To guide the preparation of blood before liver transplantation: the establishment of multiple regression equation based on R language / 中国输血杂志

ABSTRACT

【Objective】 To establish the multiple regression equation based on R language in order to guide scientific and reasonable blood preparation for clinicians before liver transplantation. 【Methods】 Basic clinical information, including gender, age and disease types, of 183 liver transplant patients were collected, and results of preoperative blood routine(MCV, MCHC, Hct, RBC, Hb and Plt), prothrombin time(PT), activated partial thromboplastin time(APTT), thrombin time(TT), fibrinogen(FIB), international normalized ratio(INR) and D dimer and antithrombin Ⅲ(AT Ⅲ), operation time, as well as intraoperative transfusion volume of red blood cells, plasma, cryoprecipitates and platelets were analyzed using R language analysis.The correlation between blood component transfusion volume and analysis factors was calculated by generalized linear function, and the regression equation for predicting blood preparation was obtained by Poisson regression analysis. 【Results】 Intraoperative transfusion rates of RBC, plasma, platelets and cryoprecipitates in liver transplantation patients were 85.79%(8.35±8.78 U), 100%(1 083±742.80 mL), 18.58%(0.26±0.60 treatment dose), and 12.02%(2.49±7.51 U), respectively. According to the analysis factors with good correlation, the prediction equations for the volume of each blood component were as follows RBC 3.348+ 1.276×Time-0.02×Hct-0.16×RBC-0.006×Hb, plasma 6.901+ 0.826 ×Time-0.003×Hb, platelets -1.275+ 1.866×Time-0.013 Hb+ 0.025×TT, and cryoprecipitates -7.183+ 2.888×Time + 0.067×MCV+ 0.029×TT. 【Conclusion】 It is of great clinical significance to use R language to carry out multivariate regression analysis and establish the prediction regression equation of blood preparation before liver transplantation, which can provide scientific, reasonable and sufficient blood supply in operation.