Quantitative structure-pharmacokinetic/pharmacodynamic relationships of corticosteroids in man.

The purpose of this study was to develop quantitative structure-activity/pharmacokinetic relationships (QSAR/QSPKR) for 11 selected corticosteroids in man. Multiple linear regression analysis with an automatic forward step-wise inclusion algorithm was used to construct QSAR/QSPKR models from molecular and submolecular descriptors that were generated using the SYBYL and KowWin computer programs. The final equations describing steroid relative receptor affinity, systemic clearance, volume of distribution, fraction unbound in plasma, and percent of oral absorption, all showed significant correlations (R(2) range 0.841 to 0.977). These relationships were crossvalidated using the leave-one-out method, and appeared to have good predictive performance (Q(2) range 0.715 to 0.912). In addition, a general method for integrating QSAR/QSPKR data to predict the time course of pharmacologic effects is presented. This approach, termed quantitative structure-pharmacodynamic relationships modeling, was successfully applied to predict the rapid cortisol suppressive effects of triamcinolone acetonide after a 2-mg intravenous bolus dose in healthy volunteers.