Diagnostic concordance between NT-proBNP and BNP for suspected heart failure

Abstract

ObjectivesBNP and NT-proBNP are viewed as comparable in their ability to diagnose and monitor HF in clinical guidelines. However, no recent large-scale study has directly established diagnostic concordance between BNP and NT-proBNP. This study sought to assess diagnostic concordance of BNP and NT-proBNP for ruling in and ruling out heart failure (HF). MethodsSimultaneous BNP and NT-proBNP testing was performed on 2729 patient samples with routinely ordered BNP testing. Hospital location, age, sex, creatinine, BNP and NT-proBNP were also recorded. Recommended cutoffs for BNP and NT-proBNP for ruling in and out HF were used for assessing diagnostic concordance and correlation. ResultsIn the ED setting, concordance between BNP and NT-proBNP was 0.695 (95% CI, 0.668–0.723) by weighted kappa using the recommended cutoffs for the acute setting. In non-ED patients, the concordance was 0.642 (95% CI, 0.580–0.705) using non-acute setting cutoffs. In the ED setting, patients with eGFR <60 mL/min/1.73m2 had lower overall concordance (0.626; 95% CI 0.580–0.672) compared to those with eGFR >60 mL/min/1.73m2 (0.707, 95% CI 0.669–0.744). Patients with an eGFR <15 mL/min/1.73m2 had a much higher ratio of NT-proBNP to BNP than patients with eGFR >60 mL/min/1.73m2 (17.0 vs. 4.7, P < .001). Linear regression revealed an r2 of 0.52 in the ED setting and 0.49 in the non-ED setting between BNP and NT-proBNP. For 368 patients with multiple measurements of natriuretic peptides, 19.7% of paired temporal measurements had an increase in one peptide and a decrease in the other. ConclusionThe current cutoffs for diagnosing HF for NT-proBNP and BNP have relatively low diagnostic concordance and correlation, particularly among patients with chronic kidney disease.