Analysis of influencing factors of massive blood transfusion during liver transplantation and constructing a predictive model / 中华器官移植杂志

ABSTRACT

Objective:To explore the influencing factors of massive blood transfusion during liver transplantation and establish a prediction model for massive blood transfusion.Methods:Clinical data were analyzed retrospectively for 103 patients undergoing liver transplantation performed by the same surgical team of our hospital in 2018 from Department of Blood Transfusion Affiliated Drum Tower Hospital Nanjing University Medical School. According to the intraoperative volume of red blood cells, they were divided into two groups of massive transfusion (≥12U, n=40) and non-massive transfusion (<12U, n=63). General patient data and preoperative laboratory parameters were compared between two groups. Logistic regression analysis was utilized for constructing a prediction model for massive transfusion during liver transplantation. Results:No significant inter-group difference existed in gender, age or blood group ( P>0.05) while preoperative diagnosis, hemoglobin (Hb), hematocrit (HCT), platelet count (PLT), international normalized ratio (INR), prothrombin time (PT), activated partial thromboplastin time (APTT), thrombin time (TT), total bilirubin (TBIL), direct bilirubin (DBIL) and end-stage liver disease (MELD) score were correlated with intraoperative massive blood transfusion ( P<0.05). By Logistic regression analysis, the predictive model of massive blood transfusion in liver transplantation was Y = 3.545-0.112 × HCT-0.009 ×P LT + 0.005×DBIL. The area under ROC curve was 0.813 and the sensitivity and specificity 80.0% and 71.4% respectively. Youden's index of 0.514 indicated that if Y value was greater than or equal to 0.514, massive blood transfusion might occur. Conclusions:Preoperative HCT, PLT and DBIL may be employed as independent predictors of intraoperative massive transfusion during liver transplantation. The prognostic model has demonstrated an excellent sensitivity and specificity.