계명의대학술지제 37 권 2 호 Keimyung Med J Vol. 37, No. 2, December, 2018 66 은승완 김태권 전재천 진상찬 최우익 계명대학교의과대학응급의학과 Early Predictors of High-Flow Nasal Cannula Oxygen Therapy Failure in the Emergency Room Seung Wan Eun, M.D., Tae Kwon Kim, M.D., Jae Cheon Jeon, M.D., Sang Chan Jin, M.D., Woo Ik Choi, M.D. Department of Emergency Medicine, Keimyung University School of Medicine, Daegu, Korea Received: October 3, 2018 Revised: October 24, 2018 Accepted: December 28, 2018, Corresponding Author: Sang Chan Jin, M.D., Department of Emergency Medicine, Keimyung University School of Medicine, 56 Dalseong-ro, Jung-gu, Daegu 41931, Korea Tel: +82-53-250-7890 E-mail: jchan98@hanmail.net The authors report no conflict of interest in this work. Copyright Keimyung University School of Medicine 2018 High flow nasal cannula (HFNC) oxygen therapy has become a substitute for other non-invasive ventilation (NIV) and mechanical ventilation in patients with respiratory failure. Despite its strong points, HFNC may result in failure of therapy, and delayed intubation leads to poor patient outcomes. This study is aimed to identify the variables that predicts HFNC oxygen therapy failure in order to increment success rate and prevent delayed intubation. The study was conducted in a retrospective manner, enrolling all non-traumatic patients over the age 20 who visited a single university-affiliated tertiary medical center emergency room and were treated with HFNC between March 1, 2016 to Feburary 28, 2017. Patients were classified into two groups: HFNC success group, and HFNC failure group: HFNC failure group was defined as patients who were treated with HFNC, and resulted in intubation and mechanical ventilation, or death. General characteristics, clinical features, and laboratory findings of the two groups were compared. Additionally, univariable logistic regression was done for statistically significant variables, and if the p-value was <0.05, multivariable logistic regression analysis was conducted. 95 patients were enrolled in the study. Repeated-measure ANOVA was conducted for serial arterial blood gas samples during therapy. 62 patients were successful in the treatment, and the treatment failed in 33 cases. After multivariate logistic regression analysis, variable that showed statistically significant association with HFNC failure was initial hematocrits (p < 0.05). Repeated-measure ANOVA revealed low bicarbonate as a predictor of
67 HFNC failure. Predictors of HFNC therapy failure are initial hematocrits, bicarbonate levels during therapy. Keywords: High flow nasal cannula, Oxygen therapy, Respiratory failure 서론 응급실은여러가지종류의 1차응급진료가이루어지는장소이며, 산소치료또한흔하게이루어진다. 응급실에서흔하게사용되는산소공급의고식적방법으로는비강캐뉼러, 단순산소마스크, 벤투리마스크및비재호흡산소마스크가있다. 적절한산소치료방법의선택은환자치료의중요한부분으로서, 이러한고식적인산소치료방법들만으로응급실환자의산소치료가충분한경우도많다. 그러나이러한산소치료방법들은흡입산소분율 (fraction of inspired oxygen, FiO 2 ) 을일정하게유지하기어렵고, 최대유속 15 L/min으로제한적이며, 공급되는산소의가온가습이부족할수있다는단점이있다 [1,2]. 고식적인산소치료방법으로적절한산소공급을할수없는경우지속적양압환기장치 (continuous positive airway pressure, CPAP) 또는이상성양압환기장치 (biphasic positive airway pressure, BiPAP) 와같은비침습적환기장치 (non-invasive ventilation, NIV) 를사용하거나기관내삽관후기계환기 (endotracheal intubation and mechanical ventilation) 또는기관절개후기계환기 (tracheostomy and mechanical ventilation) 를하는침습적인환기방법을사용한다. 고유속비강캐뉼러 (high-flow nasal cannula, HFNC) 를통한산소치료는 NIV나침습적환기장치의대안으로사용될수있다. HFNC 산소치료는성인과소아모두에서적용가능하며, 사용방법이단순하고, 환자의불편감이적을뿐만아니라특별한금기증이없고, 가습된산소를최대 60 L/min의유속으로공급할수있다는장점이있다 [2,3]. 현재 HFNC 산소치료는폐질환, 심부전, 신부전, 패혈증등여러질환에서나타나는호흡부전에서사용할수있고, 수술후호흡부전, 기관삽관발관후호흡곤란등에서널리사용되고있다. HFNC 산소치료는기관 삽관의빈도와 ICU 사망률을유의하게낮추며, 호흡을위한노력의부담을덜어주는것이많은연구를통해서확인되었다 [1,3-9]. 반면, HFNC 산소치료는저산소혈증, 고이산화탄소혈증을효과적으로교정하지못하여산소치료중기관삽관을해야되는경우가 38% 로높고, 사망률또한 30% 에달한다는보고가있다 [2,5]. HFNC 산소치료가실패하는경우기관내삽관및기계환기의적용이지연되면서환자의사망률이증가하는것으로보고된다 [2]. 이에본연구의저자들은응급실에내원한호흡부전환자에서 HFNC 산소치료의실패를초기에예측할수있는요인들을밝혀내고자하였다. HFNC 산소치료의실패인자에대한여러연구가있으나, 그결과는일관되지않다 [10-12]. 본연구의목적은이전의연구에서밝힌인자들을다시검증하고새로운위험인자를밝히는데있다. HFNC 산소치료가실패할가능성이높은환자를예측하여 HFNC 치료를선별적으로시행함으로써, HFNC 치료의성공률을높이고즉각적으로기관내삽관및기계호흡이필요한환자들에서치료가지연되지않도록하고자하였다. 재료및방법본연구는계명대학교동산의료원의의학연구윤리심의위원회 (Institutional Review Board, IRB) 기준에부합하며, 이의승인을받아진행되었다. 2016년 3월 1일부터 2017년 2월 28일까지단일삼차의료기관의응급의료센터를내원한환자중, 비외상성질환에의한호흡곤란으로 HFNC 치료를받은만 20세이상의성인들을대상으로하였다. 이들중 Glasgow Coma Scale (GCS) 15점미만으로의식상태가명료하지않은경우, 기존에기관절개를시행하여기도를유지하고있는경우, 다른비침습적환기장치를이용하기위한중재치료로써 HFNC가사용된경우, 기계환기를위한중재치료로서 HFNC가사용된경우, 기관내삽관발관이후산소치료를위해 HFNC가사용된경우, 다른원인질환으로치료받던중발생한호흡곤란의치료를위해 HFNC가사용된경우, 기계환기가필요하다고의사가판단하였으나환자나보호자가이를거부하여 HFNC가사용된경우, 치료중
68 계명의대학술지제 37 권 2 호 2018 다른기관으로전원한경우, 연구에필요한자료가부족한환자의경우는연구대상에서제외하였다. HFNC의치료실패를예측할수있는인자들을알아보기위해환자를 HFNC 치료성공군과치료실패군으로분류하였다. HFNC 치료성공군은 HFNC 치료이후비강캐뉼러, 산소마스크등의저유량산소요법으로전환한경우로, 치료실패군은 HFNC 치료중기관삽관및기계환기를시행하거나치료중사망한경우로정의하였다. 기관삽관및기계환기적응은주치의의판단에의해결정되었다. HFNC 치료성공군과실패군을비교하기위하여의무기록을이용하여환자의인구통계학적데이터, 응급실에내원했을당시의생체징후및의식수준, 일반혈액검사 (complete blood cell count, CBC), 혈청전해질수치를포함한일반화학검사결과및동맥혈가스분석자료를변수로써수집하였고, 이러한데이터를토대로 acute physiology and chronic health evaluation II (APACHE II) 점수를계산하였다. 또한증상발생으로부터응급실내원까지의시간, 응급실내원시부터 HFNC 치료시작까지의시간, 치료시작으로부터치료종료 ( 치료성공군의경우는 HFNC 치료를중단한시점, 치료실패군의경우는기관내삽관을시행한시각또는환자가사망한시점 ) 까지의시간또한변수로서연구에포함하였다. 추가로 HFNC 치료시작 1-2시간사이, 2-4시간사이, 4-10시간사이, 10-20시간사이에시행한동맥혈가스분석과혈중젖산수치를치료성공과실패의변수로분석하였다. 원인질환들에따른 HFNC의치료실패여부도확인해보고자하였으며, 치료적응원인질환은 Kang 등 [2] 이연구에사용한시스템을이용하여 6그룹으로분류하였다. 각그룹은 1) 급성호흡부전 ( 폐렴, ARDS 등 ), 2) 만성폐질환자의급성호흡부전 ( 만성폐쇄성폐질환자의급성악화, 천식급성악화등 ), 3) 심인성폐부종, 4) 신부전에의한폐부종, 5) 호흡기감염에의하지않은패혈증쇼크, 6) 기관삽관발관이후발생한호흡부전이며, 이중 6) 은연구대상에서제외하여나머지다섯그룹만을분류에사용하였다. HFNC 치료에는 Optiflow (Fisher & Paykel Healthcare, Auckland, New Zealand) 가사용되었으며, 이기기는 MaxVenturi (Maxtec, Utah, USA), MR850 heated humidifier (MR290 autofeed chamber; Fisher & Paykel Healthcare), RT 202 single-limb adult breathing circuits kit (Fisher & Paykel Healthcare), OPT 844 Optiflow nasal cannula (Fisher & Paykel Healthcare) 로구성되었다. HFNC 치료는치료시작시에 FiO2 0.4-0.8로시작하여동맥혈가스분석을일정시간간격 ( 치료시작 1-2시간사이, 2-4시간사이, 4-10시간사이, 10-20시간사이 ) 으로시행하였고, 맥박산소측정기를통해산소포화도를측정하며흡기산소유속및흡기산소분율을조절하였다. SPSS Statistics version 21.0 (SPSS Inc., Chicago, IL, USA) 로통계분석을하였다. 정규분포의확인은 Kolmogorov-Smirnov 정규성검정을사용하였다. 정규분포를따르는모수적표본의경우평균 ± 표준편차, 정규분포를따르지않는비모수적표본의경우중앙값 (2분위수값-4분위수값) 으로통계량을표기하였다. 모수적표본의연속형변수는 Student s T-test, 비모수적표본의연속형변수및범주형변수의비교에는 Mann- Whitney U-test를사용하였다. 상관관계분석은 univariate logistic regression을사용하였고, 이중유의한 (p-value <0.05) 변수는 multivariate logistic regression 분석을통하여 odds ratio (OR) 를계산하였다. 시간대별로반복측정한동맥혈가스분석및젖산의경우 repeatedmeasure ANOVA 를시행하여각군의차이를분석하였다. p-value는 95% 신뢰구간에서 0.05 미만인경우통계적의미가있다고정의하였다. Multivariate logistic regression 분석에서 HFNC 치료실패와상관관계가있는변수에대해서는 receiver operating characteristic (ROC) curve로 area under the curve (AUC) 를구했고, 치료실패를예측할수있는 cut-off point는 Youden index를사용하였다. 결과 2016년 3월 1일부터 2017년 2월 28일까지단일의료기관을방문하여 HFNC 치료를받은환자수는 215명이었으며, 이들중 120명이제외되고총 95명이연구대상으로포함되었다. HFNC 치료성공군은 62명 (65.3%) 이었으며, 치료실패군은 33명 (34.7%) 이었다. 치료성공군과치료실패군사이에통계적으로의미있는차이를보이는변수는평균
69 동맥압 (mean arterial pressure, MAP), 체온, 응급실내원시부터 HFNC 치료시작까지의시간, 응급실내원당시의 PaCO 2, PaO 2, 중탄산염, 혈중나트륨, Hematocrit (Hct) 이있었다 (p < 0.05) (Table 1, 2). 두군간에통계적으로의미있는차이를보이는변수들에대해 univariate logistic regression 분석을시행하였다. 이들변수중에서 MAP, 응급실내원시부터 HFNC 치료시작까지의시간, PaCO 2, 중탄산염, Hct가두군간에통계적으로의미있는차이를보였다. MAP는낮을수록, 치료시작까지의시간은길수록, PaCO 2 와중탄산염, Hct는낮을수록치료에실패하는경향을보였다 (OR 0.973, 1.253, 0.951, 0.923, 0.911) (p < 0.05 (Table 3). Univariate logistic regression 분석결과에서 p-value 0.05 미만인변수들에대해다시 multivariate logistic regression 분석을시행하였다. 응급실내원당시의 Hct이두군간에통계적으로유의하였다 (Table 4). 시간대별로반복측정한동맥혈가스분석결과와젖산수치를 repeated-measured ANOVA로분석하였다. 그결과 ph, PCO 2, PO 2, lactic acid, PaO 2 /FiO 2 는두군간에유의한차이를보이지않았으나, 중탄산염수치는시간에따라두군간에통계적유의성이있었다 (p < 0.05) (Fig. 1). Multivariate logistic regression 분석상 HFNC 치료실패와상관관계가있던변수는다시 ROC curve와 AUC를확인하였다. 그결과, 응급실내원당시의 Hct은 AUC가 0.650, cut-off value는 35.6( 민감도 53.23%, 특이도 75.76%) 이었다 (Fig. 2). 고찰이번연구에서저자들은 HFNC 치료초기변수중 HFNC 치료의성공과실패여부를효과적으로예측할수있는인자들을확인하고자하였다. 응급실내원시에 Hct이낮을수록치료가실패할가능성이높았다. 응급실내원당시의동맥혈가스분석에서중탄산염수치는두군간에유의한차이를보이지않았으나 HFNC 치료시작시점이후에는의미있는차이를보였다. 치료성공군에서중탄산염의수치가지속적인상승소견을보인것에비해 치료실패군은치료시작 2시간이후에중탄산염이감소하는소견을보였다. 이전의연구들에서치료실패를예측하는인자들을보면, Kim 등 [10] 은연구에서 GCS 15 미만의의식장애를가진경우, 호흡수 30회를초과하는경우 HFNC 치료에실패하는경향이있음을보여주었고, Frat 등 [11] 은고전적산소요법, NIV, HFNC를비교한연구에서치료시작 1시간후의맥박수증가가기관삽관을예측할수있는요인으로꼽았다. 저자들은 HFNC 치료를하기위해서는자발호흡및치료에대한이해와협조가필수적이라고판단하여 GCS 15점미만의의식장애는연구대상에서이미제외하였기에이에대해검증은하지않았으며, 이번연구에서호흡수와맥박수는 HFNC 치료실패와무관한것으로나타났다. PaCO 2 는호흡수에따라민감하게변하는것으로알려져있다. 본연구에서는치료성공군과실패군의호흡수에서유의한차이가없음에도치료성공군의 PaCO 2 가실패군에비해일관되게높은수치를보였던점에서, 응급실에서의호흡수의측정방법자체에오류가있었을가능성이높다. 호흡수와 HFNC 치료실패에대해서는추가적인연구가필요할것으로보인다. Putinati 등 [13] 은 COPD로인한호흡부전환자들에서 NIV 치료실패를예측하는인자로높은 APACHE II score와낮은알부민수치를제안하였으나, 본연구에서는치료성공군과치료실패군사이에서차이를보이지않았다. Kim 등 [12] 은면역억제환자의 pneumocystis pneumonia에대한연구에서치료시작전높은 SOFA (sepsis-related organ failure assessment) score가 HFNC 치료실패와연관이있으며, 중탄산염은치료실패요인이아닌것으로보고하였으나 (OR 0.84, p-value = 0.052), 본연구에서는 HFNC 치료중낮은중탄산염이실패를예측할수있는인자인것으로확인되었다. 호흡부전및쇼크상태에서는말초조직으로산소공급이원활하게일어나지않으며, 혐기성대사과정을통해젖산이생성되고대사성산증이발생한다. 신체에서는이에대한자극으로호흡중추가활성화되고과환기를하여 PaCO 2 를감소시키는호흡성보상을하게된다. 대사성산증을보상하는화학적중화제로중탄산염의소모가야기되며, 호흡성보상에의한 PCO 2 감소가신장에서의중탄산염재흡수감소로이어져혈중중탄산염감소가발생하게된다 [14]. 본연구에서 HFNC
70 계명의대학술지제 37 권 2 호 2018 Table 1. Comparison of clinical characsteristics Parameters HFNC success (n = 62) HFNC failure (n = 33) p-value Sex (male) * 43(69.3%) 21(63.6%) 0.573 Age 73.5(65-82) 77(64.5-83) 0.539 History of smoking * 32(51.6%) 13(39.4%) 0.342 Pulmonary disorders COPD * 10((16.1%) 3(9.1%) 0.344 Asthma * 1(1.6%) 1(3.0%) 0.649 ILD * 2(3.2%) 2(6.1%) 0.515 Other pulmonary disorders * 16(25.8%) 8(24.2%) 0.868 Comorbidities Hypertension * 37(59.7%) 14(42.4%) 0.110 Diabetes mellitus * 23(37.1%) 11(33.3%) 0.717 Renal disorders * 14(22.6%) 5(15.2%) 0.318 Hepatic disorders * 2(3.2%) 2(6.1%) 0.515 Cardiac disorders * 14(22.6%) 12(36.3%) 0.154 Neurologic disorders * 17(27.4%) 4(12.1%) 0.089 Malignancy * 10(16.8%) 8(24.2%) 0.339 Immunosuppression state * 3(4.8%) 2(6.1%) 0.801 Initial Vital Signs at ER Arrival Mean arterial pressure (mmhg) 98.03 ± 18.67 88.79 ± 19.05 0.025 Pulse rate (/min) 103.5 ± 23.4 99.7 ± 20.3 0.432 Respiratory rate (/min) 22(22-25) 20(20-25) 0.280 Body temperature ( ) 37(36.5-38) 36.8(36.1-37.4) 0.046 Time to ER arrival (days) 0.417(0.125-2) 1.167(0.208-3.208) 0.062 Time to therapy initition (days) 0.25(0.083-1.094) 0.854(0.249-4.168) 0.012 Time to therapy termination (days) 2.625(0.865-5) 1.25(0.311-6.396) 0.622 APACHE II score 14(12-18) 13(11-18) 0.523 * sex (male), history of smoking, pulmonary disorders, underlying diseases are expressed in n (%). These are variables that follow the normal distribution, and are expressed in average ± standard deviation. other variables are expressed in median (25 quartile-75quartile). Other pulmonary disorders: pulmonary embolism, lung infarction, active tuberculosis. Renal disorders: any chronic kidney disease regardless of need for dialysis. Hepatic disorders: any hepatitis, carriers of hepatitis virus on anti-viral therapy, fatty liver, liver cirrhosis. Cardiac disorders: any cardiomyopathies, chronic heart failure, history of coronary artery disease, patients on anti-arrythmics. Neurologic disorders: any history of stroke, dementia, Parkinson s disease, seizures, any history of spinal cord disorders, peripheral neuropathies, neuromuscular diseases. HFNC: high flow nasal cannula, COPD: chronic obstructive pulmonary disease, ILD: interstitial lung disease, MAP: mean arterial pressure, APACHE II: acute physiology and chronic health evaluation II, ER: emergency room.
71 Table 2. Comparison of laboratory results at emergency department arrival Parameters HFNC success (n = 62) HFNC failure (n = 33) p-value Initial Arterial Blood Gas Analysis ph 7.44(7.35-7.48) 7.44(7.36-7.49) 0.509 PaCO 2 (mmhg) 32(28-39.5) 28(25-32.5) 0.009 PaO 2 (mmhg) 55(47-65.5) 63(56-73) 0.013 Bicarbonate (mmol/l) 22.4(18.5-26) 19.4(16.3-22.2) 0.012 Serum lactate (mmol/l) 1.9(1.3-3.3) 1.8(1.3-3.4) 0.899 PaO 2 /FiO 2 252.4(219.0-309.5) 271.4(232.4-331.0) 0.133 Initial Laboratory Results Serum sodium (mmol/l) 136(134-138) 133(131-136) 0.005 Serum postassium (mmol/l) 4.2(3.7-4.6) 4.1(3.7-4.8) 0.953 Serum creatinine (mg/dl) 1.04(0.8-1.94) 1.24(0.8-2.05) 0.752 Hematocrit (%) 35.971 ± 6.745 31.637 ± 7.205 0.004 White blood cell count (10 3 /μl) 12.26(9.27-14.31) 12.12(7.44-14.82) 0.582 These are variables that follow the normal distribution, and are expressed in average ± standard deviation. other variables are expressed in median (25 quartile-75 quartile). PaCO 2 : arterial partial pressure of CO 2, PaO 2 : arterial partial pressure of O 2, FiO 2 : fraction of inspired oxygen concentration. Table 3. Univariate logistic regression analysis of general characteristics and initial laboratory data Parameters p-value Odds ratio (95% CI) Mean arterial pressure 0.029 0.973(0.95-0.997) Body temperature 0.084 0.651(0.4-1.059) Time to therapy initiation 0.011 1.253(1.054-1.49) PaCO 2 0.041 0.951(0.907-0.998) PaO 2 0.296 1.009(0.992-1.026) Bicarbonate 0.039 0.923(0.856-0.996) serum Na 0.057 0.919(0.843-1.002) Hct 0.008 0.911(0.851-0.976) PaCO 2 : arterial partial pressure of CO 2, PaO 2 : arterial partial pressure of O 2, Hct: hematocrit. 치료중측정된중탄산염수치가일관되게낮게측정된데에는이러한신체반응이기여한것으로추정된다. 본 연구에서 HFNC 치료중 PaO 2 및 PaO 2 /FiO 2 가개선되었음에도불구하고중탄산염이낮게측정되는것은대사성산증을교정하지못한것이 HFNC 치료실패에영향을미쳤다고해석할수있다. 호흡부전환자의인체는이러한산-염기대사및호흡반응에맞추어맥박수, 호흡수를변화시키며, 이러한인체의보상작용은개인별로, 그리고시간에따라그정도가다르게나타난다. 또한호흡부전의치료에 HFNC 외에도네뷸라이저치료, 수액요법, 중탄산나트륨투여, 승압제사용등의여러가지치료들이복합적으로이루어지기때문에중탄산염이외에활력징후나동맥혈가스분석의다른인자들은의미있는차이를나타내지못하는것으로생각된다. 낮은 Hct은입원중 NIV 치료가필요한 COPD 환자의치료결과에관한연구에서도사망과연관이있는것으로나타났다 [15]. Surviving sepsis campaign에서도패혈증환자중 mixed venous oxygen saturation (ScvO 2 ) 70% 미만이며 Hct 30% 미만인환자에서는수혈하도록
72 계명의대학술지제 37 권 2 호 2018 Table 4. Multivariate logistic regression analysis of general characteristics and initial laboratory data Parameters Category B Odds ratio 95% Confidence interval p-value Mean arterial pressure -0.017 0.983 0.956-1.01 0.220 Time to therapy initation 0.011 1.192 0.999-1.422 0.052 PaCO 2-0.054 1.011 0.945-1.081 0.752 Bicarbonate -0.073 0.948 0.847-1.061 0.351 Hct 0.199 0.929 0.868-0.996 0.038 PaCO 2 : arterial partial pressure of CO 2, Hct: hematocrit. 권장하고있다 [16]. 호흡부전환자들에게서보이는산소부족은호흡일량 (work of breathing, WOB) 을증가시키며, 이로인해주요장기들의산소요구량또한증가한다. 이때장기와조직으로의산소운반능력이뒷받침되어야하며, Hct는이와연관이있다 [17]. 낮은 Hct는가스교환에필요한적혈구의부족으로 ventilationperfusion mismatch를야기하며, 말초조직으로의산소운반능력저하로저산소증을조장하여이와같은결과가나타난것으로보인다. 이전의연구에서치료실패를예측하는요인으로꼽힌맥박수, 호흡수는두군간에유의한차이를보이지않았고, PaCO 2 는단변량분석에서는치료실패군과성공군사이에서유의한차이가있었지만, logistic regression 분석에서는경향성이없는것으로나타났다. 또한 NIV 환자를대상으로한연구에서연관성이있는것으로나타났던 APACHE II score 또한본연구에서는통계적으로의미있는차이를보이지않았다 [5,15]. PaO 2 / FiO 2 또한본연구에서는 HFNC 치료실패와의연관성을찾을수없었다. PaO 2 /FiO 2 의경우반복측정에의한변화량 (ΔPaO 2 /FiO 2 ) 을절대값과비율로두군간에비교하는추가연구를진행하였으나, 이또한유의한차이가나타나지않았다. PaO 2 /FiO 2 의경우치료시작시각부터치료종료시각까지시간의중앙값이치료성공군에서 2.625일, 치료실패군에서 1.25일이며, 각군에서최대 13일, 32일인점을고려하면, 본연구에서조사한치료시작으로부터 12시간은의미있는차이를보이기에짧은시간일가능성이높다. 본연구는몇가지한계점을가지고있다. 첫째, 본 연구는단일삼차병원에서시행되었으며, 연구에포함된환자의수가 95명의소규모였다. 둘째, 연구가후향적으로수집된자료를토대로하여, 환자선별과정에서선택편견이있을가능성이있다. 셋째, HFNC 치료의적응이기관삽관및기계환기치료와비교하여명확하게정의되지않았고, 본연구가이루어진병원에서도지정된프로토콜이없어, 전적으로주치의사의재량에의해치료가시행되어이에의한편견이발생했을수있다. 넷째, HFNC 치료진행중환자의활력징후와동맥혈가스분석이외의혈액검사결과가미비하여, 중탄산염이외에치료중에환자에게영향을미칠수있는다른인자들에대한연구를시행하지못했다. 예를들어, 본연구에서응급실내원당시의 Hct 결과가치료실패와연관성이있는것으로나타났으나, 응급실내원시부터 HFNC 치료시작시까지의시간이수일이상격차를보이는경우도있어실제로 HFNC 치료중의 Hct이 HFNC 치료의실패와연관성이있는지를확인하지못하였다. 이는추가적인연구를통해확인이필요할것으로사료된다. 다섯째, 본연구에서호흡수의측정이부정확하게이루어져자료수집단계부터결함이있었을가능성이높다. 호흡수와 HFNC 치료실패와의관계에대해서는추가적인연구가필요할것으로보인다. 요약응급실을통해 HFNC 치료를받는환자에서치료실패를예측할수있는인자는응급실내원당시의 Hct,
73 Fig. 1. Results of the values of arterial blood gas analysis during HFNC therapy. Repeated-measured ANOVA was conducted to compare between the HFNC success group and HFNC failure group. (A) ph values showed no statistical difference (p = 0.120). (B) PaCO 2 values were lower in the HFNC failure group, but statistically insignificant (p = 0.275). (C) PaO 2 values trend to show lower level in the HFNC failure group, but this difference was not statistically significant (p = 0.899). (D) Bicarbonate levels in the HFNC failure group were significantly lower than the HFNC success group (p = 0.043). (E) Lactate values differ between the two groups without statistical significance (p = 0.231). (F) PaO 2 /FiO 2 between the two groups showed no statistically significant difference (p = 0.331).
74 계명의대학술지제 37 권 2 호 2018 Fig. 2. Receiver operating characteristic (ROC) curves for HFNC failure. (A) Areas under the curve (AUC) is 0.650 (p < 0.01) for Hct at arrival to the emergency room. HFNC 치료중시행한동맥혈가스분석의중탄산염수치이다. 응급실내원당시 Hct이낮거나 HFNC 치료중동맥혈가스분석상중탄산염수치가일관되게낮게측정될경우 HFNC 치료가실패할가능성이높음을인지하고이에대비해야한다. 참고문헌 1. Roca O, Riera J, Torres F, Masclans JR. High-flow oxygen therapy in acute respiratory failure. Respir Care 2010;55:408-13. 2. Kang BJ, Koh Y, Lim CM, Huh JW, Baek S, Han M, et al. Failure of high-flow nasal cannula therapy may delay intubation and increase mortality. Intensive Care Med 2015;41:623-32. 3. Vargas F, Saint-Leger M, Boyer A, Bui NH, Hilbert G. Physiologic Effects of High-Flow Nasal Cannula Oxygen in Critical Care Subjects. Respir Care 2015;60:1369-76. 4. Peters SG, Holets SR, Gay PC. High-flow nasal cannula therapy in do-not-intubate patients with hypoxemic respiratory distress. Respir Care 2013;58:597-600. 5. Lin SM, Liu KX, Lin ZH, Lin PH. Does high-flow nasal cannula oxygen improve outcome in acute hypoxemic respiratory failure? A systematic review and meta-analysis. Respir Med 2017;131:58-64. 6. Yoo JW, Synn A, Huh JW, Hong SB, Koh Y, Lim CM. Clinical efficacy of high-flow nasal cannula compared to noninvasive ventilation in patients with postextubation respiratory failure. Korean J Intern Med 2016;31:82-8. 7. Kim ES, Lee H, Kim SJ, Park J, Lee YJ, Park JS, et al. Effectiveness of high-flow nasal cannula oxygen therapy for acute respiratory failure with hypercapnia. J Thorac Dis 2018;10:882-8. 8. Doshi P, Whittle JS, Bublewicz M, Kearney J, Ashe T, Graham R, et al. High-Velocity Nasal Insufflation in the Treatment of Respiratory Failure: A Randomized Clinical Trial. Ann Emerg Med 2018;72:73-83.e5. 9. Atwood CW Jr, Camhi S, Little KC, Paul C, Schweikert H, Macmillan NJ, et al. Impact of Heated Humidified High Flow Air via Nasal Cannula on Respiratory Effort in Patients with Chronic Obstructive Pulmonary Disease. Chronic Obstr Pulm Dis 2017;4:279-86. 10. Kim HJ, Lee DW, Lee JW, Moon HJ, Choi JH, Joeng DK, et al. Factors about Failure after High Flow Oxygen through Nasal Cannula Therapy in Hypoxic Respiratory Failure Patients at Emergency Department Presentation. J Korean Soc Emerg Med 2016;27:580-5. 11. Frat JP, Ragot S, Coudroy R, Constantin JM, Girault C, Prat G, et al. Predictors of Intubation in Patients With Acute Hypoxemic Respiratory Failure Treated With a Noninvasive Oxygenation Strategy. Crit Care Med 2018;46:208-15. 12. Kim WY, Sung H, Hong SB, Lim CM, Koh Y, Huh JW. Predictors of high flow nasal cannula failure in immunocompromised patients with acute respiratory failure due to non-hiv pneumocystis pneumonia. J
75 Thorac Dis 2017;9:3013-22. 13. Putinati S, Ballerin L, Piattella M, Panella GL, Potena A. Is it possible to predict the success of non-invasive positive pressure ventilation in acute respiratory failure due to COPD? Respir Med 2000;94:997-1001. 14. John Hall, Arthur Guyton. Guyton and Hall Textbook of Medical Physiology. 11th ed. Philadelphia, Pennsylvania: Saunders; 2005 15. Tokgoz Akyil F, Gunen H, Agca M, Gungor S, Yalcinsoy M, Sucu P, et al. Patient Outcome after Chronic Obstructive Pulmonary Disease Exacerbations Requiring Non-invasive Ventilation during Hospitalization. Arch Bronconeumol 2016;52:470-6. 16. Rhodes A, Evans LE, Alhazzani W, Levy MM, Antonelli M, Ferrer R, et al. Surviving Sepsis Campaign: International Guidelines for Management of Sepsis and Septic Shock: 2016. Intensive Care Med 2017;43:304-77. 17. Beigmohammadi MT, Hussain Khan Z, Samadi S, Mahmoodpoor A, Fotouhi A, Rahimiforoushani A, et al. Role of Hematocrit Concentration on Successful Extubation in Critically Ill Patients in the Intensive Care Units. Anesth Pain Med 2016;6:e32904.