急诊创伤并发创伤性凝血病患者预后的风险预测模型构建与验证

Construction and verification of a risk prediction model for the prognosis of patients with emergency trauma complicated by trauma-induced coagulopathy

目的探究急诊创伤并发创伤性凝血病(TIC)患者预后的影响因素,并构建风险预测模型。方法本研究为回顾性研究,收集2021年1月至2023年12月医院收治的246例急性创伤并发TIC患者临床资料,根据7∶3比例将患者分为训练集(n=172)和验证集(n=74),比较患者临床资料,以明确两组的可比性。训练集患者根据28 d内存活情况分为死亡组(n=60)和存活组(n=112),比较两组临床资料,将差异有统计学意义的变量进行Cox回归分析,得到急诊创伤并发TIC患者预后的可能影响因素,基于Cox回归分析结果构建急诊创伤并发TIC患者预后的风险预测模型,并对模型预测效能进行内部验证。结果训练集和验证集患者临床资料差异无统计学意义(P>0.05)。训练集中死亡组患者入院时损伤严重程度(ISS)评分≥25分、确诊TIC时低体温、未使用氨甲环酸治疗占比高于存活组,外周血血小板、血浆纤维蛋白原水平低于存活组,血清C-反应蛋白、血浆乳酸水平高于存活组,差异有统计学意义(均P<0.05)。Cox回归分析结果显示,入院时ISS评分≥25分、确诊TIC时低体温、血浆纤维蛋白原水平低、血浆乳酸水平高均可能为急诊创伤并发TIC患者死亡的独立危险因素(P<0.05)。绘制决策曲线发现,基于急诊创伤并发TIC患者预后影响因素构建风险预测模型,风险阈值在0.00~1.00范围内,模型临床净获益率始终大于0,最大净获益率为0.332;绘制列线图将风险预测模型可视化,结果显示,模型预测急诊创伤并发TIC患者死亡的C-index为0.820(95%CI 0.776~0.864),模型辨别度良好;校准曲线趋近于理想曲线,提示模型预测概率与实际概率具有良好的一致性;内部验证结果显示,预测模型预测急诊创伤并发TIC患者死亡的曲线下面积为0.958(95%CI 0.918~0.999),已建立的预测模型在验证集同样具有良好的预测效能。结论入院时ISS评分、确诊TIC时低体温情况、血浆纤维蛋白原、血浆乳酸均可能是急诊创伤并发TIC患者预后(生存状况)的影响因素,基于上述因素构建的风险预测模型对患者死亡风险有较好的预测价值。

Objective To explore the prognostic factors of patients with emergency trauma complicated by trauma-induced coagulopathy(TIC)and to construct a risk prediction model.Methods This study was a retrospective study,the clinical data were collected from 246 patients with emergency trauma complicated by TIC admitted to Hangzhou Ninth People′s Hospital from January 2021 to December 2023.According to a 7∶3 ratio,the patients were divided into a training set(n=172)and a validation set(n=74),and the clinical data of patients in the two groups were compared to clarify the comparability.Patients in the training set were divided into a death group(n=60)and a survival group(n=112)based on their survival status within 28 d of TIC diagnosis.The clinical data of the two groups were compared,and the variables with statistically significant differences were analyzed by Cox regression analysis to obtain the possible influencing factors of the prognosis in the patients with emergency trauma complicated by TIC.Based on the results of Cox regression analysis,a risk prediction model for the prognosis of patients with emergency trauma complicated by TIC was constructed,and the prediction efficiency of the model was verified internally.Results The difference in clinical data between the training set and the validation set was not statistically significant(P>0.05).The proportion of patients in the death group with injury severity score(ISS)score≥25 points at admission(all P<0.05),the hypothermia at the diagnosis of TIC,and without tranexamic acid treatment was higher than that in the survival group.Peripheral blood platelet and plasma fibrinogen levels were lower than those in the survival group,while serum C-reactive protein and plasma lactate levels were higher than those in the survival group(all P<0.05).The Cox regression analysis results showed that ISS score≥25 points at admission,the hypothermia at the diagnosis of TIC,low plasma fibrinogen levels and high plasma lactate levels may all be independent risk factors for the mortality in the patients with emergency trauma complicated by TIC(P<0.05).Decision curve found that the risk prediction model based on the prognostic factors of patients with emergency trauma complicated by TIC had a risk threshold in the range of 0.00-1.00,and the clinical net benefit rate of the model was always greater than 0,with a maximum net benefit rate of 0.332.A nomogram was drawn to visualize the risk prediction model.The results showed that the C-index of the model predicting the mortality of patients with emergency trauma complicated by TIC was 0.820(95%CI 0.776-0.864),and the model had good discrimination.The calibration curve approached the ideal curve,indicating good consistency between the predicted probability of the model and the actual probability of occurrence.The internal verification results showed that the area under the curve of the risk prediction model for predicting the mortality in the patients with emergency trauma complicated by TIC was 0.958(95%CI 0.918-0.999),and the established prediction model also had good predictive performance in the validation set.Conclusions The ISS score at admission,the hypothermia at diagnosis of TIC,plasma fibrinogen and plasma lactate may all be influencing factors on the prognosis(survival status)of patients with emergency trauma complicated by TIC.The risk prediction model constructed based on those factors has good predictive value for the mortality risk of the patients.