The value of acute normovolemic hemodilution (ANH) compared to preoperative autologous blood donation (PAD) in elective surgery is controversial. We therefore conducted a prospective, randomized study to compare these techniques in patients undergoing total knee arthroplasty. ANH patients underwent up to 4 units phlebotomy or to a target hematocrit level of 28% after induction of anesthesia. PAD patients were asked to donate 1 (unilateral) or 2 (bilateral, revisions) units before admission. Mean baseline hematocrit levels were not different between ANH and PAD patients (40.6 ± 4.1 vs. 38.4 ± 3.4, p = 0.09). Eight (73%) of 11 patients undergoing bilateral revision procedures received a total of 22 allogeneic blood units, whereas only 3 (14%) of 21 patients undergoing primary, unilateral procedures received a total of 3 allogeneic units (p = 0.002). We found no differences in allogeneic blood transfusions between ANH and PAD cohorts for all (n = 32) patients (1.0 ± 1.2 vs, 0.6 ± 1.4, p = 0.45), for unilateral knee (n = 21) replacement (0.25 ± 0.46 vs. 0.08 ± 0.28, p = 0.29), or for bilateral/revision (n = 11) procedures (1.9 ± 1.3 vs. 2.5 ± 1.9, p = 0.53). We conclude that each technique is equally effective in reducing allogeneic blood exposure. Patients undergoing revision or bilateral knee arthroplasties require adjunctive therapy to autologous blood procurement to further reduce allogeneic blood exposure.