13C-urea breath test for Helicobacter pylori: cut-off point determination by cluster analysis.

1. This study was performed on a large set of 13C-urea breath test results to determine the optimal cut-off point of the test for the diagnosis of Helicobacter pylori (Hp) infection. 2. The following steps were applied to three sets of urea breath test results obtained in three groups of subjects (696 adults before anti-Hp treatment, 1056 adults after anti-Hp treatment and 173 children under 17 years of age): (1) demonstrate the distribution of urea breath test results as a mix of two normal populations (Hp negative and Hp positive) by logarithmic transformation of the results in each group of subjects; (2) apply statistical cluster analysis to determine the separation point between Hp-negative and -positive populations; (3) calculate the mean and SD of each population, and use these parameters in the equation of the normal distribution to establish the frequency curves of Hp-negative and -positive populations; and (4) determine the cut-off point of the urea breath test as the intersection of the two curves, and the risks of error related to it. 3. The optimal cut-off point was found at +3.00 delta/1000, with a risk of false-negative or -positive response of the urea breath test of less than 3%. From this, a cut-off point of +3.00 delta/1000 for the 13C-urea breath test is recommended, with an indetermined zone between +2.5 and +3.5 delta/1000 to account for the spontaneous variation of 13CO2 in breath and the limits of GC-isotope ratio-MS analytical precision.