A potential morbidity of incomplete re-warming following hypothermic cardiopulmonary bypass (CPB) is cardiac arrest. In contrast, attempts to fully re-warm the patient can lead to cerebral hyperthermia. Similarly, rigid adherence to 37.0 degrees C during normothermic CPB may also cause cerebral overheating. The literature demonstrates scant information concerning the actual temperatures measured, the sites of temperature measurement and the detailed thermal strategies employed during CPB. A prospective, randomized, controlled study was undertaken to investigate the ability to manage perfusion temperature control in a group of hypothermic patients (28 degrees C) and a group of normothermic patients (37 degrees C). Eighty patients presenting for first-time, elective coronary artery bypass graft surgery (CABG) were randomly allocated to the hypothermic and normothermic groups. All surgery was performed by one surgeon and the anaesthesia managed by one anaesthetist. Temperature measurements were made at the nasopharyngeal (NP) site, as well as in the arterial line of the CPB circuit. The hypothermic group had the arterial blood temperature lowered to 25.0 degrees C to maintain the NP temperature at 28.0-28.5 degrees C. During re-warming, the arterial blood was raised to 38.0 degrees C. Meanwhile, in the normothermic group, the arterial blood temperature was raised to a maximum of 37.0 degrees C to maintain NP temperature at 36.5-37.0 degrees C. Despite strict guidelines, some patients transgressed the temperature control limits. Two patients in the hypothermic group failed to reach an NP temperature of 28.5 degrees C. Twenty-six patients were managed entirely within the control limits. During rewarming in both groups, control of both arterial and NP temperature was well managed with only 25% patients breaching the respective upper control limits. During the re-warming phases of CPB, we were unable to make any correlation between NP temperature and arterial blood temperature, using body weight or body mass index as predictors. Based on the results obtained, we recommend that strict criteria should be implemented for the management of temperature during CPB, in conjunction with more emphasis being placed on monitoring arterial blood temperature as a marker of potential cerebral hyperthermia. We should, therefore, not rely on NP temperature measurement alone during CPB.