Cisplatin and 5-fluorouracil infusion for metastatic colorectal carcinoma. Differences in survival in two patient groups with similar response rates.

Because of reported synergism between 5-fluorouracil (5-FU) and cisplatin (CDDP) in L1210 leukemic mice and activity of this combination in clinical studies, a trial was initiated in previously untreated patients with advanced colorectal carcinoma. Cisplatin at 20 mg/m2 and 5-FU as a continuous infusion at 1000 mg/m2 were both administered for 5 consecutive days every 4 weeks. Forty-one patients were treated at Memorial Sloan-Kettering Cancer Center (MSKCC) and 46 were treated by the Community Clinical Oncology Program (CCOP) physicians. A 50% reduction in measurable disease was seen in 12 of 35 (34%) MSKCC patients and in nine of 41 (22%) of the CCOP patients with 95% confidence intervals of 0.18 to 0.50 and 0.10 to 0.35 in the two groups, respectively. The predominant toxicities were as follows: nausea and vomiting, 32%; mucositis, 26%; leukocyte counts less than 2000 cells/mm3, 17%; platelet counts less than 25,000 cells/mm3, 8%; and severe neurotoxicity, 5%. Dose attenuation was similar in the two groups. The median survival was 16.4 months for the MSKCC group and 9.6 months for the CCOP group (P = 0.0003). Although the baseline characteristics (age, sex, performance status, and baseline lactic dehydrogenase [LDH] and alkaline phosphatase) were similar, on further examination differences between the two groups were evident. In the MSKCC group, 14% of patients with liver metastases had greater than 50% of their liver involved with tumor whereas this occurred in 41% of the CCOP group (P = 0.03). The LDH values greater than 500 U/l were observed in 10% of patients in the MSKCC group and in 37% of the patients in the CCOP group (P = 0.007). Characteristics which reflect the bulk of disease, such as the percent of liver involvement, need to be analyzed in order to evaluate purported survival differences in randomized and nonrandomized trials of colorectal carcinoma.