Diagnostic performance of a clinical ultrasound-based algorithm for acute heart failure in patients presenting to the emergency department with dyspnea.

Abstract

To study the diagnostic performance of an ultrasound-based algorithm that includes the deceleration time (DT) of early mitral filling to establish a diagnosis of acute heart failure (AHF) in patients who come to an emergency department because of dyspnea. Prospective analysis in a convenience sample of patients who came to a hospital emergency department with acute dyspnea. The algorithm included ultrasound findings and 4 echocardiographic findings as follows: mitral annular plane systolic excursion, Doppler mitral flow velocity, tissue Doppler imaging measure of the lateral annulus, and the DT of early mitral filling. The definitive diagnosis was made by 2 physicians blinded to each other's diagnosis and the ultrasound findings. A total of 166 adult patients with a mean (SD) age of 76 (13) years were included; 79 (48%) were women. AHF was the definitive diagnosis in 62 patients (37%). Diagnostic agreement was good between the 2 physicians (κ = 0.71). The algorithm classified all the patients, and there were no undetermined diagnoses. Diagnostic performance indicators for the ultrasound-based algorithm integrating early DT findings were as follows: area under the receiver operating characteristic curve, 0.91 (95% CI, 0.86-0.96); sensitivity, 87% (95% CI, 76%-94%); specificity, 95% (95% CI, 89%-98%); positive likelihood ratio, 18.1 (95% CI, 7.7-42.8); and negative likelihood ratio, 0.14 (95% CI, 0.07-0.26). The ultrasound-based algorithm integrating the DT of early mitral filling performs well for diagnosing AHF in emergency patients with dyspnea. The inclusion of early DT allows all patients to be diagnosed.