Background/Purpose:  Increasing evidence suggests a role for uric acid as a danger signal for the innate immune system, which could contribute to OA development and/or progression. If serum urate (SUA) is a true risk factor for OA, urate-lowering drugs could help prevent incident or progressive OA. However, SUA and knee OA are both associated with obesity; thus any association may be confounded by body mass index (BMI). Hand OA, in contrast, may be less affected by BMI, and may provide a better reflection of the systemic metabolic effects of urate. We therefore evaluated the relation of SUA to radiographic knee and hand OA (ROA) in a large community cohort unselected for OA, accounting for BMI.

Methods:  We performed a cross-sectional study among subjects of the Framingham Original Cohort 50 years and older. SUA was assessed at exams 20 and 21 (1986-1992) and ROA at exam 22 (1990-1994). ROA of each tibiofemoral joint and of each hand joint was defined as KL ≥2. We categorized SUA as: <5mg/dL, 5-<6mg/dL, 6-<7mg/dL, 7-<9mg/dL, and ≥9mg/dL. We evaluated the sex-specific relation of SUA to the prevalence of ROA using logistic regression, adjusted for age and BMI, with generalized estimating equations to account for the correlations of joints within individuals. We repeated these analyses with SUA as a continuous variable, and categorized as quintiles.

Results:  There were 329 men and 575 women with SUA and radiographic data available, among whom 89 and 156 had knee OA, and 252 and 500 had hand OA, respectively. Men with SUA >9 mg/dL had 3.3 times significantly higher prevalence of knee ROA compared with those with SUA <5 mg/dL, though the number of individuals in this category was small (table). The other SUA levels were not associated with knee or hand ROA in men or women (table), nor was there a significant test for linear trend. When SUA was assessed as a continuous variable, the prevalence of knee ROA among men and women was 1.1 times higher (95% CI 0.88-1.25 and 0.93-1.31 for men and women, respectively) for each mg/dL increment of SUA, and for hand ROA it was 1.0 (95% CI 0.94-1.10 and 0.95-1.10, respectively) (table). There was also no relation of SUA quintiles with knee or hand ROA (data not shown).  

Conclusion:  There was no dose-response relationship between SUA and ROA of the knee or hand. While men in the highest SUA level (>9 mg/dL) had significantly higher prevalence of knee ROA, given the small numbers and lack of replication in women or in hand ROA, this could be a chance finding. On the other hand, this sample had very few individuals with very high SUA levels; thus, we cannot rule out a possible threshold effect.  Repeating this study in a sample with greater numbers of individuals with higher levels of SUA is warranted. Nonetheless, it is possible that systemic urate is not wholly biologically relevant for the risk of OA, and rather urate in the joint micro-environment must be evaluated to gain insights into the role urate may play on OA pathogenesis.