Smith, J.S., O.T. Macina, N. B. Sussman, and M.H. Karol. A Structure-Activity Relationship (SAR) Model For Esters that Cause Human Skin Irritation. The Toxicologist 1999. 48(1S): 77. Abstract #361.

A structure-activity relationship (SAR) model has been developed to discriminate active from inactive esters on the basis of their ability to cause skin irritation in humans. The model is based on patch test data gathered from a critical review of the literature. Esters were selected from a broader database of human irritants and represented a large subset of homologous chemicals. Nineteen physicochemical parameters that represented transport, steric and electronic properties, were calculated for each chemical. Best subset regression analysis identified models for further analysis. Regression analyses were performed to identify all significant models (p<0.05) that had variables that were also significant (p < 0.05). These models were evaluated using linear discriminant analysis to determine if the irritant esters could be discriminated from non-irritant esters. Cross validation was employed to estimate the sensitivity and specificity of the model. The model with the greatest discriminating ability contained the parameters density. Hansen dispersion, Hansen hydrogen bonding parameters, and higher Hansen dispersion parameters. These characteristics indicate that irritant esters would tend to be less water soluble and less likely to be involved in hydrogen bonding to protein components of the skin. Thus, they are more likely to penetrate the epidermis. Absent from the model are the electronic parameters to indicate a relationship between activity and reactivity. The stability of the model was tested using random subsets of the database. The above parameters were observed with consistency over the random subsets indicating stability with regard to parameters in subsets of the data. The results imply that the model should be predictive for the irritant activity of esters toward human skin.