Heart rate at discharge in patients with acute decompensated heart failure is a predictor of mortality

Heart failure is a clinical syndrome with a poor prognosis and increasing prevalence [1]. Episodes with acute cardiac decompensation of chronic heart failure are associated with deterioration of left ventricular function and worsening of the clinical course [2]. An elevated resting heart rate is predictive for excessive morbidity and mortality for both men and women with and without cardiovascular diseases [3]. Multiple studies have confirmed this association for patients with coronary artery disease, acute myocardial infarction and heart failure [4‐6].

The deleterious effects can be explained from a mechanistic point of view: an elevated heart rate leads to ischemia, abnormal calcium handling, accelerated atherosclerosis, and increased risk of plaque rupture [7]. Furthermore, an elevated resting heart rate is associated with cardiac risk factors like high systolic blood pressure, lack of physical activity, smoking and alcohol [3]. In addition, it indicates a poor cardiac function and decreased cardiorespiratory fitness.

In heart failure, an elevated heart rate preserves cardiac output in the setting of decreased stroke volume, however, this compensatory mechanism becomes maladaptive on the long term by increasing myocardial oxygen demand and reducing coronary perfusion time. Furthermore, persistent tachycardia becomes itself involved in the development of left ventricular dysfunction or heart failure, as seen in tachycardia induced cardiomyopathy [8].

In stable chronic heart failure, a slower resting heart rate is a valuable biomarker associated with a better outcome independent of left ventricular ejection fraction when patients are in sinus rhythm [9]. Therapeutically, the beneficial role of betablocker therapy in chronic heart failure patients has been addressed to reduction in heart rate, in addition to the reduction of the incidence of arrhythmias, sudden cardiac death and providing protection from ischemia [10]. Medical therapies targeting heart rate in sinus rhythm by blocking the I_f current in pacemaker cells have additionally been proven to be advantageous in reducing cardiovascular morbidity and mortality in HFrEF patients with a heart rate ≥ 70 bpm despite optimal medical therapy in prospective multi-center studies [11, 12]. The current ESC heart failure guidelines for heart failure therefore recommend heart rate reduction with the I_f channel blocker ivabradine for symptomatic patients with chronic systolic heart failure in sinus rhythm and a heart rate of ≥ 70 bpm as an adjunct therapeutic strategy after establishment of a treatment with a maximum tolerated dose of beta-blockers, ACE-I (or ARB) and a MRA [13]. Even a co-administration strategy of ivabradine and beta-blockers during hospital admission for acute decompensation in HFrEF patients seems to be beneficial by improving systolic function, as well as functional and clinical parameters at short term [14].

While the effect of elevated heart rate for prognosis and therapy of cardiovascular morbidity and mortality is evident in chronic heart failure, data in this regard in acute heart failure are scarce and controversial. In addition, in the clinical situation of atrial fibrillation, the association of slower heart rate and improved outcome is less clear [15]. Several studies analyzing the value of heart rate on cardiovascular outcome in the context of ADHF have obtained non-conclusive results [16‐22].

In this single-center study, we sought to address the prognostic value of discharge heart rate on all-cause mortality and rehospitalization for heart failure in patients with ADHF and a reduced ejection fraction (HFrEF and HFmrEF). For this purpose, we retrospectively analyzed electronic medical records of all hospitalized patients on an all comer basis with the diagnosis of ADHF in the year 2011 who were subsequently discharged into the ambulatory care setting after successful treatment.

In this clinical investigation, we were able to demonstrate that heart rate at discharge is a predictor of mortality in patients with HFrEF and HFmrEF discharged to ambulatory care after an episode with ADHF. A heart rate ≥ 77 bpm was associated with a nearly two-fold increased mortality in this patient population. We could not detect differences for the role of this parameter in regard to the existing cardiac rhythm at discharge. In addition, this parameter was not predictive for rehospitalization.

The prognostic impact of heart rate in ADHF is still a matter of debate. In contrast to the predictive role of this biomarker in chronic systolic heart failure the role of heart rate in ADHF is much more controversial. This is partly due to differences in the time point when heart rate was measured during an acute decompensation period and focusing on different end points like in-hospital mortality and readmission in various studies [16‐22].

Risk of in-hospital mortality and particularly mortality and rehospitalization for patients hospitalized with ADHF remains high [25]. It is increased in patients with impaired metabolic status, neurohormonal activation and reduced cardiac performance, gauged by BUN, serum albumin and cholesterol levels, systolic blood pressure, heart rate, and respiratory rate [26]. Particularly, admission heart rate has been shown to be an independent risk factor for mortality during the acute phase as well as in the long term. A higher heart rate on admission was independently associated in a J-shape relationship with higher in-hospital mortality in ADHF patients with the lowest mortality seen at heart rates of 70–75 bpm [16]. Furthermore, a higher heart rate at presentation in the emergency department with ADHF was associated with an increased 7 day mortality [27]. On the other hand, higher admission heart rate can also predict survival advantage in acute HF and improve left ventricular reverse remodeling [19, 28]. However, lower heart rate is also a marker for increased in-hospital mortality in ADHF, suggesting the existence of an ideal heart rate window in these patients [18].

Some studies have correlated the difference between admission and discharge heart rate to hard cardiovascular end points [20]. Patients presenting with tachycardia and discharged with a controlled heart rate were shown to have a better outcome than those admitted non-tachycardic or discharged with a non-controlled heart rate [19]. These observations are explained by the hypothesis that elevated heart rate in the initial period of ADHF may be an indicator of preserved cardiac reserve and chronotropic competence, as the ability of the cardiovascular system to respond to this extraordinary stress situation with an adrenergic burst is preserved. Hence, heart rate recovery indicates a functioning vagal arm of the autonomous nervous system resulting in a better prognosis in patients with systolic heart failure. Nevertheless, both admission heart rate as well as heart rate difference (admission–discharge heart rate) are complicated to introduce into routine clinic algorithms for risk prediction in ADHF.

A variety of short- and long-term mortality predictors have been analyzed in patients with ADHF. Nevertheless, application of a sophisticated risk-prediction algorithm to identify patients at high risk for mortality who might benefit from aggressive monitoring and intervention using various variables suggested in the Organized Program to Initiate Lifesaving Treatment In Hospitalized Patients With Heart Failure (OPTIMIZE-HF) trial are even more cumbersome to perform in routine clinical practice [17, 25, 26].

Only few studies like ours have concentrated on heart rate at discharge for risk prediction. This parameter has several advantages: it is easy and reliably to determine, routinely available and reflects the most stable condition the patient is able to achieve and therefore may be much more relevant for the future course. Indeed, an elevated discharge heart rate is independently associated with a poor prognosis in patients revascularized with percutaneous coronary intervention for stable angina or acute coronary syndromes as well as after acute myocardial infarction [21, 29].

Our results are in line with a big retrospective cohort study of registry data showing that a higher discharge heart rate after treatment for ADHF in unselected heart failure patients is associated with an increased risk of death and rehospitalization with an even higher risk in the first 30 days after discharge [30].

The strength of our study is a high percentage of patients with guideline-based medical and device therapy and the unique long-term follow-up period with a mean duration of more than 3 years. Besides that, compared to previous studies we focused on patients with primarily systolic heart failure with an ejection fraction < 50%. The population analyzed ended up with a sicker cohort of HFrEF and HFmrEF patients with a more severely reduced mean EF. Probably due to the nearly optimal guideline directed medical therapy and the younger mean age of our study cohort, the mortality rates were noticeably lower compared to previously published studies. However, observed survival rates similar to predicted survival rates calculated with the SHFM were confirmative to exclude selection bias.

Particular differences in the study cohort might account for the differences observed in the mortality rates. In contrast to our study, the population investigated by Laskey et al. included patients who were nearly a decade older (median age 80 years) with systolic and diastolic heart failure. They had a more preserved LVEF (median 45%). In addition, this cohort included fewer ICD patients, more women and less patients with heart failure due to ischemic origin. Lastly, the follow-up period in this analysis was only 12 months. Overall, both studies underscore the positive association between discharge heart rate and mortality in patients with heart failure.

While this association is true for patients in sinus rhythm, the data in regard to Afib patients are less clear. Atrial fibrillation is not only independently associated with adverse prognosis in chronic but also in acute heart failure at least up to 1 year post-discharge [31]. Only few studies have analyzed the relationship between heart rate and mortality in ADHF patients with atrial fibrillation thereby obtaining divergent results [22, 30]. In our study, we could not detect a difference in the association of heart rate at discharge and mortality between patients presenting in SR versus those with Afib, although there was tendency that Afib was more prevalent in patients who died (p = 0.107; Table 2). A recent meta-analysis of randomized controlled trials suggests that in regard to mortality a lower heart rate in stable HFrEF patients is associated with a better prognosis only when patients were in sinus rhythm, while this association was not seen in Afib patients [32].

The optimum heart rate at discharge in respect to mortality risk for heart failure patients is not clear. We performed an exploratory analysis of the heart rate–mortality association to determine a cut-off heart rate with incremental hazard. From previous and our investigation there seems to be an upper cut-off window of 75–76 bpm. When the resting heart rate is above this rate, mortality seems to increase disproportionately. This upper cut-off window might be higher in Afib patients, although we were not able to detect significant differences in our study due to the small sample size and supposedly higher heart rate variability during Afib.

Lastly, in regard to the observed significantly higher mortality in patients taking xanthine oxidase inhibitors, indeed, in a recent systematic review and meta-analysis of uric acid-lowering agents on cardiovascular outcome in patients with heart failure treatments, allopurinol treatment was associated with a significant increase in the risk for all-cause and cardiovascular mortality [33]. Whether the observed effect of an increased mortality in patients treated with a xanthine oxidase inhibitor in our study is due to a treatment effect or elevated uric acid levels leading to medical treatment characterizes a sicker heart failure population cannot be answered with the data presented. Xanthine oxidase inhibition with allopurinol has not been demonstrated to show additional benefit nor harm in high-risk HFrEF patients with elevated uric acid levels [34].

There are a few limitations of this study. This is a single-center retrospective analysis without having access to structured registry data. Hospital factors and a rather high number of patients who had to be excluded due to incomplete chart record might have influenced the results. In addition, we had to rely on the accuracy of documentation. It cannot be fully ruled out that physiologic variables may not have been accurately recorded or been subjected to rounding which could have influenced the results. Furthermore, not all possible prognostic parameters were available in the dataset. Residual measured and unmeasured confounding factors may influence the results. Nevertheless, a considerable number of events were analyzed and we tried to exclude bias considering every patient with the diagnosis of ADHF in our study.

Taking into account heart rate at discharge in patients with ADHF and systolic heart failure may allow for greater discrimination of the highest risk patients. Further investigation should address whether heart rate reduction could be a therapeutic strategy in ADHF.