Association of Annual N-Terminal Pro-Brain Natriuretic Peptide Measurements With Clinical Events in Patients With Asymptomatic Nonsevere Aortic Stenosis: A Post Hoc Substudy of the SEAS Trial

Abstract

Importance: Recent studies have questioned the presumed low-risk status of patients with asymptomatic nonsevere aortic stenosis (AS). Whether annual N-terminal pro-brain natriuretic peptide (NT-proBNP) measurements are useful for risk assessment is unknown.

Objective: To assess the association of annual NT-proBNP measurements with clinical outcomes in patients with nonsevere AS.

Design, setting, and participants: Analysis of annual NT-proBNP concentrations in the multicenter, double-blind Simvastatin and Ezetimibe in Aortic Stenosis (SEAS) randomized clinical trial was performed. SEAS was conducted from January 6, 2003, to April 1, 2008. Blood samples were analyzed in 2016, and data analysis was performed from February 10 to October 10, 2021. SEAS included 1873 patients with asymptomatic AS not requiring statin therapy with transaortic maximal flow velocity from 2.5 to 4.0 m/s and preserved ejection fraction. This substudy included 1644 patients (87.8%) with available blood samples at baseline and year 1.

Exposures: Increased age- and sex-adjusted NT-proBNP concentrations at year 1 and a 1.5-fold or greater relative NT-proBNP concentration change from baseline to year 1. Moderate AS was defined as baseline maximal flow velocity greater than or equal to 3.0 m/s.

Main outcomes and measures: Aortic valve events (AVEs), which are a composite of aortic valve replacement, cardiovascular death, or incident heart failure due to AS progression, were noted. Landmark analyses from year 1 examined the association of NT-proBNP concentrations with outcomes.

Results: Among 1644 patients, 996 were men (60.6%); mean (SD) age was 67.5 (9.7) years. Adjusted NT-proBNP concentrations were within the reference range (normal) in 1228 of 1594 patients (77.0%) with NT-proBNP values available at baseline and in 1164 of 1644 patients (70.8%) at year 1. During the next 2 years of follow-up, the AVE rates per 100 patient-years for normal vs increased adjusted NT-proBNP levels at year 1 were 1.39 (95% CI, 0.86-2.23) vs 7.05 (95% CI, 4.60-10.81) for patients with mild AS (P < .01), and 10.38 (95% CI, 8.56-12.59) vs 26.20 (95% CI, 22.03-31.15) for those with moderate AS (P < .01). Corresponding all-cause mortality rates were 1.05 (95% CI, 0.61-1.81) vs 4.17 (95% CI, 2.42-7.19) for patients with mild AS (P < .01), and 1.60 (95% CI, 0.99-2.57) vs 4.78 (95% CI, 3.32-6.87) for those with moderate AS (P < .01). In multivariable Cox proportional hazards regression models, the combination of a 1-year increased adjusted NT-proBNP level and 1.5-fold or greater NT-proBNP level change from baseline was associated with the highest AVE rates in both patients with mild AS (hazard ratio, 8.12; 95% CI, 3.53-18.66; P < .001) and those with moderate AS (hazard ratio, 4.05; 95% CI, 2.84-5.77; P < .001).

Conclusions and relevance: The findings of this study suggest that normal NT-proBNP concentrations at 1-year follow-up are associated with low AVE and all-cause mortality rates in patients with asymptomatic nonsevere AS. Conversely, an increased 1-year NT-proBNP level combined with a 50% or greater increase from baseline may be associated with high AVE rates.

Trial registration: ClinicalTrials.gov Identifier: NCT00092677.

Conflict of interest statement

Conflict of Interest Disclosures: Dr Hadziselimovic reported receiving grants from the Research Foundation of Bispebjerg University Hospital and Gangstedfonden during the conduct of the study. Dr Olsen reported receiving diagnostic kits for NT-proBNP from Roche during the conduct of the study. Dr Kesäniemi reported receiving grants from Merck during the conduct of the study. Dr Ray reported receiving travel expenses from Merck to attend SEAS steering committee meetings during the conduct of the study. Dr Rossebø reported receiving consulting fees from Merck and Schering-Plough during the SEAS trial. Dr Willenheimer reported receiving consultant fees from Lund University Heart Health Group during the conduct of the study. Dr Nielsen reported receiving assays from Roche Diagnostics Denmark for NT-proBNP measurements during the conduct of the study and fees from Roche Diagnostics Denmark for oral presentation outside the submitted work. No other disclosures were reported.

Figures

Figure 1.. Cumulative Incidence of Aortic Valve Events During Follow-up

Patients with mild (A) and moderate (B) aortic stenosis (AS) stratified by normal vs increased adjusted levels of N-terminal pro-brain natriuretic peptide (NT-proBNP) at year 1 (n = 1644). Sensitivity analysis showing cumulative incidence of aortic valve events stratified by NT-proBNP levels and change in levels (above and below 1.5-fold) during 1 year for patients with mild (C) and moderate (D) aortic stenosis (n = 1594). NT-proBNP ratio = NT-proBNP / upper limit of normal, specific for age and sex. NT-proBNP change = (NT-proBNP at year 1) / (NT-proBNP at year 0).

Figure 2.. Restricted Cubic Splines With Hazard Ratios for Aortic Valve Events

HR indicates hazard ratio; NT-proBNP, N-terminal pro-brain natriuretic peptide.

Figure 3.. Association of N-Terminal Pro-Brain Natriuretic Peptide (NT-proBNP) Levels at Year 1 Combined With NT-proBNP Level Change During 1 Year With Outcomes

Multivariable Cox proportional hazards regression model for aortic valve events (composite of aortic valve replacement, cardiovascular death and/or hospitalization with heart failure due to aortic stenosis) and all-cause mortality as the first event during 2 years of follow-up (year 1 to year 3) for combinations of normal and increased NT-proBNP levels at year 1 and NT-proBNP level change from baseline to year 1 above or below 1.5. In addition to shown variables, the model was adjusted for age, sex, systolic blood pressure, body mass index, diuretic treatment, and creatinine level at baseline (year 0), transaortic maximal velocity at baseline and year 1, and annual change of transaortic maximal velocity between baseline and year 1. NE indicates nonestimable.