Delayed diarrhea is the main toxicity of irinotecan at the currently recommended dose of 350 mg/m2 30-minute intravenous infusion, once every 3 weeks. This phase II, multicenter, open-label, randomized study was primarily designed to evaluate the effect of a 15-day Tiorfan (racecadotril) treatment on the incidence and severity of irinotecan-induced delayed diarrhea. One hundred thirty-six patients with metastatic colorectal cancer who failed to respond to a 5-fluorouracil-based treatment received 714 cycles of irinotecan. The patients were randomly allocated either to group A (68 patients) and received Tiorfan (300 mg/day) from D0 to D15 or to group B (68 patients) with no prophylactic treatment. Delayed diarrhea occurred in 197 of 355 cycles (55%) in Group A and 203 of 344 cycles (59%) in Group B. grade III-IV diarrhea was reported in 17 of 40 compliant patients (42%) in group A and 31 of 68 evaluable patients (45%) in group B. No difference was observed between the two groups for delayed diarrhea characteristics, incidence, or severity. The response rate in 99 evaluable patients was 12.1% (6.4%-20.2%). This study has shown that Tiorfan given prophylactically at 300 mg/day has no effect on delayed diarrhea.