British Journal of Anaesthesia 104 (5): 643–7 (2010)

doi:10.1093/bja/aeq080 Advance Access publication March 30, 2010

RESPIRATION AND THE AIRWAY

Lung recruitment and positive airway pressure before extubation
does not improve oxygenation in the post-anaesthesia
care unit: a randomized clinical trial
A. B. Lumb 1, S. J. Greenhill 1 2, M. P. Simpson 1* and J. Stewart 1
1
St James’s University Hospital, Leeds Teaching Hospitals NHS Trust, Beckett Street, Leeds LS9 7TF, UK.
2
Present address: The Queen Elizabeth Hospital, Gayton Road, King’s Lynn PE30 4ET, UK
*Corresponding author. E-mail: mpsimpson@doctors.org.uk
Background. Atelectasis is known to develop during anaesthesia and after operation atelecta-
sis leads to impaired oxygenation. Lung recruitment manoeuvres, positive end-expiratory
pressure (PEEP), and continuous positive airway pressure (CPAP) have been proposed for

Downloaded from bja.oxfordjournals.org by guest on April 11, 2011

reduction of atelectasis but their benefits have not been shown to persist after operation. We
proposed that a combination of these techniques before extubation would improve oxygen-
ation after operation.
Methods. Adult patients undergoing elective surgery requiring tracheal intubation and an arterial
catheter were randomized to receive either: a lung recruitment manoeuvre of 40 cm H2O for 15
s, 30 min before the end of anaesthesia, followed by 10 cm H2O of PEEP and then 10 cm H2O of
CPAP from return of spontaneous breathing until extubation; or no lung recruitment manoeuvre,
5 cm H2O PEEP, and no CPAP. Arterial blood gases were taken at randomization and 1 h after
extubation. The primary endpoint of the study was the change in (A –a)DO2 between these times.
Statistical analysis of the two groups was done by x2 or unpaired t-test as appropriate.
Results. Twenty-two patients were recruited to each group. There were no significant differences
between the groups before randomization. There was no significant difference in the change in
(A –a)DO2 between the groups (P¼0.82).
Conclusions. Postoperative oxygenation is not improved by a combination of a lung recruitment
manoeuvre and maintenance of a positive airway pressure until extubation. Further research is
needed to elucidate the mechanism of atelectasis on emergence from anaesthesia and to evaluate
more invasive clinical strategies such as post-extubation CPAP.
Trial registered at URL http://www.controlled-trials.com
Identification number: ISRCTN32464251
(http://www.controlled-trials.com/ISRCTN32464251)
Br J Anaesth 2010; 104: 643–7
Keywords: complications, atelectasis; lung, atelectasis; ventilation, continuous positive
pressure; ventilation, positive end-expiratory pressure
Accepted for publication: February 20, 2010

Atelectasis in dependent lung units is an almost inevitable
consequence of general anaesthesia with IPPV1 2 and,
along with altered ventilation perfusion relationships,
causes an increase in alveolar to arterial oxygen partial
pressure difference (A –a)DO2. Atelectasis, which can
persist into the postoperative period,3 is resistant to simple
techniques normally used to improve lung function such
as patient posture and may contribute to postoperative
pulmonary complications.4 Studies using computerized
tomography (CT) during anaesthesia have shown that
atelectasis formation can be attenuated by the use of posi-
tive end-expiratory pressure (PEEP) and the avoidance of
high inspired oxygen fractions (FIO2).5 6 Once formed,
atelectatic lung regions may be effectively recruited by
maintaining a sustained airway pressure of 40 cmH2O for
15 s.7 This is the only lung recruitment manoeuvre that

# The Author [2010]. Published by Oxford University Press on behalf of the British Journal of Anaesthesia. All rights reserved.
For Permissions, please email: journals.permissions@oxfordjournal.org
 Lumb et al.
has been shown to be effective in CT studies,7 is safe in
clinical use,8 and improves intraoperative oxygenation.
When performed in isolation, this improvement in oxygen-
ation does not persist into the postoperative period.9
If an intraoperative recruitment manoeuvre is followed
by the use of 10 cm H2O of PEEP, atelectasis is prevented
from recurring even in the presence of a high FIO2.9 We
hypothesized that if positive airway pressure were main-
tained after a recruitment manoeuvre and throughout emer-
gence from anaesthesia, there would be a reduction in
atelectasis even in the presence of 100% oxygen and so an
improvement in oxygenation in the postoperative period.
The key period for developing atelectasis at the end of
anaesthesia is likely to be the period after discontinuation
of IPPV and PEEP but before extubation. During this
period, patients often cough and zero end-expiratory
pressure (ZEEP) is commonly used and will accentuate
flow- and volume-related airway collapse. ZEEP may be
avoided in the spontaneously breathing, intubated patient
by preventing disconnection of the breathing system from
the tracheal tube and by using the adjustable pressure lim-
iting (APL) valve to generate continuous positive airway
pressure (CPAP).
We tested the hypothesis that by performing a recruit-
ment manoeuvre near the end of the anaesthetic followed
by PEEP and then CPAP until extubation, the (A – a)DO2
1 h after extubation would be improved compared with
patients in whom these interventions were not performed.
Methods
The study was approved by the Leeds Local Research
Ethics Committee (Ref: 06/Q1205/17 COREC Locked
Code: AB/66109/2) and registered at http://www.controlled-
trials.com (Identification number: ISRCTN32464251).
Written informed consent was obtained from all the study
subjects.
Patients were recruited from elective surgical lists at the
two large teaching hospitals in the Leeds Teaching
Hospitals NHS Trust. Patients deemed eligible for
inclusion were adults undergoing elective operative pro-
cedures that were expected to last .45 min, and whose
anaesthesia technique was to include tracheal intubation,
muscle relaxation, IPPV, and the insertion of an arterial
catheter for clinical care. Patients were excluded if there
was any contraindication to the use of 10 cm H2O of
PEEP or the recruitment manoeuvre or the clinical require-
ment for the use of .5 cm H2O of intraoperative PEEP.
Patients were also excluded if any perioperative cardiovas-
cular or respiratory event occurred which the anaesthetist
with clinical responsibility for the patient thought would
make the study intervention clinically unacceptable.
Patient factors recorded before operation were age, sex,
BMI, and smoking status. Apart from the requirements for
inclusion in the study described above, the anaesthetic
644
technique was left to the discretion of the anaesthetist.
Intraoperative IPPV settings were an FIO2 of 0.3 –0.8,
PEEP 5 cm H2O, tidal volume 7 – 10 ml kg21, and venti-
latory frequency as appropriate to maintain end-tidal
carbon dioxide partial pressure of 4– 5.5 kPa. Thirty
minutes before the anticipated end of the anaesthetic, an
arterial blood sample was obtained for measurement of
arterial PO2 and PCO2, and the FIO2 was recorded. The
patients’ allocation was concealed from the investigators
until this point when an envelope prepared at the start of
the trial (by a non-investigator) bearing only a number on
the outside and the allocation group on the inside was
opened.
Fifteen minutes before the anticipated end of anaesthe-
sia, patients in the intervention group received a lung
recruitment manoeuvre consisting of lung inflation to an
airway pressure of 40 cm H2O sustained for 15 s.
Immediately after this recruitment manoeuvre, ventilation
Downloaded from bja.oxfordjournals.org by guest on April 11, 2011
continued as previously but with PEEP increased to 10 cm
H2O. IPPV continued at the same rate until evidence of
return of spontaneous ventilation. At this point, CPAP was
maintained at 10 cm H2O by adjusting the APL valve and
providing a fresh gas flow of .10 litre min21 to maintain
the positive airway pressure during inspiration. Circuit dis-
connection was avoided at all times before extubation of
the trachea to maintain the positive airway pressure.
The control group received no lung recruitment
manoeuvre and ventilation continued with the current set-
tings including any intraoperative PEEP (5 cm H2O). No
CPAP was applied on return of spontaneous ventilation,
with the APL valve left fully open. Circuit disconnection
was permitted, for example, when moving the patient.
In both groups, the FIO2 was set to 1.0 with fresh gas
flow exceeding 10 litre min21 before extubation of the
trachea. This occurred when deemed appropriate by the
anaesthetist with clinical responsibility for the patient.
In the post-anaesthesia care unit (PACU), both groups
received standard post-anaesthetic care. A Venturi mask
(Lifecare Hospital Supplies, Edinburgh, UK) providing
an FIO2 of 0.4 was used for all patients from 45 min
post-extubation until an arterial blood sample was obtained
at 1 h post-extubation for measurement of arterial PO2
and PCO2.
The (A – a)DO2 gradient at 1 h post-extubation was used
as the endpoint of the study. Arterial oxygen partial
pressure was measured from the arterial blood samples as
described above. Alveolar PO2 was calculated using the
following version of the alveolar gas equation:
PaCO2
alveolar PO2 ¼ PIO2  ð1  F IO2 ð1  RQÞÞ
RQ
RQ assumed to be 0.8.
The standard deviation (SD) of (A – a)DO2 for the pur-
poses of the power calculation was derived from preinduc-
tion values in a previous study of a similar population
 Lung recruitment and (A –a)DO2 in PACU

Consented, n = 52
Operation cancelled, n = 2
Investigator unavailable, n = 1
Required PEEP >5 cm H2O, n = 2
Randomized, n = 47

Allocated to standard group, n = 23 Allocated to intervention group, n = 24

Received allocated treatment, n = 23 Received allocated treatment, n = 22
Protocol violation circuit disconnection, n = 2

Lost to follow-up, n = 0 Lost to follow-up, n = 0

Required F IO2>0.4 in PACU, n = 1 Discontinued intervention, n = 0

Downloaded from bja.oxfordjournals.org by guest on April 11, 2011

Analysed, n = 22 Analysed, n = 22
Excluded from analysis, n = 1 Excluded from analysis, n = 2

Fig 1 Consort flow chart showing patients’ progress through the trial.

Table 1 Comparison of the two study groups showing mean (range), mean (SD)
(3.91 kPa).9 A significance level for the study of 5% or number of occurrences of factors that could potentially affect (A –a)DO2.
(P¼0.05) and a power of 80% were used for the calcu- None of the factors differed significantly between the groups, P.0.05. PACU,
lation. Two groups of 22 subjects were calculated to be post-anaesthesia care unit. Operation: AO, abdominal open; AL, abdominal
laparoscopic; P, peripheral
needed to detect a difference of 3.33 kPa (0.8 times the
previously published SD). Standard group Intervention group
Statistical analysis of the two groups was done by x2 or
Age (yr) 61.8 (42 –85) 68.3 (49 –89)
unpaired t-test as appropriate with significance assumed with BMI (kg/m2) 28.4 (7.7) 27.3 (4.5)
P,0.05 using SPSS 15.0 (SPSS Inc., Chicago, IL, USA). Sex (male/female) 18/4 14/8
ASA (I/II/III) 7/7/8 2/15/5
Smoker/non-smoker 4/18 2/20
Operation (AO/AL/P) 12/4/6 15/4/3
Duration of surgery (min) 255 (150) 319 (122)
Results (A – a)DO2 at randomization (kPa) 12.5 (8.7) 12.1 (8.4)
FIO2 at randomization 0.475 (0.119) 0.467 (0.094)
A total of 52 patients were enrolled between June 2006 Arterial PCO2 at randomization (kPa) 5.59 (0.72) 5.53 (0.80)
and February 2009 (Fig. 1). There were no significant Arterial PCO2 in PACU (kPa) 5.77 (0.61) 5.70 (0.37)
differences between the groups with regard to preoperative
characteristics, duration, and nature of operation, and arter-
ial gases at randomization and in PACU (Table 1).
The mean (SD) (A – a)DO2 during anaesthesia was 12.3 persists into the postoperative period. We have also shown
(8.4) kPa and this increased slightly in PACU to 14.1 that a lung recruitment manoeuvre followed by positive
(5.2) kPa. The change in individual (A – a)DO2 between airway pressure until extubation does not improve the
randomization and PACU displayed no particular pattern (A – a)DO2 in PACU.
(Fig. 2). In the usual absence of a significant alveolar to pulmon-
Using each patient as their own control, the mean ary capillary diffusion barrier, abnormalities of (A – a)DO2
change in (A – a)DO2 was 1.98 (6.52) in the standard group result from intrapulmonary shunting and areas of lung
and 1.52 (6.71) in the intervention group (P¼0.82). This with low ventilation– perfusion ratios.10 During anaesthe-
showed that there was no statistical difference or clinical sia, intrapulmonary shunting through areas of atelectasis is
benefit accruing from the intervention. the most significant contributor; so, in the absence of
acute lung pathology before operation, it may be assumed
that changes in (A – a)DO2 are mostly secondary to the
development of atelectasis.
Discussion Our results are comparable with a previous study which
This study has shown that patients undergoing general also used (A – a)DO2 as a marker of atelectasis,9 and
anaesthesia develop abnormal (A – a)DO2 and that this demonstrate high (A –a)DO2 values at the end of the

645

Intervention
40
(A–a)DO2 (kPa)
30
20
10
0 0
Randomization
Fig 2 Individual patients’ (A –a)DO2 at randomization and in PACU ( post-anaesthesia care unit) for each treatment group.
anaesthetic that become even higher at 1 h after operation.
As demonstrated by the large standard deviations in our
data, there is a wide variation between patients, particu-
larly during anaesthesia, results which are again compar-
able with the previous study,9 which also studied the
effects of ‘manual inflations’ every 30 min throughout
surgery and before extubation. Although details of the
manual inflation used and how long before extubation it
was applied are not clear, they, like us, failed to demon-
strate any beneficial effect on (A – a)DO2 in PACU.
It is known that CPAP at induction prevents the for-
mation of atelectasis11 and that PEEP applied after a
recruitment manoeuvre during anaesthesia prolongs the
time taken for atelectasis to reform.6 We hoped that by
applying the same principle at the end of anaesthesia and
up until the moment of extubation, we could reduce the
effects of atelectasis upon oxygenation in the early recov-
ery period. Our results show that our management strategy,
used at the end of anaesthesia, has not provided a signifi-
cant improvement in patient oxygenation in PACU. There
are a number of possible explanations for this. First,
(A – a)DO2 may be affected by the FIO2 or the arterial
PCO2,12 but this is unlikely to have affected our results as
the values for these factors were the same in each group
(Table 1). Secondly, despite using high fresh gas flow
rates, our method of applying CPAP during spontaneous
respiration may have been ineffective during inspiration,
particularly at high inspiratory flow rates. Thirdly, the
positive airway pressures used (PEEP and CPAP) after
the lung recruitment manoeuvre may have been too low;
10 cm H2O is greater than or equal to that shown to be
effective during previous studies at other periods of
anesthesia6 11 and was the maximum level we felt we
could use without risking clinically unacceptable cardio-
vascular depression. Fourthly, we did not apply CPAP by
a facemask after extubation. It was hoped that after extuba-
tion, normal vocal fold function would quickly return and
provide auto-CPAP by the normal physiological mechan-
ism of vocal fold adduction during expiration, which
retards expiration to reduce pulmonary collapse.13 It seems
that this aspect of respiratory muscle activity may not
return immediately after extubation. Finally, the
Lumb et al.
Standard
40
(A–a)DO2 (kPa)
30
20
10
PACU Randomization PACU
mechanism of atelectasis at the end of anaesthesia may
differ from that so widely studied at other times. When a
patient attempts to cough with a tracheal tube in situ, an
unphysiological phenomenon occurs. A normal cough has
Downloaded from bja.oxfordjournals.org by guest on April 11, 2011
three phases: an inspiratory phase when lung volume is
increased, a compressive phase when there is a forced
expiration against a closed glottis, and an expulsive phase
when the glottis is quickly opened and extremely high
expiratory flows generated. This differs from the expiration
reflex which occurs after direct stimulation of the upper
airway and involves only the last two phases of the cough
reflex.14 Both these protective reflexes are likely to be
active immediately before extubation, and both will be
ineffective due to the inability of the glottis to close for
the compressive phase. Instead, the expiratory muscles
will be contracting strongly against an open airway, so
actively reducing lung volume to very low levels, encoura-
ging pulmonary collapse. Although only occurring for a
short time, the positive intrathoracic pressure generated
during the compressive phase of a cough is very high, and
therefore, our strategy of 10 cm H2O positive airway
pressure will have had no impact on ameliorating the
effects of the cough reflex on lung collapse.
Our study shows that lung recruitment manoeuvres that
have been shown to improve atelectasis temporarily may
not provide a benefit that persists into the postoperative
period.
The patients we studied may have influenced the poor
oxygenation observed. In order to be included in the
study, all patients were required to have an arterial line
inserted on clinical grounds and so were mostly ASA II or
III having major and prolonged surgery, and therefore care
should be taken when applying our results to other patient
populations.
CT is the best technique for quantifying atelectasis but
this is impossible during surgery, and is difficult in the early
postoperative period when close monitoring is required. We
therefore used (A –a)DO2 as a surrogate marker of the sever-
ity of atelectasis, allowing an assessment to be made at both
the time points in which we were interested.
We did not document whether individual patients
coughed, or to what extent, so are unable to comment
646
 Lung recruitment and (A –a)DO2 in PACU
further on whether this mechanism could explain our
results. However, it is interesting to speculate whether the
trend in recent years to extubate patients when awake
enough to protect their airway from aspiration may be
causing widespread impairment of lung function in the
postoperative period.
In order to prevent the significantly impaired oxygen-
ation that occurs in PACU, further research is needed both
to elucidate the mechanism of atelectasis on emergence
from anaesthesia and to evaluate more invasive clinical
strategies.
Future studies using ventilators specifically designed to
offer CPAP while spontaneously breathing, or iatrogenic
CPAP immediately after extubation, may prove effective
in reducing atelectasis. CPAP has been shown to be an
effective treatment once postoperative hypoxia has devel-
oped,15 but a functional seal between the mask and patient
may be difficult to achieve immediately post-extubation.
In conclusion, the use of a recruitment manoeuvre and
positive airway pressure strategy, which is effective in pre-
venting atelectasis on induction and during anaesthesia,
has not been shown to improve oxygenation after operation
when used on emergence from anaesthesia and therefore
cannot be recommended.
Conflict of interest
None declared.
Funding
All funding was from departmental sources.
References
1 Warner DO, Warner MA, Ritman EL. Atelectasis and chest wall
shape during halothane anesthesia. Anesthesiology 1996; 85:
49 – 59
647
2 Lundquist H, Hedenstierna G, Strandberg A, Tokics L, Brismar B.
CT-assessment of dependent lung densities in man during general
anaesthesia. Acta Radiol 1995; 36: 626 – 32
3 Strandberg A, Tokics L, Brismar B, Lundquist H, Hedenstierna G.
Atelectasis during anaesthesia and in the postoperative period.
Acta Anaesthesiol Scand 1986; 30: 154 – 8
4 Duggan M, Kavanagh BP. Pulmonary atelectasis. A pathogenic
perioperative entity. Anesthesiology 2005; 102: 838 – 54
5 Rothen HU, Sporre B, Engberg G, Wegenius G, Högman M,
Hedenstierna G. Influence of gas composition on recurrence of
atelectasis after a re-expansion manoeuvre during general anaes-
thesia. Anesthesiology 1995; 82: 832– 42
6 Neumann P, Rothen HU, Berglund JE, Valtysson J, Magnusson A,
Hedenstierna G. Positive end-expiratory pressure prevents
atelectasis during general anaesthesia even in the presence of a
high inspired oxygen concentration. Acta Anaesthesiol Scand 1999;
43: 295 – 301
7 Rothen HU, Neumann P, Berglund JE, Valtysson J, Magnusson A,
Hedenstierna G. Dynamics of re-expansion of atelectasis during
general anaesthesia. Br J Anaesth 1999; 82: 551 – 6
8 Magnusson L, Tenling A, Lemoine R, Högman M, Tydén H,
Downloaded from bja.oxfordjournals.org by guest on April 11, 2011
Hedenstierna G. The safety of one, or repeated, vital capacity man-
euvres during general anesthesia. Anesth Analg 2000; 91: 702–7
9 Clarke JP, Schuitemaker MN, Sleigh JW. The effect of intraopera-
tive ventilation strategies on perioperative atelectasis. Anaesth
Intensive Care 1998; 26: 262 – 6
10 Lumb AB. Distribution of pulmonary ventilation and perfusion.
In: Lumb AB, ed. Nunn’s Applied Respiratory Physiology, 6th Edn.
Philadelphia: Elsevier Butterworth Heinemann, 2005; 110 – 33
11 Rusca M, Proietti S, Schnyder P, et al. Prevention of atelectasis
formation during induction of general anesthesia. Anesth Analg
2003; 97: 1835– 9
12 Kathirgamanathan A, McCahon RA, Hardman JG. Indices of pul-
monary oxygenation in pathological lung states: an investigation
using high-fidelity, computational modelling. Br J Anaesth 2009;
103: 291 – 7
13 Lumb AB. Pulmonary ventilation. In: Lumb AB, ed. Nunn’s Applied
Respiratory Physiology, 6th Edn. Philadelphia: Elsevier Butterworth
Heinemann, 2005; 76 – 91
14 Tatar M, Hanacek J, Widdicombe J. The expiration reflex from
the trachea and bronchi. Eur Respir J 2008; 31: 385 – 90
15 Squadrone V, Coha M, Cerutti E, et al. Continuous positive
airway pressure for treatment of postoperative hypoxemia: a ran-
domized controlled trial. J Am Med Assoc 2005; 293: 589 – 95