Background: The effect of packed red blood cell (PRBC) transfusion on postoperative outcomes of patients undergoing major surgery remains unclear. We sought to determine the impact of blood utilization, as well as transfusion practices, on perioperative outcomes of patients undergoing cardiothoracic-vascular (CT-V) and gastrointestinal (GI) procedures. Methods: Patients who underwent major surgical procedures at Johns Hopkins Hospital between 2009 and 2014 were identified. Data on perioperative hemoglobin (Hb) and blood utilization were obtained; transfusion strategy was categorized as liberal (Hb trigger ≥7 g/dL) vs. restrictive (Hb trigger <7 g/dL). Risk-adjusted logistic regression models and propensity score matching were used to assess the association between transfusion triggers and perioperative morbidity. Results: Among 10,163 patients undergoing either CT-V (50.9 %) or GI (49.1 %) surgery, 4401 (43.3 %) patients received PRBCs. Of the 4401 patients transfused, 71.2 % were transfused using a liberal trigger (≥7 g/dL hemoglobin), while 28.8 % had a restrictive trigger (<7 g/dL). The median number of PRBCs transfused was 3 (restrictive 5 vs. liberal 2 units). While ischemic adverse events were more common among patients undergoing CT-V surgery (17.3 %), infection was the more common complication among patients undergoing GI surgery (11.9 %). American Society of Anesthesiologist (ASA) class 3–4, Charlson score ≥3, and total units of transfused PRBCs were independently associated with overall complications (all P < 0.05). Patients in the restrictive transfusion group did not have increased risk of complications compared with the liberal transfusion group on multivariable analysis (odds ratio (OR) 1.16, 95 % confidence interval (CI) 0.98–1.38; P = 0.08) or after propensity score matching (OR 1.04, 95 % CI 0.88–1.22; P = 0.65). Conclusions: Liberal transfusion triggers after major surgery were more common than restrictive practice. Patients with restrictive transfusion trigger did not have increased risk for complications compared with patients transfused with a liberal trigger.
ASJC Scopus subject areas