The Artificial Knee An Ongoing Evolution

The Artificial Knee
An Ongoing Evolution
Peter S. Walker
123
The Artificial Knee
Center of
Femoral
femoral head
Quadriceps Lateral condyles
collateral
ligament
Greater
trochanter Anterior
Patella cruciate
Femur ligament
Patella
ligament
Medical
Anatomic
meniscus
axis of femur
Mechanical Lateral
axis of leg meniscus
Medical
collateral
ligament
Tibial metaphysis Mechanical
(cancellous bone) Medial View Lateral View
axis of femur
L M P A
Posterior
Posterior
cruciate
femoral
Anatomic and ligament
condyles
mechanical axis Mechanical
of tibia axis of leg
Tibial disphysis Posterior View

(cortical bone)
Tibia
Femur
Center point
between Patella
malleoli component
Femoral Femoral
component component Patella
(metal) flange
Patella L M P
component Tibial A
(plastic) insert
Tibial insert Tibial tray Acrylic

(plastic) cement
layers
Tibial post
Tibial tray with fins
(metal)
Tibia
Contact Tibial post Fibula
areas
Tibial component AP View Sagittal View
Top left, the overall alignment of the femur and tibia in frontal and sagittal views; top right, the
anatomical structures of the knee, excluding the muscles; bottom left, a typical artificial knee,
with the three different components, femoral, tibial, and patella; bottom right, AP and ML radio-
graphs of an artificial knee, showing the cement mantle (orange) in the sagittal view. (Figure
compiled by Jamie Liu)
Peter S. Walker
The Artificial Knee

An Ongoing Evolution
Peter S. Walker
Department of Orthopedic Surgery
New York University
New York, NY
USA
ISBN 978-3-030-38170-7 ISBN 978-3-030-38171-4 (eBook)

https://doi.org/10.1007/978-3-030-38171-4
© Springer Nature Switzerland AG 2020

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recita-
tion, broadcasting, reproduction on microfilms or in any other physical way, and transmission or infor-
mation storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publica-
tion does not imply, even in the absence of a specific statement, that such names are exempt from the
relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the
editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Contents
1 Introduction �� 1

2 Early Biomechanical Knowledge �� 13
3 The Design Process�� 33
4 Laxity and Stability �� 51
5 Fixation Using Acrylic Cement �� 69
6 Uncemented Fixation �� 85
7 Femoral Component and Patella �� 103
8 Tibial Component �� 121
9 Condylar Replacements �� 141
10 Posterior Stabilized �� 167
11 Guided Motion Knees �� 181
12 Instrumentation and Technique �� 205
13 Testing and Tribology �� 233
14 Evaluations of Designs and Techniques �� 265
15 Discussion and Future Possibilities �� 295
Glossary�� 323
Index�� 335
v
About the Author
Peter S. Walker was educated in England at the Royal

Grammar School, Newcastle, England, and then at the
University of Cambridge. After a 3-year engineering
traineeship in industry, he completed his PhD at the
University of Leeds (1966–1969) under Professors
Duncan Dowson and Verna Wright. Since then, he has
been an active participant in the field of biomechanics
of joints and the design of joint replacements, with a
major emphasis on the knee. He worked for 18 years in
the USA, 13 years in England, and then a further 20
years in the USA up to the present. He has collaborated
with surgeons and bioengineers at renowned institu-
tions, including the Hospital for Special Surgery;
Howmedica (Stryker); Brigham and Women’s Hospital,
Harvard Medical School; Royal National Orthopaedic
Hospital, University College London; and, currently,
New York University Langone Orthopedic Hospital and
NYU Tandon School of Engineering. He has coau-
thored over 300 articles with 12,000 citations. His
basic research has included innovative ideas for total
knee design, while many of the test machines and meth-
odologies have been emulated by other labs. He has
been a co-designer of several successful total knee sys-
tems; the Total Condylar, Graduated Knee System,
Kinematic & Kinemax, Smiles Bone Tumor System,
NexGen, and most recently the Persona, a leading knee
system today. He has received several awards from vari-
ous societies, the most notable being the Lifetime
Achievement Awards from ISTA in 2004 and from The
Knee Society in 2018.
vii
Introduction
Key Questions, Overview, and Layout
1
While playing tennis recently, I was struck by how well my opponent moved around
the court. At the end of the game, he remarked: “not bad for a couple of mechanical
knees.” He later explained that he had to follow a strict rehabilitation program to
reach the activity level he had before his arthritis but that it was definitely worth the
effort. After an artificial knee, most patients return to a normal everyday lifestyle
free of pain and discomfort, while participation in active sports is not uncommon.
Since the first artificial knees were implanted in the late 1960s, over 50 years ago,
the quality of the outcomes has steadily progressed until the present day. This has
been due to the concerted efforts of the surgeons and engineers who invented the
artificial knees, to the companies who produced them, and to the countless patients
who tried them, especially in the early days.
Historically, devices for relieving pain and restoring function to the arthritic knee
were implanted as far back as 1891 although these failed due to inadequate materi-
als and non-sterile surgical technique (Brand et al. 2011). In the 1940s and 1950s,
metallic resurfacing components for the distal femur and more commonly the proxi-
mal tibia were used with reasonable success. These devices were mainly used in the
less severe arthritic cases with modest functional demands and had the effect of
realigning the lower limb and reducing pain. At the other extreme, metallic hinges
with long intramedullary stems for fixation were used for the more severe cases, but
reservations about the need for resection of the major ligaments, the large amount
of bone removal, and the invasion of the intramedullary canals, limited their wide-
spread adoption. It was not until the late 1960s that artificial knees similar to those
in use today were invented. The author was working on his PhD at Leeds University
in England from 1966 to 1969 on the subject of human joint lubrication, at the same
time as the start of the artificial knee, and has worked in the latter field until the
present time.
A major aim of this book is to show how artificial knees reached the stage they
are today and what makes them work well in the majority of cases. This requires
explaining the basic principles of the designs, how the designs originated and
evolved, and the interactions between the many individuals who participated. The
© Springer Nature Switzerland AG 2020 1

P. S. Walker, The Artificial Knee, https://doi.org/10.1007/978-3-030-38171-4_1
2 1 Introduction
starting point occurred in the mid-1960s when the artificial hip had become success-
ful. It was realized that the success was due to the combination of bearing materials
with a high wear resistance and a biocompatible cement which would fix the com-
ponents rigidly to the bones. But there was quite a difference between the structure
and function of the hip and the knee. The hip was relatively simple, being a ball-in-
socket with a fixed center of rotation. The knee on the other hand was far more
complicated, relying on many structures for its stability, and capable of more move-
ments than just a simple hinge. It was not surprising that the early designers reached
several different solutions for an artificial knee. The first designs ranged from uni-
compartmental knees where small components resurfaced the femoral and bearing
surfaces but all of the ligaments were preserved, to large components which resur-
faced the entire bearing surfaces and where the cruciate ligaments were resected.
Other designs were intermediate to these two extremes.
A striking observation is that the large majority of the artificial knees used today
are of two design types which originated in the 1970s, just about 10 years after the
very first designs. These two types are the cruciate retaining (CR), which retains
only the posterior cruciate ligament, and the posterior stabilized (PS) where a cam
provides stability but the cruciate ligaments are resected. The key question addressed
in this book is: Why do these two types dominate the market, when it might be
thought that, particularly since 1980, superior designs would have emerged and
Condylar replacement knees
ACL+PCL PCL Total Posterior Guided

retaining retaining condylar stabilized motion
BCR CR TCP PS GM
Cruciates Posterior Condylar resurfacing Intercondylar Mechanism for
retained cruciate Cruciates resected stability (cam-post) guiding motion
retained
Unicondylar Ultra-
UNI congruent Rotating
UC platform
High congruence RP
Fixed Meniscal tibial surface
bearing bearing Pivots about
Medial pivot tibial center
MP
Patella Biomimetic Constrained
medially, lax
BM
laterally
Anatomic
Dome Anatomic shapes and
kinematics
Fig. 1.1 The different types of condylar replacement artificial knees. ACL anterior cruciate liga-
ment, PCL posterior cruciate ligament, BCR bi-cruciate replacement, CR posterior cruciate retain-
ing, TCP Total Condylar prosthesis, PS posterior stabilized, GM guided motion
1 Introduction 3
replaced the original ones? Could it be that the artificial knee had reached its limit
such that the performance could only be improved incrementally? That would be
surprising considering that since 1980 there has been a dramatic expansion in the
knowledge of the biomechanics of the knee, major advances in materials and surgi-
cal techniques, and a vast number of follow-up studies on the functional outcomes,
kinematics, wear, durability, and surgical technique. It is certainly true that this
body of work has led to numerous improvements. Possibly the most important
development has been polyethylene formulations which have substantially reduced
the wear. Other developments have included refinements in component shapes, mul-
tiple sizes to provide a more accurate fit for each patient, and bearing surfaces with
improved stability. Nevertheless the fact remains that the fundamental configura-
tions and mechanics of the designs used most frequently today have not changed
significantly since 1980.
While this book concentrates mainly on artificial knee design, the instrumenta-
tion and technique is also discussed, because this has a major effect on the out-
comes. A similar observation as for artificial knees can be made about the
instrumentation and technique; the majority of surgeries are performed with
mechanical instruments which are based on the same principles of alignment as in
1980 and where the construction of the instruments is similar although allowing for
greater adjustability. Computer-assisted surgery, which has been available for over
a decade, is still used in only a small percentage of cases. That may be changing
UNI Patella
CR
Fixed bearing ACL+PCL BCR
Posterior cruciate
Anterior+posterior
retained
cruciate retained
PS
Rotating Posterior
platform stabilized
UNI
Meniscal
bearing
Fig. 1.2 The main types of artificial knee which had been developed by 1980. All designs, with
the exception of the unis, have plastic patella components (arrow on the right). The medial pivot
and the ultracongruent knees (not shown here) were developed after 1980. From 1980 to the pres-
ent time, the majority of artificial knees used have been the CR and the PS
4 1 Introduction
now that robotic equipment has been introduced, but adoption is still slow despite
the advantages of accuracy of the bone cuts and soft tissue balancing.
Today, artificial knees are used in large numbers for the treatment of arthritis.
According to a recent study, 966,000 cases were performed in the USA in 2017
(Idata Research 2017). While the large majority of the cases were primary artificial
knees, revisions and unicompartmental knees were included. The numbers continue
to increase year after year due to the increasing number of individuals with arthritis,
particularly the “baby boomers” born between 1946 and 1964 who make up 29
percent of the population. There is also a great deal of emphasis today on the ben-
efits of an active lifestyle, which increases the demand for artificial knees. The more
widespread availability of the procedure is also a factor. This increased usage would
not occur if the procedure was not successful. Based on joint registries from several
countries, the failure rate of artificial knees is now very low, only about a half a
percent per year on average. The failure rate is likely to become even lower due to
the steady improvement in design, materials, surgical technique, and rehabilitation,
as well as the education and specialization of orthopedic surgeons. In parallel with
the improvements in durability, the functional results have improved also. Based on
an extensive literature of follow-up studies, the majority of those with artificial
knees are pain-free or experience only mild pain in certain activities and can carry
out activities of everyday living. Many individuals can perform more strenuous
activities including sports, although this is understandably related to overall condi-
tion, particularly muscle strength. A large percentage of individuals report that most
of the time they forget that they have an artificial knee, but not to the same extent as
for artificial hips.
This is not to say that the artificial knee procedure is without its problems. The
most serious is infection, fortunately occurring in only a fraction of a percent of
cases and which can usually be treated without removing the components. Instability
is experienced in some cases, usually soon after the procedure, and if serious enough
may require a component to be changed. Component loosening or fracture is a late
failure mode, requiring revision surgery. Wearing out of the bearing surfaces, a
common mode of failure in the early days of knee replacement, is now likely to be
rare due to the much higher wear resistance of the new polyethylene bearing materi-
als. The most common functional limitation of an artificial knee is that the range of
flexion is not as high as in a normal knee, but is sufficient for most activities.
Nonetheless getting in and out of an automobile and rising from low chairs may not
be easy. Another limitation is that, while many patients can return home 1–2 days
after the procedure and can walk shortly after, recovery can be uncomfortable and
can take several weeks or even months.
Hence it can be claimed that about 50 years after they were first invented,
artificial knees function well from a mechanical point of view. One of the early
design principles was to minimize the stresses between interfacing surfaces
which are loaded together, whether between the femoral and tibial components
or between the components and the bone surfaces. This was emphasized by Alan
Swanson, a Professor of Mechanical Engineering at Imperial College London,
who worked with Michael Freeman, an orthopedic surgeon at the London
1 Introduction 5
Hospital, on one of the first artificial knee designs in the late 1960s. A small
pebble in a shoe causes pain due to the high local stresses. High heels can make
an indentation in a soft surface. Everyone has experienced the discomfort of sit-
ting on a hard flat surface for a long period of time. Reduce the stresses by
removing the pebble, by wearing rubber-soled shoes, or by sitting on a cushion,
and the problem is solved. In an artificial knee, if the contact stresses between the
femoral and tibial components are too high, the plastic material will eventually
fail due to deformation and wear. If a component is too small in area, the under-
lying trabecular bone will be overstressed causing it to gradually fail, leading to
component subsidence. A principle such as minimizing stress needs to be incor-
porated into any type of artificial knee design.
The artificial knee itself consists of femoral, tibial, and patella components.
Some components can have more than one separate part. For example, a tibial com-
ponent usually consists of a polyethylene bearing fixed into a metal tibial tray,
which is termed modularity. Each component has a bearing surface and an interface
surface which connects to the bone. These surfaces can be further subdivided, an
example being that the femoral component has a bearing surface that interfaces with
the tibia, an area that interfaces with the patella, and a small area in between which
interfaces with both. This description highlights another principle that the sliding
surfaces must have the correct balance between stability and laxity. The stability
must be sufficient to prevent excessive sliding movements which the patient would
feel as instability. The laxity must be sufficient to allow sufficient rotation and flex-
ion movements and to prevent a feeling of stiffness.
As well as the design of the artificial knee itself, the means of implanting the
components must be considered. Principles apply just as much to the instrumenta-
tion and the technique as to the design itself. If the components are oriented to the
bones incorrectly or if the components are not securely fixed to the bones, the knee
may not function well or may prematurely fail. Hence the design of the instrumenta-
tion, the sizing and orientation of the components, and the way in which the fixation
is implemented are all subject to satisfying certain principles if the artificial knee is
to succeed. One example is that at the interface between the component and the
bone, during loaded conditions, there must be minimal relative interface motion, or
micromotion as it is called. The maximum micromotion tolerated is about one twen-
tieth of a millimeter.
Finally, the patient is part of the equation. Although many aspects of the patient
are outside of the control of the designer and the surgeon, the maximum functional
level of the artificial knee and the time taken to achieve that stage can be strongly
influenced by the instructions given to the patient and how well the patient follows
those instructions. The quality and speed of recovery after the surgery can also be
enhanced by the preparation in the weeks or months prior to the surgery. For exam-
ple, reducing weight, stopping smoking, and strengthening muscles by exercise or
physiotherapy prior to the surgery all benefit rehabilitation.
While sound principles are the basis of any design, they are not by themselves
sufficient to guarantee success in practice. Before being able to confirm that a new
design will be successful in a large population of patients, a field trial must be
6 1 Introduction
carried out. This applies to every new invention or product whether in medicine or
in engineering. An automobile has numerous separate parts and systems, each of
which will have been rigorously analyzed and tested prior to the model being
released to the consumer. An essential part of the testing involves driving the vehicle
under many conditions of road surfaces, climates, and driving styles. Because of
this process, modern automobiles have few major flaws although some problems
may emerge after product release, sometimes serious enough that a product recall is
necessary. With an artificial knee joint, trials on a specified number of patients for a
sufficient period of time are equivalent to testing an automobile on the open road.
In the early 1970s, a manufacturer could introduce a new artificial knee design to
the market with few restrictions. Inevitably this led to early failures of some designs
due to inadequate preclinical testing in the laboratory or lack of controlled clinical
trials. Today the situation is completely different. For a total knee to be approved for
clinical application, FDA requirements have to be met, consisting of an extensive
menu of tests for wear, strength, and function, to demonstrate safety and efficacy.
Nevertheless, even during the 1970s, when there was a proliferation of new designs
and few controls, there was a progressive improvement in their performance, while
unsuccessful design features were for the most part quickly weeded out. Even with-
out regulations, it was in nobody’s interest if a design performed poorly. The end
result was that even before the end of the decade, several basic design forms had
been developed that appeared to provide adequate function and showed no major
problems, at least in the short term.
Not surprisingly, some failure modes took several years to become apparent.
Wear of the polyethylene was one of these problems. It had been thought that
because of the low wear rate in artificial hips, the same would apply to knees. What
was different however was that in some knee designs, there was low conformity
between the femoral and tibial bearing surfaces leading to high contact stresses.
This led to accelerated wear, in some cases completely through the plastic layer.
Another failure mode that took several years to recognize was component loosening
due to inadequate bonding between the acrylic cement and the trabecular bone. This
was due to insufficient penetration of cement into the spaces between the trabecu-
lae, caused by not removing fluids and trabecular debris, applying the cement when
it was too “doughy,” and not applying enough pressure to force the cement into the
bone. A further reason was that offset loading conditions could actually produce
tensile stresses across part of the interface, pulling the cement out of the trabecular
spaces.
Looked at from today’s perspective, some of the deficiencies in the early designs
may seem obvious. But at that time, personal computers were not yet invented, nor
software for design, analysis, and modeling. Not only that, but biomechanical
knowledge such as the forces acting in activities, the kinematics and what controlled
it, and the geometry and mechanical properties of the tissues of the knee, was basic
and fragmentary. Designing a component which became an integral part of an ana-
tomical structure such as the knee was unfamiliar territory. The shapes of the
implants and the instruments for an artificial knee needed to account for the anat-
omy, the boundaries imposed by the tissues, and the exposure at surgery. Many
1 Introduction 7
surgeons were not familiar with engineering principles, such as the concepts of
stress and strain, forces and moments, elastic modulus, and tensile and compressive
strength, although possibly so in a conceptual or intuitive way. Finally, relatively
few testing methods had been developed for the evaluation of new design
concepts.
Those challenges were steadily overcome, particularly where there was close
collaboration between surgeons and engineers. A number of the early artificial knee
designs were surprisingly successful due to the application of engineering, ana-
tomic, and surgical principles. The process typically started with design ideas which
originated from discussions between the surgeon and the engineer. These ideas
would then have to satisfy both engineering and surgical principles at the same time.
A component that could not be easily inserted into the knee would be rejected, as
well as a means of fixation to the bone which would overstress certain areas of the
bone. There would be agreement that surgical instruments needed to produce cor-
rect lower limb alignment for uniform force distribution and ligament balancing.
Certain decisions had to be deferred pending the results of analysis or experiment,
for which the methodologies had to be devised.
CRR(%)
35
OA All revisions TKA

1976-1985 n= 2,801
30
n = 16,176
1986-1995
1996-2005 n = 54,572
25 2006-2015 n = 111,188
20
Copyright © 2017 SKAR
15
10
0
0 2 4 6 8 10 12 14 16 18 20
Year after index operation
Fig. 1.3 Cumulative revision rates for artificial knee joints (not including unicompartmental) for
four periods of 10 years. The revisions are for any reason. The revisions for the two latest 10-year
periods are much less than for the first periods. The solid colors indicate the confidence intervals.
Most of the cases were for osteoarthritis. Female/male ratios were about 60/40. Seventy hospitals
participated. Infection accounted for just less than 1% of revisions. There were small percentages
of cases where the patella was resurfaced and where uncemented components were used. Three
designs of condylar replacement knees from major companies dominated the usage. (Reprinted
with permission from the Swedish Knee Arthroplasty Registry, Otto Robertsson et al. 2017)
8 1 Introduction
The history of the development of the artificial knee has been described in a
number of well-illustrated articles and book chapters (Riley 1976; Murray 1991;
Ranawat and Ranawat 2012; Robinson 2005; Amendolo et al. 2012; Papas et al.
2018). It is interesting to compare the authors’ appraisals of the current status of the
artificial knee and predictions for the future, with due regard to the dates of the
articles. Lee Riley (1976), co-designer of one of the first metal-plastic total knees,
stated that “cobalt-chrome alloy and high-density polyethylene were acceptable
materials.…and could be firmly fixed with methylmethacrylate. A wide range of
prosthetic units are available which will provide acceptable clinical results if used
within their design limits.” David G Murray (1991), co-designer of an early rotating
knee, thought that current techniques for primary knees will provide excellent
results in 90% of cases. “An understanding of what has gone before is essential in
planning effectively for the future.” Raymond Robinson (2005), in a thorough
review, charted the time sequence of the different designs and demonstrated the
interplay between the inventors. Robinson interviewed numerous individuals who
were involved with these early developments. He concluded that collaboration
between specialists in different areas had produced successful artificial knees which
preserved cruciates or which substituted by a cam mechanism, each type with
mobile bearing options. Chitranjan Ranawat (2012), who organized the foundation
of The Knee Society in 1983, and co-authors, stated that “the future of TKA is
indeed bright,” with “future advances in non-cemented fixation, wear-resistant
polyethylene, and increased flexion range, being on the horizon.” They considered
that better outcomes will result from improved pain management, computer naviga-
tion, and robotic surgery. Most of all they emphasized “refined surgical techniques
based on sound arthroplasty principles.” This meant that not only should the bone
cuts be accurate, but the surrounding soft tissues should be well preserved and bal-
anced. Luca Amendola (2012), from the Rizzoli Institute in Bologna, focused on the
most successful designs, illustrated with remarkable semi-transparent images pro-
duced by Fabio Catani MD. It was implied that these designs emerged by learning
the lessons from the deficiencies of the very first group of designs from the early
1970s. Recently, Papas and co-authors (2018) said that “modern arthroplasty has
evolved into a procedure that can offer long-lasting results and functional improve-
ments.” They pointed out that preoperative, perioperative, and postoperative man-
agement was making major strides, including periarticular injections to reduce pain
and swelling. From all of these articles, a picture is obtained of interaction between
multiple inventors of different artificial knees, the outcomes steadily improving
over time, leading to the successful designs available today.
Moving now to the organization of the book, the chapters are written as a pro-
gression from one topic to the next, although each chapter can be read individually.
For that reason, there is a certain amount of repetition in the different chapters, but
this has been kept to a minimum. References are cited at the end of each chapter, so
that the chapters are self-contained. The number of references has been kept to a
minimum to avoid too much distraction from the text itself, and if preferred, the
reader can go through the text without stopping to look at any of the references
themselves. Today, there is a voluminous body of literature on almost every topic,
1 Introduction 9
and an exhaustive listing would have been prohibitive. The selection basis for the
early work was to search out the first articles dealing with the design and clinical
evaluation. For longer-term evaluations, recent reviews or meta-analyses were
favored, together with a selection of the highest-quality individual articles. In an
effort to avoid bias, articles with different conclusions were also selected.
On the question of “who invented things first,” unless clear-cut, there is no attempt
to specify this. The reason is that there was a great deal of interaction between early
inventors, and while certain individuals can be credited with particular designs or
ideas, the many exchanges and collaborations between those in the field were influ-
encing factors. There were various time intervals between first originating a design,
lecturing about it, publishing an article, and carrying out the first surgeries. In some
cases, a design was produced by a manufacturer immediately; in other cases, the
designer found it difficult to interest a manufacturer, resulting in a long delay.
An attempt has been made to be as objective as possible in presenting the mate-
rial in this book. However there is inevitably more detail of work the author was
personally involved with, simply because more detail was available. A lot of infor-
mation about some of the early inventions is no longer available and hence could not
be adequately covered. In addition, much of the early work was not fully described
in publications. This is in marked contrast to today where there is a steady stream of
publications to the point where it is difficult to absorb it all. One source of bias in
the book may be that much of the work described is from England, particularly the
early work, and then, later, from the USA. At the same time, there was a great deal
of innovative work on artificial knees in Europe and Japan. Some of this work may
be underrepresented. Indeed, one of the very earliest condylar knee designs was
invented in Japan. In basic research, Scandinavia and Northern Europe are well-
known for basic research in artificial knees, from the RSA method developed in
Lund, Sweden, to the iterative finite element modeling from Nijmegen, in the
Netherlands.
The figures are not referred to directly in the text, although they are placed in
relevant positions. There are two reasons for this. Firstly, the text is written as a
continuous narrative, not requiring any illustration. Secondly, the figures together
with their explicit captions can be looked at individually without having to refer to
the text. In this way, the figures provide additional perspective. Many of the figures
are photographs of individuals in the field, almost all taken by the author, usually at
conferences. Some of the pictures were taken in the 1960s, others very recently,
providing a perspective of the extended timespan of the field of artificial knees. A
number of figures were kindly provided by colleagues and are acknowledged.
Permissions from publishers were received for all figures reproduced from articles.
Terminology is not easy in a book such as this. The usual word to describe an
artificial knee is a total knee arthroplasty (TKA) or total knee. I have mostly used
the term artificial knee, because it is more colloquial and it conveys any type from a
unicompartmental to a hinge. Acronyms have been minimized in the text, although
terms such as CR and PS quickly become used more frequently. At the end of the
book, there is a detailed glossary, while at the front, there is a figure showing all
parts of the knee.
10 1 Introduction
The idea of writing a book on the artificial knee originated in late 2015. Because of
the success of several different types of design, it seemed that an explanation of the
engineering principles behind the designs would be a useful focus. Initially a lecture
was created on this theme and first presented at the Insall Traveling Fellows annual
meeting in 2016, organized by Norman Scott MD. From there, planning for the book
commenced. To obtain the initial material, the author carried out personal interviews
with early pioneers in the field or with those who were associated with them. These
were (in alphabetical order) Bob Averill, David Blaha MD, Peter Bullough MD, Al
Burstein PhD, Michael Freeman FRCS, Bill Krause PhD, John O’Connor PhD, Jim
Pritchett MD, Chit Ranawat MD, Norman Scott MD, Bahaa Seedhom PhD, and Alan
Swanson PhD. The first interview was with Alan Swanson in June 2017, at the
Institution of Mechanical Engineers in London, England. The methodology for the
interviews was to formulate questions derived from their early publications, or from
direct knowledge of their work, and then to carry out a face-to-face interview which
was recorded. Many other individuals were contacted by email or personally and pro-
vided valuable information on specific areas, for which the author is grateful. I par-
ticularly thank Thomas Einhorn MD who provided overviews about various aspects
of the book and reviewed some of the material. Many new insights and information
were uncovered during all of these exchanges such as the original ideas and how the
ideas came about, factors often not well explained in early publications. The inter-
views were particularly revealing, especially as such a long time had elapsed since the
time of the early work until now. Some of these pioneers are still active in orthopedics
today, while others are pursuing endeavors in other areas.
The author is still involved with artificial knee experimentation and design,
working in an academic setting. It needs to be stressed that the book is written
from the perspective of a biomechanical engineer who has not performed any
surgeries nor formally examined patients in a clinic. Surgeons who specialize in
artificial knees will typically perform up to several hundred surgeries each year
and over the course of several decades will follow up thousands of patients. This
will result in a perspective that cannot be experienced by an engineer. Nevertheless,
the material in this book has been accumulated since 1966 at the start of the
author’s PhD research work. Being supervised by Duncan Dowson and Verna
Wright at Leeds University in England was a unique experience at a special time
in the field of the biomechanics of joints. There was the opportunity for interac-
tion with pioneers in the field in England, such as John Scales, Michael Freeman,
Alan Swanson, John O’Connor, and Peter Bullough, but the person who stands
out the most, understandably, was John Charnley. He, more than anybody else,
started off the entire field of artificial joints. It was always a distinct honor to
meet him at his Hip Center in Wrightington or at conferences in England or the
USA.
In 1969, after completion of a PhD, the author had the privilege of being
employed at the Hospital for Special Surgery in New York. I owe the deepest
thanks to Harlan C Amstutz MD for this opportunity. Artificial hips were already
being used at the Hospital, mainly through the efforts of Philip D Wilson Jr
MD and Harlan Amstutz. This led to the interest in developing an artificial knee,
1 Introduction 11
Fig. 1.4 John Charnley (center) welcomed visitors to his Hip Center at Wrightington Hospital in
England and was interested in discussing all aspects of hip replacement and other orthopedic top-
ics. In about 1967, Bahaa Seedhom (left) and Peter Walker (right) visited to observe a total hip
operation, where the new laminar air flow unit was being used. The visit included a lab visit in
which wear tests of materials were in progress, and new instruments were being developed.
(©Peter S Walker)
work in this area already having started in England. Robert Lee Patterson MD and
Philip D Wilson Jr MD, successive Chiefs of Orthopaedics, provided a supportive
environment for the goal of developing artificial knees and other joints. The author
worked primarily with Drs. Chit Ranawat, John Insall, and Alan Inglis on this
endeavor, as well as with Fellows and Residents. After leaving the Hospital for
Special Surgery, the author worked at Cintor in Boston (now part of J&J);
Howmedica in New Jersey; Brigham & Women’s Hospital in Boston, affiliated
with Harvard Medical School and MIT; the Royal National Orthopaedic Hospital
in Stanmore, affiliated with University College London; and finally New York
University Langone Orthopedic Hospital where Joseph Zuckerman MD,
Department Chairman, has provided a stimulating academic environment in
which to work since 2003 for which I am grateful.
During all of this time, the author has participated in the Orthopaedic Research
Society (ORS) and the American Academy of Orthopaedic Surgeons (AAOS)
almost every single year since 1969. The Knee Society, founded in 1983, has been
a constant source of education and insights through its distinguished members.
Other notable society conferences have included the International Society for
Technology in Arthroplasty (ISTA) and Computer-Assisted Orthopaedic Surgery
(CAOS), as well as the European Society of Biomechanics (ESB) and the European
Orthopaedic Research Society (EORS). But it has been the innumerable contacts
and discussions with individuals in the field which has been the most valuable and
provided the most personal pleasure.
12 1 Introduction
I thank Minette Gorelik, my assistant, for typing the many revisions of each
chapter; Albright Djokoto, a student at New York University, for navigating EndNote
and RightsLink, as well as a lot of typing; and David DeSimone, librarian at NYU
Langone Orthopedic Hospital, for advice on references and sources.
Most of all I thank my wife Wuliang for traveling with me to numerous confer-
ences, for participating in every interview, for meticulously reviewing every chap-
ter, and for making it possible for me to complete this book.
References
Amendola L, Tigani D, Fosco M, Dallari D. History of condylar total knee arthroplasty. In: Recent
advances in hip and knee arthroplasty: InTechOpen; 2012. https://doi.org/10.5772/28203.
Brand RA, Mont MA, Manring MM. Biographical sketch: Themistocles Gluck (1853-1942). Clin
Orthop Relat Res. 2011;469(6):1525–7.
iData Research. Total knee replacement statistics 2017: Younger patients driving growth. 18 July
2018.
Murray DG. In: Laskin RS, editor. History of total knee replacement, Total knee replacement.
Basel, Switzerland: Springer; 1991. p. 3–15.
Papas PV, Cushner FD, Scuderi GR. The history of total knee arthroplasty. Tech Orthop.
2018;33(1):2–6.
Ranawat A, Ranawat C. The history of total knee arthroplasty. In: The knee joint: Springer; 2012.
p. 699–707.
Riley LH Jr. The evolution of total knee arthroplasty. Clin Orthop Relat Res. 1976;120:7–10.
Robertsson O, Sundberg M, Lidgren L, W-Dahl A. The Swedish knee arthroplasty. Annu Rep.
2017.
Robinson RP. The early innovators of today’s resurfacing condylar knees. J Arthroplast. 2005;20(1
Suppl 1):2–26.
Early Biomechanical Knowledge
What Was Known About Knee
2
Biomechanics by 1970?
In the context of the advances which had been made in engineering technology in
general, it may seem surprising that artificial joints for the hip and the knee were not
invented until the 1960s. After all, these devices seem simple enough, consisting
only of two main components which are fixed firmly to the bones, providing a low-
friction and wear-resistant bearing. For thousands of years, simple journal bearings
consisting of a rotating shaft in fixed bearings have been used in carts and carriages,
while a basic form of rolling element bearings consisting of rows of logs was used
for transporting enormous stones. Between 1498 and 1500, Leonardo da Vinci
invented a ball bearing where spheres were captured in cages, producing continuous
rotation by rolling rather than sliding, reducing both friction and wear. Successful
bearings were routinely used in many types of transportation. The Stockton and
Darlington steam trains started running in 1825, Ford’s production line was assem-
bling automobiles from 1913, while the first commercial airplane flight was in 1914.
Treating arthritis of the joints was regarded as a medical problem, and it seems that
there was little involvement of technical experts who could have contributed to
devising practical solutions.
Joint Lubrication
The history of how different types of bearing evolved over time has been described
in an epic book called the History of Tribology by Duncan Dowson (1998), tribol-
ogy being the science of friction, lubrication, and wear. The book was a late proj-
ect of Duncan Dowson, who from the mid-1960s was Professor of Tribology at
Leeds University in England and a pioneer in elastohydrodynamic lubrication. In
his book, Dowson included the topic of human joint lubrication because it was
quite unlike anything which had been devised in engineering. The exceedingly
low friction in animal joints was first measured in experiments performed in the

14 2 Early Biomechanical Knowledge
late 1950s (McCutchen 1962). The hypothesis was that in weight-bearing condi-
tions, water in the cartilage will be squeezed out due to the high pressure, forming
a thin film of water in between the opposing cartilage bearing surfaces. This water
film will result in almost frictionless sliding, the mechanism being called “weep-
ing lubrication.” In addition, the slippery synovial fluid, which is present in the
joint, will protect the bearing surfaces when the weeping mechanism was not
operative. In the 1960s, stimulated by McCutcheon’s work, experiments by other
researchers led to alternate theories being advanced with esoteric names: squeeze-
film lubrication, boundary lubrication, elastohydrodynamic lubrication, and
boosted lubrication by synovial fluid entrapment and enrichment. All of these
mechanisms could operate during the different functional conditions of human
joints, augmented by complex macromolecular interactions at the surfaces (Jahn
et al. 2016). However, the weeping lubrication theory, which can be likened to
walking on a film of water, has considerable appeal. Its proponent, Charles
McCutcheon, was convincing and persuasive in promoting it.
Dowson was intrigued by the extremely low friction which had been measured in
human joints, together with the durability of the cartilage bearing material which
lasted for a lifetime in the majority of people. This signified that the bearing sur-
faces were separated by a fluid film, but how could such a film be sustained in a
bearing which operated at such low sliding speeds and where the sliding direction
kept reversing? Dowson applied his theories to the problem and found that while
fluid films were predicted under certain operating conditions, much of the time there
would be boundary lubrication. It had been recognized that other factors were
involved in protecting the bearing surfaces from damage: the porosity of the articu-
lar cartilage, the hyaluronate constituents of the synovial fluid, and macromolecular
boundary effects. Could studies of human joints be applied in some way to the
design of novel bearings, possibly for resurfacing the arthritic joint? In 1966 Duncan
Dowson connected with Verna Wright, a Professor of Rheumatology at Leeds
University, to establish the “Bioengineering Group for the Study of Human Joints.”
Among the first graduates were Peter Walker, Tony Unsworth, and Bahaa Seedhom.
This was one of the early examples of a collaboration between engineering and
medical departments, to conduct basic research and solve practical problems with a
multidisciplinary approach. The activity was renamed as the Institute for Medical
and Biological Engineering and directed from 1993 to 2016 by Professor John
Fisher, who made the Institute into one of the leading centers on the tribology of
artificial joints.
Failure of the joint lubrication was seen as one of the major causes of osteoarthri-
tis of joints. For the knee, there were associated problems of instability of the joint
and misalignment of the limb. The solutions for knee arthritis in the mid-1960s were
debridement, osteotomy, tibial plateau replacement, femoral condyle resurfacing,
patella resurfacing, and metal hinges. The material used for the implants was mainly
cobalt-chrome-molybdenum alloy, trade name Vitallium. This material was well
tolerated and had high strength and wear resistance. The main limitation of all these
2 Early Biomechanical Knowledge 15
treatments was lack of complete pain relief and sometimes residual instability. Lack
of rigid fixation of the implants to the bone was a frequent cause of failure, causing
pain and bone resorption.
Joint Biomechanics
To design a successful artificial knee, knowledge of the biomechanics of the knee

both in normal and arthritic conditions was needed. While in the hip joint, the ball-
in-socket arrangement could be easily reproduced mechanically, for the knee, the
mechanics were far from simple and were not well understood. Even though biome-
chanical studies had been carried out in Europe in the nineteenth and early twentieth
centuries, this work was not widely known because of its being written in German
and French and published in documents which were only accessible to specialists.
Due to the efforts of Vera Pinskerova, Paul Maquet, Robert Furlong, and Michael
Freeman, the content of much of this work was translated into English and summa-
rized (Pinskerova et al. 2003). Many of the findings from this work could have been
directly applied to artificial knee design.
Starting with the basic studies of the geometry of the joint surfaces, this would
have application to the design of resurfacing components. The sagittal outline of the
medial femoral condyle was described by a posterior-distal circular arc and a distal-
anterior circular arc of larger radius, with a tangent point between the two arcs at
about 30 degrees of flexion. On extending the knee, the anterior arc articulated
against the distal-anterior tibial surface, elevating the femur on the tibia, acting as a
gravity brake to hyperextension as well as tightening soft tissues. As the knee was
flexed from a fully extended position, there was always an internal rotation of the
tibia about its own long axis. This caused the lateral femoral condyle to displace
posteriorly on the tibial plateau, whereas the medial femoral condyle remained at
almost the same location on the tibial surface. The lateral and medial menisci moved
with the femoral condyles. As seen on the upper surface of the tibia, the center of
the internal rotation was located on the medial side. Hence there could be consid-
ered to be two simultaneous rotations as the knee was flexed, flexion of the tibia
about a transverse axis through the center of the femoral condyles and an internal
rotation of the femur about a medial vertical axis through the tibia. The role of the
cruciate ligaments was described as a four-bar linkage, where at least some of the
fibers of the anterior and posterior cruciate were taut at all flexion angles, thereby
guiding the motion of the femur on the tibia. There was however little explanation
of how the internal rotation of the tibia was guided. Had this early work been known
when the first artificial knees were designed in the late 1960s and early 1970s, the
designs may have been very different from what they were. In particular, it is likely
that the artificial knees would have been asymmetric, with a more stable medial side
and a more mobile lateral side and with some mechanism for causing the posterior
displacement of the lateral femoral condyle during flexion.
ANT
MED
Femur flexing clockwise

A P
P s P s
A s
hA
a a a
p p
B B A B P
A P A P
I II III
Fig. 2.1 Biomechanical studies in the nineteenth and early twentieth centuries. Top: Albert fixed
the tibia, passed a rod through the femoral epicondylar axis, and, at 90-degree flexion, traced the
rod as the femur was rotated internally and externally. A medial axis is seen. Bottom: Zuppinger
proposed that the cruciate ligaments guided knee motion like a four-bar linkage. However it was
later pointed out that this is an approximation because the ligaments do not remain uniformly taut
throughout flexion (Pinskerova et al. 2003). (Reprinted with permission Clinical Orthopaedics &
Related Research)
A landmark study was published in 1941 by Brantigan and Voshell from the
University of Maryland School of Medicine. Although the paper was entitled “The
Mechanics of the Ligaments and Menisci of the Knee Joint,” rather than explaining
the knee in mechanical terms, the authors focused on the anatomy first and how that
controlled the mechanics. The authors were motivated to carry out this work due to
the confusion in the field at the time: “a study of the literature on this subject leaves
one bewildered.” No less than 100 knee specimens were studied. One original aspect
of the work was that the ligaments were treated as broad structures where different
parts could loosen and tighten and where extremes of positions occurred gradually
due to the strain-stiffening characteristics of the ligaments and menisci. The loosen-
ing of the lateral collateral ligament as flexion proceeded was noted, together with
In extension
In flexion anterior tension
fibers tense, posterior distributed
fibers relaxed throughout
width
In every angle
some part of
ligament is
tense
Posterior sliding
of ligament
in flexion
Fig. 2.2 Concepts of ligament function applied to the medial collateral ligament which consists
of multiple fibers. The fibers are tensed by different relative amounts during the flexion range. This
implies that the stiffness of the ligament varies with flexion angle. These concepts have important
application to computer modeling of the knee (Brantigan and Voshell 1941). (Reprinted with per-
mission from Journal of Bone & Joint Surgery A)
extreme posterior displacement of the lateral femoral condyle. In function, muscle

and body weight forces restrained varus-valgus motion. The menisci were observed
to cushion the contacts in hyperextension and in full flexion: “Motion of the menisci
gave good evidence that the medial femoral condyle acts more nearly as the axis of
rotation of the knee joint.” “Some portion of the medial collateral ligament is taut in
all places of extension and flexion.” “It is hardly possible to give one or more sepa-
rate or definite functions to any one ligament.” These findings are remarkably con-
sistent with those of the earlier European work described above, but with an
additional perspective from an anatomic standpoint.
While an understanding of the anatomy and internal mechanics of the knee is
essential, the functional aspects in activity are equally important. Outstanding
research on the mechanics of locomotion was published in 1947, entitled
“Fundamental Studies of Human Locomotion and Other Information Relating to

Design of Artificial Limbs.” The work was supervised by Professor H. D. Eberhart,
the Professor of Civil Engineering, University of California at Berkeley, and by
VerneInman, Professor of Orthopaedics, University of California in San Francisco.
The concept was to “bridge the gap between the engineering and medical phases.”
The work was commissioned by the National Research Council to advance the
design of artificial limbs, for which there was a substantial need following the
Second World War. The available designs of artificial limbs were not functioning
well, and trial-and-error methods had failed to make substantial improvements.
Much of the research consisted of motion analysis of level walking and stair
climbing, where displacements and rotations in all planes were measured. For the
knee during walking, it was found that there was a relative internal-external rota-
tion between the femur and tibia averaging 9 degrees. A major conclusion of the
study was: “Restrictions placed upon the normal transverse rotations will, to vary-
ing degrees, modify the synchrony and rhythm of walking. The awkwardness and
Fig. 2.3 An early gait laboratory developed in the Department of Mechanical Engineering at the
University of California, Berkeley, by Professor HD Eberhart PhD and Dr. Verne Inman MD,
PhD. The purpose was to study human locomotion with application to the design of prosthetics for
lower limb amputees (Eberhart and Inman 1947)
Walking up stairs
Time interval between phase 0.125 sec.
Subject No. 187
Walking down stairs
Fig. 2.4 Stick figures produced by Eberhart and Inman showing the successive angles of flexion
in ascending and descending stairs. In both cases, the maximum flexion angles exceeded 90
degrees (Eberhart and Inman 1947)
discomfort of a subject required to wear a leg brace, which does not provide for
rotations at the hip, knee, and foot, are no doubt due, in part, to the restriction of
these motions.” Ground-to-foot forces, both vertical and shear forces, were mea-
sured using specially designed force plates. Electromyography was used to deter-
mine muscle recruitment in the different phases of the activities. This data was
used to estimate the peak forces in the knee during walking at twice body weight
although details of the calculations were not provided. Some of the work was
made available in a publication in 1953 (Saunders et al. 1953), but the full extent
of the work was only published in a special report in 1947 to the National Research
Council. Fortunately the work was continued over many years, and eventually the
Fig. 2.5 Using advanced force plate design, Eberhart and Inman were able to measure the force
components between the foot and the ground during different activities. Force peaks at heel strike
and toe-off were clearly shown. The forces within the knee joint (estimated to be at least two body
weights) are increased due to muscle forces. The fore and aft forces were small but still significant
in that they are indicative of the shear forces within the knee joint (Eberhart and Inman 1947)
main findings were compiled in a book entitled Human Walking (Inman et al.
1981). The book contained superb mechanical depictions of the lower limb in
various activities showing flexion angles and muscle action.
Only a few years later, in 1955, Arthur Steindler, Professor of Orthopaedic
Surgery at the State University of Iowa (Buckwalter 2000), published a book enti-
tled Kinesiology of the Human Body (Steindler 1955) where he described basic
biomechanical principles of joint function. He defined rolling and sliding motions
in the knee and pointed out that contacts were not points but areas of contact due
to the deformability of the articular cartilage. He illustrated the structure of bones

and joints and showed how muscles acted at certain lever arms from the joint cen-
ters. Motions of limb segments were described as open chain or closed chain
depending on whether the foot was anchored to the ground or not. While Steindler
did not reference any studies on these topics, it is clear that he used his intuitive
knowledge based on insightful experience over many decades of orthopedic
practice.
By the early 1960s, knowledge of the anatomy and properties of the major struc-
tures around the knee was sufficient to provide a general understanding of the
mechanical function. The tubular structure of the long bones transmitted axial loads
while providing adequate bending strength. At the knee joint, the expansions of the
femoral and tibial condyles, composed of trabecular bone with a thin cortical shell,
served several functions. In the frontal plane, the condyles were sufficiently wide
such that varus-valgus moments could be supported with condylar lift-off occurring
only under extreme conditions. The posterior projections of the femoral condyles
provided enough clearance from the posterior tibial plateaus to allow for a high
range of flexion. This is particularly evident during kneeling, when a flexion angle
approaching 180 degrees can be achieved by some individuals. But even with the
large surface areas of the femoral and tibial condyles and their smooth contours, the
joint could not function as a bearing if the surfaces were bone. The contact areas
would be too small due to the high modulus of elasticity of bone, leading to small
contact areas and very high contact stresses. This would produce high bone-on-bone
friction, high local temperatures, and stick-slip motion producing a jerky action. To
prevent these problems, the bony condylar surfaces are covered with a material
specialized as a bearing, articular cartilage. Remarkably, 3–5 mm layers of this
material on the tibial and femoral condyles, as well as on the patella, provide smooth
low-friction motion with minimal surface damage during multiple activities for the
lifetime of most individuals.
But low friction itself would be insufficient in creating a long-lasting knee joint.
In the hip joint, the ball-in-socket bearing surfaces are close to being fully conform-
ing resulting in large contact areas and low contact stresses which are readily sus-
tained by the articular cartilage (Greenwald and O’Connor 1971). However, in the
knee, the lack of conformity between the femoral and tibial condyles would produce
contact areas which were too small, producing high contact stresses which would
eventually cause failure of the cartilage. It might be asked why the knee does not
have closely conforming bearing surfaces as in the hip joint. The likely explanation
is that the knee joint requires a high range of flexion as well as the ability for some
internal and external rotation, to allow for maneuverability of the leg. A conforming
uniaxial joint would not be adequate. Another requirement is that the knee needs to
prevent side-to-side bending, which is not easy considering the knee is located at the
center of a very long beam, the lower extremity. This is achieved by the large muscle
forces across the knee which press the lateral and medial condyles of the femur
against the plateaus of the tibia. A requirement in the knee joint itself is how to
allow for high flexion and internal-external rotation while maintaining large areas of
contact. The answer is the menisci which are interposed between the femoral and
Fig. 2.6 Reconstruction of the fiber pattern of the meniscus. Histological sections of the meniscus
in different planes were taken and viewed in polarized light, from which the directions of the col-
lagen fibers were deduced. Because most fibers are circumferential and the high tensile strength of
the material, it was suggested that the meniscus has a weight-bearing function between the femur
and the tibia. Such forces would have the effect of radially expanding the meniscus, causing cir-
cumferential tension (Bullough et al. 1970). (Reprinted with permission from Journal of Bone &
Joint Surgery B)
tibial condyles. Because of the deformability of the menisci and their loose periph-
eral attachments to the proximal tibia, they are able to change their shape and posi-
tion according to the location of the femur on the tibia. This is particularly so on the
lateral side where the femoral condyle slides over the convex surface of the tibia, in
contrast to the medial side where the tibial surface is somewhat dished and limits the
sliding motion.
The mechanical functions of the menisci were proposed as long ago as 1961 by
Barnett, Davies, and MacConaill in their textbook dealing with the structure and
functions of synovial joints:
• Shock absorption protecting the articular cartilage

• Increasing the congruity between the articular surfaces and increasing the stabil-
ity of the joint
• Allowing a continuation of different movements of the joint
• Prevention of undue forward gliding of the femur on the tibia, analogous to a
chock under a wheel
• Distribution of the weight across the joint by increasing the effective area of
contact
• Assisting in the lubrication of the joint
These propositions were based mainly on anatomic studies. At that time, suitable
experimental methodologies for determining the internal mechanics of the knee had
not been developed. But the insights of the authors are remarkable in that they were
all supported by subsequent experimental work after about 1970.
The first of such studies actually proposed that the menisci were weight-bearing
because the tibial cartilage beneath them was always well preserved, which could
only occur if the cartilage was loaded (Bullough et al. 1970). Additional support
for this theory was that the collagen fibers in the meniscus were predominately
circumferential indicating that the meniscus was resisting radial expansion.
Meanwhile in the 1960s, it was common practice for surgeons to remove damaged
menisci, in order to treat the pain and “catching,” not realizing that there might be
serious consequences (Smillie 1967). Surprisingly, degeneration of the cartilage
surfaces due to the elevated contact stresses and the increased sliding on the bear-
ing surfaces usually took several decades to become evident (Jackson 1968; Lanzer
and Komenda 1990).
It was realized by designers of the early artificial knees that reproducing the
complex structure and composition of articular cartilage and the menisci was
beyond the capabilities of materials science at that time, which still holds true today.
Nevertheless metal-on-polyethylene, already being used in artificial hip joints,
could still provide a bearing with a low friction coefficient of about 0.1 compared
with 0.01 of cartilage, as well as a low wear rate. It was interesting however that in
the early 1980s, artificial knees with sliding meniscus components were designed,
but using cobalt-chrome alloy and high molecular weight polyethylene, rather than
materials with more deformability.
In the anatomic knee, if the cartilage surfaces provided almost frictionless slid-
ing and the menisci reduced the contact stresses, the motion and stability were seen
to be controlled by the major ligaments. During flexion-extension, at least some of
the fibers in each ligament appeared to be taut, which led to the impression that the
ligaments guided the motion in a well-defined and reproducible way. However, this
was mainly conjecture, as up to the mid-1960s, there had been no controlled experi-
ments carried out to measure the length-tension relations of the ligaments.
Mechanical materials testing machines were only starting to be used in the field of
orthopedic biomechanics. Nor was there any data of the forces acting across the
knee, particularly the shear forces, in different activities. Consequently, ligaments
were thought of as being of a fixed length, as evidenced by the depictions of knee
motion where the ligaments were shown as bars of fixed length and single attach-
ment points, controlling the motions. Furthermore, motion was considered primar-
ily in the sagittal plane. Motions at the sliding surfaces themselves were described
as either rolling or gliding (sliding) or a combination of the two. Along with this
model of knee motion was the concept of instant centers of rotation. If a wheel is on
a fixed axle and rotates, a point on the outside of the wheel will rotate about the axle.
However if the axle moves anteriorly or posteriorly as the wheel simultaneously
rotates, the center of rotation will either be above or below the level of the axle. It is
noteworthy that the above descriptions of knee motion and the role of the ligaments
assumed that the knee was flexed passively without the high forces and moments
which occur in function. Test rigs for mechanical evaluation of knee specimens, or
means of measuring the mechanics of the living knee during activities, had not yet
been devised. Nevertheless in the field of gait analysis where the motion of limb
segments was studied, spectacular stick figures were produced using stroboscopic
photography by Muybridge, Edgerton, and others (Wing and Beek 2004), as well as
by Eberhart and Inman (1947) in their locomotion studies, described earlier in this
chapter.
Further research was carried out on the function of the ligaments in the mid-
1960s by Hallen and Lindahl from the University of Umea in Sweden. They built on
the earlier work of Brantigan and Voshell by carrying out laxity experiments and
measuring displacements and rotations. The authors considered the ligaments not as
bars of fixed length, but as extensible springs which allowed a wide range of
femoral-tibial positions, while still providing restraints at the extremes of motion
(Hallen and Lindahl 1965, 1966). They questioned the terminal external rotation of
the tibia as the leg was brought into full extension, the “screw home mechanism,”
citing some of the early work from Meyer and Fick. An angular measuring device
was used on 16 knee specimens and on 14 patients who had hip arthrodesis. In the
latter, femoral rotation could not occur, thereby making the tibial angles an accurate
indicator of rotation. While accepting that an average terminal rotation of 7 degrees
occurred in the final 20 degrees of extension, in the case of the living subjects, they
found that the subject could actively rotate the tibia such that “the outward rotation
can be more or less prevented by will power.” Even for the specimens, “depending
on the position of rotation in 20 degrees flexion and full extension, it was possible
to obtain passively an outward or an inward rotation or no rotation at all.” However
they made the important observation that when the knee was flexed through its neu-
tral path of motion at any flexion angle, rotational laxities occurred when axial
torques were applied. Hallen and Lindahl also measured the laxities after serial
sectioning of different ligaments such that by calculating the laxity increases, they
deduced which structures were important in controlling the rotation. These authors
also pointed out that full extension of the knee was a special situation where the soft
tissues of the posterior capsule tightened, applying high forces across the condyles
and hence producing varus-valgus stability. They recognized that when compressive
forces were acting due to soft tissues, body weight, or muscle action, large varus or
valgus moments were required to produce condylar lift-off. Even today, full exten-
sion is not easy to define, because the end point depends upon the extension moment
which is applied. Furthermore, the angle of flexion can be defined either by the long
axes of the entire bones or relative to the centerlines of the femur and tibia in the
10–15 cm adjacent to the knee.
Considering now the forces at the knee due to the ground-to-foot force, the
medial-lateral force distribution will depend upon the overall alignment of the lower
Fig. 2.7 Hallen and Lindahl carried out experiments in knee specimens to determine the laxities
including varus-valgus and internal-external rotations. The bones were fixed with external prongs
and the angular measurements measured with protractors. They carried out serial sectioning of
ligaments, a novel idea at the time, to determine the relative contributions to stability (Hallen and
Lindahl 1965, 1966). (Reprinted with permission from Acta Orthopaedica Scandinavica)
limb. In knees with a varus deformity, the forces on the medial side of the knee will
be higher than on the lateral side. This can then cause damage to the medial menis-
cus and cartilage, leading to yet more varus and even higher medial forces, which is
a common pathway of osteoarthritis. A common treatment for varus osteoarthritis
was osteotomy (Jackson 1968) whereby the force balance across the knee was cor-
rected by realigning the proximal tibia by removing a wedge of bone or creating a
wedge-shaped gap or even making an arc-shaped cut and rotating the proximal tib-
ial fragment. This had a sound biomechanical basis in that by reducing the medial
forces, pain relief and even restoration of cartilage were achieved. Mark
Coventry MD from the Mayo Clinic was a proponent of osteotomy and presented a
number of valuable guidelines on the technique, indications, and long-term results
(Coventry 1965; Coventry et al. 1993). Another way of realigning the knee was to
interpose metal plates between the medial condyles using either MacIntosh or
McKeever tibial plateaus. The insertion involved creating a horizontal cut across the
medial plateau. There was no direct bonding between the lower surface of the
implant and the exposed trabecular bone, but shear stability was provided by a tex-
tured surface or by keels. The metal bearing surface was smooth for articulation
with the medial femoral condyle. Clinical studies showed a correction of alignment
with pain relief, but similar failure modes as for osteotomy could occur, namely,
pain or recurrence of the deformity (Scott et al. 1985).
Returning to the question as to why it was not until the 1960s that artificial hips
and knees were introduced, the answer may lie in the traditions of orthopedic surgery
and the way in which advances were made. There was little serious collaboration
between surgeons and engineers, even though some surgeons may have had engi-
neering expertise. Prior to about 1960, collaboration was generally between surgeons
and the technical staff of orthopedic manufacturing companies on the aspects of
design and materials for implants, but the companies did not have the background or
resources for in-depth analysis or testing. However during the early 1960s, a revolu-
tionary development occurred. John Charnley, an orthopedic surgeon from England,
developed a successful artificial hip joint. Charnley was a unique individual who
worked at Wrightington Hospital, not far from Manchester in northwest England.
Wrightington itself was a small village, while the hospital was a collection of single-
story buildings originally for the treatment of tuberculosis.
Charnley’s artificial hip consisted of an intramedullary stem with a polished
sphere at the top, articulating in a high-density polyethylene socket to provide a
low-friction and wear-resistant bearing. The components were fixed to the femur
and acetabulum using acrylic cement which hardened on application at surgery.
This material had been used by Edward Haboush in 1957 at the Hospital for Joint
Diseases in New York, to fix a cap over the arthritic femoral head and neck in sev-
eral cases of hip arthritis (Haboush 1958). Mechanical problems prevented
Haboush’s work from progressing further. However, Charnley had seen this work on
a visit to New York and, when he returned to England, consulted with Dennis Smith,
a Professor of Dentistry at Manchester, about applying acrylic cement to the
Charnley hip. Acrylic was already in widespread use for dental applications (Smith
2005). The principles underpinning the fixation method were that the fixation mate-
rial would act as a grout between the components and bone rather than as an adhe-
sive and that the cement would be pressurized on insertion to provide a complete
interlock with the interstices of the trabecular bone. A crucial advantage was that the
rigid fixation of the components to the bones had the effect of eliminating the pain.
The development work was carried out in Charnley’s biomechanical laboratory
set up in the hospital. The laboratory was run on a very limited budget with used
machines donated by local factories. Who could have known what was about to be
produced! By the mid-1960s, it was evident that the artificial hip design was work-
ing well and was likely to be a successful solution to the treatment of osteoarthritis
of the hip. It might have seemed a simple matter to apply the same principles to the
design of an artificial knee joint, and while this may have been the case in terms of
the materials, it transpired that a great deal of work was needed on the design of the
artificial knee itself to make it work as well as the hip.
We now consider the further understanding of knee biomechanics and the func-
tions of the different structures, which was needed in order to design a successful
artificial knee. A notable void in the knowledge was the magnitude and direction of
the forces acting through the knee during everyday activities. In the mid-1960s,
there had been a major breakthrough in this area for the hip joint by John Paul, a
Professor of Bioengineering at Strathclyde University in Glasgow, Scotland. By
using a combination of gait analysis, ground-to-foot force measurement, electro-
myographical data, and anatomical measurements, Paul used equilibrium equations
to solve for the forces acting across the hip during walking and other activities (Paul
1969). The most important finding in the force analysis was that in the frontal plane,
the ground-to-foot force caused a moment about the center of the femoral head,
which had to be equilibrated by the abductor forces. But the moment arms were
such that the abductor force had to be twice that of the ground-to-foot force, so the
total force across the hip joint was about three times body weight, much higher than
expected.
This force analysis was a landmark achievement in biomechanics, which allowed
researchers and designers to carry out analysis and testing which had previously
been impossible. For example, Seth Greenwald at Oxford University immediately
used the data as input for his experiments on the contact areas in the hip (Greenwald
and O’Connor 1971). Following the hip work at Strathclyde University, James
Morrison, a PhD student of John Paul, applied similar methods to determine knee
forces, the results being published in 1968 and 1970. (It was surprising that
Morrison’s publications were not co-authored by Paul, who was the originator of
the methodology and supervisor of Morrison.) Applying the hip concepts to the
knee, it was apparent that in the sagittal plane, the line of action of the
Fig. 2.8 John Paul PhD,

Professor of
Bioengineering at
Strathclyde University in
Glasgow, Scotland. He was
a pioneer in the
methodology of calculating
the forces in the hip joint
in the mid-1960s. The
method was then applied to
the knee joint in the late
1960s. Over several
decades, he chaired
committees in the
International Standards
Organisation and the
Institution of Mechanical
Engineers in formulating
standards for the testing of
artificial hips and knees.
(Photo taken at his
university c 1975) (©Peter
S Walker)
5
Subject N˚ 2
Strike
Test No:
2
4 3
Heel
Joint force / Body weight
3
y
x
z
1
Reference axes
0
60 80 100 20 40 60 80
Percentage of cycle
10 JOINT FORCE AT KNEE - LEVEL WALKING
Fig. 2.9 The forces across the femoral-tibial bearing surfaces, calculated from gait analysis,
ground-to-foot forces, electromyography, and a model of the knee containing the geometry and the
ligaments. There were two force peaks just after heel strike and then a larger force peak at toe-off.
The work was carried out at the University of Strathclyde by James Morrison, supervised by John
Paul, and published in 1968 and 1970. (Reprinted with permission from Journal of Biomechanics)
ground-to-foot forces was at a distance from the femoral-tibial contact points such
that either the quadriceps or the hamstrings muscles would need to apply larger
forces than body weight to equilibrate. As for the hip, this would produce peak knee
joint forces in the region of three times body weight. These peaks occurred at heel
strike and toe-off in the stance phase of walking. In addition to these vertical forces,
anterior-posterior shear forces and internal-external torques occurred between the
femur and the tibia, although of much smaller magnitude than the axial force.
It is against the above backdrop that the first artificial knees were invented.
However, further background is needed to put into context the emergence of the
modern-day artificial knee. Metal on polymeric bearings was already in widespread
use in engineering applications. To determine a suitable polymer for the hip,
Charnley had carried out wear studies on specially designed wear machines on
numerous polymeric materials to identify those with sufficient wear resistance.
After initially choosing Teflon, which showed excessive wear in patients within
1–2 years, high-density polyethylene was subsequently found to have an acceptably
low wear rate. As well as low wear, Charnley also emphasized low friction to mini-
mize the shear stresses at the interfaces. He demonstrated the effectiveness of metal-
on-polyethylene combined with a small-diameter femoral head, using a pendulum
machine. The longer the time before the swinging ceased, the lower the frictional
torque. Charnley’s design had far lower friction compared with the McKee-Farrar
metal-on-metal hip being used at the time. Hence polished metal against high-
density polyethylene seemed the best choice for an artificial knee.
During the 1950s and 1960s when the artificial hip was being developed, hinged
knees were being used, mainly in Europe. The uncemented type such as the Walldius
were used in salvage situations and undoubtedly served a useful purpose. The cemented
systems, such as the Shiers and the Stanmore, provided immediate stability and a use-
ful range of flexion. However there was considerable reticence for widespread applica-
tion due to the large amount of resection of the distal femur and proximal tibia necessary
and the major problem of revision due to the long cemented intramedullary stems, in
cases of revision. Today, hinged designs requiring less bone resection and with rota-
tional freedom, and an anatomic patella flange, provide a useful fallback solution for
severe arthritis or after previous revisions of conventional condylar replacement knees.
Hence, the late 1960s saw the transition from the previous treatments of osteotomy,
metal spacers, and hinges to cemented metal-on-plastic condylar replacements. This
represented a major landmark in the treatment of knee arthritis. The impetus for this
transition was undoubtedly the success of the total hip which by then was being used
in increasing numbers in many countries other than England. While many of the prin-
ciples of the artificial hip could be directly translated to an artificial knee design, it was
not immediately obvious how best to design an artificial knee as evidenced by the
widely different configurations of the early designs.
It was not surprising that one of the very first knee designs was produced at
Wrightington by a visiting Fellow, Frank Gunston from Winnipeg, Canada. The Hip
Center at Wrightington was an ideal place for this initiative due to the presence of a
biomechanical laboratory where the hip development had been carried out. The knee
that was designed, called the Polycentric, could be developed and manufactured in
the laboratory. The surgical instrumentation was simple yet accurate because the
femoral components, being half-circular metal discs, were designed to be inserted
into slots in the femoral condyles, while the plastic tibial components were to be
similarly positioned in the tibial condyle. This was an innovative and fundamental
concept in design which has stood the test of time to this day: reproduce the original
contours of the bearing surfaces and preserve all of the ligaments. In parallel with
this work was that of Michael Freeman at the London Hospital and Alan Swanson at
Imperial College London. They were given the go-ahead to form a bioengineering
unit in order to start work on an artificial knee design. It was considered that surgeon-
engineer collaboration supported by a major technical university was necessary for
such an enterprise. These two early beginnings of artificial knee design were the
starting points for an immense expansion in the design and use of artificial knees, in
England, in Europe, in Japan, and then in the USA.
In parallel with this expansion, there was a need for methods of evaluating the
function of patients other than by simply taking radiographs and measuring the flex-
ion angle. When problems such as component loosening, instability, or polyethylene
wear occurred, biomechanical experimental studies were needed to address these
problems. Gait analysis was one of the methods used. Ed Chao PhD, who joined the
Mayo Clinic in 1971, established an advanced laboratory for the purpose of
Fig. 2.10 Ed Chao PhD

became Director of
Bioengineering at the
Mayo Clinic in Rochester,
NY, in 1971 and served in
that position for 20 years
before moving to Johns
Hopkins University. Ed’s
rigorous analytic, design,
and teaching abilities
enabled him to mentor
hundreds of engineering
and medical students in his
lab and introduce
innovative biomechanical
methodologies. The
achievement that he is the
most proud of is the design
of bone tumor implants for
adult and juvenile patients,
making the implants both
functional and durable. He
was a dedicated member of
the International Society of
Limb Salvage (ISOLS).
(Photo taken c 2004)
(©Peter S Walker)
evaluating the Polycentric and Geometric artificial knees which were starting to be
used by Mark Coventry, Richard Bryan, and others. Chao analyzed the data from the
motion capture system and the force plate data by defining numerous mathematical
parameters and was able to compare the function of the artificial knee patients with
that of normal subjects of a similar age range (Chao and Stauffer 1974).
In parallel with the invention of many new artificial knee designs, the 1970s saw
an overall expansion in activities relating to human joints and their artificial replace-
ments. At Case Western Reserve University in Cleveland, Ohio, Victor Frankel MD
and Albert Burstein PhD were running courses to educate surgeons in biomechanics
and even developed an instrumented tibial component to measure the knee forces
in vivo, although this work was cut short due to lack of funding. Orthopedic and
biomechanics societies were featuring this popular new area in their programs.
Many special courses and conferences were organized. At Columbia University in
New York, Van Mow PhD was formulating mathematical models for the behavior of
articular cartilage and for soft tissues such as ligaments and tendons. This work was
the foundation for further advances by researchers such as Savio Woo PhD and
Gerard Ateshian PhD. Peter Walker PhD and Keith Markolf PhD devised experi-
mental methods for measuring the laxity and stability of the knee and attempted to
match the behavior of artificial knees with that of the normal intact knee. At the
University of Leeds in England, the focus was primarily on friction and wear of
materials, for which new testing regimens were developed. Towards the end of the
1970s, finite element techniques were becoming available, which found an applica-
tion in the analysis of tibial component fixation, by Jack Lewis PhD, Roy
Crowninshield PhD, and others. Manufacturing companies played an important role
in the processes needed to produce the anatomic shapes of the implant components
as well as the instruments and in ensuring a high quality of materials to avoid failure
in service. In this way design and research in artificial knees expanded, fueled by the
increasing demand from patients and the generally successful results that were
being achieved.
What Was Known About Knee Biomechanics by 1970?
Studies on the biomechanics of the knee, which could have been applied to artificial
knee design, were carried out in Europe well before 1970, but the publications were
not available in English and were unlikely to have been widely known. Likewise, a
major study of knee motion carried out in the 1940s, which clearly showed axial
rotations, was only documented in a report to a government agency. Studies of liga-
ment function published before 1970 were available but do not seem to have influ-
enced artificial knee design. The only work that may have had an influence was the
determination of the forces in activity, carried out at Strathclyde University. For the
most part, the development of artificial knees from 1970 was carried out without the
benefit of much background knowledge. However a number of studies of knee bio-
mechanics in the 1970s were applied to design and in some cases provided crucial
design ideas and concepts. Almost no background work on instrumentation had
been carried out, other than by using the concepts of osteotomy, but even then, it
does not seem to have been directly applied to an appreciation of alignment and
force distribution.
References
Barnett C, Davies D, McConaill M. Synovial joints their structure and function. Charles C Thomas:
Springfield; 1961.
Brantigan OC, Voshell AF. The mechanics of the ligaments and menisci of the knee joint. J Bone
Joint Surg Br Vol. 1941;23(1):44–66.
Buckwalter JA. Arthur Steindler. Orthopaedic scholar, teacher, and clinician. Clin Orthop Relat
Res. 2000;374:171–82.
Bullough PG, Munuera L, Murphy J, Weinstein AM. The strength of the menisci of the knee as it
relates to their fine structure. J Bone Joint Surg. 1970;52(3):564–7.
Chao EYS, Stauffer RN. Biomechanical evaluation of geometric and polycentric knee arthro-
plasty. Conference on Total Knee Replacement, Institution of Mechanical Engineers, London,
16–18 September 1974.
Coventry MB. Osteotomy of the upper portion of the tibia for degenerative arthritis of the knee. J
Bone Joint Surg Am. 1965;47:984–90.
Coventry MB, Ilstrup DM, Wallrichs SL. Proximal tibial osteotomy. A critical long-term study of
eighty-seven cases. J Bone Jt Surg. 1993;75(2):196–201.
Dowson D. History of tribology. London: Professional Engineering Publ; 1998.
Eberhart HD, Inman VT. Fundamental studies of human locomotion and other information relat-
ing to design of artificial limbs. Report to Committee on Artificial Limbs, National Research
Council, Washington DC. 15 July 1947.
Greenwald AS, O’Connor JJ. The transmission of load through the human hip joint. J Biomech.
1971;4(6):507–28.
Haboush EJ. A new operation for arthroplasty of the hip based on biomechanics, photoelasticity,
fast-setting dental acrylic, and other considerations. 1958 [classical article]. Bulletin (Hospital
for Joint Diseases (New York, NY)). 1996;55(2):95–111.
Hallen LG, Lindahl O. The lateral stability of the knee-joint. Acta Orthop Scand. 1965;36(2):179–91.
Hallen LG, Lindahl O. The “screw-home” movement in the knee-joint. Acta Orthop Scand.
1966;37(1):97–106.
Inman VT, Ralston HJ, Todd F. Human walking. Baltimore: Williams & Wilkins; 1981.
Jackson JP. Degenerative changes in the knee after meniscectomy. Br Med J. 1968;2(5604):525–7.
Jahn S, Seror J, Klein J. Lubrication of articular cartilage. Annu Rev Biomed Eng. 2016;18:235–58.
Lanzer WL, Komenda G. Changes in articular cartilage after meniscectomy. Clin Orthop Relat
Res. 1990;(252):41–8.
McCutchen CW. The frictional properties of animal joints. Wear. 1962;5(1):1–17.
Morrison J. Bioengineering analysis of force actions transmitted by the knee joint. Bio-med Eng.
1968;3:164–70.
Morrison J. The mechanics of the knee joint in relation to normal walking. J Biomech.
1970;3(1):51–61.
Paul JP. Loading on the head of the human femur. J Anat. 1969;105(Pt 1):187–8.
Pinskerova V, Maquet P, Freeman MA. The anatomic literature relating to the knee from 1836 to
1917: an historic note. Clin Orthop Relat Res. 2003;410:13–8.
Saunders JB, Inman VT, Eberhart HD. The major determinants in normal and pathological gait. J
Bone Jt Surg Am. 1953;35A:543–58.
Scott RD, Joyce MJ, Ewald FC, Thomas WH. McKeever metallic hemiarthroplasty of the knee in
unicompartmental degenerative arthritis. Long-term clinical follow-up and current indications.
J Bone Joint Surg Am. 1985;67(2):203–7.
Smillie IS. The current pattern of internal derangements of the knee joint relative to the menisci.
Clin Orthop Relat Res. 1967;(51):117–22.
Smith DC. The genesis and evolution of acrylic bone cement. Orthop Clin North Am.
2005;36(1):1–10.
Steindler A. Kinesiology of the human body. Springfield: Publ Charles C Thomas; 1955.
Wing A, Beek P. Motion analysis: a joint centenary. Hum Mov Sci. 2004;23(5):iii–v.
The Design Process
Has a Systematic Design Process Been
3
Applied to Artificial Knees?
In this chapter, a systematic methodology for designing a new product will be

described. The product could be simple or complex, large or small, a completely
new product, or a variation of an existing product. The description of the process
will be related to the artificial knee. The development of the early artificial knee
designs will then be examined to determine if a systematic design process had been
applied.
Every object which has been invented has involved a design process of some
sort, even if that process has not been explicitly defined. Yet it might seem that
whatever design process has been used, it has often been deficient based on the
number of products which are delayed in their completion, experience cost over-
runs, are difficult to operate, do not look attractive, or do not work as intended. But
considering the large numbers of products which are streaming out into the market,
the deficient ones are a small percentage. To obtain a successful design solution
and avoid problems, care is needed in selecting the design process most applicable
to the product being designed and in implementing the process systematically
(Cross 2008).
Following a design process is valuable for a number of reasons:
• It requires a clear definition of the need, what the invention is required to do, fol-
lowed by a listing of the specifications and the constraints, called the design criteria.
• It is a systematic process that should result in a final product which best satisfies
the defined need and design criteria.
• Many projects are complex, involving different components and technologies,
and hence require a rigorous process to maintain an ordered progression of the
design over an extended time period.
• In an increasingly competitive world, it is essential that a design be completed in
a practical period of time, at an acceptable cost, and with a result that will satisfy
the needs of the end user.
• Adhering to a set of rules should maintain objectivity in decision-making and
minimize the possible negative effects of human factors.

34 3 The Design Process
An essential first step in the design process is to specify why something should
be designed at all, identifying a need. An example is the requirement to cross over a
river to access the territory beyond. It is easy to jump to the conclusion that a bridge
is needed. However, it is more logical to first consider different solutions. This
would then open up the solution to a boat, a catapult, a zipline, a crane, a cable car,
a helicopter, a tunnel, or even damming the river. Every solution can have a number
of variants. For example, a bridge can range from a simple suspended footbridge to
an elaborate arch carrying trains, automobiles, and pedestrians. The choice of the
most appropriate solution will be facilitated by specifying a set of design criteria.
These criteria can include the required usage such as pedestrians only or vehicles as
well, the volume of the usage, the cost constraints, the timetable, the durability
expected, the time to make the crossing, and aesthetics.
The above concepts can be applied to the artificial knee. The need for a treatment
for osteoarthritis of the knee has been recognized for centuries. However, it was
only in the late nineteenth century that a practical solution was formulated. The
inventor was a German surgeon, Theophilus Gluck, who proposed some type of
hinged joint which was fixed into the hollow intramedullary canals of the femur and
tibia using cement (Gluck 1891; Brand et al. 2011). Gluck stated design criteria for
a device which would treat the arthritis, a remarkable step which indicated Gluck’s
systematic way of thinking:
• The components must be secured rigidly to the bones.

• A minimum volume of bone must be removed.
• It must provide complete stability.
The device which Gluck designed was a narrow hinge made from ivory which
fitted into the intercondylar region, with stems which were cemented into the intra-
medullary canals. In his original article, Gluck does not indicate if he considered
various solutions, but he only described the solution which he chose for application
to patients. While the design may have been mechanically sound, it failed due to
infection when used in patients. Even though all of the design criteria may have
been met, the need for sterility was unknown at that time. Another criterion not
directly stated was durability. It was probably assumed that the materials were suf-
ficiently strong and rigid, and there would not be an issue with mechanical failure.
This was a reasonable assumption because the material was ivory, successfully used
in demanding mechanical applications in the wild. Another mechanical factor which
was not stated was the wear of the interfaces under repetitive sliding over a long
period of time, but the material strength and the smooth surfaces were probably
thought to provide adequate wear resistance.
To illustrate the influence of specifying the design criteria on the design solution
reached, exactly the same design as Gluck’s was reinvented 82 years later (Walker
and Shoji 1973). In the Biomechanics Laboratory at the Hospital for Special Surgery
in New York in 1970, Peter Walker; two orthopedic fellows, Hiromu Shoji and
David Mendes; and a veterinarian, Fernando Figarola, had been studying the center
of rotation of the knee as it was flexed and extended. It was found that a fixed-axis
3 The Design Process 35
Fig. 3.1 The remarkable similarity between the intercondylar linked knee designed and used in a
patient by Themistocles Gluck in 1891 (Gluck 1891; Brand et al. 2011) and a design tested in
animals at the Hospital for Special Surgery in 1973 (Walker and Shoji 1973). The similarity is
attributed to the identical set of design criteria specified prior to formulating the designs. (Reprinted
with permission from Arch fklin Chir and from Clinical Orthopaedics and Related Research)
hinge would be a reasonable approximation provided it included some freedom of

motion. An intercondylar hinge was designed and implanted in both dogs and sheep
such that the hinge axis coincided with the average axis of rotation of the knee.
However, because all of the load across the knee joint was transmitted by the hinge
and the intramedullary stems, the femoral and tibial condyles were unloaded. As a
result, the cartilage became degenerate, and the trabecular bone within the condyles
became porotic, which would probably result in later mechanical failure and would
compromise a revision. The research group at Special Surgery was not aware of
Gluck’s design, and their new design was arrived at completely independently. The
important point is that the reason for the similarity in designs was that the same
design criteria were used.
The degeneration of the condyles is an example of a phenomenon termed “unin-
tended consequences,” an undesirable outcome which was not predicted. In the case
of the design produced in the 1970s, it could have been predicted from the research
on disuse atrophy from Oxford University in England by Joseph Trueta and Peter
Bullough, who showed that areas of joints that were habitually unloaded became
degenerate, because chondrocytes and osteoblasts need mechanical loading to
maintain healthy tissue. While thinking outside of the box is usually easier in hind-
sight, it is probably true in general that the further an artificial system diverges from
the anatomical geometry and mechanics, the more likely there is to be some unin-
tended consequence.
While an intercondylar hinge fulfilled the design criteria, other solutions could
also have been formulated. For example, curved rails to resurface the femoral con-
dyles together with resurfacing components for the tibial plateaus would fulfill the
criterion of minimum bone resection and would also provide stability if the liga-
ments were still intact and functional. In the above example, if a choice had to be
Fig. 3.2 The Garry Bridge – Bridge to Nowhere. There was a need to connect the northeastern
corner of the Isle of Lewis, Scotland, to the main southern part. A reinforced concrete bridge was
designed and constructed from 1918 to 1923. The height was 50 ft, the span 100 ft, and the width
15 ft, and 140 workers were engaged in the construction. By the time the bridge was completed
however, the need was no longer there. But today, the bridge serves to reach the northern footpath
from which there is spectacular coastal scenery. (©Peter S Walker)
made between a linked intercondylar device and a condylar replacement, how

would a decision be made of the best solution? The obvious method would be to
assess how well each design satisfied the design criteria. Based on Gluck’s criteria,
the intercondylar hinge would be chosen on the basis of rigid fixation, minimum
bone removal, and complete stability. However, today, other design criteria would
be added, which would tip the decision in the other direction. Not the least of these
criteria would be to avoid invasion of the intramedullary canal due to the serious
consequences of infection, loosening, or femoral fracture.
Simplified Design Process
From the various steps which are necessary to progress from the design criteria to a
final product, a basic model of a design process can be specified and conveniently
described with a flow chart. The first step is to define the need or the problem to be
solved. This is the first creative step in design, perceiving something that is amena-
ble to a solution. In a holistic framework, the need can be defined as “a treatment for
osteoarthritis of the knee,” because this would allow for a range of solutions includ-
ing nonsurgical. On the other hand, if a company wanted a product in the already-
established artificial knee market, the need would become “an artificial knee to treat
osteoarthritis,” which would limit the solution to different configurations of artifi-
cial knees. In a field such as this, where there are already a number of design solu-
tions on the market, the need can be defined more narrowly. For example, how did
the meniscal bearing and rotating platform designs originate? For the Oxford
Formulate Select Final

Specify Design
Define the Need Multiple Solution based
Criteria
Solutions on Design Criteria
Fig. 3.3 A basic model for the design process, progressing from a definition of the need to the
final solution
unicondylar knee, the need was for a bearing which would be unconstrained ante-
rior-posterior to be consistent with the four-bar mechanism of knee motion, but
which would have low contact stresses on the plastic (it might seem that the two
requirements are incompatible, but a solution was found); for the low contact stress
(LCS) rotating platform, the need was for an artificial knee which would have con-
forming bearings to reduce wear, but would not be constrained in internal-external
rotation. To give more recent examples, the vitamin-E infused highly cross-linked
polyethylene was based on a need for a material which would have exceptional wear
resistance but which would not oxidize in the long term. The instrumented tibial
trials used for balancing at surgery were based on the need for a soft tissue balanc-
ing technique where the surgeon could obtain quantitative values with the femoral
and tibial trials in place. Such narrow definitions of the need are both creative and
valuable in that they can initiate a completely new design or way of doing
something.
Having established the need, the next step is to specify design criteria. These are
intended to provide guidelines for the proposed solutions and are the means for
comparison. The design criteria must be framed in a way that is useful both for
formulating and for evaluating design solutions. Assuming a company assembled a
group of experts in different fields as a design team, it is likely that the design crite-
ria would fall into four categories: function, durability, surgical, and commercial
viability. It is not sufficient to state that function must be “normal,” because the level
and type of function, as well as the kinematics, vary for every individual. Instead,
some type of stability criterion needs to be defined and whether this should vary for
each separate case. Durability relates to strength of the components, the fixation to
the bones, and the wear, each needing specific criteria. The surgical aspects relate to
the alignments, which require the actual values and the allowable tolerances. Soft
tissue balancing is another factor which requires definitions and tolerances. The
time factors and ease of surgery also need to be specified in some way. Commercial
factors include the costs of the implants and inventory, as well as the instruments,
and whether there are identifiable advantages over other designs or systems.
The next step is to formulate multiple design solutions. To illustrate how this
might work in practice, assume that a need is defined as “to enable a person to tra-
verse along a pavement or roadway much faster than walking.” The solution could
be a bicycle, a scooter, a skateboard, roller skates, a wheelchair, a Segway, a hover-
board, a motorcycle, an electric unicycle, or a jet-powered pack, and there will be
others. How would it be decided which of these was worth further consideration and
which should be eliminated? Assume that two of the design criteria were that the
solution had to be self-powered and easy to transport when not in use. This would
eliminate all powered vehicles, while the solution would have to be foldable and
lightweight. In practice, given the design criteria, it would have been unnecessary to
put forward powered vehicles at all, but at the initial stage, it is better to be broader.
For example, if a battery-powered device could be easily recharged by body power
rather than by plugging it in, it might be acceptable.
It can be asked if there is any systematic method of formulating design solutions.
At one extreme it can be regarded as a creative process that is more evident in cer-
tain individuals than in others. Such an ability can even be seen at an early age
where some children can draw imaginative pictures, make complicated Lego con-
structions, or create intriguing stories. At the other extreme, it can be regarded as a
process which is amenable to a logical or systematic process. For example, the bear-
ings which nature has evolved, the history of bearings in engineering, and the artifi-
cial knees already designed could be systematically searched. Most of the new ideas
of artificial knee design and technique have originated from individuals who intro-
duced new insights or concepts, without much explanation of how the design was
arrived at. Undoubtedly some of these ideas were made possible by first carrying
out experimental studies on the knee itself, for example, to determine the mecha-
nisms of stability, the function of ligaments, the effects of muscle actions, the
motion patterns, and so on. Carrying out such studies provides a perception of the
knee as both a structure and a mechanism. An obvious way to proceed is to imitate
the knee itself, a “biomimetic” approach. Biomimetics (or biomimicry) can be
looked at in two ways. Firstly, to reproduce the actual structural forms, to make the
shape of the artificial knee components exactly the same as anatomic and, if any
ligaments or menisci are deficient, to replace those also. This raises the question as
to how exactly to mimic the normal anatomy. Does it involve shape only, or are
mechanical properties included, such as modulus of elasticity, tensile strength, wear
resistance, and friction coefficient? These factors would be dealt with later in the
design process, where each solution would be subject to an evaluation of its practi-
cality. A cautionary note is that a project should not be sidetracked in solving a
problem such as developing a new material. A second approach to biomimetics is to
say that the shape of the biological structures per se does not need to be reproduced,
but the solution must perform the same function. The simplest solution for a knee
joint is a uniaxial hinge, which provides the most predominant motion, and this is
certainly an acceptable solution for some knee conditions. However, for arthritic
knees in general, a more sophisticated solution is needed which would include the
laxity and the stability behavior. In this case, the cruciate ligaments may not need to
be retained if the bearing surfaces provide the same function.
Having formulated several design solutions, the next step is to compare the solu-
tions to determine which most closely satisfies the criteria. It is useful to subdivide
the design criteria into “musts and wants.” For example, the solution must be self-
powered (a must), but it has to be lightweight (a want) and low cost (a want). Hence,
solutions which do not satisfy the “musts” are eliminated immediately, and the com-
parison of the remaining solutions is based on how well they satisfy the “wants.”
This can be accomplished by assigning scores to the different “wants,” using a

ranking between 1 and 10. The scores for the “wants” can also reflect their impor-
tance, such that weight could have a maximum score of 5, but cost a maximum of
10. During the analysis of each solution, some solutions could be modified to
increase their score. At the end of this process, the solution with the highest score
can be selected and the others eliminated.
How best can the different steps in the design process be achieved? The early
knees were usually designed by a single individual or by a surgeon working with an
engineer. This served the need of formulating a completely new design. Today,
when the need is usually to improve an existing system, a team of individuals with
different specialties is more appropriate. Whatever the case, there are certain aspects
of any new project where creative ideas are required, either in the artificial knee
itself or in the instrumentation and technique. The most efficient way to accomplish
that is for the larger group to define the need and the design criteria, but a small
subset of individuals assigned for the design itself. These individuals would be cre-
ative and motivated and would require days or weeks to reach satisfactory
solutions.
In this simplified model of the design process, analysis and testing are not
included. However these are critical steps in order to quantitatively evaluate and
compare each design. There is often confusion between designing something and
evaluating it. It is often considered that when the final design is reached, it should
then be subjected to a set of tests to verify its safety and effectiveness under service
conditions. Instead, it is more appropriate to regard the testing as a part of the design
process itself, such that a design can only be finalized after the evaluations have
been carried out. Today, with the availability of analytic software such as finite ele-
ment analysis for stress determination, and dynamic simulation for kinetics, designs
can be rigorously analyzed in a number of ways, including optimization. However
even after such a process, the final design would still need to be manufactured and
physically tested to account for materials and manufacturing factors. Such testing is
still regarded as falling within the design process because the results of the test may
require or suggest changes to the design itself.
Controlled Convergence Design Model
With the analysis and testing embodied within the design process, the simplified
model can now be extended to what has been termed a controlled convergence
model (Pugh 1991). Pugh’s model was derived from many years of industrial design
experience of both individual components and large assemblies. Later in his career,
Pugh became a Professor at Strathclyde University in Glasgow, Scotland, which
allowed him to disseminate his design knowledge further. Pugh recognized that
after initial brainstorming to produce several design solutions, after working on
these solutions for a period of time, new solutions would frequently emerge. Why
these solutions did not emerge initially is an interesting question. Part of the answer
is that once the designers start to make prototypes and analyze initial designs, it
CONTROLLED CONVERGENCE MODEL
Reject some
Specify Formulate Solutions
Define the Preliminary
Design Multiple
Need Evaluation
Criteria Solutions Add New or
Modified
In-depth Select Final Clinical Refine Final Final

Evaluation Solution Trials Solution Solution
Fig. 3.4 A model for the design process based on the controlled convergence model from Stuart
Pugh. The final three steps concerning a clinical trial have been added to make the model appli-
cable to an artificial knee or any implantable medical device
Fig. 3.5 Stuart Pugh had a

varied career in industry
and academia. In his
faculty position at the
University of Strathclyde,
Scotland, he expounded his
ideas on the design process
in a book called Total
Design: Integrated
Methods for Successful
Product Engineering,
published in 1990
stimulates the mind to generate yet newer ideas. The end result is that at each step
in the design process, the number of solutions under consideration will reduce
because some ideas will be found to be inferior, but other ideas will be added. So
long as the number of the former is greater than the latter, eventually only a single
solution will remain, hence the term used by Pugh, controlled convergence. This
gradual process is fueled by the analysis and testing which is carried out at the dif-
ferent stages.
Analysis and testing typically start from simple mechanical calculations, to anal-
ysis using software packages, to testing using purpose-built simulating machines. In
the case of artificial knees, shape and size will be formulated using databases of
bone shapes from MRI or CT scans. Stress analyses of the contact areas on the bear-
ing surfaces and on the underlying trabecular bone will be performed using finite
element analysis (FEA). Kinematics will be assessed using test machines or soft-
ware. Fatigue strength will be investigated with cyclic load machines. Wear and
friction will be measured with elaborate multichannel knee simulating machines,
wear being a process that is materials dependent and which cannot yet be predicted
by computer models.
However all of these analyses and tests are but models of reality. No amount of
testing in a computer or on laboratory test machines can completely predict the
performance in the patient. Today this statement has to be interpreted in the context
of the 50 years of experience of artificial knees in patients, with a wide range of
design types, surgical techniques, material variations, and patient activities.
Nonetheless, whenever any new aspect of a design, material, or technique is intro-
duced, there are unknown and unpredictable outcomes that might occur. The design
changes can be initiated to correct some known deficiency or to take advantage of a
new idea or technology.
When a design is considered complete, it needs to be verified by a further exten-
sion of the design process, namely, field trials. Most engineering products are tested
in this way before final release to the market. Automobiles are tested in all types of
weather and road surfaces, on the open road and in cities, with slow and aggressive
driving, and with different load distributions. Yet even after such testing, and correc-
tive measures taken to remedy problems, many recalls are reported in the media, for
deficiencies relating to accelerator pedals, air bags, braking systems, cruise con-
trols, and so on. The reason why so many of these problems were not detected when
the vehicles were field tested is that they only later occurred on a very small percent-
age of the total vehicles sold or under an unusual set of conditions. There is a paral-
lel situation with artificial knees today. Considering that around 900,000 are used
each year in the USA alone, it would be surprising if rare and unpredicted modes of
failure did not occur. Despite product deficiencies occurring even after field trials,
there would be even more deficiencies if a product was released to the market with-
out carrying out such field trials. Hence, it has to be accepted that no matter how
rigorously a product is designed and tested, there still exists the possibility of prob-
lems in service.
The controlled convergence model explicitly includes field trials, or clinical tri-
als, as applicable to medical devices. Because of the multifaceted nature of the
entire design process, each stage requires input from a group of individuals of dif-
ferent s pecialties and abilities. In the case of artificial knees as well as with most
other medical devices, there are a number of different facets of the design, in which
case the design process can be applied to each facet separately. For artificial knees,
these facets include:
• The femoral component: the shape of the bearing surfaces and patella flange
• The tibial component: the shape of the bearing surfaces and the means of fixation
to the tibial bone
• Interaction between the femoral and tibial components: contact area, laxity, and
stability
• The materials for the components: fatigue strength, friction, and wear
• The surgical instruments for preparing the bone to fit the components: ease of
use and accuracy
• The surgical technique itself: alignment, accuracy of fit of components, soft tis-
sue balancing, and fixation methodology
For each facet, specific testing methods can be used. For example, an orthopedic
company was recently developing a multi-size system of components. For the tibial
component, the optimal size and shape of the tibial fixation peg were required. One
criterion was that the peg would not impact the posterior cortex of the tibia on inser-
tion. A large database of tibial models derived from CT scans was used. In the
computer, the proximal tibial cut was made, and the best size of tibial component
was selected based on the peripheral fit. Variations in the posterior tibial slope and
the depth of cut were tested. The end result was a peg shape which would not
impinge except in rare conditions. Embodied in the analyses was a statistical model.
In recently introduced artificial knee systems, such bone databases have also been
used to define the dimensional parameters for the multiple components.
Over time, a set of tests has been defined as a final check on the safety and effi-
cacy of artificial knees. As a result of the accumulated experience of artificial knee
design over the past several decades, testing methods have been specified by the
American Society for Testing and Materials (ASTM) and the International Standards
Organization (ISO). In the USA, these tests are required by the Food and Drug
Administration (FDA) prior to using new designs in patients, while similar schemes
are used in other countries. Over time, new standards emerge, while others are
updated or abandoned. Whether or not all of the tests have been carried out prior to
design finalization, they are mandatory for a design to proceed to clinical applica-
tion. At this time, formal clinical trials are not required by the FDA for devices that
closely resemble prior clinically used devices, a 510 K designation. However it is
common practice for manufacturers to restrict the use of a new product to a limited
group of surgeons to detect if there are any deficiencies, prior to general release.
This precautionary step has obvious benefits: it avoids a large number of patients at
risk, and if a design modification was necessary, it would be expensive to implement
if there was a large inventory in the field. To the credit of the field of artificial knees,
over the past decade or more, apart from some instances of inconsistent fixation of
certain tibial component designs, there have been no major clinical problems with
artificial knees.
Design Process for the Early Artificial Knees
What was the design process used for producing the earliest knee designs? From
the late 1960s, when artificial knees were first designed, Charles Townley MD,
from Port Huron in the USA, had for many years been designing and using metal
resurfacing components for the femur and the tibia (Townley 1988). He also made
drawings of complete artificial knee designs, but these were not implemented due
to the lack of suitable materials. Jim Pritchett reported that Townley showed his
ideas to John Charnley in the early 1960s (Pritchett 2009). This may have been a
cue for Charnley to assign Gunston a short time later as a visiting Fellow to work
on the knee. Townley did not have the benefit of working in a university setting
where he might have found collaborators and technical support, but rather he
worked mainly by himself and made his own models. He did specify design criteria
for his designs however, the most important being to restore the anatomy as closely
as possible, to remove as little bone as possible, to preserve all ligaments, and to
ensure correct limb alignment at surgery. His design fulfilled all of these criteria,
and by the late 1960s, when suitable materials became available, he was able to
interest DePuy in producing his knee system, the Townley Anatomic knee. As well
as carefully developing his design over many years, other attributes led to its suc-
cess. Prior to his surgical training, Townley had spent a year at an engineering
school. When he turned to orthopedics, he carefully studied the anatomy and
mechanics of the knee and evaluated several ideas of implant designs for treating
arthritis. He made numerous models and carried out surgical testing on knee speci-
mens. He was a highly accomplished surgeon able to achieve accurate component
placement with only basic instruments (Pritchett 2009). It can be concluded that
Townley took a biomimetic approach to artificial knee design, reproducing the
geometry of the knee joint and the alignment as closely as possible, in order to
obtain normal functional results.
The second example of artificial knee design in the late 1960s was the work of
Frank Gunston, a trainee surgeon from Winnipeg, Canada, who was a Fellow with
John Charnley at Wrightington Hospital in England. He was assigned to solve the
“difficult problem of the painful and unstable knee in rheumatoid polyarthritis” by
applying “the biomedical principles and experience gained from total hip arthro-
plasty combined with an analysis of normal knee movement” (Gunston 1971). It is
notable that Gunston did not seem to have applied the work on knee anatomy and
mechanics carried out in Europe up to that time, for the reason explained in Chap.
2. Background, the inaccessibility of this work in the English literature. Instead
Gunston made some basic observations of knee motion from radiographs and con-
cluded that there was both rolling and sliding at the condylar surfaces, calling this
polycentric motion, which gave the name to his implant design. Gunston’s design
criteria were to allow for this motion, to be able to correct the level of the bearing
surfaces on each condyle, to provide rigid fixation of the components, and to avoid
invasion of the patellofemoral bearing surfaces. Charnley, having initiated the proj-
ect and encouraged the work as it proceeded, did not become involved in a major
way due to his involvement with his hip designs. Gunston worked mainly with the
technicians in the Bioengineering Laboratory where all of the development work on
Charnley’s hip was still taking place. Another design criterion which later emerged
was ease of manufacture, as it was decided to make Gunston’s knee components in
the laboratory itself. This criterion was accomplished by designing components
using stainless steel and high-density polyethylene, the same materials used in the
total hip, in shapes which could readily be made using basic workshop equipment.
No doubt minor modifications were made to the design and instrumentation based
on implanting in specimens and on initial surgical cases. The Polycentric knee
development was a result of a unique combination of factors: Charnley’s develop-
ment of the total hip, in that materials for the components and fixation were appli-
cable to the knee, access to a Bioengineering Laboratory, and the skills and full-time
dedication of Gunston who was educated in both engineering and orthopedic sur-
gery. There was an interesting sequel to the design of the Polycentric knee. On
Gunston’s return to Winnipeg, Canada, in 1969, Lowell Peterson and Richard
Bryan, from the Mayo Clinic in Minnesota, visited Gunston and asked if they could
use his design. Gunston did not have any manufacturing organized, but he agreed
that the Mayo Clinic could pursue the design on their own. They accordingly inter-
ested Howmedica in producing it, but first, Peterson, Bryan, and also Mark Coventry
carried out design modifications, notably increasing the contact area of the compo-
nents for improved fixation and wear resistance. Subsequently, these surgeons car-
ried out well over a 1000 cases. More details will be given in Chap. 9.
There was a very different situation for the development of the Freeman-Swanson
knee in London. In the interview the author held with Alan Swanson in 2017,
Swanson explained the circumstances of the collaboration. In the mid-1960s,
Michael Freeman was a recently appointed Registrar at the London Hospital and,
having seen the success of the artificial hip, became interested in developing an
artificial knee. However he realized that technical expertise would be required for
which he had the idea of forming a biomechanics unit at Imperial College, University
of London. This was an audacious idea considering his junior status at the time, but
the end result was that the Professor of Mechanical Engineering introduced Freeman
to Alan Swanson, a young lecturer at the time, and gave them permission to start a
biomechanics unit to work on knee joints. The process started with a definition of
the need: restoration of knees with severe arthritis to normal function. There is an
important contrast here with Townley’s approach which was to treat knees with
arthritis severe enough to require surgery but not with the severe deformities that
were described by Freeman. The reason was that in London and in Europe as a
whole, metallic hinges were being widely used to treat knees with severe deformi-
ties, and the idea was to develop a more conservative resurfacing type of artificial
knee to use instead of hinges. Freeman and Swanson formulated a set of design
criteria which were based on engineering principles as well as on the surgical
requirements. The principles included large surface areas for transmission of forces
to the resected bone surfaces, as well as between the metal and plastic components
themselves to minimize the wear of the plastic. Stability requirements were based
on femoral-tibial conformity, but with a spinout limit to serve the dual purpose of
setting a limit to the axial torque that could be transmitted while still allowing some
rotational freedom. Freeman and Swanson held regular meetings, in between which
models were made and ideas tested theoretically and experimentally. By carefully
defining the design criteria, they were able to limit their design solutions to one
design form with only a few minor variations.
For testing the design, they made fixtures to measure the strength of the fixation
of the tibial component to tibial bone specimens and concluded that the torque to
cause dissociation was much higher than would be encountered between the foot
and the ground under any realistic conditions. They also built a cyclic motion
machine to demonstrate that wear and deformation of the polyethylene were exceed-
ingly small, of a similar amount to that seen in artificial hip joints. The first implan-
tation in a patient was in 1970, after which the usage was expanded. In some cases,
there was side-to-side instability between the femoral and tibial components, a
degree of freedom which was unconstrained in their design. In other cases, the
patella did not track smoothly across the anterior femur component as the knee was
flexed and extended, but bumped over the projecting anterior part of the femoral
component. Determining certain deficiencies in the clinical trials resulted in changes
in the design which made the subsequent performance much more successful.
Hence Freeman and Swanson followed a design process closely, including carrying
out both laboratory evaluations and clinical trials.
The above designs involved only one or two principal designers, but with assis-
tance from technical staff at hospitals, universities, and companies. Mention should
be made of the Kodama-Yamamoto knee, designed by two orthopedic surgeons
from Matsuyama, Japan. This knee was developed in the late 1960s although details
of their design process are not available. Also in the late 1960s, there was a notable
example where a team approach was initiated for the design of the Geometric knee.
The team consisted of engineers who initiated the design, a group of surgeons
whose role was to evaluate and modify the design and develop the surgical tech-
nique, and manufacturers to produce the components and the instrumentation. The
details will be described in Chap. 9.
Progress in Design After 1970
While the late 1960s saw the emergence of only a few designs, from the early 1970s,
there was a steady expansion. Starting mainly in England, design and manufactur-
ing quickly spread to the USA, Europe, and Japan. Established orthopedic compa-
nies saw this new opportunity to expand their business. Many large companies
developed several different knee designs at the same time and would compete to
attract the most prominent surgeon designers. Many surgeons, often with the assis-
tance of engineers, were motivated to produce their own total knee design.
Orthopedic associations, such as the British Orthopaedic Association (BOA) and
the American Academy of Orthopaedic Surgeons (AAOS), started to include

sessions on total knees in their programs, while special conferences were organized
by prominent institutions and also by companies.
One of the most important such meetings was arranged by the Institution of
Mechanical Engineers (IMechE) and the British Orthopaedic Association (BOA),
held in Birdcage Walk in London in September 1974. A wide range of early designs,
most with short-term clinical results, was presented. Some of these designs were
close replicas of the anatomic joint itself, while others embodied ingenious mechan-
ical features to guide the motion and provide stability. Also, biomechanical data was
presented of the forces in the knee, the principles of stability, and testing method-
ologies. Some designs with a sound rationale showed early clinical success, while
other designs were already showing problems. At this conference, there was a bat-
tleground between the proponents of the well-established hinges and the “new-
fangled” condylar replacements. Mr. Shiers FRCS was a distinguished surgeon who
designed a cemented hinge in 1959 and used it on cases with very severe deformity,
mostly with successful results. At the conference, he pronounced that “the knee is a
hinge and the best way to treat a hinge is with another hinge.” In the ensuing years,
while the condylar replacements prevailed, hinged designs still continued to be
used. For example, they played a crucial role at the Royal National Orthopaedic
Hospital in Stanmore for the treatment of bone tumors, as well as for treating arthri-
tis with severe deformities and for revisions of failed artificial knees (Wilson et al.
1974). That hospital was the leading center in the development of implants for
enabling patients with bone tumors to walk after removal of long bone segments
including joints. The design criteria for those hinges included durable fixation,
especially because many of the patients were in their teens and 20s. Hinged design
has evolved to this day, most designs now including rotational freedom to improve
function and reduce component loosening. Their provision of immediate stability,
correct alignment determined by the stems, and ease and speed of surgery have been
their major advantages. The incidence of infection has been substantially reduced
by improvements in operating room practices.
During the 1970s and beyond, some artificial knee designs were modified, some
were withdrawn, and new designs emerged. The driving force behind this progres-
sion was the clinical experience itself but, just as importantly, the collective experi-
ence, ideas, and interaction between the different designers. For example, if a
particular design was found to have component loosening problems, observation of
other designs without such problems would result in modification by reducing the
femoral-tibial constraint or by adding more substantial fixation pegs. Patella pain
would prompt the design of a patella flange on the femoral component and a plastic
patella resurfacing component. Entrapment of cement particles in the bearing sur-
faces, which accelerated wear and increased friction, led to improved cementing
techniques and cement containment pockets in the components. The differences
between designs were largely based on the severity of the arthritis, to differences in
“philosophy,” or simply due to alternate design configurations performing equally
well up to that time. It was apparent that there were a number of design solutions
which would fulfill the need and satisfy a set of design criteria. Hence while an
engineering design process may have been followed in most cases, either explicitly
or implicitly, an additional factor was the collective experience, interaction, and
insights of particular designers.
It is also important to recognize that many new design features originated as a
result of experimental studies. While a design of a total knee can be seen as a coher-
ent model, it is composed of separate features or facets as noted previously. These
facets can relate to the component designs themselves, to the materials, to the
method of fixation, and to the instruments and surgical technique. The following
chapters will deal with the different facets individually. In an exchange with Maury
Hull PhD from University of California, Davis, he considered that the design pro-
cess can be thought of as an “interactive evolutionary process” in that the interac-
tions of the designers, new concepts or design features from particular designers,
basic and applied research, and the accumulation of clinical results, led to a steady
progress of the designs over time. By 1980 there had been a convergence in the
shapes and design configurations. Posterior cruciate retaining (CR) and posterior
stabilized (PS) were most widely used. There was also a following for unicompart-
mental knees, ACL + PCL retaining, and rotating platform. By and large, following
a design process had led to this stage.
Fig. 3.6 A design meeting for the NexGen Knee System. Left to right: Peter Walker PhD, Design
Consultant, John Insall MD, and Jorge Galante MD. Typical of the 1990s era, a large company
would set up a design team of surgeons and engineers, together with their own staff. The team
would be divided into several subgroups, as shown here, for dealing with specific design types.
(Picture taken c 1990) (© Peter S Walker)
This chapter will be concluded with a brief description of robust design, a design
concept formulated by Genichi Taguchi, a Japanese quality engineer (Taguchi and
Clausing 1990). Taguchi’s focus was how to improve the quality and reliability of
products by determining the optimal design and manufacturing parameters so that
the products would be of the highest quality with fewer problems during service
conditions. His statistical methodologies for testing were widely applied in the
automobile industry and led to the reputation of Japanese cars for their reliability
and smooth performance. Applying the concepts to artificial knees, the design itself
and the tolerances in the manufacture of both the components and instruments
would lead to more accurate alignment, less loosening, and reduced wear. The per-
formance would not be adversely affected by deficiencies of technique or by the
type of activities of the patient.
As an example, in a condylar design with double-dished partially conforming
bearing surfaces, the contacts will still be close to the “bottom of the dish,” and
there will still be adequate anterior-posterior and rotary stability, even if there is
some misalignment. In contrast, if the tibial surface is shallow or even flat in the
anterior-posterior direction, if the component is sloped too posteriorly on the prox-
imal tibia, there could be excessive posterior sliding of the contact points. If the
femoral and tibial bearing surfaces are close to being flat in the frontal plane, if
there is lift-off of one condyle due to a high varus moment, the edge of the femoral
condyle will dig into the plastic surface causing damage. In addition there could be
unstable medial or lateral sliding. Yet another example is the provision of a central
peg for fixation of the tibial component. If the cement has adequate penetration
over the entire upper surface of the tibia, the component would be well fixed.
However if substantial areas of the surface became debonded, overall loosening
would occur. In that case, a central peg would act as an insurance in case of such
surface loosening.
as a Systematic Design Process Been Applied to Artificial

H
Knees?
In most cases, a systematic process has been followed, in that the design criteria
have been specified, tests carried out, and alternate configurations considered, with
the design gradually reaching a conclusion until clinical trials were started. However
the ideas for specific design features have evolved gradually over time, due to the
results obtained and the interactions between those working in the field. The end
result was a convergence of designs up to about 1980, when only two types of
designs dominated, with a following for a few other designs. Since then, the design
process had led to other design forms with the goals of optimizing functional per-
formance, often using computer modeling for the analysis. Further advances can be
made by specifying even more stringent design criteria, based on rigorous functional
outcome measures.
References
Brand RA, Mont MA, Manring MM. Biographical sketch: themistocles gluck. Clin Orthop Relat
Res. 2011;469:1525–7.
Cross N. Engineering design methods. London: Wiley; 2008.
Gluck T. Referat über die durch das Moderne Chirurgische Experiment gewonnen positive
Resultate (Report on the positive results obtained by the modern surgical experiment). Arch
fklin Chir. 1891;41:187–239.
Gunston FH. Polycentric knee arthroplasty. Prosthetic simulation of normal knee movement. J
Bone Joint Surg. 1971;53(2):272–7.
Pritchett J. Obituary: Charles O. Townley. Clin Orthop Relat Res. 2009;67(1):308–9.
Pugh S. Total design: integrated methods for successful product engineering. Wokingham: Publ
Addison Wesley; 1991.
Taguchi G, Clausing D. Robust quality. Harv Bus Rev. 1990:65–75.
Townley CO. Total knee arthroplasty. A personal retrospective and prospective review. Clin Orthop
Relat Res. 1988;236:8–22.
Walker PS, Shoji H. Development of a stabilizing knee prosthesis employing physiological prin-
ciples. Clin Orthop Relat Res. 1973;94:222–33.
Wilson JN, Lettin AWF, Scales JT. 20 years of evolution of the Stanmore hinged total knee replace-
ment. Conference of total knee replacement, institution of mechanical engineers, London,
16–18 Sept 1974.
Laxity and Stability
How Has Laxity and Stability Been
4
Incorporated into the Designs?
The principal motion of the knee is flexion-extension about a transverse axis, but
because the femur and tibia are held together by extensible ligaments and the carti-
lage bearing surfaces are deformable, the knee behaves as a flexible hinge. This
means that displacements and rotations can take place in all directions and at all
angles of flexion, representing the laxity of the knee. Laxity is a fundamental prop-
erty of the knee which has a number of functions and consequences. When the
forces on the knee are small, the femur and tibia can be relatively positioned within
a large envelope giving the limb maneuverability, useful in confined spaces. In func-
tion, laxity allows time for muscles to activate and stabilize the external forces
before a stiff boundary is reached. If the knee was a rigid hinge, sudden external
forces would cause very high impact forces against the anatomic structures resisting
the motion.
Laxity is most commonly defined as the displacements and rotations which occur
between the femoral condyles and the tibial plateaus when forces are applied. For
example, in a seated position with the knee at 90-degree flexion, the foot can be
twisted internally or externally, causing axial rotation of the tibia relative to the
femur. If the upper part of the shank is pushed and pulled, there will be posterior and
anterior displacement of the tibia relative to the femur. If the knee is in a normal
relaxed state, that is termed a neutral position. If a force is applied to the tibia, laxity
will easily occur, but as the force is increased, there will be less and less laxity until
an end point is reached. The stiffness of the knee has steadily increased up to that
end point. The laxity can be measured by various means and related to the applied
forces. To describe the laxity quantitatively, the usual method is to define an axis
system in the proximal tibia and distal femur. In the tibia the axes will be the same
as the anatomic axis, with anterior-posterior and medial-lateral axes being mutually
perpendicular. Axes are similarly set up in the femur. However the anatomic axis
along the center of the femoral shaft is oblique to a line across the distal femoral
condyles. For that reason, the vertical femoral axis is usually defined as a line
between the center of the femoral head and the center of a line across the distal

52 4 Laxity and Stability
Rotational laxity
Ap laxity
LAT MED
Internal
laxity
neutral
Anterior External
Posterior
laxity laxity
laxity
Force
DISPLT
Laxity: low stiffness around
neutral, high stiffness at extremes
Fig. 4.1 Anterior-posterior (AP) and rotational laxity of the knee. AP laxity is produced by shear
forces, rotational laxity by torques. The central region about the neutral position has a low stiffness
where only small forces or torques are required to produce movements. As the femur moves away
from the neutral position, the stiffness becomes progressively higher. The laxities reduce when
there is a compressive force acting across the knee
femoral condyles. The displacements and rotations can then be described of the axis
system of the femur relative to the axis system in the tibia, the latter being regarded
as fixed.
Stability has no rigorous definition in that it is a clinical description rather than
a measurable quantity. The knee joint is considered to be stable when the dis-
placements and rotations during activity are within a normal boundary without
any damage to the ligaments or menisci due to excessive elongation. If the dis-
placements and rotations are increased further, soft tissue injury will occur. Due
to the injury, the knee will then be unstable, able to displace beyond the normal
boundaries even without excessive forces being applied. The three most important
instabilities are anterior-posterior displacement, internal-external rotation, and
varus-valgus rotation. These are caused by excessive stretching or tearing of one
or more of the cruciate ligaments, collateral ligaments, and posterior capsule. In
sporting activities, varus-valgus instability can be caused when there is a high
enough force applied at one side of the knee. For example, an impact force on the
outside of the knee will apply a valgus moment, causing lift-off of the medial
condyles and, if the force is sufficiently high, a tear of the medial collateral liga-
ment. If the valgus moment is accompanied by external rotation of the femur on
the tibia, the anterior cruciate can be torn also. In the absence of any injuries to the
ligaments, the knee will usually be stable to all types of loading within the normal
laxity boundaries.
Another path to instability of the knee is when the cartilage bearing surfaces of
the femur and tibia become worn, as occurs in arthritis. Normally the cartilage
4 Laxity and Stability 53
layers are about 3 mm thick on each surface. Even if only half of the layers become
worn, this represents 3 mm of looseness of the main ligaments around the joint,
which can cause instability. In anterior-posterior and rotation, the instability can
be counteracted by the additional dishing due to the wear being restricted to spe-
cific areas. Up to the late 1960s when condylar replacement artificial knees were
first designed, there had been little quantitative research on laxity and stability,
although the concepts themselves were well recognized in sports medicine. In
contrast there had been considerable attention given to the motion of the knee and
the restraining role of the ligaments as far back as the nineteenth century
(Pinskerova et al. 2003).
Remarkable work on ligament function was carried out by Brantigan and Voshell
in 1941. In this work, they used beams and pulleys to apply known forces and
moments and then measured the resulting laxities. Then in the mid-1960s, Hallen
and Lindahl (1966) reported their work on the rotational behavior of the knee and
what structures controlled it. As the knee was extended, the difference between
maximum internal and external rotation was measured as torques were applied.
They realized that it was difficult to define specifically what was internal and what
was external due to the ambiguous neutral position. However they did find that the
arc of rotational laxity moved progressively into internal rotation of the femur of the
tibia as the knee reached extension. This was called the screw-home phenomenon.
Hence by the time the first condylar knee designs emerged, there were the begin-
nings of an understanding of laxity and stability, shown to be a consequence of the
ligaments, the shape of the bearing surfaces, and the menisci which provided a
cushioning effect at the extremes of motion.
The designers of the prevailing hinged joints had no use for laxity, and deliber-
ately avoided it. Their first priorities were stability and provision of flexion-
extension. Even in designing the first condylar replacement total knee, Freeman and
Swanson (1972, 1973) concentrated on providing stability and gave relatively little
consideration to laxity. To provide stability, they decided to completely bypass the
role of the cruciate ligaments in providing anterior-posterior and rotational stability,
instead using the close conformity of the bearing surfaces. They took the dramatic
step of making the bearing surfaces of the components provide all the necessary
stability in both weight-bearing and non-weight-bearing conditions. Their decision
to resect both of the cruciates was primarily for ease of surgery, an important design
criterion. Cruciate resection was a radical step which had not been contemplated
previously.
Knees with severe varus or valgus deformities were their main focus in designing
an artificial knee for which correction of deformity using bone resections was easier
with only the collateral ligaments present. The roller-in-trough configuration
allowed for only flexion-extension motion about a single transverse axis, similar to
the motion of a hinge. However Freeman and Swanson limited the depth of the
trough such that some spin-out of the roller would occur when high torques were
applied between the femur and tibia. The idea was to prevent excessive torques from
being applied to the tibial fixation, but it had the dual effect of allowing for some
rotational laxity. Another feature of the bearing surfaces was that in full extension,
All
fibers
taut
Posterior
band
taut
Antero-
medial
band
Major taut
portion
taut
Posterior
cruciate Anterior
cruciate
Fig. 4.2 From the examination of 44 knees, ligaments were shown to consist of bundles which
were loose or taut at different flexion angles. With the knees mounted in an apparatus to control the
motions and forces, serial sectioning of ligaments was used to determine their relative function
(Girgis & Marshall 1975). (Reprinted with permission from Clinical Orthopaedics & Related
Research)
a projection from the anterior of the femoral component articulated with a flattened
region at the anterior of the tibial surface. The effect was to provide an anterior pivot
point which would limit further extension by tensing the posterior soft tissues. This
mechanism is similar to the action of the medial side of the anatomic knee, where
the upward slope of the tibial bearing surface is augmented by the anterior horn of
the medial meniscus.
A totally different approach to providing laxity and stability in a total knee was
conceived by Gunston (1971). The idea was that an artificial knee could be designed
using the same technologies that Charnley had developed for the hip. “The biome-
chanical principles and experience gained from total hip replacement arthroplasty
were combined with an analysis of normal knee replacement.” Many of the knee
cases were rheumatoid arthritis with cartilage wear but without major deformities
which likely accounts for the minimalist anatomic approach taken by Gunston. His
idea was to resurface the femoral and tibial condyles where the sagittal contours
were a close match to the original bearing surfaces. The means by which this was
achieved on the femur was by an ingenious component design in the form of discs
which were fitted in slots cut in the condyles. This gave a direct visualization of the
parallelism of the component and the original bearing surfaces. With this strategy,
the cruciate and collateral ligaments would be at their normal lengths, resulting in
normal laxity and stability for that patient. “Lateral instability (varus-valgus) was
eliminated by adjustment of the relative levels of the prosthetic articular surface.”
There is however some incompatibility of using discs which have a constant radius:
the sagittal profiles of both lateral and medial femoral condyles show a larger radius
distally than posteriorly. Due to this, the anterior of the disc component would be
embedded below the surface, but this would avoid the potential problem of impac-
tion of the patella in high flexion.
It is significant that these two earliest approaches to artificial knee design had
totally different shapes of components and the means by which laxity and stability
were achieved. The Freeman-Swanson design was a mechanistic approach relying
on the components to provide reproducible and predictable mechanics, whereas the
Gunston design was an attempt to preserve the normal laxity and stability character-
istics of each individual knee. This defined the two extremes for the design of arti-
ficial knees. The mechanistic approach was particularly applicable to large
deformities, while the anatomic approach was compatible with only mild-moderate
degeneration and cruciate preservation.
Nevertheless the mechanistic approach could in principle be applied to all cases,
and this idea was carried forwards in a different way by modifying the shape of the
bearing surfaces. This approach started in the early 1970s, based on research studies
on the factors which controlled laxity and stability. Limitations of the prior studies
of laxity and stability of the anatomic knee were that the knees were tested without
axial forces being applied, not representative of functional conditions. Moreover the
previously applied forces to cause displacements and rotations were not measured,
while it was assumed that the knees reached a relatively hard stop at the limits of the
laxity range. In addition, around 1970, concepts of knee motion still focused on the
sagittal plane. As an example, a key paper described the instant centers of rotation
as the knee flexed and extended and showed how these centers changed during the
flexion range (Frankel et al. 1971). The main point of the paper was to show that the
smoothly changing instant centers in a normal knee were disrupted in pathological
conditions. The idea of using instant centers was expanded upon in showing the
anterior-posterior motions of the contact points which occurred, based on
A - P LAXITY
Pneumatic Shear
30° Flexion, Intact
cylinder
applies axial
force to knee P-A 4.0
No load
4.0
A-P
1
/2 BW 1.4
Torque 3.8
1 BW 0.8
2.2
11/2 BW 0.6
2.2
2 BW 0.5
2.1
Fig. 4.3 Studies to measure the effect of an axial compressive force (long narrow arrow) on the
laxities. The apparatus was mounted in an Instron machine which applied torque (curved arrow) or
a shear force (straight vertical arrow). The laxity values reduced considerably with forces up to 2
body weight (Wang and Walker 1974; Hsieh and Walker 1976). The concepts were directly applied
to the design of the Total Condylar knee. (Photo of Hua Hsieh taken in 1974 at the Conference on
Total Knee Replacement, Institution of Mechanical Engineers, London, England) (Figures
reprinted with permission from Journal of Bone & Joint Surgery A) (inset ©Peter S Walker)
radiographs taken at successive flexion angles (Walker and Shoji 1973). This had
led to the proposition that tibial bearing surfaces in a condylar replacement type of
artificial knee should allow some freedom of anterior-posterior motion, whereas the
laxity was not accounted for.
Soon afterwards studies were carried out which showed the effect of axial forces
on the laxity. These studies were carried out at the Hospital for Special Surgery in
New York in the early 1970s (Wang and Walker 1974; Hsieh and Walker 1976).
Knee specimens were fixed in a test rig, and an Instron loading machine was used
to apply cyclic shear forces or torques between the femur and the tibia, while the
displacements and rotations were measured. Tests were carried out with a range of
axial forces applied to represent functional conditions. There were several impor-
tant findings. The laxity was found to be nonlinear, with a very loose region for a
small applied shear force or torque comprising most of the laxity, followed by a
relatively stiff region at the extremes of the laxity. In the loose region, there was
little resistance from ligaments, whereas at the extremes of laxity, the opposite was
the case. The implication was that there was considerable freedom of motion when
the knee was unloaded, whereas at the extremes of motion, the ligaments tighten
and prevent further movements.
When axial loads were applied to simulate function, there were major differ-
ences in the laxity behavior. As the load was increased, the laxity steadily reduced.
When the loads reached those typical of everyday activities, the laxities were
reduced to only a few millimeters of displacement and a few degrees of rotation.
These motions were well within the boundaries of the total laxities which occurred
under small axial loading. This behavior had important implications; the ligaments
themselves could not have been providing the constraint under loaded conditions,
but some other factor was providing constraint due to the axial loads. One possibil-
ity was the friction between the cartilage bearing surfaces, but this was unlikely
because with a friction coefficient of less than 0.01 in a healthy joint, an axial load
of 1000 Newtons would provide a small shear resistance of only 10 Newtons. The
factor considered to be the reason was an “uphill mechanism” caused by the dishing
between the femoral and tibial bearing surfaces: “The mechanism proposed was the
uphill movement of the femur, as the femur and the tibia were displaced or twisted
relative to one another” (Hsieh and Walker 1976). The uphill movement would be
due to the shape of the bearing surfaces and the increase due to the deformation of
the cartilage caused by the loads. The medial meniscus was also considered to
increase the effective dishing of the medial bearing surface in that it was peripher-
ally attached and would be trapped between the femoral and tibial surfaces.
Anatomic factors supporting the mechanism were that anterior displacement of the
femur on the tibia was strongly resisted by the upward anterior slope of the medial
tibial plateau, whereas axial rotation was reduced much less under load because the
femur could pivot on the medial side of the tibia, with the generally convex lateral
tibial plateau providing little resistance.
This uphill mechanism concept was later corroborated by Markolf et al. (1976,
1978). In the first of Markolf’s studies, elaborate fixtures were applied to the knees of
50 living subjects to apply and measure anterior-posterior and varus-valgus laxities.
When the subjects maximally tensed their muscles, laxities were reduced to 25–50% of
the unloaded values. In the second study on specimens: “The stiffening of the knee
(reduction of laxity) with applied joint load….suggests that there is a protective mecha-
nism for ligaments of the knee…. We believe that joint congruency, which is a function
of the contours of the tibial plateaus and femoral condyles, is a major contributing fac-
tor contributing to the stabilizing effects of the joint loading.” In all of these studies, the
constraints applied to the knee during the laxity measurements need to be defined. For
example, if the tibia is fixed and the femur is displaced in an anterior-posterior direc-
tion, if the femoral and tibial condyles are to remain in contact, there must be freedom
for a small amount of varus-valgus of the femur on the tibia. This is because the sur-
faces of the medial and lateral tibial plateaus are not parallel. The same applies to an
internal-external rotation test. This is the concept of coupled laxities which was first
recognized by Piziali et al. (1977). The consequence is that in a particular test, the other
degrees of freedom should not be constrained. This principle applies to the design of all
types of knee testing methods, whether in vivo or using laboratory machines.
Terminal
Force
stiffness
Laxity
Displacement
Neutral
stiffness
Terminal
stiffness
AP force (N) 0 90 45 20
100
Posterior AP displt (mm)

4 mm Anterior
100 N
20 45 90 0
Fig. 4.4 Experimental apparatus used by Keith Markolf PhD to measure the laxities of the knee,
both displacements and rotations. Motions in different planes were not constrained, giving more
accurate laxity values. The authors defined the load-deflection graph with a loose central portion,
and stiff portions at the extremes. The laxities and stiffnesses varied with the angle of flexion
(Markolf et al. 1976). In an in vivo study, Markolf showed that when the subjects tightened their
muscles, there was a 25–50% reduction in the AP laxity (Markolf et al. 1978). Markolf has contin-
ued to produce studies of the role of the knee ligaments in laxity and stability, working at the
UCLA Biomechanics Laboratory in Los Angeles. (Reprinted with permission from Journal of
Bone & Joint Surgery A)
The concept of laxity and stability was the basis of the Total Condylar knee
designed in 1972 by Ranawat, Insall, and Walker and first used in patients in 1973
(Walker et al. 1974). The femoral component was the starting point. In the frontal
and sagittal planes, radii similar to that of an average anatomic femur were chosen.
The tibial radius values, larger than the corresponding femoral radii, were calcu-
lated to provide the required laxity and stability based on the laxity data. The result
was femoral-tibial bearing surfaces which were partially conforming in both the
frontal and sagittal planes. In the frontal plane, there was an added advantage of this
scheme. If there was an applied varus or valgus moment sufficient to cause lift-off
of one of the condyles, there would be a stable pivot point. This would prevent side-
to-side sliding which would be felt as instability by the patient. The femoral
Fig. 4.5 Keith Markolf

PhD started his career in
biomechanics in 1970 with
studies of the spine and
then turned his attention to
the hip and knee joints. In
1976, he published the first
of several important and
high-quality studies on the
laxity and stability of the
knee, work which had a
direct application to
artificial knee design.
Since that time he has
continued to carry out
studies of the
biomechanics of the knee
and has published over 100
papers on the subject. He
currently works at the
UCLA Biomechanics
Laboratory in Los Angeles.
(Reprinted from UCLA
Biomechanics Lab Annual
Newsletter 2014–2015)
component also had an integral patella flange for which a plastic patella dome was
designed (Aglietti et al. 1975). The combination of all of these features provided a
combination of laxity and stability, with the laxity being reduced in weight-bearing
conditions. The “partially conforming double-dished bearing surfaces” substituted
for cruciate function and provided similar stability as if the cruciates were present.
These stabilities included anterior-posterior displacement and internal-external
rotation.
The principles of laxity and stability are embodied in most condylar total knees.
As a generalization, designs where both of the cruciate ligaments are preserved
(ACL + PCL type) usually have shallow tibial bearing surfaces on the basis that the
anterior-posterior stability is provided mainly by the cruciates themselves. Some
designs use shallow dishing or raised lips anteriorly and posteriorly to limit the
extremes of motion, but up to these limits, there is considerable laxity possible
between the femoral and tibial bearing surfaces. When the anterior cruciate is
resected (CR type), the tibial surface usually has moderate dishing to substitute for
the stability normally provided by the anterior cruciate. Ideally, this dishing should
include an enhanced posterior lip to prevent the femur from locating at the posterior
edge of the tibial surface, which would result in instability as well as damage to the
B1
C1 A (see detail below)
B2
M4 D2
M3
M2
M6
L1
D1 M1
T G
A2 A3
A1
Fixture A
Actuator
Fig. 4.6 Apparatus constructed to measure the nonlinear stiffness properties of the knee during
different displacements and rotations. Coupled behavior was noted where anterior-posterior dis-
placement produced internal-external rotation. Hence in testing, care is needed to not constrain the
coupled motions (Piziali et al. 1977). (Reprinted with permission from Journal of Biomechanics)
plastic. For designs with resection of both cruciate ligaments (PS or UC types), even
further dishing is provided on the tibial components in order to limit both anterior
and posterior displacements of the femur on the tibia. In the ultracongruent designs,
the tibial dishing can be extreme, restricting the laxity to only a few millimeters in
the centers of each condyle. For all types of design, the condyles of the femoral
component can be similar, averaging the anatomic values of the lateral and medial
radii. The tibial radii are determined by which ligaments are present as explained
above. Guidelines for the ideal tibial radii in the different design types were later
specified in experiments carried out using a knee simulating machine (Luger et al.
1997). This study defined the ideal radii of the tibial bearing surfaces for the
ACL + PCL, the CR and the PS types of artificial knees.
The posterior stabilized type (PS) is a special case in that the effect of the cam-
post is superimposed on the stability provided by the femoral and tibial bearing
surfaces. Near extension, the cam-post prevents posterior displacement of the femur
on the tibia. In early flexion there is no contact. After about mid-flexion when the
Fig. 4.7 John O’Connor PhD, a Professor of Mechanical Engineering at Oxford University,
England, developed the Oxford meniscal knee, together with John Goodfellow FRCS of the
Nuffield Orthopaedic Hospital. O’Connor published numerous studies of ligament modeling and
knee kinematics, emphasizing the 4-bar linkage concept. The meniscal knee, allowing freedom of
motion, was in tune with the cruciate ligaments which caused posterior displacements of the femur
with flexion. (Photo taken at interview at Oxford 2017) (©Peter S Walker)
posterior surfaces of the cam-post come into contact, anterior displacement of the
femur on the tibia is prevented. Other design intentions of the cam-post are to
increase the range of flexion and the lever arm of the quadriceps. The cam-post does
have the effect of reducing the anterior-posterior and rotational laxities in both
extension and in high flexion.
It can be proposed that if artificial knees are to function in a similar way to the
anatomic knee itself, the laxity and stability characteristics should be similar. The
solution employed for off-the-shelf designs is to choose average values. However
the laxity and stability values in anterior-posterior displacement or internal-external
rotation vary considerably between individuals. To more closely match each patient,
their laxity characteristics would need to be estimated in some way and then suit-
able radii for the bearing surfaces of the tibial component calculated. Alternatively
there could be a simple choice between high or low laxity tibial components.
One important consideration is whether the laxity and stability of the medial
and lateral sides should be different. In Chap. 2 Background, studies carried out
in Europe were referred to which showed that during flexion, the lateral contact
point displaced posteriorly much more than the medial contact point, resulting in
axial rotation with the center of rotation on the medial side. This type of motion
has been referred to as a “medial pivot.” However a medial pivot concept was not
considered for the early artificial knee designs for the two reasons explained; the
European literature was not well known, and at that time knee motion was usually
depicted in the sagittal plane only. Certainly there was recognition of the axial
Fig. 4.8 Andrew Amis PhD

was trained as an engineer at
Rolls-Royce and then gained
a PhD in Bioengineering at
Leeds University, England,
under Duncan Dowson PhD
and Verna Wright MD. In
1979, he joined
the Mechanical Engineering
Department at Imperial
College where he is still
actively involved. During
this time, he has carried out
numerous insightful studies
on the laxity and stability of
the knee and the
biomechanical functioning
of artificial knees. In
addition, he has worked with
numerous surgeons on
practical problems, often
involving the solution of
complex reconstructive
problems. Amis has been a
well-known participant in
the European Federation of
Sports Medicine Societies.
(Photo taken in 2018) (©
Peter S Walker)
rotation which occurred, but there was no consideration of asymmetry between

the lateral and medial sides of the components. Such a scheme was eventually
described in a 1992 patent (Tuke and Freeman 1993). Mike Tuke was an inventive
engineer who joined Freeman and Swanson’s Biomechanics group at Imperial
College London in the 1970s. Tuke and Freeman devised a medial pivot type of
knee replacement, as an advancement of the original Freeman-Swanson design:
“…an improved knee prosthesis which will allow the patient to move his knee
substantially in the same manner as though he had not had to undergo implanta-
tion of a knee prosthesis.” This is an ideal design criterion for an artificial knee
and is the essence of the PROM which has been introduced recently called the
Forgotten Joint Score.
The medial pivot concept for an artificial knee was also being pursued by David
Blaha in the USA. Blaha had been a Fellow with Freeman in the early 1980s. They
had numerous discussions about the asymmetry of the lateral and medial motion,
and did not accept the 4-bar linkage concept of knee motion which was described
by John Goodfellow and John O’Connor at Oxford. Further details of the medial
pivot knee will be described later in Chap. 11 Guided Motion Knees. However in
the context of laxity and stability, the lateral and medial sides behaved quite differ-
ently in terms of the laxity and stability which they provided. The medial femoral
condyle was a spherical surface, seated in a matching surface in the medial tibial
plateau. This allowed no linear laxity but allowed unconstrained rotation about a
vertical axis. In contrast, the lateral tibial plateau allowed freedom of AP motion
without any constraint to rotation about the medial side. Hence the bearing surfaces
would not provide both laxity and stability on the lateral and medial sides as in a
typical condylar replacement knee, but preferential stability on the medial side.
Constraints would still be provided by the collateral ligaments however. In an ana-
tomic knee, the deformability of the medial bearing surfaces and the meniscus and
the elasticity of the ligaments still allow some anterior-posterior displacements to
occur on the medial side, so in that sense the medial side would not act as a fixed
medial pivot, but rather as a “loose medial pivot.” But the many studies on laxity up
to about 1990 did not sufficiently describe the internal mechanics of the knee under
a wide range of loading conditions simulating different activities. Nevertheless
designing an asymmetric artificial knee was a very important step in that it defined
a completely new type of design to the existing symmetric designs. Today, there has
been renewed attention to designs with a differential between the laxities on the
lateral and medial sides.
So far only anterior-posterior and internal-external laxities have been discussed,
because this was the main focus of most of the studies. Markolf also measured
varus-valgus laxity and was the first to point out that a certain moment had to be
applied to cause lift-off of one of the condyles, after which one of the collateral liga-
ments supplemented by the cruciates would control the subsequent laxity. While
anterior-posterior and internal-external are an inherent part of normal knee motion,
varus-valgus is a mode to be avoided, because it indicates that the loading condition
is such that ligament injuries could occur.
The laxity of the knee in axial compression is usually not referred to, yet it may
have significance in that while the joint surfaces undergo deflections of several mil-
limeters during functional loading, the materials used for artificial joints only
experience small deflections. In load-deflection experiments of the anatomic knee,
due to the deformability of the menisci, Krause et al. (1976) found that energy
absorption was considerably reduced on the removal of the menisci due to their
load-bearing role. The load-deflection curve will also depend upon the thickness of
the articular cartilage on both the femur and tibia and the effective modulus of
elasticity during the repetitive loading in walking. Shepherd and Seedhom (1999)
compared the cartilage thickness in the hip, knee, and ankle and found that the
thickness was greatest in the least conforming joint, the knee. Another important
finding was that the modulus of elasticity in compression was higher the thicker
was the cartilage. The thickness was averaged for the femoral and tibial condyles,
the latter divided into the area which was covered or not covered by the menisci.
The mean femoral thickness was 1.7–2.7 mm; for the tibia, the mean thickness
ranged from 1.5 to 3.0 mm. Using MRI and stereophotogrammetry, the cartilage
thicknesses were accurately mapped over the entire surfaces of the femoral and
% load
reduction
65 % energy
reduction
55 11.6
47
38
4.8
0 45 0 45 INT RES
Degrees flexion
Fig. 4.9 Bill Krause carried out experiments showing the vital role of the menisci in attenuating
the forces and in energy absorption during axial loading. Left, two pairs of bars: The percentage
reduction in load and energy absorbed after removal of the menisci, for a given deflection value.
Right pair of bars: The contact stresses for the intact knee before (INT) and after (RES) meniscal
resection. Drawn from tabulated data (Krause et al. 1976)
tibial condyles (Cohen et al. 1999). The mean maximum thickness on the femur
and tibia were as much as 4.0 and 4.5 mm, respectively. In two different studies,
Guaon Li and colleagues from Massachusetts General Hospital in Boston used
fluoroscopy to determine the deformations of the cartilage in living subjects. In a
lunge activity, the cartilage deformation was 22–30% (Bingham et al. 2008; Liu
et al. 2010). During a gait cycle, the percentage deformation had a maximum value
of 23% compared with the resting thickness. When these values are related to the
maximum cartilage thickness of the femur and tibia combined, in the region of
6–8 mm, the linear deformations were about 1.5–2.0 mm. While these are small
amounts, they are significant in relation to the elongations of the cruciate and col-
lateral ligaments whose stiffnesses are in the region of 60 Newtons/mm. In the
resting position of the knee, the pretensions in the collateral ligaments have been
measured at 50 Newtons in each ligament. This means that during an activity cycle,
the ligament tensions could reduce to a small value. In this case the stability of the
knee would rely to a large extent on the geometry of the bearing surfaces and the
menisci, together with the stabilizing effects of the muscles themselves. Another
factor related to the deformation at the knee is the accelerations when the foot
impacts the ground. It was found that the vertical accelerations across the knee
were higher in arthritic and artificial knees, compared with normal knees (Kahn
et al. 2013).
An ideal design criterion for an artificial knee is that the laxity and stability after
the artificial knee has been implanted are the same as when the knee was in its nor-
mal healthy state. If all of the soft tissues are retained, the bearing surfaces of the
prox
Femur
thickness
0.0
med
0.7
1.4
2.1
2.8
3.5
mm
Tibia
thickness
0.0
1.0
2.0
3.0
4.0 med
5.0
6.0
mm ant
Fig. 4.10 The cartilage thickness over the femoral and tibial surfaces was mapped using a stereo-
photogrammetry technique accurate to 0.1 mm. There was considerable variation in thickness with
some indication that the thickest cartilage occurred over areas of most frequent weight-bearing
(Cohen et al. 1999). This data has implications to the deformations which occur during weight-
bearing and hence to the changing lengths of the major ligaments around the knee. Such length
changes will have significant effects on the laxity during weight-bearing phases of activity (Cohen
et al. 1999). (Reprinted with permission from Osteoarthritis & Cartilage)
Medial Lateral
980 Newtons
Unloaded
Dashed = loaded
Fig. 4.11 Knees were loaded, frozen, sectioned, and photographed. After thawing, another pho-
tograph was taken. This figure shows the regions of cartilage deformation and the magnitude.
Deformation of each cartilage layer was approximately 1 mm. The variation in cartilage thickness
can be seen, especially on the tibia. Peter Walker observed this work in progress, on a visit to
Tsunenori Takei’s lab at Shinshu University in about 1978 (Takei 1979). (Reprinted with permis-
sion from Journal of the Japanese Orthopaedic Association)
artificial knee need to substitute for the function of the articular surfaces and the
menisci. The medial side of the knee does provide anterior-posterior stability, par-
ticularly if there is a compressive force acting. Hence in a conventional condylar
knee, there should be some sagittal tibial dishing medially, even if the lateral side is
flat. If one or both cruciate ligaments are resected, the dishing will need to be
increased accordingly. Again, if the medial-lateral asymmetry is to be replicated, the
medial dishing should be greater than the medial. Based on the above criteria, some
ACL + PCL preserving designs may be under-constrained, while in cruciate
resecting designs, most are symmetric, although they will in general still provide
adequate stability. Considering that the muscles in many arthritic patients are
weaker than normal, more than anatomic constraint seems advisable. The constraint
in any particular design can be determined by the ASTM constraint test or by soft-
ware simulating that test.
ow Has Laxity and Stability Been Incorporated into

H
the Designs?
A combination of laxity and stability has been incorporated into artificial knee
designs by the radii of the femoral condyles in the frontal and sagittal view being
smaller than the radii of the tibial surfaces. The closer the radii between the femur
and tibia, the higher the stability and the lower the laxity, and vice versa. For designs
which preserve the cruciate ligaments, there should be relatively large differences in
the radii, to the point where the tibial surfaces are shallow. When the cruciates are
resected, there needs to be smaller differences in the radii, so that the bearing sur-
faces provide the stability which was normally carried by the cruciate ligaments.
20
0 4
Torque (Newton-Meter)
6
8 10
12
14
16
10 20
22
28
00 5 10 15
Axial Rotation from neutral position (deg)
force 10 12 14
1000
Shear or 4 6 8
0 16
Shear force (Newtons)
torque
20
22
500
28
Femoral roller 0
in tibial trough 0 5 10
Displacement from neutral position (mm)
15
Fig. 4.12 Equipment for measuring the laxity of total knees, designed by Walker. Knee geometry
was represented by rollers in troughs with different combinations of sagittal radii. The laxity was
determined by the difference between the femoral and tibial sagittal radii (numbers next to each
curve). A difference of 0 mm indicates complete conformity, for which displacement and rotation
were small. As the shear force or torque were increased, or as the difference in radii increased, the
laxity increased. The values shown in the graphs are for an axial force of 1000 N, about one half
the maximum force in activity. To maintain the laxities to normal levels in the absence of the cruci-
ate ligaments, moderate to high conformity is needed
References
Aglietti P, Insall JN, Walker PS, Trent P. A new patella prosthesis. Design and application. Clin
Orthop Relat Res. 1975;107:175–87.
Bingham JT, Papannagari R, Van der Velde SK, Gross C, Felson DT, Rubash HE, Li G. In vivo carti-
lage contact deformation in the healthy tibiofemoral joint. Rheumatology. 2008;47(11):1622–7.
Brantigan OC, Voshell AF. The mechanics of the ligaments and menisci of the knee joint. J Bone
Joint Surg (Br). 1941;23(1):44–66.
Cohen ZA, McCarthy DM, Kwak SD, Legrand P, Fogarisi F, Ciacchio EJ, Ateshian GA. Knee
cartilage topography, thickness, and contact areas from MRI; in-vitro calibration and in vivo
measurements. Osteoarthritis & Cartilage. 1999;7:96–109.
Frankel VH, Burstein AH, Brooks DB. Biomechanics of internal derangement of the knee.
Pathomechanics as determined by analysis of the instant centers of motion. J Bone Joint Surg
Am. 1971;53(5):945–62.
Freeman MA, Swanson SA, Zahir A. Total replacement of knee using metal polyethylene two-part
prosthesis. Proc R Soc Med. 1972;65(4):374–5.
Freeman MA, Swanson SA, Todd RC. Total replacement of the knee using the Freeman-Swanson
knee prosthesis. Clin Orthop Relat Res. 1973;(94):153–70.
Gunston FH. Polycentric knee arthroplasty. Prosthetic simulation of normal knee movement. The
J Bone Joint Surg (Br). 1971;53(2):272–7.
Hallen LG, Lindahl O. The “screw-home” movement in the knee-joint. Acta Orthop Scand.
1966;37(1):97–106.
Hsieh HH, Walker PS. Stabilizing mechanisms of the loaded and unloaded knee joint. J Bone Joint
Surg Am. 1976;58(1):87–93.
Kahn H, Walker PS, Zuckerman JD, Slover J, Jaffe F, Karia RJ, Kim JH. The potential of
accelerometers in the evaluation of stability of total knee arthroplasty. J Arthroplast. 2013
Mar;28(3):459–62.
Krause WR, Johnson RJ, Wilder DG. Mechanical changes in the knee after meniscectomy. J Bone
Joint Surg Am. 1976;58(5):599–604.
Liu F, Kozanek M, Hosseini A, Van der Velde SK, Gill TJ, Rubash HE, Li G. In vivo tibiofemoral
cartilage deformation during the stance phase of gait. J Biomech. 2010;43:658–65.
Luger E, Sathasivam S, Walker PS. Inherent differences in the laxity and stability between the
intact knee and total knee replacements. Knee. 1997;4(1):7–14.
Markolf KL, Mensch JS, Amstutz HC. Stiffness and laxity in the knee – the contribution of the
supporting structures. J Bone Joint Surg Am. 1976;58(A)(5):583–94.
Markolf KL, Graff-Radford A, Amstutz HC. In vivo knee stability. A quantitative assessment using
an instrumented clinical testing apparatus. J Bone Joint Surg Am. 1978;60(5):664–74.
Pinskerova V, Maquet P, Freeman MA. The anatomic literature relating to the knee from 1836 to
1917: an historic note. Clin Orthop Relat Res. 2003;410:13–8.
Piziali RL, Rastegar JC. Measurement of the non-linear, coupled stiffness characteristics of the
human knee. J Biomech. 1977;10(1):45–51.
Shepherd DET, Seedhom BB. Thickness of human articular cartilage in joints of the lower limb.
Ann Rheum Dis. 1999;58:27–34.
Takei T. Deformation of the articular cartilage and joint space of the human knee joint under static
load. J Japan Orthop Ass. 1979;53:577–93.
Tuke MA, Freeman MAR. ‘Knee Prosthesis’ US Patent No. 5, 219, 362. Filed Feb. 1992, issued
June 15, 1993.
ciples. Clin Orthop Relat Res. 1973;(94):222–33.
Walker PS, Wang CJ, Masse Y. Joint laxity as a criterion for the design of condylar knee prosthe-
ses. Conference on Total Knee Replacement, Institution of Mechanical Engineers, London,
UK, 16–18 September 1974.
Wang CJ, Walker PS. Rotatory laxity of the human knee joint. J Bone Joint Surg Am.
1974;56(1):161–70.
Fixation Using Acrylic Cement
What Are the Basic Principles of Cemented
5
Fixation, and How Durable Is It?
Before the artificial hip and knee were introduced, metal components for resurfac-
ing one side of the joint were being used to treat arthritis. However for resurfacing
both sides of the joint, there were two major challenges: how to fix the components
rigidly to the bones such that the fixation would be durable and identifying materials
with sufficiently low friction and wear. This chapter will describe why acrylic
cement was chosen for fixation and the way in which it should be used at surgery.
Initially the mechanics of the fixation of uncemented components will be discussed
as this led to the recognition of why some type of cementing was considered to be
necessary.
Implants Without Fixation
Hip prostheses such as the Austin Moore stem for replacing an arthritic femoral
head or for treating a femoral neck fracture were widely used in the 1950s and
1960s. Fixation of the component to the femur was by an intramedullary stem press
fit into the canal, sometimes augmented by bone graft. In many cases however, the
stem loosened in the canal due to the axial and torsional forces, causing subsidence
and pain. In the knee, uncemented metal tibial plateaus of the McKeever or
MacIntosh types were frequently used (McKeever 1960; Macintosh and Hunter
1972). Their design rationale was different. McKeever focused on biomechanical
design criteria: “There must be an optimal relation between surface area and the
range of functional stress to be borne by the prosthesis and transmittal to the bone.”
“An endoprosthesis must be self-retaining…..so that the normal forces in the joint
hold it in place.” “The direction of stress transfer between the prosthesis and the
bone must be constant.” “The stress transfer from prosthesis to bone must take place
at a single level.” This referred to a situation where a stem could transfer loads
directly to cortical bone, bypassing trabecular bone which would then experience
“stress protection” and become porotic.

70 5 Fixation Using Acrylic Cement
The emphasis of MacIntosh was: “The aims are to correct the varus or valgus
deformity and restore normal stability by taking up the slack in the collateral liga-
ments.” However these metal tibial plateaus, while finding a useful application for
certain indications, experienced an incidence of sinkage, bone resorption, and loos-
ening with some pain (Scott et al. 1985). Even the addition of anterior screws used
in Townley’s tibial plateau design (Townley 1988, originally published 1964) could
not prevent bone resorption and loosening over time. The likely reason was the
mechanics at the interface between the rigid metal component and the exposed ends
of the trabecular bone across the resected proximal tibial surface. Firstly, there
would not be a precise fit, so that the patches of contact would be overstressed;
secondly, there would be interface micromotions due to the changing compressive
and shear stresses during activities. It was clear that rigidly fixing the implant com-
ponent to the bone where the forces were uniformly distributed was essential. A
possible solution would be to use a grouting material between the implant and the
bone, where the grout would penetrate into the interstices of the trabecular bone.
The criteria for the material were biocompatibility, sufficient mechanical strength,
and practical to use at surgery.
Origin and Use of Acrylic Cement
A solution applied to a surface replacement hip was first used by Edward Haboush
(1953). Haboush was an orthopedic surgeon at the Hospital for Joint Diseases in
New York (now NYU Langone Orthopedic Hospital). Haboush was regarded as being
extrovert and inventive by his colleagues. In his paper, it is clear that he was familiar
with basic engineering principles: “I have also found it highly desirable to secure the
head or ball in position on the femur by means of a quick-setting resinous material,
such as a quick-hardening acrylic resin and also to secure the socket in position in the
pelvic arch, by means of a similar quick-setting resin which serves not only to bond
the elements together but to afford a cushioning action which absorbs and distributes
the shock throughout a greater area of the bone structure and this minimizes wear,
erosion and pain.” At surgery, a powder (polymer) and a liquid (monomer) were mixed
together, and after a few minutes, a doughy stage was reached allowing the cement to
be applied in the joint. After a few more minutes, the cement hardened, fixing the
component to the bone. This setting property of the acrylic cement allowed the sur-
geon enough time to apply the cement and position the components before the cement
hardened. Although Haboush used photoelasticity to demonstrate how cement pro-
duced uniform force transfer, mechanical problems with the hip components limited
further advancement of his work to only a few cases.
It is likely that Charnley obtained the idea of using acrylic cement from a confer-
ence in Copenhagen in 1951 or from Edward Haboush soon after (Waugh 1990).
However, Charnley subsequently carried out a great deal of experimental work and
materials development in conjunction with a manufacturing company, so that the
acrylic could be used practically for hip surgery. One important characteristic of
the acrylic was its biocompatibility in solid form. Charnley explained the role of the
5 Fixation Using Acrylic Cement 71
Thin metal shell in acetabulum

22
s
Metal shell with
collar covering the
head and neck 19 24
23
2 17
20 22
15 12
11
23
18 21
13
16 10
Components
2 Femoral head and
fixed to the 14
neck reamed to a
bones with 12 conical shape
fast-setting
dental
acrylic
cement
F
Optional screws to
enhance fixation (15,16)
Fig. 5.1 Edward Haboush from the Hospital for Joint Diseases in New York (now NYU Langone
Orthopedic Hospital) was first to use fast-setting dental acrylic cement for anchoring components
to the bone, applied to a surface replacement type of artificial hip joint. He designed several ver-
sions of the device and implanted them in patients from 1951. The figure shows his US patent filed
in February 1952 (Haboush 1953). (Reprinted with permission from Bulletin of the Hospital for
Joint Diseases)
cement in providing an accurate cast of the surface of the bone whether in the femo-
ral canal or in the acetabulum. This would transmit stress evenly over all parts of the
interface between the acrylic and cancellous bone. The cement was then used in the
total hip.
Over the next few years, Charnley studied the use of acrylic cement in the hip and
reported his findings in 1970 in a book entitled Acrylic Cement in Orthopaedic
Surgery (Charnley 1970). He prepared histological sections from the hips of deceased
patients, showing the minimal tissue reaction and durability of the interface. He also
emphasized the surgical technique and an understanding of the mechanical princi-
ples: “Failures after using acrylic cement are basically the same as the failures of all
forms of orthopaedic surgery: namely, mechanical failures from an inadequate
understanding of mechanical detail. These failures…. result from fundamental errors
in biomechanical theory underlying these procedures, errors which the concept of
cement eliminates.” An important aspect was reduction of the interface motion to
Fig. 5.2 John Charnley’s

extensive experience and
research on the use of
acrylic cement in the hip
joint set the stage for the
use of cement in the
artificial knee, by
demonstrating that it was
both safe and effective.
One disadvantage of using
cement in the hip was the
high pressures generated in
the femoral canal during
stem insertion, forcing
monomer and fat into the
circulation. However this
did not apply to the knee
where the problem was
often obtaining enough
pressure at the tibial bone
interface to achieve
sufficient cement
penetration. (Photo taken c
1980) (© Peter S Walker)
insignificant values by the interposition of cement between the implant and the tra-
becular bone, the cement interlocking with the trabeculae: “The superficial layer of
the cancellous bone in contact with the cement will move as one with the cement
surface when load is applied. The deflection of the cancellous structure under load
will take place inside the bulk of the cancellous bone. In this way we can explain the
paradox of the transmission of load from a hard to soft substance without relative
motion taking place between the surfaces in contact.” Charnley explained the mecha-
nism of cement intrusion into the cancellous bone of the walls of the femoral canal:
“The insertion of the tapered stem of the prosthesis expands the stiff dough (cement)
and injects it into the cancellous lining.” After the mid-1960s, there were an increas-
ing number of US surgeons who had spent the requisite training time at Charnley’s
hip center and were using the Charnley hip. It was not long before US companies
started to make their own artificial hip designs, and in the early 1970s, knees were to
follow. However, considering the importance of fixation to the success of the devices,
there was a surprising delay in scientific studies on the mechanics of fixation, espe-
cially for the knee. This can largely be explained by the higher priority given to the
study of the systemic effects caused by the monomer of the cement, as well as on the
effects of the heating of the cement during polymerization on the trabecular bone.
10
4 min insertion
Depth of acrylic penetration

6.5 min insertion
1 MM CAVITY SIZE
0
0 1 2 3 4 5 6 7 8
Time (s)
Fig. 5.3 Early fundamental experiments on the depth of penetration of acrylic cement under dif-
ferent conditions. The smaller the time after mixing and the less viscous, the greater the depth of
penetration. The higher the applied pressure, the greater the depth. The smaller the holes, the less
the depth (Markolf and Amstutz 1976). (Reprinted with permission from Clinical Orthopaedics &
Related Research)
At surgery, it was necessary to account for the rheological behavior of the

cement, including the variations in properties due to time, ambient temperature,
and handling. The initial liquidy state changed during only a few minutes from
low viscosity to high viscosity with rubberlike viscoelasticity. Consequently the
timing of the introduction onto the bone surface and the mechanical pressures
applied greatly affected the degree of interdigitation. In controlled experiments
using holes drilled in blocks, the depth of penetration was higher for an earlier
time of introduction, for larger hole diameters, and for larger applied pressure, as
expected (Markolf and Amstutz 1976). But after only 1–2 seconds after pressure
application, the rate of penetration slowed down significantly. This led to the
conclusion that no matter what the introduction time of the cement and placing
the component, the greatest penetration would be obtained by only a few epi-
sodes of vigorous finger packing. These experiments provided general guidelines
for the use of acrylic cement in surgical conditions, although if a specific depth
of penetration was required for a particular component, whether hip or knee, and
for cancellous bone of low or high density, more detailed experimentation would
be needed.
Cement-Bone Interface
While the above work was applied to hip replacement, in the case of the artificial
knee, although their usage expanded in the early 1970s, similar experiments were
not carried out until the late 1970s. By then, it was evident that loosening of the
tibial component was a major problem in many designs even after only 1–2 years.
Loosening was evidenced by radiolucent lines (RLLs) at the interface between the
cement and the bone on the radiographic images. In many of these cases, the RLLs
were restricted to about 1 cm width on the lateral and medial sides and did not prog-
ress in extent or thickness. This was termed a nonprogressive RLL which was not
considered to be a problem. However, in many cases, the RLLs became more exten-
sive over time, even involving the entire interface. This was called a progressive
RLL, a serious problem in that it represented a complete dissociation of the compo-
nent and cement from the underlying bone. Histological examinations indicated that
there had been a lack of cement penetration into the cancellous bone with bone
resorption beneath the cement, replaced by fibrous tissue. Beneath that was a thin
layer of new cortical bone, a subchondral layer.
Part of the explanation for the radiolucent lines was evident from biomechani-
cal experiments carried out at the Hospital for Special Surgery in the early 1970s
although the significance was not fully recognized at the time. The first piece of
evidence was that the load between the femur and the tibia in the normal knee
was transmitted over elliptical-shaped contact areas which were anteriorly
located in early flexion and posteriorly located in high flexion (Walker and Hajek
1972). This would exert a rocking effect on an artificial tibial component, trans-
mitting both tensile and compressive stresses to the underlying trabecular bone.
The density and strength of this bone were found to be highly variable, especially
in arthritic knees where deformity resulted in areas of very low density on the
relatively unloaded side of the knee (Behrens et al. 1974): “The natural trabecu-
lar pattern may have been altered to such an extent that the bone is not suited to
support the forces from insertion of a prosthetic implant the significance of this
work related….to the design and surgical technique….of condylar replacement
knee prosthesis”.
Fig. 5.4 Experiments to determine the micromotion of a tibial component relative to the tibial
surface. Loads were applied (white arrow) to a ball bearing at different AP locations along the top
of the component. Anterior A and posterior P distances were measured to indicate the relative
component-bone motions. For anterior or posterior loading, distraction of the component occurred
at the opposite end to the load, detrimental to fixation (Walker et al. 1976). (Reprinted with permis-
sion from Journal of Biomechanics)
The fixation was investigated experimentally on unicompartmental components

cemented to tibias and loaded in different ways (Walker et al. 1976). Such compo-
nents were currently part of the Unicondylar and Duo-condylar knee systems which
were being used clinically at the time. In the experiments, as an indicator of fixation,
the relative micromotions between the components and the surface of the bone were
measured. For central loading, all-plastic components deflected in a U-shape,
deforming the underlying bone as well as the component. When a metallic compo-
nent representing metal-backing was loaded centrally, there was uniform deflection
into the trabecular bone but of smaller magnitude. For off-center loading, lift-off
between the component and the bone occurred on the end opposite to the loading for
both plastic and metal components. The lift-off was even more pronounced when
the trabecular density varied across the surface. For example, on the medial con-
dyle, the central area of trabecular bone is frequently stiffer than that at the anterior
and posterior. That situation produced a “seesaw” effect on a tibial component
loaded off-center. It was recommended that the surgeon drill cement keyholes in the
bone at the anterior and posterior of the tibial surface to counteract the tilting. It was
also noted that the small surface area of unicompartmental components would also
cause higher stresses on the bone surface with potential compressive failure of tra-
beculae. These negative factors eventually led to the abandonment of separate tibial
components in favor of one-piece components which had a much larger surface
area. The obvious exception was for unicompartmental knees.
Other possible causes were investigated for the development of the RLL seen on
the radiographs of clinical cases. With components of small surface area such as in
the Polycentric knee, and in the early condylar replacements with separate medial
and lateral components, the contact stresses could be abnormally high, causing
mechanical failure of the trabecular bone immediately below the cement. The effect
would be exacerbated if the components were made from polyethylene, because
deflections beneath the loading points would cause stress concentrations beneath.
The reason for this is that for a convex femoral component loaded against a flat
polyethylene component, the contact area is small, only about 10 mm in diameter.
Due to the low modulus of elasticity of the polyethylene and the thickness of the
component, typically 8 mm or even less, the high surface stress transmits to the
underlying trabecular bone with little attenuation. Even if the component had a
metal-backing to increase rigidity, if the loading occurred near the anterior or pos-
terior edges, tilting of the component would occur with high compressive stresses
under the load and tensile stresses at the opposite edge. The tension could cause
separation of the cement from the trabecular bone, especially if the cement was not
well interdigitated.
Another factor which could cause interface failure was the amount of constraint
between the femoral and tibial components. For example, in two of the earliest
designs, the Freeman-Swanson and the Geometric, the tibial bearing surface was in
close conformity with the femoral surface. The rationale was to minimize the con-
tact stresses on the polyethylene surface, thereby reducing the wear. In addition, it
would provide both rotational and anterior-posterior stabilities, an important factor
when most of the knees treated at that time were very unstable. However, a negative
effect was that transmission of the shear and torque forces would cause high stresses
at the cement-bone interface. Even if the patients were not very active, there would
likely be sufficient episodes of twisting and sudden stops and starts, to disrupt the
interface.
Yet a further factor was the varus and valgus moments which could occur in
everyday activities. Even in normal individuals, due to the adduction moment acting
on the knee, the forces on the medial condyle are higher than on the lateral condyle
by a ratio of about 3:2. If the artificial knee components were misaligned by only a
few degrees, the ratio would be higher. In some activities, all of the load could pass
through the medial side, with lift-off occurring on the lateral side. Such uneven
loading would overstress the trabecular bone and cause compressive failure at the
interface on the medial side. This uneven varus or valgus loading condition would
affect certain designs of tibial component more than others. However the worst
effect was thought to be with all-plastic components with relatively thin plastic.
Joe Miller MD, an orthopedic surgeon, and Bill Krause PhD, a biomechanical
engineer, from McGill University in Montreal, Canada, realized that a lack of
understanding of the principles of cement fixation was having a negative effect on
the clinical results of artificial knee replacement (Krause et al. 1982a, b). According
to them there was limited appreciation of the need for cement penetration and of the
methods for achieving that: “acrylic cement applied manually in the usual fashion
to cancellous surfaces results in virtually no penetration of the cement into the inter-
faces between the boney trabeculae, and as a result, no interlock is achieved.” Such
an interface would allow micromotion to occur between the cement and the trabecu-
lar bone. In addition, the load would not be transmitted across sufficient surface area
of trabeculae. In combination, this condition would lead to bone resorption and
radiolucent lines. The solution proposed (Miller et al. 1992) was: “an improved
interlock can be produced by using acrylic cement in a liquid stage, applying it to
the boney surfaces using a pressurization system which induces an actual penetra-
tion of the cement into the interstices of the cancellous bone….This would resist
movement as a result of compression, shear and tensile loading and would result in
a more even transfer of stress from prosthesis to cement to bone”.
Krause prepared multiple specimens where metal plates of constant area were
cemented onto trabecular bone from the proximal tibia. The lower surface of the
plate was grooved to rigidly bond to the cement. Different preparation methods
were used including cleaning and drying the bone, pressurizing the cement as it
was applied, and using the cement in a more liquid or doughy form. The strength
of fixation was defined by the tensile force needed to distract the metal plate with
the cement adherent from the trabecular bone. The failure was characterized by
where the failure occurred, in the cement or in the trabecular bone either directly
below the cement or some distance below. Finger-packing doughy cement on an
untreated tibia resulted in minimal depth of penetration which increased to
2–3 mm if the bone surface was first water lavaged. The lavage removed not only
blood and fat from the bone surface, but also trabecular fragments produced dur-
ing the resection of the tibial surface. There was a further increase in cement
penetration to 5–10 mm when more liquidy cement was pressurized with a
Fig. 5.5 In the late 1970s

and early 1980s, Bill
Krause PhD worked with
Joe Miller MD at McGill
University, Montreal, on
cementing technique for
artificial knees. They
carried out the first
systematic experimentation
on the importance of
cement penetration to
achieve adequate interface
tensile strength and
developed methods for
achieving the required
cement penetration. (Photo
taken on interview at the
ORS/AAOS meeting, New
Orleans, 2018) (© Peter S
Walker)
Fig. 5.6 Different methods for achieving cement penetration into the cancellous bone, using noz-
zles to inject cement into the surface and into small holes drilled into the surface. This resulted in
a mechanically superior interface (Miller et al. 1979). (Reprinted with permission from William R
Krause PhD)
cement gun into a lavaged surface. The tensile strength to pull the cement out of
the trabeculae paralleled the depth of penetration, with the pressurized samples
giving at least twice the strength of the finger-pressed. However “….there was a
great deal of scatter in the results for each series of tests. This scatter resulted
from possible differences in porosity from sample to sample, as well as the loca-
tion of each sample.” The two most important variables affecting fixation strength
were the depth of penetration and the density of the trabecular bone. Both of
these had upper limits however. Beyond 2–3 mm of penetration, the strength did
not increase further. This was explained by observing that cement engagement of
about two levels of transversely oriented trabeculae was sufficient. With very
dense bone, the cement filaments between adjacent trabeculae had only a small
tensile strength. The opposite effect occurred with porotic bone where in this
case it was the thin trabeculae that failed in tension.
In practice there is a disadvantage to cement penetration of more than a few mil-
limeters. Firstly, there would be thermal effects on the encased trabecular bone as
the cement hardened. Secondly, if a revision was needed in the future, extra tibial
resection depth would be needed to reach new trabecular bone. However, trabecular
strength reduces steadily with depth below the surface. In addition, cement which is
liquidy enough to be applied with a syringe would spread over the edges of the bone
during application, which is difficult to remove. Probably for these reasons, applica-
tion of cement using a syringe was not generally practiced. On the other hand, the
importance of achieving several millimeters of cement penetration across the entire
tibial surface has now been recognized by most surgeons. This is usually achieved
by surface lavage, applying the doughy cement in as soft a condition as possible,
and pressurizing manually across the surface. Even so, there is often inadequate
cement penetration at the periphery because of the loss of pressure. This deficiency
is still in need of a practical solution.
Drill holes made in

Ahn. 2014 Intl Orth sclerotic bone regions
Fig. 5.7 For durable fixation of the tibial component, Krause, Krug, and Miller (1982) empha-
sized that 2–3 mm of cement penetration must be achieved across the entire tibial surface. Brushing
and water lavage were the first step. In cases where there was sclerotic bone, such as on the medial
side, small drill holes were made. On application of the cement, pressurization was applied
Fig. 5.8 An anterior

radiograph of an artificial
knee. Radiolucent lines at
the interface indicating
bone resorption and
replacement with fibrous
tissue. This is likely the
result of low cement
penetration at the sides
(green arrows), allowing
cement-bone micromotion
to occur
Back to the 1970s, it was evident that even though interface loosening was a
recognized problem with many designs, more attention was given to femoral-tibial
conformity, and enhancing fixation with large pegs, rather than focusing on cement
penetration and tensile strength of the interface. Nevertheless, for certain designs of
total knee replacement, cement fixation turned out to be adequate, with long-lasting
results even up to 20 years and more. For example, for the AGC knee, using a non-
modular tibial component with molded polyethylene, the survivorship for aseptic
loosening was 95% at a 25-year follow-up (Ritter et al. 2016). For a PS total knee
at a mean of 12 years, there had been no revisions for loosening out of 89 young (<
60 years) active patients (Meftah et al. 2016). A 20-year follow-up of the PFC
knee, designed in the 1980s, showed a survivorship for loosening of 98% (Patil
et al. 2015). A mean 15-year follow-up of a wide range of implant designs and sur-
geons showed 98.8% survivorship for loosening (Vessely et al. 2006). Even a knee
that was designed in 1973, the Total Condylar, for 45 knees in 30 surviving patients
at 20-year follow-up, only 4 had been revised. Radiolucencies were prevalent, which
could be attributed to an inadequate cement technique in that era of the surgeries,
but this had not resulted in any cause for revision (Rodriguez et al. 2001). With such
results, it might be concluded that cemented fixation, even with imperfect tech-
niques in many cases, was an effective method of fixation for the long term, for
tibial components with central pegs. For the femoral component, loosening has sel-
dom been a problem, attributed to the wraparound shape of the component, yet the
penetration depth of cement at follow-up has been found to be surprisingly small
(Howard et al. 2014).
Long-Term Behavior of Cement
Recently, Kenneth Mann PhD and colleagues from the Biomaterials Institute,
University of Syracuse, NY, investigated whether the cement-bone interface main-
tained its integrity in long-term follow-up. The basis for the work was to determine
whether cemented fixation could be considered to be satisfactory in the long term or
whether there was significant degradation over time. Autopsy specimens were
obtained, and innovative methods were developed for studying the interfaces both
morphologically and mechanically. The specimens ranged from 3- to 22-year fol-
low-up, in patients aged 54–90 years. Rectangular prisms 8 mm × 8 mm in section
were cut down into the cement below the metal tibial tray and into the underlying
trabecular bone. Micro-CT was used to make three-dimensional computer models
of the specimens. Different colors showed the trabecular bone, the cement, and the
bone cavities in the cement where the bone has once been located. A major finding
was: “Resorption of the trabecular bone below the tibial tray starts at the cement
border (at the extent of cement penetration into the bone bed) and progresses into
the cement layer. The amount of resorption increases with time in service, and may
be more prevalent in older donors, although the age effect was not statistically sig-
nificant here” (Goodheart et al. 2017). The mechanism for reduced interlock of the
cement in the bone appeared to be resorption of the bone at the boundary with the
cement, with the resorption gradually spreading into the cement over time. In the
early stages, the process would seem not to have a serious effect on the fixation.
However, the more the resorption at the interface, the weaker the resistance to shear
and tension. The resorption process was time related, but the age of the patient had
much less effect. The authors pointed to strain shielding and biological effects as
reasons for the resorption.
Since the interfaces would transmit load from the implant to the trabecular bone
beneath, it would be interesting to determine the deformations of the bone and how
much micromotion there was between the cement and the bone (Miller et al. 2016).
Specimens were placed in fixtures, and a charge-coupled device (CCD) camera
attached to a microscope was used to obtain images during the loading. Apart from
the autopsy specimens, fresh cadaver tibias were used to obtain “time of surgery” lab
specimens. The findings were as follows: “When loaded, there is a small amount of
micromotion between the trabeculae and cement, with a greater degree of micromo-
tion near the cement border compared to deep within the cement layer. With in vivo
service, local trabecular resorption causes gaps between the trabeculae and cement
LP: 85 F, 0 yrs PM: 61 M, 5 yrs
a b
PM: 87 F, 3 yrs PM: 69 F, 11 yrs
c d 1 mm
Fig. 5.9 New sectioning methods were devised for showing the penetration of the cement (brown)
within the trabecular bone (yellow) and the changes over time, on autopsy specimens. Top left a,
prepared in lab. b, c, and d show progressive loss of the embedded trabecular bone in specimens.
More initial interlock at surgery resulted in better interfaces at follow-up with less resorption and
micromotion (Miller et al. 2016). (Reproduced with permission from Journal of Orthopaedic
Research)
surface to become much larger. The larger gaps allow for more micromotion near the
cement border.” The gaps were an average of 14 and 40 microns in the lab and
autopsy specimens, respectively. Under loaded conditions the relative average micro-
motions were 1 and 10 microns. These are small values but could still contribute to
some resorption of the bone and fibrous tissue formation. To obtain further insights,
a finite element model was formulated to determine the bone strains under the differ-
ent interface conditions (Srinivasana et al. 2017). This work was carried out by a
collaboration with Nico Verdonschot PhD, Director of the Orthopaedic Research Lab
at Nijmegen University in the Netherlands. The calculated micromotions were simi-
lar to the experimental data, validating the model. It was found that the trabecular
bone at the interface with the cement was strain-shielded by as much as 35–61%.
This alone could be responsible for the gradual loss of the bone over time.
This research has produced a fascinating description and analysis of the cement-
bone interface of tibial components over an extended time period. Fixation of the
tibial component in early artificial knees was problematic. This situation was gradu-
ally improved by more robust component designs and by better cement technique.
The implant designs from the autopsy samples in the Syracuse study are very simi-
lar to, or the same as, the designs used today. Hence the results of the study can be
taken to represent today’s artificial knees. To what extent the degraded interfaces
affected the function of the donors of the samples is not evident from the papers
referenced. It is conceivable that a degraded interface could eventually affect the
function of an active patient after many years of service, but further research will be
needed to investigate that in detail.
hat Are the Basic Principles of Cemented Fixation, and How

W
Durable Is It?
The goal is to achieve a uniform depth of cement penetration of 2–3 mm around the
entire interface, in order to provide sufficient bonding to minimize micromotions
due to compressive, shear, and distraction forces. To achieve this, the trabecular
bone must be cleaned of fat, blood, and bone debris, prior to applying the cement at
sufficiently low viscosity under pressure. There is evidence of gradual degradation
of the interface in long-term follow-up, but so far there has been no sign of acceler-
ated late loosening. However further evidence is still needed of the integrity of
cemented interfaces in the long term in younger and more active patients.
References
Behrens JC, Walker PS, Shoji H. Variations in strength and structure of cancellous bone at the
knee. J Biomech. 1974;7(3):201–7.
Charnley J. Acrylic cement in orthopaedic surgery. Edinburgh: S. Livingstone Ltd.; 1970.
Goodheart JR, Miller MA, Oest ME, Mann KA. Trabecular resorption patterns of cement-bone
interlock regions in total knee replacements. J Orthop Res. 2017;35(12):2773–80.
Haboush EJ. A new operation for arthroplasty of the hip based on biomechanics, photoelasticity,
fast-setting dental acrylic, and other considerations. Bull Hosp Joint Dis. 1953;14(2):242–77.
Howard KI, Miller MA, Damron TA, Mann KA. The distribution of implant fixation for femoral
components of TKA: a postmortem retrieval study. J Arthroplast. 2014;29(9):1863–70.
Krause WR, Krug W, Miller J. Strength of the cement-bone interface. Clin Orthop Relat Res.
1982a;163:290–9.
Krause WR, Miller J, Ng P. The viscosity of acrylic bone cements. J Biomed Mater Res.
1982b;16(3):219–43.
Macintosh DL, Hunter GA. The use of the hemiarthroplasty prosthesis for advanced osteoarthritis
and rheumatoid arthritis of the knee. J Bone Joint Surg. 1972;54B(2):244–55.
Markolf KL, Amstutz HC. Penetration and flow of acrylic bone cement. Clin Orthop Relat Res.
1976;(121):99–102.
McKeever D. Clin Orthop Rel Res. 1960;18:86–95.
Meftah M, White PB, Ranawat AS, Ranawat CS. Long-term results of total knee arthroplasty in
young and active patients with posterior stabilized design. Knee. 2016;23(2):318–21.
Miller J, Krause WR, Krug WH, Kelebay LC. Low viscosity cement. Clin Orthop Rel Res.
1992;276:4–6.
Miller MA, Goodheart JR, Khechen B, Janssen D, Mann KA. Changes in micro-gaps, micromo-
tion, and trabecular strain from interlocked cement-trabecular bone interfaces in total knee
replacements with in vivo service. J Orthop Res. 2016;34:1019–25.
Patil S, McCauley JC, Pulido P, Colwell CW Jr. How do knee implants perform past the second
decade? Nineteen- to 25-year followup of the press-fit condylar design TKA. Clin Orthop Relat
Res. 2015;473(1):135–40.
Ritter MA, Keating EM, Sueyoshi T, Davis KE, Barrington JW, Emerson RH. Twenty-five-
years and greater, results after nonmodular cemented total knee arthroplasty. J Arthroplast.
2016;31(10):2199–202.
Rodriguez JA, Bhende H, Ranawat CS. Total condylar knee replacement: a 20-year followup
study. Clin Orthop Relat Res. 2001;388:10–7.
Scott RD, Joyce MJ, Ewald FC, Thomas WH. McKeever metallic hemiarthroplasty of the knee in
unicompartmental degenerative arthritis. Long-term clinical follow-up and current indications.
J Bone Joint Surg Am. 1985;67(2):203–7.
Srinivasana P, Miller MA, Verdonschot N, Mann KA, Dennis Janssen D. Strain shielding in tra-
becular bone at the tibial cement-bone interface. J Mech Behav Biomed Mater. 2017;66:181–6.
Townley CO. Articular-plate replacement arthroplasty for the knee joint. Clin Orthop Relat Res.
1964;1988(236):3–7.
Vessely MB, Whaley AL, Harmsen WS, Schleck CD, Berry DJ. The Chitranjan Ranawat award:
long-term survivorship and failure modes of 1000 cemented condylar total knee arthroplasties.
Clin Orthop Relat Res. 2006;452:28–34.
Walker PS, Hajek JV. The load-bearing area in the knee joint. J Biomech. 1972;5(6):581–9.
Walker PS, Ranawat C, Insall J. Fixation of the tibial components of condylar replacement knee
prostheses. J Biomech. 1976;9(4):269–75.
Waugh W. John Charnley the man and the hip. Chapter 11: Springer; 1990.
Uncemented Fixation
Is Uncemented Fixation Reliable
6
Enough to Replace Cemented?
In the early development of the artificial knee, fixation of the components to the
bone using acrylic cement was generally considered to be the way forward. Its suc-
cess with total hips had already been demonstrated, and it was assumed that it would
be equally successful in the knee. Furthermore, press-fit components for the knee
were not seen to be viable based on the previous experience, while cement was
becoming available. Nevertheless, a number of drawbacks of cement were being
recognized. There were thermal effects on the adjacent bone due to the heat gener-
ated during polymerization. Systemic effects, due to fat and monomer entering the
bloodstream caused by cement pressurization during component placement, were
considered a serious risk. In many cases, the bonding of the cement to the trabecular
bone was weak due to inadequate cement penetration. Localized areas of interface
failure were frequently observed on radiographs. Despite these factors, in some of
the early designs the fixation seemed to be satisfactory, at least in the short term. At
the same time, there were negative signs which indicated that fixation might not be
as durable as for the hip. The main reason was considered to be the complex forces
on the tibial component which could result in cement-bone micromotion or failure
of the underlying trabecular bone. As a result, an alternative method of fixation was
proposed. The idea was that the components would have a porous surface into which
the bone would grow, providing a rigid and long-lasting bond.
Early Experience with Porous Ingrowth
At Clemson University, South Carolina, during the 1960s, Sam Hulbert PhD
founded a Bioengineering Department and in 1969 started an annual conference
in biomaterials which eventually led to the formation of the International Society
for Biomaterials. Hulbert and his research associates started a program on porous
ceramics (Klawitter and Hulbert 1971): “Using present day techniques (includ-
ing cementing) it is usually impossible to achieve a method of permanent
mechanical fixation which will have more than a small amount of the strength of

86 6 Uncemented Fixation
Fig. 6.1 Bone ingrowth

into the porous surface of
aluminum oxide material.
Ceramic
A cylinder had been
implanted into the cortex
of a canine femur for
4 months (Klawitter and
Hulbert 1971). (Reprinted
with permission from
Journal of Biomedical
Materials Research) Bone
50 microns
the intact bone.” “One possible solution to the problem of skeletal attachment is
to provide a porous material into which bone can grow and thus create a material
capable of direct tissue attachment”. For that purpose they focused on new meth-
odologies for obtaining bone ingrowth for orthopedic implants. They chose a
calcium aluminate ceramic material and fabricated small cylindrical pellets with
different pore sizes ranging from 50 to 200 microns. For comparison, the diam-
eter of a human hair is about 180 microns, only a fifth of a millimeter. The
ceramic pellets were pressed into holes which were drilled into the mid-shafts of
canine femora, for up to 22 weeks. Elaborate methods were developed for study-
ing the tissue which infiltrated the pores in the ceramic. Using a diamond saw,
sections of only 75 microns in thickness were cut and ground. The sections were
studied using microradiography, electron microprobe analysis, and light micros-
copy. The skill and originality in developing these techniques at that time cannot
be underestimated. It was found that: “Under the conditions of the experiment,
pore interactions greater than one hundred microns (0.1 mm) were necessary to
obtain mineralized bone ingrowth.” For the ceramic samples with the highest
porosity, depths of bone ingrowth up to 1.5 mm were obtained.
This work was featured at annual conferences held at Clemson University in
the early 1970s. The meetings were presided over by the charismatic Sam
Hulbert, supported by the scientific faculty including Jerome Klawitter, Barry
Sauer, and Alan Weinstein. Considerable excitement was generated during these
meetings, porous ceramic materials being predicted to be on the verge of revolu-
tionizing the fixation of joint replacements. For practical application, Klawitter
patented an idea for a unicondylar tibial plateau consisting of an aluminum oxide
ceramic component with dense material at the upper part and porous material at
the base for bone ingrowth (Klawitter and Bhatti 1977). The idea could have
been applied to a complete tibial component covering the entire upper tibia.
However, it was not feasible to manufacture the femoral component in the same
way because there was no experience with ceramic-on-ceramic as a bearing
material in the knee with partially conforming bearing surfaces. The alternate
was to bond a thin coating of porous ceramic to a metallic femoral component
6 Uncemented Fixation 87
Fig. 6.2 The idea of a 10

tibial plateau composed of 12
a solid alumina ceramic
bearing surface integral 16
with a porous ceramic 14
base, which interfaced with
trabecular bone, was 18
patented in 1977 by
Klawitter and Bhatti. (US
Patent number
4000525A. Filed 1975,
issued 1977)
Fibula
Tibia
and a tibial tray component. This presented the technical difficulty of producing
a bond which would not dissociate over time. In addition, a thin layer could frag-
ment due to the high cyclic loads during activities. Due to these intractable dif-
ficulties, the research group shifted their attention to dental implants where solid
implants with selected areas of porous coating could be fabricated. Also, Sam
Hulbert left Clemson University in 1973 to be Dean of Engineering at Tulane
University, while other members of the group also took up new positions shortly
after Sam Hulbert left. But the pioneering work of this group set the scientific
standards for other laboratories investigating porous coatings for the fixation of
knee joint replacements.
At the same time as the Clemson University work began, Jorge Galante MD, an
orthopedic surgeon at Rush Presbyterian Hospital in Chicago, and Bill Rostoker PhD,
a Professor of Metallurgy at the University of Illinois, formed a collaboration to develop
the idea of using pads made from thin titanium fibers compressed together and sintered.
The surface of the pads consisted of irregular and elongated openings into which bone
could readily grow (Galante et al. 1971). If a pad was pressed into a cavity in a solid
implant component made from either titanium alloy or cobalt-chrome alloy, it could be
rigidly attached to the component by sintering. An important advantage of this porous
material was that it had sufficient compliance and strength to withstand cyclic loading
while maintaining mechanical integrity: “The material could not fail by crack propaga-
tion and exhibited a compliance in the same range as that of trabecular bone” (Lembert
6 mm
Fig. 6.3 The Miller-Galante knee, designed by Joe Miller, Jorge Galante, and Tom Andriacchi.
The fiber-metal porous pads were made from kinked thin titanium wires sintered together and then
sintered to the femoral and tibial surfaces. In animal experiments and in the application to bone
tumor prostheses, bone readily invaded the porous structure. The fiber diameter was approximately
0.2 mm, and the pore sizes 0.2–0.4 mm. The fiber-metal pads were then applied to a modified IB
knee and the Miller-Galante MG knee (shown here). The tibial component (not shown) had three
pegs to provide initial stability for bone ingrowth to occur
et al. 1972). In animal experiments using a femoral stem prosthesis coated with the
porous material, mean pore sizes up to 390 microns were investigated. In most areas of
the stem, bone ingrowth occurred without an interposed layer of fibrous tissue between
the bone and the fibers. The shear strength of the interface was similar to that when the
stems were fixed using acrylic cement. During subsequent years, further research on
the fiber-metal material was carried out, including application to segmental replace-
ments such as distal femurs and proximal tibias, where extensive bone ingrowth
occurred. In the late 1970s, a total hip stem with fiber-metal pads showed extensive
bone ingrowth in patients (Landon et al. 1986).
Meanwhile, in London, England, Michael Freeman FRCS had decided that
press-fit fixation where the components were directly interfaced to the bone was
viable after all, so long as there were fixation pegs to resist shear and torque forces.
Freeman designed an all-plastic tibial component with grooved pegs laterally and
medially which provided strong resistance to pull-out, as reported to the Orthopaedic
Research Society in 1981. In 1983, Freeman reported on a series of cases performed
since 1977, although not all components were uncemented and there were also
design changes during that time (Freeman et al. 1983). Comparisons between
cemented and uncemented showed similar results, but in many of the cases, there
was a postoperative range of flexion of less than 90 degrees, partially explained by
the preoperative condition of the patients. But despite the ingenuity of the design,
the overall results did not generate sufficient enthusiasm for either the design or the
fixation method, possibly because much of the interest in uncemented fixation had
already switched to porous coatings.
Returning to Chicago, based on over 10 years of laboratory research including
animal trials and experience with a total hip, Jorge Galante, in collaboration with
Fig. 6.4 An uncemented version of the Freeman-Samuelson was used from 1977. Polyethylene
pegs with circumferential groves (known as “magic pegs”) were an interference fit in holes drilled
into the trabecular surfaces. Radiographically, a layer of dense cortical bone formed around the
pegs. No gross loosening was observed. The clinical results were similar to those using cement
(Freeman et al. 1983). (Reprinted with permission Clinical Orthopaedics & Related Research)
Tom Andriacchi PhD and Joe Miller MD from Montreal, designed the Miller-
Galante knee in 1984. Andriacchi’s contribution, backed up by gait analysis studies,
was to produce a tibial surface of low constraint while preserving the cruciate liga-
ments so as to minimize the shear stresses transmitted to the interface between the
tibial component and the bone. By that time it was known that micromotion at the
interface could seriously inhibit the ingrowth of bone and lead to a fibrous tissue
interface, quite the opposite of what was intended. The mechanical fixation of the
tibial component to the resected surface of the tibia was by four porous-coated pegs
approximately 10 mm in diameter and length. However to provide a more rigid con-
nection, four long cancellous screws were passed through the center of each peg
into the underlying cancellous bone. This use of screws was evidently intuitive
because there were no recorded mechanical tests that would provide quantitative
data of reducing micromotion at the interface. It does however indicate they were
aware that in order for ingrowth to occur, a process which could take several weeks,
the porous surface had to interface with the trabecular bone without any relative
motion, whether shear or distraction. Later lab studies did show the substantial
increase in rigidity of the initial fixation provided by the screws (Volz et al. 1988).
While the Miller-Galante (MG) porous-coated knee was used extensively in the
late 1980s and into the 1990s as one of Zimmer’s premier knees, most of the usage
was in a hybrid form, where the femoral component was uncemented and the tibial
component cemented. The rationale was explained in several articles, a notable one
by Kraay et al. in 1991: “Retrieval studies of uncemented tibial components have
demonstrated that very little bone ingrowth actually occur into these implants, chal-
lenging the long-term durability of this method of fixation with both uncemented
and cemented femoral components, however, aseptic loosening has occurred only
infrequently. Recently the hybrid method of fixation, with an uncemented femoral
component and a cemented tibial component has been introduced.” However the
success of the MG knee was dampened because problems emerged with the metal-
backed patella component (Joseph and Kaufman 1990). This was a mechanical fail-
ure resulting in loosening and even fragmentation of the fiber-metal, with fibers
entering the joint itself. This problem adversely affected the follow-up results and
prompted Joseph and Kaufman to state: “This should stimulate research and analy-
sis in the laboratory so that future mistakes in design can be avoided.” While this
was a sound suggestion, artificial knee development was still a new endeavor, and
laboratory methods had not been devised for every facet of the design process.
Certain aspects, in this case the patella design, were perceived as obvious, not
requiring any elaborate testing. The main emphasis was on designing long-lasting
implants, porous coatings being seen as the way to provide that.
Another influential application of porous coating to a total knee was based on the
work of Bob Pilliar PhD and others in the late 1970s and early 1980s (Pilliar 1983).
Pilliar’s idea was to start with a femoral component and a tibial tray made from cast
Co-Cr alloy, the most common material used for total knee components at the time,
and then to apply porous coatings to the surfaces which interfaced with bone. To
produce the porous coating, powdered Co-Cr alloy was used in the form of tiny
spherical beads layered on to the surface of the implant and coated with a liquid
binder to maintain the beads in close proximity. After sintering, small neck regions
of fusion occurred between adjacent beads and between the beads and the solid
substrate. Tests on round beams with and without porous coating showed that the
process and structure of the coating did not affect the fatigue strength of the Co-Cr
specimen, an important consideration in the highly stressed regions of knee compo-
nents. Durability was attributed to the smooth connections between the beads and
the solid structure, avoiding sharp corners which would cause stress concentrations
and weaken the structure. Animal tests using plugs of porous Co-Cr implanted into
the femoral cortex showed excellent bone ingrowth for average pore sizes of 50–400
microns. The higher end of this range was chosen for application to total knees to
allow for the inevitable small micromotions at the bone-implant interface before
bone ingrowth occurred.
By 1983, there had been several years of successful experience of the Co-Cr
porous coating applied to artificial hip stems. Now it was considered the time to
apply the same technology to total knees, but a suitable artificial knee design was
needed. Pilliar had been working on the porous technology with a Canadian branch
of the British Oxygen Company, with a technically oriented sales representative,
Bob Kenna. They connected with David Hungerford MD and Ken Krackow MD,
surgeons at Johns Hopkins University, with the proposition of applying their porous
coating technology to a new knee design. In a very short time, they formulated
original ideas not only for an artificial knee design but also for the instrumentation.
The femoral component was based on the shape of the anatomic femur. The tibial
component had a metal-backing, while the plastic insert was of low conformity to
allow freedom of motion, similar to that of the Miller-Galante knee. The beaded
Co-Cr porous coating was applied to both the femoral and tibial components, as
well as to the patella component. For the instrumentation, Hungerford and Krackow
developed the Universal Instruments which were based on achieving anatomic
alignment of the components. The group met with David Fitzgerald, the CEO of
Howmedica in New Jersey, to discuss his interest in developing this new knee sys-
tem using the porous coating technology. At the time, the company’s main total
knee was the Kinematic Knee system using cemented fixation, designed by Peter
Walker, Director of Product Development, in collaboration with surgeons at the
Robert Breck Brigham Hospital in Boston: Clem Sledge, Fred Ewald, Bob Poss,
and Bill Thomas. Moreover the Kinemax Knee, an advancement of the Kinematic,
was already well under development. However, Howmedica lacked porous coating
technology which was by now being promoted by Zimmer on the Miller-Galante
knee, while other companies were also developing porous coatings. This was the
ideal opportunity for Howmedica.
Even though much preparatory work had already been carried out, production of
a new total knee system was still a formidable task. Time was of the essence because
of the gaining popularity of the Miller-Galante knee. However, David Fitzgerald
sensed that this dynamic and enthusiastic group, with an innovative approach to total
knee design and surgical technique, could make a major impact and even elevate
Fig. 6.5 The PCA knee (left) designed by David Hungerford MD, Ken Krackow MD, and Bob
Kenna. The bone-facing surfaces were coated with sintered cobalt-chrome beads for bone ingrowth.
The polyethylene insert surface was hot-pressed for a precisely contoured smooth surface. The
Universal Instruments were designed for anatomic placement of the components. In the retrieval
specimen (right), bone is seen adherent to most of the femoral surfaces, due to bone ingrowth into
the pores. In the tibia, only about 25% of the surface and the pegs have adherent bone. Fibrous
tissue is invading the pores over the remaining surface
Howmedica above Zimmer in the knee area. An entire mezzanine floor was con-
structed in the Howmedica plant, and prime staff were assigned to accomplish the
knee project in the shortest possible time. Internally, Bob Kenna single-mindedly
pushed forward the work at a rapid pace, dealing with every aspect and using the
technical knowledge he had acquired in his previous work. As well as the porous
coating technology, Kenna developed a new way of forming a smooth contoured
bearing surface on the tibial insert, a process called “hot-pressing.” The idea was to
machine the surface to an approximate shape and then to place the component into a
metal die which was heated up to mold the final surface. The surface was smooth and
shiny, giving the component a hi-tech appearance compared with the usual machined
surfaces of artificial knees at that time. The Universal Instruments was another inno-
vative aspect of the knee system, by now called the PCA knee, the porous-coated
anatomic. Hungerford and Krackow were accomplished surgeons, and mainly due to
Krackow’s mathematical background, they recognized the importance of component
alignments in relation to the applied forces during activity. For example, they pointed
out the anatomic alignment of the joint line in the frontal plane, which meant cutting
the tibia at 2–3 degrees of varus, with the distal femur cut appropriately to maintain
the correct mechanical axis of the leg. This was depicted in a colorful slide showing
that in a normal individual, the joint line was parallel to the ground during activity.
After about a year of clinical trials by Hungerford and Krackow, the use of the
PCA knee was extended to an early user group in the USA and then progressively
expanded. This was achieved by well-organized lectures, starting off in the USA,
but soon including many other countries, facilitated by Hungerford and Krackow’s
fluency in several languages. On his office wall, David Hungerford had a map of the
world with pins indicating the countries where they had given instructional lectures.
Their aim was to leave no country unvisited, a goal that was almost realized! Many
of the lecturing events were supported by other surgeons, who were early users of
the PCA knee. There was also a special group of leading surgeons in Europe, Jean-
Louis Briard being a prominent member. The quality of the presentations was of the
highest level, projecting a new way forward for artificial knee replacement, surpass-
ing the rather traditional designs and surgical techniques.
After introduction of the PCA knee, there were several follow-up studies by the
originators and by independent users in the ensuing years. One article raised an
alarm about loose beads and radiolucent lines mainly around the patella and tibial
components (Cheng and Gross 1988). In another article there was a much higher
incidence of radiolucent lines around the tibial component for an uncemented PCA
compared with the same designs used in cemented mode (Collins et al. 1991).
Subsidence anteriorly or antero-medially was also noted, possibly due to the large
amount of anterior bone resection as a consequence of a horizontal resection plane
rather than with the usual 5 degrees of posterior slope. Fibrous tissue ingrowth
rather than bone ingrowth was noted in revised tibial components, indicative of
inadequate stability of the interface especially in the time period after surgery. A
number of corrective modifications were made to the PCA knee, and a stabilized
type was added. The resulting new system, the Duracon, produced good long-term
results whether used in cemented or uncemented modes.
Fig. 6.6 At the many workshops on the PCA knee around the world, Dave Hungerford and Ken
Krackow involved local surgeons who were early users. This photo (left to right: Jean-Louis
Briard, Marc Raguet, Jean-Louis Prudhon, Gerard Deschamps) was taken in Sicilia in 1982.
(© Peter S Walker)
In 1984, Bob Volz MD, a surgeon from Tucson, Arizona, in collaboration with
Gerald Engh MD and John Moreland MD, introduced the Anatomic Modular Knee
(AMK) with porous coating. Along with several other porous-coated designs at that
time, there were reported cases with inadequate bone ingrowth on the tibial compo-
nent. To address this, Bob Volz devised a new biomechanical testing method to
investigate the initial stability of uncemented tibial components (Volz et al. 1988).
Volz had a strong interest in biomechanics and was particularly interested in improv-
ing tibial fixation which had hitherto been a problem with some uncemented com-
ponents. While porous-coated components were a possible advancement, Volz
recognized that the components must maintain initial rigid fixation before bone
ingrowth took place. The interface micromotions were compared between several
porous-coated designs: his own AMK, the Miller-Galante, the Ortholoc, and the
PCA. The experiments consisted of eccentrically loading the components cyclically
for 300,000 cycles and measuring the subsidence as well as the distraction between
the component and the bone. Volz found that uncemented component stability was
greatly enhanced by peripheral cancellous screws, an important element being pre-
vention of interface distraction on the side of the component opposite to the applied
load. “The ability of a noncemented tibial component to resist micromotion….
8mo.
D.C. p.o.
Fig. 6.7 Frontal and sagittal plane radiographs of an 8-month follow-up of a porous-coated ana-
tomic (PCA) artificial knee. Efforts have been made to take the radiograph exactly along the plane
of the base of the tibial component, such that the interface is not obscured. Close proximity of the
trabecular bone to the porous surface can be visualized (Hungerford and Kenna 1983). (Reprinted
with permission from Clinical Orthopaedics & Related Research)
would appear to be a critical determinant of the success bone might have in invading
a porous surface. The shape of the surgically prepared proximal tibia (flat) leads
little intrinsic stability to a flat metal plate placed on its surface.…Tibial designs
secured with four screws provided the best stability.” This was a vitally important
guideline for other porous-coated designs, as well as the AMK. Shortly after this
work, Walker et al. (1990) developed a method for studying the interface in response
to offset compression, shear and torque loading, using Tekscan to show the interface
micromotions which occurred. The purpose of the study was to determine which
combinations of pegs and keels would provide the least overall micromotion. A
central peg augmented with short pegs laterally and medially, or augmented with
keels, was the most effective. Screws would likely have further reduced the
micromotions.
After there had been a few years of clinical experience of the early unce-
mented designs, Myron Spector wrote a considered appraisal of the status of
component fixation using porous coatings (Spector 1987). While recognizing
that there were early problems with cement, he emphasized that successful bone
ingrowth would need certain conditions for it to succeed, such as initial mechani-
cal stabilization, not easily achieved. Also if porous-coated pegs or stems were
used, they could cause extensive stress-shielding, a very important consideration
at that time when such factors had not been considered. To avoid this, Spector
proposed the use of porous polymeric coatings such as polysulfone or polyethyl-
ene, and although he showed promising results with a hip stem, he never applied
15 Kg load MTS cross beam
Whiteside
To 3 lead
recorder AMK
Miller-Galante
To 3 lead
recorder
PCA
Fig. 6.8 Experimental determination of the stability of fixation of uncemented tibial baseplates. A
cyclic load from 5 to 115 Kg was applied for 300,000 cycles. Subsidence and lift-off were measured
with transducers. All components gradually subsided over time. The maximum lift-off was only 100
microns with the AMK due to 4 cancellous screws (see holes in baseplate) and central stem. The
other designs showed subsidence and lift-off from 200 to 500 microns. This explained the variability
in bone ingrowth, where above 100 microns micromotion could prevent ingrowth from occurring
(Volz et al. 1988). (Reprinted with permission from Clinical Orthopaedics & Related Research)
the material to the knee. In any case, there was evidence that polymeric coatings
did not have sufficient mechanical strength, while bonding to a metal substrate
was problematic.
A number of new designs subsequently emerged which were definite improve-
ments to the first uncemented designs. These included the Natural Knee designed by
Aaron Hofmann MD and Roy Bloebaum PhD, the LCS by Fred Buechel and
Michael Pappas, the AGC by Merrill Ritter, and the Ortholoc by Leo Whiteside. All
of these designs were later reported to have 94–100% survivorship at 10 years fol-
low-up (Meneghini and Hanssen 2008). The natural knee had a superior porous
coating made from cancellous-like titanium (Bloebaum et al. 1992). The LCS may
have benefited from low interface stresses due to the freedom of rotation of the tibial
insert (Buechel and Pappas 1989). In the AGC, the molded polyethylene and the
monoblock construction of the metal-reinforced tibial component would produce

less wear particles, which likely avoided bone lysis at the interfaces after up to 20
years follow-up time (Small et al. 2013). For the Ortholoc, the resected bone surface
was coated with bone slurry from the patient’s trabecular bone generated during the
surgical procedure, and the tibial component was initially stabilized with lag screws
(Whiteside 2001). The use of bone slurry was adopted for other porous-coated
designs. It is interesting that all the uncemented designs except the AMK used sev-
eral short fixation pegs on the tibial component. This is in contrast to the design of
cemented components which in most cases used the successful large central peg.
For the uncemented components, the screws essentially substituted for the central
peg which may not have been as effective in uncemented application.
In the designs where screws were used, radiographically, there was an incidence
of bone lysis around the screws. This was attributed to the accumulation of small
polyethylene particles which had migrated down the holes in the tibial tray. The
wear particles themselves were caused by backside wear between the polyethylene
insert and the base of the metal tray. This led to improvements of the insert-tray
locking mechanism and to polishing of the base of the trays. Looking at the overall
situation, Meneghini and Hanssen concluded that “cementless fixation is probably
the future of knee arthroplasty, particularly in young patients with an active life-
style.” While the main reason for their prediction was that cementless fixation would
be more durable in that group of patients, another motivation for moving in that
direction was the reduced operating time.
Scientific Evaluations
In parallel with the development of uncemented artificial knees, studies were being
conducted on the scientific aspects. The requirement for the close proximity of the
porous surface and the absence of relative motion had been shown to be essential for
bone ingrowth to occur (Martens et al. 1980). Experiments where porous surfaces
were implanted without being subjected to physiological loads, such as porous
plugs inserted into transverse holes in the femoral cortex, would give optimistic data
of the bone ingrowth which occurred, which might not apply to dynamic conditions:
“Consistent bone ingrowth was observed for implants in nonfunctional conditions if
the pore size exceeded 70 microns. Ingrowth in functional loading conditions
requires rigid initial stability of the implant and intimate contact of implant coating
and surrounding bone tissue.” This proposition was supported in later experiments
(Goodman and Aspenberg 1992) where the greater the interface motion, the more
the surrounding tissue became fibrous tissue rather than bone. This work showed
why interfaces developed in different ways, either with bone ingrowth or with
fibrous tissue, depending on the local mechanical conditions.
The above experimental methods can be regarded as preclinical evaluations,
where the intention was to reproduce the mechanical conditions believed to occur
in vivo in order to predict the relative performance of different component
designs, surgical techniques, and bone properties. Ideally, methods such as these
should have been an integral part of the design process for every new unce-
mented artificial knee design. However this was not always the case in the early
designs of uncemented knees, mainly because the equipment for carrying out
such experiments was not generally available and the methods had not been
developed until after clinical trials had been initiated. The main method for eval-
uation of the porous-bone interface was radiography, which could readily be
applied to every case postoperatively. Zonal analysis methods were defined
where the status of the interface regions in frontal and sagittal views was charac-
terized. Bone ingrowth, a fibrous interface, a sclerotic bone line, and local cysts
were features which could be discerned. Bone completely adjacent to the implant
surface was regarded as indicating bone ingrowth. However frequently there
were areas of radiolucency indicating a fibrous tissue interface, with only smaller
areas indicating bone ingrowth. The interface of the medial and lateral edges of
the component, areas with the greatest micromotion, most often showed radiolu-
cency. Progressive radiolucency over time was considered to be indicative of
loosening. This method of interface evaluation became standard and was easy to
perform routinely. It was however subject to some underestimation of radiolu-
cent lines if the x-ray beam was not exactly in line with the surface of a compo-
nent, such that radiolucency could be obscured by metal.
A totally new method for evaluating the fixation of artificial knee components
was devised in Lund, Sweden, in the early 1970s by Goran Selvik PhD, a mathema-
tician, working with Leif Ryd MD and Lars Lidgen MD (Ryd 1986). Selvik’s thesis
Fig. 6.9 Rik Huiskes PhD (left) directed a laboratory in Nijmegen, Netherlands, which developed
advanced iterative finite element computer modelling to predict the bone remodeling around implants
and the effect on fixation. He was Editor of the Journal of Biomechanics from 1979 to 2009, tripling
the journal in size. He was considered the most influential orthopedic biomechanics researcher in
Europe. Leif Ryd MD from Lund in Sweden, starting from the theoretical work of Goran Selvik,
developed the application of RSA to measuring the subsidence of artificial knee components over
time, thus predicting their longevity. The method was applied extensively to determining the efficacy
of uncemented components. RSA is still in use today (Photo c 1995). (© Peter S Walker)
was published in 1974, entitled “Roentgen stereophotogrammetry: a method for the

study of the kinematics of the skeletal system.” The generalized method, called
RSA for short, was lucidly described by Selvik as follows: “If two adjacent objects
each have at least 3 identifiable points which can be identified in any direction of
view, and two radiographs are taken from different directions, it is then possible to
use the coordinates of the 6 points to calculate the position of an axis system defined
in one of the objects relative to an axis system defined in the other object: the posi-
tion being defined by a translation between the origins of the axis systems, and an
ordered rotation of 3 angles to bring one axis system in line with the other. Because
of the divergence of the x-ray beam from a single point and the unknown location
of the points in a global axis system, a calibration process is also necessary where
additional points are included in the field of view as the x-rays are taken of the
objects. In the application of the method to studying the fixation of tibial compo-
nents, usually four (for redundancy and enhanced accuracy) tantalum beads are
embedded in the tibial component itself, while four beads are injected into the tra-
becular bone beneath the component. If biplanar radiographs are taken immediately
after surgery and then at various times of follow-up, it can be determined if the
component has moved relative to the bone, and by how much.”
The movement, or subsidence, of the component into the bone over a time
period could be expressed by the three displacements and three rotations or, in a
simpler way, by the maximum displacement of any of the beads in the component,
called “maximum total point motion” (MTPM). Another metric was “inducible
displacement,” the motion which occurred between unloaded and loaded condi-
tions, the patient at rest and then standing on the knee. During the 1980s many
RSA studies were carried out of cemented, press-fit, and porous-coated artificial
knees at different follow-up times. For cemented tibial components: “A smaller
number of prostheses….showed continuous migration, and had migrated signifi-
cantly more than others after 6 months. An indication can thus be obtained as
early as half a year.…as to whether the prosthesis will migrate continuously or
not.” This last sentence highlights the value of the RSA method, in that new
designs, such as with porous coatings, can be compared with a reference cemented
design, as to whether it had more or less migrating cases. At the time of Ryd’s
articles (1986), the conclusion for currently used porous-coated knees was: “.…
biologic ingrowth type of fixation displayed micromotion, both as migration and
inducible displacement. Stable bony ingrowth was ruled out in most cases.” “The
inducible displacement occurred within the soft tissue of the.…interface.”
“Radiolucent zones may be of both tensile and compressive origin.” This painted
a pessimistic picture of porous ingrowth, at least of the component designs and
types of porous surfaces which were investigated.
A revolutionary change from cement to porous ingrowth was not yet ready to
occur. The main reason appeared to be the inability to prevent micromotion at
the component-bone interface in the weeks after surgery, which would prevent
bone ingrowth from occurring. Instead, fibrous tissue ingrowth and not bone
ingrowth was too prevalent. Nevertheless, this referred mainly to the tibial com-
ponent, whereas in contrast, porous-coated femoral components, although less
studied, were well fixed with few reports of loosening. For that reason, from the
late 1980s, hybrid fixation was commonly adopted, although even then, there
were still many more cemented than uncemented femoral components being
used (Volz and Benjamin 1990; Kraay et al. 1991). Not surprisingly, the quest
to design a successful uncemented knee continued, but this time with the
mechanical testing methods that had been developed and the clinical follow-up
capabilities of RSA.
Since about the year 2000, there have been further advances in porous coating
technologies (Harwin et al. 2017). One example called “trabecular metal” is a tan-
talum structure that looks exactly like trabecular bone. Using this material, tibial
components were fabricated as “monoblocks” where the polyethylene insert was
fused into the porous surfaces of the metal backing. The component had substantial
pegs at the lateral and medial sides to minimize the interface micromotion caused
by varus and valgus moments. Clinical follow-ups showed interfaces with positive
signs of bone ingrowth and the absence of radiolucencies. Some companies are
experimenting with 3D printed metal backings for tibial components. Advantages of
this approach are that almost any structure can be manufactured and the coating
becomes integral with the solid metal substrate. The EOS company is the main driv-
ing force behind this method, with its advanced technology in 3D printing metals
such as cobalt-chrome and titanium. Polyetheretherketone (PEEK), a rigid plastic,
which is already used for the femoral components of a particular total knee design,
is also amenable to 3D printing manufacture and has a modulus nearer to the bone
than either metal or polyethylene.
Fig. 6.10 A porous

structure made by 3D
printing titanium on the
surface of knee
components (Stryker)
compared with trabecular
structure at the proximal
tibia (Saers et al. 2016).
Bone growing in to the
irregular 3D printed
structure is expected to
provide a bond with
minimal micromotion due
to the irregularities of the
structure. The relative scale
of the figures is
approximate. (Reprinted
Stryker and from Journal
of Human Evolution)
Despite impressive clinical results and demonstrable ingrowth of the more recent
designs, Bill Barger MD took a conservative approach, based on the successful and
reliable long-term results of cemented knees: “Until mid-to-long-term data is avail-
able that shows significant benefit in appropriate candidates, widespread adoption
of cementless total knees cannot be scientifically or economically justified” (Bargar
2015). This is a surprising viewpoint taken by the person who first introduced a
robotic system for preparing the bone for implant components. In 1986, Bill Barger,
together with “Hap” Paul, developed Robodoc for accurately shaping the femoral
canal for uncemented hip stems. Recently a new company, Think Surgical, was
formed, which has initiated clinical trials of artificial knee robot surgery in the
USA. Several companies have recently introduced robotic solutions for more accu-
rate knee surgery, applicable to both cemented and porous-coated components. It
might be thought that preparing accurate bone surfaces for porous-coated implants
would result in enhanced ingrowth. This view is supported in the conclusions of
Newman et al. (2018) who stated that: “The newer cementless TKA designs have
shown excellent survivorship, functional outcomes and satisfaction rates in not only
young but elderly populations.” But change may come about slowly due to confi-
dence and familiarity with cement, the higher implant cost of uncemented artificial
knees, and the requirement of a more exacting surgical technique.
I s Uncemented Fixation Reliable Enough to Replace

Cemented?
The two advantages of an uncemented knee are that the surgery would be quicker,
and it will potentially not be subject to the long-term degradation of a cemented
interface. At this time, there is evidence that some uncemented interfaces have over-
come the short-term loosening problems of the earlier interfaces. However before
uncemented can be confidently used on a more routine basis, further conditions are
required: the design of the component and the initial fixation method fixation must
be such that there will be no migration over the first 2 years, as shown by RSA; the
bone ingrowth must be substantial that it is likely to maintain sufficient strength in
the long-term; retrieval data is needed to determine if long-term degradation would
apply in a similar way to cemented interfaces.
References
Bargar WL. CORR Insights(R): the Mark Coventry Award: trabecular metal tibial components
were durable and reliable in primary total knee arthroplasty: a randomized clinical trial. Clin
Orthop Relat Res. 2015;473(1):43–4.
Bloebaum RD, Rubman MH, Hofmann AA. Bone ingrowth into porous-coated tibial components
implanted with autograft bone chips. J Arthroplast. 1992;7(4):483–93.
Buechel FF, Pappas MJ. New Jersey low contact stress knee replacement system. Ten-year evalu-
ation of meniscal bearings. Orthop Clin North Am. 1989;20(2):147–77.
Cheng CL, Gross AE. Loosening of the porous coating in total knee replacement. J Bone Joint
Surg. 1988;70(3):377–81.
Collins DN, Heim SA, Nelson CL, Smith P, 3rd. Porous-coated anatomic total knee arthroplasty.
A prospective analysis comparing cemented and cementless fixation. Clin Orthop Relat Res
1991(267):128-136.
Freeman MA, McLeod HC, Levai JP. Cementless fixation of prosthetic components in total arthro-
plasty of the knee and hip. Clin Orthop Relat Res. 1983;(176):88–94.
Galante J, Rostoker W, Lueck R, Ray RD. Sintered fiber metal composites as a basis for attachment
of implants to bone. J Bone Joint Surg Am. 1971;53(1):101–14.
Goodman S, Aspenberg P. Effect of amplitude of micromotion on bone ingrowth into titanium
chambers implanted in the rabbit tibia. Biomaterials. 1992;13(13):944–8.
Harwin SF, Patel NK, Chughtai M, et al. Outcomes of newer generation cementless total
knee arthroplasty: beaded periapatite-coated vs highly porous titanium-coated implants. J
Arthroplast. 2017;32(7):2156–60.
Hungerford DS, Kenna RV. Preliminary experience with a total knee prosthesis with porous coat-
ing used without cement. Clin Orthop Rel Res. 1983(176):95–107.
Joseph J, Kaufman EE. Preliminary results of Miller-Galante uncemented total knee arthroplasty.
Orthopedics. 1990;13(5):511–6.
Klawitter JJ, Bhatti NA. Ceramic prosthetic implant suitable for a knee joint plateau. US patent
4,000,525. Filed Aug. 21, 1975, Issued Jan 4, 1977.
Klawitter J, Hulbert S. Application of porous ceramics for the attachment of load bearing internal
orthopedic applications. J Biomed Mater Res. 1971;5(6):161–229.
Kraay MJ, Meyers SA, Goldberg VM, Figgie HE 3rd, Conroy PA. “Hybrid” total knee arthroplasty
with the Miller-Galante prosthesis. A prospective clinical and roentgenographic evaluation.
Landon GC, Galante JO, Maley MM. Noncemented total knee arthroplasty. Clin Orthop Relat Res.
1986;(205):49–57.
Lembert E, Galante J, Rostoker W. Fixation of skeletal replacement by fiber metal composites.
Meneghini RM, Hanssen AD. Cementless fixation in total knee arthroplasty: past, present, and
future. J Knee Surg. 2008;21(4):307–14.
Martens M, Ducheyne P, De Meester P, Mulier JC. Skeletal fixation of implants by bone ingrowth
into surface pores. Arch Orthop Trauma Surg. 1980;97(2):111–6.
Newman JM, Sodhi N, Khlopas A, et al. Cementless total knee arthroplasty: a comprehensive
review of the literature. Orthopedics. 2018;41(5):263–73.
Pilliar RM. Powder metal-made orthopedic implants with porous surface for fixation by tissue
ingrowth. Clin Orthop Relat Res. 1983;(176):42–51.
Ryd L. Micromotion in knee arthroplasty. A roentgen stereophotogrammetric analysis of tibial
component fixation. Acta Orthop Scand Suppl. 1986;220:1–80.
Saers JPP, Cazorla-Bak Y, Shaw CN, Stock JT, Ryan TM. Trabecular bone structure variation
throughout the lower limb. J Human Evol. 2016;97:97–108.
Selvik G. Roentgen stereophotogrammetry: a method for the study of the kinematics of the skeletal
system. Reprint from the original 1974 thesis. Acta Orthopaedica Scandinavica Supplementum
no. 232, Vol. 60.1989
Small SR, Ritter MA, Merchun JG, Davis KE, Rogge RD. Changes in tibial bone density mea-
sured from standard radiographs in cemented and uncemented total knee replacements after ten
years’ follow-up. Bone Joint J. 2013;95-b(7):911–6.
Spector M. Historical review of porous-coated implants. J Arthroplast. 1987;2(2):163–77.
Volz RG, Nisbet JK, Lee RW, McMurtry MG. The mechanical stability of various noncemented
tibial components. Clin Orthop Relat Res. 1988;(226):38–42.
Volz RG, Benjamin JB. The current status of total joint replacement. Investig Radiol.
1990;25(1):86–92.
Walker PS, Hsu HP, Zimmerman RA. A comparative study of uncemented tibial components. J
Arthroplast. 1990;5(3):245–53.
Whiteside LA. Long-term followup of the bone-ingrowth Ortholoc knee system without a metal-
backed patella. Clin Orthop Relat Res. 2001;388:77–84.
Femoral Component and Patella
To What Extent Do Femoral Component
7
Shapes Mimic Normal Anatomy?
The forerunner of an anatomically shaped femoral component for the knee was
the mold arthroplasty for the hip, described by Smith-Petersen and Aufranc
(1947). The word “arthroplasty” was defined as “an operative procedure under-
taken for the purpose of creating a joint mechanically suited for lasting func-
tion.” The hip mold was a thin hemispherical shell made from cobalt-chrome
alloy (Vitallium) that was interposed between the reamed surfaces of the femo-
ral head and the acetabulum. It was expected that due to the close conformity of
the surfaces, the sliding action, and the cyclic loads, hyaline articular cartridge
would form on the cancellous bone surfaces. Even though the results in the hip
were not uniformly successful, the concept was applied to the distal femur in
1958 (Hodge 1991). From molds of cadaveric femurs, thin-walled Vitallium
shells were made, with the outer surface being an anatomic replica. Both the
internal and external surfaces were polished so that new cartilage could form on
the femoral and tibial condyles. At surgery, while the distal femur was shaped to
fit the implant, there was no shaping of the proximal tibia. To provide rigid fixa-
tion of the shell, a long intramedullary stem was attached. It was notable that the
shell included the patella surface, an advantageous feature that was not included
on some of the artificial knees designed more than 10 years later. But because
there was only one size of the right and left components, it is likely that there
would be a kinematic mismatch in many cases, explaining the poor ranges of
flexion obtained, while the hard convex surfaces would inevitably overstress and
erode the cartilage and bone on the tibia. Even so, Andrew Hodge MD was justi-
fied in saying in 1991 that “Vitallium-mold arthroplasty of the knee was the
foundation for the current condylar-design femoral components of the modern
total knee arthroplasties.” The component was called the MGH femoral condyle,
referring to the Massachusetts General Hospital in Boston where it was designed.

104 7 Femoral Component and Patella
Fig. 7.1 Vitallium mold

arthroplasty, known as the
MGH femoral condyle.
The left picture shows the
anatomic shape of the
condyles and the
intramedullary stem for
fixation and alignment.
The implant was used from
about 1958. The sagittal
view radiograph shows a
long-term follow-up from
the Brigham and Women’s
Hospital in Boston. The
tibial surface has gradually
molded to the shape of the
femoral condyles.
(Reprinted with permission
from Journal of
Arthroplasty)
Early Femoral Component Designs
By the time of the landmark conference on total knee replacement, held in London
UK in 1974 and organized by the Institution of Mechanical Engineers and the
British Orthopaedic Association, numerous designs of artificial knee replacement
were already in use. Most originated in Europe and the USA but also in Japan and
other countries. At one extreme there were fixed axis metallic hinges with long
intramedullary stems for fixation. The rationale of one of these hinges, the Shiers,
was expressed by the designer as follows: “The object of the exercise is to produce
a painless knee that will bend. The simplest and strongest method of achieving this
object is by replacing the worn or damaged joint surfaces with a hinge.” At the other
extreme were the compartmental designs consisting of a metallic component which
resurfaced the femoral condyles and a plastic component resurfacing the tibial pla-
teau. Such designs could be compartmental or could resurface both the lateral and
medial sides of the knee with one-piece components. The cruciate ligaments could
be preserved or resected.
An intermediate design between the hinge and the compartmental was the
Freeman-Swanson which had been used in patients since 1970. The femoral com-
ponent was characterized by simplicity of design, strength of construction, and ease
of implantation. The straight cut across the distal end of the femur, perpendicular to
the femoral axis, highlighted the requirement that femoral components needed to be
aligned in a specified way relative to the femur. The large surface area of the femoral
7 Femoral Component and Patella 105
ANTERIO
R
POST
ERIO
R
Fig. 7.2 The original Freeman-Swanson artificial knee design, first implanted in 1970 in London,
England. The bearing was a cylindrical surface in a trough to maximize the contact area and mini-
mize the wear. There was no patella flange. The designers introduced the concepts of resecting
both cruciate ligaments to facilitate correct alignment and balancing, make a square cut across the
distal end of the femur to define frontal alignment, and make fitting of the femoral component
much easier. A few years later, a patella flange was added, together with a patella component with
close conformity to the anterior flange. (Reprinted with permission from Institution of Mechanical
Engineers, London UK)
component was intended to minimize the stresses between the component and the
femoral bone and between the femoral and tibial bearing surfaces. The design may
have been initially too simplified in that it did not have a patella groove nor an
extended trochlea for the patella to articulate against in early flexion. This initial
configuration was based on the premise that freedom of patella motion was prefer-
able to constraining the patella in a medial-lateral direction. Medial-lateral displace-
ment was similarly unconstrained between the femoral and tibial components,
which led to instability in some patients.
The first of the compartmental resurfacing knees was designed by Frank
Gunston while he was working as a fellow with John Charnley in Wrightington,
England (Gunston 1971; Gunston and MacKenzie 1976). Gunston graduated in
engineering before becoming a surgeon, thus having the ideal background to
produce an innovative knee replacement. Although Gunston intended to replace
only the femoral and tibial bearing surfaces, the configuration of the femoral
component was by no means intuitive. Rather than replacing the femoral con-
dyles with metal runners covering the entire bearing surfaces, Gunston made a
narrow half disc which fitted into a slot cut into the center of each femoral con-
dyle in the sagittal plane. This had several practical advantages. Firstly, it was
difficult to shape a worn femoral condyle by cutting away a uniform layer of
bone. It was much easier to cut a slot in the condyle using a handsaw. Secondly,
when the component was inserted into the slot in a bed of doughy cement, it
could be positioned so that the bearing surface projected 1–2 mm from the origi-
nal bearing surface which was still preserved on either side of the slot. However,
there would inevitably be some mismatch because in the sagittal plane, the distal
radius of the condyle is larger than the posterior radius on both the lateral and
medial condyles. Since the component was matched more to the smaller poste-
rior radius, the anterior part of the component would be recessed. This would
actually be an advantage in that the patella would then not impinge on the inner
edges of the component. The constant radius also made the component easy to
manufacture, important because the implants were made in Charnley’s
Biomechanical Workshop which was equipped with only basic machine tools.
The hemi-disc- shaped components would be strongly fixed to the condyle
because the large surface area at the sides of the component would efficiently
transfer axial forces in shear. For consistency, a similar design principle was used
for the tibia, where slots were cut into the tibia in an anterior-posterior direction,
although in this case the slot was v-shaped.
A limited number of Polycentric knees were carried out by Gunston at
Wrightington Hospital due to time constraints. In most of the knees, both the lat-
eral and medial sides were replaced, due to the cases being primarily rheumatoid
arthritis. When Gunston returned to Winnipeg in 1969, he chose not to pursue his
design commercially, continuing to make the components locally, but only carry-
ing out a small number of cases. However Lowell Peterson and Richard Bryan
from the Mayo Clinic in Rochester, Minnesota, met with Gunston and were inter-
ested in using the design. Initially they produced Gunston’s design in the Mayo
Clinic workshops but soon started working with companies including Howmedica.
Peterson and Bryan modified the femoral and tibial components by making them
“wide-track.” The part of the femoral component which entered the slot on the
femoral condyle was the same, but the bearing surface was widened, thinking that
this would provide better load transfer onto the femoral condyles. In follow-up
studies of the original design (Gunston and MacKenzie 1976; Lewallen et al. 1984)
and the wide-track design (Bloom and Bryan 1977), there were no fixation or bone
fracture problems with the femoral component. It is likely however that the bone
strength of the femur would have been better preserved with the wide-track ver-
sion, but this was not studied.
Because of the preservation of the normal joint geometry and the cruciate liga-
ments, it might be expected that the postoperative function of the Polycentric knee
would be close to normal. This was investigated by Ed Chao PhD, director of the
Biomechanics Laboratory at the Mayo Clinic, using an elaborate gait analysis
system (Chao and Stauffer 1974). He used electrogoniometers to measure the
rotations at the knee and force plates to measure ground reaction forces. Among
Chao’s goals were to identify the most significant gait parameters on which func-
tional performance of the total knees can be compared and provide objective cri-
teria for evaluating function. Chao found that the motion curves for a Polycentric
knee were close to those of the normal opposite side, though he noted that the
maximum flexion was less than normal, which even applies to artificial knees
today. Ed Chao was undoubtedly a pioneer in the development of the measure-
ment of objective biomechanical functional parameters in the study of artificial
knees and comparison with normal knees.
In parallel with Gunston’s work in the late 1960s, a more conventional com-
partmental design was being produced in Hamburg, called the St. Georg Sledge
(Englebrecht and Zippel 1973). “Under the guidance of the collateral ligaments
the sledge (femoral component) finds a suitable gliding area on the block (tibial
component) keeping full range of movement of the knee joint.…the surgeon.…is
not forced to consider the exact placement of two equally adjusted implants. The
prosthesis can be inserted medially, laterally or bilaterally.” Comparing the metal
hinges with the condylar resurfacing types, as well as being at opposite extremes
in size, bone removal, and canal invasion, the motions provided were completely
different. The hinge allowed only flexion-extension, while the resurfacing knees
allowed anterior-posterior and medial-lateral displacements as well as internal-
external rotation. Surgical placement of a hinge was relatively straightforward
with stability guaranteed, whereas for resurfacing, accurate component place-
ment and balancing of ligaments were essential. This was particularly the case
for the sledge knee because the surface of the tibial component was completely
flat offering no constraint in any direction. The purpose of this was so that the
femoral component can find a suitable sliding area and the surgeon will not be
forced to consider the exact placement. Even with this much freedom of motion,
the average range of flexion was only about 100 degrees in a 4-year follow-up,
the same as for the St. Georg intercondylar hinge. An important element of the
St. Georg design was that the sagittal and frontal radii of the femoral component
were closely matched to the anatomic shapes, based on measurements from 50
knees. The femoral runners themselves were a constant 2–3 mm thickness, with
two pegs for fixation, but the authors did not explain their technique for prepar-
ing the femoral condyles. A 3–10-year follow-up of 53 cases of the sledge knee
used as bicondylars gave additional insights into the performance of the femoral
component (Stockley et al. 1990). There was one case of impingement of the
femoral runner against the medial tibial eminence due to lateral subluxation, and
there was one case where the posterior part of a femoral runner fractured due to
overhang with no cement support. While small in numbers, there are certain
requirements for unicompartmental femoral components used on one or both
condyles. The different lateral and medial geometries require different shapes for
each. Sufficient sizes are necessary. The bearing geometry between the femoral
and tibial components needs to restrict medial-lateral displacement, for which
accurate instrumentation is required. There must be sufficient bending strength
based on the thickness of the runners or reinforcement fins, to carry the func-
tional forces without failure.
As well as the hinges and compartmental knees, at the 1974 London meeting,
numerous other designs were presented, some with ingenious mechanisms. Most of
the designs had similar goals: minimize the amount of bone resection needed yet
still provide sufficient stability and fixation; utilize the collateral ligaments to pro-
vide varus-valgus stability; allow for sufficient laxity for normal function; reduce
the fixation stresses on the components; and provide more inherent stability than
afforded by simple condylar resurfacing types.
The Link arthroplasty invented by Michael Devas (Devas 1974) was a narrow
all-metal hinge where a flat “finger” projecting from the tibia was housed in a circu-
lar cavity embedded in the distal femur: “While requiring no resection of the distal
femoral and proximal tibial condyles, this device still functioned mechanically as a
uniaxial hinge.” It was considered that “an old person when walking neither rotates
the knee nor locks it into full extension….therefore it has not been considered nec-
essary to build in rotation…”.
Another intercondylar design, the Attenborough stabilized gliding prosthesis
(Attenborough 1974), took an opposite approach: “Flexion and extension should
approximate to the normal knee and some rotation and lateral flexion should be pos-
sible….full extension should be completely stable....but these movements should
increase gradually with increasing flexion.…there should be anterior-posterior sta-
bility.” Hence Attenborough was advocating a restoration of the laxity and stability
characteristics of the normal knee. The femoral component was a complete resur-
facing of the condyles and patella trochlea while the tibial component was a one-
piece resurfacing with dished lateral and medial bearing surfaces, partially
conforming with the femoral condyles. Each component had a short but bulky stem
for fixation. In the center there was a loose hinged element which controlled the
motions described above. The design was a middle ground between a hinge and a
condylar replacement in that the laxity did not require long stems, but the intercon-
dylar stability substituted for the resected cruciate ligaments and to some extent
supplemented the collaterals. The design possessed many elements of future suc-
cessful total knees.
The Deane knee (Deane 1974) was another design which reproduced normal
knee laxity, stability in extension, and progressive laxity with flexion. The lack of
varus-valgus constraint allowed for components where intramedullary fixation was
not necessary. This knee actually reversed the materials, the femoral component
being plastic and the tibial being metal, which minimized the total bone resection.
An intercondylar cam produced an instant center of rotation path which changed
with flexion, a concept which was demonstrated by Frankel and Burstein in 1971.
The shape of the cam was to a large extent reproduced in the posterior stabilized or
PS designs produced in the mid- to late 1970s.
A number of condylar replacement designs had also been produced. The femo-
ral bearing surfaces were similar to the anatomic femur, while the tibial surfaces
were generally dished both laterally and medially. This was in contrast to the
anatomic tibia where the lateral condyle is convex in the sagittal plane and where
the medial surface is raised anteriorly but has only shallow concavity centrally.
However the medial concavity can be considered to be augmented by the menis-
cus, which is not the case laterally due to the mobility of the meniscus. Condylar
replacement designs at that time included the Kodama-Yamamoto, the Townley,
the Leeds (Seedhom), the Geometric, the Duo-condylar and Duo-patella, and the
Total Condylar.
Fig. 7.3 The design drawing for the first Total Condylar knee dated 1973. The component fitted
both right and left knees. Pockets for cement were provided distally and posteriorly. This version
had separate tibial plateaus. The one-piece component with a central peg was designed separately.
It is noted that computer-assisted design software did not exist at that time. The design was an
equal group effort between Chit Ranawat, John Insall, and Peter Walker
Anatomically Shaped Femoral Components
In all cases, the femoral component shape was based to a large extent on normal
anatomy. For some of the designs, the sagittal profile was obtained from radio-
graphs or from slicing knee specimens. One exception was the Geometric knee
(Coventry et al. 1972) where the surfaces were spherical, with the same radii in the
frontal and sagittal planes. The reasons were that it would present a uniform geom-
etry to the dished tibial surfaces throughout flexion and that it would be relatively
easy to machine and polish accurately. In Bahaa Seedhom’s Leeds knee design
(Seedhom 1972) however, he replicated anatomic shapes: “.…because there is a
close relation between the geometry, ligamentous attachments and the joint move-
ments, it follows that the prosthesis geometry must not depart too far from that of
the natural knee if we are to preserve the ligaments and cause as little distortion of
Fig. 7.4 Bahaa Seedhom

PhD with a casting of a
femoral component of the
anatomically shaped Leeds
knee which he designed in
the early 1970s. In his
position at University of
Leeds, England, he hosted
numerous Japanese
fellows, who carried out a
succession of innovative
biomechanical studies on
the knee. Seedhom
founded a company
making artificial ligaments,
and he continues as general
manager. He lives in a
farmhouse on the Yorkshire
Moors near Harrogate
(Photo taken at interview
in 2018). (© Peter S
Walker)
the natural movement as possible. Furthermore, it was necessary to have a large

range of flexion which meant less conformity of the bearing surfaces.” Seedhom
appreciated just how sensitive the laxity and stability were to small changes in the
lengths of the ligaments which in turn was controlled by the shape of the bearing
surfaces, especially on the femoral side. To be consistent with this, a sufficient num-
ber of femoral sizes were required. From a large number of knee specimens,
Seedhom selected representative specimens of different sizes and made exact molds
of the surfaces. From each mold, a metal femoral component was produced as a
shell of uniform thickness. The divergence of the lateral and medial condyles from
each other of 15 to 20 degrees, clearly visualized on a distal view of the femur, was
reproduced in order to provide anterior-posterior stability. An intercondylar notch in
the femoral component allowed for cruciate ligament preservation. The shape of the
femoral component was not unlike that of the MGH femoral condyle, but Seedhom
incorporated it within a full artificial knee system with multiple sizes.
Frederick Ewald MD, an orthopedic surgeon from the Robert Breck Brigham
Hospital in Boston, also invented an innovative total knee with an anatomically
shaped femoral component (Ewald 1975). He filed a US patent in 1971 which was
issued in 1974. The femoral component was in the form of a shell, on which he
added a short intramedullary stem for stabilization, but this was not intended to be
cemented. The femoral component design overall was based on the MGH femoral
condyle, described at the beginning of this chapter. In both Seedhom and Ewald’s
designs, the surface of the femoral component that interfaced with the femoral con-
dyles was anatomic in shape, and consequently the component was very difficult to
fit accurately at surgery, even though cement was used for fixation. Another diffi-
culty at surgery was balancing of all four ligaments: even though the femoral com-
ponent was anatomic in shape, the variation in shape between individual knees
could result in mismatches in ligament length and tension patterns. The difficulties
in fitting and the ligament balancing, as well as the cost of manufacturing the ana-
tomic shapes, were the main reasons why these two designs, though elegant in form,
were withdrawn only a few years after introduction. In retrospect, with moderate
tibial dishing, the designs would likely have functioned well without the cruciates
and hence would have been easier to insert surgically.
Patella Flanged Components
All of the above femoral component designs, with the exception of the Geometric
and the Duo-condylar, had a patella flange. Although there is little documentation in
the early literature of the advantages of the patella flange, it was generally recog-
nized that there would be better fixation of the femoral component to the distal
femur due to the larger surface area. It also avoided the surgical difficulty in non-
flanged designs of providing a smooth transition at the anterior edge of the compo-
nent and the anatomic trochlea above. Most importantly, it provided a smooth
continuous track for the natural or artificial patella. A cartoon of the patellofemoral
“distress” in non-flanged femoral components was drawn by Seedhom in the pro-
ceedings of the 1974 Conference on Total Knee Replacement.
For early resurfacing femoral components without patella flanges, additional
fixation was needed other than relying simply on the flat or curved surfaces of the
femoral component. The solutions used were either a central peg to be cemented a
short distance into the intramedullary canal or two short pegs cemented into the
lateral and medial condyles where the cancellous bone is sufficiently strong. Designs
which used such means were the Geometric and the Duo-condylar. Femoral compo-
nents with a patella flange may have been thought to have sufficient fixation due to
the wraparound shape from posterior to anterior, but this is not the case for two
reasons. Firstly, as the component is placed on to the distal femur at surgery, it is
difficult to pressurize the cement on the anterior and posterior surfaces. Even if
there are a few degrees of divergence of these surfaces, on insertion at surgery, the
metal component tends to scrape the cement ahead of it, rather than pressurizing it
into the cancellous bone surfaces. Secondly, when the knee is flexed past 90 degrees,
the compressive forces on the posterior condyles have a force component which
tries to push the femoral component off the end of the femur. The consequence of
these factors is that the fixation, as for the non-patella flanged components, needs to
be augmented by a central peg or two side pegs.
Patella Patella groove in

No projecting No sharp surface of femoral shell
metal which corners in path femoral shell ensures against
impacts patella of patella extended dislocation
Fig. 7.5 Guidelines for designing the trochlea of the femoral component to ensure smooth con-
tinuous sliding, low contact stresses, and stability. This was presented at the 1974 conference in
London, at a time when several implants still did not have an anatomical type of design as in the
Leeds knee (Seedhom 1974). (Reprinted with permission from Institution of Mechanical Engineers,
London UK)
This was well appreciated by Townley who designed one of the first flanged
femoral components (Townley 1988). An important feature of Townley’s femoral
component was that there were five separate facets facing the bone: the distal
surface, the anterior and posterior surfaces, and 45 degree chamfers at the anterior
and posterior. These surfaces were straight-line approximations to the sagittal
curve of the distal femur. Such facets are now standard on almost all femoral
components, because it is easier to cut straight surfaces than curved surfaces,
given the practicalities of the equipment used for joint replacement surgery. Even
now with robot-assisted cutting tools, straight lines are still quicker than curves
though not as bone preserving. Another feature which Townley included was
pockets in the femur-facing facets of the femoral component, for containment of
cement. This would give the cement mantle some integrity and avoid thin layers
which could fragment under load. Other early designs, such as the Total Condylar,
adopted the pocket feature.
Multiple Sizes
Most of the condylar designs were only made in a few sizes, mainly for economic
reasons. However, it was not sufficiently recognized that accurately reproducing the
original femoral sagittal profile was necessary for maintaining taut ligaments
throughout flexion to achieve anatomic laxity and stability in every case. Seedhom
did however provide a rationale for the number of sizes necessary based on a radio-
graphic study of 131 knees (Seedhom 1972). He found close correlations between
90
85
c b
80
b mm
75
f e d 70
a
g
65
60
70 80 90 100
a = ML width of femur (mm)
Fig. 7.6 Bahaa Seedhom carried out a systematic dimensional analysis of the femur and tibia to
determine key dimensional parameters of the components for his anatomic knee design. From 101
radiographs and 20 cadaveric knees, he determined the relations between the parameters, illustrat-
ing the deviations from an exact correlation. One example is shown for the AP width to the ML
width. To limit the AP or ML error between the component and a given knee to 3 mm, he calculated
that five sizes would be needed of the femoral component (Seedhom et al. 1972). (Reprinted with
permission from Wear)
many of the dimensional parameters and chose the medial-lateral width as the refer-
ence parameter with which to scale the other dimensions. With a 6 mm spacing in
the medial-lateral dimension between sizes, he determined that only four sizes were
needed to cover the population. While 6 mm seems large by today’s standards, the
largest error in a given knee would only be 3 mm or 1.5 mm per side. It has been
noted that Seedhom used actual cadaveric specimens to make molds for the four
sizes, so the radii of the condyles and the patella groove in the anterior-posterior
(AP) and medial-lateral (ML) views would be automatically scaled. In a separate
study however, Seedhom took sagittal and frontal slices through the condyles so that
he could calculate the contact stresses on the polyethylene surfaces for the tibia, as
the basis for wear experiments.
Today, artificial knee systems have as many as 12 sizes of femoral components,
right and left, and, in some systems, additional sizes to include different medial-
lateral widths. The dilemma in defining dimensions, notably the AP and ML, is the
variations in their ratio, which is increased when different genders and ethnicities
are included. The AP dimension is the main factor controlling kinematics, whereas
there is a condition on the ML dimension that there should be less than 2–3 mm of
overhang on any side. While the ML dimension could simply be made narrow
enough to avoid overhang, this would result in too much exposed cancellous bone,
as well as reduced varus-valgus stability. The conclusion from different studies
based on fitting of large femoral databases of femurs in the computer concluded that
the more sizes, the better but with little further gains after the 12 sizes mentioned
above (Dai et al. 2014; Morris et al. 2016).
Medial
C
B S
S
Posterior Distal Frontal view extension,

Anterior plateau
A SS
radius AB radius BC lateral condyle circular,
sloped up 11 deg
22mm 32mm medial condyle
flattened. In flexion,
22mm
both condyles circular
Lateral
Sagittal view 0 deg

flexion, medial contact
A central-anterior, lateral
contact central. 120 deg
flexion, medial contact
BC central, lateral contact
posterior
Posterior Distal Lateral meniscus

radius AB radius BC displaced posteriorly
23mm ‘large’ in high flexion
Fig. 7.7 MRI studies of knee specimens were used to characterize the shapes of the femoral and
tibial bearing surfaces, the contact point locations, and the positions of the menisci. Describing the
condylar profiles as connecting arcs is useful for practical purposes such as artificial knee design
and modelling. The tibial slope on the medial side is seen as a result of the anterior region being
sloped upwards. Some anterior-posterior restraint is provided on the medial side due to the tibial
surface shape and the relatively immobile meniscus. Medial-lateral restraint is provided by the
upward slopes on the inner tibial plateaus (Martelli and Pinskerova 2002). (Reprinted with permis-
sion from Journal of Bone & Joint Surgery B)
The sagittal profiles of the femoral condyles are likely to affect the kinematics, as
well as the contact areas. In the anatomic knee, there are differences between the lat-
eral and medial profiles (Martelli and Pinskerova 2002). Laterally, the profile which
contacts the tibia during the flexion range was of a constant radius, with only a few
millimeters distally where the radius is large. On the medial side, the distal region up
to about 30 degree flexion had an average radius of 32 mm, while from 30 degrees to
maximum flexion, the radius averaged 22 mm. Martelli and Pinskerova also pointed
out that the medial profile could be represented by a radius which gradually reduced
from distal to posterior. These authors also described the interaction of the femoral
C C
J-curve 18 G-curve
19
20.5
B
34 64 32 A 64
A
C
18
Constant Super-
B radius AB imposed
23
D
A 75 Lowest
points
Fig. 7.8 Different schemes for constructing the sagittal profile of the femoral component. J-curve
= two radii, A to B and B to C. G-curve = gradually decreasing radius from distal A to posterior
C. Constant radius = constant from distal A to posterior B. The radii in millimeters are marked
adjacent to the curve segments. When the curves are superimposed, the maximum deviation at D
is 1 mm. The lowest point is 4 mm more posterior for the constant radius
condyles with the menisci and with the very different shapes of the lateral and medial
tibial plateaus. The complex interactions are generally not reproduced by the articulat-
ing surfaces of an artificial knee, and certainly, the femoral condyles alone will not
determine the kinematics. Artificial knees use three main schemes for the sagittal pro-
files from distal to posterior. Firstly, a constant radius resembles the anatomic lateral
condyle. Secondly, a larger distal radius and a smaller posterior radius resembles the
anatomic medial condyle; this is called a J-curve. Thirdly, a gradually reducing radius
from distal to posterior again resembles the medial condyle; this is called a G-curve
(Pfitzner et al. 2018). There are an infinite number of ways to construct these curves.
However when they are superimposed, there appears to be only small differences in
the profiles. One difference however is that the lowest point on the profile which will
contact the tibial surface in early flexion may be more anterior for the J-curve and the
G-curve than for the constant radius. A final point about the profiles is that if the goal
is to mimic anatomy, the lateral condyle should be of constant radius, while the medial
condyle should be a G-curve or J-curve. But again, this would need to be matched
with asymmetric tibial surfaces.
Patellofemoral Joint
One aspect of femoral component design, the patellofemoral joint, received rela-
tively little attention in early studies. This is remarkable considering that the quad-
riceps action is the driver of function for which a smooth-acting and stable patella is
essential. It is likely that designers carried out basic measurements of the femoral
groove on radiographs or on specimens, without making that aspect into a major
study for publication. There was one exception to this, the work of Paolo Aglietti
MD and associates in the Biomechanics Laboratory of the Hospital for Special
Surgery (Aglietti et al. 1975). Aglietti, from Florence in Italy, was a surgical fellow
of John Insall, chief of the knee service. Aglietti questioned whether a patella resur-
facing component should be anatomic in shape or possibly some other shape which
would be more suitable for an artificial knee. Working with John Insall, Peter Trent,
a masters of mechanical engineering from Columbia University, and Peter Walker,
Paolo designed a resurfacing component for the patella (Aglietti et al. 1975). The
idea was that the artificial patella could be used as an independent implant or used
Fig. 7.9 Paolo Aglietti

MD was a fellow in John
Insall’s Knee Service at the
Hospital for Special
Surgery (HSS) in the early
1970s, when this picture
was taken. He participated
in the design of the early
HSS knees including the
instrumentation. He carried
out lab studies for the
design of the dome patella,
widely used today. He was
coeditor with John Insall
on the first edition of
Surgery of the Knee. Paolo
later returned to Florence
to be Professor of
Orthopedics, where he
became a founding
member of the Italian Knee
Society. He maintained a
strong relation with John
Insall and HSS throughout
his career. (© Peter S
Walker)
Dome
Cone
M L
Offset
cone
Gaussian
Anatomic
rotating
platform
Fig. 7.10 Different possible configurations for the shape of an artificial patella. The offset cone is
the closest resemblance to anatomic, but the most widely used design is the simple dome due to its
simplicity and its ability to tilt in different directions without edge loading. The shape of the inter-
condylar region (arrows) affects contact area and stability
with the Total Condylar knee which was by then being used clinically. Dimensional
parameters were measured from 80 resected patellae from surgical cases, ranging
from “normal” to severely arthritic with grossly abnormal shape. Using this data, a
dome-shaped patella was designed made from metal for use with an anatomic femur
and from polyethylene to be used with the Total Condylar. Contact locations of
normal intact knee specimens and those with a dome inserted were determined
using an acrylic casting technique. Similar contact patterns were observed. Stability
tests using a test rig were also carried out on specimens, where the Q-angle was
varied. The dome was equal in medial-lateral stability to an anatomic patella. The
conclusion was “The dome-shaped prosthesis conformed reasonably to all the
femurs we have tested so far and located well into the groove, allowing some inher-
ent stability. The symmetrical shape eliminates any ambiguity in insertion, and also
accepts twisting and tilting with no change in contact area and with little stress on
the fixation.” So even though the anatomic patella has two distinct facets, the lateral
longer than the medial, the dome was considered to be the best solution because of
its adaptability to any shape of groove in any alignment. Even though there might
be tilting or rotation, the dome will still provide coherent lateral and medial contact
areas without any edge loading. In clinical application, in artificial knees the plastic
dome patella did develop two patches of wear and deformation, indicative of
exceeding the yield stress. However at that stage, the stresses would be reduced and
would be tolerated by the plastic over a long time period without severe wear. This
was largely due to the fact that in walking, the most frequent activity, the patello-
femoral forces are less than the body weight, whereas in high flexion, even though
the forces are higher, the frequency of loading is much less. At the same time, the
intercondylar cutout, between the femoral condyles as in the femoral component of
most of today’s artificial knees, has shown disadvantages. Soft tissue distal to the
resurfaced patella has become trapped between the condyles causing a snapping or
popping as it releases, a phenomenon that has been more frequent with the posterior
stabilized (PS) designs where the intercondylar housing has been more extensive
and not sufficiently rounded.
Comparing today’s femoral component designs of standard condylar replace-
ments with those of the 1970s, the shapes have not in general evolved from being
symmetric and geometric to being asymmetric and anatomic, as might be expected.
While most designs have asymmetry in the overall peripheral shape especially in the
patella flange area, only a few designs have differences in the sagittal profiles
between the lateral and medial sides. Those designs that do have condylar asym-
metry are mostly those where one or both cruciate ligaments are preserved, whereas
for the PS designs, medial-lateral asymmetry is not considered necessary. In one
study, the differences in geometry between the femoral component and its host
Fig. 7.11 A typical

condylar femoral
component with convex
bearing surfaces to control
medial-lateral sliding, with
an anatomically shaped
trochlea for retention of the
patient’s patella or for a
plastic resurfacing, with a
right-left flange shape to
better fit the bone surface,
and with a rounded
intercondylar space for
retention of the posterior
cruciate
Fig. 7.12 The process for manufacturing femoral components from cobalt-chrome alloy, using
the “lost wax” process. Ceramic coats the original wax models. The wax is melted out, after which
molten metal is poured into the cavity. The components are mounted on a “tree,” and connecting
holes are made between components, so the molten metal will fill many components at the same
time. After the metal hardens, the ceramic is removed, leaving individual femoral components.
(Photo taken in manufacturing plant of Howmedica, now Stryker)
femur, combined with variations in surgical placement, were shown to result in

dimensional discrepancies up to 6 mm (Akbari et al. 2018). This would make it dif-
ficult to balance the knee, while kinematic abnormalities might occur. Remarkably,
the patella is symmetric in the majority of knee systems, despite the obvious asym-
metry of the anatomic patella with a longer lateral facet than medial. The likely
reason for this situation is that there have not been any compelling clinical advan-
tages in asymmetry, nor disadvantages of symmetry, to cause a widespread move-
ment to asymmetry.
o What Extent Do Femoral Component Shapes Mimic Normal

T
Anatomy?
Most of the earliest designs were simplified shapes, not resembling normal anatomy
either in the sagittal radii or in the overall contours. The exceptions to this were the
MGH femoral condyle, designed to be used without a tibial component, the Ewald
artificial knee which was based on the MGH, and the Leeds knee which was made
from actual molds of anatomic specimens. Today, cruciate sacrificing designs usu-
ally have the same sagittal profiles laterally and medially, in contrast with normal
anatomy; cruciate preserving knees usually have different lateral and medial con-
dyle shapes. While the key dimensions and multiple sizes can provide a reasonably
close match to any knee, the overall surface shapes of the femoral components are
not anatomical and can deviate by several millimeters from a host femur. The dis-
crepancies may cause some deficiencies of kinematics and function.
References
Aglietti P, Insall JN, Walker PS, Trent P. A new patella prosthesis. Design and application. Clin
Akbari S, Boulos P, Saavarsson SK, Ramm H, Fu CKJ, Miller S, Zachow S, Anglin P. Changes
in knee shape and geometry resulting from total knee arthroplasty. Proc Inst Mech Eng H.
2018;232(1):67–79.
Attenborough GG. Total knee replacement using a stabilized gliding prosthesis. Conference on
Total Knee Replacement, Institution of Mechanical Engineers, London UK, 16-18 September
1974.
Bloom JD, Bryan RS. Wide-track polycentric total knee arthroplasty. Clin Orth Rel Res.
1977;128:210–3.
Chao EY, Stauffer RN. 1974 Biomechanical evaluation of geometric and polycentric knee arthro-
plasty. Conference on Total Knee Replacement, Institution of Mechanical Engineers, London
UK, 16-18 Sept 1974.
Coventry MB, Finerman GA, Riley LH, Turner RH, Upshaw JE. A new geometric knee for total
knee arthroplasty. Clin Orthop Relat Res. 1972;83:157–62.
Dai Y, Scuderi GR, Penninger C, Bischoff JE, Rosenberg A. Increased shape and size offerings of
femoral components improve fit during total knee arthroplasty. Knee Surg Sports Traumatol
Arthrosc. 2014;22(12):2931–40.
Deane G. The Deane Knee: a new concept in knee joint design. Conference on Total Knee
Replacement, Institution of Mechanical Engineers, London UK, 16-18 September 1974.
Devas M. Link arthroplasty of the knee. Conference on Total Knee Replacement, Institution of
Mechanical Engineers, London UK, 16-18 September 1974.
Englebrecht E, Zippel J. The sledge prosthesis model St. Georg. Acta Orthrop Belg.
1973;39(1):203–9.
Ewald FC. Metal to plastic total knee replacement. Orthop Clin North Am. 1975;6(3):811–21.
Frankel VH, Burstein AH, Brooks DB. Biomechanics of internal derangement of the knee.
Pathomechanics as determined by analysis of the instant centers of motion. J Bone Joint Surg
Am. 1971;53(5):945–62.
Bone Joint Surg (Br). 1971;53(2):272–7.
Gunston FH, MacKenzie RI. Complications of polycentric knee arthroplasty. Clin Orthop Relat
Res. 1976;(120):11–7.
Hodge WA. Vitallium-mold arthroplasty of the knee. A case report with 30-year follow-up study.
J Arthroplast. 1991;6(3):195–7.
Lewallen DG, Bryan RS, Peterson LF. Polycentric total knee arthroplasty. A ten-year follow-up
study. J Bone Joint Surg Am. 1984;66(8):1211–8.
Martelli S, Pinskerova V. The shapes of the tibial and femoral articular surfaces in relation to tib-
iofemoral movement. J Bone Joint Surg (Br). 2002;84-B:607–13.
Morris WZ, Gebhart JJ, Goldberg VM, Wera GD. Implant size availability affects reproduction of
distal femoral anatomy. J Knee Surg. 2016;29(5):409–13.
Pfitzner T, Moewis P, Stein P, Boeth H, Trepczynski A, von Roth P, Duda GN. Modifications of
femoral component design in multi-radius total knee arthroplasty lead to higher lateral poste-
rior femoro-tibial translation. Knee Surg Sports Traumatol Arthrosc. 2018;26:1645–55.
Seedhom B. A technique for the study of geometry and contact in normal and artificial knee joints.
Wear. 1972;20:189–99.
Seedhom BB, Longton EB, Dowson D, Wright V. The Leeds Knee. Conference on Total Knee
Replacement, Institution of Mechanical Engineers, London UK, 16-18 September 1974.
Smith-Petersen MN, Aufranc OE, Larson CB. Principle of mold arthroplasty as applied to the hip.
Surg Clin North Am. 1947;27:1303–6.
Stockley I, Douglas DL, Elson AA. Bicondylar St. Georg sledge knee arthroplasty. Clin Orth Rel
Res. 1990;(255):228–34.
Townley CO. Total knee arthroplasty: a personal retrospective and prospective review. Clin Ortho
Rel Res. 1988;236:8–22.
Tibial Component
What Are the Best Configurations for Durability?
8
The tibial component provides a low-friction and wear-resistant bearing surface for
the femoral component, as well as constraints to displacements and rotations. In
order to avoid loosening over time, the component must have the means for rigid
fixation to the resected tibial surface.
Opposite Extremes of Tibial Component Design
The earliest tibial component designs had strikingly different configurations. The
Freeman-Swanson consisted of a plate of polyethylene covering the entire resected tib-
ial surface without provision for retaining the cruciate ligaments. This was based on the
knee pathology for which the design was intended: “the theoretical advantage of retain-
ing the cruciate ligaments was not realized in those severely disabled knees in which
these ligaments are often damaged” and “hence the largest possible area in one plane is
presented to the tibia…thus minimizing the chances of compressive failure of the bone”
(Freeman et al. 1973). On the other hand, Gunston’s Polycentric tibial components were
small polyethylene runners to be embedded in V-shaped slots in the proximal tibia. This
design resulted from an entirely different set of design criteria to Freeman and Swanson’s:
“Separate replacement of the diseased joint surfaces using cement…permits adjustment
of the relative levels of the prosthetic surfaces. Collateral ligament laxity may be reduced
and varus or valgus deformity decreased” (Gunston 1971). Freeman and Swanson
placed a higher priority on minimizing the stresses on the trabecular bone of the proxi-
mal tibia, over retaining the cruciates; whereas Gunston’s criteria were the opposite.
The Freeman-Swanson can be thought of as a mechanistic approach, where the
joint stability was provided by the component geometry, while the alignment was
based on making the bone cuts relative to defined axes in the femur and tibia. It was
assumed that “when the prosthesis has been correctly implanted the soft tissues will
be tight in extension but a little lax in flexion.” In the clinical follow-ups, some
patients reported varus-valgus instability which could be attributed to the initial
surgical technique where “the bone ends were resected at right angles to their long

122 8 Tibial Component
axis.” This would leave the limb in some varus compared with normal alignment
where the distal femoral condyles are aligned an average of 5–6° to the anatomic
axis. The Gunston design was an anatomic approach where the goal was restoration
of the levels and contours of the original bearing surfaces and correct tensing of the
ligaments to perform their normal functions. Separate components for each condyle
were important to enable individual adjustment, difficult to achieve with a single
one-piece component. At the same time, alignment of four different components
both in height, rotation, and medial-lateral position would not be easy for the sur-
geon, and there was a limited surface area of the tibial components for fixation.
Application of Anatomic Experiments
Subsequently, several tibial components intermediate to the Freeman-Swanson and

Gunston were designed, preserving a complete bridge of tibial bone in the center. At
the Hospital for Special Surgery in 1970, Peter Walker and Joe Hajek, an orthopedic
resident, determined the location of the contact areas in knee specimens throughout
flexion (Walker and Hajek 1972). Their method was to remove the menisci, fill the
cavity with soft acrylic cement, load the knee in an Instron machine, and then
remove the hardened acrylic casting. The holes in the casting on the lateral and
medial sides signified the contact areas. From a composite of the contact areas over
a full range of flexion, they made a model of a tibial component with an anterior
bridge. The tibial spines were replicated to provide medial-lateral constraint and
limit internal-external rotation: “The weight-bearing surface curves upward toward
the intercondylar area to provide medial-lateral and rotary stability, but provides no
stability in the anterior-posterior plane.” Anterior-posterior stability was to be pro-
vided by the preserved cruciate ligaments and “allow normal patterns of motion as
characterized by the variable instant centers.”
When the Duo-condylar knee was designed, however, based on the above study,
the crossbridge was abandoned to allow for individual component placement as
with the Gunston (Ranawat and Shine 1973). The provision of several thicknesses
allowed for individual tightening of each compartment. The correct medial-lateral
spacing of the two separate components at surgery was obtained by an insertion
instrument locating in holes on the front of the components. Nevertheless, the use of
separate components was discontinued after a few years due to fixation problems,
even though the components covered the full extent of the normal contact areas.
Metal Tibial Tray and Modularity
In the late 1960s, a completely new approach to tibial component design was being
developed by Nas Eftekhar, a surgeon from Columbia Presbyterian Hospital in
New York (Eftekhar 1978). Eftekhar had spent 18 months at Wrightington Hospital
in England with John Charnley. There he learned the technique of low-friction hip
arthroplasty. He overlapped with Frank Gunston and was interested in his knee
design, but considered that the small components designed by Gunston would not
8 Tibial Component 123
provide sufficient strength of fixation. On his return to the USA, Eftekhar was
appointed at Columbia Presbyterian Hospital where he established programs for
artificial hips and knees. Working in his own lab and with some involvement from
Howmedica (now Stryker), he designed the Intramedullary Adjustable knee, which
was an adaptation of Gunston’s knee. Eftekhar’s femoral component consisted of
two Gunston components without the full half-disc shape, joined by a crossbridge
on which there was an intramedullary stem. The tibial component was a metal plate
covering the entire proximal tibia where Gunston plastic components were pressed
into recesses. The metal plate had a central intramedullary stem for fixation and two
small side pegs to resist rotary forces. The design rationale was stated as: “The prin-
ciple fixation mechanism was achieved by intramedullary elements of the prosthesis
(metal stems) and the differential plastic heights were press-fitted into the metallic
component of the tibia to prevent cold flow of the plastic with the hope of prevent-
ing loosening.” Eftekhar had decided that the cruciate ligaments were not neces-
sary for an artificial knee, which made it easier to achieve balancing. The plastic
Fig. 8.1 While on a Fellowship at Charnley’s Hip Center in the late 1960s, Nas Eftekhar MD
became familiar with Gunston’s Polycentric design. On returning to New York, he worked with
Howmedica in adapting the components for stronger fixation (left). The plastic inserts could be
adjusted individually for height. The metal tray was intended to enhance fixation. Later, Eftekhar
used the same idea but this time modified the Total Condylar design, providing a metal-backed
tibial component with a one-piece plastic. It provided more rotational laxity than the Total
Condylar. This was the first metal-backed component to be designed for an artificial knee (Eftekhar
1978). (Left, Reprinted with permission from Clinical Orthopaedics and Related Research. Right,
photo of sample courtesy of William Macaulay MD)
thicknesses of 6, 10, 14, and 20 mm seem to be too widely spaced by today’s
standards. This design was used in patients from 1970 till 1974. When the Total
Condylar knee was being produced by Howmedica in 1973, Eftekhar utilized the
Total Condylar femoral component with the addition of a stem and a one-piece
plastic insert which seated into a tibial tray. But by now, Howmedica was in full
production of the Geometric knee, while the Total Condylar became an even higher
priority. Consequently, the Eftekhar knee was not widely adopted and faded out
after a few years. Another major reason was that surgeons were not ready to revert
to long cemented intramedullary stems, going back to the days of all-metal hinges.
Nevertheless Eftekhar‘s tibial components were the first metal-backed designs, with
the advantages of rigidity of fixation and interchangeability of the plastic inserts.
While Eftekhar’s tibial tray design had sufficient strength, this was not the case
for some subsequently designed metal trays. For example the Kinematic knee
showed an incidence of tray fracture of an estimated 1–2% (Abernethy et al. 1996).
Cracks initiated at the corners of a posterior rectangular-shaped notch which pro-
vided clearance for the posterior cruciate. Associated factors were lack of bone sup-
port under the metal tray and undercorrected varus deformity. Other contributing
factors were that the tray itself was too thin, in order to minimize the amount of
bone resection, and a pyramidal pattern on the base to grip the cement which acted
as a stress concentration. A finite element analysis showed that the development of
a fibrous tissue layer beneath the tray increased the stresses to the point where frac-
ture was possible (Paganelli et al. 1988). As a result of this failure mode, a study was
initiated by Peter Walker who was Director of the Biomedical Engineering
Department at the Royal National Orthopaedic Hospital in Stanmore, UK, affiliated
with University College London. The purpose of the study was to develop a testing
method for predicting whether a metal tibial tray design would fail in service. In
collaboration with surgeons, radiographic data was obtained from clinical cases
where fractures had occurred. Consistent with the above study by Paganelli, it was
found that all of the failures occurred when there had been a fibrous tissue interface
beneath the tray medially. This led to a test scenario where half of the tray was
clamped rigidly, while a cyclic load was applied to the overhanging part of the com-
ponent (Ahir et al. 1999). The result of this work was a standard adopted by the
ASTM entitled: “Cyclic fatigue testing of metal tibial tray components of total knee
replacements.”
Conformity and Cruciate Preservation
The scheme of an anterior bridge to preserve the cruciate ligaments was used in the
early 1970s in the Leeds, the Ewald, and the Geometric knee designs. The rationale
was to maintain a one-piece component, which was easier for surgery, to retain both
cruciate ligaments, and to provide extra-fixation from the bridge. Even so, compar-
ing separate components to a single component with an anterior bridge, the utility
of the bridge can be questioned, especially with an all-plastic component. The
increase in surface area for fixation is small in relation to the areas of the lateral and
Fig. 8.2 Bahaa Seedhom Ph.D. designed the Leeds knee with all-plastic tibial components where
the polyethylene was compression molded for enhanced wear resistance. There were three sizes
right and left. The peripheral shape was based on anatomic specimens. The holes in the trial com-
ponents (top right) were for drilling holes in the tibia, which lined up with fixation pegs on the base
of the components. (© Peter S Walker)
medial condyles themselves, while the removal of the anterior tibial bone weakens
the attachment of the anterior cruciate ligament, in some cases causing the ligament
to avulse. Nevertheless the advantage of joining the lateral and medial plateaus with
the consequent ease of surgery was the main priority. In the case of the Geometric,
the tibial bearing surfaces were in close conformity with the femoral surfaces with
only a small amount of mismatch to allow for some rotational laxity. It transpired
that the close conformity had two disadvantages. The average flexion angle in an
early series was only 87°, but in addition, there was an average of seven degrees of
extension lag (Skolnick et al. 1976). The likely reason for this was that in preserving
the cruciate ligaments, there needed to be provision for up to 10 mm of anterior-
posterior displacement of the femur on the tibia during the flexion range, which
would not be possible due to the high femoral-tibial conformity. The close confor-
mity between the bearing surfaces under loaded conditions would allow almost no
displacement or laxity. On the contrary, the intention was to provide a stable knee
joint by adjusting the tibial thickness: “laxity of capsular and ligamentous structures
is taken up by the so-called spacer effect of the components.”
Another effect of the high conformity was on the fixation of the components.
Even at 2 years follow-up, there was 12% of loosening while “lucent lines were
noted at the tibial cement-bone interface in all cases of loosening and in about 80%
of the remaining knees.” The authors considered that “uncorrected varus alignment
of the knee was thought to be the primary cause of loosening” with other factors
being inadequate medial bone support and incorrect component seating. In addition,
the component conformity would have caused high shear and tensile stresses at the
cement-bone interface, but the fixation pegs were relatively small. Also at that time,
there was little recognition of the need for adequate cement penetration into the
trabecular bone. After several years of widespread clinical use, the Geometric knee
was replaced by other designs. Nevertheless, the Geometric was the first total knee
in the USA to have been designed initially by an engineer, Bob Averill, followed by
bringing together a number of prominent surgeons from different institutions. Over
the first 2 years, the group refined the design and developed the surgical instrumen-
tation. There were many iterations involving cadaveric trials and mechanical test-
ing. When the knee was first introduced at the AAOS Annual Meeting in 1972, there
was an impressive exhibit with demonstrations by the developers, generating enor-
mous interest among the surgeon community.
At about the same time that the Geometric knee was introduced, Charles Townley
started using a cruciate-preserving design he had been refining for many years. The
Townley Anatomic knee was however only used by Townley in the first few years
due to the delay in finding a manufacturer who would widely promote it. Townley’s
efforts on a knee replacement design began in the early 1960s when he introduced
the articular-plate replacement, consisting of a thin metal shell for tibial resurfacing,
fixed to the bone with two screws through anterior metal lugs. At the same time,
Townley had designed an anatomically shaped femoral component, but this was not
implemented because it would need to have been made in metal and there was a
reluctance to use metal-on-metal in a non-conforming bearing. However once metal
and polyethylene knees fixed with acrylic cement had been introduced by Freeman
and Swanson in England in the late 1960s, Townley saw this as the solution.
Townley’s goal in reconstructing the knee had always been to restore normal anat-
omy and function: “optimal results require proper implant design and surgical tech-
nique that restore the normal joint anatomy mechanics and kinematics” (Townley
1985). Not surprisingly, the bearing surfaces were a close reproduction of the ana-
tomic shapes, the components were made as thin as possible, and the cruciates were
preserved. The tibial component was all-plastic with a central cutout for the cruci-
ates and an anterior bridge to join the lateral and medial sides. The tibial bearing
surfaces were shallow to allow sliding and rotations to occur between the femur and
the tibia: “For optimal postoperative motion, the normal synchronized relation
between the ligaments and the articular outlines of the femoral condyles must be
retained. This requires anatomic design contouring and accurate size selection and
technical positioning of the implants.”
The results of the Townley were better than those of the Geometric knee both in
durability and range of flexion. Townley claimed only 2% revision for loosening in
over 500 cases followed for 2–11 years. “The postoperative motion was at least 90°
in 91%, between 110–120 in 21%, and in excess of 120° in 12%.” Both the favor-
able motion and fixation results can be attributed largely to the low conformity
between the femoral and tibial components such that there would be relatively low
shear forces transmitted to the tibial component. Nevertheless the low loosening
rate is not easy to understand considering that thin plastic components are known to
deform under weight-bearing, while the stresses on the surface are transmitted
directly to the surface of the trabecular bone with little attenuation. The answer may
Fig. 8.3 Long-term follow-up of a Townley Anatomic knee. The patient’s knee was in severe
varus with a medial bone defect. The defect was filled with a graft and fixed with screws, prior to
the cementing of the entire component. The tibial component was metal-backed in the shape of a
horseshoe for preservation of both cruciate ligaments. A number of small drill holes can be seen in
the proximal tibia for cement penetration, to provide improved fixation. The patella component is
also metal-backed. (Photos courtesy James W Pritchett MD)
lie in the meticulous attention given to the surgical technique in that the target was
to align the knee to the mechanical axis with the valgus angle at the knee set on an
individual basis, and the joint line in the frontal plane set parallel to the ground. As
for the cement technique: “small one-eighth inch anchor holes are punched into the
supporting bone surfaces of the femur and tibia.” The radiographs shown by Townley
in articles do show substantial cement penetration into the trabecular bone, shown
in experiments by Krause many years later to provide a major increase in fixation
strength. Another factor was that there was only a few millimeters of proximal tibial
bone resection, preserving the strongest trabecular bone near to the surface. Townley
used his engineering knowledge and careful observation to formulate solutions
which worked.
Cloutier’s total knee design was similar to that of the Townley, but in this case,
the tibial component consisted of a metal baseplate where individual plastic bear-
ings were snap-fitted laterally and medially (Cloutier 1983). The bearing surfaces
were again shallow. The anterior bridge being made of metal could be much smaller
than for the Townley all-plastic design, although in a metal-backed construct, the
total thickness of the component is elevated due to the need for 6 mm minimum
thickness of plastic and 2–3 mm of metal-backing. Certainly, the load distribution
on the underlying trabecular bone is more uniform than for an all-plastic compo-
nent, which has generally been considered to be an advantage for long-term fixa-
tion. Cloutier emphasized that the preservation of the ACL, as well as the PCL, was
the only way in which normal kinematics could be achieved. He also pointed out
that in normal knees, when the ACL had been torn, posteromedial wear was often
seen, as well as AP instability in function. The same could occur in an artificial knee
where the anterior cruciate was resected. He used an advanced surgical technique,
incorporating ligament releases, making bone cuts perpendicular to the mechanical
axis, and tensing the ligaments with a distractor.
Total knee designs similar to those of Townley and Cloutier have been in con-
tinual use to this day, and in recent years, there has been some increased interest due
to improved surgical techniques where greater accuracy could be achieved. It is
logical that the best functional results can be obtained if the knee is restored to its
original anatomical condition, which requires both accurate shapes of implants and
a high degree of surgical precision, as well as adequate means of fixation. Hence
this type of tibial component with cruciate preservation is a viable design for achiev-
ing optimal results, at least for those patients without severe deformity and where
the cruciate ligaments are preserved and functional.
Development of the Central Peg
So far, three types of tibial components have been described: a one-piece component
covering the entire tibial surface, two separate components replacing the lateral and
medial plateaus, and two separate plateaus joined anteriorly with a bridge. Yet
another scheme, introduced in the early 1970s, was to increase the size of the anterior
bridge in the anterior-posterior direction, at the expense of resecting the anterior
cruciate ligament. This work took place at the Hospital for Special Surgery, where
Chit Ranawat, John Insall, and Allan Inglis had been using the Duo-condylar knee
for 1–2 years. Already, the deficiencies of separate plastic plateaus had been observed,
and while there were potential advantages of a one-piece component with a large
surface area such as in the Freeman-Swanson design, even there, reliable fixation had
not been achieved. Moreover, there was no desire at that stage to sacrifice both cruci-
ate ligaments. Because of the location of the posterior cruciate ligament, all that was
required for its preservation was a central notch about 15 mm in depth at the posterior
of the component. This meant that the tibial component would cover most of the
resected surface of the tibia, increasing the stiffness of the component and enhancing
fixation. It was generally assumed that the posterior cruciate was more important to
preserve because it was thicker and stronger than the anterior cruciate and prevented
the femur from sliding anteriorly on the tibia, whereas the anterior cruciate was often
absent or deficient in the arthritic knee. Another strong motivation for preserving
only the posterior cruciate was to simplify the technique. When both cruciates were
preserved, it was difficult to access the bone surfaces for preparing keyholes for pegs
in the components, removing soft tissue around the edges of the tibia, and removing
excess cement which may have squeezed out posteriorly. With the anterior cruciate
resected, the tibia could be subluxed anteriorly affording improved exposure. One of
the earliest such designs of this type was the Duo-patella, designed collaboratively
between the Hospital for Special Surgery and the Robert Brigham Hospital in Boston
(Sledge and Ewald 1979). Artificial knee designs which preserve only the posterior
cruciate are termed “cruciate retaining,” or CR, and have been widely used, initially
all-plastic, but, in recent years, predominantly metal-backed.
From the outset, the CR tibial component incorporated one vital feature, a central
fixation peg. Most other tibial component designs had various features on the base
for fixation to the proximal tibia, such as serrations or dovetails so that the cement
would lock to the component. Other designs used short projections requiring holes
or slots to be cut into the bone. The principles were explained by Swanson et al.
(1973), in terms of the various forces acting on the component: “An axial force act-
ing on the center of the component would be supported by compression stresses on
a resected surface of trabecular bone. Hence the largest surface area possible would
be preferable. Another way in which an axial force can be supported is by a tapered
peg in which case some of the axial force is supported as shear down the sides of the
peg. Indentations of some sort would increase the shear force that could be transmit-
ted. However an excessively rigid intramedullary stem may raise the danger of
localized disuse-osteoporosis, or stress protection. An anterior force on the compo-
nent would require a means of preventing the posterior of the component from dis-
tracting upwards.” Swanson correctly pointed out that on the medial side at least,
the strongest trabecular bone was located close to the center of the lateral and medial
tibial plateaus such that in the sagittal plane: “there would be a tendency for the
component to rock about a fulcrum of relatively stiff bone.” Torque acting on the
surface of the component would produce shear stresses at the cement-bone inter-
face, but such stresses could be reduced by projections into the bone which would
transmit forces on the sides. Another important point made by Swanson was that:
“tensile and shear stresses at the interfaces be as small as possible…and compres-
sive stresses are at a level that will not cause fatigue failure in the bone.” Systematic
thinking such as this had seldom been evident in total knee design. In the case of
design features for optimal fixation of the tibial component, it was scarcely raised as
a design problem requiring special attention.
The Special Surgery group was likely the first to carry out experiments to demon-
strate the efficacy of a central peg. There were several factors that led to this solution.
Observations of their own designs, the Unicondylar and Duo-condylar, as well as of
other early design such as the Freeman-Swanson, Polycentric, and Geometric,
showed that short projections into the resected proximal tibia did not provide suffi-
cient fixation, evidenced by frequent development of radiolucent lines at the bone-
cement interface. The lines would often progress over time to cover the entire
interface, such that a revision procedure was often necessary. A recommended tech-
nique for fixing compartmental components, based on laboratory experiments, was
to drill holes about 10 mm in diameter in the anterior and posterior of the resected
tibial surface (Walker et al. 1976). The cement would fill these holes and penetrate
the surrounding trabeculae, providing resistance to distraction as well as carrying
shear forces. However a reliable method was needed for rigidly fixing a one-piece
all-plastic tibial component as used for the Total Condylar or Duo-patella. Previous
studies of the density and strength of the trabecular bone in the proximal tibia had
shown that the bone in the center was the weakest, whereas the strongest bone
occurred in the central regions of the lateral and medial plateaus, the major areas of
load transmission (Behrens et al. 1974). Hence if the weakest bone at the center was
removed, it could provide access for a central stem. Metallic stems had of course
been used for decades in the hinged replacements as well as the more recent Eftekhar
knee, but these were intended to fix to the cortical bone along a length of 120 mm or
more below the tibial surface. There was however one example of an all-plastic tibial
component with a central stem only about 60 mm in length, the Attenborough
Stabilized Gliding prosthesis (Attenborough 1976). The rationale for the stem in that
design was to provide varus-valgus stability, but in an unlinked design, this would
not be a major requirement. However, if a central peg was designed with a rectangu-
lar section and cemented into a central cavity, it could possibly transmit anterior-
posterior shear forces as well as axial rotation forces. In addition it would also
transmit vertical shear forces down the trabecular bone at the sides of the cavity.
The initial central peg designed by the HSS group was 40 mm in length, with a
rectangular section. The peg was tapered posteriorly as seen in the sagittal view, so
that the posterior distal corner of the stem would not impact the posterior tibial cor-
tex. The length was based on the observation that the density of the trabecular bone
reduced steadily with depth, reaching the hollow intramedullary cavity at about
60 mm depth. Another important feature of the peg was the rectangular cavities cut
into the sides. These cavities would fill with cement and transmit the vertical shear
forces. It was proposed that even if the bone-cement interface on the horizontal sur-
face of the tibia became loose, the central peg was likely to remain firmly fixed and
hence maintain strong enough fixation of the entire component. An advantage of the
central peg was that it did not invade the strong bone on the lateral and medial sides
of the tibia. For this reason, smaller side pegs to augment the central peg were not
added.
While it seemed likely that a central peg would afford stronger fixation than
previous configurations, experimental validation was needed. The most direct
method for this was mechanical testing, but machines which simulated the variable
loading conditions of everyday use had not yet been designed. Moreover the test
would require cyclic loading of high enough magnitude over an extended time
period. Using a single station would take a great deal of time to test sufficient sam-
ples. This was the motivation for the design of a multichannel test machine. “A
9-channel machine was designed and constructed which oscillated the femoral com-
ponent back and forth through an arc of 45° in a flexion-extension plane. A force of
twice body weight was applied for one half of the cycle by air cylinders mounted
above the femoral components. The tibial components were cemented to cadaveric
tibias which were fixed to the machine base” (Walker and Hsieh 1977). Anterior-
posterior shear force components were applied to the tibial components when the
femur was in a flexed position, because the force was applied down the axis of the
femoral component itself. The test was designed to compare separate condylar com-
ponents with anterior and posterior holes for cement and a one-piece component
with a central peg. The tests showed the clear superiority of the central peg design
which remained rigidly fixed in all specimens. The separate plateaus failed by com-
pressive failure of the trabecular bone at the interface. Where there were several
keyholes for cement: “the islands of bone (between the key holes) thus created were
susceptible to weakening.” “The logical provision of a central peg, introduced into
the weak central area but flanked by strong cancellous bone, was shown to provide
the necessary strength…”
Central peg
component
Separate
components
Fig. 8.4 Testing of the fixation of tibial components using a nine-channel knee simulating machine,
at Hospital for Special Surgery. The machine flexed-extended with cyclic loading for up to 1 million
cycles. The load included shear forces during the flexed phases. All central peg components
remained well fixed. The separate components failed at different times due to failure of the trabecu-
lar bone beneath the components and between cement holes drilled to enhance the fixation (Walker
and Hsieh 1977). (Top, Reprinted with permission from Journal of Bone & Joint Surgery A)
The all-plastic tibial component with a central peg was accordingly applied to
the Total Condylar and Duo-patella knees. Follow-up studies of various central peg
designs used to this day have attested to the durability of the central peg concept
used on all-plastic components. However certain deficiencies have been noted.
Radiolucent lines have frequently developed at the sides of the horizontal tibial
surface as seen in the anterior-posterior view, signifying loss of rigid cement-bone
fixation. This has been most commonly seen on the medial side, likely due to the
higher axial forces occurring on that side. From radiographs the radiolucency has
often been seen to be associated with inadequate cement penetration. As the compo-
nent is seated at surgery and pressed down on the soft cement layer, the pressure at
the sides is much less than at the interior because the cement can readily flow out
sideways. This is a well-known characteristic of a squeeze film. Another situation
has been where the medial side of the knee has been overloaded, and there has been
gradual collapse of the trabecular bone. The plastic has permanently bent downward
due its lack of stiffness. Cement has often extruded beyond the tip of the stem into
Fig. 8.5 The design drawing of the first Total Condylar central peg tibial component dated January
1974. The radii of the condyles in frontal and sagittal planes were slightly larger than those of the
femoral component, calculated to provide the anatomic values of laxity and stability under func-
tional weight-bearing conditions. The component fitted both right and left knees, and there were
two sizes. The lower surface of the component was knurled to provide shear transmission of forces
on the cement. The sides of the peg had pockets to transmit shear forces. This one-piece central peg
was designed to provide more durable fixation than had been experienced with separate lateral and
medial components used in earlier condylar designs. The central peg design was an equal group
effort between Chit Ranawat, John Insall, and Peter Walker
the canal beneath, and while this has not caused any problem, it does signify some
loss of cement pressure and reduced penetration around the lower part of the stem.
A surgical technique introduced to prevent that was to use an impactor the same size
as the peg to punch the central bone downwards to form a bone plug. A further
observation has been that in some cases the cement has interfaced with the posterior
tibial cortex, providing a rigid axial support to the distal part of the peg. This would
tend to unload the bone at the lateral and medial sides, causing stress protection and
bone resorption. However for an all-plastic component, this effect would be small
due to the low stiffness of the plastic, but could certainly be a problem with metal-
backed components. Despite the deficiencies, fixation of the central peg tibial com-
ponents has been very reliable in the large majority of cases.
All-Plastic vs Metal-Backing
By the late 1970s, most artificial knee systems, where central pegs were used for
fixation, had shifted away from all-plastic components to metal-backed compo-
nents with tibial inserts. The main reason was the advantages of modularity and
the ability to add extension stems to the metal-backing. There was also the gen-
eral assumption that the metal-backing would transmit the forces more uni-
formly to the proximal tibial surface, thus avoiding areas of high stress beneath
the lateral and medial contact points which could occur with all-plastic compo-
nents. This seemed both intuitive and self-evident which probably accounts for
the lack of need to carry out any verification. Nevertheless, around 1980 there
was a sudden interest in using finite element analysis to justify the widespread
use of metal-backing (Bartel et al. 1982; Lewis et al. 1982; Murase et al. 1983).
A typical finding was: “…in general, metal components provide lower system
stresses than polyethylene components, particularly in the methacrylate and the
cancellous bone, and that one-piece designs gave lower cement-bone interface
stresses compared with those with separated condylar components” (Lewis
et al. 1982).
As a result of such data, all-plastic components were used mainly for older
and less active patients. But surprisingly, several recent long-term follow-up
studies have shown that the survivorship is at least as good as for metal-backed
components (Houdek et al. 2017): “…in a large cohort of patients…all-polyeth-
ylene components significantly outperform modular and monoblock metal-
backed components in terms of survival…across all age groups and a majority of
BMI’s.” This outcome data seems to contradict what would be expected from the
theoretical and experimental studies cited above. One possible explanation is that
when plastic inserts are snapped into a metal-backing, micromotion occurs at the
interface due to the cyclic forces applied, generating backside wear of the plastic
(Sisko et al. 2017). The microscopic particles so generated have been implicated
in causing osteolysis of trabecular bone at the cement-bone interface, eventually
leading to loosening. This explanation does not however apply to monoblock
metal-backed components where there is no escape route for wear particles.
Another possible explanation is that the stresses with metal-backing, spread over
a wide area, are much lower than in the normal anatomic knee in certain areas,
causing stress protection with resulting bone resorption. A stiff metal central peg
would exacerbate this effect, especially if the tip was in close proximity to the
posterior cortex of the tibia.
But whether all-plastic or metal-backed, the one-piece tibial component
with a central peg has been successful. In a review of no less than 27,000 cases
of artificial knees performed since 1999, the Swedish registry data showed a
survivorship of the tibial components, whether metal-backed or all-plastic with
central peg fixation, to be 97% at 10 years follow-up (Gudnason et al. 2014).
This is not to say that other design configurations did not have similar success.
A bi-cruciate preserving Townley design, where the tibial component consisted
of a plastic insert in a metal-backed anteriorly bridged construct, was followed
METAL UHMWP
35 mm CENTRAL POST
METAL UHMWP
70 mm CENTRAL POST
Fig. 8.6 Finite element analysis was used to compare the maximum stresses in the bone adjacent
to the tibial components. This study compared the stresses with an all-plastic component and with
a short and a long central peg. Eccentric loading was chosen to model varus loading. For the short
peg, the flexible plastic produced high stresses under the load, whereas the stresses for the metal-
backing were much less. For a longer central peg, however, the metal back produced stress protec-
tion at the tibial surface under the load, due to transfer of high loads at the lower end of the peg.
Analyses such as these accelerated the shift from all-plastic to metal-backed components starting
in the late 1970s (Murase et al. 1983). (Reprinted with permission from Journal of Biomechanics)
for 23 years in 214 knees (Pritchett 2015). The surgeon, Jim Pritchett, was an
expert in the use of this device. There were only 18 failures of the tibial com-
ponent for all causes including excessive wear, loosening, and fracture. These
levels of success of the central peg type of tibial component, as well as certain
other designs, have been the result of a steady improvement in design, tech-
nique, and materials. The application of basic engineering principles, design
methodology, and experimental, and theoretical and clinical studies, consistent
alignment, attention to adequate cement penetration, and retrieval analysis,
have all played a role.
TOTAL CONDYLAR FREEMAN-SWANSON
KINEMATIC POROUS-COATED ANATOMIC (PCA)
Fig. 8.7 Wear and damage of plastic inserts seen on retrieved tibial components of early artificial
knees. Total Condylar, large fatigue crack posteriorly resulted in corner dissociation. Some delami-
nation medial side. Freeman-Swanson, internal-external rotation about the center has produced
bow-tie delamination wear. Kinematic condylar, excessive posterior-lateral contact. PCA predomi-
nantly anterior delamination wear, possibly due to lack of posterior slope
Bone Strength Under Component
For all tibial component designs, the load is transmitted between the base of the
component and the underlying trabecular bone. The strength of this bone varies con-
siderably across the surface and with depth. The strength is highest over patches
medially and laterally, reflecting the regions of highest load transmission. As men-
tioned above, the strength in the central region is the lowest. The load distribution
across the surface will be in proportion to the stiffness which will be related to
strength. Load follows stiffness. Hvid and Hansen (1985) demonstrated that the
strength of the bone falls off rapidly with depth below the surface of the tibia. Hence
it is an advantage to minimize the depth of the bone resection. At a certain level, the
trabecular bone will be unable to withstand the stresses which are imposed upon it,
suggesting the minimum level of resection. Using CT scanning, Hvid also studied
whether the bone density changed over time after an artificial knee. He studied cases
where the knee had been in varus prior to the procedure, and hence the medial bone
was more dense than normal and the lateral bone less dense. It would be expected
136
MED
100
100
4mm 8mm
75
75 POST
50
50
25 25
0 0
BONE STRENGTH (MP a)
BONE STRENGTH (MP a)

0 0
L
10 10
10 AT
10 AT
20 )
ER
20 )
ER
A
A
30 (mm 30 (mm
20 L (m
20 L (m
40 IOR
LAT ANT
30
m
40 IOR
30
m
)
50 TER 50 TER
60 AN 60 AN
40
40
Fig. 8.8 The strength of the trabecular bone across the tibia as a function of depth below the surface. The 4 and 8 mm are depths below the bottom of the
subchondral plate which was removed. The locations of highest strength approximate to the bone under the menisci. The strength reduces considerably with
depth. The white dense bone on the proximal cut surface corresponds to the highest strength (Hvid and Hansen 1985). (Reprinted with permission from Journal
of Orthopaedic Research)
8 Tibial Component
that when the alignment was corrected at surgery, this would restore the normal lat-
eral to medial force distribution, causing the bone density to revert to normal. While
this did in fact occur, it took a considerable time, more than expected. The reduction
in strength with depth has a particular significance to unicompartmental knees, where
the surface area is relatively small. The knee that Charnley designed, the Load Angle
Inlay (described in Chap. 9), used a thin metal plate for the tibial component, which
had the advantage of being seated on the strongest bone possible.
Peripheral Shape
One of the issues in tibial component design has been the peripheral shape; is a sym-
metric shape satisfactory, or would an asymmetric shape have advantages? The aim
should be to minimize the stresses on the proximal tibia by obtaining as close a
contact to as much of the periphery as possible, thus maximizing the percentage
coverage. At the same time, there should be no overhang, especially posterolaterally
where popliteus irritation can occur or medially which can irritate the medial col-
lateral ligament. One of the first studies to investigate coverage, the fit of asymmet-
ric and symmetric components was measured on the resected bone pieces from
surgical cases (Westrich et al. 1994). The peripheral fits were found to be from 81
to 85% with no difference between symmetric and asymmetric components. A more
elaborate study was carried out on the fit of contemporary designs based on CT
scans at the resection level of 100 surgical cases (Meier et al. 2018). The compo-
nents were from 7 to 10 size systems and included symmetric and asymmetric
designs. The components were fitted based on the rotational axis: the ML axis, the
medial third axis, or the rotation of best fit. Their conclusions were: “Tibial cover-
age is similar for symmetric and asymmetric designs, but depends on the rotational
landmark for which the implant is designed. The surgeon has the option to improve
tibial coverage by optimizing placement between the two anatomic rotational align-
ment landmarks, the medial 1/3 and the ML-axis.” Most of the percentage coverage
values were between 79 and 88%, except for one design aligned along the medial
third axis which dipped to 73%. In all cases, the standard deviations indicated that
there would be an incidence of undercoverage. The ideal situation is that the sur-
geon obtains the best fit regardless of rotation. This would be acceptable for a rotat-
ing platform knee, whereas for other knee designs, it would depend how much
rotational laxity was built in. For practical reasons, there should be enough laxity to
accommodate variations in placement and between individuals, although the exact
value of laxity needed is unknown.
What Are the Best Configurations for Durability?
The most durable fixation is obtained with components which cover the entire prox-
imal tibia or which have only a small area of relief for retention of the posterior
cruciate ligament. A central post has been shown to provide durable fixation even if
the lateral and medial plateau areas become debonded. Two side posts, or keels,
transmit rotational forces for added fixation. All-plastic components and metal-
backed components have been equally durable. All-plastic components are not indi-
cated if the underlying bone is deficient, such as in cases of severe deformity. A
plastic insert must fit into the metal-backing tightly to minimize micromotion which
causes backside wear, generating small wear particles which may lead to areas of
osteolysis. Minimizing the depth of cut will result in the component being seated on
stronger bone, but care must be taken to still achieve the required cement penetra-
tion. Whether the cement needs to bond rigidly to all the surfaces of the tibial com-
ponent needs further research.
References
Abernethy PJ, Robinson CM, Fowler RM. Fracture of the metal tibial tray after kinematic total knee
replacement. A common cause of early aseptic failure. J Bone Joint Surg Br. 1996;78(2):220–5.
Ahir SP, Blunn GW, Haider H, Walker PS. Evaluation of a testing method for the fatigue perfor-
mance of total knee tibial trays. J Biomech. 1999;32(10):1049–57.
Attenborough CG. Total knee replacement using the stabilized gliding prosthesis. Ann R Coll Surg
Engl. 1976;58(1):4–14.
Bartel DL, Burstein AH, Santavicca EA, Insall JN. Performance of the tibial component in total
knee replacement. J Bone Joint Surg Am. 1982;64(7):1026–33.
Behrens JC, Walker PS, Shoji H. Variations in strength and structure of cancellous bone at the
knee. J Biomech. 1974;7(3):201–7.
Cloutier JM. Results of total knee arthroplasty with a non-constrained prosthesis. J Bone Joint
Surg Am. 1983;65(7):906–19.
Eftekhar NS. Adjustable intramedullary replacement of the knee: evolution of surgical technique
and prosthesis. Clin Orthop Relat Res. 1978;(137):235–43.
Freeman MA, Swanson SA, Todd RC. Total replacement of the knee using the Freeman-Swanson
knee prosthesis. Clin Orthop Relat Res. 1973;(94):153–70.
Gudnason A, Hailer NP, A WD, Sundberg M, Robertsson O. All-polyethylene versus metal-backed
tibial components-an analysis of 27,733 cruciate-retaining total knee replacements from the
Swedish knee arthroplasty register. J Bone Joint Surg Am. 2014;96(12):994–9.
Houdek MT, Watts CD, Wyles CC, Martin JR, Trousdale RT, Taunton MJ. Metal or modularity:
why do metal-backed tibias have inferior outcomes to all-polyethylene tibial components in
patients with osteoarthritis. J Arthroplast. 2017;32(3):836–42.
Hvid I, Hansen SL. Trabecular bone strength patterns at the proximal tibial epiphysis. J Orthop
Res. 1985;3:464–72.
Lewis JL, Askew MJ, Jaycox DP. A comparative evaluation of tibial component designs of total
knee prostheses. J Bone Joint Surg Am. 1982;64(1):129–35.
Meier M, Webb J, Collins JE, Beckman J, Fitz W. Do modern total knee replacements improve
tibial coverage? Knee Surg Sports Traumatol Arthrosc. 2018;26:3219–29.
Murase K, Crowninshield RD, Pedersen DR, Chang TS. An analysis of tibial component design in
total knee arthroplasty. J Biomech. 1983;16(1):13–22.
Paganelli JV, Skinner HB, Mote CD. Prediction of fatigue failure of a total knee replacement tibial
plateau using finite element analysis. Orthopedics. 1988;11(8):1161–8.
Pritchett JW. Bicruciate-retaining Total knee replacement provides satisfactory function and
implant survivorship at 23 years. Clin Orthop Relat Res. 2015;473(7):2327–33.
Ranawat CS, Shine JJ. Duo-condylar total knee arthroplasty. Clin Orthop Relat Res.
1973;94:185–95.
Sisko ZW, Teeter MG, Lanting BA, et al. Current total knee designs: does baseplate roughness
or locking mechanism design affect polyethylene backside wear? Clin Orthop Relat Res.
2017;475(12):2970–80.
Skolnick MD, Coventry MB, Ilstrup DM. Geometric total knee arthroplasty. A two-year follow-up
Sledge CB, Ewald FC. Total knee arthroplasty experience at the Robert Breck Brigham hospital.
Swanson SA, Freeman MA, Heath JC. Laboratory tests on total joint replacement prostheses. J
Townley CO. The anatomic total knee resurfacing arthroplasty. Clin Orthop Relat Res.
1985;(192):82–96.
Walker PS, Hsieh HH. Conformity in condylar replacement knee prosthesis. J Bone Joint Surg.
1977;59(2):222–8.
Westrich GH, Laskin RS, Haas SB, Sculco TP. Resection specimen analysis of tibial coverage in
total knee arthroplasty. Clin Orthop Relat Res. 1994;309:163–75.
Condylar Replacements
What Were the Steps Leading to the Basic
9
Condylar Replacement Knees?
When the cemented metal-polyethylene artificial hip joint began to show promising
results in the mid-1960s, it was only a matter of time before the same materials
would be used for the design of an artificial knee. From the start, it was realized that
there were significant design problems involved. In the hip, there had been many
years of experience with components which replaced the femoral head and the ace-
tabulum. For the artificial hip, it was a matter of combining the components in such
a way that the capsular structures still provided sufficient stability. Charnley
achieved this by removing the greater trochanter at the start of the procedure and
refixing it in a tightened position after the components were cemented in place. The
ball-in-socket joint itself provided considerable stability with a single rotation point.
For the knee, a similar problem needed to be solved, how to tension the ligaments
such that stability would be restored. However the bearing surfaces were noncon-
forming, allowing both sliding and rotation to occur with little inherent stability. In
addition the arthritis had usually caused an alignment deformity at the knee, with
damage to the condyles and the ligaments. These were not easy problems to solve,
but by the end of the 1960s, solutions had been reached. The first two designs, the
Gunston and the Freeman-Swanson, could not have been more different.
Gunston Polycentric Knee
Frank Gunston FRCS started his career as an engineer before switching to medicine
and then took a fellowship with Sir John Charnley at the Hip Center at Wrightington
Hospital in England to work on artificial hip joints. Noting that many of the hip
patients had severe knee problems which prevented them from returning to normal
function, Gunston was motivated to design an artificial knee, the idea receiving
Charnley’s support. The rationale for the knee design was: “The biomedical prin-
ciples and experience gained from total hip arthroplasty were combined with an
analysis of normal knee movement. The design objective was to provide an articular

142 9 Condylar Replacements
INSTANT CENTRES
0˚
0˚ 0˚
20˚
20˚
50˚
20˚
80˚
50˚
120˚
Fig. 9.1 The Gunston Polycentric knee was based on restoring the bearing surfaces only, retaining
both cruciate ligaments, and providing sufficient laxity between the femoral and tibial components
to allow for the changing instant centers of rotation during flexion. The femoral component was
cemented into narrow slots cut in the femoral condyles. The tibial components were set at the cor-
rect heights and cemented into slots in the tibial condyles (Gunston 1971). (Reprinted with permis-
sion from Journal of Bone & Joint Surgery B)
replacement procedure that would simulate the normal anatomic and functional
modalities of the knee joint as closely as possible” (Gunston 1971). Gunston stated
that from extension to 10 degrees flexion, the tibia internally rotated, analogous to
the “screw-home” movement shown in early anatomic studies, and from 10 to 20
degrees flexion the femur rolled posteriorly on the tibia, after which the same con-
tact points were maintained until full flexion. This constituted a polycentric path-
way for the instant centers of rotation, which gave the name to his design, the
Polycentric knee. Hence Gunston emphasized that the motion of the knee was not
limited to the sagittal plane and that there were different motions on the lateral and
medial sides. The idea that the artificial knee should be designed to allow for vari-
able centers of rotation and contact point locations was fundamental.
The femoral component as seen in the sagittal view was an arc of constant radius,
in contrast to the anatomic shape where the distal femur was of larger radius than
the posterior, particularly on the medial side. This compromise was made because
the component was made easily in Charnley’s Biomechanical Laboratory. The nar-
row femoral disc in the wider plastic tibial runner allowed for 20 degrees of internal-
external rotation as well as freedom of anterior-posterior displacement. The heights
of the tibial components above their respective joint surfaces were adjusted at sur-
gery to provide the correct ligament tensions and allow for a full range of flexion.
Gunston did not elaborate on how the overall leg alignment was evaluated. In
essence, the Polycentric knee aimed to restore the original profiles of the bearing
surfaces, with appropriate ligament tensions, in an effort to restore the normal
kinematics.
While only small numbers of the Polycentric knee were carried out at
Wrightington from 1968 to 1969, after Gunston’s return to Winnipeg in 1970, sur-
geons at the Mayo Clinic started using the knee after making modifications to widen
the bearing surfaces. Up to 1973, Mark Coventry and Richard Bryan carried out 713
cases but found that the technique could be exacting and difficult, mainly due to the
9 Condylar Replacements 143
lack of instrumentation (Coventry & Bryan 1974). This explained the problems they
encountered of postoperative instability and misalignment of the limb. Ed Chao
PhD, Director of the Biomechanics Laboratory at the Mayo Clinic, while not
involved with the design itself, studied the wear and deformation of the polyethyl-
ene as well carrying out gait analysis on patients. Chao’s interest was in developing
basic science methods for the clinical evaluation of artificial knees. His elaborate
methods of analysis of the gait patterns were a model for future studies in other
laboratories. Longer-term results were published of the earliest Polycentric cases,
showing that only 66% could be considered successful at 10 years follow-up, the
main problem being instability and tibial loosening, with some cases of excessive
wear and femoral condylar fractures (Lewallen et al. 1984). The tibial loosening
was evidently due to excessive stresses on the underlying trabecular bone, some-
thing that Freeman and Swanson had drawn attention to.
Freeman-Swanson Knee
While Gunston was located at Wrightington Hospital in a small village in the north-
west of England, Michael Freeman FRCS and Professor Alan Swanson PhD were at
the London Hospital and Imperial College, affiliated with the University of London.
They had set up a Biomechanics Unit in the Department of Mechanical Engineering
in 1965, with the major purpose of developing an artificial knee. Their first implant
design was relatively simple, consisting of a circular-sectioned metal femoral com-
ponent and plastic tibial blocks shaped after the MacIntosh plateau, on the lateral
and medial sides. After a small number of cases, it was found that obtaining correct
leg alignment and adjusting the heights of the separate tibial plateaus were too dif-
ficult and time-consuming, especially for knees with severe varus or valgus defor-
mities. They soon produced the new Freeman-Swanson design based on the
following design criteria (Freeman et al. 1974):
• Stability from the bearing surfaces and the collateral ligaments, with excision of
the cruciate ligaments.
• Any degree of deformity can be corrected.
• Soft tissue releases to be carried out of the collateral ligaments to correct for
varus or valgus deformities.
• Large areas of contact between the components and the cut surfaces of the bone
to minimize bone stresses.
• Large contact areas on the bearing surfaces to minimize wear.
• Full width of the bearing contact seen in the frontal plane to maximize varus-
valgus stability.
• Square bone cuts to facilitate alignment to the long axes of the femur and tibia.
The main concern was to design an artificial knee that was mechanically sound,
with a simple reproducible technique that could be used by any surgeon to treat cases
of severe arthritis. Freeman and Swanson claimed that their design had major
Fig. 9.2 Michael A R

Freeman FRCS, at the first
European Federation of
Orthopaedics &
Traumatology EFORT
meeting held in Paris in
1993. He was one of the
founding members of the
organization. Freeman and
Alan Swanson PhD
developed one of the first
metal-plastic artificial
knees, in London, UK, in
the late 1960s. Freeman
was an active figure at
conferences around the
world for many decades.
(© Peter S Walker)
advantages over both the fixed-axis metal hinges that were still being widely used and
condylar designs where separate tibial components were used medially and laterally.
For such designs, the surface area of fixation of the components to the bone was too
small, which could cause compressive failure of the supporting bone. In addition, flat
or shallow bearing surfaces would produce excessive contact stresses and wear.
From the first cases in 1970, the Freeman-Swanson knee was used by an increas-
ing number of surgeons such that over 1000 cases had been carried out by the time
of the Institution of Mechanical Engineers Conference on Total Knee Conference in
1974. The majority of the patients were said to be “free of complaint,” but few other
details were available. In a retrospective review (Bargren et al. 1983), it was reported
that the major determining factor for success or failure was the overall leg align-
ment. For cases that were in valgus, relative to an in-line femur and tibia, 89% were
Fig. 9.3 Professor Alan

Swanson PhD and Michael
Freeman FRCS designed
the Freeman-Swanson
knee in the late 1960s and
went on to design several
improved versions. Alan
Swanson was a Professor
at Imperial College
London. He now devotes
much of his time to
locating and refurbishing
woodworking tools for an
organization which sends
these to disadvantaged
regions of the world.
(Photo of Alan Swanson
taken during an interview
in 2017 at the Institution of
Mechanical Engineers in
Birdcage Walk, London,
England) (© Peter S
Walker)
successful. For varus cases, only 14% were successful. This can be attributed to the
early surgical technique where the distal femoral and proximal tibial cuts were
made perpendicular to the long axes of the bones themselves. In the case of the
femur, the cut should have been at 85 degrees, providing the normal 5 degrees of
valgus. This could explain the tendency to varus alignment which would have
caused excessive loading on the medial side.
By 1974, Freeman and Swanson had designed two other versions of the knee, one
with a patella flange, the other with slots for retention of the cruciate ligaments. Yet
other design variations were introduced in succeeding years, based on limitations of
preceding designs, as well as incorporating entirely new ideas. This process, while
progressing the design itself, had the negative effect of making it difficult to conduct
orderly clinical follow-up studies. In addition, other surgeons became uncertain
about which design they should be using, which resulted in some loss of credibility
even though a number of the design features were both innovative and beneficial.
Kodama-Yamamoto Knee
While this work on artificial knee design was proceeding in England, in the late
1960s, development was taking place in Okayama in Japan (Yamamoto 1979). Drs.
Kodama and Yamamoto, from the University Department of Orthopaedics, used a
Sbarbaro femoral shell (Sbarbaro 1974) in combination with a horseshoe-shaped
polyethylene tibial component such that the cruciate ligaments could be preserved.
The interior surface of the femoral shell fitted against surgically shaped distal femo-
ral condyles, fixation being with thin metal fins. The shell included the patella
flange. The tibial component was fixed anteriorly with two staples, preventing ante-
rior lift-off. Although they were aware of the availability of acrylic cement, they
considered that it was harmful to the bone and soft tissue. After some experience
with a congruent tibial design, they stated that “there should be some incongruity
between the femoral and tibial components, with sufficient rotatory movement
allowed in flexion, slightly restricted in extension.” Also, posterior femoral rollback
was necessary to achieve high flexion, essential for Japanese patients. In their fol-
low-up studies starting in 1970, they indicated that there was no problem with the
uncemented press-fit fixation. They did note however that complete coverage of the
tibial surface was necessary to avoid some sinkage, addressed by providing several
sizes. By the time of the 1979 review, the authors reported few problems, including
no loosening or radiolucent lines, implying the uncemented fixation was satisfac-
tory. Even so they made several design changes, partly influenced by Freeman in
London with whom Sumiki Yamamoto spent a Fellowship.
Townley Anatomic Total Knee
Charles O Townley MD, from Port Huron in Michigan, had been experimenting
with anatomically shaped components since 1953, when he started using his metal
tibial resurfacing component. Even before then he had devised a metal femoral
component, but metal-on-metal was not considered to be a viable bearing. Why a
metal-plastic version of Townley’s design was not made until 1972 is surprising,
considering that total hips were made with those materials from the early 1960s, and
by the late 1960s, the Freeman-Swanson and Gunston knees were in widespread
use, although mainly in England. The main reason was the difficulty of finding a
manufacturer who was interested. Nevertheless Townley was quite clear about the
basis of his design: “provide an articular replacement procedure that would simulate
the normal anatomic and functional modalities of the knee joint as closely as pos-
sible” (Townley 1988). His design criteria were:
• Nonconstrained articulation in all planes of motion.

• Reconstitution of the normal synchronized articular motion patterns.
• Minimization of abnormal shear and tension stresses by accurate alignment.
• Thinnest possible implants.
• Total resurfacing including patella.
Fig. 9.4 In an article called “Charles O Townley: Design by Reason,” James Pritchett MD
described the principles behind Charles Townley’s artificial hip and knee designs. The Anatomic
total knee originally had an all-plastic tibial component; later a metal-back was added. An emi-
nence of the bone was preserved in the center of the tibia so that both cruciates could be retained
(radiograph). The joint line was tilted medially by 2–3 degrees, as in the normal knee. To check the
limb alignment, Townley used a cord stretched between a marker over the center of the femoral
head, radiographically placed, and the center of the ankle (Pritchett 1988, 2015). (Photos courtesy
James Pritchett MD, Interview with James Pritchett 2017)
It was obvious that Townley regarded his new Anatomic total knee as a major
advancement over the metal tibial plate he had been using up till then, based on the
numbers used in patients. In the 17 years from 1953 to 1972, he used only 170
plates, about one every 6 weeks. Between 1972 and 1988, he had used 1700 total
knees, an average of two each week. His proposition that good results could only be
obtained by meticulous attention to accurate surgery was expressed in his statement:
“Significant abnormalities in either the size, contour, or placement of the femoral
component will distort or misplace the condylar articular outlines and place it out of
tune with the holding pattern of motion sustained by the ligaments.” His technique
emphasized accurate alignment of the components to achieve the two degrees
medial slope of the joint line with the patient standing, as well as the mechanical
axis (the line joining the center of the hip to the center of the ankle) passing 10 mm
medial to the center of the knee itself. This was tested with a length of rope, possibly
the simplest instrument ever invented! In a long-term follow-up of the Townley
Anatomic Total Knee, James Pritchett achieved 89% survivorship at 23 years fol-
low-up, with a mean flexion angle of 117 degrees (Pritchett 2015). It is not easy to
find any results which surpass these with any design of artificial knee. Undoubtedly
the surgical expertise of Pritchett was a factor in these results. The article entitled
“Design by Reason” (Pritchett 1988) was a strong tribute to Townley, written by a
surgeon who had trained under Townley and was a close colleague for his entire
surgical career. Pritchett himself was an innovator in the field of artificial knees, in
carrying out extensive clinical studies of bilateral cases to determine which designs
gave the best functional results (Pritchett 2004, 2011).
Geometric and Anametric Knees
The Geometric knee was in marked contrast to Townley’s design. Designed in

1969–1970 by Bob Averill and Alex Khowayla, design engineers at Howmedica,
Rutherford, New Jersey, it focused mainly on engineering principles, minimizing
the contact stresses on the polyethylene and providing complete anterior-posterior
and rotary stability. In these design criteria, there is a close similarity with those of
Freeman and Swanson. When the Geometric design was close to completion, a
Fig. 9.5 Robert Averill at his home in New Jersey. He and Alex Khowayla, design engineers,
developed the Geometric knee while working at Howmedica, starting in 1969. After an exhibit at
the AAOS in 1972, the knee became widely used. At bottom right, the spherical femoral runners
and matching plastic tibial condylar surfaces can be seen. Later Averill and Khowayla founded
their own company called Osteonics, which became a significant force in orthopedics. (Photo
courtesy Robert Averill)
group of prominent surgeons was invited to form a user’s group, with the main goals
of developing the instrumentation and technique and carrying out initial surgeries.
This was a remarkable way to implement an initial introduction, in that it would
have been easier to identify a single “product champion” from a well-known institu-
tion. Averill chose to form a group for several reasons. They could form a consensus
on the design and recommend modifications. They would be able to develop the
instrumentation and technique, again with a consensus. Finally, there would be a
larger volume of knees used, and a greater exposure at conferences. There was little
competition from other knees at the time; the first implantations were made in 1971.
Consequently the Geometric became widely used in the USA soon after its intro-
duction at the AAOS in 1972.
But even as early as the 1974 Total Knee Conference in London, there were
some problems reported in an article entitled “Technical pitfalls in the Polycentric
and Geometric total knee arthroplasties” authored by Mark Coventry and Richard
Bryan from the Mayo Clinic (1974). The main pitfalls concerned the technique,
mainly misalignments, lack of soft tissue balancing, and improper cementation
(Coventry et al. 1973). These problems are understandable because at that time,
there was little appreciation of the consequences of these factors, and hence the
instrumentation was not designed appropriately. In addition, due to the limited
knowledge of knee forces and kinematics, there was no accommodation in the
design for anterior-posterior and rotary laxity, which evidently led to kinematic
conflict with the cruciate ligaments, producing high tensile and compressive
stresses at the cement-bone interface. Even though the Geometric was subse-
quently modified in several ways by more than one company, it was not regarded
as easy to insert or reliable in function. The group of surgeons recognized that
fundamental changes were needed to the Geometric, and as a result, the anametric
knee was designed (Finerman et al. 1979). The femoral-tibial conformity was
reduced to increase flexion and to reduce the shear forces on the tibial fixation. A
central tibial peg and a patella flange were added. These changes resulted in
improved clinical results compared with the Geometric.
Anatomic Designs of Seedhom and Ewald
Soon after the above developments, efforts were under way in both England and the
USA to design a truly anatomic artificial knee. Bahaa Seedhom, a mechanical engi-
neer who had just completed a PhD at Leeds University in England, was offered a
position at the newly formed “Bioengineering Group for the Study of Human Joints”
which had been formed by Professors Duncan Dowson and Verna Wright. He was
asked to suggest a research topic and chose to design an artificial knee. He spent
much of his time in the Anatomy Laboratory at the Leeds General Infirmary dissect-
ing knees, determining how they functioned and how best to design an artificial
knee replacement. He concluded that the implant should be a thin shell which
required only a removal of a surface layer and hence the bearing surfaces of the
implant had to be close to the anatomic shape. Seedhom realized that the relative
geometry of the femoral and tibial condyles was the determining factor for both
kinematics and stresses (Seedhom et al. 1972). A patella flange was necessary to
provide a smooth unbroken track for the patella, which was known to carry high
forces especially in higher flexion.
The shape of the tibial surface was generated by flexing and extending the femo-
ral component to allow a range of flexion from 0 to 120 degrees, together with suf-
ficient laxity. Seedhom designed the knee based on anatomic considerations and
developed effective shape molding techniques by working with a dental technician
in the Leeds General Infirmary. Molds were used to produce femoral castings in
cobalt-chrome alloy in several sizes, right and left. After polishing of the bearing
surfaces, the appearance was exactly like an anatomic femoral condyle. Similar
meticulous methods were used for designing the tibial components, including a
technique for molding the polyethylene to improve its wear resistance (Seedhom
et al. 1973). After producing prototypes of the final design, he worked with Eddie
Longton, an orthopedic surgeon at Leeds General Infirmary, on the technique of
implanting the components into the knee. With little precedent for collaboration
between engineers and surgeons, Seedhom felt that there was a considerable barrier
to communication, because the surgeon was unfamiliar with even the basic concepts
of mechanics and materials, while Seedhom had little notion of surgical realities.
The Leeds Knee was finally manufactured and first implanted in 1972. However
after about a year, only ten cases had been performed, and the results were not
promising: “The extreme variability and sometimes disastrous results obtained have
necessarily inhibited extensive employment of this prosthesis” (Seedhom et al.
1974). This was unfortunate because the design was more sophisticated than others
at that time, and possibly if it had received sufficient priority, it could have suc-
ceeded. As a result of this situation, the manufacturer changed several times causing
delays and eventually the project became sidelined.
In the same time frame as Seedhom’s work, Fred Ewald MD, a surgeon at the
Robert Breck Brigham Hospital in Boston, developed an artificial knee almost iden-
tical to that of Seedhom, although there was no evidence of any connection between
the two projects. Ewald’s femoral component, a shell with an anatomic surface and
an intramedullary stem for fixation, was modeled after the Massachusetts General
Hospital (MGH) femoral condyle designed in Boston in the late 1950s. However
Ewald’s idea for forming the tibial surface was unique and original. He flexed and
extended the femoral component in a mechanical jig through a container of soft
wax. When the wax hardened, the surface was the most conforming surface possible
which would accommodate the relative motion between the femur and tibia, allow-
ing “…. constantly changing instantaneous centers of rotation during flexion”
(Ewald 1974). Implants were produced by Zimmer after which Ewald performed 23
cases (Ewald 1975). However Ewald decided to discontinue the use of the design
even though the early results were promising. One difference between the Seedhom
and Ewald designs was that the former retained the cruciates and the latter did not.
Even though the appearance of these designs was closely anatomic, which seems
the logical shape to make an artificial knee, it may be more than a coincidence that
neither of these designs became widely used nor were implanted in patients for
more than a few years. At that time, manufacture of the anatomical shapes was dif-
ficult and expensive because the femoral components had to be hand ground and
polished. The tibial components required a molding process which was much slower
than machining. Forming the distal end of the patient’s femur at surgery to be a suf-
ficiently accurate fit, even allowing for cement filling of any gaps, was very chal-
lenging for the surgeon requiring repeated trial and error. Even though reaming had
been used to produce accurate spherical surfaces in the hip, that could not be trans-
lated to the knee, so handcrafting with osteotomes was necessary. Also, ligament
balancing was difficult due to the relatively high femoral-tibial conformity. Finally,
even though the surfaces were anatomic in shape, for a particular knee, there could
still be a mismatch of both size and shape, especially because in Seedhom and
Ewald’s designs, only a few sizes were made of the components.
ospital for Special Surgery (HSS): Duo-condylar, Duo-patella

H
(CR), and Total Condylar
The groundwork for artificial knee design started at the Hospital for Special Surgery
(HSS) in New York during the late 1960s. Lee Ramsay Straub MD, an Attending
Surgeon at HSS, had alerted Chit Ranawat MD to the possibilities of joint replace-
ment, not only for the hip which was already being used but for the knee. They
started by using MacIntosh spacers and used that as a starting point to fabricate a
custom-made metal-plastic unicompartmental knee. This set the stage for artificial
joint developments. During the 1960s, Harlan Amstutz MD had established a
Biomechanics Laboratory to carry out research on materials for artificial hip joints.
Development of artificial joints in general was seen by the Surgeon-in-Chief, Robert
Lee Patterson, to be an area of considerable potential, so it was decided to expand
the staff of the laboratory. There were already several Universities in England where
such work was in progress. Harlan Amstutz visited Leeds University where Peter
Walker was just completing a PhD on the lubrication of joints. Walker was offered
a position at HSS which he started in late 1969. But in 1970, Harlan Amstutz was
appointed Chairman of Orthopaedics at UCLA, and the other biomechanics staff
moved there. On the other hand, Walker felt that California was too far from London,
and he asked the Director of Research at HSS, Robert Mellors, if he could continue
to work in the Biomechanics Laboratory. Accordingly he was given the opportunity
to start a new activity focused on the knee and was assigned support staff.
Coincidentally an orthopedic Resident, Joe Hajek, proposed a project to determine
the contact areas in the knee. Hajek and Walker thought that this work could have
an important application: “The significance of the study can relate to….the design
of total knee prostheses which seek to imitate the load-bearing surfaces in the natu-
ral joint” (Walker & Hajek 1972). Hence the underlying concept was to carry out
basic research first and then use the results for a practical purpose. The study that
was devised was to position a knee specimen into an Instron loading machine, pack
soft acrylic cement inside the joint, apply load, and, when the cement had hardened,
remove the acrylic casting. The castings had lateral and medial apertures showing
Fig. 9.6 Joe Hajek MD, at his home in Sharon CT where he was a practicing orthopedic surgeon
after his residency at the Hospital for Special Surgery in New York. He carried out studies with
Peter Walker on the contact areas in the knee, from which the Duo-condylar artificial knee was
designed in 1971. (Photo taken c 2010) (© Peter S Walker)
the contact areas and were obtained at a range of flexion angles. The castings were
then placed over the tibia and the femur and the outlines of the contact areas drawn
on the condyles (Walker & Hajek 1972). It was then possible to produce shapes
which would be suitable for condylar resurfacing components. Anterior bridges
were included so that the femoral and tibial components were one-piece. These
shapes were used in the design of the Duo-condylar artificial knee manufactured
using cast cobalt-chrome alloy and high-density polyethylene, which started to be
used in 1971.
Ongoing basic research on knee mechanics had confirmed that there needed to be
freedom of motion in anterior-posterior and rotation between the components, with
the stability being provided by the retained cruciate ligaments (Walker et al. 1972;
Walker 1973; Wang & Walker 1973; Wang & Walker 1974). Input in theoretical
kinematics was provided by a collaboration with Professor Barry Wolf from
Brooklyn Polytechnic Institute (now NYU Tandon School of Engineering), includ-
ing the calculation of ligament length patterns from measurement of external refer-
ence points. By now, several surgeons including Chit Ranawat, John Insall, and
Alan Inglis were using the Duo-condylar and were interested in further develop-
ments. Two surgeons from Europe, Yann Masse from Paris, France, and Adrian Bain
from Durham, England, also carried out surgeries and gave valuable input on tech-
nique and follow-up. From this early experience, it was determined that cases with
severe deformities were contraindicated with the Duo-condylar due to the difficulty
Fig. 9.7 Peter Walker

PhD in the Biomechanics
Laboratory at Hospital for
Special Surgery, working
on early knee designs
including the Duo-
condylar and an all-metal-
linked hinge. (Photo taken
1972)
of balancing ligaments. Improvements were needed in several areas. Patellofemoral

resurfacing was required. The fixation of the tibial component needed enhancement.
Fig. 9.8 Alan Inglis MD
was Director of the
Comprehensive Arthritis
Service (CAP) of the
Surgery in the 1970s. He
took part in the early
studies of the kinematics of
the knee before and after
artificial knees. He
participated in the
developments of the
Duo-condylar and
Duo-patella artificial knees
and with the surgical
techniques. (Photo
courtesy Alan Inglis Jr.)
Instrumentation was required to accurately fit the femoral component and achieve
correct alignments (Ranawat et al. 1976).
During this time at HSS, it was evident that artificial knees were a potential
breakthrough in treating the arthritic knee, and Philip D. Wilson, Jr. was assigned to
coordinate this activity. He organized a group consisting of Ranawat, Insall, Inglis,
and Walker. The group held regular meetings, as well as numerous ad hoc meetings
in the Biomechanics Laboratory. Several research projects were initiated and car-
ried out with the aid of orthopedic Residents and Fellows, including Hiromu Shoji,
Jack Shine, Paolo Aglietti, Ching-jen Wang, and Hua Hsieh. Design criteria for a
new design of artificial knee were formulated.
• Applicable to all levels of arthritis.

• Stability in anterior-posterior and internal-external rotation provided entirely by
the condylar surfaces.
• Varus and valgus stability from the collateral ligaments but with stable pivot
points in the condylar surfaces in case of lift-off.
• Normal laxity to provide similar freedom of motion as the anatomic knee.
• Sufficiently large areas of contact to avoid excessive wear.
• A patella flange on the femoral component and a patella resurfacing component.
• A one-piece tibial component which fixed rigidly to the tibia.
• Components requiring “square cuts” on the bone for ease and accuracy of
surgery.
• Instrumentation which would achieve accurate component placement and align-
ment, as well as equal tensions in the ligaments in both extension and flexion.
Some of these criteria were influenced by Michael Freeman, with whom John Insall
kept closely in touch with, having gone through Medical School together in
Cambridge and London in England.
The design of the bearing surfaces of the new knee design, to be called the Total
Condylar, resulted directly from basic research studies of the stability of the knee.
The studies were planned by Peter Walker and carried out by Ching-jen Wang, Yann
Masse, and Hua Hsieh in the Biomechanics Laboratory (Wang & Walker 1974;
Hsieh & Walker 1976). The most important finding was that while the unloaded
knee joint had considerable laxity in anterior-posterior and rotation, the laxity was
substantially reduced under axial load. This was shown to be due to an “uphill
mechanism,” whereby as the femur was displaced from the neutral position, there
was an upward displacement. Hence stability was achieved by gravity forces.
Theoretical analysis was used to define the radii of double-dished partially con-
forming bearing surfaces which provided similar stability and laxity as the anatomic
knee (Walker et al. 1974). The basic principle was that if a combined vertical and
shear force was applied to the knee during an activity, the femur would displace
horizontally on the tibial surface, reaching an equilibrium position. The displace-
ment represented the laxity which occurred. These bearing surfaces provided ana-
tomic stability without the cruciate ligaments although the collateral ligaments, if
taut, would increase their tension due to the vertical displacement of the femoral
components and further reduce the laxity.
The balance between the laxity and the stability was an essential requirement of
the bearing surfaces: less laxity and more stability than provided by shallow tibial
surfaces, but not as constrained as with fully congruent femoral and tibial surfaces.
The sagittal profile of the femoral bearing surfaces was similar to anatomic values.
Calculations were made for the radii of the tibial bearing surfaces which would
result in anatomic laxity and stability. In the frontal plane, the tibial surfaces were
dished to provide medial-lateral stability and to provide a pivot point should lift-off
occur. The final surfaces were termed “double-dished bearing surfaces with partial
femoral-tibial conformity”. With these surfaces, the contact areas on the tibial bear-
ing surfaces were also much larger than in the Duo-condylar and other knee designs
which had relatively flat tibial surfaces, which would result in reduced polyethylene
wear.
The tibial fixation requirement was solved by focusing on how best to resist the
combination of forces and moments, without requiring excess invasion of tibial
Vertical and shear

force components
Sagittal profiles of the Femur
femoral and tibial surfaces in displaces
neutral location horizontal &
vertical
Shear O
force Tibial P
Q
profile
Vertical R
force
ANT T X
S
Contact Displaced position Reaction force
of femur on tibia
Profiles of femoral and tibial surfaces in sagittal and frontal planes
Fig. 9.9 The design of the bearing surfaces of the Total Condylar knee. Based on a theoretical
analysis using force data during activities from Morrison (1970), the radii of the tibial bearing
surfaces were calculated to provide similar AP laxities and internal-external rotations as deter-
mined for the anatomic knee in laboratory experiments. The stability obtained from the bearing
surfaces was independent of any contribution from ligaments (Walker et al. 1974). (Below,
Reprinted with permission from Institution of Mechanical Engineers, London, England)
bone. The solution was a one-piece all-plastic component with a central peg, which
had a rectangular cross section. The efficacy of the central peg was demonstrated
using a specially constructed nine-channel simulating machine (Walker & Hsieh
1977). The design and testing of the central peg was described in Chap. 8.
The artificial knee design was called the Total Condylar in that it replaced all of
the bearing surfaces. To complete the design, a patella resurfacing component was
required. This was designed by Paolo Agletti MD, a Fellow from Florence, Italy. The
design was based on anatomic studies and mechanical testing. This is described in
Chap. 7. The instrumentation for the Total Condylar was designed mainly by John
Fig. 9.10 The original

Total Condylar knee
designed equally by Peter
Walker PhD, Chit Ranawat
MD, and John Insall MD
in 1973. The design was an
evolutionary step from the
Duo-condylar and
Duo-patella knees. This
knee was designed with
double-dished bearing
surfaces which under
weight-bearing conditions
provided similar laxity and
stability as the anatomic
knee
Fig. 9.11 Chitranjan Ranawat MD in his office at the Hospital for Special Surgery (HSS) in 2019.
In the early 1970s, he was a principal participant in the group who designed some of the first arti-
ficial knees in the USA. Since that time, he has been a major influence in the orthopedic world,
primarily for the knee but also for the hip and other joints. His innovative contributions have been
wide ranging. He has been recognized at HSS for his outstanding teaching contributions and has
been an authoritative participant at major conferences around the world. (© Peter S Walker)
Insall working with his Fellow, Norman Scott, and by Chit Ranawat. They devised
methodologies for achieving square cuts aligned relative to the femoral and tibial
axes. The levels of the cuts were determined by distracting the knee in extension and
flexion such that the collateral ligaments would be tightened. But despite the fact that
the tibial bearing surfaces could provide all of the necessary stability without the
cruciate ligaments, Chit Ranawat and Alan Inglis still preferred to retain the cruciates
in cases with only small amounts of deformity, which applied to many rheumatoid
arthritis cases. However the one-piece tibial component with a central peg was
incompatible with cruciate preservation where a long central slot would be required.
A compromise was made by preserving only the posterior cruciate, requiring only a
small notch at the posterior center of the tibial component, just behind the peg. This
was considered an acceptable solution because the posterior cruciate was thought to
cause posterior femoral rollback in flexion, necessary for obtaining a high flexion
range, whereas the anterior cruciate played a less important role.
The end result was that two new artificial knees were designed, the Total Condylar
where the cruciates were resected and the Duo-patella which retained the posterior
cruciate. The Duo-patella represents the CR type of artificial knee. The posterior
tibial surface in the Duo-patella was made shallower than the Total Condylar to allow
the extra posterior displacement of the contact points in high flexion. Each of these
designs was very successful clinically in long-term follow-up, although the Total
Condylar became the more widely used (Insall et al. 1976; Sledge & Ewald 1979;
Ranawat & Boachie-Adjei 1988; Rodriguez et al. 2001). In the late 1970s, these two
types of knees, posterior cruciate retaining (CR) and cruciate sacrificing (TCP),
would be combined with the Duo-condylar and Unicondylar knees and a
Fig. 9.12 John Insall in

the operating room at the
Surgery in c 1974.
Working with Fellows
including Paolo Aglietti
and Norman Scott, John
Insall was one of the main
developers of the
instrumentation for the
Total Condylar knee
replacement. He later went
on to design the posterior
stabilized knee with Al
Burstein. His unique
legacy in knee replacement
is well-known in
orthopedics. (Photo
courtesy The Knee
Society)
Stabilo-condylar design, to form a “Graduated Knee System,” the idea being that
such a range was the solution to satisfying the full scope of severities of arthritis.
The role of the manufacturers was important in the introduction of these artificial
knee designs. In the early phase, Walker had chosen a small company in Boston,
called Cintor (later acquired by Johnson and Johnson), to be the manufacturer.
However their lack of in-house metal casting and other processes led to frequent
supply delays. To solve this problem, Howmedica, located in New Jersey close to
New York, was asked to manufacture the Total Condylar. Howmedica gave the knee
a high priority in their marketing efforts. As a result, the Total Condylar rapidly
gained in clinical use, although the Graduated Knee System still retained a substan-
tial following. This reflected the general view of the surgical community that cruci-
ate retaining and sacrificing knees were both needed to treat a range of patients.
Other Designs of the 1970s and beyond
A number of other designs played an important role in the early development of the
basic condylar replacement knee. The UCI design by Ted Waugh, used from 1972,
was a condylar replacement without a patella flange. The problem of accurately
matching the metal crossbridge, joining the lateral and medial condylar to the femo-
ral trochlea, was solved in an ingenious way. A thin triangular metal blade joined the
condyles and was imbedded in the bone, creating only a small surface discontinuity,
but substantially enhancing the fixation. Late in the 1970s, the first porous-coated
knees were introduced, the Miller-Galante and the PCA, both cruciate-preserving
with low femoral-tibial conformity. Another “first” was the low-contact-stress
(LCS) mobile-bearing knee designed by Buechel and Pappas to be discussed in
Chap. 11. Also at that time, the posterior stabilized (PS) knee was designed by John
Insall and Al Burstein, the design described in the next chapter.
It can be seen that during the 1970s, the condylar replacement artificial knee had
evolved into three separate paths: designs with low femoral-tibial conformity which
preserved both cruciates; moderate femoral-tibial conformity, preserving only the
posterior cruciate; and designs with moderate to high conformity where only the
collateral ligaments were preserved. In the first type, the cruciates provided anterior-
posterior stability, while the bearing surfaces allowed freedom of motion. In the
second type, the posterior cruciate prevented excessive anterior femoral displace-
ment and produced progressive femoral rollback in flexion. In the third type, the
bearing surfaces provided both anterior-posterior and rotary stability, but there was
no specific mechanism for inducing femoral rollback in flexion. The first type usu-
ally used several small pegs for tibial fixation, the second and third types, a large
central peg. Overall the clinical results of the three types were similar. The range of
flexion was higher in the cruciate retaining designs, which achieved averages of 100
to 110 degrees. However these designs showed more instability due to the low con-
straint and the difficulty of accurate ligament balancing at surgery. Tibial loosening
was more prevalent in the designs with only small fixation pegs, whereas the central
peg proved to be more durable. The lower constraint designs showed more polyeth-
ylene wear due to the higher contact stresses and increased sliding. Another factor
claimed by some surgeons was that cruciate preservation and lower constraint led to
a more natural feel to the knee. With these different factors at play, by 1980, it was
not possible to definitively favor one type of design over the other. But the achieve-
ment of the 1970s was to produce design types that worked well, covered a wide
span of indications, and were a sound basis for any future developments.
Combining the different design concepts into one system was a reasonable step
at that time. In the mid-1970s, Peter Walker worked in Boston for a year prior to
working at Howmedica. Surgeons at the Robert Breck Brigham Hospital including
Clem Sledge, the Department Chairman, Fred Ewald, Bill Thomas, and Bob Poss
invited Peter Walker to join clinical meetings and expressed an interest in further
artificial knee developments. By the time Walker started working at Howmedica
soon after, it had been decided to design a unified system, to be called the Kinematic
Knee System. This would include an ACL + PCL, a PS, and a CR, as well as a
stabilizer. Fred Ewald was the main collaborator on this project, including the
design of the instrumentation, with an all-in-one cutting guide called the Gobot.
This instrument included distraction arms for tensioning each sides of the joint
equally. By the late 1970s, the Kinematic had become Howmedica’s flagship knee,
and the system became popular in the USA and in Europe. This relationship between
Howmedica and the Brigham and Women’s Hospital progressed further in the
1980s. The same group, using computer modeling and biomechanical testing, pro-
duced the Kinemax Knee System, a much improved Kinematic.
Tibial
surface
generated
Femoral
to allow
surface
anatomic
anatomic
motion
Fig. 9.13 Frederick Ewald MD worked at the Robert Breck Brigham Hospital in Boston, and then
at the Brigham & Women’s Hospital, affiliated with the Harvard Medical School. To the right is the
Ewald Knee, an innovative anatomic design, US patent filed in 1971 and issued in 1974. Later,
Ewald worked with Clem Sledge, Bill Thomas, Bob Poss, and Peter Walker, on the Kinematic and
the Kinemax Knee Systems. He was one of the founding members of The Knee Society and served
as President. He devised the radiographic evaluation system for the Knee Society. He was regarded
as a particularly skilled surgeon and teacher, benefiting hundreds of Residents and Fellows. (Photo
taken in 1980s) (© Peter S Walker)
Two other surgeons at the Brigham and Women’s Hospital in Boston, who were
not fully involved with the above developments, had new ideas for implant design
and surgical technique. Richard Scott and Tom Thornhill developed the Press-Fit
Condylar (PFC) knee with Johnson & Johnson, the first major knee system of that
company. One of the features was a bone-preserving cruciate-shaped tibial peg on
the metal baseplate for fixation. As a result of the clinical success, the PFC became
one of the most widely knees used in the USA (Martin et al. 1997). The long-term
results reported by an independent hospital were outstanding up to 25 years of fol-
low-up (Patil et al. 2015). Another knee system of that era was the Vanguard knee,
designed by a group of surgeons led by Adolph Lombardi and Keith Berend. The
manufacturing company was Biomet based in Warsaw Indiana. The President of the
company was Dane Miller, who graduated with a PhD in Materials Science at the
University of Cincinnatti in the mid-1970s. Due to his background, he was particu-
larly involved with this knee project. The designers had two particular goals: to
design a knee which would continue to perform well into the second decade and
provide a “continuum of constraint” so that each patient could be fitted with the best
Fig. 9.14 Richard Scott

MD and Tom Thornhill
MD, who worked at the
Brigham and Women’s
Hospital in Boston until
recently, designed the PFC
knee system in the 1980s.
The PFC became one of
the most widely used knee
systems worldwide. That
design has now been
upgraded to produce the
Attune system. Richard’s
recent book Total Knee
Arthroplasty has been a
major success, covering his
long experience of knee
replacement surgery.
(Photo taken 2016 at ISTA
Boston, MA) (© Peter S
Walker)
possible combination of laxity and stability for their condition. The first goal was
eventually realized after a recent follow-up study (Crawford et al. 2019). The sec-
ond was that tibial inserts with different slopes and conformities were available
using the same tibial baseplates, allowing choices to be made at surgery.
It is fitting to complete this chapter with a little-known artificial knee designed
by John Charnley himself in the early 1970s. Having been closely involved with
the efforts of Gunston on the Polycentric knee, Charnley felt that there were two
main deficiencies. Firstly, cutting large slots in the femoral condyles was too inva-
sive, and secondly the tibial components were of insufficient surface area and
were subject to loosening. Nevertheless he thought that the concept of small
adjustable components was viable. He made the innovative step of reversing the
materials, plastic for the femoral component and metal for the tibial component.
The femoral component was a plastic runner embedded in a slot in each femoral
condyle. Charnley had observed that the arthritic lesions on the femur were most
frequently located on the distal condyles only and did not extend posteriorly.
Therefore the femoral component was limited in its length so that it embedded
only in the distal femoral condyle, which made the component much less invasive
and easier to fit accurately. The tibial component was a flat plate with a mesial tab
which seated on a minimally resected tibial plateau. Because the plate was less
than 3 mm in thickness, only a thin tibial resection was required. This preserved
the strongest bone near to the surface for fixation and support of the tibial plate.
As in the hip, acrylic cement was used for fixation. Simple but accurate instru-
mentation was designed including routers for the distal femoral cavity. Hundreds
of cases were carried out at Wrightington Hospital, mostly in rheumatoid arthritis
patients, when both the lateral and medial sides were resurfaced. The knee was
called the load angle inlay (LAI). In 2004, at a bioengineering conference held in
Leeds, England, Bodo Purbach presented a 20–28-year follow-up of no less than
Fig. 9.15 The load angle inlay knee designed by John Charnley in the early 1970s. It was more
conservative than the Gunston Polycentric designed just a few years earlier. The plastic femoral
component resurfaced only the distal end of the femur, the most frequent site of the arthritic
lesions. The metal tibial plates required only a small bone resection. When correctly aligned these
implants survived several decades in many patients. (Photos courtesy Martyn Porter FRCS,
Wrightington Hospital, England)
746 cases performed by Charnley and other surgeons including Michael

Wroblewski (Purbach et al. 2004). The conclusion was “the LAI has given long-
lasting success in cases where the leg has stayed in physiological alignment.”
Only 22 cases were reported to have failed. Most of the failures occurred due to
the leg being in excess varus, which caused wear of the femoral component and
loosening of the tibial plate. Subluxation sometimes occurred due to the plates
being flat without any medial-lateral restraint. It was unfortunate that a full paper
on this extensive experience was not published. The only publication was of a
study of the failed cases, the authors being Julian Minns PhD, a bioengineer from
Durham, England, and Kevin Hardinge FRCS, a surgeon from Wrightington
(Minns & Hardinge 1983). The last LAI’s were implanted in 1979 after which the
knee was discontinued. But to this day, patients with a well-functioning LAI still
visit the Wrightington hospital for a long-term checkup.
The load angle inlay design was not pursued beyond the initial cases performed
at Wrightington Hospital, possibly due to the perceived limited indications. However
unicompartmental knees have retained a place in knee replacement since about
1970. The earliest designs were the St. Georg Sledge from Hamburg, Germany, the
unicondylar from HSS, and the Marmor from Los Angeles, California. Jean-Noel
Argenson from Marseilles in France was one of the first to explore the possibilities
of modular components for the lateral and medial sides as well as the patella troch-
lea (Parratte et al. 2015). After many years of experience, they found disappointing
results except for standard medial unicompartmental, mainly due to design and fixa-
tion issues of the modular components. With an improved design and the use of the
OrthoPilot Navigation System, improved results were obtained (Confalonieri &
Manzotti 2005; Biazzo et al. 2018). In recent years there have been efforts to include
compartmental knees as a way to achieve a “custom” solution, where choices are
made to replace the patella trochlear, the medial compartment, and the lateral com-
partment (Romagnoli et al. 2015). Based on gait parameters, simultaneous bi-unis
were recommended rather than staged (Marullo et al. 2019). In all of these studies,
it was evident that a limitation of the system was that there were not a large number
of cases indicated, and the surgery was technically demanding, factors which have
limited widespread adoption so far. However the clinical results have been reported
as outstanding, probably because the knees had been restored geometrically to a
close-to-normal condition, together with preservation of the cruciates.
hat Were the Steps Leading to the Basic Condylar

W
Replacement Knees?
Prior to the invention of the first condylar replacement knees, there were a number
of examples of metallic components for resurfacing either the femur or the tibia, but
not in combination. Once Charnley has succeeded in designing the cemented metal-
on-polyethylene hip, it provided the starting point for the first artificial knees. The
earliest designs were very different. Each inventor or inventors had their own pic-
ture of how the knee worked and what type of components would be needed to
restore the knee to a functional condition. In some cases basic research was carried
out first, which provided the ideas for the design of the artificial knee itself.
Competition between individual designers was certainly a factor, but so too was
sharing of ideas and experiences. Progress was rapid, with problems quickly identi-
fied and solved; a process aided by the lack of regulations such that new designs
could be quickly introduced. By 1975, several basic condylar types of artificial knee
had been invented: the ACL + PCL retaining, the PCL retaining (CR), the cruciate
sacrificing basic condylar replacement (TCP), and the unicompartmental. By 1980,
the posterior stabilizer (PS) had been added. Also, the rotating platform knee was
introduced. There had been sufficient clinical experience to show that all of these
designs were promising, but the CR and the PS were the most popular.
References
Bargren JH, Blaha JD, Freeman MA. Alignment in total knee arthroplasty. Correlated biomechani-
cal and clinical observations. Clin Orthop Relat Res. 1983;(173):178–83.
Biazzo A, Manzotti A, Confalonieri N. Bi-unicompartmental versus total knee arthroplasty: long
term results. Acta Orthop Belg. 2018;84(3):237–44.
Confalonieri N, Manzotti A. Mini-invasive computer assisted bi-unicompartimental knee replace-
ment. Int J Medical Robotics and Computer Assisted Surgery. 2005;1(4):45–50.
Coventry MB & Bryan RS. Pitfalls in the Polycentric and Geometric total knee arthroplasties.
Conference on total knee replacement, Institution of Mechanical Engineers, London, 16–18
Sept 1974.
Coventry MB, Upshaw JE, Riley LH, Finerman GA, Turner RH, Turner RH, Geometric total knee
arthroplasty. Patient data and complications. Clin Orthop Rel Res 1973;(94):177–178.
Crawford DA, Adams JB, Hurst JM, Berend KR, Lombardi AV Jr. Ten-year minimum outcomes
and survivorship with a high flexion knee system. J Arthroplast. 2019;34(9):1975–9.
Ewald FC. Joint Prostheses. United States Patent 3,798,676 filed July 9, 1971, issued March 26,
1974.
Ewald FC. Metal to plastic total knee replacement. Orthop Clin North Am. 1975;6(3):811–21.
Finerman GAM, Coventry MB, Riley LH, Turner RH, Upshaw JE. Anametric total knee arthro-
plasty. Clin Orthop Rel Res. 1979;(143):85–90.
Freeman MAR, Swanson SAV, Todd RC. Replacement of the knee with the Freeman-Swanson
prosthesis. Conference on total knee replacement, institution of mechanical engineers, London,
16–18 Sept 1974.
Bone Joint Surg. 1971;53(2):272–7.
Hsieh HH, Walker PS. Stabilizing mechanisms of the loaded and unloaded knee joint. J Bone Joint
Surg Am. 1976;58(1):87–93.
Insall JN, Ranawat CS, Aglietti P, Shine J. A comparison of four models of total knee-replacement
prostheses. J Bone Joint Surg Am. 1976;58(6):754–65.
Lewallen DG, Bryan RS, Peterson LF. Polycentric total knee arthroplasty. A ten-year follow-up
Marullo M, Vitale JA, Stucovitz E, Romagnoli S. Simultaneous bilateral unicompartmental knee
replacement improves gait parameters in patients with bilateral knee osteoarthritis. Knee.
2019;26:1413.
Minns RJ, Hardinge K. Failure of one design of surface replacement knee arthroplasty due to loos-
ening deformation and wear of the plastic femoral component. Biomaterials. 1983;4(3):147–52.
1970;3(1):51–61.
Parratte S, Ollivier M, Lunebourg A, Abdel MP, Argenson JN. Long-term results of compart-
mental arthroplasties of the knee; long term results of partial knee arthroplasty. Bone Joint J.
2015;97-B(10-Suppl A):9–15.
Patil S, McCauley JC, Pulido P, Colwell CW Jr. How do knee implants perform past the second
decade? Nineteen- to 25-year followup of the Press-fit Condylar design TKA. Clin Orthop
Relat Res. 2015;473(1):135–40. https://doi.org/10.1007/s11999-014-3792-6.
Pritchett JW. Charles O Townley: design by Reason. www.seattlejointsurgeon.com. 1988.
Pritchett JW. Patient preferences in knee prostheses. J Bone Joint Surg. 2004;86(7):979–82.
Pritchett JW. Patients prefer a bicruciate-retaining or the medial pivot total knee prosthesis. J
Arthroplast. 2011;26(2):224–8.
Pritchett JW. Bicruciate-retaining Total knee replacement provides satisfactory function and
Purbach B, Wroblewski BM, Siney PD, Fleming PA. The Charnley load angle inlay knee replace-
ment: long-term results 20–28 years. Presented at Leeds bioengineering conference, Leeds,
2004.
Ranawat CS, Boachie-Adjei O. Survivorship analysis and results of Total condylar knee arthro-
plasty. Eight- to 11-year follow-up period. Clin Orthop Relat Res. 1988;226:6–13.
Ranawat CS, Insall J, Shine J. Duo-condylar knee arthroplasty: Hospital for Special Surgery
design. Clin Orthop Relat Res. 1976;120:76–82.
Romagnoli S, Marullo M, Massaro M, Rustemi E, D’Amario F, Corbella M. Bi-unicompartmental
and combined uni plus patellofemoral replacement: indications and surgical technique. Joints.
2015;3(1):42–8.
Sbarbaro JL. Press fit implant arthroplasty of the hip and knee. J Biomed Mater Res Symposium.
1974;5(Part 2):285–88.
Seedhom BB, Longton EB, Wright V, Dowson D. Dimensions of the knee. Ann Rheum Dis.
1972;31:54.
Seedhom B, Dowson D, Wright V. Wear of solid phase formed high density polyethylene in rela-
tion to the life of artificial hips and knees. Wear. 1973;24(1):35–51.
Seedhom BB, Longton EB, Dowson D, Wright V. The Leeds Knee. Conference on total knee
replacement, institution of mechanical engineers, London, 16–18 Sept 1974.
Sledge CB, Ewald FC. Total knee arthroplasty experience at the Robert Breck Brigham Hospital.
Clin Orthop Relat Res 1979;(145):78–84.
Townley CO. Articular-plate replacement arthroplasty for the knee joint. 1964. Clin Orthop Relat
Res. 1988;(236):3–7.
Walker PS. A comparison of normal and artificial human joints. Acta Orthop Belg. 1973;39(Suppl
1):43–54.
1977;59(2):222–8.
Walker PS, Shoji H, Erkman MJ. The rotational axis of the knee and its significance to prosthesis
design. Clin Orthop Rel Res. 1972;89:160–70.
Walker PS, Wang C-J, Masse Y. Joint laxity as a criterion for the design of condylar knee pros-
theses. Conference on Total Knee Replacement, Institution of Mechanical Engineers, London,
16–18 Sept 1974.
Wang CJ, Walker PS. The effects of flexion and rotation on the length patterns of the ligaments of
the knee. J Biomech. 1973;6(6):587–96.
Wang CJ, Walker PS. Rotatory laxity of the human knee joint. J Bone Joint Surg Am.
1974;56(1):161–70.
Yamamoto S. Total knee replacement with the Kodama-Yamamoto knee prosthesis. Clin Orthop
Relat Res. 1979;145:60–7.
Posterior Stabilized
What Was the Design Rationale of the Posterior
10
Stabilized Knee?
The earliest artificial knees were of two basic types, those where the cruciate liga-
ments were preserved and those where their function was substituted by the bearing
surfaces. In the first type, the anterior-posterior (AP) stability was provided by the
cruciates, while the femoral rollback in flexion, thought to be essential for achieving
high flexion, was allowed for by shallow tibial bearing surfaces. This type of artifi-
cial knee was most suitable for arthritic knees without severe deformities. In the
second type, dishing of the tibial surfaces provided the AP stability, but femoral
rollback was limited, reducing the flexion angle in some cases. The advantage of
this type of knee was that it could be used for severe deformities. A design problem
was how to provide progressive femoral rollback with flexion without making the
tibial bearing surfaces too shallow. The answer was to use the condylar surfaces for
weight-bearing, but to use the intercondylar region for some type of mechanism
which controlled the AP displacement. There were two early examples of this solu-
tion, both presented at the 1974 Conference on Total Knee Replacement in London.
Early Designs with Intercondylar Stability
The first was the Deane knee, which actually reversed the materials, plastic for
the femoral component and metal for the tibial component (Deane 1974). The
femoral condyles blended smoothly with the patella flange, and there was a
spherical recess in the intercondylar region. The metallic tibial component had a
base which covered the tibial surface, with shallow bearing surfaces on the lat-
eral and medial sides. A sphere projected from the center of the component and
entered the recess in the femoral component. The interaction of these features
controlled the AP motion, causing posterior displacement of the femur on the
tibia with flexion. In addition, Deane considered that it was essential to repro-
duce the internal-external rotation between the femur and tibia, for balance of the
ligament forces, and to reduce stresses on the fixation of the components. This
axial rotation was allowed by the tibial sphere in the spherical housing of the

168 10 Posterior Stabilized
Femur
anterior in
extension
HOUSING
SPHERE Femoral
rollback
in flexion
Fig. 10.1 Graham Deane FRCS designed this knee based on contact area and kinematic studies
of the anatomic knee. The intercondylar post-in-housing produced progressive posterior displace-
ment of the femur in flexion. At all angles, there was freedom of internal-external rotation due to
the sphere-in-housing. The major weight-bearing was across the condyles themselves. This is one
of the earliest solutions of a condylar replacement knee which provided unlinked intercondylar
guidance to motion (Deane 1974). (Reprinted with permission from Institution of Mechanical
Engineers, London, England)
femur. The motion criteria of the artificial knee were based on biomechanical
studies of the anatomic knee, carried out by Graham Deane in his Master’s Thesis
at the University of Surrey in 1970. The knee was used successfully for a number
of years; after a period of time, cracks developed in the plastic femoral compo-
nent. The thickness of the polyethylene was inadequate to withstand the high
forces of activity. The Deane knee was the first known example of an unlinked
intercondylar mechanism providing femoral rollback and freedom of rotation
while weight-bearing on the condyles.
The Attenborough Stabilized Gliding prosthesis similarly used the intercondylar
region for a motion-guiding mechanism (Attenborough 1976). Chris Attenborough
FRCS was an inventive surgeon who practiced at the Royal East Sussex Hospital in
Hastings on the south coast of England. The ingenious mechanism produced poste-
rior femoral displacement with flexion and allowed internal-external rotation. In
addition it provided varus-valgus stability. Even though substantial cavities were
needed in the center of the distal femur and proximal tibia, the design was consid-
ered an advantage over the hinges because the femoral and tibial canals were not
invaded and the motion was more anatomic. Advantages were also claimed over the
10 Posterior Stabilized 169
Fig. 10.2 The Attenborough Stabilized Gliding knee. An innovative design with partially con-
forming condylar contacts, intercondylar AP stability, freedom of internal-external rotation, and
varus-valgus restraint. Strong fixation is provided with intramedullary pegs. It was first implanted
in January 1973 (Attenborough 1976). (Reprinted with permission from Annals of Royal College
of Surgeons of England)
early condylar replacements: “It has an advantage over unlinked condylar prosthe-
ses, namely, stability even when correcting severe deformities and in knees with
pre-existing ligamentous damage.” The Attenborough knee was first implanted in
January 1973 and then used extensively in England for a number of years. However,
a later follow-up study from Wrightington Hospital showed that there were patella
problems due to the absence of a full anterior flange, as well as loosening. Both the
Deane and the Attenborough can be credited with the first successful use of the
intercondylar space using mechanical guiding means to produce more anatomic
motion, particularly causing posterior femoral displacement in flexion and allowing
some freedom of internal-external rotation.
Stabilized Designs at HSS
In the process of developing the condylar replacements at the Hospital for Special
Surgery (HSS) in the early 1970s, Peter Walker and Hiromu Shoji, an Orthopedic
Fellow, had experimented with an intercondylar linked hinge which allowed for
some rotational laxity. Remarkably this hinge was almost identical to Gluck’s,
which he developed in 1890 as described in Chap. 3. Implantations in dogs and
sheep showed that the center of rotation of the hinge in the sagittal plane was close
to the average center of rotation in the intact joint (Walker and Shoji 1973). But at
only 6 months follow-up, the femoral and tibial condyles appeared to be arthritic
with degenerate cartilage and osteophytes, probably due to being unloaded. This
resulted in the following solution: an artificial knee design with intercondylar stabil-
ity, but where the axial load-bearing was carried on metal-on-plastic condylar sur-
faces. In other words, combine a condylar replacement knee with an intercondylar
stabilized mechanism. The first design which using this concept was called the
stabilo-condylar, made entirely in cobalt-chrome. It was implanted into a patient by
John Insall, and while it performed well functionally, it did make clicking noises.
After further development, a metal-plastic version of the knee was designed where
the tibial component had a vertical plastic post in the center, housed in a cavity in
the center of the femoral component, the objective being to provide varus-valgus
stability during the full range of flexion.
This project now became a group effort with John Insall and Chit Ranawat. After
a few refinements to the shape of the peg and housing, the femoral component was
not only prevented from sliding anteriorly but gradually displaced posteriorly with
flexion. Varus-valgus stability was still provided. The plastic peg was rounded off at
the sides to allow some freedom of axial rotation, rather than allowing only sagittal
plane motion which would make the device little different from a fixed hinge. This
design was called the Total Condylar 2, or TC2. It was not seen as a substitute for
the Total Condylar knee, but for knees where varus-valgus constraint was required,
usually arthritic knees with high preoperative deformities or where both collateral
ligaments were deficient. Extra fixation was provided by elongating the central
stems of both components.
In 1976, when Peter Walker went to work at Cintor, a small implant company in
Boston, he consolidated all of the Hospital for Special Surgery knee designs into a
Graduated Knee System with the intention of making this a major HSS knee sys-
tem. This included the Stabilo-condylar which was a modification of the TC2. The
main change was that a metal axle was fixed in the metal housing and through the
plastic post, linking the two components together. This axle was added to prevent
dislocation or hyperextension from occurring although it was probably not neces-
sary and made the surgical insertion more difficult.
After only a year in Boston, Peter Walker started in 1977 as Director of Product
Development at Howmedica, a well-established orthopedic company with extensive
manufacturing capabilities. One of the first goals was to develop a new knee system,
because the company was only making the Total Condylar knee and a few small-
volume designs in the custom products section. While in Boston, Walker had
Fig. 10.3 The Kinematic

Stabilizer, designed by
Peter Walker, maintains the
cam and the posterior of
the housing in contact
throughout flexion to
produce progressive
femoral rollback. It was a
derivative of the Total
Condylar 2 design from the
Surgery, based on research
which showed that the
center of the femur
displaced posteriorly with
flexion. It was used
surgically from 1978
developed relationships with surgeons at the Robert Brigham Hospital: Clem Sledge,
Fred Ewald, Bill Thomas, and Bob Poss. Within a year, a collaborative effort had
produced the Kinematic Knee System. This consisted of a bi-cruciate retaining type,
a posterior cruciate retaining type (CR), and an alternate to the Total Condylar where
both cruciates were resected called the Kinematic Stabilizer. The starting point for
the Stabilizer was the TC2. Rather than a tall central tibial peg in a femoral housing
which restrained varus-valgus motion, the peg and housing were reduced in height so
that the main function was control of anterior-posterior motion. Throughout the
range of flexion, the femur was constrained from anterior displacement on the tibia,
and the femur was progressively displaced posteriorly with flexion to increase the
range. The design achieved excellent function and durability in long-term follow-up
with a 95% survival rate at 10 years (Emmerson et al. 1996).
However, of the different design options in the Kinematic Knee System, the
Stabilizer was the least favorite of the Robert Brigham Group, who were more
disposed to the cruciate retaining types. This was primarily due to familiarity with
previously used designs including the Ewald anatomic design which had been
used only a few years previously. Moreover Howmedica themselves did not pro-
mote the Stabilizer as much as the other designs in the Kinematic Knee System,
mainly because the Total Condylar was still selling well and continued to be sup-
ported by John Insall and Chit Ranawat. The Kinematic Knee System however
was successful in both the USA and in Europe, especially in the UK and
Scandinavia. One of the features of this knee system was that the tibial compo-
nents were exclusively metal-backed. As described in Chap. 8, as a result of finite
element analysis studies, experimental data, and the advantages of being able to
switch the tibial inserts at surgery, metal-backing had almost completely replaced
all-plastic components by the late 1970s.
In 1976, Albert (Al) Burstein PhD was appointed as the new Director of the
Biomechanics Laboratory at the Hospital for Special Surgery. Burstein’s interest in
biomechanics began in 1963 while he was a mechanical engineering student at
Brooklyn Polytechnic Institute. He attended a course on Orthopaedic Biomechanics
given by Victor Frankel, a surgeon from the Hospital for Joint Diseases (HJD) in
New York (now New York University Langone Orthopedic Hospital). On complet-
ing his degree, Burstein accepted a position as Director of the new Biomechanics
Lab at the Hospital, established by Victor Frankel. Then in 1966, both Frankel and
Burstein left HJD to work at Case Western Reserve University in Cleveland. There
they carried out numerous biomechanical studies, mainly on fracture fixation. They
created a well-attended annual course on biomechanics, which was targeted mainly
for Residents. To go along with this course, they wrote a book entitled Orthopaedic
Biomechanics, the first book of its type which established the role of biomechanics
as an important part of orthopedic surgery. Frankel and Burstein, more than anyone,
were responsible for the expansion in activity in orthopedic biomechanics in the
ensuing years.
Fig. 10.4 Albert Burstein

PhD started his career with
Victor Frankel MD at Case
Western University in
Cleveland and established
orthopedic biomechanics
as a prominent activity,
essential to progress in the
field of artificial knee
design. One of his
significant achievements
was the design of the
Insall-Burstein knee, with
John Insall, at the Hospital
for Special Surgery in the
late 1970s. Al Burstein was
a clear-thinking designer
and lecturer who is still
involved with the field
today. (Photo c 2015)
(© Peter S Walker)
Insall-Burstein (IB) PS Knee
Soon after arriving at the Hospital for Special Surgery (HSS) in 1977, Al Burstein
appointed several PhDs with expertise in different areas. Among these new staff
were Tim Wright PhD and Claire Rimnac PhD who continue to be active in the
area of joint replacement and biomaterials today. Very soon Al Burstein was
approached by John Insall to make some modifications to the Total Condylar
knee. Although the clinical results were successful, there of the Total Condylar
knee were successful, there were two deficiencies which needed attention. The
first was that the range of flexion was only averaging 90 degrees of flexion, about
10–20 degrees less than was being achieved for the Duo-patella and for other
cruciate retaining designs being used at the time. The second limitation was that
in some cases, when the knee was flexed and then extended, the femur would slide
anteriorly on the tibia. The main reason for this was inadequate soft tissue balanc-
ing at surgery where the knee was too loose in flexion. In retrospect, both of the
deficiencies may have been ameliorated by making the tibial cut slope posteriorly
by five or six degrees, as it is anatomically. However at that time, the prevailing
technique was that the tibial resection plane was perpendicular to the long axis of
Fig. 10.5 John Insall MD

was Chief of the Knee
Service at the Hospital for
Special Surgery from 1969
to 1996 and was a major
contributor the early
artificial knee designs,
most particularly the Total
Condylar and its
instrumentation. In the late
1970s, he and Al Burstein
designed the Insall-
Burstein posterior
stabilized knee. Over
several decades John Insall
was a leading authority on
the technique of total knee
arthroplasty and trained
innumerable Residents and
Fellows, who continue his
legacy. (Photo mid 1990s)
(© Peter S Walker)
the tibia in both the frontal and sagittal planes. Within 2 years, Insall and Burstein
had developed the eponymous Insall-Burstein (IB) knee, the type now generically
called a posterior stabilized or PS knee (Insall et al. 1982). It was first used in
surgery in 1978. The development was simultaneous to that of the Kinematic
Stabilizer, but there was no connection between the two projects.
The IB knee used an intercondylar cam-post to produce increased flexion and
prevent anterior femoral displacement in high flexion. Burstein analyzed numerous
shapes of intercondylar cam-post configurations and arrived at a design where a
crossbar within an intercondylar housing in the femoral component interfaced with
a short tibial post projecting from the center of the plastic tibial component. As the
knee was flexed, contact between the crossbar and the post occurred at about 70
degrees flexion, which subsequently caused femoral rollback up to maximum flex-
ion. Also the bottom of the dish of the tibial component was more posterior than
that of the Total Condylar knee, which provided a longer lever arm for the quadri-
ceps. Burstein’s design method was impressive in that he analyzed the kinematic
motion patterns at quadruple size on sheets of transparent paper of substantial
dimensions. Computer-aided design (CAD) was not available at that time; every-
thing had to be designed with paper and pencil. An important mechanical factor
was also theorized regarding stability of the tibial component in high flexion. Due
to the femoral-tibial contact points being close to the posterior, this might cause
excessive compressive stresses on the posterior trabecular bone and uplifting of the
anterior of the tibial component. On the other hand, the anterior shear force acting
on the plastic post would have the opposite effect. Burstein showed that the resul-
tant of these two forces was a single force close to vertical acting near the center of
the component.
Another favorable mechanical feature of the IB was that the tibial component
was designed to be at five-degree posterior slope on the tibia with the contact point
of the femur on the tibia in early flexion a few millimeters posterior of the center,
resembling the normal anatomic position. The height of the lateral and medial bear-
ing surfaces was increased anteriorly to limit the anterior sliding of the femur on the
tibia in the flexion arc prior to the engagement of the cam-post. There was also an
advancement in materials, namely, molded polyethylene for the tibial component,
which was shown to have superior wear properties to the machined components
used for most artificial knees on the market at the time.
A US patent was filed for the design in 1980, which was issued in 1981 (Burstein
and Insall 1981). The fundamental mechanism was explained in the patent as fol-
lows: “The camming action between the cam follower and cam surface forces the
zones of contact between the condylar portions of the femoral component and con-
cavities in the tibial component posteriorly as the leg approaches full flexion,
thereby increasing the range of flexion without interference between posterior sur-
faces of the femur and the tibial component and preventing anterior dislocation of
the femur. The plateau of the tibial component slopes inferiorly and posteriorly, also
to increase the range of flexion without interference.” The emphasis was on preven-
tion of anterior displacement in flexion, combined with femoral rollback, to fulfil
the main two design criteria.
cam
post
Contact
posterior of
center
Cam-post
contact 80-90
deg flexion
Femoral
rollback
after
contact
Fig. 10.6 The Insall-Burstein PS total knee was designed at the Hospital for Special Surgery and
used surgically from 1978. It was intended as an improvement to the Total Condylar knee, where
the cam-post would prevent anterior sliding of the femoral component and cause rollback in high
flexion, increasing the range of flexion. (Figure on left courtesy of Fabio Catani MD, University of
Modena & Reggio Emilia. Figure on right from US Patent 4,298,992, filed 1980, issued 1981)
Zimmer, who manufactured the knee, immediately recognized the potential of

the IB knee design. While the Total Condylar had been very successful for over
5 years, clearly the next development from HSS, and from Insall and Burstein in
particular, had a high market potential. Zimmer gave strong promotional support,
while Insall and Burstein gave numerous lectures about the knee at prominent con-
ferences. By the late 1970s, there were many courses on total knee replacement run
not only by the AAOS and other organizations but by major institutions. Companies
often organized special courses featuring their new knee designers. There was con-
siderable demand for attending such courses from surgeons who were motivated to
join the fast-growing practice of artificial knee surgery. As a result of the excellent
short-term results of the IB design, the credibility of the developers, the prestige of
the Hospital for Special Surgery, and the efforts of Zimmer, within a few years, the
IB became one of the top selling knees.
The IB knee proved to be relatively easy to insert, yet forgiving due to its inherent
stability. It increased the average range of flexion by about 15 degrees and had higher
functional scores compared with the Total Condylar. It was modestly described as a
“valuable adjunct to the Total Condylar prosthesis” (Insall et al. 1982). In a long-term
follow-up of 100 cases at 15–19 years, the survivorship was 92.4% (Abdeen et al.
2010). Most of the revisions were due to patella problems. From the radiographic
examination, it was found that there is no observable osteolysis. This would have
been expected due to the accumulation of particulate debris, as has been seen with
other designs at even shorter follow-up times. In the case of the IB design, the lack
of osteolysis was attributed to the use of molded polyethylene and a monoblock con-
struction where the plastic was fused to the metal-backing at manufacture. This pre-
vented backside wear from occurring. One design detail with the IB design which
was overlooked was that the posterior of the vertical plastic post was flat. This meant
that when the femur rotated internally or externally, the contact with the transverse
metal bar was on the corners of the plastic post, causing wear and local fragmenta-
tion in some cases, as shown in subsequent retrieval studies. Other studies of the IB
design, including subsequent models to the original, have noted “patella clunk.” This
was explained as mal-tracking of the patella and entrapment of tissue due to the wide
intercondylar housing of the femoral component.
Because of the advantages of the Insall-Burstein type of design, from the early
1980s, many surgeons used it for almost all of their cases, while several other com-
panies designed their own versions, circumventing the patents for the Kinematic
Stabilizer and the Insall-Burstein. In the larger companies, the PS was usually incor-
porated into a knee system which also included the still widely used CR type of
design. One of the benefits of this was the cost advantage of being able to use the
same metal tibial tray for both designs. Much of the popularity of the IB knee was
due to Insall’s attention to teaching a succession of Residents and Fellows and also
his comprehensive book, Surgery of the Knee. Production of several editions of this
book was greatly facilitated by Norman Scott, who eventually became the main
Editor. John Insall also established a travelling fellowship program. Norman Scott
maintains the program to this day, which continues to provide so much benefit to its
participants.
Since the early 1980s, the IB posterior stabilized design has been redesigned to
take advantage of the extensive clinical experience. One example of this is that the
“patella clunk” problem was avoided when the PS knee was redesigned into the
NexGen Knee System in the late 1990s (Lonner et al. 2007). The kinematics of the
PS and its cam-post was studied in detail using fluoroscopic imaging of a loaded
lunge activity (Suggs et al. 2008). The patients in this study were Asian with an
average passive range of flexion postoperatively of 142 degrees. The cam-post
engaged at an average of 91 degrees, after which there was posterior femoral dis-
placement, contributing to the high flexion and fulfilling the design objective. The
limited internal-external rotation in the early IB models was later solved by round-
ing the tibial post. In the NexGen, different dimensional combinations were also
added to the femoral component to account for statistical variations between gen-
ders and races. This aspect of the ideal dimensions and shape for both the femoral
and tibial components was further refined when the Persona Knee System was intro-
duced in 2013. The different versions of the PS design, as well as John Insall’s many
contributions to artificial knee designs and surgery, have been described in a well-
illustrated article (Scuderi et al. 2001).
However, the above lineage of the Insall-Burstein PS knee was not the only
development of the cam-post concept. In 1996, John Insall left the Hospital for
Special Surgery (HSS) to form the Insall Scott Kelly Clinic. In the same year, Al
Burstein also left HSS, but not before starting work on a new posterior stabilized
knee. This development involved surgeons affiliated with HSS and was to be called
the Optetrak, manufactured by Exactech. The design goals were to minimize the
wear and to improve the patella tracking. Wear was considered to be affected by
contact stress. Working with Don Bartel PhD at Cornell University, a finite element
analysis determined that reducing the frontal plane radii of the bearing surfaces and
making the femoral and tibial surfaces almost conforming reduced the contact
stresses significantly (Bartel et al. 1995). In addition, Burstein worked with Steve Li
PhD, formerly a member of the HSS Biomechanics Laboratory, on a new net-shaped
molding methodology for the polyethylene. The Optetrak knee was compared with
the IB 2 knee in a study of 68 patients, with bilateral knees implanted, followed for
11 years (Robinson and Green 2011). For both knee designs, survivorship was 98%,
and there was no report of any osteolysis which may have been caused by particu-
late wear. In another study of 171 knees followed for 7 years, the range of flexion
averaged 120 degrees, a slight improvement over previous PS versions, while there
were no patella “clunk” or other patella problems reported (Ehrhardt et al. 2011).
Hence this second PS line of development met with improved results over the previ-
ous PS versions of the original Insall-Burstein.
Ultracongruent Knee
As pointed out in Chap. 3, to every design problem, there is usually more than one
solution. In this case the relevant design criteria were to provide sufficient anterior-
posterior stability and a high range of flexion in the absence of the cruciate liga-
ments. A different solution to the PS was introduced by Aaron Hofmann in the late
1990s, in the form of the Natural knee (Hofmann et al. 2010). Comparing the basic
mechanisms of the two most widely used types of design, the CR and the PS, the
sagittal radius of the CR tibial bearing surface is typically larger than that of the
PS. In other words, the CR tibial surface is shallower and less constrained in an
anterior-posterior direction, while the PS is more constrained. This seems logical
based on the absence of both cruciate ligaments in the PS design, where the cam-
post substitutes for the function of the posterior cruciate and the elevated anterior
bearing surfaces of the tibial component substitutes for the posterior cruciate. In
laboratory studies of the stability and laxity of designs with different radii, it was
determined that the boundary in the tibial radius values between the CR and the PS
had a sound rationale (Luger et al. 1997). For cruciate resection, it was stated by
Hofmann that “Relatively shallow posterior tibial curvature and a steep anterior
curvature were concluded to provide the most satisfactory combination of laxity and
stability.” But it was later determined from fluoroscopy studies of patients with CR
and PS designs that there was a limit to how shallow the posterior curvature should
be (Dennis et al. 2003). When it was too shallow, as in typical CR designs, “para-
doxical sliding” occurred. The reason was that due to the absence of the anterior
cruciate, the femoral-tibial contact point was often located too posteriorly. On the
initiation of function, the anterior shear force component caused the femur to slide
anteriorly until it was stabilized by the anterior slope of the tibial surface. But
because the anterior slope of the CR was relatively shallow, the sliding distance
could be felt by the patient as instability.
These considerations of stability and laxity occurred to Aaron Hofmann after a
decade of experience with his Natural knee, which was a CR type (Hofmann et al.
2000; Hofmann 2010). Hofmann reasoned that if the tibial bearing surface was
elevated anteriorly, it would effectively substitute for the PCL. The tibial insert was
called the Ultracongruent. “The increased anterior build-up prevents the tibia from
subluxing posteriorly through the entire arc of knee motion.” However there was no
mechanism for causing the femur to displace posteriorly during high flexion, and
this might cause a reduction in the maximum flexion angle achieved. Surprisingly,
clinical results by both Hofmann and an independent group showed that the ranges
Fig. 10.7 The Natural

knee designed by Aaron
Hofmann MD with input
on fixation aspects from
Roy Bloebaum PhD, from
the University of Utah, Salt
Lake City. The raised
anterior surface of the
tibial insert (arrows) limits
the anterior sliding of the
femoral component
particularly in early-mid-
flexion. The posterior lip
on the tibia (arrow) also
limits the AP
displacements and prevents
excessively posterior
contact. An advantage of
the design is that it can
replace both CR and PS
types of design (Hofmann
2010). (Reprinted with
permission from American
Journal of Orthopaedics)
of flexion, or the gains between pre-op and post-op, were similar for the CR and the
ultracongruent (Lutzner et al. 2017). The ultracongruent concept has a number of
advantages over a PS configuration. There is no need to cut an intercondylar hous-
ing in the distal femur, the femoral trochlea can be extended distally giving improved
patella tracking, and surgical insertion is easier. The ultracongruent concept is a
valid alternate to the cam-post, and the concept may attract increasing attention in
the future. It is still not clear however why the lack of femoral rollback does not
reduce the flexion angle achieved. It is possibly due to a relatively low posterior lip
on the tibial component, so the posterior femoral cortical impingement does not
occur too early and block flexion. Further investigation of the geometrical and surgi-
cal factors which affect the range of flexion, a very important outcome parameter for
an artificial knee, is warranted.
hat Was the Design Rationale of the Posterior

W
Stabilized Knee?
The reason that the posterior stabilized knee was designed was the limitations in the
Total Condylar knee, a flexion range in patients averaging only about 90 degrees,
and flexion instability in some cases. It was thought by John Insall that there should
be a solution to these problems, and accordingly he and Al Burstein invented the
Insall-Burstein knee. This was termed a posterior stabilized (PS) knee. The first
characteristic of the PS was that a cam-post mechanism caused femoral rollback
from mid-flexion to full flexion, rollback being thought to be the factor which pro-
duced a high flexion. The cam-post would also prevent anterior femoral subluxation
in flexion. A further principle was that the resultant of the downward compression
force near the posterior of the tibial condyles and the anterior force on the plastic
post would be a force which would be more central, so as not to cause rocking of the
tibial component. Molded polyethylene in a monoblock tibial component would
minimize polyethylene wear.
References
Abdeen AR, Collen SR, Vince KG. Fifteen-year to 19-year follow-up of the Insall-Burstein-1 total
knee arthroplasty. J Arthroplast. 2010;25(2):173–8.
Attenborough CG. Total knee replacement using the stabilized gliding prosthesis. Ann R Coll Surg
Engl. 1976;58:4–14.
Bartel DL, Rawlinson JJ, Burstein AH, Ranawat CS, Flynn WF. Stresses in polyethylene compo-
nents of contemporary total knee replacements. Clin Orthop Relat Res. 1995;317:76–82.
Burstein AH, Insall JN. Posterior stabilized total knee joint prosthesis. US Patent number
4,298,992. Filed 1980, issued 1981.
Deane G. New concept in total knee design. Conference on Total Knee Replacement. London:
Institution of Mechanical Engineers; 1974. p. 16–8.
Dennis DA, Komistek RD, Mahfouz MR. In vivo fluoroscopic analysis of fixed-bearing total knee
replacements. Clin Orthop Relat Res. 2003;410:114–30.
Ehrhardt J, Gadinsky N, Lyman S, Markowicz D, Westrich G. Average 7-year survivorship

and clinical results of a newer primary posterior stabilized total knee arthroplasty. HSS J.
2011;7:120–4.
Emmerson KP, Moran CG, Pinder IM. Survivorship analysis of the Kinematic Stabilizer total knee
replacement: a 10- to 14-year follow-up. J Bone Joint Surg British volume. 1996;78(3):441–5.
Hofmann AA, Tkach TK, Evanich CJ, Camargo MP. Posterior stabilization in total knee arthro-
plasty with use of an ultracongruent polyethylene insert. J Arthroplasty. 2000;15(5):576–83.
Hofmann AA. The design principles of the Natural-Knee system. Am J Orthod (Belle Mead, NJ).
2010;39(6 Suppl):2–4.
Insall JN, Lachiewicz PF, Burstein AH. The Posterior Stabilized condylar prosthesis: a modifica-
tion of the Total Condylar design. Two to four-year clinical experience. J Bone Joint Surg Am.
1982;64(9):1317–23.
Lonner JH, Jasko JG, Bezwada HP, Nazarian DG, Booth RE. Incidence of patellar clunk with a
modern posterior-stabilized knee design. Am J Orthop (Belle Mead NJ). 2007;36(10):550–3.
Luger E, Sathasivam S, Walker PS. Inherent differences in the laxity and stability between the
intact knee and total knee replacements. The Knee. 1997;4(1):7–14.
Lutzner J, Beyer F, Dexel J, Fritzsche H, Lutzner C, Kirschner S. No difference in range of motion
between Ultracongruent and Posterior Stabilized design in total knee arthroplasty: a random-
ized controlled trial. Knee Surg Sports Traumatol Arthrosc. 2017;25(11):3515–21.
Robinson RP, Green TM. Eleven-year implant survival rates of the all-polyethylene and metal-
backed modular Optetrak posterior stabilized knee in bilateral simulataneous cases. J
Arthroplast. 2011;8(26):1165–9.
Scuderi GR, Norman W, Scott WN, Tchejeyan GH. The Insall Legacy in Total Knee Arthroplasty.
Clin Orthop Rel Res. 2001;392:3–14.
Suggs JF, Hanson GR, Park SE, Moynihan AL, Li G. Patient function after a posterior stabi-
lizing total knee arthroplasty: cam-post engagement and knee kinematics. Knee Surg Sports
Traumatol Arthrosc. 2008;16(3):290–6.
ciples. Clin Orthop Relat Res. 1973;94:222–33.
Guided Motion Knees
Do Guided Motion Designs Have
11
Advantages over the CR and PS?
The two most popular condylar replacement knees today are the CR, where the
posterior cruciate ligament is retained, and the PS, where both cruciate ligaments
are resected and an intercondylar cam-post provides stability. The starting point of
this chapter is the condylar replacement knee, where the femoral condyles resemble
those of the anatomic femur, although with simplified geometry, while the tibial
surfaces have radii which are larger than those of the femoral surfaces in both the
frontal and sagittal planes. The collateral ligaments are retained. In such a knee, the
femur can displace anterior-posterior or medial-lateral and rotate internal-external
on the tibia, until the femur is constrained by the ligaments or the curvature of the
bearing surfaces. This applies to the PS knee, until the femoral cam interfaces with
the tibial post, after which the femur is guided to displace posteriorly. In that sense,
the PS is a guided motion knee. A rotating platform knee where the displacements
between the femur and tibia are prevented or are very limited, but where the rotation
is not constrained by the bearing surfaces but by the retained ligaments, will also be
considered as a guided motion knee.
Therefore, the definition of a guided motion knee is where mechanical features
control or guide the femoral-tibial motion, in a way which is different to the conven-
tional bearing surfaces of a condylar replacement knee; usually, a guided motion
knee will be designed to reproduce some motion characteristics of the anatomic
knee. Some guided motion knees were invented before 1980 and have been used
clinically, while others have emerged only recently and are not yet being used. In
this chapter, rotating platform, meniscal bearing, and medial pivot knees will be
included. Some guided motion knees which have similar design features to the PS
have already been discussed in the previous chapter. One of the questions that will
be addressed during this and later chapters is in what way can the kinematics of the
artificial knee be changed such that it will result in improved function or offer
advantages in other ways.

182 11 Guided Motion Knees
Freedom of Rotation
Allowing freedom of internal-external rotation was originally introduced to address

tibial loosening and to allow some freedom of rotation as in the anatomic knee. At
Syracuse University, New York, in the early 1970s, David G Murray, MD, an ortho-
pedic surgeon, and James Shaw, PhD, a mechanical engineer, had built a knee simu-
lating machine to study the rotational laxity in the anatomic knee, with a view to
applying the research to an artificial knee design (Shaw and Murray 1973, 1974).
They invented the Variable Axis knee which had a convex spherical surface in the
center of the femoral component, articulating in a concave spherical dish in the
center of the plastic tibial component. This design was unconstrained in rotation but
was constrained in the anterior-posterior and medial-lateral directions. The main
load-bearing surfaces were cylindrically shaped femoral condyles and flat tibial sur-
faces, which provided a wide base for varus-valgus stability. The knee was strongly
constructed in having a metal-backing for the tibial component and a femoral com-
ponent without an intercondylar cutout. Fixation was augmented with cemented
intramedullary stems. In a clinical follow-up (Murray 1982), a problem of the
Variable Axis was subluxation and pain in the patella, due to the projecting spheri-
cal surface in the center of the femoral component interfering with patella tracking
at higher flexion angles. As with the Total Condylar knee, there were also cases of
anterior femoral subluxation in flexion if the collaterals were excessively loose. Use
of the Variable Axis had tailed off by the end of the 1970s mainly due to these prob-
lems, as well as to the over-invasive fixation stems. The design had overcome the
loosening problem but had introduced other problems. However, the Variable Axis
knee was one of the first designs which allowed complete freedom of internal-
external rotation.
An interesting variation of the Variable Axis concept was designed in the late
1980s, called the BiSurface knee (Ueo et al. 2011). As in the Variable Axis, a spheri-
cal surface projected from the femoral component and was located in a recess in the
tibial insert. In the BiSurface, the sphere was posterior so that it did not interfere
with patella tracking. Contact between the sphere and recess only occurred after 90°
flexion so that freedom of rotation occurred in high flexion. This addressed the
requirements of the eastern lifestyle where high flexion was an everyday require-
ment. The knee has a strong following in Japan to this day, partly because the manu-
facturer, Kyocera, is a Japanese company, but also because the femoral component
is made from alumina ceramic, which reduces the wear.
Another example where rotation was unconstrained was the UCI knee (University
of California, Irvine), designed in the early 1970s by Theodore (Ted) Waugh, MD,
working with the Biomechanics Laboratory at the University of California, Irvine
(Evanski et al. 1976; Waugh et al. 1973). The UCI was an ingenious arrangement
where the femoral condyles were toroidal, while the tibial bearing surface was
shaped by the path swept out by the femoral condyles when they were rotated about
a central vertical axis. As with the Variable Axis, anterior-posterior and medial-
lateral displacements were constrained, but in the UCI, both cruciates were retained.
While the femoral component was well fixed due to a triangular prong which was
11 Guided Motion Knees 183
Fig. 11.1 The Variable

Axis knee was designed by
David G Murray, MD, and
James Shaw, PhD, in the
early 1970s. The first
design criterion was
durable fixation, achieved
with the intramedullary
stems. An additional
criterion was stability with
freedom of rotation. This
was provided by a
spherical surface projecting
from the distal femur
locating in a spherical
trough in the center of the
tibial surface. The plastic
insert was fixed with a
transverse metal pin
(Murray et al. 1982).
from Orthopaedic Clinics
of America)
impacted into the distal femur, due to the inadequate stiffness of the all-plastic tibial
component, there was a relatively high incidence of loosening and deformation.
Nas Eftekhar’s Intramedullary Adjustable knee had a similar circular track on the
tibial surface, allowing freedom of rotation. Eftekhar’s femoral component was a
Total Condylar with the addition of a central stem. One-piece plastic tibial inserts
were housed in a metal tray, which had a large central stem and two small side pegs.
Due to this rugged construction, there was only one revision at an 8-year follow-up
of 112 cases (Eftekhar 1983).
Freedom of rotation would result in reduced shear stresses at the cement-bone
interface of the tibial component and would allow for errors in the rotational
position at surgery. In the anatomic knee, however, the axis of internal-external rota-
tion, while not fixed, is biased to the medial side. Whether a central pivot is a major
disadvantage, especially after soft tissue balancing at surgery, is not known.
Reproducing the Menisci
In the early 1970s, evidence was emerging of the importance of the menisci to knee
function (Bullough et al. 1970; Walker and Erkman 1975; Krause et al. 1976). An
important mechanical property of the menisci was that they allowed the lateral and
medial femoral condyles to slide anteriorly and posteriorly on the tibial plateau,
while interposition between the femoral-tibial bearing surfaces produced congruity.
This was possible because of the deformability of the menisci, their loose peripheral
attachments, and their high tensile strength. The mobility of the lateral meniscus
was much greater than that of the medial. It was not surprising that artificial menisci
would be considered for inclusion in an artificial knee design, because the problem
of lack of congruity between the femoral-tibial bearing surfaces would be solved.
Three different types of meniscal bearing knees originated independently at
almost the same time. The first of these was at the Nuffield Orthopaedic Centre in
Oxford, England. Joseph Trueta, MD, and Peter Bullough, MD, both pathologists,
had made fundamental discoveries in the etiology of osteoarthritis of the knee con-
cerning the areas of load transmission. Also involved with this research were John
Goodfellow, FRCS, an orthopedic surgeon at the Nuffield Hospital, and John
O’Connor, PhD, a Professor of Engineering at Oxford University. There was also
collaboration with the research group at Imperial College London, notably Alice
Maroudas, PhD, a physiologist; Geoff Kempson, PhD, a mechanical engineer; and
Michael Freeman, FRCS, an orthopedic surgeon. Working at Imperial College and
the London Hospital, they were studying the properties of articular cartilage and
synovial fluid in relation to the mechanical factors which led to cartilage degenera-
tion and osteoarthritis. These collaborations are notable because they provided a
stimulating research environment.
Central to all of this work was an understanding of the anatomy of the joint and
the mechanisms of load transmission. Areas of cartilage which experienced only
small loads were subject to disuse atrophy and degeneration, while areas of exces-
sive loading degenerated due to mechanical failure (Bullough et al. 1968). This
theory was investigated for the hip joint by Seth Greenwald who developed an elab-
orate loading rig and devised novel methods using colored dyes for showing the
areas of contact as loads were progressively applied (Greenwald and Haynes 1972).
Greenwald had graduated in structural engineering from Massachusetts Institute of
Technology and was about to embark on a PhD at Oxford on the design of cooling
towers. While waiting to meet his intended supervisor in a public house in Oxford,
he fell into conversation with John O’Connor and John Goodfellow, who persuaded
him that hip joints were a far more interesting topic than cooling towers. Although
Greenwald’s work on the hip was never followed up at Oxford by parallel work on
the knee, it had a major influence on the understanding of knee mechanics. These
studies strongly suggested that in the knee, in order to maintain areas of contact
which were sufficiently large to avoid excessive contact stresses, the menisci had to
be playing a major role in force transmission.
Goodfellow had observed that osteoarthritis of the knee most frequently initiated
anteriorly on the medial side, over an area which contacted in full extension. Hence,
the problem was identified as compartmental rather than involving the whole joint.
The artificial knees which had been designed at that time, including the Freeman-
Swanson, were designed for arthritis involving the entire joint. On the other hand, a
unicompartmental knee could be used to treat the osteoarthritis at an early stage.
Several unicompartmental designs were already in use, but they had a serious prob-
lem according to Goodfellow and O’Connor (1978): the incongruent bearing sur-
faces produced small contact areas with high contact stresses between the metal and
the polyethylene, which would lead to deformation of the plastic and long-term
wear. Moreover, they had determined using the Oxford knee rig with a similar dye
technique as Greenwald’s that the contact points on the tibial surfaces displaced
posteriorly during flexion. Hence, a fully conforming unicompartmental knee where
the components were fixed to the femur and tibia was not an option.
A solution to this dilemma was elusive until O’Connor and Goodfellow attended
a conference at the University of Leeds in late 1973, where Seedhom presented his
research on the importance of the menisci in distributing the load but allowing slid-
ing motion to occur (Seedhom 1976). In a sudden insight in the car trip back to
Oxford, it was realized that a bearing which included a movable meniscus made
from polyethylene was a solution. The femoral component would be a curved metal
runner with a spherical outer surface. A polished metal plate would cover the resected
Fig. 11.2 The Oxford unicondylar knee was based on the four-bar linkage model of normal knee
mechanics consisting of the cruciates, the femur, and the tibia. The linkage produces posterior roll-
ing of the femur on the tibia while flexing. To maintain full congruency of the femoral-tibial con-
tacts in order to minimize contact stresses, a meniscal-shaped plastic bearing (yellow) was
interposed between the metal femoral component (blue), and the metal tibial plate (blue), fixed to
the resected tibial surface. The meniscus slides anterior-posterior on the tibial plate during flexion
(Goodfellow and O’Connor 1978). (Reprinted with permission from Journal of Bone & Joint
Surgery B)
medial tibial plateau. In between would be a plastic meniscus, with a spherical

upper surface to match the femoral component and a flat lower surface to match the
tibial plate. With this configuration, a full contact area could be achieved for all
flexion angles, even if displacements and internal-external rotations occurred. For
convenience of manufacture, half of the femoral component of a Geometric knee
was used, because of its spherical bearing surface. The tibial plate and the plastic
meniscus were relatively easy to manufacture (Goodfellow and O’Connor 1978).
The knee was called the Oxford meniscal bearing knee, since shortened to the
Oxford Uni.
The design had a possible failure mode in that the meniscus, being unrestrained,
could dislocate anteriorly or posteriorly. However, trials on knee specimens showed
that if the thickness of the meniscus was made in millimeter increments, selection
of the correct thickness at surgery would prevent dislocation from occurring.
Nonetheless dislocation did occur occasionally, but this was solved by slightly
increasing the lip height of the meniscus and ensuring the correct slope of the tibial
metal plate in the sagittal plane. In laboratory wear testing, and in subsequent clini-
cal application, the polyethylene wear was shown to be minimal, attributed to the
fully conforming bearing surfaces. The Oxford Uni, which has had a few minor
changes to the design and technique over the years, has continued to be widely used
and is the largest selling unicompartmental knee (Pandit et al. 2015). The long-term
results are similar to those of an artificial knee, as reported by the developing group.
However, to achieve such results, David W Murray, FRCS, has emphasized the
necessity of frequent use and acquiring expertise in case selection and surgical tech-
nique. Carrying out the procedure only occasionally leads to inferior results. A his-
tory of the Oxford knee has been documented (Price et al. 2007).
One of the first meniscal bearing artificial knees was designed by Julian Minns,
PhD, a medical physicist working at Dryburn Hospital in Durham in northeast
England. The major responsibility of Minns was to maintain the technical equipment
of the hospital, although being inventive, he found artificial joint design an intriguing
outlet. He had also formed a number of contacts in the joint replacement field includ-
ing the staff at Wrightington Hospital where John Charnley was located. Minns did
not want to design just another condylar knee, as he saw the knee as a complex
mechanism with many functional parts which he wanted to represent. He recognized
the multiple motions which occurred between the femur and the tibia and the role
that the cruciate ligaments played in guiding those motions. He also noted that wear
in total hips was still a problem even with polyethylene in the conforming Charnley
hip. He started initially with a unicompartmental design in 1978 where the anterior-
posterior motion of a plastic meniscus was guided in a T-slot (Minns and Campbell
1978). Using a photoelastic technique, he demonstrated the much lower plastic con-
tact stresses in a full conformity bearing surface compared with the stresses in a
fixed-bearing shallow polyethylene surface. Minns then moved to a meniscal bearing
artificial knee consisting of a condylar femoral component and a metal tibial base-
plate with a cutout to preserve the cruciate ligaments. Each tibial condyle had a
dovetail slot which housed a polyethylene meniscal bearing with an upper surface in
close conformity with the metal femoral condyle. As the knee was flexed and
extended, the lateral and medial menisci moved independently in their tracks.
Extensive testing was carried out, showing low wear and low torque transmission
(Minns 1989). A consequence was no loosening of tibial and femoral components in
450 cases at up to 5 years. Dislocation of the plastic bearings was thought to be
avoided because of the dovetail grooves, except in a few cases where there was
excessive laxity (Minns 1989). The Minns meniscal knee was used by a few surgeons
in England, but it was eventually discontinued for commercial reasons.
Meanwhile, in New Jersey in the USA, Fred Buechel, a surgeon from New Jersey
Medical School, and Michael Pappas, PhD, a mechanical engineer at New Jersey
Institute of Technology, had been thinking of how artificial joints could be designed
which had low contact stresses to minimize wear while still allowing freedom of
motion. In 1977 they designed a floating shoulder with an interposed polyethylene
bearing and were considering how to apply the design concept to the knee. The
main design criteria were low contact stresses but freedom of motion. They had
studied the Geometric knee which had low contacts stresses, but the motion was so
Fig. 11.3 Fred Buechel,

MD, from New Jersey
Medical School, formed a
life-long partnership with
Mike Pappas, PhD, a
Professor of Mechanical
Engineering at New Jersey
Institute of Technology.
Together, in the late 1970s,
they designed the low
contact stress (LCS)
rotating platform knee.
This unique configuration
minimized the contact
stresses, yet allowed
freedom of rotation so that
the knee could move more
naturally without
constraint. The systematic
surgical technique and
careful follow-up studies
have resulted in the
continuous clinical success
of this design to the present
day. (Photo taken at The
Knee Society 2018) (©
Peter S Walker)
restricted that loosening and limitation of motion were problems. Buechel and
Pappas initially designed a meniscal bearing knee very similar to that of Minns, but
with curved tracks instead of straight. This was consistent with an axis of rotation
close to the center of the knee (Buechel and Pappas 1986). As with the Minns, the
design was intended to retain both cruciates, but they also extended the concept to
retention of only the posterior cruciate. In both designs, the sagittal radii of the
femoral condyles were larger distally and smaller posteriorly, as for the anatomic
knee. As a result the contact areas would be slightly reduced in flexion with an
increase in contact stresses, although this would only have a small effect on the wear
because most weight-bearing activity is carried out with the knee at lower angles of
flexion. The short-term clinical results of the meniscal bearing knee reported by
Buechel were excellent, with no mechanical failures (Buechel and Pappas 1986).
Soon after, Buechel and Pappas extended their design concepts to a rotating plat-
form design. Using the same femoral component as before, the metal tibial compo-
nent was a polished flat plate covering the resected surface of the tibia, with a
tapered central peg for fixation. An interposed one-piece plastic rotating platform
was dished medially and laterally to match the femoral condyles, with a central peg
which pivoted in a hole in the central peg of the tibial plate. This configuration
allowed for freedom of internal-external rotation about a central axis. The knee was
called the LCS, low contact stress. This system successfully combined low con-
straint forces in rotation with low contact stresses. Initially this was only intended
for limited application: “the use of the rotating intermediate bearing element in
cases where cruciate function has been lost.” Gradually, however, Buechel and other
surgeons recognized the easier surgery and the more reliable results with this design.
Previous research studies on anatomic knee kinematics had shown that the rota-
tional axis of internal-external varied with the flexion angle but was biased towards
the medial side, consistent with the higher conformity on the medial side and the
mobility on the lateral side. In contrast, the axis of rotation of the LCS was central.
Despite this apparent mismatch of the kinematics imposed by the rotating platform,
the freedom of axial rotation provided advantages. In function, the knee could find
its preferred rotational position, so that the knee was unconstrained in turning and
twisting actions. In addition, freedom of rotation compensated for any rotational
error in placing the tibial component at surgery. This was useful in that there was no
general agreement for the optimal rotational position. The absence of the cruciate
ligaments with the rotating platform design was an advantage in simplifying the
surgery, and in any case, anterior-posterior stability was provided by the conforming
bearing surfaces themselves. At 10-year follow-up, the LCS rotating platform
design gave 97% survivorship (Buechel et al. 2001).
The LCS rotating platform of Buechel and Pappas built up a strong following.
The rotating platform concept was incorporated into most artificial knee systems
from several companies. The way this was done was to use the same femoral com-
ponent as for the CR design, in combination with a rotating platform tibial compo-
nent. The original purpose of the design was to minimize the wear, and this would
be valid at the time the LCS was designed. However, today’s polyethylenes are
much more wear resistant, so reducing the contact stresses is no longer such an
advantage. Even when there is relatively low conformity such as in a typical CR

design, the wear measured in tests on simulating machines and observed on retriev-
als is usually quite small, not sufficient to be a major concern. Another factor is that
it has been shown that while the wear on the condylar bearing surfaces of the rotat-
ing platform is small, when this wear is added to that between the tibial plate and
the lower surface of the plastic component, the total wear is similar to that of a
partially conforming artificial knee (Haider and Garvin 2008; Stoner et al. 2013).
Wear tracks on the tibial plate have shown circular arcs indicating rotation during
function, which has been confirmed in fluoroscopic kinematic studies (Lacour et al.
2014). Consequently the advantage of the rotating platform concept has now become
its rotational freedom, an advantage at surgery, and in the reduction in the torque
transmitted to the tibial fixation. Particularly when applied to a PS design, the rota-
tional freedom is useful in high flexion where typically a fixed-bearing PS design
can be too constrained (Meftah et al. 2012). A positive appraisal of the mobile bear-
ing concept, with a European perspective, was provided by Karol Hamelynck
(2006): “Three decades after its introduction, the mobile bearing design remains
relevant and important.”
Different Medial and Lateral Laxity and Stability
In the 1970s and 1980s, many studies were carried out on the laxity of the knee,
under both unloaded and loaded conditions (see Chap. 4), but there was little atten-
tion given to the relative laxities between the lateral and medial sides. For example,
in measuring rotational laxity, the center of rotation in the transverse plane was not
identified. In one study of knee motion controlled by the quadriceps during flexion,
there was an average of 8.6 mm posterior displacement of the center of the femur
accompanied by 15° external rotation of the femur (Rovick et al. 1991). That could
only mean an axis of internal-external rotation on the medial side of the knee, but
this was not highlighted. Nevertheless, such motion and laxity studies provided the
guidelines for a number of artificial knee designs of the CR and PS types. The
understanding of knee motion had been influenced by the four-bar linkage model in
the sagittal plane, where the femoral-tibial contact points displaced posteriorly with
flexion, guided by the cruciate ligaments (O’Connor et al. 1989). Hence, there were
two perspectives of knee mechanics, one focusing on the ligaments being elastic
and extensible allowing laxity, the other on their stiffness and in defining the motion.
In the mid-1990s, in a teaching session on knee motion at the London Hospital,
successive lectures by Michael Freeman and Peter Walker confused the audience.
Walker described rigid body motion of the femur relative to the tibia, with fixed axes
being defined in the femur and tibia. The external rotation and the posterior dis-
placement of the femur with flexion were shown separately. The posterior displace-
ment of the origin of the distal femur was referred to as “rollback,” giving the
impression that the femur displaced posteriorly on the tibia as the knee was flexed
from extension to flexion. Freeman on the other hand insisted that there was no
posterior displacement of the femur on the tibia and no rollback. Instead, he claimed
that as the femur was flexed, there were two simultaneous rotations, flexion and
axial rotation, but no displacements. However, Walker’s posterior displacement was
defined with respect to the origin of the femur relative to the origin of the tibia,
which would also occur in Freeman’s model. Both explanations of knee motion
were consistent, but it was not appreciated at the time. Nevertheless, Freeman’s
model was easier to visualize.
In subsequent studies, Freeman and other researchers further demonstrated the
medial pivot action as the knee was flexed using MRI studies of loaded and unloaded
knees for knee specimens and for living subjects (Freeman and Pinskerova 2005;
Pinskerova et al. 2004; Williams and Logan 2004). Prominent co-authors of this
work were Freeman himself; Vera Pinskerova, MD, a pathologist from Prague; and
Andy Williams, FRCS, an orthopedic resident from Saint Mary’s Hospital in
London. Their data was consistent with the anatomy of the knee whereby the medial
tibial plateau was dished and the meniscus relatively immobile, while on the lateral
side, the tibial plateau was convex, sloped downwards posteriorly, and with a mobile
meniscus. It seemed logical to apply this important finding to the design of an arti-
ficial knee. This idea was developed in England and in the USA at about the same
time, in the mid- to the late 1990s.
In England, Michael Freeman worked with Michael Tuke, PhD, a faculty mem-
ber at Imperial College, London, who later founded Finsbury Instruments. A US
patent was issued in 1993 describing a spherical conforming bearing medially and
a curved track laterally to allow internal-external rotation pivoting about the medial
mm mm
95°
M L 50
40
40 “–5°”
EF FF 5°“–5°” 31
30
EF 5° 26
27
10° 24 25
29 23
FF F 20° 22
TA +/- 1.5
24
22
23 90° 30°
21 mm 90° 45° 15
EF
110° 10
8
FF
120°
10
0
0
0 46 mm
Fig. 11.4 A basis for a medial pivot type of artificial knee. Knee specimens were flexed to differ-
ent angles and MRI scans taken. The succession of transverse flexion axes (white axis at left,
arrow) through the femoral condyles were projected on to the tibial surface. The femur appears to
pivot on the medial side as it flexes. The anterior positions near extension are due to femoral rolling
up the anterior facet on the tibial plateau (Iwaki et al. 2000). Subsequent MRI studies on subjects
showed that the axial rotation depended on different positions and muscle actions of the subjects,
but still with relatively small medial displacements. (Reprinted with permission from Journal of
Bone & Joint Surgery B)
Patella MEDIAL
1
flange
8 R
1
R R R R
Conformity in frontal Ball-in-socket

plane
6
1 LATERAL
MED LAT
r
13 12
11
2 10
Pivot Motion 2 4
along arc Ball-on-flat with anterior
point
& posterior restraint
Fig. 11.5 The first US patent for a medial pivot knee, filed in February 1992 by Michael Tuke and
Michael Freeman, of the University of London, England. The design was modified from the origi-
nal Freeman-Swanson knee by making a spherical bearing medially and by flattening the lateral
tibial surface. (US Patent 5,219,362. Filed 1992, issued 1993. Tuke and Freeman 1993)
side (Tuke and Freeman 1993). Tuke and Freeman described the invention as fol-
lows: “The present invention accordingly seeks to provide an improved knee pros-
thesis which will allow the patient to move his knee substantially in the same manner
as though he had not had to undergo implantation of a knee prosthesis”. The recently
introduced “Forgotten Joint Score” expresses a similar way of evaluating a patient’s
artificial knee.
Finsbury Instruments was already manufacturing the Freeman-Samuelson knee,
a further development of the original Freeman-Swanson. This design was then mod-
ified to produce the Medial Rotation Knee based on the patent. In a mean 6-year
follow-up (Mannan and Scott 2009), the loosening was not increased, and there was
an improved survivorship of 98% compared with the Freeman-Samuelson knee,
while the “deep medial dish provided enhanced stability without recourse to an
intercondylar cam and post.” This last phrase refers to a posterior stabilized (PS)
type of knee. Anterior-posterior stability of the knee has been noted by patients in
subsequent evaluations of the medial pivot knee, which represents an important
positive feature of this type of design.
The Medial Rotation Knee was used clinically until Finsbury Instruments was
purchased by DePuy in 2009. The knee was then redesigned by Mike Tuke who
formed a new company in 2011 called MatOrtho. The new design was called the
SAIPH Knee. Freeman and Pinskerova pursued the medial pivot concept separately
Fig. 11.6 Michael Tuke worked in Alan Swanson’s Biomechanics Laboratory in the late 1960s
and played a major role in early knee designs of Freeman and Swanson. Mike Tuke founded his
own company, Finsbury Orthopaedics in 1978, with a successful record of innovative designs for
the knee and hip. Among these designs was the Medial Rotation Knee, used now for 25 years.
Later, after the formation of a new company MatOrtho, based in Leatherhead, England, Mike Tuke
upgraded the knee, calling it the SAIFF knee. This knee continues to increase in popularity world-
wide. (Photo courtesy Michael Tuke)
through the Swiss company Medacta, calling their design the GMK Sphere (Scott
et al. 2016). Pinskerova’s explanation of the rationale was “The concept of the GMK
Sphere TKA is based on medial antero-posterior stability combined with lateral
freedom of motion, resulting in a medial center of longitudinal rotation as exhibited
by the normal knee… This movement is possible because the lateral collateral liga-
ment is loose in flexion. The combination of a fully congruent medial compartment
and an unconstrained lateral tibial surface together with tight medial collateral liga-
ment and loose lateral collateral ligament enables flexion accompanied by femoral
external rotation around the stable medial condyle” (Pinskerova 2016). This motion
was demonstrated in a computer model (Morra et al. 2012).
While the development of the medial pivot knee was proceeding in England, a
separate but related development took place in the USA, also from the mid-1990s.
The main driver of this work was David Blaha MD, from Ann Arbor, Michigan, who
had taken an interest in knee mechanics and total knee design since his residency in
the mid-1970s. At that time, Larry Matthews, MD; David Sonstegard, PhD; and
Herb Kaufer, MD, from the University of Michigan Medical Center in Ann Arbor,
were designing the Spherocentric knee (Matthews et al. 1975). As reported in a
paper to the Orthopedic Research Society (ORS) in 1975, this knee was conceived
as a combination of a hinge and a condylar replacement where a centrally embedded
ball-in-socket allowed rotational laxity such that the implant absorbed far less
energy when torque or bending were applied, compared with a fixed hinge, and
hence would move more naturally and experience less loosening. In principle the
Spherocentric had some similarity to the Attenborough, but the latter had more
Fig. 11.7 David Blaha

was one of the early
originators of the medial
pivot knee concept. His
early interest in
biomechanics was sparked
by working with Herb
Kaufer, MD, and Larry
Matthews, MD, at the
University of Michigan,
Ann Arbor in the
USA. Following a
fellowship with Mike
Freeman, he returned to
Ann Arbor to continue
experiments on knee
motion, which he applied
to a medial pivot knee
design. Working with Steve
White, this was
accomplished in the
Advance Medial Pivot
knee made by Wright
Medical. (Photo taken at
AAOS meeting, 2018) (©
Peter S Walker)
substantial fixation stems than the Spherocentric. David Blaha found an ideal envi-
ronment in the area of knee mechanics during a fellowship with Michael Freeman
at the London Hospital in the late 1970s. They were both skeptical that the knee
motion was guided by the cruciate ligaments in a four-bar link mechanism as first
proposed by Zuppinger and Strasser in the nineteenth century and later embellished
by O’Connor and others at Oxford (O’Connor et al. 1989). However, Freeman and
Blaha did not formulate any systematic experiments on which to base a more
advanced theory, other than making empirical observations that as the knee was
flexed, the medial femoral condyle appeared to be in a nearly constant position on
the tibia while the lateral condyle displaced posteriorly. On returning to the Ann
Arbor in the early 1980s, Blaha initiated a knee motion study on specimens using
Vicon equipment, which suggested that during flexion, the axis of rotation moved in
three dimensions and was not restricted to planar motion. Subsequent experiments
on knee specimens confirmed this, by using an “open chain” model where the femur
was fixed horizontally and the tibia was initially fully flexed (Blaha et al. 2003). The
tibia was then extended by pulling on the quadriceps tendon. Gradual external rota-
tion of the tibia was observed. A remarkable finding was that the motion changed
very little when the cruciates were excised, from which Blaha concluded that the
four-bar linkage model for guiding knee motion was inconsistent with the
Femoral
spherical
surface
Matching
A B tibial
spherical
surface
Fig. 11.8 David Blaha was one of the main proponents of the medial pivot concept in the
USA. This figure shows the main concepts. The basic principle was anterior-posterior stabilization
using a ball-in-socket on the medial side. (a) In a CR posterior cruciate retaining knee, just before
load acceptance. (b) On load acceptance, anterior sliding of the femur occurs. (c) In a medial pivot
knee the conforming medial side prevents the anterior sliding (Blaha 2004). (Reprinted with per-
mission from Journal of Arthroplasty)
experimental data. By 1991 a total knee design based on a more stable medial side
and an unconstrained lateral side had emerged. It took a long time to find interest
from a company, but the idea was eventually taken up by Wright Manufacturing in
Memphis, who were then making the 913 PS type design from the Hospital for
Special Surgery. The new concept became the Advance Medial Pivot due mainly to
the efforts of David Blaha and Steve White, but it was not until 1998 that the first
knees were used in patients. Other surgeons involved with the project were Murale
Jasty, Brad Penenberg, and Bill Maloney.
In the first few years of clinical application, there was uncertainty about how to
align the femoral component on the distal femur to produce the best balance of col-
lateral tightness in extension and flexion, and whether the posterior cruciate should
be retained or resected. Lack of high flexion in many cases was largely solved by
sloping the tibial cut posteriorly to the anatomic 5–7°. The idea of a medial pivot
knee was not as widely accepted as anticipated. There were several likely reasons.
The MRI studies by Freeman, Pinskerova, and others, showing the medial pivot
motion of the anatomic knee, had not yet been published, while the same applied to
Blaha’s own studies with open and closed chain experiments (Blaha et al. 2003).
Wright Manufacturing had limited marketing capabilities compared with Zimmer,
Stryker, DePuy, and Biomet. But probably the main factor was that the larger com-
panies had their flagship CR and PS designs well represented by prominent
surgeons from renowned institutions. A further factor was that there was an absence
of a sufficient body of evidence from follow-up studies that the medial pivot was
performing any better in patients than the CR and PS knees that dominated the mar-
ket. Even by 2002 (Blaha 2002), there was only one study of the Advance Medial
Pivot, in which it was stated: “In the author’s series, knees at two years follow-up
had an average Knee Society score of 91 and average 116° range of motion. In a
multicenter study of four hundred and one patients, the average range of motion was
118°. In a comparison study of forty-three patients with a medial pivot in one knee
and a CR or PS design in the other knee, a majority of patients rated the medial pivot
knee superior in ease of rehabilitation, motion, stability, and overall satisfaction.”
This latter study of bilateral cases was reported separately in a publication in
2004 (Pritchett 2004). In the same study, Pritchett also found that an ACL + PCL
retaining design was preferred over a CR or PS, while there was equal preference
for an ACL + PCL and medial pivot designs. In a further paper by Blaha (2004),
where the biomechanical principles of the medial pivot concept were expounded, he
emphasized the anterior-posterior stability throughout flexion as being the key char-
acteristic. In a later clinical paper, Pritchett considered different reasons for the
patient preferences (Pritchett 2011). In his view, the most important was the AP
stability provided, as well as an enhanced quadriceps lever arm. Other factors were
the paradoxical sliding of the CR and patella problems with the PS. A strength of
the studies by Pritchett was that these were bilateral cases, with different knee
designs in opposite sides. Hence, there is a consistency between these two authors,
although in the actual clinical trials, it was difficult to specifically identify anterior-
posterior stability from the standard clinical evaluation schemes used at the time.
The overall picture was that the medial pivot knee performed at least as well as
standard CR and PS knees and may have been superior in certain aspects. Recently
a 15-year average follow-up study was reported of 347 medial pivot knees (Macheras
et al. 2017). The average range of flexion was 120°, and the survivorship at 17 years
was 98.8%. On the important aspect of anterior-posterior stability, a major benefit
proposed for this design was “No post-operative instability was observed in our
study. All knees were objectively stable throughout the range 0–90°… The patients
in our study noted a great sense of stability, particularly in single-leg weight-bearing
positions, they felt strong on stairs and they judged their knees comfort during walk-
ing stable and natural overall… Undoubtedly the medial pivot design gives an
inherent stability to the TKA.” On the question of whether the posterior cruciate
ligament (PCL) should be retained, a difficult question when the design was intro-
duced, these authors found no difference in clinical results, but somewhat more
difficulty balancing the knee when the PCL was intact.
A different type of tibial insert where the medial side was more dished than the
lateral to provide more constraint was introduced in 2015. The medial side was not
fully conforming, in contrast to the medial pivot designs. Called the Medial
Congruent Bearing, this insert was to be used with the existing cruciate retaining
(CR) femoral component in the Persona Knee System. Designed by Steve White
and Bill Maloney, MD, this insert was a modification of Aaron Hofmann’s ultracon-
gruent insert. The Medial Congruent Bearing can be described as a “loose medial
pivot” in that there was still some anterior-posterior laxity, especially in flexion due
to the reducing radius of the femoral component. This design can be claimed to be
more anatomic than a fixed medial pivot, as the medial side of the anatomic knee
does have some anterior-posterior laxity, but much less than on the lateral side.
Guided Motion Knees with Biomimetic Motion or Geometry
In the past, the design of most standard condylar replacement knees has been based
on geometrical parameters, with ease of manufacture as an important design crite-
rion. In the sagittal plane of the femoral component, a larger distal radius has
blended into a smaller radius posteriorly. In the frontal plane, the radii have ranged
from anatomic at about 20 mm to being close to flat. The radii of the tibial surfaces
in the frontal and sagittal planes have been larger than those of the femoral compo-
nent to provide laxity in anterior-posterior and rotation. The depth and V-angle of
the patella flange has usually been based on anatomic values. Intercondylar cam-
posts have further constrained the relative motion, mainly in the anterior-posterior
direction. At surgery, ligament balancing has been carried out so that the ligaments
still perform their primary functions, particularly in providing varus-valgus stabil-
ity. Such condylar knees, even with symmetric condyle shapes, can move asym-
metrically, but not in the same way as an anatomic knee.
The design goal of Guided Motion knees is to go one step further than a standard
condylar knee in guiding or controlling the femoral-tibial motion, to more resemble
the motion of the normal anatomic knee. That said, it is not easy to define what
“normal anatomic knee motion” is, because there are many activities an individual
can perform, and every individual’s knee moves differently. There are even differ-
ences in performing the same function repetitively. Using knee specimens, the
motion has been measured in passive flexion, in flexion under quadriceps action,
and where forces act in both the quadriceps and hamstrings. In the living subject,
fluoroscopy has been used to measure motion in squatting, in leg raising, lunging,
and step and chair rise. Now that mobile fluoroscopy units are available, motion has
been measured in regular walking and in other activities. In an attempt to summa-
rize, there are certain motion characteristics that are common to everybody. For
level walking studied on a mobile fluoroscopy unit, Gray et al. (2019) stated:
“Anterior-posterior translations of the tibio-femoral contact centers were signifi-
cantly greater in the lateral compartment than the medial compartment. Posterior
translations… were associated with increasing knee flexion. The center of rotation
of the knee in the transverse plane was located predominantly on the medial side…”
In walking, however, the flexion angles during the loaded stance phase are limited
to about 20°. In a lunge activity up to 115° flexion, again using fluoroscopy, Feng
et al. (2015) found that the medial femoral condyle translated first anteriorly and
then posteriorly up to maximum flexion. The lateral side moved continuously pos-
teriorly, but with much larger displacements than on the medial side. The internal
rotation of the tibia was 27°. At maximum flexion, the lateral condyle was at the
very posterior of the tibia (risky for an artificial knee). Another important measure
of knee motion is the laxity. At a given flexion angle on the neutral path of motion,
the magnitudes of the anterior-posterior laxity and internal-external laxity are an
important parameter in defining motion. Nobody designing an artificial knee with
anatomic motion would rigidly constrain the motion to the average. The values of
laxity depend on the compressive load acting, but in all conditions, the medial side
has much less anterior-posterior laxity compared with the lateral side (Reynolds
et al. 2017). It can be concluded that if a Guided Motion knee is to mimic anatomic
motions in function, the displacements and rotations as they are known at this time
should be reproduced.
There have been several attempts to achieve anatomic knee motion. The first was
to use a “tibial surface generation” method. This was first described by Ewald and
by Seedhom in the 1970s (see Chap. 9). An anatomically shaped femoral compo-
nent was moved in a flexion path, including some laxity, to generate the tibial sur-
face. This process was later carried out by Walker and Ewald (1988) using computer
graphics, where motion data from knee specimens was input. The tibial surfaces
generated from the above were all similar, a relatively shallow lateral side and a
more concave medial side. This “tibial surface generation” method (Walker
2001) was later expanded upon by starting with different femoral component sur-
faces (Walker 2014). This included a converging condyle version, which had been
patented in 2001 by Takei. In addition, intercondylar surfaces and housings on the
femur produced ramps or projecting posts on the central tibia.
Yet further application of the “tibial surface generation” method was by a group
at Massachusetts General Hospital, including Harry Rubash, MD; Guaon Li, PhD;
and Varadarajan, PhD (Varadarajan et al. 2015; Zumbrunn et al. 2015). Starting with
a femoral component, they input motion data obtained from Guaon Li, PhD, using
fluoroscopy of normal subjects. What was unique about their method was that the
motion comprised all 6 degrees of freedom, including vertical displacements and
varus-valgus rotations. The resulting tibial surfaces were “biomimetic” in that the
medial tibial surface was moderately dished but the lateral surface was convex in the
sagittal plane. At the anterior and posterior of each surface were lips representing
the spaces that would be occupied by the menisci at the extremes of motion. The
design was called the biomimetic knee. In the version of the knee designed by
Varadarajan, the anterior cruciate ligament was resected. Nevertheless, using
KneeSim software for motion analysis, it was found that “the biomimetic knee
showed an overall medial pivot and greater rollback of the lateral femoral condyle
than the medial condyle.” In the version analyzed by Zumbrunn, the anterior cruci-
ate was retained with the proposition that there would be even further reproduction
of anatomic motion. In this version, certain nuances of the motion were found be
even closer to normal. A computer modelling approach to a similar concept was
described by Liu et al. (2012).
In about 2004, a different approach was taken to designing a guided motion knee.
It was recognized by Jan Victor, MD; Johan Bellemans, MD; Michael Ries, MD; and
others that the limited range of flexion achieved by artificial knees was inhibiting a
number of functional activities where higher flexion was required. It was proposed
that if normal kinematics could be achieved, together with other factors such as
Fig. 11.9 Tsunenori Takei 140°

designed a total knee in
2001 where the lateral
femoral condyle converged
towards high flexion. This
facilitated posterior
rollback of the lateral
femoral condyle with axial 120°
rotation, increasing the
range of flexion. (US
Patent 6,406,497 B2. Filed
2001, issued 2002)
A I
90°
B
T
A C
T
8 8b
8a
60°
30°
8 II
0°
condylar offset, flexion could be increased (Victor and Bellemans 2006). The design
group decided to use the PS concept but to make the lateral tibial surface almost flat:
“Preserving the anterior and posterior cruciate ligament has been attempted and
resulted in more normal kinematics and better flexion. Surgical feasibility limits the
potential of this solution. The other solution is to provide kinematic guidance and
stabilization by the prosthetic design through surface geometry and the cam-post
mechanism.” This arrangement produced posterior displacement with flexion, but
substantially more laterally than medially. The name of the device was the Journey
BCS knee introduced in 2005. In several clinical follow-up studies, the knee did
achieve a high range of flexion, but there were some complications of instability, pos-
sibly due to excessively posterior contacts and rotation. Accordingly, a redesign was
made by moving contacts anteriorly, the modified design being evaluated by fluoros-
copy (Zambianchi et al. 2018). The results were “The design adjustments performed
onto the second-generation prosthesis to limit excessive posterior displacement of the

femoral condyles, as well as femoral component rotation in weight-bearing condi-
tions, contributed to modify replaced knee’s kinematics during daily living activities,
as confirmed by fluoroscopy data.” This Journey design is biomimetic in certain ways,
but relying on the mechanics of a cam-post to provide much of the motion guidance
reproducibly, rather than relying on ligaments which vary with the individual.
In 2008, a biomimetic type of guided motion knee was conceived, based on a
kinematic analysis (Floerkemeier et al. 2011; Frosch et al. 2009). The originality
was that the bearing surfaces of the tibial component were asymmetric, convex for
the lateral side and concave for the medial, together with a sagittal offset between
the lateral and medial profiles. The preference for rolling rather than sliding would
reduce the wear of the polyethylene. The idea was the basis of an artificial knee
design called the Aequos knee. After clinical application, fluoroscopic studies were
carried out which showed that the patella tendon angle resembled that of normal
knees (Wachowski et al. 2011). This was due to the femoral rollback which reduced
the angle with flexion, reducing the force between the patella and the anterior femur.
A further approach was to introduce a third bearing between the lateral and
medial tibial bearing surfaces, which would replace the function of the anterior
cruciate in early flexion, by limiting the anterior-posterior displacements and main-
taining an anterior position of the femur (Walker and Borukhov 2017). The work
Anatomically shaped
Femoral Component
Continuous
groove for
Patella
AS
Central
PL surface
prevents
posterior
femoral
displacement
Sloped lateral surface
up to mid-
promotes rollback in Medial surface reproduces
flexion
higher flexion, limited by anterior slope AS and limits
posterior lip PL AP displacement
Fig. 11.10 A guided motion knee based on reproducing the stabilizing mechanisms of the natural
structures and on reproducing the normal motion patterns. There are three bearing surfaces each
performing different functions, with the medial side being more conforming than the lateral, to
produce anatomic motion, laxity, and stability characteristics (Walker and Borukhov 2017). (CAD
models by Stephen Chan)
was carried out in Walker’s Laboratory for Orthopaedic Implant Design at the NYU
Langone Orthopedic Hospital. The design was called the Replica Anatomic knee.
This guided motion design was based on replicating the laxity and stability nor-
mally provided by the cruciate ligaments and the menisci, which had been deter-
mined in previous biomechanical studies (Walker et al. 2015; Reynolds et al. 2017).
The testing method for the design was to impose a sequence of anterior-posterior
and internal-external forces on the knee, over a full flexion range and measure the
laxities. The laxity values were then compared with those from an average of 23
anatomic knees (Walker et al. 2015). It was shown that the laxity values, and the
lateral and medial displacement paths, were closer to normal than a typical PS knee.
For the latter there was much higher rotational constraint in early and late flexion,
compared with that of the anatomic knee.
Another example of a knee where the intercondylar space was used to provide
extra guidance was a Third Condyle knee designed by Bytgyqi et al. (2017). The
concept was similar to that above, but this design had already been subject to clini-
cal evaluation. It was found that while the motion was still not similar to anatomic,
the approach had further possibilities.
A refined way to address guided motion knee design was later devised by Amiri et al.
(2011) where specific kinematic functions were assigned to each region of the bearing
surfaces. The potential problem of overconstraining the knee was addressed as follows:
“the ultimate shape of the tibial component… can be defined by tuning the shapes and
clearances until an optimal compromise between guiding and laxities is formed.” The
aim of this idea was to minimize the wear of the polyethylene.
1 4
2
4 4
8
posterior
anterior
medial
lateral
medial frontal
Fig. 11.11 A Guided Motion knee where the surfaces of the femoral and tibial bearing surfaces
are shaped in different regions (numerals on figure) to provide the required guidance to motion,
together with optimizing the surfaces for wear. The target motion is close to a medial pivoting
action. Cruciate preservation is not essential, but can provide anatomic fine-tuning to the motion
(Amiri et al. 2011). (Reprinted with permission from Journal of Medical Devices ASME)
o Guided Motion Designs Have Advantages Over the CR

D
and PS?
Rotating platform and medial pivot designs have shown clinical results equivalent
to those of standard CR and PS knees. There is evidence that they provide a feeling
of anterior-posterior stability. This effect may be more apparent in a medial pivot
knee due to the medial location of internal-external rotation, whereas in a rotating
platform, the line of anterior-posterior stability is central. However, the anatomic
knee does not have absolutely fixed axes for axial rotation. Guided Motion designs
which closely replicate the shapes of the normal anatomic knee have the advantages
of a more accurate fit and correct “flow” of the soft tissues around the components.
Guidance of the motion from the bearing surfaces, without requiring the cruciate
ligaments, has the advantage of greater reproducibility and ease of surgery.
Producing more anatomic motion and laxity will provide more accurate lever arms
for the muscles. Evaluation of guided motion designs can conceivably be carried out
by a computer modeling method (Rullkoetter et al. 2017). These authors point out
the complexity of a modeling system but are optimistic that “future TKA designs
will be optimized using preclinical computational modelling.” This approach can
evaluate guided motion knees without the considerable expense of a company
financing physical production and clinical trials. It is likely that future knee systems
will include Guided Motion knees based on the evidence so far that these type of
designs more closely simulate normal kinematics in different functions, as well as
embodying normal laxity and stability. The most practical guided motion design
will be one which relies on the mechanical features to provide the anatomic motion,
laxity, and stability, without having to rely on the cruciate ligaments which require
a high accuracy of surgery which cannot always be achieved. A key question is,
however, to what extent such normal kinematics will be felt by the patient and what
type of evaluation method should be used.
References
Amiri S, Cooke TDV, Wyss UP. Conceptual design for condylar guiding features of a total knee
replacement. J Med Devices. 2011;5(2):025001.
Blaha JD. A medial pivot geometry. Orthopedics. 2002;25(9):963–4.
Blaha JD. The rationale for a total knee implant that confers anteroposterior stability throughout
range of motion. J Arthroplast. 2004;19(4 Suppl 1):22–6.
Blaha JD, Mancinelli CA, Simons WH, Kish VL, Thyagarajan G. Kinematics of the human knee
using an open chain cadaver model. Clin Orthop Relat Res. 2003;410:25–34.
Buechel FF, Pappas MJ. The New Jersey low-contact-stress knee replacement system: biome-
chanical rationale and review of the first 123 cemented cases. Arch Orthop Trauma Surg.
1986;105(4):197–204.
Buechel FF Sr, Buechel FF Jr, Pappas MJ, D’Alessio J. Twenty-year evaluation of meniscal bear-
ing and rotating platform knee replacements. Clin Orthop Relat Res. 2001;388:41–50.
Bullough P, Goodfellow J, Greenwald AS, O’Connor J. Incongruent surfaces in the human hip
joint. Nature. 1968;217(5135):1290.
Bullough PG, Munuera L, Murphy J, Weinstein AM. The strength of the menisci of the knee as it
relates to their fine structure. J Bone Joint Surg. 1970;52(3):564–7.
Bytgyqi D, Shabani B, Cheze L, Neyret P, Lustig S. Dores a third condyle TKA restore nor-
mal gait kinematics in varus knes? In vivo kinematic analysis. Arch Orthop Trauma Surg.
2017;137(3):409–16.
Eftekhar NS. Total knee-replacement arthroplasty. Results with the intramedullary adjustable total
knee prosthesis. J Bone Joint Surg Am. 1983;65(3):293–309.
Evanski PM, Waugh TR, Orofino CF, Anzel SH. UCI knee replacement. Clin Orthop Relat Res.
1976;120:33–8.
Feng Y, Tsai T-Y, Li J-S, Wang S, Hu H, Zhang C, Rubash H, Li G. Motion of the femoral condyles
in flexion and extension during a continuous lunge. J Orthop Res. 2015;33:591–7.
Floerkemeier T, Frosch KH, Wachowski M, et al. Physiologically shaped knee arthroplasty induces
natural roll-back. Technol Health Care. 2011;19(2):91–102.
Freeman MAR, Pinskerova V. The movement of the normal tibio-femoral joint. J Biomech.
2005;39:197–208.
Frosch KH, Floerkemeier T, Abisht C, Adam P, Dathe H, Fanghanel J, Srurme KL, Kubein-
Meesenburg D, Nagerl H. A novel knee endoprosthesis with a physiological joint shape. Part
1: biomechanical basics and tribological study. Unfallchirurg. 2009;112(2):1687–175.
Goodfellow J, O’Connor J. The mechanics of the knee and prosthesis design. J Bone Joint Surg.
1978;60-b(3):358–69.
Gray HA, Guan S, Thomeer LT, Schache AG, Stieger R, Pandy MG. Three-dimensional motion
of the knee-joint complex during normal walking revealed by mobile biplane x-ray imaging. J
Orthop Res. 2019;37:615–30.
Greenwald AS, Haynes DW. Weight-bearing areas in the human hip joint. J Bone Joint Surg.
1972;54(1):157–63.
Haider H, Garvin K. Rotating platform versus fixed-bearing total knees: an in vitro study of wear.
Clin Orthop Relat Res. 2008;466(11):2677–85.
Hamelynck KJ. The history of mobile-bearing total knee replacement systems. Orthopedics.
2006;29(9 Suppl):S7–12.
Iwaki H, Pinskerova V, Freeman MAR. Tibiofemoral movement 1: the shapes and rela-
tive movements of the femur and tibia in the unloaded cadaver knee. J Bone Joint Surg Br.
2000;82(8):1190–5.
Krause WR, Pope MH, Johnson RJ, Wilder DG. Mechanical changes in the knee after meniscec-
tomy. J Bone Joint Surg Am. 1976;58(5):599–604.
LaCour MT, Sharma A, Carr CB, Komistek RD, Dennis DA. Confirmation of long-term in vivo bear-
ing mobility in eight rotating-platform TKAs. Clin Orthop Relat Res. 2014;472(9):2766–73.
Liu YL, Chen WC, Yeh WL, et al. Mimicking anatomical condylar configuration into knee pros-
thesis could improve knee kinematics after TKA – a computational simulation. Clin Biomech
(Bristol, Avon). 2012;27(2):176–81.
Macheras GA, Galanakos SP, Lepetsos P, Anastasopoulos PP, Papadakis SA. A long term clinical
outcome of the Medial Pivot Knee Arthroplasty System. Knee. 2017;24(2):447–53.
Mannan K, Scott G. The Medial Rotation total knee replacement: a clinical and radiological review
at a mean follow-up of six years. J Bone Joint Surg. 2009;91(6):750–6.
Matthews LS, Sonstegard D, Kaufer H. Impact considerations in prosthetic loosening. In:
Proceedings Orthopaedic Research Society, San Fransisco, 27 February–1 March, 1975.
Meftah M, Ranawat AS, Ranawat CS. Ten-year follow-up of a rotating-platform, posterior-
stabilized total knee arthroplasty. J Bone Joint Surg Am. 2012;94(5):426–32.
Minns RJ. The Minns meniscal knee prosthesis: biomechanical aspects of the surgical procedure
and a review of the first 165 cases. Arch Orthop Trauma Surg. 1989;108(4):231–5.
Minns RJ, Campbell J. The mechanical testing of a sliding meniscus knee prosthesis. Clin Orthop
Relat Res. 1978;137:268–75.
Morra EA, Heim CS, Greenwald AS. Preclinical computational models: predictors of tibial insert
damage patterns in total knee arthroplasty: AAOS exhibit selection. J Bone Joint Surg Am.
2012;94(18):e137(131–5).
Murray DG. Total knee replacement with a variable axis knee prosthesis. Orthop Clin North Am.
1982;13(1):155–72.
O’Connor JJ, Shercliff TL, Biden E, Goodfellow JW. The geometry of the knee in the sagittal
plane. Proc Inst Mech Eng H J Eng Med. 1989;203(4):223–33.
Pandit H, Hamilton TW, Jenkins C, Mellon SJ, Dodd CA, Murray DW. The clinical outcome of
minimally invasive Phase 3 Oxford unicompartmental knee arthroplasty: a 15-year follow-up
of 1000 UKAs. Bone Joint J. 2015;97-b(11):1493–500.
Pinskerova V. GKM sphere medially stabilized knee. 2016. https://media.medacta.com/media/
gmk-sphere-publication-review-99-26sphere-pubus-rev00.pdf
Pinskerova V, Johal P, Nakagawa S, Sosna A, Williams A, Gedroyc W, Freeman MAR. Does the
femur roll-back with flexion? J Bone Joint Surg. 2004;86(6):925–31.
Price AJ, O’Connor JJ, Murray DW, Dodd CA. Goodfellow JW. A history of Oxford unicompart-
mental knee arthroplasty. Orthopedics. 2007;30(5 Suppl):7–10.
Pritchett JW. Patient preferences in knee prostheses. J Bone Joint Surg. 2004;86(7):979–82.
Arthroplast. 2011;26(2):224–8.
Reynolds RJ, Walker PS, Buza J. Mechanisms of anterior-posterior stability of the knee joint under
load-bearing. J Biomech. 2017;57:39–45.
Rovick JS, Reuben JD, Schrager RJ, Walker PS. Relation between knee motion and ligament
length patterns. Clin Biomech (Bristol, Avon). 1991;6(4):213–20.
Rullkoetter PJ, Fitzpatrick CK, Clary CW. How can we use computational modeling to improve
total knee arthroplasty? Modeling stability and mobility in the implanted knee. J Am Acad
Orthop Surg. 2017;25(Suppl 1):S33–s39.
Scott G, Imam MA, Eifert A, et al. Can a total knee arthroplasty be both rotationally uncon-
strained and anteroposteriorly stabilised? A pulsed fluoroscopic investigation. Bone Joint Res.
2016;5(3):80–6.
Seedhom BB. Loadbearing function of the menisci. Physiotherapy. 1976;62(7):223.
Shaw J, Murray DG. Knee joint simulator. Clin Orthop Rel Res. 1973;94:15–23.
Shaw JA, Murray DG. The longitudinal axis of the knee and the role of the cruciate ligaments in
controlling transverse rotation. J Bone Joint Surg Am. 1974;56(8):1603–9.
Stoner K, Jerabek SA, Tow S, Wright TM, Padgett DE. Rotating-platform has no surface damage
advantage over fixed-bearing TKA. Clin Orthop Relat Res. 2013;471(1):76–85.
Takei T. Artificial knee joint. US Patent 6,406,497. Filed July 2001, issued June 2002.
Tuke MA, MAR Freeman. Knee prosthesis. US Patent 5,219,362. Filed Feb 1992, issued June
1993.
Ueo T, Kihara Y, Ikeda N, Kawai J, Nakamura K, Hirokawa S. Deep flexion-oriented bisurface-type
knee joint and its tibial rotation that attributes its high performance of flexion. J Arthroplast.
2011;26(3):476–82.
Varadarajan KM, Zumbrunn T, Rubash HE, Malchau H, Li G, Muratoglu OK. Cruciate retaining
implant with biomimetic articular surface to reproduce activity dependent kinematics of the
normal knee. J Arthroplast. 2015;30(12):2149–53.
Victor J, Bellemans J. Physiologic kinematics as a concept for better flexion in TKA. Clin Orthop
Rel Res. 2006;452:53–8.
Wachowski MM, Fiedler C, Walde TA, Balcarek P, Schüttrumpf JP, Frosch S, Frosch KH,
Fanghänel J, Gezzi G, Kubein-Meesenburg D, Nägerl H. Construction-conditioned rollback in
total knee replacement: fluoroscopic results. Acta Bioeng Biomech. 2011;13(3):35–42.
Walker PS. A new concept in guided motion total knee arthroplasty. J Arthroplast. 2001;16(8
Suppl 1):157–63.
Walker PS. Application of a novel design method for knee replacements to achieve normal
mechanics. Knee. 2014;21(2):353–8.
Walker PS, Borukhov I. Replication and substitution of anatomic stabilizing mechanisms in a total
knee design. J Med Devices. 2017;11(4):041005.
Walker PS, Erkman MJ. The role of the menisci in force transmission across the knee. Clin Orthop
Relat Res. 1975;109:184–92.
Walker PS, Ewald FC. Bearing surface design in total knee replacement. Eng Med.
1988;17(4):149–56.
Walker PS, Arno S, Borukhov I, Bell CP. Characterizing knee motion and laxity in a testing
machine for application to total knee evaluation. J Biomech. 2015;48(13):3551–8.
Waugh TR, Smith RC, Orofino CF, Anzel SM. Total knee replacement: operative technic and pre-
liminary results. Clin Orthop Relat Res. 1973;94:196–201.
Williams A, Logan M. Understanding tibio-femoral motion. Knee. 2004;11(2):81–8.
Zambianchi F, Fiacchi F, Lombari V, Venturelli L, Marcovigi A, Giorgini A, Catani F. Changes in
total knee arthroplasty design affect in-vivo kinematics in a redesigned total knee system: a
fluoroscopy study. Clin Biomech. 2018;54:92–102.
Zumbrunn T, Varadarajan KM, Rubash HE, Malchau H, Li G, Muratoglu OK. Regaining native
knee kinematics following joint arthroplasty: a novel biomimetic design with ACL and PCL
preservation. J Arthroplast. 2015;30(12):2143–8.
Instrumentation and Technique
Are Mechanical Instruments
12
Accurate Enough?
The early devices for the treatment of osteoarthritis, tibial spacers and fixed hinges,
required little in the way of instrumentation for their insertion. For the spacers, the
tibial resection was achieved using an osteotome, with the local bone geometry as
the guide for the direction and depth of the cut. For the hinges, a saw was used to
resect the distal femur and proximal tibia with a T-shaped rod and plate for aligning
the cuts relative to the canals. The overall alignment of the component was defined
by the long intramedullary stems. Once the condylar replacement type was intro-
duced however, two problems were highlighted: how to prepare the bones for accu-
rate fitting of the component and how to achieve the required overall alignment of
the femur relative to the tibia.
Principles of Alignment and Force Transmission
To achieve these goals, the instrumentation and technique were based on a combina-
tion of anatomic and biomechanical considerations. Key reference points on the
bones were identified as the basis for reference lines or axes. In the frontal plane, the
points included the center of the femoral head, the centers of the distal femur and
proximal tibia, the extreme lateral and medial points of the distal femur and proximal
tibia, the center of the ankle level with the malleoli, and the centers of the diaphysis
of the femur and tibia. These points were then used to define the anatomic axes of the
femur and the tibia, the mechanical axis of the limb (as distinct from the mechanical
axis of the femur and tibia separately), and the joint line. In the sagittal plane, the
anatomic axis of the femur and the tibia and the tibial slope were defined. The angle
between the femoral and tibial anatomic axes in the sagittal plane denoted the flexion
angle although there is some uncertainty due to the anterior bowing of the femur.
In his book on kinesiology, first published in 1955, Arthur Steindler MD showed
the frontal plane axes as described above and concluded that “because of the obliq-
uity of the anatomic axis of the femur the greater amount of pressure is borne by the

206 12 Instrumentation and Technique
Fig. 12.1 In the early 1970s, the instrumentation for inserting the Duo-condylar knee consisted
only of two tools, one for shaping the femur and one for positioning the tibial components. The
surgeon held the femoral tool over the distal femur and marked the area which was to be recessed
about 3 mm so the component would just project from the surrounding surface. It was not easy to
match the shape of the anterior bridge. The tibial tool did position correctly spaced and parallel
tibial bearing surfaces. Very soon afterwards, alignment guides were designed to check alignment
of the components to the long axes of the femur and tibia
lateral condyle, and a greater tension is sustained by the soft structures of the medial
condyles.” (Steindler 1955). This perception was based on the assumption that a
vertical force would be applied from the top of the femoral shaft, the line of action
passing outside of the lateral condyle and hence applying the highest load on that
side of the knee. This misconception was due to the lack of biomechanical knowl-
edge at that time. By the mid-1960s however, based on radiographic observations of
arthritic knees with varus deformity, it was clear that cartilage loss and densification
of the trabecular bone were due to medial overloading, a state that became worse as
the deformity progressed. This was the basis for tibial osteotomy described by Mark
Coventry MD and others at the Mayo Clinic (Coventry 1965). By 1969, James
Morrison PhD and John Paul PhD from the University of Strathclyde, Glasgow,
Scotland, had concluded from their in vivo biomechanical studies that the larger
12 Instrumentation and Technique 207
proportion of the force on the knee passed through the medial side of the joint, even
in normal subjects (Morrison 1970). Later, Ed Chao PhD, Mark Coventry MD, and
Frank Sim MD developed software to calculate the lines of action of the resultant
force across the knee in varus deformity and calculated the ideal alignment for an
osteotomy in order to centralize the forces. In their analysis they defined the major
force of the femur on the tibia being applied along a line from the center of the
femoral head to the center of the tibial plateau.
Nevertheless, in some designs of hinged knees, the femoral and tibial stems were
co-axial, not reproducing the 5 degrees of valgus between the femur and tibia as in
the normal knee. Hence it is not surprising that in the early 1970s, when new designs
of condylar replacement knees were being introduced, there was inadequate under-
standing of the correct frontal plane alignment to apply to the surgical procedure.
Indeed at the conference on Total Knee Replacement organized by the Institution of
Mechanical Engineers and the British Orthopaedic Association in 1974, the brief
instructions for the Freeman-Swanson total knee included “The bone ends are
resected at right angles to their long axes to leave two flat cancellous surfaces paral-
lel to each other.” (Freeman et al. 1974). This would leave the knee joint in about 5
degrees of varus deformity on average. Other prominent participants at the confer-
ence, who had also developed new artificial knees, omitted to emphasize the impor-
tance of frontal plane alignment or describe how it was achieved at surgery. This
was at odds with a statement made by Ian Harrington PhD from Strathclyde
University at the same conference: “.…joint load tending to increase with increas-
ing joint angular deformity. This is more apparent with varus and flexion deformi-
ties of the knee than for valgus.…the normal pattern is to unload the lateral
component.” This was however recognized by Mark Coventry and Richard Bryan
from the Mayo Clinic in relating their experience with the Polycentric and Geometric
devices: “If the limb is in varus it must be overcorrected to 5 degrees of valgus….”
(Coventry and Bryan 1974). The development of an axial guide had made it possible
to estimate the axial alignment accurately. The consequence of too little attention
being given to frontal alignment was tibial component loosening, mostly due to
excessive varus alignment. It was this, rather than applying the earlier experience
with osteotomies or the biomechanical data, that led to the recognition of the prob-
lem and the development of reliable alignment guides.
Even so, designers of the early condylar knees seem to have relied on their own
unpublished studies of the normal alignments of the lower extremity, as there did
not appear to be any published study on the topic. As evidence of this, in the papers
of the First Scientific Meeting of the Knee Society in 1984 edited by Larry Dorr
MD, in a paper on Axial Alignment, Moreland and Hanker wrote: “Since some
authors have reported that the alignment of the lower extremity and component
positioning are important for durability….and since we were unable to find a radio-
logical study of normal alignment….”. Their study was republished a few years later
in the Journal of Bone and Joint Surgery (Moreland et al. 1987). The frontal plane
alignment does have an effect on the force distribution between the lateral and
medial sides of the knee, as implied by Moreland et al. It might be assumed that on
average, there would be equal forces through the lateral and medial sides of the
Femoral head center
Femoral mechanical axis
Knee Femoral shaft center I

physiologic Knee physiologic
valgus valgus angle II (Angle C)
angle I
(Angle D) Femoral anatomic axis II
Femoral Femoral shaft center II
anatomic
axis I
Knee center
Knee transverse axis
Angle B
Angle A
Tibial mechanical axis
Ankle center
Ankle transverse axis
Angle E
Fig. 12.2 One of the first studies where reference axes and angles were defined in the frontal
plane and average values measured. It was found that the mechanical axes of the femur and tibia
were not collinear, with an angulation of about 1.5 degrees, implying that the mechanical axis of
the limb (center of hip to center of ankle, not shown in figure) passed medial to the center of the
knee. The joint line, called the knee transverse axis, was angled at 2–3 degrees to the mechanical
axis of the tibia, a slight varus slope (Moreland et al. 1987). (Reprinted with permission from
Journal of Bone & Joint Surgery A)
knee. This was found to not be the case in a monumental study carried out by Georg
Bergmann PhD and associates at the Charite Hospital in Berlin. Total knee implants
were instrumented and the force data extracted wirelessly. The results from five
patients, carrying out activities including walking, going up and down the stairs, and
sit-to-stand, were published (Kutzner et al. 2010; Halder et al. 2012). The peak
forces were in the range of 2.2–3.5 times body weight, the highest values being for
stair ascend and descend. For an alignment between the anatomic axes of the femur
and tibia of 5 degrees, a normal mean value, 65–73% of the force passed through the
medial side. Every 1 degree more varus, the medial share value increased by 5%.
This might imply that an excess of varus would be detrimental for component wear
and loosening and the possibility of lateral lift-off. The resultant force on the tibia
was found to be predominantly in a direction in-line with the long axis. However
there were deviations from vertical in certain phases of activity, especially stair
activities, where there were shear force components. These could act in an anterior
or posterior direction on the tibia, of magnitude about 15–20% body weight. There
were even medial-lateral shear force components, somewhat less at about 10% body
weight. This data is relevant to the posterior slope of the tibial component, which
will affect the anterior shear forces as well as the femoral-tibial position. In general,
alignments of the components will affect the force distributions and hence the sta-
bility and kinematics of both the femoral-tibial joint and the patella-femoral joint.
Development of Mechanical Instruments
Applying principles of alignment to instrument design was illustrated by the work

of Charles O. Townley, a surgeon in Port Huron, Michigan, who started this work in
the late 1940s (Pritchett 1988). Before entering medical school, Townley graduated
with a degree in physical education, after which he spent a year studying mechani-
cal engineering. During his residency he formulated solutions for treating the
arthritic knee and actually made drawings of a design of total knee consisting of
metallic shells which accurately reproduced the femoral and tibial surfaces (Townley
1988). He did not implement this, however, because of the likely problems with
wear and friction, but he saw that there was utility in using only the tibial compo-
nent. His stainless steel articular-plate replacement as it was called was used from
1953 to 1972. In cases of varus or valgus deformity with wear and bone compres-
sion on one side of the tibia, Townley used different methods for correction. With
the leg straightened, the defect was filled with bone chips or by an osteotomy and
elevation of the bone near the plateau surface: “This results in a tightening of the
collateral ligament on the involved side and correction of the disturbance in the
weight-bearing line at the knee joint. It is suspected that the increased stability.…
plays a major role in the postoperative relief of pain.” However by the late 1960s,
with the availability of the materials being used for the total hip and the early knees
of Freeman-Swanson and Gunston, Townley was once again able to turn his atten-
tion to his own total knee design, to which his principles on knee arthroplasty in the
articular-plate replacement could be applied. These principles were “to provide an
articular replacement procedure that would simulate the normal anatomic and func-
tional modalities of the knee joint.” To achieve this required a combination of
implant design and technique including accurate bearing surface shapes and place-
ment, in synchrony with preservation of all of the ligaments, as well as correct
alignment of the long axes of the femur and tibia and inclination of the joint line:
“Significant abnormalities in either the size, contour or placement of the femoral

component will distort or misplace the condylar articular outline and will place it
out of tune with the holding patterns of motion sustained by the ligaments.” Townley
emphasized the inclination of the mechanical axis of the leg in the frontal plane in a
standing or functional position and the disposition of the joint line. This meant that
in a standing position, the joint lines of the right and left legs were parallel to the
ground. Consequently, the joint line was inclined at 2 degrees to the mechanical axis
and to the anatomic axis of the tibia itself. In the sagittal plane, the tibial cut should
be at the patient’s natural angle, in the range of 6–10 degrees posterior slope. It is
notable that Townley was the first to emphasize the approach of achieving normal
anatomic alignments for each patient.
Townley designed special cutting guides to achieve these various goals. To accu-
rately identify the center of the femoral head, the proximal reference point of the
mechanical axis, radiopaque markers were placed at the level of the femoral head
and a radiograph taken at the start of the operation. A length of thick cord, regarded
as an “instrument,” held taut between the centers of the femoral head and ankle,
should pass through the medial condyle of the knee. This strict anatomic approach
of Townley was intended to restore the normal ligament length patterns and, in turn,
the normal stability and laxity of the knee being operated on. It is consistent with the
data on the mechanical properties of ligaments, obtained subsequent to Townley’s
work, showing the sensitivity of knee motion to even small changes in ligament
lengths of only 1–2 millimeters. This puts into perspective the difficult task of the
implant designer and the surgeon in achieving such accuracy. One aspect that
Townley did not emphasize during the description of his technique was how to
handle deformities where normal ligament lengths could have changed, for exam-
ple, elongation of the lateral collateral and contraction of the medial collateral in
varus deformity.
The problem of implanting total knees in cases with deformity was a particular
focus of John Insall and Chit Ranawat, from the Hospital for Special Surgery: “.…
the problems of accurate placement increases with the preoperative deformity and
contracture.…” (Insall et al. 1976). Starting in about 1972, a surgical technique and
a set of instruments were developed for dealing with these situations. The Total
Condylar knee was designed to treat a full range of cases from mild to severe arthri-
tis, a strategy which had the advantage of a standard implant and technique which
could be readily learned and provide consistent results. The resection of the cruciate
ligaments made the technique easier, and because the anterior-posterior stability of
the joint was now provided by the dishing of the tibial bearing surfaces, cruciate
resection was not seen as a disadvantage. The essence of the technique was to make
the bone cuts so that the soft tissue sleeve around the knee was balanced in both
flexion and extension, rather than implanting the components to reproduce the origi-
nal bearing surfaces, as with Townley’s technique. The first step was to flex the knee
to 90 degrees; insert the rod of the first instrument, the tibial cutter, into the femoral
intramedullary canal; and then pull the tibia downwards so that the joint capsule was
tightened. The tibial cutter had a choice of horizontal slots through which to cut the
proximal tibia using an oscillating saw. The second instrument was the femoral
shaper which also had an intramedullary rod and which had slots for cutting the
anterior and posterior surfaces of the femur. The third instrument was the distal
femoral cutter that had a plate locating on the already-cut proximal tibial surface
and which was pulled down to tighten the soft tissue capsule with the knee at 0
degrees flexion. The slot in the femoral cutter was then used to make the distal
femoral cut.
It was, however, realized that when the capsule was tightened on the lateral and
medial sides, the frontal plane alignment of 5 degrees valgus between the femur and
tibia may not have been achieved: “In the more common varus deformity, the medial
ligament shortens and the lateral ligament stretches: unless corrected, this imbal-
ance will persist after replacement.” A few degrees of bone misalignment were
accepted, but for large varus deformities, the medial structures were partially
released from the tibia prior to making the femoral cuts. Valgus releases were not
recommended however. In the above procedure for making the bone cuts, in cases
of varus deformity, it usually occurred that the posterior femoral cut on the medial
condyle was a few millimeters more in thickness than the cut on the lateral condyle.
The anterior femoral cut would be parallel to the posterior cut. The result was an
external axial rotation of the femoral component on the distal end of the femur, of a
few degrees. While this rotation would reduce the steepness of the lateral face of the
patella groove at the anterior aspect of the femoral component, where the patella
articulates in early flexion, in practice this did not produce any serious
consequences.
The above two approaches to surgical technique, focusing either on anatomic
alignment of the femur to the tibia and the components to each bone or producing
equal soft tissue tensions and gaps in extension and flexion, have been the funda-
mental starting points for subsequent technique development. It was evident that
both alignment and soft tissue balancing were necessary which subsequently led to
a fusion of both concepts. This was emphasized by David Hungerford and Ken
Krackow (1984) at the first Scientific Meeting of the Knee Society: “In order for
non-constrained total knee systems to work properly, the retained ligaments must
function at normal lengths….To achieve nearly normal function each component
must be positioned not only in the correct varus-valgus, flexion-extension, and rota-
tional attitude, but also in the correct proximal-distal and anterior-posterior atti-
tude.” The mention of “non-constrained” has significance in that it was not
necessarily implying that the same rules applied to the more constrained cruciate
resecting total knee designs, which were popular at the time. In the area of align-
ment, techniques have included redefining the angles of the bone cuts as well as
attempting to improve accuracy. Hungerford and Krackow focused on the overall
alignment from full-leg standing x-rays to confirm what Townley had already
pointed out that the joint line in the frontal plane sloped medially by an average of
2–3 degrees, such that the joint line was parallel to the ground in a standing position.
Reproducing this at surgery was an important element in placement of the femoral
component on the distal femur, where the bone resected laterally and medially in
both flexion and extension was exactly the same as the thickness of the femoral
component. This was considered to be crucial because any deviation in component
Fig. 12.3 A mechanical

distractor where the gaps
on the lateral and medial
sides were measured, was
designed in the mid-1980s
by Chit Ranawat and Larry
Dorr as part of their
Precision knee instruments.
The aim was to achieve
accuracy at every step of
the procedure through
well-designed instruments
and steps in the technique.
from Springer Nature)
position would change the center of rotation of the femur producing an incorrect
ligament length pattern during flexion. Hence the instrumentation was designed to
key off the distal and posterior femoral condyles, but allowing for any cartilage
wear or bone collapse that had occurred.
There was one compromise that the distal femoral cut needed to be at the ana-
tomic angle to the femoral shaft, as indicated by both intramedullary and extramed-
ullary rods, and hence the distal cut was keyed off the most prominent condyle. This
frontal plane angle was individual to each patient, based on their long-leg radio-
graphs (Hungerford and Kenna 1983), which was an original idea at the time when
most knee systems recommended a standard angle. This aspect was addressed later
by Johan Bellemans et al. (2012), noted below. Having now positioned the femoral
and tibial components correctly, balancing was carried out in two steps. Firstly, to
achieve stability in both flexion and extension, either extra bone could be resected
from the distal femur, or a thicker tibial component could be used. Secondly, varus-
valgus instability was solved by releasing the tight side or tightening the loose side.
While the major goals of this technique were similar to those of Townley, the steps
were more systematic and the instruments more accurate. Attention was given to
avoiding an excess of varus or valgus, while the ligament balancing steps were
clearly spelled out. A limitation was that the balancing was carried out by “feel”
rather than with any instrument such as a tensor.
This deficiency was addressed by Chit Ranawat and Larry Dorr (1985), in their
“Precision Cut Instrumentation” which was applicable to mechanical alignment, an
advancement of the instruments used for the Total Condylar. By this time, several
posterior stabilized (PS) designs had emerged, for which mechanical alignment was
Fig. 12.4 Richard Bryan (left) from the Mayo Clinic and David Hungerford from Baltimore
appreciate the ancient globe at the Umberto Library in the Rizzoli Institute in Bologna, Italy.
The globe was made in 1762 by Father Rosini, based on maps and descriptions from numerous
explorers. Fabio Catani points out that Australia is not shown on the globe, a destination not
known at the time. The occasion was a conference on artificial knees, organized by Mario
Marchetti in the early 1980s. (© Peter S Walker)
also recommended: “These instruments allow for precision bone cutting and resto-
ration of the mechanical axis. The tensor allows balancing of the ligaments both in
flexion and extension”. The tensor was first used with the knee at 90 degrees flexion
after cutting the proximal tibia and the anterior and posterior femur. The lateral and
medial arms of the tensor were expanded, and if the gaps on the medial or lateral
side were unequal, the ligament on the tight side was released in small steps until
the gaps were equal. The knee was then extended with the femur at 5 degrees valgus
to the tibia and the tensor inserted. Distraction again determined any medial or lat-
eral tightness to be corrected, after which the level of the distal cut was marked on
the anterior femur. Hence, this technique achieved both the required alignments and
ligament balancing. The latter was achieved in both flexion-extension and in varus-
valgus by use of the tensor. This method of balancing was far more accurate than
using simple spacer blocks inserted between the cut surfaces, where the surgeon
assessed the relative tightness of the block in flexion and extension, and the varus
and valgus equality by applying side-to-side movements. Nevertheless, this particu-
lar tensor still relied on the surgeon assessing the tightness of the distractor wheel to
come to a “hard stop” when the ligaments were tight.
Different Alignment Goals and Techniques
The literature shows that the various alignments affect outcomes, although consid-
erably more research is needed to gain more insights. The issue is complicated
because of the wide range of conditions of the arthritic knee and the different
mechanics of the implant designs used for treatment. This means that there may not
be a “best” technique for all circumstances. The different techniques have been
thoroughly summarized in a recent article by Bill Jaffe MD from NYU Langone
Orthopedic Hospital (Jaffe et al. 2018). Following a Fellowship with John Charnley
in 1970, Bill Jaffe joined the surgical staff at NYU and maintained a strong interest
in biomechanics and biomaterials, working with scientists such as Jim Pugh PhD,
Harry Alexander PhD, and Fred Kummer PhD. The most common fundamental
tenet of surgical techniques today is to align the limb in some relation to the mechan-
ical axis: “The goal of mechanical alignment in TKA is to create a joint line that is
perpendicular to the mechanical axis of the limb, theoretically resulting in even
distribution of loads across the medial and lateral surfaces of the components.” This
differs from anatomic in that normally the joint line is at 2–3 degrees of varus to the
mechanical axis, such that the joint line is parallel to the ground in the standing
position. Making the bone cuts on the distal end of the femur and the proximal tibia,
using mechanical instruments, usually relies on an intramedullary rod in the femur
and an external alignment rod in front of the tibia. Most often, average, or close to
average, values are used for the cut angles. Cutting the tibia perpendicular to its
anatomic axis rather than at 2–3 degrees varus simplifies the surgery and avoids an
excess varus cut in case of an inaccuracy. However it is a compromise.
Jaffe then describes the different methods used for producing the bone cuts,
assuming a perpendicular tibial cut. In “measured resection” landmarks are used on
the distal femur to make the distal femoral cut and the anterior and posterior cuts.
The latter produce some internal rotation of the femoral component on the distal
femur, partly in compensation for the joint line being perpendicular to the mechani-
cal axis. A choice is made between anterior or posterior referencing, and while
posterior referencing is better for controlling the flexion angle and flexion tightness,
with multi-sized knee systems of today, it makes little difference.
Regarding the resection technique, the first was “measured resection” where the
anterior and posterior femoral cuts were made with reference to the line across the
posterior condyles (posterior condylar axis PCA), the line through the epicondyles
(epicondylar axis TEA), and the line through the base of the patella groove
(Whiteside’s line). In mechanical alignment, the anterior and posterior cuts are
rotated by about 3 degrees to compensate for the distal cut being perpendicular to
the mechanical axis. In the “gap balancing” resection technique, obtaining equal
lateral and medial soft tissue tensions is the guide for the resections.
Kinematic Alignment
Kinematic alignment requires special mention because of the increasing interest in

recent years (Howell 2019). Steve Howell MD is an orthopedic surgeon who has
collaborated on knee mechanics with Maury Hull PhD and others at the University
of California, Davis. The goal is to restore the joint lines in extension and flexion to
the normal anatomic condition. The technique has some similarity to that advocated
by Townley, and by Hungerford and Krackow, called anatomic alignment, but there
are some important differences. The method also has connections to the study by
20.0
All
17.5
15.0
12.5
Percent
10.0
7.5
5.0
2.5
0
–10 –9 –8 –7 –6 –5 –4 –3 –2 –1 0 1 2 3 4 5 6 7 8 9 10
HKA Angle (°)
Fig. 12.5 In a study of 250 young males and females (500 knees), Johann Bellemans and Jan
Victor from Ghent University, Belgium, measured the angle between the mechanical axis of the
femur and tibia (angle A in diagram by Moreland and Hanker). The mean value was about 1
degree. However there were 32% of males and 17% of females where there was a deviation greater
than 3 degrees. In positioning an artificial knee, it was questioned whether the best alignment
would take account of the angle in the particular patient or still be at the standard alignment per-
pendicular to the mechanical axis (Bellemans et al. 2012). (Reprinted with permission from
Clinical Orthopaedics & Related Research)
Bellemans et al. (2012) where it was shown that many knees have a constitutional
varus relative to the mechanical axis. The origin of kinematic alignment was the
laboratory study by Anne Hollister et al. (1993) depicting an average transverse axis
in the femur about which the tibia flexed and extended and a vertical axis through
the medial tibia about which the femur rotated internal-external. This concept can
also be connected with that of Freeman and Pinskerova (see Chap. 11). Steven
Howell MD added a third axis, transverse in the femur, about which the patella
flexes and extends (Howell 2019). These three axes are shown to be parallel and
perpendicular to the distal femoral and posterior femoral joint lines. At surgery,
distal femoral cuts are made, resecting exactly the thickness of the components,
allowing for any cartilage or bone loss. The same process is carried out for the pos-
terior femoral condyles. The tibia is cut, again restoring the line of the transverse
axis and the posterior slope. All cuts are measured and checked, adjusting if neces-
sary. It is noted that the anatomic axes in the femur and tibia, the mechanical axis,
as well as the epicondylar axis and Whiteside’s line, are not referred to. The rota-
tional positional of the tibial component is based on correct fitting of an anatomi-
cally shape component, but can also be referred to bony landmarks, the object being
to preserve the anatomic Q-angle. Having made the cuts, the balancing steps are
Lateral Femoral Condyle Medial Femoral Condyle
Fig. 12.6 Steve Howell MD from Methodist Hospital, Sacramento, and Maury Hull PhD, from
University of California Davis, developed a system of bone cuts called kinematic alignment. A
right femur (left) shows the relationships of the three kinematic axes of the knee with respect to the
joint lines of the distal and posterior femur and the position of the components. The transverse axis
in the femur about which the tibia flexes and extends is the green line. The transverse axis in the
femur about which the patella flexes and extends is the magenta line. The longitudinal axis about
which the tibia internally and externally rotates on the femur is the orange line. All three axes are
either parallel or perpendicular to the joint lines. Resecting bone, after correcting for wear, from
the distal and posterior femur equal in thickness to the condyles of the femoral component, kine-
matically aligns the femoral component. Resecting bone, after correcting for wear, from the tibia
equal in thickness to the tibial component, kinematically aligns the tibial component. The resulting
frontal plane alignment of the joint line is usually 2–3 degrees varus to the mechanical axis, as for
the anatomic knee, but varies with each individual (Howell et al. 2013). (Reproduced with permis-
sion from Knee Surgery Sports Traumatology Arthroscopy)
carried out based on a decision tree for corrective measures. For example, a knee
which was tight in extension and loose in flexion would be adjusted by a small addi-
tional resection from the distal femur. Tightness on the medial or lateral sides would
be corrected with a small wedge cut from the proximal tibia. This is consistent with
experimental and surgical data where most imbalances could be dealt with by cor-
rections of only 2 mm or 2 degrees (Walker et al. 2014). Verasense devices were
used at surgery to measure the actual compartmental forces. Anatomic values are
targeted, such as a looser lateral side than medial side in flexion. The relative medial
and lateral forces were measured as a function of overall alignment measured post-
operatively from long-leg radiographs (Shelton et al. 2017). The idea of the study
was that knees that had larger deviations from the mechanical axis, either in varus
or valgus, would show large inequalities in the contact forces at surgery. The “take-
home message” from the study was “Kinematic alignment has a high proportion of
varus and valgus outliers according to mechanical alignment criteria - Medial and
lateral tibial compartment forces of subjects with varus or valgus outlier alignment
were comparable with those with in-range alignment.” In terms of clinical results, a
10-year follow-up of 222 knees showed 98.4% survivorship, despite a high propor-
tion aligned outside of recommended limits (Howell et al. 2018). For example there
were 27% outside of +/− 3 degrees from the mechanical axis (mostly in varus) and
an almost uniform distribution of varus tibial tilt up to 6 degrees from horizontal.
Even so, a knee with a large deviation from the mechanical axis should be treated
with some caution according to Rivière et al. (2018) in which case there should be
a hybrid approach where that knee is aligned less than the deviation. This does have
a rationale for high varus deformities, where there would be substantial increase in
medial force (Halder et al. 2012), which in an artificial knee could cause excessive
wear or loosening.
Accuracy of Mechanical Instrumentation
Mechanical instrumentation has steadily evolved up to the present time, with more
precise alignment for the various angles, provision for modifying bone cuts, and
options for different levels of cut. There are two issues with any instrumentation
system: what is the accuracy which can be achieved, as measured by the deviations
from the target alignment values, and how much inaccuracy can be tolerated with-
out having an adverse effect on survivorship or function. The latter is a difficult
question because there is no compelling evidence as to what should be the correct
target, components perpendicular to the mechanical axis, or a more patient-specific
alignment.
The accuracy of the bone cuts was measured for a typical instrument system in the
1990s (Mahaluxmivala et al. 2001). While the mean values of the frontal plane angles
were close to target, the standard deviations for the femoral and tibial component
alignments were 2.2–3.9 degrees, the least value being for the tibial frontal plane.
Only 75% of overall alignment angles were within 3 degrees of target. The looseness
of the IM rod in the femoral canal was identified as large source of error. In a more
recent and often-quoted paper, 398 consecutive cases of 3 types of total knees were
carried out by a single surgeon at the Mayo Clinic in a 5-year time period (Parratte
et al. 2010). Full-length radiographs were used to measure the frontal plane angles.
Although the mean deviation from the mechanical axis was zero degrees, the standard
deviation was 2.8 degrees with a range from 8 degrees varus to 9 degrees valgus.
However the survivorship was no different between cases within 3 degrees and all
outliers of more than 3 degrees. The authors concluded nevertheless that “a neutral
mechanical axis remains a reasonable target….” In a recent revisit of the same cohort
of patients followed for a further 5 years, there were no changes in the conclusions
(Abdel et al. 2018). The effect of alignment was of great interest to Merrill Ritter MD
and colleagues in Mooresville, Indiana, using the AGC knee implant. By constantly
evaluating their results as a basis for yet further improvements in treatment modalities,
Ritter and a group of expert knee and hip surgeons made numerous contributions to
the advancement of the field. In their study of no less than 6070 knees, failures were
tested for a correlation with the femoral-tibial angles. Remarkably there were only 54
failures (0.89%) at an average of 5 years (Ritter et al. 2011). Their angles were mea-
sured on short radiographs. The mean femoral-tibial angle was 4.7 degrees of valgus,
with a standard deviation of 2.5 degrees. (The mean femoral-tibial angle in normal
knees is 5.7 degrees). The range however was from −12 degrees to +20 degrees. The
overall conclusion was “a surgeon should aim to place the tibial component at an
angle of 90 degrees from the tibial axis and the femoral component in <8.0 degrees of
valgus in order to yield an overall anatomic alignment (tibiofemoral coronal align-
ment) between 2.5 degrees and 7.4 degrees of valgus, which should be achievable
with conventional instruments.” This range is similar to that of the Mayo Clinic study,
although in Ritter’s study, the failures occurred at higher deviations.
Based mainly on observations of individual cases, many authors have suggested
that alignment errors can result in problems other than component loosening. These
problems have included patella mal-tracking, excessive polyethylene wear, instabil-
ity, and reduced flexion angle, as well as cosmetic effects. While mechanical instru-
mentation has been associated with high survivorship values, it may be that
inaccuracies have resulted in some patients not reaching their optimal functional
performance, although it has been difficult to reach definite conclusions when there
is such granularity.
Up till now the most common method for performing an artificial knee has been
the use of traditional mechanical instrumentation. It cannot be assumed that all sys-
tems are similar in their accuracy and precision or ease of use. However, overall, it
is recognized that there are certain inherent errors in the use of such instruments.
There are only a few papers to be found on the magnitude and causes of errors, so
the following list is compiled from observations made in papers, conference lec-
tures, and anecdotal discussions.
• The intramedullary rod for defining the frontal plane angle of the distal femoral
cut. In larger canals, the error can be around +/− 2 degrees. This can be even
larger if the entry point of the rod is not lined up with the center of the femoral
canal, which is about 6 mm medial to the center of the intercondylar notch.
• Angular deflection of the saw blade when passing through the slot due to clear-
ance or offset forces applied.
• Error in pinning the slotted cutting block to the bone.
• Distance of the cutting block from the starting point of the bone cut.
• Bending of the saw blade as it encounters hard bone more on one side of the
blade than the other.
• Error in placement of the distal clamp of a tibial cutting guide at the ankle.
• Lack of parallelism of the vertical rod of the tibial cutting guide to the anatomic
axis of the tibia, in the sagittal plane.
• Looseness in the modular parts of the instruments or even a misunderstanding of
the method of assembly or attachment to the bone.
• Loosening of the pins holding the slotted cutting guides, as the bone cut is being
made.
• Cutting blocks which have flat surfaces over which the blade is passed, allowing
for some tilting of the blade (although visualization may be better).
A number of these factors have been addressed in some systems, but in general,
there is still room for improvement. Hence it can be concluded that while there are
inherent errors in the use of mechanical instruments, the errors could likely be
reduced by improvements in the design and use of the cutting guides and saw
blades. The final word in this section is given by a study from the Rothman
Institute in Philadelphia (Küçükdurmaz et al. 2015). They attempted to match the
flexion and extension gaps. There were three groups of 25 patients, each group
having a different instrumentation system from a major company. Their results
were that only 44 percent of cases had all three bone cuts within 2 millimeters of
target. This is not ideal considering the sensitivity to ligaments to length changes
of only 1–2 mm.
Range of Flexion
One outcome measure, the maximum range of flexion achieved, has still not reached
anatomic levels. This could be partly explained by the fact that an artificial knee
does not sufficiently mimic the anatomic structures which it is replacing. In the
anatomic knee, flexion beyond about 120 degrees involves complex movements of
the femoral condyles on the tibial plateaus and displacement of the menisci, such
that impaction between the posterior femoral cortex and the posterior tibial plateau,
which would block flexion, is avoided (Pinskerova et al. 2009). These authors con-
cluded that “The anatomical and functional features of the arc from 120 to 160
degrees suggest that it would be difficult to design a total knee replacement giving
physiological movement.” Nevertheless there are design features of an artificial
knee which do have an effect.
Impingement of the posterior femoral cortex on the posterior edge of the tibial
insert was highlighted as a limitation to flexion in one study (Bellemans et al. 2005).
They focused on the posterior slope of the tibial cut as affecting the height of the
posterior edge of the tibial bearing surface and concluded that every 1 degree of
slope changed the flexion angle by 1.7 degrees. However simply cutting a large
slope angle beyond what is normal for the particular knee is not advisable because
the contact points between the femoral and tibial condyles displace posteriorly,
which can cause instability or extension lag (Cantin et al. 2015; Weinberg et al.
2017). Bellemans also showed how the posterior offset of the femoral condyles rela-
tive to the posterior cortex was another factor in determining when impingement
occurred, with a high offset giving the most flexion. A further possibility in limiting
flexion is abnormal tensions in the soft tissues around the implant components, in
regions where a component projects beyond the boundaries of the original femur or
tibia. Around the anterior patella flange is one area where this might occur, but there
is little literature on this aspect. Nevertheless if the preoperative flexion angle is

already limited due to changes in the soft tissue properties around the knee, it may
not be possible to regain a high flexion angle after the artificial knee.
Patient-Specific Instrumentation (PSI)
Due to the complexity and number of the mechanical instruments and the attention
needed for cleaning, sterilizing, and storing, it is not surprising that alternate meth-
ods for making the bone cuts were explored. One method for simplifying the system
was to use imaging to design and manufacture patient-specific cutting guides for
making the primary cuts of the distal femur and proximal tibia (Lombardi & Frye
2012). The idea was to generate computer models of the femur and tibia using MRI
or CT scans. From those models, the bone cuts were defined to produce the required
alignments in the frontal and sagittal planes. Cutting guides were then designed to
fit precisely on to the distal femur and proximal tibia, matching the surfaces of the
patient’s bones. To manufacture such guides with their complex geometry was a
challenge, the answer being 3D printing using rigid polymeric materials. At surgery,
the cutting guides were positioned in their unique positions on the bones and pinned
in place. This greatly simplified the surgical procedure. The method became known
as patient-specific instrumentation (PSI). As pointed out by Lombardi and Frye, an
important advantage is the ability to customize the bone cuts for each individual
patient.
The first PSI system, the Visionaire, was approved by the FDA in 2008. Since
then seven other systems have been designed and introduced. However, overall, the
P P
P
P Femoral Tibial
guide guide
Fig. 12.7 Patient-specific instrumentation (PSI) made by 3D printing for each individual patient.
CT or MRI scans are taken of the hip, knee, and ankle and reference axes determined, together with
the cutting planes of the femur and tibia. The guides are designed to be a precise fit on the distal
femur and proximal tibia (best seen on the left picture). After pinning the guides (P), holes are
drilled into the bone through the other two guide holes which define the position of a standard slot-
ted cutting guide. The number of outliers, cuts more than 3 degrees from target, was approximately
25%, at least as good as that of conventional instruments. (Ng et al. 2012). (Reproduced with
permission from Clinical Orthopaedics & Related Research)
method has not shown any striking advantages over the traditional mechanical
instrumentation, based on a number of studies. The verdict from one author Bill
Mihalko (2015) was “but until more definitive functional and patient-perceived out-
come data, along with cost effectiveness data, are obtained, we do not have adequate
evidence to recommend for or against the use of PSI for total knee arthroplasty.”
Subsequent publications provided further information. Based on a meta-analysis of
44 studies comparing almost 3000 each of PSI cases and mechanical cases
(Thienpont et al. 2017), the conclusion was “The present meta-analysis revealed an
increased risk of malalignment of the tibial component for PSI, and its impact on
implant longevity remains to be seen. This study showed that PSI only has a minor
impact on the risk of malalignment of the mechanical axis. Moreover, there is still a
lack of evidence of clinical effect associated with use of PSI.” A multicenter study
found no differences between PSI and mechanical instrumentation for alignment,
operative time, blood loss, and Knee Society Function Score (Abane et al. 2018). In
a recent meta-analysis where 38 studies were analyzed including almost 350 cases,
it was concluded that PSI does not improve patient-reported outcome measures,
surgery time, and complication rates as compared to standard techniques. While
there is much overlap in the above studies, the conclusions are the same. A caution-
ary note is that all systems do not necessarily produce the same results, some sys-
tems may be measurably superior to others in certain ways, but this is difficult to
discern from the literature.
Hence while there were no striking advantages cited, there were no major disad-
vantages in using PSI in terms of alignment accuracy and functional performance.
In that case it might be thought that the simplicity of the surgical procedure would
be a positive factor. The drawback however was that the patient had to be scheduled
for imaging a few weeks prior to surgery, and the surgeon had to study the design of
cutting guides proposed by the manufacturer, all adding time and cost. A further
point is that many surgeons were somewhat nervous about relying on the fit of the
cutting guides to the distal femur and proximal tibia to produce the correct align-
ments, without having a secondary check. It is unlikely that PSI will be abandoned,
because of its elegance, and it represents the point of view that while mechanical
instruments may produce similar accuracy, simplification would be an advantage.
Possibly modifications of PSI may be put forward to address the disadvantages.
Computer-Assisted Instrumentation
In the late 1990s, efforts were underway to develop a computer-assisted surgical

technique which could improve both accuracy and precision of the bone cuts and the
alignment. Electronic digitizers were already widely used in engineering whereby
the coordinates of a point on an object could be determined by touching the point
with a probe. This technology was applied to a surgical navigation system where
various reference points on the femur and tibia were exposed and the coordinates
determined with reference to a global axis system. The positions of the femur and
tibia in space were also determined in this axis system by fixing trackers into the
middle of the diaphyses of the femur and tibia which were optically tracked. All
points were then tracked in real time even when the femur and tibia moved. The
location of the center of the femoral head, impossible to access directly but neces-
sary for defining the mechanical axis of the leg, was determined by an ingenious
method. The femur was rotated in different directions about the hip joint, while
points on the femur were continuously tracked. The coordinates of the center of the
hip could then be calculated.
Using such a system, it was immediately evident that the alignments of the bone
cuts were more accurate and outliers were greatly reduced (Jenny and Boeri 2001).
Since the first system, several others were developed and used extensively around the
world. A recent meta-analysis from 21 studies comparing 844 knees performed by
mechanical instrumentation with 869 knees using navigation (Clayton et al. 2014)
showed that “These results demonstrate that the use of computer navigation in TKA
achieves both superior mechanical alignment and an increase in functional scores.
This increase requires the investment of additional time and financial resources.”
While navigation still has a dedicated following today, the increased surgical
time has been the main reason why the method has not become widely popular,
coupled with some skepticism as to whether the more accurate alignment makes a
significant difference in longevity and function. A positive step in recent years has
been the refinement to navigation systems in measuring the varus and valgus laxities
as the knee is put through a range of flexion. This enables the surgeon to balance the
knee in a more quantitative way than before. However if the laxities are still mea-
sured by the surgeon applying the forces by hand, there is still a subjective element
to the readings obtained.
It was inevitable that robotics would enter the arena in total knee surgery (Bargar
2007). This was initiated by an orthopedic surgeon working with a veterinary surgeon.
The orthopedist, Bill Barger, had a background in mechanical engineering before
entering the medical field. The veterinarian, Howard “Hap” Paul, was a creative and
adventurous individual who is still recognized today by the HAP Paul Award for inno-
vation, given by the International Society for Technology in Arthroplasty (ISTA) at its
annual meetings. Based on initial funding from IBM, Bargar and Paul were able to
perform their first robot hip surgery on a dog in 1990. The name they gave to their
equipment was Robodoc. Part of the driving force for choosing the hip was the enthu-
siastic movement between many surgeons around the world for uncemented custom-
ized artificial hip joints, for which a perfect fit between the implant and the canal of
the bone was required. The milling tool on the robot arm could prepare such a cavity.
During the 1990s, clinical hip trials proceeded on humans, but then in 2000, the first
total knee surgery was performed, the company making the equipment being called
Think Surgical. Concurrent with this development, Professor Brian Davies at Imperial
College, London, began development of a medical robot in 1991 called the Probot, for
prostate surgery. By 1999 Davies had developed an active constraint system called the
Acrobot for minimally invasive orthopaedic applications including total knees. After
a number of iterations of the robot, the Acrobot company was acquired by Stryker
where patent problems prevented its further application. Brian Davies is still active in
the field of medical robotics.
Fig. 12.8 Brian Davies

PhD, from the Mechanical
Engineering Department at
Imperial College London,
developed the first active
constrained robot systems
for performing artificial
knee procedures, first
performed on patients in
2002. This work set the
stage for other robotic
developments which are
gradually gaining
recognition for their
capabilities in accuracy in
effecting the optimal
surgical plan. (Photo taken
in London c 2015). (©
Peter S Walker)
There are currently several robotic systems available, described in a review by

Jacofsky and Allen (2016). Some systems were developed by enterprising indi-
viduals, either in a university setting or in a startup company. For example, Rony
Abovitz, a visionary, connected with Maurice Ferre MD in 2004 to form MAKO
Surgical, to develop the first haptic robot system, applied initially to unicompart-
mental knees. The company has since been acquired by Stryker which is concen-
trating primarily on the application of the robot for total knee replacement. Rony is
presently involved with virtual reality but not in the world of orthopedics. Dr. Tony
DiGioia, a surgeon in Pittsburgh, collaborated with faculty members in robotics at
Carnegie Mellon University, to develop a handheld robotic tool called Navio for
performing a minimally invasive knee replacement procedure. As with Bill Bargar,
DiGioia held an engineering degree before entering the medical field. A company
was formed in 2013, Blue Belt Technologies, which was acquired by Smith &
Nephew in 2015. The Navio is a handheld tool with cutting implements such as
burrs and is used in conjunction with a navigation system to shape the bone without
Fig. 12.9 An artificial knee procedure using the Mako robot (Stryker). Left, the surgeon has
planned the procedure by placing the femoral and tibial components on the images of the bones,
shown on the computer screen. The key reference points are being digitized. Right, the robot arm
with an oscillating saw mounted on the end, is being deployed to resect the posterior femoral con-
dyles. The progress of the cuts is shown on the computer screen. (Photos courtesy of Patrick Meere
MD, NYU Langone Orthopedic Hospital, New York)
Fig. 12.10 A device for

distracting the lateral and
medial condyles separately
at surgery as a guide to the
balancing required. This is
part of the Omnibot system
for carrying out an
artificial knee surgery.
Designed by Christopher
Plaskos. (Photo taken at
CAOS International
Conference, New York,
April 2019)
needing any adjunct cutting guides. Chris Plaskos PhD developed a small bone-
mounted robot which moved a cutting guide into prescribed positions, without
requiring any prior imaging. Chris was a PhD in the area of medical robotics and
co-founded OMNILife Science in 1999 to further develop this invention. A new
addition to the system has been a distractor system for balancing ligaments. The
company was acquired in 2019 by Corin Group in England. Zimmer entered the
computer-assisted surgery area in 2007 with the acquisition of ORTHOsoft Inc.,
based in Montreal. The concentration has been on smart tools such as the iAssist.
Zimmer recently acquired the ROSA robotic arm system from French-based
medtech. Rosa was originally developed for brain and spine surgery, but within a
short period of time, Zimmer developed software for artificial knee surgery. The
accuracy of the system was tested in knee specimens by eight surgeons on different
knee designs (Parratte et al. 2019). The accuracy was within about 1 degree and
1 mm. DePuy Synthes (J&J) recently acquired a Paris-based robot company
Orthotaxy. This is a compact robot, at a relatively low price, which will be easy to
operate and speed up the surgery. It is currently being developed for knee surgery.
Hence robotics for total knee surgery is an active and expanding field, currently
being evaluated in many centers. Jacofsky and Allen (2016) summarized the current
status as follows: “Regardless of what target a surgeon may desire for limb align-
ment or component position, it is clear that robotics will enable surgeons to carry
out more precise and accurate procedures on a more consistent basis with a more
patient-specific plan. In the future, robotics is moving toward becoming a valuable
adjunct to the surgeon in optimizing patient-specific arthroplasty.” The authors also
point out that there is a very high price tag for a hospital to enter this field, which
limits the level of acceptance at this time.
Another area of computer-assisted surgery which has gained support is the use of
“smart tools” incorporating inertial measurement units (IMUs) and other sensors.
While not receiving the same amount of attention as robotics, such devices offer the
advantages of simplicity, utility, and low cost, without adding time to the procedure.
They are intended to be used seamlessly in a system of mechanical instruments. The
smart tools fit in the palm of the hand and are designed to place cutting guides in a
location on the femur and tibia to make precise cuts relative to the reference axes
such as the mechanical axis. The tools are initially registered to bony landmarks and
then the cutting guides are pinned in place. Two of the most widely used devices
have received positive appraisals. In a follow-up of the KneeAlign system, Nam
et al. (2012) concluded: “This study demonstrates that accelerometer-based naviga-
tion can obtain the same degree of accuracy as previously reported with large con-
sole CAS systems, while avoiding the use of additional pin sites, arrays, or large
console platforms.” For the iAssist, Kinney et al. (2018) reached a similar conclu-
sion: “we have demonstrated that the use of a handheld guidance system in TKA
results in decreased mechanical axis variability and a decreased rate of unacceptable
postoperative alignment compared to conventional instruments, without resulting in
significant increases in operative time or blood loss.”
On the subject of computer-assisted surgery, brief mention will be made of
computer-assisted training of the surgical technique. In 2003, Noble and Conditt
devised a method whereby the surgeon was provided with a computer model of the
arthritic knee and then positioned artificial components on to the bones in the cor-
rect alignments (Noble and Conditt 2008). The surgeon then performed the proce-
dure physically on sawbones and attempted to implement the computer model. A
scoring system was devised for how accurately the dimensional parameters were
achieved. A similar method was carried further (Gunther et al. 2002) in that the
entire set of instruments and implants was modelled in the computer. The surgeon
performed the procedure in the computer, learning the sequential steps. The alter-
native was to study the manufacturer’s manual. Surgeons in the computer group
and the manual group then performed the procedure on saw bones. The computer-
trained surgeons performed in a shorter time and with fewer errors. This type of
approach has potential and is clearly more convenient and less expensive than
Computer models Surgical Patient, Cadaver

Ct Data of Implants and Instruments or Anatomic
Instruments and Implants Model
Step 1 Step 3
Operator Generates Computer Surgeon Performs the
Models to Simulate Surgery Operative Procedure
Step 4
Step 2 System Calculates Values
Surgeon sets Target of the Technical
Values for the Parameters for Operative
Technical Parameters Procedure
Step 5
Compare Target and Actual
Values of Technical Parameters
Fig. 12.11 A method devised by Philip Noble and Michael Conditt while at Baylor College of
Medicine and Methodist Hospital, Houston. The figure shows the strategy for evaluating a total
knee surgical procedure for accuracy, comparing the end result to the target values. The method
was demonstrated to be applicable to actual surgical cases or in a training environment where
surgeons work with sawbones or with leg specimens. (US patent 7,427,200 B2, filed 2003, issued
2008)
using cadaveric specimens. At the same time, there is the limitation that a computer
simulation is unlikely, in the short term at least, to simulate the anatomy and prop-
erties of the actual tissues of the knee, the methods being most applicable to the
bones themselves. Recently, the American Academy of Orthopaedic Surgeons has
initiated a study of virtual reality as a modality for surgical training. This possibil-
ity is a subject of extensive research at Imperial College, London.
Soft Tissue Balancing
One facet of surgical technique that deserves special attention is soft tissue balanc-
ing, also called ligament balancing, or simply “balancing”. Balancing is carried out
at surgery, where loads which occur during function cannot be safely applied. Hence
balancing needs to be carried out with the knee in a passive condition. Even then,
because of the weights of the thigh and shank, there are forces and moments between
the femur and tibia. In this passive condition, it was discovered in experiments at the
NYU Langone Orthopedic Hospital that there are forces acting across the lateral
and medial condyles due to “pre-tensions” in the soft tissues (Verstraete et al. 2017).
The forces were highest in extension, averaging 114 Newtons (25 pounds) on the
medial side and 63 Newtons on the lateral side. The forces reduced quickly in early
flexion and then more gradually for the remaining flexion range. The purpose of the
soft tissue tensions, it can be proposed, is to maintain the femoral and tibial con-
dyles in a defined relative position during a passive range of flexion. If the tensions
are too high, it would cause stiffness; if too low, there would be excessive laxity and
incorrect positions. When an artificial knee is placed, the objective is to place the
components on the bones so that there are normal soft tissue tensions throughout the
flexion range, to maintain the correct femoral-tibial positions. For simplicity it can
be proposed that the tensions should be equal at about 100 Newtons per condyle,
throughout flexion. This is not easy to accomplish because when a collateral liga-
ment is just tense, every millimeter of extra length causes about 50 Newtons of
force. Hence to balance the knee requires that the femoral shape and size are close
to the normal anatomy of that knee and the tibial thickness is correct. Imbalances
need to be identified at surgery and corrected by various means such as osteophyte
removal, soft tissue releases, or modifications to bone cuts (Meere et al. 2016).
While the methods of the early 1970s using simple spacer blocks served well at
that time, they may not have achieved the most consistent results. One of the indi-
viduals who recognized the complexity of balancing was Leo Whiteside MD, who
founded a bioengineering unit near his hospital in St. Louis, Whiteside Biomechanics.
Steve White PhD, Darryl Haynes PhD, and other biomechanical engineers carried
out pioneering and influential studies at this unit. In 1987 the group published a
study (Whiteside et al. 1987) showing the importance of balancing the collateral
ligaments correctly to limit varus-varus laxity, but with the effect of also controlling
rotational laxity, a phenomenon which had also been noted by John O’Connor from
Oxford in relation to unicondylar knees (Goodfellow and O’Connor 1978).
Whiteside also explained the importance of placing a femoral component precisely
matching the condyles so that the collateral ligament attachment points were cor-
rectly lined up with the sagittal contours, otherwise the ligaments will be too tight
or too loose in a part of the flexion range. This must be considered in relation to
techniques where the femoral component is positioned in relation to the mechanical
axis or some other landmarks. Whiteside (2005) emphasized the multiple soft tissue
structures across the knee, many with broad attachments, such that different fiber
bundles would become loose and tense at different flexion angles. However, inser-
tion of artificial knee components would often lead to abnormalities in these tension
patterns. This led to his concept of balancing which was to selectively release the
tightest parts of the structures (Whiteside et al. 2000 ch.26): “...release only the liga-
ments that are excessively tight, leaving those that are performing normally.” This
approach to balancing, where the tightness of small groups of fibers at different
flexion angles is accounted for, has recently been supported by Ranawat after many
years of experience with different techniques. Nevertheless, quantifying the balanc-
ing in some way during the procedure was an unsolved problem, except for the
introduction of distractor devices which could not readily be used with trial compo-
nents in place.
A different way of thinking about balancing was proposed by Asano et al. (2004).
Rather than focusing on gaps or on selective releasing, they hypothesized that the
important measure was soft tissue tensions. They devised a special instrument which
could measure tensions, the surgical goal being to achieve equal tensions on the lat-
eral and medial sides, as well as in flexion and extension. The tensions in the soft
tissues transferred directly to the condylar contacts between the femoral and tibial
components. This concept was translated into electronic instrumentation developed
recently, an instrumented trial tibial component, called the Verasense. The device
was conceived by Martin Roche MD, an orthopedic surgeon from Holy Cross
Hospital in Fort Lauderdale, and Marc Stein PhD, an electronics engineer. A com-
pany was formed by Jay Pierce, and by 2010, clinical application was initiated.
Rather than measuring ligament tensions or distraction gaps, the forces on the cut
surfaces of the lateral and medial condyles were measured by an instrumented tibial
trial, the forces being generated by the tensions in the soft tissues (Gustke et al.
Fig. 12.12 In a normal intact knee, there are forces in the collateral ligaments of about 100 N
(22.5 lbf) which apply these forces to the lateral and medial condyles (left figure). The forces
reduce with flexion, and the medial force is higher than the lateral. At surgery the surgeon can
insert an Orthosensor instrumented tibial trial which measures the condyle forces during flexion-
extension (displayed on computer screen). If the knee is unbalanced, the surgeon can make small
angular changes to the bone cuts or release tight regions of ligaments
2017). Hence by studying the lateral and medial force patterns as the knee was flexed
and extended, the surgeon could determine corrective bone cuts, changes in compo-
nent thicknesses, or soft tissue releases, in order to obtain the correct balancing. The
sensitivity of the balancing process was determined in experiments by Walker et al.
(2014) where only 2 mm or 2 degree corrections were usually the maximum needed.
This was confirmed in a series of 101 surgical cases (Meere et al. 2016). This also
reinforced the importance of accurate condylar shapes of the artificial knee compo-
nents as well as accurate placement (Zapata et al. 2019). Such small corrections
could be carried out by a “systematic selective soft tissue release” technique, although
in some cases, bony cut corrections could be made also (Peters et al. 2013).
There are now several technique options for total knee surgery. In the majority of
cases, traditional mechanical instruments are used. The main reasons are availabil-
ity, familiarity, reliability, and acceptable results. However the disadvantages
include high initial cost (usually borne by the suppliers of the artificial knees), stor-
age and cleaning requirements, and lack of sufficient accuracy and precision such
that surgery to close tolerances is not possible at present. An interesting recent alter-
nate for mechanical instruments is to make the instruments disposable using 3D
printing (Hooper et al. 2019). While disposable plastics have a negative environ-
mental impact, there are certain advantages such as elegance of pre-assembled
designs and ease of inventory handling. For accuracy and optimization at surgery,
robotic systems have a clear advantage. Adoption is so far limited due to high up-
front cost, the need for a high level of technical skill, even including a technician in
the operating environment, and so far, increased operating time. In the upcoming
years, studies on costs, outcome measures, and other aspects, are needed to provide
answers to the most appropriate ways forward.
Are Mechanical Instruments Accurate Enough?
In terms of survivorship, where the components do not loosen, the Mayo Clinic and
the Mooresville studies concluded that deviations from target limb alignments of at
least up to 3 degrees were well tolerated. It is possible that cement technique, design
of the implant, and condition of the tibial trabecular bone have more effect than
alignment. But assuming that 3 degrees was the maximum error allowed, mechani-
cal instruments are borderline for producing this accuracy routinely. Hence some
improvement in the design of the instruments is needed. An improved intramedul-
lary guide, and the smart tools navigation devices, may fill this need. In terms of
function, a limitation is that the target alignment and component positions to achieve
the optimal results in each patient, and the effect of deviation from the target, are not
known with any certainty. However for achieving a target value, computer-assisted
methods, such as navigation and robotic surgery, are more accurate. They also offer
the opportunity for testing component placements and measurement of ligament
laxity at surgery. Ongoing clinical studies will determine whether functional out-
come measures can be elevated by the use of the newer computer-assisted
techniques.
Fig. 12.13 Robert Booth

from Jefferson Health,
Philadelphia, has
performed more total knee
surgeries than any other
surgeon. His surgical
expertise and efficiency is
widely known from his
numerous lectures at
conferences around the
world. He is also an
authority on American
cultural history. (Photo at
Knee Society 2018 © Peter
Walker.)
References
Abane L, Zaoui A, Anract P, Lefevre N, Herman S, Hamadouche M. Can a single-use and patient-
specific instrumentation be reliably used in primary total knee arthroplasty? A multicenter con-
trolled study. J Arthroplast. 2018;33(7):2111–8.
Abdel MP, Ollivier M, Parratte S, Trousdale RT, Berry DJ, Pagnano MW, et al. J Bone Joint Surg
Am. 2018;100(6):472–8.
Asano H, Hoshino A, Wilton TJ. Soft-tissue tension total knee arthroplasty. J Arthroplast.
2004;19(5):558–61.
Bargar W. Robots in orthopaedic surgery: past, present, and future. Clin Orthop Rel Res.
2007;463:31–6.
Bellemans J, Colyn W, Vandenneucker H, Victor J. The Chitranjan Ranawat award: is neutral
mechanical alignment normal for all patients? The concept of constitutional varus. Clin Orthop
Relat Res. 2012;470(1):45–53.
Bellemans J, Robijns F, Duerinckx J, Banks S, Vandenneucker H. The influence of tibial slope
on maximal flexion after total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc.
2005;13(3):193–6.
Cantin O, Magnussen RA, Corbi F, Servien E, Neyret P, Lustig S. The role of high tibial osteotomy
in the treatment of knee laxity: a comprehensive review. Knee Surg Sports Traumatol Arthrosc.
2015;23(10):3026–37.
Clayton AW, Cherian JJ, Banerjee S, et al. Does the use of navigation in total knee arthroplasty
affect outcomes? J Knee Surg. 2014;27(03):171–6.
Coventry MB. Osteotomy of the upper portion of the tibia for degenerative arthritis of the knee. J
Bone Joint Surg Am. 1965;47:984–90.
Coventry MB, Bryan RS. Technical pitfalls in the polycentric and geometric total knee arthro-
plasties. Conference on Total knee replacement. London: Institution of Mechanical Engineers;
1974. p. 16–8.
Freeman MAR, Swanson SAV, Todd RC. Replacement of the knee with the Freeman-Swanson
prosthesis. Conference on Total knee replacement. London: Institution of Mechanical
Engineers; 1974. p. 16–8.
Goodfellow J, O'Connor J. The mechanics of the knee and prosthesis design. J Bone Joint Surg.
1978;60-b(3):358–69.
Gunther SB, Soto GE, Colman WW. Interactive computer simulations of knee-replacement sur-
gery. Acad Med. 2002;77(7):753–4.
Gustke KA, Golladay GJ, Roche MW, Elson LC, Anderson CR. A targeted approach to ligament
balancing using kinetic sensors. J Arthroplast. 2017;32(7):2127–32.
Halder A, Kutzner I, Graichen F, Heinlein B, Beier A, Bergmann G. Influence of limb alignment on
mediolateral loading in total knee replacement: in vivo measurements in five patients. J Bone
Joint Surg Am. 2012;94(11):1023–9.
Hollister AM, Jatana S, Singh AK, Sullivan WW, Lupichuk AG. The axes of rotation of the knee.
Hooper J, Schwarzkopf R, Fernandez E, et al. Feasibility of single-use 3D-printed instruments for
total knee arthroplasty. Bone Joint J. 2019;101(7_Supple_C):115–20.
Howell SM. Calipered kinematically aligned total knee arthroplasty: an accurate technique that
improves patient outcomes and implant survival. Orthopaedics. 2019;42(3):126–35.
Howell SM, Papadopoulos S, Kyle T, Kuznik KT, Maury L, Hull ML. Accurate alignment and
high function after kinematically aligned TKA performed with generic instruments. Knee Surg
Sports Traumatol Arthrosc. 2013;21:2271–80.
Howell SM, Shelton TJ, Hull ML. Implant survival and function ten years after kinematically
aligned total knee arthroplasty. J Arthroplast. 2018;33(12):3678–84.
Hungerford DS, Kenna RV. Preliminary experience with a total knee prosthesis with porous coat-
ing used without cement. Clin Orthop Relat Res. 1983;176:95–107.
Hungerford DS, Krackow KA. In: Dorr LD, editor. Porous coated knee replacement. Papers of the
First Scientific Meeting of the Knee Society. Baltimore, MD: Publ University Park Press; 1984.
Insall J, Ranawat CS, Scott WN, Walker PS. Total condylar knee replacement: preliminary report.
Küçükdurmaz F, Levicoff EA, Good RP, Sharkey PF. Do standard surgical guides produce accu-
rate and precise femoral bone resections during total knee arthroplasty? Surg Technol Int.
2015;27:225–32.
Jacofsky DJ, Allen M. Robotics in arthroplasty: a comprehensive review. J Arthroplast.
2016;31(10):2353–63.
Jaffe WL, Dundon JM, Camus T. Alignment and balance methods in total knee arthroplasty. J Am
Acad Orthop Surg. 2018;26(20):709–16.
Jenny JY, Boeri C. Computer-assisted implantation of total knee prostheses: a case-control com-
parative study with classical instrumentation. Comput Aided Surg. 2001;6(4):217–20.
Kinney MC, Cidambi KR, Severns DL, Gonzales FB. Comparison of the iAssist handheld guid-
ance system to conventional instruments for mechanical Axis restoration in total knee arthro-
plasty. J Arthroplast. 2018;33(1):61–6.
Kutzner I, Heinlein B, Graichen F, Bender A, Rohlmann HA, Beier A, Bergmann G. Loading of
the knee joint during activities of daily living measured in vivo in five subjects. J Biomech.
2010;43:2164–73.
Lombardi AV, Frye BM. Customization of cutting blocks: can this address the problem? Curr Rev
Musculoskelet Med. 2012;5(4):309–14.
Mahaluxmivala J, Bankes M, Nicolai P, Aldam C, Allen P. The effect of surgeon experience on
component positioning in 673 Press Fit Condylar posterior cruciate-sacrificing total knee
arthroplasties. J Arthroplast. 2001;16(5):635–40.
Meere PA, Schneider SM, Walker PS. Accuracy of balancing at total knee surgery using an instru-
mented tibial trial. J Arthroplast. 2016;31(9):1938–42.
Mihalko WM. Patient-specific cutting guides were not better than conventional instrumentation for
total knee arthroplasty. J Bone Joint Surg Am. 2015;97(22):1891.
Moreland JR, Bassett LW, Hanker GJ. Radiographic analysis of the axial alignment of the lower
extremity. J Bone Joint Surg Am. 1987;69(5):745–9.
Morrison JB. The mechanics of the knee joint in relation to normal walking. J Biomech.
1970;3(1):51–61.
Nam D, Nawabi DH, Cross MB, Heyse TJ, Mayman DJ. Accelerometer-based computer nav-
igation for performing the distal femoral resection in total knee arthroplasty. J Arthroplast.
2012;27(9):1717–22.
Ng VY, DeClaire JH, Berend KR, Gulick BC, Lombardi AV Jr. Improved accuracy of alignment
with patient-specific positioning guides compared with manual instrumentation in TKA. Clin
Orthop Relat Res. 2012;470(1):99–107.
Noble PC & Conditt M. Computer-based training methods for surgical procedures. US Patent no.
7,427,200B2. Filed 2003, issues 2008.
Parratte S, Pagnano MW, Trousdale RT, Berry DJ. Effect of postoperative mechanical axis align-
ment on the fifteen-year survival of modern, cemented total knee replacements. J Bone Joint
Surg Am. 2010;92:2143–9.
Parratte S, Price AJ, Jeys LM, Jackson WF, Clarke HD. Accuracy of a new robotically assisted
technique for total knee arthroplasty: a cadaveric study. J Arthroplast. 2019;34(11):2799–803.
Peters CL, Jimenez C, Erickson J, Anderson MB, Pelt CE. Lessons learned from selective
soft-tissue release for gap balancing in primary total knee arthroplasty: an analysis of 1216
consecutive total knee arthroplasties: AAOS exhibit selection. J Bone Joint Surg (Am Vol).
2013;95(20):e152.
Pinskerova V, Samuelson KM, Stammers J, Maruthainar K, Sosna A, Freeman MA. The knee in
full flexion: an anatomical study. J Bone Joint Surg. 2009;91(6):830–4.
Pritchett JW. Charles O Townley: Design by Reason. 1988. www.seattlejointsurgeon.com.
Ranawat CS, Dorr LD. Technique of Total knee arthroplasty with precision cut instruments. In:
Total-condylar knee arthroplasty. Berlin: Springer; 1985. p. 69–83.
Ritter MA, Davis KE, Meding JB, Pierson JL, Berend ME, Malinzak RA. The effect of alignment
and BMI on failure of total knee replacement. J Bone Joint Surg Am. 2011;93(17):1588–96.
Rivière C, Lazic S, Boughton O, Wiart Y, Vïllet L, Cobb J. Current concepts for aligning knee
implants: patient-specific or systematic? EFORT Open Rev. 2018;3(1):1–6.
Shelton TJ, Nedopil AJ, Howell SM, Hull ML. Do varus or valgus outliers have higher forces in
the medial or lateral compartments than those which are in-range after a kinematically aligned
total knee arthroplasty? Bone Joint J. 2017;99-B:1319–28.
Steindler A. Kinesiology of the human body under normal and pathological conditions. Springfield:
Publ Charles C Thomas; 1955.
Thienpont E, Schwab PE, Fennema P. Efficacy of patient-specific instruments in total knee arthro-
plasty: a systematic review and meta-analysis. J Bone Joint Surg Am. 2017;99(6):521–30.
Townley CO. Total knee arthroplasty. A personal retrospective and prospective review. Clin Orthop
Relat Res. 1988;236:8–22.
Verstraete MA, Meere PA, Salvadore G, Victor J, Walker PS. Contact forces in the tibiofemoral
joint from soft tissue tensions: implications to soft tissue balancing in total knee arthroplasty.
J Biomech. 2017;58:195–202.
Walker PS, Meere PA, Bell CP. Effects of surgical variables in balancing of total knee replace-
ments using an instrumented tibial trial. Knee. 2014;21(1):156–61.
Weinberg DS, Gebhart JJ, Wera GD. An Anatomic Investigation Into the Relationship Between
Posterior Condylar Offset and Posterior Tibial Slope of One Thousand One Hundred Thirty-
Eight Cadaveric Knees. J Arthroplasty. 2017;32(5):1659–1664.e1651.
Whiteside LA, Kasselt MR, Haynes DW. Varus-valgus and rotational stability in rotationally
unconstrained total knee arthroplasty. Clin Orthop Relat Res. 1987;219:147–57.
Whiteside L. Assess and release the tight ligament. In: Total knee arthroplasty: Springer; 2005.
p. 170–6.
Whiteside LA, Saeki K, Mihalko WM. Functional medical ligament balancing in total knee arthro-
plasty. Clin Orthop Relat Res. 2000;380:45–57.
Zapata G, Sanz-Pena I, Verstraete M, Walker PS. Effects of femoral component placement on the
balancing of a total knee at surgery. J Biomech. 2019;86:117–24.
Testing and Tribology
How Were Designs Tested for Wear,
13
Strength, and Kinematics?
Any new mechanical device needs to be thoroughly tested before it is put into gen-
eral use, especially if there would be serious consequences in the event of a failure.
Testing requires standard test methods which have been validated as representing
service conditions. Yet, in the 1960s when total hips had been used in patients for
only a few years, and artificial knees were just being introduced, standard tests for
artificial joints did not exist, even though the American Society for Testing and
Materials (ASTM) F4 Committee for Medical Devices was initiated in 1962. That
is not to say that tests were not carried out or were considered unnecessary. Some of
the early designs were indeed subjected to tests devised by their inventors, but many
other designs were introduced without any testing at all. The following describes the
rationale behind some of the earliest tests and how the test methods gradually
became more sophisticated and expanded in scope.
Tribology (Friction, Lubrication, and Wear)
The methodologies for testing the friction and wear in artificial hips and knees origi-
nated from studies of the mechanics of the anatomic joint, such as the sliding behav-
ior between the cartilage bearing surfaces. Researchers such as Charles McCutcheon
PhD (1962) built machines where specimens of cartilage were loaded on to glass
surfaces, reciprocating motion applied, and the frictional shear force measured. The
friction was measured under various loading conditions and lubricants, to enable
hypotheses to be tested regarding the lubrication mechanisms. The friction coeffi-
cients were extremely small, less than 0.01, far less than was possible using metals
and plastics. Obtaining the lowest possible friction was a major goal of John
Charnley for his artificial hip, which is one reason why the femoral head was only
22 mm in diameter. High torque from a large-diameter metal femoral head would
transmit directly to the cement-bone interface of the acetabular component, pre-
dicted to cause premature loosening. In an anatomic hip, the average femoral head
diameter is 50 mm, but if applied to an artificial hip, this would have left inadequate

234 13 Testing and Tribology
thickness of the plastic acetabular component and also produced a large frictional
torque during flexion-extension.
In the early 1960s a metal-on-metal design of total hip, the McKee-Farrar hip,
was widely used, which did use anatomic dimensions for the femoral head, because
the acetabular component was made as thin metal shell. However, Charnley consid-
ered that the large head diameter as well as the higher-friction coefficient of metal-
on-metal compared with metal-on-plastic would be major disadvantages. To
demonstrate this, Charnley designed a two-channel pendulum friction machine
which is on display at the Wrightington Hospital in England. The hip joints were
loaded, and then an arc of flexion was applied and released. The number of swings
before a hip finally came to a stop was an indicator of the friction. As expected, the
small-diameter metal-on-plastic hip kept swinging for much longer than the large-
diameter metal-on-metal hip. This demonstration was a deciding factor for many
visiting hip surgeons. Another characteristic of metal-on-metal hips was that high
torques could be caused by inadequate ball-socket clearance or by socket distortion
when it was inserted into the acetabulum. Metal-on-plastic hips on the other hand
were not sensitive to such factors.
Fig. 13.1 Peter Walker

studying the pendulum rig
designed by John Charnley
for demonstrating the
relative frictional torque
between two different
artificial hip designs.
MoM = metal-on-metal
hip, McKee-Farrar hip,
36 mm dia. MoP = metal-
on-plastic hip, Charnley,
22 mm dia. The center of
the ball-in-socket was the
pivot for the pendulum. CYL
Red weights WT are
below. The cylinders CYL
at the top apply loads to MoM
MoP
the hips. Both pendulums hip
hip
were given the same initial
displacement. The
pendulum which stopped
swinging first had the
highest frictional torque.
The Charnley hip had
lower friction than the
McKee-Farrar due to the
lower friction coefficient of
the materials and the WT
smaller diameter. (Photo
taken at the Hip Museum,
Wrightington Hospital,
England, 2017)
13 Testing and Tribology 235
Another important requirement for total hips was to minimize the wear of the
plastic component. In the Charnley hip design of the late 1950s, the use of polytet-
rafluoroethylene (PTFE) for the socket had led to excessive wear, even though the
friction was small. It is often assumed that low friction equates to low wear, but this
is not necessarily true for metal-on-plastic bearings. Charnley realized that a wear
test machine was needed which realistically simulated hip function. Rather than just
testing samples of the materials themselves, Charnley designed a four-channel
machine which tested the actual hip joints, where the machine flexed and extended
the joints under a constant load. This test led Charnley to choose high-density poly-
ethylene for the acetabular component, first implanted in 1962. Despite this back-
ground in testing of the friction and wear of hip joints, similar testing was not
recorded for Gunston’s knee, designed in the late 1960s at the same center at
Wrightington Hospital. The only reference to testing the knee was: “A program of
cadaveric trials and load tests was completed before use of the prosthesis in a
patient” (Gunston 1971). There are few records available at Wrightington on the
development of Gunston’s Polycentric knee, and it is likely that the major attention
at that time was still on the Charnley hip.
Fig. 13.2 Sir John

Charnley FRCS developed
the first successful total hip
replacement in the early
1960s, which led directly
to the development of the
artificial knee. It is
impossible to overestimate
the influence which
Charnley had on the field
of artificial joints, whether
in research, design,
technique, or teaching. His
continuous critical
appraisal and his striving
for further improvements
moved the entire field
forwards. And all this was
accomplished in
Wrightington Hospital, a
small country hospital in
North West England,
without the benefit of a
major University center.
(Photo taken at an AAOS
conference c 1990) (©
Peter S Walker)
Fig. 13.3 One of the first joint simulating machines designed to test the wear of artificial hip
joints. In the early 1960s, there were questions concerning the most efficacious materials for the
socket, polyethylene being but one possibility. The machine was designed in John Scales’s labora-
tory at the Royal National Orthopaedic Hospital, Stanmore, England, in collaboration with British
Petroleum Research laboratories. (Photo from Conference Proceedings, Lubrication & Wear in
Human Joints, Leeds, England, 1970)
Another center in the UK where there was an interest in testing artificial hip
joints in the 1960s, was the Royal National Orthopaedic Hospital at Stanmore,
located in north London and affiliated with University College London. The
Professor of Bioengineering, John Scales MD PhD, collaborated with engineers at
the British Petroleum Research Center to develop an elaborate hip simulating
machine for friction and wear testing. This machine applied cyclic loads and
motions applied along three axes to the hip joint being tested. Tests were carried out
on metal-on-metal and metal-on-polyethylene hips, including the Charnley, McKee-
Farrar, and Stanmore designs. While Scales agreed that metal-on-polyethylene pro-
duced lower friction than metal-on-metal, the volumetric wear of the latter was less,
such that Scales concluded: “We believe that it is essential to continue with studies
of all-metal bearings, particularly for those prostheses which might have to last for
upwards of forty years.” This applied mainly to bone tumor implants used in younger
patients, which had been developed.
From the 1950s, John Scales working with JT “Ginger” Wilson and other sur-
geons had developed an expertise in the design and manufacture of prostheses for
replacing large segments of bone in cases of bone tumors. The distal femoral and
proximal tibial implants included an all-metal fixed-axis knee hinge. By the 1960s,
the hinged knee joint had been developed into a stand-alone hinged joint, with plas-
tic bushings in the bearings, used for cases of advanced knee deformity and instabil-
ity. Some basic wear tests of the bushings were carried out comparing metal and
polymeric materials. It was surprising that a metal-polyethylene condylar type of
total knee was not developed at Stanmore by 1970 considering that such a design
was already being designed by Freeman and Swanson in London and by Gunston at
Wrightington. After all, considering the increasing demand for hip replacement,
surely the same would apply to the knee. The answer was the skepticism regarding
the viability of the condylar replacement type of total knee, because by the 1960s,
all-metal hinges had already become well established in England and in other
European countries with designs including the Walldius, Shiers, Stanmore, GSB,
and St. George. These hinges were thought to be a reliable solution for the typical
severe knee problems at the time. Mild to moderate arthritis was still treated with
osteotomy or conservative treatments such as debridement. However, by the early
1970s the situation started to change with condylar knees rapidly gaining ground.
At another London location, Imperial College London, laboratory tests were
developed in the mid-1960s for wear and fixation testing of the Freeman-Swanson
total knee (Swanson et al. 1973; Freeman et al. 1982). The resources of the newly
established bioengineering unit housed in the Mechanical Engineering Department
were well suited for designing and constructing the equipment and for carrying out
the tests. Nevertheless, the wear test machine was relatively simple in construction,
in that it oscillated the femoral component through a flexion arc of 90 degrees with
a constant load of 210 pounds. The tests were only run to one million cycles, which
represents less than a year of usage even for typical patients requiring a total knee.
Furthermore, the loading did not include shear or torque which would most likely
have increased the wear and deformation of the polyethylene. The rationale was:
“Flexion-extension movements….can be applied cyclically….under load. This is
the only movement essential to a knee joint.” As well as measuring the wear, tibial
fixation tests were also considered important. The reason was that the long cemented
intramedullary stems of the hinged knee joints being used at the time provided
secure fixation, whereas there was the possibility that plastic components cemented
only to the surface of the tibia could loosen. The test devised was to cement the
components into tibial specimens and then to apply an axial torque until failure of
the fixation.
To put these tests in context, while data of the forces in the hip joint during func-
tion was obtained at the Bioengineering Department, Strathclyde University,
Glasgow, Scotland, in the mid-1960s, parallel data on the knee was not published
until 1968, after the testing of the initial Freeman-Swanson knee. The knee force
studies carried out as a PhD thesis by James Morrison, under the supervision of
John Paul, significantly advanced the understanding of how the knee functioned in
real-life conditions and provided invaluable data for the design and testing of artifi-
cial knees (Morrison 1968, 1970).
The most well-known center in the field of tribology at the time was the
Mechanical Engineering Department at the University of Leeds in England.
Professor Duncan Dowson had built up a research activity in tribology and was
recognized as one of the leaders in the field with his concepts of elastohydrody-
namic lubrication. When Bahaa Seedhom PhD from Leeds University designed his
cruciate retaining anatomically shaped total knee in the early 1970s, it was the wear
of the polyethylene which was his major concern. The main reason why Seedhom
focused on wear was because he had developed a heat-forming method to shape the
Fig. 13.4 Duncan
Dowson PhD at age 80.
Dowson was a major
pioneer in the area of
tribology at the University
of Leeds, England. He
motivated many researchers
to study the tribology of
human joints and develop
artificial joints with low
friction and wear. He was a
co-organizer of the first
major Conference on Total
Knee Replacement held at
the Institution of
Mechanical Engineers in
London, in September
1974. In his later years, he
produced the authoritative
book on the history of
tribology while continuing
to participate in committee
activities of the Institution
of Mechanical Engineers.
(Photo taken at celebration
event ‘Dowson at 80’)
tibial bearing surfaces. The shapes of these surfaces were generated by moving the
metal femoral condyles through a succession of motion paths which included com-
binations of flexion-extension, internal-external rotation, and anterior-posterior dis-
placement. The resulting surfaces could not be described by circular arcs and hence
were not amenable to machining using the technologies available at that time.
Instead, hand-formed metal molds were pressed against polyethylene preforms to
produce the final surface.
Seedhom considered designing a wear testing machine for the components them-
selves, but this would have been time-consuming and expensive. Instead, he used a
tri-pin-on-disc apparatus to measure the wear of cylindrical pins made from the
heat-formed polyethylene, equipment which was already available (Seedhom et al.
1973). This test method was a standard procedure in the field of tribology where a
plastic pin was loaded onto a polished metal disc with a constant load and sliding
speed and the wear measured at time intervals. Based on the volumetric wear, the
wear coefficient for the material could be calculated.
However, the test did not embody the complex cyclic loading patterns combined
with sliding in different directions, as in the anatomic knee joint. Seedhom circum-
vented these limitations by using the wear data to predict the wear depth of the
plastic socket of a Charnley hip joint and comparing his prediction with that mea-
sured in clinical studies in which the wear was measured radiographically. On find-
ing close agreement with the actual value of 0.1 mm per year, Seedhom felt confident
enough to predict that his knee design would only wear by 1 mm depth in 20 years:
“This is thought to be an acceptable wear rate.” To provide credibility for this pre-
diction, Seedhom measured the number of steps a typical patient would take per
year. The test subjects were fitted with a pedometer which they used each day over
a period of time. Averaging the data from several subjects in each group, the steps
per year were housewives 8000, old people on vacation 6000, and university staff
11,000. By combining this data with Morrison’s force data, Seedhom was able to
calculate the force and distance terms in Archard’s wear equation: Wear vol-
ume = wear factor x load x sliding distance. (There has recently been a renewed
interest in obtaining the activity data of patients with artificial knees, by using elec-
tronic wearable sensors.)
Laxity and Stability
While Seedhom’s work on an artificial knee was proceeding, Peter Walker was
working in the Bioengineering Department at Leeds University, completing his PhD
project on the lubrication of anatomic joints. In November 1969, Walker moved to
the Hospital for Special Surgery in New York. In the design of the Total Condylar
knee in the early 1970s, Walker and colleagues took a different approach to testing,
concentrating more on the laxity and stability of the joint, rather than on the wear.
The former was regarded as a higher priority given that the designs up to that time
ENTITY TO BE
METHOD OF TESTING
TESTED
LAXITY & Specimens Constraint Robotic

STABILITY in test rig testing testing
Displacement
Pin-on-disc Multi-channel control Force
WEAR machine simulator control
Uniaxial offset FEA interface Multi-point

FIXATION loading stresses micromotion
Activity Oxford-type Computer Simulator with

KINEMATICS
simulator crouching machine simulation virtual ligaments
Single axis Multi-channel FEA stress

STRENGTH
fatigue machine fatigue machine analysis
Approx timeline 1970 1980 1990 2000 2010
Fig. 13.5 Summary of the most common testing methods of the artificial knee components used
from 1970 to the present. The different methods of testing are referred to in the text. FEA finite
element analysis
covered the range from high to low conformity, and more experimental data was
needed to determine the ideal (Walker et al. 1974). The theory behind the Total
Condylar knee was that when it was implanted, the laxity and stability behavior
would be the same as when the knee was intact, even though the cruciate ligaments
had been resected. To address this, the laxity and stability of anatomic knee speci-
mens were measured by designing test rigs where the knees were mounted into
fixtures in an Instron loading machine. Combinations of compressive force, shear
force, and torque were applied, and the resulting laxities were measured. All of the
ligaments of the specimens were intact so the laxities represented that of the entire
knee. Typical laxities for the unloaded knee were a total of 14 mm of anterior-
posterior displacement and +/− 15 degrees internal-external rotation. However,
when compressive load was applied, the laxity values reduced considerably due to
the stability provided by the dishing and compression of the bearing surfaces and
the menisci. As a result of this basic research, equations were written from which
the radii of the bearing surfaces of the Total Condylar were calculated, in order to
produce the same combination of laxity and stability as the anatomic knee.
From the mid-1970s, Keith Markolf PhD developed more elaborate testing meth-
ods for measuring the laxity of the knee, both on specimens and on patients, the
main application being to sports medicine. The equipment was ingeniously con-
structed to allow the application of forces and moments but without constraining the
resulting motions in other degrees of freedom. He defined laxity, measured in mil-
limeters or degrees, as well as stiffness, being force/displacement. Markolf used his
methods to measure the laxity and stability of the Geometric and Anametric artifi-
cial knees and the effects of the cruciates (Markolf et al. 1979): “We believe that the
characteristics should be correlated with clinical performance and….future designs
should be evaluated before they are released for clinical trial.” The ASTM have
since formulated a standard for characterizing the laxity called “Standard Test
Method for Determination of Total Knee Constraint” where the amount of con-
straint in an artificial knee is measured under different loading conditions. This
method was developed in the Orthopaedic Research Laboratories of Seth Greenwald
(Postak et al. 1994; Heim et al. 2001). With the femoral component and tibial com-
ponent set at a prescribed flexion angle, a vertical force was applied to the femoral
component. The tibial component was free to move in a horizontal plane. An ante-
rior shear force was applied to the tibia, followed by a medial-lateral force, and
finally an axial torque. The maximum forces and torques at the point of dislocation
were recorded. Of the six contemporary knee systems tested, there was a wide vari-
ation in the forces. These values were compared with those which occurred in func-
tional conditions. The application of the data was to determine whether a particular
design could be used without cruciate ligaments or needed preservation of one or
both of the cruciates, recognizing that normal function is: “dependent upon load
sharing between the implant and the surrounding ligaments….” The authors envis-
aged that different tibial inserts could be selected depending on the status of each
knee: “Implant modularity defines an ability to select different plateau geometries
for a specific femoral component and tibial tray design, extending the clinical appli-
cation of these devices.” This is the usual situation with present-day knee systems.
Fixation of Components
Another aspect of total knee design which required a testing methodology was fixa-
tion of the tibial component. The experience with the Freeman-Swanson, the
Geometric, and the Duo-condylar knees was that interface failure of the tibial com-
ponent at the cement-bone interface was all too frequent. And yet these designs
were all very different, from the solid one-piece construction of the Freeman-
Swanson to the crossbridge connection of the Geometric and to the separate pla-
teaus of the Duo-condylar. It could easily have been assumed that the loosening was
caused by shear and torque forces, especially due to the conforming geometries of
the Freeman-Swanson and Geometric. However, Peter Walker, Chitranjan Ranawat,
and John Insall proposed that an important factor was the axial compressive forces
acting at a distance from the center of the components, producing a rocking or see-
saw effect (Walker et al. 1976). Experiments were conducted in a simple test rig
where unicompartmental components were cemented on to resected tibial surfaces,
loads applied at different points on the surfaces of the component, and the distrac-
tions at the interface measured at the anterior or posterior of the components. It was
found that distractions of up to 500 microns (0.5 mm) occurred, especially when
there was nonuniform bone density from anterior to posterior. The testing protocol
was related to actual surgical cases as follows: “A radiographic study of condylar
replacement knees taken at extension and flexion, shows that in many instances the
loading point lies outside of the central region of the component.” To minimize such
interface distractions, the authors determined that anterior and posterior keyholes
for cement, or a keel on the base of the tibial component, would be effective.
This work, as well as the clinical experience of designs with one-piece or sepa-
rate medial and lateral plateaus, led to a complete rethink of tibial component
design. A one-piece component would have a more rigid construction, but there
would need to be a solid medial-lateral connection, which would eliminate designs
which retained both cruciates. Even a one-piece construction would need a rigid
connection to the tibia, rather than just drill holes or short keels and pegs. Based on
earlier studies of trabecular bone density, the central region of the tibia consisted of
relatively soft bone. This led to the idea of a central peg in this region which would
be flanked on each of its sides by stronger bone and yet would not require excavat-
ing any of the bone on the medial and lateral sides. Such a central peg component
was predicted to have more durable fixation than two separate components.
For the fixation of tibial components, another important factor was the effect of
repetitive loading, where gradual fatigue failure of the trabecular bone could occur
to produce complete loss of fixation, which would not be detected by tests where
single loads were applied. This led to the idea of a cyclic load machine which could
test multiple samples at the same time (Walker and Hsieh 1977). A nine-channel
knee simulating machine was designed, with a flexion-extension arc of 45 degrees
and where an axial force of twice body weight was applied for half the cycle.
Offsetting the axis of rotation of the femoral component caused the point of force
application on the tibia to displace anterior-posterior, reproducing the off-center
loading of the earlier static experiments. This multichannel machine allowed for
multiple tests to be carried out for millions of cycles in a practical time period. With
the one-piece central peg design, there were no interface failures, whereas with the
separate plateaus, almost every one failed by fatigue failure of the trabecular bone
around the interface.
Following this work, there were few reported studies on tibial fixation in the
second half of the 1970s, which is surprising because tibial component loosening
was a common problem with the early designs. One study compared the failure
modes of two of the earliest designs, the Polycentric and the Geometric knees, when
subjected to axial loading (Nogi et al. 1976). The components were cemented into
complete knee specimens, the femur and tibia potted in cylinders, and the load
applied through a simple press. In both designs, there was fracture of the acrylic
cement at the tibial interface. For the Polycentric, in addition, there was sinkage of
the tibial components due to compressive failure of the underlying trabecular bone.
Examples of both failure modes were seen clinically. Absence of failure of the fem-
oral component was notable, which was later explained qualitatively by the finding
that the compressive strength of the trabecular bone in the distal femur was higher
than in the proximal tibia (Colley et al. 1978). The interaction of several processes
over time was later proposed as failure mechanisms (Ducheyne et al. 1978): varus
misalignment leading to medial overload, trabecular failure causing even higher
medial loads, and bending of plastic components.
At the Orthopaedic Research Society (ORS) meeting in 1979 (Clement B Sledge
MD was President that year), Bill Krause PhD and Joe Miller MD from McGill
University in Montreal, Canada, presented their work on the importance of cement
penetration into trabecular bone and how to achieve it. The work was published a
few years later (Krause et al. 1982). It was proposed that no matter what was the
design of the tibial component, unless several millimeters of cement penetration
was achieved over the entire interface, bone resorption and fixation failure would
occur. Cement penetration was achieved by lavage cleaning and then pressurizing
using special nozzles attached to syringes (see Chap. 5). This work led to a major
change in the way in which cementing was carried out, but in certain conditions, it
was found that there was excessive cement penetration, which was considered unde-
sirable. Several years later, Larry Dorr MD, from the Good Samaritan Hospital in
Los Angeles, showed that excessive cement penetration could be avoided, by dem-
onstrating that 2–4 mm was reproducibly achieved using lavage and finger pressing
in surgical conditions (Dorr et al. 1984).
Meanwhile there was still no consensus on the optimal design for the tibial com-
ponent. At the Brigham and Women’s Hospital in the early 1980s, an apparatus for
fixation testing was constructed where tibial components were fixed to tibial speci-
mens, and a simulated femoral component was loaded from above such that combi-
nations of forces and moments were applied, representing functional loading. Using
noncontact transducers, distraction or compression of the tibial component relative
to the tibia was measured at points around the periphery (Walker et al. 1981). The
criterion was that the least deflections signified the most rigid fixation. Not surpris-
ingly, one-piece components with metal-backing and a central peg were the most
successful. Using the same test rig, Peter Brooks MD, a resident at the time,
developed wedges to buttress bone deficiencies seen in cases of severe varus or

valgus deformity (Brooks et al. 1984).
At the end of the 1980s, after the early experience with uncemented components
coated with fiber-metal on sintered beads, there was renewed interest in minimizing
the interface micromotion between the porous surface and the trabecular bone.
Several laboratories carried out tests where the tibia and its component were
mounted in a base, while a vertical rod from a loading machine applied forces to the
tibial surface, either centrally or offset to simulate varus loading. Deflections
between the component and the bone were measured using transducers. Three
examples of such tests are cited here (Volz et al. 1988: Branson et al. 1989; Whiteside
and Pafford 1989). Taking a composite of the results, it was found that offset loads
usually produced unacceptably high distractions or compressions of the component
relative to the bone. Certain combinations of pegs reduced the motions, but screws
into the cancellous bone minimized the motions and hence were adopted clinically
for most of the uncemented design systems. This is an example where a relatively
simple testing method could result in a practical solution which succeeded in prac-
tice. However the method did rely on previous basic work which highlighted the
importance of interface micromotions as a crucial experimental parameter.
Simulating Machines
We now return to the early 1970s to the development of more elaborate machines
which attempted to simulate both the loads and the motions which occurred in walk-
ing and other activities. David G Murray MD from Upstate Medical Center in
Syracuse, NY, was developing a new type of artificial knee, the Variable Axis. To
validate its mechanical efficacy, some type of testing machine was needed to evalu-
ate the wear and fixation strength. Murray collaborated with John Shaw PhD at the
Mechanical Engineering Department. They designed and constructed a simulating
machine which was based on reproducing normal knee mechanics of the lower
extremity (Shaw and Murray 1973). The data of kinematics and forces was already
available from James Morrison PhD, which was used in the design of the machine.
Instead of loading the femoral or tibial components directly, they applied a vertical
body weight to the hip joint and obtained flexion and extension motion across the
knee by applying forces to a chain attached to the tibial tubercle, which reproduced
the quadriceps mechanism. Since the muscle forces required to perform a specific
activity are a function of the joint geometry, physiological joint loading could be
achieved for any given external load by ensuring that all forces crossing the joint did
so at anatomic locations and directions. The machine was intended to simulate the
lower extremity performing a high flexion activity such as walking up and down
stairs and crouching. However, due to the limitations of the control system, only
approximate wave patterns for motion and forces could be achieved. The concept of
the Shaw-Murray dynamic machine was adopted by Ben Hillberry PhD and others
at Purdue University, Indiana (Maletsky and Hillberry 2005). The intention was to
simulate vigorous activities, which meant high forces and rates of change of joint
External load
Quadriceps
length changed
continuously
Data
recording Load applied
through ball-in-
socket
Knee specimen
with artificial
knee cemented
in place
Simulated ankle
Logic control
circuitry
Fig. 13.6 The first knee simulator of its type designed to simulate the loads and motions during
walking, such that wear and fixation of an artificial knee could be studied. A vertical external load
is applied at the top while the flexion-extension is controlled by changing the quadriceps length
during the cycle. A force is generated in the quadriceps cable to equilibrate the external force. The
direction of the external force in the frontal plane could be modified to simulate varus or valgus
misalignment (Shaw and Murray 1973). (Reprinted with permission from Clinical Orthopaedics &
Related Research)
angles, for which data was now available. The machine was also to be used with
knee specimens or with artificial joints mounted on fixtures. The authors validated
that the tibiofemoral forces were realistic, but even so, the machine simulated quad-
riceps forces only and did not include the hamstrings, which would have affected
the force vectors at the knee. Mechanical complexity limited the application of the
machine to short-term tests, such as determining joint forces and contact forces at
the knee, rather than long-term fixation or wear studies.
Recognizing the complexity of simulating actual functions, other researchers
developed test equipment which was easier to construct and operate yet could still
provide meaningful data of some aspect of joint mechanics. Such equipment was
first developed at Oxford University, called the Oxford Knee Rig. The Shaw-Murray
and the Purdue were machines in that they provided continuous motion and included
changing forces and motions. On the other hand, a rig is a construct where a knee is
subjected to loading under static conditions, even if the forces are provided by simu-
lated muscle actions. In the Oxford Rig, the tibia was fixed vertically (or at slight
flexion), the femur was flexed, a wire was attached to the quadriceps, and the other
end of the wire was attached to a weight which was hung over a pulley at the proxi-
mal end of the quadriceps. Adjusting the length of the wire produced different
Fig. 13.7 The original

Oxford Knee Rig. The
forward flexion of the tibia
was adjustable from 0 to
30 degrees. The flexion
angle was controlled by
screw adjustment of the
quadriceps length. The rig
was used to study the
instant centers of rotation
of the knee and the contact
areas in the patellofemoral
joint. A limitation was that
the maximum load that
could be applied was 75
Newtons (Goodfellow
et al. 1976). (Reprinted
Journal of Bone & Joint
Surgery B)
flexion angles. The Oxford Rig using this configuration was designed in 1976 to
carry out studies of the contact areas in the patellofemoral joint (Goodfellow et al.
1976; Harding et al. 1977). Later, John O’Connor PhD enhanced the functionality
by changing the rig to a machine, which simulated the entire lower extremity and
where flexion-extension was actively controlled by lengthening and shortening the
quadriceps length. This simulated a crouching or squatting activity, useful in that
the knee could be studied from full extension to full flexion. The mechanical com-
plexity was greatly reduced from the earlier functional simulating machines. As a
result, the Oxford-type machine was widely used in many other laboratories. Using
such a machine, Robert Singerman et al. (1995) at Case Western Reserve University,
Cleveland, OH, measured the patellofemoral forces and found that the maximum
force occurred at 70 degrees flexion and was less than the quadriceps force, peaking
at 0.8 times body weight. The center of pressure moved from the lower to the upper
patella as flexion increased. They showed the effect of the quadriceps force direc-
tion, the Q-angle, giving a tendency for the patella to sublux laterally. This had has
application to femoral component design and surgical technique. Some time later,
Ezzet et al. (2001) working with Darryl D’Lima MD and others at the Scripps Clinic
in San Diego expanded this work to an artificial patella and found improved kine-
matics with a medialized inset patella. In both of these studies, there was consider-
able variation in the data between specimens. Other researchers also used the same
type of test machine. Miller and Zhang agreed with patella medialization and found
no effect on forces when the femoral component was externally rotated, but showed
the benefits of a dome patella due to variations in tilting between knees; Rovick and
Walker at the Brigham and Women’s Hospital determined average knee motion and
used the data for the design of an asymmetric total knee.
One of the limitations of the Oxford-type machines was that the hamstring forces
were not included, which results in some errors in kinematics. An important aspect
of the design of such machines is that the knee joint should be provided with suffi-
cient freedom of motion. This concept was pointed out by Piziali (see Chap. 4). In
particular, although flexion-extension is applied by the changing lengths of the
quadriceps, the following motions need to be unconstrained so that the femur and
tibia continuously find their equilibrium positions: medial-lateral and anterior-
posterior displacements and varus-valgus and internal-external rotations. To pro-
vide this, Zavatsky (1997) determined that in the design of the test machine, the hip
should allow flexion-extension and ab-adduction, while the ankle should allow rota-
tions about three axes. An excellent review of this type of testing machine was writ-
ten by Kartik Varadarajan (2009). He and Guaon Li compared the data from the
machines to data from fluoroscopic studies of normal subjects and artificial knee
patients, to determine how relevant was the in vitro data from the Oxford-type
machines. Notwithstanding the complexity of making such a comparison due to the
numerous studies in the literature and the different designs of the machines and
methods of testing, it was concluded: “Results showed that the machines …. can
capture certain key features of closed-chain flexion-extension kinematics of both
healthy and TKA knees measured in vivo. These systems can replicate the femoral
rollback with flexion and screw-home tibial rotation between 0 and 30 degrees flex-
ion seen in healthy subjects, and the reduced femoral rollback and absence of screw
home motion in TKA patients.”
By the year 2000, a number of researchers had seen the possibilities of using
robotic equipment for testing knees. One advantage was that the existing instrumen-
tation and software built into the machines would provide data on 6 degrees of
freedom motion and forces. A review of studies carried out using robots was
reviewed in the article cited above by Varadarajan and Li. Their robot was used for
a number of studies of knee specimens and with specimens where artificial knees
were implanted. In one example (Li et al. 2004), the contact pressures between the
patella and the femoral trochlea were measured along the neutral path of motion.
The method was to position the knee in an unloaded condition at a chosen angle of
flexion and to sum the forces along the three axes. At this angle, the robot moved the
knee incrementally in all degrees of freedom to find the position where the sum of
the forces was a minimum. This defined the neutral position of the femur relative to
the tibia at that flexion angle. In this way a neutral path of motion was found for a
full flexion range. In subsequent testing where loads were applied, the displace-
ments and rotations were measured relative to the neutral path. In the above exam-
ple, both quadriceps and hamstrings forces were applied through ropes and pulleys.
It was demonstrated that patella pressures changed considerably for both muscles
compared with the quadriceps alone, from which the authors proposed that both
muscle actions were needed in such testing. One limitation of their robot, and oth-
ers, was that forces only up to about 200 Newtons could be applied, rather than the
much higher forces of activities. This limitation was later overcome by Keith
Markolf PhD and colleagues at the UCLA Orthopaedic Biomechanics Laboratory at
the UCLA Rehabilitation Center (Lin 2016). Using an industrial robot, their studies
Fig. 13.8 Researchers using a robot arm to carry out biomechanical experiments on knee speci-
mens. The robot applies and measure forces, displacements, and rotations in three dimensions to
knee specimens before and after the placement of an artificial knee. This determines how closely
the artificial knee replicates the anatomic knee. (Photo taken at the Zimmer Institute, Warsaw, IN
c 2006)
are being mainly applied to ligament reconstructions rather than artificial knees:
“The 8 foot high 2 ton robot….applies hundreds of pounds of force to a cadaver
knee specimen implanted with a custom-designed sensor that measures forces in a
knee ligament….Pairing this with a computer program that analyzes every move,
the researchers are shedding new light on how the knee works, how it gets injured
and how best to repair it. We need a big robot to simulate big forces.”
A different mechanical configuration to the above test machines was needed for
wear testing, for which physiological loading up to at least five million cycles was
required. At one cycle per second, a test would take almost 60 days of continuous
running, even more due to the need to make wear measurements at least every mil-
lion cycles. This demanded a multichannel machine. One of the first such machines
was developed in the 1990s at the Royal National Orthopaedic Hospital in Stanmore,
England. As a result of the work of John Scales and his staff, there were already a
number of test machines in the laboratory being used for fatigue testing. However,
in order to provide a more realistic simulation of in vivo conditions, the loading and
motion cycles needed to be added to reproduce everyday activities, particularly the
most frequent activity of walking. Another factor was that the sliding distances
would vary depending on the artificial knee design. For example, a design of low
conformity with posterior cruciate preservation would be expected to have greater

sliding distances than a more conforming design even if the cruciates were resected.
The machines should therefore be designed to account for whatever artificial knee
was being tested.
When Peter Walker became Director of Bioengineering in 1987, he decided to
design machines which could simulate real-life load and motion conditions. To
develop such a wear testing machine, Walker organized a consortium of manufac-
tures to contribute funding, along with the UK’s Department of Health. Academic
collaborators included Tom Andriacchi (Stanford University), John Paul (Strathclyde
University), and David Broome and Mike Dewar (University College London). The
result of this endeavor was the four-channel Stanmore Knee Simulator (Walker
et al. 1997, 2000). A key feature was that this was a “force-controlled machine,”
where the prescribed input forces were based on Morrison’s data from 25 years ago.
This difficult aspect of the machine, accounting for the type of artificial knee design
being tested, was solved with the collaboration of Hani Haider PhD, who joined the
Department at that stage. In parallel with the mountings for the femoral and tibial
components, an arrangement of springs was designed, with strong or weak stiff-
nesses depending on which of the cruciate ligaments was retained or resected. This
concept was in contrast to an alternate scheme called “displacement control” where
prescribed anterior-posterior displacements and rotation angles were imposed,
along with the flexion-extension angles. With the force-controlled concept, the knee
could determine its own displacements and rotations depending on its constraint.
The validity of this concept has been borne out by numerous kinematic studies of
total knees in patients using fluoroscopy.
John DesJardins PhD, now at Clemson University, SC, who joined the simulator
project at the validation stage, carried out studies to demonstrate the efficacy of the
method for kinematic comparisons as well as for wear testing (DesJardins et al.
2000). Sixteen of the machines were made by the lab’s accomplished technical staff
and supplied to companies and laboratories around the world. Hani Haider in his
laboratory at the University of Nebraska in Omaha later carried out numerous wear
tests of artificial knees for companies, many for FDA requirements, and it was
largely due to his efforts that the ASTM testing standard for wear testing was
adopted. He has written a comprehensive review of the field of wear testing of arti-
ficial knees, explaining the different types of wear machine and the protocols for
testing (Haider 2016). Another review article was on the comparison of the wear
between fixed bearing knees and mobile bearing knees (Haider and Garvin 2008;
Haider and Kaddick 2012). The latter paper, coauthored with Christian Kaddick
PhD, includes wear data of many different artificial knee designs, both fixed and
mobile bearings. Among the recent studies was a comparison of vitamin E polyeth-
ylene compared with previous types of the material (Haider et al. 2012).
Since that time, other knee simulating machines for wear testing have been pro-
duced. A notable example is the AMTI (Advanced Mechanical Technologies Inc.,
Watertown, MA) 6-station Knee Simulating Machine, which can be run in either
force control or displacement control modes. While the Stanmore simulator used
springs to represent ligaments, the AMTI used virtual soft tissues, a major
Femoral Component Flexion Angle

Actuator
Tibial Component
(in Fluid Bath)
A/P Force and

Axial Rotational
Shear Force Actuators
A/P Displacement and Axial

Rotational Soft Tissue
Restraint System Axial Tibial
Compressive
Force
A
Fig. 13.9 The 4-station force-controlled Stanmore Knee Simulator was designed at the Royal
National Orthopaedic Hospital, Stanmore, University College London (Walker et al. 2000). Hani
Haider and Peter Walker devised a method for replacing the constraint from soft tissues using a set
of springs (inset figure). The principal purpose of the machine was for long-term wear testing over
millions of cycles. The machine could also be used for kinematic comparisons between different
artificial knee designs (DesJardins et al. 2000). (Reprinted with permission from Elsevier Inc.)
Fig. 13.10 Hani Haider

PhD joined Peter Walker’s
lab at University College
London in 1990 where he
collaborated on the design
of the Stanmore Knee
Simulator, used for wear
and kinematics testing. In
2000, he became Director
of Research at University
of Nebraska, Omaha,
where he built up a
laboratory specializing in
wear testing and simulator
development. Haider is an
active member of the
American Society for
Testing and Materials
(ASTM). (Photo at ISTA
conference, Kyoto, Japan,
2014) (© Peter S Walker)
advancement (White et al. 2006). The concept is that each ligament is represented
by its attachment points and its stiffness characteristics. As the external forces and
motions are applied to the artificial knee being tested, the force vectors from the
ligaments are computed in real time and combined vectorially with the applied
external forces. To make the wear simulation even more realistic, Bruce White com-
piled the data of knee forces and kinematics in different activities determined by
Georg Bergmann and colleagues (Bergmann et al. 2014). White then proposed a
wear test sequence which embodied walking as the main activity, interspersed with
stair ascent and descent and other activities. This test scenario was published by
ASTM as guidelines but so far has not been incorporated within a standard. There
is naturally some reluctance from existing wear test laboratories, to implement
major changes to their existing test machines due to technical complexity and cost.
Nevertheless, individual experiments have been carried out to investigate the effect
of more elaborate and realistic testing conditions.
At the Endolab in Thansau, Germany, Christian Kaddick PhD has for many years
carried out standard testing protocols but recently explored machine modifications
to simulate highly demanding daily activities (Schwiesau et al. 2013). They
Fig. 13.11 Georg
Bergmann PhD, from
Charite Hospital, Berlin.
He, together with
colleagues, developed
instrumented artificial
knees from which the
forces on the knee were
determined for a range of
activities. His articles
summarizing all of the data
are an invaluable resource
for biomechanical testing
and computer modeling.
(Photo circa 2005) (©
Peter S Walker)
concluded that “further testing is needed to better understand the impact of each
activity and the respective influence….” At the Institute of Medical and Biological
Engineering, University of Leeds, England, John Fisher and colleagues concluded
that sophisticated control systems were needed in their electromechanical machine
to carry out such studies (Abdelgaied et al. 2017). Research in the area of wear test-
ing of artificial knee needs to account for technical complexity, high cost of equip-
ment, extended time taken for testing, techniques for accurately measuring
exceedingly low wear rates, the lack of complete biomechanical data of a range of
activities, variations in function between individual patients, and differences
between artificial knee designs and materials. To counteract these problems, studies
can be carried out on a comparative basis, for example, comparing the wear between
two designs, or a single design with different bearing materials.
A single-channel test machine called the VIVO Joint Simulator for studying the
kinematics of the knee was recently designed and introduced by the AMTI com-
pany. Forces and flexion-extension are input to the knee based on prescribed cycles,
Fig. 13.12 The AMTI

VIVO Joint Simulator for
testing knee (and other)
joints. The machine has a
6-degree-of-freedom
system for applying and
recording forces and
motions, such that different
activity cycles can be input
and the output
displacements and
rotations measured. A
unique feature is a virtual
ligament restraint system
where ligaments can be
input as defined by their
attachment points on the
femur and tibia and their
elastic properties. Multiple
fibers can be defined for a
given ligament, such as the
medial collateral, to obtain
more anatomic realism.
(Photo at AMTI Inc.,
Watertown, MA, with
permission c 2015)
but the ligaments are represented virtually, as described above (White et al. 2006).
The attachment coordinates and stiffness values of the ligaments are input, while the
software computes and applies the force vectors from the ligaments continuously
during the motion cycle. Hence the effects of the ligaments on the motion can be
readily determined and surgical aspects such as component alignments and soft tis-
sue balancing can be evaluated. This capability of modeling the ligaments virtually
is valuable for several reasons. It is not easy to identify a suitable material with a
similar load-deflection curve to actual ligaments. Attaching synthetic ligaments at
each end is problematic. Conveniently making adjustments in length to simulate
varying tightness is difficult. Measuring ligament tensions directly in situ has not
been solved in a simple way. Perhaps most importantly, ligaments are multi-bundle
with the different parts changing their tensions throughout motion, a property which
cannot be simulated practically. These problems can be solved with the virtual liga-
ments. An example of a study where the VIVO machine has been used for studies
on artificial knees has been reported (Willing and Walker 2018).
Long-Term Durability and Wear Particles
A frequently asked question about artificial knees is how long they will last, as if
there were a particular time point at which the bearing surfaces wear out or some
other definitive failure mode occurs. A better question would be: what is the prob-
ability that an artificial knee will still be in place and functioning well after a certain
period of time, say 10 years or 20 years? That is the essence of the Kaplan-Meier
estimate, a method which has been most frequently used in follow-up studies includ-
ing large rational databases. The survival probability is (the number of cases at the
start – number of failures) ÷ (the number of cases at the start). Another way to
express the data is by the revision rate, as shown by the Swedish Knee Registry (see
Chap. 1). A 0.95 probability of survival is equivalent to a 5% revision rate. In that
graph, for the most recently used artificial knee designs, the probability of survival
at 10 years was close to 0.95 and at 20 years, close to 0.92. The next question might
well be that because those figures were obtained for artificial knees performed 10
and 20 years ago would not today’s artificial knees have an even higher survivor-
ship. The overlap in the curves of the two most recent 10-year groups suggests that
the survivorship may be reaching a limit.
In the 1970s and 1980s the most frequent modes of mechanical failure were
component loosening, instability, wear-through of the polyethylene, and component
fracture. Those problems were significantly reduced by improved cement technique,
more accurate alignment, use of designs with inherent stability, more accurate soft
tissue balancing, reduced contact stresses and wear-resistant polyethylene, and
stronger components which had been thoroughly tested for fatigue strength. In these
examples, the association between the causes and the failure modes can easily be
seen. In contrast, determining the effect of wear particles has not been so easy. One
reason has been that while an accumulation of polyethylene particles could result in
bone resorption due to a biological response, resorption could also occur due to
stress protection from a component which was too stiff, leading to a reduction in the
density of the surrounding bone. Both situations could result in a symptomatically
loose component. A further point is that the survivorship values represent the aver-
age over a large number of cases, many surgeons and hospitals, and types of artifi-
cial knee designs. For a particular case, the probability of survival at 10 years could
be higher or lower based on risk factors such as inadequate cementing technique
and misalignment, as well as smoking, obesity, and high activity levels.
The design of the artificial knee has until now been in a state of continuous evo-
lution, and there is every sign that this process will continue for some time. In the
early stages, many different design configurations were introduced within only a
few years. As time went on, early failure of certain designs resulted in withdrawal
or redesign. Successful design features were incorporated into new designs. As
pointed out earlier in this book, after about only 10 years, the number of design
types had been narrowed down to only a few, although the many different manufac-
turers had their own variations. The same process occurred for the materials, with a
convergence to ultrahigh molecular weight polyethylene. Instrumentation too
became standardized, with the alignment goals, bony landmarks being more
accurately defined, and slotted cutting guides used for the bone cuts. In the ensuing
years, the main attention was on the finer points of the design, materials, and tech-
niques. The goal was to achieve the best performance, however, that was specified
and measured. The different design concepts were compared in many ways includ-
ing their stability and functional outcome. Materials were compared for wear and
strength. Techniques were compared for consistency of alignment and convenience
of use. A crucial measure was the survivorship, how long did artificial knees last
before some mode of failure occurred. While some failure modes were evident after
only a few years, others took many years to develop.
Complete wear-through of the tibial component, seen in some early designs, was
avoided by using bearing thicknesses of at least 6 millimeters, by reducing contact
stresses, and by avoiding misalignment. Nevertheless, wear still needed to be mini-
mized because of the vast numbers of particles liberated into the surrounding tissue.
In excess, this could cause an inflammatory response, and bone resorption at the
interface, leading to loosening. These phenomena were demonstrated in meticulous
scientific experiments where titanium fixtures were embedded into rabbit tibiae and
responses of bone to both motion and particular debris demonstrated (Goodman
1994). “Both micromotion and particle materials may thus play important roles in
prosthetic loosening. If excessive micromotion is present.…a fibrous tissue layer
will form at the bone-implant interface. This layer may function as a conduit for
particles....and elicit an influx of foreign body and chronic inflammatory cells that
may initiate bone resorption.” Later research showed that particle shape and size
also played a role in the magnitude of these processes. The clinical relevance of this
work was determined in a 10-year follow-up of almost 400,000 cases in the
Australian Registry (Steiger et al. 2015). During that time, there was a steady shift
in usage from ultrahigh molecular weight polyethylene (UHMWPE) to crosslinked
polyethylene. The crosslinked material was produced by gamma irradiation of the
UHMWPE. When the survivorship curves were compared for the two materials at
10 years, there was a much lower failure rate for the crosslinked material. For exam-
ple, in all of the cases combined, the revision rates were 5.8% for ultrahigh and
3.5% for crosslinked. For patients under 65 years old, the values were a similar
proportion but 50% higher. The higher revision rate of ultrahigh molecular weight
versus crosslinked polyethylene was attributed to bone lysis and loosening.
For many years it has been a requirement of the FDA that any new design,
whether or not it uses a new material, shall have a wear test carried out. The larger
companies were equipped to carry out such testing in-house, but independent labo-
ratories developed the special expertise required and offered the service to the field
as a whole. The largest such laboratories are at the University of Nebraska (Hani
Haider) and the Endolab in Bayern, Germany (Christian Kaddick). They both used
similar test methods, force-controlled testing under ISO 14243 specifications. In
combination they have measured the wear of 73 different fixed bearing designs and
50 mobile bearing designs (Haider and Kaddick 2012). The expectation was that the
mobile bearing would show less wear, because of the higher conformity and hence
lower stress of the bearing surfaces. Haider’s results were mean wear for fixed
12.7 mg/million cycles, for mobile 7.8. Kaddick’s results for mean fixed 5.8, for
mobile 5.3. The values had a wide range, however, for example, a range between 1
and 24 for the fixed bearings and from 3 to 13 for the mobile bearings: “.…no con-
sistently statistically significant difference was formed that could characterize
mobile bearings as having to have lower wear than fixed bearings.” Haider did how-
ever isolate two designs, one a fixed bearing and the other a mobile bearing, identi-
cal in materials and in most design features and showed again no significant
difference in wear (Haider and Garvin 2008). One major finding from the combined
study however was: “yet such differences were clearly featured….of significantly
lower wear in highly crosslinked UHMWPE than in the conventional UHMWPE
bearings.” This finding was corroborated in an extensive retrieval study.
Implant Retrieval and Wear
Meanwhile, since the 1970s, invaluable data on the wear of polyethylene has been
obtained from retrieval studies. Early reports featured wear completely through the
plastic insert, usually caused by misalignment and instability. For systematically
describing the wear, the early studies defined the types of wear and surface damage,
and specified different zones on the tibial component. Burnishing and abrasion
occurred due to local fatigue failure of the polyethylene and to entrapped granules
of acrylic cement. Later, large surface areas of delamination were shown, as much
as several millimeters in depth, due to coalescing of subsurface fatigue cracks
beneath the surface, most often seen in designs with low femoral-tibial conformity.
Over time, due to improved designs and materials, the wear was seen to reduce. This
was recently highlighted in an outstanding scientific exhibit shown at the American
Academy of Orthopaedic Surgeons (AAOS) annual meeting in 2018 (Van Citters
et al. 2018). The authors were from the longest running retrieval laboratory, at
Dartmouth College in Hanover, New Hampshire, founded 30 years ago by John
Collier, Professor of Engineering, and Michael Mayor, Professor of Orthopaedic
Surgery. So far over 19,000 retrievals have been analyzed, of which over 9000 were
artificial knees. They found that the volume of wear increased over time as expected,
but that the wear rate reduced when crosslinked polyethylene was introduced.
Backside wear, an important contributor to wear, was reduced when the metal tray
surface was polished. The overall conclusion was positive: “due to steady improve-
ments in designs and materials, all of the designs included in this study show a large
and significant decrease in wear rate compared to the historical reference rate.” One
caveat was noted, however, the possibility of fatigue damage at longer in vivo times
due to oxidation, even of highly crosslinked polyethylene (Currier et al. 2015).
The improvement in wear resistance from crosslinking came with a price, how-
ever, loss of ductility and resistance to fatigue crack propagation (Rimnac and
Pruitt 2008). In the case of artificial knees, this could lead to fracture of the posts
in the PS designs or loosening of the snap-in mechanism of the modular connection
in metal-backed tibial components. An extensive body of research to develop a
processing method for minimizing wear and oxidation, yet preserving mechanical
properties, has been carried out since about the year 2000, at the Harris Orthopaedics
Fig. 13.13 Michael
Mayor MD (photo) and
John Collier PhD in 1976
founded the implant
retrieval program at the
Thayer School of
Engineering, Dartmouth
NH. Since then, over 9000
knees have been studied by
the researchers including
John Collier, John Currier,
and Doug van Citters. A
steady reduction in wear
has been observed over the
years due to improvements
in polyethylene processing
and implant design. (Photo
taken at AAOS Annual
conference c 1990) (©
Peter S Walker)
Laboratory in Massachusetts General Hospital, by Orhun Muratoglu, Ebru Oral,

and their group (Rowell et al. 2015). These authors produced a review article
explaining the nuances of their processing methods and verification tests (Oral and
Muratoglu 2011). Their conclusion was that while irradiation resulted in a tenfold
reduction in wear, there was no method for preventing oxidation over time and
maintaining mechanical properties. This dilemma was solved by adding a well-
known and biocompatible antioxidant, vitamin E: “The clinical performance will
be determined using prospective, randomized clinical studies, which are under
way.” Such studies may require follow-up times of 5–10 years at least before the
effects are seen, considering that long-term prevention of oxidation is one of the
proposed benefits of the vitamin E material. Hence while there is strong evidence
that crosslinking will reduce the wear compared with UHMWPE, the wear rate of
the crosslinked material may start to increase after some years in the body. The
time frame is not known, but based on the Australian Registry data above, it seems
that there was no acceleration of wear up to 10 years. To verify that the addition of
vitamin E will prevent this increase in wear rate in vivo may take 10 years or more
Fig. 13.14 Orhun Muratoglu PhD (left) is Director of the Bill Harris Orthopaedics Laboratory at
Mass General Hospital, Boston, and has made major contributions in developing wear-resistant
formulations of polyethylene. Joshua Jacobs MD PhD is Chairman of Orthopaedics at Rush
University Medical Center in Chicago and is a specialist on the biologic effects of biomaterials
used for artificial joints. (Photo taken at the ISTA conference, Boston, 2016) (© Peter S Walker)
of study. Laboratory wear studies have, however, provided indications. In one

study (Grupp et al. 2017), a knee simulator was used to test artificial knees, with
different materials. The materials were artificially aged at the start of the test, to
simulate an extreme oxidative environment. The test were run under loading condi-
tions which included strenuous activities, not limited to the present ASTM test of
level walking. Vitamin E addition was found to prevent delamination occurring
after several million cycles. Many such tests would have to be carried out to pro-
vide a broad spectrum of conditions, but such simulator tests have an advantage of
attempting to simulate realistic conditions with a high level of usage, typical of
younger more active patients.
Computer Modeling
While the physical testing of artificial knees using equipment such as described
above continues to be widely used, much of the kinematic evaluation of artificial
knee mechanics has moved to computer simulation in recent years. In one study
(Fitzpatrick et al. 2016), the computer model was validated by comparing the output
data from a 6-degree-of-freedom test machine, for deep knee bend, gait, and step-
down. The model was also shown to be able to differentiate between different
designs “performing” these activities. Their computer model was used to carry out
constraint testing (Rullkoetter et al. 2017), as described above in relation to an
ASTM standard. The computer model was able to extend the testing to the inclusion
of soft tissues and then to muscle forces, giving a more realistic evaluation of the
Fig. 13.15 Paul
Rullkoetter PhD from the
University of Denver,
Colorado, along with
colleagues including Claire
Fitzpatrick, has developed
FEA computer models of
the knee and lower
extremity, for the
evaluation of different knee
designs in various
functions. This is a
growing area in joint
biomechanics. (Photo
taken c 2016) (© Peter S
Walker)
in vivo situation in function. A major advantage of computer modeling is that patient

variability can be taken into account (Taylor et al. 2013): “Statistical shape models
can synthetically generate thousands of representative models from a much smaller
training set.” If modeling was used to compare different designs, outputs under dif-
ferent service conditions address the goal of robust design (see Chap. 3). In the area
of artificial knee design, any design can be optimized. Comparisons can be made
between different designs and variations. The effect of patient and surgical variables
can be investigated. Different functions can be evaluated. Knees could even be cus-
tomized given the input data of the patient (Koh et al. 2019). There is a big differ-
ence between validating a computer model based on output data from a test machine
(which is itself a “model” of sorts) and predicting in vivo behavior. In the first case,
the target data is available, and the test machine has well-defined functional behav-
ior. In the second case, the system is far more complex, and the multiple variables
are not all known.
This predicament was the main purpose of “The Grand Challenge” organized by
Benjamin Fregly et al. (2012). Participants were provided with data sets: “These
PC
ALC
MPFL PCL
LCL
MCL
PL
Fig. 13.16 A computer model of the knee with artificial knee implanted, within a full skeletal
AnyBody Model of the musculoskeletal system. The model includes multiple bundle models of the
ligaments and muscles. Using an inverse kinematic-based analysis, forces and kinematics can be
predicted for different activities or artificial knee designs. In this case, the effect of tibial slope was
investigated (Marra et al. 2017). (Reprinted with permission from Journal of Biomechanical
Engineering)
data sets include tibial contact force, video motion, ground reaction, muscle EMG,
muscle strength, static and dynamic imaging, and implant geometry data.
Competition participants used their musculoskeletal models to predict tibial contact
forces without having access to the corresponding in vivo measurements.” A crucial
point is that the data was obtained from patients who had instrumented tibial com-
ponents, whether from the Berlin group (Bergmann) or the San Diego group
(D’Lima and Colwell). Hence the tibial force data, as well as the data provided to
the participants, was available for various functional activities. The results over after
several rounds of the Challenge were discussed at a symposium at the Orthopaedic
Research Society in 2019. Achieving consistency of results between different com-
putational methodologies was not straightforward due to the complexity of the mod-
els, the differences in the models themselves, and the variations of the input data
which was needed for the models but not provided. Even today, basic data of soft
tissue attachment points and geometries, and their material properties, is conspicu-
ously inadequate, and different studies frequently use different values. In the discus-
sion of the application of the work, there were two extremes. At one extreme, there
was the scientific value of a model which was able to accurately predict the tibial
forces. At the other extreme, for the purposes of defining a test, for the ASTM, for
example, the test itself should be a simple as possible such that it could be easily
implemented and be reproducible between different laboratories.
Despite the apparent objectivity of computer modeling, Rullkoetter et al. (2017)
pointed out: “Although models typically represent the articular interface well, the
interaction with the patient and the surgical process includes significant variability
and increase in complexity.” It might well be impractical to determine such vari-
ables on a given patient, even if it were of benefit to do so. But despite the limita-
tions of computer models, it is clear that there will be increasing application in
artificial knee design and evaluation. Rullkoetter stated: “It is evident that future
artificial knee designs will be optimized by using preclinical computer models.” A
limiting factor may be knowing what are the optimal design criteria.
Current Status of Testing and Standards
Today, there are published “standards” for testing all types of materials and products,
such that in order to be released to the market, both safety and efficacy must be dem-
onstrated. On a worldwide basis, the International Organization for Standardization
(ISO) publishes standards, to ensure that “products are fit for their purpose.” In the
USA, the American Society for Testing and Materials (ASTM) develops and pub-
lishes the standards. The Food and Drug Administration (FDA) determines whether
new medical devices can be released to the market based on passing ASTM standards
which have been selected by the FDA. The work of developing standards is most
often carried out by technical staff from companies and faculty members from uni-
versities on a voluntary basis. For artificial knees and the associated materials and
instrumentation, the F04 committee on Medical and Surgical Materials and Devices
is responsible. Even though most of the applications to the FDA are of devices which
are similar to others already on the market, there are often sufficient differences that
it is a time-consuming process to complete all of the testing which is necessary and
for the FDA to carry out the evaluations. In addition, devices frequently include new
design features or improved materials, which adds to the testing which needs to be
carried out. Considering the importance of testing devices such as artificial knees,
and for advancing the field as expeditiously as possible, it would be an advantage if
more resources were available for developing the methodologies for new standards
that might be deemed necessary. Advanced simulator studies and computer modeling
are among the methods that could be further developed.
In the case of artificial knees, test methods and performance standards are avail-
able for material quality, wear of the bearing surfaces, fatigue strength of the com-
ponents, strength of modular connections, and constraint during shear and torque
loading. By the time a new artificial knee has successfully passed the various tests,
it is unlikely that the device will fail in vivo for mechanical reasons. Such testing
should be regarded as part of the design process in that an unsatisfactory result of
any test should require a design modification. Nevertheless, failures still do occur
in vivo due to the design itself, to the surgical technique, or related to the patient.
For example, excessive instability can occur if the implant has relatively low inher-
ent stability and if the ligaments have not been adequately balanced at surgery and
there is weak musculature. Bone collapse or component loosening can occur if the
component alignment is incorrect by more than a few degrees and the trabecular
bone has been weakened due to the arthritis. In the long term, if the wear is
excessive due to a high activity level in a heavy patient, the accumulation of wear
particles can cause pain, swelling, and loose components.
How Were Designs Tested for Wear, Strength, and Kinematics?
In the first decade of total knee design, simple test machines were designed primar-
ily for testing laxity and stability, fixation, wear, and strength. Due to the lack of
regulations, by no means was every design evaluated. Even so, by 1980, only about
10 years after the first artificial knees were used in patients, there was a high success
rate with few designs showing poor results. This can be attributed to the gradual
improvement of test methods based on observations of clinical performance, as well
as the high incentive to all parties for successful results. Today, for a total knee to be
used in patients, extensive testing needs to be carried out as specified by the
FDA. The ASTM organization has been the main vehicle for formulating the tests.
The tests have largely been the combined work of biomechanics and biomaterials
specialists over a period of several decades. While the larger companies have suffi-
cient resources to carry out their own testing, smaller companies contract special-
ized testing laboratories. These laboratories have developed considerable expertise
and experience of a wide range of artificial knee designs. The tests will undoubtedly
become more sophisticated and numerous in the future: some may be appropriate as
new standards in that they address safety and efficacy, while others will be a means
for detailed evaluation of performance. Among the types of testing methods which
have a high potential for advancement are wear and strength testing using advanced
simulators and functional analyses using fluoroscopy and computer modeling.
References
Abdelgaied A, Fisher J, Jennings LM. A comparison between electromechanical and pneumatic-
controlled knee simulators for the investigation of wear of total knee replacements. Proc Inst
Mech Eng H J Eng Med. 2017;231(7):643–51.
Bergmann G, Bender A, Graichen F, et al. Standardized loads acting in knee implants. PLoS One.
2014;9(1):e86035.
Branson PJ, Steege JW, Wixson RL, Lewis J, Stulberg SD. Rigidity of initial fixation with unce-
mented tibial knee implants. J Arthroplast. 1989;4(1):21–6.
Brooks PJ, Walker PS, Scott RD. Tibial component fixation in deficient tibial bone stock. Clin
Colley J, Cameron HU, Freeman MA, Swanson SA. Loosening of the femoral component in sur-
face replacement of the knee. Arch Orthop Trauma Surg. 1978;92(1):31–4.
Currier BH, Currier JH, Franklin KJ, Mayor MB, Reinitz SD, Van Citters DW. Comparison of
wear and oxidation in retrieved conventional and highly cross-linked UHMWPE tibial inserts.
J Arthroplast. 2015;30(12):2349–53.
DesJardins JD, Walker PS, Haider H, Perry J. The use of a force-controlled dynamic knee simula-
tor to quantify the mechanical performance of total knee replacement designs during functional
activity. J Biomech. 2000;33(10):1231–42.
Dorr LD, Lindberg JP, Claude-Faugere M, Malluche HH. Factors influencing the intrusion of
methylmethacrylate into human tibiae. Clin Orthop Relat Res. 1984;183:147–52.
Ducheyne P, Kagan A 2nd, Lacey JA. Failure of total knee arthroplasty due to loosening and defor-
mation of the tibial component. J Bone Joint Surg Am. 1978;60(3):384–91.
Ezzet KA, Hershey AL, D’Lima DD, Irby SE, Kaufman KR, Colwell CW. Patella tracking in total
knee arthroplasty. J Arthroplasty. 2001;16(7):838–43.
Fitzpatrick CK, Maag C, Clary CW, Metcalfe A, Langhorn J, Rullkoetter PJ. Validation of a new com-
putational 6-DOF knee simulator during dynamic activities. J Biomech. 2016;49(14):3177–84.
Freeman MA, Bradley GW, Blaha JD, Insler HP. Cementless fixation of the tibial component for
the ICLH knee. J R Soc Med. 1982;75(6):418–24.
Fregly BJ, Besier TF, Lloyd DG, et al. Grand challenge competition to predict in vivo knee loads.
J Orthop Res. 2012;30(4):503–13.
Goodfellow J, Hungerford DS, Zindel M. Patello-femoral joint mechanics and pathology. 1.
Functional anatomy of the patello-femoral joint. J Bone Joint Surg. 1976;58(3):287–90.
Goodman SB. The effects of micromotion and particulate materials on tissue differentiation: bone
chamber studies in rabbits. Acta Orthopaedica Scandinavica. 1994;65(sup258):1–43.
Grupp TM, Fritz B, Kutzner I, Schilling C, Bergmann G, Schwiesau J. Vitamin E stabilised poly-
ethylene for total knee arthroplasty evaluated under highly demanding activities wear simula-
tion. Acta Biomater. 2017;48:415–22.
Bone Joint Surg. 1971;53(2):272–7.
Haider H. Tribological assessment of UHMWPE in the knee, Chapter 32. In: Kurtz SM, editor.
UHMWPE biomaterials handbook – ultra-high molecular weight polyethylene in total joint
replacement and medical devices. 3rd ed: New York, USA: Elsevier Inc; 2016. p. 559–634.
Haider H, Garvin K. Rotating platform versus fixed-bearing total knees: an in vitro study of wear.
Clin Orthop Relat Res. 2008;466(11):2677–85.
Haider H, Kaddick C. Wear of mobile bearing knees: is it necessarily less? J ASTM Int.
2012;9(2):1–10.
Haider H, Weisenburger JN, Kurtz SM, et al. Does vitamin E-stabilized ultrahigh-molecular-
weight polyethylene address concerns of cross-linked polyethylene in total knee arthroplasty?
J Arthroplast. 2012;27(3):461–9.
Harding ML, Harding L, Goodfellow JW. A preliminary report of a simple rig to aid study of the
functional anatomy of the cadaver human knee joint. J Biomech. 1977;10(8):517–23.
Heim CS, Postak PD, Plaxton NA, Greenwald AS. Classification of mobile-bearing knee designs:
mobility and constraint. J Bone Joint Surg Am. 2001;83-A Suppl 2(Pt 1):32–7.
Koh YG, Lee JA, Chung PK, Kang KT. Computational analysis of customized cruciate retain-
ing total knee arthroplasty restoration of native knee joint biomechanics. Artif Organs.
2019;43(5):504–14.
Krause WR, Krug W, Miller J. Strength of the cement-bone interface. Clin Orthop Rel Res.
1982;163:290–9.
Li G, DeFrate LE, Zayontz S, Park SE, Gill TJ. The effect of tibiofemoral joint kinematics on
patellofemoral contact pressures under simulated muscle loads. J Orthop Res. 2004;22:801–6.
Lin, Judy. 2016. UCLA researchers team up with robot for solutions to debilitating knee injuries
http://newsroom.ucla.edu/stories/ucla-researchers-team-up-with-robot-for-solutions-to-debil-
itating-knee-injuries
Maletsky LP, Hillberry BM. Simulating dynamic activities using a 5-axis simulator. J Biomech
Eng. 2005;127:123–33.
Markolf KL, Finerman GM, Amstutz HC. In vitro measurements of knee stability after bicondylar
replacement. J Bone Joint Surg Am. 1979;61(4):547–57.
Marra MA, Andersen MS, Damsgaard M, Koopman BFJM, Janssen D, Verdonschot N. Evaluation
of a surrogate contact model in force-dependent kinematic simulations of Total knee replace-
ment. J Biomech Eng. 2017;1:139(8).
McCutchen CW. The frictional properties of animal joints. Wear. 1962;5(1):1–17.
1970;3(1):51–61.
Morrison J. Bioengineering analysis of force actions transmitted by the knee joint. Biomed Eng.
1968;3:164–70.
Nogi J, Caldwell JW, Kauzlarich JJ, Thompson RC. Load testing of geometric and polycentric
total knee replacements. Clin Orthop Relat Res. 1976;114:235–42.
Oral E, Muratoglu OK. Vitamin E diffused, highly crosslinked UHMWPE: a review. Int Orthop.
2011;35(2):215–23.
Postak PD, Heim CS, Greenwald AS. The role of constraint in contemporary modular knee
designs. Surg Technol Int. 1994;3:541–6.
Rimnac C, Pruitt L. How do material properties influence wear and fracture mechanisms? J Am
Acad Orthop Surg. 2008;16:S94–S100.
Rowell SL, Reyes CR, Malchau H, Muratoglu OK. In vivo oxidative stability changes of highly cross-
linked polyethylene bearings: an ex vivo investigation. J Arthroplast. 2015;30(10):1828–34.
Rullkoetter PJ, Fitzpatrick CK, Clary CW. How can we use computational modeling to improve
Total knee arthroplasty? modeling stability and mobility in the implanted knee. J Am Acad
Orthop Surg. 2017;25(Suppl 1):S33–s39.
Schwiesau J, Schilling C, Kaddick C, et al. Definition and evaluation of testing scenarios for
knee wear simulation under conditions of highly demanding daily activities. Med Eng Phys.
2013;35(5):591–600.
Seedhom B, Dowson D, Wright V. Wear of solid phase formed high density polyethylene in rela-
tion to the life of artificial hips and knees. Wear. 1973;24(1):35–51.
Shaw J, Murray D. Knee joint simulator. Clin Orthop Rel Res. 1973;94:15–23.
Singerman R, Berilla J, Davy DT. Direct in vitro determination of the patellofemoral contact force
for normal knees. J Biomech Eng. 1995;117:9–14.
Steiger RN, Muratoglu O, Lorimer M, Cuthbert AR, Graves SE. Lower prosthesis-specific 10-year
revision rate with crosslinked than with non-crosslinked polyethylene in primary total knee
arthroplasty: 386,104 procedures from the Australian Orthopaedic Association National Joint
Replacement Registry. Acta Orthop. 2015;86(6):721–7.
Swanson SAV, Freeman MAR, Heath JC. Laboratory tests on total joint replacement prostheses. J
Bone J Surg. 1973;55B(4):759–73.
Taylor M, Bryan R, Galloway F. Accounting for patient variability in finite element analy-
sis of the intact and implanted hip and knee: a review. Int J Numer Methods Biomed Eng.
2013;29(2):273–92.
Van Citters DW, Currier JH, Currier BH, Titus AJ, Major MB, Abdul MP, Berry DJ. A retrieval
analysis of 1385 tibial inserts for wear, damage and oxidation: What factors are driving per-
formance? Scientific Exhibit SE 10: American Academy of Orthopaedic Surgeons, Annual
Meeting March 2018, Las Vegas, NV.
Varadarajan KM, Harry RE, Johnson T, Li G. Can in vitro systems capture the characteristic differ-
ences between the flexion-extension kinematics of the healthy and TKA knee? Med Eng Phys.
2009;31(8):899–906.
Volz RG, Nisbet JK, Lee RW, McMurtry MG. The mechanical stability of various noncemented
tibial components. Clin Orthop Relat Res. 1988;226:38–42.
Walker PS, Blunn GW, Broome DR, et al. A knee simulating machine for performance evaluation
of total knee replacements. J Biomech. 1997;30(1):83–9.
Walker PS, Blunn GW, Perry JP, et al. Methodology for long-term wear testing of total knee
replacements. Clin Orthop Relat Res. 2000;372:290–301.
Walker PS, Greene D, Reilly D, Thatcher J, Ben-Dov M, Ewald FC. Fixation of tibial components
of knee prostheses. J Bone Joint Surg Am. 1981;63(2):258–67.
1977;59(2):222–8.
Walker PS, Wang C-J, Masse Y. Joint laxity as a design criterion for the design of condylar knee
prostheses. Conference on total knee replacement, Institution of Mechanical Engineers,
London UK, 16–18 Sept 1974.
White BF, D’Lima D, Drueding AC, Cox J, Carignan FJ. A simulator study of TKR kinemat-
ics using modeled soft-tissue constraint: virtual soft-tissue control for knee simulation. In:
Wear of Articulating Surfaces: Understanding Joint Simulation. West Conshohocken: ASTM
International; 2006.
Whiteside LA, Pafford J. Load transfer characteristics of a noncemented total knee arthroplasty.
Clin orthop rel res. 1989;239:168–77.
Willing R, Walker PS. Measuring the sensitivity of total knee replacement kinematics and laxity to
soft tissue imbalances. J Biomech. 2018;77:62–8.
Zavatsky AB. A kinematic-freedom analysis of a flexed knee stance testing rig. J Biomech.
1997;30(3):277–80.
Evaluations of Designs and Techniques
Which Evaluation Methods Are Applicable,
14
and Are There Differences Between Designs
and Techniques?
Function Measured Using PROMs
According to the Swedish Knee Arthroplasty Register, the survivorship of artificial

knees is 92% at 20 years. Because that cannot be improved too much further, there
is an increasing focus on the functional performance. From the patient’s perspec-
tive, apart from a pain-free knee that will last indefinitely, it is not unreasonable to
expect a return to the functional level before they developed arthritis. Television
advertisements for artificial knees, where the patients are seen running, high-speed
cycling, and skiing, do not apply to everybody. However there is the suggestion that
when the patient is able to perform those sorts of activities, the artificial knee could
safely endure the forces and motions. Concurrent with the introduction of artificial
knees, evaluation schemes were devised to assess the function. Such schemes were
mostly specific to each investigator or hospital. In evaluating the Polycentric and
Geometric knees in the early 1970s, Ilstrup, Coventry, and Skolnick from the Mayo
Clinic took a statistical approach to determine the extent to which the patients had
improved on their preoperative status in several categories including function.
Gshwendt from Zurich devised a chart to show the change from pre-op to post-op,
giving a visual impression of how many patients improved and by how much. In the
late 1970s, the Hospital for Special Surgery (HSS) Knee Score was introduced,
which was based on pain, function, range of motion, muscle strength, flexion defor-
mity, and instability, with subtractions for the use of walking aids. It was not until
1989 that the Knee Society Scoring Systems were formulated for both function and
radiographic assessments (Ewald 1989). The scheme recognized that the knee itself
should be firstly assessed objectively, by measuring alignments and quality of the
interface. Secondly there should be an evaluation of function which depended on
many other factors. One of the first studies which attempted to account for the range
of activities pertinent to each individual patient was carried out by Philip Noble
Ph.D. and associates (Weiss et al. 2002). They first pointed out that “the goals of
individual patients will differ with respect to postoperative function and activity.”

266 14 Evaluations of Designs and Techniques
Fig. 14.1 Philip Noble

completed his Ph.D. at
Strathclyde University,
Glasgow, Scotland, and
later worked at Baylor
College of Medicine and
Methodist Hospital,
Houston, TX. Working on
a Committee of the Knee
Society, he advanced the
methodology for
evaluation of patients with
artificial knees by focusing
on the preferred activities
of each patient and
including the ability to
perform these activities in
the new Knee Society
Scoring System. He also
devised innovative
methodologies for tibial
fixation and for evaluating
surgical performance.
(Photo taken 2015) (©
Peter S Walker)
This led them to formulate a questionnaire which divided activities into activities of
daily living, movement and lifestyle, and exercise and sports. The patients were
asked the importance and frequency of these activities and whether the knee either
limited them or bothered them while carrying out the activities. They concluded that
“although knee replacement does not prevent patients from doing activities they
consider important, the procedure, nonetheless, does not restore the ease associated
with normal knee function.” The authors also recognized that “Additional work also
is needed to compare the responses reported in this study with those of healthy con-
trol subjects…” In other words, the patients should be realistic about what is achiev-
able in their particular case. Nevertheless, the study did imply that the artificial
knees included in their study were not providing the function and comfort level that
the patients would have if they had not developed arthritis. Noble’s work then
became the starting point for an outcomes subcommittee in The Knee Society,
chaired by Giles Scuderi MD, to develop a questionnaire for widespread use in
evaluating artificial knees. The result was the Knee Society Scoring System (KSS)
(Scuderi et al. 2012). The recorded metrics include both surgeon-reported measures
and patient-reported satisfaction and function. After several years of experience
with the system, there was a need for a shortened version to improve compliance
from patients who otherwise did not have the patience or time to complete the long
form (Scuderi et al. 2016).
KNEE SOCIETY SHORT FORM (SECOND HALF) FJS – 12. FORGOTTEN JOINT SCORE
STANDARD ACTIVITIES The following 12 questions refer to how aware you are of your
cannot do
How much does your knee no very artificial knee joint in everyday life.
(because I never
bother you during each of the bother slight moderate severe severe of knee)
following activities? 5 4 3 2 do this
1 0 Please tick one answer from each question.
13- Walking on an uneven
surface
14- Climbing up or decending
a flight of stairs
15- Getting up from a low couch

Are you aware of your artificial joint...
or a chair without arms
1. ... in bed at night?
16- Running
never almost never seldom sometimes mostly

DISCRETIONARY KNEE ACTIVITIES
2 . ... when you are sitting on a chair for more than 1 hour?
Many people consider the following activities important. Of thesse activities, which
one is the most important to you? 3 . ... when you are walking for more than 15 minutes?
(please do not write in additional activities)
4 . ... when you are taking a bath/shower?

14 Evaluations of Designs and Techniques
Swimming Weight-lifting
Golfing (18 holes) Leg Extensions 5 . ... when you are traveling in a car?
Road Cycling (>30mins) Stair-Climber
Gardening Stationary Biking / Spinning 6 . ... when you are climbing stairs?
Bowling Leg Press
Racquet Sports (Tennis, Racquetball, etc.) Jogging
7 . ... when you are walking on uneven ground?
Distance Walking Eliptical Trainer
8 . ... when you are standing up from a low-sitting position?
Dancing / Ballet Aerobic Exercises
Stretching Exercises (stretching out your muscles) 9 . ... when you are standing for long periods of time?
10 . ... when you are doing housework or gardening?
How much does your knee bother you during the activity checked above?
cannot do 11 . ... when you are taking a walk/hiking?
no bother slight moderate severe very severe (because of knee)
5 4 3 2 1 0 12 . ... when you are doing your favorite sport?
A high score 100) indicates a high degree of "forgetting” the
artificial joint, that is, a low degree of awareness.
Fig. 14.2 There is a similarity between the Knee Society Short Form (KSS) and the Forgotten Joint Score (FJS). The KSS asks “does your knee bother you in
activities?”; the FJS asks if you are aware of your artificial joint in activities. However the KSS allows the patient to choose certain activities, while the FJS
267
provides a list which almost every patient will be able to do. (Left, reprinted with permission The Knee Society; right, reprinted with permission from Journal
of Arthroplasty)
Fig. 14.3 Giles Scuderi

MD was a knee fellow
with John Insall at the
Surgery, New York, in
1987. Since that time he
has been a major educator
in multiple forums on the
principles of technique of
total knee replacement,
particular those associated
with the PS type of knee.
In 2012 Gil Scuderi and
other members of The
Knee Society formulated
the now widely used
evaluation method for total
knee outcome, the Knee
Society Knee Scoring
System. (Photo c 2017)
(© Peter S Walker)
Today the KSS is but one example of a patient-reported outcome measure

(PROM), the most widely used method of evaluating artificial knees. Such measures
are simple to carry out, require little time of the physician, include an assessment of
function, and produce a numerical score which can be compared with those of other
study groups. Typically, a patient checking in to the hospital for an artificial knee
will be asked to complete a PROM form on a tablet. Recently a survey was carried
out of clinical follow-up studies published in the major orthopedic journals over the
past 15 years (Siljander et al. 2018). A previous article was quoted as to what con-
stituted an effective PROM: “…one that adequately measures clinical results, while
also being valid, reliable, responsive and easily administered.” Siljander et al. found
that no less than 42 individual PROMs had been used in previous studies. The Knee
Society Score (KSS) was used the most frequently for total knees, but for unicom-
partmental knees, the Oxford Knee Score (OKS) was mostly used, due to the influ-
ence of the originators of the Oxford Knee from the Nuffield Hospital in Oxford,
England. Further comments by the authors included the following: “There are many
facets to a PROM, and the balance between subjective and objective scores, patient
14 Evaluations of Designs and Techniques 269
satisfaction, and factors important to the patient vs the surgeons, leave the combina-
tion endless… Furthermore an explicit definition of a successful outcome has yet to
be determined across all the PROMs.”
It is important to define exactly what is being measured in a PROM. As an exam-
ple of the type of questions, the section of the Knee Society Scoring System (KSS)
relating to function, the patient is asked: “how much does your knee bother you
during each of the following activities?” covering every day to more strenuous
activities (Scuderi et al. 2012). Similarly, in the Knee Injury and Osteoarthritis
Outcome Score (KOOS) scoring system for function, another widely used system,
the patient is asked for their experience with numerous everyday activities including
recreational: “what degree of difficulty have you experienced during the last week
due to your knee?” (Roos et al. 1998). The Forgotten Joint Score (FJS) was defined
recently to provide a simplified scoring system which focused on what really mat-
tered to the patient (Behrend et al. 2012). “The ability to forget the artificial joint in
everyday life can be regarded as the ultimate goal in joint arthroplasty.” The patient
is asked: “are you aware of your artificial knee when you are walking for more than
15 minutes?” There are 12 such questions covering common activities of everyday
living. Such an approach was embodied in Noble’s questionnaire cited above (Weiss
et al. 2002) in the following question: “does your knee feel normal to you?” PROMs
such as the above have been widely used to evaluate the outcomes of groups of
patients, with our emphasis being on the functional outcome. The PROMs can read-
ily be used with cohorts of patients to evaluate a new artificial knee design, or to
compare one design with another.
In this chapter we will concentrate on the comparisons between designs and also
techniques. Given the wide variety of configurations of artificial knees described in
previous chapters, it might be expected that some designs would function much bet-
ter than others and that it would be possible to compile a league table of functional
outcomes. Even particular types of design such as a CR or PS made by different
manufacturers might be expected to show some differences in function. There are
many examples where a manufacturer has a successful knee system, but over the
years, a number of improvements are identified. At that point, a new knee system is
designed with the improved features. A PROM could be used to determine if the
new system produced higher functional scores than the previous system. Such com-
parisons are now discussed, where PROMs have been the principal method of evalu-
ating the function.
A key question posed in the Introduction was whether any of the newer designs
have higher functional scores compared with the cruciate-retaining (CR) and
posterior-stabilized (PS) designs. To answer this is not easy for several reasons.
There are a large number of studies which have been carried out since 1980. Many
of the studies are not comparative, assessing only one particular design. The studies
have covered many designs, including different versions of each design, whether
CR, PS, or other design types. The evaluation methods have differed, whether the
KSS system or one of the other 42 systems (and still more are being added). The
conditions of the studies have been different, such as the surgical technique, the
rehabilitation regimen, and the length of follow-up. For our purpose, we will pref-
erentially use recent studies, with meta-analyses or reviews being favored, with
some attention to the size of the cohorts. Efforts have been made to choose a repre-
sentative sample of the studies, especially where there is a disparity in the results.
Studies which make useful insights or observations are also selected.
Overall Differences Between Designs
In a unique study comparing different knee types, Pritchett (2011) implanted

bilateral knees in 492 patients, with different types in opposite knees. At 2-year
follow-up, the patients were asked: “which is your better knee overall?” This eval-
uation method is similar to the Forgotten Joint Score but encapsulates everything
into one simple question. The patient had to consider all of their activities com-
bined in giving the answer. Comparing one knee with the other would be easier
and more reliable than trying to objectively assess a single knee. The bilateral case
study avoided the problem of having two separate groups, each with an artificial
knee on one side only.
In Pritchett’s study, ACL + PCL preserving (BCS) and medial pivot knees were
preferred over standard CR and PS knees. ACL + PCL and medial pivot knees were
preferred equally, as were CR and PS. Pritchett offered some explanation of his
findings: “…if the implant allows the femur to move forward on the tibia (as occurs
with CR and PS) the patient may sense this as an instability…conversely if the
prosthesis provides AP stability such as with the ACL + PCL and Medial Pivot
prostheses, patients may sense a feeling of stability…in addition patients may pre-
fer ACL + PCL prostheses because of superior proprioception.” In this study,
Pritchett chose to use a simple test, namely, asking the patients which knee they
preferred, to which the patient could provide an assessment covering all
conditions.
Unicompartmental vs Total Knee
Unicompartmental knees are expected to show higher functional scores than total
knees, based on the preservation of most of the joint except for the medial condyles,
and the reduced tissue removed during surgery. In a recent meta-analysis, where
most of the publications were since the year 2000, from the 23 selected studies,
unicompartmental knees (3254 cases) had a higher range of flexion and higher KSS
and WOMAC scores, as well as a quicker return to sports, compared with total
knees (10,640 cases) (Migliorini et al. 2019). One of the conclusions was “These
results suggest that the UKA implants are more suitable for physically active
patients, since they allow better clinical and functional performances.” Another
recent survey (Pfitzner et al. 2017) showed that while unicompartmental knees did
have higher PROM scores in follow-up studies, because their preoperative scores
UC
CR
PS
CR
PS narrow
narrow
Fig. 14.4 A state-of-the-art knee system consisting of PS (posterior stabilized), CR (posterior

cruciate retaining), UC (ultracongruent), and MC (medially congruent). The sizing schemes were
based on an analysis of a large CT database, covering gender and race. The PS and CR are in stan-
dard and narrow. Comparative outcomes between designs are still being evaluated. (Picture cour-
tesy ZimmerBiomet, Warsaw, IN)
were higher than for total knees, the gains in PROM scores were similar between the
two groups. Lyons et al. (2012) reviewed their unicompartmental (219 cases) and
total knee (5606 cases) from 1978 to 2009. Their conclusion was “consistent with
the literature, we found patients with UKA had higher absolute clinical outcome
scores for function preoperatively and at follow-up than patients with TKA. Change
in clinical outcome scores were similar in both groups.” They emphasized that it
was the increase in pre-op to post-op scores that was important; thus, the two groups
were similar in outcome. They did surmise however that there was a ceiling effect
with some of the evaluation methods. In three other recent studies, the range of
flexion was higher after a unicompartmental compared with a total knee, but the
functional scores were similar (Hauer et al. 2020; Goh et al. 2018; Kleeblad et al.
2018). Van der List et al. (2017) found that the WOMAC scores were higher for a
medial unicompartmental knee with a metal-backed tibial component, compared
with a total knee. The point about the metal-backing was that for all-plastic inlay
components, deformation and loosening occurred in some cases, causing discom-
fort and reducing the scores; the rigidity of metal-backing avoided these problems.
One unanswered question is that if a particular arthritic knee was treated with a
unicompartmental or a total knee, would the outcomes be different?
Cruciate-Retaining (CR) vs Posterior-Stabilized (PS) Knees
It is of particular interest to compare the two most widely used types of artificial
knee, the CR and the PS, with the expectation that there would be a difference
because of the differences in mechanics. However comparing the WOMAC scores
in 1287 cases in a regional data base, there were no differences in functional scores
at 5- and 10-year follow-up (Singleton et al. 2019). Also, Serna-Berna et al. (2018)
found that in a study of over 200 patients in each group of CR and PS designs, at
10-year follow-up, there were “no significant differences in functional scores, range
of motion or patient satisfaction. The superiority of one design over the other was
not found.” In comparing survivorship of over 133,000 cases in the Dutch
Arthroplasty Register, it was concluded that PS knees were 1.5 times more likely to
be revised at 8 years than CR knees (Spekenbrink-Spooren et al. 2018). The revision
rates were the highest in the younger male group. In a meta-analysis of studies up to
2015, 14 randomized control trials met certain criteria (Jiang et al. 2016). There
were 791 CR knees and 622 PS knees, with at least 6-month follow-up. “There were
no significant differences in KSS, KSPS, HSS scores… and complications. However
the PS provided better total ROM, improvement of ROM and flexion angle.” The
differences in the flexion angles were 5–10° in the different studies, which can be
attributed to the posterior femoral rollback after the cam-post engagement. Beaupre
et al. (2017) carried out a randomized study of CR vs PS followed up for 10 years.
There were no differences in the pain and function scores, nor the complications,
between the two groups. The authors found that there was a gradual reduction in
functional scores over time, a function of the aging of the patients rather than related
to the artificial knees. This is undoubtedly a factor when analyzing typical artificial
age groups where the average age is in the late 60's at the time of the proce-
dure. Finally, in one study it was claimed that a deep-dish tibial insert with a high
anterior lip provided a greater pre-op to post-op range of motion than a PS knee,
although the mean difference between the designs was only about 3° (Berend et al.
2013). At the same time, if the deep-dish design had advantages over the PS, such
as less femoral bone resection and improved patella tracking, the deep-dish insert
would not be a disadvantage, at least in flexion range.
In a recent comparative review of CR and PS, Meneghini et al. (2019) com-
bined PROMs and registry data, to conclude the following: “Equivalent functional
outcomes in multiple studies comparing CR-type bearings and PS TKA designs,
preclude the need for PS TKA designs in the modern health care environment.
Based on the available data and modern metrics, traditional PS TKA designs with
a cam-and-post articulation are not required for the majority of primary TKA.”
The authors also pointed out that CR knees with steep anterior tibial surfaces
substitute for the posterior cruciate ligament, which brings the CR into the realm
of an ultracongruent type where both cruciates are resected. A limitation of this
statement is that all CR designs are not likely to be equal, nor all PS designs, nor
all ultracongruents.
Rotating Platform Designs
Recently the Forgotten Joint Score was used to compare two rotating platform knee
systems, the classic LCS designed by Buechel and Pappas and a recent derivative,
the Attune (Behrend et al. 2019). In finding no difference in the Forgotten Joint
Scores, as well as WOMAC scores and range of motion, the authors concluded:
“The clinical relevance of this study is that it questions the importance of implant
design as the single most important factor for patient outcomes.” Nobody would
question that the type of artificial knee is but one of a number of factors which affect
the PROM score, no matter which PROM is being used. But in the example above,
the two rotating platform designs were so similar in their configuration and mechan-
ics that it is doubtful if any PROM evaluation method could distinguish between the
two designs.
Powell et al. (2018) compared the rotating platform and fixed-bearing designs of
the press-fit condylar (PFC) knee using several PROM scores. Their conclusion was
“In contrast with the established literature consensus… mobile bearing total knee
arthroplasties are superior to identical implants with fixed-bearing inserts…” They
attributed their finding to the avoidance of excessive internal rotational position of
the fixed-bearing version which evidently caused patella pain and mal-tracking.
However the superiority occurred in only some of the PROM evaluations, but not in
others. Bailey et al. (2015) reported a randomized study over four centers compar-
ing exactly the same two knee designs. At 1–2 years, there were no differences in
any clinical scores including range of flexion, KSS, Oxford, and SF-12. Again using
the same designs, although mixing between CR and PS versions, Harrington et al.
(2009) found almost identical PROM scores between rotating platform and fixed
bearing.
After many years of experience with fixed-bearing and rotating platform designs,
a review of their own cases, and of literature reports, White et al. (2015) concluded:
“Overall, both kinematic and in-vitro wear studies have highlighted certain benefits
of rotating-platforms over fixed-bearing knees in primary TKA. However, the
improved axial rotation has failed to show any benefits with respect to clinical out-
comes…” This illustrates that there may be advantages of certain designs, in this
case patella tracking or rotational placement, that do not alter the PROM score, yet
are worthwhile advantages.
Medial Pivot Designs
A medial pivot type of design might be expected to show better function over a
symmetric design, considering the asymmetry of anatomic knee motion where
the medial side has much less anterior-posterior laxity than the lateral side. An
advantage was reported by Pritchett (2011) as described in the above section
entitled “Overall Differences Between Designs.” Has the advantage been
demonstrated in subsequent studies using PROMs as the evaluation tool? In a

comparison between 76 medial pivot knees and 88 PS knees at 1 year, the medial
pivot showed a significantly higher Forgotten Joint Score (FJS), especially in
rising from a seated position (Samy et al. 2018). The authors attributed this to
the highly congruent medial side. They found no differences in range of flexion,
but no other functional evaluations were carried out. A meta-analysis between
medial stabilized (more dished medial side, less dished lateral side) and non-
medial stabilized (symmetric) showed the same 90-point Knee Society Score in
both groups (Young et al. 2018). The authors stated “…a firm conclusion cannot
be reached regarding the comparative clinical performance…” Comparing a
medial pivot design (SAIPH) with a fixed-bearing symmetric design (Triathlon),
there were no significant differences in KSS or Oxford scores (Benjamin et al.
2018). They raised the question as to why that should be the case, considering
the difference in mechanics between the two designs: “The medial pivot design
provides anteroposterior stability throughout the entire range of motion.
However, it is unclear whether this difference in design translates to clinical and
functional improvement.” From a large regional database, the 10-year survivor-
ship of the medial pivot knee (Advance) at 96.3% was not statistically different
from the other knees in the database, mainly CR and PS, at 95.7% (Bordini et al.
2016). With 150 knees in each group, the Advance medial pivot was compared
with a PFC PS knee. There were no differences in the KSS and WOMAC scores
or in the range of flexion. The flexion results were impressive at a 9° gain from
preoperative. Comparing 261 medial pivot knees with 913 PS knees, there was
no difference in range of flexion, or gain in flexion, at 1-year follow-up
(Shakespeare et al. 2006). In 98 patients, bilateral cases were analyzed, an
Advance medial pivot on one side and a PFC Sigma rotating platform on the
other side (Kim et al. 2017). At 2-year follow-up, the rotating platform results
were better than for the medial pivot in the KSS scores and the HSS scores. The
range of flexion, starting at the same average preoperatively, was higher postop-
eratively by 12°. There was no explanation given other than the excessive con-
straint on the medial side. However that does not seem valid because the rotating
platform knee has exactly the same anterior-posterior constraint but in a direc-
tion along the center of the tibia, rather than along a line on the medial side.
Katchky et al. (2019), a group of surgeons who were part of the initial design
team for the Saiph medial pivot knee, together with Michael Tuke, reported on
their 5-year follow-up of 81 cases. This was part of the “stepwise introduction” of
any new implant, recommended by Henrik Malchau. The average range of flexion
was 124°, at the high end of studies of other designs. The conclusion from the
study was “Five-year surveillance following limited release demonstrates the
safety and efficacy of the SAIPH medial ball and socket total knee arthroplasty.
Medial ball and socket kinematics facilitate ligament balancing, while limiting
the need for additional soft-tissue release.”
Guided Motion Knees
hile we have considered the medial pivot and rotating platform designs as ‘Guided
W
Motion Knees’, those guided motion types which are more biomimetic, have been
little used so far. There are two examples which can given at this time. The Aequos
G1 design which had a convex lateral tibial surface was tested in 158 patients fol-
lowed for an average of 1 year (Frosch et al. 2009). Despite the low lateral confor-
mity, the average flexion range was only 107 degrees. The Journey 2 knee with a
low conformity lateral side, was followed in a multi-center study in no less than
2000 cases (Harris et al. 2019). The results were similar to those of other more con-
ventional designs as recorded in the Australian National Joint Replacement Registry.
However, in neither of these two examples were any biomechanical functional eval-
uation methods used for analysis.
Surgical Techniques
Using PROMs as the main indicator, there have been inconclusive results concern-
ing the relative merits of different surgical techniques based on a systematic review
(Jaffe et al. 2018). With the benchmark of reproducing the mechanical axis in the
frontal plane, cases with greater than 2° error had lower KSS functional scores. On
this same topic, another author concluded: “…good alignment allows for improved
kinematics, stability and range of motion” (Longstaff et al. 2009). However using
the mechanical axis as the target has been challenged from two points of view.
Firstly, there is a variation in the location of the mechanical axis (line from the cen-
ter of the femoral head to the center of the ankle) with many male patients having a
constitutional varus. Furthermore, there is an inclination of the joint line of 2–3° to
the mechanical axis. Following these guidelines in anatomic or kinematic alignment
results in variations in the coronal alignment to the mechanical axis, but this made
no difference to the Oxford or WOMAC functional scores (Howell et al. 2018). But
according to Jaffe et al., “…long-term studies with larger numbers are required
before making any conclusions…” On computer navigation, Jaffe et al. concluded
that while there was ample evidence for more consistent alignment, there was no
evidence of improved KSS scores. Daines and Dennis (2014) reviewed their evi-
dence from two studies for the gap balancing technique versus measured resection.
The first study demonstrated that a balanced and rectangular flexion gap during
surgery was obtained much more frequently with a gap balanced technique. The
second study, using fluoroscopy, demonstrated a much high incidence of femoral
condylar lift-off for measured resection. It was concluded that gap balancing pro-
vided better frontal plane varus-valgus stability. However in these studies, PROM
data was not provided.
verview of Comparisons Using PROMs

O
The overall picture is that there are some differences in the PROM outcomes between
different designs and surgical techniques, but that in most of the studies, there were
no differences. That is surprising considering the mechanical differences between the
unicompartmentals, the CR, the PS, the rotating platforms, and the medial pivots. If
PROMs do not show functional differences between such extreme designs, it is even
more unlikely that differences will be found between different versions of the same
design type. In the studies which have not shown significant differences in function
between different designs, the reasons could include the following:
• There are indeed no significant differences, so that if a particular patient had

received a different artificial knee or technique, the result would have been the
same. While it cannot be argued that the PROM result was incorrect, a valid
question is was that PROM able to detect differences that would matter
clinically?
• The differences are small so that significance cannot be demonstrated with the
numbers typically used in studies, the studies being effectively underpowered. A
question here is what is the minimal clinically important difference?
• Most of the patients in the study groups are relatively inactive, so that even if a
design was capable of better performance, its capabilities would not be realized.
One way to counter such a result is to restrict the patient study group to active
patients, which is not easy to define, or below a certain age.
• There is a wide range in the condition of the knee preoperatively, and in the gen-
eral condition of the patient, accounting for the wide range of postoperative
scores, so that differences in scores due to the artificial knee design per se cannot
be detected. This again relates to the numbers in the study.
• There is a wide range in the condition of the opposite knee, as well as of the hips,
such that the PROM outcomes are influenced greatly by the joints other than the
artificial knee under study. It is common in follow-up studies that the conditions
of the other joint are not documented.
• The PROMs do not ask the questions which really define performance, in that they
tend to focus on pain and discomfort in performing activities, rather on the level of
performance. This is a plausible explanation and can be countered by focusing on
those parts of the scoring system which deal specifically with function.
• The PROMs have a substantial ceiling effect and are not responsive to changes.
This is a factor which has been studied by Giesinger et al. (2014).
Such limitations in the use of PROMs have been discussed in an article entitled
“The Assessment of Outcome After Total Knee Arthroplasty: Are We There Yet?”
(Hossain et al. 2015). The authors focus on the well-recognized requirement of an
evaluation system that it accounts for a wide range of more demanding activities:
“The lack of depth and breadth in the cover of all aspects of function can lead to
imprecision with a skew in the distribution of scores amongst groups of patients.
The increasingly younger patients undergoing TKA therefore achieve outcome indi-
ces which are near to the ceiling of the measurement scale. At a group level ceiling
effects produce Type-2 errors in testing a hypothesis…” In relation to PROMs as
they are at present, “there may be limitations to their use at a respondent level which
may raise concerns about whether self-reported function adequately reflects actual
function. PROM-based methodology is inherently subjective in nature and in the
context of function, measures patients’ perception of function as opposed to actual
ability.” As an alternative, “there has hence been an increasing trend by many clini-
cians to adopt more objective and exclusive methods of measuring function… The
advantages offered by such methods compared with PROMs, include greater preci-
sion, responsiveness and lack of ceiling effects.” This last comment referred to bio-
mechanical methods of evaluation such as video motion analysis, force plates,
instrumented treadmills, and accelerometer-based gait analysis. There are signs that
such a point of view is gaining more acceptance, where biomechanical measure-
ments are used to distinguish between designs and surgical techniques. The ratio-
nale is that the main role of PROMs is to provide an assessment of the patient’s
function overall, whereas the role of the biomechanical measurements is to focus on
quantitative functional aspects which reflect the specific knee being tested.
Biomechanical measurements should certainly be considered as an alternate to the
use of PROMs alone. If there is evidence of value, biomechanical measurements
can be used more extensively in studies where the mechanical aspects of function
are the main criteria for comparison. Without this approach, newer designs and
techniques which could offer important advantages may be considered ineffective if
evaluated on the basis of PROMs alone. That said, some of the limitations of
PROMs, listed above, will still apply to biomechanical measurements.
lternate Evaluation Methods

A
The following are examples of where alternate methods are applicable. A normal
gait pattern is important not only for appearance but for avoiding abnormal forces
and motions which could injure other parts of the musculoskeletal system, such as
the spine. If a particular artificial knee design was configured to produce normal gait
(a unicompartmental, or an ACL + PCL preserving, or a guided motion knee), then
gait analysis studies with ground-to-foot force data would be a suitable evaluation
method. Another suitable method would be to use mobile fluoroscopy where the
x-ray unit remains centered on the patient’s knee during the activity. The outcome
measure is the movements of the contact points on the lateral and medial condyles,
with the pattern in normal knees being the benchmark.
If a particular design or technique was intended to produce accurate soft tissue
balancing with infrequent need for recutting the bone or performing soft tissue
release, tests at surgery where the soft tissue tensions or the medial and lateral con-
tact forces are measured accurately would be an appropriate experimental method.
A new rehabilitation method might be postulated to result in a more rapid return to
normal function. Simple evaluations such as the Timed Up and Go (TUG) test or the
sit-to-stand (STS) test would be valid methods of measuring the recovery process.
Even remote methods such as activity monitoring using inertial measurement units
(IMU) electronically recorded could be used. An other example is if a new material
to significantly reduce the wear. There is no reason that this would affect the PROM
score, but the initial way of testing the wear is by a comparative study using a knee
simulating machine.
Hence the proposition is that a PROM should not always be regarded as the only
valid measure, rather that other measures can be used which will evaluate some
particular parameter which has its own inherent value, even if that parameter might
not change the PROM score in a major way.
Function Measured Using Biomechanical Methods
As described above, PROMs provide an overall assessment of which activities the

patient can achieve and the extent to which they have difficulties or feel pain during
those activities. However they are not designed to distinguish how well or how often
or how vigorously the activities take place. Several investigators have suggested that
more objective measures should be used in combination with PROMs or even as a
substitute in some cases (Mizner et al. 2011; Dobson et al. 2013; Giesinger et al.
2014; Hossain et al. 2015; Parks et al. 2015; Berliner et al. 2017; Graff et al. 2016).
The most widely used method has been gait analysis. By using this technique, the
joint angles of total knee patients during level walking and going up and down stairs
have been measured, as well as the foot-to-ground forces and the moments acting
across the knee. In some studies, electromyographic data has been used to identify
muscle activity. Such measurements have provided insight into a number of defi-
ciencies in the function of patients with artificial knees. Tom Andriacchi, a
Ascending stairs Descending stairs
Total
Condylar
Geometric
Duo-patella
Gunston
Cloutier
Normal
60 70 80 90 100 60 70 80 90 100
Knee Flexion Range (degrees)
Fig. 14.5 Thomas Andriacchi Ph.D. and colleagues showed that in ascending and descending
stairs, some designs of artificial knee had less range of flexion than normal controls. This was
attributed to abnormal femoral-tibial contact position in the sagittal plane. The ACL+PCL preserv-
ing design with low constraint and soft tissue balancing (Cloutier) had the most normal range.
Patients tended to lean forward during ascending to reduce the flexion moment. In level walking,
all designs had reduced walking speed and reduced mid-flexion angle (Andriacchi et al. 1982)
Fig. 14.6 Thomas
Andriacchi Ph.D. founded
and directed a Kinetics
Laboratory at Rush
Presbyterian Hospital,
Chicago, from 1974 to
1998. He was a major
contributor to the
Miller-Galante knee and
the NexGen CR knee. As a
Professor of mechanical
engineering and
orthopedics at Stanford
University since 1998, he
has carried out
experimental studies which
have given new insights
into the mechanical factors
associated with the
development of
osteoarthritis. He has
developed special footwear
for altering the force
distribution in the knee for
the alleviation of early
arthritis. (Photo c 2017)
biomechanical engineer at Rush Presbyterian in Chicago, and Jorge Galante, chair-

man of orthopedics, were the first to demonstrate some of the deficiencies of artifi-
cial knees in performing activities of daily living (Andriacchi et al. 1980, 1982).
While level walking was not impaired in a major way, this was not the case for
going up and down stairs. Because the center of gravity of the body was more offset
from the center of the knee in the sagittal plane in stair activities, much higher quad-
riceps muscle forces were needed to balance the external moments which were act-
ing to flex the knee. As a result, the patients had to adapt their posture to reduce the
moments. This was especially true if the artificial knee was too constrained such
that the contact points did not move posteriorly with flexion. This led Tom Andriacchi
and Jorge Galante to favor designs of low constraint with cruciate preservation, and
they produced their own design using those principles, the Miller-Galante knee.
Andriacchi also pointed out that abnormal muscle function, learned over the time
period of the patient’s arthritis, as well as weakened muscles, produced abnormal

function.
Results of gait analysis studies up to 2007 where artificial knees of various types
were compared with normal were reviewed by McClelland et al. (2007). It was
found that patients walked at a significantly slower speed than control groups. They
also walked with less total range of knee motion and a reduced range of flexion dur-
ing the loading phase of gait. Patients had less knee flexion during the swing phase
of gait compared with healthy controls. Eighty percent of controls showed a bipha-
sic sagittal knee movement pattern where the knee was flexed during the mid-stance
phase, compared with less than one third of artificial knee patients.
A recent meta-analysis of gait studies between total knees and unicompartmen-
tals, where differences in kinematics would be expected, showed that most biome-
chanical factors were similar but that the artificial knee group had significantly
shorter stride length than unicompartmental patients (Nha et al. 2018). Not a dra-
matic finding, although as Andriacchi had already pointed out, level walking is not
a major discriminator. In addition, in a review by Komnik et al. (2015), the perfor-
mance of unicompartmentals is hampered by the patients maintaining their preop-
erative motion patterns and muscle weakness. In that review, the wide variety of
techniques and parameters used in different gait analysis studies was reported. Most
studies focused on level walking, whereas more demanding activities were needed
to detect kinetic abnormalities. These abnormalities were likely due to impaired
muscle activity, based on studies where gait analysis and electromyography were
studied simultaneously (Catani et al. 2012). The authors concluded: “The present
study supports the evidence that good joint function and nearly physiological mus-
cle activity during gait can be achieved in patients undergoing UKA (unicompart-
mental knee).” For artificial knees, however, the conclusion was different. Aljehani
et al. (2019) demonstrated that in unilateral cases of artificial knee replacement, the
status of the opposite knee, often osteoarthritic to a greater or lesser degree, affected
the overall gait patterns. In fact, the gait pattern of an artificial knee more closely
matched an osteoarthritic opposite knee than a normal opposite knee. This is an
unexpected finding which has major implications, that an artificial knee only par-
tially restores the knee to normal from its preoperative arthritic condition.
Measurements of muscle strength on kinetic test machines have provided infor-
mation about the preoperative and postoperative condition of the total knee patient
(Skoffer et al. 2015). Reduced muscle strength and abnormal neuromuscular patterns
were considered the reasons why patient function after a total knee was deficient
(Yoshida et al. 2013). In studies over many years in the Physical Therapy Department
in Newark, Delaware, Lynn Snyder-Mackler Ph.D. and associates have focused on
the major effect that muscle strength and coordination have on the functional out-
comes after total knee replacement. They concluded that appropriate rehabilitation
programs pre- and postoperatively are effective in improving functional outcomes
(Pozzi et al. 2020). At 1-year follow-up, their conclusion was “on average, TKA does
not restore functional ability to the levels of age-matched subjects without joint
pathology…but…a greater proportion of patients who participated in the progressive
strengthening protocol achieved what could be considered normal clinical and
Fig. 14.7 Clifford
Colwell MD (left), from
the Scripps Clinic, Ja Jolla,
CA, has made major
contributions in research
into the biomechanics of
the knee. He collaborated
with Darryl D’Lima in the
design of implanted
instrumented artificial
knees, carried out surgical
implantations, and
measured forces in
multiple activities,
including sports. This
focused attention on the
possibilities of the artificial
knee to allow an active and
healthy lifestyle as well as
relieve pain and restore
basic functions. (Photo
taken at the ISTA meeting,
San Francisco, 2003)
functional scores.” As well as muscle strength per se, one role of an artificial knee is
to provide the optimal lever arm for the joint to effectively balance external moments
and produce desired motions. Equally important were the rehabilitation protocols.
The most striking example of research of technical and organizational excellence
in recent years was the measurement of forces in the knee by instrumented tibial
components. This was implemented in two centers, the Charite Hospital in Berlin,
Germany, and Scripps Clinic, San Diego. The work in Berlin has been described
previously in relation to the testing of implants for wear (see Chap. 13). The work
at the Scripps Clinic, led by Darryl D’Lima MD and Cliff Colwell MD focused on
the forces during everyday and sporting activities (D’Lima et al. 2011). Peak forces
in walking averaged 2.6 body weight (BW), increasing to 4 BW in jogging.
Surprisingly, similar high forces as well as high shear and torque forces occurred
Fig. 14.8 Darryl D’Lima

MD Ph.D. is director of
orthopedic research at the
Shiley Center for
Orthopaedics, Scripps
Clinic, San Diego. He was
the main designer of the
“Lab-in-a-knee” which
was an instrumented tibial
component for measuring
knee forces in vivo. That
work has resulted in
several awards from major
research organizations.
Other research areas
include soft tissue
balancing and cartilage
injury and regeneration.
(© Peter S Walker)
while performing golf strokes. Stationary cycling produced only 1 BW, indicating
its suitability for those with an arthritic knee. Wearing special variable insole shoes
offloaded the medial side of the knee by up to 19%, potentially valuable in relieving
pain in osteoarthritis. New sensor technology could make it possible to monitor
joint forces on a more routine basis.
Fluoroscopic Evaluation of Kinematics
One evaluation method for revealing the kinematics within the joint itself has
been fluoroscopy, which was first developed at the Massachusetts Institute of
Technology by Scott Banks Ph.D. and Andy Hodge MD (1996). Their purpose
in developing the method was as follows: “Accurate measurements of knee
replacement kinematics during functional activities would provide the basis for
assessing the performance of current designs and for objectively designing
devices with improved kinematics.” The method consisted of taking successive
fluoroscopic images of the knee in motion. Knowing the three-dimensional
geometry of the implant components and the relative motion between the femur
and tibia could be calculated and expressed as three rotations and three
Superior
Im
ag
e
pl
an
e
Lateral (X’,Y’)
Anterior
) (kx,ky)
Y,Z
(X,
Focus
z C pd
Fig. 14.9 Fluoroscopic imaging to determine the kinematics of the knee, often described by the
path of the contact points on the tibial surface or by movement of the transverse femoral axis rela-
tive to the tibia. A radiographic image is produced in the plane of the x-ray plate, from the point
source (focus). After correction for distortions, the orientation of each component in the global
axis frame and the orientation of the femur relative to the tibia are computed (Banks and Hodge
1996). (Reprinted with permission IEEE Transactions in Biomedical Engineering)
Fig. 14.10 Scott Banks

Ph.D. from University of
Florida in Gainesville,
working with Andy Hodge
MD from Massachusetts
General Hospital,
developed the first
fluoroscopic system for
determining the relative
three-dimensional
positions of the femoral
and tibial components,
enabling the tracking of
motion during activity. The
work was first published in
Banks’ Ph.D. thesis at
Massachusetts Institute of
Technology in 1992.
(© Peter S Walker)
Fig. 14.11 Douglas
Dennis MD works at the
Colorado Joint
Replacement Center in
Denver, Colorado, a
specialist high-volume
center in knee and hip
replacement. He holds
academic affiliations with
the universities of
Colorado and Tennessee,
in order to conduct
research on the kinematics
of joints. His fluoroscopic
research with Rick
Komistek has produced
invaluable data, which
have been widely
propagated at conferences
around the world. (Photo
taken c 2017) (© Peter S
Walker)
displacements. A requirement was that the knee had to stay in the frame of the
x-ray cameras, so the motions studied were mainly a lunge through a full flexion
angle, rising from a seated position and walking on a treadmill which is not the
same as normal walking. Important insights were discovered about total knee
function by Banks and Hodge. Many surgeons collaborated in using their meth-
odology for the evaluation of a new design or surgical technique; as examples,
“how is high flexion achieved in the anatomic and artificial knee?”, “what is the
role of femoral rollback and tibial slope?”, “do bicruciate (ACL+PCL) and
medial pivot artificial knees move similarly to the anatomic knee?”, and “why
do CR and PS knees have such variable centers of internal-external rotation?”.
Richard Komistek Ph.D. adopted the method a few years later while working in
Denver, Colorado, and carried out numerous collaborative studies with Doug
Fig. 14.12 After starting

off as a designer in
industry, Rik Komistek
Ph.D. moved to the
University of Colorado in
Denver in 1994 where he
established a fluoroscopic
evaluation laboratory in
collaboration with Doug
Dennis MD and Jim Stiehl
MD. They carried out
groundbreaking influential
studies on the motion of
different types of total
knee in various activities.
In 2002 Rik moved to
Knoxville where he
continued this work,
expanding it in engineering
sophistication with the
application of Kane’s
Dynamics to analyze the
forces in the
musculoskeletal system.
Recently his lab developed
the first mobile
fluoroscopic unit for
studying the kinetics of
functional activities.
(Photo taken 2019)
(© Peter S Walker)
Dennis MD and Jim Stiehl MD. Subsequently many other surgeons sent their
fluoroscopic data for analysis using their software system. After moving his
laboratory to Knoxville, Tennessee, Rick Komistek extended his work in several
directions such as analysis using Kane’s dynamics, but still including fluoros-
copy as the core technology.
Depicting three-dimensional motion data in a way that can be readily interpreted
is not easy. However for artificial knees, this was solved by plotting the succession
of femoral-tibial contact points on to the lateral and medial tibial surface, connected
with straight lines. This provided visualization of position, whether anterior or pos-
terior; rollback, whether uniform or different laterally and medially; and internal-
external rotation, together with the center of rotation. Possibly the most influential
fluoroscopic study was a comparison of over 800 CR and PS knees in the stance
phase of walking and deep knee bend (Dennis et al. 2003). The motion patterns of
artificial knee motions were compared with those of normal intact knees. In gait,
from heel strike to toe-off, normal knees showed posterior displacement laterally
but small medial displacements. The artificial knees showed highly variable patterns
of motion, but generally with less lateral displacement than normal. In some knees
there was anterior displacement of contacts, opposite to normal behavior. In deep
Circular axis
Epicondylar
axis
Best-fit
circles
Circular axis projected

on to tibial surface
LAT
L
MED
Posterior
reference
line
Contact
patches
Fig. 14.13 Method for describing the motion of the femur on the tibia during flexion. The circular
axis is projected on to the tibial surface plane. At a particular flexion angle, distances L and M
describe the femoral position on the tibia. The succession of the circular axis projections at a suc-
cession of discrete flexion angles describes the motion path. For a total knee, this method is similar
to describing the centers of the lateral and medial contact patches, if the circular axis of the con-
dyles of the femoral component is used. The epicondylar axis in general will not be coincident with
the circular axis
knee bend, normal knees showed 19 mm posterior lateral displacement and 3 mm
posterior medial displacement, resembling a medial pivot action. Artificial knees
with ACL + PCL showed similar patterns to normal knees. In contrast, CR knees
showed a highly variable pattern, but on average there were equal lateral and medial
displacements of the contact points. Up to 90° flexion, in 70% of cases, there was
more than 3 mm anterior femoral displacement during flexion, with an average of
8 mm, so-called paradoxical sliding. In PS knees there were again almost equal
lateral and medial displacements on average, with paradoxical sliding in only 43%
of cases, average 6 mm. The authors considered that this sliding reduced the maxi-
mum flexion achieved and reduced the quadriceps moment arm, reducing efficiency.
The authors concluded that “The sagittal plane translation patterns after an artificial
knee usually do not duplicate those of the normal knee.” The relative magnitude of
the displacements between the different designs tested could be explained largely
by the relative amount of dishing of the tibial bearing surface, the most displace-
ment for the ACL + PCL preserving, intermediate for the CR, and least for the
PS. These relative sliding distances can be explained by the different dishing
between the three types of designs. The lack of a medial pivot action as in normal
anatomic knees could be explained by the symmetric condyle shapes of the artificial
knees, compared with the asymmetric geometry in the normal knee. Recently, Rick
Komistek in Knoxville, Marcus Pandy Ph.D. in Melbourne, and William Taylor
Ph.D. in Zurich developed a mobile fluoroscopy system where the x-ray cameras
followed the center the knee. These extraordinary machines allow many types of
everyday activities to be studied, and data is just now beginning to emerge. A recent
article has reviewed numerous fluoroscopic studies which have showed characteris-
tic, and variable, patterns of motion for different implant designs and examined
whether the motion was associated with outcomes (Angerame et al. 2019). The
general findings were that there were few relations even when some functional
effect of motion might be expected.
Biomechanical Evaluation Methods for the Clinic
Due to the expense, time taken, specialized equipment, and expertise required for
the above gait analysis and fluoroscopic methodologies, they have been used mainly
in a research setting rather than in general clinical practice. Nevertheless for evalu-
ating a new artificial knee design or technique, such studies provide in-depth data
about the function of the knee. For general clinical evaluations however, as an alter-
native, a number of simpler tests have been developed and used extensively, so far
used mainly in physical therapy and rehabilitation rather than in the orthope-
dics clinic. In the Timed Up and Go (TUG) test, the test subject starts from a seated
position, they then rise, walk 3 m, turn around, walk back to the chair, and sit down
again. The test is described by Dobson (2015): “The TUG is a transition test that
assesses strength, agility and dynamic balance during multiple activities including
Sit-to-Stand, walking short distances and changing direction whilst walking.” The
measure is the time taken to complete the test. The test is simple to carry out and has
a single number to quantify the result. The sit-to-stand (STS) test has three versions,
testing knee strength or endurance, important indicators of function. In the 30-sec-
ond test, the subject stands up and sits down as many times as possible. The normal
value for somebody of average total knee age is 12–18 times. In the five-times sit-
to-stand test, the time is measured for the subject to stand-sit five times. In the right-
left comparison test, the right/left force ratio is calculated. In recent years, with the
increasing availability of wearable sensors, biomechanical methods such as a “por-
table gait lab” using accelerometers have been conveniently implemented in a clini-
cal setting (Parks et al. 2015). Ground-to-foot forces and pressures, as well as
cadence and stance times, have been measured using instrumented insoles (Stöggl
and Martiner 2017). The thin insoles only 3 mm in thickness are placed into the
normal footwear used by the test subject, with the data being transmitted wirelessly
3 mm thickness
Sit to Stand
Array of 13
force transducers
embedded in sole Dual force plates
Accelerometers
strapped above Standing
Rising
and below the Total force
knee Left & right force
Sitting
Time
Left Left
Acceleration
Step- femur tibia
down
Step-up test test
Time Time
Fig. 14.14 Equipment is now available for carrying out simple biomechanical tests in the clinic.
The Instrumented Insoles are used for measuring right and left forces and force ratios, in a Timed
Up and Go test, in walking, and in step-up/step-down tests. The dual force plates are used in a sit-
to-stand test, to measure right-left force ratios in rising and standing. The IMUs (inertial measuring
units) are used for measuring joint angles during activities and the accelerations below and above
the knee
to a computer. Several activity monitors have been introduced in recent years, with
the idea of accumulating the types and frequency of activities that the subject carries
out over a prescribed time period. These devices, which attach to a location on the
body, have the advantage of quantifying activity and alerting the physician or thera-
pist if progress is deficient by linking the data through a cell phone (Lipperts et al.
2017). In the following section, examples will be given of the application of the
evaluation methods described above, to compare different artificial knee designs.
In the above study to compare rehabilitation methods, Pozzi et al. (2020) used the
Timed Up and Go test, the stair climb test, and the 6-minute walk test. Parks et al.
(2015) used similar tests as well as a portable gait lab to track the progress of recov-
ery after a total knee, finding that most of the measures had peaked at the 4-month
follow-up time, although in general, 1 year is generally used as the minimum time
where close to full recovery has been reached. Such tests, particularly the TUG test,
have been used primarily to monitor or compare the progress of a particular therapy
or to compare one therapy to another. Examples are as follows: surgical balancing of
the medial side of the knee to avoid looseness was found to produce shorter TUG
times than when the MCL was loose; TUG times up to 3 months after surgery were
used to compare whether using a tourniquet or not at surgery made any difference to
functional recovery; recovery trajectories after different therapies, high intensity or
low intensity, were measured using TUG times; and whether there was a greater
variation in function within a patient group, before and after artificial knee surgery.
The advantage of this type of measure over a PROM is that it provides an objective
measurable quantity which signifies functional ability and is not subject to interpre-
tation by the patient. However, at this time there are relatively few available studies
where these measures have been used as the principal method to compare different
artificial knee designs or surgical procedures. Based on those studies which have
used the methods however, they do appear to be a useful adjunct to the use of PROMs.
In certain evaluations, they have been used as the primary indicator.
Function Measured in the Laboratory
To evaluate the function of an artificial knee design, the most objective method
may be a mechanical test or a computer model. This might seem to be unreason-
able in that how can an artificial knee be evaluated without being in the patient?
There is however some rationale. An automobile has performance measures of the
engine and suspension, which can be tested objectively in the factory or on a test
track. Put the car in the hands of 20 drivers, and the way the car performs will be
completely different in every case. A Maserati in the hands of a cautious driver
would be driven similarly to a family sedan. Yet each type of car will have its own
objective performance measures, no matter who drives the car or how. One obvious
way in which an artificial knee can be tested is in a knee simulating machine,
which can be used for both wear and kinematics. The Stanmore knee simulator was
used by John DesJardins, now at Clemson University, South Carolina, to test the
kinematics of different designs of artificial knee and also to study the effects of
different stiffnesses of ligament constraints as well as ligament resections
(Desjardins et al. 2000). The results revealed that certain designs might be over-
constrained, while at the other extreme, loose ligaments could result in instability
in a low-conformity design. In a recent paper, Hani Haider (2017) stated: “The
future will carry exciting developments into more sophisticated soft tissue simula-
tion for the force-control methodology. More daily activities beyond walking will
be simulated…” Crouching machines, a derivative of the original Oxford test rig,
as well as modern robotic arms, have been used in comparisons between knee
designs and found to give realistic data (Varadarajan et al. 2009). An authoritative
summary paper has been written of the knee forces during different activities mea-
sured by instrumented tibial components in vivo (Bergmann et al. 2014). This data
has been submitted to the American Society for Testing and Materials (ASTM) as
the basis for a new standard by ASTM. Although not yet implemented, the data
acts as a guideline for advanced simulation experiments. In terms of motion inputs,
it seems logical to input only the flexion-extension cycle for the different activities.
The other displacements and rotations will be determined mainly by the restraints
from the bearing surfaces of the artificial knee and the soft tissues, which is the
principle of a force-controlled machine. Advanced Mechanical Technology
(AMTI) in Watertown, MA, recently introduced the VIVO machine for studying
the kinematics of the knee. Forces and flexion-extension are input to the knee,
while the ligaments are represented virtually. With such a machine, using the
Bergmann data, the kinematics of any artificial knee can be determined under a
sequence of ligament properties, alignments, and activity scenarios. This will eval-
uate laxity, stability (and instability), and robustness of designs to extreme loading
and alignment conditions. An example of how the machine can be used to study the
effects of soft tissue variations was recently published (Willing and Walker 2018).
hich Evaluation Methods Are Applicable, and Are There

W
Differences Between Designs and Techniques?
Patient-reported outcome measures (PROMs) have been by far the most frequently
used method for evaluating artificial knees or surgical techniques. PROMs pro-
vide an overall assessment of the function of the patient and are suitable for data-
bases of large numbers of patients. However for evaluating specific aspects of
function in a more quantitative way, a broad range of biomechanical evaluation
methods is available. The methods include radiographic techniques, particularly
fluoroscopy but also MRI and RSA; gait and force analysis and muscle force mea-
surements; simple clinical tests such as TUG, STS, and activity monitoring; and
finally, computer simulations. Based on PROM data, it has been difficult to detect
systematic differences between different designs in function. However the biome-
chanical evaluation methods have detected differences between designs and tech-
niques, in gait, kinematics, forces, and other factors. Their results have provided
pointers to improvements in design, technique, and rehabilitation methods.
Biomechanical evaluation methods, whether complex and used in specialist labo-
ratories, or simple and used in a clinical setting, can provide yet further insights.
References
Aljehani M, Madara K, Snyder-Mackler L, Christiansen C, Zeni JA Jr. The contralateral knee may
not be a valid control for biomechanical outcomes after unilateral total knee arthroplasty. Gait
Posture. 2019;70:179–84.
Andriacchi TP, Andersson GB, Fermier RW, Stern D, Galante JO. A study of lower-limb mechan-
ics during stair-climbing. J Bone Joint Surg Am. 1980;62(5):749–57.
Andriacchi TP, Galante JO, Fermier RW. The influence of total knee-replacement design on walk-
ing and stair-climbing. J Bone Joint Surg Am. 1982;64(9):1328–35.
Angerame MR, Holst DC, Jennings JM, Komistek RD, Dennis DA. Total knee arthroplasty kine-
matics. J Arthroplast. 2019;34(10):2502–10.
Bailey O, Ferguson K, Crawfurd E, James P, May PA, Brown S, Blyth M, Leach WJ. No clinical
difference between fixed- and mobile-bearing cruciate-retaining total knee arthroplasty: a pro-
spective randomized study. Knee Surg Sports Traumatol Arthrosc. 2015;23(6):1653–9.
Banks SA, Hodge WA. Accurate measurement of 3-dimensional knee replacement kinematics
using single-plane fluoroscopy. IEEE Trans Biomed Eng. 1996;43(6):638–49.
Beaupre LA, Sharifi B, Johnston DWC. A randomized clinical trial comparing Posterior-Cruciate
Stabilizing vs Posterior Cruciate-Retaining Prostheses in primary total knee. J Arthroplasty.
2017;32(3):819–23.
Behrend H, Giesinger K, Giesinger JM, Kuster MS. The “forgotten joint” as the ultimate goal
in joint arthroplasty: validation of a new patient-reported outcome measure. J Arthroplasty.
2012;27(3):430–436.e431.
Behrend H, Zdravkovic V, Bosch M, Hochreiter B. No difference in joint awareness after TKA:
a matched-pair analysis of a classic implant and its evolutional design. Knee Surg Sports
Traumatol Arthrosc. 2019;27(7):2124–9.
Benjamin B, Pietrzak JRT, Tahmassebi J, Haddad FS. A functional comparison of medial pivot
and condylar knee designs based on patient outcomes and parameters of gait. Bone Joint J.
2018;100-b(1 Supple A):76–82.
Berend KR, Lombardi AV, Adams JB. Which total knee replacement implant should I pick?
Correcting the pathology: the role of knee bearing designs. Bone Joint J. 2013;95-B(11 Suppl
A):129–32.
Bergmann G, Bender A, Graichen F, et al. Standardized loads acting in knee implants. PLoS One.
2014;9(1):e86035.
Berliner JL, Brodke DJ, Chan V, SooHoo NF, Bozic KJ. Can preoperative patient-reported outcome
measures be used to predict meaningful improvement in function after TKA? Clin Orthop Relat
Res. 2017;475(1):149–57.
Bordini B, Ancarani C, Fitch DA. Long-term survivorship of a medial-pivot total knee system com-
pared with other cemented designs in an arthroplasty registry. J Orthop Surg Res. 2016;11:44.
Catani F, Benedetti MG, Bianchi L, Marchionni V, Giannini S, Leardini A. Muscle activity
around the knee and gait performance in unicompartmental knee arthroplasty patients: a com-
parative study on fixed- and mobile-bearing designs. Knee Surg Sports Traumatol Arthrosc.
2012;20(6):1042–8.
D’Lima DD, Patil S, Steklov N, Colwell CW Jr. The 2011 ABJS Nicolas Andry award: ‘Lab’-
in-a-knee: in vivo knee forces, kinematics, and contact analysis. Clin Orthop Relat Res.
2011;469(10):2953–70.
Daines BK, Dennis DA. Gap balancing vs. measured resection technique in total knee arthroplasty.
Clin Orthop Surg. 2014;6(1):1–8.
Dennis DA, Komistek RD, Mahfouz MR, Haas BD, Stiehl JB. Multicenter determination of in vivo
kinematics after total knee arthroplasty. Clin Orthop Relat Res. 2003;416:37–57.
DesJardins JD, Walker PS, Haider H, Perry J. The use of a force-controlled dynamic knee simula-
tor to quantify the mechanical performance of total knee replacement designs during functional
activity. J Biomech. 2000;33(10):1231–42.
Dobson F. Timed up and go test in musculoskeletal conditions. J Physiother. 2015;61(1):47.
Dobson F, Hinman RS, Roos EM, Abbott JH, Stratford P, Davis AM, Buchbinder R, Snyder-
Mackler L, Henrotin Y, Thumboo J, Hansen P, Benell K. OARSI recommended performance-
based tests to assess physical function in people diagnosed with hip or knee osteoarthritis.
Osteoarthr Cartil. 2013;21(8):1042–52.
Ewald FC. The knee society total knee arthroplasty roentgenographic evaluation and scoring sys-
tem. Clin Orthop Relat Res. 1989;248:9–12.
Frosch KH, Nägerl H, Kubein-Meesenburg D, Buchholz J, Dörner J, Dathe H, Hellerer O, Dumont
C, Stürmer KM. A new total knee arthroplasty with physiologically shaped surfaces. Part 2:
first clinical results. Unfallchirurg. 2009;112(2):176–84.
Giesinger K, Hamilton DF, Jost B, Holzner B, Giesinger JM. Comparative responsiveness of out-
come measures for total knee arthroplasty. Osteoarthr Cartil. 2014;22(2):184–9.
Goh GS, Bin Abd Razak HR, Tay DK, Chia SL, Lo NN, Yeo SJ. Unicompartmental knee arthro-
plasty achieves greater flexion with no difference in functional outcome, quality of life, and
satisfaction vs total knee arthroplasty in patients younger than 55 years. A propensity score-
matched cohort analysis. J Arthroplast. 2018;33(2):355–61.
Graff C, Hohmann E, Bryant AL, Tetsworth K. Subjective and objective outcome measures after
total knee replacement: is there a correlation? ANZ J Surg. 2016;86(11):921–5.
Haider H. Chapter 7.10: Wear: knee joint arthroplasty. In: Ducheyne P, Grainger DW, Healy
KE, Hutmacher DW, Kirkpatrick CJ, editors. Comprehensive biomaterials II, vol. 7. Oxford:
Elsevier; 2017. p. 152–74.
Harrington MA, Hopkinson WJ, Hsu P, Manion L. Fixed- vs mobile-bearing total knee arthro-
plasty: does it make a difference?--a prospective randomized study. J Arthroplast. 2009;24(6
Suppl):24–7.
Harris AI, Christen B, Malcorps JJ, O’Grady CP, Kopjar B, Sensiba PR, Vandenneucker H, Huang
BK, Cates HE, Hur J, Marra DA. Midterm performance of a Guided-Motion bicruciate-
stabilized total knee system: rseults from the International study of over 2000 consevutive
primary total knee arthroplasties. J Arthroplasty. 2019;34(7S):S201–S208.
Hauer G, Sadoghi P, Bernhardt GA, Wolf M, Ruckenstuhl P, Fink A, Leithner A, Gruber G. Greater
activity, better range of motion and higher quality of life following unicompartmental knee
arthroplasty: a comparative case-control study. Arch Orthop Trauma Surg. 2020;140(2):231–7.
Hossain FS, Konan S, Patel S, Rodriguez-Merchan EC, Haddad FS. The assessment of outcome
after total knee arthroplasty: are we there yet? Bone Joint J. 2015;97-b(1):3–9.
Howell SM, Shelton TJ, Hull ML. Implant survival and function ten years after kinematically
aligned total knee arthroplasty. J Arthroplast. 2018;33(12):3678–84.
Jaffe WL, Dundon JM, Camus T. Alignment and balance methods in total knee arthroplasty. J Am
Acad Orthop Surg. 2018;26(20):709–16.
Jiang C, Liu Z, Wang Y, Bian Y, Feng B, Weng X. Posterior cruciate ligament retention versus poste-
rior stabilization for total knee arthroplasty: a meta-analysis. PLoS One. 2016;11(1):e0147865.
Katchky AM, Jones CW, Walter WL, Shimmin AJ. Medial ball and socket total knee arthroplasty
five-year clinical results. Bone Joint J. 2019;101B(1_Supple_A):59–65.
Kim YH, Park JW, Kim JS. Clinical outcome of medial pivot compared with press-fit
condylar sigma cruciate-retaining mobile-bearing total knee arthroplasty. J Arthroplast.

2017;32(10):3016–23.
Kleeblad LJ, van der List JP, Zuiderbaan HA, Pearle AD. Larger range of motion and increased
return to activity, but higher revision rates following unicompartmental versus total knee
arthroplasty in patients under 65: a systematic review. Knee Surg Sports Traumatol Arthrosc.
2018;26(6):1811–22.
Komnik I, Weiss S, Pagani CF, Potthast W. Motion analysis of patients after knee arthroplasty dur-
ing activities of daily living–a systematic review. Gait Posture. 2015;41(2):370–7.
Lipperts M, Laarhoven SV, Senden R, Heyligers I, Grimm B. Clinical validation of a body-fixed
3D accelerometer and algorithm for activity monitoring in orthopaedic patients. J Orthop
Transl. 2017;11:19–29.
Longstaff LM, Sloan K, Stamp N, Scaddan M, Beaver R. Good alignment after total knee arthro-
plasty leads to faster rehabilitation and better function. J Arthroplast. 2009;24(4):570–8.
Lyons MC, MacDonald SJ, Somerville LE, Naudie DD, McCalden RW. Unicompartmental
versus total knee arthroplasty database analysis: is there a winner? Clin Orthop Relat Res.
2012;470(1):84–90.
McClelland JA, Webster KE, Feller JA. Gait analysis of patients following total knee replacement:
a systematic review. Knee. 2007;14(4):253–63.
Meneghini RM, Stefl MD, Hodge WA, Banks SA. A cam–post mechanism is no longer necessary
in modern primary total knee arthroplasty. J Knee Surg. 2019;32:710–3.
Migliorini F, Tingart M, Niewiera M, Rath B, Eschweiler J. Unicompartmental versus total knee
arthroplasty for knee osteoarthritis. Eur J Orthop Surg Traumatol. 2019;29(4):947–55.
Mizner RL, Petterson SC, Clements KE, Zeni JA Jr, Irrgang JJ, Snyder-Mackler L. Measuring func-
tional improvement after total knee arthroplasty requires both performance-based and patient-
report assessments: a longitudinal analysis of outcomes. J Arthroplast. 2011;26(5):728–37.
Nha K-W, Shon O-J, Kong B-S, Shin Y-S. Gait comparison of unicompartmental knee arthroplasty
and total knee arthroplasty during level walking. PLoS One. 2018;13(8):e0203310.
Parks NL, Whitney CE, Engh GA. Can we quantify functional improvement following total knee
arthroplasty in the clinical setting? J Knee Surg. 2015;28(6):475–82.
Pfitzner T, Perka C, von Roth P. Unicompartmental vs. total knee arthroplasty for medial osteoar-
thritis. Zeitschrift fur Orthopadie und Unfallchirurgie. 2017;155(5):527–33.
Powell AJ, Crua E, Chong BC, et al. A randomized prospective study comparing mobile-bearing
against fixed-bearing PFC Sigma cruciate-retaining total knee arthroplasties with ten-year
minimum follow-up. Bone Joint J. 2018;100-b(10):1336–44.
Pozzi F, White DK, Snyder-Mackler L, Zeni JA. Restoring physical function after knee replace-
ment: a cross sectional comparison of progressive strengthening vs standard physical therapy.
Physiother Theory Pract. 2020;36(1):122–33.
Arthroplast. 2011;26(2):224–8.
Roos EM, Roos HP, Lohmander LS, Ekdahl C, Beynnon BD. Knee Injury and Osteoarthritis
Outcome Score (KOOS)–development of a self-administered outcome measure. J Orthop
Sports Phys Ther. 1998;28(2):88–96.
Samy DA, Wolfstadt JI, Vaidee I, Backstein DJ. A retrospective comparison of a medial pivot and
posterior-stabilized total knee arthroplasty with respect to patient-reported and radiographic
outcomes. J Arthroplast. 2018;33(5):1379–83.
Scuderi GR, Bourne RB, Noble PC, Benjamin JB, Lonner JH, Scott WN. The new knee society
knee scoring system. Clin Orthop Relat Res. 2012;470(1):3–19.
Scuderi GR, Sikorskii A, Bourne RB, Lonner JH, Benjamin JB, Noble PC. The knee society short
form reduces respondent burden in the assessment of patient-reported outcomes. Clin Orthop
Relat Res. 2016;474(1):134–42.
Serna-Berna R, Lizaur-Utrilla A, Vizcaya-Moreno MF, Miralles Muñoz FA, Gonzalez-Navarro B,
Lopez-Prats FA. Cruciate-retaining vs posterior-stabilized primary total arthroplasty. Clinical
outcome comparison with a minimum follow-up of 10 years. J Arthroplast. 2018;33(8):2491–5.
Shakespeare D, Ledger M, Kinzel V. Flexion after total knee replacement. A comparison between
the Medial Pivot knee and a posterior stabilised implant. Knee. 2006;13(5):371–3.
Siljander MP, McQuivey KS, Fahs AM, Galasso LA, Serdahely KJ, Karadsheh MS. Current trends
in patient-reported outcome measures in total joint arthroplasty: a study of 4 major orthopaedic
journals. J Arthroplast. 2018;33(11):3416–21.
Singleton N, Nicholas B, Gormack N, Stokes A. Differences in outcome after cruciate
retaining and posterior stabilized total knee arthroplasty. J Orthop Surg (Hong Kong).
2019;27(2):2309499019848154.
Skoffer B, Dalgas U, Mechlenburg I, Soballe K, Maribo T. Functional performance is associated
with both knee extensor and flexor muscle strength in patients scheduled for total knee arthro-
plasty: a cross-sectional study. J Rehabil Med. 2015;47(5):454–9.
Spekenbrink-Spooren A, Van Steenbergen LN, Denissen GAW, Swierstra BA, Poolman RW,
Nelissen RGHH. Higher mid-term revision rates of posterior stabilized compared with cruci-
ate retaining total knee arthroplasties: 133,841 cemented arthroplasties for osteoarthritis in the
Netherlands in 2007-2016. Acta Orthop. 2018;89(6):640–5.
Stöggl T, Martiner A. Validation of Moticon’s OpenGo sensor insoles during gait, jumps, balance
and cross-country skiing specific imitation movements. J Sports Sci. 2017;35(2):196–206.
Van der List JP, Kleeblad LJ, Zuiderbaan HA, Pearle AD. Mid-term outcomes of metal-backed
unicompartmental knee arthroplasty show superiority to all-polyethylene unicompartmental
and total knee arthroplasty. HSS J. 2017;13(3):232–40.
Varadarajan KM, Harry RE, Johnson T, Li G. Can in vitro systems capture the characteristic dif-
ferences between the flexion–extension kinematics of the healthy and TKA knee? Med Eng
Phys. 2009;31(8):899–906.
Weiss JM, Noble PC, Conditt MA, et al. What functional activities are important to patients with
knee replacements? Clin Orthop Relat Res. 2002;404:172–88.
White PB, Ranawat AS, Ranawat CS. Fixed bearings versus rotating platforms in total knee arthro-
plasty. J Knee Surg. 2015;28(5):358–62.
Willing R, Walker PS. Measuring the sensitivity of total knee replacement kinematics and laxity to
soft tissue imbalances. J Biomech. 2018;77:62–8.
Yoshida Y, Mizner RL, Snyder-Mackler L. Association between long-term quadriceps weakness
and early walking muscle co-contraction after total knee arthroplasty. Knee. 2013;20(6):426–31.
Young T, Dowsey MM, Pandy M, Choong PF. A systematic review of clinical functional outcomes
after medial stabilized versus non-medial stabilized total knee joint replacement. Front Surg.
2018;5:25.
Discussion and Future Possibilities
How Has the Present Status Been Reached,
15
and What Further Advances Can Be Made?
It is now more than 50 years since the first cemented cobalt-chrome on polyethylene
total knees were invented. Some of the earliest designs functioned well and achieved
a high level of survivorship. For the original Total Condylar knee, of 220 knees
performed in the mid-1970s, only 14 revisions were required at an average of 11.4-
year follow-up (Rodriguez et al. 2001). For the Townley knee, survivorship was
89% at 23 years with an average flexion of 117°, the main reason for failure being
polyethylene wear (Pritchett 2015). For the AGC knee, there was 96% survivorship
at 15 years in a large cohort of over 1500 knees (Atrey et al. 2014). Over 3000 early
knee designs, including Total Condylar, Press-fit Condylar, Kinematic, Kinemax,
and AMK, performed at the Wrightington Hospital in England, showed a survivor-
ship of 89% at 10 years (Pradhan et al. 2006). For 99 cases of the Insall-Burstein
knee, the survivorship was 92% at 10 years with an average flexion of 106° (Aglietti
et al. 1999). From multiple registries, with many different designs over different
time periods, pooled data showed that 82% of knees lasted for 25 years (Evans et al.
2019). Overall, this was a remarkable achievement after little more than 10 years of
development.
By 1980, two particular design types, the posterior cruciate retaining (CR) and
the posterior stabilized (PS), were widely used and were considered successful in
terms of function and durability. Several companies manufactured their versions of
the CR and PS. These two designs dominated the market, although other design
types had not been abandoned. Designs which preserved both the ACL and PCL
were being used. Unicompartmental knees maintained their place, mainly for
medial osteoarthritis. The rotating platform mobile bearing knee had recently been
introduced and had gained a following. Stabilized designs with intercondylar varus-
valgus stability were used for severe instability or for revisions.
Since 1980, due to the continuous stream of research, clinical follow-up studies,
advances in technology, surgical specialization, and education, the quality of the
designs and techniques has steadily advanced. This has produced not only further
improvements in functional results but more consistency with fewer outliers.
Various problems and deficiencies have been identified and largely resolved. For

296 15 Discussion and Future Possibilities
example, some of the early tibial components fractured due to stress concentrations
or misalignment. Serious polyethylene wear was caused by excessive contact
stresses, often in conjunction with uncontrolled sliding. In the longer term, wear
was exacerbated by gradual oxidation of the polyethylene. Instability resulted from
bearing surfaces which were inadequately constrained in combination with soft tis-
sues which were too loose. Misalignments, most commonly excessive varus, led to
elevated medial loading, causing bone collapse, component loosening, and polyeth-
ylene wear. Loosening at the cement-bone interface, one of the most common prob-
lems, was due to inadequate cementing technique, misalignment, or inadequate
bone coverage. Patella pain and even dislocation were due to a shallow trochlea
groove on the femoral component or erroneous rotational positioning of the femoral
or tibial components. Inadequate range of flexion was caused by a femoral compo-
nent which was incorrectly sized, by a tibial component with too much posterior
constraint, or where the posterior tibial tilt angle was insufficient. “Clicking and
clunking” noises occurred, especially on PS designs, due to cam-post engagements
and to soft tissue entrapment in the intercondylar region, as well as by impacting of
the metal on the plastic, although this problem has not been associated with dimin-
ished function. Most knee designs which were prone to such problems as the above
were either removed from the market, modified, or replaced by newer designs, with
the result that the problems have been reduced to a much lower level. For example,
the NexGen CR and PS knees had a survivorship of 100% and 98% at 15-year fol-
low-up (Sartawi et al. 2018). The Triathlon knee had 98% survivorship at 10 years
(Scott et al. 2019).
The situation today (2020) is that most artificial knees are part of a system con-
sisting of several different design types, often ranging from primaries to revisions.
The American Joint Replacement Registry (AJRR), founded in 2010 and affiliated
with the American Academy of Orthopaedic Surgeons (AAOS), is accumulating
invaluable data on artificial knee usage in the USA. The data is based on about one
third of the knees used today, and the proportion continues to grow as more institu-
tions participate. Up to this time, the mean age of the patients is 66.8 years with a
standard deviation of 9.6 years. This indicates that about 16% of patients are below
57 years old, a challenging group due to their higher activity and expectation levels.
As expected, there are more female than male patients, at 59% and 41%. The distri-
butions of design types for primary knee cases confirm the persistence of types
invented in the 1970s: 50% are posterior stabilized (PS), and 40% are posterior
cruciate retaining (CR). The remaining 10% is divided between rotating platforms
(7%) and ultracongruent (3%). Unicompartmental knees are only at about 3% of
total knees despite the more conservative procedure, quicker recovery, and better
functional outcomes, although the durability has been less than that for total knees
in some series (Pfitzner et al. 2017).
The commercial exhibits at the major orthopedic conferences, notably the
American Academy of Orthopaedic Surgeons (AAOS), the European Federation of
National Associations of Orthopaedics and Traumatology (EFORT), and major
Asian meetings, show that a majority of the market is serviced by a few large com-
panies. However there are still numerous smaller companies. Many of their designs
15 Discussion and Future Possibilities 297
Fig. 15.1 David
G. Lewallen MD from the
Mayo Clinic, Rochester,
MN, is the Medical
Director of the American
Joint Replacement Registry
(AJRR). This organization
collects performance data
of artificial joints from
numerous hospitals in the
USA. The purpose is to
improve the long-term
outcomes and the quality
of care. (Photo courtesy
David Lewallen)
are standard CR or PS, where the designs are obtained from previous versions which
have a successful clinical history. In some cases, minor improvements or advance-
ments in instrumentation have been made. Other smaller companies offer special-
ized designs such as medial pivots, mobile bearings, and guided motion types, while
one company specializes in custom-made knees.
Today’s CR and PS designs, descendants of designs from the 1970s, do have
certain differences. Some are close to being anatomic in their general shape, includ-
ing the radii of the sagittal profiles of the femoral component and the peripheral
shape of the tibial component. Other designs are more geometric, such as having a
constant sagittal radius of the femoral bearing surfaces. Several systems offer a
large number of sizes and shapes to closely match any particular knee, where gender
and ethnic shape differences have been taken into account (Mahfouz et al. 2012).
Some cam-post geometries engage in early flexion in an effort to maintain stability
and provide a high flexion angle due to femoral rollback. Though many of these
differences are important, there is a general similarity in the CR and PS designs
produced by the different companies, such that there is a perception that the designs
have become generic. Typically, clinical follow-up studies of a particular CR or PS
design are often taken to represent the class as a whole. The justification is that for
the most part, the results between different CR designs or PS designs have been
similar, while even between CR and PS themselves, major differences have not been
demonstrated. This is remarkable considering the fundamental differences in design

and mechanics between a CR and a PS.
Obviously today’s designs are not exactly the same models as in the 1970s, as
there have been numerous modifications since then, but the basic design configura-
tions are clearly recognizable. How has this situation been reached? After all, there
are few products in today’s world which have remained the same over such a long
period of time. One only has to look at automobiles, bicycles, cameras, and cell
phones to realize that. Part of the answer can be found in the way in which specific
knee designs became widely used and maintained their standing.
It is instructive to identify the reasons for the success of certain designs in the
initial time period from the late 1960s up to 1980:
• A prestigious institution. In such institutions, there was invariably a well-

organized total knee service with a high volume of cases, a trainee and fellow-
ship program for carrying out follow-up studies, research facilities, and affiliation
with a university. There were resources available for new developments as well
as ongoing evaluation. Even if the institution did not originate the design them-
selves, if their surgeons used a high volume of a particular design and reported
on results at conferences, that represented strong support for the design. In this
case, the support would be strengthened because of the objectivity and lack of
any direct gain for that institution.
• A well-known and authoritative surgeon-designer. This was essential in present-
ing the design to surgeon groups at conferences and in producing publications.
An enthusiastic and confident presentation manner was also a positive factor. It
certainly helped that artificial knees were just beginning and were seen as an
important new breakthrough in the treatment of osteoarthritis of the knee, as had
already been demonstrated for the hip.
• Collaboration between surgeon and bioengineer. This combined both surgical
and technical expertise, together with the characteristic thinking patterns of the
different professions. Technical aspects of design and materials were unfamiliar
to most surgeons, while experimental and analytical support would add to the
credibility and understanding of the design. The manufacturers themselves,
while having expertise in production and materials, were generally unfamiliar
with testing methods and did not possess suitable equipment.
• A team of reputable surgeons. This signified a broad consensus of the positive
design features of the particular design. In parallel there would be less likelihood
of design flaws. Another advantage was a wider representation of the design at
conferences.
• A well-established company. This was a major factor simply because of the man-
ufacturing facilities and expertise and the sales exposure across the country. Also
the company could more easily develop relations with prominent surgeons. Such
relations were not subject to the regulatory restrictions as they are today.
• Strong company support for the design. This was a major factor, especially from
a well-established company. If the design was promoted as the flagship knee of
the company, it received high prominence on the company’s exhibit at confer-
ences, which elevated the confidence of potential surgeon users. David Fitzgerald,
President of Howmedica in the 1970s and 1980s, considered that this factor sig-
nified the company’s confidence that the product would be a clinical and com-
mercial success.
• A convincing design rationale. In the early days of artificial knee design, sound con-
cepts backed up by logical thinking and experimental data, as well as ideas which
related to anatomic mechanics, carried a great deal of weight. In this arena, technical
experts such as materials scientists and biomechanical engineers played an important
role. In contrast, a “new” or “different” idea without a great deal of background evi-
dence might attract attention, but surgeons were reluctant to use on patients.
• A well-thought-out technique with accurate and easy-to-use instruments. While
skilled surgeons were capable of inserting components requiring “hand-fitting,”
it was obvious that components with straight edges using instruments to guide
the saw cuts would result in more consistent accuracy and a shorter operating
time. Techniques that were targeted to correct alignment as well as to soft tissue
balancing were favored.
• Few problems such as loosening or instability during early clinical application by
the developers. It was in everybody’s interests to use a design that would work well
with few problems. There were many instances of designs being withdrawn due to
short-term problems, unless a convincing change in design or technique had been
introduced to solve the problem. But even this was not always easy due to lack of
confidence in the design and the availability of more successful designs on the
market.
With all of the above factors at play, it is understandable that there was a high
turnover of designs and a large disparity in the sales between one design and another
in the same company as well as between one company and another. In the early
days, some companies would even produce a design for a particular surgeon mainly
because they had a high volume of cases and the company wished to maintain good
relations with that surgeon. However some of these designs were not well thought
out and had to be withdrawn. Even the most popular designs were only made in two
or three sizes for economic reasons and because the advantages of closely matching
the anatomy were not appreciated. At conferences, there was intense rivalry between
the proponents of different concepts, such as cruciate retaining versus cruciate sac-
rificing, high or low conformity between the bearing surfaces, the best fixation peg
design for the tibial component, cement versus porous coated fixation, and whether
or not the patella should be replaced.
During the 1970s the designs gradually improved, more sizes were made, and
different design concepts were put together in “systems.” The idea of a design team
came into play as a result, combining the expertise of several surgeons and design-
ers with different areas of specialization. Knee systems typically consisted of
designs from unicompartmental to revision hinges, presented as being progressive
depending on the severity of the arthritis, especially the level of deformity and insta-
bility. Because of the large investment required to produce a system, the number of
separate design types tended to remain static.
In general, the clinical results improved, and there were no major differences
apparent in the performance between design types, although each type did have its
own nuances. By and large, there were two main “camps,” one for posterior cruciate
retention and one for posterior cruciate substitution with smaller support groups for
bi-cruciate retention, mobile bearing, and unicompartmental. In a similar way, there
were supporters of different surgical techniques, the main two approaches being gap
balancing where preference was given to equality of soft tissue tensions, and mea-
sured resection where placement of components based on anatomic landmarks was
the main goal followed by soft tissue adjustments.
By 1980, the CR and PS had become well established. The CR represented sur-
geons who preferred an anatomic approach, with the posterior cruciate seen as an
important stabilizer which would provide femoral rollback and produce a higher
range of flexion, and freedom of motion between the bearing surfaces to account for
a range of activities and differences between patients. The PS represented surgeons
who preferred a more mechanistic approach, where the surgical technique was eas-
ier and more reproducible, where stability was taken care of by the bearing surfaces
themselves, and where the cam-post would produce the rollback to increase flexion.
In parallel a great deal of attention had been given to tibial fixation, wear of the
polyethylene, stability, soft tissue balancing, and alignment, essentially refining the
designs and techniques.
Fig. 15.2 Seth Greenwald

Ph.D. founded Current
Concepts in Joint
Replacement (CCJR) in
1983, with the goal of
providing continuing
education for
reconstructive joint
surgeons, taught by the
experts in the field. Over
36 years, CCJR has been a
leading educator to tens of
thousands of surgeons
from the USA and around
the world. Seth also
founded a laboratory at
Cleveland, OH, dedicated
to the development of
ASTM standard tests for
evaluating contemporary
artificial knees. (Photo
taken at Specialty Day,
AAOS, 2018) (© Peter S
Walker)
By now, an additional criterion for the success of a new knee design had become
apparent:
• The design team. This expanded beyond a surgeon working with an engineer, or
a small group of surgeons. Now, the emphasis was on a larger group of experts,
from a number of institutions and countries. Many surgeons had gained consider-
able experience and were well known, while the same applied to biomechanical
engineers and materials scientists. The team would consist of such members. The
company would hold regular design meetings, where the team would be divided
into groups to deal with each type of design, and testing requirements would be
assigned to laboratories in the company or at universities. Some companies had
set up surgical skills facilities where instruments could be developed and knee
designs tested on specimens.
Barriers and Possibilities for Alternate Designs and Techniques
The pervasiveness of the posterior cruciate retaining (CR) design and the poste-
rior stabilized (PS) design with cruciate resection continued through the 1980s
and 1990s. During that time, there were steady improvements in the designs and
techniques but not in terms of changing the designs in any major way or introduc-
ing other design types. Comprehensive knee systems from the larger companies
dominated the market, which included revision components, in some cases using
porous materials as space fillers. The systems often included rotating platform
versions of both CR and PS designs because of the expectation that the wear
would be reduced and the function improved. While rotating platforms did gain
some popularity, by no means did it approach the volume of the standard fixed
bearing CR and PS designs. The main reasons for this were that the rotating plat-
form versions were more expensive and there was no clear indication from fol-
low-ups and experiments that function was improved or that wear was reduced.
However function started to gain more attention as a priority because 10-year
follow-up studies were available on some designs, showing failure rates of only a
few percent. Durability had therefore become a prerequisite in any design, which
it was thought could be attained by not deviating too far from the previous suc-
cessful design types. In reality, it was probably the quality of the polyethylene and
the improved cementing techniques that led to the durability, as much as the
designs themselves.
In the late 1990s, the ultracongruent design concept was introduced by Aaron
Hofmann and Roy Bloebaum from Salt Lake City (see Chap. 10 entitled “Posterior
Stabilized”). This design, with its “deep-dish” tibial surface, not requiring cruciate
preservation, was a direct challenge to the PS knee with its cam-post. The ultracon-
gruent gained some support but still did not alter the predominance of the standard
PS. It was generally considered to be too constrained with a higher risk of tibial
loosening and with a likely reduction in flexion range due to the height of the plastic
posteriorly and the lack of a mechanism for femoral “rollback” in flexion.
Nevertheless it did maintain a following and was increasingly regarded as a viable

design concept. The concept was embodied first in the Natural knee.
In the late 1990s, another new artificial knee design type, the medial pivot, was
introduced by Mike Freeman and Mike Tuke in the UK and by David Blaha in the
USA (see Chap. 11 entitled “Guided Motion Knees”). In this design, rather than
both the lateral and medial sides having the same amount of laxity, the medial side
was a ball-in-socket eliminating the anterior-posterior laxity altogether, while the
lateral side allowed almost unrestricted anterior-posterior laxity. There was biome-
chanical support for the concept in that it reproduced to a large extent the mechanics
of the normal knee as demonstrated by MRI studies on specimens and on the living
knee. At the same time, those studies did not show an absolutely fixed medial pivot,
but some deviation of the contact point on the medial side due to the deformability
of the medial meniscus and the lack of complete femoral-tibial conformity. In con-
trast, the medial pivot knee had a fully constrained ball-in-socket medially.
Nevertheless the resemblance to anatomic mechanics was an attractive feature,
together with the anterior-posterior stability it provided. Because of this, the medial
pivot was a viable option to the standard CR and PS designs. There were several
reasons why the knee did not achieve widespread adoption. In surgery, there was
some difficulty with soft tissue balancing, especially when the posterior cruciate
ligament was retained. The company which manufactured the design in the USA
was relatively small, without a prior background in artificial knees. But the major
factor was that the existing designs had a great deal of inertia, consisting of impres-
sive medium-term results from prominent institutions and individual surgeons. As a
result there was limited market penetration of the medial pivot, and this has per-
sisted up to this time.
Another factor which acted as a deterrent to the introduction of new design types
was the procedure specified by the Food and Drug Administration (FDA) to intro-
duce a new design to the market. Since 1976, a new device could be introduced to
the market if it could be shown to be substantially equivalent to a previously mar-
keted device. The FDA’s definition was: “A 510(k) requires demonstration of sub-
stantial equivalence to another legally U.S. marketed device. Substantial equivalence
means that the new device is at least as safe and effective as the predicate.” (This
regulation has recently been changed to a device which was introduced in the past
10 years.) This regulation made it relatively easy to introduce a new PS or CR knee,
even though mechanical testing was still necessary. If the new device was not sub-
stantially equivalent to a prior device, this meant complying with a complex pre-
market approval (PMA) process which included limited clinical trials, which was
expensive and which could take around 3 years or more to implement. This require-
ment undoubtedly limited the introduction of more innovative designs of artificial
knees. However a well-established company with a successful total knee system
already could decide to invest in a new design if it was considered to have a strong
market potential. The rotating platform design was an example of this, and indeed
the 7–9% of the US market, higher in certain other countries, is a sizeable market
share which justified the process.
Another factor which inhibited the emergence of a design other than the CR and
PS was the high development cost. It was necessary to have multiple sizes, right and
left, and many tibial thicknesses. Revision components also had to be provided,
which were expensive to produce and yet with a much smaller demand than the
primary knee. Hence if a company was designing and introducing a new system, the
designs that took precedence were the CR and PS. Nevertheless, the company and
the design group would often have the intention of including a “high-performance”
knee in the system, but this would be put “on the back burner” due to the sheer vol-
ume of work, time, and expense in dealing with the standard designs. In some cases,
the new standard designs were but modifications of the previous system, but with
enhancements based on the accumulated clinical experience and recent research. A
major advantage was that the new system could build on the long and successful
clinical record of the previous system. It would then be an easy transition for a sur-
geon to move to the new system. To maintain continuity, there could be similar
bearing surfaces between the new and previous designs, to the point where there
could be intermixing. On the other hand, a totally new design concept would be
unfamiliar, and many surgeons would be reluctant to use it immediately. This would
make the introduction slow, without any guarantee of it reaching economic levels.
Also linked to the high cost of developing a new design was the need for “leg-
acy,” being able to show that the design had a successful history in patients. This
would require a number of studies from different institutions, of a sizeable number
of patients, for a sufficient length of time. This led to the preference for improving
an existing design, rather than developing a new design which was substantially
different. There can be a fine line between an improvement and something novel.
Despite the above obstacles to the introduction of a new design or technique, if
the advantages were sufficiently convincing, there is the likelihood of gradual
acceptance even if the process might be slow. There are several designs already
available which are potential candidates. Unicompartmental knees have been avail-
able almost for as long as total knees, but some convincing reason would have to
emerge for an increase in usage to occur. The indications may have reached their
limit, although there is one study which suggested that 21% of arthritic knees which
were treated with a total knee could have been treated with a uni (Arno et al. 2011).
Robotic surgery for unis, with intraoperative planning to assure ideal placement and
soft tissue balancing, may increase usage to some extent. The impressive long-term
results of the Oxford group with their mobile bearing uni design have demonstrated
viability, but the results were achieved with an expert group, specializing in unis.
Few centers or individuals have so far taken that route. Using bi-unicompartmental
knees, in some cases with a separate patella flange, has the attraction of accurately
reproducing the original bearing surfaces (Romagnoli et al. 2015), but again there
are limited indications and the procedure is surgically challenging. A robotic tech-
nique could make the surgery easier.
The ACL + PCL type, sometimes called bi-cruciate retaining (BCR), has been
available since the early 1970s with the Townley and later the Cloutier. However
a ccurate component positioning is required so that all four ligaments function opti-
mally, requiring an anatomic design in sufficient sizes. Preserving the anterior
Fig. 15.3 W Norman

Scott MD was a Fellow of
John Insall in the early
1970s and contributed to
early knee designs as well
as editing several editions
of John Insall’s book
Surgery of the Knee.
Norman Scott was a
founding member of the
International Congress for
Joint Reconstruction
(ICJR) with worldwide
instructional courses
organized by local experts.
Norman focuses on
resident and fellow
training, not only through
ICJR but by his
organization of the Insall
Traveling Fellowship and
the Insall Scott Kelly
Fellowship. He is currently
a Professor at NYU
Langone Orthopedic
Hospital. (Photo taken at
the Knee Society Annual
meeting 2019) (© Peter S
Walker)
cruciate without problems has not been easy. Although there is evidence that excel-
lent results can be achieved, this has not been a consistent finding.
Rotating platform designs may have reached a stable level now that wear is no
longer regarded a major problem in CR and PS designs, but the freedom of rota-
tional position and better patella tracking are still regarded as advantages. It is pos-
sible that the early Eftekhar and UCI tibial designs with their rotatable tibial tracks
would function the same as a rotating platform, but with a lower cost of
manufacture.
Medial pivot designs still have potential based on positive results from many
users, but the smaller companies that produce these at present have much less mar-
keting force compared with the major established companies. Ultracongruent
designs are an attractive concept due to their mechanical simplicity and inherent
stability. Furthermore, the tibial inserts can conveniently fit into existing knee sys-
tems using the same tibial tray as a CR or PS. One company has produced a hybrid
design where a tibial insert has a more congruent medial side and a less congruent
lateral side, equivalent to a “loose medial pivot.” This may gain some traction
because of the convenience of being able to use it in an existing well-established
system.
There is a high potential for Guided Motion knees, where the goal is to restore
the anatomic characteristics of the normal knee joint, both in the geometry and in
the kinematics. This can be called a biomimetic approach. The guided motion knee
can achieve this due to the interaction of the femoral-tibial bearing surfaces without
requiring the cruciate ligaments. The collaterals are still required to ensure varus-
valgus stability. The guided motion knee can provide the stability normally pro-
vided by the cruciates, but with the same laxity as in the normal knee. While there
is only a small degree of laxity medially, on the lateral side, there is high laxity but
with a mechanism for inducing lateral roll forwards in extension and rollback in
flexion. Other advantages of a guided motion knee as described above are normality
of the “flow” of soft tissues around the anatomically shaped boundaries and the
anatomical “feel” to the kinematics.
This brings up the approach of custom implants, which does have a strong ratio-
nale. One question is the extent of customization which is needed. Is it sufficient to
match the patient’s main bone dimensions only, or do the components need to be
anatomically shaped also? Should the inherent stability of the bearing surfaces
match the ideal for that patient, and how is that to be measured? Wide adoption of
custom knees over time will likely depend upon showing improved functional
results of some type.
In the area of technique, standard mechanical instruments provide satisfactory
results and are generally reliable and easy to use, but there is still room for improve-
ment. They could be made much simpler and less expensive to make. The intramed-
ullary rod system for defining the frontal plane angle could be made more accurate
for all canal diameters. Centering the tibial cutting guide on the ankle could be made
more accurate. The use of smart tools is a low cost and convenient way of improving
performance of mechanical instruments.
Computer-assisted soft tissue balancing devices address a need in achieving the
goals of medial-lateral tightness in extension and flexion and have by no means
reached their full potential. It has been emphasized from the early days of knee
replacement that “it is a soft tissue operation” which means that no matter how the
components are positioned on the bones, unless the soft tissue tensions around the
knee are uniform during the full flexion range, the knee will not function optimally.
Inadequate soft tissue balancing can result in pain, tightness, a low flexion range,
and instability. Assessment of soft tissue tensions by the “feel” of the surgeon, no
matter how experienced the surgeon, has been shown to be inaccurate. Use of an
instrumented device shows that even small gap differences between the medial and
lateral sides result in sizeable differences in soft tissue tensions. This points to the
need for a quantitative means of assessment, which is likely to be more and more
recognized over time.
Computer-assisted navigation has been available for about 10 years and seems to
have reached a limit, unless it can be modified to make it easier and quicker to use.
However, experienced advocates of the system are likely to continue to use naviga-
tion as an accurate way of carrying out the surgery. There is currently a movement
towards robotics because of its many capabilities. For example, in New York State,
in 2008, only 4.3% of artificial knee surgeries used navigation or robotics, which
increased to 11.6% by 2015 (Boylan et al. 2018). Most of the use was in high-
volume hospitals, because most “early adopters” of new technology work there.
Robotic systems can offer surgical planning before or at the time of surgery, analy-
sis of the effect of varying bone cut levels and angles, soft tissue balancing assess-
ment, and validation of the surgical plan. It is not easy to predict the growth and
level of acceptance of robotic knee surgery because of the interplay between the
advantages and disadvantages. Other computer-assisted technologies such as virtual
reality await evaluation.
There are many factors involved with the successful introduction of a new artifi-
cial knee design or surgical technique, but certain criteria must be met. Anything
new has to compete with whatever is already available and therefore has to offer
demonstrable advantages. These advantages can apply to the patient (the most
important criterion), the surgeon, the hospital, and the health-care system as a
whole. Combining these, it is suggested that a new or improved artificial knee
design or technique needs to offer one or more of the following, if it is to be
successful:
• Improved durability
• Improved functional performance
• Quicker recovery time
• Increased flexion angle
• Reduced time, complexity, and cost of surgical procedure
• Research and educational support
Improved Durability
The survivorship of today’s designs is likely to be in the mid-90 percentile at

20 years, based on follow-up studies and joint registries. The first step in determin-
ing which factors will reduce failure is to examine the main causes. A recent paper
from the Hospital for Special Surgery (Pitta et al. 2018) reported that the main
causes of revision from 18,000 artificial knees performed between 2007 and 2012
were; in order of percentage; infection, instability, loosening, and stiffness.
Revisions for other reasons were at a much lower level. Infection is being addressed
by a universal decolonization protocol, control of operating room traffic, air disin-
fection, air recirculation, lavages, and antibacterial coatings on the implants. In
terms of the other factors, the authors pointed out that: “The reasons for failure for
primary TKA have evolved as implant designs, materials, and surgical techniques
have improved… New forms of polyethylene… have improved the material’s wear
behavior and thus decreased the rate of polyethylene-related failures.” In particular,
the reduction in wear rates of newer polyethylenes reduced the outright failures due
to wear-through as well as component loosening caused by polyethylene debris.
Those findings were in general agreement with other studies.
It is notable that the polyethylene improvements were widely accepted in a mat-

ter of a few years, rather than waiting for evaluations after many years of follow-up.
The initial demonstration of reduced wear was based on testing on simple recipro-
cating test machines, while specific knee designs were tested on knee simulating
machines, taking only a few months for 5 or 10 million cycles of testing. In reality
the wear test protocols do not fully simulate the variable activities typical of patients,
but since the tests compared the new material versus the previously used material,
that provided a high level of credibility. The main material improvement was poly-
ethylene with increased cross-linking, with processing methods to maintain the
mechanical properties. Later, the addition of vitamin E was shown to maintain a low
wear rate and, in addition, to reduce oxidation over time and therefore not be subject
to increased wear in the long term. While the results of such testing have been con-
sidered sufficient to justify clinical application, in vivo wear has been monitored at
progressive time intervals by retrieval studies (Rowell et al. 2015). Gradually, the
evidence is building to confirm the efficacy of these materials. An alternative way to
reduce polyethylene wear is by using oxidized zirconium-coated femoral compo-
nents, which showed reduced wear in knee simulator tests. In retrieval studies, there
was less damage on both the femoral and tibial components, validating the concept,
at least in the time frame of the study (Heyse et al. 2011). Another example of
reducing wear was a tri-condylar design where the femoral component and the tibial
baseplate were made from alumina ceramic, with a compression-molded polyethyl-
ene bearing, showing that ceramic was a promising alternate material for the femo-
ral component (Nakamura et al. 2017).
Tibial component loosening is now relatively infrequent with presently used
tibial components and techniques. However in order to sustain this, new fixation
post designs must follow known principles for success; otherwise early failure can
occur as has been found in recent years. Minimizing long-term bone loss due to
stress shielding can be avoided by avoiding excessive stiffness of the tibial base-
plate (Srinivasan et al. 2017). Adequate bonding or mechanical keying of pegs and
fins to the cement mantle can prevent interface motion and premature failure.
There are many cases where radiolucent lines develop over time, and while most
do not develop sufficient symptoms to require revision, some go on to failure. At
the same time, many of these cases may experience some pain on activity and
limitation of function. These problems can be addressed by meticulous cement
technique at surgery, where cancellous bone is exposed, fluid material is cleaned
from the surface, and sufficient pressurization is applied to obtain cement penetra-
tion to a depth of 2–3 mm across the entire interface, including the peripheries.
Uncemented fixation can potentially reduce long-term loosening should
cemented fixation be shown to be at risk in active patients. Assuring initial rigidity
of fixation is essential for the bone ingrowth to occur. Newer porous materials,
including those fabricated using 3D printing technologies, may further enhance
ingrowth capabilities. Some long-term loosening failures are due to misalignment,
most commonly where the knee has been in too much varus or where the joint line
has deviated too far from the horizontal. The misalignment has caused progressive
collapse of the bone with further misalignment and fixation failure. Such outliers of
Fig. 15.4 David Stulberg MD from Northwestern University, Chicago, has been a strong propo-
nent of the potential value of technology in arthroplasty from the 1990s when custom implants
were prominent, to today’s expansion of computer-assisted surgery. For many years David Stulberg
has been the Secretary General of the International Society for Technology in Arthroplasty (ISTA)
which has organized conferences around the world and is the largest society focusing on the tech-
nology area. (Photo taken at an ISTA meeting) (© Peter S Walker)
alignment could be reduced in number by improved surgical planning procedures

and by improved surgical techniques including computer-assisted surgery.
Instability can be due to the use of a relatively unconstrained knee in combina-
tion with inadequate soft tissue balancing (Wilson et al. 2017). Meticulous surgery,
using computer-assisted methods for bone cutting and balancing, can reduce the
instability. However, especially when mechanical instrumentation is used, artificial
knees which have built-in constraint, such as PS or guided motion types, will pro-
vide stability more consistently. At the same time, accurate soft tissue balancing is
still needed for varus-valgus stability. Stiffness, the opposite of instability, can again
be due to inadequate balancing or to incorrect placement of the components. In
addition, there are indications that biological factors play a role.
In summary, durability can potentially be improved by the use of low wear poly-
ethylenes, by tibial component designs which do not cause excessive stress protec-
tion of the underlying trabecular bone, by selection of implants with adequate
femoral-tibial stability, by a systematic technique when cement is used, by unce-
mented designs with rigid initial fixation, by accurate balancing techniques at sur-
gery, and by instrumentation which routinely provides accurate alignments.
Even then, it is hard to imagine more than a small gain in survivorship percent-
ages, due to unpredictable and outlier situations. Much of the potential gain in
durability may already be embodied in today’s designs and techniques. A further
point is that there is now more evidence of risk factors for implant failure (Bosco
et al. 2018) such as weak bone, smoking, overweight, and impact sports, allowing
behavior modification or pre-habilitation to be implemented prior to surgery and

hence reducing failures, especially those which might have occurred in the short
term.
Improved Functional Performance
In the previous chapter (Chap. 14 entitled “Comparisons and Evaluations”), meth-

ods for measuring function were discussed. At this time, PROMs are regarded as the
gold standard and are included in most joint registries. For a more objective evalua-
tion, biomechanical measures can be used. Thirdly, the function can be modeled
objectively as a property of the artificial knee in isolation or in combination with the
knee into which it is implanted. Many studies have shown that the factor which has
the greatest influence on the functional outcome is the overall physical condition of
the patient, in particular the muscular status of the lower extremities. For example,
in a study of 200 patients, half subjected to standard of care therapy and half to
progressive quadriceps-strengthening interventions, it was found that the latter
showed “enhanced clinical improvement in short and long-term, approaching the
functional level of healthy older adults. Conventional rehab did not yield similar
outcomes” (Petterson et al. 2009). In a later review of this type of therapy, it was
concluded: “Progressive exercise is critical to recovery after TKA. There is a large
decrease in quadriceps strength immediately after TKA, which is attributed to acti-
vation deficits and atrophy. This loss of strength is related to functional impairments
and biomechanical asymmetries. Progressive exercise protocols and early applica-
tion of neuromuscular electrical stimulation (NMES) should be used to attenuate
early quadriceps weakness and the associated impairments” (Pozzi et al. 2013). The
evaluation methods in the above studies, as well as PROMs, were the TUG test,
stair-climbing test, 6-minute walk test, and sit-to-stand balance test. Despite the
improved function that has been achieved with such high-intensity rehabilitation
methods, implementation involves increased physiotherapy resources, increased
cost, and compliance from the patient. The latter is impacted by the extra time com-
mitment as well as the discomfort of the therapy. Some of these disadvantages can
be overcome by the use of wearable activity monitors together with mobile applica-
tions (Chughtai et al. 2019). Such systems provide more motivation to the patient,
while monitoring the data to the physician can alert to any deficiencies. In the sys-
tem reported by Chughtai et al., KSS pain and function scores and WOMAC scores
improved substantially more than for conventional therapy.
An important question is to what extent the function can be improved by the
design of the artificial knee. In the previous chapter, several studies were cited
showing that there were no significant differences in PROMs, but yet there were
differences in biomechanical parameters. Designs which more closely simulated
normal anatomy, or which avoided excess anterior-posterior sliding, showed supe-
rior performance. To achieve a high level of function, Romagnoli et al. (2015) have
advocated the use of modular components including unicompartmental and patella
resurfacing. They point out that an accurate fit can be more readily achieved with
small components, compared with using conventional artificial knees: “…the versa-
tility of these implants is such that they can be adapted to almost all knees, allowing
the creation of custom-made prostheses that preserve, as much as possible, the
intact parts of the native knee… including the cruciate ligaments…” They propose
using modular components for broader indications than medial osteoarthritis.
Curved runners for the femoral condyles can be fitted so that the original surface
shapes of the femoral condyles are closely reproduced, given appropriate instru-
mentation and cutting tools. Special attention is needed to ensure reliable fixation of
the tibial components, but this is within the scope of today’s technologies. Despite
the resistance to expanding the use of even unicompartmental knees, the use of
small modular components should be considered further given the likely benefits to
function.
For artificial knees covering all of the bearing surfaces, custom knees designed
to match the bony geometry of the individual knee offer the possibility of improved
function due to maintaining the anatomic soft tissue envelope and muscle lever
arms. A convincing case has been made based on a dimensional analysis of 24,000
CT scans which showed that even for the multi-sized systems available today,
around 25% of cases would have errors of fit of the femoral component more than
3 mm (Meier et al. 2019). However, the custom knee design itself needs to be of a
type that would optimize function, but there have been some positive results from
recent studies.
Are there types of artificial knee designs available which will provide superior
function to the CR and PS? The alternate designs readily available today are
mobile bearing versions of the CR and PS, ACL + PCL preserving knees, ultra-
congruent knees, and medial pivot knees. So far ACL + PCL knees have not ful-
filled their promise according to a recent review paper (Barrett et al. 2017). While
acknowledging patient preference determined in the studies by James Pritchett,
this advantage has been offset by the surgical difficulties and other complications.
Medial pivot knees, also preferred by patients in the Pritchett studies, have shown
improved functional scores attributed to the AP stability. Ultracongruent knees,
including a medially congruent version, have shown at least equal functional
results compared with CR and PS, without loss of range of flexion. It has also
been shown in laboratory studies that a type of guided motion knee can provide an
even closer representation of anatomic laxity and stability, implying that the func-
tional performance should be further enhanced. Guided motion knees are at the
opposite extreme to the modular components in that the ideal blend of laxity and
stability is provided by mechanical interactions in the artificial knee. An advan-
tage of this mechanistic approach is that it requires less accuracy of surgery and
the artificial knee is therefore a more robust design. On the other hand, there could
be an issue with joint proprioception, although after reviewing the literature,
Takasaki et al. (2019) concluded that: “More high-quality studies are required to
achieve a comprehensive understanding of the effect of Proprioceptive
Neuromuscular Facilitation on the Joint Position Sense” while Wautier and
Thienpont (2017) found that proprioceptive testing with different types of total
knees was inconclusive. Implementation of both approaches, small modular com-

ponents and a guided motion knee, would require investment from companies and
objective monitoring of the functional results using both PROMs and biomechani-
cal methods.
Quicker Recovery Time
In the early days of knee replacement, continuous passive motion (CPM) was seen
as a potential method of restoring the motion of the knee and assisting with the
healing of surrounding soft tissues at their optimal physiological lengths (Salter
et al. 1980). Originally intended for healing cartilage defects, CPM was soon
applied to other applications including total knees. The patient’s leg was fitted in a
fixture which slowly flexed and extended the joint over a time period, repeated
several times each day. It seemed a logical way to speed up recovery, reduce pain
and discomfort, and achieve a higher flexion angle. Initial results showed less
thrombophlebitis and more rapid attainment of flexion after surgery, but with the
ultimate flexion angle being unchanged from conventional methods (Vince et al.
1987). After many years of trials, a review of 24 articles comprising 1445 cases
concluded: “CPM does not have clinically important effects on active knee flexion
ROM, pain, function or quality of life to justify its routine use” (Harvey et al.
2014).
Addressing the earlier time period after surgery, a major advance in the control
of pain and swelling was introduced for application to both artificial hips and knees
(Parvataneni et al. 2007). The idea behind the treatment was that local periarticular
SESSION TOPICS IN THE 36TH ANNUAL CCJR. WINTER 2019. INTERNATIONAL SOCIETY FOR TECHNOLOGY
(IN PARTNERSHIP WITH THE HIP AND KNEE SOCIETIES) IN ARTHROPLASTY - ANNUAL MEETING 2019
Ethics, Economics & OR Efficiency Innovations in TKA Kinematics
Why Knees Fail: Recognizing Causality Leads to Avoidance TKA Limb Alignment
The PCL in Contemporary TKA Modeling and Preoperative Planning
Examining Knee Arthroplasty Technologies Novel Technology for Treating Infections
The Medial Uni-Compartment Alternative: Fixed or Mobile TKA Wear and Retrievals
Mechanical Alignment Optimizes Patient Outcome TKA Robotics & Navigation Outcomes
The Cemented All Poly Tibia: Regardless of Age & BMI Ensuring Fixation of Arthroplasty Implants
The Medial Stabilized Knee: The Emerging Choice for TKA Clinical Pathways to Enhance Patient Outcomes
Periprosthetic Joint Infection Biomechanics and Preclinical Testing
Big Data: The Where, the What & the How TKA Tissue Balancing Techniques
The Unstable Knee: Manifestations, Reasons & Corrective Actions Rapid Recovery and Pain Management
Smart Trials: Taking Soft Tissue Balancing from an Art to a Science Role of Implant Design on TKA Biomechanics
Maximizing Cement Interdigitation During TKA Computational Models for Innovations in Arthroplasty
Patient Dissatisfaction Following TKA Design & Evaluation of 3D Printed Devices
Fig. 15.5 Typical program topics of a CCJR meeting organized by Seth Greenwald. The program
focuses on practical issues relevant to orthopedic surgeons specializing in artificial joints. This is
contrasted with a recent program of ISTA, where many of the topics relate to new developments
which are currently in the early stages of clinical evaluation; the attendance at the ISTA meeting is
typically divided between surgeons, researchers, and company staff
injections at various sites in and around the knee at the time of surgery would be
effective as a substitute for the then current perioperative pain control modalities.
“Analgesics and anti-inflammatories acting via different pathways is emphasized
for a multi-modal approach.” This new method was shown to have advantages
including lower pain scores from the first day onwards, quicker to straight leg raise,
going home earlier, and reduction of side effects. Adjunct measures are multimodal
pain management including regional blocks, periarticular injections, and opioid-
sparing medication protocols.
Also related to the postoperative period are programs such as the Perioperative
Orthopaedic Surgical Home (POSH) initiative. In a retrospective study of over 4000
total joint patients, patients who were assessed for modifiable risk factors and then
were treated accordingly were compared with patients who were not examined.
There was a significantly lower number of readmissions in the assessed group (Kim
et al. 2019).
In a review paper, Rutherford et al. (2017) discussed the multiple factors which
could improve the rehabilitation process. The first step was patient optimization:
“Recognising which patients are at risk for adverse outcomes after total knee is the
first step in preventing them.” This was accompanied by behavior modification and
support. Preoperative education was useful in preparing patients for the entire pro-
cess and what was expected of them. Intraoperative measures were important with
incisions which preserved quadriceps function being important. Rehabilitation was
recognized as being important, but “There is no consensus regarding the appropriate
frequency, duration or intensity of physical therapy protocols after TKA.” There is
clearly a need for additional research on methods for muscle strengthening and
coordination, which will overcome problems of patient compliance. Yet the pro-
longed time and discomfort of recovery period after an artificial knee remain as an
area for potential improvement. The tennis player referred to at the beginning of this
book, while able to walk after about 2 weeks and return to work shortly thereafter,
found it was more than 6 months before he was able to return to active sports. In the
early recovery period, he imagined 100 elastic bands around his knee giving the
sensation of tightness. Even now, several years later, he feels that the shock absorb-
ers are gone. A woman patient, whose functional requirements were less ambitious
than those of the tennis player, was nonetheless surprised and disappointed at the
level of discomfort after the procedure and the protracted recovery time. Even now
at 2 years after surgery, the range of flexion is only 100°, and she feels that the knee
is still not as good as her near-normal opposite side, especially in flexion range. This
is a perceptive observation because restoration to an almost normal condition is not
an unreasonable expectation. Retrospectively neither of these patients would have
changed their decision to undergo artificial knee surgery due to their severe pain and
disability. Their rehabilitation experiences were mirrored to a large extent by those
of Kevin Bozic MD, recounted in the monthly AAOS Now newsletter (2019). The
days after the procedure were “a lot more painful than I expected… I couldn’t even
travel in a car across town 2 weeks after surgery… I had pain past 6 weeks and
wasn’t fully recovered or functional by that time.” In his case, there was a positive
end result: “Less than a year after surgery, a hiking trip for 8–10 hours a day in rug-
ged terrain.”
Increased Flexion Angle
The normal knee achieves a high degree of flexion by the interaction of the bear-
ing surfaces and the menisci (Pinskerova et al. 2009). Firstly, there is a small
posterior displacement of the medial femoral condyle followed by lateral femoral
displacement down the slope of the lateral tibial plateau, the lateral menisci being
pushed posteriorly and downwards. Flexion from 120 to 160° was considered as a
separate arc, which the authors considered would be difficult to reproduce with an
artificial knee. The benefits of high flexion were clearly demonstrated in a study
of more than 5000 cases (Ritter et al. 2008). Pain, walking ability, and stair climb-
ing ability were enhanced for a flexion greater than 128° with a distinct falloff
below 118°.
The design of the tibial insert is expected to affect the maximum flexion angle
achieved. An insert with a high posterior lip could block the femur from flexing by
impingement against the posterior cortex. However in a clinical trial with the same
femoral component, using a high-lip ultracongruent insert, and a lower lip PS insert,
gave the same mean flexion angle of 116° at follow-up (Laskin et al. 2000). It has
been shown that artificial knee designs that allow for high flexion by modifications
to the posterior bearing surfaces of the femoral condyles and tibial insert do not
necessarily produce higher flexion. However if this type of implant is combined
with quadriceps releases from the anterior femur when necessary, high flexion val-
ues can be achieved (Tarabichi et al. 2010). In a series of 218 NexGen LPS-Flex
mobile bearing designs in 125 Asian and Middle East patients (most bilateral), 54
males and 71 females, average age 64 years, the initial average flexion angle was
125°; the range of flexion at 3–5 years follow-up was 140°. This does however need
be seen in the context of gain in flexion from pre-op, 15°. However it is still higher
than typical series of Western patients where the preoperative and postoperative
values are close to equal.
A study using fluoroscopy showed there were multiple factors affecting range of
flexion (Dennis et al. 2007): “over-stuffing of the patellofemoral joint during TKA
may have adverse effects on postoperative knee flexion… a pattern of progressive
posterior femoral translation with increasing knee flexion has been observed… in
our analysis amounting to about 1.4° flexion for every millimeter posterior femoral
translation.” Posterior tibial slope produced a similar increase in flexion. It is evi-
dent that these different factors are either design related, technique related, or a
combination. However there has not been a specific design change which has been
able to reproducibly increase flexion by a substantial amount, say at least 10°.
Nevertheless, high flexion remains an important goal for an artificial knee in that it
makes many activities of daily living much easier to perform. Identifying the factors
which affect the range of flexion and evaluating therapies which can increase flex-
ion is a worthwhile area of continuing research.
Reduced Time, Complexity, and Cost of Surgical Procedure
Considering the importance of this area, there is relatively little literature. A positive
start was made 25 years ago in a study which identified 42 steps which were carried
out during a typical procedure, from positioning the patient to closing the wound
(Dunbar and Gross 1995). Eight of the steps took significantly larger than the other
steps, but there was considerable variation in the times and between surgeons. The
need to repeat certain steps was an important factor. Robert Booth, from Jefferson
Health in Philadelphia, combined all facets of the procedure into a systematic work
flow (unpublished data). At the start of the day, the plan for each patient was dis-
cussed with the entire team. For the procedure itself, the assistants were thoroughly
familiar with the instruments and the order of the steps, so that going from one step
to the next was a seamless process. With this system the time per case was mini-
mized and the quality optimized. It would be an advantage if such a system could be
implemented in major centers, but the extensive range of procedures and the varying
staff assignments are barriers. There is benefit in the organization of the instruments
in the operating room (von Eiff et al. 2019). If the sizing of implants for each case
can be determined, the instrument trays can be supplied with far fewer instruments,
resulting in time savings in preparation, cleaning, and repackaging, as well as dur-
ing the procedure itself. This was termed size-specific trays as opposed to standard-
ized trays. The main conclusion was: “Performing 4 TKA’s in sequence generates
time saving of approximately 54 minutes which can be used for an additional
procedure.”
Can there be a reduction of time based on the type of artificial knee? Designs
that do not require meticulous accuracy and balancing will keep the time at the
lower end of the scale. Simpler instruments without unnecessary sub-components
and adjustments will reduce time and improve ease of surgery. The use of unce-
mented components could reduce the time, provided the bone preparation time was
the same as that for cemented components. The use of 3D printed instruments has
been considered and is available in a limited way with some designs (Hooper et al.
2019). There are now polymeric materials available which are biocompatible and
sufficiently strong and wear resistant to be used as cutting guides. The ideal is to
develop a set of single-use 3D printed instruments which could be used for any
design of artificial knee. The instruments would be provided in a pre-sterilized
pack or even sterilized by autoclave. This is feasible for the primary bone cuts of
the distal femur and proximal tibia, but is problematic for the femoral AP and
chamfer cuts which are in multiple sizes and different between manufacturers.
Furthermore the trial components need to be replicas of the actual final compo-
nents. The concept would however be feasible for a particular design of artificial
knee.
Can computer-assisted instrumentation reduce time, complexity, and cost? The
challenge ahead was expressed succinctly in an article entitled “The impact of smart
tools on total knee arthroplasty” (Scuderi 2007): “…there has been an exponential
rise in the number of TKA’s performed annually. The resulting demand… will
necessitate improving technology so that it can be used to treat more patients while
Fig. 15.6 A panel discussion held at the Computer-Assisted Orthopaedic Surgery (CAOS) annual
conference in New York in 2019. Such panels, with different specialists including those from
industry, are typical of today’s conferences. From left to right: Martin Roche, Chris Plaskos,
Branislav Jaramaz, Pierre Couture, Robert Cohen, Stefan Kreuzer, and Patrick Meere. (President
of CAOS 2019)
maintaining quality of care. Smart tools and robotics surgery may represent one
answer to this demand.” In other words the challenge of computer-assisted surgery
is to enable a faster procedure with greater accuracy. At this date (2019), there is
evidence that the computer-assisted smart tools have slightly reduced the surgical
time and improved the accuracy. However the robotic systems, while improving the
accuracy, have increased both the time and the cost of each procedure. It can also be
argued that complexity has been increased due to the higher technical skill and
training required both of the surgeon and the supporting staff. Imaging does allow
the surgeon greater flexibility of positioning and balancing of the implant compo-
nents, early evidence showing that this will result in better outcomes.
Another way in which computer-assisted surgery could make a difference is in
the concept of minimally invasive surgery. At present robots are being used to per-
form the same tasks as mechanical instruments. Could they be used in a new way
with newly designed components that could be inserted through very small
incisions?
Is there a possibility for a custom-made implant for every patient, made by 3D
printing either in the hospital itself or in a local facility? With newer bearing materi-
als, this might be possible in the not too distant future.
Research and Educational Support
The process by which new artificial knee designs and techniques become accepted
and used clinically is multifactorial, but the associated research studies and the
extent of publications and presentations at conferences are an essential ingredient.
This is typically a process which can take many years. Embedded within this are the
separate factors which determine the end result. These factors were listed at the
beginning of this chapter for the success of designs up to 1980: prestigious institu-
tion, authoritative designer(s), collaboration between surgeon and bioengineer,
reputable design group, well-established company, strong company support, con-
vincing design rationale, well-thought-out technique, and minimum problems in
early trials. These are all relevant today, with a greater emphasis on design teams
rather than on individuals and on a greater role for companies, particularly where
large investments are needed such as in robotic surgery. Basic research plays a role,
especially in new materials development or in biomechanical evaluations, while
large databases such as the national registries rely on new statistical methods,
including machine learning. There is a high output of research studies from univer-
sities and hospitals, funded from a wide variety of sources. Such output provides
most of the content for the lectures at conferences and publications. Based on the
increasing numbers of papers being written, it might be thought that the funding for
such work is adequate. However there is justification for a higher level of funding
from the NIH, where most of the funding for orthopedics in general is derived.
Studies of the biomechanics of the knee can be considered fundamental, which
provides a basis for the design of treatments for the arthritic knees. Devising experi-
mental methods or computer models for the evaluation of artificial knees is an area
which can minimize problems and enhance clinical results. New materials for the
bearing surfaces, for the structure of the components, and for uncemented fixation
all require basic research studies. The possibility of economically fabricating cus-
tom joints for every patient requires research in computer technology and manufac-
turing methods. The crucially important area of recovery and rehabilitation needs
further basic research.
Today, there are several educational vehicles for the presentation and evaluation
of research on artificial knees. The American Academy of Orthopaedic Surgeons
(AAOS), the largest orthopedic conference in the world, provides a holistic forum of
lectures, instructional media, scientific exhibits, and posters which include basic and
applied research studies and displays of products by companies. The annual meeting
of the American Association of Hip and Knee Surgeons (AAHKS) is a scientific
forum for implant design, outcomes, surgical technique, complications, and socio-
economic issues. The courses held during the year by the International Congress for
Joint Reconstruction (ICJR) and the Current Concepts in Joint Replacement (CCJR)
present the state of the art of the field, intended for the specialist surgeon in artificial
knees, as well as other joints. The CCJR includes debate scenarios on such topics as
CR vs PS, cemented fixation vs porous and coated, whether or not to resurface the
patella, and mechanical vs kinematic alignment. The presentations usually recognize
that each alternative is valid, but it provides the surgeon with a deeper understanding
of the nuances to enable a rational choice to be made. The ICJR often includes live
interactive surgeries which can reveal fine points and ways of overcoming problems
and choices which might arise. Most companies run instructional laboratories where
expert surgeons demonstrate new designs and techniques.
For the most part, in the above organizations, the large majority of the attendees
are orthopedic surgeons. In contrast, in the field of artificial joints, there are two
societies which represent more technical aspects and the boundary between today
and tomorrow, ISTA and CAOS. The general purpose of these societies is to show-
case and evaluate new possibilities, whether it be in methods of outcome measure-
ment, new materials and designs, or alternate surgical techniques, particularly
computer-assisted. The attendance at these conferences is divided roughly into
thirds: orthopedic surgeons, researchers and academics, and company representa-
tives. These meetings are invaluable in that they provide a multidisciplinary per-
spective of what might eventually enter the mainstream of joint replacement practice
or find a useful niche in a special application. One of the features of the meetings is
that the technical aspects can be covered in considerable depth. An important issue
is the interface between such meetings and those which deal mainly with surgical
practice. To some extent there is a dividing line where practicing surgeons do not
encounter the scientific or developmental activity. This is to a large degree inevita-
ble due to time constraints alone, although some understanding of basic biome-
chanical principles should be regarded as essential, in order to appreciate the
functional aspects of artificial knees. The main interface needs to be with those
orthopedic surgeons who are considered the opinion leaders and who make presen-
tations at clinical conferences. Such individuals provide an objective interface in
transferring appropriate new technology to a broader front.
The ORS is also involved in the area of new technology, attracting scientific
researchers mainly from academia. Computer modeling is currently a major area of
interest for determining the relative influence of variables such as geometry and soft
tissue balancing and on its potential for personalizing artificial knee design. Recently
an Implants Section was formed which has separate sessions, as well as the sessions
on implants in the general program. However, there is a greater focus on biology at
the ORS rather than on the physical sciences. A debate was recently held at the ORS
on the alternate solutions of biology versus mechanical for osteoarthritis of the knee
and hip. This is detailed below.
Biological Treatments
The possibility is often brought up that biological treatments will solve arthritis and
that artificial knees will be unnecessary. In 2017, the Orthopedic Research Society
held a debate entitled “Will Regenerative Medicine Make Orthopaedic Implants
Obsolete in Our Time?” (Johnstone et al. 2018). The title could apply to all types of
implants although the main focus was on artificial hips and knees.
For the proposition, Brian Johnstone stated as an overview: “The reality is that
orthopedic surgeons, and the orthopaedic industry, had better become smarter
biologists soon, because we are underestimating the myriad of things that regenerative
medicine will do in the long run.” He defined the purpose of regenerative medicine as
“the repair, regrowth, or replacement of damaged or diseased cells, tissues, or organs.”
With gene therapies, cell therapies, tissue engineering, and production of artificial
organs, “it is hard for me to believe that regenerative medicine will not make plastic
and metal implants obsolete in our time.” “We have to transition from traditional engi-
neering and metallurgy-based therapies to biology and biomedical engineering-based
ones.” Johnstone summarized as follows: “… metal and plastic orthopedic implants
are without question, a success story… they have plateaued in terms of significant
improvement, providing a stimulus for developing new technologies.”
Against the proposition, Joshua Jacobs stated: “Fabricating complex biological
structures to replace or restore diseased organs has been the holy grail for decades…
for orthopedic applications such as total joint replacement, there are no viable
regenerative medicine alternatives at the present time.” Jacobs then emphasized the
high performance of artificial joints: “… total joint arthroplasty with engineering
materials has set a very high bar for performance that will be difficult for regenera-
tive medicine to exceed.” Why should it not be the case that a structure made from
durable synthetic materials will surpass biological materials in a system such as a
knee joint? Jacobs also pointed out that the arthritis was, in most cases, caused by a
deficient biomechanical system: “If the aberrant biomechanics, whether static or
dynamic, are not corrected, tissue engineered constructs will likely be overloaded
and degraded as was the native tissue.” This crucial last point was made by Brandt
et al. (2008), after decades of research into osteoarthritis: “The importance of abnor-
mal joint mechanics in the etiopathogenesis of OA cannot be overemphasized. The
authors view OA as a process that is attempting to contain a mechanical problem in
the joint; OA is best defined as failed repair of joint damage that has been caused by
excessive mechanical stress. Because the body’s innate mechanisms for repairing
damaged cartilage cannot deal with the underlying mechanical abnormality, they
cannot solve the problem of OA… The common mechanical factor underlying OA
is a pathologic increase in intra-articular stress.”
Concluding Points
Today, there are well-established designs and techniques, a steadily increasing

demand worldwide, and good functional results with few complications. In this
chapter, possibilities of further advances have been discussed based on evidence
from recent published studies. These advances would improve the durability and
functional performance, provide a quicker recovery time with an increased flexion
angle, and require less time for a more accurate procedure. The most likely possi-
bilities, based on the advantages that will be obtained, are a gradual shift in the
design types used most frequently, from the CR and PS to designs such as ultracon-
gruent, medially congruent, medial pivot, and guided motion; more use of unce-
mented implants with porous coatings; and a change from manual instrumentation
to manual plus smart tools and to robotic surgery, although the latter may be focused
on major centers and for young active patients. If such changes do occur, they are
Fig. 15.7 Chitranjan S Ranawat MD has been a highly recognized surgeon at the Hospital for
Special Surgery in New York since the late 1960s. Since being one of the designers of the first
condylar knees in the early 1970s, he has continued to be a major force in the areas of design,
surgical technique, clinical follow-up, and pain management to this day. He was the main founder
of the Knee Society in 1983. His educational excellence to residents and fellows and at numerous
conferences, has been unparalleled. (Photo taken 2018) (© Peter S Walker)
likely to take a number of years to take effect due to the time taken to reach a con-
sensus that there are worthwhile benefits without any downsides. The way in which
such changes will take effect is through the well-organized and extensive research
and educational infrastructure. The increasing use of biomechanical evaluation
methods will play an important role in accelerating the decision-making, in provid-
ing more convincing functional data. Companies too will play a role in funding and
promoting the designs and techniques which are the most successful. Finance will
however play a pervasive role through the system, with some conflict between
obtaining high quality but without an increase in cost.
Are there likely to be any revolutionary changes in treating the artificial knee?
Custom knees and disposable instrumentation could be designed and made on-site
using a CNC machine for polyethylene tibial components and 3D printing for PEEK
femoral components. Implant manufacturing units would be a product sold to hospi-
tals. The regulatory and liability questions are however formidable, while on-site tech-
nical expertise would be essential. Furthermore, the functional benefits would have to
be well demonstrated. Components made from deformable polymeric materials such
as polyurethane or hydrogels are another possibility for making joints which are
deformable, allow fluid film lubrication, and giving a more natural feel to the patient,
but the wear and strength properties required may be beyond what is possible at this
time. Minimally invasive surgical techniques, widely used for internal medicine, could
result in a procedure with a much more rapid recovery and higher flexion, but this
would require a major change in the design of the components and the way they are
fitted to the femur and tibia. Could arthritis be prevented in many cases by detection at
an early stage and carrying out a procedure to treat a biomechanical abnormality such
as misalignment? Or if the knee had a ligament or meniscal injury, could the treatment
be restricted to those tissues rather than the entire joint? There is evidence from the
sports medicine area that this approach has some possibilities. Whatever changes are
made in the next few years or even the next decade, based on the changes over the past
three decades since 1980, further changes may take considerable time, but one can but
expect that the best solutions will come to the forefront and that new possibilities will
be explored.
References
Aglietti P, Buzzi R, De Felice R, Giron F. The Insall-Burstein total knee replacement in osteoar-
thritis: a 10-year minimum follow-up. J Arthroplast. 1999;14(5):560–5.
Arno S, Maffei D, Walker PS, Schwarzkopf R, Desai P, Steiner GC. Retrospective analysis of total
knee arthroplasty cases for visual, histological, and clinical eligibility of unicompartmental
knee arthroplasties. J Arthroplast. 2011;26(8):1396–403.
Atrey A, Edmondson MC, East D, Miles K, Butler-Manuel A, Ellens N. A retrospective medium-
to long-term results of 1500 AGC total knee replacements - an independent Centre functional
follow up and survivorship. J Orthop. 2014;11(1):37–42.
Barrett TJ, Shi L, Parsley BS. Bicruciate-retaining total knee arthroplasty, a promising technology,
that’s not quite there. Ann Transl Med. 2017;5(Suppl 1):S17.
Bosco JA, Harty JH, Iorio R. Bundled payment arrangements: keys to success. J Am Acad Orthop
Surg. 2018;26(23):817–22.
Boylan M, Suchman K, Vigdorchik J, Slover J, Bosco J. Technology-assisted hip and knee arthro-
plasties: an analysis of utilization trends. J Arthroplast. 2018;33(4):1019–23.
Bozic KJ. My own joint replacement was a reality check. AAOS Now, Aug 2019.
Brandt KD, Dieppe P, Radin EL. Etiopathogenesis of osteoarthritis. Rheum Dis Clin N Am.
2008;34:531–59.
Chughtai M, Kelly JJ, Newman JM, et al. The role of virtual rehabilitation in total and unicompart-
mental knee arthroplasty. J Knee Surg. 2019;32(01):105–10.
Dennis DA, Komistek RD, Scuderi GR, Zingde S. Factors affecting flexion after total knee arthro-
plasty. Clin Orthop Relat Res. 2007;464:53–60.
Dunbar M, Gross M. Critical steps in total knee arthroplasty. Int Orthop. 1995;19(5):265–8.
Evans JT, Walker RW, Evans JP, Blom AW, Sayers A, Whitehouse MR. How long does a knee replace-
ment last? A systematic review and meta-analysis of case series and national registry reports
with more than 15 years of follow-up. Lancet (London, England). 2019;393(10172):655–63.
Harvey LA, Brosseau L, Herbert RD. Continuous passive motion following total knee arthroplasty
in people with arthritis. Cochrane Database Syst Rev. 2014;(2):Cd004260.
Heyse TJ, Chen DX, Kelly N, Boettner F, Wright TM, Haas SB. Matched-pair total knee arthro-
plasty retrieval analysis: oxidized zirconium vs. CoCrMo. Knee. 2011;18(6):448–52.
Hooper J, Schwarzkopf R, Fernandez E, Buckland A, Werner J, Einhorn T, Walker
PS. Feasibility of single-use 3D-printed instruments for total knee arthroplasty. Bone Joint J.
2019;101-B(7_Supple_C):115–20.
Johnstone B, Jacobs JJ, Sandell LJ, Wilkinson LM. Regenerative medicine will make orthopaedic
implants obsolete in our time. J Orthop Res. 2018;36:2579–85.
Kim KY, Anoushiravani AA, Chen KK, Li R, Bosco JA, Slover JD, Iorio R. Perioperative orthope-
dic surgical home: optimizing total joint arthroplasty candidates and preventing readmission. J
Arthroplast. 2019;34(7S):S91–6.
Laskin RS, Maruyama Y, Villaneuva M, Bourne R. Deep-dish congruent tibial component use in
total knee arthroplasty: a randomized prospective study. Clin Orthop Relat Res. 2000;380:36–44.
Mahfouz M, Fatah EEA, Bowers LS, Scuderi G. Three-dimensional morphology of the knee
reveals ethnic differences. Clin Orthop Relat Res. 2012;470:172–85.
Meier M, Zingde S, Steinert A, Kurtz W, Koeck F, Beckmann J. What is the possible impact of
high variability of distal femoral geometry on TKA? A CT data analysis of 24,042 knees. Clin
Orthop Relat Res. 2019;477(3):561–70.
Nakamura S, Ito H, Nakamura K, Kuriyama S, Furu M, Matsuda S. Long-term durability of ceramic
tri-condylar knee implants: a minimum 15-year follow-up. J Arthroplast. 2017;32(6):1874–9.
Parvataneni HK, Shah VP, Howard H, Cole N, Ranawat AS, Ranawat CS. Controlling pain after
total hip and knee arthroplasty using a multimodal protocol with local periarticular injections:
a prospective randomized study. J Arthroplast. 2007;22(6 Suppl 2):33–8.
Petterson SC, Mizner RL, Stevens JE, et al. Improved function from progressive strengthening
interventions after total knee arthroplasty: a randomized clinical trial with an imbedded pro-
spective cohort. Arthritis Care Res. 2009;61(2):174–83.
Pfitzner T, Perca C, Roth P. Unicompartmental vs total knee arthroplasty for medial osteoarthritis.
Z Orthop Unfall. 2017;155(5):527–33.
Pinskerova V, Samuelson KM, Stammers J, Maruthainar K, Sosna A, Freeman MA. The knee in
full flexion: an anatomical study. J Bone Joint Surg. 2009;91(6):830–4.
Pitta M, Esposito DCI, Li DZ, Lee MDY, Wright TM, Padgett DE. Failure after modern total knee
arthroplasty: a prospective study of 18,065 knees. J Arthroplast. 2018;33(2):407–14.
Pozzi F, Snyder-Mackler L, Zeni J. Physical exercise after knee arthroplasty: a systematic review
of controlled trials. Eur J Phys Rehabil Med. 2013;49(6):877.
Pradhan NR, Gambhir A, Porter ML. Survivorship analysis of 3234 primary knee arthroplasties
implanted over a 26 year period. Knee. 2006;13:7–11.
Pritchett JW. Bicruciate-retaining total knee replacement provides a satisfactory function and
Ritter MA, Lutgring JD, Davis KE, Berend ME. The effect of postoperative range of motion
on functional activities after posterior cruciate-retaining total knee arthroplasty. JBJS.
2008;90(4):777–84.
Romagnoli S, Marullo M, Massaro M, Rustemi E, D’Amario F, Corbella M. Bi-unicompartmental
and combined uni plus patellofemoral replacement: indications and surgical technique. Joints.
2015;3(1):42–8.
Rowell SL, Reyes CR, Malchau H, Muratoglu OK. In vivo oxidative stability changes of highly cross-
linked polyethylene bearings: an ex vivo investigation. J Arthroplast. 2015;30(10):1828–34.
Rutherford RW, Jennings JM, Dennis DA. Enhancing recovery after Total knee arthroplasty.
Orthop Clin North Am. 2017;48(4):391–400.
Salter RB, Simmonds DF, Malcolm BW, Rumble EJ, MacMichael D, Clements ND. The biologi-
cal effect of continuous passive motion on the healing of full-thickness defects in articular car-
tilage. An experimental investigation in the rabbit. J Bone Joint Surg Am. 1980;62(8):1232–51.
Sartawi M, Zurakowski D, Rosenberg A. Implant survivorship and complication rates after total
knee arthroplasty with a third-generation cemented system: 15-year follow-up. Am J Orthop
(Belle Mead, NJ). 2018;47(3). https://doi.org/10.12788/ajo.2018.0018
Scott CEH, Bell KR, Ng RT, MacDonald DJ, Patton JT, Burnett R. Excellent 10-year patient-
reported outcomes and survival in a single-radius, cruciate-retaining total knee arthroplasty.
Knee Surg Sports Traumatol Arthrosc. 2019;27(4):1106–15.
Scuderi G. The impact of smart tools on total knee arthroplasty. Am J Orthop (Belle Mead NJ).
2007;36(9 Suppl):8–10.
Srinivasan P, Miller MA, Verdonschot N, Mann KA, Janssen D. Strain shielding in trabecular bone
at the tibial cement-bone interface. J Mech Behav Biomed Mater. 2017;66:181–6.
Takasaki H, Okubo Y, Okuyama S. The effect of proprioceptive neuromuscular facilitation on
joint position sense: a systematic review. J Sport Rehabil. 2019:1–30. https://doi.org/10.1123/
jsr.2018-0498
Tarabichi S, Tarabichi Y, Hawari M. Achieving deep flexion after primary total knee arthroplasty.
J Arthroplast. 2010;25(2):219–24.
Vince KG, Kelly MA, Beck J, Insall JN. Continuous passive motion after total knee arthroplasty.
J Arthroplast. 1987;2(4):281–4.
Von Eiff MC, von Eiff W, Roth A, Ghanem M. Process optimization in total knee arthro-
plasty procedures: impact of size-specific instrument sets on costs and revenue. Orthopade.
2019;48(11):963–8.
Wautier D, Thienpont E. Changes in anteroposterior stability and proprioception after different
types of knee arthroplasty. Knee Surg Sports Traumatol Arthrosc. 2017;25(6):1792–800.
Wilson CJ, Theodoulou A, Damarell RA, Krishnan J. Knee instability as the primary cause of fail-
ure following Total knee arthroplasty (TKA): a systematic review on the patient, surgical and
implant characteristics of revised TKA patients. Knee. 2017;24(6):1271–81.
Glossary
Many of the definitions are derived from the AAOS Comprehensive Orthopedic
Review, Biomechanics and Biomaterials.
Some definitions are worded to apply specifically to the knee joint.
Where applicable units of measurement are shown next to the term.
Units Units of measurement using the Standard

International (SI) System are:
• Length = meter m (or millimeter mm).

Mass = Kilogram kg. Time = second s.
Force = Newtons N. Pressure = Pascal
Pa. N/mm2.
• 1 inch = 25.4 mm. 1 lb = 0.454 kg.
1 pounds force = 4.45 Newtons.
100 Newtons = 22.5 lbs force. 1
MegaPascal (MPa) = 145 pounds per
square inch (psi).
• body weight (BW) typical male 170 lbs
approx. 750 Newtons
AAHKS American Association of Hip and Knee

Surgeons
AAOS American Academy of Orthopedic Surgeons
Abrasive wear Wear due to an asperity on one surface dig-
ging into the opposite surface releasing
small particles. Three-body abrasion from
loose particles intervening between sliding
surfaces.
Acceleration Rate of change of velocity with time

P. S. Walker, The Artificial Knee, https://doi.org/10.1007/978-3-030-38171-4
324 Glossary
Adhesive wear Wear at a microscopic location on the sur-

face due to adhesion against the opposite
surface, releasing a small wear particle
AJRR American Joint Replacement Register
All-plastic A component made completely of polyeth-
ylene or another polymer, usually applied to
a tibial component
Anatomic axis A line drawn on a bone in 2D or 3D showing
the centerline of the bone based on defined
end points
Anterior cruciate The ligament in the center of the knee which
limits the displacement of the femur posteri-
orly relative to the tibia or of the tibia anteri-
orly relative to the femur
Anterior displacement The motion of one bone relative to another,
along an axis defined in the anterior
direction
Anterior The front of the knee or in a frontal
direction
Arthritis A painful degenerative condition of the knee
where the cartilage bearing surfaces are
worn, often accompanied by limb
misalignment
Arthroplasty The surgical repair of a joint which can
involve the use of an artificial part or implant
Artificial knee Implantable device made from synthetic
materials to restore function to an arthritic
knee. Most commonly referred to a condylar
replacement knee (including a unicompart-
mental) but can include a hinged device
ASTM American Society for Testing and Materials
Axial force Vertical force applied in the longitudinal
direction of the anatomic axis of a bone such
as the femur or tibia
Axis system xyz Cartesian axis system fixed in the femur
or tibia defined by bony landmarks or an
axis system defined in any body. The axes
are mutually perpendicular
Bicruciate retaining (BCR) An artificial knee design where both of the
cruciate ligaments are preserved
Biomechanics Mechanical principles applied to the muscu-
loskeletal system
Biomimetic Imitating the structure or function of a bio-
logical system. Close to biomimicry
Glossary 325
Body A structure such as a bone, implant, compo-

nent. Can be depicted for the purpose of
analysis
Brittle The absence of a region of plastic deforma-
tion up to the point of fracture
Cam-post Applied to a posterior stabilized knee where
a projecting tibial post fits into a cavity in
the center of the femoral component and is
interacted on by the cavity to control motion.
Cancellous bone A network of interconnected bony spicules
to resist forces in functional directions, also
called spongy or trabecular bone
CAOS Computer-Assisted Orthopedic Surgery. An
international society
CCJR Current Concepts in Joint Replacement. An
educational organization
Ceiling effect Lack of discriminating ability of the evalua-
tion test to detect clinically relevant changes
even if the patient continues to improve
Collateral ligaments The fibrous tissue structures at the sides of
the knee which resist varus and valgus rota-
tions between the femur and the tibia
Compression The application of a force at a point or over
an area of a body
Compressive stress The force applied over a defined area,
divided by the area
Computer-assisted surgery Surgery where one or more of the surgical
instruments or guides uses computer tech-
nology, including navigation and robotics
Condylar replacement Components for resurfacing an arthritic
knee. Can cover only one half of the knee
(uni) or all of the bearing surfaces (total
knee)
Condyles The regions on the distal femur and proxi-
mal tibia which are covered with articular
cartilage and which transmit the forces and
motions
Congruent Usually applied to a joint where the oppos-
ing bearing surfaces are a close fit, such as
the ball and socket of the hip
Constraint The phenomenon whereby one body is
restricted from moving relative to another
body or the extent by which it is restricted
Contact area The area of the contact between two surfaces
which are loaded together, such as the con-
tact area in the knee or in an artificial knee
326 Glossary
Contact stress The force per unit area on a contact region

such as the contact area in the knee
Coordinate system An xyz Cartesian axis system fixed in the
femur or tibia defined by bony landmarks or
an axis system defined in a body, where all
axes are mutually at 90 ° to each other
Coronal plane A plane representing the front view of the
limb, or knee, or tibia or femur, or the entire
body (same as frontal plane)
Cortical bone Solid bone material comprising the tubular
diaphyseal sections of the femur and tibia,
also located on the outside surfaces of can-
cellous bone in the metaphysis
CR Type of artificial knee. Designation for a
condylar replacement artificial knee where
the anterior cruciate is resected but the pos-
terior cruciate is retained
Cross-shear When two opposing sliding surfaces move
in slightly different directions in succes-
sive motion cycles, causing increased
wear
Cruciate ligaments The fibrous tissue bands in the center of the
knee which resist relative anterior and poste-
rior displacements between the femur and
the tibia
Debridement Removal of damaged cartilage, soft tissue,
or bone, from inside the joint to facilitate
motion and relieve pain, usually as an
arthroscopic procedure
Deformable body An object which changes its shape, or
deforms, when forces are applied (see also
rigid body)
Degrees of freedom (dof) The number of directions where motion can
take place. A body in space has 3 linear and
3 rotational dof
Delamination wear Wear due to the coalescing of subsurface
cracks due to subsurface stresses, leading to
release of particles or flakes
Dislocation When one surface of a joint crosses the
boundaries of the opposite surface such that
it reaches an unstable state
Displacement Linear motions between one object and
another or the distance between the initial
and the final position of an object
Glossary 327
Distal The lower part of the object, e.g., the femur

or the tibia
EFORT European Federation of Orthopedics and
Traumatology
Elastic modulus A material property describing its resistance
to change of shape when loaded.
Colloquially, the “stiffness” of a material
Elastic The property of a material that allows it to
regain its shape immediately after a force is
released
Electrogoniometer A device that fits around a joint, which has
sensors to measure the angle of flexion or
other angles of motion.
Electromyography A diagnostic procedure to assess the health
of muscles and the nerve cells that control
them
EORS European Orthopedic Research Society
Equilibrium A state of a body where the sum of the forces
and moments are zero
ESB European Society of Biomechanics
Euler or Cardan angles An ordered sequence of 3 rotations of the
femur relative to the tibia or tibia relative to
femur
Extension The relative position of the femur and tibia
where the bones cannot be further extended
(note, the position can vary slightly depend-
ing on how much extending moment is
applied)
External rotation Rotation of the femur about the anatomic
axis of the tibia such that the medial femo-
ral condyle is moved anteriorly and the lat-
eral condyle posteriorly. Can also be defined
as a rotation of the tibia about its anatomic
axis such that the medial plateau moves
anteriorly and the lateral plateau moves
posteriorly
Failure A fracture of the body or material. Excessive
permanent deformation can also be regarded
as failure
Fatigue Eventual failure of a body or material after
multiple applications of a force
Finite element analysis Analysis of the stresses and strains in an
object by dividing it into multiple small
328 Glossary
shapes and applying loads and boundary

conditions
Fixation peg One or more pegs on the base of the tibial or
femoral component used for enhancing the
fixation of the component to the bone
Fixation The connection of an implant component
(e.g., femoral or tibial) to the host bone, typ-
ically on a prepared surface exposing can-
cellous bone
Fluoroscopy Medical imaging where an x-ray image is
continuous, commonly used for studying the
motion inside human joints
Force (Newtons) An entity applied to a body which causes it
to move or change its shape. Force has both
magnitude and direction
Free body diagram Depiction of all the forces and moments act-
ing on an object contained within a defined
boundary
Fretting Surface damage at contacting surfaces
which experience small relative motions
under repetitive loading
Friction coefficient The ratio of the resistive force to sliding,
divided by the force pressing the surfaces
together
Friction The resistive force between the surfaces of
two objects in contact when relative sliding
occurs
Frontal plane A plane representing the front view of the
limb, or knee, or tibia or femur, or the entire
body (same as coronal plane)
Guided Motion Knee A type of artificial knee where motion is
guided or constrained by mechanical means,
usually to provide more anatomic motion
Hardness Resistance to indentation or scratching
ICJR International Congress for Joint Replace-
ment. An educational organization
Instant center of rotation The center of rotation on a defined plane
between two bodies for a small arc of motion
Intercondylar The space in between the distal femoral con-
dyles, also called intercondylar notch or
fossa
Internal rotation Rotation of the femur about the anatomic
axis of the tibia such that the medial femo-
ral condyle moves posteriorly and the lat-
eral condyle anteriorly. Can also be defined
Glossary 329
as a rotation of the tibia about its anatomic

axis such that the medial plateau moves
posteriorly and the lateral plateau moves
anteriorly
Intramedullary Inside the hollow tubular diaphyseal portion
of the femur or tibia
ISTA International Society for Technology in
Arthroplasty
ISO International Standards Organisation
Kinematics A description of the motion of an object
without regard for how that motion is
produced
Kinetics A study of the forces that cause motions
KSS Knee Society Score. One of the most widely
used scoring system for evaluating the func-
tional performance of a knee
Lateral collateral ligament The ligament at the lateral side of the knee
which limits the varus rotation between the
femur and tibia
Lateral The outside of the knee
Laxity The displacements and rotations which
occur between the femur and tibia when
forces and torques are applied, usually at a
fixed angle of flexion
Ligament release Removing a part of a ligament from its bony
attachment or cutting some fibers of the liga-
ment, to reduce the stiffness or tightness
Linearly elastic A material or body where the deformation is
linearly proportional to the applied force
Load A colloquial term meaning the same as force
Lubrication Separation of two sliding surfaces by a fluid
film or boundary molecules to prevent or
minimize the wear of the surfaces
Mass (Kg) The amount of matter that an object con-
sists of
Maximum flexion 180 ° minus the angle between the anatomic
axes of the femur and tibia with the knee
maximally flexed
Mechanical axis of the femur The line joining the center of the femoral
head to the center of the distal femur
Mechanical axis of the lower limb The line joining the center of the femoral
head to the center of the ankle at the level of
the malleoli
330 Glossary
Medial collateral ligament The ligament at the medial side of the knee
which limits the valgus rotation between the
femur and tibia
Medial On the inside of the knee joint, facing the
opposite knee
Medial pivot A type of artificial knee where the medial
condyles are congruent and the lateral con-
dyles are incongruent. Such that internal-
external rotation pivots on the medial side
Meniscus Crescent-shaped spacers between the medial
and lateral sides of the femur and tibia which
make the condyles more congruent
Micromotion Microscopic (micron ranges) motion
between two bodies or surfaces which can
cause corrosion or wear particles. Takes
place between a plastic tibial insert and the
inside base of the metal tibial tray
Mold arthroplasty The insertion of an inert mobile mold to
guide repair between two perfectly shaped
joint surfaces
Moment arm The perpendicular distance between the
applied force and a defined point, or between
the force and the axis of rotation
Moment The product of the magnitude of the force
(Newtons) and its distance from a defined
point (millimeters)
Monoblock A component which is made from two or
more separate parts but is fused together,
e.g., a plastic tibial insert fused to a metal
baseplate
OREF Orthopaedic Research and Education
Foundation
ORS Orthopaedic Research Society
Osteotomy Cutting a wedge-shaped piece of bone from
the distal femur or proximal tibia to realign
the bone and limb
Oxidation Oxidation of polyethylene leads to embrit-
tlement and higher rates of wear including
delamination
Paradoxical sliding Abnormal and excessive anterior sliding
of the femoral component on the tibial
component during a mid-flexion range of
motion
Glossary 331
Patella flange The anterior part of a femoral component

where the patella articulates during the
range of flexion
Patella trochlea The area covered by articular cartilage on
the antero-distal end of the femur where the
patella articulates
Patella The small bone which slides on the anterior
femur during flexion-extension and trans-
mits the force between the quadriceps mus-
cle and the proximal tibia
Plastic deformation Deformation which is not recovered after
removal of the force
Plastic insert The polyethylene tibial component which
fits into the metallic tibial tray, a modular
construction
Polycentric A changing center of rotation as one body
(e.g., bone) is flexed relative to another body
Posterior cruciate ligament The ligament in the center of the knee which
limits the displacement of the femur anteri-
orly relative to the tibia, or the displacement
of the tibia posteriorly relative to the femur
Posterior tibial slope In an anatomic knee, the angle between the
line from the anterior to the posterior of the
proximal tibia and an anterior-posterior hor-
izontal line which is perpendicular to the
anatomic axis. Typically 5–7 °
Posterior The back of the knee or in a backward
direction
Pressure The force per unit area applied by a fluid
or the force per unit area applied on a
surface
PROM Patient-reported outcome measure such as
KOOS, WOMAC, KSS, to quantify the
functional performance and other
parameters
Proprioception The ability to sense position, motion, and
equilibrium; has been ascribed to the ante-
rior cruciate and other ligaments
Proximal At the upper end, usually referred to the
position on a bone
PS Type of artificial knee. Designation for a
condylar replacement artificial knee where
the anterior and posterior cruciates are
332 Glossary
resected, and stability is provided by an

intercondylar cam-post
Range of flexion The angular excursion between the femur
and tibia between maximally straight and
maximally flexed
Resultant force A single force vector representing more than
one applied force vector
Resurfacing Replacement components for the bearing
surfaces of the knee
Rigid body An object which is considered to be of con-
stant shape even when forces are applied
(see deformable body)
Rotation An angular motion (degrees) of a moving
body relative to a fixed body about a defined
axis in the fixed body
RSA Radiostereometric Analysis, used for deter-
mining motion of an artificial knee compo-
nent relative to the supporting bone, for
evaluating fixation
Sagittal radii Usually applied to the radii of the distal fem-
oral condyles seen in the sagittal view, the
radius is not constant around the profile
Sagittal The plane describing the side view of the leg
or knee
Screw-home The phenomenon where the tibia progres-
sively externally rotates relative to the
femur as the tibia is brought into full
extension
Shear stress The shear force divided by the area of sur-
face over which the force is applied
Shear A force or stress applied across a surface or
across a defined plane in a body
Snap-fit Two components fitted together by a snap
mechanism involving temporary elastic
deformation of a flexible member, usually
applied to a plastic tibial bearing snapping
into a metal tibial tray
Soft tissue Tissue excluding bone and cartilage; most
often referred to tendon, ligament, muscle,
fibrous capsule, synovial membrane
Spongy bone A network of interconnected bony spicules
called cancellous or trabecular bone
Glossary 333
Stability In function, there is no subjective feeling of

excessive femoral-tibial motion nor any
overstressing of soft tissues
Stiffness Displacement per unit applied force
Strain Change in length divided by original length
or proportionate change in length along a
defined line
Stress distribution The stress pattern across a surface or defined
plane, usually denoted with color coding,
e.g., red = high stress, blue = low stress
Stress protection An area of bone which experiences less than
its normal stress due to the load being
bypassed by an implant component
Stress Force divided by area, applied to a small ele-
ment of material, the element located in a
body or on a surface
Subluxation When one surface of a joint reaches the
boundaries of the opposite surface but before
irreducible dislocation
Survivorship Probability of an artificial knee surviving
successfully for a defined length of time,
e.g., 10 and 20 years
TCP Total Condylar Prosthesis.
Tensile strength The maximum force or stress a structure can
carry before mechanical failure, e.g., a liga-
ment or tendon
Tensile stress The force applied on a defined area which
lengthens the dimension of an object,
divided by the area
Three-body abrasion Wear caused on one or both sliding surfaces
due to the presence of a small intervening
hard particle
Tibial tray The metallic tibial component of an artificial
knee which houses the plastic bearing sur-
face or plastic insert
Tissue Materials comprising the human body,
including bone, cartilage, tendon, ligament,
muscle, synovial fluid
Torque A twisting action on a body that causes
rotation
Total knee arthroplasty Same as artificial knee
Total knee replacement Same as artificial knee
Toughness The amount of energy a material can absorb
up to the point of failure
334 Glossary
Trabecular bone A network of interconnected bony spicules

called cancellous or spongy bone
Transverse axis An axis from the medial to the lateral
side, usually applied to an axis through
the distal femur or across the proximal
tibia
Transverse plane The plane defined by the medial-lateral axis
and the anterior-posterior axis of the tibia
Ultracongruent A condylar replacement artificial knee
where the tibial surfaces are almost fully
congruent with the femoral bearing
surfaces
Unconstrained Not providing any restraint to motion, can
apply to an artificial knee which allows free-
dom of motion
Unicompartmental knee An artificial knee consisting of a metal run-
ner for a single femoral condyle and a plas-
tic bearing covering a single tibial plateau
Uphill mechanism Stability between opposing surfaces of a
joint when the surfaces distract relative to
each other against the applied force
Valgus The alignment of the femur on the tibia such
that the femoral and tibial condyles are dis-
placed medially creating a “knock-kneed”
effect
Varus The alignment of the femur on the tibia such
that the femoral and tibial condyles are dis-
placed laterally creating a “bow-legged”
effect
Vector An entity that has magnitude and direction.
Usually applied to a force or a line in space.
Velocity The rate of positional or angular change of
an object’s position with time
Viscoelastic Time-dependent deformation on application
of a force. Stress relaxion and creep are
associated terms
Vitallium The trade name for cobalt-chrome-
molybdenum alloy which has been widely
used for the manufacture of artificial knee
components (and other joints)
Wear Progressive removal of small particles of
material from a surface due to sliding of that
surface against another surface
Index
A B
ACL+PCL prostheses, 270, 278 Bi-cruciate retaining (BCR), 303, 304
ACL+PCL retaining, 47 Biomechanical knowledge
Acrylic cement, 6, 26 acrylic cement, 26
biomechanical design criteria, 69 articular cartilage, 21
cement-bone interface, 73–79 artificial knee design, 29, 31
long-term behavior of cement, 80, 81 ball-in-socket bearing surfaces, 21
micromotions, 70 boundary lubrication, 14
origin and use of, 70–73 cemented metal-on-plastic condylar
orthopaedic surgery, 71 replacements, 29
Townley’s tibial plateau design, 70 cemented systems, 29
uniform depth of cement penetration, 82 cobalt-chrome alloy, 23
Advanced Mechanical Technology cobalt-chrome-molybdenum alloy, 14, 16
(AMTI), 289 cruciate ligaments, 15
All-plastic tibial component, 88, 131 deformability, 22
American Academy of Orthopaedic Surgeons design of artificial limbs, 18, 20
(AAOS), 11, 46, 255, 256, 296, 316 distal-anterior tibial surface, 15
American Association of Hip and Knee elastohydrodynamic lubrication, 13, 14
Surgeons (AAHKS), 316 femoral and tibial condyles, 21
American Joint Replacement Registry femoral-tibial contact, 28
(AJRR), 296, 297 femoral-tibial positions, 24, 25
American Society for Testing and Materials fixation material, 26
(ASTM), 42, 248, 260, 289 fluid film, 14
Anametric knee, 148–149 force analysis, 27
Anatomic alignment, 214, 215 ground-to-foot force, 26, 27
Anatomic knee, 23 hip arthrodesis, 24
Anterior-posterior (AP), 113, 167 lateral femoral condyle, 15, 17, 18
laxity, 302 ligaments and tendons, 31
stability, 108 low-friction and wear-resistant bearing, 13
Archard’s wear equation, 239 model of knee motion, 23
Arthritic knee, 1 modern-day artificial knee, 28
Articular cartilage, 14, 21, 23 motion of limb segments, 24
Artificial knee, 1–9, 60, 62, 64, 66, 67, 113 multi-disciplinary approach, 14
design, 3, 302 orthopedic biomechanics, 23
joint, 26 sagittal outline, 15
mechanics, 257 screw home mechanism, 24
ASTM testing, see American Society for side-to-side bending, 21
Testing and Materials squeeze-film lubrication, 14
Authoritative surgeon-designer, 298 stick-slip motion, 21

P. S. Walker, The Artificial Knee, https://doi.org/10.1007/978-3-030-38171-4
336 Index
Biomechanical knowledge (cont.) Constant radius, 106, 115

tibial cartilage, 23 Constant sagittal radius, 297
tibial component fixation, 31 Continuous passive motion (CPM), 308, 311
tibial condyle, 29 Controlled convergence design model,
trabecular bone, 25 39, 41, 42
varus deformity, 25 Cruciate ligaments, 121
varus osteoarthritis, 25 Cruciate retaining (CR), 2, 3, 47, 128, 129,
varus-valgus motion, 17, 19 269, 271, 272, 295, 299
varus-valgus stability, 24 Cruciate sacrificing, 299
weeping lubrication, 14 Current Concepts in Joint Replacement
BiSurface knee, 182 (CCJR), 316
Bi-unicompartmental knees, 303 Custom knees, 305
Boosted lubrication, 14 Custom-made knees, 297
Boundary lubrication, 14
British Orthopaedic Association (BOA), 45, 46
D
Deane knee, 108, 168
C Design process
Ceramic pellets, 86 artificial knee, 34
Charge-coupled device (CCD), 80 condylar replacement, 36
Charnley’s design, 29 controlled convergence design model,
Cintor, 158 39, 41, 42
Cloutier’s total knee design, 127 definition, 33
Cobalt-chrome alloy, 23, 87 early artificial knees, 43–45
Compartmental knees, 163 fixed-axis hinge, 34–35
Compartmental resurfacing knees, 105 intercondylar hinge, 35
Component loosening, 4 intramedullary canals, 34
Comprehensive knee systems, 301 progress after 1970, 45–48
Computer-aided design (CAD), 174 simplified design process, 36–39
Computer-assisted instrumentation, 221–226 systematic process, 48
Computer-assisted methods, 308 unintended consequences, 35
Computer-assisted navigation, 305 Designs and techniques evaluations
Computer-Assisted Orthopaedic Surgery alternate evaluation methods, 277, 278
(CAOS), 11, 315, 319 biomechanical methods, 278–282, 287–289
Computer-assisted soft tissue balancing CR vs. PS knees, 272
devices, 305 fluoroscopic evaluation of kinemat-
Computer-assisted surgery, 3, 315 ics, 282–287
Computer modeling, 317 function measure, 289
Condylar knee, 196, 210 medial pivot designs, 274
Condylar replacements, 181 PROMs, 265–269, 290
anatomic designs, 149–151 rotating platform designs, 273
femoral-tibial conformity, 159 unicompartmental vs. total knee, 270, 271
Freeman-Swanson knee, 143–145 Design team, 301
geometric and anametric knees, 148–149 Diamond saw, 86
Gunston polycentric knee, 141–143 Distal femurs and proximal tibias, 88
HSS, 151–158 Duo-condylar knee, 122, 128, 129, 241
Kodama Yamamoto knee, 146 Duopatella knees, 129, 131, 157
laxity and stability, 161, 162
medica’s flagship knee, 160
metallic components, 163 E
minimally-resected tibial plateau, 162 Eftekhar‘s tibial components, 123, 124
posterior cruciate, 159 Elaborate gait analysis system, 106
tibial component, 162 Elastohydrodynamic lubrication, 13, 14
Townley Anatomic Total Knee, 146, 147 Electromyography, 19
Index 337
European Federation of National Associations Freeman-Samuelson knee, 191

of Orthopaedics and Traumatology gradual external rotation, 193
(EFORT), 296 internal-external rotation, 189–191
Ewald artificial knee, 119 loose medial pivot, 196
medial pivot knee, 194, 195
Medial Rotation Knee, 191, 192
F medial pivot knees, 181
Femoral component and patella meniscal bearing, 181
anatomically-shaped femoral compo- meniscal bearing knees
nents, 109–111 articular cartilage and synovial
femoral component designs, 104–109 fluid, 184
multiple sizes, 113–115 artificial knee, 189
patella flanged components, 111, 112 condylar femoral component, 186
patello-femoral joint, 116–119 devised novel methods, 184
thin-walled Vitallium shells, 103, 104 femoral component, 188
vitallium-mold arthroplasty, 103 flexion angle, 188
Femoral condyle, 105 geometric knee, 186
Finite element analysis (FEA), 41 medial femoral condyles, 184
Flexion deformity, 207 meniscal bearing artificial knees, 186
Flexion-extension kinematics, 246 photoelastic technique, 186
Food and Drug Administration (FDA), 42, 302 rotating platform design, 187, 188
Force analysis, 27 types of, 184
Forgotten Joint Score (FJS), 269, 270, 274 unicompartmental knee, 185, 186
Freeman–Samuelson knee, 191 Gunston design, 55, 106, 122
Freeman-Swanson design, 128 Gunston polycentric knee, 141–143, 235
Freeman-Swanson knee, 143–145 Gunston’s knee, 123
Functional performance, 309–311 Gunston’s polycentric tibial components, 121
G H
G-curve, 115 Hand-fitting, 299
Geometric artificial knees, 30 Highly-crosslinked-polyethylene, 37
Geometric knee, 45, 109, 124–126, 148–149, History of Tribology, 13
186, 187, 242, 265 Hospital for Joint Diseases (HJD), 172
Graduated Knee System, 170 Hospital for Special Surgery (HSS), 151–158,
Guided motion knees, 305 265, 319
biomimetic motion/geometry, 197 Hot-pressing, 92
cam-post mechanism, 198, 199 Human Walking, 20
lateral and medial tibial bearing
surfaces, 199, 200
mobile fluoroscopy unit, 196, 197 I
replica anatomic knee, 200 Inertial measurement units (IMU), 225, 277
sagittal planes, 196 Insall-Burstein (IB) PS knee, 173–177, 179
third condyle knee, 200 Institution of Mechanical Engineers
tibial component, 199 (IMechE), 46
CR and PS knees, 201 Instrumentation and technique, 205, 206
definition of, 181 alignment and force transmission, 205–209
freedom of rotation, 182–184 alignment goals and techniques, 214
internal-external rotation, 182 computer-assisted instrumenta-
medial and lateral laxity, and stability tion, 221–226
anatomic knee, 194 kinematic alignment, 214–217
external rotation, 189 mechanical instruments, 217–219, 229
femoral external rotation, 192, 193 alignment and soft tissue bal-
4-bar linkage model, 194, 198 ancing, 211
338 Index
Instrumentation and technique (cont.) anterior-posterior and varus-valgus

anterior-posterior stability, 210 laxities, 57, 58
design of total knee, 209 anterior-posterior displacements, 63
femoral and tibial surfaces, 209 artificial knee, 60, 62, 64, 66, 67
femoral component, 210 axial loads, 56, 57
femoral intramedullary canal, 211 biomechanical principles and expe-
long-leg radiographs, 212 rience, 55
precision cut instrumentation, 212, 213 bone resections, 53
proximal-distal and anterior-posterior definiton, 51–53
attitude, 211, 212 femoral and tibial condyles, 63, 65
radiopaque markers, 210 Gunston design, 55
varus deformity, 211 integral patella flange, 59, 60
PSI, 220, 221 internal and external rotation, 53
range of flexion, 219, 220 lateral and medial femoral condyles, 55
soft tissue balancing, 226–229 lateral instability, 55
International Congress for Joint ligament function, 53, 54
Reconstruction (ICJR), 316 load-deflection experiments, 63, 64
International Organization for Standardization medial pivot, 61, 62
(ISO), 260 principles of, 59, 61
International Society for Technology in tibial bearing surfaces, 56, 57
Arthroplasty (ISTA), 11, 222 tibial fixation, 53
International Standards Organization tibial plateaus and femoral condyles, 57, 59
(ISO), 42 tibial radius values, 58, 60
tibial surface, 54, 59, 61
Load angle inlay (LAI), 162
J Low-contact-stress (LCS) mobile-bearing
J-curve, 115 knee, 159
Joint lubrication, 14
M
K Magic pegs, 89
Kane’s dynamics, 285 MatOrtho, 191
Kinematic alignment, 214–217 Maximum micromotion, 5
Kinematic Knee System, 91, 159, 160, 171 Maximum total point motion (MTPM), 98
Kinemax Knee System, 160 McKee-Farrar metal-on-metal hip, 29
Knee arthroplasty, 209 Medial condyles, 106, 110
Knee Injury and Osteoarthritis Outcome Score Medial congruent bearing, 195
(KOOS) system, 269 Medial-lateral (ML), 113
KneeSim software, 197 Medial pivot, 61, 194, 195, 270, 297, 302, 304
Knee simulator tests, 307 Medial Rotation Knee, 191, 192
Knee Society Score (KSS), 265–268 Meniscal bearing artificial knees, 186
Kodama Yamamoto knee, 146 MGH femoral condyle, 111, 119
Miller-Galante knee, 88, 279
Minimally invasive surgery, 315
L Mobile bearings, 297
Larger distal radius, 115 Modern arthroplasty, 8
Lateral condyles, 106, 110 Modern automobiles, 6
Lateral tibial plateau, 313 Monoblock metal-backed components,
Laxity and stability 133, 134
anatomic knee, 55
anterior-posterior and internal-external
laxities, 63 N
anterior-posterior and rotational stabil- Nas Eftekhar’s intramedullary adjustable
ity, 52, 53 knee, 183
Index 339
Natural knee, 302 Robotic systems, 306

Neuromuscular electrical stimulation Roentgen stereophotogrammetry, 98
(NMES), 309
NexGen CR, 296
NexGen Knee System, 176 S
NexGen LPS-Flex mobile bearing Shaw-Murray dynamic machine, 243, 244
designs, 313 Single-use 3D printed instruments, 311, 314
Non-sterile surgical technique, 1 Sit-to-stand (STS) test, 277, 287, 288
Sledge knee, 107
Smaller posterior radius, 115
O Soft tissue balancing, 226–229
Opioid-sparing medication protocols, 308, 312 Spherocentric knee, 192
Orthopedic Research Society (ORS), 11, 192 Squeeze-film lubrication, 14
Oxford Knee Rig., 244, 245 Stair climb test, 288
Oxford Knee Score (OKS), 268
T
P Testing and tribology
Patello-femoral ‘distress, 111, 112 all-metal fixed-axis knee, 236
Patello-femoral joint, 116–119 Archard’s wear equation, 239
Patient-reported outcome measures (PROMs), artificial knee designs, 261
268, 290 cement-bone interface, 233, 234
Patient-specific instrumentation (PSI), computer modelling, 257–260
220, 221 current status, 260
Perioperative Orthopaedic Surgical Home elastohydrodynamic lubrication, 237, 238
(POSH) initiative, 308, 312 fixation of components, 241–243
Persona Knee System, 176, 195 flexion-extension movements, 237
Polycentric artificial knees, 30 four-channel machine, 235
Polycentric knee, 106, 142, 162, 265 friction coefficients, 233, 234
Polyetheretherketone (PEEK), 99 implant retrieval and wear, 255–257
Polyethylene, 6 laboratory tests, 237
pegs, 89 laxity & stability, 240
wear, 296 long-term durability and wear par-
Polymeric materials, 311, 314 ticles, 253–255
Porous plugs, 96 metal-on-metal and metal-on-polyethylene
Posterior cruciate ligament, 181 hips, 236
Posterior stabilized (PS) knee, 47, 164, 179, metal-on-plastic, 234
272, 295, 296 simulating machines, 243–252
designs, 269, 271 testing laxity and stability, 261
IB knee, 173–177 tibial bearing surfaces, 238
intercondylar stability, 167–169 time-consuming and expensive, 238
stabilized designs, HSS, 170–172 3D printed metal backings, 99
Pre-market approval process (PMA), 302 3D printing technologies, 307
Preoperative education, 312 Tibial component
Press-Fit Condylar (PFC) knee, 273 all-plastic vs. metal-backing, 133–135
Progressive radiolucency, 97 application of anatomic experiments, 122
Proximal tibia, 1 bone strength, 135–137
central peg development, 128–132
conformity and cruciate preservation, 125–128
R designs, 121, 122
Radiography, 97 durable fixation, 137, 138
Radiolucent lines (RLLs), 73, 74 loosening, 307
Recovery time, 308, 311 metal tibial tray and modularity, 122–124
Replica anatomic knee, 200 peripheral shape, 137
340 Index
Tibial coverage, 137 insert-tray locking mechanism, 96

Tibial fixation, 300 interface micromotions, 93, 94
Timed Up and Go (TUG) test, 277, 287, 288 long-lasting implants, 90
Titanium alloy, 87 metallic femoral component, 86
Total condylar knee, 117, 129, 131, 132, metal-reinforced tibial component, 96
173–175, 179, 182, 239 PCA knee, 91–94
Total condylar prosthesis, 176 porous ceramics, 85, 86
Total joint arthroplasty, 318 porous coatings, 90, 95
Total joint replacement, 318 porous structure by 3D printing, 99
Total knee arthroplasty (TKA), 9, 221, press-fit components, 85
311, 314 scientific evaluations, 96–100
Total knee designs, 128 short-term loosening problems, 100
Townley Anatomic Total Knee (ATK), 126, systemic effects, 85
127, 146, 147 tibial component, 89
Townley design, 133, 135, 148 tibial tray component, 87
Townley’s approach, 44 total knee components, 90
Townley’s femoral component, 112 Unicompartmental knees, 2, 4, 47, 270, 271,
Trabecular bone, 6, 135, 136 295, 296, 303
Trabecular metal, 99 Unicondylar, 129
U V
Ultracongruent design, 301, 304 Valgus deformity, 209
Ultracongruent knee, 177–179, 310 Variable axis knee, 182
Ultrahigh molecular weight polyethylene Varus deformity, 25, 206, 207, 210, 211
(UHMWPE), 254 Varus osteoarthritis, 25
Uncemented fixation, 307 Varus-valgus and internal-external rota-
advanced iterative finite element computer tions, 246
modelling, 97 Varus-valgus stability, 168–170, 182, 305
aluminum oxide ceramic compo- Vicon equipment, 193
nent, 86, 87 Vitallium, 14
anatomic alignment of joint line, 92 Vitallium-mold arthroplasty, 103
biomechanical testing method, 93 Vitamin E, 248, 307
calcium aluminate ceramic VIVO Joint Simulator, 251, 252
material, 86
cement-bone micromotion, 85
Co-Cr porous coating, 90, 91 W
Co-Cr specimen, 90 Wear-resistant polyethylene, 8
diamond saw, 86 Weeping lubrication, 14
fabricated small cylindrical pellets, 86
first uncemented designs, 95
fixation stability, 95 Z
hybrid method of fixation, 90 Zimmer’s premier knees, 90
initial mechanical stabilization, 94 Zonal analysis methods, 97

The Artificial Knee An Ongoing Evolution

Uploaded by

Copyright:

Available Formats

The Artificial Knee An Ongoing Evolution

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Artificial Knee An Ongoing Evolution

Uploaded by

Copyright:

Available Formats

The Artificial Knee

Tibial disphysis Posterior View

Tibial insert Tibial tray Acrylic

The Artificial Knee

ISBN 978-3-030-38170-7 ISBN 978-3-030-38171-4 (eBook)

© Springer Nature Switzerland AG 2020

Peter S. Walker was educated in England at the Royal

© Springer Nature Switzerland AG 2020 1

Condylar replacement knees

ACL+PCL PCL Total Posterior Guided

OA All revisions TKA

© Springer Nature Switzerland AG 2020 13

To design a successful artificial knee, knowledge of the biomechanics of the knee

Femur flexing clockwise

extreme posterior displacement of the lateral femoral condyle. In function, muscle

“Fundamental Studies of Human Locomotion and Other Information Relating to

Walking down stairs

to the deformability of the articular cartilage. He illustrated the structure of bones

• Shock absorption protecting the articular cartilage

Fig. 2.8 John Paul PhD,

10 JOINT FORCE AT KNEE - LEVEL WALKING

Fig. 2.10 Ed Chao PhD

What Was Known About Knee Biomechanics by 1970?

In this chapter, a systematic methodology for designing a new product will be

© Springer Nature Switzerland AG 2020 33

• The components must be secured rigidly to the bones.

hinge would be a reasonable approximation provided it included some freedom of

made between a linked intercondylar device and a condylar replacement, how

Simplified Design Process

Formulate Select Final

This can be accomplished by assigning scores to the different “wants,” using a

Controlled Convergence Design Model

CONTROLLED CONVERGENCE MODEL

In-depth Select Final Clinical Refine Final Final

Fig. 3.5 Stuart Pugh had a

Design Process for the Early Artificial Knees

Progress in Design After 1970

the American Academy of Orthopaedic Surgeons (AAOS), started to include

 as a Systematic Design Process Been Applied to Artificial

© Springer Nature Switzerland AG 2020 51

Posterior AP displt (mm)

Fig. 4.5 Keith Markolf

Fig. 4.8 Andrew Amis PhD

rotation which occurred, but there was no consideration of asymmetry between

 ow Has Laxity and Stability Been Incorporated into

Implants Without Fixation

© Springer Nature Switzerland AG 2020 69

Origin and Use of Acrylic Cement

Thin metal shell in acetabulum

Fig. 5.2 John Charnley’s

Depth of acrylic penetration

At surgery, it was necessary to account for the rheological behavior of the

The fixation was investigated experimentally on unicompartmental components

Fig. 5.5 In the late 1970s

Drill holes made in

Fig. 5.8 An anterior

Long-Term Behavior of Cement

LP: 85 F, 0 yrs PM: 61 M, 5 yrs

 hat Are the Basic Principles of Cemented Fixation, and How

What Was Known About Knee Biomechanics by 1970?

Simplified Design Process

Controlled Convergence Design Model

Design Process for the Early Artificial Knees

Progress in Design After 1970

as a Systematic Design Process Been Applied to Artificial

ow Has Laxity and Stability Been Incorporated into

Implants Without Fixation

Origin and Use of Acrylic Cement

Long-Term Behavior of Cement

hat Are the Basic Principles of Cemented Fixation, and How

Early Experience with Porous Ingrowth

I s Uncemented Fixation Reliable Enough to Replace

Early Femoral Component Designs

Anatomically Shaped Femoral Components

Patella Flanged Components

o What Extent Do Femoral Component Shapes Mimic Normal

Opposite Extremes of Tibial Component Design

Application of Anatomic Experiments

Metal Tibial Tray and Modularity

Conformity and Cruciate Preservation

Development of the Central Peg

Bone Strength Under Component

What Are the Best Configurations for Durability?

Gunston Polycentric Knee

Townley Anatomic Total Knee

Geometric and Anametric Knees

Anatomic Designs of Seedhom and Ewald

ospital for Special Surgery (HSS): Duo-condylar, Duo-patella

Other Designs of the 1970s and beyond

hat Were the Steps Leading to the Basic Condylar

Early Designs with Intercondylar Stability

Stabilized Designs at HSS

Insall-Burstein (IB) PS Knee

hat Was the Design Rationale of the Posterior