Identification of ventricular tachycardia of epicardial origin from unipolar potentials obtained at the endocardial surface: is it feasible?

VT late after myocardial infarction usually originates from the endocardial surface; subepicardial substrates are also possible. The identification of these atypical locations with endocardial mapping remains unresolved even with new mapping technologies. This study compared isopotential maps, signal morphology, and activation patterns around left endocardial breakthroughs recorded in VTs originating from the subepicardium or subendocardium after remote myocardial infarction. These results were extracted from a database of 111 tachycardias obtained at surgery in 34 patients. Mapping was performed with a 128-unipolar electrode system using an epicardial mesh and a left ventricular endocardial balloon. Subepicardial (n = 7) and subendocardial VTs (n = 10) were defined as complete superficial reentry and/or as tachycardias with a > or = 25-ms delay between the earliest activity and the breakthrough of activation on the opposite surface. A positive potential distribution covering the area of initial endocardial activity was observed in a single subepicardial VT but in none of the subendocardial ones (P = NS). R waves were observed on the earliest endocardial unipolar signals in two subepicardial VTs and five subendocardial VTs (P = NS). The area covered by the first 5-ms or 10-ms isochrone at the endocardial level was larger in subepicardial VTs than in subendocardial VTs but the difference was not significant. In conclusion, despite a wider endocardial area of early activity in VTs of subepicardial origin, no reliable criteria can be proposed to identify these tachycardias from mapping data restricted to the endocardial surface. This is probably due to highly nonuniform anisotropic propagation around the scarred tissue.