A variable rate continuous infusion of alfentanil 12.5-50 mg h-1, commenced after tracheal intubation, was used as the sole opioid anaesthetic agent in 14 patients undergoing coronary artery surgery. The infusion rate was adjusted to somatic responses and to maintain a stable systolic arterial pressure. Plasma alfentanil concentrations were determined throughout anaesthesia, including peaks of surgical stress and for 5 h after operation. Before cardiopulmonary bypass (CPB), a mean infusion rate of 36.6 +/- 8.9 mg h-1 minimized the frequency of hypertensive episodes and resulted in high plasma alfentanil concentrations which reached a mean peak, 1.76 +/- 0.46 micrograms ml-1 at maximal sternal spread. After CPB, a slower infusion rate, 6.9 +/- 5.8 mg h-1, afforded cardiovascular stability. Patients recovered consciousness 3.1 +/- 1.4 h after operation at a mean plasma alfentanil concentration of 0.27 +/- 0.13 micrograms ml-1. Apart from clinical response, parameters have not yet been identified which may optimize infusion rates.