In order to redefine the effectiveness of 5-fluorouracil (5-FU) as palliative therapy in patients with metastatic colorectal carcinoma, and to compare the effectiveness of 6-thioguanine (6-TG) with that of 5-FU, we studied 176 patients with metastatic colorectal carcinoma in a randomized prospective trial (SEG 79GI268). The pretreatment performance status of all patients was greater than 50% (ambulatory), and there was an equal distribution of patients with favorable pretreatment characteristics into each of the treatment regimens. Complete responses were only seen to 5-FU, but were obtained in only 3% of instances. The overall complete plus partial response rates were not different for 5-FU (8%) versus 6-TG (3%), or for patients who had shown prior progression on chemotherapy and who then received 6-TG in a nonrandomized fashion (7%). The time to tumor progression on each of the treatment programs was similar, 1.0 months. Survival was also similar in each regimen in the randomized study (6.3 months for 5-FU versus 7.9 months for 6-TG). However, survival was only 4.8 months for patients with previously drug-resistant tumors treated with 6-TC in the nonrandomized arm. In 16 patients failing 6-TG who then received 5-FU, there were no objective responses. Similarly, in patients failing 5-FU on this study who then received 6-TG, there were no responses in nine patients. Dose-limiting toxicity was observed in 40% to 51% of patients, and consisted of myelosuppression, vomiting, or diarrhea. It is concluded that 5-FU is a minimally effective agent in a very small number of patients with metastatic colorectal carcinoma. The drug 6-TG is equally ineffective in this setting. Alternative treatment programs to the systemic use of 5-FU should be considered in patients requiring palliative chemotherapy.