Genetic insights into resting heart rate and its role in cardiovascular disease

Abstract

Resting heart rate is associated with cardiovascular diseases and mortality in observational and Mendelian randomization studies. The aims of this study are to extend the number of resting heart rate associated genetic variants and to obtain further insights in resting heart rate biology and its clinical consequences. A genome-wide meta-analysis of 100 studies in up to 835,465 individuals reveals 493 independent genetic variants in 352 loci, including 68 genetic variants outside previously identified resting heart rate associated loci. We prioritize 670 genes and in silico annotations point to their enrichment in cardiomyocytes and provide insights in their ECG signature. Two-sample Mendelian randomization analyses indicate that higher genetically predicted resting heart rate increases risk of dilated cardiomyopathy, but decreases risk of developing atrial fibrillation, ischemic stroke, and cardio-embolic stroke. We do not find evidence for a linear or non-linear genetic association between resting heart rate and all-cause mortality in contrast to our previous Mendelian randomization study. Systematic alteration of key differences between the current and previous Mendelian randomization study indicates that the most likely cause of the discrepancy between these studies arises from false positive findings in previous one-sample MR analyses caused by weak-instrument bias at lower P-value thresholds. The results extend our understanding of resting heart rate biology and give additional insights in its role in cardiovascular disease development.

Fig. 1. Study flowchart showing the study design, in silico annotations, and functional analyses.

a Schematic overview of the study design for the discovery and replication of genetic loci associated with resting heart rate (RHR) using mixed linear models with a two-sided P-value of P < × 10⁻⁸ to define genome-wide significance. A genome-wide significant genetic variant was considered replicated if P < 0.01 in the UK Biobank and IC-RHR cohort with concordant effect sizes. The black-bordered boxes show the methodology, the red-bordered boxes show the most important results. b Analyses performed to evaluate RHR-associated genetic variants and to gain further insights into the underlying biology. c Schematic presentation of the two-sample Mendelian randomization analyses of genetically predicted RHR on mortality and cardiovascular diseases. Effect sizes were taken from the IC-RHR data to test the associations with mortality and cardiovascular diseases in the UK Biobank. Effect sizes were taken from the UK Biobank to test the association with coronary artery disease and myocardial infarction in the CARDIoGRAMplusC4D cohort, atrial fibrillation in the AFGen cohort, and any, ischemic, cardio-embolic, large artery and small vessel stroke within the MEGASTROKE consortium. BMI body mass index, GWAS genome-wide association study, HRC Haplotype Reference Panel, IC-RHR International Consortium for Resting Heart Rate, MB megabase, N sample size, Neff effective sample size, PC principal components, RHR resting heart rate, SNPs single nucleotide polymorphisms, QC quality control, 1000G = 1000 Genomes.

Fig. 2. Overview of the findings in the genome-wide association study and in silico search of candidate causal genes.

a Manhattan plot showing the −log₁₀(P-value) for the association of all genotyped or imputed genetics variants with resting heart rate (RHR) assessed using mixed linear models. Red indicates novel and internally replicated RHR-associated loci and black indicates novel but unreplicated RHR-associated loci. Dark gray indicates RHR-associated genetic variants within 1 MB of previously identified RHR-associated loci, which were internally replicated in the current study. Light gray indicates RHR-associated genetic variants within 1 MB of previously identified RHR-associated loci, which were not internally replicated in the current study. A two-sided P-value of P < 1 × 10⁻⁸ was used to define genome-wide significance. A genome-wide significant genetic variant was considered replicated if P < 0.01 in the UK Biobank and IC-RHR cohort with concordant effect sizes. b Venn diagram of the 352 identified loci. Of the 352 loci, 332 were internally replicated. c Quantile–quantile (QQ) plot of the final meta-analysis. The black dots represent the observed statistic for the genotyped genetic variants against the corresponding expected statistic. The linkage disequilibrium score regression intercept after the final meta-analysis was 1.051, suggesting little evidence of genomic inflation due to non-polygenic signal. d Venn diagram of the prioritization of the 670 unique candidate causal genes as identified by one or multiple strategies. Venn plot shows the overlap of genes tagged by one or multiple strategies, including (1) by proximity, the nearest gene or any gene within 10 kb; (2) genes containing coding variants in LD with RHR-associated variants at R² > 0.8; (3) eQTL genes in LD (R² > 0.8) with RHR-associated variants which achieved a Bonferroni corrected two-sided P = 2.65 × 10⁻⁷ and passed the HEIDI test at a P > 0.05; and (4) DEPICT genes which achieved multiple hypotheses corrected FDR < 0.05. DEPICT data-driven expression prioritized integration for complex traits, eQTL expression quantitative trait loci.

Fig. 3. Conditional analyses of tissue enrichment by DEPICT emphasizes cardiac tissue for RHR biology.

a Results of the DEPICT tissue enrichment analysis. The Y-axis shows the tissues clustered by the first MeSH term, ordered on Z-value per cluster. The X-axis shows the Z-value. A multiple comparisons corrected two-sided FDR < 0.05, corresponding to a P-value < 9.75 × 10⁻³ and Z-value of 2.585, was considered to be statistically significant. Significant tissues are plotted in red and annotated, other tissues are plotted in gray. Conditional analyses were performed by correcting for the tissue with the highest Z-value to investigate whether significant tissues were independently associated with RHR. Not a single tissue remained significant at an FDR < 0.05 after three consecutive corrections (for the heart, heart valve, and arteries). Panel b–d shows Z-values of all tissues after consecutive correction for respectively heart and heart valves, heart and arteries, and heart valve and arteries. This jointly provides information on which the other tissues are co-dependent. FDR false discovery rate.

Fig. 4. Mendelian randomization shows absence of linear and non-linear associations between genetically predicted RHR and all-cause mortality.

Linear and non-linear Mendelian randomization analyses were performed to test the association between genetically predicted RHR and all-cause mortality. a Forest plot of the linear MR analyses between genetically predicted RHR and all-cause mortality (Ncases = 16,289, Ncontrol = 396,183). With a single outcome, a two-sided P-value of P < 0.05 was considered significant. Hazard ratios and 95% confidence intervals are shown. The X-axis shows hazard ratio’s on a log₁₀ scale, the center as indicated by a gray line depicts a hazard ratio of 1. b Dose–response curve of the non-linear MR analyses between genetically predicted RHR and all-cause mortality (Ncases = 16,039, Ncontroles = 394,144). The comparisons are conducted within strata and therefore the graph provides information on the expected average change in the outcome if a person with an RHR of (say) 70 bpm instead had an RHR value of 90 bpm. The gradient at each point of the curve is the localized average causal effect. Shaded areas represent 95% confidence intervals. With a single outcome, a two-sided fractional polynomial non-linearity P-value of P < 0.05 was considered significant. The X-axis shows RHR, the Y-axis shows hazard ratios. The center as indicated by a dark gray line depicts a hazard ratio of 1. RHR resting heart rate, HR hazard ratio, CI confidence interval, MR Mendelian randomization, IVW inverse variance weighted, FE fixed effects, MRE multiplicative random effects.

Fig. 5. Mendelian randomization of genetically predicted RHR on cardiovascular diseases.

Forestplots of the linear Mendelian randomization analyses of resting heart rate (RHR) on cardiovascular diseases. Effect sizes were taken from the IC-RHR data to test the associations with mortality and cardiovascular diseases in the UK Biobank (panel a). Effect sizes were taken from the UK Biobank to test the association with cardiovascular diseases in the CARDIoGRAMplusC4D, AFGen, and MEGASTROKE consortia (panel b). Results of the MR-IVW, outlier-robust MR-Lasso, and plurality valid MR-Mix are provided. Sample sizes vary per outcome and per cohort and are shown in the figure. Odds ratios and 95% confidence intervals are shown. The X-axis shows odds ratio’s on a log₁₀ scale, the center as indicated by a gray line depicts an odds ratio of 1. RHR resting heart rate, MR Mendelian randomization, IVW inverse variance weighted multiplicative random effects, OR odds ratio, CI confidence interval.

Fig. 6. Multivariable Mendelian randomization reveals pulse pressure and atrial fibrillation as potential mediators of the association of genetically predicted RHR with ischemic and cardio-embolic stroke, respectively.

Forestplots of the results of the two-sample multivariable Mendelian randomization analyses of resting heart rate on a any stroke (Ncases = 67,162; Ncontrols = 454,450), b ischemic stroke (Ncases = 60,341; Ncontrols = 454,450) and c cardio-embolic stroke (Ncases = 9006; Ncontrols = 403,807), when using atrial fibrillation, systolic, diastolic and pulse pressure as secondary exposures. Shown in red are the univariable Mendelian randomization estimates which represent the total estimates of resting heart rate on the outcome. In black are the multivariable Mendelian randomization estimates, which show the direct effect of RHR when corrected for the secondary exposure. These results indicate that atrial fibrillation attenuates the beneficial effect of a higher resting heart rate on cardio-embolic stroke, while pulse pressure attenuates the beneficial effect on any ischemic stroke. MR-Steiger sensitivity analysis indicated that the association between the RHR-associated genetic variants and pulse pressure is unlikely mediated through RHR entirely and biological pleiotropic effects are therefore more likely to cause the attenuation of the association between RHR and stroke when correcting for pulse pressure. Odds ratios and 95% confidence intervals are shown. The X-axis shows odds ratio’s on a log₁₀ scale, the center as indicated by a gray line depicts an odds ratio of 1. RHR resting heart rate; MV multivariable, Nsnp number of SNPs.