|Systematic reviews of the clinical effectiveness and cost-effectiveness of proton pump inhibitors in acute upper gastrointestinal bleeding: PPI therapy initiated prior to endoscopy
|Leontiadis G I, Sreedharan A, Dorward S, Barton P, Delaney B, Howden C W, Orhewere M, Gisbert J, Sharma V K, Rostom A, Moayyedi P, Forman D
This review found limited evidence that proton pump inhibitor (PPI) therapy prior to endoscopy for upper gastrointestinal bleeding may reduce the incidence of stigmata of recent haemorrhage, but not other outcomes. Despite some concerns about potential differences between studies and the risk of publication bias, the findings of this well-conducted review are likely to be reliable.
To compare the effectiveness of proton pump inhibitor (PPI) therapy initiated prior to endoscopy with other treatments to reduce mortality rates in unspecified patients with acute upper gastrointestinal bleeding.
Cochrane Central Register of Controlled Trials, The Cochrane Library, MEDLINE, EMBASE and CINAHL (up to September 2005) and Current Contents (February 2004 to August 2004) and National Research Register and Cochrane Controlled Trials Register (up to 2004) were searched. The search was not restricted to English-language papers. Search terms were reported. Reference lists from retrieved articles and conference proceedings were handsearched. Experts in the field and pharmaceutical companies were contacted for ongoing and unpublished studies.
Randomised controlled trials (RCTs) that compared the effects of proton pump inhibitors with placebo or H2-receptor antagonists (H2RA) on mortality rates in unselected patients with upper gastrointestinal bleeding prior to investigative endoscopy were eligible for inclusion. The primary outcome measure was mortality within 30 days (or time point closest to 30 days) after the acute bleed. Eligible secondary outcomes were also defined.
Included studies compared intravenous omeprazole or oral lansoprazole with or without endoscopic haemostatic therapy (EHT) to placebo (intravenous mannitol or ranitidine, oral famotidine) or H2RA at varying doses and durations in hospital patients from Europe, Asia and Turkey.
Two reviewers independently selected studies for inclusion. A third reviewer was consulted where any discrepancies arose in order to reach a consensus decision.
Assessment of study quality
Two reviewers independently assessed validity using previously published criteria, including items for randomisation, allocation concealment (graded as adequate, uncertain, inadequate or not randomised), blinding, comparability of treatment groups and withdrawals.
The percentage of events for dichotomous outcomes, including mortality, re-bleeding, surgical intervention rates and the need for endoscopic haemostatic therapy were extracted and odds ratios were calculated with 95% confidence intervals (CIs). Continuous outcomes for length of hospital stay were reported as medians with range and standard deviations. Blood transfusion requirements were extracted as percentages. Two reviewers independently performed the data extraction.
Methods of synthesis
A fixed-effect model was used to pool odds ratios. Post hoc analysis was performed for blood transfusion requirements. Subgroup analyses were conducted for allocation concealment, control treatment, type and mode of proton pump inhibitor administration, application of initial endoscopic haemostatic therapy and for patients with peptic ulcer bleeding. Sensitivity analyses were also undertaken by removing individual studies. Heterogeneity was assessed using the I2 test. Publication bias was investigated using funnel plot analysis.
Results of the review
Four RCTs (n=1,512; 760 received intervention and 752 controls) were included in the meta-analysis. One further RCT (n=110 received intervention and 112 controls) was included in the subgroup analyses. The mean number of participants was 378 (range 58 to 1,147). Two RCTs reported adequate concealment. Three were double blinded. Four RCTs reported good baseline comparability between treatment groups.
There were no significant differences in mortality rates at around 30 days (four RCTs), re-bleeding (three RCTs) and surgery (three RCTs) for patients receiving proton pump inhibitor compared to controls. Sensitivity analyses did not significantly alter the results. There was evidence of publication bias for mortality at around 30 days and inconclusive for re-bleeding.
There were significantly fewer incidences of stigmata of recent haemorrhage in the proton pump inhibitor group compared to controls (odds ratio 0.67, 95% CI: 0.54 to 0.84, p=0.0005; four RCTs) at index endoscopy. Sensitivity analyses showed a non-significant result with the removal of one RCT. There was no evidence of publication bias. No significant differences were reported for transfusion requirements (three RCTs) and the number of patients requiring endoscopic haemostatic therapy at index endoscopy (two RCTs).
Subgroup analyses for patients with peptic ulcer bleeding reported a significant difference in the proportion of intervention patients with stigmata of recent haemorrhage at index endoscopy compared to controls (p=0.003) and for the need for endoscopic haemostatic therapy at the index endoscopy in intervention patients compared to placebo (p=0.002). No other subgroup analyses were significant.
There was no evidence of statistical heterogeneity for any outcome.
The use of oral proton pump inhibitors before endoscopy combined with endoscopic haemostatic therapy for patients with major stigmata of recent haemorrhage were likely to be the most cost-effective. Under base-case assumptions, this strategy was likely to be more cost-effective compared to alternatives in which no proton pump inhibitor or an intravenous proton pump inhibitor were used at any threshold over £25,000 per quality-adjusted life-year (QALY). This was the case even when short-term effects (over 28 days) were considered and at any threshold over £200 per life-year gained if long-term effects were included (when life expectancy was also taken into account).
The evidence for treatment with proton pump inhibitor therapy prior to endoscopy for upper gastrointestinal bleeding was limited and did not significantly effect clinically important outcomes such as mortality, re-bleeding or surgical intervention. There was some evidence that proton pump inhibitor therapy might reduce the proportion of patients with stigmata of recent haemorrhage at the index endoscopy, but the clinical significance of this was uncertain.
The review question was clear and was supported by appropriate inclusion criteria. A comprehensive literature search was undertaken using several electronic databases and other appropriate sources. No restrictions were imposed on language and both published and unpublished studies were sought, thereby reducing the risk of language and publication biases. However, statistical tests suggested some evidence of publication bias, although the reliability of the tests is questionable given the low number of studies included. Validity was assessed using relevant criteria and attempts were made to minimise the potential for reviewer error or bias at each stage of the review process. Appropriate methods were used to investigate statistical heterogeneity, but it may not have been appropriate to pool the results as different treatment regimens were used and thus clinical heterogeneity may have been present. Intervention and control patient characteristics were not presented, so it was unclear whether the participants were relevant to the review question. Further, the number of included studies was small and confidence intervals appeared wide. Despite the considerations mentioned, this was a generally well-conducted review and the authors’ conclusions are likely to be reliable.
Implications of the review for practice and research
Practice: The authors did not state any implications for practice.
Research: The authors stated that large RCTs were required to investigate proton pump inhibitor therapy prior to endoscopy in patients with acute upper gastrointestinal bleeding as compared to post-endoscopic initiation of proton pump inhibitor therapy or pre-endoscopic initiation of a control treatment (such as placebo or H2RA). Further RCTs were also needed to compare the effects of different doses and regimens of proton pump inhibitors. Future research may also benefit from reporting outcomes by ulcer location and exploring whether mortality was related to episodes of bleeding or unrelated causes.
Funded by the Health Technology Assessment (HTA) Programme, project number 03/12/03.
Leontiadis G I, Sreedharan A, Dorward S, Barton P, Delaney B, Howden C W, Orhewere M, Gisbert J, Sharma V K, Rostom A, Moayyedi P, Forman D. Systematic reviews of the clinical effectiveness and cost-effectiveness of proton pump inhibitors in acute upper gastrointestinal bleeding: PPI therapy initiated prior to endoscopy . Health Technology Assessment 2007; 11(51): 41-51
Other publications of related interest
Leontiadis GI, Sreedharan A, Dorward S, Barton P, Delaney B, Howden CW, Orhewere M, Gisbert J, Sharma VK, Rostom A, Moayyedi P & Forman D. Systematic reviews of the clinical effectiveness and cost-effectiveness of proton pump inhibitors in acute upper gastrointestinal bleeding. Health Technology Assessment 2007;11(51):1-145.
Subject indexing assigned by NLM
Acute Disease; Aged; Anti-Inflammatory Agents, Non-Steroidal /adverse effects /therapeutic use; Congresses as Topic; Cost-Benefit Analysis; Databases, Bibliographic; Duodenal Ulcer /complications; Endoscopy, Gastrointestinal; Gastrointestinal Hemorrhage /drug therapy /economics /prevention & control; Helicobacter Infections /drug therapy; Helicobacter pylori /drug effects; Histamine H2 Antagonists /economics /therapeutic use; Humans; Middle Aged; Peptic Ulcer Hemorrhage /drug therapy /economics /prevention & control; Proton Pump Inhibitors /economics /therapeutic use; Quality-Adjusted Life Years; Randomized Controlled Trials as Topic; Treatment Outcome; Upper Gastrointestinal Tract /drug effects
Date bibliographic record published
Date abstract record published
This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.