Background: Evidence has emerged that pioglitazone may increase the risk of bladder cancer, but the association has not been confirmed. This potential risk also has not been evaluated in users of rosiglitazone. Objective: Using Taiwan's National Health Insurance Research Database (NHIRD), this large population-based nested case-control study was conducted to explore the relationship between the use of rosiglitazone or pioglitazone and risk of bladder cancer in diabetic patients. Methods: We identified 3,412 cases of newly diagnosed bladder cancer and 17,060 controls (1:5 matched by age and sex) among a diabetic patient cohort from the NHIRD. We defined an index date for each case as the date of first hospitalization for bladder cancer. Each control was assigned the index date of their corresponding case. Multivariable conditional logistic regressions were used to estimate the association between exposure (timing and duration) to rosiglitazone or pioglitazone and bladder cancer. We defined rosiglitazone or pioglitazone exposure as "current" if the prescription duration covered the index date or ended at 90 days before, as "recent" if it ended 91-180 days before the index date, or as "past" if the last prescription ended more than 180 days before. Duration of rosiglitazone or pioglitazone use was defined based on the cumulative days of exposure prior to the index date: <1, 1-2 and ≥2 years. Results: Rosiglitazone and pioglitazone use were associated with risk of bladder cancer and the associations were stronger with a longer term of exposure (pioglitazone <1 year odds ratio [OR] 1.45 [95 % CI 1.12-1.87, p < 0.01], 1-2 years OR 1.74 [95 % CI 1.05-2.90, p = 0.03] and ≥2 years OR 2.93 [95 % CI 1.59-5.38, p < 0.01]; rosiglitazone <1 year OR 0.98 [95 % CI 0.82-1.17, p = 0.81], 1-2 years OR 1.78 [95 % CI 1.31-2.39, p < 0.01] and ≥2 years OR 2.00 [95 % CI 1.37-2.92, p < 0.01]). Conclusions: Long-term exposures to pioglitazone and rosiglitazone were associated with higher odds of bladder cancer, and the highest odds were seen in users with ≥2 years of exposure.