Publication Date



blood, transfusion, liver transplant, anemia, blood


Background: Liver transplantation is often associated with massive blood loss due to surgical complexity and the hemostatic abnormalities of end-stage liver disease. Blood transfusions have been associated with increased risk of infection, multiorgan dysfunction, graft loss and mortality.

Purpose: To determine for liver transplantation whether correlation exists between preoperative anemia and transfusion requirements, length of stay or incidence of postoperative infection.

Methods: A retrospective review of liver transplantations from Jan. 1, 2012, to June 30, 2015, was conducted. Packed red blood cell (PRBC), fresh frozen plasma (FFP), platelet and cryoprecipitate units were collected preoperatively, intraoperatively and within the first 48 hours postoperatively. Cox proportional hazards model was used to model the outcome of infection. Linear regression was used to model the outcomes of postoperative length of stay and blood use.

Results: Of the 112 patients, mean age was 56 years, mean Model for End-Stage Liver Disease score was 27 and mean preoperative hemoglobin was 10.5 g/dL. Lower preoperative hemoglobin was significantly associated with increased preoperative PRBC, platelet and cryoprecipitate use (P < 0.04) as well as increased intraoperative PRBC, FFP, platelet and cryoprecipitate use (P < 0.0001). Preoperative PRBC, FFP, and platelets as well as intraoperative PRBCs were associated with longer length of stay (P < 0.045). Each g/dL decrease in preoperative hemoglobin was associated with a 26% increased risk of infection in univariate models (hazard ratio [HR]: 1.26, P = 0.01). Longer length of stay and higher preoperative cryoprecipitate, intraoperative FFP and postoperative FFP also were associated with increased risk of infection. More units of preoperative cryoprecipitate (HR: 1.07, P < 0.01), fewer units of postoperative cryoprecipitate (HR: 0.19, P < 0.01) and more units of postoperative FFP (HR: 1.75, P < 0.01) were associated with infection in multivariable stepwise selection.

Conclusion: Lower preoperative hemoglobin was associated with increased preoperative and intraoperative transfusion requirements as well as increased postoperative infection. More preoperative cryoprecipitate units, fewer postoperative cryoprecipitate units and more fresh frozen plasma units were independent predictors of infection.