Objective: To compare structural knee joint changes in obese patients with knee osteoarthritis (OA) that after an intensive weight loss therapy were randomized to continuous dietetic support, a specialized knee exercise program, or 'no attention' for 1 year. Methods: 192 obese individuals with knee OA underwent an intensive 16-week weight loss program with subsequent randomization to one of the three treatment groups. Changes in cartilage loss, bone marrow lesions (BMLs), synovitis, and effusion were assessed using semi quantitative assessments of magnetic resonance imaging (MRI) obtained at weeks 0 and 68 applying the BLOKS score. Results: During the 52 weeks maintenance period the continuous dietary maintenance group support on average gained 1.1kg (95% CI:-0.3:2.5) body mass, the exercise group gained 6.6kg (95% CI 5.4:7.8) and the no-attention group gained 4.8kg (95% CI: 2.9:6.7). There were no statistically significant between-group differences in changes in cartilage loss, synovitis or effusion at the follow-up (analysis of covariance; ANCOVA, P>0.16), while there was an increased number of medial tibiofemoral BMLs in the exercise group (ANCOVA, P=0.015) compared to both diet (difference:-0.21 [95%CI-0.40:-0.03]) and "no attention" (difference:-0.26 [95%CI-0.44:-0.07]) groups. Conclusion: In this 1 year follow-up after weight-loss in obese knee OA patients, we found a potentially increased number of BMLs in the exercise group compared to the diet and no attention groups, with no between-group differences in changes in cartilage loss, synovitis or effusion. These findings should be interpreted with caution for exercise compliance, MRI methodology and follow-up time. (ClinicalTrials.gov identifier: NCT00655941).