PREVENTION OF 1ST BLEEDING IN CIRRHOSIS - A METAANALYSIS OF RANDOMIZED TRIALS OF NONSURGICAL TREATMENT

Objective: To assess the effectiveness of beta-blockers and endoscopic sclerotherapy in the prevention of first bleeding and reduction of mortality in patients with cirrhosis and esophagogastric varices. Data Sources: Pertinent studies were selected using MEDLINE (1980 to 1990), reference lists from published articles or reviews, and congress abstract lists. Study Selection: Randomized trials comparing beta-blockers or sclerotherapy with a nonactive treatment. Nine randomized clinical trials of beta-blockers and 19 trials of sclerotherapy were reviewed. Seven trials of beta-blockers and 15 of sclerotherapy were published as full papers. Data Extraction: Crude rates of bleeding and death in treated and control groups were extracted from each trial by three independent observers according to the intention-to-treat principle. The quality of published papers was systematically assessed and scored. Data Synthesis: The Mantel-Haenszel-Peto method was used for statistical evaluation of heterogeneity and for pooling of the results. No substantial heterogeneity was found, and the incidence of bleeding in trials of beta-blockers was significantly reduced (pooled odds ratio, 0.54; 95% Cl, 0.39 to 0.74), particularly in patients with large or medium-sized varices or in those with varices and a hepatic vein pressure gradient above 12 mm Hg; however, only a trend toward reduced mortality was obtained. Solerotherapy trials were highly heterogeneous in the direction of the treatment effects on both bleeding (pooled odds ratio, 0.6; Cl, 0.49 to 0.74) and mortality (pooled odds ratio, 0.76; Cl, 0.61 to 0.94). The quality of the trials and the rate of bleeding in the untreated groups were the major sources of heterogeneity. The favorable results of sclerotherapy were obtained in trials with high bleeding rates among controls; several of these trials had a low quality score. Conclusions: Beta-blockers may be recommended for prevention of first bleeding in cirrhotic patients with varices who have a high risk for bleeding. The effectiveness of sclerotherapy remains undetermined. Further trials in high-risk patients may prove useful if improved criteria to predict bleeding risk become available.