A high-performance liquid chromatographic (HPLC) assay was developed for the determination of the ratios of the (S)-(+) and (R)-(-) enantiomers of the anti-inflammatory drug carprofen in blood, urine, and feces. The procedure relies on: (a) extraction and purification of carprofen from biological fluids, (b) reaction of carprofen with (S)-(-)-alpha-methylbenzylamine to form the two diastereomeric (S)-(-)-alpha-methylbenzylamides via the 1,1'-carbonyldiimidazole intermediate, (c) purification of the reaction mixture by extraction of the diastereomeric derivatives into hexane at pH 11, and (d) analysis of the diastereomeric derivatives by HPLC with UV detection. The (S)-(+): (R)-(-) ratios in the blood of three subjects receiving single 100-mg oral doses of carprofen were greater than unity up to 16 hr after dosing. The mean +/- SD of the ratios in the early blood samples (0.5, 1, and 2 hr) was 1.21 +/- 0.09, while the mean of the ratios in the later blood samples (4,6,8,12, and 16 hr) was slightly higher (1.48 +/- 0.17). The blood level fall off curves for the (S)-(+) and (R)-(-) enantiomers were similar in each of the three subjects for the 4-16 hr period. The carprofen enantiomers were excreted stereoselectively by humans. An excess of the (S)-(+) enantiomer relative to the (R)-(-) enantiomer was excreted in the urine as the ester glucuronide, while unchanged (R)-(-) enantiomer predominated in the feces. The total urinary plus fecal excretion of the enantiomers (0-96 hr) revealed only a slight excess of the (S)-(+) enantiomer over the (R)-(-) enantiomer, which amounted to 2.1-49% of the dose. Since the amount of carprofen (free and glucuronide) excreted in 96 hr by the three subjects only accounted for 62-72% of the dose, no definitive statement could be made relative to the possible inversion of the carprofen chiral center.