[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 34.226.234.20. Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Figure 1.
Study Population
Study Population

CNN indicates Canadian Neonatal Network and Oxygen/RS indicates receiving supplemental oxygen and/or respiratory positive-pressure support.

Figure 2.
Associations of 6 Traditional Bronchopulmonary Dysplasia (BPD) Definitions With Adverse Outcomes at 18 to 21 Months of Age
Associations of 6 Traditional Bronchopulmonary Dysplasia (BPD) Definitions With Adverse Outcomes at 18 to 21 Months of Age

The second and third columns show the number of infants with adverse outcomes/the number of infants assessed at 18 to 21 months, with percentages in brackets for infants with or without BPD defined in the first column. The forest plots show the adjusted odds ratios (AORs; filled squares) and 95% CIs (lines). AUC indicates area under the receiver operating characteristic curve; Oxygen/RS, receiving supplemental oxygen and/or any respiratory positive-pressure support; PMA, postmenstrual age.

aAdjusted for gestational age, sex, small for gestational age, Score for Neonatal Acute Physiology II score >20, maternal education, severe intraventricular hemorrhage and/or periventricular leukomalacia, necrotizing enterocolitis, and late-onset sepsis.

bComposite outcome was defined as serious respiratory morbidity and/or neurosensory impairment at 18 to 21 months’ corrected age or death after neonatal intensive care unit discharge before 21 months’ corrected age.

Figure 3.
Association of Oxygen Use or Respiratory Support at 34 to 40 Weeks’ Postmenstrual Age With Adverse Outcomes at 18 to 21 Months of Age
Association of Oxygen Use or Respiratory Support at 34 to 40 Weeks’ Postmenstrual Age With Adverse Outcomes at 18 to 21 Months of Age

The second and third columns show the number of infants with adverse outcomes/the number of infants assessed at 18-21 months with percentages in brackets for infants with or without bronchopulmonary dysplasia (BPD) defined in the first column. The forest plots show the adjusted odds ratios (AORs; filled squares) and 95% CIs (lines). AUC indicates area under the receiver operating characteristic curve; Oxygen/RS, receiving supplemental oxygen and/or any respiratory positive-pressure support; PMA, postmenstrual age.

aAdjusted for gestational age, sex, small for gestational age, Score for Neonatal Acute Physiology II score >20, maternal education, severe intraventricular hemorrhage and/or periventricular leukomalacia, necrotizing enterocolitis, and late-onset sepsis.

bComposite outcome was defined as serious respiratory morbidity and/or neurosensory impairment at 18 to 21 months’ corrected age, or death after neonatal intensive care unit discharge before 21 months’ corrected age.

Table 1.  
Adverse Long-Term Outcomes Among the Study Population at 18 to 21 Months of Corrected Age
Adverse Long-Term Outcomes Among the Study Population at 18 to 21 Months of Corrected Age
Table 2.  
Maternal and Infant Characteristics
Maternal and Infant Characteristics
1.
Doyle  LW, Anderson  PJ.  Long-term outcomes of bronchopulmonary dysplasia.  Semin Fetal Neonatal Med. 2009;14(6):391-395.PubMedGoogle ScholarCrossref
2.
Schmidt  B, Asztalos  EV, Roberts  RS, Robertson  CM, Sauve  RS, Whitfield  MF; Trial of Indomethacin Prophylaxis in Preterms (TIPP) Investigators.  Impact of bronchopulmonary dysplasia, brain injury, and severe retinopathy on the outcome of extremely low-birth-weight infants at 18 months: results from the trial of indomethacin prophylaxis in preterms.  JAMA. 2003;289(9):1124-1129.PubMedGoogle ScholarCrossref
3.
Vohr  BR, Wright  LL, Poole  WK, McDonald  SA.  Neurodevelopmental outcomes of extremely low birth weight infants <32 weeks’ gestation between 1993 and 1998.  Pediatrics. 2005;116(3):635-643.PubMedGoogle ScholarCrossref
4.
Stoll  BJ, Hansen  NI, Bell  EF,  et al; Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network.  Trends in care practices, morbidity, and mortality of extremely preterm neonates, 1993-2012.  JAMA. 2015;314(10):1039-1051.PubMedGoogle ScholarCrossref
5.
Northway  WH  Jr, Rosan  RC, Porter  DY.  Pulmonary disease following respirator therapy of hyaline-membrane disease. Bronchopulmonary dysplasia.  N Engl J Med. 1967;276(7):357-368.PubMedGoogle ScholarCrossref
6.
Tooley  WH.  Epidemiology of bronchopulmonary dysplasia.  J Pediatr. 1979;95(5 Pt 2):851-858.PubMedGoogle ScholarCrossref
7.
Shennan  AT, Dunn  MS, Ohlsson  A, Lennox  K, Hoskins  EM.  Abnormal pulmonary outcomes in premature infants: prediction from oxygen requirement in the neonatal period.  Pediatrics. 1988;82(4):527-532.PubMedGoogle Scholar
8.
Beam  KS, Aliaga  S, Ahlfeld  SK, Cohen-Wolkowiez  M, Smith  PB, Laughon  MM.  A systematic review of randomized controlled trials for the prevention of bronchopulmonary dysplasia in infants.  J Perinatol. 2014;34(9):705-710.PubMedGoogle ScholarCrossref
9.
Jobe  AH, Bancalari  E.  Bronchopulmonary dysplasia.  Am J Respir Crit Care Med. 2001;163(7):1723-1729.PubMedGoogle ScholarCrossref
10.
Walsh  MC, Wilson-Costello  D, Zadell  A, Newman  N, Fanaroff  A.  Safety, reliability, and validity of a physiologic definition of bronchopulmonary dysplasia.  J Perinatol. 2003;23(6):451-456.PubMedGoogle ScholarCrossref
11.
Walsh  MC, Yao  Q, Gettner  P,  et al; National Institute of Child Health and Human Development Neonatal Research Network.  Impact of a physiologic definition on bronchopulmonary dysplasia rates.  Pediatrics. 2004;114(5):1305-1311.PubMedGoogle ScholarCrossref
12.
Poindexter  BB, Feng  R, Schmidt  B,  et al; Prematurity and Respiratory Outcomes Program.  Comparisons and limitations of current definitions of bronchopulmonary dysplasia for the prematurity and respiratory outcomes program.  Ann Am Thorac Soc. 2015;12(12):1822-1830.PubMedGoogle ScholarCrossref
13.
Ehrenkranz  RA, Walsh  MC, Vohr  BR,  et al; National Institutes of Child Health and Human Development Neonatal Research Network.  Validation of the National Institutes of Health consensus definition of bronchopulmonary dysplasia.  Pediatrics. 2005;116(6):1353-1360.PubMedGoogle ScholarCrossref
14.
Canadian Neonatal Network. The Canadian Neonatal Network Abstractor's Manual, version 1.3.4. http://www.canadianneonatalnetwork.org/Portal/LinkClick.aspx?fileticket=U4anCYsSN20%3D&tabid=69. Accessed December 15, 2016.
15.
Synnes  A, Luu  TM, Moddemann  D,  et al; Canadian Neonatal Network and the Canadian Neonatal Follow-Up Network.  Determinants of developmental outcomes in a very preterm Canadian cohort.  Arch Dis Child Fetal Neonatal Ed. 2016;fetalneonatal-2016-311228.PubMedGoogle Scholar
16.
Canadian Neonatal Follow-Up Network. 18-Month Corrected Age Assessment Manual, Version 5. http://www.cnfun.ca/LinkClick.aspx?fileticket=d4p7mZoXWDU%3D&tabid=68. Accessed December 15, 2016.
17.
Morris  C, Bartlett  D.  Gross motor function classification system: impact and utility.  Dev Med Child Neurol. 2004;46(1):60-65.PubMedGoogle ScholarCrossref
18.
Bayley  N.  The Bayley Scales of Infant and Toddler Development. 3rd ed. San Antonio, TX: The Psychological Corporation; 2006.
19.
Richardson  DK, Corcoran  JD, Escobar  GJ, Lee  SK.  SNAP-II and SNAPPE-II: simplified newborn illness severity and mortality risk scores.  J Pediatr. 2001;138(1):92-100.PubMedGoogle ScholarCrossref
20.
Papile  LA, Burstein  J, Burstein  R, Koffler  H.  Incidence and evolution of subependymal and intraventricular hemorrhage: a study of infants with birth weights less than 1,500 gm.  J Pediatr. 1978;92(4):529-534.PubMedGoogle ScholarCrossref
21.
Bell  MJ, Ternberg  JL, Feigin  RD,  et al.  Neonatal necrotizing enterocolitis. Therapeutic decisions based upon clinical staging.  Ann Surg. 1978;187(1):1-7.PubMedGoogle ScholarCrossref
22.
Steyerberg  EW. Evaluation of performance. In: Steyerberg  EW, ed.  Clinical Prediction Models: a Practical Approach to Development, Validation, and Updating. New York: Springer; 2009.
23.
Zou  KH, O’Malley  AJ, Mauri  L.  Receiver-operating characteristic analysis for evaluating diagnostic tests and predictive models.  Circulation. 2007;115(5):654-657.PubMedGoogle ScholarCrossref
24.
Davis  PG, Thorpe  K, Roberts  R, Schmidt  B, Doyle  LW, Kirpalani  H; Trial Indomethacin Prophylaxis in Preterms Investigators.  Evaluating “old” definitions for the “new” bronchopulmonary dysplasia.  J Pediatr. 2002;140(5):555-560.PubMedGoogle ScholarCrossref
25.
Mercier  CE, Dunn  MS, Ferrelli  KR, Howard  DB, Soll  RF; Vermont Oxford Network ELBW Infant Follow-Up Study Group.  Neurodevelopmental outcome of extremely low birth weight infants from the Vermont Oxford network: 1998-2003.  Neonatology. 2010;97(4):329-338.PubMedGoogle ScholarCrossref
26.
Maitre  NL, Ballard  RA, Ellenberg  JH,  et al; Prematurity and Respiratory Outcomes Program.  Respiratory consequences of prematurity: evolution of a diagnosis and development of a comprehensive approach.  J Perinatol. 2015;35(5):313-321.PubMedGoogle ScholarCrossref
Original Investigation
March 2017

Revisiting the Definition of Bronchopulmonary Dysplasia: Effect of Changing Panoply of Respiratory Support for Preterm Neonates

Author Affiliations
  • 1Department of Pediatrics, University of Toronto, Toronto, Ontario, Canada
  • 2Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Ontario, Canada
  • 3Maternal-Infant Care Research Centre, Department of Paediatrics, Mount Sinai Hospital, Toronto, Ontario, Canada
  • 4Department of Paediatrics, University of Western Ontario, London, Ontario, Canada
  • 5Department of Paediatrics, University of Saskatchewan, Saskatoon, Canada
JAMA Pediatr. 2017;171(3):271-279. doi:10.1001/jamapediatrics.2016.4141
Key Points

Question  What is the optimal definition of bronchopulmonary dysplasia that best predicts respiratory and neurodevelopmental outcomes in preterm infants and the postmentstrual age associated with the best predictive characteristics for serious adverse respiratory or neurosensory outcomes?

Findings  In this cohort study, receipt of supplemental oxygen and/or positive-pressure respiratory support at 40 weeks’ postmentstrual age was identified as the best predictor for serious respiratory morbidity and displayed a good ability to predict neurosensory morbidity at 18 to 21 months.

Meaning  Defining BPD by the use of oxygen alone is inadequate because oxygen and/or positive-pressure respiratory support is a better indicator of chronic respiratory insufficiency.

Abstract

Importance  Several definitions of bronchopulmonary dysplasia are clinically used; however, their validity remains uncertain considering ongoing changes in the panoply of respiratory support treatment strategies used within neonatal units.

Objective  To identify the optimal definition of bronchopulmonary dysplasia that best predicts respiratory and neurodevelopmental outcomes in preterm infants.

Design, Setting, and Participants  Retrospective cohort study at tertiary neonatal intensive care units. Preterm infants born at less than 29 weeks’ gestation between 2010 and 2011 who were admitted to neonatal intensive care units participating in the Canadian Neonatal Network and completed follow-up assessments in a Canadian Neonatal Follow-Up Network clinic at 18 to 21 months.

Exposures  Various traditional bronchopulmonary dysplasia criteria based on respiratory status at different postmenstrual ages.

Main Outcomes and Measures  Serious respiratory morbidity, neurosensory impairment at 18 to 21 months of age, and a composite outcome of respiratory or neurosensory morbidity or death after discharge. Adjusted odds ratios (AORs) and 95% CIs were calculated.

Results  Of 1914 eligible survivors, 1503 were assessed (mean gestational age was 26.3 weeks; 68% were white, 9% were black, and 23% were other race/ethnicity), 88 had serious respiratory morbidity, 257 infants had neurosensory impairment, and 12 infants died after discharge. Definitions using oxygen requirement alone as the criterion at various postmenstrual ages were less predictive compared with those using the criterion of oxygen/respiratory support (RS) (receiving supplemental oxygen and/or positive-pressure RS); among those, oxygen/RS at 36 weeks had the highest AOR and area under the curve (AUC) for all outcomes. Further analyses of oxygen/RS at each week between 34 and 44 weeks’ postmenstrual age indicated that the predictive ability for serious respiratory morbidity increased from 34 weeks (AOR, 1.8; 95% CI, 0.9-3.4, AUC, 0.721) to 40 weeks (AOR, 6.1; 95% CI, 3.4-11.0; AUC, 0.799). For serious neurosensory impairment, the AOR and AUC at 40 weeks’ PMA (AOR, 1.5, 95% CI, 1.0-2.1; AUC, 0.740) were only marginally below their peak values at 37 weeks’ PMA (AOR, 1.8; 95% CI, 1.3-2.6; AUC, 0.743).

Conclusions and Relevance  Defining bronchopulmonary dysplasia by the use of oxygen alone is inadequate because oxygen/RS is a better indicator of chronic respiratory insufficiency. In particular, oxygen/RS at 40 weeks’ PMA was identified as the best predictor for serious respiratory morbidity, while it also displayed a good ability to predict neurosensory morbidity at 18 to 21 months.

Introduction

Bronchopulmonary dysplasia (BPD) is an important morbidity in preterm infants that has short- and long-term serious adverse consequences for infants, their families, and the health care system.1-3 Approximately 45% of preterm infants born at less than 29 weeks’ gestation are diagnosed as having BPD.4 While accurate and timely diagnosis of the condition is important to identify high-risk infants in need of surveillance or special support, prevalence of BPD has also been proven to be a valuable short-term indicator for benchmarking the quality of neonatal care provided by institutions, networks, and countries.4

The term bronchopulmonary dysplasia was coined by Northway et al5 in 1967 to describe a chronic pulmonary condition observed in infants with respiratory distress syndrome and treated with high oxygen concentration and mechanical ventilation, but it has since undergone numerous revisions to accommodate different criteria.5 In 1978, a clinical definition based on the dependency on oxygen at 30 days or 1 month of age with any radiographic abnormality was proposed6 and widely used. In 1988, Shennan et al7 reported that oxygen use at 36 weeks’ postmenstrual age (PMA) had a higher accuracy for predicting long-term respiratory problems than that at 28 days of age or other PMAs, which has become the most commonly used measure to define BPD.8 In 2001, a workshop held by the National Institutes of Health proposed to define BPD as oxygen use for 28 days and categorized BPD into 3 severity levels (mild, moderate, and severe) based on oxygen use and/or respiratory support at 36 weeks’ PMA (or 56 days of age for infants at ≥32 weeks’ gestational age).9 An oxygen reduction test was also recommended to confirm physiological oxygen requirement.10,11

The validity of these BPD definitions based on respiratory status at 36 weeks’ PMA is uncertain because there have been substantial changes in the perinatal and neonatal management of preterm infants since the single-center study by Shennan et al,7 particularly in respiratory support modalities available for preterm neonates.12 This concern is particularly relevant because many previous studies used the Shennan et al definition8 and therefore did not capture infants who might be on noninvasive respiratory support but in room air, a situation that is not infrequent in modern-day neonatal intensive care units (NICUs). Although the National Institutes of Health definition13 incorporated respiratory support in room air to define severe BPD while also requiring oxygen use for 28 days after birth for BPD diagnosis, the validity of doing so has yet to be determined.

Here, we compared various traditional BPD criteria used in the literature to identify the best definition predictive of serious long-term adverse respiratory or neurosensory outcomes in preterm infants less than 29 weeks’ gestation in a large, population-based cohort. Using this definition, we conducted sensitivity analyses for the entire range of infants between 34 and 44 weeks of PMA, with the aim to identify the PMA associated with the best predictive characteristics for serious adverse respiratory or neurosensory outcomes.

Methods
Study Population and Data Collection

This retrospective cohort study included preterm infants born at less than 29 weeks’ gestation from January 2010 to September 2011 who were admitted to tertiary NICUs participating in the Canadian Neonatal Network. Those surviving until discharge or transferred to step-down units and followed up in affiliated clinics participating in the Canadian Neonatal Follow-Up Network were included. We excluded infants who had major congenital anomalies or were discharged or transferred to step-down units before 34 weeks’ PMA. Maternal and infant data were collected by trained data abstractors based on the Canadian Neonatal Network abstractor’s manual.14 All 26 Canadian Neonatal Follow-Up Network clinics provided follow-up data for this study, making up approximately 80% of neonates younger than 29 weeks’ gestation in 28 of 30 tertiary NICUs in Canada. Experienced and trained health care professionals obtained medical histories by interviews and conducted physical and neurological examinations of children at 18 to 21 months of corrected age at Canadian Neonatal Follow-Up Network clinics as previously described.15

This study was approved by the research ethics board of Mount Sinai Hospital, Toronto. Neonatal data were collected retrospectively after approval from ethics or quality improvement committee at each unit without direct consent to ensure data collection of all neonates. For follow-up data, individual written consent was obtained from parents when they presented for assessment in clinics.

Definitions Compared

Respiratory support (RS) was defined as the use of any mechanical ventilation or noninvasive respiratory support that provides positive end-expiratory pressure including continuous positive airway pressure, biphasic continuous positive airway pressure, high flow of air or oxygen (>1.5 L/min), noninvasive intermittent positive pressure ventilation, and noninvasive high-frequency oscillation. Six different criteria for classifying neonates with BPD were identified: (1) oxygen at 28 days of age, (2) receiving supplemental oxygen and/or respiratory positive-pressure support (oxygen/RS) at 28 days of age, (3) oxygen at 36 weeks’ PMA, (4) oxygen/RS at 36 weeks’ PMA, (5) oxygen at 28 days and oxygen/RS at 36 weeks’ PMA, and (6) oxygen/RS at 28 days of age and at 36 weeks’ PMA. Once the best measure among these 6 definitions was identified, sensitivity analyses were conducted to identify the PMA at which the particular measure would perform best for the prediction of serious adverse respiratory and neurosensory outcomes.

Outcomes

Three long-term adverse outcomes were assessed including serious respiratory morbidity, serious neurosensory impairment, and a composite outcome of serious respiratory morbidity and/or neurosensory impairment at 18 to 21 months corrected age or death after NICU discharge and prior to 21 months corrected age. Serious respiratory morbidity was defined as either (1) 3 or more rehospitalizations after NICU discharge owing to respiratory problems (infectious or noninfectious); (2) having a tracheostomy; (3) using respiratory monitoring or support devices at home such as an apnea monitor or pulse oximeter; and (4) being on home oxygen or continuous positive airway pressure at the time of assessment between 18 and 21 months corrected age.16 At least 3 rehospitalizations was chosen because the 95th percentile of the number of readmissions owing to respiratory problems in this cohort was 2 (which could be considered within normal limits). Neurosensory impairment was defined as having any of the following: (1) moderate to severe cerebral palsy (Gross Motor Function Classification System ≥3)17; (2) severe developmental delay (Bayley Scales of Infant and Toddler Development Third Edition [Bayley III] composite score <70 in either cognitive, language, or motor domains)18; (3) hearing aid or cochlear implant use; and (4) bilateral severe visual impairment.16

Covariates

Covariates known to affect outcomes included maternal age, hypertension, antenatal steroids, delivery mode, multiple births, sex, gestational age at birth, birth weight, small for gestational age (birth weight <10th percentile), 5-minute Apgar score less than 4, Score for Neonatal Acute Physiology II score19 greater than 20, severe cerebral injuries defined as having severe intraventricular hemorrhage (grade 3 or 4, based on the Papile et al grading),20 and/or periventricular leukomalacia, necrotizing enterocolitis (stage 2 or higher based on Bell criteria21), patent ductus arteriosus requiring surgical or medical treatment, late-onset sepsis (isolation of a pathogenic organism in blood or cerebrospinal fluid in a symptomatic neonate after 2 days of age), maternal race/ethnicity (white, black, and others), and maternal education (completing college or higher).

Statistical Analyses

Covariates were compared between infants with and without serious respiratory morbidity or neurosensory impairment using the χ2 test or t test as appropriate. Initially, multiple logistic regression models were developed to assess the association between the 6 traditional BPD definitions with long-term outcomes adjusting for potential confounders. The areas under the receiver operating characteristic curves (AUCs) of the regression models were calculated for each BPD definition.22 Area under the curve indicates the predictive ability for discrimination, a quality that expresses how well prediction models discriminate between patients with and without the outcomes of interest.22 The AUC was used to select the best regression model to predict outcomes.23 Furthermore, modified versions of the best traditional BPD definition (defined at different PMAs between 34-44 weeks) were evaluated by logistic regression analyses to decide the best timing of the assessment of respiratory status based on the associations with outcomes and the AUC. Sensitivity, specificity, and positive or negative predictive values of BPD definitions for predicting long-term outcomes were also calculated.

Infants who were discharged or transferred to step-down units without oxygen and without respiratory support between 34 and 44 weeks’ PMA were assumed to not require oxygen or respiratory support post-discharge or post-transfer. Infants who were discharged home using either oxygen or respiratory support were considered to use oxygen/RS the week following discharge. If infants were transferred to step-down units using either oxygen or respiratory support, their oxygen use and RS were considered unknown after the week of transfer. To assess the effect of these inclusion/exclusion criteria of infants discharged or transferred between 34 and 40 weeks’ PMA, 2 sensitivity analyses were conducted assuming that (1) infants discharged home on oxygen/RS were excluded from the analyses after the week of discharge or (2) infants who transferred to step-down units on oxygen/RS were considered using oxygen/RS after the week of transfer. We limited our assessment up until 44 weeks’ PMA because neonates older than 44 weeks are a very selective group of neonates with additional complications necessitating their prolonged hospital stay. All statistical analyses were conducted using SAS, version 9.3 (SAS Institute Inc).

Results

Among 2760 infants at younger than 29 weeks’ gestational age admitted to Canadian Neonatal Network NICUs during the study period, 1914 infants were eligible for this study (Figure 1). After excluding 411 infants (21%) without follow-up data, 1503 infants were included. The proportion of infants using oxygen/RS decreased from 63% to 9% from 34 to 44 weeks’ PMA (eFigure 1 in the Supplement). A total of 321 infants (21%) had a composite outcome of respiratory morbidity and/or neurosensory impairment or death after discharge at 18 to 21 months’ corrected age, among which 12 infants (1%) died after initial discharge, 88 infants (6%) had serious respiratory morbidity, and 257 infants (17%) had serious neurosensory impairment (Table 1). Table 2 shows maternal and infant characteristics of neonates with or without serious respiratory morbidity or neurosensory impairment. Infants of lower birth weight and gestational age were more likely to display serious respiratory morbidity (mean birth weight, 777 g vs 933 g; mean gestational age, 25.3 weeks vs 26.3 weeks; P < .001) and neurosensory impairment (mean birth weight, 861 g vs 944 g; mean gestational age, 25.8 weeks vs 26.4 weeks; P < .001).

Infants with oxygen/RS at 28 days or 34 to 44 weeks’ PMA had a higher likelihood of all long-term adverse outcomes (Figures 2 and 3). Four of 6 traditional BPD definitions were significantly associated with all the 3 long-term adverse outcomes after adjusting for potential confounders (Figure 2). Among them, the definition of oxygen/RS at 36 weeks’ PMA displayed the highest predictive values for serious respiratory morbidity and neurosensory impairment, suggesting oxygen requirement alone was not adequate criteria. Based on this finding, subsequent analyses focused on oxygen/RS at various PMAs between 34 and 44 weeks.

As the PMA for the assessment of oxygen/RS increased, the specificity and positive predictive values increased while the sensitivity decreased. Changes in negative predictive values were minimal (eFigure 2 in the Supplement). In most cases, oxygen/RS at 34 to 44 weeks’ PMA was significantly associated with adverse outcomes (Figure 3). For serious respiratory morbidity, the adjusted odds ratio (AOR) estimates increased from 34 to 40 weeks’ PMA and decreased after 40 weeks’ PMA (Figure 3). The AUC also increased from 34 to 40 weeks’ PMA, from 0.721 to 0.799, and peaked at 40 weeks’ PMA (Figure 3). For neurosensory impairment, the AORs and AUCs between 34 and 44 weeks’ PMA were less different and those at 40 weeks’ PMA (AOR, 1.5; 95% CI, 1.0-2.1; AUC, 0.740) were only marginally below their peak values at 37 weeks’ PMA (AOR, 1.8; 95% CI, 1.3-2.6; AUC, 0.743). Marginal variations were also observed for the composite outcome (AOR estimates, 1.6-2.3 and AUC, 0.724-0.746). Two sensitivity analyses uncovered similar results: the AOR and AUC peaked at 40 weeks’ PMA for serious respiratory morbidity and displayed minimal increases for serious neurosensory impairment or the composite outcome between 34 and 44 weeks’ PMA (eTables 1 and 2 in the Supplement).

Discussion

In this large retrospective study with comprehensive evaluation of respiratory and neurosensory impairment, we determined that oxygen/RS at 36 weeks’ PMA was the best among all contemporary definitions of BPD to predict serious severe respiratory morbidity at 18 to 21 months corrected age with the highest AUC. Moreover, oxygen/RS at 40 weeks’ PMA was most strongly associated with serious respiratory outcome among those at each of 34 to 44 weeks’ PMA. For serious neurosensory impairment and composite outcome, variations in prediction ability and strength of association with oxygen/RS at 34 to 44 weeks’ PMA were marginal, with oxygen/RS at 37 to 40 weeks’ PMA resulting in similar estimates.

Several studies have evaluated accuracy or predictive ability of various diagnostic criteria of BPD for long-term adverse respiratory outcomes. The original Shennan et al study7 evaluated the diagnostic accuracy of oxygen use at 31 to 38 weeks’ PMA for predicting adverse respiratory outcomes in the first 2 years of life. Oxygen use at 36 weeks’ PMA had the highest accuracy of 85%, with a sensitivity of 63% and specificity of 91%. The validity of this finding was limited by the fact that (1) it was a single-center study that may not have been an accurate representation of the general preterm infant population, (2) the respiratory management strategies have been changed since then, and (3) it only evaluated long-term adverse respiratory outcomes and not neurodevelopmental impairment. Davis et al24 reported that the diagnostic accuracy of oxygen use at 36 weeks’ PMA for predicting poor respiratory outcome before 18 months of corrected age (63%) was similar to those at other PMAs between 32 and 40 weeks in a cohort from a multicenter trial. In contrast, our study revealed that the AUC of oxygen/RS at 34 to 44 weeks’ PMA for predicting long-term serious respiratory morbidity was highest at 40 weeks. These differences may be owing to variations in criteria for adverse respiratory outcomes, as well as differences in indices used (accuracy vs AUC). Unlike our study, these previous studies included oxygen use at 40 weeks’ PMA or at discharge as a component of long-term adverse respiratory outcomes that did not necessarily reflect functionally important problems at follow-up. Furthermore, they also included 1 or more, or more than 1, admissions as long-term adverse respiratory outcomes, in contrast to our study in which at least 3 admissions were required to be classified as serious respiratory morbidity. A validation study of the National Institutes of Health BPD definition revealed a greater association between BPD and long-term respiratory adverse outcomes as the severity of BPD increased.13 The study also reported significant associations between other various BPD definitions and long-term adverse respiratory outcomes. However, unlike our study, it did not assess the predictive ability of these BPD definitions and did not evaluate the timing of the assessment of respiratory status other than at 28 days of age and 36 weeks’ PMA.13

Our study confirmed the significant association between BPD and neurosensory impairment that was previously reported.1-3,25 Thus far, few groups have assessed the association or predictive ability of oxygen use or RS at various PMAs for neurosensory impairment. Davis et al24 reported that the accuracy of oxygen use at 32 to 40 weeks’ PMA for predicting neurosensory impairment at 18 months of corrected age increased as the PMA for assessment increased (accuracy from 54% for the assessment at 32 weeks’ PMA to 68% for that at 40 weeks’ PMA). However, unlike our study, they did not adjust for important confounders.24 Given that other major morbidities, such as severe cerebral injuries, necrotizing enterocolitis, and sepsis, are associated with both BPD and neurosensory impairment,13,25 adjustment for these potential confounders is critical for evaluating the independent association and predictive ability of BPD for neurosensory impairment.

This study is marked by several strengths. First, the study cohort had a large sample size and contained a nearly population-based sample from Canada. Second, this study assessed not only the commonly used traditional BPD definitions but also other promising alternatives including oxygen/RS at each PMA between 34 and 44 weeks. Third, the adjustment for potential confounders, including major neonatal morbidities, enabled us to evaluate independent predictive abilities of oxygen/RS at various PMAs using AUCs of logistic regression models. Finally, the sensitivity analyses confirmed the robustness of the findings, especially for serious respiratory morbidity.

Limitations

There were several limitations in our study. First, the potential for recall bias may exist in long-term adverse respiratory morbidity findings because information was collected from parents by interviewers in follow-up visits at 18 to 21 months of age; however, recall items only included the number of hospitalizations, which is very unlikely to be falsely reported. Second, infants who were discharged home or were transferred to step-down units before 44 weeks’ PMA using oxygen/RS did not have data available on oxygen/RS after discharge/transfer. The sensitivity analysis assessed the effect of this limitation. Third, the oxygen challenge test was not mandatory to assess oxygen requirement in this study and was unlikely to have been conducted for the assessment at PMA other than 36 weeks. Although we may have overestimated oxygen use at each PMA, its effect on the association or predictive abilities was considered small. Fourth, 21% of infants did not have follow-up outcomes; however, our previous comparisons have revealed that those who were lost to follow-up were likely to be less ill15 and thus, our estimates of association may be modest. Finally, residual bias and confounding factors cannot be ruled out based on the retrospective nature of our study.

Conclusions

From a predictive ability and strength of association perspective, oxygen/RS at 40 weeks’ PMA appears to be the optimal criterion to define or diagnose BPD associated with serious respiratory morbidity. This is particularly relevant for modern neonatal care practice, whereby wide panoply of respiratory support modalities are offered to preterm neonates. Given the persistent inconsistency of BPD definitions in the literature, this unique information should be useful to develop a future consensus on a BPD definition based on outcome data. We strongly encourage incorporation of parental perspectives in defining what matters to them most as far as respiratory and neurodevelopmental outcomes are concerned in future exercises. Future large prospective cohort studies are also needed to confirm our study findings such as the Prematurity and Respiratory Outcomes Program currently under way.26

Back to top
Article Information

Corresponding Author: Prakesh S. Shah, MD, MSc, Department of Pediatrics, 19-231F, 600 University Ave, Toronto, ON M5G 1X5, Canada, (prakeshkumar.shah@sinaihealthsystem.ca).

Accepted for Publication: October 23, 2016.

Published Online: January 23, 2017. doi:10.1001/jamapediatrics.2016.4141

Author Contributions: Dr Shah had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Isayama, S. K. Lee, D. Lee, Daspal, Dunn, Shah.

Acquisition, analysis, or interpretation of data: Isayama, S. K. Lee, Yang, Shah.

Drafting of the manuscript: Isayama, Daspal.

Critical revision of the manuscript for important intellectual content: Isayama, S. K. Lee, Yang, D. Lee, Dunn, Shah.

Statistical analysis: Isayama, Yang, Shah.

Obtained funding: S. K. Lee.

Administrative, technical, or material support: Lee, Shah.

Supervision: S. K. Lee, D. Lee, Shah.

Conflict of Interest Disclosures: None reported.

Funding/Support: Organizational support was provided by the Maternal-Infant Care Research Centre at Mount Sinai Hospital in Toronto, Ontario, Canada. Neonatal follow-up data were supported by team grant FRN87518 from the Canadian Institutes of Health Research awarded to Dr Lee, and in-kind support from Mount Sinai Hospital. Dr Isayama is supported by the Ontario Graduate Scholarships program. Dr Shah holds an Applied Research Chair in Reproductive and Child Health Services and Policy Research awarded by the Canadian Institutes of Health Research (APR-126340).

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Group Information: Investigators of the Canadian Neonatal Network and Canadian Neonatal Follow-Up Network are listed here. Canadian Neonatal Network investigators: Prakesh S Shah, MD, MSc (Director, Canadian Neonatal Network and site investigator), Mount Sinai Hospital, Toronto, Ontario; Adele Harrison, MD, MBChB, Victoria General Hospital, Victoria, British Columbia; Anne Synnes, MDCM, MHSC, and Joseph Ting, MD, British Columbia Women’s Hospital, Vancouver, British Columbia; Zenon Cieslak, MD, Royal Columbian Hospital, New Westminster, British Columbia; Rebecca Sherlock, MD, Surrey Memorial Hospital, Surrey, British Columbia; Wendy Yee, MD, Foothills Medical Centre, Calgary, Alberta; Khalid Aziz, MBBS, MA, MEd, and Jennifer Toye, MD, Royal Alexandra Hospital, Edmonton, Alberta; Carlos Fajardo, MD, Alberta Children’s Hospital, Calgary, Alberta; Zarin Kalapesi, MD, Regina General Hospital, Regina, Saskatchewan; Koravangattu Sankaran, MD, MBBS, and Sibasis Daspal, MD, Royal University Hospital, Saskatoon, Saskatchewan; Mary Seshia, MBChB, Winnipeg Health Sciences Centre, Winnipeg, Manitoba; Ruben Alvaro, MD, St. Boniface General Hospital, Winnipeg, Manitoba; Sandesh Shivananda, MBBS, MD, DM, Hamilton Health Sciences Centre, Hamilton, Ontario; Orlando Da Silva, MD, MSc, London Health Sciences Centre, London, Ontario; Chuks Nwaesei, MD, Windsor Regional Hospital, Windsor, Ontario; Kyong-Soon Lee, MD, MSc, Hospital for Sick Children, Toronto, Ontario; Michael Dunn, MD, Sunnybrook Health Sciences Centre, Toronto, Ontario; Brigitte Lemyre, MD, Children’s Hospital of Eastern Ontario and Ottawa General Hospital, Ottawa, Ontario; Kimberly Dow, MD, Kingston General Hospital, Kingston, Ontario; Ermelinda Pelausa, MD, Jewish General Hospital, Montréal, Québec; Keith Barrington, MBChB, Hôpital Sainte-Justine, Montréal, Québec; Christine Drolet, MD, and Bruno Piedboeuf, MD, Centre Hospitalier Universitaire de Québec, Sainte Foy Québec; Martine Claveau, and Daniel Faucher, MD, McGill University Health Centre, Montréal, Québec; Valerie Bertelle, MD, and Edith Masse, MD, Centre Hospitalier Universitaire de Sherbrooke, Sherbrooke, Québec; Roderick Canning, MD, Moncton Hospital, Moncton, New Brunswick; Hala Makary, MD, Dr Everett Chalmers Hospital, Fredericton, New Brunswick; Cecil Ojah, MBBS, and Luis Monterrosa, MD, Saint John Regional Hospital, Saint John, New Brunswick; Akhil Deshpandey, MBBS, MRCPI, Janeway Children’s Health and Rehabilitation Centre, St. John’s, Newfoundland; Jehier Afifi, MB BCh, MSc, IWK Health Centre, Halifax, Nova Scotia; Andrzej Kajetanowicz, MD, Cape Breton Regional Hospital, Sydney, Nova Scotia; Shoo K Lee, MBBS, PhD (Chairman, Canadian Neonatal Network), Mount Sinai Hospital, Toronto, Ontario. Canadian Neonatal Follow-Up Network investigators: Thevanisha Pillay, MD, Victoria General Hospital, Victoria, British Columbia; Anne Synnes, MDCM, MHSC (Director CNFUN), British Columbia Women’s Hospital, Vancouver, British Columbia; Reg Sauvé, MD, MPh, Leonora Hendson MBBCH, MSc, Alberta’s Children’s Hospital, Foothills Medical Centre, Calgary, Alberta; Amber Reichert, MD, Glenrose Rehabilitation Hospital, Edmonton, Alberta; Jaya Bodani, MD, Regina General Hospital, Regina, Saskatchewan; Koravangattu Sankaran, MD, Royal University Hospital, Saskatoon, Saskatchewan; Diane Moddemann, MD, Winnipeg Health Sciences Centre, St. Boniface General Hospital, Winnipeg, Manitoba; Chuks Nwaesei, MD, Windsor Regional Hospital, Windsor, Ontario; Thierry Daboval, MD, Children’s Hospital of Eastern Ontario, Ottawa, Ontario; Kimberly Dow, Kingston General Hospital, Kingston, Ontario; David Lee, MD, Children’s Hospital London Health Sciences Centre, London, Ontario; Linh Ly, MD, Hospital for Sick Children, Toronto, Ontario; Edmond Kelly, MD, Mount Sinai Hospital, Toronto, Ontario; Salhab el Helou, MD, Hamilton Health Sciences Centre, Hamilton, Ontario; Paige Church, MD, Sunnybrook Health Sciences Centre, Toronto, Ontario; Ermelinda Pelausa, MD, Jewish General Hospital, Montréal, Québec; Patricia Riley, MD, Montréal Children’s Hospital, Royal Victoria Hospital, Montréal, Québec; Francine Levebrve, Centre Hospitalier Universitaire Sainte-Justine, Montréal, Québec; Charlotte Demers, Centre Hospitalier Universitaire de Sherbrooke, Sherbrooke, Québec; Sylvie Bélanger, MD, Centre Hospitalier Universitaire de Québec, Québec City, Québec; Roderick Canning, MD, Moncton Hospital, Moncton, New Brunswick; Luis Monterrosa, MD, Saint John Regional Hospital, Saint John, New Brunswick; Hala Makary, MD, Dr Everett Chalmers Hospital, Fredericton, New Brunswick; Michael Vincer, MD, IWK Health Centre, Halifax, Nova Scotia; Phil Murphy, Charles Janeway Children’s Health and Rehabilitation Centre, St. John’s, Newfoundland.

Additional Contributions: We gratefully acknowledge all investigators and data abstractors of the Canadian Neonatal and Follow-Up Networks. We also thank Natasha Musrap, PhD, from the Maternal-Infant Care Research Centre, for editorial assistance in the preparation of the manuscript. She received salaried compensation for her work from funding support (grant FRN87518).

References
1.
Doyle  LW, Anderson  PJ.  Long-term outcomes of bronchopulmonary dysplasia.  Semin Fetal Neonatal Med. 2009;14(6):391-395.PubMedGoogle ScholarCrossref
2.
Schmidt  B, Asztalos  EV, Roberts  RS, Robertson  CM, Sauve  RS, Whitfield  MF; Trial of Indomethacin Prophylaxis in Preterms (TIPP) Investigators.  Impact of bronchopulmonary dysplasia, brain injury, and severe retinopathy on the outcome of extremely low-birth-weight infants at 18 months: results from the trial of indomethacin prophylaxis in preterms.  JAMA. 2003;289(9):1124-1129.PubMedGoogle ScholarCrossref
3.
Vohr  BR, Wright  LL, Poole  WK, McDonald  SA.  Neurodevelopmental outcomes of extremely low birth weight infants <32 weeks’ gestation between 1993 and 1998.  Pediatrics. 2005;116(3):635-643.PubMedGoogle ScholarCrossref
4.
Stoll  BJ, Hansen  NI, Bell  EF,  et al; Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network.  Trends in care practices, morbidity, and mortality of extremely preterm neonates, 1993-2012.  JAMA. 2015;314(10):1039-1051.PubMedGoogle ScholarCrossref
5.
Northway  WH  Jr, Rosan  RC, Porter  DY.  Pulmonary disease following respirator therapy of hyaline-membrane disease. Bronchopulmonary dysplasia.  N Engl J Med. 1967;276(7):357-368.PubMedGoogle ScholarCrossref
6.
Tooley  WH.  Epidemiology of bronchopulmonary dysplasia.  J Pediatr. 1979;95(5 Pt 2):851-858.PubMedGoogle ScholarCrossref
7.
Shennan  AT, Dunn  MS, Ohlsson  A, Lennox  K, Hoskins  EM.  Abnormal pulmonary outcomes in premature infants: prediction from oxygen requirement in the neonatal period.  Pediatrics. 1988;82(4):527-532.PubMedGoogle Scholar
8.
Beam  KS, Aliaga  S, Ahlfeld  SK, Cohen-Wolkowiez  M, Smith  PB, Laughon  MM.  A systematic review of randomized controlled trials for the prevention of bronchopulmonary dysplasia in infants.  J Perinatol. 2014;34(9):705-710.PubMedGoogle ScholarCrossref
9.
Jobe  AH, Bancalari  E.  Bronchopulmonary dysplasia.  Am J Respir Crit Care Med. 2001;163(7):1723-1729.PubMedGoogle ScholarCrossref
10.
Walsh  MC, Wilson-Costello  D, Zadell  A, Newman  N, Fanaroff  A.  Safety, reliability, and validity of a physiologic definition of bronchopulmonary dysplasia.  J Perinatol. 2003;23(6):451-456.PubMedGoogle ScholarCrossref
11.
Walsh  MC, Yao  Q, Gettner  P,  et al; National Institute of Child Health and Human Development Neonatal Research Network.  Impact of a physiologic definition on bronchopulmonary dysplasia rates.  Pediatrics. 2004;114(5):1305-1311.PubMedGoogle ScholarCrossref
12.
Poindexter  BB, Feng  R, Schmidt  B,  et al; Prematurity and Respiratory Outcomes Program.  Comparisons and limitations of current definitions of bronchopulmonary dysplasia for the prematurity and respiratory outcomes program.  Ann Am Thorac Soc. 2015;12(12):1822-1830.PubMedGoogle ScholarCrossref
13.
Ehrenkranz  RA, Walsh  MC, Vohr  BR,  et al; National Institutes of Child Health and Human Development Neonatal Research Network.  Validation of the National Institutes of Health consensus definition of bronchopulmonary dysplasia.  Pediatrics. 2005;116(6):1353-1360.PubMedGoogle ScholarCrossref
14.
Canadian Neonatal Network. The Canadian Neonatal Network Abstractor's Manual, version 1.3.4. http://www.canadianneonatalnetwork.org/Portal/LinkClick.aspx?fileticket=U4anCYsSN20%3D&tabid=69. Accessed December 15, 2016.
15.
Synnes  A, Luu  TM, Moddemann  D,  et al; Canadian Neonatal Network and the Canadian Neonatal Follow-Up Network.  Determinants of developmental outcomes in a very preterm Canadian cohort.  Arch Dis Child Fetal Neonatal Ed. 2016;fetalneonatal-2016-311228.PubMedGoogle Scholar
16.
Canadian Neonatal Follow-Up Network. 18-Month Corrected Age Assessment Manual, Version 5. http://www.cnfun.ca/LinkClick.aspx?fileticket=d4p7mZoXWDU%3D&tabid=68. Accessed December 15, 2016.
17.
Morris  C, Bartlett  D.  Gross motor function classification system: impact and utility.  Dev Med Child Neurol. 2004;46(1):60-65.PubMedGoogle ScholarCrossref
18.
Bayley  N.  The Bayley Scales of Infant and Toddler Development. 3rd ed. San Antonio, TX: The Psychological Corporation; 2006.
19.
Richardson  DK, Corcoran  JD, Escobar  GJ, Lee  SK.  SNAP-II and SNAPPE-II: simplified newborn illness severity and mortality risk scores.  J Pediatr. 2001;138(1):92-100.PubMedGoogle ScholarCrossref
20.
Papile  LA, Burstein  J, Burstein  R, Koffler  H.  Incidence and evolution of subependymal and intraventricular hemorrhage: a study of infants with birth weights less than 1,500 gm.  J Pediatr. 1978;92(4):529-534.PubMedGoogle ScholarCrossref
21.
Bell  MJ, Ternberg  JL, Feigin  RD,  et al.  Neonatal necrotizing enterocolitis. Therapeutic decisions based upon clinical staging.  Ann Surg. 1978;187(1):1-7.PubMedGoogle ScholarCrossref
22.
Steyerberg  EW. Evaluation of performance. In: Steyerberg  EW, ed.  Clinical Prediction Models: a Practical Approach to Development, Validation, and Updating. New York: Springer; 2009.
23.
Zou  KH, O’Malley  AJ, Mauri  L.  Receiver-operating characteristic analysis for evaluating diagnostic tests and predictive models.  Circulation. 2007;115(5):654-657.PubMedGoogle ScholarCrossref
24.
Davis  PG, Thorpe  K, Roberts  R, Schmidt  B, Doyle  LW, Kirpalani  H; Trial Indomethacin Prophylaxis in Preterms Investigators.  Evaluating “old” definitions for the “new” bronchopulmonary dysplasia.  J Pediatr. 2002;140(5):555-560.PubMedGoogle ScholarCrossref
25.
Mercier  CE, Dunn  MS, Ferrelli  KR, Howard  DB, Soll  RF; Vermont Oxford Network ELBW Infant Follow-Up Study Group.  Neurodevelopmental outcome of extremely low birth weight infants from the Vermont Oxford network: 1998-2003.  Neonatology. 2010;97(4):329-338.PubMedGoogle ScholarCrossref
26.
Maitre  NL, Ballard  RA, Ellenberg  JH,  et al; Prematurity and Respiratory Outcomes Program.  Respiratory consequences of prematurity: evolution of a diagnosis and development of a comprehensive approach.  J Perinatol. 2015;35(5):313-321.PubMedGoogle ScholarCrossref
×