Machine learning to optimize the diagnostic performance of natriuretic peptides for acute heart failure across age groups

Perez Vicencio D, Doudesis D, Thurston AJ, Chenevier-Gobeaux C, Claessens YE, Lopez Ayala P, Belkin M, Wussler D, deFilippi C, Seliger S, Moe G, Fernando C, Bayes-Genis A, Pinto Y, Gaggin HK, Wiemer JC, Möckel M, Rutten JH, Gargani L, Pugliese NR, Pemberton C, Ibrahim I, Gegenhuber A, Mueller T, Neumaier M, Behnes M, Akin I, Bombelli M, Grassi G, Nazerian P, Albano G, Bahrmann P, Richards AM, McMurray JJ, Mueller C, Januzzi JL, Mills NL, Lee KK (2026)

Publication Type: Journal article

Publication year: 2026

Journal

Book Volume: 13

Journal Issue: 1

DOI: 10.1093/eschf/xvaf006

Abstract

BACKGROUND AND AIMS: N-terminal pro-B-type natriuretic peptide (NT-proBNP) concentrations are influenced by age, which may influence the diagnostic performance of this peptide. Machine learning approaches incorporating NT-proBNP and age as continuous measures may have improved diagnostic performance. METHODS: We pooled individual patient-level data for 10 369 patients [median age 73 years (25th-75th percentile: 59-82)] with suspected acute heart failure across fourteen studies. The diagnostic performance of guideline-recommended NT-proBNP thresholds (uniform rule-out threshold of 300 pg/mL and age-stratified rule-in thresholds of 450, 900, and 1800 pg/mL for patients <50, 50-75, and >75 years, respectively) and the Collaboration for the Diagnosis and Evaluation of Heart Failure (CoDE-HF) machine learning model were evaluated using random effects meta-analysis across age groups. RESULTS: Overall, 43.9% (4549/10 369) of patients had an adjudicated diagnosis of acute heart failure. The negative predictive value (NPV) of the rule-out threshold of 300 pg/mL was lower in older patients [NPV 88.7% (confidence interval (CI) 84.2-92.1%) in patients ≥80 years vs 98.9% (97.6-99.5%) <50 years]. Conversely, the positive predictive value (PPV) of age-stratified rule-in thresholds was lower in younger patients [PPV 62.0% (56.2-67.5%) in those <50 years vs 79.6% (70.7-86.3%) ≥80 years]. CoDE-HF was more accurate than guideline-recommended thresholds across all age groups, with NPV and PPV ranging from 96.4% to 99.5% (93.8-99.8% CIs) and 81.1% to 84.2% (74.7-90.4% CIs), respectively. CONCLUSION: The diagnostic performance of guideline-recommended thresholds of NT-proBNP varies significantly with age. A decision-support tool incorporating NT-proBNP with age as a continuous variable provides a more consistent and accurate approach.

Authors with CRIS profile

Involved external institutions

University of Edinburgh United Kingdom (GB) Careggi University Hospital / Azienda Ospedaliero Universitaria Careggi Italy (IT) University of Amsterdam Netherlands (NL) University of Otago New Zealand (NZ) Harvard University United States (USA) (US) Radboud University Nijmegen Medical Centre / Radboudumc of voluit Radboud Universitair Medisch Centrum (UMC) Netherlands (NL) University of Pisa / Università di Pisa (UniPi) Italy (IT) Assistance Publique-Hôpitaux de Paris (AP-HP) France (FR) Princess Grace Hospital Center (CHPG) / Centre Hospitalier Princesse Grace (CHPG) Monaco (MC) National University Hospital (NUH) Singapore (SG) Salzkammergut Klinikum Austria (AT) Universitätsspital Basel Switzerland (CH) Universitätsklinikum Mannheim / University Medical Centre Mannheim (Universitätsmedizin Mannheim) Germany (DE) University of Maryland School of Pharmacy United States (USA) (US) St. Michael's Hospital Canada (CA) Health Sciences Research Institute of the “Germans Trias i Pujol” Foundation (IGTP) Spain (ES) University of Milano-Bicocca / Università degli Studi di Milano-Bicocca, UNIMIB Italy (IT) Thermo Fisher Scientific Germany Limited & Co. KG Germany (DE) Charité - Universitätsmedizin Berlin Germany (DE) University of Glasgow United Kingdom (GB)

How to cite

APA:

Perez Vicencio, D., Doudesis, D., Thurston, A.J., Chenevier-Gobeaux, C., Claessens, Y.E., Lopez Ayala, P.,... Lee, K.K. (2026). Machine learning to optimize the diagnostic performance of natriuretic peptides for acute heart failure across age groups. ESC Heart Failure, 13(1). https://doi.org/10.1093/eschf/xvaf006

MLA:

Perez Vicencio, Daniel, et al. "Machine learning to optimize the diagnostic performance of natriuretic peptides for acute heart failure across age groups." ESC Heart Failure 13.1 (2026).

BibTeX: Download