A Novel N-Terminal Pro-B-Type Natriuretic Peptide Assay in the Early Diagnosis of Acute Heart Failure.

Background The performance of the novel NT-pro-B-type natriuretic peptide (NT-proBNP)-Access assay in the early diagnosis of acute heart failure (AHF) is unknown.Objectives The objective of the study was to assess the diagnostic accuracy of NT-proBNP-Access in patients presenting with acute dyspnea and compare it to the established NT-proBNP-Elecsys assay.Methods In a prospective multicenter diagnostic study enrolling patients presenting with acute dyspnea to the emergency department, NT-proBNP-Access was measured in a blinded fashion and compared to NT-proBNP-Elecsys concentrations. The primary endpoint was diagnostic accuracy quantified by area under the receiver operating characteristics curve (AUC). Secondary endpoints were the performance of the guideline-recommended clinical decision values (rule-out: <300 pg/mL, rule-in: age-adjusted >450/900/1,800 pg/mL) for AHF.Results Among 1,400 patients (53% AHF), the NT-proBNP-Access assay yielded significantly higher NT-proBNP concentrations vs the NT-proBNP-Elecsys assay (median 2,087 pg/mL vs 1,568 pg/mL [P < 0.001]). The NT-proBNP-Access assay had very high diagnostic accuracy (AUC: 0.914; 95% CI: 0.898-0.93), which was slightly lower than the NT-proBNP-Elecsys assay (AUC: 0.922; 95% CI: 0.908-0.937; P = 0.006). Using guideline-recommended clinical decision values, the NT-proBNP-Access assay ruled out fewer patients compared to NT-proBNP-Elecsys (18.7% vs 26.3%) with similar sensitivity (98.9% vs 98.5%). Conversely, more patients were ruled in (58.1% vs 52.1%), with lower specificity (77.6% vs 84.8%; P < 0.001). Diagnostic concordance was high (85.3%), with major mismatch (no AHF vs AHF) uncommon (0.6%), but minor mismatch (gray zone vs rule in/rule out and vice versa) common (14.1%).Conclusions The NT-proBNP-Access assay had a very high diagnostic accuracy for AHF. Levels were approximately 25% higher with NT-proBNP-Access vs NT-proBNP-Elecsys, resulting in minor diagnostic discordance in 1 of 7 patients using guideline-recommended decision values.