From January 31, 1986 to January 31, 1989, 184 eligible patients were enrolled in a randomized study of either infusional or bolus 5-fluorouracil (5-FU) for the treatment of metastatic measurable colorectal cancer. Infusion was administered at an escalated dose schedule starting at 350 mg/m2 per day for 2 weeks with a 2-week rest period on a monthly basis, while bolus 5-FU was started at 400-450 mg/m2 for 5 days every 28 days. In these chemotherapy-naive patients with good performance status, the infusion arm produced a response in 11 of 88 patients versus 6 of 82 in the bolus arm (p = 0.384). Progression free survival was significantly longer (p = 0.0139) in the infusion group (3.8 versus 2.3 months), but no significant difference in survival was observed (p = 0.207). As expected, the toxicities of the regimens were different in character, but each had approximately a 20% incidence of significant toxicities. Neither of these methods of administering fluorouracil results in an exceptional response rate, nor does the infusion have an impact on survival as compared to the bolus route. If this type of complicated infusional approach is to continue, especially with increasing dosage (which could be accomplished on our schedule), randomized studies with survival as an endpoint must remain the gold standard.