Author + information
- Received May 22, 2005
- Revision received July 9, 2005
- Accepted August 1, 2005
- Published online January 3, 2006.
- Tamara B. Horwich, MD,
- Michele A. Hamilton, MD, FACC and
- Gregg C. Fonarow, MD, FACC⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Gregg C. Fonarow, Ahmanson-UCLA Cardiomyopathy Center, UCLA Division of Cardiology CHS 47-123, 10833 Le Conte Avenue, Los Angeles, California 90095-1679.
B-type natriuretic peptide (BNP) facilitates the diagnosis of heart failure (HF) (1) and in patients with pre-existing HF, BNP correlates with HF severity, ventricular filling pressures (2,3), and HF outcomes (4). The BNP levels are affected by demographic variables such as age, gender, and ethnicity, and clinical characteristics such as hypertension, atrial fibrillation, and renal function (5–7).
Recent reports suggest that obesity, as indexed by elevated body mass index (BMI), also affects BNP levels, with lower circulating levels in those with a higher BMI. The inverse relationship between BNP and BMI has been observed in diverse populations, including healthy individuals, patients presenting to emergency departments with dyspnea, and chronic HF patients (6,8,9).
Obesity, an increasingly prevalent condition in the U.S., contributes to the development of HF (10). In patients with established HF, overweight and obesity are common, occurring in approximately 30% of patients (11), and thus further elucidation of the relationship between BNP and BMI is critical. Our study aimed to further explore the association between elevated BMI and BNP in advanced HF, focusing on the influence of co-existing factors such as gender, age, hypertension, and diabetes. We also address a clinically relevant, as yet unanswered question: does the presence of obesity affect the prognostic value of BNP in advanced HF?
The study population consisted of 316 consecutive HF patients seen at a single university medical center for HF management and/or heart transplantation evaluation between July 2000 and June 2004. Subjects with left ventricular ejection fraction (LVEF) >40%, creatinine >2.5 mg/dl, or a clinical diagnosis of acute coronary syndrome or myocarditis at time of referral were excluded. The BNP was measured in all patients within one month of initial referral date through the use of an industry-standard analytical platform (Triage, Biosite, San Diego, California). Medical record review was approved by the Medical Institutional Review Board of the University of California, Los Angeles.
Patients were divided into three BMI categories based on World Health Organization/National Institutes of Health guidelines: lean (BMI <25.0 kg/m2), overweight (BMI 25.0 to 29.9 kg/m2), and obese (BMI ≥30.0 kg/m2) (12). Differences in baseline characteristics between groups were analyzed using independent samples ttest, analysis of variance, chi-square, Kruskal-Wallis, and Mann-Whitney tests, as appropriate. The p values for analysis of variance test the hypothesis that several means are equal. Multivariate logistic regression analysis was used to estimate odds ratios (OR) for low BNP (BNP <100 pg/ml) in overweight and obese patients. Receiver operator curves for prediction of one-year outcomes by BNP were generated in each BMI category; optimal cutoff points were defined by maximization of the sum of sensitivity and specificity. Product-moment survival estimates were calculated using the Kaplan-Meier method. Cox regression models estimated the one- and two-year hazard ratios and 95% confidence intervals of BNP levels in predicting mortality. The primary end points analyzed were: 1) all-cause mortality, with censoring at the time of any heart transplantation, and 2) survival free from death or urgent transplantation (status Ia), with censoring at time of non-urgent transplantation (status Ib or II). All statistics were calculated with Statistical Package for Social Sciences version 12.0 for Windows (SPSS Inc., Chicago, Illinois).
Obesity and baseline patient characteristics
Mean age of the cohort was 53 ± 13 years, mean LVEF was 24 ± 7%, and 74% of the cohort was male. Baseline characteristics of the population according to BMI category are shown in Table 1.The HF medications were similar among the three BMI groups, with the exception of slighter higher rates of beta-blocker and statin usage in overweight and obese subjects.
Relationship between BMI and BNP levels
Circulating BNP levels were significantly lower in overweight and obese HF patients compared with lean HF patients. Median (interquartile ranges) values for the lean, overweight, and obese groups were 747 (272 to 1,300), 380 (143 to 856), and 332 (113 to 617) pg/ml, respectively (Fig. 1).The BNP levels in patients with a very low BMI (<20.0 kg/m2) compared with patients with a normal BMI (20.0 to 24.9 kg/m2) were 839 (77 to 1,300) pg/ml versus 734 (317 to 1,300) pg/ml (p = 0.40). The proportion of patients with BNP above 1,000 pg/ml was progressively lower with increases in BMI across a broad range (Fig. 2).Furthermore, the association between high BMI and low BNP was consistent in clinically significant subgroups, with the exception of diabetic patients (Fig. 3).
A multivariate logistic regression analysis was performed to determine independent predictors of low BNP (<100 pg/ml, n = 52), adjusting for age, gender, race, hypertension, diabetes, LVEF, coronary artery disease, New York Heart Association functional class, creatinine, and medications (angiotensin converting enzyme inhibitor, angiotensin receptor antagonist, beta-blocker, and aldosterone antagonist). Obesity was associated with a greater than six-fold increase in the odds of having low BNP (OR, 6.7; 95% confidence interval [CI], 1.2 to 40.4). Per unit increase in BMI, the odds of having a low BNP increased by 13% (OR, 1.1; 95% CI, 1.0 to 1.3). After excluding underweight patients (BMI < 20), higher BMI was still an independent predictor of low BNP, with 23% increased odds of low BNP per unit increase in BMI (OR, 1.2; 95% CI, 1.1 to 1.4).
BNP, HF severity, and hemodynamics
BNP, obesity, and survival
The mean follow-up time was 456 days (15.2 months). There were 93 heart transplantations (67 urgent, status Ia, and 26 non-urgent, status Ib or II), 37 deaths, and 6 patients lost to follow-up during the first year. Progressive HF death accounted for 20 deaths, where as 6 were sudden, 2 were secondary to myocardial infarction, and 9 were other or unknown. One-year unadjusted survival estimates in the lean, overweight, and obese patients were 80%, 93%, and 88%, respectively (p < 0.02).
The BNP predicted outcomes independent of BMI. A BNP above the cohort median (452 pg/ml) was associated with impaired survival in lean, overweight, and obese subjects (Table 2).On multivariate analysis, BNP remained an independent predictor of one-year mortality in each BMI category (Table 3).
The receiver operator curve analysis to assess BNP as a predictor of one-year mortality or urgent heart transplantation showed that the area under the curve was >0.7 in each BMI category (Fig. 5A).Any BNP levels above the receiver operator curve-determined cut point predicted markedly increased all-cause mortality in each BMI category (Fig. 5B).
In light of the increasing prevalence of both HF and obesity, understanding the relationship between BMI and BNP levels is integral to the successful diagnosis and management of HF. This study shows that overweight and obese HF patients (BMI ≥25.0 kg/m2) have significantly lower BNP levels compared with lean HF patients (BMI <25.0 kg/m2), confirming reports from previous investigators (6,8,9). The inverse relationship between BMI and BNP in this cohort was independent of factors known to affect circulating BNP levels, including age, race, gender, hypertension, renal function, and HF medications (5–7,9,13,14).
This study adds to the growing body of knowledge regarding natriuretic peptide levels and obesity (6,8,9). However, this is the first study to show that despite relatively less circulating BNP in HF patients who are overweight and obese, BNP retains its prognostic capacity in this cohort. The BNP not only predicted ventricular filling pressures and functional class, but also correlated with mortality at each level of BMI.
Mechanisms of low BNP in obesity
Given the association of obesity with multiple cardiovascular risk factors, the relationship between elevation of BMI and reduction in BNP is counterintuitive. Several potential mechanistic explanations deserve consideration. It is possible that the low natriuretic peptide levels in overweight and obese patients reflect less advanced stages of HF compared with lean patients. Elevated BMI was associated with improved outcomes in this cohort and has been reported previously (11). However, ventricular filling pressures were no different in obese and non-obese patients, arguing against differential ventricular wall stretch/stress as an explanation for differential BNP production and release in obese and non-obese HF patients.
Cardiac cachexia, a state characterized by weight loss and neurohumoral/cytokine activation (15,16), is another potential explanation for a low BMI-high BNP association. However, BNP was statistically similar in underweight (BMI <20.0 kg/m2) and normal-weight patients (BMI = 20.0 to 24.9 kg/m2). There is evidence that inflammatory cytokines such as tumor necrosis factor-alpha, which may elicit release of BNP from cardiomyocytes (17), are less active in states of obesity. However, an initial investigation found tumor necrosis factor-alpha and interleukin-6 to be similar in obese and non-obese HF patients (8).
Recent evidence suggests that there may be increased clearance of circulating BNP in obesity. Natriuretic peptide clearance receptors are abundant on human adipocytes (18). Additionally, the vascularity of adipose tissue may allow for increased degradation of BNP by neutral endopeptidase (19). Conversely, it is also possible that overweight and obesity are associated with less robust synthesis and/or release of BNP from the myocardium. Decreased N terminal pro-BNP and pro-atrial natriuretic peptide have been observed in obesity (6,20), which may be suggestive of decreased natriuretic peptide production.
Alternatively, decreased circulating BNP levels may not be a consequence of increased body weight or excess adiposity, but rather may be a causative factor in the genotype or phenotype that leads to development of obesity. Both ANP and BNP are now recognized to be involved in fat metabolism as stimulators of lipolysis in adipose tissue (21).
We acknowledge potential limitations to our study. Our cohort is a selected population with advanced HF referred to a tertiary center for disease management and transplantation evaluation. The number of events per BMI category was relatively low. The upper limit of detectable BNP increased from 1,300 to 5,000 pg/ml during the study period, although this adjustment likely affected all BMI categories to a similar extent. Both BMI and BNP may not have been assessed on the same day, although maximal delay was one month. We do not we have data on weight loss or measures of percent body fat, and also we did not measure additional neurohormones, cytokines, or adipokines for correlation with BNP or BMI.
In this advanced systolic HF cohort, overweight and obese patients had significantly lower circulating BNP levels compared with lean patients, and this relationship was independent of demographics, medications, or severity of HF. Despite the dampening of BNP levels in overweight and obese HF patients, BNP retained its prognostic capacity in this population, correlating not only with HF severity and hemodynamic status, but also predicting survival. These data suggest that the BNP assay is a useful tool for the management of chronic systolic heart failure at all levels of BMI, although increased risk may be seen at relatively lower levels of BNP in the obese.
This research was supported by the Ahmanson Foundation, Los Angeles, California. Dr. Horwich was funded by NIH training grant 401357JI30608. Dr. Fonarow holds the Eliot Corday Chair in Cardiovascular Medicine and Science.
- Abbreviations and Acronyms
- body mass index
- B-type natriuretic peptide
- confidence interval
- heart failure
- left ventricular ejection fraction
- odds ratio
- Received May 22, 2005.
- Revision received July 9, 2005.
- Accepted August 1, 2005.
- American College of Cardiology Foundation
- Koglin J.,
- Pehlivanli S.,
- Schwaiblmair M.,
- Vogeser M.,
- Cremer P.,
- vonScheidt W.
- Anand I.S.,
- Fisher L.D.,
- Chiang Y.T.,
- et al.
- Wang T.J.,
- Larson M.G.,
- Levy D.,
- Benjamin E.J.,
- Leip E.P.,
- Wilson P.W.,
- Vasan R.S.
- Tang W.H.,
- Girod J.P.,
- Lee M.J.,
- Starling R.C.,
- Young J.B.,
- Van Lente F.,
- Francis G.S.
- Mehra M.R.,
- Uber P.A.,
- Park M.H.,
- Scott R.L.,
- Ventura H.O.,
- Harris B.C.,
- Frohlich E.D.
- Horwich T.B.,
- Fonarow G.C.,
- Hamilton M.A.,
- MacLellan W.R.,
- Woo M.A.,
- Tillisch J.H.
- Kuczmarski R.J.,
- Flegal K.M.
- Ma K.K.,
- Ogawa T.,
- de Bold A.J.
- McCullough P.A.,
- Sandberg K.R.
- James S.K.,
- Lindahl B.,
- Siegbahn A.,
- et al.