Noninvasive tests as a substitute for histology in the diagnosis of Helicobacter pylori infection

Background: Rapid urease tests for Helicobacter pylori have a sensitivity o f 80% to 90%. Therefore histologic examination of gastric biopsies is recom mended as a "backup" diagnostic test in rapid urease test-negative patients . However, noninvasive tests (urea breath test, serology whole blood antibo dy tests) may provide a more rapid diagnosis and be less expensive but offe r similar accuracy. Methods: Sixty-seven patients (no prior treatment for H pylori, no proton p ump inhibitors, antibiotics, or bismuth within 4 weeks) undergoing endoscop y for evaluation of dyspepsia symptoms and testing rapid urease test-negati ve by antral biopsy were enrolled. All had the following tests: gastric bio psies (2 antral, 1 fundus; H&E and Alcian Yellow stain) examined for gastri tis and H pylori;C-13-UBT; capillary blood for whole blood rapid antibody t ests: FlexSure HP, QuickVue, AccuStat, and Stat-Simple Pylori; serum for Fl exSure HP; HM-CAP enzyme-linked immunoassay. H pylori infection was diagnos ed (reference standard) if chronic gastritis was present on histology and a t least 2 of the 3 following tests were positive: urea breath test, H pylor i organisms unequivocally demonstrated in biopsies on special stain, and/or enzyme-linked immunoassay. The test and treatment costs per patient were c alculated. Results: Of 67 patients with a negative rapid urease test, 4 were positive for H pylori. None had active peptic ulcer disease. Histology only identifi ed 1 patient with organisms visible on special stain. Using chronic active gastritis (neutrophilic and mononuclear infiltrate) as a diagnostic criteri on for H pylori, 6 patients would have been judged positive. However, only 2 of these were truly positive by the reference standard (positive predicti ve value 33%). Negative predictive value for presence of organisms and chro nic active gastritis was 95% and 97%, respectively. All of the noninvasive tests identified all 4 truly positive patients correctly. Urea breath test and FlexSure whole blood assay yielded a substantial number of false-positi ve results (positive predictive value 31% and 36%, respectively); positive predictive value for the other tests ranged from 50% to 80%. All tests exce pt histology had a negative predictive value of 100%. Histology was the mos t costly test (p < 0.001 compared with all other tests), followed by urea b reath test and HM-CAP serology (p < 0.001 compared with all rapid antibody tests). Conclusions: Whole blood or serum antibody testing is a rapid, accurate, an d cost-effective means for establishing H pylori status in rapid urease tes t-negative patients. Whole blood or serology rapid antibody testing should substitute for histology when the patient has not been previously treated f or H pylori.