The 13C-methacetin breath test (MBT) has been proposed for the noninvasive evaluation of the hepatic mixed function oxidase activity. Up to now, stable isotope analysis of carbon dioxide of the MBT has been carried out with isotope ratio mass spectrometry (IRMS). The aim of the present study was to test a recently developed isotope-selective nondispersive infrared spectrometer (NDIRS) in comparison to IRMS in healthy volunteers and patients with liver cirrhosis. Ten healthy volunteers (range 22 to 76 years) and ten patients with histologically proven liver cirrhosis (range 47 to 71 years; Child Pugh score A = 5, B = 3, C = 2) were studied. After an overnight fast each subject received 2 mg/kg BW of 13C-methacetin dissolved in 100 ml of tea. Breath samples were obtained before substrate administration and after 5, 10, 15, 20, 30, 40, 50, 60, 80, 100, 120, 150, 180 min. The 13C/12C-ratio was analyzed in each breath sample both by NDIRS (IRIS, Wagner Analysen Technik, Worpswede, Germany) and CF-IRMS (ABCA, Europa Scientific, Crewe, UK). Results were expressed as delta over baseline (DOB [/1000]) and as cumulative percentage doses of 13C recovered (cPDR [%]) at each time interval. Correlations between IRMS and NDIRS were tested by linear regression correlation. For measuring agreement an Altman-Bland-plot was performed. Applying correlation analysis a linear correlation was found (DOB: y = 1.068 +/- 0.0012.x + 2.088 +/- 0.2126, r = 0.98, p < 0.0001; cPDR: y = 1.148 +/- 0.0109.x + 0.569 +/- 0.172; r = 0.99, p < 0.0001). For DOB the mean difference (d) was 2.9/1000 and the standard deviation (SD) of the difference was 2.7/1000. The limits of agreement (d +/- SD) were -2.5/1000 and 8.3/1000. The comparison of DOB- and cPDR-values by NDIRS and IRMS shows a high linear correlation. However, the distance of the limits of agreement is wide. Consequently, the validity of the MBT could be influenced which could make MBT by NDIRS unprecise for exact evaluation of hepatocellular dysfunction. Further studies are necessary to determine sensitivity and specifity of the MBT with NDIRS in larger study populations.