BACKGROUND: A time-dependent decline in cerebral blood flow (CBF) has been reported in cardiac surgical patients despite stable pump flows and arterial carbon dioxide tension. Other studies have failed to support these hypothermic cardiopulmonary bypass (CPB) results, showing preservation of CBF during CPB. The purpose of the study was to define the influence of mildly hypothermic CPB duration on CBF. METHODS: Cerebral blood flow was measured using xenon-133 washout and alpha-stat blood gas management during nonpulsatile CPB. Cerebral blood flow measurements were made after the initiation of CPB and near the end of bypass during pump flows of 2.4 L.min-1.m-2. RESULTS: Fifty-two coronary artery bypass patients were studied. The average time between CBF measurements was 54 +/- 20 minutes (mean +/- standard deviation), with a range of 10 to 100 minutes. Temperature and arterial carbon dioxide tension were controlled: after the initiation of CPB, temperature was 35.5 degrees +/- 0.4 degree C and carbon dioxide tension was 37 +/- 2.8 mm Hg; whereas near the end of bypass temperature was 35.6 degrees +/- 0.5 degree C and carbon dioxide tension was 36 +/- 2.3 mm Hg. We found no correlation between CBF and time on CPB (p = 0.47; r = 0.101), in contrast to other studies suggesting that CPB duration may intrinsically affect CBF. CONCLUSIONS: Our experimental results include the following: (1) during mildly hypothermic bypass, CBF does not decrease in relation to time and (2) cerebral flow-metabolism coupling is intact at 35 degrees C.