Abstract 9485: Fragmented QRS Complexes, Conduction Disorder, and ST Segment Abnormality on Twelve-Lead Electrocardiography Predict Severe Cardiac Manifestation in Sarcoidosis

Abstract

Background: Cardiac Sarcoidosis is a potentially lethal disease, but many patients with cardiac sarcoidosis are asymptomatic and the diagnosis is difficult before severe cardiac manifestation. Therefore, we aimed to identify electrocardiographic predictors of cardiac sarcoidosis. Methods and Results: This prospective, observational cohort study included consecutive 227 patients with non-cardiac sarcoidosis who did not have cardiac manifestations including advanced atrioventricular block, ventricular tachycardia, and systolic dysfunction at baseline. We studied the association of baseline electrocardiographic findings with risk of developing cardiac manifestation. In patients with non-cardiac sarcoidosis, the abnormities were frequently found in the baseline electrocardiogram; conduction disorder in 26 patients (11%), ST segment abnormality in 23 patients (10%), and fragmented QRS complex defined as various RSR’ patterns without typical bundle branch block in 19 patients (8%). During a follow-up of 6.3±3.7 years, 11 patients developed severe cardiac involvement including advanced atrioventricular block (4 patients), ventricular tachycardia (4 patients), and systolic dysfunction (3 patients). In multivariate analyses, baseline heart rate and PR interval were associated with the increased risk of developing cardiac manifestation (Table). QRS duration or corrected QT interval were not associated with cardiac manifestation. Multivariate analyses revealed that conduction disorder, ST segment abnormality, and fragmented QRS complex were associated increased the risk of developing cardiac involvement. Conclusions: We identified the electrocardiographic predictors of severe cardiac manifestation in patients with non-cardiac sarcoidosis, suggesting that these electrocardiographic abnormalities may reflect the early stage of cardiac involvement.