13C-urea breath test threshold calculation and evaluation for the detection of Helicobacter pylori infection in children

Abstract

Background: The 13C-urea breath test (UBT) is performed in adults and children with epigastric pain for non-invasively diagnosing a suspected H. pylori infection. Criteria for UBT interpretation have not been generally agreed on and test reliability has not been established in children of different ages. This study aimed at identifying reliable UBT thresholds in children by using 251 UBTs in conjunction with reference histology and by analyzing 1232 UBTs.

Methods: At baseline and 30 and 60 minutes after the administration of 75 mg 13C-urea to children and adolescents (0.25 to 18 years of age), the differences (Delta) of 13CO2/12CO2 ratio in exhaled air (delta) were determined by mass spectrometry. UBT Deltadelta value thresholds were calculated in random subgroups and evaluated in complementary subgroups using logistic regressions on reference histology or bimodal distribution analyses of Deltadelta values from UBTs alone.

Results: Deltadelta values were higher (median, 15.4 per thousand) in positive (133/251, 53 %) than in negative histology (2.4 per thousand). At 30 minutes, the calculated cut-off was 5.3 per thousand (mean regression determination R2 = 0.91), and sensitivity (0.95), specificity (0.97), positive (0.97) and negative predictive values (0.95) were higher than at 60 minutes (threshold 6.8 per thousand, R2 = 0.85). Similar thresholds resulted from UBTs analysis (5.8 per thousand and 6.2 per thousand) when sensitivity and specificity were maximized (concordance probabilities, 0.99 and 0.99). There was no systematic age effect.

Conclusions: In children, 13C UBT cut-offs were obtained and specially validated, entailing high accuracy of non-invasively testing for gastric H. pylori infection.

Figure 1

Distribution of Δδ values in children and adolescents 30 minutes after 75 mg 13 C urea Histograms of Δδ values at 30 minutes for samples with histological validation (panel A) and for UBT samples (panel B). The nadirs of the Δδ value distributions appear to be similarly located. In panel A, the threshold Δδ value of 5.29‰ is depicted as a vertical reference line. It was calculated by logistic regression. All calculations were bootstrapped 1000-fold. In panel B, the overlaid curves correspond to two normal distributions (dotted and broken lines) and their mixed bimodal function's sum curve (continuous line). The right and left peaks presumably encompass children with and without gastric H. pylori infection, respectively. According to the goodness-of-fit-test, the empirical distribution and the mixed bimodal function were not different. The threshold value of 5.78‰ is depicted as a vertical line. It maximizes the sensitivity and specificity in discriminating between the left and right modes.

Figure 2

Δδ values and associated probability of histological H. pylori detection Probability of H. pylori detection by Δδ values. The probabilities of an H. pylori infection are shown as mean logistic functions with 95% confidence intervals of Δδ values at 30 minutes. In panel A, bars at the lower and upper border correspond to histologically absent and present H. pylori, respectively. The prevalence rate of H. pylori infection was 0.53. Probability values exceeding 1–0.53 = 0.47 (horizontal line) indicate H. pylori infection. This probability line joins the logistic function at the threshold Δδ value (vertical line) of 5.29‰. The logistic function is less determined (R² = 0.85) and shows a less steep increase in the probability of H. pylori infection at 60 minutes (not shown) than at 30 minutes (R² = 0.91, panel A). In panel B, bars depict 200 Δδ values randomly generated from the two distributions of the mean binormal model obtained by analysis of 1223 UBTs at 30 minutes. Bars at the lower and upper borders presumably correspond to histologically absent and present H. pylori, respectively.

Figure 3

Sensitivity and Specificity of the 13 C UBT by measured Δδ values Sensitivity (continuous lines) and specificity (broken lines) by Δδ values for samples with histological validation (panel A) and in UBTs alone (panel B). In panel A, discrete steps are due to discrete original data. In panel B, the continuous functions derived from the BM analysis are shown.