BACKGROUND: Loss of circadian rhythms and reduced concentrations of endogenous melatonin are common in critically ill patients. After exogenous administration, supra-physiological concentrations in serum are only ephemeral, which may explain the absence of significant therapeutic effect on sleep. The aim of this study is to describe the pharmacokinetics of enteral melatonin in critically ill patients administered in a novel regimen aiming to simulate endogenous release. METHODS: Thirteen patients in the recovery phase of critical illness were randomised to receive enteral melatonin or placebo. In the melatonin group, a total of 6 mg was administered as solution through their feeding tube, commencing with a 3 mg loading dose at 9 pm and six subsequent 0.5 mg doses hourly. The placebo was administered using a similar regimen. Serial blood samples were taken and measured using a validated chromatographic method. The concentration-time data for serum melatonin concentrations were described using non-linear mixed-effects modelling. RESULTS: The observed concentrations in the melatonin patients were significantly higher than that observed in the placebo patients. The concentrations in the patients administered melatonin were also higher than endogenous melatonin concentrations previously reported in non-critically ill patients. The patients administered melatonin had a mean clearance, volume of distribution and absorption rate constant of melatonin was 55.2 L/h, 767 L and 0.76 h-1, respectively. CONCLUSIONS: Exogenous administration of melatonin with a loading dose of 3 mg followed by an hourly dose of 0.5 mg demonstrates good oral bioavailability and results in supra-physiological and sustained concentrations of serum melatonin during 12 h overnight.