Background: Some case series and cohort studies suggest that acute GI bleeding decreases the sensitivity of Helicobacter pylori diagnostic testing.
Objective: To assess H pylori biopsy testing in patients with acute upper-GI bleeding and 1 month later.
Design: Prospective cohort study using patients as their own controls.
Setting: Urban county hospital.
Patients: Sixty-one patients with acute variceal bleeding.
Interventions: Antral and body endoscopic biopsies at admission and 1 month later.
Main outcome measurements: CLOtest and histologic examinations were performed and biopsy specimens were coded and mixed for blinded histologic examination for H pylori density and inflammation.
Results: CLOtest results changed from H pylori negative at baseline to H pylori positive at 1 month in two patients (3%), from H pylori positive to H pylori negative in 6 patients (10%), and remained the same in 53 (87%). Histologic results changed from H pylori negative at baseline to H pylori positive at 1 month in two patients (3%), from H pylori positive to H pylori negative in 5 patients (8%), and remained the same in 54 (89%). Changes occurred only in patients with low H pylori density. No significant increase in H pylori density or change in inflammatory cell infiltration was seen. CLOtest sensitivity was 8% higher with bleeding vs. 1 month after bleeding (79% vs. 71%; 95% CI of difference was -11% to 27%; i.e., maximal potential decrease in sensitivity with bleeding is 11%).
Limitations: The population is not one for which H pylori testing is recommended, and biopsy test performance was less consistent than expected.
Conclusions: Acute-GI bleeding did not decrease the sensitivity of rapid urease testing, unless the effect lasts more than 1 month. Furthermore, bleeding did not produce falsely negative histologic examinations for H pylori, decrease H pylori density, or alter inflammatory cell infiltration. However, given the lower than expected overall CLOtest sensitivity and frequent use of proton pump inhibitors for GI bleeding, histology may be preferred in this setting.