Computational Modeling for Antiarrhythmic Drugs for Atrial Fibrillation According to Genotype

Abstract

Background: The efficacy of antiarrhythmic drugs (AAD) can vary in patients with atrial fibrillation (AF), and the PITX2 gene affects the responsiveness of AADs. We explored the virtual AAD (V-AAD) responses between wild-type and PITX2 ^+/--deficient AF conditions by realistic in silico AF modeling. Methods: We tested the V-AADs in AF modeling integrated with patients' 3D-computed tomography and 3D-electroanatomical mapping, acquired in 25 patients (68% male, 59.8 ± 9.8 years old, 32.0% paroxysmal type). The ion currents for the PITX2 ^+/- deficiency and each AAD (amiodarone, sotalol, dronedarone, flecainide, and propafenone) were defined based on previous publications. Results: We compared the wild-type and PITX2 ^+/- deficiency in terms of the action potential duration (APD₉₀), conduction velocity (CV), maximal slope of restitution (Smax), and wave-dynamic parameters, such as the dominant frequency (DF), phase singularities (PS), and AF termination rates according to the V-AADs. The PITX2 ^+/--deficient model exhibited a shorter APD₉₀ (p < 0.001), a lower Smax (p < 0.001), mean DF (p = 0.012), PS number (p < 0.001), and a longer AF cycle length (AFCL, p = 0.011). Five V-AADs changed the electrophysiology in a dose-dependent manner. AAD-induced AFCL lengthening (p < 0.001) and reductions in the CV (p = 0.033), peak DF (p < 0.001), and PS number (p < 0.001) were more significant in PITX2 ^+/--deficient than wild-type AF. PITX2 ^+/--deficient AF was easier to terminate with class IC AADs than the wild-type AF (p = 0.018). Conclusions: The computational modeling-guided AAD test was feasible for evaluating the efficacy of multiple AADs in patients with AF. AF wave-dynamic and electrophysiological characteristics are different among the PITX2-deficient and the wild-type genotype models.

Keywords: PITX2; antiarrhythmic drugs; atrial fibrillation; gene; modeling.

Figure 1

Study method for the 3D left atrial modeling. The voltage map (A), CT images (B), fiber orientation (C), fibrosis (C), and LAT synchronization (D) were used for the LA modeling. The wild-type and PITX2^+/− deficiency ion currents (E,F) was implemented to simulate for wave dynamics analysis. AF pacing protocol (G) was conducted to analyze the AF initiation and maintenance.

Figure 2

Wild-type vs. PITX2^+/− baseline model analysis. (A–F) The baseline APD₉₀, CV, Smax, AF cycle length, and wave-dynamics parameters for wild-type and PITX2^+/− deficiency were measured for a comparison. The PITX2^+/− deficiency had a shorter APD₉₀ (A), lower mean Smax (C), and PS number (F) than the wild-type.

Figure 3

Characteristics of the wild-type and PITX2^+/− deficiency with the response to AADs. (A,B) The APD₉₀, CV, Smax, AF cycle length, and wave-dynamics parameters were compared with baseline after AADs. (B) Both class III and class IC increased APD₉₀, and AFCL in wild-type and PITX2⁺⁻⁻ deficiency model. Class III and class IC decreased CV in wild-type and PITX2^+/− deficiency model while both classes increased Smax in PITX2^+/− deficiency model.

Figure 4

Termination rate of the wild-type and PITX2^+/− deficiency based on the AADs. (A) Termination rate for different AADs was described. (B,C) Termination rate was compared between the wild-type and PITX2^+/− deficiency and class III and class IC. (B) Class III showed higher AF termination rate compared with class IC. (C) Class IC AADs in PITX2^+/− deficiency indicated the higher AF termination rate compared with the wild type. Class III showed the higher termination rate than class IC.