Statistical testing of drug accumulation in skin tissues by linear regression versus contents of stratum corneum lipids

Abstract

This investigation is a contribution to standardization in in vitro drug penetration measurements using excised human skin and to statistical treatment of the observations. The wide variations observed in measurements of drug accumulation in and drug permeation through the stratum corneum are caused not only by analytical errors but also by the variability of the horny layer lipid composition. The last-mentioned systematic influence can be compensated for by stepwise (multiple) linear regression using the contents of the main lipid classes as independent variables. In consequence, the S.E. of estimate given by the regression calculation is lower than the S.E. of the means of the observations. Significant differences in drug quantities accumulated in skin tissues (stratum corneum and dermis) are sensitively detected by Chow's F-test of structural change. Accumulation data of flufenamic acid and hydrocortisone penetrated from different bases are given as examples. The calculation mode is exemplarily explained and discussed. The results of the test for structural change, two-independent-groups t-test and paired-samples t-test are compared. The F-test of structural change proves to be a helpful statistical method suitable to the assessment of biopharmaceutical quality parameters and to measurements using biological materials.