King's College London

BACKGROUND: Previous research has associated repeated transurethral procedures after a diagnosis of non-muscle invasive bladder cancer (NMIBC) with increased risk of death of causes other than bladder cancer. AIM: We investigated the overall and disease-specific risk of death in patients with NMIBC compared to a background population sample. METHODS: We utilized the database BladderBaSe 2.0 containing tumor-specific, health-related and socio-demographic information for 38,547 patients with NMIBC not primarily treated with radical cystectomy and 192,733 individuals in a comparison cohort, matched on age, gender, and county of residence. The cohorts were compared using Kaplan-Meier curves and Hazard ratios (HR) from a Cox regression models. In the NMIBC cohort, we analyzed the association between number of transurethral procedures and death conditioned on surviving two or five years. RESULTS: Overall survival and survival from causes other than bladder cancer estimated with Kaplan-Meier curves was 9.3% (95% confidence interval (CI) (8.6%-10.0%)) and 1.4% (95% CI 0.7%-2.1%) lower respectively for the NMIBC cohort compared to the comparison cohort at ten years. In a Cox model adjusted for prognostic group, educational level and comorbidity, the HR was 1.03 (95% CI 1.01-1.05) for death from causes other than bladder cancer comparing the NMIBC cohort to the comparison cohort. Among the NMIBC patients, there was no discernible association between number of transurethral procedures and deaths of causes other than bladder cancer after adjustment. The number of procedures were, however, associated with risk of dying from bladder cancer HR 3.56 (95% CI 3.43-3.68) for four or more resections versus one within two years of follow-up. CONCLUSION: The results indicate that repeated diagnostic or therapeutic transurethral procedures under follow-up do not increase of risk dying from causes other than bladder cancer. The modestly raised risk for NMIBC patients dying from causes other than bladder cancer is likely explained by residual confounding.