Background Reduced heart rate (HR) changes during exercise and recovery post-exercise are strong predictors of cardiovascular (CV) death, suggesting that abnormalities in autonomic balance may precede manifestations of malignant CV events. Cardiac repolarisation is a critical component in modulating the risk of CV death. We therefore hypothesised that assessment of autonomic effects on cardiac repolarisation during exercise may help to improve CV risk prediction in the general population as it is more specific for cardiac ventricular pathophysiology compared to the autonomic effects on the sinus node.

Methods A total of 54,203 healthy individuals aged 40–69 years old without prior CV disease who had an exercise stress test from the UK Biobank study were included. HR corrected repolarisation time intervals RTc (an approximation of the QT interval) were measured from the ECG recording at rest (pre-exercise; RTc_rest), peak-exercise (RTc_ex), and 50s post-exercise recovery (RTc_rec) using a custom build algorithm. We also computed the difference between RTc_ex and RTc_rec (dRTc) as a marker of exercise associated repolarization. We finally performed a follow-up analysis to evaluate the prognostic value of these biomarkers. The endpoints studied were CV death or CV hospitalisation. Associations were tested with the Mann-Whitney and multivariate cox analysis. We then evaluated using multivariate Cox analysis whether individuals in the top 20% for were significantly more likely to suffer a CV event than those in the bottom 20%.

Results During a median follow-up time of 56 months, 1,460 (2.7%) individuals reached the endpoint. RTc_ex, RTc_rec, and dRTc were all significantly associated with the endpoint (p=4.1E-9, p<2.2E-16, and p=5.6E-5, respectively). RTc_ex, remained significantly associated with the endpoint after adjusting for age, diabetes, systolic blood pressure, body mass index, heart rate changes during exercise and recovery, and RTc_rest in the cox proportional hazards model (table 1) with a HR of 1.07 (confidence interval: 1.002–1.137. p=0.04).

Conclusion Assessment of repolarisation time during recovery improves prediction of cardiovascular death and hospitalisation in the general population independently from heart-rate changes. This work supports inclusion of clinical variables to improve personalised diagnostics and demonstrates the importance of evaluating repolarisation during exercise stress testing.

Conflict of Interest None