Bladder recurrence after radical nephroureterectomy: predictors and impact on oncological outcomes.

Objectives: To identify predictive factors of bladder recurrence after radical nephroureterectomy and to evaluate the impact of this event on oncological outcomes.

Methods: We carried out a retrospective analysis of 237 patients treated with radical nephroureterectomy for urothelial carcinoma of the upper tract at our institution from 1998 to 2011. Univariable and multivariable models evaluated the prognostic factors of bladder recurrence, and its impact on recurrence-free survival and cancer-specific survival.

Results: The median age was 69.3 years (interquartile range 60-76). With a median follow up of 44 months (interquartile range 24-79), bladder recurrence occurred in 85 patients (35.9%). A previous history of bladder cancer (P = 0.01) and the presence of concomitant carcinoma in situ (P = 0.005) remained independent predictors of bladder recurrence. The presence of bladder recurrence was not correlated with worse oncological outcomes in terms of disease recurrence (P = 0.075) and cancer-specific mortality (P = 0.06). However, the patients who experienced muscle-invasive bladder cancer recurrence had worse outcomes in terms of cancer-specific mortality (P = 0.01). Standard pathological features of aggressiveness, such as higher tumor stage (P = 0.05), higher grade (P = 0.01) and carcinoma in situ (P = 0.03), were independent predictors of muscle-invasive bladder cancer recurrence.

Conclusions: Previous history of bladder cancer, tumor location and concomitant carcinoma in situ are independent predictors of bladder recurrence in patients undergoing radical nephroureterectomy. Bladder recurrence overall does not impact the oncological outcomes, but a muscle-invasive bladder recurrence is associated with a worse cancer-specific mortality. Standard pathological features of urothelial carcinoma of the upper tract aggressiveness (pT-stage, grade) are independent predictors of muscle-invasive bladder cancer recurrence.