Background: We recently reported that exercise reduced AI-associated arthralgias. This analysis examined potential mechanisms mediating the effect on exercise on arthralgia, specifically weight, body fat and serum C-reactive protein (CRP), IL-6, and TNF-α. Methods: Women taking an AI for breast cancer and experiencing arthralgia were randomized to exercise (150 min/wk of moderate-intensity aerobic exercise and twice-weekly supervised resistance exercise) or control. Weight, dual energy X-ray absorptiometry scans, and a fasting blood sample were collected at baseline and 12-months. We used analysis of covariance, adjusted for baseline values, to evaluate mean weight, fat and inflammatory changes across groups, and used Baron and Kenny approach for mediation analysis. We evaluated moderation effects by including an interaction in the analysis. Results: We randomized 61 to exercise and 60 to control. Exercisers participated in, an average, 70% of resistance training sessions and 119 min/wk of aerobic exercise. Weight, body fat and CRP decreased at 12 months among exercisers (Table 1). Changes in these variables were not correlated with changes in arthralgia, and the interactions were not significant. After adjusting for weight, body fat and CRP, exercise decreased arthralgia by 35% vs. no change among control, p = .0059. Conclusions: Exercise decreased weight, body fat and CRP in breast cancer survivors experiencing AI-associated arthralgia. However, these changes did not mediate the effect of exercise on arthralgia. Exercise exerts beneficial effects on arthralgia via mechanisms not related to weight or systemic inflammation.