The surgical correction of congenital heart disease using haemodilution and hypothermia with cardiopulmonary bypass (CPB) may expose patients to tissue ischaemia and initiate a systemic inflammatory response, increasing the total body water and inducing impairment, especially of heart, lung and brain function. It is possible to use ultrafiltration during CPB in the rewarming phase to remove water accumulation in the third space (conventional ultrafiltration). The reduced volume of prime used in children and the ability only to filter the reservoir blood during CPB led the Great Ormond Street Group to modify the method of ultrafiltration with regards to the placement of the filter and the timing of filtration (post-CPB). The main advantage of the modified technique is the ultrafiltration of the patient. A prospective nonrandomized study has been conducted to compare conventional with conventional + modified ultrafiltration. From January 1996 to March 1998, 41 patients underwent correction of congenital heart disease and were submitted to a comparative study (homogeneous groups), using either the conventional or the conventional + modified ultrafiltration techniques. There were no technical complications, no patient required mediastinal re-exploration due to bleeding and it was possible to close all the chests. There were significant differences in the ultrafiltrate volume balance (143.3 +/- 54.3 versus 227 +/- 71.4 mL; P < 0.001) but there were no significant differences in clinical postoperative evolution between the conventional and the conventional + modified ultrafiltration.