A meta-analysis of sex differences prevalence, incidence and severity of osteoarthritis

Abstract

OBJECTIVE To resolve uncertainty regarding sex differences in osteoarthritis (OA) by performing a meta-analysis of sex differences in OA prevalence, incidence and severity. METHODS Standard search strategies for population-based studies of OA providing sex-specific data. Random effects meta-analysis to provide pooled male vs female risk and rate ratios for prevalent and incident OA, and standardized mean differences (SMD) for OA severity. Meta-regression was used to investigate sources of heterogeneity. RESULTS Males had a significantly reduced risk for prevalent OA in the knee [Risk Ratio (RR) 0.63, 95% CI 0.53-0.75] and hand [RR 0.81, 95% CI 0.73-0.90] but not for other sites. Males aged <55 years had a greater risk of prevalent cervical spine OA [RR 1.29, 95% CI 1.18-1.41]. Males also had significantly reduced rates of incident OA in the knee [Incidence Rate Ratio (IRR) 0.55, 95% CI 0.32-0.94] and hip [IRR 0.64, 95% CI 0.48-0.86], with a trend for hand [IRR 0.65, 95% confidence interval (CI) 0.35-1.20]. Females, particularly those > or = 55 years, tended to have more severe OA in the knee but not other sites. Heterogeneity in the estimates of sex differences in prevalence was substantially explained by age and other study design factors including method of OA definition. CONCLUSIONS The results demonstrate the presence of sex differences in OA prevalence and incidence, with females generally at a higher risk. Females also tend to have more severe knee OA, particularly after menopausal age. The site differences indicate the need for further studies to explore mechanisms underlying OA.