A-010 Clinical Performance of an N-Terminal Pro-B-Type Natriuretic Peptide Assay in Acute Heart Failure Diagnosis

Abstract Background We evaluated the clinical performance of Alere NT-proBNP for Alinity i with the prospective collection of specimens from patients in an emergency department (ED) with signs and symptoms consistent with a clinical suspicion of new onset or acute exacerbation of heart failure (HF). Methods Concentrations of NT-proBNP were measured in patient K2EDTA plasma samples from 17 emergency departments across the United States. A total of 2127 subjects were included in the study. For each subject, enrolled, two assigned board -certified cardiologists made a determination of HF or non-HF following the event definitions in the adjudication charter and based on their own expertise and best judgment consistent with normal standard of care. If both cardiologists did make the same determination of HF or no HF a third cardiologist made an additional determination. Patient baseline descriptive statistics of age, sex, race, and ethnicity were summarized by calculating the minimum, maximum, mean, and SD for continuous variables, or frequency and percentage for categorical variables, for all subjects combined, and by adjudicated HF and non-HF subgroups separately. Descriptive statistics of NT-proBNP concentration were calculated for all subjects, male, female, and age groups (< 50, 50-75, and > 75-year-old) separately. The sensitivity, specificity, negative predictive value (NPV), positive predictive value (PPV), positive likelihood ratio (LR+), and negative likelihood ratio (LR-) were evaluated using the age-dependent positive cutoffs (450, 900, and 1800 pg/mL), and a rule-out age-independent cutoff (300 pg/mL). The estimated area under the curve (AUC) for adjudicated HF status using Alere NT-proBNP concentration (pg/mL) was reported with 95% confidence interval (CI) using a logistic model Results Of the 2127 subjects included in the study, 880 (41.4%) were diagnosed with HF by clinical adjudication. Of the 880 HF subjects, 185 (21.0%) had new onset HF, 644 (73.2%) had worsening HF, and 51 (5.8%) were indeterminate for either new onset or worsening HF. The AUC with 95% CI for all subjects was 0.8898 (0.8760, 0.9036). Within and across age groups, the AUC with 95% CI for ROC curves was 0.8499 (0.8075, 0.8922) to 0.9271 (0.9030, 0.9513). The specificity using age-dependent positive cutoffs was 81.2% (95% CI: 79.0% to 83.3%) for all subjects, 84.0% (95% CI: 79.4% to 87.7%) for age group of 18 - < 50, 81.5% (95% CI: 78.6% to 84.0%) for age group of 50 – 75, and 74.3% (95% CI: 66.7% to 80.7%) for age group of > 75. The sensitivity was 94.2% (95% CI: 92.5% to 95.6%) and the negative predictive value (NPV) was 94.0% (95% CI: 92.2% to 95.4%) for all subjects using age-independent rule-out cutoff. The specificity in the >75-year age group was lower than observed for other age groups, likely due to a higher percentage of subjects with comorbidities, history of HF and/or eGFR < 60 mL/min/1.73m2 in this group. Conclusion The Alere NT-proBNP assay demonstrated acceptable clinical sensitivity and clinical specificity for subjects presenting to the ED with signs and symptoms of HF. The assay retained utility in patient subgroups with conditions that influence NT-proBNP concentration, and for those with gray zone results.