百草枯或敌草快中毒患者30 d预后模型的建立  

Establishment of a 30-day prognostic model of paraquat or diquat poisoning patients

作　　者：樊丹丹 高烨[1] FAN Dandan;GAO Ye(Department of Emergency Center,The First Affiliated Hospital of Xi’an Jiaotong University,Xi’an 710061,China)

机构地区：[1]西安交通大学第一附属医院急诊中心,陕西西安710061

出　　处：《西安交通大学学报（医学版）》2024年第5期837-843,共7页Journal of Xi’an Jiaotong University（Medical Sciences）

基　　金：国家自然科学基金资助项目(No.81601674)。

摘　　要：目的探讨百草枯(PQ)或敌草快(DQ)中毒患者30 d预后的影响因素,并建立二分类Logistic回归方程,以更好地判断患者的预后并提供及时有效的治疗方法。方法选取2010年1月到2021年10月西安交通大学第一及第二附属医院急诊科符合纳排标准的PQ与DQ中毒患者(109例,其中PQ中毒患者82例,DQ中毒患者27例);收集病历资料并电话随访服毒30 d时的预后,分为死亡组(n=46)及存活组(n=63);比较两组一般资料、入院1 h的血生化指标、是否血液灌流(HP)、是否HP联合持续性肾脏替代治疗(CRRT)。分析两组的独立危险因素,建立二分类Logistic回归方程模型,并对该模型进行验证。结果年龄、服毒量、乳酸值及是否HP为30 d死亡的独立危险因素(P<0.05);回归方程Y=Logit(P)=-3.588+0.05×年龄+0.031×服毒量+0.467×乳酸-1.915×是否HP(未HP赋值为0,有HP赋值为1)。该模型经Hosmer-Lemeshow拟合优度检验,χ2=7.645,自由度=8,P=0.469,该模型拟合度较好。该模型的灵敏度为86.96%,特异度为88.89%,AUC为0.93,95%CI为0.88~0.98,P<0.001。该模型校准度良好,C-index为0.931,有良好的区分度;临床决策曲线(DCA)显示该模型有较好的标准化净效益。结论本研究所建立的预测模型可以科学、有效地对PQ或DQ中毒患者进行30 d预后的预测,能指导临床医师对该类患者的治疗。Objective To explore the factors for 30-day prognosis of patients with paraquat(PQ)or diquat(DQ)poisoning so as to establish a binary Logistic equation to judge the prognosis and provide effective treatment.Methods We recruited 109 patients with PQ(n=82)or DQ(n=27)poisoning who met the inclusion and exclusion criteria in The First and The Second Affiliated Hospital of Xi’an Jiaotong University from January 2010 to October 2021,and obtained their prognosis of 30 days by medical records and telephone.The death group(n=46)and the survival group(n=63)were divided.The clinical data,biochemical data at 1 h after admission,administration of hemoperfusion(HP)and whether HP combined with continuous renal replacement treatment(CRRT)were compared between the two groups.The independent risk factors in the two groups were analyzed,and the binary Logistic regression equation model was established and verified.Results Age,dosage,lactic acid(Lac),and administration of HP were independent risk factors for 30 days mortality(P<0.05).Regression equation:Y=Logit(P)=-3.588+0.05×age+0.031×dosage+0.467×Lac-1.915×presence of HP(no HP assigned as 0,with HP assigned as 1).The model was tested by Hosmer-Lemeshow goodness of fit:χ2=7.645,freedom=8,P=0.469,which indicated that the model had a good fit.The sensitivity of the model was 86.96%,the specificity was 88.89%,area under the curve(AUC)was 0.93,the 95%confidence interval(CI)was 0.88-0.98,and the P value was<0.001.The calibration degree of the model was good,and the C-index was 0.931.The clinical decision curve analysis(DCA)showed that the model had a good standardized net benefit.Conclusion The model can scientifically and effectively predict the 30-day prognosis of patients with PQ or DQ poisoning and guide clinicians.

关 键 词：百草枯(PQ) 敌草快(DQ) 近期预后 预测模型 血液灌流(HP) 

分 类 号：R595.4[医药卫生—内科学]