The pharmacokinetics and safety of a brief i.v. infusion of l-carnitine 0, 20, 40 and 60 mg/kg have been investigated in 10 healthy subjects. The diurnal intraindividual variability of plasma carnitine was small (C. V. = 3.0, 3.9 and 3.9%, respectively), and the total 24 h excretion in urine was also small and relatively constant: 4.6, 21.5 and 13.0 mg/day in the controls vs 4.6, 20.2 and 6.0 mg/day during treatment in the three subjects to whom saline alone was administered according to a single-blind design. Therefore, the pre-dose level of carnitine was subtracted from the level after dosing for the pharmacokinetic analysis. Plasma carnitine fitted well to a three-compartment open model, with Vc of 0.11-0.20 l/kg and a t1/2 gamma of 10-23 h. The urine recovery in 24 h was 77.2-95.4%. There were no objective or subjective side-effects attributable to carnitine, so its i.v. infusion is considered to be safe.