Background: The current guidelines for muscle-invasive bladder cancer recommend the use of neoadjuvant cisplatin-based chemotherapy followed by radical cystectomy. However, a trimodal approach involving the combination of maximal transurethral resection (TUR) and combined chemoradiotherapy is an alternative in selected patients. Clinical outcomes of patients with histologic variants have not well been known. Methods: From 1990 to 2015, 148 patients with cT2-3N0M0 muscle-invasive bladder cancer underwent trimodal bladder-preserving therapy consisting of maximal TUR of the bladder tumor, intra-arterial chemotherapy and radiotherapy at our institution. We compared complete response rate (CRR) of bladder preservation, 5-yr cause-specific survival (CSS), and 5-yr overall survival (OS) for the patients with pure urothelial carcinoma (UC) or variant UC. OS and CSS were analyzed by using the Kaplan-Meier method and log-rank test. Results: The median follow-up was 38.3 months. All patients were T2-T3N0M0 (T2, n = 90; T3, n = 58). There were no significant differences in clinical characteristics between pure and variant UC groups. Eleven (7%) of the 148 patients had variant UC; 7 (64%) had UC with squamous and/or glandular differentiation, and 4 (36%) had other forms, including sarcomatoid (n = 1), plasmacytoid (n = 1), signet ring cell (n = 1), and clear cell variants (n = 1). There was no significant difference between pure UC and variant UC for CRR of bladder preservation (85% vs 82%, p = 0.66), the 5-yr CSS (88% vs 75%, p = 0.86) and the 5-yr OS (81% vs 75%, p = 0.66). Conclusions: Our findings indicate that trimodal bladder-preserving therapy can be an effective treatment option for selected muscle-invasive bladder cancer patients with variant UC.