Forty-five RCTs (n=3,236) were included in the meta-analysis. Reporting of methodological issues of included studies was generally poor, but no further details on trial quality were reported.
Traditional Chinese medicines significantly improved the complete response effects score compared to controls (RR 1.26, 95% CI 1.04 to 1.52; 23 trials) and the partial response effects score (RR 1.27, 95% CI 1.17 to 1.38; 37 trials, n=unclear). A significantly increased progressive disease rate was found in control groups compared to treatment (RR 0.54, 95% CI 0.45 to 0.64; 37 trials, n=unclear). There was no significant heterogeneity measured between trials. Traditional Chinese medicines did not significantly influence stable disease.
Traditional Chinese medicines significantly improved survival rates compared to controls at six-month outcomes (RR 1.10, 95% CI 1.04 to 1.15; 15 trials), 12 months (RR 1.26, 95% CI 1.17 to 1.36; 22 trials), 18 months (RR 1.71, 95% CI 1.002 to 2.91; four trials), 24 months (RR 1.72, 95% CI 1.40 to 2.03; 15 trials) and 36 months (RR 2.40, 95% CI 1.65 to 3.49; eight trials). Significant heterogeneity between trials was found only at 18 months (I2=70%).
Traditional Chinese medicines significantly improved abdominal pain (RR 1.50, 95% CI 1.09 to 2.07; six trials), fatigue (RR 1.54, 95% CI 1.17 to 2.01; four trials) and appetite (RR 1.53, 95% CI 1.14 to 2.05; four trials). There was no significant heterogeneity measured between trials.
Meta-regression indicated products that contained ginseng, astragalus and Mylabris had a larger treatment effect in terms of complete response (OR 1.34, 95% CI, 1.04 to 1.71) and products that contained astragalus had this effect (OR 1.35, 95% CI, 1.001 to 1.80). There was no statistical evidence of publication bias, but the authors reported that the funnel plots could not rule out publication bias (the funnel plots were not presented).