Antiarrhythmic drugs can be divided into four Vaughan Williams classes (I-IV) according to defined electrophysiological effects on the myocardium. Thus, the Vaughan Williams classification also coincides with the main myocardial targets of the antiarrhythmics, i.e., myocardial sodium-, potassium-, and calcium-channels or beta-adrenergic receptors. A more detailed characterization which is also based on the myocardial targets of a drug is given by the "Sicilian Gambit" approach of classification. Nevertheless, the appropriate drug for the management of a given clinical arrhythmia has to be chosen according to the electrophysiological effects of the respective drug. A main determinant of the antiarrhythmic or proarrhythmic properties of a drug is the frequency dependence of its electrophysiological effects. The sodium-channel blockade induced by class-I substances is enhanced with increasing heart rates. Thus, class-I antiarrhythmics can be subclassified as substances showing a more exponential, an approximately linear, or rather saturated block-frequency relation. Class-III antiarrhythmics (potassium-channel blockade) can be further differentiated according to the component of the delayed rectifier potassium current (IK) which is inhibited by a drug. Class-III drugs inhibiting selectively the rapidly activating and deactivating IKr component exhibit a marked reverse rate dependence, i.e., the drug induced prolongation of the cardiac action potential is minimized at high rates. On the other hand, during bradycardia the pronounced action potential prolongation may cause early afterdepolarizations and triggered activity leading to torsades de pointes arrhythmias (acquired QT syndrome). Class-III substances inhibiting the slowly activating IKs component are currently under investigation and are expected to show a direct rate dependence. Experimental data available so far point to an action potential prolonging effect at least independent of rate. However, it is uncertain whether proarrhythmic effects can be thus avoided, especially in light of the fact that one form of congenital QT syndrome (LQT1) seems to be linked to dysfunction of the IKs-channel.