Do associations with hand OA vary by knee osteoarthritis phenotype? Cross-sectional data from the Multicenter Osteoarthritis Study

Objective: Osteoarthritis (OA) is highly heterogeneous and has both biomechanical and systemic components that may not have the same etiology. We therefore aimed to identify specific knee OA phenotypes that may be more strongly associated with hand OA to refine the criteria used to define multi-joint OA. Design:We assessed data from the Multicenter Osteoarthritis Study (MOST). We ascertained hand OA from bilateral hand photographs; scores for each joint row were summed to yield an aggregate hand OA score. Knee OA was ascertained from bilateral posteroanterior knee radiographs read for Kellgren-Lawrence grade and individual radiographic features. We tested associations between hand and knee OA with phenotypes including symptomatic OA, hyper- and atrophic knee OA, and one excluding post-traumatic OA. Associations between hand and knee OA were assessed with logistic regression, adjusted for age. Results: We studied 2493 participants with hand and knee OA measures. Median age was 63 years with 57% women. 55% had an aggregate hand OA score >= 2; frequency of knee OA phenotypes ranged from 8% to 34%. The age-adjusted odds ratio (OR) was 1.14 (95% confidence interval (CI) = 1.04-1.26) for knee OA per standard deviation of the hand OA aggregate score. Hand OA associations with symptomatic knee OA and knee OA excluding post-traumatic knee OA were OR = 1.16 (95% CI = 1.03-1.31) and OR = 1.21 (95% CI = 1.08-1.35), respectively. No other knee OA phenotype reached statistical significance. Conclusions: Age-adjusted associations between hand and knee OA were modest and were largely similar across knee OA phenotypes.