ICU患者医院感染的死亡危险因素分析：2009年至2015年864例病例回顾

摘要

目的：探讨重症加强治疗病房（ICU）医院感染患者的死亡危险因素，以指导临床医师采取有效控制措施。方法采用回顾性队列研究方法，分析2009年6月至2015年12月河北医科大学附属衡水市哈励逊国际和平医院ICU医院感染患者的相关数据。排除非首次入住ICU、ICU住院时间＜48h、入住ICU48h内未进行首次病原学筛查、病原学资料不完整者。记录患者的性别、年龄、诊断、ICU住院时间、侵入性操作、营养状况、入ICU24h内急性生理学与慢性健康状况评分系统Ⅱ（APACHEⅡ）评分和序贯器官衰竭评分（SOFA）、感染病原体及耐药情况，以及医院感染7d的血降钙素原（PCT）水平。用二分类logistic回归分析医院感染患者的死亡危险因素；绘制受试者工作特征曲线（ROC），评估各危险因素对医院感染患者预后的预测价值。结果共864例患者发生医院感染，男性占54.75%，平均年龄（63.56±15.80）岁；存活732例（占84.72%）、死亡132例（占15.28%）。与存活组比较，死亡组患者年龄大（岁：65.47±15.32比58.15±13.27），紧急气管插管（32.58%比22.81%）、深静脉置管（83.33%比63.25%）、多重耐药菌感染（65.91%比33.20%）的比例大，ICU住院时间（d：13.56±4.29比10.29±4.32）和昏迷持续时间（d：7.36±2.46比5.48±2.14）长，白蛋白低（g/L：23.64±8.47比26.36±12.84），APACHE Ⅱ（分：19.28±5.16比17.56±5.62）、SOFA 评分（分：8.55±1.34比6.43±2.65）高，PCT高（μg/L：3.06±1.36比2.53±0.87，均P＜0.05）；而性别构成比、泌尿道置管比例差异无统计学意义（均P＞0.05）；两组前3位感染部位均为下呼吸道、泌尿道和血流感染。Logistic回归分析显示， ICU住院时间〔优势比（OR）＝2.309，95%可信区间（95%CI）＝1.231～3.473，P＝0.002〕、APACHEⅡ评分（OR＝1.683，95%CI＝1.002～9.376，P＝0.000）、SOFA 评分（OR＝2.060，95%CI＝1.208～14.309，P＝0.041）、PCT（OR＝2.090，95%CI＝1.706～13.098，P＝0.004）和多重耐药菌感染（OR＝5.245，95%CI＝2.213～35.098， P＝0.027）是医院感染患者死亡的独立危险因素。ICU住院时间、APACHEⅡ评分、SOFA评分和PCT对医院感染死亡风险预测的ROC曲线下面积（AUC）分别为0.854、0.738、0.786和0.849，最佳截断值分别为16.50d、22.45分、6.37分和3.38μg/L，敏感度分别为83.6%、90.0%、81.1%和89.6%，特异度分别为70.3%、75.6%、71.3%和85.4%。结论长期入住ICU、继发脓毒症和多器官功能障碍是ICU医院感染患者死亡的主要原因， ICU住院时间、APACHEⅡ评分、SOFA评分和PCT对医院感染死亡风险均有较好的预测价值。%Objective To investigate the mortality risk factors of nosocomial infection patients in intensive care unit (ICU), and to guide clinicians to take effective control measures. Methods A retrospectively cohort study was conducted. The relevant information of patients with nosocomial infection treated in ICU of Hengshui Harrison International Peace Hospital Affiliated to Hebei Medical University from June 2009 to December 2015 was analyzed. The patients who admitted to ICU again, with length of ICU stay less than 48 hours, without first etiology of screening within 48 hours of ICU admission, or without complete pathogenic information were excluded. The gender, age, diagnosis, length of ICU stay, invasive operation, nutritional status, acute physiology and chronic health evaluation Ⅱ (APACHEⅡ) score, sequential organ failure assessment (SOFA) score, distribution and drug resistance of the pathogens, and procalcitonin (PCT) levels at 7 days after nosocomial infection were recorded. The risk factors leading to death in patients with nosocomial infection were analyzed by logistic regression, and the receiver operating characteristic curve (ROC) was drawn to evaluate the predictive value of all risk factors on the outcome of patients with nosocomial infection. Results In 864 enrolled patients with male of 54.75% and mean age of (63.50±15.80) years, 732 (84.72%) patients survived and 132 (15.28%) died. Compared with survivors, the non-survivors had higher age (years: 65.47±15.32 vs. 58.15±13.27), incidence of urgent trachea intubation (32.58% vs. 22.81%), deep venous catheterization (83.33% vs. 63.25%), and multiple drug-resistant infection (65.91% vs. 33.20%), longer length of ICU stay (days: 13.56±4.29 vs. 10.29±4.32) and duration of coma (days: 7.36±2.46 vs. 5.48±2.14), lower albumin (g/L: 23.64±8.47 vs. 26.36±12.84), higher APACHEⅡ score (19.28±5.16 vs. 17.56±5.62), SOFA score (8.55±1.34 vs. 6.43±2.65), and PCT (μg/L: 3.06±1.36 vs. 2.53±0.87, all P < 0.05). There was no significant difference in gender and urinary tract catheterization between survivors and non-survivors (both P > 0.05). The low respiratory tract was the most common site of infection followed by urinary tract and bloodstream in both groups. It was shown by logistic regression analysis that prolonged ICU stay [odds ratio (OR) = 2.039, 95% confidence interval (95%CI) = 1.231-3.473, P = 0.002], APACHEⅡ score (OR = 1.683, 95%CI= 1.002-9.376, P = 0.000), SOFA score (OR = 2.060, 95%CI = 1.208 -14.309, P = 0.041), PCT (OR = 2.090, 95%CI = 1.706-13.098, P = 0.004), and multi-drug resistant pathogens infection (OR = 5.245, 95%CI = 2.213-35.098, P = 0.027) were independent risk factors for ICU mortality in patients with nosocomial infection. The area under ROC curve (AUC) of length of ICU stay, APACHEⅡ score, SOFA score, and PCT level for predicting death of nosocomial infection patients was 0.854, 0.738, 0.786, and 0.849, respectively, the best cut-off value was 16.50 days, 22.45, 6.37 and 3.38 μg/L, respectively, the sensitivity was 83.6%, 90.0%, 81.1%, and 89.6%, and the specificity was 70.3%, 75.6%, 71.3%, and 85.4%, respectively. Conclusions Prol onged ICU stay, nosocomial infection with secondary sepsis and multiple organ dysfunction syndrome were the leading causes of death for nosocomial infection patients in ICU. Prolonged ICU stay, APACHE Ⅱ score, SOFA score, and PCT level could effectively predict death risks for nosocomial infection patients.