To determine whether thigh muscle strength predicts the risk of undergoing knee replacement surgery, independent of radiographic severity of osteoarthritis and pain.Participants in the Osteoarthritis Initiative study who had received a knee replacement between the 12-month and 60-month follow-up visits (cases) were each matched with 1 control subject (who did not receive a knee replacement over 60 months) for age, sex, height, body mass index, baseline radiographic stage, and location of joint space narrowing. Isometric knee extensor and flexor strength was recorded biennially. The strength examination prior to knee replacement (≤2 years) was termed time 0, that 2 years prior to time 0 was termed time -2, and that 4 years prior to time 0 was termed time -4. Muscle strength in patients and controls was compared using paired t-tests and conditional logistic regression analysis adjusted for pain.One hundred thirty-six participants (60% women, mean ± SD age 65 ± 9 years, body mass index 29 ± 4 kg/m(2) ) received a knee replacement during follow-up, had strength data for at least time 0, and had a matched control. Knee extensor strength at time 0 (primary outcome measure) was significantly lower in female patients compared with controls (P < 0.001, pain-adjusted odds ratio [ORadj ] 1.72, 95% confidence interval 1.16-2.56), but no significant difference was observed in men. Results were similar for knee flexor strength at time 0 and for longitudinal change in extensor and flexor strength between time -2 and time 0. Thigh muscle strength at time -2 or time -4, or change between time -4 and time -2, did not predict the risk of undergoing knee replacement surgery in men or women.Thigh muscle strength predicted the risk of undergoing knee replacement surgery in women, but not in men. The results of this study may identify a window during which the risk of undergoing knee replacement surgery in women can be modified.