|Home | About | Journals | Submit | Contact Us | Français|
Animal models are an indispensable tool for developing and testing new clinical applications regarding the treatment of acute injuries and chronic diseases of the knee joint. Therefore, the purpose of this study was to compare the anatomy of the intra-articular structures of the human knee to species commonly used in large animal research studies.
Fresh frozen cow (n=4), sheep (n=3), goat (n=4), dog (n=4), pig (n=5), rabbit (n=5), and human (n=4) cadaveric knees were used. Passive range of motion and intra-articular structure sizes of the knees were measured, the structure sizes normalized to the tibial plateau, and compared among the species.
Statistically significant differences in the range of motion and intra-articular structure sizes were found among all the species. Only the human knee was able to attain full extension. After normalization, only the pig ACL was significantly longer than the human counterpart. The tibial insertion site of the ACL was split by the anterior lateral meniscus attachment in the cow, sheep, and pig knees. The sheep PCL had two distinct tibial insertion sites, while all the other knees had only one. Furthermore, only in human knees, both lateral meniscal attachments were located more centrally than the medial meniscal attachments.
Despite the relatively preserved dimensions of the cruciate ligaments, menisci, and intercondylar notch amongst human and animals, structural differences in the cruciate ligament attachment sites and morphology of the menisci between humans and animals are important to consider when selecting an animal model.
Injuries of the intra-articular structures of the knee - the cruciate ligaments, menisci, and cartilage - are becoming increasingly prevalent, and are often associated with severe long term consequences such as post-traumatic osteoarthritis1–3. These facts illustrate the need for new and effective treatments for these injuries4, 5. Currently, intensive research efforts have produced a considerable amount of basic science knowledge for cruciate ligament 6–8, meniscal 9, 10 and cartilage 11, 12 injuries. However, before human patients can benefit from such data, translational work in large animal models is often necessary to comply with regulatory requirements and the principles of translational medicine13.
Large animal models provide a uniform, experimental platform for development and evaluation of the effectiveness and safety of novel treatments14. Most commonly, sheep, goats, pigs, dogs, and rabbits are used as models of the human knee to test implants or to discover determinants of disease progression15–19. However, there exists no study that systematically and quantitatively compares the anatomy of the intra-articular structures of these animals with that of the human knee joint. Hence, the objective of this study was to compare the intra-articular knee anatomy of the above mentioned six animal species and the human knee in a systematic and quantitative way to define best-practice models for experimental models in knee surgery.
The study was designed to compare the anatomy and sizes of the intra-articular structures of the knees of six animal species: cow (n=4), sheep (n=3), goat (n=4), dog (n=4), pig (n=5), rabbit (n=5), with the human knee (n=5). All animal knees were retrieved from subjects undergoing euthanasia for IACUC-approved research studies. The cow specimens were from female adolescent Angus cattle, the sheep knees from female adult Columbia sheep and the goat knees from female adult Alpine goats. The dog knees were from four female adult large breed dogs, the rabbit knees from five adult female New Zealand white rabbits, and the pig knees from adult female Yorkshires. All animal knees were harvested within 2 hours of euthanasia. The human knees were obtained from a willed body program and were obtained from 4 females with an age range of 47 to 60 years. All specimens were frozen at the time of collection and maintained at −20°C until the time of analysis.
Before dissection, specimens were thawed overnight and placed in an upright stand. The passive range of motion was measured using a goniometer with 1° gradation (Quint Measuring Systems, Sam Ramon, CA). The arms of the goniometer were aligned with the femoral and tibial shafts. Prior to the measurement, the knee was moved through a full range of motion to determine the axis of the joint rotation; the center of the goniometer was then placed over this axis. All angles were measured relative to the neutral 0° axis in extension of the femur shaft. After the goniometer measurements, the surrounding skin and musculature of the knees was removed leaving the capsule intact. The dissection started anteriorly with the removal of the patella, patellar tendon, and fat pad. Extraneous capsular fascia, fatty tissue and ligamentum mucosum were removed to give a clear view of the anterior cruciate ligament (ACL) tibial insertion, posterior cruciate ligament (PCL) femoral insertion and anterior meniscal attachments. Next, the medial and lateral aspects of the knee including both collateral ligaments were dissected to reveal the medial and lateral meniscus. Removal of the posterior capsule exposed the tibial PCL origin, femoral ACL origin and the posterior meniscal attachments. Following this, the anterior attachments of the menisci were dissected from the tibia flush with the bone allowing an unobstructed view at the ACL insertion. The posterolateral meniscal attachment to the femur was transected flush with the bone to examine and remove the tibial origin of the PCL and expose the posterior insertion of the medial meniscus. Finally, the menisci were separated from their skeletal attachments. Vernier calipers (steel, double scaled, Fisher Scientific, Hanover Park, IL) were used to measure the meniscal width midway between the anterior and posterior horn as the distance from the free edge to the former capsule; the meniscal length was defined as the anterior-posterior dimension of the meniscus. ACL and PCL length and width were determined before the ligaments were excised from the knee. The length of the ACL was measured as the distance between the anterior tibial and femoral attachments, while the length of the PCL was defined as the distance between the posterior aspects of the tibial and femoral attachments. The ligament widths were measured at the mid-level of the ligaments in the sagittal plane. ImageJ 1.44 (NIH, Bethesda, MD, www.nih.gov) was used to measure the area of the insertion sites of the cruciates and menisci in the transverse plane after all soft tissue had been removed. Tibial plateau width was defined as the widest extent of the plateau in the frontal plane. The intercondylar notch width was measured at half the height from the distal most point of the femoral condyles to the apex of the notch. Each structure was measured three times and the median value was reported.
In addition to the direct measurements, all measurements were also normalized by the width of tibial plateau to facilitate comparison between knees of different sizes. This ratio between the width of the tibial plateau and structure size was referred to as the Tibial Index. Both the measured values and the values normalized by the Tibial Index were used in the comparative statistical analysis. Differences between groups were determined using an analysis of variance (ANOVA), with significance levels set at p < 0.05 for the overall ANOVA and Bonferroni adjustment for post-hoc testing to determine the significance between groups. All these calculations were performed using Intercooled STATA 10 (Statacorp LP, College Station, TX, USA).
The passive range of motion of the knees revealed that only the human knee could be brought to full extension, whereas neither of the knees of the quadruped animals reached a neutral 0° angle (Table 1). A difference in the flexion angle could only be seen in dog and rabbit knees, which had a statistically significant higher flexion angle compared to the human knees.
Photographs of the cruciate ligaments, tibial plateau, and menisci are in Figure 1.
The human ACL was significantly wider than the dog and rabbit ACL (p<0.001) and significantly longer than the goat, dog, and rabbit ACL (p<0.01).
No significant differences in the Tibial Index of the ACL width between the species were found, whereas the Tibial Index of the ACL length was significantly smaller in the human than in the pig specimens (p<0.001).
The femoral origin of the ACL was located at the posteromedial edge of the lateral condyle in all species. The tibial ACL insertion site varied among species. In the human and goat knees, the tibial insertion of the ACL was adjacent to the anterior insertion of the lateral meniscus. Cow, sheep and pig knees had the tibial insertions of the anteromedial and posterolateral ACL bundles separated by the anterior attachment of the lateral meniscus. The cow anteromedial bundle inserted across the anterior slope of the intercondylar eminence while the posterolateral bundle inserted in the center of the eminence between the tibial spines. The sheep and pig anteromedial bundles inserted at the medial edge of the anterior slope of the intercondylar eminence while the posterolateral bundles inserted on the lateral edge of the medial tibial spine deep to the anteromedial bundles and anterior lateral meniscus insertion. The ACL insertion in the rabbit knee was on the center of the tibial intercondylar eminence, posterior to the anterior lateral meniscus attachment. The dog ACL tibial insertion was on the medial slope of the intercondylar eminence.
The human PCL was significantly wider than the PCLs of all other species (p<0.01) but significantly longer only compared to the goat, dog, and rabbit PCL (p<0.001).
The normalized PCL width showed no significant differences across the species, whereas the normalized sheep and pig PCL length was significantly larger than the human (p<0.001).
The human notch width was not significantly different from the cow but was significantly larger than in the other species (p<0.001).
The Tibial Index for the notch width was significantly smaller in the sheep, goat, pig and rabbit knees when compared to the human (p<0.01).
The human medial meniscus width was significantly smaller than the cow (p<0.01) but significantly larger than the rabbit (p<0.001). With exception of the cow medial meniscus the human medial meniscus was significantly longer than the animal medial menisci.
Relative to the tibial plateau width, the human medial meniscus width was significantly smaller than the cow and pig (p<0.05).
The anterior medial meniscus bony insertion was the most anterior structure found in the knee in all species. It inserted on the anterior edge of the tibial plateau just above the tibial tuberosity in all species except the dog, where instead of individual insertions, an intermeniscal ligament connected the anterior-most sections of the medial meniscus to the lateral meniscus. The sheep and pig knees also had a small connection between the anterior medial and lateral meniscus insertions, but they did not obscure the distinct anterior insertion sites. The posterior horn of the medial meniscus inserted on the lateral edge of the posterolateral surface of the medial spine, just anterior to the PCL origin and was covered superiorly by the PCL (Figure 1, column B). The goat, dog, and rabbit insertion site areas were smaller than those noted in the human, cow, sheep, and pig knees.
The human lateral meniscus width was significantly larger than the rabbit but smaller than the cow (p<0.001). The human lateral meniscus was significantly longer than the animal lateral meniscus except for the cow (p<0.001 for all comparisons).
Neither the Tibial Index for the lateral meniscus width nor lateral meniscus length was significantly different between the species (p>0.001 for all comparisons).
The human lateral meniscus covered a smaller portion of the lateral tibial plateau when compared to all of the animal knees. Anteriorly, the lateral meniscus in human knees was attached to the lateral aspect of the lateral spine of the intercondylar eminence. The cow, sheep, and pig lateral meniscus split the ACL bundles, while the goat and dog lateral meniscus passed behind the ACL. All inserted close to the medial tibial spine. The rabbit anterior lateral meniscus was attached even more medial, adjacent to the anterior horn of the medial meniscus. Posteriorly, the menisco-femoral ligament connected the lateral meniscus to the lateral back wall of the medial femoral condyle more inferiorly in human than in the animal specimens. Human, sheep, goat, dog, and rabbit lateral meniscus had a menisco-tibial coronary ligament, attaching directly lateral to the PCL insertion, with the human, goat, and dog being less robust than the rabbit.
The cow, sheep, and goat knees did not possess a full-length fibula or proximal tibia-fibula joint. Instead a fused fibular head was attached to the lateral side of the tibial plateau and served as an attachment site for the lateral meniscus. Furthermore, all animal knees contained an intra-capsular, extra-articular, lateral long digital extensor tendon (LDET) that originated just inferior to the lateral edge of the patellar groove. The LDET was not found in the human knee. The function of the LDET appeared to be dorsiflexion of the forefoot especially during knee flexion.
In this study, significant quantitative and qualitative differences between the human intra-articular structures of the knee and the animal species were found. Quantitatively, the human ACL was longer than the goat, dog and rabbit ACLs and wider than the dog and rabbit ACLs, whereas after normalization by the tibial plateau width, only the pig ACL showed a significantly greater length when compared to the human. The human PCL was wider than all of the animal PCLs, but only longer than goat, dog and rabbit PCLs. After normalization, only the PCL length of sheep and pig differed from the human. Previous studies have quantified the length and width of the human PCL as 38 mm in length and 13 mm in width 20, measurements which are similar to those reported here of 40±7 mm and 14±5 mm, respectively. Only the cow knee had a notch width and medial meniscus length comparable to the human knee, whereas the medial and lateral meniscus width was closest to human in sheep, goats, pigs, and dogs. In prior studies of human menisci, the width of the mid-body of the medial meniscus and lateral meniscus as measured on MRI have been reported at 7.4 mm and 8.4 mm, respectively21. Those measurements were consistent with those reported here (9.5±0.6 mm and 9.8±0.7 mm for the medial meniscus and lateral meniscus, respectively)22. Qualitatively, the two bundles of the ACL (anteromedial and posterolateral bundles) and of the PCL (anterolateral and posteromedial) were not as well-defined in the human as they were in some of the animal knees. In the human and goat knees, the anterior insertion of the lateral meniscus adhered to the lateral side of the ACL which may have obscured the bundle separation in these knees. The cow, sheep, and pig knees had distinct bundles which were separated by the anterior insertion of the lateral meniscus. This is very important to consider when placing a tibial tunnel through the center of the ACL footprint while performing an ACL reconstruction on those animals, as the anterior horn of the lateral meniscus could potentially be injured. Likewise, a tibial tunnel placed posteriorly to the ACL insertion site in the goat model may also damage the anterior lateral meniscus insertion. The sheep PCL had two distinct tibial insertion sites, while all other species had only one.
Unlike the animal knees, the human knees had both horns of the lateral meniscus located more central than the medial meniscus attachments. The dog and rabbit knees did not have the anterior insertion of the lateral meniscus connecting to the ACL or separating the ACL bundles; however, it is certainly possible that these species do have cruciate bundles that are mechanically distinct even though they were not identified anatomically in this study. The posterior attachment of the lateral meniscus via the posterior menisco-femoral ligament is similar to that previously reported in dog and sheep knees23. In that study, the menisco-femoral ligament was also present in both of these species, caudal to the PCL and more oblique, features noted in our study as well. None of the four human knees in this study had anterior intermeniscal ligaments, which have been well characterized in past cadaver studies24, 25. However, in this study, the intermeniscal ligament was identified in three of the four dog knees. While the intermeniscal ligament in the human is supplemental to the bony anterior insertion sites of the menisci24, this ligament was the only anterior attachment for the medial menisci in several of the dog knees in this study. Thus, the division of an intermeniscal ligament in the dog model may destabilize the medial meniscus as the knee does not have separate bony attachments for these structures. In a prior study about the anterior insertion of the human medial meniscus, the insertion site was found to have four variations. Type I insertions were located in the flat intercondylar region of the tibial plateau; Type II occurred on the downward slope from the medial articular plateau to the intercondylar region; Type III occurred on the anterior slope of the tibial plateau; there was no firm bony insertion of the anterior horn in Type IV26. In our examination of four human knees, all the knees had Type I insertions, as did all the animal knees, with the exception of several of the dog knees which had a Type IV insertion. The variability in the insertion site anatomy in the animal species and humans may thus be important in translational studies of medial meniscus pathology.
Our measurements of the passive range of motion (Table 1) of all knees show an inherent functional difference of the knees of the quadruped animal models compared to the human knee. All animals had a physiological limit of extension at about 40° to the neutral 0° axis in extension of the femur shaft (only rabbit knees were significantly more extendable to 22°). Although this significant difference in the range of motion occurs in all the common animal models for the knee joint and therefore is less of a factor when determining the optimal animal model for the human knee, it is crucial to keep this difference in mind when translating animal data to the human case.
Limitations of the study were the exclusive study of female specimens. While this selection bias prevented our contributing to the known gender-based differences in human knee anatomy, biomechanics, and cell biology27, 28, it likely did result in a reduction of variance. Given the relatively high prevalence of knee injuries and specifically ACL ruptures in women compared to men3 was another reason to study the female knee. Another limitation in our study is the small number of specimens in each of the seven groups. A reliable statement can be given for the statistically significant comparisons, whereas not all non-significant comparisons have a power >80% and allow a correct interpretation. Apart from the specific limitations of this study, a general concern is the use of quadruped animals as knee models for the bipedal human, particularly given their range of motion differences noted in this study.
The disappearance of many of the observed differences in the cruciate and meniscal anatomy after normalization with the tibial plateau width suggests an overall conservation of relative size among species for the cruciates and menisci. Nonetheless, differences in the absolute size of the structures exist, which results in considerable changes in the load and shear the intra-articular structures are subjected to.
The overview in Table 4 evaluates these characteristics, and can act as a guideline for choosing the right model for a certain structure. Sheep and cow specimens were the best match for the human ACL in size and proportion, but had their tibial ACL insertion split by the anterior lateral meniscus, whereas the goat ACL was shorter than the human ACL, but had a similar tibial insertion site. Based on anatomy alone, sheep, cows and goats may therefore make the best models for ACL injury treatment.
Interestingly, there was less conservation of PCL dimensions among the species studied. Only the cow PCL was comparable to the human PCL in absolute length, normalized values, and anatomy of the insertion sites, whereas the human PCL was significantly wider than all of the animal specimens. This may be attributed to differences in the range of motion of bipeds compared to quadrupeds. Since the PCL prevents the knee from overextending, it might have to sustain a higher load in the human knee at full extension, in comparison to the quadruped knees which are limited in their extension (Table 1).
The size and proportion of the human medial meniscus was most similar to the sheep and goat specimens. The human lateral meniscus was most similar in size to the sheep, goat and pig specimens; however, the anatomy of the tibial insertion sites of the lateral meniscus in the goat was most comparable to that of the human knee. Considering both the medial and lateral meniscus, the animal model with the closest anatomy to the human knee was the goat.
Large animal models are a very valuable tool for developing and testing new surgical procedures. Choosing an appropriate large animal model for a certain surgical procedure, or development of a new treatment, is essential for the correct interpretation of the outcome and translation into the human model. The results of this study provide information on size and morphology of the cruciate ligaments and menisci and thus will hopefully provide some guidance on this aspect of selection of appropriate animal models for intra-articular knee surgery development.
The authors wish to thank Carla Haslauer, Linda Chao, Patrick Vavken, Eduardo Abreu, Matthew Palmer, Ashley Mastrangelo, Arthur Nedder, Mark Kelly, David Paller, David Zurakowski and Alison Biercevicz for their assistance with this project. In addition, funding was received from NIH Grants AR054099 and AR052772 (MMM) and AR049199 (BCF).
CONFLICT OF INTEREST STATEMENT
Each author certifies that he or she has no commercial associations (e.g., consultancies, stock ownership, equity interest, patent/licensing arrangements, etc) that might pose a conflict of interest in connection with the submitted article.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.