Abstract
OBJECTIVE: The present study was undertaken to assess whether the odds for incident radiographic osteoarthritis (OA) differ between men and women in regard to body mass index (BMI) and inflammatory magnetic resonance imaging (MRI) markers 1 and 2 years prior, and whether the presence of inflammation on MRI differs between normal-weight and overweight/obese individuals who develop radiographic OA up to 4 years prior. METHODS: We studied 355 knees from the Osteoarthritis Initiative study that developed incident radiographic OA and 355 matched controls. MRIs were read for effusion-synovitis and Hoffa-synovitis for up to 4 consecutive annual time points. Subjects were classified as normal-weight (BMI <25), overweight (BMI ≥25 and <30), or obese (BMI ≥30). Conditional logistic regression was used to assess odds of incident radiographic OA for effusion-synovitis and Hoffa-synovitis at 1 and 2 years prior to radiographic OA incidence (i.e., "P-1" and "P-2") considering BMI category. Bivariate logistic regression was used to assess odds of inflammation for cases only. RESULTS: One hundred seventy-eight (25.1%) participants were normal weight, 283 (39.9%) overweight, and 249 (35.1%) obese. At P-2, being overweight with Hoffa-synovitis, which had an odds ratio [OR] of 3.26 (95% confidence interval [95% CI] 1.39-7.65), or effusion-synovitis (OR 3.56 [95% CI 1.45-8.75]) was associated with greater odds of incident radiographic OA in women. For those with incident radiographic OA, there were no increased odds of synovitis in the overweight/obese subgroup for most time points, but increased odds for effusion-synovitis were observed at P-2 (OR 2.21 [95% CI 1.11-4.43]). CONCLUSION: Presence of inflammatory markers seems to play a role especially in overweight women, while obese women have increased odds for radiographic OA also in the absence of these markers.