The accumulation of activated leukocytes in the pulmonary circulation plays an important role in the pathogenesis of lung dysfunction associated with cardiopulmonary bypass. Animal studies have demonstrated that the elimination of leukocytes from the circulation reduces postoperative lung injury and improves postoperative pulmonary function. We conducted a prospective randomized clinical study to evaluate whether postoperative lung function could be improved by use of a leukocyte filter during cardiopulmonary bypass. Elective coronary artery bypass grafting was done with a leukocyte-depleting arterial blood filter incorporated in the extracorporeal circuit (14 patients, leukocyte filter group) or without the filter (18 patients, control group). Blood samples collected at intervals before, during, and after operation were used for analysis of blood cell counts, elastase concentrations, and arterial blood gases. The use of the leukocyte filter caused no significant reduction in leukocyte count (p = 0.86). There were no differences in postoperative lung function between the groups, as assessed through (1) oxygenation index (290 for leukocyte filter group compared with 329 for control group, 95% confidence interval, 286 to 372, p = 0.21), (2) pulmonary vascular resistance (p = 0.10), and (3) intubation time (16.6 hours for leukocyte filter group versus 15.7 hours for control group, 95% confidence interval, 12.1 to 19.1 hours, p = 0.72). The levels of neutrophil elastase were significantly higher at the end of cardiopulmonary bypass in the leukocyte filter group (460 microgram/L in leukocyte filter group versus 230 microgram/L in control group, 95% confidence interval, 101 to 359 microgram/L, p = 0.003). We conclude that the clinical use of the present form of leukocyte-depleting filter did not improve any of the postoperative lung function parameters analyzed in this study.