Journal of Perinatology (2006) 26, 481–485
r 2006 Nature Publishing Group All rights reserved. 0743-8346/06 $30
www.nature.com/jp
ORIGINAL ARTICLE
Comparing two methods of delivering high-flow gas therapy by
nasal cannula following endotracheal extubation: a prospective,
randomized, masked, crossover trial
DD Woodhead, DK Lambert, JM Clark and RD Christensen
Intermountain Healthcare, McKay-Dee Hospital, Ogden, UT, USA
Objective: We compared two methods of delivering high-flow gas therapy
by nasal cannula, applied immediately after planned endotracheal
extubations of NICU patients.
Study design and methods: Thirty NICU patients who were about to
be extubated from mechanical ventilation were randomized into two
groups; Group 1 received Vapotherms for the first 24 h after extubation,
then standard high-flow nasal cannula for the next 24 h, and Group 2
received standard high-flow therapy for the first 24 h, then Vapotherms
for the next 24 h. At 24 h after extubation and again 48 h after extubation,
a neonatologist who was not aware which modality the patient had been
receiving examined the nasal mucosa and applied a scoring system. A
research nurse who was unaware of the modality abstracted respiratory
rates and respiratory effort scores from a specific study-bedside record. The
experimental design was such that a patient could ‘fail’ extubation either
by reintubation for mechanical ventilation, or by rescue to the opposite
modality before completing the 24-h test period.
Results: Fifteen patients were randomized to Group 1 and 15 to Group 2.
No differences were apparent between the groups in birth weight,
gestational age, age at study entry, gender or underlying pulmonary
disorder. Respiratory rates were similar while on Vapotherms
(52±13 breaths/min, mean±s.d.) and high-flow (54±14/min). At 24 h
after starting the modality, those on Vapotherms had more normal
examinations of the nasal mucosa (2.7±1.2 vs 7.8±1.7, P<0.0005) and
lower respiratory effort scores (1.2±0.6 vs 2.0±0.9, P<0.05) than did
those on high-flow. No patients failed while on Vapotherms, but seven
failed while on high-flow (two reintubations and five rescue switches to
Vapotherms, P<0.005).
Conclusions: Among NICU patients immediately following extubation,
Vapotherms performed better than a standard high-flow nasal cannula
in maintaining a normal appearing nasal mucosa, a lower respiratory
effort score, and averting reintubation.
Correspondence: Dr RD Christensen, Intermountain Healthcare, McKay-Dee Hospital, 4403
Harrison Blvd, Ogden, UT 84403, USA.
E-mail: rdchris4@ihc.com
Received 6 February 2006; revised 3 April 2006; accepted 5 April 2006; published online 25 May
2006
Journal of Perinatology (2006) 26, 481–485. doi:10.1038/sj.jp.7211543;
published online 25 May 2006
Keywords: high-flow nasal cannula; endotracheal extubation; reintu-
bation; Vapotherms
Introduction
One potential use of a high-flow nasal cannula in the NICU is to
reduce the need for reintubation following extubation from
mechanical ventilation.1,2 The high-flow nasal cannula system has
been reported to be less invasive than nasal CPAP.1,2 However, the
high-flow cannula can dry the nasal mucosa, particularly when
gas flows exceed 1 or 2 l/min.2,3 Mucosal dryness and thick nasal
secretions sometimes occur when neonates are treated with high-
flow nasal cannula gas delivery. Vapotherms was devised as a
method, in the opinion of the company, for humidifying and
warming inhaled gas to prevent dryness of the nasal mucosa.4–6
However, to date no prospective, randomized studies have been
published directly comparing Vapotherms with a standard high-
flow nasal cannula on recently extubated neonates. To obtain this
comparison, we devised a randomized, prospective, masked,
crossover trial.
Methods
Patients were deemed eligible to participate in the study if they were
in the NICU with an endotracheal tube in place for mechanical
ventilation, and had a physician’s order to extubate to a ‘high-
flow’ (X1 l/min) nasal cannula. They were only eligible if the
attending neonatologist predicted that they were likely to need a
high-flow nasal cannula for at least the next 48 h, and if the
parent or responsible guardian signed the informed consent
document. Patients were ineligible for study participation if they
had what the attending neonatologist judged to be a lethal
congenital abnormality or if they were judged by the attending
neonatologist as likely to be transferred to another hospital before
the study (48 h) had ended.
 Vapotherms vs high-flow nasal cannula after extubation
DD Woodhead et al
482
Eligible patients were randomized before extubation, using a list
constructed by a random number table, to either Group 1 or Group
2. Group 1 patients received Vapotherms for the first 24 h after
extubation and were then crossed over to a standard high-flow
nasal cannula for the next 24 h. Group 2 patients received standard
high-flow nasal cannula treatment for the first 24 h after
extubation and then were crossed over to Vapotherms for the next
24 h. To avoid maldistribution, by chance, with a preponderance of
smaller subjects in one study group, the randomization was
blocked on birth weight, using three categories; <1000, 1000–
1500 and >1500 g. Thirty patients, 15 per group, were planned for
this pilot study, as a convenience sample, based on the projection
that about one patient per week could be studied and the desire to
conduct the study within a 6-month period.
Guidelines of the American Association of Respiratory Care were
used to assess when a patient should be extubated from
mechanical ventilation.7,8 Specific factors considered included;
acceptable blood gasses on weaning ventilator settings, adequate
breath sounds, an improving chest X-ray, and anticipated stamina
sufficient to not need mechanical ventilation any longer.
Immediately following extubation, a nasal cannula was put in
place. Depending on the size of the patient either the premature
(model MA1100A), the neonate (model MN1100B), or the infant
(model MI1300) cannula was used (manufactured by Vapotherm
Inc., Stevensville, MD, USA). The cannula tubing was then
attached to either Vapotherms or a standard high-flow system. In
each individual, the same cannula was used for the entire 48 h of
study. Adjustments in the FiO2 and gas flow were made by the
respiratory therapist and/or nurse bedside in consultation with the
neonatologist and in accordance with manufacture’s
recommendations order to provide what they interpreted as
‘optimal’ individual support to that patient. At 24 h after extubation
the nasal cannula was briefly and temporarily discontinued while a
neonatologist who was unaware of whether the patient had been
receiving high-flow or Vapotherms performed an examination of
the nasal mucosa using a speculum and otoscope. The scoring
system is shown in Table 1. Before each examination of the nasal
mucosa, the nasal prongs were slid out of the nares so that the
nasal mucosa could be seen. When the examining neonatologist
left the bedside, the respiratory therapist set up the opposite
modality for delivering nasal gas flow for the next 24-h period. A
second nasal examination was performed, using the same
procedure, after a 24-h period on the second modality. To help
keep the examining neonatologist unaware of which modality was
being used, a Vapotherms unit was always set up at the bedside of
study patients (whether they were actually receiving Vapotherms
or high-flow). In this way, it was not obvious to the examiner
whether the patient had been receiving Vapotherms or high-flow.
A research nurse (DKL) who was unaware of the randomization,
recorded each respiratory rate and respiratory effort score from a
study-specific bedside chart, on which the bedside nurse or
Journal of Perinatology
Table 1 Scoring the nasal examinations
Right nare Left nare
1 1 Completely normal
2 2 Erythematous mucosa
3 3 Erythematous and edematous mucosa
4 4 Erythema or edema and thick mucous
5 5 Occluded with thick mucous and edema and/or
hemorrhagic
A neonatologist that was unaware of whether the patient had been on Vapotherms or
standard high-flow for the past 24 h examined the nasal mucosa. Two such examinations
were performed on each patient, one after each 24-h study period. The findings of each
examination were recorded on a score sheet by circling the appropriate numbers (below).
The scores were the sum of the numbers recorded for each nare, therefore, the values
ranged from 2 (minimum) to 10 (maximum).
respiratory therapist periodically recorded respiratory rates and a
description of retractions during the two 24 h study periods. The
research nurse recorded the data in an office, where she could not
see which treatment devise the patient was receiving. The
respiratory effort score was listed as; 0 (no retractions), 1 (mild
retractions), 2 (moderate retractions) or 3 (severe retractions). Five
categories of retractions were listed: supraclavicular, suprasternal,
intercostal, substernal and subcostal. In that way respiratory effort
scores could range from a low of 0 (no retractions in any of the
five categories) to a high of 15 (three in each of the five
categories).
Reintubation of a study patient, for return to mechanical
ventilation, occurred under the direction of the clinical care team
(neonatologist, neonatal nurse practitioner, respiratory therapist
and bedside nurse). The team members were aware of whether the
patient was receiving Vapotherms or standard high-flow nasal
cannula. No set criteria were used for determining when a patient
required reintubation for mechanical ventilation. However,
considerations included a consistently rising FiO2, deteriorating
blood gases, increased respiratory effort, and increased apnea and
bradycardia requiring repeated bag/mask breaths. In a similar
fashion, a patient receiving one modality could be switched to the
opposite modality if the clinical care team judged the patient was
failing that modality, but not to the point where reintubation was
needed immediately. If patients were reintubated a tracheal aspirate
was obtained for Gram stain and culture.
Descriptive statistics were calculated using SPSS (v 13.0) for
Windows. Between group means were tested using independent
samples t-tests when parametric assumptions were met, with
Wilcoxon Rank-Sum tests used for non-parametric comparisons.
Proportions were compared between groups using w2 tests or, when
expected counts were small, Fisher’s exact test. For all tests, 2-tailed
tests were used, and a was set at 0.05. The study protocol was
approved by the Intermountain Healthcare Institutional Review
 Vapotherms vs high-flow nasal cannula after extubation
DD Woodhead et al
483

Board, and parents of the study subjects signed informed consent Thirty patients were enrolled, randomized, and begun on one
documents. modality or the other for delivering high-flow nasal gas therapy.
The characteristics of the 30 are given in Table 2. No significant
differences were observed in the characteristics of the 15
randomized to Group 1 and the 15 randomized to Group 2.
Results
Immediately following extubation the subjects were treated with
During the months of July 2005 through November 2005, the either Vapotherms (Group 1) or high-flow (Group 2), but during
families of 35 NICU patients who were receiving mechanical the first 24 h of study, seven were weaned to a low-flow (<0.5 l/
ventilation were contacted to inquire whether they were interested min) nasal cannula. Three of these were in Group 1 and four were
in learning about the extubation study. Thirty-four of the 35 in Group 2 (Table 3). These seven continued to wean from
families gave written consent for their neonate to participate in the supplemental O2, none required reintubation and none were
study, and of these, 30 were enrolled in the study. Three of the discharged home on supplemental O2. Of the remaining 23
consented patients were not enrolled because at the time of patients, 14 finished the 48-h study, but seven ‘failed’. None failed
endotracheal extubation the attending neonatologist decided that while receiving Vapotherms; all seven failures were while
the patient would not be extubated to a high-flow nasal cannula, receiving high-flow (Table 3). Two failed during the first 24 h after
but rather to a low-flow (<0.5 l/min) nasal cannula. One of the extubation due to reintubation because of increasing PCO2 and
consented patients was not enrolled because he was transferred to FiO2 and increasing atelectasis on chest X-ray, all of which were
another hospital before endotracheal extubation. No patients were unresponsive to increasing cannula gas flow (from 1 to 2 l/min).
excluded from the study on the basis that they were deemed to have Tracheal aspirates of these two following reintubation grew no
a lethal congenital abnormality or on the basis that they were organisms. Five failed during the second 24 h because they were
likely to be transferred to another hospital before the study had switched from high-flow back to Vapotherms before the 24-h
ended. study period on high-flow ended. These five were switched after
9.5 h (median; range, 1.5 to 20 h) on high-flow. Two of the five
were switched on the basis of increasing episodes of apnea and
bradycardia that subsided when switched back to Vapotherms; two
Table 2 Characteristics of the study subjects
others were switched because of increasing PCO2 and respiratory
Group 1 Group 2 rate, which improved when moved back to Vapotherms, and the
(received (received high- fifth was switched because of increasing FiO2 with diminished
Vapotherms flow first) breath sounds, which improved when moved back to Vapotherms.
first) n ¼ 15 n ¼ 15
Results of the Vapotherms and standard high-flow nasal
Birth weight (grams, mean±s.d.) 1630±812 1715±880 cannula treatment on respiratory rate, examinations of the nasal
Gestational age at delivery (weeks days; 31.0±3.6 32.0±3.1 mucosa, and respiratory effort scores, are shown in Table 4. All 30
mean±s.d.) patients had a test period on high-flow, but only 28 had a test
Age at study entry (days; median, range) 3 (1–19) 5 (1–16) period on Vapotherms. This is because two (both in Group 2)
Gender (% male) 53 67 failed high-flow within the first 24 h and were reintubated without
ever receiving a trial of Vapotherms. No differences were observed
Initial primary lung disorder: in respiratory rates during the period on Vapotherms vs the period
HMD treated with surfactant 60% 73%
HMD not treated with surfactant 33% 27%
Pneumonia/sepsis 7% F
Meconium aspiration F F Table 3 Short-term outcomes of patients randomized to Group 1 and
TTN F F Group 2
Other F F
Weaned to Failed in Crossed to Failed in
low-flow in first 24 h opposite modality second 24 h
Race/ethnicity
first 24 h at 24 h
White 80% 87%
Hispanic 20% 13% Group 1 3 0 12 5
Black F F (n ¼ 15)
American Indian F F Group 2 4 2 9 0
Pacific Islander F F (n ¼ 15)
Other F F
Group 1 – Vapotherm for first 24 h, then high-flow for second 24 h.
F ¼ zero percent. Group 2 – High-flow for first 24 h, then Vapotherm for second 24 h.

Journal of Perinatology
 Vapotherms vs high-flow nasal cannula after extubation
DD Woodhead et al
484
Table 4 Respiratory rates, nasal mucosa examination scores, and
respiratory effort scores while on Vapotherms vs while on the standard
high-flow cannula
Respiratory ratea Nasal exam Respiratory effort
during the test scoreb at the end scorec at the end
period of the test period of the test period
(Mean+s.d.) (n) (Mean+s.d.) (n) (Mean+s.d.) (n)
Vapotherms 52±13 (24) 2.7±1.2 (20) 1.2±0.6 (21)
High-flow 54±14 (27) 7.8±1.7 (16) 2.0±0.9 (16)
P-value NS <0.0005 <0.05
a
All respiratory rates (breaths per minute) charted during the test period.
b
Nasal examination scores could range from a low of 2 (normal) to a high of 10
(bilaterally occluded or hemorrhagic nares – see Table 1).
c
Respiratory effort scores tabulated retractions, and could range from a low of 0 (no
retractions) to a high of 15 (severe supraclavicular, suprasternal, intercostal, substernal,
and subcostal retractions).
(n) ¼ number of patients available to be included in this score. (Patients not available to
be included had either no nasal exam score or no respiratory effort score at the end of the
study period, because they had ‘failed’ and were either intubated or rescued to Vapotherm.
One nasal exam was inadvertently missed on one Vapotherm patient).
NS ¼ not significant (P>0.05).
on high-flow. One nasal examination was inadvertently missed in
one subject after the period on Vapotherms. At the conclusion of
24 h on Vapotherms, nares were significantly more normal
appearing (score of 2.7±1.2), than at the conclusion of 24 h on
high-flow cannula treatment (score of 7.8±1.7, P<0.0005). At the
conclusion of 24 h on Vapotherms, respiratory effort scores were
slightly lower (less retractions) than were those on high-flow
(1.2±0.6 vs 2.0±0.9, P<0.05). When respiratory effort scores were
compared before vs at the end of the 24-h period, no subjects had
an increase in score during the period on Vapotherms. In
contrast, six subjects had an increase in respiratory effort score
during the 24-h period on high-flow (P<0.05).
The nasal gas flow used during the period on Vapotherms was
3.1±0.6 l/min, compared with 1.8±0.4 l/min during the period on
standard high-flow (P<0.01). The FiO2 used during the period on
Vapotherms was 0.31±0.04, compared with 0.32±0.04 during the
period on standard high-flow. Caffeine was administered to all of
the eight extubated subjects <1000 g birth weight, to five of the 12
who were 1000–1500 g birth weight (two in Group 1 and three in
Group 2), and to none of the 10 who were >1500 g birth weight.
The subjects on caffeine treatment all received doses while on
Vapotherm and while on high-flow.
Three of the 30 patients had a documented infection during the
hospitalization. None were in proximity to the study. One was
early-onset Escherichia coli sepsis (12 days before the study). One
developed clinical sepsis with Enterobacter cloacae grown from a
tracheal aspirate through a newly placed endotracheal tube (but
not grown from blood or spinal fluid). This occurred 10 days after
completing the study. For the 8 days preceding the onset of
Journal of Perinatology
symptoms that patient was on a low-flow (0.3 l/min) nasal
cannula (not Vapotherms). The third infection was a coagulase
negative Staphylococcus grown from one of two blood cultures
eight days following completion of the study, when the patient was
receiving low-flow (0.3 l/min) nasal cannula (not Vapotherms).
Three of the 30 patients had a documented occurrence of
extraventilatory air during the hospitalization. None were in
proximity to the study. Two had a pneumothorax and one had a
pneumomediastinum. These all occurred two to three days before
extubation and thus before study-onset. No cases of pneumothorax
or pulmonary interstitital emphysemia occurred after study-onset.
Discussion
Vapotherms was developed, in the opinion of the company, to
deliver molecular water vapor through a nasal cannula with nearly
100% relative humidity at body temperature.4–6 The device was
designed to deliver gasses in the range of 5 to 40 l/min, but for use
in neonates a ‘low-flow’ vapor transfer cartridge was developed that
allows flows, according to manufacture instructions, of 1 to 8 l/
min.4,5 Vapotherms has been used recently in many NICU’s for a
variety of indications, including as a substitute for nasal CPAP
among neonates recently extubated.9 Despite its apparent
popularity, we could find no published, peer-reviewed, randomized
trials testing Vapotherms in a NICU population.9
For the present study, we focused on the problems following
extubation from mechanical ventilation, and we postulated that
Vapotherms would perform better than a standard high-flow
nasal cannula, among recently extubated neonates. The specific
performance measures we selected were; (1) appearance of the
nasal mucosa, respiratory rate, respiratory effort score, and failure
of extubation. We used a crossover design so each study patient
would be treated with both modalities. We randomized patients
regarding which of the two test treatments would be administered
first, in case the first treatment biased or conditioned the patient
regarding the response to the second treatment.
We observed that following extubation, patients placed on
standard high-flow cannula treatment were more likely to
experience an increase in retractions than were those on
Vapotherms. Furthermore, while on Vapotherms, the patients
were more likely to have a normal appearance of their nasal
mucosa and fewer extubation failures. We speculate that the better
respiratory effort scores were, at least in part, due to the higher gas
flow during the Vapotherms period. However, despite the higher
flows, the nasal mucosa during Vapotherms appeared more
normal. We speculate that this was the result of the higher
humidity and body temperature of the gas.
During the course of our study, the Centers for Disease Control
and Prevention began an investigation of Ralstonia in association
with Vapotherms use.10 This issue had not been widely recognized
as a concern during the time our study was being designed, and

consequently no cultures of the nasal mucosa, equipment, or
trachea were included in the study protocol. However, we obtained
tracheal aspirates in the two patients who were reintubated, and
both were sterile. These two had been on high-flow and not
Vapotherms at the time of the reintubation. In no case was
Ralstonia10 or Pseudomonas recovered from any culture of any of
the 30 study patients at any time during their hospitalization.
Another concern regarding Vapotherms is the unmeasured
end-expiratory distending pressures generated, and the attendant
fear that, under some circumstances, very high continuous positive
end-expiratory pressures (CPAP) can inadvertently be
administered.9,11,12 As commented on recently by Finer,12 efforts are
needed to quantify and monitor unregulated CPAP, when using
high-flow rates through a nasal cannula. We observed no cases of
pneumothorax or pulmonary interstitial emphysemia on any study
patient between study entry and discharge home, although this was
not specifically sought as part of the study design.
We recognize various shortcomings of our study. First, although
the neonatologist conducting the nasal examinations and the
research nurse abstracting the charts were effectively blinded to the
study arm, we had no way of similarly blinding the bedside nurse,
bedside respiratory therapist, and attending neonatologist. We
realize that this could introduce a source of bias. Second, we knew
that the convenience sample of only 30 subjects would be too small
to evaluate relatively rare adverse events due to one modality or the
other. Indeed, the lack of infections and pneumothoracies among
our 30 subjects during the study does not exclude these as real risks
of Vapotherms. Third, the method by which retractions are scored
(absent, mild, moderate or severe) on our bedside charts is
subjective, thus minimizing the importance of the observed better
respiratory effort scores while on Vapotherms.
In conclusion, among 30 NICU patients in a cross-over study
immediately following endotracheal extubation, it appears that
Vapotherms performed better than a standard high-flow nasal
cannula in maintaining normal appearing nasal mucosa, a lower
respiratory effort, and averting reintubation, with no recognized
complications.
Vapotherms vs high-flow nasal cannula after extubation
DD Woodhead et al
485
Acknowledgments
The authors thank the nurses and respiratory therapists at the McKay-Dee Hospital
NICU for their valuable assistance with the study. We also thank Gorgi Rigby,
RRT, LDS Hospital NICU, Salt Lake City, UT for helpful discussions about the
research project.
References
1 Sreenan C, Lemke RP, Hudson-Mason A, Osiovich H. High-flow nasal
cannulae in the management of apnea of prematurity: a comparison with
conventional nasal continuous positive airway pressure. Pediatrics 2001;
107: 1081–1083.
2 Myers TR. AARC clinical practice guideline. Selection of an oxygen delivery
devise for neonatal and pediatric patients. Respir Care 2002; 47: 707–716.
3 Kopelman A, Holbert D. Use of oxygen cannulas in extremely low birth
weight infants is associated with mucosal trauma and bleeding, and possibly
with coagulase-negative staphylococcal sepsis. J Perinatol 2003; 23: 69–97.
4 www.vtherm.com.
5 In-house literature. Vapotherm 2004. Inc. 198 Log Canoe Circle:
Stevensville, MD.
6 Waugh JB, Granger WM. An evaluation of 2 new devices for nasal high-flow
gas therapy. Respir Care 2004; 49: 902–906.
7 A Collective Task Force Facilitated by the American College of Chest
Physicians, the American Association for Respiratory Care, and the American
College of Critical Care Medicine. Evidence-based guidelines for weaning
and discontinuing ventilatory support. Respir Care 2002; 47: 69–90.
8 Durbin Jr CG, Campbell RS, Bransone RD. AARC clinical practice guideline.
Removal of the endotracheal tube. Respir Care 1999; 44: 85–90.
9 deKlerk A. Humidified high flow nasal cannulae. Hot Topics Neonatol 2005;
4–6: 138–158.
10 Centers for Disease Control and Prevention. Ralstonia associated with
Vapotherm oxygen delivery device – United States, 2005. Morb Mortal Wkly
Rep 2005; 54: 1052–1053.
11 Locke RG, Wolfson MR, Shaffer TH, Rubinstein SD, Greenspan JS.
Inadvertent administration of positive-end-distending pressure during nasal
cannula flow. Pediatrics 1993; 91: 135–138.
12 Finer NN. Nasal cannula use in the preterm infant: oxygen or pressure?
Pediatrics 2005; 116: 1216–1217.
Journal of Perinatology