We conducted a prospective, randomized, controlled trial comparing homologous blood consumption between groups of patients receiving conventional mediastinal drainage (group 1) or reinfusion of shed mediastinal blood (group 2) using hard-shell cardiotomy reservoir. One hundred consecutive patients who had elective coronary artery or valvular operations were studied. The two groups were comparable with regard to age, sex, weight, preoperative and postoperative hemoglobin levels, and surgical procedure. Group 2 patients had their shed mediastinal blood reinfused for up to 18 hours postoperatively; otherwise, the two groups were treated identically. For groups 1 and 2, average mediastinal blood losses were 705 +/- 522 and 822 +/- 445 mL and homologous blood consumption was 3.83 +/- 2.58 and 3.15 +/- 2.05 U, respectively (neither measure was significantly different). However, if blood losses exceeded 500 mL, there was a statistically significant reduction in homologous blood requirements in group 2 as compared with matched controls in group 1. This difference was most significant in patients with the greatest mediastinal losses.