BACKGROUND: Appropriate footwear is recommended for self-management of knee osteoarthritis. Shoes that reduce harmful knee loads are available, but symptomatic effects are uncertain. OBJECTIVE: To evaluate the efficacy of unloading shoes in alleviating knee osteoarthritis symptoms. DESIGN: Participant- and assessor-blinded comparative effectiveness randomized, controlled trial. (Australian New Zealand Clinical Trials Registry: ACTRN12613000851763). SETTING: Community. PARTICIPANTS: 164 persons with medial knee osteoarthritis. INTERVENTION: Walking shoes with triple-density, variable-stiffness midsoles and mild lateral-wedge insoles designed to unload the medial knee and worn daily (intervention) versus conventional walking shoes (comparator). MEASUREMENTS: Primary outcomes were pain with walking (assessed on a numerical rating scale [NRS]) and physical function (Western Ontario and McMaster Universities Osteoarthritis Index [WOMAC]) at 6 months. Secondary outcomes were knee pain and stiffness (WOMAC), average pain (NRS), intermittent and constant knee pain (Intermittent and Constant Osteoarthritis Pain questionnaire), quality of life (Assessment of Quality of Life instrument), physical activity (Physical Activity Scale for the Elderly), and global change in pain and function (Likert scales). RESULTS: A total of 160 participants (98%) completed primary outcome measures at 6 months. Changes in pain (mean difference, 0.0 units [95% CI, -0.9 to 0.8 unit]) and function (mean difference, 0.3 unit [CI, -3.2 to 3.7 units]) did not differ between groups at 6 months, with both groups showing clinically relevant improvements in function and the intervention group showing clinically relevant improvements in pain. There were no differences in secondary outcomes. Pain was globally improved in 54% of participants, and function was globally improved in 44% to 48%. Unloading shoes were not associated with increased probability of improvement (odds ratios, 0.99 [CI, 0.53 to 1.86] for pain and 0.85 [CI, 0.45 to 1.61] for function). LIMITATION: Effects on joint structure were not evaluated. CONCLUSION: Shoes with modified midsoles to unload the medial knee conferred no additional benefit over conventional walking shoes. Both improved pain and function by clinically relevant amounts. PRIMARY FUNDING SOURCE: Australian National Health and Medical Research Council.