ECG T-Wave Morphologic Variations Predict Ventricular Arrhythmic Risk in Low- and Moderate-Risk Populations.
Julia RamírezAntti M KiviniemiStefan Van DuijvenbodenAndrew TinkerPier D LambiaseHeikki V HuikuriJuha S PerkiömäkiHeikki V HuikuriMichele OriniPatricia B MunroePublished in: Journal of the American Heart Association (2022)
Background Early identification of individuals at risk of sudden cardiac death (SCD) remains a major challenge. The ECG is a simple, common test, with potential for large-scale application. We developed and tested the predictive value of a novel index quantifying T-wave morphologic variations with respect to a normal reference (TMV), which only requires one beat and a single-lead ECG. Methods and Results We obtained reference T-wave morphologies from 23 962 participants in the UK Biobank study. With Cox models, we determined the association between TMV and life-threatening ventricular arrhythmia in an independent data set from UK Biobank study without a history of cardiovascular events (N=51 794; median follow-up of 122 months) and SCD in patients with coronary artery disease from ARTEMIS (N=1872; median follow-up of 60 months). In UK Biobank study, 220 (0.4%) individuals developed life-threatening ventricular arrhythmias. TMV was significantly associated with life-threatening ventricular arrhythmias (hazard ratio [HR] of 1.13 per SD increase [95% CI, 1.03-1.24]; P =0.009). In ARTEMIS, 34 (1.8%) individuals reached the primary end point. Patients with TMV ≥5 had an HR for SCD of 2.86 (95% CI, 1.40-5.84; P =0.004) with respect to those with TMV <5, independently from QRS duration, corrected QT interval, and left ventricular ejection fraction. TMV was not significantly associated with death from a cause other than SCD. Conclusions TMV identifies individuals at life-threatening ventricular arrhythmia and SCD risk using a single-beat single-lead ECG, enabling inexpensive, quick, and safe risk assessment in large populations.
Keyphrases
- left ventricular
- heart failure
- heart rate
- cardiovascular events
- risk assessment
- catheter ablation
- ejection fraction
- heart rate variability
- aortic stenosis
- coronary artery disease
- cardiovascular disease
- type diabetes
- blood pressure
- machine learning
- dna methylation
- heavy metals
- electronic health record
- big data
- deep learning
- hypertrophic cardiomyopathy
- human health
- left atrial
- artificial intelligence