Systematic review of randomised controlled trials of Helicobacter pylori eradication therapies
Jon Deeks with David Forman, Iris Gordon (Leeds) and the Helicobacter pylori eradication review group
Numerous treatment regimens are available for the eradication of H. pylori, and several hundred studies evaluating such therapies have been published. No effective single agent treatments are available – all the regimens involve a combination of at least two, and more commonly three or four, pharmaceutical compounds, including at least one antibiotic. Therefore, many studies have looked at different drug combinations in different dosages administered for varying time periods. Several reviews and meta-analyses attempting to summarise this literature have been published but these have nearly always combined results from individual treatment arms, irrespective of the comparison being made within the specific study. These reviews have also varied in the extent to which individual studies have been evaluated for quality or to which they have restricted consideration to randomised controlled trials.
This systematic review aims to use a large international network of 17 reviewers to identify and extract data from all randomised studies of H. pylori eradication that meet predefined quality criteria. Results will then be summarised and pooled where appropriate, looking at relative treatment efficacy in all potential two-way comparisons (i.e. active treatment compared with placebo and treatment A compared with treatment B).
The review’s objectives are:
1. To assess the absolute effectiveness of H. pylori eradication regimens compared to placebo in eradicating H. pylori
2. To assess the relative effectiveness of different H. pylori eradication regimens in eradicating H. pylori
3. To compare the incidence of adverse effects associated with different H. pylori eradication regimens
The methodological challenge of this review has been to manage a complex review process with a large number of trials and reviewers while maintaining quality. This is being achieved through computerised systems to manage trial selection, data extraction and data entry. All assessments are made in duplicate by two of the 17 reviewers; 375 relevant trials have been identified.
