MINIMUM ANALYSIS REQUIREMENTS FOR THE DETECTION OF HELICOBACTER-PYLORI INFECTION BY THE C-13-UREA BREATH TEST

Background: In an attempt to identify the minimum criteria for scoring the C-13-urea breath test, we examined test results from 96 healthy s ubjects, 270 duodenal ulcer patients undergoing treatment to eradicate Helicobacter pylori, and 1000 consecutive breath tests analyzed in ou r laboratory. Methods: The (CO2)-C-13/(CO2)-C-12 isotope ratio in a ba seline sample was compared with that of samples collected at 20, 30, 4 0, and 50 min post-dosing. A positive test was defined as an average i ncrease in the ratio of 6 parts per thousand over baseline. Values for a two-sample analysis (baseline and one other) and for a single-sampl e analysis were extracted from the data and compared with the original analysis outcome. Results: Test results were negative for 186 patient s and positive for 84. The two-sample method had high specificity (94- 99 parts per thousand), sensitivity (95-99 parts per thousand), and po sitive predictive value (88-97 parts per thousand). False-positive res ults occurred most often with samples collected at 20 min and least of ten with the sample at 40 min (11 and 1, respectively). Analyses based on a single sample collected beyond 20 min, in which samples with an isotope ratio greater-than-or-equal-to -15 parts per thousand versus t he standard were defined as positive, gave comparable specificity (98- 99 parts per thousand), slightly reduced sensitivity (92-94%), and pos itive predictive value (96-99%). The analyses of 1000 consecutive brea th tests yielded approximately 1% false-positive and false-negative re sults by the two-sample method; the single-sample method yielded a 2% false-positive and false-negative result. Conclusion: The ''C-urea bre ath test can determine H. pylori infective status from a single breath sample collected at least 30 min post-dose in which the absolute C-13 abundance is greater-than-or-equal-to 15 parts per thousand versus th e standard.