In this study, an empirical model of the quantitative structure-permeability relationship (QSPR) of the transdermal delivery of non-steroidal anti-inflammatory drugs (NSAIDs) was constructed in an attempt to predict the permeability coefficients ( k P). Thirteen model NSAIDs were selected, and their in vitro permeation through the full skin of nude mice was examined. The biological parameters of transepidermal water loss (TEWL), hydration content (HD), lipid content (SB), resonance running time (RVM), and elasticity (EL) were measured. The permeability coefficients so obtained were grouped into three datasets of all model drugs and those drugs with clog P or log K o/w values of > 2 and < 2; these datasets were regressed with respect to the physical characters of molecular weight (MW) and polarity factor ( clog P or log K o/w) or the solubility parameter ( δ) of the model drugs rationally chosen to replace the polarity factor with or without taking into consideration the biological parameters of the skin. Results demonstrated that δ could be greatly improved compared to clog P and log K o/w in the regression with an adjusted R 2 of > 0.90 using the dataset of those drugs with clog P or log K o/w values of < 2, regardless of whether or not biological parameters were taken into consideration. This indicates that δ might rationally be a more-appropriate drug parameter for predicting the skin permeability of NSAIDs for transdermal delivery. A plot of observed k P versus predicted k P values by this simple empirical model of QSPR was validated to demonstrate the predictive capability of k P for transdermal delivery. In conclusion, an empirical model of QSPR to predict k P based on the hydrophilicity of the model drugs was statistically improved with δ and by taking the biological parameters of the skin into consideration.
Read full abstract