NT‐proBNP is associated with mortality and adverse cardiac events in patients with atrial fibrillation presenting to the emergency department

Abstract

Background

Atrial fibrillation (AF) is the most common cardiac arrhythmia in the emergency department. The CHA₂DS₂‐VASc score helps to predict thromboembolic risk; however, the rate of other adverse cardiac events is more difficult to predict.

Hypothesis

The biomarker N‐terminal pro‐brain natriuretic peptide (NT‐proBNP) has prognostic value in patients presenting to the emergency department with AF.

Methods

During a 1.5‐year period, a prospective study was performed in consecutive patients presenting to the emergency department with AF on the presenting electrocardiogram. At baseline, NT‐proBNP was measured. The primary endpoints were all‐cause death and major adverse cardiac events (MACE: all‐cause mortality, myocardial infarction, or revascularization).

Results

A total of 355 patients were included (mean age, 71 years; 55% male). The median duration of follow‐up was 2 years. After adjustment for baseline variables, the logNT‐proBNP was independently correlated with death (hazard ratio [HR]: 1.54, 95% confidence interval [CI]: 1.18‐1.99) and MACE (HR: 1.27, 95% CI: 1.03‐1.58). After adjustment for baseline variables, a high NT‐proBNP value (>500 pmol/L) was independently correlated with death (HR: 2.26, 95% CI: 1.19‐4.28), and for MACE a trend was seen (HR: 1.67, 95% CI: 0.96‐2.91) compared with a low value (<250 pmol/L).

Conclusions

In patients presenting to the emergency department with AF, higher NT‐proBNP values are independently associated with an increased mortality and MACE. Therefore, this biomarker may be a useful prognostic marker in the management and treatment of these patients.

1. INTRODUCTION

Atrial fibrillation (AF) is the most common arrhythmia. The prevalence is approximately 1% and is known to increase with age, with a prevalence of 9% to 15% in persons age ≥ 80 years.1, 2 Due to the aging society, AF burden will rise in the next decades. AF can cause serious complications such as heart failure (HF), cerebrovascular accidents, and death.3, 4 Therefore, therapy in patients with AF aims to reduce these complications by reducing thromboembolic risk, decreasing heart rate, and restoring sinus rhythm by chemical or electrical cardioversion. Despite these strategies, prognosis in patients with AF is variable and is strongly influenced by the underlying heart disease. The CHA₂DS₂‐VASc score helps to predict thromboembolic risk and identifies patients who need anticoagulation5; however, the occurrence of other adverse events, such as hospital admission for HF or myocardial infarction (MI), or death, is more difficult to predict.6

Cardiac biomarkers are promising tools to predict outcome in several cardiac conditions. Minor elevations in troponin I (TnI) are associated with mortality and adverse cardiac events in patients with AF.7 Natriuretic peptides, such as B‐type natriuretic peptide (BNP) and its N‐terminal fragment NT‐proBNP, have predictive value in both acute and chronic HF.8, 9 In patients without HF, natriuretic peptides predict the risk of death and cardiovascular events as well as the development of AF.10, 11 Furthermore, plasma BNP concentrations predict the risk for stroke, HF, and acute coronary syndromes (ACS) in a community‐based population of patients with AF.12 However, it is unknown whether natriuretic peptides can be used as prognostic biomarkers in patients presenting to the emergency department (ED) with AF. These patients may be in a worse condition compared with those in the outpatient clinic. Furthermore, the acute presentation is probably accompanied by higher ventricular frequencies and active comorbid conditions that may limit their predictive value. Using a prospective study design, this study tests the hypothesis that elevated NT‐proBNP levels may be predictive for adverse events in patients presenting to the ED with AF.

2. METHODS

2.1. Study population

This prospective study was performed in the Albert Schweitzer Hospital in Dordrecht, The Netherlands. The study was conducted according to the Declaration of Helsinki and approved by the local institutional review board. All consecutive patients admitted to the ED were screened for inclusion. Patients were included if AF/flutter was diagnosed on the electrocardiogram (ECG) by computer interpretation. Patients were excluded if admitted after cardiac arrest, if diagnosed with a ST‐segment elevation MI, or if they had a ventricular pacemaker rhythm.

A total of 396 patients were screened for inclusion between April 1, 2010, and August 31, 2011. In 12 patients the ECG did not show AF, and another 22 patients had missing data; thus, 362 patients were included. Seven patients were lost to follow‐up, leaving 355 patients for analysis.

2.2. Data collection

Baseline data were obtained from medical records. These included demographics, lifestyle information, medical history (including a history of cardiac events, congestive HF, and revascularization procedures), and the main diagnosis of admission to the ED as registered in the Dutch national Diagnosis Treatment Combination (DTC) database. Smoking was defined as actual smoking. Diabetes mellitus was defined as reported physician diagnosis or use of antidiabetic medication. Hypertension was defined as reported physician diagnosis or use of antihypertensive medication. Known coronary artery disease was defined as a previous MI, coronary revascularization, or a coronary angiogram with any flow‐limiting stenosis. AF was divided into 3 categories: paroxysmal AF, persistent AF, and permanent AF. The ECGs were evaluated by a trained cardiologist for the presence of AF or atrial flutter and for the ventricular response rate. X‐rays were reviewed by a radiologist for signs of congestion. Routine laboratory tests were collected, including renal function and TnI levels. Echocardiography was not routinely performed. MI was defined as described in the Universal Definition of Myocardial Infarction expert consensus document.13 Revascularization was defined as a coronary artery bypass grafting procedure or percutaneous coronary intervention. Stroke was defined as an ischemic stroke or transient ischemic attack with overt neurological symptoms.

2.3. Follow‐up and endpoints

The follow‐up period ended January 1, 2013. During follow‐up, all clinical records of outpatient visits as well as hospital admissions were reviewed. For every encounter, all available information was reviewed, such as medical history, physical examination, laboratory tests, and ECG. The primary endpoints of this study were all‐cause death and major adverse cardiac events (MACE: all‐cause mortality, MI, or revascularization). The secondary endpoints were MI, revascularization, and stroke. The follow‐up data were obtained from the hospital electronic medical record system, contact with the patients' general practitioners, or by telephone interview.

2.4. NT‐proBNP assay

At the ED, a lithium heparin anticoagulated plasma sample was obtained. NT‐proBNP concentrations were measured using a immunoassay based on luminescent oxygen channeling immunoassay (LOCI) technology according to the manufacturer's instructions.14 Because there are no uniform cutoff values to classify NT‐proBNP, we divided the values into 3 groups based upon values found in the literature correlating with high, intermediate, and low risk for adverse events and worse prognosis.15, 16 Values up to 250 pmol/L were defined as low, values between 250 and 500 pmol/L as intermediate, and values >500 pmol/L as high.

2.5. Statistical analysis

For baseline characteristics, continuous variables are presented as mean ± SD; categorical variables are expressed as numbers and percentages. Baseline characteristics between categories of NT‐proBNP were compared by means of the χ² test for categorical variables. Continuous variables were compared by analysis of variance (ANOVA). Event rates in the different groups were compared using a multivariate Cox proportional hazards model. Hazard ratios (HRs) are reported with 95% confidence intervals (CIs). Adjustments were made for age, a history of congestive HF, diabetes, hypertension, creatinine level, and TnI level on admission, the ventricular response rate, and signs of congestion on x‐ray. Event rates are graphically presented using Kaplan–Meier curves, and log‐rank tests were used to compare the different groups. A similar analysis was performed using NT‐proBNP as a continuous variable. This variable was log‐transformed, because it was not normally distributed. A P value of <0.05 was considered significant. Statistical analysis was performed using SPSS version 24 (IBM Corp., Armonk, NY). The authors had full access to the data and take responsibility for its integrity. All authors have read and agree to the manuscript written.

3. RESULTS

3.1. Baseline data

Baseline characteristics by NT‐proBNP level are displayed in the Table 1. Among the 355 patients presenting with AF, the mean age was 71 years, 55% were male, and 49% had a history of AF. In 86% a cardiac diagnosis was the reason for admission to the ED, and in 59% the main diagnosis was AF. The mean value of NT‐proBNP was 184 pmol/L. The median duration of follow‐up was 2 years. The group classified as having a low NT‐proBNP consisted of 214 (60%) patients, with 58 (16%) patients in the intermediate NT‐proBNP group and 83 (23%) patients in the high NT‐proBNP group. Patients in the high NT‐proBNP group were older and more often had a history of coronary artery disease or congestive HF. Furthermore, they had a higher creatinine level and were more likely to have signs of congestion on the x‐ray.

Table 1.

Baseline characteristics for 3 NT‐proBNP groups

	NT‐proBNP Valuea			P Value for Trend
	Low, n = 214 (60)	Intermediate, n = 58 (16)	High, n = 83 (23)
Age, y	67.3 ± 13.5	74.6 ± 12.7	77.6 ± 10.4	<0.001
Male sex	129 (60)	27 (47)	40 (48)	0.06
History of AF				0.016
Paroxysmal	41 (19)	15 (26)	13 (16)
Persistent	28 (13)	2 (3)	8 (10)
Permanent	32 (15)	14 (24)	26 (31)
History of CAD	26 (12)	12 (21)	23 (28)	0.005
History of CHF	10 (5)	9 (16)	28 (34)	<0.001
HTN	118 (55)	34 (59)	52 (63)	0.49
DM	29 (14)	16 (28)	15 (18)	0.039
Smoking	58 (27)	15 (26)	18 (22)	0.63
Ventricular response rate, bpm	121 ± 30	113 ± 30	117 ± 28	0.12
Cr, μmol/L	78.1 ± 21.3	91.0 ± 41.0	144.7 ± 112.7	<0.001
Highest Tn, ng/mL	1.04 ± 0.23	1.05 ± 0.29	1.11 ± 0.38	0.37
Congestion on x‐ray	45 (21)	21 (36)	41 (49)	<0.001

Abbreviations: AF, atrial fibrillation; CAD, coronary artery disease; CHF, congestive heart failure; Cr, creatinine; DM, diabetes mellitus; HTN, hypertension; NT‐proBNP, N‐terminal pro B‐type natriuretic peptide; SD, standard deviation; Tn, troponin.

Data are presented as n (%) or mean ± SD.

^aNT‐proBNP groups: low, <250 pmol/L; intermediate, 250–500 pmol/L; high, >500 pmol/L.

3.2. Prognostic value of NT‐proBNP

In our study population, 63 patients (18%) died during follow‐up and 83 patients (23%) experienced MACE. Ten patients had a MI, 20 were revascularized, and 19 suffered a stroke. Unadjusted, a higher absolute NT‐proBNP level was associated with higher rates of both our primary endpoints: death and MACE. After adjustment for baseline variables, the logNT‐proBNP was independently correlated with death (HR: 1.54, 95% CI: 1.18‐1.99) and MACE (HR: 1.27, 95% CI: 1.03‐1.58). Unadjusted, intermediate, and high NT‐proBNP values were both associated with death in comparison with a low value (HR: 2.86, 95% CI: 1.40‐5.85 and HR: 5.65, 95% CI: 3.16‐10.07), as it was for MACE (HR: 2.20, 95% CI: 1.22‐3.97 and HR: 3.45, 95% CI: 2.12‐5.60). After adjustment for baseline variables, a high NT‐proBNP value was independently correlated with death (HR: 2.26, 95% CI: 1.19‐4.28), and for MACE a trend was seen (HR: 1.67, 95% CI: 0.96‐2.91). The Kaplan–Meier plots for the primary study endpoints of death and MACE are shown in the Figure 1. Considering the secondary endpoints, no trends were observed between the different groups.

Figure 1.

The Kaplan–Meier plots of survival (left) and free from MACE (right) in relation to NT‐proBNP values at baseline. Low, <250 pmol/L; intermediate, 250–500 pmol/L; high, >500 pmol/L. MACE includes all‐cause mortality, MI, or revascularization. Abbreviations: MACE, major adverse cardiac events; MI, myocardial infarction; NT‐proBNP, N‐terminal pro‐brain natriuretic peptide

4. DISCUSSION

The main finding of this study is the high prognostic value of elevated NT‐proBNP in patients with AF who present to the ED. Higher NT‐proBNP values are independently associated with increased mortality and MACE. Thus, in this acute‐care setting, NT‐proBNP can provide early prognostic information in patients with AF.

Previous studies have shown that elevated NT‐proBNP levels have strong predictive value for adverse cardiac events in several populations. NT‐proBNP predicts mortality and adverse cardiac events in a community‐based sample11 and is associated with sudden cardiac death risk.17 It can predict mortality in patients with HF, especially in those with severe left ventricular dysfunction.18 In addition, in specific situations such as ACS and stable angina, increased levels of NT‐proBNP have been associated with increased mortality.19, 20 In patients with AF, the predictive value of NT‐proBNP has also been demonstrated. The biomarker substudies of the Randomized Evaluation of Long‐Term Anticoagulation Therapy (RE‐LY) and the Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation (ARISTOTLE) trials showed a strong relationship between NT‐proBNP and mortality and other adverse cardiac events in 2 large study populations.21, 22 However, these populations were not exclusively included in the ED.

In the ED, natriuretic peptides have been shown to be of predictive value in different conditions. In patients with chest pain and ACS presenting to the ED, a higher BNP was a powerful marker for cardiac mortality.23 Furthermore, in patients with acute decompensated HF, higher values of BNP and NT‐proBNP predict mortality.24 Finally, in a nonselected population presenting to the ED, higher NT‐proBNP values were related to in‐hospital and long‐term mortality.25

To our knowledge, this is the first study to investigate the prognostic value of NT‐proBNP in a nonselected group of patients presenting with AF to the ED. Because of the markedly different setting, our study adds important information to the existing body of literature in several ways. In our study, high NT‐proBNP values were seen, as well as high mortality rates. The median NT‐proBNP value was 184 pmol/L, and a mortality of 18% was observed during a follow‐up of 2 years. In comparison, in the RE‐LY and the ARISTOTLE substudies, median NT‐proBNP values of 94 pmol/L and 84 pmol/L, respectively, were seen, and the mortality was 3.3% and 3.7% per year, respectively. The main difference between these 2 studies and our study is a different population. In the RE‐LY and the ARISTOTLE substudies, the population was mainly included in the outpatient clinic, and during inclusion not every patient had AF on the ECG. In contrast, in our study all patients were included in the ED presenting acutely with relatively high ventricular rates, higher stress levels, and active comorbid conditions. Furthermore, the patients in the present study were older and were only included if AF was seen on the presenting ECG. Therefore, the patients in our study probably represent a more diseased population and may need a more aggressive treatment and more intensive follow‐up.

In HF, elevated plasma levels of natriuretic peptides are caused by increased secretion by myocytes as a reaction to wall tension, both on a ventricular and atrial level.26, 27 In AF, the mechanism of increased atrial and ventricular wall tension is probably due to higher wedge pressures as a result of absence of atrial contraction, irregularity of the ventricular contractions, and an increase in left atrial size.28 In our study, a correlation was observed between signs of congestion on the x‐ray and a higher NT‐proBNP category, as well as a history of previous HF. However, factors other than HF may play a role. It has been shown that levels of natriuretic peptides are increased in the absence of signs of HF,29 implying a direct influence of the arrhythmia on the release of NT‐proBNP by myocytes. AF causes variable filling times and irregular atrial contractions, causing local stretch at a cellular level that may cause release of natriuretic peptides such as NT‐proBNP.30 Because it is known that plasma levels of NT‐proBNP decrease in patients with paroxysmal or persisting AF after restoration of sinus rhythm,31, 32 and because acute paroxysms increase BNP levels,33 these effects appear dynamic in nature.

Natriuretic peptides have natriuretic, diuretic, and hypotensive effects and also inhibit sympathetic activity, endothelin secretion, and the renin‐angiotensin system.34 These mechanisms suggest that NT‐proBNP is an important signaling hormone of a deprived cardiac state. Many of the patients in this study may have preclinical cardiac disorders such as borderline abnormalities in intracardiac pressure, mild left ventricular dysfunction, or minor valvular dysfunction. It is known that plasma BNP can be elevated in subjects with subclinical structural heart disease.35 We therefore speculate that the elevated levels of plasma BNP in the patients with worse outcomes denote mild structural heart disease such as subclinical cardiac dysfunction, including mildly elevated intracardiac pressure and volume and, possibly, myocardial ischemia. These individuals might be prone to be at risk for HF and coronary heart disease.

NT‐proBNP is a strong predictor of adverse events and can therefore be used as a risk marker in patients with AF. Advantages in daily practice include the ease of obtaining results and its dynamic nature correlating with the state of the disease. More aggressive therapy for HF patients using BNP levels as guidance has been shown to reduce mortality.36 It may therefore be considered to use this biomarker in a similar way in patients with AF to initiate more aggressive therapies or a more intense follow‐up. This may particularly be of value in patients with new‐onset AF and no prior cardiac history, or in asymptomatic AF patients, 2 groups that might not be as aggressively managed. It would be a useful way to separate these patients into high‐ and low‐risk groups and therefore guide the clinician's approach in management. Therapies aimed at reducing wall stress might be beneficial, such as angiotensin‐converting enzyme inhibition, angiotensin II inhibitors, or aldosterone antagonists.

Despite the clear prognostic value of BNP in several studies, the European Society of Cardiology (ESC) and American Heart Association (AHA) have not adopted the routine use of biomarkers in their guidelines for follow‐up of patients with AF. This is probably due to the lack of large randomized controlled studies. Therefore, more prospective studies are required to assess the value of NT‐proBNP–guided treatment. Because AF is the most common arrhythmia in patients presenting to the ED, routine use of this biomarker in the ED might be of value.

4.1. Study limitations

Because the study population was included in the ED, results cannot be extrapolated to all AF patients, particularly not to patients in the outpatient clinic. Furthermore, our study population is relatively small and confined to a single hospital location. Therefore, larger datasets are needed to validate the NT‐proBNP cutoff values for generalizability. Results need to be interpreted in light of these limitations and the findings should be confirmed in other cohorts.

5. CONCLUSION

In patients presenting to the ED with AF, higher NT‐proBNP values are independently associated with an increased mortality and MACE. Therefore, this biomarker may be a useful prognostic marker in the management and treatment of these patients.

Conflicts of interest

The authors declare no potential conflicts of interest.

Holl MJ, van den Bos EJ, van Domburg RT, Fouraux MA, Kofflard MJ. NT‐proBNP is associated with mortality and adverse cardiac events in patients with atrial fibrillation presenting to the emergency department. Clin Cardiol. 2018;41:400–405. 10.1002/clc.22883