OBJECTIVES

It has been suggested that prevalence of Helicobacter pylori (Hp) in peptic ulcer bleeding (PUB) is lower than that in non-complicated ulcers. As Hp infection is elusive in PUB, we hypothesized that this low prevalence could be related to an insufficiently intensive search for the bacteria. The aim of the study was to evaluate whether the prevalence of Hp in PUB depends on the diagnostic methods used in a given study.

METHODS

A systematic review was performed of studies assessing the prevalence of Hp infection in patients with PUB. Data were extracted in duplicate. Univariate and multivariate random-effects meta-regression analyses were performed to determine the factors that explained the differences in Hp prevalence between studies.

RESULTS

The review retrieved 71 articles, including 8,496 patients. The mean prevalence of Hp infection in PUB was 72%. The meta-regression analysis showed that the most significant variables associated with a high prevalence of Hp infection were the use of a diagnostic test delayed until at least 4 weeks after the PUB episode-odds ratio: 2.08, 95% confidence interval: 1.10-3.93, P=0.024-and a lower mean age of patients-odds ratio: 0.95 per additional year, 95% confidence interval: 0.92-0.99, P=0.008.

CONCLUSIONS

Studies that performed a delayed test and those including younger patients found a higher prevalence of Hp, approaching that recorded in cases of non-bleeding ulcers. These results suggest that the low prevalence of Hp infection described in PUB may be related to the methodology of the studies and to patients' characteristics, and that the true prevalence of Hp in PUB is still to be determined. Our data also support the recent recommendations of the International Consensus on Non-Variceal Upper Gastrointestinal Bleeding regarding the performance of a delayed diagnostic test when Hp tests carried out during the acute PUB episode are negative.