Our novel animal model in which the medial and lateral retinacula of the patella were released showed that patellar hypermobility influenced the development of the femoral trochlea and led to trochlear dysplasia, which worsened with age. Both patellar hypermobility and patellar dislocation influenced the development of the femoral trochlea, mainly expressed as abnormal trochlear shape and cartilage thickness. The stress on the femoral trochlea greatly affected the developing trochlea. Excessive or insufficient stress may cause thinning and deformation of the cartilage and bone.
The setting of two experimental groups (PHG and PDG) and a control group (SG) increased the reliability and sensitivity of the present study. Previous research has found that excessive mechanical stress above the lateral condyle of the femur following patellar dislocation may result in a shallower trochlear groove and lower lateral condyle . However, patients with joint hypermobility have a high incidence of femoral trochlear dysplasia . In contrast to the patellar dislocation model, the patellar hypermobility model created lesser pressure on the trochlear groove and bilateral condyles in the present study.
The patellofemoral joint is attached to multiple soft tissue structures to maintain the stability of the patella and ensure the range of motion of the knee joint. The medial patellar retinaculum consists of the medial patellofemoral ligament, medial patellotibial ligament, and medial patellomeniscal ligament, and provides the lateral stability of the patella , with the medial patellofemoral ligament accounting for 53–67% of the medial restraining force . Lateral instability due to an insufficient medial patellofemoral ligament has been extensively confirmed anatomically and biomechanically . The lateral patellar retinaculum consists of the iliotibial band, lateral patellofemoral ligament, and lateral patellotibial ligament, and provides the medial stability of the patella . A cadaveric study confirmed that the lateral retinaculum restrains the lateral translation of the patella in an extended knee . Furthermore, the lateral patellofemoral ligament is important in protecting the patella from medial instability . In the present study, the release of the entire medial and lateral stability structures of the patella resulted in patellar hypermobility, loosening of the patellofemoral joint, and finally gave rise to obvious femoral trochlear dysplasia in growing rats.
The changes in the geometrical morphology of the femoral trochlea over time remain controversial. Nietosvaara et al.  measured the CSA in 50 normal children on ultrasonography and reported no significant change in the angle with increasing age. Furthermore, a retrospective analysis of magnetic resonance images of adolescents with trochlear dysplasia reported no significant differences in the shape of the dysplastic trochlea . In contrast, Øye et al.  used ultrasonography to track the femoral trochlear groove in 174 newborns until the age of 6 years, and found significant differences between the normal group and the trochlear dysplasia group in the changes in the trochlear groove angle; the trochlear groove angle of the normal group increased, while the angle of the trochlear dysplasia group decreased. However, all of the abovementioned studies measured the femoral trochlear groove as the CSA on ultrasonography or magnetic resonance images. The present study revealed that the BSA and CSA decreased significantly (by 8.1% for the BSA and by 7.9% for the CSA) with age in the normal biomechanical environment of the femoral trochlea in the SG. However, the PHG and PDG showed no significant changes in the CSA or BSA over time, which was probably due to the abnormal molding of cartilage and bone under abnormal stress, and only showed changes in the linear measurements of the femoral trochlea without changes in morphological variables such as the trochlear sulcus angle.
It remains unclear whether femoral trochlear dysplasia is congenital due to genetic factors or occurs due to stress stimulation. Glard et al.  reported that the CSA in the fetus appears to be the same as that in adults and is independent of age and sex. Similarly, Parikh et al. supported the genetic origin of trochlear dysplasia , and Miller et al.  declared that recurrent patellar dislocation appears to be inherited. In addition, Dejour et al.  reported that 96% of patients with a history of patellar dislocation have radiographic evidence of trochlear dysplasia. However, there is currently no direct evidence of the genetic origin of trochlear dysplasia. Numerous scholars believe that abnormalities in the static and dynamic relationships between the patella and femoral trochlea fail to appropriately stimulate the femoral trochlea, resulting in femoral trochlear dysplasia. The shape of developing bone is altered in response to function, and the architecture of cancellous bone changes with mechanical stress . A case report of a 16-year-old boy with trochlear dysplasia after a below-knee amputation confirmed that certain biomechanical input is essential for the formation of the trochlear groove . A linear relationship between trochlear dysplasia and the pressure in the patellofemoral joint has been demonstrated in animal models of patellar dislocation , patella alta , and patellar resection . In the current study, the PHG and PDG developed trochlear dysplasia, which was expressed as a shallower and wider trochlear groove that worsened with age. The present findings support the theory that stress on the femoral trochlea is a key factor in the development of the trochlea, and that excessive or insufficient stress leads to trochlear dysplasia.
Articular cartilage is mainly composed of chondrocytes and extracellular matrix. The synthesis and metabolism of the extracellular matrix is regulated by chondrocytes, and the main pressure-bearing structural components of chondrocytes are collagen and proteoglycan. The thickness and geometry of cartilage are related to the stress on the joint during development . In an embryonic chick model, immobilization decreases the cartilage matrix formation and mechanical properties of the tibiofemoral articular cartilage of fixed embryos compared with controls . Hagiwara et al.  confirmed that decreasing the load also creates catabolic responses in the articular cartilage of rats with the knee fixed in flexion. In agreement with these previous studies, the present study showed that the full-layer cartilage thickness of the trochlear groove in the PHG and the medial and central cartilage thicknesses in the PDG were significantly thinner at 3 weeks postoperatively compared with the SG; compared with the SG, the PHG showed a 12.6% decrease in the CTL, 13.6% decrease in the CTM, and 14.0% decrease in the CTM, while the PDG showed a 10.3% in the CTM and 16.0% decrease in the CTM. These differences may be because reduced loading created catabolic responses. However, at 6 weeks after surgical intervention, the CTC was significantly thicker in the PHG and PDG than the SG (by 17.4% in the PHG, and by 9.4% in the PDG), while the CTM was similar in the PHG, PDG, and SG. This may be due to the consistent matching morphology of the patella and trochlear groove following animal growth, resulting in a balance between cartilage catabolism and anabolism in accordance with the alterations in pressure on the cartilage. An in vitro bovine cartilage explant system showed that excessive mechanical stress damages the extracellular matrix, changes the balance of chondrocytes, and finally results in catabolism exceeding anabolism . However, caution is needed when comparing the in vivo situation to the results of in vitro experiments. In the present in vivo experiment, the CTL of the PDG was significantly thinner (by 11.3%) at 6 weeks postoperatively, which agreed exceptionally well with the findings of the previous study. Of note, there was no significant difference between the PHG and SG in the TCL and TCM. Although there was a gradual matching of the physiological morphology with growth and development as stated above, the cartilage damage caused by overload has poor repairability, while the cartilage thinning caused by load reduction has good reversibility. In healthy knee joints, the articular cartilage is usually thicker in the center of the groove, and the evaluation of the femoral groove is more accurate when it is based on the shape of cartilage than bone . Similar results were found in the present study. In addition, stress changes the bone shape by affecting osteoblasts and osteoclasts to add or remove bone to the appropriate surface . Biomechanics play an important role in bone formation, which begins with a primary cartilaginous matrix that later calcifies and forms bone via the endochondral ossification process . In the present study, the BSA and CSA in the SG were significantly smaller at 6 weeks after surgical intervention than at 3 weeks postoperatively. The alteration of the cartilage and bone of the trochlea with aging supports the previous findings.
Articular cartilage is a highly differentiated tissue that has limited regenerative capacity due to its avascularity . A previous study showed that trochlear dysplasia is improved by early reduction of patellar subluxation, which reduces the risk of secondary surgery and protects the articular cartilage . The present study simulated patellar hypermobility and patellar subluxation, both of which resulted in trochlear dysplasia. This suggests that patients with patellar hypermobility and even patients with global joint hypermobility may be at risk of developing trochlear dysplasia. If the present findings are confirmed in further studies, measurements of patellar hypermobility may become part of routine monitoring and prevention of trochlear dysplasia. In addition, in clinical operations performed to adjust the load between the patella and the femoral trochlea in adolescents, the tension of the reconstructed ligaments should be adjusted to avoid excessive or insufficient tension.
The present study has some limitations. First, the animal model does not completely match the biomechanics and anatomical structure of the human patellofemoral joint, which may lead to errors. Second, the cross-sectional design is inferior to longitudinal studies that can better assess growth over time; however, there are few methods for assessing cartilage in living animals. Third, the cartilage thickness may have been underestimated due to histological management, as the processes of fixation, decalcification, and staining may change the cartilage thickness. Fourth, there were only two assessment timepoints. In future, the number of sampling timepoints will be increased to closely observe the changes in the femoral trochlear bone and cartilage with age.