Human, rat, and dog phase I and phase II xenobiotic metabolism in precision-cut liver slices and freshly isolated hepatocytes was compared using a range of substrates. Carbamazepine (50 microM) and styrene (2 mM) were used as probes to study the maintenance of cytochrome P450 and epoxide hydrolase-mediated metabolism in male Sprague-Dawley rat, precision-cut liver slices and hepatocytes. Carbamazepine metabolism in both models resulted in the formation of the bioactive 10,11-epoxide (KM = 766 microM and Vmax = 2.5 pmol/min/mg protein in precision-cut slices). Epoxide formation was higher (2.4-fold) in hepatocytes than slices. Styrene was deactivated to styrene diol at a higher rate in hepatocytes (9.7-fold) than slices. The lower rate of metabolism in slices compared with hepatocytes confirms our previous observations using testosterone, 7-ethoxycoumarin, 1-chloro-2,4-dinitrobenzene and 2-(5'-chloro-2'-phosphoryloxyphenyl)-6-chloro-4-(3H)-quinazolinone in the rat. Testosterone 6 beta-hydroxylation in human liver slices was similar to cultured hepatocytes, but lower than in freshly isolated hepatocytes. 7-Ethoxycoumarin O-deethylation was higher in freshly isolated human hepatocytes, as was the ratio of glucuronide to 7-hydroxycoumarin. Testosterone hydroxylations, 7-ethoxycoumarin O-deethylation, and 1-chloro-2,4-dinitrobenzene conjugation were also lower in male beagle dog slices, compared with freshly isolated hepatocytes. Attempts at long-term preservation of dog liver slices using vitrification and storage for up to 9 days at -196 degrees C resulted in the retention of phase I and phase II metabolism, although conjugation was lower than in freshly prepared slices. Xenobiotic metabolism in short-term incubations is consistently lower in dog and rat precision-cut slices than in freshly isolated hepatocytes; whereas, in humans, this quantitative difference is partly hidden by the large interindividual variation.