Objectives: To compare the accuracy of the breath test using the isotope ratio mass spectrometer (IRMS) versus the nondispersive isotope-selective infrared spectrometer (NDIRS) in the diagnosis of Helicobacter pylori infection.
Method: Multicenter study in 4 Spanish hospitals. One group of dyspeptic patients who had not undergone prior eradication therapy and another group of patients with gastric ulcer or gastrointestinal bleeding due to gastroduodenal ulcer receiving H. pylori eradication therapy were included in the study. A reference standard based on histology and the rapid urease test was used. The breast test (TAU-KIT, Isomed S.L., Madrid, Spain) was performed with citric acid and 100 mg of 13C-urea. Samples of expired air were collected in tubes and bags for reading with the IRMS (ABCA, PDZ, Crewe, Manchester, England) and the NDIRS (UBiT-IR200, Otsuka Electronics, Co, Osaka, Japan), respectively. The endoscopist, pathologist and person responsible for reading the urease test and both breath tests were blinded to the results of the other diagnostic methods.
Results: Forty-one patients were included. The prevalence of H. pylori was 26%. No differences were found on comparing the mean values obtained with the IRMS and the NDIRS: 13 (standard deviation) (24) and 14 (25) delta units, respectively. The area under the ROC curve for the IRMS and the NDIRS was 0.96. The diagnostic accuracy for the best cut-off point with the IRMS and the NDIRS was, respectively: sensitivity (90 and 100%), specificity (96 and 89%), positive predictive value (90 and 77%), negative predictive value (96 and 100%), + likelihaod ratio (25 and 9.3) and (0.1 and 0). A close correlation was found between the values of the IRMS and those of the NDIRS (lineal regression equation, Y = 1.1 + 1.004. X; r = 0.97).
Conclusion: Both the spectrometers used to evaluate the breath test, the IRMS and the NDIRS, offer a high degree of accuracy in the diagnosis of H. pylori infection.