Selection Effects May Explain Smoking-related Outcome Differences After Radical Cystectomy

Abstract

The impact of smoking on mortality among patients with bladder cancer is subject to controversy. We investigated 1000 patients who consecutively underwent radical cystectomy between 1993 and 2013. Proportional hazards models for competing risks were used to study the combined effects of variables on mortality. Compared to nonsmokers, current smokers were more frequently male (35.7% vs 12.0%, p<0.0001), younger (63.5 vs 70.5 yr, p<0.0001), had a lower body mass index (26.2 vs 27.1kg/m2, p<0.0001), and suffered less frequently from cardiac insufficiency (12.7% vs 19.3%, p=0.0129). Among current smokers there was a trend towards lower bladder cancer mortality and higher competing mortality in comparison to nonsmokers. On multivariable analysis, current smoking was not a predictor of bladder cancer mortality (hazard ratio [HR] in the full model 0.76; p=0.0687) but was a predictor of competing mortality (HR in the optimal model 1.62; p=0.0044). In conclusion, this study did not confirm adverse bladder cancer–related outcome among current smokers after radical cystectomy. With a younger mean age and a male predominance, there was a trend towards lower bladder cancer mortality current smokers that was eventually neutralized by higher competing mortality, illustrating that selection effects may explain some smoking-related outcome differences after radical cystectomy. The single-center design is a study limitation. Patient summaryCurrent smokers are not at higher risk of bladder cancer after radical cystectomy but have a higher risk of competing mortality.