For MCPA the valid results increased clearly when applying the higher limit value and the range of valid data even included the range of invalid in full. This

effect is mainly due to the inclusion of absorption results obtained with six reconstructed human skin samples which were obviously higher, but based on TEWL cut-off limit 13 g m−2 h−1 classified NU7441 price as valid. The very slow penetrating test compound 14C-MCPA-2EHE showed no clear difference of absorption values in valid and invalid skin samples. This was observed with all integrity tests (Table 4, Table 5 and Table 6). Mean, min and max values did not differ significantly for the two different limit values of TWF (Table 6). However, the stricter limit value for TEER (2 kΩ) led to a different distribution (Table 4). Only 2 of 90 skin samples were classified as invalid with 1 kΩ as the limit, but 28 with 2 kΩ. Applying the limit value 2 kΩ, the majority of the reconstructed human skin samples with higher

absorption results for the test compounds were classified as invalid (23 out of 30). In contrast, five excised human skin samples were classified as invalid despite of absorption data in reasonable ranges. Analog to TEWL, differentiation with TEER (limit: 2 kΩ) and TWF resulted in obvious higher absorption means for invalid find more skin samples than for valid skin samples as well as in significant Progesterone overlapping of results. In a second step linear regression analyses for the absorption values (AD, maxKp, dependent variable y) and integrity test results (independent variable x) were used to check whether integrity tests TEER, TEWL, TWF, ISTD and BLUE are able to display minor barrier

differences between skin samples continuously. Besides human skin, rat skin was included in these analyses. Table 7 shows mean, min and max values of slopes and R2 derived from analysis for the different experimental groups. One group covers experiments using one defined combination of test compound (testosterone, caffeine, MCPA or MCPA-2EHE) and skin preparation (excised human skin, reconstructed human skin or excised rat skin). The correlations varied over a wide range for all five methods, four test compounds and three skin preparation types. The best correlations in average (R2: 0.484) and maximal (R2: 0.911) were achieved with the ISTD. Partially good correlations were observed for TEWL: the maximal R2 of 0.790 was achieved with test compound 14C-testosterone applied to reconstructed human skin. Even inverse correlations were occasionally obtained with TEWL, TEER, TWF and BLUE but not for ISTD. The dataset of the special investigation comprising all experiments with 14C-MCPA applied to undamaged and gradually damaged rat skin covers a wide range of absorption (AD 6–100%) and absorption rates (Marzulli-Class: slow to very fast) ( Marzulli and Brown, 1969).