In order to avoid the use of radioactive tracers for the determination of human ketone body turnover, we have developed a method using a primed-continuous infusion of 13C-labelled acetoacetate or D-fl-hydroxybutyrate. Determination of the mole percent enrichment of blood acetoacetate and D-fl-hydroxybutyrate was performed by gas chromatography/mass spectrometry. In the post-absorptive state, the mean total ketone body appearance rate, determined in four subjects, was 3.74~tmol.kg-l.min -1 using [3,4-~3C2] acetoacetate and 2.76 gmol.kg -1.rain -1 using [3-13C]D-fl-hydroxybutyrate, values in agreement with those reported in studies with 14C-labelled tracers. In order to evaluate the usefulness of the method for determination of ketone body kinetics in non steady-state conditions, we infused four subjects with natural sodium acetoacetate and calculated the isotopically determined total ketone body appearance rate using a single compartment model (volume of distribution 0.201/kg; functional pool fraction: 1). During the tests with [3,4-13C2] -acetoacetate, the actual infusion rates of natural acetoacetate were 7.3+0.3, 14.6+0.8, 21.9+1.2 and 10.9+0.6gmol. kg -1-min -1 whereas the corresponding isotopically determined total ketone body appearance rates were respectively 9.2_+1.0, 16.3+0.7, 23.1+1.1 and 10.7_+ 0.8 p~molkg -1. rain -1. During the tests with [3-13C]D-fl-hydroxybutyrate, the actual infusion rates were 8.4_+0.5, 16.8-+0.9, 25.2+ 1.4 and 12.6 + 0.8 p, mol. kg -1. min -1, and the isotopically determined appearance rates respectively 11.1 +0.7, 16.7_+0.7, 25.0 _+ 1.1 and 11.1 _+ 0.7 pmol. kg-1. rain-~. Thus, using either tracer we found a good agreement between acetoacetate infusion rate and tracer-determined appearance rate of ketone bodies. In conclusion, the present method may be used to determine human ketone body kinetics under steady-state and non steady-state conditions.