Anatomical Determinants of the Left Atrial Electrograms: Wall Thickness and Myocardial Fat Infiltration

Authors:
Open Access
Publisher website
Google Scholar
Abstract
Introduction: Electrogram peak‐to‐peak voltage depends on atrial wall structure and mapping configuration. We evaluated the relationship between computed tomography–derived atrial wall thickness, intramyocardial fat, myocardial muscle thickness, and electrogram voltage obtained with omnipolar, bipolar, and unipolar mapping. Methods: Consecutive patients undergoing left atrial mapping during atrial fibrillation or atypical atrial flutter ablation were retrospectively included. Pre‐procedural contrast‐enhanced CT was processed with ADAS 3D to quantify left atrial wall thickness and intramyocardial fat infiltration, from which myocardial muscle thickness was derived. High‐density electroanatomical maps were spatially co‐registered with CT. Associations between anatomical parameters and peak‐to‐peak voltage were assessed, stratifying by rhythm stability. Results: Twenty patients were included. In stable rhythms, left atrial wall thickness correlated positively with voltage for all configurations, strongest for unipolar ( β = 0.38) and omnipolar ( β = 0.23) vs. bipolar mapping ( β = 0.13) (BP–UP p < 0.0001; BP–OT p < 0.01). Intramyocardial fat showed the strongest inverse association with omnipolar ( β = –0.39), compared with bipolar ( β = –0.14) and unipolar ( β = –0.01; p ≤ 0.02). After excluding fat, muscle thickness correlated positively with OT ( β = 0.30) and UP ( β = 0.30), exceeding BP ( β = 0.14; p < 0.0001). In unstable rhythms, associations were attenuated but OT remained most responsive to fat ( β = –0.20) and muscle thickness ( β = 0.11) (both p < 0.01). Conclusions: Myocardial muscle thickness is the main determinant of local atrial voltage, while intramyocardial fat contributes to voltage attenuation. OT provides the most consistent assessment of atrial wall composition, particularly under rhythm instability.

This publication has 27 references indexed in Scilit: