High-Flow Nasal Cannulae in Very Preterm Infants After Extubation

The n e w e ng l a n d j o u r na l of m e dic i n e
original article
High-Flow Nasal Cannulae in Very Preterm

Infants after Extubation
Brett J. Manley, M.B., B.S., Louise S. Owen, M.D., Lex W. Doyle, M.D.,
Chad C. Andersen, M.B., B.S., David W. Cartwright, M.B., B.S.,
Margo A. Pritchard, Ph.D., Susan M. Donath, M.A., and Peter G. Davis, M.D.
A bs t r ac t
Background
The use of high-flow nasal cannulae is an increasingly popular alternative to nasal From the Newborn Research Centre, Royal
continuous positive airway pressure (CPAP) for noninvasive respiratory support of Womens Hospital (B.J.M., L.S.O., L.W.D.,
P.G.D.), the Departments of Obstetrics
very preterm infants (gestational age, <32 weeks) after extubation. However, data and Gynaecology (B.J.M., L.S.O., L.W.D.,
on the efficacy or safety of such cannulae in this population are lacking. P.G.D.) and Paediatrics (L.W.D., S.M.D.),
University of Melbourne, and the Critical
Care and Neurosciences Theme (B.J.M.,
Methods L.S.O., L.W.D., P.G.D.) and Clinical Epide-
In this multicenter, randomized, noninferiority trial, we assigned 303 very preterm miology and Biostatistics Unit (S.M.D.),
infants to receive treatment with either high-flow nasal cannulae (5 to 6 liters per Murdoch Childrens Research Institute,
Melbourne, VIC, the Department of Neo-
minute) or nasal CPAP (7 cm of water) after extubation. The primary outcome was natal Medicine, Womens and Childrens
treatment failure within 7 days. Noninferiority was determined by calculating the Hospital, Adelaide, SA (C.C.A.), and Wom-
absolute difference in the risk of the primary outcome; the margin of noninferiority ens and Newborn Services, Royal Brisbane
and Womens Hospital (D.W.C., M.A.P.), the
was 20 percentage points. Infants in whom treatment with high-flow nasal can- Department of Paediatrics and Child Health,
nulae failed could be treated with nasal CPAP; infants in whom nasal CPAP failed University of Queensland (D.W.C.), and the
were reintubated. University of Queensland Centre for Clini-
cal Research (M.A.P.), Brisbane, QLD all
in Australia. Address reprint requests to
Results Dr. Manley at the Newborn Research Centre,
The use of high-flow nasal cannulae was noninferior to the use of nasal CPAP, with Level 7, Royal Womens Hospital, 20 Flem-
ington Rd., Parkville, VIC 3052, Australia,
treatment failure occurring in 52 of 152 infants (34.2%) in the nasal-cannulae or at [email protected].
group and in 39 of 151 infants (25.8%) in the CPAP group (risk difference, 8.4 per-
N Engl J Med 2013;369:1425-33.
centage points; 95% confidence interval, 1.9 to 18.7). Almost half the infants in DOI: 10.1056/NEJMoa1300071
whom treatment with high-flow nasal cannulae failed were successfully treated Copyright 2013 Massachusetts Medical Society.
with CPAP without reintubation. The incidence of nasal trauma was significantly
lower in the nasal-cannulae group than in the CPAP group (P=0.01), but there were
no significant differences in rates of serious adverse events or other complications.
Conclusions
Although the result for the primary outcome was close to the margin of noninferi-
ority, the efficacy of high-flow nasal cannulae was similar to that of CPAP as respi-
ratory support for very preterm infants after extubation. (Funded by the National
Health and Medical Research Council; Australian New Zealand Clinical Trials Net-
work number, ACTRN12610000166077.)
n engl j med 369;15 nejm.org october 10, 2013 1425

The New England Journal of Medicine
Downloaded from nejm.org at RUTH LILLY MED LIBRARY on May 9, 2015. For personal use only. No other uses without permission.
Copyright 2013 Massachusetts Medical Society. All rights reserved.
I
n the United States, approximately Me thods
75,000 infants were classified as very preterm
(gestational age, <32 weeks) in 2011.1 Very Study Design and Oversight
preterm infants have substantially higher mor- From May 31, 2010, to July 3, 2012, we recruited
tality and morbidity than term infants, partly be- infants in three Australian neonatal intensive care
cause they are more prone to respiratory failure units at the Royal Womens Hospital, Melbourne;
and often require mechanical ventilation through Womens and Childrens Hospital, Adelaide; and
an endotracheal tube after birth. Once they re- the Royal Brisbane and Womens Hospital, Bris-
cover from their acute breathing problems, the bane. The trial was designed and conducted by
best way to achieve successful extubation from the authors. The human research ethics commit-
mechanical ventilation is controversial. Nasal con- tee at each center approved the trial. The study
tinuous positive airway pressure (CPAP) is known was funded by the participating neonatal units
to be superior to no positive-pressure support2 and by the National Health and Medical Research
and is the current standard of care for noninvasive Council. All authors vouch for the accuracy and
respiratory support of very preterm infants. completeness of the data and the fidelity of the
The use of high-flow nasal cannulae is an study to the protocol, which is available with the
alternative means of providing noninvasive re- full text of this article at NEJM.org. There was no
spiratory support to very preterm infants. Such commercial support for this study.
cannulae deliver heated and humidified gas at
flow rates of more than 1 liter per minute Patients
through small binasal prongs. Because high-flow Infants were eligible if they were born at a gesta-
nasal cannulae have a simpler interface with the tional age of less than 32 weeks, were receiving
infant and smaller prongs than nasal CPAP, the mechanical ventilation through an endotracheal
cannulae are perceived as easier to use, more tube, and were scheduled to undergo extubation
comfortable for the infant, and advantageous for for the first time to noninvasive respiratory sup-
motherinfant bonding.3,4 The use of high-flow port. Infants were ineligible if their gestational
nasal cannulae is increasingly popular for non- age was more than 36 weeks at the time of extu-
invasive respiratory support in neonatal intensive bation, if they were participating in a concurrent
care units around the world.3,5 In 2008, the de- study that prohibited inclusion, if they had a
vices were being used in up to two thirds of known major congenital anomaly that might af-
academic neonatal units in the United States.6 fect breathing, or if maximal intensive care was
High-flow nasal cannulae are used to prevent not being provided. All parents or guardians pro-
extubation failure in preterm infants, with pre- vided written informed consent.
scribed flow rates typically ranging from 2 to
8 liters per minute.3,7,8 However, there is little Randomization
evidence regarding the risks and benefits of this A computer-generated block-randomization se-
new technique.9,10 Several mechanisms of action quence with random block sizes was used. All
of high-flow nasal cannulae in preterm infants infants were stratified according to gestational
have been described,11 but the contribution of age (<26 weeks vs. 26 weeks) and study center.
each is undefined. Furthermore, concern has Infants who were part of multiple births under-
been raised that the generation of unregulated went individual randomization. Clinicians opened
distending pressure from high-flow nasal can- consecutively numbered, sealed, opaque enve-
nulae might injure the lungs.12,13 lopes immediately before extubation to deter-
We hypothesized that high-flow nasal cannu- mine the study-group assignment. In the case of
lae are an appropriate alternative to nasal CPAP, an unplanned extubation in an infant for whom
even if the devices do not provide superior respi- consent had already been obtained, randomiza-
ratory support. We performed a multicenter, tion occurred if the treating physician decided to
randomized, noninferiority trial of high-flow provide noninvasive respiratory support. If the
nasal cannulae, as compared with nasal CPAP, infant required resuscitation and immediate re-
as respiratory support after extubation in pre- intubation, randomization did not occur, and the
term infants born at a gestational age of less infant remained eligible for inclusion with the
than 32 weeks. next planned extubation.
1426 n engl j med 369;15 nejm.org october 10, 2013

Nasal Cannulae in Very Preterm Infants after Extubation
Study Intervention dered at the discretion of the treating team if the

Infants received oral or intravenous caffeine (ei- fraction of inspired oxygen was less than 0.3.
ther a loading dose of 20 mg per kilogram of Infants in whom treatment with high-flow
body weight or a maintenance dose of 5 to 10 mg nasal cannulae failed were treated with nasal
per kilogram) in the 24 hours before extubation. CPAP (7 cm of water), with the same pressure
The treating physician made the decision to extu- limits as those used in the CPAP group. If the
bate an infant before randomization. Infants re- CPAP treatment failed under the maximal pres-
ceived their assigned treatment immediately af- sure during the first 7 days after extubation,
ter extubation. Treatment was considered to have infants were reintubated.
failed if an infant who was receiving maximal Infants in the CPAP group were started on a
respiratory support with the assigned treatment pressure of 7 cm of water with either binasal
met one or more of the following four criteria for midline prongs (Fisher & Paykel Healthcare) or
failure within 7 days after extubation: a fraction subnasal prongs (Hudson RCI), depending on
of inspired oxygen of 0.2 or more above the base- the practice in the individual unit. Nasal CPAP
line value before extubation that was required to was generated with the use of a mechanical ven-
maintain a peripheral oxygen saturation of 88 to tilator or an underwater bubble system. Pres-
92%; a pH of less than 7.2 and a partial pressure sures were altered at the physicians discretion,
of carbon dioxide of more than 60 mm Hg on an within limits of 5 to 8 cm of water during the
arterial or free-flowing capillary blood gas sam- first 7 days after extubation. Nonsynchronized
ple; more than one apneic episode requiring in- nasal intermittent positive-pressure ventilation
termittent positive-pressure ventilation within a could be used at the time of extubation or at any
24-hour period or six or more apneic episodes later time during the admission, with a maxi-
requiring stimulation within 6 consecutive hours; mum inflating pressure of 25 cm of water and a
or an urgent need for reintubation and mechan- maximum rate of 40 inflations per minute. Once
ical ventilation, as determined by the treating infants were weaned to a nasal CPAP of 5 cm of
physician. water with a fraction of inspired oxygen of less
Infants in the nasal-cannulae group were than 0.3 for more than 24 hours, nasal CPAP could
treated with the Optiflow device, which included be discontinued. Infants in whom CPAP treatment
the MR850 humidifier and binasal infant can- failed at a maximal pressure of 8 cm of water
nulae (Fisher & Paykel Healthcare). Infants were (with or without nonsynchronized nasal intermit-
fitted with prongs that maintained a leak at the tent positive-pressure ventilation) were reintubated.
nose, with the aim of occluding approximately Infants in the CPAP group were not permitted to
half the nares. The device includes a pressure- receive treatment with high-flow nasal cannulae at
relief valve that limits circuit pressure to 45 cm any time during their admission.
of water. The starting flow rate was based on the
size of the prongs used, with 5 liters per minute Study Outcomes
for premature or neonatal prongs or 6 liters The primary outcome was treatment failure
per minute for infant, intermediate infant, within 7 days (168 hours) after extubation. Pre-
or pediatric prongs. Flow rates were altered at specified secondary outcomes included reintuba-
the physicians discretion in a stepwise fashion, tion during the primary-outcome period, death
with mandated limits between 2 liters per min- before hospital discharge, a requirement for sup-
ute and the maximum recommended for the plemental oxygen at a gestational age of 36 weeks,
prong size: 6 liters per minute for premature pneumothorax after trial entry, total days of any
and neonatal prongs, 7 liters per minute for respiratory support after trial entry, duration of
infant or intermediate infant prongs, and oxygen supplementation after trial entry, and
8 liters per minute for pediatric prongs. For length of hospital admission. Nursing and medi-
infants who were weaned to 2 liters per minute cal staff members, who were aware of study-
and who had a fraction of inspired oxygen of group assignments, regularly assessed the nasal
less than 0.3 for more than 24 hours, treatment skin and septum for trauma caused by pressure
with the high-flow nasal cannulae could be from the binasal prongs. Data included the inci-
stopped, although such cessation of therapy was dence of any nasal trauma, its cause, and the need
not mandatory, and earlier cessation was or- for change of treatment because of nasal trauma.

1099 Infants born at <32 wk of gestational age

were screened for eligibility
622 Were ineligible

530 Did not undergo ventilation
30 Died before extubation
25 Were excluded
15 Had major congenital anomaly
8 Were included in another extubation
study
2 Had gestational age 36 wk
14 Were extubated to no support
13 Received palliative care
10 Were extubated at another hospital
477 Were eligible
174 Did not undergo randomization

77 Were not approached
34 Had consent declined
28 Had medical reasons
19 Were unable to have consent provided
for social reasons
10 Had unplanned extubation
6 Had consent provided but did not undergo
randomization
3 Died before randomization
2 Had consent withdrawn
1 Missed randomization
303 Underwent randomization
152 Were assigned to high-flow nasal

151 Were assigned to nasal CPAP
cannulae
152 Were included in primary analysis 151 Were included in primary analysis
Figure 1. Enrollment and Outcomes.
Serious adverse events were predefined as where high-flow nasal cannulae had not been
pneumothorax occurring while the infants re- used previously. We prespecified the margin of
ceived the assigned treatment and death. These noninferiority for high-flow nasal cannulae as
events were reported to the principal investiga- 20 percentage points above the failure rate for
tor and the relevant ethics committee as they nasal CPAP. Since little guidance was available
occurred. Data were collected until death or the for such a comparison, the choice of margin was
first discharge home from the hospital. somewhat arbitrary. Our rationale was that many
clinicians already strongly prefer the use of high-
Statistical Analysis flow nasal cannulae, and the devices are in wide-
We estimated that extubation would have a fail- spread use because of their perceived benefits over
ure rate of 25% for very preterm infants on the nasal CPAP. We allowed the use of rescue nasal
basis of a review of 2 years of data from the lead CPAP for infants in whom treatment with high-
center (the Royal Womens Hospital, Melbourne), flow nasal cannulae failed. This procedure reflected

Table 1. Characteristics of the Infants and Their Mothers.*
Nasal-Cannulae Group CPAP Group

Characteristic (N=152) (N=151)
Mothers
White race no. (%) 127 (83.6) 120 (79.5)
Primigravida no. (%) 61 (40.1) 59 (39.1)
Exposure to antenatal glucocorticoids no. (%) 142 (93.4) 143 (94.7)
Cesarean section no. (%) 101 (66.4) 101 (66.9)
Infants
Gestational age
No. of wk 27.72.1 27.51.9
<26 wk no. (%) 32 (21.1) 31 (20.5)
Birth weight g 1041338 1044327
Male sex no. (%) 89 (58.6) 72 (47.7)
Multiple birth no. (%) 49 (32.2) 52 (34.4)
Intubated in the delivery room no. (%) 102 (67.1) 91 (60.3)
Median Apgar score at 5 min (IQR) 7 (68) 8 (68)
Surfactant treatment no. (%) 141 (92.8) 144 (95.4)
Caffeine treatment before extubation no. (%) 151 (99.3) 148 (98.0)
Median postnatal age at extubation (IQR) hr 43.2 (20.8115.7) 38.5 (22.8101.7)
Median duration of mechanical ventilation before 36 (19.5101.5) 36 (2093)
extubation (IQR) hr
pH before extubation 7.330.06 7.320.06
Partial pressure of carbon dioxide before extubation 44.29.2 43.69.0
mm Hg
Fraction of inspired oxygen before extubation 0.230.04 0.230.04
* Plusminus values are means SD. There were no significant differences between the two groups. CPAP denotes con-
tinuous positive airway pressure, and IQR interquartile range.
Race was reported by the investigators.
The Apgar score was not known for three infants.
Blood gas results were from capillary (heel-prick) or arterial blood samples.
clinical practice around the world at centers where to-treat basis, and infants remained in their as-
both treatments are commonly available. signed group for all outcomes. For the primary
The primary outcome was short-term treat- outcome and dichotomous secondary outcomes,
ment efficacy, rather than death or disability, for we calculated the difference between the groups
which a lower noninferiority margin might be using risk differences and 95% confidence inter-
recommended.14 Noninferiority was determined vals. We used chi-square tests to compare second-
by calculating the absolute risk difference and ary dichotomous outcomes and the appropriate
95% confidence interval for the primary out- parametric test (Students t-test) or nonparamet-
come. For treatment with high-flow nasal can- ric test (Wilcoxon rank-sum test) to compare
nulae to be noninferior, the upper limit of the continuous outcomes. An independent data and
95% confidence interval had to be below 20 per- safety monitoring committee performed two
centage points, and the lower limit of the 95% interim analyses, when the primary outcome
confidence interval had to be below zero.14 We was known for 100 infants and for 200 infants,
determined that a sample of 300 infants was and on the basis of each analysis recommended
required to show noninferiority for high-flow that the study continue unaltered. All analyses
nasal cannulae with a power of 87%. were performed with the use of Stata/IC soft-
All analyses were performed on an intention- ware, version 12.0 (StataCorp).

Table 2. Reasons for Treatment Failure, Reintubation, and Other Secondary Outcomes.
Nasal-Cannulae Group CPAP Group Risk Difference

Outcome (N=152) (N=151) (95% CI)* P Value
percentage points
Reason for treatment failure no./
total no. (%)
Apnea 32/52 (61.5) 25/39 (64.1) 2.6 (22.6 to 17.5) 0.80
Increase in fraction of inspired 21/52 (40.4) 20/39 (51.3) 10.9 (31.5 to 9.7) 0.30
oxygen
Respiratory acidosis 6/52 (11.5) 2/39 (5.1) 6.4 (4.7 to 17.5) 0.29
Urgent need for intubation 2/52 (3.8) 4/39 (10.3) 6.4 (17.3 to 4.5) 0.22
Reintubation within 7 days after 27 (17.8) 38 (25.2) 7.4 (16.6 to 1.8) 0.12
extubation no. (%)
Median no. of days of respiratory sup- 34 (7 to 55) 38 (11 to 57) NA 0.44
port after trial entry (IQR)
Median no. of days of oxygen therapy 38 (0 to 78) 49 (8 to 83) NA 0.15
after trial entry (IQR)
Median weight gain during first 7 days 20 (42 to 79.5) 10 (54 to 75) NA 0.39
after extubation (IQR) g
Median discharge weight (IQR) g 2886 (2515 to 3290) 2950 (2500 to 3524) NA 0.53
Median no. of days in a tertiary care 65 (43 to 91) 64 (45 to 104) NA 0.56
center (IQR)
Median no. of days in any hospital (IQR) 79 (63 to 105) 84 (65 to 106) NA 0.38
* Positive values favor the CPAP group, and negative values favor the nasal-cannulae group.
Treatment may have failed for more than one reason. In one infant in the CPAP group, treatment failed because a set
pressure of more than 8 cm of water was prescribed during the primary-outcome period.
One infant had a measured pH of exactly 7.20, and one infant had a measured pH of 7.21.
Respiratory support included the use of high-flow nasal cannulae, nasal CPAP, nonsynchronized nasal intermittent
positive-pressure ventilation, and mechanical ventilation through an endotracheal tube.
Infants who died before discharge were excluded.
This category includes only hospital days, even though some infants received oxygen therapy after being discharged.
R e sult s The most common reason for treatment fail-

ure in the two study groups was apnea (Table 2),
Study Patients with no significant between-group difference in
A total of 303 infants underwent randomization the reason for failure. Failure was most likely to
(152 to the nasal-cannulae group and 151 to the occur during the first day after extubation in the
CPAP group) during the study period (Fig. 1). The two groups (Fig. 2).
demographic and clinical characteristics of the
mothers and infants in the two groups were sim- Reintubation during the First 7 Days
ilar (Table 1). after Extubation
Of the 52 infants in whom treatment with high-
Primary Outcome flow nasal cannulae failed during the first 7 days
The use of high-flow nasal cannulae was found after extubation, 25 (48%) were successfully treat-
to be noninferior to the use of nasal CPAP by our ed with nasal CPAP or nonsynchronized nasal
definition, with treatment failure occurring in intermittent positive-pressure ventilation without
52of 152 infants (34.2%) in the nasal-cannulae reintubation. Thus, only 17.8% of infants in the
group and 39 of 151 infants (25.8%) in the CPAP nasal-cannulae group were reintubated, as com-
group (risk difference, 8.4 percentage points; pared with 25.2% of those in the CPAP group
95% confidence interval, 1.9 to 18.7). Treatment (P=0.12) (Table 2). In the latter group, 3 infants
failure stratified according to gestational-age received nonsynchronized nasal intermittent
subgroup is provided in Table S1 in the Supple- positive-pressure ventilation immediately after
mentary Appendix, available at NEJM.org. extubation, and 36 infants (23.8%) received non-

synchronized nasal intermittent positive-pressure

ventilation during the first 7 days after extuba- 100
tion; of these infants, 22 were reintubated after 90
Percentage of Infants without

treatment failure. 80 Nasal CPAP
Treatment Failure
70
Other Secondary Outcomes and Adverse Events 60 High-flow nasal cannulae
There were no significant between-group differ- 50
ences in rates of other secondary outcomes (Ta- 40
ble 2) or in rates of death or other serious adverse 30
events (Table 3). The pneumothorax rate after 20
trial entry was low in the two study groups. In 10

0
the nasal-cannulae group, infants had a signifi- 0 1 2 3 4 5 6 7
cantly lower incidence of nasal trauma than those No. of Days after Extubation
in the CPAP group (39.5% vs. 54.3%, P = 0.01), and
No. at Risk
fewer infants required a change in therapy be- High-flow nasal 152 120 116 107 104 103 102 100
cause of nasal trauma (P = 0.001). No infant in cannulae
Nasal CPAP 151 122 119 118 117 113 112 112
either group required surgical correction for na-
sal trauma. Almost half the cases of nasal trau-
Figure 2. Very Preterm Infants without Treatment Failure after Extubation.
ma in the nasal-cannulae group were diagnosed
Shown are the percentages of very preterm infants (gestational age, <32 weeks)
while the infants were receiving other types of
in whom treatment with either high-flow nasal cannulae or nasal continuous
noninvasive respiratory support. If the diagnosis positive airway pressure (CPAP) did not fail after extubation during the primary
of nasal trauma was limited to cases that were outcome period (7 days). Treatment failure occurred in 52 of 152 infants (34.2%)
diagnosed during the assigned treatment, the in the nasal-cannulae group and 39 of 151 infants (25.8%) in the CPAP group
between-group difference was greater, with a (risk difference, 8.4 percentage points).
rate of 19.1% in the nasal-cannulae group versus
53.0% in the CPAP group (P<0.001).
one of the two treatments after extubation. Sig-
Discussion nificantly more infants in the nasal-cannulae
group required reintubation within 7 days, prob-
On the basis of our prespecified definition of ably because of the reduced flow rates (<2 liters
noninferiority, the use of high-flow nasal can- per minute) that were used. In the study by Col-
nulae was noninferior to the use of nasal CPAP as lins et al.,16 which involved 132 very preterm in-
respiratory support for very preterm infants after fants, the investigators used a different high-flow
extubation, with a between-group difference of nasal-cannulae device (Vapotherm) and an in-
8.4 percentage points that favored CPAP. When creased flow rate (8 liters per minute) after extu-
treatment with high-flow nasal cannulae failed, bation. They found no significant between-group
about half the infants were successfully treated difference in the primary outcome of treatment
with CPAP without reintubation, resulting in a failure within 7 days and no significant difference
nonsignificant between-group difference in the in reintubation rates, although the authors ac-
rate of reintubation within 7 days after extuba- knowledged that the study was underpowered for
tion. Among infants in the nasal-cannulae group, these outcomes. Nasal-trauma scores were lower
there was no increase in rates of death or compli- in the nasal-cannulae group than in the CPAP
cations and there was a decrease in the rate of group, which is consistent with our findings.
nasal trauma, as compared with the CPAP group. In our multicenter trial, 63 of the infants had
The use of high-flow nasal cannulae has been a gestational age of less than 26 weeks, but the
widely adopted without evidence of safety and ef- study was not powered to evaluate the efficacy
ficacy. The results of two small, randomized trials or safety of high-flow nasal cannulae in this ex-
of these devices, as compared with nasal CPAP, tremely preterm subgroup. The treatment-failure
after extubation have been reported previously. rate was very high among these infants, regard-
Both these trials were conducted in a single center less of the assigned treatment, and the risk dif-
and had a superiority design. In the study by ference was 20 percentage points in favor of nasal
Campbell et al.,15 40 infants with a birth weight CPAP (see the Supplementary Appendix). Given
of less than 1250 g were randomly assigned to this finding, we think that clinicians should be

Table 3. Adverse Events.
Nasal-Cannulae CPAP
Group Group Risk Difference
Event (N=152) (N=151) (95% CI)* P Value
no. (%) percentage points

Death before discharge 5 (3.3) 6 (4.0) 0.7 (4.9 to 3.5) 0.75
Oxygen supplementation at gestational age of 47 (30.9) 52 (34.4) 3.5 (14.1 to 7.0) 0.51
36 wk
Glucocorticoid treatment for lung disease after 27 (17.8) 30 (19.9) 2.1 (10.9 to 6.7) 0.64
trial entry
Pneumothorax
Anytime after trial entry 1 (0.7) 4 (2.6) 2.0 (4.9 to 0.9) 0.17
During assigned treatment 0 1 (0.7) 0.6 (1.9 to 0.6) 0.32
Discharged home with oxygen 21 (13.8) 15 (9.9) 3.9 (3.4 to 11.2) 0.30
Proven sepsis after trial entry 26 (17.1) 30 (19.9) 2.8 (11.5 to 6.0) 0.54
Patent ductus arteriosus treated with medication 64 (42.1) 64 (42.4) 0.3 (11.4 to 10.8) 0.96
or surgery
Necrotizing enterocolitis stage 2 or 3 3 (2.0) 7 (4.6) 2.7 (6.7 to 1.4) 0.20
Isolated intestinal perforation 1 (0.7) 2 (1.3) 0.7 (2.9 to 1.6) 0.56
Laser surgery for retinopathy of prematurity 8 (5.3) 8 (5.3) 0.0 (5.1 to 5.0) 0.99
Intraventricular hemorrhage grade 3 or 4 3 (2.0) 8 (5.3) 3.3 (7.5 to 0.9) 0.12
Cystic periventricular leukomalacia 5 (3.3) 3 (2.0) 1.3 (2.3 to 4.9) 0.48
Nasal trauma
Any documented 60 (39.5) 82 (54.3) 14.8 (25.9 to 3.7) 0.01
Leading to change of treatment 8 (5.3) 27 (17.9) 12.6 (19.7 to 5.5) 0.001
Caused by the assigned treatment 29 (19.1) 80 (53.0) 33.9 (44.0 to 23.8) <0.001
* Positive values favor the CPAP group, and negative values favor the nasal-cannulae group.
This event was prespecified as a serious adverse event.
Infants who died before hospital discharge were excluded.
cautious before using high-flow nasal cannulae pressures.18-20 Since it was not possible to blind
as first-line respiratory support in extremely the intervention, we used prespecified, objective
preterm infants after extubation. The risk dif- criteria for treatment failure as the primary out-
ference for treatment failure among infants come to minimize bias.
with a gestational age of 26 weeks or more was We allowed the use of rescue nasal CPAP in the
5 percentage points. nasal-cannulae group only after treatment failure.
The results of our trial pertain only to infants We acknowledge that this may have influenced
who have been extubated and should not be ex- some of the secondary outcomes in the nasal-
trapolated to the use of high-flow nasal cannulae cannulae group. Because the use of nasal CPAP
as primary respiratory support after birth. We used prevented the reintubation of almost half the in-
the Fisher & Paykel high-flow nasal-cannulae sys- fants in whom treatment with nasal cannulae had
tem in our trial, but there is no evidence that any failed, it seems reasonable to use high-flow nasal
one commercially available system is superior to cannulae initially after extubation if both therapies
another.17 The flow rates for the nasal cannulae are available. In our trial, nonsynchronized nasal
that we used have been shown to result in dis- intermittent positive-pressure ventilation could be
tending pressures in vivo that are similar to, or used in infants receiving nasal CPAP. If the use
slightly lower than, commonly set nasal CPAP of nonsynchronized nasal intermittent positive-

pressure ventilation had not been permitted cannulae as primary respiratory support from
until nasal CPAP had failed, the risk difference birth will be important. We conclude that treat-
for the primary outcome might have been ment with high-flow nasal cannulae was non
smaller. inferior to the use of nasal CPAP as respiratory
When appraising the result of a noninferiority support after extubation in very preterm infants.
trial, both the risk difference and its 95% confi- Since our trial was underpowered to show non-
dence interval are considered. Although our best inferiority in infants with a gestational age of
estimation of the risk difference was 8.4 per- less than 26 weeks, the use of high-flow nasal
centage points, we acknowledge that the upper cannulae as first-line respiratory support after
limit of the 95% confidence interval was close to extubation in this extremely preterm group re-
the margin of noninferiority (20 percentage points) quires caution.
and that slightly different failure rates in either Supported by a program grant (606789) and a Centre for
group could have altered our conclusions. Clinical Research Excellence grant (546519) from the National
Health and Medical Research Council.
Further research on the use of high-flow na- Disclosure forms provided by the authors are available with
sal cannulae in preterm and term newborn in- the full text of this article at NEJM.org.
fants is required. Since such cannulae are used We thank the patients families and the staff members who
cared for them at the three participating centers, and research
with increasing frequency in the nontertiary care nurses Connie Wong, Donna Hovey, Louise Goodchild, and Ros
setting, randomized trials of high-flow nasal Lontis for their assistance in recruitment and data collection.
References
1. Hamilton BE, Martin JA, Ventura SJ. MG. Heated humidified high-flow nasal ment. JAMA 2006;295:1152-60. [Erratum,
Births: preliminary data for 2011. Nation- cannula: use and a neonatal early ex JAMA 2006;296:1842.]
al vital statistics reports. Vol. 61. No. 5. tubation protocol. J Perinatol 2007;27: 15. Campbell DM, Shah PS, Shah V, Kelly
Hyattsville, MD: National Center for Health 776-81. EN. Nasal continuous positive airway pres-
Statistics, 2012. (DHHS publication no. 8. Shoemaker MT, Pierce MR, Yoder BA, sure from high flow cannula versus Infant
(PHS) 2013-1120.) DiGeronimo RJ. High flow nasal cannula Flow for preterm infants. J Perinatol 2006;
2. Davis P, Henderson-Smart D. Post- versus nasal CPAP for neonatal respiratory 26:546-9.
extubation prophylactic nasal continuous disease: a retrospective study. J Perinatol 16. Collins CL, Holberton JR, Barfield C,
positive airway pressure in preterm in- 2007;27:85-91. Davis PG. A randomized controlled trial
fants: systematic review and meta-analysis. 9. Wilkinson D, Andersen C, ODonnell to compare heated humidified high-flow
J Paediatr Child Health 1999;35:367-71. CP, De Paoli AG. High flow nasal cannula nasal cannulae with nasal continuous
3. Hough JL, Shearman AD, Jardine LA, for respiratory support in preterm in- positive airway pressure postextubation
Davies MW. Humidified high flow nasal fants. Cochrane Database Syst Rev 2011; in premature infants. J Pediatr 2012;162(5):
cannulae: current practice in Australasian 5:CD006405. 949.e1-954.e1.
nurseries, a survey. J Paediatr Child Health 10. Manley BJ, Dold SK, Davis PG, Roehr 17. Miller SM, Dowd SA. High-flow nasal
2012;48:106-13. CC. High-flow nasal cannulae for respira- cannula and extubation success in the
4. Manley BJ, Owen L, Doyle LW, Davis tory support of preterm infants: a review premature infant: a comparison of two
PG. High-flow nasal cannulae and nasal of the evidence. Neonatology 2012;102: modalities. J Perinatol 2010;30:805-8.
continuous positive airway pressure use 300-8. 18. Spence KL, Murphy D, Kilian C, Mc-
in non-tertiary special care nurseries in 11. Shaffer TH, Alapati D, Greenspan JS, Gonigle R, Kilani RA. High-flow nasal
Australia and New Zealand. J Paediatr Wolfson MR. Neonatal non-invasive re- cannula as a device to provide continuous
Child Health 2012;48:16-21. spiratory support: physiological implica- positive airway pressure in infants. J Peri-
5. Nath P, Ponnusamy V, Willis K, Bis- tions. Pediatr Pulmonol 2012;47:837-47. natol 2007;27:772-5.
sett L, Clarke P. Current practices of high 12. Finer NN. Nasal cannula use in the 19. Kubicka ZJ, Limauro J, Darnall RA.
and low flow oxygen therapy and humidi- preterm infant: oxygen or pressure? Pedi- Heated, humidified high-flow nasal can-
fication in UK neonatal units. Pediatr Int atrics 2005;116:1216-7. nula therapy: yet another way to deliver
2010;52:893-4. 13. Finer NN, Mannino FL. High-flow continuous positive airway pressure? Pe-
6. Hochwald O, Osiovich H. The use of nasal cannula: a kinder, gentler CPAP? diatrics 2008;121:82-8.
high flow nasal cannulae in neonatal in- JPediatr 2009;154:160-2. 20. Wilkinson DJ, Andersen CC, Smith K,
tensive care units: is clinical practice con- 14. Piaggio G, Elbourne DR, Altman DG, Holberton J. Pharyngeal pressure with
sistent with the evidence? J Neonat Perinat Pocock SJ, Evans SJ. Reporting of non high-flow nasal cannulae in premature
Med 2010;3:187-91. inferiority and equivalence randomized infants. J Perinatol 2008;28:42-7.
7. Holleman-Duray D, Kaupie D, Weiss trials: an extension of the CONSORT state- Copyright 2013 Massachusetts Medical Society.
2013 icmje recommendations

The International Committee of Medical Journal Editors (ICMJE) has revised
and renamed its Uniform Requirements. The new ICMJE Recommendations
for the Conduct, Reporting, Editing and Publication of Scholarly Work
in Medical Journals are available at www.icmje.org.


High-Flow Nasal Cannulae in Very Preterm Infants After Extubation

Uploaded by

Copyright:

Available Formats

High-Flow Nasal Cannulae in Very Preterm Infants After Extubation

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

High-Flow Nasal Cannulae in Very Preterm Infants After Extubation

Uploaded by

Copyright:

Available Formats

The n e w e ng l a n d j o u r na l of m e dic i n e

High-Flow Nasal Cannulae in Very Preterm

n engl j med 369;15 nejm.org october 10, 2013 1425

1426 n engl j med 369;15 nejm.org october 10, 2013

The New England Journal of Medicine

Study Intervention dered at the discretion of the treating team if the

n engl j med 369;15 nejm.org october 10, 2013 1427

1099 Infants born at <32 wk of gestational age

622 Were ineligible

477 Were eligible

174 Did not undergo randomization

303 Underwent randomization

152 Were assigned to high-flow nasal

Figure 1. Enrollment and Outcomes.

1428 n engl j med 369;15 nejm.org october 10, 2013

The New England Journal of Medicine

Table 1. Characteristics of the Infants and Their Mothers.*

Nasal-Cannulae Group CPAP Group

n engl j med 369;15 nejm.org october 10, 2013 1429

Nasal-Cannulae Group CPAP Group Risk Difference

R e sult s The most common reason for treatment fail-

1430 n engl j med 369;15 nejm.org october 10, 2013

The New England Journal of Medicine

synchronized nasal intermittent positive-pressure

Percentage of Infants without

trial entry was low in the two study groups. In 10

n engl j med 369;15 nejm.org october 10, 2013 1431

Table 3. Adverse Events.

no. (%) percentage points

1432 n engl j med 369;15 nejm.org october 10, 2013

The New England Journal of Medicine

2013 icmje recommendations

n engl j med 369;15 nejm.org october 10, 2013 1433

You might also like