Establishment Of Risk Scoring System For Progressive Hemorrhagic Injury In Patients With Isolated Traumatic Brain Injury

Purpose:To explore risk factors of progressive hemorrhagic injury(PHI)in patients with isolated traumatic brain injury(ITBI)and to develop a prognostic model based on information available at admission.Methods:A retrospective study was conducted,the patients with ITBI,from January 1,2014 to December 31,2018 in Nan Fang Hospital department,were selected.Criteria for choosing the patients are as follows:1)admitted within 12h of trauma and undergone initial head CT examination,age from 18 to 85 years;2)A clear history of brain injury;3)Routine repeat head CT were performed within 48hours at least once;4)Abbreviated Injury Scale[AIS]score for head≥3,ASI for other areas of the body<3.The exclusions of the subjects were as follows:1)Anticoagulant drug or antiplatelet therapy was done in past 3 months;2)the subjects with known coagulation disorders,such as liver disease,hemophilia,malignant tumors,deep venous thrombosis,were excluded;3)ASI for other areas of the body≥3;4)The total volume of fluid intake is less than 2L before the hospitalization;5)Subjects with inadequate imaging;6)the subjects who fail to perform the CT examination due to death or refuse the recheck.PHI was defined as follows:(1)a new brain hematoma found on repeat head CT scan;(2)the brain hematoma volume on repeat head CT scan was increased by 25%or more compared to the initial CT scan.All patients are divided into two group,such as PHI group and Non-PHI group,respectively.Clinical data of all subjects were collected including characteristics,coagulation test and CT images,etc.Univariate and multivariate logistic analysis were used to identify independent risk factors associated with PHI.Based on proportional conversion of beta regression coefficients,points were assigned to risk factors and create a risk score of PHI.According to the total scores calculated,all subjects were divided into three groups:low-risk group(0-4points),moderate-risk group(5-10points),high-risk group(11-22points).Meanwhile,we calculate and compare the number of patients in the developed PHI,operations and clinical outcomes in three groups.Data analysis was performed using SPSS 20.0,measurement data was expressed as mean±standard deviation.Enumeration data were analyzed by Chi-square test.T-test and Chi-square test was used to analyze the differences among clinical outcomes in three groups.the Hosmer-Lemeshow test was employed to evaluate the fitting capability of the model.P<0.05 is considered statistically significant.Result:Of the 162patients,23.46%(38)experienced PHI.Univariate analysis showed that the influencing factors associated with PHI were subarachnoid hemorrhage,subdural hematoma,intracranial hematoma,midline shift,systolic blood pressure,GCS score,head AIS score,Age,PLT≤150×109/L,fibrinogen≤1.8g/L and INR≥1.2,respectively.Five independent risk factors were identified as follows:midline shift≥0.5cm(6 points),systolic blood pressure≥148mmHg(7 points),GCS<12(5 points),INR≥1.2(4 points),Fg≤1.8 g/L(4 points).The PHI rates after trauma in the three group was 8.4%,24.1%and 72%,respectively;the probability of poor prognosis was 0%,11.1%and 68%,respectively;and the rates of surgical intervention was 6.0%,14.8%and 36%,respectively;Significant statistically was observed in three groups.The risk scores system of PHI model showed good fit(Hosmer-Lemeshow,P=0.96)and discrimination(area under the ROC curve=0.802).Conclusion:Midline shift≥0.5cm,Systolic blood pressure≥148mmHg,GCS score≤12,INR≥1.2,Fg≤L8g/L are independent risk factors of PHI.A prediction model of PHI is developed and internally validated and risk score may accurately predict the probability of PHI based on five easily available information.A simple risk score based on information available at admission may accurately identify the risk for PHI after brain injury.