A MATHEMATICAL-MODEL OF DIURNAL-VARIATIONS IN HUMAN PLASMA MELATONIN LEVELS

Studies in animals and humans suggest that the diurnal pattern in plas ma melatonin levels is due to the hormone's rates of synthesis, circul atory infusion and clearance, circadian control of synthesis onset and offset, environmental lighting conditions, and error in the melatonin immunoassay. A two-dimensional linear differential equation model of the hormone is formulated and is used to analyze plasma melatonin leve ls in 18 normal healthy male subjects during a constant routine. Recen tly developed Bayesian statistical procedures are used to incorporate correctly the magnitude of the immunoassay error into the analysis. Th e estimated parameters [median (range)] were clearance half-life of 23 .67 (14.79-59.93) min, synthesis onset time of 2206 (1940-0029), synth esis offset time of 0621 (0246-0817), and maximum N-acetyltransferase activity of 7.17 (2.34-17.93) pmol . l(-1). min(-1). All were in good agreement with values from previous reports. The difference between sy nthesis offset time and the phase of the core temperature minimum was 1 h 15 min (-4 h 38 min-2 h 43 min). The correlation between synthesis onset and the dim light melatonin onset was 0.93. Our model provides a more physiologically plausible estimate of the melatonin synthesis o nset time than that given by the dim light melatonin onset and the fir st reliable means of estimating the phase of synthesis offset. Our ana lysis shows that the circadian and pharmacokinetics parameters of mela tonin can be reliably estimated from a single model.