ÌÇÐÄvlog

ImportanceÌý Active perinatal care increases the rate of survival of extremely preterm infants, but there are concerns that improved survival might increase the rate of disabled survivors.

ObjectiveÌý To determine the neurodevelopmental outcomes of a national cohort of children 6.5 years of age who had been born extremely preterm (<27 weeks’ gestational age) in Sweden.

Design, Setting, and ParticipantsÌý Population-based prospective cohort study of consecutively born extremely preterm infants. All of these infants were born in Sweden during the period from April 1, 2004, to March 31, 2007. Of 707 live-born extremely preterm infants, 486 (68.7%) survived to 6.5 years of age. These children were assessed and compared with matched controls who had been born at term. Comparison estimates were adjusted for demographic differences. Assessments ended in February 2014, and analysis started thereafter.

Main Outcomes and MeasuresÌý Cognitive ability was measured with the fourth edition of the Wechsler Intelligence Scale for Children (WISC-IV), and the mean (SD) scores of the children who had been born extremely preterm were compared with those of the controls. Clinical examinations and parental questionnaires were used for diagnosis of cerebral palsy, hearing and vision impairments, and cognition for the children who were not assessed with the WISC-IV.

ResultsÌý Of 486 eligible infants who were born extremely preterm, 441 (90.7%) were assessed at 6.5 years of age (59 by medical record review only) alongside 371 controls. The adjusted mean (SD) full-scale WISC-IV score was 14.2 (95% CI, 12.1-16.3) points lower for children who had been born extremely preterm than for controls. Cognitive disability was moderate for 18.8% of extremely preterm children and 2.2% of controls (P < .001), and it was severe for 11.1% of extremely preterm children and 0.3% of controls (P < .001). Cerebral palsy was observed in 9.5% of extremely preterm children and 0.0% of controls (P < .001), blindness was observed in 2.0% of extremely preterm children and 0.0% of controls (P < .001), and hearing impairment was observed in 2.1% of extremely preterm children and 0.5% of controls (P = .07). Overall, 36.1% (95% CI, 31.7%-40.6%) of extremely preterm children had no disability, 30.4% (95% CI 26.3%-34.8%) had mild disability, 20.2% (95% CI, 16.6%-24.2%) had moderate disability, and 13.4% (95% CI, 10.5%-16.9%) had severe disability. For extremely preterm children, moderate or severe overall disability decreased with gestational age at birth (adjusted odds ratio per week, 0.65 [95% CI, 0.54-0.79]; P < .001) and increased from 26.6% to 33.5% (P = .01) for children assessed both at 2.5 and 6.5 years.

Conclusions and RelevanceÌý Of the 441 extremely preterm infants who had received active perinatal care, 293 (66.4%) had no or mild disability at 6.5 years; of the 371 controls, 11 (3.0%) had moderate or severe disability. Disability rates at 6.5 years increased relative to the rates at 2.5 years. Results are relevant for health care professionals and planners, and for clinicians counseling families facing extremely preterm births.

Children who had been born extremely preterm are at increased risk of neurodevelopmental disabilities (NDDs), such as cerebral palsy (CP), cognitive disability, and disabilities caused by vision and hearing impairments.¹ As increasing numbers of infants born at the limit of viability are offered neonatal intensive care and as survival rates continue to increase,^2-4 extremely preterm birth has become an important public health issue. Furthermore, knowledge about the outcomes might influence resuscitation policies, and for many parents, the fear that their infant might survive with long-term NDDs is a matter possibly more important than survival.⁵ Few studies have evaluated the neurodevelopmental outcomes in contemporary cohorts of children at school age who had been born extremely preterm. One meta-analysis⁶ found 9 high-quality studies that included infants born at less than 26 gestational weeks who were evaluated at 4 to 8 years of age. According to the definitions used in these studies, severe NDD is likely to render the child highly dependent on caregivers, whereas a moderate disability, although of functional importance, is likely to allow a reasonable degree of independence. However, most studies included in the meta-analysis⁶ were small and were not population-based studies, and not all had control groups.

The Extremely Preterm Infants in Sweden Study (EXPRESS)³ includes all infants born before 27 weeks of gestation during the period from 2004 to 2007 in Sweden. The cohort was followed up at 2.5 years of age corrected for prematurity.⁷ Because school-aged children better reflect the deficits that are likely to be sustained from extreme prematurity, the aim of the present study was to determine the neurodevelopmental outcomes of children at 6.5 years of age who had been born extremely preterm and to compare these outcomes with those of a matched control group born at term.

Box Section Ref ID

Perinatal and neonatal data were prospectively collected for all 1011 infants born at 22 to less than 27 gestational weeks in Sweden.³ Of 707 live-born infants, 494 (70.0%) survived to 1 year of age; 456 of these infants were assessed at 2.5 years’ corrected age and compared with a control group of children 2.5 years of chronological age who had been born at term.⁷ Perinatal data, including determination of gestational age, have been reported.^3,8 The Regional Ethics Review Board, Lund, Sweden, approved the study; parents provided written informed consent.

The EXPRESS cohort was invited to participate in a comprehensive neurodevelopmental assessment at 6.5 years of chronological age. Eight children who had been born extremely preterm had died between 1 and 6.5 years; thus, 486 children (68.7% of all live births) were eligible for participation. The control group was recruited for the 2.5-year follow-up,⁷ and additional children were recruited at 6.5 years to substitute for attrition. Control children were selected from the Swedish Medical Birth Registry; selection criteria were a singleton birth, a gestational age of 37 to 41 weeks, and an Apgar score greater than 3 at 5 minutes, with matching of control children for place of residence, sex, day of birth, and maternal country of origin. Follow-up ended in February 2014.

The evaluation of the children who had been born extremely preterm included a clinical examination, an interview with the primary caregiver, and a psychological evaluation. Fifty-nine children who had been born extremely preterm but not physically examined at 6.5 years were assessed by medical record review; local pediatricians and rehabilitation centers provided information that enabled classification of outcomes into disability categories.

Cognitive ability was assessed by psychologists who used the Swedish version of the Wechsler Intelligence Scale for Children, fourth edition (WISC-IV).⁹ The full-scale IQ (FSIQ) provided information on general intelligence, and the 4 index scores measured specific cognitive domains. For 4 children who had been born extremely preterm but had incomplete WISC-IV assessments at 6.5 years, FSIQ scores were calculated through available subtests. Of 8 children who had been born extremely preterm and not tested with the WISC-IV at 6.5 years but had received a diagnosis from rehabilitation services of moderate or severe cognitive delay, 5 were assigned FSIQ scores of 69 (2 SDs below the normative mean), 2 were assigned FSIQ scores of 54 (3 SDs below the normative mean), and 1 was assigned an FISQ score of 39 (4 SDs below the normative mean). The FSIQ scores of children who had been born extremely preterm were related to the mean FSIQ score and the distribution of FSIQ scores of the control group.

For 11 children who had been born extremely preterm and not tested with the WISC-IV at 6.5 years but who had been physically examined, cognitive disability categories (not WISC-IV scores) were defined by using the results of the clinical examinations by the specialist in the study team and the collateral information from the medical records. The cognitive disability categories from 59 children who had been born extremely preterm but not physically examined at 6.5 years were based on medical record reviews.

Cerebral palsy was defined according to Bax et al,¹⁰ and the severity of the CP was determined through the Gross Motor Function Classification System (GMFCS).¹¹ Assessment of hearing was based on the child’s dependence on hearing aids. Visual acuity was assessed by ophthalmologists and classified according to modified World Health Organization criteria.¹² Examiners were not blinded to group status.

Severe NDD was defined as having an FSIQ score of less than the mean FSIQ score −3 SDs or as severe cognitive disability determined by a clinical examination or medical record review, or as CP (GMFCS level of ≥4), blindness (visual acuity of <20/400 in the better eye), or deafness (impairment not corrected with hearing aid). Moderate NDD was defined as having an IQ score from −3 SDs to less than −2 SDs or as moderate cognitive disability determined by a clinical examination or medical record review, or as CP (GMFCS level of 2-3), visual impairment (visual acuity of <20/63 but ≥20/400 in the better eye), or hearing impairment (hearing loss corrected with hearing aid). Mild NDD was defined as having an IQ score from −2 SDs to less than −1 SD or as mild cognitive disability determined by a clinical examination or medical record review, CP (GMFCS level of 1), or visual impairment (visual acuity of <20/40 but ≥20/63 in the better eye).

For Information regarding maternal and paternal education, socioeconomic status, and health, a questionnaire adapted from the Nordic Health and Family Questionnaire¹³ was used. For controls, the clinical examination was substituted for a parental questionnaire based on the Nordic Health and Family Questionnaire,¹³ which provided information on CP, hearing, health, and parental education. The WISC-IV and visual acuity assessments were performed for controls in the same manner as for the children who had been born extremely preterm. Cognition was assessed through a medical record review for 4 controls.

As a part of this longitudinal investigation, neuropsychological, behavioral, and motor outcomes were also evaluated. These outcomes will be reported elsewhere.

Groupwise comparisons of descriptive data were made using the χ² test. The relation between gestational age (in the extremely preterm group) and WISC-IV scores was investigated using analysis of covariance, with gestational age and parental educational levels as linear variables and maternal country of birth (non-Nordic vs Nordic) as a class variable. Agreements between NDD-classification estimates at 2.5 and 6.5 years were assessed using the Cohen κ statistic, and the differences in moderate/severe disabilities between the 2 assessments (ie, 2.5 and 6.5 years) were assessed using the McNemar test. The odds ratios (ORs) for severe/moderate disability (vs mild/no disability) among children who had been born extremely preterm vs controls were estimated using logistic regression, and adjustments were made as specified. The effect of male sex on morbidity was adjusted for gestational age (entered as a linear variable). Statistical analyses were performed using Gauss statistical software (Aptech Systems Inc []) and SPSS for Windows version 22 (SPSS Inc). A 2-tailed P < .05 was regarded as statistically significant.

Of 486 eligible children who had been born extremely preterm, 22 (4.5%) could not be traced (4 [0.8%] lived outside Sweden, 2 [0.4%] had protected identity, and 16 [3.3%] had preliminary identity numbers given at birth that did not match), and 23 (4.7%) declined participation. Thus, 441 of 486 children who had been born extremely preterm (90.7%) participated, including 59 children assessed by medical record review. Of these 441 children, 433 were assessed at 2.5 years’ corrected age. The matched control group comprised 371 children, of whom 202 participated at 2.5 years and 169 were newly recruited. The median age at assessment was 6.6 years for children who had been born extremely preterm and 6.7 years for controls.

Maternal and neonatal characteristics at birth were similar among participating and nonparticipating mothers and children who had been born extremely preterm. The disability rates at 2.5 years⁷ did not differ between children who were reassessed at 6.5 years and those who declined reassessment (n = 23). However, children who had been born extremely preterm and who had moderate/severe disabilities or mental developmental delay at 2.5 years’ corrected age⁷were more likely to be evaluated by medical record review than by physical evaluation, as were children with less educated fathers (eTable 1 in the Supplement).

The mean (SD) gestational at birth for children born extremely preterm was 25.4 (1.1) weeks. In the extremely preterm group, congenital malformations and being small for gestational age were more frequent than in the control group, parental education was lower, and more mothers were of non-Nordic origin, smokers, primiparae, or younger than 20 years of age (Table 1).

Cognition was assessed by using the WISC-IV for 371 of 441 children who had been born extremely preterm and for 367 of 371 controls. The mean FSIQ score was 16.9 points (95% CI, 14.9-18.8) lower for children who had been born extremely preterm than for controls (P < .001) (Table 2). When adjusted for baseline characteristics, the difference was marginally attenuated (14.2 [95% CI, 12.1-16.3]). The magnitude of the group differences was similar across the 4 index scores. Moderate/severe cognitive disability as measured by the WISC-IV was present in 119 children who had been born extremely preterm (32.1%) and 8 controls (2.2%).

For the 371 children who had been born extremely preterm, the mean (SD) FSIQ scores were 76 (5.4) at 22 weeks for 4 children, 75 (13.8) at 23 weeks for 37 children, 80 (13.7) at 24 weeks for 70 children, 84 (14.0) at 25 weeks for 128 children, and 88 (15.1) at 26 weeks for 132 children. The sex-adjusted FSIQ increase in score per week was 4.1 points (95% CI, 2.7-5.5) (P &±ô³Ù; .01).

Moderate/severe cognitive disability assessed by either the WISC-IV or clinical examination/medical record review was present in 132 of 441 children who had been born extremely preterm (29.9% [95% CI, 25.8%-34.4%]) and 9 of 371 controls (2.4% [95% CI, 1.1%-4.2%]) (Table 3); the unadjusted and adjusted ORs (vs no or mild disability) were 17.2 (95% CI, 8.6-34.3) and 15.3 (95% CI, 7.6-30.7), respectively.

Of the 441 children who had been born extremely preterm, 42 (9.5% [95% CI, 7.1%-12.6%]) had CP; no controls had CP (Table 3). Of these 42 children, 17 (40.5%; 3.9% of all children who had been born extremely preterm) had moderate/severe CP. Thirty-five children had spastic CP, and 7 children had CP of other types (2 ataxic, 1 dyskinetic, 2 unclassifiable, and 2 of unknown type). Thirty-nine children who had been born extremely preterm (8.8% [95% CI, 6.5%-12.1%]) and 3 controls (0.8% [95% CI, 0.3%-2.3%]) had moderate/severe neurosensory disabilities.

No or mild disability was present in 293 of 441 children who had been born extremely preterm (66.4% [95% CI, 61.9%-70.7%]) and 360 of 371 controls (97.0% [95% CI, 94.7%-98.4%]), and moderate/severe disability was present in 148 of 441 children who had been born extremely preterm (33.6% [95% CI, 29.3%-38.1%]) and 11 of 371 controls (3.0% [95% CI, 1.6%-5.3%]) (Table 3). The unadjusted OR for moderate/severe disability (vs no/mild) was marginally reduced from 16.5 (95% CI, 8.8-31.1) to 15.1 (95% CI, 8.0-28.5) when adjusted for baseline characteristics.

Of the 148 children who had been born extremely preterm and who have moderate/severe disabilities, 132 (89.2%) had a moderate/severe cognitive disability, either alone (n = 109) or in combination with a neurosensory disability (n = 23). Of the 11 controls with moderate/severe disability, 9 (81.8%) had a cognitive disability. The proportion of children who had been born extremely preterm classified as having overall moderate/severe disability (130 of 382 [34.0%]) and having had a physical evaluation was similar to that of children whose medical records were reviewed (18 of 59 [30.5%]). However, some domain-specific distributions differed (eTable 2 in the Supplement).

The outcome for the extremely preterm cohort is summarized in Table 4. The risk for moderate/severe disability among these children decreased with increasing gestational age (sex-adjusted OR per week, 0.65 [95% CI, 0.54-0.79]; P < .001), as did the risk for death or survival with severe disability at 6.5 years of age, which decreased from 92% at 22 weeks to 25% at 26 weeks (sex-adjusted OR per week, 0.50 [95% CI, 0.43-0.58]; P < .001). Malformations stratified by NDD categories are shown for these children in eTable 3 in the Supplement.

Of 433 children who had been born extremely preterm and who were assessed at 2.5 years’ corrected age, 202 (46.7%) remained at the same NDD category at 6.5 years of age (Table 5 and Table 6), 91 (21.0%) moved to a better category, and 140 (32.3%) moved to a worse category. The overall number of children with moderate/severe disabilities increased from 115 (26.6%) at 2.5 year to 145 (33.5%) at 6.5 years (P = .01). The predominant change was an increase in the number of children with a moderate/severe cognitive disability, from 83 (19.2%) at 2.5 years to 129 (29.8%) at 6.5 years (P < 01). Of 42 children who had been born extremely preterm and have CP, 14 were first detected at 6.5 years (11 had a mild disability and 3 had a moderate disability), and of 9 blind children, 4 had changed classification from moderately impaired to blind. Tables 5 and 6 show the predictive ability of the 2.5-year examination for classifying NDD at 6.5 years in our cohort.

For boys who had been born extremely preterm, the mean gestational age–adjusted FSIQ score was 4.3 points (95% CI, 1.5-7.2) lower than that for girls who had been born extremely preterm (P < .01); the proportion of boys with a moderate/severe cognitive disability was higher than the proportion of girls with a moderate/severe cognitive disability (34.9% vs 25.4%; gestational age–adjusted OR, 1.6 [95% CI, 1.0-2.4]; P < .05). Similarly, the overall percentage of boys with a moderate/severe disability was higher than the overall percentage of girls with a moderate/severe disability (38.1% vs 28.3%; gestational age–adjusted OR, 1.6 [95% CI, 1.1-2.5]; P < .05) (eTable 4 in the Supplement). There were no sex-related differences among controls.

In this national study, of 441 children born at less than 27 weeks’ gestational age, 159 (36.1%) had no disability, 134 (30.4%) had mild disability, and 148 (33.6%) had either moderate (89 [20.2%]) or severe (59 [13.4%]) overall disability of functional importance at 6.5 years. Cognitive deficits were common (89%) among children who had been born extremely preterm and who had moderate/severe disability.

The mean FSIQ score was 17 points lower for children who had been born extremely preterm than for controls. The IQ difference between children who had been born extremely preterm and controls was 24 points in the EPICure study¹⁵ and 18 points in the EPIPAGE (Etude Epidémiologique sur les Petits Ages Gestationnels)¹⁶ study (both studies had children who were <26 weeks’ gestational age), and in an Australian study¹⁷ comprising slightly more mature children (<28 weeks or <1000 g), the difference was 13 points. The extremely preterm cohort performed lower than the control group with regard to all WISC-IV index scores, suggesting a global cognitive deficit rather than impairment in any selective domain.¹⁷

In the present study, 132 of 441 children who had been born extremely preterm (30.0%) in our study had moderate/severe cognitive disability. In comparable studies, the prevalence at early school age^15,18-20 ranges from 9% in a Norwegian study²⁰ without a control group to 41% in the EPICure study.¹⁵ In the present study, the cognitive disability of children who had been born extremely preterm and who had been tested with the WISC-IV was classified relative to the mean (100.3) and SD (11.7) of the control group. Had we used test norms of a mean (SD) of 100 (15), the prevalence of moderate/severe cognitive disability as measured with the WISC-IV would have been reduced from 32% to 19% owing to the wider SD in the normative sample.⁹

In the present, study, one-third of children who had been born extremely preterm had a moderate/severe NDD. In a recent meta-analysis,⁶ the pooled rates of moderate/severe NDD at 4 to 8 years were 43% among children born at 22 weeks, 40% among children born at 23 weeks, 28% among children born at 24 weeks, and 24% among children born at 25 weeks. The 95% CIs were wide because the pooled numbers of participants were low, particularly at 22 to 23 weeks.

Neurodevelopmental outcomes of children born extremely preterm must be viewed in the context of survival. Compared with the studies included in the meta-analysis⁶ that provided survival rates based on live births, the survival rate in the EXPRESS cohort was higher³; in the EPICure study,² for instance, 11% of infants born at 23 weeks survived to be discharged home, whereas in the EXPRESS cohort,³ 52% of infants survived to 1 year. We attributed the increased survival rate in the EXPRESS cohort to active perinatal care; of infants born alive at 23 to 26 weeks of gestation, 95% were admitted for intensive care.³

Although the composite outcome of death or survival to 18 to 24 months among infants born alive is commonly reported, few studies¹⁵ monitor this outcome at 6 years. Despite favorable survival in the EXPRESS cohort, and morbidity rates being similar to those in comparable studies,^15,16,19-21 more than half of infants born at less than 25 weeks either died or survived with severe disability, clearly indicating the need for further improvements in perinatal care.

The agreement between the NDD classification at 2.5 years and the NDD classification at 6.5 years was weak. Less than half of the children who had been born extremely preterm and who have moderate/severe disability at 6.5 years were identified at 2.5 years, which suggests that children at risk might not attract timely attention. Conversely, the false positive rate was low, 15%. In the EPICure study,¹⁵ 86% of children who had been born extremely preterm and who had moderate/severe disability at 30 months were classified as severely impaired at 6 years, whereas less severe disabilities were poorly predictive. In contrast, in an Australian study,²¹ only 35% of children who had been born extremely preterm and who had severe disability at 2 years were classified as severely impaired at 8 years.

The percentage of children with moderate/severe cognitive disability increased from 2.5 to 6.5 years in our study, which might reflect a better ability to diagnose developmental changes with advancing age. Nevertheless, our finding is contrary to studies reporting lower cognitive disability rates with age.^22-24 Those studies, however, assessed developmental delay at first assessment with the second edition of the Bayley Scales of Infant Development,²⁵ which may not accurately predict cognitive outcome at school age.²⁴ The third edition of the Bayley Scales of Infant and Toddler Development²⁶ (Bayley-III), which was used at 2.5 years in the EXPRESS study,⁷ differs substantially from previous editions. Bayley-lll had strong predictive validity for cognitive outcome at 4 years in one study²⁷ but not in another.²⁸ Bayley-lll is reported to have poor sensitivity at lower scores,²⁹ which may explain the apparent increase in moderate/severe cognitive disability with age in our cohort. Moreover, the WISC-IV emphasizes elements of executive functioning. Because executive dysfunction is an area of concern for children who were born extremely preterm,^30,31 we might have diagnosed more children with cognitive disability compared with studies not using the WISC-IV.

The prevalence of CP (9.5%) in our study was low compared with the prevalence of CP in other studies (9%-20%),³² and in 60% of cases, the CP was mild. Cerebral palsy was newly detected in 14 children, of whom the majority had mild CP. This concurred with the recommendation by Hagberg et al³³ of not ruling out CP before 4 years of age. Of the 9 blind children, 4 were reclassified from moderately impaired at 2.5 years’ corrected age to blind at 6.5 years, illustrating the difficulty of performing visual examinations on toddlers.

There was a 4.3-point difference in mean FSIQ score between boys who had been born extremely preterm and girls who had been born extremely preterm, corresponding to a difference of 0.3 SD. This is a clinically important difference at the population level^22,34 that was reflected in the 60% increase in OR for moderate/severe cognitive disability in boys who had been born extremely preterm. In the EPICure study at 6 years,¹⁵ the sex-related difference of cognitive abilities was 10 points in favor of girls, and in a US study,³⁵ more boys than girls had cognitive delay (42% vs 27%) and NDD (48% vs 34%) at 18 to 24 months. We found no sex-related difference in the risk for CP, possibly owing to the small sample size.

The strengths of this study include the national, prospective, and longitudinal design. The retention rate of 90.7% was satisfactory considering the geographical dispersion and the age at follow-up. Only 23 children’s parents declined participation; the remaining nonparticipation was due to technical reasons, the most important being the inability to match children at later age to the preliminary identity numbers given at birth. Both the participating and nonparticipating children who had been born extremely preterm had similar neonatal and maternal backgrounds at 6.5 years and similar moderate/severe disability rates at 2.5 years. Some children were only assessed by medical record review. Although the overall distribution of disabilities did not differ between modes of examination, mild cognitive disability might have escaped detection in children whose medical records were reviewed.

At 6.5 years, additional controls were recruited, which might have biased the comparison with results obtained at 2.5 years. We therefore compared Bayley-III cognitive scores for the controls participating both at 2.5 and 6.5 years with the controls participating only at 2.5 years, and we compared WISC-IV scores for the new and old controls at 6.5 years. We found no substantial differences for any of the comparisons (eTable 5 in the Supplement) and deduce that no bias was introduced. Furthermore, lack of blinding, which was not possible to achieve because the study was part of a clinical follow-up, might have caused expectation bias.

The low working memory index of the control children was unexpected. There is a lack of published studies that use the Swedish version of the WISC-IV at 6 to 7 years; however, an ongoing study on children of the same age has observed low working memory scores in typically developing children (T. Klingberg, MD, PhD, written communication, October 2015), which indicates that the Swedish version may underestimate the working memory of children 6 to 7 years of age. Another ongoing study shows similar findings in term-born controls at 13 years of age (A. Farooqi, MD, PhD, written communication, March 2016). We are confident that the difference in WISC-IV scores between the 2 groups are valid despite the low working memory scores for both the control children and the children who had been born extremely preterm. Furthermore, we acknowledge that the disability criteria did not include behavior, attention, and learning disabilities that are commonly found among children who had been born extremely preterm.¹

In conclusion, two-thirds of the children who had been born extremely preterm had a normal development or mild disabilities at early school age. The disability rates were substantially higher among these children than among children born at term and were inversely related to gestational age. The outcomes were similar to those of comparable studies with lower survival rates. The disability rates at 6.5 years increased relative to the follow-up at 2.5 years’ corrected age and underscore the importance of long-term neurodevelopmental assessments of children who have been born extremely preterm, including those with apparently normal early development.

Accepted for Publication: April 25, 2015.

Corresponding Author: Fredrik Serenius, MD, PhD, Section for Pediatrics, Department of Women’s and Children’s Health, Uppsala University, S-751 85 Uppsala, Sweden (fredrik.serenius@kbh.uu.se).

Published Online: August 1, 2016. doi:10.1001/jamapediatrics.2016.1210.

Author Contributions: Drs Serenius and Källén had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Serenius, Ewald, Farooqi, Fellman, Hafström, Maršál, Olhager, Stjernqvist, Strömberg, Ådén, Källén.

Acquisition, analysis, or interpretation of data: Serenius, Ewald, Farooqi, Fellman, Hafström, Hellgren, Maršál, Ohlin, Stjernqvist, Strömberg, Ådén, Källén.

Drafting of the manuscript: Serenius, Farooqi, Hafström, Maršál, Stjernqvist, Strömberg, Ådén, Källén.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Serenius, Farooqi, Källén.

Obtained funding: Serenius, Ewald, Farooqi, Fellman, Hellgren, Olhager, Stjernqvist, Strömberg, Ådén.

Administrative, technical, or material support: Serenius, Farooqi, Fellman, Maršál, Ohlin, Olhager, Stjernqvist, Källén.

Study supervision: Serenius, Hellgren, Maršál, Stjernqvist, Ådén, Källén.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported by the Swedish Research Council (grants 2006-3858 and 2009-4250), the Uppsala-Örebro Regional Research Council (grant RFR-10324), a grant from the Research Council in the South-East Region of Sweden, and grants to Researchers in the Public Health Care from the Swedish government. Financial support was also provided through a regional agreement between the University of Umeå and the Västerbotten County Council and through a regional agreement on medical training and clinical research (ALF) between the Stockholm County Council and the Karolinska Institute. The study also received support from the Lilla Barnets Fond. Karin Källén received support from the Evy and Gunnar Sandberg Foundation and the Birgit and Håkan Ohlsson Foundation.

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

Group Information: The EXPRESS Group members are as follows: Karel Maršál, MD, PhD (Department of Obstetrics and Gynecology, Lund University, Lund, Sweden), principal investigator of the EXPRESS study; pediatricians Mats Blennow, MD, PhD; Mikael Norman, MD, PhD; Ulrika Ådén, MD, PhD; Brigitte Vollmer, MD, PhD (all from the Department of Women’s and Children’s Health, Karolinska Institutet, Stockholm, Sweden); Fredrik Serenius, MD, PhD (Department of Women’s and Children’s Health, Uppsala University, Uppsala, Sweden, and Department of Pediatrics, Institute of Clinical Sciences, Umeå University, Umeå, Sweden); Uwe Ewald, MD, PhD; Lena Hellström-Westas, MD, PhD; Gunnar Sjörs, MD, PhD; Bo Strömberg, MD, PhD (all from the Department of Women’s and Children’s Health, Uppsala University, Uppsala, Sweden); Vineta Fellman, MD, PhD; Kristina Forsblad, MD (both from the Department of Pediatrics, Clinical Sciences Lund, Lund University, Lund, Sweden); Maria Hafström, MD, PhD; Lennart Stigson, MD (both from the Department of Pediatrics, Institute of Clinical Sciences, Sahlgrenska Academy, Gothenburg University, Gothenburg, Sweden); Ulla Lindskog, MD (Pediatric Clinic, Linköping University Hospital, Linköping, Sweden); Elisabeth Olhager, MD, PhD (Department of Clinical and Experimental Medicine, Linköping University, Linköping, Sweden); Eva Lindberg, MD, PhD; Andreas Ohlin, MD, PhD (both from the Department of Pediatrics, Örebro University, Örebro, Sweden); and Aijaz Farooqi, MD, PhD (Department of Pediatrics, Umeå University, Umeå, Sweden); ophthalmologists Gerd Holmström, MD, PhD (Department of Neuroscience/Ophthalmology, Uppsala University, Uppsala, Sweden); Ann Hellström, MD, PhD (Department of Ophthalmology, Institute of Neuroscience and Physiology, Sahlgrenska Academy Gothenburg University, Gothenburg, Sweden); Peter Jakobsson, MD, PhD (Department of Ophthalmology, Linköping University, Linköping, Sweden); Kristina Tornqvist, MD, PhD (Department of Ophthalmology, Clinical Sciences Lund, Lund University, Lund, Sweden); Gunnar Lindgärde, MD, PhD; Agneta Wallin, MD, PhD (both from the Department of Ophthalmology, Karolinska Institutet, Stockholm, Sweden); Kent Johansson, MD, PhD; and Pia Lundgren, MD, PhD (both from the Department of Ophthalmology, Institute of Clinical Sciences, Umeå University, Umeå, Sweden); psychologists Karin Stjernqvist, PhD; Johanna Månsson, PhD (both from the Department of Psychology, Lund University, Lund, Sweden), Anette Carnemalm, MSc; Christina Helgason, MSc (both from the Department of Pediatrics, Skåne University Hospital, Lund, Sweden); Milly Marken, MSc; Marie Adamsson Johansson, MSc (both from the Pediatric Clinic, Umeå University Hospital, Umeå, Sweden); Ylva Fredriksson, MSC; Ingela Helling, MSc; Sandra Lindblad, MCS; Mia Thörner Vidinghoff, MCS (all from Uppsala University Children’s Hospital, Uppsala, Sweden); Eva Rehn, MSC (Queen Silvia Children’s Hospital, Gothenburg, Sweden); Kari Ylimäinen, MSC (Pediatric Clinic, Örebro University Hospital, Örebro, Sweden); Anna Lönegren, MSc; Margreth Ericsson, MSc; Anna Nyrén, MSc; Carin Johansson Wiemerö, MSc (all from the Pediatric Clinic, Linköping University Hospital, Linköping, Sweden); Birgitta Böhm, PhD (Department of Women’s and Children’s Health, Karolinska Institutet, Stockholm, Sweden); Eva Eklöf, MSC; Christina Lindqvist, MSc; Irmgard Obwexer, MSc; and Claudia Aulin-Villa, MSc (all from the Department of Child and Youth Psychiatry, Stockholm County Council, Stockholm, Sweden). Other experts were Karin Källén, PhD (Centre for Reproductive Epidemiology, Lund University, Lund, Sweden), and Nils-Eric Sahlin MD, PhD (Department of Medical Ethics, Lund University, Lund, Sweden). The nurses and local study coordinators were Barbro Fossmo, RN (Pediatric Clinic, Umeå University Hospital, Umeå, Sweden); Cecilia Ewald, RN (Uppsala University Children’s Hospital, Uppsala, Sweden); Christina Fuxin, RN (Pediatric Clinic, Linköping University Hospital, Linköping, Sweden); Lena Swartling, RN (Department of Neonatology, Karolinska University Hospital, Stockholm, Sweden); Ann-Cathrine Berg, RN (Pediatric Clinic, Skåne University Hospital, Lund, Sweden); and Cecilia Tobiasson, RN (Queen Silvia Children’s Hospital, Gothenburg, Sweden).

Additional Contributions: Technical assistance with data collection was provided by Ms Grozda Pajic, Lund University. We thank Marius Kublickas, MD, PhD, Vytas Marcincas, MSc (both from MedSciNet AB, Stockholm, Sweden), for the design and maintenance of the study database. The English language of the manuscript was revised by Sue Pajuluoma, BSc. Compensation was received from the sponsors for such contributions.