High-Flow Oxygen Therapy in Acute Respiratory Failure

Abstract

To compare the comfort of oxygen therapy via high-flow nasal cannula (HFNC) versus via conventional face mask in patients with acute respiratory failure. Acute respiratory failure was defined as blood oxygen saturation < 96% while receiving a fraction of inspired oxygen > or = 0.50 via face mask.
Oxygen was first humidified with a bubble humidifier and delivered via face mask for 30 min, and then via HFNC with heated humidifier for another 30 min. At the end of each 30-min period we asked the patient to evaluate dyspnea, mouth dryness, and overall comfort, on a visual analog scale of 0 (lowest) to 10 (highest). The results are expressed as median and interquartile range values.
We included 20 patients, with a median age of 57 (40-70) years. The total gas flow administered was higher with the HFNC than with the face mask (30 [21.3-38.7] L/min vs 15 [12-20] L/min, P < .001). The HFNC was associated with less dyspnea (3.8 [1.3-5.8] vs 6.8 [4.1-7.9], P = .001) and mouth dryness (5 [2.3-7] vs 9.5 [8-10], P < .001), and was more comfortable (9 [8-10]) versus 5 [2.3-6.8], P < .001). HFNC was associated with higher P(aO(2)) (127 [83-191] mm Hg vs 77 [64-88] mm Hg, P = .002) and lower respiratory rate (21 [18-27] breaths/min vs 28 [25-32] breaths/min, P < .001), but no difference in P(aCO(2)).
HFNC was better tolerated and more comfortable than face mask. HFNC was associated with better oxygenation and lower respiratory rate. HFNC could have an important role in the treatment of patients with acute respiratory failure.

Full text

High-Flow Oxygen Therapy in Acute Respiratory Failure
Oriol Roca MD, Jordi Riera MD, Ferran Torres MD PhD, and Joan R Masclans MD PhD
OBJECTIVE: To compare the comfort of oxygen therapy via high-flow nasal cannula (HFNC)
versus via conventional face mask in patients with acute respiratory failure. Acute respiratory
failure was defined as blood oxygen saturation <96% while receiving a fraction of inspired oxygen
>0.50 via face mask. METHODS: Oxygen was first humidified with a bubble humidifier and
delivered via face mask for 30 min, and then via HFNC with heated humidifier for another 30 min.
At the end of each 30-min period we asked the patient to evaluate dyspnea, mouth dryness, and
overall comfort, on a visual analog scale of 0 (lowest) to 10 (highest). The results are expressed as
median and interquartile range values. RESULTS: We included 20 patients, with a median age of
57 (40 –70) years. The total gas flow administered was higher with the HFNC than with the face
mask (30 [21.3–38.7] L/min vs 15 [12–20] L/min, P<.001). The HFNC was associated with less
dyspnea (3.8 [1.3–5.8] vs 6.8 [4.1–7.9], Pⴝ.001) and mouth dryness (5 [2.3–7] vs 9.5 [8 –10],
P<.001), and was more comfortable (9 [8 –10]) versus 5 [2.3– 6.8], P<.001). HFNC was associated
with higher P
aO
2
(127 [83–191] mm Hg vs 77 [64 – 88] mm Hg, Pⴝ.002) and lower respiratory rate
(21 [18 –27] breaths/min vs 28 [25–32] breaths/min, P<.001), but no difference in P
aCO
2
. CON-
CLUSIONS: HFNC was better tolerated and more comfortable than face mask. HFNC was asso-
ciated with better oxygenation and lower respiratory rate. HFNC could have an important role in
the treatment of patients with acute respiratory failure. Key words: acute respiratory failure; respi-
ratory insufficiency; treatment; oxygen therapy; high-flow nasal cannula. [Respir Care 2010;55(4):408 –
413. © 2010 Daedalus Enterprises]
Introduction
Administration of oxygen is an essential support mea-
sure to maintain proper tissue oxygenation in patients with
acute respiratory failure (ARF). The conventional oxygen
delivery methods have been nasal prongs, nose mask, and
face mask, but the oxygen provided by the conventional
systems may not suffice in patients with ARF.
High-flow oxygen (up to 50 L/min) via nasal cannula,
combined with a heated humidification system, may have
several advantages. High-flow nasal cannula (HFNC) can
decrease oxygen dilution,
1-3
reduce respiratory dead space,
2
and generate some positive airway pressure, because of the
SEE THE RELATED EDITORIAL ON PAGE 485
expiratory resistance generated by the continuous high flow
delivered.
2,3
The heated humidification can facilitate se-
cretion clearance and decrease the development of bron-
chial hyper-response symptoms.
4
HFNC has been evalu-
ated in healthy subjects,
3
patients with stable chronic
obstructive pulmonary disease,
2
and in patients after ab-
dominal surgery.
1
To our knowledge, the experience with
HFNC in patients with ARF is limited to sporadic cases.
5,6
Oriol Roca MD, Jordi Riera MD, and Joan R Masclans MD PhD are
affiliated with the Intensive Care Medicine Department, Hospital Uni-
versitari Vall d’Hebron (A
`rea General), Barcelona, Spain. Oriol Roca
MD is also affiliated with the Respiratory Medicine Department, Hos-
pital Universitari Vall d’Hebron, Barcelona, Spain, and with Centro de
Investigacio´n Biome´dica en Red de Enfermedades Respiratorias
(CIBERES), Instituto de Salud Carlos III, Spain. Ferran Torres MD PhD
is affiliated with Laboratori de Bioestadística i Epidemiologia, Univer-
sitat Auto` noma de Barcelona, and with Servei de Farmacología Clínica,
Institut d’Investigacions Biome` diques August Pi i Sunyer (IDIBAPS),
Hospital Clínic, Barcelona, Spain.
Dr Roca and Dr Masclans have disclosed relationships with Fisher &
Paykel Healthcare. Dr Riera and Dr Torres have disclosed no conflicts of
interest.
Correspondence: Oriol Roca MD, Servei de Medicina Intensiva (A
`rea
General), Hospital Universitari Vall d’Hebron, Passeig Vall d’Hebron
119-129, 08035 Barcelona, Spain. E-mail: oroca@ir.vhebron.net.
408 RESPIRATORY CARE •APRIL 2010 VOL 55 NO4

Thus, the primary aim of this study was to compare the
subjective comfort perceived by ARF patients receiving
oxygen via face mask versus via HFNC, as measured with
a visual analog scale. We previously reported this study in
an abstract.
7
Methods
The manufacturer of the HFNC system we used did not
provide any form of support or consultation in the design,
conduct, or writing of this study.
Study Population
We included 20 patients with ARF, defined as a blood
oxygen saturation (measured via pulse oximetry [S
pO
2
])
⬍96% while receiving humidified oxygen via face mask,
with a fraction of delivered oxygen (F
DO
2
)ofⱖ0.5. The
exclusion criteria were: unstable clinical status (eg, marked
changes in respiratory variables in the 60 min prior to
inclusion in the study); need for endotracheal intubation;
decreased consciousness (Glasgow coma scale ⬍14); se-
vere hemodynamic instability despite fluid therapy and
vasopressors; severe failure of ⬎2 organs apart from re-
spiratory failure; pregnancy; and lack of patient coopera-
tion.
The study was approved by the ethics committee of our
center, and patients gave their informed consent to partic-
ipate.
Study Design
This was a prospective, comparative study of sequential
interventions. The patients were resting in bed in partial
sitting position, and we recorded baseline respiratory vari-
ables. The oxygen was humidified with a bubble humidi-
fier (Respiflo Water and MN Adapter, Tyco Healthcare,
Gosport, United Kingdom) and administered via standard
face mask (Oxinova, Carburos Medica, Spain) for 30 min,
during which the patient was allowed to use a combination
of face mask and conventional nasal cannula to increase
the total oxygen flow, in which case the total flow deliv-
ered was the sum of the flow from the face mask and
cannula.
We then delivered the oxygen via HFNC (Optiflow,
Fisher & Paykel, Auckland, New Zealand) (Fig. 1) at an
initial flow of 20 –30 L/min, with an F
DO
2
identical to that
with the face mask. The total flow was increased in each
patient according to the criteria of the attending physician.
The same F
DO
2
was maintained over the entire testing
period.
For both face mask and HFNC, F
DO
2
was estimated
following the manufacturer’s specifications.
At the end of each 30-min period we asked the patient
for his or her subjective assessment of the device (see
below) and took a sample of arterial blood. Following the
two 30-min study periods we asked the patient which of
the oxygen systems he or she wanted to continue using.
Variables Measured
We recorded the Sequential Organ Failure Assessment
8
at the time of enrollment, and the Acute Physiology and
Chronic Health Evaluation score
9
at intensive care unit
(ICU) admission. We defined multiple-organ dysfunction
syndrome as failure of ⬎2 organs. We estimated the
degree of radiologic involvement as the number of af-
fected quadrants on the chest radiograph.
10
Throughout the
study periods we monitored electrocardiography variables,
blood pressure, respiratory variables, and adverse events
related to the use of the oxygen therapy systems.
We asked each patient to subjectively evaluate the 2
oxygen therapy methods, after 30 min of use in stable
conditions. With a visual analog scale that ranged from 0
(lowest) to 10 (highest), we asked the patient to rate his or
her dyspnea, mouth dryness, and overall comfort.
11
We also measured arterial blood gas values, acid-base
balance, respiratory rate, and S
pO
2
.
Statistical Analysis
The estimated sample size was 20 patients to achieve a
statistical power of 80% for detecting that an observation
Fig. 1. Diagram of the high-flow nasal cannula. A: Compressed air
source. B: Oxygen source. C: Heated humidifier. D: Nasal cannula.
HIGH-FLOW OXYGEN THERAPY IN ACUTE RESPIRATORY FAILURE
RESPIRATORY CARE •APRIL 2010 VOL 55 NO4 409

in one group was lower than in the other, at a probability
of 0.756, with the Mann-Whitney non-parametric test for
independent data, with a protection against type 1 error of
5% bilaterally. Since the study design is actually for de-
pendent data, it was expected that the statistical power
would be higher, according to the correlation found.
12
The results are expressed as the median and interquar-
tile range values, or frequencies (percentage). We com-
pared the results with the Wilcoxon non-parametric test.
The analyses were conducted with statistics software (SPSS
version 13.0, SPSS, Chicago, Illinois). Significance was
set at a 2-tailed Pvalue of .05.
Results
General Characteristics
Twenty patients were included. Fourteen (70%) were
men. The median age was 57 (40 –70) years. Table 1 shows
the cohort’s baseline characteristics. Nineteen patients
(95%) were admitted to the ICU due to hypoxemic ARF.
The median duration of ARF before inclusion in the study
was 4 (3– 8) days.
Subjective Assessment and Respiratory Variables
The results of the subjective assessment show a signif-
icant improvement in the 3 variables analyzed with the use
of HFNC (Table 2). In the evaluation of respiratory vari-
ables, there were no significant differences between the
baseline values and those at the end of 30 min of face
mask use. However, with HFNC there was a significant
increase in P
aO
2
and a reduction in respiratory rate (Fig. 2),
without hypercapnia or acidosis. Arterial blood samples
were not obtained from the 4 patients who did not have
arterial catheters. There was no significant change in mean
arterial pressure or heart rate. Once the testing period had
been completed and the patients were asked to choose
which of the 2 oxygen therapy systems they wanted to
continue using, all chose the HFNC.
Adverse Effects
Five patients (25%) reported some mild adverse effects
that may have been related to the HFNC. The most com-
mon (3 patients) was a sensation of cervical-thoracic dis-
comfort that appeared during the initial period of increas-
ing flow, and immediately disappeared when flow was
decreased. Another patient said that the gas temperature
was too high. In all the latter cases the effects appeared in
the initial few minutes of the testing period and had com-
pletely disappeared by the end of the testing period. The
other patient who reported an adverse effect had nonspecific
nasal discomfort, and nasal mucosal lesions were detected
during HFNC use. On the other hand, mucosal lesions were
seen before HFNC in another patient, probably related to the
previous use of conventional nasal cannula.
Discussion
This study presents the first results of applying humid-
ified HFNC in patients with ARF. HFNC was associated
with less dyspnea and mouth dryness, and greater overall
comfort.
The dyspnea decrease may be due to several factors.
With HFNC the higher flow delivered, the correction of
hypoxemia, and the reduction in the respiratory rate may
play an important role. Second, although there was signif-
icantly less mouth dryness, this variable was likely to have
continued improving after the 30 min of the study, because
30 min may have been too short to determine the true
effects of the heated humidification system on mouth dry-
ness. Third, the patients found HFNC significantly more
comfortable, and there may be several reasons for this.
The above-mentioned benefits surely played a part. Under-
humidified oxygen therapy is also associated with discom-
fort in critically ill patients.
13
That HFNC does not affect
speaking and allows food ingestion could also play a role.
This increased comfort is consistent with the fact that all
the patients chose to continue with HFNC after having
tried it.
Table 1. Baseline Characteristics
Patients (n)20
Age (median and IQR y) 57 (40–70)
Male (n, %) 14 (70)
APACHE II score (median and IQR) 14.0 (10.5–17.5)
SOFA score (median and IQR) 4.0 (3.2–6.0)
Multiple organ dysfunction syndrome (n, %) 4 (20)
Intrapulmonary ARF (n, %) 17 (85)
Extrapulmonary ARF (n, %) 3 (15)
Etiology of ARF (n,%)
Pneumonia 13 (65)
Extrapulmonary ALI 3 (15)
Acute lung-graft rejection 1 (5)
Severe asthma exacerbation 1 (5)
COPD exacerbation 1 (5)
Cardiogenic lung edema 1 (5)
Duration of ARF before inclusion
(median and IQR)
4 (3–8)
Number of affected quadrants on chest
radiograph (median and IQR)
3 (2–4)
APACHE II ⫽Acute Physiology and Chronic Health Evaluation II
SOFA ⫽Sequential Organ Failure Assessment
ARF ⫽acute respiratory failure
ALI ⫽acute lung injury
COPD ⫽chronic obstructive pulmonary disease
HIGH-FLOW OXYGEN THERAPY IN ACUTE RESPIRATORY FAILURE
410 RESPIRATORY CARE •APRIL 2010 VOL 55 NO4

We assessed the variables studied with visual analog
scales. These scales are solid, well recognized systems
for measuring patients’ perceptions of variables such as
dyspnea.
14-18
In fact, the visual analog scale is consid-
ered the most appropriate instrument for repeated mea-
surement of a subjective variable in a given patient, as
it is sufficiently sensitive to detect changes occurring
minute by minute.
19
The effect of HFNC on oxygenation is equally im-
portant. As compared to the conventional oxygen mask,
we found improvement in oxygenation after 30 min of
HFNC. We did not measure fraction of inspired oxygen
(F
IO
2
) because that would have required placing a pha-
ryngeal catheter, which might have generated artifacts
in the subjective evaluations. However, the oxygenation
improvement with HFNC might have been due to a
higher F
IO
2
achieved with the higher flow with HFNC.
At low flow the oxygen is diluted with ambient air
because the oxygen flow is lower than the patient’s
inspiratory flow demand. Wettstein et al
20
measured the
F
IO
2
in the pharynx during nasal cannula at low flow
(1– 6 L/min) and high flow (6 –15 L/min). There was a
progressive increase in pharyngeal F
IO
2
with each in-
crease of oxygen flow. That study analyzed only flows
up to 15 L/min, but the present study shows how this
effect may persist and possibly increase with higher
flow in patients with ARF.
Nevertheless, with a given F
IO
2
, other factors may
contribute to improved oxygenation with HFNC. First,
continuous high flow can wash-out exhaled air from the
upper airway and thus reduce dead space.
2
Second, na-
sal cannula seems to be better tolerated than face mask,
which has been reported to improve patient adherence
to therapy and oxygenation.
21
Third, the high flow ex-
erts a continuous positive airway pressure in both healthy
Table 2. Subjective Evaluation and Respiratory, Hemodynamic, and Gas-Exchange Data*
Face Mask HFNC P
Subjective Evaluation
Dyspnea 6.8 (4.1–7.9) 3.8 (1.3–5.8) .001
Mouth dryness 9.5 (8.0–10.0) 5 (2.3–7.0) ⬍.001
Overall comfort 5.0 (2.3–6.8) 9.0 (8.0–10.0) ⬍.001
Respiratory and Gas-Exchange Variables
Total oxygen flow (L/min) 15.0 (12.0–20.0) 30.0 (21.3–38.7) ⬍.001
Fraction of delivered oxygen 1.0 (0.8–1.0) 1.0 (0.8–1.0) .32
Respiratory rate (breaths/min) 28 (25–32) 21 (18–27) ⬍.001
pH
†
7.42 ⫾7.38–7.47 7.44 (7.38–7.50) .06
P
aO
2(mm Hg)
†
77 (64–88) 127 (83–191) .002
P
aCO
2(mm Hg)
†
37 (33–45) 37 (32–43) .51
HCO
3
–
(mmol/L)
†
25.0 (22.1–28.5) 24.5 (22.2–29.1) .09
Base excess (mmol/L)
†
1.0 (⫺2.3–4.8) –1.0 (–2.3 to 5.3) .055
S
pO
2(%) 95 (91–97) 98 (96–99) .002
Hemodynamic Variables
Mean arterial pressure (mm Hg) 87 (76–94) 86 (71–93) .36
Heart rate (beats/min) 94 (77–112) 85 (73–108) ⬎.99
* All values are median and interquartile range.
† Only in patients with an arterial catheter.
HFNC ⫽high-flow nasal cannula
SpO2⫽blood oxygen saturation measured via pulse oximetry
Fig. 2. Change in respiratory rate with change from face mask to
high-flow nasal cannula (HFNC).
HIGH-FLOW OXYGEN THERAPY IN ACUTE RESPIRATORY FAILURE
RESPIRATORY CARE •APRIL 2010 VOL 55 NO4 411

volunteers and patients with chronic obstructive pulmo-
nary disease,
3
who additionally showed an increase in
exercise capacity. The airway pressure generated is di-
rectly related to the flow administered; tracheal pressure
of up to 5 cm H
2
O has been recorded with a flow of
50 L/min.
3
In the present study, however, airway pres-
sure was not measured. Moreover, the reported effect
was observed with higher flows than we used in the
majority of patients in this study. Thus, it is unlikely
that a continuous positive airway pressure effect would
have contributed to improved oxygenation in our pa-
tients.
Also, the heated humidifier system may indirectly
affect oxygenation. Active humidification improves
mucociliary function, facilitates secretion clearance,
and decreases atelectasis formation, which improves
ventilation-perfusion ratio and oxygenation. However,
in the present study it is unlikely that those factors
could have affected oxygenation after only 30 min. On
the other hand, heated humidification systems may at-
tenuate the development of bronchial hyper-response
symptoms,
4
so humidified HFNC may be particularly
beneficial in patients with respiratory infection, chronic
obstructive pulmonary disease, or bronchial asthma.
The pronounced reduction in respiratory rate in the
present study is also important, particularly because it
was not associated with changes in the P
aCO
2
or pH. This
finding is consistent with other reported data
1,2
and may
have contributed to the improvement in dyspnea and
comfort.
There were some adverse effects with HFNC, but
several points should be taken into account when inter-
preting these results. First, all the adverse effects from
HFNC were mild. Second, the sensation of cervical-
thoracic discomfort disappeared when the flow was de-
creased. In addition, outside the testing period we found
that higher flows could be achieved with HFNC with no
cervical-thoracic discomfort. Similarly, the feeling of
excessive gas temperature, reported by one patient, also
disappeared during the testing period. Third, some of
the patients received oxygen at the beginning of the
study with a combination of face mask and conventional
nasal cannula. Therefore, the nasal mucosal lesion dis-
covered during HFNC could have been the result of an
additive effect of both systems, as suggested by the fact
that such a lesion was observed in one patient before
HFNC. Therefore, we believe that HFNC could be con-
sidered comfortable, especially given that after the test-
ing period all the patients chose to continue with HFNC.
Limitations
First, we did not use a randomized design; instead, pa-
tients were allowed to choose which of the 2 oxygen ther-
apy systems they wanted to continue with. Second, we
could not quantify the relative contribution of each factor
involved in the HFNC system (nasal cannula, high flow,
warm gas, humidification) to the results obtained. Third, in
an attempt to avoid generating artifacts in the subjective
evaluation, we did not measure the F
IO
2
. Fourth, the pa-
tients could not be blinded to the systems being tested.
Nevertheless, these limitations do not detract from the
favorable results obtained with HFNC in patients with
ARF in a study whose main objective was to compare the
comfort of the 2 oxygen systems.
Conclusions
HFNC can improve the clinical tolerance of oxygen
therapy in patients with ARF. In addition to the subjective
benefit, our patients showed improved oxygenation and a
reduced respiratory rate. Thus, we consider this HFNC
system useful for integrated treatment of ARF, although
further investigation is needed to determine the clinical
scenarios in which its benefits will have the greatest im-
pact.
REFERENCES
1. Lund J, Holm-Knudsen RJ, Nielsen J, Føge Jensen PB. [Nasal cath-
eters versus Hudson face mask in oxygen therapy]. Ugeskr Laeger
1996;158(28):4077-4079. Article in Danish.
2. Chatila W, Nugent T, Vance G, Gaughan J, Criner GJ. The effects of
high-flow vs low-flow oxygen on exercise in advanced obstructive
airways disease. Chest 2004;126(4):1108-1115.
3. Williams AB, Ritchie JE, Gerard C. Evaluation of a high-flow nasal
oxygen delivery system: gas analysis and pharyngeal pressures. In-
tensive Care Med 2006;32(S1):S219.
4. Moloney E, O’Sullivan S, Hogan T, Poulter LW, Burke CM. Airway
dehydration: a therapeutic target in asthma? Chest 2002;121(6):1806-
1811.
5. Lomas C, Roca O, A
´lvarez A, Masclans JR. Fibroscopy in patients
with hypoxemic respiratory insufficiency: utility of the high-flow
nasal cannula. Respiratory Medicine CME; 2009;2(3):121-124. http://
www.resmedcme.com.
6. Calvano TP, Sill JM, Kemp KR, Chung KK. Use of high-flow ox-
ygen delivery system in a critically ill patient with dementia. Respir
Care 2008;53(12):1739-1743.
7. Roca O, Masclans JR, Laborda C, Sacanell J, Serra J. High-flow
nasal cannula improves oxygen administration in hypoxemic respi-
ratory failure (abstract). Intensive Care Med 2007;33:S88.
8. Vincent JL, Moreno R, Takala J, Willatts S, De Mendoc¸a A, Bru-
ining H, et al. The SOFA (Sepsis-related Organ Failure Assessment)
score to describe organ dysfunction/failure. On behalf of the Work-
ing Group on Sepsis-Related Problems of the European Society of
Intensive Care Medicine. Intensive Care Med 1996;22(7):707-710.
9. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II:
a severity of disease classification system. Crit Care Med 1985;
13(10):818-829.
10. Murray JF, Matthay MA, Luce JM, Flick MR. An expanded defini-
tion of the adult respiratory distress syndrome. Am Rev Respir Dis
1988;138(3):720-723.
HIGH-FLOW OXYGEN THERAPY IN ACUTE RESPIRATORY FAILURE
412 RESPIRATORY CARE •APRIL 2010 VOL 55 NO4

11. Bausewein M, Farquhar M, Booth S, Gysels M, Higginson IJ. Mea-
surement of breathlessness in advanced disease: a systematic review.
C. Respir Med 2007;101(3):399-410.
12. Noether GE. Sample size determination for some sommon nonpara-
metric statistics. J Am Stat Assoc 1987;82:645-647.
13. Chanques G, Constantin JM, Sauter M, Jung B, Sebbane M, Verzilli
D, et al. Discomfort associated with under-humidified high-flow
oxygen therapy in critically ill patients. Intensive Care Med 2009;
35(6):996-1003.
14. Wilson RC, Jones PW. A comparison of the visual analogue scale
and modified Borg scale for the measurement of dyspnoea during
exercise. Clin Sci 1989;76(3):277-282.
15. Gift AG. Validation of a vertical visual analogue scale as a measure
of clinical dyspnea. Rehabil Nurs 1989;14(6):323-325.
16. Mahler DA. The measurement of dyspnea during exercise in patients
with lung disease. Chest 1992;101(5 Suppl):242S-247S.
17. Mador MJ, Kufel TJ. Reproducibility of visual analog scale mea-
surements of dyspnea in patients with chronic obstructive pulmonary
disease. Am Rev Respir Disease 1992;146(1):82-87.
18. Aitken RC. Measurement of feelings using visual analogue scales.
Proc R Soc Med 1969;62(10):989-993.
19. van der Molen B. Dyspnoea: a study of measurement instruments for
the assessment of dyspnoea and their application for patients with
advanced cancer. J Adv Nurs 1995;22(5):948-956.
20. Wettstein RB, Shelledy DC, Peters JI. Delivered oxygen concentra-
tions using low-flow and high-flow nasal cannulas. Respir Care 2005;
50(5):604-609.
21. Costello RW, Liston R, McNicholas WT. Compliance at night with
low flow oxygen therapy: a comparison of nasal cannulae and Ven-
turi face masks. Thorax 1995;50(4):405-6.
HIGH-FLOW OXYGEN THERAPY IN ACUTE RESPIRATORY FAILURE
RESPIRATORY CARE •APRIL 2010 VOL 55 NO4 413