Common Genetic Variants Modulate the Electrocardiographic Tpeak-to-Tend Interval

Abstract

Sudden cardiac death is responsible for half of all deaths from cardiovascular disease. The analysis of the electrophysiological substrate for arrhythmias is crucial for optimal risk stratification. A prolonged T-peak-to-Tend (Tpe) interval on the electrocardiogram is an independent predictor of increased arrhythmic risk, and Tpe changes with heart rate are even stronger predictors. However, our understanding of the electrophysiological mechanisms supporting these risk factors is limited. We conducted genome-wide association studies (GWASs) for resting Tpe and Tpe response to exercise and recovery in ∼30,000 individuals, followed by replication in independent samples (∼42,000 for resting Tpe and ∼22,000 for Tpe response to exercise and recovery), all from UK Biobank. Fifteen and one single-nucleotide variants for resting Tpe and Tpe response to exercise, respectively, were formally replicated. In a full dataset GWAS, 13 further loci for resting Tpe, 1 for Tpe response to exercise and 1 for Tpe response to exercise were genome-wide significant (p ≤ 5 × 10^-8). Sex-specific analyses indicated seven additional loci. In total, we identify 32 loci for resting Tpe, 3 for Tpe response to exercise and 3 for Tpe response to recovery modulating ventricular repolarization, as well as cardiac conduction and contraction. Our findings shed light on the genetic basis of resting Tpe and Tpe response to exercise and recovery, unveiling plausible candidate genes and biological mechanisms underlying ventricular excitability.

Keywords: T-peak-to-T-end interval; cardiac conduction and contraction; cardiovascular dynamics; electrocardiogram; genetic risk score; genetics; genome-wide association study; sudden cadiac death; ventricular arrhythmias; ventricular repolarization.

Figure 1

Assessment of Tpe Indices in the EST-UKB Cohort (Top) Illustration of the RR profile during the exercise stress test. (Bottom) Three averaged heartbeats are derived at rest (black filled circle), peak exercise (red filled circle), and full recovery (blue filled circle), respectively. Resting, peak exercise, and recovery Tpe intervals were derived as the temporal differences between the corresponding T-wave offset and T-wave peak timing locations. Tpe dynamics during exercise was derived by quantifying the difference between the Tpe intervals at rest (black T-wave) and at peak exercise (red T-wave), normalized by the RR change during this interval. Similarly, Tpe dynamics during recovery was derived by quantifying the difference between the Tpe intervals at peak exercise (red T-wave) and full recovery (blue T-wave), normalized by the subsequent RR change.

Figure 2

Density Plots of Tpe Phenotypes Resting Tpe (A), Tpe dynamics during exercise (B), and Tpe dynamics during recovery (C). The blue curves indicate a normal distribution using the mean and standard deviation from each distribution.

Figure 3

Analytical 5-Stage Approach Flowchart Tpe, T-peak-to-Tend interval; SNV, single-nucleotide variant. Additional information can be found in Material and Methods.

Figure 4

Overlap of Resting Tpe, Tpe Response to Exercise, and Tpe Response to Recovery Loci with Other Electrocardiogram Traits SNVs at loci with a known genome-wide significant association (from PhenoScanner or GWAS catalog) with other ECG traits are grouped accordingly. The locus names indicate the nearest coding genes. The KCNJ2 locus was shared between resting Tpe and Tpe dynamics during exercise. There was no loci overlap between Tpe response to recovery and resting Tpe or Tpe response to exercise. There was a substantial number of loci for resting Tpe that did not overlap with other ECG traits. Underlined loci are loci that have previously been associated with other ECG markers but the reported variant was not in high LD (r² < 0.8) with our lead variant, so potentially independent signals at those loci. Bold loci are loci that have not been associated with other ECG marker. †Indicates sex-specific loci. ^∗Indicates previously associated with Tpe interval in other studies (PMID: 20215044 and 22342860).