aMother had protected identity (n = 3), families moved abroad (n = 3), preliminary identity number given at birth did not match (n = 24).
The diagonal line indicates the mean of the controls and the vertical bars represent the 99% CIs of the mean values. The regression lines with 99% CIs for respective scores of children in the preterm group are based on the equations: cognitive score = 83.12 + (GA-21) × 2.517, P < .001; language score = 82.78 + (GA-21) × 3.551, P < .001; and motor score = 83.24 + (GA-21) × 2.523, P = .001. GA indicates gestational age in completed weeks.
Serenius F, Källén K, Blennow M, et al. Neurodevelopmental outcome in extremely preterm infants at 2.5 years after active perinatal care in Sweden. JAMA. 2013;309(17):doi:10.1001/jama.2013.3786.
eTable 1. Demographic and Neonatal Characteristics of the Entire EXPRESS 1-Year-Survival Cohort of Extremely Preterm Children (<27 Weeks) by Type of Follow-up at 2.5 Years Corrected Age
eTable 2. Outcome in 41 Extremely Preterm (<27 Weeks) Assessed by Chart Review Compared With Extremely Preterm Children Tested With Bayley-lll and Medical Examination
eTable 3. Relationship Between Gestational Age and Bayley III-Scores Among Infants Born Extremely Preterm (<27 Weeks)
eTable 4. Mean Bayley-lll Scores and Distribution of Overall Disability Categories in Children Born Extremely Preterm (<27 Weeks) by Gender
eTable 5. Bayley-lll Scores (Composite Cognitive, Language, and Motor) in Children Born Extremely Preterm (<27 Weeks) in Multiple Births Compared With Extremely Preterm Singletons
eTable 6. Bayley-lll Scores (Composite Cognitive, Language, and Motor) in Children Born Extremely Preterm (<27 Weeks), by Presence of Congenital Malformations
eTable 7. Bayley-lll Scores (Composite Cognitive, Language, and Motor) in Children Born Extremely Preterm (<27 Weeks) With and Without Cerebral Palsy
eTable 8. Distribution of Overall Disability* Categories in Children Born Extremely Preterm (<27 Weeks) by Gender
eTable 9. Overall Disability in Multiples Compared With Singletons Born Extremely Preterm (<27 Weeks)
Customize your JAMA Network experience by selecting one or more topics from the list below.
Serenius F, Källén K, Blennow M, Ewald U, Fellman V, Holmström G, Lindberg E, Lundqvist P, Maršál K, Norman M, Olhager E, Stigson L, Stjernqvist K, Vollmer B, Strömberg B, EXPRESS Group FT. Neurodevelopmental Outcome in Extremely Preterm Infants at 2.5 Years After Active Perinatal Care in Sweden. JAMA. 2013;309(17):1810–1820. doi:10.1001/jama.2013.3786
Importance Active perinatal care increases survival of extremely preterm infants; however, improved survival might be associated with increased disability among survivors.
Objective To determine neurodevelopmental outcome in extremely preterm children at 2.5 years (corrected age).
Design, Setting, and Participants Population-based prospective cohort of consecutive extremely preterm infants born before 27 weeks of gestation in Sweden between 2004 and 2007. Of 707 live-born infants, 491 (69%) survived to 2.5 years. Survivors were assessed and compared with singleton control infants who were born at term and matched by sex, ethnicity, and municipality. Assessments ended in February 2010 and comparison estimates were adjusted for demographic differences.
Main Outcomes and Measures Cognitive, language, and motor development was assessed with Bayley Scales of Infant and Toddler Development (3rd edition; Bayley-lll), which are standardized to mean (SD) scores of 100 (15). Clinical examination and parental questionnaires were used for diagnosis of cerebral palsy and visual and hearing impairments. Assessments were made by week of gestational age.
Results At a median age of 30.5 months (corrected), 456 of 491 (94%) extremely preterm children were evaluated (41 by chart review only). For controls, 701 had information on health status and 366 had Bayley-lll assessments. Mean (SD) composite Bayley-III scores (cognition, 94 [12.3]; language, 98 [16.5]; motor, 94 [15.9]) were lower than the corresponding mean scores for controls (cognition, 104 [10.6]; P < .001; adjusted difference in mean scores, 9.2 [99% CI, 6.9-11.5]; language, 109 [12.3]; P < .001; adjusted difference in mean scores, 9.3 [99% Cl, 6.4-12.3]; and motor, 107 [13.7]; P < .001; adjusted difference in mean scores, 12.6 [99% Cl, 9.5-15.6]). Cognitive disability was moderate in 5% of the extremely preterm group vs 0.3% in controls (P < .001) and it was severe in 6.3% of the extremely preterm group vs 0.3% in controls (P < .001). Language disability was moderate in 9.4% of the extremely preterm group vs 2.5% in controls (P < .001) and severe in 6.6% of the extremely preterm group vs 0% in controls (P < .001). Other comparisons between the extremely preterm group vs controls were for cerebral palsy (7.0% vs 0.1%; P < .001), for blindness (0.9% vs 0%; P = .02), and for hearing impairment (moderate and severe, 0.9% vs 0%; P = .02, respectively). Overall, 42% (99% CI, 36%-48%) of extremely preterm children had no disability, 31% (99% CI, 25%-36%) had mild disability, 16% (99% CI, 12%-21%) had moderate disability, and 11% (99% CI, 7.2%-15%) had severe disability. Moderate or severe overall disability decreased with gestational age at birth (22 weeks, 60%; 23 weeks, 51%; 24 weeks, 34%; 25 weeks, 27%; and 26 weeks, 17%; P for trend < .001).
Conclusions and Relevance Of children born extremely preterm and receiving active perinatal care, 73% had mild or no disability and neurodevelopmental outcome improved with each week of gestational age. These results are relevant for clinicians counseling families facing extremely preterm birth.
A proactive approach to resuscitation and intensive care of extremely preterm infants (<27 gestational weeks) has increased survival and lowered the gestational age of viability.1- 4 There are concerns that increased survival may come at the cost of later neurodevelopmental disability among survivors. Approximately 25% of extremely preterm infants born in the 1990s had a major disability at preschool age, such as impaired mental development, cerebral palsy (CP), blindness, or deafness.5,6 More recent studies report decreasing,7,8 unchanged,2 or increasing rates of neurodevelopmental disability9- 11 at preschool age compared with previous decades. The most immature infants, ie, those born before 25 weeks of gestation, continue to have high (>50%) impairment rates.12
Few outcome studies of extremely preterm infants1- 3,6,8,10,13- 15 are population based. Some may not represent recent advances in perinatal care and some are defined by birth weight, resulting in a selection bias toward growth-restricted infants. The Extremely Preterm Infants Study in Sweden (EXPRESS) is a national population-based prospective study of all infants born alive or stillborn before 27 weeks of gestation between 2004 and 2007. The high rate of infant survival in this cohort (70%) was attributed to active perinatal care.3 Active perinatal care in Sweden includes easy and free access to specialist perinatal care, centralization of extremely preterm births to level lll hospitals, a low threshold to provide life support at birth, and near universal admission of infants born at 23 to 26 weeks for neonatal intensive care.3
The aim of this prospective follow-up study was to assess neurological and developmental outcome of the EXPRESS cohort at 2.5 years corrected age in comparison with a matched control group born at term. Our primary hypotheses were that preterm infants would have poorer neurodevelopmental outcome than term infants and that outcomes would improve with increasing gestational age within the preterm group. Our secondary hypothesis was that the outcome in preterm boys would be worse than in girls and the outcome in multiple births poorer than in singleton births.
The Extremely Preterm Group. Between April 1, 2004, and March 31, 2007, research coordinators (2 from each of the 7 Swedish health regions) collected perinatal and neonatal data on all 1011 infants born before 27 completed gestational weeks.3 Gestational age was based on ultrasound dating in 95% of the pregnancies. Live birth was defined as an infant with any signs of life at birth.3 Of 707 live-born infants, 497 (70%) survived to 1 year of age, of whom 6 children died before 2.5 years corrected age. Thus, the study cohort comprised 491 children; no exclusions were made. Follow-up ended in February 2010.
The Control Group. Each preterm child was matched with 2 children in the control group, 1 for interview and Bayley-lll assessment (Bayley Scales of Infant and Toddler Development, 3rd Edition16) and 1 for interview only. A pool of 10 control participants for each preterm child was randomly selected from the Swedish Medical Birth Registry. Selection criteria were singleton at-term birth with a 5-minute Apgar score greater than 3 with matching of control participants for place of living, sex, day of birth, and maternal country of birth. If it was not possible to find 10 children who met all selection criteria, additional children were randomly selected satisfying fewer matching criteria. The 10 matched control participants were ordered according to matching quality and the parents of the first 2 children on the list were approached. If participation was declined, the second set of parents was approached and if necessary, the additional sets until 2 control participants agreed to participate or the pool of 10 children was depleted.
Clinical Assessments. At 2.5 years of corrected age (2.5 years of chronological age for control participants), certified psychologists assessed cognitive, language, and motor development with the Bayley Scales of Infant and Toddler Development (Bayley-lll). One author (K.S.) trained the psychologists and standardized their performance. Bayley-lll scales are standardized to a mean (SD) score of 100 (15). Because the Bayley-lll has not been standardized in Sweden, test scores of those in the preterm group were related to the mean (SD) of the control group's scores. In accordance with other studies,17- 19 we regarded a mean group difference greater than 5 points (equal to half a standard deviation of our control group) as clinically important.
Specialists in pediatric neurology or pediatrics examined the children in the preterm group, focusing on neuromotor function. In the same visit, parents of these children were interviewed using a structured questionnaire that included questions regarding parental education, child health, and child development. Control participants were not examined but nurses interviewed their parents by telephone or the parents completed the same questionnaire (sent by mail) used for parents of the preterm children. Maternal and paternal educational levels were assessed by SUN-codes20 representing the highest educational achievement. The codes were transferred to a quantitative scale representing years of education completed (1, ≤9 years of compulsory school; 2, 10-11 years; 3, 12-13 years; 4, 14-15 years; 5, 16 years; and 6, ≥17 years). Pediatric ophthalmologists assessed visual acuity in the preterm children. Psychologists and clinicians were not blinded to the group status because the examination of preterm children was part of a clinical follow-up and the examiners were familiar with most of the children.
Malformations were classified according to the International Classification of Diseases, Tenth Revision, but dislocation of the hip (Q65.0-Q65.5), preauricular tags (Q17.0), undescended testes (Q53.0-53.9), and patent ductus arteriosus (Q25.0) were not classified as malformations.
Forty-one children in the preterm group were assessed through chart review. Information from local pediatricians, low-vision centers, and rehabilitation centers provided information that we regarded as sufficient to allow assessment of developmental and neurosensory outcome.
The Regional Ethics Review Board, Lund University, Lund, Sweden, approved the study. The parents provided written consent for examination of their children and for data acquisition.
Children in the preterm group were compared with control participants for cognition, language, and motor function according to the Bayley-lll scales; mental developmental delay assessed by Bayley-lll; cerebral palsy (CP), visual and hearing disability, and a composite outcome of overall disabilities as described further in this section. Cognitive, motor, and language development were considered normal if the composite score on the respective Bayley-lll scale was ≥ mean-1SD, mildly impaired if the score was < mean-1SD and ≥ mean-2SD, moderately impaired if the score was < mean-2SD and ≥ mean-3SD, and severely impaired if the score was < mean-3SD. Mental developmental delay was included as an outcome to allow comparisons with studies reporting the mental developmental index (MDI) of the Bayley-ll scale.21 Mild mental developmental delay was defined as a score < mean-1SD and ≥ mean-2SD on either the cognitive or the language composite scale, moderate mental developmental delay as < mean-2SD and ≥ mean-3SD on either scale, and severe mental developmental delay was < mean-3SD on either scale.
Definition of CP was defined according to Bax et al22 and characterized as hemiplegic, diplegic, tetraplegic, ataxic, or dyskinetic. The severity of CP was classified as mild in children who were able to walk without an aid, moderate in children able to walk with an aid, and severe in children who were unable to walk even with an aid.
Children unable to fixate and follow a light with either eye were considered bilaterally blind. Children registered at low-vision centers without blindness were recorded as having moderate visual impairment. Moderate auditory impairment was defined as hearing loss corrected with a hearing aid and severe impairment as hearing loss that could not be corrected with a hearing aid (deafness).
The overall outcome was characterized as no disability, mild disability, or moderate and severe disability. Severe disability was defined as any of the following: Bayley-lll composite cognitive, language or motor score < mean-3SD, severe CP, or bilateral blindness or deafness. Moderate disability was defined as scores between −2 and −3 standard deviations from the mean of any of the Bayley-111 scales, moderate CP, and moderate visual or hearing impairment. Mild disability was defined as scores between −1 and −2 standard deviations from the mean of any of the Bayley-lll scales or mild CP. The combined category of moderate and severe disability corresponds to the concept of moderate and severe neurodevelopmental impairment in North American7,12,23- 25 and Australian8 studies.
Subgroup analyses were performed based on week of gestation, sex, and multiple births for preterm children.
Differences in the demographic characteristics between the preterm and control groups were evaluated with Fisher exact test, or with χ2 test as specified. The prevalence of neurosensory disability between the preterm group and the control group was compared with Fisher exact test. Group differences in the mean Bayley-lll composite cognitive, language, and motor scores were assessed by 1-way analysis of variance. When specified, analysis of covariance was performed when, besides Bayley scores, information on parental education (quantitative, scaled 1-6) and maternal country of birth (Nordic vs non-Nordic country) was entered into the models. Inclusion of other potential confounders (maternal age, parity, and smoking) did not change the estimate of the difference between case and control participants nor did it improve the proportion of variance. Therefore, these factors were not entered into the final models. Within the preterm group, differences in sex and between singleton and multiple births in mean Bayley-III scores were assessed by analysis of covariance adjusting for gestational age; simple linear regression analyses were used to examine the relationship between gestational age and each Bayley-lll composite score within the preterm group. Sensitivity analyses were performed using the Cook distance methods to identify outliers. Due to the number of incomplete pairings between preterm and control participants, it was not possible to perform matched analyses.
Crude odds ratios (ORs) with exact confidence intervals (CIs)26 were computed to compare the risk for the disability categories between groups. When the OR was indeterminable, the P value from Fisher exact test was computed. A 2-tailed P value of less than .01 was regarded as statistically significant. For the outcome categories, ORs were estimated using multiple logistic regression analysis adjusting for parental variables (maternal country of birth: Nordic/non-Nordic, and maternal and paternal educational level scaled 1-6), or, as specified, in comparisons within the preterm group, adjusting for gestational age. Missing information on maternal or paternal educational level was replaced by the overall mean. Due to multiple comparisons, the overall level of significance was set to 1%. With α = .01, we had 80% power to detect a true difference of 0.25 standard deviations between the mean Bayley scores among preterm and control participants. The smallest detectable OR (with α = .01 and β = .8) for severe or moderate overall disability, preterm children vs controls, was 2.6 under the assumption that this outcome occurred in 3% of control participants. Statistical analyses were made using Gauss statistical software version 10 (Gauss, Aptech Systems Inc).
Of 707 live-born preterm infants, 497 (69%) survived to 2.5 years corrected age. Of these, 30 were not eligible (Figure 1), 6 died between 1 year and follow-up, and 5 declined participation; 415 were formally assessed with a clinical examination and 399 were assessed with the Bayley-lll scale. Baseline charactistics of the extremely preterm children, the children in the control group, and their parents are shown in Table 1 and Table 2. Forty-one children were assessed by chart review and outcome data of these children are included in Table 3 (section on mental developmental delay), Table 4, Table 5, and Table 6, and shown separately in
eTable 2). Outcome data at 2.5 years of age were thus available for 456 (94%) of 491 preterm children (Figure 1).
More mothers of chart-reviewed children were younger than 20 years compared with mothers of formally assessed children (15% vs 1.9%; P < .001). Other demographic data including neonatal morbidity were
similar (eTable 1). The 5 children for whom participation was declined were all born to non-Nordic mothers and did not have severe intraventricular hemorrhage, retinopathy of prematurity, or bronchopulmonary dysplasia, as defined previously.29 Of 922 children in the control group, 701 had parental information on their health status, and of 461 children in the control group who were selected for Bayley-lll assessment, 366 were assessed (Figure 1).
In the preterm group, parental education was lower and more mothers were of non-Nordic origin than in the control group (Table 1; Table 2). The median corrected age at assessment was 30.5 months for preterm children and the median chronological age was 30.9 months for the control participants. Assessment within the planned age window of 30 ± 3 months was completed for 96% of participants in the preterm group and 91% in the control group.
Of 415 preterm children seen at follow-up, 399 completed the Bayley-111 cognitive scale, 393 the language scale, and 382 the motor scale (Table 3). In the control group, 366 children were assessed with Bayley-lll. Reasons for incomplete examinations of children in the preterm and control groups are shown in Table 3.
The mean (SD) composite cognitive score for children in the preterm group compared with those in the control group was 94 (12) vs 104 (11) (P < .001), the mean composite language score was 98 (17) vs 109 (12) (P < .001), and the mean composite motor score was 94 (16) vs 107 (14) (P < .001), respectively (Table 3). When adjusted for maternal country of birth and parental education, the mean difference in language scores was moderately attenuated. For the cognitive, language, and motor scores, the proportions of variance explained (R2) in the full models were 0.15, 0.18, and 0.16, respectively.
Normal cognitive development or mild cognitive disability was found in 88.8% (354) of children in the preterm group and in 99.5% (364) of control participants. Moderate cognitive disability was present in 5.0% (20) of children in the preterm group vs 0.3% (1) in the control group (P < .001), and severe cognitive disability was present in 6.3% (25) in the preterm group vs 0.3% (1) in the control group (P < .001). Normal language development or mild language disability was found in 83.9% (330) of children in the preterm group vs 97.5% (351) in the control group (all group comparisons, P < .001) (Table 3). Normal motor development or mild motor disability was found in 84.8% (324) of children in the preterm group and in 98.6% (348) of control participants. Moderate or severe mental developmental delay was seen in 20% (88) of children in the preterm group vs 2.8% (10) in the control group (P < .001).
Mean Bayley Composite Scores by Gestational Age.Within the preterm group, the Bayley-lll scores increased with advancing gestational age at birth by 2.5 points (99% CI, 1.0-4.0) per week for the cognitive scores (P < .001), by 3.6 points (99% CI, 1.6-5.6) per week for the language scores (P < .001), and by 2.5 points (99% CI 0.5-4.5) per week for the motor scores (P = .001) (Figure 2). Adjustment for infant sex altered the results only marginally (
The mean language score was 5 points lower in preterm boys compared with preterm girls (P value for sex comparisons <.001;
eTable 4) but there were no differences on cognitive or motor scores. There were no differences in the performance on the 3 Bayley-lll scales between singleton and multiple births (eTable 5), or between infants with or without congenital malformations (eTable 6).
Cerebral palsy was present in 32 children in the preterm group (7.0%; 99% CI, 3.9%-10.1%). Of these, 13 children had mild CP, 13 had moderate CP, and 6 had severe CP. Only 1 child in the control group had CP, which was severe (Table 4). Twenty of the children with CP in the preterm group had diplegia, 5 had hemiplegia, 3 had quadriplegia, and 4 had CP of other types (3 dyskinetic and 1 CP of unknown type). Sensory disabilities in preterm CP children were rare; 1 child was blind and deaf.
Preterm children with CP and tested with Bayley-lll (n = 26) had significantly lower mean composite scores in all 3 scales than children without CP (gestational age adjusted mean difference for cognition, 11 points lower (99% CI, 5-17), P < .001; for language, 9 points lower (99% CI, 0-17), P = .01; and for motor, 22 points lower (99% CI,
14-30), P < .001) (eTable 7).
In the whole preterm cohort, 4 children (0.9%; 99% CI, 0.1-2.4) were blind and 1 child was deaf compared with no blind or deaf children among the control participants.
The overall rate of disabilities (Table 5) includes performance on the Bayley-lll assessments, mental developmental delay, CP, and visual and hearing disabilities. Of the 456 preterm children assessed at 2.5 years corrected age (Table 5), 42.1% (192) were classified as normal (vs 78.1%  of control participants), 30.7% (140) as having had mild disabilities (vs 18.6%  of control participants), and 27.2% (124) as having moderate or severe disabilities (vs 3.3%  of control participants) (P < .001 for all comparisons between preterm and control participants).
In preterm children with severe overall disability (n = 50), severe mental developmental delay was the predominant disability found in 78% (39/50) of the children. Severe CP was found in 12% (6/50 children), blindness in 8.2% (4/50 children), and deafness in 1 child.
Overall Disabilities by Gestational Age.The proportion of children with mild or no disabilities increased from 40% at 22 weeks to 83% at 26 weeks (P < .001 for trend; Table 6). Moderate or severe disabilities decreased with advancing gestational age at birth (22 weeks, 60%; 23 weeks, 51%; 24 weeks, 34%; 25 weeks, 27% ; 26 weeks, 17% ; P < .001 for trend). For moderate disabilities, the OR for 1 week increase compared with no disability was 0.55 (99% CI, 0.39-0.79), and for severe disabilities 0.58 (99% CI, 0.39-0.86) (both P < .001).
Moderate or severe disabilities were found in 31% (77/248) of preterm boys and in 23% (47/208) of preterm girls (P = .02 adjusted for gestational age; eTable 8). There were no differences in the overall outcomes between singleton and multiple births (eTable 9).
In EXPRESS, 70% of live-born infants survived infancy,3 which is higher than reported in comparable studies.1,2,6,8,14,15 At follow-up, 42% of preterm children who survived to 2.5 years had no disability compared with 78% of control participants. For most preterm children with disabilities, disabilities were mild (31%). In contrast to studies reporting outcome unrelated to gestational age,6,13,14 we found an increase in moderate or severe disabilities with decreasing gestational age. In contrast to other studies,6,30,31 the difference in overall outcome between preterm boys and girls was not statistically significant, and we did not find differences in preterm outcomes between singleton and multiple births.
Until recently, Bayley-ll21 has been the standard tool for follow-up studies, with moderate or severe mental developmental delay being defined as a mental developmental index (MDI) of less than 70 (<mean-2 SD). Bayley-lll subdivides the Bayley-ll components for MDI into cognition and language. To be able to compare the results from Bayley-lll and Bayley-ll, the lowest score for cognition and language from the Bayley-lll scales is regarded as an estimate of MDI.8,32
Cognitive and language composite scores in preterm participants were 0.9 standard deviations lower than in control participants, and 20% had moderate or severe mental developmental delay. In a recent Australian study8 (gestational age <28 weeks), mental developmental delay was assessed at 24 months with the Bayley-lll scale and defined as in our study. The results were similar to ours with moderate or severe mental delay being found in 16%. In other studies of children born extremely preterm6,7,14 or with extremely low birth weight (<1000 g),23,24 mental developmental delay ranged from 22% to to 37%. In these studies, however, the Bayley-II Scales were used and comparisons have to be made with caution.
Results for children in the preterm group were related to those of a control group born at term that was found to have mean scores greater than 100 for all 3 Bayley-lll scales. If the performance of preterm children had been compared with published norms, the combined prevalence of moderate and severe disabilities would have been lower ( mental developmental delay: 7% vs 20%; cognitive: 5.2% vs 11%; language: 6.1% vs 16%; motor: 7.1% vs 15%). Without a study-specific control group, Bayley-lll assessments may underestimate developmental delay.33- 35
Preterm children from socially disadvantaged families may have poorer neurodevelopment.25,36,37 Despite matching, the final frequencies of low parental education and non-Nordic origin of mothers were higher in preterm infants than in control participants. After adjustment for education and maternal country of birth, the differences in mean cognitive and motor scores changed only marginally, whereas the effect on language was slightly larger, presumably because more preterm children did not have Swedish as their first language.
The prevalence of CP (7%) was lower than in other studies (9.8 to 28%).6,8,10,13,14 Almost half had mild CP with excellent prognosis for independent mobility, almost half had moderate CP with good prospects for ambulation with aid, and only a small percentage had severe CP (1.3% of all preterm children). The prevalence of blindness (0.9%) and deafness (0.2%) in the total cohort was lower than in most other studies.5- 7,9,10,13,14,24
The combined categories of moderate and severe overall disability (27%) corresponded with moderate or severe neurodevelopmental impairments in North American7,12,23- 25 and Australian9 studies. The rates were 34% and 35%, respectively, in 2 North American studies comprising extremely low-birth weight infants,23,24 and 20% in the Australian study8 of infants born at less than 28 weeks gestation, but children in these studies were more mature at birth than in our study. However, a recent UK study38 of children born at less than 27 weeks gestation and assessed at 3 years of age with Bayley-lll is comparable to our study and showed similar results; 25% had moderate or severe disability. The survival rate at 22 to 24 weeks was much lower in the UK study than in our study and the Bayley-lll results were not related to a term control group.
In our study, the less favorable outcome at 22 weeks compared with more mature infants should be viewed with caution because there were few survivors (n = 5). Thus the question of active perinatal management of these infants remains unanswered.
We attributed the high infant survival of the EXPRESS cohort to proactive perinatal care.3 It could be speculated that some of the interventions, by avoiding physiological destabilization around birth, not only increased the chances for survival but also improved the chances to survive intact or with lesser degrees of disability.39 Besides a high degree of centralization and active perinatal care, Swedish neonatal care widely utilizes family-centered developmental care and universally uses breast milk for infant nutrition. Breast milk has been shown to have a beneficial effect on cognition at preschool age17 and in our study, half of the infants were breastfed at discharge home.
The results also need to be interpreted in the light of the characteristics of the Swedish society and health care system. Sweden is a rather uniform society without extreme poverty. Antenatal care is easily accessible and utilized by close to 100% of mothers. All citizens are covered by general health insurance including a total of 480 days of parental leave after the child's birth and additional benefits for severely sick children. We believe that our results can be generalized to countries with universal access to health care and active perinatal programs.
The strengths of this study include prospective enrollment of all extremely preterm births in Sweden. Specialists in pediatrics, pediatric neurology, and ophthalmology performed the follow-up according to standardized protocols, and certified psychologists administered the Bayley-lll tests. The follow-up rate was high and a contemporary control group of children born at term was included. Children with malformations were not excluded to ensure a complete picture of the outcome. Congenital malformations are associated with neurodevelopmental disability,40 but we found no differences in the main outcome, the performance on the Bayley-lll scales, between children with and without malformations.
A potential weakness is the early age of assessment because later academic, psychological, and behavioral problems are likely41 and continued follow-up is warranted. Another possible limitation is reliance on parental questionnaires for collecting neurosensory and demographic data for the control group. However, virtually all children in Sweden attend well baby clinics and parents are informed about conditions that affect their child's health. An additional limitation is that some preterm children were not formally assessed. However, they were examined by local pediatricians and 2 of the authors (F.S. and B.S.) scrutinized the data. In some cases, the full number of control participants could not be recruited; it is unknown to what extent this influenced the characteristics of the control group. Moreover, it was not possible to blind the assessors to group status, which might have introduced expectation bias.
Improved survival did not translate into increasing disability rates, and we like others8 believe that the neurodevelopmental outcome for extremely preterm children born in the 2000s will be better than for those born in the 1990s. Nevertheless, the impact of prematurity on neurodevelopmental outcome was large, which calls for further improvements in neonatal care, such as better control of infection42 and postnatal nutrition.43 Postdischarge developmental intervention programs,44 currently not in place in Sweden, might improve outcome.
In conclusion, extremely preterm children had poorer neurodevelopmental outcome than children born at term, but 73% had no or mild disability and neurodevelopmental outcome improved with each week of gestational age. The outcome was similar or better than in comparable studies with lower survival rates. These results are relevant for clinicians counseling families facing extremely preterm birth.
Corresponding Author: Fredrik Serenius, MD, PhD, Department of Women's and Children's Health, Section for Pediatrics, Uppsala University, S-751 85 Uppsala, Sweden (firstname.lastname@example.org).
Author Contributions: Dr Serenius had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Serenius, Källen, Blennow, Ewald, Fellman, Holmström, Lindberg, Marsál, Norman, Olhager, Stjernqvist, Vollmer, Strömberg.
Acquisition of data: Serenius, Källen, Blennow, Ewald, Fellman, Holmström, Lindberg, Lundqvist, Norman, Olhager, Stigson, Stjernqvist, Vollmer, Strömberg.
Analysis and interpretation of data: Serenius, Källen, Blennow, Ewald, Fellman, Holmström, Lindberg, Lundqvist, Marsál, Norman, Olhager, Stjernqvist, Strömberg.
Drafting of the manuscript: Serenius, Källen, Blennow, Holmström, Marsál, Norman, Stjernqvist, Vollmer, Strömberg.
Critical revision of the manuscript for important intellectual content: Serenius, Källen, Blennow, Ewald, Fellman, Holmström, Lindberg, Lundqvist, Marsál, Norman, Olhager, Stjernqvist, Stigson, Vollmer, Strömberg.
Statistical analysis: Källen.
Obtained funding: Serenius, Blennow, Ewald, Fellman, Lindberg, Norman, Olhager, Vollmer, Strömberg.
Administrative, technical, or material support: Serenius, Källen, Blennow, Fellman, Holmström, Lindberg, Lundqvist, Marsál, Stigson, Stjernqvist.
Study supervision: Källen, Blennow, Marsál, Stjernqvist.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.
EXPRESS Group Members: The following researchers and clinicians contributed to the 2.5-year follow-up of children belonging to the EXPRESS cohort and of children in the control group. Main coordinator of the follow-up study, Dr Strömberg; and chief psychologist of the follow-up study, Dr Stjernqvist. Pediatricians: Drs Blennow, Norman, Vollmer, Serenius, Ewald, Strömberg, Fellman, Stigson, Olhager, and Lindberg); Department of Women's and Children's Health, Uppsala University, Uppsala (Lena Hellström-Westas, MD, PhD, and Gunnar Sjörs, MD, PhD); Department of Pediatrics, Clinical Sciences Lund, Lund University, Lund (Kristina Forsblad, MD); Department of Pediatrics, Institute of Clinical Sciences, Sahlgrenska Academy, Gothenburg University, Gothenburg (Maria Hafström, MD, PhD); Pediatric Clinic, Linköping University Hospital, Linköping (Ulla Lindskog, MD). Ophthalmologists: Dr Holmström; Department of Neuroscience/Ophthalmology, Uppsala University, Uppsala (Dordi Austeng, MD, PhD); Department of Ophthalmology, Institute of Neuroscience and Physiology, Sahlgrenska Academy Gothenburg University, Gothenburg (Ann Hellström, MD, PhD); Department of Ophthalmology, Linköping University, Linköping (Peter Jakobsson, MD, PhD); Department of Ophthalmology, Clinical Sciences Lund, Lund University, Lund (Kristina Tornqvist, MD, PhD); Department of Ophthalmology, Karolinska Institutet, Stockholm (Gunnar Lindgärde, MD, PhD, and Agneta Wallin, MD, PhD); and Department of Ophthalmology, Institute of Clinical Sciences, Umeå University, Umeå (Kent Johansson, MD, PhD). Psychologists: Dr Stjernqvist; Department of Psychology, Lund University, Lund (Johanna Månsson, MSc); Department of Pediatrics, Skåne University Hospital, Lund (Anette Carnemalm, MSc, and Christina Helgason, MSc); Pediatric Clinic, Umeå University Hospital, Umeå (Milly Marken, MSc); Uppsala University Children's Hospital, Uppsala (Ylva Fredriksson, MSc, and Ingela Helling, MSc); Queen Silvia Children's Hospital, Gothenburg (Eva Rehn, MSc); Pediatric Clinic, Örebro University Hospital, Örebro (Kari Ylimäinen, MSc); Pediatric Clinic, Linköping University Hospital, Linköping (Anna Lönegren, MSc, Margreth Ericsson, MSc, Anna Nyrén, MSc, and Carin Johansson Wiemerö, MSc); Department of Women's and Children's Health, Karolinska Institutet, Stockholm (Birgitta Böhm, MSc, PhD); and Department of Child and Youth Psychiatry, Stockholm County Council, Stockholm (Eva Eklöf, MSc, Christina Lindqvist, MSc, Irmgard Obwexer, MSc, and Claudia Aulin-Villa, MSc). Nurses, local study coordinators: Pediatric Clinic, Umeå University Hospital, Umeå (Barbro Fossmo, RN); Uppsala University Children's Hospital, Uppsala (Cecilia Ewald, RN); Pediatric Clinic, Linköping University Hospital, Linköping (Christina Fuxin, RN); Department of Neonatology, Karolinska University Hospital, Stockholm (Lena Swartling, RN); and Pediatric Clinic, Skåne University Hospital, Lund (Ann-Cathrine Berg, RN). Telephone interviews: Ms Pia Lundquist. Technical assistance in data collection: Department of Obstetrics and Gynecology, Clinical Sciences Lund, Lund University, Lund (Grozda Pajic).
Funding/Support: This study was supported by the Swedish Research Council grants 2006-3855 and 2009-4250, the Uppsala-Örebro Regional Research Council grant RFR-10324, a grant from the Research Council 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 Västerbotten County Council and through a regional agreement on medical training and clinical research (ALF) between Stockholm County Council and Karolinska Institute. The study also received support from The “Lilla Barnets Fond” Children's fund. Dr Källén reported support from the Evy and Gunnar Sandberg and from the Birgit and Håkan Ohlsson Foundations and Dr Vollmer was supported by the Marie Curie Individual Intra-European Fellowship within the EU FP6 Framework Program.
Role of the Sponsor: None of the sponsors had any role in the design of the study, analysis and interpretation of the data; or in the preparation, review, or approval of the manuscript.
Additional Contributions: We thank the EXPRESS Group members, who contributed to the study design and collection of perinatal EXPRESS data: obstetricians (Tomas Fritz, Per Åke Holmgren, Annika Jeppson, Solveig Nordén Lindeberg, Marija Simic, Magnus Westgren, Ingrid Östlund and the late Margareta Wennergren), experts in medical ethics (Anita Lundqvist, Tore Nilstun, Nils-Eric Sahlin), perinatal pathology (the late Ricardo Laurini), epidemiology (Petra Otterblad Olausson). A full list of their affiliations has been published (JAMA. 2009; 301(21):2225-2233). No compensation was received by any members of the EXPRESS Group. Stellan Håkansson, MD, PhD and Marius Kublickas, MD, PhD, are gratefully acknowledged for help in building the study database, and Hugo Lagercrantz, MD, PhD for support, advice and constructive criticism. No compensation was received by Drs Håkansson, Kublickas, or Lagercrantz. The English language of the manuscript was revised by Sue Pajuluoma, BSc, MSc. Ms Pajuluoma received small reimbursement in association with her contribution to this article.