Sensitivity Analyses for PK-QTc Modeling: Nonlinearity and Hysteresis

• Activity Number: 67
• Type: Topic Contributed
• Date/Time: Sunday, August 4, 2013 : 4:00 PM to 5:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Matthew Hutmacher*+
• Companies: Ann Arbor Pharmacometrics Group
• Keywords: Pharmacokinetics ; QT ; mixed-effects models ; population modeling

Abstract

Using pharmacokinetic (PK) or concentration data to predict cardiac risk of drug-induced QT prolongation, or heart-rate corrected QT (QTc), has increased. PK-QTc modeling can characterize prolongation in underpowered studies such as trials of patients with parallel arms, and it can provide clinically useful estimates thereof, as opposed to a general, limited hypothesis test based on the E14 Guidance-based intersection-union test (IUT). It has clinical relevance in that prolongation predictions can be made for PK changes due to, for example, amplifying drug-drug interactions or renal insufficiency. However, PK-QTc modeling has been criticized because the analyses fail sometimes to be rigorous. Statisticians have also been concerned with potentially flawed estimates due to hysteresis, a lag between the PK and its effect, or nonlinearity. Evaluation of hysteresis usually requires a PK model, predictions of which can be suspect for PK-QTc. To these ends, easy to use profile-like methods that do not use explicit PK models are proposed for checking the sensitivities of linear mixed-effects model outcomes to hysteresis and nonlinear concentration-effects.