See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/259769362

Oxygen Saturation Targeting in Preterm

Infants Receiving Continuous Positive Airway
Pressure

Article in The Journal of pediatrics January 2014

Impact Factor: 3.79 DOI: 10.1016/j.jpeds.2013.11.072 Source: PubMed

CITATIONS READS

25 62

8 authors, including:

Kathleen Lim Kevin I Wheeler

The Royal Hobart Hospital The Royal Children's Hospital
6 PUBLICATIONS 32 CITATIONS 27 PUBLICATIONS 244 CITATIONS

SEE PROFILE SEE PROFILE

Hamish D Jackson Jennifer Anne Dawson

The Royal Hobart Hospital Royal Women's Hospital in Victoria
5 PUBLICATIONS 274 CITATIONS 103 PUBLICATIONS 1,526 CITATIONS

SEE PROFILE SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate, Available from: Jennifer Anne Dawson
letting you access and read them immediately. Retrieved on: 18 July 2016
 Oxygen Saturation Targeting in Preterm Infants Receiving Continuous
Positive Airway Pressure
Kathleen Lim1, Kevin I. Wheeler, MB, ChB1, Timothy J. Gale, PhD2, Hamish D. Jackson, BMBS (Hons)1, Jonna F. Kihlstrand3,4,
Cajsa Sand3,4, Jennifer A. Dawson, PhD4,5, and Peter A. Dargaville, MD1

Objective The precision of oxygen saturation (SpO2) targeting in preterm infants on continuous positive airway
pressure (CPAP) is incompletely characterized. We therefore evaluated SpO2 targeting in infants solely receiving
CPAP, aiming to describe their SpO2 profile, to document the frequency of prolonged hyperoxia and hypoxia ep-
isodes and of fraction of inspired oxygen (FiO2) adjustments, and to explore the relationships with neonatal intensive
care unit operational factors.
Study design Preterm infants <37 weeks gestation in 2 neonatal intensive care units were studied if they were
receiving CPAP and in supplemental oxygen at the beginning of each 24-hour recording. SpO2, heart rate, and FiO2
were recorded (sampling interval 1-2 seconds). We measured the proportion of time spent in predefined SpO2
ranges, the frequency of prolonged episodes ($30 seconds) of SpO2 deviation, and the effect of operational factors
including nurse-patient ratio.
Results A total of 4034 usable hours of data were recorded from 45 infants of gestation 30 (27-32) weeks (median
[IQR]). When requiring supplemental oxygen, infants were in the target SpO2 range (88%-92%) for only 31% (19%-
39%) of total recording time, with 48 (6.9-90) episodes per 24 hours of severe hyperoxia (SpO2 $98%), and
9.0 (1.6-21) episodes per 24 hours of hypoxia (SpO2 <80%). An increased frequency of prolonged hyperoxia in
supplemental oxygen was noted when nurses were each caring for more patients. Adjustments to FiO2 were
made 25 (16-41) times per day.
Conclusion SpO2 targeting is challenging in preterm infants receiving CPAP support, with a high proportion
of time spent outside the target range and frequent prolonged hypoxic and hyperoxic episodes. (J Pediatr
2014;164:730-6).

S
upplemental oxygen therapy is an integral part of modern neonatal intensive care and improves survival of preterm infants
with respiratory dysfunction.1,2 However, excess oxygen delivery is associated with adverse outcomes, in particular reti-
nopathy of prematurity,3 and hence there is a need to continuously adjust the fraction of inspired oxygen (FiO2) so as to
avoid the extremes of oxygenation. Measurement of oxygen saturation (SpO2) by pulse oximetry is now an indispensable tool for
guiding oxygen therapy in the newborn.4,5 Several randomized trials in extremely preterm infants have highlighted the importance
of targeting an appropriate SpO2 range, with a low target range (and resultant intermittent hypoxia) associated with increased
mortality, and a high target range with an increase in the risk of retinopathy of prematurity,6,7 and bronchopulmonary dysplasia.8
The recent randomized trials also highlight the difficulties of targeting a defined SpO2 range in the preterm infant, with a
wide range of SpO2 values on histograms of pooled data.7,9 These findings reinforce those of previous observational studies
of SpO2 targeting, in which preterm infants were noted to spend extended periods outside the target range when receiving sup-
plemental oxygen.10-18 These investigations have for the most part included both ventilated infants and those receiving contin-
uous positive airway pressure (CPAP) or other less-invasive forms of support, with in one case the suggestion of increased
difficulty in maintaining the target SpO2 range in infants on CPAP.17 The only study reporting data exclusively relating to
CPAP support had a very wide acceptable SpO2 range (87%-96%), which among the 12 study infants was successfully targeted
around 80% of the time during 3 hours of recordings.10 Further studies focusing on SpO2 targeting in preterm infants on CPAP
are required as uptake of this modality of respiratory support continues to increase.19
Beyond the overall proportion of time spent outside the target SpO2 range, few
studies have reported the frequency of discrete episodes of hypoxia and hyper-
1
From the Department of Pediatrics, Royal Hobart
oxia, the severity and extent of which is clearly important in the pathogenesis Hospital and University of Tasmania; School of 2

Engineering, University of Tasmania, Hobart, Tasmania,

pings
Australia; 3The Faculty of Health Sciences, Linko
pings, Sweden; and 4Newborn
University, Linko
5
CPAP Continuous positive airway pressure Research, The Royal Womens Hospital; and The
University of Melbourne, Melbourne, Victoria, Australia
FiO2 Fraction of inspired oxygen
Supported by the Royal Hobart Hospital Research
HR Heart rate Foundation, Hobart, Australia (starter grant 12-019). J.D.
NICU Neonatal intensive care unit received an Australian National Health and Medical
RHH Royal Hobart Hospital Research Council Post Doctoral Fellowship. The authors
declare no conflicts of interest.
RWH Royal Womens Hospital
SpO2 Oxygen saturation 0022-3476/$ - see front matter. Copyright 2014 Mosby Inc.
All rights reserved. http://dx.doi.org/10.1016/j.jpeds.2013.11.072

730

of complications arising therefrom. Short-lived hypoxic epi-
sodes (SpO2 <80% for $10 seconds) occurred relatively
frequently in 2 cohorts of preterm infants on mixed modes
of respiratory support.15,16 The incidence of longer-lasting
episodes of SpO2 disturbance with potentially more impact
are less well documented. Furthermore, the influence of
operational factors in the neonatal intensive care unit
(NICU), such as the nurse/patient ratio and NICU nursing
experience, on SpO2 targeting is incompletely understood,
with a suggestion of poorer saturation targeting by nurses
looking after more than one infant being a factor.17
We sought to better understand SpO2 targeting in a robust
sample of preterm infants solely receiving CPAP. We aimed
to describe their SpO2 profile and to document the frequency
of prolonged hyperoxia and hypoxia episodes, the relation-
ships with NICU operational factors, and the frequency of
FiO2 adjustments.
Methods
We conducted a prospective observational study in preterm
infants receiving CPAP and supplemental oxygen at 2 Austra-
lian tertiary neonatal centers, Royal Hobart Hospital (RHH),
Hobart and Royal Womens Hospital, Melbourne (RWH).
Both units use CPAP for initial and step-down respiratory
support in preterm infants. CPAP is generated with either a
bubble CPAP system (Fisher and Paykel HealthCare, East
Tamaki, New Zealand) or a mechanical ventilator (Babylog
8000 plus, Drager Medical Systems Inc, Notting Hill,
Australia) and delivered via Hudson binasal prongs (Hudson
Respiratory Care, Temecula, California), or a midline CPAP
delivery system (Flexitrunk; Fisher and Paykel HealthCare)
with alternating mask and prong interfaces. At the time of
the study, the target range for SpO2 was 88%-92% (inclusive)
in both centers, with the lower SpO2 alarm limit set at 85% and
the upper limit set at 94% when in supplemental oxygen, and
at 100% when in room air. The study period was February to
November 2012. Data collection was approved by our institu-
tional ethics committees as an audit of clinical practice.
Preterm infants <37 weeks gestation were studied if they
were younger than 4 months corrected gestational age and
receiving CPAP support. At RHH, consecutive data record-
ings each of 24 hours duration were made on all such infants
during the study period so long as a study investigator was
available. A new recording was commenced each day if the
infant remained in supplemental oxygen. At RWH, data re-
cordings were made on selected infants receiving CPAP
and supplemental oxygen. The maximum number of 24-
hour recordings for any one infant was set at 25. Recording
of data continued during periods of instability and deteriora-
tion, unless intubation was required. During the data record-
ings, all aspects of clinical management were undertaken as
per usual, including nursing allocation, infant handling and
care episodes, and parental contact.
Physiological data were recorded from standard bedside
monitors at a sampling frequency of 1 Hz (RHH) or 0.5 Hz
(RWH). SpO2 and heart rate (HR) were sourced from either
Vol. 164, No. 4  April 2014
a Drager Infinity Monitor (Drager Medical Systems Inc, Not-
ting Hill, Australia) or a Masimo Radical v4 oximeter
(Masimo Corp, Irvine, California). SpO2 averaging time
was set at its minimum,20 which was 2-4 seconds (Drager In-
finity monitor) or 2 seconds (Masimo). FiO2 was continu-
ously measured with an inline oxygen analyzer (MX300-I;
Teledyne Analytical Instruments, Industry, California), sam-
pling from the inspiratory limb of the CPAP circuit. SpO2
and HR data were digital signals extracted via X5 (Drager)
or 9-pin RS232 (Masimo) cables; FiO2 was an analog signal
digitized via an analog-digital converter (model USB-6008;
National Instruments, Austin, Texas). Data were input to a
laptop computer with custom software written using Lab-
VIEW (National Instruments). Bedside staff was not specif-
ically informed of the recording and did not have access to
the recording system. Demographic and clinical details of
the study infants were collected from the patient records
and charts. For patients at the RHH site, data were collected
for each nursing shift on nurse/patient ratio and the bedside
nurses level of NICU nursing experience (years).
Statistical Analyses
Data were processed using purpose-built software, with all
recorded data used for the analysis. Periods during which
there was SpO2 signal dropout were excluded from the anal-
ysis. The proportion of time spent in hypoxia (SpO2 <88%),
normoxia (ie, target range, SpO2 88%-92%), hyperoxia
(SpO2 >92%), and eupoxia (normoxia or hyperoxia in air)
was determined. As an indicator of severity, periods of hyp-
oxia were further divided into level I: 85%-87%; level II:
80%-84%; and level III: <80%. Similarly, hyperoxic periods
were subcategorized as level I: 93%-95%; level II: 96%-
97%; or level III: $98%. Periods of hypoxic bradycardia
(SpO2 <80% and HR #100 bpm) were also identified. Fre-
quency of prolonged episodes ($30 seconds) of each level
of hyperoxia and hypoxia, and of hypoxic bradycardia, was
documented. Frequency of FiO2 adjustments (defined as at
least 0.01 change in FiO2) was determined from the FiO2
signal.
To gain an overall snapshot of SpO2 targeting in the study
population, data from all recordings in all individuals were
pooled and presented as SpO2 histograms and proportions.
Further, summary statistics were derived for the study infants
by pooling all data from each individual. Proportions of time
in the predetermined SpO2 levels were determined, along
with the frequency of prolonged hypoxic and hyperoxic epi-
sodes, and of FiO2 adjustments, in each case reported as me-
dian and IQR. The effects of NICU operational factors on
SpO2 targeting were assessed using regression analysis of
data from each nursing shift, extracted from the 24-hour re-
cordings at RHH. Input variables were nurse/patient (1:1 vs
1:2 or greater, input as ordinal variable 0, 1), number of years
of NICU nursing experience (<5 years, 5-15 years, >15 years,
input as ordinal variable 0, 1, 2), and time of shift (comparing
evening shift 3 p.m. to 11 p.m. with day shift 7 a.m. to 3 p.m.,
and night shift 11 p.m. to 7 a.m. with day shift, in both cases
input as a dichotomous variable). Outcome variables were
731
THE JOURNAL OF PEDIATRICS  www.jpeds.com Vol. 164, No. 4

100 A
90
80
70
60
50
40
30
20
10 M M M M
0
0 4 8 12 16 20 24

100 B
90
80
70
60
50
40
30
20
10 M M M M M M M
0
0 4 8 12 16 20 24
Recording duration (h)

Figure 1. Sample recordings of SpO2 (solid line, Y-axis: % saturation) and FiO2 (dashed line, Y-axis: % oxygen). A, 34 weeks
gestation infant at day 2, 8 cm H2O CPAP. B, 26 weeks gestation infant, day 14, 8 cm H2O. M, missing SpO2 signal (defaults to
zero). Gray-shaded region, desired SpO2 range (88%-92%), successfully targeted for A, 51% of the recording, and B, 32%.
Frequent and prolonged hyperoxic and hypoxic episodes are noted, and multiple FiO2 adjustments have been made to correct
SpO2 deviations.

the proportion of time in eupoxia for each shift, and the fre-
quency per shift of prolonged level 3 hypoxic and hyperoxic
episodes, as defined above. A confounding effect of corrected
gestational age was sought by regression with and without
this input variable. Comparison of the frequency of FiO2 ad-
justments between shifts was made using the Kruskal-Wallis
nonparametric ANOVA. A P < .05 was considered statisti-
cally significant.
Results
A total of 45 infants were studied, 32 at RHH, and 13 at
RWH. Median gestation was 30 weeks (IQR 27-32 weeks),
birth weight 1.3 (0.93-1.8) kg, and age at first recording 1
(0-8) days (Table I; available at www.jpeds.com). At the
onset of each recording, infants were receiving support
732
with a median CPAP level of 7 (6-8) cmH2O and FiO2 0.25
(0.22-0.29).
In all, 4113 hours of data were recorded, of which 79 hours
(1.2%) were excluded for missing SpO2 signal. Of the 4034
hours of usable data, 2971 hours (74%) were recorded
when infants required supplemental oxygen, and 1063 hours
(26%) when infants were receiving air. Per patient, a median
of 47 hours (IQR 23-66) of usable data were available (range,
5.5-578 hours). This included 23 hours (19-49) when supple-
mental oxygen was being provided (range, 1-541 hours).
Sample recordings of SpO2 and FiO2 are shown in
Figure 1. These demonstrate the brief periods of missing
SpO2 data, frequent fluctuations in SpO2, and multiple
FiO2 adjustments by the bedside staff. Despite these
adjustments, SpO2 was kept in the target range only 51%
and 32% of the time in the 2 recordings.
Lim et al
April 2014
Frequency histograms of pooled data reveal a wide spread
of SpO2 readings with a preponderance of hyperoxic values,
in part attributable to the inclusion of data while in room air
(Figure 2, A and B). Pooled data while the infant was in
supplemental oxygen show a considerable proportion of
the SpO2 signal to fall outside the intended target range,
with both significant hypoxia and hyperoxia well
represented in the histogram (Figure 2, C). On a per-
patient basis, when receiving supplemental oxygen, SpO2
was within the target range for a median of only 31% (IQR
19-39) of the time (Table II), with most readings in the
mildly hyperoxic range. Pronounced (level 3) hyperoxia
(SpO2 $98%) and hypoxia (SpO2 <80%) were noted 7.8%
and 1.1% of the time, respectively. Hypoxic bradycardia
was rare. When the infants were in room air, SpO2 was in
the target range 19% (6.6-32) of the time, with median
proportion of time at levels 1, 2, and 3 hyperoxia being
29%, 23%, and 8.6%, respectively.
Prolonged episodes of hyperoxia when in supplemental
oxygen, and of hypoxia in air or oxygen, were common
(Table II), with, on average, a period of level 3 hyperoxia
in oxygen every 30 minutes and of level 3 hypoxia every 3.3
hours. Overall, prolonged hypoxic bradycardia events were
rare, occurring more than once per 24 hours in only 7
(16%) infants.
Several operational factors were found to have an influence
on the success of SpO2 targeting. Episodes of level 3 hyper-
oxia (SpO2 $98% in oxygen) were more frequent when
nurses were allocated more than one infant (Table III).
Additionally, care provision by more experienced NICU
nurses was associated with less time in eupoxia. No effect
of the time of shift on SpO2 targeting was observed.
Corrected gestational age did not affect these results in any
meaningful way, but greater corrected gestation was
independently associated with fewer hyperoxic episodes
(P = .0027).
A total of 8168 FiO2 adjustments were made over the entire
4113 hours, with a median frequency of 25 (IQR 16-41) ad-
justments per 24 hours of individual patient data, including
12 (7-20) FiO2 increments and 14 (9-24) decrements per
24 hours. The frequency of adjustments was highly variable
per patient, ranging from 2 to 166 adjustments per patient
per 24 hours. Fewer FiO2 adjustments were made during
night shift, 0.63 (0.11-1.4) adjustments per hour, compared
with either morning shift, 1.0 (0.17-2.67) or evening shift,
1.6 (0.33-4.01), P < .05, Kruskal-Wallis ANOVA with Dunns
post test.
Discussion
Despite controversies regarding what the optimal SpO2 target
range for preterm infants should be, the importance of SpO2
is indisputable, particularly when supplemental oxygen is be-
ing delivered. We evaluated SpO2 targeting in preterm in-
fants receiving CPAP and found that when receiving
oxygen, infants spent only 31% of the time in the target
SpO2 range. Prolonged episodes of serious hypoxia, and of
Oxygen Saturation Targeting in Preterm Infants Receiving Continuous Positive Airway Pressure
ORIGINAL ARTICLES
hyperoxia in oxygen, were relatively frequent, with hypoxic
events more likely when nursing staff were allocated a greater
number of patients. Finally, we found that even on less inva-
sive respiratory support in the form of CPAP, many adjust-
ments to FiO2 were required in an effort to achieve normoxia.
Nasal CPAP is increasingly used in the early respiratory
management of preterm infants who would have previously
been intubated and ventilated,19 and hence the relevance of
information on SpO2 targeting specific to this form of sup-
port. Ours was a heterogeneous cohort from two centers,
including preterm infants of a wide spread of gestational
and postnatal ages, but nevertheless representative of the
population of infants managed with CPAP in the NICU.
Previous studies of SpO2 targeting have generally included
preterm infants both on full mechanical ventilation and
noninvasive ventilation, including CPAP.6,7,9,11-18 The gesta-
tional age ranges, SpO2 limits, and proportion of included
time in ambient air have varied widely between studies, mak-
ing comparisons difficult. What is clear is that targeting a
relatively narrow SpO2 range with a span of 5 values (eg,
88%-92%) appears to be at least as difficult in infants on
CPAP as on those receiving mechanical ventilation. Previous
comparable studies of preterm infant cohorts (including
both invasive and non-invasive support) have noted the pro-
portion of time in the target range to be 16%-55% (when in
air and supplemental oxygen)11 and 20%-50% (supple-
mental oxygen only),7 compared with 31% in the present
study. Other investigations in which infants were managed
with a wider range of acceptable SpO2 values not surprisingly
show a greater proportion of time within the target range.
One of these studied infants exclusively on CPAP, and with
a SpO2 target range of 87%-96% and noted effective targeting
by the bedside nursing staff for around 80% of the 90-minute
recording periods.10
As with previous investigations, the most common devia-
tion of SpO2 from the target range in infants receiving
oxygen was into the zone of mild hyperoxia (SpO2 93 to
95%).6,7,9,11-18 This degree of hyperoxia mostly falls within
the monitor alarm limits and is not associated with a danger-
ously elevated PaO2.21,22 The preponderance of time spent in
level 1 hyperoxia compared with level 1 hypoxia (SpO2 85 to
87%) reflects the tendency for bedside staff to target the up-
per end of the desired SpO2 range, as has been previously
noted during manual FiO2 control.13
Beyond mild hyperoxia, infants in our study also had more
severe SpO2 deviations potentially associated with unaccept-
able PaO2 values.21,22 When in supplemental oxygen, pre-
term infants on CPAP spent a median of 21% of the time
with SpO2 $96% (compared with 8.1%-45% in other studies
with mixed respiratory support),7,9,12,13 and 7.8% of the time
with SpO2 $98% (compared with 16%).13 Our study infants
demonstrated a comparatively low proportion of time with
significant hypoxia: 3.0% of time with SpO2 <85%
(compared with 3.1%-27%),7,9,13,17 and 0.9% of time with
SpO2 <80% (compared with 1.9%-9.8%).9,12,13
Unquestionably, it is the combination of pronounced and
prolonged SpO2 deviation that raises the most concern in
733
THE JOURNAL OF PEDIATRICS  www.jpeds.com Vol. 164, No. 4

10 A
9

Proportion of time (%)

8
7
6
5
4
3
2
1
0
0 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 100

16 B
14
Proportion of time (%)

12

10

0
0 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 100

10 C
9
Proportion of time (%)

8
7
6
5
4
3
2
1
0
0 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 100
SpO2 (%)

Figure 2. Pooled frequency histograms for time spent in each SpO2 band. A, Data when receiving supplemental oxygen or
ambient air; B, air; and C, supplemental oxygen. Gray-shaded region, target SpO2 range (88%-92%).

734 Lim et al
April 2014 ORIGINAL ARTICLES

Table II. SpO2 profile and hyperoxic/hypoxic episodes

Receiving oxygen (2971 h) Receiving air or oxygen (4034 h)
Proportion of Prolonged episodes Proportion Prolonged episodes
time (%) (number per 24 h) of time (%) (number per 24 h)
Hyperoxia
Level 3, SpO2 $98% 7.8 (1.3-19) 48 (6.9-90) - -
Level 2, SpO2 96%-97% 13 (7.0-20) 81 (48-161) - -
Level 1, SpO2 93%-95% 26 (20-33) 194 (155-250) - -
All levels, SpO2 $93% 59 (36-74) - - -
Normoxia/eupoxia* 31 (19-39) - 53 (40-63) -
Hypoxia
All levels, SpO2 #87% 9 (4.3-18) - 8.8 (3.1-19) -
Level 1, SpO2 85%-87% 5.5 (2.6-9.8) 38 (10-65) 5.0 (2.4-9.3) 34 (8.8-58)
Level 2, SpO2 80%-84% 2.2 (0.9-5.3) 19 (5.5-34) 2.1 (0.6-6.4) 13 (4.1-36)
Level 3, SpO2 <80% 1.1 (0.2-3.5) 9.0 (1.6-21) 0.9 (0.1-3.5) 7.3 (1.3-21)
Hypoxic bradycardia, SpO2 <80% + 0.01 (0-0.01) 0.0 (0.0-0.26) 0.02 (0- 0.14) 0.0 (0.0-0.49)
HR <100

Median (IQR) for proportion of time within predefined SpO2 levels, and frequency of prolonged ($30 s) episodes.
*Normoxia: SpO2 88%-92% when in supplemental oxygen; eupoxia: SpO2 88%-92% in oxygen, 88%-100% in air.

preterm infants, although for neither severity nor duration is
there a clear understanding of the threshold beyond which
complications are likely. In rat pups, exposure to 20-second
spells of severe hypoxia (FiO2 0.10), repeated every minute
for 2-6 hours on days 7-11 appears to impair neurotransmis-
sion and executive function in later life.23 To our knowledge,
no experimental studies have linked shorter or less-frequent
periods of hyper/hypoxia to adverse outcomes. In human in-
fants, Di Fiore et al15 found a relationship between frequent
intermittent hypoxic episodes (defined as SpO2 <80% for
$10 seconds) and later retinopathy requiring laser therapy.
In a further study the episodes identified by this definition
were noted to have an average duration of well over 20 sec-
onds.24 With these considerations in mind we chose to define
prolonged SpO2 deviations as being of duration $30 seconds
and found that even with this more stringent definition, such
episodes were relatively frequent. Unlike most other studies,
we were able to identify with certainty whether supplemental
oxygen was being administered during episodes of hyperoxia.
When in oxygen, level 3 hyperoxia episodes (SpO2 $98%)
occurred at a frequency of 2 per hour in our study, compared
with 4.1 per hour14 and 9.3 per hour10 in studies with mini-
mum episode duration defined as 10 and 5 seconds, respec-
tively.
We found that a patient allocation of more than 1 infant
per nurse was associated with a greater frequency of signifi-
cant hyperoxia (SpO2 $98%) when infants were in supple-
mental oxygen and a trend towards less eupoxia. A similar
finding was reported by Sink et al17 and is presumably attrib-
utable to the greater nursing workload and multiplicity of
tasks when caring for more than one infant on respiratory
support. The finding that less-experienced nursing staff
members were more successful in SpO2 targeting was inter-
esting. This could reflect more experienced nurses waiting a
longer time for self-recovery, or attempting other interven-
tions before adjusting delivered FiO2. Further studies will
be necessary to confirm this observation and understand its
causation.
Overall, we observed FiO2 adjustments at a frequency of
around 1 per hour, which is similar to that previously reported
for infants on CPAP10 and lower than the range of values re-
ported for ventilated infants (3.0-8.5 adjustments per

Table III. Relationship between SpO2 targeting and NICU operational factors
Outcome variable
Proportion of time Frequency of prolonged Frequency of prolonged
in eupoxia hyperoxia when in oxygen (SpO2 98%) hypoxia (SpO2 <80%)
Operational Coefficient Coefficient
factor (95% CI) P value (95% CI) P value Coefficient (95% CI) P value
Nurse/patient* 3.1 ( 8.8, 2.6) .29 13 (3.0, 23) .011 0.46 ( 1.8, 2.7) .68
(1:1 vs 1:2 or greater)
NICU nursing experience* 3.2 ( 6.1, 0.27) .033 1.8 ( 6.8, 3.2) .48 0.033 ( 1.1, 1.2) .96
(<5 y vs 5-15 y vs >15 y)
Afternoon shift (3 p.m.-11 p.m.) 1.7 ( 5.1, 1.7) .33 5.4 ( 11, 0.23) .060 0.29 ( 1.0, 1.6) .65
Night shift (11 p.m. to 7 a.m.) 0.43 ( 3.4, 2.6) .78 3.0 ( 3.0, 8.9) .33 0.64 ( 2.0, 0.70) .35

Linear regression analysis of relationship between NICU operational factors (input variables) and SpO2 targeting outcomes (dependent variables). A single value extracted for each outcome variable
per shift, all shifts treated independently, all time in oxygen and air included. A total of 32 infants were included, with 3756 h of data recording. Prolonged hyperoxia/hypoxia: episodes per shift lasting
$30 s.
*Input as ordinal variables: 0, 1 or 0, 1, 2.
Afternoon and night shifts were compared with morning shift (7 a.m. to 3 p.m.).

Oxygen Saturation Targeting in Preterm Infants Receiving Continuous Positive Airway Pressure 735
THE JOURNAL OF PEDIATRICS  www.jpeds.com
hour).13,14,18 This finding is very likely related to the greater
proportion of 1:1 nursing care among ventilated infants.
Our study has the limitations of being an observational study
in a relatively small group of infants, of whom only one-quarter
were <28 weeks gestation. Although the recording devices
were concealed behind standard equipment, it is possible that
awareness of the study on the part of the bedside nurse had
an effect on the efficacy of SpO2 targeting. Finally, even though
we consider the results to be widely generalizable, they have
most relevance to centers targeting a relatively narrow range
of SpO2 values (span of 5, eg, 88%-92%, 90%-94%).
In conclusion, SpO2 targeting is challenging in preterm in-
fants receiving CPAP support, with a high proportion of time
spent outside the target range, and frequent occurrence of
prolonged hypoxic and hyperoxic episodes. Efforts to more
effectively target SpO2, including feedback control of
FiO2,25 should be pursued vigorously. n
We thank the parents of infants involved in this study, the nursing staff
at Royal Hobart Hospital, Hobart and Royal Womens Hospital, Mel-
bourne, and Alex Nolan, Long Truong, and Cal Gerard, School of En-
gineering, University of Tasmania.
Submitted for publication Sep 16, 2013; last revision received Oct 30, 2013;
accepted Nov 27, 2013.
Reprint requests: Peter A. Dargaville, MD, Department of Paediatrics, Royal
Hobart Hospital, 48 Liverpool St, Hobart, TAS 7000, Australia. E-mail: peter.
dargaville@dhhs.tas.gov.au
References
1. Bolton DP, Cross KW. Further observations on cost of preventing retro-
lental fibroplasia. Lancet 1974;1:445-8.
2. Silverman WA. A cautionary tale about supplemental oxygen: the alba-
tross of neonatal medicine. Pediatrics 2004;113:394-6.
3. Askie LM, Henderson-Smart DJ, Ko H. Restricted versus liberal oxygen
exposure for preventing morbidity and mortality in preterm or low birth
weight infants. Cochrane Database Syst Rev 2009;CD001077.
4. Tin W, Gupta S. Optimum oxygen therapy in preterm babies. Arch Dis
Child Fetal Neonatal Ed 2007;92:F143-7.
5. Greenspan JS, Goldsmith JP. Oxygen therapy in preterm infants: hitting
the target. Pediatrics 2006;118:1740-1.
6. Carlo WA, Finer NN, Walsh MC, Rich W, Gantz MG, Laptook AR, et al.
Target ranges of oxygen saturation in extremely preterm infants. N Engl J
Med 2010;362:1959-69.
7. Stenson BJ, Tarnow-Mordi WO, Darlow BA, Simes J, Juszczak E,
Askie L, et al. Oxygen saturation and outcomes in preterm infants. N
Engl J Med 2013;368:2094-104.
8. STOP-ROP Multicenter study group. Supplemental Therapeutic Oxy-
gen for Prethreshold Retinopathy Of Prematurity (STOP-ROP), a ran-
736
Vol. 164, No. 4
domized, controlled trial. I: primary outcomes. Pediatrics 2000;105:
295-310.
9. Schmidt B, Whyte RK, Asztalos EV, Moddemann D, Poets C, Rabi Y,
et al. Effects of targeting higher vs lower arterial oxygen saturations on
death or disability in extremely preterm infants: a randomized clinical
trial. JAMA 2013;309:2111-20.
10. Urschitz MS, Horn W, Seyfang A, Hallenberger A, Herberts T, Miksch S,
et al. Automatic control of the inspired oxygen fraction in preterm in-
fants: a randomized crossover trial. Am J Respir Crit Care Med 2004;
170:1095-100.
11. Hagadorn JI, Furey AM, Nghiem TH, Schmid CH, Phelps DL,
Pillers DA, et al. Achieved versus intended pulse oximeter saturation
in infants born less than 28 weeks gestation: the AVIOx study. Pediatrics
2006;118:1574-82.
12. Laptook AR, Salhab W, Allen J, Saha S, Walsh M. Pulse oximetry in very
low birth weight infants: can oxygen saturation be maintained in the
desired range? J Perinatol 2006;26:337-41.
13. Claure N, DUgard C, Bancalari E. Automated adjustment of inspired
oxygen in preterm infants with frequent fluctuations in oxygenation: a
pilot clinical trial. J Pediatr 2009;155:640-5.
14. Claure N, Bancalari E, DUgard C, Nelin L, Stein M, Ramanathan R, et al.
Multicenter crossover study of automated control of inspired oxygen in
ventilated preterm infants. Pediatrics 2011;127:e76-83.
15. Di Fiore JM, Bloom JN, Orge F, Schutt A, Schluchter M, Cheruvu VK,
et al. A higher incidence of intermittent hypoxemic episodes is associated
with severe retinopathy of prematurity. J Pediatr 2010;157:69-73.
16. Di Fiore JM, Walsh M, Wrage L, Rich W, Finer N, Carlo WA, et al. Low
oxygen saturation target range is associated with increased incidence of
intermittent hypoxemia. J Pediatr 2012;161:1047-52.
17. Sink DW, Hope SA, Hagadorn JI. Nurse:patient ratio and achievement
of oxygen saturation goals in premature infants. Arch Dis Child Fetal
Neonatal Ed 2011;96:F93-8.
18. van der Eijk AC, Dankelman J, Schutte S, Simonsz HJ, Smit BJ. An obser-
vational study to quantify manual adjustments of the inspired oxygen
fraction in extremely low birth weight infants. Acta Paediatr 2012;101:
e97-104.
19. Morley CJ, Davis PG. Continuous positive airway pressure: scientific and
clinical rationale. Curr Opin Pediatr 2008;20:119-24.
20. Ahmed SJ, Rich W, Finer NN. The effect of averaging time on oximetry
values in the premature infant. Pediatrics 2010;125:e115-21.
21. Castillo A, Sola A, Baquero H, Neira F, Alvis R, Deulofeut R, et al. Pulse
oxygen saturation levels and arterial oxygen tension values in newborns
receiving oxygen therapy in the neonatal intensive care unit: is 85% to
93% an acceptable range? Pediatrics 2008;121:882-9.
22. Quine D, Stenson BJ. Arterial oxygen tension (PaO2) values in infants
<29 weeks of gestation at currently targeted saturations. Arch Dis Child
Fetal Neonatal Ed 2009;94:F51-3.
23. Coleman RJ, Beharry KD, Brock RS, Abad-Santos P, Abad-Santos M,
Modanlou HD. Effects of brief, clustered versus dispersed hypoxic epi-
sodes on systemic and ocular growth factors in a rat model of oxygen-
induced retinopathy. Pediatr Res 2008;64:50-5.
24. Di Fiore JM, Kaffashi F, Loparo K, Sattar A, Schluchter M, Foglyano R,
et al. The relationship between patterns of intermittent hypoxia and reti-
nopathy of prematurity in preterm infants. Pediatr Res 2012;72:606-12.
25. Claure N, Bancalari E. Automated closed loop control of inspired oxygen
concentration. Respir Care 2013;58:151-61.
Lim et al
April 2014 ORIGINAL ARTICLES

Table I. Characteristics of enrolled patients

Characteristics Median (IQR)/n (%)
Gestational age, wk 30 (27-32)
Birth weight, kg 1.3 (0.93-1.8)
Male sex 25 (56%)
Cesarean delivery 30 (67%)
Received antenatal steroids 38 (84%)
Apgar score
1 min 5 (4-8)
5 mins 8 (6-9)
Respiratory diagnosis*
RDS 33 (73%)
AOP 6 (13%)
Other 6 (13%)
Age at first recording, d 1 (0-8)
Previously intubated 5 (11%)
Caffeine administered 36 (80%)
Nasal IPPV used 1 (2%)

AOP, apnea of prematurity; IPPV, intermittent positive pressure ventilation; RDS, respiratory
distress syndrome.
*Initial diagnosis resulting in need for respiratory support.

Oxygen Saturation Targeting in Preterm Infants Receiving Continuous Positive Airway Pressure 736.e1