Concise Clinical Review

The Decision to Extubate in the Intensive Care Unit

Arnaud W. Thille1, Jean-Christophe M. Richard2,3, and Laurent Brochard2,3,4
1
Medical ICU, University Hospital of Poitiers, Poitiers, France; 2Intensive Care Division, Anesthesiology, Pharmacology and Intensive Care
Department, University Hospital, Geneva, Switzerland; 3School of Medicine, University of Geneva, Geneva, Switzerland; and 4Research Unit 955,
team 13, INSERM, University of Paris-Est, Créteil, France

The day of extubation is a critical time during an intensive care unit (NIV) in the postextubation period has further limited the
(ICU) stay. Extubation is usually decided after a weaning readiness validity of this definition. A consensus conference on weaning
test involving spontaneous breathing on a T-piece or low levels of defined success as the absence of ventilatory support during the
ventilatory assist. Extubation failure occurs in 10 to 20% of patients first 48 hours after extubation (10). Reintubation, NIV initia-
and is associated with extremely poor outcomes, including high tion, or death within 48 hours after extubation were taken to
mortality rates of 25 to 50%. There is some evidence that extubation indicate extubation failure, and these criteria were recently used
failure can directly worsen patient outcomes independently of in a prospective study on weaning (11). Death may occur in
underlying illness severity. Understanding the pathophysiology of patients who are extubated with a prior do-not-reintubate deci-
weaning tests is essential given their central role in extubation
sion. NIV can be initiated to treat postextubation respiratory
decisions, yet few studies have investigated this point. Because
distress or prophylactically before the onset of respiratory dis-
extubation failure is relatively uncommon, randomized controlled
trials on weaning are underpowered to address this issue. Moreover,
tress. In the first situation, reintubation might have been re-
most studies evaluated patients at low risk for extubation failure, quired in the absence of NIV or shortly after the time of NIV
whose reintubation rates were about 10 to 15%, whereas several initiation, although there is no strong evidence that NIV pre-
studies identified high-risk patients with extubation failure rates vents reintubation in this setting. Nevertheless, because NIV
exceeding 25 or 30%. Strategies for identifying patients at high risk may delay reintubation, the time interval needed to assess extu-
for extubation failure are essential to improve the management bation failure when NIV is used should probably be longer than
of weaning and extubation. Two preventive measures may prove 48 hours and perhaps should be 72 hours or 1 week. The use of
beneficial, although their exact role needs confirmation: one is prophylactic NIV cannot be classified as failure of extubation.
noninvasive ventilation after extubation in high-risk or hypercapnic Also, some studies focused chiefly on the occurrence of respi-
patients, and the other is steroid administration several hours before ratory distress (12). Reintubation can merely indicate poor clinical
extubation. These measures might help to prevent postextubation judgment, whereas resuming mechanical ventilation is probably
respiratory distress in selected patient subgroups. a less subjective criterion than the occurrence of respiratory dis-
tress. A consensus regarding the definition of extubation failure is
Keywords: laryngeal injury; weaning; noninvasive ventilation; organ
needed to determine the acceptable reintubation rate and to un-
dysfunction; endotracheal tube
derstand the risks associated with reintubation.
The day of extubation is a critical time during the intensive care
unit (ICU) stay in all patients surviving an episode of mechanical INCIDENCE AND IMPACT OF EXTUBATION FAILURE
ventilation. Although extubation is generally uneventful after Even among patients who meet all weaning criteria and success-
anesthesia, it is followed by a new episode of respiratory failure fully perform a weaning readiness test, 10 to 20% experience
in a substantial number of ICU patients. Very different clinical extubation failure (1–7, 13, 14) (Table 1). Failure of planned
approaches have been used to manage extubation. Not all patients extubation occurs in 5 to 10% of all intubated ICU patients,
are equal regarding the risk of reintubation, and the pathophys- a relatively low rate that hinders research into this event. In-
iology of extubation failure is incompletely understood. Conse- deed, planned extubation occurs in only 50 to 60% of ICU
quently, our knowledge about the best approaches for preventing patients (4, 6, 13) because about 30% of patients die while
and managing extubation failure remains limited. intubated (4, 6), tracheostomy may be performed without a prior
extubation attempt, about 5 to 15% of extubations are un-
DEFINING EXTUBATION FAILURE planned events (accidental or self-extubation), and some
Extubation failure is usually defined as a need for reintubation patients at the end of life undergo terminal extubation. Never-
within hours or days after planned extubation. The time interval theless, failure of planned extubation is associated with pro-
used in the definition varies from 48 hours (1–3) to 72 hours (4– longed mechanical ventilation and extremely high mortality
7) or 1 week (8, 9). The increased use of noninvasive ventilation rates of 25 to 50% (1–4, 6, 14) (Table 1). A central question
for clinicians is whether extubation failure is simply a marker of
poor prognosis or contributes to a poor prognosis. Although the
(Received in original form August 23, 2012; accepted in final form April 15, 2013) high mortality rate after failed extubation may be ascribable to
Correspondence and requests for reprints should be addressed to Laurent greater illness severity at the time of extubation, there is some
Brochard, M.D., Soins Intensifs, Hôpital Cantonal Universitaire, Rue Gabrielle- evidence that extubation failure, reintubation, and/or prolonga-
Perret-Gentil, 4, 1205 Geneva, Switzerland. E-mail: [email protected] tion of mechanical ventilation adversely affect survival indepen-
CME will be available for this article at http://ajrccm.atsjournals.org or at http:// dently of the underlying illness severity (6, 14). In the largest
cme.atsjournals.org case-series study of planned extubation, after adjustment for
Am J Respir Crit Care Med Vol 187, Iss. 12, pp 1294–1302, Jun 15, 2013 known outcome variables, reintubation was independently as-
Copyright ª 2013 by the American Thoracic Society
Originally Published in Press as DOI: 10.1164/rccm.201208-1523CI on May 3, 2013 sociated with ICU mortality (14). In a prospective observational
Internet address: www.atsjournals.org study, Thille and colleagues determined the daily sequential organ
Concise Clinical Review 1295

TABLE 1. RATES OF PLANNED EXTUBATION FAILURE AND MORTALITY

Rate of Extubation ICU Mortality in Reintubated ICU Mortality in Nonreintubated
Study (Reference) Number of Extubations Failure [% (n)] Patients [% (n)] Patients (%)

Esteban et al., 1997 (1) 397 19 (74) 27 (20) 3

Esteban et al., 1999 (2) 453 13 (61) 33 (20) 5
Epstein et al., 1997 (4) 287 14 (40) 43 (17) 12
Vallverdu et al., 1998 (3) 148 15.5 (23) 35 (8) 5.6
Thille et al., 2011 (6) 168 15 (26) 50 (13) 5
Frutos-Vivar et al., 2011 (14) 1,152 16 (180) 28 (50) 7
Funk et al., 2009 (38) 257 10 (26) Not available Not available
Tonnelier et al., 2011 (39) 115 10 (12) Not available Not available
Sellares et al., 2011 (34) 181 20 (36) Not available Not available
Peñuelas et al., 2011 (40) 2,714 10 (278) 26 (72) 5

failure assessment (SOFA) scores during the postextubation were diagnosed with upper airway obstruction (30–33). A large
period (6). The SOFA score improved on the day of extubation multicenter study evaluating the preventive efficacy of steroids
compared with the preceding days irrespective of extubation on postextubation stridor found that reintubation was directly
outcome (Figure 1). The SOFA score worsened rapidly in the ascribable to upper airway obstruction in 38% of cases (21).
failure group but continued to improve in the group with suc- These contradictory results suggest observation bias in studies
cessful extubation. This result constitutes indirect evidence that looking specifically for upper airway obstruction or frequent
extubation failure and/or reintubation per se can diminish the failure to diagnose upper airway obstruction in other studies.
chances of survival. Reintubation was significantly associated with A good marker for severe upper airway obstruction is the ab-
ventilator-associated pneumonia in several studies (6, 15), and a rea- sence of air leakage when the endotracheal tube cuff is deflated.
sonable assumption is therefore that reintubation may lead to clin- The amount of leakage can be quantified using a cuff-leak test
ical deterioration in fragile patients. The time of reintubation may to measure the difference between the insufflated volume and
also influence the outcome. Epstein and colleagues reported the expired volume in assist-control volume mode after balloon
higher mortality rates after late compared with early reintubation deflation (18, 19). A low cuff-leak volume (,110 or 130 ml)
(5). In a multicenter trial of NIV to treat postextubation respira- around the endotracheal tube before extubation may indicate
tory failure, mortality was higher in the NIV group (16), and the a high risk of upper airway obstruction (18, 19). Although the
only finding that seemed capable of explaining this mortality dif- absence of air leakage is a good predictor of upper airway ob-
ference was that time to reintubation was about 2 hours in the struction, the presence of a detectable leak does not rule out
standard-treatment group versus more than 12 hours in the NIV upper airway edema (22). The cuff-leak test is extremely useful
group. Unlike reintubation for other reasons, reintubation for because methylprednisolone therapy at least 12 hours before
transient upper airway obstruction does not seem to be associated extubation might reduce the incidence of stridor (20, 21) and
with increased mortality (5). This finding suggests that extubation the rate of reintubation (21) due to upper airway obstruction.
failure, rather than reintubation per se, is the reason for the higher The risk/benefit ratio of steroids in patients with negative cuff-
mortality rate. leak test results seems to favor steroid administration. The main
drawback of routine steroid therapy may be that steroids seem
CAUSES OF EXTUBATION FAILURE to be effective only when given several hours before extubation
and not when used only 1 hour before extubation (21, 25, 33).
The reason for extubation failure often escapes identification.
Routine steroid administration several hours before all planned
Reintubation is usually performed because of an apparently
new episode of respiratory distress, which may be related to pri-
mary respiratory failure, congestive heart failure, aspiration, in-
effective cough with airway secretion build-up, or upper airway
obstruction. Other reasons for reintubation include the onset of
new sepsis, surgical complications, acute coronary syndrome, and
neurological impairment. This multiplicity of causative factors con-
tributes to explain the clinical difficulties raised by extubation and
the persistent uncertainties about the pathophysiology of extu-
bation failure. Given the many causes for extubation failure,
data centered only on respiratory physiology may fail to con-
stitute a reliable guide for extubation decisions.
Respiratory distress can occur without lung function impair-
ment (e.g., when upper airway obstruction is unmasked by endo-
Figure 1. Changes in acute disease severity indicated by the Sequential
tracheal tube removal). Upper airway obstruction is a direct Organ Failure Assessment score (SOFA) from the day before planned
consequence of endotracheal intubation and occurs in about 5 extubation to Day 3 after extubation. SOFA scores at extubation were
to 15% of patients (17–21), being more common in women not significantly different between patients who were successfully extu-
(17, 21) and when the height/tube-diameter ratio is low (21). bated (open bars) and patients who failed extubation (solid bars). The
Other predictors have been reported, such as reason for admis- patients in both groups had improved SOFA scores within 24 hours
sion (19, 21), duration of mechanical ventilation (17, 21), and before extubation. After extubation, the SOFA scores worsened sub-
traumatic or difficult intubation (19). In studies focusing on the stantially in the patients who failed extubation and improved in the
overall causes for extubation failure, upper airway obstruction successfully extubated patients. Figure adapted by permission from
was the reason for reintubation in 10 to 20% of cases (1, 5, 35). Thille and colleagues (6). zP , 0.05 compared with the day of extu-
In studies focusing specifically on the incidence of postextuba- bation in the failed-extubation group. #P , 0.05 compared with the
tion stridor, 20 to 80% of patients who required reintubation day of extubation in the successfully extubated group.
1296 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 187 2013

extubations might result in delaying extubation until evidence spontaneous breathing trial (SBT) have been shown to predict
of systemic steroid activity is obtained. In addition, some patients failure of a weaning readiness test (29), postextubation respiratory
scheduled for extubation are not extubated and would therefore distress (30), or extubation failure (31). In a recent multicenter
receive steroids unnecessarily. A reasonable approach may be study, diuretic therapy guided by BNP values shortened the dura-
a routine cuff-leak test before extubation followed by steroid ad- tion of weaning, suggesting that inducing a negative fluid balance
ministration when this test is negative. Laryngeal edema is not the may hasten extubation (32). Lung failure may also be the reason for
only injury responsible for postextubation upper airway obstruc- extubation failure. A recent innovative study found that loss of lung
tion and/or stridor. In a prospective study of 136 patients venti- aeration as assessed by lung ultrasound during the SBT, suggesting
lated for longer than 24 hours, routine examination of the larynx lung derecruitment, predicted postextubation distress better than
after extubation showed laryngeal injuries in three-quarters of the did BNP or echocardiography (33). A minimal oxygenation thresh-
patients, with an even higher proportion after prolonged intuba- old is among the key criteria used to select patients for extubation,
tion (23). Patients failing extubation and requiring reintubation and readiness testing is usually not performed in severely hypox-
more often had granulations and vocal cord dysmotility compared emic patients. This point may explain the low predictive value of
with the other patients. These findings suggest that laryngeal inju- preextubation blood gas values for the outcome of extubation. One
ries may often go unrecognized and may participate in postextu- study showed that PaO2/FIO2 below 200 mm Hg was associated with
bation respiratory distress by increasing the work of breathing an increased risk of extubation failure in neurosurgical patients
and/or promoting aspiration via glottis dysfunction. (24), but most studies found no differences in terms of oxygenation
between patients who succeeded and those who failed extubation
(4, 6, 7, 25, 27). Patients who fail extubation have higher values of
RISK FACTORS FOR EXTUBATION FAILURE the rapid shallow breathing index (f/VT) before extubation than
AND HIGH-RISK POPULATIONS those who succeed extubation (7, 24, 25, 28), although considerable
Factors reported to be associated with extubation failure are overlap exists between these two groups. Unlike hypoxemia, hyper-
listed in Table 2. In several studies, neurological disorders (3) capnia per se may predict weaning outcomes (25, 34). One study
or impaired neurological status (24, 25) were independently as- found that hypercapnia during the SBT, defined as PaCO2 > 44 mm
sociated with extubation failure. It has also been suggested, how- Hg, was independently associated with extubation failure (25). An-
ever, that some comatose neurosurgical patients with Glasgow other study including selected patients with a high prevalence of
Coma Scale scores < 8 can be successfully extubated without chronic respiratory disorders showed that PaCO2 > 54 mm Hg dur-
delay and without an increased risk of reintubation provided ing the SBT independently predicted prolonged weaning and mor-
the airway secretions are minimal. Thus, delaying extubation tality (34). Prophylactic NIV was particularly effective in patients
may be unwarranted when an impaired neurological status is with hypercapnia (35). Therefore, hypercapnia before extubation
the only reason for considering prolonging the intubation (26). may constitute not only a valuable warning signal for an increased
Cough strength (30–32) and amount of secretions (29, 31) seem risk of prolonged weaning but also a good criterion for starting
to be good predictors of extubation failure, especially in patients prophylactic NIV (34).
with impaired neurological status, but this is a field where further An international consensus panel on weaning suggested that
objective measurements and research are needed. ventilated patients be categorized into three groups according to
The usual severity scores measured at ICU admission are poor the difficulty of their weaning process (10): “simple weaning”
predictors of extubation failure (7, 24, 27, 28) even when measured refers to patients who succeed the first weaning test and are
at the time of extubation (6, 25). The primary reason for intuba- extubated without difficulty, “difficult weaning” refers to patients
tion may help to predict the extubation outcome, but the available who fail the first weaning test and require up to three tests or
results are conflicting (4, 7). By contrast, several studies showed 7 days to achieve successful weaning, and “prolonged weaning”
higher extubation failure rates in older patients (6, 7). Thille and refers to patients who require more than 7 days of weaning after
colleagues identified a subset of patients at high risk for extuba- the first test. According to earlier studies (3, 36, 37), approxi-
tion failure. These were medical patients older than 65 years with mately 70% of mechanically ventilated patients fall into the sim-
underlying chronic cardiac or respiratory diseases, and their rein- ple weaning group. Four recent studies evaluated the proportion
tubation rate was 34% compared with only 9% in the other of patients in each group using a strategy of daily screening (34,
patients (6). In another study, a positive fluid balance on the 38–40) (Figure 2). The failure rate of the first test in these studies
day before extubation was associated with an increased risk of was 40 to 50%. Prolonged weaning was independently associated
extubation failure (7). Along the same line, high baseline levels of with increased mortality (34, 38–40) and with a significantly
B-type natriuretic peptides (BNP) or a BNP increase during a higher risk of reintubation in one study (39) and with a trend

TABLE 2. POTENTIAL RISK FACTORS FOR EXTUBATION FAILURE

Number of Episodes of
Study (Reference) Extubation Failure Risk Factors for Extubation Failure

Thille et al., 2011 (6) 26 Age . 65 yr or underlying chronic cardiorespiratory disease

Epstein et al., 1997 (4) 40 Age, APACHE II at time of weaning, and acute respiratory failure of cardiac origin
Frutos-Vivar et al., 2006 (7) 121 Pneumonia as the reason for intubation, high rapid shallow breathing index (f/VT), and positive fluid balance
Vallverdu et al., 1998 (3) 23 Neurological patients
Namen et al., 2001 (24) 44 Rapid shallow breathing index (f/VT) . 105, PaO2/FIO2 , 200 mm Hg, Glasgow Coma Scale score , 8
Mokhlesi et al., 2007 (25) 16 Abundant endotracheal secretions, Glasgow Coma Scale score < 10, PCO2 > 44 mm Hg during spontaneous
breathing trial
Smina et al., 2003 (28) 13 Peak expiratory flow < 60 L/min and rapid shallow breathing index > 100
Khamiees et al., 2001 (27) 18 Moderate or abundant endotracheal secretions, cough absent or weak, hemoglobin < 10 g/dl
Chien et al., 2008 (31) 19 Increase in B-type natriuretic peptide during a spontaneous breathing trial
Teixeira et al., 2010 (82) 31 .4.5% reduction in central venous saturation 30 min after spontaneous breathing trial initiation
Concise Clinical Review 1297

Figure 2. The proportion of patients (top), rate

of reintubation (middle), and rate of in-ICU mor-
tality (bottom) according to weaning difficulties.
The results are adapted from the four studies
evaluating three groups of ventilated patients
defined based on difficulty and duration of
weaning (34, 38–40) according to the interna-
tional conference consensus on weaning (10):
“simple weaning” (white histograms) refers to
patients who are extubated without difficulty
after the first weaning test, “difficult weaning”
(gray histograms) refers to patients who fail the
first weaning test and require up to three spon-
taneous breathing tests or 7 days to achieve
successful weaning, and “prolonged weaning”
(black histograms) refers to patients who require
more than 7 days of weaning after the first
weaning test. Using this definition, prolonged
weaning was associated with significantly higher
mortality rates in all four studies compared with
simple and difficult weaning. However, in none
of these four studies did the reintubation rate
differ among the three groups.

toward a higher risk of reintubation in two studies (34, 40). Thus, acquired polyneuromyopathy (46, 49). A recent study showed that
the classification scheme is of only moderate usefulness for pre- diaphragmatic dysfunction assessed by ultrasonography was asso-
dicting extubation failure. ciated with longer weaning times and higher reintubation rates
It has been suggested that the patient’s confidence about being (50). Diaphragmatic dysfunction at the time of extubation may
able to breathe without the ventilator may be a good predictor of correlate clinically with hypoventilation and inefficient cough,
extubation success, with the opposite being much less accurate subsequently increasing the risk of weaning failure (51).
(11). In one study, patients considered by healthcare providers to
be at high risk for extubation failure often failed extubation, but
many patients who failed were not considered to be at risk (41). WEANING READINESS TESTS
Physiological Results
UNCERTAINTIES ABOUT EXTUBATION FAILURE The ideal weaning readiness test would exhibit perfect accuracy
Extubation failure may be caused by other factors than the in predicting the tolerance of unassisted spontaneous breathing
above-mentioned predictors, such as delirium, sleep deprivation, after extubation by mimicking the postextubation physiological
adrenal insufficiency, or ICU-acquired weakness. Delirium is conditions. Thus, all patients passing the ideal weaning test
frequent in the ICU and predicts mortality (42). Acute brain would be able to maintain adequate ventilation after extubation.
dysfunction may promote extubation failure through conscious- We will discuss the predictive accuracy of current weaning tests
ness alterations, agitation, sedation induced by medications based on physiological and clinical data. A standard test for extu-
given for agitation, aspiration, and refusal of treatments includ- bation readiness is the SBT performed using the T-piece by dis-
ing NIV. Sleep is an essential physiological activity allowing connecting the patient from the ventilator and providing additional
physical, psychological, and mental recovery. No studies have oxygen. Another weaning test is performed without disconnecting
evaluated the impact of sleep quality on weaning success, al- the patient from the ventilator by using a low level of pressure sup-
though in another context poor sleep quality was associated port (PS) with or without positive end-expiratory pressure (PEEP)
with an increased risk of NIV failure in hypercapnic patients while continuously monitoring the respiratory rate and tidal vol-
(43). In a randomized study, stress-dose hydrocortisone supple- ume on the ventilator display. Cabello and colleagues compared
mentation shortened the time to extubation and increased the the SBT on a T-piece and the low-PS test (7 cm H2O) with or
rate of successful extubation among patients with adrenal insuf- without PEEP in patients with heart failure and difficult weaning
ficiency (44). However, adrenal insufficiency is difficult to detect (52). Patient effort was lower during the low-PS test than during
with the usual tests, and the steroids may have been effective the T-piece test and decreased further when PEEP was added to
because of their stimulating effects. The promising results of PS (Figure 3). These findings are consistent with previous evi-
this study have not been confirmed. ICU-acquired paresis may dence that PS and PEEP can reduce patient effort by about 30 to
occur in about 25% of patients after prolonged mechanical ven- 40% (60). An important point in the study by Cabello and col-
tilation (45) and may affect peripheral and respiratory muscles leagues is that most patients succeeded the PS test, although all
(46). It has been shown that ICU-acquired polyneuromyopathy patients failed the T-piece test (52). PS was initially introduced to
causing peripheral muscle weakness is independently associated reduce the work imposed by the ventilator circuit/valve and en-
with prolonged mechanical ventilation duration, higher ICU and dotracheal tube (53), which since then has decreased consider-
hospital mortality rates (47), weaning difficulties (46, 48, 49), and ably in parallel with technological improvements (54). Moreover,
a high risk of extubation failure (49). However, few studies have the postextubation period is characterized by relatively high
assessed the potential contribution of the inspiratory/expiratory upper airway resistance, so that the work of breathing af-
muscles to weaning or extubation difficulties in patients with ICU- ter extubation is virtually unchanged (55) or increased (56). A
1298 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 187 2013

Clinical Results
The ideal weaning readiness test would predict the tolerance of
unassisted breathing. However, the sensitivity and specificity of
weaning tests for predicting successful extubation are difficult to
assess. The extubation failure rate (i.e., the rate of false-positive
test results for predicting a successful extubation:patients toler-
ating the test but needing reintubation) is about 15%, which
makes the specificity of the test for predicting successful extuba-
tion 85%. By contrast, the proportion of patients able to tolerate
extubation despite failing the weaning test (i.e., false-negative
test results, used to determine sensitivity) is difficult to evaluate
because, for obvious ethical reasons, patients who fail a weaning
test are usually not extubated. Several types of studies have pro-
vided indirect and imperfect estimates of the false-negative rate.
Figure 3. Reduction in patient effort expressed as a percentage (mea-
1) Among patients who self-extubate at a time when they do not
sured using the esophageal pressure time product [PTP] in cm H2O/s/
min) between spontaneous breathing on a T-piece as the reference
meet weaning criteria, only 40 to 60% require reintubation.
(left), low pressure support (low PS, 7 cm H2O) with or without positive However, no information on weaning test results just before
end-expiratory pressure (PEEP) (middle), and reconnection to the ven- self-extubation is available. 2) In a relatively small study, 68%
tilator in assist-control ventilation (ACV, right). Patient effort was signif- of the patients who failed a T-piece test passed a low-PS test
icantly lower with the low-PS trial that with the T-piece trial (*P , 0.05): and were then successfully extubated, suggesting a high false-
the decreases were 231 6 29% with no PEEP and 252 6 23% with negative rate with the T-piece test (67). 3) A recent trial studied
PEEP. Figure drawn from data in Reference 52. extubation outcomes in patients who failed a SBT (9). Among
them, 42% did not require rescue NIV to treat postextubation
respiratory failure and only 37% were reintubated, indicating
common misconception is that breathing through an endotracheal limited sensitivity of the weaning test. 4) In clinical practice, the
tube necessarily increases the work of breathing, as occurs in determination of T-piece test failure is partly subjective, with
healthy volunteers breathing through a resistive tube (57). Straus many clinicians being somewhat biased toward overestimating
and colleagues showed that the work of breathing was similar be- ventilator dependency (68).
fore and immediately after extubation in 14 successfully extubated In a large, multicenter, randomized controlled trial, Esteban
patients breathing on a T-piece (55). The high upper airway resis- and colleagues compared the T-piece and low-PS tests (1). In
tance measured after extubation was due mainly to the glottis, accordance with physiological data, the proportion of patients
which was the narrowest part of the airway, whereas supraglottic who failed the first weaning test was higher with the T-piece
resistance was within the normal range (55). The high glottic resis- than with low PS. The proportion of successfully extubated
tance may be due to transient upper airway edema or to laryngeal patients after 48 hours was not significantly different between
lesions, as recently reported in many patients in the postextubation the two groups. A smaller study showed that some patients were
period (23). Therefore, a T-piece test accurately replicates the work able to pass a low-PS test immediately after failing a T-piece
of breathing required from the patient after extubation and prob- test and were then extubated with no increase in the risk of
ably constitutes a reliable assessment of the ability of an intubated extubation failure (67). Thus, the T-piece test may slightly delay
patient to maintain sufficient ventilation without assistance. By con- the identification of weaning readiness, or the low-PS test may
trast, adding even low PS levels may lead to underestimation of the carry a higher risk of reintubation. Physiological data support
extubation failure risk in some patients (58). However, the resistance the latter explanation, as indicated by Tobin in a comment on
of the endotracheal tube may be increased after several days of
mechanical ventilation (59) and may exceed that of the upper air-
ways. This is a theoretical concern that may deserve further research.
An increase in capillary pulmonary pressure can also occur
during the transition from mechanical to spontaneous ventila-
tion and may vary according to the type of weaning readiness
test (Figure 4) (52). In 1988, Lemaire and colleagues reported
the development of pulmonary edema and subsequent respi-
ratory distress shortly after the beginning of spontaneous
breathing, leading to unsuccessful weaning (60). Switching
from mechanical to spontaneous ventilation can decrease left
ventricular performance and unmask latent left ventricular
dysfunction (60) by increasing preload and afterload (61). Car-
diac dysfunction is a frequent cause for weaning test failure
(52, 62) and should be diagnosed by all available means be-
cause it can respond to diuretics and/or vasodilators and some-
Figure 4. Increase in pulmonary wedge pressure expressed as a percent-
times to coronary angioplasty in case of cardiac ischemia (63). age (measured using the pulmonary artery occlusion pressure [PAOP] in
Echocardiography can differentiate systolic and diastolic left mm Hg) between assist-controlled ventilation (ACV) as the reference (left),
ventricular dysfunction and can detect elevation of the pulmo- low pressure-support (low-PS, 7 cm H2O) with or without positive end-
nary occlusion artery pressure during the weaning test (64). It expiratory pressure (PEEP) (middle), and a T-piece (right). PAOP was sig-
has been shown that high baseline BNP or a BNP increase at nificantly higher during the T-piece and low-PS without PEEP trials than
SBT completion can predict weaning failure of cardiac origin during the low-PS with PEEP trial (*P , 0.05): the increases were 16 6
(65, 66). BNP measurement may be helpful as a first-line test 21% during the low-PS with PEEP trial, 35 6 39% during the low-PS
before an echocardiographic assessment for systolic or dia- without PEEP trial, and 41% 6 31 during the T-piece trial compared with
stolic cardiac dysfunction. ACV. Figure drawn from data in Reference 52.
Concise Clinical Review 1299

the myth of “minimal ventilator settings” (58). Nevertheless, disease, heart failure, or neuromuscular disorders (4, 6). In
there is no clinical evidence of a higher reintubation risk with patients at high risk for extubation failure and/or clinically con-
the low-PS test, and extubation failure rates have been lower sidered borderline at weaning test completion, it may be rea-
than 20 or 15% in studies of this method. The number of patients sonable to evaluate whether the patient might better tolerate
requiring reintubation remained relatively low even in large, ran- spontaneous breathing after 24 hours due to a more negative
domized controlled trials, which may have been underpowered for fluid balance or a significant neurological improvement with
this endpoint. The largest studies of weaning readiness tests stronger cough or decreased airway secretions. Thus, in some
included about 500 extubation episodes but only about 30 reintu- cases, most notably in high-risk fragile patients, it may be
bated patients in each group. Randomized controlled trials eval- worth waiting another 24 hours before reassessing the patient
uated unselected patients and, therefore, a majority of patients at for extubation.
low risk for extubation failure. The individual risk of reintubation Therapeutic NIV used in patients with postextubation respi-
may become unacceptably high only in high-risk populations. In ratory distress must be distinguished from prophylactic NIV used
two randomized controlled trials, reintubation rates were similar to prevent respiratory distress. Prophylactic NIV is the routine
using a T-piece test or a low-PS test lasting 30 minutes or 2 hours use of NIV immediately after extubation in the absence of evi-
(69). Again, these studies focused chiefly on simple weaning, dence of respiratory failure. Studies suggest that prophylactic
whereas an observational study showed that many difficult-to- NIV may help to prevent postextubation acute respiratory fail-
wean patients failed the weaning test only between 30 and ure (35, 73, 74). Prophylactic NIV is beneficial only in patients
120 minutes (3). Clinical trials are probably underpowered to at high risk for reintubation. In a recent study including more
demonstrate the safety of a procedure in high-risk patients, than 400 unselected ICU patients extubated after a successful
and their results should be extrapolated with caution. 2-hour SBT, reintubation rates were similar in patients treated
A large international survey on mechanical ventilation found with prophylactic NIV or oxygen therapy (75). Again, the extu-
that reintubation was significantly associated with the use of bation failure rate in this randomized controlled trial was low
CPAP compared with T-piece or low-PS tests (14). This result, (13.2 and 14.9%) in an unselected population (75). By contrast,
coupled to physiological data, suggests that weaning tests might Nava and colleagues found that NIV in high-risk patients de-
be done without PEEP to better detect latent cardiac dysfunction creased the need for reintubation (73). They used various cri-
and/or lung failure. Indeed, a low PEEP level in itself provides teria to define high-risk patients, such as previous weaning test
ventilatory and cardiac support, as shown by the demonstrated failure, more than one comorbidity, PCO2 .45 mm Hg after
benefits of CPAP (70), and may result in underestimation of the extubation, chronic heart failure, weak cough, or upper airway
extubation failure risk in some patients. A first-line weaning test obstruction with stridor (73). A study by Ferrer and colleagues
before extubation can probably be performed on the ventilator (74) defined patients at high risk for extubation failure using the
using a low-PS test without PEEP, but in many high-risk patients criteria identified by Epstein and colleagues (4): age older than
a prolonged T-piece test is probably more reliable for making 65 years, high severity score, or heart failure as the reason for
extubation decisions. intubation. Early NIV avoided respiratory failure after extuba-
In summary, the results of the T-piece and low-PS tests may de- tion and decreased ICU mortality without significantly decreas-
pend on the skill of the clinician taking the decision and on the prev- ing the reintubation rate (74). A subgroup analysis suggested
alence of extubation failure in a given population. T-piece test results that NIV was chiefly beneficial in hypercapnic patients with
may be too conservative if the clinicians are very cautious and/or if chronic respiratory disorders (74). In a prospective randomized
the prevalence of extubation failure is low (e.g., in postoperative controlled trial in 106 patients performed by the same group,
patients). On the other hand, the low-PS test may underestimate NIV was effective in patients who had hypercapnia at SBT
the risk of extubation failure, especially if the clinician is overop- completion and reduced Day-90 mortality (35). In contrast with
timistic or if the prevalence of extubation failure is high (e.g., in prophylactic NIV, therapeutic NIV has no proven benefit in the
patients under prolonged mechanical ventilation or having ICU- overall population of patients with postextubation acute respi-
acquired polyneuromyopathy). Keeping PEEP during the test may ratory failure (76) and can even increase the risk of death by
increase the rate of extubation failure due to lung or heart failure. delaying reintubation (16). Therapeutic NIV may decrease the
risk of reintubation after major elective abdominal surgery (77)
or lung resection (78).
SELECTING THE OPTIMAL STRATEGY IN PATIENTS Finally, few studies have reported the use of NIV as a weaning
AT HIGH RISK FOR EXTUBATION FAILURE method to hasten extubation in difficult-to-wean patients with
An international consensus panel on weaning insisted on the chronic obstructive pulmonary disease who failed a weaning test
need to perform the first weaning test as soon as the patient (9, 79–81). A recent multicenter study compared conventional
meets the following criteria (10): resolution of the initial reason weaning versus extubation followed by NIV or standard oxygen
for intubation, cardiovascular stability with minimal or no need therapy in patients who failed a 2-hour T-piece test (9). NIV re-
for vasopressors, no continuous sedation, and adequate oxygen- duced the risk of postextubation respiratory failure, but the weaning
ation defined as PaO2/FIO2 > 150 mm Hg with PEEP up to 8 cm success and reintubation rates were similar regardless of the wean-
H2O. Early identification of patients who can breathe sponta- ing method. The overall time on mechanical ventilation taking NIV
neously results in better outcomes. Daily screening followed by into account was longer in the NIV group (9). Consequently, NIV
a weaning test and then by extubation if the test is successful cannot be recommended as a weaning method in clinical practice.
can shorten the intubation time without increasing the risk of One important research objective regarding the indication for
reintubation (71, 72). In most patients, a short test (30-min low- prophylactic NIV is to better identify the high-risk population of
PS or T-piece test) is likely to be sufficient. However, there is patients most likely to benefit from this intervention. The exist-
a subgroup of easy-to-wean patients whose observed reintuba- ing data suggest that routine prophylactic NIV might be indi-
tion rate is less than 10% and another subgroup of high-risk cated immediately after extubation in hypercapnic patients.
patients who require reintubation in 20 to 30% of cases. The Further studies should evaluate whether NIV can also benefit
more challenging 2-hour T-piece test might be particularly use- patients older than 65 years with underlying chronic cardiorespi-
ful for decreasing the false-negative rate in high-risk patients ratory disease, a population identified as being at high risk for
such as elderly patients with chronic obstructive pulmonary extubation failure (6).
1300 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 187 2013

CONCLUSIONS VENTILA Group. Evolution of mechanical ventilation in response

to clinical research. Am J Respir Crit Care Med 2008;177:170–177.
In the ICU, the decision to extubate a patient is an important 14. Frutos-Vivar F, Esteban A, Apezteguia C, González M, Arabi Y, Restrepo
one. Clinical trials have established that hastening the weaning MI, Gordo F, Santos C, Alhashemi JA, Pérez F, et al. Outcome of
process is the best way to minimize the duration of mechanical reintubated patients after scheduled extubation. J Crit Care 2011;26:
ventilation. Extubation failure, which occurs in up to 20% of cases, 502–509.
has often been considered the price to pay for this approach. Sev- 15. Torres A, Gatell JM, Aznar E, el-Ebiary M, Puig de la Bellacasa J,
eral lines of evidence, however, suggest that extubation failure and González J, Ferrer M, Rodriguez-Roisin R. Re-intubation increases
reintubation can worsen outcomes and, consequently, that the rules the risk of nosocomial pneumonia in patients needing mechanical
used to make extubation decisions need to be improved, especially ventilation. Am J Respir Crit Care Med 1995;152:137–141.
regarding the identification of patients at high risk for extubation 16. Esteban A, Frutos-Vivar F, Ferguson ND, Arabi Y, Apezteguía C,
González M, Epstein SK, Hill NS, Nava S, Soares MA, et al.
failure. Future research should focus on identifying as yet unrec-
Noninvasive positive-pressure ventilation for respiratory failure
ognized factors associated with extubation failure, which might in- after extubation. N Engl J Med 2004;350:2452–2460.
clude delirium, poor sleep quality, limb muscle and diaphragmatic 17. Darmon JY, Rauss A, Dreyfuss D, Bleichner G, Elkharrat D,
weakness, and systolic or diastolic left ventricular dysfunction. Pro- Schlemmer B, Tenaillon A, Brun-Buisson C, Huet Y. Evaluation of
phylactic NIV during the postextubation period may be helpful, risk factors for laryngeal edema after tracheal extubation in adults
but studies are needed to identify the patients most likely to benefit and its prevention by dexamethasone: a placebo-controlled, double-
from this intervention. Randomized controlled trials should have blind, multicenter study. Anesthesiology 1992;77:245–251.
sufficient power to evaluate reintubation rates, as opposed to wean- 18. Miller RL, Cole RP. Association between reduced cuff leak volume and
ing rates, to better detect differences among weaning procedures. postextubation stridor. Chest 1996;110:1035–1040.
19. Jaber S, Chanques G, Matecki S, Ramonatxo M, Vergne C, Souche B,
Author disclosures are available with the text of this article at www.atsjournals.org. Perrigault PF, Eledjam JJ. Post-extubation stridor in intensive care
unit patients: risk factors evaluation and importance of the cuff-leak
test. Intensive Care Med 2003;29:69–74.
References
20. Cheng KC, Hou CC, Huang HC, Lin SC, Zhang H. Intravenous injection
1. Esteban A, Alía I, Gordo F, Fernández R, Solsona JF, Vallverdú I, of methylprednisolone reduces the incidence of postextubation stridor
Macías S, Allegue JM, Blanco J, Carriedo D, et al.; The Spanish Lung in intensive care unit patients. Crit Care Med 2006;34:1345–1350.
Failure Collaborative Group. Extubation outcome after spontaneous 21. François B, Bellissant E, Gissot V, Desachy A, Normand S, Boulain T,
breathing trials with T-tube or pressure support ventilation. Am J Respir Brenet O, Preux PM, Vignon P; Association des Réanimateurs du
Crit Care Med 1997;156:459–465. Centre-Ouest (ARCO). 12-h pretreatment with methylprednisolone
2. Esteban A, Alía I, Tobin MJ, Gil A, Gordo F, Vallverdú I, Blanch L, versus placebo for prevention of postextubation laryngeal oedema:
Bonet A, Vázquez A, de Pablo R, et al.; Spanish Lung Failure a randomised double-blind trial. Lancet 2007;369:1083–1089.
Collaborative Group. Effect of spontaneous breathing trial duration 22. Ochoa ME, Marín MdelC, Frutos-Vivar F, Gordo F, Latour-Pérez J,
on outcome of attempts to discontinue mechanical ventilation. Am J Calvo E, Esteban A. Cuff-leak test for the diagnosis of upper air-
Respir Crit Care Med 1999;159:512–518. way obstruction in adults: a systematic review and meta-analysis. In-
3. Vallverdú I, Calaf N, Subirana M, Net A, Benito S, Mancebo J. Clinical tensive Care Med 2009;35:1171–1179.
characteristics, respiratory functional parameters, and outcome of 23. Tadié JM, Behm E, Lecuyer L, Benhmamed R, Hans S, Brasnu D, Diehl
a two-hour T-piece trial in patients weaning from mechanical ventilation. JL, Fagon JY, Guérot E. Post-intubation laryngeal injuries and
Am J Respir Crit Care Med 1998;158:1855–1862. extubation failure: a fiberoptic endoscopic study. Intensive Care Med
4. Epstein SK, Ciubotaru RL, Wong JB. Effect of failed extubation on the 2010;36:991–998.
outcome of mechanical ventilation. Chest 1997;112:186–192. 24. Namen AM, Ely EW, Tatter SB, Case LD, Lucia MA, Smith A, Landry
5. Epstein SK, Ciubotaru RL. Independent effects of etiology of failure and S, Wilson JA, Glazier SS, Branch CL, et al. Predictors of successful
time to reintubation on outcome for patients failing extubation. Am J extubation in neurosurgical patients. Am J Respir Crit Care Med 2001;
Respir Crit Care Med 1998;158:489–493. 163:658–664.
6. Thille AW, Harrois A, Schortgen F, Brun-Buisson C, Brochard L. 25. Mokhlesi B, Tulaimat A, Gluckman TJ, Wang Y, Evans AT, Corbridge
Outcomes of extubation failure in medical intensive care unit patients. TC. Predicting extubation failure after successful completion of a
Crit Care Med 2011;39:2612–2618. spontaneous breathing trial. Respir Care 2007;52:1710–1717.
7. Frutos-Vivar F, Ferguson ND, Esteban A, Epstein SK, Arabi Y, 26. Coplin WM, Pierson DJ, Cooley KD, Newell DW, Rubenfeld GD.
Apezteguía C, González M, Hill NS, Nava S, D’Empaire G, et al. Implications of extubation delay in brain-injured patients meeting stan-
Risk factors for extubation failure in patients following a successful dard weaning criteria. Am J Respir Crit Care Med 2000;161:1530–1536.
spontaneous breathing trial. Chest 2006;130:1664–1671. 27. Khamiees M, Raju P, DeGirolamo A, Amoateng-Adjepong Y, Manthous
8. Demling RH, Read T, Lind LJ, Flanagan HL. Incidence and morbidity CA. Predictors of extubation outcome in patients who have successfully
of extubation failure in surgical intensive care patients. Crit Care Med completed a spontaneous breathing trial. Chest 2001;120:1262–1270.
1988;16:573–577. 28. Smina M, Salam A, Khamiees M, Gada P, Amoateng-Adjepong Y,
9. Girault C, Bubenheim M, Abroug F, Diehl JL, Elatrous S, Beuret P, Manthous CA. Cough peak flows and extubation outcomes. Chest
Richecoeur J, L’Her E, Hilbert G, Capellier G, et al. Non-invasive 2003;124:262–268.
ventilation and weaning in chronic hypercapnic respiratory failure 29. Mekontso-Dessap A, de Prost N, Girou E, Braconnier F, Lemaire F,
patients: a randomized multicenter trial. Am J Respir Crit Care Med Brun-Buisson C, Brochard L. B-type natriuretic peptide and wean-
2011;184:672–679. ing from mechanical ventilation. Intensive Care Med 2006;32:1529–1536.
10. Boles JM, Bion J, Connors A, Herridge M, Marsh B, Melot C, Pearl R, 30. Ouanes-Besbes L, Dachraoui F, Ouanes I, Bouneb R, Jalloul F, Dlala M,
Silverman H, Stanchina M, Vieillard-Baron A, et al. Weaning from Najjar MF, Abroug F. NT-proBNP levels at spontaneous breathing
mechanical ventilation. Eur Respir J 2007;29:1033–1056. trial help in the prediction of post-extubation respiratory distress.
11. Perren A, Previsdomini M, Llamas M, Cerutti B, Györik S, Merlani G, Intensive Care Med 2012;38:788–795.
Jolliet P. Patients’ prediction of extubation success. Intensive Care 31. Chien JY, Lin MS, Huang YC, Chien YF, Yu CJ, Yang PC. Changes in
Med 2010;36:2045–2052. B-type natriuretic peptide improve weaning outcome predicted by
12. Hilbert G, Gruson D, Portel L, Gbikpi-Benissan G, Cardinaud JP. spontaneous breathing trial. Crit Care Med 2008;36:1421–1426.
Noninvasive pressure support ventilation in COPD patients with 32. Mekontso Dessap A, Roche-Campo F, Kouatchet A, Tomicic V, Beduneau
postextubation hypercapnic respiratory insufficiency. Eur Respir J G, Sonneville R, Cabello B, Jaber S, Azoulay E, Castanares-Zapatero D,
1998;11:1349–1353. et al. Natriuretic peptide-driven fluid management during ventilator
13. Esteban A, Ferguson ND, Meade MO, Frutos-Vivar F, Apezteguia C, weaning: a randomized controlled trial. Am J Respir Crit Care Med 2012;
Brochard L, Raymondos K, Nin N, Hurtado J, Tomicic V, et al.; 186:1256–1263.
Concise Clinical Review 1301

33. Soummer A, Perbet S, Brisson H, Arbelot C, Constantin JM, Lu Q, 51. Matamis D, Soilemezi E, Tsagourias M, Akoumianaki E, Dimassi S,
Rouby JJ; Lung Ultrasound Study Group. Ultrasound assessment of Boroli F, Richard J-C, Brochard L. Sonographic evaluation of the
lung aeration loss during a successful weaning trial predicts postextubation diaphragm in critically ill patients: technique and clinical applications.
distress. Crit Care Med 2012;40:2064–2072. Intensive Care Med 2013;39:801–810.
34. Sellares J, Ferrer M, Cano E, Loureiro H, Valencia M, Torres A. 52. Cabello B, Thille AW, Roche-Campo F, Brochard L, Gómez FJ,
Predictors of prolonged weaning and survival during ventilator weaning Mancebo J. Physiological comparison of three spontaneous breathing
in a respiratory ICU. Intensive Care Med 2011;37:775–784. trials in difficult-to-wean patients. Intensive Care Med 2010;36:1171–1179.
35. Ferrer M, Sellarés J, Valencia M, Carrillo A, Gonzalez G, Badia JR, 53. Brochard L, Rua F, Lorino H, Lemaire F, Harf A. Inspiratory pressure
Nicolas JM, Torres A. Non-invasive ventilation after extubation in support compensates for the additional work of breathing caused by
hypercapnic patients with chronic respiratory disorders: randomised the endotracheal tube. Anesthesiology 1991;75:739–745.
controlled trial. Lancet 2009;374:1082–1088. 54. Thille AW, Lyazidi A, Richard JC, Galia F, Brochard L. A bench study
36. Brochard L, Rauss A, Benito S, Conti G, Mancebo J, Rekik N, Gasparetto of intensive-care-unit ventilators: new versus old and turbine-based
A, Lemaire F. Comparison of three methods of gradual withdrawal from versus compressed gas-based ventilators. Intensive Care Med 2009;35:
ventilatory support during weaning from mechanical ventilation. Am J 1368–1376.
Respir Crit Care Med 1994;150:896–903. 55. Straus C, Louis B, Isabey D, Lemaire F, Harf A, Brochard L. Contribution
37. Esteban A, Frutos F, Tobin MJ, Alía I, Solsona JF, Valverdú I, Fernández of the endotracheal tube and the upper airway to breathing workload.
R, de la Cal MA, Benito S, Tomás R, et al.; Spanish Lung Failure Am J Respir Crit Care Med 1998;157:23–30.
Collaborative Group. A comparison of four methods of weaning 56. Ishaaya AM, Nathan SD, Belman MJ. Work of breathing after
patients from mechanical ventilation. N Engl J Med 1995;332: extubation. Chest 1995;107:204–209.
345–350. 57. Fiastro JF, Habib MP, Quan SF. Pressure support compensation for
38. Funk GC, Anders S, Breyer MK, Burghuber OC, Edelmann G, Heindl inspiratory work due to endotracheal tubes and demand continuous
W, Hinterholzer G, Kohansal R, Schuster R, Schwarzmaier-D’Assie positive airway pressure. Chest 1988;93:499–505.
A, et al. Incidence and outcome of weaning from mechanical ventilation 58. Tobin MJ. Extubation and the myth of “minimal ventilator settings.” Am
according to new categories. Eur Respir J 2010;35:88–94. J Respir Crit Care Med 2012;185:349–350.
39. Tonnelier A, Tonnelier JM, Nowak E, Gut-Gobert C, Prat G, Renault 59. Jaber S, Pigeot J, Fodil R, Maggiore S, Harf A, Isabey D, Brochard L,
A, Boles JM, L’Her E. Clinical relevance of classification according to Louis B. Long-term effects of different humidification systems on
weaning difficulty. Respir Care 2011;56:583–590. endotracheal tube patency: evaluation by the acoustic reflection
40. Peñuelas O, Frutos-Vivar F, Fernández C, Anzueto A, Epstein SK, method. Anesthesiology 2004;100:782–788.
Apezteguía C, González M, Nin N, Raymondos K, Tomicic V, et al.; 60. Lemaire F, Teboul JL, Cinotti L, Giotto G, Abrouk F, Steg G, Macquin-
Ventila Group. Characteristics and outcomes of ventilated patients Mavier I, Zapol WM. Acute left ventricular dysfunction during
according to time to liberation from mechanical ventilation. Am J unsuccessful weaning from mechanical ventilation. Anesthesiology
Respir Crit Care Med 2011;184:430–437. 1988;69:171–179.
41. Boissier F, Ben Ghezala H, Razazi K, Brun-Buisson C, Thille AW. 61. Buda AJ, Pinsky MR, Ingels NB Jr, Daughters GT II, Stinson EB,
Predictive factors for extubation failure in medical intensive care Alderman EL. Effect of intrathoracic pressure on left ventricular
unit patients [abstract]. Intensive Care Med 2012;38:S195. performance. N Engl J Med 1979;301:453–459.
42. Ely EW, Shintani A, Truman B, Speroff T, Gordon SM, Harrell FE Jr, 62. Caille V, Amiel JB, Charron C, Belliard G, Vieillard-Baron A, Vignon
Inouye SK, Bernard GR, Dittus RS. Delirium as a predictor of P. Echocardiography: a help in the weaning process. Crit Care 2010;
mortality in mechanically ventilated patients in the intensive care 14:R120.
unit. JAMA 2004;291:1753–1762. 63. Carrié C, Bui HN, Gerbaud E, Vargas F, Hilbert G. Myocardial ischaemia
43. Roche Campo F, Drouot X, Thille AW, Galia F, Cabello B, d’Ortho MP, and weaning failure: is angioplasty the heart of the problem? Intensive
Brochard L. Poor sleep quality is associated with late noninvasive Care Med 2011;37:1223–1224.
ventilation failure in patients with acute hypercapnic respiratory 64. Lamia B, Maizel J, Ochagavia A, Chemla D, Osman D, Richard C,
failure. Crit Care Med 2010;38:477–485. Teboul JL. Echocardiographic diagnosis of pulmonary artery occlusion
44. Huang CJ, Lin HC. Association between adrenal insufficiency pressure elevation during weaning from mechanical ventilation. Crit
and ventilator weaning. Am J Respir Crit Care Med 2006;173: Care Med 2009;37:1696–1701.
276–280. 65. Grasso S, Leone A, De Michele M, Anaclerio R, Cafarelli A, Ancona G,
45. De Jonghe B, Sharshar T, Lefaucheur JP, Authier FJ, Durand-Zaleski I, Stripoli T, Bruno F, Pugliese P, Dambrosio M, et al. Use of N-terminal
Boussarsar M, Cerf C, Renaud E, Mesrati F, Carlet J, et al.; Groupe pro-brain natriuretic peptide to detect acute cardiac dysfunction during
de Réflexion et d’Etude des Neuromyopathies en Réanimation. weaning failure in difficult-to-wean patients with chronic obstructive
Paresis acquired in the intensive care unit: a prospective multicenter pulmonary disease. Crit Care Med 2007;35:96–105.
study. JAMA 2002;288:2859–2867. 66. Zapata L, Vera P, Roglan A, Gich I, Ordonez-Llanos J, Betbesé AJ. B-
46. De Jonghe B, Bastuji-Garin S, Durand MC, Malissin I, Rodrigues P, type natriuretic peptides for prediction and diagnosis of weaning
Cerf C, Outin H, Sharshar T; Groupe de Réflexion et d’Etude des failure from cardiac origin. Intensive Care Med 2011;37:477–485.
Neuromyopathies en Réanimation. Respiratory weakness is associated 67. Ezingeard E, Diconne E, Guyomarc’h S, Venet C, Page D, Gery P,
with limb weakness and delayed weaning in critical illness. Crit Care Vermesch R, Bertrand M, Pingat J, Tardy B, et al. Weaning from
Med 2007;35:2007–2015. mechanical ventilation with pressure support in patients failing a T-tube
47. Sharshar T, Bastuji-Garin S, Stevens RD, Durand MC, Malissin I, trial of spontaneous breathing. Intensive Care Med 2006;32:165–169.
Rodriguez P, Cerf C, Outin H, De Jonghe B; Groupe de Réflexion 68. Stroetz RW, Hubmayr RD. Tidal volume maintenance during weaning
et d’Etude des Neuromyopathies En Réanimation. Presence and with pressure support. Am J Respir Crit Care Med 1995;152:1034–1040.
severity of intensive care unit-acquired paresis at time of awakening 69. Perren A, Domenighetti G, Mauri S, Genini F, Vizzardi N. Protocol-
are associated with increased intensive care unit and hospital mor- directed weaning from mechanical ventilation: clinical outcome in
tality. Crit Care Med 2009;37:3047–3053. patients randomized for a 30-min or 120-min trial with pressure
48. De Jonghe B, Bastuji-Garin S, Sharshar T, Outin H, Brochard L. Does support ventilation. Intensive Care Med 2002;28:1058–1063.
ICU-acquired paresis lengthen weaning from mechanical ventilation? 70. Lenique F, Habis M, Lofaso F, Dubois-Randé JL, Harf A, Brochard L.
Intensive Care Med 2004;30:1117–1121. Ventilatory and hemodynamic effects of continuous positive airway
49. Garnacho-Montero J, Amaya-Villar R, García-Garmendía JL, Madrazo- pressure in left heart failure. Am J Respir Crit Care Med 1997;155:500–505.
Osuna J, Ortiz-Leyba C. Effect of critical illness polyneuropathy on 71. Ely EW, Baker AM, Dunagan DP, Burke HL, Smith AC, Kelly PT,
the withdrawal from mechanical ventilation and the length of stay in Johnson MM, Browder RW, Bowton DL, Haponik EF. Effect on
septic patients. Crit Care Med 2005;33:349–354. the duration of mechanical ventilation of identifying patients capable
50. Kim WY, Suh HJ, Hong SB, Koh Y, Lim CM. Diaphragm dysfunction of breathing spontaneously. N Engl J Med 1996;335:1864–1869.
assessed by ultrasonography: influence on weaning from mechanical 72. Girard TD, Kress JP, Fuchs BD, Thomason JW, Schweickert WD, Pun
ventilation. Crit Care Med 2011;39:2627–2630. BT, Taichman DB, Dunn JG, Pohlman AS, Kinniry PA, et al. Efficacy
1302 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 187 2013

and safety of a paired sedation and ventilator weaning protocol for 78. Auriant I, Jallot A, Hervé P, Cerrina J, Le Roy Ladurie F, Fournier JL,
mechanically ventilated patients in intensive care (Awakening and Lescot B, Parquin F. Noninvasive ventilation reduces mortality in acute
Breathing Controlled trial): a randomised controlled trial. Lancet 2008; respiratory failure following lung resection. Am J Respir Crit Care Med
371:126–134. 2001;164:1231–1235.
73. Nava S, Gregoretti C, Fanfulla F, Squadrone E, Grassi M, Carlucci A, 79. Nava S, Ambrosino N, Clini E, Prato M, Orlando G, Vitacca M, Brigada
Beltrame F, Navalesi P. Noninvasive ventilation to prevent respiratory P, Fracchia C, Rubini F. Noninvasive mechanical ventilation in the
failure after extubation in high-risk patients. Crit Care Med 2005;33: weaning of patients with respiratory failure due to chronic obstructive
2465–2470. pulmonary disease: a randomized, controlled trial. Ann Intern Med
74. Ferrer M, Valencia M, Nicolas JM, Bernadich O, Badia JR, Torres A. 1998;128:721–728.
Early noninvasive ventilation averts extubation failure in patients at 80. Girault C, Daudenthun I, Chevron V, Tamion F, Leroy J, Bonmarchand
risk: a randomized trial. Am J Respir Crit Care Med 2006;173:164–170. G. Noninvasive ventilation as a systematic extubation and weaning
75. Su CL, Chiang LL, Yang SH, Lin HI, Cheng KC, Huang YC, Wu CP. technique in acute-on-chronic respiratory failure: a prospective,
Preventive use of noninvasive ventilation after extubation: a prospective, randomized controlled study. Am J Respir Crit Care Med 1999;160:
multicenter randomized controlled trial. Respir Care 2012;57:204–210. 86–92.
76. Keenan SP, Powers C, McCormack DG, Block G. Noninvasive positive- 81. Ferrer M, Esquinas A, Arancibia F, Bauer TT, Gonzalez G, Carrillo A,
pressure ventilation for postextubation respiratory distress: a ran- Rodriguez-Roisin R, Torres A. Noninvasive ventilation during persistent
domized controlled trial. JAMA 2002;287:3238–3244. weaning failure: a randomized controlled trial. Am J Respir Crit Care
77. Squadrone V, Coha M, Cerutti E, Schellino MM, Biolino P, Occella P, Med 2003;168:70–76.
Belloni G, Vilianis G, Fiore G, Cavallo F, et al.; Piedmont Intensive 82. Teixeira C, da Silva NB, Savi A, Vieira SR, Nasi LA, Friedman G,
Care Units Network (PICUN). Continuous positive airway pressure Oliveira RP, Cremonese RV, Tonietto TF, Bressel MA, et al. Central
for treatment of postoperative hypoxemia: a randomized controlled venous saturation is a predictor of reintubation in difficult-to-wean
trial. JAMA 2005;293:589–595. patients. Crit Care Med 2010;38:491–496.