An exploration of a varus malaligned phenotype in knee osteoarthritis

Purpose: Multiple phenotypes characterised by different disease mechanisms have been hypothesised to explain the large variability in the knee osteoarthritis (KOA) population. Repeated evidence showed that varus malalignment alone is sufficient to determine disease progression in absence of other known risk factors. Therefore, the purpose of this study is: (1) to estimate the knee compression force (CF) during the stance phase of KOA patients with a varus malaligned phenotype (VMP) characterised by exclusive degeneration of the medial compartment; (2) to compare the estimated CF and other MRI biomarkers (i.e. bone marrow lesions [BML], meniscal damage) of this VMP with a group of controls (C), a group of KOA subjects with normal alignment (NA) and a group of KOA subjects with varus malalignment and cartilage degeneration that extends to the lateral compartment (VA). Methods: A secondary data analysis was conducted on a sample of 39 KOA subjects and 18 controls recruited for a previous study. The KOA subjects were classified using MRI cartilage biomarkers assessed with the Boston Leeds Osteoarthritis Knee Score (BLOKS) and the hip-kneeankle (HKA) angle estimated from the gait analysis data. The patients were classified in the different groups according to the following criteria: VMP (12): varus alignment (≥2°) and cartilage degeneration exclusive in the medial compartment (BLOKS ≥2 in the tibial or femoral medial compartment and BLOKS ≤1 in both femur and tibial lateral compartments). VA (17): varus alignment (≥27°) and cartilage degeneration that does not meet the VMP criteria. NA (10): normal alignment (≥2°< x < 2°). The medial and lateral CF corrected for body weight were estimated using an inverse dynamics model (AnyBody Modeling System, AnyBody Technology) . To examine between-group differences in the estimated knee CF, one-way analysis of covariance (ANCOVA) was used with walking speed included as a covariate. Sidak post hoc test and bootstrapping were performed to obtain robust p values. Differences in the presence of BML and meniscal damage were assessed using a chisquare exact test with post hoc correction for multiple comparisons. The relationship between alignment and medial CF was further analysed with a regression model introducing group membership as a moderator. Results: No differences were identified in age and disease duration between the groups. The impulse of the medial CF (overall effect of the CF over the stance time) was significantly higher (p < 0.01) in the VMP compared to all the other groups. The peak of the medial CF was higher in the VMP compared to the VA group (p < 0.05), and no statistical differences were found between the VMP and the C and NA groups. No differences were found in the impulse of the lateral compartment CF. The VMP peak of the lateral CF was the lowest among the analysed groups with a p < 0.01 when compared to the C and NA groups. Subjects in the VMP group showed a higher prevalence of meniscal maceration in the medial compartment compared to all the other groups (VMP: 92%, VA: 28%, NA 10%, C: 6%; p < 0.05); and a higher prevalence of large BML [BLOKS 2-3] in the tibia medial compartment (VMP: 83%, VA: 29%, NA 0%, C: 6%) and in the femur medial compartment (VMP: 58%, VA: 18%, NA 10%, C: 6%) compared to all the other groups (p < 0.05). No differences were identified for these features in the lateral compartment. The results of the regression analysis showed a significant moderation by group membership on the relationship between alignment and impulse of the medial CF (p < 0.01). When subjects are classified in the VMP, there is a significant relationship between alignment and impulse of the medial CF(p < 0.01). This relationship disappears in the other groups. Conclusions: The VMP showed higher medial joint force and lower lateral joint force compared to the other groups; in particular with the VA group despite no difference in alignment. The higher prevalence of meniscal maceration and large BML in the medial compartment of the VMP seems to confirm the hypothesis of a distinct biomechanical phenotype characterised by increased joint load. The alignment was related to the impulse of the medial CF only in the VMP. This suggests that in this phenotype, the malalignment and the increased force may be the main determinant of the knee disease while, in the VA group, other factors may be involved in the disease process and be responsible for the progression of the disease in the lateral compartment. Therefore, treatments aiming to reduce the knee force may see improved efficacy if tested in this phenotype. (Figure Presented).