A meta-analysis of endoscopic variceal ligation for primary prophylaxis of esophageal variceal bleeding

Research output: Contribution to journalArticle

239 Scopus citations


Despite publication of several randomized trials of prophylactic variceal ligation, the effect on bleeding-related outcomes is unclear. We performed a meta-analysis of the trials, as identified by electronic database searching and cross-referencing. Both investigators independently applied inclusion and exclusion criteria, and abstracted data from each trial. Standard meta-analytic techniques were used to compute relative risks and the number needed to treat (NNT) for first variceal bleed, bleed-related mortality, and all-cause mortality. Among 601 patients in 5 homogeneous trials comparing prophylactic ligation with untreated controls, relative risks of first variceal bleed, bleed-related mortality, and all-cause mortality were 0.36 (0.26-0.50), 0.20 (0.11-0.39), and 0.55 (0.43-0.71), with respective NNTs of 4.1, 6.7, and 5.3. Among 283 subjects from 4 trials comparing ligation with β-blocker therapy, the relative risk of first variceal bleed was 0.48 (0.24-0.96), with NNT of 13; however, there was no effect on either bleed-related mortality (relative risk [RR], 0.61; confidence interval [CI], 0.20-1.88) or all-cause mortality (RR, 0.95; CI, 0.56-1.62). In conclusion, compared with untreated controls, prophylactic ligation reduces the risks of variceal bleeding and mortality. Compared with β-blockers, ligation reduces the risk for first variceal bleed but has no effect on mortality. Prophylactic ligation should be considered for patients with large esophageal varices who cannot tolerate β-blockers. Subsequent research should further compare ligation and β-blockers to determine the effect on mortality, and measure ligation's cost-effectiveness.

Original languageEnglish (US)
Pages (from-to)802-807
Number of pages6
Issue number4
StatePublished - Jan 1 2001


ASJC Scopus subject areas

  • Hepatology

Cite this