The new england journal of medicine - ugap.fr file february 12, 2004 647 The new england journal of...

n engl j med

350;7

www.nejm.org february

12, 2004

647

The

new england

journal

of

medicine

established in 1812

february

12

,

2004

vol. 350 no. 7

Use of B-Type Natriuretic Peptide in the Evaluation and Management of Acute Dyspnea

Christian Mueller, M.D., André Scholer, Ph.D., Kirsten Laule-Kilian, B.Sc., Benedict Martina, M.D., Christian Schindler, Ph.D., Peter Buser, M.D., Matthias Pfisterer, M.D.,

and André P. Perruchoud, M.D.

abstract

From the Department of Internal Medicine,Medical Division A (C.M., K.L.-K., A.P.P.),the Department of Laboratory Medicine(A.S.), the Emergency Department (B.M.),the Institute for Social and PreventiveMedicine (C.S.), and the Division of Cardi-ology (P.B., M.P.), University of Basel,University Hospital, Basel, Switzerland.Address reprint requests to Dr. Mueller atMedizinische Klinik A, Universitätsklinik,Petersgraben 4, CH-4031 Basel, Switzer-land, or at [email protected].

N Engl J Med 2004;350:647-54.

Copyright © 2004 Massachusetts Medical Society.

background

B-type natriuretic peptide levels are higher in patients with congestive heart failurethan in patients with dyspnea from other causes.

methods

We conducted a prospective, randomized, controlled study of 452 patients who pre-sented to the emergency department with acute dyspnea: 225 patients were randomlyassigned to a diagnostic strategy involving the measurement of B-type natriuretic pep-tide levels with the use of a rapid bedside assay, and 227 were assessed in a standardmanner. The time to discharge and the total cost of treatment were the primary endpoints.

results

Base-line demographic and clinical characteristics were well matched between the twogroups. The use of B-type natriuretic peptide levels reduced the need for hospitaliza-tion and intensive care; 75 percent of patients in the B-type natriuretic peptide groupwere hospitalized, as compared with 85 percent of patients in the control group(P=0.008), and 15 percent of those in the B-type natriuretic peptide group required in-tensive care, as compared with 24 percent of those in the control group (P=0.01). Themedian time to discharge was 8.0 days in the B-type natriuretic peptide group and 11.0days in the control group (P=0.001). The mean total cost of treatment was $5,410 (95percent confidence interval, $4,516 to $6,304) in the B-type natriuretic peptide group,as compared with $7,264 (95 percent confidence interval, $6,301 to $8,227) in the con-trol group (P=0.006). The respective 30-day mortality rates were 10 percent and 12percent (P=0.45).

conclusions

Used in conjunction with other clinical information, rapid measurement of B-type natri-uretic peptide in the emergency department improved the evaluation and treatment ofpatients with acute dyspnea and thereby reduced the time to discharge and the totalcost of treatment.

n engl j med

350;7


12

,

2004

The

new england journal

of

medicine

648

eart failure is a major public

health problem. Currently, more than 15million patients have heart failure in

North America and Europe, with nearly 1.5 millionnew cases every year.

1-5

Heart failure is the mostfrequent cause of hospitalization among peopleolder than 65 years of age, and these hospitaliza-tions are an important part of the enormous cost ofthe disease. Over the past decade, the rate of hos-pitalization for heart failure has increased by 159percent.

3

It is estimated that in the United States in2001, the direct cost of the care of patients withheart failure exceeded $24 billion.

3

Therefore, cost-effective management is of paramount importance.However, the rapid and accurate differentiation ofheart failure from other causes of dyspnea remainsa clinical challenge, especially in the emergency de-partment.

4-10

After evaluating a patient’s symptoms,conducting a physical examination, and perform-ing electrocardiography and chest radiography, theclinician is often left with considerable diagnosticuncertainty, which results in misdiagnosis and de-lays the initiation of appropriate therapy.

7,10

In addi-tion, the misdiagnosis of heart failure causes mor-bidity and increases the time to discharge and thecost of treatment, because the use of a treatmentstrategy for other conditions, such as chronic ob-structive pulmonary disease, may be hazardous topatients with heart failure, and vice versa.

6,8,9

Observational studies have suggested that, whenused in conjunction with other clinical information,B-type natriuretic peptide levels may be useful inestablishing or ruling out the diagnosis of heart fail-ure in patients with acute dyspnea.

6,7,10-14

B-typenatriuretic peptide is a 32-amino-acid polypep-tide secreted by the cardiac ventricles in responseto ventricular volume expansion and pressure over-load.

15-17

The levels of B-type natriuretic peptideare elevated in patients with left ventricular dysfunc-tion, and the levels correlate with both the severityof symptoms and the prognosis.

3,6,7,9-19

However,the clinical effect of this diagnostic test on the eval-uation and treatment, outcome, and cost of treat-ment of patients with dyspnea is unknown. There-fore, we performed a randomized, controlled trial totest the hypothesis that a diagnostic strategy guid-ed by the rapid measurement of B-type natriureticpeptide levels would improve the evaluation andcare of patients with acute dyspnea who present tothe emergency department and would thereby re-duce the time to discharge and the total cost oftreatment.

setting and study population

The B-Type Natriuretic Peptide for Acute Shortnessof Breath Evaluation (BASEL) Study was a prospec-tive, randomized, controlled, single-blind study con-ducted in the emergency department of the Univer-sity Hospital in Basel, Switzerland. Patients wereevaluated in the emergency department by at leasttwo physicians: a resident in internal medicine andan internal-medicine specialist. The study investi-gators were not directly involved in patient care inthe emergency department, nor did they have any in-fluence on the decision to discharge patients fromthe ward. The study was carried out according tothe principles of the Declaration of Helsinki andapproved by the local ethics committee. Written in-formed consent was obtained from all participat-ing patients.

We screened 665 consecutive adults who pre-sented to the emergency department between May2001 and April 2002. Eligible patients were thosewho had acute dyspnea as the primary symptom,with no obvious traumatic cause of dyspnea. Pa-tients with severe renal disease (defined by a serumcreatinine level of more than 250 µmol per liter[2.8 mg per deciliter]), patients with cardiogenicshock, and patients who requested an early trans-fer to another hospital were excluded. There wereno limitations to entry according to the time of dayat which patients arrived in the emergency depart-ment or the availability of research staff.

A total of 452 patients were enrolled in the trial,and group assignment was accomplished with theuse of a computer-generated randomization schemein a 1:1 ratio without stratification. A total of 225patients were randomly assigned to be evaluatedwith the use of a diagnostic strategy that includedthe rapid bedside measurement of B-type natriuret-ic peptide levels, and 227 were assigned to be eval-uated with the use of the conventional diagnosticstrategy. The B-type natriuretic peptide was not mea-sured for clinical purposes by clinicians who treat-ed patients in the control group, nor was it mea-sured serially in either of the groups.

routine clinical assessment

All patients underwent an initial clinical assess-ment that, in general, included a clinical history tak-ing, a physical examination, electrocardiography,pulse oximetry, blood tests, and chest radiography.Echocardiography and pulmonary-function tests

h

methods

n engl j med

350;7


12, 2004

b-type natriuretic peptide in acute dyspnea

649

were strongly recommended on an outpatient ba-sis for patients who were released from the emer-gency department, as well as for the patients whowere admitted.

measurement and interpretation of b-type natriuretic peptide levels

During the initial evaluation, at the time of veni-puncture for routine blood tests, a 5-ml specimenof venous blood was collected in tubes containingpotassium EDTA. During a 15-minute period, B-typenatriuretic peptide was measured with the use ofa rapid fluorescence immunoassay (Biosite Diag-nostics). The precision, analytic sensitivity, and sta-bility of the system have been described previous-ly.

14,18

In brief, the coefficient of variation within agiven assay has been reported to be 9.5 percent,12.0 percent, and 13.9 percent for levels of 28.8,584.0, and 1180.0 pg per milliliter, respectively, andthe coefficient of variation among assays is knownto be 10.0 percent, 12.4 percent, and 14.8 percent,respectively.

14,18

The limit of analytic sensitivity wasless than 5.0 pg per milliliter, with a measurablerange of 0 to 1300 pg per milliliter.

In the group in which B-type natriuretic peptidelevels were measured, diagnostic and therapeuticdecisions were not based on the B-type natriureticpeptide levels alone; instead, this information wasconsidered in the context of the other clinical infor-mation obtained and the physicians’ clinical im-pressions, as previously described.

20

In brief, weused a B-type natriuretic peptide level of 100 pg permilliliter to separate dyspnea caused by heart fail-ure from other causes of dyspnea.

3,6,7,9-13,20

In pa-tients with a B-type natriuretic peptide level below100 pg per milliliter, the diagnosis of heart failurewas considered unlikely, and alternative causes ofdyspnea had to be investigated. In patients with aB-type natriuretic peptide level of more than 500 pgper milliliter, heart failure was considered the mostlikely diagnosis, and rapid therapy with diuretics,nitroglycerin, angiotensin-converting–enzyme in-hibitors, and morphine was recommended. For pa-tients with B-type natriuretic peptide levels between100 and 500 pg per milliliter, the protocol recom-mended the use of clinical judgment and possiblefurther diagnostic testing to rule out stable base-line left ventricular dysfunction and other condi-tions as the real cause of acute dyspnea.

3,6,7,9-14,20

No formal adjustment was recommended regard-ing the B-type natriuretic peptide cutoff values inpatients with mild chronic kidney disease. Patients

in the control group were evaluated and treated ac-cording to the most recent clinical guidelines.

4,5

end points

The time to discharge and the cost of treatmentwere the primary end points of the study. Second-ary end points included in-hospital and 30-daymortality. The time to discharge was defined as theinterval from presentation at the emergency depart-ment to discharge. Patients who died in the hospi-tal were excluded from the calculation of this endpoint. Since ratios of costs to charges have not beendefined for the majority of services and departmentsat our institution, hospital charges were used asthe most appropriate estimate of the true costs.

21,22

To avoid an imbalance owing to differences in re-imbursement or charges associated with differenttypes or classes of insurance, charges were stan-dardized according to the actual rates for patientswith general insurance who were living in Basel. Thecurrent reimbursement for the measurement ofB-type natriuretic peptide in Switzerland ($47) wasused. The time to treatment was defined as the in-terval from presentation to the initiation of the ap-propriate therapy — other than bed rest and supple-mental oxygen — according to the final dischargediagnosis. This therapy included diuretics or vaso-dilators in patients with heart failure, anticoagulantsin patients with pulmonary embolism, and inhaledbronchodilators or systemic corticosteroids in pa-tients with an exacerbation of obstructive pulmo-nary disease. All end points were assessed in a blind-ed fashion by physicians who were not involved inpatient care, with the use of all medical records per-taining to each patient.

statistical analysis

The statistical analyses were performed with theuse of the SPSS/PC software package (version 11.0,SPSS). A P value of less than 0.05 was considered toindicate statistical significance. All data were ana-lyzed according to the intention-to-treat principle.Comparisons were made with the use of the t-test,the Mann–Whitney U test, Fisher’s exact test, or thechi-square test, as appropriate. All hypothesis test-ing was two-tailed. The trial was designed to enroll222 patients in each group. This number providedthe study with a power of 80 percent to detect a re-duction in the time to discharge from 10.0 to 8.0days (20 percent) with the use of the diagnosticstrategy guided by measurement of the B-type na-triuretic peptide level. Assumptions included the

n engl j med

350;7


12

,

2004

The

new england journal

of

medicine

650

Table 1. Base-Line Characteristics of the Patients.

CharacteristicB-Type Natriuretic Peptide Group

(N=225)Control Group

(N=227)

Age — yrMean95% Confidence intervalRange

70.368.2–72.4

19–97

70.868.9–72.7

21–96

Sex — no. (%)MaleFemale

132 (59)93 (41)

130 (57)97 (43)

Smoking status — no. (%)Never smokedCurrent smokerPrevious smoker

101 (45)64 (28)60 (27)

114 (50)45 (20)68 (30)

Medical history — no. (%)

Coronary artery disease 113 (50) 112 (49)

Arterial hypertension 113 (50) 124 (55)

Diabetes mellitus 47 (21) 56 (25)

Chronic obstructive pulmonary disease 75 (33) 65 (29)

Asthma 17 (8) 12 (5)

Pneumonia 30 (13) 28 (12)

Pulmonary embolism 18 (8) 13 (6)

Other pulmonary or pleural disease 20 (9) 26 (11)

Any pulmonary disease 119 (53) 107 (47)

Depressive disorder 15 (7) 21 (9)

Stroke or peripheral vascular disease 40 (18) 49 (22)

Chronic kidney disease 56 (25) 56 (25)

Deep-vein thrombosis 19 (8) 22 (10)

Symptoms — no. (%)

Shortness of breath*While walking up a slight inclineWhile walking on level groundAt rest

32 (14)125 (56)66 (29)

33 (15)132 (58)58 (26)

Paroxysmal nocturnal dyspnea 79 (35) 87 (38)

Nocturia 60 (27) 76 (33)

Chest pain 76 (34) 78 (34)

Coughing 101 (45) 123 (54)

Expectoration 72 (32) 87 (38)

Fever 59 (26) 50 (22)

Systolic blood pressure — mm HgMean95% Confidence interval

146142–150

145141–149

Diastolic blood pressure — mm HgMean95% Confidence interval

8583–87

8683–89

Heart rate — beats/minMean95% Confidence interval

9693–99

9996–103

Temperature — °CMean95% Confidence interval

37.637.2–38.1

37.437.2–37.5

n engl j med

350;7


12, 2004


651

use of a two-tailed test, a 5 percent level of signifi-cance, and a standard deviation of 7.5 days in bothgroups.

A total of 452 patients were enrolled. The base-linecharacteristics were well matched between thestudy groups (Table 1). The mean age was 71 years.In both groups, slightly more than 40 percent of the

patients were women. The medical history includedcoronary artery disease in 50 percent of patients, hy-pertension in 52 percent, chronic obstructive pul-monary disease in 31 percent, any pulmonary dis-ease in 50 percent, and diabetes in 23 percent.

The median time from presentation at the emer-gency department to the initiation of the appropri-ate therapy according to the final discharge diag-nosis was 90 minutes in the control group and 63minutes in the B-type natriuretic peptide group

results

* Four patients in the B-type natriuretic peptide group and two patients in the control group had shortness of breath only while walking up a steep incline.

† To convert values for creatinine to milligrams per deciliter, divide by 88.4.

‡ ACE denotes angiotensin-converting–enzyme inhibitor or angiotensin-receptor blocker.

Table 1. (Continued.)

CharacteristicB-Type Natriuretic Peptide Group


(N=227)

Signs — no. (%)

Tachypnea (>20 breaths/min) 106 (47) 104 (46)

Elevated jugular venous pressure 32 (14) 32 (14)

Hepatojugular reflux 25 (11) 24 (11)

Rales 103 (46) 104 (46)

Wheezing 55 (24) 45 (20)

Hyperresonant percussion 22 (10) 17 (7)

Dullness 20 (9) 26 (11)

S

3

gallop 4 (2) 2 (1)

Cyanosis 14 (6) 19 (8)

Lower-extremity edema 73 (32) 83 (37)

Hemoglobin — g/dlMean95% Confidence interval

13.513.3–13.8

13.413.0–13.8

Serum creatinine — µmol/liter†Mean95% Confidence interval

113105–121

116109–123

Serum albumin — g/literMean95% Confidence interval

3433–34

3332–34

Medications — no. (%)

Beta-blockers 51 (23) 57 (25)

Diuretics 118 (52) 103 (45)

Nitrates 32 (14) 30 (13)

ACE inhibitors‡ 91 (40) 86 (38)

Digoxin 21 (9) 24 (11)

Aspirin 73 (32) 74 (33)

Amiodarone 19 (8) 18 (8)

Calcium-channel blockers 26 (12) 37 (16)

Inhaled bronchodilators 47 (21) 37 (16)

Inhaled corticosteroids 34 (15) 27 (12)

Oral corticosteroids 36 (16) 24 (11)

Anticoagulants 56 (25) 51 (22)

n engl j med

350;7


12

,

2004

The

new england journal

of

medicine

652

(P=0.03) (Table 2). The use of B-type natriureticpeptide levels significantly reduced the need forhospitalization and intensive care: 75 percent of pa-tients in the B-type natriuretic peptide group werehospitalized, as compared with 85 percent of thosein the control group (P=0.008), and 15 percent ofpatients in the B-type natriuretic peptide group re-quired intensive care, as compared with 24 percentof those in the control group (P=0.01).

There was a considerable range in the time todischarge, reflecting the variety of diseases respon-sible for acute dyspnea. As shown in Figure 1, thetime to discharge was significantly shorter in theB-type natriuretic peptide group (median, 8.0 days)than in the control group (median, 11.0 days; P=0.001). This difference translated into a significantdifference in the mean total cost of treatment:$7,264 in the control group, as compared with$5,410 in the B-type natriuretic peptide group (P=0.006). Twenty-one patients (9 percent) in the con-trol group died in the hospital, as compared with13 patients in the B-type natriuretic peptide group(6 percent, P=0.19).

Heart failure was the final discharge diagnosisin 45 percent of patients in the B-type natriureticpeptide group and 51 percent of patients in the con-trol group (P=0.2). Exacerbation of obstructive pul-monary disease was more commonly the cause ofacute dyspnea in the B-type natriuretic peptidegroup than in the control group (23 percent vs. 11percent, P=0.001). The treating physician indicat-ed that two causes contributed to the acute dyspneain 11 patients (5 percent) in the B-type natriureticpeptide group and in 10 patients (4 percent) in thecontrol group (P=0.81).

Clinical 30-day follow-up data were available forall patients. The rates of readmission and mortalitywithin 30 days after discharge were similarly low inthe two groups (Table 2). The 30-day mortality ratewas 10 percent in the B-type natriuretic peptidegroup and 12 percent in the control group. Amongpatients who were not initially admitted, rates of sec-ondary admission were 5 percent in the B-type natri-uretic peptide group (3 of 56 patients) and 9 per-cent in the control group (3 of 34 patients, P=0.67),and the respective 30-day mortality rates were 4 per-cent (2 deaths) and 3 percent (1 death) (P=1.00).

This randomized, controlled trial examined the ef-fect of the measurement of B-type natriuretic pep-tide levels in the emergency diagnosis of patientswith acute dyspnea. The use of B-type natriureticpeptide levels in conjunction with other clinical in-formation reduced the time to the initiation of the

discussion

* The time to treatment was defined as the interval from presentation at the emergency department to the initiation of the appropriate therapy according to the final discharge diagnosis.

† The Mann–Whitney U test was used.

‡ Fisher’s exact test was used.

Table 2. End Points.*

End Point

B-Type NatriureticPeptide Group


(N=227) P Value

Time to treatment — minMedianInterquartile range

6316–153

9020–205

0.03†

Time to discharge — daysMedianInterquartile range

8.01.0–16.0

11.05.0–18.0

0.001†

Hospitalization — no. (%) 169 (75) 193 (85) 0.008

Admission to intensive care — no. (%)

33 (15) 54 (24) 0.01

Cost of intensive care — $Median95% Confidence interval

874423–1,324

1,516989–2,043

0.07

Total treatment cost — $Median95% Confidence interval

5,4104,516–6,304

7,2646,301–8,227

0.006

In-hospital mortality — no. (%) 13 (6) 21 (9) 0.21‡

30-day mortality — no. (%) 22 (10) 28 (12) 0.45‡

30-day readmission rate —no. (%)

26 (12) 23 (10) 0.63

Figure 1. Cumulative Frequency Distribution Curve for the Time to Discharge of Patients in the B-Type Natriuretic Peptide Group as Compared with Those in the Control Group.

Cum

ulat

ive

Perc

ent

75

50

25

00 10 20 30 40 50 60

Time to Discharge (days)

Controlgroup

B-typenatriuretic

peptidegroup

P=0.001

100

n engl j med

350;7


12, 2004


653

most appropriate therapy, the need for hospitaliza-tion and intensive care, the time to discharge, andthe total cost of treatment. Given the morbidityassociated with acute dyspnea and the cost associ-ated with heart failure, chronic obstructive pulmo-nary disease, and other disorders that cause dys-pnea,

1-5,8,9,14

B-type natriuretic peptide testing islikely to be of value in the treatment of patientswith acute dyspnea. The mean total cost of treat-ment in this study was similar to the expendituresin the United States. In 1997, an estimated $5,501was spent for every hospital-discharge diagnosis ofheart failure.

3

B-type natriuretic peptide testing reduced thetotal cost of treatment by 26 percent. This findingis supported by a retrospective analysis of the costeffectiveness of the use of B-type natriuretic pep-tide levels in screening for left ventricular systolicdysfunction in the general population, which alsoshowed a 26 percent reduction in cost.

23

Our findings extend the conclusions of observa-tional studies in which the use of the measurementof B-type natriuretic peptide levels was validated bycomparison with a retrospectively adjudicated di-agnosis of heart failure by independent cardiol-ogists.

3,6,7,9-14

In the largest of these studies —the Breathing Not Properly Multinational Study— B-type natriuretic peptide levels by themselveswere more accurate than any historical or physicalfinding or laboratory value in identifying heart fail-ure as the cause of dyspnea. The diagnostic accura-cy of B-type natriuretic peptide at a cutoff value of100 pg per milliliter was 83 percent, with a sensitiv-ity of 90 percent and a specificity of 76 percent.

6,7

In our study, exacerbation of obstructive pulmo-nary disease was more often the cause of acute dys-pnea in the B-type natriuretic peptide group than inthe control group. This finding corresponds wellwith a recent observation that exacerbation of chron-ic obstructive pulmonary disease frequently escapesrecognition in the emergency department

24

and isalso in agreement with the high negative predic-tive value of the B-type natriuretic peptide level forthe diagnosis of heart failure. B-type natriureticpeptide levels below 100 pg per milliliter in a pa-tient with acute dyspnea make the diagnosis ofheart failure very unlikely and apparently help cli-nicians focus on the most common alternative di-agnosis. Obstructive pulmonary disease was presentin one third of our patients and may well have goneunrecognized as the cause of acute dyspnea in a

considerable number of patients in the controlgroup. Our findings, along with those of other in-vestigators, provide support for the inclusion ofthe measurement of natriuretic peptides in the re-cent European guidelines for the diagnosis of heartfailure.

5

A particular strength of our study is that thestudy population was highly representative of theelderly population of patients with heart failure inclinical practice.

1-3

The mean age was 71 years,nearly half the patients were women, and coexist-ing conditions were common. The rapid and accu-rate differentiation of heart failure from other caus-es of acute dyspnea in such patients is often difficult,although essential for cost-effective management.The symptoms and signs of heart failure are nei-ther sensitive nor specific and considerably overlapthose of pulmonary disease.

4-7,13

The approach tothe emergency diagnosis of acute dyspnea has beenfundamentally unchanged for decades and has beencomplemented by electrocardiography, chest radi-ography, and echocardiography for the assessmentof left ventricular function. Unfortunately, thesemethods have important limitations.

25-27

The clinical experience with B-type natriureticpeptide testing is limited. Our interpretation of thetest results was based on the data available whenthe study protocol was devised. Further studiesshould help to optimize the use of B-type natriureticpeptide measurements in clinical practice. The useof normal values corrected for age and sex may rep-resent a clinically significant advance, since the lev-els of B-type natriuretic peptide increase with ageand are higher in women than in men.

28

Moreover,a heart-failure diagnosis nomogram has been devel-oped.

8

In patients with severe renal disease, B-typenatriuretic peptide levels are increased. The meanB-type natriuretic peptide level in patients with anoncardiac cause of dyspnea and an estimated glo-merular filtration rate of less than 60 ml per minuteper 1.73 m

2

of body-surface area was nearly 300 pgper milliliter in the Breathing Not Properly Multi-national Study.

29

Therefore, higher cutoff valuesneed to be identified for this important patientpopulation.

In conclusion, we found that when used in con-junction with other clinical information, rapid mea-surement of B-type natriuretic peptide levels in theemergency department improves the care of pa-tients with acute dyspnea and thereby reduces thetime to discharge and the total cost of treatment.

n engl j med

350;7


12

,

2004

654


Supported by research grants from the Swiss National ScienceFoundation, the Swiss Heart Foundation, the Novartis Foundation,the Krokus Foundation, and the University of Basel (to Dr. Mueller).Diagnostic devices and reagents (Triage) were provided by Biosite,San Diego, Calif.

We are indebted to the emergency department staff at UniversityHospital Basel for their valuable efforts, to all participating patients,and to Drs. Barbara Frana, Daniel Rodriguez, and Bruno Schurterfor their help with data management.

references

1.

McCullough PA, Philbin EF, Spertus JA,et al. Confirmation of a heart failure epi-demic: findings from the Resource Utili-zation Among Congestive Heart Failure(REACH) study. J Am Coll Cardiol 2002;39:60-9.

2.

Redfield MM. Heart failure — an epi-demic of uncertain proportions. N Engl JMed 2002;347:1442-4.

3.

Heart disease and stroke statistics —2003 update. Dallas: American Heart Asso-ciation, 2002.

4.

Hunt SA, Baker DW, Chin MH, et al.ACC/AHA guidelines for the evaluation andmanagement of chronic heart failure in theadult: executive summary: a report of theAmerican College of Cardiology/AmericanHeart Association Task Force on PracticeGuidelines (Committee to Revise the 1995Guidelines for the Evaluation and Manage-ment of Heart Failure). Circulation 2001;104:2996-3007.

5.

Task Force for the Diagnosis and Treat-ment of Chronic Heart Failure, European So-ciety of Cardiology. Guidelines for the diag-nosis and treatment of chronic heart failure.Eur Heart J 2001;22:1527-60. [Erratum, EurHeart J 2001;22:2217-8.]

6.

Maisel AS, Krishnaswamy P, NowakRM, et al. Rapid measurement of B-type na-triuretic peptide in the emergency diagnosisof heart failure. N Engl J Med 2002;347:161-7.

7.

McCullough PA, Nowak RM, McCord J,et al. B-type natriuretic peptide and clinicaljudgment in emergency diagnosis of heartfailure: analysis from Breathing Not Proper-ly (BNP) Multinational Study. Circulation2002;106:416-22.

8.

Bales AC, Sorrentino MJ. Causes of con-gestive heart failure: prompt diagnosis mayaffect prognosis. Postgrad Med 1997;101:44-9, 54-6.

9.

Wuerz RC, Meador SA. Effects of pre-hospital medications on mortality andlength of stay in congestive heart failure.Ann Emerg Med 1992;21:669-74.

10.

Morrison LK, Harrison A, Krishna-swamy P, Kazanegra R, Clopton P, Maisel A.Utility of a rapid B-natriuretic peptide assay

in differentiating congestive heart failurefrom lung disease in patients presentingwith dyspnea. J Am Coll Cardiol 2002;39:202-9.

11.

Davis M, Espiner E, Richards G, et al.Plasma brain natriuretic peptide in assess-ment of acute dyspnea. Lancet 1994;343:440-4.

12.

McDonagh TA, Robb SD, Murdoch DR,et al. Biochemical detection of left-ventricu-lar systolic dysfunction. Lancet 1998;351:9-13.

13.

Dao Q, Krishnaswamy P, Kazanegra R,et al. Utility of B-type natriuretic peptide inthe diagnosis of congestive heart failure inan urgent-care setting. J Am Coll Cardiol2001;37:379-85.

14.

Cabanes L, Richaud-Thiriez B, Fulla Y,et al. Brain natriuretic peptide blood levelsin the differential diagnosis of dyspnea.Chest 2001;120:2047-50.

15.

Yasue H, Yoshimura M, Sumida H, et al.Localization and mechanism of secretion ofB-type natriuretic peptide in comparisonwith those of A-type natriuretic peptide innormal subjects and patients with heart fail-ure. Circulation 1994;90:195-203.

16.

Hama N, Itoh H, Shirakami G, et al.Rapid ventricular induction of brain natri-uretic peptide gene expression in experi-mental acute myocardial infarction. Circula-tion 1995;92:1558-64.

17.

Wiese S, Breyer T, Dragu A, et al. Geneexpression of brain natriuretic peptide inisolated atrial and ventricular human myo-cardium: influence of angiotensin II and di-astolic fiber length. Circulation 2000;102:3074-9.

18.

Cheng V, Kazanegra R, Garcia A, et al.A rapid bedside test for B-type peptide pre-dicts treatment outcomes in patients admit-ted for decompensated heart failure: a pilotstudy. J Am Coll Cardiol 2001;37:386-91.

19.

Koglin J, Pehlivanli S, Schwaiblmair M,Vogeser M, Cremer P, von Scheidt W. Role ofbrain natriuretic peptide in risk stratifica-tion of patients with congestive heart fail-ure. J Am Coll Cardiol 2001;38:1934-41.

20.

Mueller C, Buser P. B-type natriureticpeptide (BNP): can it improve our manage-

ment of patients with congestive heart fail-ure? Swiss Med Wkly 2002;132:618-22.

21.

Weinstein MC, Siegel JE, Gold MR,Kamlet MS, Russell LB. Recommendationsof the Panel on Cost-effectiveness in Healthand Medicine. JAMA 1996;276:1253-8.

22.

Siegel JE, Weinstein MC, Russell LB,Gold MR. Recommendations for reportingcost-effectiveness analysis. JAMA 1996;276:1339-41.

23.

Nielsen OW, McDonagh TA, Robb SD,Dargie HJ. Retrospective analysis of thecost-effectiveness of using plasma natri-uretic peptide in screening for left ventricu-lar systolic dysfunction in the general popu-lation. J Am Coll Cardiol 2003;41:113-20.

24.

Incalzi RA, Fuso L, Serra M, et al. Exac-erbated chronic obstructive pulmonary dis-ease: a frequently unrecognized condition.J Intern Med 2002;252:48-55.

25.

Rihal CS, Davis KB, Kennedy JW, GershBJ. The utility of clinical, electrocardio-graphic, and roentgenographic variables inthe prediction of left ventricular function.Am J Cardiol 1995;75:220-3.

26.

Gillespie ND, McNeill G, Pringle T,Ogston S, Struthers AD, Pringle SD. Crosssectional study of contribution of clinical as-sessment and simple cardiac investigationsto diagnosis of left ventricular systolic dys-function in patients admitted with acutedyspnoea. BMJ 1997;314:936-40.

27.

Zile MR, Gaasch WH, Carroll JD, et al.Heart failure with a normal ejection frac-tion: is measurement of diastolic functionnecessary to make the diagnosis of diastolicheart failure? Circulation 2001;104:779-82.

28.

Redfield MM, Rodeheffer RJ, JacobsenSJ, Mahoney DW, Bailey KR, Burnett JC Jr.Plasma brain natriuretic peptide concentra-tion: impact of age and gender. J Am CollCardiol 2002;40:976-82.

29.

McCullough PA, Duc P, Omland T, et al.B-type natriuretic peptide and renal func-tion in the diagnosis of heart failure: an analy-sis from the Breathing Not Properly Multi-national Study. Am J Kidney Dis 2003;41:571-9.

Copyright © 2004 Massachusetts Medical Society.

Heart Failure

Predischarge B-Type Natriuretic PeptideAssay for Identifying Patients at High Risk ofRe-Admission After Decompensated Heart FailureDamien Logeart, MD,* Gabriel Thabut, MD,† Patrick Jourdain, MD,§ Christophe Chavelas, MD,*Pascale Beyne, PHD,‡ Florence Beauvais, MD,* Erik Bouvier, MD,* Alain Cohen Solal, MD, PHD*Clichy and Pontoise, France

OBJECTIVES The aim of this study was to determine the value of serial B-type natriuretic peptide (BNP)assay for predicting post-discharge outcome of patients admitted for decompensatedcongestive heart failure (CHF).

BACKGROUND Patients hospitalized for decompensated CHF are frequently re-admitted. Thus, identifica-tion of high-risk patients before their discharge is a major issue that remains challenging.B-type natriuretic peptide measurement could be useful.

METHODS Serial BNP measurements were performed from admission to discharge in two samples ofconsecutive patients. Survivors were monitored for six months; the main end point combineddeath or first re-admission for CHF.

RESULTS Among the 105 survivors of the derivation study, all serial BNP values, percentage change inBNP levels, and predischarge Doppler mitral pattern correlated with the outcome. Incontrast, clinical variables and left ventricular ejection fraction were poorly predictive. Thepredischarge BNP assay had the best discriminative power (area under the receiver operatingcharacteristic [ROC] curve � 0.80) and remained the lone significant variable in multivariateanalysis (hazard ratio [HR] � 1.14 [95% confidence interval {CI}, 1.02 to 1.28], p � 0.027).Among the 97 survivors of the validation study, the predischarge BNP assay was also the mostpredictive parameter (area under the ROC curve � 0.83). The risk of death or re-admissionincreased in stepwise fashion across increasing predischarge BNP ranges (p � 0.0001). Afteradjustment for baseline covariables, the HRs were 5.1 [95% CI 2.8 to 9.1] for BNP levelsbetween 350 and 700 ng/l and 15.2 [95% CI 8.5 to 27] for BNP levels �700 ng/l, comparedwith BNP �350 ng/l.

CONCLUSIONS High predischarge BNP assay is a strong, independent marker of death or re-admission afterdecompensated CHF, more relevant than common clinical or echocardiographic parametersand more relevant than changes in BNP levels during acute cares. (J Am Coll Cardiol 2004;43:635–41) © 2004 by the American College of Cardiology Foundation

Heart failure (HF) is a major reason for hospitalization andrepresents a huge cost for national health care budgets. Hos-pitalization for decompensated congestive heart failure (CHF)carries a poor vital prognosis, with frequent subsequent re-admissions (1–5). It has recently been shown that outcome canbe improved in high-risk patients by treatment intensificationand home-based interventions (6,7). However, there is cur-rently no simple clinical criterion or score for predicting earlyoutcome after discharge (8) and, thus, for identifying patientswho require such caution. Doppler echocardiography is largelyused in hospitalized patients and can help to stratify patients.Indeed, Doppler analysis of left ventricular (LV) filling param-eters has been linked to outcome after myocardial infarctionand in dilated cardiomyopathy (9,10), including in the settingof hospitalization (11).

The B-type natriuretic peptide (BNP) serum level is apromising cardiac marker in various HF settings, especiallybecause of the advent of rapid assays. Indeed, BNP is secretedby overloaded LV, and its blood level is related to severity ofLV dysfunction and CHF. B-type natriuretic peptide hasdiagnostic value in acute dyspnea (12–14); BNP also hasprognostic value after myocardial infarction (15,16), as well asin outpatients with chronic CHF and systolic dysfunction(17–19). Despite potential interest, the prognostic relevance ofBNP assay has been poorly investigated in the setting ofpatients admitted for decompensated CHF.

The aims of this study were to determine the prognosticvalue of serial BNP assay in consecutive patients hospi-talized for severely decompensated CHF for the predic-tion of early death or re-admission for CHF, and tocompare it to common clinical and echocardiographiccharacteristics.

METHODS

Study design. The study was approved by local ethicscommittees. Consecutive patients admitted to the cardiol-

From *Service de Cardiologie, †Service de Pneumologie, ‡Laboratoire de Bio-chimie, Hopital Beaujon, Clichy, France; and §Service de Cardiologie, Hopital ReneDubos, Pontoise, France. Supported, in part, by a grant from the French Federationof Cardiology, Paris, France. Drs. Logeart and Thabut contributed equally to thiswork. Marc A. Silver, MD, acted as guest editor for this paper.

Manuscript received March 9, 2003; revised manuscript received September 10,2003, accepted September 17, 2003.

Journal of the American College of Cardiology Vol. 43, No. 4, 2004© 2004 by the American College of Cardiology Foundation ISSN 0735-1097/04/$30.00Published by Elsevier Inc. doi:10.1016/j.jacc.2003.09.044

ogy department for decompensated CHF were enrolled,except in the following circumstances: acute myocardialinfarction, severe valve disease, surgical patients, or pooradherence to therapy. The diagnosis of decompensatedCHF was confirmed by two senior cardiologists using thegenerally accepted Framingham criteria and corroborativeinformation including the hospital course and results offurther cardiac tests. On admission, a blood sample wascollected for BNP measurement. Standard treatment wasprescribed daily by senior cardiologists. Blood samples werealso collected every day for serial BNP assay. A final bloodsample was collected either on the day of discharge or on theday before discharge (predischarge values). B-type natri-uretic peptide assays were performed after discharge ofpatient, and results were kept blinded until the end of thestudy. Doppler echocardiograms were performed beforedischarge, in order to assess LV ejection fraction (LVEF) aswell as Doppler mitral inflow pattern and systolic pulmo-nary artery pressure. The discharge was decided by twosenior cardiologists in charge of the HF unit, using clinicalexamination, biological tests (except BNP measurements),electrocardiogram, and chest radiograms. Patients weredischarged when they presented no more sign of decom-pensation, stable blood pressure as well as renal function,and optimal achievement of diuretics and angiotensin-converting enzyme inhibitor dosages. Outcome during thesix months after discharge was determined in every case, bytelephoning the patient or the general practitioner. Themain end point combined death or first unscheduled re-admission for CHF. Re-admission for CHF was defined byhospitalization for decompensated CHF.Patients. In the derivation study, 127 patients were en-rolled in a single center (Beaujon Hospital); 13 weresubsequently excluded because of requirement for emergenttransplantation, transfer to another hospital, absence ofpredischarge BNP measurement, or rectified non-CHFdiagnosis, and, thus, 114 patients were finally included inthe derivation study. In the validation study, 109 patientswere included in another center (Pontoise Hospital).BNP assay. Blood was collected into tubes containingpotassium EDTA (1 mg/ml blood), and plasma was storedat �80°C for blinded BNP assay with the Triage BNP test(Biosite Diagnostics Inc., San Diego, California). Triage isa point-of-care method based on fluorescence immunoas-

say, and can be used to quantify BNP in whole blood orplasma. At time of the study, the working range of theTriage BNP assay was 5.0 to 1,300 ng/l, and samples withvalues exceeding 1,300 ng/l were, thus, diluted with normalplasma and retested.Doppler echocardiography. Doppler echocardiographicexaminations were performed with a Hewlett Packard(Andover, Massachusetts) Sonos 5500 machine equippedwith a 2.5-MHz probe. Each examination was recorded onvideotape for subsequent blind analysis. Pulsed Doppleranalysis of mitral inflow included measurements of themitral valve early peak filling velocity (E), the late peakfilling velocity (A), the E to A ratio, and the decelerationtime of E-wave (DTE), and yielded three patterns: 1) an“impaired relaxation” pattern (E/A ratio �1); 2) a “restric-tive” pattern when the E/A ratio was �2, or between 1 and2 with an E-wave deceleration time (DT) �130 ms, or DT�130 ms alone in case of atrial fibrillation; and 3) a“pseudonormal” or “normalized” pattern when the E/A ratiowas between 1 and 2 and the E-wave DT �130 ms. Themitral Doppler pattern was unavailable in eight patientsbecause of poor echogenicity, tachycardia, permanent pac-ing, or mitral prosthesis. Systolic pulmonary arterial pres-sure was calculated from the velocity of tricuspid regurgita-tion, when present. The LVEF was estimated by Simpson’smethod.Statistical analysis. Categorical data are presented as num-bers (percent), and continuous data as means � SD.Log-transformed values for BNP were used in these anal-yses to reduce the effects of skewness in the distribution ofBNP values. Student t test and Fisher exact test were usedwhen indicated. Group comparisons of BNP values weremade by using analysis of variance with the Newman-Keulspost hoc test. We used Cox proportional hazards regressionmodels to examine the relation of clinical variables, BNPlevels, and echocardiographic findings with the incidence ofprimary end point (death or re-admission for CHF) or alonere-admission for CHF within the first month and at sixmonths after discharge. B-type natriuretic peptide levelswere evaluated both as a continuous variable (with increasesin risk calculated per increment of 100 ng/l) and as a categoricalvariable (based on distribution quartiles). Percentage change inBNP level ([BNP “admission” � “predischarge ” BNP]� 100/BNP “admission”) was also evaluated as a contin-uous variable (with decreases in risk calculated per decreaseof 10%) and as a categorical variable. Analyses were adjustedfor the following baseline covariables: age, diabetes mellitus,and LVEF (well-known predictors of mortality), inotropicdrug use, and Doppler mitral pattern (associated withoutcome in this study). Receiver operating characteristics(ROC) curves were constructed to illustrate various cut-offvalues of BNP. P values �0.05 were considered significant.Analyses were performed using STATA 8.0 software forWindows (Stata Corporation, College Station, Texas).

Abbreviations and AcronymsBNP � B-type natriuretic peptideCHF � congestive heart failureDTE � deceleration time of E-waveHF � heart failureHR � hazard ratioLV � left ventricle/ventricularLVEF � left ventricular ejection fractionROC � receiver operating characteristic

636 Logeart et al. JACC Vol. 43, No. 4, 2004BNP Assay and Heart Failure Outcome February 18, 2004:635–41

RESULTS

Derivation study: outcome and predictors of outcome.All the 114 patients were admitted because of severelydecompensated CHF (New York Heart Association classIV); 83 patients had pulmonary edema, 12 had cardiogenicshock, and 24 required inotropic agents and/or mechanicalventilation. Main characteristics are summarized in Table 1.Nine patients died of refractory CHF during the initialhospital stay. The 105 surviving patients were dischargedhome in New York Heart Association class II to III and hadno more symptoms at rest, no rales, no gallop, no severehypotension (defined as systolic blood pressure �80 mmHg). Mean BNP levels were 1,015 � 604 ng/l at admission,881 � 615 ng/l at 24 h, 638 � 560 ng/l at 48 h, and 457 �402 ng/l predischarge (Fig. 1). B-type natriuretic peptidelevels did not decrease, or fell by �50 ng/l, in 11 patientsbetween admission and discharge. Predischarge Doppler-echocardiographic examination showed a systolic dysfunc-tion (LVEF �45%) in 70% of patients, “impaired relax-ation,” “pseudonormal,” and “restrictive” mitral patterns in38%, 33%, and 29% of patients, respectively.

During the six months of follow-up, 12 patients died, and

39 were re-admitted for unscheduled CHF; 29% of theseevents occurred during the first month, and the mean timefrom discharge to the first event was 72 � 48 days. Table 2shows the results of univariate Cox analysis for each clinicalvariable, creatininemia, echocardiographic findings, and se-rial BNP measurements as predictors of death or re-admission. Among clinical variables, only the use of inotro-pic drugs was associated with adverse outcome. Amongechocardiographic variables, LVEF was poorly predictivewhile the predischarge Doppler mitral pattern was stronglyassociated with death or re-admission. Serial BNP measure-ments were also predictive of outcome as well the percent-age change between admission and subsequent assays. Thepredischarge BNP level was the value most strongly associ-ated with death or re-admission and the most discriminative(area under ROC curve � 0.80 [0.71 to 0.89] vs. 0.69 and0.68 for previous BNP assays and 0.76 for percentagechange).

In multivariate analysis including clinical variable, echo-cardiographic findings, the predischarge BNP level, and thepercentage change in BNP levels, only the predischargeBNP level remained significant (hazard ratio [HR] � 1.14[1.02 to 1.28], p � 0.027). Using the predischarge BNPlevel (as a continuous variable and after adjustment forcovariables), similar results were obtained for: 1) death orre-admission at one month (HR � 1.17 [1.06 to 1.28], p �0.002); and 2) re-admission at six months (HR � 1.25 [1.16to 1.34], p � 0.001). At last, a BNP level of 350 ng/l wasfound to have the best compromise between sensitivity andspecificity for predicting death or re-admission at sixmonths (Fig. 2).Validation study. Main characteristics of the validationsample are summarized in Table 1. Patients enrolled in thissample differed from those enrolled in the derivation sampleonly for gender and inotropic drugs use. Among the 109included patients, 12 died during the hospitalization, ninedied during the six months follow-up, and 26 were rehos-pitalized for decompensated HF. Mean BNP levels were941 � 526 ng/l at admission, 693 � 440 ng/l at day 1, and441 � 501 mg/l at discharge. Among serial BNP measure-ments, predischarge BNP level remains a strong predictor ofdeath or re-admission in the validation sample (area underROC curve � 0.83 [0.69 to 0.97]). Event-free patients hadmean predischarge BNP levels of 247 � 201 ng/l comparedwith 908 � 809 ng/l for patients with adverse events (p �0.001). Applying the predefined BNP cut-off level of 350ng/l, death or re-admission at six months was predictedwith a sensitivity of 80% and a specificity of 88%.

Figure 3 shows Kaplan-Meier curves according to pre-defined BNP cut-off (350 ng/l); predischarge BNP levels�350 ng/l strongly related to death or re-admission (HR �12.6 [5.7 to 28.1], p � 0.0001), and the rate of eventsreached 23.5% at one month and 79.4% at six months,compared with 0% and 12.7% for predischarge BNP levels�350 ng/l.

Table 1. Main Characteristics of the Patients in the TwoGroups and Adverse Events After Discharge

DerivationStudy

(n � 114)

ValidationStudy

(n � 109) p Value

Age (yrs) 69.4 � 14.4 70.9 � 13.3 0.45Age �75 yrs (n, %) 50 (44%) 32 (43%) 0.85Gender (male/female) 79/35 56/53 0.003Ischemic etiology (n, %) 44 (39%) 45 (42%) 0.50LVEF 37.5 � 14.9 31.8 � 14.5 0.17Use of inotropes (n, %) 25 (22%) 34 (32%) 0.07In-hospital death (n, %) 9 (8%) 12 (11%) 0.64BNP at admission (ng/ml) 1,015 � 604 941 � 526 0.68BNP at discharge (ng/ml) 457 � 451 441 � 501 0.80

Outcome after discharge n � 105 n � 97Cardiac death (n, %) 12 (11%) 9 (9%) 0.64First re-admission (n, %) 39 (37%) 26 (27%) 0.21

BNP � B-type natriuretic peptide; LVEF � left ventricular ejection fraction.

Figure 1. Box plots showing median predischarge levels of B-type natri-uretic peptide (BNP) from admission to discharge (derivation study). Meanlevels are indicated in the box.

637JACC Vol. 43, No. 4, 2004 Logeart et al.February 18, 2004:635–41 BNP Assay and Heart Failure Outcome

Graded relation between predischarge BNP levels andoutcome. We constructed Kaplan-Meier curves with thepredischarge BNP values from the whole population (i.e.,derivation and validation samples) (Fig. 4). The risk ofdeath or re-admission increased in stepwise fashion acrossincreasing predischarge BNP ranges: �350 ng/l, 350 to 700

ng/l, �700 ng/l (p � 0.001); this relation was unchangedafter adjustment for age, LVEF, diabetes mellitus, and theuse of inotropic drugs. Patients with BNP �350 ng/l hadthe best outcome (16.2% of events at 6 months) comparedwith patients with BNP between 350 and 700 ng/l (60.0%,HR 5.1 [2.8 to 9.1]) and patients with BNP �700 ng/l(92.7%, HR 15.2 [8.5 to 27]).

DISCUSSION

This prospective study documents and validates that highpredischarge blood BNP levels are a strong predictor ofshort-term death or re-admission after acute hospital carefor decompensated CHF. This relationship between predis-charge BNP and outcome is graded. The prognostic infor-mation of predischarge BNP assay is greater than arecommon clinical variables, BNP change during the in-hospital stay, and Doppler echocardiographic findings. Thetwo samples consisted of unselected patients with a widerange of ages and LVEF values (preserved LVEF, i.e.,�0.45, in 30% of patients), similar to that encountered inroutine practice, and successively included in two differenthospitals. The very high rate of post-discharge adverseevents was in keeping with the results of other studies ofunselected community-dwelling patients (1–5,20). Stratifi-cation for subsequent treatment (titration of medicationssuch as diuretics, visits, and so forth) is, therefore, impor-

Table 2. Univariate Cox Analysis Showing the Association Between Various Variables and theRisk of Death or Re-Admission During the Six Months of Follow-up in the Derivation Study

Variables HR 95% CI p Value

Age (yrs), per five-year increase 1.03 0.93–1.13 0.6Gender (male) 0.87 0.49–1.56 0.6Ischemic 1.14 0.65–2.0 0.6Diabetes 1.16 0.58–2.32 0.7Atrial fibrillation 0.61 0.30–1.26 0.2Creatininemia 1.02 0.98–1.06 0.3Inotropic use 2.34 1.26–4.34 0.007LVEF 0.95 0.78–1.16 0.6Doppler mitral pattern

“Abnormal relaxation” pattern 1“Pseudo-normal” pattern 3.27 1.43–7.49 0.005“Restrictive” pattern 6.30 2.78–14.25 0.0001

Serial BNPBNP admission, per 100 ng/l increase 1.06 1.03–1.10 0.0001BNP day 2, per 100 ng/l increase 1.07 1.03–1.11 0.0001BNP predischarge, per 100 ng/l increase 1.22 1.15–1.30 0.0001Percentage decrease BNP, per 10% decrease 0.84 0.77–0.90 0.0001

BNP predischarge (per quartile of distribution)First quartile (0–130) 1Second quartile (130–370) 2.94 0.92–9.37 0.07Third quartile (370–660) 4.42 1.45–13.45 0.009Fourth quartile (660–1,725) 13.77 4.71–40.23 0.0001

Percentage decrease BNP (per quartile of distribution)First quartile (�38%) 1Second quartile (38%–63%) 0.94 0.49–1.81 0.85Third quartile (63%–79%) 0.40 0.19–0.87 0.02Fourth quartile (�79%) 0.18 0.07–0.48 0.001

BNP � B-type natriuretic peptide; CI � confidence interval; HR � hazard ratio; LVEF � left ventricular ejection fraction;Percentage decrease BNP � (BNP admission � predischarge BNP)/BNP admission.

Figure 2. Receiver operating characteristic (ROC) curve for predischargeB-type natriuretic peptide (BNP) cut-off values and association with deathor re-admission at six months (derivation study).


tant. In our study, clinical variables were poorly predictive ofpost-discharge outcome, with the exception of inotropic useduring acute care. All patients were judged to be stable atdischarge, even though, respectively, 15% of them werere-admitted or died during the first month and more than40% during the first six months after discharge. In practiceit is difficult to evaluate, using clinical criteria, the stabilityof such weakened and sometimes bedridden patients afterseveral days of aggressive treatment. Indeed, clinical judg-ment correlates poorly with tests of cardiac function (21,22).Moreover, a number of tests with known prognostic value,such as invasive hemodynamic measurements and stresstesting of aerobic capacity, are unusable in many of thesepatients. Nevertheless, Doppler-echocardiographic exami-nation results in criteria of prognostic significance, whichcan be easily obtained, even in this setting, and are widely

admitted. While LVEF had poor prognostic value in thetwo population samples of our study, the Doppler mitralinflow pattern was predictive, and “restrictive” and “pseudo-normal” patterns before discharge were associated withadverse outcome. The “restrictive” pattern is a widely usedprognostic index in various settings (9,10,23–25), whereas a“pseudonormal” pattern has rarely been linked to pooroutcome of CHF (11).

B-type natriuretic peptide measurement, a simple biolog-ical test, is appropriate to this acute setting. In univariateCox analysis, all BNP measurements from admission todischarge, as well as the percentage change, were signifi-cantly associated with the risk of death or re-admission afterdischarge, but the predischarge assay was the most valuable.After adjustment for age, diabetes mellitus, and LVEF(usual predictors of mortality), and for inotropic drug

Figure 3. Kaplan-Meier curves showing the cumulative incidence of death or re-admission according to predischarge B-type natriuretic peptide (BNP)cut-off value of 350 ng/l in the validation study; p � 0.0001.

Figure 4. Kaplan-Meier curves showing the cumulative incidence of death or re-admission according to predischarge B-type natriuretic peptide (BNP)ranges (�350, 350 to 700, �700 ng/l) in the whole population; p � 0.001 for the trend among BNP ranges. On the right, hazard ratios are shown foreach BNP range.


requirements (the only predictive clinical variable foundhere), predischarge BNP levels remained strongly predictiveof death or re-admission after both one month and sixmonths of follow-up. Recently, changes in the BNP levelduring early aggressive treatment were closely associatedwith falling pulmonary wedge pressure in patients treatedfor decompensated CHF (26). In addition, Cheng et al. (27)reported that changes in serial BNP levels during hospitalcare were predictive of outcome, but BNP was not com-pared with other parameters, and end points combinedin-hospital deaths and post-discharge events. In this study,very high mean BNP levels at discharge (more than 1,500ng/l) and no decrease during treatment in patients who diedor were re-admitted suggest that the patients were veryseverely sick. In our study, the percentage change in BNPlevels during acute care had less prognostic value than thepredischarge measurement alone. B-type natriuretic peptidelevels are mainly determined by LV wall stress. The predis-charge BNP level may, thus, reflect the decrease in LVfilling pressure and the degree of hemodynamic stability,which is achieved after acute treatment, as well as theseverity of underlying diastolic dysfunction, and is a majorpredictor of early outcome. Interestingly, BNP level andDoppler mitral pattern have similar significance from amechanistic point of view, as LV filling pressure and LVwall stress are major determinants of both BNP levels (28)and Doppler mitral pattern (29,30). The E-wave DTcorrelated negatively with BNP levels in our study (data notshown), and the predischarge “restrictive” Doppler pattern(suggesting lasting high LV filling pressure or severe dia-stolic dysfunction) was associated with the highest BNPlevels at discharge. Finally, Doppler analysis of mitral inflowadded no significant prognostic information to BNP mea-surement when the two parameters were included in mul-tivariate Cox models. A predischarge BNP level at 350ng/ml appeared the more relevant cut-off to predict adverseoutcome in our two samples of patients. In fact, the risk ofdeath or re-admission increased in stepwise fashion acrossincreasing predischarge BNP ranges; a dramatic increase inthe risk of events was observed from 350 ng/l, and BNP�700 ng/l was associated with a major risk (31% of death orre-admission for CHF at one month and 93% at sixmonths). These levels, like those reported by Cheng et al.(27), are higher than the values generally reported inoutpatients (17,18). Among 85 selected outpatients withsymptomatic CHF and systolic dysfunction, Tsutamoto etal. (17) found that the median BNP level (73 ng/l) distin-guished subsequent survivors from nonsurvivors. Berger etal. (19) recently proposed a cut-off of 130 ng/l. Thesedifferences in BNP cut-offs reflect differences in the dura-tion of follow-up and in the study populations (e.g.,hospitalized patients vs. outpatients).Clinical implications. The high rates of early re-admissionobserved in this study suggest that many CHF patients aredischarged without sufficient circulatory stabilization, de-spite the clinician’s impression to the contrary. The single

predischarge BNP assay is strongly predictive of earlyoutcome, regardless of the initial BNP level (and in-hospitalchanges) and echocardiographic findings; thus, it appears tobe a simple and reliable test to identify the highest riskpatients. Further studies are required to determine if serialBNP assay can improve patient management, for exampleby determining the optimal timing of discharge and subse-quent care requirements, as recently suggested (31).

AcknowledgmentThe authors thank David Young for re-styling the manu-script.

Reprint requests and correspondence: Dr. Damien Logeart,Service de Cardiologie, Hopital Beaujon, 100 Boulevard du GalLeclerc, 92110 Clichy, France. E-mail: [email protected].

REFERENCES

1. Cowie MR, Fox KF, Wood DA, et al. Hospitalization of patients withheart failure: a population-based study. Eur Heart J 2002;23:877–85.

2. Jong P, Vowinckel E, Liu PP, et al. Prognosis and determinants ofsurvival in patients newly hospitalized for heart failure. Arch InternMed 2002;162:1689–94.

3. Philbin EF, Rocco TA, Lindenmuth NW, Ulrich K, Jenkins PL.Clinical outcomes in heart failure: report from a community hospital-based registry. Am J Med 1999;107:549–55.

4. Haldeman GA, Croft JB, Giles WH, et al. Hospitalization of patientswith heart failure: national hospital discharge survey 1985–1995. AmHeart J 1999;137:352–60.

5. Krumholz HM, Parent EM, Tu N, et al. Readmission after hospital-ization for congestive heart failure among Medicare beneficiaries. ArchIntern Med 1997;157:99–104.

6. Rich MW, Beckham V, Wittenberg C, et al. A multidisciplinaryintervention to prevent the readmission of elderly patients withcongestive heart failure. N Engl J Med 1995;333:1190–5.

7. Stewart S, Pearson S, Horowitz JD. Effects of a home-based inter-vention among patients with congestive heart failure discharged fromacute hospital care. Arch Intern Med 1998;158:1067–72.

8. Krumholz HM, Chen YT, Wang Y, Vaccarino V, Radford MJ,Horwitz RI. Predictors of readmission among elderly survivors ofadmission with heart failure. Am Heart J 2001;139:72–7.

9. Vanoverschelde JL, Raphael DA, Robert AR, et al. Left ventricularfilling in dilated cardiomyopathy: relation to functional class andhemodynamics. J Am Coll Cardiol 1990;15:1288–95.

10. Pozzoli M, Traversi E, Cioffi G, et al. Loading manipulations improvethe prognostic value of Doppler evaluation of mitral flow in patientswith chronic heart failure. Circulation 1997;95:1222–30.

11. Whalley GA, Doughty RN, Gamble GD, et al. Pseudonormal mitralfilling pattern predicts re-admission in patients with congestive heartfailure. J Am Coll Cardiol 2002;39:1787–95.

12. Maisel AS, Krishnaswamy P, Nowak RM, et al., for the Breathing NotProperly Multinational Study Investigators. Rapid measurement ofB-type natriuretic peptide in the emergency diagnosis of heart failure.N Engl J Med 2002;347:161–7.

13. Morrison KL, Harrison A, Krishnaswamy P, et al. Utility of a rapidB-natriuretic peptide (BNP) assay in differentiating CHF from lungdisease in patients presenting with dyspnea. J Am Coll Cardiol2002;39:202–9.

14. Logeart D, Saudubray C, Beyne P, et al. Comparative value of Dopplerechocardiography and B-type natriuretic peptide assay in the etiologicdiagnosis of acute dyspnea. J Am Coll Cardiol 2002;40:1794–800.

15. Arakawa N, Nakamura M, Aoki H, Hiramori K. Plasma brainnatriuretic peptide concentrations predict survival after acute myocar-dial infarction treatment of heart failure guided by plasma aminoterminal brain natriuretic peptide (N-BNP) concentrations. J Am CollCardiol 1996;27:1656–61.


16. Richards AM, Nicholls MG, Yandle TG, et al. Plasma N-terminalpro-brain natriuretic peptide and adrenomedullin: new neurohormonalpredictors of left ventricular function and prognosis after myocardialinfarction. Circulation 1998;97:1921–9.

17. Tsutamoto T, Wada A, Maeda K, et al. Attenuation of compensationof endogenous cardiac natriuretic peptide system in chronic heartfailure: prognostic role of plasma brain natriuretic peptide concentra-tion in patients with chronic symptomatic left ventricular dysfunction.Circulation 1997;96:509–16.

18. Koglin J, Pehlivanli S, Schwaiblamir M, et al. Role of brain natriureticpeptide in risk stratification of patients with congestive heart failure.J Am Coll Cardiol 2001;38:1934–41.

19. Berger R, Huelsman M, Strecker K, et al. B-type natriuretic peptidepredicts sudden death in patients with chronic heart failure. Circula-tion 2002;105:2392–7.

20. Senni M, Tribouilloy CM, Rodeheffer RJ, et al. Congestive heartfailure in the community: a study of all incident cases in OlmstedCounty, Minnesota, in 1991. Circulation 1998;98:2282–9.

21. Stevenson LW, Perloff JK. The limited reliability of physical signs forestimating hemodynamics in chronic heart failure. JAMA 1989;261:884–8.

22. Remes J, Miettinen H, Reunanen A, et al. Validity of clinical diagnosisof heart failure in primary health care. Eur Heart J 1991;12:315–21.

23. Xie GY, Berk MR, Smith MD, et al. Prognostic value of Dopplertransmitral flow patterns in patients with congestive heart failure. J AmColl Cardiol 1994;24:132–9.

24. Giannuzzi P, Temporelli PL, Bosimini E, et al. Independent andincremental prognostic value of Doppler-derived mitral decelera-

tion time of early filling in both symptomatic and asymptomaticpatients with left ventricular dysfunction. J Am Coll Cardiol1996;28:383–90.

25. Tabet JY, Logeart D, Geyer C, et al. Comparison of the prognosticvalue of left ventricular filling and peak oxygen uptake in patientswith systolic heart failure. Eur Heart J 2000;21:1864 –71.

26. Kazanegra R, Cheng V, Garcia A, et al. A rapid test for B-typenatriuretic peptide correlates with falling wedge pressures in pa-tients treated for decompensated heart failure: a pilot study. J CardFailure 2001;7:21–9.

27. Cheng VL, Krishnaswamy P, Kazanegra R, et al. A rapid bedside testfor B-type natriuretic peptide predicts treatment outcomes in patientsadmitted with decompensated heart failure. J Am Coll Cardiol2001;37:386–91.

28. Maeda K, Takayoshi T, Wada A, et al. Plasma brain natriureticpeptide as a biochemical marker of high left ventricular end-diastolicpressure in patients with symptomatic left ventricular dysfunction. AmHeart J 1998;135:825–32.

29. Nishimura RA, Tajik AJ. Evaluation of diastolic filling of left ventriclein health and disease: Doppler echocardiography is the clinician’sRosetta stone. J Am Coll Cardiol 1997;30:8–18.

30. Appleton CP, Firstenberg MS, Garcia MJ, et al. The echo-Dopplerevaluation of left ventricular diastolic function: a current perspective.Cardiol Clin 2000;18:513–46.

31. Troughton RW, Frampton CM, Yandle TG, et al. Treatment of heartfailure guided by plasma amino terminal brain natriuretic peptide(N-BNP) concentrations. Lancet 2000;355:1126–30.


EXPEDITED REVIEW

Plasma Brain Natriuretic Peptide-GuidedTherapy to Improve Outcome in Heart FailureThe STARS-BNP Multicenter Study

Patrick Jourdain, MD,*†‡ Guillaume Jondeau, MD, PHD,§ François Funck, MD,‡ Pascal Gueffet, MD,�Alain Le Helloco, MD,¶ Erwan Donal, MD,¶ Jean F. Aupetit, MD,# Marie C. Aumont, MD,§Michel Galinier, MD, PHD,** Jean C. Eicher, MD,†† Alain Cohen-Solal, MD, PHD,‡‡Yves Juillière, MD, PHD§§

Paris, Pontoise, Nantes, Rennes, Lyon, Toulouse, Dijon, and Vandoeuvre Les Nancy, France

Objectives The aim of this multicenter study was to evaluate the prognostic impact of a therapeutic strategy using plasmabrain natriuretic peptide (BNP) levels.

Background The prognosis of chronic heart failure (CHF) remains poor, even among patients treated in specialized depart-ments.

Methods A total of 220 New York Heart Association functional class II to III patients considered optimally treated withangiotensin-converting enzyme inhibitors (ACEIs), beta-blockers, and diuretics by CHF specialists were random-ized to medical treatment according to either current guidelines (clinical group) or a goal of decreasing BNPplasma levels �100 pg/ml (BNP group). Outpatient visits were scheduled every month for 3 months, then every3 months. The primary combined end point was CHF-related death or hospital stay for CHF.

Results Both groups were similar for baseline clinical and biological characteristics. Left ventricular ejection fraction wasslightly lower in the BNP group than in the clinical group (29.9 � 7.7% vs. 31.8 � 8.4%, p � 0.05). At the endof the first 3 months, all types of drugs were changed more frequently in the BNP group. Mean dosages of ACEIsand beta-blockers were significantly higher in the BNP group (p � 0.05), whereas the mean increase in furo-semide dosage was similar in both groups. During follow-up (median 15 months), significantly fewer patientsreached the combined end point in the BNP group (24% vs. 52%, p � 0.001).

Conclusions In optimally treated CHF patients, a BNP-guided strategy reduced the risk of CHF-related death or hospital stayfor CHF. The result was mainly obtained through an increase in ACEI and beta-blocker dosages. (J Am CollCardiol 2007;49:1733–9) © 2007 by the American College of Cardiology Foundation

Optimal treatment of heart failure includes angiotensin-converting enzyme inhibitors (ACEIs), beta-blockers, spi-ronolactone, and diuretics, as recommended by the interna-tional guidelines (1,2). Treatment optimization is currentlybased on physician experience and patient tolerance.

Chronic heart failure (CHF) remains the main reason forhospital stay among patients �65 years of age and isassociated with high mortality (3,4). However, implemen-tation of guidelines remains imperfect (5).

Brain natriuretic peptide (BNP) is a 32-amino acidprotein secreted by cardiac ventricles (6–9). The diagnosticand prognostic value of BNP plasma levels in CHF issupported by many studies (10–15). Most drugs used totreat heart failure significantly reduce BNP levels (16–18).In the Val-HeFT (Valsartan Heart Failure Trial), onlypatients receiving valsartan in addition to conventionaltherapy presented with a significant reduction in plasmaBNP levels compared with the placebo group (19). Recentdata from RALES (Randomized Aldactone EvaluationStudy) showed that spironolactone can also markedly reduceBNP levels in patients with severe heart failure (20). Theeffect of beta-blockers on BNP remains controversial. In

From the *René Descartes University, Paris, France; †Georges Pompidou Hospital,Paris, France; §Department of Cardiology, Bichat University Hospital, Paris, France;‡‡Department of Cardiology, Centre Hospitalier Universitaire de Lariboisière, Paris,France; ‡Hospitalier de Rene Dubos, Pontoise, France; �Centre Hospitalier Univer-sitaire de Nantes, Nantes, France; ¶Centre Hospitalier Universitaire de Rennes,Rennes, France; #Hopital Saint Luc Lyon, Lyon, France; **Department of Cardiol-ogy, Centre Hospitalier Universitaire de Toulouse, Toulouse, France; ††CentreHospitalier Universitaire de Dijon, Dijon, France; and the §§Department ofCardiology, Centre Hospitalier Universitaire de Vandoeuvre Les Nancy, VandoeuvreLes Nancy, France. The STARS study was supported by an unrestricted grant fromBiosite Inc. (San Diego, California) to the French Working Group on Heart Failure.Gary Francis, MD, FACC, served as Guest Editor for this article.

Manuscript received September 12, 2006; revised manuscript received October 10,2006, accepted October 22, 2006.

Journal of the American College of Cardiology Vol. 49, No. 16, 2007© 2007 by the American College of Cardiology Foundation ISSN 0735-1097/07/$32.00Published by Elsevier Inc. doi:10.1016/j.jacc.2006.10.081

randomized trials, patients withthe most marked neurohormonalactivation and the highest BNPlevels at enrollment seemed tobenefit most from beta-blockertherapy (21–23). More recent datashowed a significant N-terminalpro-BNP reduction in patientsreceiving carvedilol or metopro-lol (24).

Plasma BNP level monitoringhas been proposed for treatmentoptimization in patients withheart failure (25–27). However,these studies included few pa-

tients, most of whom did not receive beta-blocker therapy.Thus, the aim of this French multicenter randomized studywas to evaluate the benefit of BNP-guided therapy onoutcome in CHF patients in clinical practice.

Methods

Study population. The patients were included by CHFspecialists from 17 university hospitals. All investigators hadexpertise in CHF management and were members of theWorking Group on Heart Failure (WGHF) of the FrenchSociety of Cardiology.

The inclusion criteria were as follows: patients older than18 years with symptomatic (New York Heart Association[NYHA] functional class II to III) systolic heart failuredefined by left ventricular ejection fraction (LVEF) �45%assessed by echocardiography using the American Society ofEchocardiography guidelines, in stable condition (no hos-pital stay in the previous month), and treated by optimalmedical therapy according to the European guidelines at thetime of the study (1); dosages of medications were to bestable for at least 1 month before inclusion. Patients had toreceive diuretics, ACEIs, or angiotensin II-receptor block-ers (ARB) at the maximal tolerated dosage unless docu-mented intolerance and beta-blockers approved for CHF(carvedilol, bisoprolol, and metoprolol XR-CL), at themaximal tolerated dosage unless documented intolerance orspecific contra-indication. All patients signed informedconsent.

The exclusion criteria were as follows: acute coronarysyndrome within 3 months, chronic renal failure (plasmacreatininemia �250 �mol/l), documented hepatic cirrhosis,asthma, or chronic obstructive pulmonary disease.Study design. Patients were randomized into 2 groups:

• BNP group: medical therapy was increased with the aimof lowering plasma BNP levels (target �100 pg/ml)(1,10); each class of therapy could be modified accordingto the judgment of the investigator.

• Clinical group: medical therapy was adjusted accordingto the opinion of the investigator, on the basis of thephysical examination and usual paraclinical and biologi-

cal parameters. The investigators were not allowed tomeasure plasma BNP level.

Outpatient visits were scheduled every month for the first 3months and every 3 months thereafter. During the titrationphase (3 months), each visit included physical examination,electrocardiogram recording, and blood sample measure-ments. Plasma sodium levels, renal function, and hemoglo-bin were measured in both groups, whereas plasma BNPlevel was measured only in patients belonging to BNPgroup. During the follow-up phase, each visit included aclinical examination. The BNP was systematically measuredonly in the BNP group.

The protocol of the study was written by the WGHF ofthe French Society of Cardiology and approved by theCCPPRB (French legal ethics committee). Each patientgave written informed consent.Plasma BNP level measurement. Venous blood samplewas taken after 20 min of rest in the supine position andcollected on an EDTA tube. The BNP was immediatelymeasured with the immunofluorometric triage method (28)(Biosite Inc., San Diego, California). The BNP triage assayvariability was 10% for a threshold of 800 pg/ml and 15%for a threshold of 100 pg/ml.Primary and secondary end points. The primary compos-ite end point was defined as unplanned hospital stays forheart failure or death related to heart failure. Hospital stayswere decided by the physician in the emergency room, notinvolved in the study, and unaware of plasma BNP levels.The BNP plasma levels were kept in a specific patient chartand accessible only to investigator. Patients also wereblinded for BNP results. The secondary end points wereall-cause death, death related to heart failure, all-causehospital stay, and hospital stay for heart failure.Statistical analysis. Intent-to-treat principles were used.Continuous variables are presented as mean � SD. Com-parisons between continuous variables were made withStudent parametric unpaired t test for between groupcomparison and paired t test for 6-month versus baselinecomparison. Proportions were compared with the chi-square test. A statistical difference was considered signifi-cant if p � 0.05. Survival was evaluated with the Kaplan-Meier method. Differences in survival between the 2 groupswere compared with log-rank analysis. Statistical analyseswere performed with the SPSS version 11.0 software (SPSSInc., Chicago, Illinois).

Results

Baseline. Each group included 110 patients. Mean age ofthe patients was 65 � 5 years; 73% were male. Completebaseline patient characteristics are reported in Table 1.Patients had severe heart failure with a mean LVEF of 30 �8% and a mean left ventricular end-diastolic diameter of67 � 12 mm.

The 2 groups were comparable in terms of left ventricularend-diastolic diameter, NYHA functional class, hemody-

Abbreviationsand Acronyms

ACEI � angiotensin-converting enzyme inhibitor

ARB � angiotensin II-receptor blockers

BNP � brain natriureticpeptide

CHF � chronic heart failure

LVEF � left ventricularejection fraction

NYHA � New York HeartAssociation

1734 Jourdain et al. JACC Vol. 49, No. 16, 2007The STARS-BNP Study April 24, 2007:1733–9

namic parameters, QRS duration, and biological parame-ters. However, percentage of active smokers (p � 0.03) andmean LVEF (p � 0.05) were lower in the BNP group thanin the clinical group. Both groups were also similar forbackground medical therapies. At baseline, medical treat-ment was optimized in the population: 99% of patientsreceived an ACEI or an ARB at 94% of the recommendeddosage (Table 1). Similarly, 94% of patients received abeta-blocker at 58% of the recommended dosage. All thepatients received furosemide. Therefore, the backgroundtherapy was truly optimized before entry into the study.Titration phase: treatment adjustment during the first 3months. Changes in treatment occurred more frequently inthe BNP group compared with the clinical group, 134versus 66 occasions (p � 0.05). In the BNP group, thetreatment was adjusted according to the plasma BNP levelin 79% of the cases (106 of 134 treatment changes) (Fig. 1).

The most frequently changed pharmacological drugswere diuretics (41% of the changes in the BNP group vs.39% of the changes in the clinical group, p � NS) (Fig. 2).The mean increase in furosemide dosage was 9 � 20 mg andwas similar in both groups. However, all types of drugs werechanged more frequently in the BNP group (furosemide:55% vs. 26% of the patients; spironolactone: 17% vs. 7% ofthe patients; ACEI or ARB: 21% vs. 9% of the patients;beta-blockers: 36% vs. 20% of the patients).

During the first 3 months, the mean daily dosage ofACEI increased in the 2 groups (from 94% to 98% of therecommended dosage in the clinical group, p � NS, and

from 94% to 106% of the recommended dosage in the BNPgroup, p � 0.05) as well as the mean daily dosage ofbeta-blockers (from 57% to 67% of the recommendeddosage in the clinical group, p � 0.05, and from 58% to 77%of the recommended dosage in the BNP group, p � 0.05).However, at the end of the first 3 months, the mean dosagesof ACEIs and beta-blockers, expressed as percentage of

Figure 1 Number of Changes inMedical Therapy During the First 3 Months

Changes were related to high brain natriureticpeptide (BNP) value according to protocol or clinical reasons.

Population Characteristics at Baseline According to the Group Allocation

Table 1 Population Characteristics at Baseline According to the Group Allocation

Clinical (n � 110) BNP (n � 110) p Value

Age (yrs) 66 � 6 65 � 5 NS

Male/female patients (n) 62/48 65/45 NS

Smoking (%) 53 39 0.03

Hypertension (%) 30 30 NS

Diabetes (%) 19 16 NS

Dyslipidemia (%) 60 46 NS

Duration of heart failure (months) 29 31 NS

Ischemic heart failure (%) 48 55 NS

Weight (kg) 77 � 17 76 � 18 NS

NYHA functional class 2.21 � 0.62 2.29 � 0.60 NS

Heart rate (beats/min) 69 � 13 68 � 13 NS

QRS duration (ms) 118 � 40 119 � 43 NS

Natremia (mmol/l) 137 � 13 137 � 13 NS

Creatininemia (�mol/l) 97 � 40 92 � 40 NS

LVEF (%) 31.8 � 8.4 29.9 � 7.7 0.05

LVEDD (mm) 69 � 11 67 � 12 NS

ACEI or ARB (%) 99 99 NS

At recommended daily dose (1) (%) 94 94 NS

Beta-blocker (%) 97 99 NS

At recommended daily dose (1) (%) 58 59 NS

Spironolactone (%) 22 25 NS

Furosemide (%) 100 100 NS

Mean daily dose (mg) 52 � 60 50 � 48 NS

ACEI � angiotensin-converting enzyme inhibitor; ARB � angiotensin receptor blocker; BNP � brain natriuretic peptide; LVEDD � left ventricularend-diastolic diameter; LVEF � left ventricular ejection fraction; NYHA � New York Heart Association.

1735JACC Vol. 49, No. 16, 2007 Jourdain et al.April 24, 2007:1733–9 The STARS-BNP Study

recommended dosage, were significantly higher in the BNPgroup than in the clinical group (p � 0.05).Titration phase: clinical and biological variations. Dur-ing the first 3 months, NYHA functional class improved inboth groups. There was no difference for mean NYHA classfor the 2 groups at the end of the titration phase (2.1 � 0.5vs. 2.0 � 0.4, p � NS). Blood pressure, heart rate, andweight decreased in a comparable way in the 2 groups(Fig. 3). Similar clinical adverse events rates occurred in the2 groups.

Creatininemia (Fig. 3) and azotemia tended to increaseduring the titration phase in the 2 groups (p � NS).Increase in creatininemia by more than 30% was observed in9% of the control group and in 7% of the BNP group (p �NS). There was no significant variation in potassium orsodium levels.Follow-up. No patient was lost to follow-up (Fig. 4).During follow-up (minimum 6 months, median 15months), the primary composite end point (unplannedhospital stays for heart failure or death related to heartfailure) was observed in 83 occasions in the 220 patients(38%) (25 of 110 [24%] in the BNP group vs. 57 of 110[52%] in the clinical group, p � 0.001). By log-rankanalysis, event-free survival was also significantly better in

Figure 3 Evolution of Systolic Blood Pressure, Heart Rate, Weight, and Creatininemia During the First 3 Months

Hemodynamic and biological variations during titration phase. BNP � brain natriuretic peptide; M � month.

Figure 2 Changes in Medical Therapy During theTitration Phase in BNP Group and Clinical Group

Therapeutic modifications during titration phase. ACEI �

angiotensin-converting enzyme inhibitor; BNP � brain natriuretic peptide.


the BNP group (84%) than in the clinical group (73%) (p �0.01) (Fig. 4).

All-cause death was not significantly different betweenthe groups (7 in the BNP group vs. 11 in the clinical group,p � NS). Death related to heart failure was observed in 3patients in the BNP group versus 9 patients in the clinicalgroup. Six patients died from non–CHF-related causes (2 ofcancer, 3 of pulmonary embolism, 1 of endocarditis).

All-cause hospital stays were not significantly differentbetween the 2 groups: 60 patients in the clinical groupversus 52 in the BNP group (p � NS).

Hospital stays for heart failure were observed in 22patients in the BNP group versus 48 patients in the clinicalgroup (p � 0.0001). Two patients in the BNP group versus10 patients in the clinical group were hospitalized twice ormore for acute heart failure decompensation (p � 0.02).

In the BNP group, mean plasma BNP levels significantlydecreased during follow-up from 352 � 260 pg/ml atbaseline to 284 � 180 pg/ml at 3-month follow-up (p �0.03). Accordingly, the target (�100 pg/ml) was reached in16% of patients at baseline to 33% at 3-month follow-up(p � 0.04) (Fig. 5).

Discussion

This study has shown that CHF treatment optimizationbased on plasma BNP levels is associated with a lower risk

of death related to heart failure or hospital stay related toheart failure than the usual strategy based on clinicalexpertise. The benefit was observed despite optimization oftherapy in both groups by cardiologists specialized incongestive heart failure before entry into the study (morethan 95% of the patients received a recommended tripletherapy: ACEI or ARB � beta-blocker � furosemide).

Suboptimal dosages of beta-blockers were reported in theEuroHeart Failure Survey (29), much lower than thatprescribed at baseline in our study. However, in our study,dosage at baseline remained below those reached duringrecent multicenter beta-blocker trials (30,31), and inclusioninto the study led to significant increase in diuretics, ACEIs,and beta-blockers in both groups. Optimization is actually along-term aim, and a patient’s condition varies from time totime. Beta-blocker and ACEI increases were well tolerated.

Previous studies have examined BNP variations and theirprognostic implications during medical therapy. In the firstlandmark study, Throughton et al. (26) observed a signifi-cant reduction in all-cause mortality and in heart failure-related hospital stay with BNP-guided medical therapy.However, the sample size was limited (69 patients), andmedical therapy mainly consisted of ACEI and diureticswithout beta-blockers.

Brain natriuretic peptide is a noninvasive marker ofventricular myocyte stretch and correlates with left end-diastolic pressure (9,32). High BNP values can identifypatients with high left end-diastolic pressures who wouldbenefit from medical optimization (32). Interestingly, in ourstudy, diuretic dosages and weight variations were similar inboth groups, whereas plasma BNP level is dependent onblood volume.

Figure 4 Event-Free (Hospital Stay for Heart Failure or DeathRelated to Heart Failure) Survival in the 2 Groups

Event-free survival. BNP � brain natriuretic peptide.

Figure 5 Plasma BNP Level in BNP Group During TitrationPhase and % of Patients Reaching Target BNP Value

BNP � brain natriuretic peptide; M � month.


It had been expected that a high plasma BNP level wouldlead to an increase in diuretic dosage. In fact, the use ofplasma BNP levels led also to an increase in ACEIs andbeta-blockers. During the trial, we found that the associa-tion of abnormal BNP levels plus clinical examination wasmore meaningful to the cardiologist and led to moreoptimization in medications than clinical examinationalone. Whether knowledge of BNP plasma level is beneficialthrough better evaluation of the CHF status of the patientor acts as a supplementary stimulus for increasing all drugsremains to be determined.Study limitations. In our study, the benefit of the BNP-guided strategy was tested after medical therapy had beenoptimized by highly qualified cardiologists. The benefitmight actually be larger in routine clinical practice.

Similarly, baseline LVEF tended to be lower in the BNPgroup, which would be expected to increase the number ofclinical events. This actually strengthens the results of ourstudy.

As in other randomized studies, our population wasyoung, mostly male, and limited to systolic dysfunction(LVEF �45%). Medical registers demonstrate that CHFpatients are older and frequently female and diastolic dys-function is increasingly prevalent. Our results only apply tothe selected population included.

Conclusions

In optimally treated CHF patients with left ventricularsystolic dysfunction, a BNP-guided strategy reduces theincidence of a combined end point (death and hospital stayrelated to heart failure) compared with a standard strategy.The result is mainly obtained through an increase in ACEIand beta-blocker dosages.

AcknowledgmentsThe following is a list of the French investigators involvedin the STARS-BNP study: National Coordinator: PatrickJourdain (Pontoise); Investigators: Marie-Claude Aumont(Paris), Jean-François Aupetit (Lyon), Jean-Marc Davy(Montpellier), Michel Desnos (Paris), Erwan Donal(Poitiers), Jean-Christophe Eicher (Dijon), François Funck(Pontoise), Pascal De Groote (Lille), Michel Galinier(Toulouse), Pierre Gibelin (Nice), Jean-Yves Gueffet(Nantes), Guillaume Jondeau (Boulogne-Billancourt), YvesJuillière, Chairman of the French Working Group on HeartFailure (Nancy), Yves Kettelers (Armentières), MichelKomajda (Paris), Alain Le Helloco (Rennes), and DamienLogeart (Clichy).

Reprint requests and correspondence: Dr. Patrick Jourdain,Hopital de Pontoise, CH Rene Dubos, Heart Failure Department,6 Avenue de France, Pontoise, 95301 France. E-mail: [email protected].

REFERENCES

1. The Task Force for the Diagnosis and Treatment of Chronic HeartFailure. ESC guidelines for the diagnosis and treatment of chronicheart failure. Eur Heart J 2001;22:1527–60.

2. Hunt SA, Baker DW, Chin MH, et al. ACC/AHA guidelines for theevaluation 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.Circulation 2001;104:2996–3007.

3. Cleland JGF, Swedberg K, Follath F, et al. The EuroHeart Failuresurvey program—a survey on the quality of care among patients withheart failure in Europe. Part 1: patient characteristics and diagnosis.Eur Heart J 2003;24:442–63.

4. McCullough PA, Philbin EF, Spertus JA, et al. Confirmation of aheart failure epidemic: findings from the Resource Utilization AmongCongestive Heart Failure (REACH) study. J Am Coll Cardiol2002;39:60–9.

5. Redfield MM. Heart failure, an epidemic of uncertain proportions.N Engl J Med 2002;347:1442–4.

6. Nakao K, Ogawa Y, Suga SI, et al. Molecular biology and biochemistryof the natriuretic peptide system. J Hypertension 1992;10:907–12.

7. Yasue H, Yoshimura M, Sumida H, et al. Localization and mechanismof secretion of B-type natriuretic peptide in comparison with those ofA-type natriuretic peptide in normal subjects and patients with heartfailure. Circulation 1994;90:195–203.

8. Hosoda K, Nakao K, Mukoyama M, et al. Expression of brainnatriuretic peptide gene in human heart: production in the ventricle.Hypertension 1991;17:1152–6.

9. Levin ER, Gardner DG, Samson W. Natriuretic peptides. N EnglJ Med 1998;339:321–8.

10. Maisel AS, Krishnaswamy P, Nowak RM, et al. Rapid measurementof B-type natriuretic peptide in the emergency diagnosis of heartfailure. N Engl J Med 2002;347:161–7.

11. McCullough PA, Nowak RM, McCord J, et al. B-type natriureticpeptide and clinical judgment in emergency diagnosis of heart failure:analysis from Breathing Not Properly (BNP) multinational study.Circulation 2002;106:416–22.

12. Stanek B, Frey B, Hulsmann M, et al. Prognostic evaluation ofneurohumoral plasma levels before and during beta-blocker therapy inadvanced left ventricular dysfunction. J Am Coll Cardiol 2001;38:436–42.

13. Koglin J, Pehlivanli S, Schwaiblmair M, et al. Role of brain natriureticpeptide in risk stratification of patients with congestive heart failure.J Am Coll Cardiol 2001;38:1934–41.

14. Bettencourt P, Ferreira A, Dias P, et al. Predictors of prognosis in mildto moderate heart failure. J Card Fail 2000;6:306–13.

15. Gardner RS, Ozalo F, Murday AJ, Robb SD, McDonagh TA.N terminal pro-brain natriuretic peptide. A new gold standard inpredicting mortality in patients with advanced heart failure. EurHeart J 2003;24:1735– 43.

16. Yoshimura M, Mizuno Y, Nakayama M, et al. B-type natriureticpeptide as a marker of the effects of enalapril in patients with heartfailure. Am J Med 2002;112:716–20.

17. Ferreira A, Bettencourt P, Dias P, et al. Neurohormonal activation,the renal dopaminergic system and sodium handling in patients withsevere heart failure under vasodilator therapy. Clin Sci (Lond) 2001;1:100–5.

18. Murdoch DR, McDonagh TA, Byrne J, et al. Titration of vasodilatortherapy in chronic heart failure according to plasma brain natriureticpeptide concentration: randomized comparison of the hemodynamicand neuroendocrine effects of tailored versus empirical therapy. AmHeart J 1999;138:1126–32.

19. Latini R, Mason S, Anand I, et al. Effects of valsartan on circulatingbrain natriuretic peptide and norepinephrine in symptomatic chronicheart failure: the Valsartan Heart Failure Trial (Val-HeFT). Circulation2002;106:2454–8.

20. Tsutamoto T, Wada A, Maeda K, et al. Effect of spironolactone onplasma brain natriuretic peptide in patients with congestive heartfailure. J Am Coll Cardiol 2001;37:1228–33.


21. Bettencourt P, Frioes F, Azevedo A, et al. Prognostic informationprovided by serial measurements of brain natriuretic peptide in heartfailure. Int J Cardiol 2004;93:45–8.

22. Richards AM, Doughty R, Nicholls MG, et al. Plasma N-terminalpro-brain natriuretic peptide and adrenomedullin: prognostic utilityand prediction of benefit from carvedilol in chronic ischemic leftventricular dysfunction. J Am Coll Cardiol 2001;37:1781–7.

23. Richards A, Doughty R, Nicholls M. Neurohumoral prediction ofbenefit from carvedilol in ischaemic left ventricular dysfunction.Circulation 1999;99:786–92.

24. Fung JW, Yu CM, Yip G, et al. Effect of beta-blockade (carvedilol ormetoprolol) on activation of the renin-angiotensin-aldosterone system andnatriuretic peptides in chronic heart failure. Am J Cardiol 2003;92:406–10.

25. Cheng V, Kazanagra R, Garcia A, et al. A rapid bedside test for B-typepeptide predicts treatment outcomes in patients admitted for decom-pensated heart failure: a pilot study. J Am Coll Cardiol 2001;37:386–91.

26. Troughton RW, Frampton CM, Yandle TG. Treatment of heartfailure guided by plasma aminoterminal brain natriuretic peptide(N-BNP) concentrations. Lancet 2000;355:1126–30.

27. Richards M, Throughton RW. NT pro-BNP in heart failure therapydecision and monitoring. Eur J Heart Failure 2004;6:351–4.

28. Vogeser M, Jacob K. B-type natriuretic peptide (BNP)—validation ofan immediate response assay. Clin Lab 2001;47:29–33.

29. The Study Group of Diagnosis of the Working Group on HeartFailure of the European Society of Cardiology. The EuroHeart FailureSurvey program—a survey on the quality of care among patients withheart failure in Europe. Part 2: treatment. Eur Heart J 2003;24:464–74.

30. Packer M, Coats AJ, Fowler MB, et al. Effect of carvedilol on survivalin severe chronic heart failure. N Engl J Med 2001;344:1651–8.

31. Poole-Wilson PA, Swedberg K, Cleland JGF, et al. Comparison ofcarvedilol and metoprolol on clinical outcomes in patients with chronicheart failure in the Carvedilol Or Metoprolol European Trial(COMET): randomised controlled trial. Lancet 2003;362:7–13.

32. Maeda K, Tsutamoto T, Wada A. Plasma brain natriuretic peptide asa biochemical marker of high left ventricular end-diastolic pressure inpatients with symptomatic left ventricular dysfunction. Am Heart J1998;135:825–32.


based CHDR assessment system is feasible, has beenadopted by many physicians, and could be used widely.

References1. Expert Panel on Detection, Evaluation and Treatment of High Blood Choles-

terol in Adults. Summary of the second report of the National CholesterolEducation Program (NCEP) Expert Panel on Detection, Evaluation, andTreatment of High Blood Cholesterol in Adults (Adult Treatment Panel II).JAMA 1993;269:3015–23.

2. British Cardiac Society, British Hyperlipidaemia Association, British Hyper-tension Society. Joint British recommendations on prevention of coronarydisease in clinical practice. Heart 1998;80(Suppl2):S1–29.

3. Wood D, De Backer G, Faergeman O, Graham I, Mancia G, Pyorala K.Prevention of coronary heart disease in clinical practice: recommendationsof the Second Joint Task Force of European and other Societies on CoronaryPrevention. Atherosclerosis 1998;140:199–270.

4. National Heart Foundation. Clinical guidelines for the assessment andmanagement of dyslipidaemia. N Z Med J 1996;109:224–32.

5. Anderson KM, Odell PM, Wilson PWF, Kannel WB. Cardiovascular diseaserisk profiles. Am Heart J 1991;121:293–8.

6. Ramsay LE, Haq IU, Jackson PR, Yeo WW, Pickin DM, Payne JN. Targetinglipid-lowering drug therapy for primary prevention of coronary disease: anupdated Sheffield table. Lancet 1996;348:387–8.

7. Jones AF, Walker J, Jewkes C, Game FL, Bartlett WA, Marshall T, Bayly GR.Comparative accuracy of cardiovascular risk prediction methods in generalpractice patients. Heart 2001;85:37–43.

8. Bayly GR, Bartlett WA, Davies PH, Husband D, Haddon A, Game FL, Jones AF.Laboratory-based calculation of coronary heart disease risk in a hospitaldiabetic clinic. Diabet Med 1999;16:697–701.

9. Fraser CG, Harris EK. Generation and application of data on biologicalvariation in clinical chemistry. CRC Crit Rev Clin Lab Sci 1989;27:409–37.

10. Durrington PN, Prais H, Bhatnagar D, France M, Crowley V, Khan J, MorganJ. Indication for cholesterol-lowering medication: comparison of risk-assess-ment method. Lancet 1999;353:278–81.

11. Isles CG, Ritchie LD, Murchie P, Norrie J. Risk assessment in primaryprevention of coronary heart disease: randomized comparison of threescoring methods. BMJ 2000;320:690–1.

12. Jones AF. Statins and hypercholestrolaemia: UK Standing Medical AdvisoryCommittee Guidelines. Lancet 1997;350:1174–5.

13. Jones AF, Game FL. Cardiovascular risk assessment. Mod HypertensManage 1999;1:10–3.

14. Hingorani AD, Vallance P. A simple computer program for guiding manage-ment of cardiovascular risk factors and prescribing. BMJ 1999;318:101–5.

15. The National Service framework of coronary heart disease. Modern stan-dards and service models. London: Department of Health, 2000.

Evaluation of a New, Rapid Bedside Test for Quantita-tive Determination of B-Type Natriuretic Peptide, YurikoFischer,1* Karsten Filzmaier,2 Hugo Stiegler,1 Jurgen Graf,2

Simone Fuhs,2 Andreas Franke,2 Uwe Janssens,2 and Axel M.Gressner1 (1 Institut fur Klinische Chemie und Pathobio-chemie, and 2 Medizinische Klinik I, Universitatsklinikumder RWTH Aachen, Pauwelsstrasse 30, 52074 Aachen,Germany; * author for correspondence: fax 49-241-8888512, e-mail [email protected])

Congestive heart failure (CHF) is among the most fre-quently encountered cardiac diagnoses, with an estimatedprevalence of 1% (1 ). Plasma concentrations of B-typenatriuretic peptide (BNP) and its precursor, N-terminalpro-BNP (NT pro-BNP), are increased in patients withCHF and have been shown to accurately predict clinicalseverity and left ventricular ejection fraction (LVEF) aswell as morbidity and mortality in those patients (2, 3). Amajor limitation in the routine determination of natri-uretic peptides is the time-consuming nature of analyticaltechniques, e.g., extraction procedures, long incubation

times, or radioactive labeling (4–8). A rapid bedside testfor determination of BNP was introduced recently (TriageBNP; Biosite Diagnostics, San Diego, CA). We evaluatedthe assay and compared it, in samples from patients withsuspected CHF, with other assay systems for determina-tion of either BNP (Shionoria BNP; CIS Diagnostics) or NTpro-BNP (Roche Diagnostics, Tutzing, Germany).

The Triage BNP test is an immunofluorometric assayfor quantitative determination of BNP in EDTA-anticoag-ulated whole blood or plasma (9, 10). A murine recombi-nant polyclonal antibody is bound to the fluorescent label,and a murine monoclonal antibody against the disulfidebond-mediated ring structure of BNP 32 is bound to thesolid phase.

Briefly, after the addition of 250 mL of EDTA-anticoag-ulated whole blood, the plasma is separated and allowedto react with fluorescent antibody conjugates within areaction chamber. After an incubation period, complexesof the analyte and fluorescent antibody conjugates arecaptured on a detection lane. The concentration of BNP inthe specimen is proportional to the fluorescence bound tothe detection lane, which is determined quantitatively bya handheld fluorescence instrument (Triage Meter) asdescribed in detail elsewhere (11 ). The time from appli-cation of the sample until the result is reported is ;15min.

The Shionoria BNP test is a one-step immunoradiomet-ric assay that uses two different monoclonal antibodiesthat recognize the C-terminal structure and the disulfidebond-mediated ring structure of BNP 32, respectively.This test requires ;20 h. NT pro-BNP(1–76) was deter-mined by a recently developed research enzyme immu-noassay that uses a polyclonal antibody specific for theN-terminal sequence (amino acids 1–21) of NT pro-BNPas well as a polyclonal antibody specific for the middlesequence (amino acids 30–38) of NT pro-BNP. The assaytime is ;2 h, and no extraction is required. Both assayshave been described in detail elsewhere (6, 12–14).

The within-run (n 5 11) and between-day (n 5 11 days)imprecision of the Triage BNP assay were determined atthree different concentrations that were prepared bymixing patient samples with high and low BNP concen-trations. Within-run and between-day studies usedEDTA-anticoagulated whole blood and EDTA plasma,respectively, each with a single assay lot and unchangedcalibration.

For determination of the detection limit, defined as thelowest concentration that could be differentiated fromzero (mean 1 3 SD), a zero calibrator was measured 20times with four lots of reagents on 5 consecutive days.

Plasma samples from 50 apparently healthy individuals(30 males, 20 females; age range, 20–60 years) weremeasured, and the median and 97.5th percentile concen-trations were calculated.

The study population consisted of 100 patients withunderlying cardiac disease and suspected CHF. Clinicalseverity was characterized according to the functionalclassification system of the New York Heart Association(NYHA) in stages I to IV (15 ). The LVEF was evaluated

Clinical Chemistry 47, No. 3, 2001 591

with biplane transthoracic echocardiography by the mod-ified Simpson rule using second harmonic imaging (16 ).Five patients were excluded because of insufficient sam-ple material or missing clinical data. A total of 95 patientsremained in the study group (Table 1). All patients gaveinformed consent. The study was approved by the localethics committee.

On the day of echocardiography, peripheral venousblood was collected into sampling tubes containing EDTAas the anticoagulant. Within 20 min after venipuncture,BNP concentrations were determined by the Triage sys-tem. The remaining sample material was centrifuged, andEDTA plasma was frozen at 280 °C until further deter-mination.

Correlation of BNP Triage with Shionoria BNP and NTpro-BNP was evaluated by the statistical procedure byPassing and Bablok (17 ). For calculation of assay compar-ison, results of the 95 study patients and the 50 healthysubjects were included (n 5 145). Results of ROC analysisare expressed as areas under the individual ROC curves,including the 95% confidence interval. The diagnosticsensitivity of each assay was calculated as the number oftest results with a BNP or NT pro-BNP value equal to orhigher than a specific cutoff value among all patients withan LVEF #50%, whereas diagnostic specificity was de-fined as the number of test results below the cutoff valueamong all patients with an LVEF .50%.

For imprecision, the within-run CVs of the Triage BNPassay were 9.4%, 13%, and 15% at 40, 450, and 800 ng/L,

respectively. At the same mean concentrations, the be-tween-day CVs were 11%, 13%, and 16%, respectively.The detection limit was 6 ng/L. The 97.5th percentile ofthe healthy subjects was 82 ng/L for the Triage BNPassay, 64 ng/L for the Shionoria BNP, and 57 pmol/L forthe NT pro-BNP ELISA.

Comparison of results by regression analysis (Fig. 1)revealed the following: Triage BNP 5 1.579(ShionoriaBNP) 2 2.947 (r 5 0.963); Triage BNP 5 1.198(NT pro-BNP) 1 1.419 (r 5 0.947); NT pro-BNP 5 1.309(ShionoriaBNP) 2 3.780 (r 5 0.948). The 95% confidence intervals forslope and intercept were as follows: Triage BNP vsShionoria BNP, 1.490–1.659 and 24.909 to 21.884; TriageBNP vs NT pro-BNP, 1.136–1.270 and 20.664 to 3.500; NTpro-BNP vs Shionoria BNP, 1.251–1.388 and 26.493 to21.004. Comparison of BNP and NT pro-BNP concentra-tions with NYHA classes revealed only moderate corre-lations for all assays: rs 5 0.689 (Triage BNP), 0.667(Shionoria BNP) and 0.632 (NT pro-BNP); P ,0.001; n 595. Nevertheless, BNP and NT pro-BNP values showed astrong inverse correlation with LVEF: rs 5 20.816 (TriageBNP), 20.762 (Shionoria BNP), and 20.718 (NT pro-BNP);P ,0.001; n 5 95.

ROC analysis regarding impaired ventricular function,defined as LVEF #50%, revealed areas under the ROCcurve (95% confidence interval) of 0.91 (0.83–0.98) forTriage BNP, 0.88 (0.80–0.95) for Shionogi BNP, and 0.86(0.77–0.94) for NT pro-BNP, respectively. At a cutoffvalue of 130 ng/L, the Triage assay had a sensitivity of

Table 1. Patient characteristics according to LVEF (n 5 95).LVEF

>50% 40–50% 30–39% <30%

n 53 14 14 14Age, mean (range), years 60 (19–85) 65 (52–86) 67 (58–76) 61 (41–75)M/F 37/16 11/3 8/6 8/6NYHA classificationa

I 38 (72%) 2 (14%)II 8 (15%) 8 (57%) 7 (50%) 5 (36%)III 7 (13%) 5 (36%) 4 (29%) 5 (36%)IV 1 (7%) 1 (7%) 4 (28%)

Underlying cardiac diseaseb

CADc 29 (55%) 11 (79%) 10 (71%) 2 (21%)HTN 5 (9%) 3 (21%) 3 (21%)DCM 1 (2%) 1 (7%) 3 (21%) 8 (57%)ARRH 11 (21%) 2 (14%) 3 (21%) 3 (21%)VHD 3 (6%) 2 (14%) 1 (7%)Otherd 16 (30%) 1 (7%) 1 (7%) 1 (7%)

BNPe

Triage, ng/L 60 (28–109) 333 (217–616) 482 (229–1087) 811 (500–1420)Shionoria, ng/L 45 (18–83) 186 (92–415) 242 (121–541) 375 (261–892)NT pro-BNP, pmol/L 64 (31–104) 277 (106–650) 321 (195–798) 538 (312–973)a Absolute values and percentages given.b Suspected or diagnosed underlying cardiac disease in patient population. Absolute values and percentages given, because of comorbidity numbers, may exceed

total number of patients.c CAD, coronary artery disease; HTN, hypertension; DCM, dilated cardiomyopathy; ARRH, arrhythmias; VHD, valvular heart disease.d Myocarditis, pericarditis, pulmonary embolism, pneumonia, chronic obstructive pulmonary disease.e Values for BNP and NT pro-BNP are given as median and 25th and 75th percentiles.

592 Technical Briefs

93% and a specificity of 79% to detect patients withimpaired LVEF. Similar results were obtained for theother assays: for the Shionogi BNP, at a cutoff of 100ng/L, the sensitivity was 80% and the specificity was 79%;for the NT pro-BNP assay, at a cutoff of 100 pmol/L, thesensitivity was 90% and the specificity was 66%.

The Triage BNP assay seems to be a sensitive screeningmethod to decide which patient with suspected CHFwarrants further investigation, particularly when assess-ment by echocardiography is not readily available. Themajor advantage of the Triage test system compared withthe other investigated assays is its rapid and accuratemeasurement of BNP from whole blood with 24-h avail-ability in a routine laboratory or at the point of care.Additional time-consuming preparation, centrifugation,extraction, and incubation steps can be omitted. Rapidmeasurement of either BNP or NT pro-BNP seems useful,especially in the triage of patients with suspected CHFpresenting to the emergency room, where these resultsadd important information for faster diagnostic evalua-tion (8, 18). It might also be useful in patients admittedwith decompensated CHF as a previous study has shownthat changes in BNP during treatment are a strong pre-dictor for mortality and early readmission (19 ). Further-more, the recently introduced BNP-guided therapy forchronic CHF will require faster determination of theanalyte (20 ). Nevertheless, additional clinical trials needto be designed to clarify the detailed clinical benefit inpatients diagnosis, management, and therapy as well asthe cost-effectiveness of this marker (8 ).

We gratefully acknowledge VIVA Diagnostika Germany,Biosite Diagnostics Europe, and Roche Diagnostics Ger-many for providing reagents free of cost.

References1. Cowie MR, Mosterd A, Wood DA, Deckers JW, Poole-Wilson PA, Sutton GC,

Grobbee DE. The epidemiology of heart failure. Eur Heart J 1997;18:208–25.

2. Tsutamoto T, Wada A, Maeda K, Hisanaga T, Mabuchi N, Hayashi M, et al.Plasma brain natriuretic peptide level as a biochemical marker of morbidityand mortality in patients with asymptomatic or minimally symptomatic leftventricular dysfunction. Comparison with plasma angiotensin II and endo-thelin-1. Eur Heart J 1999;20:1799–807.

3. Richards AM, Nicholls MG, Yandle TG, Frampton C, Espiner EA, Turner JG, etal. Plasma N-terminal pro-brain natriuretic peptide and adrenomedullin: newneurohormonal predictor of left ventricular function and prognosis aftermyocardial infarction. Circulation 1998;97:1921–9.

4. Mair J, Friedl W, Thomas S, Puschendorf B. Natriuretic peptides in assess-ment of left-ventricular dysfunction. Scand J Clin Lab Invest 1999;59(Suppl230):132–42.

5. Yandle T, Fisher S, Espiner E, Richards M, Nicholls G. Validating aminoter-minal BNP assays: a word of caution. Lancet 1999;353:1068–9.

6. Karl J, Borgya A, Gallusser A, Huber E, Krueger K, Rollinger W, Schenk J.Development of a novel, N-terminal-proBNP (NT-proBNP) assay with a lowdetection limit. Scand J Clin Lab Invest 1999;59(Suppl 230):177–81.

7. Missbichler A, Hawa G, Woloszczuk W, Schmal N, Hartter E. Enzymeimmunoassays for proBNP fragments (8–29) and (32–57). J Lab Med1999;4:241–4.

8. Clerico A, Del Ry S, Giannessi D. Measurement of cardiac natriuretichormones (atrial natriuretic peptide, brain natriuretic peptide, and relatedpeptides) in clinical practice: the need for a new generation of immunoassaymethods. Clin Chem 2000;46:1529–34.

9. Buechler KF, Dwyer BP, Noar B, Tadesse L, Moi S. A fluorescence energytransfer detection system for immunoassays of biological samples [Ab-stract]. Clin Chem 1997;43(Suppl 6):S136.

10. McPherson PH, Sundquist AR, Anderberg JM, Lesefko SM, Nakamura KK,Buechler JA, et al. A rapid, quantitative point-of-care assay system for themeasurement of B-type natriuretic peptide in blood [Abstract]. Clin Chem2000;46(Suppl 6):A84.

11. Apple FS, Christenson RH, Valdes R Jr, Andriak AJ, Berg A, Duh SH, et al.Simultaneous rapid measurement of whole blood myoglobin, creatinekinase MB, and cardiac troponin I by the triage cardiac panel for detectionof myocardial infarction. Clin Chem 1999;45:199–205.

12. Shimitsu H, Aono K, Masuta K, Asada H, Misaki A, Teraoka H. Stability ofbrain natriuretic peptide (BNP) in human blood samples. Clin Chim Acta1999;285:169–72.

13. Niederau C, Fischer Y, Stiegler H, Kolbe-Busch S, Haass M, Karl J, et al.Stability of N-terminal pro-brain natriuretic peptide and brain natriureticpeptide in different sampling media and varying sample handling [Abstract].Clin Chem 1999;45(Suppl):A142.

14. Mukoyama M, Nakao K, Hosoda K, Suga S, Saito Y, Ogawa Y, et al. Brainnatriuretic peptide as a novel cardiac hormone in humans. Evidence for an

Fig. 1. Method comparison using the Passing and Bablok (17) regression model.(A), Triage BNP (ng/L) vs Shionoria BNP (ng/L): Triage BNP 5 1.579(Shionoria BNP) 2 2.947; r 5 0.963; n 5 145. (B), Triage BNP (ng/L) vs NT pro-BNP (pmol/L):Triage BNP 5 1.198(NT pro-BNP) 1 1.419; r 5 0.947; n 5 145. The solid line represents the regression line; the dotted line denotes the identity y 5 x. Note thatfor better visualization, concentrations .1000 ng/L (pmol/L) are not shown.

Clinical Chemistry 47, No. 3, 2001 593

exquisite dual natriuretic peptide system, atrial natriuretic peptide and brainnatriuretic peptide. J Clin Invest 1991;87:1402–12.

15. The Task Force on Heart Failure of the European Society of Cardiology.Guidelines for the diagnosis of heart failure. Eur Heart J 1995;16:741–51.

16. Jensen-Urstad K, Bouvier F, Hojer J, Ruiz H, Hulting J, Samad B, et al.Comparison of different echocardiographic methods with radionuclide imag-ing for measuring left ventricular ejection fraction during acute myocardialinfarction treated by thrombolytic therapy. Am J Cardiol 1998;81:538–44.

17. Passing H, Bablok W. A new biometrical procedure for testing the equality ofmeasurements from two different analytical methods. J Clin Chem ClinBiochem 1983;21:709–20.

18. Maisel A, Dao P, Krishnaswamy P, Kazanegra R, Lenert L, Clopton P. Utilityof B-natriuretic peptide (BNP) in the diagnosis of congestive heart failure inan urgent care setting [Abstract]. Eur Heart J 2000;21(Suppl):291.

19. Maisel A, Cheng V, Krishnaswamy P, Kazanegra R, Garcia A, Gardetto N. Arapid bedside test for B-type natriuretic peptide predicts treatment out-comes in patients with decompensated heart failure [Abstract]. Eur Heart J2000;21(Suppl):585.

20. Troughton RW, Frampton CM, Yandle TG, Espiner EA, Nicholls MG, RichardsAM. Treatment of heart failure guided by plasma aminoterminal brainnatriuretic peptide (N-BNP) concentrations. Lancet 2000;355:1126–30.

Measurement of Low Apolipoprotein Concentrations byOptimized Immunoturbidimetric Applications, MustafaPorsch-Ozcurumez,1* Sabine Westphal,2 and Claus Luley2

(1 Institute of Clinical Chemistry, University of Regens-burg, Franz-Josef-Strauss-Allee 11, 93042 Regensburg,Germany; 2 Institute of Clinical Chemistry, University ofMagdeburg, Leipziger Strasse 44, 39120 Magdeburg, Ger-many; * address correspondence to this author at: Univer-sitatsklinikum Regensburg, Institut fur Klinische Chemieund Blutbank, Franz-Josef-Strauss-Allee 11, 93042 Re-gensburg, Germany; fax 49-941-944-6202, e-mail [email protected])

Apolipoproteins define the functional properties of li-poprotein particles. In addition to their stabilizing fea-tures, several apolipoproteins have ligand functions.Some are also responsible for the modulation of enzymesinvolved in the homeostasis of lipid metabolism. Thus,apolipoprotein concentrations may provide essential in-formation about lipid metabolism and associated dis-eases. There is evidence, for example, that apolipoproteinconcentrations provide clinically relevant informationconcerning risk for coronary heart disease (1 ).

The determination of apolipoproteins in distinct li-poprotein subfractions narrows the conclusions that canbe drawn for risk assessment as determined by or thediscriminative power of HDL- and LDL-cholesterol (2, 3).In research, the determination of apolipoproteins encom-passes their measurement in cell culture supernatants (4 ),lipoprotein subfractions separated by ultracentrifugation(5 ), isotachophoresis (6 ), immunoaffinity chromatogra-phy (7 ), or size-exclusion chromatography (8 ). The mea-surement of apolipoproteins in samples derived fromsuch procedures requires appropriate methods reliable atconcentrations below the physiological range.

Five analytical techniques are commonly used to quan-tify apolipoproteins. No delipidation is necessary in anyof the advanced assays.

The use of radioactive reagents in RIAs is problematic

(9 ). Radial immunodiffusion (RID) is simple to performbut time-consuming. Difficulties can occur when analyz-ing lipemic sera if the diffusion of the lipoprotein particlesis complicated by particle size (5 ). Electroimmunodiffu-sion using the Laurell-Rocket technique (9 ) requires lesstime than RID, but large amounts of antibodies areneeded. ELISAs provide several advantages (10 ), includ-ing good precision and a sensitivity comparable to RIAs.ELISAs are useful in routine clinical determinations be-cause of the availability of automated methods. Nephe-lometry and immunoturbidimetry provide additional ad-vantages in apolipoprotein measurement (11, 12).Whereas ELISAs are highly sensitive, nephelometry andimmunoturbidimetry are superior with respect to preci-sion, time, and cost. Therefore, nephelometry and immu-noturbidimetry seem to be the most suitable methods forroutine analysis of apolipoproteins.

To improve the relatively low sensitivity comparedwith ELISA, we optimized and evaluated immunoturbi-dimetric applications for low concentrations of the mostfrequently measured apolipoproteins, i.e., apolipoprotein(Apo)A-I, ApoA-II, ApoB, ApoC-III, and ApoE, attaining12.5- to 45-fold higher sensitivities for these assays.

Commercially available assays for ApoA-I and ApoB(“ApoA1, immunologischer Trubungstest” and “ApoB,immunologischer Trubungstest”) were obtained fromRolf Greiner Biochemica and modified to measure phys-iological concentrations of ApoA-II, ApoC-III, and ApoE.Assays were performed on a Hitachi 911 automatedanalyzer (Boehringer Mannheim).

The application characteristics of assays for physiolog-ical applications and the final low-range conditions aregiven in Table 1. Initially, samples were incubated for 5min with buffer 1 (“Immunofluid”; Greiner Biochemica)containing different concentrations of polyethylene glycol(PEG 10000) dissolved and stabilized in 100 mmol/L Trisbuffer (pH 7.5) as provided ready to use by the manufac-turers. Sample volumes were increased 10- to 12.5-fold inlow-range applications. Commercially available humanpolyclonal goat antibodies specific against ApoA-I, A-II,B, C-III, and E (Rolf Greiner Biochemica) were usedwithout additional dilution in all assays. Antisera wereadded after the first incubation step. The resulting absor-bance was determined after an incubation interval ofanother 5 min.

Calibrators and control sera were purchased from Behr-ingwerke [“N-Apolipoprotein” and “ApolipoproteinControl Serum CHD (human)”]. Analytical values ofApoA-I and ApoB were based on IFCC reference prepa-rations (13 ). For ApoE, reference values provided by themanufacturers for nephelometry were used. The ApoC-IIIconcentration of the N-Apolipoprotein standard was re-peatedly measured by RID, and the mean value was usedfor calibration. Calibrators for the low-range applicationswere within the experimentally determined linearityrange (see below) starting with a dilution of 1:16 of theN-Apolipoprotein standard. In all assays, a six-pointcalibration was performed. Sample concentrations wereestimated by the logit-log method.

594 Technical Briefs

The new england journal of medicine - ugap.fr file february 12, 2004 647 The new england journal of...

Documents

Transcript of The new england journal of medicine - ugap.fr file february 12, 2004 647 The new england journal of...