Objective: To compare biomechanical and clinical outcome of laterally wedged insoles (LWI) and an ankle-foot orthosis (AFO) in patients with medial knee osteoarthritis. Design: Single-centre, block-randomized, cross-over controlled trial. Setting: Outpatient clinic. Subjects: About 39 patients with symptomatic medial knee osteoarthritis. Interventions: Patients started with either LWI or AFO, determined randomly, and six weeks later changed to the alternative. Main measures: Change in the 1st maximum of external knee adduction moment (eKAM) was assessed with gait analysis. Additional outcomes were other kinetic and kinematic changes and the patient-reported outcomes EQ-5D-5L, Oxford Knee Score (OKS), American Knee Society Clinical Rating System (AKSS), Hannover Functional Ability Questionnaire – Osteoarthritis and knee pain. Results: Mean age (SD) of the study population was 58 (8) years, mean BMI 30 (5). Both aids significantly improved OKS (LWI P = 0.003, AFO P = 0.001), AKSS Knee Score (LWI P = 0.01, AFO P = 0.004) and EQ-5D-5L Index (LWI P = 0.001, AFO P = 0.002). AFO reduced the 1st maximum of eKAM by 18% ( P < 0.001). The LWI reduced both maxima by 6% ( P = 0.02, P = 0.03). Both AFO and LWI reduced the knee adduction angular impulse (KAAI) by 11% ( P < 0.001) and 5% ( P = 0.05) respectively. The eKAM (1st maximum) and KAAI reduction was significantly larger with AFO than with LWI ( P = 0.001, P = 0.004). Conclusions: AFO reduces medial knee load more than LWI. Nevertheless, no clinical superiority of either of the two aids could be shown.