Classic or simplified LV5FU2 regimen: Multivariate analysis from a phase III study in metastatic colorectal cancer in elderly patients.

3550 Background: In the early 2000s, classic LV5FU2 (C) (folinic acid, 5FU bolus, then 5FU infusion on D1 and D2) was replaced with simplified LV5FU2 (S) (folinic acid and 5FU bolus on D1 only), considered as effective and less toxic. No trial proved this assertion. The LV5FU2 companion in the FOLFIRI or FOLFOX regimen was C or S. The FFCD 2001-02 study compared in a 2 x 2 factorial design, in not-pretreated elderly patients (75+) with metastatic colorectal cancer, C or S, with or without irinotecan. No significant differences in PFS and OS were observed in the comparison with or without irinotecan. The median OS was 15.2 months in C versus 11.4 months in S, HR = 0.71 (0.55–0.92) and objective response rate was 37.1% in C vs S 25.6% in S, p = 0.004. The aim of this study was to present the factors associated with these differences. Methods: Prognostic factors associated with OS were studied using a Cox model. The multivariate analysis used the significantly different items from the univariate analysis and the differences observed at the inclusion. For each of these items, a subgroup analysis was performed. The second- and third-line treatments were analysed. Results: The 282 patients from the intent-to-treat study were included in the model. In OS, the prognostic factors were C versus S, number of metastatic sites, alkaline phosphatases (AP) and CEA. The interaction test in each subgroup for OS was not significant but C was significantly better in the following subgroup: age > 80 years, male, Karnofsky 100%, 1-2 Charlson index, AP ≤ 2N, leucocyte count > 11,000, CEA > 2N, CA 19-9 ≤2N. No differences were observed in the NCI toxicities but 130 serious adverse events in S versus 102 in C. A second-line was used for 55% patients in C, 46% in S, 81% of them with oxaliplatin or irinotecan in C, 76% after S. The third-line administration (20%) and targeted therapy (15%) were similar in C and S. Conclusions: C-LV5FU2 was superior both in subgroups with better and lower prognostics and this difference cannot be explained by an imbalance between the populations. The toxicity was not higher and a second-line was more often possible after C. The switch from C to S without scientific proof was perhaps a mistake in our practices. Clinical trial information: NCT00303771.