Modeling and Interpreting QTc Prolongation in Clinical Pharmacology Studies
- PDF / 13,436,441 Bytes
- 9 Pages / 648 x 864 pts Page_size
- 41 Downloads / 192 Views
Scott D. PUttOfSOB GlaxoSmithKlinc Pharmaceuticals, USA
B y r r Jomos Pfizer Pharmaceuticals, United Kingdom N6vlmO hiiff GlaxoSmithKline Pharmaceuticals, USA
Koy Words ECG; QTc; Mod&;
ICH Cwrrrpdnco Allrorr Scott pattnson, GlaxoSmithKlinc Pharmaceuticals, 1250 South CdlegcvillcRoad, Cdlegm*lle,PA 19426 (mail: [email protected]). The data andfindings of this article a n npromccedfrom Ref 6 with pmnission of CRC Pnss.
I N T E R V A L P R O L O N G A T I O N 137
Modeling and Interpreting QTc Prolongation in Clinical Pharmacology Studies
INTRODUCTION The safety of new drug products has received much attention in the public recently (1, 2). Even prior to the recent cardiac safety findings of the popular COX2 class of drug, regulators were intensifylng requirements around cardiac safety of new drugs, at least as relates to proanythmic potential measured by QTc prolongation. Indeed, through the International Conference on Harmonisation (ICH) process (3). regulators have been calling for a thorough or definitive QTc assessment study to be conducted prior to approval, preferably before phase I11 trials are initiated. A handful of these thorough QTc trials has been conducted and reported to date. However, industry sponsors and regulators alike have yet to fully review existing data alongside the current ICH El4 guidance (3) to help refine the statistical approaches for such trials currently suggested. In this article, we use two examples of studies conducted to develop a basic framework for deriving statistics to describe the results. The studies are complex in design and are typically conducted in 30-50 healthy volunteers under highly controlled conditions and include at least two dosing regimens of the new drug along with placebo and active control agent known to affect QTc.
We next review a statistical framework to formally test a series of hypotheses relating to the two doses of the new drug and in some instances to the active control as well. We also show how the current end point described in the guidance relates to well-established statistical tests and discuss the type 1 and type 2 properties of these methods.
DATA, MODEL, AND RESULTS To illustrate an approach to the modeling of QTc data, we consider some data from previous trials. Two crossowr data sets were selected for use as examples in this article to provide a range of example effect sizes in QTc prolongation relative to changes in electrocardiogram (ECG) sampling procedures and sample sizes. Normal healthy volunteers are generally (3)the population dosed in such studies as it is felt that QTc prolongation observed in that population does not pose a great risk, and findings are readily applicable to patient populations. Both studies presented were fully randomized crossover designs in normal healthy volunteers. In both cases, the objective of the trial was to detect changes in QTc induced by the study drug over and above those introduced by a control agent, and ECGs were manually overread by a qualified, blinded cardiologist. Fridericia's correc
Data Loading...
