OUP user menu

Resting heart rate as a predictive risk factor for sudden death in middle-aged men

Xavier Jouven, Mahmoud Zureik, Michel Desnos, Claude Guérot, Pierre Ducimetière
DOI: http://dx.doi.org/10.1016/S0008-6363(01)00230-9 373-378 First published online: 1 May 2001

Abstract

Objective: A relative hyperadrenergic tone related to abnormalities of the autonomic nervous system is suspected in the mechanisms of sudden death. Therefore, we assessed the role of an elevated basal heart rate in the occurrence of sudden death in a long-term cohort study. Methods: 7746 subjects aged 42–53 years, underwent ECG and physical examination conducted by a physician under standardized conditions, provided blood samples for laboratory tests, and answered questionnaires administered by trained interviewers. The vital status was obtained from specific inquiries up to the time of retirement and then by death certificates. Men with known ischemic heart disease were further excluded from analysis which was conducted on the 7079 remaining subjects. Results: After an average follow-up period of 23 years, there were 2083 deaths, among which were 603 cardiovascular deaths including 118 sudden deaths and 192 following myocardial infarction. The crude risk of sudden death increased linearly with the level of resting heart rate and the risk in men in the highest quintile of heart rate was 3.8 fold than in those in the lowest quintile, whereas rates were approximatively twice higher for fatal myocardial infarction, cardiovascular and total mortality (all P<0.01). When age, body mass index, systolic blood pressure, tobacco consumption, parental history of myocardial infarction and parental history of sudden death, cholesterol level, diabetic status, and sport activity were simultaneously entered into the survival model, resting heart rate remained an independent risk factor for sudden death (P = 0.03) but not for fatal myocardial infarction. Conclusion: An elevated heart rate at rest was confirmed as an independent risk factor for sudden death in middle-aged men.

Keywords
  • Heart rate (variability)
  • Sudden death

Time for primary review 34 days.

1 Introduction

Sudden death remains a major and still unsolved problem in industrialized countries. Abnormalities of the autonomic nervous system are suspected in the mechanisms of sudden cardiac death, and substantial data are accumulating that implicate abnormalities of both the sympathetic and parasympathetic systems, which modulate heart rate and heart rate variability [1]. Data that support either increased sensitivity to sympathetic input or abnormal sympathetic function issue from trials with patients, usually with ischemic heart disease and/or heart failure, or experiments with animals [2–6]. This question was addressed in the general population through heart rate measurements. Whereas some studies have found heart rate to be an independent risk factor for cardiovascular disease [7–9], others have shown the association to be dependent on confounding risk factors [10]. A particular association between heart rate and sudden death was described in three studies [7–9].

Since then, a systematic search for potentially specific risk factors of sudden death was undertaken from the cohort of asymptomatic middle-aged men enrolled in the Paris Prospective Study 1 and followed up for more than 20 years. Parental history of sudden death was disclosed as such a specific factor [11], and the particular susceptibility of diabetics to sudden death was emphasized [12]. In the present work, the role of elevated heart rate at rest was investigated as a predictor of the occurrence of sudden death in that population.

2 Methods

Details of the Paris Prospective Study 1 recruitment, design, and procedures have been described elsewhere [13]. Briefly, after their oral informed consent was obtained, the examination of 7746 native Frenchmen employed by the Paris Civil Service and aged 42–53 years, was carried out from 1967 to 1972. Subjects underwent ECG and physical examination conducted by a physician under standardized conditions (heart rate was clinically determined by measurement of the radial pulse during a 1-min recording, after a 5-min rest in supine position), provided blood samples for laboratory tests, and answered questionnaires administered by trained interviewers regarding sociodemographic factors, family and personal medical history, and smoking habits. Subjects were asked for a parental myocardial infarction, for parental age at death and whether the death was sudden or not. Diabetic status was defined as known reported diabetes whether treated or not. A further examination was proposed every year for 5 years after inclusion, during which subjects underwent a physical examination and provided blood samples for laboratory tests in the same standardized conditions. Tobacco consumption was the average consumption in the 5 years preceding the screening. The investigation conforms with the principles outlined in the Declaration of Helsinki.

2.1 Follow-up

Until retirement, the administrative department in charge of the population provided a list of deceased subjects annually. All available data relevant to the causes of death were collected from specific inquiries, i.e. medical records from hospital departments or general practitioners indicated by the relatives of the deceased. The data were then reviewed by an independent medical committee. After retirement, causes of death were obtained from death certificates. The eighth revision of the International Classification of Diseases [14] was used for coding. Sudden death (code 798.1) was defined as a natural death, occurring within 1 h of the onset of acute symptoms. Fatal myocardial infarction (code 410–414) was coded only if the death was found to be strictly related to a myocardial infarction.

The deadline of the follow-up period was January 1, 1994. The vital status could not be obtained for 355 subjects (4.6%). Men with diagnosis of ischemic heart disease (myocardial infarction, angina) established at the beginning of the study from personal medical history, clinical examination and ECG were excluded from analysis. The present analysis was conducted on the 7079 remaining subjects.

2.2 Statistical analysis

ANOVA and χ2 analysis were used for global comparisons between groups. Relative risks of mortality were adjusted for confounding factors and estimated by Cox proportional hazard model. For quantitative explanatory variables, risk ratios were reported for 1 S.D. change (standardized risk ratios: sRR). SAS procedures (Statistical Analysis System, Cary, NC, USA) were used for analysis.

3 Results

Among 7079 men followed for an average follow-up period of 23 years, there were 2083 deaths, among which 603 cardiovascular deaths distributed as follows: 118 sudden deaths (19.6%), 192 fatal myocardial infarctions (31.8%), 31 deaths from cardiac failure (5.1%), 110 deaths from other cardiac causes (18.2%), 100 deaths from stroke (16.6%) and 52 deaths from other vascular causes (8.6%).

The characteristics of the subjects are given in Table 1, and all reported parameters were significantly different between the following three groups of subjects: those who died from sudden death, from myocardial infarction, and all other subjects taken as controls, whether or not alive at the end of follow-up. In men who died suddenly, a particularly elevated rate of parental sudden death and diabetic status and low rate of sport activity could be noticed. The prevalence of ECG abnormalities was not significantly different between the three groups: frequent premature ventricular complexes were present at rest in 1.1% in those who died suddenly vs. 1.5% in those who died from myocardial infarction, and 1.2% in controls; left complete or incomplete bundle branch block was present in 1.1% vs. 0% and 0.9%; right complete or incomplete bundle branch block in 9.0% vs. 5.2 and 8.7%, and electrical left ventricular hypertrophy in 2.1% vs. 1.4 and 0.8%, respectively)

Fig. 1 shows that the crude risk of sudden death increased linearly with the level of resting heart rate, and the relative risk associated with the highest quintile of heart rate was 3.8 in comparison with the lowest quintile, whereas such a relative risk was approximatively 2 for fatal myocardial infarction, cardiovascular and total mortality (all P<0.01).

Table 2 gives the standardized univariate risk ratios for, respectively, sudden death and fatal myocardial infarction associated with each parameter reported in Table 1. Heart rate, systolic and diastolic blood pressure, tobacco consumption, cholesterol level were risk factors for both sudden death and fatal myocardial infarction. Body mass index was a risk factor for sudden death but did not reach significance for fatal myocardial infarction (P = 0.06), whereas parental sudden death, diabetic status and the absence of sport activity were risk factors for sudden death but not for fatal myocardial infarction.

View this table:
Table 1

Comparison of the characteristics at inclusion of subjects who died of sudden death or myocardial infarction during follow up and all other subjects taken as controlsa

Sudden death (n = 118)Fatal myocardial infarction (n = 192)Controls (n = 6769)P
Age (years)47.9±1.848.1±1.947.6±1.90.002
Parental sudden death (%)18.6 (n = 22)9.9 (n = 19)10.6 (n = 718)0.02
Parental myocardial infarction (%)8.5 (n = 10)12.5 (n = 24)6.5 (n = 442)0.004
Sport activity (%)7.6 (n = 9)13.5 (n = 26)14.9 (n = 1009)0.04
Diabetic status (%)7.6 (n = 9)3.6 (n = 7)2.6 (n = 175)0.003
Tobacco consumption (g/day)15.3±9.513.9±10.811.7±10.50.0001
Body mass index (kg/m2)27.2±3.526.3±3.325.9±3.30.0001
Resting heart rate (bpm)72.6±11.771.2±11.968.5±10.20.0001
Systolic blood pressure (mmHg)148±25147±24137±200.0001
Diastolic blood pressure (mmHg)87±1786±1579±130.0001
Total cholesterol (mg/dl)246±45236±51222±420.0001
  • a Data are expressed as mean±S.D. when appropriate. Sport activity applies to subjects who performed ≥1 h regular sports per week. Tobacco consumption is the average consumption (g/day) in the 5 years preceding the screening.

When age, body mass index, systolic blood pressure (or diastolic blood pressure), tobacco consumption, parental history of myocardial infarction and parental history of sudden death, cholesterol level, diabetic status, and sport activity were simultaneously entered into the survival model (Table 3), mean resting heart rate remained an independent risk factor for sudden death (sRR, 1.28; 95% CI, 1.06–1.61) but not for fatal myocardial infarction (sRR, 1.05; 95% CI, 0.95–1.16). The association between sport activity became nonsignificant whereas diabetic status (P = 0.02) and a parental history of sudden death (P = 0.007) remained independent risk factors for sudden death only. Tobacco consumption, systolic blood pressure and cholesterol, were independent risk factors for both sudden death and fatal myocardial infarction. Results were similar when diabetic subjects where excluded from the analysis. As shown in Fig. 1, the proportion of subjects who died suddenly differ proportionally with the quintiles of heart rate.

Fig. 1

Relative risks associated with quintiles of resting heart rate for total and cardiovascular mortality, sudden death and fatal myocardial infarction (quintile 1 taken as reference).

View this table:
Table 2

Standardized relative risk of sudden death and fatal myocardial infarction according to characteristics at inclusion (univariate analysis)a

Sudden deathFatal myocardial infarction
RR (95% CI)PRR (95% CI)P
Age1.13 [0.94–1.35]NS1.19 [1.02–1.39]0.02
Parental sudden death (0–1)1.95 [1.23–3.10]0.0050.97 [0.60–1.55]NS
Parental myocardial infarction (0–1)1.30 [0.68–2.49]NS2.09 [1.36–3.20]0.0008
Sport activity (0–1)0.38 [0.06–0.98]0.041.55 [0.90–2.64]NS
Diabetic status (0–1)3.23 [1.64–6.37]0.00071.47 [0.69–3.12]NS
Tobacco consumption1.79 [1.31–2.40]0.00031.55 [1.19–2.03]0.001
Body mass index1.45 [1.21–1.68]0.00011.14 [0.97–1.33]0.06
Resting heart rate1.42 [1.22–1.64]0.00011.35 [1.20–1.49]0.0001
Systolic blood pressure1.54 [1.35–1.74]0.00011.51 [1.32–1.69]0.0001
Diastolic blood pressure1.56 [1.36–1.80]0.00011.46 [1.31–1.63]0.0001
Total cholesterol1.52 [1.34–1.79]0.00011.34 [1.18–1.56]0.0001
  • a Tobacco consumption is the average consumption (g/day) in the 5 years preceding the screening. Increased risk of an event for 1 S.D. increase in variables, including age (S.D.=1.9 years), body mass index (S.D.=3.3 kg/m2), tobacco consumption (S.D.=10.5 g/day), heart rate (S.D.=10.2 bpm), systolic blood pressure (S.D.=20 mmHg), diastolic blood pressure (S.D.=13 mmHg), and cholesterol (S.D.=42 mg/dl).

In order to have a better estimate of resting heart rate in individuals, we computed the mean value of heart rate measured at five consecutive annual examinations in 3754 subjects who completed all of them. The results obtained when the 5-year mean heart rate was used in the Cox analysis for sudden death prediction were very similar to these of Table 3.

4 Discussion

After 23 years of follow-up of the cohort of the Paris Prospective Study 1, an elevated resting heart rate was confirmed as an independent risk factor for sudden death but not for fatal myocardial infarction in middle-aged men free of known cardiovascular disease, after adjustment for a large set of cardiovascular risk factors, including parental sudden death and diabetes, which were recently described as particularly associated with sudden death. Our results are consistent with these obtained from three cohort studies (from Chicago [7], from Framingham [8] and from the British Regional Heart Study [9]), which found that heart rate was independently associated with sudden death, and that its association with other cardiovascular deaths was largely explained by classical cardiovascular risk factors. Adjustment for diabetes was not made in the Chicago studies and in the British Regional Heart Study, whereas adjustment for physical activity was not made in the Framingham study and in the Chicago studies, and adjustment for body mass index was not made in the Framingham study and in the British Regional Heart Study. Parental sudden death history was included in the multivariate model solely in our study, and it remained independently associated with sudden death, suggesting that the association between heart rate and sudden death may be only partially explained by parental influence.

View this table:
Table 3

Standardized relative risk of sudden death and fatal myocardial infarction according to characteristics at inclusion (multivariate analysis)a

Sudden deathFatal myocardial infarction
RR (95% CI)PRR (95% CI)P
Age0.98 [0.81–1.18]NS1.01 [0.88–1.15]NS
Parental sudden death (0–1)2.02 [1.21–3.36]0.0070.71 [0.41–1.23]NS
Parental myocardial infarction (0–1)1.23 [0.63–2.42]NS2.60 [1.65–4.01]0.0001
Sport activity (0–1)0.43 [0.08–1.42]NS1.14 [0.90–2.12]NS
Diabetic status (0–1)2.46 [1.29–5.38]0.021.44 [0.63–3.27]NS
Tobacco consumption1.29 [1.10–1.48]0.0021.30 [1.16–1.42]0.001
Body mass index1.19 [1.02–1.68]0.040.98 [0.92–1.06]NS
Resting heart rate1.28 [1.06–1.61]0.031.05 [0.95–1.16]NS
Systolic blood pressure1.21 [1.06–1.37]0.021.16 [1.04–1.29]0.01
Total cholesterol1.25 [1.14–1.82]0.00011.16 [1.10–1.62]0.01
  • a Tobacco consumption is the average consumption (g/day) in the 5 years preceding the screening. Increased risk of an event for 1-S.D. increase in variables, including age (S.D.=1.9 years), body mass index (S.D.=3.3 kg/m2), tobacco consumption (S.D.=10.5 g/day), heart rate (S.D.=10.2 bpm), systolic blood pressure (S.D.=20 mmHg), and cholesterol (S.D.=42 mg/dl).

Cardiovascular mortality may be affected by heart rate in different ways. A high resting heart rate may increase myocardial oxygen demand, and decrease coronary blood flow by decreasing diastolic filling time [15]. It may reflect a poor physical fitness and a poor overall health. The association between elevated heart rate and sudden death may be due also to an underlying cardiomyopathy which may not be clinically detectable at an early stage but which may play a role of a confounding factor associated both with an increased heart rate at rest and the occurrence of sudden death during follow-up. Coronary disease is very common in sudden death patients after 40 years of age, and it may certainly represent the main substrate for the occurrence of sudden death [16,17]. In our study, subjects who died suddenly had many risk factors for coronary disease at inclusion in the study (higher body mass index, systolic blood pressure, tobacco consumption and cholesterol level). Other cardiomyopathies such as hypertrophic [18], dilated [3], or arrhythmogenic right ventricular [19] cardiomyopathies may be associated with an increased rate of sudden death, but due to their low prevalence in the population, they could only contribute in a small way to our results.

A high basal heart rate may also reflect abnormalities of the autonomic nervous system, and an increased sympathetic input in subjects with elevated heart rate at rest but without any underlying cardiomyopathy at inclusion in the study is highly likely. During follow-up, some of these subjects will develop coronary disease. We may suspect that a basal hyperadrenergic tone would lead to sudden death in ischemic conditions, whereas other subjects with the same ischemic conditions but without hyperadrenergic tone would suffer a myocardial infarction with less probability of ventricular arrhythmias.

Generally, adrenergic stimulation in vivo shortens ventricular refractoriness, it increases excitability and the propensity of the ventricle to develop arrhythmias. It may also produce afterdepolarizations and triggered activity. Vagal stimulation modulates these sympathetic effects; it prolongs refractoriness, reduces the effect of adrenergic stimulation on cardiac excitability, and increases electrical stability [1].

Some nonpharmacologic interventions may modify the autonomic nervous system activity. Repeated physical exercise leads to physiologic changes (including a decrease in heart rate) that are currently referred to exercise conditioning, provided that the activity is of sufficient intensity, frequency, and duration. Parasympathetic activity is increased and sympathetic activity is decreased after exercise conditioning [20]. Exercise conditioning may therefore be advised, especially in subjects at high risk of sudden death.

A basal heart rate decrease may also be obtained with drugs. The pharmacological reduction in heart rate and its consequences on sudden cardiac death following myocardial infarction has been extensively reviewed by Haverkampf and Breithardt [21]. Beta-blockers slow heart rate by inhibiting cardiac adrenergic stimulation, and patients with higher initial heart rate before therapy tend to have a greater decline. After myocardial infarction, sudden death was reduced by 30% in a total of 25 000 patients from 24 trials [22,23]. This reduction might be explained by the effect of β-blockers on ventricular fibrillation, which is the most common cause of death in the first 24 h following an acute myocardial infarction [24]. Treatment with a non-selective β-blocker (carvedilol) was associated with a decrease in mortality due to heart failure and sudden death in congestive heart failure patients, and the benefit was greater in patients with a baseline resting heart rate higher than 82 bpm [25]. Some calcium antagonists and antiarrhythmic drugs have bradycardiac effects, but they have restricted indications. In subjects with heart failure following myocardial infarction, the effect of angiotensin-converting-enzyme inhibitors in heart rate reduction and sudden death is discussed [26].

Due to adverse effects, the indication for drugs that slow heart rate must be limited to patients with an identified cardiomyopathy, and since up to 50% of sudden deaths in men occur in persons without knowledge of cardiovascular disease, drugs have a low potential for prevention of sudden death in the population.

Finally, subjects with diabetes or glucose intolerance are prone to earlier development of coronary and extra coronary macrovascular and microvascular diseases, leading to autonomic neuropathy. This could contribute to the higher risk of sudden death observed in diabetics, although a high resting heart rate and diabetic status remained independently associated with sudden death in our study.

4.1 Study limitations

The common working definition of sudden death is a natural death, occurring within 1 h of the onset of acute symptoms. The impact of the definition of sudden death on the study results has been already largely discussed in the Paris Prospective Study I [11] and the same limitations apply. Although many sudden deaths are instantaneous or unwitnessed, the elapsed time can be at least roughly estimated [27]. It is likely that we had some inaccurate coding for sudden death from death certificates, mainly because of imprecise knowledge of the elapsed time. Sensibility and positive predictive value of death certificates compared with autopsies range from 70 to 90%, and generally underestimate the number of deaths from ischemic cardiomyopathy [28–31]. However, inaccurate coding and misclassification cases are likely to be conservative and to decrease the power of our results.

The fact that heart rate is highly variable and depending on the subject's conditions should be emphasized. No threshold of elevated heart rate can be determined as abnormal for a given subject, and this restricts at present the use of elevated heart rate as a practical mean to assess the risk of sudden death in clinical practice. Further studies are needed (for example by combining basal heart rate level with other parameters related to heart rate variability) to help to identify high risk subjects in the population.

5 Conclusion

Parental sudden death [11], diabetic status [12] and elevated resting heart rate are independent risk factors particularly associated with sudden death in middle-aged men free of known cardiovascular disease. The practical use of elevated heart rate as a predictor of sudden death is probably restricted, but the heart rate may be a target for the prevention of sudden death [21]. Some form of primary prevention such as exercise conditioning, might therefore be advised in subjects identified at high risk of sudden death.

References

  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
View Abstract