Identification of risk factors for allogenic transfusion in cardiac surgery from an observational study

OBJECTIVE: To determine perioperative variables for predicting allogenic transfusion in adult cardiac surgery. STUDY DESIGN: Prospective study. PATIENTS: We included 335 consecutive patients undergoing cardiac surgery between February and April 2001. METHODS: Perioperative variables were prospectively collected in a database. For each patient who received transfusion, hemoglobin threshold for transfusion and total number of units of red cell concentrates were collected. Univariate and multivariate analysis were performed. RESULTS: The two strategies for blood conservation which were predominantly used were aprotinin therapy (78%) and blood salvage from the extracorporeal circuit (68%). During perioperative period, 42% of patients [95% CI: 37-47%] received allogenic transfusion. The haemoglobin threshold for transfusion was 7.4 +/- 1.1 and 8.0 +/- 0.7 g x dl(-1) in operating room and in intensive care unit, respectively. On average, 3.4 +/- 2.7 units of red cell concentrates were transfused perioperatively per patient. Using multivariate analysis, perioperative allogenic transfusion was significantly associated with the following variables: preoperative haemoglobin level < 12 g x dl(-1) (odds ratio 8.9; p = 0.001), emergency procedure (odds = 3.7, p = 0.01), reoperation (odds ratio = 3.3; p = 0.002), chronic obstructive pulmonary disease (odds ratio = 2.5; p = 0.03) and complex surgery (odds ratio = 2.4; p = 0.01). The age, the gender, and body mass index were only independent risk factors by univariate analysis. CONCLUSION: In despite of techniques to limit requirement of allogenic transfusion, a large proportion of cardiac surgical patients remains transfused. Independent risk factors of perioperative transfusion are haemoglobin level < 12 g x dl(-1), emergency procedure, reoperation, chronic obstructive pulmonary disease and complex surgery.