OBJECTIVE: . An abnormally high knee adduction moment increases the medial tibiofemoral compartment load at the knee during gait, and is an important biomechanical marker of joint pathology. This cross-sectional study examines the relationship between the knee adduction moment and knee pain in middle-aged women without radiographic knee osteoarthritis (OA). METHODS: Three-dimensional Vicon gait analyses were performed on 20 women who had knee pain but no radiological evidence of joint pathology. RESULTS: In multivariate analysis, the peak knee adduction moment during the late stance phase of gait was inversely associated with knee pain [beta: -10.1 (95% CI -17.6, -2.7), p = 0.01] after adjustment for body mass index (BMI) and age. This explained that the knee adduction moment during late stance contributed 32% of the variance in knee pain. The peak knee adduction moment during early stance was not significantly associated with knee pain prior to and after adjustment for BMI and age. CONCLUSION: There is a significant inverse association between the peak knee adduction moment during late stance and the amount of knee pain experienced by women without radiographic evidence of joint pathology. This may represent a compensatory mechanism to reduce medial tibiofemoral joint load in the setting of knee pain.