J. Paediatr.

Child Health (1998) 34, 147–150

Postextubation chest X-rays in neonates: A routine

no longer necessary
MW DAVIES and DW CARTWRIGHT
Grantley Stable Neonatal Unit, Royal Women’s Hospital, Herston, Queensland, Australia

Objectives: To ascertain the incidence of postextubation atelectasis (PEA) in neonates, to delineate any objective differences
between those infants with PEA and those without, and to see if any of those differences were predictive of the need for a
postextubation chest X-ray (CXR).
Methods: This is a retrospective review of all infants ventilated in 1994. For each separate period of extubation the medical,
physiotherapy and nursing notes were examined. Data were collected on birthweight, gestational age, duration of ventilation,
age at extubation, ventilation requirements pre-extubation, pre- and postextubation arterial carbon dioxide tensions (PaCO )
2
and oxygen requirements, the number of episodes of bradycardia and apnoea, the pulse and respiratory rates pre- and
postextubation, and the use of nasal continuous positive airway pressure (NCPAP). It was routine practice throughout 1994 for
all ventilated babies to have a CXR 6 h postextubation. Each postextubation CXR was examined by one of the authors (MWD)
for the presence of atelectasis and other diagnoses. PEA was defined as any atelectasis present on the postextubation CXR
that was not present on the pre-extubation CXR.
Results: The overall incidence of any PEA was 2.5% (6/236). In those babies with PEA, the increase in oxygen requirement
at 1 and 6 h postextubation was higher (change in inspired oxygen (DFiO ) of 0.05 vs 0.015, P=0.043 and DFiO of 0.045 vs
2 2
0.0, P=0.033, respectively). There was a higher incidence of the need for NCPAP some time after extubation (2/4 vs 9/163,
P<0.001). No infant with PEA required reintubation and ventilation.
Conclusions: In this nursery the incidence of PEA is low with no significant morbidity. Postextubation CXRs should be
performed on only those infants who have an increase in oxygen requirement postextubation or become symptomatic with new
or increasing respiratory distress, and to follow up atelectasis on the most recent pre-extubation CXR.

Key words: atelectasis; chest X-ray; intermittent positive pressure ventilation; intubation; newborn; postextubation.

Routine chest X-ray (CXR) after extubation of ventilated METHODS

neonates is common practice in many neonatal intensive
care nurseries. In our institution a CXR is planned 6 hours
postextubation with the principal aim being to diagnose
postextubation atelectasis (PEA). Postextubation atelectasis
has been a significant problem in neonatal intensive care over
the last three decades, with studies reporting an incidence as
high as 40.7% in 1977.1 The most recent study in 1993
reported an incidence of 10.8%.2 When the incidence of PEA
was at its peak in the 1970s, it was associated with significant
morbidity with reintubation rates in cases of PEA ranging from
44% to 64%.1,3,4 It is our impression that the incidence of PEA
has been declining over time and that this may warrant a
review of the practice of routine postextubation chest
X-rays.
The aim of this study was to ascertain the incidence and
pattern of postextubation atelectasis in our population of
ventilated neonates, to see if there were any differences
between those infants who had PEA and those who did not,
and to see if any of those differences were predictive of the
need for a postextubation CXR.
Correspondence: Dr MW Davies, Department of Neonatal Medicine,
Royal Women’s Hospital, 132 Grattan Rd, Carlton 3053, Victoria,
Australia.
Accepted for publication 18 September 1997.
Study design
This study is a retrospective review of all infants who were
ventilated in our Intensive Care Nursery (ICN) in 1994. The medical
records and postextubation CXRs of all babies were examined.
Each separate period of ventilation was ascertained, and for each
period the medical, physiotherapy and nursing notes and nursing
observation charts were examined. Data were collected on
birthweight, gestational age, duration of each particular episode
of ventilation, the age at extubation, ventilation requirements pre-
extubation, pre- and postextubation arterial carbon dioxide levels
(PaCO ) and pre- and postextubation oxygen requirements. Data
2
were also collected on the number of episodes of bradycardia
pre-extubation and postextubation and the number of apnoeic
episodes at 1, 6, 12, and 24 h postextubation, the pulse and
respiratory rates pre-extubation and at 1, 6, 12, and 24 h
postextubation. Other data were collected on the presence of
atelectasis on the pre-extubation CXR, chronic lung changes on
the postextubation CXR, whether the infant had ever had a patent
ductus arteriosus (PDA) or septicaemia and whether the infant
was commenced on nasal continuous positive airway pressure
(NCPAP) immediately after extubation or at some other time in
the first 24 h postextubation.
Postextubation atelectasis was defined as any atelectasis
present on the postextubation CXR that was not present on the
148
pre-extubation CXR. All CXRs were reviewed by a single
investigator (MWD). Atelectasis was defined as any opacification
on the CXR with loss of lung volume. Volume loss is characterised
by changes in the position of fissures, the mediastinum, hilar
shadows or the hemidiaphragms; crowding of the ribs; or
compensatory overinflation of the remaining aerated segments in
the affected lobe.5,6 We did not include the diffuse patchy
changes seen in evolving chronic lung disease (CLD). For each
episode of extubation the routine postextubation CXR was viewed,
as well as the CXR just prior to extubation, to determine if any
atelectasis present on the postextubation film was present
pre-extubation. It was routine practice throughout 1994 for all
ventilated babies to have a CXR 6 h postextubation. There was
no routine practice to perform a pre-extubation CXR, the most
recent film prior to extubation was reviewed.
Care protocols
Routine practices in our ICN throughout 1994 consisted of: (i)
almost all (97.3%) babies intubated in our nursery had oral
endotracheal tubes (ETT), (ii) all intubated babies had regular
ETT suction (with passage of the suction catheter to the tip of
the ETT and no further) at least every 4 h (or more frequently if
required clinically) during the period of intubation, (iii) enteral
feeds were ceased for 4 h prior to and 12 h after extubation, (iv)
a physiotherapy regimen (routinely administered if the baby had
been intubated for more than 48 h or less than 48 h with a large
amount of ETT secretions) that included 3-hourly treatments pre-
extubation with mask percussion for 1–2 min, ETT suction and
positioning to facilitate right upper lobe (RUL) drainage, then
three treatments every 2 h postextubation with chest percussion,
suction and positioning for RUL drainage, (v) NCPAP was
commenced routinely postextubation for any baby under 1200 g—
this protocol was overridden at the discretion of the consultant
neonatologist in charge of the infant (63/90 [70%] of infants
<1200 g were commenced on NCPAP immediately postextub-
ation), (vi) Aminophylline was used routinely (commenced prior
to extubation) in infants less than 32 weeks gestation, and to
treat troublesome apnoea in more mature infants.
Statistical analysis
Statistical analyses of numerical data comparing infants with
PEA with those with no atelectasis were performed using the
Mann–Whitney test. Categorical data in Table 2 were analysed
using the uncorrected Chi-squared test.7 A P-value <0.05 was
assumed to represent statistical significance.
RESULTS
There were 366 episodes of ventilation of 312 babies in our ICN
in 1994. Thirty-six of these ventilated infants were either not
extubated (e.g. transferred elsewhere ventilated) or died whilst
intubated or immediately after extubation. There were therefore
330 episodes of extubation, however, 83 babies (25%) did not
have a postextubation CXR. This left 247 babies who were
extubated and had a routine postextubation CXR. (Table 1).
There were 11 babies who had atelectasis both pre- and
postextubation, thus leaving 236 episodes of extubation.
The incidence of PEA as defined above was 2.5%. That is
six cases of PEA out of 236 episodes of extubation.
MW Davies and DW Cartwright
Table 1 Summary of patient numbers
Total number of infants ventilated (312)
366 Episodes Of ventilation
Number of episodes where the infant was not
extubated
(i.e. transferred to another centre or had a long-term 3
tracheostomy)
Number of infants who died during or immediately 33
after an episode of ventilation
Number of episodes of extubation (excluding above) 330
330 Extubations
Number of episodes of extubation with no 83
postextubation CXR
Number of postextubation CXRs 247
247 Postextubation CXRs
No atelectasis 230
Atelectasis present pre- and postextubation 11
Postextubation atelectasis (PEA) 6*
*Incidence of PEA per episode of ventilation, 2.5% (i.e. 6/236).
The 11 infants who had atelectasis present on the pre- and
postextubation CXR did not meet our definition of PEA (as
above). Altogether there were 13 infants who had atelectasis
present on the pre-extubation CXR, and 11 (85%) of these had
atelectasis on the postextubation CXR.
The most frequent site for PEA in this study was the right
upper lobe (RUL), 4/6 cases (67%)—3/6 cases involving the
whole RUL (50%) and 1/6 cases involving multiple segments in
the RUL (17%). One case had collapse of both lungs, and one
case had multiple areas of atelectasis on both sides of the chest.
The results comparing the infants with PEA with those with
no atelectasis are summarised in Table 2. The median PaCO
2
pre-extubation was higher in the PEA group (46 vs 38, P=
0.039). The median increase in oxygen requirement at 1 h
postextubation was higher in the PEA group (change in inspired
oxygen (DFiO ) of 0.05 vs 0.015, P=0.043). Similarly, the median
2
increase in oxygen requirement at 6 h postextubation was
higher in the PEA group (DFiO of 0.045 vs 0, P=0.033), Table 2.
2
If there was PEA then the infant was more likely to need NCPAP
after already being extubated to not have NCPAP (2/4 vs 9/163,
P<0.001), Table 2.
In the babies with PEA there was no significant difference
between pre- and postextubation PaCO . Postextubation pulse
2
and respiratory rates were the same in both groups, as was
the frequency of episodes of apnoea and bradycardia.
No infants with PEA required reintubation and ventilation for
deterioration in respiratory status in the 24 h after extubation,
and in only two out of the six babies with new PEA was there
any documented change in management following the postex-
tubation CXR.
DISCUSSION
Postextubation atelectasis has been a problem in neonatal
intensive care since ventilation of neonates became established
practice in the 1960s and 70s. The incidence of PEA in the
1970s was as high as 40.7%.1 The importance of diagnosing
PEA when it occurred was emphasised because of its significant
morbidity—studies reported an incidence of the need for
reintubation in cases of PEA from 44% reported by Finer et al.
in 1979,3 to as high as 64% reported by Wyman and Kuhns in
Routine postextubation CXRs 149

Table 2 Results: infants with no atelectasis compared with infants with postextubation atelectasis (PEA)

No atelectasis PEA
n=230 n=6 P-value

Median (interquartile range) Mann–Whitney test

Birthweight (g) 1673 (991–2400) 1290 (1104–1401) NS
Gestational age at birth (w) 31 (27–34) 29 (26.5–30) NS
Duration of intubation(h) 86 (54–140) 96 (53–149) NS
Age at extubation (days) 5 (4–9) 8 (6.3–13.5) NS
Weight at extubation (g) 1695 (1113–2410) 1273 (1104–1528) NS
Pre-extubation IPPV
rate (breaths min−1) 10 (10–10) 10 (10–14) NS
Pre-extubation PIP
(cm H O) 14 (13–15) 14 (13–15) NS
2
Pre-extubation PEEP
(cm H O) 4 (4–5) 4 (4–4) NS
2
Pre-extubation PaCO
2
( mmHg) 38 (33–42) 46 (42.8–48.8) 0.039
Post-extubation PaCO
2
( mmHg) 38 (34–42) 43 (41.3–44.8) NS
Difference between
pre- and post- PaCO
2
(mmHg) 1 (−3–4) −1.5 (−6.3–2.3) NS
Pre-extubation FiO 0.21 (0.21–0.25) 0.21 (0.21–0.21) NS
2
Postextubation FiO
2
at 1 h 0.25 (0.21–0.29) 0.26 (0.24–0.35) NS
at 6 h 0.21 (0.21–0.25) 0.26 (0.22–0.31) NS
at 12 h 0.21 (0.21–0.26) 0.24 (0.21–0.28) NS
at 24 h 0.21 (0.21–0.25) 0.21 (0.21–0.26) NS
Increase in FiO
2
at 1 h 0.015 (0.0–0.05) 0.05 (0.025–0.13) 0.043
at 6 h 0.0 (0.0–0.02) 0.045 (0.01–0.07) 0.033
at 12 h 0.0 (0.0–0.01) 0.025 (0.0–0.05) NS
at 24 h 0.0 (0.0–0.01) 0.0 (0.0–0.038) NS
n (%) X2 test§
Atelectasis present on 2 (1) 0 (0) NS
pre-extubation CXR?
CXR changes consistent with CLD 35 (15) 1 (17) NS
present at extubation?
Ever had a patent ductus 63 (27) 3 (50) NS
arteriosus?
Ever had septicaemia? 19 (8) 0 (0) NS
Postextubation NCPAP started 67 (29) 2 (33) NS
immediately postextubation?
NCPAP commenced some time 9/163 (6) 2/4 (50) <0.001
after extubation?‡

*NS, not significant.

†IPPV, intermittent positive pressure ventilation; PIP, peak inspiratory pressure; PEEP, positive end expiratory pressure.
‡Infants originally extubated without NCPAP who later required NCPAP (therefore excludes infants who were commenced on NCPAP immediately
after extubation).
§Uncorrected Chi-squared test.7

1977.1 Because of this high incidence of PEA and subsequent
morbidity, many nurseries performed routine postextubation
CXRs to diagnose PEA in an attempt to treat atelectasis with
more aggressive physiotherapy or to assess results of
intervention programmes such as routine chest physiotherapy
pre- and postextubation.
The most recent study was by Odita et al.,2 who reported an
incidence of PEA of 10.8% in 1993. In our study the incidence
of PEA was 2.5%. This is therefore a much lower incidence of
PEA than in all previously published reports. This compares
with earlier studies (Wyman and Kuhns1 and Odita et al.2) who
also defined PEA as only those cases where the atelectasis
was not present pre-extubation. Two of the infants with PEA
(33%) needed to commence NCPAP within 24 h postextubation,
however, no infant with PEA had a deterioration in respiratory
status requiring reintubation and ventilation. Although these
numbers are small, this represents a clinically significant
decrease in severe morbidity associated with PEA.
We attribute the decrease in the incidence and morbidity of
PEA to improvements in the care of ventilated neonates over
the last two decades. Improvements such as minimising
bronchial damage by only passing the suction catheter to the
tip of the ETT and no further, routine pre- and post extubation
physiotherapy,3,8 enhancement of mucociliary clearance of
bronchial secretions with attention to the temperature and
humidification of ventilator inspired gases,9,10 and the use of
nasopharyngeal CPAP immediately postextubation in VLBW
infants,11 all of which are routine practices in our ICN.
150
Previously PEA has been associated with a variety of factors,
only some of which were supported by our data. Prematurity
and lower birthweight,12 prolonged duration of intubation,1,12
sepsis2 and the occurrence of a patent ductus arteriosus2 have
all been associated with PEA, however, we found no such
association.
One previous study found a strong association between the
presence of atelectasis whilst intubated and subsequent PEA.12
However, our definition excludes those cases where atelectasis
was evident on the pre-extubation CXR. Nevertheless, the
presence of atelectasis on the pre-extubation CXR is strongly
predictive of atelectasis persisting postextubation—85% of
cases of atelectasis on the pre-extubation CXR showed
persistence of the atelectasis postextubation.
We found that PEA was associated with a small but
statistically significant rise in the pre- to postextubation
difference in FiO at 1 and 6 h postextubation. This is not
2
surprising given the ventilation/perfusion mismatch that occurs
with lung atelectasis.13 The significant increase in the need for
intervention with NCPAP is also not surprising, as the presence
of atelectasis will cause a decrease in overall lung compliance13
with subsequent increased respiratory effort or distress. All the
babies with PEA who required intervention with NCPAP did so
on clinical grounds because of increasing respiratory effort
or distress.
The pattern of PEA seen in our study shows a preponderance
for the RUL with 67% of cases involving this site alone. Other
studies have shown a similar affinity for right-sided collapse2,12,14
related to the greater degree of trauma that the right main
bronchus receives because of accidental low placement of
endotracheal tubes and damage secondary to suction cath-
eters,14 causing increased bronchial secretions,1,15 mucosal
oedema15 and loss of normal ciliated epithelium,14 thus causing
bronchial obstruction and subsequent collapse.
With any investigation one must ensure that any benefits
gained outweigh the disadvantages. With such a low incidence
of PEA, it does not seem that we can justify continuing to
perform routine postextubation CXRs. There were six cases of
PEA in a 12-month period, however, on two of these occasions
the CXR was performed early on clinical grounds (i.e. increasing
respiratory effort or distress). Of the remaining four cases, one
infant had chest physiotherapy continued beyond the period
for routine postextubation physiotherapy and three infants
had no documented change in management with no ill
effect.
CONCLUSION
In this nursery the incidence of PEA is low with no significant
morbidity. We have found that postextubation factors associated
MW Davies and DW Cartwright
with PEA include an increased oxygen requirement in the first
few hours postextubation, and the requirement for intervention
with NCPAP for increased respiratory effort or distress. We
therefore recommend that postextubation CXRs are no longer
necessary, and that they should be performed only on those
infants who have an increase in oxygen requirement postextub-
ation or become symptomatic with new or increasing respiratory
distress. We also recommend a follow-up postextubation CXR
on those infants known to have atelectasis on the most recent
pre-extubation CXR.
REFERENCES
1 Wyman ML, Kuhns LR. Lobar opacification of the lung after tracheal
extubation in neonates. J. Pediatr. 1977; 91: 109–12.
2 Odita JC, Kayyali M, Ammari A. Post-extubation atelectasis in
ventilated newborn infants. Pediatr. Radiol. 1993; 23: 183–5.
3 Finer NN, Moriartey RR, Boyd J, Phillips HJ, Stewart AR, Ulan O.
Postextubation atelectasis: A retrospective review and a prospective
controlled study. J. Pediatr. 1979; 94: 110–13.
4 Fox WW, Berman LS, Dinwiddie R, Shaffer TH. Tracheal extubation
of the neonate at 2–3 cm H O continuous positive airway pressure.
2
Pediatrics 1977; 59: 257–61.
5 Kuhn JP. Disorders of the peripheral airways and disturbances of
pulmonary aeration. In: Silverman FN, Kuhn JP, eds. Essentials of
Caffey’s Pediatric X-ray Diagnosis. Year Book Medical Publishers,
Chicago, 1990; 253–69.
6 Edwards DK. The child who wheezes. In: Hilton SVW, Edwards DK,
eds. Practical Pediatric Radiology, 2nd edn. WB Saunders,
Philadelphia, 1994; 85–111.
7 Haviland MG. Yates’s correction for continuity and the analysis of
2×2 contingency tables. Stat. Med. 1990; 9: 363–7.
8 Vivian-Beresford A, King C, MacAuley H. Neonatal post-extubation
complications: the preventative role of physiotherapy. Physiotherapy
Can. 1987; 39: 184–90.
9 Dick W. Aspects of humidification: Requirements and techniques.
Int. Anesthesiol. Clin. 1974; 12: 217–39.
10 Tarnow-Mordi WO, Reid E, Griffiths P, Wilkinson AR. Low inspired
gas temperature and respiratory complications in very low birth
weight infants. J. Pediatr. 1989; 114: 438–42.
11 Engelke SC, Roloff DW, Kuhns LR. Postextubation nasal continuous
positive airway pressure. Am. J. Dis. Child 1982; 136: 359–61.
12 Whitfield JM, Jones MD. Atelectasis associated with mechanical
ventilation for hyaline membrane disease. Crit. Care Med. 1980;
8: 729–32.
13 Hazinski TA. Atelectasis. In: Chernick V, Kendig EL, eds. Kendig’s
Disorders of the Respiratory Tract in Children, 5th edn. WB
Saunders, Philadelphia, 1990; 509–16.
14 Roper PC, Vonwiller JB, Fisk GC, Gupta JM. Lobar atelectasis after
nasotracheal intubation in newborn infants. Aust. Paediatr. J. 1976;
12: 272–5.
15 Spitzer AR, Fox WW. Postextubation atelectasis—the role of oral
vs nasal endotracheal tubes. J Pediatr 1982; 100: 806–10.