Urinary tract infection-like symptom is associated with worse bladder cancer outcomes in the Medicare population: Implications for sex disparities

Objectives: To determine the time to bladder cancer diagnosis from initial infection-like symptoms and its impact on cancer outcomes. Methods: Using Surveillance, Epidemiology and End Results-Medicare, we designed a retrospective cohort study identifying beneficiaries aged >= 66 years diagnosed with bladder cancer from 2007 to 2009. Patients were required to have a hematuria or urinary tract infection claim within 1 year of bladder cancer diagnosis (n = 21 216), and have 2 years of prior Medicare data (n = 18 956) without any precedent hematuria, bladder cancer or urinary tract infection claims (n = 12 195). The number of days to bladder cancer diagnosis was measured, as well as the impact of sex and presenting symptom on time to diagnosis, pathology, and oncological outcomes. Results: The mean time to bladder cancer diagnosis was 72.2 days in women versus 58.9 days in men (P < 0.001). A logistic regression model identified the greatest predictors of >= pT2 pathology were both women (odds ratio 2.08, 95% confidence interval 1.70-2.55) and men (odds ratio 1.71, 95% confidence interval 1.49-1.97) presenting with urinary tract infection. Cox proportional hazards analysis identified an increased risk of mortality from bladder cancer and all causes in women presenting with urinary tract infection (hazard ratio 1.37, 95% confidence interval 1.10-1.71, and hazard ratio 1.47, 95% confidence interval 1.28-1.69) compared with women with hematuria. Conclusions: Women have a longer interval from urinary tract infection to diagnosis of bladder cancer. Urinary tract infection presentation can adversely affect time to diagnosis, pathology and survival. Time to diagnosis seems not to be an independent predictor of bladder cancer outcomes.