Background and importance: Diagnosing acute heart failure (AHF) is difficult in elderly patients presenting with acute dyspnea to the emergency department.

Objectives: To assess the diagnostic accuracy of NT-proBNP, high-sensitivity cardiac troponin-I (Hs-cTnI), soluble ST2 (ST2), galectin-3 and CD146 alone and in combination for diagnosing AHF in elderly patients presenting with acute dyspnea to the emergency department.

Design, settings and participants: This was a prospective, multicenter study performed between September 2016 and January 2020, including elderly patients presenting with acute dyspnea to the emergency department of 6 French hospitals.

Intervention: Measurement of NT-proBNP, hs-cTnI, ST2, galectin-3 and CD146.

Outcome measure and analysis: The reference standard, AHF, was adjudicated by two independent physicians based on ED and hospitalization clinical, biological (excluding biomarkers), radiological and echocardiography data (performed by a cardiologist in the cardiology department specifically for this study). Three exploratory methods (two using a cross-sectional approach with logistic regression and counting all biomarker combinations, and one using a sequential approach with gray zone optimizations) were applied to create comprehensive combinations of the 5 biomarkers for measuring diagnostic accuracy.

Main results: Two hundred thirty-eight patients (median age of 85 years, IQR = 8) were analyzed, and 110 (46%) were diagnosed with AHF. The accuracies of NT-proBNP, CD146, hs-cTnI, galectin-3, and ST2 were 0.72 [95% confidence interval (CI) 0.66-0.77], 0.63 (95% CI 0.57-0.69), 0.59 (95% CI 0.53-0.65), 0.55 (95% CI 0.49-0.61) and 0.51 (95% CI 0.45-0.57), respectively. Regardless of the approach used or how the 5 biomarkers were combined, the best accuracy for diagnosing AHF (0.73, 95% CI 0.67-0.78) did not differ from that of NT-proBNP alone.

Conclusion: In this study, NT-proBNP alone exhibited the best diagnostic accuracy for diagnosing AHF in elderly patients presenting with acute dyspnea to the emergency departments. None of the other biomarkers alone or combined improved the accuracy compared to NT-proBNP, which is the only biomarker to use in this setting.