Seventeen RCTs met the inclusion criteria (n=3,529; range 84 to 573). Data were not available for two RCTs, therefore 15 RCTs were included in the analysis (n=3,292; range 84 to 573). All scored two or three on the Jadad scale. Only seven RCTs reported adequate randomisation methods.
Adjuvant chemotherapy improved the five-year overall survival rate (RR 0.90, 95% CI: 0.84, 0.96, p=0.001; 15 RCTs) and disease-free survival rate (RR 0.87, 95% CI: 0.78, 0.96, p=0.005; 7 RCTs). No statistically significant heterogeneity between studies was observed for either analysis.
Results remained significant when Western and Asian subgroups were analysed separately. When subgrouped by treatment regimen, the meta-analysis suggested slightly better five-year overall survival rates only with mitomycin-containing regimens, with or without anthracycline. The numbers of studies within these subgroups were small.
The funnel plot showed no evidence of publication bias.