We studied the predictive performance of the Minto pharmacokinetic model during cardiopulmonary bypass in patients undergoing cardiac surgery. Patients received remifentanil target-controlled infusion using the Minto model during total intravenous anaesthesia with propofol. From 56 patients, 275 arterial blood samples were drawn before, during and after bypass to determine the plasma concentration of remifentanil, and the predicted concentrations were recorded at each time. For pooled data, the median prediction error and median absolute prediction error were 21.3% and 21.8%, respectively, and 22.1% and 22.3% during bypass. Both were 148.4% during hypothermic circulatory arrest and measured concentrations were more than three times greater than predicted (26.9 (17.0) vs. 7.1 (1.6) ng.ml ). The Minto model showed considerable bias but overall acceptable precision during bypass. The target concentration of remifentanil should be reduced when using the Minto model during hypothermic circulatory arrest.