OBJECTIVE
To perform a systematic review and a meta-analysis of diagnostic accuracy of the different tests aimed to detect Helicobacter pylori infection in patients with upper gastrointestinal bleeding (UGIB).
METHODS
Selection of studies: assessing the accuracy of H. pylori diagnostic methods in patients with UGIB. Search strategy: electronic bibliographical searches. Data extraction: independently done by two reviewers. Data synthesis: meta-analyses of the different tests were performed combining the sensitivities, specificities, and likelihood ratios (LRs) of the individual studies.
RESULTS
Studies showed a high degree of heterogeneity. Pooled sensitivity, specificity, LR+ and LR– (95% confidence interval (CI)) for the different methods were: Rapid urease test (16 studies/1,417 patients): 0.67 (0.64–0.70), 0.93 (0.90–0.96), 9.6 (5.1–18.1), and 0.31 (0.22–0.44). Histology (10 studies/827 patients): 0.70 (0.66–0.74), 0.90 (0.85–0.94), 6.7 (2.5–18.4), and 0.23 (0.12–0.46). Culture (3 studies/314 patients): 0.45 (0.39–0.51), 0.98 (0.92–1.00), 19.6 (4–96), and 0.31 (0.05–1.9). Urea breath test (8 studies/520 patients): 0.93 (0.90–0.95), 0.92 (0.87–0.96), 9.5 (3.9–23.3), and 0.11 (0.07–0.16). Stool antigen test (6 studies/377 patients): 0.87 (0.82–0.91), 0.70 (0.62–0.78), 2.3 (1.4–4), and 0.2 (0.13–0.3). Serology (9 studies/803 patients): 0.88 (0.85–0.90), 0.69 (0.62–0.75), 2.5 (1.6–4.1), and 0.25 (0.19–0.33).
CONCLUSION
Biopsy-based methods, such as rapid urease test, histology, and culture, have a low sensitivity, but a high specificity, in patients with UGIB. The accuracy of 13C-urea breath test remains very high in these patients. Stool antigen test is less accurate in UGIB. Although serology seems not to be influenced by UGIB, it cannot be recommended as the first diagnostic test for H. pylori infection in this setting.
