where
E = Pharmacodynamic effect metric i.e. , AUEC ;
D = Duration of exposure (min) to the TC;
Emax = Maximum possible value for E ;
ED50 = Dose duration necessary to achieve 50% of the Emax response.
Although the FDA’s VCA guidance recommends the use of naive pooling or Non-Linear Mixed Effect modelling (NLME) to fit the pharmacodynamic response data (22,27,29,30), naive pooling does not take into consideration the inter-individual variability, which may not accurately represent the study population (29,30). Therefore, P-Pharm software which uses NLME modelling with the likelihood estimation (30,31) is a more appropriate method for the determination of population parameters such as ED50 and Emax (30).
Plots of individual a-scale readings vs the assessment time points over a period of 24 h were constructed and the areas under the effect curve (AUEC0-24 ) values for each subject at each dose duration for the respective TCPs were determined using the linear trapezoidal method from the baseline adjusted and untreated site corrected a-scale values (26,32). AUEC0-24 valuesvs the respective dose durations were used to estimate theEmax model that best described the data based on the minimum Akaike Information Criterion (AIC ).
A Student’s t -test was used to compare theEmax values of TC API solutions and TCPs. The TC API solutions and the TCPs were compared amongst themselves using a paired t -test.
The TC APIs were compared with the TCPs containing the corresponding corticosteroids using a one-way analysis of variance (ANOVA) to determine the presence/absence of significant differences in theEmax parameters between the TCPs and their corresponding TCs.