THE PREDICTION OF SKIN PERMEABILITY BY USING PHYSICOCHEMICAL DATA

A database on physicochemical properties and skin permeation compiled by Health Canada was analysed by using linear regression analysis. The correlation between permeability coefficient (K-p) and the octanol-wa ter partition coefficient (K-ow) has been improved by grouping the com pounds according to their respective molar volumes. Linear regression analysis of the individual groups has demonstrated a positive correlat ion for the majority of the groups, with the compounds in the lowest m olar volume range (less than or equal to 75 Angstrom(3)) having the be st correlation (r(2) = 0.86), and the compounds in the highest molar v olume range (greater than or equal to 301 Angstrom(3)) being the least well-correlated (r(2) = 0.55). Due to the diversity of the chemicals used in this analysis, and the statistically significant correlations obtained, this model could permit the prediction of skin permeation of a wide variety of chemical compounds. Although of a simplistic nature , and not yet experimentally validated, this quantitative structure-ac tivity relationship may be useful for predicting human skin permeabili ty coefficients for compounds that fall within the constraints of this data set.