OBJECTIVE To evaluate the appropriateness of autologous blood (AB) transfusion during radical retropubic prostatectomy in relation to the cardiopulmonary risk of the patient. METHODS We reviewed the medical records of 100 patients with American Society of Anesthesiologists status I, II, or III who underwent radical retropubic prostatectomy under general or combined general and epidural anesthesia. All patients had donated 2 units (U) of autologous blood, received 0, 1, or 2 U of autologous blood perioperatively, and received no allogeneic blood. Patients were placed in three cardiopulmonary risk groups on the basis of risk factors or documented cardiopulmonary disease. The low-risk group was assigned a target discharge hematocrit of 24% or less; moderate-risk, 25% to 28%; and high-risk, 29% or greater. The appropriateness of transfusion was determined by whether patients' hematocrit was in their group's preassigned range at discharge. RESULTS On the basis of discharge hematocrit, significantly more low-risk patients underwent inappropriate transfusion than moderate-risk (64% versus 26%, P = 0.006) or high-risk (64% versus 13%, P = 0.001) patients. Seventy-five AB units were discarded and at least 53 U were inappropriately transfused. We found an increase in the number of units of autologous blood transfused when a larger estimated blood loss was reported (P < 0.001). The estimated charge for the units discarded and inappropriately transfused exceeded $12,000. CONCLUSIONS Sixty-four percent of autologous blood units were discarded or inappropriately transfused during radical retropubic prostatectomy. Transfusion of autologous blood was not governed by cardiopulmonary risk stratification. If the decision to transfuse had been based on cardiopulmonary risk factors instead of estimated blood loss, fewer patients would have received autologous blood.