中华急诊医学杂志  2021, Vol. 30 Issue (9): 1106-1112   DOI: 10.3760/cma.j.issn.1671-0282.2021.09.011
术后低氧血症患者高流量氧疗治疗失败的早期预测
刘韬滔1 , 赵沁宇2 , 杜斌3     
1. 北京医院 国家老年医学中心 中国医学科学院老年医学研究院 外科ICU 100730;
2. 澳洲国立大学工程与计算机科学学院,堪培拉2600,澳大利亚;
3. 中国医学科学院 北京协和医学院 北京协和医院内科ICU 100730
摘要: 目的 探讨术后合并低氧血症患者高流量氧疗治疗失败的早期预测指标。方法 回顾性队列研究重症监护医疗信息数据库Ⅳ(MIMIC-Ⅳ)中术后撤机时合并低氧血症(100 mmHg < PaO2/FiO2≤300 mmHg)并接受高流量吸氧治疗的成人患者,根据治疗48 h是否再次气管插管分为治疗成功组和失败组。采用机器算法XGBoost模型分析患者撤机后48 h再插管危险因素。根据危险因素制定高流量氧疗治疗失败预测指标,记录上述指标从撤机前至撤机后48 h的动态改变。采用t检验比较撤机成功和失败患者在各时间段预测指标差异,比较各时间段预测指标与基线数据差异。计算撤机前后4 h指标预测48 h再插管的准确性,计算受试者工作特征曲线(ROC)面积,与呼吸浅快指数(定义为呼吸频率与潮气量之比)和ROX指数(定义为脉氧饱和度与吸氧体积分数之比除以呼吸频率)进行比较。结果 共筛查524 520份住院记录,最终纳入患者318例,48 h再插管患者38例(11.95%)。机器算法XGBoost模型预测撤机失败的特征重要性依次为撤机前机械通气时间、体质量指数、简化急性生理评分Ⅱ、心率(HR)、氧分压(PaO2)、平均动脉压、潮气量、年龄、脉氧饱和度(SpO2)、呼吸频率。根据以上特征重要性,以HR/PaO2和HR/SpO2作为48 h再插管预测指标。根据撤机前4 h数据,HR/PaO2和HR/SpO2的ROC曲线下面积(AUC)为0.640和0.617,高于呼吸浅快指数(AUC=0.537)及ROX指数(AUC=0.539)。根据撤机后4 h数据,HR/SpO2的AUC为0.657,高于ROX指数(AUC=0.587)。撤机后4 h,HR/SpO2由基线数据上升至1.2时,预测48 h再插管特异度可达92%。高流量氧疗治疗组患者撤机后4 h内,撤机失败患者的HR/SpO2较撤机成功患者差异有统计学意义(1.02 vs 0.92, P < 0.05),同时段ROX指数改变差异无统计学意义(8.14 vs 9.27, P > 0.05)。在撤机后8~12 h,撤机失败患者与撤机成功患者比较,HR/SpO2与ROX指数差异均有统计学意义(均P < 0.05)。结论 对于术后低氧血症患者,HR/SpO2比ROX指数能更早更准确地预测高流量吸氧治疗失败,但两者的临床价值尚需进一步评估。
关键词: 高流量吸氧    手术    低氧血症    重症监护医疗信息数据库Ⅳ    
Early predictors of high flow oxygen treatment failure for post-operation patients with hypoxemia
Liu Taotao1 , Zhao Qinyu2 , Du Bin3     
1. Department of Surgical Intensive Care Unit, Beijing Hospital, National Center of Gerontology, Institute of Geriatric Medicine, Chinese Academy of Medical Sciences, Beijing 100730, China;
2. College of Engineering and Computer Science, Australian National University, Canberra 2600, Australia;
3. Department of Medical Intensive Care Unit, Peking Union Medical College Hospital, Peking Union Medical College, Chinese Academy of Medical Sciences, Beijing 100730, China
Abstract: Objective To explore the early predictors of high flow oxygen treatment failure for post-operation patients with hypoxemia. Methods The post-operation adult patients with hypoxemia (100 mmHg < PaO2/FiO2≤300 mmHg) received high flow nasal cannula (HFNC) oxygen were retrospectively screened in the Medical Information Mart for Intensive Care IV (MIMIC-IV) database. The patients were assigned to the treatment success or failure group according to whether receiving reintubation with 48 h after extubation. The risk factors of 48-h reintubation were screened and analyzed by extreme gradient boosting (XGBoost) algorithm. And the predictors were formulated according to the risk factors. The changes of predictors were collected from extubation to 48 h later. The predictors were compared at different time points after extubation between patients weaning successfully and failed with t test. The values at different time after extubation were also compared to the baseline data. The areas under the receiver operating characteristic (ROC) curve were calculated for 48-h reintubation prediction according to values at 4 h before and after extubation, which were compared with those of rapid shallow breathing index (RSBI) and ROX index. The RSBI was defined as the ratio of respiratory rate to tidal volume. The ROX index was defined as the ratio of SpO2/FiO2 to respiratory rate. Results A total of 524 520 medical records were screened and 318 patients were included. There were 38 patients (11.95%) received reintubation within 48 h. According to the XGBoost model, the important features of 48-h reintubation were the duration before extubation, body mass index, simplified acute physiology scoring II, heart rate (HR), PaO2, mean blood pressure, tidal volume, age, SpO2 and respiratory rate. Thus HR/PaO2 and HR/SpO2 were formulated as predictors for 48-h reintubation according to the above features. The areas under the ROC of HR/PaO2 and HR/SpO2 were 0.640 and 0.617 for 48-h reintubation prediction according values at 4 h before extubation, which were larger than those of RSBI (0.537) and ROX index (0.539). According values at 4 h after extubation, the area under the ROC of HR/SpO2 was 0.657, which was larger than that of ROX index (0.587). When the HR/SpO2 reached 1.2 at 4 h after extubation, the specificity for 48-h reintubation was 92%. There was significant difference of HR/SpO2 at 4 h after extubation between patients weaning successfully and failed (1.02 vs 0.92, P < 0.05), and no significant difference of ROX index at the same time (8.14 vs 9.27, P > 0.05). There were significant differences of HR/SpO2 and ROX index at 8 to 12 h after extubation between the two groups (both P < 0.05). Conclusions HR/SpO2 is more early and accurate in predicting HFNC failure than ROX index for post-operation patients with hypoxemia. However, both the predictors should be further evaluated.
Key words: High-flow oxygen treatment    Surgery    Hypoxemia    Medical Information Mart for Intensive Care IV    

近年来国内外已有大量临床研究报道了经鼻高流量氧疗(high flow nasal cannula oxygen,HFNC)对于术后低氧血症患者的治疗效果[1-5]。研究发现HFNC在一定适应证之内,其治疗效果不亚于无创机械通气[6]。但有研究发现患者接受高流量吸氧治疗失败可能导致延迟插管,增加病死率[7]。因此有必要对高流量吸氧治疗效果进行早期预测。有研究提出ROX指数(脉氧与吸氧体积分数和呼吸频率的比值)可用于预测高流量氧疗失败,但其准确性还需要进一步验证[8]

本研究回顾性分析重症监护医疗信息数据库Ⅳ(Medical Information Mart for Intensive Care Ⅳ,MIMIC-Ⅳ)中的术后撤机患者,通过机器学习算法分析治疗失败患者的特征重要性,探索高流量氧疗失败的早期预测指标,并与ROX指数比较预测准确性。

1 资料与方法 1.1 一般资料

MIMIC-Ⅳ数据库包括2008—2019年重症监护病房(intensive care unit, ICU)患者的高质量医学信息数据[9],由贝斯以色列女执事官医学中心创建,基于MIMIC-Ⅲ数据库进行更新。该数据库已通过伦理审查,于2020年9月开放,笔者已获得数据授权。

纳入标准:术后患者机械通气撤机时合并低氧血症(100 mmHg < PaO2/FiO2≤300 mmHg)(1 mmHg=0.133 kPa);年龄≥18岁;合并或不合并高碳酸血症;撤机后接受高流量吸氧治疗。

排除标准:气管切开;意外拔管;撤机后交替接受高流量氧疗和无创机械通气的患者。

1.2 研究方法

患者根据撤机后48 h有无再次气管插管分为撤机成功组和撤机失败组。记录以下数据:年龄、性别、体质量指数(body mass index,BMI)、入ICU后24 h简化急性生理评分Ⅱ(simplified acute physiology scoring Ⅱ,SAPS-Ⅱ)、机械通气时间、拔管前至撤机后48 h生理参数及血气分析、病死率、住ICU时间、住院时间。

采用机器算法分析患者撤机后48 h再行插管的危险因素,由此制定高流量氧疗失败的预测指标。计算撤机前后4 h基线数据预测48 h再插管的准确性,计算受试者工作特征曲线(receiver operating characteristic curve,ROC)下面积,与呼吸浅快指数(the rapid shallow breathing index,RSBI)及ROX指数进行比较。呼吸浅快指数=呼吸频率/潮气量;ROX指数=脉氧/吸入氧体积分数/呼吸频率[10]。记录两组患者从撤机前至撤机后48 h生理参数和预测指标动态改变。比较撤机成功和撤机失败患者在各时间段预测指标差异,比较各时间段预测指标与基线数据差异。

1.3 统计学方法

所有数据采用R(3.6.1)统计分析,正态分布的计量资料采用均数±标准差(Mean±SD)描述,非正态分布的计量资料采用中位数(四分位数)[M(QL, QU)]描述。组间比较分别采用两独立样本t检验和Mann-Whitney U检验,率的比较采用χ2检验。绘制ROC曲线后,根据约登指数确定截断值,计算特异度和灵敏度。使用机器学习算法模型(the XGBoost model)分析再插管危险因素[11],计算SHAP(SHapley Additive explanation)绝对值在样本水平的均数。以P < 0.05为差异有统计学意义。

2 结果

总共筛查524 520份住院记录,其中20 165例计划拔管患者,最终纳入术后撤机时合并中重度低氧血症(100 mmHg < PaO2/FiO2≤300 mmHg)并接受高流量吸氧治疗患者318例。48 h再插管患者为38例(11.95%),28 d病死率为21例(6.60%),患者基线资料见表 1

表 1 治疗成功组与失败组患者基线数据和预后 Table 1 Baseline data and prognosis of patients in the treatment success and failure groups
指标 合计(n=318) 成功组(n=280) 失败组(n=38) 检验值t/F/χ2 P
年龄[岁, M(QL, QU)] 68.15(58.58, 78.17) 68.39(59.53, 78.24) 64.47(48.52, 77.10) 2.374 0.123
男性(例,%) 210(66.04) 180(64.29) 30(78.95) 2.587 0.108
BMI (kg/m2, Mean±SD) 30.81±6.71 31.14±6.83 28.42±5.25 2.840 0.006
SAPS-Ⅱ(Mean±SD) 43.03±13.24 43.08±13.59 42.63±10.36 0.239 0.812
机械通气时间[h, M(QL, QU)] 26.83(9.82, 88.72) 24.87(9.14, 82.97) 53.92(15.89, 108.83) 3.354 0.067
查尔斯合并症指数[M(QL, QU)] 7(6, 9) 7(6, 9) 9(8, 11) 2.354 0.086
手术种类(例,%)
  心胸及大血管手术 62(19.50) 52(16.35) 10(26.32) 24.354 < 0.01
  腹部外科手术 41(12.89) 33(10.38) 8(21.05) 1.279 0.276
  神经外科手术 114(35.85) 106(37.86) 8(21.05) 4.109 0.043
  骨科手术 44(13.84) 40(14.28) 4(10.53) 0.397 0.529
  其他 57(17.92) 49(15.41) 8(21.05) 0.618 0.432
撤机前生理参数  
  心率(次/min,Mean±SD) 84.08±13.87 83.19±13.39 90.62±15.70 2.787 0.008
  呼吸频率(次/min,Mean±SD) 19.32±4.28 19.31±4.30 19.35±4.19 0.053 0.958
  潮气量(mL,Mean±SD) 513.21±125.91 509.84±124.24 543.93±138.96 1.222 0.231
  平均动脉压(mmHg,Mean±SD) 78.98±12.37 78.63±12.10 81.50±14.13 1.190 0.240
  pH (Mean±SD) 7.40±0.06 7.41±0.06 7.40±0.07 0.688 0.495
  PaO2[mmHg, M(QL, QU)] 92.50(80.50, 109.00) 93.62(80.00, 110.00) 89.92(82.62, 104.88) 0.284 0.594
  PaCO2 (mmHg, Mean±SD) 39.90±6.78 39.96±6.79 39.46±6.76 0.430 0.669
  SpO2[%, M(QL, QU)] 97.00(95.20, 98.50) 97.00(95.22, 98.50) 96.50(95.12, 98.12) 0.208 0.649
  PaO2/FiO2[mmHg, M(QL, QU)] 201.29(164.00, 238.56) 200.62(162.00, 238.50) 208.00(170.10, 231.62) 0.267 0.606
预后
  28 d病死率(例,%) 21(6.60) 18(6.43) 3(7.89) 1.248 0.726
  住院时间[d, M(QL, QU)] 13.84(8.22, 21.59) 12.70(7.81, 19.13) 22.66(15.07, 31.09) 23.279 < 0.01
  住ICU时间[d, M(QL, QU)] 7.41(4.08, 14.45) 6.52(3.85, 12.84) 14.41(8.91, 20.79) 26.118 < 0.01
注:BMI为体质量指数,SAPS-Ⅱ为简化急性生理评分Ⅱ,PaO2为氧分压,PaCO2为二氧化碳分压,SpO2为脉氧饱和度,FiO2为吸氧体积分数

机器算法XGBoost模型预测撤机失败的特征重要性依次为撤机前机械通气时间、BMI、SPAP-Ⅱ、心率(HR)、氧分压(PaO2)、平均动脉压、潮气量、年龄、脉氧饱和度(SpO2)、呼吸频率等,见图 1。根据以上特征重要性,构建HR/PaO2和HR/SpO2作为48 h再插管预测指标。

BMI为体质量指数,SAPS-Ⅱ为简化急性生理评分Ⅱ,PaO2为氧分压,SpO2为脉氧饱和度,PaCO2为二氧化碳分压,COPD为慢性阻塞性肺疾病,SHAP绝对值即SHapley Additive explanation 图 1 机器算法XGBoost模型预测48 h再插管特征重要性 Fig 1 The importance of features in machine algorithm XGBboost model for predicting 48-h reintubation

治疗失败组患者在撤机后48 h内,HR/PaO2和HR/SpO2增加,ROX指数降低,见图 2

图 2 两组患者HR/PaO2、HR/SpO2和ROX指数撤机后48 h改变曲线 Fig 2 The changes of HR/PaO2, HR/SpO2 and ROX index within 48 h after weaning in the treatment sucess and failure groups

患者撤机后4 h内,撤机失败患者的HR/SpO2较撤机成功患者增加,差异有统计学意义(1.02 vs 0.92, P < 0.05),同时段ROX指数下降,但差异无统计学意义(8.14 vs 9.27, P > 0.05)。在撤机后8~12 h,治疗失败患者与撤机成功患者比较,HR/SpO2(1.00 vs 0.93, P < 0.05)与ROX指数(7.86 vs 9.13, P < 0.05)差异均有统计学意义,见表 2

表 2 两组患者撤机后48 h监测指标改变 Table 2 Changes of monitoring indicators within 48 h after weaning in the treatment sucess and failure groups
指标 撤机失败(n=38)   撤机成功(n=280)
撤机前4 h 撤机后4 h 20~24 h 36~40 h   撤机前4 h 撤机后4 h 20~24 h 36~40 h
心率(次/min,Mean±SD) 90.62±15.70a 96.10±16.69a 100.58±14.78ab 115.05±9.92ab   83.19±13.39 87.97±14.47b 85.50±14.49 85.72±14.88b
呼吸频率(次/min,Mean±SD) 19.35±4.19 22.90±4.39b 24.58±5.85ab 26.88±8.09   19.31±4.30 21.37±4.74b 21.09±4.60b 21.24±4.96b
平均动脉压(mmHg,Mean±SD) 81.50±14.13 81.46±15.73 82.53±14.12 81.92±12.14   78.63±12.10 79.96±12.69 79.48±11.90 79.04±11.71
pH (Mean±SD) 7.40±0.07 7.37±0.12 7.37±0.10 7.42±0.08   7.41±0.06 7.41±0.07 7.43±0.06b 7.45±0.06b
PaO2[mmHg, M(QL, QU)] 89.92 (82.62, 104.88) 94.00 (75.50, 123.33) 82.00 (72.00, 83.75)b 75.50 (68.75, 82.25)   93.62 (80.00, 110.00) 83.00 (70.50, 98.50)b 78.00 (70.50, 87.00)b 83.50 (69.00, 109.00)b
PaCO2 (mmHg, Mean±SD) 39.46±6.76 41.82±7.80 39.71±10.52 47.50±4.95   39.96±6.79 39.38±7.03 37.82±8.68 37.91±8.68
SpO2[%, M(QL, QU)] 96.50 (95.12, 98.12) 95.12 (94.06, 95.96)b 94.25 (93.90, 95.95)b 95.00 (94.50, 96.25)   97.00 (95.22, 98.50) 95.00 (93.75, 96.64)b 95.00 (93.75, 96.50)b 95.25 (93.50, 97.00)b
PaO2/FiO2[mmHg, M(QL, QU)] 208.00 (170.10, 231.62) 151.79 (139.00, 172.56)b 86.32 (85.66, 95.68)b 75.50 (68.75, 82.25)ab   200.62 (162.00, 238.50) 135.00 (99.50, 173.50)b 136.67 (90.31, 178.86)b 152.00 (120.00, 172.86)
ROX指数[M(QL, QU)] 8.17 (6.59, 9.74) 8.14 (6.54, 10.42) 6.07 (5.66, 6.80)ab 5.77 (4.68, 6.77)ab   9.43 (7.31, 11.44) 9.27 (6.60, 11.28) 8.91 (6.62, 10.66)b 9.47 (7.00, 11.31)
HR/PaO2 (Mean±SD) 0.98±0.23a 1.06±0.37 1.28±0.29b 1.64±0.49   0.89±0.28 1.09±0.33b 1.13±0.60b 1.08±0.36b
HR/SpO2 (Mean±SD) 0.94±0.16a 1.02±0.18a 1.07±0.16ab 1.22±0.10ab   0.86±0.14 0.92±0.15b 0.90±0.15b 0.90±0.16b
注:PaO2为氧分压,PaCO2为二氧化碳分压,SpO2为脉氧饱和度,FiO2为吸氧体积分数,HR为心率,与治疗成功患者比较,aP < 0.05;与撤机前4 h比较,bP < 0.05

撤机前4 h时HR/PaO2和HR/SpO2预测48 h再插管ROC曲线下面积(AUC)为0.640和0.617,高于呼吸浅快指数(AUC=0.537)及ROX指数(AUC=0.539)。撤机后4 h后HR/SpO2预测48 h再插管的AUC为0.657,高于ROX指数(AUC=0.587)。撤机4 h后HR/SpO2由基线值升至1.2时,预测48 h再插管特异度为92%,见表 3图 3

表 3 预测指标预测48 h再插管准确性比较 Table 3 Comparison of index prediction accuracy of 48 h reintubation
指标 AUC(95% CI P 截断值 约登指数 灵敏度 特异度 阳性预测值 阴性预测值
撤机前4 h
  HR/PaO2 0.640(0.570~0.699) < 0.01 0.873 0.266 0.602 0.664 0.175 0.934
  HR/SpO2 0.617(0.543~0.686) < 0.01 0.830 0.199 0.732 0.467 0.148 0.932
  呼吸浅快指数 0.537(0.459~0.618) < 0.01 48.40 0.113 0.408 0.704 0.145 0.904
  ROX指数 0.539(0.471~0.602) < 0.01 0.107 0.154 0.634 0.520 0.143 0.918
撤机后4 h
  HR/SpO2 0.657(0.585~0.716) < 0.01 1.203 0.326 0.400 0.926 0.462 0.907
  ROX指数 0.587(0.529~0.645) < 0.01 6.376 0.021 0.211 0.800 0.139 0.875
HR为心率,PaO2为氧分压,SpO2为脉氧饱和度,AUC为曲线下面积

图 3 撤机后4 h时HR/SpO2和ROX指数预测48 h再插管ROC曲线 Fig 3 The ROC curves of HR/SpO2 and ROX index predicting 48-h reintubation according to values at 4 h after weaning
3 讨论

本研究筛选了MIMIC-Ⅳ中超过50万份病例,最终纳入318例撤机时合并中重度低氧血症的术后患者接受HFNC序贯治疗。在ICU患者中,撤机后再插管的发生率一般为10%[12],但是对于合并危险因素接受HFNC序贯治疗的患者,再插管率可以达到20%。在本研究中,接受高流量氧疗治疗患者48 h再插管率为11.95%,这与之前的两项随机对照研究结果相似[1, 6]。根据既往研究,再插管患者中有半数再插管时间集中于撤机后48 h内,一般把48 h内再插管视作撤机失败[13]。研究发现,高流量吸氧治疗失败导致的48 h之后的延迟再插管可能会增加患者病死率[7],因此有必要对这一类患者进行早期预测。

RSBI常被作为患者撤机时施行自主呼吸试验的筛查内容,通常把RSBI≥105作为撤机失败的预测指标之一。但研究表明基线水平的RSBI对于撤机失败的预测价值较差[14-16],而且高流量氧疗时通常不监测潮气量,不利于动态观测RSBI。ROX指数是近几年提出的对于高流量氧疗失败的预测指标,由SpO2、FiO2和呼吸频率组成[17]。这三项参数都易于获得,方便撤机后监测,其动态改变有助于评估高流量氧疗是否成功。一般认为,ROX指数大于4.88时,提示高流量氧疗效果较好,而小于3.85时,提示有治疗失败的风险[8]。但是ROX指数没有纳入患者心率改变,在呼吸衰竭早期,心率是一个敏感的生理参数,有研究指出联合心率改变对ROX指数进行修正,其预测高流量氧疗失败的准确性更佳[18]

本研究通过机器算法XGBoost模型研究了在撤机后48 h高流量氧疗失败的危险因素。患者自身危险因素包括BMI、SAPS-Ⅱ、年龄等,生理参数依次为HR、PaO2、平均动脉压、潮气量、SpO2、呼吸频率。在撤机前4 h的基线数据中,HR和PaO2比潮气量和呼吸频率具有更强的特征重要性。笔者希望通过两个生理参数构建一个更准确同时易获取的预测指标,因此依据循环系统最具特征重要性的心率和呼吸系统最具特征重要性的PaO2来构建观察指标HR/PaO2。另外SpO2作为易获取的常规监测指标,将HR/SpO2也作为再插管预测指标。

绘制预测指标随时间改变曲线发现,撤机后48 h内,在治疗失败患者中HR/PaO2和HR/SpO2较治疗成功患者增高明显,ROX指数下降。以上改变可以从患者病理生理机制做出解释:高流量吸氧治疗通过提供高流量空氧混合气体和低水平的呼气末正压[19-21],对于治疗有效的患者,可以减少其呼吸做功,降低心率和呼吸频率[22-23]。绘制ROC曲线发现,根据撤机后4 h生理参数,HR/SpO2指数预测撤机后48 h再插管的AUC大于ROX指数,但均不足0.7,提示这两项预测指标的准确性均不够理想。分析原因,可能与心率与呼吸频率等生理参数在高流量氧疗失败时灵敏度较好而特异度较差有关。

撤机后4 h,撤机失败患者与撤机成功患者比较,HR/SpO2差异有统计学意义;而同一时间段撤机失败患者ROX指数为8.14,较撤机成功患者相比,差异无统计学意义。到撤机后8~12 h,撤机失败患者HR/SpO2和ROX指数与撤机成功患者比较,均发生显著改变。在撤机后24 h,HR/SpO2和ROX指数较基线数据改变均超过10%。总之,观察HR/SpO2和ROX指数在撤机后的动态改变均有利于早期发现高流量氧疗失败的患者,而HR/SpO2比ROX指数预测撤机失败的时机更早。

本研究具有一定局限性。⑴作为回顾性研究,撤机成功组和撤机失败组患者每日接受高流量氧疗治疗的时长和设置参数并不明确,这会对患者氧疗效果产生影响。⑵每例患者接受高流量氧疗适应证并不完全明确。由于2型呼吸衰竭是否可以作为高流量吸氧治疗的适应证尚不明确[24-25],因此在制定纳入标准时亦未排除2型呼吸衰竭患者。⑶患者的基础疾病、手术类型等信息会影响高流量氧疗治疗效果预测准确性,但HR/SpO2和ROX指数等仅由生理参数组成,并不包含以上重要信息。因此,简单的生理参数改变只能为临床医生判断治疗效果提供参考,通过医疗大数据建立算法模型,将患者个人信息和各项生理参数均纳入其中,将会是更具前景的人工智能预测方法。

综上所述,对于术后低氧血症患者,HR/SpO2比ROX指数能更早更准确地预测高流量吸氧治疗失败,但两者的临床价值尚需进一步评估。

利益冲突   所有作者均声明不存在利益冲突

参考文献
[1] Hernández G, Vaquero C, Colinas L, et al. Effect of postextubation high-flow nasal cannula vs noninvasive ventilation on reintubation and postextubation respiratory failure in high-risk patients: a randomized clinical trial[J]. JAMA, 2016, 316(15): 1565-1574. DOI:10.1001/jama.2016.14194
[2] 孙波, 张天卿, 胡雪忠, 等. 经鼻高流量氧疗在ICU气管插管拔管后呼吸衰竭患者中的应用[J]. 中华急诊医学杂志, 2020, 29(2): 279-283. DOI:10.3760/cma.j.issn.1671-0282.2020.02.0033
[3] 杨逢露, 吴春双, 曹夏婧, 等. 经鼻高流量氧疗在颈椎损伤伴高位截瘫患者术后脱机中的应用[J]. 中华急诊医学杂志, 2019, 28(8): 1005-1009. DOI:10.3760/cma.j.issn.1671-0282.2019.08.017
[4] 王可, 徐思成, 左蕾, 等. 经鼻高流量吸氧治疗急性呼吸衰竭的系统评价[J]. 中华急诊医学杂志, 2017, 26(8): 879-884. DOI:10.3760/cma.j.issn.1671-0282.2017.08.009
[5] 卢骁, 徐善祥, 钱安瑜, 等. 经鼻高流量氧疗在临时ICU病房治疗重症新型冠状病毒肺炎的经验[J]. 中华急诊医学杂志, 2020, 29(4): 509-510. DOI:10.3760/cma.j.issn.1671-0282.2020022.005
[6] Stéphan F, Barrucand B, Petit P, et al. High-flow nasal oxygen vs noninvasive positive airway pressure in hypoxemic patients after cardiothoracic surgery: a randomized clinical trial[J]. JAMA, 2015, 313(23): 2331-2339. DOI:10.1001/jama.2015.5213
[7] Kang BJ, Koh Y, Lim CM, et al. Failure of high-flow nasal cannula therapy may delay intubation and increase mortality[J]. Intensive Care Med, 2015, 41(4): 623-632. DOI:10.1007/s00134-015-3693-5
[8] Spinelli E, Roca O, Mauri T. Dynamic assessment of the ROX index during nasal high flow for early identification of non-responders[J]. J Crit Care, 2020, 58: 130-131. DOI:10.1016/j.jcrc.2019.08.013
[9] Johnson A, Bulgarelli L, Pollard T, et al. MIMIC-IV[DB/OL]. PhysioNet, 2020[2020-08-13]. https://doi. org/10.13026/a3wn-hq05.
[10] Roca O, Messika J, Caralt B, et al. Predicting success of high-flow nasal cannula in pneumonia patients with hypoxemic respiratory failure: The utility of the ROX index[J]. J Crit Care, 2016, 35: 200-205. DOI:10.1016/j.jcrc.2016.05.022
[11] Chen TQ, Guestrin C. XGBoost: a scalable tree boosting system[C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. San Francisco California USA. New York, NY, USA: ACM, 2016. DOI: 10.1145/2939672.2939785.
[12] Miltiades AN, Gershengorn HB, Hua M, et al. Cumulative probability and time to reintubation in US ICUs[J]. Crit Care Med, 2017, 45(5): 835-842. DOI:10.1097/ccm.0000000000002327
[13] Béduneau G, Pham T, Schortgen F, et al. Epidemiology of weaning outcome according to a new definition. the WIND study[J]. Am J Respir Crit Care Med, 2017, 195(6): 772-783. DOI:10.1164/rccm.201602-0320OC
[14] Krieger BP, Isber J, Breitenbucher A, et al. Serial measurements of the rapid-shallow-breathing index as a predictor of weaning outcome in elderly medical patients[J]. Chest, 1997, 112(4): 1029-1034. DOI:10.1378/chest.112.4.1029
[15] Patel KN, Ganatra KD, Bates JHT, et al. Variation in the rapid shallow breathing index associated with common measurement techniques and conditions[J]. Respir Care, 2009, 54(11): 1462-1466.
[16] Frutos-Vivar F, Ferguson ND, Esteban A, et al. Risk factors for extubation failure in patients following a successful spontaneous breathing trial[J]. Chest, 2006, 130(6): 1664-1671. DOI:10.1378/chest.130.6.1664
[17] Roca O, Caralt B, Messika J, et al. An index combining respiratory rate and oxygenation to predict outcome of nasal high-flow therapy[J]. Am J Respir Crit Care Med, 2019, 199(11): 1368-1376. DOI:10.1164/rccm.201803-0589OC
[18] Goh KJ, Chai HZ, Ong TH, et al. Early prediction of high flow nasal cannula therapy outcomes using a modified ROX index incorporating heart rate[J]. J Intensive Care, 2020, 8: 41. DOI:10.1186/s40560-020-00458-z
[19] Groves N, Tobin A. High flow nasal oxygen generates positive airway pressure in adult volunteers[J]. Aust Crit Care, 2007, 20(4): 126-131. DOI:10.1016/j.aucc.2007.08.001
[20] Parke R, McGuinness S, Eccleston M. Nasal high-flow therapy delivers low level positive airway pressure[J]. Br J Anaesth, 2009, 103(6): 886-890. DOI:10.1093/bja/aep280
[21] Chikata Y, Onodera M, Oto J, et al. FIO2 in an adult model simulating high-flow nasal cannula therapy[J]. Respir Care, 2017, 62(2): 193-198. DOI:10.4187/respcare.04963
[22] Corley A, Caruana LR, Barnett AG, et al. Oxygen delivery through high-flow nasal cannulae increase end-expiratory lung volume and reduce respiratory rate in post-cardiac surgical patients[J]. Br J Anaesth, 2011, 107(6): 998-1004. DOI:10.1093/bja/aer265
[23] Wagstaff TA, Soni N. Performance of six types of oxygen delivery devices at varying respiratory rates[J]. Anaesthesia, 2007, 62(5): 492-503. DOI:10.1111/j.1365-2044.2007.05026.x
[24] Yang PL, Yu JQ, Chen HB. High-flow nasal cannula for acute exacerbation of chronic obstructive pulmonary disease: a systematic review and meta-analysis[J]. Heart Lung, 2021, 50(2): 252-261. DOI:10.1016/j.hrtlng.2020.12.010
[25] Pantazopoulos I, Daniil Z, Moylan M, et al. Nasal high flow use in COPD patients with hypercapnic respiratory failure: treatment algorithm & review of the literature[J]. COPD, 2020, 17(1): 101-111. DOI:10.1080/15412555.2020.1715361