Patients undergoing coronary artery surgery run a certain risk of developing myocardial infarction in situations with increased myocardial oxygen demand due to e.g. elevations in heart rate and blood pressure. After cardiopulmonary bypass (CPB) there is, however, also the risk of graft occlusion. The present study evaluated the haemodynamic effects of a sequential anaesthesiological technique using halothane 0.5-1.5% in combination with 50% nitrous oxide and droperidol 0.1 mg X kg b.w.-1 before CPB followed by fentanyl 0.2 mg X h-1 in continuous i.v. infusion and diazepam 10-15 mg during and after bypass. Fourteen patients were studied. In seven patients (Group I) halothane was discontinued immediately before CPB and in the following seven patients (Group II) 10-15 min before bypass. The aim was to depress moderately the inotropic state before bypass and to have a normalized myocardial oxygen demand after CPB in order to promote a good flow in the grafts. In Group II mean left ventricular stroke work index (LVSWI) was 0.54-0.79 J X m-2 after bypass as compared to 0.45-0.51 before at comparable filling pressure. Mean left ventricular power index (LVPI) increased from 0.42-0.55 W X m-2 before to 0.73-1.08 after CPB. The patients in Group I showed a similar pattern although with a less marked difference in LVSWI and LVPI values before and after bypass. No correlation was seen between oxygen delivery and oxygen uptake either below or above an oxygen delivery of 15 mmol X min-1 X m-2.