What's new?

We showed that heart rate (HR) above 70 bpm assessed using an automated office blood pressure measurement (AOBPM) is associated with an increased risk of cardiovascular events in individuals with and without a prior diagnosis of cardiovascular disease. When AOBPM is used to estimate HR, an increased risk of cardiovascular events is observed at lower HR values than currently suggested. Although high HR is related to elevated risk of cardiovascular events, the risk for myocardial infarction may be increased in patients with prior cardiovascular disease and low HR values.

Introduction

Elevated heart rate (HR) is associated with increased cardiovascular morbidity and mortality in the general population and patients previously diagnosed with cardiovascular disease (CVD).^1-3 A post hoc analysis of data from the LIFE (Losartan Intervention For End point) study revealed that a 10 bpm increase in HR results in a 16% increase in the risk of cardiovascular death and 25% higher risk of all‑cause mortality.⁴ In the post hoc analysis of the VALUE (Valsartan Antihypertensive Long‑term Use Evaluation) trial, participants with HR in the highest quintile had a higher risk for heart failure (HF), sudden cardiac death, myocardial infarction (MI), stroke, and all‑cause mortality by 1.93 (95% CI, 1.45–2.55), 1.49 (95% CI, 0.97–2.28), 1.38 (95% CI, 1.03–1.85), 1.49 (95% CI, 1.1–2.00), and 1.98 (95% CI, 1.64–2.38), respectively, as compared with participants with HR in the lowest quintile.⁵ In both LIFE and VALUE studies, HR was measured using an electrocardiogram (ECG).

Although ECG most accurately determines HR, some trials also assessed the relationship between HR and cardiovascular events, when HR was considered to be equivalent to the pulse rate obtained during blood pressure (BP) measurements.⁶

Additionally, HR obtained during office BP measurement (OBPM) or 24‑hour ambulatory BP monitoring (ABPM) is associated with cardiovascular outcomes. Based on a large dataset from the Spanish Ambulatory Blood Pressure Registry, Böhm et al⁷ inferred that increased HR measured during OBPM is associated with an elevated risk for all‑cause and cardiovascular death. Similarly, elevated HR recorded using ABPM over 24 hours, activity, and night‑time rest periods were factors associated with increased risk of all‑cause death or cardiovascular death.⁷ Interestingly, the authors found that office HR was higher than the mean 24‑hour HR; however, no relevant difference between the office and mean day HR was observed.

Some data also indicated that HR measured during home BP monitoring (HBPM) in participants without significant arrhythmias is associated with an increased risk of cardiovascular events.⁸ Using data from the Ohasama study, Hozawa et al⁸ showed that increased HR measured during HBPM is associated with higher cardiovascular mortality independently from HBPM values.

Current evidence supporting the use of HR measured during automated office blood pressure measurements (AOBPM) to predict cardiovascular events is lacking. AOBPM is an increasingly used and promising method for BP evaluation.^9,10 Due to multiple, unattended BP measurements and less susceptibility to white‑coat effects, AOBPM is believed to provide lower BP levels than traditional OBPM.¹¹ To our knowledge, HR assessed using AOBPM has not yet been compared with the values obtained using other measurement methods; however, the assumption that AOBPM provides the same values might not be justified. Therefore, this study aimed to examine the impact of HR assessment with AOBPM on the risk of cardiovascular events in participants with and without prior CVD diagnosis.

Patients and methods

Results

Comparison of participants with and without prior cardiovascular disease

Among the 9361 SPRINT participants, 1877 had prior CVD. They were older and more commonly men, smokers, and had a history of chronic kidney disease more often than those without CVD diagnosis at baseline.

The participants with a history of CVD also had lower baseline HR, SBP, and DBP than those without prior CVD diagnosis. During the trial, CE events were more than twice more common in the participants with than without a CVD history (10.9% vs 4.8%). All individual components of CE were also more commonly recorded when CVD was present at baseline. The participants with prior CVD had also lower in‑trial HR and DBP than those without prior CVD diagnosis, whereas in‑trial SBP was similar in both groups. More individuals with prior CVD diagnosis than without CVD received treatment with β-blockers at baseline (60.4% vs 29.3%; P <⁠0.001). There was no difference in the use of nondihydropyridine calcium channel blockers (eg, diltiazem or verapamil) between the groups (5.1 vs 5.2%; P = 0.76, respectively). Only 1 person received digoxin at the study beginning and had prior CVD. In Figure 1 the distribution of baseline HR in the participants with and without a history of CVD disease is presented. Lower values and closer to the central value distribution of baseline HR in the participants with prior CVD (Figure 1) are probably associated with extensive use of HR lowering drugs (β-blockers) in this group.

A detailed comparison of the SPRINT participants based on a history of CVD is presented in Table 1.

Table 1. Comparison of participants with and without prior diagnosis of cardiovascular disease

Parameter		Participants without CVD (n = 7484)	Participants with CVD (n = 1877)	P value
Data are presented as mean (SD), median (interquartile range), or number (percentage) of patients. SI conversion factors: to convert glucose to mmol/l, multiply by 0.0555; HDL and total cholesterol to mmol/l, by 0.0259; triglycerides to mmol/l, by 0.113, creatinine to µmol/l, by 88.4. Abbreviations: ASA, acetylsalicylic acid; BMI, body mass index; CE, composite clinical end point event; CKD, chronic kidney disease; CVD, cardiovascular disease; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; HDL, high‑density lipoprotein; HR, heart rate; MI, myocardial infarction; SBP, systolic blood pressure
Baseline heart rate, bpm		66.9 (11.6)	63.6 (11)	<⁠0.001
Age, y		67.5 (9.4)	69.7 (9.5)	<⁠0.001
Female sex		2822 (37.7)	510 (27.2)	<⁠0.001
Black race		2482 (33.2)	465 (24.8)	<⁠0.001
Smoking status	Current smoker	976 (13)	264 (14.1)	<⁠0.001
	Former smoker	3047 (40.7)	926 (49.3)
	Never smoker	3440 (46)	682 (36.3)
BMI, kg/m2		30 (5.8)	29.4 (5.6)	0.001
Previous CKD		2002 (26.8)	644 (34.3)	<⁠0.001
eGFR, ml/min/1.73 m2		72.5 (20.6)	68.8 (20.4)	<⁠0.001
Creatinine, mg/dl		1.1 (0.3)	1.1 (0.4)	<⁠0.001
Total cholesterol, mg/dl		194.5 (39.8)	172.7 (41.8)	<⁠0.001
HDL cholesterol, mg/dl		53.5 (14.7)	50.3 (13.2)	<⁠0.001
Triglycerides, mg/dl		107 (77–151)	106 (77–147)	0.42
Serum glucose, mg/dl		98.6 (13.4)	99.5 (14.1)	0.02
On statin		2757 (37.2)	1297 (69.5)	<⁠0.001
On ASA		3341 (44.8)	1415 (75.7)	<⁠0.001
Baseline SBP, mm Hg		140 (15.5)	138.5 (15.8)	<⁠0.001
Baseline DBP, mm Hg		78.9 (11.7)	74.9 (12.3)	<⁠0.001
Mean in‑trial SBP, mm Hg		129.5 (9.8)	129.1 (9.9)	0.11
Mean in‑trial DBP, mm Hg		72.8 (9.1)	69.5 (9.4)	<⁠0.001
In‑trial HR, bpm		68.2 (9.3)	64.7 (8.9)	<⁠0.001
Allocation to intensive treatment arm		3738 (49.9)	940 (50.1)	0.94
CE		357 (4.8)	205 (10.9)	<⁠0.001
MI		134 (1.8)	79 (4.2)	<⁠0.001
Acute coronary syndrome other than MI		37 (0.5)	43 (2.3)	<⁠0.001
Stroke		93 (1.2)	39 (2.1)	0.008
Heart failure exacerbation		101 (1.3)	61 (3.2)	<⁠0.001
Cardiovascular death		59 (0.8)	43 (2.3)	<⁠0.001
All‑cause death		246 (3.3)	119 (6.3)	<⁠0.001

The relationship between heart rate and cardiovascular event risk

Event rates of CE, MI, and CVD death were evaluated based on the presence of CVD at baseline and baseline HR (Figure 2). In the participants both with and without CVD history, the annual event rate of CE was the highest with HR values ranging from 70 to 80 bpm. Similarly, when HF events and cardiovascular deaths were considered, the highest event rate was observed at HR of 70 to 80 bpm. The highest event rate for MI was observed in the participants with prior CVD and HR below 50 bpm. In the individuals without prior CVD, the event risk rose together with growing HR values. The event rates of acute coronary syndromes other than MI, stroke, HF exacerbation, and total death are presented in Supplementary material, Figure S1.

In the Cox proportional hazard model including age, sex, race, baseline SBP, baseline HR, current smoking status, allocation to intensive treatment arm, statin treatment, use of aspirin, total cholesterol, serum glucose concentration and eGFR, the interaction term between HR and CVD history was nonsignificant in relation to the CE risk (P = 0.76), stroke (P = 0.51), HF exacerbation (P = 0.37), cardiovascular death (P = 0.12), and total death (P = 0.20). However, the interaction term was significant when the risk of MI was analyzed (P = 0.049).

Figure 3 presents hazard ratio plots against baseline HR from the models including age, sex, race, baseline SBP, baseline HR, current smoking status, allocation to intensive treatment arm, statin treatment, use of aspirin, total cholesterol, serum glucose concentration, and eGFR to predict CE, MI, and CVD death. Similar plots showing the risk of acute coronary syndromes other than MI, stroke, HF exacerbation, and total death are presented in Supplementary material, Figure S2.

To depict the relationship between baseline HR and CE, MI, and cardiovascular death, 3 additional proportional hazard models were created. We assessed the impact of baseline HR higher than 70 bpm on the risk of CE and cardiovascular death in Model A and Model C, respectively (Table 3). Due to increasing risk of MI at lower baseline HR values, especially pronounced in the participants with prior CVD, Model B included also baseline HR lower than 60 bpm. All the models were adjusted for age, sex, race, renal function, history of CVD, smoking status, baseline SBP, allocation to intensive treatment arm, statin or aspirin treatment, total cholesterol, and serum glucose. In Models A and C, baseline HR above 70 bpm was associated with increased risk of CE or cardiovascular death, as compared with baseline HR up to 70 bpm (hazard ratio, 1.32; 95% CI, 1.1–1.58; P = 0.003 and hazard ratio, 1.9; 95% CI, 1.24–2.9; P = 0.003, respectively). As for the risk of MI, we showed that baseline HR lower than 60 bpm and higher than 70 bpm was associated with increased risk in comparison with HR within the range of 60 to 70 bpm, by 1.41 with 95% CI of 1.01–1.97 (P = 0.045) and by 1.42 with 95% CI of 1–2.02 (P = 0.05), respectively.

Table 3. Proportional hazard models evaluating the impact of baseline heart rate above 70 bpm on composite clinical end point event (Model A), myocardial infarction (Model B), and cardiovascular death (Model C)

Variable	Hazard ratio (95% CI)	P value
Clinical composite end point was defined as MI, acute coronary syndrome other than MI, heart failure exacerbation, stroke, or cardiovascular death. a In comparison with never smokers b In comparison with baseline HR in the range of 60–70 bpm Abbreviations: see Table 1
Model A – prediction of CE
Age, y	1.05 (1.04–1.06)	<⁠0.001
Female sex	0.72 (0.6–0.88)	0.001
Black race	1.05 (0.86–1.28)	0.64
Baseline SBP, mm Hg	1.01 (1–1.01)	0.01
Former smokera	1.2 (1–1.45)	0.06
Current smokera	2.21 (1.69–2.87)	<⁠0.001
Allocation to intensive treatment arm	0.74 (0.62–0.87)	<⁠0.001
On statin	1.05 (0.87–1.27)	0.59
On aspirin	1.15 (0.96–1.38)	0.13
Total cholesterol, mg/dl	1 (1–1.01)	0.01
Glucose, mg/dl	1 (1–1.01)	0.28
eGFR, ml/min/1.73 m2	0.99 (0.99–0.99)	<⁠0.001
Baseline HR >70 bpm	1.32 (1.1–1.58)	0.003
History of CVD	2.07 (1.71–2.49)	<⁠0.001
Model B – prediction of MI
Age, y	1.04 (1.03–1.06)	<⁠0.001
Female sex	0.63 (0.45–0.87)	0.005
Black race	0.87 (0.62–1.22)	0.41
Baseline SBP, mm Hg	1.01 (1–1.02)	0.04
Former smokera	1.43 (1.04–1.96)	0.03
Current smokera	3.25 (2.15–4.9)	<⁠0.001
Allocation to intensive treatment arm	0.81 (0.62–1.06)	0.13
On statin	1.13 (0.83–1.53)	0.44
On aspirin	1.06 (0.78–1.42)	0.72
Total cholesterol, mg/dl	1 (1–1.01)	0.008
Glucose, mg/dl	1 (0.99–1.01)	0.69
eGFR, ml/min/1.73 m2	0.99 (0.98–1)	0.04
Baseline HR >70 bpmb	1.42 (1–2.02)	0.05
Baseline HR <⁠60 bpmb	1.41 (1.01–1.97)	0.045
History of CVD	2.03 (1.5–2.76)	<⁠0.001
Model C – prediction of cardiovascular death
Age, y	1.05 (1.02–1.07)	<⁠0.001
Female sex	0.43 (0.26–0.72)	0.001
Black race	1.56 (0.99–2.47)	0.06
Baseline SBP, mm Hg	1.02 (1.01–1.03)	0.004
Former smokera	1.48 (0.94–2.33)	0.09
Current smokera	2.06 (1.07–3.98)	0.03
Allocation to intensive treatment arm	0.57 (0.38–0.86)	0.007
On statin	1.31 (0.84–2.07)	0.24
On aspirin	0.96 (0.62–1.48)	0.84
Total cholesterol, mg/dl	1 (1–1.01)	0.80
Glucose, mg/dl	1.01 (0.99–1.02)	0.39
eGFR, ml/min/1.73 m2	0.98 (0.97–0.99)	<⁠0.001
Baseline HR >70 bpm	1.9 (1.24–2.9)	0.003
History of CVD	2.28 (1.47–3.53)	<⁠0.001

Discussion

This is the first study to show that elevated HR assessed using AOBPM is associated with a higher risk of cardiovascular events in participants with and without prior CVD. Additionally, an elevated risk for MI is present especially in the participants with prior CVD and low HR.

The negative impact of increased HR on cardiovascular events confirms previous study results.^1-3. However, it is worth paying attention to this risk factor, especially since the topic of HR control in the treatment of arterial hypertension has returned in recent years. For example, in a recent trial of 3463 patients in China, observed for 2.3 years, increased HR was found to be an independent risk factor for peripheral artery disease development.¹³ In another recent study,¹⁴ it was concluded that elevated HR (measured using ECG) is associated with incident HF in people with hypertension but without prior CVD (eg, HF, MI, stroke). The authors showed that HR higher than 79 bpm is associated with a higher risk of HF, but it should be noticed that even lower HR (74–79 bpm) was associated with an elevated risk for HF (by 1.65 with 95% CI of 1.07–2.55) after adjustment for age, sex, SBP, high‑density lipoprotein cholesterol, low‑density lipoprotein cholesterol, smoking status, alcohol consumption, physical exercise, presence of diabetes, use of antihypertensive agents, BMI, and eGFR. Contrary to these results, we were only able to observe an increase in HF decompensation events in the individuals with prior cardiovascular disease (Supplementary material, Figure S2). Moreover, in a recently published study from Tunisia,¹⁵ HR above 80 bpm was identified as an independent predictor of poor blood pressure control increasing the risk of cardiovascular events by 1.59 (95% CI, 1.48–1.71).The method of HR measurement was not described in the study, however, the authors did not use AOBPM to assess HR.

The effect of HR in the context of lower blood pressure target has not been explored in the randomized clinical trial design. In our previous analysis,¹⁶ also based on the SPRINT study, we showed that despite good BP control, HR among the participants treated to reach an SBP goal below 120 mm Hg remained a significant risk factor. In addition, in this subgroup, more potent relationship between the increased HR and cardiovascular events was observed than in the group with less controlled blood pressure.¹⁶ With reference to this conclusion, our finding presented here, that the risk increases even from HR of 70 bpm (not from 80 bpm, as indicated by the European Society of Cardiology [ESC] recommendations), is of particular importance to the physicians. This conclusion remains in line with the results of a meta‑analysis of 45 nonrandomized prospective cohort studies by Zhang et al.¹⁷ The authors showed that HR higher than 70 bpm is associated with unfavorable outcome regarding both cardiovascular and noncardiovascular events.

In a recent review describing the impact of elevated HR on cardiovascular risk in hypertensive participants, apart from genetic factors leading to elevated HR and BP, the authors underlined specific roles of sympathetic nervous system activity.¹⁸ HR is known to be strongly influenced by the balance between the sympathetic and parasympathetic nervous system activity. In the participants with elevated HR, higher norepinephrine concentration and increased muscle sympathetic nerve activity were found.^19,20 The sympathetic overactivity is also a known risk factor for adverse cardiovascular effects, including white‑coat hypertension, masked hypertension, progression of hypertension‑mediated end‑organ damage, and other hypertension‑related conditions.^18,21

We would like to emphasize that proving an association between HR measured using AOBPM and cardiovascular risk might be of special significance, because HR is lower during an unattended AOBPM due to decreased sympathetic activity.²² It may also indicate that the HR assessment using AOBPM is a more sensitive tool for predicting cardiovascular complications than HR obtained by the conventional method.

In the ESC recommendations, HR above 80 bpm is considered a cardiovascular risk modifier. In our study, an increase in cardiovascular risk was observed at lower HR values. Similar observations were found when considering the effect of HR assessed by HBPM or ABPM on cardiovascular events.^8,23 This finding can also be explained by the fact that during unattended AOBPM sympathetic activity is lower than during OBPM.

Regarding the balance between the sympathetic and parasympathetic systems, sympathetic activation is decreased during night rest. Our results are of interest in the context of other studies showing that nocturnal nondipping of HR is an independent predictor of cardiovascular events.²⁴

To the best of our knowledge, only 1 study has presented the impact of AOBPM HR on the occurrence of cardiovascular events. Wang et al²⁵ showed that in Chinese octogenarians elevated AOBM HR is associated with all‑cause and cardiovascular mortalities. However, the trial also showed that low HR values were associated with increased risk for all‑cause and cardiovascular mortalities. The lowest risk of these end points was observed when HR remained in the range of 77 to 85 bpm.

In our trial, the risk of MI rose in the participants with low HR but only in the group with prior CVD. As proposed by Boudoulas et al,²⁶ this phenomenon may be explained by increased central aortic pressure. In the participants with prior CVD, arterial stiffness is more pronounced and results in higher pulse wave velocity. Similarly, the pulse wave velocity reflected from the peripheral arterial circulation to the root of the aorta is higher than in the participants without CVD. Consequently, when HR is slow, the diastolic period is prolonged much more than the systolic period as compared with normal HR. Both these situations, that is, the higher velocity of reflected pulse wave and elongation of the diastolic period in comparison with the systolic period, result in the arrival of the reflected pulse wave at the root of the aorta in the systole instead of the diastole. Under such circumstances, the central aortic BP is elevated and coronary blood flow is reduced. Hence, diminished coronary blood flow may result in myocardial ischemia and infarction, especially in the participants with prior CVD diagnoses. Recently, Mthembu et al²⁷ found that the age of participants without CVD was the most important factor influencing the effects of low HR on the reflected pulse wave velocity and central BP. In our analysis, the participants without CVD were older than those with the CVD history; however, we speculate that the CVD diagnosis may have a similar or stronger effect on the reflected pulse wave velocity than age. Fortunately, increased MI risk was observed in the participants at a very low HR, rarely seen in clinical practice.

In the patients with stable coronary disease it is recommended to maintain resting HR in the range of 55 to 60 bpm.²⁸ These suggestion is based on the results of 2 studies. In the first study,²⁹ data of the patients with HR below 62 bpm were used as a reference. The multivariable Cox model showed only that the risk for total mortality or cardiovascular mortality was increased when HR ranged from 71 to 76 bpm or above, in comparison with HR below 62 bpm.²⁹ The second study³⁰also failed to show any difference in the risk of nonsudden death from MI between the participants with HR below 60 bpm and HR in the range of 60 to 64 bpm. In fact, none of these studies investigated if (or how) low HR may be harmful. Taking into account our results, we are convinced that the lack of differences in survival may be due to increased risk of MI at lower HR values. Moreover, several studies in patients with coronary artery disease demonstrated an increased risk of unfavorable cardiovascular events when HR was low. In the post hoc analysis of the INVEST (International Verapamil‑SR/trandolapril Study), the authors showed a J‑shaped relationship between HR and the risk of composite adverse events (death, nonfatal MI, or nonfatal stroke) in patients with prior MI.³¹ Similarly, the data from PROVE‑IT TIMI 22 trial³² indicated that both low and high HR was associated with an increased risk of composite clinical end point defined as death from any cause, MI, unstable angina requiring rehospitalization, revascularization after 30 days, or stroke. The analysis showed nadir at 68 bpm.

Several limitations of our analysis should be pointed out. The main limitation is its post hoc design. Furthermore, premature ending of the SPRINT could influence the results of this analysis. Second, HR measurements may be affected by the fact that only approximately 50% of AOBPM were taken as unattended.³³ However, the effects of the presence of an observer on HR assessed by AOBPM are unknown. We should also underline that not only HR, but also SBP and DBP values are assessed using AOBPM, and may be slightly different than the values obtained with other methods of BP estimation according to the results of other trials.³⁴ Our analysis did not include the effects of rate‑control drug prescription. Such an analysis was not performed, because some participants could receive such a treatment temporarily as antihypertensive therapy (eg, β-blockers).³⁵ We also did not evaluate the influence of potential arrhythmia, for example, atrial fibrillation.