Firstly, the ability for binary differentiation of human skin samples
was evaluated for the three standard tests TEER, TEWL and TWF. Therefore, we differentiated valid and invalid excised or reconstructed human skin samples according to the standard limit values for human skin set 1 kΩ, 10 g m−2 h−1 and 2.5 ∗ 10−3 cm h−1 and for TEER, TEWL Alectinib clinical trial and TWF, respectively. In addition one further limit value was used for each test. Based on the outcome, these were more liberal for TEWL (13 g m−2 h−1) and TWF (4.5 ∗ 10−3 cm h−1), yet more strict for TEER (2 kΩ). The minimum (min), maximum (max) and mean absorption results (maxKp and AD) were calculated separately for the defined valid and invalid groups. Furthermore we plotted the single cell results for the defined valid and invalid skin samples. Next, the ability of all five integrity tests (TEER, TEWL, TWF, ISTD and BLUE) to detect and explain minor differences in barrier function was investigated by correlation analyses. For this task, rat skin was included, basically, to make use of the in theory lower donor variability of rat skin for the special investigation in which rat skin was systematically damaged to various grades. For the correlation analyses we grouped all experiments using the same test compound (caffeine, testosterone, MCPA or MCPA-EHE) and barrier system (human, rat or reconstructed human skin) together. Groups with at least 10 single data points
were used for linear regression analysis of integrity test results (independent variable x) against absorption results (AD and maxKp, dependent variable y).
All data points were included independent of valid or invalid selleck products classification. Slopes and correlation coefficients (R2) were reported for evaluation. Min, max and mean values were calculated for each integrity test, but only R2 from correlations with the correct algebraic sign were used. To assess Rapamycin price the variability of the methods and the effect of the human donor, overall, inter- and intra-donor variabilities were calculated for the different methods. Overall variability is given as the variation coefficient (CV, often referred to as the relative standard deviation (SD)) of all skin samples used, inter-donor variability is given as CV calculated with the mean values for each donor and intra-donor variability which corresponds to the method variability is given as the pooled, average, CV for each donor weighted by the number of replicates. If from one human donor both, full-thickness and dermatomed skin, was used, the underlying means and variabilities were calculated separately. For ISTD and the general in vitro dermal absorption method, only the pooled CV could be calculated due to the various kinds of ISTDs and test compounds used. Underlying means were calculated separately for each ISTD or test compound. Since energy spectra of 14C and 3H overlap, a LSC method was used that compensates for the influence of the other isotope.