Objective: To assess the effect of lateral wedge insoles compared with flat control insoles on improving symptoms and slowing structural disease progression in medial knee osteoarthritis. Design: Randomised controlled trial. Setting: Community in Melbourne, Australia. Participants: 200 people aged 50 or more with clinical and radiographic diagnosis of mild to moderately severe medial knee osteoarthritis. Interventions: Full length 5 degree lateral wedged insoles or flat control insoles worn inside the shoes daily for 12 months. Main outcome measures: Primary symptomatic outcome was change in overall knee pain (past week) measured on an 11 point numerical rating scale. Primary structural outcome was change in volume of medial tibial cartilage from magnetic resonance imaging scans. Secondary clinical outcomes included changes in measures of pain, function, stiffness, and health related quality of life. Secondary structural outcomes included progression of medial cartilage defects and bone marrow lesions. Results: Between group differences did not differ significantly for the primary outcomes of change in overall pain (-0.3 points, 95% confidence intervals -1.0 to 0.3) and change in medial tibial cartilage volume (-0.4 mm3, 95% confidence interval -15.4 to 14.6), and confidence intervals did not include minimal clinically important differences. None of the changes in secondary outcomes showed differences between groups. Conclusion: Lateral wedge insoles worn for 12 months provided no symptomatic or structural benefits compared with flat control insoles. Trial registration: Australian New Zealand Clinical Trials Registry ACTR12605000503628 and ClinicalTrials.gov NCT00415259.