Abstract—The Na+channel blocking activity and the antiarrhythmic effects of riluzole, and established anticonvulsants (lamotrigine and lifarizine) and class 1 antiarrhythmics (lidocaine, flecainide and disopyramide) were studied under in vitro and in vivo conditions. Guinea‐pig cardiac Purkinje fibres were superfused with Tyrode solution and electrically driven for recording action potentials with intracellular microelectrodes. In these preparations paced at 1 Hz, all compounds tested produced concentration‐dependent (0.3–100 μM) reductions in the maximum rate of depolarization of the action potential (Vmax). For riluzole, phenytoin and carbamazepine this effect was frequency‐independent (0.5–6 Hz) but for lamotrigine, lifarizine, lidocaine, flecainide and disopyramide it was frequency‐dependent. In anaesthetized rats, riluzole, in contrast to flecainide, did not delay the appearance of aconitine‐induced arrhythmias. Riluzole (0.3‐3.9 mg/kg, i.v.) also lacked notable cardiac electrophysiological effects in anaesthetized dogs. At an i.v. dose of 3.0 mg/kg riluzole failed to restore a normal sinus rhythm in conscious dogs with polymorphic arrhythmias produced by ligation of the left anterior descending coronary artery 24 h earlier. These results indicate that riluzole, phenytoin and carbamazepine, unlike lamotrigine, lifarizine and flecainide, block cardiac Na+channels in a frequency‐independent manner. This property may account for the lack of antiarrhythmic activity of riluzole, phenytoin and carbamazepine in animal models of arrhythmias that respond to class I antiarrhythmic drugs. It may also account for the clinical observation that riluzole does not seem to cause the unfavourable electrocardiographic changes characteristic of drugs that block cardiac Na+channels in a frequency‐dependent manner.