围手术期外科危重患者无创通气失败的 多因素回归分析

Failure factors of noninvasive positive pressure ventilation for perioperative critical ill patients in surgery intensive care unit

目的：分析外科危重患者围手术期无创正压通气（NPPV）治疗失败的独立危险因素，以期指导围手术期危重患者NPPV的临床应用。方法通过电子病历检索系统，回顾性收集2004年1月至2016年1月北京大学第一医院外科重症监护病房收治的围手术期因急性呼吸功能不全接受NPPV治疗患者的临床资料，按NPPV后是否行有创通气将患者分为成功组与失败组。记录患者一般情况、手术类型、急性生理学与慢性健康状况评分系统Ⅱ（A PACHEⅡ）评分、急性呼吸功能不全的原因和分型（肺源性、心源性、严重全身感染引起的过度通气、其他；Ⅰ型呼衰、Ⅱ型呼衰、单纯高碳酸血症等）、NPPV时呼吸机参数、NPPV前及治疗1h的生命体征及血气分析结果、NPPV相关并发症及预后指标。结果最终纳入189例患者，成功组109例（57.7%），失败组80例（42.3%）。与成功组比较，失败组患者年龄大（岁：69.14±14.24比62.33±16.03）、APACHEⅡ评分高（分：16.74±8.00比11.76±5.53），术后并发肺部感染的比例大（45.0%比24.8%），NPPV时最高压力支持（PS）水平高〔cmH2O（1cmH2O＝0.098kPa）：12.38±2.08比11.29±2.18〕、最高吸入氧浓度高（FiO2：0.666±0.201比0.506±0.166），NPPV前后pH差值小（0.01±0.07比0.03±0.06），NPPV前后氧合指数（PaO2/FiO2）差值小〔mmHg（1mmHg＝0.133kPa）：-10.53±57.01比-59.47±71.14〕，差异均有统计学意义（均P＜0.05）；两组患者性别、既往史、手术类型、呼吸功能不全分型均无差异。将单因素分析中有统计学意义的指标纳入多因素Logistic回归分析发现，APACHEⅡ评分高〔优势比（OR）＝0.888，95%可信区间（95%CI）＝0.837～0.942，P＜0.001〕、NPPV时所需FiO2过高（OR＝0.955，95%CI＝0.934～0.976，P＜0.001）、NPPV后1hPaO2/FiO2升高不明显（OR＝0.989，95%CI＝0.982～0.995，P＜0.001）、术后并发肺部感染（OR＝0.246，95%CI＝0.129～0.470，P＜0.001）是NPPV治疗失败的独立危险因素。与成功组比较，失败组持续NPPV时间明显缩短〔d：7.5（1.6，21.0）比12.0（4.0，35.5），P＜0.01〕，总住院时间略有延长〔d：34.0（22.0，67.5）比28.0（19.5，42.5）， P＞0.05〕，但重症加强治疗病房（ICU）住院时间明显延长〔d：9.5（6.0，16.0）比5.0（3.0，8.0），P＜0.01〕，住院病死率也明显升高（67.5%比2.8%，P＜0.01）。结论 NPPV治疗围手术期外科危重患者呼吸功能不全安全有效；患者APACHEⅡ评分高、NPPV时所需FiO2高、NPPV后1hPaO2/FiO2升高不明显、术后并发肺部感染是导致NPPV失败的独立危险因素。

Objective To evaluate the independent risk factor of noninvasive positive pressure ventilation （NPPV） treatment failure for perioperative critical ill patients in surgical intensive care unit （SICU）, and guide the clinical application of NPPV in perioperative critical patients. Methods Patients undergoing NPPV due to acute respiratory insufficiency admitted to SICU of Peking University First Hospital from January 2004 to January 2016 were retrospectively analyzed by electronic medical record retrieval system. According to whether invasive ventilation was needed finally or not, perioperative patients treated with NPPV were divided into the success group and failure group. The patients＆#39; perioperative data were recorded including general state, the type of operation, acute physiology and chronic health evaluation Ⅱ （APACHEⅡ） score, the cause and the type of acute respiratory insufficiency （pulmonary, cardiogenic, excessive ventilation caused by severe systemic infection, etc.; type Ⅰ respiratory failure, type Ⅱ respiratory failure, simple hypercapnia, etc.）, the related parameters of ventilator during NPPV, vital signs and blood gas analysis before and 1 hour after NPPV, NPPV related complications and prognostic indicators. Results 189 patients were enrolled, 109 patients in success group （57.7%）, and 80 in failure group （42.3%）. Compared with success group, patients in failure group were older （years： 69.14±14.24 vs. 62.33±16.03）, had higher APACHE Ⅱ score （16.74±8.00 vs. 11.76±5.53） and incidence of pulmonary infection after operation （45.0% vs. 24.8%）, and higher pressure support （PS） [cmH2O （1 cmH2O = 0.098 kPa）： 12.38±2.08 vs. 11.29±2.18] and fraction of inspired oxygen （FiO2： 0.666±0.201 vs. 0.506±0.166） during NPPV, lower differences in pH value （0.01±0.07 vs. 0.03±0.06） and oxygenation index [PaO2/FiO2 （mmHg, 1 mmHg = 0.133 kPa）： -10.53±57.01 vs. -59.47±71.14] before and after NPPV with statistically significant differences （all P 〈 0.05）. There was no significantly difference in gender, past medical history, the type of operation, and the type of acute respiratory insufficiency in two groups. The parameters showed significant differences in univariate analysis were collected for multivariate Logistic regression analysis, which found that higher APACHE Ⅱ score [odds ratio （OR） = 0.888, 95% confidence interval （95%CI） = 0.837-0.942, P 〈 0.001], higher FiO2 setting during NPPV （OR = 0.955, 95%CI = 0.934-0.976, P 〈 0.001）, lack of improvement of PaO2/FiO2 at 1 hour after NPPV （OR = 0.989, 95%CI = 0.982-0.995, P 〈 0.001）, and pulmonary infection after operation （OR = 0.246, 95%CI = 0.129-0.470, P 〈 0.001） were the independent risk factors of NPPV treatment failure. Duration of NPPV in the failure group was significantly shorter than that of successful group [days： 7.5 （1.6, 21.0） vs. 12.0 （4.0, 35.5）, P 〈 0.01], total length of hospital stay was slightly prolonged [days： 34.0 （22.0, 67.5） vs. 28.0 （19.5, 42.5）, P 〉 0.05], but the length of intensive care unit （ICU） stay was significantly prolonged [days： 9.5 （6.0, 16.0） vs. 5.0 （3.0, 8.0）, P 〈 0.01], and the in-hospital mortality was significantly increased （67.5% vs. 2.8%, P 〈 0.01）. Conclusions NPPV may be an effective and security method for treatment of acute respiratory insufficiency in perioperative critical patients. The independent risk factors associated with failure of NPPV were higher APACHE Ⅱ score, higher FiO2 during NPPV, lack of improvement PaO2/FiO2 at 1 hour after NPPV, and pulmonary infection after operation.