OSTEOARTHRITIS IS A MULTI-JOINT DISEASE. OR IS IT?

INTRODUCTION

OA frequently affects both the hip and knee joints; however, most prognostic studies evaluate these joints in isolation. Given the biomechanical and systemic connections between them, this joint-specific focus may overlook important patterns of disease progression. A better understanding of combined hip and knee OA progression could support the development of more accurate prediction models and treatment strategies.

OBJECTIVE

To determine whether combined hip and knee OA progression exhibits a distinct phenotype compared to isolated OA progression.

METHODS

This study used the OAI data, which comprises data from participants aged 45–79 years at risk of developing knee OA. The dataset features bilateral posteroanterior knee radiographs and standardized weight-bearing anteroposterior pelvic radiographs obtained at the baseline and 48-month follow-up visits. Minimal JSW (mJSW) was measured manually for knees and automatically for hips. OA progression was defined as JSN of ≥0.5 mm in either hip or ≥0.7 mm in either knee. Participants with no JSN in any of the four joints at the 48-month follow-up were excluded. The selected participants were classified as having isolated (either hip or knee) or combined (both hip and knee) progression. A logistic regression model incorporating clinical and structural baseline features was used to explore associations with combined progression compared to isolated progression. Baseline radiographic OA (ROA) status of each of the four joints was classified as no ROA (0), early ROA (1) and definite ROA (2) based on the KLG and modified Croft grades. Adjusted odds ratios (aOR) and their 95% confidence intervals, estimated using bootstrapping with 10,000 iterations, and the goodness-of-fit of the model were assessed.

RESULTS

Among the 1,190 included participants with any ROA progression (mean age 61.5 ± 8.8 years; BMI 29.1 ± 4.5; 55.1% female), 281 (23.6%) showed combined ROA progression. The co-occurrence of baseline hip and knee ROA grades was reviewed descriptively (Table 1). The observed relatively small prevalence of combined progression in this population prevented including these ROA status interaction effects within our model. The logistic model showed improved fit over an intercept-only model (likelihood ratio test, p < 0.0001), and acceptable goodness-of-fit (Hosmer-Lemeshow test, p = 0.40). Several baseline features were associated with higher odds of combined progression compared to isolated, including older age, female sex, varus knee alignment in the right knee, higher mJSW in the hip, and having definite ROA in the left hip (Figure 1). Interestingly, having ROA in the right hip or valgus knee alignment in the left knee would decrease the odds of combined progression.

CONCLUSION

Our findings suggest that combined hip and knee OA progression may represent a distinct clinical phenotype with identifiable characteristics. While the model showed statistically significant associations and demonstrated adequate fit, these results should be interpreted cautiously. Larger and more diverse datasets are needed to further validate these findings and investigate the heterogeneity of multi-joint OA progression.