A population Pharmacokinetic Turnover and Surge‐Function Model for Describing Melatonin Biological Rhythm in Healthy Male Subjects

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 = KIN [1 + (AMP/(((t − T0)/WID)N + 1))] − KOUTA, using NONMEM. Estimates were sought for the typical population parameter values and the intersubject variability (CV%) of; baseline amount of melatonin (AB), elimination rate constant (KOUT), peak amplitude (AMP), peak width (WID), acrophase (T0). The model stability was validated using a nonparametric bootstrap (100 samples with replacement). Population typical values (CV%) were: AB; 103 ng (103%), KOUT; 0.255 h−1 (57.8%), AMP; 34.9 (108%), WID; 2.3 h (20.9%), T0; 14.8 h (4.4%). Interoccasion variability (CV%) for; AB (12.3%), KOUT (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. © 2008 Wiley‐Liss, Inc. and the American Pharmacists Association J Pharm Sci 98:782-790, 2008