OBJECTIVE:To determine the effect of weight loss and exercise interventions on serum leptin and to investigate the relationship of physical function and osteoarthritis (OA) severity with serum leptin in older overweight and obese adults with knee OA. In addition, the study examined if serum leptin predicts weight loss. DESIGN: Longitudinal, controlled clinical trial of weight loss and exercise interventions. SUBJECTS: Community dwelling, older, overweight and obese adults (n ¼ 316; 460 years of age; body mass index Z28.0 kg m À2 ) with symptomatic knee OA and self-reported difficulty in performing selected physical activities were recruited. INTERVENTIONS: Participants were randomized into one of four groups for the 18-month study duration: Healthy Lifestyle Controls, Dietary Weight Loss (Diet), Exercise Training (Exercise), and a combination of Dietary Weight Loss and Exercise Training (Diet þ Exercise). The weight loss goal for the two Diet groups was 5% from baseline at 18 months. Participants in the Exercise groups were trained for 3 days week À1 , 60 min day À1 . MEASUREMENTS: Body weight, body mass index, serum leptin, physical function, and OA severity were measured at baseline, 6 months, and 18 months. RESULTS: Diet and Diet þ Exercise groups lost 5.3 and 6.1% of their weight, respectively, at 18 months with the Exercise group losing 2.9%. There was a significant main effect of weight loss on serum leptin with a decrease in serum leptin averaged across the 6-and 18-month time points for the Diet and Diet þ Exercise groups compared to the other two groups (b ¼ 0.245; Po0.01). No main effect for exercise training was observed. Serum leptin was related to self-reported physical function. In all participants, a mixed model analysis demonstrated that lower levels of baseline serum leptin predict larger weight loss (b ¼ À2.779; P ¼ 0.048). CONCLUSION: Decreases in serum leptin may be one mechanism by which weight loss improves physical function and symptoms in OA patients.