Seventy-seven patients with advanced urothelial cancer were treated with methotrexate, vinblastine, adriamycin, and cisplatin (M-VAC). Of these 77 patients, 65 could be evaluated for response and 74 for toxicity. Response rates were 65% in the primary organs (62% in the renal pelvis and ureter, 67% in the bladder), 68% in the lymph nodes, 60% in the lung, 25% in the bone and 14% in the liver. Complete responses were noted in 11 patients (17%) and partial responses in 26 patients for an overall response rate of 57% (95% confidence limits 45 to 69%). The median durations of response were 11 months for complete response patients and 7 months for partial response patients. Of the 65 patients 20 (31%) are alive, and 1-, 2-, and 3-year survival rates were 65%, 37%, and 25%, respectively. While survival rates of responders were higher than those of nonresponders with a statistical significance until 15 months, no significant differences were observed in survival rates between these two groups in the subsequent period. The M-VAC regimen was used for 15 patients as a neoadjuvant chemotherapy. Of the 15 patients, 8 responded and primary organs were preserved in 6 of the 8 responders. Histological effects classified according to Oboshi-Shimosato's criteria were G.I in 9, G.IIA in 3, G.IIB in 1, and G.IVC in 2. There were no significant differences in survival rates according to responses and histological effects. Factors related to response were analyzed with a multiple logistic regression model on 54 patients treated with intravenous administration of drugs and whose histological type was transitional cell carcinoma. The analysis results indicate that the presence of distant metastases is an important factor in predicting poor efficacy. Sixteen of 74 patients (22%) had white blood cell count of less than 1,000 cells per mm3 in the first cycle, while the decrease of platelet count was mild in degree compared with that of the white blood cell count. Patients with elevations of serum creatinine, GOT, and GPT were low in frequency, and toxic symptoms were controllable. Factors significantly related to the occurrence of side effects were sex, performance status, prior radiotherapy, prior chemotherapy, and the method of drug administration. Among these factors, prior radiotherapy was related to severe decrease of white blood cell count. While an excellent overall response rate was provided with the M-VAC regimen, disadvantages of the present regimen were low effectiveness in the bone and liver, and short duration of response.(ABSTRACT TRUNCATED AT 400 WORDS)