Mature results of adjuvant colon cancer trials from the fluorouracil-only era.

In this issue of the Journal, Labianca et al. (1) report the results of a randomized trial comparing intraportal chemotherapy, systemic chemotherapy, or the combination of both regimens as adjuvant therapy for patients with Dukes’ stage B or C colon carcinoma. 5-Fluorouracil at a dosage of 500 mg/m of body surface area with 5000 IU of heparin was given through the portal vein for 7 days, starting on the day of surgery (IP regimen). Systemic therapy involved intravenous administration of leucovorin at 100 mg/m over a 2-hour period followed by bolus 5-fluorouracil at 370 mg/m daily on days 1–5, repeated every 28 days for six cycles (SY regimen). Over 6 years, 1199 patients were randomly assigned to treatment during surgery, and 1084 patients deemed eligible were analyzed. A higher percentage of patients in the systemic therapy arm appeared to have a performance status of 1–2 and Dukes’ stage C colon cancer than in the other two arms (Table 1). Intraportal therapy was not started in 7%–10% of patients, whereas systemic chemotherapy was not started in 13%–15% of patients. No statistically significant difference was noted in 5-year survival among the treatment arms. The rationale for administration of 5-fluorouracil by portal venous infusion is that, although most large metastases obtain their blood supply predominantly from the arterial circulation, the portal circulation may be important for micrometastases. The potential benefit of portal perfusion as adjuvant therapy has been tested in patients with colon and rectal cancer. In 1997, the Liver Infusion Meta-analysis Group reported on 3499 patients enrolled in 10 studies that evaluated portal venous infusion of cytotoxic drugs for 5–7 days in the immediate postoperative period (5fluorouracil–heparin in eight trials and 5-fluorouracil–heparin plus mitomycin C in two trials) versus no additional therapy (2). When all patients were included (intention to treat), 42.1% of patients allocated to portal venous infusion had died compared with 47% in the control group, representing a 13.6% reduction in the annual odds of death (P .006). Data on ineligible patients could not be provided for three trials, which introduces potential bias. Analysis restricted to the seven trials for which complete data were available revealed a mortality reduction of 10% (P .07). The 5-year survival curve for all patients randomly assigned to treatment groups showed that the curves overlapped for the first 2 years and diverged thereafter. Since most recurrences occur within 2 years of initial surgery, one would expect the curves to separate during this period if there were a beneficial treatment effect. This result raises the possibility that the late divergence of the curves may be due to chance. Finally, there was no consistent association of portal vein infusion therapy on decreased risk of developing liver metastases among the trials. The AXIS collaborators recently reported the results of a trial of adjuvant radiotherapy and 5-fluorouracil administered by portal venous infusion in patients with colorectal cancer (3). Over an 8-year period, 3583 patients were randomly assigned to surgery with or without 5-fluorouracil (1 g plus 5000 IU of heparin) daily for 7 days. Patients with rectal cancer could be randomly assigned to radiation therapy or to no radiation either