From American Heart Journal

Effect of Body Mass Index on Natriuretic Peptide Levels in Patients With Acute Congestive Heart Failure: A ProBNP
Investigation of Dyspnea in the Emergency Department (PRIDE) Substudy
Daniel G. Krauser, MD; Donald M. Lloyd-Jones, MD, ScM; Claudia U. Chae, MD, MPH; Renee Cameron, MS; Saif Anwaruddin,
MD; Aaron L. Baggish, MD; Annabel Chen, MD; Roderick Tung, MD; James L. Januzzi Jr., MD
Posted: 05/18/2005; American Heart Journal. 2005;149(4):744-750. © 2005 Mosby, Inc.
Abstract and Introduction

Abstract

Background: Obesity is associated with lower B-type natriuretic peptide (BNP) levels in healthy individuals and patients with
chronic congestive heart failure (CHF). Neither the mechanism of natriuretic peptide suppression in the obese patient nor whether
obesity affects natriuretic peptide levels among patients with acute CHF is known.

Methods: The associations of amino-terminal pro-BNP (NT-proBNP), BNP, and body mass index (BMI) were examined in 204
subjects with acute CHF. Multivariable regression analyses were performed to identify factors independently related to NT-proBNP
and BNP levels.

Results: Across clinical strata of normal (< 25 kg/m2), overweight (25-29.9 kg/m2), and obese (≥ 30 kg/m2) patients, median NT-
proBNP and BNP levels decreased with increasing BMI (both P values < .001). In multivariable analyses adjusting for covariates
known to affect BNP levels, the inverse relationship between BMI and both NT-proBNP and BNP remained ( P < .05 for both). Using
a cut point of 900 pg/mL, NT-proBNP was falsely negative in up to 10% of CHF cases in overweight patients (25-29.9 kg/m2) and
15% in obese patients (≥ 30 kg/m2). Using the standard cut point of 100 pg/mL, BNP testing was falsely negative in 20% of CHF
cases in both overweight and obese patients. The assays for NT-proBNP and BNP exhibited similar overall sensitivity for the
diagnosis of CHF.
Conclusions: When adjusted for relevant covariates, compared with normal counterparts, overweight and obese patients with acute
CHF have lower circulating NT-proBNP and BNP levels, suggesting a BMI-related defect in natriuretic peptide secretion. NT-
proBNP fell below the diagnostic cutoff for CHF less often than BNP in overweight and obese individuals; however, when used as a
diagnostic tool to identify CHF in such patients, both markers may have reduced sensitivity.

Introduction

Obesity is a major risk factor for development of congestive heart failure (CHF).[1—3] The clinical diagnosis of acute CHF is often
challenging,[4] and this challenge is accentuated in the obese patient, in whom history and physical examination may be limited in
sensitivity and specificity. Therefore, improvements in the diagnostic evaluation of obese patients with suspected CHF are needed.

Recent attention has focused on the use of B-type natriuretic peptide (BNP) measurement for the diagnosis of CHF. This class of
biomarkers has proven useful for identifying patients with CHF, even when standard clinical assessment is not informative.[5] Thus,
BNPs may hold special promise for diagnosis of CHF in obese patients. B-type natriuretic peptide is a 32-amino-acid protein
synthesized by myocardium. BNP is originally derived from an intracellular 108 amino acid precursor, which is cleaved into 2
fragments, yielding a 76-amino-acid N-terminal fragment (NT-proBNP) and BNP. Both peptides are secreted and are measurable in
peripheral blood specimens of patients with acute CHF.[6, 7]
B-type natriuretic peptide levels were recently shown to be inversely proportional to body mass index (BMI) in individuals without
heart failure[8, 9] and those with chronic CHF,[10] possibly because of either increased peptide clearance[11—14] or reduced secretion of
natriuretic peptides in obese patients.[15, 16] Whether BMI would be expected to affect the sensitivity of the marker in the setting of
acute CHF is not known. Furthermore, data are sparse regarding the effect of BMI on NT-proBNP levels. Accordingly, we performed
the present study in an effort to clarify the association between BMI and natriuretic peptides in patients with acute CHF and to attempt
to elucidate the mechanism of reduced natriuretic peptide levels in the obese state.

Methods

Study Sample

The details of the PRIDE study have recently been described.[7] The study examined 600 patients with acute dyspnea to evaluate the
role of NT-proBNP levels in the diagnosis of acute CHF. For each subject, a study physician was provided with all hospital records
pertaining to the subject, starting from the time of emergency department presentation through to the results of the 60-day follow-up.
These records included (when available) office notes, hospital discharge records, and results of all laboratory and imaging tests. In
addition, at 60 days an attempt was made to contact every patient and review subsequent medical course after presentation. Using the
data available from presentation through the 60-day review, the physician, blinded to the results of NT-proBNP testing, assigned the
cause for each patient's presentation. In cases where the diagnosis was unclear or in doubt, an adjudicated diagnosis was rendered by a
second physician, using the Framingham Heart Study criteria for diagnosis of CHF. The institutional review board approved all
investigational procedures involved in this study.

Of the patients enrolled, 209 were judged to have acute CHF as the cause of their dyspnea, and therefore were eligible for this
analysis. Five patients were excluded because of missing covariate data, leaving 204 patients for the present analysis. Body mass
index was calculated for these patients as weight (in kilograms) divided by the square of height (in meters). Patients with diabetes
mellitus were identified by their use of insulin and/or oral diabetes medications. Patients with hypertension were identified by systolic
blood pressure ≥ 140 mm Hg or diastolic blood pressure ≥ 90 mm Hg at presentation, or by their historical use of antihypertensive
therapy. Left ventricular ejection fraction was determined from historical information as assessed within 6 months by echocardiogram
or echocardiographic assessment during the index presentation. Echocardiographic assessment was not performed on all subjects.
Creatinine clearance was calculated using the Cockcroft-Gault equation[17] and the Modification of Diet in Renal Disease (MDRD)
Study Group equation.[18]

Plasma Natriuretic Peptide Measurements

Blood was collected from enrolled patients upon presentation to the emergency department, within 2 hours after administration of any
loop diuretic. After collection, EDTA plasma treated blood samples were processed and frozen for later analysis. NT-proBNP and
BNP were measured using commercially available immunoassays (Elecsys ProBNP, Roche Diagnostics, Indianapolis, Ind; ADVIA
Centaur BNP, Bayer Diagnostics, Tarrytown, NY), using established methodology.

Statistical Analyses

Body mass index was treated as both a continuous and a categorical variable using the World Health Organization/National Institutes
of Health classification scheme: normal < 25 kg/m2, overweight 25 to 29.9 kg/m2, and obese ≥ 30 kg/m2.[19] Comparisons of clinical
characteristics between patients in each BMI category were performed using χ2 tests for categorical data and the Kruskal-Wallis test
for continuous data.

In the initial analyses, NT-proBNP and BNP were considered as continuous variables and were log-transformed to achieve normality.
Univariable and multivariable linear regressions were performed using log-transformed NT-proBNP and BNP separately as the
dependent variables. Covariates examined for inclusion in the multivariable models were age, gender, symptom severity (defined
using New York Heart Association [NYHA] heart failure classification), pulse rate, serum creatinine level, creatinine clearance,
history of cardiomyopathy, atrial fibrillation at presentation, prior CHF, prior myocardial infarction, initial troponin T level (Elecsys
Troponin T STAT, Roche Diagnostics), outpatient loop diuretic use before presentation, diabetes mellitus, and hypertension.[7]

Covariates that were strongly associated with natriuretic peptide levels ( P < .10 in univariate analyses) were included in the
multivariate analysis. In additional models, we used categories (normal, overweight, or obese) to represent BMI. All models were
fitted, found to be appropriate, and tested for first-order interactions. Based on the findings of interactions in the models, we created
stratified models to clarify the effect of BMI on natriuretic peptide levels.

To examine the potential false-negative rate for natriuretic peptides among overweight and obese patients with CHF, we performed
logistic regression analyses to determine predictors of NT-proBNP and BNP levels below previously described diagnostic thresholds.
We evaluated several different diagnostic thresholds for the definition of a low NT-proBNP, including the current standard cut points
of < 125 pg/mL for patients aged < 75 years, and < 450 pg/mL for patients aged ≥ 75 years ("age-stratified" cut points), as well as a
single cut point of < 900 pg/mL for all patients, as previously identified.[7] A low BNP was defined as < 100 pg/mL as previously
defined.[6] Odds ratios (ORs) and 95% CIs for having low natriuretic peptide levels were estimated, according to BMI category, with
BMI < 25 kg/m2 as the referent group. Odds ratios were adjusted for the same covariates as in the linear models. Sensitivities of both
NT-proBNP and BNP for each BMI category were calculated using the above cut points.

All analyses were performed with the use of STATA software, version 8 (STATA Corp, College Station, Tex). A 2-sided P value <
.05 was considered to indicate statistical significance.

Results

Overall, 59 (29%) participants were overweight, 64 (31%) participants obese, and 81 (40%) were normal. Body mass index ranged
from 15.5 to 51.7 kg/m2 (mean ± SD = 27.6 ± 6.6 kg/m2). Morbidly obese patients (BMI > 40 kg/m2) accounted for 14 (7%) of the
204 study subjects. Characteristics of the study sample by clinical strata of BMI are shown in Table I . Overweight and obese patients
were younger, had higher creatinine clearance, and a higher prevalence of diabetes (all P < .05) than patients with normal BMIs.

Table I. Baseline Patient Characteristics Expressed as a Function of BMI

Median NT-proBNP and BNP levels were lower in overweight and obese individuals compared with individuals with normal BMI ( P
< .001 for both, Figure 1). When considered as a continuous variable and across deciles of BMI, both NT-proBNP and BNP were
inversely associated with BMI ( P < .001) (Figure 2), with a 12% decrease in both NT-proBNP and BNP levels associated with each
BMI decile increase and a 6% decrease associated with each BMI unit increase.

Figure 1. Box plots demonstrating unadjusted median levels of A , NT-proBNP and B , BNP among patients with acute CHF. Boxes
represent interquartile ranges; and whiskers, the highest and lowest value in each group.
Figure 2. Relationship between BMI and log-adjusted natriuretic peptide levels. For each decile increase in BMI a 12% decrease in
natriuretic peptide levels was observed, with a corresponding 6% drop with each 1 unit increase in BMI ( r2 for NT-proBN P =
0.0856, P < .001; r2 for BN P = 0.1091, P < .001).

Obesity and Plasma Natriuretic Peptide Levels

In univariate analysis, several variables were significantly associated with NT-proBNP levels: age ( P = .02), NYHA class ( P = .02),
serum creatinine level ( P < .001) and calculated creatinine clearance ( P < .001), history of prior CHF ( P = .03), history of prior
cardiomyopathy ( P = .04), and BMI ( P < .001). The following variables were significant univariate predictors of BNP levels: age ( P
= .05), gender ( P = .008), serum creatinine level ( P < .001), history of prior CHF ( P = .07), history of prior cardiomyopathy ( P =
.01), calculated creatinine clearance ( P < .001), and BMI ( P < .001). While both the Cockroft-Gault and MDRD equations were used
for the estimation of calculated creatinine clearance, for both NT-proBNP and BNP, the Cockroft-Gault equation demonstrated best fit
in the univariate models, and was thus used in the multivariable model.

In multivariable analyses, after adjustment for relevant covariates, BMI (as a continuous variable) was still inversely associated with
plasma natriuretic peptide levels ( Table II ), with a 3% decrease associated with each 1 unit increase in BMI. In the models predicting
NT-proBNP, there were significant interactions between BMI and age, creatinine clearance, and gender. In stratified models, the
effect of BMI was only significant in young (age < 50 years) female patients with CHF ( P = .002), who accounted for only 16.7% of
the study sample. In the models predicting BNP levels, there were significant interactions between BMI and age or creatinine
clearance. Examination of stratified models revealed that most of the effect of BMI on BNP levels was observed among older patients
(age > 50 years) with low creatinine clearance ( P = .007), a patient subgroup accounting for 39.2% of the study sample.

Table II. Multivariable Linear Regression Results

Obesity and Natriuretic Peptide Sensitivity for CHF Diagnosis

Using the standard age-stratified cut points for NT-proBNP, none of the overweight or obese patients had a false-negative NT-
proBNP value (100% sensitivity), although across all patients in the PRIDE study, use of these NT-proBNP cut points was associated
with a specificity of 62%, reflecting their value as rule-out thresholds.
After multivariable adjustment, a BMI of 25 to 29.9 kg/m2 was associated with having a 2.7-fold increase in the odds of having a BNP
level < 100 pg/mL (95% CI 1.0-7.3, P = .05) but no significant increase in the odds of having an NT-proBNP < 900 pg/mL (OR 1.40,
95% CI 0.39-5.1, P = .60). Obesity was associated with a similar rate of false-negative BNP levels (OR 2.44, 95% CI 0.90-6.6, P =
.08) and false-negative NT-proBNP level < 900 pg/mL (OR 3.1, 95% CI 1.03-9.61, P = .046). Overall, for each 1 unit increase in
BMI, there was an 8% greater chance of having a BNP < 100 pg/mL and a 10% greater chance of having an NT-proBNP < 900
pg/mL ( P < .001 for both) ( Table III ).

Table III. The Influence of BMI on the Odds of Having an NT-ProBNP < 900 pg/mL or BNP < 100 pg/mL in the Setting of
Acute CHF

While 20% of CHF patients in both the overweight and obese groups had BNP values < 100 pg/mL, 10% of overweight CHF patients
had an NT-proBNP < 900 pg/mL, and 15% of obese patients had an NT-proBNP < 900 pg/mL ( Table IV ). For patients with BMI >
25 kg/m2 (overweight plus obese groups), the overall sensitivities for BNP and NT-proBNP (using a cut point of < 900 pg/mL) were
80% and 87%, respectively ( P = .17).

Table IV. Distribution of Plasma NT-ProBNP and BNP Values (pg/mL) Expressed as a Function of BMI
Discussion

Obesity affects more than 40 million Americans, and between 280000 and 325000 deaths are attributed to obesity annually in the
United States.[20] Both figures are rapidly increasing. Obesity-related illnesses, such as diabetes and hypertension, are risk factors for
CHF,[1—3] which represents a major source of morbidity and mortality among overweight and obese patients. As overweight and obese
patients with dyspnea are often the most difficult to assess clinically, BNP assays have the potential to play a major role in the
diagnosis of acute CHF in this patient population. However, an inverse association between BMI and natriuretic peptide levels has
been noted in several groups, including healthy individuals,[8] nondyspneic obese individuals,[9] as well as patients with chronic
CHF,[10] hypertension[12] and acute coronary syndromes.[21] In the present study of patients with acute CHF, we now demonstrate this
inverse relationship between BMI and BNP, but also demonstrate the same relationship between BMI and NT-proBNP.

We found an independent, inverse relationship between BMI and both NT-proBNP and BNP levels in patients with acute CHF, with
overweight and obese individuals demonstrating significantly lower natriuretic peptide levels in the setting of acute CHF than
individuals with a normal BMI. This finding is robust, even after adjustment for factors known to affect natriuretic peptide levels, and
for differences in clinical variables between obese and nonobese individuals. Our data demonstrate that BMI should be taken into
consideration when interpreting natriuretic peptide results, and that previously established optimal diagnostic thresholds for CHF may
be less valid for overweight and particularly obese individuals. Indeed, our data suggest that obesity is associated with a 2- to 3-fold
higher odds of having a natriuretic peptide value below these optimal thresholds (< 900 pg/mL for NT-proBNP and < 100 pg/mL for
BNP) for the diagnosis of CHF. That the previously established age-stratified rule-out thresholds for NT-proBNP had 100%
sensitivity suggests the possibility that a lower cut point for overweight and obese patients may be reasonable to consider.

Our results may shed light on the mechanisms for reduced natriuretic peptide levels in obesity. Natriuretic peptide clearance receptors
(NPR-Cs) are abundant in adipose tissue, and thus it has been suggested that adipocytes or adipocyte products such as neutral
endopeptidases (NEPs) may participate in the clearance of BNP from the circulation.[10—12, 14, 22]
However, as NT-proBNP is not
cleared by NPR-C or NEP degradation, one would not expect NT-proBNP levels to be affected by this "natriuretic peptide handicap"9
if it was because of increases in such clearance mechanisms. In adjusted multivariable models, NT-proBNP decreased with increasing
BMI in a manner nearly parallel to the decrease of BNP. These results lend support to the hypothesis that reduced secretion of
natriuretic peptides from either diminished myocardial hormone release[15] or impaired synthesis[16] is the major cause of the inverse
relationship between natriuretic peptides and BMI. A similar inverse relationship between BMI and N-terminal atrial natriuretic
peptide was observed in obese individuals without overt CHF,[9] further supporting the hypothesis of reduced release of natriuretic
peptides with increasing BMI. Further data are needed to understand the relative importance of these mechanisms.

Although the sensitivities of both biomarker assays were negatively affected by rising BMI, we detected differences between NT-
proBNP and BNP. In particular, in overweight patients (BMI 25-29.9 kg/m2) using cut points offering optimal balance of sensitivity
and specificity in acute CHF for both markers,[6, 7] NT-proBNP (10%) was less prone than BNP (20%) to fall below the diagnostic cut
point for acute CHF. Furthermore, interaction testing suggested that the association between BMI and natriuretic peptide levels varied
between the 2 markers, with NT-proBNP levels affected by BMI in only a small subset (16.7%) of our patient sample, specifically
women aged < 50 years with CHF. Conversely, BNP levels were most affected by BMI in patients aged > 50 years with impaired
renal function, who represented 39.2% of the study sample. Thus, in the balance, the results may favor NT-proBNP over BNP for the
diagnosis of acute CHF in overweight and obese patients.

The strengths of our study include an adequate sample size to adjust for factors that influence natriuretic peptide levels, the
measurement of both NT-proBNP and BNP in the same sample of patients with adjudicated acute CHF, and the ascertainment and
adjustment for several relevant covariates affecting NT-proBNP and BNP levels. There were several limitations, however. Left
ventricular ejection fraction, an important predictor of NT-proBNP and BNP levels, was not included as a covariate in our
multivariable analysis because it was not measured on all patients. However, a history of prior cardiomyopathy was used to take
systolic dysfunction into account when analyzing the data, and no difference existed between the 3 groups when the available ejection
fraction data were expressed as a function of BMI. All patients in this study had acute CHF, and thus the sample as a whole had
relatively high plasma natriuretic peptide levels, which were distributed in a nonnormal manner. Log adjustments were used to
account for the skewed distribution of NT-proBNP and BNP levels. Nevertheless, the consistency of our results across both NT-
proBNP and BNP and the similarity to non-CHF studies of the BNPs[8, 9, 12, 21] support the validity of our clinical findings.

Our findings may have several implications. First, above-normal BMI should be taken into consideration when interpreting natriuretic
peptide results in the setting of suspected CHF; the best diagnostic thresholds for CHF may be different for overweight and obese
individuals than for normal individuals. This observation merits further investigation. Second, one may hypothesize that the favorable
response to exogenous natriuretic peptide therapy may be magnified in the setting of obesity.[23] Lastly, and importantly, similar to
overweight and obese patients without acute CHF,[9] the inverse relationship of both NT-proBNP and BNP with BMI in acute CHF
supports the hypothesis that there is a derangement in natriuretic peptide synthesis or secretion in overweight and obese individuals
with acute CHF, with consequent abnormalities in salt and water handling, and exaggerated activation of the sympathetic and renin-
angiotensin systems among such patients.

The results of this study may assist physicians to prevent misdiagnosis in obese patients with CHF. In addition, the finding of reduced
natriuretic peptide release in overweight and obese patients supports the possibility of a novel mechanism of hypertension and
vascular dysregulation in these subjects and may identify a potential therapeutic target to manage obesity-related cardiovascular
disease.

References

1. Lauer MS, Anderson KM, Kannel WB. The impact of obesity on left ventricular mass and geometry: the Framingham Heart
Study. JAMA 1991;266:231-236.
2. Schunkert H. Obesity and target organ damage: the heart. Int J Obes 2002;26:S15-S20.
3. Kenchaiah S, Evans JC, Levy D. Obesity and the risk of heart failure. N Engl J Med 2002;347:305-313.
4. Sparrow N, Adlam D, Cowley A. The diagnosis of heart failure in general practice: implications for the UK National Service
Framework. Eur J Heart Fail 2003;5:349-354.
5. McCullough PA, Nowak RM, McCord J. B-type natriuretic peptide and clinical judgment in emergency diagnosis of heart
failure: analysis from Breathing Not Properly (BNP) Multinational Study. Circulation 2002;106:416-422.
6. Maisel AS, Krishnaswamy P, Kazanegra R. Rapid measurement of B-type natriuretic peptide in the emergency diagnosis of
heart failure. N Engl J Med 2002;347:161-167.
7. Januzzi JL, Camargo CA, Anwaruddin S. N-terminal ProBNP for emergency evaluation of shortness of breath: the ProBNP
Investigation of Dyspnea in the Emergency Department (PRIDE) Study. J Am Coll Card 2004;43:325(a).
8. Wang TJ, Larson MG, Levy D. Impact of age and sex on plasma natriuretic peptide levels in healthy adults. Am J Cardiol
2002;90:254-258.
9. Wang TJ, Larson MG, Levy D. Impact of obesity on plasma natriuretic peptide levels. Circulation 2004;109:594-600.
10. Mehra MR, Uber PA, Park MH. Obesity and suppressed B-type natriuretic peptide levels in heart failure. J Am Coll Cardiol
2004;43:1590-1595.
11. Sarzani R, Dessi-Fulgheri P, Paci VM. Expression of natriuretic peptide receptors in human adipose and other tissues. J
Endocrinol Invest 1996;19:581-585.
12. Dessi-Fulgheri P, Sarzani R, Tamburrini P. Plasma atrial natriuretic peptide and natriuretic peptide gene expression in
adipose tissue of normotensive and hypertensive obese patients. J Hypertens 1997;15:1695-1699.
 13. Dessi-Fulgheri P, Sarzani R, Rappelli A. The natriuretic peptide system in obesity-related hypertension: new pathological
aspects. J Nephrol 1998;11:296-299.
14. Sarzani R, Dessi-Fulgheri P, Salvi F. A novel promoter variant of the natriuretic peptide clearance receptor gene is associated
with lower atrial natriuretic peptide and higher blood pressure in obese hypertensives. J Hypertens 1999;17:1301-1305.
15. Licata G, Volpe M, Scaglione R. Salt-regulating hormones in young normotensive obese subjects: effects of saline load.
Hypertension 1994;23:120-124.
16. Morabito D, Vallotton MB, Lang U. Obesity is associated with impaired ventricular protein kinase C-MAP kinase signaling
and altered ANP mRNA expression in the heart of adult Zucker rats. J Investig Med 2001;49:310-318.
17. Charleson HA, Bailey RR, Stewart A. Quick prediction of creatinine clearance without the necessity of urine collection. N Z
Med J 1980;92:425-426.
18. Levey AS, Bosch JB, Lewis JB. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new
prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med 1999;130:461-470.
19. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults. Bethesda (Md):
National Heart Lung, and Blood Institute; 1998.
20. Allison DB, Fontaine KR, Manson JE. Annual deaths attributable to obesity in the United States. JAMA 1999;282:1530-
1538.
21. James SK, Lindahl B, Siegbahn A. N-terminal pro-brain natriuretic peptide and other risk markers for the separate prediction
of mortality and subsequent myocardial infarction in patients with unstable coronary artery disease: a Global Utilization of
Strategies to Open occluded arteries (GUSTO)-IV substudy. Circulation 2003;108:275-281.
22. Sarzani R, Paci VM, Dessi-Fulgheri P. Comparative analysis of atrial natriuretic peptide receptor expression in rat tissues. J
Hypertens Suppl 1993;11:S214-S215.
23. Dessi-Fulgheri P, Sarzani R, Serenelli M. Low calorie diet enhances renal, hemodynamic, and humoral effects of exogenous
atrial natriuretic peptide in obese hypertensives. Hypertension 1999;33:658-662.

Authors and Disclosures

Daniel G. Krauser, MD,a Donald M. Lloyd-Jones, MD, ScM,b Claudia U. Chae, MD, MPH,a Renee Cameron, MS,a Saif
Anwaruddin, MD,a Aaron L. Baggish, MD,a Annabel Chen, MD,a Roderick Tung, MD,a and James L. Januzzi Jr., MD.a

a
Department of Medicine, Massachusetts General Hospital and Harvard Medical School, Boston, Mass
b
Department of Preventive Medicine, Feinberg School of Medicine, Northwestern University, Chicago, Ill.

Funding Information
This study was supported by a grant from Roche Diagnostics, Indianapolis, Ind
Reprint Address
James L. Januzzi Jr, MD, Massachusetts General Hospital, Bulfinch 019, 55 Fruit Street, Boston, MA 02114. Email address:
jjanuzzi@partners.org (James L. Januzzi)
American Heart Journal. 2005;149(4):744-750. © 2005 Mosby, Inc.
Copyright © Mosby-Year Book, Inc. All rights reserved.