Safety of Flecainide and Propafenone in Patients With Atrial Fibrillation and Structural Heart Disease

Class Ic antiarrhythmic drugs (AADs) have been contraindicated in patients with structural heart disease since the 1991 Cardiac Arrhythmia Suppression Trial (CAST), which demonstrated increased mortality in post-myocardial infarction patients with ventricular ectopy and reduced ejection fraction. Contemporary guidelines extend this prohibition broadly to coronary artery disease (CAD), heart failure, and other structural heart disease, yet CAST enrolled a narrow population, and the totality of post-CAST evidence across structural heart disease has never been comprehensively synthesized.

To systematically evaluate the safety of class Ic AADs (flecainide and propafenone) compared with alternative antiarrhythmic strategies in patients with atrial fibrillation (AF) and structural heart disease, with CAD as the principal prespecified subgroup.

We searched PubMed, Embase, Cochrane Library, and Web of Science from inception through May 2026 for studies comparing class Ic AAD use with class III AADs, amiodarone, or no antiarrhythmic therapy in patients with AF and structural heart disease (CAD, heart failure, or left ventricular hypertrophy). The primary endpoint was ventricular arrhythmia (VA); secondary endpoints were all-cause mortality and major adverse cardiovascular events (MACE). The primary analysis pooled all structural heart disease populations; CAD was the principal prespecified subgroup, and a test for subgroup differences (CAD vs. non-CAD) was performed. Pooled hazard ratios (HRs) with 95% confidence intervals (CIs) were calculated using restricted maximum likelihood (REML) random-effects models with the Hartung-Knapp adjustment. Heterogeneity was assessed with I² and Cochran Q statistics.

Seven studies (N = 35 886) met inclusion criteria; six contributed to quantitative pooling. In the primary analysis across all structural heart disease populations, class Ic AAD use was not associated with increased ventricular arrhythmia risk (pooled HR 0.73; 95% CI 0.51-1.03; p = 0.065; I² = 88.6%; k = 5), and no individual study reported a statistically significant increase in VA risk. All-cause mortality was numerically lower but did not reach significance after the Hartung-Knapp adjustment (HR 0.47; 95% CI 0.21-1.01; k = 5); MACE was significantly lower with class Ic use (HR 0.54; 95% CI 0.39-0.75; p = 0.010; k = 4). The CAD subgroup was directionally concordant and showed no harm (VA HR 0.80 [95% CI 0.54-1.17]; mortality HR 0.65 [95% CI 0.22-1.91]; MACE HR 0.54 [95% CI 0.29-0.99]). Tests for subgroup differences between CAD and non-CAD populations were nominally significant for VA (p = 0.009) and mortality (p = 0.037) but not MACE (p = 0.92). Substantial heterogeneity was observed across all endpoints.

Across structural heart disease populations, including a well-powered CAD subgroup, class Ic antiarrhythmic therapy for AF was not associated with excess ventricular arrhythmia risk, and no subgroup demonstrated harm. The absence of a proarrhythmic signal across diverse study designs and substrates provides supportive, though not definitive, hypothesis-generating evidence regarding the safety of these agents in carefully selected patients with stable disease and preserved ventricular function. Prospective randomized trials are needed to confirm these findings. PROSPERO Registration: CRD420261411804.