RESUMO
BACKGROUND: It has been debated which diagnostic test should be preferred for the diagnosis of Helicobacter pylori in patients with gastroduodenal diseases. METHODS: The H. pylori infection was diagnosed prospectively in 97 untreated patients. H. pylori was diagnosed by means of tests based on five different principles: 1) culture, 2) microscopy (HLO), 3) urease activity (urea breath test (UBT) and urease test on biopsy specimens (CLO test)), 4) DNA detection (polymerase chain reaction (PCR)), and 5) IgG antibody detection (enzyme-linked immunosorbent assay (EIA) and Western blotting (WB)). RESULTS: This study showed that two positive tests out of five tests, based on different principles, were most reliable for predicting the H. pylori infection. Most tests had specificities and predictive values for a negative result greater than 90%. The most important difference between the tests was the sensitivity and the predictive value for a positive result (PPV). WB, HLO, UBT, and PCR had sensitivities and PPV greater than 75%. CONCLUSIONS: The non-invasive tests UBT and WB are reliable, both alone and in combination, and they are recommended for the pre-endoscopic diagnosis of H. pylori. WB is recommended as a confirmative test for antibody detection by EIA. When patients have an upper endoscopy, we recommend taking biopsy specimens for culture and histology because of the additional information obtained about susceptibility, virulence determinants, and morphology, including the degree of inflammation.