Original Contribution

N-Terminal pro B-type Natriuretic peptide testing for short-term prognosis in breathless older adults^B

Camille Chenevier-Gobeaux MD^c, Jean-Christophe Allo MD^d, Martine Arthaud MD^b,

R. Achkar MD^a, Yan-Eric Claessens MD, PhD^d, O.G. Ekindjian MD, PhD^d,

Bruno Riou MD, PhD^a, Patrick Ray MD^a,?

^aDepartment of Emergency Medicine and Surgery, Centre Hospitalo-Universitaire Pitie-Salpetriere, Assistance-Publique Hopitaux de Paris (AP-HP), Universite Pierre et Marie Curie Paris 6, 47-83 boulevard de l’hopital, 75013 Paris, France ^bLaboratory of Emergency Biology, Centre Hospitalo-Universitaire Pitie-Salpetriere, Assistance-Publique Hopitaux de Paris (AP-HP), Universite Pierre et Marie Curie Paris 6, 47-83 boulevard de l’hopital, 75013 Paris, France

^cDepartment of Biochemistry A, Hopital Cochin, Assistance Publique-Hopitaux de Paris (AP-HP), 27 rue du Faubourg Saint-Jacques, 75679 Paris Cedex 14, France

^dDepartment of Emergency Medicine, Hopital Cochin, Assistance Publique-Hopitaux de Paris (AP-HP), 27 rue du Faubourg Saint-Jacques, 75679 Paris Cedex 14, France

Received 18 April 2007; revised 23 August 2007; accepted 23 August 2007

Abstract

Background: Amino-terminal pro-brain natriuretic peptide (NT-proBNP) is useful for the triage of patients with dyspnea. Our aim was to determine whether Nt-proBNP levels could predict in-hospital outcome in breathless elderly patients.

Methods: At admission, NT-proBNP plasma concentrations were determined in 324 dyspneic patients aged 75 years and older. The association between NT-proBNP values and in-hospital mortality was assessed.

Results: Median NT-proBNP concentrations were not different in deceased patients (n = 43, 13%) compared to that of survivors (n = 281, 87%) (4354 vs 2499 pg/mL, respectively; P = .06). To predict in-hospital mortality, the optimum threshold of NT-proBNP was 3855 pg/mL, as defined by the receiver operating characteristic (ROC) curve, with a nonsignificant area under the ROC curve of 0.59. Mortality was significantly higher in patients (n = 139) with NT-proBNP levels 3855 pg/mL or higher (17.9% vs 9.7%, P = .045). After multivariate analysis, NT-proBNP level 3855 pg/mL or higher at admission was predictive of mortality (odds ratio, 2.41; 95% confidence interval, 1.02-5.68; P = .04).

Conclusion: NT-proBNP higher than 3855 pg/mL is associated with in-hospital mortality in patients aged 75 years and older admitted for dyspnea.

(C) 2008

^? The authors declare that they have no competing interests. The manufacturer (Roche Diagnostics) provided the diagnostic tests free of charge.

* Corresponding author. Service d’Accueil des Urgences, Groupe

Hospitalier Pitie-Salpetriere, 47-83 boulevard de l’hopital, 75013 Paris, France. Tel.: +33 1 42 17 72 49; fax: +33 1 42 17 72 64.

E-mail address: [email protected] (P. Ray).

Introduction

Congestive heart failure (CHF) is the leading reason for hospital admission among elderly patients and accounts for half of acute shortness of breath (SOB) in elderly patients [1-3]. Despite advances in treatment, patients admitted with

0735-6757/$ – see front matter (C) 2008 doi:10.1016/j.ajem.2007.08.027

decompensated heart failure present significant hospital mortality and early readmission rates [4-6].

B-type Natriuretic Peptide is a cardiac neurohor- mone secreted from membrane granules in the cardiac ventricles in response to ventricular Volume expansion and pressure overload. B-Type natriuretic peptide concentrations have been shown to be a reliable marker of CHF, even in elderly patients [4,5]. The amino-terminal fragment of BNP (NT-proBNP), which is split from pre-proBNP, also circulates in human plasma and thus can aid in diagnosing CHF with comparable accuracy [6-8]. Natriuretic peptides are of prognostic value in acute coronary syndrome (ACS)

[9] and chronic heart failure [10].

Previous studies have suggested that measuring BNP at admission to an ED might be of prognostic value in a middle- aged population [11-13]. Recently, an elevated admission BNP level was shown to be an independent predictor of in-hospital mortality in patients presenting an acute decom- pensated heart failure [14]. Despite its potential interest to emergency physicians, the prognostic relevance of NT- proBNP has been poorly investigated in the setting of elderly atients.

The aim of the study was therefore to evaluate the prognostic value of NT-proBNP concentrations as predictors of outcome in elderly patients admitted for acute SOB.

Patients and methods

This work was a post hoc analysis of data previously published in 2 studies [5,7]. Briefly, these 2 studies included patients with SOB admitted to the ED of 2 teaching hospitals and evaluated the diagnostic accuracy of both BNP and NT- proBNP. The present study was approved by the ethics committees of both institutions and waived informed consent was authorized because the routine care of patients was not modified.

The criteria for inclusion in the present study were age 75 years or older, consultation at the ED, acute SOB as the prominent presenting complaint. A final diagnosis of cardiac-related dyspnea was decided by 2 independent experts (respirologist, cardiologist, emergency physicians), including all initial findings at the ED (medical history, physical examination, electrocardiogram, chest x-ray, and blood analysis), the results of other investigations, such as echocardiography, performed during hospitalization, and the summary chart.

Usual medical care, including diagnostic workup (eg, helical thoracic computed tomography for suspected pul- monary embolism, etc), and emergency treatments (eg, nitrate and/or diuretics for CHF) decided by the emergency physician were guided by practices and guidelines [15]. The duration of hospitalization, the rates of admission to intensive care unit (ICU), and in-hospital mortality rates were recorded for all patients.

NT-proBNP measurements

At admission in the ED, NT-proBNP was drawn before any treatment. NT-proBNP was measured using a Roche Diagnostics NT-proBNP assay on an Elecsys 2010 analyzer (Roche Diagnostics, Meylan, France). Plasma samples were incubated with biotinylated polyclonal antibody plus a different polyclonal antibody labeled with a ruthenium complex. Both antibodies are directed to the NT-proBNP (amino acids 1-76 of proBNP) region. After incubation, the bound fraction was separated with streptavidin-coated microparticles and quantified by chemiluminescence. Data provided by Roche Diagnostics show that total assay accuracy ranges from 1.8% at 6781 pg/mL to 2.7% at 175 pg/mL, with a measuring range from 5 to 35000 pg/mL [8]. The biologist who performed the assays was unaware of the diagnosis of dyspnea, and the emergency physician and

experts were unaware of the NT-proBNP results.

Calculation of estimated glomerular filtration rate

Estimated glomerular filtration rate values (ml/ [min 1.73 m²]) were calculated using the Levey-modified formula: 186 ? [serum creatinine (mg/L)^-1.154] ? [age (years)^-0.203]; the calculated values were then multiplied by 0.742 for women [16].

Cardiac troponin I measurements

Cardiac troponin I measurements were decided by the physicians in charge according to clinical picture and recommendations. At admission, cTnI determinations were performed using either a Dimension Xpand HM (Dade Behring, Paris-La Defense, France; N b 0.20 ng/mL) [5] or ACS 180 assay (Bayer Diagnostics, Puteaux, France; N b 0.20 ng/mL) [7]. For statistical analyses, cTnI values were pooled and considered as positive (>=0.20 ng/mL) or negative.

Statistical analysis

Data are expressed as mean +- SD or median (and its 95% confidence interval [CI] or interquartile range) for non- Gaussian variables. Univariate comparisons between patients who died and those who survived were performed with the ?2 test for categorical variables and with the nonparametric Mann-Whitney U test for continuous vari- ables. Group comparisons of continuous variables were performed with 1-way analysis of variance. When multiple group comparisons were carried out, a Bonferroni correction was applied to the results. Fisher exact method was used to compare 2 proportions. The receiver operating characteristic (ROC) curve was used to determine the best threshold for NT-proBNP values which would be predictive of in-hospital mortality. The best threshold was that which minimized the

distance to the ideal point (sensitivity = specificity = 100%) on the ROC curve. The area under ROC curves (AUC) and its 95% CI were calculated. Sensitivity, specificity, positive and negative predictive values, and accuracy were calculated for this threshold value.

The whole study population was then stratified according to this optimum cut-off point. Finally, a forward logistic regression was performed to assess variables associated with death. For this analysis, NT-proBNP levels were evaluated as categorical variables based on the optimum cut-off point previously determined by ROC curves. We adopted a conservative approach and only included significant pre- operative variables in the univariate analysis (P value of entry b.10), age, sex, all previous medical history, PaO2, PaCO2, eGFR, cTnI, admission to the ICU, and the final diagnosis determined by experts. Interactions were not tested. Kaplan-Meier estimates of the distribution of times from baseline to death were computed, and the groups were compared using the log-rank test.

All statistical comparisons were 2-tailed and a P value of less than .05 was required to reject the null hypothesis.

Table 1 Main characteristics of patients

Statistical analysis was performed using StatView for Windows (version 5.0, Acton, MA) and GraphPad Prism for ROC curves (San Diego, CA).

Results

During the study period (18 months), 395 dyspneic patients older than 75 years consulted and underwent NT- proBNP measurement at admission. Patients were hospita- lized in situ (n = 290), in another institution (n = 34), or were directly discharged at home from the ED (n = 71, 18%). The latter were excluded from the statistical analysis. Thus, 324 patients aged 75 years or older were included and analyzed. Table 1 reports the main characteristics of the 324 patients. Eighty patients (24.7%) presented with hypoxemic (PaO2 <=55 mm Hg) and 72 patients (22.2%) with

hypercapnic (PaCO2 >=45 mm Hg) acute respiratory failure.

Fifty-three patients (16.4%) were admitted to the ICU, and the median (min-max) length of stay was 14 (1-194) days.

In-hospital death (n = 43)		Survivors (n = 281)	P value
Demographic data
Age (y)	85 +- 6	85 +- 6	.841
Women	26 (60.5%)	157 (55.8%)	.688
Chest pain	2 (6.0%)	15 (5.3%)	.64
Previous medical history
Hypertension	22 (51.6%)	150 (53.4%)	.914
Coronary artery disease	16 (37.2%)	95 (33.8%)	.791
Heart failure	7 (16.3%)	74 (26.3%)	.219
COPD	12 (27.9%)	64 (22.8%)	.585
Malignancy	7 (16.3%)	35 (12.5%)	.652
Biological results
arterial pH	7.40 +- 0.08	7.42 +- 0.06	.048
PaCO2 (mm Hg)	39.4 +- 13.1	39.7 +- 9.7	.885
PaO2 (mm Hg)	65.6 +- 17.3	67.0 +- 22.5	.679
Bicarbonates (mmol/L)	24.6 +- 6.7	25.3 +- 4.4	.387
eGFR (mL/[min 1.73 m2])	52.0 +- 23.8	53.3 +- 19.7	.690
cTnI >=0.2 ng/mL a (n = 243)	14 (43.8%)	40 (18.9%)	.003
Final diagnosis
Congestive heart failure	22 (51.2%)	149 (53.0%)	.949
Acute on chronic respiratory disease b	11 (25.6%)	77 (27.4%)	.948
Other cardiac diseases c	7 (16.3%)	36 (12.8%)	.702
Others respiratory diseases including asthma or bronchitis	2 (3.9%)	28 (10.0%)	.403
Sepsis (including infectious pneumonia, severe sepsis and septic shock)	19 (4.7%)	67 (23.8%)	.009
COPD indicates chronic obstructive pulmonary disease. Results are expressed as numbers (percentages), means (+-SD), or medians (min-max). a Only 243 patients did undergo Troponin testing (32 deceased in hospital, 211 survived). b Including COPD. c Including acute coronary syndrome, pulmonary embolism, and myopericarditis; because several causes could be present in the same patient, the percentages do not total 100%.

Mortality reached 12.9% in patients with CHF (n = 171), 11.3% in patients with acute chronic respiratory disease (n = 88), and 22.1% in patients presenting with sepsis (including infectious pneumonia, severe sepsis, and septic shock, n =

86). Patients who died (n = 43, 13.3%) presented more frequently with lower arterial pH, elevated cTnI (performed in 243 patients [75.0%]), and a final diagnosis of sepsis than patients who survived (n = 281, 86.7%) (Table 1).

Median NT-proBNP concentrations were not different in deceased patients (n = 43, 13%) compared to that of survivors (n = 281, 87%) (4354 vs 2499 pg/mL, respectively; P = .06) (Fig. 1). NT-proBNP concentrations did not differ between patients who were hospitalized in a medical department less than 7 days (n = 98), more than 7 days (n

= 173), and directly admitted to the ICU (n = 53) (511 pg/mL [864-7356] vs 5501 pg/mL [746-8299] vs 4159 pg/mL [657-

9458] respectively; P = .54). NT-proBNP concentrations were poorly sensitive and specific to predicting in-hospital mortality, as reflected by a nonsignificant AUC of 0.59 (95% CI, 0.50-0.67; P = .064) with an optimum threshold value of 3855 pg/mL (sensitivity, 60%; specificity, 58%; positive predictive value, 58%; negative predictive value, 59%; diagnostic accuracy, 59%).

Fig. 2 shows the Kaplan-Meier curve for patients who were stratified into 2 groups according to the above- determined threshold value for NT-proBNP. The in-hospital mortality was higher in patients (n = 139) with NT-proBNP levels higher than 3855 pg/mL or higher than in patients (n = 185) with NT-proBNP levels less than 3855 pg/mL

Fig. 1 Box-and-whiskers for NT-proBNP values according to outcome. Median values are indicated in the boxes.

Fig. 2 Kaplan-Meier curves showing survival according to the NT-proBNP cut-off value of 3855 pg/mL.

(17.9% vs 9.7% respectively; P = .045). Mortality in patients with positive cTnI (n = 54) was higher when patients had NT-proBNP >=3855 pg/mL than in patients with NT-proBNP less than 3855 pg/mL (32.5% vs 7.1%; P =

.082). Finally, after forward logistic regression analysis, only

3 variables remained significant predictors of mortality (Table 2): NT-proBNP level >=3855 pg/mL at admission, cTnI level >=0.2 ng/mL at admission, and sepsis as the final diagnosis.

Discussion

In our studied population (ie, breathless patients aged 75 years and older), we found that NT-proBNP level higher than 3855 pg/mL was an independent predictor of in-hospital mortality.

Previous studies had suggested that NT-proBNP levels could be a reliable predictor of poor outcome in patients with acute CHF [17-21]. In 96 patients hospitalized with acute CHF, O’Brien et al [17] reported that the AUC for NT-proBNP levels at admission was correct regarding the prediction of death or heart failure. However, it was inferior to that for predischarge NT-proBNP levels (0.70 vs 0.87; P b .05) [17]. Bayes-Genis et al [19] reported that the percentage reduction in NT-proBNP levels during admission for Acutely decompensated heart failure appeared to be the best predictor of Cardiovascular death during a long follow-up period. However, these studies evaluated serial measurements of natriuretic peptides as predictors of poor outcome and not a single measurement at admission.

The primary objective of our study was to investigate whether a single measurement of NT-proBNP performed at admission to the ED of breathless patients aged 75 years and older could predict in-hospital mortality. To our knowledge, few studies have evaluated the prognostic usefulness of

Table 2 Univariate and multivariate analysis for independent prediction of in-hospital death
		In hospital death, n (%)	Univariate analysis			Multivariate analysis
			Odds ratio (95% CI)	P value		Odds ratio (95% CI)	P value
	NT-proBNP (pg/mL)			.045			.044
	b3855 (n = 185)	18 (9.7%)	1			1
	>=3855 (n = 139) cTnI a (ng/mL)	25 (17.9%)	2.03 (1.06-3.89)	.004		2.41 (1.02-5.68)	.019
	b0.2 (n = 189)	18 (9.5%)	1			1
	>=0.2 (n = 54)	14 (25.9%)	3.33 (1.53-7.25)			2.68 (1.17-7.69)
	Sepsis			.009			.022
	No (n = 237)	24 (10.1%)	1			1
	Yes (n = 87)	19 (22.1%)	2.52 (1.30-4.88)			2.66 (1.15-6.13)
Results are expressed as numbers (percentages) or odds ratio and its 95% CI. a Cardiac troponin I results missing in 81 (25.0%) patients.

natriuretic peptide measurements at admission in the ED. Januzzi et al [20] suggested that NT-proBNP strongly predicted the likelihood of short-term mortality in subjects with acute dyspnea, with a more than 5-fold increase in the risk of death within 76 days among those with markedly elevated NT-proBNP concentrations. Our results demon- strated that, even in patients presenting with a greater age and decreased renal function (mean eGFR of 52 mL/min.m^-2]), NT-proBNP N3855 pg/mL was associated with in-hospital mortality. Gegenhuber et al [21] evaluated NT-proBNP in dyspneic patients in terms of a distinction between those who died and those survived at 1 year. The AUC value was 0.691 (95% CI, 0.630-0.747), with a cut-off NT-proBNP concen- tration of 2060 pg/mL (sensitivity, 58%; specificity, 70%). Conversely, we did not find an accurate threshold value of NT-proBNP, usable at the bedside (AUC, 0.59). It should be noted that the previous studies included middle-aged populations and excluded patients with severe renal dysfunction. That should explain the differences between our findings and previous results.

Previous studies suggested that a Multimarker approach might enable risk stratification in non-ST-elevation acute coronary syndromes [10] or CHF [11,12]. In 98 consecutive patients hospitalized for worsening chronic heart failure (mean age, 69 years), Ishii et al [12] demonstrated that Cardiac troponin T (cTnT) levels N0.033 pg/mL and BNP N440 pg/mL at admission were correlated with an incre- mental increase in mortality and future cardiac events. We, like others [12,22,23], suggested that the subgroup of patients with elevated cTnI levels and elevated NT-proBNP may have a higher mortality.

Limitations of our study

We are aware of some limitations to our study.

Because the study was observational, some data concern- ing vital clinical signs are lacking. However, the aim of this study was to investigate the prognostic rather than the diagnostic accuracy of NT-proBNP.

Although the outcome revealed increased in-hospital mortality among those with elevated NT-proBNP levels and elevated troponin, cTnI measurement was not performed in all patients, thus avoiding any strong conclusion. Further studies are warranted to confirm the usefulness of multimarker approach to predict mortality in elderly dyspneic patients.

Conclusion

Our results demonstrate that NT-proBNP level higher than 3855 pg/mL at admission was associated with in- hospital mortality in patients aged 75 years and older admitted for SOB. However, no threshold value was sufficiently accurate to be useful at the bedside.

Acknowledgments

We acknowledge the assistance of the other investigators in this study: M Bennaceur, B Madonna-Py, F Maziere, M Bendahou, O Sulkowski, Y Zhao (Department of Emergency Medicine and Surgery at the Pitie-Salpetriere Hospital), and S Voyer, C Ginsburg (Department of Emergency Medicine at the Cochin Hospital).

References

1. Rich MW. Heart failure in the 21st century: a cardiogeriatric syndrome. J Gerontol A Biol Sci Med Sci 2001;56:M88-96.
2. Executive summary of the guidelines on the diagnosis and treatment of acute heart failure: The Task Force on Acute Heart Failure of the European Society of Cardiology. Eur Heart J 2005;26:384-416.
3. Rich MW, Beckham V, Wittenberg C, et al. A multidisciplinary intervention to prevent the readmission of elderly patients with congestive heart failure. N Engl J Med 1995;333:1190-5.
4. Maisel AS, Krishnaswamy P, Nowak RM, et al. Rapid measurement of B-type natriuretic peptide in the emergency diagnosis of heart failure. N Engl J Med 2002;347:161-7.
5. Ray P, Arthaud M, Lefort Y, et al. Usefulness of B-type natriuretic peptide in elderly patients with acute dyspnea. Intensive Care Med 2004;30:2230-6.
6. Berdague P, Caffin PY, Barazer I, et al. Use of N-terminal prohormone brain natriuretic peptide assay for etiologic diagnosis of acute dyspnea in elderly patients. Am Heart J 2006;151:690-8.
7. Chenevier-Gobeaux C, Claessens YE, Voyer S, et al. Influence of renal function on N-terminal pro-brain natriuretic peptide (NT-proBNP) in patients admitted for dyspnoea in the emergency department: comparison with Brain natriuretic peptide . Clin Chim Acta 2005;361:167-75.
8. Januzzi Jr JL, Camargo CA, Anwaruddin S, et al. The N-terminal pro- BNP investigation of dyspnea in the emergency department study. Am J Cardiol 2005;95:948-54.
9. Morrow DA, de Lemos JA, Blazing MA, et al. Prognostic value of serial B-type natriuretic peptide testing during follow-up of patients with unstable coronary artery disease. JAMA 2005;294:2866-71.
10. Sabatine MS, Morrow DA, de Lemos JA, et al. Multimarker approach to risk stratification in non-ST elevation acute coronary syndromes: simultaneous assessment of troponin I, C-reactive protein, and B-type natriuretic peptide. Circulation 2002;105:1760-3.
11. Fonarow GC, Horwich TB. Combining natriuretic peptides and necrosis markers in determining prognosis in heart failure. Rev Cardiovasc Med 2003;4:S20-S208.
12. Ishii J, Nomura M, Nakamura Y, et al. Risk stratification using a combination of cardiac troponin T and brain natriuretic peptide in patients hospitalized for worsening chronic heart failure. Am J Cardiol 2002;89:691-5.
13. Harrison A, Morrison LK, Krishnaswamy P, et al. B-type natriuretic peptide predicts future cardiac events in patients presenting to the emergency department with dyspnea. Ann Emerg Med 2002;39:131-8.
14. Fonarow GC, Peacock WF, Phillips CO, et al. Admission B-type natriuretic peptide levels and in-hospital mortality in acute decom- pensated heart failure. J Am Coll Cardiol 2007;49:1943-50.
15. Hunt SA, Baker DW, Chin MH, et al. ACC/AHA guidelines for the evaluation and management of chronic heart failure in the adult:

executive summary. A report of the American College of Cardiology/ American Heart Association Task Force on Practice Guidelines (Committee to revise the 1995 Guidelines for the Evaluation and Management of Heart Failure). J Am Coll Cardiol 2001;38:2101-13.

1. Levey AS, Coresh J, Balk E, et al. National kidney foundation practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Ann Intern Med 2003;139:137-47.
2. O’Brien RJ, Squire IB, Demme B, et al. Pre-discharge, but not admission, levels of NT-proBNP predict adverse prognosis following acute LVF. Eur J Heart Fail 2003;5:499-506.
3. Richards AM, Doughty R, Nicholls MG, et al. Plasma N-terminal pro- brain natriuretic peptide and adrenomedullin: prognostic utility and prediction of benefit from carvedilol in chronic ischemic left ventricular dysfunction. Australia-New Zealand Heart Failure Group. J Am Coll Cardiol 2001;37:1781-7.
4. Bayes-Genis A, Lopez L, Zapico E, et al. NT-ProBNP reduction percentage during admission for acutely decompensated heart failure predicts long-term cardiovascular mortality. J Card Fail 2005; 11:S3-S8.
5. Januzzi JL, van Kimmenade R, Lainchbury J, et al. NT-proBNP testing for diagnosis and short-term prognosis in acute destabilized heart failure: an international Pooled analysis of 1256 patients: the International Collaborative of NT-proBNP Study. Eur Heart J 2006; 27:330-7.
6. Gegenhuber A, Mueller T, Dieplinger B, et al. B-type natriuretic peptide and amino terminal proBNP predict one-year mortality in short of breath patients independently of the baseline diagnosis of acute destabilized heart failure. Clin Chim Acta 2006;370:174-9.
7. Binder L, Pieske B, Olschewski M, et al. N-Terminal pro-brain natriuretic peptide or troponin testing followed by echocardiography for risk stratification of acute pulmonary embolism. Circulation 2005; 112:1573-9.
8. Christ M, Laule K, Klima T, et al. Multimarker strategy for risk prediction in patients presenting with acute dyspnea to the emergency department. Int J Cardiol 2007 [electronic publication ahead of print].