In isolated hepatocytes the availability of intracellular glucose appears to be a key factor controlling the rate of xenobiotic glucuronidation during hypoxia. This study in the isolated perfused rat liver examines the effect of both a 24-hr fast and removal of glucose (8 mM) from liver perfusate on the elimination of bolus doses of harmol (20 mumol) under normoxic and hypoxic conditions. In the preparations used in these experiments, harmol glucuronide is the major metabolite (greater than 80%) with the remainder being sulphate. During normal oxygenation, in the livers from fed rats, harmol was rapidly eliminated (t1/2 = 4.2 +/- 0.4 min; mean +/- SD, N = 4). Fasting led to a small reduction in harmol elimination rate (t1/2 = 5.6 +/- 0.4 min; P less than 0.025) while removal of glucose from perfusate made no difference in either fed or fasted preparations. In the same livers, a second bolus dose of harmol was given during hypoxia. This produced a modest decline in harmol elimination in fed rats (t1/2 = 7.1 +/- 2.0 min; P less than 0.05). However, in fasted rats there was a striking reduction in harmol elimination (t1/2 = 109.8 +/- 54.0 min; P less than 0.025). The removal of glucose from perfusate made no significant difference to these results (t1/2 = 253 +/- 209 min in fasted preparations, P greater than 0.1). In all preparations, reoxygenation resulted in a rapid recovery of drug elimination. We conclude that nutritional state is important in determining the impact of hypoxia on harmol elimination by the liver. This study suggests that clinically significant reductions in xenobiotic glucuronidation are most likely to occur in poorly nourished or fasted subjects who became hypoxaemic.