We used the mass isotopomer distribution analysis (MIDA) technique to measure endogenous synthesis of plasma cholesterol in vivo in rats and normal human subjects. Sodium [1-13C]- or [2-13C]acetate was infused, and plasma free cholesterol was analyzed by gas chromatography-mass spectrometry. Frequencies of mass isotopomers M0-M4 (mass-to-charge ratio 368-372) were quantified. The enrichment of the true precursor for cholesterol synthesis (acetyl-coenzyme A in contributing tissues) was determined using the MIDA method. This technique remains mathematically valid even if more than one tissue contributes to circulating free cholesterol. The fractional contribution (f) from endogenous synthesis to free cholesterol in normal women (n = 5) was 2.48 +/- 0.39% after 7 h in the postabsorptive state and 1.27 +/- 0.41% after 8 h of refeeding. In ad libitum-fed rats (n = 12), f was 2.89 +/- 0.44% after 12 h, whereas administration of recombinant tumor necrosis factor increased this value fourfold. Next, the rate constant (k) for removal of labeled free cholesterol from plasma was calculated. Higher masses (M2-M4) were followed to avoid the problem of persistent label incorporation. During the 60 h after cessation of [13C]acetate infusions, k was 0.02490 +/- 0.00298/h in humans. Using these values of k and f, absolute cholesterogenesis was 568 +/- 55 mg/day in normal women (follicular menstrual phase), similar to prior estimates based on whole body sterol balances. Women also exhibited a diurnal variation for endogenous cholesterol synthesis (34.6 +/- 5.4 mg/h nighttime vs. 15.9 +/- 5.2 mg/h daytime) consistent with current knowledge about rhythms in cholesterogenesis. Checks on the model were internally consistent (e.g., comparisons among different isotopomers for calculating precursor enrichment). We conclude that fractional and absolute endogenous cholesterol synthesis can be measured using stable isotopes in vivo by the MIDA technique.