Randomized Controlled Trial of Oxygen Saturation Targets in Very PretermInfants: Two Year Outcomes

Brian A. Darlow, MD, FRACP1, Simone L. Marschner, MSc2, Mark Donoghoe, BSc2, Malcolm R. Battin, MD, MPH, FRACP3,

Roland S. Broadbent, MB, ChB, FRACP4, Mark J. Elder, MD, FRANZCO, FRACS5, Michael P. Hewson, MB, ChB, FRACP6,

Michael P. Meyer, MD, FRACP7, Alpana Ghadge, PhD2, Patricia Graham, BN1, Nicolette J. McNeill, Dip N1,

Carl A. Kuschel, MB, ChB, FRACP3,8, and William O. Tarnow-Mordi, MB, ChB, FRCPCH9, on behalf of the

Benefits Of Oxygen Saturation Targeting-New Zealand (BOOST-NZ) Collaborative Group*

Objective To assess whether an oxygen saturation (SpO2) target of 85%-89% compared with 91%-95% reducedthe incidence of the composite outcome of death or major disability at 2 years of age in infants born at <28 weeks’gestation.Study design A total 340 infants were randomized to a lower or higher target from <24 hours of age until 36weeks’ gestational age. Blinding was achieved by targeting a displayed SpO2 of 88%-92% using a saturationmonitor offset by �3% within the range 85%-95%. True saturations were displayed outside this range. Follow-up at 2 years’ corrected age was by pediatric examination and formal neurodevelopmental assessment. Majordisability was gross motor disability, cognitive or language delay, severe hearing loss, or blindness.Results The primary outcome was known for 335 infants with 33 using surrogate language information. Targetinga lower comparedwith a higher SpO2 target range had no significant effect on the rate of death ormajor disability at 2years’ corrected age (65/167 [38.9%] vs 76/168 [45.2%]; relative risk 1.15, 95%CI 0.90-1.47) or any secondary out-comes. Death occurred in 25 (14.7%) and 27 (15.9%) of those randomized to the lower and higher target, respec-tively, and blindness in 0% and 0.7%.Conclusions Although there was no benefit or harm from targeting a lower compared with a higher saturation inthis trial, further information will become available from the prospectively planned meta-analysis of this and 4 othertrials comprising a total of nearly 5000 infants. (J Pediatr 2014;-:---).

See editorial, p ���

xygen is the most common treatment given to very preterm (<28 weeks’ gestation [VP]) infants. Like any drug, too

Omuch and too little can cause harm, but it remains unclear where the balance should lie to minimize the competingadverse outcomes of retinal damage, chronic lung disease (CLD), neurodevelopmental impairment, and death.1

Prior to the 1990s, oxygen therapy in VP infants was usually monitored by either continuous or intermittent arterial oxygentension, and a target range of 50-80 mm Hg was widely accepted, although this was based on professional opinion.2 Through

From the 1Department of Pediatrics, University of Otago,Christchurch, New Zealand; 2National Heath andMedical Research Council Clinical Trials Center,University of Sydney, Sydney, Australia; 3NewbornServices, Auckland City Hospital, Auckland, NewZealand; 4Women’s and Children’s Health, University ofOtago, Dunedin, New Zealand; 5Academic Departmentof Surgery, University of Otago, Christchurch, NewZealand; 6Neonatal Intensive Care Unit, Wellington

the 1990s, most neonatal intensive care units (NICUs) adopted pulse oximeteroxygen saturation (SpO2) monitoring because it was simple and noninvasiveand reduced complications, including the need for blood transfusions. Withfew data to guide SpO2 targets many clinicians chose a range of 90%-95%,although surveys showed high and low limits varied from 100% to 80%.3

Several studies, including 2 randomized controlled trials in convalescing in-fants,4,5 suggested lower targets might be beneficial. An observational studyfrom England reported that units targeting a SpO2 range of 80%-90% from

Regional Hospital, Wellington South, New Zealand;7KidzFirst, Middlemore Hospital, Otahuhu, Auckland,New Zealand; 8Neonatal Services, The Royal Women’sHospital, Melbourne, Victoria, Australia; and 9WINNERCenter for Newborn Research, National Health andMedical Research Council Clinical Trials Center,University of Sydney, Sydney, Australia

*A list of members of the BOOST-NZCollaborativeGroupis available at www.jpeds.com (Appendix).

Funded by New Zealand Health Research Council (05/145) and the Child Health Research Foundation (CureKids). The authors declare no conflicts of interest.

Registered with Australian New Zealand Clinical TrialsRegistry (http://www.anzctr.org.au/):ACTRN012605000253606.

0022-3476/$ - see front matter. Copyright ª 2014 Mosby Inc.

All rights reserved. http://dx.doi.org/10.1016/j.jpeds.2014.01.017

BII Bayley Scales of Infant

Development, Second Edition

BIII Bayley Scales of Infant

Development, Third Edition

BOOST Benefits Of Oxygen Saturation

Targeting

CLD Chronic lung disease

COT Canadian Oxygen Trial

CP Cerebral palsy

CTC Clinical Trials Centre

IPD Individual patient data

NICU Neonatal intensive care unit

ROP Retinopathy of prematurity

RR Relative risk

SpO2 Oxygen saturation

SUPPORT Surfactant, Positive Pressure, and

Pulse Oximetry Randomized

Trial

VP Very preterm

1

THE JOURNAL OF PEDIATRICS � www.jpeds.com Vol. -, No. -

shortly after birth in VP infants had lower rates of severe reti-nopathy of prematurity (ROP) and CLD than did unitsadopting a range of 94%-98% but with no differences inmortality or cerebral palsy (CP) at 1 year.6 A single-unit studyfrom California reported that lowering the SpO2 target to85%-93% for infants of <32 weeks’ gestation and to 83%-93% for the smallest and sickest infants was associated witha decrease in severe ROP and trends to increased survivalin infants with birth weights <1250 g.7 But while many NI-CUs adopted lower SpO2 targets in these VP infants, there re-mained concerns that the risks and benefits of thistherapeutic drift were unknown and particularly whethershort-term benefits such as less ROP might be achieved atthe expense of worse long-term neurodevelopmental out-comes.8,9

Because obtaining funding for a single large trial seemedunlikely, separate trials were planned and ultimately 5 werefunded. There was agreement from the outset to adopt similarprotocols and to plan for a prospective individual patient data(IPD) meta-analysis following publication of the individualtrials.10 The hypothesis for the Benefits Of Oxygen SaturationTargeting (BOOST)-New Zealand trial was that a lower SpO2target range (85%-89%), compared with a higher target range(91%-95%), would reduce the incidence of a compositeoutcome of death or major disability at 2 years of age cor-rected for prematurity in infants of <28 weeks’ gestation.

Methods

This was a double-blind randomized controlled trial in 5 ofthe 6 regional NICUs in New Zealand. The BOOST-NewZealand coordinators in Christchurch managed the trial incollaboration with the National Health andMedical ResearchCouncil Clinical Trials Centre (CTC), Sydney, Australia,which was responsible for randomization and data analysis.The BOOST-New Zealand study was monitored by the inde-pendent data and safetymonitoring committee appointed forthe Australian BOOST II study. The study was approved bytheMulti-region Ethics Committee of theMinistry of Health.

Eligible infants were <28 weeks’ gestation, <24 hours ofage, and either born in or transferred into a trial NICU.Exclusion criteria were a congenital anomaly affectingoxygenation or long-term development, imminent death,or the inability to follow up at 2 years (principally non–En-glish-speaking parents or known to be moving overseas).Written informed parental consent was obtained.

Randomization occurred before 24 hours of age via tele-phone call to the CTC. Computer-generated randomizationlists were prepared by an independent CTC statistician.Randomization was stratified by NICU, sex, gestation <26or $26 weeks, and inborn or outborn. Multiple births wererandomized separately. Each center had a pool of oximeters,identified by a unique number and maintained by the studycoordinator in conjunction with the independent oximetercontrol center at Technical Services, Canterbury DistrictHealth Board, who alone were aware of the offset. Fifty studyoximeters, 25 with a +3% and 25 with a �3% masked offset

2

within the range of 85%-95%, were leased fromMasimo Cor-poration (Irvine, California).The SpO2 target for all infants was 88%-92% when the in-

fant was receiving supplemental oxygen. Nursing staff re-corded the inspired oxygen concentration and SpO2 hourly,and the stored histogram of the percentage of time the infantspent in different saturation bands over the past 24 hours wasdisplayed and recorded at midnight each day. The infant re-mained on the assigned study oximeter for at least the first 2weeks of life and until 36 weeks’ gestational age unless notrequiring supplemental oxygen or respiratory support andwith a SpO2 >96% for >95% of the time for 3 days. Study oxi-meters were prominently labeled and only used in the study.When first attached to the infant, this was after a gap of atleast 5 minutes without saturation monitoring. When thestudy ended and the study oximeter was removed, therewas a gap of at least 30 minutes before a standard nonstudyoximeter was attached. If deemed clinically necessary tomonitor the infant with a nonstudy monitor without theoffset, such as during surgery, the protocol required at leasta 5-minute gap between discontinuing the study monitorand attaching a standard monitor.The monitor alarm settings were recommended (but not

mandated) to be at 87% and 93% when in supplemental ox-ygen. The SpO2 levels were stored every 10 seconds. Thestored data were downloaded every 14 days by the researchnurse in each collaborating NICU. If an infant was trans-ferred to a lower level (level II) NICU before 36 weeks’ gesta-tional age, the same study monitor and protocol were used.Before the study, the collaborating centers, including level

II centers, were visited by a study coordinator, who under-took education sessions on all aspects of the trial, optimizingcompliance with saturation targets, viewing the oximeter his-togram, and downloading stored data. A research nurse ineach center facilitated ongoing education and surveillanceand coordinated data collection and oximeter downloadsand sent via courier oximeters to Christchurch, via courier,when an infant completed the study. Unlike similar studiesin Australia and the United Kingdom11 and in Canada,12

the New Zealand trial did not change the oximeter softwareduring the conduct of the trial.Research nurses recorded all data using standardized defi-

nitions. Data were entered into the CTC web-based datacollection (Inform) system. The BOOST-New Zealand coor-dinators audited data from each site at least annually. All in-fants were assessed as close to 2 years’ corrected age aspossible based on pediatric examination and a neurodevelop-mental assessment by a registered psychologist. All caregiversand assessors remained blinded as to the treatment group ofthe infant.The prespecified primary outcome for the study was a

composite of death or major disability at 24 months’ cor-rected age. Major disability was originally defined as any ofa Mental Developmental Index on the Bayley Scales of InfantDevelopment, Second Edition13 (BII) <70, CP with GrossMotor Function Classification System level$2, severe visualloss (legally blind, <6/60 vision), or deafness requiring

Darlow et al

- 2014 ORIGINAL ARTICLES

hearing aids. Because the Bayley Scales of Infant Develop-ment, Third Edition (BIII)14 scales become standard inNew Zealand before follow-up started, the neurodevelop-mental criteria were changed, in line with the Neonatal Oxy-gen Prospective Meta-analysis Collaboration,10 to $1 BIIIDevelopmental Assessment composite cognitive score <85or a BIII Developmental Assessment composite languagescore <85. When an infant could only be assessed in a levelII center that used the BII scales, an Mental DevelopmentalIndex <70 was used for the primary outcome.

Johnson found a vocabulary of <18 words to be a valid in-dicator of significant cognitive delay at 2 years (sensitivity0.81, specificity 0.71).15 Where the BII or BIII was missingin a child without CP, blindness, or deafness, we defined“major disability” as the more stringent <10 words(including signed) by parent report at the pediatric assess-ment. If none of these data were available, the primary endpoint was considered missing.

Secondary outcomes were severe ROP ($stage 3, or retinalsurgery), oxygen dependency or respiratory support at 36weeks’ gestational age, days of continuous positive airwaypressure, days of endotracheal intubation, days of oxygena-tion in both hospital and days at home, a patent ductus arte-riosus diagnosed by echocardiography and requiringtreatment, necrotizing enterocolitis requiring surgery or acause of death, weight at 2 years’ corrected age, readmissionsto hospital by 2 years’ corrected age, any respiratory illness(home oxygen, asthma, chest symptoms more than once aweek) between discharge and 2 years, CP, mean cognitionscore and mean language score on BIII, brain injury (intra-ventricular hemorrhage $grade 3, periventricular leukoma-lacia, or porencephalic cysts or contributing to death), anddeath from pulmonary causes from 4 weeks of age to 2 years’corrected age.

Statistical analysesAll analyses were undertaken using the intention-to-treatprinciple. The effect of SpO2 on the probability of death ormajor disability was assessed using generalized estimatingequations with a binomial model and a log link function ac-counting for the dependency of multiple births. Results arepresented as relative risk (RR) with 95% CIs and 2-sided Pvalues. Binary secondary variables were assessed similarly,and continuous secondary outcomes were tested using a ttest.

To compare the SpO2 values in the 2 groups, data fromcomplete days when on supplemental oxygen were adjustedback to actual values without the offset, and the median valuewas determined for each infant. These values were plotted byintention-to-treat group and the distributions compared byusing Wilcoxon rank-sum test.

At the time our study was planned, 200 VP infants wereadmitted for NICU care annually in New Zealand with80% survival to discharge.16 With a sample size of 340 in-fants, BOOST-New Zealand had 80% power (2P = .05) todetect an absolute 15% increase or decrease in the compositeoutcome of death or major disability at 2 years (from 37% to

Randomized Controlled Trial of Oxygen Saturation Targets in Ver

52% or 37% to 23%). This would mean 1 less infant who diedor was disabled for every 6 infants managed in the optimalrange and would have similar power to detect a reductionin severe ROP from 10% to 2% and in CLD from 40% to26%. The study was also funded on the basis that it wouldcontribute to the planned IPD meta-analysis of 5000 infants,where the primary outcome was estimated as 42% and whichwould be powered to detect a 4% absolute difference in theprimary outcome.10

Results

A total of 340 infants were enrolled from September 2006 tothe end of December 2009. Enrollment was 85% of thoseeligible and approached for consent (Figure 1). Thebaseline demographic and clinical characteristics by SpO2target group were similar and are shown in Table I.The primary outcome was available for 167 of 170 (98.2%)

assigned to the lower target and 168 of 170 (98.8%) assignedto the higher target. Of those alive at 2 years, neurodevelop-mental outcome was assessed on the basis of the BIII scales in118 of 146 (81%) and 120 of 143 (84%) and on the basis ofthe BII scales in 13 (9%) and 12 (8%); 144 and 142, respec-tively, had pediatric examination forms completed. The useof <10 words was used to determine if a child had majordisability in a total of 33 children, with only 1 in the lowertarget group and 2 in the higher target groupmeeting this cri-terion.Table II shows the primary outcome and individual

criteria for major disability by intention-to-treat groups.Targeting a lower compared with a higher SpO2 target rangehad no significant effect on the rate of death or majordisability at 2 years’ corrected age, taking account of thedependency of multiples. Seven children, 4 in the lower and3 in the higher target group, died after discharge from theprimary hospital admission and before 24 months of age.There were 16 children with CP but only 12 with GrossMotor Function Classification System level $2 and 1 childdied before 24 months of age. Only 1 child had bilateralblindness. We also undertook a secondary sensitivityanalysis in which the primary composite outcome excludesthe 33 children where the language test was used as asurrogate for the BIII result. This made no difference to theresults (lower target 64/151 [42.4%] vs higher target 74/151[49.0%]; RR 1.15, 95% CI 0.90-1.46; P = .27).There were no significant differences between the groups

on any binary or continuous secondary outcomes(Tables III and IV; available at www.jpeds.com). The causeof death is shown in Table V (available at www.jpeds.com),with time of death spread throughout the neonatal courseand not significantly different between treatment groups(log rank P = .74).The actual median recorded SpO2 values for complete days

in supplementary oxygen are shown in Figure 2. Althoughthe distributions were significantly different for the low andhigh target groups (P < .001), there was considerable overlap.

y Preterm Infants: Two Year Outcomes 3

Figure 1. Eligibility, enrollment, and study completion. Fiveinfants who were still alive had discontinued study oximetermonitoringbefore 36weeks’ gestational age: 3 in the lower and2 in the higher target group, and all in error. *Not meeting in-clusion criteria, n = 49 (39 imminent death, 2 congenital ab-normalities, 8 unable to follow-up [no English spoken orreturning overseas]). **Other reasons, n = 139 (116missed [notapproachable for consent], 16 equipment constraints, 4enrolled in other trials, 3 mothers unwell in intensive care unit]).

Table I. Baseline characteristics by SpO2

SpO2 targets

Lower85%-89%(n = 170)

Higher91%-95%(n = 170)

All, No. 170 (100%) 170 (100%)Male, No. 90 (52.9%) 90 (52.9%)Birth weight, mean g (SD) 873 (202) 884 (186)Gestational age, mean wk (SD) 26.1 (1.23) 26.1 (1.19)Multiple birth, No. 46 (27.1%) 46 (27.1%)Outborn, No. 11 (6.5%) 13 (7.6%)No steroids, No. 20 (11.8%) 18 (10.6%)Incomplete steroids, No. 48 (28.2%) 41 (24.1%)Complete steroids, No. 94 (55.3%) 103 (60.6%)Steroids >7 days earlier, No. 8 (4.7%) 8 (4.7%)Vaginal birth, No. 74 (43.5%) 75 (44.1%)Instrumental birth, No. 1 (0.6%) 4 (2.4%)Cesarean, in labor, No. 46 (27.1%) 42 (24.7%)Cesarean, no labor, No. 49 (28.8%) 49 (28.8%)Admission temperature,

mean �C (SD)36.4 (1.01) 36.4 (0.89)

Apgar score at 5 min,median (IQR)

8 (6-9) 8 (7-9)

First base excess,median mmol/L (IQR)

�4.2 (�6.8 to �2.2) �4 (�6.3 to �1.9)

Nasal CPAP,* No. 59 (34.7%) 50 (29.4%)Ventilation by endotracheal

tube,* No.111 (65.3%) 119 (70.0%)

Inspired oxygen concentration,*median % (IQR)

21 (21-25) 21 (21-24)

CPAP, Continuous positive airway pressure.*At randomization.

THE JOURNAL OF PEDIATRICS � www.jpeds.com Vol. -, No. -

Discussion

In this multicenter randomized controlled trial comparing alower (85%-89%) and higher (91%-95%) SpO2 target rangefrom soon after birth in infants <28 weeks’ gestation, weshowed no difference between the groups in the primaryoutcome of death or major disability at 2 years of age cor-rected for prematurity. There was also no significant differ-ence in any secondary outcome, including mortality, severeROP, and CLD.

Two large trials of SpO2 targeting in infants <28 weeksgestation with a similar protocol to the present study haverecently reported outcomes at 18-22 months of age.12,17,18

Both the Surfactant, Positive Pressure, and Pulse OximetryRandomized Trial (SUPPORT) from North America18 andthe Canadian Oxygen Trial (COT) from North America, Eu-rope, Israel, and Argentina12 also reported no differences be-tween the target groups in the composite outcome of death orneurodevelopmental impairment at 18-22months of age cor-rected for prematurity. However, in SUPPORT, there was asignificantly higher mortality associated with the lower satu-ration target (22.1% vs 18.2%, RR 1.25, 95% CI 1.00-1.55; P= .046),18 which was similar to that previously reported athospital discharge.17

Following publication of SUPPORT hospital outcomes17

and review by their data monitoring committees, the similarBOOST II trials in Australia and the United Kingdom termi-nated recruitment early19 and, combined with the presentstudy, which had already completed recruitment, reportedoutcomes at hospital discharge.11 The trials in Australia and

4

the United Kingdom had upgraded the saturation monitorsoftware20 roughly midway through recruitment. All infantsin the BOOST-New Zealand trial were treated with oximetersthat did not have the upgraded software. Among the 1177 in-fants in the combined BOOST trials cared for with the up-graded monitors, the lower saturation target significantlyincreased mortality (23.1% vs 15.9%, RR 1.45, 95% CI 1.15-1.84; P = .002).11 The COT also upgraded the monitor soft-ware midway through recruitment but found no evidencethat this affected outcomes.12

Strengths of our study include the high recruitment rateand the use of saturation monitors with the same softwarethroughout. Although the follow-up rate was high, the useof <10 words was used as a surrogate for a BIII test in 33 chil-dren. The major weakness of the study is the relatively smallnumber of participants and thus reduced power to detectsmall but possibly clinically important differences inoutcome. However, the study did have sufficient power todetect a reduction in severe ROP from 10% to 2%, whichis the reduction observed by Chow et al7 when that unitadopted a lower saturation target. At the outset of the trial,we stated our intention to pool our data in a prospectiveIPD meta-analysis of 5000 randomized infants,10 and ourparticipation will help translate its results to the New Zealandpopulation and facilitate early uptake of any recommendedchange in practice.As with previous published studies, there was considerable

overlap between the actual median SpO2 values achieved by

Darlow et al

Table II. Primary outcome of death or disability at 24 months

Outcome

SpO2 targets, n/total

Unadjusted RR (95% CI)* P value*Lower

85%-89%Higher

91%-95%

Death or major disability 65/167 (38.9) 76/168 (45.2) 1.15 (0.90-1.47) .26ComponentsDeath before 24 mo 25/170 (14.7) 27/170 (15.9) 1.10 (0.68-1.78) .69CP with GMFCS $2 5/144 (3.5) 7/142 (4.9) 1.40 (0.46-4.32) .56BIII cognitive <85† 20/117 (17.1) 14/114 (12.3) 1.52 (0.68-3.38) .31BIII language <85† 33/110 (30.0) 32/106 (30.2) 1.03 (0.68-1.57) .88BII MDI <70 2/13 (15.4) 2/12 (16.7) 1.08 (0.21-5.71) .93Bilateral blindness 0/143 (0) 1/138 (0.7) 0.99 (0.98-1.01) .32Deaf 2/135 (1.4) 1/136 (0.7) 0.50 (0.05-5.36) .56

GMFCS, Gross Motor Function Classification System; MDI, Mental Developmental Index.Only unadjusted data are shown because the relatively small sample size means it is not feasible to adjust for stratification factors (hospital, sex, gestation <26 or$26 weeks, inborn, or outborn).*Allowing for dependency of multiples.†118 (lower target) and 120 (higher target) had a BIII undertaken but some had missing data.

- 2014 ORIGINAL ARTICLES

both groups, and the peak values were 91% and 93.5%, whichcompares with 91% and 94% in SUPPORT17 and 90% and94% in COT after the first 12 hours in the study.12 Beforethe present study commenced, 1 site (Christchurch) contrib-uted to a multicenter study of compliance with SpO2 targets

21

to examine closely how to improve compliance. Reinforcingcompliance with the target of 88%-92% was a focus of ourlocal research nurses, and there was daily feedback to nursescaring for study infants from the display of the stored 24-hour SpO2 histograms. Compliance was also aided by thefact that 88%-92% was the standard target for other preterminfants not in the study but requiring supplementary oxygen.Themedian inspired oxygen concentration at study entry was21% when nearly one-third of infants were being treated withnasal continuous positive airway pressure. When preterm in-fants are in low levels of supplementary oxygen, nurses willhave less ability to avoid SpO2 values above target, whilealso preventing hypoxemia, and target ranges may be easierto achieve when more intensive respiratory support isrequired.22

What SpO2 target policy should we adopt pending theIPD meta-analysis? Although the present study and theCOT study both showed no benefit or harm from either

Figure 2. Actual median SpO2 for complete days in supple-mentary oxygen by target group.

Randomized Controlled Trial of Oxygen Saturation Targets in Ver

group, most commentators23-26 have argued that thetarget range of 85%-89% should be avoided because thiswas associated with a higher mortality in SUPPORT17

and in the pooled Australian and UK BOOST trials whenthe upgraded monitor was used.11 A lower saturationtarget in the first few weeks of life may reduce the risk ofROP,27 but if it also increases mortality, this would notbe acceptable. Before the trial, a target range of 88%-92% was most commonly adopted in New Zealand, butthis range has not been compared in studies, so adoptingthe higher 90%-94% would seem reasonable until moreevidence is available.23-26 An upper target of 94% wouldallow the high alarm to be set at 95% because there is justi-fiable concern that exceeding 95% saturations will increasethe risk of ROP and measures should be in place to mini-mize this.In this randomized controlled trial in infants of <28 weeks’

gestation, targeting a saturation range of 85%-89%compared with 91%-95% when in supplementary oxygenwas associated with no difference in the primary outcomeof death or disability at 2 years of age or in any secondary out-comes. Full interpretation of the results should await theplanned IPD meta-analysis of this and at least 4 other trialswith a similar protocol. n

We are grateful to Cynthia H. Cole, MD (Boston Medical Center), forproviding impetus for the initial study concept. Technical assistancewith the oximeters and maintenance of the monitor pool was providedby Tony Blackler and Chas Perkins (Canterbury DHB, Christchurch,New Zealand). Wendy Hague, PhD, and Adrienne Kirby, MSc,NHMRC (Clinical Trials Centre, Sydney, Australia), provided assis-tance with organizational and with statistical issues, respectively. Weare also grateful to the many nurses, pediatricians, ophthalmologists,and psychologists who contributed to the day-to-day running of thetrial and outcomes assessments. Lastly, and most importantly, we arevery grateful to the parents and families who willingly allowed theirnewborn infant to participate in the trial.

Submitted for publication Sep 18, 2013; last revision received Nov 26, 2013;

accepted Jan 9, 2014.

Reprint requests: Brian A. Darlow, MD, Department of Paediatrics, University

of Otago Christchurch, PO Box 4345, Christchurch 8140, New Zealand. E-

mail: brian.darlow@otago.ac.nz

y Preterm Infants: Two Year Outcomes 5

THE JOURNAL OF PEDIATRICS � www.jpeds.com Vol. -, No. -

References

1. Silverman WA. A cautionary tale about supplemental oxygen: the alba-

tross of neonatal medicine. Pediatrics 2004;113:394-6.

2. American Academy of Pediatrics. American College of Obstetricians and

Gynecologists. Guidelines for perinatal care. 5th ed. Elk Grove Village,

IL: American Academy of Pediatrics; 2002.

3. Vijayakumar E, Ward GJ, Bullock CE, Patterson ML. Pulse oximetry in

infants of < 1500 gm birth weight on supplemental oxygen: a national

survey. J Perinatol 1997;17:341-5.

4. Askie LM, Henderson-Smart DJ, Irwig L, Simpson JM. Oxygen-satura-

tion targets and outcomes in extremely preterm infants. N Engl J Med

2003;349:959-67.

5. The STOP-ROP Multicenter Study Group. Supplemental Therapeutic

Oxygen for Prethreshold Retinopathy of Prematurity (STOP-ROP), A

randomized, controlled trial. I: Primary outcomes. Pediatrics 2000;

105:295-310.

6. Tin W, Milligan DW, Pennefather P, Hey E. Pulse oximetry, severe reti-

nopathy, and outcome at one year in babies of less than 28 weeks gesta-

tion. Arch Dis Child Fetal Neonatal Ed 2001;84:F106-10.

7. Chow LC, Wright KW, Sola A , CSMC Oxygen Administration Study

Group. Can changes in clinical practice decrease the incidence of severe

retinopathy of prematurity in very low birth weight infants? Pediatrics

2003;111:339-45.

8. Tin W, Wariyar U. Giving small babies oxygen: 50 years of uncertainty.

Semin Neonatol 2002;7:361-7.

9. Cole CH, Wright KW, Tarnow-Mordi W, Phelps DL. Resolving our un-

certainty about oxygen therapy. Pediatrics 2003;112:1415-9.

10. Askie LM, Brocklehurst P, Darlow BA, Finer N, Schmidt B, Tarnow-

Mordi W, et al. NeOProM: Neonatal Oxygen Prospective Meta-

analysis Collaboration study protocol. BMC Pediatr 2011;11:6.

11. The BOOST II UK, Australia and New Zealand Collaborative Groups

Stenson BJ, Tarnow-Mordi WO, Darlow BA, et al. Oxygen saturation

and outcomes in preterm infants. N Engl J Med 2013;368:2094-104.

12. 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.

13. Bayley N. Bayley Scales of Infant Development, 2nd edition manual. San

Antonio, TX: Psychological Corp; 1993.

14. Bayley N. Manuel for the Bayley Scales of Infant and Toddler Develop-

ment. 3rd ed. San Antonio, TX: Psychological Corp; 2006.

6

15. Johnson S, MarlowN,Wolke D, Davidson L, Marston L, O’Hare A, et al.

Validation of a parent report measure of cognitive development in very

preterm infants. Dev Med Child Neurol 2004;46:389-97.

16. Cust AE, Darlow BA, Donoghue DA. Outcomes for high-risk New Zea-

land newborn infants in 1998-99: a population based, national study.

Arch Dis Child Fetal Neonatal Ed 2003;88:F15-22.

17. SUPPORT Study Group of the Eunice Kennedy Shriver NICHD

Neonatal Research Network Carlo WA, Finer NN, Walsh MC, et al.

Target ranges of oxygen saturation in extremely preterm infants. N

Engl J Med 2010;362:1959-69.

18. Vaucher YE, Peralta-Carcelen M, Finer NN, Carlo WA, Gantz MG,

Walsh MC, et al. Neurodevelopmental outcomes in the early CPAP

and pulse oximetry trial. N Engl J Med 2012;367:2495-504.

19. Stenson B, Brocklehurst P, Tarnow-Mordi W , U.K. BOOST II trial,

Australian BOOST II trial, New Zealand BOOST II trial. Increased 36-

week survival with high oxygen saturation target in extremely preterm

infants. N Engl J Med 2011;364:1680-2.

20. Johnston ED, Boyle B, Juszczak E, King A, Brocklehurst P,

Stenson BJ. Oxygen targeting in preterm infants using the Masimo

SET Radical pulse oximeter. Arch Dis Child Fetal Neonatal Ed

2011;96:F429-33.

21. Hagadorn JI, Furey AM, Nghiem T-H, Schmid C, Phelps DL, Pillers D-

A, et al. Achieved versus intended pulse oxygen saturation in infants less

than 28 weeks gestation:the AVIOx study. Pediatrics 2006;118:1574-82.

22. Sink DK,Hope SA, Hagadorn JL. Nurse:patient ratio and achievement of

oxygen saturation goals in premature infants. Arch Dis Child Fetal

Neonatal Ed 2011;96:F93-8.

23. Bancalari E, Claure N. Oxygen targets and outcomes in premature in-

fants. JAMA 2013;309:2161-2.

24. Polin RA, Bateman D. Oxygen-saturation targets in preterm infants. N

Engl J Med 2013;368:2141-2.

25. Sweet DG, Carnielli V, Greisen G, Hallman M, Ozek E, Plavka R, et al.

European consensus guidelines on the management of neonatal respira-

tory distress syndrome in preterm infants–2013 update. Neonatology

2013;103:353-68.

26. Saugstad OD, Aune D. Optimal oxygenation of extremely low birth

weight infants: a meta-analysis and systematic review of the oxygen

saturation target studies. Neonatology 2013 Nov 15 [Epub ahead of

print].

27. Chen ML, Guo L, Smith LEH, Dammann CEL, Dammann O. High or

low oxygen saturation and severe retinopathy of prematurity: a meta-

analysis. Pediatrics 2010;125:e1483-92.

Darlow et al

Appendix

In addition to the authors, the following individuals contrib-uted to the BOOST-NZ Collaborative Group:

Data and Safety Monitoring Committee (shared with theBOOST II Australia trial)—University of Sydney, NSW,Australia. Andrew Berry, MB, BS, AM (Chair); Universityof Newcastle, NSW, Australia: Catherine A. D’Este, PhD;University of Glasgow, United Kingdom: Forrester Cock-burn, MD; Clinical Trials Centre, University of Sydney,NSW, Australia: Ian Marschner, PhD.

Christchurch Women’s Hospital, Christchurch, New Zea-land: Jennifer E. Gardner, PGDipHealSci; Middlemore Hos-pital, South Auckland, New Zealand: Tracey Bushell, BA;Auckland City Hospital Auckland, New Zealand: SabineHuth, BA; Wellington Women’s Hospital, Wellington, NewZealand: Frances E. Trist, BEd, BA; Dunedin Hospital, Dun-edin, New Zealand: Frances T. McCaffrey, BN.

Table III. Binary secondary clinical outcomes

SpO2, n/total (%) RR

Lower Higher Effect (95% CI) P

ROP ($stage 3) 23/160 (14.4) 20/156 (12.8) 0.89 (0.51-1.56) .69CLD (oxygen dependency at 36 wk) 58/153 (37.9) 70/148 (47.3) 1.25 (0.96-1.63) .10Patent ductus arteriosus (required medical or surgical treatment) 104/170 (61.2) 90/170 (52.9) 0.87 (0.72-1.04) .12Necrotizing enterocolitis requiring surgery or leading to death 15/170 (8.8) 12/170 (7.1) 0.80 (0.39-1.66) .55IVH ($grade 3) 17/170 (10.0) 14/170 (8.2) 0.82 (0.42-1.62) .57Proved infection 77/170 (45.3) 70/170 (41.2) 0.91 (0.71-1.16) .44Death from pulmonary causes before hospital discharge 3/170 (1.8) 1/170 (0.6) 0.33 (0.04-3.17) .30Readmission to hospital 106/149 (71.1) 96/146 (65.8) 0.92 (0.79-1.08) .32Any respiratory illnesses following discharge 68/149 (45.6) 64/146 (43.8) 0.96 (0.75-1.24) .76

IVH, Intraventricular hemorrhage.

- 2014 ORIGINAL ARTICLES

Randomized Controlled Trial of Oxygen Saturation Targets in Very Preterm Infants: Two Year Outcomes 6.e1

Table IV. Continuous secondary clinical outcomes

SpO2

Lower n Higher n P*

Days of ETT in hospital, median (IQR)† 7 (2-18) 129 5.5 (2-22) 124 .65Days of CPAP in hospital, median (IQR)† 48 (33-61) 149 49 (40-60) 146 .07Days on oxygen in hospital, median (IQR)† 41 (17-86) 139 58 (22-91) 138 .13Age at death, d, predischarge, mean (SD) 45.8 (50.6) 21 35.5 (52.1) 24 .51Length at discharge, cm, mean (SD)† 47.6 (4.04) 135 48.2 (3.28) 126 .24Weight at discharge, g, mean (SD)† 3055 (664) 149 3107 (754) 146 .53Head circumference atdischarge, cm, mean (SD)†

33.9 (1.62) 144 33.7 (1.83) 141 .45

Days on oxygen at home, median (IQR) 125 (78-230) 29 194 (75-288) 39 .38Number of times admitted, median (IQR) 2 (1-4) 139 1 (0-3) 137 .11Height at 2 y, cm, mean (SD) 50.7 (38.7) 132 48.8 (39.4) 134 .69Weight at 2 y, cm, mean (SD) 11 (4.16) 139 10.2 (4.73) 141 .13Head circumference at 2 y, cm, mean (SD) 31.6 (21.2) 120 31.3 (21) 123 .91Age at death, d, all before 2 y 94.3 (146) 25 69.7 (119) 27 .51

ETT, endotracheal intubation.*P value calculated using t test.†Data from infants surviving to discharge only.

Table V. Primary cause of death by SpO2 target

Primary cause of death

SpO2

Lower Higher

Congenital abnormality 1* 1Respiratory distress syndrome 2 2CLD 4* 4Pneumonia 2† 1*IVH/brain injury 2 3*Septicemia/meningitis 4 6Necrotizing enterocolitis 3 6Pulmonary hemorrhage 3 1Sudden infant death syndrome 0 1*Other 4 2Total 25 27

*One infant died postdischarge.†Two infants died postdischarge.

THE JOURNAL OF PEDIATRICS � www.jpeds.com Vol. -, No. -

6.e2 Darlow et al