尿百草枯浓度联合中毒时间及肌酐清除率对急性百草枯中毒患者预后的评估价值

Prognostic value of urine paraquat concentrations combined with poisoning time and creatinine clearance rate ;on prognosis in patients with acute paraquat poisoning

目的：探讨尿百草枯（PQ）浓度联合中毒时间及肌酐清除率（CCr）对急性百草枯中毒（APP）患者预后的评估价值。方法采用回顾性病例对照研究方法，分析2014年3月至2016年5月中国医科大学附属盛京医院急诊科收治的96例APP患者的临床资料，收集患者性别、年龄、体重、尿PQ浓度（半定量比色法检测）、中毒时间（服毒至尿检时间）及CCr，计算中毒指数（中毒指数＝尿PQ浓度×中毒时间/CCr）和简化中毒指数（简化中毒指数＝尿PQ浓度×中毒时间）；记录患者院内死亡情况并电话随访，按中毒后2个月内预后分为死亡组和存活组。比较两组尿PQ浓度、中毒指数及简化中毒指数的差异；采用二分类logistic回归分析影响预后的危险因素；用受试者工作特征曲线（ROC）和诊断性试验分析各指标对预后的评估价值。结果与存活组比较，死亡组患者尿PQ浓度〔mg/L：30.00（10.00，100.00）比10.00（3.00，10.00）〕、中毒指数〔mg·h^-1·μmol^-1：12.72（1.86，33.75）比0.56（0.18，1.12）〕及简化中毒指数〔mg·h^-1·L^-1：600.00（150.00，1000.00）比60.00（18.00，120.00）〕均明显增高（均P＜0.01）。Logistic回归分析结果显示，尿PQ浓度〔优势比（OR）＝1.046，95%可信区间（95%CI）＝1.006～1.087，P＝0.022〕和中毒指数（OR＝1.353，95%CI＝0.029～1.815， P＝0.031）是影响APP患者预后的独立危险因素。ROC曲线及诊断性试验分析结果显示，中毒指数评估APP患者预后的ROC曲线下面积（AUC）最大（为0.902），最佳临界值＞1.23mg·h^-1·μmol^-1时，预测死亡的敏感度为90.91%，特异度为73.08%；尿PQ浓度评估预后的AUC为0.759，最佳临界值＞20.00mg/L时，敏感度为63.64%，特异度为76.92%；简化中毒指数评估预后的AUC为0.846，最佳临界值＞135.00mg·h^-1·L^-1时，敏感度为81.82%，特异度为76.92%。结论通过尿PQ浓度联合中毒时间及CCr计算所得的中毒指数对APP患者预后具有评估价值，且中毒指数较单用尿PQ浓度的预测价值更高。

Objective To evaluate the prognostic value of urine paraquat （PQ） concentrations combined with poisoning time and creatinine clearance rate （CCr） on prognosis of patients with acute paraquat poisoning （APP）. Methods A retrospective case control study was conducted. Clinical data of 96 patients with APP admitted to Department of Emergency of Shengjing Hospital of China Medical University from March 2014 to May 2016 were analyzed. The gender, age, body weight, urine PQ concentrations （determined by semi-quantitative colorimetric method）, poisoning time （time from oral poison to urine detection） and CCr of patients were collected, and poisoning index （poisoning index = urine PQ concentrations × poisoning time/CCr） and simplified poisoning index （simplified poisoning index = urine PQ concentrations × poisoning time） were calculated. The patients were divided into death group and survival group according to 2-month outcome after poisoned with clinical data and telephone follow-up. The urine PQ concentrations, poisoning index, and simplified poisoning index between the two groups were compared. Binary classification logistic regression was used to analyze the risk factors affecting prognosis. Receiver-operating characteristic curve （ROC） and diagnostic test were used to analyze the prognostic value of the parameters. Results Compared with survival group, the urine PQ concentrations [mg/L： 30.00 （10.00, 100.00） vs. 10.00 （3.00, 10.00）], poisoning index [mg·h^-1·μmol^-1： 12.72 （1.86, 33.75） vs. 0.56 （0.18, 1.12）], and simplified poisoning index [mg·h^-1·L^-1： 600.00 （150.00, 1 000.00） vs. 60.00 （18.00, 120.00）] in death group were significantly increased （all P 〈 0.01）. It was shown by logistic regression analysis that both urine PQ concentrations [odds ratio （OR） = 1.046, 95% confidence interval （95%CI） = 1.006-1.087, P = 0.022] and poisoning index （OR = 1.353, 95%CI = 0.029-1.815, P = 0.031） were independent risk factors affecting the prognosis of patients with APP. It was shown by ROC curve and diagnostic test that the poisoning index had greater area under ROC curve （AUC was 0.902） for evaluating the prognosis of patients with APP. When the best cut-off value was greater than 1.23 mg·h^-1·μmol^-1, the sensitivity was 90.91%, and the specificity was 73.08%. The AUC of urine PQ concentrations for evaluating the prognosis was 0.759. When the best cut-off value was greater than 20.00 mg/L, the sensitivity was 63.64%, and the specificity was 76.92%. The AUC of simplified poisoning index for evaluating the prognosis was 0.846. When the best cut-off value was greater than 135.00 mg·h^-1·L^-1, the sensitivity was 81.82%, and the specificity was 76.92%. Conclusion The poisoning index calculated with urine PQ concentrations combined with poisoning time and CCr has prognostic value for prognosis of APP patients, and the prognostic value of poisoning index is greater than that of the urine PQ concentrations alone.