Background/aims: The detection of Helicobacter pylori (H. pylori) antigen in stool by conventional ELISA is a reliable non-invasive method for the diagnosis of H. pylori infection in untreated patients. Recently, rapid in-office stool tests have been developed for the same purpose.
Methodology: We have prospectively evaluated the performances of a commercially available enzyme-linked immunoassay (Novitec EIA) and a rapid near-patient immunochromatographic stool test (Stick H. pyl) for the detection of H. pylori stool antigen. Fifty H. pylori positive and 50 negative patients were included. H. pylori infection was diagnosed by using histology, rapid urease test and urea breath test. Patients were classified as positive if two of the three tests were positive and negative if all the three tests were negative. Testing was carried out according to the manufacturer's instructions.
Results: Novitec EIA had 8% equivocal results. If they were interpreted as negative the sensitivity, specificity, positive predictive value, negative predictive value and overall diagnostic accuracy were 82%, 86%, 86%, 83% and 84% and if as positive 88%, 76%, 79%, 86% and 82% respectively. ROC curve analysis showed a cut-off value of 0.144 for our population. The corresponding numbers for this cut-off value were: 82%, 94%, 93%, 84% and 88%. The respective numbers for Stick H. pyl were 78%, 78%, 76%, 78% and 79%.
Conclusions: Novitec EIA performed well in this cohort of Greek patients and demonstrated a high specificity and positive predictive value when we adjusted the cut-off at 0.144. Performance of Stick H. pyl was sub-optimal.