Recent investigations have shown that human skin permeabilities may not always follow a Gaussian-normal distribution as usually assumed. Permeability coefficients (K-p) for the model hydrophilic permeant, 5-fluorouraciI (5-FU), and the model lipophilic permeant, estradiol (ES), were demonstrated to follow more closely log-normal distributions. The present study has evaluated the effect of terpene penetration enhancer treatment on the distribution of 5-FU and ES permeabilities in human skin. Control permeability coefficients (K-p), enhanced permeability coefficients (K-pe) and enhancement ratios (ER) were tested for both Gaussian-normality and log-normality. K-pe and ER data, obtained from 5-FU experiments employing 22 different terpene enhancer formulations and one vehicle control, were pooled using standardised values, to obtain 188 replicate measurements for statistical analysis. Similarly, K-pe and ER data, collected from ES experiments using 10 different terpene treatments and two vehicle controls, were pooled to obtain 69 replicate measurements. Statistical analysis revealed that K-p, K-pe and ER data for 5-FU tended to follow a log-normal distribution. In contrast, the K-p, K-pe and ER data for ES followed, more closely, a Gaussian-normal distribution. Furthermore, the variability of the ES data was smaller in comparison with 5-FU data. It is possible that differences in data distribution may be linked to the physico-chemical properties of 5-FU and ES. The results of this study suggest that, when control permeabilities follow a log-normal distribution, as was the situation for 5-FU, it is likely that enhancer treatment will produce permeabilities and enhancement factors which are also log-normally distributed. In these instances, all data are most accurately represented by geometric means and geometric standard errors.
