Abstract
Electrocardiogram (ECG)-based detection of left ventricular systolic dysfunction (LVSD) has poor specificity and positive predictive value, even when including major ECG abnormalities, such as left bundle branch block (LBBB) within the criteria for diagnosis. Although machine-read ECG algorithms do not provide information on LVSD, advanced ECG (A-ECG), using multiparameter scores, has superior diagnostic utility to strictly conventional ECG for identifying various cardiac pathologies, including LVSD. Methods: We evaluated the diagnostic utility of A-ECG in a case-control study of 40 patients with LVSD (LV ejection fraction < 50% by echocardiography), due to non-ischemic cardiomyopathy (NICM), and 39 other patients without LVSD. Diagnostic sensitivity and specificity for LVSD were determined after applying a previously validated probabilistic A-ECG score for LVSD to stored standard (10 s) clinical 12L ECGs. In 25 of the NICM patients who had serial ECGs and echocardiograms, changes in the A-ECG score versus in echocardiographic LV ejection fraction were also studied to determine the level of agreement between the two tests. Results: Analyses by A-ECG had a sensitivity of 95% for LVSD (93% if excluding N = 11 patients with LBBB) and specificity of 95%. In the 29 NICM patients without LBBB who had serial ECGs, sensitivity improved to 97% when all ECGs were considered. By comparison, human readers in a busy clinical environment had a sensitivity of 90% and specificity of 63%. A-ECG score trajectories demonstrated improvement, deterioration or no change in LVSD, which agreed with echocardiography, in 76% of cases (n = 25). Conclusion: A-ECG scoring detects LVSD due to NICM with high sensitivity and specificity. Serial A-ECG score trajectories also represent a method for inexpensively demonstrating changes in LVSD. A-ECG scoring may be of particular value in areas where echocardiography is unavailable, or as a gatekeeper for echocardiography.
Highlights
Since the advent of the electrocardiogram (ECG) over a century ago, its utility as a cost-effective and accessible diagnostic tool has been proven repeatedly
For these reasons the ECG has become a standard of care in the investigation of patients with suspected left ventricular systolic dysfunction (LVSD), and is a valuable tool in primary healthcare [1]
We have demonstrated that advanced ECG (A-ECG) scoring, a novel method of analysis of standard 12L ECG, has high sensitivity and specificity for detecting mechanical LVSD in a cohort with non-ischemic cardiomyopathy (NICM), exceeding that described in previous studies
Summary
Since the advent of the electrocardiogram (ECG) over a century ago, its utility as a cost-effective and accessible diagnostic tool has been proven repeatedly. Through the application of rules and recognition of certain patterns, expert readers of ECGs can infer a wide variety of cardiac diagnoses. For these reasons the ECG has become a standard of care in the investigation of patients with suspected left ventricular systolic dysfunction (LVSD), and is a valuable tool in primary healthcare [1]. Due to the abundance of patients presenting with dyspnoea, but relative inaccessibility of echocardiography and delays in laboratory testing, the ECG represents one of the most cost-effective point-of-care diagnostic instruments available in the community. Modern advances in digital recording, signal processing, and machine learning have significantly improved accuracy of diagnostic ECG algorithms
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
