OBJECTIVE To evaluate the safety and preliminary efficacy of escalating doses of hemoglobin raffimer (Hemolink) with intraoperative autologous blood donation for coronary artery bypass graft (CABG) surgery. METHODS Randomized, controlled, single-blind phase II dose escalation trial. METHODS Multi-institutional university setting. METHODS Adult patients (n = 60) undergoing elective CABG surgery. METHODS After induction of anesthesia, autologous whole blood was collected to achieve a hemoglobin of 7 g/dL on cardiopulmonary bypass. Patients were randomized to receive either hemoglobin raffimer (treatment) or 6% hetastarch (control) in sequential escalating dose blocks of 250 mL, 500 mL, or 750 mL. After return of autologous blood, allogeneic red blood cells were transfused according to predetermined hemoglobin triggers. RESULTS Safety parameters (vital signs, hematology, blood chemistry, coagulation, and adverse events) were monitored from randomization through week 4 postdischarge. Serious adverse events were distributed evenly between the 2 groups of patients. Elevated blood pressure was more frequent in the treatment group (16/28 mmHg v 9/32 mmHg, p = 0.036). In the group of 40 patients in the 750-mL dose block, 8 of the 18 treatment patients and 4 of the 22 control patients avoided allogeneic red blood cell transfusion (p = 0.093). Median volume of allogeneic red blood cells transfused was lower in treated subjects compared with controls (p = 0.042). CONCLUSIONS Hemoglobin raffimer is well tolerated and may be effective in reducing transfusion for patients undergoing CABG surgery. Although perioperative hypertension was more frequent in the treated patients, blood pressure management prevented serious adverse sequelae. Definitive evaluation of efficacy in a larger phase III trial is warranted.