BACKGROUND: Thiazolidinedione (TZD) treatment has been associated with fractures. The purpose of this study was to examine the association between TZD treatment and fractures in type 2 diabetic patients. METHODS: Using data from Translating Research into Action for Diabetes, a multicenter prospective observational study of diabetes care in managed care, we conducted a matched case-control study to assess the odds of TZD exposure in patients with type 2 diabetes with and without fractures. We identified 786 cases based on fractures detected in health plan administrative data. Up to four controls without any fracture diagnoses were matched to each case. Controls were matched on health plan, date of birth within 5 yr, sex, race/ethnicity, and body mass index within 5 kg/m(2). We performed conditional logistic regression for premenopausal and postmenopausal women and men to assess the odds of exposure to potential risk factors for fracture, including medications, self-reported limited mobility, and lower-extremity amputations. RESULTS: We found statistically significant increased odds of exposure to TZDs, glucocorticoids, loop diuretics, and self-reported limited mobility for women 50 yr of age and older with fractures. Exposure to both loop diuretics and TZDs, glucocorticoids, and insulin and limited mobility and lower-extremity amputation were associated with fractures in men. CONCLUSION: Postmenopausal women taking TZDs and the subset of men taking both loop diuretics and TZDs were at increased risk for fractures. In postmenopausal women, risk was associated with higher TZD dose. No difference between rosiglitazone and pioglitazone was apparent.