Large volume acute normovolemic hemodilution in patients undergoing cardiac surgery with intermediate-high risk of transfusion: A randomized controlled trial.

STUDY OBJECTIVE:To investigate whether large volume acute normovolemic hemodilution (L-ANH), compared with moderate acute normovolemic hemodilution (M-ANH), can reduce perioperative allogeneic blood transfusion in patients with intermediate-high risk of transfusion during cardiac surgery with cardiopulmonary bypass (CPB). DESIGN:Prospective randomized controlled trial. SETTING:University hospital. PATIENTS:Patients with transfusion risk understanding scoring tool ("TRUST") ≥2 points undergoing cardiac surgery with CPB in the Second Affiliated Hospital of Zhejiang University from May 2020 to January 2021 were included. INTERVENTIONS:The patients were randomly assigned with a 1:1 ratio to M-ANH (5 to 8 mL/kg) or L-ANH (12 to 15 mL/kg). MEASUREMENTS:The primary outcome was perioperative red blood cell (RBC) transfusion units. The composite outcome included new-onset atrial fibrillation, pulmonary infection, cardiac surgery associated acute kidney injury (CSA-AKI) class ≥2, surgical incision infection, postoperative excessive bleeding, and resternotomy. MAIN RESULTS:Total 159 patients were screened and 110 (55 L-ANH and 55 M-ANH) were included for final analysis. Removed blood volume of L-ANH is significantly higher than M-ANH (886 ± 152 vs. 395 ± 86 mL, P < 0.001). Perioperative RBC transfusion was median 0 unit ([25th, 75th] percentiles: 0-4.4) in M-ANH group vs. 0 unit ([25th, 75th] percentiles: 0-2.0) in L-ANH group (P = 0.012) and L-ANH was associated with lower incidence of transfusion (23.6% vs. 41.8%, P = 0.042, rate difference: 0.182, 95% confidence interval [0.007-0.343]). The incidence of postoperative excessive bleeding was significantly lower in L-ANH vs. M-ANH (3.6% vs. 18.2%, P = 0.029, rate difference: 0.146, 95% confidence interval [0.027-0.270]) without significant difference for other second outcomes. The volume of ANH was inversely related to perioperative RBC transfusion units (Spearman r = -0.483, 95% confidence interval [-0.708 to -0.168], P = 0.003), and L-ANH in cardiac surgery was associated with a significantly reduced risk of perioperative RBC transfusion (odds ratio: 0.43, 95% confidence interval: 0.19-0.98, P = 0.044). CONCLUSIONS:Compared with M-ANH, L-ANH during cardiac surgery inclined to be associated with reduced perioperative RBC transfusion and the volume of RBC transfusion was inversely proportional to the volume of ANH. In addition, LANH during cardiac surgery was associated with a lower incidence of postoperative excessive bleeding.