C-13 mixed-triglyceride breath test: Isotope selective non-dispersive infrared spectrometry in comparison with isotope ratio mass spectrometry in volunteers and patients with chronic pancreatitis

Background: The C-13 mixed-triglyceride breath test (MTB) has been proposed for the non-invasive assessment of duodenal pancreatic lipase activity. Un til now, stable isotope analysis of CO2 of the MTB has been carried out wit h isotope ratio mass spectrometry (IRMS). The aim of the present study was to compare MTB results by using the new non-dispersive infrared spectrometr y (NDIRS) and the IRMS. Methods: Ten healthy volunteers and 10 patients wit h chronic pancreatitis and exocrine insufficiency were studied. After an ov ernight fast each subject received a test meal containing 250 mg 1,3 distea ryl, 2[C-13] octanoyl glycerol. Breath samples were taken at base line and at 30-min intervals over a period of 6 h postprandially. The C-13/C-12 rati o was determined in each breath sample by NDIRS and CF-IRMS as delta values . Results were expressed as delta over base line (DOB (parts per thousand)) and as cumulative percentage dose of C-13 recovered (cPDR (%)). Correlatio ns between IRMS and NDIRS were tested by linear regression analysis. For me asuring agreement an Altman-Bland plot was performed. Results: A Linear cor relation was found (DOB: y = 0.645 +/- 0.040 x + 1.496 +/- 0.089, r = 0.70, P < 0.0001; cPDR: y = 1.269 +/- 0.031 x + 2.010 +/- 0.353, r = 0.93, P < 0 .0001). For DOE the mean difference (d) was 1.0 parts per thousand, and the standard deviation (s) of the difference was 1.3 parts per thousand. The l imits of agreement (d +/- 2 s) were -1.6 parts per thousand and 3.6 parts p er thousand. Conclusion: The comparison of DOE and cPDR values by NDIRS and IRMS shows a moderate to good linear correlation. However, the distance of the limits of agreement is rather wide. Consequently, the validity of the MTB is diminished, which makes MTB by NDIRS less suitable for exact evaluat ion of noninvasive assessment of duodenal pancreatic lipase activity. Furth er studies are necessary to determine sensitivity and specifity of the MTB with NDIRS in larger study populations.