Beat-to-beat variability of ventricular repolarization reveals sex-specific instability and proarrhythmic risk

Beat-to-beat QT interval variability (QTV) is a well-established marker of increased vulnerability to ventricular arrhythmias; however the underlying electrophysiological mechanisms remain poorly understood. In this study, we employed sex-specific, physiologically detailed computational models of human ventricular myocytes to investigate the role of dynamical repolarization instability under baseline and pharmacologically perturbed conditions. Action potential duration (APD) variability was quantified as a cellular-level surrogate for QTV and evaluated in relation to restitution kinetics, drug risk classification and biological sex. Female models exhibited significantly greater APD variability than male models, consistent with clinical observations of sex differences in QTV. Exposure to high-risk torsadogenic compounds further amplified APD variability, whereas drugs with low proarrhythmic potential produced only modest effects. Intermediate-risk agents elicited APD variability patterns aligned with previously reported machine learning-based predictions of torsades de pointes (TdP) risk. Notably increased APD variability strongly correlated with steeper APD restitution slopes, reflecting enhanced dynamical instability. Importantly elevated variability was observed even in the absence of early afterdepolarizations (EADs), underscoring its potential as a sensitive, early marker of proarrhythmic susceptibility. These findings provide mechanistic evidence linking dynamical instability to QTV and establish sex as a critical modulator of arrhythmogenic drug response. KEY POINTS: Beat-to-beat QT interval variability (QTV) is an established clinical marker of arrhythmic risk, but the mechanistic link between QTV and arrhythmia remains poorly understood. Using sex-specific, physiologically detailed human ventricular myocyte models, we show that QTV arises from dynamical repolarization instability, with action potential duration (APD) variability closely tracking the slope of the APD restitution curve. Female models exhibited greater baseline APD variability than male models, in agreement with findings in human ventricular cardiomyocyte experiments, and exposure to high-risk torsadogenic drugs further amplified this variability in a concentration-dependent manner, even in the absence of early afterdepolarizations (EADs). APD variability emerged as an indicator of dynamical instability and was predictive of arrhythmogenic susceptibility across a range of pacing rates and pharmacological conditions. Our findings support its integration into safety pharmacology frameworks to improve current arrhythmia risk assessment.