早期累积液体平衡和血电解质水平对重型创伤性脑损伤患者短期预后的影响

Impacts of early cumulative fluid balance and serum electrolyte level on the short-term prognosis of patients with severe traumatic brain injury

目的 观察重型创伤性脑损伤(sTBI)患者早期累积液体平衡和血清电解质水平变化,探讨二者对sTBI患者短期预后的影响。方法 2022年1月—2023年12月郑州大学第一附属医院诊治sTBI患者121例,根据入院28 d内生存情况分为生存组83例和死亡组38例。比较2组入院时急性生理学与慢性健康状况Ⅱ(APACHEⅡ)评分、入院72 h内液体出入量情况(第1、2、3个24 h液体总入量、总出量,48、72 h累积液体平衡,48、72 h液体积累指数)、入院72 h内动脉血气电解质水平(入院时及入院48 h、72 h血钠、血氯水平,入院48、72 h血钠、血氯变化值)。采用多因素logistic回归分析sTBI患者入院28 d内预后不良的影响因素;绘制ROC曲线,评估入院48、72 h影响因素及相应的回归模型预测sTBI患者入院28 d内预后不良的效能;采用Hosmer-Lemeshow拟合优度检验评价模型的校准能力。结果 121例sTBI患者28 d内病死率为31.4%。死亡组APACHEⅡ评分[23.50(22.00,28.00)分]高于生存组[16.00(14.00,19.00)分](Z=7.378,P<0.001)。死亡组第1、2、3个24 h液体总入量,48、72 h累积液体平衡,48、72 h液体积累指数,入院时及入院48、72 h血钠、血氯水平,入院48、72 h血钠、血氯变化值均高于生存组(P<0.05),第1、2、3个24 h液体总出量与生存组比较差异均无统计学意义(P>0.05)。以入院48 h影响因素为自变量,APACHEⅡ评分(OR=1.370,95%CI:1.099~1.707,P=0.005)、48 h液体积累指数(OR=2.463,95%CI:1.336~4.541,P=0.004)、入院48 h血钠变化值(OR=1.235,95%CI:1.045~1.461,P=0.013)是sTBI患者入院28 d内预后不良的影响因素。回归模型1公式为:ln(p/1-p)分别为APACHEⅡ评分、48 h液体积累指数、入院48 h血钠变化值。APACHEⅡ评分、48 h液体积累指数、入院48 h血钠变化值、回归模型1分别以21.00分、0.104、5.5 mmol/L、0.234为最佳截断值,预测sTBI患者入院28 d内预后不良的AUC分别为0.918(95%CI:0.870~0.966,P<0.001)、0.905(95%CI:0.841~0.968,P<0.001)、0.912(95%CI:0.846~0.977,P<0.001)、0.979(95%CI:0.958~1.000,P<0.001),灵敏度分别为84.2%、86.8%、84.2%、97.4%,特异度分别为81.9%、86.7%、92.8%、90.4%。Hosmer-Lemeshow拟合优度检验结果显示,回归模型1拟合效果较好(χ^(2)=2.419,P=0.965)。以入院72 h影响因素为自变量,APACHEⅡ评分(OR=1.362,95%CI:1.102~1.682,P=0.004)、72 h液体积累指数(OR=1.149,95%CI:1.058~1.247,P<0.001)是sTBI患者入院28 d内预后不良的影响因素。回归模型2公式为:ln(p/1-p)(x_(1)、x_(2)分别为APACHEⅡ评分、72 h液体积累指数)。APACHEⅡ评分、72 h液体积累指数、回归模型2分别以21.00分、0.080、0.424为最佳截断值,预测sTBI患者入院28 d内预后不良的AUC分别为0.918(95%CI:0.870~0.966,P<0.001)、0.942(95%CI:0.887~0.997,P<0.001)、0.968(95%CI:0.924~1.000,P<0.001),灵敏度分别为84.2%、94.7%、94.7%,特异度分别为81.9%、92.8%、94.0%。Hosmer-Lemeshow拟合优度检验结果显示,回归模型2拟合效果较好(χ^(2)=6.673,P=0.572)。结论 对早期血流动力学稳定的sTBI患者,入院时APACHEⅡ评分、48 h和72 h液体积累指数高及48 h血钠升高幅度大者短期死亡风险增大;以上述指标构建的回归模型预测sTBI患者短期预后有较高价值。

Objective To observe the changes of early cumulative fluid balance and serum electrolyte level in patients with severe traumatic brain injury(sTBI),and to explore their impacts on the short-term prognosis of sTBI patients.Methods Totally 121 patients with sTBI were treated in the First Affiliated Hospital of Zhengzhou University from January,2022 to December,2023,and were divided into survival group(n=83)and death group(n=38)based on the prognosis within 28 d after admission.The acute physiology and chronic health evaluationⅡ(APACHEⅡ)scores on admission,fluid intake and outflow in 72 h after admission(total fluid intake and outflow in the first,second and third24 h,cumulative fluid balance in 48 and 72 h,fluid accumulation index in 48 and 72 h),arterial blood gas electrolyte levels in 72 h after admission(blood sodium and chlorine levels on admission and in 48 and 72 h after admission,and changes in blood sodium and chlorine levels in 48 and 72 h after admission)were compared between two groups.Multivariate logistic regression analysis was done to identify the influencing factors of poor prognosis in sTBI patients within 28 d after admission.ROC curves were plotted to evaluate the efficiencies of influencing factors in 48 and 72 h after admission and corresponding regression models on predicting poor prognosis of sTBI patients within 28 d after admission.The Hosmer-Lemeshow goodness of fit test was done to evaluate the calibration ability of the model.Results The mortality rate within 28 d in 121 patients with sTBI was 31.4%.The APACHE II score was higher in death group[23.50(22.00,28.00)]than that in survival group[16.00(14.00,19.00)](Z=7.378,P<0.001).The volumes of total fluid intake in the first,second and third 24 h,the cumulative fluid balance in 48 and 72 h,the fluid accumulation indexes in 48 and 72 h,the blood sodium and chlorine levels on admission and in 48 and 72 h after admission,and the changes in blood sodium and chlorine levels in 48 and 72 h after admission were higher in death group than those in survival group(P<0.05),and there were no significant differences in the volumes of total fluid output in the first,second and third 24 h between two groups(P>0.05).Using the influencing factor in 48 h after admission as the independent variable,the APACHEⅡscore(OR=1.370,95%CI:1.099-1.707,P=0.005),48-h fluid accumulation index(OR=2.463,95%CI:1.336-4.541,P=0.004),and 48-h blood sodium change value(OR=1.235,95%CI:1.045-1.461,P=0.013)were the influencing factors of poor prognosis within 28 d after admission in sTBI patients.The formula for regression model 1 was ln(p/(1-p))=9.275+0.31x_(1)+0.902x_(2)+0.211x_(3)(x_(1)-x_(3) represents APACHEⅡscore,48-h fluid accumulation index,and 48-h blood sodium change value on admission,respectively).When the optimal cut-off values of APACHEⅡscore,48-h fluid accumulation index,blood sodium change in 48 h after admission,and regression model 1 were 21.00,0.104,5.5 mmol/L and 0.234,the AUCs for predicting poor prognosis of sTBI patients within 28 d after admission were 0.918(95%CI:0.870-4.966,P<0.001),0.905(95%CI:0.841-0.968,P<0.001),0.912(95%CI:0.846-0.977,P<0.001),and 0.979(95%CI:0.958-1.00,P<0.001).with the sensitivities of84.2%,86.8%,84.2%and 97.4%,and the specificities of 81.9%,86.1%,92.8%and 90.4%,respectively.The results of the Hosmer-Lemeshow goodness of fit test showed that regression model 1 had a good fitting effect(χ^(2)=2.419,P=0.965).Using the influencing factor in 72 h after admission as the independent variable,the APACHEⅡscore(OR=1.362,95%CI:1.102-1.682,P=0.004)and 72-h fluid accumulation index(OR=1.149,95%CI:1.058-1.247.P<0.001)were the influencing factors of poor prognosis within 28 d after admission in sTBI patients.The formula for regression model 2 was ln(p/(1-p))=-0.820+0.309x_(1)+0.139x_(2)(x_(1)-x_(2) represents APACHEⅡscore,and72-h fluid accumulation index,respectively).When the optimal cut-off values of APACHEⅡscore,72-h fluid accumulation index,and regression model 2 were 21.00,0.080 and 2.424,the AUCs for predicting poor prognosis of sTBI patients within 28 d after admission were 0.918(95%CI:0.870-0.966,P<0.0001),0.942(95%CI:0.887-0.997,P<0.001)and 0.968(95%CI:0.924-1.000,P<0.001),with the sensitivities of 84.2%,94.7%and 94.7%,and the specificities of 81.9%,92.8%and 94.0%,respectively.The results of the Hosmer-Lemeshow goodness of fit test showed that regression model 2 had a good fitting effect(χ^(2)=6.673,P=0.572).Conclusion In sTBI patients with stable hemodynamics in the early stage,the high APACHEⅡscore,high 48-and 72-h fluid accumulation indexes,and significant increase in blood sodium in 48 h after admission increase the risk of short-term death,and the regression model based on the above indicators has a high value to the prediction of the short-term prognosis of sTBI patients.