Phenotypic but not genetically predicted heart rate variability associated with all-cause mortality

Abstract

Low heart rate variability (HRV) has been widely reported as a predictor for increased mortality. However, the molecular mechanisms are poorly understood. Therefore, this study aimed to identify novel genetic loci associated with HRV and assess the association of phenotypic HRV and genetically predicted HRV with mortality. In a GWAS of 46,075 European ancestry individuals from UK biobank, we identified 17 independent genome-wide significant genetic variants in 16 loci associated with HRV traits. Notably, eight of these loci (RNF220, GNB4, LINCR-002, KLHL3/HNRNPA0, CHRM2, KCNJ5, MED13L, and C160rf72) have not been reported previously. In a prospective phenotypic relationship between HRV and mortality during a median follow-up of seven years, individuals with lower HRV had higher risk of dying from any cause. Genetically predicted HRV, as determined by the genetic risk scores, was not associated with mortality. To the best of our knowledge, the findings provide novel biological insights into the mechanisms underlying HRV. These results also underline the role of the cardiac autonomic nervous system, as indexed by HRV, in predicting mortality.

Fig. 1. Mirrored Manhattan plots of the GWAS of HRV traits.

a RMSSD and RMSSDc and b SDNN and SDNNc. The red horizontal line represents the genome-wide significance threshold. Genes closest to the independent lead SNPs are indicated for the loci that were genome-wide significantly associated with the trait. Novel loci are highlighted in red. RMSSD root mean square of successive differences, RMSSDc hear rate-corrected root mean square of successive differences, SDNN SD of normal-to-normal intervals, SDNNc heart rate-corrected SD of normal-to-normal intervals.

Fig. 2. Venn-diagram of lead SNPs shared by the HRV traits.

RMSSD root mean square of successive differences, RMSSDc heart rate-corrected root mean square of successive differences, SDNN SD of normal-to-normal intervals, SDNNc heart rate-corrected SD of normal-to-normal intervals.

Fig. 3. Phenotypic associations of HRV traits with all-cause and cardiovascular mortality.

lnRMSSD log-tranformed root mean square of successive differences, lnRMSSDc log-tranformed corrected root mean square of successive differences, lnSDNN log-transformed SD of normal-to-normal intervals, lnSDNNc log-transformed corrected SD of normal-to-normal intervals, CI confidence interval.

Fig. 4. Kaplan–Meier curves for risk of mortality among participants in quartiles of HRV traits.

a RMSSD and all-cause mortality; b RMSSD and cardiovascular mortality; c RMSSDc and all-cause mortality; d RMSSDc and cardiovascular mortality. RMSSD root mean square of successive differences, RMSSDc corrected root mean square of successive differences.

Fig. 5. Kaplan–Meier curves for risk of mortality among participants in quartiles of HRV traits.

a SDNN and all-cause mortality; b SDNN and cardiovascular mortality; c SDNNc and all-cause mortality; d SDNNc and cardiovascular mortality. SDNN SD of normal-to-normal intervals, SDNNc corrected SD of normal-to-normal intervals.