We present a sensitive tracer method, suitable for in vivo human research, that uses beta-[(14)C]carotene coupled with accelerator mass spectrometry (AMS) detection. Using this approach, the concentration-time course of a physiological (306 microgram 200 nCi) oral dose of beta-[(14)C]carotene was determined for 209 days in plasma. Analytes included beta-[(14)C]carotene, [(14)C]retinyl esters, [(14)C]retinol, and several [(14)C]retinoic acids. There was a 5.5-h lag between dosing and the appearance of (14)C in plasma. Labeled beta-carotene and [(14)C]retinyl esters rose and displayed several maxima with virtually identical kinetic profiles over the first 24-h period; elevated [(14)C]retinyl ester concentrations were sustained in the plasma compartment for >21 h postdosing. The appearance of [(14)C]retinol in plasma was also delayed 5.5 h postdosing and its concentration rose linearly for 28 h before declining. Cumulative urine and stool were collected for 17 and 10 days, respectively, and 57.4% of the dose was recovered in the stool within 48 h postdosing. The stool was the major excretion route for the absorbed dose. The turnover times (1/k(el)) for beta-carotene and retinol were 58 and 302 days, respectively. Area under the curve analysis of the plasma response curves suggested a molar vitamin A value of 0.53 for beta-carotene, with a minimum of 62% of the absorbed beta-carotene being cleaved to vitamin A.In summary, AMS is an excellent tool for defining the in vivo metabolic behavior of beta-carotene and related compounds at physiological concentrations. Further, our data suggest that retinyl esters derived from beta-carotene may undergo hepatic resecretion with VLDL in a process similar to that observed for beta-carotene.