To assess the effects of experimental liver injury on caffeine metabolism, 1 muCi/kg b.w. of [3-methyl 14C]-caffeine (together with 5 mg/kg b.w. of the cold compound) was injected i.p. to four different experimental groups and respective controls of unanesthetized male Sprague-Dawley rats. Exhaled 14CO2 was completely collected during 4 h and peak exhalation rate and fraction of dose recovered were calculated. 1/3 hepatectomy affected 14CO2 exhalation to a limited extent, decreasing solely peak exhalation rate (p < 0.05 compared to sham-operated controls). 2/3 hepatectomy, on the other hand, resulted in significant reduction (p < 0.01) in both peak exhalation rate (by 59%) and fraction of dose recovered (by 47%), that were proportionate to the loss of liver mass (59%). End-to-side portocaval shunt led to the well-documented hepatic "atrophy", liver weight being diminished on average to 50% within 2 weeks of surgery; however, reductions in peak exhalation rate (by 75%) and fraction of dose recovered (by 64%) were even more pronounced. Finally, 48 h bile duct ligation was equivalent to "functional 2/3 hepatectomy", peak exhalation rate (by 65%) and fraction of dose recovered (by 56%) being markedly diminished despite increased liver weight. These results indicate that 14CO2 exhalation curves following administration of specifically labelled caffeine are quantitative indicators of acute or chronic loss of functioning liver mass. In addition, the 3-demethylation pathway appears to be particularly sensitive to the inhibitory effects of cholestasis on microsomal function.