PURPOSE: Previous studies have reported an increased risk of bladder cancer (BC) in IC/BPS patients. In this study, we re-examined the BC risk in a population based IC/BPS cohort to assess the potential detection bias caused by some IC/BPS patients diagnosed who might already have co-existing BC. METHODS: We performed a retrospective cohort study based on a Research Database by extracting IC/BPS patients diagnosed within years 2002-2013. The patients in the study cohorts were identified based on at least 2 IC/BPS diagnoses and excluded patients with BC occurred before IC/BPS diagnosis. The primary outcome was BC events detected. Propensity scores (PSs) were calculated for matching IC/BPS cohort with non-IC/BPS cohort on a 1:1 basis. Cox proportional hazard regression analysis was then used to compare hazard ratios of BC development between 2 cohorts. RESULTS: By excluding patients with BC diagnosed within 1 year after IC/BPS diagnosis, the study cohort was insignificantly different from the PS-matched control (Model 1, p = 0.219) but significantly different from the non-PS-matched control (Model 2, p <
0.001). However, when including patients with BC diagnosed within 1 year after IC/BPS diagnosis, the study cohort was significantly different from both PS-matched (Model 3, p = 0.002) and non-PS-matched (Model 4, p <
0.001) controls, indicating that excluding patients with BC diagnosed within 1 year after IC/BPS diagnosis and adopting PS matching method greatly reduce the BC detection bias. CONCLUSIONS: IC/BPS is not associated with BC. The detection bias of previous studies may result from inadequate recruitments of study cohorts or improper matching of control cohorts.