Journal of Critical Care (2012) 27, 434–439

High-flow nasal therapy in adults with severe acuterespiratory infection☆

A cohort study in patients with 2009 influenza A/H1N1v

Jordi Rello⁎, Marcos Pérez, Oriol Roca, Garyphallia Poulakou, Jéssica Souto,César Laborda, Joan Balcells, Joaquim Serra, Joan Ramon MasclansCRIPS investigators

Vall d'Hebron University Hospital, P. Vall d'Hebron 119-129. Annexe 5a planta-AG. E08035, Barcelona, Spain

Cth

B

0h

Keywords:Acute respiratory failure;High-flow nasal cannula;Pandemic 2009 influenzaA/H1N1v;

Severe acute respiratoryinfection (SARI);

Viral pneumonia

AbstractPurpose: The experience with high-flow nasal cannula (HFNC) oxygen therapy in severe acuterespiratory infection (SARI) is limited. The objective was to assess the effectiveness of HFNC oxygentherapy in adult patients with SARI by confirmed 2009 influenza A/H1N1v infection (by real-timereverse transcription polymerase chain reaction testing).Material and Methods: A single-center post hoc analysis of a cohort of intensive care unit patientsadmitted with SARI due to 2009 Influenza A/H1N1v was done. High-flow nasal cannula (Optiflow;Fisher & Paykel, Auckland, New Zealand) was indicated in the presence of acute respiratory failurewhen the patient was unable to maintain a pulse oxymetry more than 92% with more than 9 L/min ofoxygen using a standard face mask conventional delivery systems. Nonresponders were defined by theirneed of subsequent mechanical ventilation.Results: Twenty-five nonintubated adult patients were admitted for SARI (21 pneumonia). Twentywere unable to maintain pulse oxymetry more than 92% with conventional oxygen administration andrequired HFNC O2 therapy, which was successful in 9 (45%). All 8 patients on vasopressors requiredintubation within 24 hours. After 6 hours of HFNC O2 therapy, nonresponders presented a lower PaO2/fraction of inspired oxygen (median, 135 [interquartile range, 84-210] vs 73 [56-81] mm Hg P b .05)and needed higher oxygen flow rate. No secondary infections were reported in health care workers. Nonosocomial pneumonia occurred during HFNC O2 therapy.Conclusion: High-flow nasal cannula O2 therapy appears to be an innovative and effective modality forearly treatment of adults with SARI.© 2012 Elsevier Inc. All rights reserved.

☆ Conflicts of interest: Dr Masclans and Dr Roca have received honoraria as lecture fees from Fisher & Paykel. This study has been funded in part withIBERES funding (PCI Neumonia) and AGAUR (2009SGR01226). Fisher & Paykel did not have access to the manuscript and have not participated neither ine study design or interpretation of results. Dr Pérez, Dr Poulakou, Dr Souto, Dr Laborda, Dr Balcells, and Dr Rello have no conflict of interest to disclose.⁎ Corresponding author. Critical Care Department, Vall d'Hebron University Hospital, P. Vall d'Hebron 119-129. Annexe 5a planta-AG. E08035,

arcelona, Spain. Tel.: +34 932 74 62 09; fax: +34 932 74 60 62.E-mail address: jrello@crips.es (J. Rello).

883-9441/$ – see front matter © 2012 Elsevier Inc. All rights reserved.ttp://dx.doi.org/10.1016/j.jcrc.2012.04.006

435HFNC in adults with severe acute respiratory infection

1. Introduction

Severe acute respiratory infection (SARI) due to pan-demic 2009 influenza A/H1N1v infection was characterizedby a rapid acute respiratory failure (ARF), preceded by 3 to 5days of flu-like symptoms. Data from large cohorts duringthe pandemic period indicated that 30% of all hospital-admitted patients required intensive care unit (ICU)admission with more than 60% of them being subsequentlymechanically ventilated [1,2].

High-flow nasal cannula (HFNC) O2 therapy is thenewest and least known noninvasive option for the clinicalmanagement of patients with ARF. Current heated/humid-ified high flow (up to 50 L/min) devices have achieved agreat degree of efficacy and comfort in delivering high-flowoxygen compared with conventional oxygen therapy [3].Furthermore, improvement in oxygenation with HFNC hasbeen described in patients with mild to severe ARF [3-5]. Forthese reasons and the comfort provided to the patient, HFNCO2 therapy appears to be an innovative modality for earlytreatment of adult patients with severe ARF. Although it hasthe potential to reduce the need of mechanical ventilation(MV), its indications remain speculative. Information onSARI due to influenza A/H1N1v is limited to 5 patients [4].

Our hypothesis was that HFNC O2 therapy could be anefficient noninvasive intervention and that it might alleviatethe need for MV in some patients with SARI. To evaluatethis hypothesis, the clinical course of nonintubated patientswith SARI was recorded, with specific focus on thosereceiving HFNC O2 therapy who were also compared withcontrol patients already mechanically ventilated on admis-sion. Primary end points were the need for MV and ICUmortality. Secondary objectives of the analysis were toidentify a subset of patients most likely to benefit fromHFNC O2 therapy and to anticipate outcomes.

2. Methods

This study represents a post hoc analysis of a prospec-tively assessed cohort of adult patients admitted with ARFdue to 2009 influenza A/H1N1v infection (in the generalICU of Vall d'Hebron University Hospital in Barcelona, alarge tertiary university hospital) from September 1, 2009, toJanuary 31, 2011. Patients reported in this study were alsoreported to a large national registry, which received ethicsboard approval in July 2009 (ref. 07/J23). Informed consentwas waived because of the observational nature of the study.Clinical and epidemiological parameters of this cohort ofadult patients, in comparison with pediatric patients admittedwith the same indication in Vall d'Hebron UniversityHospital, have been reported elsewhere [6].

Inclusion criteria were ARF and laboratory confirmationof 2009 influenza A/H1N1v infection by use of real-time

reverse transcription polymerase chain reaction testing onnasopharyngeal samples or endobronchial secretions [7,8].Methodology is described elsewhere [6]. Exclusion criteriawere age younger than 18 years, hypercapnia, or lack ofconfirmation of 2009 A/H1N1v infection.

Demographic data, comorbidities, clinical and laboratoryfeatures, severity indices, and radiologic findings wererecorded. Clinical and epidemiological criteria were usedaccording to World Health Organization definitions [7,8].The definitions of community-acquired pneumonia, second-ary bacterial pneumonia, and hospital-acquired pneumoniawere based on 2007 American Thoracic Society (ATS) andInfectious Disease Society of America (IDSA) guidelines[9]. Primary viral pneumonia was defined in patientspresenting during the short-term phase of influenza virusillness with adult respiratory distress syndrome and unequiv-ocal alveolar opacification involving 2 or more lobes, withnegative respiratory and blood bacterial cultures. Dailyclinical and radiologic assessment was performed by criticalcare specialists. Hospital-acquired pneumonia was definedaccording with 2005 ATS/IDSA guidelines [10]. Baro-trauma was defined as visible pneumothorax in the chest x-ray. Obese patients were defined as those with body massindex over 30 kg/m2 [11]. Consensus criteria were used forthe definitions of shock and shock requiring administrationof vasopressors [12]. Severity of illness at ICU admissionwas assessed with Acute Physiology and Chronic HealthEvaluation (APACHE) II score [13], which was determinedin all patients within the first 24 hours; organ failure wasassessed using the Sequential Organ Failure Assessment(SOFA) scoring system [14].

Standard and droplet control measures were adopted for allpatients included. Patients were isolated for at least 7 daysafter onset of illness. Health care workers used specialpersonal protective equipment (PPE), such as filtering facepiece (FFP3) masks, gowns, gloves, and eye protection. Aftereach use, PPE was appropriately and safely disposed of afteruse and careful hand washing was performed. Secondaryinfections among health care workers were self-reported.

The HFNC O2 device (Optiflow system, MR850 heatedhumidified RT202 delivery tubing, and RT050/051 nasalcannula; Fisher & Paykel Healthcare Ltd, Auckland, NewZealand) were applied to provide optimal humidity (37°C, 44mg/L) was indicated in the presence of ARF when the patientwas unable to maintain a pulse oxymetry (SpO2) more than92% with more than 9 L/min of oxygen on conventionaloxygen administration using a standard face mask (Oxinova,Carburos Medica, Spain) with a bubble humidificator(Respiflo Water and MN Adapter; Tyco Healthcare,Gosport, United Kingdom). Low-resistance nasal cannulaavailable can deliver up to 50 L/min of totally humidified gasadmixture. The fraction of inspired oxygen (FIO2) and flowrate were adjusted to individual patient needs with a targetSpO2 of 95%. Parameters used to assess respiratory failurewere respiratory rate (RR), PaO2/FIO2 ratio, SpO2, and PaCO2.

436 J. Rello et al.

Parameters used to asses the level of respiratory supportprovided were FIO2 and total delivered flow.

Clinically relevant outcomes were HFNC O2 treatmentfailure and ICU mortality. Criteria for intubation wereinability to correct oxygenation (PaO2 b80 mm Hg withHFNC at 30 L/min and FIO2 of 1), respiratory acidosis(PaCO2 N50 mm Hg with pH b7.30), RR more than 30breaths per minute, cardiac arrest/arrhythmias, shock, andinability to clear secretions.

Data were expressed as median (interquartile range[IQR]) or frequency (percentage). Fisher exact test and χ2

test is used to carry out comparisons between categoricalvariables, as appropriate; comparisons between groups areperformed with 1-way analysis of variance and Mann-Whitney U test. Kaplan-Meier curve was plotted for delayedintubation. P ≤ .05 was considered to be statisticallysignificant. Data analysis was conducted using SPSS 18.0software (SPSS, Inc, Chicago, Ill).

3. Results

Thirty-five adult patients with confirmed 2009 influenzaA/H1N1v infections required ICU admission because ofSARI with severe hypoxemia or acute respiratory distresssyndrome. All patients received oral oseltamivir andparenteral antibiotics. Ten patients (28.6%) already requiredintubation before ICU admission. Data for intubated patientswere compared with patients requiring HFNC O2 treatment.High-flow nasal cannula O2 therapy patients were signifi-cantly younger (37 [29-47] vs 50 [44-56] years, P b .05),

Fig. 1 Flowchart of 35 patients with SARI due to 20

with a lower number of organs affected (2 [1-2] vs 3 [2-4], Pb .05) and lower severity of illness, as measured byAPACHE II score (11 [8-14] vs 25 [13-31], P b .05) andSOFA score (3 [3-5] vs 8 [4-12], P b .01).

Among the cohort of nonintubated patients with SARI,only 5 (20%) maintained an SpO2 more than 92% withconventional oxygen therapy, and 20 (80%) underwentHFNC O2 therapy (Fig. 1). No patient underwent noninva-sive ventilation (NIV). Indeed, in these 5 patients who weresuccessfully treated with conventional oxygen therapy,baseline total delivered gas flow was significantly lower(15 [11-18] vs 30 [20-30] L/min, P = .02) with the sameestimated FIO2 (1 [0.5-1] vs 1 [0.7-1], P = .76) to maintainsimilar oxygen saturation measured by SpO2 (97% [96%-98%] vs 96% [90%-97%], P = .23) on ICU admission.These 5 patients were discharged without intubation. AtICU admission, those patients treated with successfulconventional oxygen therapy tended to have lower medianRR (24 [22-26] vs 31 [24-34] breaths per minute, P = .09)compared with those who further required treatment withHFNC O2 therapy.

Baseline characteristics of the overall population withARF and the study cohort are presented in Table 1.Median flow used was 30 L/min. Among the 20 patientswith HFNC, 18 had ARF caused by viral pneumonia (5with bacterial coinfection), and HFNC was successful in 7(39%). The remaining 2 presented SARI without chest x-ray opacities, and HFNC was successful in both. High-flow nasal cannula was set with a median FIO2 of 1 (0.75-1) and flow of 30 L/min (20-33 L/min). A successfuloutcome of HFNC treatment was achieved in 9 patients(45%) of the total cohort 20 that received this modality of

09 influenza A H1N1v requiring ICU admission.

Table 1 Baseline characteristics of the whole populationincluded and 25 nonintubated patients with 2009 influenzaA/H1N1v infection at ICU admission

Total(N = 35)

Nonintubated(n = 25)

Age (y) 43 (33-53) 38 (30-51)Sex (male) 19 (52.8 %) 14 (56%)ComorbiditiesObesity (BMI N30 kg/m2) 8 (22.9%) 5 (20%)Chronic lung disease 11 (31.4%) 9 (36%)Immune suppression 3 (8.6%) 3 (12%)Pregnancy 3 (8.6%) 2 (8%)Time from symptom onsetto ICU admission (d)

5 (3-7) 5 (2-7)

At the ICU admissionAPACHE II score 12.5 (8.75-24.25) 10.5 (7-16.25)SOFA 4 (3-5.75) 3 (3-4.5)Shock 15 (42.9%) 9 (36%)No. of organs affected 2 (1-2.75) 1 (1-2)Pneumonia 20 (57%) 14 (56%)Bacterial coinfection 10 (28.6%) 7 (28%)ICU LOS (d) 10 (3-27.75) 7 (3-31.5)Hospital LOS (d) 12 (8-32.25) 10 (7.25-53.5)ICU mortality 5 (14.3%) 3 (12%)

Data are expressed as median (IQR) or frequency (percentage). BMIindicates body mass index; LOS, length of stay.

Fig. 2 Kaplan-Meier curve cumulative probability for intubationin failing patients with HFNC.

437HFNC in adults with severe acute respiratory infection

therapy (Fig. 1). Table E1 presents a comparativedescription of clinical and epidemiological characteristicsof patients successfully treated and failing on HFNC.Baseline characteristics with statistically significant differ-ences were as follows: (1) the presence of shock requiringadministration of vasopressors was a significant determi-nant of nonresponders (72.8% vs 0%, P b .01) and (2)SOFA and APACHE II scores at ICU admission, with aSOFA score value of 4 or more (72.7% vs 11.1%, P b .05)and an APACHE II score 12 or more (60% vs 11.1%,P b .05), being associated with failure of HFNC. All

Table 2 Respiratory parameters 6 hours after initiation of HFNCin 20 patients with SARI

Patients withHFNC success(n = 9)

Patients withHFNC failure(n = 11)

RR (breaths/min) 21 (20-25) 21 (12-31)PaO2/FIO2(mm Hg) ⁎

135 (84-210) ⁎ 73 (56-81) ⁎

SpO2 (%) 98 (97-98) 94 (91-99)PaCO2 (mm Hg) 38 (36-38) 37 (28-43)Total deliveredflow (L/min) ( ⁎⁎)

25 (20-30) 30 (30-35)

Data are expressed as median (IQR) or frequency (percentage).⁎ P b .05 for the comparison between patients with successful

HFNC vs failure of HFNC.⁎⁎ P = .12 for the comparison between patients with successful

HFNC vs failure of HFNC.

patients treated with HFNC O2 therapy who requiredvasopressors needed intubation and MV (Table E2).Indeed, for HFNC patients with vasopressors, therelative risk ratio for MV (HFNC failure) was 4 (95%confidence interval, 1.5-10.6).

In contrast, the presence of chronic respiratory diseasessuch as asthma or chronic obstructive lung disease (COPD)was more prevalent in responders (55.5% vs 0%, P b .01)(Table E1). Among the 5 patients with chronic respiratorydiseases, 2 of them had exacerbation of the underlyingrespiratory disease without chest x-ray opacities, 1 had 3quadrants affected, and the remaining 2 patients had 4 chestx-ray quadrants affected.

Six hours after HFNC treatment onset, patients who didnot respond to HFNC therapy tended to require 20% higherdelivered flows (Table 2). Moreover, the PaO2/FIO2 ratio wassignificantly ameliorated in patients with successful HFNCcompared with patients with ultimate HFNC failure (135[84-210] vs 73 [56-81] mm Hg, P b .05). Intolerance wasnever a cause of HFNC cessation. No other side effects werereported, including secondary infections among health careworkers. None of the patients with HFNC developedhospital-acquired pneumonia during HFNC.

All patients with successful HFNC survived, whereasHFNC failure was associated with an ICU mortality of27.3%. The median (IQR) period to MV in patients whofailed HFNC oxygen treatment was 18 hours (6-38 hours).Intubation was required within 48 of ICU admission in allpatients with HFNC failure (Fig. 2), but the relative delay inintubation was not associated with different mortality.

4. Discussion

This is the first study reporting HFNC used in a series ofadult patients as early therapy for severe ARF caused by

438 J. Rello et al.

2009 influenza A/H1N1v infection. Our findings suggestthat all patients with inotropes required intubation.Furthermore, patients without cardiovascular compromiseand an RR above 30 breaths per minute on conventionaloxygen therapy, upon ICU admission, may benefit of aHFNC trial. Six hours after onset of treatment, failingpatients demonstrated worse oxygenation (PaO2/FIO2 b100mm Hg) and required 20% higher oxygen flow. Patientswith chronic respiratory comorbidities also benefit fromHFNC O2 therapy.

High-flow nasal cannula O2 therapy failure was notassociated with complications, such as pneumonia orbarotrauma. Furthermore, HFNC O2 therapy was not relatedwith any secondary infections among health care workers.However, no follow-up was performed, including serologiesof all health care workers in charge of these patients. Ascould happen with other techniques, such as high-frequencyoscillatory ventilation, that have been used in patients withH1N1 [15], this study demonstrates that HFNC therapy is notharmful when it is applied using adequate PPE and carefulhand washing.

Moreover, it should be taken into account that HFNC hasbeen previously associated with greater comfort and lessdegree of dyspnea and mucosal drying compared withconventional oxygen therapy [3]. Apart from correction ofhypoxemia and reduction of RR, delivering of adequatelyhumidified gas may play a role in these results [16].Furthermore, nasal cannula allows nasogastric tube feedingand does not affect phonation, physiotherapy, and coughing.

There are preliminary reports about the use of HFNC inneonates and children [17], but indications in adults remainunclear. In a recent randomized, comparative studyperformed in a cardiothoracic and vascular ICU, patientswith mild ARF receiving HFNC had significantly fewerhypoxemic episodes and tended to need less noninvasiveventilation [5]. Patients with severe ARF may also takebenefit from HFNC O2 therapy, manifested as improvedoxygenation and a reduction in RR without hypercapnicacidosis [3]. More recently, HFNC was associated with abeneficial effect on clinical signs and oxygenation in a cohortof patients with ARF [4]. High-flow nasal cannula O2

therapy improves oxygenation by various mechanisms, suchas decreasing oxygen dilution, reducing dead space, andincreasing end-expiratory lung volume and tidal volume[18,19].

Patients with an RRmore than 30 beats per minute may bean early indicator for the early use of HFNC O2 therapy. Ourfindings suggest that early indication of HFNC O2 therapymay be based on RR, and it should be explored in amulticenter, randomized clinical trial. Furthermore, patientsavoiding intubation can be identified early. These patientshad higher PaO2/FIO2 ratio values after 6 hours of treatmentwith HFNC O2 therapy. Similar results were reported in arecently published observational study [4] that included 38patients with ARF due to different organisms. Indeed, 2important issues should be taken into account. First, the

clinical presentation and course of influenza A/H1N1vpneumonia is significantly different from interpandemiccommunity-acquired pneumonia. Second, only 5 patientswith ARF due to 2009 influenza A/H1N1v infection wereincluded. For these reasons, some other factors that maypredict the need ofMV could be described in this subgroup ofpatients with ARF. In this sense, patients with lower oxygenflow needs were more likely to avoid intubation. On the otherhand, all patients with shock were HFNC nonresponders andrequired intubation. Thus, patients on inotropes do not benefitfrom this therapy and require immediate intubation.

A subset most likely to benefit from HFNC oxygentherapy was patients with chronic respiratory comorbidities,such as asthma or COPD, regardless of the type of clinicalpresentation. In fact, all 5 patients with this comorbidity whostarted HFNC responded successfully. However, none ofthem presented with hypercapnic acidosis. Although benefitsof administering an adequately humidified gas could beinteresting in these patients, HFNC should be carefullyapplied, probably using lower flow rates than in otherpatients, trying to avoid hyperinflation, and, in case ofpatients with COPD, not correcting their baseline hypox-emia. In this sense, HFNC has been previously safely used instable patients with COPD with severe air entrapmentdemonstrating an improvement in oxygenation and exerciseperformance [20].

Limitations of this study are that it is an observationalsingle-center study in adult patients with SARI not specificallydesigned to analyze the effect of HFNC O2 therapy; this is apost hoc analysis, and no baseline gas exchange values wereavailable at admission for all patients. However, the RR hasbeen shown to be a sensitive indicator for patients who wouldbenefit from HFNC O2 therapy for influenza A/H1N1vinfection. This is consistent with a preliminary report insuggesting that HFNCmight reduce intubation rate in childrenwith viral bronchiolitis [21]. Sample size is small, and a type IIerror cannot be ruled out. In spite of this, many comparisonswere already statistically significant. Despite of limitations,this is the largest study describing the clinical impact of HFNCon SARI caused by influenza A/H1N1v infection. Arandomized clinical trial for the use of HFNC O2 therapy inARF should be encouraged, but in reality, it may not beavailable soon. Meanwhile, our findings suggest that HFNCO2 therapy appears to be an innovative and effective modalityfor early treatment of adults with SARI.

Supplementary materials related to this article can befound online at http://dx.doi.org/10.1016/j.jcrc.2012.04.006.

Appendix. List of CRIPS investigators

Roser AnglèsElisabet GallartRosa Maria GraciaMercedes PalomarIsabel Porta

439HFNC in adults with severe acute respiratory infection

Maria Alba RieraJudith SacanellTeresa PontBárbara BorgattaSimone GattarelloAna Parra CastilloPurificación PérezAlejandra García-RocheElisabeth PapiolAna Sánchez-CorralJordi Riera del BríoElsa Sofía Da Palma AfonsoDavid TéllezLaura RuanoSofía Pérez-Hortiguela

References

[1] Rodríguez A, Lisboa T, Rello J. Pandemic influenza A (H1N1)v in theintensive care unit: what have we learned? Arch Bronconeumol2010;46(Suppl. 2):24-31.

[2] Rodríguez A, Martin-Loeches I, Bonastre J, Olaechea P, Álvarez-Lerma F, Zaragoza R, et al. First influenza season after the 2009pandemic influenza: report of the first 300 ICU admissions in Spain.Med Intensiva 2011;35:208-16.

[3] Roca O, Riera J, Torres F, Masclans JR. High-flow oxygen therapy inacute respiratory failure. Respir Care 2010;55:408-13.

[4] Sztrymf B, Messika J, Bertrand F, Hurel D, Leon R, Dreyfuss D, et al.Beneficial effects of humidified high flow nasal oxygen in critical carepatients: a prospective pilot study. Intensive Care Med 2011;37:1780-6.

[5] Parke RL, McGuinness SP, Eccleston ML. A preliminary randomizedcontrolled trial to assess effectiveness of nasal high-flow oxygen inintensive care patients. Respir Care 2011;56(3):265-70.

[6] Poulakou G, Souto J, Balcells J, Pérez M, Laborda C, Roca O, et al.First influenza season after the 2009 pandemic influenza: character-istics of ICU admissions in adults and children in Vall d'HebronHospital. Clin Microbiol Infect 2012;18:374-80.

[7] CDC protocol of real-time RTPCR for influenza A (H1N1). Genève:World Health Organization; 2009. http://www.who.int/csr/resources/publications/swineflu/realtimeptpcr/en/index.html (last accessed 01March 2011).

[8] Guidance for clinicians on the use of rapid Influenza diagnostic testsfor the 2010–2011 Influenza season. http://www.cdc.gov/flu/professionals/diagnosis/clinician_guidance_ridt.htm.

[9] Mandell LA,Wunderink RG, Anzueto A, Bartlett JG, Campbell GD, DeanNC, et al. Infectious Diseases Society of America; American ThoracicSociety: Infectious Diseases Society of America/American ThoracicSociety Consensus Guidelines on the management of community-acquiredpneumonia in adults. Clin Infect Dis 2007;44(Suppl. 2):S27-72.

[10] American Thoracic Society, Infectious Diseases Society of America.Guidelines for the management of adults with hospital-acquired,ventilator-associated, and healthcare-associated pneumonia. Am JRespir Crit Care Med 2005;171:388-416.

[11] Díaz E, Rodríguez A, Martin-Loeches I, Lorente L, del Mar Martín M,Pozo JC, et al. Impact of obesity in patients infected with 2009influenza A (H1N1). Chest 2011;139:382-6.

[12] Calandra T, Cohen J. International Sepsis Forum Definition ofInfection in the ICU Consensus Conference. The international sepsisforum consensus conference on definitions of infection in the intensivecare unit. Crit Care Med 2005;33(7):1538-48.

[13] Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: aseverity of disease classification system. Crit Care Med 1985;13:818-29.

[14] Moreno R, Takala J, Willatts S, De Mendonça A, Bruining H, et al.The SOFA (Sepsis-Related Organ Failure Assessment) score todescribe organ dysfunction/failure. Intensive Care Med 1996;22:707-10.

[15] Ramsey CD, Funk D, Miller III RR, Kumar A. Ventilator managementfor hypoxemic respiratory failure attributable to H1N1 novel swineorigin influenza virus. Crit Care Med 2010;38(4 Suppl):e58-65.

[16] Chanques G, Constantin JM, Sauter M, Jung B, Sebbane M, Verzilli D,et al. Discomfort associated with under-humidified high-flow oxygentherapy in critically ill patients. Intensive Care Med 2009;35:996-1003.

[17] Spentzas T, Minarik M, Patters AB, Vinson B, Stidham G. Childrenwith respiratory distress treated with high-flow nasal cannula. JIntensive Care Med 2009;24:323-8.

[18] Corley A, Caruana LR, Barnett AG, Tronstad O, Fraser JF. Oxygendelivery through high-flow nasal cannulae increase end-expiratorylung volume and reduce respiratory rate in post-cardiac surgicalpatients. Br J Anaesth 2011;107:998-1004.

[19] Dysart K, Miller TL, Wolfson MR, Shaffer TH. Research in high flowtherapy: mechanisms of action. Respir Med 2009;103:1400-5.

[20] Chatila W, Nugent T, Vance G, Gaughan J, Criner GJ. The effects ofhigh-flow vs low-flow oxygen on exercise in advanced obstructiveairways disease. Chest 2004;126:1108-15.

[21] Schibler A, Pham TM, Dunster KR, Foster K, Barlow A, Gibbons K,et al. Reduced intubation rates for infants after introduction of high-flownasal prong oxygen delivery. Intensive Care Med 2011;37:847-52.