1. The clearance of dihydroartemisinin (DHA) in control and malaria-infected (MI) rats was investigated using the isolated perfused rat liver (IPRL) model and hepatic microsomal studies. 2. In the recirculating IPRL, clearance of DHA was reduced from a mean (s.d.) of 8.2+/-1.8 ml min(-1) in controls (n=8) to 6.0+/-1.0 ml min(-1) in MI (n=8; P<0.01). Clearance in control livers was similar to the perfusion flow rate, suggesting a high hepatic extraction ratio for DHA. 3. Single-pass IPRL studies in controls (n=8) showed that DHA bioavailability at 1.3, 8 and 38 microm was 0.026+/-0.020, 0.043+/-0.025 and 0.14+/-0.06, respectively (P<0.001 for 8 microM vs 38 microM). In MI livers (n=5), DHA bioavailability at 8 and 38 microM was 0.18+/-0.07 and 0.40+/-0.08, respectively (P=0.002). Bioavailability was higher in the MI group than in controls (P=0.01 at 8 microM and P<0.001 at 38 microM). DHA-glucuronide was the sole biliary metabolite. 4. Hepatic microsomal studies of DHA-glucuronide formation showed a significantly lower Vmax but no significant change in Km, in MI compared to control livers (n=6). Intrinsic metabolic clearance (Vmax/Km) was higher in control than in MI livers (5.2+/-1.3 and 2.5+/-1.4 microl min(-1) mg(-1), respectively; P=0.006). 5. These studies demonstrate that DHA has a high, concentration-dependent hepatic extraction ratio that is reduced by 20-30% in the P. berghei rodent malaria model. The impaired hepatic clearance of DHA in MI is attributable to a reduction in intrinsic metabolic clearance.