Eur Respir J 2006; 28: 588595

DOI: 10.1183/09031936.06.00150705
CopyrightERS Journals Ltd 2006

Noninvasive ventilation for prevention of

post-extubation respiratory failure in obese
patients
A.A. El Solh*, A. Aquilina*, L. Pineda*, V. Dhanvantri*, B. Grant*,# and P. Bouquin*

ABSTRACT: Current recommendations for management of obese patients post-extubation are

based on clinical experience and expert opinions. It was hypothesised that the application of
noninvasive ventilation (NIV) during the first 48 h after extubation in severely obese patients
would reduce post-extubation failure and avert the need for reintubation.
Following protocol-driven weaning trials, 62 consecutive severely obese patients (body mass
index o35 kg?m-2) were assigned to NIV via nasal mask immediately post-extubation and
compared with 62 historically matched controls who were treated with conventional therapy. The
primary end-point was the incidence of respiratory failure in the first 48 h post-extubation.
Compared with conventional therapy, the institution of NIV resulted in 16% (95% confidence
interval 2.929.3%) absolute risk reduction in the rate of respiratory failure. There was a significant
difference in the intensive care unit and lengths of hospital stay between the two groups.
Subgroup analysis of hypercapnic patients showed reduced hospital mortality in the NIV group
compared with the control group.
In conclusion, noninvasive ventilation may be effective in averting respiratory failure in severely
obese patients when applied during the first 48 h post-extubation. In selected patients with
chronic hypercarbia, early application of noninvasive ventilation may confer a survival benefit.
KEYWORDS: Noninvasive ventilation, obesity, obstructive sleep apnoea, reintubation
umerous studies have highlighted the
complexity of respiratory management of
critically ill obese patients during the
period following liberation from mechanical
ventilation [1, 2]. The development of respiratory
instability, episodic desaturation during supine
position, and the respiratory depressant effects of
sedatives and opioid analgesia predispose these
patients to prolonged periods of apnoeas,
hypoxia and severe hypercapnia culminating in
respiratory failure. Yet the incidence of these
serious complications is not well documented.
The rate of reintubation post-extubation in
severely obese patients has been reported at 8
14% among patients undergoing mechanical
ventilation for .48 h [3, 4], but these estimates
are likely to underestimate the true incidence of
respiratory failure in this population.

recent investigation has documented a beneficial

effect of administering NIV immediately postextubation in selected patients at increased risk
for extubation failure [8]. FERRER et al. [8] have
shown not only a reduction in the rate of
respiratory failure but also a decrease in the rate
of intensive care unit (ICU) mortality. As severely
obese patients are considered at high risk of
developing respiratory complications [1, 9], the
current authors instituted a protocol requiring
early application of NIV post-extubation. It was
hypothesised that the preventive use of NIV
would reduce the incidence of respiratory failure
and the rate of reintubation. To test this hypothesis, an observational study was conducted
comparing NIV therapy to historical controls of
conventional medical therapy during the first
48 h post-extubation.

Noninvasive ventilation (NIV) has been considered a promising therapy by an International

Consensus Conference to avert respiratory failure
after weaning [5]. However, randomised controlled studies in nonobese patients have shown
mixed results. While earlier studies failed to
show any benefit from rescue NIV [6, 7], a more

METHODS
Patient population
In a prospective study beginning in September
2004, involving 52 beds in three ICUs, all severely
obese patients (body mass index (BMI) o35
kg?m-2) requiring endotracheal mechanical ventilation for at least 48 h at a university-affiliated

588

VOLUME 28 NUMBER 3

AFFILIATIONS
*Western New York Respiratory
Research Center, Depts of Medicine
and of Biostatistics, Division of
Pulmonary, Critical Care, and Sleep
Medicine, University at Buffalo, and
#
Veterans Affairs Medical Center,
Buffalo, NY, USA.
CORRESPONDENCE
A.A. El Solh
Division of Pulmonary
Critical Care
and Sleep Medicine
Erie County Medical Center
462 Grider Street
Buffalo
NY 14215
USA
Fax: 1 7168986139
E-mail: [email protected]
Received:
December 21 2005
Accepted after revision:
May 11 2006
SUPPORT STATEMENT
None of the authors have a financial
relationship with a commercial entity
that has interest in the subject of this
manuscript. This study was
supported by a grant from Research
for Health in Erie County, NY, USA.

European Respiratory Journal

Print ISSN 0903-1936
Online ISSN 1399-3003

EUROPEAN RESPIRATORY JOURNAL

A.A. EL SOLH ET AL.

tertiary care centre were eligible for enrolment. Exclusion

criteria included recent facial, cranial or cervical spine surgery,
neuromuscular disorders, uncontrolled delirium, active upper
gastrointestinal bleeding and unplanned extubation. Patients
who had a do-not-resuscitate order were also excluded. For
those patients with more than one ICU admission, only the
first event was included in the analysis to ensure independence
of observation. A historic control group of patients who met
the same inclusion and exclusion criteria was chosen from
severely obese patients admitted to the ICU in the preceding 29
months. The NIV and the historic group were matched for age
(5 yrs), BMI (2 kg?m-2), and Acute Physiology and Chronic
Health Evaluation (APACHE II) score (2) [10]. For each
patient in the NIV group, matching was performed with a
patient from the historic group according to the following
hierarchy: baseline APACHE II score, BMI and age. When
matching was unsuccessful at end of this process, it was
started again with extension of the age for matching (5 yrs)
and then extension of the range of BMI (2 kg?m-2). If there
was more than one match, the patient in the historic group was
selected by the best match according to the matching
hierarchy. Using this algorithm, matches for all the NIV
patients were identified.
Study approval was obtained from the local Institutional
Review Board prior to initiation of the study. Written informed
consent was waived owing to the observational nature of the
study.
Weaning protocol
Starting in February 2002, weaning trials were conducted
according to a previously established protocol [11]. In brief, a
weaning trial was initiated once the patient was fully awake,
responsive, and had fulfilled the following criteria: arterial
oxygen tension (Pa,O2) .9.31 kPa, with inspiratory oxygen
fraction (FI,O2) f0.35 and positive end-expiratory pressure
f5 cmH2O; rectal temperature .36.6uC; a stable cardiovascular system (i.e. systolic blood pressure .100 mmHg, heart
rate ,120 beats?min-1); and absence of metabolic acidosis and
withdrawal of all vasopressors for .12 h. The screen for
readiness to wean was performed between 07:00 h and 09:00 h
based on a previously published weaning protocol shown to
reduce duration of ventilation [11]. During screening, the ratio
of respiratory frequency to tidal volume (fR/VT) was assessed
[12]. If fR/VT was ,105 min-1?L-1, a spontaneous breathing
trial was attempted for 30 min on continuous positive airway
pressure (CPAP), with 5 cmH2O pressure support if the
endotracheal tube was f8 mm. Extubation was then ordered
by the attending physician, otherwise the patient was placed
back on mechanical ventilation if within 30 min of initiating
spontaneous breathing trial (SBT) any of the following criteria
were met: 1) change in heart rate of .20 beats?min-1 from
baseline prior to initiating SBT persisting for o5 min; 2)
systolic blood pressure of ,90 mmHg or change of
.30 mmHg after initiating SBT persisting for o5 min; 3)
arterial oxygen saturation (Sa,O2) ,90%, Pa,O2 ,7.98 kPa, or pH
f7.35; and 4) respiratory muscle fatigue or increased work of
breathing suggested by the use of accessory respiratory
muscles, paradoxical motion of the abdomen, or retraction of
the intercostal spaces.
EUROPEAN RESPIRATORY JOURNAL

PREVENTION OF POST-EXTUBATION FAILURE

Post-extubation, patients were placed on conventional medical

therapy, including oxygen therapy to maintain Sa,O2 at .90%,
bronchodilators, chest physiotherapy and other treatments
directed by the primary-care physicians.
Noninvasive ventilation
In July 2004, the current authors adopted a protocol in ICUs to
administer NIV to severely obese patients for the first 48 h
immediately post-extubation. NIV was delivered using the
bilevel positive airway pressure mode (BiPAP S/T-D
Ventilatory Support System; Respironics Inc., Murrysville,
PA, USA) via a nasal mask. Nasogastric tubes were removed
prior to institution of NIV. The inspiratory and expiratory
pressures were initially set at 12 cmH2O and 4 cmH2O,
respectively, in a spontaneous mode [13]. The pressure settings
were increased gradually to the patients tolerance with the
aim of achieving a respiratory rate ,25 breaths?min-1 and Sa,O2
.90%. The fractional concentration of oxygen was also titrated
to maintain an Sa,O2 .90%. Subjects were advised to use NIV
continuously with periods of rest up to 60 min every 6 h to
receive meals, scheduled medications or nursing care. Regular
examination of the facial skin was made by the respiratory
therapist to prevent skin ulceration from the tight-fitting mask.
NIV was maintained for the first 48 h, then the decision to
maintain or discontinue NIV was left to the admitting
physician.
Medical care was provided by the same intensivist team,
critical care nurses and respiratory therapists during the study
period.
Criteria for respiratory failure and reintubation
Reintubation was performed immediately post-extubation
when any of the following major clinical events occurred:
respiratory or cardiac arrest; irregular respiratory rate associated with loss of consciousness or gasping for air; and severe
haemodynamic instability without response to fluids and
vasoactive drugs [14]. In addition, respiratory failure was
defined as the presence or persistence after 1 h post-extubation
of any of the following parameters requiring reintubation or
rescue NIV: 1) hypercapnia (arterial pH ,7.35 along with an
increase in carbon dioxide arterial tension (Pa,CO2) of .20%
from the time of extubation); 2) hypoxaemia defined as Sa,O2 by
pulse oximetry ,90% or Pa,O2 ,7.98 kPa with FI,O2 .0.5;
3) decreased consciousness or psychomotor agitation rendering the patient unable to tolerate NIV; 4) clinical signs
suggestive of respiratory muscle fatigue and/or increased
work of breathing, such as the use of respiratory accessory
muscles, intercostal indrawing, or paradoxical motion of the
abdomen; 5) severe delirium or agitation; and 6) inability to
clear secretions.
Data collection
Sociodemographic and clinical data were analysed. These
included age, sex, BMI, reason for mechanical ventilation,
APACHE II score on admission to the ICU and at the time of
liberation from mechanical ventilation, and duration of
mechanical ventilation before extubation. Respiratory rate,
heart rate, spontaneous tidal volume and the rapid shallow
breathing index were obtained prior to extubation. The causes
for respiratory failure were classified as described in the
VOLUME 28 NUMBER 3

589

PREVENTION OF POST-EXTUBATION FAILURE

A.A. EL SOLH ET AL.

preceding section. In the historic group, the need for rescue

NIV in the first 48 h post-extubation was considered a failure
of the standard medical therapy. Complications recorded postextubation included hospital-acquired pneumonia and nosocomial bloodstream infection irrespective of its origin. These
were defined according to the Centers for Disease Control
criteria [15]. At the time of discharge, the total lengths of stay
and hospital mortality rate were recorded.
Statistical analysis
The primary end-point was the incidence of respiratory failure
in the first 48 h post-extubation. After reviewing reintubation
rates in critical care units in this high-risk group over a 1 yr
period, 62 patients were required in each group in order to
detect a 20% absolute reduction in the risk of respiratory
failure with the application of prophylactic NIV relative to
25.8% in the conventionally treated group, with a power of
80% and a type I error of 5%. Secondary outcomes were lengths
of ICU and hospital stay and hospital mortality.
Results are expressed as meanSD. Continuous variables for
the two groups were compared with an unpaired t-test for
normally distributed data or the MannWhitney U-test
otherwise. Qualitative or categorical variables were compared
with the Chi-squared test or Fishers exact test. A difference
was considered statistically significant when the alpha probability was ,0.05 (all two-tailed).
RESULTS
Over the period from February 2002 to June 2004, 101 severely
obese patients who required endotracheal intubation for
o48 h were identified. Eighteen had tracheostomy placed
without extubation trial, one was transferred to another
facility, and two had a terminal weaning. Eighty eligible
patients were then considered for inclusion. Seven patients
were excluded for the following reasons: uncontrolled delirium (n53), unplanned extubation (n52), facial surgery (n51),
and cranial surgery (n51), leaving 73 patients in a pool to
access for matching to the NIV group. From July 2004 to
September 2005, 83 severely obese patients required intubation
o48 h. Fourteen patients had a tracheostomy and one had a
terminal weaning. Six patients met one of the exclusion
criteria: unplanned extubation (n53), do-not-resuscitate order
(n51), uncontrolled delirium (n51), and facial surgery (n51).
The remaining 62 patients of the NIV group were then matched to
the conventionally treated participants (fig. 1). The two groups
were similar in age, sex, BMI, underlying comorbidities and
APACHE II score (table 1). Similarly, the causes for endotracheal
mechanical ventilation were comparable in both groups.
Obstructive apnoea was documented in 22% of the NIV group
and 27% of the conventional medical therapy group. The
corresponding meanSD apnoeahypopnoea indices were
45.47.5 events?hr-1 (range 14112) and 40.86.3 events?hr-1
(range 694; p50.64). None of the patients in the historic group
had CPAP therapy in the first 48 h post-extubation.
Table 2 displays the physiological indices of participants prior
to extubation. In the NIV group, the meanSD inspiratory and
expiratory positive airway pressures were 17.23.5 cmH2O
(range 1226) and 7.21.6 cmH2O (range 512), respectively.
The meanSD use of NIV was 16.22.6 h?day-1 for the first
48 h post-extubation. Nine patients complained of oral dryness
590

VOLUME 28 NUMBER 3

due to an inability to keep a tight seal. Four patients responded

to placement of a chinstrap and five were switched to a full
face mask. Two patients had abrasion of the dorsum nasae but
none of the patients developed evidence of gastric distension,
aspiration or epistaxis. Twenty-three patients remained on the
bilevel positive pressure ventilation beyond the 48 h; of these,
17 showed evidence of hypercapnia during a trial of
spontaneous breathing. In comparison, the meanSD inspiratory and expiratory positive airway pressures for the conventionally treated group were 15.53.5 cmH2O (range 1022)
and 6.42.2 cmH2O (range 410), respectively, but the
difference was not statistically different between the two
groups (p50.17 for inspiratory positive airway pressures and
p50.2 for expiratory positive airway pressures, respectively).
In contrast, the average use of NIV was significantly lower at
9.85.8 h (p,0.001) as eight of the 12 severely obese patients
who received rescue NIV required reintubation within the first
24 h after the onset of respiratory failure.
The institution of NIV post-extubation resulted in a 16% (95%
confidence interval (CI) 2.929.3%) absolute reduction in the
risk of respiratory failure compared with the conventionally
treated group (10 versus 26%; relative risk (RR) 0.38; 95% CI
0.160.89). Table 3 outlines the causes of respiratory failure in
both groups. As NIV rescue therapy resulted in three patients
avoiding reintubation out of the 12 receiving standard therapy,
the rate of reintubation was not statistically significant between
the two groups (10 versus 21%; RR 0.46; 95% CI 0.191.14). The
majority of events responsible for respiratory failure occurred
in the first 24 h as 19 out of the 22 cases of respiratory failure
were reported 8.64.9 h post-extubation. The time to respiratory failure post-extubation was longer for the NIV group (16.0
8.9 h) relative to those assigned to the standard therapy
(8.66.1 h) but the difference fell short of statistical significance (p50.14).
Although the rate of nosocomial bloodstream and respiratory
infections was not significantly different between the two
groups, patients who received NIV therapy had shorter ICU and
hospital lengths of stay compared with those who were assigned
to the conventional medical therapy (p,0.001 and p50.007,
respectively). Hospital mortality, however, was comparable
between the NIV group and the conventional therapy group.
In a post hoc analysis of the 47 patients who had hypercapnia
during a trial of spontaneous breathing, the incidence of
respiratory failure was significantly reduced in those assigned
to the NIV group (p50.03; table 4). There was also a trend
towards a reduction in the frequency of reintubation (p50.1).
Bloodstream and respiratory infections were observed in four
(16%) out of 25 in the NIV group with hypercapnia and 12
(55%) out of 22 in the conventional therapy group (p50.01). As
a result, the mortality rate was significantly higher for patients
with hypercapnia assigned to conventional therapy (p50.03).
Respiratory failure with prolonged mechanical ventilation was
responsible for eight out of the 11 deaths in the conventional
therapy group, compared with two out of the four deaths in
the hypercapnic NIV group (p50.56).
DISCUSSION
The main finding of the present study is that the use of NIV
therapy in severely obese patients resulted in a 16% absolute
EUROPEAN RESPIRATORY JOURNAL

A.A. EL SOLH ET AL.

PREVENTION OF POST-EXTUBATION FAILURE

Severely obese requiring

endotracheal intubation 48 h
(n=101)

Severely obese requiring

intubation 48 h
(n=83)

Not eligible (n=15)

Tracheostomies (n=14)
Terminal weaning (n=1)

Eligible patients
(n=69)

Not eligible (n=21)

Tracheostomies (n=18)
Terminal weaning (n=2)
Transferred (n=1)

Eligible patients
(n=80)

Excluded (n=6)
Unplanned extubation (n=3)
Delirium (n=1)
DNR (n=1)
Facial surgery (n=1)

NIV
(n=62)
Excluded (n=7)
Unplanned extubation (n=2)
Delirium (n=3)
Cranial surgery (n=1)
Facial surgery (n=1)

Respiratory failure
(n=16)

Respiratory failure
(n=6)

Reintubation
(n=6)

Reintubation
(n=4)

Rescue NIV
(n=12)

No reintubation
(n=3)

FIGURE 1.

Conventional therapy
(n=73)
62 matched to NIV

Reintubation
(n=9)

Flow diagram of the study population. DNR: do not resuscitate; NIV: noninvasive ventilation.

reduction in the rate of respiratory failure post-extubation

compared with conventional medical therapy. NIV was
notably more beneficial in those who showed evidence of
persistent hypercapnia during weaning trials.
Earlier investigations suggested that the prophylactic use of
NIV in morbidly obese patients during the first 24 h postoperatively significantly reduced pulmonary dysfunction after
gastroplasty and accelerated re-establishment of pre-operative
pulmonary function [13]. JORIS et al. [13] demonstrated that the
application of bilevel positive airway pressure set at 12 cmH2O
and 4 cmH2O significantly improved peak expiratory flow
EUROPEAN RESPIRATORY JOURNAL

rate, forced vital capacity and Sa,O2 on the first post-operative

day. This improvement was attributed to a combined effect of
improved lung inflation, prevention of alveolar collapse and
reduced inspiratory threshold load. However, the study did
not investigate the potential benefit of bilevel positive airway
pressure on the incidence of post-operative pulmonary
complications nor did it assess the impact on mortality or
length of stay.
Subsequent studies that have examined the role of NIV in the
post-extubation period have assessed outcomes only after
respiratory failure has occurred. Two randomised controlled
VOLUME 28 NUMBER 3

591

PREVENTION OF POST-EXTUBATION FAILURE

TABLE 1

A.A. EL SOLH ET AL.

Baseline characteristics of the study population

NIV

Subjects n

Conventional therapy

p-value

62

62

Age yrs

47.611.7

50.112.5

Sex M/F

37/25

33/29

0.59

47.410.7

45.67.6

0.29

BMI kg?m-2

0.27

Comorbidities
Chronic heart diseases#

6 (10)

8 (13)

0.78

Chronic pulmonary diseases"

13 (21)

16 (26)

0.67

Hypertension

39 (63)

30 (48)

0.15

Diabetes mellitus

33 (53)

27 (43)

0.37

Obstructive sleep apnoea

14 (23)

17 (27)

0.68

Underlying causes for mechanical ventilation

Cardiac failure

15 (24)

11 (18)

0.51

8 (13)

14 (22)

0.24

33 (53)

27 (44)

0.37

3 (5)

4 (6)

0.83

Sepsis
Respiratory failure
Gastrointestinal+
Neurological1
APACHE II on admission

3 (5)

6 (10)

0.49

21.84.6

23.15.7

0.17

Data are presented as meanSD and n (%), unless otherwise stated. NIV: noninvasive ventilation; M: male; F: female; BMI: body mass index; APACHE: Acute Physiology
and Chronic Health Evaluation. #: chronic heart diseases include coronary artery disease, valvular heart diseases and cardiomyopathy of any cause; ": chronic pulmonary
diseases refer to the presence of obstructive lung disease, chronic hypercapnia or interstitial or occupational lung disease; +: underlying gastrointestinal causes for
mechanical ventilation included pancreatitis, diffuse colitis and cholecystitis; 1: underlying neurological causes for mechanical ventilation included cerebrovascular
accidents and seizure disorders.

trials [6, 7] involving a heterogenous group of patients found

that the use of NIV post-extubation did not reduce the rate of
reintubation when compared with standard medical therapy.
In contrast, the current study found that early intervention
with NIV post-extubation reduced the frequency of respiratory
failure and demonstrated a trend toward avoiding reintubation. The difference in these observations can be explained by
the fact that the current protocol required immediate institution of NIV post-extubation rather than intervention after
clinical signs of respiratory failure became evident. Once
respiratory distress ensues, delay in reversing respiratory
TABLE 2

failure has been shown to contribute to organ dysfunction and

poor outcomes [16]. Along this line, JIANG et al. [17] examined
the efficacy of prophylactic NIV in a randomised trial by
assigning 93 patients to either biphasic positive airway
pressure or unassisted oxygen therapy, but the difference in
rates of reintubation in the trial was not statistically significant
between the two groups. However, the study recruited
participants indiscriminately, including those with
unplanned extubation. The relevance of patient selection was
highlighted in two recently published trials [18, 19]. In a
multicentre randomised controlled trial, NAVA et al. [18] found

Characteristics of the patients at the time of extubation

NIV

Subjects n

Conventional therapy

p-value

62

62

7.93.6

8.84.2

0.23

14.83.4

15.53.5

0.32

Heart rate min-1

96.918.8

99.319.9

0.51

Respiratory rate min-1

23.74.3

22.64.7

0.43

fR/VT ratio min-1?L-1

74.611.6

71.513.7

0.16

Arterial pH

7.390.05

7.40.06

0.58

Pa,CO2 mmHg

47.18.9

44.87.7

0.13

Pa,O2 mmHg

8314

8412

0.67

21538

22141

0.38

Duration of mechanical ventilation days

APACHE II at the time of extubation
Physiological indices during
spontaneous breathing trial

Pa,O2/FI,O2

Data are presented as meanSD, unless otherwise stated. NIV: noninvasive ventilation; APACHE: Acute Physiology and Chronic Health Evaluation; fR/VT: ratio of
respiratory frequency to tidal volume; Pa,CO2: carbon dioxide arterial tension; Pa,O2: oxygen arterial tension; FI,O2: inspiratory oxygen fraction. 1 mmHg50.133 kPa.

592

VOLUME 28 NUMBER 3

EUROPEAN RESPIRATORY JOURNAL

A.A. EL SOLH ET AL.

TABLE 3

PREVENTION OF POST-EXTUBATION FAILURE

Post-extubation complications of critically ill morbidly obese patients

NIV

Subjects n

Conventional therapy

62

p-value

62

Respiratory failure

6 (10)

16 (26)

0.03

Reintubation

6 (10)

13 (21)

0.14

Hypoxia

2 (3)

3 (5)

Hypercapnia

2 (3)

9 (15)

Respiratory muscle fatigue

1 (2)

2 (3)

Causes of respiratory failure

Haemodynamic instability

1 (2)

Inability to clear secretions

1 (2)

1 (2)

Hospital-acquired pneumonia

Delirium

3 (5)

9 (15)

Bloodstream infection

2 (3)

5 (8)

0.13
0.44

ICU stay days

11.87.9

18.211.2

,0.001

Hospital stay days

20.610.6

26.011.3

0.007

Hospital mortality

8 (13)

15 (24)

0.17

Data are presented as n (%) and meanSD, unless otherwise stated. NIV: noninvasive ventilation; ICU: intensive care unit.

prophylactic NIV to be more effective in preventing reintubation than standard medical therapy in patients at risk. FERRER et
al. [19] reached a similar conclusion by showing that the early
use of NIV averted respiratory failure in selected patients
considered at risk for respiratory failure post-extubation.
Another potential explanation for the observed benefit in the
NIV cohort pertains to the fact that this study population is
considered at high risk of obstructive sleep apnoea (OSA).
Although the authors were aware of the presence of OSA in
25% of the study population, unrecognised OSA has been
implicated in worsening respiratory failure and unexpected
transfer to the ICU after liberation from mechanical ventilation
[20, 21]. GUPTA et al. [22] reported 24% serious complications
post-hip and -knee replacement in 101 patients with a
diagnosis of OSA within 72 h post-extubation including
reintubation and urgent CPAP. In the absence of high-quality
evidence supporting the routine use of CPAP in severely obese
patients with OSA after liberation from mechanical ventilation,
the early application of NIV in the current study might have
TABLE 4

Characteristics of patients with hypercapnia

during spontaneous breathing trial
NIV

Subjects n

Conventional therapy

p-value

25

22

55.66.6

52.95.7

Respiratory failure

3 (12)

10 (45)

0.03

Reintubation

3 (12)

8 (36)

0.10

14.59.8

22.114.6

0.04

4 (16)

11 (50)

0.03

Pa,CO2 mmHg

ICU length of stay

0.19

days
Hospital mortality

Data are presented as meanSD and n (%), unless otherwise stated. NIV:
noninvasive ventilation; Pa,CO2: carbon dioxide arterial tension; ICU: intensive
care unit. 1 mmHg50.133 kPa.

EUROPEAN RESPIRATORY JOURNAL

averted potential worsening in respiratory status of OSA

patients from residual effects of sedatives and narcotics that
could have developed hours post-extubation.
Most of the published studies assessing the use of NIV in acute
respiratory failure have relied on oro-nasal or facial masks [6,
7, 18, 19]. Although there are no published studies, to the
current authors knowledge, on the superiority of one method
over the other in the obese population, prior experience
suggests a better tolerance for nasal masks in the morbidly
obese, because these tend to be less claustrophobic. One
controlled trial comparing the efficacy of nasal and oro-nasal
masks in 26 nonobese patients with stable hypercapnia caused
by chronic obstructive pulmonary disease (COPD) or restrictive disease found that the oro-nasal mask was more effective
in lowering Pa,CO2 but the nasal mask was better tolerated than
either the nasal pillow or the oro-nasal mask [23]. However,
another preliminary report from a controlled trial comparing
nasal and oro-nasal masks found that the rates of decline of
Pa,CO2 and respiratory rate were independent of mask type
when the masks were used for patients with respiratory
distress [24]. Owing to the ongoing debate, further studies are
needed to elucidate the optimal interface for the most effective
response in obese patients with respiratory failure.
The results of the present study point to preferential benefit of
NIV in severely obese patients with underlying hypercapnia.
The effectiveness of NIV in reducing the rate of hypercapnic
respiratory failure has been suspected from earlier studies
involving patients with COPD exacerbations [25, 26]. In a
nonrandomised trial, HILBERT et al. [26] reported 47% attributable risk reduction in the need for endotracheal intubation for
hypercapnic COPD patients assigned to NIV post-extubation.
A significant decrease in mean duration of ventilatory
assistance for the treatment of post-extubation distress and in
the length of ICU stay was also noted. Comparable to the
observations presented here, NIV was more effective in
patients who exhibited hypercapnia during the spontaneous
breathing trial. However, the present data extended those
VOLUME 28 NUMBER 3

593

PREVENTION OF POST-EXTUBATION FAILURE

A.A. EL SOLH ET AL.

findings by demonstrating a reduced rate of bloodstream and

hospital-acquired pulmonary infections, which were translated
into shorter length of stay and improved hospital mortality. It
is noteworthy that aside from possible random effects as a
consequence of the limitations of post hoc analysis, the current
study was not powered to address the outcome of prophylactic
NIV in hypercapnic obese patients. This hypothesis should be
investigated in a future prospective, randomised controlled
trial.
Among the limitations of the present study is the use of
historically matched controls [27]. A frequently cited source of
bias with historical controls is the mere fact that implementing
a prospective study may improve patient outcome. Moreover,
historical comparisons of two cohorts are influenced by
changes in disease patterns and treatment strategies, and are
considered to favour the group treated with the new method.
However, except for the application of NIV therapy, the
presence of protocols-driven ICU management coupled with
low staff turnover in the chosen institution would militate
against this possibility. Moreover, differences in pulmonary
function among participants might favour one group over the
other, particularly if an obstructive pattern happens to be more
frequent in the NIV group. In the absence of pulmonary
function tests, this possibility should be taken into consideration when interpreting the results of the study. Another
limitation of this study pertains to the potential of high
prevalence of sleep apnoea in the study population, which
might introduce a bias in favour of the NIV group. Pending a
randomised controlled trial, the risk of respiratory complications and poor outcome of prolonged ventilation in this highrisk group population should be balanced against the observed
benefits of a historical controlled investigation.
In conclusion, prophylactic use of noninvasive ventilation in
severely obese patients post-extubation is warranted. Early
application of noninvasive ventilation may be effective in
averting respiratory failure before the development of respiratory distress and may be responsible for decreasing mortality
in selected patients with chronic hypercapnia.
ACKNOWLEDGEMENTS
The authors are indebted to the respiratory therapist unit, the
nursing staff of the intensive care units, and to the patients for
their cooperation in the completion of this study.

REFERENCES
1 El Solh A, Sikka P, Bozkanat E, Jaafar W, Davies J. Morbid
obesity in the medical ICU. Chest 2001; 120: 19891997.
2 Taylor RR, Kelly TM, Elliott CG, Jensen RL, Jones SB.
Hypoxemia after gastric bypass surgery for morbid
obesity. Arch Surg 1985; 120: 12981302.
3 Blouw E, Rudolph A, Narr B, Sarr M. The frequency of
respiratory failure in patients with morbid obesity undergoing gastric bypass. AANA J 2003; 71: 4550.
4 Gaszynski T, Gaszynski W, Strzelczyk J. Critical respiratory events in morbidly obese. Twoj Magazyn Medyczny
Chirurgia 2003; 3: 5558.
5 American Thoracic Society, European Respiratory Society,
European Society of Intensive Care Medicine, and Societe de
Reanimation de Langue Franccaise. International Consensus
594

VOLUME 28 NUMBER 3

10

11

12

13

14

15

16

17

18

19

20

Conferences in Intensive Care Medicine: noninvasive positive pressure ventilation in acute respiratory failure. Am J
Respir Crit Care Med 2001; 163: 283291.
Esteban A, Frutos-Vivar F, Ferguson ND, et al.
Noninvasive positive-pressure ventilation for respiratory
failure after extubation. N Engl J Med 2004; 350: 24522460.
Keenan SP, Powers C, McCormack DG, Block G.
Noninvasive positive-pressure ventilation for postextubation respiratory distress: a randomized controlled trial.
JAMA 2002; 287: 32383244.
Ferrer M, Valencia M, Bernardich O, et al. Non-invasive
ventilation in the prevention of extubation failure. A
randomized clinical trial. Am J Respir Crit Care Med 2004;
169: A523.
Bercault N, Boulain T, Kuteifan K, Wolf M, Runge I,
Fleury J. Obesity-related excess mortality in an adult
intensive care unit: a risk-adjusted matched cohort study.
Crit Care Med 2004; 32: 9981003.
Knaus WA, Draper EA, Wagner DP, Zimmerman J.
APACHE II: a severity of disease classification system.
Crit Care Med 1985; 13: 818829.
MacIntyre N, Cook D, Wesley E, et al. Evidence-based
guidelines for weaning and discontinuing ventilatory
support: a collective task force facilitated by the
American College of Chest Physicians, the American
Association for Respiratory Care and the American
College of Critical Care Medicine. Chest 2001; 120: Suppl.
6, 375S395S.
Ely E, Baker A, Evans G, Haponik E. The prognostic
significance of passing a daily screen of weaning parameters. Intensive Care Med 1999; 25: 581587.
Joris J, Sottiaux T, Chiche J, Desaive C, Lamy M. Effect of
bi-level positive airway pressure nasal ventilation on the
postoperative pulmonary restrictive syndrome in obese
patients undergoing gastroplasty. Chest 1997; 111: 665670.
Ferrer M, Esquinas A, Arancibia F, Gonzalez G, Alarcon A,
Torres A. Noninvasive ventilation during persistent
weaning failure. A randomized controlled trial. Am J
Respir Crit Care Med 2003; 168: 7076.
Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM.
CDC definitions for nosocomial infections. J Infect Control
1988; 16: 128140.
Epstein SK, Ciubotaru RL. Independent effects of etiology
of failure and time of reintubation on outcome for patients
failing extubation. Am J Respir Crit Care Med 1998; 158:
489493.
Jiang J, Kao S, Wang S. Effect of early application of
biphasic positive airway pressure on the outcome of
extubation in ventilator weaning. Respirology 1999; 4:
161165.
Nava S, Gregoretti C, Fanfulla F, et al. Noninvasive
ventilation to prevent respiratory failure after extubation
in high-risk patients. Crit Care Med 2005; 33: 24652470.
Ferrer M, Valencia M, Nicolas JM, Bernadich O, Badia J,
Torres A. Early noninvasive ventilation averts extubation
failure in patients at risk: a randomized trial. Am J Respir
Crit Care Med 2006; 173: 164170.
Erman MK, Loube DI. Perioperative complications in
obstructive sleep apnea patients. Sleep Breathing 1997; 1:
310.
EUROPEAN RESPIRATORY JOURNAL

A.A. EL SOLH ET AL.

PREVENTION OF POST-EXTUBATION FAILURE

21 Benumof JL. Obstructive sleep apnea in the adult obese

patient: implications for airway management. J Clin Anesth
2001; 13: 144156.
22 Gupta R, Parvizi J, Hanssen A, Gay P. Postoperative
complications in patients with obstructive sleep apnea
syndrome undergoing hip or knee replacement: a casecontrol study. Mayo Clin Proc 2001; 76: 897905.
23 Navalesi P, Fanfulla F, Frigeiro P, Gregoretti C, Nava S.
Physiologic evaluation of noninvasive mechanical ventilation delivered with three types of masks in patients with
chronic hypercapnic respiratory failure. Chest 2000; 28:
17851790.
24 Kwok H, McCormack J, Cece R, Houtchens J, Hill N.
Controlled trial of oronasal versus nasal mask ventilation in

the treatment of acute respiratory failure. Crit Care Med

2003; 31: 468473.
25 Plant PK, Owen JL, Elliott MW. Early use of non-invasive
ventilation for acute exacerbations of chronic obstructive
pulmonary disease on general respiratory wards: a multicentre randomised controlled trial. Lancet 2000; 355:
19311935.
26 Hilbert G, Gruson D, Portel L, Gbikpi-Benissan G,
Cardinaud JP. Noninvasive pressure support ventilation
in COPD patients with postextubation hypercapnic
respiratory insufficiency. Eur Respir J 1998; 11: 13491353.
27 Sacks H, Chalmers T, Smith H. Randomized versus
historical controls for clinical trials. Am J Med 1982; 72:
233240.

EUROPEAN RESPIRATORY JOURNAL

VOLUME 28 NUMBER 3

595