Extubation from low-rate intermittent

positive airways pressure versus

extubation after a trial of endotracheal
continuous positive airways pressure in
intubated preterm infants
Davis PG, Henderson-Smart DJ
Background - Methods - Results - Characteristics of Included Studies - References - Data
Tables & Graphs

Cover sheet
Title
Extubation from low-rate intermittent positive airways pressure versus extubation after a
trial of endotracheal continuous positive airways pressure in intubated preterm infants

Reviewers
Davis PG, Henderson-Smart DJ

Dates
Date edited: 27/05/2003
Date of last substantive update: 03/07/2001
Date of last minor update: 17/03/2003
Date next stage expected 28/08/2004
Protocol first published:
Review first published: Issue 3, 1998

Contact reviewer
Dr Peter G Davis, MD, MBBS
Consultant Paediatrician
Division of Paediatrics

Royal Women's Hospital

132 Grattan St
Melbourne
Victoria AUSTRALIA
3053
Telephone 1: +61 3 93442000 extension: 2130
Facsimile: +61 3 93471761
E-mail: pgd@unimelb.edu.au

Contribution of reviewers
Internal sources of support
Royal Women's Hospital, Melbourne, AUSTRALIA
University of Melbourne, Melbourne, Australia, AUSTRALIA
Royal Prince Alfred Hospital, Sydney, AUSTRALIA
Mothers and Infants Health Foundation of NSW, Sydney, AUSTRALIA
Murdoch Children's Research Institute, Melbourne, AUSTRALIA

External sources of support

Department of Health and Aging, Commonwealth Government, Canberra, AUSTRALIA

What's new
This review updates the previous version of "Extubation from low-rate intermittent positive
airways pressure versus extubation after a trial of endotracheal continuous positive airways
pressure in intubated preterm infants" last updated in The Cochrane Library, Issue 4, 2001.
A repeat literature search showed no new trials eligible for inclusion and there have been
no substantive changes to the review.

Dates
Date review re-formatted: 27/09/1999
Date new studies sought but none found: 17/03/2003
Date new studies found but not yet included/excluded: / /
Date new studies found and included/excluded: / /
Date reviewers' conclusions section amended: / /
Date comment/criticism added: / /
Date response to comment/criticisms added: / /

Text of review
Synopsis
No evidence that time on endotracheal CPAP (continuous low pressure rather than
intermittent breaths from the ventilator) before taking preterm babies off a ventilator helps
them adjust to breathing on their own.
Babies in neonatal intensive care often need help to breathe, sometimes via an endotracheal
tube (through the windpipe) connected to a ventilator (machine). It has been thought that it
might help a baby adjust to breathing after ventilation if there was a period of CPAP
(continuous positive airways pressure) before extubation (coming off the ventilator).
However, there have also been concerns that this may be too much hard work for the baby,
and may cause harm. The review of trials found that a trial of CPAP first before extubation
does not improve the baby's ability to breathe on their own.

Abstract
Background
Failure of extubation and subsequent reintubation may result in additional stress and trauma
to the premature infant. Testing infants about to be extubated with a period of endotracheal
CPAP has been suggested as a method of demonstrating readiness for extubation. However,
this process has been criticized as increasing the neonate's work of breathing and perhaps
increasing the likelihood of extubation failure.

Objectives
In premature infants having their endotracheal tube removed, is direct extubation from low
rate intermittent positive pressure ventilation (IPPV) more successful than that following a
period of endotracheal continuous positive airway pressure (CPAP)?

Search strategy
The standard search strategy of the Cochrane Neonatal Review Group as outlined Tn the
Cochrane Library was used. This included searches of the Oxford Database of Perinatal
Trials, Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library),
MEDLINE, previous reviews including cross references, abstracts, conferences, symposia
proceedings, expert informants and journal hand searching mainly in the English language.
These searches were updated in March 2003.

Selection criteria

All trials using random or quasi-random allocation of premature infants to endotracheal

CPAP or direct extubation following a period of IPPV were included.

Data collection & analysis

Data were extracted using standard methods of the Cochrane Collaboration and its Neonatal
Review Group, with separate evaluation of trial quality and data extraction by each author
and synthesis of data using relative risk.

Main results
Direct extubation from low rate ventilation is associated with a trend to increased chance of
successful extubation when compared to extubation after a period of endotracheal CPAP,
RR 0.45 (0.19,1.07), RD -0.103 (-0.200,-0.006), NNT 10 (5,167). When only truly
randomized trials are considered, this result becomes both statistically significant and
clinically important, RR 0.10 (0.01,0.78), RD -0.201 (-0.319,-0.083), NNT 5 (3,12). Similar
differences are seen for the secondary outcome, apnea.

Reviewers' conclusions
Preterm infants no longer requiring endotracheal intubation and IPPV should be directly
extubated without a trial of ETT CPAP.

Background
Endotracheal intubation and intermittent positive pressure ventilation (IPPV) of infants
with respiratory failure provides life-saving support until the underlying pulmonary
pathology has resolved spontaneously or been successfully treated. Weaning of ventilatory
support occurs during the resolution phase and when an infant is able to generate sufficient
respiratory effort, the endotracheal tube (ETT) is removed.
Minimising the duration of endotracheal intubation is thought to reduce the risks of chronic
lung disease, sepsis and upper airway trauma. A trial of endotracheal continuous positive
airway pressure (ETT CPAP) prior to extubation has been advocated by Nugent and others
(Nugent 1988) to demonstrate the readiness of an infant to breath without IPPV. However,
the same authors go on to suggest that in small infants, the added dead space and resistance
to gas flow imposed by the ETT may increase the work of breathing (Nugent 1988). Le
Souef and others (LeSouef 1984) demonstrated a substantial decrease in resistance of the
respiratory system following extubation. They also showed that diaphragmatic activity was
greater when an infant was on ETT CPAP and concluded that the presence of the ETT
increased the work of breathing. Lopes 1981 suggested that diaphragmatic fatigue might
precede the development of apnea in some infants.
Failure of extubation and subsequent reintubation may result in additional stress and
physical trauma to the infant. Maximising the likelihood of successful extubation is

therefore an important goal. Failure of extubation may be indicated by increasing rates of

apnea and bradycardia, increasing oxygen requirements or respiratory acidosis.
Previous reviews suggest that factors such as the use of postextubation nasal CPAP (Davis
1997), the administration of methylxanthines (Henderson-Smart 1997) and in some cases
dexamethasone (Davis P 1997), improve the likelihood of successful extubation. The use of
these therapies has implications for this comparison of extubation from low rate vs ETT
CPAP.

Objectives
In premature infants having their endotracheal tube removed, is direct extubation from low
rate ventilation more successful than extubation following a period of endotracheal CPAP?
Success was defined as avoidance of endotracheal reintubation for treatment of
apnea/bradycardia, increasing oxygen requirements and respiratory acidosis.
Subgroup analyses were planned to determine whether there were any differences between
the group as a whole and subgroups of very low birth weight (VLBW) infants, ie those <
1500g, and extremely low birth weight (ELBW) infants, ie those < 1000g. Analysis
according to level of CPAP used (less than 5 cm water vs 5 cm or greater) was also to be
undertaken if differences between studies existed.
Sensitivity analysis was planned to investigate the role of potential confounding variables
such as the use of postextubation NCPAP, methylxanthines and dexamethasone. Sensitivity
analysis according to quality of trial was to be performed if differences existed.

Criteria for considering studies for this review

Types of studies
All trials using random or quasi-random patient allocation were included.

Types of participants
Premature infants who received IPPV and were about to have their endotracheal tube
removed.

Types of interventions
The use of endotracheal CPAP for at least 6 hours prior to extubation.

Types of outcome measures

The primary outcome of interest was the need for reintubation of the trachea.

Secondary outcomes thought worthy of assessment included: rates of chronic lung disease,
growth, long term neurological development, rates of intraventricular haemorrhage and
radiological outcomes such as post-extubation atelectasis or collapse. Incidence of apneic
events was added as a post hoc assessment using the definition of greater than 0.5 events
per hour as indicating significant apnea.

Search strategy for identification of studies

See: Cochrane Neonatal Review Group search strategy
The standard search strategy of the Cochrane Neonatal Review Group as outlined in The
Cochrane Library was used. This included searches of the Oxford Database of Perinatal
Trials, The Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane
Library, Issue 1, 2003): text words - extubation and neonatal}, MEDLINE {PubMed - 1966
to March 2003: Infant, Newborn (explode) [MeSH heading] and Positive Pressure
Respiration (explode) [MeSH heading] and extubation [text word]}, previous reviews
including cross references, abstracts, conferences, symposia proceedings, expert informants
and journal hand searching mainly in the English language.

Methods of the review

Criteria and methods used to assess the methodological quality of the included trials:
The standard method of the Cochrane Collaboration and its Neonatal Review Group was
used. The methodological quality of each trial was reviewed by the second author blinded
to trial authors and institutions.
Methods used to collect data from the included trials:
Each author extracted data separately before comparison and resolution of differences.
Methods used to synthesise data:
Standard method of the Neonatal Review Group with the use of relative risk (RR) and risk
difference (RD). The number needed to treat (NNT) was calculated as the inverse of the
RD.
Subgroup analysis was planned to determine whether there was a difference in response
between the group of premature infants as a whole and VLBW and ELBW infants.
Sensitivity analysis was planned to investigate the role of potential confounding variables
such as the use of nasal CPAP post-extubation, methylxanthines and dexamethasone.
Analysis on the basis of trial quality, ie randomized or quasi-randomized, was performed
on a post hoc basis.

Description of studies
Using the search strategy described above, three clinical trials were identified which
addressed the issue of whether extubation of infants directly from IPPV was more
successful than extubation following a short period of endotracheal CPAP. In two of the

studies (Kim 1989 and Kim 1987) the ETT CPAP group was designated as the control
group. This convention is followed throughout this review. A full description of each study
is included in the Table, Characteristics of Included Studies. No trials were excluded from
the review and no ongoing trials were identified.
Kim 1987 and Tapia 1995 recruited infants less than 1250g and 1500g respectively. Kim
1989 excluded infants with size 2.5 ETTs and therefore enrolled a population that was both
heavier (mean birthweight 1.93kg) and more mature (mean gestational age 33 weeks).
Almost all infants studied were treated prophylactically with aminophylline and all were
weaned to minimal levels of IPPV before study commencement. Surfactant usage was
confined to a small subgroup of babies recruited in the latter stages of the trial of Tapia. All
three studies excluded infants with conditions which might make successful extubation less
likely, eg neuromuscular disorders, patent ductus arteriosus, CNS depressant medication.
Following removal of their endotracheal tube, infants in the trials of Kim 1989 and Kim
1987 were placed in headbox oxygen. The trial of Tapia 1995 allocated infants to either
headbox oxygen or nasal CPAP following direct extubation and to headbox only following
ETT CPAP. The use of dexamethasone, either as a short course to facilitate extubation or as
a longer course as treatment for chronic lung disease, is not recorded in any of the trials.
All three trials used low levels of ETT CPAP, ranging from as low as 2 cm water in the trial
of Kim and Boutwell, to a maximum of 4 cm in the trial of Tapia.
The criteria for reintubation were similar across the studies. These included frequent apnea
(greater than 2/hour for Kim 1989 and Kim 1987 and greater than 3/hour for Tapia) or
apnea requiring positive pressure ventilation and respiratory acidosis. Tapia had the
additional indication for reintubation of the need for greater than 60% inspired oxygen
concentration.
It was possible to derive the proportion of infants having clinically important apneic events,
ie more frequently than 0.5/hour, in the trials of Kim 1989 and Kim 1987. This was not
possible in the trial of Tapia 1995 in which only the presence or absence of apnea was
noted.
No trials reported rates of IVH, CLD, growth or Xray changes determined throughout the
in-hospital stay or long term neurodevelopmental follow-up.

Methodological quality of included studies

Full details are provided in the Table, Characteristics of Included Studies.
Method of subject allocation:
The trials of Kim 1989 and Kim 1987 were randomized controlled trials although only the
former specified the use of sequentially numbered sealed envelopes. The trial of Tapia 1995
used a method of allocation in which subjects were sequentially allocated to each of 3
groups (communication with author).

Masking of caregivers:
The nature of the intervention meant that caregivers were aware of the group of assignment.
Completeness of outcome assessment:
This was greater than 90% in all trials.
Masking of outcome assessors:
This was not performed.

Results
There was no disagreement between reviewers with respect to quality assessment done
independently.
The use of a period of ETT CPAP prior to removal of the endotracheal tube confers no
advantage over direct extubation. Using the summary statistic, relative risk (RR), there is a
trend favouring direct extubation which does not reach statistical significance 0.45 (0.19,
1.07); however, the risk difference (RD) shows a significant difference of -0.103 (-0.200,0.006) giving a number needed to treat (NNT) of 10 (5,167). When the quasi-randomized
trial of Tapia is excluded a statistically significant benefit of direct extubation is observed:
summary RR 0.10 (0.01,0.78), RD -0.201 (-0.319,-0.083) and NNT 5 (3,12).
Sub-analysis on the basis of care received after removal of the endotracheal tube shows that
the benefits of direct extubation are sustained when extubation is to headbox oxygen - RR
0.21 (0.06.0.72), RD -0.151 (-0.246,-0.057), NNT 7 (4,18). There were no trials in which
both groups were managed with NCPAP after removal of the endotracheal tube. There is no
difference between direct extubation and ETT CPAP groups when the direct extubation
group only is managed with NCPAP. This result should be interpreted with caution given
that the same trial of Tapia was the "outlier" in the meta-analysis of trials evaluating the
usefulness of post-extubation NCPAP (Davis and Henderson-Smart) which concluded that
NCPAP was more effective than headbox oxygen.
The use of ETT CPAP is associated with a trend towards increased frequency of apnea
when all three trials are considered. Exclusion of the trial by Tapia on the basis of quality
again leads to this result achieving statistical significance: RR 0.15 (0.03,0.77), RD -0.202
(-0.337,-0.067), NNT 5 (3,15).
Planned subgroup analysis on the basis of the level of ETT CPAP applied (high or low) was
not performed because all trials used a pressure less than 5 cm water.

Discussion
The use of a trial of ETT CPAP, at pressures of 2 to 4 cm water, as the final stage of
weaning ventilator rate prior to extubation is not supported by this meta-analysis.
Successful extubation is more likely if infants are extubated directly from a low rate of

IPPV. Physiological reasons for this result have been mentioned previously - when small
infants breathe through endotracheal tubes with narrow diameters, airway resistance and
therefore work of breathing increase to a point where spontaneous respirations are unable to
be sustained. It is possible that higher levels of CPAP might be more effective but there are
no data available to evaluate this.
The potential confounding variables, use of methylxanthines and post-natal steroids, were
not able to be assessed because of the almost universal usage of the former and lack of
usage of the latter. Nasal CPAP has been shown in a previous review to be effective in
facilitating extubation. Its impact on the question posed in the current review is not able to
be determined. The two trials that were truly randomized did not use NCPAP for postextubation care and the quasi-randomized trial of Tapia used this method of treatment only
for infants directly extubated from IPPV.
Direct subgroup analysis on the basis of birthweight is not possible using the published data
from these trials. However inferences may be drawn on the basis of the trials of Kim 1989
and Kim 1987. The former trial enrolled only infants with a size 2.5 mm endotracheal tube
and the latter excluded these infants and enrolled only infants having 3.0 and 3.5 mm
diameter tubes. Since choice of size of endotracheal tube is influenced mainly by weight
these trials provide information relevant to the question of whether the effect of ETT CPAP
differs according to weight of the neonate. It appears that ETT CPAP is more strongly
associated with adverse outcomes in smaller infants (with smaller endotracheal tubes) than
it is in larger infants.

Reviewers' conclusions
Implications for practice
Premature infants no longer requiring endotracheal intubation and IPPV should be directly
extubated without a trial of ETT CPAP.

Implications for research

Future trials may address the issue of whether higher levels of ETT CPAP are useful in this
situation.

Acknowledgements
The reviewers wish to acknowledge the additional information kindly provided by Dr.
Tapia in regard to his 1995 trial.

Potential conflict of interest

None

Characteristics of included studies

Allocation
concealme
nt

Stud
Methods
y

Intervention
Participants
Outcomes
s

Kim 1) Blinding
1987 of
randomizati
on - yes:
method not
specified
2) Blinding
of
intervention
- no
3) Complete
followup yes
4) Blinding
of outcome
measuremen
t - no

Included:
infants
<1250g,
ventilated
more than 12
hours but less
than 28 days
via 2.5mm
endotracheal
tube. Almost
all (23/27)
treated with
aminophyllin
e. Reached
minimal
levels of
ventilation rate 6 to 10,
PIP < 16,
PEEP = 3
(>1kg) or 2
(<1kg).
Excluded:
infants with
neuromuscul
ar disorders
or on CNS
depressant
medication.

Direct
extubation
from a
ventilator
rate of
between 6
and 10 or
extubation
following a 6
hour trial on
endotracheal
CPAP (at a
level = PEEP
before
randomizatio
n)

Failure = need
to resume IPPV
because either
1) 2 or more
episodes of
apnea or
bradycardia
requiring
vigorous
stimulation
within 30
minutes, 2) any
use of IPPV for
profound apnea
or bradycardia
or 3) respiratory
acidosis with
pH < 7.3 (CO2
not specified)

Length of B
time after
extubation
during
which an
infant
could fail
is not
specified.
Obligatory
crossover
after
failure.

Kim 1) Blinding
1989 of
randomizati
on - yes:
sequentially
numbered
sealed
envelopes
2) Blinding
of

Included: all
newborn
infants
intubated
with a size
3.0 or 3.5
ETT and
ventilated >
12 hours. All
infants < 37

Direct
extubation
from IPPV or
extubation
following 6
hours of ETT
CPAP (3 cm
water). Both
groups
extubated to

Resumption of
IPPV for either
:
1) 2 episodes of
severe
apnea/bradycard
ia requiring
vigorous
stimulation
within 30

Infants
A
stratified
according
to diameter
of ETT ie
3.0 or 3.5
mm.
After
failure,
infants

Notes

intervention
- no
3) Complete
followup yes
4) Blinding
of outcome
measuremen
t - no

Tapi 1) Blinding
a
of
1995 randomizati
on - no:
sequential
assignment
of infants to
each of 3
groups
2) Blinding
of
intervention
- no
3) Complete
follow-up yes (3/90
lost after
randomizati
on because
accidental
extubation)
4) Blinding
of outcome
measuremen
t - no

weeks
headbox
gestation
oxygen.
received
theophylline.
Excluded:
infants with
neuromuscul
ar disorders,
apnea and
bradycardia
in 24 hours
preextubation,
CNS
depressant
medication.

minutes
2)
apnea/bradycard
ia requiring
IPPV
3) Respiratory
acidosis with
pH < 7.30
Assessed for the
6 hour period
following
randomization.

were
crossed
over to the
alternate
strategy.

Included:
preterm
infants <
1500g
admitted to
either of 2
participating
centres,
ventilated for
> 48 hours
and weaned
to minimal,
defined
levels of
ventilatory
support
Excluded:
infants with
sepsis,
necrotizing
enterocolitis,
symptomatic
patent ductus
arteriosus,
major
congenital
malformation
s,
neuromuscul
ar disorders,

Failure of
extubation
indicated by the
need for the
reventilation
within 72 hours
for
1) apnea frequent or
severe,
2) respiratory
acidosis (pH <
7.25 and pCO2
> 60, or
3) need for
inspired oxygen
concentration >
60%.

QuasiC
randomize
d design.
Allocation
by
"systemati
c
sampling"
was in fact
sequential
assignment
to 3 groups
in rotation
(informatio
n from
author).
Surfactant
usage was
confined to
one of two
centres
involved
and for
less than
one third
of the
trial's
recruitmen
t time.

Direct
extubation
from IPPV to
either
headbox or
NCPAP or
extubation to
headbox
following 12
-24 hours of
ETT CPAP
(3-4 cm
water).

grade 3 or 4
intraventricul
ar
hemorrhage
or other
severe
neurological
damage.
PIP = peak inspiratory pressure; PEEP = positive end expiratory pressure; CNS = central
nervous system

References to studies
References to included studies
Kim 1987 {published data only}
Kim EH, Boutwell WC. Successful direct extubation of very low birth weight infants from
low intermittent mandatory ventilation rate. Pediatrics 1987;80:409-414.
Kim 1989 {published data only}
Kim EH. Successful extubation of newborn infants without preextubation trial of
continuous positive airway pressure. J Perinatol 1989;9:72-76.
Tapia 1995 {published data only}
Tapia J, Bancalari A, Gonzalez A, Mercado M. Does continuous positive airways pressure
(CPAP) during weaning from intermittent mandatory ventilation in very low birthweight
infants have risks or benefits? A controlled trial. Ped Pulmonol 1995;19:269-279.
* indicates the primary reference for the study

Other references
Additional references
Davis 1997
Davis PG, Henderson-Smart DJ. Prophylactic post-extubation nasal CPAP in preterm
infants (Cochrane Review). In: The Cochrane Library, Issue 4, 1997. Oxford: Update
Software.

Davis P 1997
Davis PG, Henderson-Smart DJ. Intravenous dexamethasone in neonates being extubated
(Cochrane Review). In: The Cochrane Library, Issue 4, 1997. Oxford: Update Software.
Henderson-Smart 1997
Henderson-Smart DJ, Davis PG. Prophylactic methylxanthine for extubation in preterm
infants (Cochrane Review). In: The Cochrane Library, Issue 4, 1997. Oxford: Update
Software.
LeSouef 1984
LeSouef PN, England SJ, Bryan AC. Total resistance of the respiratory system in preterm
infants with and without an endotracheal tube. J Pediatr 1984;104:108-111.
Lopes 1981
Lopes JM, Muller NL, Bryan MH, Bryan AC. Synergistic behaviour of inspiratory muscles
after diaphragmatic fatigue in the newborn. J Appl Physiol 1981;51:547-551.
Nugent 1988
Nugent J, Matthews BJ, Goldsmith JP. Pulmonary care. In: Goldsmith and Karotkin,
editor(s). Assisted ventilation of the neonate. Philadelphia: WB Saunders, 1988:103-4.

Other published versions of this review

Davis 1998
Davis PG, Henderson-Smart DJ. Extubation from low-rate intermittent positive airway
pressure versus extubation after a trial of endotracheal continuous positive airway pressure
in intubated preterm infants (Cochrane Review). In: The Cochrane Library, Issue 3, 1998.
Oxford: Update Software.
Davis 2001
Davis PG, Henderson-Smart DJ. Extubation from low-rate intermittent positive airways
pressure versus extubation after a trial of endotracheal continuous positive airways pressure
in intubated preterm infants (Cochrane Review). In: The Cochrane Library, Issue 4, 2001.
Oxford: Update Software.

Comparisons and data

01 Extubation from IPPV vs ETT CPAP

01.01 Extubation failure

01.02 Extubation failure (extubated to headbox)
01.03 Extubation failure (extubated to NCPAP)
01.04 Extubation failure (true RCTs only)
01.05 Apnea
01.06 Apnea (true RCTs only)

Comparison or outcome
Studies
01 Extubation from IPPV vs ETT CPAP
01 Extubation failure
3
02 Extubation failure (extubated to
3
headbox)
03 Extubation failure (extubated to
1
NCPAP)
04 Extubation failure (true RCTs
2
only)
05 Apnea
3
06 Apnea (true RCTs only)
2

Participants

Statistical method

Effect size

174

RR (fixed), 95% CI

0.45 [0.19, 1.07]

145

RR (fixed), 95% CI

0.21 [0.06, 0.72]

57

RR (fixed), 95% CI

1.69 [0.55, 5.14]

87

RR (fixed), 95% CI

0.10 [0.01, 0.78]

174
87

RR (fixed), 95% CI
RR (fixed), 95% CI

0.66 [0.35, 1.25]

0.15 [0.03, 0.77]

Notes
Published notes
Amended sections
Cover sheet
Abstract
Search strategy for identification of studies
References to studies
Other references

Contact details for co-reviewers

Prof David J Henderson-Smart
Director
NSW Centre for Perinatal Health Services Research

Queen Elizabeth II Institute for Mothers and Infants

Building DO2
University of Sydney
Sydney
NSW AUSTRALIA
2006
Telephone 1: +61 2 93517318
Telephone 2: +61 2 93517728
Facsimile: +61 2 93517742
E-mail: dhs@perinatal.usyd.edu.au
This review is published as a Cochrane review in The Cochrane Library 2003, Issue 3,
2003 (see www.CochraneLibrary.net for information). Cochrane reviews are regularly
updated as new evidence emerges and in response to comments and criticisms, and The
Cochrane Library should be consulted for the most recent version of the Review.