment is the product of one of the two peripheral forces and its perpendicular distance from the
axis of rotation, the point of application of the middle force (Figure 12.3). If loading continues to
the yield point, the structure should break at the application point of the middle force, assuming
that the structure is homogeneous and symmetrical. Typical three-point bending fractures include
boot top tibial fractures sustained by snow skiers.

1

A. W

HOLE

B

ONE

1. Sample Preparation

Bone specimens to be tested should be harvested as soon as possible postmortem. They are
disarticulated from surrounding joints, wrapped in saline-soaked towels, and sealed in plastic bags
with soft tissues left intact.

9

They are then frozen at 20

C until testing. Bones are allowed to thaw
and reach 20 to 22

C (room temperature) before testing. After removal from the freezer, specimens
should be maintained wet, in saline, during all further stages of tissue preparation. For ease of
application, soft tissues are usually removed prior to application of xation devices, but for all
cases of bone testing in bending the soft tissues should be removed so that a soft tissue component
is not included in testing values. The time between thawing and testing should be kept to a minimum.
One author recommends that no more than 12 h occur between removal from the freezer and
completion of a test so that the conditions of fresh bone are as closely simulated as possible.

10

2. Testing Methods

Typically, the support span for bending testing of an entire bone extends from metaphysis to
metaphysis (Figure 12.4).

11,12

Segments of bone such as the diaphysis can be tested as well.

6

In

FIGURE 12.3

Diagram of cross-sectional distances measured to calculate the area moment of inertia:

X

1

=
lateral-to-medial total width (minimum outside radius),

X

2

= lateral-to-medial medullary canal width (minimum
inside radius),

Y

1

= dorsal-to-palmar total width (major outside radius), and

Y

2

= dorsal-to-palmar medullary
canal width (major inside radius) for the anterior-to-posterior axis.

0266/ ch12/frame Page 209 Thursday, October 21, 1999 3:07 PM

210

Mechanical Testing of Bone and the BoneImplant Interface

such cases, the support span typically extends from each end of the segment with just enough bone
extending over the end supports to ensure good contact. The support should be strong enough to
withstand the forces necessary to test the bone in bending, wide enough to support the bone width,
and of sufcient length for the area of interest to be contained within the support span. The end
supports should be smooth, at, and perpendicular to the horizontal axis. Typically, the samples
tested are close enough in size such that a single support designed for the average bone can be
used, although adjustable end supports have been used as well.

10,12

Three-point bending requires the use of an actuator with a single point of application. The
application point of the actuator should be located on the midline between the two end supports
(Figure 12.3). The actuator is applied according to the desired direction of bending. For example,
posterior to anterior bending is performed with the anterior surface of the bone facing downward
on the support and the load applied on the posterior aspect of the bone. Lateromedial bending is
performed with the load applied on the lateral surface.
The testing loading rate depends on the experimental study as discussed above and data obtained
from preliminary trials. These authors prefer to test bone nondestructively by applying load over
a single-cycle ramp function at a constant displacement rate of 10 mm/min until a maximum
displacement of 0.8 mm is obtained. To test specimens to failure, the same displacement rate is
used but with no displacement end limit.

3. Data Collection and Calculation

Force and crosshead displacements are recorded from the materials testing system during testing.
The authors collect data at 0.1-s intervals and record it on a personal computer directly linked to
the materials testing system. One of the most important points relevant to data collection is that
the frequency of collection should be high enough to reect continuous data collection as closely
as possible. Loaddeformation curves are generated for all tests. For bending tests, load is repre-
sented as the bending moment (Nm). Stiffness values are calculated as the slope of the linear
regression t from the straight portion of the curve. Failure is generally dened as the point at
which the bone fractures or the central load applicator visibly crushes the underlying cortex.
Most equations to calculate the structural and material properties of bone in bending are based on
long prismatic beams where the beam is initially straight, the cross section of the beam does not vary

FIGURE 12.4

Three-point bending occurs when three forces acting on a bone produce two equal moments.
Each moment is the product of one of the two peripheral forces and its perpendicular distance from the point
of application of the middle force.

0266/ ch12/frame Page 210 Thursday, October 21, 1999 3:07 PM

Bending Tests of Bone

211

along its length, and the beam is made of an isotropic, homogeneous, linearly elastic material. Bone
does not conform to many of these assumptions, but calculations based on these equations provide a
means for comparison between studies. Three properties of bones commonly calculated for three-point
bending include area moment of inertia, breaking strength, and modulus of elasticity.

6,7,10,13-16

Bending moment is determined by the formula:
M = FL/4 (12.1)
where

M

= bending moment (Nm),

F

= applied force,

L

= distance between two end supports.

17

Bending moment allows comparisons of force to be made between bones of different lengths.
Stress is dened as force per unit area. The area moment of inertia accounts not only for
differences in area, but also for differences in shape of the bone cross section through which the
force is applied. Bending stress can be calculated from the exure formula:

S

=

Mc/I

(12.2)
where

S

= stress (N/m

2

),

M

= bending moment on the cross section,

c

= distance from the farthest
point in the cross section to the neutral axis, and

I

= area moment of inertia.
Again, assuming prismatic beam theories, the area moment of inertia can be calculated for a
whole bone in posterior-to-anterior bending with the formula:

I

=

!

/64(

X

1

Y

1
3



X

2

Y

2
3

) (12.3)
where

I

= area moment of inertia,

X

1

= lateral-to-medial total width (minimum outside radius),

X

2

=
lateral-to-medial medullary canal width (minimum inside radius),

Y

1

= anterior-to-posterior total
width (major outside radius), and

Y

2

= anterior-to-posterior medullary canal width (major inside
radius) for the anterior-to-posterior axis (Figure 12.4).

6,8,10,12,16

When bending is performed about the opposite axis,

x

and

y

values should be interchanged.
For the most applicable results, specimens should be measured where the fracture originates through
the cortex.

16

Stress allows comparisons to be made between strengths of bones that differ in length, size,
and shape. For three-point bending, breaking stress is calculated with the formula:
BS

max

=

M

max

C/I

= (

FLC

/4

I

) (12.4)
where BS = breaking stress (N/m

2

),

F

= load at failure,

L

= distance between end supports, and

C

=

y

1

/2 = distance from the centroid to the surface (see Figure 12.4), and

I

= area moment of
inertia.

16

Bending stress is calculated using Equation 12.4 at a point prior to the failure point.
The modulus of elasticity for a bone in three-point bending is calculated with the formula:

E

= (BS

max

L

3

)/(48

Id

a

) (12.5)
where

E

= modulus of elasticity, BS

max

= breaking strength,

d

a

= the deformation at the point of
load application measured as actuator displacement, and

L

= distance between end supports.

8,12,15,18

Strain takes into account the amount of bending or deformation that occurs in the bone as it
is being tested. Strain is unitless since it is the change in length per unit length. The formula for
calculation of strain in three-point bending is

"

= (12

d

a

C

)/

L

2

(12.6)

0266/ ch12/frame Page 211 Thursday, October 21, 1999 3:07 PM

212

Mechanical Testing of Bone and the BoneImplant Interface

where

d

a

= the deformation at the point of load application measured as actuator displacement,

C

=

y

1

/2 = distance from the centroid to the surface, and

L

= distance between end supports.

18

B. C

ANCELLOUS

B

ONE

The same principles of harvesting and preservation for whole bone apply to cancellous bone.
Samples of cancellous bone may be prepared at the same time as collection or after thawing.

19,20

The shape of the samples is generally cylindrical, but other shapes are used, and the area moment
of inertia is calculated accordingly. Depending on the directional strength of interest, the samples
may be cut longitudinally, transversely, or in any trabecular orientation.

19-22

Often, samples from
different areas within the bone are collected.

20,21

Cancellous bone samples are usually cored from
the appropriate bone while using a physiological solution lavage.

19,20

In a typical longitudinal section
preparation, a stainless-steel coring bit is used to extract cylindrical dowels of cancellous bone
beginning at the articular surface and continuing proximally in a plane perpendicular to the ground,
usually with the bone in a normal standing weight-bearing orientation. A physiological solution is
used during drilling to moisten and cool the specimen. Ideally, it is forced through the drill bit
while drilling. The cylinders are then cut to the appropriate length using a low-speed saw with
parallel cutting blades again while using a physiological lavage solution. Care is taken to exclude
the subchondral bone. Cylinder length and diameter are normally measured using a caliper. Testing
methods are similar to those described for whole-bone three-point bending. The size and speci-
cations of the testing support correspond to the size of the specimens. Loading rates are again
dependent upon the chosen model (i.e., fracture vs. physiological loading). A sample loading rate
for cancellous bone is 0.05 mm/sec.

20

Data collection for cancellous bone is similar to that of whole bone. Once again, the frequency
of data collection should reect continuous collection as closely as possible. Calculation of data
for a solid cylinder or rectangle generally requires fewer manipulations than that of whole bone.
Bending moment, bending and breaking stresses, and modulus of elasticity can be calculated from
Equations 12.1, 12.4, and 12.5, respectively. The formula for area moment of inertia for a solid
cylinder is:

I

cylinder

=

!

r

4

/4 (12.7)
where

r

= radius of the sample.

4

The formula for area moment of inertia of a solid rectangle is

I

rectangle

=

bh

3

/12 (12.8)
where

b

= base and

h

= height.

4

C. C

ORTICAL

B

ONE

Collection and preservation of bones for cortical sample collection are the same as for whole bones.
Long bones are usually sectioned with a band saw prior to removal of cortical samples.

23,24

The
cortex can then be divided into regions and the samples subsequently lathed into right cylinders,
or cylinders can be cored from diaphyseal sections using a diamond-tipped hole saw.

23,24

As
before, physiological solution is applied to the bone during all trimming to provide moisture and
cooling. Cortical samples are often collected from different areas of the bone for comparative
purposes.

24

Samples are typically oriented with the long axis of the bone, although any direction
is possible. The same methods of testing whole and cancellous bone samples in three-point
bending described in Sections II.A and II.B, respectively, apply to cortical bone samples. The
same methods and formulas as for cancellous bone described in Section II.B accomplish data
collection and calculation.

0266/ ch12/frame Page 212 Thursday, October 21, 1999 3:07 PM

Bending Tests of Bone

213
III. FOUR-POINT BENDING
Four-point bending occurs when two force couples acting on a structure produce two equal
moments. A force couple refers to a pair of parallel forces of equal magnitude but opposite direction
applied to a structure (Figure 12.5). The bending moment magnitude is the same throughout the
area between the force couples; hence, the structure being tested should fracture at its weakest
point. This arrangement is advantageous for testing where one might be uncertain about the strongest
or weakest point and does not wish to inuence the test by locating the maximum bending moment
at a specic place. A clinical example of a four-point bending fracture is a femoral fracture through
a previous fracture site resulting from one force couple formed by the posterior knee joint capsule
and tibia and the other by the femoral head and hip joint capsule.
A. WHOLE BONE
Whole bone sample preparation is described in Section II.A.
The testing methods described in Section II.A for loading rates, application direction, and
support span construction are essentially the same for three- and four-point bending. The major
difference between three- and four-point bending is the construction of the actuator. The arms of
the actuator are usually spaced such that the area of interest is located between them (see
Figure 12.5). This ensures that the bending moment is uniform throughout that area. The size and
material structure of the support span and actuator are again dictated by the specimen to be tested.
For accurate results, the two central points of load application must contact the bone at the same
time. Due to the irregular surface shape of some bones, the span between the actuator arms and
hence the span tested may be limited. All points in contact with bone should be smooth and rounded
to prevent stress concentration.
Data collection is the same as described for three-point bending in Section II.A. Calculations
for some of the biomechanical parameters of interest are different due to differences in load
application. Equations 12.2 and 12.3 can be used to calculate bending stress and area moment of
inertia, respectively. For four-point bending, bending moment is calculated with the formula:
M = FL/6 (12.9)
FIGURE 12.5 Four-point bending occurs when two force couples acting on a structure produce two equal
moments. The bending moment magnitude is the same throughout the area between the force couples.
0266/ ch12/frame Page 213 Thursday, October 21, 1999 3:07 PM
214 Mechanical Testing of Bone and the BoneImplant Interface
where M = bending moment (Nm), F = applied force, and L = distance between two end supports.
Breaking stress is calculated with the formula:
BS
max
= M
max
C/I = (F/2a)C/I (12.10)
where BS = breaking stress (N/m
2
), F = load at failure, a = distance from the end support to the
nearest point of load application, C = y
1
/2 = distance from the centroid to the surface (see
Figure 12.4), and I = area moment of inertia.
6
Bending stress is calculated using Equation 12.10 at
a point prior to the failure point. The modulus of elasticity is calculated with the formula:
E = [BS
max
/2a
2
(3L 4a)]/(6Id
a
) (12.11)
where E = modulus of elasticity, BS
max
= breaking stress, a = distance from the end support to the
nearest point of load application, L = distance between end supports, I = area moment of inertia,
and d
a
= the deformation at the point of load application measured as actuator displacement.
6
B. CANCELLOUS BONE
Cancellous bone sample preparation is described in Section II.B for three-point bending.
29
Testing
methods are the same as described in Section II.B for three-point bending with the exception of
the actuator, which must be modied for four-point bending as described previously for whole-
bone four-point bending. Data collection is the same as described for three-point bending in
Section II.A. The same basic formulas are used for biomechanical parameter calculation as in whole
bone four-point bending with differences owing to the shape. Equation 12.9 is used to calculate
bending moment. Equation 12.7 or 12.8 is used to calculate area moment of inertia of the sample
depending on its shape. Equations 12.10 and 12.11 are used to calculate breaking and bending
stresses and modulus of elasticity, respectively.
C. CORTICAL BONE
Cortical bone sample preparation is described in Section II.C for three-point bending. Testing
methods are the same as described in Section II.C for three-point bending with actuator modica-
tions as described previously. Data collection and calculation is accomplished by the same methods
and formulas for cancellous bone described in Section III.B.
25
IV. CANTILEVER TEST
Cantilever bending refers to a loading arrangement in which one end of the specimen is rigidly
xed while the other end is completely free. The bending moment varies from a maximum at the
xed end to zero at the force application point (Figure 12.6). This particular bending test may be
applied to a whole bone or to part of a whole bone as well as to cancellous and cortical specimens.
A. WHOLE BONE
Whole bone sample collection and preservation is described in Section II.A. For cantilever testing,
one end of the bone must be xed with the area of interest free. This may be accomplished in a
number of ways, the most common of which is to pot one end of the bone into a container. Potting
materials include such substances as polyester resins and low-melting alloys.
26,27
Care must be
taken to ensure that the material used to pot the bone does not affect the bone substance during
the polymerization phase. The container may be made from anything ranging from PVC tubing to
stainless steel, depending on the size of the specimen and the forces to be applied.
26-29
It is important
0266/ ch12/frame Page 214 Thursday, October 21, 1999 3:07 PM
Bending Tests of Bone 215
to pot the bone centrally and to direct it parallel to the longitudinal axis of the potting container
so that the orientation of applied forces does not vary between specimens when the sample is
clamped in the mechanical testing system. A part of a whole bone may be prepared for cantilever
bending as well. A specic example is application of vertical force to the femoral head of an intact
proximal femur. The distal aspect of the femur is potted with the same principles of potting materials
and orientation as for a whole-bone sample.
28
The majority of cantilever bending testing is performed with the force applied at the point on
the free end most distant from the xed end of the bone so that the highest bending moment is
obtained and the majority of the sample is included in the testing. The forces can be applied at
different levels on the sample, however, with care taken to ensure that they are not applied directly
on the area of interest. The side of the bone to which the load is applied depends on which surfaces
are to be under tensile or compressive loads and thus which bending direction is desired. Specimens
can also be oriented at different angles within the materials testing system to simulate various
anatomical positions.
27
Cantilever bending requires an actuator with a single load application point.
The actuator size and shape is dependent on the bone being tested. As described previously for
other testing structures, the actuator should be smooth and of adequate material properties to
withstand applied loads. Loading rates depend on the desired model and the bone being testing as
described before. Examples of loading rates include 5 mm/min displacement for cantilever testing
of human phalangeal specimens and 16 N/s loading for cantilever testing of intact rat femora.
26,27
Data collection is the same as described for three-point bending in Section II.A. Area moment
of inertia is calculated using Equation 12.3. The formula for bending moment for cantilever
bending is:
M = FL (12.12)
where M = bending moment (Nm), F = applied force, and L = distance between xed end and
applied load.
4
Breaking stress is provided by the formula:
BS
max
= (F)(d)(C/2I) = F sin #/A + [(F cos #)d]C/I (12.13)
FIGURE 12.6 Cantilever bending refers to a loading arrangement in which one specimen end is rigidly fixed
while the other end is completely free. The bending moment varies from a maximum at the xed end to zero
at the force application point.
0266/ ch12/frame Page 215 Thursday, October 21, 1999 3:07 PM
216 Mechanical Testing of Bone and the BoneImplant Interface
where BS
max
= breaking stress (N/m
2
), F = load at failure, # = angle between fracture surface and
the vertical, A = total cross-sectional area, d = distance between the xed end and the applied load,
C = distance from the centroid to the surface, and I = area moment of inertia.
28
Bending stress is
calculated using Equation 12.13 at a point prior to the failure point.
B. CANCELLOUS BONE
Cancellous bone sample preparation is described in Section II.B for three-point bending. Potting
procedures are the same as for cantilever testing of whole-bone specimens in Section IV.A. Can-
tilever testing methods for cancellous bone are the same as those for whole bone described in
Section IV.A. with considerations made for size and shape of the samples. Data collection is the
same as described for three-point bending in Section II.A. Area moment of inertia, bending moment,
and bending and breaking stresses are calculated using Equations 12.7 or 12.8, 12.12, and 12.13,
respectively.
C. CORTICAL BONE
Cortical bone sample preparation is described in Section II.C. for three-point bending. Potting
procedures are the same as for cantilever testing of whole-bone specimens in Section IV.A. Canti-
lever testing methods for cancellous bone are the same as those for whole bone described in
Section IV.A. with considerations made for size and shape of the samples. Data collection and
calculation for cantilever bending of cortical bone are the same as for cancellous bone described
in Section IV.B.
REFERENCES
1. Nordin, M. and Frankel, V.H., Biomechanics of bone, in Basic Biomechanics of the Musculoskeletal
System, Nordin, R.W. and Frankel, V.H., Eds., Lea & Febiger, Philadelphia, 1989, 3.
2. Trostle, S.S. and Markel, M.D., Fracture biology, biomechanics, and internal xation, Vet. Clin. North
Am. Food Anim. Pract., 12, 19, 1996.
3. Markel, M.D., Fracture biomechanics, in Equine Fracture Repair, Nixon, Ed., W.B. Saunders, Phil-
adelphia, 1996, 10.
4. Tencer, A.F., Johnson, K.D., Kyle, R.F., and Fu, F.H., Biomechanics of fractures and fracture xation,
in Instructional Course Lectures, American Academy of Orthopaedic Surgeons, Heckman, J.D., Ed.,
Rand McNally, Taunton, MA, 1993, 19.
5. Ashman, R.B., Experimental techniques, in Bone Mechanics, Cowin, S.C., Ed., CRC Press, Boca
Raton, FL, 1989, 75.
6. Hanson, P., Markel, M., and Vanderby, R., Jr., Diaphyseal structural properties of equine long bones,
Am. J. Vet. Res., 56, 233, 1995.
7. McDuffee, L.A., Stover, S.M., Taylor, K.T., and Les, C.M., An in vitro biomechanical investigation
of an interlocking nail for xation of diaphyseal tibial fractures in adult horses, Vet. Surg., 23, 219,
1994.
8. Kasra, M., Vanin, C.M., MacLusky, N.J., et al., Effects of different estrogen and progestin regimens
on the mechanical properties of rat femur, J. Orthop. Res., 15, 118, 1997.
9. Walter, M.C., Smith, G.K., and Newton, C.D., Canine lumbar spinal internal xation techniques, Vet.
Surg., 15, 191, 1986.
10. Bynum, D., Ledbetter, W.B., Boyd, C.L., and Ray, D.R., Flexural properties of equine metacarpus,
J. Biomed. Mater. Res., 5, 63, 1971.
11. Combs, N.R., Kornegay, E.T., Lindemann, M.D., et al., Calcium and phosphorus requirement of swine
from weaning to market weight: II. Development of response curves for bone criteria and comparison
of bending and shear bone testing, J. Anim. Sci., 69, 682, 1991.
12. Jarvinen, T.L., Sievanen, H., Kannus, P., and Jarvinen M., Dual-energy X-ray absorptiometry in
predicting mechanical characteristics of rat femur, Bone, 22, 551, 1998.
0266/ ch12/frame Page 216 Thursday, October 21, 1999 3:07 PM
Bending Tests of Bone 217
13. Currey, J.D., The mechanical consequences of variation in the mineral content of bone, J. Biomech.,
2, 1, 1969.
14. Pope, M.H. and Outwater, J.O, Mechanical properties of bone as a function of position and orientation,
J. Biomech., 7, 61, 1974.
15. Simkin, A. and Robin, G., The mechanical testing of bone in bending, J. Biomech., 6, 31, 1973.
16. Specht, T.E., Miller, G.J., and Colahan, P.T., Effects of clustered drill holes on the breaking strength
of the equine third metacarpal bone, Am. J. Vet. Res., 51, 1242, 1990.
17. Brennan, J.J. and Aherne, F.X., The effect of dietary calcium and phosphorus levels on performance,
bone bending moment and the severity of osteochondrosis and lameness in boars and gilts slaughtered
at 100 or 130 kg body weight, Can. J. Anim. Sci., 66, 777, 1986.
18. Crenshaw, T.D., Peo, E.R., Jr., Lewis, A.J., et al., Inuence of age, sex and calcium and phosphorus
levels on the mechanical properties of various bones in swine, J. Anim. Sci., 52, 1319, 1981.
19. Kaneps, A.J., Stover, S.M., and Lane, N.E., Changes in canine cortical and cancellous bone mechanical
properties following immobilization and remobilization with exercise, Bone, 21, 419, 1997.
20. Oden, Z.M., Selvitelli, D.M., Hayes, W.C., and Meyers, E.R., The effect of trabecular structure on
DXA-based predictions of bovine bone failure, Calcif. Tissue Int., 63, 67, 1998.
21. Banse, X., Delloye, C., Cornu, O., and Bourgois, R., Comparative left-right mechanical testing of
cancellous bone from normal femoral heads, J. Biomech., 29, 1247, 1996.
22. Wohl, G.R., Loehrke, L., Watkins, B.A., and Zernicke, R.F., Effects of high-fat diet on mature bone
mineral content, structure, and mechanical properties, Calcif. Tissue Int., 63, 74, 1998.
23. Courtney, A.C., Hayes, W.C., and Gibson, L.J., Age-related differences in post-yield damage in human
cortical bone. Experiment and model, J. Biomech., 29, 1463, 1996.
24. Les, C.M., Stover, S.M., Keyak, J.H., et al., The distribution of material properties in the equine third
metacarpal bone serves to enhance saggital bending, J. Biomech., 30, 355, 1997.
25. Bigot, G., Bouzidi, A., Rumelhart, C., and Martin-Rosset, W., Evolution during growth of the mechan-
ical properties of the cortical bone in equine cannon-bones, Med. Eng. Phys., 18, 79, 1996.
26. Battraw, G.A., Miera, V., and Anderson, P.L., et al., J. Biomed. Mater. Res., 32, 285, 1996.
27. Campbell, J.T., Schon, L.C., Parks, B.G., et al., Mechanical comparison of biplanar proximal closing
wedge osteotomy with plantar plate xation vs. crescentic xation for the correction of metatarsus
primus varus, Foot Ankle Int., 19, 293, 1998.
28. Hou, J.C., Salem, G.J., Zernicke, R.F., and Barnard, R.J., Structural and mechanical adaptations of
immature trabecular bone to strenuous exercise, J. Appl. Physiol., 69, 1309, 1990.
29. Salem, G.J., Zernicke, R.F., Martinez, D.A., and Vailas, A.C., Adaptations of immature trabecular
bone to moderate exercise: geometrical, biochemical, and biomechanical correlates, Bone, 14, 647,
1993.
0266/ ch12/frame Page 217 Thursday, October 21, 1999 3:07 PM
0266/ ch12/frame Page 218 Thursday, October 21, 1999 3:07 PM

219

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

13

Torsional Testing of Bone

Benjamin R. Furman and Subrata Saha

CONTENTS

I. Introduction ..........................................................................................................................219
II. Theory of Torsion for Cylinders..........................................................................................220
III. Effect of Structural Defects .................................................................................................221
IV. Fixation to the Test System.................................................................................................222
V. Automated Mechanical Testing Systems.............................................................................222
A. Servohydraulic Systems .................................................................................................222
B. Electromechanical Systems............................................................................................226
C. System Selection ............................................................................................................226
VI.

In Vitro

Testing Methodology for Servohydraulic Systems................................................226
A. Adjusting the Crosshead Position..................................................................................226
B. Internal Controller Calibration.......................................................................................227
C. External Load Calibration..............................................................................................227
D. Test Preparation: The Ramp Test...................................................................................227
E. General Ramp Testing Procedure ..................................................................................227
F. Test Preparation: The Fatigue Test ................................................................................228
G. General Fatigue Testing Procedure................................................................................229
H. The Creep Test ...............................................................................................................229
VII. Summary ..............................................................................................................................229
References ......................................................................................................................................229
Further Reading..............................................................................................................................230

I. INTRODUCTION

During normal daily activities, the skeletal system is subjected to a complex system of loading
exerted by the forces of gravity and the muscles attached to the bones. Such loading modes include
tensile, compressive, bending, and torsional forces applied to the bones of the skeletal system.
Therefore, in evaluating the tolerance limits of bones, it is important to determine the failure
behavior of bones under all of these loading conditions. This chapter will discuss the testing of
bones in torsion. This is important as many of the long bones as well as the spine are often subjected
to a signicant amount of torsional load. However, only limited information is available in the
literature on the mechanical behavior of bones under torsion.

1-3

This is true for the torsional testing
of whole bones as well as for testing machined compact and cancellous bone samples.

3,4

This is
partly because most mechanical testing machines (both screw driven and servohydraulic) available
in engineering schools in this country and abroad are suitable for tension, compression, or bending
tests. One needs to use a specially designed test setup or a biaxial servohydraulic mechanical testing
system for torsional testing of bones. Although screw-driven torsional testing machines exist, they are
very rare. On the other hand, a pendulum type of torsional impact tester was popularized by Frankel
and Burstein

5

in the 1960s and is a common mechanical testing machine in many biomechanics

0266/ ch13/frame Page 219 Thursday, October 21, 1999 4:34 PM

220

Mechanical Testing of Bone and the BoneImplant Interface

laboratories in this country, particularly in medical schools.

5-7

However, unless instrumented, such
an impact tester only provides the information on the total energy to failure. When instrumented
with a dynamic load cell and a suitable rotational measurement device, a pendulum-type torsional
tester can provide information regarding the rotational stiffness, maximum torque, and the maximum
rotational angle before failure.
It should be pointed out that for testing torsional properties of metallic implants, certain standards
have been adopted by the American Society for Testing and Materials (ASTM).

8

However, no such
standards exist for the mechanical testing of bones. Thus, when mechanical testing of bones is being
planned, one should consider using the loading rate which would approximate the loading situation
simulating the

in vivo

condition being examined.

9

Moreover, one should also remember that as bone
is a viscoelastic material, mechanical property data generated from different experiments can only be
compared if the loading rates employed are similar. It should also be noted that torsional testing is
important in the evaluation of many surgical constructs using orthopaedic implants.

10-12

II. THEORY OF TORSION FOR CYLINDERS

A diaphyseal segment from a long bone might be grossly approximated as a hollow cylindrical
shaft made from a homogeneous, linear elastic material. Such a shaft might have a certain inner
radius,

r

i

, an outer radius,

r

o

, and a length,

L

. If one end,

A

, of the shaft is xed, and a torsional
force,

T

, is applied to the opposite end,

B

, then end

B

will rotate in its own plane through some
angle

with respect to end

A

, as illustrated in Figure 13.1.
In order to nd the shear stress,

, in the material at any radius within the cross section of the
shaft, the following simple formula is used as a guide

13

:
(13.1)
where

J

is the polar moment of inertia, which for a hollow cylinder is equal to

(

r

o

4



r

i

4

)/2, and

is a specied radius, bounded by

r

o

and

r

i

. Thus, the maximum shear stress,

max

is given by
(13.2)
The relative angle between ends

A

and

B

is similarly given by
(13.3)

FIGURE 13.1

Simple torsional rotation of the free end of a hollow cylinder.

=
T
J

max
=
Tr
J
o
=
TL
JG

0266/ ch13/frame Page 220 Thursday, October 21, 1999 4:34 PM

Torsional Testing of Bone

221

where

G

is the elastic modulus of the bone in shear.
Since a cylindrical shaft is perfectly round in its perpendicular cross section, all cross sections
along the entire length of the shaft will remain planar and parallel to one another. Any deviation
in form away from this idealized cylindrical form will cause out-of-plane deformation to occur as
torsional load is applied. Realizing that bones are not normally perfectly cylindrical in form, the
authors emphasize that Equations 13.1 and 13.2 can only give rough approximations of the behavior
observed in real bones. A closer approximation can be obtained from computed mechanical models
using actual mechanical test data, especially when examining localized behaviors within a complex
formation. An example of such a detailed analysis can be found in the work of Levenston et al.

14

III. EFFECT OF STRUCTURAL DEFECTS

The theories described by Equations 13.1 and 13.2 are only valid for closed-section structures such
as

intact

cortical bone. However, it is important to bear in mind two important structural deviations
from the norm: so-called open structures, and closed structures containing small defects or holes.
First, consider that the torsional rigidity of

any

cylindrical structure, no matter how irregular,
will be considerably reduced by a longitudinal cut or slit that continues along the entire or part of
the length of the shaft. Such a slit is usually referred to as an opening in the structure.
To return again to an idealized case where there are ideal hollow cylinders, one open and one
closed, which are otherwise identical in cross section (Figure 13.2), the open structure will have a
torsional load-carrying capacity that is reduced (multiplied) by a factor of

t

/3

r

, where

t

is the wall
thickness of the cylinder and

r

is its mean radius, with respect to the closed structure. One should
always account for this difference, using the theoretical value as a rough guideline wherever a
complete opening exists in the test specimen. This is especially important, for example, when
comparing device designs for intramedullary nails. Exaggerated cases of open structure include
devices having C-, U-, V-, and I-shaped cross sections.
Also of concern are holes, or incomplete openings, in the wall of a cylindrical material structure.
Examples include pinholes in cortical bone or screw holes in prosthetic devices. Such defects will
considerably reduce torsional strength as a matter of principle.

7,15,16

It is frequently useful in bio-
mechanical testing to evaluate the effects of round holes placed in cortical bone to predict the relative
likelihood of

in vivo

injury due to pin and/or screw placement. Most often, the effect of holes on the
strength of the material is referred to as a

stress concentration

effect, whereby the localized stress in
the material immediately surrounding the defect is higher than that predicted for the bulk of the material
structure.

17-19

Perfectly round holes will increase the localized stress in the material by a certain

factor

,

max

/

nominal

, which is equal to 3 for an idealized structure loaded in simple torsion.

20

Other holes will
have different stress concentration factors, depending on their size, shape, and orientation in combina-
tion with the loading mode for the structure. The intuitive conclusion is that peak failure loads should
be expected to be considerably reduced, and this must always be planned for when setting up the
expected ranges for load and deformation during a mechanical test.

FIGURE 13.2

Cross-sectional view of a closed (a) and an open (b) cylinder.

0266/ ch13/frame Page 221 Thursday, October 21, 1999 4:34 PM

222

Mechanical Testing of Bone and the BoneImplant Interface

Wherever it is desirable to predict the behavior of whole bones, which have complex geometries
and orthotropic material properties, it is most convenient for the surgeon to investigate empirical
results from mechanical testing. Mechanical testing, and torsional testing in particular, may be
performed with a wide variety of equipment on a wide variety of bones even those with unusual
geometries using a wide variety of loading modes. Please see the Further Reading list at the
end of this chapter for a few examples.

IV. FIXATION TO THE TEST SYSTEM

For

in vitro

torsional testing, it is usually helpful to embed, or pot, a portion of the bone into a
moldable material such that intimate xation can be achieved. This embedding material may, in
turn, be held within a metal sleeve, cap, or other similar device in order to mate appropriately with
the test machine.
The ideal embedding material should be easily formed, adhere well to the bony surface, become
perfectly rigid once set, and be easily removable. While it is unrealistic to expect all of these
characteristics from one material, there are a few materials available which may approximate to
the ideal properties.
Epoxies and other self-curing, thermosetting polymers are perhaps the best embedding
materials. They are available in liquid, paste, and dough forms that can all be suitably formed into
close apposition with the bony surface. They undergo a low degree of shrinkage during curing and
are also adhesive, thereby forming a strong bond with both the bone and any supporting metal
devices once it is fully cured. The resulting construct is rigid enough for all practical purposes. If
additional stiffness is required, reinforcing glass particles or short bers may be added to the epoxy.
Epoxies have the additional advantage of resistance to moisture. Once the epoxy has set, it can be
completely submerged in saline without fear of loosening.
Many previous investigators have used polymethylmethacrylate (PMMA) bone cement as a
embedding material because it is perceived to be compatible with bone. While it does have
excellent working properties and sets up to be reasonably rigid in a short period of time, it may
not be cost effective to use when compared with the hardware store epoxies. Moreover, the
exotherm generated by curing PMMA is much higher than that of the epoxies.
Beyond the polymers, another choice of embedding material is Woods metal. This alloy has
a melting temperature of 70

C, which is accessible with a small warming plate. Unfortunately, the
temperature is also high enough to cause protein denaturation; however, the metal quenches quite
rapidly. Woods metal offers the advantage of higher elastic modulus, as compared with the
thermosetting polymers, while still conforming readily to irregular bony structures and being easy
to use.

21

In addition to or in place of embedding materials, pins and clamps may be used to hold the
epiphyses of long bones. If used alone, they have the advantage of being easily removed. The
xtures will necessarily cause pinhole and clamping stresses at the attachment points; however,
this should not cause any great difculty if the test length of the specimen is far away from the
attachment points and is expected to fail more readily than the xed ends. Figure 13.3 shows an
example of such xation of a spine segment in a clamp, and Figure 13.4 shows the results of such
torsional testing of a thoracolumbar spine segment using such a clamp.

22

If combined with an
embedding material, pins can provide an additional mechanical interlock with the xture.

V. AUTOMATED MECHANICAL TESTING SYSTEMS
A. S

ERVOHYDRAULIC

S

YSTEMS

The most exible and commonly used mechanical testing systems for examining the torsional
properties of materials under a wide variety of loading conditions are modern servohydraulic

0266/ ch13/frame Page 222 Thursday, October 21, 1999 4:34 PM

Torsional Testing of Bone

223

machines. All servohydraulic testing machines have a large load frame, a substantial hydraulic
pump, a valve-controlled actuator piston, a linear variable differential transducer (LVDT) to measure
the axial motion, and a load transducer or load cell to measure the axial thrust.

23

FIGURE 13.3

A thoracolumbar spine segment held by clamps for torsional testing in a servohydraulic
mechanical testing system.

FIGURE 13.4

A comparison of the torsional stiffnesses of spine segments as evaluated by torsional testing.
(Data from Reference 22.)

0266/ ch13/frame Page 223 Thursday, October 21, 1999 4:34 PM

224

Mechanical Testing of Bone and the BoneImplant Interface

Biaxial testing systems are a special class of servohydraulic machines which can exert and
respond to the torsional loads in addition to the axial loads (Figure 13.5).

23

In addition to the
components mentioned above, these machines require a torsional actuation mechanism, a rotational
variable differential transducer (RVDT) to measure the angular rotation, and a torque cell to measure
the applied torque.
The LVDT, RVDT, load cell, and torque cell are all electromechanical transducers that yield
to mechanical deformation and produce an electrical signal output which can be used as the

feedback

signal. Other transducers that generate useful feedback signals include resistive strain
gauges and extensometers that can be applied externally to the specimen. In exactly the opposite
way, the

actuator

responds to an electrical signal, called the

command

signal, by producing a
mechanical motion. All servohydraulic control systems operate by carefully balancing the com-
mand and feedback signals to produce motion when necessary and generate information about
the loaddeformation response of the specimen. The machines cannot operate without the presence
of a specimen because the specimen serves as a link between the command and feedback systems,
as shown in the simplied diagram (Figure 13.6).
Testing is generally conducted under three types of control, or control mode:

load

,

deformation

,
and

strain

. Most testing is conducted under displacement control; however, there are situations
where the other two control modes are quite useful.
In load control mode, the system controller allows a specied command signal (voltage) to be
applied to the actuator, which responds immediately by moving. The system controller oversees
the test by continually monitoring the load/torque cell feedback. After some time, when a specied

FIGURE 13.5

A new biaxial servohydraulic mechanical testing machine (Instron model 8874) with hydraulic
grips. The electronic control panels are shown in the bottom right corner.

0266/ ch13/frame Page 224 Thursday, October 21, 1999 4:34 PM

Torsional Testing of Bone

225

feedback from the load/torque cell is detected by the system controller, the controller will no longer
signal the actuator to move, and the loading cycle is completed. The difference between the starting
load and the total expected load is called the

load range

for the test. The rate at which voltage is
applied to the actuator by the system controller is determined by the machine operator and is called
the

loading rate

. The voltage may vary linearly, sinusoidally, or according to some other continuous
function. This type of control mode is quite useful for fatigue testing, where load is varied cyclically
over a specied range.
In deformation control mode, the machine operator again determines the rate at which voltage
is applied to the controller, but in this case the controller input is a signal from the LVDT/RVDT.
The total expected actuator motion is called the

displacement range

, and the controller voltage is
based on the

displacement rate.

For biological materials, which are generally much more compliant
than the machine itself, the machine displacement may be considered approximately equal to the
overall specimen deformation. Once a test has begun, feedback from the LVDT/RVDT is recorded
and measured vs. either (1) the displacement feedback from the LVDT/RVDT or (2) the specimen
strain feedback from an external transducer. The resulting plot indicates how well the specimen
has withstood a given loading. Most tests performed in this way are run slowly, or

quasi-statically

.
After the specimen has failed, the actuator continues to move until a displacement limit has been
reached. It is usually convenient to stop the test manually.
In strain control mode, the controller will apply a command signal until a specied level of
feedback from an external strain gauge or extensometer is detected. Fatigue testing can be performed
over a specied

strain range

, and some machine operators prefer this control method to load control.
The feedback signals from the electromechanical transducers are usually amplied by circuitry
in the controller, and the amplied signals can be used to drive a host of plotters to generate curves
of applied load vs. linear displacement, torque vs. angular rotation, etc. The most sophisticated
systems read the output signals by computer, and the associated software packages are usually
quite convenient for calculating the work of fracture and the elastic moduli for a given test.
Biaxial servohydraulic testing systems (see Figure 13.5) often require a signicant amount of oor
space and other resources to operate but can give a wealth of information in return. More recently,
system manufacturers have been introducing smaller, more specialized machines to handle small-scale
testing. It is important to match machine size with the expected range of specimens to be tested.
Another point of signicance is that servohydraulic machines are very quick to respond to
controller input and are capable of generating extremely high or extremely low loading rates. This
can be useful for simulating impact in one test and creep in the next. It should be noted here,
however, that no mechanical testing system may be able to match perfectly the desired loading
prole with respect to time, whether it be linear, sinusoidal, or some other prole. The most
sophisticated servohydraulic systems offer digital control and very ne sensitivity; however, simpler

FIGURE 13.6

Control/feedback linkage including specimen.

0266/ ch13/frame Page 225 Thursday, October 21, 1999 4:34 PM

226

Mechanical Testing of Bone and the BoneImplant Interface

systems can often be used with good success depending upon the desired accuracy. These include
the manually operated pendulum machines described in the introduction as well as electromechan-
ical systems made from inexpensive equipment. Electromechanical systems make up the second
most versatile class of automated testing equipment.

B. E

LECTROMECHANICAL

S

YSTEMS

For the purpose of testing long bones, some researchers have devised cost-effective, purpose-built
torsional machines from existing motor-driven equipment (such as a lathe). It has been demonstrated
that these systems can reliably and consistently perform torque-to-failure tests when using angular
testing rates in the range of 3 to 12

/second.

24

One group of Finnish researchers used a particularly cost-effective approach.

25

They replaced
the original motor of an existing lathe with an overscaled asynchronous motor capable of producing
continuous torque up to 250 Nm. Generally, electric motors operate best at a nearly xed rotational
speed, and so the group was careful to select a motor rated for maximal torque at a desirable speed.
They chose an angular rate of 6

/s. (It is to be understood that no electric motor has a perfectly
linear response, and yet the results can be reliable if the average speed of the motor falls within
the range described above and has minimal deviation.) The lathe was equipped with a torque cell
at the xed end. The total machine inaccuracy, including the linearity error of the motor, strain
gauge sensitivity, repeatability error, sensor asymmetry, and output baseline instability, was deter-
mined to be less than 1.0%. The whole-method error for torsional testing of sheep tibiae was
reported to be 3.0%. This is sufciently accurate for the routine testing of long bones.
It should be noted here that very high loading rates can increase the apparent torque required
to fracture long bones since less time is allowed for microcracks to develop and to dissipate fracture
energy.

6

While it can be useful to test bones at a high rate to simulate transient loading conditions

in vivo

, it would be unwise to try to simulate slow fracture with an abbreviated test.

C. S

YSTEM

S

ELECTION

Custom-built equipment is naturally less exible but can be less expensive than a complex servo-
hydraulic system. The researcher must therefore determine the type of tests necessary to fulll both
the short- and long-term goals. When selecting a test machine, one should rst consider the
geometric needs and constraints: (1) What are the longest and shortest bones to be tested? (2) Will
there ever be any bulky augmentation, such as external xators, applied to the bone? Second,
determine precisely what loading modes are most important: (1) Will the testing be limited to
simple torsion? (2) Will biaxial loading be necessary in future studies? (3) Will there be a need for
transient loads, such as impact, to be applied at any point during any of the tests?

VI.

IN VITRO

TESTING METHODOLOGY
FOR SERVOHYDRAULIC SYSTEMS
A. A

DJUSTING



THE

C

ROSSHEAD

P

OSITION

The crosshead of the machine may be lowered or raised in order to bring the actuator and load cell
closer together or farther apart, respectively. It is the relative position of the crosshead that lends
versatility to servohydraulic machines in terms of the range of specimen sizes that can be tested.
The actual testing ranges are generally much smaller.
Lowering of the crosshead is performed by the action of gravity alone, and the hydraulic pump
should not be turned on during this operation. Lowering is achieved by releasing the hydraulic
clamp and lowering valve in succession. Alternatively, the crosshead may be raised only when the

0266/ ch13/frame Page 226 Thursday, October 21, 1999 4:34 PM

Torsional Testing of Bone

227

hydraulic pump is

on

. Again, the hydraulic clamp is released, followed by the raising valve. The
rate of crosshead motion is controlled by the degree to which the valves are opened and is stopped
by closing the valves and the clamp. It is important to close but not to overtighten the valves.

B. I

NTERNAL

C

ONTROLLER

C

ALIBRATION

Many recent control systems are self-calibrating, but older units will occasionally require that the
internal feedback and command signals be brought into alignment as a taring operation whenever
xtures and/or specimens are changed. Follow the manufacturers instructions carefully to determine
whether and how an adjustment should be made. For most tests, it is desirable for the load, torque,
and displacements to begin at zero.

C. E

XTERNAL

L

OAD

C

ALIBRATION

Another important calibration determines whether the output device (plotter, computer, etc.) reg-
isters a known load correctly. All load/torque cells will have a range over which they should respond
linearly to applied forces, and that range should not be exceeded. The use of three or more
different dead loads will allow the checking of the accuracy of the readout as well as the linearity
of the load/torque cell. If the readout is not accurate, then the output device must be adjusted.
If the readout is not linear to within a specied error, then the load/torque cell should be checked
before testing.

D. T

EST PREPARATION: THE RAMP TEST
A number of settings can be made before the machine is actually turned on. These are detailed as
follows. For a linear ramp test, the system controller function generator output must be set to
increase linearly with time. This is true whether the test is to be controlled by load or by displace-
ment. Make sure that the function selector of the function generator is switched to the ramp function.
Afterward, select the mode of the test as being positive or negative, depending on which direction
of actuator motion is desired.
Destructive tensile, compressive, torsional, and combined tests are often performed to determine
the load under which yielding or ultimate failure takes place. In order to conduct this type of test,
the test is often run in the displacement control mode. This type of control results in a continuous
change in the position of the actuator with respect to time, and the displacement rate is predeter-
mined before testing begins. The displacement rate is given as a measure of the displacement per
unit time and is determined by reverse-calculation from the available controller settings.
The full displacement range of many actuators is less than 4 in. axially and 45 rotationally.
If torsional testing is to be combined with tension or compression, the percentage of the total
axial displacement range to be used is sometimes manually selected. This is important when the
displacement range of the test to be conducted is much less than the overall range of the device.
The reason is that the feedback signal from the LVDT will be relatively weak at small fractions
of its total range, and the controller must therefore be set to amplify the signal. Generally, it is
important to match the axial displacement capabilities of the machine with the type of tests to
be done, but reliable data can be obtained using as little as 10% of the available displacement.
Finally, the plotter resolution can be adjusted for the desired output scale over the expected range
of the test.
E. GENERAL RAMP TESTING PROCEDURE
The hydraulic pump should be turned on and allowed to build pressure for a few minutes. While
the pump is building pressure, the specimen may be xed appropriately to the load/torque cell of
0266/ ch13/frame Page 227 Thursday, October 21, 1999 4:34 PM
228 Mechanical Testing of Bone and the BoneImplant Interface
the machine. Afterward, it is desirable to re-zero the load and displacement feedback signals at the
controller. Next, the actuator may be turned on. It may jump slightly upon being turned on;
however, the operator should also be certain that the function generator of the controller is not
running. Once the hydraulic actuator is on, the actuator piston can be lowered or raised until the
upper xture is in the appropriate position for mounting the remaining free end of the specimen.
If careful preparation has been made, the piston should not have to move very far.
Note: Safety glasses must be worn any time the actuator is on. This is of extreme importance
for the safety of all personnel working directly with or in the close vicinity of the machine.
Prescription glasses are acceptable for this purpose only if they have polycarbonate lenses with
side shields.
Once the actuator is turned on under active displacement control, any change in the command
signal from the controller will cause the actuator to move. Engaging the function generator of the
controller begins the test. Before operating the function generator, be sure that the load feedback
at the start of the test is at the desired level. If the test is to be performed after some preload, make
sure to apply it manually prior to starting the function generator. Likewise, if the test xtures have
generated any unwanted preload, it may be manually relaxed.
Note: In some cases it is appropriate to condition the specimen by applying a small fraction
of the expected load and then releasing it. This is especially important for any specimen that is
difcult to seat in the testing xtures.
Again, ensure that the ramp function of the function generator is selected and that the test will
occur in the appropriate direction (+ or ). Double check all controller settings, paying special
attention to the function generator ramp rate. Finally, prepare the recording device and begin the
test. Be ready to use any available emergency stop switches in case of difculty. At the completion
of the test, stop the function generator and wait for it to reset to its original position.
F. TEST PREPARATION: THE FATIGUE TEST
Fatigue testing is almost always performed with a sinusoidally varying command signal, and the
function generator of the controller should be set to operate in that capacity with the desired
frequency and load/strain ranges. Many controllers operate based on the half-cycle amplitude of
the function. Be sure to check the manufacturers instructions for calculating the function parameters
correctly. Fatigue tests can be conducted in either load or strain control mode. However, it may be
easier to start it in displacement control mode and then transfer to load or strain control after the
preloading conditions have been achieved.
Rotational and axial fatigue can both be performed in ve different capacities: positive-positive,
negative-negative, zero-positive, zero-negative, and positive-negative. As an example, take the case
of positive-positive torsional fatigue under load control. The specimen will experience positive
torsional loading at all times. A positive torsional preload, or mean level, is placed on the specimen,
and subsequent loading cycles are superimposed over that level. For example, the preload may be
equal to 5 Nm while the half-cycle amplitude is 1 Nm. The test will then cycle between 4 and 6
Nm. Negative-negative tests behave the same way, but in the reverse direction. Zero-positive and
zero-negative tests have no preload. Positive-negative tests may use any type of preload (or none
at all) as long as the torsional direction reverses at some point in each cycle.
Load-controlled tests will obtain their feedback from the load/torque cell, while strain-
controlled tests require the use of external gauges. The latter must be calibrated to work properly
with the system controller and its function generator. As with the ramp test, the displacement ranges
should also be set to provide the appropriate amplication of the LVDT/RVDT feedback signal if
the displacement is to be monitored.
0266/ ch13/frame Page 228 Thursday, October 21, 1999 4:34 PM
Torsional Testing of Bone 229
G. GENERAL FATIGUE TESTING PROCEDURE
Begin the test in displacement control mode, as with the ramp test. Preload the specimen to the
desired level, and adjust the mean level of the load/strain controller to match the preload. Once the
preload has been applied, the control modes may be switched. After changing control modes, the
function generator of the controller may be started. This begins the test. After the test specimen
has failed, the machine will continue to run unless axial and rotational displacement limits have
been applied. These should be available and apparent on the controller. Refer to the manufacturers
instructions for setting the limits. Once a limit has been exceeded, the actuator will turn itself off.
Cycle counters and oscilloscopes can be easily used to monitor the load/torque cell feedback
instantaneously. Alternatively, computer software acquisition will allow a more complete feedback
history to be recorded.
H. THE CREEP TEST
The creep test is the simplest test to operate from a machine perspective, yet it can be the most
difcult test to instrument. It will always be performed under load control. As with the fatigue test,
however, the servohydraulic machine may be difcult to start in the load control mode. Rather, the
test may be started under stroke control, and the operator may manually adjust the controller to
obtain a desired static preload. Once the preload is achieved, the machine may be switched to load
control, at which point the machine will maintain the set preload for as long as is desired. Over
time, the bone will begin to creep, and the resulting machine displacement or strain feedback signals
can be monitored. Since creep is a very slow and small-scale phenomenon, the displacement
feedback will most likely be of little use. Resistive strain gauges, on the other hand, can provide
more detailed information about what changes have taken place over a small region of the bone.
These gauges must be carefully placed over the region of interest with cyanoacrylate adhesives.
They are delicate instruments, and it is most useful to monitor the small feedback signals with a
computerized acquisition system.
Some authors have used a specially designed apparatus to study the torsional creep behavior
of compact bone.
26
Such equipment can be built inexpensively by applying the torque by means
of dead loads with pulleys and lever arms.
VII. SUMMARY
Torsional testing is a uniquely capable technique for examining the in vitro mechanical properties
of a wide variety of bones. Servohydraulic testing equipment can be a straightforward means to
obtain a large amount of torsional data using different loading modes. Electromechanical and
manually operated machines can be produced economically for purpose-specic torsional studies.
Gross approximations of torsional stress in diaphyseal bone may be calculated using simple linear
elastic mechanic theories; however, more accurate and detailed theoretical treatments may require
the use of computational mathematics. Empirical torsional testing is often the most efcient means
for the surgeon to obtain useful mechanical data. Torsional data may correlate well with many
common injury scenarios.
REFERENCES
1. Evans, F.G., Mechanical Properties of Bone, Charles C Thomas, Springeld, IL, 1973.
2. Evans, F.G., Stress and Strain in Bones, Charles C Thomas, Springeld, IL, 1957.
3. Hayes, W.C. and Carter, D.R., Biomechanics of bone, in Skeletal Research: An Experimental
Approach, Simmons, D.J. and Kunin, A.S., Eds., Academic Press, New York, 1979, 263.
0266/ ch13/frame Page 229 Thursday, October 21, 1999 4:34 PM
230 Mechanical Testing of Bone and the BoneImplant Interface
4. Saha, S., Dynamic strength of bone and its relevance, in Osteoarthromechanics, Ghista, D.N., Ed.,
McGraw-Hill, New York, 1982, 1.
5. Burstein, A.H. and Frankel, V.H., A standard test for laboratory animal bone, J. Biomech., 4, 155, 1971.
6. Sammarco, G.J., Burstein, A.H., Davis, W.L., and Frankel, V.H., The biomechanics of torsional
fractures: the effect of loading on ultimate properties, J. Biomech., 4, 113, 1971.
7. Medige, J., Mindell, E.R., and Doolittle, T., Remodeling of large, persistent bone defects, Clin.
Orthop., 69, 275, 1982.
8. ASTM, Standard test method for measuring the torsional properties of metallic bone screws, in 1997
Book of ASTM Standards, 13.01, ASTM, West Conshohocken, PA, 1997, 958.
9. Archdeacon, M.T., Davy, K.J., and Jepsen, K.J., Time dependent damage accumulation in bovine
cortical bone loaded in torsion, Orthop. Trans., 21, 731, 199798.
10. Bankston, A.B., Keating, M., and Saha, S., The biomechanical evaluation of intramedullary nails in
distal femoral shaft fractures, Clin. Orthop., 276, 272, 1992.
11. Hajek, P.D., Bicknell, H.R., Bronson, W.E., et al., Clinical and biomechanical analysis of one vs. two
distal screws in the treatment of femoral shaft fractures with locked intramedullary nails, J. Bone
Joint Surg., 75A, 519, 1993.
12. Albright, J.A., Thompson, T., and Saha, S., The principles of internal xation, in Orthopaedic Mechanics:
Procedures and Devices, Ghista, D.N. and Roaf, R., Eds., Academic Press, New York, 1978, 124.
13. Beer, F.P. and Johnston, E.R., Jr., Mechanics of Materials, McGraw-Hill, New York, 1992, 114.
14. Levenston, M.E., Beaupr, G.S., and Van der Meulen, M.C.H., Improved method for analysis of whole
bone torsion tests, J. Bone Miner. Res., 9, 1459, 1994.
15. Brooks, D.B., Burstein, A.H., and Frankel, V.H., The biomechanics of torsional fractures the stress
concentration effect of a screw hole, J. Bone Joint Surg., 52A, 507, 1970.
16. Clark, C.R., Morgan, C., Sonstegard, D.A., and Mathews, L.S., The effect of biopsy-hole shape and
size on bone strength, J. Bone Joint Surg., 59A, 213, 1977.
17. Currey J., Stress concentration in bone, Q. J. Microsc. Sci., 103, 111, 1962.
18. Nowinski, J.L., Effects of holes and perforations on the strength and stress distribution in bone
elements, in Osteoarthromechanics, Ghista, D.N., Ed., Hemisphere Publishing, Washington, D.C.,
1982, 180.
19. Saha, S., Stress concentration in bone: an experimental and theoretical investigation, in Biomedical
Engineering II: Recent Developments, Hall, C.W., Ed., Pergamon Press, New York, 1983, 367.
20. Ugural, A.C. and Fenster, S.K., Advanced Strength and Applied Elasticity, Prentice-Hall, Upper Saddle
River, NJ, 1995.
21. Won, H.Y., Lounci, S., Chen, D., et al., Inuence of bounding conditions on torsional structural testing
of canine tibial diaphysis, in Proceedings of the Combined Orthopaedic Research Societies Meeting,
September 2830, Hamamatsu, Japan, 1998.
22. Lipka, J.M., Saha, S., Keating, E.M., and Albright, J.A., The biomechanical analysis of a simulated
spondylolysis fracture and its contribution to lumbar spine rigidity, in Biomedical Engineering V:
Recent Developments, Saha, S., Ed., Pergamon Press, New York, 1986, 521.
23. Instron, Guide to Advanced Materials Testing, Instron Corporation, Canton, MA, 1997.
24. Strmberg, L. and Dlen, N., Experimental measurement of maximum torque capacity of long bones,
Acta Orthop. Scand., 47, 257, 1976.
25. Jms, T. and Jalovaara, P., A cost-effective, accurate machine for testing the torsional strength of
sheep long bones, Med. Eng. Phys., 18, 433, 1996.
26. Lakes, R. and Saha, S., Long term torsional creep in compact bone, J. Biomech. Eng., 102, 178, 1980.
FURTHER READING
Brnemark, R., hrnell, L.-O., Nilsson, P., and Thomsen, P., Biomechanical characterization of osseointegra-
tion during healing: an experimental in vivo study in the rat, Biomaterials, 18, 969, 1997.
Cervantes, C., Badison, J. B., Miller, G.J., and Casar, R.S., An in vitro biomechanical study of a multiplanar
circular external xator applied to equine third metacarpal bones, Vet. Surg., 25, 1, 1996.
Dueland, R.T., Berglund, L., Venderby, R., and Chao, E.Y., Structural properties of interlocking nails, canine
femora, and femur-interlocking nail constructs, Vet. Surg., 25, 386, 1996.
0266/ ch13/frame Page 230 Thursday, October 21, 1999 4:34 PM
Torsional Testing of Bone 231
Farfan, H.F., Cossette, J.W., Wells, R.V., and Kraus, H., The effects of torsion on the lumbar intervertebral
joints: the role of torsion in the production of disc degeneration, J. Bone Joint Surg., 52A, 468, 1970.
Hopper, S.A., Schneider, R.K., Ratzlaff, M.H., et al., Effect of pin hole size and number on in vitro bone
strength in the equine radius loaded in torsion, Am. J. Vet. Res., 59, 201, 1998.
Malkani, A.L., Voor, M.J., Fee, K.A., and Bates, C.S., Femoral component revision using impacted morsellised
cancellous graft: a biomechanical study of implant stability, J. Bone Joint Surg., 78B, 973, 1996.
Netz, P., Eriksson, K., and Strmberg, L., Material reaction of diaphyseal bone under torsion: an experimental
study on dogs, Acta Orthop. Scand., 51, 223, 1980.
Seltzer, K.L., Stover, S.M., Taylor, K.T., and Willits, N.H., The effect of hole diameter on the torsional
mechanical properties of the equine third metacarpal bone, Vet. Surg., 25, 371, 1996.
0266/ ch13/frame Page 231 Thursday, October 21, 1999 4:34 PM
0266/ ch13/frame Page 232 Thursday, October 21, 1999 4:34 PM

233

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

14

Indentation Testing of Bone

Brodie E. McKoy, Qian Kang, and Yuehuei H. An

CONTENTS

I. Introduction ..........................................................................................................................233
II. Major Indentation Tests .......................................................................................................233
A. Macroindentation Tests ..................................................................................................233
B. Microhardness Tests .......................................................................................................235
C. Nanoindentation Tests ....................................................................................................236
III. Macroindentation Tests Using a Cylindrical Indenter.........................................................236
References ......................................................................................................................................238

I. INTRODUCTION

Hardness of a solid material is dened as its resistance to penetration by another solid body.
Hardness or indentation tests measure hardness by driving an indenter with a specied geometry
into a sectional surface of the material. The tests can be categorized based on the geometry and/or
the size of the indenter employed. With different geometries there are Brinell, Rockwell, Vickers,
and Knoop indenters (Figure 14.1). Based on the size of the indenter, macroindentation, micro-
hardness, and nanoindentation are dened. All of these tests are used for biomechanical studies of
bone. Each of these methods assesses bone structures at different scales based on the sizes of the
specimens and the indenters.
Bone is a hierarchical structure.

1

In order to understand its mechanical properties as a whole,
one must understand the mechanical properties of its constituent parts. The various levels of
organization of bone can be tested by the different categories of indentation tests. Based on the
work by Rho et al.

1

and Hofer et al. (Chapter 8), the hierarchical structure of bone can be simplied
into four levels: (1) macrostructure: cancellous or cortical bone; (2) microstructure (1 to 500

m):
Haversian systems, osteons; (3) submicrostructure/nanostructure (200 nm to 1

m): billar collagen
and lamella; and (4) subnanostructure (less than 200 nm): molecular structure of constituent ele-
ments. Macroindentation, microhardness, and nanoindentation tests can be used to study the macro-
structure, microstructure, and submicrostructure of bone, respectively. Hardness values differ
according to the bone structure being indented. Since it is not possible to extrapolate the mechanical
properties of the bone from a single indentation test, it is ideal to perform different tests at various
levels of the bone structure.

II. MAJOR INDENTATION TESTS
A. M

ACROINDENTATION

T

ESTS

Macroindentation testing evaluates bone mechanical properties at the macrostructural level. Bone
at this level is considered either cortical or cancellous. These types of bone are distinguished most
readily by their amount of porosity.

2,3

Denitive differentiation is achieved only by microscopic

0266/ ch14/frame Page 233 Thursday, October 21, 1999 4:38 PM

234

Mechanical Testing of Bone and the BoneImplant Interface

evaluation of the tissue microstructure. The mechanical properties of bone at this level, including
hardness, are important for bone-related research as well as clinical purposes. Numerous investi-
gators have studied the mechanical properties of bone using this type of indentation testing. Several
different tests are of historical as well as functional interest.
The rst indentation test was reported in 1900 by Brinell and co-workers using a spherical
indenter (a steel ball).

4

The Brinell hardness value can be obtained by the following equations:
Brinell Hardness (BH) =

P/A

(N/mm

2

) (14.1)
where

P

is the applied load and

A

is indentation area measured after withdrawing the ball. Following
the pioneering work of Brinell, Lexer in 1929 evaluated the hardness of bone macroscopically.

5

Using a 3-mm steel ball and a large load, this investigator produced indentations in bone to assess
its hardness. He found no variation of hardness with illness or age. The design of these original
tests have persisted throughout the twentieth century with only minor modications.
The Brinell method has been used to measure the hardness of cancellous bone of human patella

4

and fracture callus material.

6,7

Equation 14.1 was used by Bjrkstrom and Goldie.

4

For the method
used by Markel et al.,

6,7

only a simple value of indentation (N/mm) was drawn from the test when the
indentation depth was maintained the same for all of the specimens. A modied Brinells hardness test
was reported by Aro et al.

8

using the following equation for calculation of hardness value:
Modied Brinell Hardness (MBH) = 2

P

/

dD

(N/mm

2

) (14.2)
where

P

is the applied load,

d

is the diameter of the indenter, and

D

is the depth of the indentation.
The Rockwell supercial hardness test is another indentation test which measures hardness of
materials at the macrostructural level.

9

Originally, this test was used to test metals with a rough
surface nish. Doppler modied this test to allow the testing of cancellous bone samples by
decreasing the major load.

10

A minor load (3 kg) is rst applied and the resulting indenter position
is used as the zero point. Next, the major load (5 kg) is applied and then removed. The permanent
indentation depth produced by the major load is measured.

11,12

Recently, several investigators have
used this technique to measure the hardness of bovine and human bone.

9,13

As in the Brinell hardness
test, a spherical indenter is used. However, the incremental depth is used as hardness number
without considering the indentation area.

FIGURE 14.1

The typical indentations of common indentation tests.

0266/ ch14/frame Page 234 Thursday, October 21, 1999 4:38 PM

Indentation Testing of Bone

235

The osteopenetrometer reported by Sneppen and Hvid et al.

14-16

also measures the hardness
of bone at the macrostructural level. The osteopenetrometer was originally developed as purely a
mechanical construction which measured the force of penetration of a needle into bone by means
of strain gauges, and depth of penetration by means of a differential transformer. The current
systems contain a hydraulic system between the recording unit and the needle.

17

The osteo-
penetrometer operates by recording the force necessary to penetrate the cancellous bone with a
2.5-mm-diameter pointed needle, with the shaft milled to 2.3 mm to avoid friction, as a function
of the depth of penetration. This test is an indentation test with a small indenter (a 90

conical
prole) which travels into the bone up to 6 to 12 mm. The evaluation program disregards the
initial 1.5 mm depth interval, which represents increasing contact with the measuring prole;
the average force per unit area of the measuring needle of the following consecutive 2 mm depth
intervals (2 to 5 mm) is used to represent the penetration strength of that interval. Agreement has
been found between the penetration test and conventional compression test on the same bony
structure. Hvid et al.

18

found correlation coefcients of close to 0.90 when ultimate stress data
was compared between the osteopenetrometer and conventional compression tests. The advantage
of this method is that it can be used during surgery. Intraoperative recording of bone strength may
be benecial in individualizing the choice of prosthetic design and postoperative rehabilitation
programs. The authors used this method for measuring the bone strength of human upper tibia
and calcaneous.
The most popular macroindentation test uses at-ended cylindrical indenters.

19-21

It has been
used for examining the mechanical properties of cancellous bone from different subjects. In this
indentation test, an indenter is driven into a sectional bone surface. Although the failure mechanisms
are more complicated and less clear than the conventional compression test, it is useful for
examining the mechanical properties of small cancellous bones of different species. Because of the
ease of specimen fabrication, the use of the indentation test has increased in recent years.

22

The
test is simpler than the compression test which uses cubed or cylindrical samples. Only a at sample
surface of minimal thickness is needed for indentation tests. Recent reports describe the use of the
indentation test for measuring the mechanical properties of rat,

23

rabbit,

22

canine,

24

and bovine

25

cancellous bone.

B. M

ICROHARDNESS

T

ESTS

The microhardness test is an indentation test using a specially designed testing device with a very
small indentor.

26

Microhardness indentations range from 20 to 150 mm in length. At this range,
this test evaluates materials at the microstructural level of bone such as individual osteons or
Haversian systems. In 1936, Lips and Sack

27

introduced the rst metallurgical microscope and with
this microhardness testers became available.
Vickers and Knoop are the two main types of microhardness indenters used today. The Vickers
has a pyramidal diamond indenter with an apical angle of 136

between the faces. Vickers can also
be used as a macrohardness test. The Knoop indenter has a rhombic-shaped pyramidal diamond,
with a longitudinal angle of 172

30

and a transverse angle of 130

. Both methods are mainly
used to examine the hardness of metals, plastics, composites,

27

or polymers. Several investigators
have reported on the application of microhardness tests to bone samples

5,28-30

including cortical
bone,

31,32

trabecular bone,

33

woven callus bone,

34,35

and bone adjacent to endosseous implants.

36,37

Using the Vickers and Knoop indenters, the hardness values have units of kg/mm

2

. The Vickers
microhardness number (VHN or HV) is calculated using the following equation:

38,39


VHN = HV = 1.8544

P

/

D

2

(14.3)

0266/ ch14/frame Page 235 Thursday, October 21, 1999 4:38 PM

236

Mechanical Testing of Bone and the BoneImplant Interface

where

P

is the applied load in kg and

D

is the mean of the length of

d

1

and

d

2

(the two diagonals
of the indentation) in mm. The Knoop microhardness number (HK) is calculated using the following
equation:

31,40

(14.4)
where

P

is the applied load in kg,

A

is the unrecovered projected area of indentation (mm

2

),

L

is
the measured length (mm) of the long diagonal, and

C

(0.07028) is a constant for the indenter
relating the projected area of the indentation to the square of the length of the long diagonal.

C. N

ANOINDENTATION

T

ESTS

Nanoindentation tests allow very small microstructural features to be studied. This method can
evaluate bone at the submicrostructural level which includes individual lamella of an osteon and
trabeculae.

41,42

The development of nanoindentation over the past several years has allowed the
evaluation of component properties (

in situ

). This technique allows the intrinsic properties of
individual submicrostructural components to be studied without the inuences of inhomogeneities
in the macro- and microstructure.
Recently, nanoindentation has been used to study individual trabeculae of human vertebrae and
tibia.

41,42

Elastic moduli and the hardness of individual lamella within an osteon have been evaluated.
Investigators have used this technique to show differences in hardness in both transverse and
longitudinal directions of bone. Even though this test is currently performed predominately with
dry bone, nanoindentation has greatly increased knowledge of bone mechanical properties.

III. MACROINDENTATION TESTS USING A CYLINDRICAL INDENTER

Numerous investigators have used this method to assess the hardness of bone. In humans, the
indentation stiffness of the patella,

5

distal femur,

20

tibial plateau,

21

distal tibia,

43

and head and neck
of the femur

44

has been evaluated. Several animal models have been studied with the at-ended
indenter. This method of indentation is accepted by many researchers.
The indentation test using a at-ended cylindrical indenter has been applied to different animal
bones in the authors laboratory, including epiphysometaphyseal cancellous bones of rats, rabbits,
dogs, and goats.

22,23,45

The methods of performing the indentation tests in these various animals are
similar. In the authors laboratory, the selected bones (rat, rabbit, dogs, and cows) are rough-cut
using a band saw. They are then ground on a rotating wheel grinder to a proscribed level in the
cancellous bone at the epiphysis or metaphysis to create a surface for testing.
After the rst surface is created, two different methods can be used to position the specimen
on the testing platform. In the rst method, a parallel cut is made next to the rst ground surface
to create a second surface to be set against the specimen-holding platform. The second method
involves potting the specimen. Instead of performing the second cut, the specimen can be potted
in dental stone or plaster of paris for positioning on the platform. Both of these methods have been
used successfully in the authors laboratory. The latter is especially suitable for small specimens

.

In the authors laboratory, a mechanical testing system (MTS, Minneapolis, MN) is used for
the indentation test. It is operated in displacement control, and is calibrated using an extensometer.
After the specimen to be tested is prepared, the platform holding the specimen is leveled to ensure
that the loading is perpendicular to the specimen surface to be tested. A cylindrical stainless-steel
indenter with a at-ended surface, ranging from 1.3 to 5.0 mm in diameter, is used. After the
specimen is positioned on the platform and the indenter adjusted close to the specimen surface,
the indenter is driven into the bone at a constant slow rate (1 mm/min in the authors laboratory).
HK
P
A
---
P
CL
2
--------- = =

0266/ ch14/frame Page 236 Thursday, October 21, 1999 4:38 PM

Indentation Testing of Bone

237

The loading is stopped manually when the loaddisplacement curve passes the ultimate load (the
highest point of the curve).
Using this indentation test, the ultimate load, stiffness, ultimate strength (ultimate stress), and
elastic modulus of the bone can be obtained directly or calculated from the loaddeformation curve.
The ultimate indentation strength is calculated using the following equation:

= 4

P

/

d

2

(14.5)
where

P

is the ultimate indentation load and d is the diameter of the indenter. By utilizing the
formula developed by Timoshenko and Goodier

46

and validated recently by Sumner et al.,

24

the
local modulus of elasticity (

E

) for each test site is calculated with the following formula:

E

=

S

(1

2

)/

d

(14.6)
where

S

is the indentation stiffness (N/mm) and d (mm) is the diameter of the indenter. As used
by Sumner et al.

24

and Aitken et al.,

43

the Poissons ratio (

)

is assumed to be 0.2 according to Vasu
et al.

47

Recent reports describe the use of the indentation test for measuring the mechanical properties
of rat,

23

rabbit,

22

canine, and bovine

48

cancellous bones.



Differences in mechanical properties were
found between epiphyseal cancellous bones at different locations. The reason for this phenomenon
is the functional difference between the different locations.

49

For example, the humeral head bears
less load compared with the femoral head, so the ultimate strength of the cancellous bones of
humeral head is less than that of the femoral head. Correlation has been obtained between inden-
tation depth (at 50 N load) and ultimate strength (

R



= 0.937,

p

<0.05), meaning that with the
increase of ultimate strength the indentation depth or deformation decreased proportionally.
Sumner et al.

24

have veried that the data obtained from the indentation tests correlate well
with that from conventional compressive tests. Kang et al.

25

showed that ultimate load, stiffness,
and ultimate strength, measured by the indentation test were higher than those measured by the
compression test. This group also found a signicant correlation between compression testing and
indentation testing with correlation values of 0.823 to 0.952. Although the ultimate strength and
elastic modulus of different cancellous bones from different subjects are not the same, they generally
fall into a certain range, like that of compression tests. According to the data pooled from the
literature

21,24,25,43,50

and the data generated from this study, the ultimate strength of cancellous bones
obtained by the indentation test ranges from 38 to 71 MPa. The wide range of these values is not
surprising and is due to different subjects and different locations. Even with the conventional
compressive test, the range of elastic modulus was much larger, ranging from several to 3000
MPa.

48

Trabecular bone modulus can vary 100-fold from one location to another even within the
same metaphysis.
Indentation testing with a at-ended cylindrical indenter has been used often due to its several
advantages. Indentation testing may be more representative of the

in vivo

condition, which can be
considered as a constrained compression test. The mechanical properties (mostly obtained from
compression testing) of a cube or cylinder bone sample separated from the bone such as the femur
or tibia are not the same as that when the cube or cylinder were in the bone tissue. The preparation
of bone required with compression testing introduces errors. With indentation testing, the bone can
be tested as a whole and numerous areas of the same bone may be tested. The indentation test is
a simple procedure. Only a at surface of the sample is needed for testing, and it is less invasive
than the conventional compression test. The indentation test makes mechanical testing on smaller
bones feasible. There have been no reports to the authors knowledge of attempts to study rat bones
using compression tests, possibly because the bone size is too small. Because the structure of
cancellous bone is anisotropic and heterogeneous, which is more apparent for smaller bones,
indentation testing may be more appropriate. Also, fewer variables are involved with indentation

0266/ ch14/frame Page 237 Thursday, October 21, 1999 4:38 PM

238

Mechanical Testing of Bone and the BoneImplant Interface

testing compared with conventional compression testing. When the conditions of the test machine
are the same, indentation tests only require the specimen deformation and the surface area of the
indenter and the load to be known. When using a compression test with cylindrical samples, the
length (which cannot be easily controlled), end surface area, and the deformation have to be known.
Finally, macroindentation using a cylindrical indenter requires only a conventional material testing
machine available in most material testing laboratories.

REFERENCES

1. Rho, J.Y., Kuhn-Spearing, L., and Zioupos, P., Mechanical properties and the hierarchical structure
of bone,

Med. Eng. Phys.,

20, 92, 1998.
2. Carter, D.R. and Hayes, W.C., The compressive behavior of bone as a two-phase porous structure,
J. Bone Joint Surg., 59A, 954, 1977.
3. Gibson, L.J., The mechanical behavior of cancellous bone, J. Biomech., 18, 317, 1985.
4. Bjrkstrom, S. and Goldie, I.F., Hardness of the subchondral bone of the patella in the normal state,
in chondromalacia, and in osteoarthrosis, Acta Orthop. Scand., 53, 451, 1982.
5. Weaver, J.K., The microscopic hardness of bone, J. Bone Joint Surg., 48A, 273, 1966.
6. Markel, M.D., Wikenheiser, M.A., and Chao, E.Y., A study of fracture callus material properties:
relationship to the torsional strength of bone, J. Orthop. Res., 8, 843, 1990.
7. Markel, M.D., Wikenheiser, M.A., and Chao, E.Y., Formation of bone in tibial defects in a canine
model. Histomorphometric and biomechanical studies, J. Bone Joint Surg., 73A, 914, 1991.
8. Aro, H.T., Wippermann, B.W., Hodgson, S.F., et al., Prediction of properties of fracture callus by
measurement of mineral density using micro-bone densitometry, J. Bone Joint Surg., 71A, 1020, 1989.
9. Houde, J., Marchetti, M., Duquette, J., et al., Correlation of bone mineral density and femoral neck
hardness in bovine and human samples, Calcif. Tissue Int., 57, 201, 1995.
10. Doppler, R.A., Bone Mechanical Property Determination through Supercial Rockwell Hardness
Testing, M.S. thesis, Worcester Polytechnic Institute, Worcester, MA, 1993.
11. Standards, A., Standard test methods for Rockwell hardness and Rockwell supercial hardness of
metallic materials, ASTM Standards Specications, E18, ASTM, Philadelphia, 1989, 176.
12. Evans, F.G. and Lebow, M., Regional differences in some of the physical properties of the human
femur, J. Appl. Physiol., 3, 563, 1951.
13. Duquette, J., Lin, J., Hoffman, A., et al., Correlations among bone mineral density, broadband ultra-
sound attenuation, mechanical indentation testing, and bone orientation in bovine femoral neck
samples, Calcif. Tissue Int., 60, 181, 1997.
14. Sneppen, O., Christensen, P., Larsen, H., and Vang, P.S., Mechanical testing of trabecular bone in
knee replacement, Int. Orthop., 5, 251, 1981.
15. Hvid, I., Trabecular bone strength at the knee, Clin. Orthop., 227, 210, 1988.
16. Jensen, N.C., Madsen, L.P., and Linde, F., Topographical distribution of trabecular bone strength in
the human os calcanei, J. Biomech., 24, 49, 1991.
17. Hvid, I., Andersen, K., and Olesen, S., Cancellous bone strength measurements with the osteo-
penetrometer, Eng. Med., 13, 73, 1984.
18. Hvid, I., Jensen, J., and Nielsen, S., Bone strength measurements at the proximal tibia. Penetration
tests and epiphyseal compressive strength, Int. Orthop., 10, 271, 1986.
19. Josechak, R.G., Finlay, J.B., Bourne, R.B., and Rorabeck, C.H., Cancellous bone support for patellar
resurfacing, Clin. Orthop., 192, 1987.
20. Nakabayashi, Y., Wevers, H.W., Cooke, T.D., and Grifn, M., Bone strength and histomorphometry
of the distal femur, J. Arthroplasty, 9, 307, 1994.
21. Behrens, J.C., Walker, P.S., and Shoji, H., Variations in strength and structure of cancellous bone at
the knee, J. Biomech., 7, 201, 1974.
22. An, Y.H., Kang, Q., and Friedman, R.J., Mechanical symmetry of rabbit bones studied by bending
and indentation testing, Am. J. Vet. Res., 57, 1786, 1996.
23. An, Y.H., Zhang, J.H., Kang, Q., and Friedman, R.J., Mechanical properties of rat epiphyseal cancel-
lous bones studied by indentation test, J. Mater. Sci. Mater. Med., 8, 493, 1997.
0266/ ch14/frame Page 238 Thursday, October 21, 1999 4:38 PM
Indentation Testing of Bone 239
24. Sumner, D.R., Willke, T.L., Berzins, A., and Turner, T.M., Distribution of Youngs modulus in the
cancellous bone of the proximal canine tibia, J. Biomech., 27, 1095, 1994.
25. Kang, Q., An, Y.H., and Friedman, R.J., The mechanical properties and bone densities of canine
cancellous bones, J. Mater. Sci. Mater. Med., 9, 1998.
26. Wassell, R.W., McCabe, J.F., and Walls, A.W., Subsurface deformation associated with hardness
measurements of composites, Dent. Mater., 8, 218, 1992.
27. Lips, E.M. and Sack, J., A hardness tester for microscopical objects, Nature, 138, 328, 1936.
28. Ramrakhiani, M., Pal, D., and Murty, T.S., Micro-indentation hardness studies on human bones, Acta
Anat., 103, 358, 1979.
29. Ramrakhiani, M., Pal, D., and Datta, S.C., Effect of heating on the hardness of human bone, Acta
Anat., 108, 316, 1980.
30. Amprino, R., Investigations on some physical properties of bone tissue, Acta Anat., 34, 161, 1958.
31. Riches, P.E., Everitt, N.M., Heggie, A.R., and McNally, D.S., Microhardness anisotropy of lamellar
bone, J. Biomech., 30, 1059, 1997.
32. Ziv, V., Wagner, H.D., and Weiner, S., Microstructuremicrohardness relations in parallel-bered and
lamellar bone, Bone, 18, 417, 1996.
33. Lenart, G., Toth, I., and Piner, J., Experiments on the hardness of bone by Vickers microhardness
measurements, Acta Biochem. Biophys. Acad. Sci. Hung., 3, 1968.
34. Huja, S.S., Qian, H., Roberts, W.E., and Katona, T.R., Effects of callus and bonding on strains in
bone surrounding an implant under bending, Int. J. Oral Maxillofac. Implants, 13, 630, 1998.
35. Blackburn, J., Hodgskinson, R., Currey, J.D., and Mason, J.E., Mechanical properties of microcallus
in human cancellous bone, J. Orthop. Res., 10, 237, 1992.
36. Stea, S., Savarino, L., Toni, A., et al., Microradiographic and histochemical evaluation of mineraliza-
tion inhibition at the bone-alumina interface, Biomaterials, 13, 664, 1992.
37. Stea, S., Visentin, M., Savarino, L., et al., Microhardness of bone at the interface with ceramic-coated
metal implants, J. Biomed. Mater. Res., 29, 695, 1995.
38. Simske, S.J. and Sachdeva, R., Cranial bone apposition and ingrowth in a porous nickel-titanium
implant, J. Biomed. Mater. Res., 29, 527, 1995.
39. Evans, G.P., Behiri, J.C., Currey, J.D., and Boneld, W., Microhardness and Youngs modulus in
cortical bone exhibiting a wide range of mineral voume fractions, and in a bone analogue, J. Mater.
Sci. Mater. Med., 1, 38, 1990.
40. Huja, S.S., Katona, T.R., Moore, B.K., and Roberts, W.E., Microhardness and anisotropy of the vital
osseous interface and endosseous implant supporting bone, J. Orthop. Res., 16, 54, 1998.
41. Rho, J.Y., Tsui, T.Y., and Pharr, G.M., Elastic properties of human cortical and trabecular lamellar
bone measured by nanoindentation, Biomaterials, 18, 1325, 1997.
42. Lee, F.Y., Rho, J.Y., Harten, R., Jr., et al., Micromechanical properties of epiphyseal trabecular bone
and primary spongiosa around the physis: an in situ nanoindentation study, J. Pediatr. Orthop., 18,
582, 1998.
43. Aitken, G.K., Bourne, R.B., Finlay, J.B., et al., Indentation stiffness of the cancellous bone in the
distal human tibia, Clin. Orthop., 264, 1985.
44. Hardinge, M.G., Determination of the strength of the cancellous bone in the head and neck of the
femur, Surg. Gynecol. Obstetr., 89, 439, 1949.
45. An, Y.H., Kang, Q., and Friedman, R.J., et al., Do mechanical properties of epiphyseal cancellous
bones vary? J. Invest. Surg., 10, 221, 1997.
46. Timoshenko, S.P. and Goodier, J.N. Axisymmetric stress and deformation in a solid of revolution.
Load distributed over a part of a boundary of an innite solid, in Theory of Elasticity, Clark, B.J. and
Maisel, J.W., Eds., McGraw-Hill, New York, 1970, 380.
47. Vasu, R., Carter, D.R., and Harris, W.H., Stress distributions in the acetabular region I. Before and
after total joint replacement, J. Biomech., 15, 155, 1982.
48. Kang, Q., An, Y.H., and Friedman, R.J., Effects of multiple freezingthawing cycles on ultimate
indentation load and stiffness of bovine cancellous bone, Am. J. Vet. Res., 58, 1171, 1997.
49. Goldstein, S.A., The mechanical properties of trabecular bone: dependence on anatomic location and
function, J. Biomech., 20, 1055, 1987.
50. Saitoh, S., Nakatsuchi, Y., Latta, L., and Milne, E., An absence of structural changes in the proximal
femur with osteoporosis, Skeletal Radiol., 22, 425, 1993.
0266/ ch14/frame Page 239 Thursday, October 21, 1999 4:38 PM
0266/ ch14/frame Page 240 Thursday, October 21, 1999 4:38 PM

241

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

15

Penetration Testing of Bone
Using an Osteopenetrometer

Ivan Hvid and Frank Linde

CONTENTS

I. Introduction ..........................................................................................................................241
II. Guidelines for Penetration Testing of Trabecular Bone......................................................241
A. The Osteopenetrometer: Designed for

in Vivo

Testing .................................................241
B. Laboratory Penetration Testing......................................................................................245
References ......................................................................................................................................246

I. INTRODUCTION

Penetration testing was originally developed to characterize the quality of soil. A measuring prole
is driven from the surface into the structure or material under investigation, while the force necessary
to advance the probe is recorded as a function of the depth of penetration.
The motivation to develop and use this particular method of mechanical testing was a desire
to characterize trabecular bone mechanically at the knee joint during total knee arthroplasty. This
work was initiated in the 1970s, when mechanical loosening was still a signicant problem in the
emerging semi- and nonconstrained total knee designs, the long-term function of which were
dependent upon the mechanical quality of tibial and femoral condylar trabecular bone.

1

As it turned
out, the problem of mechanical loosening was minimized by a combination of improved implant
design and development of instrumentation to assure proper implant alignment. Even with these
renements, however, the strength of condylar trabecular bone remains an important factor for the
risk of mechanical loosening.

2

Reduced physical activity, low body weight, and rheumatoid arthritis
were shown to be related to reduced trabecular bone strength around the knee.

2,3

Trabecular bone
strength was universally (normal knees, osteoarthrosis, rheumatoid arthritis) shown to diminish
quite signicantly with the distance from the subchondral bone plate,

2-4

indicating that bone
resection during total knee replacement should be minimized.
In the laboratory, the method was used to obtain closely spaced bone strength measurements
to describe accurately the variation of bone strength across the surface of joints.

3-6

The ndings
correlated qualitatively to ndings of gait analysis,

7,8

tending to justify the simplifying assumptions
inherent in the model solutions in these kinds of studies.

II. GUIDELINES FOR PENETRATION TESTING OF TRABECULAR BONE
A. T

HE

O

STEOPENETROMETER

: D

ESIGNED



FOR



IN

V

IVO

T

ESTING

Since the idea was to develop a tool to measure bone strength intraoperatively, the device had to
be quite small to facilitate handheld operation. It should be able to endure repeated exposure to
high pressure and high temperature for sterilization purposes.

0266/ ch15/frame Page 241 Thursday, October 21, 1999 4:41 PM

242

Mechanical Testing of Bone and the BoneImplant Interface

A number of mechanical restraints were necessary. Since counterpressure must be exerted by
the operator (Figure 15.1), the maximal measurable force could not exceed 500 to 600 N. Accord-
ingly, the penetration needle had to be relatively thin. This was also desirable from the point of
view that minimal damage should be inicted on the bone structure. A measuring prole diameter
of 2.5 mm (4.9 mm

2

projected cross section, Figure 15.2) was nally chosen after some experi-
mentation. The speed of penetration was to be kept constant since bone is a viscoelastic substance,
so that mechanical properties are strain rate dependent. In fact, penetration strength was found to
be penetration rate (speed) dependent in laboratory studies.

9

The penetration speed used routinely
in the authors clinical and laboratory studies was 1 mm s

1

.
The rst prototype in clinical use was a mechanical device.

1

It was soon abandoned because
it tolerated sterilization procedures poorly. A hydraulically powered penetrometer, using sterile
demineralized water as hydraulic uid, was then developed (Figure 15.3).

9

This osteopenetrometer
needed hydraulic uid rells between every few sterilizations, but was otherwise stable and rela-
tively easy to use. It was indirectly powered by a computer-controlled electromotor, and the same
custom-built computer was used to calibrate the device automatically and store a series of 12
measuring cycles.
The measurement obtained was the force of penetration as a function of the depth of penetration
(Figure 15.4). Penetration strength (MPa) was reported as the force of penetration (N) averaged
over an arbitrary depth interval, and normalized for the projected cross-sectional area of the needle
(in mm

2

). Penetration tests as hardness tests do not reect any well-dened property of bone.
They do, however, correlate quite closely to some important mechanical properties of bone.

5,9

A
systematic series of tests was done relating the results of penetration tests to those of compression
tests on unconned machined specimens in a regular materials testing machine. Such a comparison
presents a crucial inherent problem in that both types of tests are destructive. However, using
two different approaches

5,9

yielding very similar results, it was possible to establish empirical
relationships to yield strength, ultimate strength, Youngs modulus, and energy absorption as
determined from regular materials testing. These observations were later conrmed using bone

FIGURE 15.1

Sketch of operation of the handheld osteopenetrometer for clinical, intraoperative use. The
operator must resist the reaction force used to drive the measuring probe into the trabecular bone.

0266/ ch15/frame Page 242 Thursday, October 21, 1999 4:41 PM

Penetration Testing of Bone Using an Osteopenetrometer

243

density measurements obtained by quantitative computed X-ray tomography scanning as a link
between the two modes of mechanical testing.

10

Obviously, the measurement obtained is dependent upon the design of the needle used to
penetrate the trabecular bone. A systematic approach to nd a suitable needle design was under-
taken.

9

Blunt needles and needles with a stem diameter identical to that of the base of the measuring
prole resulted in empirical relationships to ultimate strength and ultimate strain energy that were

FIGURE 15.2

Sketch of preferred needle design.

FIGURE 15.3

Design diagram of osteopenetrometer developed for clinical measurement.

0266/ ch15/frame Page 243 Thursday, October 21, 1999 4:41 PM

244

Mechanical Testing of Bone and the BoneImplant Interface

functions of the depth of penetration (higher penetration strength values relative to ultimate prop-
erties were obtained from deep portions of the penetration path than from supercial portions).
This was interpreted to be a result of impaction of bone beneath the blunt needle and frictional
forces along the stem of the needle, respectively. The needle design nally chosen for future
measurements (see Figure 15.2) featured a 90

conical measuring prole with a base diameter of
2.5 mm, and the needle shaft milled to 2.3 mm, since this needle produced indistinguishable
regression lines from supercial and deep aspects of the bone.
In the laboratory, the osteopenetrometer could be mounted on a heavy metal frame. In this way,
it was more or less comparable with a regular materials testing machine with the crosshead equipped
with a needle. Early on, it was shown that the 2.5-mm needle resulted only in very localized damage
to the trabecular bone structure, leaving trabeculae outside the actual measuring pathway of the
needle virtually unharmed.

1

Therefore, the method would be suitable to obtain fairly closely spaced
measurements which would be of considerable interest in establishing a deeper insight in the
patterns of load transmission across joints, particularly the rather complex knee joint. It was found
that 5 mm between the center of each measurement was sufcient to avoid measurable deterioration
of the mechanical properties of trabecular bone located within the connes of four penetration tests
in a squared pattern,

5

and cross talk between penetration tests using this minimal distance between
individual measurements could not be demonstrated.

3

Following transverse resection of the joint surface just below the subchondral bone plates of
the condyles, measurements were taken across the proximal tibial epiphysis using penetration tests
5 mm apart in a squared pattern. This revealed a smooth topographical pattern of strength with a
higher medial than lateral peak strength localized centrally and posteriorly, respectively. Toward
the periphery of the condyles, bone strength diminished progressively to very low values near the
edges, indirectly demonstrating the effectiveness of the menisci in distributing the forces at the
normal knee joint and minimizing strength/stiffness gradients in joint cartilage and subchondral
bone.

3

Viewed in concert with the marked reduction of strength away from the joint surface (about
50% over 10 mm at the stronger central parts of the condyles) certain proximal tibial fracture
patterns can be explained.

FIGURE 15.4

Typical measurement curve. An example of data extraction from the curve is shown. Initial
1.5 mm of penetration disregarded this is needed to engage the leading prole of the needle fully with
bone. The curve is then further subdivided in arbitrary intervals, depending upon the depth resolution desired.
Finally, force is averaged over each depth interval, and the resulting force normalized with respect to the
projected cross-sectional area of the measuring prole. When force is measured in N and the cross-sectional
area expressed in mm

2

, the strength expression will be in MPa.

0266/ ch15/frame Page 244 Thursday, October 21, 1999 4:41 PM

Penetration Testing of Bone Using an Osteopenetrometer

245

One more problem related to this unique needle design needs to be addressed. As with other
mechanical tests not performed on individual trabeculae, the strength of the trabecular structure
rather than of the bone as such is measured. Therefore, the bone is assumed to be a continuum. To
meet this assumption, a linear distance of ten intercept lengths (bonemarrow interfaces as seen
microscopically), corresponding to ve intertrabecular widths (approximately 5 mm in the human
proximal tibia) must be engaged by the measuring device according to Harrigan et al.

11

Penetration
tests on human trabecular bone using a 2.5-mm-diameter needle obviously violates the continuum
assumption. However, when measuring human trabecular bone around the knee with this particular
needle dimension, the estimated error is moderate, less than

10%.

12

This error could be huge if
measurements were done in other parts of the human skeleton: consider the aged human vertebra
where trabeculae may be several millimeters apart.
The osteopenetrometer as described above is not commercially available. Modern imaging
techniques offer the opportunity of assaying bone strength by measuring bone density very accu-
rately.

10,13-15

With present-day technology, these methods of assaying bone strength are probably as
accurate as

in vivo

osteopenetrometer testing, and, in addition, they are noninvasive. The doses of
ionizing radiation necessary to obtain these measurements are decreasing, while the structural
resolution is increasing, leading to even better noninvasive assays of the mechanical properties of
bone. Accordingly, although the use of the osteopenetrometer has produced some interesting results
in the past, it must be regarded as obsolete for clinical use.

B. L

ABORATORY

P

ENETRATION

T

ESTING

As mentioned above, penetration tests can be performed on a regular materials testing machine. The
needle must be mounted on the crosshead of the machine. The needle can then be advanced into
trabecular bone samples at a constant crosshead speed. The load cell is most conveniently mounted
between the crosshead and the needle. The bony structure being studied must be rigidly supported.
Measurements should usually be obtained by advancing the needle at right angles to the bone
surface. This will limit the usefulness of the method or call for elaborate adjustable xation devices,
e.g., in measuring the bone strength distribution under curved surfaces such as the femoral condyles
and the femoral head. Failure to engage the surface of the bone specimen at right angles, especially
when the bone is relatively strong, may lead to bending of the needle and invalidation of the
measurement. When trying to measure extremely strong bone, e.g., the sclerotic subchondral bone
in severe osteoarthrosis, the needle may buckle, again invalidating the measurement. To deal with
such problems, one might consider the use of a needle of heavier design. However, it should be
remembered that the penetration test has not been validated measuring very dense bone, and
therefore the results obtained may not be readily interpretable. Other methods should be considered.
Indentation testing, for instance, can be used to obtain a dense pattern of measurement, and a quasi-
continuous measurement deep to the surface could be established by removing bone successively
from the surface doing repeated measurements as needed.
The use of larger needles or penetrators has been reported,

16

but has not been validated. In
the authors opinion, large needle penetration tests are not useful for laboratory use, since better
methods exist.
The authors believe that the penetration test has been very useful in describing topographical
variations of trabecular bone strength in subchondral bone regions, i.e., close to joints, and has
provided valuable (indirect) information on the mode of load transmission through various joints.
For such purposes this method can be recommended, and is still being used.

17

The method does
provide quasi-three-dimensional information on the mechanical strength of trabecular bone not
readily obtainable by other methods. While the standard method of mechanically testing trabe-
cular bone as a structure is likely to remain compression tests on machined specimens large
enough to satisfy the continuum assumption, it is worth remembering that this method is blind
to variations within the test volume. One signicant accomplishment of the penetration test was

0266/ ch15/frame Page 245 Thursday, October 21, 1999 4:41 PM

246

Mechanical Testing of Bone and the BoneImplant Interface

to document the 50% decrease of bone strength within the rst few millimeters of subchondral
bone at the proximal tibia, which was not revealed by previous studies using standard testing of
machined specimens.

REFERENCES

1. Sneppen, O., Christensen, P., Larsen, H., and Vang, P-S., Mechanical testing of trabecular bone in
knee replacement,

Int. Orthop.,

5, 251, 1981.
2. Hvid, I., Trabecular bone strength at the knee,

Clin. Orthop.

, 227, 210, 1988.
3. Hvid, I. and Hansen S.L., Trabecular bone strength patterns at the proximal tibial epiphysis,

J. Orthop.
Res.,

3, 464, 1985.
4. Hvid, I. and Hansen, S.L., Subchondral bone strength in arthrosis. Cadaver studies of tibial condyles,

Acta Orthop. Scand.,

57, 47, 1986.
5. Hvid, I., Cancellous bone at the knee: a comparison of two methods of strength measurement,

Arch.
Orthop. Trauma Surg.,

104, 211, 1985.
6. Hvid, I., Rasmussen, O., Jensen, N.C., and Nielsen, S., Trabecular bone strength proles at the ankle
joint,

Clin. Orthop.,

199, 306, 1985.
7. Morrison, J.B., The mechanics of the knee joint in relation to normal walking,

J. Biomech.,

3, 51, 1970.
8. Harrington, I.J., Static and dynamic loading patterns in knee joints with deformities,

J. Bone Joint
Surg.,

65A, 247, 1983.
9. Hvid, I., Andersen, K., and Olesen, S., Cancellous bone strength measurements with the osteopene-
trometer,

Eng. Med.,

13, 73, 1984.
10. Bentzen, S.M., Hvid, I., and Jrgensen, J., Mechanical strength of tibial trabecular bone evaluated by
X-ray computed tomography,

J. Biomech.,

20, 743, 1987.
11. Harrigan, T.P., Jasty, M., Mann, R.W., and Harris, W.H., Limitations of the continuum assumption in
cancellous bone,

J. Biomech.,

21, 269, 1988.
12. Hvid, I., Mechanical Strength of Trabecular Bone at the Knee, thesis, University of Aarhus, Laege-
foreningens Forlag, Aarhus, Denmark, 1988, 15.
13. Hvid, I., Bentzen, S.M., and Jrgensen, J., Bone density at the proximal tibia after total knee
replacement. A two-year follow-up study using quantitative X-ray computed tomography,

Acta Orthop.
Scand.,

59, 567, 1988.
14. Petersen, M.M., Nielsen, P.T., Lauritzen, J.B., and Lund, B., Changes in bone mineral density of the
proximal tibia after uncemented total knee arthoplasty. A 3-year follow-up of 25 knees,

Acta Orthop.
Scand.,

66, 513, 1995.
15. Petersen, M.M., Jensen, N.C., Gehrchen, P.M., and Nielsen, P.K., The relation between trabecular
bone strength and bone mineral density assessed by dual photon and dual energy X-ray absorptiometry
in the proximal tibia,

Calcif. Tissue Int.

, 59, 311, 1996.
16. Yuzuki, O., Study on the mechanical strength and the inner structure of the knee,

J. Jpn. Orthop.
Assoc.,

52, 537, 1978.
17. Mller-Gerbl, M., Dalstra, M., Ding, M., et al., Distribution of strength and mineralization in the
subchondral bone plate of human tibial heads,

Trans. Eur. Orthop. Res. Soc.

, 8, 94, 1998.

0266/ ch15/frame Page 246 Thursday, October 21, 1999 4:41 PM

247

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

16

Microhardness Testing of Bone

Sarandeep S. Huja, Thomas R. Katona, and W. Eugene Roberts

CONTENTS

I. Introduction ..........................................................................................................................247
II. Microhardness Indenters ......................................................................................................248
III. Factors Affecting Microhardness of Bone...........................................................................250
A. Specimen Preparation.....................................................................................................250
1. Storage Temperature..................................................................................................250
2. Fixation......................................................................................................................250
3. Embedding in Resin and Inltration.........................................................................251
4. Degreasing/Defatting.................................................................................................251
5. Drying vs. Testing Wet..............................................................................................251
B. Anatomical Variables......................................................................................................251
1. Type of Bone Cortical, Trabecular, and Woven Bone Callus .............................251
2. Anatomical Location of Bone and Site Within Bone ..............................................251
3. Age, Sex ....................................................................................................................251
4. Disease States............................................................................................................252
C. Testing Variables ............................................................................................................252
1. Specimen Thickness..................................................................................................252
2. Indentation Spacing...................................................................................................252
3. Load Dependency......................................................................................................252
4. Duration of Dwell .....................................................................................................252
5. Dimension of Indentation and Magnication of the Optical System......................252
IV. Guidelines for Microhardness Testing.................................................................................253
V. Measuring Microhardness Anisotropy.................................................................................253
VI. Conclusion............................................................................................................................254
Acknowledgments ..........................................................................................................................255
References ......................................................................................................................................255

I. INTRODUCTION

Hardness is a reection of the resistance to penetration or indentation. Two types of hardness tests
are recognized based on the magnitude of the indentation load. Macrohardness testing involves
using loads greater than 1 kg, while microhardness tests use loads less than 1 kg.

1

Macrohardness
testing was used in the past to gain information about the mechanical properties of bone.

2

However,
microhardness testing is routine and preferred for testing small bone specimens. Others classify
hardness testing into microhardness, low-load hardness, and standard hardness tests.

3

Recently,
nanoindentation techniques have also been applied to bone.

4

For perspective, the size of an osteon
is about 200 to 300

m, and each lamella in bone is 3 to 7

m wide.

5

Microhardness indentations
range from 20 to 150

m in length, while nanoindentations can be of the order of 1

m. Bone is

0266/ ch16/frame Page 247 Thursday, October 21, 1999 4:46 PM

248

Mechanical Testing of Bone and the BoneImplant Interface

a hierarchal structure

6

and it is possible that hardness values will differ with the particular bone
structure (e.g., collagen ber, lamella, osteon) being indented.
The literature reports a number of studies detailing the experimental basis of microhardness
testing,

1,3,7,8

the application of microhardness testing to bone,

9,10

and dental enamel.

11

Some studies
attempt to relate microhardness to mechanical properties.

12-15

A review of the above-cited articles
should provide the reader with a comprehensive understanding of the use and limitations of
microhardness testing of biological tissues.
The intent of this chapter is to aid researchers in developing a suitable protocol for conducting
bone microhardness tests. Factors that could inuence results will be discussed. In addition, the
reader will be referred throughout the chapter to articles that provide information about specic
important factors.
The prototype for microhardness testers became available in 1936 with the introduction of a
metallurgic microscope by Lips and Sack.

16

(See Lysaght

7

for a description of the evolution of hardness
and microhardness testers.) At the outset, microhardness testing enabled metallurgists and engineers
to test brittle materials and microconstituents of metals and alloys which could not be examined by
macrohardness tests. Bone microhardness testing was rst reported by Carlstrom.

17

He measured
Vickers microhardness on a human femur section that contained three groups of Haversian systems
with differences in mineralization of about 10%. Carlstrom concluded that there was a close relation-
ship between mineralization, as measured by microradiographic density, and microhardness. His
Vickers microhardness numbers ranged from 29.5 to 38.3 kg/mm

2

. Similar observations about min-
eralization were made by Weaver.

9

While mineralization is an important determinant of microhardness,
Amprino

10

suggests that other factors such as collagen ber orientation may also inuence micro-
hardness. Subsequent to these early studies, the microhardnesses of cortical bone,

18,19

trabecular
bone,

9,20

woven bone callus,

21,22

bone adjacent to endosseous implants,

21,23,24

bone with an Haversian
system,

9

and dental enamel

11

have been measured. In addition, microhardness has been used to test
in-plane anisotropy of cortical bone and parallel-bered bone in different animal species.

10,18,19,21

In
many of these studies, there would not have been other means of obtaining information about the
mechanical properties of the bone structures. Constraints imposed by specimen size or shape would
have precluded conventional mechanical testing (e.g., preparing dumbbell/coupon-shaped specimens).
Therein lies the uniqueness and applicability of microhardness testing.

II. MICROHARDNESS INDENTERS

Vickers and Knoop are the two types of microhardness indenters that are commonly used. In 1920,
the Vickers indenter was developed in England and was initially used for macrohardness testing.

25

In 1939, the Knoop indenter was introduced by the National Bureau of Standards in the United
States.

26

The instrument of choice has partly been determined by the part of the world in which
the studies have been conducted. Since most, but not all,

9

of the early bone microhardness studies
were conducted in Europe,

10,17,20

the literature reports mostly Vickers microhardness results.
The Vickers indenter is a square-shaped diamond (Figure 16.1a) with face angles of 136

. The
Knoop indenter (Figure 16.1b) is a rhombic-shaped pyramidal diamond, with a longitudinal angle
of 172

30

and a transverse angle of 130

. For both indenters, the microhardness number is the
ratio of applied load to indented area, and they have the units of kg/mm

2

.
The Vickers microhardness number (HV) is given by
2

P

sin(

/2)/

d

2

(16.1)
where

P

is the applied load in kg,

d

is the mean of the length of

d

1

and

d

2

in mm, and

is 136

.
The Knoop microhardness number (HK) is given by

P/A

=

P/C

(

d

1

)

2

(16.2)

0266/ ch16/frame Page 248 Thursday, October 21, 1999 4:46 PM

Microhardness Testing of Bone

249
FIGURE 16.1

Schematics of (a) Vickers and (b) Knoop indenter geometries and their indentations in the
cross-hatched surfaces. The angles between the faces of the indenters are indicated. The

d

1

/

d

2

ratios are 1 : 1
and 1 : 7 for the Vickers and Knoop indenters, respectively. With the same load, the Vickers indenter penetrates
approximately two times deeper than the Knoop indenter.

0266/ ch16/frame Page 249 Thursday, October 21, 1999 4:46 PM

250

Mechanical Testing of Bone and the BoneImplant Interface

where

P

is the applied load in kg, and

A

is the unrecovered projected area (mm

2

). The unrecovered
area is calculated from the length of the long diagonal

d

1

and a constant

C

(0.07028) which relates
the length of the long diagonal to the projected area.
The depths of the Vickers and Knoop indentations are

th and

th of the respective lengths
of their diagonals. For a given material, the Knoop diagonal is approximately three times as long
as the Vickers, while the Vickers indenter penetrates nearly twice as deep as the Knoop indenter.
Microhardness and macrohardness indentations can be made using the same Vickers indenter. This
is not the case with the Knoop indenter, which is used primarily for microhardness tests.
The Knoop indenter is, in essence, a specialized case of the Vickers indenter. Knoop and his
colleagues attempted to develop a universal indentation scale.

8

Knoop wanted a

d

1

/

d

2

ratio between
5 and 10. After experimentation they chose a ratio of 7. A ratio of 5 may have allowed elastic
recovery, and a ratio of 10 decreased sensitivity due to lack of ability to locate the ends of the
indentations. With the Knoop indenter, the maximum strain (deformation) is along its width, the
least is at the tips of its length.

8

In making comparisons between the two indenters, it has been suggested that the Knoop indenter
cuts, while the Vickers deforms.

1,10,27

It has also been stated that the Knoop indenter allows elastic
recovery of the impression to occur primarily along the shorter diagonal. This is because the stresses
are distributed in such a manner that only the minor axis dimensions are subject to recovery.

27

Since the long diagonal dimension is affected minimally by elastic recovery, it is used to calculate
HK.

7,8

With lighter loads on the Knoop indenter, inated microhardness numbers may result for
two reasons. Lighter loads allow for greater elastic recovery of the long diagonal,

7

and, second,
the recovery makes the actual ends of the diagonal more difcult to locate under the microscope.
Bckle

3

disputes that the change in indentation dimensions is related to the elastic properties of
the material. He states that elastic recovery with a (Vickers) indenter is negligible along the diagonals,
but it can occur along the depth of the indentation. He ascribes changes in the diagonal length to
secondary causes such as formation of pile-ups, the reecting power of the specimen, the numerical
aperture of the objective. Pileups are plastic ow of the material in the vicinity of the indentation.
Thus, Bckle is of the opinion that operator and machine error are responsible for dimensional changes.

III. FACTORS AFFECTING MICROHARDNESS OF BONE

It is important to understand the effects of the following factors prior to conducting bone micro-
hardness tests.

A. S

PECIMEN

P

REPARATION

1. Storage Temperature

It is recommended that bone specimens be stored in saline-soaked gauze at 20

C prior to mechan-
ical testing.

28

Effects of storage on microscopic hardness have been examined.

9

It was found that
rapid freezing and storage of bone at 20

C in a sealed container had no appreciable effect on
hardness, provided the surface is prepared within 1 h of testing.

2. Fixation

The effect xation has on microhardness is unclear. It is thought that xation causes cross-linking
of collagen. Amprino,

10

citing Rssle

29

and Lexer,

30

suggested that brief (i.e., 24 h vs. weeks)
xation in formalin has no effect on hardness. However, Weaver

9

found that storage in 10% neutral
formalin for 24 h caused a 20% increase in microhardness. Blackburn et al.

22

demonstrated that
xation in 4% formaldehyde for 24 h at room temperature had no effect on microhardness. As this
is a variable which may affect microhardness, it is prudent to eliminate it by using unxed specimens
when possible.

0266/ ch16/frame Page 250 Thursday, October 21, 1999 4:46 PM

Microhardness Testing of Bone

251

3. Embedding in Resin and Inltration

The literature does not make a clear distinction between

embedding

and

inltration

of bone specimens
in resin. While testing trabecular bone specimens or callus, it is benecial to provide support by
embedding the specimen. This, however, is different from inltration, in which the objective is to allow
the methyl methacrylate to occupy the smallest spaces (e.g., lacunae) within the substance of bone.
Embedding (in reality, inltrating) bone in methyl methacrylate has been reported to increase
microhardness by 30 to 40%.

9

When unxed and noninltrated specimens were used in studies of
different animal species

18,21

similar cortical bone Knoop values were obtained, suggesting the
relative importance of these two variables. However, when specimens are inltrated, the Knoop
microhardness number does not increase.

31

It is therefore preferable to use noninltrated specimens
for microhardness testing.

4. Degreasing/Defatting

Defatting allows for sharper indent outlines. Defatting with a chloroform/methanol mixture had no
effect on microhardness.

22

5. Drying vs. Testing Wet

There is a general consensus that drying bone specimens prior to testing will increase the microhardness
values substantially. The chief reason to dry specimens is to provide clear indentation outlines, thereby
allowing accurate measurements. Drying causes a signicant increase in microhardness in primary and
secondary bone the higher the drying temperature (38

C, 60

C, 120

C for a few hours), the greater
the increase.

10

This is in agreement with Blackburn et al.

22

who dried their specimens at 40

C prior to
testing. This resulted in an increase in microhardness by approximately 20%. These ndings are also
in agreement with Evans and Lebow

2

who found that air drying (105

C) increased elastic modulus by
17.6% and macrohardness by 54.3%. While some investigators dry the specimens prior to testing, it
would be preferable to use specimens that have not been subjected to heat. In addition, however, it has
been found that microhardness of rehydrated bone is very close to that of wet fresh bone.

10

B. A

NATOMICAL

V

ARIABLES

1. Type of Bone Cortical, Trabecular, and Woven Bone Callus

When using a Vickers indenter, trabecular bone has lower microhardness than cortical bone.

9,20

Blackburn et al.

22

demonstrated that the microhardness of nodular-type microcallus and trabecular
bone were similar. However, the mean microhardness values for the former are lower, but with
more variability. Huja et al.

21

examined woven bone calluses that stabilize endosseous implants at
12 weeks postimplantation in dog midfemoral diaphysis. They found that periosteal and endosteal
callus had signicantly lower HK values than the layer of rapidly remodeling bone adjacent to the
implant. Both the callus and implant adjacent bone, in turn, had lower HK than the cortical bone.

2. Anatomical Location of Bone and Site Within Bone

Ramarakhiani et al.

32

found HV to be different for different embalmed dry bones of the human
skeleton. Riches et al.

18

and Weaver

9

found that HV was lower in the epiphyseal and metaphyseal
regions than in the diaphysis. HV was also uniform throughout the diaphysis.

3. Age, Sex

Two studies

9,32

report no sex specic differences in microhardness of cortical bone. However,
Weaver

9

was able to demonstrate an increase in microhardness with age.

0266/ ch16/frame Page 251 Thursday, October 21, 1999 4:46 PM

252

Mechanical Testing of Bone and the BoneImplant Interface

4. Disease States

Amprino

10

examined the microhardness of bone specimens in patients with Pagets disease, renal
rickets, and osteogenesis imperfecta. Others found no detectable change in the microhardness of
osteoporotic bone.

9

In addition, hardness of subchondral bone in chondromalacia and osteoarthrosis
has been examined.

33

C. TESTING VARIABLES
The precision, bias, repeatability, and reproducibility of Vickers and Knoop microhardness
indentations for metallic specimens of varying hardnesses have been determined.
34
Similarly,
factors that determine the accuracy of indentation microhardness tests have been discussed.
35
With instruments available today, the rate of load application, duration of dwell, and removal of
the load from the specimen are usually automated. Other factors that should be taken into
consideration are discussed below.
1. Specimen Thickness
Generally, the thickness of the specimen should be at least ten times the depth of the anticipated
impression.
7
2. Indentation Spacing
When making multiple tests of the same specimen, it is important not to place additional indenta-
tions within the strain eld of previous indentations. It has been suggested that additional Vickers
indentations be made no closer than 1.5 to 2 the length of the original indentation.
7,10
3. Load Dependency
By using a number of different materials, it has been shown that HK increases as the load decreases.
In contrast, HV shows the opposite trend.
1
These effects are only seen at relatively low (below
approximately 50 g) loads for both indenters. Amprino
10
addressed the question of whether similar
HV values can be obtained from a given bone structure under different loads (15, 25, 50, 100, 200,
300 g). He found that with increasing load there was a slight tendency to a lowering of the
microhardness value. Ramarakhiani et al.
32
measured HV using 5, 10, 20, 30, 50, 75, 100, 130,
150 g loads and found that HV increased with loads to 50 g, and then became independent of load
in subsequent measurements. In another study,
21
loads of 25 and 50 g did not reveal statistically
signicant (p = 0.74) differences in HK values of cortical bone. Other studies have also rigorously
addressed this issue.
8,34
For cortical bone, a 50-g load seems appropriate.
4. Duration of Dwell
Mott
25
stated that the reason for applying a load for a standard time is to ensure static conditions;
i.e., there should be no ow/creep of a material leading to increase in the size of an indentation.
This becomes more important for viscoelastic materials. It has been stated that consistent results
were obtained on bone specimens by leaving the indenter in place for longer than 3 to 4 s.
9
A 10
to 15 s dwell on bone specimens seems to be adequate. Additional information is available on the
effects of this factor.
3,8
5. Dimension of Indentation and Magnication of the Optical System
When making pilot indentations in a test material, a balance between the size of the indentation
and the microscope magnication must be struck. As a general rule, the larger the indentation, the
0266/ ch16/frame Page 252 Thursday, October 21, 1999 4:46 PM
Microhardness Testing of Bone 253
greater the accuracy.
1
In bone specimens, the dimension of the structure may limit the size of the
indentation. If an indentation is too large for a given magnication, the load must be decreased.
For example, a 25-g load with a Knoop indenter on cortical bone may result in a small (e.g., 70 m)
indentation, necessitating a higher magnication (e.g., 400). A 50-g load and 200 magnication
could then be chosen. To allow for comparisons between studies, it is imperative to specify the load.
IV. GUIDELINES FOR MICROHARDNESS TESTING
The following guidelines for testing bone specimens are recommended. Specimens, frozen during
storage, must be thawed. Water lubrication during cutting is advisable, particularly if an
implantbone block is being sectioned. It is advisable to test bone specimens which have not been
xed in formalin or inltrated with resin. Generally, the bone specimens must be mounted in a
resin block to provide support for polishing and testing.
A good surface polish is essential. Procedures similar to those used to obtain a metallurgic
nish are recommended. A method that has been used successfully to polish cortical bone specimens
is described below.
21
A rotary wheel (Vari/Pol, VP-150, Leco, St. Joseph, MI) set at 150 rpm with
successive 400-, 800-, and 1200-grit SiC papers and water lubrication is used. The 400-grit paper
is used for a brief interval, with light pressure, as it can leave deep gouges in the bone. Frequently
changing the orientation of the specimen ensures a good nish without scratches. The bone specimen
is examined under a stage microscope to conrm a smooth surface. The bone is polished further
using a napless Metcloth (Buehler polishing cloth, #40-7158, Lake Bluff, IL) with diluted water
slurry of 1 m aluminum oxide paste (Buehler Micropolish C alpha Alumina, Lake Bluff, IL). The
specimen is sonicated in cleaning solution (Ultramet Sonic, Buehler, Lake Bluff, IL). Blotting with
paper towels for 10 min to obtain a dry surface will remove excess moisture.
The specimen is then mounted in a leveling vise of the microhardness testing machine. The
vise allows the indenter to penetrate the surface at right angles. The loading mechanism should be
free of vibration and any effects of inertia. An accurate optical system for measuring the indentations
should be available with most indenters. Under the microscope, the operator selects the area where
the indentation is to be made. Automatic test cycle features ensure standardization of the test cycle.
The rate of descent of the indenter, duration of dwell, and return of the indenter to its original
position is uniform from one test to the next. Changing the direction of specimen illumination can
make a difference in the clarity of the indentations and should be standardized to obtain the best
view. With the Knoop and Vickers indenter, a difference of more than 20% between the two mirror
halves of the indentation around the axes is usually an indication that the indenter is not entering
perpendicular to the specimen surface. Asymmetric readings should be discarded.
A weeping problem has been observed with bone specimens that have been blotted dry
without heat.
21,22
This is due to the exudate from the specimen. Weeping is recognized by the
observation of indistinct indentation margins or the inability to focus on the surface. Reblotting
the specimen for about 3 min with paper towels adequately resolves this problem. In most testers,
the HV or HK is calculated automatically after the load and length of the indentation are provided.
Routine maintenance and calibration of the equipment by a qualied technician is important
for consistent results. Intraobserver variability should be evaluated prior to starting the research
project. Close to 5% error on bone and stainless steel indentations has been reported.
21,34
V. MEASURING MICROHARDNESS ANISOTROPY
Measurement of anisotropy has been conducted with conventional mechanical tests,
36
acoustic
techniques,
37
and microhardness indenters.
19
With the latter, anisotropy has been investigated using
two methods. First, multiple indentations can be made on a single surface by orientating the indentor
in specic directions
18,21
(e.g., relative to the lamellae.). Second, a cubic bone specimen
19
can be
0266/ ch16/frame Page 253 Thursday, October 21, 1999 4:46 PM
254 Mechanical Testing of Bone and the BoneImplant Interface
machined to coincide with the orthogonal directions of bone, and indentations made on the different
surfaces can be compared.
Amprino
10
demonstrated that Vickers microhardness measurements on collagen bers were
independent of the Vickers pyramid face alignment. In his experiment, indentations were made so
that collagen bundles from avian long bone (ethanol xed and air dried) were cut either (1) with
one diagonal of the Vickers parallel to the collagen bundle and the other perpendicular to it, or
(2) both diagonals crossing at 45 to the long axis of the collagen bers. Amprino ascribed his
results partly to the characteristic shape of the Vickers pyramid; which does not act as a wedge-
shaped tip, and therefore pressures prevail over splitting components. This suggests that the Vickers
indenter can only provide one hardness measurement per surface. Similarly, Weaver
9
made Vickers
indentations along the transverse and longitudinal axes of the same Haversian system in a bular
cortex. He did not nd microhardness differences in the two directions of his inltrated sections.
In contrast, the Knoop indentor has been successfully used to detect in-plane anisotropy.
18,21
Riches
et al.
18
were able to detect a microhardness anisotropy ratio of approximately 1.3 in the transverse
vs. the longitudinal direction on specimens of rat tibiae. Similarly, Huja et al.,
21
using a Knoop
indenter, demonstrated microhardness anisotropy in lamellar bone near the endocortical surfaces
of the femur. Riches et al.
18
suggest that the Knoop indenter is more sensitive than a Vickers indenter
in detecting anisotropy of bone.
Amprino
10
also made measurements on two perpendicular surfaces. In these sections the
collagen bers were cut either perpendicular or along their long axes. Various animal species were
tested and the HV microhardness reading was higher by approximately 20 to 25% on the surface
where the collagen bers were cut perpendicular to their long axes. Ziv et al.
19
used a Vickers
indenter on cubic specimens and were able to demonstrate anisotropy in three orthogonal directions
on parallel-bered bone. Although these authors acknowledged that the building blocks of lamellar
bone are anisotropic, they found lamellar bone to be isotropic. This was probably due to the large
size of their indentations.
VI. CONCLUSION
Microhardness testing has been used to measure the resistance to indentation of various bone types.
It is generally acknowledged that microhardness results are not easily interpreted
14
and that hardness
represents a combination of several material properties.
25
What these microhardness measurements
mean remains unclear. However, relative comparisons between two microhardness measurements
using the same methodology are accurate.
However, there are at least three studies
11,12,15
which have found strong correlations between
microhardness and variables such as volume fraction of mineral, Youngs modulus, yield stress,
and calcium content. Currey and Brear,
12
referring to the relationship between hardness and elastic
modulus, state that hardness will not, of course, actually determine the other mechanical properties;
these mechanical relationships doubtless indicate a common relationship with another variable,
almost certainly mineral content of the tissues. Equations to estimate elastic modulus and yield
stress from microhardness are available.
12,13
These equations have been derived for a range of
mineralization levels. However, these formulae can only be used if identical methodology to that
used in the original study is utilized. No such equations are available for the Knoop indenter.
These limitations do not detract from the usefulness of microhardness testing. Instead, infor-
mation gained from microhardness tests could be corroborated by other methods to study the
mechanical properties of bone and could serve to develop newer methods of investigating micro-
structural properties of bone. One hopes the methods will lead to a better understanding of the
complicated structure of bone and the mechanical properties of various subunits of bone, including
the level of interaction between the collagen and hydroxyapatite crystals. Microhardness tests can
be valuable tools in bone biomechanics, once the limitations of the tests are fully understood.
0266/ ch16/frame Page 254 Thursday, October 21, 1999 4:46 PM
Microhardness Testing of Bone 255
ACKNOWLEDGMENTS
The authors are extremely grateful to Dr. David B. Burr and Dr. B. Keith Moore for their invaluable
assistance. This work was supported by National Institutes of Dental Research Grant PHS R55
DE09822.
REFERENCES
1. Fee, A.R., Segabache, R., and Tobolski, E.L., Hardness testing, in Metals Handbook, Vol. 8, American
Society for Metals Mechanical Testing, Metals Park, OH, 1985, 87.
2. Evans, F.G. and Lebow, M., Regional differences in some of the physical properties of the human
femur, J. Appl. Physiol., 3, 563, 1951.
3. Bckle, I., Progress in micro-indentation hardness testing, Metall. Rev., 4, 49, 1959.
4. Rho, J-Y., Tsui, T.Y., and Pharr, G.M., Elastic properties of human cortical and trabecular lamellar
bone measured by nanoindentation, Biomaterials, 18, 1325, 1997.
5. Martin, R.B. and Burr, D.B., Structure, Function and Adaptation of Compact Bone, Raven Press, New
York, 1989, 186.
6. Katz, J.L., Hierarchical modeling of compact Haversian bone as a ber reinforced material, in
Advances in Bioengineering, Mates, R.E and Smith, C.R., Eds., ASME, New York, 1976, 18.
7. Lysaght, V.E., Indentation Hardness Testing, Reinhold, New York, 1949, 234.
8. Knoop, P., Peters, C.G., and Emerson, W.B., A sensitive pyramidal-diamond tool for indentation
measurements, J. Res. Natl. Bur. Stand., 23, 39, 1939.
9. Weaver, J.K., The microscopic hardness of bone, J. Bone Joint Surg., 48A, 273, 1966.
10. Amprino, R., Investigations on some physical properties of bone tissue, Acta Anat., 34, 161, 1958.
11. Featherstone, J.D., ten Cate, J.M., Shariati, M., and Arends, J., Comparison of articial caries-like
lesions by quantitative microradiography and microhardness proles, Caries Res., 17, 385, 1983.
12. Currey, J.D. and Brear, K., Hardness, Youngs modulus and yield stress in mammalian mineralized
tissue, J. Mater. Sci. Mater. Med., 1, 14, 1990.
13. Evans, G.P., Behiri, J.C., Currey, J.D., and Boneld, W., Microhardness and Youngs modulus in
cortical bone exhibiting a wide range of mineral volume fractions, and in a bone analogue, J. Mater.
Sci. Mater. Med., 1, 38, 1990.
14. Hodgskinson, R., Currey, J.D., and Evans, F.G., Hardness, an indicator of the mechanical competence
of cancellous bone, J. Orthop. Res., 7, 754, 1989.
15. Evans, G.P., Behiri, J.C., and Boneld, W., Microhardness, Youngs modulus and mineral content in
bone and bone analogue, in Implant Materials in Biofunction; Advances in Biomaterials, Vol. 8, de
Putter, C., de Lange, G.L., de Groot, K., and Lee, A.J., Eds., Elsevier Science, Amsterdam, 1988, 311.
16. Lips, E.M. and Sack, J., A hardness tester for microscopical objects, Nature, 138, 328, 1936.
17. Carlstrom, D., Micro-hardness measurements on single Haversian systems in bone, Acta Anat., 15,
171, 1952.
18. Riches, P.E., Everitt, N.M., Heggie, A.R., and McNally, D.S., Microhardness anisotropy of lamellar
bone, J. Biomech., 30, 1059, 1997.
19. Ziv, V., Wagner, H.D., and Weiner, S., Microstructure-microhardness relations in parallel-bered and
lamellar bone, Bone, 18, 417, 1996.
20. Lenart, G., Toth, I., and Pinter, J., Experiments on the hardness of bone by Vickers microhardness
measurements, Acta Biochim. Biophys. Acad. Sci. Hung., 3, 205, 1968.
21. Huja, S.S., Katona, T.R., Moore, B.K., and Roberts, W.E., Microhardness and anisotropy of the vital
osseous interface and endosseous implant supporting bone, J. Orthop. Res., 16, 54, 1998.
22. Blackburn, J., Hodgskinson, R., Currey, J.D., and Mason, J.E., Mechanical properties of microcallus
in human cancellous bone, J. Orthop. Res., 10, 237, 1992.
23. Stea, S., Tarabusi, C., Ciapetti, G., et al., Microhardness evaluation of the bone growing into porous
implants, J. Mater. Sci. Mater. Med., 3, 252, 1992.
24. Stea, S., Visentin, M., Savarino, L., et al., Microhardness of bone at the interface with ceramic-coated
metal implants, J. Biomed. Mater. Res., 29, 695, 1995.
25. Mott, B.W., Measuring hardness, New Sci., 9, 103, 1964.
0266/ ch16/frame Page 255 Thursday, October 21, 1999 4:46 PM
256 Mechanical Testing of Bone and the BoneImplant Interface
26. Lysaght, V.E., Indentation hardness testing, Reinhold Publishing, New York, 1949, 188.
27. Anusavice, K.J., Phillips Science of Dental Materials, 10th ed., W.B. Saunders, Philadelphia, 1996,
69.
28. Turner, C.H. and Burr, D.B., Basic biomechanical measurement of bone: a tutorial, Bone, 14, 595,
1993.
29. Rssle, R., Untersuchungen ber Knochenhrte, Beitr. Pathol. Anat., 77, 174, 1927.
30. Lexer, E.W., Untersuchungen ber die Knochenhrte des Humerus, Z. KonstLehre, 14, 227, 1929.
31. Chachra, D., Turner, C.H., Dunipace, A.J., and Grynpas, M.D., The effect of uoride treatment on
bone mineral in rabbits, Calcif. Tissue Int., 64, 345, 1999.
32. Ramarakhiani, M., Pal, D., and Murty, T.S., Micro-indentation hardness studies on human bones, Acta
Anat., 103, 358, 1979.
33. Bjrkstrm, J. and Goldie, I.F., Hardness of the subchondral bone of the patella in the normal state,
in chondromalacia, and in osteoarthrosis, Acta Orthop. Scand., 55, 453, 1982.
34. Vander Voort, G.F., Results of an ASTM E-4 Round-Robin on the precision and bias of measurements
of microindentation hardness impressions, in Factors That Affect the Precision of Mechanical Tests,
Papirno, R. and Weiss, H. C., Eds., ASTM, Baltimore, 1989, 3.
35. Tobolski, E.L., Factors that affect the accuracy of indentation hardness tests, in Factors That Affect
the Precision of Mechanical Tests, Papirno, R. and Weiss, H.C., Eds., ASTM, Baltimore, 1989, 46.
36. Reilly, D.T., Burstein, A.H., and Frankel, V.H., The elastic modulus for bone, J. Biomech., 7, 271, 1974.
37. Ashman, R.B., Cowin, S.C., Van Buskirk, W.C. and Rice, J.C., A continuous wave technique for the
measurement of the elastic properties of cortical bone, J. Biomech., 17, 349, 1984.
0266/ ch16/frame Page 256 Thursday, October 21, 1999 4:46 PM

257

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

17

Nanoindentation Testing
of Bone

Jae-Young Rho and George M. Pharr

CONTENTS

I. Introduction ..........................................................................................................................257
II. Sample Preparation ..............................................................................................................258
III. Indentation Procedures.........................................................................................................259
IV. Analyzing Nanoindentation LoadDisplacement Data .......................................................260
V. Determination of Load Frame Compliance and Diamond Area Function .........................264
VI. Limitations............................................................................................................................265
VII. Results for Measurements in Bone......................................................................................266
Acknowledgments ..........................................................................................................................268
References ......................................................................................................................................268

I. INTRODUCTION

Nanoindentation, developed over the last 10 years, is now used widely in the materials science
community for probing the mechanical properties of thin lms, small volumes, and small micro-
structural features.

1,2

The availability of depth-sensing nanoindentation instruments with capabilities
for measuring displacements on the order of nanometers makes it possible to study the mechanical
properties of thin lms and other nely structured materials where small volumes need to be
probed.

3

From analysis of nanoindentation loaddisplacement data, it is possible to derive values
of the elastic modulus,

1,2

hardness,

1,2

and properties associated with time-dependent deformation
such as strain rate sensitivity and the stress exponent for creep.

4

Since nanoindentation can be used
to probe a surface and map its properties on a spatially resolved basis, often with a resolution of
better than 1 mm, it can be used to measure the properties of small microstructural features such
as individual osteons and trabeculae.

5-8

The mechanical properties of bone have been extensively studied at the macrostructural and
microstructural level. Conventional mechanical testing has a lower limit of a few hundreds of
microns at best. Microindentation measures a dimension of a few tenths to hundreds of microns,

9,10

while nanoindentation can study on a 1

m or smaller length scale. Since many important micro-
structural components of bone have dimensions of only a few microns or less, the nanoindentation
technique can be used to investigate the mechanical properties of osteonal, interstitial, and trabecular
lamellar bone at the largely unexplored micron and submicron level. As an example, the nanoin-
dentation method has been used to examine variations in the individual lamellar properties within
osteons, as a function of distance from the osteonal center. To the best of the authors knowledge,
in only one previous study has mechanical property measurement of individual lamellae within
osteons been attempted,

11

and it met with little success due to technical difculties. The nano-
indentation method also potentially allows one to examine the properties in different directions and
therefore explore elastic anisotropy, even in very small specimens such as individual trabeculae.

0266/ ch17/frame Page 257 Thursday, October 21, 1999 4:53 PM

258

Mechanical Testing of Bone and the BoneImplant Interface

At present, microstructurally based theories of bone behavior are difcult to test due to a lack of
microscopic mechanical property data.

12

Characterization of lamellar bone properties by nano-
indentation methods would thus aid in further theoretical developments of the mechanisms of
fracture in bone.
Because of experimental complications associated with testing in liquid environments and
keeping specimens wet during testing, most nanoindentation work has focused on dried bone.

5

It
is well documented, however, that the mechanical properties of bone show notable changes after
dehydration.

13-16

In general, drying increases the Youngs modulus of bone, decreases its toughness,
and reduces the strain to fracture. An important question thus arises of the degree to which
nanoindentation mechanical property measurements are affected by drying. To this end, a series of
tests was recently conducted on wet and dry bovine femur using a special testing xture in which
specimens could be tested fully immersed in a liquid.

17

Drying was found to increase the elastic
modulus by 9.7% for interstitial lamellae and 15.4% for osteons. The hardness was found to increase
by 12.2% for interstitial lamellae and 17.6% for osteons.
The nanoindentation techniques elucidated in this chapter can be used to study mechanical
properties at boneimplant interfaces, pathological metabolic bone diseases, the properties of woven
bone and calcied cartilage in fracture callus, the properties of the growth plate, and biological
aspects of implant loosening. Such results would ll a gap in the present knowledge by providing
a better understanding of the mechanical function of bone at the microstructural and submicro-
structural level.

II. SAMPLE PREPARATION

Nanoindentation specimens are normally in the form of polished, metallographic samples mounted
in plastic mounts or glued to metal disks. The specimen holder is a metal plate into which holes
have been drilled (Figure 17.1). Sample mounts are tted into the holes and are held in place with
set screws.
To test dried bone, bone sections can be dehydrated in a series of alcohol baths and embedded
without vacuum in epoxy resin at room temperature. In the case of trabecular bone, a low-
viscosity epoxy is preferred to inltrate and provide support for the porous network. The
embedded samples are metallographically polished to produce the smooth surfaces needed for
nanoindentation testing. After being ground using silicon carbide abrasive papers of decreasing
grit size (600, 800, and 1200 grit) under deionized water, the specimens are polished on micro-
cloths with successively ner grades of alumina powder, the nest being 0.05 mm grit. The last
polishing step is on plain microcloth under deionized water, and the specimens are cleaned
ultrasonically to remove surface debris.

FIGURE 17.1

Schematic diagram of a typical specimen holder.

0266/ ch17/frame Page 258 Thursday, October 21, 1999 4:53 PM

Nanoindentation Testing of Bone

259

Wet specimens can be tested using a specially designed test xture such as that shown in
Figure 17.2, for which cortical bone specimens can be tested without embedding in plastic. The
specimen is held securely in place by two set screws with the polished surface extending slightly
above the aluminum mount. An appropriate solution is added to the test xture until the specimen
is completely immersed, but to avoid problems caused by water rising too far up the indenter shaft,
the levels must be controlled carefully so that water extends only 50 to 100 mm above the specimen
surface. At this level, the indenter tip is immersed during testing, but water does not rise high
enough up the indenter shaft to damage the testing system. The proper water level can be determined
by using an optical microscope to focus separately on the specimen surface and the water surface,
noting when the two focal planes are separated by 50 to 100 mm. Reducing the surface tension of
the water by adding a drop of liquid dishwashing detergent to the water aids in keeping water on
the specimen surface and reduces the force needed for the indenter to penetrate the water surface
at the beginning of a test. The latter effect helps in identifying the point of rst surface contact
during indenter approach to the specimen surface.

III. INDENTATION PROCEDURES

Experiments elucidated here were conducted using a commercially available nanoindenter, a
schematic illustration of which is shown in Figure 17.3. This fully automated hardness testing
system makes small indentations at precise positions on a specimen surface while continuously
monitoring the loads and displacements of the indenter with resolutions of 75 nN and 0.04 nm,
respectively. The specimens are held on an

xyz

table whose position relative to the microscope
or the indenter is controlled with a joystick. The spatial resolution of the position of the table
in the

xy

plane is 400 nm. However, the optical microscope used for locating indentation sites
may limit the positioning resolution to roughly 1 mm. The apparatus is enclosed in an insulated
cabinet to provide thermal stability and suspended on a pneumatic antivibration table to isolate
it from external vibrations.
A typical indent cycle is (1) approach to locate the specimen surface as accurately as possible;
(2) load at a set rate to a set force or displacement limit; (3) hold a short time for the system to
equilibrate fully before proceeding to an unloading; (4) unload at the same rate to typically 90%
of maximum load; (5) hold for a relatively longer time to calculate the thermal drift; and
(6) complete unload.
As an example, tests are typically conducted with a Berkovich diamond indenter (a three-sided
pyramid) in load control using the loadtime sequence such as that shown in Figure 17.4. As seen
in the plot, it is useful to load and unload the indenter several times and include intervening periods

FIGURE 17.2

Mounting stage used for testing wet specimens. (From Rho, J. and Pharr, G.,

J. Mater. Sci.
Mater. Med.,

1999. With permission.)

0266/ ch17/frame Page 259 Thursday, October 21, 1999 4:53 PM

260

Mechanical Testing of Bone and the BoneImplant Interface

during which the load is held constant. The multiple loading scheme is used to examine the extent
to which the loaddisplacement data are inuenced by viscoelastic deformation. Figure 17.5 pre-
sents a typical set of loaddisplacement data obtained from a test conducted in an osteon of dried
human bone. After the initial loading, in which most of the plastic deformation occurs and the
hardness impression is formed, there is a hysteresis in the data which indicates that a non-negligible
portion of the indentation displacement is viscoelastic. Viscoelastic deformation is also apparent
in the creep displacements observed during the constant load hold period at peak load. Viscoelastic
deformation is important because analysis procedures used to obtain mechanical properties from
the nanoindentation loaddisplacement data are premised on the notion that the upper portion of
the unloading curve is dominated by elastic rather than viscoelastic recovery. Thus, to minimize
the effects of viscoelasticity and creep on property measurements, a relatively long constant load
hold period should be inserted prior to the nal unloading, during which the viscoelastic deformation
can diminish to a negligible rate. The second constant load hold period, near the end of the test at
90% of the peak load, is used to establish the rate of thermal expansion or contraction of the testing
apparatus to correct the displacement data for thermal drift. Viscoelastic recovery is signicant
during the initial portion of the hold period, but after a period of time (about 200 s in the experiment
shown in Figure 17.5) the viscoelastic effects become relatively small compared with the thermal
effects. Data from the latter portion of the second hold period are used to establish the thermal
drift rate and correct the displacement data. To minimize thermal drift effects, care must be taken
to let the system thermally equilibrate for an extended period of time prior to testing. The indentation
is then completely withdrawn. The upper half of the data obtained in the nal unloading segment
is used to determine the hardness and elastic modulus.

1

Fused silica, which exhibits elastic isotropy
and has a relatively low modulus-to-hardness ratio, is used to calibrate the indenter shape function.
The elastic modulus of fused silica is 72 GPa, and its Poissons ratio is 0.17. Figure 17.6 shows a
plot of elastic modulus of fused silica as a function of depth. The elastic modulus of fused silica
is independent of indentation depth.

IV. ANALYZING NANOINDENTATION LOADDISPLACEMENT DATA

The most common method for analyzing nanoindentation loaddisplacement data is that of Oliver
and Pharr

1

which expands on ideas developed by Doerner and Nix,

2

but is not constrained by the
assumption of a at punch indenter geometry. A typical loaddisplacement indentation curve is
shown in Figure 17.7. During loading, both elastic and plastic deformation occur under the indenter
as the contacts are changed with increasing depth. The unloading part of the curve is dominated
by elastic displacements. The hardness is found by computing the mean pressure under the indenter

FIGURE 17.3

The mechanical properties microprobe.

0266/ ch17/frame Page 260 Thursday, October 21, 1999 4:53 PM

Nanoindentation Testing of Bone

261

at the point of maximum load. This requires a knowledge of the contact area at that point. Because
the unloading curve is dominated by elastic displacement, it is possible to determine the elastic
modulus of the material being indented from the slope of the unloading curve.

1,18

Figure 17.8 shows a cross section of an indentation and identies the parameters used in the
analysis. At any time during loading, the total displacement

h

is written as

h

=

h

c

+

h

s

, (17.1)
where

h

c

is the vertical distance along which contact is made (called contact depth) and

h

s

is the
displacement of the surface at the perimeter of the contact. At peak load, the load and displacement
are

P

max

and

h

max

, respectively. Upon unloading, the elastic displacements are recovered, and when
the indenter is fully withdrawn, the nal depth of the residual hardness impression is

h

f

.
The analysis begins by tting the unloading curve to the power-law relation

P

=

b

(

h



h

f

)

m

, (17.2)

FIGURE 17.4

The indenter loadtime sequence used for all nanoindentation testing.

FIGURE 17.5

Typical loaddisplacement data (tested in an osteon).

0266/ ch17/frame Page 261 Thursday, October 21, 1999 4:53 PM

262

Mechanical Testing of Bone and the BoneImplant Interface

FIGURE 17.6

Elastic modulus of fused silica as a function of depth.

FIGURE 17.7

A schematic representation of load vs. indenter displacement showing quantities used in the
analysis as well as a graphical interpretation of the contact depth.

FIGURE 17.8

A schematic representation of a section through an indentation showing various quantities
used in the analysis.

0266/ ch17/frame Page 262 Thursday, October 21, 1999 4:53 PM

Nanoindentation Testing of Bone

263

where

P

is the indentation load,

h

is the displacement,

b

and

m

are empirically determined tting
parameters,

h

f

is the nal displacement after complete unloading determined by curve tting.
The unloading stiffness,

S

, that is, the slope of the unloading curve during the initial stages of
unloading, is then established by differentiating Equation 17.2 at the maximum depth of pene-
tration,

h

=

h

max

, giving
, (17.3)
The depth along which contact is made between the indenter and the specimen,

h

c

,

can also be
estimated from the loaddisplacement data using
, (17.4)
where

P

max

is the peak indentation load and

is a constant which depends on the geometry of the
indenter

= 1 for the at punch

;

= 0.75 for the paraboloid of revolution (Berkovich indenter);

= 0.72 for the conical indenter. With these basic measurements, the projected contact area of the
hardness impression,

A,

is derived by evaluating an empirically determined indenter shape function
at the contact depth,

h

c

; that is,

A

=

f

(

h

c

).
Finally, the hardness,

H,

and effective elastic modulus,

E

eff

, are derived from
(17.5)
and
, (17.6)
where

is 1 for circular (at punch), 1.034 for triangular (Berkovich indenter), and 1.012 for
square (Vickers indenter). Note that the values of

for the triangular and square geometries deviate
from the circular one by only 3.4 and 1.2%.

19

An effective modulus is used in the analysis to account for the fact that elastic deformation
occurs in both the specimen and the indenter. The effective modulus is related to the specimen
modulus through
, (17.7)
where

E

and

are indentation modulus and Poissons ratio for the specimen, and

E

i

and

i

are the
same quantities for the indenter. The elastic properties of the diamond indenter,

i

and

E

i

,

are 0.07
and 1140 GPa, respectively. We assume that Poissons ratio for bone is

= 0.3. A sensitivity study
showed that varying

in the range 0.2 to 0.4 changed the measured values of

E

by no more than 8%.
The attractiveness of this approach is that direct observation and measurement of the contact
area is not needed for the evaluation of

H

and

E,

thus facilitating property measurement from very
small indentations. Clearly, however, the accuracy with which

H

and

E

can be measured depends
on how well Equations 17.3 through 17.6 describe the indentation deformation behavior. In this
S
dP
dh
mb h h
h h
f
m
= =
=

max
max
( )
1
h h
P
S
c
=
max
max

H
P
A
=
max
E
S
A
eff
=
1
2

1 1 1
2 2
E E E
i
i eff
=

+


0266/ ch17/frame Page 263 Thursday, October 21, 1999 4:53 PM

264

Mechanical Testing of Bone and the BoneImplant Interface

regard, it is important to note these equations were derived by Sneddon,

20

assuming elastic defor-
mation only. One important way in which an elastic solution fails to describe properly the elas-
tic/plastic behavior observed in indentation concerns the pileup or sink-in of material around the
indenter. In the purely elastic contact solution, material always sinks in, while for elastic/plastic
contact, material may either sink in or pile up. Since this has important effects on the indentation
contact area, it is not entirely surprising that the OliverPharr method has been found to work well
for hard ceramics, in which sink-in predominates, but signicant errors can be encountered when
the method is applied to soft metals that exhibits extensive pileup.

21

Pileup leads to contact areas
that are greater than the cross-sectional area of the indenter at a given depth. These effects lead to
errors in the absolute measurement of mechanical properties by nanoindentation.

3

However, the
elastic and plastic properties of bone are such that pileup is not an important issue, and can largely
be ignored (see Figure 17.10).

V. DETERMINATION OF LOAD FRAME COMPLIANCE
AND DIAMOND AREA FUNCTION

A method for calibrating the shape of the indenter tip and load frame compliance has been developed
by Oliver and Pharr.

1

It is based on the measurement of contact stiffness and requires no imaging
of the indentations. The area function for a perfect Berkovich indenter can be described by

A

(

h

c

) = 24.5

h

c

2

. (17.8)
However, all Berkovich indenters have some deviation from this geometry at the tip. Since the
measured indentation displacement is the sum of the displacements in the specimen and the load
frame, the load frame compliance must be known to determine specimen displacements accurately.
To a good approximation, the load frame and the specimen can be modeled as two springs in series,
for which case the total measured compliance is

C

=

C

s

+

C

m

,

(17.9)
where

C

s

is the compliance of the specimen (the contact compliance), and

C

m

is the compliance
of the testing machine. The contact compliance depends on the contact area at that point,

A

. Since
the specimen compliance during elastic contact is given by the inverse of the contact stiffness,

S,
Equations 17.6 and 17.9 combine to yield
. (17.10)
Equation 17.10 shows that a plot of C vs. A
1/2
is linear. The intercept of the plot is the load frame
compliance, C
m
, and the slope is proportional to 1/E
eff
. The best values of C
m
are obtained when
the second term on the right-hand side of Equation 17.10 is small, i.e., for large indentations. To
implement the procedure, a series of indentations are made in aluminum, some of which can be
quite large due to its low hardness. For the larger aluminum indentations, the area function is
assumed to be that for a perfect Berkovich indenter (Equation 17.8), which provides a rst estimate
of the contact area. Initial estimates of C
m
and E
eff
are then obtained by plotting C vs. A
1/2
. By
using these values, the contact area is computed by rewriting Equation 17.10 as
. (17.11)
C C
E A
m
= +
1
2
1

eff
A
E C C
m
=

4
1 1
2 2
eff
( )
0266/ ch17/frame Page 264 Thursday, October 21, 1999 4:53 PM
Nanoindentation Testing of Bone 265
An initial guess at the area function is made by tting the A vs. h
c
data to a relationship such as
(17.12)
where n
1
through n
8
are constants.
The procedure is not complete at this stage because the exact form of the area function inuences
the values of C
m
and E
eff
, so using the new area function the procedure is applied again and iterated
several times until convergence is achieved. To determine area function at shallower depths, a
similar method involving indentations in fused quartz is used (fused quartz is much harder than
aluminum, so it is easier to make small indenters), assuming the same machine compliance as
measured in the aluminum experiments. This area function is assumed to describe the tip geometry
completely, and thus it yields the contact area of any indentation made to a known depth. In this
approach, the cross-sectional area of the tip and the contact area of the indenter and the specimen,
at a given depth, are treated implicitly as if they are the same.
3
However, the tip-end geometry
needs to be known precisely to determine the true contact area between the indenter and the material
being tested when the contact area is small, or the machine compliance must be accurately
determined where the specimen is stiff.
22
Figure 17.9 shows a plot of the measured compliance vs. contact area for a sample of fused
quartz. We noted that Equation 17.10 is very well obeyed and that the elastic modulus of quartz
can be found from the slope of the curve.
VI. LIMITATIONS
The analytical methods outlined here are derived under the assumption that the material is
homogeneous and isotropic in its elastic properties. When used to measure the elastic modulus
of an anisotropic material such as bone, the modulus derived from the method is an average of
the anisotropic elastic constants biased toward the modulus in the direction of testing.
5
The
relationship between the elastic modulus measured by nanoindentation using the OliverPharr
method and the anisotropic elastic constants has been examined by Vlassak and Nix.
23,24
They
showed that the elastic modulus obtained from nanoindentation techniques and the actual Youngs
modulus are the same only for elastically isotropic media. For materials with lower symmetry,
elastic anisotropy affects the indentation measurements. The inuence of anisotropy on
FIGURE 17.9 Unloading compliance vs. contact area for the indentation of fused quartz. Equation 17.10 is
well obeyed. The effective elastic modulus can be found from the slope of the curve.
A h h n h n h n h n h
c c c c c c
( ) .
/ / /
= + + + + 24 5
2
1
1
2
1 2
3
1 4
8
1 128
0266/ ch17/frame Page 265 Thursday, October 21, 1999 4:53 PM
266 Mechanical Testing of Bone and the BoneImplant Interface
experimental measurements has been studied by Hay et al.
25
who examined single crystals of -
silicon nitride with the long axis of the crystal oriented parallel to the c-direction of the hexagonal
crystal structure. They found that the elastic modulus obtained from nanoindentation in the
longitudinal direction underestimates the single-crystal Youngs modulus while in the transverse
direction, the nanoindentation elastic modulus overestimates the single-crystal Youngs modulus.
Hay et al.
25
also found that the elastic modulus obtained from nanoindentation is dominantly
controlled by the elastic properties in the indentation direction, and is only weakly inuenced by
properties in the transverse directions. However, without some prior knowledge of the type of elastic
anisotropy and an estimate of the associated anisotropic elastic constants, it is difcult to estimate how
large the inuences of anisotropy may be. This problem needs to be investigated further.
VII. RESULTS FOR MEASUREMENTS IN BONE
Recently, the elastic properties of several microstructural components of dry human vertebrae
and tibiae have been investigated in the longitudinal and transverse directions using nanoinden-
tation.
6
The bone sections were prepared in the same way as mentioned above. A total of 373
indentations were produced in this study. In the vertebrae, three to ve indentations were made
in 12 separate trabeculae in the longitudinal direction (i.e., perpendicular to the transverse section)
and 29 separate trabeculae in the transverse direction. For the tibiae, three to ve indentations
were made in 15 osteons and 14 interstitial lamellae in the longitudinal direction. In the transverse
direction (i.e., perpendicular to the longitudinal section), it was difcult to distinguish osteons
microstructurally from interstitial lamellae (Figure 17.10), so individual measurements were not
possible. Consequently, the transverse direction data represent an average over all the indentations
in the longitudinal section without regard to their exact locations in the two microstructural
components. Basic modulus and hardness results are summarized in Table 17.1. The elastic
modulus was found to be highest for the interstitial lamellae, intermediate for osteonal lamellae,
and lowest for trabecular bone. As regards direction, moduli determined in the longitudinal
direction were always greater than those determined in the transverse direction for a given
microstructural constituent. The ratio of elastic moduli in the longitudinal to transverse directions
was found to be 1.35 for osteons and 1.29 for trabecular bone. The hardnesses follow in similar
order. The ratio of hardness in the longitudinal to transverse direction is 1.10 for osteons and
1.19 for trabeculae. The mean values of elastic moduli for all of the bone components were
found to be statistically different (p < 0.05). Therefore, for a given microstructural component,
there is an anisotropy between the longitudinal and transverse directions. Similar conclusions
hold for the hardnesses.
More recently, the nanoindentation method was used to examine variations in the individual
lamellar properties within osteons, as a function of distance from the osteonal center (Haversian
canal).
7
Elastic moduli and hardnesses of the individual lamellae within the osteon decrease
monotonically as a function of increasing distance from the osteonal center. The mechanical
properties of osteons are signicantly lower than those of the interstitial bone (p < 0.0001). There
were statistically signicant differences (p = 0.0005 and 0.0004, respectively) between lamellar
properties obtained from the two innermost osteonal lamellae (E = 20.8 1.3 GPa and H = 0.65
0.06 GPa) and those from the two outermost osteonal lamellae (E = 18.8 1.0 GPa and H =
0.55 0.05 GPa). The ratio (E
1
/E
2
) of the elastic moduli of the outermost osteonal lamella (E
1
)
(considered to be the soft part of the osteons) and that of interstitial bone (E
2
) was approximately
0.7. These results contradict expectations based on the current understanding of the physiology and
the mechanism of osteonal remodeling, but they may have important implications for the mechanical
contribution of individual osteons for bone biomechanics.
0266/ ch17/frame Page 266 Thursday, October 21, 1999 4:53 PM
Nanoindentation Testing of Bone 267
TABLE 17.1
Mean Elastic Moduli and Hardnesses for the Microstructural Components of Cortical
and Trabecular Bone as Measured in This Study
Bone Type
Direction to Be
Tested
No. of
Subjects
No. of
Indentations
No. of
Microstructural
Components
Elastic
Modulus,
GPa
(SD)
Hardness,
GPa
(SD)
Cortical bone Longitudinal
Osteons 2 72 15 22.4 (1.2) 0.617 (0.039)
Interstitial lamellae 2 58 14 25.7 (1.0) 0.736 (0.044)
Transverse 2 58 13 16.6 (1.1) 0.564 (0.034)
Trabecular bone Longitudinal 3 53 12 19.4 (2.3) 0.618 (0.061)
Transverse 7 132 29 15.0 (2.5) 0.515 (0.082)
Note: Standard deviations are shown in parentheses (SD). The transverse cortical bone data represent an average over all
the transverse section indentations without regard to their exact locations in the two microstructural components.
Source: Rho, J. et al., J. Biomed. Mater. Res., 45, 4854. 1999 John Wiley & Sons, Inc. With permission.
FIGURE 17.10 Typical nanoindentation of interstitial lamellar bone of tibia. (a) Scanning electron micro-
graph (5000) and (b) atomic force microscopy.
0266/ ch17/frame Page 267 Thursday, October 21, 1999 4:53 PM
268 Mechanical Testing of Bone and the BoneImplant Interface
ACKNOWLEDGMENTS
This research was sponsored by a faculty summer research fellowship from the Oak Ridge Institute
for Science and Technology (ORISE) and the National Institutes of Health AR45297 (JY). Ana-
lytical instrumentation for the nanoindentation testing was provided by the Division of Materials
Sciences, U.S. Department of Energy, under Contract DE-AC05-96OR22464 with Lockheed Martin
Energy Research Corp. and through the SHaRE Program under Contract DE-AC05-76OR00033
between the U.S. Department of Energy and Oak Ridge Associated Universities.
REFERENCES
1. Oliver, W.C. and Pharr, G.M., An improved technique for determining hardness and elastic modulus
using load and displacement sensing indentation experiments, J. Mater. Res., 7, 1564, 1992.
2. Doerner, M.F. and Nix, W.D., A method for interpreting the data from depth-sensing indentation
instruments, J. Mater. Res., 601, 1986.
3. McElhaney, K.W., Vlassak, J.J., and Nix, W.D., Determination of indenter tip geometry and indentation
contact area for depth-sensing indentation experiments, J. Mater. Res., 13, 1300, 1998.
4. Mayo, M.J., Siegel, R.W., Liao, Y.X., and Nix, W.D., Nanoindentation of nanocrystalline ZnO,
J. Mater. Res., 7, 973, 1992.
5. Rho, J.Y., Tsui, T.Y., and Pharr, G.M., Elastic properties of human cortical and trabecular lamellar
bone measured by nanoindentation, Biomaterials, 8, 1325, 1997.
6. Rho, J.Y., Roy, M.E., Tsui, T.Y., and Pharr, G.M., Elastic properties of microstructural components
of human bone measured by nanoindentation, J. Biomed. Mater. Res., 45, 48, 1999.
7. Rho, J.Y., Zioupos, P., Currey, J.D., and Pharr, G.M., Variations in the individual lamellar properties
within osteons by nanoindentation, Bone, in press.
8. Roy, M.E., Rho, J.Y., Tsui, T.Y., et al., Mechanical and morphological variation of the human lumbar
vertebral cortical and trabecular bone, J. Biomed. Mater. Res., 44, 191, 1999.
9. Blackburn, J., Hodgskinson, R., Currey, J.D., and Mason, J.E., Mechanical properties of microcallus
in human bone, J. Orthop. Res., 10, 237, 1992.
10. Ziv, V., Wagner, H.D., and Weiner, S., Microstructuremicrohardness relations in parallel-bered and
lamellar bone, Bone, 18, 417, 1996.
11. Ascenzi, A., Benvenuti, A., and Bonucci, E., The tensile properties of single osteonic lamellae:
technical problems and preliminary results, J. Biomech., 5, 29, 1982.
12. Prendergast, P.J. and Huiskes, R., Microdamage and osteocyte-lacuna strain in bone: a microstructural
nite element analysis, J. Biomech. Eng., 118, 240, 1996.
13. Evans F.G., Mechanical Properties of Bone, Charles C Thomas, Springeld, IL, 1973.
14. Yamada, H., Strength of Biological Materials, Williams & Wilkins, Baltimore, MD, 1970.
15. Townsend P.R. and Rose, R.M., Buckling studies of single human trabeculae, J. Biomech., 8, 199, 1975.
16. Currey, J.D., The effects of drying and re-wetting on some mechanical properties of cortical bone,
J. Biomech., 21, 439, 1988.
17. Rho, J.Y. and Pharr, G.M., Effects of drying on the mechanical properties of bovine femur measured
by nanoindentation, J. Mater. Sci. Mater. Med., 10, 485, 1999.
18. Nix, W.D., Elastic and plastic properties of thin lms on substrates: nanoindentation techniques, Mater.
Sci. Eng., A234, 37, 1997.
19. Pharr, G.M., Oliver, W.C., and Brotzen, F.R., On the generality of the relationship among contact
stiffness, contact area, and elastic modulus during indentation, J. Mater. Res., 7, 613, 1992.
20. Sneddon, I.N., The relation between load and penetration in the axisymmetric Boussinesq problem
for a punch of arbitrary prole, Int. J. Eng. Sci., 3, 47, 1965.
21. Bolshakov, A. and Pharr, G.M., Inuences of pileup on the measurement of mechanical properties by
load and depth sensing indentation techniques, J. Mater. Res., 13, 1049, 1998.
22. Hainsworth, S.V., Chandler, H.W., and Page, T.F., Analysis of nanoindentation loaddisplacement
loading curves, J. Mater. Res., 11, 1987, 1996.
23. Vlassak, J.J. and Nix. W.D., Indentation modulus of elastically anisotropic half spaces, Philos. Mag.
A, 67, 1045, 1993.
0266/ ch17/frame Page 268 Thursday, October 21, 1999 4:53 PM
Nanoindentation Testing of Bone 269
24. Vlassak, J.J. and Nix, W.D., Measuring the elastic properties of anisotropic materials by means of
indentation experiments, J. Mech. Phys. Solids, 42, 1223, 1994.
25. Hay, J.C., Sun, E.Y., Pharr, G.M., et al., Elastic anisotropy of -silicon nitride whiskers, J. Am. Ceram.
Soc., 81, 2661, 1998.
0266/ ch17/frame Page 269 Thursday, October 21, 1999 4:53 PM
0266/ ch17/frame Page 270 Thursday, October 21, 1999 4:53 PM

271

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

18

Single Osteon
Micromechanical Testing

Maria-Grazia Ascenzi, Alessandro Benvenuti,
and Antonio Ascenzi

CONTENTS

I. Introduction ..........................................................................................................................271
II. Elements of Micromechanical Testing ................................................................................272
III. Testing Methods ...................................................................................................................274
A. Tension Test....................................................................................................................274
B. Compression Test ...........................................................................................................276
C. Shearing Test ..................................................................................................................278
D. Pin Test .......................................................................................................................280
E. Bending Test...................................................................................................................281
F. Torsion Test ....................................................................................................................282
1. Frasca Method...........................................................................................................283
2. A. Ascenzi Method....................................................................................................283
G. Cyclic Loading in Tension and Compression................................................................284
IV. Acquired Knowledge and Pending Questions.....................................................................287
Acknowledgments ..........................................................................................................................288
References ......................................................................................................................................288

I. INTRODUCTION

Because sample dimensions may affect the mechanical behavior of a material, mechanics
differentiates into micro- and macromechanics. The mechanics of biological materials, micro-
and macrobiomechanics, can be viewed as a fundamental expression of skeleton physiology
and pathology

1,2

and, hence, as complements of skeletal anatomy. This chapter treats the
microbiomechanics of single secondary osteons.
While the observations of Galileo Galilei

3

in the 1600s can be considered to have initiated bone
macromechanics, bone micromechanics began much later, between 1937 and 1942, with the studies
conducted at the Anatomy School of Bologna, according to Evans.

4

Such studies did not, however,
include techniques for isolation of single microscopic structural units.
The justication for studying bone micromechanics is found in bone morphology. The bone
component of bone samples is not at all structured as a simple aggregation of molecular and
supramolecular organic and inorganic units, which results in a homogeneous macroscopic entity.
Petersen

5

rst introduced the concept that bone structures can be viewed as a four-order hierarchy,
arranged in decreasing size. The rst order comprises the structures corresponding to gross shape
and differentiation between compact and cancellous bone. The second order of compact bone
includes Haversian systems, cylindrical lamellar systems, and additional related structures, e.g.,
bone marrow. The third order consists of collagen brils which lie in the ground substance; one

0266/ ch18/frame Page 271 Friday, October 22, 1999 8:50 AM

272

Mechanical Testing of Bone and the BoneImplant Interface

of their roles, in association with osteocytes, is the constitution of lamellae. The fourth order consists
of the molecular patterning between organic and inorganic substances. This four-order hierarchy
forms through a slow, continuous process of replacement of primary or woven bone, which begins
as woven bone is formed, in the fetus or soon after birth, and continues throughout life.
In view of the structural hierarchical orders described, the mechanical analysis of a rst-order
sample is possible only with knowledge of the properties of the three lower orders; conversely,
from knowledge of properties of the three lower structural orders, mechanical behavior of the rst-
order sample can be derived. This is the case under both normal and pathological conditions so
long as the pathology does not alter the structural conguration of the four orders.
An additional motivation to study the mechanics of osteons resides in the lack of knowledge
about the structure of some of the microscopic entities which comprise them, such as lamellae in
alternate osteons. Increased knowledge of osteon mechanics may shed light on the microstructural
problems of such entities.

II. ELEMENTS OF MICROMECHANICAL TESTING

The methods for single osteon micromechanical testing share three well-established steps:
(1) osteon selection; (2) sample isolation; (3) sample loading by an apparatus crafted in accor-
dance with sample shape and dimensions, as well as load type and magnitude. Listed here are
aspects of these steps, which apply to each of the micromechanical testing methods to be
presented individually.
Any human or vertebrate secondary osteon may be selected and isolated. Ortner and Yong

6

and
Frasca et al.

7

describe methods for isolating whole osteons. However, whole osteons display irreg-
ular shape, and such irregularity prevents accurate mechanical loading and comparison of mechan-
ical loading results. Consideration of osteon samples of simple and same geometric shape, rather
than whole osteons, allows comparison of mechanical loading results. Human osteons, the osteons
considered in this chapter, have a somewhat cylindrical shape around a vascular canal. The osteon
sample shape that suitably represents osteons and their mechanical properties is accordingly a
cylindrical shape around a central vascular canal.
The selection of specic osteons for isolation depends on two factors: (1) sample dimensions,
especially when the experimentation requires samples of approximately the same size; (2) sample
type, described in terms of the two principal bone mechanical components which present three
characteristics. The two components are collagen bundles and hydroxyapatite crystallites, and the
three characteristics are distribution of collagen bundles, orientation of collagen bundles, and
amount of hydroxyapatite crystallites.
Osteons are selected from one or more sections of fresh compact bone, not previously xed.
Sections are obtained by grinding on a glass plate while air cooling to prevent heating of material
or by using a rotating-saw microtome equipped with a continuous waterspout to prevent heating
of material. Sections of the same thickness are generally cut in parallel (longitudinal sections) or
perpendicularly (transverse or cross sections) to the axes of the osteons.
The classic theory categorizes osteons structurally according to the types of coaxial lamellae
which comprise them. Two lamellar types are evidenced in bone sections by the polarizing micro-
scope. Classical theory supposes that lamellae which appear bright in longitudinal sections and
dark or extinct in cross sections under the polarizing microscope consist of collagen bundles parallel
to the osteon axis. They are called longitudinal lamellae. Classical theory supposes that lamellae
which appear dark in longitudinal sections and bright in cross sections under the polarizing
microscope consist of collagen bundles perpendicular to the osteon axis. They are called transverse
or circularly bered lamellae. In the last 20 years, Frasca et al.,

8

A. Ascenzi and Benvenuti,

9,10

and
Giraud-Guille

11

reported the additional presence of oblique bundles with discrete or continuous
orientations relative to osteon axis. Pending denitive ascertainment, the terms

transverse

or

circularly bered

lamella are used.

0266/ ch18/frame Page 272 Friday, October 22, 1999 8:50 AM

Single Osteon Micromechanical Testing

273

Among the various arrangements of longitudinal and transverse lamellae, those characterizing
two typical osteon types were chosen for the experiments that this chapter describes. In the rst
type, collagen bundles have a marked longitudinal spiral course, with the pitch of the spiral changing
so slightly that the bundles in one lamella make an angle of approximately 0

with the bundles of
the next lamella; under the polarizing microscope, such osteons appear homogeneously bright in
longitudinal section and dark in cross section with few bright lamellae at the periphery and around
the Haversian canal. A few, very thin, incomplete bright lamellae may also be present. Such osteons
are called longitudinal osteons (Figure 18.1a). In the second type, bundles



in one lamella make an
angle of approximately 90

with the bundles of the next lamella; under the polarizing microscope,
such osteons show an alternation of bright and dark lamellae in both longitudinal and cross section
(Figure 18.1b), as reported by Gebhardt

12

and Amprino,

13

and they are called alternate osteons. In
only a few studies was a third typical osteon type selected, that called transverse or circular. In the
third type, collagen brils have a marked transverse spiral course in successive lamellae; under the
polarizing microscope, such osteons appear homogeneously dark in longitudinal section and bright
in cross section (Figure 18.1c).
The three mentioned osteon types are representative of osteons because they are outstanding
appearances of a continuous distribution of osteon structures. The continuity of the distribution is
due to the varying mesh of longitudinal and transverse lamellae.
The degree of calcication of single osteons is assessed by means of the microradiographic tech-
nique. On a micro-X-ray, the calcium amount in each unit is revealed by a shade ranging from light to
dark gray. Since a micro-X-ray is a magnied negative lm, the image of osteonic section at initial
(nal, respectively) stage of calcication is dark (light, respectively) gray. Osteons at an intermediate
stage of calcication show an intermediate shade of gray. When osteons at either initial or nal stage
of calcication are desired, their recognition is straightforward. To assess exactly intermediate stages
of calcication, a scale of comparison is necessary. In this case, a scale made of very thin, superimposed
aluminum sheets is microradiographed together with the bone section on the same X-ray lm.
After complete isolation of a cylindrical sample, the regularity of the sample is assessed in
terms of the position and orientation of the canal. The distance between the vascular canal and
external surface of the sample at various rotational angles and levels is required to be constant.
The percent of discarded samples may reach 93 to 95% of prepared samples.
Skilled technicians designed and made the osteon loading apparatus which this chapter
describes. Such apparatus proved useful and manageable to A. Ascenzis research group. Apparatus
based on other principles, but of similar sensitivity, could, of course, also be manufactured.
Osteon samples are loaded under physiological conditions, i.e., after rehydration by soaking
in saline solution. Currey veried the validity of this method.

14

The experiments are preferably performed at a temperature of 20 to 22

C.

FIGURE 18.1

The three paradigmatic osteon types referenced in this chapter as they appear after isolation
and in cross section under a polarizing microscope: (a) longitudinal; (b) alternate; (c) transverse.

0266/ ch18/frame Page 273 Friday, October 22, 1999 8:50 AM

274

Mechanical Testing of Bone and the BoneImplant Interface

III. TESTING METHODS

Methods for the study of mechanical properties of isolated osteon samples follow. Results are
briey reported after each mechanical test description to facilitate understanding of the method.

A. T

ENSION

T

EST

Design and preparation of osteon samples for axial tensional loading was a time-consuming process
which involved diverse and repeated attempts.
Initially, it seemed unfeasible to obtain osteon samples of cylindrical shape around a vascular
canal. Hence, hemidiametral osteon sections, 50 to 60

m thick, were prepared. This type of sample
at rst presented some advantages: preparation by dissection and assessment of osteon type were
straightforward. Tension testing showed that longitudinal osteons were stronger in tension than
alternate osteons, as anticipated by structural considerations.

15,16

However, stressstrain curves of
alternate osteon samples showed frequent abrupt changes in slope, so-called knees, at low stresses,
similar to those recorded in some cross-ply laminates. Electron microscope examination revealed
that such knees were due to failure of interbrillar cementing substance in the lamellae whose
collagen bundles were mainly perpendicularly oriented to the loading direction.

17

Hemidiametral
sample preparation eliminated the longitudinal initial stress that ordinarily protects alternate osteons
at low axial stresses.

18-21

Hence, this kind of sample was inadequately representative of osteon
mechanical properties.
Studies were accordingly redirected toward development of a method for isolation of a cylin-
drical sample around the vascular canal which would also allow certain assessment of osteon type.
This goal was achieved and the method follows.
A longitudinal section of thickness slightly larger than the osteon diameter, i.e., 300 to 350

m,
is cut from a bone shaft. It is difcult to locate longitudinal and alternate osteons on the section:
the overlapping of concentric lamellae may mean that fewer or no dark lamellae are visible under
a polarizing microscope; therefore, an alternate osteon might appear bright and be mistaken as a
longitudinal one. Denite identication of osteon type is possible only after testing. It is accom-
plished by observing under a polarizing microscope a thin cross section cut with a microscopic
drill. Longitudinal and alternate osteons, at initial and nal stages of calcication, are tentatively
chosen.
Isolation of osteon samples from sections occurs in two steps. In Step 1, a dental drill is inserted
into the body of a microscope in place of the tube. The bone section is rmly secured on the
microscope stage, perpendicularly oriented with respect to the drill. As the drill and microscope
stage are operated simultaneously, the device allows longitudinal cuts so as to coarsely free an
osteon sample around the vascular canal from the surrounding material. The shape of the sample
is thinner in the middle portion, which is a parallelepiped, and larger at the rectangular cross-
sectional ends (Figure 18.2a). Such lugs serve to secure the sample to the loading device when
inserted into the loading device rectangular jaws. The total length of the sample measures 500

m,
the maximum osteonic stretch that does not contain Volkmanns canals, which would create
discontinuities. Samples are air-cooled during coarse isolation to avoid overheating.
In Step 2, the middle portion of the sample is transformed from a parallelepiped to a
cylindrical shape whose axis coincides with the axis of the vascular canal (Figure 18.2b) by
means of a microgrinding lathe designed by A. Ascenzis research group and manufactured by
CECOM Company.
The microgrinding lathe features two jaws, powered by an electric motor and synchronously
rotating around a common axis; they are used to secure the osteon sample. Once the shape of the
sample has been corrected, the rotating axis of the system should coincide with the axis of the
vascular canal. A small steel blade, whose edge measures nearly 500

m in length, produces the
grinding effect. The blade moves forward or backward and its movements are monitored by a

0266/ ch18/frame Page 274 Friday, October 22, 1999 8:50 AM

Single Osteon Micromechanical Testing

275

micrometer. The maximum penetration of the blade into the sample roughly corresponds to the
minimum osteon radius in order to eliminate signicant residues of adjacent structures. Samples
are air-cooled during grinding to avoid overheating. At this point, complete isolation of a cylindrical
sample is achieved (Figure 18.3). The transverse area of the cylindrical portion of the sample is
used to calculate the ultimate tensile stress.
The loading apparatus consists of an extensometer furnished with a microwave micrometer
which allows loading of osteon samples and measurement of sample elongations. It was devised
by A. Gozzini, late holder of the chair of Solid State Physics at the Scuola Normale Superiore, Pisa.

22

The informing principle of the apparatus follows. Let

h

and

D

be height and diameter, respec-
tively, of a metallic cylinder whose cavity functions as a resonator for electromagnetic waves. The
resonance frequency,

f

o

, of this cavity will depend on

h

,

D

, and the electromagnetic eld congu-
ration inside its cavity, called mode. When the Te

012

mode is the chosen conguration type, the
following relation holds:

23

f

o

2

= (

C/D

)

2

(1.488 + (

C/h

)

2

) (18.1)
where

C

is the velocity of light in a vacuum.

FIGURE 18.2

Diagram showing the preparation steps of a cylindrical osteon sample for tensional loading.
(a) Step 1. The sample has been isolated from the bone section by means of a dental drill. The sample is still
surrounded by bone material, is thinner in the middle portion, and has a rectangular cross section.
(b) Step 2. The material surrounding the middle portion of the sample has now been removed by means of a
special microgrinding lathe. The sample shows a middle cylindrical shape. (From Ascenzi, A., Baschieri, P.,
and Benvenuti, A., The torsional properties of single selected osteons,

J. Biomech.

, 27, 875, 1994. With
permission from Elsevier Science.)

FIGURE 18.3

Osteon sample with its lugs. (From Ascenzi, A., Baschieri, P., and Benvenuti, A., The torsional
properties of single selected osteons,

J. Biomech.

, 27, 875, 1994. With permission from Elsevier Science.)

0266/ ch18/frame Page 275 Friday, October 22, 1999 8:50 AM

276

Mechanical Testing of Bone and the BoneImplant Interface

Figure 18.4 shows a diagram of the apparatus used to measure the elongation of osteon samples
under tensile load. Each of the two ends of the sample is secured into a jaw; one of the two jaws
supports a Plexiglas disk whose lower metal-coated surface forms the upper plane of the cylindrical
metal cavity. The disk, which may slide into the vertically oriented cavity, has a very thin nylon
thread attached to its center. The thread runs along the longitudinal axis of the cavity and exits
through a small hole at the center of the lower xed end of the cylindrical metal cavity. To exert
traction on the osteon sample, an increasing number of weights is attached to the free end of the
thread. As the osteon sample elongates, the sample pulls the lug afxed to the upper disk of the
cylindrical cavity; the upper disk slides into the cavity reducing the height of the cavity; such
reduction produces a change in resonance frequency in the cavity. Therefore, as elongation of
sample (

l

) takes place, the height of the cavity reduces by the same amount (

l

) and resonance
frequency changes by

f

o



in the cavity. In this apparatus,

l

and

f

o



are



small with respect to

h

.
Equation 18.1 gives, for small

l

s,

l

=

f

o



(1+1.488(

h/

)

2

)(

h/f

o

) (18.2)
A pulse technique

22

is used to measure

f

o

.



An oscilloscope, based on Equation 18.2 and inserted
in the apparatus, transforms

f

o


values into

l

values.



The pulse technique has the advantages of
being relatively simple to use, highly sensitive, and accurate. Accuracy in osteon sample elongation
measurements is within 1%; accuracy partially depends on exact alignment of the osteon sample
and apparatus.
In accordance with the results obtained on hemidiametral osteon samples, longitudinal osteon
cylindrical samples resist tension better than alternate osteon cylindrical samples at both initial and
nal stages of calcication. A structural explanation may reside in the supposed diversity of the
orientation of collagen bundles in alternate osteon samples, especially with regard to the presence
of transverse bundles.

B. C

OMPRESSION

T

EST

Over the years, two methods have been developed to load and measure sample length changes in
compression. In contrast with tension methods, both compressive methods are equally viable and
the rst one presented allows for recognition of osteon type before sample isolation.
Thickness of bone sections may measure at most 500

m to be transparent enough to allow
examination under the polarizing microscope. For thicker sections, bone transparency may be
somewhat improved by soaking the section in bromoform, if necessary. A transverse section of
long bone diaphysis 500

m thick contains many straight, nonbranching portions of osteons which
may be suitable for testing, once isolated.

24

Longitudinal, alternate, and transverse osteons at both initial and nal degree of calcication
were selected on transverse sections 500

m thick.
The specially designed device used for sample isolation consists of a very thin, carefully
sharpened steel needle inserted off-center in a dental drill. As the drill turns, the tip of the needle
describes a circle whose diameter may be adjusted to match the diameter of the sample. When the
needle rotation axis is perpendicular to bone section surfaces, i.e., coincides with the osteon axis,
the tip of the needle cuts an osteon sample of cylindrical shape with walls of uniform thickness
just inside its limits. To ensure that the needle rotation axis is perpendicular to bone section surfaces,
the drill handle is inserted in a microscope body in place of the tube and the section is rmly
secured onto the microscope stage. Coarse microscope adjustment easily controls the needle
movements into the bone section. Osteon cutting is controllable by watching the operation through
a stereoscopic microscope. Figure 18.5 shows the shape of samples prepared with the above-
described technique.

0266/ ch18/frame Page 276 Friday, October 22, 1999 8:50 AM

Single Osteon Micromechanical Testing

277

Length and diameter of samples are accurately measured by means of an eyepiece micrometer.
In experiments, the ratio of length to diameter ranged between 2.5 to 1 and 3 to 1.

24

A microcompressor furnished with a microwave micrometer based on the cavity and pulse
technique described in Section III.A above was used to load the specimens.
Figure 18.6 shows the apparatus diagram. The lower end of the sample is supported by the upper
base of the xed cylinder. The upper end is in contact with the push cylinder whose upper surface
is the lower plane of the cylindrical cavity, which functions as a resonator for electromagnetic waves.

FIGURE 18.4

Diagram of cavity used as extensometer. From top to bottom: CA, cylindrical appendage;
UJ, upper jaw with its screw; S, sample; SR, screw ring; LJ, lower jaw with its screw; DFUP, disk forming
the upper plane of the cavity; NT, nylon thread; C, cavity; DFLP, disk forming the lower plane of the cavity;
W, weights. (From Ascenzi, A., Bonucci, E., and Checcucci, A.,

Studies on the Anatomy and Function of Bone
and Joints

, F.G. Evans, Ed., Springer-Verlag, Berlin, 1966. With permission from Springer-Verlag, New York.
From Ascenzi, A. and Bonucci, E., The tensile properties of single osteons,

Anat. Rec.

, 158, 375, 1967, The
Wistar Institute. Reprinted by permission of Wiley-Liss, Inc., a division of John Wiley & Sons, Inc.)

0266/ ch18/frame Page 277 Friday, October 22, 1999 8:50 AM

278

Mechanical Testing of Bone and the BoneImplant Interface

The light (3.5 g) push cylinder can move freely inside the vertically oriented cavity. When weights
are added to the thread attached to the push cylinder, the osteon sample undergoes a shortening
which makes the cylindrical cavity lower plane slide, thus increasing the cavity height. Because
sample shortening (

l

) is equal in magnitude to cavity height elongation (

l

), sample shortenings
are deduced from resonant frequency changes through Equation 18.2. Length change measurements
are accurate up to 1%.
Longitudinal and alternate osteon samples show a compressive behavior opposite to tensional
behavior. Modulus of elasticity and ultimate tensile strength values are maximum, intermediate,
and minimum, respectively, for transverse, alternate, and longitudinal osteon types at the same
degree of calcication. Modulus of elasticity and ultimate tensile strength values increase as
calcication progresses.
The second compression test method differs from the previous one because the osteon sample
shape is the sample shape used for tension loading, namely, a cylindrical central portion with
rectangular cross-sectional ends. Consequently, the sample is secured to the loading apparatus by
means of jaws as opposed to simply laying it in contact with the xed cylinder surface and the
push cylinder surface. The use of jaws allows loading under both tension and compression and, in
particular, under tensioncompression cyclic loading (see Section III.G). The disadvantage of the
second method remains that denite identication of osteon type occurs only after loading.

C. S

HEARING

T

EST

A. Ascenzi and Bonucci

25

have tested osteon samples loaded centrally and off-center with respect
to their axes by means of a technique that may be viewed as a double-shearing strength test.
Longitudinal, alternate, and transverse osteons at both initial and nal degree of calcication
were selected on 300-

m-thick transverse sections. Each chosen osteon must pass through the
section as shown by the appearance of its sectioned ends on both sides of the section.

FIGURE 18.5

Isolated osteon sample to be loaded in compression, as seen in transmitted light. (From
Ascenzi, A. and Bonucci, E., The compressive properties of single osteons,

Anat. Rec.

, 61, 377, 1968, The
Wistar Institute. Reprinted by permission of Wiley-Liss, Inc., a division of John Wiley & Sons, Inc.)

0266/ ch18/frame Page 278 Friday, October 22, 1999 8:50 AM

Single Osteon Micromechanical Testing

279

Unlike previously described testing, sample isolation for a shearing test is part of sample
loading. Hence, for shearing, isolation is examined in association with the apparatus description.
Figure 18.7 illustrates the method. It shows a compact bone transverse section with an osteon
oriented perpendicularly to the upper and lower surface of the section itself. The section is supported
by a rigid plane with a hole at its center. A steel cylinder, whose axis is aligned to the osteon axis
of the sample, rests on the osteon upper end and acts as a punch when loaded with weights. In A.
Ascenzis groups investigation, the punch diameter was slightly smaller than the osteon diameter;
although it may be smaller, according to the shearing surfaces to be studied. As the punch is loaded
by adding small weights, the punch presses on the osteon; as soon as the punch displacement
caused by an additional weight stopped, additional weights were added. At rst, the punch induces

FIGURE 18.6

Diagram of cavity used as microwave micrometer and testing apparatus in compression. UPC,
upper plane of cavity; C, cavity; LPC, lower plane cavity; PC, push cylinder; S, sample; FC, xed cylinder;
NT, nylon thread; W, weights. (From Ascenzi, A. and Bonucci, E., The compressive properties of single
osteons,

Anat. Rec.

, 161, 377, 1968, The Wistar Institute. Reprinted by permission of Wiley-Liss, Inc., a
division of John Wiley & Sons, Inc.)

0266/ ch18/frame Page 279 Friday, October 22, 1999 8:50 AM

280

Mechanical Testing of Bone and the BoneImplant Interface

a deformation, then the connections between adjacent lamellae, at the immediate periphery of the
punch, break. Consequently, as soon as ultimate shearing strength is reached and connections break,
a cylindrical osteon sample begins to slip through the lower surface of the section and through the
supporting plane hole (Figure 18.8). The diameter of the osteon sample coincides with the diameter
of the punch.
A microwave micrometer based on cavity and pulse technique measures the progressive descent
of the punch until ultimate strength is reached. The principle of the microwave micrometer is
described in Section III.A.
Longitudinal osteons are the osteons least able to resist shearing stress. This suggests that bone
compactness of the other osteon types is strengthened by circular collagen bundles. Ultimate
strength and elasticity modulus in shear increases as calcication proceeds. Shearing properties of
osteons loaded centrally with respect to their axes seem to be mainly related to lamellar structure
as a whole. In fact, osteon samples loaded off-center are not cylindrical in shape and they fracture.

D. P

IN

T

EST

This is a method for measuring the resistance of osteon walls to internal pressure. It was specially
devised for micromechanical testing and is inspired by the standard pin test for copper and
copper-alloy tubing expansion testing.

26

It consists of loading of an osteon sample by progressive
penetration of a steel cone into the vascular canal until fracture occurs.

27

Longitudinal, alternate, and transverse osteons at both initial and nal degree of calcication
were selected on 100-

m-thick transverse sections. Sections are cut by means of a rotating saw to
ensure constant osteon sample height.
FIGURE 18.7 Illustration of criteria used to investigate osteon sample shearing strength. From top to bottom:
P, punch; BS, bone section; O, osteon; SP, supporting plane; H, hole. (From Ascenzi, A. and Bonucci, E., The
shearing properties of single osteons, Anat. Rec., 172, 499, 1972, The Wistar Institute. Reprinted by permission
of Wiley-Liss, Inc., a division of John Wiley & Sons, Inc.)
0266/ ch18/frame Page 280 Friday, October 22, 1999 8:50 AM
Single Osteon Micromechanical Testing 281
Cylindrically shaped samples are obtained by using the rst technique described in
Section III.B. Uniform thickness of each osteon sample is checked with special care before the
pin test: the outer and inner diameter measurements of each sample are checked at least eight
times at 45 displacement.
Substitution of the punch in the apparatus for osteon testing in compression (Section III.B)
with a steel cone yields an apparatus for the pin test. The angular width of the steel cone measures
35. The choice of 35 allows the most accurate comparison between results, according to the
A. Ascenzi groups experience. Progressive cone loading by means of weights causes the cone to
press with increasing force on the upper circumferential end of the vascular canal. As soon as the
cone displacement caused by an additional weight stopped, additional weights were added. At rst,
osteon deformation occurs and then lamellae increasingly fracture starting with lamellae closest to
the vascular canal. A microwave micrometer based on cavity and pulse technique measures the
progressive descent of the steel cone until ultimate strength is reached. The principle of the
microwave micrometer is described in Section III.A. To allow comparison between loaddeection
data of samples, dilating strength and strain are computed. Dilating strength equals P/(r(R r))
and dilating strain equals d/(R r) where P is the load, r and R are inner and outer osteon radii,
respectively, and d is the radius increase due to cone forward movement.
Transverse, alternate, and longitudinal osteon samples show, respectively, maximum, inter-
mediate, and minimum ultimate dilating strength. This is further direct evidence that the presence
of transverse lamellae increases osteon wall resistance.
E. BENDING TEST
Distribution of longitudinal and transverse lamellae in long bone sections is compatible with the
force distribution usually acting on long bone under normal conditions.
28-33
Femoral lamellar
distribution has been found to vary in accordance with a bending deformity consequent to severe
rickets suffered in infancy.
34
Hence, a relationship exists in femur between modied microscopic
FIGURE 18.8 Fully calcified alternate osteon sample which has slipped out of an axially loaded bone section
during shear testing. (From Ascenzi, A. and Bonucci, E., The shearing properties of single osteons, Anat. Rec.,
172, 499, 1972, The Wistar Institute. Reprinted by permission of Wiley-Liss, Inc., a division of John Wiley
& Sons, Inc.)
0266/ ch18/frame Page 281 Friday, October 22, 1999 8:50 AM
282 Mechanical Testing of Bone and the BoneImplant Interface
structure and macroscopic features under bending. Such results motivated A. Ascenzi et al.
35
to
explore micromechanical properties of a single osteon loaded in bending.
Longitudinal and alternate osteons at nal degree of calcication were selected on 500-m-
thick transverse sections of a femoral diaphysis.
Cylindrically shaped samples are obtained by using the rst technique described in
Section III.B. Sample axis coinciding with the axis of the vascular canal is checked with particular
care. The outer and inner diameter measurements of each sample need to be checked at least eight
times at 45 displacement.
A 500-m-long sample is laid on the edges of two small steel strips carefully placed on exactly
the same horizontal plane. Since the distance between the edges measures 400 m, an osteon sample
is supported by each surface edge over a length of 50 m. In this way, there are no constraints to
angular rotation of the sample at the points of support, because both sample ends are free to move
while the sample is loaded in bending (Figure 18.9a). The central part of the sample is loaded by
means of a steel point which has linear contact with the sample (Figure 18.9b). Avoidance of
longitudinal or rotational displacement of the sample relative to the point during loading is necessary
for testing accuracy.
A microwave micrometer based on the cavity and pulse technique measures the advancement
of the steel point. The principle of the microwave micrometer is described in Section III.A. Meas-
urement of changes in bending is accurate within 1%.
Note that because the length of the sample is short with respect to the its external diameter, a
large shearing stress is generated at the failure site. Such shearing stress makes for an improper
bending test. However, this occurrence seems negligible because the aim of the present investigation
is to provide comparative, as opposed to absolute, behavior in bending.
Ultimate bending load, ultimate bending deformation, elastic modulus, and rupture modulus
are either obtained experimentally or computed from experimental data by means of beam theory.
Longitudinal osteon samples are less able to withstand bending than alternate osteon samples.
An interpretation resides in the very few transverse collagen bundles present in longitudinal
osteons. In other words, alternate osteon samples are stiffer in bending because they are struc-
turally more compact.
F. TORSION TEST
Frasca et al.
7,36
developed a method for torsional loading of whole osteons and A. Ascenzi et al.
37
developed a method for torsional loading of osteon samples.
FIGURE 18.9 (a) Fully calcified alternate osteon sample ready to be loaded in bending. (b) Fully calcified
alternate osteon sample at ultimate bending strength. (From Ascenzi, A., Baschieri, P., and Benvenuti, A., The
bending properties of single osteons, J. Biomech., 23, 763, 1990. With permission from Elsevier Science.)
0266/ ch18/frame Page 282 Friday, October 22, 1999 8:50 AM
Single Osteon Micromechanical Testing 283
1. Frasca Method
Single osteons and osteon groups were chosen without regard to structural type and degree of
calcication. Single osteons and osteon groups measuring approximately 5 to 8 mm in length are
dissected from longitudinal and transverse bone sections by a method which separates each sample
at its natural boundaries. Microfractures are produced and slowly propagated along the boundaries
of each sample using scalpel and tweezers under a 30 stereoscopic microscope. Samples are tested
wet and dry. A laboratory-built microtorsional device is used for testing. In essence, it consists of an
R-F coil immersed in a 5000 gauss magnetic eld and connected to an oscillator. Angular amplitudes
are measured by reecting a laser light spot from a rotating mirror. Results indicate strain and frequency
dependence of shear storage modulus of single osteons and osteon groups. They establish differences
in wet and dry torsional behavior of both single osteons and osteon groups.
2. A. Ascenzi Method
A. Ascenzi et al.
35
have found that it is necessary to prepare the sample very accurately and to use
a highly sensitive apparatus. Longitudinal and alternate osteons at nal degree of calcication were
tentatively selected on longitudinal sections 300 to 350 m thick. Denitive identication of osteon
type is possible only after loading. Osteon samples are isolated from sections by means of the
second method described in Section III.B. In addition to usual sample regularity requirements, each
sample needs to be carefully examined under an optical microscope to exclude the possibility of
interference of small surface defects with experimental results. A special torsional device was
manufactured according to researchers specications by the CECOM Company (Figure 18.10).
This device consists of a xed axis and a rotational axis horizontally aligned; each axis is furnished
with a set of jaws that grips the osteon sample during testing. The center of each set of jaws lies
on the associated axis. By means of a stereoscopic microscope, the rotational axis is veried to
coincide with the osteon axis; i.e., the center of each jaw needs to correspond to one end of the
axis of the osteon sample canal. The jaws are not free to move axially. This establishes an axial
loading effect which could inuence absolute measurements, but here such loading effect may be
neglected because this investigation concerns comparative measurements.
FIGURE 18.10 Simplified diagram of torsional device. Fixed axis (FA) and rotation axis (RA), each with
its jaw (J) for securing sample to device; two hard metal wedges of pendulum loading system (MW); wheel
(W) around which the tungsten thread with weights is attached; axis of pendulum (AP); mirror (M). (From
Ascenzi, A., Baschieri, P., and Benvenuti, A., The torsional properties of single selected osteons, J. Biomech.,
27, 875, 1994. With permission from Elsevier Science.)
0266/ ch18/frame Page 283 Friday, October 22, 1999 8:50 AM
284 Mechanical Testing of Bone and the BoneImplant Interface
One set of jaws is xed, while the other turns in synchrony with a wheel measuring 61 mm in
diameter. To reduce the rotating friction of the turning jaw to a negligible minimum, a pendulum
loading system is adopted whose fulcrum is the tip of two hard metal wedges. The maximum
oscillation of the pendulum is 55. A tungsten thread, whose section measures 20 m in diameter,
winds around the rim of the wheel as 0.1 g weights are attached one after the other at one end of
the tungsten thread until sample failure occurs. The interval between the application of two
consecutive weights is kept constant at 4 s. The weight which produces an unchecked twisting of
the sample; i.e., a sudden nonstop acceleration of the system, is considered to be the one responsible
for osteon failure. The angles through which a specimen twists during a test are measured optically;
such optical measure is based on the reection of a laser beam from a small mirror inserted in the
rotating set of jaws. The variations in the torsional angle are read on a graduated scale placed
160 cm from the device.
Precision and accuracy of the graduated scale and apparatus coincide, as checked by applying
standard experimental procedures.
Ultimate torque, ultimate angular deection, shear modulus, energy absorbed up to failure, and
torsional shear stress are either obtained experimentally or computed from experimental data by means
of beam theory. Ultimate torque and shear modulus values are higher for longitudinal osteon samples
than alternate osteon samples. Moreover, osteon shear modulus resulting from such testing is as much
as four times larger than shear modulus of macroscopic samples.
38,39
Lakes
40
explains that such moduli
are actually in agreement through Cosserat elasticity theory. More precisely, slender specimens are
stiffer than thick ones and the lower stiffness in thick specimens is attributed to localized slippage at
the cement lines. Cosserat elasticity theory is a suitable model to represent presence (absence, respec-
tively) of slippage under torsion in macrosamples (osteons, respectively) because it allows a moment
per unit area in addition to the usual force per unit area of classic elasticity theory.
G. CYCLIC LOADING IN TENSION AND COMPRESSION
Availability of the technique which provides a cylindrical osteon sample with rectangular cross-
sectional ends suggested to A. Ascenzis research group to load osteon samples cyclically in tension
and compression and to investigate their hysteresis loops.
41,42
Longitudinal and alternate osteons at
nal degree of calcication were tentatively selected on longitudinal sections 300 to 350 m thick.
Denitive identication of osteon type is possible only after loading. The technique of Section III.A
is used here.
A. Gozzini, late holder of the chair of Solid State Physics at the Scuola Normale Superiore,
Pisa, designed and assembled the apparatus. In essence, it comprises two parts: (1) a microwave
micrometer based on cavity and pulse technique; (2) an electromechanical device functioning as a
transducer to allow sample cyclic loading.
The microwave micrometer is an improved version of previously described apparatuses for
loading in tension and in compression. The working frequency, f
o
, was reduced from 24 GHz (l =
1.25 cm) to 14 GHz (l = 2.2 cm). A higher-quality klystron, which generates the microwaves, and
crystal detectors had become available by 1985 in the latter frequency region; they ensure an
increase in sensitivity and stability despite the increase in cylinder height from 1.78 to approximately
2 cm. The two hollow cylinders of the apparatus (one acting as a measuring cavity and the other
used for comparison) were obtained from a single copper block to eliminate temperature differences
between them and to minimize mechanical noises.
The transducer causing changes in cavity height (l) was replaced by a more sophisticated
electromechanical device whose description follows. Figure 18.11 shows a diagram of the electro-
mechanical device. It consists of a measuring cavity (C
m
) with two terminal walls: one terminal
wall is mobile and can either penetrate into or withdraw from the cavity. The terminal wall is set
in motion by an asynchronous motor (AM) whose rotating direction is reversible. The motor allows
rotation of a metal bar (B) bearing both right and left screw threads connected to form a continuous
0266/ ch18/frame Page 284 Friday, October 22, 1999 8:50 AM
Single Osteon Micromechanical Testing 285
track. As result of bar rotations, a pin (P) inserted into the screw track undergoes alternating
movements. The pin exerts alternating compression and decompression on each spring (T and T
1
)
coiled around the metal bar.
The coiled springs act on a slide (S) attached to the mobile terminal wall of the cavity C
m
. The
slide moves on two rectilinear rails of cemented steel through four ball bearings. One osteon sample
lug is secured to the mobile part (R) and the other lug to the immobile part (H) of the device.
Therefore, the moving slide exerts tension (compression, respectively) on the sample as the spring
T (T
1
, respectively) undergoes compression.
When tension (compression, respectively) is applied, the sample elongates (shortens, respec-
tively), the mobile wall of the cavity advances (recedes, respectively), and cavity height decreases
(increases, respectively). Since the length change of the sample is equal in magnitude to the cavity
height change, the length change of the sample is deduced from the corresponding resonance
frequency change of the cavity by means of Equation 18.2 of Section III.A.
Reduction of slide friction to a minimum and careful preparation of jaws to achieve perfect
alignment between sample and jaws without sliding are crucial to accuracy.
Here the load limit is deduced from preliminary trials so that the stress limit, which varies from
osteon to osteon because of the section variability, approximately corresponds to a middle value
between the proportional stress limit and ultimate strength. An electronic system, which carries out
the following operations, was devised: (1) recording of complete turns and fraction of turns;
(2) recording of turns of the motor for each displacement in either direction; (3) memorization and
comparison of computed data; (4) transmission of an electric signal to reverse motor rotation; and
(5) provision to a recorder Y-axis of an electronic potential difference, proportional to the displace-
ment and consequently to the load acting on the sample.
Sample deection is registered by transformation of pulses coming from the two resonance
cavities C
m
and C
c
into a difference of electric potential. Because the apparatus can detect fractional
changes in resonance frequency on the order of 10
8
, sample length changes as small as a few
ngstroms can be measured. Since bone sample measurements require a sensitivity much lower
than a few ngstroms, high accuracy is assured.
FIGURE 18.11 Simplified diagram of mechanical component of electromechanical device for osteon sample
tensioncompression cyclic loading and transformation of sample length changes into frequency variations of
electromagnetic waves in resonator cavities. O, osteon sample with its lugs into securing jaws H and R; C
m
,
measuring cavity; C
c
, comparing cavity; AM, asynchronous motor; B, bar; P, pin; T and T
1
, coiled springs;
S, slide. (From Ascenzi, A., Benvenuti, A., Mango, F., and Simili, R., Pinching in longitudinal and alternate
osteons during cyclic loading, J. Biomech., 18, 391, 1985. With permission from Elsevier Science.)
0266/ ch18/frame Page 285 Friday, October 22, 1999 8:50 AM
286 Mechanical Testing of Bone and the BoneImplant Interface
A direct check for the apparatus is provided by loading of thin calibrated samples of a variety
of material with predictable hysteresis loops, under the same experimental conditions as osteon
samples. In addition, the conclusion is drawn that incidental variations in temperature do not
appreciably affect test results.
The investigation of the A. Ascenzi group was performed with a mean loadingunloading cycle
duration of 105 s, mean load rate of 0.05 N/s, and maximum load limit in both tension and
compression of 1 N. The cycles applied to each sample varied in number; they were interrupted
before spontaneous specimen rupture. Interruption is necessary because denitive identication of
osteon type is possible only before rupture, but after testing.
A mathematical model of the hysteresis loops of osteon samples served to show that, as the
number of cycles increases, strain limit increases, stiffness degrades, and energy absorption
increases, all because of the increasing extent of lesions and possible increasing buckling. At equal
degrees of calcication, the strain limit is greater in compression than in tension for longitudinal
osteon samples because the high number of longitudinal bers renders longitudinal osteon samples
more resistant to tension than compression. The strain limit is smaller in compression than in
tension for alternate osteon samples because the transverse and oblique collagen bundles minimize
buckling and render alternate osteon samples more resistant to compression than tension.
In addition to stiffness degradation, another degradation effect occurs. Hysteresis loops are S-
shaped or pinched (Figures 18.12 and 18.13) and such pinching degrades as the number of cycles
increases. Pinching is due to temporary repair under compression and resolution under tension of
lesions, and consequent abrupt changes in stiffness. Pinching degradation is less signicant and
energy absorption is more signicant for longitudinal than alternate osteon samples. An interpre-
tation is that, in longitudinal osteons, the simpler structure allows for easier repair and resolution
of lesions while the lack of transverse collagen bundles increases the extent of the lesions.
FIGURE 18.12 (a) Idealized curvilinear hysteresis model of curve prior to cycling and at the first cycle.
Compressive half-cycle BE and tensional half-cycle EH are S-shaped or pinched as consequence of load
reversal in both compression and tension. (b) Idealized bilinear hysteresis model of curve prior to cycling and
at the rst cycle used for structural interpretations. (From Ascenzi, A., Benvenuti, A., Bigi, A., et al., Calcif.
Tissue Int., 62, 266, 1998. With permission from Springer-Verlag, New York.)
0266/ ch18/frame Page 286 Friday, October 22, 1999 8:50 AM
Single Osteon Micromechanical Testing 287
IV. ACQUIRED KNOWLEDGE AND PENDING QUESTIONS
Analysis of secondary compact bone focused on the learning of osteon function through mechanical
testing of osteonic lamellar structure is painstaking and time-consuming research which would be
of limited value if not justied by a large breadth of useful results. Although this chapters focus
is mechanical testing of single osteons, a few words to summarize results to date and to list pending
questions are included to stimulate researchers interest.
1. The countless types of human osteons result from arrangements of longitudinal and
transverse lamellae. Longitudinal (transverse, respectively) lamellae appear dark (bright,
respectively) in cross section under a polarizing microscope. Longitudinal lamellae resist
tension and torsion better than transverse lamellae. Under compression, shearing, and
bending, transverse lamellae resist loading better than longitudinal lamellae.
2. The distribution of osteonic lamellae and interstitial bone in cross sections of long bone
shafts is not random.
3. The distribution of osteonic lamellae in cross sections of long bone shafts shows a pattern
compatible with the shaft shape and the distribution of forces usually operating on long
bone under both normal and pathological condition. In general terms, there is indication
of a high incidence of longitudinal (transverse, respectively) lamellae in bone sectors
loaded in tension (compression, respectively).
FIGURE 18.13 Stressstrain hysteresis loop of longitudinal osteon. Strain limit increases more rapidly and
maximum curvature is smaller during compression than tension. (From Ascenzi, A., Benvenuti, A., Mango,
F., and Simili, R., Pinching in longitudinal and alternate osteons during cyclic loading, J. Biomech., 18, 391,
1985. With permission from Elsevier Science.)
0266/ ch18/frame Page 287 Friday, October 22, 1999 8:50 AM
288 Mechanical Testing of Bone and the BoneImplant Interface
4. Under normal conditions, the progressive increase of longitudinal lamellar concentra-
tion from the inner to the outer edge of a section

is hypothesized as related to a condition
of initial stress that would reduce the extent of the response of long bones to forces
acting on them.
5. Variation in osteonic lamellar distribution of a long bone shaft is a sophisticated
process, as proved by slight structural asymmetries found in shafts of subjects who
had shown some simple asymmetric use of limbs during life.
6. The high concentration of transverse lamellae along the elongated sides of elliptical
cross sections of a pathologically bowed shaft increases the bending stiffness of the
deformed bone.
7. Pending questions are (a) the relation between osteonic lamellar distribution and
geometry of the bone shaft loaded in bending and torsion, with respect to the mechan-
ical properties of longitudinal and transverse lamellae; (b) the relation between osteonic
lamellar distribution and geometry of the bone shape in determining fractures, espe-
cially in bending, with respect to the mechanical properties of longitudinal and trans-
verse lamellae; and (c) the structural classication of osteons in diaphysis of
vertebrates, whose lamellar structure somewhat differs from that of humans, to increase
accuracy of interpretations.
43,44
ACKNOWLEDGMENTS
Research Grant 97.03992.04 from the National Research Council of Italy supported this work. The
authors thank Lucio Virgilii for gure preparation.
REFERENCES
1. Ascenzi, A., The micromechanics vs. the macromechanics of cortical bone A comprehensive
presentation, J. Biomech. Eng., 110, 357, 1988.
2. Ascenzi, A., Boyde, A., Portigliatti-Barbos, M., and Carando, S., Micro-biomechanics vs. Macro-
biomechanics in cortical bone. A micromechanical investigation of femurs deformed by bending, J.
Biomech., 20, 1045, 1987.
3. Ascenzi, A., Biomechanics and Galileo Galilei, J. Biomech., 26, 95, 1993.
4. Evans, F.G., Mechanical Properties of Bone, Charles C Thomas, Springeld, IL, 1973.
5. Petersen, H., Die Organe des Skeletsystems, in Handbuch der mikroskopischen Anatomie des Menschen,
Mllendorff (v.), W., Ed., Springer, Berlin, 1930.
6. Ortner, D.J. and Yong, D., A precision microdissection procedure for undecalcied bone thin sections,
Calcif. Tissue Res., 17, 169, 1975.
7. Frasca, P., Harper, R.A., and Katz, J.L., Isolation of single osteons and osteon lamellae, Acta Anat.,
95, 122, 1976.
8. Frasca, P., Harper, R.A., and Katz, J.L., Collagen ber orientation in human secondary osteons, Acta
Anat., 98, 1, 1977.
9. Ascenzi, A. and Benvenuti, A., Orientation of collagen bers at the boundary between two successive
osteonic lamellae and its mechanical interpretation, J. Biomech., 19, 455, 1986.
10. Ascenzi, A. and Benvenuti, A., The revisited osteon, in preparation.
11. Giraud-Guille, M.-M., Twisted plywood architecture of collagen brils in human compact bone
osteons, Calcif. Tissue Int., 42, 167, 1988.
12. Gebhardt, W., ber funktionell wichtige Anordnungsweisen der feineren und grberen Bauelemente
des Wirbeltierknochens. II. Spezieller Teil 1. Der Bau der Haversschen Lamellensysteme und seine
funktionelle Bedeutung, Arch. Entwmech. Org., 20, 187, 1906.
13. Amprino, R., Reported in Trattato di Istologia, Levi, G., Ed., Unione Tipograco-Editrice Torinese,
Turin, 1946, 513.
0266/ ch18/frame Page 288 Friday, October 22, 1999 8:50 AM
Single Osteon Micromechanical Testing 289
14. Currey, J.D., The effect of drying and re-wetting on some mechanical properties of cortical bone,
J. Biomech., 21, 439, 1988.
15. Ascenzi, A., Bonucci, E., and Checcucci, A., The tensile properties of single osteons studied using a
microwave extensometer, in Studies on the Anatomy and Function of Bone and Joints, Evans, F.G.,
Ed., Springer-Verlag, Berlin, 1966.
16. Ascenzi, A. and Bonucci, E., The tensile properties of single osteons, Anat. Rec., 158, 375, 1967.
17. Ascenzi, A. and Bonucci, E., Mechanical similarities between alternate osteons and cross-ply lami-
nates, J. Biomech., 9, 65, 1976.
18. Ascenzi, A. and Benvenuti, A., Evidence of a state of initial stress in osteonic lamellae, J. Biomech.,
10, 447, 1977.
19. Ascenzi, A. and Benvenuti, A., Evidence of a state of initial stress in osteonic lamellae, Acta Orthop.
Belg., 46, 580, 1980.
20. Ascenzi, M.-G., A rst estimation of prestress in so-called circularly bered osteonic lamellae,
J. Biomech., 32, 935, 1999.
21. Ascenzi, M.-G., A rst estimate of prestress in so-called circularly bered osteonic lamellae, Abstracts
of the 11th conference of the European Society of Biomechanics, J. Biomech., 31 (Suppl. 1), 22, 1998.
22. Battaglia, A., Bruin, F., and Gozzini, A., Microwave apparatus for the measurement of the refraction
dispersion and absorption of gases at relatively high pressure, Nuovo Cimento, 7, 1, 1958.
23. Montgomery, G., Technique of Microwave Measurements, McGraw-Hill, New York, 1947.
24. Ascenzi, A. and Bonucci, E., The compressive properties of single osteons, Anat. Rec., 161, 377, 1968.
25. Ascenzi, A. and Bonucci, E., The shearing properties of single osteons, Anat. Rec., 172, 499, 1972.
26. ASTM, Standards expansion (pin test) of copper and copper-alloy tubing, Designation: B 15358, 3,
17, 1958.
27. Ascenzi, A. and Bonucci, E., Relationship between ultrastructure and pin test in osteons, Clin.
Orthop., 121, 275, 1976.
28. Portigliatti-Barbos, M., Bianco, P., and Ascenzi, A., Distribution of osteonic and interstitial compo-
nents in the human femoral shaft with reference to structure, calcication, and mechanical properties,
Acta Anat., 15, 178, 1983.
29. Boyde, A., Bianco, P., Portigliatti-Barbos, M., and Ascenzi, A., Collagen orientation in compact bone:
I. A new method for the determination of the proportion of collagen parallel to the plane of compact
bone sections, Metab. Bone Dis. Relat. Res., 5, 299, 1984.
30. Portigliatti-Barbos, M., Bianco, P., Ascenzi, A., and Boyde, A., Collagen orientation in compact bone:
II. Distribution of lamellae in the whole of the human femoral shaft with reference to its mechanical
properties, Metab. Bone Dis. Relat. Res., 5, 309, 1984.
31. Carando, S., Portigliatti-Barbos, M., Ascenzi, A., and Boyde, A., Orientation of collagen in human
tibial and bular shaft and possible correlation with mechanical properties, Bone, 10, 139, 1989.
32. Carando, S., Portigliatti-Barbos, M., Ascenzi, A., et al., Macroscopic shape of, and lamellar distribution
within, the upper limb shafts, allowing inferences about mechanical properties, Bone, 12, 265, 1991.
33. Portigliatti-Barbos, M., Carando, S., Ascenzi, A., and Boyde, A., On the structural symmetry of human
femurs, Bone, 8, 165, 1987.
34. Ascenzi, A., Improta, S., Portigliatti-Barbos, M., et al., Distribution of lamellae in human femoral
shafts deformed by bending with inferences on mechanical properties, Bone, 8, 319, 1987.
35. Ascenzi, A., Baschieri P., and Benvenuti, A., The bending properties of single osteons, J. Biomech.,
23, 763, 1990.
36. Frasca, P., Harper, R.A., and Katz, J. L., Strain and frequency dependence of shear storage modulus
for human single osteons and cortical bone microsample-size and hydration effects, J. Biomech., 14,
679, 1981.
37. Ascenzi, A., Baschieri P., and Benvenuti, A., The torsional properties of single selected osteons,
J. Biomech., 27, 875, 1994.
38. Pfafrod, G.O., Knets, I.V., Saulgozis,Y.Z, et al., Age-related changes in the strength of compact bone
tissue under torsion, Polym. Mech., 3, 493, 1975 [in Russian].
39. Evans, P.G., Relations between torsion properties and histology of adult human compact bone,
J. Biomech., 11, 157, 1978.
40. Lakes, R., On the torsional properties of single osteons, J. Biomech., 28, 1409, 1995.
0266/ ch18/frame Page 289 Friday, October 22, 1999 8:50 AM
290 Mechanical Testing of Bone and the BoneImplant Interface
41. Ascenzi, A., Benvenuti, A., Mango, F., and Simili, R., Mechanical hysteresis loops from single osteons:
technical devices and preliminary results, J. Biomech., 18, 391, 1985.
42. Ascenzi, A., Ascenzi M.-G., Benvenuti, A., and Mango, F., Pinching in longitudinal and alternate
osteons during cyclic loading, J. Biomech., 30, 689, 1997.
43. Riggs, C.M., Lanyon, L.E., and Boyde, A., Functional associations between collagen bre orientation
and locomotor strain direction in cortical bone of the equine radius, Anat. Embryol., 187, 231, 1993.
44. Riggs, C.M., Vaughan, L.C., Evans, G.P., et al., Mechanical implications of collagen bre orientation
in cortical bone of the equine radius, Anat. Embryol., 187, 239, 1993.
0266/ ch18/frame Page 290 Friday, October 22, 1999 8:50 AM

291

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

19

Micromechanical Testing
of Single Trabeculae

Peter L. Mente

CONTENTS

I. Introduction ..........................................................................................................................291
II. Testing Methods ...................................................................................................................292
A. Three-Point Bending Tests.............................................................................................292
B. Tensile Testing................................................................................................................294
C. Cantilever Beam Testing ................................................................................................295
III. Testing Considerations .........................................................................................................300
A. Specimen Geometry .......................................................................................................300
B. Grips and Edge Effects ..................................................................................................301
C. Tissue Hydration ............................................................................................................302
D. Apparatus Checks...........................................................................................................302
IV. Results ..................................................................................................................................302
References ......................................................................................................................................304

I. INTRODUCTION

Trabecular (or cancellous) bone is the relatively porous bony tissue found in the marrow cavities
at the ends of the long bones and within the vertebrae and other small bones; the trabecular bone
is encased by a denser and more highly structured cortical bone shell. The individual trabeculae
that make up the trabecular tissue tend to have one of two basic shapes, either slender rods (or
beamlike struts) or relatively thin, at plates. These struts and plates form a three-dimensional
scaffolding within the cortical shell. Structurally, trabecular bone helps to distribute loading from
the articular surfaces to the cortical shell. Trabecular tissue is also very important for the stability
and xation of joint implants and other xation devices placed into bone.
Because of its structural importance, many studies have focused on measuring the material
properties of trabecular tissue. The majority of these studies have tested cubic or cylindrical
specimens with dimensions on the order of 5 to 10 mm, and therefore have measured the bulk
material properties of the tissue (see Chapter 11). The bulk material property combines the contri-
butions of trabecular architecture (shape and spatial distribution of the trabecular scaffolding),
porosity of the tissue, and the material properties of the bony tissue itself. It is important to make
a distinction between the trabecular bone bulk modulus and the tissue modulus. Throughout this
chapter,

tissue modulus

will be used to refer to the modulus of the bony material that makes up
individual trabeculae and

bulk modulus

will refer to the modulus calculated from larger samples
(including the effects of architecture and porosity). Originally, many models of trabecular bone
assumed that the tissue modulus of trabecular bone was the same as the tissue modulus of cortical
bone and that the bulk tissue stiffness of trabecular bone could thus be modeled by assuming that
trabecular bone was simply a more porous cortical bone tissue. As models of trabecular bone
become more rened and begin to include the internal trabecular architecture,

1-7

knowledge of the

0266/ ch19/frame Page 291 Friday, October 22, 1999 9:14 AM

292

Mechanical Testing of Bone and the BoneImplant Interface

mechanical properties of the trabecular tissue has become more important. In order to determine
if trabecular bone tissue is different from cortical bone tissue, and to understand how this might
affect these models of the trabecular tissue, it is necessary to be able to perform mechanical testing
on individual trabeculae and measure the modulus of the trabecular tissue itself.
The modulus of cortical bone tissue has been characterized extensively, in part because its size
and relatively solid nature allow specimens to be machined into relatively standard test congura-
tions for tensile and compressive testing (see Chapter 11). This, however, is not possible with
trabecular bone tissue; the extremely small size and ill-dened shape of individual trabeculae make
determining the modulus of trabecular tissue an extremely difcult task. Testing of these small,
irregularly shaped specimens requires the use of nonstandard measurement techniques. In this
chapter mechanical testing methods that have been employed to test individual trabeculae will be
examined. These nonstandard techniques generally require specially built testing apparatus and
special consideration of the specimen geometry. The rst attempts to determine the modulus of
individual trabeculae used a buckling analysis where the modulus was determined based on the
critical buckling load during the controlled compression of individual trabecular specimens.

8,9

More
recent testing methods which this chapter will cover have used more conventional engineering
beam analysis to determine the modulus of this tissue. Three different mechanical testing techniques
will be discussed: three-point bending, tensile testing, and cantilever beam testing. The three-point
bending tests and tensile testing techniques use specimens of known or assumed geometry which
allow the calculation of the tissue modulus using general engineering beam equations. The cantilever
beam method uses a nite-element model of the test in order to determine the modulus of the
specimen by matching model and experimental results.

II. TESTING METHODS
A. T

HREE

-P

OINT

B

ENDING

T

ESTS

Kuhn et al.

10

and Choi et al.

11

tested rectangular beam specimens of trabecular bone in three-point
bending to measure their tissue modulus. Their trabecular beam specimens were created by micro-
machining individual trabeculae down to a uniform rectangular cross section. To obtain trabecular
specimens with uniform cross sections, blocks of bulk trabecular bone were rst sectioned to a
thickness of 75 to 100

m using a Buehler Isomet and a low-speed diamond cutting blade (Buehler,
Lake Bluff, IL), creating sections of tissue with two at and parallel faces. Trabeculae that were
continuous and homogeneous were dissected from the sections and placed in the holding xture
of a specially designed miniature milling machine using pressure plates lined with sandpaper. In
the micromilling machine the two remaining uncut sides of the trabeculae were milled parallel to
each other to a thickness of 50 to 200

m. Specimens obtained in this manner had an overall length
on the order of 1500 to 2000

m. Cutting and machining was done under irrigation to prevent the
specimens from drying out. Specimen dimensions were conrmed using digitized photographs of
the trabecular beams taken at a magnication of 60

.
In a three-point bending test, a specimen is placed across two supports and a force is applied
midway between the supports, causing the specimen to deect (Figure 19.1). The modulus of beams
tested in this manner can be calculated using the equation:
(19.1)
where

E

= Youngs (tissue) modulus,

P

= applied load,

= displacement at the point of application
of the load (midpoint of the beam),

L

= distance between the supports,

I

= specimen moment of
inertia (for a rectangular specimen this is equal to

bh

3

, where

b

is the width of the specimen
and

h

is its thickness).
E
P L
I

3
48

0266/ ch19/frame Page 292 Friday, October 22, 1999 9:14 AM

Micromechanical Testing of Single Trabeculae

293

Because relatively slender specimens were tested, Kuhn et al.

10

and Choi et al.

11

employed a
modication of this general equation that includes terms that take into account the concentrated
nature of the applied force.

12

(19.2)
The testing apparatus used in these two studies was a modied microscope base (Figure 19.2).
The two supports were xed to the top of a load cell approximately 1 mm apart and the load cell
in turn was attached to the specimen stage of the microscope. The loading head (positioned midway
between the supports) was attached to the microscope stationary body tube. Specimens were
deformed by moving the specimen stage up against the stationary loading head, using a stepper
motor attached to the ne adjustment knob of the microscope. This allowed the specimens to be
loaded at a controlled rate. Rates used in these tests were 0.081 mm/s

10

and 0.11 mm/s.

11

Vertical
displacement of the stage relative to the loading head (which represents the specimen displacement
at the point of application of the load) was measured using a linear variable differential transformer
(LVDT) with a full-scale range of

2.5 mm; the load cell that was used had a maximum compressive
range of 0.6 N. The maximum load applied in the experiments was 0.2 to 0.3 N, and the maximum
resulting specimen deection at the point of application of the force was approximately 0.1 mm.
The applied force (

P

) and specimen deection at the point of application of the force (

) were
acquired by computer and the slope (

P

/

) calculated by tting a straight line to the experimental
data. The experimentally measured slope, along with the measurements of the specimen geometry,
allowed the trabecular tissue modulus to be determined from Equation 19.2.

FIGURE 19.1

Schematic diagram of a three-point bending test geometry.

FIGURE 19.2

Schematic diagram of the three-point bending test apparatus used by Kuhn et al.

10

and Choi
et al.

11

(From Choi, K., Kuhn, J.L., Ciarelli, M.J., and Goldstein, S.A.,

J. Biomech.,

23, 1103, 1990. With
permission.)
E
P L
I
h
L
h
L
+

_
,

_
,

1
]
1

3
2 3
48
1 2 85 0 84 . .

0266/ ch19/frame Page 293 Friday, October 22, 1999 9:14 AM

294

Mechanical Testing of Bone and the BoneImplant Interface

B. T

ENSILE

T

ESTING

Tensile tests on individual trabeculae have been performed by two groups, Ryan and Williams,

13

and Rho et al.

14

Standard tensile test specimens are normally machined or cut along the middle
segment of their free length so that stress and strain measurements can be made on an area of
reduced cross-sectional area. In this way, measurements are made well away from the xed ends,
where edge effects and stress concentrations due to the grips inuence the general homogeneous
stress state. The modulus for a standard specimen tested in tension is given by
(19.3)
where

E

= Youngs (tissue) modulus,

P

= applied load,

L

= reference length,

= elongation of the
reference length

L,

and

A

= specimen cross-sectional area.
Both groups tested rodlike specimens excised from the bulk trabecular tissue using a scalpel
under a dissecting microscope. Specimens were chosen to be as straight and uniform as possible.
The specimens tested by Ryan and Williams

13

were approximately 0.3 mm in diameter with lengths
on the order of 4 mm. Rho et al.

14

tested specimens with diameters and lengths of approximately
0.18 and 2.3 mm, respectively. The difference in size of tested specimens reects the species and
location from which specimens were obtained; Ryan and Williams

13

tested trabeculae from the
distal end of bovine femurs and Rho et al.

14

tested specimens collected from a proximal human tibia.
Because individual trabeculae are so small, test specimens cannot be easily machined to create an
area of reduced cross section, nor can specimens be obtained that are long enough for displacement
changes to be measured away from the grips to avoid edge effects. Relative displacement measurements
on trabeculae specimens tested in tension have therefore been made over the entire specimen length,
using grip displacement as a measurement of specimen elongation. The cross-sectional area measure-
ment used for these specimens represents the average cross-sectional area along the length. Although
the cross-sectional area is generally elliptical with a major and minor radius, for modeling purposes
the cross-sectional area was assumed to be circular with a constant radius over the entire free length
between the grips. Ryan and Williams

13

noted that at a given cross section, the maximum and minimum
radius could differ by as much as a factor of 2, although both studies indicated that measured specimen
diameters varied by less than 10% along their test length.
Ryan and Williams

13

tested their specimens by clamping each end of their rodlike specimens
between plates lined with 320-grit sandpaper (Figure 19.3). A set screw between the pair of grips
set the free length at 1.524 mm for all specimens. The specimens were placed in the grips so that
their longitudinal axes were aligned as closely as possible with the direction of loading. Two guide
rods running between the grips maintained the orientation of the specimen and helped to prevent
bending of the specimens during testing. Specimens were distracted at a strain rate of approximately
0.001

/s. A load cell (maximum load 6 N) and an extensometer (maximum displacement 2.5 mm)
attached across the two grips were used to measure the force and specimen elongation, respectively.
Specimen elongation ranged from 0.04 to 0.32 mm and the applied tensile force ranged from 2.2
to 5.5 N. The resulting forcedisplacement curves were plotted on an

xy

plotter and the slope of
the curves (

P

/

) measured, using a straight line drawn on the graph, to determine specimen stiffness.
The specimen stiffness along with the measured geometry (specimen free length and cross-sectional
area) were then used in Equation 19.3 to calculate the tissue modulus.
The grips used in the tensile testing performed by Rho et al.

14

were rods with hollow ends
into which the ends of the trabeculae were xed using a cyanoacrylate glue (Figure 19.4). This
attachment method required the specimens to be dry to ensure a good glue bond (specimens
were not rewetted before testing). Specimens were aligned in the grips under a stereomicroscope
to ensure they were straight and to minimize the bending moments that would develop from
nonaxial loading. Specimens were loaded at a strain rate of 5.5

/s using a motor geared to a
slider that advanced the grip (see Figure 19.4). A purpose-built load cell was used to measure
E
P L
A

0266/ ch19/frame Page 294 Friday, October 22, 1999 9:14 AM

Micromechanical Testing of Single Trabeculae

295

load and an extensometer (maximum range 3 mm) attached to the two grip rods was used to
measure specimen elongation. Load and elongation were recorded on an

xy

plotter; maximum
applied loads were on the order of 1 N and maximum specimen elongation was approximately
0.01 mm. As with Ryan and Williams,

13

the slope of the forcedisplacement curve (

P

/

) along
with the measured geometry (specimen free length and cross-sectional area) were used in
Equation 19.3 to calculate the tissue modulus.

C. C

ANTILEVER

B

EAM

T

ESTING

The third testing method that has been used to measure trabecular tissue modulus is cantilever
beam testing.

15

A cantilever beam has one end xed and the other end free, and loads are applied
along the length of the specimen perpendicular to the long axis of the specimen, causing the
specimen to bend. Trabecular bone specimens were tested in this conguration using a two-step
procedure: rst individual trabeculae were mechanically loaded and their bending response
recorded, then a three-dimensional nite-element model of the trabeculae was created and subjected

FIGURE 19.3

(A) Schematic diagram of the tensile testing apparatus used by Ryan and Williams.

13

(From
Ryan, S.D., Tensile Testing of Individual Bovine Trabeculae, Masters thesis, Northeastern University, Boston,
1988. With permission). (B) A detail of the grip assembly used to hold specimens. (From Ryan, S.D. and
Williams, J.L., Tensile testing of rodlike trabeculae excised from bovine femoral bone,

J. Biomech.

, 22, 351,
1989. With permission.)

0266/ ch19/frame Page 295 Friday, October 22, 1999 9:14 AM

296

Mechanical Testing of Bone and the BoneImplant Interface

to simulated test conditions. The tissue modulus was determined by changing the modulus of the
trabecular bone elements in the nite-element model until the displacement of the model matched
the displacement of the specimen measured during experimental testing, for a given applied load.
Since nite-element models were created for each trabecular specimen tested, it was not
necessary to select specimens with a uniform cross section; in fact, both at, platelike specimens
and rodlike specimens could be tested using this methodology. Specimens of trabecular bone were
obtained by cutting open the end of a long bone to expose the internal trabecular scaffolding.
Sections of the bulk trabecular tissue were cut from the bone using a scalpel and the marrow was
removed using a water pick. When the marrow was removed, the trabecular scaffolding could be
visualized and individual trabeculae excised using a scalpel. The main selection criteria for spec-
imens was overall length. Side struts on the trabeculae that did not contribute to the bending stiffness
were trimmed, to eliminate features that would add complexity to the nite-element model. For
platelike trabeculae, the specimen width was trimmed to a maximum of 4.0 mm so that specimens
would t in the test apparatus. Trimming of the platelike specimens was also used to eliminate
holes and other defects in the specimens to facilitate later modeling; specimen thickness was not
altered, only specimen width.
Epoxy bases to hold the trabeculae were created using 8-mm-diameter electron microscopy
specimen capsules. Epoxy (Shell Epon 828 and Hysol HD 3404 hardener in a 10 : 1 ratio) was
poured into the capsules and allowed to set. Once the epoxy was hardened, the top surface was
squared off and the top 3 mm turned down to a diameter of 5 mm. The remaining ledge on the
base was used as a landmark from which geometry measurements could be referenced. A score
along the length of the base served as an angular reference landmark to dene the specimen
orientation. A hole 4 mm in diameter by 3 mm deep was drilled into the top of the base to accept
the specimens. The ends of the trabeculae were xed by lling the hollowed out top of the epoxy

FIGURE 19.4

Schematic diagram of the tensile testing apparatus used by Rho et al.,

14

including a detail of
the glue attachment of a specimen to the grip rods. (From Rho, J.Y., Ashman, R.B., and Turner, C.H., Youngs
modulus of trabecular and cortical bone material: ultrasonic and microtensile measurements,

J. Biomech.

, 26,
111, 1993. With permission.)

0266/ ch19/frame Page 296 Friday, October 22, 1999 9:14 AM

Micromechanical Testing of Single Trabeculae

297

base with fresh epoxy and pushing the trabecula specimens into the epoxy and holding them straight
while the epoxy cured. This particular epoxy took 2 h to cure under a heat lamp. During the curing
time the trabeculae dried out; to rewet the specimens they were soaked for a minimum of 1 h in
saline before testing. When the epoxy set, it wicked up around the base of the specimen, creating
a variably sized epoxy meniscus. The size and shape of the meniscus was found to have a signicant
effect on the specimen bending response and had to be included in the nal models.
Specimens were tested by attaching the epoxy base to an

xy

platform with the trabeculae
standing vertically (Figure 19.5). The trabeculae were positioned against a horizontally oriented
force probe with a full scale rating of 0.6 N (Model UC3; Statham Instruments Inc., Hato Rey,
PR). The height of the force probe above the machined ledge on the epoxy base was measured and
the specimen orientation about its vertical axis was adjusted to place the attest contact area against
the force probe. Since specimens were not symmetric and did not have a regular geometry, the
direction of loading relative to the specimen had a signicant effect on the specimen bending
response. The reference line on the epoxy base allowed the rotational position of the specimen
relative to the force probe to be recorded. The amount of contact between the specimen and the
force probe was also recorded for use in the later modeling.
Specimen loading was produced by displacing the base relative to the xed force probe (see
Figure 19.5). As the base was displaced by hand, no consistent loading rate was used in these tests.
The displacement of the base relative to the force probe was measured with an LVDT with a

0.6 mm
linear range (Model 025 MHR, Schaevitz, Pennsauken, NJ). The LVDT was held against the force
probe by a calibrated compression spring (spring constant 0.019 N/mm). Each trial was run to a
maximum displacement of 0.4 mm; the resulting applied force ranged from 0.01 to 0.09 N. The
forcedisplacement response was recorded on an x-y plotter. A straight line was drawn to best
approximate the experimental forcedisplacement plots and the slope of line calculated. The
stiffness of the spring holding the LVDT against the force probe was subtracted from the experi-
mentally measured slope to determine the specimen stiffness (

P

/

). Specimens were loaded at
several different heights to create several different load cases, and each test was repeated several

FIGURE 19.5

Schematic diagram of the cantilever beam test apparatus used by Mente and Lewis. (From
Mente, P. L. and Lewis, J. L., Experimental method for the measurement of the elastic modulus of trabecular
bone tissue,

J. Orthop. Res.

, 7, 456, 1989. With permission.)

0266/ ch19/frame Page 297 Friday, October 22, 1999 9:14 AM

298

Mechanical Testing of Bone and the BoneImplant Interface

times to obtain an average response. Specimens were rewetted periodically throughout the testing
procedure to prevent them from drying over the course of the tests.
The nite-element models of the individual trabeculae were created after mechanical testing
using photographs of the trabecular cross sections taken as the specimens were ground down. In
order to prepare a specimen to be ground down it was bleached, encased in a hollow cylinder that
was then lled with epoxy colored with carbon black, and subjected to a vacuum to remove any
air bubbles from the epoxy. This procedure produced excellent contrast between the white bleached
bone tissue, the black potting epoxy, and the lighter epoxy of the meniscus and base. The cylinders
encasing the trabeculae were engraved internally with four longitudinal lines at 90

intervals to act
as markers for a superimposed grid (1.27 mm Bausch and Lomb reticule); the lines on the cylinder
were aligned with the reference line on the epoxy base of the specimen. The reticule provided a
consistent local

xy

coordinate system and length standard in each cross-sectional photograph so
that successive levels could be properly aligned with each other. The specimens were ground down,
by hand, with 600-grit silicon carbide paper under water irrigation at intervals of 0.25 mm until
the meniscus was reached, at which point they were ground down at intervals of 0.125 mm. Parallel
surfaces were obtained by advancing the epoxy-encased specimen up through the center of an
aluminum annulus by means of a screw advance (accuracy = 0.025 mm) and then grinding the
specimen down until the annulus was reached. Specimens were ground down until the machined
ledge on the base was reached; this allowed the

z

-position of each section to be referenced relative
to the landmarks used during testing. Knowing the

z

-position of the local coordinate system of
each section, which was dened by the reticule in the photographs, a global coordinate system for
the specimens could be dened with its origin at the center of the epoxy base on the plane created
by the ledge of the base. This global coordinate system allowed the test geometry to be transferred
to the nite-element model.
Tracings of the cross-sectional outline of the trabeculae were made and then sectioned into
three to ve quadrilaterals. The corner points of these quadrilaterals were digitized into the computer

FIGURE 19.6

View of a cantilever beam trabecular specimen and its corresponding three-dimensional finite-
element model. (From Mente, P.L. and Lewis, J.L., Experimental method for the measurement of the elastic
modulus of trabecular bone tissue,

J. Orthop. Res.

, 7, 456, 1989. With permission.)

0266/ ch19/frame Page 298 Friday, October 22, 1999 9:14 AM

Micromechanical Testing of Single Trabeculae

299

and connected together vertically to the corresponding quadrilaterals on neighboring levels to form
three-dimensional brick elements (Figure 19.6). These elements were further subdivided into addi-
tional elements to ensure that there would be a minimum of four elements in the bending direction.
Three different materials were dened in the nite-element models: epoxy, air (used to model holes
in the specimens), and the bony tissue. The elastic modulus and Poissons ratio used for epoxy
elements were 3.1 GPa and 0.36, respectively, and for air elements 1 Pa and 0.30, respectively.
The initial material properties used for the trabecular tissue was a tissue modulus of 15.0 GPa and
a Poissons ratio of 0.28. Fixation at the base of the specimen was simulated by constraining all
of the nodes on the lowest cross section against displacements in all three directions. The trabecular
model was rotated to match the specimen rotation during testing and a horizontal force was applied
at the nodes that most closely approximated the contact area between the specimen and the force
probe at the height recorded during testing. The magnitude of the force was divided equally among
each of the nodes.
The nodal displacements at the point of application of the force, obtained from the nite-
element solution (FEM displacement), were compared with the corresponding displacement cal-
culated using the experimentally measured slope (F

displacement). A discrepancy in these
displacements would indicate that the initial tissue modulus used was incorrect. Because of the
approximate linearity of the nite-element nodal displacements with the elastic modulus that was
used for the bone tissue elements (FEM modulus), the actual tissue modulus (the modulus that
would make the two displacements equal) could be determined using the equation:
(19.4)
In addition to testing trabecular bone, small cortical bone and aluminum rectangular cantilever
beam specimens were also tested. These specimens were machined to a uniform rectangular cross
section (approximate dimensions 0.5

3.0

10.0 mm). This allowed the experimental procedure
(mechanical testing and nite-element modeling) to be tested using specimens with a known
geometry and modulus. For regular rectangular beam specimens tested as cantilever beams the
modulus can be calculated using the equation:
(19.5)
where

E

= Youngs modulus,

P

= applied load,

= displacement at the point of application of the
load,

L

= distance from the xed end of the specimen to the applied load, and

I

= specimen moment
of inertia (for a rectangular specimen this is equal to

bh

3

, where

b

is the width of the specimen
and

h

is its thickness)
Being able to calculate the modulus directly allowed the accuracy of the apparatus and the
nite-element models to be determined. Tests of aluminum specimens showed an error of only
about 4 to 5% in calculated modulus compared with reference values. Tests on cortical bone
specimens showed modulus differences on the order of 17% compared with published values; this
larger error for calculated cortical bone tissue modulus probably reects the fact that the modulus
of cortical bone tissue is in general much more variable. These tests indicated that the test apparatus
was sufciently accurate to test small, relatively compliant specimens.
Several areas of potential error in nite-element modeling were also quantied. The sensitivity
of the model to changes in the position of the digitized cross sections was found to be about 5%.
Changes in the positioning of the load on a given level changed the calculated results 2 to 8%,
while moving the load between adjacent vertical levels resulted in changes of about 10%. Changes
in specimen rotation were found to cause changes of about 5%. These sources of error were found
Tissue modulus
FEM displacement
F displacement
---------------------------------------------
FEM modulus =
E
P L
I

3
3

0266/ ch19/frame Page 299 Friday, October 22, 1999 9:14 AM

300

Mechanical Testing of Bone and the BoneImplant Interface

to be generally small and represent the largest errors that might be induced in translating the
specimen geometry into a nite-element model. The most signicant effects were found to result
from changing the representation of the epoxy meniscus. Not including the epoxy meniscus at the
base of the specimen in the models, for example, resulted in changes of between 30 and 70% in
predicted modulus. This indicates the importance of this feature, which models the xation condi-
tions, to the specimen response, and underscored the importance of creating an accurate model of
the specimen and the test conditions.

III. TESTING CONSIDERATIONS
A. S

PECIMEN

G

EOMETRY

One of the major difculties in measuring the tissue modulus of individual trabeculae is dealing
with the irregular shape of the available specimens. In the studies that have been performed this
has been dealt with in several different ways; each method has its advantages and disadvantages.
In tensile tests a uniform circular cross section was assumed for the entire specimen. While this
greatly simplies the analysis and allows a simple beam equation to be used to calculate the tissue
modulus, it also introduces errors that result from the changes in the specimen cross section along
its length and in the accuracy of determining an average radius. Ryan and Williams
13
and Rho
et al.
14
have estimated that these effects result in errors in modulus calculations of about 10%.
Irregularities in specimen geometry, particularly deviations along the axial loading direction, result
in specimens that are not subjected to pure axial tension loads. In these specimens bending loads
will be introduced that may also inuence the specimens response.
The problems that result from irregularities of geometry can be avoided by machining specimens
into a standard uniform geometry, as was done by Kuhn et al.
10
and Choi et al.
11
This is the normal
practice in standard material testing procedures. The difculty that arises from doing this with
trabecular tissue is that the available material is small and inhomogeneous. As specimens are
machined, voids and defects in the tissue can become more pronounced. In addition, this process
can interrupt the tissue microstructure. Bone tissue is a composite material with hydroxyapatite
crystals deposited on a collagen scaffolding. The tissue material properties are dependent on the
properties of the hydroxyapatite and the collagen as well as the internal architecture of the scaf-
folding. In trabecular tissue the collagen architecture is much less aligned than in cortical bone;
machining away part of a trabeculae may disrupt this internal architecture resulting in changes to
the material properties of the trabecular tissue.
These problems can be circumvented by modeling the complete trabecula as was done in
bending tests using the nite-element method. Finite-element modeling of the trabeculae allows
the irregular shape of the specimens to be taken into account and allows specimens to be tested
without disturbing their internal architecture. While nite-element modeling has the potential to
represent the specimen geometry accurately, the creation of individual models for each specimen,
particularly when each specimen has to be ground down in order to dene its geometry, is an
extremely time-consuming procedure. In addition, proper element sizes and shapes must be
generated to obtain reliable results, as the accuracy of the models is highly dependent on the
mesh density. Recent advances in -CT technology and nite-element preprocessing may make
this procedure more robust and reliable by allowing models on the single trabecula scale to be
scanned and a nite-element mesh generated automatically.
1-7
The speed and storage capacity of
new computers and nite-element algorithms also mean that ner models can be created, elim-
inating problems related to poor mesh density. These advances may eventually make this proce-
dure quick and reliable enough to allow testing of signicant specimen numbers, making this
method more appealing.
There is also some indication that there is a size effect in specimen testing.
11
These effects may
arise due to inhomogeneties in the material properties of the tissue, increased inuence of structural
0266/ ch19/frame Page 300 Friday, October 22, 1999 9:14 AM
Micromechanical Testing of Single Trabeculae 301
defects as specimen size becomes small, specimen surface-to-volume ratio, or machining effects.
Choi et al.
11
have observed this size effect in cortical bone specimens, although similar results were
not seen by Rho et al.
14
Nonetheless, it should be remembered that bone tissue is an inhomogeneous
material and special consideration should be given to the size and shape of specimens being tested.
B. GRIPS AND EDGE EFFECTS
Another testing concern is stress concentrations due to concentrated loading and edge effects around
supports and grips. These effects were probably most severe for the test geometry used by Ryan
and Williams
13
where the small free length relative to the diameter of the specimen in the grip
meant that edge effects could be felt over a large percentage of the specimen length. While
elongation measurements in standard testing procedures are made on a subset of the specimen
length away from the grips to avoid these effects, this is not a practical solution for testing small
trabeculae. It should, however, be noted that stress calculations based on the applied loading
probably underestimate the tissue stresses that are experienced throughout a large percentage of
the specimen length.
Another problem that needs to be addressed with grips is specimen slipping. Ryan and
Williams
13
noted that specimen slipping could be detected by sudden changes in the
forcedisplacement tracings; slipping was also checked by placing a mark on the specimen where
it was xed in the grips so that any specimen movement occurring during their test could be
noted after the tests. In their experience, approximately 5 to 6% of the specimens tested expe-
rienced slipping in the grips. Tests where slipping was suspected were discarded. While glue can
be used to eliminate specimen slipping as done by Rho et al.,
14
the glue may be absorbed into
the specimen, thus altering its mechanical properties. Rho et al.
14
used ultrasonic techniques to
quantify the effect this had on specimen modulus calculations and found an approximate 4%
difference between specimens before and after they were inltrated with glue. These differences
were, however, only measured on dried specimens.
In three-point bending tests, the shear contributions due to the concentrated nature of the load
have been incorporated into Equation 19.2. However, Lotz et al.
16
have indicated that local plastic
deformation around loading heads and supports can also have an effect on measured specimen
deection when overall displacements are small. They recommended using an experimentally
determined correction factor to measure these local deformation effects. This correction factor ()
is determined by testing a completely supported specimen (a specimen resting on a rigid support)
in compression with a single indenter. The magnitude of local deformation can be quantied by
measuring the slope of the resulting forcedisplacement curve (p/d):
(19.6)
This can be incorporated into the analysis by adjusting the specimen stiffness measurement (P/)
used in Equation 19.3 according to the equation:
(19.7)
The factor of 1.5 in this equation is used for three-point bending to take into account that the two
bottom supports only see half the total force of the load head so local deformation effects due to
the bottom supports will be half that due to the single loading point on the top. While this effect
may have an inuence on three-point bending tests, it is unlikely that this effect will be as signicant
in cantilever beam tests. The specimen displacement relative to the applied force is much greater
1

p
d
P P
P

( ) 1 5 .
0266/ ch19/frame Page 301 Friday, October 22, 1999 9:14 AM
302 Mechanical Testing of Bone and the BoneImplant Interface
in cantilever bending than in three-point bending; thus, any local deformations will have a much
smaller effect on measurements of the overall specimen stiffness.
Edge effects that are most signicant for cantilever beam specimens are due to the meniscus
that develops at the base of cantilever beam specimens when they are potted in epoxy. Although
these effects can be taken into account by including the meniscus in the nite-element model, the
meniscus does add to the complexity of the models and the time necessary to create and solve
them. The nite-element method does take into account the stress concentrations that arise from
geometric factors such as grips or material changes (provided appropriate boundary conditions are
imposed), which makes it a good method for dealing with edge effects.
C. TISSUE HYDRATION
Tissue hydration has been shown to be a signicant factor in tissue modulus, with dried specimens
as much as 24% stiffer than wet specimens.
9
Maintaining specimen hydration is a particularly
signicant problem for small specimens that can dry out very quickly. Specimens should be kept
frozen when in storage and every effort should be made to keep them moist throughout both
preparation and testing. Kuhn et al.
10
and Choi et al.
11
kept specimens between saline-soaked lter
paper to prevent dehydration. Mente and Lewis
15
kept their specimens submerged in a saline bath
to keep them moist before testing. This was necessary, in part, to rehydrate specimens that dried
out during preparation. Specimen drying can be a problem for testing congurations that require
that the specimen be bonded to a grip or support. Specimens, for example, may need to be dried
before they are glued to ensure good glue adhesion, or the specimen may dry out while the glue
or epoxy sets. In these cases it is possible to rehydrate specimens by soaking them in saline,
although no data exist on whether drying and rehydrating of specimens has an effect on the resulting
measured tissue modulus. In general, it is prudent to avoid letting the specimens dry out if possible
and to keep them moist throughout the testing procedure.
D. APPARATUS CHECKS
Because the apparatus used to test trabeculae are usually built for this single purpose, it is
necessary to ensure the accuracy of the apparatus and its ability to handle the small load, and
displacements that must be applied or measured, as well as the generally high compliance of the
small specimens being tested. It is necessary to ensure that the compliance of the testing system
is very much lower than the compliance of the specimens. Many measuring devices used to
measure small loads are by nature fairly compliant and deformations in these elements may be
a signicant percentage of the deformation of the specimen. If there is signicant compliance,
or elements of the test system introduce loads to the system, these sources need to be quantied
and subtracted from the specimen response. Any nonstandard test system also requires calibration
and verication; this can usually be done by testing homogeneous specimens of known modulus
to verify the testing procedures. The specimens chosen should be similar in size and stiffness to
the trabeculae to be tested. Testing should include both the mechanical test apparatus and the
equations or procedure used to analyze the results.
IV. RESULTS
Results to date (Table 19.1) indicate a modulus of trabecular bone tissue that ranges from approx-
imately 1.0 to 8.0 GPa (for specimens tested wet). Reported values for cortical bone tissue are on
the order of 20 GPa.
17
While the amount of available data is still quite small and does not provide
a basis for making any statement about the modulus of trabecular bone tissue in general, these
studies do suggest that trabecular bone tissue modulus is less stiff than the cortical bone tissue
modulus. There are several possible explanations for the differences in these tissues. There may
0266/ ch19/frame Page 302 Friday, October 22, 1999 9:14 AM
Micromechanical Testing of Single Trabeculae 303
T
A
B
L
E

1
9
.
1
E
l
a
s
t
i
c

M
o
d
u
l
u
s

o
f

T
r
a
b
e
c
u
l
a
r

B
o
n
e

T
i
s
s
u
e
S
o
u
r
c
e

o
f

T
i
s
s
u
e
C
o
n
d
i
t
i
o
n

o
f

T
e
s
t

S
p
e
c
i
m
e
n
N
o
.

T
e
s
t
e
d
T
e
s
t
i
n
g

M
e
t
h
o
d
L
o
a
d
i
n
g

o
r

S
t
r
a
i
n

R
a
t
e
A
p
p
r
o
x
i
m
a
t
e

D
i
m
e
n
s
i
o
n
s
,

m
m
C
a
l
c
u
l
a
t
e
d

T
i
s
s
u
e

M
o
d
u
l
u
s

(
G
P
a
)

M
e
a
n

(
S
t
D
)
R
e
f
.
H
u
m
a
n

m
e
d
i
a
l

t
i
b
i
a
l

p
l
a
t
e
a
u
D
r
y
9
T
r
a
b
e
c
u
l
a
r

b
u
c
k
l
i
n
g
0
.
1
3
4
6

m
m
/
s
L

=

1
.
0

1
.
8
1
4
.
1
9
W
e
t
9
t

=

0
.
0
9

0
.
1
4
1
1
.
3
w

=

0
.
2
5

1
.
0
H
u
m
a
n

d
i
s
t
a
l

f
e
m
u
r
D
r
y
4
T
r
a
b
e
c
u
l
a
r

b
u
c
k
l
i
n
g
U
n
k
n
o
w
n
L

=

2
.
4

3
.
0
8
.
7

(
3
.
2
)
8
t

=

0
.
1
3
w

=

0
.
8

1
.
2
H
u
m
a
n

f
e
m
u
r

(
d
r
i
e
d
)
R
e
w
e
t
t
e
d
3
C
a
n
t
e
l
e
v
e
r

b
e
a
m

w
i
t
h

F
E
A
U
n
k
n
o
w
n
L

=

5

1
0
6
.
2

(
1
.
2
)
1
5
H
u
m
a
n

t
i
b
i
a

(
f
r
e
s
h
)
3
1
1
.
2

(
1
0
.
1
)
C
o
m
b
i
n
e
d
6
8
.
7

(
7
.
0
)
H
u
m
a
n

i
l
i
a
c

c
r
e
s
t
6
3

y
e
a
r
s

o
l
d
W
e
t
2
9
T
h
r
e
e
-
p
o
i
n
t

b
e
n
d
i
n
g
0
.
8
1

m
m
/
s
L

=

1
.
5

2
.
0
4
.
2

(
2
.
0
)
1
0
2
3

y
e
a
r
s

o
l
d
1
3
h

=

0
.
0
5

0
.
2
0
3
.
0

(
1
.
6
)
C
o
m
b
i
n
e
d
4
2
b

=

0
.
0
5

0
.
2
0
3
.
8
B
o
v
i
n
e

d
i
s
t
a
l

f
e
m
u
r
W
e
t
3
8
T
e
n
s
i
o
n
0
.
0
1

/
s
L

=

4
.
0
0
.
8

(
0
.
4
)
1
3
r

=

0
.3
H
u
m
a
n

p
r
o
x
i
m
a
l

t
i
b
i
a
V
e
r
t
i
c
a
l
W
e
t
1
3
T
h
r
e
e
-
p
o
i
n
t

b
e
n
d
i
n
g
0
.
1
1

m
m
/
s
L

=

1
.
5

2
.
0
4
.
9

(
1
.
8
)
1
1
H
o
r
i
z
o
n
t
a
l
7
h

=

0
.
0
5

0
.
2
0
3
.
8

(
0
.
5
)
C
o
m
b
i
n
e
d
2
0
b

=

0
.
0
5

0
.
2
0
4
.
6

(
1
.
3
)
H
u
m
a
n

t
i
b
i
a
D
r
i
e
d
2
0
T
e
n
s
i
o
n
5
.
5

/
s
L

=

2
.
3
1
0
.
4

(
3
.
5
)
1
4
r

=

0
.
0
9
H
u
m
a
n

t
i
b
i
a
W
e
t
2
3
4
-
p
o
i
n
t

b
e
n
d
i
n
g
0
.
0
2

N
/
s
e
c
L

=

1
.
5

2
.
0

m
m
5
.
7

(
1
.
3
)
1
8
h

=

0
.
0
7

0
.
1
4
0

m
m
b

=

0
.
0
7

0
.
1
4
0

m
m
0266/ ch19/frame Page 303 Friday, October 22, 1999 9:14 AM
304 Mechanical Testing of Bone and the BoneImplant Interface
be differences in the degree of mineralization of the two tissues. The microarchitecture of the tissue
including the organization or orientation of the collagen ber network and the relationship between
the ber network and the mineralization may also affect the modulus of the two tissues. Cortical
bone tissue is much more organized than trabecular bone; it is built up around the osteon with
concentric lamellae of mineralized collagen bers which tend to alternate direction between the
lamellae. There is no such organized structure in the trabecular tissue. Collagen orientation is more
random and cement lines indicating the juncture between remodeled areas are randomly dispersed
through the tissue.
18
How these architectural differences may inuence the modulus is unknown,
but they do suggest a structural basis for differences in tissue modulus between cortical and
trabecular bone.
REFERENCES
1. Hollister, S.J., Brennan, J.M., and Kikuchi, N., A homogenization sampling procedure for calculating
trabecular bone effective stiffness and tissue level stress, J. Biomech., 27, 433, 1994.
2. Ladd, A.J. and Kinney, J.H., Numerical errors and uncertainties in nite-element modeling of trabe-
cular bone, J. Biomech., 31, 941, 1998.
3. Ladd, A.J., Kinney, J.H., Haupt, D.L., and Goldstein, S.A., Finite-element modeling of trabecular
bone: comparison with mechanical testing and determination of tissue modulus, J. Orthop. Res., 16,
622, 1998.
4. Muller, R. and Ruegsegger, P., Analysis of mechanical properties of cancellous bone under conditions
of simulated bone atrophy, J. Biomech., 29, 1053, 1996.
5. Muller, R. and Ruegsegger, P., Three-dimensional nite-element modeling of noninvasively assessed
trabecular bone structures, Med. Eng. Phys., 17, 126, 1995.
6. Ulrich, D., van Rietbergen, B., Weinans, H., and Ruegsegger, P., Finite element analysis of trabecular
bone structure: a comparison of image-based meshing techniques, J. Biomech., 31, 1187, 1998.
7. van Rietbergen, B., Weinans, H., Huiskes, R., and Odgaard, A., A new method to determine trabecular
bone elastic properties and loading using micromechanical nite-element models, J. Biomech., 28,
69, 1995.
8. Runkle, J.C. and Pugh, J., The micromechanics of cancellous bone. II. Determination of the elastic
modulus of individual trabeculae by a buckling analysis, Bull. Hosp. Joint Dis., 36, 2, 1975.
9. Townsend, P.R., Rose, R.M., and Radin, E.L., Buckling studies of single human trabeculae,
J. Biomech., 8, 199, 1975.
10. Kuhn, J.L., Goldstein, S.A., Choi, K., et al., Comparison of the trabecular and cortical tissue moduli
from human iliac crests, J. Orthop. Res., 7, 876, 1989.
11. Choi, K., Kuhn, J.L., Ciarelli, M.J., and Goldstein, S.A., The elastic moduli of human subchondral,
trabecular, and cortical bone tissue and the size-dependency of cortical bone modulus, J. Biomech.,
23, 1103, 1990.
12. Timoshenko, S.P. and Goodier, J.N., Theory of Elasticity, 3rd ed., New York, McGraw-Hill, New York,
1970.
13. Ryan, S.D. and Williams, J.L., Tensile testing of rodlike trabeculae excised from bovine femoral bone,
J. Biomech., 22, 351, 1989.
14. Rho, J.Y., Ashman, R.B., and Turner, C.H., Youngs modulus of trabecular and cortical bone material:
ultrasonic and microtensile measurements, J. Biomech., 26, 111, 1993.
15. Mente, P.L. and Lewis, J.L., Experimental method for the measurement of the elastic modulus of
trabecular bone tissue, J. Orthop. Res., 7, 456, 1989.
16. Lotz, J.C., Gerhart, T. N., and Hayes, W.C., Mechanical properties of metaphyseal bone in the proximal
femur, J. Biomech., 24, 317, 1991.
17. Cowin, S.C., The mechanical properties of cortical bone tissue, in Bone Mechanics, Cowin, S.C., Ed.,
CRC Press, Boca Raton, FL, 1989, 97.
18. Choi, K. and Goldstein, S.A., A comparison of the fatigue behavior of human trabecular and cortical
bone tissue, J. Biomech., 25, 1371, 1992.
0266/ ch19/frame Page 304 Friday, October 22, 1999 9:14 AM

305

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

20

Strain Gauge Measurements
from Bone Surfaces

John A. Szivek and Vasanti M. Gharpuray

CONTENTS

I. Introduction ..........................................................................................................................305
II. History of Strain Gauge Use in Biomechanics ...................................................................306
III. Components of a Strain Gauge Measuring System............................................................308
IV. Strain Gauge Selection.........................................................................................................308
A. Foil Grid Material ..........................................................................................................308
B. Gauge Backing Material ................................................................................................310
C. Gauge Length, Width, and Pattern.................................................................................310
D. Other Gauge Options .....................................................................................................311
V. Lead Wires and Transcutaneous Connectors.......................................................................312
VI. Preparation of Gauges Before Attachment ..........................................................................313
VII. Attachment of Gauges..........................................................................................................314
A. Attachment for

in Vitro

Studies .....................................................................................314
B. Attachment for

in Vivo

Studies......................................................................................314
C. Preparation of CPC-Coated Gauges for Extended

in Vivo

Strain Measurement..........316
VIII. Measurement Collection ......................................................................................................316
IX. Data Analysis .......................................................................................................................317
References ......................................................................................................................................318

I. INTRODUCTION

There are several applications for strain measurement in orthopaedic biomechanics and biomaterials
research.

1

These include testing and development of prosthetic and implantable devices,

2-5

determining
the mechanical properties of biological structures,

6,7

determining bone strains during physiological
activities,

8-10

determining strain levels that lead to an adaptive bone remodeling response,

11-13

and
validating or complementing computational simulations.

14,15

Although various strain measurement
techniques such as photoelasticity, brittle coatings, holographic interferometry, and thermographic stress
analysis have been established, the most common is electric resistance strain gauging.

16

Strain gauging offers many advantages over other techniques.

17

Gauges are very sensitive and
can measure strains directly. When properly protected, gauges can measure strains in a wide variety
of corrosive environments and can be continuously monitored in a dynamic loading situation. The
primary disadvantage of strain gauges is that they measure strain at discrete points only. One must
have a good idea of the critical regions in the test specimen prior to gauge attachment and testing.
In addition, strain measurement must be used in conjunction with material properties derived from
separate tests to determine stresses. Thus, strain gauges are commonly used to measure strains in
a few critical regions, and are used in conjunction with material properties and analytical models
to draw inferences about the stress state of the bone.

0266/ ch20/frame Page 305 Friday, October 22, 1999 9:20 AM

306

Mechanical Testing of Bone and the BoneImplant Interface

The most common use for strain gauges

in vitro

has been to determine strain and stress
distributions in cadaver bone under various loading conditions and with various implants in place.
However, it is extremely difcult to accurately simulate muscle and joint loads in animals during
physiological activities with a mechanical testing apparatus. For this reason,

in vitro

studies of
animal whole bones are much less common than

in vitro

studies of human whole bones.

In vivo

strain gauging studies are more commonly used to study whole bones in animal models.
One challenge faced when performing

in vivo

strain gauging is the selection of a consistent
zero strain point. Since bones are always loaded by muscles in living animals it is not possible to
dene a posture in which the bone strain is zero. Gauges can be used to measure strain changes
relative to some established zero during a specic activity. In this chapter, techniques for both

in vitro

and

in vivo

strain measurement will be discussed. Additional information about errors and
pitfalls that the user should be aware of and some criteria for gauge, adhesive, and protective coating
selection will be provided.

II. HISTORY OF STRAIN GAUGE USE IN BIOMECHANICS

The fundamental principle underlying the electric resistance strain gauge is that when a wire is
stretched, its length increases and its cross-sectional area decreases causing a change in its electrical
resistance.

18

The equation relating the length change to the resistance change is
(20.1)
where

R

= resistance (ohm),

= a material property of the conductor called resistivity (ohmcm),

L

= length of the conductor (cm), and

A

= cross-sectional area of the conductor (cm

2

). In a properly
calibrated system, one can determine the change in length of a wire by measuring its change in
electrical resistance.
The rst recorded use of a wire strain gauge was in 1931 when the deformation of pillars was
measured by using wires stretched between them.

19

The bonded wire gauge was rst used by Clark
and Datwyler

20

in 1938 to measure strains during tensile impact loading and was patented by
Edward Simmons in 1942.

21

Advanced fabrication techniques led to the development of foil strain
gauges made of self-temperature-compensated alloys.

22

The foil is photoetched in a grid pattern
onto a plastic backing which makes it easier to handle. The backing also makes attaching the foil
onto the test specimen easier. The foil pattern can vary in size and shape and, if necessary, several
gauges can be stacked in different orientations on top of each other to create rosettes.
Gurdjian and Lissner

23

rst described the use of a strain gauge on living bone in 1944. The
wire strain gauge, which was bonded with polymethylmethacrylate (PMMA) to the cranium of an
anesthetized test animal, was used to collect

in vivo

bone deformation measurements. In this early
study, it took 2 h for the methacrylate to cure thoroughly and the site had to be kept exposed
because of the lack of waterproong. In a review of stress analysis techniques used on bone, Evans

24

reported that he, Coolbaugh, and Lebow had strain-gauged a canine tibia and collected the rst
bone strains during gait.
The earliest review of gauge preparation and bonding techniques was provided by Roberts.

25

He described a number of ways to waterproof, wire, and bond strain gauges to bone in living
animals. The most important suggestions from his review were that bone surfaces should be prepared
by degreasing with alcohol, ether, or acetone and that a cyanoacrylate adhesive was a better choice
for bonding gauges than dental cement (PMMA) or epoxy because it bonded almost instantaneously.
He also recommended obtaining lead wire isolation by using wires coated with Teon

(E.I. du
Pont de Nemours and Company, Inc., Wilmington, DE)



with an additional silastic layer or silastic
tubing around them and using subcutaneous strain relief loops.
R
L
A
=

0266/ ch20/frame Page 306 Friday, October 22, 1999 9:20 AM

Strain Gauge Measurements from Bone Surfaces

307

At about the same time, biochemical studies examining cyanoacrylates for other tissue-bonding
applications showed that these adhesives are broken down quickly

in vivo

,

26

but that the butyl form
lasts longer and is better tolerated than the methyl, ethyl, or propyl cyanoacrylates.

27

Lanyon and Smith

28

were the rst to publish a thorough description of gauge preparation and
placement techniques as part of an

in vivo

strain gauge study. In their study, polyester-backed wire
gauges were attached with isobutyl 2-cyanoacrylate to the tibiae of sheep and

in vivo

strain
measurements were collected during gait. However, this study was plagued with problems caused
by the electric circuitry and recording devices. An AC-based signal conditioner used in conjunction
with high-resistance semiconductor gauges eliminated these problems.

29

Foil gauges were rst used
in 1972.

30

They were noted to be easier to place and to cause less damage to the bone and
surrounding tissues.

31

Studies using foil gauges were extended to several types of animals and to
humans. Gauges attached with isobutyl 2-cyanoacrylate were used to collect

in vivo

measurements
from the human tibial shaft during gait.

32

The earliest reported use of strain gauges in a human (although it was attached to a prosthesis
and not directly to bone) was by Rydell in 1966.

33

In that study, strain gauges were attached to a
femoral prosthesis before it was implanted. The instrumented prosthesis was used to determine the
forces acting on the femoral head.
Evidence that gauges placed

in vivo

could measure physiological bone deformation was reported
when strains were collected and related to simultaneous accelerometer readings

28



and bone surface
electrical potential readings.

34,35

One of these studies

35

showed a linear relationship between bone
surface charge and bone strain during gait. Later, optical extensometry was used during mechanical
loading of freshly explanted bones from test animals which had been used for

in vivo

measurements
for validation of

in vivo

strain measurement accuracy.

36

Carter et al.

37

also showed that accurate strain
measurements could be collected after strain-gauged, explanted bones were soaked in saline for a week.

In vitro

, strain gauges have been used for bone strain measurements since the late 1950s.

38

Lambert

39

reported that gauges could be attached directly to the surface of moist cortical bone and
described methods for waterproong the gauges and lead wires. In the Lambert study, the defor-
mation and weight-bearing functions of the bula and tibia were measured.
In some of these studies, bone surfaces were dried prior to bonding of gauges. In addition,
procedures such as surface cleaning, scraping, degreasing, and dehydration were used prior to gauge
placement. A report on gauge application to bone

in vitro

showed that techniques such as scraping
and degreasing create visible surface damage to bone.

40

The authors concluded that these techniques
were acceptable for

in vitro

preparations and produced good bonding. However, they are too
aggressive for

in vivo

studies. The surface damage created can be expected to cause a remodeling
or healing response which would alter bone properties and could recruit cells which degrade and
remove the adhesive more quickly.
Around 1979, widespread interest in measurement of physiological strains and advances in
strain gauge and adhesive technology led to a proliferation of bone strain measurement studies.

In
vitro

procedures were used to determine the mechanical behavior of bone under various loading
conditions.

41-44

They were also used to assess the effect of disease, surgical procedures, or implants
on strain patterns.

3,45-47



In vivo

strain measurements were used to compare loading pattern changes
to anatomical and histological characteristics of bone,

9,12,39,48-50

to examine evolutionary aspects of
bone shape,

10,49-51

and to correlate implant-induced bone shape changes with implant-induced strain
changes.

52-55

Bone strains were also measured during exercise protocols in a variety of mammals
and birds

56-61

and to study remodeling in microgravity environments.

62,63

The aspect of the

in vivo

strain gauging procedure that still needs signicant development is
bonding. Cyanoacrylate adhesives degrade within the bodys environment in as little as 2 to 3 weeks
leading to measurement inaccuracies. An ongoing search for better methods of attaching strain
gauges to bone

in vivo

has led to three approaches. One approach is the use of dental adhesives
originally used on enamel surfaces. Page et al.

8

reported the successful use of this approach to bond
gauges to femora in a dog model.



The second approach which has been successfully used in animals

0266/ ch20/frame Page 307 Friday, October 22, 1999 9:20 AM

308

Mechanical Testing of Bone and the BoneImplant Interface

and humans utilizes PMMA but requires that the bone surface be abraded or drilled and roughened.

64

A more promising approach for very long-term sensor attachment is the use of calcium phosphate
ceramic (CPC) particles bonded to the strain gauge sensing surface.

65-68

This approach has been
successfully used to collect strain measurements after as long as 16 weeks

in vivo

.

68

In addition,
recent studies

69

have shown that the bond strength of the CPC gauge interface and the boneCPC
interface remain constant for up to 9 months

in vivo.

70

III. COMPONENTS OF A STRAIN GAUGE MEASURING SYSTEM

Beckwith et al.

71

have outlined several factors for successful



strain gauge measurement:
Accurate operation and stability of the gauge itself;
Selecting appropriate lead wires (and an adequate percutaneous conduit when

in vivo

monitoring is performed);
Providing protection of the gauge from the environment;
Utilizing a strong, stable adhesive to bond the gauge to the test specimen and a repro-
ducible attachment technique;
Measuring with a stable, low-noise electrical circuit to provide adequate measurement
accuracy.
A good understanding of each of these factors is important, since gauge performance can be
optimized for a particular environment by the correct choice of these characteristics.

IV. STRAIN GAUGE SELECTION

The rst step in projects requiring strain measurement is the selection of suitable gauges. The choices
to be made include (1) the foil material and resistance; (2) the backing material; and (3) the gauge
length and pattern (Figure 20.1). A comprehensive examination of all gauge types available is beyond
the scope of this chapter but is available in texts

72,73

and in technical literature

22,74

provided by
manufacturers. However, a brief review of some criteria for appropriate gauge selection will be given.

A. F

OIL

G

RID

M

ATERIAL

The three most important requirements that a foil must satisfy are (1) low temperature sensitivity;
(2) high strain sensitivity or gauge factor (GF); and (3) high electric resistivity.

71

Commercial gauges,
which satisfy most of these requirements, are usually made of a nickel-copper alloy such as Con-
stantan, Advance, or Monel or a nickelchromium alloy such as Karma, Isoelastic, or Nichrome V.
Table 20.1 gives the compositions and properties of some of the more commonly used alloys.

FIGURE 20.1

Schematic showing a typical strain gauge.

0266/ ch20/frame Page 308 Friday, October 22, 1999 9:20 AM

Strain Gauge Measurements from Bone Surfaces

309

Temperature sensitivity can cause errors in strain gauge measurements because temperature
uctuations induce changes in the resistance of the gauge and GF. Thus, if the temperature of the
environment changes after the gauge is installed and balanced, the indicator will register strains
even in the absence of loading changes. This reading is called

thermal output

or

apparent



strain

.
Thermal output can be quite large when gauges are used at high temperatures but is relatively small
(approximately 0.5

F) when gauges are used between room temperature and body temperature.
While this may not appear to be of concern in biomechanical testing where the temperature is held
constant by body uids, one must be aware that friction between overlying tissue and gauges can
cause a localized temperature increase. To minimize thermal output, alloys are processed to reduce
the temperature sensitivity of the foil material. These self-temperature-compensated (STC) gauges
are available in different grades, which are based on the thermal expansion coefcient of the test
material.
The sensitivity of the foil material is dened by GF, which is the ratio of the relative change
in resistance to the relative change in length of the material.
(20.2)
Clearly, the higher GF, the more sensitive the gauge. GF depends on the material, size and geometry
of the grid, the STC number, and the temperature, and is determined experimentally for each gauge.
Of the commonly available foil materials, the Isoelastic alloy usually has the highest GF (approxi-
mately 3.5), while for most other materials GF varies from 1.9 to 2.5. Unfortunately, the Isoelastic
alloy is not available in STC form, and can be used only under dynamic loading conditions in which
a zero reference is not required. One has to compromise between choosing the highest GF and the
best STC. Constantan or Advance are good foil material choices. These materials have good thermal
stability and a linear strain sensitivity over a large range of strain.

72

This makes them an ideal choice

TABLE 20.1
Common Foil Grid Materials and Their Properties

22,71

Grid Material Composition Gauge Factor Resistivity

.cm

10

6

Nichrome V 80% Nickel
20% Chromium
2.0 108
Constantan, Copel, Advance 45% Nickel
55% Copper
2.0 49
Isoelastic 55.5% Iron
36% Nickel
8% Chromium
0.5% Molybdenum
3.5 112
Karma 74% Nickel
20% Chromium
3% Aluminum
3% Iron
2.4 130
Manganin 84% Copper
12% Manganese
4% Nickel
0.47 48
Platinum-Iridium 95% Platinum
5% Iridium
5.1 24
Monel 67% Nickel
33% Copper
1.9 42
GF =
dR R
dL L
/
/

0266/ ch20/frame Page 309 Friday, October 22, 1999 9:20 AM

310

Mechanical Testing of Bone and the BoneImplant Interface

for strain measurement in biomechanics, since such gauges can be accurately calibrated over a narrow
range and the calibration constant will remain the same over a much wider range.
The highest-resistance strain gauges offer the advantages of reduced heat generation within the
foil and they aid in heat dissipation, provide for lead wire desensitization, and improve the signal-
to-noise ratio. Since bone has poor thermal conductivity, a high-resistance gauge (i.e., 1000 or
5000

) would be the best choice to minimize thermal drift. Matching of the thermal expansion
coefcient of the gauge with that of the bone will also reduce errors due to thermal expansion of
the gauge. Fortunately, since most testing with bone is performed with the bone and gauges
immersed in body-temperature uids (both

in vivo

and

in vitro

), thermal changes are relatively
insignicant. Thus, the very high resistance gauges (i.e., greater than 1000

), which are also more
expensive, are not essential. Gauges are most commonly available with 120 and 350

resistances.
The 120-

gauges provide adequate measurements. In cases in which low excitation voltages and
the least thermal effects are desired, higher-resistance gauges are a better choice even though gauge
selection is limited and prices are higher.

B. G

AUGE

B

ACKING

M

ATERIAL

The three functions of the gauge backing material are to provide ease of handling, to provide a
means of attaching the foil to the test specimen, and to act as an insulator between the test material
and the foil. The most common backing materials are phenolic impregnated paper, polyimide,
epoxy-type plastic lms, and glass berreinforced epoxy-phenolic. The choice of the backing
material is not completely independent of the foil and adhesive.
Most gauges are made on a polyimide backing. This is the backing material of choice for bone
strain measurements because it is tough and exible and can tolerate deformations up to 20%
without experiencing any loss of integrity. Its exibility is an asset when placing gauges on surfaces
with a small radius of curvature and when collecting strains from very exible bones.

C. GAUGE LENGTH, WIDTH, AND PATTERN
Strain gauges are available with gauge lengths that range from about 0.2 to about 100 mm. When
selecting a gauge length and width, it is important to remember that a gauge averages the strain
it measures over its surface area. Critical regions on a test specimen may experience high strains
and high strain gradients. Measurements from a strain gauge in regions with large strain gradients
can be unrepresentative of the true strain (Figure 20.2). Errors in excess of 20% have been noted
when relatively large gauges are placed in regions of high strain gradients.
75,76
For this reason it
is critical that the narrowest grid width appropriate be selected. In addition, bone specimens
often have curved surfaces, and bone properties vary as a function of position along and across
the bone. Narrow gauges can help minimize measurement inaccuracies caused by these factors.
When selecting small gauges, the length should also be considered carefully. Gauges less than
3 mm in length can be difcult to handle, install, and wire, and have poorer heat dissipation
properties. In addition, their maximum elongation, measurement stability during static loading,
and endurance during fatigue loading are poorer than those of larger gauges.
22
A compromise
must often be made between a small gauge length to minimize gradient errors and a larger gauge
length to minimize thermal drift errors. The same considerations should be given to gauge width
selection. A smaller width should be chosen to minimize averaging errors caused by high strain
gradients, while a larger width should be chosen when heat dissipation and gauge stability are
considered essential.
In addition to length, width, and pattern choices, gauges are also available in single and
multiple grids. Single-grid gauges (also called uniaxial gauges) are often used when the strain
along one direction is of primary interest. Similarly, double-grid or biaxial gauges (i.e., gauges
that have two strain sensing elements oriented perpendicular to each other) are often used when
0266/ ch20/frame Page 310 Friday, October 22, 1999 9:20 AM
Strain Gauge Measurements from Bone Surfaces 311
the strains along two perpendicular directions are of interest. Rosette gauges, which have three
or more elements, provide the most complete information about the strain state of a point on the
bone. They can be used to determine strains along specic directions as well as the principal
strains. Misalignment errors are less critical if rosettes are used for measuring principal strains.
76-78
However, it should be noted that accurate principal strain measurements are dependent on accurate
gauge orientation measurements. As such, careful alignment of gauges should not be abandoned
when rosettes are utilized. Their relative insensitivity to misalignment makes these the gauges
of choice when it is difcult to align gauges accurately along specic axes or landmarks or when
only peak strains are of interest. Although the alignment of the rosettes during installation is not
critical, as already noted, the accurate measurement of gauge orientation is critical in order to
be able to complete the strain analysis. The equations used to assess the principal strains and
principal strain directions are dependent on the type of rosette utilized and can be found in many
fundamental mechanics books.
72,79
Rosettes are generally available in several different congurations (Figure 20.3) and the choice
of a conguration usually depends on the proposed location of the gauge on the test specimen.
Another choice that an investigator must make when using rosettes is whether to use a single-plane
rosette (i.e., one in which all elements are in one plane) or a stacked rosette (i.e., one in which the
elements are stacked on top of each other) (see Figure 20.3). The advantage of the single-plane
rosette is that it has better heat dissipation properties and thus its measurements are more accurate
when it is used on relatively at specimens experiencing small strain gradients. Single-plane rosettes
are more exible than stacked rosettes and can be contoured to a specimen surface better. However,
principal strain measurements can be extremely inaccurate when used on curved surfaces with large
strain gradients because each element collects strain from a different location. The stacked rosette
measures strains in different directions at a single point, and is much more suitable in areas with
high strain gradients and on curved surfaces with an uneven or small radius of curvature.
D. OTHER GAUGE OPTIONS
When selecting gauges for bone strain measurements, it is valuable to select a gauge type which
offers the option of large solder dots, integral strain relief tabs, and an encapsulated face. This
FIGURE 20.2 Schematic showing how gauge length can cause averaging errors in regions of high strain
gradients. A hole in a test specimen causes a stress concentration at its edge. Since a strain gauge has a nite
length, it measures an average strain over its length, and thus does not give a measure of the true strain near
the hole.
0266/ ch20/frame Page 311 Friday, October 22, 1999 9:20 AM
312 Mechanical Testing of Bone and the BoneImplant Interface
combination of options allows ease of soldering, which ultimately results in stronger, more elec-
trically conductive solder joints, a sensor that is more immune to measurement errors caused by
wire motion, and a more waterproof sensor wire system.
V. LEAD WIRES AND TRANSCUTANEOUS CONNECTORS
The rst step in gauge preparation is wiring. This must be done on the bench top prior to gauge
attachment to bone since in situ soldering is cumbersome for in vitro tests and impossible during
sterile surgery. In order to minimize contamination of gauges, they should be placed on a clean
glass surface. One edge of each gauge can be held in place with a tape that leaves minimal residue.
An invisible Scotch

tape (3M, St. Paul, MN) can be used. The low-tack surface of a Post-it

(3M,
St. Paul, MN) note leaves even less residue, although it holds the gauge less securely. If gauges
without integral strain relief tabs have been purchased, it is advisable to attach strain relief tabs to
the gauge (Technical Tip, TT-604, Measurements Group, Raleigh, NC) and then attach wires to
the tabs. In this way any wire motion will not be transferred to the gauge.
For in vitro applications, any exible, relatively low resistance wire may be used (see
Section VIII). In most cases two wires are sufcient, one attached to each solder tab. Three-wire
systems, which are used to provide lead wire temperature compensation, are not essential because
the temperature of the testing setup is usually maintained constant. Wires used in vivo must satisfy
some additional requirements. First, they must be very exible so that motion of muscles does not
put large stresses on the gaugewire junction. Multiconductor cables with multistrand high-con-
ductivity, low-oxygen copper wires can be obtained from Cooner Wire (Chatsworth, CA). A shielded
two-conductor cable such as the NMUF 2/30-4046 SJ (Cooner Wire, Chatsworth, CA) works well
for single-element gauges, or one such as NMUF 3/30-4046 SJ (Cooner Wire, Chatsworth, CA)
can be used if a three-wire system is to be implemented. For small animal models the wires from
this cable can be used without the shield and cable cover. The six conductor NMUF 6/30-4046 SJ
(Cooner Wire, Chatsworth, CA) is a good choice for use with rosette gauges.
If wires are to be buried in vivo for any length of time prior to monitoring, the ends of the
wires or cables can be waterproofed using the same procedure used for waterproong gauges
(discussed in Section VI). In some cases, individual wires can be led transcutaneously and left
exposed for later monitoring. In rat models wires can often be left exposed in this fashion for
several weeks. Connectors can then be attached at the time of monitoring and removed when
monitoring is complete, leaving transcutaneous wires exposed for future monitoring sessions. If
measurements are to be collected often for extended periods, transcutaneous multipin connectors
may be anchored in the skin.
62
However, this requires additional surgical manipulation to place the
connection system.
In cases in which a few wires will be chronically transcutaneous, it is advisable to make a
small puncture through the skin with a new, ne-tipped scalpel blade. A purse string suture can be
used to hold the skin closed around each wire. A drop of cyanoacrylate can be used to seal the
skinwire junction. If transcutaneous connectors are to be used, silicone anges on the connector
are valuable, allowing them to be sutured to the skin. For a connector placed in the cranium, PMMA
FIGURE 20.3 Some configurations of strain gauge rosettes.
0266/ ch20/frame Page 312 Friday, October 22, 1999 9:20 AM
Strain Gauge Measurements from Bone Surfaces 313
can be used to attach it to surrounding bony tissues and the scalp can be closed around the connector.
If a large number of strain gauges are to be implanted, a large number of wires must be brought
out through the skin. In such cases, heat-shrinkable Teon tubing can be used to hold all the wires
together, and a percutaneous device (PMMA post with a stainless steel mesh ange) can be placed
at the exit site of the wires.
80
Alternatively, an exit site for each wire has been used successfully
68
but requires that wires be kept 2 to 3 cm apart.
Connector selection is dependent on the monitoring system used to collect measurements. Widely
available, inexpensive nine-pin connectors can be used to connect up to four gauges to monitoring
equipment. If gauges are to be monitored using externally attached miniature radio telemetry or the
connection requirements mandate the use of very small, multipin connectors, the C-46 series Konigsberg
implantable seven-pin connectors (Konigsberg Instruments, Pasadena, CA) work well.
Surgical approaches will be dependent on the gauge placement site. In addition to planning the
approach to reach the bone site of interest, planning to allow wire placement is essential. In small
animal models, because of space constraints, it is often easier to place the wires in the positions
that they will occupy at the completion of surgery before attempting to attach the gauges to bone.
In all cases it will be necessary to route the wires subcutaneously to a site which the animal cannot
easily reach. In most animals the best site is at the nape of the neck or the crown of the skull. With
primates, additional restraints to prevent them from manually destroying wires and connectors are
also needed. In these animal models an implantable telemetry system may provide the best solution.
These systems are best housed in the abdominal cavity of small animals, although miniaturization
has made some systems so small they can be left adjacent to the gauge site.
Another common problem is infection at transcutaneous leads. In larger animal models trans-
cutaneous leads and transdermal connectors necessitate careful wound care to prevent infections
and require protection of leads or connectors. Recent developments in implantable telemetry
systems should solve these problems. However, telemetry systems are still relatively noisy compared
with hard-wired measurement systems and do require the use of higher-resistance gauges to reduce
bridge power requirements.
VI. PREPARATION OF GAUGES BEFORE ATTACHMENT
For the results of an in vitro test to be representative of the in vivo behavior of bone, the in vivo
conditions must be simulated as closely as possible. Although early studies showed that there was
no difference between the deformation behavior of live and dead bone,
81,82
recent studies have
shown that if bone is allowed to dry, its properties change signicantly.
83
For this reason, in vitro
strain gauge measurements are generally made while the bone is kept moist or immersed in a
physiological solution, and strain gauges used for in vivo measurements are constantly exposed to
body uids. Therefore, gauges used for both in vitro and in vivo tests must be encapsulated in a
protective waterproof coating.
One of the challenges of strain measurement collection is preventing uid inltration into the
gaugewire junction. For short-term in vitro testing, this requirement can be satised by purchasing
strain gauges with attached lead wires that are completely encapsulated. If inltration is a problem
after immersion in uid, an additional coating of nitrile rubber may be applied. For long-term
testing in saline, the gauge and gaugewire junction should be coated with a lm of Teon. Coatings
of nitrile rubber and polysulde epoxy should be applied over the Teon lm with each coating
extending beyond the previous one. It is also advisable to coat the lead wires with a thin coat of
nitrile rubber. It is important to note that some of these coatings take up to 24 h to cure completely.
If a gauge is installed on a bone surface prior to waterproong, the specimen may dehydrate during
the coating-curing period.
In vivo, the waterproong technique used should not only keep wires and gauges insulated but
should also produce a gaugewire assembly that is relatively bioinert so that the tissue reaction to
the waterproofed gauge is minimal. Handling of gauges during surgery also presents unique
0266/ ch20/frame Page 313 Friday, October 22, 1999 9:20 AM
314 Mechanical Testing of Bone and the BoneImplant Interface
waterproong challenges. Two drawbacks of the traditional coating materials used for waterproof-
ing, such as epoxy and silicone, are that these materials can induce a biological response and that
they create assemblies that are stiff and bulky. Loads on the gauge assemblies caused by muscles
can create large interfacial shear stresses at the gaugebone interface leading to adhesive failures.
Thus, the coating must be exible, relatively bioinert, and waterproof and must protect the gauges
and wires from damage caused by handling during surgery. A multilayer coating developed by
Szivek
54
has provided waterproong for up to 16 weeks during in vivo strain measurement collec-
tion.
56,68
After wiring the gauge (Figure 20.4A), the rst step in the coating process is to strengthen
mechanically and waterproof the gaugewire junction (Figure 20.4B) with a small bead of PMMA
(Kerr Fastcure Sybron, Romulus, MI). Once the PMMA has set and dried for 4 h, the gaugewire
junction is coated with a layer of M Coat B (Measurements Group, Raleigh, NC), an air-drying
nitrile rubber, and allowed to dry for an additional 24 h (Figure 20.4C). The nitrile rubber acts as
a coupling agent between the next waterproong coating and the wire insulation and manufacturers
polyimide gauge backing. A thin layer of M Coat D (Measurements Group, Raleigh, NC), an air-
drying acrylic, is placed on the gauges next and allowed to dry for 24 h (Figure 20.4D). At this
point an ultrane-point permanent marker can be used to assign the gauge a number or to mark
the orientation of specic elements of a rosette. The acrylic coating provides a waterproof coating
on the gauge and gaugewire junction but is brittle and easily damaged by surgical instruments.
The nal coating is a layer of M Coat A (Measurements Group, Raleigh, NC), an air-drying
polyurethane, which is allowed to cure for 7 days (Figure 20.4E). The primary purpose of this
coating is to protect the acrylic layer from surface damage.
In preparation for surgery, the wired gauges must be sterilized. Ethylene oxide sterilization
assures minimum thermal damage to gauges and coatings. Sterile wired gauges should be double
aerated since residual ethylene oxide trapped in the wires can leach out in vivo and cause tissue
irritation and necrosis.
VII. ATTACHMENT OF GAUGES
A. ATTACHMENT FOR IN VITRO STUDIES
Cyanoacrylates, such as M Bond 200 (Measurements Group, Raleigh, NC), are the most common
adhesives used in vitro. The bone is rst cleaned of all soft tissue and the periosteum is removed.
Depending on the nature of the tests, specimens may be machined from the cleaned bone, or the
whole bone used as is. The location on the specimen at which the gauge is to be bonded is marked
and then dried thoroughly. Aggressive degreasing and scraping as suggested by Wright and Hayes
40
is not necessary for satisfactory adhesive performance. However, if the gauge is to be applied in
an area containing periosteal blood vessels or in an area in which the bone is very thin, the area
should be sealed and prepared for the adhesive. This is usually done by applying a thin layer of
the adhesive catalyst and allowing it to dry. Then one or two drops of the adhesive are spread thinly
over the bone surface and allowed to dry for about 5 min. Next the area is smoothed down with a
ne grit (300 to 400) silicon-carbide paper.
43
To attach the gauge, one drop of the adhesive is placed
on the prepared surface and the gauge is held against the surface for 1 min using nger pressure.
This is enough to develop a strong bond, and the bone can then be returned to its saline environment.
Care must be taken to align the gauge correctly during the installation. An established technique
such as the one described by Wright and Hayes
40
may be used.
B. ATTACHMENT FOR IN VIVO STUDIES
The most commonly used adhesives for in vivo applications are the cyanoacrylates. Isobutyl
cyanoacrylate bonds quickly and, when used at relatively dry sites, may hold gauges in place for
several weeks. A single-component form of this adhesive such as VetBond (3M Animal Care
0266/ ch20/frame Page 314 Friday, October 22, 1999 9:20 AM
Strain Gauge Measurements from Bone Surfaces 315
Products, St. Paul, MN) or an adhesive which comes with a separate accelerator such as M-Bond
200 can be used. When using the one-component system it is best to place a small amount of
adhesive on the bone surface prior to gauge attachment and allow it to dry. This procedure will
seal the bone surface. This region should be scraped slightly prior to gauge attachment. A drop of
adhesive should then be placed on the gauge and the gauge pressed against the bone surface for
1 min. The tip of a blunt surgical instrument can be used to hold miniature gauges in place while
the glue sets. Care should be taken not to let any glue get on instruments, gloves, or tissues.
FIGURE 20.4 This series of diagrams shows the steps used to wire and waterproof a strain gauge for use
in vivo. Note that the PMMA in B is tapered to cover the gaugewire junction but also to maintain the exibility
of the gauge. The rubber layer should adequately cover the junction and a short region of the cable housing.
The remaining two waterproong layers should cover the junction and the entire gauge.
0266/ ch20/frame Page 315 Friday, October 22, 1999 9:20 AM
316 Mechanical Testing of Bone and the BoneImplant Interface
One should be aware that cyanoacrylate adhesives experience biodegradation and that mea-
surements collected from glued gauges may vary in accuracy from day to day. Adhesive degradation
mandates the testing of gauge accuracy following the completion of in vivo measurements using
some form of bench-top testing. Three approaches have been reported. The rst involves loading
the bones with a known load, collecting strains from the gauge which had been used for in vivo
bone strain measurements, then replacing the gauge with a freshly glued gauge in the same location
and retesting the bone.
30
The second procedure utilizes a second strain measurement technique
(either optical or electronic extensometry) to collect bone strain while simultaneously collecting
strain measurements from the gauge which had been used for in vivo bone strain measurements.
38,39
The third technique requires explanting and bench-top testing both the bone which was originally
gauged and the contralateral bone.
63-68
Validity of strain comparisons to measurements from freshly
attached gauges on contralateral bones, during cantilever bending, have been conrmed.
84,85
C. PREPARATION OF CPC-COATED GAUGES FOR EXTENDED IN VIVO STRAIN MEASUREMENT
As noted earlier, cyanoacrylate bonding of gauges to bone in vivo is acceptable for short-term
measurement. However, adhesive failure precludes the use of these systems for long-term measure-
ment. CPC-coated gauges can be used instead. After wiring and waterproong gauges using the
three-stage process described earlier, gauges should be sanded with 600-grit carbide paper on their
sensing surfaces. This surface is coated with a layer of 15 wt.% Udel

, a medical-grade polysulfone
(Amoco, Huntington Beach, CA) dissolved in 1,1,2,2-tetrachloroethane (Eastman Kodak Company,
Rochester, NY). Coated gauges must then be baked for 1 h at 90C. This rst-baked polysulfone
layer is then sanded with 1200-grit carbide paper. Another coating of polysulfone is applied,
followed by CPC particles sprinkled onto the moist polysulfone. A combination of an amorphous
tricalcium phosphate with a particle size of 9 7 mm and a microcrystalline CPC with a particle
size of approximately 561 85 mm (Biointerfaces, San Diego, CA) have produced an ideal
combination of bone bonding and gaugeCPC bond stability in vivo.
42
Coated gauges will then be
baked for 5 h at 90C. Following this preparation process, gauges must be sterilized with ethylene
oxide and double-aerated prior to implantation.
VIII. MEASUREMENT COLLECTION
In order to measure strains from gauges, the wire leads from the gauge must be connected to a
signal conditioning and recording system. A hard-wired system can successfully be used for this
purpose. However, if extensive freedom of movement is required, a radio telemetry system may
be a better choice. Hard-wired systems are typically made up of a signal conditioning circuit which
provides a balance capability (often in the form of a Wheatstone bridge) to offset the initial resistance
of the gauge, lead wires, and connectors. They generally also contain an electronic signal calibration
capability that can be used to place a known resistance into the circuit so that voltage signals from
the signal conditioning circuit can be accurately converted to strains. A recording system attached
to the signal conditioning circuit should be able to record the voltage changes resulting from strain
changes accurately. A continuous recording device such as a multipen recorder with a broad
frequency response and high slew rate can be used. Recording systems will be a function of the
needs of the experiment, however. For most in vivo applications a recording device which can
accurately monitor strain changes at sampling rates of no less than 1000 Hz should be selected.
Alternatively, a digital recording system provides more data reduction and presentation exibility.
A system that has been precongured such as the Measurements Group System
6000 (Measurements Group, Raleigh, NC) is a good choice if one computer is likely to be dedicated
solely to strain and load measurement collection. LabView (Signalogic, Dallas, TX) software
used in conjunction with a data acquisition board (National Instruments Corporation, Austin, TX)
is a more exible data collection system that may be a better choice if a variety of measurements
0266/ ch20/frame Page 316 Friday, October 22, 1999 9:20 AM
Strain Gauge Measurements from Bone Surfaces 317
from various types of test devices will be used. LabView is more difcult to congure, however,
and a signicant time commitment may be required to get this software congured to the specic
needs of individual users. A number of other data collection systems are also available and may
do an equally acceptable job. When evaluating a system for in vivo strain measurement collection,
the sampling rate (which should not be less than 1000 Hz) and number of channels it can simul-
taneously accommodate are of primary consideration. Real-time strain display and the ability to
monitor other equipment, such as video or electromyograms (EMGs) which could be used for limb
tracking, are also valuable features to have in a hard-wired measurement system.
If the measurements to be collected require that the test animal or subject have complete freedom
of movement, a radio telemetry system can provide signal conditioning and measurement trans-
mission to a receiver and recording device. These systems are for the most part custom-made and
have often been made by investigators who collect the strain measurements. Companies such as
Konigsberg (Pasadena, CA), MicroStrain (Burlington, VT), and CME Telemetrix (Waterloo,
Ontario, Canada) have built biomedical measurement systems in the past and have produced custom-
made radio telemetry systems. The transmission signal stability and measurement reproducibility
of digital radio telemetry systems are generally far superior to analog radio telemetry systems.
However, currently available subminiature implantable radio transmitters suffer from slow sampling
rates and remote powering challenges.
Whether the system chosen to collect measurements is a hard-wired or radio telemetry system,
measurements should be checked for validity prior to measurement collection. Custom-made
systems which do not have built-in signal calibration capabilities must be bench-top calibrated
using an external calibration procedure. The validity of measurements made by a system can be
checked by comparing strains collected from metal test specimens with well-characterized material
properties against either (1) strains computed using beam theory or (2) strains measured with
another device such as an extensometer.
Finally, in preparation for in vivo strain measurement a zero strain state must be selected.
Although no concensus on an ideal point has been reached, zeroing of signals is often carried out
with the test animal in a resting posture.
54
Alternatively, the zero can be selected as the strain state
during swing phase
56
or when a particular parameter is at a minimum.
86
In this case the strains
must be computed following measurement collection relative to the swing phase strain state.
IX. DATA ANALYSIS
Analysis of uniaxial strains in one location generally provides little information about the strain
state of a bone. Multiple gauges around a bone section and along the length of a long bone offer
much more information about its strain state. A linear elastic response of bone material properties
in relatively uniform cross sections of long bones can be assumed at the load rates associated with
gait. As such, the strain state of these bones can be interpolated from measured strains using simple
beam analysis. If an implant is present, or if relatively complex regions of a bone are analyzed,
nite-element analysis must be used in conjunction with strain measurements to evaluate the strain
state of the bone between gauges.
Analysis of strains from rosette gauges allows evaluation of strains along each element of the
gauge and also allows evaluation of maximum and minimum principal strains at the measurement
location. The equations to calculate principal strains are dependent on the type of gauge used and
can be found in texts that have chapters describing strain analysis.
72
Strains can also be converted
into stress if an accurate value for modulus of the bone under consideration is available. One should
keep in mind that bone is not isotropic and that its properties vary as a function of location. Whether
single-element or rosette gauges are used for in vivo studies, investigators must make assumptions
about the zero strain point. Careful consideration of these factors and the limitations of the in vitro
and in vivo strain measurement process will provide a reasonable understanding of the deformation
of bone in various situations.
0266/ ch20/frame Page 317 Friday, October 22, 1999 9:20 AM
318 Mechanical Testing of Bone and the BoneImplant Interface
REFERENCES
1. Little, E.G. and Finlay, J.B., Perspectives of strain measurement techniques, in Strain Measurement
in Biomechanics, Miles, A.W. and Tanner, K.E., Eds., Chapman & Hall, London, 1992, 1.
2. Finlay, J.B., Bourne, R.B., Landsberg, R.P., and Andreae, P., Pelvic stresses in vitro, I: malsizing of
endoprostheses, J. Biomech., 19, 708, 1986.
3. Finlay, J.B., Bourne, R.B., Landsberg, R.P., and Andreae, P., Pelvic stresses in vitro, II: a study of
the efcacy of metal-backed acetabular prostheses, J. Biomech., 19, 715, 1986.
4. Hasenkam, J.M., Nygaard, H., Paulsen, P.K., et al., What force can the myocardium generate on a
prosthetic mitral valve ring? An animal experimental study, J. Heart Valve Dis., 3, 324, 1994.
5. Nagata, H., Schendel, M.J., Transfeldt, E.E., and Lewis, J.L., The effects of immobilization of long
segments of the spine on the adjacent and distal facet force and lumbosacral motion, Spine, 18, 2471,
1993.
6. Bowman, S.M., Zeind, J., Gibson, L.J., et al., The tensile behavior of demineralized bovine cortical
bone, J. Biomech., 29, 1497, 1996.
7. Roth, V. and Mow, V.C., The intrinsic tensile behavior of the matrix of bovine articular cartilage and
its variation with age, J. Bone Joint Surg., 62, 1102, 1980.
8. Page, A.E., Jasty M., Harrigan, T. P., et al., Determination of loading parameters in the canine hip in
vivo, J. Biomech., 26, 571, 1993.
9. Hylander, W.L., Johnson K.R., and Crompton A.W., Loading patterns and jaw movements during
mastication in Macaca fascicularis: a bone-strain, electromyographic, and cineradiographic analysis,
Am. J. Phys. Anthropol., 72, 287, 1987.
10. Swartz, M.S., Bennett, M.B., and Carrier, D.R., Wing bone stresses in free ying bats and the evolution
of skeletal design for ight, Nature, 359, 726, 1992.
11. Lanyon, L.E. and Baggott, D.G., Mechanical function as an inuence on the structure and form of
bone, J. Bone Joint Surg., 58, 436, 1976.
12. Goodship, A.E., Lanyon, L.E., and McFie, H., Functional adaptation of bone to increased stress: an
experimental study, J. Bone Joint Surg., 61A, 539, 1979.
13. Mikic, B. and Carter, D.R., Bone strain gage data and theoretical models of functional adaptation,
J. Biomech., 28, 465, 1995.
14. Cheal, E.J., Hayes, W.C., White, A.A., and Perren, S.M., Three-dimensional nite element analysis
of a simplied compression plate xation system, J. Biomech. Eng., 106, 295, 1984.
15. Rohlmann, A., Mossner, U., Bergmann, G., and Kolbel, R., Finite element analysis and experimental
investigation in a femur with hip endoprosthesis, J. Biomech., 16, 727, 1983.
16. Miles, A.W. and Tanner, K.E., Eds., Strain Measurements in Biomechanics, Chapman & Hall, London,
1992.
17. Dabestani, M., In vitro strain measurement in bone, in Strain Measurement in Biomechanics, Miles,
A.W. and Tanner, K.E. Eds., Chapman & Hall, London, 1992, 58.
18. Thompson, K., On the electro-dynamic qualities of metals, Philos. Trans. R. Soc. London, 146, 649,
1856.
19. Eaton, E.C., Resistance strain gauge measures stresses, Concr. Eng. News Rec., 107, 615, 1931.
20. Clark, D.S. and Datwyler, G., Stressstrain relations under tension impact loading, Proc. ASM, 38,
98, 1938.
21. Simmons, E.E., U.S. Patent No. 2292549, 1942.
22. Measurements Group, TN-505, Technical Notes, Raleigh, NC, 1996.
23. Gurdjian, E.S. and Lissner, H.R., Mechanism of head injury as studied by the cathode ray oscilloscope:
preliminary report, J. Neurosurg., 1, 393, 1944.
24. Evans, F.G., Methods of studying the biomechanical signicance of bone form, Am. J. Phys. Anthro-
pol., 11, 413, 1953.
25. Roberts, V.L., Strain-gauge techniques in biomechanics, Exp. Mech., 19A, 1966.
26. Cameron, J.J., Boodward, S.C., Pulaski, E.J., et al., The degradation of cyanoacrylate tissue adhesives,
Surgery, 58, 424, 1965.
27. Leonard, F., Kulkarni, R.K., Brandes, G., et al., Synthesis and degradation of poly(alkylalphacy-
anoacrylates), J. Appl. Polym. Sci., 10, 259, 1966.
0266/ ch20/frame Page 318 Friday, October 22, 1999 9:20 AM
Strain Gauge Measurements from Bone Surfaces 319
28. Lanyon, L.E. and Smith, R.N., Bone strain in the tibia during normal quadrupedal locomotion, Acta
Orthop. Scand., 41, 238, 1970.
29. Lanyon, L.E., Strain in sheep lumbar vertebrae recorded during life, Acta Orthop. Scand., 42, 102, 1971.
30. Cochran, G.V., Implantation of strain gages on bone in vivo, J. Biomech., 5, 119, 1972.
31. Lanyon, L.E., Analysis of surface bone strain in the calcaneus of sheep during normal locomotion,
J. Biomech., 6, 41, 1973.
32. Lanyon, L.E., Hampson W.G., Goodship A.E., and Shah J.S., Bone deformation recorded in vivo from
strain gauges attached to the human tibial shaft, Acta Orthop. Scand., 46, 256, 1975.
33. Rydell, N.W., Forces acting on the femoral head-prosthesis. A study of strain gauge supplied prostheses
in living persons, Acta Orthop. Scand. 37 (Suppl. 88), 1, 1966.
34. Cochran, G.V., A method for direct recording of electromechanical data from skeletal bone in living
animals. Technical notes, J. Biomech., 7, 563, 1974.
35. Lanyon, L.E. and Hartman, W., Strain related electrical potentials recorded in vitro and in vivo, Calcif.
Tissue Res., 22, 315, 1977.
36. Baggott, D.G. and Lanyon, L.E., An independent post-mortem calibration of electrical resistance strain
gauges bonded to bone surfaces in vivo, J. Biomech., 10, 615, 1977.
37. Carter, D.R., Smith, D.J., Spengler, D.M., et al., Measurement and analysis of in vivo bone strains on
the canine radius and ulna, J. Biomech., 13, 27, 1980.
38. Hirsch, C. and Frankel, V., The reaction of the proximal end of the femur to mechanical forces, in
Biomechanical Studies of the Musculoskeletal System, Evans, F.G., Ed., Charles C Thomas, Spring-
eld, IL, 1961.
39. Lambert, K.L., The weight-bearing function of the bula, J. Bone Joint Surg., 53A, 507, 1971.
40. Wright, T.M. and Hayes, W.C., Strain gage application on compact bone, J. Biomech., 12, 471, 1979.
41. Huiskies, R., On the modeling of long bones in structural analysis, J. Biomech., 15, 65, 1982.
42. Gies, A.A. and Carter, D.R., Experimental determination of whole long bone sectional properties,
J. Biomech., 15, 297, 1982.
43. Finlay, J.B., Bourne, R.B., and McLean, J.A., Technique for the in vitro measurement of principal
strains in the human tibia, J. Biomech., 15, 723, 1982.
44. Kennedy, J.G., Carter, D.R., and Caler, W.E., Long bone torsion, II. A combined experimental and
computational method for determining an effective shear modulus, J. Biomech. Eng., 107, 189, 1985.
45. Arner, M. and Hagberg, L., Wrist exion strength after excision of the pisiform bone, Scand. J. Plast.
Reconstr. Surg., 18, 241, 1984.
46. Bourne, R.B., Finlay, J.B., Papadopoulos, P., et al., In vitro strain distribution in the proximal tibia.
Effect of varus-valgus loading in the normal and osteoarthritic knee, Clin. Orthop., 188, 285, 1984.
47. Bourne, R.B., Finlay, J.B., Papadopoulos, P., and Andreae, P., The effect of medial menisectomy on
strain distribution in the proximal part of the tibia, J. Bone Joint Surg., 66, 1431, 1984.
48. Lanyon, L.E. and Bourn, S., The inuence of mechanical function on the development and remodeling
of the tibia, J. Bone Joint Surg., 61A(2), 263, 1979.
49. Rubin, C.T. and Lanyon, L.E., Limb mechanics as a function of speed and gait: a study of functional
strains in the radius and tibia of horse and dog, J. Exp. Biol., 101, 187, 1982.
50. Rubin, C.T. and Lanyon, L.E., Dynamic strain similarity in vertebrates; an alternative to allometric
limb bone scaling, J. Theor. Biol., 107, 321, 1984.
51. Hylander, W.L. and Johnson, K.R., In vivo bone strain patterns in the zygomatic arch of macaques
and the signicance of these patterns for functional interpretation of craniofacial form, Am. J. Phys.
Anthropol., 102, 203, 1997.
52. Carter, D.R., Vasu, R., Spengler, D.M., and Dueland, R.T., Stress elds in the unplated and plated
canine femur calculated from in vivo strain measurements, J. Biomech., 14, 63, 1981.
53. Lanyon, L.E., Paul, I.L., Rubin, C.T., et al., In vivo strain measurements from bone and prosthesis
following total hip replacement, J. Bone Joint Surg., 63A, 989, 1981.
54. Szivek, J.A., A Quantitative Study of the Effect of Strain Redistribution on Bone Remodelling, Ph.D.
thesis, University of Toronto, Canada, 1984.
55. Szivek, J.A., Magee, F.P., Weng, M.S., and Johnson, E.M., Comparative in vivo strain measurement
from a two year implanted and unimplanted canine hip, Trans. Int. Soc. Biomater., Kyoto, Japan, 1988.
56. Szivek, J.A., Johnson, E.M., and Magee, F.P., In vivo strain analysis of the greyhound femoral
diaphysis, J. Invest. Surg., 5, 91, 1992.
0266/ ch20/frame Page 319 Friday, October 22, 1999 9:20 AM
320 Mechanical Testing of Bone and the BoneImplant Interface
57. Loitz, B.J. and Zernicke, R.F., Strenuous exercise-induced remodeling of mature bone: relationships
between in vivo strains and bone mechanics, J. Exp. Biol., 170, 1, 1992.
58. Biewener, A.A. and Bertram, J.E., Skeletal strain patterns in relation to exercise training during growth,
J. Exp. Biol., 185, 51, 1993.
59. Biewener, A.A. and Bertram, J.E., Structural response of growing bone to exercise and disuse, J. Appl.
Physiol. 76, 946, 1994.
60. Skerry, T.M. and Lanyon, L.E., Interruption of disuse by short duration walking exercise does not
prevent bone loss in sheep calcaneous, Bone, 16, 269, 1995.
61. Turner, C.H., Yoshikawa, T., Forwood, M.R., et al., High frequency components of bone strain in dogs
measured during various activities, J. Biomech., 28, 39, 1995.
62. Keller, T.S. and Spengler, D.M., In vivo strain gage implantation in rats, J. Biomech., 15, 911, 1982.
63. Szivek, J.A., Wilson, D.L., Anderson, P.L., and DeYoung, D., Development of a model for the study
of in vivo bone strains during normal and microgravity environments, J. Appl. Biomater., 6, 203, 1995.
64. Burr, D.B., Milgrom, C., Fyhrie, D., et al., In vivo measurement of human tibial strains during vigorous
activity, Bone, 18, 405, 1996.
65. Szivek, J.A., Gealer, R.G., Magee, F.P., and Emmanual, J., Preliminary development of a hydroxya-
patite-backed strain gauge, J. Appl. Biomater., 1, 241, 1990.
66. Maliniak, M.M., Szivek, J.A., and DeYoung, D.W., The development of hydroxyapatite coated strain
gauges for long-term in vivo bone loading response measurements, J. Appl. Biomater., 4, 143, 1993.
67. Szivek, J.A., Anderson, P.L., and DeYoung, D.W., Evaluation of factors affecting bonding rate of
calcium phosphate ceramic coatings for in vivo strain gauge attachment, J. Biomed. Mater. Res., 33,
121, 1996.
68. Szivek, J.A., Anderson, P.L., and DeYoung, D.W., In vivo strain measurements collected using calcium
phosphate ceramic bonded strain gauges, J. Invest. Surg., 10, 263, 1997.
69. Battraw, G.A., Szivek, J.A., and Anderson, P.L., Interface strength studies of calcium phosphate
ceramic-coated strain gauges, J. Biomed. Mater. Res., 43, 462, 1998.
70. Szivek, J.A., Battraw, G.A., and Anderson, P.L., Calcium phosphate ceramic/polysulfone/polyimide
interface characteristics after 4, 6 and 9 months in vivo, Trans. Surf. Biomater., 163, 1998.
71. Beckwith, T.G., Marangoni, R.D., and Lienhard, J.H., Mechanical Measurements, Addison-Wesley,
New York, 1993.
72. Dalley, J.W. and Riley, W.F., Experimental Stress Analysis, McGraw-Hill, New York, 1991.
73. Murray, W.M. and Miller, W.R., The Electric Resistance Strain Gauge: An Introduction, Oxford
University Press, Oxford, U.K., 1992.
74. BLH Electronics Inc., Catalogue 100-6, Canton, MA, 1998.
75. Perry, C.C., The electric resistance strain gauge revisited, Expl. Mech. 24, 286, 1984.
76. Pople, J., DIY strain gauge transducers, Strain, 16, 23, 1980.
77. Perry, C.C., Strain gauge misalignment errors, Instrum. Control Syst., 42, 137, 1969.
78. Tuttle, M.E. and Brinson, H.F., Resistance foil strain gauge technology as applied to composite
mechanics, Expl. Mech., 24, 54, 1984.
79. Malvern, L.E., Introduction to the Mechanics of a Continuous Medium, Prentice-Hall, Englewood
Cliffs, NJ, 1969.
80. Hiatt, M.J., In Vivo Strains in Bone near Transcortical Implants, Thesis, Clemson University, Clemson,
SC, 1996.
81. Greenberg, S.W., Gonzalez, G., Gurdjian, E.S., and Thomas, L.M., Changes in the physical properties
of bone between the in vivo, freshly dead and embalmed conditions, in 12th Stapp Car Crash
Conference, Society of Automotive Engineers, New York, 1968, 272.
82. Evans, F.G., The Mechanical Properties of Bone, Charles C Thomas, Springeld, IL, 1973.
83. Broz, J.J., Simske, S.J., Greenberg, A.R., and Luttges, M.W., Effects of rehydration state on the exural
properties of whole mouse long bones, J. Biomech. Eng., 115, 447, 1993.
84. Kersey, R.C., Szivek, J.A., and Sacoman, D.M., Symmetry of biomechanical properties of canine
femora, J. Appl. Biomater., 5, 99, 1994.
85. Battraw, G.A., Miera, V., Anderson, P.L., and Szivek, J.A., Symmetry of biomechanical properties in
rat femora, J. Biomed. Mater. Res., 32, 285, 1996.
86. Bessman, E.S., Carter, D.R., McCarthy, J.C., and Harris, W.H., Accuracy enhancement of in vivo bone
strain measurements and analysis, J. Biomech. Eng., 104, 226, 1982.
0266/ ch20/frame Page 320 Friday, October 22, 1999 9:20 AM

321

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

21

Screw Pullout Test for Evaluating
Mechanical Properties of Bone

Matthew S. Crum, Franklin A. Young, Jr., and Yuehuei H. An

CONTENTS

I. Introduction ..........................................................................................................................321
II. Examples ..............................................................................................................................321
A. Screw Pullout Strength and Bone Structure..................................................................321
B. Effects of Vertebral Cortices and Insertion Angle on Screw Pullout Strength.............323
C. Effects of Preservation Methods on Bone Screw Pullout Strength ..............................326
D. Effects of Bone Density on Screw Pullout Strength.....................................................327
III. Summary ..............................................................................................................................328
References ......................................................................................................................................328

I. INTRODUCTION

Screw pullout testing refers to the measurement of the force required to pull out a screw inserted
in a bone specimen. Analysis of the test allows determination of the optimum screw size, insertion
technique, angle of penetration, and screw hole preparation method. While all of these variables
play a major role in the success of the screw xation, another parameter that must be explored is
the impact of bone structure on the force required to pull the screw. This chapter primarily focuses
on the effects of bone microstructure, density, and pathological conditions on the screw pullout
test. The chapter introduces the methods of testing the mechanical variables of bone using one
screw, but will not discuss the methods for testing the effects of xation of different type screws
in one bone (for the latter, see Chapter 36).

II. EXAMPLES
A. S

CREW

P

ULLOUT

S

TRENGTH



AND

B

ONE

S

TRUCTURE

The screw pullout test basically involves the insertion of screws of various sizes and at several
different angles into a bone specimen. This analysis is necessary due to the heterogeneous structure
of the bone. If the bone were a simple homogeneous material, the location and angle of insertion
would be irrelevant. However, because bone consists of a complex columnstrut arrangement of
trabeculae (Figure 21.1B), the direction and location of screw penetration are vital parameters that
must be considered, evaluated, and optimized.
Recently, An et al.

1,2

investigated the correlation between the angle of screw insertion and
trabeculae orientation. Because more trabeculae lie in the weight-bearing, or vertical, direction,
they hypothesized that the screw would have an optimum strength at 0

to vertical (parallel to
trabeculae), minimum strength at 90

(perpendicular to trabeculae), and an intermediate strength
at 45

(Figure 21.1C and D). To insure that the angle of screw insertion would be the only

0266/ ch21/frame Page 321 Friday, October 22, 1999 9:24 AM

322

Mechanical Testing of Bone and the BoneImplant Interface

variable tested in the experiment, An et al. used identical screws (40 mm long

3.5 mm thread
diameter self-tapping cortical screws). The samples were small cubes (at least 2.5

2.5

4.0
cm) extracted from adult bovine femoral condyles at two different levels (Figure 21.1A). The
corners of the specimens used for the 45

insertion were chamfered to ensure optimum penetration
(Figure 21.1D).

FIGURE 21.1

Illustration of a test for screw-holding powers of bovine cancellous bone from different
directions. (A) The sampling site and orientation. (B) The structure of the bone block is idealized into a
column-strut model. The columns indicate thicker trabeculae at the weight-bearing direction and the struts
represent the trabecular connections between columns (column trabeculae). It is assumed that the columns
are thicker and stronger and the struts are thinner and weaker. (C and D) The directions of screw insertion.
(From An, Y.H. and Draughn, R.A., Mechanical properties and testing methods of bone, in

Animal Models in
Orthopaedic Research,

An, Y.H. and Friedman, R.J., Eds., CRC Press, Boca Raton, FL, 1999. With permission.)

0266/ ch21/frame Page 322 Friday, October 22, 1999 9:24 AM

Screw Pullout Test for Evaluating Mechanical Properties of Bone

323

Screw pullout tests were performed on the different specimens (with the screw inserted at
different angles to the main trabecular direction) using a mechanical testing machine (MTS System
810, Minneapolis, MN), at a displacement rate of 1 mm/min. The loaddisplacement curve was
obtained using a chart recorder. The slope of the linear portion of the curve (before the yield stress
was achieved and permanent deformation occurred) was investigated to determine the stiffness of
the bonescrew interface. From these measurements, the ultimate strength

(MPa) was calculated
using the following formula:

=

P

/

dh

(21.1)
where

P

(N) was the ultimate applied load,

d

(mm) was the major diameter of the screw, and

h

(mm) was the length of the effective threads in the screw in the cancellous bone. An ANOVA test
was then used to determine the holding strengths for each of the three angles of insertion. X rays
were also taken of 3-mm bone slices from the specimens for radiographic analysis. Sections 5 mm
thick were then cut from the 3-mm bone slices, dehydrated, and stained for further analysis under
a light microscope.
Histological analysis showed trabeculae lying in the weight-bearing direction (Figure 21.2B).

1,2

A histological image of the cross section of the specimen revealed a honeycomb-like structure,
which represented the tops of the trabeculae (Figure 21.2A). The oblique orientation of trabeculae
shown in Figure 21.2C represented the 45

insertion. The ANOVA clearly supported the hypothesis.
The screw insertions at 0

had signicantly higher strengths because they were in the weight-
bearing direction (where the plates were subjected to compressive loading). Conversely, those
penetrating at 90

performed the worst due to the bending moments they experienced. The 45

screws were affected by both compressive and bending loads but to a lesser extent. Hence, this
orientation possessed an intermediate strength. The maximum stresses for the 0, 45, and 90

orientations were 54

5.4, 43

3.9, and 37

5.0 MPa at applied loads of 840

82, 655

59,
and 560

76 N, respectively.
In addition, a nite-element analysis (FEA) was performed on the specimens to investigate
further the bone microstructure and its impact on pullout strength.

3

A 60

section was chosen for
modeling because the bonescrew samples are axisymmetric. A model consisting of 2982 elements
was employed in the 0 and 90

testing. The 45

analysis involved 5964 six-noded pentahedral
elements. The model was subjected to 3.6, 2.8, and 2.4 N concentrated loading at the implant
surface over 39 nodes about 360

. The results clearly showed a strong correlation in strength and
ultimate load between the FCA data and the measurements from the pullout test (Table 21.1). This
experiment also presents a strong argument on the effects of trabeculae orientation on the ultimate
pullout power of screws.

B. E

FFECTS



OF

V

ERTEBRAL

C

ORTICES



AND

I

NSERTION

A

NGLE



ON

S

CREW

P

ULLOUT

S

TRENGTH

A similar study was performed by Horton et al.

4

Their analysis consisted of two experiments. The
rst of these introduces another variable affecting the screw pullout test, unicortical vs. bicortical
predrilling. Unicortical and bicortical drilling simply refer to the screw predrill depth. A 4.5-mm-
diameter drill bit was used to penetrate ve cadaver thoracic vertebrae to two different depths. Half
of the specimens were drilled only in the rst cortex (unicortical) and the other half were drilled
completely through the rst and second cortices (bicortical). These samples were then subjected
to pullout tests. Horton et al. observed no statistical variation in screw pullout strengths between
those samples subjected to bicortical rather than unicortical predrilling.
The second experiment performed by Horton et al.

4

offers more insight into the signicance
of screw orientation and the effects of vertebral bone structure. Using cadaver thoracic spines
Horton et al. employed screw insertion at traditional midbody, superior endplate, superior oblique,
and inferior oblique orientations (Figure 21.3) on 48 vertebrae. Although the 0

screw orientation

0266/ ch21/frame Page 323 Friday, October 22, 1999 9:24 AM

324

Mechanical Testing of Bone and the BoneImplant Interface

of An et al. (screw insertion in the weight-bearing direction) was not studied, Horton et al. did
experiment with off-center drilling locations (Figure 21.3B) and insertion angles below horizontal
(Figure 21.3D). Allowing all insertion techniques and preparations (with the exception of the screw
angle) to remain analogous, the researchers inserted a Danek TSRH 6.5-mm-diameter vertebral
screw at the various orientations until only one thread was visible between the specimen surface
and screw head, indicating the same depth. The results revealed the strongest orientation to be the
superior oblique and the weakest alignment to be the inferior oblique.
The poor performance of screws inserted in the traditional midbody and inferior oblique
orientations can once again be attributed to the microstructure of the bone sample. By common
knowledge the trabeculae are thicker and more robust at the ends of the vertebrae (Figure 21.4).

FIGURE 21.2

Histological sections showing the relations between the directions of screw insertion and
trabecular orientations. (A) The axial view of 0

insertion; (B) side view of 90

insertion; and (C) side view
of 45

insertion.

0266/ ch21/frame Page 324 Friday, October 22, 1999 9:24 AM

Screw Pullout Test for Evaluating Mechanical Properties of Bone

325

TABLE 21.1
Mechanical Parameters of the Pullout Test and FEA Results
(mean

SEM,

N

= 16, ANOVA test for repeated measures)

Testing Method Direction
Ultimate Load
(N)
Stiffness
(N/mm)
Strength
(MPa)

Pullout test 0

818

82 2,833

297 54.9

5.4
45

649

59 2,237

196 42.8

3.9
90

563

76 1,955

219 37.1

5.0

P

value <0.0001 <0.001 <0.0001
Finite element 0

840 18,000 53
Analysis 45

655 16,200 43
90

560 14,800 37

FIGURE 21.3

Different screw orientations in a vertebral body (Adapted from Horton et al.

4

)

0266/ ch21/frame Page 325 Friday, October 22, 1999 9:24 AM

326

Mechanical Testing of Bone and the BoneImplant Interface

The midsection of the specimen consists of looser, generally mechanically weaker trabeculae. In
addition, vertebrae are thinner in the middle than at the upper and lower ends. Hence, not only are
the screws being forced into less bone from a dimensional standpoint, they are also penetrating
less dense bone. Therefore, the performance of screws inserted in the traditional midbody orientation
will quite obviously not fare as well as those screws piercing denser and more massive portions
of the vertebrae. The thicker arrangement of trabeculae at the ends of the vertebral bone would
also explain why Horton et al. saw a higher pullout strength in the superior endplate orientation.
Without a careful morphological evaluation of vertebral bone, however, there is no explanation for
the poor performance of the inferior oblique insertion.
Clearly, based on work by Horton et al.,

4

An et al.,

1,2

and Nicholson et al.

5

at least the angle
in which the screw is inserted with respect to the trabeculae orientation and the density of the
bone are two determining factors on the holding power of screw xation. However, there are
additional specimen characteristics that must be taken into account. These variables include the
duration and methods utilized in bone preservation as well as the overall condition, health, and
age of the bone specimen.

C. E

FFECTS



OF

P

RESERVATION

M

ETHODS



ON

B

ONE

S

CREW

P

ULLOUT

S

TRENGTH

Roe et al.

6

investigated the effects of four different preservation methods on the mechanical prop-
erties of canine femoral bones. One femur of each animal was chosen as experimental and the
other represented its matched control. The control specimens were cooled to 4

C and subjected to
mechanical testing. The rst group, aseptic collection, consisted of 15 limbs prepared for aseptic
surgery with the femurs removed. The epiphyseal sections and all soft tissues were extracted and
the bone was sealed in sterilized polyethylene tubing and stored at 20

C. Group 2 (ethylene oxide)
was composed of 15 cleanly harvested bones dried for 18 to 24 h at room temperature, packaged
in polyethylene tubing, and sterilized in 12% ethylene oxide for 2 h, aerated for 8 h, and stored at
room temperature. Group 3 (chemical sterilization) contained 15 cleanly harvested femurs immersed
in a methanol:chloroform solution (1 : 1) for 24 h, and moved to a ltered solution of 10 m

M

iodoacetic acid in 0.1

M

phosphate-buffered NaCl solution for 48 h before they were nally sealed
in sterilized polyethylene tubing and stored at 20

C. The fourth group (chemical sterilization and
partial decalcication) consisted of 15 cleanly harvested bones immersed in a methanol:chloroform
solution (1 : 1) for 4 h, then in 0.6

N

hydrochloric acid solution for 24 h, and nally in lter-
sterilized, 10 m

M

iodoacetic acid and 10 m

M

sodium azide in 0.1

M

phosphate-buffered NaCl

FIGURE 21.4

Macrophotograph of a vertebral section showing differences in bone density (trabeculae)
between middle and end portion. (From Brien, E. and Healey, J.H., Orthopedic aspects of osteoporosis, in

Osteoporosis Diagnosis and Treatment

, Sartoris, D.J., Ed., Marcel Dekker, New York, 1996. With permission.)

0266/ ch21/frame Page 326 Friday, October 22, 1999 9:24 AM

Screw Pullout Test for Evaluating Mechanical Properties of Bone

327

solution for 72 h. The bones were then stored identically to the Group 3 and 4 samples. The storage
times were 1, 16, and 32 weeks for each treatment.
When subjected to compressive loading to failure the aseptically collected bones stored for
1 week were signicantly stronger than the control bones. The comparison after 16 and 32 weeks
was much less apparent. Similarly, there was no substantial difference in screw pullout testing for
this group at 1, 16, or 32 weeks of storage. Bones sterilized in ethylene oxide withstood compressive
loading slightly better than the aseptically collected specimens for 1-week-old samples, but did not
fare as well after 16 and 32 weeks. Group 2 was also much weaker in the pullout test for the
32-week-old specimens. Chemically treated bones showed no variation between experimental and
control lots during testing after 1 week. However, the compressive and screw pullout tests showed
that the strength of the specimens decreased considerably after 16 and 32 weeks. Partially decal-
cied, chemically sterilized samples were considerably (48 to 82%) weaker than the control
specimens for all cases.
Roe et al.

6

concluded that the mechanical properties of the bone are considerably affected by
the techniques used for storage. Among the different mechanisms, the clearest is the effect of partial
decalcication on the weakened mechanical properties of the bone. Other potential reasons for the
changes of mechanical properties include loss of water vapor, dehydration, and enzyme activity
during storage.

D. E

FFECTS



OF BONE DENSITY ON SCREW PULLOUT STRENGTH
The bones of middle-aged adults are much stronger than the soft bones of children and weak bones
of the elderly. While the youthful bones will eventually grow strong, old bones are more fragile
and often become susceptible to the effects of osteoporosis. Numerous studies have been performed
to show the effects of osteoporosis on the strength of the bone. Osteoporosis, experienced most
often by postmenopausal females, occurs when a loss of estrogen results in an imbalance between
osteoblast and osteoclast production and function. The equilibrium is lost as osteoclast activity
(bone resorption) overruns osteoblast production and the bone mineral density (BMD) and mechan-
ical strength are signicantly reduced.
7,8
How does one secure a screw in weak bone? The effects
of osteoporotic bone on the screw pullout test are quite signicant and must also be examined.
Halvorson et al.
9
studied the effects of bone mineral density and the techniques of screw
insertion on the screw-holding power. Ten experimental groups were used. The rst six lots were
classied as untapped, 5.5 mm tap, and 6.5 mm tap for both normal and osteoporotic bone samples,
respectively. The other four groups had the following characteristics: stripped and packed with
bone chips (normal), osteoporotic, one screw/one hook, and one screw/two hooks.
The results of Halvorson et al.
9
clearly showed the effects of osteoporosis on bone mineral
density. The average bone mineral density in normal spines (1.17 0.08 g/cm
2
) was signicantly
higher than that for the osteoporotic spines (0.08 0.05 g/cm
2
). The difference between healthy
bone and osteoporotic bone is even more apparent in the average pullout force withstood by each
specimen before fracture. The normal bones fared much better with an average pullout force of
1540 361 N. The osteoporotic samples, on the other hand, withstood only 206 159 N of applied
force. Halvorson et al. also showed that screws inserted in 5.5 or 6.5 mm tapped holes were
signicantly stronger than untapped samples.
From this study the substantial impact that bone mineral density has on screw pullout strength
is apparent. The bone specimens from normal spines had a pullout strength almost 7.5 times greater
than the osteoporotic specimens. Because the performance of the osteoporotic samples was signif-
icantly worse than that of the normal specimens, special screws may have to be designed and
employed at the boneimplant interface for victims of osteoporosis. Clinically, the bone micro-
structure and density of all implant recipients should therefore be inspected carefully before a screw
design is chosen.
0266/ ch21/frame Page 327 Friday, October 22, 1999 9:24 AM
328 Mechanical Testing of Bone and the BoneImplant Interface
III. SUMMARY
This chapter briey introduces several applications of testing bone mechanical strength using the
screw pullout test. Other factors that inuence the magnitude of the pullout force a bone specimen
can experience include the application of a resin around the screw, the size of the sample and the
method of xation, as well as the dimensions of the screw (such as thread prole) and the depth
of its penetration. Although these variables have signicant effects on the screw pullout test, they
are not directly related to the impact of the bone on the pullout test. These factors are discussed
in Chapter 36 of this book.
REFERENCES
1. An, Y.H., Kang, Q., Friedman, R.J., and Young, F.A., Comparison of screw pullout strength from
weight-bearing and non-weight-bearing directions in the cancellous bone of bovine tibial plateau, in
The 20th Annu. Meet. Am. Soc. Biomech., Atlanta, Ga, 1996, 181.
2. An, Y.H., Kang, Q., Friedman, R.J., and Young, F.A., The effect of microstructure of cancellous bone
on screw pullout strength, Trans. Soc. Biomater., 20, 385, 1997.
3. An, Y.H., Young, F.A., Kang, Q., and Williams, K.R., Effects of microstructure of cancellous bone
on screw pullout strength, Unpublished data, 1999.
4. Horton, W.C., Blackstock, S.F., Norman, J.T., et al., Strength of xation of anterior vertebral body
screws, Spine, 21, 439, 1996.
5. Nicholson, P.H., Cheng, X.G., Lowet, G., et al., Structural and material mechanical properties of
human vertebral cancellous bone, Med. Eng. Phys., 19, 729, 1997.
6. Roe, S.C., Pijanowski, G.J., and Johnson, A.L., Biomechanical properties of canine cortical bone
allografts: effects of preparation and storage, Am. J. Vet. Res., 49, 873, 1988.
7. Dickenson, R.P., Hutton, W.C., and Stott, J.R., The mechanical properties of bone in osteoporosis,
J. Bone Joint Surg., 63B, 233, 1981.
8. Van Audekercke, R. and Van der Perre, G., The effect of osteoporosis on the mechanical properties
of bone structures, Clin. Rheumatol., 13 (Suppl. 1), 38, 1994.
9. Halvorson, T.L., Kelley, L.A., Thomas, K.A., et al., Effects of bone mineral density on pedicle screw
xation, Spine, 19, 2415, 1994.
0266/ ch21/frame Page 328 Friday, October 22, 1999 9:24 AM

329

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

22

Viscoelastic Properties
of Bone and Testing Methods

Naoki Sasaki

CONTENTS

I. Introduction ..........................................................................................................................329
II. Viscoelasticity and Viscoelastic Analysis ............................................................................330
A. General Remarks ............................................................................................................330
B. Viscoelastic Analysis......................................................................................................331
III. Viscoelastic Properties of Bone...........................................................................................333
A. Creep Properties .............................................................................................................333
B. Stress Relaxation Properties...........................................................................................334
C. Determination of the Relaxation Mechanism on the Basis of the
Empirical Relaxation Modulus ......................................................................................336
1. KWW-Type Relaxation .............................................................................................336
2. The Debye-Type Relaxation Process (Process II) ....................................................337
IV. Method of Testing ................................................................................................................339
A. Method and Apparatus ...................................................................................................339
1. Torsion Pendulum Used by Lakes et al....................................................................339
2. Compression Tester Used by Lugassy and Korostoff ..............................................341
B. Specimen Preparation.....................................................................................................341
C. Apparatus........................................................................................................................342
1. Torsion Tester ............................................................................................................342
2. Cantilever Bending System.......................................................................................343
3. Three-Point Bending System....................................................................................344
4. Data Acquisition........................................................................................................345
5. Fast Measurements....................................................................................................345
V. Summary and Future Directions of Research .....................................................................347
References ......................................................................................................................................347

I. INTRODUCTION

According to Ferry, viscoelasticity of a material is dened as follows.

1

The theory of elasticity deals
with mechanical properties of elastic solids, for which, in accordance with Hookes law, stress is always
directly proportional to strain in small deformations but independent of the strain rate. The theory of
hydrodynamics deals with properties of viscous liquids, for which, in accordance with Newtons law,
stress is always directly proportional to the strain rate but independent of the strain itself. These
categories are idealizations; however, it is true that the behavior of many solids approaches Hookes
law for innitesimal strains and the behavior of many liquids approaches Newtons law for innitesimal
rates of strain. Deviations from ideal behavior are observed under other conditions. Even if both strain
and rate of strain are innitesimal, a system may exhibit behavior that combines liquid-like and solid-
like characteristics. Materials whose behavior exhibits such characteristics are called viscoelastic.

0266/ ch22/frame Page 329 Friday, October 22, 1999 9:31 AM

330

Mechanical Testing of Bone and the BoneImplant Interface

In many of the materials of interest in physics, as well as of practical importance in engineering,
viscoelastic anomalies are negligible or of only minor signicance. Although the foundations of
phenomenological theory of linear viscoelasticity have provided valuable information about, for
example, the structure of metals, the deviations from perfect elasticity here are small. In polymers,
on the other hand, mechanical behavior is dominated by viscoelastic phenomena which are often
truly spectacular. Each exible threadlike polymer molecule has a number of internal degrees of
freedom compared with materials containing molecules of low molecular weight and is continually
changing the shape of its contour as it wriggles and writhes with its thermal energy. Rearrangements
on a local scale are relatively rapid, but they are very slow on a long-range scale. Under stress, a
new assortment of congurations is obtained; the response to the local aspects of the new distribution
is rapid, the response to the long-range aspects is slow, and, all told, there is a very wide and
continuous range of time covering the response of such a system to external stress. This distribution
of response times in a system brings about the viscoelasticity of polymers.

1

Viscoelasticity has been regarded as a characteristic feature of polymeric materials. Almost all
biological materials are made of polymers. This is the reason many biological materials manifest
viscoelasticity. The major constituents of bone are stiff hydroxyapatite-like mineral and pliant
collagen. Because of the viscoelastic nature of collagen bers in the bone matrix, bone itself has
remarkable viscoelasticity.
However, detailed experimental works on the viscoelasticity of bone have only been carried
out recently, despite the fact that it has been known for a long time that bone has viscoelasticity.

26

A considerable amount of work has been done on some of the mechanical properties of bone,

7

especially its strength and its modulus of elasticity. Most of the strength tests have been static, but
some work has been done on the dynamic and fatigue strength of bone. Thus, at the present time,
there is still little known about the viscoelastic properties of bone.
There seems to be at least three reasons viscoelastic properties of bone have not been an
important subject for many investigators. The rst reason is that an elastic approximation for bone
had been working well for actual bone statics and a deviation of the mechanical properties of bone
from elasticity was regarded as a scatter of data generally observed in biological systems. The
second reason is that the biological importance of the viscoelasticity has not been recognized. It
is difcult to nd a situation in the lives of animals that is similar to even simple viscoelastic
experiments, such as creep and stress relaxation. The third reason is the dearth of information
available on how constituents of bone are mechanically combined. After knowledge about the
histological structure of bone accumulated, the contribution of a collagen matrix to the mechanical
properties of bone was taken into consideration. Lees

8

proposed that the mechanical properties of
a collagen matrix change with mineral content. Sasaki and Yoshikawa,

6

furthermore, showed that
the viscoelastic properties of bone also change as a function of mineral content.
In a viscoelastic material, some of the elastic energy generated by an external force applied to
the material is dissipated as heat. All living things make use of this dissipating mechanism of energy
for survival. In bone, dissipated energy may contribute to the driving force in the remodeling process
of bone.

9

By elucidating the mechanism by which viscoelasticity of bone is generated, it may be
possible to understand the physicochemical origin of Wolffs law.

10

In this chapter, only the
viscoelastic properties of cortical bone are discussed; the mechanical properties of trabecular bone
are discussed in another chapter.

II. VISCOELASTICITY AND VISCOELASTIC ANALYSIS
A. G

ENERAL

R

EMARKS

Viscoelasticity as reviewed here is limited to static experiments (creep properties and stress relax-
ation of bone), although the dynamic mechanical properties of bone are also an important issue.

0266/ ch22/frame Page 330 Friday, October 22, 1999 9:31 AM

Viscoelastic Properties of Bone and Testing Methods

331

The ultrasonic method (a high-frequency limit of dynamic mechanical measurement) is reviewed
in another chapter.
The objectives of the phenomenological viscoelasticity investigation are (1) to understand
phenomenological relations between excitation and the response of the material to it in order to
nd a constitutive equation of the material and (2) to visualize molecular events occurring in the
material during the process. For description of the constitutive equation, a relaxation function and
a creep function must be determined (Figure 22.1). Both functions are dened as follows. When a
step strain

was applied to the material, the stress response

(

t

) of the material was expressed as

(

t

) =

E

0

[1

(

t

)]

0

=

E

(

t

)

0

(22.1)
where

E

0

is the glass modulus,

(

t

) is the relaxation function and

E

(

t

) is the relaxation modulus.
When a step stress

0

was applied to the material the strain response

(

t

) was described as

(

t

) =

J

0

[1 +

(

t

)]

0

=

J

(

t

)

0

(22.2)
where

J

0

is the glass compliance,

(

t

) is the creep function, and

J

(

t

) is the creep compliance. In
order to determine if these functions can be used for accurate estimation of the viscoelasticity of
a material, stress relaxation and creep experiments have been performed by many investigators.

B. V

ISCOELASTIC

A

NALYSIS

Analysis based on the theory of linear viscoelasticity was used to determine the relationships
between material functions,

(

t

),

E

(

t

),

(

t

), and

J

(

t

), and molecular events occurring in bone from
data obtained from viscoelastic experiments. The elementary process of stress relaxation is ideal-
istically described by the so-called Maxwell element, and that of the creep process is described by
the Voigt element (Figure 22.2). Debye used an exponential-type decay function to describe the
time course of relaxation with a single relaxation time, as in the case of the Maxwell element.
Relaxation described by an exponential decay function is called a Debye-type relaxation. Material
functions for both elements are related to a single exponential type function,

e

t

/

, where

is a
characteristic time of the process.
In the actual case of a viscoelastic material, no single element can describe the actual relaxation
or creep process. Relaxation that cannot be described by a simple exponential decay is referred to

FIGURE 22.1

Stress relaxation (A)



and creep measurements (B).

0266/ ch22/frame Page 331 Friday, October 22, 1999 9:31 AM

332

Mechanical Testing of Bone and the BoneImplant Interface

as non-Debye relaxation. There is no established method for determining the decay function for
non-Debye relaxation. Combinations of Maxwell elements or Voigt elements have usually been
used to describe the non-Debye-type viscoelastic properties of a material. For a more complicated
case, the conventional approach to non-Debye relaxation is to consider a distribution of Debye-
type relaxing elements, each with its own relaxation time.
Then the relaxation modulus

E

(

t

) is expressed by the relaxation spectrum

H

(ln

)

d

ln

as

1

(22.3)
One course of mechanical analysis of a material is to obtain the relaxation spectrum from stress
relaxation data.
A relaxation spectrum, however, does not contain more information than the original data, as
it is simply a mathematical transform of the original, invariably non-Debye, data. It is still difcult
to deduce molecular events occurring in relaxation. A number of empirical functions for non-Debye
relaxation processes have been proposed and have been applied to specic samples of materials.
Recently, the universality of non-Debye-type relaxation phenomena in condensed matter has
been demonstrated. This universality means that almost all of the non-Debye relaxation processes
can be described by a few simple relaxation functions. Jonscher

11

and his colleagues demonstrated
that relaxation of a variety of materials can be described by a combination of the power-law relation
of time or frequency.

11,12

Williams and Watts

13

empirically obtained a response function which is
called a stretched exponential. It was shown that this function is identical to the Kohlrausch
fractional exponential function, which was rst applied to estimate the stress relaxation and plastic
deformation of glass bers by Frederick Kohlrausch in 1863:

14,15

E(t)

exp[

(t/

], (0 <

1) (22.4)

FIGURE 22.2

The Maxwell element and the elementary process of stress relaxation (A) and



the Voigt
element and the elementary creep process



(B).
E t ( ) H ln ( ) t ( ) exp ln d

+

=

0266/ ch22/frame Page 332 Friday, October 22, 1999 9:31 AM

Viscoelastic Properties of Bone and Testing Methods

333

where

is a characteristic time of the relaxation and

is a parameter describing the shape of the
time dependence of the relaxation modulus. Recently, this function has been used to describe the
relaxation phenomena of glass-forming amorphous polymers.

16

Several models have been proposed
for explaining the relaxation phenomenon that have a relaxation function of the same type as that
of Equation 22.4.

17-19

III. VISCOELASTIC PROPERTIES OF BONE
A. C

REEP

P

ROPERTIES

In Figure 22.3, the results of creep experiments conducted by Currey

2

on cortical bone from bovine
tibia and metacarpal bone have been reproduced. A cantilever bending test method was employed.
According to Currey, the creep process can be divided into two subprocesses: the creep response
that occurs almost immediately after stress application, and the creep response that occurs after
the rst process and continues for more than 10 days. The former is dominated by elastic defor-
mation. Currey concluded that the second creep process is anelastic, not plastic. Anelasticity is a
recoverable strain that appears over a certain period of time. Lakes and Saha

20

also performed creep
experiments in conjunction with observations using an optical microscope. They also observed the
same two creep processes. For the process that occurred after several minutes and continued for
more than 10 days, they found simultaneous cement-line slippage in the micrograph. They con-
cluded that creep that occurred over a long period of time is attributed to large-scale structural
change in bone, such as cement-line slippage. A creep process that originates from a structural
change in the material usually causes an unrecoverable strain. The results by Lakes and Saha
contrasted with those of Currey on this point. Lakes and Saha also observed creep fracture phe-
nomena at comparatively large stress values for the creep.

21

The difference between the results of
Currey and those of Lakes and Saha may have originated from the stress values applied. A creep
function was not determined in either study, but a review of their experimental data showed that
creep functions in both cases could not be described by a simple exponential function. Rimnac
et al.

22

empirically determined the strain rate of a steady creep process that appeared after an
immediate deformation process of bovine femoral cortical bone as a function of temperature, applied
stress, and the volume fraction of Haversian bone in the specimen. In a steady creep process,
viscosity is dominant and the strain can be described as

(

t

) =

t

, where

is a constant.

FIGURE 22.3

Creep process in bovine cortical bone. Data points were replotted from Currey.

2

0266/ ch22/frame Page 333 Friday, October 22, 1999 9:31 AM

334

Mechanical Testing of Bone and the BoneImplant Interface

B. S

TRESS

R

ELAXATION PROPERTIES
Figure 22.4 shows the results of stress relaxation experiments conducted by Lugassy and Korostoff
23
on bovine femoral cortical bone. With the inuence of specimen orientation, the effects of loading
speed and amount of deformation were examined. The general features of stress relaxation of
cortical bone are rapid relaxation in the early stage of relaxation and then gradual relaxation over
a long period of time. These correspond well to the characteristic feature of creep in cortical bone.
In the case of relaxation over a long period of time, relaxation modulus can be described by a
simple exponential decay function (Debye-type relaxation). According to Tobolskys method,
24
which is usually applied to polymer systems, the relaxation process can be divided into two parts:
an initial rapid relaxation process lasting for up to 100 min followed by a process of gradual
relaxation. The relaxation time for the latter process was determined to be in the order of 10
6
s.
For the former process, description of the relaxation by a simple exponential decay was found to
be impossible. Instead, Lugassy and Korostoff found that the early stage of the relaxation process
in bone could be described by a linear relation between the relaxation modulus E(t) and the logarithm
of time:
E(t) = a b log t (22.5)
where a and b are constants that are independent of time. This equation was derived in a procedure
for obtaining a relaxation time spectrum using Alfreys approximation method. The linear relation
indicates that this part of stress relaxation in bone is reminiscent of the wedge-type region of the
relaxation spectrum in amorphous polymeric materials, although obtained by approximation.
25,26
Lakes and Katz
4
presented eight decades of relaxation spectra for human tibial cortical bone
and bovine femoral cortical bone from both stress relaxation and dynamic mechanical measure-
ments. Figure 22.5 shows the relaxation spectra for human bone and bovine bone. The spectra were
estimated by using Alfreys approximation. In accordance with the results by Lugassy and Korost-
off,
23
it was shown that relaxation can be described by at least two relaxation processes. For human
bone, the relaxation process having a single relaxation time of about 5 10
4
sec was most
remarkable, and there were several small processes on the short time-side foot of it. Lakes and
Katz estimated the contribution of thermoelasticity in osteons and lamellae toward tan of the
FIGURE 22.4 Stress relaxation in bovine femoral cortical bone. Data points were replotted from Lugassy
and Korostoff.
23
0266/ ch22/frame Page 334 Friday, October 22, 1999 9:31 AM
Viscoelastic Properties of Bone and Testing Methods 335
system, and they assigned the small peaks in the spectrum as contributions from the inhomogeneous
structure in bone.
4
For bovine bone, there were no remarkable relaxation processes over a long
duration as there are in human bone. Rather, the spectra increased monotonously with relaxation
time. This difference, however, may be due to the fact that the stress relaxation measurement of
bovine bone was performed only up to 5 10
4
s, and a remarkable peak may appear at a time
more than 10
5
s, as was suggested by the results of Lugassy and Korostoff.
23
As described above, the general features of stress relaxation in cortical bone are rapid, non-
Debye relaxation in the early stage (process I) and gradual relaxation of the Debye type over a
long period of time (process II). The total relaxation shear modulus, G(t), can be described by an
empirical equation such as
G(t)/G
0
= A
1
f(t/
1
) + A
2
exp (t/
2
) (22.6)
FIGURE 22.5 Normalized relaxation spectra, H
r
(t), with respect to initial relaxation modulus values of
(A) human tibial and (B) bovine femoral cortical bone. (Adapted from Lakes and Katz.
8
)
0266/ ch22/frame Page 335 Friday, October 22, 1999 9:31 AM
336 Mechanical Testing of Bone and the BoneImplant Interface
where G
0
is the glass modulus of bone, A
1
and A
2
are the non-Debye and Debye fractions, respectively,
against the relaxation as a whole,
2
is the relaxation time for process II, and f(t/
1
) is the non-Debye
function to be determined, where
1
is the relaxation time for process I. Sasaki et al.
27
proposed that
process I can be described by the KohlrauschWilliamsWatts (KWW) function:
f(t/
1
) = exp [(t/
1
)

] (0 < 1). (22.7)

Then, for all the relaxation process, the following empirical equation was presented:
G(t)/G
0
= A
1
exp [(t/
1
)

] + A
2
exp (t/
2
) (A
1
+ A
2
= 1, 0 < 1). (22.8)
Figure 22.6 shows a comparison of the experimental data points with estimated values obtained by
using Equation 22.8. The results show that the equation can be used to describe accurately the
relaxation process as a whole for bone.
In Figure 22.7, the relaxation curve (logarithm of averaged relaxation modulus vs. time) by
Lugassy and Korostoff for bovine femur is replotted.
23,27
The data points were normalized by
reference to the initial stress. In this gure, the solid curve represents values obtained from
Equation 22.8 for compressive stress relaxation. The data points corresponded well to the values
obtained from the equation. The parameters are listed in the gure. A combination of the KWW
function and the Debye function was recently proposed to describe the relaxation phenomena in
amorphous polymers.
28
C. DETERMINATION OF THE RELAXATION MECHANISM ON THE BASIS
OF THE EMPIRICAL RELAXATION MODULUS
1. KWW-Type Relaxation
The mechanism of relaxation in bone was investigated on the basis of the empirically determined
relaxation modulus. The mechanical properties of collagen bril have been shown to be greatly
affected by the water content,
29-31
(in g/g bone), while those of hydroxyapatite are thought to be
unaffected. These different responses in mechanical properties to a change in water content com-
petitively determine the mechanical properties of wet bone. Sasaki and Enyo
32
carried out torsional
FIGURE 22.6 Stress relaxation of bovine femoral cortical bone. Data points were replotted from Sasaki
et al.
27
The solid line represents values obtained from Equation 22.8.
0266/ ch22/frame Page 336 Friday, October 22, 1999 9:31 AM
Viscoelastic Properties of Bone and Testing Methods 337
stress relaxation measurements in bone samples of various water contents. All of the obtained
relaxation curves could be described by Equation 22.8. Figure 22.8 shows the water content depen-
dence of G(t) at indicated times. In the early stage, up to 10
3
s after strain application, G(t) decreased
with , while in the later stage, at 10
4
to approximately 10
5
s, curves of G(t) vs. almost leveled
off. Thus, the collagen matrix is believed to be responsible for the relaxation of bone in the early
stage. However, in the later stage G(t) values are very similar, indicating that the modulus, in this
stage of hydration, is not so sensitive to . This leads to a preliminary conclusion that KWW
relaxation is closely related to the mechanical properties of the collagen matrix.
Figure 22.8B shows the creep process of collagen ber in saline, as was observed by a time-
resolved X-ray diffraction method where strain was determined for the HodgePetruska
D-period.
33,34
Creeplike change in D-period strain,
D
, corresponds well with the KWW-type stress
relaxation in collagen ber soaked in saline, as shown in Figure 22.8C.
35
The change in
D
was
found to be related to molecular rearrangement in the ber caused by externally applied force.
36,37
Therefore, the KWW-type stress relaxation in bone is also thought to originate from a molecular
rearrangement in collagen ber in a bone matrix.
2. The Debye-Type Relaxation Process (Process II)
Because of the similarity in the time course, process II is related to the nonelastic deforming process
in the creep results for bone.
20
Such a nonelastic deformation of inhomogeneous material in the
solid state by an external force has been discussed generally on the basis of the concept relating
to the crack nucleation and propagation. Here, process II is discussed on the basis of the microcrack
nucleation concept at a stress-concentrated area, such as the tip of an osteosyte lacuna in bone. As
a consequence of stress concentration around the tip of the crack, nucleation of small cracks near
the tip occurs. The nucleation reduces the concentrated stress, and this can be the elementary
process of stress relaxation in bone.
In general, the nucleation rate, K
n
, is written as
38
K
n
= A exp[(U )/kT] (22.9)
where A is a constant, U is the activation energy of the relaxation, is the function representing
the contribution of applied stress and then the concentrated local stress
L
to the free energy
FIGURE 22.7 Fitting of values obtained from Equation 22.8 to the data obtained by Lugassy and Korostoff.
23

(Adapted from Sasaki et al.
27
)
0266/ ch22/frame Page 337 Friday, October 22, 1999 9:31 AM
338 Mechanical Testing of Bone and the BoneImplant Interface
FIGURE 22.8 (A) Water content dependence of relaxation shear modulus at indicated time after the strain
application. (Adapted from Sasaki and Enyo
32
) (B) Creep in the D-period of tendon collagen. Data points
were replotted from Sasaki et al.
34
(C) Stress relaxation of tendon collagen. Data points were replotted from
Azuma and Hasegawa.
35
The solid curve represents values obtained from Equation 22.8 tted to the data.
34
0266/ ch22/frame Page 338 Friday, October 22, 1999 9:31 AM
Viscoelastic Properties of Bone and Testing Methods 339
of the system, and k is the Boltzmann constant. On the basis of the nucleation concept, is
described as
39,40
~ ln
L
. (22.10)
By using this relation, the nucleation rate is rewritten as
K
n
= A exp(U/kT)
L

(22.11)
where A and are constants, which are determined according to the material. As
L
is related to the
macroscopic stress via the stress concentration coefcient,
L
can be estimated from the relaxation
stress. When 1, the stress relaxation rate determined by the nucleation process is written as
K
n
= (d/dt) (22.12)
and the relaxation process then becomes the Debye type. In this case, the relaxation was conrmed
empirically to be the Debye type, and should therefore be almost unity.
In order to establish the characterization of (
L
) relating to water content, the initial value
of the shear modulus G
0
RD
( = A
2
G
i
, G
i
; initial relaxation shear modulus) for the Debye-type
relaxation in bone was plotted in Figure 22.9A as a function of water content. The values have
been normalized, and the modulus value for the sample with the smallest water content should
therefore be unity. It is clear that G
0
RD
reduces monotonously and almost linearly with . The shear
modulus value for collagen lm has also been reported to depend almost linearly on in the range
of 0.05 < < 0.2.
29
Assuming that
L
G
0
RD
() = ( and being constants that are independent
of and , > 0) as an empirical relation and utilizing the relation in Equation 22.11, the relaxation
rate of the Debye process K
D
( K
n
) can be described as
K
D
= (a b)

(22.13)
where a and b are constants that are independent of . Empirically, 1, then
K
D
~ a b (22.14)
This expectation can be compared with the empirical results of the authors experiments.
The relaxation time,
2
of the Debye process is equal to K
D
1
. In Figure 22.9B, the relaxation rate
of the Debye relaxation, K
D
(= 1/
2
), is plotted against . K
D
decreases linearly with . The results
shown in this gure conrm the validity of the relation in Equation 22.14. This indicates that
the effect of on
2
can be explained by the formulation of a crack nucleation process with the
help of empirical data.
IV. METHOD OF TESTING
A. METHOD AND APPARATUS
In order to investigate the viscoelastic properties of bone, a viscoelasticity testing machine is not
needed. For example, an Instron testing machine can be used to detect the anelasticity of bone.
23
Here, the apparatus used in past studies is described.
1. Torsion Pendulum Used by Lakes et al.
3
Figure 22.10 shows a schematic representation of the biaxial torsion pendulum developed by
Lakes et al.
3
for examining the viscoelastic nature of cortical bone. A specimen, shaped into a
0266/ ch22/frame Page 339 Friday, October 22, 1999 9:31 AM
340 Mechanical Testing of Bone and the BoneImplant Interface
waisted cylinder, is xed between clamps and exposed to a slight extending force along the
specimen and torsion axes. The torsional strain is generated by a torsion motor consisting of a
DC magnet assembly and a driving coil rotor. One end of the rotor shaft is connected to the
upper clamp, and the other end is connected to a torsional strain sensor using a linear variable
FIGURE 22.9 (A) G
0
RD
plotted against water content, . Data were adapted from Sasaki and Enyo.
32

(B) Relaxation rate for process II, plotted against . Data points were replotted from Sasaki and Enyo.
32
0266/ ch22/frame Page 340 Friday, October 22, 1999 9:31 AM
Viscoelastic Properties of Bone and Testing Methods 341
differential transformer (LVDT). Torsional force is detected by a torque sensor mounted between
the frame and the lower clamp of the specimen. The rise time of a step function angular
displacement is electronically controlled to 60 to 100 ms for zero to nal value of angle. By
operating the driving coil with the output from an oscillator, dynamic mechanical measurements
up to a maximum of 50 Hz are possible. Some results obtained by using this apparatus were
presented in a previous section.
2. Compression Tester Used by Lugassy and Korostoff
Strain application and force detection were performed by the use of an Instron testing machine.
23
The strain value was signalized by an LVDT, which was attached to an upper ram. A specimen
holder for the relaxation experiments of bone is schematically shown in Figure 22.11. A cylinder-
shaped specimen was prepared for a compressive stress relaxation experiment. The specimen
was placed between the upper and lower rams. The lower ram was connected to a load cell. The
specimen was bathed in Ringers solution. Some results obtained by using this apparatus were
presented in a previous section.
B. SPECIMEN PREPARATION
Cortical bone specimens for viscoelastic measurements are prepared as follows. From the middi-
aphysis of a long bone, rectangular plates of cortical bone are cut using a diamond saw under tap
water. (The specimen must be shaped as a rectangular plate to t the apparatus.) The plate is then
shaped to a suitable size by using emery paper, and nal polishing is done using apatite paste. The
surface of the specimen is then observed under an optical microscope to conrm that there are no
scratches. The suitable specimen size is discussed in the apparatus section. The actual size of the
specimen is determined by observation using a traveling microscope.
FIGURE 22.10 Schema of the torsion pendulum used by Lakes et al.
3
0266/ ch22/frame Page 341 Friday, October 22, 1999 9:31 AM
342 Mechanical Testing of Bone and the BoneImplant Interface
C. APPARATUS
1. Torsion Tester
The relaxation shear modulus is usually determined by a torsion tester. Figures 22.12A and B show,
respectively, a cross-sectional view and the top view of a torsion tester of the type invented by
Kusano and Murakami
41
and modied for experiments on bone.
27,32
The bone specimen used in
this apparatus is a plate with dimensions of 8 40 0.5 mm. The torsion axis is parallel to the
longer edges of the specimen plate. The suitable size of the specimen for this system is estimated
according to the formula:
ab
3
[1 (0.63025/u)] = (3LF/G), u = a/b (22.15)
FIGURE 22.11 Schema of the specimen holder used by Lugassy and Korostoff.
23
FIGURE 22.12 Schema of the torsion tester: (A) a cross-sectional view along the torsion axis and
(B) a top view.
0266/ ch22/frame Page 342 Friday, October 22, 1999 9:31 AM
Viscoelastic Properties of Bone and Testing Methods 343
and by using the average bone modulus value G in the literature and the resolution of the
sensitivity of the force sensor, where a and b are the width and thickness, respectively; and L,
F, and are the length of the specimen, the recovering force, and torsional deformation in radians,
respectively. The bone specimen, S, is set between an upper clamp, UC, and a lower clamp, LC,
in a temperature- and humidity-controlled chamber, H. The temperature is controlled by an
electric heater and circulation of tap water around the chamber. For the regulation of humidity,
the ambient atmosphere of the specimen was exposed to saturated salt solution in the chamber.
The lower clamp is xed to the frame, and the upper clamp is connected to the force translational
rod, R. A rigid arm, A, is attached to the top of the rod, and a force transducer, TD, is connected
to the end of the arm. The transducer is mounted on a turntable, T. When T is rotated around
the sample axis, A, R, and UC also rotate around the same axis. To achieve smooth rotation, two
ball bearings (B1; NSK 6000; B2, NSK 5208) and a needle bearing (B3, 30203) are used, as
shown in the gure. These bearings also enable the transducer TD to detect the recovering force
generated in the bone specimen against the shearing deformation applied by the rotation of T.
Torsional loading is applied within a few seconds by rotating ( = 10 on average) the turntable,
T, on which the force sensor, TD, is mounted. Shear strain is xed by a stopper screw, ST.
According to Kusano and Murakami shear-modulus values smaller than 10
2
GPa were obtained.
41
In the present form, shear strain is applied manually. However, by using an automatic turntable
with a stepping motor, all the measurements could be automatically performed.
2. Cantilever Bending System
One of the most commonly used methods for measuring the relaxation Youngs modulus of bone
is cantilever bending. Figure 22.13 shows a top view of the cantilever bending system used for
determining the relaxation Youngs modulus function of cortical bone.
6,27
This system is also set
in a temperature- and humidity-regulated chamber. The bone specimen used in this system is also
a plate of similar dimensions to those of the plate used in the torsion tester. The suitable size of
the specimen for this system is estimated according to the formula:
ab
3
= 4FL
3
/Ed (22.16)
and by using the average bone modulus value E in the literature and the resolution of the sensitivity
of the force sensor, where a and b are the width and thickness respectively, and L, F, and d are the
length of the specimen, the recovering force, and the bending deformation, respectively. The sample
plate, S, is xed to one end of a clamp, C. A force transducer, TD, is mounted on the sliding table,
T. Bending deformation is applied to the specimen by the probe tip of the force transducer, and
the recovering force is detected while keeping the probe of the transducer perpendicular to the
FIGURE 22.13 Schema of a cantilever bending tester.
0266/ ch22/frame Page 343 Friday, October 22, 1999 9:31 AM
344 Mechanical Testing of Bone and the BoneImplant Interface
specimen plate. Deformation is applied within about 1 s. The deformation thus applied is held
constant during the duration of the test. Care must be taken to ensure that the bones are not strained
into the plastic deformation range. In the present form, the bending deformation is applied manually.
However, by using an automatic pulse stage with a stepping motor, all measurements could be
automatically performed.
3. Three-Point Bending System
A three-point bending system is also used for determining the relaxation Youngs modulus function.
Figure 22.14 shows a side view of the three-point bending tester used in the authors laboratory. A
force transducer, TD [LTS-1LK, Kyowa Electric Works (KEW), Japan] is mounted on a Sigma
Koki (SK) auto pulse z-stage, S. The pulse stage is controlled by an SK-stage controller (Mark-
12) with a terminal box (SK CSG-5151). For the force application to the specimen, the stage is
lowered perpendicular to the plate surface. In order for the force to be applied just at the center of
the fulcrums, the specimen holder can be moved by the XY-stage on which the holder is mounted.
The suitable size of the specimen, the width and thickness, and the interfulcrum distance L of this
system are estimated according to the formula:
(22.17)
FIGURE 22.14 Schema of a three-point bending tester.
ab
FL
Ed
3
3
4
=
0266/ ch22/frame Page 344 Friday, October 22, 1999 9:31 AM
Viscoelastic Properties of Bone and Testing Methods 345
and by using the average bone modulus value E in the literature and the resolution of the
sensitivity of the force sensor, where F and d are the recovering force and the bending defor-
mation, respectively.
4. Data Acquisition
The signal from the sensor used in each apparatus is rst passed to a KEW data logger UCAM
10A and then through the data logger to a personal computer. Figure 22.15 shows a block
diagram of the measuring system. The data acquisition program is operated before applying the
strain to the specimen and a baseline of the recovering force is detected by reading the signal
value. Measurements are made at 1-s intervals for 30 min after the strain application, and then
the sampling interval is increased. The measurements are performed for more than 10
5
s after
strain application.
5. Fast Measurements
The three-point bending system can be used to measure the relaxation modulus in an early stage
of stress relaxation in bone.
42
The results of experiments showed that there is a relaxation
process in the early stage that cannot be described by the empirical relaxation modulus,
Equation 22.8.
The relaxation Youngs modulus was measured by three-point bending of a rectangular sample
plate, as shown in Figure 22.16A. A KEW LTS-1K strain gauge transducer was used as both the
deformation generator and the force sensor. The LTS-1K gauge was set on an automicrostage,
which was operated by a rectangular pulse generated by a function synthesizer (Toa FS-111A) to
guarantee identical deformation for a series of measurements. Bending deformation was detected
by a Yokogawa Electric Works (YEW, Musashino, Japan) 3612 strain transducer with a YEW
3624-05 sensor of the noncontacting type. Over a 50-s time interval, an electric signal from the
LTS-1K was passed on to a KEW DPM-613A strain amplier, then to an Autonics S-210 autodig-
itizer, and nally to a personal computer using a GP-IB. The electric signal was recorded every
FIGURE 22.15 Block diagram of the data acquisition system.
0266/ ch22/frame Page 345 Friday, October 22, 1999 9:31 AM
346 Mechanical Testing of Bone and the BoneImplant Interface
0.05 s by the S210. From 50 s to 5 10
4
s, electric signals generated from LTS-1K were directly
recorded by a KEW data logger (UCAM 10A) and nally passed on to a microcomputer. A
maximum strain of less than 0.05% was applied within 0.025 s. This value is well within the
frequency range examined by other authors in dynamic studies. Figure 22.16B shows the stress
response from a high-carbon stainless steel plate, which can be regarded as an ideally elastic
material, compared with a bone. From the stress response by the stainless steel plate, it was found
that there was no overshoot in the applied strain.
FIGURE 22.16 Schema of the apparatus used for fast measurement (A) and a typical loadtime curve
showing the stress response of a stainless steel plate (B).
0266/ ch22/frame Page 346 Friday, October 22, 1999 9:31 AM
Viscoelastic Properties of Bone and Testing Methods 347
V. SUMMARY AND FUTURE DIRECTIONS OF RESEARCH
Although the viscoelasticity of bone has long been regarded as a phenomenon of minor signicance,
the authors group found, by using a simple experimental method, that the viscoelasticity of bone is
as remarkable as that of polymeric materials. However, most techniques currently used to investigate
the viscoelastic properties of bone are still phenomenological ones. In order to understand molecular
events occurring in the viscoelastic process, structural measurements during the process should be
performed. On the other hand, mechanical measurements are methods of materials science, but bone
is a living tissue. One of the biggest problems in bone mechanics is the adaptation or remodeling of
bone. Recent investigations have focused on the residual stress in bone as a result of remodeling.
43
Such a residual stress, although very small compared with the forces applied to an animal in its daily
life, will also be relaxed by the viscoelastic nature of bone. The longest relaxation time of bone, ~10
6
s, is roughly comparable with the remodeling cycle in bone. Further investigation of the viscoelasticity
of bone may lead to an understanding of the remodeling process and mechanism of bone tissue.
REFERENCES
1. Ferry, J.D.,Viscoelastic Properties of Polymers, 3rd ed., Wiley & Sons, New York, 1980.
2. Currey, J.D., Anelasticity in bone and echinoderm skeletons, J. Exp. Biol., 43, 279, 1965.
3. Lakes, R.S., Katz, J.L., and Sternstein, S.S., Viscoelastic properties of wet cortical bone. 1. Torsional
and biaxial studies, J. Biomech., 12, 657, 1979.
4. Lakes, R.S. and Katz, J.L., Viscoelastic properties of wet cortical bone. 2. Relaxation mechanisms,
J. Biomech., 12, 679, 1979.
5. Lakes, R.S. and Katz, J.L., Viscoelastic properties of wet cortical bone. 3. A nonlinear constitutive
equation, J. Biomech., 12, 689, 1979.
6. Sasaki, N. and Yoshikawa, M., Stress relaxation in native and EDTA-treated bone as a function of
mineral content, J. Biomech., 26, 77, 1993.
7. Currey, J.D., The Mechanical Adaptations of Bones, Princeton University Press, Princeton, NJ, 1984.
8. Lees, S., Sonic velocity and the ultrastructure of mineralized tissues, in Calcied Tissue, Hukins,
D.W.L., Ed., Macmillan Press, London, 1989.
9. Levenston, M.E. and Carter, D.R., An energy dissipation-based model for damage stimulated bone
adaptation, J. Biomech., 31, 579, 1998.
10. Roesler, H., Some historical remarks on the theory of cancellous bone structure (Wolffs law), in
Mechanical Properties of Bone, Cowin, S.C., Ed., American Society of Mechanical Engineers, 1981.
11. Jonscher, A.K., Dielectric Relaxation in Solids, Chelsea Dielectrics, London, 1983.
12. Hill, R.M. and Dissado, L.A., Relaxation in elastic and viscoelastic materials, J. Mater. Sci., 19, 1576,
1983.
13. Williams, G. and Watts, D.C., Non-symmetrical dielectric relaxation behavior arising from a simple
empirical decay function, Trans. Faraday Soc., 66, 80, 1970.
14. Tschoegl, N.W., The Phenomenological Theory of Linear Viscoelastic Behavior An Introduction,
Springer-Verlag, Heidelberg, 1987.
15. Scher, H., Shlesinger, M.F., and Bendler, J.T., Time-scale invariance in transport and relaxation, Phys.
Today, 44, 26, 1991.
16. Legrand, D.G., Olszewski, W.V., and Bendler, J.T., Anelastic response of bisphenol-A polycarbonate,
J. Polym. Sci. B Polym. Phys. Ed., 25, 1149, 1985.
17. Palmer, R.G., Stein, D.L., Abraham, E., and Anderson, P.W., Models of hierachically constrained
dynamics for glassy relaxation, Phys. Rev. Lett., 53, 958, 1984.
18. Klafter, J. and Shlesinger, M.F., On the relationship among three theories of relaxation in disordered
systems, Proc. Natl. Acad. Sci. U.S.A., 83, 848, 1986.
19. Ngai, K.L., Evidence for universal behavior of condensed matter at low frequencies/long times, in
Non-Debye Relaxation of Condensed Matter, Ramakrishnan, T.V. and Lakshimi, M.R., Eds., World
Science, Singapore, 1987.
0266/ ch22/frame Page 347 Friday, October 22, 1999 9:31 AM
348 Mechanical Testing of Bone and the BoneImplant Interface
20. Lakes, R. and Saha, S., Cement line motion in bone, Science, 204, 501, 1979.
21. Lakes, R. and Saha, S., Long-term torsional creep in compact bone, J. Biomech. Eng., 102, 178, 1980.
22. Rimnac, C.M., Petko, A.A., Santner, T.J., and Wright, T.M., The effect of temperature, stress and
microstructure on the creep of compact bovine bone, J. Biomech., 26, 219, 1993.
23. Lugassy, A.A. and Korostoff, E., Viscoelastic behavior of bovine femoral cortical bone and sperm
whale dentin, in Research in Dental and Medical Materials, Korostoff, E., Ed., Plenum Press, New
York, 1969.
24. Tobolsky, A.V. and Murakami, K., Existence of a sharply dened maximum relaxation time for
monodisperse polystyrene, J. Polym. Sci., 40, 443, 1959.
25. Tobolsky, A.V., Elastoviscous properties of polysiobutylene. VI. Relation between stress relaxation
modulus and dynamic modulus, J. Am. Chem. Soc., 74, 3786, 1952.
26. Tobolsky, A.V., Stress relaxation studies of viscoelastic properties of polymers, J. Appl. Phys., 27,
673, 1956.
27. Sasaki, N., Nakayama, Y., Yoshikawa, M., and Enyo, A., Stress relaxation function of bone and bone
collagen, J. Biomech., 26, 1369, 1993.
28. Wagner, H. and Richert, R., Dielectric relaxation of the electric eld in poly(vinyl acetate): A time
domain study in the range 10
3
10
6
s, Polymer, 38, 255, 1997.
29. Tanioka, A., Jojima, E., Miyasaka, K., and Ishikawa, K., Effect of water on the mechanical properties
of collagen lms, Biopolymers, 11, 1489, 1973.
30. Nomura, S., Hiltner, A., Lando, J.B., and Baer, E., Interaction of water with native collagen, Biopoly-
mers, 16, 231, 1977.
31. Pineri, M.H., Escoubes, M., and Roche, G., Water-collagen interactions: calorimetric and mechanical
experiments, Biopolymers, 17, 2799, 1979.
32. Sasaki, N. and Enyo, A., Viscoelastic properties of bone as a function of water content, J. Biomech.,
28, 809, 1995.
33. Hodge, J.A. and Petruska, J.A., Recent studies with the electron microscope on ordered aggregates
of the tropocollagen macromolecule, in Aspects of Protein Structure, Ramachandran, G.N., Ed.,
Academic Press, New York, 1963.
34. Sasaki, N., Shukunami, N., Matsushima, N., and Izumi, Y., Time-resolved X-ray diffraction from
tendon collagen during creep using synchrotron radiation, J. Biomech., 32, 285, 1999.
35. Azuma, T. and Hasegawa, M., A rheological approach to the architecture of arterial walls, Jpn. J.
Physiol., 21, 27, 1971.
36. Sasaki, N. and Odajima, S., Elongation mechanism of collagen brils and force-strain relations of
tendon at each level of structural hierarchy, J. Biomech., 29, 1131, 1996.
37. Mosler, E., Folkhard, W., Geercken, W., et al., Stress-induced molecular rearrangement in tendon
collagen, J. Mol. Biol., 182, 589, 1985.
38. Yokobori, T., An Interdisciplinary Approach to Fracture and Strength of Solids, Wolters-Noordhoff
Ltd., Groningen, the Netherlands, 1968.
39. Yokobori, T., Failure and fracture of metals as nucleation processes, J. Phys. Soc. Jpn., 7, 44, 1952.
40. Yokobori, T., Creep fracture of copper as nucleation process, J. Phys. Soc. Jpn., 7, 48, 1952.
41. Kusano, T. and Murakami, K., Apparatus for stress-relaxation experiments, Bull. Res. Inst. Non-
Aqueous Soln., 199, 1970.
42. Goto, T., Sasaki, N., and Hikichi, K., Early stage stress-relaxation in compact bone, J. Biomech., 32,
93, 1999.
43. Todoh, M., Tadano, S., Shibano, M., and Ukai, T., Polychromatic X-ray measurements of anisotropic
residual stress in bovine femoral bone, Trans. Jpn. Soc. Mech. Eng. (Ser. A), 65, 406, 1999.
0266/ ch22/frame Page 348 Friday, October 22, 1999 9:31 AM

349

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

23

Observation of Material Failure
Mode Using a SEM with a Built-In
Mechanical Testing Device

Rong-Ming Wang and Yuehuei H. An

CONTENTS

I. Introduction ..........................................................................................................................349
A. Scanning Electron Microscopy ......................................................................................349
B. Low-Vacuum Scanning Electron Microscopy ...............................................................350
C. SEM with an In-Scope Loading System.......................................................................350
II. Mechanical Loading within the SEM Scope.......................................................................351
III. Observation of Material Fracture Mode ..............................................................................352
IV. Potential Applications in Bone-Related Research...............................................................354
References ......................................................................................................................................355

I. INTRODUCTION

This chapter introduces a technology for potential combined testing of bone properties and observation
of failure mode. There are now available low-vacuum scanning electron microscopes (LV SEM, e.g.,
Model: JSM-5600LV, JEOL, Tokyo, Japan) equipped with a sample loading stage and an energy
dispersive X-ray spectrometer (Figure 23.1). These systems allow simultaneous observation of changes
in surface morphology of a specimen under loading; fracture modes of materials under tensile, com-
pression, and possibly bending load; measurement of mechanical parameters of the sample; and
elemental analysis of chemical composition. The systems have great potential for the determination of
mechanical properties and failure modes of biological materials such as bone, cartilage, or ligament.
Since no data on bone, cartilage, or ligament are available from these types of instruments, the principles
of the system and the potential uses are discussed using some data on inorganic materials.

A. S

CANNING

E

LECTRON

M

ICROSCOPY

Scanning electron microscopy (SEM) has had widespread use in biology, physics, chemistry, and
materials science in recent years owing to the rapid developments in its easily operated multifunc-
tions. With the development of computer technology, computer-controlled SEMs have been devel-
oped and applied in many elds.
In SEM, a focused electron beam is continuously scanned, in a rasterlike pattern, over the
inclined surface of a specimen, causing both backscatter of some of the primary electrons, and
ejection of secondary electrons. These secondary electrons are attracted to a positively charged
collector, and the signal is amplied and displayed on a cathode-ray tube in which a beam spot is
scanned in synchronization with the primary beam on the specimen. The magnication of the image
on the screen of the cathode-ray tube can readily be controlled by varying the area of specimen

0266/ ch23/frame Page 349 Friday, October 22, 1999 9:36 AM

350

Mechanical Testing of Bone and the BoneImplant Interface

covered by the raster. The electron-optical arrangement is similar to that of the microprobe analyzer,
and, for convenience in manipulating the specimen, the nal (or objective) lens must have a
longer working distance, of about 20 mm. Three or more lenses are therefore commonly employed
in the illuminating system.

B. L

OW

-V

ACUUM

S

CANNING

E

LECTRON

M

ICROSCOPY

In recent years, the LV SEM has been developed, which allows specimen observation without the
chemical xation, dehydration, drying and coating needed for conventional high-vacuum SEM.
The LV SEM neutralizes the charge of nonconductive specimens by utilizing ions generated by
the interaction between residual gas molecules in the specimen chamber and the electrons used for
observation. For this purpose, the pressure of the specimen chamber can be raised to approximately
270 Pa by using differential pumping. The use of backscattered electrons for imaging allows not
only specimen surfaces but also inner structures to be observed.

1

In the low-vacuum mode, the specimen chamber is kept in a low vacuum (high pressure) to
observe a nonconductive specimen without conductive coating. Since an energy dispersive spec-
trometer and one channel of wavelength spectrometer can be attached to the microscope, the
application of this instrument can be expanded from high-resolution morphological observation in
the high-vacuum mode and nonconductive specimen observation in the low-vacuum mode for
elemental analysis.
The LV SEM mode is used for observing backscattered electron images. This type of image
contains both composition contrast, which shows the distribution of different substances in
the specimen, and contrast caused by specimen topography. The LV SEM allows nonconductive
specimens to be observed in their original state. Since there is no need to lower the accelerating
voltage, the instrument allows elemental analysis with an energy dispersive X-ray spectrometer
(EDS).
The LV SEM also allows the observation of specimens that contain water or are stained
(biological tissues such as bone and cartilage). Even specimens that are difcult to pretreat, such
as aqueous microorganisms, can be observed with an LV SEM. For this purpose, such specimens
are rapidly frozen in liquid nitrogen and then freeze-dried on the standard specimen stage in the
specimen chamber of an LV SEM. It is also possible to coat freeze-dried specimens for observation
in the HV SEM mode, or to preserve the specimens for later observation.

C. SEM

WITH



AN

I

N

-S

COPE

L

OADING

S

YSTEM

The JSM-5600LV instrument in the authors laboratory (Beijing Institute of Aeronautical Materials)
is equipped with a loading stage (or tensile stage) and an EDS (see Figure 23.1). The instrument
allows observation of surface morphology with X, Y, and R (rotation) movement, and observation of

FIGURE 23.1

A JSM-5600LV and EDS combination system.

0266/ ch23/frame Page 350 Friday, October 22, 1999 9:36 AM

Observation of Material Failure Mode Using a SEM with a Built-In Mechanical Testing Device

351

fracture modes of materials under tensile, compression, and possibly bending load. Most operations,
such as auto-gun-alignment, autofocusing, auto-astigmatism-correction, and auto-contrast/brightness-
adjustment can be controlled using a personal computer.

II. MECHANICAL LOADING WITHIN THE SEM SCOPE

The loading stage makes possible the observation of the fracture process of a material. The stage
is used to observe the dynamic change in a specimen by applying tensile or compression load. The
amount of load can be measured in real time with a load cell. Table 23.1 gives the typical
specications of a tensile stage.
Figure 23.2 shows a typical diagram of the specimen holder in a JEOL tensile stage, which
consists of xed and movable ends. The movable side can be controlled by a motor at a constant
velocity, e.g., 0.05 mm/s, or manually controlled. The amount of load is measured in real time with
a load cell. The accuracy of the cell is 1 kg.
Figure 23.3 shows a typical tensile specimen size and attachment to the stage. The width of
the specimen at the center should be less than 5 mm. The size can be modied to ensure the fracture
of the specimen within 200 kg load. The specimen can also be compressed in the stage. Figure 23.4
shows a typical compression specimen setup.

TABLE 23.1
Typical Specications of a JEOL Tensile Stage

Specimen movement range X = 5~+20 mm
Y = 3~+3 mm
T = 5

~+45

Working distance (WD) 20 mm
Specimen size For tension: 5(W)

29(L)

1(T) mm
For compression: 5(W)

20(L)

1(T) mm
Tensile load 200 kg at maximum
Tensile amount 20 mm at maximum
Driving methods Automatic or manual
Driving speed 0.05~1.0 mm/min (4-steps; 50 Hz)
0.06~1.2 mm/min (4-steps; 60 Hz)
Alarm buzzer Active when tension/compression range exceeded

FIGURE 23.2

Diagram of the specimen holder in a JEOL tensile stage.

0266/ ch23/frame Page 351 Friday, October 22, 1999 9:36 AM

352

Mechanical Testing of Bone and the BoneImplant Interface

After the tension or compression specimen has been installed in the stage, the load is applied
and the changes in the specimen observed. Specimen position can be changed with the stage controls
if the viewing eld shifts with application of load. From the observed image, the size of the area
supporting the load can be calculated. Combined with the measured load, the strength and modulus
values can be obtained. The fracture mode of the material can also be obtained from the observed
image and EDS can be used for chemical analysis.

III. OBSERVATION OF MATERIAL FRACTURE MODE

The authors have studied the dynamic fracture of steels containing nonmetallic inclusions using
the SEM-in-scope loading system. Figure 23.5 shows a typical fracture process of 38CrMoAl steel
containing aluminum nitride (AlN) particles. Cracks start at sharp corners of the inclusions per-
pendicular to the tensile force and propagate along the interface between the inclusions and the
matrix.
Xiao and Han

2

have investigated the microstructure change in a directionally solidied Ni

3

Al
base alloy IC6 with 0.12 wt.% Y and 0.1~0.2 wt.% Si addition. Figure 23.6 shows a typical sec-
ondary electron image of the modied alloy. A nodular precipitate with bright edge is seen in the

FIGURE 23.3

Schematic diagram of a tensile specimen on the loading stage.

FIGURE 23.4

Schematic diagram of a compression specimen on the loading stage.

0266/ ch23/frame Page 352 Friday, October 22, 1999 9:36 AM

Observation of Material Failure Mode Using a SEM with a Built-In Mechanical Testing Device

353

interdendritic area. The EDS/spectrum from the nodular phase is shown in Figure 23.7. The chem-
ical composition is summarized in Table 23.2. Combined with transmission electron microscopy
(TEM) analysis, the precipitate is identied as Mo

6

(Ni

0.75

(Si,Al)

0.25

)

7

.

FIGURE 23.5

Fracture process of an AlN particle in the matrix of 38CrMoAl steel. (a) The morphology of
AlN before tensile loading. (b) The crack initially appears at the sharp corner in AlN perpendicular to the
tensile force. (c) During propagation of the crack, the interfaces between the AlN particle and the matrix
separate perpendicular to the direction of the tensile load.

TABLE 23.2
Chemical Composition of the Precipitate
in Figure 23.6

Ni Mo Al Si

Atom percentage (%) 40.2 47.2 2.4 10.2
Weight percentage (%) 32.6 62.5 0.9 4.0

0266/ ch23/frame Page 353 Friday, October 22, 1999 9:36 AM

354

Mechanical Testing of Bone and the BoneImplant Interface

Yamashita and Kameyama

3

studied the microstructures of oil/water (O/W) emulsions using
cryoscanning electron microscopy. The emulsions used in the study were oil-in-water types of
surfactant-cetostearyl alcoholwater systems. A eld emission SEM (FESEM) was used to inves-
tigate the microstructure of the emulsions because the procedures for specimen preparation were
much simpler. A conventional method, the freeze-fracture method, was used to fracture the emulsion
with a cooled knife in the sample stage of the SEM. A spherical oil droplet was revealed on the
uneven fracture surface (Figure 23.8).

IV. POTENTIAL APPLICATIONS IN BONE-RELATED RESEARCH

The technology presented in this chapter has great potential for applications in bone-related
research. Cortical or cancellous bone cylinders or cubes can be compressively loaded and the

FIGURE 23.6

Typical secondary electron (backscattered) image of the directionally solidified Ni

3

Al base alloy
IC6 with 0.12 wt.% Y and 0.1~0.2 wt.% Si addition. (The image was contributed by Xiao, B. and Han, Y.F.)

FIGURE 23.7

Energy dispersive X-ray spectrum of the nodular phase in Figure 23.6. (The graph was
contributed by Xiao, B. and Han, Y.F.)

0266/ ch23/frame Page 354 Friday, October 22, 1999 9:36 AM

Observation of Material Failure Mode Using a SEM with a Built-In Mechanical Testing Device

355

dynamic morphological changes during deformation and failure can be monitored. Bone can also
be cut into dumbbell-shaped specimens for observation of the fracture process under tensile force.
With a self-designed xture, bending tests of bone or tensile tests of boneimplant interface
specimens could also be done.
The loading stage within the scope is similar to the frame of a standard uniaxial materials
testing machine. With appropriate xtures, most tests commonly used for measuring mechanical
properties of bone, tendon and ligament, skin, and fascia could be performed in the SEM. Research-
ers could test the mechanical properties of bone or other biological materials, observe their mor-
phological changes under loading, and possibly collect data on chemical composition of the samples
at the same time.

REFERENCES

1. Product Manual, JSM-5600LV/JSM-5600 Scanning Electron Microscopes, JEOL, Tokyo, Japan, 1998.
2. Xiao, B. and Han, Y.F., Study on precipitates due to the addition of silicon in yttrium modied Ni-
Al-Mo-Cu alloy IC6, unpublished data.
3. Yamashita, M. and Kameyama, K., Cryo-scanning electron microscopy of microstructures of O/W
emulsions, JEOL News, 33E, 28, 1998.

FIGURE 23.8

A cryo-SEM image of an oil/water emulsion fractured in the SEM. The partial sphere in the
center is a frozen oil droplet. (From Yamashita, M. and Kameyama, K.,

JEOL News,

33E, 28, 1998. With
permission.)

0266/ ch23/frame Page 355 Friday, October 22, 1999 9:36 AM

0266/ ch23/frame Page 356 Friday, October 22, 1999 9:36 AM

357

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

24

Ultrasonic Methods
for Evaluating Mechanical
Properties of Bone

Jae-Young Rho

CONTENTS

I. Introduction ..........................................................................................................................357
II. Sample Preparation ..............................................................................................................360
III. Testing Methods ...................................................................................................................361
IV. Data Collection and Calculation..........................................................................................362
V. Results for Measurements in Bone......................................................................................366
Acknowledgment............................................................................................................................369
References ......................................................................................................................................369

I. INTRODUCTION

Ultrasonic techniques have signicant advantages over mechanical testing methods in the determi-
nation of the elastic properties of bone, in that they are able to use smaller, more simply shaped
specimens.

1-6

Table 24.1 summarizes some of the differences between mechanical testing methods
and the ultrasonic testing method.

6

Also, several anisotropic properties can be measured from a
single specimen. Due to inhomogeneity, anisotropy, and size limitations of bone, these advantages
are signicant.
The elastic properties can be deduced from velocity measurements of shear and longitudinal
waves propagating in particular directions in the bone specimens if the density and the elastic
anisotropy of bone are specied. The relations between velocity and elastic properties follow from
the theory of small-amplitude elastic wave propagation in anisotropic solids.

7,8

There are two modes
of wave propagation, characterized by overall specimen geometry. The rst case is referred to as bulk
wave propagation, as overall specimen geometry tends to innity. Since the cross-sectional dimension
is large, the wave does not perceive the solid boundaries. The second case, where overall specimen
geometry tends to 0, is called bar wave propagation.

2

In this case the entire cross section is excited
by the passing wave. The bar wave velocity can be written in terms of Youngs modulus directly as
. (24.1)
The bulk wave velocity is
, (24.2)
v
bar
E

---
v
bulk
k

---

0266/ ch24/frame Page 357 Friday, October 22, 1999 9:42 AM

358

Mechanical Testing of Bone and the BoneImplant Interface

T
A
B
L
E

2
4
.
1
C
o
m
p
a
r
i
s
o
n

b
e
t
w
e
e
n

M
e
c
h
a
n
i
c
a
l

T
e
s
t
i
n
g

a
n
d

U
l
t
r
a
s
o
n
i
c

T
e
s
t
i
n
g

M
e
t
h
o
d
s

T
e
n
s
i
l
e
C
o
m
p
r
e
s
s
i
v
e
B
e
n
d
i
n
g
U
l
t
r
a
s
o
n
i
c

S
p
e
c
i
m
e
n

s
h
a
p
e
D
i
f

c
u
l
t

t
o

m
a
c
h
i
n
e

s
p
e
c
i
a
l
i
z
e
d

s
h
a
p
e
s

f
o
r

m
o
u
n
t
i
n
g
.
R
i
g
h
t

c
y
l
i
n
d
e
r
s

o
r

c
u
b
e
s
;

p
a
r
a
l
l
e
l

f
a
c
e
s

a
r
e

c
r
i
t
i
c
a
l
.
R
e
c
t
a
n
g
u
l
a
r

p
a
r
a
l
l
e
l
e
p
i
p
e
d
;
l
e
n
g
t
h
/
c
r
o
s
s

s
e
c
t
i
o
n

r
a
t
i
o

i
s

c
r
i
t
i
c
a
l
.
C
y
l
i
n
d
e
r
s

o
r

p
a
r
a
l
l
e
l
e
p
i
p
e
d
;
p
a
r
a
l
l
e
l

f
a
c
e
s

n
o
t

n
e
c
e
s
s
a
r
i
l
y

c
r
i
t
i
c
a
l
;
l
e
s
s

c
o
m
p
l
i
c
a
t
e
d

s
h
a
p
e
.
A
n
i
s
o
t
r
o
p
i
c

e
l
a
s
t
i
c

p
r
o
p
e
r
t
i
e
s
T
h
r
e
e

o
r
t
h
o
g
o
n
a
l

s
p
e
c
i
m
e
n
s

f
o
r

t
h
r
e
e

m
o
d
u
l
i
;

s
h
e
a
r

m
o
d
u
l
i

p
o
s
s
i
b
l
e

i
f

c
r
o
s
s

s
e
c
t
i
o
n

i
s

r
o
u
n
d
;

P
o
i
s
s
o
n

s

r
a
t
i
o

p
o
s
s
i
b
l
e

w
i
t
h

b
i
a
x
i
a
l

e
x
t
e
n
s
o
m
e
t
e
r
.
T
h
r
e
e

o
r
t
h
o
g
o
n
a
l

m
o
d
u
l
i

f
r
o
m

c
u
b
e
;

m
a
y

b
e

p
o
s
s
i
b
l
e

t
o

m
e
a
s
u
r
e

P
o
i
s
s
o
n

s

r
a
t
i
o

w
i
t
h

e
x
t
e
n
s
o
m
e
t
e
r
,

b
u
t

a
s

y
e
t

n
o

r
e
p
o
r
t
s

o
f

P
o
i
s
s
o
n

s

r
a
t
i
o

m
e
a
s
u
r
e
d

i
n

t
h
i
s

w
a
y
.
T
h
r
e
e

o
r
t
h
o
g
o
n
a
l

s
p
e
c
i
m
e
n
s

f
o
r

t
h
r
e
e

m
o
d
u
l
i
;

d
e
t
e
r
m
i
n
a
t
i
o
n

o
f

P
o
i
s
s
o
n

s

r
a
t
i
o

i
n

p
u
r
e

b
e
n
d
i
n
g

t
e
s
t
s
;

t
h
i
s

m
e
t
h
o
d

r
e
q
u
i
r
e
s

s
p
e
c
i
m
e
n
s

w
i
t
h

r
e
l
a
t
i
v
e
l
y

l
a
r
g
e

c
r
o
s
s

s
e
c
t
i
o
n
s
.
T
h
r
e
e

m
o
d
u
l
i
,

t
h
r
e
e

s
h
e
a
r

m
o
d
u
l
i
,

s
i
x

P
o
i
s
s
o
n

s

r
a
t
i
o
s

p
o
s
s
i
b
l
e

f
r
o
m

a

c
u
b
e

a
s

s
m
a
l
l

a
s

1
0

m
m

f
o
r

c
a
n
c
e
l
l
o
u
s

b
o
n
e

a
n
d

5

m
m

f
o
r

c
o
r
t
i
c
a
l

b
o
n
e
.
N
o
t
e
s
I
f

i
n
d
u
c
e
d

b
e
n
d
i
n
g

i
s

a
c
c
o
u
n
t
e
d

f
o
r
,

a
n
d

i
f

s
t
r
a
i
n

i
s

m
e
a
s
u
r
e
d

w
i
t
h

a
n

e
x
t
e
n
s
o
m
e
t
e
r
,

t
h
i
s

t
e
c
h
n
i
q
u
e

c
a
n

b
e

a
c
c
u
r
a
t
e
;

t
e
n
s
i
l
e

t
e
s
t
i
n
g

i
s

m
o
s
t

c
o
m
m
o
n

m
e
t
h
o
d

o
f

m
e
a
s
u
r
i
n
g

e
l
a
s
t
i
c
i
t
y

o
f

e
n
g
i
n
e
e
r
i
n
g

m
a
t
e
r
i
a
l
s
.
T
h
i
s

t
e
c
h
n
i
q
u
e

c
a
n

b
e

a
c
c
u
r
a
t
e

i
f

f
a
c
e
s

a
r
e

p
a
r
a
l
l
e
l

a
n
d

i
f

s
t
r
a
i
n

i
s

m
e
a
s
u
r
e
d

w
i
t
h

a
n

e
x
t
e
n
s
o
m
e
t
e
r

i
n
s
t
e
a
d

o
f

p
l
a
t
e
n

m
o
t
i
o
n
.
C
o
m
p
r
e
s
s
i
v
e

t
e
s
t
i
n
g

i
s

l
e
s
s

c
o
m
m
o
n

f
o
r

e
n
g
i
n
e
e
r
i
n
g

m
a
t
e
r
i
a
l
s
,

b
u
t

A
S
T
M

s
t
a
n
d
a
r
d
s

h
a
v
e

b
e
e
n

w
r
i
t
t
e
n

a
n
d

a
r
e

u
s
e
d

f
o
r

r
i
g
i
d

p
l
a
s
t
i
c
s
;

t
h
e

A
S
T
M

s
u
g
g
e
s
t
e
d

s
p
e
c
i
m
e
n

s
i
z
e

i
s

1
2

1
2

5
0

m
m
,

n
o
t

c
u
b
i
c

s
p
e
c
i
m
e
n
s
.
F
o
r

d
e
t
e
r
m
i
n
a
t
i
o
n

o
f

e
l
a
s
t
i
c

c
o
n
s
t
a
n
t
s
,

s
e
v
e
r
a
l

s
e
r
i
e
s

o
f

s
p
e
c
i
m
e
n
s

w
i
t
h

d
i
f
f
e
r
e
n
t


h
/
l


r
a
t
i
o

a
r
e

n
e
c
e
s
s
a
r
y
.
I
n
a
c
c
u
r
a
c
i
e
s

o
c
c
u
r

d
u
e

t
o

s
p
e
c
i
m
e
n

m
i
s
a
l
i
g
n
m
e
n
t
,

f
r
i
c
t
i
o
n

a
t

t
h
e

l
o
a
d

p
o
i
n
t
s
,

i
m
p
r
e
c
i
s
e

s
t
r
a
i
n

m
e
a
s
u
r
e
m
e
n
t
,

i
n
a
d
e
q
u
a
t
e


h
/
l


r
a
t
i
o
,

a
n
d

e
l
a
s
t
i
c
-
p
l
a
s
t
i
c

d
e
f
o
r
m
a
t
i
o
n
;

t
h
e

l
i
m
i
t
a
t
i
o
n
s

i
m
p
o
s
e
d

b
y

t
h
e
o
r
e
t
i
c
a
l

c
o
n
s
i
d
e
r
a
t
i
o
n
s

m
u
s
t

b
e

t
a
k
e
n

i
n
t
o

a
c
c
o
u
n
t
.
A
c
t
u
a
l

p
a
t
h

l
e
n
g
t
h

i
s

u
n
k
n
o
w
n

u
n
l
e
s
s

s
p
e
c
i
m
e
n

s
h
a
p
e

i
s

s
i
m
p
l
e
;

p
a
t
h

l
e
n
g
t
h

i
s

d
e
t
e
r
m
i
n
e
d

b
y

a
v
e
r
a
g
i
n
g

t
h
e

a
c
t
u
a
l

l
e
n
g
t
h
s
.
T
h
e

v
e
l
o
c
i
t
y

o
f

p
r
o
p
a
g
a
t
i
o
n

o
f

a
n

u
l
t
r
a
s
o
n
i
c

w
a
v
e

c
a
n

b
e

d
e
p
e
n
d
e
n
t

o
n

t
h
e

f
r
e
q
u
e
n
c
y

o
f

o
s
c
i
l
l
a
t
i
o
n
.
P
u
r
e

l
o
n
g
i
t
u
d
i
n
a
l

a
n
d

s
h
e
a
r

w
a
v
e
s

p
r
o
p
a
g
a
t
e

o
n
l
y

i
n

d
i
r
e
c
t
i
o
n
s

p
a
r
a
l
l
e
l

t
o

a
x
i
s

o
f

m
a
t
e
r
i
a
l

s
y
m
m
e
t
r
y
.

S
o
u
r
c
e
:

R
h
o
,

J
.
Y
.
,


U
l
t
r
a
s
o
n
i
c
s
,


3
4
,

7
7
7
,

1
9
9
6
.

W
i
t
h

p
e
r
m
i
s
s
i
o
n
.

0266/ ch24/frame Page 358 Friday, October 22, 1999 9:42 AM

Ultrasonic Methods for Evaluating Mechanical Properties of Bone

359

where

k

is the bulk modulus. The bulk modulus for an isotropic material is
. (24.3)
Here,

is Poissons ratio. This phenomenon refers to geometric dispersion in which the velocity
of propagation depends on the external dimensions of the material in which the wave is propa-
gated.

2,9,10

The ultrasonic technique has been used to measure both the elastic modulus of the
cancellous bone structure (structural modulus) and the elastic modulus of trabecular bone material
(material modulus).

11

The wavelength of the ultrasonic wave must be larger than the characteristic
dimension of the material in order to measure structural modulus. Conversely, ultrasonic waves
with wavelengths shorter than the characteristic dimension of the material would be less affected
by the elasticity of the structure, and would propagate at velocities determined by the elasticity of
trabecular material.

11

Several investigators have assumed cortical bone specimens to be orthotropic with nine
independent elastic constants.

6,12-15

Consideration must be given to the different structures of cortical
and cancellous bone. Techniques for the ultrasonic velocity measurement for cortical bone have
utilized frequencies between 2 and 10 MHz. These relatively high frequencies allow accurate
determination of the time delay due to propagation through cortical specimens with dimensions as
small as 5 mm.

2

Both cylindrical and cubic specimen shapes have been used, but cubic specimens
offer the possibility of velocity measurement in several different directions. The use of ultrasound
to measure the elastic properties of the porous structure of cancellous bone is a much more complex
problem. Lower-frequency waves (50 to 100 kHz) of longer wavelength are necessary to determine
the elastic properties of the cancellous bone structure due to its high porosity.

3,11

Porosity can be
quantied by measuring the void fraction, dened as the ratio of void volume to total volume. Void
fractions range from 0.5 to 0.8 for cancellous bone. A direct consequence of the large void fraction
is that ultrasonic waves are much more strongly attenuated in cancellous bone than in cortical bone.
In other words, the velocity of propagation of an ultrasonic wave is strongly dependent on the
frequency of oscillation. The dispersive nature of cancellous bone must be analyzed carefully in
order to obtain meaningful values for its elastic properties.
The shear moduli and Poissons ratio for cancellous bone have rarely been measured using
ultrasound, since bar wave propagation techniques may not be adequate to measure Poissons ratio
for the cancellous bone. Ashman et al.

2

and Rho

6

show that the bulk mode works well for the
measurement of the orthotropic elastic properties of cortical bone. In the same context, bulk wave
propagation in cancellous bone may provide an adequate method for measuring Poissons ratio.
Bulk wave propagation in cubic cancellous bone specimens measuring approximately 10 mm on
a side may be achieved by raising the frequency of the ultrasonic waves from 50 to approximately
500 kHz. This would reduce the wavelength from approximately 40 to about 4 mm. To achieve
this task, one should be careful not to increase the frequency so much as to cause continuum
breakdown, where the waves begin to propagate along individual trabeculae.

16

Ultrasonic velocity
measurements should be made at intermediate frequencies between 50 and 500 kHz to determine
at what frequency a transition between bulk and bar waves occurs.
In a solid material, as mentioned above, ultrasonic velocity is equal to the square root of the ratio
of the elastic modulus to density. Whether or not this is true in a porous material like cancellous bone
is critically dependent on the wavelength of sound in relation to the bone specimen and microstructural
dimensions.

11,17

Even though Ashman and Rho

11

have found good agreement between the ultrasonically
determined and mechanically measured elastic modulus of cancellous bone structure, the conicting
results for high-frequency sound velocities relative to the microscopic modulus (i.e., trabecular bone
material) have suggested that the theory must be modied.

17

These modications need to deal with
k
E
=
3(1 2 )

0266/ ch24/frame Page 359 Friday, October 22, 1999 9:42 AM

360

Mechanical Testing of Bone and the BoneImplant Interface

cancellous bone structure and describe the relationship of high-frequency clinical measurements to
bone microstructure. One reason for the conicting results is the unknown actual path length.
Ultrasonic velocity data at mixed cortical and cancellous bone sites have been reported. Langton
et al.

18

suggest that the path must be linear. However, McCartney et al.

19

demonstrated a nonlinear
transverse path of ultrasound in thick-walled cylinders. It is essential to know the exact path of the
ultrasound beam in bone, as the determination of ultrasonic velocity is based on measuring the
time of ight and the distance traveled.

II. SAMPLE PREPARATION

Cortical bone specimens are obtained from the diaphysis of long bones. Positions along the length
of the bone are standardized by using external anatomical landmarks on the surface of the bone to
dene specimen locations. For an example, a femoral bone is placed posterior quadrant down on
a at surface. Three anatomical structures, the quadrant tubercle and the medial and lateral knee
condyles, come into contact with the surface. These contact points are used to dene the length of
the bone. The length is then divided into any number of increments at which transverse sections
are cut. The anterior, medial, posterior, and lateral quadrants are then marked on the transverse
sections. Specimens can then be located with respect to these four locations on each transverse
section. A three-step cutting procedure is used to fabricate cortical specimens of the desired
geometry (side dimensions on the order of 5 mm). Bones are cleaned of attached muscles prior to
cutting. The rst step is to make rough cuts with a band saw to obtain the transverse sections along
the length of the bone. Although it is easy to burn bone using a band saw, rough cutting affects an
area only 1 to 2 mm into the bone from the cut.

4

In the second step, the cutting process consists
of machining cubic specimens from the anterior, posterior, medial, and lateral quadrants at the
several sections along the length of the bone. This is accomplished with a

-in. diameter end mill
while the bone is kept wet. Materials burned by the rst step can be removed during this step.
Since a particular geometry proposed by Van Buskirk et al.

14

is utilized to obtain all necessary
velocity measurements from one specimen, the third step consists of cutting the edges of the cube
off at 45

angles with a v using a slow-speed diamond cutter

2

(Figure 24.1). Specimens axes are
aligned with the axial (

x

3

), the circumferential (

x

2

), and the radial (

x

1

) axes for each bone.
Cancellous bone specimens can be obtained from the ends of long bones. The inhomogeneity
of cancellous bone and the fact that cancellous elastic properties are less well documented require
a greater number of cancellous specimens. Location of the specimens is standardized on several
12-mm slabs using plastic templates

3

and care is taken to ensure that specimen orientation is
recorded. The 10

10

10 mm specimens are cut from the 12-mm sections with a slow-speed diamond
wafering saw ensuring no dense epiphyseal bone is included in the 10

10

10 mm cubes. Diamond
wafering blades are particularly good for making smooth and parallel cuts. Specimens axes are aligned

FIGURE 24.1

Cubic specimen developed by Van Buskirk et al.

14

(From Rho, J.,

Ultrasonics,

34, 777, 1996.
With permission.)

0266/ ch24/frame Page 360 Friday, October 22, 1999 9:42 AM

Ultrasonic Methods for Evaluating Mechanical Properties of Bone

361

with the superior-inferior (

x

3

), medial-lateral (

x

2

), and anterior-posterior (

x

1

) axes for each bone.
Bone marrow is removed from each cube with a water jet prior to analysis, and specimens are kept
moist in saline at all times between testing. Individual trabeculae (0.3 mm diameter by 2 mm length)
can be isolated using a stereomicroscope, scalpel, and forceps.

III. TESTING METHODS

The general method used to measure the ultrasonic wave velocities is a pulse transmission
technique. The basis of this technique is the time delay of propagation between the transmitted
and received ultrasonic waves. From the time delay and the thickness of the specimen, the
ultrasonic velocity can be calculated. To measure ultrasonic velocities of cortical bone, a fre-
quency of 2.25 MHz is used. To measure ultrasonic velocities of cancellous bone structure, 50
kHz is used. At 50 kHz, the wavelength of the ultrasonic wave is much longer than the charac-
teristic dimension of the cancellous bone structure. These specic frequencies are optimized by
comparing elastic properties determined ultrasonically at several frequencies for cortical or
cancellous bone with those measured by mechanical testing techniques. When comparing the
ultrasonically measured elastic properties with those determined mechanically, the ultrasonic
strain rate is often brought up. It would appear that the strain rate is relatively high, but strain
rate is change in length/original length/time. For 2.25 MHz the time is indeed short, but so is
the change in length. The values of elastic properties determined ultrasonically are in fact similar
to those obtained at quasi-static rates.

2,3,11

By using a pulse generator, ultrasonic waves produced by the transducers are recorded by an
oscilloscope which allows the time delay between the transmitted and received waves to be
determined. The time delay for both longitudinal and shear measurements is measured from the
start of the input signal to the start of the output signal. The rst received point is taken to be the
beginning of the output signal (Figure 24.2). Longitudinal and shear transducers of 2.25 MHz are
used to measure the velocity of the cortical bone specimens and trabecular bone material (material
modulus), while those of 50 kHz are used to measure the structural velocity of the cancellous bone.
Coupling of the transducers to the bone is done with water for the longitudinal mode transducers
and with a viscous liquid acoustic coupler for the transverse mode transducers.
The density of bone must be used in the calculation of the elastic coefcients from the wave
velocities. The density range of cortical bone from 1700 to 2000 kg/m

3

is determined by
Archimedes principle. The density range of the cancellous structure (often referred to as the
apparent density) from 50 to 1000 kg/m

3

is determined by weighing the cancellous structure after
removing the bone marrow and dividing by the volume of the overall physical dimensions, including
pores. The specimens are centrifuged to remove residual water from the intertrabecular spaces prior
to weighing.

FIGURE 24.2

Graphic representation for measuring time delay with a pulse propagation method.

0266/ ch24/frame Page 361 Friday, October 22, 1999 9:42 AM

362

Mechanical Testing of Bone and the BoneImplant Interface

IV. DATA COLLECTION AND CALCULATION

Cortical bone is an inhomogeneous anisotropic, viscoelastic material. However, it is reasonable to
model cortical bone as linearly elastic and anisotropic.

1,14,15,21

The constitutive equation, Hookes
law, for a linear anisotropic elastic material is

i



=

c

ij

j

(24.4)
where the terms

i

and

j

represent components of the stress and strain tensors, respectively, and

c

ij



is the matrix of elastic coefcients. The matrix of elastic coefcients for an orthotropic material
is written as follows:
. (24.5)
There are 12 nonzero components of which 9 are independent. Equation 24.4 can be rewritten in
terms of the compliance matrix,

s

ij

.

j



=

s

ij

i

(24.6)
The compliance matrix for an orthotropic material is expressed in terms of elastic properties such
as Youngs modulus, shear modulus, and Poissons ratio and is
. (24.7)
where

E

i



is the Youngs moduli in the

i

direction,

ij

is Poissons ratio for strain in the

j

direction
with stress applied in the

i

direction, and

G

ij

are the shear moduli in the

ij

plane (

i, j

= 1, 2, or
3). The matrix of elastic coefcients and the compliance matrix are symmetric; hence,
. (24.8)
[ ] c
c c c
c c c
c c c
c
c
c
ij
=

11 12 13
12 22 23
13 23 33
44
55
66
0 0 0
0 0 0
0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
[ ] [ ] c s
E E E
E E E
E E E
G
G
G
ij
ij

= =

1
1
21
2
31
3
12
1 2
32
3
13
1
23
2 3
23
31
12
1
0 0 0
1
0 0 0
1
0 0 0
0 0 0
1
0 0
0 0 0 0
1
0
0 0 0 0 0
1

ij
i
ji
j
E E
=

0266/ ch24/frame Page 362 Friday, October 22, 1999 9:42 AM

Ultrasonic Methods for Evaluating Mechanical Properties of Bone

363

The relationship between

c

ij

and

s

ij

can be accomplished in closed form as outlined in Table 24.2.
However, the matrix of elastic coefcients and the compliance matrix must be positive.

1,21

Therefore,
to satisfy this requirement,

s

11

,

s

22

,

s

33

,

s

44

,

s

55

,

s

66

> 0,
(24.9)
and
Similar requirements must be satised by the components of the matrix of elastic coefcients.

E

1

,

E

2

,

E

3

,

G

23

,

G

31

,

G

12

> 0
(1

23

32

), (1

13

31

), (1

12

21

) > 0 (24.10)
and
1

12

21

23

32

31

13
2
21

32

13
> 0.
These relations provide an additional check on the experimentally obtained values of the elastic
coefcients. Relations between velocities and elastic coefcients for an orthotropic material are
listed in Table 24.3. These relationships are the solutions of the equations of motion of elastic
TABLE 24.2
Analytical Expressions for the Inversion between
the Orthotropic Matrix of Elastic Coefcients and
the Orthotropic Compliance Matrix
Source: Ashman, R.B., Ph.D. dissertation, Tulane University, New
Orleans, LA, 1982. With permission.
c
s s s
11
22 33 23
2
=

c
s s s s
12
13 23 12 33
=

c
s
44
44
1
=
c
s s s
22
33 11 13
2
=

c
s s s s
13
12 23 13 22
=

c
s
55
55
1
=
c
s s s
33
11 22 12
2
=

c
s s s s
23
12 13 33 11
=

c
s
66
66
1
=
= + s s s s s s s s s s s s
11 22 33 11 23
2
22 13
2
33 12
2
12 23 22
2
s s
s s
s s
s s
s s
s s
11 12
12 22
22 23
23 33
11 13
13 33
0 0 0 > > > , , ,
s s s
s s s
s s s
11 12 13
12 22 23
13 23 33
0 > .
0266/ ch24/frame Page 363 Friday, October 22, 1999 9:42 AM
364 Mechanical Testing of Bone and the BoneImplant Interface
waves in a unbounded medium.
1,8,14
It should be noted that the velocities associated with the diagonal
directions are not necessarily purely longitudinal or purely transverse.
1
The velocities measured in
these directions are called quasi-longitudinal and quasi-transverse.
A total of 18 different velocities can be measured, three in each of six directions using the
specimen shape developed by Van Buskirk et al.
14
shown in Figure 24.1. However, only nine
velocities could be measured from tibial cortical specimens due to the thin cortex, the shape of
which is shown in Figure 24.3. Two of the nine independent orthotropic elastic constants (c
12
and
c
13
) cannot be obtained using this specimen shape, but can be obtained by relationships assuming
transverse isotropy: c
13
= c
23
and c
66
= (c
11
c
12
). However, seven elastic coefcients are sufcient
TABLE 24.3
Relationships between the Measured Velocities and
the Terms of the Matrix of Elastic Coefcients
2,6,14
Relations permitting to check the assumption of orthotropy:
v
i
: velocity of a longitudinal wave in i direction
v
ij
: velocity of a transverse wave traveling in i direction, with particle
motion in j direction
v
ij/k
: velocity of a transverse wave traveling in (i + j)/ direction with
particle motion in k direction
v
ij/ij
: velocity of a transverse or longitudinal wave traveling in (i + j)/
direction with particle motion in the ij plane; velocities are quasi-
longitudinal or quasi-transverse
Source: Rho, J., Ultrasonics, 34, 777, 1996. With permission.
c v
11 1
2
=
c v
22
=
2
2
c v
33
=
3
2
c v v
44
= =
23
2
23
2
c v v
55
= =
13
2
31
2
c v v
66
= =
12
2
12
2
c c c v c c v c
12 11 66 22 66 66
( 2 ) ( 2 ) = + + +
12 12
2
12 12
2
/ /
c c c v c c v c
13 11 55 33 55 55
( 2 ) ( 2 ) = + + +
13 13
2
13 13
2
/ /
c c c v c c v c
23 22 44 33 44 44
( 2 ) ( 2 ) = + + +
23 23
2
23 23
2
/ /
c c v
55 66
2 + =
23 1
2
/
c c v
44 66
2 + =
13 2
2
/
c c v
44 55
2 + =
12 3
2
/
2
2
0266/ ch24/frame Page 364 Friday, October 22, 1999 9:42 AM
Ultrasonic Methods for Evaluating Mechanical Properties of Bone 365
to determine whether the elastic symmetry of the material is orthotropic, transversely isotropic, or
isotropic.
Cancellous bone is also considered linearly elastic and anisotropic.
22
The most straightforward
relationships between velocities and elastic properties are derived for waves of long wavelength
propagating in bar-shaped specimens of narrow cross section.
7
Longitudinal velocities v
i
of waves
propagating in the i direction and shear wave velocities v
ij
for propagation in the i direction with
particle motion in the j direction are related to Youngs moduli and shear moduli by
E
i
= v
i
2
(24.11)
and
G
ij
= v
ij
2
= v
ji
2
(24.12)
where is the apparent density of the structure, E
i
is the elastic modulus in the i direction, and G
ij
is the shear modulus. Two requirements must be satised for correct application of Equations 24.11
and 24.12. First, the wavelength must be greater than the characteristic dimensions of the structure,
if the structure is to be considered a continuum. Second, the wavelength must be larger than the
cross-sectional dimensions of the specimen. Wave propagation fullling these requirements is often
referred to as bar wave propagation. In order for an ultrasonic wave to propagate at a velocity
governed by the elastic properties of the gross structure, the rst requirement must be satised.
Harrigan et al.
16
presented arguments that the 5-mm characteristic dimension is the lower limit at
which cancellous bone can be considered a continuum. The 50-kHz frequency of the transmitted
ultrasonic wave, with measured velocities between 1000 and 1600 m/s, predicts a wavelength
(velocity/frequency) on the order of 20 mm, well above the 5-mm characteristic dimension of the
cancellous structure and the 10-mm cross-sectional dimension of the specimen.
FIGURE 24.3 Specimen shapes used to obtain the necessary velocities for the calculation of the anisotropic
elastic coefcients. v
i
is velocity of a longitudinal wave in the i direction. v
ij
is velocity of a transverse wave
in i direction with particle motion in j direction. v
ij/ij
is velocity of a transverse or longitudinal wave in the (i
+ j)/ direction with particle motion in the ij plane (quasi-longitudinal or quasi-transverse velocity). (From
Rho, J., Ultrasonics, 34, 777, 1996. With permission.)
2
0266/ ch24/frame Page 365 Friday, October 22, 1999 9:42 AM
366 Mechanical Testing of Bone and the BoneImplant Interface
V. RESULTS FOR MEASUREMENTS IN BONE
Eight right femorae were obtained from frozen unembalmed human cadavers. None of the indi-
viduals had a history of bone disease. Cortical bone specimens were obtained from the four
specimens from each of three transverse sections at 30, 50, and 70% of total bone length
(Figure 24.4). There is a greater variability along the length of a whole bone than around its
circumference (the variation in the mechanical properties around the circumference being less than
10%) which could be due to a changing thickness in the longitudinal direction. The elastic moduli
FIGURE 24.4 The position of femoral cortical and cancellous specimens. The three directions are coincident
with the long axis of the bone; the one and two directions are the radial and circumferential directions for
cortical bone and anterioposterial and mediolateral directions for cancellous bone, respectively.
0266/ ch24/frame Page 366 Friday, October 22, 1999 9:42 AM
Ultrasonic Methods for Evaluating Mechanical Properties of Bone 367
TABLE 24.4
Elastic Properties (standard deviation in parentheses)
of Human Femoral Cortical Bone
Elastic Properties Minimum Maximum Average (SD)
E
1
(GPa) 7.4 16.2 11.6 (1.9)
E
2
(GPa) 8.1 16.4 12.2 (1.9)
E
3
(GPa) 11.4 31.9 19.9 (2.7)
G
12
(GPa) 2.3 5.6 4.0 (0.7)
G
13
(GPa) 3.6 6.3 5.0 (0.7)
G
23
(GPa) 3.9 6.9 5.4 (0.7)

12
0.267 0.563 0.417 (.070)

13
0.143 0.323 0.232 (.039)

21
0.323 0.548 0.436 (.50)

23
0.147 0.302 0.225 (.032)

31
0.294 0.496 0.396 (.053)

32
0.309 0.411 0.362 (.020)
(kg/m
3
) 1453 1996 1811 (148)
Note: The direction 3 is coincident with the long axis of the bone; the
directions 1 and 2 are the radial and circumferential directions, respec-
tively. Values of elastic properties are expressed in GPa.
TABLE 24.5
Elastic Properties (standard deviation in parentheses)
of Human Proximal Femoral Cancellous Bone
Level Position E
1
E
2
E
3
I 1 1584 (485) 1620 (501) 2212 (663)
2 1698 (628) 1646 (653) 2244 (760)
3 2114 (633) 2061 (647) 2710 (745)
4 2036 (620) 1846 (685) 2476 (760)
II 5 860 (705) 804 (706) 1453 (644)
6 620 (326) 590 (392) 1488 (810)
III 7 299 (162) 570 (257) 902 (436)
8 321 (357) 426 (299) 801 (297)
9 310 (164) 482 (204) 1170 (363)
10 261 (174) 540 (220) 753 (440)
IV 11 142 (85) 227 (193) 565 (220)
12 175 (139) 274 (299) 622 (312)
13 220 (144) 308 (223) 792 (453)
14 286 (238) 366 (277) 808 (599)
V 15 257 (132) 309 (282) 764 (505)
16 217 (196) 242 (235) 518 (282)
Note: The direction 3 is coincident with the long axis of the bone; the
directions 1 and 2 are the anterioposterior and mediolateral directions,
respectively. Values of elastic properties are expressed in MPa.
0266/ ch24/frame Page 367 Friday, October 22, 1999 9:42 AM
368 Mechanical Testing of Bone and the BoneImplant Interface
in the radial or circumferential directions correlate to the axial/longitudinal modulus, but the
correlation coefcients are relatively low (0.4 and 0.2, respectively). However, all three perpendic-
ular moduli correlated even better with the density of the material, on the order of 0.5 to 0.8.
Cancellous bone specimens are obtained from both the proximal and distal femorae. Tables 24.4
through 24.6 provide the values in elastic properties of cortical and cancellous bone along the
length and around the quadrant of the bone.
5
Mechanical properties vary signicantly around the
periphery and along the length, and there is a pronounced anisotropy of cancellous structure in the
highly mechanically loaded bone. Overall, the differences between the mechanical properties in
cancellous bone are much broader than those in cortical bone.
The pulse transmission ultrasonic technique has been undertaken to characterize the actual path-
way and the wavelength dependence in relation to the bone specimen and microstructural dimen-
sions.
23
From one human right tibia, 20 individual trabeculae and 20 cylindrical cancellous bone
specimens are obtained. The average velocity through individual trabecular bone is 2901 161 m/s
(mean SD), while the mean velocity through cylindrical cancellous bone specimens is 2717 171
m/s. Thus, the velocity through the cylindrical cancellous bone specimens is underestimated as 6.4%
as much as that through individual trabeculae. As expected, the actual pathway of trabecular bone
material measurement is slightly longer than the average length of the specimen. It may be postulated
that ultrasonic energy has propagated along trabeculae, the shortest pathway. The wavelengths, deter-
mined from the measured velocities ranging from 2600 to 3300 m/s, are on the order of 1 mm. The
1-mm wavelength is larger than the cross-sectional dimension of the trabeculae (0.1 to 0.5 mm);
hence, bar wave propagation can be applied to calculate the elastic modulus. Also, since the 1-mm
wavelength is smaller than the 5-mm characteristic dimension of the structure, the wave may be
considered not to be affected by the cancellous bone structure.
TABLE 24.6
Elastic Properties (standard deviation in parentheses)
of Human Distal Femoral Cancellous Bone
Level Position E
1
E
2
E
3
I 1 606 (212) 604 (267) 1627 (396)
2 1093 (474) 1129 (526) 2328 (818)
3 1654 (692) 1173 (485) 2676 (1037)
4 1071 (329) 1054 (348) 1941 (511)
5 1123 (351) 1096 (378) 2468 (911)
6 1118 (361) 1076 (295) 2110 (517)
7 1828 (562) 1842 (529) 3098 (1144)
II 8 420 (217) 416 (227) 1241 (338)
9 264 (170) 280 (206) 967 (367)
10 599 (367) 669 (392) 1516 (797)
11 571 (435) 527 (353) 1409 (562)
12 224 (114) 183 (122) 683 (357)
13 312 (285) 386 (441) 810 (584)
14 917 (460) 643 (366) 2320 (958)
15 599 (273) 437 (269) 1487 (718)
16 723 (535) 762 (609) 1991 (1233)
III 17 190 (161) 171 (158) 678 (454)
18 363 (365) 285 (295) 1212 (957)
Note: The direction 3 is coincident with the long axis of the bone; the
directions 1 and 2 are the anterioposterior and mediolateral directions,
respectively. Values of elastic properties are expressed in MPa.
0266/ ch24/frame Page 368 Friday, October 22, 1999 9:42 AM
Ultrasonic Methods for Evaluating Mechanical Properties of Bone 369
ACKNOWLEDGMENT
This work was supported in part by grants from the Whitaker Foundation.
REFERENCES
1. Ashman, R.B., Ultrasonic Determination of the Elastic Properties of Cortical Bone: Techniques and
Limitations, Ph.D. dissertation, Tulane University, New Orleans, LA, 1982.
2. Ashman, R.B., Cowin, S.C., Van Buskirk, W.C., and Rice, J.C., A continuous wave technique for the
measurement of the elastic properties of cortical bone, J. Biomech., 17, 349, 1984.
3. Ashman, R.B., Rho, J.Y., and Turner, C.H., Anatomical variation of orthotropic elastic moduli of the
proximal human tibia, J. Biomech., 22, 895, 1989.
4. Ashman, R.B., Experimental techniques, in Bone Mechanics, Cowin, S.C., Ed., CRC Press, Boca
Raton, FL, 1989.
5. Rho, J.Y., Mechanical Properties of Cortical and Cancellous Bone, Ph.D. dissertation, University of
Texas Southwestern Medical Center, Dallas, TX, 1991.
6. Rho, J.Y., An ultrasonic method for measuring the elastic properties of human tibial cortical and
cancellous bone, Ultrasonics, 34, 777, 1996.
7. Kolsky, H., Stress Waves in Solids, Dover Publications, New York, 1963.
8. Hearmon, R.F.S., An Introduction to Applied Anisotropic Elasticity, Oxford University Press, Oxford,
U.K., 1961.
9. Nigro, N.J., Wave propagation in anisotropic bars of rectangular cross section. I. Longitudinal wave
propagation. J. Acoust. Soc. Am., 46, 958, 1968.
10. Thurston, R.N., Elastic waves in rods and clad rods, J. Acoust. Soc. Am., 64, 137, 1978.
11. Ashman, R.B. and Rho, J.Y, Elastic moduli of trabecular bone material, J. Biomech., 21, 177, 1988.
12. Yoon, H.S. and Katz, J.L., Ultrasonic wave propagation in human cortical bone I. Theoretical
considerations for hexagonal symmetry, J. Biomech., 9, 409, 1976.
13. Yoon, H.S. and Katz, J.L., Ultrasonic wave propagation in human cortical bone II. Measurements
of elastic properties and microhardness, J. Biomech., 9, 459, 1976.
14. Van Buskirk, W.C., Cowin, S.C., and Ward, R.N., Ultrasonic measurement of orthotropic elastic
constants of bovine femoral bone, J. Biomech. Eng., 103, 67, 1981.
15. Van Buskirk, W.C. and Ashman, R.B., The elastic moduli of bone, in Mechanical Properties of Bone,
Cowin, S.C., Ed., AMD Vol. 45, American Society of Mechanical Engineers, New York, 1981.
16. Harrigan, T.P., Jasty, M., Mann, R.W., and Harris, W.H., Limitations of the continuum assumption in
cancellous bone, J. Biomech., 21, 269, 1988.
17. Martin, R.B., Determinants of the mechanical properties of bones, J. Biomech., 24 (Suppl. 1), 79, 1991.
18. Langton, C.M., Ali, A.V., Riggs, C.M., et al., A contact method for the assessment of ultrasonic
velocity and broadband attenuation in cortical and cancellous bone, Clin. Phys. Physiol. Meas., 11,
243, 1990.
19. McCartney, R.N., Jeffcott, L.B., and McCarthy, R.N., Transverse path of ultrasound waves in thick-
walled cylinders, Med. Biol. Eng. Comput., 33, 551, 1995.
20. Knets, I.V., Mechanics of biological tissues: a review, Polym. Mech. [translation of Mekhanika
Polimerov], 13, 434, 1978.
21. Jones, R.M., Mechanics of Composite Materials, McGraw-Hill, New York, 1975.
22. Carter, D.R. and Hayes, W.C., The compressive behavior of bone as a two-phase porous structure,
J. Bone Joint Surg., 59A, 954, 1977.
23. Rho, J.Y., Characterization of ultrasonic pathway and wavelength dependence in determining elastic
properties using ultrasound method, Med. Biol. Eng. Comp., 36, 57, 1998.
0266/ ch24/frame Page 369 Friday, October 22, 1999 9:42 AM
0266/ ch24/frame Page 370 Friday, October 22, 1999 9:42 AM

371

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

25

Evaluating Mechanical
Properties of Bone Using
Scanning Acoustic Microscopy

Charles H. Turner and J. Lawrence Katz

CONTENTS

I. Introduction ..........................................................................................................................371
II. Low-Frequency Studies........................................................................................................373
A. Acoustic Impedance Method .........................................................................................373
B. Transmission Velocity Measurements............................................................................375
III. High-Frequency Studies.......................................................................................................377
References ......................................................................................................................................383

I. INTRODUCTION

Scanning acoustic microscopy (SAM) was introduced in the early 1970s by Quate and colleagues

1

at Stanford University. Since then a number of individually constructed and commercially available
reection-mode SAMs have been developed. SAM works on the principle of having a beam of
acoustic waves, generated by a piezoelectric transducer, travel through a high-quality single-crystal
lens (usually sapphire) that focuses the waves onto an interface. Part of the wave is reected back
toward the source and part of the wave is transmitted across the interface (Figure 25.1). The former
wave is what is analyzed to determine the mechanical properties of the reecting surface at the
interface. The fraction of the wave reected from the interface will depend on the acoustic imped-
ance differences of the materials comprising the interface.
Acoustic impedance, generally denoted by

Z,

is the product of local density (

) times acoustic
velocity (

v

), i.e.,

Z

=

v,

where the units for

Z

are called Rayleighs, abbreviated as Rayls.

Z

is
usually given in terms of MRayls, 1 MRayl = 10

6

kg m

2

s

1

. For optimum results a liquid medium
is required to couple the acoustic lens to the material being imaged. Thus, if distilled (or deionized)
water is used as the couplant uid for studying bone, the fractional value of the reected wave will
be determined by solving the equation for the reection coefcient

r

, dened as
(25.1)
where

Z

1

and

Z

2

are the acoustic impedances of the water and bone, respectively. By using an
average value of

Z

2

= 8 MRayls for bone and

Z

1

= 1.5 MRayls for water at 20

C,

r

= 0.68, or 68%
of the incident sound wave will be reected back through the same lens to the piezoelectric
transducer now acting as a receiver. A graph of the reection coefcient vs. acoustic impedance
indicating the relative location of various materials is given in Figure 25.2.
r =

+


1 2
2 1

0266/ ch25/frame Page 371 Friday, October 22, 1999 9:47 AM

372

Mechanical Testing of Bone and the BoneImplant Interface

The amplitude of the reected wave arriving at the transducer will have been attenuated during
its propagation through the water couplant; this attenuation increases with increased frequency of
the acoustic wave. It is the amplitude of this attenuated wave that is stored as a voltage in the
electronic system of the SAM and then displayed as a shade of gray proportional to the voltage.
In the scanning mode, the lens and/or specimen raster relative to one another over a preprogrammed
number of pixels resulting in an image in shades of gray (or pseudo-color) of the relative acoustic
impedance properties of the material over the area being scanned.
The attenuation of the reected beam makes it somewhat more difcult to use SAM with soft
tissues whose values of

Z

generally are close to that of the water couplant, so that the amplitude
of the signal arriving at the transducer is relatively weak, potentially limiting the ability to discrim-
inate between aspects of the image.
Using acoustic impedance information in the SAM image to determine the mechanical prop-
erties of the material being scanned requires some intermediate steps. Generally, the mechanical

FIGURE 25.1

Sketch of reflected and transmitted acoustic waves at an interface between two materials,
including denition of acoustic impedance,

Z,

and reection coefcient,

r

.

A

R

and

A

I

are the amplitudes of
the reected and incident waves, respectively.

FIGURE 25.2

Plots of reflection and transmission curves for

r

vs.

Z,

indicating properties of different
materials. (Adapted from Katz, J.L. and Meunier, L.,

J. Biomech. Eng.,

115, 543, 1993. With permission.)

0266/ ch25/frame Page 372 Friday, October 22, 1999 9:47 AM

Evaluating Mechanical Properties of Bone Using Scanning Acoustic Microscopy

373

properties derivable from acoustic wave propagation studies that are of interest are the technical
moduli: Youngs modulus, E; Bulk modulus, K; shear modulus, G; and Poissons ratio,

, treating
the material as isotropic and homogeneous; or the elastic stiffness coefcients,

C

ij

, for an anisotropic
material. In either case the constants,

C,

are related to density,

, times acoustic velocity squared,
v

2

, i.e.,

C

=

v

2

as is obtained by solving the wave equation.

2

Clearly, this is closely related to the
acoustic impedance

Z

=

v,

as both depend on the density and elastic properties of the material.
Thus, if one can determine the proportionality between the two it is possible to obtain the mechanical
properties of the biological material from the SAM image at a much ner scale, down to micro-
meters, than is possible by mechanical testing.
The spatial resolution of the lens system at high frequencies depends on the operating
frequency, the coupling medium, and the quality and properties of the transducer. The equation
demonstrating the relation between the spatial resolution and frequency is

v/f

=

, where

is
the wavelength,

f

the frequency, and

v

the velocity of sound. The higher the frequency, the shorter
the wavelength, thus the ner will be the spatial resolution. Unfortunately, the higher the
frequency, the greater the attenuation of the acoustic wave into the body of the material being
examined. This does limit studies of the deep interior of high-

Z

materials such as bone, ceramics,
and metals as opposed to low-

Z

materials such as soft tissues and polymers. For more details
on the physics of SAM see Briggs.

3

This chapter describes several techniques for measuring the mechanical properties of bone
using acoustic microscopy. The methods can be divided into low-frequency and high-frequency
approaches. Low-frequency approaches use acoustic waves ranging in frequency from 30 to 100
MHz. At low frequencies, mechanical properties of bone can be measured using acoustic impedance
or transmission velocity methods. The spatial resolution of these methods ranges from 30 to 100

m.
High-frequency approaches range in frequency from 100 MHz to 2 GHz. These approaches offer
ne spatial resolution (as good as 0.5

m), but at high frequencies the attenuation of the acoustic
waves is very high so transmission velocity methods cannot be used. Instead, mechanical properties
can be measured using acoustic impedance or

xz

curves.

II. LOW-FREQUENCY STUDIES
A. A

COUSTIC

I

MPEDANCE

M

ETHOD

Meunier, Katz, and co-workers in Paris developed a low-frequency pulse-mode SAM for studying
the mechanical properties of bone, skeletal biomaterials, and the boneimplant interface. With this
instrument it was possible to obtain the amplitude of the rst echo of the longitudinal sound wave
reected back into the lens from the specimen surface pixel by pixel. Using this technique Meunier
et al.

4

compared their

Z

measurements and calculation of

r

with literature values of

v

for several
materials. In addition, Meunier et al.

4

showed that there was excellent correlation between the SAM
maps of acoustic impedance and the elastic stiffness coefcients determined using traditional bulk
ultrasonic wave propagation for all four aspects (anterior, lateral, posterior, medial) of femoral
cortical bone specimens from the middiaphyses from both a 25- and a 74-year-old male. The
relationship between

r

and the elastic stiffness in gigapascals for a variety of materials, including
a number of different bone specimens is shown in Figure 25.3; note that the bone data are nearly
linear in the gure. Zimmerman et al.

5

studied the remodeling of femoral cortical bone around hip
implants. In addition to using SAM for studying local inhomogeneities at a microstructural level,
bulk ultrasonic wave propagation and histology were used to measure elastic properties and geo-
metric characteristics, respectively.
Descriptions of additional low-frequency SAM experiments of cortical bone material properties
are presented by Zimmerman et al.

6

The rst describes studying the variations in SAM properties
at each of the four quadrants within each cross section of femoral cortical bones at each of ten
locations along the length of the bones from three elderly males and three elderly females. In

0266/ ch25/frame Page 373 Friday, October 22, 1999 9:47 AM

374

Mechanical Testing of Bone and the BoneImplant Interface

addition to the acoustic properties peaking at midshaft, it was found that the posterior quadrants
had signicantly lower acoustic impedances than the three other quadrants for the six most proximal
sections. At the more distal locations the acoustic impedance values appeared to converge to one
another. These ndings were found to be in essential agreement with the measurements of both
Ashman et al.

7

and Meunier et al.

8

of the longitudinal elastic stiffness coefcient,

C

33

, and density
variations along the femur. Both the Zimmerman et al.

6

and Meunier et al.

8

studies showed that
most bone sections are quite heterogeneous; i.e., even sections with very low average acoustic
properties had numerous locations with high acoustic properties. Thus, the bone remodeling that
may create regions of decreased elastic properties does not appear to be uniform throughout the
cortical regions of bone.
Meunier et al. also performed a SAM study on human tibial cortical bone similar to their study
of human femoral bone.

8

In addition to studying specimens from all four quadrants of the midshafts
of 20 pairs of human femurs from both males and females of varying ages, they also studied
specimens cut from six different aspects of the midshafts of tibial bones from the same cadavers
from which the femoral specimens were obtained. Six aspects were used here because of the near-
triangular cross section of the tibiae; this comprised one specimen from each of the three sides of
the tibia and one from each corner of the triangular shape. Results similar to those cited above for
the femoral specimens were observed for the tibial specimens as well.

8

A SAM experiment studying the relationship between acoustic impedance and modulus of
elasticity is described by Zimmerman et al.

6

for bovine bone. Flattened dumbbell-shaped specimens
of bovine femoral bone were prepared from bone sections soaked in varying amounts of NaF. After
mechanically testing each specimen to failure they were prepared for low-frequency SAM analysis.
Due to the NaF treatment a wide variation of specimen densities was available, and thus it was
possible to obtain a linear relationship between SAM-measured acoustic impedance and mechanical
testing values of Youngs moduli, with an

r

2

= 0.672 over a range of stiffness moduli from 7.5 to
22.5 GPa. This value is lower than the

r

= 0.99 for a second-order regression between the coefcient
of reection from SAM measurements and the stiffness coefcient, along the femoral bone axis,

C

33

, from transmission measurements by Meunier et al.,

8

and the results cited above from Meunier
et al.

4

However, it supports the consistency of the relationship between the acoustic impedance and
mechanical properties of bone. These observations present a strong argument supporting the con-
tention that acoustic impedance is directly related to a material property of bone. Indeed, Yoon and

FIGURE 25.3

Reflection coefficient,

r

, vs. elastic stiffness for a number of materials including bone.

0266/ ch25/frame Page 374 Friday, October 22, 1999 9:47 AM

Evaluating Mechanical Properties of Bone Using Scanning Acoustic Microscopy

375

Katz

9

showed how the elastic stiffness coefcient,

C

ij

,

for bone could be used to calculate the
technical moduli, i.e., Youngs, bulk, and shear moduli for Haversian bone.
For acoustic impedance measurements, specimens can be prepared either fresh or embedded,
either in epoxy or polymethylmethacrylate (PMMA). Whether fresh or embedded, the specimen
is cut with a diamond blade under owing water and then polished with successively ner and
ner grit, down to 0.05

m for the high-resolution studies and of the order of 1

m for the low-
frequency studies. Ultrasonic cleaning is used to clean the surface of both bone particles and
polishing debris. Specimens are kept in a uid environment and refrigerated until being thawed
slowly for the SAM studies.

B. T

RANSMISSION

V

ELOCITY

M

EASUREMENTS

The preferred orientation of mineral crystals and collagen brils within bone give the tissue
directional variation in elastic properties or anisotropy. This anisotropy can be quantied using
acoustic velocity measurements. The measurement requires cross-sectional and longitudinal sec-
tions cut from a bone specimen. The thickness of these sections should be 300 to 500

m. In
previous studies

10-13

acoustic velocity was measured using a scanning acoustic microscope (UH3,
Olympus, Tokyo, Japan) with a 50-MHz transducer (V-390, Panametrics, Waltham, MA). This
system has a spatial resolution of about 60

m. Velocity is measured using the pulse-echo method
(Figure 25.4). As noted above, the elastic stiffness coefcient for a specimen is calculated as

C

=

v

2

, where

is the density of the specimen and

v

is the measured acoustic velocity. The accuracy
of the acoustic method was veried by measuring the elastic constants of materials with known
properties (Figure 25.5). The coefcient of variation for the velocity measurement is 0.3%.

13

An
anisotropy ratio (AR) can be calculated from velocities measured along the symmetry axes of the
bone (e.g., longitudinal and transverse directions for a long bone).
AR = [

V

L

/V

T

]

2

(25.2)
where

V

L

is the longitudinal velocity and

V

T



is the transverse velocity.

FIGURE 25.4

Acoustic velocity measurement. A specimen is cut 300 to 500

m thick using a wire saw,
submerged in distilled water, and xed to the bottom of a chamber in a scanning acoustic microscope. A
50-MHz transducer and lens is used to generate and receive acoustic waves in pulse-echo mode. Delay time
between waves reected at the surface (A) and those reected at the bottom of the specimen (B) are measured
at ve sites for each specimen. Acoustic velocity is calculated as twice the specimen thickness divided by the
delay time. (From Hasegawa, K. et al.,

Bone,

16, 85, 1995. With permission.)

0266/ ch25/frame Page 375 Friday, October 22, 1999 9:47 AM

376

Mechanical Testing of Bone and the BoneImplant Interface

The acoustic method allows anisotropy measurements of the collagen and mineral components
of bone tissue. For measurement of the collagen matrix, specimens are demineralized by submerging
in 10% ethylenediaminetetraacetic acid (EDTA) with 0.1

M

Tris buffer adjusted to pH 7.4 for 72 h.
This procedure removes all of the inorganic material from the bone.

13

For measurement of the bone
mineral, specimens are deproteinized by submerging in 2.5% sodium hypochlorite for 72 h. This
procedure removes the protein matrix from the tissue. Verication of deproteinization can be veried
by subsequent demineralization, which should leave no organic residue.

13

The xation of bone tissue with 10% buffered formalin can affect the acoustic properties of
the bone matrix. Takano et al.

12

showed that formalin xation had no effect on the acoustically
derived elastic coefcients of monkey cortical bone, either before or after deproteinization. How-
ever, formalin xation affected the elastic coefcients of the demineralized bone. Demineralized
bone was signicantly stiffer in both the longitudinal and transverse directions after formalin xation
(Table 25.1). Embedding bone specimens in plastic also can affect the acoustic properties of the
bone matrix. Embedding has no effects on the AR of bone but can affect the acoustic properties
of demineralized bone (Table 25.2).
Initial studies of canine femoral bone

10

showed that the anisotropy of the bone matrix is
considerably different from the anisotropy of the bone mineral (Figure 25.6). Similar results were
observed by Takano et al.
12
for human, monkey, cow, and rat bone (Figure 25.7). In particular, the
anisotropy of the collagen matrix in rat bone is nearly isotropic, AR = 1. This reects the more
random, woven nature of primary rat bone. Nevertheless, the mineral in rat bone has considerable
anisotropy, indicating that mineral anisotropy and collagen anisotropy are not necessarily related.
To a great extent the anisotropy of bone tissue reects predominant loading patterns. In a recent
study, bone tissue anisotropy ratios were measured in dog radii from which the in vivo longitudinal
strain patterns had been measured. The average AR in the posterior cortex was 1.28 0.01, and
the average AR in the anterior cortex was 1.43 0.01 (these values are signicantly different at
p < 0.0001). Interestingly, the posterior cortex is loaded predominantly in compression and the
anterior cortex is loaded in tension. Takano et al.
11
found a strong correlation between the peak in
vivo strains and the AR (Figure 25.8).
FIGURE 25.5 Verification of the acoustic method. For comparison with measured values, elastic constants of
metals and plastics were calculated from mechanical properties provided by the manufacturers. The elastic
constants of cortical bone (canine) were measured using a standard technique reported previously.
10
The slope of
the regression line was not signicantly different from one (p > 0.8, t-test), and the y-intercept was not signicantly
different from zero (p > 0.7, t-test). (From Turner, C.H. et al., J. Biomech., 32, 437, 1999. With permission.)
0266/ ch25/frame Page 376 Friday, October 22, 1999 9:47 AM
Evaluating Mechanical Properties of Bone Using Scanning Acoustic Microscopy 377
III. HIGH-FREQUENCY STUDIES
The initial SAM studies of the acoustic properties of osteons and osteon lamellae were reported
by Katz and Meunier
15
using bovine femoral cortical bone. Specimens were prepared as described
above and imaged using the burst mode at both 400 and 600 MHz in an Olympus UH3 SAM.
Three important characteristic traits of osteonic bone were observed for the rst time. First, adjacent
TABLE 25.1
Effects of Formalin Fixation on Acoustic
Properties of Bone before and after
Demineralization or Deproteinization
C
L
(GPa) C
T
(GPa) AR
(1) Fresh 39.21 0.21 25.15 0.18 1.56 0.01
(2) Fixed 39.28 0.34 25.32 0.20 1.55 0.01
(3) Fresh-demin 4.37 0.02 3.70 0.03 1.18 0.01
(4) Fixed-demin 4.78 0.05 4.01 0.03 1.19 0.01
(5) Fresh-deprot 23.85 0.37 14.83 0.34 1.62 0.04
(6) Fixed-deprot 23.39 0.26 14.56 0.44 1.63 0.04
12 p = 0.96 p = 0.26 p = 0.24
34 (demin) p < 0.001 p < 0.001 p = 0.37
56 (deprot) p = 0.28 p = 0.62 p = 0.84
Mean SEM. C
L
, longitudinal elastic coefcient; C
T
, transverse
elastic coefcient; AR, anisotropy ratio; demin, demineralization
with EDTA; deprot, deproteinization with NaOCl. p values are
from paired t-tests.
TABLE 25.2
Effects of Embedding on Acoustic Properties of
Rat Femoral Bone before and after
Demineralization or Deproteinization
V
L
(m/s) V
T
(m/s) AR
(1) Nonembedded 4246 17 3782 20 1.25 0.02
(2) Embedded 4245 20 3818 11 1.24 0.01
(3) Nonembed-demin 1752 22 1718 33 1.04 0.04
(4) Embed-demin 1825 10 1802 10 1.03 0.01
(5) Nonembed-deprot 3385 35 2742 42 1.57 0.05
(6) Embed-deprot 3356 24 2739 21 1.51 0.03
12 p = 0.94 p = 0.12 p = 0.63
34 (demin) p < 0.01 p < 0.05 p = 0.80
56 (deprot) p = 0.50 p = 0.95 p = 0.31
Mean SEM. V
L
, longitudinal acoustic velocity; V
T
, transverse
acoustic velocity; AR, anisotropy ratio; demin, demineralization
with EDTA; deprot, deproteinization with NaOCl. p values are
from paired t-tests.
0266/ ch25/frame Page 377 Friday, October 22, 1999 9:47 AM
378 Mechanical Testing of Bone and the BoneImplant Interface
lamellae within osteons as well as in the endosteal region alternated regularly in light and dark
gray levels. The gray level of a pixel is proportional to the reection coefcient of the material at
the point it is being examined; i.e., it is a measure related to the acoustic impedance of the material.
The larger the reection coefcient, the lighter the gray level; i.e., a white spot represents a larger
value of Z at that pixel than does a gray spot, while a black spot would represent the lowest value
of Z in the

image. Such variations in gray levels on a SAM micrograph for normal bone, corre-
sponding to variations in Z = v, can arise due to either a change in local material density (essentially
mineral content) and/or changes in elastic properties. The changes themselves can be caused by a
number of factors: different orientation of apatite crystallites relative to the bone axis and/or different
amount of crystallites in adjacent lamellae. Both of those factors would affect the elastic properties
and thus the velocity; the latter factor only would affect the density. Therefore, a complementary
technique, such as backscatter scanning electron microscopy (B-SEM) or microcomputed tomog-
raphy, must be used to determine possible variations in density. This alternation in gray levels
implies an alternation between relative compliancy and stiffness in adjacent lamellae. The outermost
FIGURE 25.6 Anisotropy ratios (velocity divided by transverse velocity) for canine femoral bone.
Demineralization (a) signicantly decreased the anisotropy ratio at angles 0, 10, 20, 30, 40, and 50, while
deproteinization (b) changed the anisotropy ratio only at 30 off-axis. Asterisks indicate signicant differences
(p < 0.05) by paired t-tests. (From Turner, C.H. et al., Bone, 17, 85, 1995. With permission.)
0266/ ch25/frame Page 378 Friday, October 22, 1999 9:47 AM
Evaluating Mechanical Properties of Bone Using Scanning Acoustic Microscopy 379
lamella of each osteon in an image always had a dark gray level, indicating that it was a more
compliant lamella. Also, the outermost lamella of adjacent abutting osteons always exhibited
essentially the same dark gray level to the degree that the osteons appeared to interdigitate (blend)
at their borders. The B-SEM micrographs showed that the osteons are structurally distinct even
though they overlap in acoustic impedance (and thus essentially in elastic properties). This will be
illustrated in the following discussion of similar SAM observations made with human femoral
cortical bone specimens prepared in the same manner described above.
Figure 25.9A is a SAM micrograph of the same specimen of human femoral cortical bone, cut
transversely to the bone axis, described in previous studies by Katz and Meuiner,
15,16
using a 600
MHz burst-mode lens (120 aperture angle). Clearly, the same three observations made with the
canine femoral cortical bone are made here, i.e., alternating gray levels, outermost lamellae dark,
FIGURE 25.7 Anisotropy ratios (AR, means and SEM) for minerialized tissues from rat, cow, human, and
monkey femora. AR is given for the mineralized tissue before and after treatment with EDTA (demineralization)
and NaOCl (deproteinization). AR for mineralized tissues varied signicantly among the groups (p < 0.001,
by ANOVA). AR for demineralized tissue also varied signicantly among the groups (p < 0.001), but AR for
deproteinized tissue did not vary among the groups (p > 0.6). (From Takano, Y. et al., J. Bone Miner. Res.,
11, 1292, 1996. With permission.)
FIGURE 25.8 The anisotropy ratio of the bone tissue was strongly correlated with the measured peak
longitudinal strains in the radii of dogs. Anisotropy ratio tended to be greater in regions under tensile (positive)
strains and lowest in regions under compressive (negative) strains. (From Takano, Y. et al., J. Orthop. Res.,
17, 59, 1999. With permission.)
0266/ ch25/frame Page 379 Friday, October 22, 1999 9:47 AM
380 Mechanical Testing of Bone and the BoneImplant Interface
and interdigitation of the gray levels of the outermost lamellae of adjacent abutting osteons.
Figure 25.9B is a backscatter SEM of the same area shown in Figure 25.9A. It is clear here as well
that the adjacent abutting osteons are separated as opposed to the interdigitation (overlap or
blending) of gray levels seen on Figure 25.9A. The interpretation of the gray levels in the B-SEM
is as follows: the lighter the pixel, the higher the relative density; the darker the pixel, the lower
the relative density. This apparent blending of the acoustic (and therefore elastic) properties of
the outermost lamellae of adjacent abutting osteons may also impact on the modeling of bone as
a hierarchical structurematerials composite.
17
Instead of requiring a compliant interface (matrix)
between stiff hollow bers (osteons) in the model, it may be that each osteon provides its own
compliant matrix via the outermost lamellae, which thus provides the net effect of osteons (stiff
solid hollow cylinders) interfacing with one another through the interdigitating compliant properties
of the adjacent abutting outermost lamellae.
FIGURE 25.9 (A) 600 MHz burst-mode (aperture angle 120, full scale x dimension is 1 mm) SAM micrograph
of a portion of human femoral cortical bone cut transverse to the bone axis. (B) B-SEM micrograph of the same
region shown in Figure 25.4a. The clear delineation of individual Haversian structures is observed here.
0266/ ch25/frame Page 380 Friday, October 22, 1999 9:47 AM
Evaluating Mechanical Properties of Bone Using Scanning Acoustic Microscopy 381
Figure 25.10A is a 600-MHz burst-mode (120 aperture; full-scale x dimension 250 m) of the
dark osteon seen in the upper left-hand region of Figure 25.9A; the alternation of gray levels in
adjacent lamellae is clearly delineated. Figure 25.10B is the xz SAM image along a line through
the midpoint of Figure 25.10A (full-scale x dimension is 250 m; full-scale z is 50 m). xz images
are obtained as follows: the lens is vibrated along the x direction (left to right) while the stage is
held stationary at a location along the y direction (front to back). The lens is lowered (defocused)
to a specic depth below the surface (the z direction) and then scans upward in z while the lens
vibrates in x. The resulting image is an acoustic interference prole in the xz plane providing
additional information about the acoustic properties of the material at each pixel along the x line.
The level nature of the top of the broad band shows that the specimen surface is perpendicular to
FIGURE 25.10 (A) 600-MHz burst-mode (aperture angle 120, full scale x dimension is 250 m) SAM
micrograph of the dark Haversian system seen in the upper left-hand guadrant of Figure 25.9A clearly showing
the variations in gray levels of alternate lamellae. (B) xz SAM image along the line through the midpoint
on Figure 25.10A (full scale x is 250 m; full scale z is 50 m). The level white band shows the sample was
at and well aligned perpendicular to the lens. Slight incursions in the at surface reect, in part, polishing
artifacts and changes in material properties.
0266/ ch25/frame Page 381 Friday, October 22, 1999 9:47 AM
382 Mechanical Testing of Bone and the BoneImplant Interface
the lens, a necessary condition to ensure that experimental artifacts are minimized. The slight
incursions in the broad band reect a combination of property differences and possible polishing
artifacts. The separation between the solid broad white band and the scattered narrower white lines
below is related to the average subsurface longitudinal acoustic velocity.
The xz image is a generalized example of a form of interference pattern between longitudinal and
surface waves generated when a lens whose aperture angle is greater than a critical angle is defocused
and raised while the specimen is motionless. The signal that is produced as the lens is raised is known
FIGURE 25.11 (A) 600-MHz burst-mode (aperture 120, full scale x dimension is 250 m) SAM micro-
graph of a trabecular portion of a specimen of sheep femoral condyle showing clearly the same alternations
in gray levels in the lamellar structure as observed in the osteonic lamellae of Figure 25.10A. (B) 600-MHz
burst-mode (aperture angle 120, full scale x dimension 100 m) SAM of the lower left-hand quadrant of
Figure 25.11A. The disruption of the lamellar structure and properties at the tip of the lunate-shaped defect
is observed.
0266/ ch25/frame Page 382 Friday, October 22, 1999 9:47 AM
Evaluating Mechanical Properties of Bone Using Scanning Acoustic Microscopy 383
as a V(z) curve. Spectral analysis (Fourier transform) of this V(z) curve will provide the Rayleigh wave
velocity for high-Z materials such as ceramics and metals. However, for bone because of the high
attenuation due to porosity, and because the low Rayleigh velocity puts the Rayleigh angle outside
the useful aperture angle of a water-coupled lens,
3
it has not proved possible to obtain a Rayleigh
wave velocity. However, Briggs also points out that with care, at higher resolution in reection, lateral
longitudinal waves can be made to perform a role very similar to that of the Rayleigh waves in stiffer
materials.
3
This last condition is made use of in analyzing the xz curves.
The same alternations in lamellae gray levels in cortical Haversian bone is observed in trabecular
bone. Figure 25.11A is a 600-MHz burst-mode (120 aperture; full scale x dimension 250 m) of
cancellous bone from a sheep femoral condyle. Figure 25.11B is a SAM micrograph taken with
the same lens, over a scan range of 100 m in the x dimension, of the region surrounding the
lunate-shaped defect seen in the lower left-hand corner of Figure 25.11A. In addition to the same
gray level alternations in adjacent lamellae, the effect of the defect on the structural organization
and properties is also clearly delineated. Similar observations of the alternation in lamella gray
levels have recently been made in trabecular bone from a human femoral condyle.
18
REFERENCES
1. Lemons, R.A. and Quate, C.F., Acoustic microscope-scanning version, Appl. Phys. Lett., 24, 163, 1974.
2. Beyer, R. and Lechter, S., Physical Ultrasonics, Academic Press, New York, 1969.
3. Briggs, G.A.D., Acoustic Microscopy, Clarendon Press, Oxford, 1992.
4. Meunier, A., Katz, J.L., Christel, P., and Sedel, K., A reection scanning acoustic microscope for
bone and bone-biomaterials interface studies, J. Orthop. Res., 6, 770, 1988.
5. Zimmerman, M.C., Meunier, A., Katz, J.L., and Christel P., The evaluation of cortical bone remodeling
with a new ultrasonic technique, IEEE Trans. Biomed. Eng., 37, 433, 1990.
6. Zimmerman, M.C., Harten, R.D., Shieh, S.J., et al., Techniques and applications of scanning acoustic
microscopy in bone remodeling studies, in Oxford Textbook of Orthopaedic Trauma, J.A. Buckwalter
et al., Eds., Oxford University Press, in press.
7. Ashman, R.B., Cowin, S.C., van Buskirk, W.C., and Rice, J.C., A continous wave technique for the
measurement of the elastic properties of cortical bone, J. Biomech., 17, 349, 1984.
8. Meunier, A., Riot, O., Katz, J.L., et al., Inhomogeneities in anisotropic elastic constrants of cortical
bone, IEEE 1989 Ultrasonic Symposium, IEEE, New York, 1015, 1989.
9. Yoon, H.S. and Katz, J.L., Ultrasonic wave propagation in human cortical bone. II. Measurement of
elastic properties and microhardness, J. Biomech., 9, 459, 1976.
10. Turner, C.H., Chandran, A., and Pidaparti, R.M., The anisotropy of osteonal bone and its ultrastructural
implications, Bone, 17, 85, 1995.
11. Takano, Y., Turner, C.H., Forwood, M.R., et al., Elastic anisotropy and collagen orientation of osteonal
bone are dependent on the mechanical strain distribution, J. Orthop. Res., 17, 59, 1999.
12. Takano, Y., Turner, C.H., and Burr, D.B., Mineral anisotropy in mineralized tissues is similar among
species and mineral growth occurs independently of collagen orientation in rats: results from acoustic
velocity measurements, J. Bone Miner. Res., 11, 1292, 1996.
13. Hasegawa, K., Turner, C.H., Recker, R.R., et al., Elastic properties of osteoporotic bone measured by
scanning acoustic microscopy, Bone, 16, 85, 1995.
14. Turner, C.H., Rho, J., Takano, Y., et al., The elastic properties of trabecular and cortical bone tissues
are similar: results from two microscopic measurement techniques, J. Biomech., 32, 437, 1999.
15. Katz, J.L. and Meunier, A., Scanning acoustic microscope studies of the elastic properties of osteons
and osteon lamellae, J. Biomech. Eng., 115, 543, 1993.
16. Katz, J.L. and Meunier, A., Material properties of single osteons and osteonic lamellae using high
frequency scanning acoustic microscopy, in Bone Structure and Remodeling, Odgaard, A. and Weinans,
H., Eds., World Scientic Publishing, Singapore, 1995, 157.
17. Katz, J.L., On the anisotropy of Youngs modulus of bone, Nature, 283, 105, 1980.
18. Bumrerraj, S., M.S. project, Department of Biomedical Engineering, Case Western Reserve University,
Cleveland, OH, May, 1999.
0266/ ch25/frame Page 383 Friday, October 22, 1999 9:47 AM
0266/ ch25/frame Page 384 Friday, October 22, 1999 9:47 AM

385

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

26

Peripheral Quantitative
Computed Tomography
for Evaluating Structural
and Mechanical Properties
of Small Bone

Jos Luis Ferretti

CONTENTS

I. Introduction ..........................................................................................................................385
II. What Is a pQCT Machine?..................................................................................................386
III. What Does pQCT Measure?................................................................................................388
IV. How Does a pQCT Machine Operate?................................................................................390
A. ContMode.......................................................................................................................391
B. PeelMode........................................................................................................................391
C. CortMode........................................................................................................................392
V. What Kind of Variables and Bone Properties Are Measured by pQCT? ...........................393
A. Indicators of Bone Mass.............................................................................................393
B. Indicators of Bone Apparent Density.........................................................................393
C. Indicators of Bone Architectural Quality ......................................................................393
D. Indicators of the Whole-Bone Quality...........................................................................394
VI. How Can the Raw Information Provided by pQCT Be Applied? ......................................395
A. Analysis of the Pathogenesis of Some Experimental Effects
on Bone Biomechanics...................................................................................................395
B. Analysis of the Mechanostat Status by Means of Distribution/Quality Curves .......397
C. Analysis of the Architectural Efciency of Bone Material Distribution
within Bone Diaphyses ..................................................................................................397
D. Analysis of Threshold-Dened ROIs When Cortical and Trabecular Bone
Are Indistinguishable .....................................................................................................400
E. Analysis of Muscle/Bone Interrelationships..................................................................400
VII. Concluding Remarks............................................................................................................400
References ......................................................................................................................................402

I. INTRODUCTION

Quantitative computed tomography (QCT) technology offers many opportunities for the inves-
tigation of bone biomechanics because it determines some indicators of bone properties that are

0266/ ch26/frame Page 385 Friday, October 22, 1999 10:02 AM

386

Mechanical Testing of Bone and the BoneImplant Interface

relevant to bone strength, namely, the mass, the mechanical quality, and the spatial distribution
of bone material.

1-11

Peripheral QCT was developed as a small-eld, high-resolution extension of the existing QCT
systems,

12-15

to measure the peripheral skeleton with a substantial improvement of the image
denition. Currently available pQCT machines perform transverse scans of a wide size range of
regions of interest, from large body segments such as the whole human head, neck, or thigh, to
tiny excised bones such as mouse femurs or vertebrae. The gantry size of the machines limits the
size of the specimen or body segment to be studied.
Stratec/Norland (Pforzheim, Germany) produces the widest range of pQCT machines in the
market. These were developed from the original, XCT-960 model (the only pQCT device with FDA
approval by 1993

15,16

), which allowed a research option for studying relatively small bones. Some
of these models are especially adapted for that purpose. The XCT Research SA (SA refers to small
animals) is designed to study bones of animals ranging in size from sheep or primates to rats

in
vivo

or

in vitro

, and the high-resolution XCT Research M (M refers to mouse) is used to study
specimens ranging from mice to ferrets, even

in vivo

. A special achievement, the XCT Microscope,
was also developed for quasi-histomorphometric analysis of small excised bones at a very high
resolution, up to allowing the measurement of individual trabeculae. Scanco Medical AG (Bassers-
dorf, Switzerland) produces ultrahigh-resolution (microtomographic) machines

17,18

as the

-CT-20,
which is able to analyze two- and three-dimensional histomorphometrical images of trabecular
bone specimens but cannot determine bone density values. Some disadvantages of the high-
resolution equipment are the high cost and the long determination time. Table 26.1 shows some
distinctive characteristics of the equipment described.

II. WHAT IS A pQCT MACHINE?

A basic pQCT machine consists of two major components: a scanner unit and a control/analysis
computer system. The scanner unit contains (1) a source that emits a very narrow X-ray beam;
(2) a detector of the emitted radiation xed at a short distance from the source, which can measure
the intensity of the radiation and the attenuation produced by the tissues studied; and (3) a
mechanical system allowing radial, transverse, and axial displacements of the sourcedetector
couple in order to achieve different scanning positions of the bone (Figure 26.1A). The bone sample

TABLE 26.1
Major pQCTs on the Market

Machine and Sample Voxel Max SV Min Slice
Characteristics Diameter, mm Max Size,

m Path, mm Thickness

XCT 960, research option, good for large bones 88 90690 120 2.50 mm
XCT Research SA, formerly 960 A, wide size 90 90500 180 0.75 mm
thickness range, variable slice capacity, avoids
repositioning
XCT Research M, narrower, softer X-ray beam,
higher and controlled resolution
50 70500 180 0.55 mm
XCT Microscope, higher mechanical precision, 50 30300 180 0.130.50 mm
smaller sourcecenter distance, special for thin
cortical shells, quasi-histomorphometric scan,
three-dimensional network reconstruction
Scanco

-CT 20, ultrahigh denition, static histo- 17 814 50 30

m
morphometry allowed, three-dimensional network
reconstruction, no bone densitometry allowed

0266/ ch26/frame Page 386 Friday, October 22, 1999 10:02 AM

Peripheral Quantitative Computed Tomography

387

is centrally located between the source and the detector with the aid of special supports adapted
to

in vivo

or

in vitro

conditions (Figure 26.1B). As a rst step, when the measurement is started,
successive transverse displacements of the sourcedetector couple, repeated after small axial dis-
placements, produce a computed radiograph or scout-view of the bone piece along its longitudinal
axis, resembling a standard densitometric (DEXA) picture (Figure 26.1C). Reference points on the
screen allow selection of any convenient position along the bone axis. At that point, a series of
transverse measurements are then performed after successive, partial rotation displacements of the
couple until completion of a 180

excursion (translate/rotate mechanism).
The computer system controls the complete scanning procedure and integrates the information
obtained into a composed image of the bone slice, the tomographic scan that is shown on the
screen (Figure 26.2), and analyzes that scan as described below. The scan is integrated by the
computer system through a complex procedure known as backprojection with ltration. The
complete eld scanned is divided into a number of tiny square areal units (pixels). As the tomo-
graphic bone slice has a predetermined, constant thickness, the pixels actually represent the bases
of volume units which are more properly referred to as voxels. The number of voxels per eld
is a xed characteristic of each type of machine. The operator can change the size of the eld, but
not the number of voxels in the eld. Therefore, the smaller the eld selected, the smaller the
voxels and the greater the information contained per unit of eld area (i.e., the resolution of the
machine), and vice versa. For the earlier Stratec machines the voxel resolution was

of the
variable scanning diameter.

15

The new machines provide 1024

1024 voxels. Some improvements
such as automatic contextual segmentation algorithms to identify bone compartments

19

or post-
processing of the images in a separate workstation with a region grow and skeletonization step

20

may approach some histomorphometric assessments of bone structure.

FIGURE 26.1

(A) Translational/rotational displacements of a pQCT machine (XCT-960). (Courtesy of
Stratec Medizintechnik GmbH, Germany.) (B) Standard support and tube provided for positioning rodent
limbs and excised long bones for a pQCT measurement in the XCT-960 A. (C) Example of a pQCT scout-
view (human radius). (Courtesy of Stratec Medizintechnik GmbH, Germany.) (D) Didactical representation
of the meaning and calculation of the rectangular (related to

x

and

y

axes, relevant to bending analyses) and
polar versions (related to

z

axis, relevant to torsional analyses) of the cross-sectional moments of inertia of
a bone diaphysis (see Equation 26.2).

0266/ ch26/frame Page 387 Friday, October 22, 1999 10:02 AM

388

Mechanical Testing of Bone and the BoneImplant Interface

III. WHAT DOES pQCT MEASURE?

As an essentially absorptiometric technique, the pQCT measures only the attenuation of the radiation
passing through the whole tomographic slice. This is the information from which the scan is performed.
The magnitude of the attenuation depends on both the mass and the number of electrons in the atoms
of the elements present in the tomographic slice, so that bone minerals are most relevant for such
measurement. The attenuation is expressed in cm

1

units as the linear



attenuation coefcient (

, a
variable depending on the absorbing material) of the relationship between the intensities (

I

o

, I

) of the
emitted radiation before and after absorption, according to the equation

I

=

I

o



D

d

(26.1)
where

D

is the density of the absorber in g/cm

3

, and

d

is its layer thickness in cm.

FIGURE 26.2

Examples of pQCT scans of the femoral distal metaphyses from the immobilized (IM) or
overloaded legs (OL) of rats otherwise untreated or given anabolic doses of hPTH(1-38) for 75 days (P),
withdrawn with no other treatment (Pw), or sequenced by risedronate (PR,R) during further 90 days

30,31,40

obtained with an adapted XCT-960 machine.

0266/ ch26/frame Page 388 Friday, October 22, 1999 10:02 AM

Peripheral Quantitative Computed Tomography

389

The integration algorithm calculates the attenuation coefcient value corresponding to every par-
ticular voxel. This is a correlate of the amount (content) of absorbing matter in the voxel. However,
the machine is unable to assess the absolute mineral density value of each voxel as a direct mea-
surement. A special hydroxyapatite phantom is provided for this purpose and provides a particularly
stable reference that does not require calibration too frequently.

21

The phantom allows the machine to
transform the attenuation coefcients of every voxel into the volumetric mineral content (vBMC)
and volumetric mineral density (vBMD = vBMC/voxel volume) data that are provided as the out-
come. The machine is also able to assign different, default (rainbow scheme) or customizable voxel
colors to successive ranges of attenuation/density values. Thus, the integrated image of the bone slice
can be given a semiquantitative densitometric aspect (see Figure 26.2).
The machines are arranged for assigning automatically a zero attenuation/density value to every
voxel in the image corresponding to a measurement of pure fat.

22

Therefore, any above-zero
attenuation measured for a given voxel should represent a certain content of matter denser than the
fat it contains. Regardless of whether that corresponds to a soft or a hard (mineral-containing)
tissue, or a mixture of both, the automatic reference to the phantom scale forces the machine to
take and express that value as a mineral content. The energy of the radiation used is specically
selected in order to minimize the error involved in such estimations of mineral, as well as to
optimize the separation of bone from soft tissues. On summing up the individual values assigned
to each voxel, the machine can determine the mineral mass or vBMC and the vBMD of the whole-
bone slice or of a selected region of it. The vBMD represents the vBMC/bone volume relationship
of the region and is usually expressed in mg/cm

3

.
It is critical to interpret correctly what kind of bone density the machine assesses. No available
absorptiometric device can measure the true density of an ideal piece of pure solid, absolutely
pore-free bone substance (a hardly available specimen, if ever!). The established value of about
1.9 g of matter/cm

3

of solid tissue volume for such a material could only be estimated by calculation
from what is known about its fractional composition. Practically, assuming it to be 58% matrix +
42% mineral, the following approximate estimation can be derived.
Matrix density = 1.0 g/cm

3

0.58 = 0.58 g/cm

3

Mineral density = 3.2 g/cm

3

0.42 = 1.34 g/cm

3

Specic density of the solid bone matter = 1.92 g/cm

3

(matrix + mineral composite)
What is actually measurable in practice is the density of the whole bone including all hard
substance, cells, vessels, marrow, etc. This is known as the apparent or Archimedean bone density.
This concept is valid for the full range of bone porosity of either woven or lamellar, cortical or
trabecular bone tissue, and also applies to bones as organs or to whole skeletons. The pQCT
machines assess only this aspect of bone density in the bone slice. Moreover, they merely compare
the absorptiometric data to a phantom scale of volumetric mineral densities. For that reason their
outcome (vBMD) is expressed as a mineral mass per volume unit of whole-bone tissue.
Despite that limitation, the vBMD value could be regarded as a noninvasive indicator of the
mechanical quality (specic stiffness, or elastic modulus) of the solid bone tissue. In fact, it
represents at least one (the mineral amount per unit volume) of the most relevant determinants of
that property. In support of that assumption, a close correlation between the chemically assessed
vBMD (ash content per unit of bone volume) and the mechanically determined elastic modulus
of the solid (cortical) bone tissue has been veried repeatedly.

23-25

It has also been demonstrated that the pQCT-assessed vBMD is more representative of the
actual bone material quality than the areal bone mineral density (BMD) expressed as a mass

0266/ ch26/frame Page 389 Friday, October 22, 1999 10:02 AM

390

Mechanical Testing of Bone and the BoneImplant Interface

of mineral per (square) unit of projected bone area provided by DEXA.

1-9,26,27

A further advantage
of the pQCT-assessed vBMD values is that they are independent of the rotation of the bone
slice during the measurement, as well as (to some extent) of the bone size or shape. In addition,
the information provided by the pQCT determinations can be processed by the machine in order
to (1) distinguish between trabecular and cortical bone in many instances,

3,22,28,29

and (2) cal-
culate a number of variables which describe many aspects of bone architecture,

4-11,22,29-40

as
commented below.
The machine is blind concerning the type of tissue under measurement in each voxel because
the vBMD is calculated from the attenuation coefcient value of the whole voxel, which reects
its bulk mineral mass content. In other words, voxels with trabecular bone of relatively low porosity
and high true mineral density could not be distinguished from those containing cortical bone
with relatively high microporosity if both of them show a similar attenuation coefcient. This pitfall
gives rise to the so-called partial volume effect which is discussed below.
However, both the distinction between bone- and no-bone-containing voxels and the histo-
logical limitation between trabecular and cortical tissue are usually clear enough in practice to
overcome the above inconveniences. The menu of operator-dened modes provided by the
Special Analysis Software in the Norland/Stratec machines deals generally quite well with that
purpose, as described below.

IV. HOW DOES A pQCT MACHINE OPERATE?

Before starting the measurement, the operator must check the apparatus through the quality assur-
ance steps, and then adjust the measurement parameters such as bone length, voxel size, number
of steps in the scout-view scan, position of the reference line, number of sites to measure, and their
distances to the reference line. These procedures vary for the different machines.
Some recommendations have to be followed concerning the management of the samples. As
attenuation values for air are lower than for fat, it is convenient to avoid any interference from
the air that may surround or be inside the bone sample, which would produce negative density
values. Although this may not be relevant for

in vivo

or whole-limb studies, it is recommended
that operators (1) avoid measuring dried bones containing some air within the pores, as well as
(2) immerse the fresh bones into water in tubes (see Figure 26.1B) during any

in vitro

determi-
nation. Freezing the samples should not interfere the absorptiometric determinations, provided
that air was not allowed to go into the bone cavity. However, it is not advisable to freeze the
bones for long periods of time (say, more than 1 month) if investigators also aim to perform
mechanical testing on the same samples.
Positioning the sample is critical for the measurement. Some special devices as limb clamps
for

in vivo

studies (see Figure 26.1B) can be provided on request, but special supports or cradles
usually would have to be prepared by each laboratory according to the particular aim of the study.
The software helps to improve the reproducibility by detecting the same region of interest from
one measurement to the next by testing the number of voxels.

15,16

The reproducibility of the
determinations concerning repositioning should be assessed by calculating the corresponding coef-
cients of variation, for which quite low values have been reported.

26,27,41,42

Different eld/voxel sizes can be conveniently selected according to the bone studied in order
to optimize the resolution of the measurement. Ideally, the bone slice should t within the scanned
eld allowing just a minimal margin all around. Care should be taken to prevent the bone image
from reaching the margins of the eld; otherwise important sources of errors in the measurement
could be introduced.
The rst step of the determination is to obtain the scout-view, which is automatically made
by the machine, and to dene the scan sites on the screen. This can be arranged either symmet-
rically or asymmetrically concerning the starting reference line. However, there is a length
limitation for the scout analysis (see Table 26.1) that restricts the performance of multiple slices

0266/ ch26/frame Page 390 Friday, October 22, 1999 10:02 AM

Peripheral Quantitative Computed Tomography

391

in a single measurement. Following the scout-view, a single command initiates the whole CT
measurement automatically under screen control to detect the misplacement of the sample within
the eld. The image obtained is then displayed, ready for analysis (see Figure 26.2). Special
tools (Loop functions) help when performing the standardized or customized analysis of the
images. The most critical requisite for performing a good analysis (and hence that most subjected
to discussion and further elaboration in the future) is the image denition concerning both the
boundary between the bone section and the surrounding tissues and the one between the cortical
and trabecular bone regions.
The Special Analysis Software in the Norland/Stratec machines (reference is made here to the
newest available versions, 5.20 for XCT 960 A and 5.40 for XCT SA) allows customizing the
methods (Modes) for separating (1) the soft tissue from the outer edge of the bone (ContMode);
(2) the trabecular region from the cortical shell within the ContMode-dened bone to obtain a
trabecular and a cortical-subcortical value (PeelMode); and (3) the cortical bone from the trabe-
cular bone to obtain a cortical bone value (CortMode), as follows.

A. C

ONT

M

ODE

A preselected or user-dened attenuation threshold (the THBD threshold) causes the machine to
eliminate automatically any voxel with an attenuation value below that level. A 0.5 cm

1

THBD
value is recommended for that purpose (at this point it is possible to set the attenuation thresholds
directly in milligrams per cubic centimeter as a resource to standardize the procedure for the
different machines). As the machine automatically works centripetally in this mode until a complete
bone perimeter is dened from the outside, every soft tissue surrounding the bone is removed that
way (ContMode 1). For relatively large, regularly shaped bones, a couple of iterative contour-
detection procedures can also test the neighboring voxels with respect to either default or user-
dened thresholds. This algorithm can be additionally activated in order to optimize the selection
of a particular voxel as a boundary one (ContModes 2 and 3).

B. P

EEL

M

ODE

Working now from the outside bone edge in, this algorithm concentrically peels away an operator-
dened percent of the outside area of the bone (usually 55% for the human radius; different values
should be tried for animal bones). This default procedure (PeelMode 1) peels away (1) the outside
cortical shell; (2) an inner area that is part cortical and part trabecular; and (3) a small portion of
the inner area that is purely trabecular. It works well when the trabecular region is regularly shaped
and hence the remaining, inner region can be considered to be purely trabecular.
PeelMode 1 may be unsatisfactory otherwise, including when the trabecular bone area must
be analyzed as completely as possible. In these cases an operator-dened, inner attenuation
threshold (the THBD2 threshold) should be selected (proposedly, 0.63 cm

1

for rat bones) to
separate a trabecular (attenuation values lower than threshold) and a cortical-subcortical
region (values higher than threshold). A ltering process is used to ignore any high-attenuation
voxel isolated within the trabecular area as well as in areas that are not continuous (PeelMode
2). If the operator wants also to eliminate the subcortical region as completely as possible, an
additional peeling of the remaining area down to an indicated percentage can be ordered (PeelM-
ode 3). An additional percentage of peeling can also be performed as an attempt to eliminate
completely any high-density voxel inside the trabecular area which could inuence the trabecular
density (PeelMode 4).
When the cortical shell is well dened in the scan, the cortical/trabecular boundary can also
be iteratively dened by testing the maximal attenuation gradient between the successive voxels
working from the outside in, perpendicularly to the tangents to every voxel on the outer bone edge.
This way the trabecular/cortical boundary is dened as a continuous chain of selected, high-slope

0266/ ch26/frame Page 391 Friday, October 22, 1999 10:02 AM

392

Mechanical Testing of Bone and the BoneImplant Interface

voxels around a complete circle (PeelMode 5). This procedure works well for

in vivo

rat studies.
Additional peeling as in PeelModes 3 and 4 can also be commanded (PeelModes 6 and 7). These
modes may be useful when PeelMode 5 worked well for performing serial studies or if a low
resolution had to be employed, respectively.
Abnormalities such as a coupling of a high attenuation value and a low trabecular density, or
a high-density pocket within a low-trabecular-density region, can be identied by displaying the
highest attenuation coefcient in trabecular bone, which allows one to see the density shifts
(PeelMode 20).

C. C

ORT

M

ODE

The cortical shell can be separated by selecting an adequate threshold (the THCRT2 threshold;
a 0.93 cm

1

value is recommended for rat bones, but in skeletally immature animals younger
than 9 months it should be lowered to around 0.76 cm

1

). The machine will automatically
peel away every voxel showing an attenuation value below the selected one, i.e., the trabecular
bone region (CortMode 1). When working at low resolutions, the remaining voxels showing a
comparatively low attenuation with respect to their neighbors can also be additionally peeled
away in order to ignore any low-density point within the cortical area (CortMode 2). An iterative
contour detection algorithm, analogous to that employed in PeelMode 5 for the trabecular region,
can also be employed to dene the cortical shell from both the outer and inner sides (CortMode
3). This procedure may work well for lower resolutions such as those employed in

in vivo

studies.
The cortical region can also be dened after setting the outside bone edge by CortMode 1,
working in, selecting a maximal attenuation for the inner threshold rst and then lowering it
until the whole cortex is displayed (CortMode 4). In many instances, the selected modes succeed
in dening the trabecular and cortical regions of interest (ROIs), which are shown on the screen.
Then, special functions as CalcBD and CortBD yield the results of total, trabecular, cortical-
subcortical, and cortical vBMD (in mg/cm

3

) and the cross-sectional areas of the corresponding
bone portions (in mm

2

) as determined.
A two-color histogram represents the proportional distribution for total and trabecular voxels
(the latter are characterized as those featuring within the trabecular region dened by the
PeelMode) according to their attenuation values. The greater the accumulation of trabecular
voxels in the low-attenuation region of the histogram, the better the boundary denition
achieved between the trabecular and cortical ROIs. By modifying the selected thresholds the
operator can improve that denition up to a certain limit. Bone images could, hence, be classied
as apparently well or poorly dened concerning the sharpness (or the reliability) of the
trabecular/cortical boundary. The standard analyses of cortical and trabecular bone provided by
the machines obviously refer to the rst case.
Nevertheless, whatever the method employed for dening the cortical and trabecular regions,
the outcome will always be affected by the partial volume effect (PVE).

22,43

The PVE is a source
of error in the determination of the bone areas derived from the unavoidable inclusion of voxels
that are not lled with mineralized tissue. It derives from the blindness of the machine as
commented above and may lead to underestimation of the regional vBMD measurements that may
exceed 15%. By optimizing the above modes, as well as by selecting the smallest voxel sizes
available, the PVE can be minimized, although it raises a severe limitation for the analyses of
relatively thin cortical shells unless the denition achieved was actually very high.
The only way to test the PVE interference in the interpretation of the pQCT data is by measuring
bones of the same kind but of different size and cortical thickness. On plotting the assessed cortical
vBMD data (

y

) against the cortical thickness (

x

), a constant (horizontal) relationship for the greater
values of the latter will be observed; however, toward the lowest end of the range the points should
tend to shift to the lower-left region of the graph. This would point out the existence of the PVE
and allow measurement of its relative signicance.

0266/ ch26/frame Page 392 Friday, October 22, 1999 10:02 AM

Peripheral Quantitative Computed Tomography

393

V. WHAT KIND OF VARIABLES AND BONE PROPERTIES
ARE MEASURED BY pQCT?

Many bone variables can be measured by pQCT in different long bones from small animals, principally
in the midshafts (in order to correlate the data with those of bending tests

37

) in the metaphyseal regions
(where trabecular bone and remodeling are present in most species

44

), in femoral necks, and, tentatively,
in vertebral bodies and hemimandibles, too. The available data can be classied as follows according
to their relevance as indicators of the different aspects of bone mechanical quality, namely, mass,
apparent density, architectural design, and structural stiffness and strength.

A. I

NDICATORS



OF

B

ONE

M

ASS

The vBMC of the trabecular, cortical, and total bone regions (0.6 to 1.9%

in vitro

and

in vivo

precision, 1.7 to 3.2% with repositioning with the XCT 960 A), as well as the cortical bone area
(1.3 to 2.4%, and average 3.8%, respectively), reect the amount of mineralized tissue in the
corresponding parts of the bone section.

17,45

Close correlations have been found between the pQCT
and histomorphometrically assessed values of cortical area and other variables in rat tibiae and
femorae.

40,42

These indicators, especially those for cortical bone, should be regarded as relevant to
bone strain and structural stiffness and strength in longitudinal compression.

46-52

The trabecular
bone area (as arbitrarily dened by the above modes) is not directly suitable for a proper bio-
mechanical estimation of bone quality. The total bone area (calculated as the whole solid section
area within the outer edge of the bone, regardless of the inner tissue structure) should only provide
information on bone size.

B. I

NDICATORS



OF

B

ONE

A

PPARENT

D

ENSITY

The vBMD of the



trabecular region can be measured very precisely (0.4 to 0.6% in rats

in vitro

and

in vivo

, 2.9% with repositioning

8

) and with a high statistical power, especially at the distal-
femoral and proximal-tibial metaphyses. Changes can be detected within a few weeks after gona-
dectomy or treatment with raloxifene or PTH in small groups of rodents, much sooner than when
employing DEXA.

8,9,27,28,41,53-55

Rather than to represent a true material property, the trabecular
vBMD should be regarded as a correlate of the structural stiffness and strength of the trabecular
network, i.e., a mechanical quality indicator expressed at the tissue level of complexity.

56,57

The vBMD of the cortical region (0.6 to 1.9% precision

in vitro

and

in vivo

, 3.3% with
repositioning

8,10,53

), represents the true vBMD of the solid bone substance. Therefore, it may be
regarded as indicative of one of the chief determinants of the intrinsic stiffness (elastic modulus)
of the solid bone tissue.

7,23-25

Moreover, it has been shown to correlate well with the bone-breaking
force in femoral diaphyses and necks of rats and mice and in the human radius.

11,29,51
However,
especially when measuring thin cortices at low resolutions, PVE-derived errors in the cortical area
measurements (not in the vBMC determination) may lead to underestimation of the vBMD.
22,41,51,58
High-resolution techniques should overcome this problem.
17
Changes in the cortical vBMD are
much slower than those in the trabecular density.
8
No directionality could be ascribed to these qualitative indicators because bone material or
structural anisotropy is disregarded by the absorptiometric determinations. The total vBMD of the
slice (a combination of the trabecular and cortical vBMDs) should not be an indicator of any
particular bone mechanical property, analogously to the areal BMD provided by DEXA.
20
C. INDICATORS OF BONE ARCHITECTURAL QUALITY
The equatorial and polar second moments of inertia of the cross-sectional bone area (xCSMI,
pCSMI; CV = 6.4% with the XCT-960 A in mice) are relevant to long-bone stiffness and strength
0266/ ch26/frame Page 393 Friday, October 22, 1999 10:02 AM
394 Mechanical Testing of Bone and the BoneImplant Interface
in bending and torsion, respectively.
1-7,29-37,39,47,52,59,60
The Norland/Stratec machines calculate the
CSMIs of the total or cortical bone areas as
CSMI = (A
i
d
i
2
) (26.2)
where A
i
is the area of an individual voxel within the bone section and d
i
is the distance from the
center of that area to the reference, bending (x, y) or torsion (z) axis (see Figure 26.1D). As seen
from the formula, the CSMI values (given in mm
4
) increase linearly with bone mass (A), but are
also exponentially proportional to the distance (d
2
) of the bone material from the reference axis.
Therefore, the equatorial and polar CSMIs are true indicators of the bending or torsional stiff-
ness/strength of the assayed long bone, respectively, regardless of the bone material quality.
7,10,11
The machines can also measure the cross-sectional diameters, endosteal and periosteal perim-
eters, and average cortical thickness. These variables may help to evaluate the modeling-derived
changes provoked by growth or by anabolic or anticatabolic treatments.
8
However, they do not
affect the quality of the architectural design or the mechanical competence of long bones as much
as do the CSMIs.
1-3,7
D. INDICATORS OF WHOLE-BONE QUALITY
The mechanical properties of the whole bones as measured by pQCT exceed the possibilities of
standard densitometry.
5-8,48
The structural stiffness and strength of long hollow tubular structures
are generally proportional to the product CSMI E (E being the elastic modulus of the material
of which the structure is made
61
). The elastic modulus can only be assessed mechanically, but it
could be reasonably estimated by the apparent mineral density of the solid bone.
23-25
Hence, on
replacing E by the pQCT-assessed vBMD of the cortical bone in Equation 26.2, the authors group
developed the so-called bone strength index (BSI) as the product
BSI = xCSMI cortical vBMD (26.3)
The authors group has shown that this BSI correlates strongly with the actual three-point
bending strength (breaking force) in a large sample of rat femur shafts, regardless of bone size and
experimental conditions (Figure 26.3).
5
Conversely, weak correlations were observed between the
breaking force and the DEXA-assessed, areal BMD of the central diaphyseal region of the same
bones. This is evidence of the greater ability of the tomographic BSI to describe the bone strength
in the assayed conditions. Besides this BSI (xBSI), the pQCT machines also calculate its bending
y and torsion versions (yBSI, pBSI, for which a mechanical validation is still needed), by using
the yCSMI or the pCSMI (see Figure 26.1D) instead of the xCSMI in Equation 26.3, respectively.
The BSI concept should not be freely generalized to the analysis of any bone region or method
of deformation. Specially adapted formulae should be derived in order to achieve further, suitable
BSIs for every method of bone deformation applied to every skeletal region of interest.
5,11
The
Norland/Stratec machines provide also another kind of BSI for long bones, the stress/strain index
(SSI),
62,63
calculated as
(26.4)
where d
Mx
is the maximal distance from a voxel to the polar (z) axis in the image, and vBMD
Mx
is the maximal value the cortical vBMD could theoretically assume (i.e., 1200 mg/cm
3
). This SSI
was proposed to reect the long-bone strength more generally than the above BSIs and as such it
should be validated in future investigations. These BSIs or SSIs do not take into account any other
SSI =
pCSMI cortical vBMD
vBMD
Mx Mx

i
d
0266/ ch26/frame Page 394 Friday, October 22, 1999 10:02 AM
Peripheral Quantitative Computed Tomography 395
relevant factors to bone material quality. They ignore the many microstructural determinants,
including fatigue damage, that may also affect the mechanical ability of the bone tissue. Those
indexes may be useful, however, provided that these factors can be assumed to remain unaffected
by the assayed treatments.
VI. HOW CAN THE RAW INFORMATION
PROVIDED BY pQCT BE APPLIED?
The standard information provided by pQCT studies in animals may be used for descriptive pur-
poses.
64
However, it can also be used for other purposes,
4,6,7,45,60
as the following examples suggest.
A. ANALYSIS OF THE PATHOGENESIS OF SOME EXPERIMENTAL EFFECTS ON BONE BIOMECHANICS
Data from separate pQCT evaluations of trabecular and cortical bone and the material and architectural
properties of bones can be correlated with those from biomechanical studies in order to investigate
the pathogenesis of any change in the whole-bone quality. Figure 26.4, upper, shows the ovariectomy
(OX)-induced impairment and the alendronate-induced protection of the bending strength in rat
FIGURE 26.3 Upper: Close, linear correlation between a BSI (x) assessed from pQCT scans of the midshafts
and the actual fracture load in bending (y) of 206 femurs from rats of different ages and sizes,
5
treated with
dexamethasone
8,36
or aluminum hydroxide,
65
or studied as controls. Lower: Lack of correlation of the DEXA-
assessed BMD of the central diaphyses with the same indicator of bone strength of the same bones. (From
Feretti et al., Bone, 18, 2, 97, 1996. Reproduced by permission of Elsevier Science.)
0266/ ch26/frame Page 395 Friday, October 22, 1999 10:02 AM
396 Mechanical Testing of Bone and the BoneImplant Interface
FIGURE 26.4 Upper: Effects of peripubertal ovariectomy alone (OX) or immediately followed by 5 or
25 g/kg sc 2/week of alendronate for 6 months on the diaphyseal strength (fracture load) and architecture
(body-weight adjusted CSMI, Ix) and the material quality (calculated elastic modulus, E) and vBMD of cortical
bone of rat femurs.
65,66
Asterisks express the statistical signicance of the intergroup differences. Lower:
Correlation between the breaking force and the pQCT-assessed BSI of the bones from the OX- and 25-g-
alendronate-treated animals of the same experiment. Values for sham controls are represented by their 95% C.I.
0266/ ch26/frame Page 396 Friday, October 22, 1999 10:02 AM
Peripheral Quantitative Computed Tomography 397
femurs.
65,66
This effect must have resulted from changes in cortical material quality or in the diaphyseal
architecture, or both. The lack of changes in the pQCT-assessed diaphyseal CSMIs in any group ruled
out bone architecture as a source of such an effect. Parallel changes in the elastic modulus of the
cortical tissue pointed to bone material quality as the cause of the observed variation in the whole-
bone quality. However, the cortical vBMD did not vary between groups. Therefore, the changes in
bone material quality must be ascribed to effects on some of the mineralization-unrelated, microstruc-
tural components of bone tissue that are known to affect bone material quality.
In support of that conclusion, the correlation between the measured breaking force and the
calculated BSI of the same bones showed that bones from both the OX and risedronate-protected
rats varied in opposite directions from the natural association between these variables (Figure 26.4,
lower). The reason for that behavior should be a change in some factor(s) relevant to bone material
quality that the BSI calculation disregarded (i.e., unrelated to bone mineralization). Findings like
this may help to explain some differences between drug effects on fracture incidence and on the
DEXA-assessed areal BMD in human studies.
B. ANALYSIS OF THE MECHANOSTAT STATUS BY MEANS OF DISTRIBUTION/QUALITY CURVES
An opposite behavior of parameters of material quality (E) and architectural cross-sectional design
(CSMIs) of femur shafts from growing rats of two different lines have been described.
1,3
This
reected a normalization of the bone strength to the body weight of the animals, due to a feedback
regulation of diaphyseal modeling by a function of the bone strain history (Frosts mechanostat
theory).
7,45,67-73
The negative hyperbolic functions describing that relationship were called distri-
bution/quality curves.
4,6,7
It was also found that the same interrelationships could be shown by
plotting the CSMIs vs. the vBMD of the cortical bone as a correlate of the mechanical indicator,
E, in many instances.
1-3,29-36,65,66,74,75
The effects of some treatments on the control of bone quality by the mechanostat can be described
by such graphs in a useful way. On the one hand, displacements of the points along the curve showing
no departure from the normal relationship would reect an indemnity of the mechanostatic control.
On the other, any shift of the data to the upper-right or to the lower-left of the graph should indicate
an anabolic (or anticatabolic) or a catabolic (or antianabolic) shift of the mechanostat set point,
4,6,7
respectively. In this way the effects of many treatments on rat bones have been
described.
29-36,39,65,66,74,75
As examples, a negative interaction of dexamethasone and a positive inu-
ence of anabolic PTH alone or combined with risedronate on bone mechanostat are summarized.
Dexamethasone administration to growing rats reduced all material (vCtBMD), architectural
(CSMIs), and mechanical properties (breaking force) of femur shafts in a dose-dependent fash-
ion.
36,39
Figure 26.5a shows the antianabolic shift induced to the bone mechanostat set point. A
three-dimensional representation (Figure 26.5b) shows the combined, dose-dependent, negative
inuence of changes in both vCtBMD and CSMI on bone strength.
Low, intermittent doses of hPTH(1-38) given for 75 days to rats with a right hind limb
immobilization and a mechanical overloading of the other leg enhanced all femur CSMI, vCtBMD,
and bending breaking strength.
30,31,40
The distribution/quality curves (Figure 26.5c) showed that
these effects (1) reected an anabolic interaction with the mechanostat set point; (2) were of a
transient nature; (3) were maintained by a sequential administration of a remodeling inhibitor, the
bisphosphonate olpadronate (an anti-catabolic interaction); and (4) were potentiated by the mechan-
ical overload.
C. ANALYSIS OF THE ARCHITECTURAL EFFICIENCY OF BONE MATERIAL DISTRIBUTION
WITHIN BONE DIAPHYSES
This interesting feature of bone modeling can be assessed by plotting any of the CSMIs (y) vs. the
cortical area of the same scan (x; Figure 26.6a). The higher the slope of the correlation, the better
0266/ ch26/frame Page 397 Friday, October 22, 1999 10:02 AM
398 Mechanical Testing of Bone and the BoneImplant Interface
F
I
G
U
R
E

2
6
.
5
(
a
)

D
i
s
t
r
i
b
u
t
i
o
n
/
q
u
a
l
i
t
y

g
r
a
p
h

s
h
o
w
i
n
g

t
h
e

a
n
t
i
a
n
a
b
o
l
i
c

e
f
f
e
c
t

o
f

d
e
x
a
m
e
t
h
a
s
o
n
e

o
n

r
a
t

f
e
m
u
r
s
.
2
9
,
3
9

C
o
n
t
r
o
l

a
n
i
m
a
l
s

a
r
e

r
e
p
r
e
s
e
n
t
e
d

b
y

t
h
e

9
5
%

C
.
I
.

o
f

t
h
e
d
a
t
a
.

(
b
)

A

t
h
r
e
e
-
d
i
m
e
n
s
i
o
n
a
l

r
e
p
r
e
s
e
n
t
a
t
i
o
n

o
f

t
h
e

c
o
m
b
i
n
e
d

i
m
p
a
c
t

o
f

c
h
a
n
g
e
s

i
n

b
o
n
e

a
r
c
h
i
t
e
c
t
u
r
e

(
x
C
S
M
I
)

a
n
d

c
o
r
t
i
c
a
l

m
a
t
e
r
i
a
l

q
u
a
l
i
t
y

(
a
s

a
s
s
e
s
s
e
d

b
y

t
h
e

v
B
M
D
)
i
n

t
h
e

s
a
m
e

b
o
n
e
s
.

(
F
r
o
m

F
e
r
r
e
t
t
i

e
t

a
l
.
,

B
o
n
e

1
6
,

1
,

1
1
9
,

1
9
9
5
.

R
e
p
r
o
d
u
c
e
d

b
y

p
e
r
m
i
s
s
i
o
n

o
f

E
l
s
e
v
i
e
r

S
c
i
e
n
c
e
.
)

(
c
)

D
i
s
t
r
i
b
u
t
i
o
n
/
q
u
a
l
i
t
y

c
u
r
v
e
s

s
h
o
w
i
n
g

t
h
e

t
r
a
n
s
i
e
n
t
,
a
n
a
b
o
l
i
c

i
n
t
e
r
a
c
t
i
o
n

o
f

h
P
T
H

w
i
t
h

t
h
e

m
e
c
h
a
n
o
s
t
a
t
i
c

c
o
n
t
r
o
l

o
f

b
o
n
e

a
r
c
h
i
t
e
c
t
u
r
e

(
u
p
p
e
r
)

a
n
d

t
h
e

m
a
i
n
t
a
i
n
a
n
c
e

o
f

t
h
a
t

e
f
f
e
c
t

b
y

t
h
e

s
e
q
u
e
n
t
i
a
l

a
d
m
i
n
i
s
t
r
a
t
i
o
n

o
f

r
i
s
e
d
r
o
n
a
t
e
(
l
o
w
e
r
)

a
s

d
e
s
c
r
i
b
e
d

i
n

F
i
g
u
r
e

2
6
.
2
.
3
0
,
3
1
,
4
0

D
i
f
f
e
r
e
n
c
e
s

b
e
t
w
e
e
n

O
L

a
n
d

I
M

l
e
g
s

r
e

e
c
t

t
h
e

p
o
s
i
t
i
v
e

i
n
t
e
r
a
c
t
i
o
n

o
f

t
h
e

m
e
c
h
a
n
i
c
a
l

u
s
a
g
e

w
i
t
h

t
h
o
s
e

e
f
f
e
c
t
s
.
0266/ ch26/frame Page 398 Friday, October 22, 1999 10:02 AM
Peripheral Quantitative Computed Tomography 399
F
I
G
U
R
E

2
6
.
6
(
a
)

S
c
h
e
m
a
t
i
c

r
e
p
r
e
s
e
n
t
a
t
i
o
n

o
f

t
h
e

a
s
s
e
s
s
m
e
n
t

o
f

d
i
f
f
e
r
e
n
c
e
s

i
n

t
h
e

e
f
f
i
c
i
e
n
c
y

o
f

t
h
e

d
i
a
p
h
y
s
e
a
l

a
r
c
h
i
t
e
c
t
u
r
a
l

d
e
s
i
g
n

(
C
S
M
I
)

p
e
r

u
n
i
t

o
f

a
v
a
i
l
a
b
l
e

c
o
r
t
i
c
a
l
m
a
t
e
r
i
a
l
.
3
8

T
h
e

h
i
g
h
e
r

t
h
e

s
l
o
p
e

o
f

t
h
e

c
o
r
r
e
l
a
t
i
o
n
,

t
h
e

g
r
e
a
t
e
r

t
h
e

e
f

c
i
e
n
c
y

o
f

b
o
n
e

m
e
c
h
a
n
o
s
t
a
t

t
o

s
t
i
m
u
l
a
t
e

a
n
d

o
r
i
e
n
t

b
o
n
e

m
o
d
e
l
i
n
g
.

(
b
)

D
i
s
t
a
l
-
m
e
t
a
p
h
y
s
e
a
l

s
c
a
n
s

o
f
t
h
e

f
e
m
u
r
s

f
r
o
m

h
y
p
e
r
c
a
l
c
e
m
i
c
,

t
u
m
o
r
-
t
r
a
n
s
p
l
a
n
t
e
d

m
i
c
e

u
n
t
r
e
a
t
e
d

(
l
e
f
t
)

o
r

t
r
e
a
t
e
d

w
i
t
h

o
l
p
a
d
r
o
n
a
t
e

(
r
i
g
h
t
)

o
b
t
a
i
n
e
d

w
i
t
h

a
n

X
C
T
-
9
6
0

A

s
h
o
w
i
n
g

t
h
e

i
m
p
o
s
s
i
b
l
e

d
i
s
t
i
n
c
t
i
o
n
b
e
t
w
e
e
n

t
r
a
b
e
c
u
l
a
r

a
n
d

c
o
r
t
i
c
a
l

b
o
n
e
.
7
6

(
c
)

D
i
s
t
r
i
b
u
t
i
o
n

c
u
r
v
e
s

o
f

t
h
e

c
o
u
n
t
e
d

v
o
x
e
l
s

o
f

c
r
o
s
s
-
s
e
c
t
i
o
n
a
l

a
r
e
a

(
C
S
A
)
.

(
d
)

t
h
e

v
B
M
C

o
f

t
h
e

a
n
i
m
a
l
s

f
r
o
m

t
h
e

s
a
m
e

s
t
u
d
y
,
t
h
r
o
u
g
h
o
u
t

t
h
e

a
v
a
i
l
a
b
l
e

r
a
n
g
e

o
f

T
H
B
D
2

t
h
r
e
s
h
o
l
d
s
.
0266/ ch26/frame Page 399 Friday, October 22, 1999 10:02 AM
400 Mechanical Testing of Bone and the BoneImplant Interface
the architectural design that the same bone mass achieved. This method allowed a description of
gender-related differences in the design of human bones
38
and is most suitable for analyzing the
interactions of the mechanical usage of the limb with the effects of treatments on bone architecture
and strength in animal models.
D. ANALYSIS OF THRESHOLD-DEFINED ROIS WHEN CORTICAL AND
TRABECULAR BONE ARE INDISTINGUISHABLE
When the distinction between cortical and trabecular bone is difcult but still possible, one should
use the function Loop, working rst at a xed THBD, ContMode and CortMode and varying the
THBD2 to measure the trabecular bone. A second Loop should then be made at a xed THBD,
ContMode, THBD2, and Cortmode and varying the THCRT2 to measure the cortical bone. The
software allows changing the ROI and the methods of analysis of the images as desired while
working with a given image. However, the parameters selected to perform the image itself cannot
be changed a posteriori, so one has to be careful when setting the conditions at which the image
is made. On following the above procedure, when a treatment blunts any distinction between cortical
and trabecular bone, its effects can be properly analyzed by displaying the changes in the bone
variables (y) along the whole range of attenuation thresholds (x) at which the machine is able to
work, so the effects are described as differences between the distribution curves of the assessed
variables in the threshold-dened ROIs. The resulting type of bone is thus dened by the
vBMD of the corresponding ROI, and the effects can be described according to the THBD2
threshold range at which they were most evident in each group.
As an example, the authors group failed to dene the olpadronate-induced protection of
metaphyseal bone of tumor-implanted, hypercalcemic mice as related to pQCT-assessed changes
in trabecular tissue by performing the traditional pQCT measurements in PeelMode 1
(Figure 26.6b). However, on varying the THBD2 threshold (PeelMode 2; Figure 26.6c) it was
possible to show that the differences between the tumor- and olpadronate-induced effects varied
widely,
76
so one could conclude that (a) the tumor reduced the bone mass and made the cortical
bone undetectable, and (b) olpadronate increased bone mass above normal values by protecting
that trabecular bone from remodeling. Those conclusions could not have been derived from
standard pQCT determinations at a xed THBD2 threshold. This novel pQCT application is
proposed as a useful tool in skeletal research, avoiding many false negative results.
E. ANALYSIS OF MUSCLE/BONE INTERRELATIONSHIPS
The ability of pQCT machines to measure also the cross-sectional muscle areas allows evaluation of
some muscle/bone interrelationships that are essential for assessing the state of bone mechanostat in
different experimental conditions. In human studies, a close, linear relationship has been shown
between bone CSMIs or BSIs (y) and the force or the pQCT-assessed, cross-sectional area of the
regional muscles (x) in normal men and women. The slope of that relationship changed signicantly
after menopause (Figure 26.7, top).
45,60,62,63,77
This offers the basis for distinguishing between disuse
osteopenias (in which the mechanostat is still working properly) and true osteoporosis (in which the
mechanostat set point is offset, most commonly because of an endocrine disorder; Figure 26.7,
bottom).
60,67,68
The eld is open for a wide variety of animal studies concerning this attractive proposal.
VII. CONCLUDING REMARKS
Many new opportunities are open to future research that uses pQCT in animal models. pQCT can
provide exclusive information concerning many biomechanical aspects of bone that are highly
specic to the skeletal region studied. For that reason, special recommendations should be followed
in future studies. They should avoid the understandable temptation to (1) extrapolate the pQCT
0266/ ch26/frame Page 400 Friday, October 22, 1999 10:02 AM
Peripheral Quantitative Computed Tomography 401
FIGURE 26.7 Top: Muscle/bone interrelationships as described in humans by the correlation between the
tibial SSI and the cross-sectional muscle area of the calf of normal males and premenopausal females.
77
Center:
The same data from postmenopausal females depart signicantly from the above correlation. Bottom: Sche-
matic representation of zones of normal mechanostasis (i.e., muscle/bone interrelationships under normal
control by bone mechanostat) and biomechanical incompetence (because of a shift in the bone mechanostat
set point), as derived from the evidence shown in the analogous graphs at the left. According to Frosts
principles,
78
this proposal (which can be tested employing animal models) should help to achieve a tomographic
differentiation between osteopenias and osteoporoses.
0266/ ch26/frame Page 401 Friday, October 22, 1999 10:02 AM
402 Mechanical Testing of Bone and the BoneImplant Interface
data of a given skeletal region to a different site and (2) compare the pQCT performance with those
of other techniques that measure different aspects of bone mass as the standard densitometry
does especially if these are inadequately taken as gold standards.
REFERENCES
1. Ferretti, J.L., Spiaggi, E., Capozza, R., et al., Interrelationships between geometric and mechanical
properties of long bones from three rodent species with very different biomass. Phylogenetic impli-
cations, J. Bone Miner. Res., 7(S2), S423, 1992.
2. Ferretti, J.L., Capozza, R., Mondelo, N., and Zanchetta, J., Interretionships between densitometrical,
geometric and mechanical properties of rat femurs. Inferences concerning mechanical regulation of
bone modeling, J. Bone Miner. Res., 8, 1389, 1993.
3. Ferretti, J.L., Capozza, R., Mondelo, N., et al., Determination of femur structural properties by
geometric and material properties as a function of body weight in rats. Evidence of a sexual dimor-
phism, Bone, 14, 265, 1993.
4. Ferretti, J.L., Perspectives of pQCT technology associated to biomechanical studies in skeletal research
employing rat models. Bone, 17(4S), 353S, 1995.
5. Ferretti, J.L., Capozza, R., and Zanchetta, J., Mechanical validation of a tomographic (pQCT) index
for the noninvasive assessment of rat femur bending strength, Bone, 18, 97, 1996.
6. Ferretti, J.L., Noninvasive assessment of bone architecture and biomechanical properties in animals
and humans employing pQCT technology, J. Jpn. Soc. Bone Morphom., 7, 115, 1997.
7. Ferretti, J.L., Biomechanical properties of bone, in Bone Densitometry and Osteoporosis, Genant,
H.K., Guglielmi, G., and Jergas, M., Eds., Springer, Berlin, 1998, 143.
8. Gasser, J.A., Assessing bone quantity by pQCT, Bone, 17, 145S, 1996.
9. Gasser, J.A., Quantitative assessment of bone mass and geometry by pQCT in rats in vivo and site
specicity of changes at different skeletal sites, J. Jpn. Soc. Bone Morphom., 7, 107, 1997.
10. Jms, T., Jalovaara, P., Peng, Z., et al., Comparison of three-point bending test and peripheral
quantitative computed tomography analysis in the evaluation of the strength of mouse femur and tibia,
Bone, 23, 155, 1998.
11. Jms, T., Tuukkanen, J., and Jalovaara, P., Femoral neck strength of mouse in two loading congu-
rations: method evaluation and fracture characteristics, J. Biomech., 31, 723, 1998.
12. Brner, W., Grehn, S., and Moll, E., Messung der Absorption des Fingerknochens mit einem
I
125
-Proscanner. Quantitative Methode zur Erkennung der Osteoporose, Fortschr. Rntgenstr., 110,
378, 1969.
13. Regsegger, P., Elsasser, U., Anliker, M., et al., Quantication of bone mineralization using computed
tomography, Radiology, 121, 93, 1976.
14. Schneider, P. and Berger, P., Bone density determination using quantitative computed tomography
and a special purpose scanner, Nuklearmediziner, 2, 145, 1988.
15. Schneider, P. and Reiners, C., Peripheral quantitative computed tomography, in Bone Densitometry
and Osteoporosis, Genant, H.K., Guglielmi, G., and Jergas, M., Eds., Springer, Berlin, 1998, 349.
16. Guglielmi, G., Schneider, P., Lang, T., et al., Quantitative computed tomography at the axial and
peripheral skeleton, Eur. Radiol., 7(S2), S332, 1997.
17. Kinney, J.H., Lane, N., and Haupt, D., in vivo three-dimensional microscopy of trabecular bone,
J. Bone Miner. Res., 10, 264, 1995.
18. Mller, R., Hildebrand, T., and Regsegger, P., Non-invasive bone biopsy: a new method to analyse
and display the three-dimensional structure of trabecular bone, Phys. Med. Biol., 39, 145, 1994.
19. Helterbrand, J.D., Higgs, R., Iversen, P., et al., Application of automatic image segmentation to tibiae
and vertebrae from ovariectomized rats, Bone, 21, 401, 1997.
20. Gordon, C.L., Webber, C., Adachi, J., and Christoforou, N., in vivo assessment of trabecular bone
structure at the distal radius from high-resolution computed tomography images, Phys. Med. Biol.,
41, 495, 1996.
21. Regsegger, P. and Kalender, W., A phantom for standardization and quality control in peripheral
bone measurements by pQCT and DEXA, Phys. Med. Biol., 38, 1963, 1993.
0266/ ch26/frame Page 402 Friday, October 22, 1999 10:02 AM
Peripheral Quantitative Computed Tomography 403
22. Augat, P., Gordon, C.L., Lang, T., et al., Accuracy of cortical and trabecular bone measurements with
peripheral quantitative computed tomography (pQCT), Phys. Med. Biol., 43, 2873, 1998.
23. Burstein A. H., Zika, J., Heiple, K., and Klein, L., Contribution of collagen and mineral to the elastic-
plastic properties of bone, J. Bone Joint Surg., 57A, 956, 1975.
24. Currey, J.D., The mechanical consequences of variation in the mineral contents of bone, J. Biomech.,
2, 1, 1969.
25. Currey, J.D., The effect of porosity and mineral content on the Youngs modulus of elasticity of
compact bone, J. Biomech., 21, 131, 1988.
26. Sato, M., Comparative X-ray densitometry of bones from ovariectomized rats, Bone, 17(4S), 157S,
1995.
27. Sato, M., Kim, J., Short, L., et al., Longitudinal and cross-sectional analyses of raloxifene effects on
tibiae from ovariectomized aged rats, J. Pharmacol. Exp. Ther., 272, 1252, 1995.
28. Breen, S.A., Millest, A., Loveday, B., et al., Regional analysis of bone mineral density in the distal
femur and proximal tibia using peripheral quantitative computed tomography in the rat in vivo, Calcif.
Tissue Int., 58, 449, 1996.
29. Ferretti, J.L, Gaffuri, O., Capozza, R., et al., Dexamethasone effects on structural, geometric and
material properties of rat femur diaphyses as described by peripheral quantitative computerized
tomography (pQCT) and bending tests, Bone, 16, 119, 1995.
30. Capozza, R.F., Ferretti, J.L., Ma, Y.F., et al., Tomographic (pQCT) and biomechanical effects of
hPTH(1-38) on chronically immobilized or overloaded rat femurs, Bone, 17(4S), S233, 1995.
31. Capozza, R.F., Aplicacin de Recursos Originales Biomecnicos, Radio-densitomtricos e Histomor-
fomtricos a la Evaluacin de los Efectos seos de la Parathormona, Ph.D. thesis, University of
Buenos Aires, Buenos Aires, 1998.
32. Cointry, G.R., Mondelo, N., Zanchetta, J., et al., Intravenous olpadronate restores ovariectomy-affected
bone strength. A mechanical, densitometric and tomographic (pQCT) study, Bone, 17(4S), S373, 1995.
33. Cointry, G.R., Estudio de la Curva Completa de Efectos de Bisfosfonatos de Tercera Generacin sobre
la Biomecnica, Densitometra e Histomorfometra en Huesos de Rata, Ph.D. thesis, University of
Buenos Aires, Buenos Aires, 1997.
34. Ferretti, J.L., Lin, B., Ke, H., et al., Geometric, material, structural and histomorphometric properties
of femurs and tibiae of ovariectomized rats as affected by leg immobilization or overloading, Calcif.
Tissue Int., 54, 349, 1994.
35. Ferretti, J.L., Mondelo, N., Peluffo, V., et al., Sub-chronic effects of high doses of mildronate on
femur densitometric (DEXA), tomographic (pQCT) and mechanical properties in young rats, Bone
Miner., 25(S2), S12, 1994.
36. Ferretti, J.L., Mondelo, N., Capozza, R., et al., Effects of large doses of olpadronate (dimethyl-
pamidronate) on mineral density, cross-sectional architecture, and mechanical properties of rat femurs,
Bone, 16(4S), 285S, 1995.
37. Ferretti, J.L., Frost, H. M., Gasser, J., et al., Perspectives in osteoporosis research: its focus and some
insights from a new paradigm, Calcif. Tissue Int., 57, 399, 1995.
38. Ferretti, J.L., Zanchetta, J., Capozza, R., et al., The premenopausal accumulation of bone material per
unit muscle mass in women would be of little mechanical relevance, Osteoporosis Int., 8(S3), 37, 1998.
39. Gaffuri, O.H., Desarrollo y Aplicacin de Recursos Biomecnicos, Densitomtricos e Histomorfo-
mtricos a la Evaluacin Farmacolgica del Efecto de los Glucocorticoides de Uso Clnico Corriente,
Ph. D. thesis, University of Buenos Aires, Buenos Aires, 1994.
40. Ma, Y.F., Ferretti, J.L., Capozza, R., et al., Effects of on/off anabolic hPTH and remodeling inhibitors
on metaphyseal bone of immobilized rat femurs. Tomographical (pQCT) description and correlation
with histomorphometric changes in tibial cancellous bone, Bone, 17(4S), S321, 1995.
41. Breen, S.A., Loveday, B., Millest, A., and Waterton, J., Stimulation and inhibition of bone formation:
use of peripheral quantitative computed tomography in the mouse in vivo, Lab. Anim., 32, 467, 1998.
42. Mhlbauer, R.C., Schenk, R., Chen, D., et al., Morphometric analysis of gastrectomy-evoked osteope-
nia, Calcif. Tissue Int., 62, 323, 1998.
43. Links, J.M., Beach, L., Subramaniam, B., et al., Edge complexity and partial volume effects,
J. Comput. Assist. Tomogr., 22, 450, 1998.
44. Frost, H.M. and Jee, W.S.S., On the rat model of human osteopenias and osteoporoses, Bone Miner.,
18, 227, 1992.
0266/ ch26/frame Page 403 Friday, October 22, 1999 10:02 AM
404 Mechanical Testing of Bone and the BoneImplant Interface
45. Frost, H.M., Ferretti, J.L., and Jee, W.S., Some roles of mechanical usage, muscle strength, and the
mechanostat in skeletal physiology, disease, and research, Calcif. Tissue Int., 62, 1, 1998.
46. Brinckmann, P., Biggemann, M. and Hilweg, D., Prediction of the compressive strength of human
lumbar vertebrae, Clin. Biomech., 4(S2), S1, 1989.
47. Crenshaw, T.D., Peo, E., Lewis, A., and Moser, B., Bone strength as a trait for assessing mineralization
in swine. A critical review of techniques involved, J. Anim. Sci., 53, 827, 1981.
48. Faulkner, K.G., Gluer, C., Majumdar, S., et al., Noninvasive measurements of bone mass, structure,
and strength. Current methods and experimental techniques, Am. J. Roentgenol., 157, 1229, 1991.
49. Hansson, T., Roos, B., and Nachemson, A., The bone mineral content and ultimate compressive
strength of lumbar vertebrae, Spine, 5, 46, 1980.
50. Hayes, W.C., Piazza, S., and Zysset, P., Biomechanics of fracture risk prediction of the hip and spine
by quantitative computed tomography, Radiol. Clin. North Am., 29, 1, 1991.
51. Louis, O., Boulpaep, F., Willnecker, J., et al., Cortical mineral content of the radius assessed by
peripheral pQCT predicts compressive strength on biomechanical testing, Bone, 16, 375, 1995.
52. Turner, C.H. and Burr, D.B., Basic biomechanical measurements of bone. A tutorial, Bone, 14, 595,
1993.
53. Jerome, C.P., Johnson, C., and Lees, C., Effect of treatment for 6 months with human parathyroid
hormone 1-34 peptide in ovariectomized cynomolgus monkey (Macaca fascicularis), Bone, 17(4S),
415S, 1995.
54. Rosen, H.N., Tollin, S., and Balena, R., Differentiating between orchidectomized rats and controls
using measurements of trabecular bone density: a comparison among DXA, histomorphometry, and
peripheral quantitative computerized tomography, Calcif. Tissue Int., 57, 35, 1995.
55. Takada, M., Engelke, K., Hagiwara, S., et al., Accuracy and precision study in vitro for peripheral
quantitative computed tomography, Osteoporosis Int., 6, 207, 1996.
56. Mosekilde, Li., Mosekilde, Le., and Danielsen, C., Biomechanical competence of vertebral trabecular
bone in relation to ash density and age in normal individuals, Bone, 8, 79, 1987.
57. Mosekilde, Li., Danielsen, C., and Gasser, J., The effect on vertebral bone mass and strength of long term
treatment with antiresorptive agents (estrogen and calcitonin), human parathyroid hormone (1-38), and
combination therapy, assessed in aged ovariectomized rats, Endocrinology, 134, 2126, 1994.
58. Fujii, Y., Miyauchi, A., Takagi, Y., et al., Fixed ratio between radial cortical volume and density
measured by peripheral quantitative computed tomography (pQCT) regardless of age and sex, Calcif.
Tissue Int., 56, 586, 1995.
59. Burr, D.B. and Martin, R.B., The effects of composition, structure and age on the torsional properties
of the human radius, J. Biomech., 16, 603, 1983.
60. Ferretti, J.L., Schiessl, H., and Frost, H. M., On new opportunities for absorptiometry, J. Clin.
Densitom., 1, 41, 1998.
61. Baker, J. L. and Haugh, C., Mechanical properties of bone: a review, Trans. Am. Soc. Agric. Eng.,
22, 678, 1979.
62. Schiessl, H., Ferretti, J.L., Tysarczyk-Niemeyer, G., and Willnecker, J., Noninvasive bone strength
index as analyzed by peripheral quantitative computed tomography (pQCT), in Paediatric Osteology.
New Developments in Diagnostics and Therapy, Schnau, E., Ed., Elsevier, Amsterdam, 1996, 141.
63. Schiessl, H., Ferretti, J.L., Tysarczyk-Niemeyer, G., et al., The role of the muscles to the mechanical
adaptation of bone, in Advances in Osteoporosis, Vol. 1, Lyritis, G.P., Ed., Hylonome, Athens, 1998, 53.
64. Beamer, W.G., Donahue, L., Rosen, C., and Baylink, D.J., Genetic variability in adult bone density
among inbred strains of mice, Bone, 18, 397, 1996.
65. Chiappe, A., Iorio, B., Alvarez, E., et al., Alendronate protects rat femur against the negative impact
of ovariectomy. A densitometric (DXA), tomographic (pQCT) and biomechanical study, Bone, 17,
601, 1995.
66. Chiappe, A., Efectos del Alendronato sobre la Estructura y la Biomecnica sea, Ph.D. thesis,
University of Buenos Aires, Buenos Aires, 1998.
67. Burr, D.B. and Martin, R.B., Mechanisms of bone adaptation to the mechanical environment, Triangle
(Sandoz), 31, 59, 1992.
68. Carter, D.R., Mechanical loading history and skeletal biology, J. Biomech., 20, 1095, 1987.
69. Frost, H.M., Ed., The Intermediary Organization of the Skeleton, Vols. I and II, CRC Press, Boca
Raton, FL, 1986.
0266/ ch26/frame Page 404 Friday, October 22, 1999 10:02 AM
Peripheral Quantitative Computed Tomography 405
70. Frost, H.M., The mechanostat: a proposed pathogenetic mechanism of osteoporoses and the bone
mass effects of mechanical and nonmechanical agents, Bone Miner., 2, 73, 1987.
71. Frost, H.M., Structural adaptations to mechanical usage (SATMU): 2. Redening Wolffs law: the
bone modeling problem, Anat. Rec., 226, 403, 1990.
72. Lanyon, L.E., Functional strain in bone tissue as an objective, and controlling stimulus for adaptive
bone remodelling, J. Biomech., 20, 1083, 1987.
73. Rubin, C.T., McLeod, K., and Bain, S., Functional strains and cortical bone adaptation. Epigenetic
assurance of skeletal integrity, J. Biomech., 23, 43, 1990.
74. Cointry, G.R., Negri, A., Vzquez, S., et al., Densitometric, tomographic (pQCT), histomorphometric
and biomechanical changes produced by chronic aluminum intoxication in rat femurs and tibiae, Bone
Miner., 25(S2), S13, 1994.
75. Ferretti, J.L., Effects of bisphosphonates on bone biomechanics, in Bisphosphonate on Bones, Bijvoet,
O.L., Caneld, R., Fleisch, H., and Russel, R., Eds., Elsevier, Amsterdam, 1995, 211.
76. Parma, M., Schneider, P., Piccinni, E., et al., Threshold-dened ROI analysis of bisphosphonate effects
on bone structure in tumor-implanted mice employing pQCT, Bone, 23(5S), S479, 1998.
77. Roldn, E.J., Prez Lloret, A., and Ferretti, J.L., Olpadronate: a new amino-bisphosphonate for the
treatment of medical osteopathies, Expert Opinion Invest. Drugs, 7, 1521, 1998.
78. Frost, H.M., Dening osteopenias and osteoporoses. Another view (with insights from a new para-
digm), Bone, 20, 385, 1997.
0266/ ch26/frame Page 405 Friday, October 22, 1999 10:02 AM
0266/ ch26/frame Page 406 Friday, October 22, 1999 10:02 AM

407

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

27

Computer Modeling
for Evaluating Trabecular
Bone Mechanics

Rakesh Saxena and Tony S. Keller

CONTENTS

I. Introduction ..........................................................................................................................408
II. Image Array Acquisition......................................................................................................409
A. Quantitative Serial Sectioning and Serial Grinding (QSSG) ........................................409
Plate 27.1 QSSG Applied ....................................................................................410
B. Computed Tomography (CT) .........................................................................................410
C. Magnetic Resonance Imaging (MRI) ............................................................................412
III. Image Processing .................................................................................................................413
A. Data Types......................................................................................................................413
B. Displaying Rasters..........................................................................................................413
C. Data Conversion and Feature Extraction.......................................................................414
Plate 27.2 Image Processing Applied: Thresholding ..........................................414
D. Volume of Interest Selection..........................................................................................414
Plate 27.3 Image Processing Applied: VOI Selection.........................................415
E. Raster-to-Vector Conversion and Image Rendering ......................................................415
Plate 27.4 Image Processing Applied: Image Rendering....................................417
IV. Finite Element Volume Meshing .........................................................................................419
A. Smooth Surface Tetrahedral FE Mesh Modeling..........................................................420
B. Voxel-Based FE Mesh Modeling...................................................................................420
Plate 27.5 FE Mesh Generation Applied: Tetrahedral and Voxel Elements.......422
V. FE Stress Analysis................................................................................................................423
A. Analysis Assumptions ....................................................................................................423
B. The Uniaxial Compressive Test .....................................................................................424
C. Boundary Conditions......................................................................................................424
D. FE Algorithm..................................................................................................................424
E. Validation of FE Model: Convergence ..........................................................................425
F. Validation of FE Model: FE Analysis ...........................................................................425
Plate 27.6 FE Stress Analysis Applied: Tetrahedral Model Convergence..........425
G. Computation of Stresses and Strains .............................................................................428
Plate 27.7 FE Stress Analysis Applied: StressStrain Intensity
and Poissons Ratio ............................................................................428
Plate 27.8 FE Stress Analysis Applied: Tetrahedral vs. Brick Model ................429
H. Determination of Tissue Modulus .................................................................................430

0266/ ch27/frame Page 407 Friday, October 22, 1999 10:06 AM

408

Mechanical Testing of Bone and the BoneImplant Interface

VI. Morphological Analysis .......................................................................................................430
A. The Mean Intercept Length (MIL) Method...................................................................430
Plate 27.9 Morphological Analysis Applied: MIL Method ................................431
B. The Volume Orientation (VO) Method..........................................................................433
C. The Star Volume Distribution (SVD) Method...............................................................433
D. The Star Length Distribution (SLD) Method................................................................433
VII. Summary ..............................................................................................................................434
References ......................................................................................................................................434

I. INTRODUCTION

A knowledge of mechanical properties of trabecular bone is essential in understanding problems
of age-related bone fracture, prosthesis loosening, and bone remodeling. Over the years, researchers
have attempted, both experimentally and theoretically, to relate bone mechanical properties (stiff-
ness, strength) to bone physical properties (bone density, bone volume, bone architecture, etc.).
Numerous experimental studies have shown that both apparent modulus and strength have a power-
law relationship with apparent density, and that the elastic mechanical properties of bone improve
as the apparent density increases.

1-5

The elastic properties of bone are also dependent upon a number
of the factors including anatomical site, tissue heterogeneity, architecture, loading direction, and
age. Other bone material properties, tissue modulus and Poissons ratio, also have a direct impact
on the apparent modulus and strength. In the past, research in this eld has utilized

in vivo

,

in vitro

,
and simple mathematical modeling techniques.
More recently, powerful computers with large memory, speed, and multiple processors com-
bined with efcient imaging schemes have made it possible to perform large-scale, anatomically
precise three-dimensional (3D) numerical modeling of trabecular bone. Such 3D models of trabe-
cular bone are usually analyzed by nite-element (FE) stress analysis. The advantages of computer
modeling of trabecular bone are as follows:
1. Physical properties can be related to mechanical properties noninvasively.
2. Parametric studies can be performed very easily (for example, variations of Poissons
ratio value on apparent modulus).
3. Precise 3D modeling can be used to quantify bone architecture.
4. Heterogeneity and anisotropy can be addressed by analyzing 3D models from different
anatomical sites and by varying the loading direction, respectively.
5. Tissue modulus can be predicted.
Computer modeling of trabecular bone not only provides an alternative method to obtain
correlations between bone physical properties and mechanical properties, but also allows one to
study experimental artifacts such as boneplaten friction and local damage and increased compli-
ance associated with cut surfaces of bone. Such artifacts can be problematic in mechanical tests,
since they are well-known sources of measurement error.
A schematic illustration of the key steps involved in computer modeling of trabecular bone is
shown in Figure 27.1. These are image array acquisition, image processing, volume of interest
(VOI) selection, surface mesh generation, FE volume mesh generation, FE analysis, and morphological
analysis. Each of these processes is summarized in the following sections. Various stages of the
modeling and analysis process will be illustrated with a simple unit cell model of trabecular bone.
Special sections identied as Plates are used to highlight modeling and analysis processes applied
to the trabecular bone unit cell model.

0266/ ch27/frame Page 408 Friday, October 22, 1999 10:06 AM

Computer Modeling for Evaluating Trabecular Bone Mechanics

409

II. IMAGE ARRAY ACQUISITION

Image rendering is basically data visualization, and requires two-dimensional (2D) data sets which
are transformed into a set of graphic primitives. Biological data are generally acquired using medical
imaging hardware or histological techniques. Image data obtained using medical imaging hardware
is in the form of planar or 2D slices of the patients anatomy. There are several noninvasive medical
imaging techniques which can be used to gather 3D data, including for example, X-ray, tomography,
radionuclide imaging, ultrasound, and magnetic resonance imaging. Most of these noninvasive
imaging methods are based on the principles of probing fundamental properties of matter through
mechanical or electromagnetic waves.
Information obtained from quantication of these properties is utilized in reconstructing a 2D
image, which allows one to visualize various tissues in the body in some manner. For example, an
attenuation in X-ray intensity proportional to the tissue density leads to different gray levels of the
image on photographic lm. On the other hand, water content variations inherent in different
biological tissues result in different resonant peaks in magnetic resonance imaging.
In this chapter the most common histological and noninvasive techniques for acquisition of
bone image data, i.e., quantitative serial sectioning and serial grinding (QSSG), computed
tomography imaging (CT and

-CT), and magnetic resonance imaging (MRI), will be discussed
in more detail.

A. Q

UANTITATIVE

S

ERIAL

S

ECTIONING



AND

S

ERIAL

G

RINDING

(QSSG)

Quantitative serial sectioning and serial grinding are histological techniques and therefore require
removal of a specimen of bone tissue from the body. One of the essential elements of images
derived from QSSG and other imaging techniques is adequate contrast between the tissue of interest
(e.g., bone) and surrounding tissues (e.g., marrow, ligaments, muscle, etc.). Good contrast or
physical dissimilarity between bone and the surrounding tissues is important so that the feature of
interest (i.e., bone) can be extracted from the image background. Noisy or otherwise poorly

FIGURE 27.1

Schematic illustration of the steps used to generate and analyze a three dimensional model
of trabecular bone (1 = > smooth surface tetrahedral mesh and 2 = > voxel based brick/tetrahedral mesh).

0266/ ch27/frame Page 409 Friday, October 22, 1999 10:06 AM

410

Mechanical Testing of Bone and the BoneImplant Interface

contrasted images make feature extraction (contouring in 2D, surface rendering in 3D) extremely
difcult and may lead to an inaccurate description of the anatomy.
To enhance contrast and facilitate feature extraction in trabecular bone one needs to replace
the bone marrow with some medium which is dissimilar in color value from bone. For this purpose,
the marrow is removed from the bone using a high-pressure water jet, defatted in acetone (24 h),
bleached in dilute ammonia solution (24 h), and dried (100

C for 1 h). This produces a trabecular
bone specimen which is chalk-white in appearance (Figure 27.2A). The cleaned and dried specimen
is then embedded in black-pigmented epoxy resin. A low-speed centrifuge can be used to facilitate
inltration of the resin throughout the porous network of trabeculae. This results in a histological
preparation which has excellent contrast between bone and marrow spaces (Figure 27.2B), and
from which the bone features can be accurately extracted using a process called thresholding
(Figure 27.2C).
In order to capture the bulk architecture of the bone, the specimen is serially sectioned using
a microtome (Figure 27.2D) or serially ground using a Computer Numerically Controlled (CNC)
milling machine.
The surface of each slice is then recorded using a frame capture camera connected to a computer
frame grabber, with care taken to ensure that each slice is spatially registered in an identical manner
so that image distortion between adjacent slices is minimized. This is readily accomplished by
xing the recording device (charged-coupled device or CCD camera) over the specimen.

Plate 27.1 QSSG APPLIED

A specimen of human (60-year-old male) L4 vertebral trabecular bone (9

9 mm) was removed
from the vertebral centrum using an Isomet low-speed diamond saw (Buehler, Lake Bluff, IL). The
specimen had an apparent density of 0.3 g/cm

3

(

a

= bone mass/bulk volume) and bone volume fraction
(

V

f

= bone volume/bulk volume) of 11%. The specimen was embedded in a black polyester resin and
was serially sliced along the superior-inferior axis at 20-

m intervals using a Reichert-Jung

polycut
E microtome (Leica Microsystems, Inc., Deereld, IL). Then, 16-bit color video images of each sectioned
surface were recorded at an image resolution of 20 mm/pixel, using a CCD camera and a PC image capture
card (TARGA 16 graphic board, Truevision, Inc., Indianapolis, IN). MIPS (Map and Image Processing
Software) program (MicroImages, Inc., Lincoln, NE) was used to create a 3D binary image array comprising
of 201 serial (20

20

m) resolution slices.

6

This process is summarized in Figure 27.2.

B. C

OMPUTED

T

OMOGRAPHY

(CT)

In conventional radiography, X rays diverging from a source pass through the body projecting an
image of the skeleton, organs, air spaces, and any existing tumors onto a sheet of lm. However,
this technique has certain limitations, namely, a 3D object is projected onto a 2D plane. Thus, many
planes are superimposed onto one, and the depth information is lost. Also, conventional X rays
cannot differentiate between soft tissues. The tomography (the Greek word

tomo

means a cut
section) technique overcomes the limitation of depth loss by taking an X-ray projection of a single
plane. This is done by simultaneously moving the X-ray source in one direction and photographic
lm in the opposite direction. Thus, if the patients body can be regarded as a series of planes
parallel to the lm, there is only one plane that remains stationary with respect to the lm as the
lm moves; hence, only this plane remains sharply focused and the other planes are blurred.
In 1917, Radon

7

proved that a 2D or 3D object can be reconstructed uniquely from the innite
set of all its projections. Based on this principle, techniques have been developed to reconstruct
images from projections. In CT a 2D image is reconstructed by multiple projections of a single
plane. CT offers much greater sensitivity and so enables soft tissue to be differentiated from hard
tissue. In conventional quantitative CT (QCT), the images have an out-of-plane resolution of 1000
to 2000

m and an in-plane resolution of approximately 500

m. Such image resolutions are
incapable of resolving trabecular bone structures, which can be 150

m in diameter and smaller.

0266/ ch27/frame Page 410 Friday, October 22, 1999 10:06 AM

Computer Modeling for Evaluating Trabecular Bone Mechanics

411
F
I
G
U
R
E

2
7
.
2

T
r
a
b
e
c
u
l
a
r

m
o
r
p
h
o
l
o
g
y
,

q
u
a
n
t
i
t
a
t
i
v
e

s
e
r
i
a
l

s
e
c
t
i
o
n
i
n
g
,

a
n
d

i
m
a
g
i
n
g

p
r
o
c
e
d
u
r
e
.

(
A
)

B
o
n
e

e
m
b
e
d
d
e
d

i
n

b
l
a
c
k

p
o
l
y
e
s
t
e
r

r
e
s
i
n
(
t
o
p
)

a
n
d

c
l
e
a
n
e
d
,

d
r
i
e
d

a
n
d

b
l
e
a
c
h
e
d

t
r
a
b
e
c
u
l
a
r

b
o
n
e

s
p
e
c
i
m
e
n
;

(
B
)

U
n
t
h
r
e
s
h
o
l
d
e
d
,

2
4
-
b
i
t

i
m
a
g
e

o
f

h
u
m
a
n

l
u
m
b
a
r

v
e
r
t
e
b
r
a
l

t
r
a
b
e
c
u
l
a
r

b
o
n
e
s
u
r
f
a
c
e
;

(
C
)

t
h
r
e
s
h
o
l
d
e
d

b
i
n
a
r
y

i
m
a
g
e

o
f

(
B
)
,

B
o
n
e

w
h
i
t
e

p
i
x
e
l
s
,

m
a
r
r
o
w

b
l
a
c
k

p
i
x
e
l
s
;

(
D
)

e
x
p
e
r
i
m
e
n
t
a
l

s
e
t
u
p

i
l
l
u
s
t
r
a
t
i
n
g

t
h
e

p
o
l
y
c
u
t
-
E

h
a
r
d

t
i
s
s
u
e

m
i
c
r
o
t
o
m
e

a
n
d

h
i
g
h

r
e
s
o
l
u
t
i
o
n

(
u
p

t
o

2
0
4
8

2
0
4
8

p
i
x
e
l
s
)

f
r
a
m
e

c
a
p
t
u
r
e

c
a
m
e
r
a

(
J
V
C

T
K
7
1
0
0
U
)

u
s
e
d

t
o

s
e
r
i
a
l
l
y

s
e
c
t
i
o
n

a
n
d
i
m
a
g
e

t
h
e

t
r
a
b
e
c
u
l
a
r

b
o
n
e

s
p
e
c
i
m
e
n
s
.

T
h
e

s
e
c
t
i
o
n
i
n
g

a
n
d

i
m
a
g
i
n
g

p
r
o
c
e
s
s

i
s

f
u
l
l
y

a
u
t
o
m
a
t
e
d
.

0266/ ch27/frame Page 411 Friday, October 22, 1999 10:06 AM

412

Mechanical Testing of Bone and the BoneImplant Interface

Indeed, a study investigating the morphology of the lumbar vertebral centrum indicated that the
optimal image resolution for human trabecular bone was 25

m or better.

8

More recently, peripheral
QCT (pQCT) and micro-CT (

-CT) imaging techniques have been developed which can provide
high-resolution images of small specimens of trabecular bone. In

-CT, resolutions of 20

m are
possible (Figure 27.3). The advantages of

-CT are it is nondestructive, less time-consuming than
QSS, and has better resolution. By using

-CT techniques it is also possible to obtain

in vivo

images
of rodent skeletal tissues.

C. M

AGNETIC

R

ESONANCE

I

MAGING

(MRI)

This technique is based on the principle of nuclear magnetic resonance (NMR). According to this
principle, resonant peaks are obtained from a material placed in an electromagnetic eld as energy
is absorbed by spin ips of protons. These resonant peaks provide a unique energy signature of a
particular substance. When the human body is subjected to a strong magnetic eld, the NMR signal
is due predominantly to water protons. Hence, water content differences in various tissues can be
detected by ips of protons present in the uid-containing tissues. The energy signature can be
processed into an X-ray-like image. MRI has the advantage of being able to penetrate bony and air-
lled structures with negligible attenuation and with minimal artifacts. Also, unlike conventional X-ray

FIGURE 27.3

XCT images (256

256 pixels, 22

22

m resolution) and FE mesh of mouse femur: (a) slice
160 from a set of 234 slices at 20

m increments; (b) TBMAS volumetric rendering; and (c) detail; (d) the
FE mesh.

0266/ ch27/frame Page 412 Friday, October 22, 1999 10:06 AM

Computer Modeling for Evaluating Trabecular Bone Mechanics

413

and computed tomography, MRI uses nonionizing radiation and is minimally invasive. Formerly, MRI
was used mainly for imaging the brain and the spinal cord. However, with the advent of faster
computers and imaging techniques, MRI has been extended to the chest and abdomen where thorax
and blood vessel motion artifacts had previously limited the application of this technique.

III. IMAGE PROCESSING

Historically, pictures, images, maps, and charts have been made of paper and ink or lm. Prior to
the computer revolution, scientists used stereology tools (which have been around since the early
1950s and 1960s) to analyze 2D data sets obtained photographically from scanning electron
microscopes and other available histological tools. In the case of X-ray lms, CT scans, and MRI
scans, microcomputers have turned paper and lm into 2D and 3D image processing and graphical
information systems. A desktop computer can manipulate 2D and 3D images to make their storage,
retrieval, and interpretation easier and more accurate for researchers and technicians alike.
Today sophisticated image processing techniques (many of which were originally developed
for military applications) greatly simplify the task of displaying, converting, and interpreting
complex trabecular bone data sets derived from QSSG, CT, and MRI techniques.

A. D

ATA

T

YPES

Three-dimensional image data of an object obtained using QSSG/QCT/MRI techniques generally
comprise a series of planar or 2D arrays of a single data type. A single, related, 2D grouped set of
numbers of a single data type is dened as a raster or raster object. Each number represents the
value of some parameter (color, attenuation coefcient, etc.), and its position in the group represents
its relative position to the other values. A raster format, therefore, breaks an image into a grid of
equally spaced pieces called pixels and records color information for each pixel. Most image
processing software packages can display images from raster objects of any data type (1 to 64 bit
integers or oating point numbers, typically 8, 16, or 24 bit). For example, if every pixel is 12 bits,
then the color value will be an integer number between 0 and 4095 (2

12

1). If every pixel is 1 bit,
then the image is binary (0 or 1) or black and white.
One of the main problems with image les is the variety of standard and proprietary le formats
in use. The format of a le is very important because it allows interchange of les between people
and programs. Image les normally contain a header which includes the le format version, a
description of data (bits per pixel, spatial dimensions) and the structure of the data (compressed or
not), followed by the data section. The simplest le format for storage of image data is a simple
integer array which, as the name implies, simply contains a bitmap or row

column array of digital
encoded information and no header. Other standard bitmap le formats which contain header infor-
mation include TARGA (.TGA, Truevision Inc., Indianapolis, IN) and TIFF (.TIF, tag image le
format, Adobe Developer Association, Mountain View, CA). File size is also an important consider-
ation when working with images. Assuming, for example, a raster object comprising a 512

512
pixel array where every pixel is encoded with 12 bits, then the size of a simple integer array le will
be 524,288 bytes (512 kilobytes) since each pixel must be saved in 2-byte words; 8-bit pixel data can
be stored as 1-byte words, which reduces the bitmap le size to 262,144 bytes (256 kilobytes). Simple
array, TARGA, and TIFF les, can be compressed to reduce le size, and recent developments in
data compression have enabled bitmap le sizes to be drastically reduced, often as much as 90%.

B. D

ISPLAYING

R

ASTERS

The ability to display meaningfully or numerically represent a raster object of a given data type is
one of the most important elements of image processing. Indeed, if one wishes to extract quantitative
information accurately from a raster object (trabecular bone features, for example), then good contrast

0266/ ch27/frame Page 413 Friday, October 22, 1999 10:06 AM

414

Mechanical Testing of Bone and the BoneImplant Interface

or dissimilarity between elements in the raster is an essential and prerequisite requirement for the
object. As described in Section II.A of this chapter, the rst step in acquiring a raster object repre-
sentation of a histological sample of bone is to ensure that there is good contrast between the bone
and background (marrow spaces) prior to capturing the image. One can also tune signals associated
with CT and MRI systems in order to obtain higher-contrast raster objects. Moreover, it is possible
to numerically alter the contrast of the raster object after the image has been captured. Many software
packages (TBMAS, Musculoskeletal Research Laboratory, Burlington, VT; NIH Image, Scion Cor-
poration, Frederick, MD; MIPS, MicroImages, Inc., Lincoln, NE; and others) enable the user to
modify a raster object digitally using standard contrast methods such as linear, normalized, equalized,
or logarithmic contrast enhancement. Such digital processes can improve the contrast of the raster
object,

9

but it is always a good idea to start with the best possible contrast image.

C. D

ATA

C

ONVERSION



AND

F

EATURE

E

XTRACTION

Before one can perform morphological and/or FE analyses of trabecular bone, planar (2D) or
multiplanar (3D) raster objects must be converted into binary or vector representations. Thus, some
method to separate the object(s) of interest is required. This process is called feature extraction
and involves isolating areas in raster images into surface features based on similarities and differ-
ences associated with the raster data values. In the case of trabecular bone, extraction of the
trabecular bone features from the surrounding marrow spaces is of principal interest. Feature
extraction is generally a semiautomated process which involves thresholding or setting a data
conversion separation limit such that any value above the designated limit or threshold is assigned
a 1 value (in binary thresholding, a 1) and any incoming value below or equal to the limit is
assigned another value (in binary thresholding, a 0) (Plate 27.2). Thus, thresholding of 2 or
higher-bit raster images generates a binary raster. Binary rasters are particularly useful for identi-
fying features such as edges, which is an important step in raster-to-vector conversion. Binary
rasters are also a standard input le format for stereologic analysis which is used to quantify bone
morphology. Raster-to-vector conversion (creation of surface and volumetric meshes) and quanti-
cation of bone morphology (structural properties) are described later in this chapter.

Plate 27.2 IMAGE PROCESSING APPLIED: THRESHOLDING

A typical unthresholded section of a small region of trabecular bone is shown in Figure 27.4a. A binary
representation of the trabecular bone was produced using an iterative thresholding algorithm that
determined the optimal color level separating the bone and marrow. The threshold optimization proce-
dure consisted of computing a color histogram of the captured image (Figure 27.4c) and applying an
iterative scheme (Figure 27.4d) that provided increasingly cleaner extractions of the bone. A convenient
threshold to start the algorithm could be

T

o



= (

P

B

+ P

M

)/2, where

P

B

and

P

M



correspond to the peaks
of bone and marrow, respectively. The iterative scheme converged to an optimal color threshold level
(

P

opt

= 76) after four iterations. By using the optimal color threshold value, each pixel in the original,
unthresholded image was assigned to a black (0, marrow) or white (1, bone) color value based upon
whether its value was less than or greater than

P

opt

, respectively. The resulting thresholded image is
shown in Figure 27.4b.

D. V

OLUME



OF

I

NTEREST

S

ELECTION

A smaller volume of interest (VOI) within a larger volume is often desired and can be selected by
specifying its size and location within the raster array. Reasons for selecting a smaller volume are
as follows:
1. Regions near the edges are usually not dened clearly.
2. The maximum size of the VOI is limited by computing resources available.

0266/ ch27/frame Page 414 Friday, October 22, 1999 10:06 AM

Computer Modeling for Evaluating Trabecular Bone Mechanics

415

Plate 27.3 IMAGE PROCESSING APPLIED: VOI SELECTION

A small 1.5

1.5

2.0 mm subvolume or unit cell of trabecular bone is used to illustrate basic image
rendering, FE mesh generation, FE stress analysis, and morphological analyses applied to trabecular
bone (detailed in the following sections of this chapter). Note, however, that a larger VOI should be
used in order to satisfy continuum requirements.

11

A general rule of thumb is to select a volume
comprising ve intertrabecular distances. A custom program,

Trabecular Bone Morphology and Analysis
System

(TBMAS),

12

was used to select the VOI (Figure 27.5).

E. R

ASTER

-

TO

-V

ECTOR

C

ONVERSION



AND

I

MAGE

R

ENDERING

Raster data is a useful format for displaying planar or 2D data, but is less useful for visualization
of multiplanar or 3D data arrays. Moreover, computer-aided design (CAD) and FE mesh generation
packages rely on a vector description of an object. Thus, it is desirable to convert raster data into
vector data. Raster-to-vector conversion is relatively straightforward in 2D, but is much more
complicated for 2D structures, particularly nonhomogeneous, porous, open-cell lattice structures
such as trabecular bone.
Raster-to-vector conversion or image rendering involves converting the raster object (pixels)
into an object comprising nodes, lines, and polygons (vector). One of the simplest vector le formats
consists of a list of nodes followed by a list of elements (connectivity). For example, a simple 1

1 unit cube is represented in vector format as follows:

FIGURE 27.4

(a) Unthresholded image of a section of vertebral centrum; (b) image (a) thresholded at color
value = 76; (c) the color histogram; (d) the iterative thresholding algorithm.

10

0266/ ch27/frame Page 415 Friday, October 22, 1999 10:06 AM

416

Mechanical Testing of Bone and the BoneImplant Interface

/* x, y, and z coordinates of 8 vertices */
0 0 0
1 0 0
1 1 0
0 1 0
0 0 1
1 0 1
1 1 1
0 1 1
/* Connectivity information of 12 triangles that comprise the cube surface */
1 2 5
6 5 2
2 3 6
7 6 3
3 4 8
8 7 3
5 4 1
4 5 8
4 2 1
4 3 2
5 6 8
6 7 8
In the unit cube example, the numeric values of the vertices and elements are space-delimited
integer values. Comma- and tab-delimited integers and real numbers can also be used. Standard
vector formats include stereo lithography (STL), initial graphics exchange specication (IGES),
drawing exchange format (DXF), and many others.
There are two main types of image rendering

surface rendering

and

volume rendering

.
Surface rendering, as the name suggests, describes only the surface of an object and not the interior.
However, if one is interested in looking inside an object, for example at various tissues inside a
skull, this is done by volume rendering.

FIGURE 27.5 Volume selector tool used to select a volume of interest (1.5 mm 1.5 mm 2.0 mm shaded
region) from the 20 m voxel resolution dataset (7.00 mm 6.20 mm 4.02 mm volume). (From Saxena, R.
et al., Comput. Methods Biomech. Biomed. Eng., 2, 287, 1999. With permission.)
0266/ ch27/frame Page 416 Friday, October 22, 1999 10:06 AM
Computer Modeling for Evaluating Trabecular Bone Mechanics 417
Lorensen and Cline
13
proposed a surface rendering technique to create an isosurface from a
discrete 3D raster data set called marching cubes (MC). The MC algorithm involved spanning the
3D raster data set by a 3D volume consisting of eight-node voxels (volume cells) and then
determining how the isosurface (3D contour of constant scalar value) data intersect the 12 edges
of each cube by checking the value at each of the eight cube vertices. An intersection between the
isosurface and the edge is implied only if the isosurface values lies between the two vertex values
of any edge. Figure 27.6 shows the main steps in an MC-based surface mesh generation scheme.
The Visualization ToolKit (VTK)
14
is a powerful graphic package that utilizes the MC algorithm
to generate a triangular surface mesh separating regions of different isodensities (surfaces of
constant density values). The code can written in either C, C++, or Tcl.
Plate 27.4 IMAGE PROCESSING APPLIED: IMAGE RENDERING (Figure 27.7)
Tcl Code
# Get the interactor ui
source vtkInt.tcl;
source colors.tcl;
# Create the render master
vtkRenderMaster rm;
set renWin [rm MakeRenderWindow];
set ren1 [$renWin MakeRenderer];
set iren [$renWin MakeRenderWindowInteractor];
FIGURE 27.6 Surface mesh generation on VTK (* indicates optional steps).
0266/ ch27/frame Page 417 Friday, October 22, 1999 10:06 AM
418 Mechanical Testing of Bone and the BoneImplant Interface
# A virtual light for 3D rendering
vtkLight lgt;
# Read the binary slice data
vtkVolume16Reader v16;
v16 SetDataDimensions 77 77;
[v16 GetOutput] SetOrigin 0.0 0.0 0.0;
v16 SwapBytesOn;
v16 SetFilePrex data/trab/z60im;
v16 SetImageRange 0 101;
v16 SetDataAspectRatio .02 .02 .02;
# Apply marching cubes (MC) algorithm
vtkMarchingCubes iso;
iso SetInput [v16 GetOutput];
# Specify the isosurface value for tissue
iso SetValue 0 1;
# Merge duplicate points and remove
# degenerate primitives
vtkCleanPolyData clean;
clean SetInput [iso GetOutput];
# Apply connectivity to extract largest
# region
vtkPolyConnectivityFilter connect;
connect SetInput [clean GetOutput];
connect ExtractLargestRegion;
# Reduce the number of triangles in a mesh
vtkDecimate deci;
deci SetInput [clean GetOutput];
# Adjust point positions using Laplacian
# smoothing
vtkSmoothPolyFilter smooth;
smooth SetInput [deci GetOutput];
# Map vtkPolyData to graphics primitives
vtkPolyMapper isoMapper;
FIGURE 27.7 The surface mesh model of the unit cell (1.5 1.5 2.0 mm
3
) of trabecular bone (20,545
nodes and 41,544 triangles). (From Saxena, R. et al., Comput. Methods Biomech. Biomed. Eng., 1999. With
permission.)
0266/ ch27/frame Page 418 Friday, October 22, 1999 10:06 AM
Computer Modeling for Evaluating Trabecular Bone Mechanics 419
isoMapper SetInput [connect GetOutput];
isoMapper ScalarsVisibleOff;
# Represent an object in a rendered scene
vtkActor isoActor;
isoActor SetMapper isoMapper;
set isoProp [isoActor GetProperty];
eval $isoProp SetColor 0.286 0.843 1.;
# Write vtk polygonal data
vtkPolyWriter writer;
writer SetInput [connect GetOutput];
writer SetFilename trab_z60im_out;
writer Write;
# Add the actors to the renderer,
# set the background and size
$ren1 AddActors isoActor;
$ren1 AddActors textActor;
$ren1 SetBackground 0 0 0;
$ren1 AddLights lgt;
$renWin SetSize 600 600;
$ren1 SetBackground 1 1 1;
$renWin DoubleBufferOff;
set cam1 [$ren1 GetActiveCamera];
$cam1 Elevation 0;
$cam1 SetViewUp 0 1 0;
$cam1 Zoom 1.0;
$cam1 Azimuth 0;
eval lgt SetPosition [$cam1 GetPosition];
eval lgt SetFocalPoint [$cam1 GetFocalPoint];
# Render the image
$iren SetUserMethod {wm deiconify .vtkInteract};
$renWin Render;
puts Done;
IV. FINITE ELEMENT VOLUME MESHING
The FE method is a convenient numerical technique to analyze a complex structure of trabecular bone
for which a closed-form solution using analytical methods is impossible. In recent years, detailed 3D
FE models of trabecular bone have been investigated.
15-17
In these FE models (usually constructed
from CT raster arrays), trabecular bone is represented by a 3D array of identical eight-node brick
elements also called voxels. The main advantages of voxel-based FE models are (1) voxel models
can be generated extremely fast, and (2) very efcient stress analysis schemes can be implemented
using voxel elements. However, voxel elements lead to jagged edges due to protruding vertices of the
cubes at the surface, which results in errors in the computation of local stresses and strains.
More recently, volumetric modeling of trabecular bone has been accomplished using four-node
tetrahedral elements.
18,19
Using the marching cubes principle, Mller and Regsegger
18
developed a
volumetric marching cubes (VOMAC) algorithm. When applied to a 3D discrete data set VOMAC
generates tetrahedrons, in contrast to the MC approach which generates surface triangles. Thus, instead
of dening a triangle surface conguration for every voxel, a tetrahedron subvolume is assembled
inside the enclosed volume by the VOMAC algorithm. By using this technique the bone surface can
be represented more accurately; however, the mesh generation time required is much longer.
Differences in apparent moduli derived from the two approaches have been shown to range
from 2 to 27.5% for a simplied 2D model of a plate with a centered hole.
20
In general, however,
0266/ ch27/frame Page 419 Friday, October 22, 1999 10:06 AM
420 Mechanical Testing of Bone and the BoneImplant Interface
the apparent mechanical behavior of trabecular bone can be predicted reasonably well by large-
scale, voxel-based FE (LS-FE) schemes.
21,22
The trabecular bone unit cell will be modeled and analyzed using both voxel and surface
meshing schemes.
A. SMOOTH SURFACE TETRAHEDRAL FE MESH MODELING
For the unit cell trabecular bone example, the surface mesh was created using a custom FE mesh
generator 3dMesh.
23
3dMesh generates four-node tetrahedral elements from a surface mesh input le.
3dMesh is a grid-based system that deploys a uniform volume mesh spanning the domain
(Figure 27.8a). The domain node material classication is determined using a boundary-line crossing
routine.
24
Each domain node, D, is the local origin of the vector, E. This E vector is oriented in the
positive x-direction with a magnitude equal to the length of the body diagonal of the model. This
length ensures that the end point of E lies outside the domain. Every time E intersects a boundary
element, the counters of the two materials associated with that boundary element are incremented.
After all the boundary elements are checked for intersection with E, the material counter showing an
odd number of crossings dictates the material zone of node D. If all the counters are even, then the
node is external to the domain and it is classied with a 0 (Figure 27.8b). The external material
boundary counter, material 0, is not included in the analysis. Areas of special interest (bone surfaces
in this example) are specied for mesh renement. The important features of 3dMesh are as follows:
1. Gauge Size: The default mesh gauge size is a user specied parameter which corresponds
to an edge length of a tetrahedron.
2. Renement: Renement about the bone surfaces is performed by creating a bounding
rectangular parallelepiped about each surface platelet or triangle. A renement length
parameter is specied by the user. Each bounding parallelepiped is enlarged in all
directions by this renement length. All existing volume tetrahedrons whose centroid
lies within any expanded parallelepiped get tagged for renement. Each tagged tetrahe-
dron is subdivided into eight smaller tetrahedrons. These rened elements get knitted
into the parent mesh such that the mesh remains consistent and contiguous (Figure 27.8c).
3. Multiple Materials: The algorithm can handle multiple material regions including, for
example, bone marrow elements.
4. FE Surface Mesh: The FE surface mesh can also be generated for comparison to the
input surface (VTK mesh). Several noncommercial (CUBIT, QMG, GENIE++) and
commercial (ALGOR, ANSYS, MARC) mesh generators are currently available. A
l i st descri bi ng and compari ng vari ous mesh generat ors i s l ocat ed at :
<http://www.andrew.cmu.edu/user/sowen/softsurv.html>.
B. VOXEL-BASED FE MESH MODELING
Brick FE meshes, in which the elements are eight-node cubes, can be generated directly from a
3D voxel data set (Figure 27.9). The rst step is to assign the voxel dimensions. For example, a
100 100 100 mm voxel will have 5
3
or 125 pixels if the data set has a resolution of 20 mm in
all three directions. Next, the data set is divided into a 3D cubic grid consisting of voxels of identical
shapes and sizes. Each voxel in the data set is then checked to determine if it is bone or not
bone (i.e., marrow). A common criterion is that if more than half of the volume (i.e., more than
50% of the pixels in the voxel) are bone pixels, then the voxel is considered to be bone; otherwise,
the voxel is assigned to marrow. This is a straightforward method and a brick mesh can be generated
extremely fast a few seconds compared with a few hours for a smooth surface FE mesh of
similar size and resolution. Note that a tetrahedral mesh can be generated directly from the brick
mesh by creating ve tetrahedra from each brick element.
0266/ ch27/frame Page 420 Friday, October 22, 1999 10:06 AM
Computer Modeling for Evaluating Trabecular Bone Mechanics 421
An important step after (or during, depending upon the algorithm) the FE meshing procedure
is to extract the largest region. This step ensures that small regions not connected to the largest
region are removed. Unconnected regions would produce singularities during FE analysis. A good
check for the validity and quality of surface generation and FE mesh generation schemes is to
compute the bone volume fraction (V
f
or BVF) and surface area (A
s
) at each step and compare it
with the original value. Also, the aspect ratio (the ratio between the longest to shortest dimensions
of the element) of elements in FE mesh should not be too high. The optimum aspect ratio at any
location within the grid depends largely upon the difference in rate of change of displacements in
different directions. If displacements vary at about the same rate in each direction, the closer the
FIGURE 27.8 Generation of smooth tetrahedral FE mesh using 3dMesh: (a) a node grid bounds the surface
mesh of the unit cell; (b) node classication scheme; and (c) mesh renement. (i) Each tetrahedron has six
midside nodes added and is rened into eight elements; breakdown pattern for tetrahedron containing (ii) three;
(iii) two; and (iv) one midside node.
0266/ ch27/frame Page 421 Friday, October 22, 1999 10:06 AM
422 Mechanical Testing of Bone and the BoneImplant Interface
aspect ratio to unity, the better the quality of the solution.
25
As a general rule, the aspect ratio should
be kept below 10 for deformation analysis and below 3 for stress analysis.
26
Plate 27.5 FE MESH GENERATION APPLIED:
TETRAHEDRAL AND VOXEL ELEMENTS
Figure 27.10a shows a 1.5 1.5 2.0 mm 3dMesh tetrahedral FE model generated from the trabecular
bone subvolume created by the VOI routine. Initially, the mesh is created by dividing the bone volume
into equal-size hexahedrons (eight-node cubes). Each cube is then broken down into ve tetrahedra.
As the meshing algorithm progresses, however, the size of the tetrahedra may vary. The size of
tetrahedral elements can also be varied by specifying alternate mesh density in a local region. The
maximum number of elements that can be generated by 3dMesh is limited only by machine memory.
The four-node tetrahedral meshes were generated by 3dMesh on a 200-MHz Silicon Graphics Chal-
lenger system with 256 MB memory. An eight-node brick element FE mesh was generated directly
from the binary image les selected using the VOI routine. The FE mesh consisted of 52,293 nodes
and 37,126 brick elements (Figure 27.10b). The mesh had a resolution of 20 m, and the model had a
V
f
= 7.05%.
FIGURE 27.9 Voxel-based FE mesh generation.
0266/ ch27/frame Page 422 Friday, October 22, 1999 10:06 AM
Computer Modeling for Evaluating Trabecular Bone Mechanics 423
V. FE STRESS ANALYSIS
A. ANALYSIS ASSUMPTIONS
The most common FE analysis is linear static. A linear static analysis involves computation of stresses,
displacements, and strains due to static loads. The term static implies that the loads do not change with
time. This means that the effects of inertia and damping forces are considered insignicant. The term
linear refers to the case when nonlinearities due to plasticity, large deection, large strain, in-plane
effects, contact surfaces, creep and relaxation effects can be either linearized or ignored. Often, the
material is assumed to be isotropic (no directional dependence) and homogeneous (properties do not
vary spatially within the structure). An isotropic material has only two independent constants appearing
in the constitutive stressstrain relations. These two constants are Youngs modulus (E) and Poissons
ratio (). For the case of a uniaxial compression test, these two constants appear as follows:
(27.1)
(27.2)
where and are the strain and stress, respectively.
Youngs modulus gives a measure of the stiffness of the material. For trabecular bone the tissue
modulus (E
t
) has been reported to vary from 1 to 20 GPa.
27
For comparison, E varies between 197
GPa and 207 GPa for steel and steel alloys and is about 70 GPa for aluminum. A tensile load on
the specimen causes a lateral contraction, whereas a compressive load causes a lateral extension
this is called the Poisson effect. Poissons ratio relates the lateral or transverse deformations (
yy
,
zz
)
to axial or longitudinal deformations (
xx
) in a uniaxial test as follows:
(27.3)
The positive deniteness of strain energy puts a limit on these two constants:
E > 0 (27.4a)
FIGURE 27.10 FE models of unit cell of trabecular bone: (a) a coarse tetrahedral model (810 nodes, 1710
tetrahedrons); (b) the voxel-based brick model (52,293 nodes, 37,126 bricks).

xx xx
E
=
1

yy zz xx
E
= =
( )
1

yy xx
=
0266/ ch27/frame Page 423 Friday, October 22, 1999 10:06 AM
424 Mechanical Testing of Bone and the BoneImplant Interface
and
1 < <1/2 (27.4b)
The value = 1/2 implies elastic incompressibility, whereas negative values of are virtually
unknown, except for special cellular solids such as reentrant foams.
28
Values of for trabecular
bone have been reported to range from zero to just less than one.
29
Most structural FE analyses
assume a Poissons ratio value of 0.3. Poissons ratio for steels varies between 0.265 and 0.305,
and cork has a Poissons ratio close to zero.
B. THE UNIAXIAL COMPRESSIVE TEST
Using a linear, static analysis, the FE model is subjected to a compressive load by xing one face
and applying displacement boundary condition to the opposite face. The trabecular bone tissue is
generally assumed to have a uniform tissue modulus of E
t
= 5 to 10 GPa. The choice of the tissue
modulus E
t
(10 GPa) is arbitrary since the actual tissue modulus can be estimated by simply scaling
the apparent modulus (E
a
) obtained from FE analysis to the E
a
obtained experimentally. This will
be covered in a later section.
Whereas the tissue modulus is a material property, the apparent modulus is dened as
(27.5)
where E
a
is the apparent modulus, RF is the total reaction force, A
f
is the area of the face of the
unit cell upon which the displacement boundary conditions are applied, and is the applied strain
(typically 1%). To calculate the apparent modulus, only the reaction force on the bone is generally
considered. This is a reasonable assumption since the trabecular bone tissue is much stiffer than
other constituents present (e.g., marrow).
C. BOUNDARY CONDITIONS
To compare the results of FE analysis to results obtained experimentally, it is important to ensure that
the boundary conditions (BCs) used for the FE model accurately represent the experimental test setup.
For example, to simulate a 1% strain compression test with zero friction between the test platens and
the bone, the nodes on one face of the FE model should be prescribed using a displacement corre-
sponding to 1% strain ( = 0.01), while the displacements of the nodes on opposite face should be
constrained only in the loading direction. The remaining model nodes should be unconstrained.
D. FE ALGORITHM
The main computation in a linear displacement-based FE analysis is to solve the following set of
linear simultaneous equations:
[K]{U}={R} (27.6)
where [K] is the global stiffness matrix, {U} is the unknown displacement vector, and {R} is the
applied load vector. A nonstationary iterative scheme such as preconditioned conjugate gradient
(PCG) method can be efciently implemented to solve Equation 27.6. The advantage of this method
is that it converges to the solution very fast. A preconditioner such as the diagonal preconditioner
(main diagonal of [K]) accelerates the solution convergence. However, if the main diagonal of [K]
does not have much variability as when all the voxels are of identical shape and size, another
E
RF
A
a
f
=

0266/ ch27/frame Page 424 Friday, October 22, 1999 10:06 AM

Computer Modeling for Evaluating Trabecular Bone Mechanics 425
preconditioner known as incomplete Cholesky (IC) factorization should be used. The number of
computations per iteration are greater for IC factorization, but this extra work is more than repaid
by a reduction in the iteration count.
30
Many commercial FE packages are readily available. Cosmos/M (Structural Research and
Analysis Corp., V.2.0, Los Angeles, CA) can analyze models with up to 256,000 nodes/elements,
and requires a minimum of 128 Megabytes of RAM (random access memory). For smaller models
(up to 64,000 node/elements) the memory required is 32 Megabytes. Other commercial packages
include ADINA, ANSYS, PATRAN, NASTRAN, and MARC. A World Wide Web address that
lists a number of popular packages is <http://skyscraper.fortunecity.com/copland/949/fem-
info/femsivu1.htm>.
E. VALIDATION OF FE MODEL: CONVERGENCE
A continuous structure has innite degrees of freedom. On the other hand, an FE model of the
same structure will have nite degrees of freedom; hence, an FE model will be stiffer than the
original structure. As the number of elements in the FE model is increased, it becomes less stiff
and closer to the original structure. However, increasing the number of elements also increases the
computational cost. Hence, the minimum number of elements should be determined for a desired
accuracy. One criterion for determining the minimum number of elements (or FE model resolution)
is the convergence of apparent modulus. Another important numeric quantity is von Mises stress,
which gives a measure of deformation energy in a structure as follows:
(27.7)
(27.8)
where
VM
is the von Mises stress (Equation 27.7), U
def
is the deformation energy, and
i
, i = 1,2,3,
are the principal normal stresses. The von Mises stress can also be used to determine the validity
and convergence of the FE model, and is often used as a yield failure criterion. Another common
failure criterion is called the Tresca criterion.
F. VALIDATION OF FE MODEL: FE ANALYSIS
The validity of the FE analysis scheme can be checked using the following equation.
17
(27.9)
where is the average tissue stress,
a
is the apparent stress (
a
= RF/A
f
), and V
f
is the bone
volume fraction (BVF).
Plate 27.6 FE STRESS ANALYSIS APPLIED:
TETRAHEDRAL MODEL CONVERGENCE
Figures 27.11a and b and show the tetrahedral FE model of the unit cell and the deformed structure,
respectively. An FE model should be able to capture the anatomical details of the original specimen
and should also satisfy some convergence criterion. In order to determine the appropriate mesh resolution
required, nine FE models with increasing mesh resolutions were generated from the same MC triangular

VM
=
( )
+
( )
+
( )
1
3
1 2
2
2 3
2
3 1
2

U
E
def VM
=
+
( )
3
2
1 2

t
a f
V
=
1

t
0266/ ch27/frame Page 425 Friday, October 22, 1999 10:06 AM
426 Mechanical Testing of Bone and the BoneImplant Interface
surface mesh model (Table 27.1 and Figure 27.11). Figure 27.12a shows how the resolution improves
as the number of elements in the FE mesh is increased. Since the mesh resolution is proportional to
the cube root of the number of elements, the rate of improvement decreases with an increase in the
number of elements. The constant of proportionality between mesh resolution and number of elements
will depend upon the bone volume fraction. As the mesh resolution of the FE models was varied from
32 to 107 m, the discrepancy in the BVF was found to be about 2%. In this example the tetrahedral
mesh generation time increases linearly as the number of tetrahedral elements are increased.
By using COSMOS/M linear static analysis, each FE unit cell model was subjected to a compressive
load by xing one face and applying a 1% strain = 0.01) to the opposite face. The trabecular bone
was assumed to have a constant tissue modulus of 10 GPa. The computational time for the linear static
analysis exhibited a bimodal solution time with respect to the number of elements. The computational
time increased signicantly for models with more than 25,000 elements.
TABLE 27.1
Details of Nine FE Models
Label
No. of
Nodes
No. of
Tetrahedrons
Mesh Density
(No. of Tetrahedrons
per mm
3
of
Bone Volume)
A-1
A-2
A-3
A-4
A-5
A-6
a
A-7
A-8
A-9
810
1,891
2,522
3,839
5,402
7,642
9,820
13,040
15,641
1,700
4,456
6,208
10,208
15,056
22,450
30,346
41,337
50,897
3,911
10,251
14,281
23,483
34,635
51,645
69,809
95,093
117,085
a
The convergent model chosen for detailed analysis.
Source: Saxena, R. et al., Comput. Methods Biomech. Biomed.
Eng., 2, 287, 1999. With permission.
FIGURE 27.11 (a) Undeformed FE 4-node, tetrahedral mesh of unit cell shown with constrained (left) and
displaced nodes (model A-6) and (b) deformed FE mesh (displacement magnication 5). (From Saxena, R.
et al., Comput. Methods Biomech. Biomed. Eng., 2, 287, 1999. With permission.)
0266/ ch27/frame Page 426 Friday, October 22, 1999 10:06 AM
Computer Modeling for Evaluating Trabecular Bone Mechanics 427
Figures 27.12b and c show the dependence of apparent modulus and average von Mises stress on the
number of tetrahedral elements. Both plots converge for a model with about 22,000 elements. For the
converged model (A-6), an apparent modulus value of 61 MPa is obtained, which corresponds to a
mesh resolution of 42 m or a mesh density of approximately 50,000 elements/mm
3
of bone tissue
volume. Figure 27.12b shows apparent modulus plots obtained for compressive strains applied in two
perpendicular directions denoted by (*) and (o). Note that the moduli values in the two directions are
different, and indicate that the unit cell has different stiffnesses in the two directions. The convergence
trend, however, is similar in the two cases.
For the convergent FE model (A-6), the ratio is 12.427 and 1/V
f
is 12.421 which provides
additional validation for the FE model and the computational scheme.
The actual experimental apparent modulus value of the original specimen (11 11 11 mm) was 173
MPa.
31
In this case, the discrepancy between the experimental and numerical results is likely due to a
FIGURE 27.12 Results for tetrahedral models: (a) FE mesh resolution improves with increasing number of
tetrahedrons. (b) Convergence plot and directional dependence of apparent modulus. Approximately 22,000
elements produced a converged result. This corresponds to a mesh density of about 50,000, four-node tetrahedral
elements for the 1.5 1.5 2.0 mm
3
(9.66% bone volume fraction) model examined. (c) The average (over all
the tetrahedra) von Mises stress showed similar convergence characteristics as the apparent modulus. (a and c,
from Saxena, R. et al., Comput. Methods Biomech. Biomed. Eng., 2, 287, 1999. With permission.)

t a
0266/ ch27/frame Page 427 Friday, October 22, 1999 10:06 AM
428 Mechanical Testing of Bone and the BoneImplant Interface
number of factors, most notably the assumption of a 10-GPa constant isotropic tissue modulus and
specimen size. As mentioned earlier, the tissue modulus of trabecular bone has been reported to vary
from 1 to 20 GPa. Therefore, the apparent modulus for a linear analysis will lie in the range of 6 to
120 MPa. Furthermore, only a small subvolume (1.5 1.5 2.0 mm
2
) of the original one cubic
centimeter (10 10 10 mm
3
) mechanical test specimen was analyzed. Moreover, this subvolume had
a bone volume fraction of 9.66% compared with 11% in the original experimental test specimen, and
would therefore be expected to yield a lower apparent stiffness. Assuming a cubic relationship between
E and density, one can compute a density adjusted E value for the numerically analyzed subvolume.
In this case, the adjusted value is 90 MPa, which is closer to the experimental value (173 MPa).
G. COMPUTATION OF STRESSES AND STRAINS
Besides computing the apparent modulus, another important objective of FE analysis is to determine
the local stresses and strains. Noteworthy in this regard is the fact that a very small apparent strain
may cause large local stresses and strains. Such stress and strain concentrations are hypothesized
to result in local failure of the trabecular bone lattice.
17-19
In addition, these local stresses and strains
are very important parameters from the point of view of bone remodeling. Normalized local stress
or stress intensity is given by
(27.10)
Similarly, normalized local strain or strain intensity is given by
(27.11)
Strain shapes can also be calculated using Lodes parameter (LP):
32
(27.12)
where
i
= ln(1 + e
i
), i = 1, 2, 3, are principal logarithmic strains and e
i
, i = 1, 2, 3, are the principal
strains. Negative values of LP represent local constriction and positive values represent local attening.
Plate 27.7 FE STRESS ANALYSIS APPLIED:
STRESSSTRAIN INTENSITY AND POISSONS RATIO
For the model with a mesh density of 50,000 elements/mm
3
(FE model A-6), the normal stress varied
from 400 MPa (extreme compressive value) to 100 MPa (extreme tensile value). Compared with the
computed compressive apparent stress value (0.6 MPa), these stress values indicate that while the
apparent stress is quite low, the local tissue stresses are very high almost 200 times in tension and
600 times in compression (Figure 27.13a). A similar nding was evident in the strain plot where local
strains were seen to vary from 0.5% (extreme tension) to nearly 4.0% (extreme compression) and
indicates that the local compressive strains are four times the applied strain (Figure 27.13b). These
ndings agree with previous studies.
17
There was less than a 5% variation in the apparent modulus when Poissons ratio values were varied
between 0.1 and 0.4 (Figure 27.13c). Poissons ratio values greater than 0.4, however, had a much more
marked effect on the apparent modulus for which there was a roughly 25% increase for a corresponding
increase in Poissons ratio. This analysis (linear static, axial compression) indicates that the structural

int
=
xx
a

int
=
xx
applied
LP =
2
2

1 3
1 3
0266/ ch27/frame Page 428 Friday, October 22, 1999 10:06 AM
Computer Modeling for Evaluating Trabecular Bone Mechanics 429
stiffness of trabecular bone is relatively insensitive to Poissons ratio values less than or equal to 0.4.
However, the choice of Poissons ratio for values greater than 0.4 appears to be much more critical
from the standpoint of model predictions. The result of FE analyses are dependent upon other factors,
including the choice of boundary conditions and element type.
Plate 27.8 FE STRESS ANALYSIS APPLIED: TETRAHEDRAL vs. BRICK MODEL
The stress intensity and strain intensity values were signicantly higher for tetrahedral elements than
for brick elements. Figure 27.14a shows the variation in Lodes parameter (LP) values obtained for the
tetrahedral and brick element meshes. Both models show a similar LP trend, and the majority of elements
were associated with positive LP values. Positive LP values indicate that the 1% applied compressive
strain resulted in more attening than constriction of the elements.
A frequency histogram of the local stress values in the load direction is shown in Figure 27.14b. A
symmetric peak at zero stress value indicates the presence of both tensile and compressive stresses in
response to bending of the trabecular struts. The smaller peak in the negative stress region corresponds
to elements that are mainly axially compressed in the load direction. Note that there were a signicantly
lower number of brick elements that experienced compressive stresses.
FIGURE 27.13 Linear stress analysis results for convergent tetrahedral model (A-6): (a) local stress intensity;
(b) local strain intensity; (c) dependence of apparent modulus on Poissons ratio. (From Saxena, R. et al.,
Comput. Methods Biomech. Biomed. Eng., 2, 287, 1999. With permission.)
0266/ ch27/frame Page 429 Friday, October 22, 1999 10:06 AM
430 Mechanical Testing of Bone and the BoneImplant Interface
H. DETERMINATION OF TISSUE MODULUS
The experimental tissue modulus of trabecular bone ( ) can be estimated by scaling the FE
determined value ( ) to the apparent modulus obtained experimentally ( ). Assuming similar
specimen geometry and similar boundary conditions, the tissue modulus is obtained as follows:
(27.13)
VI. MORPHOLOGIC ANALYSIS
The term morphology refers to the study of form and structure of an organism. It was mentioned
earlier that the mechanical properties (apparent modulus and strength) of trabecular bone are
strongly related to the apparent density. Typically, the mechanical properties are related to density
using a power function (y = ax
b
).
2-6,27
However, the value of the exponent (b) may vary signicantly
and power functions relating modulus and strength to apparent density generally only explain 60
to 80% of the variance in the data.
5
The unexplained variance has been hypothesized to be due to
variations in the architecture of the trabecular bone lattice. Therefore, a central objective of
morphological analysis is to obtain additional structural indexes which can be used to improve
predictions of bone mechanical properties. In the following sections, several standard methods to
quantify the architecture of a trabecular bone are presented.
33-38
A. THE MEAN INTERCEPT LENGTH (MIL) METHOD
In the MIL method, the number of intersections between a linear grid and bonemarrow interface
are counted as a function of the orientation of the grid (Figure 27.15a). The MIL function is given by
(27.14)
where L is the total line length, I is the number of intersections, and is the orientation of the
grid. Whitehouse
35
showed that the MIL plotted as a radius at the angle of measurement t the
FIGURE 27.14 Comparison of tetrahedral and brick models: (a) variation of LP values for the two mesh
models; (b) frequency plot of local normal stress. Normalized Number of Elements = Number of Elements/
Total Number of Elements.
E
t
exp
E
a
FE
E
a
exp
E
E
E
E
t
a
a
t
exp
exp
=
FE
FE
MIL

( ) =
( )
L
I
0266/ ch27/frame Page 430 Friday, October 22, 1999 10:06 AM
Computer Modeling for Evaluating Trabecular Bone Mechanics 431
equation of an ellipse very closely. Harrigan and Mann
36
generalized this result in 3D to an ellipsoid
and dened a material anisotropy tensor (M) given by
(27.15)
where n is the unit vector, in the direction in which the measurement was made.
Cowin
39
dened a fabric tensor (H) which is related to M by
(27.16)
Along any particular direction H will increase with an increase in modulus value in that direction.
Thus, the extreme maximum and extreme minimum values of H correspond to the extreme maxi-
mum and extreme minimum values of modulus.
Plate 27.9 MORPHOLOGICAL ANALYSIS APPLIED: MIL METHOD
The MIL method was applied to compute the morphological parameters of the trabecular bone unit
cell model. By using a 3D version of the directed secant method, the binary image of the unit cell was
scanned by a 3D test grid at randomly determined orientations ( and

rotations about x-axis and
z-axis, respectively) using analysis specications given in Table 27.2. The intersections were recorded
when the binary value of the current voxel differed from that of the previous voxel. Morphological and
anisotropy parameters were then computed from intersection data using on the parallel plate model (see
Table 27.2). The locus of the end points of the MIL vectors issuing from a common center was plotted
and tted to an ellipsoid of general formula:
(27.17)
where A F are the ellipsoid coefcients and n
i
are the direction cosines of the MIL vector L.
Equation 27.17 is identical to Equation 27.15. The material anisotropy tensor [M] is given by
(27.18)
FIGURE 27.15 Principles for determining architectural anisotropy: (a) MIL measurement; (b) VO measure-
ment; and (c) SVD and SLD measurement.
1
2
L n M n =
H M =
1 2
An Bn Cn Dn n En n Fn n L
1
2
2
2
3
2
1 2 1 3 2 3
2
1 + + + + + =
M [ ]
A D E
D B F
E F C
=
0266/ ch27/frame Page 431 Friday, October 22, 1999 10:06 AM
432 Mechanical Testing of Bone and the BoneImplant Interface
A multivariate linear least-square tting technique was used to t the data by solving the linear system
(27.19)
where N contains the projection data constituting of n
i
, X is the column vector of the ellipsoid
coefcients, and R is a column vector of 1/L
2
. The solution is formulated as
(27.20)
The resulting 3D ellipsoid is obtained by tting Equation 27.17 to the locus of experimental MIL
vectors, and is illustrated in Figure 27.16.
TABLE 27.2
MIL Results for Unit Cell Model of Trabecular Bone
12
Analysis Specications
Total grid size 77 77 98
Sphere centered at (38, 38, 49)
Sphere radius 37
Number of MIL rotations 144
Pixel size (mm
3
/voxel) 0.02 0.02 0.02
Volume
Bone volume (BV) (mm
3
) 0.16
Total volume (TV) (mm
3
) 1.70
BV/TV (BVF) 0.0939
Intersections
Test line spacing (mm) 0.020
Number of test lines 4281
Average test line length (mm) 28.17
Total number of intersections 798109
Number of intersections per test line length (mm
1
) 1.5477
Morphology Parameters
Bone surface/bone volume (BS/BV) (mm
1
) 32.9464
Trabecular thickness (mm) 0.0607
Trabecular number (mm
1
) 1.5477
Trabecular spacing (mm) 0.5854
Anisotropy Parameters Experimental
Principal MILs (mm) 0.2411, 0.1994, 0.1063
Mean MIL (mm) 0.1268
SD of MIL (mm) 0.0305
Kurtosis 0.8397
Skew of MILs 0.9703
Orientation 1, () (11.38, 55.59)
Orientation 2, () (154.69, 151.05)
Orientation 3, () (61.91, 123.43)
Degree of anisotropy MIL 1 0.3230
Degree of anisotropy MIL 2 0.0940
Degree of anisotropy MIL 3 0.4169
N X R [ ]{ } = { }
X
N R
N N
T
T
{ } =
[ ] { }
[ ] [ ]
0266/ ch27/frame Page 432 Friday, October 22, 1999 10:06 AM
Computer Modeling for Evaluating Trabecular Bone Mechanics 433
B. THE VOLUME ORIENTATION (VO) METHOD
In the VO method, a local volume orientation is dened at a point. The local VO can be described as
the orientation of the longest intercept length through the point as shown in Figure 27.15b. A statistic
of a sample of local volume orientations is given by an orientation matrix called the VO fabric tensor (V).
C. THE STAR VOLUME DISTRIBUTION (SVD) METHOD
This method is related to both the VO method and the star volume method (Figure 27.15c). The
star volume is dened as the average unobscured volume at any point, and is given mathematically as
(27.21)
A star volume component is dened as the volume at a particular orientation at a point given by
(27.22)
An SVD fabric tensor (S) is described based on the variation of the star volume component
with orientation.
D. THE STAR LENGTH DISTRIBUTION (SLD) METHOD
Instead of computing the volume as in the SVD method, the length is computed in the SLD method.
The star length component (s) is given by
(27.23)
where the SLD fabric tensor is denoted by L.
FIGURE 27.16 MIL data fitted to an ellipsoid.

V
L =
1
3
3

V
L ( ) = ( )
1
3
3
s
n
L
n
i
i
( ) = ( )
=
1
1
0266/ ch27/frame Page 433 Friday, October 22, 1999 10:06 AM
434 Mechanical Testing of Bone and the BoneImplant Interface
van Rietbergen et al.
40
performed a 3D FE analysis of trabecular bone specimens obtained from
the single vertebral body of a whale, and computed the corresponding fabric tensors using the
above four methods. Compliance matrices (inverse of stiffness) were computed from the fabric
tensors. These were compared with the compliance constants determined from the FE analysis and
excellent correlations were found (Table 27.3). Correlations for volume-based fabric measures were
found to be slightly better than for MIL-based measures of fabric.
VII. SUMMARY
Recent advances in imaging techniques, 3D visualization software, and efcient FE analysis algo-
rithms have made it possible to perform large-scale investigations of the apparent mechanical
behavior of trabecular bone. FE analysis results indicate that small values of apparent strains can
lead to very large values of local stresses and strains in the trabecular bone lattice. Such stress and
strain concentrations are hypothesized to produce local failures. The effects of architecture on the
apparent mechanical properties of trabecular bone can be studied utilizing 2D and 3D morphologic
analysis techniques on the 3D models of the trabecular bone. By utilizing both experimental and
numeric approaches, the tissue modulus value can be estimated. In the future, higher-resolution
CT images and efcient FE mesh generation schemes will make anatomically precise numerical
analysis a routine and powerful tool to study the mechanical behavior of complex lattice structures
such as trabecular bone.
REFERENCES
1. McElhaney, J.H., Fogle, J.L., Melvin, J.W., et al., Mechanical properties of cranial bone, J. Biomech.,
3, 495, 1970.
2. Carter, D.R. and Hayes,W.C., The compressive behavior of bone as a two-phase porous structure,
J. Bone Joint Surg., 59-A, 954, 1977.
3. Gibson, L.J., The mechanical behavior of cancellous bone, J. Biomech., 18, 317, 1985.
4. Rice, J.C., Cowin, S.C., and Bowman, J.A., On the dependence of the elasticity and strength of
cancellous bone on apparent density, J. Biomech., 21, 155, 1988.
5. Keller, T.S., Predicting the compressive mechanical behavior of bone, J. Biomech., 27, 1159, 1994.
6. Zhu, M., Keller, T.S., Moeljanto, E., and Spengler, D.M., Multiplanar variations in the structural
characteristics of cancellous bone, Bone, 15, 251, 1994.
TABLE 27.3
Correlations between Compliance Components
Obtained from the FE Analyses and Those
Predicted from MIL, VO, SVC, and SLD
MIL VO SVD SLD
0.968
0.924
0.967
0.973
0.959
0.981
0.974
0.958
0.982
0.980
0.957
0.982
Source: van Rietbergen, B. et al., J. Orthop. Res., 16, 23, 1998.
1
1
E

ij
i
E
1
G
ij
0266/ ch27/frame Page 434 Friday, October 22, 1999 10:06 AM
Computer Modeling for Evaluating Trabecular Bone Mechanics 435
7. Radon, J., On the determination of functions from their integrals along certain manifolds, Ber. Verh.
Saechs. Akad. Wiss. Leipzig, Math. Phys. Kl., 69, 262, 1917.
8. Keller, T.S., Moeljanto, E., Main, J.A., and Spengler, D.M., Distribution and orientation of bone in
the human lumbar vertebral centrum, J. Spinal Dis., 5, 60, 1992.
9. Hall, E.L., Computer Image Processing and Recognition, Academic Press, New York, 1979.
10. Trussell, H.J., Comments on Picture thresholding using an iterative selection method, IEEE Trans.
Syst. Man Cybern., 9, 311, 1979.
11. Harrigan, T.P., Jasty, M., Mann, R.W., and Harris, W.H., Limitations of the continuum assumption in
cancellous bone, J. Biomech., 21, 269, 1988.
12. Keller, T.S., Trabecular Bone Morphology and Analysis System (TBMAS), Department of Mechanical
Engineering, the University of Vermont, Burlington, 1998.
13. Lorensen, W.E. and Cline, H.E., Marching cubes: a high resolution 3D surface construction algorithm,
Comput. Graphics, 21, 163, 1987.
14. Schroeder, W., Martin, K., and Lorensen, B., The Visualization Toolkit An Object-Oriented
Approach to 3D Graphics, Prentice-Hall, Upper Saddle River, NJ, 1997.
15. Fyhrie, D.P., Hamid, M.S., Kuo, R.F., and Lang, S.M., Direct three-dimensional nite element analysis
of human vertebral cancellous bone, Trans. Orthrop. Res. Soc., 17, 551, 1992.
16. Hollister, S.J., Brennan, J.M., and Kikuchi, N., A homogenization sampling procedure for calculating
trabecular bone effective stiffness and tissue level stress, J. Biomech., 27, 433, 1994.
17. van Rietbergen, B.,Weinans, H., Huiskes, R., and Odgaard, A., A new method to determine trabecular
bone elastic properties and loading using micromechanical nite-element models, J. Biomech., 28,
69, 1995.
18. Mller, R. and Regsegger, P., Three-dimensional nite-element modelling of noninvasively assessed
trabecular bone structures, Med. Eng. Phys., 17, 126, 1995.
19. Saxena, R., Keller, T.S., and Sullivan, J.M., A three-dimensional nite element scheme to investigate
the apparent mechanical properties of trabecular bone, Comput. Methods Biomech. Biomed. Eng., 2,
287, 1999.
20. Guldberg, R.E. and Hollister, S.J., Finite element solution errors associated with digital image-based
mesh generation, ASME Bioeng. Conf., 28, 147, 1994.
21. Jacobs, C.R., Davis, B.R., Rieger, C.J., et al., Accurate quantication of cancellous bone tissue
modulus can be made using experimentally measured apparent stiffness and large-scale nite-element
modeling, Trans. ORS, 23, 111, 1998.
22. van Rietbergen, B., Ulrich, D., Pistoia, W., et al., Prediction of trabecular bone failure parameters
using a tissue failure criterion, Trans. ORS, 23, 550, 1998.
23. Sullivan, J.M., Jr., Charron, G., and Paulsen, K.D., A three-dimensional mesh generator for arbitrary
multiple material domains, Finite Elem. Anal. Des., 25, 219, 1997.
24. Lo, S.H., A new mesh generation scheme for arbitrary planar domains, Int. J. Num. Meth. Eng., 21,
1403, 1985.
25. Desai, C.S. and Abel, J.F., Introduction to the Finite Element Method, Van Nostrand Reinhold, New
York, 1972.
26. Spyrakos, C.C., Finite Element Modeling in Engineering Practice, West Virginia University Press,
Morgantown, 1994.
27. Rho, J.Y., Ashman, R.B., and Turner, C.H., Youngs modulus of trabecular and cortical bone material:
ultrasonic and microtensile measurements, J. Biomech., 26, 111, 1993.
28. Lakes, R., Foam structures with a negative Poissons ratio, Science, 235, 1038, 1987.
29. Keaveny, T.M., Borchers, R.E., Gibson, L.J., and Hayes, W.C., Theoretical analysis of the experimental
artifact in trabecular bone compressive modulus, J. Biomech., 26, 599, 1993.
30. Golub, G.H. and Ortega, J.M., Scientic Computing and Differential Equations: An Introduction to
Numerical Methods, Academic Press, San Diego, CA, 1992.
31. Zhu, M., Keller, T.S., and Spengler, D.M., Effects of specimen load-bearing and free surface layers
on the compressive mechanical properties of cellular materials, J. Biomech., 27, 57, 1994.
32. Lode, W., Versuche ber den Einuss der mittleren Hauptspannung auf das Fliessen des Metalle Eisen,
Kaupfer, and Nickel, Z. Phys., 36, 913, 1926.
33. Weibel, E.R. and Elias, H.E., Quantitative Methods in Morphology, Springer, Berlin, 1967.
34. Underwood, E.E., Quantitative Stereology, Addison-Wesley, Reading, MA, 1970.
0266/ ch27/frame Page 435 Friday, October 22, 1999 10:06 AM
436 Mechanical Testing of Bone and the BoneImplant Interface
35. Whitehouse, W.J., The quantitative morphology of anisotropic trabecular bone, J. Microsc., 101, 153,
1974.
36. Harrigan, T.P. and Mann, R.W., Characterization of microstructural anisotropy in orthotropic materials
using a second rank tensor, J. Mater. Sci., 19, 761, 1984.
37. Kuo, A.D. and Carter, D.R., Computational methods for analyzing the structure of cancellous bone
in planar sections, J. Orthop. Res., 9, 918, 1991.
38. Odgaard, A., Three-dimensional methods for quantication of cancellous bone architecture, Bone, 20,
315, 1997.
39. Cowin, S.C., Bone Mechanics, CRC Press, Boca Raton, FL, 1989.
40. van Rietbergen, B., Odgaard, A., Kabel, J., and Huiskes, R., Relationships between bone morphology
and bone elastic properties can be accurately quantied using high-resolution computer reconstruc-
tions, J. Orthop. Res., 16, 23, 1998.
0266/ ch27/frame Page 436 Friday, October 22, 1999 10:06 AM
Section III
Methods of Mechanical Testing
of the BoneImplant Interface
0266/ ch27/frame Page 437 Friday, October 22, 1999 10:06 AM
0266/ ch27/frame Page 438 Friday, October 22, 1999 10:06 AM

439

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

28

Factors Affecting the Strength
of the Bone-Implant Interface

Brodie E. McKoy, Yuehuei H. An, and Richard J. Friedman

CONTENTS

I. Introduction ..........................................................................................................................439
II. Implant and Local Factors ...................................................................................................440
A. Implant Surface Congurations .....................................................................................440
B. Implant Surface Coatings...............................................................................................443
C. Implant Geometry ..........................................................................................................444
D. Material Properties of Implant .......................................................................................445
E. Quality of Cement and Cementing Procedures .............................................................445
F. Particle Debris ................................................................................................................446
G. Implant Design and Stress Shielding.............................................................................447
III. Medications and Other Local Factors..................................................................................448
A. Systemic Hormones........................................................................................................448
B. Nonsteroidal Anti-Inammatory Drugs .........................................................................448
C. Irradiation .......................................................................................................................449
D. Electromagnetic or Electrical Stimulation.....................................................................449
E. Local Use of Growth Factors or Hormones ..................................................................450
F. Other Factors ..................................................................................................................451
IV.

In Vitro

Factors.....................................................................................................................451
A. Storage before Testing....................................................................................................451
B. Specimen Preparation.....................................................................................................451
V. Surgical and Hospital Factors ..............................................................................................452
A. Surgical Skill ..................................................................................................................452
B. Complications.................................................................................................................452
C. Postoperative Care..........................................................................................................453
VI. Patient Factors ......................................................................................................................453
A. Age .................................................................................................................................453
B. Lifestyle..........................................................................................................................454
C. Inammation...................................................................................................................454
D. Osteopenia or Osteoporosis ...........................................................................................455
VII. Summary ..............................................................................................................................455
References ......................................................................................................................................455

I. INTRODUCTION

The modern era of total joint arthroplasty began in the 1960s with Sir John Charnleys development
of total hip replacement consisting of a stainless steel femoral head articulating with a polyethylene
acetabular implant, both secured to supporting bone by polymethylmethacrylate (PMMA) cement.

0266/ ch28/frame Page 439 Friday, October 22, 1999 10:09 AM

440

Mechanical Testing of Bone and the BoneImplant Interface

Now, 40 years later, 15- to 20-year success rates close to 90% are being reported for total hip and
total knee arthroplasty. The high success rate of this early technique of xation makes it the gold
standard to which all subsequent methods should be compared. Infection and material failure
presented early problems; however, these complications have been greatly reduced, and the surgeries
have gained widespread popularity with consistently reliable results.
Despite great strides in the development of better component designs, biomaterials, and surgical
techniques, the life span of cemented total joint-arthroplasty is limited, among other things, by the
long-term mechanical properties of PMMA and polyethylene debris. The clinically signicant
loosening rates of cemented implants, especially in younger, active persons, have led many inves-
tigators to pursue methods of cementless xation. New biomaterials and innovative designs are
two promising approaches which will improve this type of xation.
As research and development of cementless xation continues, a great deal of attention is being
focused on the boneimplant interface and the factors affecting its strength. Several methods of
achieving implant xation with long-term stability have been introduced, including biological
xation, achieved with ingrowth into a porous coating; press-t xation, in which a nonporous-
coated implant is surrounded by brous tissue; and chemical xation, where a bioactive ceramic
chemically bonds to surrounding bone. All of these approaches have merit and each approach has
its proponents and detractors.

1

Early results with each method have been promising, but only time
will reveal long-term complications and rates of success.
Not only do the implant and local conditions affect the boneimplant interface, but numerous
other factors exist which may have an impact on long-term xation in either a positive or negative
way. Modalities such as radiation, hormones, and various medications have the potential to interfere
with implant xation. Numerous growth factors have been studied and used in tumor and spinal
surgery to promote bone ingrowth. Some of these growth factors used locally may enhance bonding
between the prosthesis and bone. Surgical technique as well as various patient factors including
age and lifestyle also affect the all-important boneimplant interface.

II. IMPLANT AND LOCAL FACTORS
A. I

MPLANT

S

URFACE

C

ONFIGURATIONS

The primary function of the boneimplant interface is to provide a safe and effective load transfer
from the prosthesis to the bone. Cementless implants may be xed to bone by both osseointegration
and bone ingrowth. Branemark

2,3

originally coined the term

osseointegration

in the dental literature
to refer to the intimate contact of bone tissue with the surface of a titanium implant.

2

The term
bone ingrowth refers to the actual bone formation within the porous structure of the implant.

4,5

The
surface congurations of the implant play a role in initial stability and type of xation achieved.

6

Porous-coated surfaces are used to obtain biological xation in cementless surgery. In recent
years, these have become increasingly popular in young, active patients.

7,8

Several different porous
surfaces have been developed with slight differences in the strength of xation to bone. Friedman
et al.

9

compared the histological and mechanical characteristics of arc-deposited titanium implants
in rabbits to one layer cobaltchromium beads, three layers of cobaltchromium beads, plasma-
sprayed cobaltchromium, and uncoated titanium alloy (Figure 28.1). The physiological response
to a porous-coated implant is similar to the healing cascade of cancellous bone defects, with newly
formed tissue lling in the void spaces of the porous material.

10,11

The initial stability of the
prosthesis determines the clinical success by allowing bony ingrowth instead of brous tissue. This
bony ingrowth gives signicant strength to the boneimplant interface. The shear strength at the
boneimplant interface after ingrowth equals or exceeds trabecular bone.

12-15

Galante found bone
ingrowth at 6 weeks in wire mesh surfaces to be as stable as a bone cement interface.

13

Too much relative motion between the implant and host bone and the quality of the initial press-
t in an uncemented arthroplasty are important factors leading to ingrowth of brous connective tissue

0266/ ch28/frame Page 440 Friday, October 22, 1999 10:09 AM

Factors Affecting the Strength of the Bone-Implant Interface

441

rather than bone into porous surfaces.

16,17

The ingrowth of bone is necessary for initial stability
and subsequent long-term xation.

18,19

Micromotion of as little as 100 to 500

m between bone
and implant is sufcient to inhibit bone ingrowth and result in a brous membrane forming between
the two, resulting in a mechanically unstable implant.

20

Porous coatings used to promote bone ingrowth include both ceramics and metals. Currently,
the majority of implants use porous metal surfaces, created either by sintering cobaltchrome
powder or beads or by diffusion bonding of titanium wire mesh.

21,22

In addition, plasma spraying
material onto metal surfaces is another method used to produce a porous-like surface. The surface
of implants manufactured in this manner is highly textured, but the actual porosity is limited.

23,24

Pore sizes have been extensively studied and found to affect the strength of xation. Several
investigators have studied pore sizes of less than 100

m and found that increasing pore size
correlated with increasing strength of xation.

25,26

One study showed no relationship between pore
sizes in the range of 150 to 400

m and interface shear strength.

27,28

From these and other studies,
it can be concluded that the optimal range for pore size to obtain maximal strength at the
boneimplant interface is from 100 to 400

m.

FIGURE 28.1 (

A) Scanning electron micrographs of five types of porous coatings. The bar represents 1 mm.
Cr/Co-1 = one layer of cobaltchromium beads, Cr/Co-3 = three layers of cobaltchromium beads, Ti-Arc =
arc-deposited titanium, Cr/Co-PS = plasma-sprayed cobalt chromium, and Ti alloy = grit-blasted titanium
alloy. (B) Shear strengths of the same implants at 6 and 12 weeks. Bar denotes SD of mean. (From Friedman
R.J., et al.

J. Orthop. Res.

, 14, 456, 1996. With permission.)

0266/ ch28/frame Page 441 Friday, October 22, 1999 10:09 AM

442

Mechanical Testing of Bone and the BoneImplant Interface

Biegler et al.

29

studied the effects of porous coating and loading conditions on femoral stem
stability. This group examined the bone prosthesisinterface behavior under different load types using
nite-element analysis. Torsional loads such as stair climbing were found to contribute to a greater
degree of implant micromotion than stance loading. It was concluded that contact at the boneimplant
interface was more dependent on load type than either surface coating or implant geometry.
Smooth surface implants have been used in cemented arthroplasties. Some investigators have
used the term

press-t

to describe the xation of a smooth prosthesis without porous ingrowth or
cement. However, press-t is an engineering term which denotes a mechanical joining of two parts
based on contact pressure. As Figure 28.2 illustrates, a nail driven into a block of wood is an
example of press-t.

30

Thus, either a smooth or porous structure can be press-t by the true
meaning. A brous tissue may form around smooth implants in cementless surgery largely as a
result of low initial stability.

31

It is this intervening tissue that may weaken the boneimplant
interface. Lee et al.

6

compared initial stability of uncemented porous-coated and smooth implants
in synthetic femurs. This group found that porous-coated implants had greater initial stability with
less axial micromotion compared with smooth implants. These ndings suggest that surface con-
guration plays a major role in initial stability and thus strength of the boneimplant interface.
Impaction bone-grafting has been studied by several investigators in different models with
varying results. Segawa et al.

32

found tibial components placed with impaction grafting to have
greater micromotion and less initial stability. Meding et al.

33

found worrisome results in femoral
stems xed with cement and impaction grafting. However, several groups have shown increased
initial stability and strength of the boneimplant interface.

34-36

Longer-term

in vivo

studies should
be undertaken to evaluate the compaction technique and results.
Grooves and other surface modications have been studied in cementless press-t techniques
as well as cemented techniques.

37,38

These prostheses rely on a macrointerlock with bone or cement.
Wang et al.

39

found that as surface roughness increased there was a corresponding increase in
interfacial shear strength. This group also showed surface nish to be more important in
stemcement interface shear strength than cement interface porosity.

40

Many investigators feel these

FIGURE 28.2

A nail driven into a block of wood with a predrilled hole is an example of press-fit.
Carpenters predrill holes to prevent the wood from splitting when nailed. (Adapted from Blaha, J. D., Press-
t femoral components, in

The Adult Hip

, Callaghan, J. J., et al., Eds., Lippincott-Raven, Philadelphia, 1998.)

0266/ ch28/frame Page 442 Friday, October 22, 1999 10:09 AM

Factors Affecting the Strength of the Bone-Implant Interface

443

types of surface congurations may be superior to porous coating for two reasons: (1) some retrieval
data show that less than 30 to 60% of porous coated specimens are occupied by bone and (2) these
congurations do not lessen the fatigue strength as porous coatings might. The initial strength of
the boneimplant interface appears to be good with these surface modications. However, there is
considerable debate concerning the lasting implant stability.

B. I

MPLANT

S

URFACE

C

OATINGS

Various surface coatings have been extensively studied and can markedly affect the strength of the
boneimplant interface. Chemical xation is a technique where bioactive ceramics chemically bond
to surrounding bone.

41

Ceramics are a group of materials made from ne powder. In the biomaterials
literature, ceramics comprise all nonmetallic and inorganic materials. Oxide ceramics and calcium
phosphate ceramics are the two materials most often used in joint arthroplasty.
Oxide ceramics have found a use in joint surgery as femoral head components which articulate
with a polyethylene acetabular component and acetabular liners. In contrast, calcium phosphate
ceramics have been used as bone substitution material and coatings for joint prostheses.

42

Tricalcium
phosphate and hydroxyapatite are the two most common calcium phosphate ceramics used.
Hydroxyapatite has been studied for over 20 years and is referred to as a bioactive ceramic
because of the biological response it generates after implantation.

43

The biocompatibility of this
substance has been documented in many

in vivo

and

in vitro

studies.

44,45

Originally studied for
dental and oral surgery applications, hydroxyapatite has been used to coat metal prostheses to
provide a chemical bond between the implant and bone.

46,47

This coating allows an intimate union
between the bone and prosthesis as seen in Figure 28.3.

48

Many methods have been developed to apply hydroxyapatite powder to a metal surface. The
technique of plasma spraying has been used most extensively.

49

This is done by introducing the
hydroxyapatite powder into a ame that directs the particles for deposit on the metal surface. Ion
sputtering, dip-coating-sintering, and hot isostatic pressing are other coating methods being investigated.
Any coating applied to a metal may affect the fatigue strength of that metal. Thin coatings,
between 30 and 90

m, are recommended, because they do not affect the mechanical properties
of the substrate metal.

50

Mechanical testing of the coating applied to metal has demonstrated that
the mechanical properties of the coating are greatly improved in shear and tension, with shear
increased three to four times.

51

The hydroxyapatite may form a chemical bond with the titanium
or chromium-cobalt alloy to which it is applied.

50

FIGURE 28.3

Photomicrograph of an implant coated with hydroxyapatite. Note the intimate union of
trabecular bone and the prosthesis with the 50-

m hydroxyapatite layer.

48

0266/ ch28/frame Page 443 Friday, October 22, 1999 10:09 AM

444

Mechanical Testing of Bone and the BoneImplant Interface

The mechanical advantages of hydroxyapatite-coated implants have been studied in various
animal models.

52

When hydroxyapatite-coated implants are compared with noncoated devices of
the same design, the maximum xation strength is shown to be increased and the time required to
achieve adequate xation strength is shortened.

53,54

Thus, the implant forms a faster and stronger
bond to surrounding bone. Histological studies have conrmed these ndings. Sections through
the boneimplant interface of a hydroxyapatite-coated implant show increased amounts of
boneimplant contact through preferential deposition of new bone both on the surface of the implant
as well as on the host bed.

55

This interface does not contain a brous tissue membrane as seen in
some smooth implants.
Numerous other studies have documented that hydroxyapatite coatings increase the amount of
bone ingrowth and osseointegration.

56,57

The maximal interface strength with these coatings is
achieved at 6 weeks, compared with 12 to 16 weeks for porous ingrowth. The strength of the
interface itself has been shown to increase more than threefold with hydroxyapatite coating.
The strength of the boneimplant interface with hydroxyapatite coating has been investigated
by several groups. Mechanical testing to failure in shear has demonstrated that failure occurs in
the bone, but not at the hydroxyapatitebone interface or between the substrate metal and
hydroxyapatite.

58

The strong bond made to bone and the increase in ultimate strength are sec-
ondary to the fact that hydroxyapatite is osteophilic and provides a scaffolding on which bone
can proliferate and bond.

59

Small gaps between the implant and coating or defects in the coating
did not alter these ndings.

60

Hydroxyapatite coating can also alter the strength of the bone-implant interface by its ability
to ll defects. This bioactive coating has been shown to ll in defects up to 2 mm in size with bony
ingrowth.

61,62

Thus, hydroxyapatite increases the contact between the prosthesis and bone, leading
to a greater interface stability. In contrast, a brous membrane will form in porous material if the
gap is greater than 300 mm, leading to a much weaker interface.

63

Numerous investigations have been carried out to test the strength of hydroxyapatite prostheses
in the canine model.

64,65

These implants withstood higher compression, shear, and tensile forces
than did noncoated controls. In addition, the bonding occurred in a shorter time. Bone scans
performed at 6 months showed an equilibrium between implant integration and bone metabolism.
Histological examination also showed superior results, with no intervening brous tissue between
the bone and hydroxyapatite coating. Again, close apposition between the bone and implant were
not required to obtain a stable interface.
Another type of coating applied in arthroplastic surgery using cement is a precoating of the
implant with PMMA. Mounting evidence points to the cement mantle as being the reason for
failures in these types of xations. Debonding at this interface appears to be a key problem.

66

To
circumvent this, many groups have begun precoating stems with cement which allows a chemical
bond to the cement mantle. This technique markedly improves the strength of the interface.
Bone marrow cells seeded in a bone ller have been shown to have osteoconductive properties

in vivo

.

67

This bone tissue engineering is in the early stages of development. In the future, this may
prove useful in reconstruction surgery to treat large bone defects or to coat prostheses to promote
increased bone ingrowth into porous surfaces.

C. I

MPLANT

G

EOMETRY

Implant geometry varies between different manufacturers. The effect of different geometric patterns
on implant stability and strength of the boneimplant interface has been evaluated.

68,69

For cemented
prostheses, sharp edges should be avoided as these produce stress risers that may initiate a fracture
of the cement mantle. Noncircular shapes, such as rounded rectangles and ellipses, help improve
rotational stability of the implant with the cement mantle. The implant should ll at least 80% of
the cross section of the medullary canal with an optimal cement mantle of 4 mm proximally and
2 mm distally.

70

Implants that allow these dimensions of cement in neutral placement lessen the

0266/ ch28/frame Page 444 Friday, October 22, 1999 10:09 AM

Factors Affecting the Strength of the Bone-Implant Interface

445

chance of localized weakness and fracture of the cement mantle. For cementless xation, geometry
is important as well. The size and shape of femoral components should be optimal to provide both
axial and rotational stability.
Schneider et al.

71

evaluated the degree of micromotion with different geometric-shaped femoral
components.



Three straight and one curved prostheses were submitted to

in vitro

tests in autopsy
specimens. Dynamic axial and torsional loads were applied. They found micromotion at the
boneimplant interface was smallest for the curved and highest for the straight femoral components.
Initial stability allows less micromotion and increases the bond between the prostheses and the
host bone. Prosthesis stability was found by Kelysey and Goodman

72

to be more important to
surgeons than ease of implantation and removal or cost.
The geometry of acetabular cups inuences stability. Ries et al.

73

studied the effect of cementless
cup geometry on strain distribution and press-t stability. This group found a nonhemispheric cup
to provide greater stability. The gradual transition from a hemisphere at the dome to a larger
peripheral dimension maximized peripheral strains.



The geometry of all implanted components on
both sides of the joint is important in stability of the prosthesis.

D. M

ATERIAL

P

ROPERTIES



OF

I

MPLANT

The material properties of the implant play a critical role in the life and strength of the
boneimplant interface. The properties of the material strongly inuence the design choices as
well as the eventual device performance. These properties are specic to each material used in
joint surgery.
Corrosion, the electrochemical dissolution of a metal, is one such property. Particles can be
generated by corrosion.

74

A second property, fretting, refers to small cyclic motions, usually less
than 100

m, of one surface relative to another. These two processes may be synergistic. Fretting
can disrupt the oxide lms which protect cobaltchrome and titanium-based alloys from corrosion.
This disruption may dramatically accelerate the corrosion rate.
Modular implants demonstrate fretting corrosion.

75

Modular implants often have different metals
in contact with each other. Levine and Staehle

75

reported that corrosion rates were increased when
CoCrMo was coupled with TiAlV vs. when it was coupled to itself. Another study reported less
corrosion of TiAlV when coupled with CoCrMo than with itself.

76

The amount of corrosion a
metal undergoes is specic to the metal itself. Fretting corrosion leads to the generation of particles
which may cause aseptic loosening and weakening of the boneimplant interface.
Another property of the implant which affects interface strength is the strength of the implant.
Implant fractures can cause motion and failure leading to decreased stability.

77,78

In comparing
different materials, the fatigue life is important. This is dened as the number of cycles needed to
cause failure at a xed cyclic stress. Sandborn et al.

79

evaluated a unicompartmental knee implant
retrieved after 11 years

in situ.

The specimen was a cast stainless steel femoral component which
had a fracture 1 cm from the anterior tip. Microscopic analysis showed a high inclusion content
and inhomogeneous grain size which likely contributed to the fracture.

E. Q

UALITY



OF

C

EMENT



AND

C

EMENTING

P

ROCEDURES

Since the early days of Sir John Charnley, the cemented prosthesis still remains the gold standard
of xation techniques. The cement used in arthroplasty is PMMA. This acrylic cement as well as
the methods used to prepare it have undergone extensive research in recent years. Both the quality
of cement as well as the cementing procedure affect the strength of the boneimplant interface.
A great deal of information has accumulated regarding the factors that either adversely affect or
enhance the use of PMMA. The cementing techniques have gone through several generations of
improvement. The most current techniques are referred to as third generation and include (1) canal
plugging; (2) reduced cement porosity; (3) canal preparation; and (4) cement column pressurization.

80

0266/ ch28/frame Page 445 Friday, October 22, 1999 10:09 AM

446

Mechanical Testing of Bone and the BoneImplant Interface

The presence of porosity in cement has been shown by many investigators to cause a decrease
in strength.

81-83

Porosity reduction has been shown to improve the fatigue life of PMMA. Cement
porosity can be decreased to less than 1% by mixing in a vacuum of approximately 500 mmHg.
Centrifugation also decreases the porosity considerably (Figure 28.4).

84

Both of these techniques
require additional equipment but they are easily performed and provide an increase of 10 to 20%
in strength. Cement column pressurization also has been shown to increase strength by up to
20% by reducing porosity.
The quality of the cement is another important factor in strength. Any impurities will affect
the fatigue strength of the compound. However, PMMA in general has not improved signicantly
when compared with the original cement used by Charnley. Much of the quality problems arise
with cement preparation and insertion of the mixture. Canal lavage helps to clear the area of
unwanted debris. Inclusion of blood or tissue can cause up to a 77% loss of strength in the
cement.

85

All of these methods improve the cement and help to strengthen the interface in
cemented prosthesis.
The analysis of the interfaces in cement prosthesis is complex. No true prosthesis-to-bone
interface exists because of the intervening PMMA. Instead, two interfaces are present, one between
the prosthesis and the cement and the other between the cement and the bone. This added variable
makes individual interface analysis difcult. Because of this, the strength of the cement is studied
most exhaustively and is considered by some to be an expanded interface.

F. P

ARTICULATE

D

EBRIS

Osteolysis, or periprosthetic bone loss, is a signicant cause of decreased strength of the
boneimplant interface. This condition is usually recognized roentgenographically either as diffuse
cortical thinning or as focal cystic lesions (Figure 28.5). The pathogenesis of focal osteolysis has
not been fully elucidated. In 1977, Wilbert and Semlitsch

86

were among the rst to suggest that a
macrophage response to particulate debris was an important factor leading to aseptic loosening.
The nature of the interface between the host and the implant was described by Goldring et al.

87

They described a synovial-like membrane at the bonecement interface in patients with a loose

FIGURE 28.4

Polished stained cement specimens hand-mixed, centrifuged, and vaccum-mixed demonstrat-
ing amount of porosity present. (From Wixson, R.L

., J. Arthroplasty

, 2, 144, 1987. With permission.)

0266/ ch28/frame Page 446 Friday, October 22, 1999 10:09 AM

Factors Affecting the Strength of the Bone-Implant Interface

447
total hip prosthesis. Large amounts of prostaglandin E2 (PGE2) and collagenase were found in this
membrane which was believed to produce bone resorption. Many investigators have since studied
these intercellular mediators and their relationship to aseptic loosening.
88-90
The generation of particulate debris and its migration into the synovial cavity and periprosthetic
space stimulates macrophage recruitment and proliferation. This is followed by phagocytosis and
secretion of various cytokines and intercellular mediators like interleukin-1 (IL-1) and prostaglandin
E2 (PGE2). These substances, in turn, stimulate osteoclastic bone resorption.
The particles may originate from various sites. In acetabular components, particles may come
from the articular surface of ultrahigh molecular weight polyethylene (UHMWPE), the nonarticular
side of UHMWPE, at the metal shell or at the xation screws. In the femoral component, the stem,
porous coating, articular surface, and modular connection may all be the culprit. In addition, surgical
tools, bone, remnants from the surface processing of the prosthetic device, and the catalyst used
in the processing of UHMWPE can all be sources of particulate debris. The predominant particle
appears to be UHMWPE, mostly from the articular surface.
The plethora of literature on particulate debris and its relationship to periprosthetic loosening
illustrates its importance. Osteolysis around the implant signicantly weakens the boneimplant
interface. As research continues in this area, understanding of the interaction of this debris and its
relationship to the boneimplant interface will continue to increase.
G. IMPLANT DESIGN AND STRESS SHIELDING
An engineering-related problem of design in arthroplasty is stress shielding which might eventually
undermine the mechanical stability of the prosthesis. Implants are made from metallic materials
which are usually an order of magnitude, or more, stiffer than the bone which they contact. This leads
FIGURE 28.5 AP radiograph of the right hip of a patient with a total joint arthroplasty demonstrating
radiotranslucent periprosthetic osteolysis around the femoral component (see arrows).
0266/ ch28/frame Page 447 Friday, October 22, 1999 10:09 AM
448 Mechanical Testing of Bone and the BoneImplant Interface
to a large percentage of the load normally transmitted by the bone being borne by the prosthesis instead.
This shielding can lead to bone disuse atrophy.
91,92
Bone remodeling is very sensitive to stress.
In designing implants, the modulus of elasticity is a key component. A decrease in the modulus
leads to a decrease in stress in the stem and an increase in stress transmitted by the bone. In addition,
increasing the length and cross-sectional diameter of the stem increase the stress in the stem and
result in an increase in stress shielding. The bending stiffness of a stem is proportional to the fourth
power of the diameter, so a small increase in diameter leads to a large increase in rigidity.
Several investigators have identied the proximal medial cortex as the area most affected by
stress shielding in both cemented and cementless implants.
93,94
Because of this nding, some
cemented femoral components have been designed with a collar to load this medial area axially.
95
However, the role of a collar in preventing loosening has not been clearly established. In
cementless components, a collar is more controversial because it may prevent complete seating
of the implant.
In the acetabulum, ultrahigh-weight polyethylene causes peak stress to develop in the pelvic
bone. However, a metal-backed design with a polyethylene liner reduces the areas of high stress
and causes a more uniform distribution. The preservation of subchondral bone and thick-walled
cups leads to decreased peak stress levels and may cause stress shielding in this area. The implant
designs mentioned above play a crucial role in stress shielding and bone remodeling which deter-
mine the strength of the boneimplant interface.
III. MEDICATIONS AND OTHER LOCAL FACTORS
A. SYSTEMIC HORMONES
Systemic hormone deciencies may signicantly alter the strength of the boneimplant interface.
The hormone studied most extensively has been estrogen.
96
Total joint arthroplasties are often
performed in postmenopausal women so this issue is of extreme importance.
Martin et al.
96
studied the effects of estrogen deciency in the growth of tissue into porous
titanium implants.

In this study ve dogs were ovariectomized and ve dogs received a sham
operation. Later, 4 months after the operation, titanium implants were placed in the humerus of
each dog; 2 months later the implants were harvested and pushout tests and histological studies
were performed. The pushout strength of the implants from the ovariectomized dogs was 31% less
than the controls. The estrogen-decient dogs had increased brous tissue on histological exam.
Increased brous tissue at the boneimplant interface in estrogen deciency is a key determinant
to the decreased strength. Numerous investigators have implicated brous tissue in the failure of
implants.
97,98
Not only are implants which rely on bone ingrowth affected by estrogen deciencies,
but cemented and press-t implants are as well. The latter two of these implants rely on healthy,
dense bone to form a strong xation. Deciencies in estrogen cause an osteoporotic bone which
may not be suitable for xation. The effects of estrogen and other systemic hormones at this level
are tremendous. Clearly, more work needs to be done to dene better the exact effects of hormones
on the boneimplant interface.
B. NONSTEROIDAL ANTI-INFLAMMATORY DRUGS
Porous-coated implants rely on bone ingrowth to achieve implant stabilization. There is a concern
that the use of certain medications, including nonsteroidal anti-inammatory drugs (NSAIDS), may
adversely affect bone ingrowth.
99
Indomethacin, a commonly prescribed NSAID, is a potent inhib-
itor of fracture healing and bone remodeling in therapeutic doses.
100102
On a histological basis, the
process of bone ingrowth is similar to fracture healing and factors that inhibit one may inhibit the
other. NSAIDS are used clinically to inhibit heterotopic ossication. In addition, they are often
administered for postoperative pain control, inammatory arthritis, gout, and osteoarthritis.
0266/ ch28/frame Page 448 Friday, October 22, 1999 10:09 AM
Factors Affecting the Strength of the Bone-Implant Interface 449
Keller et al., using a rabbit model with quantitative histomorphometry, showed that indometha-
cin, compared to a placebo, caused a signicant decrease in the amount of ingrowth into a porous
titanium cylinder at 8 weeks, independent of pore size.
103
Follow-up studies with porous-coated
chromiumcobalt cylinders in the same animal model showed that therapeutic dosages of
indomethacin, aspirin, and ibuprofen all cause a statistically signicant decrease in bone ingrowth,
as compared with a control group.
104
A doseresponse effect was found for the rst two NSAIDS,
with a greater inhibitory effect resulting from high doses. Longo et al.
105
found the same results
with Naproxen.
Cook et al.
106
also studied the effects of indomethacin on xation and found the adverse effects
in interface shear strengths only at early periods.

This group evaluated the specimens at longer
postoperative intervals and determined the effects to be transient and minimal. Based on their
ndings, this group felt the possible benets of postoperative use of NSAIDS, such as indomethacin,
outweigh the minimal risk of transient decrease in attachment strength.
Until the mechanism of inhibition of bone ingrowth by NSAIDS is better dened and under-
stood, caution should be exercised when using these medications immediately after surgery in
patients with a porous-coated implant. If bone is inhibited from lling the pores, then brous tissue
will develop and ll this area instead. Once this has occurred, it is unlikely that bone will be able
to replace the brous tissue when the inhibitory medicines are withdrawn. As previously stated,
brous tissue weakens the interface between the implant and bone and can lead to early failure.
108
C. IRRADIATION
Not only is indomethacin used to prevent heterotopic bone formation in patients undergoing an
uncemented hip replacement, but irradiation is an alternative method. Of patients undergoing these
surgeries, 10 to 15% may develop heterotopic bone which can be prevented with 1000 rads of
cobalt irradiation.
109
Although this treatment has been shown to be effective, recent studies suggest
that it may weaken the interface of a porous-coated implant.
Wise et al.
110
implanted porous titanium ber metal plugs in dogs which were subsequently
radiated with 10 Gy given in four daily fractions of 2.5 Gy.

Pullout strength tests of the irradiated
implants were performed at 6 weeks survival. A 19% reduction was found in irradiated implants
compared with controls. Metabolic activity monitored with bone scans was also signicantly
reduced in the irradiated extremity compared with the nonirradiated side. No histological evaluations
of the boneimplant interface were performed.
Similar results have been produced by other investigators.
111-112
By using similar implants in
dogs, dosages of 1000 rads were found to decrease the xation strength early on by inhibiting
bone ingrowth. Histological and mechanical tests documented these ndings. A dose-dependent
effect on ingrowth was found. The strength of the interface at 2 and 4 weeks was unchanged
when the treatment consisted of only 500 rads; however, the amount of bone formation was
slightly decreased.
While studies are being performed to look at the effects of lower single doses, the long-term
effects of irradiation on the boneimplant interface are not known. Recommendations for patients
who are felt to require postoperative irradiation are for cemented arthroplasty or shielding of porous-
coated implants to minimize effects to the interface. Further studies are needed to help clarify the
long-term effects of irradiation on the strength of the boneimplant interface.
D. ELECTROMAGNETIC OR ELECTRICAL STIMULATION
In both experimental and clinical settings, electrical osteogenesis has been reported to be effective
on bone healing of nonunion or ununited fractures.
113-115
Several investigators have looked at the
effect of electromagnetic stimulation on bone ingrowth into porous implants.
116
By using canine
models, an acceleration of bone ingrowth by electrical stimulation has been demonstrated.
117-119
0266/ ch28/frame Page 449 Friday, October 22, 1999 10:09 AM
450 Mechanical Testing of Bone and the BoneImplant Interface
This increase in bony ingrowth undoubtedly leads to greater stability and strength at the
boneimplant interface.
Shinizu et al.
120
evaluated the effect of pulsing electromagnetic elds on bone ingrowth into
porous hydroxyapatite and porous tricalcium phosphate implants in rabbit tibiae. They quantitated
the biological response using a scanning electron microscope. The morphometrical ndings in the
hydroxyapatite pores showed a much greater amount of bone ingrowth in the pulsing electromag-
netic eld group compared with the controls at 3 to 4 weeks postimplantation. However, no
signicant difference was found in the tricalcium phosphate pores.
120
Histological ndings supported
the morphometric data. The differences in the two materials tested were felt to be secondary to the
differences in pore sizes of the two ceramics.
Several investigators have studied electrical elds to stimulate bone ingrowth.
121
In a double
blind study, Schutzer et al.
122
evaluated the effects of capacitively coupled electrical elds on bone
ingrowth into porous-surfaced prosthesis in dogs. The electrical eld was induced by an external
source delivering voltage through skin electrodes. No signicant growth into bone was seen at the
end of 6 weeks.
Electromagnetic stimulation appears to have a positive effect on the boneimplant interface.
The increased bone ingrowth caused by these forces is most striking at 3 to 4 weeks after surgery.
Capacitively coupled electrical elds have not been shown to improve ingrowth. Additional studies
are required to delineate the biochemical mechanisms by which increased osteogenesis may be
promoted by electromagnetic stimulation.
E. LOCAL USE OF GROWTH FACTORS OR HORMONES
Growth factors, such as platelet-derived growth factor (PDGF), broblast growth factor (FGF),
epidermal growth factor (EGF), and transforming growth factor (TGF-), are found in all tissues
and regulate cell-to-cell metabolism. These naturally occurring proteins are mitogenic and chemo-
tactic for the mesenchymal cells which contribute to new bone formation. These factors may
accelerate and/or enhance bone ingrowth into porous implants leading to increased strength at the
boneimplant interface.
123
Preliminary studies in a dog model have demonstrated that the percentage
of bone ingrowth into a porous titanium implant coated with equal amounts of PDGF, EGF, and
TGF- is almost doubled.
Friedman et al.
124
have shown TGF-, PDGF, and EGF in combination increase the rate of fracture
healing in rats. Increased bone, cartilage, brous tissue, and callous area were seen at 1 week
postoperative in rats receiving growth factor compared with controls. This indicates an accelerated
rate of fracture repair. Because the process of fracture healing and bone ingrowth are similar, these
growth-promoting factors may play a role in the long-term xation of porous coated implants.
Lynch et al.
125
demonstrated an increased bone ingrowth to dental implants stimulated with
PDGF and IGF-1 (insulin growth factor 1) in dogs. These implants were press-t into the root
canal.

Recently, three separate groups have shown TFG- adsorbed to ceramic-coated implants
inserted into trabecular bone of dogs can stimulate anchorage, bone ingrowth, and gap healing
of implants.
126-128
However, mechanical pushout tests showed no difference in xation strength.
Wang et al.
129
found basic FGF (bFGF) to promote ingrowth into porous structures coated with
hydroxyapatite.
Several recent studies have evaluated the effects of growth hormone on bone ingrowth, but
with conicting results. Guicheux et al.
130
found that human growth hormone increased bone
ingrowth.

However, Blom et al.
131
found that human growth hormone inhibited bone formation in
grooves coated with calcium phosphate.

Morberg et al.
132
felt negative results reported may have
been secondary to antibody formation against the growth hormone injected. This group used
transgenic mice overexpressing bovine growth hormone and found increased bone-to-metal contact.
From these ndings, it was felt that systemic administration of growth hormone in humans may
improve implant integration into bone.
0266/ ch28/frame Page 450 Friday, October 22, 1999 10:09 AM
Factors Affecting the Strength of the Bone-Implant Interface 451
More research needs to be done in this area. Only growth factors from the TGF- superfamily
(TGF-s and bone morphogenetic proteins) and bFGF have been demonstrated to possess substan-
tial in vivo bone stimulatory capacity. A major challenge for the clinical use of these growth factors
is an appropriate delivery system which lasts long enough (2 weeks) and has no side effects on
bone formation. However, these proteins will undoubtedly play a large role in cementless xation
in the future because of their positive effect at the boneimplant interface.
F. OTHER FACTORS
Mechanical vibration, ultrasound, and biphosphonates are several other modalities that may affect
bone ingrowth but have not been extensively studied. Usui et al.
133
studied the effects of mechanical
vibration on bone ingrowth into porous hydroxyapatite implant and fracture healing in a rabbit
model. The effects of 20 and 60 min vibration per day was evaluated. They found that the mechanical
vibration used in the study had a benecial effect on callous volume, but did not appear to promote
bone ingrowth into the porous implant.
Low-intensity ultrasound has been shown to be an effective means of accelerating fracture
healing.
134,135
Tanzer et al.
136
used a canine model to determine if noninvasive ultrasound could
inuence the rate or extent of growth into a porous-coated implant. Porous-coated transcortical
implants were inserted bilaterally into the femur of dogs. One femur served as a control and the
contralateral side received daily ultrasound for 2, 3, or 4 weeks. The ultrasound-stimulated
implants showed an 18% increase in bone ingrowth compared with the control side in the rst
2 to 3 weeks. Further animal and clinical study is needed to assess the clinical application for
noncement porous implants.
Biphosphonates have been demonstrated to block the action of osteoclasts and inhibit wear
debris-mediated bone resorption in a canine model. Recently, Shanbhag et al.
137
have shown that
biphosphonates can enhance the net bone formation in canine total hip components. This nding
coupled with the ability of biphosphonates to inhibit osteolysis may prove to be invaluable in the
long-term stability of total joints. Each of the modalities mentioned needs additional animal and
clinical testing to determine its efcacy.
IV. IN VITRO FACTORS
A. STORAGE BEFORE TESTING
In mechanical testing of bones, specimens often have to be frozen and thawed. Several investigators
have found no signicant effects on the mechanical properties of bones frozen at 20C for less
than 9 months and then thawed to 22C (room temperature).
138,139
In fact it has been shown that
multiple freezingthawing cycles do not signicantly affect the properties of bone.
140,141
Kang
et al.
142
studied the effects of multiple freezingthawing cycles on bovine tibial cancellous bone.
They found these cycles did not affect the ultimate indentation load or stiffness of the bone.
However, they suggest that bone specimens should be frozen and thawed in saline.
Thus far, no published data could be found on the effects of freezing on the mechanical properties
of the boneimplant interface. Some may infer from previous work cited that storage will have little
affect on the boneimplant interface, but this needs to be documented in animal models.
B. SPECIMEN PREPARATION
In order to study properly the mechanical properties of the boneimplant interface, specimen
preparation must be precise and adequate. The results obtained are only as good as the method
used to obtain them. Implants are inserted into bone at various angles and orientations although
they were aimed at certain directions. When harvesting the implant to test on a mechanical testing
system, such as pushout test or pullout test, it is important that the removal cuts be made
0266/ ch28/frame Page 451 Friday, October 22, 1999 10:09 AM
452 Mechanical Testing of Bone and the BoneImplant Interface
perpendicular to the implant and followed by precise ne cuts and grinding. If this is not done
properly, it will not be possible to align the implant in the loading direction. Loading forces
must be applied parallel to the implant to get a true biomechanical measure of the strength of
the boneimplant interface. If the force applied to the implant is oblique, then erroneous data
will be obtained.
Hartman et al.
143
evaluated the effects of pretreatment sterilization and cleaning methods on
material properties and osseoinductivity. The implants were analyzed with electron spectroscopy.
Argon radiofrequency glow discharge and ultraviolet chambertreated implants were associated
with rapid bone ingrowth and maturation of new bone. Conventional steam sterilization was
associated with less new bone ingrowth. These different methods of implant sterilization affect
bone ingrowth into porous substances and thus the strength of the boneimplant interface.
V. SURGICAL AND HOSPITAL FACTORS
A. SURGICAL SKILL
The xation strength of an implant either cemented or uncemented is highly dependent upon the
technical skill of the operating surgeons. Better long-term results are found in major medical centers
where more surgeries are performed each year. Total joint arthroplasties are technical surgeries
with much room for operator error. The constant evolution of information and skills required
demands that the practicing orthopaedic surgeon stay up-to-date.
A thorough preoperative evaluation and radiographs are essential for a successful outcome. A
skillful surgeon begins with both. Careful patient evaluation prevents catastrophic or even fatal
complications. Radiographs enable the surgeon to determine the correct size prosthesis for the best
t. Poorly tting prosthesis leads to motion which will weaken the boneimplant interface.
Many surgical approaches and techniques are used in arthroplasty. Surgeons should individu-
alize the operation according to their clinical and educational experiences. Specic techniques for
implantation vary according to the type of skeletal xation, shape of femoral component, length
of stem, and assembly of any modular components. The surgeon should be thoroughly familiar
with all instrumentation prior to the procedure.
Numerous aspects of total joint arthroplasty require exacting skill. For example, while reaming
the medullary canal for a cementless implant, care must be taken not to penetrate the femoral
cortex. Skill is also required in positioning the acetabular cup.
144
Malposition of the cup may cause
the neck of the femoral component to impinge on the margin of the socket and transfer excess
stress to the cup which may lead to loosening. In cemented surgeries, surgical skill is necessary in
the preparation and insertion of cement. Improperly prepared cement can lead to increased porosity
and weakness at the bonecementimplant interfaces. Also, skill is needed in centering the pros-
thesis to obtain a uniform cement mantle. Cracks may develop if the mantle is not uniform. These
are only a few of the many aspects of joint arthroplasty which require surgical skill to ensure a
strong boneimplant interface.
B. COMPLICATIONS
Infections and dislocations are two complications which may affect the long-term stability of the
prosthesis. Postoperative infections can be catastrophic. Bacteria may grow in the biolm on
biomaterials making eradication difcult. Approximately 1% of all total hip arthroplasties become
infected within 1 year.
145
Both acute and chronic infections often present a treatment dilemma.
Component loosening can sometimes be seen on X-ray examination in chronic infections. This
loosening weakens the boneimplant interface.
Dislocations are another complication seen after total hip arthroplasty. The incidence is close
to 3%.
146,147
Several risk factors have been documented, including (1) history of previous surgery;
0266/ ch28/frame Page 452 Friday, October 22, 1999 10:09 AM
Factors Affecting the Strength of the Bone-Implant Interface 453
(2) posterior approach; (3) faulty positioning of components; (4) inadequate soft tissue tension;
and (5) noncompliance. Dislocations may initially cause pain and swelling. If found quickly, the
boneimplant interface will not be affected. However, the joint mechanics may be altered if the
joint remains dislocated for extended periods of time, thus affecting the strength of the
boneimplant interface.
C. POSTOPERATIVE CARE
Many different rehabilitation programs are accepted for postoperative care. Regaining motion and
function and returning to independent living are vital elements. Postoperatively in hip arthroplasty,
the hip is placed in approximately 15 of abduction with a pillow, and the leg is stabilized so
rotation does not occur.
148
Most patients can begin limited mobilization on the rst postoperative
day. On the second day after the operation, patients are allowed to sit on the side of the bed and
begin gait training.
Time to weight-bearing differs depending on the type of xation used. With cementless, porous
ingrowth implants, many recommend limited weight-bearing for 6 to 8 weeks.
149
If implants were
cemented, weight-bearing as tolerated is permitted. Failure to follow these recommendations may
weaken the boneimplant interface and affect the stability of the implant.
Sun
150
studied the effects of load bearing on bone ingrowth into porous-coated implants in
rabbits. Sun found stimulation of load bearing beneted bone ingrowth and no weight-bearing
resulted in brous tissue forming in pores after 12 weeks. These results point to the importance of
some bearing in porous-coated implants. However, full weight-bearing may be detrimental, causing
a failure of xation.
Patients are often discharged when they are able to get up and out of bed on their own. This
may be as early as postoperative day 3 to 6. Patients should receive printed instructions to take
home. Any positions of instability noted in the operating room should be supplied to the patient.
All of these specics are aimed at ensuring stability at the boneimplant interface in order to have
a successful outcome.
VI. PATIENT FACTORS
A. AGE
Age is a primary determinant of whether a surgeon will use a cemented or cementless xation.
The effect of age on cementless xation is of paramount importance. Bone and mineral metab-
olism are known to change with age.
151
Most studies of biological xation are carried out in
young animals, or animals of unknown age, to show osseointegration of an implant. Very few
studies have evaluated how these implants are xed in the elderly and the consequences of aging
on the boneimplant stability. Since most orthopaedic implants are placed in aged skeletons, this
is a matter of utmost importance.
Magee et al.
152
evaluated bone ingrowth into transcortical porous-coated plugs in old (6.6 years)
and young (2.3 years) greyhounds.

Gender and ovarian status were not considered. Regardless, this
group reported low interface shear stress in older animals. In another study, Nakajima et al.
153
studied implant xation in young and old mongrel dogs. Age was dened based on teeth wear and
again gender and ovarian status were not considered. This group found a signicantly inferior callus
quality in old dogs in the early postoperative period (3 to 6 weeks).
In contrast, Eckhoff et al.
154
evaluated bone ingrowth in hydroxyapatite-coated and uncoated
implants in young and old sheep. These animals were of known age and reproductive status.
Surprisingly, they found comparable xation of implants in the young and old sheep. This
group felt age was not a dominant factor in the development or failure of implant osseointe-
gration. A case report supporting this study was presented by Soballe et al.
155
The case is a
0266/ ch28/frame Page 453 Friday, October 22, 1999 10:09 AM
454 Mechanical Testing of Bone and the BoneImplant Interface
retrieval of a hydroxyapatite-coated femoral prosthesis from a 98-year-old osteoporotic female
12 weeks after implantation. Histological analysis of the specimen showed a large amount of
bone in direct contact with the prosthesis despite age, gender, and presence of osteoporosis in
the patient.
Due to the conicting reports of the effects of age on boneimplant strength, further study is
needed in this area. All of the previous studies have involved animals of different breeds with
numerous variables. It is hoped, in the near future, that studies may document ndings in humans.
B. LIFESTYLE
Lifestyle of the patient is a consideration when deciding on the type of implant to use in joint
arthroplasty. Most surgeons will choose a biologically xed implant in more active individuals.
One reason for this is that if revision surgery must be done, uncemented prostheses are much easier
to work with. Patients who are not active tend to be elderly and a cemented prosthesis is a better
choice. The activity of the patient will determine the amount of stress put on the boneimplant
interface. Large amounts of stress may lead to failure. However, some stress is needed for bone to
remain healthy and not atrophy. Thus, the lifestyle of the patient is not only important in choosing
the type of xation but also in ultimate strength of the boneimplant interface.
C. INFLAMMATION (RHEUMATOID AND OTHER INFLAMMATORY CONDITIONS)
Few studies have been carried out to study the effects of inammation on bone ingrowth. An et al.
48
evaluated the interface between implanted cylinders with three different surface textures in the
distal femur of rabbits following the induction of inammatory arthritis.

Inammatory arthritis was
induced with injections of carrageenan to the knee joint. The contralateral knee joint was used as
a control. This group found the thickness and number of trabeculae less on the arthritic side of all
three texture types compared with controls. Mechanical testing also showed the shear strength of
the interface was weaker on the inammatory side. From these ndings, they concluded that
inammatory arthritis may inuence the quality of local bone and thus compromise the bone
apposition and mechanical stability of the interface between implant and bone. Branemark and
Thomsen
156
found similar trends in rabbit models.
Inammation from many different clinical conditions affects the longevity of implanted pros-
thesis. Dorr et al.
157
reported a higher failure rate for implants in patients with inammatory
conditions of the hip. Numerous problems such as infection are commonly seen in chronic inam-
matory diseases. This may be due to the disease itself or partly because of the medications many
of these patients take. Steroids are commonly prescribed for many patients with these disorders.
Rheumatoid arthritis is a common disease in orthopaedic patients. Poss et al.
158
reported an
increased risk of complications in patients with this condition. In their study, 96% of patients with
rheumatoid arthritis clinically improved after total hip replacement. However, 40% of patients had
subsidence of the femoral component and 78% increased radiolucency of the acetabular component
at 6 to 10 years after surgery. Infection rate was also doubled in the rheumatoid group vs. the
osteoarthitic controls.
Rheumatoid arthritis patients often have an activity-limiting disease. This limitation of stress
at the boneimplant interface may in fact lead to a longer implant survivorship compared with
other conditions such as osteoarthritis. Sochart and Porter
159
reviewed 226 Charnley low-friction
arthroplasties in 161 patients with congenital dislocation of the hip, osteoarthritis, and rheumatoid
arthritis. They found rheumatoid arthritis patients had the lowest prevalence of loosening and
revision of the acetabular component.
Systemic lupus erythematosis (SLE) is another chronic, multisystemic inammatory disorder
usually affecting women in their second or third decade. Hanssen et al.
160
evaluated hip arthroplasty
in patients with SLE and found an increased rate of wound complications and deep infections.

In
0266/ ch28/frame Page 454 Friday, October 22, 1999 10:09 AM
Factors Affecting the Strength of the Bone-Implant Interface 455
another study, Huo et al.
161
found loosening of femoral and acetabular components to be a problem.
Based on these and other studies, reamed uncemented or hybrid total hip arthroplasties are recom-
mended for this patient population.
These clinical studies show the numerous problems in patients with inammatory conditions
who have arthroplasty. The increased rate of infection and associated problems in these patients
undoubtedly affect the boneimplant interface. Further research is needed in this area to dene
better the mechanisms and its signicance.
D. OSTEOPENIA 45OR OSTEOPOROSIS
Osteopenia and osteoporosis afict many patients, particularly the elderly. Since total joint arthro-
plasty is often performed in this age group, the effects of these conditions on implant stability is
signicant. These conditions may ultimately affect the boneimplant interface and success of the
implant. Several investigators have evaluated osteoporosis and its effects on osseointegration.
162,163
Fujimoto et al.
164
evaluated the effects of osteoporosis on osseointegration of titanium implants in
rabbits.

Osteoporosis was created in these animals by administering steroids. Bone density was
used to evaluate osteoporosis in these models. These investigators found a signicantly lower
removal torque in the osteoporotic group than in controls.
In another study, Mori et al.
165
studied the reaction of boneimplant interface in osteoporotic
animal models. The experimental model consisted of a rabbit which was ovariectomized and fed
a low-calcium diet. Dual energy X-ray absorptiometry was used to document the bone mineral
density.
166
In this study, new bone formation was delayed in the osteoporotic animals but did occur
after 12 weeks. From this they concluded that osseointegration occurs in osteoporotic bone but the
healing period is extended. Thus, conditions affecting the density and quantity of bone are important
factors in the boneimplant interface strength.
VII. SUMMARY
As the number and sophistication of orthopaedic implants is growing, a great deal of research is
focusing on the boneimplant interface. The factors affecting the strength of this interface are
important not only to surgeons implanting the prosthesis but also to researchers studying the
biomechanical properties of various articial components. As discussed, numerous factors have
proved to have a signicant impact either positive or negative on the strength of the interface.
Various implant surface congurations, coatings, geometric designs, and implantation tech-
niques affect the nal stability of the prosthesis. Medications commonly used by surgeons, such
as NSAIDs or hormones, may decrease the bony ingrowth into porous implants. Patient factors
including age, lifestyle, and osteoporosis also can dramatically affect the strength of the
boneimplant interface. In addition, appropriate storage and specimen preparation can inuence
the data obtained in mechanical testing of this area.
It is the responsibility of the surgeon and the scientist to be cognizant of the numerous factors
affecting the strength of the boneimplant interface. Continued research, especially mechanical
testing, needs to be done to clarify further as well as discover new factors that play a role in the
strength of the boneimplant interface.
REFERENCES
1. Park, J.B., Orthopaedic prosthesis xation, Ann. Biomed. Eng., 20, 583, 1992.
2. Branemark, P.I., Osseointegration and its experimental background, J. Prosthet. Dent., 50, 399, 1983.
3. Branemark, P.I., Zarb, G.A., and Albektsson, T., Tissue Integrated Prostheses: Osseointegration in
Clinical Dentistry, Quintessence, Chicago, 1985.
0266/ ch28/frame Page 455 Friday, October 22, 1999 10:09 AM
456 Mechanical Testing of Bone and the BoneImplant Interface
4. Sumner, D.R. and Galante, J.O., Bone ingrowth, in Surgery of the Musculoskeletal System, Evarts,
C.M., Ed., Churchill Livingstone, New York, 1990, 151.
5. Sumner, D.R., Klenapfel, H., and Galante, J.O., Metallic implants, in Bone Grafts: From Basic Science
to Clinical Application, Habal, M.B. and Reddi, A.H., Eds., W.B. Saunders, Philadelphia, 1993, 619.
6. Lee, T.Q., Danto, M.I., and Kim, W.C., Initial stability comparisons of modular hip implants in
synthetic femurs, Orthopedics, 21, 885, 1998.
7. Jarcho, M., Calcium phosphate ceramics as hard tissue prosthetics, Clin. Orthop., 157, 259, 1981.
8. Jarcho, M., Kay, J.F., Gumaer, K.L., et al., Tissue, cellular and subcellular events at a boneceramic
hydroxyapatite interface, J. Bioeng., 1, 79, 1992.
9 Friedman, R.J., An, Y.H., Ming, J., et al., Inuence of biomaterial surface texture on bone ingrowth
in the rabbit femur, J. Orthop. Res., 14, 455, 1996.
10. Bobyn, J.D., Pilliar, R.M., Cameron, H.U., and Weatherly, G.C., The optimum pore size for the xation
of porous-surfaced metal implants by the ingrowth of one, Clin. Orthop., 150, 263, 1980.
11. Young, F.A., Spector, M., and Kresch, C.H., Porous titanium endosseous dental implants in Rhesus
monkeys: microradiography and histological evaluation, J. Biomed. Mater. Res., 13, 843, 1979.
12. Galante, J. and Rostoker, W., Fiber metal composites in the xation of skeletal prostheses, J. Biomed.
Mater. Res. Symp., 4, 43, 1973.
13. Galante, J., Rostoker, W., Lueck, R., and Ray, R.D., Sintered ber metal composites as a basis for
attachment of implants to bone, J. Bone Joint Surg., 53A, 101, 1971.
14. Heck, D.A., Nakajima, I., Kelly, P.J., and Chao, E.Y., The effect of load alteration on the biological
and biomechanical performance of a titanium ber-metal segmental prosthesis, J. Bone Joint Surg.,
68A, 118, 1986.
15. Galante, J., Laing, P.G., and Lautenschlager, E., Biomaterials, AAOS Instr. Course Lect., 24, 1, 1975.
16. Aspenberg, P., Goodman, S., Toksvig-Larsen, S., et al., Intermittent micromotion inhibits bone
ingrowth. Titanium implants in rabbits, Acta Orthop. Scand., 63, 141, 1992.
17. Friedman, R.J., Advances in biomaterials and factors affecting implant xation, AAOS Instr. Course
Lect., 41, 127, 1992.
18. Branson, P.J., Steege, J.W., Wixson, R.L., et al., Rigidity of internal xation with uncemented tibial
knee implants, J. Arthroplasty, 4, 21, 1989.
19. Soballe, K., Hansen, E., Rasmussen, H., et al., Hydroxyapatite coating enhances xation of porous
coated implants, Acta Orthop. Scand., 61, 299, 1990.
20. Soballe, K., Hansen, E.S., Rasmussen, H., et al., Tissue ingrowth into titanium and hydroxyapatite-
coated implants during stable and unstable mechanical conditions, J. Orthop. Res., 10, 285, 1992.
21. Kohn, D.H. and Ducheyne, P., A parametric study of the factors affecting the strength of porous coated
Ti-6Al-4V implant alloy, J. Biomed. Mater. Res., 24, 1483, 1990.
22. Takeuchi, M.J., Kelman, D.C., and Smith, T.S., The effects of sintering heat treatments on the fatigue
strength of Ti-6Al-7NB, Trans. Soc. Biomater., 14, 291, 1991.
23. Kelman, D.C., Takeuchi, M.J., and Smith, T.S., Effect of metallic plasma sprayed coatings on high
cycle fatigue, Trans. Soc. Biomater., 14, 290, 1991.
24. Kummer, F.J., Ricci, J.L., and Blumenthal, N.C., RF plasma treatment on metallic implant surfaces,
J. Appl. Biomater., 3, 39, 1992.
25. Robertson, D.M., Pierre, L., and Chehal, F., Preliminary observations of bone ingrowth into porous
material, J. Biomed. Mater. Res., 10, 335, 1976.
26. Weish, R.P., Pilliar, R.M., and Macnab, I., Surgical implants. The role of surface porosity in xation
to bone and acrylic, J. Bone Joint Surg., 53A, 963, 1971.
27. Bobyn, J.D, Pilliar, R.M., Cameron, H.U., and Weatherly, G.C., The optimum pore size for the xation
of porous-surfaced metal implants by the ingrowth of bone, Clin. Orthop., 150, 263, 1980.
28. Cook, S.D., Walsh, K.A., and Haddad, R.J., Jr., Interface mechanics and bone growth into porous Co-
Cr-Mo alloy implants, Clin. Orthop., 193, 163, 1985.
29. Biegler, F.B., Reuben, J.D., Harrigan, T.P., et al., Effect of porous coating and loading conditions on
total hip femoral stem stability, J. Arthroplasty, 10, 839, 1995.
30. Shigley, J.E., Mechanical Engineering Design, 3rd ed., McGraw-Hill, New York, 1977.
31. Vernon-Roberts, B. and Freman, M.A.R., The tissue response to total joint replacement prostheses,
in The Scientic Basis of Total Joint Replacement, Swanson, S.A.V. and Freeman, M.A.R., Eds., John
Wiley, New York, 1977, 86.
0266/ ch28/frame Page 456 Friday, October 22, 1999 10:09 AM
Factors Affecting the Strength of the Bone-Implant Interface 457
32. Segawa, H., Lew, W.D., Bourgeault, C., et al., Micromotion study of the preparation method of the
proximal tibia for primary tibial components, Trans. Orthop. Res. Soc., 24, 866, 1999.
33. Meding, J.B., Ritter, M.A., Keating, E.M., and Faris, P.M., Impaction bone-grafting before insertion
of a femoral stem with cement in revision total hip arthroplasty. A minimum two-year follow-up study,
J. Bone Joint Surg., 79A, 1834, 1997.
34. Green, J.R., Nemzek, J.A., Arnoczky, S.P., et al., The effect of bone compaction on early xation of
porous-coated implants, J. Arthroplasty, 14, 91, 1999.
35. Yu, L., Clark, J.G., Dai, Q.G., et al., Improving initial mechanical xation of porous coated femoral
stem by a cancellous bone compaction method, Trans. Orthop. Res. Soc., 24, 863, 1999.
36. Chareancholvanich, K., Bechtold, J.E., Soballe, K., et al., Compaction of existing cancellous bone in
the primary setting enhances interfacial shear strength in vivo, Trans. Orthop. Res. Soc., 24, 865, 1999.
37. Feighan, J.E., Goldberg, V.M., Dawy, D., et al., The inuence of surface-blasting on the incorporation
of titanium-alloy implants in a rabbit intramedullary model, J. Bone Joint Surg., 77A, 1380, 1995.
38. Goldberg, V.M., Stevenson, S., Feighan, J., and Davy, D., Biology of grit-blasted titanium alloy
implants, Clin. Orthop., 319, 122, 1995.
39. Wang, J.S., Taylor, M., Flivik, G., and Lindgren, L., Inuence of stem roughness on the static shear
strength of the stem-cement interface, Trans. Orthop. Res. Soc., 24, 872, 1999.
40. Wang, J.S., Flivik, G., Taylor, M., and Lindgren, L., Stem-cement interface shear strength is inuenced
by stem surface nish but not cement interface porosity, Trans. Orthop. Res. Soc., 24, 871, 1999.
41. Nakamura, T., Yamamuro, T., Higashi, S., et al., A new glass ceramic for bone replacement: evaluation
of its bonding ability to bone tissue, J. Biomed. Mater. Res., 19, 685, 1985.
42. Uamamuro, T., Hench, L., and Wilson, J., Eds., CRC Handbook of Bioactive Ceramics, Vols. 1 and
2, CRC Press, Boca Raton, FL, 1990.
43. Pazzaglia, U.E., Brossa, F., Zatti, G., et al., The relevance of hydroxyapatite and spongious titanium
coatings in xation of cementless stems. An experimental comparative study in rat femur employing
histological and microangiographic techniques, Arch Orthop. Trauma Surg., 117, 279, 1998.
44. Lemons, J.E., Hydroxyapatite coatings, Clin. Orthop., 235, 220, 1988.
45. Lemons, J.E., Bioceramics: is there a difference? Clin. Orthop., 261, 153, 1991.
46. Davies, J.E., Ed., The BoneBiomaterial Interface, University of Toronto Press, Toronto, 1991.
47. Hench, L.L. and Paschal, H.A., Direct chemical bond of bioactive glass ceramics to bone and tissue,
J. Biomed. Mater. Res., 4, 25, 1973.
48. An, Y.H., Friedman, R.J., Jiang, M., et al., Bone ingrowth to implant surfaces in an inammatory
arthritis model, J. Orthop. Res., 16, 576, 1998.
49. de Groot, K., Geesink, R.G., Klein, C.P., et al., Plasma sprayed coatings of hydroxylapatite, J. Biomed.
Mater. Res., 21, 1275, 1987.
50. Geesink, R.G., de Groot, K., and Klein, C.P., Chemical implant xation using hydroxylapatite coatings,
Clin. Orthop., 25, 147, 1987.
51. Cook, S.D., Thomas, K.A., Kay, J.F., et al., Hydroxyapatitecoated titanium for orthopaedic implant
applications, Clin. Orthop., 232, 225, 1988.
52. Soballe, K., Hansen, E.S., Helle, B.R., et al., Tissue ingrowth into titanium and hydroxyapatite-coated
implants during stable and unstable mechanical conditions, J. Orthop. Res., 10, 285, 1992.
53. Thomas, K.A, Kay, J.F., Cook, S.D., et al., The effect of surface macrotexture and hydroxylapatite
coating on the mechanical strengths and histologic proles of titanium implant materials, J. Biomed.
Mater. Res., 21, 1395, 1987.
54. Thomas, K.A., Cook, S.D., Kay, J.F., et al., Attachment strength and histology of hydroxylapatite
coated implants, Biomater. Med. Devices Artif. Organs, 14, 73, 1986.
55. Cook, S.D., Thomas, K.A., Kay, J.F., et al., Hydroxyapatite-coated porous titanium for use as an
orthopaedic biologic attachment system, Clin. Orthop., 230, 303, 1988.
56. Richardson, D.C., The response of cancellous and cortical canine bone to hydroxylapatite coated
titanium rods, Trans. Soc. Biomater., 12, 176, 1989.
57. Magee, F.P., Kay, J.F., and Hedley, A.K., Interface strength and histology of hydroxylapatite coated
and surface textured titanium, Trans. Soc. Biomater., 12, 173, 1989.
58. Manley, M.T., Kay, J.F., Uratsuji, M., et al., Hydroxylapatite coatings applied to implants subjected
to functional loads, Trans. Soc. Biomater., 10, 210, 1987.
59. Lemons, J.E., Hydroxyapatite coatings, Clin. Orthop., 235, 202, 1988.
0266/ ch28/frame Page 457 Friday, October 22, 1999 10:09 AM
458 Mechanical Testing of Bone and the BoneImplant Interface
60. Cook, S.D. and Thomas, K.A., Coating defects in hydroxylapatite coated implants, Trans. Soc.
Biomater., 12, 172, 1989.
61. Noble, P.C., Alexander, J.W., Lindahl L.J., et al., The anatomic basis of femoral component design,
Clin. Orthop., 235, 148, 1988.
62. Soballe, K., Hansen, E.S., Rasmussen, H.B., et al., Enhancement of osteopenic and normal bone
ingrowth into porous coated implants by hydroxyapatite coating, Trans. Orthop. Res. Soc., 14, 554,
1989.
63. Harris, W.H., White, R.E., McCarthy, J.C., et al., Bony ingrowth xation of the acetabular component
in canine total hip arthroplasty, Clin. Orthop., 176, 7, 1983.
64. Thomas, K.A., Cook, S.D, Haddad, R.J., et al., Biologic response to hydroxylapatite coated titanium
hips, J. Arthroplasty, 4, 43, 1989.
65. Mahomed, N., Maistrelli, G., and Fornasier, V., Interfacial shear strength of hydroxyapatite coated
hip implants, Orthop. Trans., 14, 710, 1990.
66. Schmalzried, T.P., Kwong, L.M., Jasty, M., et al., The mechanism of loosening of cemented acetabular
components in total hip arthroplasty, analysis of specimens retrieved at autopsy, Clin. Orthop., 274,
60, 1992.
67. de Bruijn, J.D., van den Brink, I., Mendes, S., et al., Bone induction by implants coated with cultured
osteogenic bone marrow cells and in vitro formed bone matrix, Trans. Orthop. Res. Soc., 24, 888, 1999.
68. Jarvis, W.C., Biomechanical advantages of wide-diameter implants, Compend. Contin. Educ. Dent.,
18, 687, 1997.
69. Alfaro-Adrian, J., Gill, H.S., and Murray, D.W., The inuence of prosthetic design on migration. A
study of Charnley and Exeter THR with RSA, Trans. Orthop. Res. Soc., 24, 919, 1999.
70. Canale, S.T., Ed., Campbells Operative Orthopaedics, Mosby, St. Louis, MO, 1998, 308.
71. Schneider, E., Christian, K., Eulenberger, J., et al., Comparative study of initial stability of cementless
hip prostheses, Clin. Orthop., 248, 200, 1989.
72. Kelsey, D. and Goodman, S.B., Design of the femoral component for cementless hip replacement:
the surgeons perspective, Am. J. Orthop., 26, 407, 1997.
73. Ries, M.D., Harbaugh, M., Shea, J., and Lambert, R., Effect of cementless acetabular cup geometry
on strain distribution and press-t stability, J. Arthroplasty, 12, 207, 1997.
74. Jacobs, J.J., Gilber, J.l., and Urban, R.M., Corrosion of metal implants, in Advances in Operative
Orthopaedics, Vol. 2, Staufer, R.N., Ed., Mosby, St. Louis, MO, 279, 1994.
75. Levine, D.L. and Staehle, R.W., Crevice corrosion in orthopaedic implant metals, J. Biomed. Mater.
Res., 11, 553, 1977.
76. Budinski, G.K., Tibological properties of titanium alloys, Wear, 151, 203, 1991.
77. Woolson, S.T., Milbauer, J.P., Bobyn, J.D., et al., Fatigue fracture of a forged cobaltchro-
miummolybdenum femoral component inserted with cement. A report of ten cases, J. Bone Joint
Surg., 79A, 1842, 1997.
78. Trousdale, R.T., Berry, D.J., Jacob, J., and Gilbert, J.L., Fracture of an acetabulum component inserted
without cement, J. Bone Joint Surg., 79A, 901, 1997.
79. Sandborn, P.M., Cook, S.D., Kester, M.A., and Haddad, R.J., Jr., Fatigue failure of the femoral
component of a unicompartmental knee, Clin. Orthop., 222, 249, 1987.
80. Miller, J.J., Review of Orthopaedics, 2nd ed., W.B. Saunders, Philadelphia, 1996.
81. Bayne, S.C., Lautenschlager, E.P., Greener, E.H., and Myer, P.R., Jr., Degree of polymerization of
acrylic bone cement, J. Biomed. Mater. Res., 9, 27, 1975.
82. Cameron, U.H., Mills, R.H., Jackson, R.W., and Macnab, I., The structure of polymethylmethacrylate
cement, Clin. Orthop., 100, 287, 1974.
83. Haas, S.S., Brauer, G.M., and Dickson, G., A characterization of polymethylmethacrylate bone cement,
J. Bone Joint Surg., 57A, 380, 1975.
84. Wixson, R.L., Lautenschlager, E.P., and Novak, M.A., Vacuum mixing of acrylic bone cement,
J. Arthroplasty, 2, 144, 1987.
85. Gruen, T.A., Markolf, K.L., and Amstutz, H.C., Effects of laminations and blood entrapment on the
strength of acrylic bone cement, Clin. Orthop., 119, 250, 1976.
86. Willert, H.G. and Semlitsch, M., Reactions of the articular capsule to wear products of articial joint
prostheses, J. Biomed. Mater. Res., 11, 157, 1977.
0266/ ch28/frame Page 458 Friday, October 22, 1999 10:09 AM
Factors Affecting the Strength of the Bone-Implant Interface 459
87. Goldring, S.R., Schiller, A.L., Roelke, M., et al., The synovial-like membrane at the bonecement
interface in loose total hip replacements and its proposed role in bone lysis, J. Bone Joint Surg., 65A,
575, 1983.
88. Goodman, S.B., Chin, R.C., Chiou, S.S., et al., A clinical pathologic biochemical study of the mem-
brane surrounding loosened and nonloosened total hip arthroplasties, Clin. Orthop., 244, 182, 1989.
89. Apple, A.M., Sowder, W.G., Hop, C.N., and Herman, J.H., Production of mediators of bone resorption
by prosthesis associated pseudomembranes, Trans. Orthop. Res. Soc., 13, 362, 1988.
90. Mather, S.E., Emmanuel, J., Magee, F.P., et al., Interleukin and prostaglandin E2 in failed total hip
arthroplasty, Trans. Orthop. Res. Soc., 14, 489, 1989.
91. Chrisman, O.D. and Snook, G.A., The problem of refracture of the tibia, Clin. Orthop., 60, 217, 1968.
92. Dencker, H., Refracture of the shaft of the femur, Acta Orthop. Scand., 35, 16, 1964.
93. Engh, C.A., Hip arthroplasty with a Moore prosthesis with porous coating a ve year study, Clin.
Orthop., 176, 52, 1983.
94. Engh, C.A., Bobyn, J.D., and Glassman, A.H., Porous coated hip replacement. A study of factors
governing bone ingrowth, stress shielding, and clinical results, J. Bone Joint Surg., 69A, 45, 1987.
95. Manley, P.A., Vanderby, R., Jr., Kohles, S., et al., Alterations in femoral strain, micromotion, cortical
geometry, cortical porosity, and bony ingrowth in uncemented collared and collarless prostheses in
the dog, J. Arthroplasty, 10, 63, 1995.
96. Martin, R.B., Paul, H.A., Bargar, W.L., et al., Effects of estrogen deciency on growth of tissue into
porous titanium implants, J. Bone Joint Surg., 70A, 4, 1988.
97. Paul, H.A. and Bargar, W.L., Histologic changes in the dog femur following total hip replacement
with current cementing techniques, J. Arthroplasty, 1, 5, 1986.
98. Rose, R.M., Martin, R.B., Orr, R.B., and Radin, E.C., Architectural changes in the proximal femur
following prosthetic insertion: preliminary observations of an animal model, J. Biomech., 17, 241,
1984.
99. Newman, N.M. and Ling, R.S., Acetabular bone destruction related to nonsteroidal antiinammatory
drug, Lancet, 2, 11, 1985.
100. Keller, J., Bayer-Kristensen, I., Bak, B., et al., Indomethacin and bone remodeling. Effect on cortical
bone after osteotomy in rabbits, Acta Orthop. Scand., 60, 119, 1989.
101. Ro, J., Sudmann, E., and Martin, P.F., Effect of indomethacin on fracture healing in rats, Acta Orthop.
Scand., 47, 588, 1976.
102. Longo, J.A., Hedley, A.K., Weinstein, A.M., et al., Comparative effects of EHDP, radiation therapy
and indomethacin on bone ingrowth in rabbits, Trans. Soc. Biomater., 8, 161, 1985.
103. Keller, J.C., Trancik, T.M., Young, F.A., et al., Effects of indomethacin on bone ingrowth, J. Orthop.
Res., 7, 28, 1989.
104. Trancik, T., Mills, W., and Vinson, N., The effect of indomethacin, aspirin, and ibuprofen on bone
ingrowth into a porouscoated implant, Clin. Orthop., 249, 113,.1989.
105. Longo, J.A., Magee, F.P., Van De Wyngaerde, D.P., et al., Naproxen and skeletal implant interfaces,
in Interfaces in Medicine and Mechanisms, William, K.R., Toni, A., Middleton, J., and Pallotti, G.,
Eds., Elsevier, New York, 1991.
106. Cook, S., Barrack, R., Dalton, J.E., et al., Effects of indomethacin on biological xation of porous
coated titanium implants, J. Arthroplasty, 10, 3, 1995.
107. Engh, C.A., Glassman, A.H., and Suthers, K.E., The case for porous coated implants, the femoral
side, Clin. Orthop., 261, 63, 1990.
108. Skriptz, R. and Aspenberg, P., Tensile bond between bone and titanium: a reapproval of osseointe-
gration, Acta Orthop. Scand., 69, 315, 1998.
109. Ayers, D.C., Pellegrini, V.D., and Evarts, C.M., Prevention of heterotopic ossication in highrisk
patients by radiation therapy, Clin. Orthop., 263, 87, 1991.
110. Wise, M.W., Robertson, I.D., Lachiewicz, P.F., et al., The effects of radiation therapy on the xation
strength of an experimental porouscoated implant in dogs, Clin. Orthop., 261, 276, 1990.
111. Sumner, D.R., Turner, T.M., Pierson, R.H., et al., Effects of radiation on xation of non-cemented
porous-coated implants in a canine model, J. Bone Joint Surg., 72A, 1527, 1990.
112. Chin, H.C., Frassica, F.J., Markel, M.D., et al., The effects of therapeutic doses of irradiation on
experimental bone graft incorporation over a porous-coated segmental defect endoprosthesis, Clin.
Orthop., 289, 254, 1993.
0266/ ch28/frame Page 459 Friday, October 22, 1999 10:09 AM
460 Mechanical Testing of Bone and the BoneImplant Interface
113. Basset, C.A.L., Pilla, A.A., and Pawluk, R.J., A nonoperative salvage of surgically resistant pseudar-
throses and non-unions by pulsing electromagnetic elds, Clin. Orthop., 124, 128, 1977.
114. Basset, C.A.L., Caulo, N., and Kort, J., Congenital pseudarthroses of the tibia. Treatment with
pulsing electromagnetic elds, Clin. Orthop., 154, 136, 1981.
115. Basset, C.A.L., Mitchell, S.N., and Gaston, S.R., Treatment of ununited tibial diaphyseal fractures
with pulsing electromagnetic elds, J. Bone Joint Surg., 63A, 511, 1981.
116. Ijiri, K., Matsunaga, S., Fukuyama, K., et al., The effect of pulsing electromagnetic eld on bone
ingrowth into a porous coated implant, Anticancer Res., 16, 2853, 1996.
117. Park, J.B. and Kenner, G.H., Effect of electrical stimulation on the tensile strength of the porous
implant and bone interface, Biomater. Med. Devices Artif. Organs, 3, 233, 1975.
118. Park, J.B. and Kenner, G.H., Effect of electrical stimulation on the interfacial tensile strength and
amount of bone formation, Biomater. Med. Devices Artif. Organs, 4, 225, 1976.
119. Weinstein, A.M., Klawitter, J.J., Cleveland, T.W., and Amoss, D.C., Electrical stimulation of bone
growth into porous Al
2
O
3
, J. Biomed. Mater. Res., 10, 231, 1976.
120. Shinizu, T., Zerwekh, J.E., Videman, T., et al., Bone ingrowth into porous calcium phosphate ceramics:
inuence of pulsing electromagnetic eld, J. Orthop. Res., 6, 248, 1988.
121. Ducheyne, P., Ellis, L.Y., Pollack, S.R., et al., Field distributions in the rat tibia with and without a
porous implant during electrical stimulation: a parametric modeling, IEEE Trans. Biomed. Eng., 39,
1168, 1992.
122. Schutzer, S.F., Jasty, M., Bragdon, C.R., et al., A double-blind study on the effects of a capacitively
coupled electrical eld on bone ingrowth into porous-surfaced canine total hip prostheses, Clin.
Orthop., 260, 297, 1990.
123. Canalis, E., Effects of growth factors on bone cell replication and differentiation, Clin. Orthop., 193,
246, 1985.
124. Friedman, R.J., Acurio, M.T., Davis, R., et al., Effects of growth factors and indomethacin of fracture
healing, Trans. Orthop. Res. Soc., 17, 421, 1992; Orthop. Trans., 16, 514, 1992.
125. Lynch, S.E., Buser, D., Hernandez, R.A., et al., Effects of platelet derived growth factor/insulin like
growth factor 1 combination on bone around titanium dental implants: results of a pilot study in beagle
dogs, J. Periodontol., 62, 710, 1991.
126. Lind, M., Overgaurd, S., Soballe, K., et al., Transforming growth factor beta enhances xation of
ceramic coated implants, Trans. Orthop. Res. Soc., 192, 1995.
127. Sumner, D.R., Turner, T.M., Durchio, A.F., et al., Enhancement of bone ingrowth by transforming
growth factor beta, J. Bone Joint Surg., 77A, 1135, 1995.
128. Martin, L., Soren O., Boonsri, O., et al., Transforming growth factor beta stimulates bone ongrowth
to weight loaded tricalcium phosphate coated implants, J. Bone Joint Surg., 78B, 3, 1996.
129. Wang, J.S. and Aspenberg, P., Basic broblast growth factor promotes bone ingrowth in porous
hydroxyapatite, Clin. Orthop., 333, 252, 1996.
130. Guicheux, J., Gauthier, O., Aguado, E., et al., Growth hormone-loaded macroporous calcium phos-
phate ceramic: in vitro biopharmaceutical characterization and preliminary in vivo study, J. Biomed.
Mater. Res., 40, 560, 1998.
131. Blom, E.J., Verheij, J.G., de Blieck-Hogervorst, J.M., et al., Cortical bone ingrowth in growth hormone-
loaded grooved implants with calcium phosphate coatings in goat femurs, Biomaterials, 19, 263, 1998.
132. Morberg, P.H., Isaksson, O.G., Johansson, C.B., et al., Improved long-term boneimplant integration.
Experiments in transgenic mice overexpressing bovine growth hormone, Acta Orthop. Scand., 68,
344, 1997.
133. Usui, Y., Zerwekh, J.E., Vanharanta, H., et al., Different effects of mechanical vibration on bone
ingrowth into porous hydroxyapatite and fracture healing in a rabbit model, J. Orthop. Res., 7, 559,
1989.
134. Kristiansen, T.K., Ryaby, J.P., McCabe, J., et al., Accelerated healing of distal radius fractures with
the use of specic, low intensity ultrasound, J. Bone Joint Surg., 79A, 961, 1997.
135. Heckman, J.D., Ryaby, J.P., McCabe, J., et al., Acceleration of tibial fracture-healing by noninvasive,
low-intensity pulsed ultrasound, J. Bone Joint Surg., 76A, 26, 1994.
136. Tanzer, M., Harvey, E., Kay, A., et al., Effect of noninvasive low intensity ultrasound on bone growth
into porous-coated implants, J. Orthop. Res., 14, 901, 1996.
0266/ ch28/frame Page 460 Friday, October 22, 1999 10:09 AM
Factors Affecting the Strength of the Bone-Implant Interface 461
137. Shanbhag, A.S., May, D., Kovach, C., et al., Enhancing net bone formation in canine total hip
components with biphosphonates, Trans. Orthop. Res. Soc., 24, 255, 1999.
138. Panjabi, M.M., Krag, M., Summers, D., et al., Biomechanical time-tolerance of fresh cadaveric human
spine specimens, J. Orthop. Res., 3, 292, 1985.
139. Roe, S.C., Pijanowski, G.J., and Johnson, A.L., Biochemical properties of canine cortical bone
allografts: effects of preparation and storage, Am. J. Vet. Res., 49, 873, 1988.
140. Linde, F. and Sorensen, H.C., The effect of different storage methods on the mechanical properties
of trabecular bone, J. Biomech., 26, 1249, 1993.
141. Sedlin, E.D. and Hirsch, C., Factors affecting the determination of the physical properties of femoral
cortical bone, Acta Orthop. Scand., 37, 29, 1966.
142. Kang, W., An, Y.H., and Friedman, R.J., Effects of multiple freezingthawing cycles on ultimate
indentation load and stiffness of bovine cancellous bone, Am. J. Vet. Res., 58, 1171, 1997.
143. Hartman, L.C., Meenaghan, M.A., Schaaf, N.G., and Hawker, P.B., Effects of pretreatment sterilization
and cleaning methods on materials properties and osseoinductivity of a threaded implant, Int. J. Oral
Maxillofac. Implants, 4, 11, 1989.
144. Adler, E., Stuchin, S.A., and Kummer, F.J., Stability of press-t acetabular cups, J. Arthroplasty, 7,
295, 1992.
145. Fitzgerald, R.H., Total hip arthroplasty sepsis, prevention and diagnosis, Orthop. Clin. North Am., 23,
259, 1992.
146. Khan, M.A., Brakenbury, P.H., and Reynolds, I.S., Dislocations following total hip arthroplasty,
J. Bone Joint Surg., 63B, 214, 1981.
147. Kristiansen, B., Jorgensen, L., and Holmich, P., Dislocations following total hip replacement, Arch.
Orthop. Trauma Surg., 103, 375, 1985.
148. Canale, S.T., Ed., Campbells Operative Orthopaedics, Mosby, Philadelphia, 1998, 296.
149. Nercessian, O.A. and Joshi, P.R., General principles of surgical technique, in The Adult Hip, Callaghan,
J.J., Rosenberg, A.G., and Rubash, H.E., Eds., Lippincott-Raven, Philadelphia, 1998, 951.
150. Sun, Y., [The effects of early stage load-bearing on bone ingrowth of porous coated implant in rabbits],
Chung Hua I Hsueh Tsa Chih (Taipei), 72, 265, 1992.
151. Syftestad, G.T. and Urist, M.R., Bone aging, Clin. Orthop., 16, 288, 1982.
152. Magee, F.P., Long, J.A., and Hedley, A.K., The effect of age on the interface between porous coated
implants and bone, Orthop. Trans., 13, 455, 1989.
153. Nakajima, I., Dai, K.R., Kelly, P.J., and Chao, E.Y., The effect of age on bone ingrowth into titanium
ber-metal segmental prosthesis: an experimental study in a canine model, Orthop. Trans., 6, 296, 1985.
154. Eckhoff, D.G., Turner, A.S., and Aberman, H.M., Effect of age on bone formation around orthopaedic
implants, Clin. Orthop., 312, 253, 1995.
155. Soballe, K., Gotfredsen, K., Brockstedt-Rasmussen, H., et al., Histologic analysis of a retrieved
hydroxyapatite-coated femoral prosthesis, Clin. Orthop., 272, 255, 1991.
156. Branemark, R. and Thomsen, P., Biomechanical and morphological studies on osseointegration in
immunological arthritis in rabbits, Scand. J. Plast. Reconstr. Hand Surg., 31, 185, 1997.
157. Dorr, L.D., Takei, G.K., and Conaty, J.P., Total hip arthroplasties in patients less than 45 years old,
J. Bone Joint Surg., 65A, 474, 1983.
158. Poss, R., Maloney, J.P., Ewald, F.C., et al., Six- to eleven-year results of total hip arthroplasty in
rheumatoid arthritis, Clin. Orthop., 182, 109, 1984.
159. Sochart, D.H. and Porter, M.L., The long-term results of Charnley low-friction arthroplasty in young
patients who have congenital dislocation, degenerative osteoarthritis, or rheumatoid arthritis, J. Bone
Joint Surg., 79A, 1599, 1997.
160. Hanssen, A.D., Cabanela, M.E., and Michet, C.J., Jr., Hip arthroplasty in patients with systemic lupus
erythematosis, J. Bone Joint Surg., 69A, 807, 1987.
161. Huo, M.H., Salvati, E.A., Brone, M.G., et al., Primary total hip arthroplasty in systemic lupus erythe-
matosis, J. Arthroplasty, 7, 51, 1992.
162. Dorr, L.D., Arnala, I., Faugere, M.C., and Malluche, H.H., Five-year post-operative results of cemented
femoral arthroplasty in patients with systemic bone disease, Clin. Orthop., 259, 114, 1990.
163. Hayashi, K., Uenoyama, K., Mashima, T., and Sugioka, Y., Remodelling of bone around hydroxyapatite
and titanium in experimental osteoporosis, Biomaterials, 15, 11, 1994.
0266/ ch28/frame Page 461 Friday, October 22, 1999 10:09 AM
462 Mechanical Testing of Bone and the BoneImplant Interface
164. Fugimoto, T., Niimi, A., and Ueda, M., Effects of steroid-induced osteoporosis on osseointegration
of titanium implants, Int. J. Oral Maxillofac. Implants, 13, 183, 1998.
165. Mori, H., Manabe, M., Kurachi, Y., and Nagumo, M., Osseointegration of dental implants in rabbit
bone with low mineral density, J. Oral Maxillofac. Surg., 55, 351, 1997.
0266/ ch28/frame Page 462 Friday, October 22, 1999 10:09 AM

463

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

29

Implant Pushout
and Pullout Tests

Aivars Berzins and Dale R. Sumner

CONTENTS

I. Introduction ..........................................................................................................................463
II. Features Shared by Pushout and Pullout tests.....................................................................464
III. Factors Specic to Pushout Tests ........................................................................................465
IV. Factors Specic to Pullout Tests..........................................................................................469
V. Development of Nondestructive and Other Test Modes .....................................................469
Acknowledgments ..........................................................................................................................470
References ......................................................................................................................................472

I. INTRODUCTION

Pushout and pullout tests are commonly used to test the

ex vivo

mechanical competence of
biological xation of orthopaedic and dental implants via evaluation of the shear strength of the
boneimplant interface. One of the reasons these tests are widely used is the relative simplicity
of the test protocol, which usually requires a uniaxial materials testing machine to be operated
under displacement control using a simple support jig for the pushout test or a hookup system
for the pullout test.
The most common applications for these tests include testing for the effects of implant material,
surface texture, cross-sectional geometry, porosity, and surface composition in the context of
cementless xation by bone ingrowth or bone apposition to the implant. Other studies have
investigated the effects of the presence of a bone graft, bone graft substitutes, or anabolic growth
factors on or in the space around the implant. Several studies have utilized pushout or pullout tests
to address the effects of anti-inammatory and anticoagulant medication, irradiation, bisphospho-
nates, and local electrical stimulation on the strength of the boneimplant interface. There are a
number of studies where pullout tests have been used to test xation strength of orthopaedic devices,
including screws used for adjuvant xation of articial joints or spinal hardware and intramedullary
rods for fracture xation. A few studies have used pushout tests to test either the cementbone or
cementmetal interfaces in cemented joint replacement.
The purpose of this chapter is to provide a practical guide to conducting these tests. The list
of citations is extensive, but not exhaustive. In addition to the specic papers cited herein, there
are a number of reviews covering topics beyond the scope of this chapter, including animal
models,

1

factors known to augment or inhibit cementless xation,

2,3

and certain aspects of implant
design.

4

This review has limited its citations to peer-reviewed journal articles. Several groups
have multiple publications using the same technique. For these groups, representative papers
have been cited.

0266/ ch29/frame Page 463 Friday, October 22, 1999 10:14 AM

464

Mechanical Testing of Bone and the BoneImplant Interface

II. FEATURES SHARED BY PUSHOUT AND PULLOUT TESTS

Both types of tests are usually used on cylindrical implants that had been implanted

in vivo

and
left

in situ

for various lengths of time. In general, the retrieved specimens are tested either in the
fresh condition or after having been frozen because histological processing methods (e.g., xation
in formalin) alter the mechanical properties of bone and soft tissue. Whether tested fresh or frozen,
the specimens should not be allowed to dry out during the test.
With both pushout and pullout tests, a load is applied to the implant via a device connected to
the crosshead of the materials testing machine and a forcedisplacement curve is recorded
(Figure 29.1). This curve typically shows the force along the vertical axis and the position of the
crosshead along the horizontal axis. The test is run until the boneimplant interface fails, easily
recognized as a sudden displacement of the implant within the host bone. The most commonly
calculated end point of these tests is the ultimate shear strength of the interface (

u

). Its value is
calculated by dividing the maximum pushout or pullout force (

F

max

) by the nominal interface area (

A

):

u



=

F

max

/

A

(29.1)
For a cylindrical implant,

A

=

DL

(29.2)
where

D

is the outer diameter of the cylindrical implant and

L

is the length of the implant in contact
with bone. For transcortical implants, the value of

L

is measured on the specimen, usually at multiple
points. For implants cut into slab sections, the value of

L

is simply the thickness of the slab. For
pullout specimens

L

is usually the length of the implant (or in the case of porous-coated implants,
the length of the porous-coating region). Thus, the results are usually reported in terms of force

FIGURE 29.1

A typical loaddisplacement curve from a pushout or pullout test. F = the maximum force
applied to the implant during the test; E

= apparent shear stiffness (the slope of the loaddisplacement curve
in its linear region); EA = energy absorption to failure. (Modied from Sballe, K. et al.,

J. Arthroplasty,

6,
307, 1991. With permission.)

0266/ ch29/frame Page 464 Friday, October 22, 1999 10:14 AM

Implant Pushout and Pullout Tests

465

per unit area (i.e., the maximum shear stress). As a simplied measure for relative comparisons of
two implants with the same test conditions, geometry, and nominal bone contact area, one can
simply use

F

max

. Certain authors also report the results in terms of the yield strength, i.e., the force
at which the loaddeformation curve becomes nonlinear.
In addition to the ultimate shear strength of the interface, quantities called the apparent shear
stiffness, the apparent elastic modulus, or the interface stiffness have been calculated as either

F/d

(29.3)
or as
(

F/A

)/

d

(29.4)
where

d

is the displacement. These calculations are based on the assumption that the interface is
innitely thin. However, shear depends upon an angular deformation and the interface must therefore
have a nite thickness.

5,6

The denition of this thickness is somewhat arbitrary, but in reality is
represented by the difference in the diameter of the implant and, for pushout tests, the diameter of
the hole in the support jig (i.e., the clearance listed in Table 29.1, see below). An exact formula
for calculating the interface shear modulus in pullout tests has recently been derived and validated,

5

and would work equally well for pushout tests:

G

=

F

((ln

R

2

ln

R

1

)/2

dL

) (29.5)
where

G

is the interface shear modulus,

F

is the applied force,

R

1

is the radius of the implant,

R

2

is the radius of the support hole, and

d

and

L

are as dened above (see also Reference 6). Energy
absorption at the interface to failure can be measured as the area under the forcedisplacement curve.
It must be understood that the introduction of the displacement measure, as recorded from the
crosshead of the materials testing machine, leads to the potential for a systematic error related to
the compliance of the testing setup. Thus, the compliance of the setup should be tested rst in
order to adjust the displacement readings. This is probably more critical for pullout tests than for
pushout tests because the test setup tends to be less rigid with pullout tests (see below). An alternative
method to using the displacement at the crosshead is to take measurements from a transducer in
close proximity to the interface.

5

In addition, the rate of displacement could affect the results because bone is a viscoelastic
material. On the one hand, Shirazi-Adl et al.

7

found that the pullout force of screws from poly-
urethane models was not dependent on the displacement rate, which was varied from 0.6 to
600 mm/min in their study. On the other hand, Raab et al.

8

showed a clear strain rate dependence
in their system for testing the implantbonecement interface. Specically, there was minimal effect
for strain rates varying from 10

3

/s to 10

1

/s (roughly, a displacement rate of 1 to 100 mm/min for
the implants tested), but an increasingly signicant effect at faster strain rates.

III. FACTORS SPECIFIC TO PUSHOUT TESTS

Typical anatomical locations for pushout specimens are the cortex of femur, tibia, or mandible. In
the long bones, this implant placement is often referred to as transcortical. Occasionally, speci-
mens placed in a trabecular bone bed are tested with this method. After retrieval the specimen is
prepared for analysis. For the transcortical implants, this involves trimming or machining of bone
so that the specimen can be properly placed in the testing jig (Figure 29.2). For the implants placed
in trabecular bone, a slab section is usually prepared. With both types of implants, the specimen
is placed on a support jig with clearance for the implant.

0266/ ch29/frame Page 465 Friday, October 22, 1999 10:14 AM

466

Mechanical Testing of Bone and the BoneImplant Interface

T
A
B
L
E

2
9
.
1
S
u
m
m
a
r
y

o
f

T
e
s
t

C
o
n
d
i
t
i
o
n
s

f
o
r

P
u
s
h
o
u
t

T
e
s
t
s

R
e
f
.
Y
e
a
r
R
e
l
a
t
e
d

R
e
f
s
.
I
m
p
l
a
n
t

S
i
z
e

(
m
m
)

a

S
u
p
p
o
r
t

C
l
e
a
r
a
n
c
e

(
m
m
)
D
i
s
p
l
a
c
e
m
e
n
t

R
a
t
e

(
m
m
/
m
m
)
O
u
t
p
u
t
s

2
8
1
9
7
1
2
9
3
.
2

1
2
.
7
N
A
0
.
5
S
h
e
a
r

s
t
r
e
n
g
t
h
3
0
1
9
7
2
6
.
3
5

6
.
3
5
0
.
1
2
7
0
.
1
2
7
S
h
e
a
r

s
t
r
e
n
g
t
h

a
t

y
i
e
l
d
,

m
a
x
.

s
h
e
a
r

s
t
r
e
n
g
t
h
3
1
1
9
7
3
4
.
5

6
.
4

4
.
8

6
.
4
N
A
1
.
2
7
S
h
e
a
r

s
t
r
e
n
g
t
h
3
2
1
9
7
6
4
.
8

1
0
N
A
5
S
h
e
a
r

s
t
r
e
n
g
t
h
3
3
1
9
7
6
9

2
0
M
i
n
i
m
a
l
1
.
2
7
S
h
e
a
r

s
t
r
e
n
g
t
h
3
4
1
9
7
8
~
5

2
N
A
N
A
S
h
e
a
r

s
t
r
e
n
g
t
h
3
5
1
9
8
0
4
.
5

9
0
.
2
5
0
.
5
S
h
e
a
r

s
t
r
e
n
g
t
h
3
6
1
9
8
0
6
.
5

8
1
.
9
1
.
2
7

o
r

0
.
5
S
h
e
a
r

s
t
r
e
n
g
t
h
3
7
1
9
8
1
1
0

4
N
A
1
.
2
7
S
h
e
a
r

s
t
r
e
n
g
t
h
,

i
n
t
e
r
f
a
c
e

s
t
i
f
f
n
e
s
s
3
8
1
9
8
1
6
.
4

6
.
4
5
N
A
0
.
0
4
9
8
S
h
e
a
r

s
t
r
e
n
g
t
h
,

a
p
p
a
r
e
n
t

e
l
a
s
t
i
c

m
o
d
u
l
u
s
8
1
9
8
1
3
9
6
.
4

1
2
.
8
0

?
1

b

C
r
i
t
i
c
a
l

v
a
l
u
e

o
f

t
h
e

s
t
r
a
i
n

e
n
e
r
g
y

r
e
l
e
a
s
e

r
a
t
e
4
0
1
9
8
2
1
0
.
5

1
2
.
5

2
0
N
A
2
5
.
4
S
h
e
a
r

s
t
r
e
n
g
t
h
1
9
1
9
8
4
6

~
8
N
A
1
.
2
S
h
e
a
r

s
t
r
e
n
g
t
h
,

i
n
t
e
r
f
a
c
e

s
t
i
f
f
n
e
s
s

(
r
e
p
e
t
i
t
i
v
e

l
o
a
d
i
n
g
/
l
o
a
d

c
o
n
t
r
o
l
)
4
1
1
9
8
5
4
.
7
5

o
r

5

1
0
N
A
0
.
5
S
h
e
a
r

s
t
r
e
n
g
t
h
4
2
1
9
8
5
6

~
9
N
A
1
2
.
7
S
h
e
a
r

s
t
r
e
n
g
t
h
,

i
n
t
e
r
f
a
c
e

s
t
i
f
f
n
e
s
s

(
r
e
p
e
t
i
t
i
v
e

l
o
a
d
i
n
g
/
l
o
a
d

c
o
n
t
r
o
l
)
4
3
1
9
8
5
4
4
,
4
5
6

~
9
0
.
4
1
.
2
7
S
h
e
a
r

s
t
r
e
n
g
t
h
,

s
h
e
a
r

s
t
i
f
f
n
e
s
s
9
1
9
8
5
6

~
8
0
.
1
5
0
.
1
2
7
S
h
e
a
r

s
t
r
e
n
g
t
h
,

s
h
e
a
r

s
t
i
f
f
n
e
s
s

(
r
e
p
e
t
i
t
i
v
e

l
o
a
d
i
n
g
/
l
o
a
d

c
o
n
t
r
o
l
)
4
6
1
9
8
6
8

1
2
.
7

6
.
3
5
N
A
1
2
7
S
h
e
a
r

s
t
r
e
n
g
t
h
,

i
n
t
e
r
f
a
c
e

s
t
i
f
f
n
e
s
s
,

e
n
e
r
g
y

t
o

f
a
i
l
u
r
e
4
7
1
9
8
9
5

1
0
N
A
0
.
1
S
h
e
a
r

s
t
r
e
n
g
t
h
4
8
1
9
9
0
2
6
,
4
9

5
3
6

4
0
.
5
5
S
h
e
a
r

s
t
r
e
n
g
t
h
,

a
p
p
a
r
e
n
t

s
h
e
a
r

s
t
i
f
f
n
e
s
s
,

e
n
e
r
g
y

a
b
s
o
r
p
t
i
o
n

0266/ ch29/frame Page 466 Friday, October 22, 1999 10:14 AM

Implant Pushout and Pullout Tests

467

5
4
1
9
9
2
2
7
5
.
9
5

9
0
.
1
5
1
.
2
7
S
h
e
a
r

s
t
r
e
n
g
t
h
5
5
1
9
9
2
6
.
3

~
6
.
5
N
A
2
5
.
4
S
h
e
a
r

s
t
r
e
n
g
t
h
1
1
1
9
9
3
6

1
3
1
0
.
5
S
h
e
a
r

s
t
r
e
n
g
t
h
5
6
1
9
9
3
4
.
7
6

1
2
N
A
0
.
2
S
h
e
a
r

s
t
r
e
n
g
t
h
5
7
1
9
9
3
2
.
5

2
0
N
A
1
S
h
e
a
r

s
t
r
e
n
g
t
h
,

e
n
e
r
g
y

a
b
s
o
r
p
t
i
o
n
5
8
1
9
9
5
5
.
5

8
1
.
0
2
5
1
S
h
e
a
r

s
t
r
e
n
g
t
h
5
9
1
9
9
5
6
.
8

1
0

o
r

1
5
N
A
0
.
5
S
h
e
a
r

s
t
r
e
n
g
t
h
6
0
1
9
9
5
4

1
5
N
A
1
S
h
e
a
r

s
t
r
e
n
g
t
h
6
1
1
9
9
5
5
.
9
5

2
2
0
.
1
5
1
.
3
S
h
e
a
r

s
t
r
e
n
g
t
h
6
2
1
9
9
5
6

1
3
0
.
5
0
.
5
S
h
e
a
r

s
t
r
e
n
g
t
h
6
3
1
9
9
5
6
4
2
.
8

6
N
A
1
S
h
e
a
r

s
t
r
e
n
g
t
h
1
0
1
9
9
5
2

5

c
o
n
i
c
a
l

s
h
a
p
e
~
1
.
5
0
.
5
S
h
e
a
r

s
t
r
e
n
g
t
h
6
5
1
9
9
5
6
.
5

o
r

6
.
6

2
4

c

N
A
0
.
0
5
S
h
e
a
r

s
t
r
e
n
g
t
h
6
6
1
9
9
6
7

1
5

1
0
N
A
0
.
2
5
M
a
x
i
m
u
m

f
o
r
c
e
6
7
1
9
9
6
1
0

6
0
.
1
5
1
.
2
7
S
h
e
a
r

s
t
r
e
n
g
t
h
6
8
1
9
9
7
5
.
3

5
.
3
5
N
A
1
H
o
l
d
i
n
g

s
t
r
e
n
g
t
h
,

h
o
l
d
i
n
g

s
t
r
e
s
s
6
9
1
9
9
7
6
.
4

5
N
A
0
.
5
S
h
e
a
r

s
t
r
e
n
g
t
h
7
0
1
9
9
7
6

2
0
~
1
.
0
0
.
5
S
h
e
a
r

s
t
r
e
n
g
t
h
7
1
1
9
9
8
8

1
0

4
0
.
5
0
.
5
S
h
e
a
r

s
t
r
e
n
g
t
h
7
2
1
9
9
8
7
3
4

9

o
r

1
0
N
A
2
P
u
s
h
o
u
t

f
o
r
c
e
N
A

=

i
n
f
o
r
m
a
t
i
o
n

w
a
s

n
o
t

a
v
a
i
l
a
b
l
e

f
r
o
m

t
h
e

p
u
b
l
i
c
a
t
i
o
n
.

a

S
o
m
e

i
m
p
l
a
n
t
s

w
e
r
e

s
e
c
t
i
o
n
e
d

b
e
f
o
r
e

t
e
s
t
i
n
g
;

t
h
e
r
e
f
o
r
e
,

o
n
l
y

t
h
e

t
e
s
t
e
d

s
i
z
e

i
s

l
i
s
t
e
d

(
o
u
t
s
i
d
e

d
i
a
m
e
t
e
r

l
e
n
g
t
h
)
.

b

C
a
l
c
u
l
a
t
e
d

f
r
o
m

a

r
e
p
o
r
t
e
d

s
t
r
a
i
n

r
a
t
e

o
f

1
.
3

1
0

3

/
s
,

a
s
s
u
m
i
n
g

t
h
e

o
r
i
g
i
n
a
l

l
e
n
g
t
h

w
a
s

1
2
.
8

m
m
.

c

E
s
t
i
m
a
t
e
d

f
r
o
m

a

r
a
d
i
o
g
r
a
p
h

i
n

t
h
e

m
a
n
u
s
c
r
i
p
t

(
e
x
a
c
t

v
a
l
u
e

n
o
t

s
p
e
c
i

e
d
)
.

0266/ ch29/frame Page 467 Friday, October 22, 1999 10:14 AM

468

Mechanical Testing of Bone and the BoneImplant Interface

Accurate alignment among the plunger, the specimen, and the support jig is crucial to prevent
the implant from jamming within its track of exit. Thus, additional features of the support xture
can be incorporated to allow for self-alignment of the longitudinal axis of the specimen with the
longitudinal axis of the plunger. In cases where a slab section containing part of the original implant
is tested, it is important to have accurate cutting jigs so that the slab section can be cut perpendicular
to the long axis of the implant. Another important test condition is perfect seating of the specimen
on the support xture. Machining the base of the specimen to assure ideal t and alignment with
the support jig has been used in several studies.

9-11

In addition, the specimen can be afxed to the
support jig by cementing.

11

As noted, the strength of xation is calculated as the ultimate shear stress of the interface. This
simplied stress calculation assumes a uniform stress distribution at the interface. In reality,
however, the stress distribution along the interface is not uniform. Thus, it becomes crucial to avoid
comparisons between two implants with dissimilar interface stress distributions, i.e., implants with
different geometries or testing conditions that could affect the distribution of the interface stresses.
In order to establish grounds for standardization of these tests, two theoretical studies have
addressed the inuence of pushout test conditions using the nite-element method. Harrigan et al.

12

investigated the effect of irregular stress distribution due to stress concentration in the region
between the base of the specimen and the support xture. They concluded that pushout test results
depend heavily on uniform support of the specimen (i.e., close apposition of two at surfaces).
Dhert et al.

13

addressed a number of pushout test variables, including clearance of the implant
within the hole in the support jig, the elastic modulus of the implant, cortical thickness (length of
the implant), and diameter of the implant. By analyzing stress distributions at the interface, they
found that the clearance of the hole is the most critical parameter, followed by the elastic modulus
of the implant. Based on this nite-element study, these authors recommended a clearance of at

FIGURE 29.2

Schematic of a typical mechanical test fixture used in pushout tests. The support tube has an
inner diameter greater than the diameter of the implant, creating a clearance. (Modied from Cook, S.D. et al.,

J. Arthroplasty,

10, 351, 1995. With permission.)

0266/ ch29/frame Page 468 Friday, October 22, 1999 10:14 AM

Implant Pushout and Pullout Tests

469

least 0.7 mm to minimize nonuniform stress distribution and to compare only those implants which
are made from materials with similar elastic moduli. According to this nite-element study, vari-
ations in the length and diameter of the implants did not signicantly affect the interface stress
distribution and, thus, could be neglected. However, it should be noted that implants with a larger
length-to-diameter ratio would tend to be less stable during the test and more prone to develop
misalignment with the axis of the applied pushout force.
A wide range of implant sizes and testing conditions have been used in pushout tests (see
Table 29.1). Many of these appear to have been conducted in contrast to the recommendations from
the theoretical studies. Thus, caution is needed when comparing results among investigators or
when exact values are needed for modeling interface conditions, as might be done in a nite-
element analysis of a hip or knee replacement. Comparisons within a study are probably valid.

IV. FACTORS SPECIFIC TO PULLOUT TESTS

Typical implants used with pullout tests are cylindrical; however, there have been reports about
implants with elliptical

14

or trapezoid

15

shapes and combined conicalcylindrical shapes.

16

Unlike
the implants for the pushout test protocol, the pullout implants are usually placed in anatomical
locations allowing for signicant contact with cancellous bone by being aligned with the long
axis of the bone in the proximal or distal femur, proximal tibia, or proximal humerus. The greatest
advantage of the pullout test setup is that the presence of either a exible linkage

17

or an

XY

slide table

16,18

allows the specimen to self-align during the test, thus eliminating concerns about
potential misalignment between the axes of the implant and the pullout force. In a typical pullout
test, the bone part of the specimen is potted in a self-hardening polymer cement and attached to
the materials testing machine (Figure 29.3). A pulling rod is then attached to the implant, usually
via a threaded piece on the implant, and the implant is pulled out, typically under displacement
control of the material testing machine.
Similar to the reports on pushout tests, there are a wide variety of test conditions for pullout
test protocols with a potential to bias the test results (Table 29.2). After testing a number of
pullout conditions using polyurethane foam and cadaver bones, Shirazi-Adl et al.

7

found that
material arrangements and boundary conditions can substantially alter the pullout force. Thus,
comparisons of results of pullout tests with different design arrangements should be approached
with caution. Berzins et al.

5

used polyurethane foam to test for the effects of implant geometry
using combinations of three diameters (5 to 9.7 mm) and three lengths (15 to 48 mm) for a total
of nine implant sizes. It was found that the interface shear modulus was signicantly affected
by the length of the implants, but the ultimate xation strength was unaffected by implant diameter
and length in the range studied.

V. DEVELOPMENT OF NONDESTRUCTIVE AND OTHER TEST MODES

Because the standard pushout and pullout tests are destructive, the ability to examine the intact
interface after the test is lost. In the presence of accurate direct displacement monitoring at the
interface, it is possible to develop a nondestructive pushout or pullout protocol to measure the
interface shear modulus.

5,19

To date, the accuracy of measuring the shear modulus nondestructively
has now been validated using a model system (polyurethane foam),
5
supporting the concept of
using a nondestructive test to characterize the mechanical and morphological properties of the intact
interface in the same specimen.
19
In addition, the shear modulus of the interface may prove to be
more reective of the type of tissue at the interface than the ultimate shear strength.
Torque tests have been reported by several groups (Table 29.3). These tests have the theoretical
advantage of a uniform stress eld at the interface, but the experimental setup is more elaborate
than that required for a pushout or pullout test. The recent paper by Brnemark et al.
6
provides a
0266/ ch29/frame Page 469 Friday, October 22, 1999 10:14 AM
470 Mechanical Testing of Bone and the BoneImplant Interface
useful description of how interfacial shear strength and the torsional shear modulus can be calcu-
lated. Another test method has been to test the tensile bonding strength of cortical bone to a
transcortical implant.
20
Finally, a method for testing the fatigue properties of the interface between an implant and
bone cement was described nearly 20 years ago
8
and a method for testing the fatigue properties
of porous surfaces was described nearly 15 years ago,
21
but to the authors knowledge the fatigue
properties of the cementless boneimplant interface have received little attention.
7,22
With the
growing interest in the role of fatigue on bone remodeling in general
23,24
and the probable
importance of the fatigue properties of the bone supporting an implant for the long-term main-
tenance of a biological interface,
25
one can only presume that more effort along these lines will
be made in the future.
ACKNOWLEDGMENTS
Research was supported by NIH Grants AR16485 and AR42862.
FIGURE 29.3 Schematic of a mechanical test fixture used in a pullout test in which the investigators
calculated the interface shear modulus (longitudinal cross section). Note: 1 = aluminum mold, 2 = cylindrical
implant, 3 = foam model (bone, in a biological sample), 4 = extended yoke, 5 = sliding sleeve held against
the foam (bone) surface by springs, 6 = extensometer, 7 = foam (bone) layer in which the interface shear
modulus is measured, F = tensile force applied to the setup (Modied from Berzins, A. et al., J. Biomed.
Mater. Res., 34, 337, 1997. With permission.)
0266/ ch29/frame Page 470 Friday, October 22, 1999 10:14 AM
Implant Pushout and Pullout Tests 471
TABLE 29.2
Summary of Test Conditions for Pullout Tests
Ref. Year
Related
Refs.
Implant
Size
(mm)
c
Displacement
Rate
(mm/mm) Outputs
17 1971 9.52 12.7 12.7 Shear strength
74 1972 9 50 125 Shear strength
75 1973 22.225 staples 1.27 Maximum force
76 1977 8.5 or 9 85 2,700 Shear strength
77 1980 Plate pulloff test (exact dimensions
not given)
0.5 Fixation (tensile) strength
78 1984 ~4.7 19 50 Maximum force
79 1985 5 10 60 Shear strength
80 1986 3.2 ~40 12.7 Shear strength
81 1987 82,83 10 50 0.25 Shear strength
84 1990 10 5 5 (rectangular) NA Shear strength
85 1990 10 17 90.6 kgf/s
(load control)
Maximum force
86 1992 7 30 0.25 Shear strength
87 1992 2.65 NA 1.52 Maximum force
14 1993 ~4 NA (cylindrical or elliptical) 2 Shear strength
88 1994 5 25 2 Shear strength
7 1994 5.75 to 8.75 40 9.6 Pullout force
18 1995 5 10 2 Shear strength
16 1995 NA 3 Shear strength
15 1995 7 10 5 (trapezoid) 2 Shear load, duration of peak
89 1995 90 Pedicle screws and hooks 0.5 Maximum force
91 1995 6.5 55 10 Maximum force
92 1996 6.257 3040 60 Pullout force
93 1997 3.255 12 2 Pullout force, shear strength
5 1997 59.7 1548 0.25 Shear strength, shear
modulus
6 1997 94 2 NA 0.5 Pullout load
95 1998 ~3.1 6 0.5 Pullout load
96 1998 5 25 2 Shear strength
NA = information was not available from the publication.
a
Some implants were sectioned before testing; therefore, only the tested size is listed (outside diameter length).
TABLE 29.3
Summary of Test Conditions for Torque Tests
Ref. Year
Related
Refs.
Size of Tested
Implant,
a
mm
Displacement
Rate Outputs
30 1972 6.35 9.525 NA Maximum torque
97 1988 98,99 16 ~30 15/min Torsional stiffness, maximum torque,
energy absorption
100 1992 14 4 10 57/min Torsional shear strength, energy absorption
101 1995 3.5 10 NA Removal torque
6 1997 94 2 NA 2/min Break point torque, maximum torque
102 1997 3.75 10 na Failure torque
95 1998 ~3.1 6 2/min Yield torque, maximum torque, rotation at yield,
rotation at maximum
NA = information was not available from the publication.
a
Some implants were sectioned before testing; therefore, only the tested size is listed (outside diameter length).
0266/ ch29/frame Page 471 Friday, October 22, 1999 10:14 AM
472 Mechanical Testing of Bone and the BoneImplant Interface
REFERENCES
1. Sumner, D.R., Turner, T.M., and Urban, R.M., Animal models of bone ingrowth and joint replacement,
in Animals Models in Orthopedic Research, An, Y.H. and Friedman, R.J., Eds., CRC Press, Boca
Raton, FL, 1999, 407.
2. Kienapfel, H. and Griss, P., Fixation by ingrowth, in The Adult Hip, Callaghan, J.J., Rosenberg, A.G.,
and Rubash, H., Eds., Lippincott-Raven, Philadelphia, 1998, 201.
3. Sumner, D.R., Bone ingrowth: implications for establishment and maintenance of cementless porous-
coated interfaces, in Orthopaedic Knowledge Update: Hip and Knee Reconstruction, Callaghan, J.J.,
Dennis, D.A., Paprosky, W.G., and Rosenberg, A.G., Eds., American Academy of Orthopaedic Sur-
geons, Rosemont, IL, 1995, 57.
4. Pilliar, R.M., Porous-surfaced metallic implants for orthopedic applications, J. Biomed. Mater. Res.,
21(A1), 1, 1987.
5. Berzins, A., Shah, B., Weinans, H., and Sumner, D.R., Nondestructive measurements of implantbone
interface shear modulus and effects of implant geometry in pullout tests, J. Biomed. Mater. Res., 34,
337, 1997.
6. Brnemark, R., Ohrnell, L.O., Nilsson, P., and Thomsen, P., Biomechanical characterization of osseoin-
tegration during healing: an experimental in vivo study in the rat, Biomaterials, 18, 969, 1997.
7. Shirazi-Adl, A., Dammak, M., and Zukor, D.J., Fixation pullout response measurement of bone screws
and porous-surfaced posts, J. Biomech., 27, 1249, 1994.
8. Raab, S., Ahmed, A.M., and Provan, J.W., The quasistatic and fatigue performance of the
implant/bonecement interface, J. Biomed. Mater. Res., 15, 159, 1981.
9. Thomas, K. A. and Cook, S. D., An evaluation of variables inuencing implant xation by direct bone
apposition, J. Biomed. Mater. Res., 19, 875, 1985.
10. Wie, H., Hero, H., Solheim, T., et al., Bonding capacity in bone of HIP-processed HA-coated titanium:
mechanical and histological investigations, J. Biomed. Mater. Res., 29, 1443, 1995.
11. Hayashi, K., Inadome, T., Mashima, T., and Sugioka, Y., Comparison of boneimplant interface shear
strength of solid hydroxyapatite and hydroxyapatite-coated titanium implants, J. Biomed. Mater. Res.,
27, 557, 1993.
12. Harrigan, T.P., Kareh, J., and Harris, W.H., The inuence of support conditions in the loading xture
on failure mechanisms in the pushout test: a nite element study, J. Orthop. Res., 8, 678, 1990.
13. Dhert, W.J., Verheyen, C.C., Braak, L.H., et al., A nite element analysis of the pushout test: inuence
of test conditions, J. Biomed. Mater. Res., 26, 119, 1992.
14. Cook, S.D., Salkeld, S.L., Gaisser, D.M., and Wagner, W.R., An in vivo analysis of an elliptical dental
implant design, J. Oral Implantol., 19, 307, 1993.
15. Li, Z., Kitsugi, T., Yamamuro, T., et al., Bone-bonding behavior under load-bearing conditions of an
alumina ceramic implant incorporating beads coated with glass-ceramic containing apatite and wol-
lastonite, J. Biomed. Mater. Res., 29, 1081, 1995.
16. Maruyama, M., Hydroxyapatite clay used to ll the gap between implant and bone, J. Bone Joint
Surg., 77B, 213, 1995.
17. Galante, J., Rostoker, W., Lueck, R., and Ray, R.D., Sintered ber metal composites as a basis for
attachment of implants to bone, J. Bone Joint Surg., 53A, 101, 1971.
18. Dean, J.C., Tisdel, C.L., Goldberg, V.M., et al., Effects of hydroxyapatite tricalcium phosphate coating
and intracancellous placement on bone ingrowth in titanium bermetal implants, J. Arthroplasty, 10,
830, 1995.
19. Anderson, R.C., Cook, S.D., Weinstein, A.M., and Haddad, R.J., An evaluation of skeletal attachment
to LTI pyrolytic carbon, porous titanium, and carbon-coated porous titanium implants, Clin. Orthop.,
182, 242, 1984.
20. Park, J.B., Salman, N.N., Kenner, G.H., and Von Recum, A.F., Preliminary studies on the effects of
direct current on the bone/porous implant interfaces, Ann. Biomed. Eng., 8, 93, 1980.
21. Manley, M.T., Kotzar, G., Stern, L.S., and Wilde, A., Effects of repetitive loadings on the integrity of
porous coatings, Clin. Orthop., 217, 293, 1987.
22. Lim, T.-H., An, H.S., Hasegawa, T., et al., Prediction of fatigue screw loosening in anterior screw
xation using dual energy X-ray absorptiometry, Spine, 20, 2565, 1995.
0266/ ch29/frame Page 472 Friday, October 22, 1999 10:14 AM
Implant Pushout and Pullout Tests 473
23. Burr, D.B., Martin, R.B., Schafer, M.B., and Radin, E.L., Bone remodeling in response to in vivo
fatigue microdamage, J. Biomech., 18, 189, 1985.
24. Martin, B., Mathematical model for repair of fatigue damage and stress fracture in osteonal bone,
J. Orthop. Res., 13, 309, 1995.
25. Sumner, D.R., Bone remodeling of the proximal femur, in The Adult Hip, Callaghan, J.J., Rosenberg,
A.G., and Rubash, H., Eds., Lippincott-Raven, New York, 1998, 211.
26. Sballe, K., Hansen, E.S., Brockstedt-Rasmussen, H., et al., Fixation of titanium and hydroxyapatite-
coated implants in arthritic osteopenic bone, J. Arthroplasty, 6, 307, 1991.
27. Cook, S.D., Barrack, R.L., Dalton, J.E., et al., Effects of indomethacin on biologic xation of porous-
coated titanium implants, J. Arthroplasty, 10, 351, 1995.
28. Welsh, R.P., Pilliar, R.M., and MacNab, I., Surgical implants: the role of surface porosity in xation
to bone and acrylic, J. Bone Joint Surg., 53A, 963, 1971.
29. Cameron, H.U., Pilliar, R.M., and MacNab, I., The rate of bone ingrowth in porous metal, J. Biomed.
Mater. Res., 10, 295, 1976.
30. Predecki, P., Auslaender, B.A., Stephan, J.E., et al., Attachment of bone to threaded implants by
ingrowth and mechanical interlocking, J. Biomed. Mater. Res., 6, 401, 1972.
31. Nilles, J.L., Coletti, J.M., and Wilson, C., Biomechanical evaluation of bone-porous material interfaces,
J. Biomed. Mater. Res., 7, 231, 1973.
32. Robertson, D.M., St. Pierre, L., and Chahal, R., Preliminary observations of bone ingrowth into porous
materials, J. Biomed. Mater. Res., 10, 335, 1976.
33. Weinstein, A.M., Klawitter, J.J., Cleveland, T.W., and Amoss, D.C., Electrical stimulation of bone
growth into porous A1
2
O
3
, J. Biomed. Mater. Res. Symp., 10, 231, 1976.
34. Young, S.O., Park, J.B., Kenner, G.H., et al., Dental implant xation by electrically mediated process.
I. Interfacial strength, Biomater. Med. Dev. Artif. Organs, 6, 111, 1978.
35. Bobyn, J.D., Cameron, H.U., Pilliar, R.M., and Weatherly, G.C., The optimum pore size for the xation
of porous-surfaced metal implants by the ingrowth of bone, Clin. Orthop., 150, 263, 1980.
36. Ducheyne, P., Hench, L.L., Kagan, II, A., et al., Effect of hydroxyapatite impregnation on skeletal
bonding of porous coated implants, J. Biomed. Mater. Res., 14, 225, 1980.
37. Clemov, A.J., Weinstein, A.M., Klawitter, J.J., et al., Interface mechanics of porous titanium implants,
J. Biomed. Mater. Res., 15, 73, 1981.
38. Colella, S.M., Miller, A.G., Stang, R.G., et al., Fixation of porous titanium implants in cortical bone
enhanced by electrical stimulation, J. Biomed. Mater. Res., 15, 37, 1981.
39. Ahmed, A., Raab, S. and Miller, J.E., Metal/cement interface strength in cemented stem xation,
J. Orthop. Res., 2, 105, 1984.
40. Rhinelander, F.W., Stewart, C.L., Wilson, J.W., et al., Growth of tissue into a porous, low modulus
coating on intramedullary nails: an experimental study, Clin. Orthop., 164, 293, 1982.
41. Brooker, A.F. and Constable, D., Bone ingrowth into titanium grooves: a comparison of surfaces for
biologic xation, Adv. Orthop. Surg., 125, 1985.
42. Cook, S.D., Walsh, K.A., and Haddad, R.J., Interface mechanics and bone growth into porous Co-
Cr-Mo alloy implants, Clin. Orthop., 193, 271, 1985.
43. Thomas, K.A., Cook, S.D., Renz, E.A., et al., The effect of surface treatments on the interface
mechanics of LTI pyrolytic carbon implants, J. Biomed. Mater. Res., 19, 145, 1985.
44. Thomas, K.A., Kay, J.F., Cook, S.D., and Jarcho, M., The effect of surface macrotexture and hydrox-
ylapatite coating on the mechanical strengths and histologic proles of titanium implant materials,
J. Biomed. Mater. Res., 21, 1395, 1987.
45. Cook, S.D., Thomas, K.A., Kay, J.F., and Jarcho, M., Hydroxyapatite-coated porous titanium for use
as an orthopedic biologic attachment system, Clin. Orthop., 230, 303, 1988.
46. Heck, D.A., Nakajima, I., Kelly, P.J., and Chao, E.Y., The effect of load alteration on the biological
and biomechanical performance of a titanium ber-metal segmental prosthesis, J. Bone Joint Surg.,
68A, 118, 1986.
47. Oonishi, H., Yamamoto, M., Ishimaru, H., et al., The effect of hydroxyapatite coating on bone growth
into porous titanium alloy implants, J. Bone Joint Surg., 71B, 213, 1989.
48. Sballe, K., Hansen, E.S., Brockstedt-Rasmussen, H., et al., Hydroxyapatite coating enhances xation
of porous coated implants: a comparison in dogs between press t and noninterference t, Acta Orthop.
Scand., 61, 299, 1990.
0266/ ch29/frame Page 473 Friday, October 22, 1999 10:14 AM
474 Mechanical Testing of Bone and the BoneImplant Interface
49. Sballe, K., Hansen, E.S., Brockstedt-Rasmussen, H., et al., Bone graft incorporation around titanium-
alloy and hydroxyapatite-coated implants in dogs, Clin. Orthop., 274, 282, 1992.
50. Sballe, K., Hansen, E.S., Brockstedt-Rasmussen, H., et al., Gap healing enhanced by hydroxyapatite
coating in dogs, Clin. Orthop., 272, 300, 1991.
51. Lind, M., Overgaard, S., Ongpipattanakul, B., et al., Transforming growth factor- stimulates bone
ongrowth to weight-loaded tricalcium phosphate coated implants, J. Bone Joint Surg., 78B, 377, 1996.
52. Lind, M., Overgaard, S., Nguyen, T., et al., Transforming growth factor- stimulates bone ongrowth:
hydroxyapatite-coated implants studied in dogs, Acta Orthop. Scand., 67, 611, 1996.
53. Overgaard, S., Lind, M., Glerup, H., et al., Porous-coated vs. grit-blasted surface texture of hydroxya-
patite-coated implants during controlled micromotion, J. Arthroplasty, 13, 449, 1998.
54. Cook, S.D., Thomas, K.A., Dalton, J E., et al., Hydroxylapatite coating of porous implants improves
bone ingrowth and interface attachment strength, J. Biomed. Mater. Res., 26, 989, 1992.
55. Luckey, H.A., Lamprecht, E.G., and Walt, M.J., Bone apposition to plasma-sprayed cobalt-chromium
alloy, J. Biomed. Mater. Res., 26, 557, 1992.
56. Wang, B.C., Lee, T.M., Chang, E., and Yang, C.Y., The shear strength and the failure mode of plasma-
sprayed hydroxyapatite coating to bone: the effect of coating thickness, J. Biomed. Mater. Res., 27,
1315, 1993.
57. Shen, W.J., Chung, K.C., Wang, G.J., et al., Demineralized bone matrix in the stabilization of porous-
coated implants in bone defects in rabbits, Clin. Orthop., 293, 346, 1993.
58. Friedman, R.J., Bauer, T.W., Garg, K., et al., Histological and mechanical comparison of hydroxya-
patie-coated cobalt-chrome and titanium implants in the rabbit femur, J. Appl. Biomater., 6, 231, 1995.
59. Rashmir-Raven, A.M., Richardson, D.C., Aberman, H.M., and DeYoung, D.J., The response of
cancellous and cortical canine bone to hydroxylapatite-coated and uncoated titanium rods, J. Appl.
Biomater., 6, 237, 1995.
60. Hetherington, V.J., Lord, C.E., and Brown, S.A., Mechanical and histological xation of hydroxyla-
patite-coated pyrolytic carbon and titanium alloy implants: a report of short-term results, J. Appl.
Biomater., 6, 243, 1995.
61. Callahan, B.C., Lisecki, E.J., Banks, R.E., et al., The effect of warfarin on the attachment of bone to
hydroxyapatite-coated and uncoated porous implants, J. Bone Joint Surg., 77A, 225, 1995.
62. Inadome, T., Hayashi, K., Nakashima, Y., et al., Comparison of boneimplant interface shear strength
of hydroxyapatite-coated and alumina-coated metal implants, J. Biomed. Mater. Res., 29, 19, 1995.
63. Li, J., Fartash, B., and Hermansson, L., Hydroxyapatite-alumina composites and bone-bonding,
Biomaterials, 16, 417, 1995.
64. Li, J., Liao, H., Fartash, B., et al., Surface-dimpled commercially pure titanium implant and bone
ingrowth, Biomaterials, 18, 691, 1997.
65. Eckhoff, D.G., Turner, A.S., and Aberman, H.M., Effect of age on bone formation around orthopaedic
implants, Clin. Orthop., 312, 253, 1995.
66. Berzins, A., Sumner, D.R., Wasielewski, R.C., and Galante, J.O., Impacted particulate allograft for
femoral revision total hip arthroplasty: in vitro mechanical stability and effects of cement pressuriza-
tion, J. Arthroplasty, 11, 500, 1996.
67. Walenciak, M.T., Zimmerman, M.C., Harten, R.D., et al., Biomechanical and histological analysis of
an HA coated, arc deposited CPTi canine hip prosthesis, J. Biomed. Mater. Res., 31, 465, 1996.
68. Vemuganti, A., Siegler, S., Abusaeh, A., and Kalidindi, S., Development of self-anchoring bone
implants. II. Boneimplant interface characteristics in vitro, J. Biomed. Mater. Res., 38, 328, 1997.
69. Lewis, C.G., Jones, C.L., and Hungerford, D.S., Effects of grafting on porous metal ingrowth,
J. Arthroplasty, 12, 451, 1997.
70. Nakashima, Y., Hayashi, K., Inadome, T., et al., Hydroxyapatite-coating on titanium arc sprayed
titanium implants, J. Biomed. Mater. Res., 35, 287, 1997.
71. Vercaigne, S., Wolke, J.G., Naert, I., and Jansen, J.A., Histomorphometrical and mechanical evaluation
of titanium plasma-spray-coated implants placed in the cortical bone of goats, J. Biomed. Mater. Res.,
41, 41, 1998.
72. Ogiso, M., Yamamura, M., Kuo, P.T., et al., Comparative pushout test of dense HA implants and HA-
coated implants: ndings in a canine study, J. Biomed. Mater. Res., 39, 364, 1998.
73. Ogiso, M., Yamashita, Y., and Matsumoto, T., Microstructural changes in bone of HA-coated implants,
J. Biomed. Mater. Res., 39, 23, 1998.
0266/ ch29/frame Page 474 Friday, October 22, 1999 10:14 AM
Implant Pushout and Pullout Tests 475
74. Lembert, E., Galante, J., and Rostoker, W., Fixation of skeletal replacement by ber metal composites,
Clin. Orthop., 87, 303, 1972.
75. Cameron, H.U., Pilliar, R.M., and MacNab, I., The effect of movement on the bonding of porous
metal to bone, J. Biomed. Mater. Res., 7, 301, 1973.
76. Ducheyne, P., DeMester, P., and Aernoudt, D., Inuence of a functional dynamic loading on bone
ingrowth into surface pores of orthopedic implants, J. Biomed. Mater. Res., 11, 811, 1977.
77. Bobyn, J.D., Pilliar, R.M., Cameron, H.U., et al., The effect of porous surface conguration on the
tensile strength of xation of implants by bone ingrowth, Clin. Orthop., 149, 291, 1980.
78. McLaughlin, R.E., Reger, S.I., Bolander, M., and Eschenroeder, H.C., Enhancement of bone ingrowth
by the use of bone matrix as a biologic cement, Clin. Orthop., 183, 255, 1984.
79. Barth, E., Ronningen, H., and Solheim, L.F., Comparison of ceramic and titanium implants in cats,
Acta Orthop. Scand., 56, 491, 1985.
80. Berry, J.L., Geiger, J.M., Moran, J.M., et al., Use of tricalcium phosphate or electrical stimulation to
enhance the boneporous implant interface, J. Biomed. Mater. Res., 20, 65, 1986.
81. Rivero, D.P., Skipor, A.K., Singh, M., et al., Effect of disodium etidronate (EHDP) on bone ingrowth
in a porous material, Clin. Orthop., 215, 279, 1987.
82. Rivero, D.P., Fox, J., Skipor, A.K., et al., Calcium phosphate-coated porous titanium implants for
enhanced skeletal xation, J. Biomed. Mater. Res., 22, 191, 1988.
83. Sumner, D.R., Turner, T.M., Pierson, R.H., et al., Effects of radiation on xation of non-cemented
porous-coated implants in a canine model, J. Bone Joint Surg., 72A, 1527, 1990.
84. Ducheyne, P., Beight, J., Cuckler, J., et al., Effect of calcium phosphate coating characteristics on
early post-operative bone tissue ingrowth, Biomaterials, 11, 531, 1990.
85. Wise, M.W., III, Robertson, I.D., Lachiewicz, P.F., et al., The effect of radiation therapy on the xation
strength of an experimental porous-coated implant in dogs, Clin. Orthop., 261, 276, 1990.
86. Kienapfel, H., Sumner, D.R., Turner, T.M., et al., Efcacy of autograft and freeze-dried allograft to
enhance xation of porous coated implants in the presence of interface gaps, J. Orthop. Res., 10, 423,
1992.
87. Chae, J.C., Collier, J.P., Mayor, M.B., et al., Enhanced ingrowth of porous-coated CoCr implants
plasma-sprayed with tricalcium phosphate, J. Biomed. Mater. Res., 26, 93, 1992.
88. Tisdel, C.L., Goldberg, V.M., Parr, J.A., et al., The inuence of hydroxyapatite and tricalcium-
phosphate coating on bone growth in titanium ber-metal implants, J. Bone Joint Surg., 76A, 159,
1994.
89. Berlemann, U., Cripton, P., Nolte, L.P., et al., New means in spinal pedicle hook xation. A bio-
mechanical evaluation, Eur. Spine J., 4, 114, 1995.
90. Berlemann, U., Cripton, P.A., Rincon, L., et al., Pullout strength of pedicle hooks with xation screws:
inuence of screw length and angulation, Eur. Spine J., 5, 71, 1996.
91. Lim, T.H., An, H.S., Evanic, C., et al., Strength of anterior vertebral screw xation in relation to bone
mineral density, J. Spinal Disord., 7, 121, 1995.
92. Kwok, A.W., Finkelstein, J.A., Woodside, T., et al., Insertional torque and pullout strengths of conical
and cylindrical pedicle screws in cadaveric bone, Spine, 21, 2429, 1996.
93. Kido, H., Schulz, E.E., Kumar, A., et al., Implant diameter and bone density: effect on initial stability
and pullout resistance, J. Oral Implantol., 23, 163, 1997.
94. Ohrnell, L.O., Brnemark, R., Nyman, J., et al., Effects of irradiation on the biomechanics of osseoin-
tegration. An experimental in vivo study in rats, Scand. J. Plast. Reconstr. Surg. Hand Surg., 31, 281,
1997.
95. Brnemark, R., hrnell, L.O., Skalak, R., et al., Biomechanical characterization of osseointegration:
an experimental in vivo investigation in the beagle dog, J. Orthop. Res., 16, 61, 1998.
96. Jinno, T., Goldberg, V.M., Davy, D., and Stevenson, S., Osseointegration of surface-blasted implants
made of titanium alloy and cobalt-chromium alloy in a rabbit intramedullary model, J. Biomed. Mater.
Res., 42, 20, 1998.
97. Okada, Y., Suka, T., Sim, F.H., et al., Comparison of replacement prostheses for segmental defects of
bone: different porous coatings for extracortical xation, J. Bone Joint Surg., 70A, 160, 1988.
98. Chin, H.C., Frassica, F.J., Markel, M.D., et al., The effects of therapeutic doses of irradiation on
experimental bone graft incorporation over a porous-coated segmental defect endoprosthesis, Clin.
Orthop., 289, 254, 1993.
0266/ ch29/frame Page 475 Friday, October 22, 1999 10:14 AM
476 Mechanical Testing of Bone and the BoneImplant Interface
99. Roy, R.G., Markel, M.D., Lipowitz, A.J., et al., Effect of homologous brin adhesive on callus
formation and extracortical bone bridging around a porous-coated segmental endoprosthesis in dogs,
Am. J. Vet. Med., 54, 1188, 1993.
100. Cook, S.D., Baffes, G.C., Palafox, A.J., et al., Torsional stability of HA-coated and grit-blasted titanium
dental implants, J. Oral Implantol., 18, 354, 1992.
101. Gotfredsen, K., Wennerberg, A., Johansson, C., et al., Anchorage of TiO
2
-blasted, HA-coated, and
machined implants: an experimental study with rabbits, J. Biomed. Mater. Res., 29, 1223, 1995.
102. Peb, P., Barbot, R., Trinidad, J., et al., Countertorque testing and histomorphometric analysis of
various implant surfaces in canines: a pilot study, Implant Dent., 6, 259, 1997.
0266/ ch29/frame Page 476 Friday, October 22, 1999 10:14 AM

477

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

30

The Validity
of a Single Pushout Test

Wouter J.A. Dhert and John A. Jansen

CONTENTS

I. Introduction ..........................................................................................................................477
II. Practical Aspects of Pushout and Tensile Testing...............................................................479
III. Factors Affecting Pushout and Tensile Testing ...................................................................482
A. Experimental Setup ........................................................................................................482
B. Implant Properties ..........................................................................................................482
C. Biological Parameters.....................................................................................................483
IV. Failure Mode Analysis .........................................................................................................485
V. Practical Recommendations and Conclusions.....................................................................485
References ......................................................................................................................................487

I. INTRODUCTION

The long-term clinical success of hard-tissue implants is based on the presence and maintenance
of a proper bone response.

1

Basically, two types of bone response to implant materials can be
distinguished. The rst type involves the formation of a collagenous connective tissue capsule
around the implant.

2

The second type of bone response is characterized by a direct boneimplant
contact, without an intervening connective tissue layer, or even by direct bone bonding.

3-5

Generally, this second type of bony xation is considered to be the preferable situation for
implant xation.
Various histological and histomorphometrical techniques can be used to analyze and quantify
the bone response to implanted biomaterials. However, a thorough histological description cannot
be used as the only criterion for implantbone biocompatibility. For example, histological
procedures do not provide information about the boneimplant interfacial strength and perhaps
the development of chemical bonding between implant and surrounding bone. Since a major
purpose in implant-related research in orthopaedic surgery is to achieve a proper xation of the
implant in the surrounding bone (as a result of bone bonding or bone ingrowth), a mechanical
test that actually measures this implant xation is desired. Various test methods are now available
for the mechanical evaluation of retrieved implants. Two tests that have been developed to provide
the investigator with absolute numbers for the mechanical xation of an implant in the surround-
ing bone are the pushout test and the tensile test. Both tests measure the force that is necessary
to move an implant in the surrounding tissue (usually cortical or trabecular bone), and allow the
calculation of a so-called pushout or tensile strength. Pushout and tensile tests are simple tests
which can easily be performed with limited biomechanical knowledge and low-cost instruments.
On the other hand, the ease of use of these tests in experiments on implant fixation does not
mean that tests performed by various investigators in various experiments are comparable. When
pushout data from various studies are compared (see Table 30.1), a large scatter in data is seen,

0266/ ch30/frame Page 477 Friday, October 22, 1999 10:19 AM

478

Mechanical Testing of Bone and the BoneImplant Interface

even for identical implant materials. The problem of comparability and reproducibility of a
pushout test has been subject to various discussions in the past,

6,7

and in 1989, Black stated that
when a pushout test is used, the test conditions need to be described thoroughly.

8

It is now obvious that pushout and tensile tests are valuable instruments for the study of implant
xation in bone. However, past experience indicates that comparison of results from different
investigators is difcult, and that variations in results can most likely not only be attributed to
differences in bone bonding ability between implants but also to variations in methods of testing.
The present chapter discusses various biomechanical and biological aspects that determine the
validity of pushout and tensile testing.

TABLE 30.1
Example Overview of Various Studies on Press-Fit-Inserted Implants in Which Pushout
Strengths Were Measured

Substrate
Surface
Treatment
Surface
Texture Follow-up Animal
Bone
Type
Shear
Strength Ref.

Polyurethane HA coated

4
12
Rabbit Cortical 5.9
8.2
9
c.p. Ti Noncoated
HA coated
Porous coated 3
6
12
3
6
12
Dog
Dog
Cortical
Cortical
7.7
12.6
18.1
7.5
14.2
17.9
10
c.p. Ti Noncoated Bead/grit blasted 3
6
10
32
3
6
10
32
Dog Cortical 1.5
1.5
1.0
1.2
4.0
7.0
7.3
6.0
11
Ti6Al4V Noncoated
HA coated
Bead/grit blasted 12
26
12
26
Goat
Goat
Cortical
Cortical
7.4/6.7
9.6/7.5
13.3/11.1
17.3/13.3
12
Ti6Al4V Noncoated
HA coated
Bead/grit blasted 6
12
6
12
Dog
Dog
Cortical
Cortical
49.1
54.8
13
Ti6Al4V Noncoated
HA coated
Porous coated 4
4
Dog
Dog
Trabecular
Trabecular
7.0
6.7
14
c.p.Ti Noncoated
HA coated
Macrotexture 3
5
10
3
5
10
Dog
Dog
Cortical
Cortical
4.4
4.8
10.5
6.0
9.6
14.1
15
c.p. Ti Noncoated
Ti coated
Bead/grit blasted 12
12
Goat
Goat
Cortical
Cortical
2.9
6.9
16
The variation in pushout data can most likely not only be explained by differences between implant materials and/or surfaces.

0266/ ch30/frame Page 478 Friday, October 22, 1999 10:19 AM

The Validity of a Single Pushout Test

479

II. PRACTICAL ASPECTS OF PUSHOUT AND TENSILE TESTING

The design of a pushout test is based upon a cylindrical implant that is placed in cortical or trabecular
bone (Figure 30.1). The implant is prepared of the specic material to be tested, and has a certain
diameter (usually 2 to 5 mm), length (usually 3 to 7 mm), and surface characteristics (coating and
surface topography).
For an extensive description of the practical aspects of the pushout test, see Chapter 29, Implant
Pushout and Pullout Tests. Briey, the following procedure is used for pushout testing. After
surgical placement of the implant in the bone, the wound is closed, with or without suturing the
periosteum over the implant top. The implant is then allowed to heal, and after a certain follow-
up period, the animal is sacriced and the bone with the implant(s)

in situ

is excised. By using a
diamond saw, specimens are removed from the retrieved bones, taking care not to damage the
interface and ensuring adequate bone remains next to the implant. Some authors use only part of
an implantbone specimen by making a cross section,

14,17

but in most studies the entire implant is
used in a pushout test. Figure 30.2 shows a schematic drawing of a pushout setup. To prevent
disturbance of the test by periosteal or endosteal bone that has grown over the implant top/bottom
surface, this tissue should be removed by polishing or grinding. This will also allow for an accurate
alignment of the specimen (loaded surface perpendicular to boneimplant interface). The specimen
is positioned on a support jig above a hole which is larger than the implant diameter, and pushed
out of the surrounding bone by a force that is applied from the opposite side of the implant. The
pushout load is applied by a test bench with a constant displacement speed of 0.5 to 1.0 mm/min.
The resulting forces are registered by a load cell, and a loaddisplacement curve is recorded
(Figure 30.3). The peak force on the loaddisplacement curve is considered as the pushout force
and is usually the force at which the implant starts to move in the surrounding bone. A pushout
test measures the force that is necessary to move the implant in the surrounding bone (ultimate
load), and since from a theoretical standpoint this is a shear force, subsequently an interface shear
strength is calculated:

=

F/A

(30.1)
where

is the interface shear strength (Pa),

F

is the ultimate load (N), and

A

is the implant surface (mm

2

).

FIGURE 30.1

Photograph of surgical procedure where various cylindrical implants were placed in the
diaphyseal femoral cortex of a goat. After a dened healing period, the mechanical xation of the implants
in the cortical bone is measured in a pushout test.

0266/ ch30/frame Page 479 Friday, October 22, 1999 10:19 AM

480

Mechanical Testing of Bone and the BoneImplant Interface

It is clear from the formula that besides the ultimate load

F

that was measured, the implant surface

A

will determine the pushout strength. Although most scientic reports only mention the pushout
strength, it is important to explain how the surface area was measured. The surface area

A

should
represent the actual interface area, i.e., only that part of the implant that was xed in the bone. From
a macroscopic point of view,

A

is the part of the implant that is surrounded by bone, but from a
microscopic point of view,

A

can also be considered as that part of the implant that is actually in
contact with the implant surface. The latter is the percentage of boneimplant contact, and can only
be determined by histomorphometry. When transcortical implants are tested, upgrowth of woven or
trabecular bone at the endosteal side of the implant has most likely occurred, which will have an
unknown inuence on the ultimate load (Figure 30.4). It will be technically difcult to remove this
endosteal bone without damaging the interface in the cortical zone. Therefore, although this can be
subject to dispute, the authors recommend neglecting this endosteal bone and calculate

A

as follows
(30.2)

FIGURE 30.2

Schematical drawing of the pushout test. F = force applied on implant; C = cortical bone;
J = support jig; y = position along interface; t = cortical thickness; d = implant diameter; x = clearance of hole
in support jig.

FIGURE 30.3

Example of a loaddisplacement curve, as registered during a pushout test.
A dh =

0266/ ch30/frame Page 480 Friday, October 22, 1999 10:19 AM

The Validity of a Single Pushout Test

481

where

d

is the implant diameter (mm) and is the average thickness of the cortex (mm), taken
from several measurements at distance of the implant, so upgrown bone at the endosteal side is
excluded (Figure 30.5).
To measure (predominantly) shear strengths with pushout tests, it was suggested by
Steinemann

18

to use a model for tensile testing. A hole is reamed in, but not through, the cortical
bone of a test animal, and one end of a cylinder-shaped implant is placed in the hole. The implant
can be kept in position by either a spring with screws next to the implant, or by a cerclage wire.
A silastic spacer is placed around the implant to prevent upgrowth of bone. Hypothetically, this
test has the advantage of measuring only tensile strength, being more reproducible, and allowing
for a better alignment during the test. Nevertheless, calculations show that for cylindrical implants
the distribution of the tensile stresses along the interface is still not uniform, unless the implant is
shaped according to very specic standards.

19,20

Further, previous

in vivo

results have shown that
data from this tensile test model do differ with respect to data from a pushout model. It takes, for

FIGURE 30.4

Graphical representation of a cross section of an implant in cortical bone, after a healing
period. At the periosteal side, bone has grown over the implant, and at the endosteal side, upgrowth of woven
bone along the implant surface is present. Prior to pushout testing, the periosteal bone should be removed by
polishing or grinding.

FIGURE 30.5

Schematic drawing of a piece of semicircular cortical bone prior to pushout testing. Arrows
indicate the four positions at which the cortical thickness can be measured to calculate the pushout strength.
h

0266/ ch30/frame Page 481 Friday, October 22, 1999 10:19 AM

482

Mechanical Testing of Bone and the BoneImplant Interface

instance, a considerable follow-up period (>6 weeks in goats) to get a measurable tensile
strength.

19,20

Nevertheless, the authors think that the model is very promising for tensile testing of
implant xation.

III. FACTORS AFFECTING PUSHOUT AND TENSILE TESTING

Despite the use of standardized and appropriate pushout and pullout models, the investigator needs
to be aware that there are a lot of experimental variables, which can affect the nally measured
boneimplant xation strength. These inuences can be test, implant, or biologically dependent.

A. E

XPERIMENTAL

S

ETUP

As explained by the formulae in the previous paragraph, pushout strength is calculated as shear
force per unit area. From a biomechanical point of view, the use of this formula is justied when
actual forces are measured that represent the boneimplant interface, and when these forces are
shear forces. By dividing the pushout force by the contact area, local stresses along the interface
are averaged and no longer recognized. Only for situations where there is an equal distribution
of stresses along the interface does the calculated pushout strength also represent local stresses.
From a theoretical standpoint, local peak stresses along the interface could initiate interface
debonding/local implant loosening, after which the peak stresses move to another point that is still
bonded/xed, and the same process of following local areas of peak stresses is repeated until the
implant is entirely loosened. Since these peak stresses are not recognized in a single pushout test,
a pushout test setup should pursue an equal distribution of stresses along the interface. To gain
insight in these interface stress distributions, several studies calculated interface stress distributions
using nite-element models. In a study by Dhert et al.,

21

it was demonstrated that the implant
diameter and the cortical thickness did not have much inuence, but that the Youngs modulus of
the implant (see next paragraph) and the clearance of the hole in the support block (Figure 30.6)
were of major impact for the interface stress distribution. This clearance is nothing more than the
oversize of the hole in the support block as compared with the implant diameter. Several investi-
gators mistakenly believe that the better the t of the implant in the hole, the more accurate the
test will be. Shirazi-Adl

22,23

reported that to get a uniform stress distribution along the interface,
the bone surrounding the implant should be supported instead of using a support block under the
implant. This will, however, introduce practical difculties in the execution of the test.

B. I

MPLANT

P

ROPERTIES

Implant xation can be considered as the (im)possibility of motion of an implant in the surrounding
bone, and is basically a result of three parameters: friction, mechanical interlock, and chemical
bonding. When two surfaces that are completely smooth and at are tested in shear, the main
parameter that will prevent movement is friction. However, when these surfaces are not smooth or
have a certain macrotexture, mechanical interlock will be another factor that determines shear
movement. In view of this, distinction has to be made between macro- and microtexture. Macro-
texture, for example, is provided by screw threads and porous bead-shaped coatings. Microtexture
is related to surface dimensions in the order of magnitude of 1 to 40

m, like that created by grit-
blasting procedures. Although macrotextured implants have been subjected to pushout testing,

15

it
must be emphasized that pushout tests are not suited to measure the mechanical bone interfacial
strength of such implants. Bone ingrowth in the recesses of the macrotexture will prevent a reliable
interfacial disruption at the boneimplant interface.
Besides the surface condition of the implant material, specic bulk mechanical properties can
affect the pushout or pullout test. As regards material-related mechanical aspects, various classes
of implant materials have to be distinguished, e.g., metals, ceramics, and polymers. The modulus

0266/ ch30/frame Page 482 Friday, October 22, 1999 10:19 AM

The Validity of a Single Pushout Test

483

of elasticity of metallic implant materials is in general above that of bone, while the modulus of
elasticity of polymers is usually below that of bone. This will result in different patterns of load
transfer through the implant, and in differences in implant deformation. Clearly, this can hamper
pushout testing. Finite-element studies demonstrated that interface stress distributions are inuenced
by the Youngs modulus of the implants.

21

At lower Youngs moduli (e.g., for polymer implants),
interface stress distributions become less uniform, and the higher stresses are located at the side
from which the load is applied to the implant (Figure 30.7). Ceramics such as hydroxyapatite can
give a favorable bone response and these materials are used frequently in mechanical interface
tests. However, ceramics have intrinsic mechanical shortcomings,

24

such as a low tensile and
bending strength. This low tensile strength limits the use of a tensile test for ceramic materials.
Pushout tests are also limited as implant failure during the test can hamper the procedure.

C. B

IOLOGICAL

P

ARAMETERS

In addition to mechanical testing, bone contact measurements are used to evaluate how bone tissue
responds to an implant surface. They are also supposed to give an indication of implant xation
since it seems logical that implants with a high percentage of bone contact will be rmly xed into
the bone. To investigate a possible relationship with the most common test for implant xation,
the pushout test, the authors plotted corresponding pushout and bone contact data from former
studies

12,25

against each other, and this is presented in Figure 30.8. The correlation coefcient for
these data was 0.38. Thus the expected high correlation between bone contact and pushout strength
was not present in these data. This is very remarkable. Two explanations for this discrepancy can
be given: rst, the unequal distribution of stresses along the interface during a pushout test

21

already
indicates that there cannot be a linear relationship between bone contact and pushout strength.
Second, the biomechanical characteristics of the bone tissue in contact with the implant might also
vary (e.g., in mineralization, collagen content, etc.), which is conrmed by Corten et al.

26

Bone
mineral density measurements using dual X-ray absorptiometry (DEXA) around implants inserted
into the lateral and medial femoral condyles of goats revealed that the bone of the medial condyle
showed a higher density than the lateral condyle. Bone contact measurements showed that the
implants in the medial condyle demonstrated more bone contact compared with similar implants
in the lateral condyle. Evidently, this is due to the X-shaped anatomy of the limbs of goats. Moreover,
it can be supposed that this higher density will affect the biomechanical evaluation of the bone

FIGURE 30.6

Interface stress distributions, as calculated with a finite-element analysis, for implant-support
jig distances of 0.1 and 1.0 mm (Ti6Al4V implant; implant diameter of 5.0 mm; cortical thickness of 3.0 mm).

Y

-axis represents position along the interface from 0 mm (endosteal side) to 3 mm (periosteal side). The graph
demonstrates that a closer t of the support jig results in high stress peaks at the periosteal side of the implant.

0266/ ch30/frame Page 483 Friday, October 22, 1999 10:19 AM

484

Mechanical Testing of Bone and the BoneImplant Interface

response. These ndings also prove that care has to be taken of an inuence of the biomechanical
properties and loading forces of the bone at the implantation site.
A second relevant biological consideration is the freshness of the specimens, as used for
pushout or pullout studies. Considering the transportation and maintenance of the retrieved samples,
an uniform method of preservation has to be recommended. Concerning this storage, the best
method is to cool the specimens on ice (0

C) until measurement. However, the samples should
never be frozen,

27

nor xed in formalin

21

prior to testing. With regard to freshness, fresh is dened
as tested within 4 h after explantation.

6

FIGURE 30.7

Interface stress distributions, as calculated with a finite-element analysis, for implants with
a Youngs modulus comparable to Ti6Al4V (

E

= 100 GPa) and for implants with a low (

E

= 2 GPa) Youngs
modulus (clearance of hole in support jig of 0.7 mm; implant diameter of 5.0 mm; cortical thickness of
3.0 mm).

Y

-axis represents position along the interface from 0 mm (endosteal side) to 3 mm (periosteal side).
The stress distribution is less uniform for implants with a low Youngs modulus, resulting in high local interface
stresses at the side of the implant from which the load is applied. This could result in a stepwise implant
loosening during the test at relative low loading.

FIGURE 30.8

Scatter plot of pushout strengths as a function of corresponding bone contact measurements of
the same implants. Data represent a total of 96 transcortical implants in the femur and humerus of goats, 12 and
25 weeks after implantation. Although there is clearly a trend visible, the actual correlation is only moderate.

0266/ ch30/frame Page 484 Friday, October 22, 1999 10:19 AM

The Validity of a Single Pushout Test

485

IV. FAILURE MODE ANALYSIS

At present, no standard pushout and tensile test methods are available to evaluate the mechanical
behavior of the implanttissue interface. Therefore, it is essential that in publications as much
detail as possible is provided about the test conditions.

8,21

With regard to this, besides mechan-
ical testing, data also have to be provided about the mode of failure. Otherwise, a correct
interpretation of the measured shear or tensile forces is impossible. The actual mode of failure
evaluation can be most simply done by scanning electron microscopy (SEM). Following the
mechanical testing, the samples are xed, dehydrated, and embedded in methylmethacrylate.
After polymerization, the specimens are hemisectioned longitudinally through the implant.
After polishing the specimens, the failure site resulting from the mechanical testing can be
examined with an SEM.
The relevance of a failure mode analysis can be demonstrated with two recent examples.
First, Vercaigne et al.

16

inserted three types of titanium plasma-spray-coated implants with a
roughness value (Ra) of 16, 21, and 38

m, respectively, in the tibial cortical bone of goats.
Mechanical pushout evaluation was performed 3 months after implantation. As expected, the
pushout values showed an increase with increasing surface roughness. On the other hand,
failure mode analysis revealed that the fracture line around the roughest implants was located
in the bone and not at the boneimplant interface. This phenomenon is explained by the
previously mentioned parameter of mechanical interlocking. In the Vercaigne study, the rough-
est surface provided the strongest mechanical interlock, and thus resulted in the highest pushout
value. As shown in Figure 30.9, the pushout test did not measure the actual strength of the
boneimplant interface, but the mechanical properties of the bone were the major determining
factors. So the measured failure load can be considered only as some kind of general measure
of the amount of xation of the implant in the bone. The second example is dealing with
plasma-sprayed hydroxyapatite (HA) coatings. These coatings are applied on titanium implants
to improve the poor fatigue properties of HA ceramic with maintenance of the biological
advantage.

28

Installation of plasma-sprayed HA-coated implants in the tibial cortex of goats
resulted, after 3 months of implantation, in mean pushout strengths of about 15 MPa.

12

However,
fracture analysis demonstrated that the fracture site was in some situations located at the
coatingtitanium interface and not at the coatingbone interface (Figure 30.10). Consequently,
the measured data not only represent the boneimplant interface, but also the strength of the
substratecoating xation.
The two above-mentioned examples clearly demonstrate that under comparable test conditions
two completely different mechanical structures were measured (i.e., the mechanical properties of
bone itself, and the strength of the coatingtitanium interface), and that none of these structures
measured the actual boneimplant interface. Of course, this raises doubts about the reliability of a
pushout test to measure the strength of boneimplant interface, which concerns the clinical rele-
vance. Therefore, a pushout test should be used to obtain data on implant xation in a more general
context. Subsequent failure analysis should give the user additional information on the actual mode
of xation of the implant. This will put the interpretation of single pushout data and comparison
with other studies in perspective.

V. PRACTICAL RECOMMENDATIONS AND CONCLUSIONS

Pushout and tensile testing can be effective in demonstrating differences in the interfacial bonding
strength for various types of bone implants. Unfortunately, the simplicity of these methods is
associated with an easy inclusion of errors, which affect the accuracy of the nally obtained
data. Therefore, it has to be recommended that researchers use a strictly designed protocol for
their mechanical tests and mention as much detail as possible.

8,21

It has to be recognized that
most pushout tests do not measure the shear strength of the boneimplant interface, but report

0266/ ch30/frame Page 485 Friday, October 22, 1999 10:19 AM

486

Mechanical Testing of Bone and the BoneImplant Interface

a value that, although called pushout strength, is a combination of various xation principles
(friction, mechanical interlock, and chemical bonding). From a clinical perspective, such pushout
data still are valuable, since these will provide an indication of how strongly an implant is
anchored in the implant bed. However, since so many parameters are of signicant inuence on
the pushout test, extrapolation to other studies is complex and can result in misinterpretation,
unless experiments were performed under identical conditions (which will hardly ever happen
in practice). Therefore, the authors strongly recommend that investigators design pushout exper-
iments such that comparisons within a single study are possible. A thorough and detailed
description of the pushout test should be provided. Still, investigators should recognize that even
within one experiment, implant properties, biological parameters, and test conditions may vary
and prevent a proper comparison of pushout data. In view of this, to obtain the best possible
impression about implant xation, the authors recommend always combining a mechanical test
with histomorphometrical measurements.

FIGURE 30.9

Example of failure mode analysis after a pushout test for two characteristic implant surfaces:
(A) nonmacrotextured surface showing interface failure (arrows) at the bonecoating interface; (B) macro-
textured surface showing failure of the bone (arrows) next to the interface.

0266/ ch30/frame Page 486 Friday, October 22, 1999 10:19 AM

The Validity of a Single Pushout Test

487

REFERENCES

1. Zoldos, J. and Kent, J.N., Healing of endosseous implants, in

Endosseous Implants for Maxillofacial
Reconstruction

, Block, M.S. and Kent, J.N., Eds., W.B. Saunders, Philadelphia, 1995, 40.
2. Weiss, C.M., Tissue integration of dental endosseous implants: description and comparative analysis
of the bro-osseous integration and osseous integration systems,

J. Oral Implantol.,

12, 169, 1986.
3. Linder, L., Osseointegration of metallic implants. I. Light microscopy in the rabbit,

Acta Orthop.
Scand.,

60, 129, 1989.
4. Brnemark, P.I., Introduction to osseointegration, in

Tissue Integrated Prostheses, Osseointegration
in Clinical Dentistry

, Brnemark, P.I., Zarb, G.A., and T.A., Eds., Quintessence, Chicago, 1985, 11.
5. Cook, S.D., Thomas, K.A., Dalton, J.E., et al., Hydroxylapatite coating of porous implants improves
bone ingrowth and interface attachment strength,

J. Biomed. Mater. Res.,

26, 989, 1992.
6. Groot de, K., Letter to the editor

, J. Biomed. Mater. Res.,

23, 1367, 1989.
7. Thomas, K. A., Letter to the editor,

J. Biomed. Mater. Res.,

23, 1367, 1989.
8. Black, J., Pushout tests

, J. Biomed. Mater. Res.,

23, 1243, 1989.
9. Boone, P.S., Zimmerman, M.C., Gutteling, E., et al., Bone attachment to hydroxyapatite coated
polymers,

J. Biomed. Mater. Res.

, 23, 183, 1989.
10. Cook, S.D., Thomas, K.A., Kay, J.F., and Jarcho, M., Hydroxyapatite-coated porous titanium for use
as an orthopedic biologic attachment system,

Clin. Orthop.

, 303, 1988.
11. Cook, S.D., Thomas, K.A., Kay, J.F., and Jarcho, M., Hydroxyapatite-coated titanium for orthopedic
implant applications,

Clin. Orthop.

, 225, 1988.
12. Dhert, W.J., Klein, C.P., Wolke, J.G., et al., A mechanical investigation of uorapatite, magnesium-
whitlockite, and hydroxylapatite plasma-sprayed coatings in goats,

J. Biomed.



Mater. Res.

, 25, 1183,
1991.
13. Geesink, R.G., de Groot, K., and Klein, C.P., Chemical implant xation using hydroxyl-apatite
coatings. The development of a human total hip prosthesis for chemical xation to bone using
hydroxyl-apatite coatings on titanium substrates,

Clin. Orthop.

, 147, 1987.
14. Soballe, K., Hansen, E.S., Brockstedt-Rasmussen, H., et al., Hydroxyapatite coating enhances xation
of porous coated implants. A comparison in dogs between press t and noninterference t,

Acta
Orthop. Scand.

, 61, 299, 1990.
15. Thomas, K.A., Kay, J.F., Cook, S.D., and Jarcho, M., The effect of surface macrotexture and hydrox-
ylapatite coating on the mechanical strengths and histologic proles of titanium implant materials,

J. Biomed. Mater. Res

., 21, 1395, 1987.
16. Vercaigne, S., Wolke, J.G., Naert, I., and Jansen, J.A., Histomorphometrical and mechanical evaluation
of titanium plasma-spray-coated implants placed in the cortical bone of goats,

J.



Biomed. Mater. Res.

,
41, 41, 1998.

FIGURE 30.10

Example of failure mode analysis after a pushout test of a plasma-sprayed calcium
phosphatecoated Ti6Al4V implant. The fracture is visible at the coatingtitanium interface. c = coating; b =
bone; s = soft tissue; ti = titanium alloy implant; f = fracture as a result of the pushout test.

0266/ ch30/frame Page 487 Friday, October 22, 1999 10:19 AM

488

Mechanical Testing of Bone and the BoneImplant Interface

17. Soballe, K., Hansen, E.S., Brockstedt-Rasmussen, H., et al., Bone graft incorporation around titanium-
alloy- and hydroxyapatite-coated implants in dogs,

Clin. Orthop.

, 282, 1992.
18. Steinemann, S.G., Eulenberger, J., Maeusli, P.A., and Schroeder, A., Adhesion of bone to titanium,
in

Biological and Biomechanical Performance of Biomaterials. Advances in



Biomaterials

, Christel,
P., Meunier, A., and Lee, A.J., Eds., Elsevier Science Publishers, Amsterdam, 1986,
19. Brunski, J.B., Elias, J., Edwards, J.T., et al., Tensile bonding at bonebiomaterial interfaces. An

in
vivo

test and related nite stress analysis,

Trans. Soc. Biomater.

, 17, 15, 1991.
20. Kangasniemi, I., Development of Calcium Phosphate Ceramic Containing Bioactive Glass Compos-
ites, Thesis Biomaterials, University of Leiden, the Netherlands, 1993.
21. Dhert, W.J., Verheyen, C.C., Braak, L.H., et al., A nite element analysis of the pushout test: inuence
of test conditions,

J. Biomed. Mater. Res

., 26, 119, 1992.
22. Shirazi-Adl, A., Finite element stress analysis of a pushout test. Part 1: Fixed interface using stress
compatible elements,

J. Biomech. Eng.

, 114, 111, 1992.
23. Shirazi-Adl, A. and Forcione, A., Finite element stress analysis of a pushout test. Part II: Free interface
with nonlinear friction properties,

J. Biomech. Eng.

, 114, 155, 1992.
24. Aoki, H.,

Science and Medical Applications of Hydroxyapatite

, Takajima Press System Center Co.,
Tokyo, 1991,
25. Dhert, W.J., Klein, C.P., Jansen, J.A., et al., A histological and histomorphometrical investigation of
uorapatite, magnesiumwhitlockite, and hydroxylapatite plasma-sprayed coatings in goats,

J. Biomed.
Mater. Res.

, 27, 127, 1993.
26. Corten, F.G., Caulier, H., van der Waerden, J.P., et al., Assessment of bone surrounding implants in
goats:

ex vivo

measurements by dual X-ray absorptiometry,

Biomaterials

, 18, 495, 1997.
27. Thomas, K.A., Cook, S.D., Dalton, J.E., and Baffes, G.C., Effects of storage upon interface attachment
characteristics: testing of fresh vs. frozen samples,

Trans. Soc. Biomater.

, 16, 251, 1990.
28. Groot de, K., Geesink, R., Klein, C.P., and Serekian, P., Plasma sprayed coatings of hydroxylapatite,

J. Biomed. Mater. Res.

, 21, 1375, 1987.

0266/ ch30/frame Page 488 Friday, October 22, 1999 10:19 AM

489

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

31

Tensile Testing
of BoneImplant Interface

Takashi Nakamura and Shigeru Nishiguchi

CONTENTS

I. Introduction ..........................................................................................................................489
II. Tensile Test Method (Nakamuras Method) ........................................................................489
A. Implantation....................................................................................................................489
B. Sample Preparation ........................................................................................................490
C. Measurement of Tensile Failure Load ...........................................................................491
D. Advantages and Disadvantages......................................................................................491
E. Previous Results Using the Detachment Test ................................................................493
F. A Second Type of Tensile Test ......................................................................................493
III. Other Tensile Tests...............................................................................................................494
IV. Conclusion............................................................................................................................496
References ......................................................................................................................................496

I. INTRODUCTION

Tests of interfacial bonding strength have been pursued using a variety of test methods, implant
geometries, and surface congurations. One type of test is a torque test, frequently involving screw-
shaped implants in the rabbit tibia or femur.

1

Another type of interfacial test is a pushout or shear
test. In the typical form of this test, cylindrical implants are placed through the cortex of a long
bone, such as the femur of a dog, and then pushed out following animal sacrice.

2

Besides torque and pushout tests, another method for investigating interfacial strength is the
tensile test, in which loads are applied in a direction normal to the tissuebiomaterial interface. It
is considered that in the tensile test, the strength of a direct biochemical bonding between the bone
and the biomaterial can be measured, thereby eliminating the resistive force due to surface roughness
of the biomaterial.

3

However, among various tensile testing methods, there is no standard method
in widespread use at the present time.
This chapter describes in detail the tensile testing method the authors have developed,

3

dis-
cussing its advantages and one disadvantage. The methods are compared to the tensile test methods
used by other research groups.

II. TENSILE TEST METHOD (NAKAMURAS METHOD)
A. I

MPLANTATION

The biomaterial to be tested is shaped into a rectangular plate, of size 15

10

2 mm. A rabbit
is anesthetized by intravenous injection of sodium pentobarbital (50 mg/kg). The operation is
performed under standard aseptic conditions with local administration of 0.25% xylocaine. For

0266/ ch31/frame Page 489 Friday, October 22, 1999 10:28 AM

490

Mechanical Testing of Bone and the BoneImplant Interface

implantation of the implant into the proximal tibial condyle, a 3-cm skin incision is made on the
anteromedial aspect of the tibial metaphysis of the rabbit, and the skin, muscle, and periosteum
are retracted. By using a dental burr and saline coolant, a hole (15

2 mm) is made parallel to the
longitudinal axis of the tibial metaphysis with the proximal end located 5 mm distal to the joint.
The medial cortex is drilled as slowly as possible so that the heat generated does not damage the
remaining bone. At this stage, care is taken to ensure that the hole for the implant is kept parallel
to the frontal plane of the tibia. After making the hole, a test insertion of the implant into the hole
is attempted. If the implant cannot be inserted, the hole is enlarged little by little.
When the preparation of the hole in the medial cortex is complete, the hole in the tibial
metaphysis for implantation is extended from the medial cortex to the lateral cortex. After the
medial cortex has been penetrated, the dental burr immediately reaches the inner side of the lateral
cortex because of the scant amount of cancellous bone in the proximal tibial metaphysis. Since the
lateral cortex cannot be visualized directly due to constant bleeding from the medullary cavity of
the tibia, careful cutting is required at this stage. Once again, care is taken to ensure that the hole
in the lateral cortex is kept parallel to the frontal plane of the tibia. The lateral hole is enlarged
step by step and it is conrmed whether the implant can be inserted into the tibia through both
medial and lateral cortices. The hole is irrigated with saline before the plate is implanted in the
frontal plane, perforating the tibia and protruding from both tibial cortices (Figure 31.l). The muscle
and skin are closed in layers.

B. S

AMPLE

P

REPARATION

The rabbit is killed at a given postimplantation time. A segment of the tibia containing the implant
is excised (Figure 31.2a) and the bone is sectioned transversely with a disk cutter by parallel cuts
2 mm above and below each end of the plate (Figure 31.2b). The bony tissues located on the lateral
sides of the plate are removed with a dental burr. Cuts of the same width as those of the implants
are made with a dental burr, extending from both ends of the bone tissue to the ends of the implant
(Figure 31.2c). After such sectioning, the bony segments on either side of the implant are not
directly connected but are now joined only through the intervening implant. Extreme care is taken
to remove the bone completely from the sides of the implant and to keep the bones and implant
moist throughout the dissection and subsequent testing.

FIGURE 31.1

Implantation method: implant is inserted into rabbit tibia from medial side through both
medial and lateral cortices parallel to the frontal plane of the tibia. The plate protrudes from both cortices.

0266/ ch31/frame Page 490 Friday, October 22, 1999 10:28 AM

Tensile Testing of BoneImplant Interface

491

C. M

EASUREMENT



OF

T

ENSILE

F

AILURE

L

OAD

Each segment is held by a hook that is connected to an Instron-type testing machine (either
Autograph Model S-500, Shimadzu Seisakujo Ltd., Kyoto, Japan, or Model 1011, Aikoh Engineer-
ing Co. Ltd., Nagoya, Japan; Figures 31.2d and 31.3). The plate is placed horizontally and pulled
in opposite directions at a crosshead speed of 3.5 cm/min. The load at which an implant is detached
from the bone or at which the bone is broken is designated as failure load. In the authors study,
the pressure or tensile stress cannot be calculated because the true contact area cannot be measured
adequately by X-ray or light microscopic techniques. Thus, the results of this method are shown
by a unit of force, such as kgf or N, not by a unit of stress, such as MPa.

D. A

DVANTAGES



AND

D

ISADVANTAGES

This method possesses four advantages. First, the detachment test is simple and easy to perform
and quite reproducible. On account of the relatively low invasive nature of the surgical procedure,
this test method can be performed on both tibias. Second, since the shape of the biomaterial to be
tested is rectangular (15

10

2 mm), it is relatively easy to prepare the material for implantation

FIGURE 31.2

Schematic drawing of preparation of tensile specimen. (a) Retrieve the proximal tibia con-
taining the implant. (b) Cut the tibia at 2 mm proximal and distal at the end of the plate. (c) Remove bone
connecting anterior and posterior cortices. (d) Apply tensile load through hooks holding anterior and posterior
cortices until the bone is detached from the plate.

0266/ ch31/frame Page 491 Friday, October 22, 1999 10:28 AM

492

Mechanical Testing of Bone and the BoneImplant Interface

and testing. Third, the implant has two boneimplant interfaces, only one of which is broken after
a single tensile test. Thus, on the remaining boneimplant interface, one can either perform a second
tensile test or carry out a histological examination. Finally, this tensile test has been performed since
1985, mainly at the authors institute, in order to evaluate the bone-bonding abilities of many different
kinds of biomaterials. The accumulated results are quite useful for comparing the bone-bonding

FIGURE 31.3

Photographs of detaching test. (A) Overall configuration. (B) Close-up of the hooks and the bone.

0266/ ch31/frame Page 492 Friday, October 22, 1999 10:28 AM

Tensile Testing of BoneImplant Interface

493

strengths of different biomaterials. One can better estimate the bone-bonding ability of a new
material by comparing its performance with the results obtained from preexisting biomaterials.
On the other hand, this tensile test method has one disadvantage; the force intensity or tensile
stress cannot be calculated because the true contact area cannot be measured easily.

E. P

REVIOUS

R

ESULTS

U

SING



THE

D

ETACHMENT

T

EST

As mentioned previously, since 1985, when Nakamura et al. developed this testing method, several
different researchers at Kyoto University have carried out this test in order to evaluate the bone-
bonding strength of different biomaterials, such as bioactive ceramics,

3,4,5

bioinert ceramics,

3

tita-
nium,

6,7

bioactive cements,

8

and titanium coated with bioactive materials.

7

These results are sum-
marized in Table 31.1. Since the tensile tests in these studies are based completely upon the
detachment test method of Nakamura, the results can be compared directly with each other.

F. A S

ECOND

T

YPE



OF

T

ENSILE

T

EST

Yoshii et al.

9

examined the bonding strength of A-W glass ceramic to the surface of bone cortex.
They used semicolumnar-shaped blocks (4

10

8 mm) with holes 2 mm in diameter at their
centers (Figure 31.4), made of A-W glass ceramic and xed with screws on the surface of the tibias
of rabbits. The tensile load required to detach the implant from the surface of the bone cortex was
measured at 2, 4, 8, and 25 weeks after the implantation.
Mature male rabbits each weighing 3.0 to 3.5 kg were anesthetized with sodium pentobarbital
(50 mg/kg). The operations were performed under aseptic conditions. A skin incision (2 cm long)
was made on the anteromedial aspect of the proximal tibial metaphysis of the rabbit. The cortex
of the tibia was exposed subperiosteally about 10

10 mm. A hole 1.8 mm in diameter was made
with a drill at the center of the exposed cortex, penetrating into the posterior cortex. The implant

TABLE 31.1
Bone-Bonding Strength of Biomaterials Using the Nakamuras Tensile Test Method
(mean value)

4 weeks,
kgf
8 weeks,
kgf
10 weeks,
N
16 weeks,
kgf 25 weeks

AW-glass ceramic (AW-GC)

3

7.43 8.15 kgf
Dense hydroxyapatite

3

6.28 7.90 kgf
45S5 Bioglass

3

2.75
Alumina

3

0.13

-Calcium phosphate

5

31.65 47.04 N

-Tricalcium phosphate (TCP)

5

72.81 71.34 N
Tetracalcium phosphate

5

43.22 62.03 N
Ti6Al4V

7

0.0014 0.029 0.334 2.852 kgf
Bioglass-coated Ti6Al4V

7

1.04 2.72
AW-coated Ti6Al4V

7

2.03 2.39
TCP-coated Ti6Al4V

7

3.91 4.23
Alkali- and heat-treated Ti

6

2.71 4.13
Pure titanium

6

0.02 0.33
Demineralized bone powder + AW-GC

4

7.21 9.00 11.91 kgf
PMMA cement

8

0.29 0.2 N
Bioactive bone cement
(bis-GMA) + (MgO-CaO-SiO

2

-P

2

O

5

-CaF

2

) glass powder

8

29.52 33.42 N
(bis-GMA) + (MgO-CaO-SiO

2

-P

2

O

5

-CaF

2

) glass ceramic powder

8

41.48 41.27 N
(bis-GMA) + (MgO free CaO-SiO

2

-P

2

O

5

-CaF

2

) glass powder

8

28.22 33.64 N

0266/ ch31/frame Page 493 Friday, October 22, 1999 10:28 AM

494

Mechanical Testing of Bone and the BoneImplant Interface

was placed on the at surface of the medial cortex and xed with a stainless steel screw 2 mm in
diameter (Figure 31.5). The screw was driven securely into the drilled hole through the hole in the
implant. The wound was cooled and cleaned with sterile saline solution during the operation, before
being closed in layers.
At 2, 4, 8, and 25 weeks after implantation, the tibia was excised with the implant and sectioned
transversely 3 mm proximal and distal to the implant. The bony tissue that had grown over the
margins of the implant was removed completely in order to prepare the implant so that it was in
contact with bone only at its base. The screw was removed gently.
The implant and the piece of bone were held by the use of hooks. The hooks were connected
to an Instron-type tensile load tester (Autograph S-500, Shimadzu, Kyoto, Japan). The implant was
pulled perpendicularly to the base of the implant contacting the bony cortex (Figure 31.6). The
crosshead speed was set at 3.5 cm/min. The load required to detach the implant from the bone, or
to break the bone, was measured. In this type of tensile test, the bonding area between bone and
implant is the same as the area of implant surface. Thus, the tensile strength can be calculated if
necessary, by dividing tensile force by implant surface area.

III. OTHER TENSILE TESTS

Lin et al.

10

carried out another type of tensile strength testing in order to evaluate bonding strength
between bone and a plasma-sprayed, hydroxyapatite (HA)-coated titanium and titanium alloy. Their
implant had a semicircular shape, 5.5 mm in diameter and 2.2 mm thick. Both sides of the implant

FIGURE 31.4

Implant. (a) Schema of the implant. (b) The base of the implant (approximately 70 mm

2

).

FIGURE 31.5

The method of fixation of the implant to the surface of the medial cortex of the tibia with a screw.

0266/ ch31/frame Page 494 Friday, October 22, 1999 10:28 AM

Tensile Testing of BoneImplant Interface

495

had a thin layer of plasma-sprayed HA coating. They used male Japanese white rabbits. About
2 cm below the knee, a 2.2-mm-long semicircular piece of the diaphyseal cortex was removed,
using a parallel dental disk cutter and a dental burr under saline coolant. The implant was placed
into this defect using light thumb pressure. After sacrice, the left tibia was removed and sectioned
into 3-cm-long segments with the implant in the middle. The extra bony tissue surrounding the
implant was carefully removed using a dental burr, so that only the distal and proximal sides of
the bone contacted the implant. A 1-mm hole was made in both of the bone segments, and ligatures
were attached and xed in the clamps of a Universal test machine (Autograph, DCS-5000,
Shimadzu, Kyoto, Japan). A pulling force was applied to the boneimplant segment in a direction
perpendicular to the boneimplant interfaces, with cross-head speed set at 2.5 mm/min. The load
that broke the boneimplant interface was automatically recorded.
To calculate the boneimplant interfacial area, the bony side of the sample was measured after
tensile fracture by means of image analysis. The whole bony surface that contacted the implant
was used as the interfacial area. The tensile strength was determined by dividing the load obtained
during testing by the interfacial area.
The advantages of the testing method used by Lin et al.

11

are that tensile testing can be performed
twice with one sample and, because the bonding area between bone and implant can be determined
using an image analysis system, the true tensile strength can be obtained.
The disadvantages of their tensile testing method are that a small implant of complicated
shape has to be employed. In addition, the operative procedure is difcult to perform and a
precise preparation of the bony recess is necessary in order to obtain an optimal t between bone
and the implant. Furthermore, this method has not been used in any other study and therefore
lacks the accumulated data from the results on other biomaterials using the one method. Another
concern about their method is whether the contact area between the bone and the plate can be
determined exactly.
Previously, Lin et al.

11

evaluated the tensile strength of the interface between hydroxyapatite
and bone using a rabbit tibia model. However, in that study, they used an implant of yet another
geometry, namely, a 3-mm-long semicylindrical shape (U-shape).
Several other research groups have developed their own tensile testing methods. Steinemann
et al.

12

performed tensile testing to evaluate the bond strength between bone and titanium implants
of different surface roughness. They placed disk-shaped implants on the attened surface of the

FIGURE 31.6

Tensile interface testing method. The implant and the piece of bone were held with hooks.
The bone was pulled perpendicularly to the surface of the implant.

0266/ ch31/frame Page 495 Friday, October 22, 1999 10:28 AM

496

Mechanical Testing of Bone and the BoneImplant Interface

monkey ulna and held the implants in place with a spring. Tensile testing was performed 3 to 7
months after implantation. The load was applied with a crosshead displacement speed of 1 mm/min.
Edwards et al.

13

also carried out a study on the tensile strength of the boneHA interface. They
used HA disks (5.25 mm in diameter and 2.30 mm thick) and xed these disks on the attened
surface of the cortex of the rabbit femur. An additional small bony plate was used to stabilize each
test specimen in position on the bone. At the time of testing, the bony plate was removed. A tensile
load was applied through a stainless steel cable through the hole in the top of the Ti6Al4V part
that had been bonded to the HA disk. The stroke control was set at 1 mm/min and the greatest load
during the test was reported as the failure load.
Kangasniemi et al.

14

examined the tensile strength of uorapatite and coatings sprayed with
HA-plasma. A cylinder 5 mm in diameter and 3 mm in height was coated with bioactive material.
The cylinder was implanted on the attened surface of the cortex of the femur of goats and was
xed with wire bound around the femur. In tensile testing, the pulloff wires were attached to the
implant, and the whole system was placed in the support jig with wires downward so that the bone
could be left resting freely during testing. A crosshead speed of 0.1 mm/min was used in the
Hounseld testing apparatus.
As reviewed above, all these tensile testing methods are characterized by differences in exper-
imental animals, implantation sites, surgical techniques, implant-xation methods, testing machines,
and crosshead speeds. Furthermore, most testing methods were used in one study only, and the
results cannot be compared easily with other studies.

IV. CONCLUSION

In recent years, several tensile testing methods have been developed to evaluate the bonding strength
between bone and biomaterials. However, a standardized method, with respect to choice of animal,
implantation site, surgical technique, method of implant xation, testing device, and crosshead
speed has not yet been established. In order to evaluate the bonding abilities of biomaterials and
to compare the bonding strengths of biomaterials with each other, it is desirable to establish a
standard method of tensile testing as soon as possible. The testing method described is simple to
perform. The method has the additional advantage that it has already been used to accumulate
signicant amounts of data on many different kinds of biomaterials. The authors believe that it is
an appropriate method of choice for tensile testing of biomaterials.

REFERENCES

1. Johansson, C. and Albrektsson, T., Integration of screw implants in rabbit: a 1-year follow-up of
removal torques of Ti implants,

Int. J. Oral Maxillofac. Implants

, 2, 69, 1987.
2. Cook, S.D., Thomas, K.A., Kay, J.F., and Jarcho, M., Hydroxyapatite-coated titanium for orthopedic
implant application.,

Clin. Orthop.

, 232, 225, 1988.
3. Nakamura, T., Yamarumo, T., Higashi, S., et al., A new glass-ceramic for bone replacements: evaluation
of its bonding ability to bone tissue,

J. Biomed. Mater. Res

., 23, 631, 1985.
4. Kotani, S., Yamamuro, T., Nakamura, T., et al., Enhancement of bone bonding to bioactive ceramics
by demineralized bone powder,

Clin. Orthop.,

278, 1311, 1992.
5. Kitsugi, T., Nakamura, T., Oka, M., et al., Bone-bonding behavior of plasma-sprayed coatings of
Bioglass, AW-glass ceramic, and tricalcium phosphate on titanium alloy,

J. Biomed. Mater. Res.

, 30,
261, 1996.
6. Nishiguchi, S., Nakamura, T., Kobayashi, M., et al., Effect of heat treatment on bone bonding ability
of heat-treated titanium,

Biomaterials

, 29, 491, 1999.
7. Takatsuka, K., Yamamuro, T., Nakamura, T., and Kokubo, T., Bone-bonding behavior of titanium
alloy evaluated mechanically with detaching failure load,

J. Biomed. Mater. Res.

, 29, 157, 1995.

0266/ ch31/frame Page 496 Friday, October 22, 1999 10:28 AM

Tensile Testing of BoneImplant Interface

497

8. Tamura, J., Kitsugi,T., Iida, H., et al., Bone bonding ability of bioactive bone cements,

Clin. Orthop.

,
343, 183, 1997.
9. Yoshii, S., Kakutani,Y., Yamamuro, T., et al., Strength of bonding between A-W glass-ceramic and
the surface of bone cortex,

J. Biomed. Mater. Res.

, 22(3 Suppl.), 327, 1988.
10. Lin, H., Xu, H., and de Groot, K., Tensile strength of the interface between hydroxyapatite and bone,

J. Biomed. Mater. Res.

, 26, 7, 1992.
11. Lin,H., Xu, H., Zhang, X., and de Groot, K., Tensile tests of interface between bone and plasma-
sprayed HA coating-titanium implant,

J. Biomed. Mater. Res.

, 43, 113, 1998.
12. Steinemann, S.G., Eulenberger, J., Maeusli, P.A., and Schroeder, A., Adhesion of bone to titanium,
in

Biological and Biomechanical Performance of Biomaterials

, Christel, P., Meunier, A., and Lee,
A.J., Eds., Elsevier Science, Amsterdam, 1986.
13. Edwards, J.T., Brunski, J.B., and Higuchi, H.W., Mechanical and morphologic investigation of the
tensile strength of a bone-hydroxyapatite interface,

J. Biomed. Mater. Res.

, 36, 454, 1997.
14. Kangasniemi, I.M., Verheyen, C.C., Velde, E.A., and de Groot, K.,

In



vivo

tensile testing of uorapatite
and hydroxyapatite plasma-sprayed coatings,

J. Biomed. Mater. Res.

, 28, 563, 1994.

0266/ ch31/frame Page 497 Friday, October 22, 1999 10:28 AM

0266/ ch31/frame Page 498 Friday, October 22, 1999 10:28 AM

499

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

32

Fracture Toughness Tests
of the BoneImplant Interface

Xiaodu Wang, Kyriacos A. Athanasiou, and C. Mauli Agrawal

CONTENTS

I. Introduction ..........................................................................................................................499
II. Compact Sandwich Fracture Toughness Test ......................................................................502
A. Specimen Preparation.....................................................................................................502
B. Test Procedure ................................................................................................................503
C. Data Interpretation..........................................................................................................504
D. Limitations......................................................................................................................507
III. Mixed-Mode Sandwich Fracture Toughness Test ...............................................................507
A. Specimen Preparation.....................................................................................................508
B. Test Procedure ................................................................................................................508
C. Data Interpretation..........................................................................................................509
D. Limitations......................................................................................................................510
IV. Short-Bar and Short-Rod Fracture Toughness Test.............................................................510
A. Determination of

G

IC

Using a Short-Bar Specimen......................................................510
1. Specimen Preparation................................................................................................510
2. Test Procedure ...........................................................................................................511
3. Data Interpretation.....................................................................................................512
B. Determination of

K

IC

Using a Short-Rod Specimen .....................................................512
1. Specimen Preparation................................................................................................512
2. Test Procedure ...........................................................................................................513
3. Data Interpretation.....................................................................................................513
C. Limitations......................................................................................................................513
References ......................................................................................................................................513

I. INTRODUCTION

Interfacial strength between bone and implants is critical to the success of both orthopaedic and
dental implant systems. To evaluate an implant system, it is imperative for researchers to assess
the intrinsic interfacial strength between bone and implant. Currently, there are no standardized
test techniques. However, some interfacial fracture toughness tests have been developed and are
used in orthopaedic and dental research. This chapter presents three relatively well-dened inter-
facial fracture toughness test specimens: compact sandwich specimen, mixed mode specimen, and
short-bar or short-rod specimen.
Interfacial debonding between tissue and implant surfaces is a major reason for failure of both
orthopaedic and dental implant systems. The mechanical strength of boneimplant interfaces not
only has a direct impact on short-term success, but also inuences long-term stability of the
prosthesis.

9,15

There are many variables that inuence the mechanical integrity of the interface such

0266/ ch32/frame Page 499 Friday, October 22, 1999 10:31 AM

500

Mechanical Testing of Bone and the BoneImplant Interface

as interfacial bonding, the intrinsic properties of implanted materials and surrounding tissues, and
structural characteristics and external loading modes. For instance, porous surfaces and bioactive
materials have been proved to be benecial for the stability of implant systems due to enhanced
interfacial bonding between implant and bone tissue. On the other hand, wear debris from the
articulating surfaces of total joint prostheses can induce bone loss (osteolysis) around the implant
and consequently lead to loosening of the prosthesis. Furthermore, high-cycle, low-magnitude
loading may cause a combination of fatigue failures at the interface, in the implant, and in the
surrounding bone tissues. To understand fully the implications of all these variables on the intrinsic
fracture properties of the interface, it is imperative that a suitably reliable and accurate mechanical
test be developed.
Over the years, the pushout test has been commonly used for evaluating the interface

1,5,6,15

because it is straightforward and simple. Use of simple tensile tests also has been suggested by
some researchers.

9,10,16

From an engineering perspective, both pushout and tensile tests are akin to
structural tests, which yield the nominal strength of interfaces, but not the intrinsic interfacial
strength. In recent years, fracture mechanics approaches have drawn an interest in testing interfacial
bonding of biological systems. Because fracture mechanics estimates the interfacial strength in
terms of the energy or stress intensity required to propagate an interfacial crack along the inter-
face,

6,7,13

this critical energy or stress intensity may be a fundamental and meaningful measure of
the intrinsic fracture properties of the interface. There are two major parameters commonly used
to assess the fracture toughness: critical stress intensity factor,

K

C

,

and critical energy release rate
(driving force),

G

C

.

Figure 32.1 shows a schematic representation of these parameters.

K

C

denes
the magnitude of stress at the crack tip from where the crack starts to propagate, while

G

C

gives
a measure of energy which is needed to create new fracture surfaces concomitant with an initial
crack extension (

a

). As illustrated in Figure 32.2, there are three modes of loading that a crack
can experience. In Mode I loading, the principal load is applied normal to the crack plane; since
the crack tends to open under such loading, this type of loading is also known as the opening mode.
Mode II is characterized by the sliding of one crack face against the other in the direction of crack
propagation, and so it is regarded as the shearing mode. Mode III refers to a twisting mode because
the shear load is applied in a direction normal to the crack propagation direction. In reality, an
interfacial crack can be loaded in any of these modes, or a combination of two or three of them.
In general, a crack tends to propagate along the path of least resistance, which is not necessarily
conned to its initial plane. However, in the case of interfacial fractures of boneimplant interfaces,
crack propagation commonly occurs at the interfaces.
There are no standard interfacial fracture toughness tests for biological systems. Among the
techniques that have been developed thus far, sandwich (Figure 32.3) and the short-rod or short-
bar (Figure 32.4) specimens are relatively well dened. The short-rod or short-bar specimen has
been widely used in dental biomaterials research,

6,8,12

while the sandwich specimen was developed

FIGURE 32.1

A conceptual diagram showing definitions of the fracture toughness parameters.

K

C

is the
critical stress intensity factor, and

G

C



is the critical strain energy release rate.

0266/ ch32/frame Page 500 Friday, October 22, 1999 10:31 AM

Fracture Toughness Tests of the BoneImplant Interface

501

for biological systems in general.

18,19

The distinguishing feature of these specimens is that a layer
of one material is sandwiched between two layers of another material with a precrack lying at one
of the two interfaces. In fact, these specimens were originally developed for testing adhesive joints
and interfacial bonding strength of composite laminates. These techniques have been adapted to
testing the interfacial bonding strength between bone and implant.
It should be borne in mind that these fracture toughness tests have their own limitations, because
they are developed mainly based on linear elastic fracture mechanics. First of all, these tests are
not suitable for testing material systems with large plastic deformation, although some corrections
can be made to take into account small-scale plastic deformation at the vicinity of the crack tip.
Second, these tests cannot be used to assess time-dependent fracture behavior. If viscoelastic
materials are involved in the material system, a slow and constant loading speed is recommended
for these tests. Third, these tests can be used only for those interfaces that are formed between

FIGURE 32.2

Three loading modes that a crack can experience: (a) opening mode (tension); (b) shear mode;
and (c) tear mode.

FIGURE 32.3

A schematic representation of the compact sandwich specimen. This specimen shares a similar
overall conguration with the standard compact tension specimen, but has a bilayer specimen sandwiched
between two holders. The holder material must be the same, or similar to one of the layers of the bilayer
specimen.

0266/ ch32/frame Page 501 Friday, October 22, 1999 10:31 AM

502

Mechanical Testing of Bone and the BoneImplant Interface

relatively smooth surfaces; otherwise, most assumptions on which the techniques are derived become
invalid. Finally, these tests require relatively rigorous specimen preparation procedures, such that well-
developed skills and hands-on experience are needed to ensure the accuracy of these tests.

II. COMPACT SANDWICH FRACTURE TOUGHNESS TEST

The compact sandwich specimen is designed to measure interfacial fracture toughness under an
opening mode (Mode I, tension). Except for the interlayer (see Figure 32.3), this specimen is similar
to a homogeneous compact tension specimen described in ASTM standard E399-90.

2

A. S

PECIMEN

P

REPARATION

A compact sandwich specimen consists of two components: a bilayer specimen and two holders
(see Figure 32.3). The bilayer component comprises the interface between bone and a biomaterial
of interest. As far as the overall geometry is concerned, the proportions and tolerances of the
specimen follow those recommended by the ASTM standard for compact tension specimens.

2

This
technique requires that the holder material be the same as or similar to either the bone or the
biomaterial in the bilayer component. As shown in Figure 32.3, the layer in the bilayer which has
material properties different from the holders is dened as the interlayer.
The interface can be formed either

in vivo

or

in vitro

depending on the purpose of ones study.
Figure 32.5 shows an example of preparing

in vivo

formed bilayer specimens.

1

A coupon made of
the test biomaterial had dimensions of 20

3.5

2 mm. This specimen was implanted with one
surface xed ush against a defect made on the proximal tibia of a rabbit. Briey, the rabbit was
anesthetized and the incision region was shaved and sterilized. In order to ensure a consistent
surgical outcome, a custom-designed xture was utilized to create the defect. The leg was then
secured to the xture, draped, and a longitudinal incision was made just below the tibial tuberosity
to expose the subperiosteal surface of the medial aspect of the proximal tibia (Figure 32.5). Then,
a shallow, trough-shaped groove (about 0.2 mm deep) was created on the subperiosteal surface of
the tibia using a 4-mm carbide router. Finally, the implant was secured in the groove using sutures
with the surface of interest against the defect in the bone. The incision was closed in a routine
fashion. In order to facilitate initiation of the precrack at the boneimplant interface, one third of

FIGURE 32.4

Schematic diagrams of the short bar (a) and short rod (b) specimens used for testing
dentindental composite interfaces. A chevron notch is made at the interface between dentin and the
biomaterial tested.

0266/ ch32/frame Page 502 Friday, October 22, 1999 10:31 AM

Fracture Toughness Tests of the BoneImplant Interface

503

the surface of these implants were covered with a thin layer of bone wax (Ethicon Inc., Somerville,
NJ) so as to prevent bone ingrowth from occurring. At sacrice, the implant was excised with the
bone tissue attached. All extra tissue was carefully removed using a diamond circular saw and a
bench-top milling machine to form a rectangular bilayer specimen. Extra precautions were taken
during the specimen preparation to avoid damage to the boneimplant interface. The bilayer
specimen was cemented between the two holders using a cyanoacrylate adhesive to form the
compact sandwich specimen.

19

The overall dimensions of the specimen were 20

19

3.5 mm.
The holder was made of bovine femur to ensure that no signicant differences in elastic properties
would occur between the holders and the bony layer of bilayer specimens; such differences would
violate the assumptions on which this technique is based and consequently affect accuracy of the
test. In this case, the interlayer was the biomaterial to be tested.
The next example considers a bilayer specimen consisting of an

in vitro

formed bone-to-bone
cement interface (Figure 32.6).

19

In this test, the interfacial strength of bone-to-bone cement was
tested. Bone coupons were obtained from a bovine femur and machined to nal dimensions of
21.9

3.5

2.0 mm. To prepare the bone cement, liquid monomer and polymer powder were mixed
at room temperature (23

C) following the manufacturers instruction. The cement was then poured
on the top of the bone coupon which was placed in the base of a mold (see Figure 32.6). Pressure
was applied uniformly through the mold cap for approximately 20 min. Because it is difcult to
obtain very accurate specimen dimensions from the molding process, the bilayer specimens were
initially made larger than the required size and machined to the nal desired dimensions. The
bilayer was carefully cemented between two PMMA (polymethylmethacrylate) holders using a
cyanoacrylate adhesive (Quicktite, Loctite Corp.). Any extra adhesive on either side of the
specimen was polished off using ne sandpaper. A starter notch was cut at the interface using a
circular saw with a 0.5-mm-thick blade, then a precrack was introduced at the tip of the notch
using a sharp razor blade. In this case, the interlayer material was bone.
Several precautions need to be taken to ensure testing success. First, the elastic properties (i.e.,
elastic modulus and Poissons ratio) of the holder must be the same as or similar to the bone specimens
or biomaterial to be tested, as the governing equations are derived based on this assumption. Second,
the accuracy of specimen shape and size is critical to the results obtained using the compact sandwich
specimen because the results are very sensitive to these parameters. Finally, extreme care has to be
taken during specimen preparation to avoid damaging the interface of interest.

B. T

EST

P

ROCEDURE

Precaution needs to be taken while mounting the compact sandwich specimen in a material testing
machine not to prestress the interface. The specimen is loaded until failure of the interface. Clevises
(U-shaped xtures) are needed to facilitate loading of the compact sandwich specimens. The design

FIGURE 32.5

An example showing the preparation of an

in vivo

formed bilayer specimen. An implant was
placed against a shallow slot previously made on the medial subperiosteal surface of the proximal tibia in a
rabbit. Sutures were used to secure the implant in position.

0266/ ch32/frame Page 503 Friday, October 22, 1999 10:31 AM

504

Mechanical Testing of Bone and the BoneImplant Interface

of the clevises is outlined in ASTM standard E399-90.

2

A water-soluble dye can be used at the
crack tip to facilitate the measurement of the crack length (

a

) under an optical microscope. The
specimen width (

W

) and thickness (

B

) can be determined using a digital caliper micrometer. To
avoid time-dependent effects (viscoelasticity), the loading rate at the crack tip, dened as the
increment of stress intensity factor in unit time (

K/s

), should be kept low and constant (e.g., 0.15
to 0.30 MPa /s) by adjusting the crosshead speed of the test machine. The curve of force vs.
displacement is recorded, and the peak load is determined as the critical load,

P

C

.

C. D

ATA

I

NTERPRETATION

The sandwich technique was originally developed to test engineering material systems.

11,16

Based
on the same concept, this technique was successfully adapted to measuring tissuebiomaterial
interfacial bonding strength.

18,19

The interfacial fracture toughness or critical stress intensity factor
(

K

C

) can be calculated using the following equation developed in a previous study:

19

(32.1)
where

K

C

is the interfacial fracture toughness,

P

C

is the critical load,

is a scale factor determined
by elastic properties of bone and the biomaterial bilayer materials,

is a correction factor which
is a function of the normalized interlayer thickness

h/W

,

19



B

is the thickness of specimen,

W

is the
specimen width, and the remaining parameter

Y

is a constant determined by

W

and initial crack
length

a

. Parameter

can be determined using
(32.2)
where

and

are Dundurs parameters, which estimate the elastic mismatch across an interface
of two dissimilar materials

4

and are given by

FIGURE 32.6

An example of preparing bone-to-bone cement bilayer specimen. Bone cement was press
molded onto a bone coupon to form a bilayer specimen in a mold.
m
K
P Y
B W
C
C
=

1
1
2

0266/ ch32/frame Page 504 Friday, October 22, 1999 10:31 AM

Fracture Toughness Tests of the BoneImplant Interface

505

(32.3)
(32.4)
where

E

1

,

E

2

, and

1

,

2

are elastic moduli and Poissons ratios of the interlayer and the other
material of the bilayer specimen, respectively. Additionally, in Equation 32.1,

is the correction
factor for the nite interlayer thickness h, and is expressed as a polynomial function of h/W (in
the range of 0.0 <

h/W

<0.12) as
(32.5)
Here,

h

is the thickness of the interlayer.
Finally,

Y

can be directly calculated using the standard equation for compact tension specimens
recommended in ASTM standard E 399-90,

2

(32.6)
From

K

C

,

the critical strain energy release rate can be calculated as follows:
(32.7)
where

is a parameter determined by one of the Dundurs parameters of material mismatch,

:
(32.8)
An extremely important aspect of compact sandwich specimens is the existence of complex
loading modes at the crack tip. Unlike homogeneous test specimens, the interfacial crack of compact
sandwich specimens is actually loaded under both tension and shear loads because of the material
mismatch across the interface.

9

However, the interfacial fracture toughness cannot be simply
separated into conventional Mode I (tension) and Mode II (shear).

9

To address this issue, a nominal-
phase angle,

, can be used to assess the ratio of the shear component with respect to the tensile
component of the interfacial fracture toughness.
(32.9)

=

+
E E
E E
1 2
2
2 1
2
1 2
2
2 1
2
1 1
1 1
( ) ( )
( ) ( )

=

+
E E
E E
1 2 2
2
2 1 1
2
1 2
2
2 1
2
1 2 1 2
2 1 2 1
( ) ( )
( ) ( )
=

1 00 14 07 136 0 538 8
2 3
. . . .
h
W
h
W
h
W
Y
a
W
a
W
a
W
a
W
a
W
a
W
= +

2
0 886 4 64 13 32 14 72 5 6
1
2 3 4
3 2
. . . . .
/
G
E E
E E
K
C C
=
+
1 2
2
2 1
2
1 2
2
2 1 1
2
( ) ( )
cosh ( )

1
2
1
1
ln
=

tan
1
Shear component of
Tensilecomponent of
K
K
C
C

0266/ ch32/frame Page 505 Friday, October 22, 1999 10:31 AM

506

Mechanical Testing of Bone and the BoneImplant Interface

can be determined using the following procedure:

19

First, a material system having very small

(< 0.01) is selected arbitrarily as control, and then the relative discrepancy in phase angle,

,

between the test specimen and the control is calculated using the following equation:
(32.10)
where

is an auxiliary parameter determined by

and

(Table 32.1),

is the parameter outlined
in Equation 32.8, and

h

is the interlayer thickness for the test specimen.

0

,

h
0
, and
0
are the values
for the control. Then, can be obtained as
(32.11)
where
0
(=
0
+
0
ln r/h
0
]) is the phase angle of the control as outlined in a previous study.
19
Although r is an arbitrary number in this case,
9
for simplicity r is xed at of h
0
. Table 32.2
shows the values of , , , , and for the material systems often seen in orthopaedic applications.
In fact, the loading modes at the crack tip are primarily determined by the material mismatch, and
interlayer thickness, h, of the compact sandwich specimen.
TABLE 32.1
Values of as a Function of and

0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8
0.4 2.2 3.5
0.3 3.0 4.0 3.3 1.4
0.2 3.6 4.1 3.4 2.0 0.3 3.3
0.1 4.0 4.1 3.3 2.0 0.1 2.3 5.5 10.8
0.0 4.4 3.8 2.9 1.6 0.0 2.1 4.7 8.4 14.3
0.1 2.3 1.1 0.5 2.3 4.5 7.4 11.6
0.2 1.3 3.0 4.9 7.3 10.5
0.3 5.8 7.8 10.4
0.4 11.1
This table was reproduced based on the work by Suo and Hutchinson.
11
TABLE 32.2
Values of , b, , , and for Some BoneBiomaterial Systems
Holder/Interlayer E (GPa) ()
Human bone/CoCr 11.8/234 0.25/0.30 0.907 0.304 0.10 1.450 2
PMMA/bovine bone 2.8/20 0.35/0.25 0.764 0.164 0.051 1.336 3.5
Bovine/human bone 20/11.8 0.25/0.25 0.258 0.086 0.027 0.865 2.7
Ti alloy/bovine bone 110/20 0.30/0.25 0.70 0.24 0.078 0.564 9.1
CoCr/human bone 234/11.8 0.30/0.25 0.907 0.304 0.10 0.320 12
PMMA: polymethylmethacrylate; CoCr: cobaltchrome alloy; Ti: titanium.


=
( )
0 0
0
ln / h h
= +
0

0266/ ch32/frame Page 506 Friday, October 22, 1999 10:31 AM

Fracture Toughness Tests of the BoneImplant Interface 507
D. LIMITATIONS
The compact sandwich test is not valid for measuring the fracture toughness of very rough or
irregular interfaces. Currently, there are no accurate means to assess the fracture toughness of such
interfaces. In addition, linear elasticity and isotropy are assumed for this test. However, the elastic-
plastic behavior at the crack tip may play an important role in determining interfacial fracture
toughness. Liechti and Liang
5
have shown that the plastic zone at the interfacial crack tip might
be extremely large if a shear component is present, and it may become acute when the interlayer
is much more compliant compared with the holders. Also, this test does not account for visco-
elasticity, a time-dependent factor, which may also affect the measurement of interfacial fracture
properties. To avoid such effects, one should use a constant loading speed for all specimens to be
compared. Because specimen sizes may vary to accommodate the diverse sizes of in vivo samples,
the overall specimen size may affect the accuracy of the test. A previous study has shown that the
effect of overall specimen size can be ignored as long as the specimen dimensions are proportional.
19
In addition, elastic properties of bone and the biomaterial of the bilayer specimen are needed for
determining the interfacial fracture toughness. Although in most cases these properties can be
obtained from standard mechanical tests or from the literature, it is cumbersome to perform
additional experiments.
III. MIXED-MODE SANDWICH FRACTURE TOUGHNESS TEST
The mixed-mode test for measuring interfacial fracture toughness is an extension of the aforemen-
tioned sandwich technique. In this test, the interfacial crack is loaded in both tension and shear
simultaneously. The specimen conguration is shown in Figure 32.7. The mixed mode specimen
has a different geometry compared with the compact sandwich specimen.
FIGURE 32.7 A schematic diagram of the mixed-mode sandwich specimen. A bilayer specimen is sand-
wiched between two holders. The specimen is mounted onto a loading xture through the holes in the holders.
0266/ ch32/frame Page 507 Friday, October 22, 1999 10:31 AM
508 Mechanical Testing of Bone and the BoneImplant Interface
A. SPECIMEN PREPARATION
Similarly to the compact sandwich specimen, the mixed-mode specimen also consists of two
components: a bilayer test specimen and two holders (see Figure 32.7). There are three pinholes
in each holder to facilitate load transfer from the loading xture to the specimen (Figure 32.8).
The procedure for preparing the bilayer specimen is the same as described for the compact
sandwich specimen. In this test, the bilayer specimen is carefully cemented between the two
holders using a cyanoacrylate adhesive. The extra adhesive on both sides of the specimen is
polished off using ne sandpaper. A starter notch is cut at the interface using a circular saw with
a 0.5-mm-thick blade. Then, a precrack can be introduced at the tip of the notch using a sharp
razor blade.
B. TEST PROCEDURE
The test specimen is mounted onto a two-piece buttery-shaped loading xture, which can alter
the loading direction with respect to the interfacial crack plane as shown in Figure 32.8. The loading
direction is changed by selecting corresponding loading holes which are evenly distributed along
the perimeter of the xture. Clevises (U-shaped xtures) are needed to facilitate loading of the
mixed-mode sandwich specimen. The design of the clevis should follow ASTM standard E399-90.
2
A water-soluble dye can be used at the crack tip to facilitate the measurement of the initial crack
length (a) under an optical microscope. The specimen width (W) and thickness (B) can be deter-
mined using a caliper micrometer. In order to avoid time-dependent effects (e.g., viscoelasticity),
the loading rate at the crack tip, dened as the increment of stress intensity factor in a unit time
(K/s), should be kept slow and constant (e.g., 0.15 to 0.30 MPam/s) by adjusting the crosshead
speed of the test machine. The critical load, P
C
, can be determined as the peak load from the force
vs. displacement curve recorded.
FIGURE 32.8 A custom-designed loading fixture (butterfly-loading fixture) to facilitate variation of load-
ing direction () with respect to the crack plane. Trough-shaped mounting pins are used to reduce the
contact area between the pins and holes to avoid uneven load transfer to the test specimens. The lateral
movement of the specimen is restricted by the middle holes, and axial movement of the specimen is restricted
by the side holes.
0266/ ch32/frame Page 508 Friday, October 22, 1999 10:31 AM
Fracture Toughness Tests of the BoneImplant Interface 509
C. DATA INTERPRETATION
The mixed-mode test is designed to assess the interfacial fracture toughness of a bonebiomaterial
system under a mixture of different loading modes (shear and tension). Its overall effect can be assessed
using a mixed-mode fracture toughness, K
C
, which can be calculated using the following equation:
17

(32.12)
where P
C
is the critical load, a is the initial crack length, W and B are the width and thickness of
the specimen, is the loading angle, and A
I
and A
II
are functions of a and W:
10
(32.13)
(32.14)
Parameter is the correction factor for nite interlayer thickness, h, and is expressed as a
function of h/W and loading angle, .
17
(32.15)
where p and k are functions of loading angle, , and are given as:
(32.16)
(32.17)
Here, is in radians. The interfacial fracture toughness of the mixed-mode specimen cannot be
simply separated into conventional Mode I (tension) and Mode II (shear).
9
As outlined for the
compact sandwich specimens, the nominal phase angle, , for the mixed-mode specimen can be
calculated as:
(32.18)
where is the loading angle, is an auxiliary parameter determined by and , is determined
by as outlined in Equation 32.8, and h is the interlayer thickness for the test specimen, while

0
,
0
,
0
,
0
, and h
0
are the values for the control. For the mixed-mode specimens, the loading
modes at the crack tip are primarily affected by the material mismatch across the interface and
the interlayer thickness, h.
K
P a
WB
A A
C
C
=
( )
+
( )

I II
cos sin
2 2
A
a
W
a
W a
a
W a
a
W a
I
=

1
1
0 26 2 65
1 0 55 0 08
2
. .
. .
A
a
W
a
W a
a
W a
a
W a
II
=

1
1
0 23 1 40
1 0 67 2 08
2
. .
. .
=
( )
e
p h W
k
p = + 5 056 1 20 266
2
. .
k = 0 777 0 124 0 028
2
. . .

= + +

0 0 0
0
0
ln
h
h
0266/ ch32/frame Page 509 Friday, October 22, 1999 10:31 AM
510 Mechanical Testing of Bone and the BoneImplant Interface
D. LIMITATIONS
The mixed-mode specimen has the same limitations outlined for the compact sandwich specimen.
IV. SHORT-BAR AND SHORT-ROD FRACTURE TOUGHNESS TEST
The short-bar or short-rod specimen was rst developed by Barker
3
to evaluate engineering mate-
rials. Because it has a small and simple shape, and no precracking procedure is needed, this
technique has been successfully adapted for testing of interfacial bonding at dentin/biomaterial
interfaces.
6,8,14
The short-bar or short-rod specimen consists of two halves with a bonebiomaterial
interface between, as shown in Figure 32.4. Half of the specimen is made of the biomaterial, and
the other half is a bonebiomaterial laminate. Thus far, two different procedures have been used;
6,8,14
one of them using the short-bar specimen to determine the critical strain energy release rate, G
C
,
and the other using the short-rod specimen to assess the critical stress intensity factor, K
C
, of the
interfaces. The following are examples of these two procedures.
A. DETERMINATION OF G
IC
USING A SHORT-BAR SPECIMEN
1. Specimen Preparation
Lin and Douglas
6
proposed a procedure to determine the critical strain energy release rate, G
C
,
using a short-bar specimen. In this test, equal compliance for the two halves of the specimen was
required. Specically, the two halves of the short bar had the same loaddeformation relationship.
This was achieved by experimentally determining the height of each half of the short-bar specimen
which would give the same compliance. Figure 32.9 shows the schematic diagram of the short-bar
specimen used in their study. Two steps were involved for this procedure; rst, a short-bar specimen
was made by cementing a layer of dentin between two blocks of dental composite (biomaterial).
A chevron notch was made in the middle of the dentin layer. This specimen was loaded at a loading
rate of 0.001 mm/s until the specimen split into two halves. One half of the fractured specimen
was used as the dentincomposite laminate for fabricating the interfacial fracture toughness test
specimens. The fractured surface of dentin of the dentincomposite specimen was wet polished to
a at surface using 600-grit sandpaper. The load vs. mouth-opening displacement curve of the
specimen was recorded and the slope of the linear part of the curve was used to calculate the
compliance of the dentincomposite laminate, C
1
, as
(32.19)
where is the mouth-opening displacement of the notch and P is the load. C
1
was used to determine
the height of the other half of the specimen (composite), which should have the same compliance
as the dentincomposite laminate. To do so, another set of short-bar specimens of the composite
were made with the same specimen length (W) and width (D) but varying specimen height (H)
compared with the dentincomposite specimen described above. The compliance of these specimens
was measured and the height of the specimens was adjusted until the compliance of the specimen
was equal to C
1
. Finally, the dental composite half with the new height was molded on the
dentincomposite laminate to form the nal short-bar specimen. Teon tape was used to cover all
but the V-shaped dentin-bonding surface of the dentin-composite laminate during the molding
process to form a chevron notch.
To calculate fracture toughness using the short-bar specimen, a parameter called the mini-
mum geometry factor coefcient, Y
m
, must be determined. Lin and Douglas
6
used a three-
dimensional nite-element model for this purpose. This model was generated to simulate a
C
P
1
=

0266/ ch32/frame Page 510 Friday, October 22, 1999 10:31 AM

Fracture Toughness Tests of the BoneImplant Interface 511
short-bar specimen of the dental composite with the height determined earlier. A previous study
has shown that the values of the geometry factor coefcient of the specimen, Y, can be calculated
using the following equation:
3

(32.20)
where is the normalized crack length (a/W),
0
is the normalized initial crack length (a
0
/W),
1
is the normalized chevron-notch length (a
1
/W), a is the crack length from loading line to crack
front, a
0
is the initial crack length from loading line to the vertex of the chevron, a
1
is the maximum
depth of chevron anks from the loading line to the base of the chevron, C is the specimen
compliance, D is the specimen thickness, and E is the elastic modulus of the composite. Since D,
E,
0
, and
1
are constants, Y is a function of dC/d and . By altering the crack length in the
model, the compliance (C) of the specimens was calculated with respect to the normalized crack
length (). The curve of the compliance vs. the normalized crack length was then used to determine
the values of dC/d with respect to the normalized crack length (). Finally, the value of Y
m
was
determined from the curve of Y vs. . In the case of this example, the value of Y
m
was determined
as 17.3 at the normalized crack length = 0.48.
2. Test Procedure
The short-bar specimen is mounted onto a material-testing machine using a custom-designed xture,
in which the specimen is held in knife-edged loading arms as shown in Figure 32.9. Precautions need
to be taken to ensure that the orientation of the loading line is perpendicular to the chevron-notch
fracture plane of the specimen. The test is run at a constant loading rate of 0.001 mm/s by adjusting
the crosshead speed of the test machine. Several loadingunloading cycles are performed during the
test until the maximum failure load is reached. The cycle which includes the maximum load is dened
as the effective cycle. Load and displacement values are recorded to plot the curve of load vs. mouth-
FIGURE 32.9 An example of the short-bar specimen used by Lin and Douglas.
6
The specimen was loaded
through two knife-edged loading arms, which were placed between the two lips at the head of the specimen.
Y
DE dC
d
=

2
1 0
0
1 2


/
0266/ ch32/frame Page 511 Friday, October 22, 1999 10:31 AM
512 Mechanical Testing of Bone and the BoneImplant Interface
opening displacement. From the curve, the average load of the effective cycle was determined as the
critical load, P
C
. Details for obtaining the critical load can be found elsewhere.
3
3. Data Interpretation
The critical strain energy release rate, G
IC
, for a short-bar specimen can be calculated using the
following equation:
6
(32.21)
where P
C
is the critical load measured at time of fracture, D is the specimen thickness for the short-
bar specimen, E is the elastic modulus of the composite, W is the specimen length, and Y
m
is the
minimum geometric factor coefcient.
B. DETERMINATION OF K
IC
USING A SHORT ROD SPECIMEN
1. Specimen Preparation
Tam and Pilliar
12,14
have used short-rod specimens (Figure 32.10) to assess the interfacial bonding
of dentin and dental biomaterials. In their studies, slices were sectioned from animal teeth and cut
with a high speed burr, under constant water irrigation, to a rectangular coupon (7 4 1.5 mm).
The slice was inserted into a stainless steel mold, and the dental composite was packed into the
mold to form a half specimen. The dentin surface of the half specimen was wet polished with
600-grit sandpaper, rinsed, and dried. The dried dentin surface was then pre-treated, dried with
compressed air, and coated with primer. Thereafter, the half specimen was placed in another mold
and a spacer made of highly polished stainless steel was placed on top of it to form a chevron-
shaped notch after molding. The adhesive agent was brushed onto the dentin surface and light-
cured without prior air thinning. The dental composite was then packed into the remainder of the
mold. After complete polymerization was ensured, the specimen was separated from the mold. The
minimum geometry factor coefcient, Y
m
, was required for the short-rod specimen in order to
calculate fracture toughness values. In this test, Y
m
was determined experimentally with the com-
pliance calibration curves of the specimens using a similar procedure outlined earlier.
FIGURE 32.10 An example of the short-bar specimen developed in previous studies.
12
The specimen was
loaded through the holes in the two loading anges at the head of the specimen.
G
P Y
ED W
IC
C m
=
2 2
2
0266/ ch32/frame Page 512 Friday, October 22, 1999 10:31 AM
Fracture Toughness Tests of the BoneImplant Interface 513
2. Test Procedure
The short-rod specimen is mounted on a mechanical testing machine using clevises. Precaution is
taken to avoid prestressing the interface during the process. A load is applied to the specimen at a
loading rate of 0.5 mm/min until failure of the interface. The peak load, P
C
, is recorded at the time
of failure. The initial crack length (a
0
), specimen length (W), and specimen diameter (D) were
determined by the dimensions of the mold.
3. Data Interpretation
The critical stress intensity factor, K
IC
, can be obtained from this test using the following
equation:
12,15

(32.22)
where P
C
is the peak load measured at time of fracture, D is the diameter of the short-rod specimen,
W is the specimen length, and Y
m
is the minimum stress intensity factor coefcient.
C. LIMITATIONS
A drawback of the short-bar or short-rod specimen is that for each conguration and material
combination preliminary experiments are needed to determine the specimen geometry, to ascertain
the geometry factor coefcient, and to ensure equal compliance for the two halves of the specimen.
Thus, investigators must be quite experienced in specimen preparation. They also must be quite
knowledgeable in the nuances of fracture mechanics. The short-bar and short-rod specimens, which
are similar to sandwich specimens, are developed based on the assumption of linear elasticity. Thus,
this approach is not valid for a specimen with large plastic deformation. Additionally, Liechti and
Liang
5
have shown that the plastic zone at the interfacial crack tip might be extremely large if there
is a shear component present, and it may become acute when the material mismatch is signicant
across the interface. Moreover, this test is valid only for measuring the fracture toughness of smooth
interfaces because it has been developed based on this assumption. Also, this approach does not
account for viscoelasticity. To avoid such effects, one should use a constant loading speed for all
specimens to be compared.
REFERENCES
1. Agrawal, C.M., Wang, X., and Mabrey, J., Bone ingrowth into porous coated implants impregnated
with PLA-PGA and TGF-, unpublished.
2. ASTM, E399-90 Standard test method for plane strain fracture toughness of metallic materials, in
1993 Annual Book of ASTM Standards, ASTM, West Conshohocken, PA, 1993.
3. Barker, L.M., Theory for determining K
IC
from small, non-LEFM specimens, supported by experi-
ments on aluminum, Int. J. Fract., 15, 515, 1979.
4. Dundurs, J., Edge-bonded dissimilar orthogonal elastic wedges under normal and shear loading,
J. Appl. Mech., 36, 650, 1969.
5. Liechti, K.M. and Liang, Y.-M., The interfacial fracture characteristics of biomaterial and sandwich
strip blister specimens, Int. J. Fract., 55, 95, 1992.
6. Lin, C.P. and Douglas, W.H., Failure mechanism at the human dentin-resin interface: a fracture
mechanics approach, J. Biomech., 27, 1037, 1994.
7. Pilliar, R.M., Davies, J.E. and Smith, D.C., The bonebiomaterial interface for load-bearing implants,
MRS Bull., 155, 55, 1991.
K
P Y
D W
IC
C m
=
0266/ ch32/frame Page 513 Friday, October 22, 1999 10:31 AM
514 Mechanical Testing of Bone and the BoneImplant Interface
8. Pilliar, R.M., Smith, D.C., and Maric, B., Fracture toughness of dental composites determined using
the short-rod fracture toughness test, J. Dent. Res., 65, 1308, 1986.
9. Rice, J.R., Elastic fracture mechanics concept for interfacial cracks, J. Appl. Mech., 55, 98, 1988.
10. Richard, H.A., Examination of brittle fracture criteria for overlapping mode I and mode II loading
applied to cracks, in Proceedings of the International Conference on Application of Fracture Mechan-
ics to Materials and Structures, Freiburg, F.R.G., 1983, 309.
11. Suo, Z. and Hutchinson, J.W., Sandwich test specimen for measuring interface crack toughness, Mater.
Sci. Eng., A107, 135, 1989.
12. Tam, L.E. and Pilliar, R.M., Fracture toughness of dentin/resin-composite adhesive interfaces, J. Dent.
Res., 72, 953, 1993.
13. Tam, L.E. and Pilliar, R.M., Effects of dentin surface treatments on the fracture toughness and tensile
bond strength of a dentincomposite adhesive interface, J. Dent. Res., 73, 1530, 1994.
14. Tam, L.E. and Pilliar, R.M., Fracture surface characterization of dentin-bonded interfacial fracture
toughness specimens, J. Dent. Res., 73, 607, 1994.
15. Wang, C.T. and Pilliar, R.M., Short-rod elastic-plastic fracture toughness test using miniature speci-
mens, J. Mater. Sci., 24, 2391, 1989.
16. Wang, J.S. and Suo, Z., Experimental determination of interfacial toughness curves using Brazil-nut
sandwiches, Acta Metall. Mater., 38, 1279, 1990.
17. Wang, X. and Agrawal, C.M., Mixed mode fracture toughness of bonebiomaterial interfaces, Trans.
Orthop. Res. Soc., 22, 46, 1997.
18. Wang, X., Subramanian, A., Dhanda, R., and Agrawal, C., Testing of bonebiomaterial interfacial
bonding strength: a comparison of different techniques, J. Biomed. Mater. Res., 33, 133, 1996.
19. Wang, X.D. and Agrawal, C.M., Interfacial fracture toughness of tissuebiomaterial systems,
J. Biomed. Mater. Res., 38, 1, 1997.
0266/ ch32/frame Page 514 Friday, October 22, 1999 10:31 AM

515

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

33

In Vitro

Measurements
of Implant Stability

Aivars Berzins and Dale R. Sumner

CONTENTS

I. Introduction ..........................................................................................................................515
II. Stability Measurements and the Number of Measured Motion Components ....................516
III. Stability Measuring Systems and the Loading Setups........................................................521
IV. Conclusion............................................................................................................................523
Acknowledgments ..........................................................................................................................523
References ......................................................................................................................................523

I. INTRODUCTION

The purpose of this chapter is to review some of the basic principles involved in performing

in
vitro

tests of implant stability. After introducing the topic, the motions analyzed in the different
setups are rst considered. Then, the various approaches to applying loads to implant and host
bone are discussed. In general, citations are restricted to papers that have appeared in peer-
reviewed journals.
From the early days of joint replacement surgery, there has been interest in determining how
prosthetic design and surgical technique inuenced xation of the implant to the host bone.

1

This
has been investigated by measuring the relative motion between the bone and the implant. Since
the variety of prosthetic designs has increased and concepts of biological xation have evolved,

in
vitro

measurements are commonly used to test the initial stability of orthopaedic devices implanted
in cadaver bones or synthetic bones. Technically similar tests have also been applied to

postmortem

retrieved specimens to evaluate the long-term stability

ex vivo

.
Femoral or acetabular components of hip replacement systems and femoral or tibial components
of knee replacement systems are the most frequently tested implants, mainly due to concerns about
loosening. A number of studies have used

in vitro

stability tests not only to address xation design
features but also to test new surgical techniques, particularly those used in revision surgery in the
presence of insufcient bone stock. In addition to tests on human cadaver bones, numerous inves-
tigators have used animal models to study implant stability, predominantly with hip and knee
replacement systems.
It is worth mentioning at this point that implant stability has also been investigated

in vivo

in
patients, most commonly with a technique called roentgen stereophotogrammetric or radiostereo-
metric analysis (RSA).

2

In general, this technique has a detection limit of 0.2 to 0.3 mm and is far
superior to conventional radiographic assessment, but is one to two orders of magnitude less
sensitive than many of the

in vitro

techniques described below. RSA and the techniques described
below can measure both the inducible or recoverable motion and the nonrecoverable motion (often
referred to as migration). The recoverable motion is the difference in position of the implant from
the unloaded to the loaded state. After a few load cycles this value usually becomes constant, but

0266/ ch33/frame Page 515 Friday, October 22, 1999 10:34 AM

516

Mechanical Testing of Bone and the BoneImplant Interface

typically is below the detection limit of RSA. The nonrecoverable motion is the difference in
position of the implant from the rst to the last load cycle.

In vivo

, it is easier to measure migration than the recoverable motion because of the limited
sensitivity of RSA and because measurement of recoverable motion requires the subject or inves-
tigator to load and unload the joint repeatedly, and it is difcult to do this reproducibly in the living
subject. In contrast,

in vitro

one can measure either quantity, although it is often simpler to measure
the recoverable motion because only a limited number of load cycles are required (often as few as ten).

II. STABILITY MEASUREMENTS AND THE NUMBER
OF MEASURED MOTION COMPONENTS

In vitro

implant stability testing setups can rst be characterized by the number of separate motion
components (degrees of freedom) which the motion monitoring system is capable of sensing and
recording. Table 33.1 summarizes an immense amount of information and is arranged from relatively
simple setups (e.g., one translation or one rotation) to the most complete description of motion (three
translations and three rotations) for femoral stems, tibial trays, and acetabular components. With the
simplest systems, a motion at a specic location is measured by determining the displacement at that
location (Figure 33.1). In this situation, no assumptions are needed, although it should be recognized
that the investigator has had to decide

a priori

the most important motion component and its location.
To measure all six components of motion, one typically uses an array of dial gauges, exten-
someters, linear variable differential transformers, or eddy current transducers (Figure 33.2). These
are grounded to the surrounding bone in order to sense the displacement of a target device rigidly
attached to the implant. From these motions, the motion at a given point on the implant with respect
to the bone is then calculated. In this type of system, the assumption is made that the bone and
implant behave as two rigid bodies (i.e., neither deform while being loaded). While this is obviously
not always true globally (e.g., along the length of the femoral stem in hip replacement), it has been
shown that the assumption allows for accurate local estimates of motion.

3,4

If one is interested in
motion at different parts of the boneimplant interface (e.g., the proximal and distal aspects of the
femoral stem) and it is known that signicant deformation of the bone and/or implant occurs
between these sites, then the array of transducers must be set up at each local site.

3

In this situation,

FIGURE 33.1

Test setup used by Charnley and Kettlewell in a paper published in 1965.

1

In this case the
authors were interested in axial motion at the distal aspect of the femoral component in cemented hip
replacement. (From Charnley, J. and Kettlewell, J.,

J. Bone Joint Surg.,

47B, 56, 1965. With permission.)

0266/ ch33/frame Page 516 Friday, October 22, 1999 10:34 AM

In Vitro

Measurements of Implant Stability

517
FIGURE 33.2

Setup for a complete six-degree-of-freedom measurement system. (A), left, shows a femur
with an implant

in situ

and the target device attached to the distal aspect of the implant through a hole in the
cortex. (A), right, shows that the six linear variable displacement transducers were attached to the bone via a
rigid collar that did not interfere with the motion of the implant or target device. In the (B) schematic, the
attachment of the six linear variable displacement transducers to the femoral cortex is not shown. Note that
the target was rigidly attached to the implant and that the transducers measured motion of this target. Based
on these measurements, three translations (PA = posterior-anterior, ML = medial-lateral, and SI = superior-
inferior) and three rotations (in the sagittal, transverse, and coronal planes) of the point and axes, respectively,
dened by the intersection of the three planes were calculated. (From Berzins, A., et al.,

J. Orthrop. Res.,

11,
758, 1993. With permission. The published matrix used in the calculations contained a typographic error and
an erratum was published in the

J. Orthop. Res

., 13, 151, 1995).

0266/ ch33/frame Page 517 Friday, October 22, 1999 10:34 AM

518

Mechanical Testing of Bone and the BoneImplant Interface

TABLE 33.1
Summary of Implants Tested, Sensors Used, Number of Motion Components Determined
and the Location of the Measurement Sites

Ref. Implants Tested Sensors Used
Number of Motion
Components Determined
Location of
Measurement Sites

1 Femoral stem Dial gauge Axial translation Distal tip of the stem
46 Femoral stem Extensometer Axial translation Proximal and
distal stem
14 Femoral stem LVDT Axial translation At the crosshead
38 Canine femoral stem LVDT One translation (medial-lateral) Proximal stem
47 Tibial component Two LVDTs Axial translation Medial and
lateral plateau
48 Femoral stem LVDT Transverse rotation

a

Prosthetic neck
49 Femoral stem LVDT Transverse rotation

a

Prosthetic neck
50 Femoral stem LVDT Transverse rotation

a

Prosthetic neck
24 Femoral stem LVDT Transverse rotation

a

At the crosshead
51 Femoral stem LVDT Transverse rotation

a

Prosthetic neck
52 Femoral stem LVDT Transverse rotation

a

Prosthetic neck
16 Tibial component Pressure-sensitive lm Transverse interface motion Implantbone
interface
53 Tibial component LVDT Loaddisplacement curve At the crosshead
21 Femoral stem Video digitizing system Axial translation; transverse
rotation

a

Proximal stem
54 Femoral stem LVDT and rotational
transducer

b

Axial translation; transverse
rotation

a

At the crosshead
41 Canine femoral stem Two extensometers Two translations (axial and
medial-lateral)
Proximal stem
(anterior and
posterior)
55 Canine femoral stem Two eddy current
transducers
Two translations (axial and
medial-lateral)
Proximal stem
35 Femoral stem Extensometer Two translations (axial and
medial-lateral)
Proximal stem
(anterior and
posterior)
56 Femoral stems Extensometer Two translations (axial and
medial-lateral)
Proximal stem
(anterior and
posterior)
39 Femoral implant
(stemless)
Four LVDTs Two translations (anterior-
posterior and medial-lateral)
Prosthetic neck
34 Femoral stem Three strain gauge
displacement transducers
Two translations (axial and
medial-lateral); transverse
rotation

a

Proximal stem
57 Femoral stems Dial gauge, extensometer Two translations (axial and
medial-lateral); transverse
rotation

a

Proximal stem
36 Femoral stem Extensometers Three translations; transverse
rotation
Proximal and
distal stem
58 Femoral stem Six LVDTs Three translations (axial-
proximal, anterior-posterior
and medial-lateral distal);
transverse rotation (proximal)
Proximal and
distal stem
59 Femoral stem Five dial gauges Three translations proximal;
two translations distal
(no axial)
Prosthetic head;
proximal and distal
stem

0266/ ch33/frame Page 518 Friday, October 22, 1999 10:34 AM

In Vitro

Measurements of Implant Stability

519

42 Femoral stems Six eddy current transducers Three translations and two
rotations (transverse and
frontal)
Proximal stem
60 Femoral stem Seven LVDTs Three translations distal; two
translations proximal (no
axial); transverse rotation
Proximal and
distal stem
61 Femoral stem Six LVDTs Three translations proximal;
two translations distal;
transverse rotation

b

Proximal and distal
stem; prosthetic neck
9 Femoral stems Two linear resistive
transducers, LVDT; rotary
potentiometer; four strain
gauge-based interface
transducers
Three translations and
transverse rotation of the
distal femur; interface motion
in two planes
(sagittal and frontal)
Distal femur; proximal
and distal stem
(interface motion)
40 Femoral stems Six LVDTs Three translations and three
rotations
Distal stem
3 Femoral stems Six LVDTs Three translations and three
rotations
Proximal and distal
stem
33 Femoral stem Seven LVDTs Three translations and three
rotations
Proximal and distal
17 Femoral stem
(goat model)
RSA Three translations; three
rotations
Femoral stem
62 Tibial component Three LVDTs Axial translation Anterior, medial, and
posterior aspects
63 Tibial component Four liquid metal strain
gauges
Axial translation Anterolateral,
anteromedial,
posterolateral, and
posteromedial
aspects
64 Simulated tibial
component
Four strain gauge
transducers
Axial translation Anterolateral,
anteromedial,
posterolateral, and
posteromedial
aspects
65 Tibial component Four LVDTs Axial translation Anterior, posterior,
medial, and
lateral rims
20 Tibial component Video camera via a
stereomicroscope
Interface motion Implantbone
interface
66 Tibial component Three LVDTs Two translations (axial in two
locations and medial-lateral)
Lateral plateau,
medial rim
23 Tibial component LVDTs Axial translation; posterior-
anterior translation
(separate tests)
Anterolateral and
posteromedial
plateau, medial and
lateral rims
45 Tibial component LVDTs Axial translation; posterior-
anterior translation; transverse
rotation

a

(separate tests)
Lateral plateau, medial
and lateral rims

TABLE 33.1 (continued)
Summary of Implants Tested, Sensors Used, Number of Motion Components Determined
and the Location of the Measurement Sites

Ref. Implants Tested Sensors Used
Number of Motion
Components Determined
Location of
Measurement Sites

0266/ ch33/frame Page 519 Friday, October 22, 1999 10:34 AM

520

Mechanical Testing of Bone and the BoneImplant Interface

if it were important to have a map of the motions everywhere at the interface, then a nite-element
model would be needed,

5,6

with the actual

in vitro

measurements serving as a validation of the
model. If one is interested in micromotion at only a few relevant locations of the boneimplant
interface, it is possible to use custom-built sensors specically designed to measure three-
dimensional micromotion at the point of their location.

7-9

In the presence of a full six-degree-of-freedom motion data set and a model of the geometry
of both the implant and bone, it is possible to calculate the exact amount of displacement for

22 Tibial component Two dial gauges; four
LVDTs
Posterior-anterior translation;
axial translation
Medial and lateral
rims, anterior, lateral,
posterior,
posteromedial points
4 Canine tibial
component
Six LVDTs Three translations, three
rotations
Central aspect of the
component
10 Tibial component Seven LVDTs Three translations, three
rotations
Medial and lateral
rims
44 Acetabular
component
Load cell Force required to loosen
the component
11 Acetabular
component
Preset angular velocity in a
displacement- controlled
test
Twist angle to loosen
the component
Actuator (estimated)
13 Acetabular
component
Rotational transducer

b

Polar rotation (displacement
controlled loading)
Actuator
67 Acetabular
component
Two extensometers Interface motion at two
locations
Dome of the cup,
posterior wall
68 Acetabular
component
Strain gauge displacement
transducer
Translation within the
equatorial plane (superior-
lateral)
Superior-lateral edge
of the cup
69 Acetabular
component
Three LVDTs Interface motion within
equatorial plane
Adjacent to
implantbone
interface
70 Acetabular
component
Three eddy current
transducers
Axial translation At the cup edge
71 Acetabular
component
Six LVDTs Three translations (resultant
translation reported)
Anteromedial and
posterolateral rim
19 Acetabular
component
(goat model)
RSA Three translations;
three rotations
Acetabular component

Note:

The table is organized from simple systems (e.g., determination of one motion component) to complex (determination
of all six motion components).

a

Indicates that rotations were reported as projected linear displacements in

m within arbitrary distance from the axis of
rotation.

b

Indicates that the transducer is a part of the materials testing machine and senses the displacement of the actuator or the
crosshead.
LVDT = linear variable displacement transducer; RSA = roentgen stereophotogrammetric analysis.

TABLE 33.1 (continued)
Summary of Implants Tested, Sensors Used, Number of Motion Components Determined
and the Location of the Measurement Sites

Ref. Implants Tested Sensors Used
Number of Motion
Components Determined
Location of
Measurement Sites

0266/ ch33/frame Page 520 Friday, October 22, 1999 10:34 AM

In Vitro

Measurements of Implant Stability

521

any specic point of the boneimplant interface (keeping in mind the limitation of the rigid-
body assumption, discussed above).

3,4,10

Thus, maps of each of the components of motion or of
various combinations can be displayed to help interpret how the component moves with respect
to the bone (see, e.g., Reference 4). In addition, the number of motion descriptors can be reduced
to a single measure of interface motion (e.g., the resultant translation) that can then be used
for interpretation.
Between the one-motion and complete six-degree-of-freedom systems, there are a variety of
motion-sensing setups that provide a subset of motions (see Table 33.1). Most of these include at
least an axial displacement and/or rotational motion around the longitudinal axis for hip femoral
components and subsidence and lift-off for knee tibial components.
In addition to motions determined from static testing (e.g., the application of one or a very
limited number of load cycles), recoverable and nonrecoverable motion can be characterized from
cyclic tests. The nonrecoverable motion should be analogous to migration measured

in vivo

by
RSA. Thus, some authors report both the recoverable (inducible) motion and the nonrecoverable
(migratory) motion.
While most studies use the implantbone displacement to describe implant stability, there
are a few studies where the ultimate force required to loosen the implant has been used to
characterize implant stability.

11-15

In addition, imprints on a precalibrated pressure-sensitive lm
have been used to describe a stability of tibial components of total knee replacements,

16

RSA
has been used to evaluate acetabular component and femoral stem stability

ex vivo

in a goat
model

17-19

and direct visualization by a video camera was used to document boneimplant
interface movement in a canine tibial hemiarthroplasty model

20

and in femoral stems implanted
in cadaveric human bones.

21

III. STABILITY MEASURING SYSTEMS AND THE LOADING SETUPS

The value of results from

in vitro

stability tests depends on the representativeness of the loading
protocol with respect to the known magnitude and direction of joint reaction forces measured

in
vivo

. It is generally agreed that using simplied loading protocols with high forces acting in a
nonphysiological direction can result in misleading test results. As an example, testing prosthetic
components by application of an isolated load in the transverse direction without a normal axial
component of load will almost certainly lead to the failure of implant xation

22-24

in a fashion which
is not comparable to loosening mechanisms observed

in vivo

.
Ideally, resultant joint reaction forces are preferred to isolated components of load. On the one
hand, the reports on force measurements obtained from telemetrized hip prostheses provide an
excellent database for hip loading conditions in humans,

25-31

canines,

26,32

and sheep.

26

On the other
hand, attempts to replicate the exact

in vivo

loading conditions

in vitro

can lead to the introduction
of nonphysiological constraints and excessive stresses imposed by the test setup. Because the loads
affecting the hip joint are fairly well understood and have received the most attention of those
investigating implant stability, the trade-offs involved, using the hip as an example, will be illus-
trated. This discussion is followed by much briefer descriptions of loading of the acetabular
components and the tibial component.
In a simple setup for testing the stability of the femoral stem in hip replacement, the distal
femur typically is rigidly xed to maintain the shaft angle against the test loads simulating the
stance phase of level walking or stair climbing.

3,17,18,21,33,34

By using this type of loading setup,
despite the perfect projection of the resultant force against the proximal joint surface of the implant,
the whole femoral shaft is subjected to bending and torsion due to its distal constraint.
So, even though the loads applied to the head of the implant may be physiological, the moments
created by the joint forces are always balanced by the activity of the muscles crossing the hip and
knee joints

in vivo

and not by rigid xation of the distal femur as is often the case

in vitro

. Therefore,
if such a loading protocol is used to test the implantbone displacement one has to recognize the

0266/ ch33/frame Page 521 Friday, October 22, 1999 10:34 AM

522

Mechanical Testing of Bone and the BoneImplant Interface

potential bias due to excess deformation of the proximal femur. With these setups, it is not
uncommon for the specimen to fail (i.e., the femur to break) as a result of stress concentration at
the distal xity.
More accurate reproductions of hip loading conditions can be achieved by simulation of the
abductors and adding one degree of rotational freedom (hinge) to the distal femur for level
walking.

35-38

Hip loading conditions for fast walking have been replicated using two actuators and
a system of cables simulating the action of three muscles.

39

Stair climbing can be represented by
a combination simultaneous axial and torsion load

36

or by simulation of the abductors and extensors
in stair climbing

35,40

(Figure 33.3). Similar setups with added support to the proximal femur can
be used to reach realistic peak joint reaction forces without the risk of losing the specimen due to
femoral fracture in bending. Three muscle groups have been used to simulate femoral component
loading conditions in the canine.

41

However, these more sophisticated loading regimes have draw-
backs simulating muscle forces through the application of nylon straps

36,42

or steel cables

40

attached to the proximal femur may lead to nonphysiological local strain concentrations in the bone.
The next logical step toward more accurate loading simulation would be a complete release of
the distal femur (except for the projected ground reaction force) in the presence of perfectly balanced
simulation of all major muscle groups. This type of setup has been used to investigate the bio-
mechanical effects of relocation of the hip center,

43

but has not been used in conjunction with
implant stability experiments. Technically, it would be difcult to have the complex instrumentation
needed for both.
Acetabular components of hip replacement are typically subject to either oblique axial loads
or to torque tests.

44

Since the axial load is estimated to be the predominant component of the
resultant force within the knee joint, most tibial components are tested by eccentric application of
an axial load, with only a few studies adding torsion or a transverse load component.

45

FIGURE 33.3

Schematic of loading setups to simulate stair climbing originally published by Burke et al.

35

(A) and Berzins et al.

40

(B). Note for (B), 1 = load cell, 2 = loading plate, 3 = gluteus medius cable, 4 =
gluteus maximus cable, and 5 = the distal support (which allowed rotation). ((A) from Burke, D.W. et al.,

J. Bone Joint Surg.,

73B, 33, 1991. With permission. (B) from Berzins, A., et al.,

J. Arthroplasty,

11, 500,
1996. With permission.)

0266/ ch33/frame Page 522 Friday, October 22, 1999 10:34 AM

In Vitro

Measurements of Implant Stability

523

There certainly is no single perfect loading apparatus for a given joint. After all, an implant
tested in a cadaver bone is just a model. So, it is up to the researcher to understand the inherent
limitations and to design the experiment and the experimental set up to permit specic questions
to be addressed.
Another issue that merits consideration is the number of load cycles, frequency, and rate of
load application. In the studies concerned with recoverable motions of the implant against the bone,
it is important to apply a sufcient number of cycles to accomplish the initial settling and to record
the motion after it has become repeatable.

4

If the study is more concerned with migration (non-
recoverable motion) between the implant and bone, a larger number of load cycles should be used
to simulate the shift in implant position after hundreds or even thousands of

in vivo

load cycles.
Loading frequency should always be representative of physiological loading rates. In studies where
a large number of loading cycles are applied and time constraints call for an accelerated test protocol,
the higher test speeds involving dynamic loading conditions (inertial forces, etc.) should be avoided.

IV. CONCLUSION

The authors would like to emphasize that

in vitro

experiments of implant stability fall at the
crossroads of two complex issues deciding how to make the stability measurements and deciding
how to apply the loads to the system. In terms of implant motion, one can measure from one to
six components of motion. In terms of loading, one can apply a simple unidirectional load or
attempt more physiological loading. The complexity of the experimental setup increases signi-
cantly, as the experiment becomes more realistic. It is always important to keep in mind that the
research question being asked determines how complex the experimental setup needs to be.

ACKNOWLEDGMENTS

Research was supported by NIH Grants AR16485 and AR42862.

REFERENCES

1. Charnley, J. and Kettlewell, J., The elimination of slip between prosthesis and femur,

J. Bone Joint
Surg.

, 47B, 56, 1965.
2. Nilsson, K.G. and Karrholm, J., RSA in the assessment of aseptic loosening,

J. Bone Joint Surg.

,
78B, 1, 1997.
3. Berzins, A., Sumner, D.R., Andriacchi, T.P., and Galante, J.O., Stem curvature and load angle inuence
the initial relative boneimplant motion of cementless femoral stems,

J. Orthop. Res.

, 11, 758, 1993.
4. Sumner, D.R., Berzins, A., Turner, T.M., et al., Initial

in vitro

stability of the tibial component in a
canine model of cementless total knee replacement,

J. Biomech.

, 27, 929, 1994.
5. Keaveny, T.M. and Bartel, D.L., Effects of porous coating, with and without collar support, on early
relative motion for a cementless hip prosthesis,

J. Biomech.

, 26, 1355, 1993.
6. Weinans, H., Huiskes, R., and Grootenboer, H.J., Quantitative analysis of bone reactions to relative
motions at implantbone interfaces,

J. Biomech.

, 26, 1271, 1993.
7. Buhler, D.W., Oxland, T.R., and Nolte, L.P., Design and evaluation of a device for measuring three-
dimensional micromotions of press-t femoral stem prostheses,

Med. Eng. Phys.

, 19, 187, 1997.
8. Schneider, E., Schonenberger, U., Giraud, P., and Burgi, M., [Initial stability and pelvic deformation
in cemented and non-cemented acetabular components],

Orthopade,

21, 57, 1992 [in German].
9. Schneider, E., Eulenberger, J., Steiner, W., et al., Experimental method for the

in vitro

testing of the
initial stability of cementless hip prostheses,

J. Biomech.

, 22, 735, 1989.
10. Stern, S. H., Wills, R. D., and Gilbert, J. L., The effect of tibial stem design on component micromotion
in knee arthroplasty,

Clin. Orthop.

, 345, 44, 1997.
11. Ohlin, A. and Balkfors, B., Stability of cemented sockets after 314 years,

J. Arthroplasty

, 7, 87, 1992.

0266/ ch33/frame Page 523 Friday, October 22, 1999 10:34 AM

524

Mechanical Testing of Bone and the BoneImplant Interface

12. Jansson, V., Zimmer, M., Kuhne, J.H., and Sailer, F.P. [Initial stability of an implanted cement-canal
prosthesis. Results in experimental studies on human cadaver femurs],

Z. Orthop.,

131, 377, 1993 [in
German].
13. Curtis, M.J., Jinnah, R.H., Wilson, V.D., and Hungerford, D.S., The initial stability of uncemented
acetabular components,

J. Bone Joint Surg.

,



74B, 372, 1992.
14. Whiteside, L.A., Amador, D., and Russell, K., The effects of the collar on total hip femoral component
subsidence,

Clin. Orthop.

, 231, 120, 1988.
15. Kendrick, J.B., Noble, P.C., and Tullos, H.S., Distal stem design and the torsional stability of cement-
less femoral stems,

J. Arthroplasty

, 10, 463, 1995.
16. Walker, P.S., Hsu, H.P., and Zimmerman, R.A., A comparative study of uncemented tibial components,

J. Arthroplasty

, 5, 245, 1990.
17. Schreurs, B.W., Huiskes, R., Buma, P., and Sloof, T. J., Biomechanical and histological evaluation of
a hydroxyapatite-coated titanium femoral stem xed with an intramedullary morselized bone grafting
technique: an animal experiment on goats,

Biomaterials

, 17, 1177, 1996.
18. Schreurs, B.W., Buma, P., Huiskes, R., et al., Morselized allografts for xation of the hip prosthesis
femoral component,

Acta Orthop. Scand.

, 65, 267, 1994.
19. Schimmel, J.W., Buma, P., Versleyen, D., et al., Acetabular reconstruction with impacted morselized
cancellous allografts in cemented hip arthroplasty,

J. Arthroplasty

, 13, 438, 1998.
20. Stulberg, B.N., Watson, J.T., Stulberg, S.D., et al., A new model to assess tibial xation. II. Concurrent
histologic and biomechanical observations,

Clin. Orthop.

, 263, 303, 1991.
21. Lee, T.Q., Lewis, D.A., Barnett, S.L., et al., Structural integrity of implanted WMT innity hip and
osteonics omnit hip in fresh cadaveric femurs, J. Biomed. Mater. Res., 39, 516, 1998.
22. Miura, H., Whiteside, L.A., Easley, J.C., and Amador, D.D., Effects of screws and a sleeve on initial
xation in uncemented total knee tibial components, Clin. Orthop., 259, 160, 1990.
23. Yoshii, I., Whiteside, L.A., Milliano, M.T., and White, S.E., The effect of central stem and stem length
on micromovement of the tibial tray, J. Arthroplasty, 7, 433, 1992.
24. Phillips, T.W., Nguyen, L.T., and Munro, S.D., Loosening of cementless femoral stems: a biomechan-
ical analysis of immediate xation with loading vertical, femur horizontal, J. Biomech., 24, 37, 1991.
25. Bergmann, G., Graichen, F., Rohlmann, A., and Linke, H., Hip joint forces during load carrying, Clin.
Orthop., 335, 190, 1997.
26. Bergmann, G., Siraky, J., Rohlmann, A., and Koelbel, R., A comparison of hip joint forces in sheep,
dog and man, J. Biomech., 17, 907, 1984.
27. Bergmann, G., Graichen, F., and Rohlmann, A., Hip joint loading during walking and running,
measured in two patients, J. Biomech., 26, 969, 1993.
28. Bergmann, G., Graichen, F. and Rohlmann, A., Is staircase walking a risk for the xation of hip
implants? J. Biomech., 28, 535, 1995.
29. Davy, D.T., Kotzar, G.M., Brown, R.H., et al., Telemetric force measurements across the hip after
total arthroplasty, J. Bone Joint Surg., 70A, 45, 1988.
30. Kotzar, G.M., Davy, D.T., Berilla, J., and Goldberg, V.M., Torsional loads in the early postoperative
period following total hip replacement, J. Orthop. Res., 13, 945, 1995.
31. Kotzar, G.M., Davy, D.T., Goldberg, V.M., et al., Telemeterized in vivo hip joint force data: a report
on two patients after total hip surgery, J. Orthop. Res., 9, 621, 1991.
32. Page, A.E., Allan, C., Jasty, M., et al., Determination of loading parameters in the canine hip in vivo,
J. Biomech., 26, 571, 1993.
33. Gilbert, J.L., Bloomfeld, R.S., Lautenschlager, E.P., and Wixson, R. L., A computer-based bio-
mechanical analysis of the three-dimensional motion of cementless hip prostheses, J. Biomech., 25,
329, 1992.
34. McKellop, H., Ebramzadeh, E., Niederer, P.G., and Sarmiento, A., Comparison of the stability of
press-t hip prosthesis femoral stems using a synthetic model femur, J. Orthop. Res., 9, 297, 1991.
35. Burke, D.W., OConnor, D.O., Zalenski, E.B., et al., Micromotion of cemented and uncemented
femoral components, J. Bone Joint Surg., 73B, 33, 1991.
36. Callaghan, J.J., Fulghum, C.S., Glisson, R.R., and Stranne, S.K., The effect of femoral stem geometry
on interface motion in uncemented porous-coated total hip prostheses: comparison of straight-stem
and curved-stem designs, J. Bone Joint Surg., 74-A, 839, 1992.
0266/ ch33/frame Page 524 Friday, October 22, 1999 10:34 AM
In Vitro Measurements of Implant Stability 525
37. Maloney, W.J., Jasty, M., Burke, D.W., et al., Biomechanical and histologic investigation of cemented
total hip arthroplasties, Clin. Orthop., 249, 129, 1989.
38. Finkelstein, J.A., Anderson, G.I., Waddell, J.P., et al., A study of micromotion and appositional bone
growth to a canine madreporic-surfaced femoral component, J. Arthroplasty, 9, 317, 1994.
39. Munting, E. and Verhelpen, M., Fixation and effect on bone strain pattern of a stemless hip prosthesis,
J. Biomech., 28, 949, 1995.
40. Berzins, A., Sumner, D.R., Wasielewski, R.C., and Galante, J.O., Impacted particulate allograft for
femoral revision total hip arthroplasty: in vitro mechanical stability and effects of cement pressuriza-
tion, J. Arthroplasty, 11, 500, 1996.
41. Jasty, M., Krushell, R.J., Zalenski, E., et al., The contribution of the nonporous distal stem to the
stability of proximally porous-coated canine femoral components, J. Arthroplasty, 8, 33, 1993.
42. Walker, P.S., Schneeweis, D., Murphy, S., and Nelson, P., Strains and micromotion of press-t femoral
stem prostheses, J. Biomech., 20, 693, 1987.
43. Doehring, T.C., Rubash, H.E., Shelley, F.J., et al., Effect of superior and superolateral relocations of
the hip center on hip joint forces. An experimental and analytical analysis, J. Arthroplasty, 11, 693,
1996.
44. Volz, R.G. and Wilson, R.J., Factors affecting the mechanical stability of the cemented acetabular
component in total hip replacement, J. Bone Joint Surg., 59, 501, 1977.
45. Kraemer, W.J., Harrington, I.J., and Hearn, T.C., Micromotion secondary to axial, torsional, and shear
loads in two models of cementless tibial components, J. Arthroplasty, 10, 227, 1995.
46. Markolf, K.L., Amstutz, H.C., and Hirschowitz, D.L., The effect of calcar contact on femoral com-
ponent micromovement, J. Bone Joint Surg., 62A, 1315, 1980.
47. Lee, R.W., Volz, R.G., and Sheridan, D.C., The role of xation and bone quality on the mechanical
stability of tibial knee components, Clin. Orthop., 273, 177, 1991.
48. Sugiyama, H., Whiteside, L.A., Engh, C.A., and Otani, T., Late mechanical stability of the proximal
coated AML prosthesis, Orthopedics, 17, 583, 1994.
49. Sugiyama, H., Whiteside, L.A., and Kaiser, A.D., Examination of rotational xation of the femoral
component in total hip arthroplasty. A mechanical study of micromovement and acoustic emission,
Clin. Orthop., 249, 122, 1989.
50. Sugiyama, H., Whiteside, L.A., and Engh, C.A., Torsional xation of the femoral component in total
hip arthroplasty, Clin. Orthop., 275, 187, 1992.
51. Ohl, M.D., Whiteside, L.A., McCarthy, D.S., and White, S.E., Torsional xation of a modular femoral
hip component, Clin. Orthop., 287, 135, 1993.
52. Otani, T., Whiteside, L.A., White, S.E., and McCarthy, D.S., Reaming technique of the femoral
diaphysis in cementless total hip arthroplasty, Clin. Orthop., 311, 210, 1995.
53. Channer, M.A., Glisson, R.R., Seaber, A.V., and Vail, T.P., Use of bone compaction in total knee
arthroplasty, J. Arthroplasty, 11, 743, 1996.
54. Mahomed, N., Schatzker, J., and Hearn, T., Biomechanical analysis of a distally interlocked press-t
femoral total hip prosthesis, J. Arthroplasty, 8, 129, 1993.
55. Manley, P.A., Vanderby, R., Kohles, S., et al., Alterations in femoral strain, micromotion, cortical
geometry, cortical porosity, and bony ingrowth in uncemented collared and collarless prostheses in
the dog, J. Arthroplasty, 10, 63, 1995.
56. Fischer, K.J., Carter, D.R., and Maloney, W.J., In vitro study of initial stability of a conical collared
femoral component, J. Arthroplasty, 7 (Suppl.), 389, 1992.
57. Maloney, W.J., Jasty, M., Burke, D.W., et al., Biomechanical and histologic investigation of cemented
total hip arthroplasties. A study of autopsy-retrieved femurs after in vivo cycling, Clin. Orthop., 249,
129, 1989.
58. Hua, J. and Walker, P.S., Relative motion of hip stems under load. An in vitro study of symmetrical,
asymmetrical, and custom asymmetrical designs, J. Bone Joint Surg., 76A, 95, 1994.
59. Whiteside, L.A. and Easley, J.C., The effect of collar and distal stem xation on micromotion of the
femoral stem in uncemented total hip arthroplasty, Clin. Orthop., 145, 1989.
60. Otani, T., Whiteside, L.A., White, S.E., and McCarthy, D.S., Effects of femoral component material
properties on cementless xation in total hip arthroplasty, J. Arthroplasty, 8, 67, 1993.
0266/ ch33/frame Page 525 Friday, October 22, 1999 10:34 AM
526 Mechanical Testing of Bone and the BoneImplant Interface
61. Whiteside, L.A., White, S.E., Engh, C.A., and Head, W., Mechanical evaluation of cadaver retrieval
specimens of cementless bone-ingrown total hip arthroplasty femoral components, J. Arthroplasty, 8,
147, 1993.
62. Kaiser, A.D. and Whiteside, L.A., The effect of screws and pegs on the initial xation stability of an
uncemented unicondylar knee replacement, Clin. Orthop., 259, 169, 1990.
63. Branson, P.J., Steege, J.W., Wixson, R.L., et al., Rigidity of initial xation with uncemented tibial
knee implants, J. Arthroplasty, 4, 21, 1989.
64. Wyatt, R.W., Alpert, J.P., Daniels, A.U., et al., The effect of screw xation on initial rigidity of tibial
knee components, J. Appl. Biomater., 2, 109, 1991.
65. Shimagaki, H., Bechtold, J.E., Sherman, R.E., and Gustilo, R.B., Stability of initial xation of the
tibial component in cementless total knee arthroplasty, J. Orthop. Res., 8, 64, 1990.
66. Volz, R.G., Nisbet, J.K., Russel, W.L., and McMurtry, M.G., The mechanical stability of various
noncemented tibial components, Clin. Orthop., 226, 38, 1988.
67. Kwong, L.M., OConnor, D.O., Sedlacek, R.C., et al., A quantitative in vitro assessment of t and
screw xation on the stability of a cementless hemispherical acetabular component, J. Arthroplasty,
9, 163, 1994.
68. Lachiewicz, P.F., Suh, P.B., and Gilbert, J.A., In vitro initial xation of porous-coated acetabular total
hip components. A biomechanical comparative study, J. Arthroplasty, 4, 201, 1989.
69. Won, C.H., Hearn, T.C., and Tile, M., Micromotion of cementless hemispherical acetabular compo-
nents. Does press-t need adjunctive screw xation? J. Bone Joint Surg., 77B, 484, 1995.
70. Perona, P.G., Lawrence, J., Paprosky, W.G., et al., Acetabular micromotion as a measure of initial
implant stability in primary hip arthroplasty, J. Arthroplasty, 7, 537, 1992.
71. Stiehl, J.B., MacMillan, E., and Skrade, D.A., Mechanical stability of porous-coated acetabular
components in total hip arthroplasty, J. Arthroplasty, 6, 295, 1991.
0266/ ch33/frame Page 526 Friday, October 22, 1999 10:34 AM

527

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

34

In Vitro

Testing of the Stability
of Acetabular Components

James R. Davis, Robert A. Lofthouse, and Riyaz H. Jinnah

CONTENTS

I. Introduction. .........................................................................................................................527
II. Various Methods and Parameters in the Literature .............................................................530
A. Torsional Stability ..........................................................................................................530
B. Offset Cup Loading........................................................................................................531
C. Axial Loading in a Superioposteriomedial Direction 30

to the Plane
of Cup Rim, Adduction Torsion.....................................................................................532
D. Axial Load Orientated at 12

of Flexion and 21

of Adduction...................................533
E. Axial Load at 10

of Flexion and 15

of Adduction.....................................................534
F. Axial Load with the Pelvis Orientated in the Anatomical Position..............................535
G. Repetitive 100,000 Cycle Axial Loading in Full Extension, Load to Failure
in Same Orientation .......................................................................................................536
H. Finite-Element Analysis to Calculate Acetabular Deformation ....................................537
I. Pullout Force ..................................................................................................................537
III. Conclusion............................................................................................................................538
References ......................................................................................................................................538

I. INTRODUCTION

Measurement of acetabular stability is difcult to perform

in vivo

. Every day in the clinical situation,
surgeons make a judgment on whether or not an acetabular component is stable. If there is doubt,
a number of dome screws may be used in conjunction with the cup. The decision when to insert
screws is arbitrary. If asked closely about screw insertion and the rationale behind it, many surgeons
will answer that if it feels stable on insertion they will not use screws, but if the position of the
cup has to be changed due to initial malpositioning, then screws will be inserted. Many will always
use screws and others will select a prosthesis with ns, pegs, spikes, or altered geometries in an
attempt to gain increased initial stability. The use of screws in the revision setting is more prevalent,
especially when using modied implants such as the oblong or bilobed cups.

1

The use of cups with dome screw holes has been of concern. Polyethylene will creep through
these holes when repetitively loaded and generate wear debris which may induce pelvic osteolysis
and lead to premature failure.

2

If correctly placed, the use of screws will increase the initial stability,
but they may loosen causing extensive cavitary acetabular defects which can result in a more
difcult revision procedure. Screw placement is also not without immediate hazards; in the litera-
ture, there are reported cases of vital vessel or organ damage.

3,4

Concerns also arise with respect
to screw fretting at the prosthetic interface as this may cause debris generation and promote a
histological response resulting in component loosening.

5

0266/ ch34/frame Page 527 Friday, October 22, 1999 10:36 AM

528

Mechanical Testing of Bone and the BoneImplant Interface

Initial acetabular prosthetic stability is important for optimal biological xation.

6,7

Figures for
the level of micromotion at the cupbone interface preventing osseointegration and predisposing
to brous xation vary between 40

m and 1 mm.

8-10

Gaps between the prosthesis and bone are
theoretically not so problematic; in a canine model with a stable implantbone environment, gaps
of 1.5 mm have been bridged by new bone.

11

The correlation between the canine model and humans
is, however, questionable with respect to predicting the osseous response to implants.

12

Stability from bony ingrowth occurs by the apposition and generation of new bone, which
incorporates into pores or surface defects on the acetabular cup surface. The most important factor
for providing optimum conditions for this to occur is stability at the boneprosthesis interface.

7,10,13

This, in conjunction with pore sizes of 200 to 400

m, availability of cancellous bone, and close
proximity of the implant to bone provide the most likely environment for successful integration.

7,13-15

Retrieval studies of porous-surfaced implants have conrmed this.

15-19

Prosthetic designs that pre-
load the bone in compression and eliminate tensile and shear loading at the interface are likely to
provide the best initial xation and the correct conditions for bony ingrowth.

20

In animal studies,
weight-bearing stresses are important for the stimulation of bony ingrowth into acetabular implants.
There is an optimal conguration for these force vectors and, if imbalance exists, they can lead to
abnormal micromotion and disruption of early trabecular ingrowth.

21

Other factors inuencing the
process of ingrowth include porous surface geometry, hydroxyapatite coating, accuracy of press-
t, and implant design.

13,22-24

The difference in the modulus of elasticity of the acetabular prosthesis and bone at the
boneprosthetic interface will lead to differential strains across the area of the cup. Most uncemented
acetabular prostheses are signicantly stiffer than the surrounding bone. It has been shown that on
axial loading of a press-t acetabular component without augmentation, more motion is consistently
detected at the ilium than the pubis or ischium, respectively. This suggests that the ilium is the
least supportive structure contributing to initial stability. Furthermore, the ilium bears the greatest
load across the hip joint and gains the majority of its load-bearing capabilities from an intact
subchondral bone plate.

25

This reinforces the importance of retaining the subchondral plate when
implanting an acetabular component.
Direct measurement of the true motion at the interface is difcult. Most studies rely on detecting
movement in a rigid cupjig construct and assume this reects the movement at the interface.

9,21,26,27

To measure the motion between cup and subchondral bone the strain gauge would need to be
applied directly at the interface. This is not possible, so the closest is to measure the motion between
the cup and the bone just behind the subchondral plate. Many authors thus incorporate in their
results the deformation at three sites: compression or torque in the prosthesis itself, deformation
in cancellous bone, and the true movement at the boneprosthesis interface. Consideration needs
to be given in respect to the changing Youngs modulus of cancellous bone; as it is crushed, the
lattice structure becomes more compact and therefore stiffer. This deformation may be plastic and
not recoverable. If more than one load is applied to the cancellous bone, the subsequent results
will include less displacement at this level.

28

There are reported differences between embalmed and fresh acetabula and their behavior under
load. Embalmed specimens display changes in mechanical properties; commonly, the modulus of
elasticity of cortical bone is increased and its compressive strength diminished.

21,29

The bone mineral
density becomes an important factor when comparisons between pelvises are made. This requires
standardization as the extrapolation of stability in pelvises with different densities may not be linear.
There are also inherent variations between cadaveric pelvises. To some extent this has been addressed
by comparing different implants in the two acetabula of one pelvis (matched pairs). However, this does
not address the effect of macroscopic or microscopic individual local acetabular defects.

30

To overcome the variables in cadaveric bones, comparisons between prostheses can be made
by standardizing the material into which the prosthesis is implanted. Many studies have used foam
blocks into which the acetabular prosthesis is seated. Comparisons can be made by either varying

0266/ ch34/frame Page 528 Friday, October 22, 1999 10:36 AM

In Vitro

Testing of the Stability of Acetabular Components

529

the implant diameters with the same acetabular prosthesis or maintaining constant cup diameters
and comparing different cup designs.

31

One important comparative parameter that requires consistency is the external diameter of the
acetabular components. As the radius becomes greater, the surface area and the linear rim length
in contact with bone increase, and this may inuence the initial stability, particularly in torsion.

30

To compare a cohort of the same prosthesis, the cup size requires standardization. Problems can
arise with preparation of the acetabulum. It is known that acetabula reamed to cancellous bone
offer less support to the uncemented press-t cup than those where the subchondral bone is left
intact.

32

Choice of acetabula to test becomes critical, the demographics of normal acetabular size
show a large variance and a certain number of pelvises will be unsuitable because the acetabular
preparation to accept the same diameter of cup across the experimental cohort will be different.
Additionally, those with similar diameters may show disparities in the amount and thickness of the
subchondral bone plate remaining after preparation.
Implantation technique is important; the method requires consistency and repeatability and
preferably should be performed by the same surgeon.

32,33

The acetabula require reaming accu-
rately to remove remaining hyaline cartilage but to leave the subchondral bone plate, particularly
in the acetabular dome, relatively undisturbed. The dome in conjunction with the posterior
column offers extremely important means of acetabular stability. The acetabulum acts as a
dynamic structure in obtaining stability, which differs from that of a xed-ring construct. Under
load, the dome is compressed while the columns are pushed apart by tensile stresses.

34

Implan-
tation of the cup is critical as bone has inherent viscoelasticity, meaning that the stiffness of
bone varies depending on the rate of loading. The faster the load rate, the stiffer and thus more
brittle the bone. If a cup is inserted with a heavy blow, macro- or microfractures may be created,
weakening the cup bone construct. Conversely, if the cup is not seated properly, the interface
may not be optimized.
Testing of the implantbone construct requires isolating the acetabulum and holding it in such
a way that the inherent elasticity of the pelvis is minimized and that the deformation detected will
be at the bone-prosthesis interface. This requires mounting, usually in resin, gypsum, or low-
melting-point metal. The orientation for each specimen needs to be consistent to eliminate errors.
The implant must be tested under the same conditions in a standard manner and determination of
the parameters measured must be established prior to the test to prevent cupbone interface
disturbance and, thus, inaccurate data acquisition.
Guidelines for the amount of torque force experienced at the cupbone interface vary
depending on bearing surface, femoral head size, cup radius, and individual patient characteris-
tics.

9

Torsional moments of 6.8 Nm are not uncommon for metal-on-metal bearing surfaces,
with polyethylenemetal forces about half of that value. These forces are thought to be the most
damaging to the boneimplant interface; their values increase as the static load across the
acetabulum increases.

35,36

Compressive forces between prosthetic femoral head and cup have
been found to be as high as 4.3 times body weight in the middle-aged man.

36

The strains imparted
on the prosthesis bone interface can be presumed to be signicantly higher in the young adult.

9

When using an instrumented total hip arthroplasty the forces experienced during passive move-
ments of physical therapy were as high as 50% of body weight, and during protected weight
bearing with two crutches forces between 85 and 200% of body weight were observed.

9

The
frictional moments experienced by the acetabular prosthesis have been reported to be between
0.4 to 10.8 Nm; these are the baselines on which many trials are founded.

1,37

The importance
of initial protected weight bearing is conrmed by studies showing that by increasing the applied
static load from 110 to 840 N the frictional torque forces increase by 2.5 to 3.5 times.

38

Increasing
the femoral head diameter has a limited effect on the frictional torque forces, progression from
22.5 to 50 mm elevates them by 1.5 times. These forces are magnied when there is no lubrication,
which is seen in prolonged weight bearing without movement.

36

0266/ ch34/frame Page 529 Friday, October 22, 1999 10:36 AM

530

Mechanical Testing of Bone and the BoneImplant Interface

II. VARIOUS METHODS AND PARAMETERS IN THE LITERATURE
A. T

ORSIONAL

S

TABILITY

A number of studies have measured this variable which was felt to be a reproducible means of
assessing acetabular component xation

in vitro

.

32,33,39

In one study 52-mm porous-coated cups
with two pegs were implanted in fresh frozen cadaveric acetabula which had been underreamed
by 1, 2, 3, and 4 mm. The cups were implanted at 45

of abduction and 10

of anteversion. Seating
of the cup was measured through the dome holes in the cup with a modied depth gauge. No
screws were used to augment the initial stability. The acetabula were mounted in resin and tested
on a material testing machine (MTS). A ten-sided foil, shaped to t the metal shells, was machined.
To determine the initial stability a compressive load of 686 N was applied perpendicular to the cup
rim to simulate loading comparable to a 70-kg human. A torque of 6.8 Nm (within physiological
limits) was applied to the cup as a triangular wave over 3 s for ve cycles. The angular deformation
was then measured. Second, a 100-Nm force was applied over a 0.5-s ramp and the moments
required to produce 1

, 2

, and 20

of rotation of the prosthesis in the bone were recorded.

26

Measurement of the displacement was achieved through the actuator attached to the MTS. This
assumes that the acetabulumpelvis-resin mount does not move in relation to the cup and that the
deformation is occurring only at the boneprosthesis interface.
A second study compared the porous-coated two-pegged cup to a porous-coated cup with screw
holes. The cups were implanted using the manufacturers guidelines into one pelvis. This was
repeated ve times in ve fresh-frozen cadaver pelvises. The authors used meticulous care to spare
the subchondral bone. A prosthesis of the same nal diameter as the nal reamer used was implanted.
Three screws were used to x the second cup. The same surgeon implanted all prostheses. The
hemipelvises were then embedded into epoxy resin blocks known to have negligible torsional
instability. A four-pronged removable jig was constructed from two orthogonally placed stainless
steel pins locked into a hemispherically surfaced aluminum block. The pins were placed in predrilled
slots on the implanted acetabular prostheses after mounting on the MTS (Figure 34.1).
The loading cycle consisted of an axial 686-N force directed medially with a 6.8-Nm moment
applied in both directions over 1.5 s for ve cycles. The resultant angular deformation was then
measured. The construct was then tested to failure over a 0.5-s period with 100 Nm moment.
Failure was dened as 20

of rotation of the prosthesis within the pelvis.

31

This was deter-
mined to be an excessive level and a failure limit of 2

was adopted which approximates to 1-mm
linear displacement.

9

Again the degree of movement that is taken to cause failure of optimum xation is too great,
as the displacement is being measured indirectly by assuming that the acetabulum remains stationary

FIGURE 34.1

Testing configuration for measurement of torsional stability.

0266/ ch34/frame Page 530 Friday, October 22, 1999 10:36 AM

In Vitro

Testing of the Stability of Acetabular Components

531

while the cup rotates within it. This is not measuring the motion at the interface but is giving a
relative value. If the specimens are mounted in the same way and the pelvises are all matched, then
the errors are limited. The study was performed with matched pairs standardizing the variables to
some degree. Direct medial axial compression does not simulate the direction of physiological load
bearing

in vivo

. There is no mention of the torsional stiffness of the jig used to impart the moment
to the cup or its t within the machined slots; these could be sources of error.
The third study of torsional stability investigated the stability of three component designs, one
with dome screws, one with three-spike xation, and the third with two-peg xation. Six of each were
implanted into 8 fresh-frozen and 12 embalmed cadaveric acetabula. All prostheses were implanted
in accordance with the manufacturers guidelines by one surgeon. An anteroposterial (AP) radiograph
of each pelvis was then taken and the acetabula removed from the pelvis, potted in acrylic resin and
placed on the testing frame. Each liner was drilled and two screws were used to secure the polyethylene
to the outer shell. The liner was then drilled and tapped to accept a torque wrench. The torque wrench
was instrumented with four strain gauges and secured to the liner with four screws. A 1.25-cm steel
rod was screwed to the torque wrench at its center of rotation and passed through a pillow block,
which was bolted to the test frame. Load was delivered at 4.06 cm/min over a 25

arc with a moment
arm of 25 cm. The load curve had reached a plateau by the end of each arc

28

(Figure 34.2).
One must not forget the stability issues of cemented prostheses whose problem is not so much
the initial but the long-term stability. However, comparative studies of torsion have been done
between autopsy retrieval specimens and cups cemented into cadaveric pelvises. In one such study
Charnley cups were cemented into 11 cadaveric acetabula. The acetabula were prepared with a
hemispherical reamer preserving the subchondral bone plate. Ten holes were then drilled with a
diameter of 4 and 10 mm deep and the cups were implanted after cement pressurization. Eight
holes were drilled in the periphery of the cup to accept a holder through which torque could be
imparted to the specimen. The acetabula were mounted in a cubiform mount. A torque force was
applied to the cup by means of an AC motor revolving at an angular velocity of 2.12

/s. The forces
were recorded a load cell. Instability was dened as the point at which the torquedisplacement
curve reached a plateau.

40

B. O

FFSET

C

UP

L

OADING

This form of loading is thought to be a more stringent test of initial xation than axial load or
torsion.

7

This mechanism tests the resistance of the interface to shear loads. In one study, metal-

FIGURE 34.2

Testing in torque. A strain gauge instrumented torque wrench is mounted to the acetabular
prosthesis. A 1.25-cm bolt was passed through the pillow block after attachment to the strain gauge. Load
was delivered to the torque wrench by the MTS.

0266/ ch34/frame Page 531 Friday, October 22, 1999 10:36 AM

532

Mechanical Testing of Bone and the BoneImplant Interface

backed acetabular cups were implanted into a dense, homogeneous, isotropic closed-cell polyeth-
ylene foam. The compressive modulus and Poissons ratio of this were within the range for
cancellous bone.

29,41-45

A 54-mm hemispherical socket was reamed and debris removed with com-
pressed air. A custom jig and centering hole were employed to assure consistency of implantation.
Five different cup designs were tested, three with external threads, one with four external ns, and
a press-t design. The nned cup was tested in two ways, loading over one n, then loading between
ns. The press-t cup was loaded in three ways; with no screws, with three dome screws, and
nally with three peripheral screws. For each trial a new polyethylene block was used.
The threaded cups were inserted using the rotary capabilities of the MTS with minimal axial
compression. The cups were inserted to a nal torque of 125 Nm. Press-t cups were slowly
lowered until fully seated.
The cups were then tested with an axial load applied to the periphery of the cup until shear
failure. A load of 0 to 2000 N applied over a 3-min period producing a torsional moment of 54 Nm
was used. Displacement was measured continuously at the point of load application by the MTS.
Each cup was tested in seven freshly prepared polyethylene blocks with a minimum of ve
repetitions of the loading cycle to each specimen.

31

In a second study six different 56-mm cups were seated in articial bone blocks according to
the manufacturers guidelines. A radiograph was then obtained to conrm seating of each component
and an offset load applied to the rim of the cup until failure.

46

Studies in prosthetic bones can only at best give relative values of stability. Screw placement
not only depends on the position within the cup but the quality of the surrounding bone. One can
standardize for position in this model but it is not applicable to the clinical situation. Displacement
measured at the point of load application will give very rough values of the actual interface
displacement as it is purely a two-dimensional measure. The prosthetic bone will, however, have
the same characteristics between samples. One concern is that the cups have different geometries
and will occupy a different volume in the same-sized block of prosthetic material, thus compro-
mising the standard behavior of the material. Important information can be gained by testing in
the prosthetic bone model; however, as already discussed, bone is a heterogeneous, anisotropic
material and behaves differently from homogeneous substitutes.

C. A

XIAL

L

OADING



IN



A

S

UPERIOPOSTERIOMEDIAL

D

IRECTION

30

TO



THE

P

LANE



OF

C

UP

R

IM

,
A

DDUCTION

T

ORSION

In this experiment, 12 fresh-frozen cadaver pelvises were obtained; each hemipelvis was mounted
in low-melting-point metal with the acetabula parallel to the horizontal plane. The acetabula were
then reamed to subchondral bone level, a trial implant was then implanted to assess t, and if the
t was not uniform the cavity was re-reamed. A component with a 2-mm peripheral are was
implanted. The size of the are was 2 mm greater than the last reamer used. Three titanium screws
were inserted in a triangular conguration either centrally or at the rim of the prosthesis. The rim
screws were cortical and the dome screws cancellous. The screws were tightened and, if any stripped,
a supplementary screw was repositioned and tightened with a torque wrench to 5 in.-lb. Six linear
variable differential transformers (LVDT) were attached between the bone and the rim of the
implant. The same individual placed each LVDT, in reference to anatomical landmarks. Each
transformer was capable of resolution to 1

m. The implants were tested in axial load to 500 N
and then 1000 N in a superioposteriomedial direction at 30

to the plane of the rim of the implant.
The axial load was applied through a 28-mm femoral head. The implant was then tested in adduction
torsion from 1.1 to 1.1 Nm and 22.2 to 22.2 Nm. One of two different acetabular defects was
then created; either 50% of the acetabular rim (peripheral) or 50% of the dome of the acetabulum
(central), using an osteotome and oscillating saw. The size of the defect was recorded by lining
the acetabulum with pressure-sensitive lm both before and after the defect had been created. Cups
were then implanted into the defective acetabula; the cups implanted into the acetabula with

0266/ ch34/frame Page 532 Friday, October 22, 1999 10:36 AM

In Vitro

Testing of the Stability of Acetabular Components

533

peripheral defects were secured with central screws and those with central defects with peripheral
screws. The acetabula were all rotated by 30

to allow the screws to gain purchase in fresh bone.
The specimens were then tested as before.
Each acetabulum was used for these experiments: rim and center screws in intact acetabulum,
either rim or central screws with a created defect. The order of screw placement, either central or
peripheral, was reversed in each specimen.
The effect of larger acetabular cups were then assessed by reaming the acetabulum to 66, 72,
and 78 mm. Bone apposition was determined with pressure-sensitive lm.
The bone mineral density was determined in all specimens.
In order to simplify data collection the data from the LVDT showing the greatest displacement
were analyzed. For comparison of torsion the most distal transducer was used and for axial load the
most proximal. The torque or axial load that produced 50

m of motion and the maximal implant
motion were used for comparison. The screw torques for each specimen were added together.

30

The preparation of test specimens in this study appears to be meticulous. The number of tests
performed on a single specimen is excessive. The axial load is smaller than that expected in clinical
practice. Torsion studies were performed after axial loading; it is unlikely that the values gained
in the torsional experiment would be similar if this part had been performed rst. It is difcult then
to create a defect in the acetabula that is consistent within the test group. Retesting the acetabula
with defects by rotating the cup through 30

to achieve screw purchase in fresh bone is difcult.
There is no way of knowing that the screw is not crossing a previous tract and thus will offer less
support than in fresh bone. Prior testing has also already compromised the remaining acetabular
surface and the values for stability are likely to be lower than those in a fresh specimen. This
method of testing, however, gives a more realistic picture of what a surgeon might face in the
clinical environment.

D. A

XIAL

L

OAD

O

RIENTATED



AT

12

OF

F

LEXION



AND

21

OF

A

DDUCTION

Eight fresh frozen and 12 embalmed cadaver pelvises were randomly distributed into three test
groups. Three acetabular prosthesis were tested; one with dome screws, one with three-spike
xation, and the third with two-peg xation. Six of each type of prosthesis were implanted according
to the manufacturers guidelines. The acetabula were removed and potted in acrylic resin. The
construct was then placed in the MTS and oriented so that the resultant force would act at 12

of
exion and 21

of adduction of the femur. This is the orientation achieved during postoperative
partial weight bearing with ambulatory supports.

47

Load was applied to the construct via the
appropriately sized femoral head at a rate of 2.44 mm/min until a resultant load of 100 kg was
reached. Motion was sensed by a strain gauge transducer mounted between two 1.5-mm threaded
rods 13 mm apart and implanted either side of the prosthetisbone interface. The transducer
consisted of a 0.25-mm-thick and 5.5-mm-wide strip of brass bent into a rectangle. A gap of 4 mm
was present at the free end of the rectangle. Four foil-type strain gauges were mounted to the
transducer and connected in a Wheatstone bridge conguration. The transducer was calibrated by
attaching two similar rods to the MTS and advancing the actuator at 0.001 in. and the resolution
of the system determined at 10

m (Figure 34.3).

28

There is no information given on the diameters of prosthesis implanted and these may vary
both within and between the study cohorts. The selected load of 100 kg represents about 1.4 times
body weight and the predicted partial weight-bearing values are between 0.5 to two times body
weight.

36,43

The values obtained for motion again are not pure interface values and reect both
compression of the prosthesis but also that of subchondral cancellous bone. The measurement was
only recorded in one direction and may be misleading. Theoretically, six degrees of freedom with
rotation about three perpendicular axes and translation along these could be present. No mention
is made of whether the acetabula were matched pairs and it is difcult to compare embalmed with
fresh-frozen pelvises because of their mechanical differences.

0266/ ch34/frame Page 533 Friday, October 22, 1999 10:36 AM

534

Mechanical Testing of Bone and the BoneImplant Interface

E. AXIAL LOAD ORIENTATED AT 10 OF FLEXION AND 15 OF ADDUCTION
Three fresh cadaver pelvises were harvested. High-contrast radiographs were taken in ve
projections: AP, iliac oblique, obturator oblique, pelvic inlet, and outlet. This ruled out any
underlying skeletal disease. The AP projection was used as a template for the correct acetabular
component size. The hemipelvises were then mounted by the iliac crests in polymethylmethacry-
late. Each acetabulum was under-reamed by 2 mm, preserving 50% of the subchondral bone
plate and the cups were implanted at 45 of abduction and 20 of anteversion. An acetabular
impactor and mallet were used to seat the prostheses. Prior to implantation the cups were modied
slightly; the polar screw hole at the apex of the dome and one peripheral screw hole were bored
and threaded. Threaded metal pins were then screwed into place and passed through overdrilled
holes in the pelvis. The dome pin passed through the medial wall and the peripheral screw
through the posterior wall of the acetabulum. Cylindrical brass platforms were then placed around
each motion pin and attached to the bone with epoxy-cement; these platforms served as a reference
point against which movement of the pins could be measured. Extensometers were used to assess
the relative movement between pin and platform (Figure 34.4). Motion of the implant was
measured in three planes under simulated stance loads: anterioposterior tilt, representing move-
FIGURE 34.3 Testing under physiological load.
FIGURE 34.4 Motion pin and brass used to detect movement between the cup and bone.
0266/ ch34/frame Page 534 Friday, October 22, 1999 10:36 AM
In Vitro Testing of the Stability of Acetabular Components 535
ment in anteversion and retroversion; lateral tilt, representing movement in abduction-adduction
plane; and axial motion, representing movement parallel to the motion of the axis of the
micromotion pin. Assessment of seating of the component was carried out in two ways:
(1) visualization of the contact through the screw holes and (2) 5-mm disks of pressure-sensitive
lm were placed at multiple sites in the acetabulum prior to implantation. Five standard radio-
graphic projections were taken. The specimens were then mounted in the loading jig oriented in
the position of single-legged stance with 15 of abduction and 10 of forward exion relative to
the vertically directed load. A load of 200 lb resulting in a joint reaction force of approximately
250 lb was applied with an appropriately sized femoral head. The relative motion of the prosthesis
in relation to the bone was assessed. The cups were then secured with four 6.5-mm cancellous
screws, two in the vertical position, one in the resultant position, and one in an anteriosuperior
direction. All screws were tightened to a torque of 12 in.-lb. A second set of radiographic
projections was taken to assess cup position. When measurements with four screws had been
made, the screws were successively removed to give values for three, two, and one screw. After
testing was complete, the acetabular component was removed and reamed to give a 1-mm press-
t and then 0 mm of discrepancy. The experimental procedure was repeated.
49
Errors are inherent in the results obtained with this testing regime. Implantation of pressure-
sensitive lm will to some degree compromise the xation, particularly in the press-t cups. Some
problems are encountered when performing multiple experiments on one specimen, as not only
will the screw holes alter the performance of the acetabular prosthetic interface but re-reaming the
same acetabulum will give results that are likely to be different from those achieved in a fresh
specimen. Initially, 50% of the subchondral bone plate was preserved. After reaming to leave no
discrepancy between cup and acetabulum it is likely that very little would be left, altering the
mechanical properties signicantly. One may assume from the methodology that as the AP radio-
graph was used as a template for the correct size of prosthesis that the component size may not be
the same between the pelvises. The measurement apparatus is ingenious but again is not measuring
the deformation purely at the boneprosthesis interface but also the deformation of the prosthesis
and subchondral cancellous bone.
F. AXIAL LOAD WITH THE PELVIS ORIENTATED IN THE ANATOMICAL POSITION
In this experiment, 11 fresh-frozen cadaver pelvises were selected and the acetabula evaluated
to remove those with fractures or lesions. The posterior aspect of each pelvis was mounted in
gypsum dental stone to minimize rigid-body motion during loading. The acetabula were then
reamed to the medial pelvic wall with size-matched reamers and debris cleared with pressurized
water. Three types of acetabular prosthesis were implanted: (1) a porous-coated titanium press-
t hemispherical cup; (2) a porous-coated titanium cup with dome screws; and (3) a porous-
coated titanium cup with three spikes. For comparison a cemented metal-backed prosthesis was
also tested. The cups were implanted in accordance with the manufacturers guidelines. To gain
consistency with positioning a preangled rod was used to place the cups in 40 abduction and
15 anteversion. Seating was obtained with a soft mallet. The specimens were then mounted on
the MTS and a 32-mm femoral head with a 125 neckshaft angle was seated in the acetabulum
in anatomical orientation. The rst four pelvises tested incorporated pressure-sensitive lm at
the boneprosthesis interface to verify the position of load application. Three 1.5 1.5 cm
aluminum blocks were secured to the periacetabular bone with small Kirschner wires and cement.
Eddy current transducers were used to measure the movement of the cup rim relative to these
plates. The sensitivity of these transducers was 1 m. The rst transducer was placed slightly
posterior to the ilium and the remaining two staggered by 120 (Figure 34.5).
Five cycles of axial load of 0 to 2354 N at crosshead speed of 10 to 20 mm/min was selected
to seat the prosthesis initially, followed by three consecutive trials from 0 to 2354 N at a crosshead
speed of 0.5 mm/min increasing in 392-N increments. The prosthesis was then augmented in the
0266/ ch34/frame Page 535 Friday, October 22, 1999 10:36 AM
536 Mechanical Testing of Bone and the BoneImplant Interface
same sample with one, then two 4.5-mm screws. The experimental procedure was then repeated.
Each type of testing was carried out on the same acetabulum.
36
The experiments appear to utilize the same acetabulum for each test, so here again the
mechanical behavior of the bone will change once it has been loaded in the rst experiment.
It is difcult then to compare the results of one- and two-screw xation accurately in the same
sample. The screws were tightened until snug; there is a wide variation in the torque that is
imparted to screws and the security in bone is reliant on this. Once the prosthesis has been
tested, there is no doubt that the purchase of the screw will have changed. There is great difculty
in measuring the movement at the interface. This form of data collection has the same inherent
errors as previously discussed. Implantation of pressure-sensitive lm will also affect the
interface performance.
G. REPETITIVE 100,000 CYCLE AXIAL LOADING IN FULL EXTENSION,
LOAD TO FAILURE IN SAME ORIENTATION
In this experiment, 33 embalmed cadaver hemipelvises were mounted in gypsum dye stone to a
neutral frame. Five different types of porous-coated acetabular components were selected for
testing, 1-mm press-t, 1-mm press-t with peg xation, 1-mm press-t with screw xation,
line-to-line press-t with screw xation, and 1.5-mm peripheral expansion press-t. Each pros-
thesis was implanted in accordance with the manufacturers guidelines. Micromotion and sub-
sidence of the cup were measured three dimensionally by six LVDTs accurate to 3 m. Two
aluminum blocks were mounted to the anteriomedial and posteriolateral metal rim of each cup.
Six LVDTs were attached to the cup with two cube transducer holders. The LVDTs were xed
perpendicular to the three sides of the block to determine three-dimensional movement
(Figure 34.6). The construct was then placed on the MTS and loaded with the hip in full extension.
The specimens were loaded with a sinusoidal prole from a preload of 10 kg to a peak of 100 kg
at 5 Hz for 100,000 repetitions. At the completion of this loading cycle static load-to-failure
analysis was repeated at successive loads of 150, 200, 250, and 300 kg. Five cycles at each
increment were performed until gross motion was detected.
21
This study uses a sophisticated method of three-dimensional cup movement. However, again
it is not measuring the interface displacement directly, but is testing only one parameter in each
sample. The peak load is slightly lower than those predicted postoperatively and the loading is
performed in extension, which is an unlikely position.
FIGURE 34.5 Transducers plates in the ilium, publis, and ischium.
0266/ ch34/frame Page 536 Friday, October 22, 1999 10:36 AM
In Vitro Testing of the Stability of Acetabular Components 537
H. FINITE-ELEMENT ANALYSIS TO CALCULATE ACETABULAR DEFORMATION
Acetabular strain distribution was studied using nite-element analysis (FEA) applied to different cup
designs: 1- and 2-mm oversized hemispherical cups, an elliptical shape, a dual-radius geometry, a
cup with a hemispherical dome and a wider outer cup diameter, and a cup with low-prole geometry.
The magnitude and distribution of acetabular strains were assessed with an axisymmetric nite-
element model. The shape of the bone section was based on the geometry described by Pedersen
et al.
50
The cup was xed by the ilium and was displaced toward the center of the dome until contact
occurred at the medial wall of the acetabulum. The acetabular notch was not included. This increases
the rigidity of the acetabulum; to compensate, the modulus of elasticity of cortical bone was reduced
over that used in two-dimensional FEA.
34
A ve-stage linear transition from the cortical rim to the
subchondral bone was added. The Poissons ratio was set at 0.3. The nite-element mesh was created
with I-DEAS (version V.I, Structural Dynamics Research, Milford, OH) and solved using ABAQUS
(version 5.4-1, Hibbitt, Karlsson, and Sorensen, Pawtucket, RI). The acetabular prosthesis was mod-
eled as rigid and the coefcient of friction between bone and cup set at 0.15. Higher values for this
have been previously reported.
51
Increasing the frictional coefcient led to incomplete seating of the
cup in this model. Overall acetabular deformation was determined by calculating the strain energy
absorbed by the nite model. Strain energy is the recoverable work due to deformation.
2
As acetabular strains were calculated in a nite-element model the values obtained might not
be those seen in vivo. It does, however, compare the differences in the geometries of each cup
under stringent conditions. The main problem is that it cannot measure the initial stability of the
boneprosthetic interface.
I. PULLOUT FORCE
Aluminum models of the ve-cup geometries in the previous section were machined. The models were
coated with glass beads with no porous coating in order to assess the effect of the geometric differences.
Each prosthesis was implanted into foam bones after reaming on a drill press which produced consistent
hemispherical shapes. A rod was threaded into the center of each cup and mounted on the MTS. A
gradually increasing tensile force was then applied to the rod until failure.
2
The results obtained cannot
be directly translated to the clinical situation but it isolates the cup geometries as the only variable.
The method of failure is not physiological unless a constrained acetabular liner were to be implanted.
FIGURE 34.6 Setup for three-dimensional displacement detection.
0266/ ch34/frame Page 537 Friday, October 22, 1999 10:36 AM
538 Mechanical Testing of Bone and the BoneImplant Interface
III. CONCLUSION
There is no fault-free method of testing the stability of the acetabular cup. None of the published
methods can measure the micromotion at the prosthesisbone interface directly. All rely on indirect
methods. It is noticeable in the literature that authors have become aware of this and are employing
many varieties of strain gauges and congurations to record values for micromovement. Method-
ology does require elimination of variables by using consistent techniques of both implantation
and testing. It is tempting to test more than one parameter on one acetabulum. It is imperative that
testing conditions and design are standardized, as by doing this comparisons between centers and
thus different prostheses can be made, which at the moment is difcult because of the wide variety
of parameters measured. If a single test method were adopted, results could become more consistent
and thus accurate. The authors feel that torsional stability is the most important parameter and
therefore test this at their institution.
REFERENCES
1. Cameron, H.U., Modied cups, Orthop. Clin. North Am., 29, 277, 1998.
2. Ries, M.D., Harbaugh, M., Shea, J., and Lambert, R., Effect of cementless acetabular cup on strain
distribution and press t stability, J. Arthroplasty, 12, 207, 1997.
3. Keating, E.M., Ritter, M.A., and Faris, P.M., Structures at risk from medially placed acetabular screws,
J. Bone Joint Surg., 72A, 509, 1990.
4. Zalenski, E., Jasty, M., Page, A., et al., Micromotion of porous-surfaced, cementless prosthesis fol-
lowing 6 months of in vivo bone growth in a canine model, presented at The 35th Annual Meeting of
the Orthopaedic Research Society, Las Vegas, NV, February 69, 1989.
5. Letournel, E., Failures of biologically xed devices: causes and treatment, in The Hip, Proceedings
of the 14th Open Scientic Meeting of the Hip Society, Mosby, St. Louis, MO, 1987, 318.
6. Gallante, J.O., Overview of current attempts to eliminate methyl methacrylate, in The Hip, Proceedings
of the 11th Open Scientic Meeting of the Hip Society, Mosby, St. Louis, MO, 1984, 181.
7. Pilliar, R.M., Bony ingrowth into porous metal coated implants, Orthop. Rev., 9, 85, 1980.
8. Burke, D.W., Bragdon, C.R., OConnor, D.O., et al., Dynamic measurement of interface mechanics
in vivo and the effect of micromotion on bone ingrowth into a porous surface device under controlled
loads in vivo, Orthop. Res. Soc. Trans., 16, 103, 1991.
9. Clarke, H.J., Jinnah, R.H., Warden, K.E., et al., Evaluation of acetabular stability in uncemented
prostheses, J. Arthroplasty, 6, 335, 1991.
10. Pilliar, R.M., Lee, J.M., and Maniatopoulos, C., Observations on the effect of movement on bone
growth into porous-surfaced implants, Clin. Orthop., 208, 108, 1986.
11. Haddad, R.J., Cook, S.D., and Thomas, K.A., Biological xation of porous-coated implants: current
concepts review, J. Bone Joint Surg., 69A, 1459, 1987.
12. Phillips, T.W., Johnston, G., and Wood, P., Selection of an animal model for resurfacing hip arthro-
plasty, J. Arthroplasty, 2, 111, 1987.
13. Cameron, H.U., Pilliar, R.M., and McNab, I., The effect of movement on the bonding of porous metal
to bone, J. Biomed. Mater. Res., 7, 301, 1973.
14. Bobyn, J.D., Pilliar, R.M., Cameron, H.U., and Weatherly, G.C., The optimum pore size for the xation
of porous surfaced metal implants by ingrowth of bone, Clin. Orthop., 150, 263, 1980.
15. Pilliar, R.M., Powder metal-made orthopaedic implants with porous surface for xation by tissue
ingrowth, Clin. Orthop., 176, 42, 1983.
16. Bobyn, J.D., Pilliar, R.M., Cameron, H.U., and Weatherly, G.C., Osteogenic phenomena across
endosteal bone implant spaces with porous surfaced intramedullary implants, Acta Orthop. Scand.,
52, 145, 1981.
17. Cook, S.D., Barrack, R.L., Thomas, K.A., and Haddad, R.J., Quantitative analysis of tissue growth
into human porous total hip components, J. Arthroplasty, 3, 249, 1988.
18. Cook, S.D., Renz, E.A., Barrack, R.L., et al., Clinical and metallurgical analysis of retrieved internal
xation devices, Clin. Orthop., 194, 236, 1985.
0266/ ch34/frame Page 538 Friday, October 22, 1999 10:36 AM
In Vitro Testing of the Stability of Acetabular Components 539
19. Cook, S.D., Thomas, K.A., and Haddad, R.J., Jr., Histologic analysis of retrieved human porous-
coated total joint components, Clin. Orthop., 234, 90, 1988.
20. Bobyn, J.D., Pilliar, R.M., Cameron, H.U., et al., The effects of porous surface conguration on the
tensile strength of xation of implants by bony ingrowth, Clin. Orthop., 149, 291, 1980.
21. Stiehl, J.B., MacMillan, E., and Skrade, D.A., Mechanical stability of porous-coated acetabular
components in total hip arthroplasty, J. Arthroplasty, 6, 295, 1991.
22. Hedly, A.K., Kabo, N., Kim, W., et al., Bone ingrowth xation of a newly designed acetabular
component in a canine model, Clin. Orthop., 176, 12, 1983.
23. Jasty, M. and Harris W.H., Observation on factors controlling bony ingrowth into weight-bearing
porous canine total hip replacement, in Non-Cemented Total Hip Arthroplasty, Fitzgerald, R., Ed.,
Raven Press, New York, 1988.
24. Pilliar, R.M., Cameron, H.U., Welsh, R.P., and Binnington, A.G., Radiographic and morphologic
studies of load-bearing porous surface structural implants, Clin. Orthop., 156, 249, 1981.
25. Bourne, R.B., Finlay, J.B., Rorabeck, C.H., and Landsberg, R.P., The effect of endoprosthetic mismatch
and metal or non-metal-backed acetabular components on in vitro pelvic stresses, in The Hip, Pro-
ceedings of the 13th Open Scientic Meeting of the Hip Society, Mosby, St. Louis, 1985, 114.
26. Curtis, M.J., Jinnah, R.H., Wilson, V.D., and Hungerford, D.S., The initial stability of uncemented
acetabular components, J. Bone Joint Surg., 74B, 372, 1992.
27. Perona, P.G., Lawrence, J., Paprosky, W.G., et al., Acetabular micromotion as a measure of initial
implant stability in primary hip arthroplasty, J. Arthroplasty, 7, 537, 1992.
28. Lachiewicz, P.F., Suh, P.B., and Gilbert, J.A., in vitro initial xation of porous-coated acetabular total
hip components a biomechanical comparative study, J. Arthroplasty, 4, 201, 1989.
29. Evans, F.G., Mechanical Properties of Bone, Charles C Thomas, Springeld, IL, 1972.
30. Hadjari, M.H., Hollis, J.M., Hoffman, O.E., et al., Initial stability of porous coated acetabular implants,
Clin. Orthop., 307, 117, 1994.
31. Litsky, A.S. and Pophal, S.G., Initial mechanical stability of acetabular prostheses, Orthopedics, 17,
53, 1994.
32. Andersson, G.B., Freeman, M.A., and Swanson, S.A., Loosening of the cemented acetabular cup in
total hip replacement, J. Bone Joint Surg., 54B, 590, 1972.
33. Oh, I., A comprehensive analysis of the factors affecting acetabular cup xation and design in total
hip replacement arthroplasty: a series of experimental and clinical studies, in The Hip, Proceedings
of the 11th Open Scientic Meeting of the Hip Society, Mosby, St. Louis, MO, 1984, 129.
34. Vasu, R., Carter, D.R., and Harris, W.H., Stress distribution in the acetabular region: I. Before and
after total joint replacement, J. Biomech., 15, 155, 1982.
35. Rydell, N.P., Forces acting on the femoral head-prosthesis: a study on strain gauge supplied prosthesis
in living persons, Acta Orthop. Scand., 37(Suppl.) 88, 1966.
36. Simon, S.R., Paul, I.L., Rose, R.M., and Radin, E.L., Stiction friction of total hip prostheses and
its relationship to loosening, J. Bone Joint Surg., 57A, 226, 1975.
37. Wilson, J.N. and Scales, J.T., Loosening of total hip replacements with cement xation, Clin. Orthop.,
72, 145, 1970.
38. Ma, S.M., Kabo, J.M., and Amstutz, H.C., Frictional torque in surface and conventional hip replace-
ment, J. Bone Joint Surg., 65A, 366, 1985.
39. Volz, R.G. and Wilson, R.J., Factors affecting the mechanical stability of the cemented acetabular
component in total hip replacement, J. Bone Joint Surg., 59A, 501, 1977.
40. Ohlin, A. and Balkfors, B., Stability of cemented sockets after 314 years, J. Arthroplasty, 7, 87, 1992.
41. Brown, T.D. and Ferguson, A.B., Jr., Mechanical property distributions in the cancellous bone of the
human proximal femur, Acta Orthop. Scand., 51, 429, 1980.
42. Carter, D.R. and Hayes, W.C., The compressive behavior of bone as a two phase porous structure,
J. Bone Joint Surg., 59, 954, 1977.
43. Lindahl, O., Mechanical properties of dried defatted spongy bone, Acta Orthop. Scand., 47, 11, 1976.
44. McElhaney, J.H. and Byars, E.F., Dynamic response of biologic materials, American Society of
Mechanical Engineers, Publ. 65-WA/HUF 9, 1, 1965.
45. Van Audekercke, R. and Martens, M., Mechanical properties of cancellous bone, in Natural and Living
Biomaterials, Hastings, G.W. and Ducheyne, P., Eds., CRC Press, Boca Raton, FL, 1984, 89.
0266/ ch34/frame Page 539 Friday, October 22, 1999 10:36 AM
540 Mechanical Testing of Bone and the BoneImplant Interface
46. Bear, B., Bostrom, M., and Sculco, T., Press t acetabular cups, an evaluation of initial torsional
stability, AAOS Scientic Exhibit, San Francisco, CA, 1997.
47. Engh, C.A., Bobyn, J.D., and Matthews, J.G., III, Biologic xation of a modied Moore prosthesis,
in The Hip, Proceedings of the 11th Open Meeting of the Hip Society, Mosby, St. Louis, MO, 1984, 95.
48. Bergmann, G., Rohlmann, A., and Graichen, F., Hip joint forces during physical therapy after joint
replacement, Orthop. Trans., 14, 303, 1990.
49. Kwong, L.W., OConnor, D.O., Sedlacek, R.C., et al., A quantitative in vitro assessment of t and
screw xation on the stability of a cementless hemispherical acetabular component, J. Arthroplasty,
9, 163, 1994.
50. Pedersen, D.R., Crowninsheild, R.D., Brand, R.A., and Johnston, R.C., An axisymmetric model of
acetabular components in total hip arthroplasty, J. Biomech., 15, 305, 1982.
51. Shirazi-Adl, A., Dammak M., and Paiement G., Experimental determination of friction characteristics
at the trabecular bone/porous-coated metal interface in cementless implants, J. Biomed. Mater. Res.,
27, 167, 1993.
0266/ ch34/frame Page 540 Friday, October 22, 1999 10:36 AM

541

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

35

In Vitro

Testing of the Stability
of Femoral Components

Sanjiv H. Naidu, Fadi M. Khoury, and John M. Cuckler

CONTENTS

I. Introduction ..........................................................................................................................541
II. Methods ................................................................................................................................542
A. Materials and Methods...................................................................................................542
B. Derivation of Displacement and Rotation Equations ....................................................544
C. Rotation ..........................................................................................................................544
D. Vertical (Superior-Inferior) Translation .........................................................................545
E. Medial-Lateral Displacement .........................................................................................546
F. Anterior-Posterior Displacement....................................................................................546
III. Results ..................................................................................................................................546
IV. Discussion ............................................................................................................................547
References ......................................................................................................................................549

I. INTRODUCTION

The success of a cementless femoral component during

in vivo

service is a direct function of
osseointegration. Initial stability of cementless femoral devices is paramount for biological ingrowth
leading to mechanical xation.

1

A universal conclusion of clinical experiences with cementless
xation is that immediate stable xation is fundamental to a successful hip arthroplasty.

2,3

Primary xation of a femoral stem is accomplished by surgical preparation of the bone cavity to
optimally t the implant. Impaction of the femoral stem during implantation leads to a press-t between
the implant and the bone interface; the intention is to create an interface with minimal relative motion,
a condition crucial to the permanent xation of the implant stem with the bone. If such a primary
condition of minimal relative motion is met, secondary xation with bone ingrowth will follow, ensuring
a successful outcome. Burke and associates

4

showed that rotational moments caused by out-of-plane
forces sustained in activities such as stair climbing cause interface slippage of an order of magnitude
greater than coronal, in-plane loading. Cook

5

concluded that lack of proximal t and poor bone quality
signicantly increased the magnitude of stem micromotion when torsional moment was applied.
It has long been recognized that relative motions of human joints involve six degrees of
freedom.

6

The following setup provides for rigorous measurement of micromotion in three dimen-
sions: translations measured along and rotations are measured about the three axes, accounting for
all six degrees of freedom of the femoral stem motion relative to the femur. A modular femoral
stem with metaphyseal augmentation pads and distal cylindrical sleeves was used to vary the
proximal and distal canal ll in composite femurs. The modular system allows one to vary the
proximal and distal t/ll values reliably; by further adding consistency in bony anatomy and
material properties, the following experimental setup with composite bones clearly delineates the
role of canal t and ll on the initial stability of femoral stems.

0266/ ch35/frame Page 541 Friday, October 22, 1999 10:54 AM

542

Mechanical Testing of Bone and the BoneImplant Interface

II. METHODS
A. M

ATERIALS



AND

M

ETHODS

Test xtures that will attach to the femur and prosthesis were designed to permit the measurement
of the micromotion of the prosthesis relative to the femur. The test xtures consist of several parts
that, when fastened to the femur and the prosthetic femoral head, enable the calculation of the
motion of the prosthesis relative to the femur at a critical location, namely, the prosthetic center
and the bone interface. Six linear variable displacement transducers (LVDTs) were positioned as
shown in Figure 35.1 (RDP Electronics, Phoenixville, PA). The following will describe the manner
in which the test xtures are assembled together for data acquisition.
The plate that attaches to the femoral head is formed by fastening, with two screws, the two
half plates illustrated in Figure 35.2A. Both half plates were designed and machined such that the
cavity that forms when both plates are fastened together has the exact geometric conformation of the
femoral head. This will allow a very rigid connection of the top plate around the prosthetic head.
The bottom plate shown in Figure 35.2B fastens to the femur via long screws that all meet at
the midpoint of the plate. This midpoint coincides with both the axis of the implanted prosthesis
and the midpoint of the top plate. Gauges 5 and 6, of the six LVDTs, are fastened vertically at the
top plate such that the gauges are in contact with the upper surface of the bottom plate. The top
and bottom plates are aligned by two long screws to ensure that the plates are parallel in assembly.
The plate in Figure 35.2C fastens vertically to the medial edge of the bottom plate such that
the midpoint between gauges 1 and 2 is aligned with the horizontal plane bisecting the top plate.
The plate in Figure 35.2D mounts vertically onto the medial edge of the top plate. Gauges 1 and
2 are fastened to the plate in Figure 35.2C such that they are inferior and superior to the horizontal
bisecting plane of the top plate.

FIGURE 35.1

Experimental apparatus for evaluating the initial stability of a cementless femoral stem in
anatomically consistent composite bones. The prosthesis plate is xed to the prosthetic head. The femoral
bone clamps afx the bottom plate to the femur. LVDTs attached to the two plates measure the relative motion
between the prosthesis and bone. (From Naidu, S.H. et al.,

Am. J. Orthop.,

25, 829, 1996. With permission).

0266/ ch35/frame Page 542 Friday, October 22, 1999 10:54 AM

In Vitro

Testing of the Stability of Femoral Components

543
F
I
G
U
R
E

3
5
.
2

P
l
a
t
e

f
i
x
t
u
r
e
s

t
o

h
o
l
d

t
h
e

f
e
m
o
r
a
l

c
o
m
p
o
n
e
n
t
:

(
A
)

T
h
e

f
e
m
o
r
a
l

h
e
a
d

o
r

t
h
e

t
o
p

p
l
a
t
e

w
h
e
r
e

L
V
D
T
s

5
,

a
n
d

6

a
r
e

f
a
s
t
e
n
e
d
;

(
B
)

t
h
e

b
o
t
t
o
m

p
l
a
t
e
w
i
t
h

s
c
r
e
w

c
l
a
m
p
s

f
o
r

h
o
l
d
i
n
g

t
h
e

f
e
m
u
r
;

(
C
)

p
l
a
t
e

t
o

h
o
l
d

L
V
D
T
s

1

a
n
d

2
;

(
D
)

t
h
i
s

p
l
a
t
e

m
o
u
n
t
s

v
e
r
t
i
c
a
l
l
y

o
n
t
o

t
h
e

m
e
d
i
a
l

e
d
g
e

o
f

t
h
e

t
o
p

p
l
a
t
e
;

(
E
)

t
w
o

o
f
t
h
e
s
e

p
l
a
t
e
s

a
r
e

m
o
u
n
t
e
d

m
e
d
i
a
l

a
n
d

l
a
t
e
r
a
l

t
o

t
h
e

f
e
m
u
r

t
o

h
o
l
d

L
V
D
T
s

3

a
n
d

4
,

r
e
s
p
e
c
t
i
v
e
l
y
;

(
F
)

p
l
a
t
e

t
o

t
r
a
n
s
f
e
r

l
o
a
d

f
r
o
m

t
h
e

I
n
s
t
r
o
n

c
r
o
s
s
h
e
a
d

t
o

t
h
e

f
e
m
o
r
a
l
h
e
a
d
.

(
F
r
o
m

N
a
i
d
u
,

S
.
H
.
,


A
m
.

J
.

O
r
t
h
o
p
.
,


2
5
,

8
2
9
,

1
9
9
6
.

W
i
t
h

p
e
r
m
i
s
s
i
o
n
.
)

0266/ ch35/frame Page 543 Friday, October 22, 1999 10:54 AM

544

Mechanical Testing of Bone and the BoneImplant Interface

The two plates shown in Figure 35.2E are fastened vertically to the anterior edge of the bottom
plate, anterior to the coronal plane of the femur, such that gauges 3 and 4 align with the top plate.
Gauges 3 and 4 are placed such that they are medial and lateral, respectively, to the sagittal plane
of the femur. The load transfer from the Instron 1331 to the femoral head is facilitated by the
component illustrated in Figure 35.2F.
Modular femoral stems (Smith and Nephew, Memphis, TN) were implanted in composite bones
(Pacic Research Labs, Vashon Island, WA) with a section modulus equal to that of human femora.

7

They were used for consistency in anatomy and in mechanical properties. In all, 20 bones were
randomly prepared to receive one of four prosthesis congurations. A size 14 large femoral stem
was modied by either adding a 6-mm proximal anterior pad or a 17-mm distal sleeve, by adding
both, or by using only a 0-mm anterior pad without a distal sleeve. Therefore, ve femurs, and
thus ve tests for each of the following congurations, were prepared and mechanically tested:
0/S = 0-mm proximal anterior pad, no distal sleeve;
6/S = 6-mm proximal anterior pad, no distal sleeve;
0/+S = 0-mm proximal anterior pad, 17-mm distal sleeve;
6/+S = 6-mm proximal anterior pad, 17-mm distal sleeve.
The 20 bones were randomly prepared to receive ve stems of each of the above four
congurations. Osteotomy was made 1.5 cm above the lesser trochanter, and serial reaming with
appropriate rigid reamers and broaching was done. The reconstructed femora were aligned in an
aluminum pot and cemented in place with polyester resin. They were then cyclically loaded on
a servohydraulic Instron 1331 (Canton, MA) between 130 and 1500 N at 1 H; the assembly was
positioned in 7

of adduction and loaded in neutral to mimic level walking and at 20

of exion
to mimic stair climbing. Before mechanical testing, prosthetic seating was conrmed by radio-
graph as shown in Figure 35.3.
After mechanical testing, one 6/+S, and one 0/S were sectioned on the accutome without
embedding. The sections were radiographed and analyzed on the Macintosh image analysis system
to determine t/ll values. Fit was determined as the percentage of prosthetic surface in close
apposition with the endosteum (<1 mm); ll was calculated as the percentage of the endosteal canal
that the prosthesis occupied. Data from four proximal sections and four distal sections were averaged
to obtain the t/ll values for each of the four prosthetic congurations.

B. D

ERIVATION



OF

D

ISPLACEMENT



AND

R

OTATION

E

QUATIONS

In these equations, the prosthesis and the bone are considered to be rigid bodies. The design of the
test assembly necessitated the LVDTs to be centered about a point in the center of the upper plate.
In order to give a better estimate of the relative motion at the real interfaces, motions were calculated
using the point in the center of the lower plate of the test assembly as a reference. In each direction,
the translation at this point is the translation in the center of the top plate plus an additional
displacement due to the rotation of the prosthesis.

C. R

OTATION

Rotation in each plane is found from the arctangent of the difference between the two appropriate
LVDTs and the distance between them. Transverse plane rotations are found using LVDTs 3 and
4 (

T

); coronal from 1 and 2 (

c

); and sagittal from 5 and 6 (

S

). The calculation is identical in
each plane. For example, if

S

56

is the distance between LVDTs 5 and 6, the rotation in the coronal
plane (

c

) is

c



= arctan (#6 #5)/

S

56

(35.1)

0266/ ch35/frame Page 544 Friday, October 22, 1999 10:54 AM

In Vitro

Testing of the Stability of Femoral Components

545

D. V

ERTICAL

(S

UPERIOR

-I

NFERIOR

) T

RANSLATION

This is the easiest case, since there is no contribution from a rotation. That is, the vertical displace-
ment of the prosthesis is the same for all the points along the axis of the prosthesis. Since LVDTs
5 and 6 were centered about this axis, the vertical displacement is given by

SI

= (#5 + #6)/2 (35.2)

FIGURE 35.3

Anteroposterior (A) and lateral (B) radiographs of the prosthesis with 6-mm anterior proximal
pad without the distal sleeve, conrming prosthetic seating in the composite bones. (From Naidu, S.H. et al.,

Am. J. Orthop.,

25, 829, 1996. With permission).

0266/ ch35/frame Page 545 Friday, October 22, 1999 10:54 AM

546

Mechanical Testing of Bone and the BoneImplant Interface

E. M

EDIAL

-L

ATERAL

D

ISPLACEMENT

This displacement is found by adding the mediolateral displacement in the top plate to a translation
resulting from rotation in the coronal plane. If

h

is the distance between the top plate and the
bottom plate, and

S

12

is the distance between LVDTs 1 and 2, then the translation along the
mediolateral plane is

ML

= (#1 + #2)/2 +

h

tan

c

= (#1 + #2)/2+

h

(#2 #1)/

S

12

(35.3)

F. A

NTERIOR

-P

OSTERIOR

D

ISPLACEMENT

Anteroposterior displacement is calculated in the same manner using LVDTs 3 and 4 for the
translational component and the rotation in the sagittal plane (LVDTs 5 and 6). The rotational
component is subtracted in this case since a positive rotation in the coronal plane results in a
posterior translation. If

S

56

is the distance between LVDTs 5 and 6, then

AP

= (#3 + #4)/2

h

tan

S

= (#3 + #4)/2

h

(#6 #5)/

S

56

(35.4)

III. RESULTS

From Table 35.1 it is clear that none of the four groups showed signicant differences in translational
interfacial micromotion when tested with neutral loading and exion loading. However, it is evident
that prostheses in all four groups showed increased translational micromotion when tested in exion,
compared with neutral loading. While the mean anteroposterior displacements in exion ranged
from 56.4 to 72.4

m, it ranged from 8.7 to 12.2

m in neutral loading.
Similarly, Table 35.2 shows that rotation of the prosthesis is consistently increased in all three
planes when tested in exion for all four congurations. This increase is signicant along both the
prosthesis vertical and anteroposterior axes (

p

< 0.05). The mean rotation (

10

3

) in exion along
the vertical axis ranged from 43.5 to 74.5; whereas in neutral loading, mean vertical rotation
(

10

3

) ranged from 6.3 to 13.4. The anteroposterior rotation in exion also showed a six- to
sevenfold increase when compared with neutral loading.
Although the rotation in exion consistently increased for all four congurations in comparison
with neutral loading, there are some notable differences in rotation between the four groups in
exion. Prosthesis with the larger proximal pad, without the distal sleeve, 6/-S, showed a mean
rotation (

10

3

) of 43.5, whereas the one with the smaller proximal pad, without the distal sleeve,

TABLE 35.1
Relative Micromotion (mm

SD) at the Femoral Stem Bone Interface

Vertical (

Z

Axis)

Mediolateral (

X

Axis)

Anteroposterior (

Y

Axis)
Group Neutral

n

= 5 Flexon

n

= 5 Neutral

n

= 5 Flexion

n

= 5 Neutral

n

= 5 Flexion

n

= 5

0/S 13.9

0.4 19.8

1.0 44.9

1.4 47.4

2.3 8.7

1.4 56.4

8.9

a

6/S 13.6

0.7 19.0

0.5 45.9

2.4 47.5

1.2 10.1

0.9 72.4

6.1

a

0/+S 17.1

0.5 22.3

1.4 46.7

1.2 45.7

1.8 10.8

1.4 69.4

4.7

a

6/+S 16.5

1.0 21.7

2.9 49.9

1.9 51.2 3.5 12.2 1.9 67.0 10.3
a
Note: For group descriptions, see text in methods section.
a
Signicantly higher than displacements along Y axis in neutral loading (one-way ANOVA p < 0.05).
Source: Naidu, S.H., Am. J. Orthop., 25, 829, 1986. With permission.
0266/ ch35/frame Page 546 Friday, October 22, 1999 10:54 AM
In Vitro Testing of the Stability of Femoral Components 547
0/S, showed an average rotation of 56.4 along the vertical axis. Although this difference is not
statistically signicant, all ve tests in the 6/S group consistently showed less rotation than did
those of the 0/S group.
When the 17-mm distal sleeve was added to the smaller, 0-mm proximal pad stem, the 0/+S
group showed a mean rotation ( 10
3
) of 74.5 along the vertical axis; this is a denite increase
over its sleeveless counterpart, 0/S, which showed a mean rotation of 56.4 ( 10
3
). When the
distal sleeve was added to the larger 6-mm proximal pad augmented stem, the 6/+S stem showed
a signicant increase in mean rotation along its vertical axis at 73.1 10
3
, compared with its
sleeveless counterpart, which displayed vertical rotation of 43.5 10
3
.
From Table 35.3, it is apparent that proximal ll is essentially unchanged at 63% for all the
congurations. Addition of the 17-mm distal sleeve increased the distal canal ll from 71 to 96%.
Although proximal ll was unchanged by the 6-mm proximal pad, proximal t was markedly
improved from 6 to 37%. Distal t was increased from 16 to 57% by adding the distal sleeve.
IV. DISCUSSION
The increase in anteroposterior displacements in exion may be explained by the increase in the
posteriorly directed component of the loading force when the potted femur is exed. Although no
other axis shows such a dramatic increase in translational micromotion, rotational instability is
evident along both the anteroposterior and vertical axes. The six- to sevenfold increase in both
anteroposterior and vertical rotation when compared with neutral loading suggests that the pros-
TABLE 35.2
Relative Rotation at the Bone Stem Interface (degrees 10
3
SD)
Vertical (Z Axis) Mediolateral (X Axis) Anteroposterior (Y Axis)
Group Neutral n = 5 Flexon n = 5 Neutral n = 5 Flexion n = 5 Neutral n = 5 Flexion n = 5
0/S 8.4 2.0 56.4 8.9
a
60.2 2.6 75.6 5.6 7.0 0.3 49.0 2.9
a
6/S 6.3 1.0 43.5 3.6
a,b
62.3 3.8 72.2 2.9 6.4 1.0 47.8 3.2
a
0/+S 12.1 1.8 74.5 5.9
a
65.8 2.5 79.3 6.5 7.7 0.9 53.9 2.3
a
6/+S 13.2 2.4 73.1 12.6
a,b
69.9 4.2 84.8 11.4 8.4 1.0 57.0 5.0
a
a
Signicantly higher than its neutral rotation along the same axis (one-way ANOVA p < = 0.05).
b
Signicant intergroup difference (one-way ANOVA p < = 0.05).
Source: Naidu, S.H., Am. J. Orthop., 25, 829, 1996. With permission.
TABLE 35.3
Intramedullary Canal Fit and Fill Percentages
Conguration Proximal Fill Distal Fill Proximal Fit Distal Fit
0/S 63% n = 3 71% n = 4 6% n = 3 16% n = 4
6/S 63% n = 4 71% n = 4 37% n = 4 16% n = 4
0/+S 63% n = 3 96% n = 4 6% n = 3 57% n = 4
6/+S 63% n = 4 96% n = 4 37% n = 4 57% n = 4
Note: Congurations are the same as the group descriptions in the methods section.
Source: Naidu, S. H., Am. J. Orthop., 25, 829, 1996. With permission.
0266/ ch35/frame Page 547 Friday, October 22, 1999 10:54 AM
548 Mechanical Testing of Bone and the BoneImplant Interface
thesis is unstable in more than one dimension when out-of-plane forces are applied. Burke and
colleagues, in a cadaver femora study, reported that torsional interface slippage increased by a
factor of ten in uncemented devices, when the implant bone assembly was loaded on the materials
testing system assembly in 30 of exion.
Crowninshield and colleagues
8
estimated the greatest loads on the hip to be during stair
climbing, when the resultant was predicted to be more than seven times the body weight, with the
component of force acting on the pelvis in the anteroposterior direction being up to ve times the
body weight. They commented that this force affects the torque that is transmitted to the bonepros-
thesis interface and may be related to implant loosening. Our observation of increased rotation in
exion loading in more than one axis is consistent with their ndings.
Although most in vitro studies on implant stability use cadaver femora, the validity of our
model is conrmed by Walker and Robertsons
9
experiments. They tested the so-called exact t
stem in cadaver femora, both in vertical loading and with an added out-of-plane force. In vertical
loading, similar to coronal neutral stance loading, the anteroposterior displacement ranged from
8.8 to 26.3 m; however, when they added a 100 N anterior force to the 1000 N vertical load, the
anteroposterior displacements increased to a range of 42.2 to 51.0 m. The vertical rotational angles
increased from 19 10
3
to 73 10
3
by adding the anterior force. Since the measured translational
micromotion and rotational angles are of the same magnitude under similar loading regimens, the
use of composite bones for initial stability study is justied.
Gilbert and colleagues,
10
in a cadaver femora study, compared a porous-coated anatomical stem
and a custom prosthesis with higher proximal ll; the authors demonstrated no signicant difference
between the porous-coated anatomical stem and the custom prosthesis in both migration and
micromotion. The authors hypothesized that stability appears to be provided by the relatively small
areas of cortical contact between the stem and the bone. The authors data appear consistent with
their observations. The 0-mm proximal pad gave a proximal t value of 6%; adding a 6-mm
proximal pad increased the t to 37%. This improved t decreased the prosthesis rotation along
its vertical axis, as shown in Table 35.3.
Huiskes,
11
in his study of cementless stems using nite-element analysis, hypothesized that
press-t devices require elastic subsidence to develop interface compression to balance the axial
joint compressive forces. Since the proximal femoral mediolateral stem curvature dictates the
amount of mediolateral compressive forces generated, it follows that the anteroposterior taper angle
will inuence the amount of anteroposterior compressive forces generated to stabilize the prosthesis.
Since there is no ingrowth, initial stability provided by the larger anterior pad depends solely on
prosthetic elastic subsidence and the shear stress generated by the porous proximal coating
(Figure 35.4). With the larger anterior pad, the anteroposterior taper angle increases, thereby
allowing more anteroposterior compressive forces to be generated. These larger anteroposterior
compressive forces reduce the prosthesis rotation about the vertical axis.
The addition of the distal sleeve increased the average distal ll to 96%, and the average
distal t to 57%. From Table 35.3 it is evident that the addition of the distal sleeve signicantly
increased the rotation of the 6/+S group compared with the 6/S group along the prosthesis
vertical axis. The distal canal was rigidly reamed to 17 mm, and, therefore, distal xation is
unlikely to be the reason for the increase in rotation. The increase in distal t probably
diminishes the elastic subsidence of the stem. Because of decreased stem subsidence, the effect
of proximal pad wedging is effectively negated. Since the composite bone was reamed anteriorly
to accommodate the proximal pad, the observed rotational instability of the 6/+S conguration
is not surprising.
In the above in vitro study, the authors demonstrate that the modular uncemented femoral stem
investigated is more unstable when subject to exion loading than in neutral loading. This study
also suggests that it is probably the proximal t, and not the ll, that contributes to initial stability
of an uncemented system where stabilization by bony ingrowth has yet to occur.
0266/ ch35/frame Page 548 Friday, October 22, 1999 10:54 AM
In Vitro Testing of the Stability of Femoral Components 549
REFERENCES
1. Pilliar, R.M., Lee, J.M., and Maniatopoulos, C., Observations on the effect of movement on bone
ingrowth into porous surfaced implants, Clin. Orthop., 208, 108, 1986.
2. Engh, C.A. and Bobyn, J.D., The inuence of stem size and the extent of porous coating on femoral
bone resorption after primary cementless hip arthroplasty, Clin. Orthop., 1988; 231, 1988.
3. Engh, C.A., Bobyn, J.D., and Glassman, A.H., Porous coated hip replacement: the factors governing
bone ingrowth, stress shielding, and clinical results, J. Bone Joint Surg., 69B, 45, 1987.
4. Burke, D.W., OConnor, D.O., Zalenski, E.B., and Harris, W.H., Biomechanics of the femoral total
hip components in simulated stair climbing, Trans. Orthop. Res. Soc., 13, 345, 1988.
5. Cook, S.D., Walsh, K.A., Haddad, R.J., Jr., Interface mechanics and bone growth into porous Co-Cr-
Mo alloy implants, Clin. Orthop., 193, 271, 1985.
6. Noyes, F.R., Grood, E.S., and Torizilli, P.A., Current concepts review. The denition of terms for
motion and position of the knee and injuries of the ligaments, J. Bone Joint Surg., 71B, 465, 1989.
7. Miller, D., Saw Bones 1994, Pacic Research Labs, Washington, 1994, 22.
8. Crowninshield, R.D., Johnston, R.C., Andrews, J.G., and Brand, R.A., Biomechanical investigation
of the human hip, J. Biomech., 11, 75, 1978.
9. Walker, P.S. and Robertson, D.D., Design and fabrication of cementless hip stems, Clin. Orthop., 235,
25, 1988.
FIGURE 35.4 Load transfer at the wedge cylinder interface in uncemented stems. (A) With a porous coat,
shear stress (the force tangent to the wedge surface, divided by the area on which the force acts; shown by
the angled arrows) is developed to equilibrate the compressive force which is the weight applied. (B) Since
there is no bony ingrowth, initial equilibrium relies on the magnitude of the vertical component (N) of the
normal force (N, the force perpendicular to the surface of the wedge) of the compressive interface stress. The
wedge must subside to develop the required compressive stress values. (From Naidu, S.H. et al., Am. J. Orthop.,
25, 829, 1996. With permission.)
0266/ ch35/frame Page 549 Friday, October 22, 1999 10:54 AM
550 Mechanical Testing of Bone and the BoneImplant Interface
10. Gilbert, J.L., Bloomeld, R.S., Lautenschlager, E.P., et al., An in vitro investigation of the migration
and micromotion of custom and off-the-shelf cementless prostheses, Trans. Orthop. Res. Soc., 16,
543, 1991.
11. Huiskes, R., The various patterns of press t, ingrown, and cemented femoral stems, Clin. Orthop.,
261, 27, 1990.
0266/ ch35/frame Page 550 Friday, October 22, 1999 10:54 AM

551

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

36

Screw Pullout Testing

Lisa A. Ferrara and Timothy C. Ryken

CONTENTS

I. Introduction ..........................................................................................................................551
II. Screw Anatomy ....................................................................................................................552
III. Screw Mechanics..................................................................................................................554
A. The Mechanics of Materials ..........................................................................................554
B. Screw Strength ...............................................................................................................554
C. Holding Strength of a Screw .........................................................................................554
D. Increasing the Holding Strength ....................................................................................555
1. Thread Depth.............................................................................................................556
2. Depth of Screw Penetration ......................................................................................557
3. Unicortical vs. Bicortical Purchase...........................................................................557
4. Triangulation of the Screws ......................................................................................557
E. Torque Insertion .............................................................................................................557
F. Thread Angulation..........................................................................................................558
G. Bone Mineral Content ....................................................................................................558
H. Screw Performance Based on Implant Material............................................................559
IV. Experimental Methodology for Screw Pullout Testing.......................................................559
A. Testing Apparatus...........................................................................................................559
B. Gripping Fixtures ...........................................................................................................560
C. Specimen Preparations ...................................................................................................560
1. Synthetic Bone Specimens........................................................................................560
2. Bone Specimens ........................................................................................................561
D. Screw Insertion Technique.............................................................................................562
E. Testing Protocol..............................................................................................................563
F. Grouping Specimens ......................................................................................................563
G. Data Analysis..................................................................................................................563
References ......................................................................................................................................564

I. INTRODUCTION

The bonescrew interface is a critical component of spinal stabilization. In general, bone has a
smaller modulus of elasticity than the screw and is therefore weaker than the various metals used
in screw constructs. Screw loosening as a result of bony failure can signicantly compromise the
stabilizing effect of the implant.

1,2

Differences between the elastic moduli of bone and metal can
have detrimental effects regarding stability. Placement of a signicantly stiffer penetrating implant
into bone disperses the forces nonuniformly, and regions of increased stress result within the screw
and within the bone. Eventually, the bone will fail due to microfracturing of the trabeculae or
fracturing within the cortical bone, or the screw will break at the region of peak stress. Therefore,
the quality of the bone and the biomechanics of screw xation must be respected when using

0266/ ch36/frame Page 551 Friday, October 22, 1999 11:01 AM

552

Mechanical Testing of Bone and the BoneImplant Interface

instrumentation to stabilize an injury. This symbiotic relationship must exist between the bone and
the screw for a successful fusion or stabilization to occur.
To enhance screw purchase to the bone and increase resistance to pullout, certain design features
of the screw have been implemented. Screw diameter, in particular the major diameter, screw design,
insertion techniques, and bone integrity all affect the strength of the screw xation.

3-7

The clinical
objective is, theoretically, to increase the volume of bone the screw can hold between the threads.
This relationship of bone volume to screw thread contributes to the degree of purchase the screw
has to the bone.
Many types of screws are used in bone xation. The standard machine screw design is used for
hard materials that are less compressible, such as metals and cortical bone. Cortical screws are used
in hard, compact, dense bone and possess the machine screw design with shallow threads that have
approximately a 60

thread angle.

2

Tapping the insertion hole for cortical screw placement is optimal
to avoid microfracturing within the bone and increased insertional torques which could result



in screw
breakage.

1,2,6,8

The self-tapping machine screw is used for minimizing the steps during cortical bone
screw insertion while optimizing the cutting capabilities. Self-tapping screws have a tapered outside
diameter to eliminate additional steps of bone preparation to receive the screw. These screws tap
during insertion and have leading-edge utes to guide the debris and maintain it within the conne
of the screwbone interface region. The self-tapping machine screw is conventionally used in cortical
bone and provides similar stabilizing properties as screws that are placed after tapping.

1,9

The standard wood screw design is used for softer compressible materials such as cancellous
bone. Its deep thread design and tapering permits easy insertion into the material. This design forms
threads within the material by compressing the surrounding material during the insertion process.

6

Tapping is not recommended in cancellous bone because it can weaken the boneimplant interface
and decrease the pullout strength of the screw.

9,10

Screw insertion into cancellous bone and resistance to pullout requires that an optimal amount
of bone be captured between the screw threads. The structural characteristics of the screw and the
integrity of the cancellous bone can signicantly alter the performance of the screw. Enhancement
of the screw performance within the bone can be achieved by altering the thread depth, shape, and
interthread distance of the screw. This enables a larger volume of bone to surround the screw,
thereby increasing the resistance to backout.

II. SCREW ANATOMY

Screws function as xation devices that stabilize bony abnormalities and injuries.

3,6,11,12

Understand-
ing the biomechanical principles of screw xation and its stabilizing features, such as compression
generation, can help reduce failure rates. The function of the screw is to change rotational motion
into translational motion while providing mechanical stability to the injured site.

6,8,13

In order to
achieve correction of the deformity in bone, there are some basic design features of the screw that
contribute to maintaining the mechanical stability within bone. The main components of the screw
consist of the head, the core, and the threads (Figure 36.1).

1,2,6

Each component plays a crucial role
in the performance of the screw.

6

The head (Figure 36.1) of the screw serves to transmit the insertion torque to the core and
threads. It also functions as a stop when it comes into contact with the surface of the bone, hence
ceasing the translational motion of the screw. Once the translational motion has stopped, the screw
generates a compressive force. Continued rotation of the screw can result in failure of the screw
or the material into which it is inserted.

6

The core (see Figure 36.1) is the main shaft of the screw around which the threads wrap. Geomet-
rically, the screw has a major and minor diameter. The major diameter, or outside diameter, is the largest
diameter of the screw and is measured from the outside tip of one side of the thread to the outside tip
of the other side. The minor diameter, or inner diameter, is the smallest diameter of the screw measured
across the base of the threads. It is the minor diameter that is referred to as the core of the screw.

0266/ ch36/frame Page 552 Friday, October 22, 1999 11:01 AM

Screw Pullout Testing

553

The thread (see Figure 36.1) is dened as the ridge in the form of a helix on the external or
internal surface of a cylinder that does the work of an inclined plane.

1,2

Pitch (Figures 36.1 and 36.2) is the distance between the threads on the screw, provided the
measurements are taken consistently from the base or apex of one thread to the exact point on the
adjacent thread. Pitch can also be dened as the inverse of the number of threads per inch (T.P.I.):
Pitch = 1/T.P.I. (36.1)

FIGURE 36.1

Anatomy of the screw. The three main components of the screw are the head, core (minor
diameter), and thread. The major diameter is the outer diameter of the screw measured at the crest of the
threads. Length of the engagement of the screw is measured by the number of threads seated in the material.

FIGURE 36.2

The pitch is the distance between the threads of the screw and can be calculated from the
inverse of the number of threads per inch. The single depth of the thread is calculated from the root to the
crest of the thread.

0266/ ch36/frame Page 553 Friday, October 22, 1999 11:01 AM

554

Mechanical Testing of Bone and the BoneImplant Interface

The lead is the distance a screw advances with one turn.

1,2

This is equal to the pitch (distance
between threads).
Thread depth is the distance from the apex of the thread to the base of the thread (see
Figure 36.1). The depth of the thread inuences the purchase of the screw into the bone and can
signicantly inuence pullout resistance.

1,2,8,13-15

III. SCREW MECHANICS
A. T

HE

M

ECHANICS



OF

M

ATERIALS

The boneimplant interface in screw xation is the site of greatest load transfer. The load is
transferred from the head of the screw along the core and into the material in which the screw
is placed. The elasticity or stiffness of a material can be expressed as the amount of elastic
deformation that results from an applied force. This is termed the elastic modulus of the material
and describes the stress (force per unit of cross-sectional area) per unit of strain (the linear
deformation per unit of original length).

2

There are three types of elastic moduli that exist:
Youngs modulus, bulk modulus, and shear modulus. All are dependent on the type of loading
the material endures.
A high modulus of elasticity implies a stiffer material. A low modulus of elasticity suggests
that the material is less rigid. Given this information, bone screws are commonly constructed from
various grades of titanium or 316L stainless steel with high moduli of elasticity ranging from 100
to 200 GPa.

2

Steel and titanium are roughly eight times stiffer than bone, whose modulus of elasticity
is approximately 20 GPa. A material whose modulus of elasticity is similar to that of a screw



will
transfer the force to the screw core. As the core is deformed from the loading, a force is generated
between the screw head and the material, creating a region of high stress localized at the head of
the screw. When a material with a lower modulus of elasticity is used with screw insertion, the
material is deformed, transferring the force between the head of the screw and the material.

6,14

Applying this to bone suggests that any loosening that occurs at the bonescrew interface will
result in loss of compression, leading to loss of xation.

B. S

CREW

S

TRENGTH

Complex forces act upon a screw once it is inserted into bone (Figure 36.3). Therefore, it is essential
that the screw be of sufcient strength to accommodate these forces. The bending strength of a
screw can be expressed in terms of the section modulus (

Z

).

1,16

Z

= (

D

3

)/32 (36.2)
In bending, the strength of a screw is proportional to the third power of its minor diameter.

1,2

Thus, the core diameter of a screw signicantly affects the bending strength of the screw. A small
increase in the diameter can yield a large increase in the bending strength of the screw.

C. H

OLDING

S

TRENGTH



OF



A

S

CREW

The holding power of a screw in bone is associated with the mechanical properties of the screw
itself and the shear strength of the bone.

17

The main constituents of screw pullout resistance with
respect to screw design are thread depth and the major diameter.

18

Other variables, however,
contribute to the pullout behavior of the screw, such as the extent of cortical purchase, depth of
screw penetration, thread angulation, pitch diameter, screw placement within the bone, physical
changes to the screw or bone between the time of insertion and the time of withdrawal, speed at
which the screw is withdrawn, the occurrence of predrilled holes, and the quality of the

0266/ ch36/frame Page 554 Friday, October 22, 1999 11:01 AM

Screw Pullout Testing

555

bone.

6,8,10,14,19,20-22,24

Consideration of all of these variables will minimize screw failure and loss of
bone stabilization. When a screw is inserted into a material such as bone, pullout failure of the
screw implies failure of the bone as well. As the screw is loosened and begins to back out, the
bone breaks and the screw toggles within the bone yielding a large void. The trabecular structure
within cancellous bone forms a matrix. When the screw toggles, it breaks the individual trabeculae,
disrupting the natural matrix, and enlarges the void between each trabecula. However, in soft
uniform materials, the void is smaller and similar in size to the outer diameter of the screw. It is
a concept best visualized by placing a screw into a soft uniform material, such as wax (Figure 36.4).
When the screw is pulled out, a smooth cylinder is left inside the block of wax due to the screw
threads shearing the wax from the side wall as it is pulled out.

8,16

The area of the hole is that of a
cylinder; area = depth

perimeter. Therefore, the maximum holding strength a screw can have is
(area of the cylinder)

(shear strength of the material).

D. I

NCREASING



THE

H

OLDING

S

TRENGTH

The repetitive cyclical loading from motion in the human body can contribute to screw loosening
in bone. Many factors can affect the loss of screw xation in bone.

1,2,10,14,20-25

Failures can be
attributed to poor screw design, application of an improper screw type for the material of choice,
misconception of the forces applied on the screw, shallow screw threads that do not grasp a sufcient
amount of bone to prevent backout, and other limiting factors that do not respect the symbiotic
relationship that must exist between the screw and the material. Fortunately, alterations in screw
design and an understanding of the mechanical principles of the material used for screw xation
can provide better resistance to screw loosening and reduce the risk of stabilization failure. These
issues are discussed in detail below.

FIGURE 36.3

The forces acting upon a screw thread.

= the frictional coefficient between the screw and
the bone,

F

b

= the reaction force of the bone to the inclined plane of the screw thread,

= thread angle,

d

m

=
mean total width of the screw,

Pxd

m

= bone surface transversed by the screw thread in one rotation,

F

t



= total
reaction force of the bone to the screw insertion,

F

s

= axial insertion force of the screw,

P

= pitch.

0266/ ch36/frame Page 555 Friday, October 22, 1999 11:01 AM

556

Mechanical Testing of Bone and the BoneImplant Interface

1. Thread Depth

Screw pullout resistance can be greatly inuenced by the thread design (Figure 36.5). The volume
of bone the threads capture contributes to the pullout performance of the screw. The pitch can alter
the amount of bone captured by changing the distance between one thread and the adjacent thread.
The thread depth can alter the bone volume obtained by modifying the length of the thread. A
deeper thread allows more bone to reside between each thread and increases pullout resistance.

FIGURE 36.4

Screw pullout from a soft homogenous material. After screw withdrawal, a smooth cylinder
is left within the wax block due to shearing of the screw threads from the side walls of the wax.

FIGURE 36.5

The difference in thread depth can significantly alter the pullout strength of a screw. Altering
the thread depth while keeping the pitch constant will change the amount of bone that surrounds the threads
of the screw. A greater pullout resistance exists for greater thread depth (A). Less bone surrounds the screw
in (B); hence, pullout resistance is decreased.

0266/ ch36/frame Page 556 Friday, October 22, 1999 11:01 AM

Screw Pullout Testing

557

2. Depth of Screw Penetration

The length of the screw has a dramatic effect on the pullout strength (Figure 36.6). Deeper screw
insertion has been shown to increase pullout resistance and strength in exion and extension
loading.

26

The benets of using a longer screw must be balanced against increased operative risk.
Radiographic assessment may help reduce this possibility.

3. Unicortical vs. Bicortical Purchase

Since cortical bone is denser than cancellous bone and has a signicantly greater pullout strength,
studies have shown bicortical screw purchase to be superior in pullout strength to unicortical screw
purchase.

4,22,27

The two layers of cortex surrounding the cancellous bone provide greater stability
to the screw with respect to wobbling and pullout than when only one cortical surface is purchased
(Figure 36.6).

27

4. Triangulation of the Screws

Bone is a heterogeneous material. The trabecular architecture of cancellous bone varies and tends
to be denser proximal to the cortical margins.

28

Triangulation of the screws into these concentrated
regions of bone will signicantly increase pullout resistance.

2,3,29-31

The toeing-in of the screws
prevents backing out along the direction of the applied axial force. It also purchases a larger wedge
of bone between the two screws requiring more force to pull the screws out (Figure 36.7).

3,29,31

Triangulated constructs are biomechanically benecial and offer stronger opposition to the forces
that cause the screw to back out.

E. T

ORQUE

I

NSERTION

The torsional forces during screw insertion have been shown to correlate with pullout strength.
Some studies have shown that high insertional torques measured during screw insertion into material
correlate very well with high pullout strengths for certain screws and uniform materials such as
synthetic bone blocks and calf vertebrae.

7,14

However, pullout studies conducted in cadaver bone
did not yield strong correlations between insertional torque and pullout strength.

21,32

The peak torque that is generated during screw insertion can depend on screw type, screw
diameter, screw design, and whether the screw hole was tapped prior to insertion.

12,21,24,32,33

A screw
with a larger diameter in composite bone is more likely to engage the region of denser cortical
bone. In contrast to this, a smaller diameter screw placed in composite bone would yield lower

FIGURE 36.6

Increased depth of screw penetration is achieved with longer screws. Screw pullout strength
is signicantly increased with longer screw penetration for both a unicortical or bicortical purchase. However,
bicortical screw purchase is superior to unicortical purchase in pullout performance. The ideal situation is the
use of a longer screw to obtain bicortical purchase, provided there are no neurological risks.

0266/ ch36/frame Page 557 Friday, October 22, 1999 11:01 AM

558

Mechanical Testing of Bone and the BoneImplant Interface

insertional torques because it primarily engages cancellous bone.

10,21,32

Tapping of the screw hole
prior to insertion also yields lower peak torques during insertion.

21,34

F. T

HREAD

A

NGULATION

The thread angle denes the shape of the thread (Figure 36.8). This is measured as an angle that
deviates from a perpendicular line drawn to the shaft of the screw.

2

Altering this angle changes the
shape of the threads and alters the amount of bone that the screw grasps directly. The larger the
bone volume the screw carries within each thread, the greater the pullout resistance. However, there
is a balance between optimal thread depth, shape, pitch, and thread angle. A large thread depth
with few threads per inch may not provide optimal resistance to screw backout.

G. B

ONE

M

INERAL

C

ONTENT

When a metal screw is inserted into bone, the shear strength of the screw is considerably greater
than the bone. Thus, it is the bone that fails initially in pullout testing. The more dense the bone,
the smaller the risk of bone failure, and the higher the pullout resistance.

3,17,31

A bending force
applied to a screw in any material creates a stress concentration at the initial two or three threads
of the screw. This region is often the area where screw breakage or failure (pullout) will occur.

1,2

It is the supercial two or three threads of the screw that are responsible for the transfer of the

FIGURE 36.7

(A) Triangulation of the screws, or toeing-in, provides increased screw purchase and resis-
tance to pullout. Angulation of the screw resists backing out along the direction of the applied load. Further-
more, a large wedge of bone surrounds each of the angled screws, which increases the resistance to screw
pullout. (B) If the two triangulated screws are cross-linked, there is an even greater resistance to pullout. It
is far more difcult to pull out bilateral screws simultaneously. If the screws are angled toward the cortical
margins while xed to a plate, greater screw purchase will be acquired due to the increased density of the
bone as it approaches the cortex. Resistance to pullout will be greater as well, due to the larger quantity of
bone surrounding the angled screw.

0266/ ch36/frame Page 558 Friday, October 22, 1999 11:01 AM

Screw Pullout Testing

559

load to the bone. Therefore, thread conguration and bone density correlate well with pullout
resistance, especially at the bonescrew interface.

H. S

CREW

P

ERFORMANCE

B

ASED



ON

I

MPLANT

M

ATERIAL

Screw material can signicantly impact performance. Currently, 316L stainless steel with a
480-MPa ultimate tensile strength is the conventional material used in biological implants.

2

Titanium
has gained popularity due to its resistance to corrosion within the human body and the similarities
in terms of mechanical properties to that of stainless steel.
Bioabsorbable screws have been developed in an effort to increase patient rehabilitation time
postoperatively. Polylactic acid is a common component of these types of screws.

In vivo

degra-
dation of these screws occurs over time by the process of hydration, depolymerization, loss of
morphological supporting structure, absorption, and elimination.

33

Pena et al.

33

conducted a tensile
pullout study on bioabsorbable screws and compared their results with metal screws. The metal
screws performed favorably and had signicantly higher insertional forces than the bioabsorbable
screws. They also demonstrated signicantly greater failure loads than the bioabsorbable screws.

IV. EXPERIMENTAL METHODOLOGY FOR SCREW PULLOUT TESTING

The purpose of quantifying the pullout strength of screws is to measure their ability to attach and
hold an object.

35

There are several ways to quantify screw performance and the holding strength
of the screw implant. Quantication of the tensile forces required to pull a screw out of a particular
material determines the pullout strength of that screw. Measurement of the axial forces necessary
for insertion of the screw as it is tightened into a material will also describe biomechanical
characteristics of the screw.

A. T

ESTING

A

PPARATUS

Any testing apparatus is suitable to perform a screw pullout study, provided it has a testing platform
that can move at a constant rate and has an accuracy of

1% after calibration.
35
A uniaxial materials
testing system capable of operating in tension and compression that is hydraulically or electrically
controlled is the highest-quality testing apparatus.
FIGURE 36.8 The thread angle defines the shape of the thread. Altering the thread angle changes the shape of
the thread and the amount of bone that will be captured. In turn, this will alter the pullout strength of the screw.
0266/ ch36/frame Page 559 Friday, October 22, 1999 11:01 AM
560 Mechanical Testing of Bone and the BoneImplant Interface
B. GRIPPING FIXTURES
The xtures that are used to grip the screw and the material into which it is inserted can dramatically
affect the results of a pullout study. According to the American Society for Testing and Materials
standards (ASTM), a true pullout assessment of a screw would require that the grips used to grasp
the screw head be shaped to t accurately and that the grips provide a true axial load.
35
An ideal
withdrawal study performed on screws should allow the grips to pivot or toggle. This would reduce
any residual stresses that may occur within a rigid gripping device and diminish the transfer of these
stresses to the screw during the pullout. Universal ball joints and pinned joints are commonly used
as mounting devices because of their mobility, standard design, and ease of use (Figure 36.9). Com-
patibility of these joints to standard machinery makes them adaptable to any materials testing system
with the aid of attachment xtures. However, there are pullout studies that do not require gripping
devices to toggle during tensile loading. An example of this would be a triangulated screw study using
xtures that toggle. Hence, the effect of angulation observed on the pullout strength would be reduced.
C. SPECIMEN PREPARATIONS
1. Synthetic Bone Specimens
Often materials of uniform density are used during screw pullout studies to provide a standardized
experiment. Currently, rigid polyurethane foams are used as bone substitutes for pullout studies.
The uniformity of the material and consistent mechanical properties similar to that of human bone
make it a good substitute for screw strength quantication. It can be purchased in a variety of densities
that represent good bone integrity, as well as osteoporotic bone models. The uniform nature eliminates
FIGURE 36.9 Illustration of pullout fixtures: (A) universal ball joint for the top grip with the bottom grip
xed to the platen of the testing apparatus, allowing the top grip to toggle during the application of tensile
forces, and (B) universal ball joint for the top and bottom grip allowing both to toggle during the application
of tensile forces. Use of these devices eliminates residual stresses that occur within the pulling xtures
during tension.
0266/ ch36/frame Page 560 Friday, October 22, 1999 11:01 AM
Screw Pullout Testing 561
the variability that exists between normal bone samples and provides a standardized method for
determining screw pullout resistance. Therefore, it can provide an accurate assessment of the pullout
strength of the screw. Use of synthetic bone is limited in its clinical applicability because bone is
an inhomogeneous substance in most regions of the body.
A general rule of thumb when determining the size of the synthetic bone sample is to maintain
a constant pattern of failure within the synthetic bone by keeping the screw diameter approximately
less than 5% of the circumference of the synthetic bone block.
11,19
When mounting the synthetic
bone into the bottom grip on the testing apparatus, it is crucial to apply a uniform pressure across
the material (Figure 36.10). This can be done by using plates that are similar in size to the synthetic
bone block attached to the clamping arms of the vise. The plates distribute the forces applied by
the vise screws over a larger surface area and transmit these forces uniformly across the block. The
ideal situation is to have four plates clamping four sides of the block for a uniform application of
pressure. This eliminates discrepancies in the screw pullout data due to regions of high stresses
within the synthetic bone blocks caused by nonuniform clamping.
2. Bone Specimens
In order to assess the clinically relevant nature of screw xation, pullout experiments are
commonly conducted in various types of bone. Bovine, porcine, and unembalmed cadaver bone
FIGURE 36.10 Diagram of the test sample mounted into clamping fixtures. These fixtures are positioned
into the machine grips and a tensile force is applied at a uniform rate. The platens separate and the top grip
pulls the screw out of the material. The load cell of the testing apparatus records the force required to separate
the screw from the material. The peak force during pullout usually occurs within the rst few threads that are
pulled out of the material.
0266/ ch36/frame Page 561 Friday, October 22, 1999 11:01 AM
562 Mechanical Testing of Bone and the BoneImplant Interface
are common bone sample types. However, bone is not homogeneous in composition. In the
vertebral bodies, the bone density is signicantly higher at the cortical margins and lowest in
the center of the body.
28
Human cortical bone, such as that found in the long bones of the
appendages, is extremely dense and can generate screw pullout strengths approximately four
times greater than that of cancellous bone.
7,17,18,20,36-39
It is essential to prepare each of the bone samples meticulously, especially when using cadaver
bone samples.
38
All soft tissue should be cleanly dissected off the bone in order to provide a better
gripping surface. The oils secreted from the musculature can cause the bone sample to slip out of
the grips during the tensile pull. Although Kaab et al.
40
found no signicant changes in pullout
strength with cadaver bone that was allowed to undergo structural degradation caused by time-
dependent autolytic processes, Simonian et al.
5
found that the process of freeze-drying the bone
signicantly weakened the pullout resistances of the screw. Therefore, it is best to conduct pullout
studies with bone specimens that have been subjected to minimal thawing and freezing cycles and
have not undergone any processing treatments.
A signicant factor in the holding strength of bone instrumentation is the quality of the bone
itself. There are many techniques currently used to assess the integrity of the bone samples.
4,36
Dual
X-ray absorptiometry (DEXA), quantitative computed tomography (QCT), and magnetic resonance
imaging (MRI) are used clinically to study the bone mineral density for studies on osteoporo-
sis.
4,19,20,22,28
DEXA is a common scanning method for determining bone quality because of its ease
of use, speed, and accurate assessment. However, it provides an overall average bone mineral
density (BMD) for the regions of bone measured. QCT is another common method used for
determining BMD. When bone specimens are scanned with this technique, it is suggested that a
phantom be used with each scan taken. The advantage of this technique is the ability to measure
distinct regions of bone for the quantication of BMD. This can yield a more localized BMD for
the region where the screw will be placed. Proper CT algorithms should be used for cancellous
and cortical bone and the regions measured within the bone can be calibrated against the phantom
for each scan. MRI is the most accurate method for BMD quantication and can be used to
characterize trabecular architecture.
36
The inhomogeneity of tissue induces signal intensity differ-
ences in response to the magnetic stimulation; thus differences in the trabeculae can be visualized
and quantied for specic regions of interest.
The manner in which the specimens are secured into the bottom xture that sits in the grip of
the testing apparatus can affect the pullout resistances of the screw.
38
The xture is responsible for
holding the specimen in place and resisting the tensile forces placed on the screw. In contrast, the
grip is attached to the testing apparatus and holds onto the xture housing the testing specimen.
Typically, the specimens are secured in a vise and may initially be embedded in a polyester resin
(see Figure 36.10). The resin provides additional support to the bone specimen and can prevent
rotation of the specimen when clamped within the vise. It also serves to distribute the axial forces
evenly across the surface between the specimen and xture by allowing a greater surface area. The
extensive porosity of bone can prompt intrusion of the resin into it. Pfeiffer et al.
38
demonstrated
this occurrence by examining the trabecular architecture after failure, and characterizing the bio-
mechanical behavior of screw pullout in bone that had been invaded with resin. Their data dem-
onstrated signicantly greater pullout resistances with bone that had been intruded by the resin
than those without resin intrusion. Therefore, if an embedding material is to be used, the bone
specimens should be wrapped in a substance that will not allow the material, such as latex or plastic
wrapping, to seep into the bone.
D. SCREW INSERTION TECHNIQUE
To initiate screw insertion, the screw holes can be drilled with bits that are approximately 10%
larger than the minor diameter of the screw without compromising pullout strength.
11,19
If the
0266/ ch36/frame Page 562 Friday, October 22, 1999 11:01 AM
Screw Pullout Testing 563
bone is cortical bone or has a cortical shell, it is suggested that the cortex be drilled and tapped
to avoid microfracturing at this layer. Since the cortical shell is usually at the bonescrew interface
and may surround the initial few threads of the screw, microfractures could reduce the pullout
strength. Guides must be used to ensure the proper orientation of the screw within the bone. If
the screws are to be placed at an angle, special guides must be constructed to ensure the screw
will follow the correct insertion angle. Radiographs can document the screw orientation and
screw depth within the test sample. After insertion, the peak insertional torque should be
quantied by using a nely graded torque wrench.
E. TESTING PROTOCOL
The test specimen should be placed onto the loading platform of the testing machine into a top
and bottom grip that is in perfect alignment. This avoids bending moments and additional stresses
transferred to the screw as a result of misalignment. Depending on the type of testing apparatus,
the protocol for screw pullout should be withdrawal of the screw using an axial force in tension
moving at a constant rate. This causes the platens on which the grips are secured to separate at
a uniform speed. For material testing systems such as an MTS model (MTS Systems, Inc.,
Minneapolis, MN), the test should be performed under stroke control (actuator displacement).
According to ASTM standards,
35
the basic speed of withdrawal should be the application of the
tensile force throughout the test at a uniform rate of platen separation of 0.10 in./min 25%
(2.54 mm/min).
35
Some common pull rates range from 0.1 to 5 mm/s depending on the type of
screw and the material into which it is inserted.
4,19,34,38
It is recommended that a screw be used
only once within a pullout study. Repetition of pullout on a screw can weaken and dull the
threads.
35
Placement of this used screw into another piece of bone will reduce its purchase
capability and may decrease its pullout resistance.
F. GROUPING SPECIMENS
Randomization of the test specimens, especially if bone samples are used in the study, is
recommended to ensure an unbiased statistically relevant study. Differences in gender, age, bone
densities, anatomical alignment, and size can all be randomized for each testing group. Test
samples should be selected to cover an appropriate range of densities and properties that would
be clinically pertinent. The number of tests conducted should be extensive enough to provide
reliable results. ASTM standards suggest that the coefcient of variation from screw testing
ranges from 15 to 30%.
35
An optimal sample size of pullout testing for each group of screws is
to conduct at least ten replications. Often, a power analysis can be conducted to validate the
sample size.
G. DATA ANALYSIS
Statistical comparisons between testing groups and the pullout results commonly use a Students
t-test analysis and/or a multivariate analysis of variance to detect differences between the testing
groups and the corresponding results. To dene the relationship among bone mineral density,
insertional torque, and pullout force, a regression analysis or multivariate regression analysis should
be conducted. The regression analysis is dened here on the amount of linearity that the pullout
force has when compared with one of the other variables, such as bone mineral density. It will
determine if there is a strong statistical association between bone mineral density and screw pullout
strength. An r value closer to 1 characterizes a strong relationship between the two variables,
whereas an r value of 0 indicates no correlation. A multivariate regression analysis will evaluate
the relative contribution of both dependent variables, bone mineral density and insertion torque,
on the variability observed with the independent variable, pullout force.
0266/ ch36/frame Page 563 Friday, October 22, 1999 11:01 AM
564 Mechanical Testing of Bone and the BoneImplant Interface
REFERENCES
1. Bennett, G.J., Materials and materials testing, in Spinal Instrumentation, Benzel, E.C., Ed., American
Association of Neurological Surgeons, Chicago, 1994, 31.
2. Benzel, E.C., Implantbone interfaces, in Biomechanics of Spine Stabilization: Principles and Clinical
Practice, Benzel, E.C., Ed., McGraw-Hill, New York, 1995, 127.
3. Ruland, C.M., McAfee, P.C., Warden, K.E., and Cunningham, B.W., Triangulation of pedicular
instrumentation. A biomechanical analysis, Spine, 16, S270, 1991.
4. Ryken, T.C., Clausen, J.D., Traynelis, V.C., and Goel, V.K., Biomechanical analysis of bone mineral
density, insertion technique, screw torque, and holding strength of anterior cervical plate screws,
J. Neurosurg., 83, 325, 1995.
5. Simonian, P.T., Conrad, E.U., Chapman, J.R., et al., Effect of sterilization and storage treatments on
screw pullout strength in human allograft bone, Clin. Orthop., 302, 290, 1994.
6. Uhl, R.L., The biomechanics of screws, Orthop. Rev., 18, 1302, 1989.
7. Zdeblick, T.A., Kunz, D.N., Cooke, M.E, and McCabe, R., Pedicle screw pullout strength. Correlation
with insertional torque, Spine, 18, 1673, 1993.
8. Asnis, S.E., Ernberg, J.J., Bostrom, M.P., et al., Cancellous bone screw thread design and holding
power, J. Orthop. Trauma, 10, 462, 1996.
9. Boyle, J.M., Frost, D.E., Foley, W.L., and Grady, J.J., Torque and pullout analysis of six currently
available self-tapping and emergency screws, J. Oral Maxillofac. Surg., 51, 45, 1993.
10. Phillips, J.H. and Rahn, B.A., Comparison of compression and torque measurements of self-tapping
and pretapped screws, Plast. Reconstr. Surg., 83, 447, 1989.
11. Pratt, W.B. and Yazdani, S., Laboratory testing of bolts and screws in cancellous bone, Orthop. Rev.,
18, 1073, 1989.
12. Wittenberg, R.H., Lee, K.S., Shea, M., et al., Effect of screw diameter, insertion technique, and bone
cement augmentation of pedicular screw xation strength, Clin. Orthop., 296, 278, 1993.
13. Chapman, J.R., Harrington, R.M., Lee, K.M., et al., Factors affecting the pullout strength of cancellous
bone screws, J. Biomech. Eng., 118, 391, 1996.
14. Daftari, T.K., Horton, W.C., and Hutton, W.C., Correlations between screw hole preparation, torque
of insertion, and pullout strength for spinal screws, J. Spinal Disord., 7, 139, 1994.
15. Horton, W.C., Blackstock, S.F., Norman, J.T., et al., Strength of xation of anterior vertebral body
screws, Spine, 21, 439, 1996.
16. Brod, J.J., The concepts and terms of mechanics, Clin. Orthop., 146, 9, 1980.
17. Stromsoe, K., Kok, W.L., Hoiseth, A., and Alho, A., Holding power of the 4.5 mm AO/ASIF cortex
screw in cortical bone in relation to bone mineral, Injury, 24, 656, 1993.
18. Skinner, R., Maybee, J., Transfeldt, E., et al., Experimental pullout testing and comparison of variables
in transpedicular screw xation. A biomechanical study, Spine, 15, 195, 1990.
19. DeCoster, T.A., Heetderks, D.B., Downey, D.J., et al., Optimizing bone screw pullout force, J. Orthop.
Trauma, 4, 169, 1990.
20. Halvorson, T.L., Kelley, L.A., Thomas, K.A., et al., Effects of bone mineral density on pedicle screw
xation, Spine, 19, 2415, 1994.
21. Kohn, D. and Rose, C., Primary stability of interference screw xation. Inuence of screw diameter
and insertion torque, Am. J. Sports Med., 22, 334, 1994.
22. Maiman, D.J., Pintar, F.A, Yoganandan, N., et al., Pullout strength of Caspar cervical screws, Neuro-
surgery, 31, 1097, 1992.
23. Ellis, J.A. and Laskin, D.M., Analysis of seating and fracturing torque of bicortical screws, J. Oral
Maxillofac. Surg., 52, 483, 1994.
24. Schwimmer, A., Greenberg, A.M., Kummer, F., and Kaynar, A., The effect of screw size and insertion
technique on the stability of the mandibular sagittal split osteotomy, J. Oral Maxillofac. Surg., 52,
45, 1994.
25. Soshi, S., Shiba, R., Kondo, H., and Murota, K., An experimental study on transpedicular screw
xation in relation to osteoporosis of the lumbar spine, Spine, 16, 1335, 1991.
26. Krag, M.H., Beynnon, B.D., Pope, M.H., and DeCoster, T.A., Depth of insertion of transpedicular
vertebral screws into human vertebrae: effect upon screwvertebra interface strength, J. Spinal Disord.,
1, 287, 1988.
0266/ ch36/frame Page 564 Friday, October 22, 1999 11:01 AM
Screw Pullout Testing 565
27. Ryken, T.C., Goel, V.K., Clausen, J.D., and Traynelis, V.C., Assessment of unicortical and bicortical
xation in a quasistatic cadaveric model. Role of bone mineral density and screw torque, Spine, 20,
1861, 1995.
28. Keaveny, T.M. and Hayes, W.C., A 20-year perspective on the mechanical properties of trabecular
bone, J. Biomech. Eng., 115, 534, 1993.
29. Baldwin, N.G. and Benzel, E.C., Sacral xation using iliac instrumentation and a variable-angle screw
device. Technical note, J. Neurosurg., 81, 313, 1994.
30. Benzel, E.C. and Baldwin N.G., Crossed-screw xation of the unstable thoracic and lumbar spine,
J. Neurosurg., 82, 11, 1995.
31. Hadjipavlou, A.G., Nicodemus, C.L., Al-Hamdan, F.A., et al., Correlation of bone equivalent mineral
density to pullout resistance of triangulated pedicle screw construct, J. Spinal Disord., 10, 12, 1997.
32. Kwok, A.W., Finkelstein, J.A., Woodside, T., et al., Insertional torque and pullout strengths of conical
and cylindrical pedicle screws in cadaveric bone, Spine, 21, 2429, 1996.
33. Pena, F., Grontvedt, T., Brown, G.A., et al., Comparison of failure strength between metallic and
absorbable interference screws. Inuence of insertion torque, tunnel-bone block gap, bone mineral
density, and interference, Am. J. Sports Med., 24, 329, 1996.
34. Hearn, T.C., Schatzker, J., and Wolfson, N., Extraction strength of cannulated cancellous bone screws,
J. Orthop. Trauma, 7, 138, 1993.
35. Standard Test Methods for Mechanical Fasteners in Wood. Annual Book of ASTM Standards, Vol. 4,
ASTM, Philadelphia, 1999, 268.
36. Eysel, P., Schwitalle, M., Oberstein, A., et al., Preoperative estimation of screw xation strength in
vertebral bodies, Spine, 23, 174, 1998.
37. Lieberman, I.H., Khazim, R., and Woodside. T., Anterior vertebral body screw pullout testing. A
comparison of Zeilke, Kaneda, Universal Spine System, and Universal Spine System with pullout-
resistant nut, Spine, 23, 908, 1998.
38. Pfeiffer, M., Gilbertson, L.G., Goel, V.K., et al., Effect of specimen xation method on pullout tests
of pedicle screws, Spine, 21, 1037, 1996.
39. Sell, P., Collins, M., and Dove, J., Pedicle screws: axial pullout strength in the lumbar spine, Spine,
13, 1075, 1988.
40. Kaab, M.J., Putz, R., Gebauer, D., and Plitz, W., Changes in cadaveric cancellous vertebral bone
strength in relation to time. A biomechanical investigation, Spine, 23, 1215, 1998.
0266/ ch36/frame Page 565 Friday, October 22, 1999 11:01 AM
0266/ ch36/frame Page 566 Friday, October 22, 1999 11:01 AM

567

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

37

Finite Element Analysis
for Evaluating Mechanical
Properties of the
BoneImplant Interface

Keith R. Williams

CONTENTS

I. Introduction ..........................................................................................................................567
II. Modeling and Methods ........................................................................................................569
A. Axial Cantilever .............................................................................................................569
B. Dental Implants in the Mandible ...................................................................................569
C. Knee Ligament Connector in Cancellous Bone ............................................................572
III. Results ..................................................................................................................................573
A. Axial Model....................................................................................................................573
B. Mandible Model .............................................................................................................574
C. Knee Ligament Connector .............................................................................................575
IV. Discussion ............................................................................................................................577
V. Conclusion............................................................................................................................578
References ......................................................................................................................................579

I. INTRODUCTION

An example of a mechanical interaction between the implant and bone occurs following the loading
of titanium dental pins surgically implanted in the mandible. Titanium pins are now used routinely
for xed and removable bridge and single implant restorations in edentulous and partially dentate
patients. The clinical implantation technique has remained as established by Brnemark

1

and
colleagues some 20 years ago, although the actual pin design has undergone various modications
by subsequent workers. The ability of titanium to osseointegrate

2,3

into bone has resulted in very
high success rates for this type of clinical restoration. For these reasons this chapter will concentrate
mainly on assessing the mechanical stability of titanium dental implants in bone, although a static
analysis of a knee ligament prosthesis in cancellous bone is also investigated.
Osseointegration is dened as a direct contact between an implant and adjacent hard tissue.

1-3

The chemical and biological processes that take place at the interface between a solid implant and
biological tissue depend on various factors. These include the implant biocompatibility, the presence
of infection, the interval between the operation and the distribution of load to the pin, and the clinical
procedure followed. Variations in these factors can lead to different rates of osseointegration and
degrees of stability of the pin in the surrounding tissue, ranging from complete bone xation to brous
encapsulation. It is therefore important that an assessment be made of the state of pin stability before

0266/ch37/frame Page 567 Monday, October 25, 1999 10:12 AM

568

Mechanical Testing of Bone and the BoneImplant Interface

xed or removable metal work or other attachments are made to titanium implants. Indeed, a non-
invasive technique whereby an assessment of the xity of the implant in tissue with time would appear
an ideal clinical situation. It is believed that the natural frequency of vibration of the implant in the
surrounding tissue offers the best method of estimating this interaction. However, natural frequency
analysis alone provides no indication of the type and quality of bone supporting the implant. X-ray
radiography does not offer a high enough resolution to be able to indicate the bone quality at the
interface; hence, the current adoption of the nite-element approach to the problem.
Certain noninvasive methods for the evaluation of fracture healing of long bones have been
developed over the past 10 years

4-7

relying on vibrational methods

8-10

for the assessment of bone
healing with time. Essentially, two techniques have been devised; in the rst, the resonances are
measured by the impedance probe technique, while the second relies on impulse excitation by
hammer impact. Both techniques give good reproducibility when the experimental conditions are
idealized in the laboratory on cadaver material.

5

A similar approach has been attempted within the broad area of dental treatment. The equipment
developed again makes use of an impulse excitation in order to assess the state of periodontal
health. The test relies on measurement of the deceleration of an impacting probe on the surface of
the tooth and provides a subjective measure of the viability of the periodontium. This test method
may be useful in assessing the preliminary stages of osseointegration since it is probable that the
material properties of the initial amorphous zone surrounding the implant may be similar to those
of the periodontium

11

resulting in similar displacements and decelerations. However, when an
implant is supported by bone with signicantly higher elastic moduli and density, the probe method
becomes unreliable. This unreliability may be due to the low frequency of probe excitation (4 Hz).
Thus, vibrational methods are being developed

12

relying on higher excitation frequencies in order
to measure implant xity in bone. Indeed, natural frequency analysis of the implant in tissue offers
a more sensitive means of measuring the implant movement following the application of physio-
logically acceptable excitation loads.
These vibrational techniques have been simulated using nite-element methodologies.

13,14

Results from these computer models suggest the natural frequencies of vibration of the implant are
extremely sensitive to the xity conditions at the implantbone interface or the degree of osseo-
integration. These simulation studies provide a reasonable understanding of the range and impor-
tance of the variables that require measuring experimentally and clinically in order to allow an
assessment of bone healing and eventual osseointegration following surgery. The nite-element
method has also been used to study the behavior of endosseous implants in tissue and related
problems,

15-22

as well as the biomechanics of the osseointegration process.

23-26

In this work, the natural frequencies and mode shapes of vibration of a single titanium implant
xed in various media and the mandible have been calculated using a cantilever bar attached to
the dental pin and excited using an attached piezoelectric plate.

27

Additionally, due to the clinical
complexities in attaching such a cantilever

in situ,

a simpler impulse technique has also been
examined in which the dental pin was excited directly by hammer impact and the resulting dynamic
response observed. In this way, the natural frequencies of vibration may be identied with healing
time and hence the exact level of integration existing around the implant. It is possible using modern
numerical analysis techniques to discretize the various tissues and their anatomy existing around
the implant and assign various mechanical properties to these volumes in order to evaluate the
degree of osseointegration or pin xity with time. Since the discretized volumes may be of the
order of size of trabecular processes, a correlation may be established between the pin frequency
and its xity in the bone. Furthermore, the numerical procedures allow a variety of bone properties
to be input to the model ranging from isotropic to more realistic anisotropic nonlinear behavior.
In this work, both homogeneous and recently measured orthotropic properties

28

have been used. A
reasonably simple discretization of trabecular bone based on a plate model in contact with a titanium
implant has also been examined and the resultant interfacial stresses and strains compared with
recently acquired experimental data.

0266/ch37/frame Page 568 Monday, October 25, 1999 10:12 AM

Finite Element Analysis for Evaluating Mechanical Properties of the BoneImplant Interface

569

Accordingly, a variety of possible implant support conditions, bone types and properties, and
bone adaptation parameters have been set up and interfacial stresses, displacements, and natural
frequencies calculated. The sensitivity of the natural frequencies to these variable conditions have
been analyzed over a time interval sufcient to be able to offer data that lead to a clinical
interpretation of the osseointegration condition.

II. MODELING AND METHODS

The experimental methodology for the measurement of the natural frequencies of vibration of
implants in supporting media has been the subject of a recent study.

27

Essentially, the setup consists
of a small cantilever between 15 and 22 mm long and 2 by 2.6 mm in cross section, to which are
attached piezoelectric plates and which is then screwed onto the implant. The external frequency
excitation is then increased until the rst natural frequency of vibration of the pin is reached and
then recorded. Various cantilever designs have been produced, but in this work an axial and right-
angled cantilever attachment to the dental implant has been modeled.

A. A

XIAL

C

ANTILEVER

The nite-element discretization initially modeled the dental implant as an axially symmetric
arrangement using axisymmetric triangular and quadrilateral elements (Figure 37.1). Pin support
was initially provided by assigning various levels of spring xity to the threaded portion of the
implant over a range 1 to 10

6

N/mm (10

3

to 10

3

N/m). Subsequently, in order to increase the
sensitivity of the model to the support provided by the surrounding tissue a full three-dimen-
sional (3D) discretization was established using six- and eight-noded continuum elements
incorporating both axial and right-angled cantilevers (Figure 37.1 shows the axial model). The
renement of this mesh allowed the tissue properties to be changed between that of collagenous
ber to cortical bone at the implantbone interface and also allowed the incorporation of crestal
and basal bone elements adjacent to the implant in order to simulate a more realistic

in vivo

model. The volume of bone was gradually extended radially from the implant to a maximum
of 22 mm in order to simulate increasing support and the possibility of bone growth around the
dental pin during healing.

B. D

ENTAL

I

MPLANTS



IN



THE

M

ANDIBLE

A 3D discretization of a section of the human mandible is shown in Figure 37.2 with both a single
implant and an implant with right-angled cantilever attachment.
The basic geometric model reported in the present work derives from the anatomical measure-
ment of several mandibles by tomographic scanning. The mandible is assumed to be symmetric,
and accordingly only one half is discretized with the inclusion of two osseointegrated but uncon-
nected dental pins as shown in Figure 37.2. This gure illustrates the conditions for the cantilever
setup on the dental pin. The local

z

direction is along the axis of the dental pins while the

x

direction
is parallel to the crestal ridge and the

y

direction is orthoradial at the symmetry interface. The
material properties assumed for pin, cortical, and cancellous bone are taken from those indicated
earlier.

28

Fixity of the model discretization is assumed at the symmetry interface and all nodes
surrounding the condyle head.
Initially, a simple eigenvalue extraction of the discretized model was carried out in order to
provide the rst few natural frequencies as calculated for the axial model (see Figure 37.2).
However, this numerical approach is not sufciently sensitive to the local pinbone interface
condition in the current model and yields only bending modes for the complete mandible. Accord-
ingly, a dynamic analysis procedure was implemented.

0266/ch37/frame Page 569 Monday, October 25, 1999 10:12 AM

570

Mechanical Testing of Bone and the BoneImplant Interface

FIGURE 37.1

Axial cantilever attached to a dental implant. Axisymmetric and 3D discretization (mode 1
resonance dotted).

FIGURE 37.2

Section of the human mandible with single dental implants, one with attached cantilever.

0266/ch37/frame Page 570 Monday, October 25, 1999 10:12 AM

Finite Element Analysis for Evaluating Mechanical Properties of the BoneImplant Interface

571

In order to examine the natural frequency of vibration of the cantilever model, an impulse
excitation was applied at the end of the cantilever (node 8499) in the

z

direction rather than the
use of piezoelectric plates described earlier. The impulse length,

t

o

, was varied between 5

10

4

and 5

10

6

sec in order to assess the effect of impulse strength on frequency. With the cantilever
removed as indicated in Figure 37.2 impulse loads of 100 N were applied in two different
directions, the rst in an orthoradial direction (local

y

direction) and the second series in a
direction along the cortical crestal ridge (essentially local

x

direction). Under these conditions
the impulse length was varied between 10

4

and 2

10

6

s. In all these initial loading conditions
the pin is assumed xed to the surrounding bone; i.e., osseointegration is assumed to have taken
place. Subsequently, partial osseointegration of the implant was modeled by incorporating spring
elements at the pinbone interface.
Prior to implementing the dynamic approach, a forced harmonic response analysis was com-
pleted which provided vibration data over the whole of the discretized volume allowing the
movement of each node to be identied. This approach provides valuable information on the
displacement at the pinbone interface at the rst few natural frequencies for the pincantilever
model unlike the initial natural frequency extraction procedure of the whole discretized volume.
The equation of motion of a system with one degree of freedom and an excitation

F

(

t

) is
(37.1)
where

m

is the mass matrix,

c

is the damping matrix,

k

is the stiffness matrix,

x

is the displacement,



is the velocity, and is the acceleration.
The system response due to a unit impulse input with zero initial conditions is called its impulse
response. A unit impulse is impossible to realize in applications but an excitation can be considered
an impulse if its duration is very short compared with the natural period of the system.

29

If a shock

F

(

t

) is applied to a mass

m,

the equation of motion is that given by Equation 37.1
and the response

x

(

t

) is
(37.2)
where
(37.3)
For a shock of time length

t

o

and for

t

>

t

o

the maximum value of

x

(

t

) can be expressed as
(37.4)
which gives the residual shock spectrum since it occurs after the shock has terminated,
where
or more simply
mx cx kx F t ( )
x x
1 f
n
( )
x t ( ) F t ( )h t ( ) t d
0
t

h t
m
e t
t
( ) =
1

sin
x
F k
r
r r
o
max
cos =

2
1 2
2

r

i

n
------
n
2
,
k
m
----
n
t
o
,

r
---- and
n
, 2 f
n


0266/ch37/frame Page 571 Monday, October 25, 1999 10:12 AM

572

Mechanical Testing of Bone and the BoneImplant Interface

(37.5)
for
where

d

max

is the maximum displacement,

d

i



is the initial displacement,

T

is the initial period of
vibration = 2

n

,

and

d

max

/d

i

is the magnication factor, R.

C. K

NEE

L

IGAMENT

C

ONNECTOR



IN

C

ANCELLOUS

B

ONE

The value of this type of study is quantication of the holding of a screw in cancellous bone
with different bone trabeculae directions. The nite-element analysis should help in allowing
modications to the screw design as well as providing an indication of the xation in tissue. It
is also of value to be able to quantify stresses at the microstructural level in order to gain
knowledge of possible mechanically induced bone remodeling or excessive loading leading to
local bone fracture
Previously, effective stiffness and microstructural stresses have been approximated using a
representative volume element (RVE) approach.

30

This involves modeling and analyzing small
volumes of the composite material under a set of arbitrary boundary conditions. Other investigators
of the mechanical properties of trabecular bone have idealized the bone geometry as intersecting
plates

31

or as spherical voids within a cube.

32

Each of these approaches can only supply upper or
lower bounds on the effective stiffness because boundary conditions are unknown.
The current methodology again relies on a RVE approach using plates, but here the boundary
conditions should be better dened from a knowledge of the experimentally available pullout test
results. The nite-element discretizations are based on radiographic and histological observations
of the trabeculae directions and dimensions. The intersecting trabecula plates were arranged to be
at 0

, 45

, and 90

to the direction of loading of the bone. The model employed quadrilateral brick
and pentahedral continuum elements. Elastic properties of the trabeculae were varied between 10
and 20 GPa and the intratrabecular property as 1 GPa. These two values provided global trabecular
bone elastic behavior of

E

= 12.3 GPa = 0.26
and
E = 7.9 GPa = 0.24
The failure loads obtained from the pullout tests
33
were used in the model as applied loads for
each trabecula direction, thereby allowing location and magnitude of stresses at the screwbone
interface. An elastic modulus and Poissons ratio of E = 7.9 GPa and = 0.24 were judged to be
most representative of bovine bone properties and were used subsequently in the stress and dis-
placement calculations. The bonescrew assembly was axisymmetric, allowing computation to be
reduced by modeling a 60 section of the whole. The nite-element discretization for the 0 and
90 calculations used a model containing 2982 elements while it was necessary to use 5964
pentahedral elements to model the 45 trabeculae. Concentrated loads of 3.6, 2.8, and 2.4 N were
applied to each of the 39 nodes at the screw surface providing total loads over the modeled screws
of 840, 655, and 560 N for the 0, 45, and 90 trabecula directions, respectively. The total loads
were obtained from the screw pullout tests. The nodes at the base of the trabecular bone block
were xed in the x, y, and z global directions.
d
d
t
T
i
o max
sin = 2
t
T
o

1
2
0266/ch37/frame Page 572 Monday, October 25, 1999 10:12 AM
Finite Element Analysis for Evaluating Mechanical Properties of the BoneImplant Interface 573
III. RESULTS
A. AXIAL MODEL
Figure 37.3 shows the change of the rst natural frequency with bone density/modulus and the position
of bone support around the implant for the axisymmetric case. Following a shallow maximum at a
position 5 mm from the pin, the frequencies gradually fall away as the volume of bone surrounding
the implant increases. The frequencies are also seen to increase with bone modulus in the range 2
10
9
to 20 10
9
Pa. The effect of changing the bone properties from homogeneous to orthotropic were
investigated using a right-angled cantilever arrangement similar to that shown in Figure 37.1. The
results are shown in Figure 37.4 which indicates a rapid rise in frequency as the modulus of the
supporting tissues changes from that near brous material to cortical bone. Superimposed on this graph
is the frequency when the whole-bone block assumes orthotropic properties (diamonds).
FIGURE 37.3 Variatiozn in natural frequency of vibration of a dental implant with bone modulus and bone
support.
FIGURE 37.4 First natural frequency of the implant in a cylindrical bone block (Figure 37.1) for changing
homogeneous and orthotropic bone properties.
0266/ch37/frame Page 573 Monday, October 25, 1999 10:12 AM
574 Mechanical Testing of Bone and the BoneImplant Interface
B. MANDIBLE MODEL
The forced harmonic response spectrum is shown in Figure 37.5 and indicates a rst natural
frequency of the pin of about 3.6 kHz.
Table 37.1 shows the results of the dynamic analysis of the pin/cantilever model over the
impulse duration periods from 5 10
4
to 5 10
6
s. These data are plotted in Figure 37.6 where
the change in frequencies with impulse duration are indicated. The corresponding results for this
dynamic analysis of pin only are shown in Table 37.2 for excitation in the local x and y directions.
The variation in frequency with pulse length ratio for the pin with the cantilever removed is shown
in Figure 37.6.
Partial integration of the implant was modeled using spring elements at the boneimplant
interface. The change in frequency for different spring support conditions for an impulse strength
of t = 5.0 10
5
is shown in Figure 37.7.
FIGURE 37.5 Forced harmonic response for the pincantilever in mandible discretization with excitation
at the cantilever extremity indicating the rst natural frequency.
TABLE 37.1
Impulse Response Values for Pin and Cantilever
t
0
10
6
(s) T (s) d
i
(mm) d
max
(mm) f
n
(kHz)

100 N Load
5 3.5 10
4
0.55 10
4
0.75 10
4
1.36 2.86 0.014 71.4
10 3.5 10
4
1.0 10
4
1.53 10
4
1.53 2.86 0.028 35.0
50 3.8 10
4
3.5 10
4
6.7 10
4
1.91 2.63 0.130 7.69
100 4.4 10
4
8.5 10
4
11.2 10
4
1.31 2.27 0.230 4.34
200 5.8 10
4
14.2 10
4
14.6 10
4
1.03 1.72 0.345 2.9
500 11.0 10
4
17.5 10
4
16.0 10
4
0.91 0.91 0.450 2.22
10 N Load
100 4.4 10
4
0.85 10
4
112 10
4
132 2.27 0.230 4.34
R
d
max
d
i
----------

t
o
T
---

n

i
------

n
------
0266/ch37/frame Page 574 Monday, October 25, 1999 10:12 AM
Finite Element Analysis for Evaluating Mechanical Properties of the BoneImplant Interface 575
C. KNEE LIGAMENT CONNECTOR
The experimental results of the screw pullout tests indicate a stress of 54.9 MPa obtained when
the trabeculae are orientated in the same direction as the implant axis. This direction is the normal
load-bearing direction of the bone. At 90 to this direction, a stress of 37.1 MPa was obtained and
an intermediate value of 42.8 MPa at a 45 direction. The nite-element model was employed to
explain these failures at a microscopic level.
Following an examination of the types of stress present at the screwbone interface, it was
concluded that an effective shear stress best explained failure. An example of this stress distribution
at the bonescrew interface for the 0 trabeculae direction is shown in Figure 37.8. Relative stress
values are shown as contours and the bone plates of the model can be observed. The highest stresses
in the bone occur at the minor diameter of the screw threads and fall away rapidly along the interface
FIGURE 37.6 Change in implant frequency with impulse strength for the pincantilever and pin-only model
(full integration).
TABLE 37.2
Impulse Response Values for the Pin-Only Discretization (100 N load)
t
0
10
6
(s) T (s
1
) d
i
(mm) d
max
(mm) f
n
(kHz)

Impulse Local x Direction
2 6 10
5
1.4 10
6
3.5 10
6
2.5 16.6 0.033
5 6.0 10
5
3.5 10
6
8.5 10
6
2.4 16.6 0.09
10 6.3 10
5
8.7 10
6
13.7 10
6
1.57 15.8 0.16
50 1.25 10
4
24 10
6
30 10
6
1.25 7.9 0.40
100 1.8 10
4
34 10
6
25 10
6
0.73 5.56 0.54
Impulse Local y Direction
2 8 10
5
1.25 10
6
3.7 10
6
2.84 12.5 0.03
5 7.5 10
5
3.0 10
6
8.3 10
6
2.76 13.3 0.07
10 7.5 10
5
7.5 10
6
14.5 10
6
1.93 13.3 0.13
50 15.0 10
5
40.0 10
6
22 10
6
0.55 6.7 0.33
100 2.2 10
4
50 10
6
28 10
6
0.56 4.55 0.45
R
d
max
d
i
----------

t
o
T
---

n

i
------

0266/ch37/frame Page 575 Monday, October 25, 1999 10:12 AM

576 Mechanical Testing of Bone and the BoneImplant Interface
FIGURE 37.7 Change in implant frequency for the partially integrated condition.
FIGURE 37.8 Effective stress contours at the implant trabecular bone interface. Screw pullout calculation
for a simple plate discretization of trabecular bone.
0266/ch37/frame Page 576 Monday, October 25, 1999 10:12 AM
Finite Element Analysis for Evaluating Mechanical Properties of the BoneImplant Interface 577
away from the nodal support positions. The maximum effective stress predicted by the nite-element
analysis model was for the 0 trabecula direction. The value of 54 MPa is in good agreement with
experimentally measured values. For the 90 trabecula direction an effective stress of 37 MPa was
calculated and 43 MPa for the intermediate 45 direction.
IV. DISCUSSION
The natural frequencies of vibration of cantilever sections can be associated normally with three
distinct modes of propagation, namely, transverse, torsional, and longitudinal. The lowest natural
frequencies in a solid beam are due to resonant transverse modes, followed by torsional and
longitudinal modes. In this work, the initial dental pin/axial cantilever nite-element discretization
was modeled using axisymmetric solid elements. The transverse and torsional vibration modes are
therefore excluded and only the higher-frequency longitudinal mode allowed.
The frequencies calculated for the axisymmetric pinbone block shown in Figure 37.1 are
longitudinal modes and have been reduced by a value of 17.5 in Figure 37.3
13
for comparison with
the nonaxisymmetric cases plotted subsequently.
Nevertheless, the natural frequencies of vibration of the implant plotted in this gure suggest a
marked sensitivity to the type and quality of supporting bone. Thus, when the bonetissue modulus or
density surrounding the pin increases by a factor of ten, the rst natural frequency increases by 1000
Hz. Additionally, the natural frequencies are dependent on the volume and quality of bone supporting
the pin. If, as suggested clinically, cortical bone gradually builds around the implant during the healing
process, then it would appear about 5 mm thickness is sufcient for optimum support (see Figure 37.3).
The calculated natural frequencies with changing homogeneous bone modulus for the full 3D
discretization of pin, cantilever, and supporting bone are shown in Figure 37.4. In this analysis, the
bone was fully xed at the periphery of the bone block. Figure 37.4 also shows the effects of
changing to orthotropic bone properties. These new frequency values are superimposed on the
homogeneous data already indicated. The frequencies shown are somewhat different in magnitude
compared with the axial cantilever model shown in Figure 37.1, due to the different geometric
conditions imposed in the analysis. From these data, it is possible to calculate rates of pin integration
using frequency plots such as Figure 37.4, since it is to be expected that increasing implant support
is a result of increasing bone density and modulus. Thus, if a period of time P is required to obtain
a nal bone density/modulus during the healing period, then Figure 37.4 (i.e., frequency vs.
modulus) may be converted into a plot indicating the change in frequency with time. Implant
osseointegration plots similar to Figure 37.4 suggest substantial pin support is obtained early in
the healing process and little change occurs after approximately 40% of the elapsed healing time.
In the earlier section of this chapter, the natural frequencies of vibration of a single dental
implant with attached cantilever in a cylindrical block of bone were calculated. For a completely
osseointegrated pin with approximately 5 mm of surrounding bone, a rst natural frequency of
about 3.3 kHz was obtained. A forced harmonic response analysis for a similar implantcantilever
setup in the human mandible shown in Figure 37.2 yields a rst natural frequency of about 3.6
kHz (see Figure 37.5) which compares favorably with the earlier simplied model. The result
suggests that a bone model discretization local to the implant may be adequate for dynamic analysis.
From a clinical perspective, the attachment of devices to a dental implant can prove problem-
atical as well as involving patient stress. Thus, in the dynamic analysis approach, the cantilever
has been removed and an impulse excitation applied directly to the top of the dental implant. This
work follows on from an earlier analysis
34
where this methodology was attempted purely as a means
of examining the type of implant response to be expected in a clinical situation. However, in this
earlier work,
34
an arbitrary impulse excitation was applied at the top of the pin and while still
providing valuable displacement information may not necessarily yield the correct natural frequen-
cies required for comparison with a parallel clinical study.
0266/ch37/frame Page 577 Monday, October 25, 1999 10:12 AM
578 Mechanical Testing of Bone and the BoneImplant Interface
In the present work, this problem has been addressed by examining a range of impulse durations/
strengths for both the cantilever model and pin-only model shown in Figure 37.2. A listing of all the
frequency information obtained from a range of impulse values is shown in Table 37.1 for the cantilever
conguration. Figure 37.6 shows the change in frequency with impulse strength when the cantilever
is excited at the extremity. Clearly, true dynamic frequency values are only obtained at short impulse
durations (~ 5 10
6
s) and small displacements. As the impulse duration increases, the rst natural
frequency falls to unrealistic values near 1 kHz. The frequency obtained at an impulse of 5 10
6
s
(~3.0 kHz) is in close agreement with that obtained from the forced harmonic analysis (see Figure 37.5)
and the calculated value from the pincantileverbone block model. The effect of changing the impulse
load from 100 to 10 N has no effect on the frequency values, only on pin displacement.
The frequency data for the pin-only conguration (Figure 37.2) are shown in Table 37.2. These
data were obtained from impulse loads applied in the local x and y directions. Due to the varying
stiffness values of bone in the different directions,
34
it would be expected that different frequency
values emerge from this analysis. This appears to be the case with the frequency peaking at 16.6
kHz in the x direction where more bone support for the pin is to be expected compared with the
orthoradial direction (y) with a maximum frequency of 13.3 kHz
The frequencies are plotted against pulse length in Figure 37.6 for both loading directions. The
displacement and frequency values are sensitive to pulse duration, and as for the cantilever model,
small impulse durations are required to establish the correct frequency values, in this case near 2 10
6
s.
The frequency data calculated above was based on full integration of the implant with bone.
Any changes in frequency calculated with time thus demonstrate a change in the type and quality
of bone supporting the implant. In order to explore the condition of less than full integration, spring
elements were included at the implantbone interface. By varying the value of the spring constants,
a range of partial integration conditions were simulated. Figure 37.7 demonstrates the changing
implant frequency with changing spring support for an intermediate impulse loading of t
o
= 5
10
5
s. In these partial integration calculations, the supporting bone properties were maintained
constant. The calculations indicate the extreme sensitivity of the natural frequency of vibration of
the implant to the support offered by the tissue at the implantbone interface.
The mechanical data values used for the axial cantilever and mandible models outlined above
use essentially homogeneous and orthotropic bone properties for estimating the implant xity. This
is an idealized situation since it is known that bone properties are markedly directional.
A relatively simple perpendicular plate model for cancellous bone has yielded excellent agree-
ment with experiment (Figure 37.8). The model was effective in explaining the quantitative differ-
ences observed in screw pullout tests and also explained the radiographic features observed.
33
The
calculated stiffness values were somewhat higher than those measured experimentally, which was
attributed to the rigid boundary conditions imposed in the model discretization. However, the
calculations conrm the directionality of properties of cancellous bone and demonstrate a signicant
difference in screw-holding ability depending upon the direction of the cancellous bone spicules.
V. CONCLUSION
The dynamic analysis by impulse excitation of a dental pin osseointegrated in bone has indicated
a sensitivity of the implant response to both impulse duration and direction. The analysis provides
response data from the implant for a range of impulse strengths. These data may be used with
clinically measured experimental data for a correct interpretation of implant osseointegration, rate
of bone healing, or the onset of infection or pin integration failure. Renements to these models
should include discretization of bone volumes closer to the implant which will allow more-detailed
stress evaluation, a more exact representation of the bone structure, and relaxation of the rigid
boundary conditions which produce high stiffness values. It is the intention in future work to develop
and rene further the mesh by the introduction of material and geometric nonlinearities in order to
simulate the early stages of bone healing. These new procedures will allow full validation of the
numerical technique and extend its use to a noninvasive clinical assessment of osseointegration.
0266/ch37/frame Page 578 Monday, October 25, 1999 10:12 AM
Finite Element Analysis for Evaluating Mechanical Properties of the BoneImplant Interface 579
REFERENCES
1. Brnemark, P.-I., Osseointegration and its experimental background, J. Prosthet. Dent., 50, 399, 1983.
2. Albrektsson, T., Hansson, H.A., and Ivarsson, B., Interface analysis of titanium and zirconium bone
implants, Biomaterials, 6, 97, 1985.
3. Brnemark, P.-I., Hansson, B.O., and Adell, R., Osseointegrated implants in the treatment of the
edentulous jaw. Experience from a 10 year period, Scand. J. Plast. Reconstr. Surg., 16 (Suppl.), 1977.
4. Collier, R.J., Nadan, O., and Thomas, T.G., The mechanical resonances of a human tibia: Part I
in vitro, J. Biomech., 15, 545, 1882.
5. Christensen, A.B., Ammintzboll, F., Dyrnbye, C., et al., Assessment of tibial stiffness by vibration testing
in situ. I. Identication of mode shapes in different supporting conditions, J. Biomech., 19, 53, 1986.
6. Collier, R.J. and Donarski, R.J., Non-invasive method of measuring resonant frequency of a human
tibia in vivo. Parts I and II, J. Biomed. Eng., 9, 321, 1987.
7. Bull, D.R., Nokes, L.D.M., and Frampton, R., Detection of internal xation plate loosening by means
of an analysis of vibratory response, J. Biomed. Eng., 13, 321, 1991.
8. Doherty, W.M., Dynamic Response of Human Tibia, Ph.D. dissertation, University of California,
Berkeley, 1971.
9. Jurist, J.M. and Kianion, K., Three models of the vibrating ulna, J. Biomech., 6, 331, 1973.
10. Lewis, J.L., A dynamic model of a healing fractured long bone, J. Biomech., 9, 17, 1975.
11. Albrektsson, T. and Hansson, H.A., An ultrastructural characterisation of the interface between bone
and sputtered titanium or stainless steel surfaces, Biomaterials, 7, 201, 1986.
12. Meredith, N., Alleyne, D., and Cowley P., Quantitative determination of the stability of the implanttis-
sue interface using resonance frequency analysis, Clin. Oral Implants Res., 7, 1, 1996.
13. Williams, K.R. and Natali, A.N., Simulation of the stability of dental implants in bone by natural
frequency analysis, in 2nd Symposium on Computer Methods in Biomechanics and Biomedical
Engineering, Middleton, J.M., Pande, G.R., and Jones, M., Eds., Edward Arnold, London, 1995, 415.
14. Williams, K.R. and Williams, A.D.C., Impulse response of a dental implant in bone by numerical
analysis, Biomaterials, 715, 18, 1997.
15. Awadalla, H.A., Azarbal, M., Ismail, Y.H., and el-Ibiari, W., Three-dimensional nite element stress
analysis of a cantilever xed partial denture, J. Prosthet. Dent., 68, 243, 1992.
16. Chen, J., Lu, X., Paydar, N., et al., Mechanical simulation of the mandible with and without an
endosseous implant, Med. Eng. Phys., 16, 53, 1994.
17. Hart, R.T., Hennebel, V.V., Thongpreda, N., et al., Modeling the biomechanics of the mandible: a
three-dimensional nite element study, J. Biomech., 25, 261, 1992.
18. Holmes, R.T., Grigsby, W.R., Goel, V.K., and Keller, J.C., Comparison of stress transmission in the
IMZ implant system with polyoxymethylene or titanium intramobile element: a nite element stress
analysis, Int. J. Oral Maxillofac. Implants, 7, 450, 1992.
19. Ko, C.C., Kohn, D.H., and Hollister, S.J., Micromechanics of implant/tissue interfaces, J. Oral
Implantol., 18, 220, 1992.
20. Meijer, H.J., Starmans, F.J., Bosman, F., and Steen, W.H.A., A comparison of three nite-element
models of an edentulous mandible provided with implants, J. Oral Rehabil., 20, 147, 1993.
21. Meijer, H.J., Starmans, F.J., Bosman, F., and Steen, W.H.A., A three dimensional nite element analysis
of bone around dental implants in an edentulous mandible, Arch. Oral Biol., 38, 491, 1993.
22. Rieger, M.R., Adams, W.K., and Kinzel, G.L., A nite element survey of eleven endosseous implants,
J. Prosthet. Dent., 63, 457, 1990.
23. Richter, E.J., Basic biomechanics of dental implants in prosthetic dentistry, J. Prosthet. Dent., 61,
602, 1989.
24. Hart, R.T., Davy, D.T., and Heiple, K.G., A computational method for stress analysis of adaptive
elastic materials with a view toward applications in strain-induced bone remodeling, J. Biomech. Eng.,
106, 342, 1984.
25. Natali, A.N., Nonlinear interaction phenomena between pin and bone, Clin. Mater., 9, 109, 1992.
26. Natali A.N., and Meroi, E., Pinbone interface: a geometric and material nonlinear analysis, Proc.
Interfaces 90, Elsevier Science, London, 1991, 104.
27. Meredith, N., Cawley, P., and Alleyne, D., The application of modal vibration analysis to study bone
healing in vivo, J. Dent. Res., 73, 793, 1994.
0266/ch37/frame Page 579 Monday, October 25, 1999 10:12 AM
580 Mechanical Testing of Bone and the BoneImplant Interface
28. Natali, A.N. and Meroi, E., A review of the mechanical properties of bone as a material, J. Biomed.
Eng., 11, 266, 1989.
29. Crede, C.E., Shock and Vibration Concepts in Engineering Design, McGraw-Hill, New York, 1965,
138.
30. Hashin, Z., Analysis of composite materials a survey, J. Appl. Mech., 50, 481, 1983.
31. Pugh, J.W., Rose, R.M., and Rodin, E.L., A structural model for the mechanical behavior of trabecular
bone, J. Biomech., 6, 657, 1973.
32. Beaupre, G.S. and Hayes, W.C., Finite element analysis of a three-dimensional open-celled model for
trabecular bone, J. Biomech. Eng., 107, 249, 1985.
33. Young, F.A., private communication.
34. Natali, A.N., Meroi, E., Williams, K.R., and Calabrese, L., Investigation of the integration process of
dental implants by means of a numerical analysis of dynamic response, Dent. Mater., 13, 325, 1997.
0266/ch37/frame Page 580 Monday, October 25, 1999 10:12 AM

581

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

38

Fatigue Testing of Bioabsorbable
Screws in a Synthetic
Bone Substrate

William S. Pietrzak, David R. Sarver, and David H. Kohn

CONTENTS

I. Introduction ..........................................................................................................................581
II. Background ..........................................................................................................................582
A. Bioabsorbable Polymer Technology ..............................................................................582
B. Lactosorb

Copolymer

in Vitro

Strength-Loss Prole..................................................583
C. Mathematical Model of Screw Fixation of a Mandibular Osteotomy..........................584
D. Fatigue Testing of Absorbable Materials.......................................................................585
III. Fatigue Testing of Bioabsorbable ScrewSynthetic Bone ..................................................586
A. Methods and Materials...................................................................................................586
B. Results ............................................................................................................................587
C. Discussion.......................................................................................................................588
IV. Conclusions ..........................................................................................................................589
References ......................................................................................................................................589

I. INTRODUCTION

The use of rigid internal xation (RIF) to maintain the apposition and stability of skeletal fractures
and osteotomies during osseous union is one of the primary techniques employed by orthopaedic
and craniomaxillofacial surgeons. Historically, such xation devices, consisting of plates, screws,
and pins, were made of medical-grade metals such as stainless steel, titanium and its alloys, and
cobaltchromium alloys. Despite the success of this mode of xation, there has been increasing
concern over certain consequences that may originate from the use of these implants. The high
stiffness of these materials, relative to bone, may cause stress shielding in the vicinity of the implant,
thereby weakening the bone; implant migration and growth restriction, particularly in pediatric
patients, can be an issue; long-term tissue reactions to the presence of metal implants have been
documented; palpability can be problematic; key radiographic information can be obscured by their
presence; and they can interfere with therapeutic applications of irradiation.

1-7

All of the above can
result in the need for a secondary surgical procedure to remove the implant with its attendant risks
and costs. In an attempt to address this wide spectrum of issues, a new class of RIF devices has
evolved over the past 30 years, and has achieved status as a viable alternative to metallic internal
xation over the past 10 years.

8-12

These devices, fabricated from bioabsorbable polymers, gradually
degrade commensurate with tissue healing and are eliminated by the body. In this fashion, a
temporary implant serves a temporary xation need without the requirement of further surgical
intervention to remove the implant.

0266/ ch38/frame Page 581 Friday, October 22, 1999 11:15 AM

582

Mechanical Testing of Bone and the BoneImplant Interface

There is increasing interest in the use of absorbable xation of mandibular fractures and
osteotomies, including absorbable screw xation of sagittal split mandibular osteotomies for the
correction of occlusal deciencies.

13-17

Not only is this a signicant load-bearing application, but
the dynamic, cyclic nature of the applied loads mandate that not only must this xation be capable
of withstanding the magnitude of the expected loads, but also the fatigue characteristics of such
xation must be demonstrated to be capable of withstanding the expected number of cycles until
osseous union occurs. Absorbable implants present special considerations with respect to charac-
terization of their fatigue properties, and, to date, few studies have investigated the subject.

18

Presented here is a basic background of absorbable technology and the unique methods employed
to determine the fatigue characteristics of such implants in a synthetic bone substrate. Clinical
relevance to the case of sagittal split mandibular xation is discussed, although the concepts
presented are broad and have general applicability.

II. BACKGROUND
A. B

IOABSORBABLE

P

OLYMER

T

ECHNOLOGY

Ideally, an absorbable polymer xation implant should provide adequate xation for approximately
6 to 8 weeks to permit primary healing to occur.

19

During this period, a decrease in xation strength
due to implant degradation should be matched by a corresponding increase in strength from the
healing biological union. The degrading implant should not cause any deleterious physiological
reaction necessitating surgical intervention. The implant should fully absorb and be totally elimi-
nated. Once gone, the body should forget that the implant was ever there.

20

In general, absorbable polymers degrade,

in vivo

, in a two-phase process.

5,20

In the rst phase,
water molecules inltrate the polymer and hydrolyze the polymer chains, thereby reducing the
molecular weight and causing a reduction in mechanical strength of the implant. As the polymer
degrades and fragments, a cellular macrophage response ensues, with the macrophages phagocy-
tizing the polymer debris, metabolically converting them to water and carbon dioxide which is
eliminated via the kidneys and the lungs.
These degradation phases result in a timed sequence of molecular weight reduction, strength
loss, and mass loss, often with a signicant difference in the times between strength loss and mass
loss.

20

Consequently, the implant will generally lose all mechanical strength before signicant mass
is lost. Many factors inuence the rate of polymer degradation, including the chemical identity and
microstructure of the polymer, the molecular weight, the size and shape of the implant, the
vascularity of the implant site, the tissue environment (bone or soft tissue), and the state of stress
application to the implant.

20-23

The strength-loss proles of absorbable implants are often tested under controlled laboratory
conditions that simulate the

in vivo

environment, i.e., pH 7.4 phosphate-buffered saline (PBS) at
37

C.

24

The implants, submersed in the buffer either alone or attached to a substrate, are periodically
removed from the bath for mechanical testing. While such testing probably provides a good
approximation to the expected initial

in vivo

strength-loss proles, it cannot mimic the

in vivo

mass-loss proles due to the lack of a metabolizing, cellular response.
The most commonly used absorbable polymers are from the class of constituents known as
alpha-hydroxy acids.

25

Such compounds include lactic acid (both

D

and

L

), glycolic acid, and

p

-dioxanone, with polymers of

L

-lactic acid, glycolic acid, and copolymers thereof the most
prevalent. Historically, homopolymers of glycolic acid (PGA) and

L

-lactic acid (PLLA) have
dominated clinical use. Late inammatory reactions in response to these degrading polymers have
been documented in several clinical series.

23,26-31

As the homopolymers tend to possess a high degree
of crystallinity, crystalline release is thought to play a role in this reaction.

23,26-29

Additionally, PGA,
which degrades very quickly, may overwhelm the local ability of the body to clear the acidic

0266/ ch38/frame Page 582 Friday, October 22, 1999 11:15 AM

Fatigue Testing of Bioabsorbable Screws in a Synthetic Bone Substrate

583

degradation products, and may be a factor in these reactions. There is a trend toward the increasing
use of amorphous copolymers which appear to have little or no incidence of these late reactions.
One such copolymer, LactoSorb

(Biomet, Inc., Warsaw, IN), is composed of 82%

L

-lactic acid
and 18% glycolic acid and has been used in over 10,000 clinical cases for internal xation. This
material retains signicant strength for 6 to 8 weeks,

24

requires approximately 12 months to
completely absorb,

19,32,33

and results in no late inammatory response as demonstrated in both
animal testing

19,32

and human clinical series.

33

Coupled with appropriate implant design, LactoSorb

appears to be suitable for many internal xation needs. The absorbable screw implants discussed
below are made from LactoSorb

copolymer.

B. L

ACTOSORB

C

OPOLYMER



IN

V

ITRO

S

TRENGTH

-L

OSS

P

ROFILE

The modes by which screws can experience load

in situ

include axial pullout or extraction forces,
bending or exural forces, and shear forces. For sagittal split mandibular osteotomies, it is the
resistance to shear that is the primary material property of interest.

13

To determine the

in vitro

shear
strength-loss prole of LactoSorb

copolymer, a model system was tested. This model consisted
of small-diameter (1.5-mm) cylindrical rods of LactoSorb

copolymer that provided a uniform
cross-sectional area for the tests. The cylinders were submersed in pH 7.4 PBS at 37

C for
predetermined periods up to 12 weeks, then removed from the bath and mechanically tested at the
various intervals.
Shear strength testing consisted of placing the cylinders through a continuous hole drilled through
a sequence of three steel plates, placed directly alongside each other, as shown in Figure 38.1. The
clearance between the plate holes and the cylinders was sufciently tight to just permit introduction
of the cylinders without notable excess clearance. The two outer plates were xed to the base of the
mechanical test system (Sintech 1/S, Sintech, Inc., Research Triangle Park, NC) while the central
plate was attached to the movable crosshead of the apparatus. With the crosshead engaged to move
upward at the rate of 5.0 mm/min, forcedisplacement curves were recorded. Shear strength was
dened as the peak shear stress at failure (severance of the cylinder) as calculated by the Testworks
software (Sintech, Inc. Research Triangle Park, NC). Five specimens were tested in this fashion at
each

in vitro

time point. The statistical signicance of differences between means was determined
by ANOVA and subsequent StudentNewmanKeuls test at

p

< 0.05.

FIGURE 38.1

Schematic showing the measurement of shear strength of the bioabsorbable LactoSorb

cylinders.

0266/ ch38/frame Page 583 Friday, October 22, 1999 11:15 AM

584

Mechanical Testing of Bone and the BoneImplant Interface

Figure 38.2 shows the

in vitro

shear strength-loss prole. In this model, the shear strength
increased signicantly over the rst 8 weeks and did not signicantly diminish, relative to initial
values, until after 10 weeks of hydrolysis in PBS. Although not fully understood, this phenomenon
of initial strength increase has also been observed by the authors when performing similar tests on
LactoSorb

copolymer screws (unpublished data). This may be due to the occurrence of two
competing reactions, with hydrolytic strength loss dominating after several weeks.

C. M

ATHEMATICAL

M

ODEL



OF

S

CREW

F

IXATION



OF



A

M

ANDIBULAR

O

STEOTOMY

During orthognathic mandibular surgery to correct occlusal deciencies, the mandibular ramus is
bilaterally sagittally split. The distal fragment, or the body of the mandible, is then advanced or
set back. The fragments are typically internally xed with three screws on each side. Generally,
the patient is placed in maxillomandibular xation (MMF) and on a soft diet for a period of time,
followed by MMF removal and a program of progressively more demanding rehabilitation.

34

The analysis of the state of stress in the screws used to x sagittal split mandibular osteotomies
is complex, as has been derived by Kohn et al.

13

In its simplest form, the distal bone fragment can
be treated as a cantilever beam bolted onto a rigid plate. A measure of the stability of such xation
can be dened as the deection of the tip of the beam under the inuence of the loading pattern,
with greater stability associated with less deection. A mathematical model was derived for such
a loading pattern that related the deection of the tip of the distal bone fragment to the applied
load, considering such factors as bone geometry, modulus, load lever arm, screw size, shear strength,
and screw pattern. Typically, there is a critical screw in the xation pattern that sustains the
greatest shear stress and governs the stability of the xation. The position of the critical screw
varies with the screw pattern employed. Consequently, the equation relating the load on the distal
fragment to the deection of the tip of the distal fragment is dependent upon the xation screw
conguration. For a linear pattern, the relationship may be expressed as

total

=

L

tan {cos

1

[1 (

R

2

/4 +

RS

+

S

2

)

D

2

/2

l

2

G

2

]} +

PL

3

/3

E

(38.1)

R

=

PL/Al

(38.2)

S

=

P

/3

A

(38.3)
where

total

is the total deection of the tip of the distal fragment;

P

is the load on the distal fragment;

L

is the distance between the point of load application and the centroid of the xation screw pattern,

FIGURE 38.2

Peak shear stress of LactoSorb

cylinders

in vitro

. Error bars represent

1 SD.

0266/ ch38/frame Page 584 Friday, October 22, 1999 11:15 AM

Fatigue Testing of Bioabsorbable Screws in a Synthetic Bone Substrate

585

i.e., the lever arm;

D

is the screw diameter;

G

is the shear modulus of the screws;

l

is the interscrew
distance;

A

is the screw cross-sectional area;

E

is the elastic modulus; and I is the cross-sectional
moment of inertia of the distal bone fragment.

13

The model was validated by empirically testing both LactoSorb

copolymer and titanium screws
in various substrates, including cadaver mandibles and synthetic mandibles. It was found that the
dominant effect causing deection was rotation of the distal fragment about the centroid of the screw
pattern, in response to the shear deformation of the screws. Bending of the distal bone fragment under
the action of the applied load had an effect that was three orders of magnitude less. Consequently, in
this type of model, the quality of the substrate was not important, and, in fact, comparing predicted
deections with both synthetic and natural substrates yielded no signicant differences.

13

D. F

ATIGUE

T

ESTING



OF

A

BSORBABLE

M

ATERIALS

In general, a material specimen, exposed to uctuating loads and deformations above a certain
threshold, will eventually fail suddenly and catastrophically after a period of time. This process is
known as fatigue.

35

The loads and deformations that cause fatigue failure are typically far below
those that cause static failure, or failure after the application of a single stress cycle. The mechanism
behind this process is the initiation and propagation of cracks to an unstable size. Once this critical
size of defect is reached, the next cycle of applied stress causes catastrophic failure. Fatigue behavior
is highly statistical in nature since the nucleation sites for crack initiation include material defects
which, themselves, are statistically distributed. Many engineered parts, including implants, undergo
cyclic loads in actual use conditions, making an understanding of their fatigue behavior an essential
part of the design process.
The formal theory and practice of fatigue analysis is complex and beyond the scope of this
chapter.

35

In general, the number of applied load cycles to failure (

fatigue life

) will increase with
decreasing load, or stress, applied per cycle. Some materials may approach a

fatigue limit,

which
is dened as the stress cycle at which the specimen can survive an innite number of cycles, i.e.,
does not exhibit fatigue. Specimens can be loaded in a variety of congurations (tension, compres-
sion, shear, torsion, bending, etc.). The load cycles can be tailored by adjusting the minimum and
maximum load values, the frequency, the amplitude prole, etc. A given specimen will typically
be cycled until failure. Alternatively, a run-out can be established in which the test on a given
specimen is terminated if it survives a predetermined number of load cycles,

N

.
Testing the fatigue characteristics of absorbable polymers presents special challenges due to
the deteriorating physical properties of the polymer in a hydrolytic environment. Thus, a fatigue
test performed on an absorbable polymer in dry air, where the changes in the material properties
are attributed only to the applied cyclic load, will be considerably different than a test performed
while the specimens are submersed in aqueous buffer, because of the combination of degradation
due to cyclic loading and hydrolysis. Both of these effects may be synergistic; that is, the formation
of microcracks during cyclic loading may create pathways to facilitate the penetration of water into
the interior of the specimen, thereby increasing the rate of hydrolysis. Conversely, during hydrolysis,
absorbable polymers typically embrittle, which would be expected to reduce fatigue performance.
Consequently, the concept of a fatigue limit would not be expected since pronounced degradation
would ultimately lead to failure regardless of the magnitude of the applied cyclic load.
A novel approach to the determination of the fatigue properties of absorbable polymer speci-
mens is to preincubate the specimens,

in vitro

, in a load-free manner to create a given amount of
hydrolytic degradation. Then, the specimens are cyclically loaded over a period of time that is
relatively small compared with the unloaded incubation time in an effort to obtain fatigue snap-
shots of the material as hydrolysis is occurring. Thus, a composite curve can be constructed that
can indicate the point of hydrolysis at which the number of cycles to failure is adversely affected.
If an absorbable polymer implant is to be used as an osteosynthesis fracture xation device,
then the implant should be inserted into an appropriate substrate, with the implant(s) and substrate

0266/ ch38/frame Page 585 Friday, October 22, 1999 11:15 AM

586

Mechanical Testing of Bone and the BoneImplant Interface

tested as a complex. For an osteosynthesis device, one choice of substrate is bone. However, there
are many problems with the use of this substrate (see below) that make desirable the use of a
synthetic substrate. Following are the details of a unique procedure to determine the fatigue
characteristics of absorbable screws in synthetic bone substrate.

III. FATIGUE TESTING OF BIOABSORBABLE SCREWSYNTHETIC BONE
A. METHODS AND MATERIALS
The absorbable osteosynthesis implants used were 2.5-mm-diameter by 11-mm-long fully threaded
screws fabricated from LactoSorb

absorbable copolymer (supplied by Lorenz Surgical, Jackson-

ville, FL). The substrate was solid, polyurethane synthetic cortex (Pacic Research, Vashon Island,
WA) with an elastic modulus of approximately 15 GPa. The synthetic cortical substrate was 3.2 mm
in thickness and cut into rectangles of dimensions 26 mm wide by 37.5 mm long. Two of the
cortical specimens were overlapped by 13 mm and xed together using three of the screws in a
linear pattern, directed perpendicular to the long axis of the cortical specimens, as shown in
Figure 38.3. Screws were administered by drilling 2.2-mm-diameter pilot holes through the over-
lapping segments and tapping the holes to a threadform identical to that of the screws. The screw
heads had a twist-off drive mechanism that sheared off once the screws were fully seated. Thus,
all screws experienced the same nal insertion torque. Screws were used positionally without
lagging with an interscrew distance of approximately 7 mm. At the end of each of the two cortical
specimens (opposite the center of the construct), a 6.5-mm-diameter hole was drilled to permit
attachment of each end to the mechanical test apparatus with pins.
The constructs were rst submersed in a circulating, temperature-controlled bath of PBS (pH
7.4, 37C), without load, from 0 to 8 weeks prior to fatigue testing. Fatigue testing was then
performed with the constructs submersed in a similar bath with the two ends of the constructs
attached to the stationary base and movable crosshead of the servohydraulic custom-testing machine
(Endura-Tec, Eden Prairie, MN), respectively.
During the fatigue test, a sinusoidal tensile load was applied to the construct with magnitudes
from 4.5 to 45.5 kg
f
(corresponding to 11.6 to 116 MPa of shear stress based on a screw minor
diameter of 2.2 mm) at the rate of 1 Hz. The test apparatus was under closed-loop load control;
i.e., the apparatus generated whatever displacement was required to maintain this load cycle. The
bodies of the screws at the interface between the blocks were in shear. Due to the nite thickness
FIGURE 38.3 Schematic of synthetic substrateabsorbable screw constructs.
0266/ ch38/frame Page 586 Friday, October 22, 1999 11:15 AM
Fatigue Testing of Bioabsorbable Screws in a Synthetic Bone Substrate 587
of each synthetic cortex, there was a distance of 3.2 mm between their centers. The high stiffness
and minimal relative motion of the cortical blocks ensured that the screws did not bend during
cyclic loading.
Three to ve repetitions were performed for each period of pre-soaking the unloaded spec-
imens. Individual runs were terminated when the screws failed by shear (a displacement greater
than 2 mm detected) or a maximum of 260,000 cycles (run-out) was attained. Statistical signicance
of the differences between means was determined by ANOVA and subsequent application of the
StudentNewmanKeuls test at p < 0.05.
B. RESULTS
Table 38.1 summarizes the data while Figure 38.4 shows a plot of the log of the number of cycles
corresponding to run-out (2.6 10
5
cycles) or failure as a function of unloaded incubation time.
The specimens attained 5.1 to 5.4 log units of cycles up to at least 4 weeks of unloaded incubation
in PBS. After 8 weeks of unloaded incubation in buffer the number of cycles to failure was
signicantly less than those of all of the preceding time points. The trendline shows a knee in
the curve at a time equivalent to about 6 weeks of exposure to PBS, meaning that the corre-
sponding degree of hydrolysis is sufcient to affect the fatigue characteristics of the LactoSorb

specimens under the conditions tested. Failure of the substrate was not observed at any of the
time periods.
TABLE 38.1
Summary of Fatigue Test Results
Nonload PBS
Incubation Time
(weeks)
Average
No. Cycles S.E.M. N
Log
(no. cycles)
0 2.60 10
5
0 5 5.41
1 1.75 10
5
5.25 10
4
3 5.24
2 2.58 10 2.06 10
3
5 5.41
3 1.72 10
5
3.30 10
4
4 5.24
4 1.36 10
5
1.73 10
4
4 5.13
8 8.64 10
2
7.2 10
1
3 2.94
FIGURE 38.4 Plot of the log of the number of cycles to run-out (2.5 10
5
) or failure as a function of weeks
of initial unloaded incubation in PBS (pH 7.4, 37C).
0266/ ch38/frame Page 587 Friday, October 22, 1999 11:15 AM
588 Mechanical Testing of Bone and the BoneImplant Interface
C. DISCUSSION
Synthetic substrates provide a number of advantages over cadaver bone for the testing of xation
implants. First, the quality of cadaver bone varies widely, requiring a large number of specimens
to be tested per conguration to establish signicant differences. Second, xation implants are
often used in relatively young patients whose bone quality can be poorly represented by the often
fragile, osteoporotic bone characteristic of the elderly donors from whom most cadaver bone is
derived. Third, cadaver bone is typically obtained fresh-frozen, hence has not been sterilized,
creating stringent handling requirements for the prevention of disease transmission. Fourth, for a
long-term in vitro study to be performed, deterioration of the properties of the cadaver bone over
time must be considered.
The use of synthetic substrates in the testing of internal xation screws is an established
method.
36-40
Such substrates, typically polymeric, have been used as models for many skeletal sites,
including the mandible, the pedicles of the spine, the femur, and the metatarsals.
38,39,41,42
In one
study, the interfemur mechanical variability for the synthetic models was 20 to 200 times less than
that of cadaver femurs.
42
The polyurethane synthetic cortex used in the current study had a similar
modulus to that of cortical bone, i.e., about 15 GPa for the synthetic vs. about 12 to 17 GPa for
natural bone.
43
It has been stated that plastic models may simulate the mechanical properties of
natural bone in tests in which only elastic deformation of the substrate is achieved.
41
During the
fatigue testing, the substrate did not undergo noticeable plastic deformation since no alteration of
the screw hole geometry was seen. Consequently, for the purpose of generating preliminary fatigue
data on absorbable polymer screws used for the xation of sagittal split mandibular osteotomies,
the synthetic substrate model served as a reasonable approximation to natural healthy bone.
Maximum mean unilateral molar bite forces measured in skeletally normal adults have been
reported as ranging from approximately 30 to 86 kg.
44-46
Patients requiring orthognathic surgery
typically exhibit lower bite forces than normal controls.
47
For patients requiring mandibular setback
surgery, mean maximal molar bite forces were measured at 13.7 kg before surgery, 7.6 kg at MMF
removal, and 14.2, 19.7, and 26.1 kg at 3, 6, and 12 months postsurgery, respectively.
48
In normal
chewing, the forces exerted on the occlusal surface seldom exceed 5 to 7 kg.
49
As such, the load
range of 4.5 to 45 kg used in the current study may be considered to be the worst case since patients
undergoing postoperative rehabilitation following mandibular surgery are unlikely to exceed normal
chewing activities. It has been reported that a typical bite cycle in a normal patient is about one
per second.
50
Assuming that a patient eats three meals per day and is chewing an average of 10 min
per meal, then this corresponds to about 2000 masticatory cycles per day. Over the course of
8 weeks, this represents 112,000 cycles. Thus, the 1 Hz load frequency and the run-out limit of
260,000 cycles chosen for the fatigue test were also clinically relevant.
At the rate of 1 Hz, approximately 3 days are required to accumulate 260,000 cycles. From
Figure 38.2, it is apparent that little hydrolysis occurs over this short an interval. Consequently,
Figure 38.4 may be interpreted as a true composite curve whereby each data point represents the
fatigue characteristics of the LactoSorb

copolymer screwsynthetic bone constructs for a given

level of hydrolysis.
Under the conditions of the fatigue test employed, it is apparent that the fatigue properties of
the LactoSorb

copolymer screws diminish greatly after the equivalent of about 6 weeks of

hydrolysis in simulated in vivo conditions. Tested statically, the LactoSorb

copolymer cylinders
did not begin to lose strength until after 8 weeks. This may be evidence of a synergistic effect of
hydrolysis and cyclic loading.
The major goal of this study was to develop a novel method of characterizing the fatigue
characteristics of an absorbable screwsynthetic bone substrate system. The composite curve that
was created, showing fatigue snapshots corresponding to various degrees of screw hydrolysis,
demonstrated the interrelated dynamics of hydrolysis and cyclic loading. It was demonstrated that
xation of synthetic bone substrate with three 2.5-mm-diameter LactoSorb

copolymer screws
0266/ ch38/frame Page 588 Friday, October 22, 1999 11:15 AM
Fatigue Testing of Bioabsorbable Screws in a Synthetic Bone Substrate 589
could withstand cyclic loads nearly one order of magnitude greater than typical loads developed
during chewing. Additionally, even after considerable hydrolysis, the screws withstood the number
of chewing cycles expected from a patient during healing. This provides strong evidence of the
utility of this method of xation for sagittal split mandibular osteotomies. This method provides a
unique analytical tool upon which future studies can be based.
There were, however, several limitations in the current study that can be addressed with more
sophisticated methods. First, as was discussed above, the mathematical analysis of the stability of
screw xation of sagittally split mandibular osteotomies is complex, requiring that a series of vector
equations be solved to characterize the shear stress pattern in the critical screw in response to the
moment placed on the distal fragment. In the current study, a simple model was chosen for the
preliminary study of the fatigue properties of an absorbable screwsynthetic bone complex. The
substrate segments were loaded in tension without an applied moment. As such, the stresses resisted
by each of the three screws were equivalent to each other, unlike the clinical situation in the
mandible. The testing, as it was performed, may have more correlation with the stress pattern in
screws used to x fractures of the extremities and other regions. Second, soaking the implants,
unloaded, in buffer prior to accelerated fatigue testing does not replicate the short periods of cyclic
loading during eating followed by long periods of non-load-bearing between meals characteristic
of the clinical pattern. The interplay between hydrolysis and fatigue would be expected to differ
between the two cases. Third, only one set of loading parameters was used in the study. A different
loading cycle may have produced a different composite curve than the one shown in Figure 38.4.
Although the applied loads were extreme and can be considered to be worst case, future studies
should include the generation of a family of composite curves in which each individual curve
corresponds to a different cyclic load cycle. In this manner, the variables of hydrolysis, loading
pattern, and number of cycles to failure could be combined.
An important concept governing the use of absorbable xation devices is that of variable load
sharing. When an osteotomy or fracture is united with absorbable xation, initially the absorbable
xation provides 100% of the support. After healing occurs and the absorbable devices are elimi-
nated, all of the support is provided by the osseous union. At intermediate times, the load will be
partially shared between the healing union and the residual strength of the absorbable implants.
This important load-sharing concept cannot be mimicked by any in vitro study with cadaver or
synthetic bone substrates.
IV. CONCLUSIONS
The fatigue testing of bioabsorbable xation devices presents signicant challenge due to the time-
varying properties of the materials under the inuence of hydrolysis. To avoid the limitations of a
cadaver bone substrate, a synthetic bone substrate that resembled the modulus of natural cortical
bone was used in conjunction with absorbable copolymer screws. Under the test conditions
employed, it appeared that the fatigue characteristics of the absorbable screws did not begin to
diminish greatly until about 6 weeks of hydrolysis had occurred. The results indicate that these
absorbable screws may be useful to x sagittal split mandibular osteotomies based on documented
loading patterns in the mandible.
REFERENCES
1. Viljanen, J., Kinnunen, J., Bondestam, S., et al., Bone changes after experimental osteotomies xed
with absorbable self-reinforced poly-L-lactide screws or metallic screws studied by plain radiographs,
quantitative computed tomography and magnetic resonance imaging, Biomaterials, 16, 1353, 1995.
2. Velkes, S., Nerubay, J., and Lokiec, F., Stress fracture of the proximal femur after screw removal,
Arch. Orthop. Trauma Surg., 115, 61, 1996.
0266/ ch38/frame Page 589 Friday, October 22, 1999 11:15 AM
590 Mechanical Testing of Bone and the BoneImplant Interface
3. Ebraheim, X.R., Nadaud, M.C., and Phillips, E.R., Local tissue of the lumbar spine response to
titanium platescrew system. Case reports, Spine, 21, 871, 1996.
4. Thomas, K.A., Cook, S.D., Harding, A.F., and Haddad, R.J., Jr., Tissue reaction to implant corrosion
in 38 internal xation devices, Orthopedics, 11, 441, 1988.
5. Pietrzak, W.S., Sarver, D.R., and Verstynen, M.L., Bioabsorbable polymer science for the practicing
surgeon, J. Craniofac. Surg., 8, 87, 1997.
6. Jorgenson, D.S., Mayer, M.H., and Ellenbogen, R.G., et al., Detection of titanium in human tissues
after craniofacial surgery, Plast. Reconstr. Surg., 99, 976, 1997.
7. Rubin, J.P. and Yaremchuk, M.J., Complications and toxicities of implantable biomaterials used in
facial reconstructive and aesthetic surgery: a comprehensive review of the literature, Plast. Reconstr.
Surg., 100, 1336, 1997.
8. Cutright, D.E., Hunsuck, E.E., and Beasley, J.B., Fracture reduction using a biodegradable material,
polylactic acid, J. Oral Surg., 28, 393, 1971.
9. Viljanen, V.V. and Lindholm, T.S., Background of the early development of absorbable xation
devices, Tech. Orthop., 13, 117, 1998.
10. Pietrzak, W.S., Verstynen, M.L., and Sarver, D.S., Bioabsorbable xation devices: status for the
craniomaxillofacial surgeon, J. Craniofac. Surg., 8, 92, 1997.
11. Middleton, J.C. and Tipton, A.J., Synthetic biodegradable polymers as medical devices, Med. Plast.
Biomater., Mar/Apr, 30, 1998.
12. Barber, F.A., Resorbable xation devices: a product guide, Orthopedics, Spec. Ed., 4, 11, 1998.
13. Kohn, D.H., Richmond, E.M., Dootz, E.R., et al., In vitro comparison of parameters affecting the
xation strength of sagittal split osteotomies, J. Oral Maxillofac. Surg., 53, 1374, 1995.
14. Bos, R.R., Rozema, F.R., Boering, G., et al., Bio-absorbable plates and screws for internal xation
of mandibular fractures. A study in six dogs, Int. J. Oral Maxillofac. Surg., 18, 365, 1989.
15. Suuronen, R., Comparison of absorbable self-reinforced poly-L-lactide screws and metallic screws in
the xation of mandibular condyle osteotomies: an experimental study in sheep, J. Oral Maxillofac.
Surg., 49, 989, 1991.
16. Suuronen, R., Pohjonen, T., Vasenius, J., and Vainionp, S., Comparison of absorbable self-reinforced
multilayer poly-l-lactide and metallic plates for the xation of mandibular body osteotomies: an
experimental study in sheep, J. Oral Maxillofac. Surg., 50, 255, 1992.
17. McManners, J., Moos, K.F., and El-Attar, A., The use of biodegradable xation in sagittal split and
vertical subsigmoid osteotomy of the mandible: a preliminary report, Br. J. Oral Maxillofac. Surg.,
35, 401, 1997.
18. Gerlach, K.L. and Derichs, D., Comparative studies on the fatigue strength of absorbable polymers
used in oral and maxillofacial surgery, Dtsch. Zahnarztl. Z., 43, 376, 1988 [in German].
19. An, Y.H, Friedman, R.J., Powers, D.L., et al., Fixation of osteotomies using bioabsorbable screws in
the canine femur, Clin. Orthop., 355, 300, 1998.
20. Pietrzak, W.S., Sarver, D., and Verstynen, M., Bioresorbable implants: practical considerations, Bone,
19(Suppl. 2), 109S, 1996.
21. Athanasiou, K.A., Agrawal, C.M., Barber, F.A., and Burkhart, S.S., Orthopaedic applications for PLA-
PGA biodegradable polymers, Arthroscopy, 14, 726, 1998.
22. Hutmacher, D., Hurzeler, M.M., and Schliephake, H., A review of material properties of biodegradable
and bioresorbable polymers and devices for GTR and GBR applications, Int. J. Oral Maxillofac.
Implants, 11, 667, 1996.
23. Bucholz, R.W., Henry, S., and Henley, M.B., Fixation with bioabsorbable screws for the treatment of
fractures of the ankle, J. Bone Joint Surg., 76A, 319, 1994.
24. Pietrzak, W.S., Sarver, D.R., Bianchini, S.D., and DAlessio, K., Effect of simulated intraoperative
heating and shaping on mechanical properties of a bioabsorbable fracture plate material, Appl. Bio-
mater., 38, 17, 1997.
25. Simon, J.A., Ricci, J.L., and DiCesare, P.E., Bioresorbable fracture xation in orthopedics: a com-
prehensive review. Part I. Basic science and preclinical studies, Am. J. Orthop., 26, 665, 1997.
26. Bstman, O.M., Pihlajamaki, H.K., Partio, E.K., and Rokkanen, P.U., Clinical biocompatibility and
degradation of polylevolactide screws in the ankle, Clin. Orthop., 320, 101, 1995.
0266/ ch38/frame Page 590 Friday, October 22, 1999 11:15 AM
Fatigue Testing of Bioabsorbable Screws in a Synthetic Bone Substrate 591
27. Eitenmuller, J., David, J., Pommer, A., and Muhr, G., Die Versorgung von Sprunggelenksfrakturen
unter Verwendung von Platten und Schrauben aus resorbierbarem Polymermaterial, in Abstracts of
53 Jahrestagung der Deutschen Gesellschaft fr Unfallheilkunde, No. 64, Springer, Berlin, 1990.
28. Simon, J.A., Ricci, J.L., and DiCesare, P.E., Bioresorbable fracture xation in orthopaedics: A com-
prehensive review Part II. Clinical studies, Am. J. Orthop., 26, 754, 1997.
29. Bergsma, J.E., de Bruijn, W.C., Rozema, F.R., et al., Late degradation tissue response to poly
(L-lactide) bone plates and screws, Biomaterials, 16, 25, 1995.
30. Bstman, O., Hirvensalo, E., Makinen, J., and Rokkanen, P., Foreign body reactions to fracture xation
implants of biodegradable synthetic polymers, J. Bone Joint Surg., 74B, 592, 1990.
31. Casteleyn, P., Handelberg, F., and Haentjens, P., Biodegradable rods vs. Kirschner wire xation of
wrist fractures, J. Bone Joint Surg., 76A, 319, 1994.
32. Eppley, B.L. and Reilly, M., Degradation characteristics of PLLA-PGA bone xation devices,
J. Craniofac. Surg., 8, 116, 1997.
33. Eppley, B.L., Prevel, C.D., and Sadove, A.M., Resorbable bone xation: does it have a role in
craniomaxillofacial trauma? J. Craniomaxillofac. Trauma, 2, 56, 1996.
34. Salyer, K.E., Aesthetic Craniofacial Surgery, J.B. Lippincott, Philadelphia, 1989, chap. 7.
35. Collins, J.A., Failure of Materials in Mechanical Design Analysis, Prediction, Prevention, John
Wiley & Sons, New York, 1981, chaps. 7 and 11.
36. DeCoster, T.A., Heetderks, D.B., and Downey, D.J., et al., Optimizing bone screw pullout force,
J. Orthop. Trauma, 4, 169, 1990.
37. Hearn, T.C., Schatzker, J., and Wolfson, N., Extraction strength of cannulated cancellous bone screws,
J. Orthop. Trauma, 7, 138, 1993.
38. McKinley, T.O., McLain, R.F., Yerby, S.A., et al., The effect of pedicle morphometry on pedical screw
loading. A synthetic model, Spine, 22, 246, 1997.
39. Ziccardi, V.B., Scheider, R.E., and Kummer, F.J., Wurzburg lag screw plate vs. four-hole miniplate
for the treatment of condylar process fractures, J. Oral Maxillofac. Surg., 55, 602, 1997.
40. Thompson, J.D., Benjamin, J.B., and Szivek, J.A., Pullout strengths of cannulated and noncannulated
bone screws, Clin. Orthop., 341, 241, 1997.
41. Landsman, A.S. and Chang, T.J., Can synthetic bone models approximate the mechanical properties
of cadaveric rst metatarsal bone? J. Foot Ankle Surg., 37, 122, 1998.
42. Cristofolini, L., Vicconti, M., Cappello, A., and Toni, A., Mechanical validation of whole bone
composite femur models, J. Biomech., 29, 525, 1996.
43. Hayes, W.C. and Bouxsein, M.L., Biomechanics of cortical and trabecular bone: implications for
assessment of fracture risk, in Basic Orthopaedic Biomechanics, 2nd ed., Mow, V.C. and Hayes, W.C.,
Eds., Lippincott-Raven, Philadelphia, 1997, 85.
44. Hellsing, E. and Hagberg, C., Changes in maximum bite force related to extension of the head, Eur.
J. Orthodontol., 12, 148, 1990.
45. Floystrand, F., Kleven, E., and Oilo, G., A novel miniature bite force recorder and its clinical
application, Acta Odontol. Scand., 40, 209, 1982.
46. Waltimo, A. and Kononen, M., A novel bite force recorder and maximal isometric bite force values
for health young adults, Scand. J. Dent. Res., 101, 171, 1993.
47. Iwase, M., Sugimori, M., Kurachi, Y., and Nagumo, M., Changes in bite force and occlusal contacts
in pateints treated for mandibular prognathism by orthognathic surgery, J. Oral Maxillofac. Surg., 56,
850, 1998.
48. Kim, Y.G. and Oh, S.H., Effect of mandibular setback surgery on occlusal force, J. Oral Maxillofac.
Surg., 55, 126, 1997.
49. De Boever, J.A., McCall, W.D., Jr., Holden, S., and Ash, M.M., Jr., Functional occlusal forces: an
investigation by telemetry, J. Prosthet. Dent., 40, 326, 1978.
50. Youssef, R.E., Throckmorton, G.S., Ellis, E., and Sinn, D.P., Comparison of habitual masticatory
cycles and muscle activity before and after orthognathic surgery, J. Oral Maxillofac. Surg., 55, 699,
1997.
0266/ ch38/frame Page 591 Friday, October 22, 1999 11:15 AM
0266/ ch38/frame Page 592 Friday, October 22, 1999 11:15 AM

593

0-8493-0266-9/00/$0.00+$.50
2000 by CRC Press LLC

39

Testing Intervertebral Stability
after Spinal Fixation

Kenneth S. James and A. U. Daniels

CONTENTS

I. Introduction ..........................................................................................................................593
II. Spine Testing Methods.........................................................................................................594
A. Intervertebral Motion Measurement Devices ................................................................594
B. Spine Testing Load- and Displacement-Control Methods .......................................594
III. Spine Motion Control and Analysis System.......................................................................595
A. Spine Motion Control System .......................................................................................595
B. Relative Intervertebral Micromotion Measurement.......................................................596
C. Anatomical Characterization..........................................................................................598
D. Load Measurement System............................................................................................599
E. Data Acquisition.............................................................................................................599
F. Data Analysis and Reduction.........................................................................................599
IV. An Example Scoliosis Instrumentation Testing..............................................................600
A. Overview.........................................................................................................................600
B. Specimen Preparation.....................................................................................................601
C. Test Modes .....................................................................................................................601
D. Data Analysis and Reduction.........................................................................................602
E. Construct Stiffness .........................................................................................................602
F. Motion outside the Construct.........................................................................................602
G. Motion within the Construct ..........................................................................................603
V. Summary ..............................................................................................................................605
References ......................................................................................................................................605

I. INTRODUCTION

Implant components used to stabilize the spine are generally referred to as spinal instrumentation,
and the affected portion of the spine with instrumentation in place is described as a spine construct.
Too much intervertebral motion occurring in a spine construct is deleterious to achieving spine
fusion that the surgeon initiates to maintain long-term correction of spinal injury or deformity. By
measuring the amount of motion present within a spinal construct in the laboratory, it is possible
to assess the contribution that instrumentation design and specic construct components make to
the intervertebral stability the instrumentation provides.
Spinal instrumentation load-to-failure experiments,

1-3

individual component testing,

4,5

and con-
struct fatigue tests

6-8

are excellent mechanisms for initial screening of new instrumentation designs.
Less strong, less durable, less secure designs are weeded out early by these testing protocols.
Subsequent spinal construct intervertebral stability testing is an essential additional step in pursuing
a meticulous, clinically relevant spinal instrumentation analysis. Intervertebral stability can be

0266/ch39/frame Page 593 Tuesday, October 26, 1999 12:45 PM

594

Mechanical Testing of Bone and the BoneImplant Interface

dened as the amount of intervertebral displacement that occurs when a construct is subjected to
a loaddisplacement regimen believed to be clinically relevant.
Ashman et al.,

9

Smith,

10

Wilke et al.,

11

and Panjabi

12

have reviewed a number of

in vitro

methods
for applying and measuring spinal loads and intervertebral displacements in instrumented spines.
Limitations and pitfalls of such systems have similarly been discussed.

13,14

Many investigators have
adapted standard material testing devices to load and move spines. Others have developed an
apparatus specically for testing spines. This chapter describes the authors methodology for
providing an accurate means to control gross spine movements while concurrently measuring
intervertebral micromotion and the sacral load prole. The spine motion control system developed
here is capable of moving uninstrumented and instrumented cadaver spines through physiological
ranges of exion, extension, lateral bending, and torsion. Parameters such as cycling frequency and
motion excursion amplitude are controlled through a custom software interface. Intervertebral
micromotion is recorded in three dimensions with a custom-built optoelectronic laser/charge-
coupled device (CCD) system. Motion and load data are computed and recorded concurrently in
real time by an Intel microprocessor-based personal computer. Differences in spinal instrumentation
and stability are assessed by comparing intervertebral motion and load data.

II. SPINE TESTING METHODS
A. I

NTERVERTEBRAL

M

OTION

M

EASUREMENT

D

EVICES

Displacement transducers can be divided into linked and nonlinked types. Linked types generally
are electromechanical and introduce the possibility of altering normal spinal motion through
their linkages from one vertebra to another. A popular linked motion transducer is the LVDT
(linear variable differential transducer). Panjabi et al.

15

produced a displacement transducer
system based on LVDTs. Another popular linked motion transducer is the electrogoniometer.
This mechanism is capable of measuring relative vertebral movement in all six degrees of freedom
by using linkages connected to the vertebral bodies and potentiometers oriented in each plane
of motion.

16

Nonlinked motion measurement methods do not affect the intervertebral motion and are
potentially superior to linked methods. Systems have been devised using optoelectronic and
stereophotogrammetry techniques.

17,18

Stereophotogrammetry has the disadvantage that the accu-
racy is inversely related to the viewing volume.

19

Historically, optoelectronic systems had proved
too bulky for measuring intervertebral micromotion. This changed with the development of the
laser-CCD devicebased motion measurement system.

17,20

This device made very accurate, non-
contacting three-dimensional measurement of multilevel intervertebral micromotion possible.
Advantages of this system include miniaturization of the components, the elimination of the need
for mechanical linkages between the bodies, and a high degree of accuracy which is not dependent
on viewing volume.

19,21

B. S

PINE

T

ESTING

L

OAD

-

AND

D

ISPLACEMENT

-C

ONTROL

M

ETHODS

The spine is one of the most complex mechanical structures of the body and is able to produce
and undergo complicated three-dimensional movements.

In vitro

spine testing protocols attempt
to replicate the

in vivo

situation. However, simulating active paraspinous musculature, intraab-
dominal pressure, intercostal muscles, and the dynamic uid exchange through the disk and
vertebra, for example, is currently not practical in the laboratory. Only the mechanical contribu-
tions of the passive bony and cartilaginous components of the spine to strength or stability can
generally be tested.
One spine testing method available to the researcher is to control the load applied to the spine,
e.g., 10 Nm exion moment, and measure the resulting vertebral displacements a load-controlled

0266/ch39/frame Page 594 Tuesday, October 26, 1999 12:45 PM

Testing Intervertebral Stability after Spinal Fixation

595

experiment. This is a popular testing method, partly arising from the relative ease with which load-
controlled experiments can be performed.

9,12,15,22

However, the alternative testing approach, con-
trolling gross spine displacement and measuring the resulting loads and intervertebral motion a
displacement-control protocol offers the benets of better simulating

in vivo

motions, has the
best supporting data for most biomechanical situations and, consequently, offers the opportunity
for obtaining data of greater clinical relevance.

23, 24

For the present displacement control studies to mimic,

in vitro

, the natural movements of the
spine, a motion control system was built to guide the spine through dened physiological ranges
of exion, extension, lateral bending, and torsion. The system duplicates

in vitro

the direction and
amount of vertebral motions seen

in vivo

on radiographs by applying the loading needed to achieve
natural spine movement. What is more, controlling gross spine motion and measuring the resultant
loading and intervertebral motion grounds the resulting intervertebral stability data in the clinical
realm of actual patient spine bending or torsional movements.

III. SPINE MOTION CONTROL AND ANALYSIS SYSTEM
A. S

PINE

M

OTION

C

ONTROL

S

YSTEM

Located at the center of the spine motion load frame (Figure 39.1) are two 4-in. steel specimen
mounting cups. The inferior cup is xed to a stepper motor (25,000 motor steps per 360

of axial

FIGURE 39.1

The spine motion control system is shown here with a single laser/CCD unit attached to the
mounted spine. See text for details. (From James, K.S., Ph.D. thesis, University of Utah, Salt Lake City, 1995.)

0266/ch39/frame Page 595 Tuesday, October 26, 1999 12:45 PM

596

Mechanical Testing of Bone and the BoneImplant Interface

rotation; Compumotor PC23, Parker Hannifan Corp, Petaluma, CA); (hereafter referred to as the
torsional stepper motor), and the superior cup is housed within a rigid, movable crosshead structure.
Each end of the crosshead is rigidly mounted to an aluminum carriage (Thomson System 2DA,
Port Washington, NY) with four close-tolerance, low-friction, linear bearings. Each carriage slides
on a vertically mounted dual-shaft rail assembly. One of these horizontally sliding carriages is
mounted to a drive screw, which is driven by a second identical stepper motor (25,000 motor steps
per 2.54 cm of linear motion). The other horizontally sliding carriage travels freely upon its rail.
The result is a horizontally driven crosshead which is free to move passively in the vertical direction.
A gimbal system within the crosshead allows the superior mounting cup to pivot about two
orthogonal axes (exion/extension and lateral bending) while remaining xed about the torsional
axis. The horizontal crosshead, free to move in the vertical direction, is initially counterbalanced
to apply zero axial load to the spine specimen. Static weights may be applied to the crosshead to
provide vertical distractive or compressive loads, if required.
To take advantage of the accuracy of the actuators and achieve reproducible mechanical
performance, a rigid loading frame supports the aforementioned horizontally mounted dual shaft
rail assemblies. The height of the mounted rails can be adjusted to accommodate spine specimens
of any length normally seen in humans, e.g., the entire thoracolumbar spine.
To produce exion or extension, the spine specimen is placed in the mounting cups such that
the anterior-posterior (AP) plane of the spine corresponds with the plane of primary crosshead
movement. Under microcomputer control, the linear motion stepper motor is actuated, moving the
crosshead and therefore the cephalad-most vertebral body at a software-controllable frequency,
number of cycles, and range of motion. The inferior cup is held xed by the locked torsional stepper
motor, thereby immobilizing the sacral region of the spine specimen, much the same way the pelvis
does in the intact human.
Lateral bending of the spine specimen is made possible by rotating the spine 90

such that the
AP plane of the spine is perpendicular to the primary crosshead movement. In addition, the most
caudal vertebral body can be subjected to cyclic torsion about the spine axis by the torsional motion
stepper motor, either while the spine is held in place or during bending motion, thereby simulating
trunk rotation. The motors are controlled by a custom-designed software interface and can resolve
the position of the horizontal crosshead to within

0.005 mm, and the position of the torsional
motor to

5 arc-min.

B. R

ELATIVE

I

NTERVERTEBRAL

M

ICROMOTION

M

EASUREMENT

Three-dimensional intervertebral micromotion is measured with an optical transducer system con-
sisting of a mutually orthogonal cluster of three miniature AlGaAS laser diodes (LDs) (ML4102A,
Mitsubishi Electronic America, Inc., Sunnyvale, CA) which is mounted to one vertebral body with
bone screws and is directed at a corresponding array of three mutually orthogonal plane CCDs
(CCD222, Fairchild Weston Systems, Inc., Sunnyvale, CA) mounted to an adjacent body
(Figure 39.2, 39.3). The CCDs are light-sensitive video screens with over 185,000 (380

488)
sensing elements in a rectangle of 8.8

11.4 mm, a density high enough to detect the motion of
a collimated laser beam at a resolution of 30

m. The collimated laser beams effectively originate
from the same point in space. Both the CCDs and LDs are packaged in custom-made brackets
occupying the volume of a 4-cm cube, and weigh approximately 80 g. As one vertebra moves
relative to another, the three points illuminated by the lasers on the CCDs change accordingly, and
uniquely determine the spatial orientation of the laser array with respect to the CCD array. The
maximum translation measurable in any one direction is 8.8 mm. If the axis of rotation is located
at the laser origin, the system is capable of measuring angulation up to 20

. In practice, however,
the axis of rotation is located within the vertebral bodies, permitting angulation up to 10

to be
measured. With proper mounting, the entire typical range of intervertebral motion of the lumbar
or thoracic spinal motion segments can be measured.

25

0266/ch39/frame Page 596 Tuesday, October 26, 1999 12:45 PM

Testing Intervertebral Stability after Spinal Fixation

597

The static orientations of the LD and CCD, with respect to an anatomically based coordinate
system of the vertebral body to which they are mounted, are determined prior to dynamic testing
of the spine utilizing a three-dimensional digitizer (see below). The relative three-dimensional
motion of a vertebral pair itself is derived through a series of coordinate transformations which
relate the motion sensed by the laser/CCD arrays to the actual relative motion of the vertebral
bodies. The mathematical basis of this system is thoroughly described by Dean et al.

17

The system
is calibrated with a six-degree-of-freedom micrometer (Edmund Scientic, Barrington, NJ). Tests
with this micrometer have established the accuracy of the system to be 50

m in translation and
0.1

in rotation.

17

Newer, higher-resolution CCDs (over a million pixels) now available offer the

FIGURE 39.2

The laser/CCD motion measurement arrays used to measure relative intervertebral motion.
(From James, K.S., Ph.D. thesis, University of Utah, Salt Lake City, 1995.)

FIGURE 39.3

The laser/CCD motion measurement system mounted on prepared human cadaver spines, to
measure relative intervertebral motion as the spine specimens are moved through physiological motion excursions.

0266/ch39/frame Page 597 Tuesday, October 26, 1999 12:45 PM

598

Mechanical Testing of Bone and the BoneImplant Interface

promise of even greater accuracy. It is possible to use multiple LD/CCD systems to measure
intervertebral motion at several levels simultaneously.

C. A

NATOMICAL

C

HARACTERIZATION

To be of value, motion data recorded by the measurement system must be related to the anatomy
of the vertebral bodies and their initial relative positions. Many researchers reference motions and
positions to the center of the vertebral body, and a vertical axis which passes through it identies
the orientation of the body in space.

9,22

A method for repeatably determining a coordinate system for the vertebral bodies is based on
measuring external geometries with a highly accurate three-dimensional coordinate measuring
machine (Figure 39.4, Microval, Brown & Sharpe, North Kingstown, RI). The anterior surface of
each vertebral end plate is digitized down to the bases of the transverse processes. A centroid is
calculated based upon each semiellipse, and a line through the two centroids represents the vertical
coordinate axis, with the midpoint as the local origin. The tip of the spinous process is digitized
to provide a posterior planar direction and subsequently to establish the anterior axis. The third
axis, in the left lateral direction, is easily resolved from the previously determined axes.
Once the CCDs and lasers are mounted anterolaterally to their respective vertebral bodies, their
positions are digitized relative to the spine anatomy. The nal relationship appears as rotation
matrices and translation vectors

17

which are used in online calculations to resolve dynamically
relative vertebral body position in real time.
If the vertebral bodies are treated as rigid bodies, the relative translation between any two
anatomical points can be calculated so long as the relative positions of these points, e.g., points on
the facet joint, have likewise been digitized (Figure 39.5).

FIGURE 39.4

The three-dimensional, relative position relationships between the vertebral bodies and the
corresponding mounted laser and CCD arrays are established with a highly accurate three-dimensional
digitizer. The geometry of the vertebral bodies is digitized to approximate the vertebral body centers. Other
anatomical points are digitized to establish the orientation of the local coordinate system for each vertebral
body. Similarly, when relative motion between two specic points is of interest, e.g., the facet joint, additional
points are digitized on the adjacent vertebrae.

0266/ch39/frame Page 598 Tuesday, October 26, 1999 12:45 PM

Testing Intervertebral Stability after Spinal Fixation

599

D. L

OAD

M

EASUREMENT

S

YSTEM

A load cell (AMTI, Newton, MA) is mounted between the inferior mounting cup and the torsional
motion stepper motor. It is capable of measuring three orthogonal forces and three orthogonal
moments, thereby providing a complete load prole at the sacrum. The resolution of these forces
and moments are 0.075 N and 0.015 Nm, respectively, and are recorded through a multichannel
analog-to-digital board (Keithley Instruments, Cleveland, OH) mounted in the computer.

E. D

ATA

A

CQUISITION

For spine specimens subjected to dynamic motion, rates of computerized data acquisition per motion
cycle depend on microprocessor speed, the number of vertebral levels under evaluation, and the
number of other parameters measured (e.g., resultant loads and gross spine position). One can
routinely sample and compute motion data from three spinal levels plus sacral load data concurrently
at 15 Hz with an Intel microprocessor-based personal computer.

F. D

ATA

A

NALYSIS



AND

R

EDUCTION

In quantifying three-dimensional vertebral motion, every researcher is faced with the difcult
problem of interpreting the results in a manner which is clinically relevant and facilitates specimen-
to-specimen comparisons. The coordinate system which best relates the spatial representations of
the vertebral bodies will be one whose separate planes correspond to the common clinical anatomical
planes, namely, the midsagittal, transverse, and frontal. The authors have oriented their coordinate
system with the

x

-axis aligned anteriorly, the

y

-axis laterally, and the

z

-axis pointed in the superior
direction (Figure 39.6). The separate components of the translation vector therefore correspond to
motion parameters familiar to the clinician, while the magnitude of the resultant vector can serve
to quantify the complete amount of translation.
The rotational matrix can be somewhat more challenging to interpret due to the complex nature
of three-dimensional rotation and the noncommunitivity of sequential rotation.

12,15,22

While other
researchers have referenced the rotational matrix as axial rotations, the authors have chosen a
projection method similar to Pearcy

26

which provides a unique solution to the interpretation of the
matrix in parameters familiar to the clinician in viewing radiographic lms, which are also projec-

FIGURE 39.5

Relative vertebral motion is commonly referenced to the vertebral body centers. However,
because the vertebrae are treated as rigid bodies, the motion between other sites, e.g., the facet or midsagittal
anterior disk, can also be calculated so long as the positional relationships have been established during the
digitization step. For instance, calculating the translations at additional sites is helpful in analyzing the motion
present at the fusion site.

0266/ch39/frame Page 599 Tuesday, October 26, 1999 12:45 PM

600

Mechanical Testing of Bone and the BoneImplant Interface

tions.

17

Rotations of the coordinate system of the superior vertebra are projected onto the planes
of the inferior coordinate system, dened as the midsagittal plan, transverse plane, and frontal plane.
By using this projection method, each axis of one vertebral body provides a different rotation
parameter when projected onto the appropriate plane of the adjacent body (see Figure 39.5). To
measure exion/extension, for example, a line parallel to the vertical axis of the body is transformed
into the

xz

plane of the adjacent body, which approximates the midsagittal plane. The angles
between the lines comparing the neutral position with the exed and extended spine dene the
relative angular motion of exion and extension, respectively. Similarly, the anterior axis projected
onto the transverse (

xy

) plane quanties torsion of the spinal column, and the lateral axis projected
onto the frontal (

yz

) plane quanties lateral bending.

IV. AN EXAMPLE SCOLIOSIS INSTRUMENTATION TESTING
A. O

VERVIEW

The evolution of scoliosis instrumentation has been marked recently by the introduction of bulkier,
increasingly stiff implants. Presumably, these stiffer constructs are intended to allow less intervertebral
motion within the instrumented area which, in turn, provides a more favorable environment con-
ducive to the formation of the bony fusion mass needed to maintain correction of the deformity.
The question presents itself During patient movement, is there a signicant difference in the
amount of intervertebral motion present within scoliosis instrumentation systems of varying com-
plexities, sizes, and stiffnesses?

FIGURE 39.6

To aid in interpreting the three-dimensional rotation matrix, rotations of the coordinate system
of the superior vertebra are projected onto the planes of the inferior coordinate system, dened as the
midsagittal plane, transverse plane, and frontal plane.

0266/ch39/frame Page 600 Tuesday, October 26, 1999 12:45 PM

Testing Intervertebral Stability after Spinal Fixation

601

In this example, the authors investigated how a spectrum of scoliosis instrumentation systems
differs in their respective stiffnesses and their abilities to restrict intervertebral motion. This study
assessed the initial effect that Harrington (the historic standard), Cotrel Dubousset (CD), and Luque
instrumentation have on both intervertebral micromotion and spine stiffness when these instrumentation
systems span the spinal levels commonly instrumented in scoliosis correction. In addition, a protocol
was employed to examine how removal of selected components in the CD constructs decreased stiffness
and whether this correlated with a measured increase in intervertebral motion. This tear-down protocol
helps in determining specic component contribution to the stability of the scoliosis construct.
The change in relative intervertebral motion immediately caudal to the instrumentation was
also investigated. In an instrumented patient, intervertebral motion in the lumbar spine increases
to accommodate the exibility lost in the fused segments. Clinically, the additional demand placed
on the unfused motion segments has been implicated in accelerated adjacent disk degeneration and
higher incidences of low back pain.

27, 28

B. S

PECIMEN

P

REPARATION

Six human cadaver spines (one female, age 57 and ve males, ages 19, 36, 37, 39, and 47) were
harvested and frozen within 24 h after death at 20

in evacuated, heat-sealed polyethylene bags.
Prolonged storage under these conditions and subsequent thawing has little effect on a spines
biomechanical properties.

29

Gross dissection involved removal of the cervical spine and the distal
sacrum and pelvis from the specimen. The ribs were resected such that 5 cm remained in order to
preserve the costotransverse joint and the associated ligament. All remaining associated musculature
(except for the intercostal muscle associated with the remaining ribs) was removed. All ligamentous
structures were left intact, and AP and lateral radiographs were taken to assess any degeneration or
bony abnormalities. The specimens present a limitation for scoliosis studies in that they are normal
rather than scoliotic. However, spine biomechanics researchers generally agree that this does not seem
likely to change the relative ability of the instrumentation to control intervertebral motion.
The most cephalic (T1) and caudal (S1 and S2) vertebral bodies of a given spine were potted
in 10-cm-diameter aluminum cups with a low-melting-point lead alloy. The specimens were potted
such that the axial rotation axis of the spine is perpendicular to the L3/L4 disk. Metal pins and
screws inserted through the distal and proximal end structures assured a stable xation throughout
the duration of the test.
While thawed, specimens are kept moist with saline. Loosely wrapping the specimen in
household plastic lm helps to prevent the specimen from drying.

C. T

EST

M

ODES

The spine was xed with commercially available Harrington distraction/compression, CD and Luque
instrumentation from T4 to L2 in a conguration typical for a right thoracic idiopathic scoliosis curve.
The instrumentation test sequence performed on each spine was as follows: Harrington distraction/com-
pression system, full CD construct, CD construct without devices for transverse traction (DTTs), a
single left CD rod, and a full Luque construct. The instrumentation testing protocol had to be ordered
in this manner and not randomized in order to minimize the tissue damage caused by one instrumen-
tation system that would interfere with the xation of another system tested.
Three laser/CCD units were used to make intervertebral motion measurements at two levels
within each construct, T7T8 and T10T11, and at a level immediately outside the applied instru-
mentation, L2-L3 (see Figure 39.3A). The spines were cycled at 0.1 Hz to gross bending angles
of 15

exion, 10

extension, 15

right and left bending, and 15

right and left torsion. Clinical
radiographic studies have shown that normal patients can voluntarily achieve these angles. Each
specimen was precycled 20 times for each of the instrumentation congurations and gross motion
imposed with intervertebral motion recorded on the subsequent cycle. A steep acceleration ramp

0266/ch39/frame Page 601 Tuesday, October 26, 1999 12:45 PM

602

Mechanical Testing of Bone and the BoneImplant Interface

was used to ensure that induced movements were at a constant velocity throughout the majority
of a motion cycle. The neutral position used to differentiate between exion and extension, right
and left torsion, and right and left bending was dened as the position where no moments or lateral
forces were measured at the base of the spine. During testing, sacral load data were recorded
concurrently with intervertebral motion data. Data was recorded at 15 Hz. This yielded 150 load
and position readings per spine level per motion cycle.

D. D

ATA

A

NALYSIS AND REDUCTION
The resultant relative motion between adjacent vertebra is quite small. Maximum primary projection
angles are typically less than 3, and the translation vector less than 3 mm. This conrms the need
for a sensitive, accurate motion transducer system.
Primary angular micromotion was consistent with the induced gross motion. In other words,
when the spine was placed in exion and extension, the majority of the angular motion was seen
in the midsagittal projection plane. Likewise, for lateral bending the primary angular motion was
seen in the frontal plane, and induced torsion resulted in primary angulation in the transverse plane.
Subsequent analyses focused on the changes measured in primary angulation. Similarly, although
the load acquisition system provides a complete description of sacral loading, the primary moment
measured is deemed sufcient for comparing construct stiffness.
Motion and stiffness data were normalized to the uninstrumented spine data. Intervertebral
micromotion seen at the spinal levels monitored are present for selected gross spine motions: torsion
and lateral bending at T7/T8 and T10/T11 and exion, extension, lateral bending, and torsion at
L2/L3. Flexion and extension data are not presented for the thoracic segments because the facet
orientation inherently limited the motion in these planes to an extent below the measurement
capabilities of the system.
E. CONSTRUCT STIFFNESS
In general, the application of the instrumentation systems increased average spine stiffness by 70%
(Figure 39.7) over the uninstrumented spines. The elegantly simple Harrington rods which are
attached to the spine only at the rod ends matched the stiffness of Luque instrumentation which,
in contrast, is attached to the spinal column at every level within the construct. The greatest disparity
in spine stiffness was in torsion, where spines instrumented with CD instrumentation were over
twice as stiff as the average Harrington and Luque constructs.
Removal of the DTTs and subsequently the right rod of the CD construct resulted in stiffness
decreases of approximately 15 and 55%, respectively, for all motion excursions. Notably, the one-
rod CD construct approached the stiffness of the two-rod Harrington and Luque constructs. This
testing method found that the CD instrumentation design, utilizing a combination of pedicle and
laminar hooks (5 hooks per rod) and the largest diameter rods (6.8 mm compared with 6.4 mm
diameter rods) offered the most resistance to an imposed displacement. In fact, that data suggest
it may be overdesigned and could be streamlined.
F. MOTION OUTSIDE THE CONSTRUCT
Intervertebral motion (angulation and translation) at L2/L3, the segment immediately caudal to the
construct, typically increased 30% upon application of the spinal instrumentation systems
(Figure 39.8) except for torsion, where motion increased upward of 65% as compared with the
uninstrumented spines. Torsional motion is greatly increased because most of the thoracic spine
that normally accommodates torsion had been instrumented. This measured increase in motion was
not a function of the instrumentation conguration used since each of the instrumentation systems
were grossly comparable in xing the intervertebral motion within the construct. Each construct,
therefore, displaced similar amounts of motion caudal to the instrumentation.
0266/ch39/frame Page 602 Tuesday, October 26, 1999 12:45 PM
Testing Intervertebral Stability after Spinal Fixation 603
By using such a displacement control protocol it is possible in vitro to measure magnitudes of
displaced motion systematically as a function of the number of levels fused and other instrumen-
tation designs. Clinically, only fusions extending below L2 have been implicated in increased back
pain. Consequently, it should be possible to establish motion baselines beyond which accelerated
degeneration and low back pain may result.
G. MOTION WITHIN THE CONSTRUCT
Intervertebral micromotion measurements made within the constructs at T7/T8 and T10/T11
revealed no signicant difference between the Harrington, CD, and Luque instrumentation
systems (Figures 39.8 and 39.9). Typically, primary angulation was reduced 40 to 50% and
translation of the vertebral body centers 10 to 18%. The contribution of the DTTs and the second
rod to the CD construct to restricting micromotion was unremarkable. The one-rod CD construct
was, on average, equivalent to the two-rod Harrington construct in managing intervertebral motion
within the instrumentation.
It is helpful to analyze motion at sites other than the vertebral body centers, e.g., the fusion site,
in order to appreciate movements not immediately apparent in the data.
30
For example, in an extreme
case, it is possible for a vertebra to rotate about its body center. In this case no relative translation,
only rotation, would be measured. This can be misleading because at the facet, for example, where
fusion is to occur, there is signicant translation (see Figure 39.5). Studying the motion at different
landmarks across the intervertebral joints can provide more clinically useful information regarding
spinal motion than simply studying the translations of one point, especially the vertebral body centers.
The authors nd that referencing intervertebral micromotion to more than one landmark helps in
comprehending the complex three-dimensional translations and angulations measured. Referencing
motion to the area where the fusion mass forms is of special importance.
FIGURE 39.7 Average stiffness for all instrumentation configurations and motions tested. For full constructs,
a 70% increase in overall spine stiffness is typically measured. Removal of instrumentation components was
accompanied by a decrease in construct stiffness, with a one-rod construct being about half as stiff as the
two-rod construct. (From James, K.S., Ph.D. thesis, University of Utah, Salt Lake City, 1995.)
0266/ch39/frame Page 603 Tuesday, October 26, 1999 12:45 PM
604 Mechanical Testing of Bone and the BoneImplant Interface
FIGURE 39.8 Normalized angulation measured inside and outside the constructs tested. Each construct
provided comparable intervertebral stability within the construct. Similarly, motion increases caudal to the
constructs were comparable for each type of instrumentation. (From James, K.S., Ph.D. thesis, University of
Utah, Salt Lake City, 1995.)
FIGURE 39.9 Through coordinate transformations, it is possible to measure the translations between any
two digitized or calculated points on adjacent vertebrae. Here, the relative translation between the vertebral
body centers (T7/T8) and facet are compared for right bending. Differences between instrumentation types
are accentuated when the translation at the facet is calculated and analyzed. This is important because
translation at the fusion site can be of greater clinical signicance.
0266/ch39/frame Page 604 Tuesday, October 26, 1999 12:45 PM
Testing Intervertebral Stability after Spinal Fixation 605
To measure motion at the facet, the relative location of the right side facet to the vertebral body
centers was digitized in three dimensions. The maximum translation at the facet joint and the
vertebral body center was measured as the spine was moved from the neutral position to the gross
motion extremes. Figure 39.9 shows the translation between the vertebral body centers and the
facet at T7/T8 for the case of right bending. Clearly, the translation is markedly different depending
on the reference site. There are differences in translation at the facet site that are not readily
appreciated when simply considering motion at the vertebral body centers. Translation at the facet
was reduced 25 to 40% by the instrumentation systems compared with 10 to 18% when the relative
motion between the vertebral body centers was considered. In general, there are greater translations
at the facet than at the body center because the axis of rotation of the vertebral bodies is closer to
the body center than the facet. More importantly, these motions reect the translation of the facet
joint to be fused. At the facet, the CD system appears somewhat better in inhibiting motion when
compared with the Harrington and Luque systems.
Similar to the stiffness results, the one-rod CD construct proved comparable with the two rod
Harrington and Luque systems in intervertebral xation at T7/T8 and T10/T11. This again suggests
the possibility of streamlining the CD design. Such data, if borne out in additional testing, raise
fundamental questions regarding the signicance of the added implants that are present when one
uses traditional, two-rod constructs with cross-linking. At the clinical extreme, one might achieve
the same fusion rate with a single rod as with a double rod while at the same time minimizing
hooks and increasing the surface area available for fusion. These data are germane, however, only
to the restriction of intervertebral micromotion and the respective stiffness of the constructs. The
questions of yield strength, fatigue life, and loosening of a single-rod construct have to be considered
in further studies.
V. SUMMARY
This approach to evaluating the stability provided by spinal constructs employs instrumented human
cadaver spines which are moved through physiological ranges of motion under displacement control.
Loading and intervertebral motion are measured concurrently. Systematically analyzing a variety
of instrumentation systems and the contribution of individual components to the overall stability
of the construct provides a means of characterizing and optimizing instrumentation design. Although
the example given compared posterior scoliosis instrumentation systems, the same biomechanical
testing and analysis methods are also readily adapted to anterior instrumentation, spinal fracture
models, and spinal fracture instrumentation congurations.
REFERENCES
1. Gurr, K., McAfee, P., and Shih, C., Biomechanical analysis of posterior instrumentation systems after
decompressive laminectomy: an unstable calf-spine model, J. Bone Joint Surg., 70A, 680, 1988.
2. Jacobs, R., Nordwall, A., and Nachemson, A., Reduction, stability, and strength provided by internal
xation systems for thoracolumbar spinal injuries, Clin. Orthop., 171, 300, 1982.
3. Wenger, D., Carollo, J., Wilkerson, J., and Herring, J., Laboratory testing of segmental spinal instru-
mentation vs. traditional Harrington instrumentation for scoliosis treatment, Spine, 7, 265, 1982.
4. Krag, M., Beynnon, B., Pope, M., and DeCoster, T., Depth of insertion of transpedicular vertebral
screws into human vertebrae: effect upon screwvertebra interface strength, J. Spinal Disord., 1, 287,
1989.
5. Freedman, L., Houghton, G., and Evans, M., Cadaveric study comparing the stability of upper
distraction hooks used in Harrington instrumentation, Spine, 11, 579, 1986.
6. Sutterlin, C., McAfee, P., Warden, K., et al., A biomechanical evaluation of cervical spinal stabilization
methods in a bovine model: static and cyclic loading, Spine, 13, 795, 1988.
0266/ch39/frame Page 605 Tuesday, October 26, 1999 12:45 PM
606 Mechanical Testing of Bone and the BoneImplant Interface
7. Ashman, R., Birch, J., Bone, L., et al., Mechanical testing of spinal instrumentation, Clin. Orthop.,
227, 113, 1988.
8. Gaines, R., Carson, W., Satterlee, C., and Groh, G., Experimental evaluation of seven different spinal
fracture internal xation devices using nonfailure stability testing: the load sharing and unstable
mechanism concepts, Spine, 16, 903, 1991.
9. Ashman, R., Bechtold, J., Edwards, W., et al., In vitro spinal arthrodesis implant mechanical testing
protocols, J. Spinal Disord., 2, 274, 1989.
10. Smith, T., In vitro spinal biomechanics: experimental methods and apparatus, Spine, 16, 1204, 1991.
11. Wilke, H.J., Wenger, K. and Klaes, L., Testing criteria for spinal implants: recommendations for the
standardization of in vitro stability testing of spinal implants, Eur. Spine J., 7, 148, 1998.
12. Panjabi, M., Biomechanical evaluation of spinal xation devices: I. A conceptual framework, Spine,
10, 1129, 1988.
13. Adams, M.A., Mechanical testing of the spine An appraisal of methodology, results and conclusion,
Spine, 20, 2151, 1995.
14. Kostuik, J. and Smith, T., Pitfalls of biomechanical testing, Spine, 16, 1233, 1991.
15. Panjabi, M.M., Krag, M.H., and Goel, V.K., A technique for measurement and description of three-
dimensional six degree of freedom motion of a body joint with an application to the spine, J. Biomech.,
14, 447, 1981.
16. Wilke, H.J., Claes, L., Schmitt, H., and Wolf, S., A universal spine tester for in vitro experiments with
muscle force simulation, Eur. Spine J., 3, 91, 1994.
17. Dean, J.C., Wilcox, C.H., Daniels, A.U., et al., A new method for assessing relative dynamic motion
of vertebral bodies during cyclic loading in vitro, J. Biomech., 24, 1189, 1991.
18. Goel, V., Nye, T., Clark, C., et al., A technique to evalute an internal spinal device by use of the
Selspot system: an application to Luque closed loop, Spine, 12, 150, 1987.
19. Fioretti, S., Germani, A., and Tommaso, L., Sterometry in very close-range sterophotogrammetry with
non-metric cameras for human movement analysis, J. Biomech., 18, 831, 1985.
20. van Wagoner, E., Method and Apparatus for Sensing and Measuring Relative Position and Motion
between Two Points, U.S. Patent 5150169, 1992.
21. Goel, V.K., Clark, C.R., Harris, K.G., and Schulte, B.S., Kinematics of the cervical spine: effects of
multiple total laminectomy and facet wiring, J. Orthop. Res., 6, 611, 1988.
22. Panjabi, M.M., Brand, R.A., and White, A.A., Mechanical properties of the human thoracic spine as
shown by three-dimensional load displacement curves, J. Bone Joint Surg., 58A, 642, 1976.
23. Goel, V.K., Wilder, D.G., Pope, M.H., and Edwards, W., Biomechanical testing of the spine: load-
controlled vs. displacement-controlled analysis, Spine, 20, 2354, 1995.
24. James, K.S., Scoliosis Construct and Burst Fracture Stability: In Vitro and Clinical Results Suggesting
More Conservative Treatment Strategies for Scoliosis and Burst Fractures, Ph.D. thesis, University
of Utah, Salt Lake City, 1995.
25. White, A.A. and Panjabi, M.M., The basic kinematics of the human spine. A review of past and current
knowledge, Spine, 3, 12, 1978.
26. Pearcy, M.J., Stereo radiography of lumbar spine motion, Acta Orthop. Scand., 56 (Suppl. 212), 7,
1985.
27. Fabry, G., Backpain after Harrington rod instrumentation for idiopathic scoliosis, Spine, 14, 620, 1989.
28. Hayes, M.A., Tompkins, S.F., Herndon, W.A., et al., Clinical and radiological evaluation of lumbo-
sacral motion below fusion levels in idiopathic scoliosis, Spine, 13, 1161, 1988.
29. Smeathers, J.E. and Joanes, D.N., Dynamic compressive properties of human lumbar intervertebral
joints: a comparison between fresh and thawed specimens, J. Biomech., 21, 425, 1988.
30. Goodwin, R.R., James, K.S., Daniels, A.U., Stans, A.A., and Dunn, H.K., A new approach in analysis
of intervertebral micromotion: multiple reference point evaluation of spinal instrumentation. Trans.
39th Annual Meeting of the Orthopedic Research Society, San Francisco, CA, February 1518, 1993.
0266/ch39/frame Page 606 Tuesday, October 26, 1999 12:45 PM

607

Appendix 1
Unit Conversions

TABLE A1.1 Force

lb gm kg N dyne

1 lb 1 453.6 0.4536 4.448 4.448

10

5

1 gm 2.205

10

3

1 10

3

9.807

10

3

980.7
1 kg 2.205 10

3

1 9.087 9.807

10

5

1 N 0.2248 1.020

10

2

0.1020 1 10

5

1 dyne 2.248

10

6

1.020

10

3

1.020

10

6

10

5

1

TABLE A1.2 Stress

psi gm/mm

2

dyne/cm

2

kg/cm

2

Pa MPa

1 psi 1 0.703 6.895

10

4

7.031

10

2

6.895

10

3

6.895

10

3

1 gm/mm

2

1.422 1 9.807

10

4

0.1 9.807

10

3

9.807

10

3

1 dyne/cm

2

1.450

10

5

1.020

10

5

1 1.020

10

6

0.1 10

5

1 kg/cm

2

14.226 10 9.807

10

5

1 9.807

10

4

9.807

10

2

1 Pa
1 N/m

2

1.450

10

4

1.020

10

4

10 1.020

10

5

1 10

6

1 MPa
1 MN/m

2

1 N/mm

2

145.05 1.020

10

2

10

7

10.20 10

6

1

TABLE A1.3 Length

mm cm m in. ft

1 mm 1 10

1

10

3

3.937

10

2

3.281

10

3

1 cm 10 1 10

2

0.397 3.281

10

2

1 m 10

3

10

2

1 39.37 3.281
1 in. 25.40 2.540 2.540

10

2

1 8.333

10

2

1 ft 3.048

10

2

30.48 0.3048 12 1

TABLE A1.4 Area

mm

2

m

2

in.

2

ft

2

1 mm

2

1 10

6

1.550

10

3

1.076

10

5

1 m

2

10

6

1 1550 10.76
1 in.

2

6.452

10

2

6.452

10

4

1 6.944

10

3

1 ft

2

9.29

10

4

9.290

10

2

144 1

0266/Ap1/frame Page 607 Wednesday, October 27, 1999 12:57 PM

0266/Ap1/frame Page 608 Wednesday, October 27, 1999 12:57 PM

609

Appendix 2
Useful Journals and Books
on Bone Mechanics

USEFUL JOURNALS

Acta Orthopaedica Scandinaviaca

(12 issues, Scandinavian University Press)

Biomaterials

(24 issues, Elsevier Science)

Bone

(12 issues, Elsevier Science)

Calcied Tissue International

(12 issues, Springer-Verlag, New York)

Clinical Biomechanics

(10 issues, Elsevier Science)

Clinical Orthopaedics and Related Research

(12 issues, Lippincott Williams & Wilkins)

Journal of Applied Biomechanics

(4 issues, Human Kinetics)

Journal of Applied Physiology

(12 issues, American Physiological Society)

Journal of Arthroplasty

(6 issues, Churchill Livingstone)

Journal of Biomechanics

(12 issues, Elsevier Science)

Journal of Biomechanical Engineering

(4 issues, American Society of Mechanical Engineers)

Journal of Biomedical Materials Research

(16 issues, John Wiley & Sons)

Journal of Biomedical Materials Research

(Applied Biomaterials) (4 issues, John Wiley & Sons)

Journal of Bone and Joint Surgery (Am)

(12 issues, Journal of Bone and Joint Surgery, Inc.)

Journal of Bone and Joint Surgery (Br)

(6 issues, British Editorial Society of Bone and
Joint Surgery)

Journal of Engineering in Medicine Part H

(4 issues, Mechanical Engineering Publications Ltd.)

Journal of Materials Science Materials in Medicine

(4 issues, Chapman & Hall, CRC Press)

Journal of Orthopaedic Research

(6 issues, Journal of Bone and Joint Surgery, Inc.)

Medical & Biological Engineering & Computing

(6 issues, Peter Peregrinus Ltd.)

Medical Engineering & Physics

(10 issues, Elsevier Science)

SELECTED BOOKS ON BONE BIOMECHANICS

Lucas, G.L., Cooke, F.W., and Friis, E.A.,

A Primer of Biomechanics

. Springer-Verlag, New
York, 1999
Cowin, S.C.,

Bone Mechanics

. CRC Press, Boca Raton, 1st edition 1989, 2nd edition 2000
Bruce, M.R. and Burr, D.B.,

Skeletal Tissue Mechanics.

Springer-Verlag, New York, 1998,
392 pages
Evans, F.G.,

Mechanical Properties of Bone,

Charles C Thomas, Springeld, IL, 1973

0266/Ap1/frame Page 609 Wednesday, October 27, 1999 12:57 PM

0266/Ap1/frame Page 610 Wednesday, October 27, 1999 12:57 PM

611

Index

A

A. Ascenzi method torsion test, 283284
Absorptiometry
property evaluation methods, 109, 111
techniques for measuring density, 46
Accuracy in testing, 193, 194196
Acetabular component stability
factors affecting

in vitro

testing
bony ingrowth, 528
implantation technique, 529
modulus of elasticity differences, 528
sample geometry, 529
testing orientation, 529
torque forces involved, 529
true motion measurements, 528
variables in test samples, 528529

in vivo

stabilization with screws, 527
testing methods
adduction, 10 axial load, 534535
adduction, 12 axial load, 533
adduction, 30 axial load, 532533
axial load from anatomical positions, 535536
axial loading from repetitive cycles, 536
offset cup loading, 531532
pullout force, 537
strain distribution using FEA, 537
torsional stability, 530531
Acoustic impedance, 371372, 373375
Activity levels and mechanical properties of bone, 74
Adhesives for strain gauges, 314316
Age
bone fragilities from, 143
boneimplant strength and, 453454
bone mechanical properties and, 6668
bone structure and, 8, 9
Alcohol consumption and bone health, 76
Alpha-hydroxy acids in bioabsorbable polymer technology,
582
Alternate lamellae, 15, 4849, 273
American Society for Testing and Materials (ASTM)
testing designations, 177
Anelasticity, 333
Animal models in testing, 198199
Architectural level of bone, 137138
Areal moment of inertia for bone, 2930, 211
Arthritis affect on bone mechanical properties, 7576
Articular cartilage, 7
ASTM (American Society for Testing and Materials)
testing designations, 177
Autoclaving, 77
Axial cantilever test of boneimplant interface, 569,
573574

B

Backing material for strain gauges, 310
Bar wave velocity, 357
Bending tests of bone
cantilever beam
cancellous bone, 216
cortical bone, 216
single trabeculae, 296300
viscoelastic properties, 343344
whole bone, 214216
description of loads, 207209
four-point bending
cancellous bone, 214
cortical bone, 214
whole bone, 161, 213214
loading modes, 3031
single osteons methods, 281282
three-point bending
cancellous bone, 212
cortical bone, 212
whole bone, 209212
Berkovich diamond indenter, 259260, 263, 264
Biaxial servohydraulic testing systems, 224, 225
Bioabsorbable polymer technology fatigue testing
materials used, 582583
screw xation model, 584585
with synthetic bone interface, 586589
testing procedures, 585586

in vitro

strength-loss prole, 583584

in vivo

degradation process, 582
Biophosphonates and bone, 76, 451
BMC (bone mineral content)
absorptiometric techniques for measuring, 46, 111
dened, 45, 107
BMD (bone mineral density)
absorptiometric techniques for measuring, 46, 111
dened, 4445, 107
pQCT used to measure, 388390, 393395
Boiling/autoclaving, 77
Bone composition and structure
mechanical properties affected by, 6872
mineral substance, 56
organic matrix, 45
osteons
lamellae, 1315
microradiographic density, 13
remodeling, 13
structure, 1113
trabeculae, 1516
types of bone
cancellous, 10
compact, 9

0266/index/frame Page 611 Friday, October 22, 1999 11:27 AM

612

Mechanical Testing of Bone and the BoneImplant Interface

parallel-bered, 11
primary and secondary, 11
woven, 1011
whole bone description, 68
Boneimplant interface
acetabular component stability
factors affecting

in vitro

testing, 522, 527529
testing methods, 531537
fatigue testing of bioabsorbable screws
materials used, 582583
screw xation model, 584585
with synthetic bone interface, 586589
testing procedures, 585586

in vitro

strength-loss prole, 583584

in vivo

degradation process, 582
femoral component stability test
discussion, 547549
displacement equation derivation, 545546
overview, 541
results, 546547
rotation equation derivation, 544
test setup, 542544
nite-element analysis
axial cantilever test, 569
axial model results, 573574
discussion, 577578
knee ligament connector results, 575576
knee ligament connector test, 572
mandible model results, 574575
mandible model test, 569572
overview, 568569
vibrational techniques, 568
fracture toughness test
compact sandwich, 502507
crack loading modes, 500
interfacial debonding impact, 499500
mixed-mode sandwich, 507510
short-bar, critical strain energy release rate,
510512
short-rod, critical stress intensity factor, 512513
structural test use, 500
technique overview, 500502
intervertebral stability
background, 593594
scoliosis instrumentation testing, 600605
spine testing methods, 594600
pushout and pullout tests
acetabular

in vitro

testing, 537
features shared by tests, 464465
nondestructive test modes, 469470
overview, 463
pullout tests factors, 469, 471
pushout tests factors, 465, 469
validity of pushout tests,

see

Validity of single
pushout tests
screw pullout tests
experimental methodology, 559563
screw anatomy, 552554
screw design features, 552
screw mechanics, 554559
stability measurements

in vitro

acetabular component tests, 522
load cycle considerations, 523
loading simulation techniques, 522
motion components summary table, 518520
overview, 515516
realistic load scenarios importance, 521
test setup, 516517, 520521
strength of
implant and local factors, 440448
medications and local factors, 448451
patient factors, 453455
surgical and hospital factors, 452453

in vitro

factors, 451452
synthetic material testing, 163164
tensile test method
advantages and disadvantages of, 491493
detachment test results, 493
failure load measurement, 491
implantation, 489490
sample preparation, 490
strength test, 493494
variations of, 494496
Bone ingrowth,

see

Boneimplant interface
Bone mass and weightlessness, 7475
Bone mineral content (BMC)
absorptiometric techniques for measuring, 46, 111
dened, 45, 107
Bone mineral density (BMD)
absorptiometric techniques for measuring, 46, 111
dened, 4445, 107
pQCT used to measure, 388390, 393395
Bone models, synthetic
cortical bone substitutes
implant testing, 163164
mechanical properties evaluation, 161163
organic sample difculties, 159
preparation for testing, 161
test objectives, 159160
usefulness, 160
trabecular bone substitutes
difculties with, 165
foam preparation, 166168
foam testing, 169
foam use, 165166
Bone morphogenetic protein, 5
Breaking strength, 196197, 211
Breaking stress, 214, 215
Brinell hardness value, 234
Buckling, 37
Bulk modulus denition, 291, 359
Bulk wave propagation, 357, 359
Bulk wave velocity, 357

C

Calcitonin and bone metabolism, 75
Calcium intake and bone health, 76
Calcium phosphate ceramic (CPC)
as implant surface coatings, 443

0266/index/frame Page 612 Friday, October 22, 1999 11:27 AM

Index

613

for strain gauge bonding, 308, 316
Calcium phosphate hydroxyapatite, 5
Calibration of equipment, 227
Cancellous bone,

see also

Trabecular bone
cantilever test, 216
composition, 10, 7072
density, 107108, 109
four-point bending, 214
mechanical properties
anisotropy and heterogeneity, 5556
bone density, 5052
material properties, 5657
microstructure of trabecular bone, 5255
structure, 50, 51
nondestructive testing
challenges of standard testing, 151152
compression test guidelines, 152154
variation in test technique, 155
three-point bending, 212
ultrasonic measurement techniques, 359
Cantilever bending tests
cancellous bone, 216
cortical bone, 216
of single trabeculae
nite-element model development, 298299
xation, 296297
loading, 297298
potential errors, 299300
specimen selection, 296
tissue modulus determination, 299
viscoelastic properties, 343344
whole bone, 214216
CD (Cotrel Dubousset) instrumentation, 601
Cement lines and tissue-level testing, 139
Cholesky (IC) factorization, 425
CLSM (confocal laser scanning microscopy), 107
Coarse-ber bone, 1011
Collagen brils, 4
Compact bone, 7, 9
Compact sandwich fracture toughness test
data interpretation, 504506
limitations, 507
specimen preparation, 502503
test procedure, 503504
Compliance
dened, 35
local frame, 264265
machine, 125126
strain measurements and, 192193
Compression tests,

see also

Tensile and compression testing
advantages over tensile tests, 187
fracture of brittle materials, 187
grip stresses and, 195
methods for single osteons, 276278
nondestructive testing of cancellous bone
challenges of standard testing, 151152
test guidelines, 152154
variation in test technique, 155
setup considerations, 186
specimen-xture interface considerations, 126
tester for viscoelastic properties, 341
trabecular bone and, 186187
Compressive loading, 30
Computed tomography (CT), 110111, 410, 412;

see also


Peripheral quantitative computed tomography
(pQCT)
Computer modeling of trabecular bone
advantages of, 408
nite element stress analysis
analysis assumptions, 423424
boundary conditions, 424
FE algorithm, 424425
stress and strain computations, 428430
tissue modulus determination, 430
uniaxial compressive test, 424
validation by convergence, 425
validation by FE analysis, 425428
nite element volume meshing
approaches to, 419420
surface modeling, 420
voxel-based modeling, 419, 420423
image array acquisition
computed tomography, 410, 412
described, 409
magnetic resonance imaging, 412413
QSSG, 409410
image processing
data conversion, 414
data types, 413
image rendering, 416419
raster display, 413414
raster-to-vector conversion, 415416
volume of interest selection, 414415
morphologic analysis methods
MIL, 430433
SLD, 433434
SVD, 433
VO, 433
schematic of steps involved, 408409
Confocal laser scanning microscopy (CLSM), 107
ContMode in pQCT, 391
Controllers, 89
Cortical bone
cantilever test, 216
composition, 6870
density, 4346, 69, 107108, 109
described, 7
four-point bending, 214
level in bone hierarchy, 143144
mechanical properties
anisotropy and heterogeneity, 4748
bending, 43, 45, 46
density, 4346
porosity, 47
single osteons and microspecimens, 4850
strength and elastic modulus ranges, 43, 44
synthetic materials as models
implant testing, 163164
mechanical properties evaluation, 161163
organic sample difculties, 159
preparation for testing, 161

0266/index/frame Page 613 Friday, October 22, 1999 11:27 AM

614

Mechanical Testing of Bone and the BoneImplant Interface

test objectives, 159160
usefulness of, 160
three-point bending, 212
ultrasonic measurement techniques, 359
viscoelastic properties
analysis, 331333
apparatus, 342346
relaxation mechanism determination, 331,
336339
specimen preparation, 341
stress relaxation, 331333, 334336
testing methods, 339341
CortMode in pQCT, 392
Cotrel Dubousset (CD) instrumentation, 601
CPC (calcium phosphate ceramic)
as implant surface coatings, 443
for strain gauge bonding, 308, 316
Cracking, 37
Creep, 35, 229, 331, 333
Critical strain energy release rate, 500, 510512
Critical stress intensity factor, 37, 500, 512513
Crosshead position adjustment, 226227
CT (computed tomography), 110111, 410, 412;

see also


Peripheral quantitative computed tomography
(pQCT)
Cutting and machining of bone, 179
Cyclic loading tests for single osteons
electromechanical device, 284285
microwave micrometer component, 284
testing approach, 285286
test results, 286

D

Daro foam, 161, 165, 166168
Data management systems, 91
Debye-type relaxation, 331, 337, 339
Density of bone
absorptiometric techniques for measuring, 46, 111
cancellous, 5052
cortical, 4346, 69
dened, 4445, 107108
pQCT used to measure, 388390, 393395

in vitro

measuring of, 108109
Dental treatments and boneimplant interface
axial cantilever test, 569
axial model results, 573574
discussion, 577578
mandible model results, 574575
mandible model test, 569572
overview, 568569
vibrational techniques, 568
Diaphyses, 67
Displacement, 25
Drying of bone, 7879, 130, 258
Dual-energy absorptiometry, 111

E

Elasticity, 3335
Elastic limit, 184
Elastic modulus,

see also

Youngs modulus
from nanoindentation methods, 265266
ranges for cortical bone, 43, 44
of trabecular bone, table, 303
Elderly bone structure, 8
Electromagnetic/electrical stimulation and boneimplant
strength, 449450
Electric resistance strain gauging, 305
Electromechanical testing machines, 88, 226, 284285
Embalming/xation, 76
Endurance limit of bone, 37
Energy absorption of test material, 197
Epiphyses, 7
Epoxies for xation
described, 222
ber-reinforced
implant testing, 163164
preparation for testing, 161
torsional testing, 162163
usefulness of, 160
Equipment for bone tests,

see also

Evaluation methods for
mechanical properties
automated systems for torsional testing
electromechanical, 226
servohydraulic, 222226
Berkovich diamond indenter, 259260, 263, 264
biaxial servohydraulic systems, 224, 225
compression tester, 341
electromechanical testing machines, 88, 226, 284285
extensometers, 89, 191192, 276
at-ended cylindrical indenters, 235, 236238
hardness/indentation testing devices, 9395
Knoop indenter, 235236, 248249, 254
mechanical testing machines, 9295
microgrinding lathe, 274275
micromechanical testing machines, 93
osteopenetrometer, 235
potting materials, 9597
pQCT machine, 386387
screw driven test machines, 194
servohydraulic testing machines, 88, 222226
specimen preparation tools, 9798
spine testing instrumentation, 595596
tensile and compression testing, 190192
test grips, 192
torsion pendulum, 339341
Vickers indenter, 235, 248249, 254, 263
for viscoelastic testing, 342345
Estrogen deciency and boneimplant strength, 448
Evaluation methods for mechanical properties,

see also


Equipment for bone tests
absorptiometry, 111
computed tomography, 110111, 410, 412
computer modeling,

see

Computer modeling of
trabecular bone
density and material content
denitions and methods, 107109

0266/index/frame Page 614 Friday, October 22, 1999 11:27 AM

Index

615

in vitro

density values, 108110
macromeasurements of dimensions
direct, 104
microspecimens, 105
orientator, 105

in vivo

, 105
x-ray image based, 104105
magnetic resonance imaging, 111, 412413
micromeasurement of structure
confocal microscopy, 107
histomorphometry, 105106
ridge number density, 107
scanning electron microscope, 107
pQCT
bone density measurement, 388390
equipment options, 386
information applications, 395400
machine description, 386387
machine operation, 390392
variables and bone properties measured, 393395
scanning acoustic microscope
acoustic impedance, 371372
described, 371
high-frequency studies of osteonic bone, 377383
low-frequency studies, 373377
mechanical properties determination, 372373
spatial resolution of lens system, 373
uses of, 141
selection table, 112
ultrasound
data collection and calculation, 362365
description and uses, 111, 113
elastic properties determination, 357
for fracture healing, 451
measurement results, 366368
vs. mechanical testing methods, 358
sample preparation, 360361
testing methods, 361
Exercise and mechanical properties of bone, 74
Extensometers, 89, 191192, 276

F

Fatigue and failure characteristics, 3637
Fatigue testing
bioabsorbable screws
materials used, 582583
screw xation model, 584585
with synthetic bone interface, 586589
testing procedures, 585586

in vitro

strength-loss prole, 583584

in vivo

degradation process, 582
for tensile and compressive loads, 188189
for torsional loads, 228229
FEA,

see

Finite element analysis
Femoral component stability

in vitro

testing
discussion, 547549
methods
displacement equation derivation, 545546
rotation equation derivation, 544
test setup, 542544
overview, 541
results, 546547
torsional testing of synthetic, 162163
Fiber-reinforced epoxy bones
implant testing, 163164
preparation for testing, 161
torsional testing, 162163
usefulness of, 160
Finite element analysis (FEA)
acetabular deformation calculation using, 537
of boneimplant interface
axial cantilever test, 569
axial model results, 573574
discussion, 577578
knee ligament connector results, 575576
knee ligament connector test, 572
mandible model results, 574575
mandible model test, 569572
overview, 568569
vibrational techniques, 568
described, 3839
limitations to, 300
in micromechanical testing of bone, 298299
stress analysis of trabecular bone
algorithm, 424425
analysis assumptions, 423424
boundary conditions, 424
stress and strain computations, 428430
tissue modulus determination, 430
uniaxial compressive test, 424
validation by convergence, 425
validation by FE analysis, 425428
volume meshing
approaches to, 419420
surface modeling, 420
voxel-based modeling, 419, 420423
Fin whale (

Tympanic bulla

), 69
Fixation
affect on mechanical properties, 76
for cantilever bending tests, 296297
epoxies used for, 222
ber-reinforced epoxies, 160164
methods for torsional testing, 222
PMMA used for, 126, 222, 445446
Flat bones, 7
Flat-ended cylindrical indenters
advantages of, 235, 237238
methodology, 236237
Fluoride and bone, 76
Foil grid material for strain gauges, 308310
Force and displacement, 25
Force tranducers, 89
Four-point bending tests
cancellous bone, 214
cortical bone, 214
whole bone, 161, 213214
Fracture, 35, 36
Fracture toughness test, boneimplant interface
compact sandwich
data interpretation, 504506

0266/index/frame Page 615 Friday, October 22, 1999 11:27 AM

616

Mechanical Testing of Bone and the BoneImplant Interface

limitations, 507
specimen preparation, 502503
test procedure, 503504
crack loading modes, 500
interfacial debonding impact, 499500
mixed-mode sandwich
data interpretation, 509510
limitations, 510
specimen preparation, 508
test procedure, 508
short-bar, critical strain energy release rate
data interpretation, 512
limitations, 513
specimen preparation, 510511
test procedure, 511512
short-rod, critical stress intensity factor, 512513
structural test use, 500
technique overview, 500502
Frasca method torsion test, 283
Freezing effects on bone, 7778, 121122

G

Geometry of bone specimens, 179181
Grips and edge effects in testing, 301302
Grip stresses in compression tests, 195
Growth cartilage, 7
Growth factors/hormones
boneimplant strength and, 450451
bone metabolism and, 75

H

Hardness/indentation testing devices, 9395;

see also


Indenters
Harrington distraction/compression test, 601
Harvesting of bone and mechanical testing, 120
Haversian systems,

see

Osteons
Hierarchy of bone and mechanical properties
architectural level, 137138
architecture overview, 135136
computational methods
bi-level methods, 145146
limitations of, 146
material property estimates, 145
indentation tests,

see

Indentation testing
lamellar level, 140141
multilevel investigations examples
cortical bone, 143144
trabecular bone, 142143
tissue level, 138140
ultrastructual level, 141142
whole-bone level, 136137
Histomorphometry, 105106
Hole effects on strength, 79, 221
Hookes law, 34, 362
Hormones effect on mechanical properties, 75
Hydraulic power supply units (HPS), 88
Hydroxyapatite-coated implants
description and uses, 443444
failure mode analysis in pushout tests, 485
tensile testing of, 494496
Hysteresis loop, 33

I

IC (incomplete Cholesky) factorization, 425
IGF (insulin-like growth factors) and bone metabolism, 75
Implant-bone interface,

see

Boneimplant interface
Implant pushout and pullout tests,

see

Pushout and pullout
tests
Indentation testing,

see also

Penetration testing for
trabecular bone
macroindentation tests
Brinell hardness value, 234
cylindrical indenter use, 236238
at-ended cylindrical indenters, 235
osteopenetrometer, 235
Rockwell supercial hardness test, 234
microhardness tests, 235236
nanoindentation tests, 236
Indenters
Berkovich diamond, 259260, 263, 264
cylindrical, 235, 236238
hardness/indentation testing devices, 9395
Knoop, 235236, 248249, 254
Vickers, 235, 248249, 254, 263
Indomethacin and boneimplant strength, 448449
Inertia property, 29
Infant bone structure, 8
Insulin-like growth factors (IGF) and bone metabolism, 75
Interfacial strength,

see

Boneimplant interface
Intervertebral stability testing
background, 593594
scoliosis instrumentation testing
construct stiffness, 602
data analysis and reduction, 602
motion, outside construct, 602603
motion, within construct, 603605
overview, 600601
specimen preparation, 601
test modes, 601602
spine testing methods
anatomical characterization, 598
data acquisition, 599
data analysis and reduction, 599600
instrumentation design, 595596
loaddisplacement control, 594595
load measurement, 599
motion measurement devices, 594
relative micromotion measurement, 596598
Irradiation and boneimplant strength, 79, 449

K

Kashin-Beck, 76
Knee ligament connector test, 572, 575576

0266/index/frame Page 616 Friday, October 22, 1999 11:27 AM

Index

617

Knoop indenter, 235236, 248249, 254
KWW-type relaxation, 336337, 338
Kyniklometry, 105

L

Lactosorb copolymer, 583584, 586, 588
Lamellar bone,

see also

Osteons
in bone hierarchy, 140141
described, 11
mechanical properties of, 15
optical patterns, 14, 377383
organization of collagen brils, 1314
types of arrangements in osteons, 1415, 4849,
272273
Linear variable differential transducer (LVDT), 223
Load cells, 89
Load-displacement curve (P-D), 25, 124125, 177
Load units, 88
Local frame compliance, 264265
Long bones regions, 67
Longitudinal lamellae, 15, 4849, 273
Luque instrumentation, 601
LVDT (linear variable differential transducer), 223

M

Machine compliance and testing, 35, 125126;

see also


Compliance
Macroindentation tests
Brinell hardness value, 234
at-ended cylindrical indenters, 235
advantages of, 237238
methodology, 236237
osteopenetrometer, 235
Rockwell supercial hardness test, 234
Magnetic resonance imaging (MRI), 111, 412413
Mammalian animals bone material properties, 68, 69
Mandible model test, 569572, 574575
Marrow, 7
mechanical testing and, 194
removing, 108
Material failure observations using SEM
described, 349350
with in-scope loading system, 350351
low-vacuum, 350
material fracture mode observations, 352354
mechanical loading observations, 351352
research applications, 354355
uses of, 107
Mean intercept length (MIL) analysis method, 430434
Mechanical loosening, 241
Mechanical properties of bone
cancellous
anisotropy and heterogeneity, 5556
density, 5052
material properties, 5657
microstructure of trabecular bone, 5255
structure, 50, 51
continuum assumption, 5758
cortical
anisotropy and heterogeneity, 4748
bending properties, 43, 45, 46
density, 4346
human vs. bovine, 44
porosity, 47
single osteons and microspecimens, 4850
strength and elastic modulus ranges, 43, 44
hierarchical levels of bone, 42
term denitions, 66

in vitro

factors affecting
boiling/autoclaving, 77
drying/freeze-drying, 7879
embalming/xation, 76
hole effects on strength, 79
sterilization, 79
storage, 7778

in vivo

factors affecting
age, 6668
arthritis, 7576
composition, 6872
function, 7275
hormones, 75
other systemic factors, 76
sex, 68
species, 68, 69
Mechanical testing concepts
compliance
dened, 35
load frame, 264265
machine, 125126
strain measurements and, 192193
elasticity, 3335
fatigue characteristics, 3637
force and displacement, 25
fracture energy, 35, 36
hierarchical approach,

see

Hierarchy of bone and
mechanical properties
loading modes
bending, 3031
compressive, 30
multiaxial, 32
static or cyclic, 3233
torsional, 3132
material dimensions concepts, 2930
modeling and simulation, 3839
statics and dynamics, 2425
strength, 33
stress and strain
dened, 2628
measuring on bone, 28
stressstrain curve, 28
symbols used, 24
toughness, 35
viscoelasticity, 35
Mechanical testing considerations
factors affecting results
machine compliance, 35, 125126
specimen end effects, 129
specimen-xture interface, 126127

0266/index/frame Page 617 Friday, October 22, 1999 11:27 AM

618

Mechanical Testing of Bone and the BoneImplant Interface

specimen size and geometry, 127129
strain rate, 129130
testing conditions, 130
harvesting of bone, 120
preservation of bone, 120122
sources of bone, 119120
specimen preparation, 122123
testing procedures, 123125
Medications affecting boneimplant strength
biophosphates, 451
electromagnetic/electrical stimulation, 449450
growth factors/hormones, 450451
irradiation, 449
NSAIDS, 448449
systemic hormones, 448
ultrasound, 451
vibrations, 451
Medulla, 7

Mesoplodon densirostris

(whale), 69
Metaphyses, 7
Microgrinding lathe, 274275
Microhardness tests
anisotropy measurements, 253254
factors affecting microhardness of bone
anatomical variables, 251252
specimen preparation, 250251
testing variables, 252253
guidelines for, 253
indentation testing, 235236
indenters, 248250
overview, 141, 248
usefulness of, 254
Micromechanical testing for single osteons
bending test, 281282
compression test, 276278
cyclic loading, 284287
elements of, 272273
pin test, 280281
shearing test, 278280
summary, 287288
tension test, 274276
torsion test, 282284
Micromechanical testing machines, 93
Micromechanical testing of single trabeculae
cantilever beam
nite-element model development, 298299
xation, 296297
loading, 297298
potential errors, 299300
specimen selection, 296
tissue modulus determination, 299
elastic modulus table, 303
results, 302304
tensile test, 294295
terminology, 291
testing considerations
apparatus checks, 302
grips and edge effects, 301302
specimen geometry, 300301
tissue hydration, 302
three-point bending, 292293
Microradiography, 13
Microstructural level of bone tests, 235236
MIL (mean intercept length) method, 430433
Mineral substance of bone, 56
of bone tissue, 6970
and mechanical strength, 45
Mixed-mode sandwich fracture toughness test
data interpretation, 509510
limitations, 510
specimen preparation, 508
test procedure, 508
Modied Brinell Hardness, 234
Modulus of elasticity
dened, 197
of trabecular bone, 303
whole bone bending tests, 211, 214
Moments, 2930, 211
MRI (magnetic resonance imaging), 111, 412413
Multiaxial loading, 32
Multiaxial machines, 92

N

Nakamuras tensile test method
advantages and disadvantages of, 491493
detachment test results, 493
failure load measurement, 491
implantation, 489490
sample preparation, 490
tensile strength test, 493494
variations of, 494496
Nanoindentation tests, 140141
diamond area function, 264265
indentation procedures, 259260
limitations due to anisotropy, 265
load-displacement data analysis, 260264
load frame compliance, 264265
measurement results, 266267
overview, 236, 257258
sample preparation, 258259
Noncollagenous components, 45
Nondestructive testing of cancellous bone
challenges of standard testing, 151152
compression test guidelines, 152154
friction effects, 195
variation in test technique, 155
Nonsteroidal anti-inammatory drugs (NSAIDS), 448449

O

OA (osteoarthritis), 7576
Offset cup loading and acetabular component stability,
531532
OI (osteogenesis imperfecta), 144
Organic matrix of bone, 45
Orientator techniques, 105
Osseointegration, 440, 444, 567
Osseous trabeculae, 7
Osteoarthritis (OA), 7576

0266/index/frame Page 618 Friday, October 22, 1999 11:27 AM

Index

619

Osteogenesis imperfecta (OI), 144
Osteons,

see also

Lamellar bone
composition and structure
lamellae, 1315
microradiographic density, 13
remodeling, 13
structure, 1113
harvesting of, 123
lamellae gray levels observed with SAM, 377383
mechanical properties of types, 4850
micromechanical testing for single
bending test, 281282
compression test, 276278
cyclic loading, 284287
elements of, 272273
pin test, 280281
shearing test, 278280
summary, 287288
tension test, 274276
torsion test, 282284
nanoindentation tests, 236, 266267
tensile and compression testing for single
anisotrophic considerations, 181182
axes of symmetry determination, 182
bending corrections, 184
compression testing, 186187
fatigue testing, 188189
shape and gauge region, 183
tensile testing, 183186
whole bone specimens, 189190
tension test for single
elongation measurements, 276
loading apparatus, 275276
sample preparation, 274275
Osteopenetrometer for in vivo penetration testing
macroindentation testing, 235
measurement distances, 244
measurements obtained, 242243
needle design, 243244
needle dimension and errors, 245
tool requirements and design, 242
Osteoporosis
bone fragilities from, 143, 199
factor in boneimplant strength, 455
screw pullout strength and, 327
Oxide ceramics as implant surface coatings, 443

P

Parallel-bered bone, 11
Parallelism index, 186
Parathyroid hormone and bone metabolism, 75
PeelMode in pQCT, 391392
Penetration testing for trabecular bone, see also Indentation
testing
laboratory testing, 245
usefulness of, 245246
in vivo
measurement distances, 244
measurements obtained, 242243
needle design, 243244
needle dimension and errors, 245
tool requirements and design, 242
Percent elongation, 197
Peridontal health, 568
Periosteum, 7
Peripheral quantitative computed tomography (pQCT)
bone density measurements, 388390
equipment options, 386
information applications
architectural efciency analysis, 397400
mechanostat status analysis, 397
muscle/bone interrelationships, 400
pathogenesis analysis, 395397
threshold-dened ROI analysis, 400
machine description, 386387
machine operation
ContMode, 391
CortMode, 392
PeelMode, 391392
settings, 390391
specimen preparation, 390
variables and bone properties measured
apparent density, 393
architectural quality, 393394
mass, 393
whole-bone quality, 394395
Pin test methods for single osteons, 280281
PMMA (polymethylmethacrylate) for xation, 126, 222,
445446
Poissons ratio
for bone, 263
dened, 29
for uniaxial compression tests, 423
Polar moment of inertia, 30
Polymethylmethacrylate (PMMA) for xation, 126, 222,
445446
Porosity of cortical bone, 47, 6869
Potting of specimens
for cantilever bending tests, 214215
materials, 9597
for tensile and compressive tests, 181
whole-bone level, 189
pQCT, see Peripheral quantitative computed tomography
Precision in testing, 193194
Preservation methods for bone, 120122, 178179
Press-t denition, 442
Primary and secondary bone, 11
Proportional-elastic limit, 33, 184
Pushout and pullout tests, see also Screw pullout tests
for acetabular in vitro testing, 537
features shared by, 464465
nondestructive test modes, 469470
overview, 463
pullout tests factors, 469, 471
pushout tests factors, 465, 469
validity of pushout tests
factors affecting testing, 482484
failure mode analysis, 485
inconsistency in experimental results, 477478
practical aspects of testing, 479482

0266/index/frame Page 619 Friday, October 22, 1999 11:27 AM

620

Mechanical Testing of Bone and the BoneImplant Interface

Q

Quantitative computed tomography (QCT), 385386; see
also Peripheral quantitative computed
tomography (pQCT)
Quantitative serial sectioning and serial grinding (QSSG),
409410

R

Radiostereometric analysis (RSA), 515, 521
Radio telemetry systems for strain measurement, 317
Ramp test, 227228
RA (rheumatoid arthritis)
boneimplant strength and, 454
bone mechanical properties and, 7576
Raster display, 413414
Raster-to-vector conversion, 415416
Relaxation, see Stress relaxation
Remodeling of osteons, 13
Representative volume element (RVE), 572
Rheumatoid arthritis (RA)
boneimplant strength and, 454
bone mechanical properties and, 7576
Ridge number density (RND), 107
Rockwell supercial hardness test, 234
Rosette gauges, 311
Rotational variable differential transducer (RVDT), 224
RSA (radiostereometric analysis), 515, 521
RVE (representative volume element), 572

S

Sagittal split mandibular osteotomies
materials used, 583
screw stress, 584585
SAM, see Scanning acoustic microscope
Sample-holding devices, 8991
Scanning acoustic microscope (SAM)
acoustic impedance, 371372
described, 371
high-frequency studies of osteonic bone, 377383
low-frequency studies
acoustic impedance method, 373375
anisotropy measurements, 375377
mechanical properties determination, 372373
spatial resolution of lens system, 373
uses of, 141
Scanning electron microscope (SEM)
described, 349350
with in-scope loading system, 350351
low-vacuum, 350
material fracture mode observations, 352354
mechanical loading observations, 351352
research applications, 354355
uses of, 107
Scoliosis instrumentation testing
construct stiffness, 602
data analysis and reduction, 602
motion, outside construct, 602603
motion, within construct, 603605
overview, 600601
specimen preparation, 601
test modes, 601602
Screw driven test machines, 194
Screw pullout tests, see also Pushout and pullout tests
experimental methodology
data analysis, 563
gripping xtures, 560
insertion technique, 562563
specimen grouping, 563
specimen preparation, 560562
testing apparatus, 559
testing protocol, 563
screw anatomy, 552554
screw design features, 552
screw mechanics
bending strength, 554
bone mineral content, 558559
holding strength, described, 554555
holding strength, factors affecting, 555557
implant material and, 559
material properties, 554
thread angulation, 558
torque insertion, 557558
strength tests, 55
density effects, 327
orientation to bone structure, 321323
preservation methods effects, 326327
vertebral cortices and insertion angle effects,
323326
SEM, see Scanning electron microscope
Servohydraulic testing machines
control modes, 224225
described, 88, 222224
limits and response rate, 225226
in vitro testing methodology
calibration of equipment, 227
creep test, 229
crosshead position adjustment, 226227
fatigue test, 228229
ramp test, 227228
Sex and bone mechanical properties, 68
Shearing test for single osteons, 278280
Shear stresses, 2627
Short-bar fracture toughness test
data interpretation, 512
limitations, 513
specimen preparation, 510511
test procedure, 511512
Short bones, 7
Short-rod fracture toughness test, 512513
Single-axis machines, 92
Single-photon absorptiometry, 111
SLD (star length distribution) method, 433434
SLE (systemic lupus erythematosis) and boneimplant
strength, 454455
Software programs for testing, 91
Specimen end effects and testing, 129
Specimen preparation

0266/index/frame Page 620 Friday, October 22, 1999 11:27 AM

Index

621

compact sandwich fracture toughness test, 502503
mechanical testing, 122123
microhardness testing, 250251
mixed-mode sandwich fracture toughness test, 508
nanoindentation tests, 258259
potting materials, 9597
pQCT, 390
SAM tests, 375, 376
scoliosis instrumentation testing, 601
screw pullout tests
organic bone, 561562
synthetic bone, 560561
single pushout tests, 479
tensile and compression testing
cutting and machining, 179
geometry, 179181
potting, 181
preservation, 178179
tensile tests of boneimplant interface, 490
tools for, 9798
ultrasound tests, 360361
Specimen size/geometry and testing, 127129
Spine testing, see Intervertebral stability testing
Star length distribution (SLD) method, 433434
Star volume distribution (SVD) method, 433
Static or cyclic loading, 3233
Sterilization, 79
Steroid administration and bone metabolism, 75
Stiffness of materials, 3334, 125126
Storage methods for bone, 7778, 178
Strain and stress
and compliance, 192193
dened, 2628
errors in measurement, 195196
measuring on bone, 28
stressstrain curve, 28, 184186
Strain gauge measurements
attachment of gauges
CPC-coated, 308, 316
for in vitro studies, 314
for in vivo studies, 314316
background, 305
data analysis, 317
factors in, 308
gauge selection criteria
backing material, 310
dimensions, 310311
foil grid material, 308310
history of use in biomechanics, 306308
lead wires, 312313
measurement collection, 316317
pre-attachment preparation, 313314
transcutaneous connectors, 312313
Strain rate and testing, 129130
Strain rate power function, 191
Strength denition, 33
Strength of boneimplant interface
implant and local factors
cement quality, 445446
design and stress shielding, 447448
geometry, 444445
material properties, 445
particulate debris, 446447
surface coatings, 443444
surface congurations, 440443
medications and local factors
biophosphates, 451
electromagnetic/electrical stimulation, 449450
growth factors/hormones, 450451
irradiation, 449
NSAIDS, 448449
systemic hormones, 448
ultrasound, 451
vibrations, 451
patient factors
age, 453454
inammation, 454455
lifestyle, 454
osteopenia/osteoporosis, 455
surgical and hospital factors
complications, 452453
postoperative care, 453
surgical skill, 452
in vitro factors, 451452
Stress and strain
and compliance, 192193
dened, 2628
errors in measurement, 195196
measuring on bone, 28
stress-strain curve, 28, 184186
Stress concentration effects of holes, 79, 221
Stress fracture, 3637, 75
Stress relaxation, 35, 331333
for cortical bone, 334336
Debye-type, 331, 337, 339
KWW-type, 336337, 338
Stress-strain curve, 28, 184186
Structural density, 108
SVD (star volume distribution) method, 433
Synthetic materials as bone models
cortical bone substitutes
implant testing, 163164
mechanical properties evaluation, 161163
organic sample difculties, 159
preparation for testing, 161
test objectives, 159160
usefulness of, 160
trabecular bone substitutes
difculties with, 165
foam preparation, 166168
foam testing, 169
foam use, 165166
Systemic hormones and boneimplant strength, 448
Systemic lupus erythematosis (SLE) and boneimplant
strength, 454455

T

Temperature and bone testing, 7778, 121122, 130
Tensile and compression testing
animal models, 198199

0266/index/frame Page 621 Friday, October 22, 1999 11:27 AM

622

Mechanical Testing of Bone and the BoneImplant Interface

considerations for
mechanical properties of specimens, 177178
standard procedures use, 176177
structural levels, 176
equipment
elements of, 190191
extensometers, 89, 191192, 276
load control, 191
strain rate choice, 191
test grips, 192
guidelines for, 199200
measurement procedures
accuracy, 193, 194196
compliance, 192193
mechanical and structural properties
determination, 196197
precision, 193194
toe region compensation, 197
validation, 193
pushout test model, 481482
single osteon methods
anisotrophic considerations, 181182
axes of symmetry determination, 182
bending corrections, 184
compression testing, 186187
fatigue testing, 188189
shape and gauge region, 183
tensile testing, 183186
whole bone specimens, 189190
specimen preparation
cutting and machining, 179
geometry, 179181
potting, 181
preservation, 178179
Tensile loading, 30
Tensile strength, 197
Tensile tests
boneimplant interface
advantages and disadvantages of, 491493
detachment test results, 493
failure load measurement, 491
implantation, 489490
sample preparation, 490
tensile strength test, 493494
variations of, 494496
dimension standards, 183
fracture variables, 186
micromechanical testing devices, 93
of single trabeculae, 294295
specimenxture interface considerations, 126127
stress-strain curve, 184186
Tension test for single osteons
elongation measurements, 276
loading apparatus, 275276
sample preparation, 274275
Test grips, machine, 192
Testing facilities and instruments, see Equipment for bone
tests
3dMesh, 420
Three-point bending tests
cancellous bone, 212
cortical bone, 212
of single trabeculae, 292293
viscoelastic properties, 344345
whole bone
data collection and calculation, 210212
sample preparation, 209
testing methods, 209210
Thyroid hormone and bone metabolism, 75
Tibia torsional testing, 163
Tissue level in bone hierarchy, 138140
Tissue modulus denition, 291
Toeing-in of screws, 557
Toe region compensation, 198
Torsional loading, 3132
Torsional testing of bone
acetabular stability
factors affecting in vitro testing, 527529
testing methods, 530537
automated systems
electromechanical, 226
servohydraulic, 222226
system selection, 226
background, 219220
of ber-reinforced epoxies, 162163
xation methods, 222
methods for single osteons, 282284
structural defects effects, 221222
theory of torsion for cylinders, 220221
viscoelastic properties, 342343
in vitro testing methodology
calibration of equipment, 227
creep test, 229
crosshead position adjustment, 226227
fatigue test, 228229
ramp test, 227228
Torsion pendulum, 339341
Toughness, 35
Trabecular bone, see also Cancellous bone
age and gender comparisons experiments, 142143
cantilever bending tests of single
nite-element model development, 298299
xation, 296297
loading, 297298
potential errors, 299300
specimen selection, 296
tissue modulus determination, 299
compression tests and, 186187
computer modeling of mechanical properties
nite element stress analysis, 423430
nite element volume meshing, 419423
image array acquisition, 409413
image processing, 413419
morphologic analysis, 430434
schematic of steps involved, 408409
histomorphometry used to measure, 106
material properties, 5657
micromechanical testing of single
cantilever beam, 297300
elastic modulus table, 303
results, 302304
tensile test, 294295

0266/index/frame Page 622 Friday, October 22, 1999 11:27 AM

Index

623

terminology, 291
testing considerations, 300302
three-point bending, 292293
microstructure of, 5255
nondestructive testing friction effects, 195
penetration testing in vivo
measurement distances, 244
measurements obtained, 242243
needle design, 243244
needle dimension and errors, 245
tool requirements and design, 242
stiffness denition, 177
stress analysis
analysis assumptions, 423424
boundary conditions, 424
FE algorithm, 424425
stress and strain computations, 428430
tissue modulus determination, 430
uniaxial compressive test, 424
validation by convergence, 425
validation by FE analysis, 425428
structure and function, 1516
synthetic materials as models
difculties with, 165
foam preparation, 166168
foam testing, 169
foam use, 165166
ultrasound used to measure elastic properties, 359
Transcutaneous connectors for strain gauge measurements,
312313
Transverse lamellae, 15, 4849, 273
Tympanic bulla (n whale), 69

U

Ultimate strength, 196197
Ultrasonic methods for property evaluation
data collection and calculation
cancellous bone, 365
cortical bone, 362
elastic coefcients matrix, 362365
description and uses, 111, 113
elastic properties determination, 357, 359361
for fracture healing, 451
measurement results, 366368
vs. mechanical testing methods, 358
sample preparation, 360361
testing methods, 361
Ultrastructual level in bone hierarchy, 141142
Uniaxial compressive test, 424

V

Validity of single pushout tests
factors affecting testing
biological parameters, 483484
experimental setup, 482
implant properties, 482483
failure mode analysis, 485
inconsistency in experimental results, 477478
practical aspects of testing
implant surface area calculation, 480481
specimen preparation, 479
tensile test model, 481482
recommendations, 485486
Vibrational techniques for implant evaluation, 568
Vickers indenter, 235, 248249, 254, 263
Viscoelasticity, 35
Viscoelastic properties of cortical bone
analysis, 331333
apparatus for viscoelastic testing
cantilever bending, 343344
data acquisition, 345
fast measurements, 345346
three-point bending, 344345
torsion tester, 342343
creep, 331, 333
described, 329331
relaxation mechanism determination
Debye-type, 331, 337, 339
KWW-type, 336337, 338
specimen preparation, 341
stress relaxation, 331333, 334336
testing methods
compression tester, 341
torsion pendulum, 339341
Volume orientation (VO) method, 433
Volumetric marching cubes (VOMAC), 419
Von Mises stress, 425
Voxel-based FE models
advantages of, 419
mesh modeling, 420423

W

Wall resistance test for osteons, 280281
Wave propagation modes, 357, 359
Weightlessness and bone mass, 7475
Wet or dry samples and testing, 130
Whale (Mesoplodon densirostris), 69
Whole bone
cantilever test, 214216
described, 68
four-point bending, 161, 213214
level in bone hierarchy, 136137
macroindentation tests, 233235
modulus of elasticity, 211, 214
quality measured with pQCT, 394395
tensile and compression testing, 189190
three-point bending
data collection and calculation, 210212
sample preparation, 209
testing methods, 209210
Wolffs Law, 74
Woods metal for xation, 222
Woven bone, 1011

0266/index/frame Page 623 Friday, October 22, 1999 11:27 AM

624

Mechanical Testing of Bone and the BoneImplant Interface

Y

Yield point, 185, 196
Yield strength, 33
Youngs modulus, see also Elastic modulus
dened, 34, 177
interface stress distributions and, 483
storage effects on, 178
for uniaxial compression tests, 423

Z

Zinc deciency and bone health, 76

0266/index/frame Page 624 Friday, October 22, 1999 11:27 AM