Resting Heart Rate and Ischemic Stroke in Patients with Heart Failure

Approximately 5.7 million adults are living with heart failure in the United States, and the prevalence is increasing [1]. Among heart failure patients, approximately half have reduced left ventricular ejection fraction (LVEF). Heart failure with reduced LVEF (HFrEF) is a major cause of mortality and hospital admission, and a high resting heart rate (RHR) is a strong predictor of mortality and hospital re-admission in patients with HFrEF [2,3,4]. Beta-adrenoceptor-blocking agents (beta-blockers) are now well established as a mandatory therapy in patients with HFrEF [5,6,7,8], and part of their beneficial effect may depend on RHR reduction [9,10,11]. However, the relationship between RHR and ischemic stroke risk in HFrEF is still unclear, although HFrEF is associated with an increased risk for cardioembolic stroke [12,13]. The aim of this study was to investigate the relationship between RHR and ischemic stroke in HFrEF patients in sinus rhythm who were treated with beta-blockers.

We analyzed data from the Warfarin versus Aspirin in Reduced Cardiac Ejection Fraction (WARCEF) trial, which compared warfarin and aspirin in a double-blind, randomized design [14]. The results of the primary analysis have been previously published [15]. WARCEF obtained data from 168 centers in 11 countries, and enrolled 2,305 patients with follow-up periods of up to 6 years (mean 3.5 ± 1.8 years). Patients were >18 years of age and had normal sinus rhythm and LVEF ≤35% within 3 months before randomization. All patients were treated with guideline-recommended medical treatment. Since approximately 90% of patients were treated with beta-blockers, which affect RHR, we limited our analysis to only these patients. A total of 243 patients who were not treated with beta-blockers were excluded from the study. Two others were excluded because they did not have RHR information. The final sample for analysis thus included 2,060 patients. RHR was obtained from the baseline electrocardiogram. The study was approved by the institutional review boards and ethics boards of participating centers.

The clinical characteristics included in our study were age, sex, hypertension, diabetes mellitus, current smoking status and alcohol consumption, history of myocardial infarction (MI), history of atrial fibrillation (AF), prior stroke/transient ischemic attack (TIA), education level, New York Heart Association (NYHA) class, systolic and diastolic blood pressures (BP), pulse pressure, medications, implantable cardioverter defibrillator, and LVEF. LVEF on either quantitative echocardiography, radionuclide or contrast ventriculography was obtained in all patients within 3 months before randomization.

Follow-up was performed monthly by telephone or in person. An in-person assessment was conducted quarterly for clinical evaluation. Stroke was defined as a clinically relevant new lesion detected on computed tomography or magnetic resonance imaging or, in the absence of a new lesion, clinical findings that were consistent with the occurrence of clinical stroke and that lasted for longer than 24 h.

Categorical variables are presented as number/total number (%) and compared by quartiles of RHR using chi-square test or Fisher exact test. Continuous variables are presented as mean ± SD and compared using the ANOVA F test. Kaplan-Meier (KM) estimates for ischemic stroke, stratified by quartiles of RHR, were also calculated. Univariable and multivariable linear regression analyses were conducted to determine the variables independently associated with RHR. Univariable and multivariable Cox proportional hazards regression analyses were then performed to identify the association between RHR and ischemic stroke. The linearity of associations was assessed using restricted cubic splines and, if a trend of non-linearity was found (p < 0.10), a linear spline or quadratic or cubic polynomial transformation was chosen based on the univariable Akaike's information criterion (AIC). To compare the association of RHR with ischemic stroke in aspirin- and warfarin-treated subgroups, we added treatment and its interaction with RHR to the Cox models. Multivariable analyses were performed in 2 models. Model 1: adjustment by variables with significant association with ischemic stroke in univariable Cox proportional hazard analysis, and model 2: adjustment as in model 1 plus variables with significant association with RHR in univariable linear regression analysis. Missing values for the covariates were imputed using means for continuous variables and model values for categorical variables. Differences were considered significant at p < 0.05 2-sided. Statistical analyses were performed using SAS 9.4 software (SAS Institute, Cary, NC, USA).

The study sample had a mean age of 60 ± 11 years and a mean heart rate of 71 beats/min (median 70 beats/min, 25th-75th percentile 64-80 beats/min). Baseline characteristics stratified by RHR quartiles are shown in Table 1. The factors independently associated with RHR in a multivariable linear regression analysis were age (β-coefficient -0.132, p < 0.001), diabetes mellitus (β-coefficient 2.188, p < 0.001), current smoking (β-coefficient 2.614, p < 0.001), current alcohol consumption >2 oz/day (β-coefficient -2.037, p < 0.001), history of MI (β-coefficient -2.085, p < 0.001), history of AF (β-coefficient -3.448, p = 0.008), NYHA class (β-coefficient 1.609, p = 0.003), diastolic BP (β-coefficient 0.169, p < 0.001), pulse pressure (β-coefficient -0.061, p = 0.001), diuretics (β-coefficient: 2.037, p = 0.002), and LVEF (β-coefficient -0.134, p < 0.001; Table 2).

During 3.5 ± 1.8 years of follow-up, 77 patients (5.3% from KM curve) developed ischemic stroke. The highest incidence of ischemic stroke (21/503 [KM 6.9%]) was observed in the lowest RHR quartile (RHR <64 beats/min) compared to other groups; 22/573 (KM 5.3%) in 64-70 beats/min, 13/465 (KM 3.5%) in 71-79 beats/min, and 21/519 (KM 5.4%) in RHR >79 beats/min (p = 0.693).

A trend of nonlinear association between RHR and stroke was found (p = 0.08), and a linear spline with knot at RHR = 64 beats/min was selected as the best model based on AIC. The hazard ratio (HR) plot is displayed in Figure 1. Results from Cox models are presented in Table 3. Overall, low RHR was a significant predictor of stroke (adjusted p = 0.044 in model 1). In particular, the risk of stroke increased with decreasing RHR for RHR values <64 beats/min (adjusted HR 1.07, p = 0.013), while it was unaffected above that value (adjusted HR 0.99, p = 0.355). Ischemic stroke was also significantly associated with the history of stroke or TIA (adjusted HR 3.42, p < 0.001) and LVEF (quadratic model, overall p < 0.001, 75th vs. 25th percentile adjusted HR 0.82). Even after adjustment for additional variables with significant association with RHR (age, diabetes mellitus, current smoking, current alcohol consumption, history of MI, history of AF, NYHA class, diastolic BP, pulse pressure, diuretics, and statin), RHR remained significantly associated with ischemic stroke (Table 3, model 2).

Table 4 shows the relationship between RHR and stroke risk in warfarin- and aspirin-treated subgroups. The interaction model results showed that lower RHR was significantly associated with ischemic stroke among patients randomized to aspirin (p = 0.039), whereas there was no relationship in those randomized to warfarin (p = 0.408). This result persisted after adjustment for variables with significant association with RHR (Table 4, model 2). In patients with RHR <64 beats/min, significantly lower ischemic stroke rate was observed in the warfarin group compared to the aspirin group (5/235 [KM 4.4%] vs. 16/268 [KM 9.2%], p = 0.034), whereas ischemic stroke rate did not differ between the 2 treatment groups in patients with RHR ≥64 beats/min (22/790 [KM 3.6%] vs. 34/767 [KM 6.0%], p = 0.086).

This study demonstrates for the first time that low RHR is associated with a higher incidence of ischemic stroke among patients with HFrEF who are in sinus rhythm and are treated with currently recommended medical regimen, including beta-blockers.

In patients with HFrEF who are in sinus rhythm, a high RHR is associated with increased mortality and hospital re-admissions [2,3,4], and beta-blockers substantially improve the outcome [5,6,7,8]. Although the benefits of beta-blockers may not be entirely related to RHR reduction, several meta-analyses have shown a stronger effect on survival for RHR rather than the beta-blockers dose achieved [9,10,11]. As a result, it has become a common clinical assumption that the beneficial effect of beta-blockers depends on, or is heralded by, their RHR-lowering effect: the slower the RHR, the greater the benefit. However, although HFrEF is associated with an increased risk for cardioembolic stroke [12,13], the relationship between RHR and ischemic stroke has not been fully investigated in patients with HFrEF in sinus rhythm. Here, we report for the first time that low RHR was associated with a higher incidence of ischemic stroke in patients with HFrEF in sinus rhythm. Our finding was unexpected, because low RHR is usually associated with lower mortality and rate of hospital admission in these patients [2,3,4]. The relationship between heart rate and stroke still remains unclear in other clinical settings, where conflicting results have been reported [16,17,18,19,20,21]. Mao et al. [18] showed that high RHR increased the risk of stroke in 169,871 general Chinese adults ≥40 years. Similarly, data from patients with stable coronary artery disease and hypertension demonstrated that high RHR was associated with an increased risk of stroke [19,20]. More recently, the Reasons for Geographic and Racial Differences in Stroke study conducted in 24,730 subjects without history of stroke showed that each 10 beats/min increase in heart rate was associated with a 10% increase in the risk of stroke [21]. In contrast, reports from the general French population and the Women's Health Initiative Study did not show an association between RHR and stroke [16,17].

The underlying mechanisms of our finding are unclear, but several potential explanations can be hypothesized. Recent studies have reported that low RHR is associated with higher incidence of AF development in various populations [22,23,24]. Bohm et al. [22] reported that RHR lower than 60 beats/min was associated with increased incidence of AF in 27,064 patients with high cardiovascular risk during a mean follow-up period of 4.7 years. O'Neal et al. [23] also reported that RHR lower than 60 beats/min was an independent risk for AF development in 5,226 elderly individuals from the general population. In our study, patients with low RHR may have more frequently developed transient episodes of AF during follow-up, which might be involved in their higher risk for ischemic stroke. Another possible mechanism could be an increase in central aortic pressure secondary to heart rate lowering. Bradycardia leads to dyssynchrony or uncoupling between outgoing and reflected waves, thereby elevating the central aortic pressure. In the Conduit Artery Functional Evaluation study [25], significantly higher central aortic systolic BP was observed with beta-blocker treatment compared with calcium channel blocker treatment despite similar effect on peripheral BP, resulting in a higher incidence of stroke in patients with hypertension. Finally, an increase in pulse pressure may also contribute to the development of stroke in HFrEF patients with low RHR. Because the mean arterial pressure is a product of cardiac output (heart rate stroke volume) and peripheral vascular resistance, low RHR should result in higher stroke volume to maintain cardiac output. A higher stroke volume, in turn, causes elevated pulse pressure which has been recognized as an independent predictor of stroke [26,27]. Indeed, RHR was negatively correlated with pulse pressure in our population. However, pulse pressure was not significantly associated with ischemic stroke.

Interestingly, when patients were divided into subgroups on the basis of assigned antithrombotic treatment, low RHR was associated with ischemic stroke in the aspirin group, but not in the warfarin group. In patients with RHR <64 beats/min, patients treated with warfarin had significantly lower ischemic stroke rate than those with aspirin. This result suggests that systemic anticoagulation may counteract the risk of stroke associated with low RHR, and also suggests a potential thromboembolic component for the stroke mechanism. Furthermore, it may indicate that warfarin treatment may be preferable to aspirin treatment for stroke prevention in patients with low RHR.

Among medications that affect heart rate, ivabradine is a novel HF medication that specifically inhibits the I_f current in the sinoatrial node, thereby lowering the heart rate without affecting the other aspects of cardiac function [28]. In Systolic Heart failure treatment with the ‑_f inhibitor ivabradine Trial, RHR reduction with ivabradine was associated with 26% risk reduction of first HF hospitalization, and 11% risk reduction of first all-cause hospitalization [29]. Because of the study period of WARCEF trial (from October 2002 through January 2010), ivabradine was not used. Future studies are needed to investigate the association of RHR with ischemic stroke in patients treated with ivabradine.

Our study has several limitations. Because we enrolled patients with HFrEF in sinus rhythm, the results may not be generalizable to patients with HF with preserved LVEF and to those with AF. Because of the absence of information on beta-blocker dose, we cannot evaluate a possible effect of different doses on our results. However, several meta-analyses have shown a stronger relationship between the RHR and prognosis in HF than between beta-blockers dose and prognosis [9,10,11]. Finally, although we performed multivariable analyses adjusting for ischemic stroke risk factors and variables associated with RHR, we cannot rule out the possibility of unmeasured confounders playing a role in the observed associations.

In contrast to the beneficial effect of lower RHR on mortality and hospital admissions, lower RHR increased the risk of ischemic stroke in HFrEF patients in sinus rhythm treated with beta-blockers. Further studies are required to evaluate the mechanisms for the increased risk of ischemic stroke in patients with low RHR.

We thank Michelle Bierig, RDCS, and Rui Liu, MD, for their help with the echocardiographic measurements.

The study was funded by U01-NS-043975 (Dr. Shunichi Homma) and U01-NS-039143 (Dr. John L.P. Thompson) from the National Institute of Neurological Disorders and Stroke.

Dr. Shunichi Homma reports receiving payment from AGA Medical (now St. Jude Medical) for his work as a member of a data and safety monitoring board and consulting fees from Boehringer Ingelheim. Dr. Bruce Levin reports receiving consulting fees from United Healthcare. Dr. John R. Teerlink reports receiving research grants or consulting fees from Amgen, Bayer, Cardio3 Bioscience, Cytokinetics, Mast Therapeutics, Medtronic, Novartis, St. Jude, and Trevena. Dr. Arthur J. Labovitz reports receiving grant support from Boehringer Ingelheim and BMS Pfizer, lecture fees from Boehringer Ingelheim, and fees for the development of educational presentations from the American College of Cardiology. Dr. Stefan D. Anker reports receiving consulting fees from Vifor, Bayer, Janssen, Novartis, Relypsa, ZS-Pharma, and Thermo Fisher; grant support from Vifor Pharma, and Abbott Vascular; and lecture fees from Vifor, Novartis, and Stealth Peptides. Dr. Piotr Ponikowski reports receiving consulting fees from Bayer, Boehringer Ingelheim, Coridea, Corthera, Johnson & Johnson, Pfizer, Respicardia, and Vifor Pharma; grant support from Vifor Pharma on behalf of himself and his institution; and lecture fees from Abbott, Boehringer Ingelheim, Merck Serono, Pfizer, Respicardia, Sanofi-Aventis, Servier, and Vifor Pharma. Dr. Gregory Y.H. Lip reports receiving consulting fees from Bayer/Janssen, Astellas, Merck, Sanofi, BMS/Pfizer, Biotronik, Medtronic, Portola, Boehringer Ingelheim, Microlife, and Daiichi-Sankyo; speakers bureau fees from Bayer, BMS/Pfizer, Medtronic, Boehringer Ingelheim, Microlife, Roche, and Daiichi-Sankyo; and payment for the development of educational presentations from Bayer, Boehringer Ingelheim, and Merck. The other authors report no conflicts.