Atrial pacing was used for preoperative evaluation of six patients with recent anterior myocardial infarction (MI) (e.g. within 6 weeks) scheduled for abdominal emergency surgery. Central and coronary haemodynamics were used to compare changes in myocardial work and oxygenation with alternations of the non-invasive variables rate pressure product (RPP) (systolic blood pressure X heart rate), triple product (TP) (systolic blood pressure X heart rate X mean pulmonary arteriolar occlusion pressure) and ST-T segments (lead V5). There was good correlation between myocardial oxygen consumption and rate pressure product and triple product during pacing to stable angina pectoris. ST-T-segment depressions were recorded already at moderate chest discomfort and correlated well with a decrease in coronary vascular resistance. Changes in myocardial oxygen consumption induced by combined thoracic epidural analgesia (T3-4 to L1-2) and light general anaesthesia with nitrous oxide and fentanyl were poorly correlated with changes in rate pressure product or triple product. ST-T-segment depressions were recorded on five occasions in four of the patients, all in association with intubation and/or extubation. Only on one of these occasions could RPP or TP have indicated that myocardial oxygen demand exceeded supply. On the other four occasions, it was probable that myocardial ischaemia was induced by transient arterial hypoxaemia. The V5 ST-T-segment was the most sensitive non-invasive variable to monitor. The anaesthetic method was safe in all patients, as judged by good intraoperative cardiovascular stability, low morbidity and absence of intra- or postoperative reinfarction.