Unexpected events have occurred in cardiology over the last 4 years. A study by the Heart and Lung Institute of the National Institute of Health in the mid-1980s showed to great surprise that class I antiarrhythmic drugs given to patients with ventricular arrhythmias following myocardial infarction, instead of preventing deaths, actually increased the number of patients dying. Since then, a series of studies has consistently confirmed this original observation. The problem for psychiatry is that the tricyclic antidepressant (TCA) drugs are also class I antiarrhythmics. There is every reason to believe that a similar increased risk of death would exist with the TCAs. It is, therefore, important for psychiatrists to understand the cause and magnitude of this excess in deaths. Evidence to date would suggest that all class I compounds, despite being powerful antiarrhythmics under usual physiological conditions, become proarrhythmic under anoxic conditions. Such conditions would exist in ischaemic heart disease during angina and, particularly, myocardial infarction. How this might alter our use of TCAs and whether this happens with the selective serotonin reuptake inhibitors is discussed.