burns 39 (2013) 236–242

Available online at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/burns

Cough strength, secretions and extubation outcome in burn

patients who have passed a spontaneous breathing trial

Sarah T. Smailes a,b,*, Andrew J. McVicar b, Rebecca Martin a

a
St Andrew’s Centre for Plastic Surgery and Burns, Broomfield Hospital, Chelmsford, Essex CM1 7ET, United Kingdom
b
Anglia Ruskin University, Chelmsford, Essex, United Kingdom

article info abstract

Article history: The aim of this study was to develop a clinical prediction model to inform decisions about
Accepted 28 September 2012 the timing of extubation in burn patients who have passed a spontaneous breathing trial
(SBT). Rapid shallow breathing index, voluntary cough peak flow (CPF) and endotracheal
Keywords: secretions were measured after each patient had passed a SBT and just prior to extubation.
Cough strength We used multiple logistic regression analysis to identify variables that predict extubation
Secretions outcome. Seventeen patients failed their first trials of extubation (14%). CPF and endotra-
Extubation cheal secretions are strongly associated with extubation outcome ( p < 0.0001). Patients with
Predictor CPF  60 L/min are 9 times as likely to fail extubation as those with CPF > 60 L/min (risk
Spontaneous breathing trial ratio = 9.1). Patients with abundant endotracheal secretions are 8 times as likely to fail
extubation compared to those with no, mild and moderate endotracheal secretions (risk
ratio = 8). Our clinical prediction model combining CPF and endotracheal secretions has
strong predictive capacity for extubation outcome (area under receiver operating charac-
teristic curve = 0.96, 95% confidence interval 0.91–0.99) and therefore may be useful to
predict which patients will succeed or fail extubation after passing a SBT.
# 2012 Elsevier Ltd and ISBI. All rights reserved.

we demonstrated that the negative consequences associated

1. Introduction
Decisions regarding the timing extubation are important
aspects of the intensive care management of patients [1].
This is because increasing duration of mechanical ventilation
itself is associated with increased mortality [2]. Therefore it is
of paramount importance that patients are liberated from
mechanical ventilation at the earliest opportunity. However,
patients who are prematurely extubated and underwent
reintubation have poorer outcomes than those who succeed
such that reintubation has been shown to be associated with
increased mortality in medical and surgical patients [3–5]. The
reasons for this have been attributed to the development of
new complications after reintubation [5]. In a previous study,
with extubation failure in intensive care patients also apply to
burn patients because those who underwent reintubation had
a significantly prolonged duration of ventilation, intensive
care and hospital length of stay and extubation failure was
associated with increased mortality [6].
Liberation of patients from mechanical ventilation consists
of two processes firstly, removal of the ventilatory support so
that the patient breathes without assistance and secondly,
removal of the endotracheal tube or extubation. Assessment
of patients’ capacity for spontaneous breathing has been
extensively studied through evaluation of the spontaneous
breathing trial protocol and this has been shown to decrease
the duration of ventilation in medical and surgical intensive
care patients [7–9]. In earlier work, we also demonstrated that

* Corresponding author at: Burn Intensive Care Unit, St Andrew’s Centre for Plastic Surgery and Burns, Broomfield Hospital, Court Road,
Chelmsford, Essex CM1 7ET, United Kingdom. Tel.: +44 1245 516037; fax: +44 1245 516007.
E-mail address: sarah.smailes@meht.nhs.uk (S.T. Smailes).
0305-4179/$36.00 # 2012 Elsevier Ltd and ISBI. All rights reserved.
http://dx.doi.org/10.1016/j.burns.2012.09.028
 burns 39 (2013) 236–242
burn patients who passed a 30 minute spontaneous breathing
trial have a significantly shorter duration of ventilation than
their retrospective controls [10].
Over the last 20 years many physiological measures have
been proposed as predictors of weaning outcome. Commonly
used weaning indices, e.g. negative inspiratory force (NIF) and
rapid shallow breathing index (RSBI – respiratory rate: tidal
volume ratio) have been reported to have only modest
predictive capacity for extubation outcome in ICU patients
[11,12]. For successful extubation to occur the patient, having
passed a spontaneous breathing trial, must be able to
maintain a patent airway by demonstrating adequate cough
strength and gag reflexes, and secretion volumes must be
manageable. For this reason cough strength and quantity of
secretions are commonly incorporated into the decision
making process when assessing a patient’s readiness for
extubation and these factors have performed well as extuba-
tion predictors in studies with other patient populations who
have passed a spontaneous breathing trial [13–17]. Evidence
supporting voluntary cough strength as a good predictor of
extubation outcome is provided by Smina et al. and Salam
et al. who identified that medical and cardiac patients with
weaker cough strength as measured by cough peak flow
(CPF)  60 L/min were 5 times as likely to fail extubation
compared to patients with a CPF > 60 L/min [13,14]. A
subsequent study by Beuret et al. evaluated cough strength
as a predictor of extubation outcome in intensive care patients
and found that a patient’s inability to cough to order or a peak
cough expiratory flow optimal cut off value of 35 L/min
predicted extubation failure [16].
Studies evaluating endotracheal secretions as a predictor
of extubation outcome include Salam et al. who instigated a
protocol involving collecting endotracheal secretions in a
sputum trap and performing suction every hour. That study
identified that patients were 3 times as likely to fail extubation
if their secretions volumes were >2.5 ml/h [14]. A subsequent
study by Mokhlesi et al. found that moderate or copious
endotracheal secretions, measured by suction frequency
according to nurses discretion, predicted extubation failure
in 122 medical and surgical patients [15].
At present there are no studies that have focused on
predictors of extubation outcome in burn patients. In the
current study we evaluate a quantitative method of measuring
voluntary cough strength (cough peak flow – CPF), rapid
shallow breathing index (RSBI) and endotracheal secretions as
predictors of extubation outcome in burn patients. The
objective is to develop a clinical prediction model to be used
at the bedside to assist with decision-making regarding the
timing of extubations in burn patients who have passed a SBT.
2. Methods
This is a prospective observational study. The Local Research
Ethics Committee and Research Ethics Board of the hospital
approved the study and permission was obtained from each
patient at the time of recruitment. All adult patients
(>18 years) undergoing endotracheal intubation and
ventilation > 24 h in our burn intensive care unit between
July 2005 and December 2010 were eligible for the study.
237
Patients were studied when they had successfully completed a
30 min spontaneous breathing trial and when they were about
to be extubated. In our ICU ventilation weaning is guided by a
non-mandatory protocol that is performed by nurses, a
physiotherapy consultant and anaesthetic specialist registrars
who are supervised by consultant intensivists. All patients are
screened daily for their readiness for a spontaneous breathing
trial by the presence of all of the following: (1) resolution of the
disease process necessitating ventilation; (2) PEEP  5 cm H2O;
(3) pressure support  10 cm H2O for at least 30 min; (4) no
further need for vasoactive or intravenous sedative drugs
(non-escalating doses were permitted); (5) awake and respon-
sive – Richmond agitation sedation scale 0–1 and able to
complete 3 simple tasks (open/close eyes, open/close mouth,
poke tongue out of mouth); (6) temperature < 39 8C for the
preceding 12 h; (7) haemoglobin > 7 g/dL; (8) no significant
airway swelling, cuff leak evident; (9) anaesthetist agrees the
patient is in a stable condition and is ready to be weaned from
the ventilator.
Spontaneous breathing trial assessments are performed
with 40% oxygen using a T-Tube with venturi device, in the
upright position (458) for 30 min arterial blood gas analysis is
performed at the end of the SBT or if indicated before. The SBT
is terminated by the anaesthetist if any of the following are
observed: (1) RR > 35 for 5 min or longer; (2) SpO2 < 90%; (3)
HR > 140 bpm or an increase or decrease > 20% baseline; (4)
systolic blood pressure > 180 mmHg or <90 mmHg; (5) signs of
increased work of breathing-accessory muscle use, paradoxi-
cal breathing movements, intercostal retractions, nasal
flaring; (6) agitation diaphoresis or signs of anxiety despite
attempts by the caregivers to alleviate this; (7) PaO2  8 kPa on
FiO2  0.4. A failed SBT is followed by a period of rest for at
least 24 h on the ventilator.
When patients had passed a spontaneous breathing trial
and extubation was being considered, the physiological
indices under investigation were taken. The patients’ were
asked to take a deep breath in and to cough as strongly as
possible into a mini – Wright peak flow metre via a Filta–
Guard breathing filter (intersurgical) attached. Patients were
instructed to cough ‘‘maximally’’ and the best of three
attempts was recorded. In order to calculate the RSBI minute
ventilation was measured using a calibrated spirometer
(Wrights) with a Filta–Guard breathing filter (intersurgical)
attached. The scoring system for endotracheal secretions
was based on suction frequency for the preceding 6 h to a
spontaneous breathing trial (Table 1). The number of passes
becomes important if a patient is suctioned every 3 h, to
group the patient into either mild or moderate secretions.
Due to the potentially deleterious side effects of performing
suction, it was not performed according to a study protocol
but when it was deemed to be necessary by the caregivers
according to clinical signs e.g. guided by ventilation and gas
exchange parameters and audible secretions in the endotra-
cheal tube. A minimum suction frequency of 6 hourly applies
in our unit to assess for and ensure endotracheal tube
patency. This approach to performing suction reflects
commonly accepted practice in ICUs and therefore our
method of scoring endotracheal secretions is relevant to
clinical practice. All patients received active humidification
whilst being ventilated. The consultant intensivists making
238 burns 39 (2013) 236–242
the decisions regarding the timing of extubation were blinded
to the CPF and RSBI measures.
The decision to undertake extubation, for a patient who
passes a SBT, is made by the consultant intensivist. Patients
who remain extubated at 48 h are classified as a successful
extubation. Patients who undergo reintubation within 48 h are
classified as extubation failures. Patients were followed until
discharge from the ICU or death. Patients were excluded if
they received ventilation for less than 24 h, died without
undergoing extubation or were accidentally or self extubated.
Patients who underwent tracheostomy without prior extuba-
tion were excluded. Also all patients with non-burn causes for
their skin loss were excluded.
The following data were gathered for each patient: age, %
total burn surface area (TBSA), incidence of smoke inhalation
injury, extubation outcome, duration of ventilation (DMV –
days ventilation from intubation to first extubation and after
reintubation for extubation failure patients) and ICU length of
stay (ICU LOS). The data from the patient’s first extubation was
included in the analyses.
The main outcome of interest was extubation outcome.
2.1. Statistical analysis
All data were entered into a statistical package for analyses
(Medcalc). The appropriate measure of central tendency was
used for the demographic, physiological measures and
outcome variables for the cohort according to whether the
variables were normally distributed. The mean was used to
describe the central tendency for variables that were normally
distributed. The median was used for variables that were not
normally distributed. An independent Samples t test was used
to compare means and a Mann–Whitney rank sum test was
used to compare medians. For categorical data a chi square
test and Fisher’s exact test (if n  5) were used to compare
numbers of patients in the extubation failure and success
groups. Spearman’s rank correlation coefficient was calculat-
ed to determine the presence of statistical associations
between the RSBI, CPF and secretions measures and the
patients’ extubation outcome. Risk ratios for extubation
failure (with 95% confidence intervals) were calculated for
CPF and endotracheal secretions. To develop a clinical
predictive model for extubation outcome whilst adjusting
for the confounding effects of the other variables multiple
logistic regression analysis using a stepwise regression
algorithm was performed for the CPF, RSBI and endotracheal
secretions scores and demographic data. The model fit was
determined using Akaike’s information criterion which
combines the usual likelihood function with a penalty term
that ensures that only variables that have a substantial effect
Table 1 – Scoring system for endotracheal secretions.
Secretions Description Frequency of suction
score
1 No Every 4–6 h
2 Mild Every 3–4 h, 1 pass
3 Moderate Every 2–3 h, 2 or more passes
4 Abundant Every hour or more frequent
on the performance of the model are included. Additionally,
the model was tested on the first time failure patients’
subsequent extubations. Receiver operating characteristic
(ROC) curve analyses of the variables were performed to
identify a critical threshold with the greatest diagnostic
accuracy for extubation outcome. Statistical significance
was placed at 0.05.
3. Results
One hundred and twenty-five patients met eligibility criteria,
passed a spontaneous breathing trial and were extubated
during the study period. These patients contributed a total of
135 extubations. Seventeen (14%) of the 125 patients failed
their first extubation. Three first time failures succeeded their
second extubation. Fourteen first time failures underwent
tracheostomy to facilitate weaning and 13 of these success-
fully weaned from ventilation. Five patients whose first
extubation was successful were reintubated between 6 days
to 2 weeks afterwards for sepsis and all of these were
subsequently successfully extubated. Two patients whose
first extubations were successful underwent subsequent
surgical procedures and received mechanical ventilation for
>24 h afterwards, these patients extubated successfully.
Table 2 illustrates that there are no significant differences
in the age, TBSA burn size or incidence of smoke inhalation
injury between the patients who succeeded and failed
extubation. RSBI and endotracheal secretions measures were
obtained for all of the patients in the cohort. However, CPF
could be obtained in all but 1 patient who was unable to
produce a cough effort to command and so the CPF score was
excluded for this patient. RSBI and endotracheal secretions
values are significantly higher in the patients who failed
extubation whereas the CPF values are significantly lower in
the patients who failed extubation. The main reason for
reintubation was hypoxaemia (9), respiratory distress/clinical
signs of increased work of breathing (5), excessive secretions
(2) and need for airway protection/stridor (1).
In terms of outcomes, the patients who failed extubation
remained ventilated for a significantly prolonged period of
time when compared to the patients who succeeded.
Similarly, the patients who failed extubation stayed in the
intensive care unit significantly longer than the patients who
were successfully extubated. There were 6 deaths, 1 patient in
the extubation failure group and 5 patients in the extubation
success group. There is no significant difference in the
mortality between the 2 groups.
Table 3 illustrates that CPF is significantly positively
correlated with extubation success. Conversely, endotracheal
secretions and RSBI are significantly negatively correlated
with extubation success. These results establish statistical
associations between CPF, secretions and RSBI with extuba-
tion outcome, the strongest association is between endotra-
cheal secretions and extubation outcome and the weakest is
between RSBI and extubation outcome.
Table 4 shows the predictive characteristics of CPF and
endotracheal secretions for extubation failure in terms of risk
ratios. Patients who score a CPF  60 L/min are nine times as
likely to fail extubation as those who score a CPF > 60 L/min
 burns 39 (2013) 236–242 239

Table 2 – Patient characteristics and outcomes stratified according to extubation outcome.

Variable Extubation failure n = 17 Extubation success n = 108 P
Age 44 (27–70.4) 38 (34–43.6) 0.23
TBSA burn size (%) 30.4 (18.3–42.4) 35.1 (31.3–38.9) 0.38
Smoke inhalation injury (no. of patients) 8 42 0.7
Cough peak flow (l/min) 74.2 (52–96.5) 125.8 (118.7–133) <0.0001
Endotracheal secretions 3 (3–4) 3 (2–3) <0.0001
RSBI (breaths/min per l) 61 (38.2–86.3) 41.7 (36.5–46.1) 0.01
Days mechanical ventilation 20 (13.3 –28) 6 (4–7) <0.0001
ICU LOS/% burn 1.03 (0.8–2.3) 0.48 (0.4–0.7) 0.0004
Deaths (no. of patients) 1 5 0.6
Values are expressed as mean or median as appropriate with 95% confidence intervals; Figures expressed in bold denote statistical
significance. RSBI, rapid shallow breathing index.

success decrease by 0.005 whereas for every 1 unit increase in

Table 3 – Rank correlation analysis of the variables for CPF the odds of extubation success increase by 1.06.
extubation outcome. The ROC curve analysis for the clinical prediction model
Variable Spearman’s rank correlation P combining the 2 variables CPF and endotracheal secretions is
coefficient (Rho) (95% CI) illustrated in Table 6. The overall diagnostic accuracy for the
Cough peak flow 0.38 (0.22–0.52) <0.0001 model is strong as illustrated by the area under the ROC curve
(CPF) (L/min) (AUC = 0.96 – Table 5). The model estimates the probability of
Endotracheal 0.43 ( 0.56 to 0.28) <0.0001 extubation success and the optimum cut off or threshold to
secretions separate estimated successes and failures is placed at 52.1%.
RSBI 0.23 ( 0.39 to 0.04) 0.01 This threshold generates strong predictive characteristics as
95% CI, 95% confidence interval. Figures expressed in bold denote illustrated by its sensitivity (0.95), specificity (0.88) and
statistical significance. predictive values (0.84, 0.86). These values indicate that the
model has strong predictive capacity for extubation outcome
in burn patients who have passed a spontaneous breathing
Table 4 – Predictive characteristics of CPF and secretions trial. Additionally, there were no substantial differences in the
for extubation failure. predictive capacity of the clinical prediction model when
Variable Risk ratio (95% CI) P tested on repeat extubations of first time failures.
CPF  60 L/min 9.1 (4–20.6) <0.0001
Secretions score = 4 (abundant) 8 (4–16) <0.0001
4. Discussion
Figures expressed in bold denote statistical significance.

Deciding when to extubate a burn patient is a critical decision

(RR = 9.1). The extubation failure rate for patients who do not
cough to order or who have a CPF  60 L/min is 59% compared
to an extubation failure rate of 6.5% for patients who have a
CPF of >60 L/min. Patients who have abundant endotracheal
secretions are 8 times as likely to fail extubation as those who
have moderate, mild or no endotracheal secretions (RR = 8).
The extubation failure rate for patients with abundant
secretions is 75% compared to 9.5% for those with moderate,
mild or no secretions.
Table 5 shows the most predictive model for extubation
outcome which combines the variables CPF with endotracheal
secretions. Although the RSBI was entered in the multiple
logistic regression analysis, it was removed from the model as
its regression coefficient was not significantly different from
zero ( p = 0.3). The CPF/endotracheal secretions model has a
highly significant regression coefficient ( p < 0.0001). Within
this model each of the variables independently predicts
extubation outcome as denoted by their significant regression
coefficients. The ROC curve analysis for the model indicates
that it has good diagnostic accuracy for the prediction of
extubation outcome as illustrated by the area under the ROC
curve (AUC = 0.96) and its 95% confidence interval. Table 5 also
illustrates the odds ratios for the variables, for every 1 unit
increase in endotracheal secretions the odds of extubation
because the associated risks, complications and costs of
unnecessarily prolonging mechanical ventilation must be
balanced against those of premature extubation leading to
extubation failure. The SBT is a standard assessment of
patients’ readiness for removal of ventilatory support but it is
not a good indicator of their ability to maintain a clear airway,
a requirement for successful extubation. This study has
demonstrated that voluntary CPF and endotracheal secretions
are important and potent predictors of extubation outcome in
burn patients who have passed a spontaneous breathing trial
according to strict protocol.
We have demonstrated that CPF is strongly positively
associated with extubation success and independently pre-
dicts extubation outcome in the 2 variable predictive model
(CPF and endotracheal secretions). Therefore we support the
use of this simple, reproducible measure of voluntary cough
strength in the assessment of burn patients’ readiness for
extubation. This finding supports other similar findings where
CPF was found to be associated with extubation failure
[13,17,18]. The current study has demonstrated that patients
with a CPF of 60 L/min are nine times as likely to fail
extubation as those with a CPF > 60 L/min. Smina et al. and
Salam et al. also identified a CPF threshold of 60 L/min but
they identified that patients with a CPF  60 L/min were only
240 burns 39 (2013) 236–242

Table 5 – Multiple logistic regression analysis of all the variables for extubation outcome.
Model Regression Odds ratio Regression Area under ROC
coefficient for (95% CI) coefficient for curve (AUC)
independent model (P) for model (95% CI)
variable (P)
Cough peak flow 0.0001 1.06 (1.03–1.09) <0.0001 0.96 (0.91–0.99)
Endotracheal secretions 0.0004 0.005 (0.0002–0.08)
Figures expressed in bold denote statistical significance.

Table 6 – ROC curve analysis for the predictive model; CPF and endotracheal secretions.
Model Optimal threshold Sensitivity Specificity Positive predictive Negative predictive
for model (%) (%) (%) value (%) value (%)
CPF + endotracheal secretions 52.1 95.0 88.2 84.0 86.0

five times as likely to fail extubation. The difference in risk
ratio is likely to be attributed to different patient populations
(medical/cardiac versus burn) or to varying spontaneous
breathing trial protocols leading to inconsistent exposure to
spontaneous breathing and levels of patient fatigue. The SBT
practiced in both of those studies were of varying durations
(30–120 min) and levels of support (either T-Piece or pressure
support 7 cm H2O) compared to a strict protocol of 30 min
spontaneous breathing using a T-Piece being used in the
current study. A later study by Beuret et al. identified a much
lower CPF threshold for predicting extubation failure in 130
ICU patients. They found that CPF was the sole factor
significantly associated with extubation failure and patients
who did not cough to order or those whose CPF  35 L/min
were nearly seven times as likely to fail extubation. Again, the
different risk ratios and threshold for CPF between Beuret et al.
and the current study could be attributed to different patient
populations or to SBT protocols. That study utilised a T-Piece
but the duration of the SBT was not specified.
Endotracheal secretion is a potent and independent
predictor of extubation outcome in our burn patients. This
finding is in accordance with Salam et al. and Mokhlesi et al.
but the methods used for scoring endotracheal secretions in
those studies were different. Salam et al. employed a protocol
where sputum was collected in a trap 2–3 h before extubation,
staff were instructed to suction the patient every hour and the
sputum volume was calculated as an average per hour (taking
into account saline instillations). They identified that secre-
tions independently predicted extubation failure and
patients with a secretions volume of >2.5 mL/h were 3 times
as likely to fail extubation. We have demonstrated that
endotracheal secretions measured by suction frequency and
number of passes, administered according to the clinical
judgement of the caregiver, independently predicts extuba-
tion outcome. This is an important funding since there are
deleterious side effects of suction and therefore it is
recommended practice that it is performed only when
secretions are present and not routinely, when perhaps it
is not needed [19]. Similar to our protocol, Mokhlesi et al. [15]
scored endotracheal secretions as suction frequency which
was determined by nurses, and they identified that moderate
to copious secretions independently predicted reintubation
in 122 medical and surgical intensive care patients. Mokhlesi
defined moderate secretions as suction every 1–2 h and
abundant secretions as suctioning several times per hour.
Moderate secretions was not predictive of extubation failure
in our study but our definition of ‘‘moderate’’ secretions was
suction every 2–3 h, more comparable to ‘‘minimal’’ secre-
tions in Mokhlesi et al. Our findings do not support those of
Smina et al. [13] who did not find an association between
magnitude of endotracheal secretions and extubation out-
come in their cohort of 95 medical and cardiac patients. They
explain this finding as being due to the physicians not being
blinded to the endotracheal secretions measurements and a
lack of standardised methodology for collecting secretions.
Although we identified that the RSBI was significantly
higher in burn patients who failed extubation and that it is
positively associated with extubation failure, it is the weakest
predictor when compared to CPF and endotracheal secretions.
This finding supports literature and studies that suggest that
RSBI is not a good sole predictor of extubation outcome [11,12].
In accordance with our findings, Frutos-Vivar et al. had also
identified that the RSBI was associated with extubation
outcome, but they found other important risk factors for
extubation failure, i.e. increasing age, a positive fluid balance
and pneumonia as the cause of respiratory failure in 900
medical patients [20]. It is possible that the RSBI does not
independently predict extubation outcome in our study
because it does not index the patient’s capacity to maintain
a clear and unobstructed airway. Indeed, other studies have
also identified that the RSBI has drawbacks as an extubation
predictor because it has poor specificity [21].
We have developed a simple, bedside assessment tool in
the form of a predictive model to estimate the probability of
extubation success in burn patients, having passed a
spontaneous breathing trial, based on reproducible CPF and
endotracheal secretions measures. The ROC analysis for the
model identifies good diagnostic accuracy (AUC = 0.96; 95%
CI = 0.91–0.99), sensitivity, specificity and predictive values for
extubation outcome. It seems clear from our findings that a
patient who has abundant secretions and whose CPF  60 L/
min is at high risk of extubation failure and, conversely, a
patient who has minimal secretions and a CPF > 60 L/min has
a low risk of extubation failure. We envisage that our
predictive model will provide useful information for such
patients and those where decisions about the timing of
 burns 39 (2013) 236–242
extubation are less clear, thus providing more objective
guidance for decisions regarding the timing of extubation.
Mokhlesi et al. also developed a similar ‘‘clinical prediction
rule’’ based upon 3 factors; hypercapnia, moderate to copious
endotracheal secretions and low Glasgow coma scale to
inform decisions regarding the timing of extubation [15].
There are 3 weaknesses to this study, firstly there is no
prospective validation set to test performance of the predictive
model developed in this study but we have prospectively
validated the importance of 2 extubation predictors identified
in other studies and settings [13–18] to a new population-burn
patients. Indeed, the validation and importance of objective
measures of voluntary cough strength (CPF) and endotracheal
secretions in other patient groups was only just starting to be
published when this study commenced and so it is possible
that these factors were not incorporated into extubation
decisions with burn patients.
Secondly, although the researcher did not share the details
of the scoring system for endotracheal secretions, endotra-
cheal secretions scores, CPF or RSBI measures with the
intensivists making the decisions regarding the timing of
extubation, the suction frequency is readily available on the
patient’s electronic record. However, patients with abundant
secretions were still extubated and so this does not seem to
have affected the intensivist’s decision making.
Thirdly, is the development of an overoptimistic predictive
model based on the relatively small data set and small number
of extubation failures which has led to wider confidence
intervals than desired for some of the variable estimates.
Although we used Akaike’s Information Criterion for the
multiple logistic regression analysis to discourage over fitting
and we tested the model on the patients’ subsequent
extubations, if future patients have very different character-
istics to our cohort the model’s predictions may be less
accurate.
The above limitations notwithstanding, this study has
demonstrated that simple and reproducible methods of
measuring voluntary cough strength and endotracheal secre-
tions, CPF and suction frequency, are important predictors of
extubation outcome in burn patients who have passed a SBT.
Our method of scoring endotracheal secretions is easily
transferable to other intensive care units as the score does
not involve any additional measures or change in suction
practice. However, voluntary CPF seems to vary between
different populations of critically ill patients [15] and so our
findings are not generalisable to these, but they may be
generalisable to similar populations of adult burn patients with
TBSA burn size 4–95%. Our clinical prediction model combining
CPF and endotracheal secretions may be useful to inform
decisions about the timing of extubation in burn patients.
5. Conclusion
In conclusion, this study has demonstrated that voluntary CPF
and endotracheal secretions contribute to extubation out-
comes in burn patients. Our simple and reproducible methods
of measuring these variables may provide clinicians in other
burn centres with inexpensive assessment tools that deter-
mine extubation readiness.
241
Our predictive model that determines the probability of
extubation success may also provide useful information
regarding the timing of extubation in burn patients who have
passed a SBT.
Acknowledgements
The authors would like to thank Professor Peter Dziewulski for
his support and also for reading the manuscript.
references
[1] Epstein SK. Weaning from ventilatory support. Curr Opin
Crit Care 2009;1536–43.
[2] Esteban A, Anzueto A, Frutos F, Alia I, Brochard L, Stewart.
et al. Characteristics and outcomes in adult patients
receiving mechanical ventilation. A 28 day international
study. J Am Med Assoc 2002;287:345–55.
[3] Torres A, Gatell JM, Aznar E, El-Ebiary M, Puig de la
Bellacasa J, Gonzalez J, et al. Reintubation increases the
risk of nosocomial pneumonia in patients needing
mechanical ventilation. Am J Respir Crit Care Med
1995;152:137–41.
[4] Epstein SK, Ciubotaru RL, Wong JB. Effect of failed
extubation on the outcome of mechanical ventilation.
Chest 1997;112:186–92.
[5] Frutos-Vivar F, Esteban A, Apezteguria C, Gonzalez M, Arabi
Y, Restrepo M, et al. Outcome of reintubated patients after
scheduled extubation. J Crit Care 2011;26:502–9.
[6] Smailes ST, Martin RV, McVicar AJ. The incidence and
outcome of extubation failure in burn intensive care
patients. J Burn Care Res 2009;30:386–92.
[7] Esteban A, Frutos F, Tobin M, Alia I, Salsona JF, Vallverdu I,
et al. A comparison of four methods of weaning patients
from mechanical ventilation. N Engl J Med 1995;332:345–
50.
[8] Ely EW, Baker AM, Dunagan DP, Burke HL, Smith AC, Kelly
PT, et al. Effect on the duration of mechanical ventilation of
identifying patients capable of breathing spontaneously. N
Engl J Med 1996;335:1864–9.
[9] Esteban A, Alia I, Tobin M, Gil A, Gordo F, Vallverdu I, et al.
Effect of spontaneous breathing trial duration on outcome
of attempts to discontinue mechanical ventilation. Am J
Respir Crit Care Med 1999;159:512–8.
[10] Smailes ST, Martin RV, McVicar AJ. Evaluation of the
spontaneous breathing trial in burn intensive care patients.
Burns 2009;35:665–71.
[11] MacIntyre NR, Cook DJ, Ely Jr EW, Epstein SK, Fink JB,
Heffner JE, et al. Evidence-based guidelines for weaning
and discontinuing ventilatory support: a collective task
force facilitated by the American College of Chest
Physicians; American Association for Respiratory Care;
College of Critical Care Medicine. Chest 2001;120:375S–95S.
[12] El- Khoury MY, Panos RJ, Ying J, Almoosa KF. Value of the
PaO2: FiO2 and rapid shallow breathing index in predicting
successful extubation in hypoxaemic respiratory failure.
Heart Lung 2010;39:529–36.
[13] Smina M, Salam A, Khamiees M, Gada P, Amoateng-
Adjepong Y, Manthous CA. Cough peak flows and
extubation outcomes. Chest 2003;124:262–8.
[14] Salam A, Tilluckdharry L, Amoateng-Adjepong Y,
Manthous CA. Neurologic status, cough, secretions
and extubation outcomes. Intensive Care Med
2004;30:1334–9.
242 burns 39 (2013) 236–242
[15] Mokhlesi B, Tulaimat A, Glickman TJ, Wang Y, Evans AT,
Corbridge TC. Predicting extubation failure after successful
completion of a spontaneous breathing trial. Respiratory
Care 2007;52:1710–7.
[16] Beuret P, Roux C, Auclair A, Nourdine K, Kaaki M, Carton M-
J. Interest of an objective evaluation of cough during
weaning from mechanical ventilation. Intensive Care Med
2009;35:1090–3.
[17] Khamiees M, Raju P, DeGirolamo A, Amoateng-adjepong Y,
Manthous CA. Predictors of extubation outcome in patients
who have passed a trial of spontaneous breathing trial.
Chest 2001;120:1262–70.
[18] 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;161:1530–6.
[19] AARC Clinical Practice Guidelines. Endotracheal suctioning
of mechanically ventilated patients with artificial airways.
Respir Care 2010;55(6):758–64.
[20] Frutos-Vivar F, Ferguson ND, Esteban A, Epstein SK, Arabi
Y, Apezteguia C, et al. Risk factors for extubation failure in
patients following a successful spontaneous breathing
trial. Chest 2006;130:1664–71.
[21] Epstein SK. Etiology of extubation failure and the predictive
value of the rapid shallow breathing index. Am J Respir Crit
Care Med 1995;152:545–9.