Detection of Helicobacter pylori Antigen in Stool by Enzyme Immunoassay

ABSTRACT

Invasive techniques for diagnosis of Helicobacter pylori (H. pylori) infection require an endoscopic examination which is expensive and inconvenient and may cause complications. Stool cultures for H. pylori or a direct detection of H. pylori antigen in stools by PCR are expensive, tedious, and have a low sensitivity. We recently used an enzyme immunoassay (EIA) to detect H. pylori antigen in stool specimens. A total of 41 patients were seen at Inha University Hospital, Inchon, Korea between September and October 1998. There were 26 men and 15 women who had an average age of 37.6 years which ranged from 5 to 71 years in the present study. All of these patients came to the hospital complaining of an upper abdominal discomfort and were subjected to endoscopy and biopsies. Fifteen had a gastric ulcer, 13 had a duodenal ulcer, 1 had an early gastric cancer, and there were 12 chronic gastritis patients as shown by endoscopy. The biopsy specimens were examined by histology, CLOTM test, and cultures and these results were used as gold standards. Stool specimens were tested for the H. pylori antigen by EIA. A dual wavelength cut-off of 0.100 that was recommended by the manufacturer gave a good performance (87.1% sensitivity, 100% specificity, 100% positive predictive value, 71.4% negative predictive value, and a 90.2% efficiency). But the adjusted cut-off value using the receiver operating characteristic curve improved the performance of the test (using the cut-off value of 0.024, the sensitivity, specificity, PPV, NPV, and efficiency were 100%, 90.0%, 96.9%, 100%, and 97.6% respectively). Re-evaluation of the cut-off value may be needed for Korean patients. This technique is non-invasive, rapid, easy-to-use, and shows good performance characteristics for diagnosis of H. pylori infections. Therefore, this technique may be a substitute for gastric endoscopy especially in children and some patients who are unable to tolerate an endoscopic examination and it may be substituted for a serologic test in epidemiological research.