The use of adjuvant chemotherapy (AC) in pure urothelial carcinoma of the bladder is established. Regarding variant histology, there is a gap in knowledge concerning the optimal treatment after radical cystectomy (RC). The objective of this study was to assess the effect of AC on overall survival (OS) in patients who had pure urothelial carcinoma, urothelial carcinoma with concomitant variant histology, or another pure variant histology.

Within the National Cancer Data Base, 15,397 patients who underwent RC for nonmetastatic, localized carcinoma of the bladder and had positive lymph nodes (T2N+) or locally advanced stage (≥T3N0/N+) were identified, excluding those who had previously received neoadjuvant chemotherapy. Multivariable Cox regression models were used to examine the specific effect of AC on OS stratified by each distinct histologic subtype, including pure urothelial carcinoma, micropapillary or sarcomatoid differentiation, squamous cell carcinoma, adenocarcinoma, and neuroendocrine tumors. To account for immortal time bias, Cox regression analyses and Kaplan-Meier analyses were conducted with a landmark at 3 months.

In multivariable landmark analyses, AC compared with initial observation was associated with an OS benefit for patients who had pure urothelial carcinoma (hazard ratio, 0.87; 95% confidence interval, 0.82-0.91), whereas no differences were observed with regard to those who had variant histology.

Multivariable Cox regression landmark analysis revealed a survival benefit from AC for patients with a pure urothelial carcinoma. However, a survival benefit of AC for patients who had urothelial carcinoma with concomitant variant histology or other pure variant histology was not demonstrated.