The results of this study support the initial hypothesis. Severe distal femoral varus with an aLDFA of ≥ 103° caused medial luxation of the patella when other anatomical factors of MPL were controlled. Thus, the results presented in this study lead the authors to suggest that femoral varus can lead to medial patellar luxation. This finding is in good agreement with previous studies that recommend distal femoral osteotomy is performed to correct MPL in large-breed dogs when the aLDFA is ≥ 102° [11, 13]. Interestingly, this is the first report to evaluate the biomechanical effect of distal femoral varus on the stability of the patella.
The biomechanics of the patellofemoral joint are complex, and standardization of the factors related to the joint is difficult in cadaveric and in vivo studies. Moreover, it is considered impossible to experimentally measure the whole three-dimensional state of the forces within the joint. In human medicine, FE-based computer modeling has been broadly used to investigate the biomechanics of knee joints to overcome the limitations and difficulties faced in cadaveric or in vivo studies [15, 16, 19, 22, 23]. Additionally, computer modeling is a method that avoids ethical problems related to animal use and has been increasingly used in veterinary research [20, 21, 24–26]. The current study employed computer modeling to evaluate patellar stability in the femoral trochlea with various degrees of distal femoral varus while controlling factors such as the trochlear groove shape, medial displacement of the tibial tuberosity, and internal rotation of the tibia.
The reaction force (kN) was calculated when the patella contacted the trochlea of the femur. When the aLDFA was larger than 103°, the reaction force became zero and then became a negative value during the simulation (Fig. 3). These aforementioned results indicate that a normal trochlear groove and medial ridge can keep the patella in the appropriate position when the aLDFA is less than or equal to 102°. At aLDFAs exceeding this angle, the trochlea is no longer able to withstand the shear force generated by the quadriceps mechanism between the patella and trochlea. These findings are consistent with those in clinical studies that reported a high success rate of MPL correction using DFO with an aLDFA larger than 102° [9, 13, 14].
A reaction force of 0 kN indicates patella luxation because the two elements, the femur and the patella, no longer react with each other. The negative rcforce was generated because the contact between the femur and the patella occurred in opposite directions. After luxation, the patella remained in lateral contact with the outside of the trochlear groove, and therefore, the rcforce was generated. However, since the patella was already luxated, this phenomenon was not clinically significant.
The total simulation time was set to 50 seconds because whether or not luxation occurred, time was required for the patellar motion to stabilize; thus, the time was set to 50 seconds to allow time for the motion of the patella to stabilize on all of the aLDFA models.
In terms of the outcome of luxation time, the trend was that the greater the aLDFA angle was, the shorter the time until luxation occurred, but some angles did not follow the trend consistently (Fig. 3). Patella luxation occurred at 19 seconds at 104° and approximately 15 seconds at 108°, whereas luxation occurred at approximately 20 seconds at 105° and 106° and 16 seconds at 109°. At an aLDFA of 110°, temporary luxation occurred at 11 seconds, showing an unstable state, and complete luxation occurred at 15 seconds.
In the reconstructed canine stifle model used in this study, there was no articular cartilage, synovial fluid, or other soft tissue to buffer the impact generated when the hard cortical bones of the patella and the trochlear groove collided with each other. Therefore, the above time discrepancies may have been due to the rebound phenomenon occurring as soon as the patella hit the medial ridge of the groove. Nevertheless, the fact that soft tissues were not modeled alone does not fully explain why the models with aLDFAs of 104° and 108° dislocated faster than those with other aLDFAs, and we could not determine the exact cause for this particular angle. In addition, patellar rotation, which is not possible with normal anatomy, is present after luxation, and the presence of muscles, ligaments, and the joint capsule prevent this rotation in vivo.
Our second objective was to determine the magnitude of aLDFA that would significantly increase the risk of patellar luxation in dogs. According to the normal aLDFA reported in large-breed dogs [27] and the aLDFA of MPL with a grade of less than 2 reported in small-breed dogs [10, 28], angles of 95° and 98° were considered angles to be unaffected by varus deformities. Therefore, we assumed the critical angle that causes and increases the risk of patellar luxation to be ≥ 100°. However, we did not evaluate an aLDFA of 99°, which was within the range of values reported in grade 3 MPL cases in small-breeds, which was 100.53° ± 2.05°.10 Although we can deduce the trend in patellar luxation as the varus angle changes, this point is a limitation of our study and needs to be carefully interpreted.
There are notable limitations of the present study. The FE model was not validated due to the lack of a comprehensive ex vivo testing model or motion analysis method to replicate the entire 3D biomechanics of the canine patellofemoral joint. Additionally, the FE model was derived from an individual Beagle and did not include other factors affecting the MPL, such as cartilage, soft tissue stabilizers such as the components of the lateral aspect (vastus lateralis and lateral patellofemoral ligament), shapes of the femoral trochlea, and different positions of the tibial tuberosity. Thus, the results may not be generalizable to all dogs, and the calculated luxation risk may have been overestimated. Therefore, the aLDFA value in MPL obtained here should be considered as an indication of a trend caused by bony deformities and not as an absolute criterion. However, it is interesting that the results of this study are in good agreement with clinical recommendations for DFO in large-breed dogs with MPL [9, 13, 14, 29].