ABSTRACT
Objective:To systematically review the efficacy and safety of acid-suppressive therapy including proton pump inhibitors (PPI) and histamine 2 receptor antagonists (H2RA) and compare with those of placebo or blank control in the postoperative patients with stress ulcer bleeding (SUB) to provide evidence-based reference for clinical practice.Methods:The Cochrane library,Medline,Embase,CBM,CNKI,VIP,Wan Fang Data,clinicaltrials.gov,ISRCTN Register and WHO ICTRP were searched.Only randomized controlled trials (RCTs) of acid-suppressive therapy compared with placebo or blank control for postoperative stress ulcer bleeding were selected with bleeding rate,mortality,adverse drug reactions,gastric juice pH and length of stay as the indices.After the quality evaluation and data extraction,Meta-analysis was performed by using Stata12.0 statistics software.The results were expressed as relative risk(RR) and its corresponding 95% confidence interval(CI).Funnel plot and Eggers test were used to determine the publication bias;and then Grade approach was applied to assess the confidence in the evidence for each outcome.Results:Totally 15 trials enrolling 971 patients were selected,and acid-suppressive therapy was more effective than placebo or blank control in reducing the risk of stress ulcer bleeding,overt upper gastrointestinal bleeding and clinical important bleeding(RR 0.29,95% CI:0.19-0.45;RR 0.25,95%CI:0.10-0.64;RR 0.36,95%CI:0.17-0.77)(moderate),however,there was no statistical significance in the incidence of occult bleeding,mortality and adverse drug reactions (RR 0.79,95%CI:0.41-1.50;RR 0.49,95%CI:0.17-1.38;RR 0.78,95%CI:0.33-1.85,very low confidence).The subgroup analysis of drug classification,operation types and administration juncture showed that the incidence of SUB using PPI (RR=0.27) was lower than that using H2RA (RR=0.33);that of heart surgery (RR=0.20) was lower than that of general surgery (RR=0.31) and neurosurgery(RR=0.37);that of postoperative administration (RR=0.26) was lower than that of preoperative administration (RR=0.23).Conclusion:Acid-suppressive drugs seem to be more effective than placebo or blank control in reducing the risk of bleeding without significant increase of the risk of mortality or adverse drug reactions.The robustness of the conclusion is limited because of the low quality of the trial methodology,incomplete outcome indicators and lack of safety indices for pneumonia and clostridium diffcile-associated infection.Trials with high-quality and larger sample size are still needed to verify its clinical effects.