A Case Study of PK-PD Modeling and Exposure-Response Prediction for Count Data
Xuezhou Mao, Sanofi  *Hui Quan, sanofi  Lin Wang, Sanofi  Lynn Wei, Sanofi 

Keywords: plasma concentration, annualized relapse rate, negative binomial, recurrent event, risk ratio, prediction

Even with two doses of the experimental drug in a Phase III study, it may still be difficult to determine the final commercial dose. In such a scenario, with plasma concentration data collected in the Phase III study, a PK/PD modeling approach can be applied to predict treatment effects at different concentration levels. Through an established relationship between concentration and dose, the treatment effects of doses not studied in the Phase III studies can be predicted. Then the results can be applied to justify the final dose selection. In this presentation, a trial example with count data as the primary endpoint is used to illustrate the application of such a technique for dose selection. Several PK/PD models such as over-dispersion Poisson model, negative binomial model and other models for recurrent events are considered. The negative binomial model with the capability of within-treatment assessment and between-treatment comparison is preferable because of its simplicity and other properties.