QTc Model Misspecification May Cause Spurious Carryover Effects

• Activity Number: 468
• Type: Contributed
• Date/Time: Wednesday, August 4, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Xiaodong Li*+ and Andrea Maes and Deborah Panebianco
• Companies: Merck & Co., Inc. and Novartis Pharmaceuticals Corporation and Merck & Co., Inc.
• Keywords: Crossover ; Type I Error ; Mixed Effects Model

Abstract

A QT/QTc study is conducted to demonstrate that a drug does not prolong the QT interval. The most frequently employed QTc study design is a crossover with placebo, active control and therapeutic and supratherapeutic doses of the test drug. We consider such a crossover with one predose baseline and multiple postdose measurements captured throughout the dosing interval in each treatment period. A repeated measures mixed model with fixed effects for period, treatment, time, treatment-by-time interaction, and random subject effect has been used to analyze QTc change from baseline. We show this model is not appropriate for estimating the crossover's effects. Further, simulations have demonstrated a Type I error rate >60% for carryover when such a term is added to this model. We propose an alternative mean model to appropriately estimate the effects and provide a valid test for carryover.