Furan is a volatile solvent and chemical intermediate that is hepatotoxic and hepatocarcinogenic in rats and mice but is not mutagenic or DNA-reactive. Furan hepatotoxicity requires cytochrome P450 2E1 bioactivation to cis-2-butene-1,4-dial. We have previously shown that furan biotransformation kinetics determined with freshly isolated rat hepatocytes in vitro accurately predict furan pharmacokinetics in vivo [Kedderis et al. (1993) Toxicol. Appl. Pharmacol. 123, 274], suggesting that furan biotransformation kinetics determined with freshly isolated mouse or human hepatocytes can be used to develop species-specific pharmacokinetic models. Hepatocytes from male B6C3F1 mice or human accident victims (n = 3) were incubated with furan vapors to determine the kinetic parameters for furan bioactivation and compared to our previous data for rat hepatocytes. Isolated hepatocytes from all three species rapidly metabolized furan (Vmax of 48 nmol/hr/10(6) mouse hepatocytes, 19-44 nmol/hr/10(6) human hepatocytes, and 18 nmol/hr/10(6) rat hepatocytes) with high affinity (KM ranging from 0.4 to 3.3 microM). The hepatocyte kinetic data and physiological parameters from the literature were used to develop dosimetry models for furan in mice and people. The hepatocyte Vmax values were extrapolated to whole animals assuming 128 x 10(6) hepatocytes/g rodent liver and 137 x 10(6) hepatocytes/g human liver. Simulations of inhalation exposure to 10 ppm furan for 4 hr indicated that the absorbed dose (mg/kg), and consequently the liver dose of cis-2-butene-1,4-dial, was approximately 3- and 10-fold less in humans than in rats or mice, respectively. These results indicate that the target organ concentration, rather than the exposure concentration, is most appropriate for interspecies comparison of dose. The initial rates of furan oxidation in rat, mouse, and human liver were approximately 13-, 24-, and 37-fold greater than the respective rates of blood flow delivery of furan to the liver after 4-hr exposures to < or = 300 ppm. One important consequence of blood flow limitation of furan bioactivation is that the amount of toxic metabolite formed in the liver will be unaffected by increases in Vmax due to the induction of cytochrome P450 2E1. Therefore, the interindividual variations observed in cytochrome P450 2E1 activity among human populations would not be expected to have a significant effect on the extent of furan bioactivation in people. These considerations may be important for human cancer risk assessments of other rapidly metabolized rodent carcinogens.