Left Ventricular Ejection Fraction Assessment by Emergency Physician-Performed Bedside Echocardiography: A Prospective Comparative Evaluation of Multiple Modalities

Background: Although there is some support for visual estimation (VE) as an accurate method to estimate left ventricular ejection fraction (LVEF), it is also scrutinized for its subjectivity. Therefore, more objective assessments, such as fractional shortening (FS) or e-point septal separation (EPSS), may be useful in estimating LVEF among patients in the emergency department (ED). Objective: Our aim was to compare the real-world accuracy of VE, FS, and EPSS using a sample of point-of-care cardiac ultrasound transthoracic echocardiography (POC-TTE) images acquired by emergency physicians (EPs) with the gold standard of Simpson's method of discs, as measured by comprehensive cardiology-performed echocardiography. Methods: We conducted a single-site prospective observational study comparing VE, FS, and EPSS to assess LVEF. Adult patients in the ED receiving both POC-TTE and comprehensive cardiology TTE were included. EPs acquired POC-TTE images and videos that were then interpreted by 2 blinded EPs who were fellowship-trained in emergency ultrasound. EPs estimated LVEF using VE, FS, and EPSS. The primary outcome was accuracy. Results: Between April and May 2018, 125 patients were enrolled and 113 were included in the final analysis. EP1 and EP2 had a kappa of 0.94 (95% confidence interval [CI] 0.87-1.00) and 0.97 (95% CI 0.91-1.00), respectively, for VE compared with gold standard, a kappa of 0.40 (95% CI 0.23-0.57) and 0.38 (95% CI 0.18-0.57), respectively, for EPSS compared with gold standard, and a K of 0.70 (95% CI 0.54-0.85) and 0.66 (95% CI 0.50-0.81), respectively, for FS compared with gold standard. Sensitivity of severe dysfunction was moderate to high in VE (EP1 85% and EP2 93%), poor to moderate in FS (EP1 73% and EP2 50%), and poor in EPSS (EP1 11% and EP2 18%). Conclusions: Using a real-world sample of POC-TTE images, the quantitative measurements of EPSS and FS demonstrated poor accuracy in estimating LVEF, even among experienced sonographers. These methods should not be used to determine cardiac function in the ED. VE by experienced physicians demonstrated reliable accuracy for estimating LVEF compared with the gold standard of cardiology-performed TTE. (C) 2021 Elsevier Inc. All rights reserved.

Automated summary

This study compared the accuracy of visual estimation (VE), fractional shortening (FS), and e-point septal separation (EPSS) in estimating left ventricular ejection fraction (LVEF). It was found that VE was relatively accurate, while EPSS and FS demonstrated poor accuracy.