Академический Документы
Профессиональный Документы
Культура Документы
S2
Volume 71 Supplement 2 Pages ii1ii35
Thorax
AN INTERNATIONAL JOURNAL OF RESPIRATORY MEDICINE
thorax.bmj.com
Appendix 1 Guideline Group Members Figures Receive regular table of contents by email.
Figure 1 Summary for providing acute non-invasive Register using this QR code.
Tables ventilation
Table 1 SIGN grades of recommendations Figure 2 Guide to initial settings and aims with
Table 2 SIGN levels of evidence invasive mechanical ventilation
Table 3 Technical issues: a guide for when NIV is Figure 3 The three phases of patient management in
failing acute hypercapnic respiratory failure.
Boxes Web appendices
Box 1 Indications for IMV in AECOPD Web appendix 1 Clinical questions and literature
Box 2 Risk Factors for Extubation Failure following search strategy
IMV Web appendix 2 Evidence tables
Box 3 Essential requirements for a NIV service Web appendix 3 Slide set available on BTS website
Box 4 Discharge checklist after an episode of
AHRF
Additional material is SUMMARY OF RECOMMENDATIONS ECG monitoring is advised if the patient has a
published online only. To view Principles of mechanical ventilation pulse rate >120 bpm or if there is dysrhythmia
please visit the journal online
(http://dx.doi.org/10.1136/ Modes of mechanical ventilation or possible cardiomyopathy.
thoraxjnl-2015-208209). Recommendation
1. Pressure-targeted ventilators are the devices of
choice for acute NIV (Grade B). Supplemental oxygen therapy with NIV
For numbered afliations see
Good practice points Recommendations
end of article.
Both pressure support (PS) and pressure control 4. Oxygen enrichment should be adjusted to
Correspondence to modes are effective. achieve SaO2 8892% in all causes of acute hyper-
Dr A C Davidson, Only ventilators designed specically to deliver capnic respiratory failure (AHRF) treated by NIV
BTS, 17 Doughty Street, NIV should be used.
London WC1N 2PL, UK; (Grade A).
craigdavidson@doctors.org.uk 5. Oxygen should be entrained as close to the
Choice of interface for NIV patient as possible (Grade C).
Healthcare providers need to Good practice points
use clinical judgement,
Recommendation
knowledge and expertise when 2. A full face mask (FFM) should usually be the As gas exchange will improve with increased
deciding whether it is rst type of interface used (Grade D). alveolar ventilation, NIV settings should be opti-
appropriate to apply Good practice points mised before increasing the FiO2.
recommendations for the A range of masks and sizes is required and staff The ow rate of supplemental oxygen may need to
management of patients. The
recommendations cited here involved in delivering NIV need training in and be increased when ventilatory pressure is increased
are a guide and may not be experience of using them. to maintain the same SaO2 target.
appropriate for use in all NIV circuits must allow adequate clearance of Mask leak and delayed triggering may be caused
situations. The guidance exhaled air through an exhalation valve or an by oxygen ow rates >4 L/min, which risks pro-
provided does not override the moting or exacerbating patient-ventilator asyn-
responsibility of healthcare
integral exhalation port on the mask.
professionals to make decisions chrony. The requirement for high ow rates
appropriate to the should prompt a careful check for patient-
Indications for and contra-indications to NIV in
circumstances of each patient, ventilator asynchrony.
in consultation with the patient AHRF
A ventilator with an integral oxygen blender is
and/or their guardian or carer. Recommendation
recommended if oxygen at 4 L/min fails to
3. The presence of adverse features increase the
Received 17 December 2015 maintain SaO2 >88%.
risk of NIV failure and should prompt consider-
Accepted 10 January 2016
ation of placement in high dependency unit
(HDU)/intensive care unit (ICU) (Grade C).
Humidication with NIV
Good practice points
Recommendation
Adverse features should not, on their own, lead
6. Humidication is not routinely required
to withholding a trial of NIV.
(Grade D).
The presence of relative contra-indications
Good practice point
necessitates a higher level of supervision, consid-
Heated humidication should be considered if the
eration of placement in HDU/ICU and an early
patient reports mucosal dryness or if respiratory
http://dx.doi.org/10.1136/ appraisal of whether to continue NIV or to
thoraxjnl-2016-208281
secretions are thick and tenacious.
convert to invasive mechanical ventilation
(IMV).
Bronchodilator therapy with NIV
Monitoring during NIV Good practice points
Good practice points Nebulised drugs should normally be adminis-
Oxygen saturation should be continuously tered during breaks from NIV.
To cite: Davidson AC, monitored. If the patient is dependent on NIV, bronchodila-
Banham S, Elliott M, et al. Intermittent measurement of pCO2 and pH is tor drugs can be given via a nebuliser inserted
Thorax 2016;71:ii1ii35. required. into the ventilator tubing.
Davidson AC, et al. Thorax 2016;71:ii1ii35. doi:10.1136/thoraxjnl-2015-208209 ii1
BTS guidelines
26. Severe acidosis alone does not preclude a trial of NIV in an severe facial deformity, xed upper airway obstruction, facial
appropriate area with ready access to staff who can perform safe burns (Grade D).
endotracheal intubation (Grade B). 35. Intubation is indicated where risk/benet analysis by an
27. The use of NIV should not delay escalation to IMV when experienced clinician favours a better outcome with IMV than
this is more appropriate (Grade C). with NIV (Grade D).
28. The practice of NIV should be regularly audited to maintain
standards (Grade C). Outcome following NIV or IMV in AECOPD
Recommendations
Starting NIV in COPD 36. Prognostic tools may be helpful to inform discussion regard-
Good practice points ing prognosis and with regard to the appropriateness of IMV
Arterial blood gas (ABG) measurement is needed prior to but with the caveat that such tools are poorly predictive for
and following starting NIV. individual patient use (Grade B).
Chest radiography is recommended but should not delay ini- 37. Clinicians should be aware that they are likely to underesti-
tiation of NIV in severe acidosis. mate survival in AECOPD treated by IMV (Grade B).
Reversible causes for respiratory failure should be sought and 38. Clinicians should discuss management of possible future epi-
treated appropriately. sodes of AHRF with patients, following an epsiode requiring
At the start of treatment, an individualised patient plan ventilatory support, because there is a high risk of recurrence
(involving the patient wherever possible) should document (Grade B).
agreed measures to be taken in the event of NIV failure.
Acute asthma
Prognostic features relating to use of NIV in COPD Recommendations
Recommendations 39. NIV should not be used in patients with acute asthma
29. Advanced age alone should not preclude a trial of NIV exacerbations and AHRF (Grade C).
(Grade A). 40. Acute (or acute on chronic) episodes of hypercapnia may
30. Worsening physiological parameters, particularly pH and complicate chronic asthma. This condition closely resembles
respiratory rate (RR), indicate the need to change the manage- COPD and should be managed as such (Grade D).
ment strategy. This includes clinical review, change of interface,
adjustment of ventilator settings and considering proceeding to Non-CF bronchiectasis
endotracheal intubation (Grade A). Recommendations
Good practice point 41. In patients with non-CF bronchiectasis and AHRF, con-
If sleep-disordered breathing pre-dates AHRF, or evidence of it trolled oxygen therapy should be used. (Grade D)
complicates an episode, the use of a controlled mode of NIV 42. In patients with non-CF bronchiectasis, NIV should be started
overnight is recommended. in AHRF using the same criteria as in AECOPD (Grade B).
43. In patients with non-CF bronchiectasis, NIV should usually
Duration of NIV in COPD be tried before resorting to IMV in those with less severe
Recommendation physiological disturbance (Grade C).
31. NIV can be discontinued when there has been normalisation 44. In non-CF bronchiectasis, the patients clinical condition
of pH and pCO2 and a general improvement in the patients prior to the episode of AHRF, and the reason for the acute
condition (Grade B). deterioration, should be evaluated and used to inform the
Good practice points decision about providing IMV (Grade C).
Time on NIV should be maximised in the rst 24 h depend- Good practice points
ing on patient tolerance and/or complications. In patients with non-CF bronchiectasis, the precipitating
NIV use during the day can be tapered in the following cause is important in determining short-term prognosis.
23 days, depending on pCO2 self-ventilating, before being Health status prior to the episode of AHRF is an important
discontinued overnight. predictor of outcome.
50. NIV should almost always be trialled in the acutely unwell PEEP in the range of 510 is commonly required to increase
patients with NMD or CWD with hypercapnia. Do not wait for residual volume and reduce oxygen dependency in both
acidosis to develop (Grade D). patient groups.
51. In patients with NMD or CWD, NIV should be considered
in acute illness when vital capacity (VC) is known to be <1 L Obesity hypoventilation syndrome
and RR >20, even if normocapnic (Grade D). Recommendations
52. In patients with NMD or CWD, consider controlled ventila- 56. Controlled oxygen therapy should be used in patients with
tion as triggering may be ineffective (Grade D). OHS and AHRF (Grade D).
53. In NMD or CWD, unless escalation to IMV is not desired 57. In patients with OHS, NIV should be started in AHRF using
by the patient, or is deemed to be inappropriate, intubation the same criteria as in AECOPD (Grade B).
should not be delayed if NIV is failing (Grade D). 58. NIV is indicated in some hospitalised obese hypercapnic
Good practice points patients with daytime somnolence, sleep disordered breathing
Individuals with NMD and CWD who present with AHRF and/or right heart failure in the absence of acidosis (Grade D).
should not be denied acute NIV.
NIV is the ventilation mode of choice because patients with NIV settings and placement in OHS
NMD or CWD tolerate it well and because extubation from Good practice points
IMV may be difcult. High inspiratory positive airway pressure (IPAP) and expira-
In patients with NMD or CWD, deterioration may be rapid tory positive airway pressure (EPAP) settings are commonly
or sudden, making HDU/ICU placement for therapy more required in patients with OHS (eg, IPAP>30, EPAP>8).
appropriate. Volume control (or volume assured) modes of providing NIV
In patients with NMD or CWD, senior/experienced input is may be more effective when high ination pressures are
needed in care planning and is essential if differences in required.
opinion exist or develop between medical staff and patient
representatives. NIV failure in OHS
In patients with NMD, it should be anticipated that bulbar Good practice points
dysfunction and communication difculties, if present, will Fluid overload commonly contributes to ventilatory failure in
make NIV delivery difcult, and may make it impossible. patients with OHS, and its degree is easily underestimated.
Discussion about NIV and IMV, and patients wishes with Forced diuresis may be useful.
respect to cardiopulmonary resuscitation, should occur as As the risk of NIV failure is greater, and intubation may be
part of routine care of patients with NMD or CWD. more difcult, placement in HDU/ICU for NIV is
In patients with NMD or CWD, nocturnal NIV should recommended.
usually be continued following an episode of AHRF, pending
discussion with a home ventilation service. Discontinuing NIV in OHS
Good practice points
NIV failure and discontinuing NIV following recovery in NMD and NIV can be discontinued, as in patients with AECOPD.
CWD Many patients with AHRF secondary to OHS will require
Good practice points long-term domiciliary support (CPAP or NIV).
In patients with NMD or CWD, intolerance of the mask and Following an episode of AHRF referral to a home ventilation
severe dyspnoea are less likely to cause NIV failure. Bulbar service is recommended.
dysfunction makes NIV failure more likely.
Deterioration in patients with NMD or CWD may be very IMV strategy in OHS
sudden. Difculty achieving adequate oxygenation or rapid Good practice points
desaturation during a break from NIV are important warning In patients with OHS, pressure controlled MV is recom-
signs. mended initially.
In patients with NMD or CWD, the presence of bulbar dys- In patients with OHS, high PEEP settings may be needed to
function, more profound hypoxaemia or rapid desaturation recruit collapsed lung units and correct hypoxaemia.
during NIV breaks, suggests that placement in HDU/ICU is In patients with OHS, a forced diuresis is often indicated.
indicated.
IMV in NMD/CWD Weaning from IMV
Recommendations Introduction
54. In patients with NMD or CWD, senior staff should be Recommendations
involved in decision-making, in conjunction with home mechan- 59. Treating the precipitant cause of AHRF, normalising pH,
ical ventilation specialists, if experience is limited, and especially correcting chronic hypercapnia and addressing uid overload
when the appropriateness of IMV is questioned (Grade D). should all occur before weaning is started (Grade D).
55. Advance care planning, particularly around the potential 60. A brain natriuretic peptide (BNP)-directed uid manage-
future use of IMV, is recommended in patients with progressive ment strategy should be considered in patients with known left
NMD or CWD. This may best be supported by elective referral ventricular dysfunction. (Grade B)
to a home ventilation service (Grade D).
Weaning methods
IMV strategy in NMD and CWD Recommendations
Good practice points 61. Assessment of the readiness for weaning should be under-
Patients with NMD usually require low levels of PS. taken daily (Grade C).
Patients with chest wall deformity usually require higher 62. A switch from controlled to assisted IMV should be made as
levels of PS. soon as patient recovery allows (Grade C).
ii4 Davidson AC, et al. Thorax 2016;71:ii1ii35. doi:10.1136/thoraxjnl-2015-208209
BTS guidelines
63. IMV patients should have a documented weaning plan Good practice points
(Grade B). A 24 bedded designated NIV unit (located within a medical
high dependency area or within a respiratory ward with
Assessing readiness for discontinuation of mechanical ventilation enhanced stafng levels) provides a sound basis for the provi-
Recommendation sion of NIV in a DGH serving a population of 250 000 and
64. A 30 min spontaneous breathing trial (SBT) should be used with an average prevalence of COPD.
to assess suitability for extubation (Grade B). Areas providing NIV should have a process for audit and
65. Factors including upper airway patency, bulbar function, interdisciplinary communication.
sputum load and cough effectiveness should be considered prior
to attempted extubation (Grade D). Palliative care and advanced care planning
Recommendations
Outcome following extubation 78. Clinicians delivering NIV or IMV should have ready access
Recommendation to palliative medicine (Grade D).
66. Care is needed to identify factors that increase the risk of 79. Multidisciplinary advance care planning should be an
extubation failure so that additional support, such as NIV or integral part of the routine outpatient management of progres-
cough assist, can be provided (Grade B). sive or advanced disease and care plans should be reviewed on
presentation during an episode of AHRF (Grade D).
Weaning protocols 80. The use of NIV may allow time to establish patient prefer-
Recommendations ence with regard to escalation to IMV. (Grade D)
67. Although an organised and systematic approach to weaning
is desirable, protocols should be used with caution in patients End of life care
with AHRF (Grade B). Good practice points
68. The use of computerised weaning cannot be recommended Although removal of the NIV mask may be agreed as prefer-
in AHRF (Grade D). able, a dignied and comfortable death is possible with it in
place.
Use of NIV in the ICU Clinicians delivering NIV or IMV should have training in
Planned NIV to speed weaning from IMV end-of-life care and the support of palliative care teams.
Recommendation
69. NIV is recommended to aid weaning from IMV in patients Novel therapies
with AHRF secondary to COPD (Grade B). Extracorporeal CO2 removal (ECCO2R)
70. In other causes of AHRF, NIV may have a role in shortening Recommendations
the duration of IMV when local expertise in its use exists (and of 81. If local expertise exists, ECCO2R might be considered:
cough assist when indicated) and there are features present that If, despite attempts to optimise IMV using lung protective
indicate extubation is likely to be successful (Grade D). strategies, severe hypercapnic acidosis ( pH<7.15) persists
(Grade D);
When lung protective ventilation is needed but hypercapnia
NIV in high-risk patients is contraindicated, for example, in patients with coexistent
Recommendation brain injury (Grade D);
71. Prophylactic use of NIV should be considered to provide For IMV patients awaiting a lung transplant (Grade D).
post-extubation support in patients with identied risk factors Good practice point
for extubation failure (Grade B). ECCO2R is an experimental therapy and should only be used
by specialist intensive care teams trained in its use, and where
NIV as rescue therapy post-extubation additional governance arrangements are in place, or in the
Recommendations setting of a research trial.
72. NIV should not be used routinely for unexpected post-
extubation respiratory failure (Grade B). Helium/oxygen ventilation
73. In COPD, a trial of NIV may be justied for unexpected Recommendation
post-extubation respiratory failure where local expertise exists 82. Heliox should not be used routinely in the management of
(Grade D). AHRF (Grade B).
increased likelihood of future life-threatening attacks.8 The Ovid EMBASE, EMSCO CINAHL, Ovid PsycINFO and the
same risks apply to AHRF complicating CF and bronchiectasis, Cochrane Library (including the Cochrane Database of
although this has not been formally reported. In the neuromus- Systematic Reviews, the Database of Abstracts of Reviews of
cular and CWDs, including morbid obesity, respiratory pump Effects and the Cochrane Central Register of Controlled Trials).
failure is often insidious in its onset, but AHRF may be acute An initial search was carried out in November 2010, using a
and unexpected. Acute on chronic decompensated episodes of combination of indexed and free text terms dening the clinical
AHRF are more common and normally indicate the future need questions that had been agreed as important in formulating
for domiciliary NIV. guidelines in AHRF. It was limited to studies after 1990, on
adults, in journals published in English and where at least an
Intended use and target audience of the guideline abstract was available. The searches identied a total of 582
A central theme of the guideline is to promote integration in the potential papers, which were subsequently supplemented by
planning and delivery of NIV and IMV in AHRF. Despite evi- publications known to members or resulting from additional
dence demonstrating the value of non-invasive ventilation in the searches undertaken by the writing groups after 2010. The lit-
management of AHRF, its introduction into routine clinical erature search was run again in September 2013, for relevant
practice in the UK has not delivered the expected patient benet publications between 2010 and 2013, yielding a further 308
and it is likely that NIV provision has, inadvertently, reduced potentially relevant references. Additional references were sub-
access to IMV in AECOPD and the other causes of AHRF. The sequently included from personal collections.
introduction, in hospitals accepting acute admissions, of an
adequately resourced and integrated AHRF patient pathway is Appraisal of the literature
strongly recommended in the expectation that this will lead to Appraisal was performed using the criteria stipulated by the
improved clinical outcomes and patient experiences. AGREE collaboration. Each paper was appraised by at least two
The target audience for the guideline is medical, nursing and reviewers. The writing lead for each section read the title and
physiotherapy staff working in emergency receiving rooms, abstract of papers identied and agreed with at least one
medical assessment units, admission wards, respiratory wards member of each writing group on whether such a paper was
and in high dependency and critical care units. The guideline denitely relevant, possibly relevant or not relevant, to the
applies to adults. For information on NIV in children with section. The criteria used were that the paper addressed a clin-
neuromuscular weakness, see the BTS guideline Respiratory ical question, the study method used was satisfactory and that
Management of Children with Neuromuscular Weakness.9 the paper was available in English.
Full papers were obtained for all relevant or possibly relevant
Areas not covered by the guideline abstracts. Two members for each section independently appraised
The guideline does not cover the management of AHRF due to each paper, using the SIGN critical appraisal checklists. An evi-
cardiac failure, trauma or acute brain injury. The guideline refers dence level was assigned to each study using SIGN methodology
to domiciliary NIV but does not aim to provide guidance on (table 2). These evidence levels are shown in the evidence tables
this. The use of non-invasive ventilation is more extensively presented in the online supplementary appendix 3.
covered than IMV because the evidence and the clinical experi-
ence in its use is recent and because the technical aspects con- Considered judgement and grading of recommendations
cerning IMV are well covered by standard texts. The guideline group used the evidence tables to judge the body
of evidence and to develop recommendations for this guideline.
Units Where evidence was lacking, expert opinions were obtained by
Intrathoracic pressure and pressures relating to mechanical consensus. The following were considered in the grading of the
ventilation are presented as cm H2O. ABG tensions are recommendations: the number of studies and number of
presented as kPa. patients providing evidence, the applicability of such evidence,
and whether generalisable to the patient groups in the guideline
Guideline group members and to UK practice and the degree of strength as judged by the
A list of Guideline Group members and BTS Standards of Care consistency of evidence obtained to support recommendations.
Committee members who assisted with the production of the
guideline is given in appendix 1.
The Guideline Group members adhered to the BTS and ICS Table 2 SIGN levels of evidence
policies for the Declaration of Interests and, where appropriate,
1++ High-quality meta-analyses, systematic reviews of RCTs, or RCTs with a
specic relevant interests are declared in appendix 1. very low risk of bias
1+ Well-conducted meta-analyses, systematic reviews or RCTs with a low
Methods and terminology risk of bias
The guideline has been produced according to the BTS 1 Meta-analyses, systematic reviews or RCTs with a high risk of bias
Guideline Production manual and adheres to the criteria set out 2++ High-quality systematic reviews of case control or cohort or studies
in the AGREE II instrument.10 11 High-quality casecontrol or cohort studies with a very low risk of
confounding or bias and a high probability that the relationship is causal
Clinical questions and literature search 2+ Well-conducted casecontrol or cohort studies with a low risk of
confounding or bias and a moderate probability that the relationship is
Clinical questions were gathered in the PICOT (Patient,
causal
Intervention, Comparison, Outcome and Time) format to dene
2 Casecontrol or cohort studies with a high risk of confounding or bias
the scope of the guideline and inform the literature search. and a significant risk that the relationship is not causal
Systematic electronic database searches were conducted in order 3 Non-analytic studies, eg, case reports, case series
to identify potentially relevant studies for inclusion in the guide- 4 Expert opinion
line. For each clinical question, the following databases were
RCT, randomised controlled trial.
searched: Ovid MEDLINE (including MEDLINE In-Process),
Davidson AC, et al. Thorax 2016;71:ii1ii35. doi:10.1136/thoraxjnl-2015-208209 ii7
BTS guidelines
Recommendations were graded from A to D, using the SIGN inspiratory effort. Pressure-controlled ventilation (PCV) is the
Grading System (table 2), as indicated by the strength of the evi- term used to describe a mode in which the operator sets the
dence as listed in the tables. Important practical points that lack inspiratory pressure, the length of inspiration and the inspiratory
research evidence were highlighted as Good Practice Points. rate. In the spontaneous/timed (S/T) mode (also known as assist
Good practice points control), a backup rate is set by the operator. If the patients RR is
Recommended best practice based on the clinical experience of slower than the backup rate, machine-determined breaths will be
the guideline development group. delivered (ie, controlled ventilation). If the patient breathes faster
than the backup rate, no machine determined breaths will be
Drafting the guideline delivered and all breaths will be triggered (or assisted). The pro-
The Guideline Group corresponded regularly. The initial portion of controlled and assisted breaths often varies, depending
meeting took place in October 2009, and subsequent meetings on the patients state of alertness and respiratory drive.
of the full committee occurred in June and November 2010, Trigger sensitivity refers to the effort required by the patient
September 2011, and March and September 2012. Draft docu- to initiate, or trigger, the ventilator. The lower the trigger sensi-
ments were reviewed by the BTS Standards of Care Committee tivity, the greater effort the patient needs to make to trigger a
at meetings in 2013 and 2014, and a nal draft was produced supported breath. Different trigger settings may be required for
with the help and collaboration of members of the BTS individual causes of AHRF (see Management of hypercapnic
Standards of Care Committee in September 2014 to March respiratory failure section).
2015. The guideline was made available for public consultation S/T is the NIV mode most commonly employed in treating
on the BTS website from 7 May to 12 June 2015. The revised AHRF. There have been no trials comparing PS ventilation and
document was reviewed by the BTS Standards of Care PCV in the treatment of AHRF. Bench studies suggest that venti-
Committee in September 2015 and nal approval for publica- lators designed specically for NIV have superior performance
tion was given in November 2015. over standard ICU ventilators used to deliver NIV, particularly
in the presence of signicant leak.1722 The extent to which
PRINCIPLES OF MECHANICAL VENTILATION individual types of ICU ventilators (set in the NIV mode) can
Modes of mechanical ventilation compensate for leak and the adequacy of patient triggering
There are two basic modes of providing mechanical ventilation. varies.23 Generally, ICU ventilators appear more prone to
In volume-targeted, the operator sets the tidal volume to be patient-ventilator asynchrony than home care ventilators.24
delivered and the duration of inspiration (Ti). The ventilator Evidence statement
generates whatever pressure is necessary to deliver this volume Most RCTs that demonstrate an advantage to NIV in AHRF
within this time. In pressure-targeted, the operator sets the have used pressure targeted ventilators (Level 1+).
inspiratory pressure. The volume of air the patient receives is a Recommendation
function of the impedance to ination of the lungs and chest 1. Pressure targeted ventilators are the devices of choice for
wall and the inspiratory time. The Ti should be of sufcient acute NIV (Grade B).
length to achieve an adequate volume and at a frequency that Good practice points
allows the patient time to fully exhale. Both PS and pressure control modes are effective.
The terminology used for pressure-targeted ventilation can Only ventilators designed specically to deliver NIV should
cause confusion. In bi-level ventilation, one pressure is set for be used.
inspiration (IPAP) and a second pressure for expiration (EPAP).
The difference between the two is the level of ventilatory assist- Choice of interface for NIV
ance or PS. This mode is most commonly used for NIV. The The FFM is the most suitable interface, as mouth breathing pre-
same term, CPAP with Pressure Support, can be used to dominates in AHRF. To accommodate the natural diversity of
describe a mode of invasive ventilation and/or non-invasive ven- the human face, a range of shapes and sizes of FFM should be
tilation on some ICU ventilators. The operator sets an incremen- available. Reported studies suggest that different types of inter-
tal inspiratory pressure above the CPAP setting rather than faces do not affect outcome, but the trials have been small and
setting an absolute level of inspiratory pressure. comparison of masks has been inadequately powered to detect a
Pressure-targeted ventilation has a number of advantages. difference.2537 The helmet interface, which covers the whole
First, the pressure delivered is constant and this avoids the head, is an alternative to an FFM,3844 but triggering is ineffect-
sudden and uncomfortable pressure increase that occurs with ive. Patients may report about noise caused by turbulence within
volume control. Second, pressure-targeted ventilation compen- the helmet,45 and it is not possible to provide humidied gases
sates for air leak,12 13 which is an inevitable consequence of the because of rain out in the helmet. A mask that covers the
interfaces used for NIV. Third, positive pressure throughout whole of the face (including the eyes, but not the ears) is useful
expiration (EPAP) ushes exhaled CO2 from the mask and distal when air leak remains excessive or when nasal bridge ulceration
ventilator tubing,14 15 aids triggering (see below) and counter- develops,46 and is sometimes better tolerated by the confused or
acts the tendency for upper airway collapse during expiration. agitated patient. In those who nd the FFM claustrophobic or
Pressure ventilators have been used in almost all of the rando- distressing, experienced practitioners may consider using a nasal
mised controlled trials (RCTs) in AHRF.16 In the UK, volume mask or nasal pillows. Mouth leak limits the effectiveness of
ventilators are rarely employed (outside of specialist centres) nasal interfaces during sleep and nasal pillows are more easily
and will not be considered further in this guideline. dislodged than the FFM.
In the Spontaneous (S) mode (also known as assist mode), the Ventilators designed for NIV usually employ a single lumen
ventilator delivers assisted breaths in response to patient inspira- circuit whereas IMV ventilators use a dual lumen circuit (separ-
tory effort. If the patient fails to make adequate inspiratory effort, ate tubing for inhalation and exhalation). In the former, a mask
no ventilator support is delivered. By contrast, in the timed (T) with an integral exhalation port is commonly used. If not, an
mode (also known as control mode), the ventilator delivers exhalation port needs to be inserted into the ventilator circuit
breaths at a rate set by the operator regardless of patient close to the mask. A minimum EPAP of 3 cm is required to vent
ii8 Davidson AC, et al. Thorax 2016;71:ii1ii35. doi:10.1136/thoraxjnl-2015-208209
BTS guidelines
exhaled air.14 47 The website http://ersbuyersguide.org/ offers impairment make NIV more difcult to apply but should not
information on NIV interfaces that are currently available. preclude its use.
Evidence statement There is less haemodynamic compromise with NIV than with
An FFM is the interface of choice for general/non-specialist use IMV, and hypotension should rarely preclude using NIV.
(Level 4). Signicant arrhythmia, especially if causing hypotension, may
Recommendation tip the balance towards preferring intubation as, in these cir-
2. An FFM should usually be the rst type of interface used cumstances, cardioversion may be indicated.
(Grade D). An acute pneumothorax should be drained before applying
Good practice points NIV. If it is too small to allow the safe placement of a chest
A range of masks and sizes is required, and staff involved in drain (or is suspected to be chronic) NIV may proceed with
delivering NIV need training in and experience of using careful monitoring. Using a lower ination pressure seems theor-
them. etically sensible but is without evidence. If the patient deterio-
NIV circuits must allow adequate clearance of exhaled air rates, NIV should be discontinuedin case it is contributing to
through an exhalation valve or an integral exhalation port on the development of a tension pneumothoraxand an urgent
the mask. chest radiograph obtained.
Vomiting has been considered a contra-indication. The key
Indications for and contra-indications to NIV in AHRF issue is whether the NIV mask can be rapidly removed, that is,
The indication for NIV will vary according to the underlying an assessment of whether the patient can signal the need to
cause, severity of illness and associated complicating factors. vomit. Marked abdominal distension may sometimes precipitate
Broad criteria can be used and are summarised in gure 1, and AHRF in individuals at risk, for example in COPD or morbid
further discussed in Management of hypercapnic respiratory obesity. Management should then address the underlying cause
failure section. Severe facial deformity, xed upper airway of abdominal distension and manage the risk of vomiting by
obstruction or facial burns, will occasionally make NIV impos- inserting a nasogastric tube. Similarly, in the at-risk patient,
sible. A number of other contra-indications have been suggested hypercapnic respiratory failure may complicate the later stages
(see gure 1).48 These have most often been employed as exclu- of pregnancy (eg, kyphoscoliosis or muscular dystrophy). NIV is
sion criteria in clinical trials rather than being denitively shown ideally suited to manage this complication. The need for NIV
to result in a worse outcome.16 Some of the criteria have been should be electively assessed (by nocturnal monitoring), but
challenged. For instance, coma has been regarded as an absolute mask ventilation can be initiated during delivery should respira-
contra-indication, because of its associated loss of airway protec- tory distress develop in an at-risk patient.
tion, but Diaz et al49 report similar outcomes with NIV in those The presence of copious secretions increases the risk of treat-
with a Glasgow Coma Score <8 as the outcomes found in more ment failure,50 but NIV may also improve the ability to clear
alert patients. Similarly, confusion, agitation and cognitive secretions and improve alveolar ventilation.51 52
Respiratory arrest or peri-arrest have been considered as abso- receiving NIV.62 This is usually easily achieved in AECOPD, but
lute contra-indications as NIV is intended to supplement spon- severe hypoxaemia may complicate AHRF in other causative
taneous breathing. However, as bag and mask ventilation (itself diseases such as CWD.
a form of NIV) is used as a prelude to intubation, a short trial As for the best method of supplying oxygen, Padkin and
of NIV by an experienced operator, can be justied while Kinnear,63 in a study of patients who were not acutely unwell,
paying special attention to the risk of glottic occlusion. reported no difference in inspired content whether delivered
In summary, the presence of adverse features is an indication for directly into the NIV mask or into the ventilator tubing close to
more intense monitoring and placement within HDU/ICU the mask. Introducing oxygen at the ventilator end of the tubing
rather than a contra-indication per se. was less effective. The mean FiO2 achieved was 31% at 1 L/min,
Evidence statement 37% at 2 L/min, 40% at 3 L/min and 44% at 4 L/min. Flow
There are few absolute contra-indications to a trial of NIV but rates >4 L/min provided minimal additional increase. Kaul64
some adverse features, especially when combined, require more found that the higher the inspiratory pressure, the less add-
caution and more intense monitoring (Level 4). itional benet resulted from higher ow rates (because higher
The presence of adverse features increases the risk of NIV pressures increase leak). High ow rates also resulted in delay
failure (Level 2++). triggering the ventilator. As this risks promoting patient ventila-
Recommendation tor asynchrony, technically advanced NIV ventilators that allow
3. The presence of adverse features increases the risk of NIV precise oxygen blending (and a higher FiO2 enrichment) are a
failure and should prompt consideration of placement in HDU/ safer and more appropriate alternative when hypoxaemia is
ICU (Grade C). severe.
Good practice points Evidence statements
Adverse features should not, on their own, lead to withhold- In AHRF-targeted oxygen therapy (SaO2 8892%) reduces mor-
ing a trial of NIV. tality (Level 1+).
The presence of relative contra-indications necessitates a When providing NIV, oxygen enrichment is best given at or
higher level of supervision, consideration of placement in near the mask (Level 3).
HDU/ICU and an early appraisal of whether to continue Recommendations
NIV or to convert to IMV. 4. Oxygen enrichment should adjusted to achieve SaO2 88
92% in all causes of AHRF being treated by NIV (Grade A).
Monitoring during NIV 5. Oxygen should be entrained as close to the patient as possible
Continuous monitoring of oxygen saturation is essential. (Grade C).
Repeated measurement of ABG tensions will be required and Good practice points
can be assessed by capillary sampling or intermittent arterial As gas exchange will improve with increased alveolar ventila-
puncture, noting that capillary sampling is less painful for the tion, NIV settings should be optimised before increasing the
patient.53 54 One advantage of HDU/ICU placement may be to FiO2.
allow the safe use of an indwelling arterial line for blood sam- The ow rate of supplemental oxygen may need to be
pling. Transcutaneous pCO2 (TcpCO2) monitoring is a com- increased when ventilatory pressure is increased to maintain
monly employed investigation in home ventilation units and the the same SaO2 target.
devices are increasingly being employed in hospitals. Mask leak and delayed triggering may be caused by oxygen
Small studies have reported on its use in acute respiratory ow rates >4 L/min, which risks promoting or exacerbating
acidosis.5557 A study by van Oppen et al58 reported on patientventilator asynchrony. The requirement for high ow
10 patients receiving acute NIV and demonstrated that TcpCO2 rates should prompt a careful check for patientventilator
monitoring is reliable over 12 h and provides an adequate asynchrony.
estimation of pH. Further studies are needed to assess the role A ventilator with an integral oxygen blender is recommended
of transcutaneous CO2 monitoring. if oxygen at 4 L/min fails to maintain SpO2 >88%.
ECG monitoring is advised for all patients with a tachycardia
>120 bpm, dysrhythmia or known cardiomyopathy. As in all Humidication with NIV
severely ill patients, serial vital signs (and National Early There is no evidence to guide the use of humidication in acute
Warning Scores, where implemented) should be recorded. NIV. Heated humidication may reduce upper airway resistance
Good practice points and increase comfort when leak is high.65 In short-term studies,
Oxygen saturation should be continuously monitored. heated humidication reduces upper airway dryness,66 67 which
Intermittent measurement of pCO2 and pH is required. might improve tolerance and aid secretion clearance, but this
ECG monitoring is advised if the patient has a pulse rate has not been proven. Humidication should only be considered
>120 bpm or if there is dysrhythmia or possible when upper airway dryness is a problem or secretions are dif-
cardiomyopathy. cult to expectorate.
Evidence statement
Supplemental oxygen therapy with NIV No evidence exists to guide humidication practice in acute
There are no trials to guide the use of oxygen enrichment. It is NIV (Level 4).
well recognised that hyperoxygenation is harmful in the self- Recommendation
ventilating patient with AHRF.5961 In the absence of harm 6. Humidication is not routinely required (Grade D).
from modest hypoxaemia, and to avoid confusion that might Good practice point
arise from having different target saturations in different condi- Heated humidication should be considered if the patient
tions, a saturation range of 8892% is recommended in all reports mucosal dryness or if respiratory secretions are thick
patients with AHRF either spontaneously breathing or when and tenacious.
Bronchodilator therapy with NIV pharmacological case for preferring an opiate to a benzodiazep-
As part of a PhD thesis, Kaul68 found that nebulised bronchodi- ine (because the latter promotes upper airway obstruction
lators given concomitantly with NIV in stable patients produced through inhibiting the pharangeal dilating muscles) but con-
less benet than when given while patients were breathing spon- cluded that studies to date have been too small, have used differ-
taneously. Brief discontinuation of NIV for the administration ent drugs and therapy regimes and employed a variety of
of bronchodilators appears to be safe.69 Acccordingly, broncho- outcome measures.78 Guidance on the use of sedation within
dilator therapy is probably better given during breaks in NIV. hospitals might be expected to improve patient safety when
This may also facilitate coughing and the clearing of respiratory implemented.79
secretions. If discontinuing NIV results in patient distress, it Evidence statements
should be continued and a nebuliser sited proximally in the Patient distress is common in AHRF and often made initially
circuit.70 worse by applying NIV (Level 4).
Good practice points There is inadequate evidence to guide the use of sedation/anxio-
Nebulised drugs should normally be administered during lysis in acute NIV. Their use in a critical care setting is reported
breaks from NIV. to improve outcome and reduce patient distress (Level 2).
If the patient is dependent on NIV, bronchodilator drugs can Recommendations
be given via a nebuliser inserted into the ventilator tubing. 7. Sedation should only be used with close monitoring
(Grade D).
8. Infused sedative/anxiolytic drugs should only be used in an
Sedation with NIV HDU or ICU setting (Grade D).
Patient agitation and distress are common in AHRF and may be 9. If intubation is not intended should NIV fail, then sedation/
made worse by the application of NIV before gas exchange has anxiolysis is indicated for symptom control in the distressed or
improved and the patient has sensed a reduction in the work of agitated patient (Grade D).
breathing. Despite this, sedatives/anxiolytics and/or opiates are Good practice point
infrequently used due to concern about depressing respiratory In the agitated/distressed and/or tachypnoeic individual on NIV,
drive. This is understandable if NIV is delivered in an inappro- intravenous morphine 2.55 mg ( benzodiazepine) may
priate environment that is unable to provide continuous moni- provide symptom relief and may improve tolerance of NIV.
toring and that does not have the ready availability of medical
staff to perform safe intubation if needed. On the contrary,
relieving patient distress is an important goal and might be NIV complications
expected to increase comfort and the success of NIV. In a 2007 The reported rate of complications varies widely. One review
survey of members of the critical care assemblies of the gives an incidence between 30% and 50%,80 the range partly
American College of Chest Physicians and the European depending on how a complication is dened. Extended duration
Respiratory Society, respondents reported using sedatives or of NIV, patient agitation and the frequent need to adjust mask
opiates in only 25% of cases and 21% stated they had never t are all associated with an increase in rate/severity of
used either.71 The risk of respiratory depression was given as mask-related problems.
the reason for non-use. Individual practice was highly variable Nasal bridge ulceration is the most common problem (510%)
and, as the response rate was poor (42% European, 14% North and may be severe enough to result in NIV failure.81 Over-
American), the conclusions reported are more qualitative than tightening is a common cause. NIV masks are designed to mould
quantative. When treatment was given it was mostly by bolus to the face when pressurised which over-tightening impairs.
injection and rarely according to a sedation protocol. Greater Should signs of skin trauma become apparent, a barrier dressing
experience in the use of NIV and being a critical care clinician and a strategy of regular breaks and alternating between two
increased reported use of opiates/sedation. interface types should be used. Latex allergy occasionally results
In the 2013 BTS audit, involving 2693 cases, NIV failed to in orid skin reactions. Some patients seem especially prone to
reverse AHRF in 30% of patients.5 Agitation was reported as mask-related rash even in the absence of allergy. Topical steroids
the principal reason in 31% of these. Sedation was attempted may be indicated and/or antibiotics if the wound becomes
in 84%. No details are available on what agents were used, or infected.
outcome in those so treated. As 91% of all NIV treatments were NIV may cause severe gastric distension. It usually indicates
provided outside of the HDU/ICU, it appears sedation is now poor coordination between patient and ventilator and it may be
more commonly employed but in a potentially unsafe necessary to insert a nasogastric tube. Sinus or ear discomfort
environment. and nasal mucosal congestion or drying/ulceration can all occur.
In the ICU setting, case series have reported that infusions of The value of humidication in preventing these side effects is
propofol,72 dexmedetomidine73 and remifentanyl74 75 are safe, uncertain but water-based nasal gels and topical corticosteroids
improve comfort and reduce the failure rate of NIV. Senoglu or decongestants can be used. Petroleum-based emollients
et al76 compared infusions of dexmedetomidine and midazolam should not be used with supplemental oxygen.
in 45 AECOPD cases with AHRF, using a protocol aiming at a An acute pneumothorax may be life-threatening but difcult
standard degree of sedation. No differences were found in to detect. The development of unexplained agitation/distress
effectiveness between the two agents. There were no signicant or chest pain requires this complication to be excluded.82
adverse events and no patient failed to improve with NIV. In Co-existent interstitial lung disease or previous episodes of
another report, the addition of infused dexmedetomidine to a spontaneous or ventilator-induced pneumothorax increase the
standard protocol of as needed bolus intravenous midazolam risk. Using a lower IPAP to avoid large tidal volumes, and a
and fentanyl, given according to a sedation protocol, failed to lower EPAP to avoid signicantly increasing end-expiratory
show benet, but sedation goals were readily achieved and there lung volume (EELV), are logical but not evidence based. If a
was good NIV tolerance and success with the standard proto- pneumothorax develops, intercostal drainage is usually
col.77 A review of sedation to facilitate NIV tolerance makes the required.
Davidson AC, et al. Thorax 2016;71:ii1ii35. doi:10.1136/thoraxjnl-2015-208209 ii11
BTS guidelines
Good practice points pressure and volume has become blurred and hybrid modes
Minor complications are common but those of a serious combine aspects of both. Most patients with AHRF do not
nature are rare. Patients should be frequently assessed to require sophisticated modes of providing IMV.
identify potential complications of NIV. Initially, when airway resistance is high and/or compliance is
Care is needed to avoid overtightening of masks. low (eg, in asthma, CF or bronchiectasis) a period of mandated or
Previous episodes of ventilator-associated pneumothorax controlled mechanical ventilation, often combined with deep
warrant consideration of admission to HDU/ICU and use of sedation to reduce spontaneous breathing effort, allows time for
NIV at lower than normal inspiratory pressures. bronchodilators, steroids and antibiotics to treat airway inamma-
The development of a pneumothorax usually requires inter- tion, overcome infection and for bronchial toilet to be provided.
costal drainage and review of whether to continue with NIV. These considerations also variably apply to the restrictive causes of
AHRF. In addition, poor triggering, because of muscular weak-
Sputum retention ness, is a risk in patients with NMD in whom a prolonged period
Sputum retention can be a precipitant for AHRF, can cause NIV of controlled mechinical ventilation may be necessary. In all
to fail and is a common reason for respiratory distress post- patients with AHRF, allowing restorative sleep is important.9193
extubation in patients initially managed by IMV. Excessive Management should shift towards supporting rather than
sputum production characterises bronchiectasis and CF, and mandating the pattern of ventilation as recovery begins. If there
complicates some patients with AECOPD. Promoting sputum is adequate spontaneous effort, and the RR is not excessive, a
clearance can be particularly challenging in those with NMD switch to PS is recommended to reduce the need for sedation
and in the morbidly obese. Techniques, such as manually and also as the risk of respiratory muscle wasting may be
assisted cough and mechanical insufationexsufation (MI-E), reduced by establishing early spontaneous breathing. The
aid sputum clearance in patients with NMD.83 84 However, in a concept that suppressing spontaneous breathing is causally
study including patients with either scoliosis or COPD, MI-E related to diaphragm wasting is contentious in the literature.
reportedly had no benet.85 In another RCT, the use of MI-E Space constraints prevent a fuller examination. One initially
reduced post-extubation respiratory failure in a mixed group of compelling human study that claimed to have demonstrated
patients including some with AHRF.86 This study also provided disuse atrophy was subsequently critised because the dia-
NIV to those in respiratory distress. The reader is referred phragms had been denervated.94 A study in patients with adult
to the BTS Physiotherapy Guidelines87 for more detailed respiratory distress syndrome (ARDS) reported that those
information. patients allowed to breathe spontaneously had less need for sed-
Mini-tracheostomy facilitates secretion clearance in the spon- ation than patients treated with controlled IMV, a reduced
taneously breathing patient88 and may have a role when sputum requirement for vasopressors, fewer days of ventilatory support,
retention is thought to be a major determinant of AHRF, such earlier extubation and a shorter length of ICU stay.95 This strat-
as in CF. It is not an easy technique to perform in the anxious egy has not been assessed in AHRF.
and breathless patient and training opportunites are rare. Evidence statements
Clinicians who insert percutaneous tracheostomies are best Establishing early spontaneous breathing reduces the need for
placed to provide a service and the HDU/ICU is the best envir- sedation, improves cardiac function and reduces the duration of
onment in which to perform mini-tracheostomy. In an attempt IMV in ARDS (Level 1).
to avoid intubation, a combination of respiratory support by Recommendations
NIV and suctioning via a mini-tracheostomy has been described. 12. Spontaneous breathing should be established as soon as pos-
This probably only has application if IMV is not desired by the sible in all causes of AHRF (Grade C).
patient as, in most such cases, IMV offers more chance of a suc- 13. Controlled IMV may need to be continued in some patients
cessful outcome. In the patient initially managed by IMV, a due to severe airow obstruction, weak muscles leading to poor
mini-tracheostomy may be inserted at the time of endotracheal triggering or to correct chronic hypercapnia (Grade C).
tube decannulation in patients with a high secretion load and/or Good practice point
a poor cough. In obstructive diseases, controlled IMV should be continued
Evidence statements until airway resistance falls.
Manual-assisted cough and MI-E are safe methods for aiding
secretion clearance (Level 1+). Invasive ventilation strategy
MI-E is more effective than manual-assisted cough in patients In obstructive causes, tidal volume (Vt) is limited by the airow
with stable NMD (Level 2+). obstruction and compounded by the mechanical disadvantage of
Mini-tracheostomy is a useful bedside procedure that can mark- hyperination. The use of high ination pressures, to achieve a
edly improve secretion clearance, but requires patient cooperation normal Vt, risks dynamic hyperination.96 It most dramatically
and a skilled operator to be performed safely (Level 4). occurs soon after intubation but may develop on switching ven-
Recommendations tilation mode, for example, from controlled to assisted
10. In patients with NMD, mechanical insufation and exsufa- ventilation.97
tion should be used, in addition to standard physiotherapy tech- The adverse consequences of hyperination include baro-
niques, when cough is ineffective and there is sputum retention trauma, impaired gas exchange and patient discomfort. The
(Grade B). increased intrathoracic pressure impedes venous return and
11. Mini-tracheostomy may have a role in aiding secretion clear- increases right ventricular afterload with a resulting fall in
ance in cases of weak cough (NMD/CWD) or excessive amounts cardiac output and hypotension.98
(COPD, CF), (Grade D). Prolonging expiratory time limits gas trapping and is achieved
by shortening the inspiratory time and reducing the minute
Modes of IMV volume, an approach recommended in airow obstruction.99 100
Critical care ventilators are complex devices capable of deliver- If signicant gas trapping still occurs, the recommendation is to
ing multiple modes.89 90 The traditional divide between use a lower than normal Vt in combination with a low RR and
ii12 Davidson AC, et al. Thorax 2016;71:ii1ii35. doi:10.1136/thoraxjnl-2015-208209
BTS guidelines
a more prolonged expiratory phase.99 100 This can often only Recommendations for IMV in obstructive disease
be achieved using a controlled ventilation mode combined with 14. During controlled ventilation, dynamic hyperination
deeper levels of sedation. On switching to PS (assist) during should be minimised by prolonging expiratory time (I:E ratio
recovery, the inspiratory pressure needs to be sufcient to 1:3 or greater) and setting a low frequency (1015 breaths/min)
provide adequate tidal volume but not excessive. Settings there- (Grade C).
fore need to be individually adjusted and require regular review. 15. Permissive hypercapnia (aiming for pH 7.27.25) may be
In ARDS, over-distention and repetitive recruitment/ required to avoid high airway pressures when airow obstruc-
de-recruitment of lung units causes alveolar damage (so-called tion is severe (Grade D).
ventilator-induced lung injury) and may even provoke systemic 16. Carbonic anhydrase inhibitors should not be used routinely
inammation.101 One explanation for improved outcome with in AHRF. (Grade C).
low Vt ventilation (<6 mL/kg), compared with conventional
practice, may be avoidance of ventilator-induced lung injury.102 Positive end expiratory pressure
The ARDS literature provides evidence for permissive hypercap- PEEP is an area of physiology that causes confusion among
nia, demonstrating that a pH above 7.2 is well tolerated.103 This healthcare professionals. The best way to set optimal PEEP
is the consensus target when pH control is difcult.89 90 remains contentious. Simply stated, PEEP shifts the lungs to a
Allowing permissive hypercapnia will result in cerebral vasodila- more compliant portion of the pressurevolume curve. In
tion and a rise in intracranial pressure and may also compromise restrictive causes of AHRF, lung volume is usually reduced and
myocardial contractility. Attempts to raise pH to >7.2 may, there may be dependent lung that is poorly ventilated or in
however, compound hyperination and barotrauma. In ARDS, a which there is no effective alveolar ventilation. In these circum-
peak airway pressure of 30 cm is the usual trigger for employing stances, increasing external PEEP increases Vt for a given
permissive hypercapnia, a strategy that reduces mortality.104 inspiratory pressure, will reduce pCO2 and improve oxygen-
In AECOPD, attempts to rapidly restore pO2 and pCO2 to ation. In obstructive disease, PEEP improves expiratory airow,
normal are unnecessary. Although there is little evidence to limits dynamic hyperination and improves alveolar ventila-
provide guidance, it is suggested that the higher the pre-morbid tion.108 109 Dynamic hyperination may be suspected by a pro-
pCO2 (inferred by a high admission bicarbonate), the higher the gressive fall in tidal volume with constant ventilator pressure
target pCO2 should be. Recovery from extreme levels of hyper- settings (or, with volume control, an increase in ination pres-
capnia is recognised.105 Any metabolic causes of acidosis, for sure) and by signs of increasing patient distress such as tachycar-
example, from insulin insensitivity or excessive B2 stimulated dia and hypotension.
glycogenolysis, should be treated separately. The degree of intrinsic PEEP (iPEEP) can be estimated by
In NMD, an adequate tidal volume can be achieved with rela- examination of the expiratory ow curve and pressure110 or be
tively low ination pressures (eg, 1015), but higher pressure is measured invasively.111 112 Active expiratory muscle contraction,
needed in CWD because of reduced chest wall compliance. common in airow obstruction, will articially increase appar-
Lung recruitment strategies (ie, increasing PEEP) should be con- ent iPEEP. Levels of iPEEP in obstructive airways disease have
sidered when there is persisting hypoxia and/or evidence of pre- been reported to range from 4.6 to 13.6 cm H2O.113
mature small airway closure in dependent lung tissue. Setting the PEEP level in excess of iPEEP may be deleterious.
Controlled MV may need to be continued in NMD when trig- This has led to the recommendation that PEEP be set at 50
gering is likely to be inadequate or tiring. 80% of iPEEP.114 115 However, as the severity of airway obstruc-
Reducing the bicarbonate buffering capacity will require a tion in small airways will vary throughout the lung, a variable
period of relative hyperventilation when hypercapnia is chronic. response to increasing the PEEP might be anticipated. If, on
The resulting urinary bicarbonate loss resets central respiratory balance, an increase in ePEEP were to reduce overall airway
drive. Carbonic anhydrase inhibitors can be used but caution is resistance then EELV will fall even though ePEEP apparently
needed as high doses produce unpredictable effects through exceeds iPEEP.116 117
central stimulation of breathing.106 107 Intrinsic PEEP is a pressure that must be overcome by patient
Evidence statements effort before a breath can be triggered. It is, therefore, an
In ARDS, a low Vt strategy improves survival (Level 1+). inspiratory threshold load and may lead to ineffective triggering
In airow obstruction, prolonging the expiratory time reduces and patient discomfort. Offsetting iPEEP by increasing the ven-
dynamic hyperination (gas-trapping) (Level 2+). tilator PEEP will then reduce the effort of triggering and
improve patientventilator asynchrony.118120 It is important to monitor the effect on patient comfort and RR. If inadequate PS
appreciate that the same pathophysiological processes occur is given, the breathing rate will fall. The detection of the more
during treatment with NIV when a higher EPAP setting may subtle forms of patientventilator asynchrony requires examin-
improve triggering, patient comfort and oxygenation. ation of the pressure/ow waveforms.135 The most sensitive
Evidence statement measure of patientventilator asynchrony is by simultaneous
In obstructive causes of AHRF, PEEP may increase tidal volume, recordings of diaphragm electrical activity and pressure changes
improve compliance and reduce airow obstruction (Level 2+). in the oesophagus.134 Flow rather than pressure triggers reduce
Setting PEEP greater than iPEEP can be harmful (Level 2+). the incidence of asynchrony,136 137 as has the move away from
In restrictive causes of AHRF, PEEP may assist in lung recruit- volume-controlled ventilation.138 139
ment, improve compliance and correct hypoxaemia (Level 3). Proportional assist ventilation (PAV) and neurally adjusted
Recommendation ventilatory assist (NAVA) are modes that are being assessed as
17. Applied ePEEP should not normally exceed 12 cm ways to reduce patientventilator asynchrony. With PAV, the
(Grade C). degree of pressure support is determined, on a breath by breath
basis, by the patients inspiratory effort.140142 Compared with
Sedation in IMV PS, PAV has been reported to reduce the probability of returning
Patients receiving IMV require sedation, especially before stabil- to a controlled mode and the incidence of patientventilator
ity is achieved.89 Most ICUs use Propofol or a benzodiazepine, asynchrony.143 In NAVA, the ventilator attempts to match neural
either alone or in combination with an opioid. Benzodiazepines drive by adjusting the degree of PS (within safe limits), using the
with inactive metabolites and/or short acting synthetic opioids electrical activity of the diaphragm to drive the ventilator.
have been recommended to avoid over-sedation.121 122 Studies comparing patientventilator interaction show a reduc-
Although sedation increases IMV tolerance, over-use is asso- tion in triggering delay with NAVA, reduced cycling delay and a
ciated with adverse outcomes such as prolonged duration of reduction in asynchrony events.144 145 Uncertainties persist on
IMV, increased ICU length of stay and delirium.123 how to adjust the NAVA level and this technical issue is currently
To avoid this, withholding of further sedation until an object- frustrating efforts to demonstrate clinical benet.
ive degree of wakefulness develops has been investigated. In two It is important to emphasise that patient ventilator asynchrony
trials, this strategy was shown to reduce duration of IMV and is common with NIV. While the same principles apply it has
ICU length of stay.124 125 Studies employing sedation protocols been less frequently recognised or investigated. It can critically
targeting specic (higher) levels of alertness have also reported a affect the success of NIV and the patient experience (see below).
reduction in duration of IMV, ICU and hospital length of Evidence statements
stay.126129 However, a meta-analysis of RCTs on sedation Patientventilator asynchrony is common and deleterious, and
breaks demonstrated safety but failed to conrm benet,130 and can be minimised through informed adjustment of ventilator
a more contemporary RCT, combining protocolised sedation settings (Level 2+).
with daily breaks, also found no benet.131 No study has shown Proportional and NAVA have been shown experimentally to
harm from sedation breaks. The effect of stopping or reducing reduce ventilator asynchrony but have yet to improve patient
sedation on patient experience has not been reported. outcome (Level 2+).
Evidence statements Recommendations
Daily interruption of sedation is safe and may reduce the dur- 19. Ventilator asynchrony should be considered in all agitated
ation of IMV and ICU length of stay (Level 1+). patients (including NIV) (Grade C).
Sedation protocols that target specic levels of alertness may 20. As patients recover from AHRF, ventilator requirements
reduce duration of IMV and ICU length of stay (Level 1+). change and ventilator settings should be reviewed regularly
Recommendation (Grade C).
18. Sedation should be titrated to a specic level of alertness
(Grade B). Use and timing of a tracheostomy
It is accepted that translaryngeal intubation beyond 10 days can
Patientventilator asynchrony be detrimental.146 147 Historically, it was believed that early
Patientventilator asynchrony is common and increases patient tracheostomy reduced ventilator time and ICU length of stay.148
discomfort, the work of breathing, the need for sedation, the A survey of ICU physicians in 2005 found that 61% of respon-
incidence of confusion, the need for tracheostomy and the mor- dents would perform a tracheostomy without rst performing a
tality rate.132 133 The commonest cause is ineffective triggering trial of extubation and 50% favoured tracheostomy insertion
due to either respiratory muscle weakness and/or excessive within the rst week.149 Two large multicentre studies have
effort required to overcome iPEEP and trigger a breath.134 failed to show benet from tracheostomy performed within
Trigger failure is more common during sleep and more likely if 7 days of admission.150 151 A subsequent meta-analysis also
hypercapnia persists by day. A hybrid mode, such as PS with a reported no effect on the incidence of ventilator-associated
mandatory backup rate is recommended in these circumstances pneumonia or mortality,152 although less sedation was required
to avoid pCO2 increasing during sleep. after a tracheostomy had been inserted. Tracheostomy carries a
Auto triggering refers to inappropriately delivered breaths morbidity and mortality risk at the time of insertion153 and sub-
being provided by the ventilator. It can be provoked by patient sequently.154 A UK national report has highlighted the risk of
movement, suctioning, coughing and swallowing, and is more critical airway incidents in patients with tracheostomies.155
likely when the trigger sensitivity is set too high. Both a delay in Accordingly, consideration of the risk and benet should be
the onset of a triggered breath or an inadequate amount of PS undertaken before proceeding to insert a tracheostomy and due
to sufciently augment inspiratory ow can lead to an unpleas- consideration should be given to using NIV post-extubation to
ant sensation best described as air hunger. This can be difcult avoid a tracheostomy. This is particularly the case in progressive
to detect or for the patient to report. Experienced NIV practi- NMD/CWD when tracheostomy insertion carries the risk of
tioners may trial increasing trigger sensitivity and/or PS, and permanence. These aspects, and the evidence summarised
ii14 Davidson AC, et al. Thorax 2016;71:ii1ii35. doi:10.1136/thoraxjnl-2015-208209
BTS guidelines
and/or use of IMV reported has led to the suggestions that the Starting NIV in COPD
clinical environment in which NIV was delivered was inad- Recommendations regarding investigations before starting NIV
equate for the level of patient complexity/acidbase disturbance, are based on expert opinion. ABG measurement is required to
that there was an over-reliance on the effectiveness of NIV and diagnose and quantify the severity of AHRF, and a chest radio-
an under recognition of NIV failure.7 Similar conclusions can be graph is needed to seek evidence of causation or complications.
drawn from BTS NIV audits. In the most recent survey, carried To avoid any delay in giving ventilatory support, NIV should be
out in 2013, median pH was 7.24 and yet NIV was provided initiated in extreme acidosis ( pH<7.25) without waiting for a
outside of HDU/ICU in 91%. In the AECOPD group (61% of chest X-ray. Other investigations (eg, full blood count (FBC), U
AHRF cases), overall mortality was 28% in those admitted to +E, ECG) should be performed and treatment directed at any
HDU/ICU and 40% for those admitted to admission wards.5 reversible factors contributing to AHRF. In some cases, echocar-
See Care planning and delivery of care section for further diography may be indicated to exclude acute pulmonary
consideration of the possible unintended consequences of the oedema. As is further discussed in Care planning and delivery
introduction of NIV in managing AHRF. of care section, it is recommended that an action plan be agreed
In around 20% of AHRF cases secondary to AECOPD, opti- in the event of NIV failure and that this is documented at the
mised medical therapy, which includes targeting an oxygen sat- start of treatment.
uration to 8892%, will result in normalisation of arterial Good practice points
pH.2 62 Established guidance is therefore to await improvement ABG measurement is needed prior to and following starting
and initiate NIV if, after 60 min, the following are present: pH NIV.
<7.35, pCO2 > 6.5 kPa and RR >23 breaths/min.1 48 Chest radiography is recommended but should not delay ini-
There is some evidence that NIV may also be benecial in tiation of NIV in severe acidosis.
patients with hypercapnia in the absence of acidosis. A study Reversible causes for respiratory failure should be sought and
from China162 showed a reduction in the need for endotracheal treated appropriately.
intubation in a subgroup analysis of patients with hypercapnia At the start of treatment, an individualised patient plan
but a pH >7.35 (9/80 vs 2/71, p=0.04). However, length of (involving the patient whenever possible) should document
stay and duration of NIV were longer than in a similar UK agreed measures to be taken in the event of NIV failure.
study,163 and there was a high incidence of side effects, particu-
larly gastric distension (23%), despite low ination pressure
being used (IPAP 124). It is unclear if this study is applicable Prognostic features relating to use of NIV in COPD
to UK practice. The 2003 UK National COPD audit165 demonstrated a higher
There is insufcient evidence to support the use of absolute hospital mortality in patients with a lower admission pH and
values of pH or pCO2 as an indication for IMV rather than oxygen saturation, higher urea, lower albumin and older age
NIV.164 Nevertheless, a pH of 7.25 has been suggested as a (see below for further discussion), irrespective of treatment
threshold level below which IMV should be considered. NIV modality. Increased base excess (indicating chronicity of hyper-
may still be effective at reversing such severe acidosis but the capnia), MRC dyspnoea index and RR are additional prognostic
failure rate is higher.158 variables.165 166 The presence of pulmonary consolidation on
Evidence statements X-ray and impaired consciousness level (GCS<8) increase the
Optimal medical therapy, including controlled oxygen therapy, NIV failure rate,167 although successful outcome despite
leads to a resolution of respiratory acidosis in 20% of indivi- impaired consciousness has been reported.49 168
duals with AECOPD (Level 1+). In contrast, Nava et al169 reported a good outcome for
Compared with standard medical therapy, NIV improves sur- patients aged >75 years, in terms of intubation avoidance and
vival, reduces the need for endotracheal intubation, reduces reduced mortality with NIV. Others have also achieved satisfac-
complications and reduces length of stay (Level 1+). tory results in the elderly.170 However, in a retrospective ana-
There is no lower limit of pH below which a trial of NIV has lysis of 240 ward-based cases from a single centre, age
been shown to be harmful (Level 2++). >75 years was associated with poorer outcomes with NIV.171
Continued use of NIV when the patient is deteriorating, rather The NCROP audit, which collected data on 9716 AECOPD
than escalating to IMV, increases mortality (Level 2+). admissions, reported mortality at 12% when the presentation
Audit data show that real world outcomes do not reproduce pH was the lowest value reached, 24% when acidosis increased
those demonstrated in the RCTs (Level 2+). after presentation and 33% when acidosis only developed after
One risk of an expansion of ward-based rather then HDU/ICU admission.7 These ndings reect a combination of increasing
provision of NIV may be greater delay in expert review and/or severity of illness and a lack of response to standard medical
escalation to IMV (Level 4). treatment. In addition, delay in providing therapeutic NIV and/
Recommendations or IMV contributed. The audit also highlighted that a coinci-
24. For most patients with AECOPD, the initial management dent metabolic acidosis was an adverse nding.
should be optimal medical therapy and targeting an oxygen sat- Once NIV has been initiated, a reduction in RR and improve-
uration of 8892% (Grade A). ment in pH within 4 h predicts NIV success.172 Associated fea-
25. NIV should be started when pH<7.35 and pCO2>6.5 kPa tures are a reduction in signs of respiratory distress, reduced
persist or develop despite optimal medical therapy (Grade A). anxiety or agitation and a decrease in heart rate. In one large
26. Severe acidosis alone does not preclude a trial of NIV in an study, Confalonieri et al173 showed that, if pH <7.25 and RR
appropriate area with ready access to staff who can perform safe >35 persist, NIV failure is likely. Worsening acidosis, after
endotracheal intubation (Grade B). initial improvement with NIV, is also associated with a worse
27. The use of NIV should not delay escalation to IMV when prognosis.174176 In a case series published by Moretti et al,
this is more appropriate (Grade C). 20% of patients deteriorated after initially improving with NIV.
28. The practice of NIV should be audited regularly to maintain In these circumstances, prognosis was poor whether patients
standards (Grade C). were subsequently intubated or continued with NIV.177
ii16 Davidson AC, et al. Thorax 2016;71:ii1ii35. doi:10.1136/thoraxjnl-2015-208209
BTS guidelines
Roche Campo et al93 found that polysomnographic evidence uid retention, obstructive sleep apnoea (OSA) or OHS. One
of severe sleep disturbance in patients with COPD with AHRF study suggested there may be an advantage to employing NIV
correlated with a poor outcome and Gursel et al,91 reporting on for longer than the conventional 3 days.162 More trial data are
a retrospective analysis of patients with COPD and OHS treated needed to guide optimal withdrawal of NIV.
in an ICU setting, found better outcome in patients receiving Evidence statement
pressure control rather than overnight PS. Clinical research in In clinical trials, NIV has been discontinued when there has
stable sleep hypoventilation also suggests that limiting the been normalisation of pH and pCO2 and a general improve-
increase in hypercapnia during sleep is important and that a ment in the patients condition (Level 1+).
controlled ventilation mode may be more advantageous than the Recommendation
assist mode.178 31. NIV can be discontinued when there has been normalisation
Evidence statements of pH and pCO2 and a general improvement in the patients
Advanced age is not an important determinant of outcome with condition (Grade B).
NIV treatment of AHRF (Level 1+). Good practice points
An improvement in physiological parameters, usually within 1 Time on NIV should be maximised in the rst 24 h depend-
2 h, particularly pH and RR, predict a successful outcome from ing on patient tolerance and/or complications.
NIV treatment (Level 1+). NIV use during the day can be tapered in the following 2
Worsening of physiological parameters, particularly pH and RR, 3 days, depending on pCO2 self-ventilating, before being dis-
is predictive of an increased risk of death and/or requirement continued overnight.
for intubation (Level 1+).
Recommendations Optimising NIV delivery and technical considerations
29. Advanced age alone should not preclude a trial of NIV The commonest reasons for failure of NIV are excessive mask
(Grade A). leak, insufcient PS and ventilator patient asynchrony. If the PS
30. Worsening physiological parameters, particularly pH and is inadequate, alveolar ventilation will not be signicantly
RR, indicate the need to change the management strategy. This increased. This may be detected by a lack of augmentation of
includes clinical review, change of interface, adjustment of venti- chest and abdominal wall movement. National NIV audits have
lator settings and considering proceeding to endotracheal intub- revealed that inadequate IPAP is often used in AECOPD.3 4 In
ation (Grade A). general, while a patient might be started on NIV with an IPAP
Good practice point of 15, this should be progressively increased to reach an IPAP of
If sleep-disordered breathing pre-dates AHRF, or evidence of it 2030 within 1030 min, the need for higher pressure and a
complicates an episode, the use of a controlled mode of NIV more rapid escalation being indicated by patient size and more
overnight is recommended. severe acidosis, respectively.
In the presence of persisting hypoxaemia, that is thought
Duration of NIV in COPD unrelated to sputum retention, the EPAP may need to be
Normalisation of pH and a pCO2 <6.5 are commonly used as a increased in an attempt to recruit areas of poorly ventilated
guide to the discontinuation of NIV. Restoring respiratory drive lung. (It may also be appropriate if there is a degree of upper
will require a prolonged period of NIV to reduce the pCO2 airway obstruction). If ineffective, or if this results in distress,
than to correct the acidosis. senior review is indicated while the FiO2 is temporarily
The optimal amount of NIV in the initial period, and the increased.
most effective way to withdraw it as the patient improves, have Leak should always be minimised by mask adjustment and/or
not been examined in published trials. As the work of breathing by changing the mask type. Positional upper airway obstruction
falls and acute hyperination reverses, as a result of treatment may result in ineffective NIV. It is often indicated by variable
with steroids, antibiotics and intense broncholdilator therapy, mask leak. Care is needed to ensure head exion is avoided,
unsupported alveolar ventilation will return towards normal.
The more orid the evidence for infection precipitating AHRF,
the more likely there is to be full reversal. Normalisation of
pCO2 may not be possible in some patients, particularly those Table 3 Technical issues: a guide for when NIV is failing
who show evidence of chronic hypercapnia at presentation. Problem Cause(s) Solution (s)
In most RCTs, the intention has been that patients should
Ventilator cycling Inspiratory trigger Adjust trigger
receive semicontinuous NIV for the rst 24 h. The amount of independently of patient sensitivity is too Reduce mask leak
NIV actually delivered, when this has been reported, has been effort high
less than planned, from a median of 20 h in one study179 to 7 h Excessive mask leak
in another.163 Conventional practice is to gradually reduce the Ventilator not triggering Excessive mask leak Reduce mask leak
amount of time on NIV, with increasingly prolonged periods of despite visible patient effort Inspiratory trigger Adjust trigger
self-ventilation during the day, while continuing with NIV over- sensitivity too low For NM patients consider
switch to PCV
night. Monitoring of pCO2 on and off NIV is a useful guide to
Inadequate chest expansion Inadequate Tidal Increase IPAP. In NM or
how quickly the withdrawal of NIV can proceed. despite apparent triggering volume chest wall disease
Transcutaneous pCO2 measurement may facilitate this better consider longer Ti
than continuing with arterial or capillary sampling. A gradual Chest/abdominal paradox Upper airway Avoid neck flexion
reduction of ventilator pressures, and a switch to PS or a reduc- obstruction Increase EPAP
tion in backup rate, should mirror patient recovery. Attempts to Premature expiratory effort Excessive Ti or IPAP Adjust as necessary
adjust ventilator settings to achieve patient comfort remain by patient
important. Those with a less clear infective cause for AHRF, EPAP, expiratory positive airway pressure; IPAP, inspiratory positive airway pressure;
and/or evidence of chronicity of hypercapnia, should be assessed NIV, non-invasive ventilation; NM, neuromuscular; PCV, pressure-controlled
ventilation.
for alternative or additional causative factors such as marked
Davidson AC, et al. Thorax 2016;71:ii1ii35. doi:10.1136/thoraxjnl-2015-208209 ii17
BTS guidelines
In summary, an admission with AHRF is a critical point in the deteriorate rapidly, to require high ination pressures and a high
natural history of COPD, with a more accelerated decline in most inspired oxygen concentration. Trialling NIV therefore carries
patients following recovery. It indicates a high risk of recurrence of signicant risk. The patient with brittle asthma or a very short
AHRF and poor long-term prognosis. Survival data vary between history suggesting hyperacute brochospasm, especially when
published studies, presumably a reection of differences in case oxygen toxicity in transit is implicated, might justify a trial of
mix. Survival might be better if NIV is successfully employed for NIV in the resucitation area but, in all other circumstances, ven-
AHRF than if IMV is used. An important management point is that tilatory support should be by intubation and provision of IMV.
the rst episode of AHRF should prompt a discussion about life- The overall invasive management of acute severe asthma is
style, patient wishes for management of future episodes and discus- similar to that in AECOPD but a higher SaO2 target of 96% is
sions about end-of-life care generally. advised. For more specialist consideration, the reader is referred
There is evidence of prognostic pessimism among clinicians to standard textbooks or recent reviews.
caring for patients with AECOPD. In an outcome study of 517 Acute (and chronic) hypercapnia may complicate chronic
patients, 62% survived to 180 days, yet overall predicted survival asthma, a condition that shares many features of COPD, such as
at the time of admission was 49%.182 For those considered to be chronic hyperination, persistent and only partially reversible
in the worst prognostic group (a survival rate of 10%), 40% recov- airow obstruction, mucus hypersecretion and infective exacer-
ered. Accordingly, as survival from AECOPD becomes less likely, bations. As the pathophysiology is similar, the guidance on the
clinicians become worse at prediction and err on the side of under- use of NIV and IMV in AECOPD applies to the chronic asth-
estimating survival. By implication, it is likely that patients who matic with AHRF.
might otherwise survive are currently being denied admission to Evidence statements
ICU because their survival potential is underestimated. There is insufcient evidence to support the use of NIV in
Importantly, from a patient perspective, Wildman et al193 reported AHRF in acute asthma (Level 3).
that the majority of patients surviving IMV for AHRF had stable IMV in acute asthma carries a very low mortality rate. Most
and acceptable QoL despite poor health status and 96% stated asthma deaths relate to presentation in extremis or a failure to
they would opt for IMV again under similar circumstances. immediately implement IMV when indicated rather than a
Evidence statement failure of IMV per se (Level 2+).
There are validated tools for the assessment of prognosis in Recommendations
stable and exacerbating COPD populations but, on their own, 39. NIV should not be used in patients with acute asthma
they are unreliable for individual prognostication (Level 2++). exacerbations and AHRF (Grade C).
Physicians underestimate survival potential in AECOPD treated 40. Acute (or acute on chronic) episodes of hypercapnia may
by IMV (Level 2+). complicate chronic asthma. This condition closely resembles
The majority of patients with COPD or chronic asthma who COPD and should be managed as such (Grade D).
receive IMV would elect to receive it again (Level 2+).
An episode requiring ventilatory support generally indicates Non-CF bronchiectasis
advanced disease with a high risk for future episodes of AHRF Recurrent episodes of hypercapnic respiratory failure may char-
and a limited prognosis (Level 2++). acterise bronchiectasis with periods of good or acceptable
Recommendations quality of life (QoL)/health status in the intervening months or
36. Prognostic tools may be helpful to inform discussion regard- years. In some, domiciliary ventilation will be indicated for
ing prognosis and the appropriateness of IMV with the caveat symptoms of sleep disordered breathing. There are no RCTs of
that such tools are poorly predictive for individual patient use NIV versus IMV in acute exacerbations of bronchiectasis. The
(Grade B). recommendations regarding NIV for AECOPD are appropriate
37. Clinicians should be aware that they are likely to underesti- although there is the additional challenge of excessive and dif-
mate survival in AECOPD treated by IMV (Grade B). cult to clear sputum. NIV may relieve breathlessness and help
38. Clinicians should discuss management of possible future epi- patients to participate more effectively with physiotherapy. A
sodes of AHRF with patients following an epsiode requiring mini-tracheostomy, or other techniques to aid sputum clearance,
ventilatory support because there is a high risk of recurrence may be indicated.202
(Grade B). There is scant data on outcomes for AHRF in non-CF bron-
chiectasis. In a retrospective review203 of patients managed by
Acute asthma NIV (n=31) or IMV (n=26) for AHRF, the NIV group had less
Five small RCTs194198 of NIV in acute asthma have been pub- severe physiological disturbance. There was no difference in
lished. Four were conducted in the ED and one in a respiratory hospital mortality between the two groups (26% and 27%). The
ICU. Importantly, none of the RCTs included patients with NIV failure rate (need for intubation, or death in the ICU) was
hypercapnia and intubation rates were low.199 Most showed 33%. Using logistic regression, the APACHE II score was the
treatment with NIV led to a faster improvement in FEV1 and a only predictor of hospital mortality (OR 1.19 per point) and
shorter ICU/hospital stay. They all have important design weak- the severity of hypoxia ( pO2/FiO2 ratio) the only predictor of
nesses. The trial by Soma et al197 lacked a second control arm NIV failure (OR 1.02/mm Hg decrease). Hospital mortality of
(conventional inhaled bronchodilators) and the trial by Brandao patients with AHRF secondary to bronchiectasis is therefore
et al194 did not give systemic steroids. No information was pro- approximately 25% whether management is by NIV or IMV.
vided about acceptability of NIV to patients. The only study When selectively applied, NIV fails in one-third and this is pre-
reporting use in AHRF asthma was a retrospective cohort study dicted by the degree of hypoxaemia. Similar criteria should be
by Meduri et al200 of 17 patients with a mean pH of 7.25. NIV used as in AECOPD when deciding appropriateness of intub-
was reported to be successful in avoiding intubation in 15. ation: health status, comorbidities, previous episodes of IMV
The use of NIV in acute asthma, particularly AHRF, needs to and patient preferences. Evidence of an acute precipitating
be set in the context of a very low mortality rate with IMV.201 factor (infection) should favour intubation, as reversibility is
There is also the potential for patients with acute asthma to more likely than in progressive chronic hypercapnia.
Davidson AC, et al. Thorax 2016;71:ii1ii35. doi:10.1136/thoraxjnl-2015-208209 ii19
BTS guidelines
performed. Such patients may have established pulmonary In patients with NMD or CWD, senior/experienced input is
hypertension, chronic hypoxaemia and polycythaemia. needed in care planning and is essential if differences in
In contrast to AECOPD, where the degree of acidosis is more opinion exist or develop between medical staff and patient
important than the degree of hypercapnia, any elevation of representatives.
pCO2 in NMD/CWD may herald an impending crisis.213 In patients with NMD, it should be anticipated that bulbar
Patients have a reduced respiratory reserve but may initially dysfunction and communication difculties, if present, will
sustain sufcient alveolar ventilation to maintain normal carbon make NIV delivery difcult and may make it impossible.
dioxide tension. Minor infection, such as coryza, may be pro- Discussion about NIV and IMV, and patients wishes with
vocative and over the next 2472 h progressive hypercapnia respect to cardiopulmonary resuscitation, should occur as
may develop. Tolerance of acute and chronic hypercapnia varies part of routine care in patients with NMD or CWD.
considerably. Some patients are excessively sleepy with minimal In patients with NMD or CWD, nocturnal NIV should
elevation of pCO2, while others remain alert despite much usually be continued following an episode of AHRF pending
more severe hypercapnia. NIV should be considered in any discussion with a home ventilation service.
breathless/acutely unwell patient with NMD/CWD before
respiratory acidosis develops. NIV failure and discontinuing NIV following recovery in NMD and CWD
In the absence of bulbar dysfunction, NIV is usually well tol- Decisions regarding resuscitation and intubation can be particu-
erated in the restrictive causes of AHRF. Unless there is signi- larly challenging as little or no evidence exists for most of the
cant skeletal deformity, a low degree of PS (eg, a pressure causative conditions, communication with the patient may be dif-
difference of 812 cm) is needed in NMD. By contrast, in cult and/or cognition be impaired and there may be unreasonable
severe kyphoscoliosis, an IPAP >20, and sometimes up to 30, expectation on the part of families and carers. A resuscitation plan
may be required because of the high impedance to ination. is important but may be difcult to negotiate. Inability to clear
Expiratory ow is normally not limited in either restrictive cat- secretions is a common cause of NIV failure. This may result from
egory and the inspiratory/expiratory time (IE) ratio for the an excessive volume of secretions or from a combination of
backup rate (or PCV) should initially be set at 1:1 to allow an limited inspiratory capacity, expiratory muscle weakness and
adequate time for inspiration. Bulbar dysfunction renders effect- bulbar dysfunction. Specialist advice and experience is required to
ive NIV more difcult to achieve, requires a higher EPAP to manage NIV in the presence of bulbar dysfunction and to provide
overcome upper airway obstruction and needs special attention effective cough assistance.215 As with all patients, good communi-
to aid cough and the clearing of upper and lower airways. cation is important. As this may be a challenge, it is another reason
Clinical experience in providing NIV is needed to best titrate for seeking specialist help and advice. Enlisting the help of normal
the EPAP. A modest increase in the domiciliary ventilator set- carers may be useful because they may engender more reassurance
tings is advised in the case of home mechanical ventilation to patients and be better at aiding sputum clearance.
patients being admitted with AHRF. Recovery usually takes longer than in AECOPD, so that step-
While triggering is usually normal in CWD, it is commonly ping down the time on NIV should proceed more slowly, and
inadequate in the other restrictive conditions. Many patients NIV will need to be continued overnight. The higher the pres-
with NMD nd PCV more comfortable and this may also more entation HCO3, the longer the period of relative hyperventila-
effectively control nocturnal hypoventilation. tion required to reduce buffering capacity. A target pCO2
Evidence statement around 6.5 kPa self ventilating is recommended. Following
There are no trials comparing NIV with IMV in AHRF in recovery, the majority of individuals with NMD or CWD will
NMD or CWD. Domiciliary NIV is effective in treating chronic require NIV at home. NIV should continue overnight until dis-
hypercapnia, improves long-term survival and preserves a good cussion with a home ventilation service.
or acceptable QoL (Level 4). Good practice points
Recommendations In patients with NMD or CWD, intolerance of the mask and
49. Controlled oxygen therapy should be used in patients with severe dyspnoea are less likely to cause NIV failure. Bulbar
NMD or CWD and AHRF (Grade D). dysfunction makes NIV failure more likely.
50. NIV should almost always be trialled in the acutely unwell Deterioration in patients with NMD or CWD may be very
patient with NMD or CWD with hypercapnia. Do not wait for sudden. Difculty achieving adequate oxygenation or rapid
acidosis to develop (Grade D). desaturation during a break from NIV are important warning
51. In patients with NMD or CWD, NIV should be considered signs.
in acute illness when VC is known to be <1 L and RR >20, In patients with NMD or CWD, the presence of bulbar dys-
even if normocapnic (Grade D). function, more profound hypoxaemia or rapid desaturation
52. In patients with NMD or CWD, consider controlled ventila- during NIV breaks suggests that placement in HDU/ICU is
tion as triggering may be ineffective (Grade D). indicated.
53. In NMD and CWD, unless escalation to IMV is not desired
by the patient or is deemed to be inappropriate, intubation IMV in NMD/CWD
should not be delayed if NIV is failing (Grade D). Many clinicians have limited experience of managing NMD and
Good practice points CWD. There is the danger of underestimating survival potential
Individuals with NMD and CWD who present with AHRF in the face of severe general disability. Patient choice and seeking
should not be denied acute NIV. the views of advocates when communication with the patient is
NIV is the ventilation mode of choice because patients with difcult are paramount. Discussion with a specialist centre on
NMD or CWD tolerate it well and because extubation from both the delivery of IMV and weaning is recommended.
IMV may be difcult. The risk of sudden deterioration is greater due to reduced
In patients with NMD or CWD, deterioration may be rapid respiratory reserve, impaired cough, cardiomyopathy ( possibly
or sudden, making HDU/ICU placement for therapy more undiagnosed) and sometimes communication challenges.
appropriate. Intubation practice, elective or in AHRF, varies between centres
Davidson AC, et al. Thorax 2016;71:ii1ii35. doi:10.1136/thoraxjnl-2015-208209 ii21
BTS guidelines
and between countries. For instance, in motor neurone disease 55. Advance care planning, particularly around the potential
(MND), elective intubation is reported to occur in 0.8% future use of IMV, is recommended in patients with progressive
(Ireland), 6% (USA) and 10.6% (Italy) of cases.214 NMD or CWD. This may best be supported by elective referral
Outcome data following IMV are limited to case series in to a home ventilation service (Grade D).
MND and OHS. These reports usefully illustrate shared issues
in progressive NMD and many patients with advanced CWD. IMV strategy in NMD and CWD
One report of outcome in MND following intubation for In patients with NMD without signicant chest wall distortion,
AHRF highlighted that 50% of patients were undiagnosed at the impedance to ination is low. It is rarely necessary to use an
the time of intubation, only 17% weaned and few left hos- IPAP above 20. It should initially be set at 10 and increased
pital.216 Recently, Sancho et al217 reported a median survival of according to the resulting tidal volume. In contrast, patients
1 year in patients intubated after failing acute NIV. Chio et al214 with kyphoscoliosis usually require high ination pressures.
reported on 1260 MND cases, over an 8-year period, from a Expiration is generally not ow limited but impedence is typic-
single Italian neurology centre; 134 patients received IMV, ally high so that an I:E ratio of 1 to 1 is recommended in both
which was initiated as an emergency in 40%. Median survival diagnostic groups.
was 250 days. Death occurred in hospital in 20%, at home in When lung volume is reduced, there is radiological evidence
48% and in a nursing home in 32%. Neither patient experience of lobar collapse or unexplained hypoxia, the PEEP setting on
nor economic analysis was reported. the ventilator may need to be increased up to or above
The outcome of patients with MND referred to a specialist 10 cm.14 15 Adjustments should be individualised according to
weaning service in the UK was examined by Chadwick et al.218 ventilatory parameters (RR, dynamic compliance, plateau pres-
Thirty patients had been transferred over a 15-year period. sure) and patient comfort.
Diagnosis followed intubation in 17. In 14 patients, extubation Good practice points
to long-term NIV was possible, of whom 9 were non-bulbar Patients with NMD usually require low levels of PS.
cases and 10 returned home. Thirteen remained tracheostomy Patients with chest wall deformity usually require higher
ventilated, of whom 9 were bulbar and 7 returned home. Median levels of PS.
survival from tracheal intubation was 13.7 months (95% CI 0 to PEEP in the range 510 is commonly required to increase
30.8) for those known to have MND and 7.2 months (95% CI residual volume and reduce oxygen dependency in both
5.1 to 9.4) for those not previously diagnosed. patient groups.
There has been a call for the value of IMV in MND to be
re-evaluated both as an elective policy and at the time of crisis.219 Obesity hypoventilation syndrome
In many of the other NMDs, for example, acid maltase deciency In obese patients, hospitalised for any reason, the presence of
and Duchene Muscular Dystrophy, a more prolonged survival hypercapnia increases morbidity and mortality.220 Despite this,
rate with a good QoL is to be expected following recovery from currently, there is a lack of evidence to guide treatment of either
AHRF, and an aggressive approach to managing it is, in the chronic hypercapnia or AHRF complicating obesity. One non-
opinion of the guideline group, more justied than has historically randomised trial suggested that long-term survival is better in
been the case in the UK. It is also what most patients and their those who accept treatment for sleep disordered breathing com-
families want. Expert experience is that the majority of patients pared with those who do not.221 Severe OSA is the principal
will survive a period of IMV. Comorbidity, especially associated cause of hypercapnia, but hypoventilation also results from the
cardiomyopathy, is important prognostically. The weaning process mechanical effect of obesity.222
is often prolonged but, in the absence of severe bulbar dysfuntion, Presentation with acute on chronic respiratory failure is more
many can be safely extubated onto NIV and avoid a tracheostomy. common than de novo AHRF but the precipitant cause for
Should this fail, and a tracheostomy be required, specialist centre destabilisation may be unclear. Not uncommonly, chronic hyper-
experience is that subsequent decannulation is possible in most. capnia is unexpectedly revealed peri-operatively following
While long-term survival may be limited, QoL may be acceptable routine or emergency surgery in an obese patient not known to
and health status may improve with domicilary NIV. This is par- have OHS. The possibility of OSA/OHS in the morbidly obese
ticularly the case in the more slowly progressive NM conditions (BMI >35) needs to be borne in mind by surgical and anaes-
and in stable CWD. In the latter group, even advanced pulmonary thetic teams.
hypertension may resolve. In the absence of evidence, we recommend that the indica-
Evidence statements tions for NIV in the breathless obese patient should be the same
There are national (and centre) differences in use of IMV in as in AECOPD ie pCO2 >6.5 and pH <7.35. Additionally, NIV
AHRF complicating motor neurone disease (Level 3). should be considered in any patient admitted to hospital with a
The diagnosis of motor neurone disease, and other neuromuscu- raised pCO2 who is excessively somnolent or when there is evi-
lar conditions, is sometimes only made after admission to the dence of uid retention. Following recovery, patients will need
ICU for IMV (Level 3). to be referred to an HMV centre. Patients with OHS can some-
De-cannulation of a tracheostomy is more difcult when there is times be switched to CPAP at a later date.
bulbar disease (Level 3). Evidence statements
Planned elective domiciliary NIV is preferable to crisis In patients with OHS, NIV is indicated if there is respiratory
management in NMD and CWD. This reduces the risk of acute acidosis using the same criteria as in AECOPD (Level 1).
presentation and provides a proven alternative to IMV which In the absence of acidosis, NIV may be indicated in some hyper-
risks prolonged or permanent tracheostomy ventilation (Level 3). capnic and/or somnolent obese patients (Level 2+).
Recommendations Recommendations
54. In patients with NMD or CWD, senior staff should be 56. Controlled oxygen therapy should be used in patients with
involved in decision-making, in conjunction with home mechanical OHS and AHRF (Grade D).
ventilation specialists, if experience is limited, and especially when 57. In patients with OHS, NIV should be started in AHRF,
the appropriateness of IMV is questioned (Grade D). using the same criteria as in AECOPD (Grade B).
ii22 Davidson AC, et al. Thorax 2016;71:ii1ii35. doi:10.1136/thoraxjnl-2015-208209
BTS guidelines
58. NIV is indicated in some hospitalised obese hypercapnic IMV strategy in OHS
patients with daytime somnolence, sleep disordered breathing Intubation can be challenging and patient deterioration may be
and/or right heart failure in the absence of acidosis (Grade D). rapid. There is also a higher risk of aspiration. Pressure control
is recommended until stability has been achieved and should be
NIV settings and placement in OHS initially set at 20 and increased according to the resulting tidal
Obese patients with severe AHRF have a signicant risk, volume. Inspiratory pressure in excess of 30 may be required.
despite receiving NIV, of sudden deterioration and are likely To recruit collapsed lung, PEEP may need to be 1015 cm.14 15
to be difcult to intubate (see below). Upper airway obstruc- It should be adjusted according to ventilatory parameters (RR,
tion is common and will be more apparent during sleep. It dynamic compliance, plateau pressure) and patient comfort.
may persist, despite increasing the EPAP, as indicated by inter- Good practice points
mittent abdominothoracic paradox during NIV assisted In patients with OHS, pressure-controlled MV is recom-
breaths. Another clue is intermittent mask leak that accompan- mended initially.
ies obstructed inspiration. A more upright position may help, In patients with OHS, high PEEP settings may be needed to
but an EPAP in the 1015 range is often required. Expert recruit collapsed lung units and correct hypoxaemia.
assessment is recommended to titrate the EPAP. Tidal volume In patients with OHS, forced diuresis is often indicated.
may be compromised by high level EPAP and, in some, the
impedance to ination is very high and an IPAP of >30 is WEANING FROM IMV
required.217 Prolonging Ti will increase the resulting Vt deliv- Introduction
ered so a I:E ratio of 1:1 is advised. If the resulting Vt is still Weaning is dened as the progressive reduction of ventilatory
inadequate, consideration should be given to using volume- support leading up to extubation. Delayed weaning complicates
controlled ventilation or a volume-assured mode,92 although 6% of patients managed by IMV but consumes 37% of ICU
the benets of the latter are currently unproven. Different resources.223 In one study, up to 50% of patients who self-
EPAP settings may be appropriate depending on sleep/awake extubated did not require re-intubation,224 implying that many
state. patients are treated with IMV for longer than necessary. Clinical
Good practice points criteria to be met before starting weaning are detailed
High IPAP and EPAP settings are commonly required in below:225 226
patients with OHS (eg, IPAP >30, EPAP >8). Adequate oxygenation: PaO2/FiO2 ratio >27 kPa (200 mm Hg)
Volume control (or volume assured) modes of providing NIV FiO2 <0.5
may be more effective when high ination pressures are PEEP <10 cm H2O
required. Adequate alveolar ventilation ( pH >7.3, pCO2 <6.5 kPa).
Fluid balance should also be optimised. The detrimental
NIV failure in OHS effect of excess hydration is now recognised in sepsis227 and in
In patients with OHS, the same indicators suggest a failing acute lung228 and kidney injury.229 A positive uid balance
patient and the same troubleshooting solutions apply as in adversely affects alveolar ventilation, oxygenation, weaning pro-
AECOPD (see table 3). Fluid retention is common and its gress and extubation outcome.224 230 Brain Natriuretic Peptide
extent is commonly under-estimated. It may be in excess of (BNP) has been reported to predict failure to wean and corre-
20 L. Achieving an SaO2 8892% may be difcult and relates to lates with weaning duration; a BNP-directed uid management
collapse of dependent lung and/or reects underlying pulmon- strategy has been reported to shorten time to extubation, par-
ary vascular disease. Sudden and precipitous falls in oxygenation ticularly in patients with left ventricular dysfunction.231
may follow temporary removal of NIV. If high EPAP settings fail Evidence statements
to improve the A-a gradient, a ventilator offering oxygen blend- Easily measured clinical parameters indicate when weaning can
ing may be required. Difculty in clearing secretions may con- start (Level 2+).
tribute to poor gas exchange. Excess uid administration may delay weaning or contribute to
Good practice points its failure (Level 2++).
Fluid overload commonly contributes to ventilatory failure in In left ventricular dysfunction, a BNP-directed uid manage-
patients with OHS and its degree is easily underestimated. ment strategy has been shown to shorten the duration of IMV
Forced diuresis may be useful. (Level 2).
As the risk of NIV failure is greater, and intubation may be Recommendations
more difcult, placement in HDU/ICU for NIV is 59. Treating the precipitant cause of AHRF, normalising pH,
recommended. correcting chronic hypercapnia and addressing uid overload
should all occur before starting weaning (Grade D).
Discontinuing NIV in OHS 60. A BNP-directed uid management strategy should be
During wakefulness, weaning of NIV should proceed as in considered in patients with known left ventricular dysfunction
AECOPD. NIV overnight should be continued pending discus- (Grade B).
sion with the local home ventilation service. Other aspects, such
as consideration of bariatric surgery and optimal EPAP settings Weaning methods
when returning home, are important aspects of continuing care. Despite several multinational studies, there is no consensus as to
Good practice points the optimal weaning method. Brochard et al232 reported that
NIV can be discontinued as in patients with AECOPD. progressively reducing PS was better than other weaning
Many patients with AHRF secondary to OHS will require methods. Subsequent trials have reported that daily (or multiple)
long-term domiciliary support (CPAP or NIV). T piece trials (SBTs) are as effective as PS weaning.233 234 It is
Following an episode of AHRF, referral to a home ventilation likely that patient-specic characteristics are more important
service is recommended. than the weaning protocol in determining the duration of
weaning. There is agreement that the Synchronised Intermittent reduces the risk of early failure. Late extubation failure is more
Mandatory Ventilation method is inferior to PS and T piece complex in aetiology and more than one cause may be present.
weaning. It is also accepted that a formalised weaning plan, and The causes are summarised below:244
staff familiarity with the approach adopted on the ICU, are Capacityload imbalance: patients with severe airow
important factors to improve successful weaning.235 obstruction or neuromuscular weakness;
Evidence statement Impaired bulbar function: aspiration of upper airway secre-
Progressive reduction of PS and daily SBTs are satisfactory tions, impaired gas exchange and/or obstructed breathing;
methods of weaning (Level 1+). Ineffective cough: typically in NMD/CWD but also in other
Recommendations patients with AHRF;
61. Assessment of the readiness for weaning should be under- Non-respiratory issuesmyocardial ischaemia/left ventricular
taken daily (Grade B). dysfunction, encephalopathy/delirium or severe abdominal
62. A switch from controlled to assisted IMV should be made as distension.
soon as the patient recovery allows (Grade C). Evidence statement
63. IMV patients should have a documented weaning plan Patient, clinical and ventilatory factors aid the identication of
(Grade B). patients at increased risk of extubation failure (Level 2+).
Recommendation
Assessing readiness for discontinuation of mechanical 66. Care is needed to identify factors that increase the risk of
ventilation extubation failure so that additional support, such as NIV or
SBTs are used to assess readiness to resume normal breathing. cough assist, can be provided (Grade B).
During the SBT, a patient breathes with minimal or no PS
(dened as <8). A successful trial requires the absence of Weaning protocols
respiratory distress. Failure of an SBT may be dened by subject- Weaning protocols that specify the steps to follow during
ive (comfort) or objective (deterioration in gas exchange or mea- weaning have been claimed to reduce the duration of IMV,
sured ventilator parameters) criteria.232 233 Studies have shown increase the success of extubation, reduce unplanned or acciden-
that the majority of SBT failures occur within 30 min.236 237 tal extubation and reduce the tracheostomy rate, ventilator-
Repeated failure of SBT should lead to consideration of other associated complications and costs, compared with usual
methods of weaning.238 239 care.245 The studies summarised in this review were, however,
It is important to note that the criteria that dene success of not specic to AHRF. Most were performed in the USA, where
an SBT do not necessarily reect the likelihood of successful differences in supervision of patient management exist com-
extubation. About 10% of patients who successfully manage an pared with the UK. There is also marked variation in the
SBT will fail to maintain adequate gas exchange and/or develop weaning methods and protocols between the studies reported. A
signs of distress following extubation.240 An SBT assesses the European study reported that a weaning protocol did not
balance of respiratory load to capacity of the respiratory reduce ventilation time.246 Computer-automated weaning, in
muscles but does not take into account other factors that may which adjustment in pressure settings occurs in response to
affect success such as upper airway patency, bulbar function, changes in patient parameters, has been compared to
sputum load or effectiveness of cough.240 professional-led weaning. One multicentre RCT found that dur-
Evidence statement ation of weaning was reduced.247 A second study reported no
A SBT is useful in assessing load/capacity but does not predict difference in weaning duration between automated weaning and
the success of extubation (Level 1+). weaning by an experienced nurse.248 There is currently insuf-
Recommendation cient evidence to support the use of automated weaning over
64. A 30 min SBT should be used to assess suitability for extuba- clinical/nurse-led protocols.
tion (Grade B). Evidence statements
65. Factors including upper airway patency, bulbar function, Weaning protocols may reduce the duration of IMV and ventila-
sputum load and cough effectiveness should be considered prior tor associated pneumonia (Level 1+).
to attempted extubation (Grade D). There is conicting evidence regarding the value of computer-
automated weaning (Level 1).
Outcome following extubation
Successful extubation is dened as the absence of the need for
ventilatory support for 48 h. Patients receiving post-extubation
NIV (see below) are classied as weaning in progress.241 Box 2 Risk factors for extubation failure following
Much of the evidence regarding the prediction of the risk of invasive mechanical ventilation (IMV)
post-extubation failure has come from trials of relatively short
duration IMV and with a mixture of underlying patholo- Positive uid balance
gies.235 242 243 Several risk factors have been identied. Raised rapid shallow breathing index during spontaneous
The more adverse factors present, the greater the risk of breathing trial
extubation failure. Risk factors for extubation failure are shown Pneumonia or pulmonary disease as the cause requiring IMV
in box 2.235 242 243 Increased age
Respiratory distress may occur early or develop later on after Prolonged duration of IMV
extubation. Early failure commonly results from loss of airway Anaemia
patency, for example, from upper airway oedema that becomes Increased severity of illness
evident following removal of the endotracheal tube.244 Patients Low albumin
with NMD are at risk of early extubation failure due to bulbar Previous failed extubation
dysfunction and/or ineffective cough despite a successful SBT. Bulbar dysfunction
The planned use of NIV and an MI-E following extubation
ii24 Davidson AC, et al. Thorax 2016;71:ii1ii35. doi:10.1136/thoraxjnl-2015-208209
BTS guidelines
Recommendations re-intubation, 9 h longer than the control group. This may have
67. Although an organised and systematic approach to weaning contributed to the worse outcome. Post hoc analysis suggested a
is desirable, protocols should be used with caution in patients benet with NIV post-extubation in patients with COPD.
with AHRF (Grade B). Evidence statement
68. The use of computerised weaning cannot be recommended The use of NIV as rescue therapy for unexpected post-
in AHRF (Grade D). extubation respiratory failure does not improve outcome and
may be detrimental (Level 1+).
Use of NIV in the ICU Recommendations
Planned NIV to speed weaning from IMV 72. NIV should not be used routinely for unexpected post-
In an uncontrolled study on lung transplantation, NIV was extubation respiratory failure (Grade B).
found to speed extubation and reduce the time spent invasively 73. In COPD, a trial of NIV may be justied for unexpected
ventilated and the attendant complications.249 Subsequent post-extubation respiratory failure where local expertise exists
studies have compared the use of NIV with conventional (Grade D).
weaning in patients who have failed an SBT. Benet was demon-
strated in patients with underlying COPD.50 250 These studies CARE PLANNING AND DELIVERY OF CARE
utilised NIV at high levels of PS and for longer than 24 h. NIV Appropriate care environments for the delivery of NIV
weaning was reported to confer no benet in a subsequent A study by Roessler and colleagues from Germany randomised
study.251 A subsequent Cochrane review concluded that the use 51 patients to either out-of-hospital NIV or standard medical
of NIV to speed weaning in patients with COPD reduced mor- treatment. Out-of-hospital NIV was reported to be feasible, safe
tality and the incidence of pneumonia without increasing the and effective.255 A survey of French mobile ICUs also suggests
need for re-intubation.252 that NIV and CPAP can be safely employed pre-hospital in acute
Evidence statements cardiogenic pulmonary oedema but not in other causes of
NIV has been shown to accelerate weaning from IMV in the respiratory failure.256 Further evaluation of out-of-hospital NIV
patient with COPD failing an SBT (Level 1+). in AHRF is required.
Recommendation NIV is commonly initiated in the ED, but given the other pri-
69. NIV is recommended to aid weaning from IMV in patients orities and pressures on emergency resuscitation areas patients
with AHRF secondary to COPD (Grade B). should be transferred as soon as practicable to an environment
70. In other causes of AHRF, NIV may have a role in shortening appropriately staffed and equipped to provide on going care. A
the duration of IMV when local expertise in its use exists (and prospective observational study of 245 patients attending 24
of cough assist when indicated) and there are features present hospital EDs in Australia identied the staff responsible for NIV
that indicate extubation is likely to be successful (Grade D). set-up.257 This was equally distributed between nursing and
medical personnel. Hess et al258 conducted a survey of 132 aca-
NIV in high-risk patients demic EDs in the USA, and concluded that, although NIV was
NIV has been assessed in patients who have passed an SBT but widely available, physician condence/competence was a barrier
who have risk factors for extubation failure such as age to optimal application. A survey of NIV use in UK EDs found a
>65 years, poor cough, cardiac and respiratory comorbidity, wide variety of practice and suggested the need for a specic
and hypercapnia (while ventilated and/or pre-existing). NIV was ED guide for NIV.259 A pro forma-based COPD management
reported to reduce the re-intubation rate and mortality in one tool, supported by targeted education, was reported to improve
study,253 and has been reported to be effective where obesity ED care including the use of NIV.260
(BMI >35 kg/m2) is an additional adverse feature.254 Previous guidelines have recommended limiting the number
Evidence statements of areas providing NIV to ensure that staff perform it suf-
NIV may be effective in reducing respiratory failure, ciently regularly.48 Suitable sites need to be able to provide an
re-intubation and mortality in COPD (Level 1 +) and patients NIV service 24/7 and integration with critical care services is
with increased BMI (Level 2+). essential. If NIV is provided in more than one area within a hos-
Planned post-extubation NIV reduces mortality, ICU and hos- pital, protocols and guidelines should be shared.261 Current
pital length of stay and the incidence of ventilator-associated NICE Quality Standards for COPD include guidance on how to
pneumonia (Level 1). benchmark NIV provision.262 The requirements for an NIV
Recommendation service are summarised in box 3.
71. Prophylactic use of NIV should be considered to provide For all but the mildest cases, Nava and Hill263 recommend
post-extubation support in patients with identied risk factors that NIV be delivered in a level 2 facility with enhanced stafng
for extubation failure (Grade B). levels. A survey carried out in 1999 found that NIV was pro-
vided in level 2/3 facilities in most western European coun-
NIV as rescue therapy post-extubation tries.264 In contrast, NIV has been delivered in admission or
A number of RCTs have examined the use of NIV as an respiratory ward settings in the UK. This may partially account
unplanned rescue treatment for post-extubation respiratory dis- for the poor performance and high mortality rates associated
tress. One multicentre RCT reported that patients who passed with use of NIV reported by audits.35 The 2009 ICS recom-
an SBT but who then developed post-extubation respiratory mendations265 reiterate level 2 as the appropriate clinical envir-
failure had an increased ICU mortality rate if treated with NIV onment for NIV and the 2008 joint BTS, ICS and RCP guide
as opposed to re-intubation.181 This study has been criticised as on the use of NIV in COPD with AHRF266 recommends one
few patients were treated in each participating centre, despite a nurse for every 2 NIV cases, especially during the rst 24 h of
long recruitment period, raising the suspicion that lack of famil- treatment.
iarity with NIV may have resulted in it being poorly applied. Despite this, in the 2013 BTS NIV audit, 91% of patients
The patients who failed NIV and went on to require intubation were treated on general medical wards despite 43% having
also received long periods of ineffective NIV before pH<7.25.5 This was associated with a low intubation rate and
Davidson AC, et al. Thorax 2016;71:ii1ii35. doi:10.1136/thoraxjnl-2015-208209 ii25
BTS guidelines
Figure 3 The three phases of patient management in acute hypercapnic respiratory failure.
ii26 Davidson AC, et al. Thorax 2016;71:ii1ii35. doi:10.1136/thoraxjnl-2015-208209
BTS guidelines
As discussed in the Management section, patient outcomes 77. Initial care plans should should include robust arrangements
reported in UK national audits are notably worse than would for escalation, anticipating that up to 20% of AHRF cases should
be expected from trial data and facilities for provision of NIV, be managed in a level 2 or 3 environment (Grade C).
and evidence of consultation with the ICU, are frequently Good practice points
limited or inadequate.3 4 7 Important deciences that have A 24 bedded designated NIV unit (located within a medical
been identied include delays in commencing ventilatory high dependency area or within a respiratory ward with
support, under-recognition of more complex acidbase distur- enhanced stafng levels) provides a robust basis for the provi-
bances, use of inadequate ventilatory pressures, rare use of a sion of NIV in a DGH serving a population of 250 000 and
different mask when NIV is failing, lack of progression from with an average prevalence of COPD.
NIV to IMV and lack of consultation in decision-making. The Areas providing NIV should have a process for audit and
preponderance for application of NIV in lower level facilities interdisciplinary communication.
than elsewhere in the world (outlined above) along with evi-
dence of a lack of integration with the ICU,3 4 indicates that
attention directed at organisational factors are needed and are Palliative care and advanced care planning
highly likely to improve patient outcome and experience in It is recognised that palliative interventions may be appropriate
AHRF. and yet be provided at the same time as therapies intended to
NIV facilities need to encompass adequate capacity, and the prolong life.272 Accordingly, employing NIV as part of care that
expertise and associated stafng levels, to deal with complex aims to relieve distress and has escalation limits may be entirely
critically ill patients who have a signicant risk of death. To be justied.
effective, the NIV service needs to have good operational links Effort is needed to establish patient preferences with respect
to the ICU in the expectation that 10% to 20% of NIV-treated to intubation or resuscitation status. Momen et al,273 in a sys-
patients should be managed in HDU/ICU and that many will be tematic review of end-of-life conversations in COPD, found
potential candidates for IMV. The case for a specically identi- considerable variation among patients in the desire to discuss
ed and appropriately staffed and equipped area for providing end of life. Almost 50% of patients did not wish to have such a
NIV is strongly supported by the evidence. In some European conversation and there was a preference to wait until the disease
countries, NIV services are provided in a Respiratory was advanced, with patient perception that this implied the
Intermediate Care Unit.264 last few days of life. Advance directives/living wills assist health-
Evidence statements care providers in tailoring clinical response and support.274 The
A care environment with either level 2 or 3 stafng favours a importance of actively involving patient/family, especially
successful outcome from NIV therapy (Level 2+). regarding do not attempt cardiopulmonary resuscitation
Coordination between the ICU and ward areas improves (DNACPR) orders, are highlighted in revised recommendations
outcome in AHRF (Level 3). following a judicial ruling.275 The essential element is that,
Organisational aspects are pivotal in achieving best outcomes while patients cannot insist on CPR being performed, the
(Level 4). matter should be discussed. Perceived patient distresswhich
Recommendations might be exacerbated by such discussionis no longer regarded
74. NIV services should operate under a single clinical lead with sufcient grounds for not raising the issue. When the risk of
formal working links with the ICU (Grade D). causing physical or psychological harm is present, attempts
75. The severity of AHRF, and evidence of other organ dysfunc- should be made to talk to a healthcare advocate. The enormous
tion, should inuence the choice of care environment (Grade C) . challenges in this serve to emphasise the crucial nature of active
76. NIV should take place in a clinical environment with and ongoing communication strategies. Chakrabarti reported
enhanced nursing and monitoring facilities beyond those of a interviews with 50 patients with stable COPD and found that
general medical ward (Grade C). discussion and demonstration of NIV equipment altered future
treatment perceptions and willingness to consider an advance
directive.276
Sinuff et al277 reported clinician attitudes to NIV in patients
with acute respiratory failure and do not intubate/do not resusci-
Box 4 Discharge checklist after AHRF tate instructions. While about 60% of physicians considered
that NIV should be discussed in this context, 85% of respiratory
therapists (those actually administering NIV) felt NIV should be
Arrange early specialist review, pulmonary rehabilitation and
actively promoted. This may reect a lack of condence and
help with smoking cessation as indicated.
understanding, among physicians, of the potential for NIV to
Consider early home visit, for example, outreach COPD
relieve distress and be effective even in advanced disease. In
team/community nurses
Denmark, 15% of patients with do not intubate instructions,
Discuss future care planning with patient/family and inform
and who received NIV, survived at least a year with COPD and
community services of the result of such discussion
congestive heart failure the most favourable underlying
Provide warning card/inform ambulance services regarding
diagnoses.278
future need for controlled oxygen therapy
Evidence statements
Consider referral to home NIV service, for example,
In advanced disease, care planning should ideally predate acute
neuromuscular disease (NMD) cases or suspected sleep
presentation or commence as early as possible on presentation
disordered breathing
with AHRF (Level 4).
Review reasons/route of admission and consider methods to
Health professionals experienced in NIV delivery have a more
improve if these were problematic
positive view of the benet of NIV and perceive patient treat-
Learn from any identied mistakes through multiprofessional
ment wishes more postively than do clinicians with less experi-
review
ence of NIV (Level 4).
Davidson AC, et al. Thorax 2016;71:ii1ii35. doi:10.1136/thoraxjnl-2015-208209 ii27
BTS guidelines
Recommendations The authors indicated that the use of sedatives, anxiolytics and
78. Clinicians delivering NIV or IMV should have ready access opiates could have been improved, emphasising that palliative
to palliative medicine (Grade D). care training or support is needed to achieve best practice.
79. Multidisciplinary advance care planning should be an inte- In conclusion, the role of NIV in achieving a good death
gral part of the routine outpatient management of progressive may currently be underutilised and there may be a lack of
or advanced disease and care plans should be reviewed on pres- appreciation that a peaceful death can occur while receiving sup-
entation during an episode of AHRF (Grade D). portive ventilation.
80. The use of NIV may allow time to establish patient prefer- Evidence statements
ence with regard to escalation to IMV (Grade D). The concerns of patients with COPD towards their end of life,
centre on high levels of physical symptoms, especially breath-
End-of-life care lessness (Level 3).
A questionnaire study of 118 patients with COPD, carried out in Clinicians often consider withdrawal of assisted ventilation
Canadian teaching hospitals,279 reported that patients with (NIV/IMV) as more challenging than removal of other life
COPD were less interested in prognosis, CPR, IMV or referral to support techniques (Level 4).
palliative care than were patients with metastatic cancer. In Good practice points
another study, comparing QoL between patients with advanced Although removal of the NIV mask may be deemed as pref-
COPD and patients with cancer, patients with COPD reported erable, a dignied and comfortable death is possible with it
higher levels of physical discomfort with uncontrolled shortness in place.
of breath in 78%.280 A recent review of the 4 RCTs that have Clinicians delivering NIV or IMV should have training in
explored whether NIV relieves dyspnoea in AECOPD concluded end-of-life care and the support of palliative care teams.
that benet was likely but that study limitations constrained a
condent conclusion.281 With regard to physical symptoms,
breathlessness and fatigue are dominant in AECOPD. Attention NOVEL THERAPIES
to secretion clearance is an additional major concern in bronchi- Extracorporeal CO2 removal
ectasis and CF and for many with NMD. Ability to communicate, The technical aspects of providing prolonged extracorporeal
to feel safe, to be individually respected and enabled to retain membrane oxygenation (ECMO) or CO2 removal (ECCO2R)
control are common psychological needs. have advanced in recent years. Both are being increasingly inves-
Patients receiving NIV as ceiling care, who fail to improve tigated in refractory respiratory failure including AHRF. NICE
will need appropriate end-of-life attention, including appropri- has issued general guidance on the use of ECCO2R advising
ate sedation/relief of distress. It is important that if withdrawal that it should only be used in patients with potentially reversible
of NIV is decided on, that this is achieved with minimum dis- hypercapnic respiratory failure or those being considered for
tress to the patient and their relatives. The BMA guidance on lung transplantation.290
end-of-life care in 2007 did not address withdrawal of assisted ECCO2R uses a gas exchange membrane to provide partial
ventilation.282 Although withholding and withdrawing are con- CO2 clearance, from 30% to 50% of the bodys production,
sidered ethically equivalent,283 for many individuals, including depending on blood ow and membrane efciency. Removing
clinicians, discontinuing mechanical ventilation is felt be emo- carbon dioxide extracorporeally reduces the native pulmonary
tionally different to, for instance, stopping haemodialysis. This minute ventilation required to maintain an acceptable PaCO2.
may be because of the immediacy of the consequence.284 A This offers the potential benets of either enabling protective
Japanese study reported interviews with 35 critical care physi- mechanical ventilation or providing an alternative to mechanical
cians and found withdrawing ventilation was regarded differ- ventilation in selected patients (such as those with COPD).
ently to stopping other life-sustaining measures because of There is little evidence of clear benets to patients of ECCO2R
concern over an abrupt and distressing demise.285 There was a at present. In moderately hypoxaemic ARDS, one RCT demon-
desire for a soft landing, with a slow and gradual death per- strated that lower tidal volumes and ventilatory pressures could
ceived as natural. The ATS Clinical Policy Statement of 2007 be achieved, but this failed to translate into a meaningful
provides comprehensive guidance on withdrawal of mechanical improvement in patient outcome.291 Larger studies are planned
ventilation, including symptom management of the dying in the UK and Europe. In patients with COPD, there are no
patient. It emphasises that decision-making is a process requiring RCTs exploring the role of ECCO2R. One retrospectively
frequent discussion with patient, family, health professionals matched cohort study compared outcomes between groups of
and others.286 Pro-active family-centred conferences allow time patients with AECOPD who had an inadequate response to NIV.
for families to adjust and provision of literature on bereavement Twenty ve potential candidates who had failed to improve with
reduces the risk of subsequent emotional morbidity.287 NIV were compared with historical controls treated in the same
In practical terms, progressive reduction of pressure/backup hospitals matched by the GenMatch process. Despite signicant
rate to achieve CO2 narcosis/coma and an alternative strategy of improvements in acidosis and respiratory distress, the trial failed
extubation when intubated or removal of NIV have both been to show benet in the primary outcome of need for intubation.
described. In the former scenario, Cox et al288 suggest initial The complication rate with ECCO2R was high (52%) and this
weaning of oxygen over 10 min with appropriate adjustment to contributed to the need for intubation.292
opiate or anxiolytic medication. Once patient comfort is The devices available for ECCO2R have evolved over time.
assured, it is suggested that mandatory ventilation is withdrawn Early CO2 removal membranes were pumpless, required arterial
and PS reduced to zero over 510 min. Kuhnlein et al289 con- and venous cannulation and used the patients own cardiac
ducted structured interviews with 29 families regarding the cir- output to drive blood through the membrane. This resulted in
cumstances of dying in MND patients receiving NIV. Seventeen signicant shunting of cardiac output and the danger of limb
caregivers described the nal stages and eventual death as ischaemia. An alternative approach is to take blood from a dual-
peaceful. Eleven of the patients died peacefully while using lumen large bore cannula sited in a central vein and pump it
NIV. Choking sensation was evident in some bulbar patients. through the membrane. The advantages of the veno-venous
ii28 Davidson AC, et al. Thorax 2016;71:ii1ii35. doi:10.1136/thoraxjnl-2015-208209
BTS guidelines
12
technique are principally lack of effect on cardiac output and University of Plymouth, Plymouth, UK
13
reduced complications, particularly limb ischaemia. Department of Physiotherapy, Guys and St Thomas NHS Foundation Trust, St
Thomas Hospital, London, UK
Evidence statements 14
Aintree University Hospital, Liverpool, UK
Extra-corporeal CO2 removal devices can reduce PaCO2 and 15
CLAHRC, Imperial College, London, UK
16
minute volume (Level 2). Queens Medical Centre, Nottingham, UK
17
Veno-venous extra-corporeal CO2 removal in patients with Royal College of Physicians, London, UK
AECOPD and an inadequate response to NIV has not been
shown to reduce intubation rate and is associated with a 52%
Twitter Follow BTS at @BTSrespiratory, Ben Creagh-Brown at @SPACeRGroup and
complication rate (Level 2). James Dodd at @theotherdodd
Recommendations
Acknowledgements The authors acknowledge assistance with developing the
81. If local expertise exists, ECCO2R might be considered: aims and structure of the guidelines provided by Martin Wildman, Simon Baudouin,
If, despite attempts to optimise IMV using lung protective Robert Winter, Sara Bolton, Mandy Odell, David Smith and the members of the BTS
strategies, severe hypercapnic acidosis ( pH <7.15) persists Standards of Care Committee.
(Grade D); Collaborators BTS Standards of Care Committee members: Colin Gelder, James
When lung protective ventilation is needed but hypercapnia Dodd and Caroline Patterson.
is contraindicated, for example, in patients with coexistent Contributors ACD chaired the guideline group, and led the drafting and revision
brain injury (Grade D); of the document. He has nal responsibility for the guideline. SB, ME, DK, CG and
For IMV patients awaiting a lung transplant (Grade D). AG drafted and revised the paper. CC, BC-B, JD, TF, BF, LM, LMcD, RP, CP, MS and
LT conducted appraisal of the literature and provided draft sections of the document.
Good practice point
ECCO2R is an experimental therapy and should only be used Representation Dr Bob Winter and Sara Bolton represented the Intensive Care
Society, Dr Bernard Fox represented the College of Emergency Medicine, Dr Daniel
by specialist intensive care teams trained in its use and where Kennedy represented the Royal College of Anaesthetists and Surgeon Captain Lynn
additional governance arrangements are in place or in the Thomas represented the Royal College of Physicians.
setting of a research trial. Competing interests ACD declares being paid as a consultant to Smith Medical
between 2008 and 2013. ME declares he has received an honorarium, and travel
Helium/oxygen ventilation and subsistence expenses for speaking at a meeting in Australia organised by
When mixed with oxygen (Heliox), the lower density of helium Resmed, a Respiratory Sleep and Ventilation company. He has received an
honorarium and travel expenses for speaking at a meeting in London organised by
reduces resistance in the large airways where ow is predomin-
Phillips Respironics, a Respiratory Sleep and Ventilation company. He has received
antly turbulent compared to air/oxygen ventilation. It therefore travel and subsistence expenses for speaking at a meeting in China organised by
has a theoretical advantage in obstructive causes of AHRF.293 Curative Medical Inc, a Respiratory Sleep and Ventilation company. He has received
Heliox has been studied in combination with both NIV and travel expenses for speaking at meetings in India organised by Phillips Respironics, a
IMV. It increases the delivered dose of bronchodilators and has Respiratory Sleep and Ventilation company. AG declares being paid as a consultant
and receiving honoraria and travel expenses for speaking at meetings organised by
been reported to improve symptoms and physiological variables Armstrong Medical Ltd in the UK, between 2014 and 2015.
in spontaneously breathing asthmatics.294 295 At oxygen concen-
Provenance and peer review Not commissioned; internally peer reviewed.
trations >40%, Heliox has no benet compared with oxygen
air mixtures.296 A large RCT in AECOPD found that Heliox in
combination with NIV did not reduce rates of intubation, dur- REFERENCES
ation of ventilatory support or mortality.297 Heliox has been 1 British Thoracic Society Standards of Care Committee. Non-invasive ventilation in
reported to reduce pCO2 and airway pressures in intubated acute respiratory failure. Thorax 2002;57:192211.
2 Plant PK, Owen J, Elliott MW. One year period prevalance study of respiratory
patients with severe asthma298 but a subsequent meta-analysis acidosis in acute exacerbation of COPD; implications for the provision of
concluded that it did not affect outcome.299 An uncontrolled non-invasive ventilation and oxygen administration. Thorax 2000;
study reported that Heliox improved patient comfort in the 55:5504.
presence of post-extubation respiratory distress when stridor 3 Davidson AC. Report British Thoracic Society NIV Audit 2011. https://www.
was present.300 brit-thoracic.org.uk/document-library/audit-and-quality-improvement/audit-reports/
bts-adult-niv-audit-report-2011/
Evidence statement 4 Davies M. Report British Thoracic Society NIV Audit 2012. https://www.
The use of Heliox does not reduce rates of intubation and brit-thoracic.org.uk/document-library/audit-and-quality-improvement/audit-reports/
length of IMV, nor does it reduce mortality in patients of bts-adult-niv-audit-report-2012/
AECOPD or asthma (Level 1+). 5 Davies M. Report British Thoracic Society NIV Audit 2013. https://www.
brit-thoracic.org.uk/document-library/audit-and-quality-improvement/audit-reports/
Recommendation bts-adult-niv-audit-report-2013/
82. Heliox should not be used routinely in the management of 6 Davidson AC. Towards a comprehensive ventilatory strategy for acute exacerbations
AHRF (Grade B). of COPD. JICS 2008;9:57.
7 Roberts CM, Stone RA, Buckingham RJ, et al. Acidosis, non-invasive ventilation
and mortality in hospitalised COPD exacerbations. Thorax 2011;66:438.
Author afliations 8 Gupta D, Keogh B, Chung KF, et al. Characteristics and outcome for admissions to
1
British Thoracic Society, London, UK adult, general critical care units with acute severe asthma: a secondary analysis of
2
St Jamess University Hospital, Leeds, UK the ICNARC Case Mix Programme Database. Crit Care 2004;8:R11221.
3
Barts Health NHS Trust, London, UK 9 Hull J, Aniapravan R, Chan E, et al. British Thoracic Society guideline for
4
Department of Respiratory Research, University Hospitals of Coventry and respiratory management of children with neuromuscular weakness. Thorax
Warwickshire NHS Trust, Coventry, UK 2012;67(Suppl 1):i140.
5
Shefeld Teaching Hospitals NHS Foundation Trust, Shefeld, UK 10 British Thoracic Society. BTS guideline production manual. London: BTS, 2014.
6
Department of Respiratory, Scottish Pulmonary Vascular Unit, Glasgow, UK 11 AGREE. AGREE II Instrument. Secondary AGREE II Instrument 2010. http://www.
7
Royal Surrey County Hospital NHS Foundation Trust and Faculty of Health and agreetrust.org/agree-ii/
Medical Sciences, University of Surrey, Guildford, UK 12 Mehta S, McCool FD, Hill NS. Leak compensation in positive pressure ventilators:
8
Academic Respiratory Unit, University of Bristol, Bristol, UK a lung model study. Eur Respir J 2001;17:25967.
9
North Bristol Lung Centre, North Bristol NHS Trust, Bristol, UK 13 Storre JH, Bohm P, Dreher M, et al. Clinical impact of leak compensation during
10
University Hospital of South Manchester NHS Foundation Trust, Manchester, UK non-invasive ventilation. Respir Med 2009;103:147783.
11
Emergency Department, Manchester Royal Inrmary, Central Manchester University 14 Ferguson GT, Gilmartin M. CO2 rebreathing during BiPAP ventilatory assistance.
Hospitals NHS Foundation Trust, Manchester, Manchester, UK M13 9WL Am J Respir Crit Care Med 1995;151:112635.