Cognitive Development and Quality of Life Associated With BPD in 10-Year-Olds Born Preterm

The Extremely Low Gestational Age Newborns (ELGAN) study is a multicenter, prospective, observational study of the risk of structural and functional neurologic disorders in extremely preterm infants.³² A total of 1506 infants born before the 28th week of gestation were enrolled during the years 2002–2004, and 1198 survived to 10 years; 966 children had data on inflammatory related proteins and were recruited for an age-appropriate neurocognitive assessment. Of these 966 children, 889 (92%) returned for follow-up; 882 had diagnosis of BPD at 36 weeks’ corrected gestational age. In this report, we compare the prevalence of low scores on assessments of cognition (Differential Ability Scales II [DAS II], n = 863), executive function (NEuroPSYchological Assessment II [NEPSY II], n = 834), academic achievement (Wechsler Individual Achievement Test-III [WIAT III], n = 854, and Oral and Written Language Scales [OWLS], n = 842) limitations, and quality of life (Pediatric Quality of Life Inventory [PedsQL], n = 848), as well as undesirably high scores on assessments of social functions (Social Responsiveness Scale [SRS], n = 842) among children with a diagnosis of BPD (Supplemental Table 1). Enrollment and consent procedures for this study were approved by the institutional review boards of all participating institutions. Demographic, pregnancy, and newborn variables have been reported elsewhere.³² 

The physiologic definition of BPD was first published after the initiation of the ELGAN study.³³ Consequently, information about the results of a room air challenge were not available for most ELGAN study subjects. We modified Jobe’s³⁴ definition to allow us to base severity of BPD on data we collected at 36 weeks’ postmenstrual age (PMA) about oxygen supplemental and need for ventilation assistance. Our definitions are as follows: no to mild BPD (no oxygen), moderate (oxygen only), and severe (oxygen and ventilator assistance). Our definition of severe BPD differs from that of the Eunice Kennedy Shriver National Institute of Child Health and Human Development consensus definition of severe BPD in that it is only a subset of the Eunice Kennedy Shriver National Institute of Child Health and Human Development consensus definition of severe BPD.

Families willing to participate were scheduled for 1 visit, during which all of the measures reported here were administered in 3 to 4 hours, including breaks. The assessments were selected to provide the most comprehensive information about cognitive, academic function. While the child was tested, the parent or caregiver completed a general health questionnaire (medical diagnoses and receipts of medications), language, behavior, and quality of life. Questionnaires were also provided to the child’s school teacher to obtain teacher-reported behavioral status. Examiners who were unaware of the child’s medical history assessed neurodevelopment in several clinically important domains.

General cognitive ability or IQ was assessed with the DAS II verbal and nonverbal reasoning scales.³⁵ Expressive and receptive language skills were evaluated with the OWLS, in which semantic, morphologic, syntactic, and pragmatic production and comprehension of elaborated sentences are assessed.³⁶ The WIAT III provides standard scores in word recognition and decoding, spelling, and numeric operations.³⁷ Attention and executive function were assessed with both the DAS II and the NEPSY II.^38,39 

Children’s motor function was assessed with Gross Motor Function Classification System (GMFCS). A child was classified as level III or higher if he or she needed mobility assistance (level III: walks using a hand-held mobility device; level IV: self-mobility with limitations, may use powered mobility; level V: transported in a manual wheelchair).⁴⁰ 

Children’s communication function was assessed with the Communication Function Classification System (CFCS).⁴¹ A child was classified as level 3 or higher if he or she needed assistance in communicating with partners (level III: communicates with familiar partners effectively but needs assistance with unfamiliar partners; level IV: inconsistently sends and receives messages with familiar partners; level V: seldom sends and receives messages).

Seizure identification was achieved via a screening questionnaire followed by an open-ended interview conducted by an epileptologist.⁴² 

If the child screened positive for autism on the parent-completed Social Communication Questionnaire (SCQ), the parent was then asked to complete the Autism Diagnostic Interview, Revised (ADIR).⁴³ Children who met the modified ADIR criteria for autism spectrum disorder (ASD) were administered the Autism Diagnostic Observation Schedule 2 (ADOS2).⁴⁴ All children who met standardized criteria for ASD on both ADIR and ADOS2 were classified as having ASD.

Children were also assessed with the SRS,^44,45 the Dean Handedness Survey^46,47 and questionnaires for the Manual Ability Classification System,⁴⁸ the GMFCS,⁴⁰ the PedsQL,^49,50 the Children’s Communication Checklist 2 (CCC2),⁵¹ and the SCQ.⁵² 

We evaluated the null hypothesis that BPD is not associated with any cognitive, executive function, language, academic achievement, neurologic, social, or behavioral outcomes. We began by assessing correlates of BPD, including the maternal demographic and newborn characteristics at birth (Supplemental Table 2). Then we evaluated the overall distribution of each assessment at age 10 years, including the 25th, 50th, and 75th centile values (Fig 1). Because we did not evaluate a term comparison group, we relied on comparisons to historical normative samples that are described by the authors of the assessments we used.^35,–38 

For assessment of neurocognitive and social function, we created logistic regression models of the risk of a score 1 or more SDs less than the normative mean of each assessment (that is, z scores ≤−1). These models, which included potential confounders (maternal education, mother’s eligibility for government-provided medical care insurance, delivery for preeclampsia or fetal indication, gestational age, and birth weight z score), allowed us to calculate odds ratios indicating the strength of the association between the severity of BPD and each outcome. Unadjusted forest plots are available in Supplemental Figs 6–9.

Of the 863 children who had an assessment of IQ at or near age 10 years, 9% (n = 78) had severe BPD (at 36 weeks’ PMA required ventilation assistance as well as supplemental oxygen), 43% (n = 372) had moderate BPD (required supplemental oxygen, but not ventilation assistance), and 48% (n = 413) did not have BPD. A higher prevalence of BPD was also associated with younger maternal age at the time of delivery and absence of college degree. Children delivered for maternal or fetal indications had a higher prevalence of BPD and severe BPD than children delivered for spontaneous indications. Overall, the lower the gestational age, birth weight, and birth weight z score, the higher the prevalence of BPD and severe BPD (Supplemental Table 2).

The central line in the box of the box-and-whisker plots indicates the median (50th centile), whereas the top of the box indicates the 75th centile and the bottom of the box indicates the 25th centile. Children with severe BPD consistently have the lowest scores on cognitive function (DAS II), language and academic achievement (OWLS, WIAT III), and executive function (NEPSY II) (Fig 1).

Eight percent to 10% of children who did not have BPD had intellectual disability as measured by DAS II component scores >2 SDs less than the expected mean (ie, z scores ≤−2). In contrast, 19% to 22% of children who had moderate BPD had such low scores, and 30% to 32% of children who had severe BPD had such low scores. This gradient of the lowest prevalence of a z score ≤−2 among those who did not have BPD and the highest prevalence among infants who had severe BPD applied to language functioning using the OWLS and academic achievement in reading, mathematics, and spelling using the WIAT III, and the opposite gradient was observed for z scores >1 for DAS II, OWLS, and WIAT III components (Supplemental Table 3).

This gradient of z scores along the path from no BPD to severe BPD also applied to DAS II working memory and the following components of the NEPSY II: auditory attention, auditory response set, inhibition inhibition, inhibition switching, animal sorting, inhibition naming, arrows, and visuo-motor function. This BPD gradient was also present for less severely low scores (ie, z scores >−2, ≤−1) on the NEPSY II geometric puzzles assessment. The gradient observed between severe BPD and no or mild BPD occurred for z scores >1 (Supplemental Table 4).

After adjustment for potential confounding factors (gestational age category, low birth weight z score category, delivery for maternal or fetal indications, and maternal fever with 48 hours of delivery), odds ratios and 95% confidence intervals of z scores ≤−1 on each DAS II and NEPSY II neurocognitive assessment, these elevations were attenuated, but by and large they remained statistically significant for severe BPD, whereas for moderate BPD the following were no longer statistically significant: OWLS listening comprehension and oral expression, DAS II working memory, and NEPSY II auditory response set, inhibition inhibition, animal sorting, and arrows (Fig 2; unadjusted odds ratios can be found in Supplemental Fig 6).

When children who had an IQ of <70 were eliminated from this sample, after adjustment for potential confounding factors, the risks of low scores on verbal and nonverbal reasoning, the DAS II working memory, auditory response set, and inhibition inhibition; OWLS oral expression; and WIAT III word reading components were no longer statistically significant among children with severe BPD. In contrast, children with moderate BPD were at increased risk of low scores on NEPSY II inhibition switching and WIAT III numerical operations components (Fig 3; unadjusted odds ratios can be found in Supplemental Fig 7).

The typical BPD gradient was seen for ADOS2-based diagnoses of ASD (13%, 9%, 4%), motor impairments (ie, higher classification on the Manual Ability Classification System and the GMFCS), communication impairment (highest classification on the CFCS for children), and lower quality of life (ie, higher scores on the PedsQL inventory, including physical, social, school, and psychosocial functioning). Lower quality of emotional function was similarly associated with both severe and moderate BPD (30%–31% vs 24%). Children who had severe BPD had higher prevalence than others of a seizure, epilepsy, receipt of medication for seizures, and left handedness (Supplemental Table 5).

For children who had severe BPD, these elevations were somewhat less prominent after adjustment for potential confounding factors, but the elevated odds of a high T-score on the total score, as well as the social cognition, social communication, and social motivation components, remained statistically significant (Fig 4; unadjusted odds ratios can be found in Supplemental Fig 8). The risk of a “positive” SCQ assessment also remained significantly elevated. After adjustments, children with moderate BPD were also at significantly increased risk of low scores on the speech, coherence, and initiation components (Fig 5; unadjusted odds ratios can be found in Supplemental Fig 9).

In this cohort of 10-year-old children born before 28 weeks of gestation, those who had BPD were at increased risk of neurocognitive, behavioral, and social dysfunctions. Children with severe BPD were more likely than other children to have low scores on cognitive, language, academic achievement, executive, emotional, social, and physical functioning. Their parents rated them as experiencing a lower quality of life. The multitude of dysfunctions we found at 10 years in children with BPD, even after adjusting for potential confounders, were most prominent in children who experienced severe BPD.

Lower cognitive function and academic achievement have been found repeatedly among children who had BPD than among their peers.^4,10,–14 Similar to our study, the majority of the studies of school-aged children found that children with BPD have higher rates of executive function limitations compared with their peers.^16,–20 The main difference between our study and other studies is the modification of classification of BPD by Jobe.³⁴ We stratified our BPD into moderate with supplemental oxygenation and severe with ventilatory support and supplemental oxygen, whereas in other studies, researchers combined all categories of BPD into with and without BPD.

Prematurity may be the main reason for the development of BPD. Other contributions might come from an increase in reactive oxygen species.^53,54 

Children who experience sepsis^55,56 and those who acquire other nosocomial infections⁵⁷ are at risk for BPD. Infections in the perinatal period are associated with white matter abnormalities and later motor and cognitive impairments.^58,–61 

Assisted ventilation appears to increase the probability of systemic inflammation,⁶² which in turn is associated with increased risk of BPD.²² Lung inflammation contributes to the pathogenesis of BPD and may be accompanied by systemic responses. Systemic inflammation appears to put newborns at increased risk of multiple indicators of brain injury.^63,–69 

BPD risk is heightened among infants born small for gestational age,^70,71 who are also at increased risk of delayed development.⁷² The complexity of the situation is suggested by the observation that newborn lambs who are growth restricted are at increased risk of ventilation-induced brain damage.⁷³ 

Children whose pulse oximeter oxygen saturation was in the range of 85% to 89% were not at increased risk of disability.⁷⁴ Thus, it is unlikely that low oxygen saturation contributes to what we found.

Pain-related stress in the very preterm newborn has been associated with altered brain structure and function.⁷⁵ These findings have encouraged efforts to minimize pain, including that associated with prolonged assisted ventilation.⁷⁶ Among children born very premature, those exposed to sedatives and/or opioid drugs for >7 days performed as well on neurodevelopmental assessments at age 5 as their 1457 peers who were not so exposed.⁷⁷ In light of “limited evidence of benefit and potential for harm,”⁷⁸ the use of opioids for agitation resulting from mechanical ventilation remains controversial.⁷⁹ 

Lagercrantz⁸⁰ has suggested that “more skin-to-skin care for several hours per day and talking and singing to the infant a lot may help” diminish the occurrence of ASD among children on the ventilator. If this hypothesis is correct, some of the social dysfunctions we identified might be easily preventable.

The high risk of dysfunctions among children who had BPD, especially severe BPD, is in keeping with the hypothesis that multiple risk factors together and/or in sequence might be especially damaging to the developing brain.^81,–84 Extreme prematurity and/or vulnerability and fetal growth restriction are among the earliest risk factors, whereas systemic inflammation might be the most obvious of the subsequent risk factors.^83,85 Part of the process might reflect secondary exacerbation of damage,^86,87 although the sequence of events in some situations likely begins with sensitization.⁸⁸ 

Elimination of children who had a median IQ of <70 from the sample did not change our findings appreciably, suggesting that the functions we assessed are not due to low IQ.^89,–91 Even though children with severe BPD were more likely to have low scores on cognitive, language, academic achievement, executive emotional, social and physical functioning, and diminished quality of life, approximately one-half of our cohort had scores in the normal range for academic achievement (Supplemental Table 3).

For our current study, we wanted to see to what extent BPD predicted later dysfunctions regardless of the contributions of antecedents. We recognize that systemic inflammation and other correlates of ventilation and BPD might have contributed to what we found.^24,63,92 

Our study has several strengths. First, in contrast to a majority of the existing studies, our sample size is large, making it unlikely that we have missed important associations because of lack of statistical power or that we claimed associations that might reflect stability of small numbers. Second, we selected infants on the basis of gestational age, not birth weight, to minimize confounding due to factors related to growth restriction.⁹³ Third, we collected our data prospectively. Fourth, attrition at neurocognitive assessment was modest. Fifth, the individuals who evaluated study participants at 10 years of age were not aware of their medical histories. The weaknesses of our study are those of all observational studies. We are unable to distinguish between causation and association as explanations for what we found.

Ten years after they were born extremely preterm in this cohort, those who had BPD were at increased risk of cognitive, language, and executive dysfunctions, as well as low scores on assessments of academic achievement, social skills, and quality of life. These limitations were especially prominent among those who were ventilator-dependent at 36 weeks’ PMA. Almost half of children who had severe BPD had scores in the normal range for academic achievement.

 
• ADIR

  Autism Diagnostic Interview, Revised

 
• ADOS2

  Autism Diagnostic Observation Schedule 2

 
• ASD

  autism spectrum disorder

 
• BPD

  bronchopulmonary dysplasia

 
• CCC2

  Children’s Communication Checklist 2

 
• CFCS

  Communication Function Classification System

 
• DAS II

  Differential Ability Scales II

 
• ELGAN

  Extremely Low Gestational Age Newborns

 
• GMFCS

  Gross Motor Function Classification System

 
• NEPSY II

  NEuroPSYchological Assessment II

 
• OWLS

  Oral and Written Language Scales

 
• PedsQL

  Pediatric Quality of Life Inventory

 
• PMA

  postmenstrual age

 
• SCQ

  Social Communication Questionnaire

 
• SRS

  Social Responsiveness Scale

 
• WIAT III

  Wechsler Individual Achievement Test-III

The ELGAN Study Investigators were as follows: Janice Ware, Taryn Coster, Brandi Henson, Rachel Wilson, Kirsten McGhee, Patricia Lee, Aimee Asgarian, Anjali Sadhwani, Boston Children’s Hospital (Boston, MA); Ellen Perrin, Emily Neger, Kathryn Mattern, Jenifer Walkowiak, Susan Barron, Tufts Medical Center (Boston, MA); Jean Frazier, Lauren Venuti, Beth Powers, Ann Foley, Brian Dessureau, Molly Wood, Jill Damon-Minow, University of Massachusetts Medical School (Worcester, MA); Richard Ehrenkranz, Jennifer Benjamin, Elaine Romano, Kathy Tsatsanis, Katarzyna Chawarska, Sophy Kim, Susan Dieterich, Karen Bearrs, Yale University School of Medicine (New Haven, CT); T. Michael O’Shea, Nancy Peters, Patricia Brown, Emily Ansusinha, Ellen Waldrep, Jackie Friedman, Gail Hounshell, Debbie Allred, Wake Forest University Baptist Medical Center (Winston-Salem, NC); Stephen C. Engelke, Nancy Darden-Saad, Gary Stainback, University Health Systems of Eastern Carolina (Greenville, NC); Diane Warner, Janice Wereszczak, Janice Bernhardt, Joni McKeeman, Echo Meyer, North Carolina Children’s Hospital (Chapel Hill, NC); Steve Pastyrnak, Wendy Burdo-Hartman, Julie Rathbun, Sarah Nota, Teri Crumb, Helen DeVos Children’s Hospital (Grand Rapids, MI); Madeleine Lenski, Deborah Weiland, Megan Lloyd, Sparrow Hospital (Lansing, MI); Michael D. Schreiber, Scott Hunter, Michael Msall, Rugile Ramoskaite, Suzanne Wiggins, Krissy Washington, Ryan Martin, Barbara Prendergast, Megan Scott, University of Chicago Medical Center (Chicago, IL); and Judith Klarr, Beth Kring, Jennifer DeRidder, Kelly Vogt, William Beaumont Hospital (Royal Oak, MI).

We thank the children and their families who participated in this study. We also acknowledge the contributions of the ELGAN Study Investigators.