Session P71.5

APART: A Novel Method for Poincaré Plot Shape Quantification Demonstrates Cardiac Tissue Repolarization Inhomogeneities Induced by Drugs

S Mensing*, JT Limberis, GA Gintant, A Safer

Abbott GmbH & Co. KG
Ludwigshafen am Rhein, Germany

Preclinical safety evaluation of drug compounds includes the assessment of drug induced effects on repolarization representing one part of the integrated risk assessment of cardiac safety (ICH S7B). The use of variable pacing protocols to mimic sinus arrhythmia generates Poincaré Plots of the repolarization duration (APD90(i) vs APD90 (i+1)) which exhibit a cyclic structure. In order to characterize and quantify the specific shapes of the plots, the Algorithm for Pre-/Clinical Analysis of Repolarization Trajectories (APART) has been developed. It estimates the medial trajectory and its belief limits thus allowing a statistical comparison of Poincaré Plots for tissues exposed to different compounds and conditions. Comparable features are: (1) median location of APD trajectory (=length of APD), (2) size of APD trajectory (=variability featuring electro-instability) and (3) shape of APD trajectory (a second measure of electro-instability). A proof of concept study (N=3) has been conducted in order to investigate the repolarization changes induced by the reference agent Moxifloxacin (3 and 30µg/ml). A pacing protocol derived from a resting conscious telemetrized dog was applied. Due to prominent RR variations during sinus arrhythmia and the correlation of basic cycle length and repolarization duration, the Poincaré Plot of APD90 exhibits the cyclic structure for which we provide a novel method to quantify. The evaluation of shape changes in the Poincaré Plot demonstrates differences in drug responses by two different cardiac tissues. The median curve perimeter for the papillary muscle changed to 93% (SD: 2) and 105% (SD: 19) under 3µg/ml and 30µg/ml drug relative to control. The perimeter of the medial curve for the Purkinje fiber trajectory increased to 114% (SD: 5) and 182% (SD: 16) under 3µg/ml and 30µg/ml drug relative to the perimeter under control conditions. Shapes of trajectories were not influenced. The use of a variable pacing protocol in combination with the APART method allows the quantification of three electro-instability features influenced by drug treatment.

(Abstract Control Number: 204)