Brain natriuretic peptide (BNP) is a marker of systolic and diastolic dysfunction and a strong predictor of mortality in heart failure patients. The present study aimed to assess the relationship of BNP with aortic stenosis (AS) severity and prognosis. The cohort comprised 289 high-risk patients with severe AS who were referred for transcatheter aortic valve implantation. Patients were divided into tertiles based on BNP level: I (n = 96); II (n = 95), and III (n = 98). Group III patients were more symptomatic, had higher Society of Thoracic Surgeons and EuroSCORE scores, and had a greater prevalence of renal failure, atrial fibrillation, and previous myocardial infarction; lower ejection fraction and cardiac output; and higher pulmonary pressure and left ventricular end diastolic pressure. The degree of AS did not differ among the 3 groups. Stepwise forward multiple regression analysis identifies ejection fraction and pulmonary artery systolic pressure as independent correlates with plasma BNP. Mortality rates during a median follow-up of 319 days (range 110 to 655) were significantly lower in Group I compared with Groups II and III, p <0.001. After multivariable adjustment, the strongest correlates for mortality were renal failure (hazard ratio 1.44, p = 0.05) and medical/balloon aortic valvuloplasty (HR 2.2, p <0.001). Mean BNP decreased immediately after balloon aortic valvuloplasty from 1,595 ± 1,229 to 1,252 ± 1,076, p = 0.001 yet increased to 1,609 ± 1,264, p = 0.9 at 1 to 12 months. After surgical aortic valve replacement, there was a nonsignificant, immediate decrease in BNP level from 928 ± 1,221 to 896 ± 1,217, p = 0.77, continuing up to 12 months 533 ± 213, p = 0.08. After transcatheter aortic valve implantation, there was no significant decrease in BNP immediately after the procedure; however, at 1-year follow-up, the mean BNP level decreased significantly from 568 ± 582 to 301 ± 266 pg/dl, p = 0.03. In conclusion, a high BNP level in high-risk patients with severe AS is not an independent marker for higher mortality. BNP level does not appear to be significantly associated with the degree of AS severity but does reflect heart failure status.
The value of brain natriuretic peptide (BNP) in the clinical care of patients with a high surgical risk and symptomatic aortic stenosis (AS) is unclear. This study was undertaken primarily to investigate the clinical, echocardiographic, and hemodynamic correlates, and mortality rates associated with plasma BNP levels in such patients. Secondarily, we aimed to examine changes in plasma BNP levels over time after treatment with transcatheter aortic valve implantation (TAVI), surgical aortic valve replacement (SAVR), and balloon aortic valvuloplasty (BAV).
Methods
This prospective cohort study was approved by the Institutional Review Board of the MedStar Health Research Institute. From 2007 to 2010, we enrolled 289 patients with symptomatic, severe AS (aortic valve area <1 cm 2 or aortic valve index <0.5 cm 2 /m 2 ). All had been referred for TAVI consideration because of high operative risk, and included all such patients in whom BNP level was measured at baseline. For analysis, patients were grouped into tertiles based on their plasma BNP level.
Suitability for TAVI was carefully considered in all 289 patients by a multidisciplinary team consisting of an interventional cardiologist, a cardiac surgeon, and the patient’s primary cardiologist. TAVI was performed in 105 patients (36.3%) and surgical valve replacement in 25 (8.6%). The remaining 159 patients did not undergo either form of valve replacement for various clinical reasons. BAV was undertaken in 127 patients (43.9%), and 32 (11.0%) were managed medically.
TAVI was performed with the Edwards SAPIEN valve (Edward LifeSciences, Irvine, California); valve size was selected based on the annulus size measured by transesophageal echocardiography. Transfemoral arterial access was first considered, but in cases of peripheral vascular disease or small caliber iliac arteries, the transapical approach was used. BAV was performed according to standard techniques via the retrograde femoral approach. SAVR was performed under standard anesthesia and following standard surgical techniques, extracorporeal circulation, and myocardial protection methods. The Society of Thoracic Surgeons and logistic EuroSCOREs were calculated using Web-based systems ( http://209.220.160.181/STSWebRiskCalc261/de.aspx and http://www.euroscore.org , respectively). All patients had transthoracic echocardiography and Doppler studies. Left ventricular ejection fraction was calculated with biplane Simpson methods; transvalvular pressure gradient was determined by the Bernoulli formula, and aortic valve area was calculated by the continuity equation based on measurement of left ventricular outflow tract diameter as measured in the parasternal long-axis view. Outflow tract velocity was measured using pulse-wave Doppler, and maximum transvalvular velocity was measured using the continuous-wave Doppler signal.
Right heart catheterization was performed with a 7Fr Swan-Ganz catheter (Edwards Lifesciences, Irvine, California). We carefully measured right atrial pressures (amplitude of a and v waves and mean pressure), right ventricular systolic and diastolic pressures, pulmonary artery pressures (systolic, diastolic, and mean), and pulmonary capillary wedge pressures (a and v waves and the mean pressure). Cardiac output was determined by the thermodilution method. Left heart catheterization was carried out from the femoral artery. Aortic valve area was calculated with the Gorlin formula.
Venous blood samples were drawn from antecubital vein into chilled ethylenediaminetetraacetic acid tubes, placed immediately on ice and centrifuged within 20 minutes at −4°C. Serum BNP level was determined using a fluorescent immunoassay (Triage BNP Test, Biosite Diagnostics, San Diego, California).
All calculations were performed using SAS version 9.1 (SAS Institute Inc., Cary, North Carolina). Continuous variables are presented as mean ± SD, and categorical variables as proportions and percentages. Days of follow-up are presented as median and interquartile range. Differences between continuous variables were tested for significance by analysis of variance. Differences between paired variables were tested by paired Student’s t test. Categorical variables were compared using the chi-square test or Fisher’s exact test as indicated. Significance was set at p <0.05.
Associations among mortality and clinical, laboratory, echocardiographic, and hemodynamic variables were studied with Cox proportional hazard regression analysis. To determine independent correlates of mortality, we initially performed a univariable Cox proportional hazard regression analysis using all variables recorded in the clinical laboratory, echocardiographic, and hemodynamic categories. In each group, all univariable correlates of mortality with a p value ≤0.05 were then used in a stepwise multivariable Cox regression model.
Cumulative survival curves were constructed using the Kaplan-Meier method, and the significance was assessed using the log-rank statistic. The relationship and variability between BNP and clinical, echocardiographic, and hemodynamic variables were analyzed by Pearson correlation analysis.
Results
Baseline characteristics are displayed in Table 1 . Patients in Group III (highest BNP level) were more symptomatic, had higher Society of Thoracic Surgeons and EuroSCOREs, lower body mass indices, and a greater prevalence of renal failure, atrial fibrillation, and previous myocardial infarction.
| Baseline Characteristic | BNP Tertile | p Value | ||
|---|---|---|---|---|
| Lowest (n = 96) | Middle (n = 95) | Highest (n = 98) | ||
| Age (yrs ± SD) | 82.9 ± 6.4 | 83.5 ± 6.6 | 81.0 ± 11.3 | 0.09 |
| Male patients | 39 (40.6%) | 45 (47.3%) | 43 (43.9%) | 0.46 |
| Society of Thoracic Surgeons score | 10.6 ± 4.2 | 13.3 ± 6.1 | 14.1 ± 7 | <0.001 |
| Logistic EuroSCORE | 38.6 ± 19.6 | 46.3 ± 21.2 | 51.4 ± 23.1 | <0.001 |
| New York Heart Association class IV | 45 (46.8%) | 58 (61.0%) | 79 (80.6%) | <0.001 |
| Body mass index (kg/m 2 ) | 29.0 ± 7.9 | 26.6 ± 5.6 | 26.2 ± 6.8 | 0.02 |
| Diabetes mellitus | 30 (31.2%) | 30 (31.5%) | 41 (41.8%) | 0.20 |
| Systemic hypertension ∗ | 83 (86.4%) | 80 (84.2%) | 88 (89.7%) | 0.24 |
| Coronary artery disease | 57 (59.3%) | 56 (58.9%) | 59 (60.2%) | 0.84 |
| Chronic obstructive pulmonary disease | 13 (13.5%) | 24 (25.2%) | 18 (18.4%) | 0.11 |
| Renal failure (CrCl <60 ml/min) | 32 (33.3%) | 39 (41.0%) | 59 (60.2%) | <0.001 |
| Previous cerebral vascular accident | 23 (23.9%) | 19 (20.0%) | 15 (15.3%) | 0.29 |
| Atrial fibrillation | 28 (29.1%) | 44 (46.3%) | 31 (31.6%) | 0.01 |
| Peripheral vascular disease | 27 (28.1%) | 30 (31.5%) | 29 (29.6%) | 0.90 |
| Previous coronary artery bypass grafting | 31 (32.3%) | 29 (30.5%) | 28 (28.5%) | 0.94 |
| Previous myocardial infarction | 13 (13.5%) | 27 (28.4%) | 35 (36.8%) | 0.004 |
∗ History of systemic hypertension diagnosed and/or treated with medication or currently being treated with diet and/or medication by a physician.
Laboratory parameters, echocardiographic data, invasive hemodynamic data, and treatment modalities are presented in Table 2 . Importantly, the severity of the AS was similar in all BNP tertiles. Left ventricular ejection fraction and cardiac output were lower in Group III. Taken together with the greater elevation of pulmonary artery and mean right pressures, these findings reflect reduced cardiac performance in this group. As might be expected, the sicker patients in Group III were less often selected for TAVI/SAVR and more frequently defaulted to BAV ( Table 2 ).
| Variable | BNP Tertile | p Value | ||
|---|---|---|---|---|
| Lowest (n = 96) | Middle (n = 95) | Highest (n = 98) | ||
| Laboratory values | ||||
| BNP (pg/ml) | 273 ± 122 | 878 ± 272 | 2,761 ± 1,127 | <0.001 |
| Hemoglobin level (mg %) | 12.5 ± 7.9 | 11.36 ± 1.7 | 11.2 ± 1.7 | 0.10 |
| Creatinine (mg %) | 1.2 ± 1.2 | 1.3 ± 0.8 | 1.8 ± 1.6 | <0.001 |
| hs-CRP (mg/dl) | 9.0 ± 16.3 | 23.8 ± 36.0 | 35.0 ± 39.6 | 0.009 |
| Sodium (md/dl) | 138.2 ± 3.8 | 137.4 ± 4.4 | 136.3 ± 4.7 | 0.004 |
| Echocardiographic data | ||||
| Ejection fraction (%) | 55.4 ± 15.1 | 43.5 ± 17.2 | 35.5 ± 16.4 | <0.001 |
| Pulmonary artery systolic pressure (mm Hg) | 48.6 ± 20.6 | 51.4 ± 18.8 | 56.1 ± 15.9 | 0.03 |
| Aortic valve area (cm 2 ) | 0.68 ± 0.15 | 0.68 ± 0.16 | 0.71 ± 0.16 | 0.22 |
| Maximum velocity across aortic valve (m/s) | 4.2 ± 0.6 | 4.0 ± 0.8 | 3.8 ± 0.7 | <0.001 |
| Mean gradients across aortic valve (mm Hg) | 45.4 ± 16.6 | 43.6 ± 17.7 | 37.1 ± 15.1 | 0.002 |
| Peak gradients across valve (mm Hg) | 72.3 ± 24.3 | 68.6 ± 26.8 | 60.9 ± 24.8 | 0.001 |
| Hemodynamic parameters | ||||
| Right atrium pressure (mm Hg) | 9.0 ± 4.3 | 9.8 ± 5.0 | 13.3 ± 11.9 | 0.004 |
| Pulmonary artery systolic pressure (mm Hg) | 51.6 ± 21.5 | 53.8 ± 16.4 | 66.3 ± 40.1 | <0.001 |
| Pulmonary capillary wedge pressers (mm Hg) | 20.2 ± 9.3 | 21.5 ± 7.4 | 22.3 ± 8.1 | 0.35 |
| Left ventricular end diastolic pressure (mm Hg) | 18.2 ± 5.5 | 22.4 ± 8.1 | 21.1 ± 6.1 | 0.01 |
| Cardiac output (L/min) | 4.6 ± 1.3 | 4.2 ± 1.1 | 3.7 ± 1.2 | <0.001 |
| Aortic valve area (cm 2 ) | 0.64 ± 0.2 | 0.63 ± 0.2 | 0.62 ± 0.2 | 0.64 |
| Mean gradients across aortic valve (mm Hg) | 52.8 ± 22.8 | 48.1 ± 21.5 | 41.4 ± 20.3 | 0.003 |
| Treatment modality | ||||
| Medical | 5 (5.2%) | 9 (9.5%) | 18 (18.3%) | <0.001 |
| BAV | 26 (27.1%) | 37 (38.9%) | 64 (65.3%) | |
| SAVR | 11 (11.5%) | 11 (11.6%) | 3 (3.1%) | |
| TAVI | 54 (56.2%) | 39 (41.0%) | 12 (12.2%) | |
No association was found between BNP level and aortic valve area (r = −0.01, p = 0.81). A significant correlation was found, however, among BNP level and clinical, echocardiographic, and hemodynamic evidence of heart failure: (1) ejection fraction (r = −0.47, p <0.001), (2) pulmonary pressure (r = 0.33, p ≤0.001); (3) cardiac output (r = −0.26, p <0.001), and (4) New York Heart Association class IV (p <0.001). Stepwise forward multiple regression analysis identified ejection fraction and pulmonary artery systolic pressure as independent correlates of plasma BNP. Moreover, patients who died had a median BNP level of 1,100 pg/dl (range 497 to 2,200) compared with 588 pg/dl (range 320 to 1,480) in survivors (p <0.001).
During a median follow-up of 319 days (range 110 to 655), 142 patients died (49.1%); the cause was cardiovascular in 64 (45.1%), noncardiovascular in 47 (33.1%), and unknown in 31 (21.8%). Mortality in the first, second, and third tertiles was 34 (35.4%), 48 (50.5%), 60 (61.2%), respectively, p <0.001. Figure 1 shows Kaplan-Meier survival curves for all tertiles. Factors associated with mortality are presented in Table 3 . In the univariable Cox analysis, BNP level was associated with mortality (HR 1.47, p <0.001); however, after multivariable adjustment, the only independent correlates for mortality were renal failure (hazard ratio 1.44, p = 0.05) and exclusion from a valve replacement procedure (hazard ratio 2.2, p <0.001).
