Bayesian Approach to Assay Sensitivity Analysis of Thorough QT Trials

Abstract

One of the analyses recommended in ICH E14 Guidance (International Conference on Harmonisation, 2005) after the test drug is shown to be negative in QT interval prolongation after subjects treated with the test drug is an assay sensitivity analysis of a positive control drug with known effect on QT prolongation. The assay sensitivity is validated if the response profile is shown to be expected and the QT interval after administration of the positive control drug is shown to be at least 5 ms more than placebo. The negative result of the test treatment is validated if the prolongation of the positive control is verified among the study subjects. One of the most frequently used positive control drugs in thorough QT (TQT) trials is moxifloxacin. In order to improve the efficiency of the study and to reduce the number of subjects exposed to moxifloxacin, we explore the potential sample size reduction with a Bayesian approach to the assay sensitivity utilizing the data of historical TQT trials. We derived the distribution of moxifloxacin-induced QT prolongation based on 14 crossover trials and six parallel trials. The estimated distribution is used as a prior distribution to assess the posterior probability that the moxifloxacin-induced QT prolongation is larger than 5 ms. Sample size based on such Bayesian approach will be compared with the conventional frequentist approach for efficiency assessment.