In this retrospective analysis we tested the hypothesis that aprotinin doses of more than 6 x 10(6) kallikrein inhibiting units (KIU) per patient may be more effective in reducing bleeding compared with the high-dose regimen of 5-6 x 10(6) KIU aprotinin. The aprotinin doses administered for 8281 adult cardiac surgical patients were correlated to body weight and time of operation and calculated in KIU per kg body weight and minute of operation. Linear and logistic regression models were designed to detect potential associations between dose and postoperative bleeding, transfusion, and other covariates. The 6-h chest tube drainage in the lowest quartile dosing group was 447 +/- 319 mL (mean +/- sd) compared with 360 +/- 290 mL in the highest quartile dosing group (P < 0.001). The proportion of patients requiring allogeneic blood transfusion was reduced from 55% to 47% comparing the lowest with the highest dosing group (P < 0.01). Aprotinin dose was also an independent predictor for rethoracotomy for surgical hemostasis (1.9% in the highest quartile to 2.4% in the lowest dosing quartile; P < 0.01). The risk of renal failure requiring dialysis (2.3% in the highest dosing group vs 3.3% in the lowest dosing group; P < 0.01) or impairment of renal function (creatinine increase of >or=2 mg/dL postoperatively, 6.4% in the highest dosing group vs 10.0% in the lowest dosing group; P < 0.01) was lower with higher doses of aprotinin. Thus, there was no association between aprotinin dose and renal function. Our results support the hypothesis that a more individualized aprotinin regimen with potentially higher doses may optimize the effectiveness of aprotinin therapy in cardiac surgery.