Twenty-three RCTs (n=1,860) were included in the review. The sample sizes ranged from 28 to 171 participants.
Of the 16 trials published as articles, half fulfilled more than four of the six quality measures, but none met all of them. The investigator was blinded in only 2 trials. The baseline characteristics of the treatment and control groups were balanced in all trials. The median follow-up periods of trials comparing combination therapy with endoscopic therapy and with drug therapy were 18.7 and 24 months, respectively. Publication bias was observed for re-bleeding, but not for mortality.
Combination therapy was more effective than endoscopic therapy at reducing re-bleeding (RR 0.68, 95% CI: 0.52, 0.89) and mortality (Peto OR 0.78, 95% CI: 0.58, 1.07), although the mortality result was not statistically significant. Substantial clinical and statistical heterogeneity was found for re-bleeding (I2=61%; p<0.001), but not for mortality (I2=0%; p=0.85). In the meta-regression analyses, no covariate was related to treatment effect for re-bleeding; for mortality, trials with longer follow-up showed an effect size close to that observed in the overall analysis, whereas trials with shorter follow-up did not show any effect. Combination therapy seemed to reduce the mortality rate in trials with intention-to-treat designs.
Combination therapy was more effective than drug therapy for preventing re-bleeding (RR 0.71, 95% CI: 0.59, 0.86) and mortality (Peto OR 0.70, 95% CI: 0.46, 1.06), although the mortality result was not statistically significant. No significant heterogeneity was found.
Combination therapy also reduced variceal re-bleeding and variceal recurrence. Asthenia was the most commonly reported β-blocker-related side-effect. Further analyses were reported.