During cardiopulmonary bypass (CPB) an adequate reservoir volume is maintained by the addition of crystalloid, colloid or packed cells to the reservoir. This volume contributes to the overall perioperative positive fluid balance. We studied the effect of the preoperative addition of either 75 g albumin, or 50 g mannitol followed by 50 g at commencement of rewarming or both of the above to a bypass circuit prime of lactated Ringer's solution (LR) on intraoperative fluid balance, postoperative indices of oxygenation and time to extubation. The study was a prospective, randomized, single-blinded controlled trial of 103 patients undergoing cardiac surgery requiring CPB. There was a large and highly significant reduction in volume of fluid added to the reservoir during CPB (2137 ± 1499 ml versus 144 ± 230 ml), the fluid balance during bypass, including prime volume (3236 ± 650 ml versus 5876 ± 1465 ml), and perioperative fluid balance (4470 ± 936 ml versus 7023 ± 1760 ml) in the group receiving both mannitol and albumin in the pump prime compared with the group receiving only lactated Ringer's solution. There were no differences between the groups with respect to both measured indices of oxygenation measured on return to ICU (alveolar-arterial oxygen tension difference (D_A-aO₂) or arterial oxygen tension to inspired oxygen fraction ratio (P_aO₂/F _iO₂), or time from ICU admission to extubation.