RESUMO
PURPOSE: The negative influence of perioperative transfusion of packed red blood cells on the prognosis of various malignancies is the focus of recent research interest. The development of a propensity score for the prediction of perioperative transfusion of packed red blood cells (pRBCs) and the identification of independent risk factors for survival, that can either be known prior to or during surgery in patients undergoing pancreaticoduodenectomy for pancreatic head cancer are the two objectives of this study. METHODS: Logistic regression analyses and Cox regression modeling were used to identify independent risk factors for perioperative transfusion of pRBCs and to determine individual risk factors for patient survival. A total of 101 adult patients who underwent surgery between 01/01/2016 and 12/31/2020 were investigated in a single-center retrospective analysis. RESULTS: Preoperative hemoglobin levels (OR: 0.472, 95%-CI: 0.312-0.663, p < 0.001) and extended resections (OR: 4.720, 95%-CI: 1.819-13.296, p = 0.001) were identified as independent risk factors for perioperative transfusion of pRBCs, enabling the prediction of pRBC transfusion with high sensitivity and specificity (AUROC: 0.790). The logit of the derived propensity model for the transfusion of pRBCs (HR: 9.231, 95%CI: 3.083-28.118, p < 0.001) and preoperative Body Mass Index (BMI) (HR, 0.925; 95%-CI: 0.870-0.981, p = 0.008) were independent risk factors for survival. CONCLUSIONS: Low preoperative hemoglobin levels, low BMI values, and extended resections are significant risk factors for survival that can be known and thus potentially be influenced prior to or during surgery. Patient blood management programs and prehabilitation programs should strive to increase preoperative hemoglobin levels and improve preoperative malnutrition.