The review text stated that 26 RCTs were eligible for meta-analysis, but the review abstract stated that 25 trials with 2,159 patients were eligible. Details of 26 trials were presented in tables. There was no evidence of publication bias.
Pharmacotherapy versus endoscopic procedures (12 RCTs, n=1,252 patients): There was no statistically significant difference between pharmacotherapy and endoscopic therapy in re-bleeding, variceal re-bleeding, all-cause mortality, or mortality due to re-bleeding. Significant heterogeneity was found for re-bleeding (p=0.001) and re-bleeding due to varices (p=0.001).
Pharmacotherapy plus endoscopic procedure versus endoscopic procedure alone (reported on different occasions as 13 and 14 RCTs, n=1,069 patients): Pharmacotherapy plus endoscopic procedure was associated with a statistically significant reduction in re-bleeding (RR 0.62, 95% CI 0.52 to 0.74) and variceal re-bleeding (RR 0.60, 95% CI 0.44 to 0.82) compared with endoscopy alone. There was no statistically significant difference between endoscopic procedures with versus without pharmacotherapy, in all-cause mortality nor mortality due to re-bleeding. No significant heterogeneity was found for any of these analyses.