Direct Comparison of Adjacent Endocardial and Epicardial Electrograms: Implications for Substrate Mapping

BACKGROUND Analysis of unipolar voltage maps has been used to detect epicardial scar, but data to define optimal parameters to identify scar remote from the recording site is limited. This study compares the characteristics of electrograms at endocardial sites adjacent to abnormal epicardial sites. METHODS AND RESULTS Data obtained from endocardial and epicardial electroanatomical maps of 31 patients with scar-related ventricular tachycardia were reviewed. Five hundred twenty-three pairs of endo- and epicardial points were selected according to predefined criteria. The endocardial points adjacent to epicardial scar (bipolar voltage < 1.5 mV) had smaller unipolar voltage than those distant from epicardial scar (P < 0.001). In multivariable analysis, unipolar voltage was the only endocardial electrogram predictor of epicardial scar (P < 0.001, OR 0.94, 95% CI 0.93 to 0.97). An endocardial unipolar amplitude < 4.4 mV in the right ventricular (RV) (sensitivity 93%, specificity 76%) and < 5.1 mV in the left ventricular (LV) (sensitivity 91%, specificity 75%) was the optimal cutoff predicting epicardial scar. Applying these thresholds to electroanatomical maps, revealed a good match between endocardial unipolar abnormality and epicardial scar for 67% of LV and 75% of RV maps, respectively, but notably poor matches occurred in 8 (29%) maps (7 with nonischemic cardiomyopathy). Site-by-site correlations were better for ischemic than nonischemic cardiomyopathy. CONCLUSIONS This study supports the contention that unipolar electrograms are capable of indicating overlying epicardial scar during endocardial mapping, but illustrates limitations that appear to differ with nonischemic as compared to ischemic cardiomyopathy. The presence of epicardial arrhythmia substrate cannot be excluded by analysis of unipolar endocardial maps in some patients.