If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
We aimed to describe the determinants of discharge heart rate in acute coronary syndrome patients and assess the impact of discharge heart rate on 5-year mortality in hospital survivors.
French Registry of Acute ST-Elevation or non-ST-elevation Myocardial Infarction (FAST-MI) 2005 is a nationwide French registry that included all consecutive patients with acute myocardial infarction over 1 month in 223 institutions in 2005. Discharge heart rate was recorded in 3079 patients discharged alive; all had 5-year follow-up. Logistic regression was used to detect predictors of high heart rate at discharge. Cox's proportional hazards model was used to assess the hazard ratio for mortality at 5 years. Heart rate was categorized into 4 groups by quartiles (<60, 61-67, 68-75, >75 beats per minute). High heart rate was defined as ≥75 beats per minute. Landmark analysis was performed at 1 year.
Independent predictors of heart rate ≥75 beats per minute at discharge were female sex, ST-segment elevation myocardial infarction, diabetes, chronic obstructive pulmonary disease, bleeding/transfusion during hospitalization, left ventricular dysfunction, renal dysfunction, and prescription (type, but not dose category) of beta-blockers at discharge. Discharge heart rate was significantly related to mortality at 1 year (hazard ratio 1.13; 95% confidence interval, 1.03-1.24 per 10 beats per minute, P = .02); this was confirmed by landmark analysis, with a 39% increase (hazard ratio 1.39; 95% confidence interval 1.05-1.84) in the risk of 1-year death for discharge heart rate ≥75 beats per minute vs <75 beats per minute. This relationship was no longer significant between 2 and 5 years.
After acute myocardial infarction, patients discharged with high heart rate (≥75 beats per minute) are at higher risk of death during the first year, but not later, irrespective of beta-blocker use.
We found female sex, ST-elevation myocardial infarction, diabetes, chronic obstructive pulmonary disease, bleeding/transfusion during hospitalization, left ventricular dysfunction, renal dysfunction, and prescription of beta-blockers at discharge to be independent predictors of heart rate >75 beats per minute at discharge after myocardial infarction.
Discharge heart rate was significantly related to 1-year mortality.
Patients discharged with high heart rate are at higher risk of death during the first year.
Over the last 2 decades, there has been growing interest in the use of heart rate as a marker for risk stratification in cardiovascular diseases, and as a prognostic factor for global and cardiovascular mortality in healthy subjects,
At a distance from an episode of acute coronary syndrome, the utility of controlling elevated heart rate with new drugs in addition to beta-blockers has been shown in patients with impaired left ventricular ejection fraction.
few data are available regarding the long-term impact of discharge heart rate. So far, only one publication has reported a negative impact of discharge heart rate on prognosis in patients with ST-elevation myocardial infarction (STEMI) treated with primary percutaneous intervention,
and no information is available for other types of myocardial infarction.
Therefore, the aims of this study were to describe the determinants of discharge heart rate in STEMI and non-STEMI (NSTEMI) patients and to assess the impact of discharge heart rate on 5-year mortality in hospital survivors, whatever their management.
Patients and Methods
The design of the French Registry of Acute ST-Elevation or non-ST-elevation Myocardial Infarction (FAST-MI) 2005 registry has previously been described in detail.
Briefly, the objective of this registry was to evaluate practices for myocardial infarction management in “real life” conditions, and to measure their relationship with acute and long-term outcomes of patients admitted to coronary care units for myocardial infarction in France, irrespective of the type of health care establishment to which the patients were admitted (public university teaching hospitals, public nonacademic hospitals, or private for-profit clinics). Among the 374 centers that treated patients with acute coronary syndrome, 223 (60%) participated in this registry. One physician responsible for the study was recruited at each center and data for each patient meeting the inclusion criteria during the 1-month study recruitment period were collected by dedicated research technicians sent on site for the purposes of the study. Patient care at each center was performed according to usual practice. All patients admitted in a participating center for acute myocardial infarction with symptom onset ≤48 hours were prospectively included in the registry for a 1-month period beginning in October 2005, with a 1-month extension for diabetic patients.
Previous cardiovascular and noncardiovascular history, risk factors (smoking status, hypertension or treated hypertension, dyslipidemia or treated dyslipidemia, family history, diabetes mellitus), and clinical course during hospital stay, including symptoms, Killip class, therapeutic management during the first 48 hours, during hospital stay, and at discharge, were recorded for each patient. Admission heart rate was the rate measured at admission, and discharge heart rate was the last heart rate recorded before discharge.
Impaired left ventricular function was defined as being present when measured left ventricular ejection fraction was ≤40%, or when loop diuretics were prescribed at discharge when left ventricular ejection fraction was not recorded. Beta-blockers at discharge were recorded as a binary variable (yes/no) according to 5 sub-categories: atenolol, metoprolol, acebutolol, bisoprolol, and others. The dose of beta-blockers prescribed at discharge was recorded and transformed into “low,” “intermediate,” and “maximal” dose categories, according to the tertiles of the recommended dose for each beta-blocker.
The outcome variable was all-cause mortality during 5 years of follow-up. Vital status was assessed by consulting the death registry at the registrar's offices of the patient's birthplace, or by writing to the general practitioner or cardiologist, or by direct contact with the patients or their families. Follow-up was centralized at the French Society of Cardiology and completed by a dedicated research assistant. The rate of patients lost to follow-up was 0.3% at 1 year, 2% at 3 years, and 5% at 5 years.
Baseline characteristics are described as mean ± SD for continuous variables, and absolute number and percentage for categorical variables. Groups were compared with the Kruskal-Wallis and chi-squared tests, as appropriate. Trends were tested by the Jonckheere-Terpstra test or the Cochran-Armitage test.
Heart rate is expressed as a continuous variable; or by categories, per increment of 10 beats per minute; or by groups, low vs high heart rate (≥75 beats per minute = “high heart rate” group). The relative reduction of heart rate from admission to discharge was calculated as [(admission − discharge)/admission].
The correlation between admission and discharge heart rate was tested, and the part of variance of discharge heart rate explained by admission heart rate was expressed by r2.
Logistic regression was used to detect predictors of high heart rate at discharge. Variables entered in the multivariate analysis were: type of infarction (STEMI or NSTEMI); age; sex; cardiovascular risk factors; previous infarction, stroke or heart failure; history of cancer or chronic pulmonary disease; estimated glomerular filtration rate using the Modification of Diet in Renal Disease formula; clinical conditions at admission (heart rate, systolic blood pressure, Killip Class; hemoglobin level); left ventricular (LV) dysfunction; prehospital, acute, and discharge treatments; use of coronary angiography; revascularization; and the Thrombolysis in Myocardial Infarction major bleeding during hospitalization.
To test the incremental predictive value of discharge heart rate, on top of established predictors, we compared the discriminatory capacity and net reclassification index of the Global Registry of Acute Coronary Events (GRACE) risk score (for 6-month mortality) with a model combining the GRACE risk score and discharge heart rate.
Time to event curves are presented using the Kaplan-Meier method and compared with the log-rank test. Cox's proportional hazards model, adjusted for the same variables as in the logistic regression model, was used to assess the hazard ratio for mortality at 5 years. A landmark analysis was performed at 1 year. The proportionality of hazards assumption was tested for heart rate as both a continuous and a categorical variable using Schoenfeld residuals.
Interactions between heart rate and age ≥75 years, diabetes, sex, history of chronic obstructive pulmonary disease (COPD), LV dysfunction, beta-blockers at discharge, and atrial fibrillation were tested by adding a multiplicative variable to the model.
For all tests, P < .05 was considered significant. All analyses were performed using SAS software, version 9.3 (SAS Institute Inc., Cary, NC).
Patient Characteristics at Discharge
Among the 3670 patients included in the FAST-MI registry, 3319 (94.4%) were discharged alive, and heart rate at discharge was available in 3079, of whom 220 (7%) had atrial fibrillation. The median heart rate at discharge was 66 beats per minute (interquartile range 60-75). There was a significant trend toward higher discharge heart rate with older age; female sex; diabetes; hypertension; smoking; peripheral artery disease; COPD; Killip class ≥2; heart rate at admission; LV ejection fraction; impaired LV function (Table 1); no use of beta-blockers, statins, or clopidogrel; and use of loop diuretics, with increasing quartiles of heart rate at discharge (Table 1, Table 2). There was a weak correlation between discharge heart rate and admission heart rate (correlation coefficient = 0.20, r² = 0.04) and LV ejection fraction (correlation coefficient = 0.18, r² = 0.03). The median relative reduction in heart rate was −40% (interquartile range −72%-25%).
Table 1Baseline Characteristics, Risk Factors, Previous History, and Conditions at Admission According to Discharge Heart Rate
Q1 <60 bpm (n = 892)
Q2 60-66 bpm (n = 685)
Q3 67-75 bpm (n = 792)
Q4 >75 bpm (n = 710)
P ANOVA/ Chi-Squared
66 ± 13
66 ± 14
66 ± 14
67 ± 14
Peripheral artery disease
Reperfusion therapy (among STEMI)
Admission HR (n = 2576)
76 ± 20
77 ± 19
80 ± 21
86 ± 20
SBP on admission (n = 2574)
142 ± 27
142 ± 28
144 ± 33
139 ± 30
Admission Killip class ≥2
GRACE score (n = 2507)
142 ± 34
143 ± 33
147 ± 35
155 ± 37
LVEF (%) (n = 2014)
54 ± 12
54 ± 13
52 ± 13
49 ± 14
ANOVA = analysis of variance; bpm = beats per minute; CABG = coronary artery bypass graft; CHF = congestive heart failure; COPD = chronic obstructive pulmonary disease; GRACE = Global Registry of Acute Coronary Events; HR = heart rate; LVEF = left ventricular ejection fraction; MI = myocardial infarction; PCI = percutaneous coronary intervention; Q = quartile; STEMI = ST segment elevation MI; SBP = systolic blood pressure.
Among patients treated with beta-blockers at discharge (n = 2267), bisoprolol was prescribed in 36%, atenolol in 31%, acebutolol in 21%, metoprolol in 8%, and others in the remaining 5%. Verapamil was prescribed in 87 (2.7%) of the patients, mainly in those not receiving beta-blockers. The dose of beta-blocker at discharge, available for 2231 patients (90%), was classified as low in 636 (31%), intermediate in 1130 (51%), and maximal in 418 patients (19%), respectively. Average heart rate decreased from 72 to 69, 66, and 65 beats per minute for no beta-blockers, low, intermediate, and maximal dose of beta-blockers, respectively, (P < .001 for trend). Compared with patients without beta-blockers (and taking into account the dose actually used), heart rate at discharge was reduced by 6.6 beats per minute with acebutolol, 5.4 beats per minute with atenolol, 3.7 beats per minute with metoprolol, and by 4.6 beats per minute with other beta-blockers. The relative reduction in heart rate (compared with admission) was significant for the 5 beta-blocker groups and in the group without beta-blockers. The magnitude of the reduction was greater among those with beta-blockers (range 8 to 15 beats per minute across the 5 beta-blocker groups, reduction of 8 beats per minute in patients not receiving beta-blockers or verapamil).
Incremental Predictive Capacity
The GRACE risk score has an excellent discriminatory capacity for 1-year mortality. The addition of discharge heart rate to the GRACE risk score provided a modest, albeit significant improvement in the C-statistic (Figure 1). The absolute net reclassification index was 64% (P < .001). When translated into categories (±10% mortality at 1 year), the addition of discharge heart rate to the GRACE model allowed adequate reclassification of 5% of the whole population: 4% (113/2730) among survivors and 1% (3/261) among patients who died within 1 year.
Independent Predictors of Heart Rate at Discharge
Multivariate analysis showed that predictors of discharge heart rate ≥75 beats per minute were: female sex, STEMI presentation, history of diabetes, COPD, major bleeding or transfusion during hospitalization, LV dysfunction, Killip class >2 at admission, renal dysfunction, and prescription (type, but not dose category) of beta-blockers at discharge (Figure 2).
Discharge Heart Rate and Mortality during Follow-Up
Discharge heart rate was a correlate of mortality at 1 and 5 years: at 1 year, in the lowest quartile of heart rate, mortality was 6.7%, compared with 7.7% for the second, 8.7% for the third, and 13.2% for the highest quartile (P < .001 for trend). A significant trend toward increased mortality was observed at 5 years (respectively, 20.0%, 23.1%, 25.7%, and 30.3%, for lowest to highest quartiles, P < .001 for trend) (Table 2). Univariate Kaplan-Meier curves showed a difference in outcomes according to quartiles of discharge heart rate (P < .001 by log-rank test for both 1- and 5-year mortality) (Figure 3).
Cox multivariate analysis showed that age ≥75 years, LV dysfunction, Killip class >2 at admission, history of congestive heart failure, stroke, peripheral artery disease or infarction, diabetes, cancer, hypertension, COPD, glomerular filtration rate, beta-blockers, angiotensin-converting enzyme inhibitors, and aspirin at discharge were predictors of 5-year mortality. Heart rate at discharge was related to 1-year mortality after adjustment for these variables (hazard ratio 1.13; 95% confidence interval [CI], 1.03-1.24 per 10 beats per minute, P = .02). The negative impact of high heart rate in these patients held true, mainly over the first year, regardless of whether discharge heart rate was considered continuously, by quartiles, or according to a threshold heart rate of 75 or 70 beats per minute (Table 3). This was also shown by landmark analysis, with a 39% increase (hazard ratio 1.39; 95% CI, 1.05-1.84) in the risk of 1-year death when discharge heart rate was ≥75 beats per minute vs <75 beats per minute (or an increase of 13% [hazard ratio 1.13; 95% CI, 1.04-1.24] in risk of death per increase of 10 beats per minute). This relationship was no longer significant between 2 and 5 years (Figure 4).
Table 3Association between Discharge Heart Rate and Outcomes: Hazard Ratio from Adjusted Cox Model for 1-Year Mortality, 5-Year Mortality, and 5-Year Mortality among Survivors after 1 Year
Discharge Heart Rate
Outcomes and Time
No. of Events
Continuous: per 10 bpm
>1 to 5 year mortality
Categories: ± 75 bpm (upper quartile)
>1 to 5 year mortality
Categories: ± 70 bpm
>1- to 5-year mortality
Categories: Q1 (<60 bpm) vs Q4 (>75 bpm)
>1- to 5-year mortality
Relative reduction of heart rate (per 5% increase)
>1- to 5-year mortality
Continuous: per 10 bpm patients with LV dysfunction
>1- to 5-year mortality
bpm = beats per minute; CI = confidence interval; LV = left ventricular; Q = quartile; RH = relative hazard.
Impact of Discharge Heart Rate on Mortality Across Subgroups
The impact of discharge heart rate ≥75 beats per minute on 5-year mortality held true regardless of sex, history of diabetes, age (±75 years), or use of beta-blockers at discharge. Conversely, there was a significant interaction for STEMI and LV dysfunction, and of borderline significance for history of COPD and atrial fibrillation. Thus, the negative impact of elevated heart rate was observed only in patients with LV dysfunction or with NSTEMI (Figure 5).
This is the first study to date describing the determinants of discharge heart rate after all types of acute myocardial infarction. We found elevated discharge heart rate to be associated with an increased risk of 5-year death, driven by higher mortality during the first year. These results raise the question of the determinants of discharge heart rate, the magnitude and timing of its impact on mortality, and the clinical implications.
In our study with contemporary management, average discharge heart rate was 66 beats per minute, and heart rate >75 beats per minute was observed in 1 in 4 patients. Higher discharge heart rate has been reported in previous studies,
suggesting that our population had effective beta-blocker treatment at discharge.
The determinants of discharge heart rate have not been previously studied. In FAST-MI, elevated discharge heart rate was associated with female sex, diabetes, COPD, LV dysfunction, failure to prescribe beta-blockers, and bleeding during hospitalization, but not with admission heart rate. Among the predictors, some were expected, like sex, COPD, or lack of beta-blockers. The higher heart rate in diabetic patients could be the result of autonomic cardiac dysfunction,
in line with the lesser degree of previous angina, despite more diffuse coronary lesions in this group. Similarly, the association between high discharge heart rate and in-hospital bleedings suggests persistently low hemoglobin level in these patients, although discharge hemoglobin was comparable between patients with vs without bleeding. Interestingly, a reduction in heart rate at discharge compared with admission was observed in almost all patients, including those without a beta-blocker or other heart rate-lowering treatments. Among patients with impaired LV function, in whom low heart rate has been shown to be beneficial,
This high heart rate was an independent risk marker for mortality. More recently, an association between discharge heart rate and risk of death was observed in patients treated with percutaneous coronary intervention for STEMI
Our data confirm these results and extend them to other types of myocardial infarction, mainly NSTEMI patients. Finally, the results were observed irrespective of the management strategy (noninvasive, reperfusion, or revascularization), and regardless of whether heart rate was considered as a continuous or categorical variable.
The magnitude of the association between heart rate and mortality was much lower than that previously reported. We observed a 47% increase in the risk of death at 1 year in all patients with discharge heart rate >70 beats per minute, which is far from the 3.16-fold increase reported previously in STEMI patients.
Our results indicate that heart rate at discharge had no impact on outcomes in patients with preserved LV function but was associated with an 80% increase in risk if discharge heart rate was ≥75 beats per minute, or an 18% increase per 10 beats per minute, close to the 16% reported by Fosbol et al
in patients with LV dysfunction. To date, lowering heart rate has been shown to be beneficial only in patients with LV dysfunction but not in all patients with coronary artery disease. Our data suggest that elevated heart rate has no impact in patients without heart failure.
The landmark analysis showed that the largest part of the relationship between high heart rate and mortality was observed during the first year after discharge. This temporal effect has never been studied specifically before. Indeed, only one study
had long-term follow-up, but the number of deaths was limited (n = 83/1436, 6% at 4 years) as compared with our data (n = 763/3079, 25% at 5 years), precluding any landmark analysis. The lack of relationship after 1 year does not necessarily mean that high heart rate does not have a deleterious impact, because heart rate may change over a 5-year period of time.
Study Strengths and Limitations
This study carried out in late 2005 reflects contemporary practices and outcomes in patients hospitalized for acute myocardial infarction, with 60% of all cardiac care units in France participating in the survey. Our findings are derived from observational analyses, which are subject to well-known limitations. First is the potential for confounding by measured or unmeasured variables, which cannot be completely ruled out, even after multivariable adjustment. A second limitation of the study is the lack of repeated heart rate measures during follow-up, particularly because the rate of beta-blocker use decreased over time. Lastly, the absence of any prognostic significance of discharge heart rate beyond 1 year, or in patient subsets, does not preclude a possible prognostic role over a longer period of time or with a larger population.
After acute myocardial infarction, patients discharged with high heart rate (≥ 75 beats per minute) are at higher risk of death during the first year, but not later, irrespective of beta-blocker use.
This simple clinical index should not be overlooked by physicians caring for acute myocardial infarction patients, particularly when LV dysfunction is present.
The authors thank Fiona Ecarnot (EA3920, University Hospital Jean Minjoz, Besancon, France) for translation and editorial assistance.
Paffenbarger Jr., R.S.
Heart rate and cardiovascular mortality: the Framingham Study.
Funding: FAST-MI (French Registry of Acute ST-Elevation or non-ST-elevation myocardial infarction) 2005 is a registry organized and implemented by the French Society of Cardiology. FAST-MI 2005 was supported by unrestricted grants from Pfizer and Servier , and an additional grant from the Caisse Nationale d'Assurance Maladie des Travailleurs Salariés (national social security system).
Conflict of Interest: None.
Authorship: All authors had access to the data and a role in writing the manuscript.