ORIGINAL RESEARCH

Airway Management Strategies for Brain-injured Patients Meeting

Standard Criteria to Consider Extubation
A Prospective Cohort Study
Victoria A. McCredie1,2,3*, Niall D. Ferguson1,2,4,5,6,7*, Ruxandra L. Pinto2,3, Neill K. J. Adhikari1,3,4,8,
Robert A. Fowler1,2,3,4,8, Martin G. Chapman1,3, Althea Burrell9, Andrew J. Baker1,10, Deborah J. Cook11,
Maureen O. Meade11, and Damon C. Scales1,2,3,4,8; for the Canadian Critical Care Trials Group
1
Interdepartmental Division of Critical Care and 2Institute for Health Policy, Management and Evaluation, University of Toronto, Toronto,
Ontario, Canada; 3Department of Critical Care Medicine, Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada; 4Department
of Medicine and 5Department of Physiology, University of Toronto, Toronto, Ontario, Canada; 6Division of Respirology, Department of
Medicine, University Health Network and Mount Sinai Hospital, Toronto, Ontario, Canada; 7Toronto General Research Institute, Toronto,
Ontario, Canada; 8Sunnybrook Research Institute, Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada; 9Division of
Respirology, Department of Medicine, Markham Stouffville Hospital, Markham, Ontario, Canada; 10Department of Critical Care
Medicine, St. Michael’s Hospital, Toronto, Ontario, Canada; and 11Department of Medicine and Department of Clinical Epidemiology
and Biostatistics, McMaster University, Hamilton, Ontario, Canada
ORCID ID: 0000-0002-1590-2846 (V.A.M.).

Abstract
Rationale: Patients with acute brain injury are frequently capable of
breathing spontaneously with minimal ventilatory support despite
persistent neurological impairment.
Objectives: We sought to describe factors associated with extubation
timing, success, and primary tracheostomy in these patients.
Methods: We conducted a prospective multicenter observational
cohort study in three academic hospitals in Toronto, Canada.
Consecutive brain-injured adults receiving mechanical ventilation for
at least 24 hours in three intensive care units were screened by study
personnel daily for extubation consideration criteria. We monitored all
patients until hospital discharge and used logistic regression models to
examine associations with extubation failure and delayed extubation.
Measurements and Main Results: Of 192 patients included,
152 (79%) were extubated and 40 (21%) received a tracheostomy without
an extubation attempt. The rate of extubation failure within 72 hours was
32 of 152 (21%), which did not vary significantly between those extubated
before (early; 6 of 37; 16.2%), within 24 hours (timely; 14 of 70; 20.0%), or
more than 24 hours after meeting criteria to consider extubation (delayed;
12 of 45; 26.7%; P = 0.49). Delayed extubation was associated with lower
a Glasgow Coma Scale (GCS) score at the time of consideration of
extubation, absence of cough, and new positive sputum cultures. Factors
independently associated with successful extubation were presence of
cough (odds ratio [OR], 3.60; 95% confidence interval [CI], 1.42–9.09),
fluid balance in prior 24 hours (OR, 0.75 per 1-L increase; 95% CI, 0.57–
0.98), and age (OR, 0.97 per 10-yr increase; 95% CI, 0.95–0.99). A higher
GCS score was not associated with successful extubation.
Conclusions: Extubation success was predicted by younger age,
presence of cough, and negative fluid balance, rather than GCS score
at extubation. These results do not support prolonging intubation
solely for low GCS score in brain-injured patients.
Keywords: acute brain injury; extubation; tracheostomy;
mechanical ventilation; weaning

(Received in original form August 21, 2016; accepted in final form October 21, 2016 )
*V.A.M. and N.D.F. share joint first authorship.
Supported by the Faculty of Medicine Dean’s Fund and Connaught Matching Grant, University of Toronto. D.C.S. was supported by a Fellowship in
Translational Health Research from the Physicians’ Services Incorporated Foundation. D.J.C. is a Canada Research Chair of the Canadian Institutes of Health
Research. The authors have no other disclosures and no competing interests to declare. The opinions, results, and conclusions reported in this article are
those of the authors and are independent of funding sources.
Ethics Approval: This study was approved by the research ethics board of each institution; all waived the need for informed consent.
Ann Am Thorac Soc Vol 14, No 1, pp 85–93, Jan 2017
Copyright © 2017 by the American Thoracic Society
DOI: 10.1513/AnnalsATS.201608-620OC
Internet address: www.atsjournals.org

McCredie, Ferguson, Pinto, et al.: Airway Management Strategies for Brain-Injured Patients 85
 ORIGINAL RESEARCH

Acutely brain-injured patients are frequently
intubated and mechanically ventilated
(1). In contrast to the general medical–
surgical intensive care unit (ICU)
population, these patients usually do not
have a primary respiratory indication for
intubation and ventilatory support (2).
Even those with persistent and severe
neurological compromise often appear
capable of breathing spontaneously (3, 4),
Methods Table 1. Criteria for extubation
consideration
Enrollment Criteria
Consecutive patients admitted to three d Glasgow Coma Scale score > previous
mixed medical–surgical and neuroscience day
d Intracranial pressure , 20 mm Hg
ICUs in tertiary academic centers were
d Cerebral perfusion pressure . 60 mm Hg
screened for study eligibility in 2005. All d PaO /FIO . 200 mm Hg
ICUs were staffed by intensivists who 2 2
d PEEP < 5 cm H2O
made primary decisions about airway d Temperature , 38.58 C
management. Patients were enrolled if they d Mean arterial pressure . 60 mm Hg and

a state that would usually lead to prompt
extubation in a general ICU patient (5).
However, relatively little evidence is
available to guide extubation decision
management in the acutely brain-injured
patient who is tolerating minimal
ventilatory support (6).
In this setting clinicians may extubate
promptly, they may delay extubation, or
they may proceed to primary tracheostomy
without an extubation attempt because of
concerns about poor pulmonary hygiene or
the risk of aspiration pneumonia in patients
with a diminished level of consciousness.
The extubation failure rate in this
population may be higher than that of the
medical–surgical ICU population (7–12),
and failed extubation has been correlated
with the development of pneumonia and
increased mortality (8, 13–17). Conversely,
delaying extubation probably increases the
risk of ventilator-associated pneumonia
(12).
The optimal timing of extubation
of acutely brain-injured patients, once
standard criteria to consider extubation are
met, is unknown. It is also unclear
whether primary tracheostomy is preferable
to a trial of elective extubation. We therefore
set out to determine the rate and factors
associated with successful extubation and
extubation delay in this clinical setting.
In addition, we examined the impact
of extubation timing and decisions to
extubate or perform a tracheostomy on
clinical outcomes in the ICU and hospital.
Some of these results have been
previously reported in the form of an
abstract (18).
had an acute brain injury (see additional
methods in the online supplement), had
been receiving mechanical ventilation for
more than 24 hours, were over 16 years of
age, and had not been previously enrolled
in this study. We subsequently excluded
patients who died before extubation or
tracheostomy, who self-extubated, or
who were extubated in association with
withdrawal of life-sustaining therapies. The
study was approved by the research ethics
board of each institution; all waived the
need for informed consent.
Data Collection
Once patients were enrolled, we collected
baseline and demographic data and
screened daily for criteria to consider
extubation (Table 1) (12, 19). Patients
meeting these criteria were assessed daily
for physiological parameters known or
suspected to influence extubation outcome:
24-hour fluid balance, 24-hour sputum
and suction count, presence of gag reflex,
spontaneous cough or cough with
suctioning, rapid shallow breathing index
(frequency/tidal volume), minute
ventilation, vital capacity, negative
inspiratory force, and the presence of an
endotracheal tube cuff leak (see the online
supplement for additional methodology).
We deemed extubation to be prompt if it
occurred either before study criteria for
extubation consideration were met (early
extubation group) or within 1 day of
meeting these criteria (timely extubation
group). Patients extubated more than 1 day
after meeting criteria to consider extubation
were defined as delayed (delayed extubation
no vasoactive drugs
d No surgery planned in next 72 h
d No order for hyperventilation
d Has passed spontaneous breathing trial or
tolerates pressure support < 7 cm H2O
Definition of abbreviations: PaO2/FIO2 = partial
pressure of arterial oxygen/fraction of inspired
oxygen; PEEP = positive end-expiratory
pressure.
group). Extubation failures included any
reintubation within 72 hours; the reason for
reintubation was documented.
Outcome Measures
We recorded the following clinical outcomes
for all patients: extubation failure (in those
extubated), ventilation days, ICU and
in-hospital mortality, ICU and in-hospital
lengths of stay, and hospital-acquired
pneumonia (see the online supplement
for pneumonia diagnostic criteria).
Statistical Analysis
Descriptive statistics were reported as means
with standard deviations and medians with
interquartile ranges (IQRs) for continuous
data, and counts and percentages for
categorical data. Comparisons between
groups were made by x2 or Fisher exact
test for categorical variables and by t test
or Wilcoxon rank-sum test as appropriate.
A multivariable logistic regression
model was used to identify independent
factors at time of readiness and baseline
covariates associated with extubation
success and delay; only those patients who
were electively extubated were considered
in this model (n = 152). Covariates

Author Contributions: V.A.M., N.D.F., and D.C.S. contributed to the literature search, study design, data analysis, statistical analysis, data interpretation, writing, and
critical revision. A.B. contributed to the literature search, data analysis, data interpretation, writing, and critical revision. R.L.P. contributed to the data analysis and critical
revision. N.K.J.A., R.A.F., M.G.C., A.J.B., D.J.C., and M.O.M. contributed to the study design, data collection, data analysis and interpretation, and critical revision. All
authors listed have contributed sufficiently to the project to be included as authors, and all those who are qualified to be authors are listed in the author byline.
Correspondence and requests for reprints should be addressed to Niall D. Ferguson, M.D., M.Sc., Toronto General Hospital, 585 University Avenue, 11-PMB-
120, Toronto, ON, M5G 2N2 Canada. E-mail: n.ferguson@utoronto.ca
This article has an online supplement, which is accessible from this issue’s table of contents at www.atsjournals.org.

86 AnnalsATS Volume 14 Number 1 | January 2017

 ORIGINAL RESEARCH
thought to be clinically important or with
a P value less than 0.2 in univariable
analysis were potentially considered
(age, Glasgow Coma Scale [GCS] score,
presence of cough or gag, fluid balance,
and PaO2/FIO2 [partial pressure of arterial
oxygen/fraction of inspired oxygen]) (20).
We excluded PaO2/FIO2 from the multiple
logistic regression model as it was not
significant in the presence of the other
covariates (P = 0.51) and had a number
of missing data points (n = 16 patients).
Gag reflex and cough (spontaneous
or provoked) were highly correlated
(correlation coefficient, .0.8); we
15 Excluded from cohort
9 Self-extubated/unplanned extubation
4 Transferred to another hospital
1 Tracheostomy before screening
1 Mechanical ventilation <24 hours
45 Excluded from cohort
Died prior to extubation or extubated in
association with withdrawal of
life-sustaining therapies
40 Primary tracheostomy
20 Early
2 Timely
15 Delayed
1 Wrong data
2 Didn’t meet extubation readiness criteria
in the first 28 days
120 Successfully extubated
10 Reintubated after >72 hrs
4 Secondary
6 Extubated
tracheostomy
Figure 1. CONSORT diagram for NEURO pilot study. CONSORT = Consolidated Standards of Reporting Trials; NEURO = Neurological ICU Extubation
Strategy Utilisation and Reintubation Outcomes. When patients met enrollment criteria they were screened daily for criteria to consider extubation. When a cohort
inclusion/exclusion outcome was attained patients underwent daily observation of additional airway and ventilatory parameters and pneumonia.
McCredie, Ferguson, Pinto, et al.: Airway Management Strategies for Brain-Injured Patients
therefore considered models with Results
gag reflex and cough separately and
retained the model with the lowest Of 252 patients enrolled with acute brain
Akaike information criterion (21). We injury, 192 met all criteria for cohort
used a manual backward elimination inclusion (Figure 1). Elective extubation
technique retaining all variables with occurred in 152 of 192 patients (79%)
P , 0.2 in the final multivariable model and primary tracheostomy occurred in 40
and forcing the GCS score in the model, (21%); an additional 12 patients (8% of
as decided a priori (22). those electively extubated) received a
Model fit was assessed using the secondary tracheostomy. More than 80%
Hosmer–Lemeshow test for goodness of fit. of all patients reached consideration of
All data were analyzed with SAS software extubation criteria by 5 days (110 of
(version 9.3; SAS Institute, Cary, NC), and 132 patients; Figure 2). Early extubation
a two-tailed P value less than 0.05 was (i.e., extubation before study criteria to
considered to be statistically significant. consider extubation were met) occurred
252 Enrolled* *Enrollment criteria:
Age >16 years
Mechanical ventilation >24 hours
Acute brain injury new on hospital
admission+
Cohort inclusion/exclusion
outcome attained **
+ Acute brain injury:
Subarachnoid hemorrhage
Traumatic brain injury
Ischaemic stroke
192 Included in cohort Subdural hematoma
Epidural hematoma
Post-craniotomy
Intracerebral hemorrhage
Global cerebral ischemia
Status epilepticus
Meningitis/encephalitis
Brain abscess
Other
** Inclusion outcome:
Elective extubation
Primary tracheostomy
Exclusion outcome:
152 Elective extubations Unplanned extubation
37 Early extubation Death prior to extubation or
70 Timely extubation extubated in association with
45 Delayed extubation withdrawal of life-sustaining
treatment
32 Reintubated ≤72 hours
8 Secondary
24 Extubated
tracheostomy
87

1.0
Proportion meeting criteria to consider extubation
0.8
0.6
0.4
0.2
0.0 192
0 10
Time from intubation to meeting criteria to consider extubation
Figure 2. Days to meeting criteria to consider extubation from day of intubation. 1Patients censored
if extubation attempted or tracheostomy performed before meeting readiness criteria. Note: 132 of
192 patients met extubation criteria overall.
in 37 of 152 elective extubation patients
(24%). The majority of these patients (27
of 37; 73%) met all but 1 or 2 of the 11
preset study criteria; the criteria absent
most commonly were pressure support less
than or equal to 7 cm H2O (27 of 37; 73%)
and PaO2/FIO2 greater than or equal to
200 mm Hg (11 of 37; 30%).
Compared with patients who received a
primary tracheostomy, patients who were
electively extubated had higher admission
GCS and lower APACHE II (Acute
Physiology and Chronic Health Evaluation
II) scores. There were also differences
between these groups for indications for
mechanical ventilation. Acute brain injury
was less commonly the primary reason for
intubation among those who received a
primary tracheostomy (73%) compared
with those who were electively extubated
(92%, P = 0.002) (Table 2).
Factors Associated with
Extubation Delay
Of those electively extubated, extubation
occurred promptly, before or within 1 day
of meeting criteria to consider extubation,
in 107 patients (70%), and was delayed
in the other 45 patients (30%) (Figure 3;
and see Table E1 in the online supplement).
88
Censored
26 7 3 0 with prompt extubation (2 of 45 [4%] vs.
20 30 40
Delayed extubation was associated with
a lower median GCS score at the time of
meeting criteria to consider extubation,
no endotracheal cuff leak, and a higher
PaCO2 (Table E2). In the multivariable
analysis, lower GCS motor score at the
time of meeting criteria to consider
extubation, absence of spontaneous cough,
and new positive sputum cultures were
associated with delayed extubation (Tables E3
and E4). Extubation failure rates were
similar between those extubated promptly
(20 of 107, 19%) versus those who were
delayed (12 of 45, 27%; P = 0.27).
Factors Associated with
Successful Extubation
Most patients (120 of 152; 79%) remained
extubated for 72 hours after elective
extubation. In patients who failed
extubation, the most common reasons
for reintubation were airway problems
secondary to upper airway obstruction,
secretions, or decreased level of
consciousness (18 of 32, 56%). The
remaining patients failed extubation because
of respiratory failure (10 of 32, 31%) or other
issues unrelated to airway or respiratory
parameters (4 of 32, 13%). Table 3 reports
variables that were associated with
AnnalsATS Volume 14 Number 1 | January 2017
ORIGINAL RESEARCH
successful extubation in the unadjusted
analysis. Only three factors associated
with successful extubation were identified
in the final multivariable regression model
(Table 3): younger age, presence of
spontaneous or provoked cough, and lower
total fluid balance for the previous day.
Neurological status, as assessed by GCS
score at admission, on the day of meeting
criteria to consider extubation, or on
the day of extubation was not predictive
of extubation success (Tables 3 and 4,
Figure 4, and Tables E5–E7). A sensitivity
analysis treating GCS score as a categorical
variable produced similar results.
ICU and In-Hospital Mortality
Delayed extubation was not associated with
increased ICU mortality when compared
3 of 107 [3%]; P = 0.63) (Table E8), but was
associated with higher in-hospital mortality
(8 of 45 [18%] vs. 4 of 107 [4%]; P , 0.01).
ICU mortality was lower in patients who
were successfully extubated compared with
those who failed extubation (1 of 120 [1%]
vs. 4 of 32 [13%]; P = 0.01), but in-hospital
mortality was similar (Table 4). Among
all patients who met study criteria to
consider extubation, the ICU mortality rate
was 4 of 26 (15%) in the group failing
extubation and 5 of 17 (29%) in the
primary tracheostomy group (P = 0.44)
(Table E8); in-hospital mortality rates were
also similar between these two groups.
ICU and In-Hospital Length of Stay
Patients with delayed extubation had longer
median ICU stays (13 [IQR, 8–17] vs.
6 [IQR, 4–13] d; P , 0.01) and longer
median in-hospital lengths of stay
(28 [IQR, 22–42] vs. 22 [IQR, 15–31] d;
P = 0.03) compared with patients who were
extubated promptly (Table E8). Patients
who failed extubation had longer median
ICU stays than successfully extubated
patients (14 [IQR, 8–18] vs. 7 [IQR, 4–14]
d, respectively; P , 0.01). However
in-hospital lengths of stay were similar
(Table 4).
Days of Mechanical Ventilation
By definition, patients with delayed
extubation experienced significantly more
days of ventilation after meeting criteria to
consider extubation (median, 5 [IQR, 4–7]
vs. 1 [IQR, 0–2] d for patients extubated
promptly; P , 0.01) (Table E9). The
primary tracheostomy group accumulated
 ORIGINAL RESEARCH

Table 2. Baseline characteristics of enrolled patients We found three independent factors

Characteristic
Enrolled patients, n (%)
Age (yr), mean (SD)
Male, n (%)
Cause of neurological injury,* n (%)
Traumatic brain injury
Subarachnoid hemorrhage
Subdural hematoma
Intracerebral hemorrhage
Postcraniotomy
Other†
Primary neurological cause for
mechanical ventilation,
n (%) (n = 189)
Acuity scores
Admission total GCS, median
(IQR) (n = 192)
Abbreviated GCS on day of
readiness, median (IQR)‡
Abbreviated GCS on day of
extubation/tracheostomy,
median (IQR)
APACHE II, mean (SD) (n = 191)
Hunt and Hess grade, median
(IQR) (n = 77)
Injury severity score, mean
(SD) (n = 100)
Definition of abbreviations: abbreviated GCS = eye and motor components only of the Glasgow Coma
Scale; APACHE II = Acute Physiology and Chronic Health Evaluation II; IQR = interquartile range; total
GCS = total Glasgow Coma Scale.
*More than one cause of neurological injury possible.
†
Other includes ischemic stroke, epidural hematoma, global cerebral ischemia, status epilepticus,
meningitis/encephalitis, and brain abscess.
‡
Patients were included if they met readiness criteria.
significantly more ventilation days than
the failed extubation group from time of
intubation to extubation or tracheostomy
(median, 13 [IQR, 10–19] vs. 5 [IQR, 3–8]
d; P , 0.01). Of patients in the primary
tracheostomy group who met consideration
of extubation criteria (n = 17), the median
duration of mechanical ventilation was 6
(IQR, 4–14) days.
Pneumonia
During the 14 days after meeting criteria
to consider extubation, 12 of 32 patients
(38%) who failed extubation and 12 of
120 patients (10%) who were successfully
extubated developed pneumonia (P , 0.01)
(Table 4). Over the same time frame, 13 of
107 promptly extubated patients (12%)
and 11 of 45 patients (24%) with delayed
extubation developed pneumonia (P = 0.06)
(Table E9). Among 17 patients who
received a primary tracheostomy after
meeting criteria to consider extubation,
6 (35%) developed pneumonia (Table E8).
associated with extubation success (lower
Elective Primary P Value age, a less positive fluid balance, and
Extubation Tracheostomy presence of cough). Fluid balance was
previously shown to predict extubation
152 40 success among general ICU patients who
50 (19) 53 (22) 0.52 have passed a spontaneous breathing trial (3).
103 (68) 27 (68) 0.97 Likewise, in the medical ICU higher cough
peak flow and lower secretion burden
46 (30) 15 (38) 0.38
39 (26) 12 (30) 0.58 both predicted extubation success (24).
36 (24) 14 (35) 0.15 In a more comparable population
24 (16) 9 (23) 0.32 of neurosurgical patients, Coplin and
14 (9) 1 (3) 0.20 colleagues also found that the presence
40 (26) 7 (18) 0.25
137 (92; n = 149) 29 (73; n = 40) 0.002 of spontaneous cough at the time of
extubation was associated with successful
extubation (12).
The role that level of consciousness
7 (4–9; n = 152) 5 (4–8; n = 40) 0.01 should play in making extubation decisions
9 (7–10) 7 (5–9) ,0.001 for brain-injured patients remains
controversial. Patients are frequently
9 (8–10) 6 (4–9) ,0.001 intubated and ventilated because of coma
(25), so it may be intuitive that they should
14 (6; n = 151) 23 (10; n = 40) ,0.001 be awake before we consider extubation.
3 (2–4; n = 66) 4 (3–4; n = 11) 0.50 In a prior study of factors associated
with extubation success among 100
24 (13; n = 80) 26 (13; n = 20) 0.51 neurosurgical patients, Namen and
colleagues found that GCS score on the day
of extubation was associated with successful
extubation (OR, 1.24; 95% CI, 1.1–1.4, for
every increment in GCS score) (10). This
study was confounded, however, by
inclusion of patients who were being
extubated in the context of withdrawal of
life support. Many patients who failed
extubation were not reintubated and were
Discussion allowed to die; these patients were likely
overrepresented in the population with
This is the largest study to date examining low GCS score on extubation day.
extubation delays and failures in acutely In contrast, Coplin and coworkers did
brain-injured patients. We found that not find GCS score to be predictive of
among those not selected for primary extubation failure in a group of 136 acutely
tracheostomy, factors associated with brain-injured patients who met standard
extubation success included lower age, a weaning criteria before extubation (12). In
less positive fluid balance, and presence of that study, 10 of 11 patients (91%) with
cough, but not level of consciousness. In GCS score not exceeding 4 and 39 of
addition, a strategy of delaying extubation 49 patients (80%) with GCS score not
was not associated with an increase in exceeding 8 were successfully extubated.
extubation success. The lowest mortality More recently, Anderson and coauthors
rates were observed among patients who examined factors associated with
were promptly and successfully extubated, extubation failure in 339 individuals in
although this likely stems from an inherent one neurocritical care unit (23). In their
selection bias and identifies a group of multivariable model, a GCS score 7T–9T
less sick patients. at the time of enrollment was associated
Our observed rate of extubation failure with increased risk of extubation failure
at 72 hours was 21%, which may seem high compared with 11T (OR, 2.91; 95% CI,
compared with expected rates in general 1.19–7.12). This analysis was limited by
ICU patients, but this is similar to previously the fact that a trial of extubation was not
reported rates ranging from 15 to 36% attempted in individuals with a GCS
among brain-injured patients (7–12, 23). score less than 7T.

McCredie, Ferguson, Pinto, et al.: Airway Management Strategies for Brain-Injured Patients 89

45
40
35
30
Patients (n)
25
20
15
10
0
1 2
Figure 3. Days from meeting criteria to consider extubation to actual extubation or primary tracheostomy. *Patients who never met criteria to consider
extubation before their actual extubation attempt (early extubation) or primary tracheostomy. Note: If a patient met criteria for consideration of extubation
on day of actual extubation, days to extubation = 1.
We excluded patients who self-
extubated and those whose extubation was
in the context of the withdrawal of life-
sustaining treatment, as these patients are
not relevant to the clinician trying to decide
whether it is safe to electively extubate.
Despite forcing GCS score into our model,
we observed no association between GCS
score—either total, or motor component—
and extubation failure. However, given the
low number of selected patients electively
extubated with very low GCS score in our
study, the extubation successes we observed
in these patients do not guarantee the safety
of extubating all such patients.
Clinicians caring for a patient with
a neurologically stable brain injury who
has passed a spontaneous breathing trial
but who is still not awake must choose
either to extubate, to perform a primary
tracheostomy instead, or to wait to extubate
or perform tracheostomy later. The option
of primary tracheostomy has been
advocated for neurologically injured
90
3 4 5 6 7 8 9 10 11 13 14 15 17
Days to extubation or primary tracheostomy
patients (26). Proposed benefits include
decreased risks of aspiration pneumonia
and tracheal injury secondary to prolonged
intubation and reduced duration of
mechanical ventilation and ICU length of
stay (27, 28).
In our study, patients who received
primary tracheostomy had prolonged
lengths of stay in the ICU, and had
similar ICU and in-hospital mortality
compared with patients who failed
extubation. Tracheostomy patients tended
to have lower GCS scores and higher
APACHE II scores on admission to the
ICU; thus, these observed outcome
differences likely are confounded by more
severe initial injury. However, many
patients (43%) in the primary tracheostomy
group met prespecified criteria to consider
extubation criteria before tracheostomy.
These patients tended to have low but
stable GCS scores, and some of them
may have been candidates for elective
extubation.
AnnalsATS Volume 14 Number 1 | January 2017
ORIGINAL RESEARCH
Extubation
Primary tracheostomy
24 27 28 50 *
We observed that extubation was
delayed in 30% of cases and was predicted by
lower GCS scores, absence of spontaneous
cough, and new positive sputum cultures
at time of meeting criteria to consider
extubation. A strategy of delaying
extubation did not increase the chance
of extubation success. More cases of
pneumonia were observed in the delayed
extubation group, but this difference did not
reach statistical significance. These results
corroborate those of Coplin and colleagues,
who reported a delayed extubation rate
of 27%; in comparison with promptly
extubated patients, delayed patients in that
study had lower GCS scores at extubation,
acquired pneumonia more frequently, but
did not have increased rates of extubation
failure (12). It has been suggested that
extubation delay results in part from
clinicians’ desire to avoid failed extubation
in patients with decreased levels of
consciousness (6, 12). The results of this
study suggest that delaying extubation on
 ORIGINAL RESEARCH

Table 3. Factors associated with successful extubation at time of meeting extubation readiness criteria

Variable Univariable Analysis Multivariable Regression Model

n OR 95% CI P Value OR* 95% CI P Value

Age (10-yr increase) 152 0.70 0.56–0.88 0.003 0.75 0.59–0.97 0.03
Male sex 152 0.64 0.27–1.56 0.33
Primary neurological indication for ongoing 149 0.74 0.16–3.59 0.71
mechanical ventilation
Admission acuity
APACHE II on admission 151 0.98 0.92–1.05 0.60
Admission total GCS 152 0.91 0.80–1.04 0.19 0.92 0.79–1.08 0.30
Neurological indicators
ICP (1-mm Hg increase) 42 0.96 0.88–1.04 0.32
GCS motor 152 1.08 0.64–1.81 0.78
GCS total 152 1.15 0.91–1.44 0.24
Airway parameters
Gag reflex present 152 2.25 0.99–5.12 0.05
Cough present† 152 3.28 1.43–7.55 0.005 3.60 1.42–9.09 0.01
ETT cuff leak present 152 1.23 0.55– 2.75 0.61
New positive sputum cultures 152 0.53 0.18–1.52 0.24
Sputum/suction count 142 1.00 0.98–1.03 0.91
Ventilatory parameters
PEEP (1-cm H2O increase) 151 0.87 0.68–1.12 0.28
PaCO2 (1-mm Hg increase) 135 1.03 0.97–1.10 0.29
RSBI (10-breaths/min/L increase) 146 1.13 0.93–1.38 0.22
PaO2/FIO2 (10-unit increase) 136 1.04 0.99–1.08 0.07
Minute ventilation (1-L/min increase) 126 0.95 0.80–1.12 0.53
Systemic parameters
Temperature ( 8 C) 151 1.34 0.69–2.59 0.39
Fluid balance (1-L increase)‡ 142 0.75 0.58–0.96 0.02 0.75 0.57–0.98 0.03
Inotrope/vasopressor use 152 2.21 0.27–18.38 0.46
Hemoglobin (10-g/L increase) 149 0.77 0.62–0.95 0.02

Definition of abbreviations: APACHE II = Acute Physiology and Chronic Health Evaluation II; CI = confidence interval; ETT = endotracheal tube;
GCS = Glasgow Coma Scale; ICP = intracranial pressure; OR = odds ratio; PaO2/FIO2 = partial pressure of arterial oxygen/fraction of inspired oxygen;
PEEP = positive end expiratory pressure; RSBI = rapid shallow breathing index.
*Odds ratios were estimated using a multivariable logistic regression model with successful extubation as the outcome of interest.
†
Spontaneous or provoked cough with suctioning.
‡
Yesterday’s 24-hour fluid balance.

Table 4. Outcomes for successful versus failed extubation

Outcomes All Successful Extubation Failed Extubation* P Value

(n = 152) (n = 120) (n = 32)

Death, n (%)
ICU 5 (3) 1 (1) 4 (13) ,0.01
Hospital 12 (8) 7 (6)† 5 (16)‡ 0.13
Days of stay, median (IQR)
Days in ICU 8 (4–14) 7 (4–14) 14 (8–18) ,0.01
Days in hospital 25 (15–36) 24 (16–39) 25 (15–31) 0.68
Total days of MV,x median (IQR) 5 (3–8) 5 (3–8) 4 (3–8) 0.09
Days of MV from meeting readiness criteria for 1 (1–3) 1 (0–3) 2 (1–5) 0.20
extubation, median (IQR)jj
Pneumonia, n (%) 24 (16) 12 (10) 12 (38) ,0.01
Pneumonia sensitivity analysis,¶ n (%) 28 (18) 16 (13) 12 (38) ,0.01
Secondary tracheostomy, n (%) 12 (8) 4 (3) 8 (25) ,0.01

Definition of abbreviations: ICU = intensive care unit; IQR = interquartile ratio; MV = mechanical ventilation.
*Extubation failure within 72 hours.
†
Four patients missing.
‡
One patient missing.
x
Mechanical ventilation from time of intubation to extubation or tracheostomy (includes all living and dead patients).
jj
Early extubations before meeting criteria for consideration of extubation were assigned 0 days.
¶
Sensitivity analysis of maximum number of pneumonias diagnosed using all positive chest X-rays for infection identified by one or more physicians and
clinical parameters in keeping with infection.

McCredie, Ferguson, Pinto, et al.: Airway Management Strategies for Brain-Injured Patients 91
 ORIGINAL RESEARCH

A
80 Successful extubation

70 Failed extubation

Tracheostomy
60

50
Patients (n)

40

30

20

10

0
1 2 3 4 5 6
GCS motor

B
50
Successful extubation
45 Failed extubation

40 Tracheostomy

35

30
Patients (n)

25

20

15

10

0
2 3 4 5 6 7 8 9 10
GCS motor & eye

Figure 4. (A) Glasgow Coma Scale (GCS) motor score at time of elective extubation or primary tracheostomy. (B) Abbreviated GCS score at time of
elective extubation or primary tracheostomy.

the grounds of a low GCS score alone may
not be warranted, and may be harmful if a
potential association between extubation
delay and increased risk of pneumonia is
verified.
In our study, patients who failed
extubation had an increased risk of ICU
death, yet delaying extubation was
associated with an increased risk of
in-hospital death. Because of the small
number of outcome events we were not able
to perform multivariable analyses to adjust
for differences in baseline factors between
these patient populations. Independent
associations between extubation failure and
increased mortality have been noted
previously, although largely in the medical
ICU (6, 8, 12, 13, 17). Whether this
association is truly causal or just a strong
marker of otherwise unmeasured severity of
illness remains uncertain. The absolute
difference in mortality that we observed
between successful and unsuccessful
extubation (10–12%) is considerably lower
than that noted in many medical ICU
studies (13, 17). We speculate that
extubation failure in brain-injured patients,
which is driven by different causes, may
have less impact on subsequent outcomes
than in medical ICU patients.
This observational study precludes
causal inferences; however, the associations
we have observed reflect real world practice.
Treatment indication bias may have
influenced our results, and we were unable
to adjust for baseline differences between
groups, suggesting cautious interpretation
of our findings. The parameters used to
characterize groups, such as the GCS score
and measures of airway function, are subject
to interobserver variability and may be
insensitive indicators of complex clinical
situations. In addition, the GCS score is
imperfect and vulnerable to fluctuations
over time, but because no “gold standard”

92 AnnalsATS Volume 14 Number 1 | January 2017

 ORIGINAL RESEARCH

exists for the evaluation of consciousness,
the GCS score is currently the most widely
used measure in clinical studies (29). The
criteria used to determine consideration
of extubation, and the definitions of
extubation failure and extubation delay,
have not been conclusively validated,
although there is support for these
definitions in the literature (3, 10, 12).
Finally, it is possible that clinical practice
has evolved since the assembly of our
cohort. However, we conducted a systemic
review and found no new studies that
clearly identified an optimal airway
management strategy for this population
of patients (30).
Conclusions
In conclusion, our multicenter observational
study does not support the use of GCS scores
to help predict which brain-injured patients
will fail extubation within 72 hours. When
facing airway management decisions in a
brain-injured patient who has passed a
spontaneous breathing trial, clinicians
should consider the use of airway protection
assessments, such as the presence of cough,
as part of the decision-making process.
Baseline differences notwithstanding,
neither a strategy of extubation delay or
primary tracheostomy was found to
improve outcomes. A randomized
trial is needed to rigorously evaluate
the impact of these different airway
management strategies on important
patient outcomes. n
Author disclosures are available with the text
of this article at www.atsjournals.org.

References Group. Effect of spontaneous breathing trial duration on outcome of

1 Esteban A, Frutos-Vivar F, Muriel A, Ferguson ND, Peñuelas O, Abraira V,
Raymondos K, Rios F, Nin N, Apezteguı́a C, et al. Evolution of mortality
over time in patients receiving mechanical ventilation. Am J Respir Crit
Care Med 2013;188:220–230.
2 Nyquist P, Stevens RD, Mirski MA. Neurologic injury and mechanical
ventilation. Neurocrit Care 2008;9:400–408.
3 Frutos-Vivar F, Ferguson ND, Esteban A, Epstein SK, Arabi Y,
Apezteguı́a C, González M, Hill NS, Nava S, D’Empaire G, et al. Risk
factors for extubation failure in patients following a successful
spontaneous breathing trial. Chest 2006;130:1664–1671.
4 Pelosi P, Ferguson ND, Frutos-Vivar F, Anzueto A, Putensen C,
Raymondos K, Apezteguia C, Desmery P, Hurtado J, Abroug F, et al.;
Ventila Study Group. Management and outcome of mechanically
ventilated neurologic patients. Crit Care Med 2011;39:1482–1492.
5 MacIntyre NR, Cook DJ, Ely EW Jr, Epstein SK, Fink JB, Heffner JE, Hess
D, Hubmayer RD, Scheinhorn DJ; American College of Chest
Physicians; American Association for Respiratory Care; American
College of Critical Care Medicine. Evidence-based guidelines for
weaning and discontinuing ventilatory support: 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. Chest 2001;120(6, Suppl):375S–395S.
6 Epstein SK. Weaning from ventilatory support. Curr Opin Crit Care
2009;15:36–43.
7 Vidotto MC, Sogame LC, Calciolari CC, Nascimento OA, Jardim JR. The
prediction of extubation success of postoperative neurosurgical patients
using frequency–tidal volume ratios. Neurocrit Care 2008;9:83–89.
8 Vallverdú I, Calaf N, Subirana M, Net A, Benito S, Mancebo J. Clinical
characteristics, respiratory functional parameters, and outcome of a
two-hour T-piece trial in patients weaning from mechanical ventilation.
Am J Respir Crit Care Med 1998;158:1855–1862.
9 Navalesi P, Frigerio P, Moretti MP, Sommariva M, Vesconi S, Baiardi P,
Levati A. Rate of reintubation in mechanically ventilated neurosurgical
and neurologic patients: evaluation of a systematic approach to
weaning and extubation. Crit Care Med 2008;36:2986–2992.
10 Namen AM, Ely EW, Tatter SB, Case LD, Lucia MA, Smith A, Landry S,
Wilson JA, Glazier SS, Branch CL, et al. Predictors of successful
extubation in neurosurgical patients. Am J Respir Crit Care Med
2001;163:658–664.
11 Ko R, Ramos L, Chalela JA. Conventional weaning parameters do not
predict extubation failure in neurocritical care patients. Neurocrit
Care 2009;10:269–273.
12 Coplin WM, Pierson DJ, Cooley KD, Newell DW, Rubenfeld GD. Implications
of extubation delay in brain-injured patients meeting standard weaning
criteria. Am J Respir Crit Care Med 2000;161:1530–1536.
13 Rothaar RC, Epstein SK. Extubation failure: magnitude of the problem,
impact on outcomes, and prevention. Curr Opin Crit Care 2003;9:59–66.
14 Hsieh AH, Bishop MJ, Kubilis PS, Newell DW, Pierson DJ. Pneumonia
following closed head injury. Am Rev Respir Dis 1992;146:290–294.
15 Esteban A, Alı́a I, Tobin MJ, Gil A, Gordo F, Vallverdú I, Blanch L, Bonet
A, Vázquez A, de Pablo R, et al.; Spanish Lung Failure Collaborative
attempts to discontinue mechanical ventilation. Am J Respir Crit Care
Med 1999;159:512–518.
16 Esteban A, Alı́a I, Gordo F, Fernández R, Solsona JF, Vallverdú I,
Macı́as S, Allegue JM, Blanco J, Carriedo D, et al.; Spanish Lung
Failure Collaborative Group. Extubation outcome after spontaneous
breathing trials with T-tube or pressure support ventilation. Am J
Respir Crit Care Med 1997;156:459–465.
17 Epstein SK, Ciubotaru RL, Wong JB. Effect of failed extubation on the
outcome of mechanical ventilation. Chest 1997;112:186–192.
18 Ferguson ND, Adhikari NKJ, Scales D, Fowler RA, Chapman M, Baker AJ,
Cook DJ, Meade MO; Canadian Critical Care Trials Group. Neurological
ICU Extubation Strategy Utilisation and Reintubation Outcomes
(NEURO) pilot study [abstract]. Intensive Care Med 2006;32:S110–S118.
19 Ely EW, Baker AM, Dunagan DP, Burke HL, Smith AC, Kelly PT,
Johnson MM, Browder RW, Bowton DL, Haponik EF. Effect on
the duration of mechanical ventilation of identifying patients capable
of breathing spontaneously. N Engl J Med 1996;335:1864–1869.
20 Maldonado G, Greenland S. Simulation study of confounder-selection
strategies. Am J Epidemiol 1993;138:923–936.
21 Burnham KP, Anderson DR. Multimodel inference understanding AIC and
BIC in model selection. Sociol Methods Res 2004;33:261–304.
22 Vittinghoff E, McCulloch CE. Relaxing the rule of ten events per variable in
logistic and Cox regression. Am J Epidemiol 2007;165:710–718.
23 Anderson CD, Bartscher JF, Scripko PD, Biffi A, Chase D, Guanci M,
Greer DM. Neurologic examination and extubation outcome in the
neurocritical care unit. Neurocrit Care 2011;15:490–497.
24 Salam A, Tilluckdharry L, Amoateng-Adjepong Y, Manthous CA.
Neurologic status, cough, secretions and extubation outcomes.
Intensive Care Med 2004;30:1334–1339.
25 Esteban A, Anzueto A, Frutos F, Alı́a I, Brochard L, Stewart TE, Benito S,
Epstein SK, Apezteguı́a C, Nightingale P, et al.; Mechanical
Ventilation International Study Group. Characteristics and
outcomes in adult patients receiving mechanical ventilation: a
28-day international study. JAMA 2002;287:345–355.
26 Alali AS, Scales DC, Fowler RA, Mainprize TG, Ray JG, Kiss A, de
Mestral C, Nathens AB. Tracheostomy timing in traumatic brain
injury: a propensity-matched cohort study. J Trauma Acute Care
Surg 2014;76:70–76, discussion 76–78.
27 Mascia L, Corno E, Terragni PP, Stather D, Ferguson ND. Pro/con clinical
debate: tracheostomy is ideal for withdrawal of mechanical ventilation
in severe neurological impairment. Crit Care 2004;8:327–330.
28 Koh WY, Lew TW, Chin NM, Wong MF. Tracheostomy in a neuro-
intensive care setting: indications and timing. Anaesth Intensive Care
1997;25:365–368.
29 Teasdale G, Maas A, Lecky F, Manley G, Stocchetti N, Murray G. The
Glasgow Coma Scale at 40 years: standing the test of time. Lancet
Neurol 2014;13:844–854.
30 McCredie VA, Alali AS, Scales DC, Adhikari NK, Rubenfeld GD, Cuthbertson
BH, Nathens AB. Effect of early versus late tracheostomy or prolonged
intubation in critically ill patients with acute brain injury: a systematic
review and meta-analysis. Neurocrit Care [online ahead of print] 6 Sep
2016; DOI: 10.1007/s12028-016-0297-z.

McCredie, Ferguson, Pinto, et al.: Airway Management Strategies for Brain-Injured Patients 93