Left Ventricular Ejection Fraction Assessment in Older Adults: An Adjunct to Natriuretic Peptide Testing to Identify Risk of New-Onset Heart Failure and Cardiovascular Death?

Study Questions:

Does the assessment of left ventricular ejection fraction (LVEF) aid in heart failure (HF) and cardiovascular mortality risk assessment above that of N-terminal pro-B-type natriuretic peptide (NT-proBNP) levels in older adults?


Serum NT-proBNP levels and LVEF were assessed in 4,137 adults ≥65 years without HF symptoms enrolled into the Cardiovascular Health Study. LVEF was assessed by echocardiogram at baseline and 5 years after enrollment. Patients were categorized as having normal (LVEF ≥55%), borderline (≥45% to <55%), or subnormal (<45%) systolic function. Patients had an abnormal NT-proBNP if the level was ≥190 pg/ml (sample 70th percentile). The primary outcomes of interest were incident HF and cardiovascular mortality during the period of follow-up, assessed using Cox modeling (hazard ratio [95% confidence interval] provided). Areas (AUC) under the receiver operating characteristic curves were compared for LVEF and BNP risk stratification.


The mean ± standard deviation patient age was 72.7 ± 5.5 years, 59% were female, and 13% were African American. At baseline, LVEF was subnormal in 2.6% of patients, borderline in 5.1%, and normal in 92%. Median [25th, 75th] NT-proBNP levels were 111 [56, 219] and 29.5% had a level ≥190 pg/ml. Over a median 10.7 years of follow-up, 1,112 (27%) patients developed new HF symptoms, and 893 (22%) died of cardiovascular causes. The risk of HF was 2.95 [2.6-3.3] and 2.4 [2.0, 2.9] higher for those with NT-proBNP ≥190 pg/ml and LVEF <55%, respectively. Compared with patients who had a normal LVEF and NT-proBNP at baseline, the adjusted risk of new-onset HF was 1.3 [0.92, 1.7] for those (n = 135) with a low NT-proBNP and LVEF <55%, 2.1 [1.8, 2.4] for those (n = 1,037) with high NT-proBNP and LVEF ≥55%, and 2.7 [2.1, 3.4] for those (n = 162) with high NT-proBNP and LVEF <55%. Likewise, compared with patients who had a normal LVEF and NT-proBNP at baseline, the adjusted risk of death was 1.7 [1.2, 2.3] for those with a low NT-proBNP and LVEF <55%, 1.9 [1.6, 2.3] for those with high NT-proBNP and LVEF ≥55%, and 3.0 [2.3, 3.8] for those with high NT-proBNP and LVEF <55%. Overall, the AUCs for predicting HF development-based LVEF, NT-proBNP, and LVEF plus NT-proBNP were 0.679, 0.719, and 0.723, respectively (p = 0.14 for addition of LVEF to NT-proBNP). For mortality, corresponding AUCs were 0.726, 0.759, and 0.761, respectively (p = 0.25 for addition of LVEF to NT-proBNP). Among patients whose NT-proBNP increased ≥25% from baseline to ≥190 pg/ml on follow-up, an LVEF <55% was also associated with increased risk of cardiovascular death. However, an abnormal LVEF again failed to improve risk stratification for either endpoint.


The authors concluded that LVEF assessment added to NT-proBNP does not routinely improve HF prognostication in older adults.


Predicting HF development and mortality risk can be challenging. In this observational cohort study, an abnormal LVEF in asymptomatic older adults was associated with a worse outcome. However, echocardiography did not improve prognostication above that of patient demographics and baseline or follow-up NT-proBNP levels. The authors argue that LVEF likely adds little accuracy to NT-proBNP risk assessment because an incident decline in LVEF is not common (8% of older adults) and absolute LVEF percentage may not have a strong influence on HF symptom development or survival. However, foregoing echocardiography in asymptomatic patients with abnormal BNPs may prevent introduction of interventions with proven mortality benefit (evidenced-based medications or ICD therapy) in systolic HF that would not be necessarily indicated in patients with diastolic HF and normal LVEF.

Clinical Topics: Anticoagulation Management, Heart Failure and Cardiomyopathies, Noninvasive Imaging, Acute Heart Failure, Heart Failure and Cardiac Biomarkers, Echocardiography/Ultrasound

Keywords: Follow-Up Studies, Area Under Curve, Demography, Systole, Natriuretic Peptides, Biological Markers, Cardiology, Heart Failure, Peptide Fragments, Stroke Volume, Ventricular Function, Confidence Intervals, ROC Curve, Diastole, Risk Assessment, United States, Echocardiography

< Back to Listings