B-type natriuretic peptide (BNP) levels are elevated in patients with aortic stenosis (AS) and decrease acutely after replacement of the stenotic valve. The long-term prognostic value of BNP after transcatheter aortic valve implantation (TAVI) and the relative prognostic utility of single versus serial peri-interventional measurements of BNP and N -terminal prohormone BNP (NT-pro-BNP) are unknown. This study sought to determine the impact of BNP levels on long-term outcomes after TAVI and to compare the utility of BNP versus NT-pro-BNP measured before and after intervention. We analyzed 340 patients with severe AS and baseline pre-TAVI assessment of BNP. In 219 patients, BNP and NT-pro-BNP were measured serially before and after intervention. Clinical outcomes over 2 years were recorded. Patients with high baseline BNP (higher tertile ≥591 pg/ml) had increased risk of all-cause mortality (adjusted hazard ratio 3.16, 95% confidence interval 1.84 to 5.42; p <0.001) and cardiovascular death at 2 years (adjusted hazard ratio 3.37, 95% confidence interval 1.78 to 6.39; p <0.001). Outcomes were most unfavorable in patients with persistently high BNP before and after intervention. Comparing the 2 biomarkers, NT-pro-BNP levels measured after TAVI showed the highest prognostic discrimination for 2-year mortality (area under the curve 0.75; p <0.01). Baseline-to-discharge reduction, but not baseline levels of BNP, was related to New York Heart Association functional improvement. In conclusion, high preintervention BNP independently predicts 2-year outcomes after TAVI, particularly when elevated levels persist after the intervention. BNP and NT-pro-BNP and their serial periprocedural changes provide complementary prognostic information for symptomatic improvement and survival.
Transcatheter aortic valve implantation (TAVI) has evolved into a reliable treatment method in high-risk or inoperable patients with severe, symptomatic aortic stenosis (AS). B-type natriuretic peptide (BNP) and N -terminal prohormone BNP (NT-pro-BNP) increase in response to pressure or volume overload of the myocardium and are established as powerful predictors of adverse outcomes across the spectrum of cardiovascular disease. The possible predictive value of natriuretic peptides (NPs) has been previously examined in patients treated with TAVI, but important unaddressed issues were raised. First, previous findings were discordant with some studies reporting no independent impact of NP elevation on outcomes after TAVI. Second, previous studies were limited by relatively small patient numbers and shorter duration of follow-up. Third, although NPs undergo substantial changes immediately after replacement of the stenotic valve, previous reports focused on the prognostic impact of preintervention measurements or measurements obtained later during patient follow-up. Although serial reductions of NPs after therapeutic interventions provide incremental prognostic information in patients with heart failure and acute coronary syndromes, the relative predictive utility of BNP measured before versus early after TAVI has not been defined. Against this background, the purpose of this study was to determine the long-term prognostic impact of BNP and to compare the predictive value of pre and postintervention levels and of early periprocedural changes of BNP versus NT-pro-BNP in patients with severe AS who underwent TAVI.
Methods
This is a retrospective analysis of prospectively collected data. All patients with severe native-valve AS (defined as indexed aortic valve area [AVA] ≤0.6 cm 2 /m 2 or mean gradient >40 mm Hg) who underwent TAVI at our institution from August 2007 to August 2012 were entered into a dedicated database. Of 500 consecutive patients treated with TAVI, 340 patients with measurement of BNP within 7 days before the intervention were included in the present analysis. The study conforms to the guiding principles of the Declaration of Helsinki and was approved by the local Ethics Committee. All patients provided written informed consent for prospective follow-up.
Patients underwent right and left heart catheterization for hemodynamic assessment before TAVI. Aortic valve gradients were measured by pullback technique from the left ventricle to the ascending aorta. Calculation of the AVA was derived from the Gorlin equation and was indexed by body surface area. Standardized transthoracic echocardiographic examination was performed before TAVI, and left ventricular ejection fraction (LVEF) was calculated using the biplane Simpson method. Severity of aortic valve regurgitation was evaluated using spectral and color Doppler images and semiquantitatively graded as mild, moderate, and severe according to guidelines. TAVI was performed using standard techniques. Vascular access was transfemoral using the Medtronic CoreValve Revalving System (Medtronic, Inc., Minneapolis, Minnesota) or the Edwards SAPIEN XT valve (Edwards Lifesciences, Irvine, California), transapical for Edwards SAPIEN XT, or trans-subclavian using the Medtronic CoreValve prosthesis.
BNP was measured in 340 patients before intervention on hospital admission (1.5 ± 1.3 days before TAVI) using a chemiluminescent microparticle immunoassay (ARCHITECT BNP assay; Abbott Laboratories Diagnostics Division, Abbott Park, Illinois). In addition, BNP levels were measured in 253 patients, and both BNP and NT-pro-BNP were measured in 219 patients, after the intervention (on the days of discharge, 6.3 ± 2.8 days after TAVI).
Patients were followed throughout 2 years, and no patient was lost to follow-up. Adverse cardiac and cerebrovascular events were assessed in hospital, and regular follow-up was performed at 1 month, 1 year, and 2 years by means of a clinical visit or a standardized telephone interview. All events were adjudicated by a clinical event committee. An academic clinical trials unit (CTU Bern, Bern University Hospital, Switzerland) was responsible for central data audits and maintenance of the dedicated database. Clinical end points were defined according to the Valve Academic Research Consortium (VARC)-2 criteria. Primary end points were all-cause and cardiovascular death at 2 years. Secondary end points included the VARC-2–defined clinical efficacy end point (composite of all-cause death, stroke, hospitalizations for valve-related symptoms or worsening congestive heart failure, New York Heart Association [NYHA] class III or IV, and valve-related dysfunction defined as mean aortic valve gradient >20 mm Hg, effective orifice area <0.9 to 1.1 cm 2 , Doppler velocity index <0.35 m/sec, and/or moderate or severe prosthetic valve regurgitation). Because repeat hospitalizations were consistently recorded during 1 year after intervention in this cohort, the VARC-2–defined clinical efficacy end point is reported at 1 year. In addition, NYHA class at follow-up and change of NYHA functional class status from baseline to 2 years were recorded.
Baseline BNP was categorized as low, defined as the lower 2 tertiles (<591 pg/ml), or high, defined by the higher tertile of all measurements (≥591 pg/ml). Our primary assessment was the impact of baseline BNP levels on outcomes. A secondary goal compared the prognostic performance of single (before TAVI) versus serial measurements of BNP and the performance of BNP versus NT-pro-BNP in the subset of patients with serial assessment of both peptides. On the basis of previous reports examining serial changes of NPs in patients with heart failure and acute coronary syndromes, we explored outcomes in relation to (1) any decrease compared to any increase or no change of BNP and (2) presence of persistently high BNP ≥591 pg/ml both before and after TAVI (high → high) versus high baseline BNP but subsequently low post-TAVI BNP (high → low) versus low baseline BNP.
Continuous variables are summarized as mean ± standard deviation and categorical variables as actual numbers and percentages. Event rates at 2 years are reported using time-to-first-event data and graphically presented using Kaplan–Meier curves with incidence rates calculated from life tables, censoring patients at death, at the last follow-up performed, or at withdrawal of consent (whichever occurred first). Event rates at 2 years were compared using Cox’s regressions. Hazard ratios (HRs) with 95% confidence intervals (CIs) and p values were derived from the Wald chi-square test. Adjusted HRs are based on inverse probability of treatment weighing in 100 multiple imputation data sets using chained equations, using baseline variables with a univariate and/or multivariate effect (p <0.2) on all-cause mortality at 2 years. These included age, gender, renal failure, LVEF, body mass index, chronic obstructive pulmonary disease, atrial fibrillation, previous myocardial infarction, NYHA class III or IV, logistic EuroScore, and Society of Thoracic Surgeons score. The prognostic discrimination of before and after TAVI levels of BNP and NT-pro-BNP was assessed by receiver operating characteristics (ROC) curves; the chi-square test was used to determine the equality of the ROC-estimated area under the curve (AUC). Two-sided p values <0.05 were considered statistically significant. All analyses were performed with Stata version 13.1 (StataCorp, College Station, Texas).
Results
Baseline characteristics are summarized in Table 1 . Patients with high versus low baseline BNP were more frequently men with NYHA class III or IV symptoms and had lower body mass index, lower LVEF, and increased Society of Thoracic Surgeons scores and EuroScores. Procedural characteristics of TAVI interventions did not differ between groups ( Table 2 ). Interventions were performed by means of the femoral route in 91% of patients.
Variable | All patients (n=340) | Baseline BNP | P value | |
|---|---|---|---|---|
| Low (n=227) | High (n=113) | |||
| Age (years) | 83.2 ± 4.8 | 82.9 ± 4.8 | 83.6 ± 4.7 | 0.10 |
| Female | 195 (57%) | 143 (63%) | 54 (48%) | 0.05 |
| Body mass index (kg/m²) | 26.7 ± 5.1 | 27.7 ± 4.9 | 24.5 ± 4.3 | <0.001 |
| Cardiovascular risk factors | ||||
| Diabetes mellitus | 91 (27%) | 62 (28%) | 29 (26%) | 0.48 |
| Hypercholesterolemia | 207 (61%) | 142 (62%) | 65 (58%) | 0.32 |
| Arterial hypertension | 278 (82%) | 188 (83%) | 90 (80%) | 0.71 |
| Medical history | ||||
| Previous myocardial infarction | 53 (16%) | 28 (12%) | 25 (22%) | 0.004 |
| Previous PCI | 81 (24%) | 51 (22%) | 30 (27%) | 0.91 |
| Renal failure (GFR<60ml/min/1.73m²) | 227 (67%) | 136 (60%) | 91 (81%) | <0.001 |
| Chronic obstructive pulmonary disease | 56 (16%) | 44 (19%) | 12 (11%) | 0.023 |
| Atrial fibrillation | 94/310 (30%) | 54/205 (26%) | 40/105 (38%) | 0.037 |
| Functional class | ||||
| NYHA III/IV | 224 (66%) | 138 (61%) | 86 (77%) | <0.001 |
| Risk assessment | ||||
| Logistic EuroScore | 23.0 ± 13.5 | 19.3 ± 11.2 | 30.2 ± 14.8 | <0.001 |
| STS Score | 6.3 ± 4.1 | 5.8 ± 4.0 | 7.4 ± 4.1 | <0.001 |
| Hemodynamic variables | ||||
| Aortic valve area (cm 2 ) | 0.60 ± 0.22 | 0.61 ± 0.21 | 0.57 ± 0.22 | 0.13 |
| Indexed aortic valve area (cm 2 /m 2 ) | 0.33 ± 0.12 | 0.34 ± 0.12 | 0.31 ± 0.12 | 0.13 |
| Mean aortic valve gradient (mmHg) | 44.2 ± 17.4 | 44.9 ±16.6 | 42.7 ± 18.9 | 0.28 |
| Peak aortic valve gradient (mmHg) | 69.3 ± 25.7 | 70.7 ± 24.3 | 66.9 ± 27.9 | 0.32 |
| Echocardiographic variables | ||||
| LV end-diastolic diameter (mm) | 48.9 ± 9.9 | 45.4 ± 7.9 | 50.9 ± 10.4 | 0.007 |
| LV end-systolic diameter (mm) | 34.6 ± 12.4 | 28.2 ± 7.5 | 38.0 ± 13.2 | <0.001 |
| LV ejection fraction (%) | 53.0 ± 15.3 | 60.2 ± 10.3 | 45.9 ± 16.2 | <0.001 |
| Natriuretic peptides | ||||
| BNP, pg/ml | 605.7 ± 787.6 | 214.9 ± 145.6 | 1390.7 ± 950.5 | <0.001 |
| NT-pro-BNP, pg/ml | 5453.2 ± 8059.6 | 1833.3 ± 1830 | 12908.3 ± 10481 | <0.001 |
| Variable | Baseline BNP | P value | |
|---|---|---|---|
| Low (n=227) | High (n=113) | ||
| Access route | |||
| Femoral | 210 (93%) | 99 (88%) | 0.16 |
| Apical | 16 (7%) | 11 (10%) | 0.40 |
| Subclavian | 1 (0%) | 3 (3%) | 0.11 |
| Valve type | |||
| Medtronic CoreValve | 135 (59%) | 78 (69%) | 0.096 |
| Edwards Sapien XT Valve | 92 (41%) | 35 (31%) | 0.096 |
| Revascularization | |||
| Concomitant PCI | 40 (18%) | 24 (21%) | 0.46 |
| Procedural complications | |||
| Access vessel complication | 21 (9%) | 9 (8%) | 0.84 |
| Moderate/severe (grade ≥2) post-procedure AR | 27 (12%) | 17 (15%) | 0.80 |
| Severe (grade 3) post-procedure AR | 1 (0%) | 0 (0%) | 1.00 |
Overall mortality was 15.3% at 1 year and 23.8% at 2 years. After multivariate adjustments, high baseline BNP was associated with a higher risk of all-cause death (p <0.001) and cardiovascular death at 2 years (p <0.001; Table 3 , Figure 1 ). The VARC-2 clinical efficacy end point at 1 year occurred more frequently in patients with high compared to those with low baseline BNP (Odds ratio [OR] 2.41, 95% CI 1.50 to 3.88; p <0.001). There was no difference with regard to hospitalization for valve-related symptoms or worsening heart failure (OR 0.83, 95% CI 0.28 to 2.41; p = 0.73) or NYHA class III or IV (OR 0.94, 95% CI 0.38 to 2.33; p = 0.89) in relation to baseline BNP.
| Outcome | Baseline BNP | Adjusted Hazard Ratio (95% CI) | P value | |
|---|---|---|---|---|
| Low (n=227) | High (n=113) | |||
| Death | 40 (18.9%) | 41 (37.4%) | 3.16 (1.84-5.42) | <0.001 |
| Cardiovascular death | 22 (10.7%) | 33 (31.2%) | 3.37 (1.78-6.39) | <0.001 |
| Myocardial infarction | 5 (2.3%) | 3 (3.3%) | 0.61 (0.11-3.42) | 0.57 |
| Stroke | 5 (2.5%) | 11 (10.8%) | 8.44 (2.67-26.67) | <0.001 |
| Death, myocardial infarction or stroke | 46 (21.5%) | 46 (41.7%) | 2.61 (1.50-4.55) | 0.001 |
