Annals of Biomedical Engineering, Vol. 43, No. 10, October 2015 ( 2015) pp.

2467–2476
DOI: 10.1007/s10439-015-1292-9

Biomechanical Characterization of a Model of Noninvasive, Traumatic

Anterior Cruciate Ligament Injury in the Rat
TRISTAN MAERZ,1,2,3 MICHAEL D. KURDZIEL,1,2 ABIGAIL A. DAVIDSON,1 KEVIN C. BAKER,1,2
KYLE ANDERSON,5 and HOWARD W. T. MATTHEW3,4
1
Orthopaedic Research Laboratories, Beaumont Hospital, 3811 W 13 Mile Road, Royal Oak, MI, USA; 2Department of
Orthopaedic Surgery, Oakland University William Beaumont School of Medicine, Rochester, MI, USA; 3Department of
Biomedical Engineering, Wayne State University, Detroit, MI, USA; 4Department of Chemical Engineering & Materials Science,
Wayne State University, Detroit, MI, USA; and 5Department of Orthopaedic Surgery, Beaumont Hospital, Royal Oak, MI,
USA
(Received 17 November 2014; accepted 25 February 2015; published online 17 March 2015)

Associate Editor Michael R. Torry oversaw the review of this article.

Abstract—The onset of post-traumatic osteoarthritis following trauma due to sporting injury, motor vehicle
(PTOA) remains prevalent following traumatic joint injury accidents, or other injurious events. It has been esti-
such as anterior cruciate ligament (ACL) rupture, and animal mated that PTOA constitutes approximately 12% of
models are important for studying the pathomechanisms of
PTOA. Noninvasive ACL injury using the tibial compression all OA cases,6 and there is a 20–50% risk of developing
model in the rat has not been characterized, and it may PTOA following any joint injury.9 The severity and
represent a more clinically relevant model than the common rate of onset is directly related to the severity of trau-
surgical ACL transection model. This study employed four ma,2 and few interventional clinical treatments have
loading profiles to induce ACL injury, in which motion shown repeatable success.22 Rupture of the anterior
capture analysis was performed, followed by quantitative
joint laxity testing. High-speed, high-displacement loading cruciate ligament (ACL) is a well-described injury
repeatedly induces complete ACL injury, which causes leading to the onset of PTOA.22 As the major stabilizer
significant increases in anterior-posterior and varus laxity. of anterior translation of the tibiofemoral joint, the
No loading protocol induced valgus laxity. Tibial internal ACL is critical in maintaining knee stability during
rotation and anterior subluxation occurs up to the point of biomechanical loading of the joint. High-magnitude
ACL failure, after which the tibia rotates externally as it
subluxes over the femoral condyles. High displacement was and high-rate anterior translation and/or rotation of
more determinative of ACL injury compared to high speed. the tibia with respect to the femur cause ACL sprains
Low-speed protocols induced ACL avulsion from the or ruptures. Excessive inflammation due to damage to
femoral footprint whereas high-speed protocols caused either ligament, cartilage, meniscus, and synovium tissue in
midsubstance rupture, avulsion, or a combination injury of addition to adverse biomechanical loading of these
avulsion and midsubstance rupture. This repeatable, nonin-
vasive ACL injury protocol can be utilized in studies tissues following destabilization of the knee are pos-
assessing PTOA or ACL reconstruction in the rat. tulated factors contributing to the onset of PTOA.
2,7,23
Though the gold-standard treatment of ACL
Keywords—Anterior cruciate ligament, Biomechanics, Animal ruptures is ACL reconstruction in efforts to remove
models, Noninvasive injury, Post-traumatic osteoarthritis. damaged tissues and restabilize tibiofemoral articula-
tion, the onset of PTOA following ACL rupture is
reported to be as high as 90% regardless of treat-
INTRODUCTION ment.10,19,21,29,33,38
Small-animal models of PTOA following ACL in-
Post-traumatic osteoarthritis (PTOA) is the onset jury are commonly utilized to study the onset, pro-
and progression of osteoarthritis (OA)-like symptoms gression, and potential treatment of PTOA. The most
widely published of these models is surgical ACL
transection,11–14,24,34 which acutely destabilizes the
Address correspondence to Kevin C. Baker, Orthopaedic Re-
knee joint, causing adverse biomechanical joint load-
search Laboratories, Beaumont Hospital, 3811 W 13 Mile Road,
Royal Oak, MI, USA. Electronic mail: [email protected] ing and the subsequent onset of OA-like symptoms
2467
0090-6964/15/1000-2467/0  2015 Biomedical Engineering Society
2468 MAERZ et al.

such as the loss of proteoglycan content,1,11 osteophyte protocol. Immediately after euthanasia, both hind
formation,12 altered biomechanical properties of ar- limbs were dissected to remove the skin, and small
ticular cartilage,1 loss/thinning of articular cartilage,11 incisions were made over the lateral aspect of the fe-
and chondrocyte death.24 However, surgical transec- moral metaphysis and the lateral aspect of the tibial
tion may fail to accurately represent the native biolo- tuberosity using a #15 scalpel. A 1.00 mm Kirschner
gical response after high-magnitude tibiofemoral wire (K-Wire) was used to drill a hole into the femoral
loading and subsequent ACL injury. Furthermore, metaphysis and tibial tuberosity, and a 1.00 mm pin
incision of the skin and joint capsule may induce bio- attached to a 4.00 mm hemispherical retroreflective
logical cascades not present in human ACL injury, marker was rigidly inserted into each hole. One addi-
thus introducing confounding effects. tional reference marker was drilled into the greater
Only a few studies have utilized noninvasive loading trochanter of the femur, and another reference marker
protocols to induce ACL injury in small animals. Onur was attached to the testing fixture holding the animal’s
et al. applied a cyclic compressive load to the distal paw.
femur of the flexed knee joint in mice until ACL rup-
ture occurred, which induced OA-like degenerative
changes at 8 weeks.30 Tang et al. used a rotational
Mechanical Loading Protocol
mechanical load in a rat to characterize pro-inflam-
matory cytokine and proteinase expression following The fixture and testing protocol utilized in this study
noninvasive ACL rupture.37,39 Christiansen and are modified from a previously-published tibial com-
coworkers used a mouse model in which they applied pression knee injury model in mice.8 A custom fixture
tibial compression via the paw with the knee flexed at was rigidly mounted onto the stage of an electrome-
~90. Each leg was subjected to a compressive load of chanical materials testing system (MTS Insight 5, Eden
12 N, causing transient anterior subluxation of the Prairie, MN, USA). The bottom fixture supports the
tibia relative to the femur and ultimate failure of the animal in a prone position, and a knee stage with an
ACL.8 They demonstrated a loss of trabecular bone 8.5 mm wide, 25 mm long, and 3.5 mm deep trough is
volume, proteoglycan loss, chondrocyte atrophy, and positioned at the center of the actuator. The knee stage
loss of the superficial zone of cartilage by 56 days. limits medial–lateral motion while providing room for
Another study by the same group examined loading anterior subluxation of the tibia relative to the femur.
rate-dependent injury modes in the same mechanical The top fixture was mounted at the center of the ac-
loading protocol in mice and found that low loading tuator, and it rigidly supports the rat’s hind paw in 30
rates caused bony avulsion at the ACL insertion of dorsiflexion without allowing medial or lateral
whereas high loading rates caused midsubstance rup- motion. Resultant knee flexion after positioning of the
ture of the ACL.20 animal is ~ 100 (Fig. 1a).
The rat is commonly utilized as a model of PTOA as Four displacement-controlled loading protocols
well ACL reconstruction,5,15,28 but, to date, no group has were employed to induce tibial-compression knee in-
employed the tibial compression ACL injury model in the jury and/or ACL injury (n = 9 limbs per group/pro-
rat and characterized joint biomechanics during injury. tocol). A 3 N preload was held for 10 s, followed by 10
Since the mouse is generally considered too small to preconditioning cycles of 1–5 N at 0.5 Hz. A 15 N
perform studies involving ACL reconstruction, a preload ramp at 0.1 mm/s was then applied, and im-
clinically-relevant model of ACL injury is needed in the mediately after reaching a 15 N load, a displacement-
rat. To this end, the purpose of this study was to biome- controlled ramp using one of four testing conditions is
chanically characterize a noninvasive ACL injury in the applied: high and low speed (8 and 1 mm/s, respec-
rat using a tibial compression protocol. Specifically, we tively), and high and low endpoint displacement (3 and
sought to investigate the repeatability of four loading 2 mm, respectively). The four resultant testing groups
protocols at varying speed and endpoint displacement, were: high-speed, high-displacement (HSHD); high-
perform motion capture to quantity joint motion during speed, low-displacement (HSLD); low-speed, low-dis-
loading, and quantify post-loading laxity. placement (LSLD), and low-speed, high-displacement
(LSHD). These loading parameters were chosen based
on preliminary tests performed at varying speeds and
MATERIALS AND METHODS endpoint displacement. These preliminary experiments
demonstrated that displacements higher than 3 mm
Specimen Preparation
increased the incidence of fracture while 2 mm was
Adult, female Lewis rats (200–250 g) were eu- found to be the minimum displacement necessary to
thanized by CO2 asphyxiation under an Institutional induce ACL injury. Load, displacement, and time data
Animal Care and Use Committee (IACUC)-approved was sampled at 200 Hz during each test.
 Biomechanics of Noninvasive Rat ACL Injury 2469

malized to the position of the distal femoral marker at

each data point. Relative tibiofemoral motion was
analyzed in all three 2D anatomical planes (transverse,
coronal, and sagittal) as well as in 3D. Motion was
analyzed at the point of peak load, which constituted
either ACL injury or the highest load when injury was
not successfully induced, as well as motion between the
point of peak load and maximum 3D joint displace-
ment, which represented joint motion after ACL injury
when injury was successful. Motion was also analyzed
as total motion (i.e., total distance traveled by the tibia
relative to the femur) as well as net motion.

Quantitative Joint Laxity Testing

Following mechanical loading, limbs were disar-
ticulated at the hip joint, and approximately 1 cm of the
FIGURE 1. Mechanical loading and laxity testing fixtures.
proximal femur was exposed to facilitate clamping.
(a) To induce ACL injury, a compressive displacement is Anterior-posterior (AP) laxity was assessed using a
applied to the tibia with the knee flexed to ~100. Medial–lat- subluxation test in which the femur was rigidly clamped,
eral motion of the knee is constrained with a trough in the
knee stage, and the paw is fully constrained and flexed to 30
the knee was flexed to 90, and the paw was fixed in a
of dorsiflexion. (b) Following injury loading, Anterior-poste- clamp in 30 of dorsiflexion (Fig. 1b). AP laxity testing
rior (AP) laxity was quantified with a subluxation test by the was performed with a 10 s 0.3 N preload, 10 precondi-
application of a constant tibial compressive load. (c) Varus
and valgus laxity are assessed by the application of a con-
tioning cycles between 0.1 and 0.5 N at a rate of 0.25 Hz,
stant load applied at the distal tibia to induce either varus or another 10 s 0.3 N preload, and a ramp to 2 N at
valgus joint stress. DF, dorsiflexion; Flex, flexion; MCL, 0.25 mm/s, which was held for 10 s. AP laxity was ex-
medial collateral ligament; LCL, lateral collateral ligament.
pressed as net displacement due to the applied 2 N load
as well as compliance during the applied load, expressed
in mm/N. After AP laxity testing, the paw was disar-
Motion Tracking and Analysis
ticulated at the ankle joint, and a 1 mm pinhole was
During the duration of each mechanical test, motion drilled anterior-to-posterior in the distal tibia. The femur
of each reflective marker was tracked using a four- was rigidly mounted in the clamp, and the distal tibia
camera, 5 megapixel motion capture system (Oqus 5, was attached to a linear bearing with a 0.8 mm stainless
Qualisys, Göteborg, Sweden). Data was collected steel pin via the drilled pin hole. The linear bearing was
continuously at a 179 Hz and analysis was performed attached to the actuator, which applied a downward
using both the Qualisys Track Manager software and load to assess varus laxity and an upward load to assess
Matlab (v2013a, Mathworks, Natick, MA). Ex- valgus laxity (Fig. 1c). The linear bearing ensured that
periments performed to determine the resolution of linear actuator motion resulted in arc motion of the tibia
this system using the same reflective markers, the same in order to eliminate joint distraction, which would skew
camera-to-object distance, and displacement speeds of load-controlled testing. Varus and valgus laxity testing
both 1 and 8 mm/s found that the system has a was performed with a 10 s 0.3 N preload, 10 precondi-
resolution of 279 lm with a 98.07% accuracy at 8 mm/ tioning cycles between 0.1 and 0.5 N at a rate of 0.25 Hz,
s, and a 49.02 lm resolution with 99.31% accuracy at another 10 s 0.3 N preload, and a ramp to 1 N at
1 mm/s. The motion-capture system was synchronized 0.25 mm/s, which was held for 10 s. Varus and valgus
to the materials testing system using an external circuit laxity were expressed as net displacement due to the
to capture time-synchronized data between the two applied 2 N load in the respective directions as well as
systems. Following acquisition, data was converted compliance during the applied load. Healthy, non-
into a calibrated coordinate system where the z direc- loaded specimens were used as laxity controls.
tion constitutes the vertical axis of the actuator (i.e.,
the direction of the applied load along the tibia), the
Micro Computed Tomography and Gross Dissection
x axis constitutes the medial–lateral direction with re-
spect to the animal, and the y axis constitutes the Following mechanical loading and assessment of joint
cranial-caudal direction with respect to the animal. To laxity, limbs were imaged using micro computed to-
analyze the motion of the tibia relative to the femur, mography (lCT) at 70 kVp, 114 lA at an isotropic voxel
the position of the proximal tibial marker was nor- size of 36 lm (lCT40, Scanco Medical, Brüttisellen,
2470 MAERZ et al.

Switzerland). A blinded investigator assessed each image types were observed: a true midsubstance rupture with
set for gross fractures of the tibia or femur, and bony ligamentous tissue fully attached to both the femoral
avulsions of the ACL were identified by small bone pieces and tibial footprints, where the rupture was found to
in the center of the joint near the femoral notch. be closer to the femoral footprint in all specimens. The
Following imaging, each sample was dissected to second injury type was a combination avulsion and
confirm ACL injury, to identify and confirm potential midsubstance rupture where one bundle of the ACL
fractures, ruptures, or avulsions, and to characterize avulsed from the femoral footprint and the other
the location and/or type of injury. bundle exhibited a midsubstance rupture. The third
injury type was a full avulsion of the femoral attach-
ment where both bundles remained attached to bone,
Statistical Analysis
and the avulsion footprint could be visualized on the
Statistical analysis was performed in SPSS (v22, IBM, femur. Of the 9 injured specimens in HSHD, 2 speci-
Armonk, NY). The normality and equal variance as- mens exhibited midsubstance rupture of the ACL, 4
sumptions were assessed using the Shapiro–Wilk test and specimens in HSHD exhibited the combination rup-
Levene’s test, respectively. Differences in independent ture-avulsion injury, and 3 specimens exhibited full
normally distributed and non-normally distributed vari- avulsions. In two of the combination injury specimens,
ables were compared between groups using one-way we were able to confirm that the anteromedial bundle
analysis of variance (ANOVA) and Kruskal–Wallis tests, was the avulsed bundle, and the posterolateral bundle
respectively. Multiple comparisons were performed with was ruptured. Of the 3 injured specimens in HSLD, 2
a Sidak p value correction at a = 95%. Differences in exhibited the combination avulsion-rupture injury and
independent, normally distributed and non-normally 1 exhibited a midsubstance rupture. All 5 of the 5 in-
distributed variables between two groups (e.g., injured vs. jured specimens in LSHD were full avulsions. lCT
non-injured specimens) was performed using t tests and data confirmed avulsion injuries by the appearance of
Mann–Whitney U tests, respectively. The association a small bony piece in the femoral notch (Figs. 2b and
between complete ACL injury and testing speed and 2c), whereas no bone pieces were visualized in the joint
displacement was assessed using v2 tests. P values lower in specimens that experienced midsubstance rupture.
than 0.05 were considered significant. Gross dissection did not reveal disruption of the
PCL, LCL, or MCL in any testing group. The origin
and insertion of each ligament was observed to be in-
RESULTS tact with no evidence of macroscopic disruption to the
ligament midsubstance. No evidence of gross articular
Complete Injury and Fracture Incidence
cartilage injury was noted on the femoral condyles of
Complete ACL injury occurred in 100% of specimens any specimen. Furthermore, we did not observe any
in HSHD (9/9), 33.33% (3/9) in HSLD, 55.56% (5/9) in structural damage to subchondral or epiphyseal bone
LSHD, and 0% in LSLD (0/9). Injury was significantly on lCT images of the femur and tibia.
associated with higher testing speed (v2 = 5.46,
p = 0.019) and higher endpoint displacement (v2 =
Joint Motion During Tibial Compression Loading
13.48, p < 0.001). No bony diaphyseal or metaphyseal
fractures occurred during loading, and no fractures were Relative tibiofemoral joint motion of all tested
identified on lCT images. However, a distal femoral specimens is summarized in Table 1. In general, the
physeal fracture/slip occurred in one specimen in each of motion in the transverse plane was most pronounced.
HSLD, LSHD, and LSLD during mechanical loading. HSHD exhibited significantly higher total coronal
These fractures were confirmed on lCT images, indicating motion and transverse motion after peak load com-
separation of the distal femoral physis with a concomitant pared to HSLD. LSHD had significantly higher total
coronally-oriented bony fracture of the distal femoral sagittal motion compared to HSLD. HSHD had the
epiphysis (Fig. 2a). All three of these fractures were sev- highest 3D motion compared to other groups, but
eral millimeters away from the pinholes drilled for these differences were not significant due to the
retroreflective marker placement, and lCT images did not relatively large standard deviations when grouping
indicate fracture propagation to the pinholes. both complete-injury and uninjured specimens. When
comparing only specimens that exhibited complete
ACL injury, more pronounced differences in relative
Loading-Dependent Injury Types
tibiofemoral joint motion are observed between groups
ACL injury was confirmed with gross observation (Table 2). HSHD exhibited significantly higher total
of the dissected joint (Fig. 3). Of the 17 specimens that coronal motion, higher total 3D motion, and higher
experienced complete ACL injury, three major injury transverse motion after peak load compared to HSLD.
 Biomechanics of Noninvasive Rat ACL Injury 2471

FIGURE 2. lCT imaging following loading. Three specimens exhibited physeal displacement (a, white arrow) with a concomitant
distal femoral epiphyseal fracture (a, red arrow). Avulsions were noted by small pieces of bone in the joint space, apparent near the
femoral footprint on both sagittal (b, red arrow) and coronal slices (c, red arrow).

over the femoral condyles with a rotational compo-

nent. Specimens that did not exhibit complete injury,
however, exhibited only tibial internal rotation with
less pronounced tibial subluxation (Figs. 4c and 4d).
Motion after the point of peak loading is markedly
lower in specimens that did not exhibit ACL failure,
most notably in the medial–lateral direction, and no
external rotation past the point of the start of loading
is noted.
The most profound differences in relative tibiofemoral
joint motion were observed when grouping high dis-
placement groups (HSHD, LSHD) and low displacement
groups (HSLD, LSLD). There was a significant differ-
ence in total 3D motion (high-displacement: 7.17 mm ±
FIGURE 3. Complete ACL injury was confirmed in dissected 1.29; low-displacement: 5.19 mm ± 2.38, p = 0.008), 3D
joints. The intact ACL spans the joint from the anterior tibia to the
medial wall of the lateral femoral condyle (a). Complete injury of motion after peak load (high-displacement: 2.84 mm ±
the ACL (b) occurred by either midsubstance rupture, bony 1.57; low-displacement: 1.70 mm ± 1.19, p = 0.036),
avulsion from the femoral footprint, or a combination injury motion after peak load in the transverse plane
where one bundle of the ACL was avulsed and the other bundle
exhibited a midsubstance rupture. F, femur; T, tibia; P, patella; (high-displacement: 2.69 mm ± 1.66; low-displacement:
Med, medial; Lat, lateral; ACL, anterior cruciate ligament. 1.22 mm ± 0.88, p = 0.014), total motion in the trans-
verse plane (high-displacement: 6.58 mm ± 1.37; low-
LSHD had significantly higher total sagittal motion displacement: 4.41 mm ± 2.27, p = 0.004), motion after
compared to HSLD. peak load in the coronal plane (high-displacement:
Complete-injury specimens had significantly higher 2.74 mm ± 1.49; low-displacement: 1.43 mm ± 1.04,
total motion after peak load in the transverse plane p = 0.016), total motion in the coronal plane (high-
(2.58 mm ± 1.65) compared to specimens that did displacement: 4.70 mm ± 1.07; low-displacement:
exhibit complete injury (1.25 mm ± 0.97, p = 0.031). 3.20 mm ± 1.76, p = 0.013), total motion in the sagittal
There was significantly higher net displacement in the plane (high-displacement: 5.46 mm ± 0.92; low-dis-
coronal plane in injured specimens (2.15 mm ± 0.87) placement: 3.72 mm ± 1.76, p = 0.002), and motion
compared to non-injured specimens (1.45 mm ± 0.69, after peak load in the sagittal plane (high-displacement:
p = 0.034). Joint motion plots indicate that complete- 1.75 ± 1.11; low-displacement: 0.95 ± 0.85; p = 0.037).
injury specimens exhibited tibial internal rotation up to Differences between grouped high and low speed groups
the point of ACL failure, and after failure, the tibia were less pronounced, and the only difference was ob-
displaced caudally, anteriorly and exhibited external served in transverse motion after peak load (high speed:
rotation (Figs. 4a and 4b), indicating tibial subluxation 2.59 mm ± 1.64; Low speed: 1.35 ± 1.13, p = 0.043).
2472 MAERZ et al.

TABLE 1. Relative tibiofemoral joint motion during injury loading (all specimens)

Transverse motion
Coronal motion (mm ± SD) (mm ± SD) Sagittal motion (mm ± SD) 3D motion (mm ± SD)

Total After peak load Total After peak load Total After peak load Total After peak load

HSHD 4.98 ± 1.19a 3.20 ± 1.47 6.69 ± 1.77 3.40 ± 1.50a 5.10 ± 0.83 1.73 ± 0.81 7.25 ± 1.63 3.23 ± 1.26
HSLD 2.96 ± 1.70a 1.53 ± 1.01 4.25 ± 2.56 1.37 ± 0.98a b
3.61 ± 2.05 1.00 ± 0.72 5.00 ± 2.61 1.85 ± 1.08
b
LSHD 4.28 ± 0.77 2.05 ± 1.34 6.44 ± 0.70 1.62 ± 1.37 5.93 ± 0.85 1.76 ± 1.49 7.07 ± 0.78 2.34 ± 1.88
LSLD 3.49 ± 2.00 1.33 ± 1.15 4.57 ± 2.18 1.04 ± 0.81 3.82 ± 1.61 0.91 ± 1.04 5.38 ± 2.34 1.55 ± 1.35

SD: standard deviation.

a
HSHD and HSLD differ.
b
HSLD and LSHD differ.

TABLE 2. Relative tibiofemoral joint motion during rupture loading (complete-injury specimens only)

Transverse motion
Coronal motion (mm ± SD) (mm ± SD) Sagittal motion (mm ± SD) 3D motion (mm ± SD)

Total After peak load Total After peak load Total After peak load Total After peak load

HSHD 4.98 ± 1.19a 3.20 ± 1.47 6.69 ± 1.77 3.40 ± 1.50a 5.10 ± 0.83 1.73 ± 0.81 7.25 ± 1.63a 3.23 ± 1.26
HSLD 2.80 ± 0.75a 1.48 ± 1.07 3.91 ± 1.49 1.21 ± 1.02a 3.44 ± 1.17b 1.01 ± 0.82 4.41 ± 1.48a 1.65 ± 1.15
LSHD 4.40 ± 0.30 2.04 ± 1.03 6.39 ± 1.80 1.75 ± 1.45 5.73 ± 0.98b 1.42 ± 1.30 7.01 ± 0.87 1.99 ± 1.60
LSLD N/A N/A N/A N/A N/A N/A N/A N/A

N/A in LSLD because no specimens ruptured in LSLD.

SD: standard deviation.
a
HSHD and HSLD differ.
b
HSLD and LSHD differ.

Joint Laxity and varus compliance (injury: 5.61 mm/N ± 1.30;

non-injury: 3.86 mm/N ± 1.51, p < 0.001) compared
Quantitative joint laxity testing results of all speci-
to non-injured specimens. There was no difference in
mens are summarized in Table 3. When all specimens
valgus displacement or valgus compliance between
are analyzed, only HSHD exhibited a significant in-
injured and non-injured specimens.
crease in AP and varus displacements compared to
There were marked differences in AP and varus laxity
uninjured limbs, with corresponding decreases in re-
when comparing high and low displacement groups.
spective compliance. There were no significant differ-
High-displacement had significantly larger AP dis-
ences in valgus laxity between any groups. When
placement (high-displacement: 0.991 mm ± 0.53; Low-
comparing only specimens that exhibited complete
displacement: 0.444 mm ± 4.3, p = 0.003), AP com-
ACL injury, HSHD, HSLD, and LSHD all had sig-
pliance (high-displacement: 0.457 mm/N ± 0.20; low-
nificant increases in AP net displacement, AP compli-
displacement: 0.217 ± 0.19, p = 0.006), varus displace-
ance, varus displacement, and varus compliance
ment (high-displacement: 3.29 mm ± 0.93; low-dis-
compared to unloaded controls. (Table 4) No group
placement: 2.55 mm ± 1.03, p = 0.031), and varus
exhibited significant increases in valgus displacement
compliance (high-displacement: 4.96 mm/N ± 1.39;
or valgus compliance. There were no differences in any
low-displacement: 3.83 mm/N ± 1.75, p = 0.021). Val-
laxity variable between groups when comparing only
gus displacement or compliance did not vary as a
specimens that exhibited complete injury.
function of displacement. When comparing high and
Specimens that exhibited complete injury had sig-
low speed groups, there were significant differences in
nificantly higher AP displacement (injured: 1.16 mm ±
AP displacement (high-speed: 0.873 mm ± 0.45; low-
0.4; non-injured: 0.257 mm ± 0.11, p < 0.001) and AP
speed: 0.569 mm ± 0.61, p = 0.026) and AP compli-
compliance (injured: 0.541 mm/N ± 0.12; non-injured:
ance (high-speed: 0.447 mm/N ± 0.22; low-speed:
0.128 mm/N ± 0.06, p < 0.001) compared to non-in-
0.228 mm/N ± 0.18, p = 0.004), but no difference was
jured specimens. Specimens that had complete injury
found in varus displacement, varus compliance, valgus
had significantly higher varus displacement (injury:
displacement, or valgus displacement.
3.64 mm ± 0.78; non-injury: 2.18 mm ± 0.69, p < 0.001)
 Biomechanics of Noninvasive Rat ACL Injury 2473

FIGURE 4. Joint motion in the coronal plane (a, c) and transverse plane (b, d) in specimens that exhibited complete ACL injury (a,
b) and specimens that did not exhibit complete injury (c, d). The tibia internally rotates up to the point of peak load or ACL failure,
but tibial external rotation with pronounced anterior and caudal subluxation is noted only in specimens that exhibited ACL failure
and complete injury.

TABLE 3. Laxity Testing Results

Anterior-posterior laxity Varus laxity Valgus laxity

Net displ. Compliance Net displ. Compliance Net displ. Compliance

(mm ± SD) (mm/N ± SD) (mm ± SD) (mm/N ± SD) (mm ± SD) (mm/N ± SD)

Control 0.350 ± 0.12 0.162 ± .05 2.20 ± 0.27 3.15 ± 0.59 2.50 ± 0.39 3.98 ± 0.98
HSHD 1.07 ± 0.27* 0.566 ± 0.13* 3.53 ± 0.98* 5.26 ± 1.33* 2.58 ± 0.34 3.74 ± 0.53
HSLD 0.648 ± 0.53 0.313 ± 0.23 2.94 ± 0.97 4.57 ± 2.00 2.60 ± 0.35 3.55 ± 0.56
LSHD 0.900 ± 0.73 0.335 ± 0.20 3.02 ± 0.85 4.63 ± 1.48 2.61 ± 0.66 3.86 ± 1.29
LSLD 0.239 ± 0.10 0.120 ± 0.05 2.16 ± 0.98 3.09 ± 1.14 3.12 ± 0.99 4.54 ± 1.30

Displ: displacement; SD: standard deviation.

* Significant difference to control.

DISCUSSION between surgical and nonsurgical treatment.19,22 This

indicates that the acute time frame following traumatic
PTOA is a prevalent and debilitating condition joint injury may play a crucial role in the onset and
which develops following joint injury. The incidence of progression of PTOA. As such, animal models accu-
PTOA following ACL rupture has been reported as rately reproducing the biologic cascades after joint
high as 90%,10,21,29,33,38 and the incidence of PTOA injury are critical, but the most prevalent model for
following ACL rupture does not appear to vary PTOA in the rat is surgical ACL transection. To date,
2474 MAERZ et al.

TABLE 4. Laxity testing results (complete-injury specimens only)

Anterior-posterior laxity Varus laxity Valgus laxity

Net displ. Compliance Net displ. Compliance Net displ. Compliance

(mm ± SD) (mm/N ± SD) (mm ± SD) (mm/N ± SD) (mm ± SD) (mm/N ± SD)

Control 0.350 ± 0.12 0.162 ± .05 2.20 ± 0.27 3.15 ± 0.59 2.50 ± 0.39 3.98 ± 0.98
HSHD 1.07 ± 0.27* 0.566 ± 0.13* 3.53 ± 0.98* 5.26 ± 1.33* 2.58 ± 0.34 3.74 ± 0.53
HSLD 1.20 ± 0.48* 0.578 ± 0.12* 4.07 ± 0.33* 6.72 ± 1.66* 2.81 ± 0.17 3.63 ± 0.25
LSHD 1.31 ± 0.60* 0.472 ± 0.07* 3.55 ± 0.53* 5.57 ± 0.81* 2.50 ± 0.21 3.52 ± 0.26
LSLD N/A N/A N/A N/A N/A N/A

N/A in LSLD because no specimens ruptured in LSLD.

Displ: displacement; SD: standard deviation.
* Significant difference to control.

models of noninvasive ACL injury in the rat are and a combination avulsion-rupture injury in our high
sparse, and, to this end, the purpose of this study was speed group. Although differences in joint biome-
to biomechanically characterize the tibial compression chanics between the mouse and rat may cause some
model in the rat. We found that complete ACL injury discrepancies in injury type, this finding likely indicates
can be repeatedly induced with a 3 mm compressive that 8 mm/s represents the lower end of the loading
displacement of the tibia at 8 mm/s when the knee is rate range necessary to induce midsubstance ACL
flexed to 100. ACL injury causes marked increases in rupture, and future studies are necessary to determine
AP and varus laxity, and joint displacement in the which loading rate repeatedly induces ACL rupture in
transverse and coronal planes was the most con- the rat. Furthermore, although Lockwood et al.20 did
tributory to ACL injury in our model. not observe extensive differences in OA-like changes
Only a few studies have utilized noninvasive joint between avulsion and midsubstance rupture injury
loading to induce ACL rupture in the rat, and no study modes, additional investigation should determine
has utilized the tibial compression model in the rat. whether the extent of in vivo degenerative joint changes
Tang et al. utilized a rotational model to study post- varies between the avulsion, midsubstance rupture,
injury protein expression.37 In this model, ACL injury and combination injury observed in our study.
was induced by rotating the tibia and femur in oppo- ACL rupture in humans occurs most prevalently in
site directions in 120 of joint extension. No charac- non-contact scenarios during rapid deceleration or
terization of the injury or post-injury joint mechanics change of direction,3,4 and several biomechanical mo-
was performed. The tibial compression model has been tions have been shown to induce ACL rupture, in-
previously employed in the mouse. Christiansen et al. cluding anterior tibial translation due to high-
applied a 12 N tibial compressive load to induce a magnitude tibiofemoral compression,26 internal tibial
noninvasive ACL injury,8 which induced OA-like rotation,27 and combined tibial rotation and valgus
changes such as cartilage degeneration and subchon- joint stress.4,27 Meyer et al. have shown that tibial
dral bone changes. The load was applied at 1 mm/s compression causes transient anterior subluxation of
and the authors stated that bony avulsion of the ACL the tibia relative to the femur.27 In their study, internal
occurred in all specimens. A follow-up study by that tibial rotation was noted to occur during pre-failure
group indicated that true ACL rupture was achieved at compressive loading, and following ACL rupture, the
higher speed (500 mm/s) whereas lower speed injury tibia rotated externally. In our study, we observed that
(1 mm/s) caused a ‘‘disruption of the ACL with an joint motion in the transverse and coronal planes was
avulsion fracture from the posterior femur.’’ 20 Inter- markedly higher in specimens that exhibited complete
estingly, histologic evidence of articular cartilage de- ACL injury compared to specimens that did not ex-
generation did not vary between injury mode (i.e., hibit injury. The medial–lateral motion component of
avulsion vs. mid-substance rupture, and only acute these planes was most pronounced, and motion plots
trabecular bone changes were impacted by injury indicate that this is tibial rotation and/or translation.
mode). Our data corroborates that loading rate de- Due to anatomic size restrictions, it was only possible
termines injury type. We also observed ACL avulsion to place one marker in each bone, and we, therefore,
at 1 mm/s, but since the 8 mm/s loading rate in our cannot definitively conclude whether rotation or
‘‘high speed’’ group was markedly lower than the translation predominated. However, our results of
500 mm/s ‘‘high speed’’ group used by Lockwood motion in the medial–lateral direction after rupture are
et al.,20 we observed avulsion, midsubstance rupture, consistent with Meyer’s biomechanics study of human
 Biomechanics of Noninvasive Rat ACL Injury 2475

ACL rupture, indicating that not only vertical and tures and bony avulsions, but we cannot exclude the
anterior tibial displacement occurs during and after possibility of missing microfractures or very small
ACL rupture. We observed that only specimens that avulsions given the lCT imaging parameters. Due to
experienced ACL injury exhibited increased LCL lax- size restrictions and the risk of fracture, we were only
ity, a concomitant joint injury also observed in humans able to place one retroreflective marker in the tibia and
following ACL rupture, although rarer than MCL in- femur, which limited the information able to be
jury.17,18 Since total motion after peak load in the ascertained from motion capture data. While we did
transverse plane as well as net displacement in the not observe any gross injury to or disruption of the
coronal plane was significantly higher in injured spe- PCL, MCL, and LCL, we cannot rule out mi-
cimens, we conclude that these motions cause LCL crostructural damage. Since increased varus laxity was
injury, and tibial external rotation immediately after observed, we conclude that LCL damage was induced
ACL failure is the most likely cause. Meyer et al. did but unable to be macroscopically visualized. Lastly, as
not investigate LCL injury following ACL rupture, nor this is a biomechanics study of cadaveric rats, the
were joint laxity tests performed, and we cannot con- biomechanical behavior of live tissue may not be fully
clude whether our proposed mechanism of LCL injury represented in our data set, and no information about
is fully representative of concomitant LCL injury in the actual impact of this ACL injury on PTOA can be
humans. Future studies assessing the exact pathome- gathered. Further studies are underway in our insti-
chanism of LCL injury in both our models and human tution to assess degenerative joint changes in the rat
injury are necessary. following noninvasive ACL injury using the model
We did not measure any increases in valgus laxity presented in this study.
due to ACL injury. Medial collateral ligament (MCL) This study biomechanically characterized the tibial
injury is a common concomitant injury during ACL compression model of ACL injury in the rat. A re-
rupture in humans,35,40 notably during the common producible ACL injury can be induced by the appli-
‘‘valgus collapse’’ mode of ACL injury involving the cation of a tibial axial displacement of 3 mm at 8 mm/s
pivot shift mechanism.4,16,31 In a biomechanics study when the knee is flexed at ~100, and this injury causes
assessing relative ACL and MCL strain during concomitant LCL injury due to post-failure tibiofe-
simulated landing-induced ACL rupture in human moral motion in the transverse and coronal planes.
cadaveric specimens, Quatman et al. found that MCL Motion-capture data indicates that this injury is simi-
strain is most pronounced in multiplanar tibiofemoral lar to human modes of ACL injury, and this injury
loading, which includes anterior shear, abduction, and model may be applied in future rat studies of PTOA
internal tibial rotation.32 They stated that the MCL and/or ACL reconstruction.
strain observed in their study is not sufficient to induce
MCL injury, but a valgus and internal rotation mo-
ment are necessary to strain the MCL. Their simulated
landing model allowed unconstrained tibial rotation
and varus-valgus joint deformation at 25 of flexion. REFERENCES
Our loading protocol rigidly fixed the paw at 100 of
knee flexion, which inherently constrains tibial internal 1
Altman, R., et al. Biomechanical and biochemical prop-
rotation and valgus deformation (i.e., tibial abduc- erties of dog cartilage in experimentally induced os-
tion). Other studies have indicated that concomitant teoarthritis. Ann. Rheum. Dis. 43:83–90, 1984.
2
MCL injury during ACL rupture is induced by valgus Anderson, D. D., et al. Post-traumatic osteoarthritis: im-
proved understanding and opportunities for early inter-
loading,25,36 and we, therefore, conclude that valgus
vention. J. Orthop. Res. 29:802–809, 2011.
injury was not induced in our model due to the lack of 3
Boden, B. P., et al. Mechanisms of anterior cruciate liga-
valgus loading and fully unconstrained tibial internal ment injury. Orthopedics 23:573–578, 2000.
4
rotation. Boden, B. P., et al. Noncontact anterior cruciate ligament
This study has several limitations. Our motion injuries: mechanisms and risk factors. J. Am. Acad. Orthop.
Surg. 18:520–527, 2010.
capture data is inherently limited by the system’s 5
Brophy, R. H., et al. Effect of short-duration low-magni-
resolution at a given speed. At a capture rate of 179 Hz tude cyclic loading versus immobilization on tendon-bone
and with 4 mm retroreflective markers, our resolution healing after ACL reconstruction in a rat model. J. Bone
was 279 lm with a 98.07% accuracy at 8 mm/s, and a Joint Surg. 93:381–393, 2011.
6
49.02 lm resolution with 99.31% accuracy at 1 mm/s. Brown, T. D., et al. Posttraumatic osteoarthritis: a first
estimate of incidence, prevalence, and burden of disease. J.
Inherently higher error at the higher loading rate could
Orthop. Trauma 20:739–744, 2006.
not be avoided, and high-speed groups may therefore 7
Buckwalter, J. A., and T. D. Brown. Joint injury, repair,
have higher error. Secondly, we utilized lCT imaging and remodeling: roles in post-traumatic osteoarthritis.
at a resolution of 36 lm to assess specimens for frac- Clin. Orthop. Relat. Res. 423:7–16, 2004.
2476 MAERZ et al.
8 25
Christiansen, B., et al. Musculoskeletal changes following Mazzocca, A. D., et al. Valgus medial collateral ligament
non-invasive knee injury using a novel mouse model of post- rupture causes concomitant loading and damage of the
traumatic osteoarthritis. Osteoarthr. Cartil. 20:773–782, anterior cruciate ligament. J. Knee Surg. 16:148–151, 2003.
26
2012. Meyer, E. G., and R. C. Haut. Excessive compression of
9
Dirschl, D. R., et al. Articular fractures. J. Am. Acad. the human tibio-femoral joint causes ACL rupture. J.
Orthop. Surg. 12:416–423, 2004. Biomech. 38:2311–2316, 2005.
10 27
Gillquist, J., and K. Messner. Anterior cruciate ligament Meyer, E. G., and R. C. Haut. Anterior cruciate ligament
reconstruction and the long term incidence of gonarthrosis. injury induced by internal tibial torsion or tibiofemoral
Sports Med. 27:143–156, 1999. compression. J. Biomech. 41:3377–3383, 2008.
11 28
Guilak, F., et al. Mechanical and biochemical changes in Mifune, Y., et al. Tendon graft revitalization using adult
the superficial zone of articular cartilage in canine ex- anterior cruciate ligament (ACL)-derived CD34+ cell
perimental osteoarthritis. J. Orthop. Res. 12:474–484, 1994. sheets for ACL reconstruction. Biomaterials 34:5476–5487,
12
Hashimoto, S., et al. Development and regulation of os- 2013.
29
teophyte formation during experimental osteoarthritis. Myklebust, G., and R. Bahr. Return to play guidelines
Osteoarthr. Cartil. 10:180–187, 2002. after anterior cruciate ligament surgery. Br. J. Sports Med.
13
Jansen, H., et al. Effects of low-energy NMR on post- 39:127–131, 2005.
30
traumatic osteoarthritis: observations in a rabbit model. Onur, T. S., et al. Joint instability and cartilage compres-
Arch. Orthop. Trauma Surg. 131:863–868, 2011. sion in a mouse model of posttraumatic osteoarthritis. J.
14
Jansen, H., et al. Detection of vascular endothelial growth Orthop. Res. 32:318–323, 2014.
31
factor (VEGF) in moderate osteoarthritis in a rabbit Quatman, C. E., and T. E. Hewett. The anterior cruciate
model. Ann. Anat. 194:452–456, 2012. ligament injury controversy: is ‘‘valgus collapse’’ a sex-
15
Kadonishi, Y., et al. Acceleration of tendon–bone healing specific mechanism? Br. J. Sports Med. 43:328–335, 2009.
32
in anterior cruciate ligament reconstruction using an en- Quatman, C. E., et al. Preferential loading of the ACL
amel matrix derivative in a rat model. J. Bone Joint Surg. compared with the MCL during landing a novel in sim
Br. 94:205–209, 2012. approach yields the multiplanar mechanism of dynamic
16
Koga, H., et al. Mechanisms for noncontact anterior cru- valgus during ACL injuries. Am. J. Sports Med.
ciate ligament injuries knee joint kinematics in 10 injury 42:177–186, 2013. doi:10.1177/0363546513506558.
33
situations from female team handball and basketball. Am. Roos, H., et al. Osteoarthritis of the knee after injury to the
J. Sports Med. 38:2218–2225, 2010. anterior cruciate ligament or meniscus: the influence of time
17
Lee, J. K., et al. Anterior cruciate ligament tears: MR and age. Osteoarthr. Cartil. 3:261–267, 1995.
34
imaging compared with arthroscopy and clinical tests. Ruan, M. Z., et al. Quantitative imaging of murine os-
Radiology 166:861–864, 1988. teoarthritic cartilage by phase-contrast micro-computed
18
Levine, J. W., et al. Clinically relevant injury patterns after tomography. Arthritis Rheum. 65:388–396, 2013.
35
an anterior cruciate ligament injury provide insight into Sankar, W. N., et al. Combined anterior cruciate ligament
injury mechanisms. Am. J. Sports Med. 41:385–395, 2013. and medial collateral ligament injuries in adolescents. J.
19
Linko, E., et al. Surgical versus conservative interventions Pediatr. Orthop. 26:733–736, 2006.
36
for anterior cruciate ligament ruptures in adults. Cochrane Shin, C. S., A. M. Chaudhari, and T. P. Andriacchi. The
Database Syst. Rev. 2:CD001356, 2005. effect of isolated valgus moments on ACL strain during
20
Lockwood, K. A., et al. Comparison of loading rate-de- single-leg landing: a simulation study. J. Biomech.
pendent injury modes in a murine model of post-traumatic 42:280–285, 2009.
37
osteoarthritis. J. Orthop. Res. 32:79–88, 2014. Tang, Z., et al. Contributions of different intraarticular
21
Lohmander, L., et al. High prevalence of knee os- tissues to the acute phase elevation of synovial fluid MMP-
teoarthritis, pain, and functional limitations in female 2 following rat ACL rupture. J. Orthop. Res. 27:243–248,
soccer players twelve years after anterior cruciate ligament 2009.
38
injury. Arthritis Rheum. 50:3145–3152, 2004. Von Porat, A., E. M. Roos, and H. Roos. High prevalence
22
Lohmander, L. S., et al. The long-term consequence of of osteoarthritis 14 years after an anterior cruciate ligament
anterior cruciate ligament and meniscus injuries os- tear in male soccer players: a study of radiographic and
teoarthritis. Am. J. Sports Med. 35:1756–1769, 2007. patient relevant outcomes. Ann. Rheum. Dis. 63:269–273,
23
Mankin, H., A. Grodzinsky, and J. A. Buckwalter. Ar- 2004.
39
ticular cartilage and osteoarthritis. In: Orthopaedic Basic Wu, Y., et al. The profile of MMP and TIMP in injured rat
Science: Foundations of Clinical Practice, edited by T. ACL. Mol. Cell Biomech. 7:115–1124, 2009.
40
Einhorn. Chicago, IL: American Academy of Orthopaedic Yoon, K. H., J. H. Yoo, and K.-I. Kim. Bone contusion
Surgeons, 2007. and associated meniscal and medial collateral ligament in-
24
Martin, J., and J. Buckwalter. Post-traumatic osteoarthri- jury in patients with anterior cruciate ligament rupture. J.
tis: the role of stress induced chondrocyte damage. Bone Joint Surg. 93:1510–1518, 2011.
Biorheology 43:517–521, 2006.