Prevalence of Autism

Spectrum Disorder in Preterm

Infants: A Meta-analysis
Sachin Agrawal, FRACP,​a Shripada C. Rao, FRACP,​a,​b Max K. Bulsara, PhD,​c Sanjay K. Patole, DrPHa,​b

CONTEXT: Evidence is emerging that preterm infants are at risk for autism spectrum disorder
abstract
(ASD).
OBJECTIVES: To conduct a systematic review and meta-analysis to estimate the prevalence of
ASD in preterm infants.
DATA SOURCES: Medline (via PubMed and Ovid), Embase, PsycINFO, and relevant conference
proceedings were searched in May 2017.
STUDY SELECTION: Original studies in which researchers report on the prevalence of ASD using
diagnostic tests in children born preterm were included. Studies in which researchers used
only ASD screening tools were excluded.
DATA EXTRACTION: Relevant data were extracted independently by 3 authors.

RESULTS: Researchers in a total of 18 studies (3366 preterm infants) used ASD diagnostic
tools. The median gestation, birth weight, and age at assessment were 28.0 weeks (range:
25.1–31.3 weeks), 1055 g (range: 719–1565 g), and 5.7 years (range: 1.5–21 years),
respectively. Meta-analysis revealed that the overall prevalence rate for ASD was 7%
(95% confidence interval: 4% to 9%). The funnel plot and Egger’s test revealed that there
was probably no evidence of publication bias.
LIMITATIONS: The limitations were significant heterogeneity and a lack of studies from middle-
and low-income countries.
CONCLUSIONS: The prevalence of ASD is significantly high in the preterm population. Adequate
resources are needed to improve the outcomes of these children.

aNeonatal Unit, King Edward Memorial Hospital for Women and Perth Children Hospital, Nedlands, Western Australia, Australia; bSchool of Medicine, University of Western Australia,
Crawley, Australia; and cDepartment of Biostatistics, Institute for Health Research, University of Notre Dame, Western Australia, Australia

Drs Rao and Agrawal proposed the hypothesis and idea for the systematic review and literature search, reviewed studies for inclusion, extracted and checked the
data, wrote the first draft of the manuscript, and contributed to the development and analysis plan; Dr Patole reviewed studies for inclusion, interpreted the results,
reviewed and edited the manuscript, and contributed to the development and analysis plan; Dr Bulsara conducted statistical analyses, interpreted the results, edited
the manuscript, and contributed to the development and analysis plan; and all authors approved the final manuscript as submitted.
The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or
approval of the manuscript; or decision to submit the manuscript for publication.
DOI: https://​doi.​org/​10.​1542/​peds.​2018-​0134
Accepted for publication Jun 13, 2018
Address correspondence to Shripada C. Rao, FRACP, Perth Children’s Hospital, Hospital Ave, Nedlands, WA 6009, Australia. E-mail: shripada.rao@health.wa.gov.au
PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275).

To cite: Agrawal S, Rao SC, Bulsara MK, et al. Prevalence of Autism Spectrum Disorder in Preterm Infants: A Meta-analysis. Pediatrics. 2018;142(3):e20180134

Downloaded from www.aappublications.org/news by guest on November 22, 2018

PEDIATRICS Volume 142, number 3, September 2018:e20180134 REVIEW ARTICLE
Advances in neonatal intensive by intellectual disability or global of prevalence were also included.
care have improved the survival of developmental delay.‍13 ASD is Studies in which researchers
preterm infants.‍1–‍‍ 4‍ It is well known associated with substantial lifetime described the prevalence of ASD in
that very preterm and extremely costs to individuals, their families, the general population were included
preterm infants carry a high risk and the community.‍14 A recent if specific information on the preterm
of long-term neurodevelopmental study revealed that the median population was available (ie, the
morbidities.4,​5‍ Recent literature family cost of ASD was $34 900 number of preterm infants with ASD,
reveals that even moderate- to late- (Australian dollars) per annum. For total number of preterm infants
preterm infants are also vulnerable each additional symptom reported, tested, and tool that was used). The
for such adverse outcomes.‍6 Of an extra $1400 per family per annum studies could have been published in
late, evidence is emerging that was incurred. Delay in diagnosis was peer-reviewed journals or presented
prematurity and being of low birth associated with a modest increase as conference abstracts. The
weight are risk factors for later in the number of ASD symptoms, diagnosis of ASD should have been
development of autism spectrum which indirectly impacts the cost based on diagnostic tests rather than
disorder (ASD).‍7–9 ‍ The prevailing of ASD.‍14 Researchers in multiple screening tools. Preterm infants were
consensus favors a multifactorial narrative reviews have discussed defined as neonates born before 37
pathogenesis for ASD. It is thought the risk factors and prevalence of weeks’ gestation.‍18 Low birth weight
that ASD develops in individuals with ASD in the preterm population‍15,​16; was defined as birth weight <2500 g.
an underlying biologic vulnerability however, there are no meta-analyses
who experience varying degrees or systematic reviews. We aimed Exclusion Criteria
of exogenous stressors during a to conduct a systematic review
Studies in which information
critical period of brain development and meta-analysis to describe the
regarding the number of preterm
during intrauterine and immediate prevalence of ASD in preterm infants
infants with ASD, total number of
postnatal life.‍10 It is also thought and analyze the relation between
preterm infants tested, and tools
that whatever initiates the preterm gestational age, birth weight, and
used to diagnose ASD were not
birth process might also initiate prevalence of ASD.
available were excluded. Studies
abnormal pathways of brain
in which researchers used only
development, thereby creating a
METHODS screening tools to diagnose ASD were
“perfect storm” of interactions
excluded. Multiple publications from
among genetic background and Guidelines from the Joanna Biggs the same cohort of preterm infants
environmental exposures that Institute were followed for conducting were included only if any additional
initiates a fetal immune response and reporting this systematic review.‍17 information was available; otherwise,
and leads to both preterm birth and Ethics approval was not required. they were considered to be duplicates,
altered development of the brain
and information was used only once.
and neural connectivity.‍11 Injury to Review Questions
Animal studies, letters to the editor,
the cerebellum in preterm infants
Review questions included the reviews, case reports, and series were
is more common than previously
following: excluded but read in detail to identify
thought and is known to be associated
potentially eligible studies.
with an increased risk of ASD.‍12 1. What is the prevalence of ASD in
preterm infants when diagnostic
Search Strategy
According to the Diagnostic and tools are used?
Statistical Manual of Mental Disorders, 2. Is there a relation between The databases PubMed, Embase,
Fifth Edition (DSM-V), the criteria gestational age, birth weight, and Medline (via Ovid), and PsycINFO
for ASD are as follows: (1) persistent prevalence of ASD? were searched from inception until
deficits in social communication May 2017. Abstracts of conference
and social interactions in multiple 3. Does the prevalence of ASD proceedings (such as the Perinatal
contexts; (2) restricted, repeated decrease if children with a Society of Australia and New Zealand,
patterns of behavior, interest, or disability are excluded? the European Academy of Pediatric
activities; (3) symptoms must be Societies, and the British Maternal and
Inclusion Criteria
present in the early developmental Fetal Medicine Society) were searched
period; (4) symptoms cause Original studies in which researchers in Embase. Electronic abstracts
significant impairment in social, report on the prevalence of ASD in from the Pediatric Academic Society
occupational, and other current children who were born preterm and other conference proceedings
areas of functioning; and (5) were included. Studies in which were also searched. The reference
disturbances are not explained data allowed for the calculation lists of all included full-text articles

Downloaded from www.aappublications.org/news by guest on November 22, 2018

2 AGRAWAL et al
were searched to identify any birth, study population, mean age at 95% CIs were reported. In addition,
studies missed in the initial search. diagnosis, exclusion of children with the proportions with their 95% CI
Reviewers (S.R. and S.A.) conducted congenital anomalies and syndromes, values from individual studies were
the literature search independently. and exclusion of children with also presented in a forest plot. Meta-
No language restriction was applied. major disability), method used for regression analysis was conducted
We searched Medline via Ovid using diagnosing ASD, number of children to find the relationship between the
the following terminology: [Preterm (former preterm infants) evaluated, prevalence of ASD, gestational age,
infant.mp. or exp Infant, Premature/ number of children (former preterm and birth weight. The results of the
OR low birth weight infant.mp. or infants) who tested positive on meta-regression analysis are given as
exp Infant, Low Birth Weight/ OR diagnostic tests, and prevalence of regression coefficients with 95% CIs.
exp Infant, Very Low Birth Weight/ ASD based on diagnostic tests. Funnel plots were used for assessing
or exp Infant, Premature/ or very publication bias. Eggers’s test was
Risk of bias was assessed by
low birth infant.mp. or exp Infant, used as a formal test of funnel plot
using the quality assessment tool
Low Birth Weight/ or exp Infant, asymmetry and publication bias.‍20
for prevalence studies from the
Premature, Diseases/] AND [autism. For studies in which researchers
Joanna Briggs Institute‍19 by using
mp. or exp Autistic Disorder/ OR exp reported only the median, range, or
the following criteria: (1) Was the
Asperger Syndrome/ or exp Autistic interquartile range (IQR), we used
sample frame appropriate to address
Disorder/ or exp Child Development the method of Wan et al‍21 to estimate
the target population? (2) Was the
Disorders, Pervasive/ or pervasive the sample means. All statistical
study population sampled in an
developmental disorder.mp. or exp analyses were conducted by using
appropriate way? (3) Was the sample
Developmental Disabilities/]. The Stata 15 SE (Stata Corp, College
size adequate? (4) Were the study
other databases were searched Station, TX). We planned 2 subgroup
subjects and settings described in
with similar terms. References analyses. The first was based on the
detail? (5) Was the data analysis
were compiled and managed by presence of disability, and the second
conducted with sufficient coverage of
using EndNote 8 (Thomson Reuters, was based on the age at assessment
the identified sample? (6) Were valid
Toronto, Canada), with duplicate (<3 years versus ≥3 years). For the
methods used for identification of
citations being removed by using this first subgroup analyses, the presence
the condition? (7) Was the condition
software. of either a genetic syndrome or
measured in a standard, reliable way
for all the participants? (8) Was there neurodevelopmental disability or
Study Selection
appropriate statistical analysis? both was considered a disability.
Two reviewers (S.R. and S.A.) Current opinion is that early
(9) Was the response rate adequate,
screened the titles and abstracts of diagnosis and intervention, especially
and if not, was the low response rate
the potentially relevant studies. At at <3 years of age, may improve the
managed appropriately? The criteria
this stage, the search was purposely outcomes of children with autism.
were rated as either yes, no, not clear,
broad to allow for the inclusion This is supported by the results of
or not applicable. When necessary,
of all potentially relevant studies. a recently published randomized
authors of the included studies were
All studies that met the inclusion controlled trial, in which children
contacted, requesting additional
criteria at this stage were reviewed (2–4 years old) who were randomly
information from their studies.
by the third author (S.K.P.) to ensure assigned to receive early intervention
appropriateness for inclusion in Statistical Analysis showed significant improvements
the final analysis. Disagreements compared with those who received
of eligibility were reconciled by We used the absolute number of usual care.‍22,​23
‍ Hence, we chose a
discussion among 3 authors. observed events and calculated the cutoff of 3 years of age for the second
proportions and 95% confidence preplanned subgroup analysis.
Data Abstraction and Risk-of-Bias intervals (CIs), assuming a binomial
Assessment distribution. A logistic normal
Tools Used to Diagnose ASD in the
A standardized data abstraction random effects model was fitted. Included Studies
form was used to enter the study- For CIs for the pooled estimate,
related variables. Two independent we used a variance stabilizing The tools used to diagnose ASD
reviewers (S.A. and S.C.R.) abstracted Freeman–Tukey transformation. included the following:
the data from each included study. Heterogeneity was assessed by 1. The Autism Diagnostic
The following data were abstracted: using the standard χ2 test and the I2 Observational Schedule
study identification (ie, author, statistic. When data could be pooled, (ADOS)‍24–‍ 26
‍ is a semistructured,
journal, country, and year), study meta-analysis was conducted, and standardized assessment designed
characteristics (participant year of relevant summary statistics and for use with individuals who

Downloaded from www.aappublications.org/news by guest on November 22, 2018

PEDIATRICS Volume 142, number 3, September 2018 3
 are referred for possible ASD. hyperactivity symptoms and developmental pediatricians,
A protocol of activities or social any resultant impairment. The and child neurologists are
presses is administered in ∼45 different sorts of information are the specialists who can make
minutes, and then items are brought together by a computer comprehensive assessments to
scored on a 4-point scale, with program that also predicts likely diagnose ASD.‍37
0 indicating “no abnormality of diagnoses. These computer-
type specified” and 3 indicating generated summary sheets and
“moderate to severe abnormality.” diagnoses form a convenient RESULTS
It is designed for use in the starting point for experienced A total of 18 studies in which
diagnostic evaluation of children clinicians (child psychiatrists researchers reported the prevalence
(12 months to 16 years old) who and psychologists), who decide of ASD were included.‍38–‍‍‍‍‍‍‍‍‍‍‍‍‍ 55
‍ ‍Figure 1
are referred for suspected ASD; whether to accept or overturn provides details of the study selection
the computer diagnosis (or lack process. Three authors responded
2. The Autism Diagnostic Interview–
of diagnosis) in light of their and provided additional information
Revised (ADI-R),​‍27,​28
‍ a revision
review of all the data, including for this systematic review.41,​44,​ ‍ 46
‍
of the Autism Diagnostic
transcripts. Finally, a diagnosis
Interview, is a semistructured, The total sample size was 3366
is allocated in accordance with
investigator-based interview for participants from 18 studies (‍Table 1).
DSM-IV, DSM-V, or ICD-10 criteria.
caregivers of children and adults The median sample size in the
Multiple studies have revealed
for whom autism or pervasive included studies was 98 (IQR
excellent agreement between the
developmental disorders 61–177; range: 53–857). The median
DAWBA and the ADOS or ADI-R
are a possible diagnosis. It is gestational age was 28.0 weeks (IQR:
for the diagnosis of ASD‍30 and
appropriate for children with 27.2–30.3 weeks; range 25.1–31.3
their use in clinical settings‍31,​32
‍ ;
mental ages from ∼18 months weeks). The median birth weight
into adulthood and is linked to 4. The Childhood Autism Rating was 1055 g (IQR: 965–1201 g; range
International Classification of Scale (CARS)‍33–‍‍ 36
‍ is used from the 719–1565 g). The median age at
Diseases, 10th Revision (ICD-10) age of 2 years to late adulthood. assessment was 5.7 years (IQR: 3.7–
and Diagnostic and Statistical It is a clinical rating scale for the 11 years; range: 1.5–21 years). The
Manual of Mental Disorders, Fourth trained clinician to rate items median number of preterm infants
Edition (DSM-IV) criteria; that are indicative of ASD after diagnosed with ASD in individual
direct observation of the patient studies was 8 (IQR 3–16; range: 1–61).
3. The Development and Wellbeing
and to determine the symptom
Assessment (DAWBA)‍29 is a Meta-analysis
severity through quantifiable
novel package of questionnaires,
ratings. It is used to evaluate The overall prevalence rate of ASD
interviews, and rating techniques
the following functional areas: was 7% (95% CI: 4% to 9%; ‍Fig 2);
designed to generate ICD-10 and
relating to people; body use; visual there was significant statistical
DSM-IV psychiatric diagnoses
response; listening response; heterogeneity (I2 = 84.82%; ‍Fig 2).
on 5- to 16-year-old patients.
taste, smell, and touch response;
Interviewers administer a
use of verbal communication and Subgroup Analysis
structured interview to parents
nonverbal communication; and
about psychiatric symptoms and Subgroup analysis of studies in
level and consistency of intellectual
their resultant impact. When which children with disabilities were
response. The revised version is
definite symptoms are identified excluded revealed a prevalence rate of
effective in discriminating between
with the structured questions, ‍ –‍ 47,​
9% (95% CI: 6% to 11%).‍39,​45 ‍ 50,​51,​
‍ 55‍
children with autism and those
interviewers use open-ended Subgroup analysis of studies in which
with severe cognitive deficits and
questions and supplementary children with disabilities were not
distinguishing mild-to-moderate
prompts to get parents to describe excluded revealed a prevalence
from severe autism; and
the problems in their own words. rate of 6% (95% CI: 2% to 10%).
These descriptions are transcribed 5. Team assessment (psychologist, Subgroup analysis of studies in
verbatim by the interviewers but psychiatrist, and developmental which children were assessed at <3
are not rated by them. A similar pediatrician) according to years revealed a prevalence rate of
interview is administered to DSM-IV or DSM-V criteria for 7% (95% CI: 4% to 9%). Subgroup
11- to 16-year-old patients. In diagnosis of ASD is used. The analysis of studies in which children
addition, teachers complete a Centers for Disease Control and were assessed at ≥3 years revealed
brief questionnaire covering the Prevention recommend that a prevalence rate of 7% (95% CI: 4%
main conduct, emotional, and child psychiatrists, psychologists, to 10%).

Downloaded from www.aappublications.org/news by guest on November 22, 2018

4 AGRAWAL et al
 Risk of Bias
The risk of bias was low in the
majority of the domains in the
included studies (‍Table 2).

DISCUSSION
Our systematic review that included
3366 preterm infants from 18 studies
revealed that when diagnostic tools
were used, the overall prevalence rate
of ASD was 7% (95% CI: 4% to 9%).
This equates to ∼900 000 additional
children each year who will develop
ASD given that globally, ∼15 million
infants are born preterm (before 37
weeks’ gestation), of whom 13 million
survive.‍57–‍ 59
‍ Our findings confirm
that the prevalence of ASD in preterm
infants is considerably higher than in
the general population, in which the
overall prevalence has been reported
to be 0.76%.‍60 Researchers in another
study reported a prevalence of 1.46%
(range: 0.57%–2.19%) in the general
pediatric population aged 8 years.61
Although there are many tests used
to screen children for ASD,​‍62 it is
important not to label children who
test positive on screening tests as
having ASD. Such children should
always be assessed with definitive
tests before confirming them as
having ASD. Expert committees (the
Child Development and Behavior
FIGURE 1 Special Interest Group of the Chapter
Flow diagram of the study selection. of Community Child Health, the
Pediatrics and Child Health Division,
Sensitivity Analysis Publication Bias and the Royal Australasian College
The results of the study by Verhaeghe The funnel plot and Egger’s test of Physicians) have advised to use
et al‍42 appeared to be an outlier; revealed that there was probably no clinical judgement when suspecting
hence, sensitivity analysis was publication bias (P = .294; ‍Fig 3). a child as having ASD and to not rely
conducted by excluding it. The solely on screening tools.‍63 A recent
prevalence rate for ASD was then 6% Meta-regression survey of health professionals in
(95% CI: 4% to 8%), and statistical Meta-regression analysis Australia revealed inconsistencies
heterogeneity decreased from 85% revealed no significant association in ASD assessment practices across
to 74%. Nine of 53 (17%) children in between gestational age the states and between the private
that study had significant cerebellar (regression coefficient: −0.0075; and public service settings.‍64 Only
hemorrhage in the neonatal period, 95% CI: −0.0234 to 0.0083; half of the respondents reported
which could have contributed to the P = .315; ‍Fig 4) or birth weight that they include a standardized
high prevalence of ASD; it is well (regression coefficient: −0.0001; objective assessment tool, such as the
known that cerebellar hemorrhage 95% CI: −0.0003 to 0.00001; ADOS, in ASD assessments. The study
and anomalies are associated with an P = .068; ‍Fig 5) and the prevalence raised concerns that clinicians may
increased risk of ASD.‍12,​56
‍ of ASD. not be practicing in a manner that is

Downloaded from www.aappublications.org/news by guest on November 22, 2018

PEDIATRICS Volume 142, number 3, September 2018 5
6
TABLE 1 Characteristics of Included Studies
Serial Study, Year of Birth Study Age at Exclusion of Exclusion Diagnostic Tool Used No. Children No. Former Prevalence
No. Location Population Assessment Children With of Children (Former Preterm Infants Based on
Congenital Anomaly With Major Preterm Diagnosed With Diagnostic
or Syndromes Disability Infants) ASD Tool, %
Evaluated
1 Indredavik 1986–1988 BW <1500 g 14 y (mean Yes No Interview by 2 psychiatrists and 56 1 2
et al,​‍38 Mean BW 1174 g; 14.1 y; SD conclusions drawn according
Norway SD 233 g 0.3 y) to DSM-IV criteria
Mean GA 28.8
wk; SD 2.7 wk
2 Pinto-Martin 1984–1989 BW <2000 g 21 y Not clear No ADI-R, ADOS 623 14 2.2
et al,​‍39 Mean BW 1475.3
United g; SD 353.3 g
States Mean GA 31.2
wk; SD 3.1 wk
3 Johnson et al,​‍48 1995 GA <26 wk Median 10.9 y Yes No The DAWBA was administered to 219 16 8
United parents via telephone interview
Kingdom or online version, then
and Ireland summary sheets and clinical
transcripts of the interview
were reviewed by 2 child and
adolescent psychiatrists, who
assigned DSM-IV and ICD-10
consensus diagnoses.
4 Treyvaud 2001–2003 GA <30 wk 7y Not clear No The DAWBA was completed online 177 8 4.5
et al,​‍49 (mean GA by parents. Raters were 2
Australia 27.5 wk; SD clinical psychologists with
1.94 wk); diagnostic experience who
BW <1250 g completed the online DAWBA
(mean BW rater training and assigned
975 g; SD diagnoses according to DSM-IV
223 g) criteria.
5 Yaari et al,​‍45 2009–2013 GA 24–34 wk 18 mo Yes Yes ADOS Toddler 99 8 8
Israel Mean GA 31.28
wk (SD 2.57
wk); mean

Downloaded from www.aappublications.org/news by guest on November 22, 2018

BW 1541.38 g
(SD 474.32 g;
range:
490–2400 g)
6 Mettelman NA GA ≤30 wk 4y NA NA CARS administered by 174 17 10
et al,​‍50a psychologist while examining
the children

AGRAWAL et al
 TABLE 1 Continued
Serial Study, Year of Birth Study Age at Exclusion of Exclusion Diagnostic Tool Used No. Children No. Former Prevalence
No. Location Population Assessment Children With of Children (Former Preterm Infants Based on
Congenital Anomaly With Major Preterm Diagnosed With Diagnostic
or Syndromes Disability Infants) ASD Tool, %
Evaluated
7 Dudova et al,​‍40 2012–2013 BW <1500 g 2y Not clear Yes ADOS and clinical evaluation by 2 75 8 10.6
Czech Mean BW 1033.1 g; psychiatrists
Republic SD 293.1 g
Mean GA 28.4
wk; SD 2.8 wk
8 Joo et al,​‍51 NA GA <36 wk 3–6 y (mean Not clear Yes CARS Korean version 58 1 1.7
South (mean GA 3.7 y; SD
Korea 30.6 wk; SD 0.7 y)
3.2 wk); mean

PEDIATRICS Volume 142, number 3, September 2018

BW 1565 g
(SD 603.9 g)
9 Ochiai et al,​‍41 2001–2005 BW <1500 6–9 y Yes Not clear Clinical evaluation by psychiatrists 77 4 5.1
Japan (median
1034 g; range
484–1494 g);
mean BW
1055 g (SD
296 g)
Mean GA 29 wk
6 d; SD 24.6 d
10 Kihara and 2001–2005 Median BW 3–6 y Not clear No Clinical assessment by pediatric 321 35 10.9
Nakamura,​‍52 951 g; IQR neurologist using DSM-IV
Japan 726–1219 g criteria
Median GA
27.4 wk; IQR
25.1–30.5 wk
11 Verhaeghe 1999–2000 GA <27 wk 11–15 y Not clear No ADOS, ADI-R 53 21 40
et al,​‍42 BW 400–1200 g (mean 12.6
Belgium (mean BW y; SD 1.03 y)
791.75 g; SD
179.08 g)

Downloaded from www.aappublications.org/news by guest on November 22, 2018

12 Gray et al,​‍43 2007–2009 GA <30 wk; 2y Yes No Clinical by developmental 97 1 1
Australia mean GA 27.6
wk; SD 2 wk
Mean BW 1072 g; Pediatrician according to DSM-IV
SD 326 g criteria
13 Yang et al,​‍53 1996–1999 Mean BW 1200.9 g; 13.4 y Not clear No Diagnoses of ASD were based on 61 2 3.3
Taiwan SD 218.8 g combined information from
diagnostic tools, observations
of the subjects, and parental
report and labeling according
to DSM-IV Text Revision criteria.

7
8
TABLE 1 Continued
Serial Study, Year of Birth Study Age at Exclusion of Exclusion Diagnostic Tool Used No. Children No. Former Prevalence
No. Location Population Assessment Children With of Children (Former Preterm Infants Based on
Congenital Anomaly With Major Preterm Diagnosed With Diagnostic
or Syndromes Disability Infants) ASD Tool, %
Evaluated
14 Mohammed 2012–2013 BW 620–1220 g 3y Not clear Not clear Clinical assessment in the 107 5 4.7
et al,​‍54 GA 27–33 wk comprehensive autism
Saudi program included a
Arabia multidisciplinary team of
psychologists, behavioral
pediatricians, and speech
therapists; ASD diagnosis was
based on DSM-IV criteria.
15 Ikejri et al,​‍55 2004–2007 GA <33 wk 30–74 mo Not clear Yes Clinical assessment by 2 59 9 15.2
Japan (mean BW (mean 49.1 pediatricians using DSM-IV Text
857 g; SD mo; SD 15.7 Revision criteria
367.6 g); mo)
mean GA 27
wk (SD 2.2
wk; range
474–1448 g)
16 Joseph et al,​‍46 2002–2004 GA <28 wk 10 y No Yes ADI-R, ADOS 2 857 61 7.1
United
States
17 Pritchard 2006–2008 GA <29 wk 2–4 y Not clear No ADOS Generic 169 3 1.8
et al,​‍44 Mean GA 26.7
Australia wk; SD 1.4 wk
18 Padilla et al,​‍47 2004–2007 GA <27 wk 6.5 y Yes Yes Clinical assessment using DSM-IV 84 10 11.9
Sweden (mean 25.1 and ICD-10 criteria
wk; SD 1.1
wk)
Mean BW
718.8 g; SD
148.5 g
ADOS 2, Autism Diagnostic Observation Schedule, Second Edition; BW, birth wt; GA, gestational age; NA, not available.
a Conference abstracts only.

Downloaded from www.aappublications.org/news by guest on November 22, 2018

AGRAWAL et al
 2 years of age, routine screening
for ASD in this highly vulnerable
population is not yet a standard
practice even in high-income
countries. Given the shortage of
experts in diagnosing and managing
ASD, innovative strategies will be
required within existing health
systems. These could include
the training of neonatologists,
pediatricians, and developmental
psychologists in administering the
screening tools at routine follow-up
visits. Children who test positive on
screening tools and those in whom
the diagnosis is suspected on the
basis of clinical judgement could be
referred for further assessment by
experts using definitive diagnostic
tools.
The median age at assessment in
our review was 5.7 years, which is
FIGURE 2 considered to be late. Early diagnosis
Forest plot used to assess the prevalence of ASD. ES, effect size. and intervention have the potential
to limit the severity of symptoms and
improve the outcomes of children
with ASD.‍22,​65,​
‍ 66‍
Children with ASD require the
services of multiple specialists,
including pediatricians, general
practitioners, psychologists,
psychiatrists, occupational therapists,
behavioral therapists, nutritionists,
physiotherapists, audiologists, and
many more. Funding agencies should
provide adequate finances to support
these services.
Our meta-regression analysis
revealed no significant association
between gestational age, birth
weight, and prevalence of ASD
in preterm infants. The probable
reason for not reaching statistical
FIGURE 3 significance may be the small sample
Diagnostic funnel plot with pseudo–95% confidence limits for publication bias. ES, effect size.
size. Contrary to our expectations,
we found no significant difference in
consistent with international best- The large proportion of preterm
the prevalence of ASD when infants
practice guidelines or statements infants who subsequently develop
with disabilities were included or
from professional bodies.‍64 The use ASD necessitates an adequate
excluded, but the sample size was
of nonstandardized methods causes allocation of resources to facilitate
small to make firm conclusions on
difficulty in calculating the true diagnosis and rehabilitation.
this.
prevalence of ASD and hence makes Although there are well-established
it difficult to allocate appropriate neurodevelopmental programs to An important observation in our
resources from funding agencies. manage preterm infants until at least review was the total absence of

Downloaded from www.aappublications.org/news by guest on November 22, 2018

PEDIATRICS Volume 142, number 3, September 2018 9
 ASD. In a review of ASD studies
from 9 LMICs, Samms-Vaughan‍69
identified a lack of resources as the
main barrier to early diagnosis. In
a comprehensive review of autism-
related research from Sub-Saharan
Africa, Franz et al70 found 53
publications, but no epidemiologic
or early intervention studies were
identified. Durkin et al‍71 suggested
that barriers in LMICs include the
high cost of proprietary tools for
diagnosing ASD and the high cost
of training professionals to use the
tools. They recommended open-
source and open-access models to
facilitate global collaboration and
training.‍71
FIGURE 4 Researchers in a recent systematic
Correlation between gestational age and prevalence of ASD. review identified 28 studies from
LMICs in which researchers reported
on 18 different screening tools that
were used to assess children for
ASD. None of the included studies
in that review had a specific focus
on preterm infants.‍72 Given that
∼90% of preterm deliveries occur
in LMICs‍57 and the survival of
preterm infants in those countries is
increasing rapidly,​‍73–75
‍ the number
of children who will need assessment
and management of ASD are expected
to increase in the coming years.
Hence, the implementation of optimal
strategies in LMICs is urgently
needed.

FIGURE 5 The limitations and strengths of

Correlation between birth weight and prevalence of ASD. our systematic review need to be
discussed. (1) There was significant
ASD studies of preterm infants from by their pediatricians that “children clinical and statistical heterogeneity.
low- and middle-income countries should not be compared to each The heterogeneity was mainly
(LMICs). Because of various barriers, other” and that “boys have a slower related to the age at assessment
children with ASD often do not development rate or are more (which ranged from 1.5 to 21
receive a confirmatory diagnosis agitated than girls.” In addition, years), the type of diagnostic tool
until preschool age, especially in they reported that most mothers used, and the sample size (which
LMICs. In a Brazilian survey, Ribeiro described their interactions with the ranged from 53 to 857). Because
et al‍67 found that most mothers pediatricians as negative experiences heterogeneity was anticipated, we
sought help within 3 months of and felt discouraged from expressing used a random-effects model‍76 for
first suspicion and mentioned their their concerns again.‍66 In a study the meta-analysis. The high level of
concerns to pediatricians, but only conducted in Vietnam, Van Cong statistical heterogeneity was due to
one-third of such children were et al‍68 found that limited resources, the varied prevalence of ASD in the
referred for further evaluation of a lack of awareness, and a lack of included studies. Even after removing
ASD. The authors reported that in official government policy were an extreme outlier, the heterogeneity
many instances, mothers were told barriers to the early diagnosis of remained high; hence, caution is

Downloaded from www.aappublications.org/news by guest on November 22, 2018

10 AGRAWAL et al
 TABLE 2 Risk of Bias in the Included Studies
Studies Was the Was the Study Was the Sample Were the Study Was the Data Were the Valid Was the Condition Was There Was the Response
Sample Frame Population Size Adequate? Subjects and Analysis Methods Used for Measured in a Appropriate Rate Adequate,
Appropriate Sampled in an Settings Described Conducted Identification of Standard, Reliable Statistical and If Not, Was
to Address the Appropriate Way? in Detail? With Sufficient the Condition? Way for All the Analysis? the Low Response
Target Population? Coverage of the Participants? Rate Managed
Identified Sample? Appropriately?
Indredavik et al‍38 Yes Yes No Yes Yes Yes Yes Yes Yes
Pinto Martin et Yes Yes Yes Yes Yes Yes Yes Yes Yes
al‍39
Johnson et al‍48 Yes Yes Yes Yes Yes Yes Yes Yes Yes
Treyvaud et al‍49 Yes Yes Yes Yes Yes Yes Yes Yes Yes
Yaari et al‍45 Yes Yes Yes Yes Yes Yes Yes Yes Yes
Mettelman et al‍50a Not available Not available Not available Not available Not available Not available Not available Not available Not available

PEDIATRICS Volume 142, number 3, September 2018

Dudova et al‍40 Yes Yes No Yes Yes Yes Yes Yes Yes
Joo et al‍51 Yes Yes No Yes No Yes Yes Yes Yes
Ochiai et al‍41 Yes Yes Yes Yes Yes Yes Yes Yes Yes
Kihara and Yes Yes Yes Yes Yes Yes Yes Yes Yes
Nakamuar‍52
Verhaeghe et al‍42 Yes Yes No Yes Yes Yes Yes Yes Yes
Gray et al‍43 Yes Yes Yes Yes Yes Yes Yes Yes Yes
Yang et al‍53 Yes Yes No Yes No Yes Yes Yes No
Mohammed et al‍54 Yes Yes Yes Yes Yes Yes Yes Yes Not clear
Ikejiri et al‍55 Yes Yes No Yes Yes Yes Yes Yes Yes
Joseph et al‍46 Yes Yes Yes Yes Yes Yes Yes Yes Yes
Pritchard et al‍44 Yes Yes Yes Yes Yes Yes Yes Yes Yes
Padilla et al‍47 Yes Yes No Yes Yes Yes Yes Yes Yes
a Conference-only abstract.

children.

Downloaded from www.aappublications.org/news by guest on November 22, 2018

CARS: C
ADI-R: A

DSM-V: D
DSM-IV: D
CONCLUSIONS

Scale

country
ABBREVIATIONS

Edition
preterm population.

ICD-10: I nternational

10th Revision
CI: c onfidence interval

 iagnostic and
Given that ASD results in

IQR: interquartile range

ADOS: Autism Diagnostic

Manual of Mental
 utism Diagnostic
Interview–Revised
an enormous burden on the

diagnosis and treatment are

DAWBA: Development and

are the rigorous methodology,

high in the preterm population.

Adequate resources are needed

Statistical Manual of
Observational Schedule

Disorders, Fifth Edition

patients, caregivers, health care

to improve the outcomes of these

LMIC: low- and middle-income

ASD: a utism spectrum disorder
high-income countries, the results

The major strengths of our review

Wellbeing Assessment
 hildhood Autism Rating
urgently needed, especially in the

Classification of Diseases,
 iagnostic and Statistical
the first systematic review in which
of this review are not generalizable.

Mental Disorders, Fourth

researchers address the issue of the
reasonably large sample size for the

of publication bias. Ours is probably

‍
primary analysis, and likely absence

strategies to implement a definitive

(2) Because all the studies were from

prevalence of ASD in preterm infants.

professionals, and communities,​‍77–‍ 79

The prevalence of ASD is significantly

needed when interpreting the results.

11
Copyright © 2018 by the American Academy of Pediatrics
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Dr Agrawal received grants from the Telethon Foundation of Western Australia to conduct research evaluating neurodevelopmental outcomes of
preterm infants receiving probiotic supplementation. This systematic review was part of his Telethon fellowship training.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

REFERENCES
1. Ancel PY, Goffinet F, Kuhn P, et al; born preterm-role of exposure 18. Goldenberg RL, Culhane JF, Iams
EPIPAGE-2 Writing Group. Survival and to perinatal inflammation. Front JD, Romero R. Epidemiology and
morbidity of preterm children born Neurosci. 2013;7:123 causes of preterm birth. Lancet.
at 22 through 34 weeks’ gestation 9. Wang C, Geng H, Liu W, Zhang G. Prenatal, 2008;371(9606):75–84
in France in 2011: results of the perinatal, and postnatal factors 19. Munn Z, Moola S, Lisy K, Riitano
EPIPAGE-2 cohort study. JAMA Pediatr. associated with autism: a meta-analysis. D, Tufanaru C. Methodological
2015;169(3):230–238 Medicine (Baltimore). 2017;96(18):e6696 guidance for systematic reviews of
2. Anderson JG, Baer RJ, Partridge JC, observational epidemiological studies
10. Limperopoulos C. Autism spectrum
et al. Survival and major morbidity reporting prevalence and cumulative
disorders in survivors of extreme
of extremely preterm infants: a incidence data. Int J Evid-Based
prematurity. Clin Perinatol.
population-based study. Pediatrics. Healthc. 2015;13(3):147–153
2009;36(4):791–805, vi
2016;138(1):e20154434
11. Erdei C, Dammann O. The perfect storm: 20. Egger M, Davey Smith G, Schneider
3. Santhakumaran S, Statnikov Y, Gray M, Minder C. Bias in meta-analysis
preterm birth, neurodevelopmental
D, Battersby C, Ashby D, Modi N; detected by a simple, graphical test.
mechanisms, and autism causation.
Medicines for Neonates Investigator BMJ. 1997;315(7109):629–634
Perspect Biol Med. 2014;57(4):470–481
Group. Survival of very preterm infants
12. Limperopoulos C. Extreme prematurity, 21. Wan X, Wang W, Liu J, Tong T. Estimating
admitted to neonatal care in England
cerebellar injury, and autism. Semin the sample mean and standard
2008-2014: time trends and regional
Pediatr Neurol. 2010;17(1):25–29 deviation from the sample size,
variation. Arch Dis Child Fetal Neonatal
median, range and/or interquartile
Ed. 2018;103(3):F208–F215 13. American Psychiatric Association. range. BMC Med Res Methodol.
4. Sharp M, French N, McMichael Diagnostic and Statistical Manual of 2014;14:135
J, Campbell C. Survival and Mental Disorders (DSM-5). Washington,
neurodevelopmental outcomes in DC: American Psychiatric Association; 22. Pickles A, Le Couteur A, Leadbitter
extremely preterm infants 22-24 2013 K, et al. Parent-mediated social
weeks of gestation born in Western communication therapy for
14. Horlin C, Falkmer M, Parsons R, young children with autism
Australia. J Paediatr Child Health. Albrecht MA, Falkmer T. The cost of
2018;54(2):188–193 (PACT): long-term follow-up of a
autism spectrum disorders. PLoS One. randomised controlled trial. Lancet.
5. Pierrat V, Marchand-Martin L, Arnaud 2014;9(9):e106552 2016;388(10059):2501–2509
C, et al; EPIPAGE-2 Writing Group. 15. Mahoney AD, Minter B, Burch K, Stapel-
Neurodevelopmental outcome at 2 23. Sigafoos J, Waddington H. 6
Wax J. Autism spectrum disorders
years for preterm children born at year follow-up supports early
and prematurity: a review across
22 to 34 weeks’ gestation in France autism intervention. Lancet.
gestational age subgroups. Adv
in 2011: EPIPAGE-2 cohort study. BMJ. 2016;388(10059):2454–2455
Neonatal Care. 2013;13(4):247–251
2017;358:j3448 24. Luyster R, Gotham K, Guthrie W, et al.
16. Ng M, de Montigny JG, Ofner M, Do MT.
6. Cheong JL, Doyle LW, Burnett The Autism Diagnostic Observation
Environmental factors associated with
AC, et al. Association between Schedule-toddler module: a new
autism spectrum disorder: a scoping
moderate and late preterm birth module of a standardized diagnostic
review for the years 2003-2013.
and neurodevelopment and social- measure for autism spectrum
Health Promot Chronic Dis Prev Can.
emotional development at age 2 years. disorders. J Autism Dev Disord.
2017;37(1):1–23
JAMA Pediatr. 2017;171(4):e164805 2009;39(9):1305–1320
17. Joanna Briggs Institute. The Joanna
7. Maramara LA, He W, Ming X. Pre- 25. Lord C, Rutter M, DiLavore PC, Risi
Briggs Institute reviewer’s manual
and perinatal risk factors for S, Gotham K, Bishop SL. Autism
2014: the systematic review of
autism spectrum disorder in a Diagnostic Observation Schedule
prevalence and incidence data.
New Jersey cohort. J Child Neurol. (ADOS): Manual. Torrance, CA: Western
Available at: http://​joannabriggs.​
2014;29(12):1645–1651 Psychological Services; 2001
org/​assets/​docs/​sumari/​
8. Meldrum SJ, Strunk T, Currie A, ReviewersManual_​The-​Systematic-​ 26. Lord C, Risi S, Lambrecht L, et al.
Prescott SL, Simmer K, Whitehouse AJ. Review-​of-​Prevalence-​and-​Incidence-​ The autism diagnostic observation
Autism spectrum disorder in children Data.​pdf. Accessed May 10, 2017 schedule-generic: a standard measure

Downloaded from www.aappublications.org/news by guest on November 22, 2018

12 AGRAWAL et al
 of social and communication deficits 36. Dawkins T, Meyer AT, Van Bourgondien 46. Joseph RM, O’Shea TM, Allred EN,
associated with the spectrum ME. The relationship between the et al. Prevalence and associated
of autism. J Autism Dev Disord. Childhood Autism Rating Scale: Second features of autism spectrum disorder
2000;30(3):205–223 Edition and clinical diagnosis utilizing in extremely low gestational age
the DSM-IV-TR and the DSM-5. J Autism newborns at age 10 years. Autism Res.
27. Rutter M, Le Couteur A, Lord C. Autism
Dev Disord. 2016;46(10):3361–3368 2017;10(2):224–232
Diagnostic Interview—Revised. Vol 29.
Los Angeles, CA: Western Psychological 37. Centers for Disease Control and 47. Padilla N, Eklöf E, Mårtensson GE, Bölte
Services; 2003:30 Prevention. Autism spectrum disorder S, Lagercrantz H, Ådén U. Poor brain
(ASD): screening and diagnosis. growth in extremely preterm neonates
28. Lord C, Rutter M, Le Couteur A. Autism
Available at: https://​www.​cdc.​gov/​ long before the onset of autism
Diagnostic Interview-Revised: a revised
ncbddd/​autism/​screening.​html . spectrum disorder symptoms. Cereb
version of a diagnostic interview for
Accessed November 20, 2017 Cortex. 2017;27(2):1245–1252
caregivers of individuals with possible
pervasive developmental disorders. 38. Indredavik MS, Vik T, Heyerdahl S, 48. Johnson S, Hollis C, Kochhar P,
J Autism Dev Disord. 1994;24(5):659–685 Kulseng S, Fayers P, Brubakk AM. Hennessy E, Wolke D, Marlow N.
Psychiatric symptoms and disorders Psychiatric disorders in extremely
29. Goodman R, Ford T, Richards H,
in adolescents with low birth weight. preterm children: longitudinal finding
Gatward R, Meltzer H. The Development
Arch Dis Child Fetal Neonatal Ed. at age 11 years in the EPICure study.
and Well-Being Assessment:
2004;89(5):F445–F450 J Am Acad Child Adolesc Psychiatry.
description and initial validation of an
2010;49(5):453–463.e1
integrated assessment of child and 39. Pinto-Martin JA, Levy SE, Feldman
adolescent psychopathology. J Child JF, Lorenz JM, Paneth N, Whitaker 49. Treyvaud K, Ure A, Doyle LW, et al.
Psychol Psychiatry. 2000;41(5): AH. Prevalence of autism spectrum Psychiatric outcomes at age seven
645–655 disorder in adolescents born for very preterm children: rates and
weighing <2000 grams. Pediatrics. predictors. J Child Psychol Psychiatry.
30. McEwen FS, Stewart CS, Colvert E,
2011;128(5):883–891 2013;54(7):772–779
et al. Diagnosing autism spectrum
disorder in community settings 40. Dudova I, Kasparova M, Markova D, et 50. Mettelman B, Bode M, Gross SJ. O-049
using the Development and Well- al. Screening for autism in preterm parent’s over identification of autism
Being Assessment: validation in a UK children with extremely low and very spectrum disorder in extremely
population-based twin sample. J Child low birth weight. Neuropsychiatr Dis premature infants at four years of
Psychol Psychiatry. 2016;57(2):161–170 Treat. 2014;10:277–282 age. Arch Dis Child. 2014;99(suppl
2):A40–A41
31. Aebi M, Kuhn C, Metzke CW, Stringaris 41. Ochiai M, Ichiyama M, Iwayama M,
A, Goodman R, Steinhausen HC. Sakai Y, Yoshida K, Hara T. Longitudinal 51. Joo JW, Choi JY, Rha DW, Kwak
The use of the development and study of very low birth weight EH, Park ES. Neuropsychological
well-being assessment (DAWBA) infants until 9years of age; attention outcomes of preterm birth in children
in clinical practice: a randomized deficit hyperactivity and autistic with no major neurodevelopmental
trial. Eur Child Adolesc Psychiatry. features are correlated with their impairments in early life. Ann Rehabil
2012;21(10):559–567 cognitive functions. Early Hum Dev. Med. 2015;39(5):676–685
2015;91(12):783–786
32. Fombonne E, Simmons H, Ford T, 52. Kihara H, Nakamura T. Early
Meltzer H, Goodman R. Prevalence of 42. Verhaeghe L, Dereu M, Warreyn P, De standard development assessment
pervasive developmental disorders Groote I, Vanhaesebrouck P, Roeyers characteristics in very low birth
in the British nationwide survey of H. Extremely preterm born children at weight infants later classified with
child mental health. Int Rev Psychiatry. very high risk for developing autism autism spectrum disorder. Early Hum
2003;15(1–2):158–165 spectrum disorder. Child Psychiatry Dev. 2015;91(6):357–359
Hum Dev. 2016;47(5):729–739
33. Schopler E, Reichler RJ, DeVellis RF, 53. Yang P, Chen YH, Yen CF, Chen HL.
Daly K. Toward objective classification 43. Gray PH, Edwards DM, O’Callaghan Psychiatric diagnoses, emotional-
of childhood autism: Childhood Autism MJ, Gibbons K. Screening for autism behavioral symptoms and functional
Rating Scale (CARS). J Autism Dev spectrum disorder in very preterm outcomes in adolescents born
Disord. 1980;10(1):91–103 infants during early childhood. Early preterm with very low birth
Hum Dev. 2015;91(4):271–276 weights. Child Psychiatry Hum Dev.
34. Schopler E, Reichler R, Renner B. The
2015;46(3):358–366
Childhood Autism Rating Scale (CARS) 44. Pritchard MA, de Dassel T, Beller
Manual. Los Angeles, CA: Western E, et al. Autism in toddlers 54. Mohammed HS, Wahass SH, Mahmoud
Psychological Services; 1988 born very preterm. Pediatrics. AA. Incidence of autism in high risk
2016;137(2):e20151949 neonatal follow up. Neurosciences
35. Schopler E, Van Bourgondien ME,
(Riyadh). 2016;21(1):43–46
Wellman GJ, Love SR. Childhood 45. Yaari M, Yitzhak N, Harel A, et al.
Autism Rating Scale, (CARS2). 2nd ed. Stability of early risk assessment for 55. Ikejiri K, Hosozawa M, Mitomo S,
Torrance, CA: Western Psychological autism spectrum disorder in preterm Tanaka K, Shimizu T. Reduced growth
Services; 2010 infants. Autism. 2016;20(7):856–867 during early infancy in very low birth

Downloaded from www.aappublications.org/news by guest on November 22, 2018

PEDIATRICS Volume 142, number 3, September 2018 13
 weight children with autism spectrum Rockville, MD: Agency for Healthcare in low resource settings: challenges
disorder. Early Hum Dev. 2016;98:23–27 Research and Quality; 2016 and opportunities to enhance research
56. Rogers TD, McKimm E, Dickson PE, 63. Silove N, Blackmore R, Warren A, Gibbs and services worldwide. Autism Res.
Goldowitz D, Blaha CD, Mittleman G. Is V, Roberts J. A consensus approach 2015;8(5):473–476
autism a disease of the cerebellum? for the paediatrician’s role in the 72. Stewart LA, Lee LC. Screening for
An integration of clinical and pre- diagnosis and assessment of autism autism spectrum disorder in low- and
clinical research. Front Syst Neurosci. spectrum disorders in Australia. middle-income countries: a systematic
2013;7:15 Available at: https://​www.​racp.​edu.​au/​ review. Autism. 2017;21(5):527–539
57. Blencowe H, Cousens S, Chou D, et al; docs/​default-​source/​advocacy-​library/​
73. Gladstone M, Oliver C, Van den Broek
Born Too Soon Preterm Birth Action pa-​pol-​a-​consensus-​approach-​for-​the-​
N. Survival, morbidity, growth and
Group. Born too soon: the global paediatricians-​role-​in-​the-​diagnosis-​
developmental delay for babies born
epidemiology of 15 million preterm and-​assessment-​of-​autism.​pdf.
preterm in low and middle income
births. Reprod Health. 2013;10(suppl Accessed November 22, 2017
countries - a systematic review of
1):S2 64. Taylor LJ, Eapen V, Maybery MT, et al. outcomes measured. PLoS One.
58. Liu L, Oza S, Hogan D, et al. Global, Diagnostic evaluation for autism 2015;10(3):e0120566
regional, and national causes of spectrum disorder: a survey of health
74. Lee AC, Katz J, Blencowe H, et al; CHERG
under-5 mortality in 2000-15: an professionals in Australia. BMJ Open.
SGA-Preterm Birth Working Group.
updated systematic analysis with 2016;6(9):e012517
National and regional estimates of
implications for the sustainable 65. Lazaratou H, Economou M, Dikeos D. term and preterm babies born small
development goals. Lancet. The necessity of early intervention in for gestational age in 138 low-income
2016;388(10063):3027–3035 autism. J Pediatr. 2017;184:240–241 and middle-income countries in 2010.
59. Blencowe H, Lee AC, Cousens S, 66. Torjesen I. Early intervention may limit Lancet Glob Health. 2013;1(1):e26–e36
et al. Preterm birth-associated severity of autism symptoms. BMJ. 75. Smid MC, Stringer EM, Stringer JS. A
neurodevelopmental impairment 2016;355:i5776 worldwide epidemic: the problem and
estimates at regional and global levels challenges of preterm birth in low-
67. Ribeiro SH, Paula CS, Bordini D, Mari
for 2010. Pediatr Res. 2013;74(suppl and middle-income countries. Am J
JJ, Caetano SC. Barriers to early
1):17–34 Perinatol. 2016;33(3):276–289
identification of autism in Brazil. Rev
60. Baxter AJ, Brugha TS, Erskine HE, Bras Psiquiatr. 2017;39(4):352–354 76. Higgins JP, Thompson SG, Spiegelhalter
Scheurer RW, Vos T, Scott JG. The DJ. A re-evaluation of random-effects
68. Van Cong T, Weiss B, Toan KN, et al. Early
epidemiology and global burden of meta-analysis. J R Stat Soc Ser A Stat
identification and intervention services
autism spectrum disorders. Psychol Soc. 2009;172(1):137–159
for children with autism in Vietnam.
Med. 2015;45(3):601–613
Health Psychol Rep. 2015;3(3):191–200 77. Lavelle TA, Weinstein MC, Newhouse
61. Christensen DL, Baio J, Van Naarden
69. Samms-Vaughan ME. The status JP, Munir K, Kuhlthau KA, Prosser LA.
Braun K, et al; Centers for Disease
of early identification and early Economic burden of childhood autism
Control and Prevention (CDC).
intervention in autism spectrum spectrum disorders. Pediatrics.
Prevalence and characteristics of
disorders in lower- and middle-income 2014;133(3). Available at: www.​
autism spectrum disorder among
countries. Int J Speech-Language pediatrics.​org/​cgi/​content/​full/​133/​3/​
children aged 8 years–Autism and
Pathol. 2014;16(1):30–35 e520
Developmental Disabilities Monitoring
Network, 11 sites, United States, 2012. 70. Franz L, Chambers N, von Isenburg 78. Leigh JP, Du J. Brief report: forecasting
MMWR Surveill Summ. 2016;65(3):1–23 M, de Vries PJ. Autism spectrum the economic burden of autism in 2015
disorder in Sub-Saharan Africa: a and 2025 in the United States. J Autism
62. McPheeters ML, Weitlauf A, Vehorn A,
comprehensive scoping review. Autism Dev Disord. 2015;45(12):4135–4139
et al. Screening for Autism Spectrum
Disorder in Young Children: A Res. 2017;10(5):723–749 79. Vourda A, Geitona M. Socioeconomic
Systematic Evidence Review for the 71. Durkin MS, Elsabbagh M, Barbaro J, burden of autism in children in Greece.
US Preventive Services Task Force. et al. Autism screening and diagnosis Value Health. 2015;18(7):A748

Downloaded from www.aappublications.org/news by guest on November 22, 2018

14 AGRAWAL et al
Prevalence of Autism Spectrum Disorder in Preterm Infants: A Meta-analysis
Sachin Agrawal, Shripada C. Rao, Max K. Bulsara and Sanjay K. Patole
Pediatrics 2018;142;
DOI: 10.1542/peds.2018-0134 originally published online August 3, 2018;

Updated Information & including high resolution figures, can be found at:
Services http://pediatrics.aappublications.org/content/142/3/e20180134
References This article cites 71 articles, 11 of which you can access for free at:
http://pediatrics.aappublications.org/content/142/3/e20180134#BIBL
Subspecialty Collections This article, along with others on similar topics, appears in the
following collection(s):
Developmental/Behavioral Pediatrics
http://www.aappublications.org/cgi/collection/development:behavior
al_issues_sub
Autism/ASD
http://www.aappublications.org/cgi/collection/autism:asd_sub
Permissions & Licensing Information about reproducing this article in parts (figures, tables) or
in its entirety can be found online at:
http://www.aappublications.org/site/misc/Permissions.xhtml
Reprints Information about ordering reprints can be found online:
http://www.aappublications.org/site/misc/reprints.xhtml

Downloaded from www.aappublications.org/news by guest on November 22, 2018

Prevalence of Autism Spectrum Disorder in Preterm Infants: A Meta-analysis
Sachin Agrawal, Shripada C. Rao, Max K. Bulsara and Sanjay K. Patole
Pediatrics 2018;142;
DOI: 10.1542/peds.2018-0134 originally published online August 3, 2018;

The online version of this article, along with updated information and services, is
located on the World Wide Web at:
http://pediatrics.aappublications.org/content/142/3/e20180134

Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it
has been published continuously since 1948. Pediatrics is owned, published, and trademarked by
the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois,
60007. Copyright © 2018 by the American Academy of Pediatrics. All rights reserved. Print ISSN:
1073-0397.

Downloaded from www.aappublications.org/news by guest on November 22, 2018