Hao Wei, Feng Jiancheng, Luo Hongliang, Ma Ruifang, Lu Xuan, Xiong Dongsheng, Liu Yong
Department of Neurosurgery, Ordos Central Hospital, Ordos, 017000, China.
Department of Neurosurgery, Tianjin First Central Hospital, Tianjin, 300052, China.
Sci Rep. 2025 Aug 11;15(1):29313. doi: 10.1038/s41598-025-15287-z.
To construct a precise and personalized nomogram model and assess the risk factors associated with deep vein thrombosis (DVT) in patients undergoing (traumatic brain injury) TBI. Clinical data from TBI patients between January 2015 and January 2020 were retrospectively gathered. Divided into the model training set and the model validation set in chronological order. The risk factors for DVT were analyzed using LASSO regression and multifactor logistic regression. Post-modeling assessments were conducted for differentiation, consistency, and clinical efficacy. LASSO regression results showed that Age, BMI, smoking history, balance of intake and output, interval between operation and injury, preoperative D-dimer, preoperative FIB, and preoperative PT were the risk factors of DVT in patients with TBI after surgery (P < 0.05). The nomograph model was constructed using the above 8 risk factors. The AUC of the training set and validation set models were 0.833 (0.790-0.876) and 0.815 (0.748-0.882) respectively, and the Brier values of the training set and verification set were 0.157 and 0.165 respectively, indicating that the calibration of the model was good. Clinical decision curves for both sets confirmed the model's high net benefit, indicating its effectiveness. Age, BMI, smoking history, balance of intake and output, interval between operation and injury, preoperative D-dimer, preoperative FIB, and preoperative PT are identified as significant risk factors for DVT development in TBI patients. The risk prediction model exhibits robust consistency and prediction efficiency, offering valuable insights for medical practitioners in early identification and targeted invervention for high-risk TBI patients prone to DVT.
构建精确的个性化列线图模型，并评估创伤性脑损伤（TBI）患者深静脉血栓形成（DVT）的相关危险因素。回顾性收集2015年1月至2020年1月期间TBI患者的临床资料，并按时间顺序分为模型训练集和模型验证集。采用LASSO回归和多因素逻辑回归分析DVT的危险因素。对模型进行区分度、一致性和临床疗效的后建模评估。LASSO回归结果显示，年龄、BMI、吸烟史、出入量平衡、手术与受伤间隔时间、术前D-二聚体、术前纤维蛋白原（FIB）和术前凝血酶原时间（PT）是TBI术后患者发生DVT的危险因素（P＜0.05）。利用上述8个危险因素构建列线图模型。训练集和验证集模型的AUC分别为0.833（0.790-0.876）和0.815（0.748-0.882），训练集和验证集的Brier值分别为0.157和0.165，表明模型的校准良好。两组的临床决策曲线证实了模型的高净效益，表明其有效性。年龄、BMI、吸烟史、出入量平衡、手术与受伤间隔时间、术前D-二聚体、术前FIB和术前PT被确定为TBI患者发生DVT的重要危险因素。该风险预测模型具有较强的一致性和预测效率，为医务人员早期识别和针对性干预易发生DVT的高危TBI患者提供了有价值的见解。
