APACHEⅡ评分和血清白蛋白联合分析在评估急性重症有机磷中毒患者预后中的价值研究

Research on the value of APACHEⅡscore and serum albumin combined analysis in evaluating the prognosis of patients with acute severe organophosphorus poisoning

目的:探讨急性生理学与慢性健康状况(APACHEⅡ)评分、血清白蛋白(ALB)预测急性重症有机磷中毒(ASOPP)患者预后的价值,研究两者联合应用在预后判断中的改善效果。方法:选用2015年12月—2019年12月期间我院ICU住院治疗的47例ASOPP患者,根据救治结果将患者分为死亡组13例和存活组34例。比较两组患者APACHEⅡ评分、ALB的差异,分析APACHEⅡ评分、ALB及预后的相关性。基于Logistic回归分析建立APACHEⅡ评分、ALB联合预测模型,绘制两者联合及单项预测的受试者工作特征(ROC)曲线,分析各曲线的敏感度、特异度、最佳截断值、曲线下面积(AUC)。利用DeLong方法、净重新分类指数(NRI)、整合鉴别改善指数(IDI)3种方法来评价APACHEⅡ评分联合ALB预测预后的改善效果。结果:不同预后两组患者APACHEⅡ评分、ALB比较差异有统计学意义(P<0.05),Spearman相关性分析显示APACHEⅡ评分与预后呈强相关(r=0.634,P<0.01),ALB与预后呈中等程度相关(r=-0.582,P<0.01)。APACHEⅡ评分对ASOPP患者死亡预测的AUC为0.908(P<0.01,95%CI:0.788~0.973),21.0分为最佳截断值,其敏感度为92.31%,特异度为79.41%;ALB对患者死亡预测的AUC为0.876(P<0.01,95%CI:0.747~0.954),32.3g/L为最佳截断值,其敏感度为69.23%,特异度为94.12%;APACHEⅡ评分联合ALB预测死亡的AUC为0.957(P<0.01,95%CI:0.854-0.995),敏感度为100.00%,特异度为82.35%。虽然APACHEⅡ评分联合ALB预测的AUC与APACHEⅡ评分的AUC比较差异无统计学意义(P>0.05),但联合预测相对APACHEⅡ评分重新分类的连续性NRI为0.940(95%CI:0.423~1.457,P<0.01),IDI为0.110(95%CI:0.012~0.208,P=0.027),重新分类正确的比例和整体判别能力有一定的提高,具有统计学意义。结论:入院当天的APACHEⅡ评分、ALB测定结果均可用于判断ASOPP患者预后病死率,ALB可提高APACHEⅡ评分预测患者预后的准确性,且ALB在临床上易获取,值得临床推广。

Objective:To explore the value of acute physiology and chronic heath(APACHEⅡ)score and serum albumin(ALB)in predicting the prognosis of patients with acute severe organophosphorus poisoning(ASOPP),and to study the improvement effect of the combination of the two in prognosis judgment.Methods:Forty-seven patients with ASOPP who were treated in ICU of our hospital from December 2015to December 2019 were selected.According to the treatment results,the patients were divided into death group(13cases)and survival group(34cases).APACHEⅡscore and ALB were compared between the two groups,and the correlation between APACHEⅡscore,ALB and prognosis were analyzed.The new combined prediction model was established by APACHEⅡscore and ALB based on Binary Logistic regression analysis.The ROC curves of APACHEⅡscore,ALB and the new prediction model were plotted to analyze the area under the curve(AUC),sensitivity,specificity and optimum cutoff value.The new prediction model was compared with APACHEⅡscore by the methods of DeLong,net reclassification improvement(NRI)and integrated discrimination improvement(IDI)to calculate the improvement.Results:There were significant differences in APACHEⅡscore and ALB between the two groups with different prognosis(P<0.05).Spearman correlation analysis showed that APACHEⅡscore was strongly correlated with prognosis(r=0.634,P<0.01),ALB was moderately correlated with prognosis(r=-0.582,P<0.01).The AUC of APACHEⅡscore was 0.908(P<0.01,95%CI:0.788-0.973),the optimum cutoff value was 21.0,the sensitivity was 92.31%,the specificity was 79.41%.The AUC of ALB was 0.876(P<0.01,95%CI:0.747-0.954),the optimum cutoff value was 32.3g/l,the sensitivity was 69.23%,the specificity was 94.12%.The AUC of the combined prediction model of APACHEⅡscore and ALB was 0.957(P<0.01,95%CI:0.854-0.995),the sensitivity was 100.00%,and the specificity was 82.35%.Although the AUC of combined prediction was not significantly higher than that of APACHEⅡscore(P>0.05).Compared with APACHEⅡscore,the continuous NRI of combined prediction was 0.940(95%CI:0.423-1.457,P<0.01),and the IDI was 0.110(95%CI:0.012-0.208,P=0.027),respectively.Conclusion:The results of APACHEⅡscore and ALB can be used to predict the prognosis and mortality of patients with ASOPP.ALB can improve the accuracy of APACHEⅡscore in predicting the prognosis of patients,and ALB is easy to obtain in cliniccal practice,which is worthy of clinical promotion.