Background: Several studies have investigated the possible role of the adjuvant chemotherapy after curative resection for gastric cancer failing to show a clear indication; previous meta-analyses suggested small survival benefit of adjuvant chemotherapy, but the statistical methods used were open to criticisms.

Materials and methods: Randomised trials were identified by means of Medline and CancerLit and by selecting references from relevant articles. Systematic review of all randomised clinical trials of adjuvant chemotherapy for gastric cancer compared with surgery alone, published before January 2000, were considered. Pooling of data was performed using the fixed effect model. Death for any cause was the study endpoint. The hazard ratio and its 95% confidence intervals (95% CI), derived according to the method of Parmar, were the statistics chosen for summarising the relative benefit of chemotherapy versus control.

Results: Overall 20 articles (21 comparisons) were considered for analysis. Three studies used single agent chemotherapy, seven combination of 5-fluorouracil (5-FU) with anthracyclin, ten combination of 5-FU without anthracyclines. Information on 3658 patients, 2180 deaths, was collected.

Chemotherapy reduced the risk of death by 18% (hazard ratio 0.82, 95% CI: 0.75–0.89, P < 0.001). Association of Anthracyclines to 5-FU did not show a statistically significant improvement when compared with the effect of the other regimens.

Conclusions: Chemotherapy produces a small survival benefit in patients with curatively resected gastric cancer. However, taking into account the limitations of literature based meta-analyses, adjuvant chemotherapy is still to be considered as an investigational approach.