Enantioselective pharmacokinetics and absorption of eflornithine in the rat was investigated using population pharmacokinetic modeling and a modified deconvolution method. Bidirectional permeability of L- and D-eflornithine was investigated in Caco-2 cells. The rat was administered racemic eflornithine hydrochloride as a single oral dose [40-3,000 mg/kg bodyweight (BW)] or intravenously (IV) (100-2,700 mg/kg BW infused over 60-400 min). Serial arterial blood samples were collected and L- and D-eflornithine were quantitated with a previously published chiral bioanalysis method. The D:L concentration ratio was determined in rat faeces. Intravenous L-and D-eflornithine plasma concentration-time data was analyzed using population pharmacokinetic modeling and described with a 3-compartment pharmacokinetic model with saturable binding to one of the peripheral compartments. Oral plasma concentration-time data was analyzed using a modified deconvolution method accounting for nonlinearities in the eflornithine pharmacokinetics. Clearance was similar for both enantiomers (3.36 and 3.09 mL/min). Oral bioavailability was estimated by deconvolution at 30 and 59% for L- and D-eflornithine. The D:L concentration ratio in feces was 0.49 and the Caco-2 cell permeability was similar for both enantiomers (6-10 × 10(-8) cm/s) with no evident involvement of active transport or efflux. The results presented here suggest that the difference in the bioavailability between eflornithine enantiomers is caused by a stereoselective difference in extent rather than rate of absorption. The presented modified deconvolution method made it possible to account for the non-linear component in the suggested three-compartment pharmacokinetic model thus rapidly estimating eflornithine oral bioavailability.