Class Ic antiarrhythmic drugs are effective in the treatment of atrial fibrillation, but their mechanism of action is unknown. In previous work, we have found that flecainide causes tachycardia-dependent increases in atrial action potential duration (APD) and effective refractory period (ERP) by reducing APD accommodation to heart rate. The present study was designed to evaluate the efficacy and mechanisms of action of flecainide in an experimental model of sustained atrial fibrillation (AF). AF was produced by a brief burst of atrial pacing in the presence of vagal stimulation and persisted spontaneously until vagal stimulation was stopped. The actions of flecainide at two dose levels were compared with those of isotonic glucose placebo in each dog, with a randomized order of blinded drug administration. Flecainide terminated AF in all 16 dogs, while glucose was effective in none (p less than 0.0001). Flecainide increased atrial ERP and reduced conduction velocity in a tachycardia-dependent manner. Doses of flecainide that converted AF resulted in larger changes in ERP than in conduction velocity, increasing the minimum path-length capable of supporting reentry (wavelength). In addition, flecainide reduced regional heterogeneity in ERP and wavelength, an action opposite that of vagal stimulation. Atrial epicardial mapping with a 112-electrode atrial array was used to study the mechanism of flecainide action on AF. Under control conditions, multiple small zones of reentry coexisted. Flecainide progressively increased the size of reentry circuits, decreased their number, and slowed the frequency of atrial activation until the arrhythmia finally terminated; all changes were compatible with an increase in wavelength. We conclude that flecainide terminates atrial fibrillation in this experimental model by causing tachycardia-dependent increases in atrial ERP, which increase the wavelength at the rapid rates characteristic of AF to the point that the arrhythmia can no longer sustain itself.