The role of temporary percutaneous endocardial pacing has been examined in a retrospective analysis of all paced patients admitted to one coronary care unit over a 6 year period. The majority of 162 cases (84.6%) were paced for complete heart block complicating acute myocardial infarction. These patients had a higher incidence of previous hypertension, myocardial infarction and diabetes, compared to matched controls (P less than 0.05, less than 0.02 and less than 0.001, respectively). Admission blood glucose levels were also higher (P less than 0.05). The in-hospital mortality was high (46.7%), especially for those with anterior myocardial infarction (74.5%). Twenty-five (15.4%) patients without recent myocardial infarction were paced for symptomatic brady-dysrhythmias, usually due to chronic complete heart block (Lenegre's disease) or sick sinus syndrome. Most later required permanent pacing. Complications of temporary pacing were more frequent in those who died, the most common being dysrhythmias during pacemaker insertion. Review of our cases suggests that whilst facilities for temporary pacing were extremely valuable, many cases treated were not haemodynamically compromised and probably did not require pacing. Guidelines should be established on coronary care units to prevent the unnecessary morbidity, mortality and expense of the procedure.