The methodology for measuring in vivo oxidation of substrates labeled with the nonradioactive carbon isotope 13C has been developed with isotope ratio mass spectrometry. The use of 13C offers the possibility of utilizing CO2 breath tests in infants, children, pregnant women, and all subjects in whom 14CO2 breath tests cannot be used. The excretion of 140 nmol/kg-hr of 13CO2 produced from the oxidation of the labeled substrate could be detected with 95% confidence during a total CO2 excretion of 9 mM/kg-hr. The precision of CO2 breath tests using 13C is limited by the natural fluctuations of the ratio of 13C/12C in expired CO2, which occur with a standard deviation of 0.72%, or approximately 7 parts 13CO2 per 10(6) parts expired CO2. Larger excursions in the ratio were observed if the subjects ate shortly before or during the breath test. Clinically significant diagnostic tests can reasonably be expected to require the excretion of 2 to 20 times as much labeled CO2, or 0.28 to 1.4 micronM/kg-hr.