The cardiovascular side effects of older antidepressants, such as tricyclic antidepressants, are well established and are known to be linked to their capacity to inhibit cardiac and vascular ion channels. Newer compounds, such as selective serotonin reuptake inhibitors, mirtazapine and venlafaxine, have been reported to have a more benign cardiovascular profile, although they also share antagonistic properties with regard to voltage-dependent ion channels in different tissues. The electrophysiological effects that antidepressants exert on ion channels may affect the cardiac action potential (AP), lengthening both depolarization and repolarization phases, widening the QRS complex, prolonging the QT interval or causing Brugada-like electrocardiogram patterns. Lengthening of the depolarization phase can slow conduction through the His-Purkinje system and myocardium, while slowing repolarization can lead to early after depolarizations and Torsade de Pointes (TdP). In this review, we discuss data from experimental animal models regarding the effects of antidepressants on the cardiac AP, as well as antidepressant-induced QT prolongation in humans and sudden death in patients treated with antidepressants. It appears that although various experimental studies may lead to an understanding of the mechanisms involved in the modulation of cardiac electrical activity, there are significant discrepancies between in vitro data describing the action of antidepressants on the AP, data from clinical trials on QT prolongation by antidepressants and risk of TdP. The role of genetic polymorphisms of potassium-channel-encoding genes in determining the individual risk of cardiac arrhythmias and the limits of QT use as a marker of risk are discussed. Extensive pharmacokinetic and pharmacodynamic studies are required to determine the doses and plasma ranges of each drug that are associated with the greatest risk of arrhythmic complications.