Am Fam Physician. 2005;72(4):683
Clinical Question: Does treatment of patients with bleeding ulcers decrease the rates of mortality or rebleeding, or the need for surgery?
Setting: Inpatient (any location)
Study Design: Meta-analysis (randomized controlled trials)
Synopsis: Increasing stomach pH in patients with an acute bleeding peptic ulcer will enhance platelet function and inhibit fibrinolysis, but do these actions result in better patient outcomes? Researchers combined the results of 21 randomized controlled trials that compared a proton pump inhibitor (PPI; omeprazole [Prilosec] or lansoprazole [Prevacid]) with placebo or a histamine H2 blocker in 2,915 patients with an endoscopically verified bleeding ulcer. Studies were identified by searching four databases, conference proceedings, and by contacting pharmaceutical companies. The studies included use of oral and intravenous therapy. Two authors independently determined whether each study met their inclusion criteria, which included at least one instance of death, rebleeding, or surgery in the study.
Overall, treatment with a PPI did not decrease mortality rates, although it decreased the likelihood of rebleeding or surgery. In the high-quality studies, one episode of rebleeding was avoided for every 10 patients treated (95% confidence interval [CI], 6 to 25) and one fewer patient required surgery for every 25 patients who received treatment (95% CI, 14 to 50). There was no difference in outcomes when comparing oral treatment with intravenous treatment, in higher quality studies, or in studies that used concealed allocation compared with those for which allocation was unknown. There also was no difference in outcomes when evaluating only the results of studies using high-dose continuous intravenous administration. There was no significant heterogeneity among the studies and there was some evidence of publication bias; studies showing no benefit were underrepresented.
Bottom Line: Neither oral nor intravenous use of PPIs decreases the risk of dying as the result of peptic ulcer bleeding. The likelihood of rebleeding or the need for surgery is reduced, with one episode of rebleeding avoided in every 10 patients treated and one surgery avoided for every 25 patients treated. (Level of Evidence: 1a)