Meta-analysis. combination endoscopic and drug therapy to prevent variceal rebleeding in cirrhosis

Background: Combining endoscopic therapy and [beta]-blockers may improve outcomes in patients with cirrhosis and bleeding esophageal varices. Purpose: To assess whether a combination of endoscopic and drug therapy prevents overall and variceal rebleeding and improves survival better than either therapy alone. Data Sources: MEDLINE, EMBASE, the Cochrane Central Register of Controlled Trials, the Cochrane Database of Systematic Reviews, and conference proceedings through 30 December 2007. Study Selection: Randomized trials comparing endoscopic plus [beta]-blocker therapy with either therapy alone, without language restrictions. Data Extraction: Two reviewers independently extracted data on interventions and the primary study outcomes of overall rebleeding and mortality. Metaregression and stratified analysis were used to explore heterogeneity. Data Synthesis: 23 trials (1860 patients) met inclusion criteria. Combination therapy reduced overall rebleeding more than endoscopic therapy alone (pooled relative risk, 0.68 [95% CI, 0.52 to 0.89]; [/.sup.2] - 61%) or [beta]-blocker therapy alone (pooled relative risk, 0.71 [CI, 0.59 to 0.86]; [/.sup.2] = 0%). Combination therapy also reduced variceal rebleeding and variceal recurrence. Reduction in mortality from combination therapy did not statistically significantly differ from that from endoscopic (Peto odds ratio, 0.78 [CI, 0.58 to 1.07) or drug therapy (Peto odds ratio, 0.70 [CI, 0.46 to 1.06]). Effects were independent of the endoscopic procedure (injection sclerotherapy or banding). No trial-level variable associated with the effect was identified through metaregression or stratified analysis. Limitation: Statistically significant heterogeneity in trial quality and evidence for selective reporting and publication bias were found. Conclusion: A combination of endoscopic and drug therapy reduces overall and variceal rebleeding in cirrhosis more than either therapy alone.