A randomized phase II trial of capecitabine and two different schedules of irinotecan in first-line treatment of metastatic colorectal cancer: efficacy, quality-of-life and toxicity

Background: To determine the efficacy, impact on quality-of-life (QoL) and tolerability of two different irinotecan administration schedules in combination with capecitabine as first-line treatment of metastatic colorectal cancer. Patients and methods: We carried out a randomized phase 11 trial to select one of the following treatment regimens for further investigation: weekly irinotecan at a dose of 70 mg/m(2) days 1, 8, 15, 22, 29 (arm A) or 3-weekly irinotecan at a dose of 300/240mg/m(2) day 1 and days 22 (arm B) in combination with capecitabine 1000 mg/m(2) twice daily days 1-14 and days 22-35 every 6 weeks. Results: Seventy-five patients with good performance status entered the trial. The two arms were well balanced for relevant patient and disease characteristics. The most frequent toxic effects were grade 3/4 diarrhea (arm A: 34%, 13: 19%), grade 3/4 neutropenia (A: 5%, 13: 19%) and grade 2/3 alopecia (A: 26%, 13: 65%). Other grade 3/4 toxic effects were rare (<5%). Response rates were 34% [95% confidence interval (CI) 20% to 51%] in arm A and 35% (95% CI: 20% to 53%) in arm B. Median time to progression was 6.9 (4.6-10.1) and 9.2 (7.9-11.5) months and median overall survival was 17.4 (12.6-23.0+) and 24.7 (16.3-26.4+) months. Patients with an objective tumor response reported better physical well-being (P<0.01), mood (P<0.05), functional performance (P<0.05) and less effort to cope (P<0.05) compared with the non-responders and stable disease patients. Conclusions: The primary end point of this study was the objective response rate and based on the statistical design of the trial, the 3-weekly irinotecan schedule was selected over weekly irinotecan administration. The 3-weekly irinotecan schedule also seemed advantageous in terms of grade 3/4 diarrhea, time to progression, overall survival and patient convenience, but the study was not designed to detect differences in these parameters. In addition, tumor response was shown to have a beneficial effect on QoL indicators.