High-dose opioids are advocated for paediatric cardiac surgery to suppress stress responses but they can produce unwanted side effects. There are no data on the dose-dependent effects of opioids on the stress response on which to base rational opioid administration. We conducted a dose ranging study on 40 children less than 4 yr undergoing elective open heart surgery using one of five fentanyl doses: 2, 25, 50, 100 or 150 mu g kg(-1) before surgery. The standardized anaesthetic also included pancuronium and isoflurane. Blood samples were taken at induction, before incision, after sternotomy, immediately before, and at the end of cardiopulmonary bypass. Patients in the 2 mu g kg(-1) group had significant rises in pre-bypass glucose (P<0.01), pre- and post-bypass cortisol (P<0.01), and pre- and post-bypass norepinephrine (P<0.01). No significant rise occurred in glucose, cortisol and catecholamines in any of the higher dosage groups. Patients in the 2 mu g kg(-1) group had significantly higher mean systolic blood pressure (P<0.02) and heart rate (P<0.04). A balanced anaesthetic containing fentanyl 25-50 mu g kg(-1) is sufficient to obtund haemodynamic and stress responses to the pre-bypass phase of surgery. Higher doses of fentanyl (100 and 150 mu g kg(-1)) offer little advantage over 50 mu g kg(-1), and can necessitate intervention to prevent hypotension.