Three hundred and sixty-six consecutive patients admitted to a department of cardiology were evaluated for drug events as a cause of admission. The drug events considered were adverse drug reactions (ADR) and dose-related therapeutic failures (DTF). 'Definite' or 'probable' drug events accounted for 15 admissions (4.1%, 95% confidence limits 2.3-6.7%), of which eleven were ADR and four were DTF. With the inclusion of six 'possible' drug events, the rate of drug-related hospitalizations (DRH) was 5.7%. DRHs were characterized by a preponderance of acute admissions and elderly patients. Hypokalaemia (less than 3.5 mM) was observed in 27 (16%) patients receiving diuretics, and could be related to four cases of arrhythmias (two 'probable' and two 'possible' ADR). The average serum potassium level was similar in diuretic treated patients with or without drugs to counteract hypokalaemia, irrespective of the drugs chosen. Among the 15 'definite'/'probable' DRHs, five were considered to be due to an error in prescription, and a further five cases were judged to have been avoidable had appropriate measures been taken by prescribing physicians. A DRH educational intervention programme should primarily deal with non-compliance or with prescription of diuretics or digoxin, since these problems constitute the majority of cases of DRH. No specific group of doctors could be targeted as responsible for DRH, avoidable or not.