RESEARCH ARTICLE

Does the Addition of Non-Invasive Ventilation during

Pulmonary Rehabilitation in Patients with Chronic
Obstructive Pulmonary Disease Augment Patient
Outcome in Exercise Tolerance? A Literature Review
Eve Corner, BSc (Hons) MCSP1,* & Rachel Garrod, PhD MSc MCSP2
1
Chelsea and Westminster Hospital Foundation Trust, London

2
Dulwich Hospital, East Dulwich Grove, London

Abstract
Background. Non-invasive ventilation (NIV) during exercise in patients with chronic obstructive pulmonary
disease (COPD) has been shown to increase exercise time and intensity. Feasibly then, NIV during pulmonary
rehabilitation will enhance post-rehabilitation training effects. The purpose of this review is to systematically con-
sider and critique the literature concerning the effects of NIV, when used during an exercise programme in COPD
patients on exercise tolerance. Method. An electronic literature search was completed and the reference lists of
the articles that fitted the following inclusion criteria were screened. Studies that used any mode of NIV during an
exercise programme with a primary outcome measure focusing on exercise tolerance and were written after the
year 2000 to reflect current practice. Studies that were not written in English or had been included in previous
literature reviews were excluded. The studies were then critically appraised and assigned a level of evidence based
upon Scottish Intercollegiate Guidelines Network. Results. Twenty-eight articles were screened, of which six fitted
the inclusion criteria. The methodological quality ranged from level 1− to 1+. All but one study by Bianchi et al.
(2002) demonstrated a statistically significant improvement in exercise tolerance with the addition of some form
of NIV during pulmonary rehabilitation. The benefits may be greater in patients with more severe airway obstruc-
tion as determined by Forced Expiration Volume (FEV), (%pred). Conclusion. This review would suggest that
NIV may allow an increased exercise intensity and duration during pulmonary rehabilitation in patients with
moderate to very severe COPD, (according to the American Thoracic Society guidelines), resulting in a greater
training effect and a prolonged exercise capacity. Further research looking at long-term follow-up is recommended.
Copyright © 2009 John Wiley & Sons, Ltd.

Received 9 July 2009; Accepted 24 October 2009

Keywords
COPD; non-invasive ventilation; pulmonary rehabilitation

*Correspondence
Eve Corner BSc (Hons) MCSP, Chelsea and Westminster Hospital Foundation Trust, London.
Email: evelyn.corner@chelwest.nhs.uk

Published online 23 December 2009 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/pri.451

Physiother. Res. Int. 15 (2010) 5–15 © 2009 John Wiley & Sons, Ltd. 5
Does NIV during Pulmonary Rehabilitation Improve Exercise Tolerance in COPD?
Introduction
Chronic obstructive pulmonary disease (COPD) is
increasingly becoming recognised as a systemic disease,
principally defined by irreversible airway obstruction.
It results in debilitating breathlessness, muscle wasting
and reduced exercise tolerance (Wooters et al., 2002;
Wooters, 2002). Pulmonary rehabilitation is an effec-
tive strategy in the management of COPD (Ries, 2007).
By training a patient at a moderate to high intensity
during pulmonary rehabilitation, statistically signifi-
cant improvements in cardiovascular fitness, muscle
strength and HRQofL (Health related quality of life) are
frequently demonstrated (Ambrosino, 2006; Ries et al.,
2007; Casaburi et al., 1991). However, exercising at this
intensity can be difficult for the moderate to severe
COPD patient because of dyspnoea.
In COPD, expiratory airflow is limited by obstruc-
tion leading to an increase in functional residual capac-
ity (FRC) termed hyperinflation. This has a detrimental
effect on the length–tension relationship of the dia-
phragm and the alveoli by preventing full elastic recoil
at end expiration (De Troyer, 1997). This creates an
above atmospheric intrapulmonary end expiratory
pressure (intrinsic positive end-expiratory pressure
[PEEPi]). As a result of PEEPi, inspiratory airflow will
only be generated when a negative pressure is created
overriding that of the PEEPi (Ranieri et al., 1996). The
overall effect is to increase work of breathing.
During exercise, these effects are magnified because
of a greater oxygen demand. Respiratory rate is
increased leading to a reduction in expiratory time and
frequent shallow breaths. As a result, the FRC and expi-
ratory muscle activity increase. The combined effect is
a higher PEEPi and dynamic hyperinflation. This creates
severe dyspnoea on minimal exertion (Polkey et al.,
1996), preventing patients from reaching the desired
intensity of training and limiting the exercise effect.
Ranieri et al. (1996) suggest that the use of extrinsic
PEEP (PEEPe), for example NIV, can overcome the
effects of PEEPi. It does this by reducing the pressure
differential between the alveoli and the mouth on end
expiration, thus limiting the demand on the inspiratory
muscles. Keilty et al. (1994) demonstrated a similar
effect with inspiratory pressure support. They con-
cluded that off-loading the inspiratory muscles during
physical activity allows for a longer exercise time.
In 2002, van’t Hul et al. completed a systematic
review of the literature investigating the immediate
6 Physiother. Res. Int. 15 (2010) 5–15 © 2009 John Wiley & Sons, Ltd.
E. Corner and R. Garrod
effects of NIV on exercise endurance and dyspnoea.
Their conclusion showed possible increases in immedi-
ate exercise intensity and duration with NIV in situ
during a one off exercise session. Their discussion iden-
tified a further need to research the additional benefits
of NIV as an adjunct to pulmonary rehabilitation on
long-term exercise tolerance. The hypothesis would
then follow that if the duration and intensity of each
individual exercise session is increased with NIV, then
the magnitude of training-induced effects would be
greater post rehabilitation. The aim of this current
review therefore was to critically appraise recent studies
concerned with the use of NIV during a rehabilitation
training programme. This review complements and
adds additional data to that of van’t Hul et al. (2002)
and is required in light of current practice.
Method
Literature review
An electronic literature search was completed, using the
following search strategy: Exercis* OR pulmonary reha-
bilitation AND COPD AND BiPap OR NIV OR CPAP
OR non-invasive ventilation OR PAV OR proportional
assist ventilation. The following search engines were
used: Medline (OVID) NCBI (pubmed); the Cochrane
central register of controlled trials and systematic
reviews; the National Health Service (NHS) database.
Two thousand six hundred sixty-six titles were identi-
fied and their abstracts were screened for relevance
using the following inclusion criteria: any type of NIV
(i.e. continuous positive airway pressure, NIV, propor-
tional assist ventilation [PAV], bilevel positive airway
pressure [BiPAP] or inspiratory pressure support [IPS])
used during an exercise programme or pulmonary
rehabilitation where the primary outcome is concerned
with exercise tolerance/time. Titles and abstracts were
excluded using the following criteria: articles not
written in English, articles published before 2000 and
were previously included in a literature review. The full
papers of 26 abstracts were retrieved and six articles
were then included in the full review.
Methodological analysis
The articles were critiqued using the Scottish Intercol-
legiate Guidelines Network (SIGN) methodological
checklists (SIGN, 2009a) which enable the categoriza-
tion of research studies into levels of evidence (Table
E. Corner and R. Garrod Does NIV during Pulmonary Rehabilitation Improve Exercise Tolerance in COPD?
Table 1. SIGN levels (for RCTs)
1++ High quality meta-analysis, systematic reviews of
randomized controlled trials (RCTs), or RCTs with a very
low risk of bias.
1+ Well-conducted meta-analyses, systematic reviews or RCTs
with a low risk of bias.
1− Meta-analyses, systematic reviews or RCTs with a high risk
of bias
2++ High quality systematic reviews of case control or cohort
studies.
2+ Well-conducted case control or cohort studies with a low
risk of confounding or bias with a moderate possibility
that the relationship is causal.
2− Case control or cohort studies with a high risk of
confounding or bias with a significant risk that the
relationship is not causal.
3 Non-analytic studies e.g. case reports or case series.
4 Expert opinion.
SIGN = Scottish Intercollegiate Guidelines Network; RCT = random-
ized control trial.
1). The strength and validity of the study can then be
evaluated.
SIGN’s objective is to ‘create national clinical guide-
lines containing recommendations for effective practice
based on current evidence’ (SIGN 2009b). Study details
are described in Table 2.
Results and discussion
Ventilation modes and settings
The studies critiqued focus on three different modes of
ventilation: BiPap, PAV and IPS. Table 3 shows each
mode and setting of ventilation used in each study.
Bianchi et al. (2002) used PAV (Respironics BiPap
Vision). Participants were randomised into training
unsupported or with PAV. The method of randomiza-
tion was not stated and was completed after their initial
exercise testing leaving the process open to bias. They
titrated the ventilator to support the subjects at 80% of
their maximum tolerable flow assist and volume assist
(Table 3). Bianchi et al. (2002) concluded ‘no addi-
tional physiological benefit in comparison with exercise
training alone’.
This result is in contrast to Hawkins et al. (2002) who
also used PAV via the same make of ventilator. Their
subjects were randomised via sealed envelopes into
either training unsupported or with PAV. Hawkins
et al. (2002) again titrated the flow and volume assist
to 80% of the patient’s highest tolerated pressures
(Table 3). However, they fail to mention whether
Physiother. Res. Int. 15 (2010) 5–15 © 2009 John Wiley & Sons, Ltd.
expiratory positive airway pressure (EPAP) was used
making full comparison to Bianchi et al.’s study (2002)
difficult. It was concluded that ‘PAV enables a higher
intensity of training in patients with severe COPD,
leading to greater improvements in maximum exercise
capacity with evidence of true physiological adaptation’
(Hawkins et al., 2002).
This varying result may be explained by the ventila-
tor settings. In the Hawkins et al. study (2002), the
mean volume assist was nearly double that of Bianchi
et al.’s (2002); however, the flow assist setting was
similar. The differences in ventilator settings may have
to do with greater disease severity in Hawkins et al.
(2002) cohort (see Table 3). PAV is intended to supply
pressure support proportional to patient effort. The
flow assist setting is designed to off-load the resistive
component of breathing and the volume assist the
elastic component (Appendini, 2003). Patients with
more severe airway impairment will suffer from
a greater degree of dynamic hyperinflation and,
subsequently, a greater PEEPi (elastic component).
Therefore, a larger volume assist would be required
to override the positive intrapulmonary pressure.
van’t Hul et al. (2006) used IPS during a cycling
exercise regime. This study is the only study to use IPS
and a sham intervention. Patients were randomised
using sealed envelopes into the sham group (IPS at
pressures of 5 cmH2O) or the experimental group (IPS
at pressure of 10 cmH2O). The IPS machine was a flow
triggered pressure controlled ventilator (see Table 3).
Significantly greater improvements were shown in the
incremental shuttle walk and the constant load cycle
endurance test in IPS10 group compared with the IPS5
group (p < 0.05). Inspiratory pressure support of
10 cmH2O may therefore be considered an effective
adjunct during high intensity training. van’t Hul et al.
(2006) and Bianchi et al. (2002) are the only studies in
this review that report intention to treat analysis further
validating their findings.
Comparison between van’t Hul et al. (2006) and
Bianchi et al.’s (2002) studies is difficult because of the
different ventilators and modes. van’t Hul et al. (2006)
validated the use of IPS5 as a sham in a previous study
(van’t Hul et al. 2002) subjects completed exercise
endurance testing whilst breathing IPS5, IPS10 and no
IPS. There was a significant improvement in exercise
time whilst exercising with IPS10, but no significant
difference between IPS5 and no IPS (p < 0.001). This
data suggests that the use of pressure of 5 cmH2O may
7
8
Table 2. Study characteristics

Reference SIGN Number FEV1 % pred Patient Population Intervention NIV settings Outcome Measure Result (taken directly from paper)
level (completed)

Bianchi et al. 1− 33 (19) FEV1 47.7 + −18.7% Consecutive male Duration: 6 weeks FA 3.5 + −1.6 cm H20.1−1 Exercise tolerance ‘55% in the PAV group reached
(2002) pred in PAV stable COPD Frequency: Three 3-hour VA 6.6 + −2.2 cm H2O−1 (Symptom limited 80% of baseline peak workload
group, 40.1 + patients referred sessions EPAP of 2 cmH2O incremental exercise and 50% in the SB.’
−12.3% pred SB to an outpatient Setting: outpatient MDT testing and 6-minute ‘6MWD went from 488 + −80 to
group PRP from PRP. walking distance 504 + −83 m in the PAV group
January 1998 to Programme: 30 minutes Dyspnoea (BDI)) and from456 + −80 to 503 +
April 2000 in continuous cycling, all HRQ of L (SGRQ) −83 in the SB group (p =
Gussaco, Italy. body resistance exercises 0.002).’
and education. ‘55% in PAV group and 40% in
Intensity: at 50–70% the SB group showed a post
maximal load PRP increase in peak work load
Conditions: of more than 50% (p = 0.65)’
1) SB ‘TDI improvement significantly in
2) PAV both groups.’
‘22% in the PAV group and 50%
in the SB group increased
6MWD by over 54 m, an
estimated threshold of clinical
importance’ (Bianchi 2002)

Van ‘t Hul et al. 1+ 37 (29) FEV1 < 60% pred, COPD patients in Duration: 8 weeks IPS of 3 L/min and a Fi02 ISW In the first week, there was a
(2006) Pa02 > 8 KPa, Amsterdam aged Frequency: 45 minutes, of 0.21 experimental SGRQ significantly higher training
Does NIV during Pulmonary Rehabilitation Improve Exercise Tolerance in COPD?

peak MV < 50 L/ 40 to 75 years thrice a week
min with ventilatory Programme: Cycling
limited exercise calibrated cycle ergometer
capacity. Intensity: started at 65% of
Wmax for up to 15
minutes, then the
intensity was increased by
5% of Wmax.
Conditions: Randomised to
sham or experimental.
group Constant load cycle intensity increase in the IPS10
— IPS of 10 cmH2O cycle endurance test group.
sham group at 75% of Wpeak Incremental Shuttle Walking
— IPS of 5 cmH2O. distance was significantly higher
in the IPS10 group (31 + −21 m
Vs 14 + −31 m; p < 0.05). This
was also significant when
intention to treat analysis was
performed.
Cycle endurance time was longer
in the IPS10 group compared
with IPS5 (7.4 + −5.4 vs. 3.9 +
−6.0 minutes; p = 0.07)
No statistical change in SGRQ.

Physiother. Res. Int. 15 (2010) 5–15 © 2009 John Wiley & Sons, Ltd.
E. Corner and R. Garrod
 Hawkins et al. 1− 29 (19) (FEV1 26(7)% pred Patients with severe Duration: 6 weeks FA 3.6 + −0.7 cmH2O/l/s Venous plasma lactate PAV group were training at a
(2002) in PAV group COPD attending Frequency: 30 minutes, VA 12.7 + −1.5 cm H2O/l conc higher work rate at the end of
and 28(7)% pred a chest clinic at thrice week Maximal symptom the intervention.
in control.) Kings College Programme: Cycling, no limited incremental Mean peak work rate increased by
Hospital London education element to the cycle test (Wpeak) 14.5% and 32.0% in the
over 2 years. programme. CWR unassisted group and PAV
two women (both Intensity: 70% of peak work MV group respectively (p = 0.005)
in unassisted rate until symptom rating HR Mean exercise duration in the
E. Corner and R. Garrod

group) of 5 on the Borg scale was CWR test increased from 8.7 to
17 men achieved. Intensity was 17.2 minutes in the PAV group
increased progressively by and from 6.9 to 13.0 minutes in
5 W when able to sustain the unassisted group (p < 0.005
exercise for 30 minutes vs. baseline). No significant
Conditions: difference was found between
1) PAV the groups.
2) Unassisted Plasma lactate concentration was
reduced by 30% (p = 0.002) in
the assisted group and 11% (p =

Physiother. Res. Int. 15 (2010) 5–15 © 2009 John Wiley & Sons, Ltd.
0.08) in the unassisted group
versus baseline. (not significant)
Dyspnea and RR reduced in both
groups.

Reuveny et al. 1+ 24 (19) FEV1 < 40% pred Severe Stable COPD Duration: 8 weeks EPAP 2 cmH20, IPAP Incremental cycle BiPAP group; Exercise tolerance
(2005) patients at the Frequency: 45 minutes, 7–10 cmH2O ergometry increased
pulmonary twice a week according to patient plasma lactate acid ‘VO2 max increased by 23 + −16%
institute at Sheba Programme: Treadmill comfort concentration (p < 0.005)’
Medical Centre, training expiratory fraction of ‘AT increased by 11 + −12%
Israel. Intensity: 65–70% of their O2 and CO2 (p < 0.05)’
predetermined maximum Peak VO2 ‘Oxygen uptake per heartbeat
Conditions: Anaerobic threshold increased by 20 + −19%
1) BiPap (AT) (p < 0.05)’
2) Unassisted Lactate concentration did not
All had supplemental change
oxygen (sats > 92%) ‘Max MV increased by 17 + −20%
(p < 0.005)’
‘Vt max increased by 16 + −12%
(p < 0.005)’
No physiological improvements
reached significance in the
control group (Reuveny et al.
2005)

9
Does NIV during Pulmonary Rehabilitation Improve Exercise Tolerance in COPD?
10
Table 2. Continued

Reference SIGN Number FEV1 % pred Patient Population Intervention NIV settings Outcome Measure Result (taken directly from paper)
level (completed)
Johnson et al. 1− 41 (32) FEV1 < 50% pred Patients referred to Duration: 6 weeks IPAP 8–12 cm H20 and Maximal graded No immediate effect with Heliox.
(2002) PRP between Frequency: 20 minutes, EPAP 2 cmH2O exercise treadmill Immediate significant increase in
March 1996 and twice a week test with heliox, air maximal workload following
January 1997 Programme: Treadmill and NIPPV. training with NIPPV (>2.5
from a training and education minutes longer).
population of sessions All had a significant improvement
military Intensity: 50–60% of their in exercise time and workload
beneficiary’s in predetermined maximum after the 6-week programme
San Antonio, ‘TX Conditions: (p < 0.000001)
metropolitan 1) Unassisted training The NIPPV group had a
areas.’ 2) Humidified heliox 10 L/ significant improvement
min via non-rebreather compared to the unassisted
(79% helium/21% 02) group on immediate exercise
3) Nasal NIPPV with air time (p = 0.016)

Costes et al. 2− 14 (14) FEV1 31.9% + −9.2 Patients enrolled in Duration: 8 weeks EPAP increased with Incremental exercise ‘Significant increase in exercise
(2003) pred the pulmonary Frequency: 30 minutes, tolerance level from testing tolerance in both groups’
rehab training thrice a week 4–8 mmHg ‘CWR exercise: +81% (p < 0.05) in
programme in Programme: cycling IPAP as tolerated the control group and +116%
France Intensity: 60% of their peak (p < 0.001) in the NIV group.’
Does NIV during Pulmonary Rehabilitation Improve Exercise Tolerance in COPD?

V02 The between group difference

Conditions: in CWR was not significant.
1) Unassisted ‘The application of ventilatory
2) BiPAP support during exercise
significantly increased CWR
exercise duration by 72%
(p < 0.05)’

SIGN = Scottish Intercollegiate Guidelines Network; FEVI = Forced expiratory volume in 1 second; COPD = Chronic obstructive pulmonary disease; MDT = Multi-disciplinary team; NIV = Non-invasive ventila-
tion; FA = Flow assist; EPAP = Expiratory positive airway pressure; SB = Spontaneously breathing; PAV = Proportional assist ventilation; IPS = Inspiratory pressure support; BDI = Baseline Dyspnoea Index;
HRQ of L = Health related quality of life; SCRQ = St. George’s Respiratory Questionnaire; ISW = Incremental Shuttle Walk; 6 MWD = 6 Metre walking distance; TDI = Transitional Dyspnoea Index; CWR =
Constant work rate; MV = Minute Volume; HR = Heart rate; WMAX = Maximal workload; SATS = Oxygen saturations; PRP = Pulmonary rehabilitation programme; W = Watts; VA = Volume assist; NIPPV
= Non-invasive positive pressure ventilation.

Physiother. Res. Int. 15 (2010) 5–15 © 2009 John Wiley & Sons, Ltd.
E. Corner and R. Garrod
E. Corner and R. Garrod Does NIV during Pulmonary Rehabilitation Improve Exercise Tolerance in COPD?

Table 3. Summary of ventilator settings and modes breathing room air. The greater ventilatory response
resulted in greater tidal volumes, expired volumes and
Author Ventilation modes and settings
a lower peak end tidal carbon dioxide PetCO2. The
Bianchi et al. (2002) PAV: 6.6 ± 2.2 cmH2O−1 volume assist and validity of this study is enhanced through blinding of
3.5 ± 1.6 cmH2O−1 flow assist. The
researchers from patient group allocation. van’t Hul
ventilator had a built in EPAP of
2 cmH2O. et al. (2006) was the only other researcher to address
Costes et al. (2003) BiPap: EPAP 4–8 cmH2O, IPAP pressures this. van’t Hul et al. (2006) also blinded the participants
not stated to group allocation; however, having been pre-tested on
Hawkins et al. PAV: 3.6 ± 0.7 cmH2O/l/s flow assist and both IPS settings, the effectiveness of this may be
(2002) 12.7 ± 1.5 cmH2O/l volume assist. No
EPAP
questioned.
Johnson et al. (2002) BiPap: EPAP 2 cmH2O and an IPAP of 8 to Johnson et al. (2002) compared patients in a pulmo-
12 cmH2O, humidified heliox or air 10l/ nary rehabilitation programme whilst training on a
minute via and non-rebreath mask. treadmill with BiPap, humidified heliox or humidified
Rueveny et al. Bipap; EPAP of 2 cmH2O and an IPAP of 7
air. There was no acute effect of heliox but an immediate
(2005) to 10 cmH2O
van’t Hul et al. IPS: Sham: IPS 5 cmH2O improvement in exercise intensity and duration with
(2006) IPS: Experimental: 10 cmH2O BiPap. However, whether this lead to a greater training
PAV = Proportional assist ventilation; BiPap = Bilevel positive airway
effect at the end of the six week programme is unclear
pressure; EPAP = Expiratory positive airway pressure; IPAP = as regression to the mean errors may have skewed
inspiratory positive airway pressure; IPS = Inspiratory pressure findings.
support. Costes et al. (2003) used BiPap as an adjunct to a
pulmonary rehabilitation programme. They consecu-
Table 4. Mean FEV 1 (%pred) tively assigned patients to train either with NIV or
spontaneously breathing. Costes et al. (2003) fail to
Author Disease severity (FEV1 (%pred))
report the IPAP settings used. The EPAP (Inspiratory
Bianchi et al. (2002) PAV: 47.7 ± 18.7 FEV1 (% pred) positive airway pressure) is documented as 4 to
SB: 40.1 ± 12.3 FEV1 (% pred)
8 cmH2O, which is considerably higher than the other
Costes et al. (2003) Bipap: 31.3 ± 11.9 FEV1 (% pred) studies using BiPap, which may possibly lead to an
SB: 31.7 ± 6.5 FEV1 (%pred) increase in FRC as the EPAP is likely to exceed the
Hawkins et al. (2002) PAV: 26.0 ± 7.0 FEV1 (%pred) PEEPi (Ranieri et al., 1996). The lack of knowledge
SB: 28.0 ± 7.0 FEV1 (%pred) about the IPAP setting makes comparisons and repeti-
Johnson et al. (2002) SB: 30.7 ± 11.3 FEV1 (%pred) tion of the study impossible.
Heliox: 34.1 ± 12.8 FEV1 (%pred) Costes et al. (2003) showed an improvement in exer-
Bipap: 31.6 ± 9.3 FEV1 (%pred) cise tolerance whilst training with BiPap compared with
Rueveny et al. (2005) Bipap: 32.0 ± 4 FEV1 (%pred) spontaneously breathing; however, these differences
SB: 33.0 ± 9.0 FEV1 (%pred) were not significant. The study is a cohort study there-
van’t Hul et al. (2006) IPS10: 41.0 ± 10.0 FEV1 (%pred) fore lacking randomization, blinding and concealment
IPS5: 38.0 ± 9.0 FEV1 (% pred) and had the smallest study population.
FEV = Forced expiratory volume in 1 second; PAV = Proportional
assist ventilation; IPS = Inspiratory pressure support; SB = Spontane- Participants
ously breathing.
Disease severity (Table 4)
imply that the pressures used by Bianchi et al. (2002) Most participants for the studies are classed as
were insufficient to achieve a significant effect. ‘severe’ to ‘very severe’ according to the American Tho-
Reuveny et al. (2005) randomly assigned patients via racic Society (ATS) guidelines (ATS, 2009). Table 4
sealed envelopes into a treadmill training programme demonstrates FEV1 (%pred) of the cohort for each
with or without Bipap (see Table 3 for settings). They study. Only Bianchi et al. (2002) and van’t Hul et al.
concluded that Bipap ‘augmented training effect’ and (2006) included participants in the ‘moderate’ category.
resulted in a higher ventilatory response during exercise Bianchi et al. (2002) states that the intended population
when compared with exercising whilst spontaneously were mild COPD patients. However, none of their

Physiother. Res. Int. 15 (2010) 5–15 © 2009 John Wiley & Sons, Ltd. 11
Does NIV during Pulmonary Rehabilitation Improve Exercise Tolerance in COPD?
cohort fitted these criteria. Hawkins et al.’s (2002)
cohort is more severe than Bianchi et al.’s (2002), which
may explain why the addition of PAV showed a positive
outcome in this study. Exercise induced dynamic
hyperinflation and dyspnoea is likely to be more pro-
nounced in patients with severe airway disease and
therefore reduced to a greater extent with addition of
PAV. This may also explain why the tolerated ventilator
settings differed so greatly between to two groups.
It is beyond the scope of this paper to statistically
analyse the difference in FEV1 (%pred) between the two
studies. However, van’t Hul et al. (2006) found signifi-
cant improvement with IPS in a similar cohort to
Bianchi et al. (2002) who found no benefit with PAV.
While comparison is difficult because of the different
modes of ventilation, these conflicting results may indi-
cate that IPS is more effective for the moderate severity
COPD patient. Alternatively, these conflicting results
may be a result of lower pressures in Bianchi et al.’s
study (2002). Furthermore it should be noted that
patients from the study of Bianchi et al. (2002) stopped
exercising because of dyspnoea, and not leg fatigue,
supporting the fact that they had more severe ventila-
tory impairment.
Only Bianchi et al. (2002) and van’t Hul et al. (2006)
have acknowledged MRC scores/ATS stages as a guide
to disease severity. However, although there is a general
trend between MRC (Medical Research Council)
scores/ATS stages and FEV1 (%pred), a lower FEV1 (%
pred) may not necessarily correlate with poorer exercise
tolerance (Pineda et al., 1984). The Bianchi et al. (2002)
and van’t Hul et al. (2006) studies were also the only
two that included the St Georges Respiratory Question-
naire (SGRQ). Neither study was able to demonstrate
significant changes in the SGRQ score. This may be due
to the small sample sizes leading to a type 11 error
(Molken et al., 1999).
Recruitment and study population (Table 5)
For all studies, the attrition rate significantly impacted
on the sample size, and in all studies, the sample size was
relatively small. This leaves ambiguity over the power of
the studies to draw meaningful conclusions.
Johnson et al. (2002) had the largest cohort; however,
study participants are all ex-military which may reduce
its environmental validity. The studies were completed
internationally, which, whilst enabling generalization
of the findings, does limit extrapolation about cost and
practicality in the UK clinical setting.
12
E. Corner and R. Garrod
Table 5. Population and recruitment methods
Author Population and recruitment methods
Bianchi et al. (2002) [3] 33 (19) consecutive male stable
COPD patients referred to an
outpatient PRP from January 1998
to April 2000 in Gussaco Italy.
Costes et al. (2003) [10] 14 (14) patients enrolled in the
pulmonary rehab training
programme in France
Hawkins et al. (2002) [12] 37 (29) COPD patients in
Amsterdam aged 40 to 75 with
ventilatory limited exercise
capacity.
Johnson et al. (2002) [13] 41 (32) patients referred to PRP
between March 1996 and January
1997 from a population of military
beneficiary’s in San Antonio, TX
metropolitan areas.
Rueveny et al. (2005) [21] 24 (19) severe stable COPD patients
at the pulmonary institute at Sheba
Medical centre in Israel.
van’t Hul et al. (2006) [26] 27 (19) patients with severe COPD
attending a chest clinic at Kings
College Hospital London over 2
years.
Only two of the participants were
women (both in unassisted group).
The number within the brackets indicates the number of participants
that completed that study.
COPD = chronic obstructive pulmonary disease.
Training: regimes, duration and intensity
(Table 6)
For the details of training regimes, see Table 6. All
researchers exercised the patients at 50–70% of their
Wpeak, in line with current recommendations (ATS,
2009; British Thoracic Society [BTS], 2001). Reuveny
et al. (2005) and Johnson et al. (2002) are the only
researchers to train the patients on a treadmill. Tread-
mill training, compared with cycling, is considered a
transferable and favoured training regime in the UK.
The duration of the programmes ranged from 6 to 8
weeks in line with the joint American College of Chest
Physicians and The American Association of Cardio-
vascular and Pulmonary Rehabilitation Guidelines
(ACCP/AACVPR, 2006). However, they also suggest
that programmes exceeding 12 weeks ‘produce greater
sustained benefits than shorter programmes.’
Johnson et al. (2002) exercised his patients for 20
minutes twice a week, it is questionable as to whether
this is sufficient time to elicit the desired effect and may
Physiother. Res. Int. 15 (2010) 5–15 © 2009 John Wiley & Sons, Ltd.
E. Corner and R. Garrod Does NIV during Pulmonary Rehabilitation Improve Exercise Tolerance in COPD?

Table 6. Summary of training regimes

Author Training regime, duration and intensity

Bianchi et al. (2002) Duration: 6 weeks

Frequency: Three 3-hour sessions
Setting: outpatient MDT PRP. Programme: 30 minutes continuous cycling, all body resistance exercises and
education.
Intensity: at 50–70% maximal load

Costes et al. (2003) Duration: 8 weeks

Frequency: 30 minutes, thrice a week
Programme: cycling
Intensity: 60% of their peak V02

Hawkins et al. (2002) Duration: 6 weeks

Frequency: 30 minutes, thrice a week
Programme: Cycling, no education element to the programme.
Intensity: 70% of peak work rate until symptom rating of 5 on the Borg scale was achieved. Intensity was
increased progressively by 5 W when able to sustain exercise for 30 minutes

Johnson et al. (2002) Duration: 6 weeks

Frequency: 20 minutes, twice week
Programme: Treadmill training and education sessions
Intensity: 50–60% of their predetermined maximum

Rueveny et al. (2005) Duration: 8 weeks

Frequency: 45 minutes, twice a week
Programme: Treadmill training
Intensity: 65–70% of their predetermined maximum

van’t Hul et al. (2006) Duration: 8 weeks

Frequency: 45 minutes, thrice a week
Programme: Cycling calibrated cycle ergometer
Intensity: started at 65% of Wmax for up to 15 minutes, then the intensity was increased by 5% of Wmax.

MDT = Multi-disciplinary team; PRP = Pulmonary rehabilitation programme.

explain his inconclusive results. As this is an American
study, it was compared with the ACCP/AACVPR
guidelines (2009) which do not specify frequency
or duration of each exercise session. However, The
Australian Lung Foundation and The Australian
Physiotherapy Association (ALF/APA, 2009) recom-
mend a minimum of 30 minutes lower limb endurance
training at least 4–5 times a week. Additionally, the BTS
(BTS, 2009) recommend a minimum of 20–30 minutes,
three times a week. Johnson et al.’s (2002) exercise
regime failed to meet these standards. The rest of the
study programmes were in accordance with the evi-
denced-based BTS guidelines (BTS, 2001).
Compliance and practicality (Table 7)
One notable problem as demonstrated in Table 7 is
the drop-out rate because of discomfort with the ven-
tilator, either because of the mask or the ventilation
settings. Bianchi et al.’s (2002) study had the highest
drop-out for this reason. This is unfortunate, particu-
larly in light of the fact that one of the aims of PAV was
to increase synchrony between the patient and ventila-
tor (Ambrosino et al. 2002), which would hopefully
increase patient comfort. Problems may have arisen
because of the difficulties in setting the appropriate
flow and volume assist.
Reuveny et al. (2005) found a similar lack of compli-
ance with the Bipap masks. This was not demonstrated
in Johnson et al.’s (2002) study where nasal masks were
used instead of full facemasks, possibly explaining the
differing drop-out rate.
One important factor discussed in all studies is the
cost and practicality of NIV in a pulmonary rehabilita-
tion programme. The increases in cost because of
equipment and time required to set up the NIV will
have implications that cannot be ignored when apply-
ing this in a clinical setting.

Physiother. Res. Int. 15 (2010) 5–15 © 2009 John Wiley & Sons, Ltd. 13
Does NIV during Pulmonary Rehabilitation Improve Exercise Tolerance in COPD? E. Corner and R. Garrod

Table 7. Drop-out rate

Author Drop out active Reasons for drop out

Bianchi et al. (2002) PAV: 8 (44%) Discomfort due ‘mask and/or ventilatory setting’ (n = 5)
SB: 5 (33%) Exacerbation COPD (n = 7)
Exercise induced hypertension (n = 3)
Coronary disease (n = 2)

Costes et al. (2003) Bipap: 0(0%) Not applicable

SB: 0(0%)

Hawkins et al. (2002) PAV: 4 (26%) ‘Exacerbation of underlying disease (n = 5)’ ‘Non-compliance with exercise programme
SB: 6 (40%) (n = 4)’ ‘Non-pulmonary hospitalization (n = 2)’

Rueveny et al. (2005) Bipap: 3 (25%) Non-compliance with Bipap mask (n = 3) Lung transplant (n = 1) Back pain (n = 1)
SB: 2 (17%)

van’t Hul et al. (2006) IPS10: 4 Fatigue (n = 1)

IPS5: 4 CVA (n = 1)
Exacerbation COPD (n = 6)

Johnson et al. (2002) SB: 2 (15%) ‘Exertional angina’ (n = 1)

Bipap: 4 (27%) ‘Congestive heart failure’ (n = 1)
Heliox:1 (9%) ‘Flare of chronic liver disease’ (n = 1)
‘COPD exacerbation’ (n = 1)
‘Tibial fracture’ (n = 1)
‘Scheduling conflict’ (n = 1)
‘Non-compliance with exercise programme’ (n = 1)

COPD = Chronic obstructive pulmonary disease; CVA = Cerebral vascular accident.

Conclusions and implications Acknowledgement

In conclusion, the use of NIV as an adjunct to pulmo-
nary rehabilitation may augment improvements in
exercise tolerance. However, the research validity is
questionable because of a significant lack of blinding,
concealment, ill-defined randomization methods and
low numbers of participants. Both PAV and BiPAP
appear to cause problems with compliance because of
discomfort and the optimal mode of ventilation is still
open to debate. It appears that higher pressures are
more effective provided they can be tolerated.
It would be reasonable to suggest, based on these
studies, that severe COPD patients benefit more from
NIV during exercise than milder patients. NIV may
however be impractical and time consuming especially
in the NHS setting. Numbers needed to treat and cost–
benefit analysis are required.
Further research
Further research of long term follow up (e.g. 6–12
months) is recommended to see if the additional train-
ing benefits are maintained.
We gratefully acknowledge the support of the Faculty
of Health and Social Care Sciences St. George’s and
Kingston Universities.
REFERENCES
Ambrosino N, Rossi A. Proportional assist ventilation
(PAV): a significant advance or a futile struggle between
logic and practise. Occasional Review. Thorax 2002; 57:
272–276.
Ambrosino N. Assisted ventilation as an aid to exercise
training: a mechanical doping? European Respiratory
Journal 2006; 27: 3–5.
American Thoracic Society (ATS) (homepage online).
(2009) (Available at: http://www.thoracic.org/sections/
copd/for-health-professionals/definition-diagnosis-
and-staging/assessment-of-severity-staging.html,
accessed April 2009)
The Australian Lung Foundation and The Australian
Physiotherapy Association (2009) Australian Pulmo-
nary Rehabilitation Guidelines. (Available at: http://www.
pulmonaryrehab.com.au/index.asp?page=51, accessed
12 May 2009)

14 Physiother. Res. Int. 15 (2010) 5–15 © 2009 John Wiley & Sons, Ltd.
E. Corner and R. Garrod Does NIV during Pulmonary Rehabilitation Improve Exercise Tolerance in COPD?
Appendini L. Proportional assist ventilation: back to the
future. Editorial. Respiration 2003; 70: 345–346.
Bersten A, Soni N. Oh’s Intensive Care Manual, 6th edn.
Philadelphia: Elsevier, Butterworth Heinemann, 2009.
Bianchi L, Foglio K, Porta P, Baiardi P, Vitacca M, Ambro-
sino N. Lack of additional effect of adjunct of assisted
ventilation to pulmonary rehabilitation in mild COPD
patients. Respiratory Medicine 2002, 96: 359–367.
British Thoracic Society (BTS). (2001) (Available at:
http://www.brit-thoracic.org.uk/clinical-information/
pulmonary-rehabilitation/pulmonary-rehabilitation
-guidelines.aspx, accessed 13 May 2009)
Casaburi R, Patessio A, Ioli F, Zanaboni S, Donner CF,
Wasserman K. Reduction in exercise lactic acidosis and
ventilation as a result of exercise training in patients
with obstructive lung disease. American Review of
Respiratory Disease 1991; 143: 9–18.
Costes F, Agrest A, Court-Fortune I, Roche F, Vergnon
J-M, Barthélémy JC. Non-invasive ventilation during
exercise training improves exercise tolerance in patients
with chronic obstructive pulmonary disease. Journal
Cardiopulmonary Rehabilitation 2003; 23: 307–331.
De Troyer A. Effect of hyperinflation on the diaphragm.
European Respiratory Journal 1997; 10: 708–713.
Hawkins P, Johnson LC, Nikoletou D, Hamnegård C-H,
Sherwood R, Polkey MI. Proportional assist ventilation
as an aid to exercise training in severe chronic obstruc-
tive pulmonary disease. Thorax 2002; 57: 853–859.
Johnson J, Gavin D, Adams-Dramiga S. Effects of training
with Heliox and non-invasive positive pressure ventila-
tion on exercise ability in patients with severe COPD.
Chest 2002; 122: 464–472.
Keilty SEJ, Ponte J, Fleming A, Moxham J. Effect of inspira-
tory pressure support on exercise tolerance and breath-
lessness in patients with severe stable chronic obstructive
pulmonary disease. Thorax 1994; 49; 990–994.
Molken M, Roos B, Van Nord JA. An empirical compari-
son of the St Georges Respiratory Questionnaire (SGRQ)
and the Chronic Respiratory Disease Questionnaire
(CRG) in a clinical trial setting. Thorax 1999; 54(11):
995–1003.
Pineda H, Haas F, Axen K, Haas A. Accuracy of pulmonary
function tests in predicting exercise tolerance in chronic
Physiother. Res. Int. 15 (2010) 5–15 © 2009 John Wiley & Sons, Ltd.
obstructive pulmonary disease. Chest 1984; 86(4):
564–567.
Polkey MI, Kyroussis D, Mills GH Somfay A, Casaburi R.
Inspiratory pressure support reduces slowing of inspira-
tory muscle relaxation rate during exhaustative tread-
mill walking in severe COPD. American Journal of
Respiratory and Critical Care Medicine 1996; 154:
1146–1150.
Ranieri VM, Dambrosio M, Brenza N. Intrinsic PEEP and
cardiopulmonary interaction in patients with COPD
and acute ventilatory failure. European Respiratory
Journal 1996; 9(6): 1283–1291.
Reuveny R, Ben-Dov I, Gaides M, Reichert N. Ventilatory
Support during training improves training benefit in
severe chronic airway obstruction. Israel Medical Asso-
ciation Journal 2005; 7: 151–155.
Ries AL, Bauldoff GS, Carlin BS, Casaburi R, Emery C,
Mahler D, Make B, Rochester C, ZuWallack R, Herre-
rias C. (2007) Pulmonary rehabilitation*: joint ACCP/
AACVPR Evidence-Based Clinical Practice Guidelines.
Chest 2007; 131(5 suppl): 45–425.
Scottish Intercollegiate Guidelines Network (SIGN)
(2009a) Critical Appraisal: Notes and Checklists. (Avail-
able at: http://www.sign.ac.uk/methodology/checklists.
html, accessed April 2009)
Scottish Intercollegiate Guidelines Network (SIGN)
(2009b) What is SIGN? (Available at: http://www.sign.
ac.uk/about/introduction.html, accessed April 2009)
van’t Hul A, Kwakkel G, Gosselink R. The acute effects of
non-invasive ventilatory support during exercise on
exercise endurance and dyspnoea in patients with
chronic obstructive pulmonary disease: a systematic
review. Journal of Cardiopulmonary Rehabilitation
2002; 4: 290–297, July/August.
van’t Hul AJ, Gosselink R, Hollander AP, Postmus PE,
Kwakkal G. Acute effects of inspiratory pressure support
during exercise in patients with COPD. European
Respiratory Journal 2004; 23: 34–40.
Wouters EFM, Creutzberg EC, Schols AMWJ. Systemic
effects in COPD. Chest 2002; 121(5): 127s–130s.
Wouters EFM. Chronic obstructive pulmonary disease
5: systemic effects of COPD. Thorax 2002; 57:
1067–1070.
15