Association Between Race and Radiographic, Symptomatic, and Clinical Hand Osteoarthritis: A Propensity Score–Matched Study Using Osteoarthritis Initiative Data

Objective: To determine the associations between Black race and the presence of radiographic, symptomatic, and clinical hand osteoarthritis (OA). Methods: Using available hand radiographs from the Osteoarthritis Initiative cohort (total 4,699; n = 849 Black subjects [18.1%], n = 3,850 non-Black subjects [81.9%]), a propensity score–matching method was used to match Black subjects with non-Black subjects for known potential risk factors of hand OA (age, sex, body mass index, smoking status, cardiovascular disease, osteoporosis, excessive occupation- or recreation-related hand use, and knee OA). Posteroanterior radiographs of subjects’ dominant hands were reviewed by a musculoskeletal radiologist in a blinded manner. To assess the severity of hand OA, the modified Kellgren/Lawrence (K/L) radiographic OA scoring scale (grades 0–4) was used, and the presence of erosive OA in the hand joints was recorded. Associations between race and the severity of hand OA (measured as the summed modified K/L grade), presence of radiographic hand OA (modified K/L grade ≥2), presence of erosive hand OA, presence of symptomatic hand OA (radiographic OA with hand pain), and presence of clinical hand OA (indicated by clinical findings of Heberden's nodes in the hands) were studied using regression models. In these models, beta coefficients or odds ratios (ORs) with 95% confidence intervals (95% CIs) were calculated for the associations between Black race and any of these radiographic and symptomatic hand OA phenotypes. Results: Black subjects had less severe hand OA (β = –1.93 [95% CI –2.53, –1.34]), as well as a lower risk of developing radiographic hand OA (OR 0.79 [95% CI 0.66, 0.94]), erosive hand OA (OR 0.23 [95% CI 0.11, 0.47]), symptomatic hand OA (OR 0.63 [95% CI 0.49, 0.82]), and clinical hand OA (OR 0.49 [95% CI 0.41, 0.60]), as compared to non-Black subjects. Conclusion: In contrast to the well-established association between Black race and knee or hip OA, the findings of this study suggest that the risk of hand OA is lower in Black subjects compared to non-Black subjects, which is not mediated by known hand OA risk factors. Future studies are warranted to determine the mediating protective factors for hand OA among Black subjects.