This review concluded that the combination of folinic acid, 5-fluorouracil, oxaliplatin and irinotecan (FOLFOXIRI) significantly improved progression-free survival, survival, response, and complete-removal rates, but it was more toxic than the standard combination of folinic acid, 5-fluorouracil, and irinotecan (FOLFIRI). These conclusions areunreliable primarily because they were derived from two small trials with methodological limitations.

To assess the relative effectiveness of the combination of folinic acid, 5-fluorouracil, oxaliplatin and irinotecan (FOLFOXIRI), compared with the standard combination of folinic acid, 5-fluorouracil, and irinotecan (FOLFIRI) for the first-line treatment of metastatic colorectal cancer.

PubMed, EMBASE, and Cochrane Central Register of Controlled trials (CENTRAL) were searched (dates unspecified), using reported search terms.

Randomised controlled trials that compared FOLFOXIRI with FOLFIRI as a first-line treatment for metastatic colorectal cancer were eligible for inclusion. Progression-free survival and the response rate were the primary outcomes, with overall survival, complete removals (R0 resections), and toxicity as secondary outcomes.

The median patient age was 65 years. The drug regimens were comparable with similar doses of the drugs delivered in the first three days of each 15-day cycle.

The number of reviewers assessing study eligibility was not reported.

The ability of the trial methods to limit selection, performance, attrition and detection bias was assessed. Two reviewers assessed study validity.

Categorical outcomes were tabulated to allow the calculation of odds ratios and associated 95% confidence intervals. The time-to-event data were tabulated to calculate hazard ratios and associated 95% confidence intervals.

Two reviewers independently extracted the data.

Effects were pooled using Mantel-Haenszel fixed-effect models. Heterogeneity was quantified using Ι².

Two trials (527 patients) were included. The authors rated both trials as at a moderate risk of bias as the methods used to randomise patients were unclear and neither trial used blinding.

FOLFOXIRI significantly increased the progression-free survival (HR 0.72, 95% CI 0.60 to 0.88) and overall survival (HR 0.71, 95% CI 0.55 to 0.91; equivalent to between two and four months additional survival). Heterogeneity was substantial for progression-free (Ι²=51%), but not overall survival.

FOLFOXIRI significantly increased toxicity causing neutropenia, thrombocytopenia, nausea, diarrhoea, and neurological events, but not anaemia, fatigue, or febrile neutropenia. The response rates (OR 2.05, 95% CI 1.44 to 2.91) and R0 resections (OR 2.79, 95% CI 1.39 to 5.58) were higher with FOLFOXIRI treatment.

FOLFOXIRI significantly improved progression-free survival, survival, response, and R0 resection rates, but it was more toxic than FOLFIRI.

There were numerous uncertainties in this review. Three major sources of trial information were searched, but the strategy might have been insufficient to identify all eligible trials. The number of reviewers assessing eligibility, and the methods used to reduce bias, were not stated. The use of a fixed-effect model may have been inappropriate, where heterogeneity was moderate to substantial.

Trial quality was assessed and a moderate risk of bias was found. The authors acknowledged that the inclusion of only two trials, with small samples, was a major limitation, but the uncertainty this raised was not reflected in their conclusions. There was also a danger of confusing statistical with clinical significance, given the small difference in overall survival.

The reliability and generalisability of the conclusions are low.

Practice: The authors stated that FOLFOXIRI was an option for first-line treatment of metastatic colorectal cancer patients especially if they were young and had potentially resectable disease.

Research: The authors did not state any implications for research.

Not stated.