Van de Sandt, J., N. Cnubben, G. Elliott, W. Meling, A. De Bie, A. Gilde, W. Leeman, E. Geerlings, J. Kruse, and B. Hakkert. ATLA 1999. 27: 97.

In order to study the predictive value of in vitro percutaneous penetration measurements. It is important to compare in vitro and in vivo data obtained under identical experimental conditions. Therefore, in the comparative studies presented here, experimental conditions were rigidly standardized with respect to dose, vehicle (60% aqueous ethanol) and exposure duration (4 hours). Two pesticides with different physicochemical characteristics were used as test compounds: 2-isopropoxyphenyl N-methyl carbamate (propoxur; log Po/w 1.56, MW 209.2, dose 150 ug/cm2); and ortho-phenylphenol (OPP, log Po/w 3.28, MW 170.2, dose 120 ug/cm2). In vivo studies were performed in Wistar rats and human volunteers, while in vitro experiments were carried out using viable skin membranes containing part of the dermis (rat and human), non-viable epidermal membranes (rat and human) and the perfused pig ear model. In vitro-in vivo comparisons were based on the potentially available dose (applied minus dislodged dose after exposure for 4 hours), systemically available amount after 24 hours (amount in receptor fluid (in vitro) or in urine (in vivo), lag time and maximal flux. In vivo absorption kinetics of propoxur differed considerably between rat and human with respect to lag time, maximal flux and systemically available amount (20.8% and 5.6% after 24 hours, respectively). Based on the potentially absorbed dose, all in vitro models closely resembled the human in vivo situation. However, on the basis of the maximal flux and systemically available amount, the in vitro models overestimated human in vivo penetration by a factor of 2 to 3 (viable skin membranes), 2 to 10 (perfused pig ear) and 5 to 26 (epidermal membranes). In vivo percutaneous penetration of OPP in rats was approximately 1.5 times higher than in humans (urinary excretion after 24 hours of 42.9% and 29.0%, resp). Again, all in vitro models closely resembled the human in vivo situation, based on the potentially absorbed dose. Penetration of OPP through epidermal membranes and the perfused pig ear, correlated well with the human in vivo situation on the basis of maximal flux, but was overestimated on the basis of the systemically available amount. Viable skin membranes clearly underestimated human in vivo penetration at early time-points. Interestingly, this difference disappeared at later time-points, indicating that the dermis acted as a reservoir leading to a delayed release of OPP into the receptor fluid. In conclusion, percutaneous penetration of propoxur and OPP differed from each other with respect to both the in vivo rat –in vivo human relationship and the in vitro-in vivo relationship. For both test substances, all in vitro models closely predicted the human in vivo situation on the basis of the potentially absorbed dose. In contrast, large differences were observed between the in vitro models with respect to the amount of test substance actually reaching the receptor fluid.