Antidepressant drugs are currently the mainstay of treatment for all but the mildest forms of depression. Their effectiveness in the management of depressive illness is undisputed and their effectiveness in preventing suicide, while not proven, may be assumed. Nevertheless, of all the drugs that are taken in lethal overdose, prescribed antidepressants are among the most common. Epidemiological studies from several countries have provided evidence of marked differences in overdose toxicity between drug classes and, in some cases, between individual drugs within a class, with some of the older tricyclic antidepressants (TCAs) being the most toxic. Over 80% of all deaths arising from overdose of antidepressant medication in the UK between 1987 and 1992 were caused by 2 drugs: amitriptyline and dothiepin. Taken alone, this figure conveys little information about the toxicity of either drug. However, when considered within an epidemiological context, the evidence suggests that both drugs are highly toxic in overdose, a conclusion that is supported by animal studies of the toxicity of TCAs and by clinical evidence of overdose toxicity. This paper reviews the epidemiological evidence concerning the acute toxicity of antidepressant drugs and considers the interplay of factors that contribute to the toxicity which occurs when they are taken in acute overdose. The inherent toxicity of the drug appears to be the crucial factor and, although less well researched, prescribing practices and perception of toxicity are probable contributory factors.