OBJECTIVE: Obese patients with knee osteoarthritis (OA) are encouraged to lose weight to obtain symptomatic relief. Risk of vascular events is higher in people with OA compared to people without arthritis. Our aim in this randomized trial was to compare changes in cardiovascular disease (CVD) risk-factors, nutritional health, and body composition after 1-year weight-loss maintenance achieved by [D]diet, [E]knee-exercise, or [C]control, following weight loss by low-energy-diet. DESIGN AND METHODS: Obese individuals (n = 192, >50 years) with knee OA, 63 years (SD 6), weight 103.2 kg (15.0), body-mass index 37.3 kg/m(2) (4.8), were enrolled into a 68-week weight-loss trial. RESULTS: Mean changes in weight, in D, E, and C were -11.0, -6.3, and -8.3 kg (P = 0.002). Reduction in waist circumference in D, E, and C were -8.4, -4.6, and -7.0 cm (P = 0.007). D reduced waist circumference significantly more than E: -3.8 cm (95%CI -6.2 to -1.4; P = 0.0024). There was no difference between the groups in changes in CVD risk factors; blood pressure, triglycerides, and cholesterol. Nutritional health was improved in all groups. For markers of bone, no statistical difference was found between the groups. CONCLUSIONS: Dietary support, or control, maintained improvements in cardiovascular risk factors to the same extent and none of the interventions had a detrimental effect on bone.