Development and evaluation of an in vitro method for prediction of human drug absorption

The aim of the present work was to assess the actual suitability and general applicability of a new in vitro permeation method based on an appositely developed artificial membrane to evaluate and predict drug absorption potential. The proposed method was employed to evaluate the apparent permeability of a set of 21 structurally diverse drugs having very different solubility and permeability properties, covering the whole range of fraction absorbed in humans (Fa from 13 to 100%); 13 of the drugs in this study were part of the list suggested by FDA for validation of in vitro permeation methods. An excellent linear correlation (R2 = 0.957) was obtained between artificial membrane apparent permeability and human absorption data in the whole range of Fa values examined (including all the drugs belonging to the above FDA list), indicating the good predictive ability of the proposed method not only for highly absorbed hydrophobic compounds but also, differently from other in vitro permeation methods, for poorly or middling permeable drugs. The predictive ability of the new method was greater than those obtained for the same set of drugs with PAMPA and Caco-2 permeability literature data, probably due to the poor sensitivity of these methods towards hydrophilic drugs. The better performance of our artificial membrane was attributed to the hydrophilic nature of the support that, differently from the commonly used hydrophobic supports, offers less resistance to permeation of hydrophilic compounds. A comparison of permeation data of theophylline, ketoprofen, aciclovir and furosemide (selected, respectively, as models of I–IV BCS classes) obtained using a Caco-2 cell based dynamic method and the developed artificial membrane and the corresponding Fa values in humans further confirmed the suitability of the proposed permeation method as predictor of the oral absorption of passively absorbed drugs.