Stool testing for Helicobacter pylori infection: yet another noninvasive alternative

Abstract

A prospective, multicenter study was performed in 11 European centers to evaluate the accuracy of stool testing for active Helicobacter pylori (H. pylori) infection, using the HpSA ELISA antigen assay. The accuracy of this test was assessed in a large number of patients both before and after treatment by comparing results with a rigidly defined gold standard consisting of gastric antral and body biopsies with normal and special stains, as well as culture and rapid urease testing. The accuracy of the stool test was also compared with the currently accepted best noninvasive test, the carbon-13 urea breath test. Five hundred and one treatment naive patients (276 men, mean age 52 yr) were tested after endoscopy (491 patients had evaluable results). The sensitivity and specificity of the stool test were 94.1% (the 95% confidence interval was 90.6-96.6%) and 91.8% (87.3-95.1%), respectively. Pretreatment sensitivity and specificity for breath testing were similar (95.3% [92.2-97.5%] and 97.7% [94.8-99.3%], respectively). One hundred and seven infected patients were reassessed 4 wk after undergoing therapy for H. pylori infection. The posttreatment sensitivity and specificity of the HpSA assay was also excellent, although the confidence intervals were significantly wider because of the smaller number of patients (90% [68.3-98.9%], and 95.3% [88.5-98.7%], respectively). Posttreatment sensitivity and specificity for urea breath testing were also similar (90% [68.3-98.8%] and 98.9% [93.8-100%], respectively). Intercenter variability of test results did not reach statistical significance with either testing method. The authors concluded that HpSA stool testing is a reliable and easy-to-use method for diagnosing H. pylori infection in both treatment naïve and posttreatment patients that compares well with carbon-13 urea breath testing. They site specific advantages of the HpSA assay to include: 1) a more simple sampling method (only one stool specimen is required); 2) the lack of a requirement for trained personnel at the testing site; and 3) there is no need for expensive equipment.