Background: Non-selective beta-blockers are used as a first-line treatment for primary prevention in patients with medium- to high-risk oesophageal varices. The effect of non-selective beta-blockers on mortality is debated and many patients experience adverse events. Trials on banding ligation versus non-selective beta-blockers for patients with oesophageal varices and no history of bleeding have reached equivocal results.
Objectives: To compare the benefits and harms of banding ligation versus non-selective beta-blockers as primary prevention in adult patients with endoscopically verified oesophageal varices that have never bled, irrespective of the underlying liver disease (cirrhosis or other cause).
Search methods: In Febuary 2012, electronic searches (the Cochrane Hepato-Biliary Group Controlled Trials Register, The Cochrane Library, MEDLINE, EMBASE, and Science Citation Index Expanded) and manual searches (including scanning of reference lists in relevant articles and conference proceedings) were performed.
Selection criteria: Randomised trials were included irrespective of publication status, blinding, and language.
Data collection and analysis: Review authors independently extracted data. All-cause mortality was the primary outcome. Intention-to-treat random-effects and fixed-effect model meta-analyses were performed. Results were presented as risk ratios (RR) and 95% confidence intervals (CI) with I(2) statistic values as a measure of intertrial heterogeneity. Subgroup, sensitivity, regression, and trial sequential analyses were performed to evaluate the robustness of the overall results, risks of bias, sources of intertrial heterogeneity, and risks of random errors.
Main results: Nineteen randomised trials on banding ligation versus non-selective beta-blockers for primary prevention in oesophageal varices were included. Most trials specified that only patients with large or high-risk oesophageal varices were included. Bias control was unclear in most trials. In total, 176 of 731 (24%) of the patients randomised to banding ligation and 177 of 773 (23%) of patients randomised to non-selective beta-blockers died. The difference was not statistically significant in a random-effects meta-analysis (RR 1.09; 95% CI 0.92 to 1.30; I(2) = 0%). There was no evidence of bias or small study effects in regression analysis (Egger's test P = 0.997). Trial sequential analysis showed that the heterogeneity-adjusted low-bias trial relative risk estimate required an information size of 3211 patients, that none of the interventions showed superiority, and that the limits of futility have not been reached. When all trials were included, banding ligation reduced upper gastrointestinal bleeding and variceal bleeding compared with non-selective beta-blockers (RR 0.69; 95% CI 0.52 to 0.91; I(2) = 19% and RR 0.67; 95% CI 0.46 to 0.98; I(2) = 31% respectively). The beneficial effect of banding ligation on bleeding was not confirmed in subgroup analyses of trials with adequate randomisation or full paper articles. Bleeding-related mortality was not different in the two intervention arms (29/567 (5.1%) versus 37/585 (6.3%); RR 0.85; 95% CI 0.53 to 1.39; I(2) = 0%). Both interventions were associated with adverse events.
Authors' conclusions: This review found a beneficial effect of banding ligation on primary prevention of upper gastrointestinal bleeding in patient with oesophageal varices. The effect on bleeding did not reduce mortality. Additional evidence is needed to determine whether our results reflect that non-selective beta-blockers have other beneficial effects than on bleeding.