Abstract:[Objective] To establish an appropriate and effective blood transfusion evaluation model by analyzing the influencing factorsof perioperative blood transfusionin spinal surgery, so as to provide appropriate suggestions and evidence for clinical practice, optimize bloodtransfusion management, save blood resources to the maximum extent and reduce transfusion- related risks. [Methods] A total of 268 pa-tients who received spinal surgery in our hospital from January 2020 to June 2021 were retrospectively analyzed and divided into transfusiongroup and non-transfusion group according to perioperative blood transfusion or not. The basic demographic data, comorbidities, preopera-tive status and operative documents were recorded. The independent impacting factors of blood transfusion were analyzed by univariate com-parison and binary multiple logistic regression, while prediction model of blood transfusion volume was constructed by multiple linear regres-sion. [Results] In term of univariate comparison, the transfusion group proved significantly higher than the non-transfusion group in ratio oftrauma, liver disease, the lumbar surgically involved, decompression and fusion, and no anticoagulants used (P<0.05). The former was signifi-cantly worse than the latter in the preoperative ASA grade and incision healing grade (P<0.05). In addition, the transfusion group had signifi-cantly lower preoperative Hb, HCT and Alb, whereas significantly greater number of segments surgically involved, operative time, intraopera-tive blood loss and postoperative drainage volume than the non-transfusion group (P<0.05). However, there were no statistically significantdifferences between the two groups in age, sex, BMI, diabetes, hypertension, and kidney disease, and preoperative PT and APTT (P>0.05).As results of logistic regression, the intraoperative blood loss (OR=1.010, P<0.05) and operation time (OR=1.020, P<0.05) were the indepen-dent risk factor for blood transfusion, while the preoperative HCT (OR=0.844, P<0.05) was a protective factor for blood transfusion. Regard-ing multiple stepwise regression of 51 patients who received transfusion, the number of segments surgically involved (Xs) (B=2.118, P=0.001) and preoperative Alb (Xa) (B=-0.439, P=0.017) were the main factors of predicating blood transfusion volume (Y), with the regres-sion equation as follows: Y=16.063+2.118Xs-0.439Xa. [Conclusion] Improving patients' preoperative condition and reducing surgical trau-ma is the way to reduce blood transfusion in spinal surgery. The number of segments surgically involved and preoperative Alb are importantfactors in the evaluation and transfusion volume.