Back

A Role For Electrocardiograms in Aortic Stenosis Screening: A Retrospective Trial

Pulaski, M.; Newadkar, A.; Richie, I.; Urnes, C.; Joseph, J.; Ho, I.; Matthews, R.; Zaman, J. A. B.

2024-04-14 cardiovascular medicine
10.1101/2024.04.10.24305584 medRxiv
Show abstract

BackgroundAortic stenosis affects 1 in 50 adults over age 65 and is associated with significant morbidity and mortality. Machine learning has identified an association between right-sided precordial U waves and moderate to severe aortic stenosis. No study has explored the role of ECG screening by primary care physicians for patients with unknown aortic stenosis status. MethodsA retrospective single center cohort analysis performed by non-cardiologists identified right-sided precordial U waves on ECGs in fifty adults ages 65 to 89. Following identification, reviewers were unblinded to echocardiograms to determine whether there was an association between right-sided precordial U waves and aortic stenosis severity. Fifty age- and gender-matched patients without right-sided precordial U waves comprised the control group. ResultsChi-squared analysis revealed a significant association between right-sided precordial U waves and severity of aortic stenosis ({chi}2 = 16.77, df = 3, p < 0.001). Multinomial logistic regressions demonstrated no relationship between categorical SBP (< 125, 126 - 145, > 145) and aortic stenosis (p = 0.35), but increasing categorical age (65-73, 74-81, 82-89) was associated with moderate to severe aortic stenosis (p = 0.002). ConclusionsOur data highlights right-sided precordial U wave identification by non-cardiologists as a novel objective adjunct to the physical examination in detection of a medical condition which portends significant morbidity and mortality if left untreated. These findings motivate a prospective randomized clinical trial in the utility of right-sided precordial U waves in screening for aortic stenosis in primary care settings.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.