A significant increase in the dose intensity of chemotherapy with fluoropyrimidines and platinum complexes has resulted from selective circadian timing and/or circadian modulation of the infusion rate. The relevance of such chronopharmacologic strategy for improving the outcome of metastatic colorectal cancer was evaluated in an extended Phase II clinical trial involving 93 patients. Of these, 49% previously had received chemotherapy and/or radiation therapy. The drugs 5-fluorouracil (5-FU, 700 mg/m2/d) and folinic acid (FA, 300 mg/m2/d) combined with oxaliplatin (l-OHP, a nonnephrotoxic platinum complex, 25 mg/m2/d) were infused continuously for 5 days every 3 weeks. In a pilot randomized study, the infusion of all three drugs at a constant rate resulted in World Health Organization (WHO) Grade 3 or 4 toxicity in all four patients compared with no such toxicity in four patients if the infusion rate was modulated according to circadian rhythms. In this Phase II trial, drug delivery was modulated sinusoidally over the 24-hour day with peak flow rates at 4 AM for 5-FU and FA and at 4 PM for l-OHP, using an ambulatory programmable-in-time pump. All patients and 784 of 839 courses (93%) were evaluable for toxicity. Dose-limiting toxicities (WHO Grade 2 to 4) included diarrhea (19% of courses) and vomiting (35% of courses). In addition, WHO Grade 2 to 4 hematologic or mucosal toxicity, respectively, occurred in 2.5% and 7% of courses. Two toxic deaths were encountered. Peripheral sensory neuropathy led to discontinuation of l-OHP in 14 patients after 7 to 12 courses; it completely disappeared within 3 months. Fifty-four of the 93 patients had an objective response (58%; 95% confidence limits, 48% to 68%), irrespective of previous treatment or prior documented progression while receiving standard chemotherapy with 5-FU and FA or continuous 5-FU. Complete responses (CR) were seen in 6 patients (4 of which were proved histologically) and, after surgery, in 12 additional patients (overall CR rate, 18 of 93 [19%]; 95% confidence limits, 11% to 27%). Median progression-free survival (PFS) and overall survival were, respectively, 10 and 15 months, irrespective of prior therapy. Both PFS and survival were significantly longer in patients with a good performance status (PS, 0 or 1, by WHO criteria; respectively, 12 and 21 months) than in patients with poor PS (respectively, 8 and 10 months; P less than 0.01, by log-rank test). This chronopharmacologic protocol may have circumvented, to some extent, both the natural and acquired resistance of colorectal cancer to chemotherapy.