Population Modeling Approach for Human Cardiac Arrhythmia Risk Prediction

Cardiovascular disease is the number 1 killer in industrialized countries, with sudden cardiac death due to ventricular arrhythmias representing a major component. To reduce sudden cardiac death, accurate risk prediction and development of effective preventive treatments remain major challenges. In this study, we explored the possibility of using a population-based computational modeling approach to perform virtual clinical trials for antiarrhythmic drug discovery and drug safety testing.

We developed genetically diverse populations of 1-dimensional cardiac tissue models for both normal hearts and hearts with long QT syndromes (LQT1, LQT2, and LQT3) based on matching the models to the clinically measured distributions of corrected QT intervals for each condition.

Using a doubling of the L-type calcium current to mimic sympathetic stress, the population models exhibited a similar incidence of arrhythmias as observed in corresponding clinical studies for each condition. We demonstrated that the model populations (1) accurately predicted arrhythmia risk under normal and diseased conditions; (2) could be used to assess the effectiveness of a therapeutic strategy, namely shifting the steady-state inactivation curve of the L-type calcium current; and (3) accurately predicted the cardiotoxicity of a series of drugs when compared with their known clinical profiles.

The population-based modeling approach outlined here shows promise as a computational platform that can directly take advantage of data from human clinical studies to improve arrhythmia risk prediction, test antiarrhythmic therapies, and assess cardiotoxicity of drugs.