Concise Definitive Review R.

Phillip Dellinger, MD, FCCM, Section Editor

Noninvasive ventilation in acute respiratory failure

Nicholas S. Hill, MD; John Brennan, MD; Erik Garpestad, MD; Stefano Nava, MD

Background: Noninvasive ventilation has assumed an impor- exacerbations, pneumonia, acute lung injury, or acute respiratory
tant role in the management of respiratory failure in critical distress syndrome; during bronchoscopy; or as a means of
care units, but it must be used selectively depending on the preoxygenation before intubation in critically ill patients with
patient’s diagnosis and clinical characteristics. severe hypoxemia.
Data: We review the strong evidence supporting the use of Conclusion: Noninvasive ventilation has assumed an impor-
noninvasive ventilation for acute respiratory failure to prevent tant role in managing patients with acute respiratory failure.
intubation in patients with chronic obstructive pulmonary disease Patients should be monitored closely for signs of noninvasive
exacerbations or acute cardiogenic pulmonary edema, and in ventilation failure and promptly intubated before a crisis de-
immunocompromised patients, as well as to facilitate extubation velops. The application of noninvasive ventilation by a trained
in patients with chronic obstructive pulmonary disease who re- and experienced intensive care unit team, with careful patient
quire initial intubation. Weaker evidence supports consideration selection, should optimize patient outcomes. (Crit Care Med
of noninvasive ventilation for chronic obstructive pulmonary dis- 2007; 35:2402–2407)
ease patients with postoperative or postextubation respiratory KEY WORDS: noninvasive ventilation; acute respiratory failure;
failure; patients with acute respiratory failure due to asthma intubation prevention

N ill patients (i.e., those with a pH ⬍7.2)

oninvasive ventilation (NIV), pists (2). This review will focus on the
a form of ventilatory support evidence supporting the use of NIV in var- has been controversial. However, recent
that avoids airway invasion, ious forms of respiratory failure (Table 1), studies demonstrate that outcomes of se-
has seen increasing use in selection of appropriate patients, and tech- vere COPD exacerbations are no worse if
emergency departments and intensive niques for successful implementation. treated with NIV than with endotracheal
care units (ICUs) in recent years, based intubation (10, 11), indicating that an
on the results of clinical trials showing NIV FOR SPECIFIC TYPES OF initial trial with NIV is not deleterious,
improved outcomes in certain types of even in severely ill COPD patients.
ACUTE RESPIRATORY FAILURE:
acute respiratory failure (ARF) (1). NIV Asthma. Evidence is weaker for the
usually refers to the provision of inspira-
EVIDENCE AND
use of NIV in asthma patients with acute
tory pressure support plus positive end- RECOMMENDATIONS
respiratory failure. An uncontrolled study
expiratory pressure (PEEP) via a nasal or (12) observed improved gas exchange and
face mask. Although continuous positive Hypercapnic Respiratory Failure avoidance of intubation in 15 of 17 pa-
airway pressure (CPAP) does not actively tients with status asthmaticus, and all
assist inspiration and is not a ventilatory Chronic Obstructive Pulmonary Dis-
ease. NIV should be considered first-line patients survived. A subsequent random-
support mode, it is considered a form of ized pilot study in 33 patients with acute
NIV here when used as a therapy for ARF. therapy in the management of ARF
caused by chronic obstructive pulmonary asthma but not ARF showed improved
The successful application of NIV re- flow rates and decreased hospitalizations
quires the training and collaboration of disease (COPD) exacerbations based on
evidence derived from multiple random- with NIV vs. sham NIV (13). However, Dr.
an experienced ICU team, including in- Ram and colleagues (14) concluded that
tensivists, nurses, and respiratory thera- ized trials (2–5). Meta-analyses by Dr.
Lightowler and colleagues (6) and Dr. large randomized controlled trials (RCTs)
Keenan and colleagues (7) show reduced are needed before recommending NIV use
intubation rates, hospital lengths of stay, in status asthmaticus. A trial of NIV can be
From the Division of Pulmonary, Critical Care, and and mortality with NIV use. The strength considered in asthmatics who fail to re-
Sleep Medicine, Tufts-New England Medical Center, spond adequately to initial bronchodilator
of the evidence justifies the application of
Boston, MA (NSH, JB, EG); and the Respiratory Inten-
NIV as a standard of care in appropriately therapy to improve air flow obstruction and
sive Care Unit, Fondazione Salvatore Maugeri, Istituto
di Ricovero e Cura a Carattere Scientifico, Istituto selected patients with hypercapnic ARF decrease the work of breathing. Patients
Scientifico di Pavia, Pavia, Italy (SN). due to COPD. Although considered a con- should be monitored closely and intubated
Consultant fees and research grants received from traindication to NIV in the past, hyper- promptly if there is no improvement in the
ResMed (NSH) and Respironics (NSH, SN).
capnic coma in patients with COPD can first hour or two, because these patients
For information regarding this article, E-mail:
[email protected] be treated with NIV as successfully as in can deteriorate rapidly.
Copyright © 2007 by the Society of Critical Care noncomatose patients (8, 9). Facilitating Extubation in COPD. Fa-
Medicine and Lippincott Williams & Wilkins The use of NIV as an alternative to cilitation of extubation in patients with
DOI: 10.1097/01.CCM.0000284587.36541.7F endotracheal intubation in more severely ARF due to COPD exacerbations is an-

2402 Crit Care Med 2007 Vol. 35, No. 10

Table 1. Noninvasive ventilation for various types of acute respiratory failure (ARF): Evidence for been identified as a risk factor for NIV
efficacy and strength of recommendation failure (27). Two thirds of patients with
severe community-acquired pneumonia
Type of ARF Level of Evidencea Strength of Recommendationb
required intubation after being started on
Hypercapnic respiratory failure NIV in one cohort study, even though
COPD exacerbation A Recommended those who succeeded with NIV had very
Asthma C Option good outcomes (28). An RCT on patients
Facilitation of extubation (COPD) A Guideline with severe community-acquired pneu-
Hypoxemic respiratory failure
Cardiogenic pulmonary edema A Recommended
monia showed that NIV reduced intuba-
Pneumonia C Option tion rates, ICU length of stay, and
ALI/ARDS C Option 2-month mortality rate, but only in the
Immunocompromised A Recommended subgroup with underlying COPD (29).
Postoperative respiratory failure B Guideline Another RCT on patients with hypoxemic
Extubation failure C Guideline
Do not intubate status C Guideline respiratory failure (30) showed that NIV
Preintubation oxygenation B Option reduced the need for intubation among
Facilitation of bronchoscopy B Guideline patients with pneumonia (26% vs. 73% in
the conventional therapy group), but a
COPD, chronic obstructive pulmonary disease; ALI, acute lung injury; ARDS, acute respiratory more recent RCT (31) testing NIV as an
distress syndrome.
a alternative to invasive ventilation in pa-
A, multiple randomized controlled trials and meta-analyses; B, more than one randomized,
controlled trial, case control series, or cohort studies; C, case series or conflicting data; b recom-
tients with various types of ARF found
mended, first choice for ventilatory support in selected patients; Guideline, can be used in appropriate that the subgroup with pneumonia did
patients but careful monitoring advised; Option, suitable for a very carefully selected and monitored very poorly, with all eight patients ran-
minority of patients. domized to NIV requiring intubation. The
scant and conflicting data do not support
the routine use of NIV in patients with
severe pneumonia, with the exception of
other application of NIV supported by lowing will consider studies focusing on
patients with underlying COPD. How-
strong evidence. In intubated patients the subcategories.
ever, a cautious trial of NIV may be con-
with COPD and hypercapnic respiratory Cardiogenic Pulmonary Edema. The
sidered in patients with pneumonia
failure who failed a single (15) or re- use of NIV or CPAP in patients with cardio-
deemed to be excellent candidates, but
peated T-piece trials (16) and were extu- genic pulmonary edema is supported by
they need careful monitoring, because
bated to NIV or continued on invasive multiple randomized trials (17–23). The
the risk of failure is high.
ventilation and weaned according to a main physiologic benefit from NIV or CPAP
Acute Lung Injury/Acute Respiratory
standard pressure support protocol, ran- in these patients is likely due to an increase
Distress Syndrome. Studies on NIV to
domized trials showed an increased in functional residual capacity that reopens treat acute lung injury (ALI) and acute
weaning rate at 28 days, decreased dura- collapsed alveoli and improves oxygenation. respiratory distress syndrome (ARDS)
tions of mechanical ventilation and ICU This also increases lung compliance and have reported failure rates ranging from
stay, and reduced rates of nosocomial reduces work of breathing. The increased 50% to ⬎80% (27, 30, 32, 33), but no
pneumonia and 60-day mortality. Based intrathoracic pressure also can improve RCTs have focused on ALI/ARDS exclu-
on these findings, patients intubated for cardiac performance by decreasing ventric- sively. Independent risk factors for NIV
hypercapnic respiratory failure due to ular preload and afterload. failure in this group of patients include
COPD who fail spontaneous breathing tri- Several meta-analyses have shown severe hypoxemia, shock, and metabolic
als should be considered for a trial of extu- equivalent reductions in intubation and acidosis (33). A recent prospective multi-
bation to NIV. This approach should be mortality rates with CPAP and NIV for center survey found that when NIV was
reserved for patients who are good candi- cardiogenic pulmonary edema, although used as first-line therapy for selected ALI/
dates for NIV in other respects and who are the reduction in mortality with NIV was ARDS patients (those with ⬎2 organ fail-
able to tolerate levels of pressure support not statistically significant in some, prob- ures, hemodynamic instability, or en-
easily administered via mask (i.e., ⱕ15 cm ably because there were fewer NIV than cephalopathy were excluded), 54%
H2O). In addition, they should not have CPAP studies (24, 25). However, several avoided intubation and had excellent out-
been a difficult intubation. studies have shown more rapid reduc- comes (34). Predictors of NIV failure were
tions in respiratory rate and dyspnea with Simplified Acute Physiology Score II ⬎34
Hypoxemic Respiratory Failure NIV than with CPAP alone (20, 26). Thus, and PaO2/FIO2 ⱕ175 after the first hour of
either NIV or CPAP can be used to treat therapy. Thus, although NIV cannot be
Hypoxemic ARF is defined by a PaO2/ cardiogenic pulmonary edema with equal recommended as routine therapy for ALI/
FIO2 ratio ⬍200 and a variety of different expectations of success. Some recom- ARDS, these data support a cautious trial
non-COPD etiologies. It is a very broad mend starting with CPAP, because it is a in highly selected patients with a Simpli-
category, and studies focusing on it are simpler and potentially less expensive fied Acute Physiology Score II ⱕ34 and
difficult to apply to individual patients. therapy, with pressure support added if readiness to promptly intubate if oxygen-
The concern is that responses within an patients remain dyspneic or hypercapnic ation fails to improve sufficiently within
etiologic subcategory— harmful or bene- on CPAP alone. the first hour.
ficial— could be obscured by responses in Pneumonia. Pneumonia has been a Respiratory Failure in Immunocom-
other subcategories. Therefore, the fol- challenge to treat noninvasively and has promised Patients. RCTs in recipients of

Crit Care Med 2007 Vol. 35, No. 10 2403

solid-organ or bone-marrow transplants therapy) as soon as patients developed ratory failure are at high risk of oxygen
who developed hypoxemic respiratory signs of extubation failure (43). Surpris- desaturations during intubation. A recent
failure have found decreased intubation ingly, not only did NIV fail to reduce RCT of such patients (49) showed that
and ICU mortality rates and shorter ICU reintubations, but its use also was asso- preoxygenation with NIV before intubation
lengths of stay in patients treated with ciated with increased ICU mortality, resulted in improved oxygen saturation
NIV as compared with conventional ther- thought to be related to delays in needed during and after intubation and decreased
apy (35, 36). Similar findings have been reintubation. Only 10% of patients in this the incidence of oxygen desaturations be-
reported for acquired immune deficiency trial had COPD, however. Two subse- low 80% during intubation. This approach
syndrome patients in a nonrandomized quent RCTs (44, 45) on patients deemed is promising but should be further stud-
study (37). The reduced mortality is likely to be at high risk for extubation failure ied before routine use can be recom-
related to reduced infectious complica- found that NIV reduced the need for re- mended.
tions associated with NIV use compared intubation and ICU mortality, but in one Fiberoptic Bronchoscopy. A RCT has
with endotracheal intubation, including of the studies the hypercapnic subgroup shown that CPAP alone (up to 7.5 cm
ventilator-associated pneumonia, other (mainly COPD patients) had most of the H2O) improves oxygenation and reduces
nosocomial infections, and septic shock benefit (45). These data support the use postprocedure respiratory failure in pa-
(38). These data support NIV as the pre- of NIV in patients at high risk of extuba- tients with severe hypoxemia undergoing
ferred initial ventilatory modality for tion failure, particularly if they have bronchoscopy (50). In an RCT (51) of 26
these patients, to avoid intubation and its COPD, congestive heart failure, and/or patients with hypoxemia (PaO2/FIO2 ratio
associated risks. hypercapnia. However, early indiscrimi- ⬍200), NIV increased PaO2/FIO2 by 82%
nate use in all patients with risk factors is compared with a 10% worsening in the
Postoperative Respiratory discouraged. Patients with extubation conventional oxygen therapy group. Suc-
Failure failure treated with NIV should be moni- cessful bronchoscopy during NIV also has
tored closely and delays in needed intu- been reported in hypercapnic COPD pa-
Both CPAP and NIV have shown ben- bation avoided. tients with pneumonia. NIV improved oxy-
efit in the postoperative period. When gen saturation, and all 10 patients tolerated
used prophylactically after major abdom- Palliative Care and the procedure without complications (52).
inal surgery (39) or thoracoabdominal Do-Not-Intubate Status The evidence supports the use of NIV dur-
aneurysm repair (40), CPAP (10 cm H2O) ing fiberoptic bronchoscopy, especially
A prospective cohort series (46) eval-
reduces the incidence of hypoxemia, when risks of intubation are deemed high,
uated 114 patients with acute respiratory
pneumonia, atelectasis, and intubations such as in immunocompromised patients
failure and a status of do not intubate.
compared with standard treatment. In or in those with bleeding diatheses. How-
Overall, 43% of the patients survived the
the only RCT of NIV in the postoperative ever, the ICU team must be prepared for
hospitalization, and those with diagnoses
setting, patients with hypoxemic respira- the possibility of emergent intubation.
of cardiogenic pulmonary edema and
tory failure after lung resection had re-
COPD had hospital survival rates ⬎50%.
duced intubation and mortality rates if PATIENT SELECTION
The presence of a cough and an awake
treated with NIV when compared with
mental status imparted a favorable prog-
standard management (41). Because Selection of appropriate patients is an
nosis. Another prospective cohort series
these studies have examined different important skill that is key to the success-
also showed favorable success rates in do
techniques following various surgeries, ful application of NIV. In brief, it is a
not intubate patients with COPD and car-
firm specific recommendations cannot be clinical judgment that takes into consid-
diogenic pulmonary edema, but a high
made. However, the data lend support to eration the etiology of the ARF and evi-
failure rate in patients with hypoxemic
the use of CPAP or NIV in postoperative dence for efficacy (Table 1). Good candi-
respiratory failure, postextubation fail-
patients, either prophylactically in high- dates for NIV have a demonstrable need
ure, and end-stage cancer (47). Depend-
risk patients or as an early therapy of for ventilatory assistance and no contra-
ing on patient and/or family wishes, a
respiratory insufficiency. indications to NIV (Table 2). In addition,
trial of NIV can be considered in do-not-
consideration of predictors of success and
intubate patients, but the goals of therapy
Postextubation Respiratory failure may be helpful (Table 3); intuba-
should be clear (48). If the patient and/or
Failure tion is preferred if the likelihood of fail-
family desire prolonged survival, then use
ure is too great. Numerous studies have
should be reserved primarily for COPD
Extubation failure is associated with demonstrated that the response to NIV
and congestive heart failure patients. On
high morbidity and mortality, and NIV after the first hour or two is the best
the other hand, if the goal is to palliate, to
has been suggested as a way to avoid predictor of eventual outcome (27, 34,
relieve dyspnea, or to delay death so that
re-intubation and improve outcomes. An 53). Therefore, when in doubt, a brief,
affairs can be settled, then NIV can be
earlier RCT found no reduction in rein- cautious trial of NIV can be attempted,
used for these as well as other diagnoses.
tubations among patients who developed with plans to intubate if the patient fails
However, it should be reassessed fre-
respiratory distress within 48 hrs of ex- to improve sufficiently.
quently and stopped if the goal of pallia-
tubation, although few patients with
tion is not being met.
COPD were included in this study and the NIV TECHNIQUES
level of pressure support used may have
Other ICU Applications of NIV
been subtherapeutic (42). Another RCT Although technical aspects such as
attempted to prevent extubation failure Preoxygenation Before Intubation. choice of interface and ventilator settings
by starting NIV (or continuing standard Critically ill patients with hypoxic respi- are clearly important to NIV success, the

2404 Crit Care Med 2007 Vol. 35, No. 10

Table 2. General guidelines for selection of patients for noninvasive ventilation In a study designed to determine op-
timal NIV settings in patients with ALI, a
1. Need for ventilatory assistance? 2. Contraindications for NIV?
PEEP of 10 cm H2O improved PaO2/FIO2
Moderate to severe dyspnea Respiratory arrest
Tachypnea (⬎24 for hypercapnic, 30 for hypoxemic) Medically unstable more than a PEEP of 5 cm H2O (58). A
Accessory muscle use Unable to protect airway pressure support of 15 cm H2O and a
Abdominal paradox Excessive secretions PEEP of 5 cm H2O decreased PaCO2, re-
PaCO2 ⬎45 mm Hg, pH ⬍7.35 Agitated, uncooperative spiratory rate, and work of breathing, and
PaO2/FIO2 ⬍200 Recent UGI or airway surgery
Unable to fit mask improved dyspnea more than a pressure
support of 10 cm H2O and PEEP of 10 cm
NIV, noninvasive ventilation; UGI, upper gastrointestinal. H2O, even though PaO2/FIO2 was better
with the higher PEEP. This study shows
Table 3. Predictors of failure: Noninvasive ven- Table 4. Monitoring of noninvasive ventilation that adjusting settings may require a bal-
tilation (NIV) for acute respiratory failurea for acute respiratory failure ancing of beneficial and adverse effects to
come up with the best combination for a
COPD Subjective given patient.
Air leaking Mask comfort Monitoring of NIV. To assure the suc-
APACHE II ⱖ29b Tolerance of ventilator settings
Asynchrony Respiratory distress cess of NIV, close monitoring is neces-
Copious secretions Physical findings sary, especially during the initiation pe-
Glasgow Coma Score ⱕ11b Respiratory rate riod (Table 4). Favorable subjective
Lack of “compliance” or “tolerance” Other vital signs responses—including tolerance of the
pH ⬍7.25b Accessory muscle use
Respiratory rate ⱖ35 breaths/minb Abdominal paradox
mask and air pressure and reduction of
Hypoxemic respiratory failure Ventilator parameters respiratory distress and effort—are im-
ALI/ARDS Air leaking portant to establish early. Air leaking
SAPS II ⱖ35 Adequacy of pressure support should be sought and minimized, and gas
Metabolic acidosis Adequacy of PEEP exchange should be stabilized and im-
PaO2/FIO2 ⱕ146 (or ⱕ175 for ARDS) after 1 Tidal volume (5–7 mL/kg)
hr of NIV Patient-ventilator synchrony proved.
Pneumonia Gas exchange The location of NIV delivery also is
Severe hypoxemia Continuous oximetry (until stable) important to assure adequate monitor-
Shock ABGs, baseline and 1–2 hrs, then as indicated ing. Most studies have monitored pa-
Location
Usually ICU or respiratory care unit to start
tients in ICUs or respiratory step-down
COPD, chronic obstructive pulmonary dis-
General ward may be OK if patient stable units, but some have reported successful
ease; APACHE II, Acute Physiology and Chronic
Health Evaluation II; ALI, acute lung injury; Depends on monitoring needs of patients and application of NIV on general medical
ARDS, acute respiratory distress syndrome; SAPS monitoring capabilities of unit wards (59, 60). Monitoring should be tai-
II, Simplified Acute Physiology Score II. lored to the acuity of illness. If the patient
a PEEP, positive end-expiratory pressure; becomes unstable within minutes of
Based on composite from Ambrosino Thorax
ABGs, arterial blood gases; ICU, intensive care
‘94Confalonieri, Rana, Antonelli ICM 2001, An- mask removal, then close monitoring is
unit.
tonelli CCM 2006; b likelihood of failure ⬎50% mandatory, ideally in an ICU or respira-
if any three and 82% if all four present at base- tory unit.
line; ⬎75% if any three and 99% if all four critical care ventilators now offer NIV
present after 2 hrs of NIV. modes that provide pressure support ven- CONCLUSION
tilation with leak compensation and si-
lencing of nuisance alarms, little data are Strong evidence from randomized tri-
human factor is important, as well. As available to demonstrate clinical efficacy. als supports the use of NIV in the man-
caregivers in one ICU gained experience The determination of optimal ventilator agement of ARF to prevent endotracheal
with NIV, success rates remained stable settings also has been inadequately stud- intubation in patients with COPD exacer-
even as more severely ill patients were ied, but is usually a process of balancing bations or acute cardiogenic pulmonary
treated (54). the ability to reduce work of breathing by edema, and in immunocompromised pa-
Selection of an Interface. Selection of providing an adequate level of pressure tients, as well as to facilitate extubation
a properly fit and comfortable interface is support (usually ⬎8 –10 cm H2O) against in patients with COPD. NIV should be
critical to NIV success. Although patients the discomfort and greater air leaking contemplated in patients with postopera-
rate nasal masks as more comfortable imposed by higher pressures. A common tive respiratory failure or at high risk for
than full face masks (55), an RCT dem- mistake is to start with a low inspiratory postextubation respiratory failure who
onstrated that full face masks are better pressure to facilitate tolerance and then are otherwise good candidates for NIV,
tolerated than nasal masks for ARF be- to fail to titrate pressures upward to re- and as a means of preoxygenating criti-
cause of less air leaking through the duce respiratory effort. cally ill patients with hypoxemia before
mouth (56). In COPD, the inspiratory threshold intubation. NIV can be considered in pa-
Selection of a Ventilator and Ventila- load imposed by intrinsic PEEP is a major tients with asthma exacerbations, pneu-
tor Settings. Either critical care or bilevel component of the work of breathing. Ex- monia, and ALI/ARDS, although the sup-
positive-pressure ventilators can be used ternal PEEP, usually not higher than 5 porting evidence is fairly weak; these and
to administer NIV to ARF patients; no cm H2O to 7 cm H2O, should be provided other acutely ill patients should be mon-
study has demonstrated superiority of with pressure support to these patients to itored closely for signs of NIV failure until
one type over the other. Although many minimize inspiratory effort (57). stabilized. If there are signs of NIV fail-

Crit Care Med 2007 Vol. 35, No. 10 2405

ure, patients should be intubated invasive positive pressure ventilation in sta- continuous positive airway pressure in acute
promptly before a crisis develops. The tus asthmaticus. Chest 1996; 110:767–774 pulmonary edema. Crit Care Med 1997; 25:
application of NIV by a trained and expe- 13. Soroksky A, Stav D, Shpirer I: A pilot pro- 620 – 628
rienced ICU team, with careful patient spective, randomized, placebo-controlled 27. Antonelli M, Conti G, Moro ML, et al: Pre-
trial of bi-level positive airway pressure in dictors of failures of noninvasive positive
selection, should optimize patient out-
acute asthmatic attack. Chest 2003; 123: pressure ventilation in patients with acute
comes. 1018 –1025 hypoxemic respiratory failure: A multi-
14. Ram FS, Wellington S, Rowe B, et al: Non- center study. Intensive Care Med 2001; 27:
invasive positive pressure ventilation for 1718 –1728
REFERENCES treatment of respiratory failure due to severe 28. Jolliet P, Abajo B, Pasquina P, et al: Non-
acute exacerbations of asthma. Cochrane Da- invasive pressure support ventilation in se-
1. Mehta S, Hill NS: Noninvasive ventilation: tabase Syst Rev 2005; 1:CD004360 vere community-acquired pneumonia. In-
State of the art. Am J Respir Crit Care Med 15. Nava S, Ambrosino N, Clini E, et al: Non- tensive Care Med 2001; 27:812– 821
2001; 163:540 –577 invasive mechanical ventilation in the wean- 29. Confalonieri M, Potena A, Carbone G, et al:
2. Brochard L, Isabey D, Piquet J, et al: Reversal ing of patients with respiratory failure due to Acute respiratory failure in patients with se-
of acute exacerbations of chronic obstructive chronic obstructive pulmonary disease: A vere community-acquired pneumonia. A pro-
lung disease by inspiratory assistance with a randomized study. Ann Intern Med 1998; spective randomized evaluation of noninva-
face mask. N Engl J Med 1990; 323: 128:721–728 sive ventilation. Am J Respir Crit Care Med
1523–1530 16. Ferrer M, Esquinas A, Arancibia F, et al: 1999; 160:1585–1591
3. Bott J, Carroll MP, Conway JH, et al: Ran- Noninvasive ventilation during persistent 30. Ferrer M, Esquinas A, Leon M, et al: Nonin-
domised controlled trial of nasal ventilation weaning failure. A randomized controlled vasive ventilation in severe hypoxemic respi-
in acute ventilatory failure due to chronic trial. Am J Respir Crit Care Med 2003; 168: ratory failure. A randomized clinical trial.
obstructive airways disease. Lancet 1993; 70 –76 Am J Respir Crit Care Med 2003; 168:
341:1555–1557 17. Bersten AD, Holt AW, Vedig AE, et al: Treat- 1438 –1444
4. Kramer N, Meyer TJ, Meharg J, et al: Ran- ment of severe cardiogenic pulmonary 31. Honrubia T, Garcia Lopez FJ, Franco N, et al:
domized, prospective trial of noninvasive edema with continuous positive airway pres- Noninvasive vs. conventional mechanical
positive pressure ventilation in acute respi- sure delivered by face mask. N Engl J Med ventilation for acute respiratory failure.
ratory failure. Am J Respir Crit Care Med 1991; 325:1825–1830 Chest 2005; 128:3916 –3924
1995; 151:1799 –1806 18. Lin M, Yang Y, Chiany H, et al: Reappraisal of 32. Rocker GM, Mackensie MG, Williams B, et al:
5. Plant PK, Owen JL, Elliott MW: Early use of continuous positive airway pressure therapy Noninvasive positive pressure ventilation.
non-invasive ventilation for acute exacerba- in acute cardiogenic pulmonary edema: Successful outcome in patients with acute
tions of chronic obstructive pulmonary dis- Short-term results and long-term follow-up. lung injury/ARDS. Chest 1999; 115:173–177
ease on general respiratory wards: A multi- Chest 1995; 107:1379 –1386 33. Rana S, Hussam J, Gay P, et al: Failure of
centre randomised controlled trial. Lancet 19. Nava S, Carbone G, DiBattista N, et al: Non- non-invasive ventilation in patients with
2000; 355:1931–1935 invasive ventilation in cardiogenic pulmo- acute lung injury: Observational cohort
6. Lightowler J, Wedjicha JA, Elliott MW, et al: nary edema: A multicenter randomized trial. study. Crit Care 2006; 10:R79
Non-invasive positive pressure ventilation to Am J Respir Crit Care Med 2003; 168: 34. Antonelli M, Conti G, Esquinas A, et al: A
treat respiratory failure resulting from exac- 1432–1437 multiple-center survey on the use in clinical
erbations of chronic obstructive pulmonary 20. Crane SD, Elliott MW, Gilligan P, et al: Ran- practice of noninvasive ventilation as a first-
disease: Cochrane systematic review and domised controlled comparison of continu- line intervention for acute respiratory dis-
meta-analysis. BMJ 2003; 326:185–187 ous positive airways pressure, bilevel non- tress syndrome. Crit Care Med 2007; 35:
7. Keenan SP, Sinuff T, Cook DJ, et al: Which invasive ventilation, and standard treatment 18 –25
patients with acute exacerbation of chronic in emergency department patients with 35. Antonelli M, Conti G, Bufi M, et al: Nonin-
obstructive pulmonary disease benefit from acute cardiogenic pulmonary oedema. vasive ventilation for treatment of acute re-
noninvasive positive pressure ventilation? A Emerg Med J 2004; 21:155–161 spiratory failure in patients undergoing solid
systematic review of the literature. Ann In- 21. Pang D, Keenan SP, Cook DJ, et al: The effect organ transplantation: A randomized trial.
tern Med 2003; 138:861– 870 of positive airway pressure on mortality and JAMA 2000; 283:235–241
8. Gonzalez Diaz G, Carillo A, Perez P, et al: the need for intubation in cardiogenic pul- 36. Hilbert G, Gruson D, Vargas F, et al: Nonin-
Noninvasive positive-pressure ventilation to monary edema. Chest 1998; 114:1185–1192 vasive ventilation in immunosuppressed pa-
treat hypercapnic coma secondary to respi- 22. Rasanen J, Heikkila J, Downs J, et al: Con- tients with pulmonary infiltrates and acute
ratory failure. Chest 2005; 127:952–960 tinuous positive airway pressure by face respiratory failure. N Engl J Med 2001; 344:
9. Scala R, Naldi M, Archinucci I, et al: Nonin- mask in acute cardiogenic pulmonary 481– 487
vasive positive-pressure ventilation in pa- edema. Am J Cardiol 1985; 55:296 –300 37. Confalonieri M, Calderini E, Terraciano S, et
tients with acute exacerbations of COPD and 23. Lin M, Chiang HT: The efficacy of early con- al: Noninvasive ventilation for treating acute
varying levels of consciousness. Chest 2005; tinuous positive airway pressure therapy in respiratory failure in AIDS patients with
128:1657–1666 patients with acute cardiogenic pulmonary Pneumocystis carinii pneumonia. Intensive
10. Conti G, Antonelli M, Navalesi P, et al: Non- edema. J Formos Med Assoc 1991; 90: Care Med 2002; 28:1233–1238
invasive vs. conventional mechanical ventila- 736 –743 38. Nourdine K, Combes P, Carton MJ, et al:
tion in patients with chronic obstructive pul- 24. Masip J, Roque M, Sanchez B, et al: Nonin- Does noninvasive ventilation reduce the ICU
monary disease after failure of medical vasive ventilation in acute cardiogenic pul- nosocomial infection risk? A prospective
treatment in the ward: A randomized trial. monary edema. JAMA 2005; 294:3124 –3130 clinical survey. Intensive Care Med 1999; 25:
Intensive Care Med 2002; 28:1701–1707 25. Winck JC, Azevedo LF, Costa-Pereira A, et al: 567–573
11. Squadrone E, Frigerio P, Fogliati C, et al: Efficacy and safety of non-invasive ventila- 39. Squadrone V, Coha M, Cerutti E, et al: Con-
Noninvasive vs. invasive ventilation in COPD tion in the treatment of acute cardiogenic tinuous positive airway pressure for treat-
patients with severe acute respiratory failure pulmonary edema—a systematic review and ment of postoperative hypoxemia: A random-
deemed to require ventilatory assistance. In- meta-analysis. Crit Care 2006; 10:R69 ized controlled trial. JAMA 2005; 293:
tensive Care Med 2004; 30:1303–1310 26. Mehta S, Jay GD, Woolard RH, et al: Ran- 589 –595
12. Meduri GU, Cook TR, Turner RE, et al: Non- domized, prospective trial of bilevel versus 40. Kindgen-Milles D, Müller E, Buhl R, et al:

2406 Crit Care Med 2007 Vol. 35, No. 10

 Nasal continuous positive airway pressure re- 48. Curtis RJ, Cook DJ, Sinuff T, et al: Noninva- tilation in treating COPD patients over 8
duces pulmonary morbidity and length of sive positive pressure ventilation in critical years. Intensive Care Med 2003; 29:419 – 425
stay following thoracoabdominal aortic sur- and palliative care settings: Understanding 55. Navalesi P, Fanfulla F, Frigerio P, et al: Phys-
gery. Chest 2005; 128:821– 828 the goals of therapy. Crit Care Med 2007; iologic evaluation of noninvasive mechanical
41. Auriant I, Jallot A, Herve P, et al: Noninvasive 35:932–939 ventilation delivered with three types of
ventilation reduces mortality in acute respi- 49. Baillard C, Fosse JP, Sebbane M, et al: Non- masks in patients with chronic hypercapnic
ratory failure following lung resection. Am J invasive ventilation improves preoxygenation respiratory failure. Crit Care Med 2000; 28:
Respir Crit Care Med 2001; 164:1231–1235 before intubation in hypoxic patients. Am J 1785–1790
42. Keenan SP, Powers C, McCormack DG, et al: Respir Crit Care Med 2006; 174:171–177 56. Kwok H, McCormack J, Cece R, et al: Con-
Noninvasive positive-pressure ventilation for 50. Maitre B, Jaber S, Maggiore SM, et al: Con- trolled trial of oronasal versus nasal mask
postextubation respiratory distress: A ran- tinuous positive airway pressure during fi- ventilation in the treatment of acute respi-
domized controlled trial. JAMA 2002; 287: beroptic bronchoscopy in hypoxemic pa- ratory failure. Crit Care Med 2003; 31:
3238 –3244 tients. A randomized double-blind study 468 – 473
43. Esteban A, Frutos-Vivar F, Ferguson ND, et al: using a new device. Am J Respir Crit Care 57. Appendini L, Patessio A, Zanaboni S, et al:
Noninvasive positive-pressure ventilation for
Med 2000; 162:1063–1067 Physiologic effects of positive end– expiratory
respiratory failure after extubation. N Engl
51. Antonelli M, Conti G, Rocco M, et al: Nonin- pressure and mask pressure support during
J Med 2004; 350:2452–2460
vasive positive-pressure ventilation vs. con- exacerbations of chronic obstructive pulmo-
44. Nava S, Gregoretti C, Fanfulla F, et al: Non-
ventional oxygen supplementation in hypox- nary disease. Am J Respir Crit Care Med
invasive ventilation to prevent respiratory
emic patients undergoing diagnostic 1994; 149:1069 –1076
failure after extubation in high-risk patients.
bronchoscopy. Chest 2002; 121:1149 –1154 58. L’Her E, Deye N, Lellouche F, et al: Physio-
Crit Care Med 2005; 33:2465–2470
45. Ferrer M, Valencia M, Nicolas JM, et al: Early 52. Da Conceicao M, Genco G, Favier JC, et al: logic effects of noninvasive ventilation dur-
noninvasive ventilation averts extubation Fiberoptic bronchoscopy during noninvasive ing acute lung injury. Am J Respir Crit Care
failure in patients at high risk: A randomized positive-pressure ventilation in patients with Med 2005; 172:1112–1118
trial. Am J Respir Crit Care Med 2006; 173: chronic obstructive lung disease with hypox- 59. Plant PK, Owen JL, Elliott MW: Early use of
164 –170 emia and hypercapnea. Ann Fr Anesth Re- non-invasive ventilation for acute exacerba-
46. Levy MM, Tanios MA, Nelson D, et al: Out- anim 2000; 19:231–236 tions of chronic obstructive pulmonary dis-
comes of patients with do-not-intubate or- 53. Confalonieri M, Garuti G, Cattaruzza MS, et ease on general respiratory wards: A multi-
ders treated with noninvasive ventilation. al: Italian noninvasive positive pressure ven- centre randomised controlled trial. Lancet
Crit Care Med 2004; 32:2002–2007 tilation (NPPV) study group. A chart of fail- 2000; 355:1931–1935
47. Schettino G, Altobelli N, Kacmarek RM: Non- ure risk for noninvasive ventilation in pa- 60. Farha S, Ghamra ZW, Hoisington ER, et al:
invasive positive-pressure ventilation re- tients with COPD exacerbation. Eur Respir J Use of noninvasive positive-pressure ventila-
verses acute respiratory failure in select “do- 2005; 25:348 –355 tion on the regular hospital ward: Experience
not-intubate” patients. Crit Care Med 2005; 54. Carlucci A, Delmastro M, Rubini F, et al: and correlates of success. Respir Care 2006;
33:1976 –1982 Changes in the practice of non-invasive ven- 51:1237–1243

Crit Care Med 2007 Vol. 35, No. 10 2407