The optimal screening strategy to prevent sudden cardiac death (SCD) in athletes remains unknown. Pre-participation screening with electrocardiogram (ECG) remains controversial. The utility and accuracy of limb-lead (LL) ECG alone in identifying cardiac abnormalities associated with SCD has not been studied. This study was a comparative secondary data analysis, comparing the interpretation accuracy of 4 physicians evaluating publicly available ECGs of the most common cardiac conditions associated with SCD in athletes. Each physician interpreted a total of 100 ECGs: 50 normal ECGs (25 LL and 25 standard 12L) and 50 abnormal ECGs (25 LL and 25 standard 12L). The agreement between LL ECGs and 12L ECGs was assessed by Cohen's kappa coefficient and the accuracy of identifying an abnormal ECG was compared across LL and 12L ECGs using a chi-squared test. Inter-rater reliability was assessed by estimating the Fleiss's kappa coefficient. The sensitivity of LL ECG and 12L ECG was identical at 86%. The specificity of LL ECG was 75% (95% CI = 65% to 83%) and 12L ECG was 82% (95% CI = 73% to 89%). Substantial agreement was seen between LL ECG and 12L ECG interpretation across all readers (k = 0.63; 95% CI = 0.49 to 0.77). Interpretation accuracy was 81% (95% CI = 74% to 86%) and 84% (95% CI 78% to 89%) using LL ECG and 12L ECG, respectively (p = 0.43). In conclusion, the accuracy, sensitivity, and specificity were high and comparable for both LL ECG and 12L ECG in identifying cardiovascular conditions associated with SCD. Agreement between LL ECG and 12L ECG was substantial.
Copyright © 2021. Published by Elsevier Inc.