Score Indices for Predicting Electrophysiologic Outcomes in Patients with Unexplained Syncope

Abstract

Our ability to predict a positive electrophysiologic (EP) study in the evaluation of unexplained syncope is suboptimal. In patients with unexplained syncope, we defined clinical predictors of bradyarrhythmia and ventricular tachycardia (VT) diagnosed during EP study, constructed diagnostic score indices for bradyarrhythmia and VT, and evaluated the predictive power of each score index. All patients evaluated in the Arrhythmia Clinic for unexplained syncope from January 1, 1996, through December 31, 1998, were identified and enrolled in the study. Five hundred eight patients (325 [64%] men; mean+/-SD age, 64+/-17 years) underwent EP testing. We analyzed elements from historical data and noninvasive laboratory findings as predictors of bradyarrhythmia and VT diagnosed on EP study. Fifty-eight patients (11%) had sinus node dysfunction, 94 (19%) had atrioventricular (AV) node disease, 92 (18%) had His-Purkinje system disease, and 101 (20%) had VT. Models were fit using logistic regression analysis. Predictors were assigned weighted scores, and a score index was formulated. The area under the curve associated with sinus node dysfunction, AV node disease, His-Purkinje system disease, and VT models was 0.64, 0.60, 0.84, and 0.60, respectively. We have constructed diagnostic score indices for EP outcomes of bradyarrhythmia and VT in syncope. Of all the score indices, the model for His-Purkinje system disease has the highest predictive power.