急性百草枯中毒患者预后影响因素分析

Analysis of prognostic factors in patients with acute paraquat poisoning

目的探索简便、高效辨别急性百草枯中毒患者预后的主要影响因素。方法采用回顾性研究方法,选取2010年1月至2019年12月在广州市第十二人民医院住院治疗的201例急性百草枯中毒患者为研究对象,依据出院时病情及中毒后60 d的随访结果,分为存活组(78例)和死亡组(123例)。比较两组患者基本信息、诊疗方案、入院时实验室检验指标的差异,采用二分类logistic回归分析筛选出可独立预测中毒结局的主要影响因素,受试者工作特征曲线(ROC)判定主要影响因素辨别急性百草枯中毒患者预后的效能。采用t检验、χ^(2)检验。结果存活组男33例、女45例,年龄(30.41±14.04)岁;死亡组男61例、女62例,年龄(31.86±14.15)岁。存活组患者百草枯摄入量[(20.95±23.14)ml]、出现呕吐症状比例[70.51%(55/78)]、白细胞计数(WBC)[(12.39±4.93)×10^(9)/L]、血红蛋白(Hb)[(136.56±18.04)g/L]、凝血酶原时间(PT)[(12.49±3.52)s]、谷丙转氨酶(ALT)[(18.30±10.76)U/L]、谷草转氨酶(AST)[(23.44±11.85)U/L]、血糖[(7.09±2.21)mmol/L]、尿素[(4.96±3.48)mmol/L]、肌酐[(98.03±79.33)μmol/L]、肌酸激酶(CK)[(187.78±178.06)U/L]、肌酸激酶同工酶(CKMB)[(18.88±15.50)U/L]、乳酸脱氢酶(LDH)[(190.92±60.50)U/L]及尿蛋白秩次(60.65)均低于死亡组[(62.07±50.59)ml、83.74%(103/123)、(23.15±12.05)×10^(9)/L、(142.32±20.64)g/L、(13.50±3.13)s、(94.04±144.42)U/L、(130.94±184.19)U/L、(8.43±3.49)mmol/L、(9.55±8.17)mmol/L、(249.98±196.40)μmol/L、(336.36±261.96)U/L、(43.91±43.74)U/L、(346.01±196.50)U/L、126.59],差异均有统计学意义(均P<0.05)。存活组患者动脉血二氧化碳分压(PaCO_(2))[(4.33±0.58)kPa]、血碳酸氢根浓度(HCO_(3)^(-))[(21.12±2.36)mmol/L]均高于死亡组[(3.81±1.25)kPa、(16.85±4.75)mmol/L],差异均有统计学意义(均P<0.05)。二分类logistic回归分析显示,百草枯摄入量(OR=1.022,P<0.05)、HCO_(3)^(-)(OR=0.763,P<0.05)、AST(OR=1.059,P<0.05)、血糖(OR=1.235,P<0.05)、尿蛋白(OR=1.878,P<0.05)是影响急性百草枯中毒患者预后的关键因素。进一步ROC分析显示,5个指标在预测急性百草枯中毒患者预后中均具有较高效能,其中预测价值最高的指标为AST[曲线下面积(AUC)为0.87,最佳诊断值为35.30 U/L],灵敏度最高的指标为尿蛋白(灵敏度0.94),特异度最高的指标为HCO_(3)^(-)和AST(特异度均为0.96);阳性预测值最高的指标为HCO_(3)^(-)和AST(阳性预测值均为0.96),阴性预测值最高的指标为尿蛋白(阴性预测值为0.86)。结论百草枯摄入量、HCO_(3)^(-)、AST、血糖、尿蛋白是预测急性百草枯中毒结局的主要影响因素,具有较高的预测效能和灵敏度、特异度,可用于独立辨别急性百草枯中毒患者预后。

Objective To explore the main factors affecting the prognosis in patients with acute paraquat poisoning simply and efficiently.Methods Using the method of retrospective study,201 patients with acute paraquat poisoning who were hospitalized in Guangzhou Twelfth People's Hospital from January 2010 to December 2019 were selected as the research objects,and according to the conditions at discharge and the follow-up results 60 days after poisoning,the subjects were divided into a survival group(78 cases)and a death group(123 cases).The differences of basic information,diagnosis and treatment plans,and laboratory test indicators at admission between the two groups were compared.The binary logistic regression analysis was used to screen out the main influencing factors that could independently predict the outcome of poisoning.The receiver operating characteristic curve(ROC)was used to determine the effectiveness of the main influencing factors in identifying the prognosis in patients with acute paraquat poisoning.t test andχ^(2) test were used.Results In the survival group,there were 33 males and 45 females,aged(30.41±14.04)years.In the death group,there were 61 males and 62 females,aged(31.86±14.15)years.In the survival group,the paraquat intake[(20.95±23.14)ml],proportion of vomiting symptom[70.51%(55/78)],white blood cell count(WBC)[(12.39±4.93)×10^(9)/L],hemoglobin(Hb)[(136.56±18.04)g/L],prothrombin time(PT)[(12.49±3.52)s],alanine transaminase(ALT)[(18.30±10.76)U/L],aspartate aminotransferase(AST)[(23.44±11.85)U/L],blood glucose[(7.09±2.21)mmol/L],urea nitrogen[(4.96±3.48)mmol/L],creatinine[(98.03±79.33)μmol/L],creatine kinase(CK)[(187.78±178.06)U/L],creatine kinase isoenzyme(CKMB)[(18.88±15.50)U/L],lactate dehydrogenase(LDH)[(190.92±60.50)U/L],and urine protein average ranks(60.65)were lower than those in the death group[(62.07±50.59)ml,83.74%(103/123),(23.15±12.05)×10^(9)/L,(142.32±20.64)g/L,(13.50±3.13)s,(94.04±144.42)U/L,(130.94±184.19)U/L,(8.43±3.49)mmol/L,(9.55±8.17)mmol/L,(249.98±196.40)μmol/L,(336.36±261.96)U/L,(43.91±43.74)U/L,(346.01±196.50)U/L,and 126.59],with statistically significant differences(all P<0.05).The blood partial pressure of carbon dioxide(PaCO_(2))[(4.33±0.58)kPa]and blood bicarbonate(HCO_(3)^(-))concentration[(21.12±2.36)mmol/L]in the survival group were higher than those in the death group[(3.81±1.25)kPa and(16.85±4.75)mmol/L],with statistically significant differences(both P<0.05).Binary logistic regression analysis showed that paraquat intake(OR=1.022,P<0.05),HCO_(3)^(-)(OR=0.763,P<0.05),AST(OR=1.059,P<0.05),blood glucose(OR=1.235,P<0.05),and urinary protein(OR=1.878,P<0.05)were key indicators of prognosis in patients with acute paraquat poisoning.Further ROC analysis showed that the five indexes had high efficacies in predicting the prognosis in patients with acute paraquat poisoning,among which AST was the index with the highest predictive value[the area under the curve(AUC)was 0.87,and the optimal diagnostic value was 35.30 U/L],urinary protein was the index with the highest sensitivity(0.94),HCO_(3)^(-)and AST were the indexes with the highest specificity(both 0.96),HCO_(3)^(-)and AST were the indexes with the highest positive predictive value(both 0.96),and urinary protein was the index with the highest negative predictive value(0.86).Conclusions Paraquat intake,HCO_(3)^(-),AST,blood glucose,and urinary protein are the main influencing factors for predicting the outcome of acute paraquat poisoning.They have high predictive efficacy,sensitivity,and specificity,and can be used to independently identify the prognosis in patients with acute paraquat poisoning.