A conscious dog model was used to study pharmacokinetics of bupivacaine after a short infusion (SI) (15 min) and a prolonged infusion (PI) (24 hours). Bupivacaine was infused in six mongrel dogs at least 10 days after implantation of femoral arterial and venous catheters. Each dog received both the SI and the PI in a random crossover design at a one week interval. Bupivacaine concentration was measured in serum sampled during the SI, during the last hour of the PI, and at frequent intervals during eight hours after cessation of infusion. Indocyanine green (ICG) clearance also was measured during the last hour of the PI and 90 min after cessation of both the SI and the PI. The terminal half-life (T1/2Z) of bupivacaine increased after the PI compared with the SI, 167 +/- 86 vs 53 +/- 13 min, respectively (mean +/- SD; P less than 0.05), and total body clearance (Cl) decreased, 3.4 +/- 1.2 vs 9.5 +/- 4.5 ml.min-1.kg-1, (P less than 0.05), although the volumes of distribution (VZ and VSS) did not change. A decrease in hepatic blood flow did not cause the decrease in Cl because ICG clearance did not change during the three sets of measurements. Thus, the observed increase in T1/2Z and decrease in Cl after the PI as compared with the SI are due to a decrease in hepatic intrinsic clearance of bupivacaine. Differences in the kinetic profile of the two enantiomers of bupivacaine cannot be excluded as a cause. We conclude that the extrapolation of kinetic data of bupivacaine obtained after a short infusion (or a bolus injection) to prolonged dosage must be done with care.