OBJECTIVE: To assess whether knee extensor strength or hamstring:quadriceps (H:Q) ratio predicts risk for incident radiographic tibiofemoral and incident symptomatic whole knee osteoarthritis (OA) in adults ages 50-79 years. METHODS: We followed 1,617 participants (2,519 knees) who, at the baseline visit of the Multicenter Osteoarthritis (MOST) Study, did not have radiographic tibiofemoral OA and 2,078 participants (3,392 knees) who did not have symptomatic whole knee OA (i.e., did not have the combination of radiographic OA and frequent knee symptoms). Isokinetic strength was measured at baseline, and participants were followed for development of incident radiographic tibiofemoral OA, or incident symptomatic whole knee OA at 30 months. Generalized estimating equations accounted for 2 knees per subject, and multivariable models adjusted for age, body mass index (BMI), hip bone mineral density, knee surgery or pain, and physical activity score. RESULTS: In the studies of incident radiographic and incident symptomatic knee OA, mean +/- SD ages were 62.4 +/- 8.0 years and 62.3 +/- 8.0 years, respectively, and mean +/- SD BMI scores were 30.6 +/- 5.8 kg/m(2) and 30.2 +/- 5.5 kg/m(2), respectively. Knee extensor strength and H:Q ratio at baseline significantly differed between men and women. Neither knee extensor strength nor the H:Q ratio was predictive of incident radiographic tibiofemoral OA. Compared with the lowest tertile, the highest tertile of knee extensor strength protected against development of incident symptomatic whole knee OA in both sexes (adjusted odds ratio 0.5-0.6). H:Q ratio was not predictive of incident symptomatic whole knee OA in either sex. CONCLUSION: Thigh muscle strength does not appear to predict incident radiographic OA, but does seem to predict incident symptomatic knee OA.