Epicardial Fibrosis Explains Increased Endo-Epicardial Dissociation and Epicardial Breakthroughs in Human Atrial Fibrillation

Abstract

Background: Atrial fibrillation (AF) is accompanied by progressive epicardial fibrosis, dissociation of electrical activity between the epicardial layer and the endocardial bundle network, and transmural conduction (breakthroughs). However, causal relationships between these phenomena have not been demonstrated yet. Our goal was to test the hypothesis that epicardial fibrosis suffices to increase endo-epicardial dissociation (EED) and breakthroughs (BT) during AF.

Methods: We simulated the effect of fibrosis in the epicardial layer on EED and BT in a detailed, high-resolution, three-dimensional model of the human atria with realistic electrophysiology. The model results were compared with simultaneous endo-epicardial mapping in human atria. The model geometry, specifically built for this study, was based on MR images and histo-anatomical studies. Clinical data were obtained in four patients with longstanding persistent AF (persAF) and three patients without a history of AF.

Results: The AF cycle length (AFCL), conduction velocity (CV), and EED were comparable in the mapping studies and the simulations. EED increased from 24.1 ± 3.4 to 56.58 ± 6.2% (p < 0.05), and number of BTs per cycle from 0.89 ± 0.55 to 6.74 ± 2.11% (p < 0.05), in different degrees of fibrosis in the epicardial layer. In both mapping data and simulations, EED correlated with prevalence of BTs. Fibrosis also increased the number of fibrillation waves per cycle in the model.

Conclusion: A realistic 3D computer model of AF in which epicardial fibrosis was increased, in the absence of other pathological changes, showed increases in EED and epicardial BT comparable to those in longstanding persAF. Thus, epicardial fibrosis can explain both phenomena.

Keywords: EED; atrial fibrillation; breakthrough waves; computer models; fibrosis; transmural conduction.

FIGURE 1

Visualization of anatomical model of the atria, used for simulations, and three layers of fiber orientations. (A) Anterior view of the model and epicardial layer of fiber orientations, indicated by blue lines. (B) Posterior view of the model, with epicardial layer of fiber orientations. (C,D) Endocardial layer of fiber orientations. (E,F) Fiber orientations in endocardial and epicardial bundles, including BB (orange). (G) Trabecular network of the LAA. (H) Trabecular network of the RA, with 20 pectinate muscles (interior wide-angle view).

FIGURE 2

Posterior and anterior view of the atria with patchy and uniform fibrosis (white shows fibrotic tissue and dark red shows normal tissue). (A) Moderate fibrotic model with patchy fibrosis pattern (posterior view). (B) Moderate fibrotic model with patchy fibrosis pattern (anterior view). (C) Moderate fibrotic model with uniform fibrosis pattern (posterior view). (D) Moderate fibrotic model with uniform fibrosis pattern (anterior view). (E) A cross-section view of the atria with uniform fibrosis.

FIGURE 3

Example of a spiral wave initiation. (A) Temporary line of block (green line). (B) Activation during the first 150 ms, after pacing near the mitral ring, with the block line in place. (C) The next 140 ms, after the block line was removed. Activation first crosses the former block line in the area between the superior pulmonary veins. (D–F) The next two cycles of the reentry. The colorbar indicates the activation time, and the scales are in milliseconds.

FIGURE 4

The endo–epicardial electrode contains two identical plaques of 8 × 8 unipolar electrodes.

FIGURE 5

Electrophysiological parameters during AF. (A) Conduction velocity (CV) in patients. (B) CV in simulations with endomysial (uniform) fibrosis. (C) CV in simulations with patchy fibrosis. (D) Snapshot of activation in a control simulation. Brighter colors indicate higher transmembrane potentials. The epicardial layer is shown in a semi-transparent manner so that EED can be appreciated. (E) Snapshot of a severe fibrotic simulation with patchy fibrosis distribution. (F) Number of endocardial and epicardial waves per cycle in patient recordings. (G) Number of epicardial waves per cycle in simulations with uniform fibrosis. (H) Number of epicardial waves per cycle in simulations with patchy fibrosis. ^∗indicates significant difference from the control group (p < 0.05).

FIGURE 6

Simultaneous endo–epicardial mapping in a patient’s RA during AF. Dotted line, epicardial activation; dashed line, endocardial activation. Depicted are six simultaneous electrograms, three on each exactly opposing side of the atrial wall. At points A, H, and I, almost synchronous activity was seen with slightly earlier activation at the endocardium, while at B and C, endo- and epicardial activation were out of phase.

FIGURE 7

An example of a breakthrough occurring in the epicardial layer in a patient. (A) A wave enters the endocardial layer (red) (a′, b′, and c′). The epicardium is not activated (a, b, c). The left panels show the local electrograms. The right panels show simultaneous endo–epicardial activation maps at three different time instants. (B) Epicardial breakthrough (b) resulting from propagation from the endocardium. (C) The breakthrough in the epicardial layer spreads further in synchrony with the endocardium.

FIGURE 8

(A) An example of a simulated epicardial breakthrough indicated by a red arrow. (B) Red lines indicate clipping planes. (C) Clipped right atrium at the breakthrough location. (D) A fibrillation wave propagating through an endocardial bundle (black arrow). (E) The fibrillation wave propagated transmurally from the endocardial bundle to the epicardial surface. (F) Appearance of the epicardial breakthrough (red arrow).

FIGURE 9

Examples of simulated BTs. Snap shots of BTs in (A) a control simulation, (B) a moderate fibrotic simulation with patchy fibrosis pattern, (C) a severe fibrotic simulation with uniform fibrosis, and (D) a severe fibrotic simulation with patchy fibrosis. The colorbar indicates the transmembrane voltage. (E) EED in patient recordings. (F) EED in simulations. *indicates significant difference from control in patchy fibrosis (p < 0.05) and ⁺ indicates significant difference from control in uniform fibrosis (p < 0.05). Number of BTs per cycle in panel (G) in patient recordings and (H) simulations.

FIGURE 10

(A) Correlation between EED and BT incidence in patient recordings. Pearson’s correlation r = 0.61 for both endo- and epicardial BT, p < 0.05. (B) Correlation between EED and breakthrough incidence in uniform fibrosis (r = 0.51, p < 0.05) and (C) patchy fibrosis simulations (r = 0.69, p < 0.05).