Gonadotropin-releasing hormone (GnRH) agonists decrease bone mineral density, a surrogate for fracture risk, in men with prostate cancer. We conducted a claims-based cohort study to characterize the relationship between GnRH agonists and risk for clinical fractures in men with nonmetastatic prostate cancer. Using medical claims data from a 5% national random sample of Medicare beneficiaries, we identified a study group of men with nonmetastatic prostate cancer who initiated GnRH agonist treatment from 1992 to 1994 (n = 3,887). A comparison group of men with nonmetastatic prostate cancer who did not receive GnRH agonist treatment during the study period (n = 7,774) was matched for age, race, geographic location, and comorbidity. Clinical fractures were identified using inpatient, outpatient, and physician claims during 7 years of follow-up. In men with nonmetastatic prostate cancer, GnRH agonists significantly increased fracture risk. The rate of any clinical fracture was 7.88 per 100 person-years at risk in men receiving a GnRH agonist compared with 6.51 per 100 person-years in matched controls (relative risk, 1.21; 95% CI, 1.14 to 1.29; P < .001). Rates of vertebral fractures (relative risk, 1.45; 95% CI, 1.19 to 1.75; P < .001) and hip/femur fractures (relative risk, 1.30; 95% CI, 1.10 to 1.53; P = .002) were also significantly higher in men who received a GnRH agonist. GnRH agonist treatment independently predicted fracture risk in multivariate analyses. Longer duration of treatment conferred greater fracture risk. GnRH agonists significantly increase risk for any clinical fracture, hip fractures, and vertebral fractures in men with prostate cancer.