Nineteen randomised trials were included in the review (n=5,895).
3G agents were associated with improvement in one-year survival compared with best supportive care (risk difference 7%, 95% CI: 2% to 12%; five trials, n=1,029). Number needed to treat was 14. Response was greater with 2G platinum-based therapy compared with 3G monotherapy (risk difference was -6%, 95% CI: -11% to 0%; four trials, n=871).
3G combination regimens including platinum-based compounds had a higher response rate than 2G platinum based regimens (risk difference 12%, 95% CI: 10% to 15%; 12 trials, n=3,995). The number needed to treat was eight. There was no evidence of statistically significant heterogeneity or publication bias for these outcomes. There was a high degree of statistical heterogeneity for one-year survival when 3G combination regimens including platinum-based compounds were compared with 2G platinum-based regimens (p=0.1).
No difference in one-year survival was found between 3G monotherapy and 2G platinum-based combined regimens.