按柏林新标准分析急诊脓毒症患者发生急性呼吸窘迫综合征的危险因素

Analysis of the risk factors of acute respiratory distress syndrome of Berlin new definition in patients with sepsis in emergency department

目的：探讨急诊脓毒症患者发生急性呼吸窘迫综合征（ARDS）的早期危险因素。方法回顾性分析中国医科大学附属第一医院急诊科收治的312例脓毒症患者，采用柏林新标准，根据是否发展至ARDS分为ARDS组和非ARDS组。收集两组患者的年龄、性别、生命体征、实验室数据、基础疾病、急诊脓毒症死亡风险评分（MEDS）、肺损伤预测评分（LIPS）等指标，对各项因素进行单因素分析；将单因素分析中有统计学意义的变量进行多因素logistic回归分析；绘制受试者工作特征曲线（ROC），评估各指标对ARDS的预测价值。结果312例脓毒症患者中35例发生ARDS，发生率为11.2%。35例ARDS患者中轻度ARDS 10例，中度ARDS 18例，重度ARDS 7例。单因素分析显示：年龄（t＝-2.134，P＝0.035）、氧合指数（t＝-4.245，P＝0.001）、乳酸值（Lac，t＝6.245，P＜0.001）、血管活性药物（χ2＝4.261，P＝0.026）、休克（χ2＝4.386，P＝0.021）、MEDS（t＝4.021， P＝0.045）、LIPS（t＝5.569，P＜0.001）、肺部感染（χ2＝4.289，P＝0.025）、机械通气（χ2＝6.245，P＝0.001）与ARDS的发生有关。不同Lac水平患者ARDS发生率不同，低水平Lac（＜2.0 mmol/L）者ARDS发生率为5.00%（3/16），中水平Lac（2.0～3.9 mmol/L）者ARDS发生率为9.46%（14/148），高水平Lac（≥4.0 mmol/L）者ARDS发生率为17.31%（18/104）。进一步行多因素logistic回归分析显示：LIPS〔优势比（OR）＝5.124，95%可信区间（95%CI）＝3.642～10.153，P＝0.002〕、 Lac（OR＝18.180，95%CI＝7.677～32.989，P＜0.001）为发生ARDS的独立危险因素。ROC曲线下面积（AUC）显示，LIPS、 Lac预测ARDS发生的价值较大，LIPS的AUC为0.725，截断值为7分，LIPS≥7分时的敏感度为71.0%，特异度为75.6%；Lac的AUC为0.793，截断值为4.2 mmol/L， Lac≥4.2 mmol/L时的敏感度为72.1%，特异度为81.9%。结论 LIPS、Lac是急诊脓毒症患者发展为ARDS的高危因素，可以为临床早期预测ARDS的发生提供参考。

Objective To discuss the risk factors of acute respiratory distress syndrome （ARDS） in patients with sepsis in emergency department.Methods 312 patients with sepsis admitted to Department of Emergency of China Medical University Affiliated First Hospital were retrospectively analyzed, and they were divided into two groups according to development of ARDS, which was defined according to the Berlin new definition. The age, gender, vital signs, laboratory results, underlying disease, the mortality in emergency department sepsis （MEDS） score and lung injury prediction score （LIPS） were collected. Univariate analysis was done for each parameter. Statistical significance results were evaluated by multivariate logistic regression analysis. Receiver operating characteristic （ROC） curve was plotted to analyze the predictive value of the parameter for ARDS.Results The incidence of sepsis-related ARDS was 11.2% （35/312）. Within 35 cases of ARDS, there were 10 cases of mild ARDS, 18 cases of moderate ARDS, and 7 cases of severe ARDS. Univariate analysis showed that age （t = -2.134,P = 0.035）, oxygenation index （t = -4.245, P = 0.001）, arterial lactate （Lac,t = 6.245,P〈 0.001）, drugs for vascular diseases （χ2 = 4.261,P = 0.026）, shock （χ2 = 4.386,P = 0.021）, MEDS （t = 4.021,P = 0.045）, LIPS （t = 5.569,P〈 0.001）, lung infections （χ2 = 4.289, P = 0.025）, and mechanical ventilation （χ2 = 6.245,P = 0.001） were related to ARDS. The incidence of ARDS was different in different levels of Lac, which was 5.00% （3/16） at low level of Lac （〈 2.0 mmol/L）,9.46% （14/148） at middle level of Lac （2.0-3.9 mmol/L） and 17.31% （18/104） at high level of Lac （≥4.0 mmol/L）. It was shown by＆amp;nbsp;multivariate logistic regression analysis that LIPS [odds ratio （OR） = 5.124, 95% confidence interval （95%CI） =3.642-10.153,P = 0.002], Lac （OR = 18.180, 95%CI = 7.677-32.989,P〈 0.001） were independent risk factors for ARDS. It was shown by area under ROC （AUC） that the predictive value of LIPS and Lac in ARDS occurrence was significant. AUC of LIPS was 0.725, the cut-off value was 7, when LIPS≥7, the sensitivity was 71.0%, specificity was 75.6%. AUC of Lac was 0.793, the cut-off value was 4.2 mmol/L, when Lac≥ 4.2 mmol/L, the sensitivity was 72.1%, and specificity was 81.9%.Conclusion LIPS and Lac are independent risk factors of ARDS in patients with sepsis in emergency department, which may be a reference for the early clinical diagnosis of ARDS.