Indexed on: 09 May '08Published on: 09 May '08Published in: Journal of Pharmaceutical Sciences
A turnover model in combination with an empirical surge function based on a chromatographic peak detection algorithm was used to model the population pharmacokinetics of melatonin in 32 healthy males (aged 20-30 years, weight 62-85 kg) who had 15 blood samples drawn (11:00 to 08:00) on three occasions, separated by 2-4 weeks. Serum melatonin concentrations were measured by RIA. A pharmacokinetic model with a surge function was fitted to the data; dA/dt = K(IN) [1 + (AMP/(((t - T(0))/WID)(N) + 1))] - K(OUT) A, using NONMEM. Estimates were sought for the typical population parameter values and the intersubject variability (CV%) of; baseline amount of melatonin (A(B)), elimination rate constant (K(OUT)), peak amplitude (AMP), peak width (WID), acrophase (T(0)). The model stability was validated using a nonparametric bootstrap (100 samples with replacement). Population typical values (CV%) were: A(B); 103 ng (103%), K(OUT); 0.255 h(-1) (57.8%), AMP; 34.9 (108%), WID; 2.3 h (20.9%), T(0); 14.8 h (4.4%). Interoccasion variability (CV%) for; A(B) (12.3%), K(OUT) (29.5%) and AMP (9.1%) was much less than the corresponding intersubject variability. This approach has potential for application in clinical studies designed to characterize abnormal melatonin rhythms.