The purpose of this phase II trial was to compare the efficacy, safety and pharmacokinetics of four irinotecan schedules for the treatment of metastatic colorectal cancer. In total, 174 5-fluorouracil pretreated patients were randomised to: arm A (n = 41), 350 mg m-2 irinotecan as a 90-min i.v. infusion q3 weeks, arm B (n = 38), 125 mg m-2 irinotecan as a 90-min i.v. infusion weekly x 4 weeks q6 weeks, arm C (n = 46), 250 mg m-2 irinotecan as a 90-min i.v. infusion q2 weeks, or arm D (n = 49), 10 mg m -2 day-1 irinotecan as a 14-day continuous infusion q3 weeks. No significant differences in efficacy across the four arms were observed, although a shorter time to treatment failure was noted for arm D (1.7 months, P = 0.02). Overall response rates were in the range 5-11%. Secondary end points included median survival (6.4-9.4 months), and time to progression (2.7-3.8 months) and treatment failure (1.7-3.2 months). Similarly, there were no significant differences in the incidence of grade 3-4 toxicities, although the toxicity profile between arms A, B, and C and D did differ. Generally, significantly less haematologic toxicity, alopecia and cholinergic syndrome were observed in arm D, however, there was a trend for increased gastrointestinal toxicity. Irinotecan is an effective and safe second-line treatment for colorectal cancer. The schedules examined yielded equivalent results, indicating that there is no advantage of the prolonged vs short infusion schedules. © 2004 Cancer Research UK.