外科ICU患者脱机后序贯经鼻高流量氧疗的特点及治疗失败的危险因素分析

Characteristics and failure risk factors of sequential high-flow nasal cannula oxygen therapy after weaning from invasive ventilation in patients of surgical intensive care unit

目的分析有创机械通气脱机后序贯经鼻高流量氧疗(HFNC)的特点及治疗失败的危险因素.方法选择2016年6月1日至2018年5月31日北京大学人民医院外科重症医学科(ICU)有创机械通气脱机后序贯HFNC治疗的患者.收集患者一般临床资料、呼吸治疗相关参数、呼吸生理相关指标、循环相关指标及结局指标,分析脱机后序贯HFNC的治疗特点.根据HFNC失败与否将患者分为HFNC失败组与HFNC成功组,比较两组各项指标的差异;采用多因素Logistic回归分析HFNC治疗失败的独立危险因素,并对各危险因素和回归模型预估值预测治疗失败的价值进行受试者工作特征曲线(ROC)分析.结果共纳入99例患者,男性61例,中位年龄67.0(57.0,76.0)岁.HFNC初始治疗流速为50(50,60)L/min,吸入氧浓度(FiO2)为0.50(0.40,0.60).有18例患者(18.2%)HFNC治疗失败.与HFNC成功组相比,HFNC失败组序贯器官衰竭评分(SOFA)更高〔分:4(3,5)比2(1,3),P<0.01〕,治疗前血B型利钠肽(BNP)水平更高〔ng/L:647.2(399.2,1 331.3)比127.2(55.2,369.5),P<0.01〕,HFNC治疗30 min呼吸频率(RR)和心率(HR)更快〔RR(次/min):26(22,28)比19(17,21),HR(次/min):105(97,107)比85(77,90),均P<0.01〕,平均动脉压(MAP)更高〔mmHg(1 mmHg=0.133 kPa):104.3(101.7,110.7)比92.3(88.3,97.7),P<0.01〕,氧合指数(PaO2/FiO2)更低〔mmHg:207.3(185.8,402.8)比320.2(226.2,361.5),P<0.05〕.多因素Logistic回归分析显示,SOFA评分〔优势比(OR)=2.818,P=0.022,β=1.036〕、治疗前BNP(OR=1.002,P=0.033,β=0.002)和治疗30 min HR(OR=1.140,P=0.032,β=0.131)为HFNC治疗失败的独立危险因素.ROC曲线分析显示,SOFA、治疗前BNP、治疗30 min HR和Logistic回归模型预估值预测HFNC失败的ROC曲线下面积(AUC)分别为0.840、0.859、0.860和0.962,均有较好的预测价值(均P<0.01).结论脱机后序贯HFNC治疗是ICU常用的氧疗手段之一,但并不是所有患者都能从中获益;SOFA评分、治疗前BNP和治疗30 min HR为外科ICU有创机械通气患者脱机后序贯HFNC治疗失败的独立危险因素,且各独立危险因素及回归模型均可预测HFNC治疗成功与否.

Objective To investigate the characteristics and failure risk factors of sequential high-flow nasal cannula oxygen therapy (HFNC) after weaning from invasive ventilation. Methods The patients who received sequential HFNC after weaning from invasive ventilation admitted to surgical intensive care unit (ICU) of Peking University People's Hospital from June 1st 2016 to May 31st 2018 were retrospectively analyzed. Clinical variables, respiratory therapy parameters, respiratory variables, cardiac variables and outcomes were reviewed and analyzed. Treatment characteristics of HFNC after weaning was analyzed. Patients were divided into HFNC success group and HFNC failure group according to the failure of HFNC, and the differences between the two groups were compared. The independent risk factors of HFNC treatment failure were analyzed by Logistic regression analysis. The value of predictive treatment failure of risk factors and regression models were analyzed by receiver operating characteristic (ROC) curve. Results A total of 99 patients were included, 61 men, and the median age was 67.0 (57.0, 76.0) years old. The medianinitial HFNC flow was 50 (50, 60) L/min, and inspired oxygen concentration (FiO2) was 0.50 (0.40, 0.60). Eighteen patients experienced HFNC failure (18.2%). Compared with the HFNC success group, the sequential organ failure assessment (SOFA) score in the HFNC failure group was higher [4 (3, 5) vs. 2 (1, 3), P < 0.01], B type natriuretic peptide (BNP) before HFNC therapy were significant higher [ng/L: 647.2 (399.2, 1 331.3) vs. 127.2 (55.2, 369.5), P < 0.01], and respiratory frequency (RR) and heart rate (HR) were significant faster, mean arterial pressure (MAP) was significant higher, oxygen index (PaO2/FiO2) was significant lower after 30 minutes HFNC treatment [RR (times/min): 26 (22, 28) vs. 19 (17, 21), HR (bpm): 105 (97, 107) vs. 85 (77, 90), MAP (mmHg, 1 mmHg = 0.133 kPa): 104.3 (101.7, 110.7) vs. 92.3 (88.3, 97.7), PaO2/FiO2 (mmHg): 207.3 (185.8, 402.8) vs. 320.2 (226.2, 361.5), all P < 0.05]. It was shown by multiple Logistic regression analysis that the SOFA score [odds ratio (OR)= 2.818, P = 0.022,β= 1.036], BNP before HFNC treatment (OR = 1.002, P = 0.033,β= 0.002) and HR after HFNC treatment 30 minutes (OR = 1.140, P = 0.032,β= 0.131) were independent risk factors for HFNC treatment failure. It was shown by ROC curve that the area under the ROC curve (AUC) for the prediction of HFNC failure was 0.840, 0.859, 0.860 and 0.962 for SOFA, BNP before HFNC treatment, HR after HFNC treatment 30 minutes, and regression model, all had good forecast values (all P < 0.01). Conclusions HFNC is one of the commonly used oxygen therapy methods in the ICU, but not all patients who are treated as a sequential therapy after invasive mechanical ventilation weaning can benefit from it. SOFA score, BNP before HFNC treatment and HR after 30 minutes HFNC treatment were independent risk factors of HFNC failure. Each independent risk factor and regression model can predict the success of HFNC treatment.