Neurologic and Developmental Disability

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/8096383
Neurologic and Developmental Disability at Six Years of Age after Extremely

Preterm Birth
Article in New England Journal of Medicine · February 2005

DOI: 10.1056/NEJMoa041367 · Source: PubMed
CITATIONS READS
1,686 1,426
4 authors, including:
Neil Marlow Dieter Wolke

University College London The University of Warwick
573 PUBLICATIONS 28,504 CITATIONS 544 PUBLICATIONS 26,009 CITATIONS
SEE PROFILE SEE PROFILE
Muthanna Samara
Kingston University London
62 PUBLICATIONS 3,991 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Promoting academic resilience in preterm children with adaptive computerized training - a multi-center randomised controlled trial View project
SNAP Trial View project
All content following this page was uploaded by Neil Marlow on 31 May 2014.
The user has requested enhancement of the downloaded file.

new england
The
journal of medicine
established in 1812 january 6 , 2005 vol. 352 no. 1
Neurologic and Developmental Disability

at Six Years of Age after Extremely Preterm Birth
Neil Marlow, D.M., Dieter Wolke, Ph.D., Melanie A. Bracewell, M.D.,
and Muthanna Samara, M.Sc., for the EPICure Study Group*
abstract
background
Birth before 26 weeks of gestation is associated with a high prevalence of neurologic From the School of Human Development,
and developmental disabilities in the infant during the first two years of life. University of Nottingham, Nottingham
(N.M., M.A.B.); and the Unit of Perinatal
and Pediatric Epidemiology, Department of
methods Community-based Medicine, University of
We studied at the time of early school age children who had been born at 25 or fewer Bristol, Bristol, and the Department of Psy-
chology, University of Hertfordshire, Hatfield
completed weeks of gestation in the United Kingdom and Ireland in 1995. Each child (D.W., M.S.) — all in the United Kingdom.
had been evaluated at 30 months of age. The children underwent standardized cogni- Address reprint requests to Dr. Marlow at
tive and neurologic assessments at six years of age. Disability was defined as severe (in- the Academic Division of Child Health, Lev-
el E East Block, Queens Medical Centre,
dicating dependence on caregivers), moderate, or mild according to predetermined Nottingham NG7 2UH, United Kingdom,
criteria. or at [email protected].
Drs. Marlow and Wolke contributed equally

results to this article.
Of 308 surviving children, 241 (78 percent) were assessed at a median age of six years
and four months; 160 classmates delivered at full term served as a comparison group. *The investigators and study coordinators
are listed in the Appendix.
Although the use of test reference norms showed that cognitive impairment (defined
as results more than 2 SD below the mean) was present in 21 percent of the children N Engl J Med 2005;352:9-19.
born extremely preterm (as compared with 1 percent in the standardized data), this val- Copyright © 2005 Massachusetts Medical Society.
ue rose to 41 percent when the results were compared with those for their classmates.
The rates of severe, moderate, and mild disability were 22 percent, 24 percent, and 34
percent, respectively; disabling cerebral palsy was present in 30 children (12 percent).
Among children with severe disability at 30 months of age, 86 percent still had moder-
ate-to-severe disability at 6 years of age. In contrast, other disabilities at the age of 30
months were poorly predictive of developmental problems at 6 years of age.
conclusions
Among extremely preterm children, cognitive and neurologic impairment is common
at school age. A comparison with their classroom peers indicates a level of impairment
that is greater than is recognized with the use of standardized norms.
n engl j med 352;1 www.nejm.org january 6, 2005 9
Downloaded from www.nejm.org at WASHINGTON UNIV SCH MED MEDICAL LIB on February 22, 2006 .
Copyright © 2005 Massachusetts Medical Society. All rights reserved.
The new england journal of medicine
er of the class identify three children of the same
a n increased prevalence of cog-

nitive impairment and poorer education-
al achievement has been repeatedly ob-
served among school-age children of extremely low
birth weight, as compared with those born at full
sex and race or ethnic group as the index child, with
birthdays close to that of the index child, and we
then selected a control at random from the three; if
a child identified by the teacher for comparison had
term.1-4 Such children were born before the wide been born preterm, a different child was selected.
introduction of antenatal treatment with corticoste- Race or ethnic group was assigned by the teacher
roids and surfactants. These agents are important on the basis of knowledge of the children and their
determinants of the increased survival of extremely families. A total of 160 children selected as com-
preterm infants5,6 and might be expected to im- parison (control) subjects who were born at full
prove long-term outcomes. term were evaluated. In another 44 cases, the head
In a previous report in the Journal, we described teachers or the parents refused to take part in the
the outcomes at 30 months of age (corrected for study, in 2 cases the head teachers did not find a
prematurity) of a cohort of infants born at 25 or suitable match to the index child, and in one case
fewer completed weeks of gestation in 1995 in the the child was assessed outside the school, without
United Kingdom and Ireland (the EPICure Study).7 a classmate.
More than 60 percent of the children in the study All parents gave written informed consent, and
cohort were exposed to antenatal treatment with the study was approved by the Trent Multicenter
steroids, and 84 percent received surfactant.8 At 30 research ethics committee and the local education
months of age, 24 percent of the survivors had se- authorities in Scotland. The study investigators
vere disabilities. were responsible for the identification of the orig-
The high prevalence of disability at 30 months inal study cohort and the studies of children up to
of age made it important to assess this cohort fur- 2.5 years of age and again at 6 years of age, and
ther, at a later age, when the degree of disability can the developmental panel performed the data col-
be more clearly defined and is more likely to be pre- lection.
dictive of problems that will continue throughout
childhood and into later life. In this report, we de- assessment
scribe the outcomes among this cohort at six years The 207 children in mainstream schools were eval-
of age, when the children were involved in full-time uated by means of a clinical examination includ-
education. ing neuropsychological assessment. Children with
disabilities in special-needs schools were evaluat-
methods ed without the use of a comparison child by means
of an appropriate assessment. The developmental
study subjects panel included seven experienced developmental
We identified all extremely preterm children (ges- pediatricians and eight psychologists, who received
tation at birth, no more than 25 weeks and 6 days) formal training in performing the assessments.
who were born in the United Kingdom and Ireland All reached the required level of competence (agree-
between March and December 1995.8,9 Of 308 chil- ment of more than 80 percent with an independent
dren known to have been alive at 30 months, all had observer for videotaped tasks) before commencing
survived to 6 years of age; 15 of these children were the study assessments. The assessors were unaware
living outside the United Kingdom or Ireland. Of of the neonatal courses of the children they evalu-
the remaining 293 children, 241 (82 percent) partic- ated and were not informed as to which children
ipated in this study at a median age of six years and were preterm and which were the controls.
four months (range, five years and two months to We classified the children into four functional
seven years and three months). Of the 241 partic- groups of disability on the basis of the definitions
ipants, 34 were being educated at schools for chil- we have used previously.10 A disability was defined
dren with special needs, 3 were in a special-needs as severe if it was considered likely to make the child
class attached to a mainstream school, and 204 highly dependent on caregivers and if it included
were in a class in a mainstream school. nonambulant cerebral palsy, an IQ score more than
For each child in a class in a mainstream school, 3 SD below the mean, profound sensorineural hear-
we sought a classmate matched for age and sex to ing loss, or blindness. A disability was defined as
serve as a control. We requested that the head teach- moderate if reasonable independence was likely to
10 n engl j med 352;1 www.nejm.org january 6 , 2005
disability and extreme prematurity
be reached and if it included ambulant cerebral pal- GraphPad Software). All statistical tests were two-
sy, an IQ score 2 to 3 SD below the mean, sensori- sided.
neural hearing loss that was corrected with a hear-
ing aid, and impaired vision without blindness. Mild results
disability included neurologic signs with minimal
functional consequences or other impairments We had previously evaluated 283 children at 30
such as squints or refractive errors. Cerebral palsy months of age corrected for prematurity.7 Of the
was classified independently of the degree of dis- 47 children not assessed again at six years of age,
ability, and the classification was made retrospec- 17 (36 percent) had been classified as severely dis-
tively, at the completion of the study, according to abled, 12 (26 percent) had other disabilities, and 18
the description of functions for each limb,11 by (38 percent) had no disability. As compared with
two assessors. those who had been assessed, these 47 children were
When a cognitive assessment was appropriate, more likely to have young mothers (21 years of age
it was made with the use of the Kaufman Assess- or younger; 24 percent vs. 11 percent, P<0.01). The
ment Battery for Children (K-ABC).12 Among the distribution of neonatal complications, other socio-
41 index children whose severe cognitive impair- economic factors, and outcomes at 30 months of
ment or disability precluded the use of this assess- age corrected for prematurity was similar in the two
ment tool, either the Griffiths Scales of Mental groups. Another five children who were not evalu-
Development13 (35 children) or the neuropsycho- ated at 30 months were included in the analysis. The
logical instrument known as NEPSY14 (6 children) 241 children assessed for this report were repre-
was used, and the results for these children were sentative of the whole population of survivors (308)
substituted for the missing values in the Mental Pro- with regard to birth weight, gestational age, and
cessing Composite of K-ABC to produce an overall several perinatal variables (see the Supplementary
cognitive score. Children with a score below 40 (the Appendix, available with the full text of this article
lowest score in the K-ABC) were assigned a score at www.nejm.org).
of 39. No other substitution of data values was made
in other K-ABC scales. We measured cognitive im- cognitive performance
pairment with the use of reference groups consist- Table 1 shows the mean (±SD) scores for all the
ing of both the published test norms (based on children according to sex. Figure 1 shows individ-
children born in the late 1970s, for the K-ABC) and ual cognitive scores according to sex and gestation-
contemporary classmates, given the well-described al age at birth. The mean difference in scores for
rise in IQ scores over time.15-17 Cognitive perfor- overall cognitive ability between the extremely pre-
mance was classified as severely impaired if the term children and the comparison group of class-
score was more than 3 SD below the mean, moder- mates was 24 points (95 percent confidence inter-
ately impaired if it was more than 2 but not more val, 20 to 27). Whereas in the group of classmates
than 3 SD below the mean, and mildly impaired if the scores of boys were similar to those of girls,
it was more than 1 but not more than 2 SD below among the extremely preterm children boys had
the mean. lower scores than girls (mean difference, 10 points);
the interaction between sex and group was signif-
statistical analysis icant (P=0.002). These effects remained after the
Data were collected on standardized forms and en- exclusion of children with physical disability (mean
coded for computerized analysis with the use of difference between the preterm group and the com-
SPSS for Windows software (version 11). The as- parison group, 20 points; 95 percent confidence
sessment data for each child were examined be- interval, 17 to 23) and those who could not com-
fore they were combined with the data set of the plete the K-ABC (data not shown). Although there
previous main study for analysis. Categorical out- appeared to be a reduction in the scores with de-
comes were compared with use of chi-square tests creasing gestational age at birth from 25 to 23 weeks
for trends, as appropriate, or Fisher’s exact test. (Fig. 1), this trend was not significant after adjust-
Continuous outcomes were compared with use of ment for sex.
independent Student’s t-tests. Differences in pro- The scores of extremely preterm children were
portions and odds ratios were calculated from con- significantly lower than those of control children
tingency tables with the use of Instat (version 3.02, in all the K-ABC subscales, and boys had consistent-
Table 1. Overall Cognitive Scores and Composite Score Profiles for Extremely Preterm Children and Classmates According to Sex.*
Mean Difference
in Scores Odds Ratio
Scale Classmates Extremely Preterm Children (95% CI) (95% CI)
Serious Serious
No. Score Impairment No. Score Impairment
% %
Overall cognitive score† 160 105.7±11.8 1.3 241 82.1±19.2 41 24 (20 to 27) 56 (13 to 250)
Boys 71 105.9±12.8 1.4 122 77.1±19.6 49 29 (24 to 34)
Girls 89 105.5±11.0 1.1 119 87.2±17.4 32 18 (14 to 22)
Mean difference between extremely 10 (5 to 15)
preterm girls and boys
Mental processing composite 159 105.7±11.8 1.3 200 88.4±12.7 29 17 (15 to 20) 32 (8 to 143)
Boys 71 105.9±12.8 1.4 92 85.9±11.5 33 20 (16 to 24)
Girls 88 105.5±11.0 1.1 108 90.5±13.4 26 15 (12 to 19)
Simultaneous processing 159 104±11.5 1.3 201 86.3±12.8 32 18 (15 to 21) 37 (9 to 143)
Boys 71 105.2±12.3 1.4 93 84.1±11.9 35 21 (17 to 25)
Girls 88 103.1±10.8 1.1 108 88.2±13.2 29 15 (11 to 18)
Sequential processing 159 106.5±12.5 1.3 202 94.5±13.9 15 12 (9 to 15) 14 (3 to 59)
Boys 71 105.4±13.0 1.4 93 92.2±12.8 16 13 (9 to 17)
Girls 88 107.5±12.1 1.1 109 96.5±14.5 14 11 (7 to 15)
Achievement scale 159 100.0±15.0 4.5 205 78.7±21.4 27 21 (17 to 25) 8 (3 to 18)
Boys 71 99.4±15.9 5.6 96 75.6±20.7 32 24 (18 to 30)
Girls 88 100.5±14.3 3.5 109 81.4±21.7 22 19 (14 to 24)
Mean difference between extremely 6 (¡0.1 to 12)
* The Kaufman Assessment Battery for Children (K-ABC)12 comprises two summative scales: the Mental Processing Composite, a global mea-
sure of cognitive ability in two scales, sequential processing and simultaneous processing; and the Achievement Scale, an assessment of
knowledge of facts, language concepts, and school-related skills (range of possible scores, 40 to 150). These scales were standardized to a
mean (±SD) of 100±15, on the basis of results in children born in the late 1970s. Serious impairment is defined as a score more than 2 SD be-
low the mean score for the comparison group, equivalent to the categories of moderate impairment and severe impairment. Odds ratios for
the risk of serious impairment were calculated with the use of logistic regression, with sex as a covariate. Plus–minus values are means ±SD.
CI denotes confidence interval.
† For the overall cognitive score, the range of possible scores is from 39, lowest, to 150, highest.
ly lower subscale scores than girls (Table 1). Among On the basis of the K-ABC reference data, 21
extremely preterm children, the mean scores for percent of the extremely preterm children were
simultaneous processing were 8 points lower than classified as having moderate or severe cognitive
those for sequential processing, and the achieve- impairment (Table 2), as compared with none of
ment scores were significantly lower than the Men- the children in the standardized data group. In con-
tal Processing Composite scores (by 10 points trast, as compared with the classmates used as the
among boys, and 9 points among girls). In contrast, reference group, 41 percent of extremely preterm
the control group had similar scores across the four children scored more than 2 SD below the mean, as
scales in the K-ABC. compared with 2 percent of classmates in the com-
parison group (Table 2). With the use of the controls

as the reference group for all subsequent compari- 150
sons, extremely preterm children were significant- 140

ly more likely to have moderate-to-severe overall 130
cognitive deficits. As compared with extremely pre- 120
term girls, extremely preterm boys were more than 110
twice as likely to have serious impairment as shown 100
in the overall cognitive scores (Table 1). Among
Cognitive Score
90
extremely preterm children, those from multiple 80
births had a risk of disability that was similar to
70
that of singleton births (data not shown).
60
50
neuromotor function
40
Of the 241 extremely preterm children assessed, 49
(20 percent) had spastic or dyskinetic cerebral pal- 30
sy and a further 9 had abnormal neurologic signs 20

(hypotonia), whereas none in the control group had 10
cerebral palsy or abnormal neurologic conditions. 0
Boys Girls Boys Girls Boys Girls Boys Girls
Thirty-five children (15 percent) had signs of spas- (N=7) (N=17) (N=37)(N=36) (N=78)(N=66) (N=71)(N=89)
tic diplegia, four had hemiplegia, nine had quadri-
≤23 wk 24 wk 25 wk Comparison
plegia, and one child had dyskinetic cerebral palsy. Gestational Age (completed wk) Group
Of the 35 children, 15 were not walking (severe mo-
tor disability, 6 percent) and another 15 with cere- Figure 1. Cognitive Scores for 241 Extremely Preterm Children and 160 Age-
bral palsy were independently walking but with ab- Matched Classmates Who Were Full Term at Birth, According to Sex and Com-
normal gait (moderate motor disability). The mean pleted Weeks of Gestation.
cognitive score for these 30 children was 49±17. The scores are Kaufman Assessment Battery for Children scores for the Men-
tal Processing Composite or developmental scores according to the Griffiths
The remaining 19 children with cerebral palsy had
Scales of Mental Development and NEPSY (possible range, 39 to 150); higher
no evidence of clinically important functional diffi- scores indicate better function. Bars indicate mean values. The dashed line is
culties relating to gait or hand use identified in the the mean of the standardized data (normal population).
assessment; and a mean cognitive score of 81±19.
Twelve children with diplegia (34 percent) had
mild motor disability, as compared with only 7 per-
cent of children with other types of cerebral palsy. whom required hearing aids. In contrast, hearing
Cerebral palsy was more common among boys than impairment was present in three control children,
girls (26 percent vs. 14 percent; odds ratio for the one of whom required a hearing aid.
risk of cerebral palsy, 2.1; 95 percent confidence in-
terval, 1.1 to 4.1), as was cerebral palsy with disabil- overall classification
ity (odds ratio, 2.2; 95 percent confidence interval, Two of the 160 comparison children had a moderate
0.9 to 5.4). disability. The outcome for the extremely preterm
cohort is summarized in Table 2 in each of four do-
sensory morbidity mains (neuromotor, cognition, hearing, and vision)
Four children were blind, and two could see only and according to gestational age and sex in Table 3.
light (severe disability, 2 percent) (Table 2); five of These data were combined with birth information
these children had received treatment for retino- to derive overall outcomes at each gestational age
pathy of prematurity. Many children had other, less on the basis of live births or admission to neonatal
severe visual impairments: squint was present in intensive care units (Table 4).
58 (24 percent), and 58 (24 percent) wore eyeglass-
es, as compared with 1 control child with a squint change in disability profile
and 6 with refractive errors. Seven children had pro- between assessments
found hearing loss that could not be corrected with Of the 241 children in this study, 236 had also been
hearing aids (severe disability, 3 percent) and 17 among the group assessed at 30 months of age.7
children had lesser degrees of hearing loss, 7 of Outcome at 30 months of age corrected for pre-
Table 2. Neurocognitive Function and Degree of Disability at Six Years of Age among 241 Extremely Preterm Children and 160 Classmates
Born at Full Term.*
Disability† Comparison with Standardized Data Comparison with Classmates
Comparison Extremely Preterm Comparison Extremely Preterm

Group Group Group Group
no. % (95% CI) no. % (95% CI) no. % (95% CI) no. % (95% CI)
Severe
Cerebral palsy, nonambulatory 0 15 6 (4–10) 0 15 6 (4–10)
IQ >3 SD below mean
Range, 39–54 0 27 11 (8–16)
Range, 39–69 0 50 21 (16–26)
Profound sensorineural hearing loss 0 7 3 (1–6) 0 7 3 (1–6)
Blind 0 6 2 (1–5) 0 6 2 (1–5)
Any severe disability 0 32 13 (9–18) 0 53 22 (17–28)
Moderate
Abnormal neurologic findings with 0 17 7 (4–11) 0 17 7 (4–11)
functional loss but ambulatory
IQ >2 to 3 SD below mean
Range, 55–69 0 23 10 (6–14)
Range, 70–81 2 1 (0–4) 48 20 (15–26)
Sensorineural hearing loss corrected 1 1 (0–3) 7 3 (1–6) 1 1 (0–3) 7 3 (1–6)
with hearing aids
Impaired vision but ability to see 0 11 5 (2–8) 0 11 5 (2–8)
Any moderate disability 1 1 (0–3) 27 11 (8–16) 2 1 (0–4) 57 24 (18–30)
Mild
Neurologic signs, minimal functional 0 26 11 (7–15) 0 26 11 (7–15)
impairment
IQ >1 to 2 SD below mean
Range, 70–84 3 2 (0–5) 61 25 (20–31)
Range, 82–94 23 14 (9–21) 75 31 (25–37)
Mild hearing impairment 2 1 (0–4) 10 4 (2–7) 2 1 (0–4) 10 4 (2–7)
Squint or refractive error 7 4 (2–8) 69 29 (23–35) 7 4 (2–8) 69 29 (23–35)
Any mild disability 9 6 (3–10) 71 29 (24–36) 28 18 (12–24) 83 34 (29–41)
No disability 150 94 (89–97) 111 46 (40–53) 120 75 (68–81) 48 20 (15–25)
* IQ scores for extremely preterm infants were classified with the use of the original test standardization norms or results obtained for class-
mates. IQ scores are presented as ranges of the overall cognitive scores in each category; for standardization norms, these are scores with
a standardization mean (±SD) of 100±15, and for classmates a standardization mean of 106±12. Among extremely preterm children, 50 had
scores more than 2 SD below the mean as compared with the standardized data and 98 had scores more than 2 SD below the mean as com-
pared with classmates.
† A severe disability was defined as one that was likely to make the child highly dependent on caregivers, a moderate disability as one that would
probably allow a reasonable degree of independence to be reached, and a mild disability as the presence of neurologic signs with minimal
functional consequences or with other impairments, such as squint or refractive error.
maturity was divided into three functional out- fant Development were used to assign a value for
comes, on the basis of professional consensus10 — cognitive disability without the use of a compari-
namely, severe disability, other disabilities, and no son or control group. The category of “other dis-
disability. The criteria for severe disability used at abilities” included any disability or impairment in
30 months of age were the same as those used at the four domains that was not covered by the cate-
6 years of age, except that the Bayley Scales of In- gory of “severe” disability.
Table 3. Severity and Type of Disability at Six Years of Age among Extremely Preterm Children, According to Gestational Age at Birth and Sex.*
Total
Domain ≤23 Wk 24 Wk 25 Wk (N=241)
Boys Girls All Boys Girls All Boys Girls All

(N=7) (N=17) (N=24) (N=37) (N=36) (N=73) (N=78) (N=66) (N=144)
number of children (percent)
Overall cognition
No disability (score, >94) 0 6 (35) 6 (25) 4 (11) 11 (31) 15 (21) 20 (26) 27 (41) 47 (33) 68 (28)
Mild disability (score, 82–94) 1 (14) 3 (18) 4 (17) 13 (35) 12 (33) 25 (34) 24 (31) 22 (33) 46 (32) 75 (31)
Moderate disability (score, 70–81) 3 (43) 5 (29) 8 (33) 6 (16) 7 (19) 13 (18) 16 (21) 11 (17) 27 (19) 48 (20)
Severe disability (score, ≤69) 3 (43) 3 (18) 6 (25) 14 (38) 6 (17) 20 (27) 18 (23) 6 (9) 24 (17) 50 (21)
Neuromotor
No disability 5 (71) 13 (76) 18 (75) 21 (57) 30 (83) 51 (70) 58 (74) 56 (85) 114 (79) 183 (76)
Abnormal signs, minimal 2 (29) 2 (12) 2 (8) 4 (11) 4 (11) 8 (11) 11 (14) 5 (8) 16 (11) 26 (11)
functional loss
Cerebral palsy with disability, 0 1 (6) 3 (12) 6 (16) 0 6 (8) 5 (6) 3 (5) 8 (6) 17 (7)
ambulatory
Cerebral palsy, nonambulatory 0 1 (6) 1 (4) 6 (16) 2 (6) 8 (11) 4 (5) 2 (3) 6 (4) 15 (6)
Hearing
No disability 6 (86) 15 (88) 21 (88) 32 (86) 30 (83) 62 (85) 73 (94) 61 (92) 134 (93) 217 (90)
Mild hearing loss 1 (14) 1 (6) 2 (8) 3 (8) 2 (6) 5 (7) 2 (3) 1 (2) 3 (2) 10 (4)
Use of hearing aid, but hears 0 0 0 0 2 (6) 2 (3) 2 (3) 3 (5) 5 (3) 7 (3)
Profound hearing loss 0 1 (6) 1 (4) 2 (5) 2 (6) 4 (5) 1 (1) 1 (2) 2 (1) 7 (3)
Vision
No disability 5 (71) 6 (35) 11 (46) 20 (54) 20 (56) 40 (55) 53 (68) 51 (77) 104 (72) 155 (64)
Squint or refractive error 2 (29) 7 (41) 9 (38) 11 (30) 14 (39) 25 (34) 21 (27) 14 (21) 35 (24) 69 (29)
Visually impaired, not blind 0 2 (12) 2 (8) 4 (11) 1 (3) 5 (7) 3 (4) 1 (2) 4 (3) 11 (5)
Severe blindness 0 2 (12) 2 (8) 2 (5) 1 (3) 3 (4) 1 (1) 0 1 (1) 6 (2)
Overall disability
None 0 3 (18) 3 (12) 2 (5) 8 (22) 10 (14) 16 (21) 19 (29) 35 (24) 48 (20)
Mild 1 (14) 5 (29) 6 (25) 11 (30) 15 (42) 26 (36) 24 (31) 27 (41) 51 (35) 83 (34)
Moderate 3 (43) 6 (35) 9 (38) 9 (24) 7 (19) 16 (22) 19 (24) 13 (20) 32 (22) 57 (24)
Severe 3 (43) 3 (18) 6 (25) 15 (41) 6 (17) 21 (29) 19 (24) 7 (11) 26 (18) 53 (22)
* Cognitive impairment was defined with the use of contemporary classmates as a reference group.
Figure 2 shows the relationship between the specificity (93 percent). Thirty-eight percent of chil-
findings at 30 months of age and at 6 years of age dren classified as having “other disabilities” and 24
for the 236 extremely preterm children assessed at percent of those categorized as having “no disabil-
both ages. Severe disability at 30 months was high- ity” at 30 months had moderate or severe disability
ly predictive of outcome at 6 years (P<0.001). Of at the assessment at 6 years of age.
63 children classified as having severe disability at
30 months, 86 percent had either severe or moder- discussion
ate disability at 6 years. The category of severe dis-
ability at 30 months had low sensitivity (50 percent) We observed a high prevalence of disability at early
for moderate or severe disability at 6 years but good school age among a cohort of children born be-
Table 4. Summary of Outcomes among Extremely Preterm Children.* Severe Moderate Mild No
disability disability disability disability
22 Wk 23 Wk 24 Wk 25 Wk 65
Outcome (N=138) (N=241) (N=382) (N=424) 60
55
number (percent) 50
Outcome at 6 Yr (%)
45
Died in delivery room 116 (84) 110 (46) 84 (22) 67 (16)
40
Admitted to NICU 22 (16) 131 (54) 298 (78) 357 (84) 35
30
Died in NICU 20 (14) 105 (44) 198 (52) 171 (40) 25
Survived to discharge 2 (1) 26 (11) 100 (26) 186 (44) 20
15
Died after discharge 0 1 (0.4) 2 (0.5) 3 (0.7) 10
5
Lost to follow-up 0 3 (1) 25 (7) 39 (9)
0
Severe Moderate or mild None
At 6 yr of age
Disability at 30 Mo
Had severe disability 1 (0.7) 5 (2) 21 (5) 26 (6)
Had moderate disability 0 9 (4) 16 (4) 32 (8) Figure 2. Severity of Disability at 30 Months of Age Cor-
rected for Prematurity and at 6 Years among 236 Chil-
Had mild disability 1 (0.7) 5 (2) 26 (7) 51 (12) dren Who Were Assessed at Both Ages.
Survived without impairment A total of 63 children had severe disability, 82 other dis-
abilities (moderate or mild), and 91 no disability.
As a percentage of live 0 3 (1) 10 (3) 35 (8)
births
As a percentage of NICU 0 3 (2) 10 (3) 35 (10) tremely preterm children were found to have general
admissions
cognitive deficits (41 percent). Thus, important dif-
Survived without severe or ferences may be underestimated if preterm children
moderate disability
are not compared with a contemporary comparison
As a percentage of live 1 (0.7) 8 (3) 36 (9) 86 (20) group.17,19
births This finding of substantial general cognitive def-
As a percentage of NICU 1 (5) 8 (6) 36 (12) 86 (24) icit is unlikely to be attributable to selective dropout
admissions of high achievers. Our analysis indicates that there
was no material difference with regard to medical
* NICU denotes neonatal intensive care unit. variables, growth, or early disability between those
lost to follow-up and those assessed. Rather, drop-
fore 26 weeks of gestation whom we had evaluat- out was more likely to occur in families of young
ed at 30 months of age corrected for prematurity. mothers and to be associated with social disadvan-
Cognitive impairment was the most common dis- tage, which would be expected to reduce cognitive
ability of the four domains assessed — neuromo- performance17,20,21; we may therefore have under-
tor, cognition, hearing, and vision. estimated the frequency of cognitive deficit in our
As compared with the normative data used to population. However, our classroom controls may
standardize the K-ABC, which may be considered to represent a relatively healthy group, because only
be equivalent to “historical controls” from the late peers from mainstream schools were included. The
1970s, the data on this group of children showed United Kingdom and Ireland have a policy that aims
that they do not seem to be faring much worse than at integrating children with special needs into main-
has been reported previously for extremely preterm stream education, and only children with a very high
or extremely-low-birth-weight children at a simi- degree of special-needs requirements are educated
lar age1,17,18: 21 percent had scores more than 2 SD in special schools. Thus, we would expect only
below the mean (Table 2). All these studies includ- slight overestimations of differences between ex-
ed more mature infants and extremely preterm chil- tremely preterm children and the classmates who
dren. However, as compared with a more relevant served as controls.
control group of contemporary classmates who The proportion of children who were consid-
were born at full term, almost twice as many ex- ered to have a severe disability (22 percent at 6 years
of age) was similar to that at 30 months of age. There are few reports of outcomes at school age
The category of severe disability that we used at 30 among children born extremely preterm.26 Direct
months was identified by consensus as an impor- comparisons between previous reports and the
tant outcome classification that was likely to imply present study are made difficult by the lack of con-
substantial ongoing disability.10 The value of this sistency in the measures of outcome. In four reports
outcome classification as a high-risk category was spanning the decade from 1984 through 1994, the
confirmed in the current study, because 86 percent prevalence of severe disability (variously defined) at
went on to have severe or moderate disability at six three to five years of age ranged from 25 to 60 per-
years of age. The outcomes of children whose dis- cent among survivors born at less than 26 weeks
abilities were assessed as less severe at 30 months of gestation.27-30 The numbers of children includ-
of age appeared similar to those of children with- ed in these studies were small, ranging from 14 to
out a documented disability at that assessment. 77. In contrast, a study conducted in Stockholm of
Twelve percent of the children assessed had ce- births during the period from 1990 to 1992 reported
rebral palsy with moderate or severe motor disabil- severe disability in 14 percent of three-year-old
ity, but these children represent only half of those children (4 of 29) born at 24 weeks or less of ges-
observed to have abnormal neurologic signs. Chil- tation and in 9 percent of children (14 of 148) born
dren with moderately or severely disabling cerebral at 25 or 26 weeks.31 None of these studies used
palsy were also more likely to have cognitive im- classmates as a control group.
pairment, as compared with children with other The rate of disability may also be dependent on
motor disabilities who had a similar frequency of the willingness of the attending clinical staff to in-
cognitive impairment as children without cerebral stitute intensive care at the time of an extremely pre-
palsy. Nonetheless, the cognitive scores of children term birth32; comparisons across studies are thus
with cerebral palsy and mild motor disability were also limited by the lack of detailed information on
lower than those of the remaining children without the liveborn population from which such infor-
signs of neurologic abnormalities. This finding is mation is derived and on the delivery-room policy
consistent with previous reports of the cognitive with regard to instituting intensive care. The rate of
function of preterm children born at later gesta- moderate or severe disability that we observed —
tional ages.22,23 46 percent among our assessed cohort — is simi-
The differences between the sexes in outcome lar to that in earlier reports from the United King-
that was noted among extremely preterm children dom, the United States, and Australia for births in
are concordant with previous observations of ex- 1984, 1990 through 1994, and 1991 to 1992, respec-
cess mortality8 and morbidity7,8 among extremely tively.27-30 Furthermore, these studies all classified
preterm boys, as compared with girls. Sex differ- outcomes with the use of standardization data with-
ences in cognitive measures appeared to be great- out correction for contemporary performance, and
er at 6 years of age (10 points) than at 30 months therefore do not indicate any improvement in prog-
(6 points), in contrast to the absence of sex differ- nosis over the decade spanned by these reports. Re-
ences among the classmates of the extremely pre- ports of outcomes based on birth weight for chil-
term children. The concept of biologic vulnera- dren with a birth weight of 1000 g or less are more
bility among boys has been debated,24 although a common than studies based on gestational age,26
range of adverse perinatal outcomes has been asso- but studies based on birth weight are not directly
ciated with male sex, including death, perinatal comparable to ours, because they include not only
brain injury, cerebral palsy, delayed preterm lung extremely preterm children but also children of
maturation, preterm birth, and stillbirth.24 Sub- higher gestational age with intrauterine growth re-
stantial differences between the sexes in cognitive striction.
functioning have been reported in some cohorts25 The prevalence of cerebral palsy among extreme-
but not in others.17 Our finding of a significant ly preterm children ranged from 16 to 21 percent in
increase in the risk of disability — in particular, four recent reports.27,28,31,33 Although we report a
the risk of severe disability, cerebral palsy, and low similar prevalence (20 percent), the proportion of
scores for cognitive functioning — supports the children with cerebral palsy and severe or moder-
concept that male sex is an important biologic risk ate motor disability is lower (12 percent). We be-
factor in extremely preterm infants that may have lieve this difference is clinically important, because
implications for clinical practice. a neurologic abnormality that does not interfere
with gait or with gross manipulative skills allows learning difficulties, such as in language or be-
the child substantial independence and a better havioral and social areas,34 are important addition-
quality of life. al predictors of academic achievement. Although
The present findings in a large cohort of chil- the majority of children born at fewer than 26 weeks
dren born at fewer than 26 weeks of gestation indi- of gestation have cognitive scores within the low-
cate that adverse cognitive sequelae are a more fre- er normal range, close follow-up of these children
quent outcome among such infants than among is warranted with regard to their future academic1
more mature preterm populations. Further analy- and psychological34,35 difficulties.
sis will be needed to determine whether the cog- Supported by BLISS, the premature baby charity; the Health
nitive impairments may explain the educational Foundation; and WellBeing of Women.
We are indebted to the EPICure Study Group, which includes pe-
difficulties that have been reported for extremely- diatricians in 276 maternity units in the United Kingdom and Ireland
low-birth-weight children1 and whether specific who contributed data to the study and whose help was invaluable.
ap p e n d i x
The following investigators and study coordinators participated in the EPICure Study Group: K. Costeloe (London), A.T. Gibson (Sheffield),
E.M. Hennessy (London), N. Marlow (Nottingham), A.R. Wilkinson (Oxford), D. Wolke (Bristol). Developmental panel: pediatricians: M.
Bracewell, M. Cruwys, R. MacGregor, L. McDonald, M. Morton, M. Morris, S. Thomas; psychologists: E. Luck, C. Bamford, H. Betteridge,
H. Bruhn, S. Johnson, I. Magiati, M. Morahan; I. Tsverik; M. Samara (psychological data analysis); H. Palmer (study administrator).
refer enc es
1. Saigal S, den Ouden L, Wolke D, et al. National Perinatal Epidemiology Unit and ioral sequelae of perinatal insults and early
School-age outcomes in children who were the former Oxford Regional Health Author- family adversity at 8 years of age. J Am Acad
extremely low birth weight from four inter- ity. March 1994. Oxford, England: NPEU and Child Adolesc Psychiatry 2000;39:1229-
national population-based cohorts. Pediat- ORHA, 1995. 37.
rics 2003;112:943-50. 11. Evans P, Alberman E, Johnson A, Mutch 21. Taylor HG, Burant CJ, Holding PA, Klein
2. Whitfield MF, Grunau RV, Holsti L. L. Standardisation of recording and report- N, Hack M. Sources of variability in sequelae
Extremely premature (< or = 800 g) school- ing cerebral palsy. Dev Med Child Neurol of very low birth weight. Neuropsychol Dev
children: multiple areas of hidden disability. 1987;29:272. Cogn Sect C Child Neuropsychol 2002;8:
Arch Dis Child Fetal Neonatal Ed 1997;77: 12. Kaufman A, Kaufman N. Kaufman 163-78.
F85-F90. Assessment Battery for Children (K-ABC). 22. Nelson KB, Ellenberg JH. Children who
3. Taylor HG, Klein N, Hack M. School- Circle Pines, Minn.: American Guidance Ser- “outgrew” cerebral palsy. Pediatrics 1982;69:
age consequences of birth weight less than vice, 1983. 529-36.
750 g: a review and update. Dev Neuropsy- 13. Griffiths R. The abilities of young chil- 23. Drillien CM. Abnormal neurologic signs
chol 2000;17:289-321. dren. Somerset, U.K.: Young and Son, 1970. in the first year of life in low-birthweight
4. Hack M, Flannery DJ, Schluchter M, 14. Korkman M, Kirk U, Kemp S. Manual for infants: possible prognostic significance.
Cartar L, Borawski E, Klein N. Outcomes in the NEPSY: a developmental neuropsycho- Dev Med Child Neurol 1972;14:575-84.
young adulthood for very-low-birth-weight logical assessment. San Antonio, Tex.: Psy- 24. Kraemer S. The fragile male. BMJ 2000;
infants. N Engl J Med 2002;346:149-57. chological Testing Corporation, 1998. 321:1609-12.
5. Fanaroff AA, Poole K, Duara S, et al. 15. Wolke D, Ratschinski G, Ohrt B, Riegel 25. Brothwood M, Wolke D, Gamsu H, Ben-
Micronates: 401-500 grams: the NICHD Neo- K. The cognitive outcome of very preterm son J, Cooper D. Prognosis of the very low
natal Research Network Experience 1996- infants may be poorer than often reported: birthweight baby in relation to gender. Arch
2001. Pediatr Res 2003;54:398A. abstract. an empirical investigation of how method- Dis Child 1986;61:559-64.
6. Lorenz JM, Wooliever DE, Jetton JR, ological issues make a big difference. Eur 26. Hack M, Fanaroff AA. Outcomes of chil-
Paneth N. A quantitative review of mortality J Pediatr 1994;153:906-15. dren of extremely low birthweight and ges-
and developmental disability in extremely 16. Flynn J. Searching for justice: the discov- tational age in the 1990’s. Early Hum Dev
premature newborns. Arch Pediatr Adolesc ery of IQ gains over time. American Psychol- 1999;53:193-218.
Med 1998;152:425-35. ogist 1999;54:5-20. 27. Emsley HC, Wardle SP, Sims DG, Chis-
7. Wood NS, Marlow N, Costeloe K, Gib- 17. Wolke D, Meyer R. Cognitive status, lan- wick ML, D’Souza SW. Increased survival
son AT, Wilkinson AR. Neurologic and devel- guage attainment, and prereading skills of and deteriorating developmental outcome in
opmental disability after extremely preterm 6-year-old very preterm children and their 23 to 25 week old gestation infants, 1990-4
birth. N Engl J Med 2000;343:378-84. peers: the Bavarian Longitudinal Study. Dev compared with 1984-9. Arch Dis Child Fetal
8. Costeloe K, Hennessy E, Gibson AT, Mar- Med Child Neurol 1999;41:94-109. Neonatal Ed 1998;78:F99-F104.
low N, Wilkinson AR. The EPICure study: 18. Bhutta AT, Cleves MA, Casey PH, Cra- 28. Piecuch RE, Leonard CH, Cooper BA,
outcomes to discharge from hospital for dock MM, Anand KJ. Cognitive and behav- Kilpatrick SJ, Schlueter MA, Sola A. Out-
infants born at the threshold of viability. ioral outcomes of school-aged children who come of infants born at 24-26 weeks’ ges-
Pediatrics 2000;106:659-71. were born preterm: a meta-analysis. JAMA tation: II. Neurodevelopmental outcome.
9. Jacobs SE, O’Brien K, Inwood S, Kelly EN, 2002;288:728-37. Obstet Gynecol 1997;90:809-14.
Whyte HE. Outcome of infants 23-26 weeks’ 19. Aylward GP, Pfeiffer SI, Wright A, Ver- 29. Doyle LW. Outcome at 5 years of age of
gestation pre and post surfactant. Acta Pae- hulst SJ. Outcome studies of low birth weight children 23 to 27 weeks’ gestation: refining
diatr 2000;89:959-65. infants published in the last decade: a meta- the prognosis. Pediatrics 2001;108:134-41.
10. Disability and perinatal care: a report analysis. J Pediatr 1989;115:515-20. 30. Johnson A, Townshend P, Yudkin P, Bull
of two working groups convened by the 20. Laucht M, Esser G, Baving L, et al. Behav- D, Wilkinson AR. Functional abilities at age
4 years of children born before 29 weeks of New Jersey and the Netherlands: outcomes Behavioural problems in children who weigh
gestation. BMJ 1993;306:1715-8. and resource expenditure. Pediatrics 2001; 1000 g or less at birth in four countries. Lan-
31. Finnstrom O, Gaddlin PO, Leijon I, Sam- 108:1269-74. cet 2001;357:1641-3.
uelsson S, Wadsby M. Very-low-birth-weight 33. D’Angio CT, Sinkin RA, Stevens TP, et al. 35. Botting N, Powls A, Cooke RW, Marlow
children at school age: academic achieve- Longitudinal, 15-year follow-up of children N. Attention deficit hyperactivity disorders
ment, behavior and self-esteem and relation born at less than 29 weeks’ gestation after and other psychiatric outcomes in very low
to risk factors. J Matern Fetal Neonatal Med introduction of surfactant therapy into a birthweight children at 12 years. J Child Psy-
2003;14:75-84. region: neurologic, cognitive, and educa- chol Psychiatry 1997;38:931-41.
32. Lorenz JM, Paneth N, Jetton JR, den tional outcomes. Pediatrics 2002;110:1094- Copyright © 2005 Massachusetts Medical Society.
Ouden L, Tyson JE. Comparison of manage- 102.
ment strategies for extreme prematurity in 34. Hille ET, den Ouden AL, Saigal S, et al.
View publication stats

Neurologic and Developmental Disability

Uploaded by

Copyright:

Available Formats

Neurologic and Developmental Disability

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Neurologic and Developmental Disability

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Neurologic and Developmental Disability at Six Years of Age after Extremely

Article in New England Journal of Medicine · February 2005

Neil Marlow Dieter Wolke

SEE PROFILE SEE PROFILE

SNAP Trial View project

The user has requested enhancement of the downloaded file.

Neurologic and Developmental Disability

Drs. Marlow and Wolke contributed equally

n engl j med 352;1 www.nejm.org january 6, 2005 9

er of the class identify three children of the same

a n increased prevalence of cog-

10 n engl j med 352;1 www.nejm.org january 6 , 2005

n engl j med 352;1 www.nejm.org january 6, 2005 11

12 n engl j med 352;1 www.nejm.org january 6 , 2005

parison group (Table 2). With the use of the controls

sons, extremely preterm children were significant- 140

sy and a further 9 had abnormal neurologic signs 20

n engl j med 352;1 www.nejm.org january 6, 2005 13

Disability† Comparison with Standardized Data Comparison with Classmates

Comparison Extremely Preterm Comparison Extremely Preterm

14 n engl j med 352;1 www.nejm.org january 6 , 2005

Boys Girls All Boys Girls All Boys Girls All

n engl j med 352;1 www.nejm.org january 6, 2005 15

16 n engl j med 352;1 www.nejm.org january 6 , 2005

n engl j med 352;1 www.nejm.org january 6, 2005 17

18 n engl j med 352;1 www.nejm.org january 6 , 2005

n engl j med 352;1 www.nejm.org january 6, 2005 19

You might also like