Generalizability of toxicity data from oncology clinical trials to clinical practice: toxicity of irinotecan-based regimens in patients with metastatic colorectal cancer

BACKGROUND The relevance of oncology trial results to clinical practice depends on whether the trial participants are similar to the actual population of patients receiving treatment for the malignancy and whether the patients are treated similarly in both circumstances. Chemotherapy treatments may be more toxic in patients of advanced age and poor performance status-patients typically excluded from clinical trials. METHODS In a retrospective chart review that included all non-trial patients with metastatic colorectal cancer treated with irinotecan-based chemotherapy from January 2004 to September 2006 at our institution, we quantified and subsequently compared the toxicity rates of the irinotecan regimens in clinical practice with published toxicity rates from corresponding phase iii clinical trials. The primary endpoint was the incidence of grades 3 and 4 diarrhea. RESULTS The study included 203 patients, and the irinotecan regimens considered included FOLFIRI [irinotecan, leucovorin, 5-fluorouracil (5fu)],IFL (bevacizumab, irinotecan, 5FU, leucovorin),XELIRI (capecitabine, 3-weekly irinotecan), andirinotecan monotherapy. The rates of grades 3 and 4 diarrhea for FOLFIRI, IFL, XELIRI, and irinotecan monotherapy in clinical practice were 10%, 15%, 17%, and 21% as compared with 10%, 23%, 20%, and 31% respectively in clinical trials. When only patients meeting trial performance status and age criteria were analyzed, the rates of grades 3 and 4 diarrhea by regimen were 11%, 20%, 19%, and 26% respectively. CONCLUSIONS Overall, the toxicity rates for FOLFIRI and irinotecan monotherapy in non-trial patients were not statistically different from the rates quoted in published clinical trials.