Brain-derived neurotrophic factor (BDNF) is a member of the neurotrophin family, which regulates neuronal differentiation and functions. Low levels of BDNF are because of psychological stress and potentially play a role in the pathogenesis of depression and cognition disorders. Because psychological stress and depression are associated with increased risk of heart failure (HF), the pathogenic link between HF and psychological status has attracted clinical attention. We hypothesized that plasma BDNF levels might be decreased in patients with HF and that BDNF could be a key factor associated with HF. We evaluated plasma BDNF levels in 242 patients with HF and 80 subjects without HF who are age and gender matched. Plasma BDNF levels were significantly lower in patients with HF (3,712 pg/ml [2,124 to 6,180]) than those without HF (7,247 pg/ml [5,388 to 9,255], p <0.001) and lower in patients with HF with the New York Heart Association functional class III than class I (p = 0.01) and class II (p <0.001). Log BDNF levels correlated negatively with log B-type natriuretic peptide (r = −0.203, p = 0.03) in patients with HF. Of 61 acute decompensated patients with HF, plasma BDNF levels were significantly higher at discharge (4,194 pg/ml [2,356 to 6,916]) compared with those at admission (2,749 pg/ml [1,380 to 4,161], p = 0.003). Multivariate logistic regression analysis identified log BDNF level as a significant correlate with the presence of HF (odds ratio 0.82; 95% confidence interval 0.76 to 0.91, p <0.001). In conclusion, plasma BDNF levels were decreased in patients with HF and associated with HF severity. BDNF could be a potentially clinically useful biomarker of HF reflecting possible cardio-neuronal linkage.
Brain-derived neurotrophic factor (BDNF) is a member of the neurotrophin family, which regulates neuronal differentiation and affects the development and function of the nervous system. Low levels of BDNF are caused by psychological stress through the hypothalamic-pituitary-adrenal (HPA) axis and play a role in the pathogenesis of depression and cognition disorders. Several basic experiments reported that BDNF also enhances angiomyogenesis, contributes to survival of endothelial cells during cardiovascular development, and provides protection against cardiac remodeling after myocardial infarction. Therefore, it seems that BDNF plays an important role not only in the central nervous system but also in the cardiovascular system. Because psychological stress and depression are associated with increased risk and poor prognosis of cardiovascular disease and heart failure (HF), the pathogenic link between these diseases and mental status including depression have attracted clinical attention. It is reported that low-serum or plasma BDNF levels are associated with higher risk of future coronary and stroke events ; however, the BDNF levels in patients with HF have not been fully evaluated. The present study was designed to evaluate plasma BDNF levels in patients with HF, changes in plasma BDNF levels during management of acute decompensated HF (ADHF), and identify factors that determine such levels.
Methods
The study protocol was approved by the ethics committee of our institution. We retrospectively analyzed 242 patients with HF (mean age 71 ± 12 years, men 66%) who were admitted to Kumamoto University Hospital and the National Cerebral and Cardiovascular Center from October 2009 to August 2014. Of them, 134 patients were admitted to the emergency department because of ADHF, and the remaining patients were admitted for evaluation of the clinical condition or optimize the treatment strategy of HF as scheduled. The following inclusion criteria were applied in patient selection: (1) ≥20 years, (2) met the criteria for the diagnosis of HF set by the American College of Cardiology/American Heart Association, (3) diagnosis of ADHF was based on the Framingham criteria, (4) provided written informed consent for participation and publication of research data, and (5) clinically confirmed absence of acute coronary syndrome. Patients were classified as having HF of ischemic origin if they had a history of myocardial infarction or significant coronary artery disease.
We also included in this study a control group of 80 subjects without HF who are age and gender matched. These subjects were admitted to our hospital with suspected angina pectoris and later confirmed by coronary angiography to have no significant coronary stenosis.
Blood samples were collected from all patients under clinically stable condition. In patients with ADHF, we could obtain paired blood samples from 61 patients at the emergency department and at discharge after treatment for ADHF. The plasma samples were kept frozen at −80°C until analysis. Plasma BDNF levels were measured using the Human BDNF Quantikine ELISA Kit (R&D Systems, Minneapolis, Minnesota). The intra-assay and interassay variability were 3.8% to 6.2% and 7.6% to 11.3%, respectively. Plasma B-type natriuretic peptide (BNP) levels were measured using the MI02 Shionogi BNP kit (Shionogi, Osaka, Japan). The glomerular filtration rate (GFR) was calculated using the level-modified Modification of Diet in Renal Disease modified for Japanese. In this study, patients with HF and left ventricular ejection fraction (LVEF) >50% were classified as having “HF with preserved EF” and the remaining with HF and LVEF ≤50% were classified as having “HF with reduced EF.”
After blood sampling, patients with HF were followed in the outpatient clinic. All-cause mortality was identified by searching the medical records and confirmed through direct contact with the patients or relatives and caring physicians. Patients with HF were divided into low and high BDNF groups based on the median concentration of BDNF (3,712 pg/ml).
Normally distributed data were presented as mean ± SD, whereas data of variables with skewed distribution were expressed as medians with interquartile ranges. Differences between groups were examined by the Student’s t test or the Mann-Whitney U test for unpaired data. Categorical values were presented as numbers (percentage) and compared by the chi-square test or Fisher’s exact test as appropriate. Data of multiple groups were compared by both the Kruskal-Wallis and Mann-Whitney U tests with Bonferroni adjustment. Variables with skewed distribution were transformed logarithmically before Pearson’s correlation to fulfill the conditions required for the type of analysis. Linear relations between log BDNF levels and key variables were analyzed by univariate and multiple regression analyses. Univariate logistic regression analysis was performed to identify significant parameters for the presence of HF. Then, multivariate logistic regression analysis was performed by stepwise multivariate analysis. A forward stepwise algorithm (p <0.10 for entry and p <0.05 for stay) was used. The Hosmer-Lemeshow statistic was applied to assess model calibration. Because BNP is an established biomarker for the diagnosis of HF, we decided to exclude BNP in this regression analysis. Kaplan-Meier curve for all-cause mortality was compared between low BDNF and high BDNF groups. Univariate and multivariate Cox regression analyses were used to assess the prognostic association of BDNF levels. A 2-tailed value of p <0.05 was considered statistically significant. All statistical analyses were performed with the Statistical Package for Social Sciences, version 19 (SPSS Inc., Chicago, Illinois).
Results
Table 1 lists the baseline characteristics of the subjects with and without HF. The plasma BDNF levels were significantly lower in patients with HF than those without HF (3,712 pg/ml [2,124 to 6,180] vs 7,247 pg/ml [5,388 to 9,255], p <0.001). Among the study subjects, plasma BDNF levels were significantly lower in loop diuretics and aldosterone antagonists users compared with nonusers (3,582 pg/ml [2,228 to 5,609] vs 5,961 pg/ml [3,710 to 8,504], p <0.001; and 3,445 pg/ml [2,105 to 5,489] vs 5,817 pg/ml [3,642 to 8,297], p <0.001, respectively] and significantly higher in statin users compared with nonusers (4,920 pg/ml [3,201 to 7,966] vs 3,711 pg/ml [2,067 to 6,734], p = 0.002).
| Variable | Heart failure group (n=242) | Non- heart failure group (n=80) | p value |
|---|---|---|---|
| Age (years) | 71±12 | 69±9 | 0.19 |
| Males | 159 (66%) | 48 (60%) | 0.36 |
| Body mass index (kg/m 2 ) | 22.4±3.7 | 23.7±3.4 | 0.004 |
| New York Heart Association functional class (I/II/III) | 24/155/63 | Not available | |
| Hypertension ∗ | 158 (65%) | 60 (75%) | 0.11 |
| Diabetes mellitus | 91 (38%) | 29 (36%) | 0.83 |
| Dyslipidemia † | 152 (63%) | 72 (90%) | <0.001 |
| Prior heart failure hospitalization | 118 (49%) | Not available | |
| Ischemic etiology | 85 (35%) | Not available | |
| Laboratory data | |||
| Plasma brain-derived neurotrophic factor (pg/ml) | 3712 [2124-6180] | 7247 [5388-9255] | <0.001 |
| Plasma B-type natriuretic peptide (pg/ml) | 238 [117-427] | 20 [13-35] | <0.001 |
| Hemoglobin (g/dl) | 12.4±1.9 | 13.4±1.6 | <0.001 |
| Serum sodium (mEq/L) | 138±4 | 140±2 | <0.001 |
| Estimated glomerular filtration rate (ml/min/1.73m 2 ) | 48.0±19.5 | 66.7±16.1 | <0.001 |
| High density lipoprotein-cholesterol (mg/dl) | 49.4±31.2 | 48.9±11.2 | 0.89 |
| Low density lipoprotein-cholesterol (mg/dl) | 90.8±33.3 | 86.3±28.3 | 0.27 |
| Triglyceride (mg/dl) | 102.7±47.2 | 113.9±44.6 | 0.06 |
| High-sensitivity C-reactive protein (mg/ml) | 0.16 [0.05-0.47] | 0.05 [0.02-0.11] | <0.001 |
| Imaging findings | |||
| Left ventricular end-diastolic dimension (mm) | 54.9±10.8 | 44.9±5.2 | <0.001 |
| Left ventricular ejection fraction (%) | 41.8±14.2 | 62.7±6.0 | <0.001 |
| Heart failure with preserved ejection fraction | 66 (27%) | Not available | |
| Medications | |||
| Loop diuretics | 159 (66%) | 3 (4%) | <0.001 |
| β-blockers | 185 (76%) | 48 (60%) | 0.004 |
| Angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers | 211 (87%) | 40 (50%) | <0.001 |
| Aldosterone antagonists | 145 (60%) | 5 (6%) | <0.001 |
| Digitalis | 32 (13%) | 0 (0%) | 0.001 |
| Statins | 120 (50%) | 64 (80%) | <0.001 |
∗ Defined as treated with antihypertensive drugs or blood pressure ≥140/90 mm Hg.
† Defined as treated with lipid lowering drugs or low-density cholesterol ≥ 140 mg/dL and/or serum triglyceride ≥150 mg/dL and/or high-density cholesterol < 40 mg/dL.
Plasma BDNF levels were lower in patients with the New York Heart Association functional class III (2,563 pg/ml [1,580 to 4,284]) than those with the New York Heart Association functional class I (4,622 pg/ml [1,830 to 7,712], p = 0.01), class II (4,052 pg/ml [2,558 to 6,732], p <0.001), and subjects without HF (p <0.001; Figure 1 ). Among patients with HF, plasma BDNF levels were significantly lower in those with previous HF hospitalization, older age (≥75 years), renal dysfunction (median estimated GFR <47 ml/min/1.73 m 2 ), and high BNP levels (median ≥238 pg/ml; Figure 2 ). In contrast, plasma BDNF levels were not influenced by gender (p = 0.80), presence or absence of diabetes mellitus (p = 0.99), hypertension (p = 0.30), ischemic or nonischemic origin (p = 0.38), and preserved or reduced LVEF (p = 0.76).
Table 2 details the results of univariate and multivariate linear regression analyses for log BDNF in patients with HF. Log BDNF levels correlated negatively with log BNP levels (r = −0.203, p = 0.002; Figure 3 ). To examine the factors associated with log BDNF levels, we performed multiple regression using the forced inclusion model with age, body mass index, hemoglobin, estimated GFR, triglyceride, and log BNP levels. The analysis identified log BNP levels (β = −0.143, p = 0.04) as the only independent and significant parameter that correlated with log BDNF levels ( Table 2 ).
| Variable | Univariate Analysis | Multivariate Analysis | ||
|---|---|---|---|---|
| r | p value | β | p value | |
| Age | -0.142 | 0.03 | -0.048 | 0.51 |
| Body mass index | 0.126 | 0.05 | 0.068 | 0.31 |
| Log (B-type natriuretic peptide) | -0.203 | 0.002 | -0.143 | 0.04 |
| Hemoglobin | 0.126 | 0.05 | 0.054 | 0.44 |
| Serum sodium | 0.000 | 0.99 | ||
| Estimated glomerular filtration rate | 0.137 | 0.03 | 0.042 | 0.59 |
| High density lipoprotein-cholesterol | 0.086 | 0.18 | ||
| Low density lipoprotein-cholesterol | 0.064 | 0.32 | ||
| Triglyceride | 0.130 | 0.04 | 0.064 | 0.34 |
| Log (high-sensitivity C-reactive protein) | -0.017 | 0.80 | ||
| Left ventricular end-diastolic dimension | 0.054 | 0.40 | ||
| Left ventricular ejection fraction | -0.028 | 0.66 | ||
We used univariate and multivariate logistic regression analyses to identify the parameters associated with the presence of HF in the study patients. Multivariate analysis identified log BDNF level (odds ratio 0.82; 95% confidence interval [CI] 0.76 to 0.91, p <0.001; Hosmer-Lemeshow chi-square = 0.99 and p = 0.87) as an independent and significant correlate with the presence of HF ( Table 3 ).
| Variables | Univariate Analysis | Multivariate Analysis | ||
|---|---|---|---|---|
| OR (95% CI) | p value | OR (95% CI) | p value | |
| Age (per year) | 1.01 (0.99-1.04) | 0.25 | Not selected | |
| Males (yes) | 1.28 (0.76-2.15) | 0.36 | Not selected | |
| Body mass index (per kg/m 2 ) | 0.91 (0.85-0.97) | 0.005 | Not selected | |
| Hypertension (yes) | 0.63 (0.35-1.11) | 0.11 | Not selected | |
| Diabetes mellitus (yes) | 0.94 (0.56-1.60) | 0.83 | Not selected | |
| Log (brain-derived neurotrophic factor) (per 0.1 pg/ml) | 0.83 (0.79-0.88) | <0.001 | 0.82 (0.76-0.91) | <0.001 |
| Hemoglobin (per g/dl) | 0.76 (0.66-0.88) | 0.005 | 0.59 (0.43-0.80) | 0.001 |
| Serum sodium (per mEq/L) | 0.81 (0.73-0.89) | <0.001 | Not selected | |
| Estimated glomerular filtration rate (per ml/min/1.73 m 2 ) | 0.95 (0.93-0.96) | <0.001 | 0.96 (0.94-0.99) | 0.004 |
| High density lipoprotein-cholesterol (mg/dl) | 1.00 (0.99-1.01) | 0.89 | Not selected | |
| Low density lipoprotein-cholesterol (mg/dl) | 1.01 (1.00-1.01) | 0.27 | Not selected | |
| Triglyceride (mg/dl) | 1.00 (0.99-1.00) | 0.07 | Not selected | |
| High-sensitivity C-reactive protein (per mg/dl) | 1.48 (0.92-2.38) | 0.11 | Not selected | |
| Left ventricular ejection fraction (%) | 0.84 (0.80-0.89) | <0.001 | 0.80 (0.75-0.85) | <0.001 |
| Loop diuretics (yes) | 50.00 (15.2-166.67) | <0.001 | 20.00 (2.72-142.86) | 0.003 |
| β-blockers (yes) | 2.16 (1.26-3.70) | 0.005 | Not selected | |
| Angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers (yes) | 6.80 (3.82-12.20) | <0.001 | 5.13 (1.01-27.78) | 0.049 |
| Aldosterone antagonists (yes) | 22.22 (8.77-58.82) | <0.001 | Not selected | |
| Statins (yes) | 0.25 (0.13-0.50) | <0.001 | 0.04 (0.01-0.23) | <0.001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
