Optimization of a vehicle mixture for the transdermal delivery of melatonin using artificial neural networks and response surface method

The objective of this study was to optimize a suitable vehicle composition, using response surface method (RSM) and artificial neural networks (ANN), for the transdermal delivery of melatonin (MT). MT is a hormone produced by the pineal gland that influences mammalian sleep and reproductive patterns. A successful treatment for sleep disorders can be developed if MT is delivered with a rate at which it is produced in the body (endogenous rhythm). Prominent hepato-gastrointestinal first-pass metabolism and short half-life of MT in the body, limits the ability of oral route to mimic the endogenous rhythm. Transdermal route is supposed to avoid first-pass metabolism, and maintain steady-state plasma MT concentrations for a required period of time. However, MT by itself can not pass through the dense lipophilic matrix of stratum corneum. Hence solvents like water (W), ethanol (E), propylene glycol (P), their binary and ternary mixtures were employed to increase MT flux and reduce lag time. Special quartic model (RSM) and δ back-propagation algorithm (ANN) were employed as prediction tools. W:E:P (20:60:20)>W:E (40:60)>W:P (50:50) were predicted as the effective vehicles. W:E:P was considered as the best vehicle, both in terms of flux (12.75 μg/cm² per h) and lag time (5 h). RSM and ANN prediction of the best mixtures coincided very well. The ability of these tools to summarize various responses (solubility, flux, and lag time) with respect to vehicle composition enabled us to study the inter-relativity between the responses. Copyright (C) 1999 Elsevier Science B.V.