Risk factors analysis and predictive model development and validation for trauma-induced coagulopathy in pediatric patients with moderate-to-severe traumatic brain injury
10.3760/cma.j.cn501098-20250421-00241
- VernacularTitle:中重型创伤性脑损伤患儿并发早期创伤性凝血病的危险因素及其预测模型构建与验证
- Author:
Yuchen LIU
1
;
Yi ZHONG
;
Hailing YANG
;
Zhenjiang BAI
;
Feng LIU
;
Hangzhou WANG
Author Information
1. 苏州大学附属儿童医院神经外科，苏州 215025
- Publication Type:Journal Article
- Keywords:
Brain injuries;
Child;
Nomograms;
Blood coagulation disorder
- From:
Chinese Journal of Trauma
2025;41(8):754-763
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To analyze risk factors for early trauma-induced coagulopathy (TIC) in pediatric patients with moderate-to-severe traumatic brain injury (msTBI), develop a predictive model and evaluate its predictive performance.Methods:A retrospective cohort study was conducted to analyze the clinical data of 290 pediatric patients with msTBI who were admitted to Children′s Hospital of Soochow University between January 2016 and December 2024, including 188 boys and 102 girls, aged 0.2-15.7 years [5.2(2.8, 9.3)years]. Based on the coagulation test results at admission, the patients were divided into TIC group ( n=162) and non-TIC group ( n=128). The patients were randomly allocated into training set ( n=203) and validation set ( n=87) at a ratio of 7∶3. Demographic characteristics, clinical data, vital signs, imaging findings, arterial blood gas analysis results, and coagulation profiles of the patients were collected. Univariate analysis and Lasso regression analysis were used to identify risk factors associated with early TIC in children with msTBI and multivariate Logistic regression analysis was performed to determine independent risk factors and construct a predictive model. The model′s discrimination and calibration were evaluated using the area under the receiver operating characteristic (ROC) curve (AUC), Hosmer-Lemeshow (H-L) test, and calibration curve. Its clinical utility was assessed through decision curve analysis (DCA). Results:Significant differences were observed between the TIC group and non-TIC group in terms of age, weight, time from injury to admission, child′s Glasgow coma scale (CGCS) score, pediatric trauma score (PTS), shock index, heart rate, respiratory rate, systolic blood pressure, Rotterdam CT score, intraventricular hemorrhage, cerebral contusion, brain herniation, long bone fracture, pelvic fracture, hemopneumothorax, pulmonary contusion, intra-abdominal organ injury, actual bicarbonate, base excess, base excess in the extracellular fluid, blood glucose, hemoglobin (Hb), osmolarity, blood calcium, anion gap, blood lactate, prothrombin time, activated partial thromboplastin time, international normalized ratio, and platelet count ( P<0.05). With coagulation-related variables excluded, the following features were identified with Lasso regression including CGCS score, PTS, heart rate, systolic blood pressure, long bone fracture, blood glucose, and Hb. Multivariate Logistic regression analysis revealed that CGCS score≤8 points ( OR=3.05, 95% CI 1.65, 5.63), PTS>5 points ( OR=0.45, 95% CI 0.23, 0.89), systolic blood pressure ( OR=0.98, 95% CI 0.97, 0.99), blood glucose ( OR=1.09, 95% CI 1.01, 1.17), and long bone fracture ( OR=2.47, 95% CI 1.13, 5.42) were influencing factors for early TIC in children with msTBI ( P<0.05). The regression equation of the predictive model was established as follows: Logit[ P/(1- P)]=1.01×"CGCS score≤8 points"-0.69×"PTS>5 points"- 0.02×"systolic blood pressure"+0.89×"long bone fracture"+0.08×"blood glucose"+1.32. The ROC curve analysis showed that the training set had an AUC of 0.86 (95% CI 0.78, 0.94), with sensitivity and specificity of 76.6% and 92.5%, while the AUC was 0.80 (95% CI 0.74, 0.86), with sensitivity and specificity of 75.7% and 79.6% in the validation set. H-L test results showed a χ2 value of 8.18 ( P=0.416) in the training set and 5.30 ( P=0.216) in the validation set. The calibration curves for both sets demonstrated good agreement with the actual curves, indicating that the predicted probabilities closely matched the observed probabilities. The DCA results indicated that both the training set and validation set demonstrated positive net benefits within threshold probabilities ranges of 10%-100% and 15%-96%. Conclusions:Independent risk factors for early TIC in pediatric msTBI patients include CGCS score≤ 8 points, PTS≤5 points, low systolic blood pressure, long bone fracture, and high blood glucose. The predictive model constructed based on these factors demonstrates favorable predictive performance and clinical application value.
