Assessment of drug permeability distributions in two different model skins

Past in vitro studies with human skin have indicated that drug permeability coefficient (K-p) distributions do not always follow a Gaussian-normal pattern. This has major statistical implications, exemplified by the fact that use of t-tests to evaluate significance is limited to normally distributed populations. Percutaneous absorption research often involves using animal or synthetic skins to simulate less readily available human skin. However, negligible work has been performed on assessing the permeability variabilities of these model membranes. This paper aims to fill this gap. To this end, four studies were undertaken representing two different drugs (caffeine and testosterone) with each drug penetrating through two different model skins (silicone membrane and pig skin). It was determined that in the silicone membrane studies, both compounds' K-p distributions could be fitted to a normal pattern. In contrast, in the pig skin studies, there were notable differences between each drug. While the testosterone K-p values could be fitted to a normal distribution, this was not possible with the caffeine K-p data, which could be fitted to a log-normal distribution. There is some evidence from the literature as well as physicochemical considerations that these outcomes may reflect general trends that are dependent upon both membrane and penetrant properties. (c) 2005 Elsevier B.V. All rights reserved.