Antenatal Ultrasound and Risk of Autism Spectrum Disorders

Article (PDF Available)inJournal of Autism and Developmental Disorders 40(2):238-45 · October 2009with238 Reads
DOI: 10.1007/s10803-009-0859-4 · Source: PubMed
We evaluated antenatal ultrasound (U/S) exposure as a risk factor for autism spectrum disorders (ASD), comparing affected singleton children and control children born 1995-1999 and enrolled in the Kaiser Permanente health care system. Among children with ASD (n = 362) and controls (n = 393), 13% had no antenatal exposure to U/S examinations; case-control differences in number of exposures during the entire gestation or by trimester were small and not statistically significant. In analyses adjusted for covariates, cases were generally similar to controls with regard to the number of U/S scans throughout gestation and during each trimester. This study indicates that antenatal U/S is unlikely to increase the risk of ASD, although studies examining ASD subgroups remain to be conducted.
Antenatal Ultrasound and Risk of Autism Spectrum Disorders
Judith K. Grether Æ Sherian Xu Li Æ
Cathleen K. Yoshida Æ Lisa A. Croen
Ó Springer Science+Business Media, LLC 2009
Abstract We evaluated antenatal ultrasound (U/S)
exposure as a risk factor for autism spectrum disorders
(ASD), comparing affected singleton children and control
children born 1995–1999 and enrolled in the Kaiser Per-
manente health care system. Among children with ASD
(n = 362) and controls (n = 393), 13% had no antenatal
exposure to U/S examinations; case–control differences in
number of exposures during the entire gestation or by tri-
mester were small and not statistically significant. In
analyses adjusted for covariates, cases were generally
similar to controls with regard to the number of U/S scans
throughout gestation and during each trimester. This study
indicates that antenatal U/S is unlikely to increase the risk
of ASD, although studies examining ASD subgroups
remain to be conducted.
Keywords Antenatal ultrasound Obstetrical ultrasound
Autism and other autism spectrum disorders (ASD) are
behaviorally defined conditions, characterized by impair-
ments in communication and social skills accompanied by
repetitive and stereotypic behaviors. Available evidence
indicates that the initiating process leading to ASD typi-
cally begins during fetal development (Bauman and
Kemper 2005; Rodier 2004), but, little is currently known
about causal factors, despite considerable research in recent
years. While twin and family studies support a strong
genetic underpinning for ASD (reviewed in Muhle et al.
2004), there is also evidence to support a role for non-
genetic or environmental factors, probably operating
through interaction with genetic susceptibility (reviewed in
Newschaffer et al. 2002, and in Lawler et al. 2004). During
the past two decades, as the proportion of the childhood
population diagnosed with ASD has increased in the United
States and elsewhere, there has been considerable specu-
lation about environmental factors that may be contributing
to the increase.
Initially introduced into obstetrical (OB) practice in the
1970s (Huang 1994), antenatal ultrasonography (U/S) has
become commonplace as a means to visualize fetal growth
and development; multiple U/S examinations during
pregnancy are now routine in many settings, increasing the
number and duration of fetal exposures. Early human
studies that examined potential risk associated with ante-
natal U/S have focused primarily on neonatal outcomes
(Scheidt et al. 1978; Stark et al. 1984; Lyons et al. 1988;
Newnham et al. 1993; Bellieni et al. 2005). In general,
findings have not indicated adverse neonatal consequences,
but, with more widespread and frequent examinations
and advances in technology, continuing studies have
been called for (Huang 1994; Mole 1986; Barnett 2002;
Marinac-Dabic et al. 2002; Hershkovitz et al. 2002; Rados
2004; Bly and Van den Hof 2005).
More recently, studies beyond the neonatal period have
suggested that dyslexia (Stark et al. 1984), speech delay
(Campbell et al. 1993), and non-right handedness (Salvesen
J. K. Grether (&)
Environmental Health Investigations Branch, California
Department of Public Health, 850 Marina Bay Parkway, P-3,
Richmond, CA 94804, USA
S. X. Li C. K. Yoshida L. A. Croen
Division of Research, Kaiser Permanente of Northern California,
Oakland, CA, USA
J Autism Dev Disord
DOI 10.1007/s10803-009-0859-4
et al. 1993; Kieler et al. 1998b, 2001) may be associated
with fetal U/S exposure; no differences in other behavioral
markers have been reported (Stark et al. 1984; Salvesen
et al. 1993; Kieler et al. 1998a; Newnham et al. 2004;
reviewed in Salvesen 2007). Impairments in communica-
tion are a defining characteristic of ASD and non-right
handedness has been reported to be more common in
children with autism (Gillberg 1983; Lewin et al. 1993;
Dane and Balci 2007).
While antenatal U/S has been suggested as a possible
causal factor for autism (Rodgers 2006), no studies
designed to test the hypothesis of an association have yet
been reported. We here report a case–control study
designed to evaluate this hypothesis.
Subjects for this study were singleton children enrolled in
Kaiser Permanente of Northern California (KPNC), a large
integrated health care delivery system that serves more
than three million members living in a 14-county region in
Northern California. All subjects were born January 1,
1995 through December 31, 1999 in a KPNC hospital to
mothers who were members of the health plan throughout
their entire pregnancy; all subjects remained in the health
plan for at least 2 years after birth.
Cases were defined as singleton children with at least
one diagnosis of an ASD (i.e., autism, Asperger’s Disorder,
or Pervasive Developmental Disorder Not Otherwise
Specified (PDD-NOS)) based on International Classifica-
tion of Diseases criteria ((ICD-9-CM 299.0 or 299.8) and
recorded in KPNC clinical databases. Cases were identified
by electronically scanning KPNC clinical databases con-
taining all diagnoses made at outpatient visits occurring at
plan facilities and outside approved facilities and recorded
between January 1995 and November 2002. One singleton
control per case was randomly selected from the cohort of
KPNC births without an ASD diagnosis, frequency mat-
ched to cases on gender, birth year, and hospital of birth.
Following identification of cases and controls, siblings of
case and control children born at a KPNC facility between
1990 and 1999 were identified from KPNC files. Since the
focus of this study was on maternal factors during preg-
nancy, siblings were defined as births to the same mother,
without regard to paternity.
Clinical data were obtained from multiple sources
within KPNC to identify antenatal U/S examinations, date
and time of scan, gestational age at time of scan, type of
scan (American Medical Association Current Procedural
Terminology, 4th Edition (CPT-4), codes 76805, 76815,
76830, 76856, 76946), and related data. Abstraction of
maternal medical records was conducted by trained KPNC
medical records abstractors for all subjects. Data for U/S
examinations conducted within KPNC were also obtained
from the KPNC electronic imaging databases. For each
child, maternal hard-copy and electronic records were
searched to obtain any indication that an antenatal U/S
examination was conducted. Based on CPT-4 codes for
type of scan, all obstetrical and non-obstetrical pelvic U/S
were included; U/S examinations for which CPT-4 codes
were missing were assumed to be obstetrical U/S exam-
inations and included in the analysis if the scan date was
between the last menstrual period and the date of delivery.
Among children with ASD, 25% of scans during this time
period had missing information for type of scan, and
among controls, 26%. Children for whom no antenatal U/S
examinations were recorded were considered to have no
scans. Two scans conducted on the same date at the same
time were counted as one scan and two scans conducted on
the same date at different times or, if time was unknown,
with different procedure codes, were counted as two scans.
Inconsistencies and discrepancies in reporting of U/S data
from multiple sources were manually reviewed and
resolved through re-review of medical records as neces-
sary. The one mother–child pair identified with a pulsed
Doppler examination was excluded from all analyses, as
different risks may be associated with this type of
Estimated gestational age at birth (EGA
) was
obtained from the Kaiser Infant Cohort File, based on a
previously designed algorithm that included consideration
of aberrant gestational ages and birth weights. Estimated
gestational age at time of antenatal U/S (EGA
) was
obtained using an algorithm that gave priority to the ges-
tational age established at time of U/S based on fetal
measurements and, if unavailable, calculated EGA based
on the date of last menstrual period recorded at the time of
U/S scan. For children with documentation of U/S scans
but for whom EGA
was missing, we calculated EGA
by subtracting the number of completed weeks between
birth date and scan date from the gestational age at
birth (EGA
). Data on EGA
were then classified by
trimester (1st trimester = 0–12 weeks, 2nd trimester =
13–24 weeks, 3rd trimester C 25 weeks).
Demographic data were obtained from birth certificates
to evaluate possible confounding by parity, birth year, birth
weight, and maternal education, age, and race/ethnicity.
Statistical Analysis
For case–control analysis of demographic characteristics,
we used chi-square tests of association for categorical
variables and t-test for continuous variables, with
p \ 0.05 indicating statistical significance. For unadjusted
analysis of the number of U/S scans performed during the
J Autism Dev Disord
pregnancy and within each trimester, we used the chi-
square test for trend, comparing case and control fre-
quencies across number of scans treated as a categorical
variable. The chi-square trend test does not require the
assumption of normality and tests for increasing risk with
increased exposure. Adjusted logistic regression models
were used to compute odds ratios (OR) for ASD and 95%
confidence intervals (95% CI) associated with the number
of antenatal U/S examinations. Adjusted models were run
encompassing the entire gestational period and within
strata defined by trimester and gender. The total number
of any antenatal U/S examinations was treated as a con-
tinuous exposure variable and, in separate models, as a
categorical variable (reference = 0 vs. 1, 2, or C3).
Demographic variables significantly associated with case–
control status were considered as possible confounders
and included in the adjusted logistic regression models, as
were the matching variables for control selection. Since
our study is the first to directly address the concerns about
a possible link between ultrasound and ASD, statistical
significance was evaluated without correction for multiple
comparisons to enhance the likelihood of finding signifi-
cant associations that could be tested in future studies.
Statistical Analysis Software (SAS) was used for all
To minimize possible confounding by unmeasured
maternal or familial factors, analyses were initially
restricted to cases from simplex families (only one child
with ASD) and controls from families without another
child with ASD. Further analyses were then conducted that
included multiplex case families (n = 11). To assure
independence of observations, one child was randomly
selected from families with two siblings in the case group.
Three controls with an affected sibling were excluded from
all analyses.
This study was approved by the California State Com-
mittee for the Protection of Human Subjects and the
Institutional Review Board of KPNC.
Final sample size was N = 362 children with ASD from
simplex families (N = 373 simplex ? multiplex) and
N = 393 control children. Mothers of children with ASD
had significantly greater mean maternal age than mothers
of children in the control group, but maternal age was not
different when evaluated in 5-year age categories
(Table 1). Mothers of children with ASD had more years of
education than mothers of children in the control group and
the mean gestational age at birth of children with ASD was
somewhat shorter than for control children, approaching
statistical significance. Gender, birth weight, parity, year of
birth, and maternal race/ethnicity were not different for
children with ASD compared to controls (Table 1).
Thirteen percent of children with ASD from simplex
families and 12.5% of control children had no exposure to
U/S examinations during gestation. The majority of U/S
examinations were conducted in the second trimester;
77.9% of cases and 79.2% of controls had one or more
scans during this time. In the first trimester and the third
trimester, approximately 28% of both cases and controls
had at least one scan. In unadjusted analyses using a chi-
square trend test, the number of U/S examinations was not
significantly associated with ASD status during the entire
gestational period or for any trimester (Table 2).
Logistic regression models were adjusted for maternal
education, maternal age (continuous), gestational age, birth
hospital, birth year, and gender (for models that included
children of both genders). In models with U/S frequency
treated as a continuous variable, no significant or consistent
increased risk of ASD was observed with increasing
numbers of antenatal U/S examinations for the total ges-
tation or for any trimester (Table 3). Within gender strata,
no increased risk of ASD was observed with increasing
numbers of U/S examinations for either male or female
children, with the exception of females in the 2nd trimester
(AOR = 2.49, 95% CI 1.20, 5.15; Table 3). In models that
treated frequency as a categorical variable, we found no
statistically significant elevation in risk with increasing
numbers of antenatal U/S examinations for the total ges-
tation or for any trimester. When boys or girls were eval-
uated separately in models that treated frequency as a
categorical variable, no significant elevations in risk were
seen for the total pregnancy or for any trimester (data not
All results were similar for the total sample when the 11
children with autism from multiplex families were added to
the case group (data not shown).
Since being introduced into obstetrical care in the 1970s,
antenatal ultrasound examinations have become widely
accepted in obstetrical practice as a routine, non-invasive
tool for determining the size, location, number, and age of
fetuses, and for detecting fetal malformations and intra-
uterine growth retardation. Multiple U/S examinations
during pregnancy are now common throughout the indus-
trialized world, with many women having three or more
U/S scans during a normal pregnancy (Marinac-Dabic et al.
2002; Hershkovitz et al. 2002). Ultrasound uses high-fre-
quency sound waves converted to electric impulses to form
an image. Ultrasound energy absorbed locally has the
potential for localized loss of cells and tissues if there is a
J Autism Dev Disord
sufficient rise in temperature or sufficient cavitation from
U/S-induced pressure changes and gas bubbles (Mole
1986; Barnett 2002; Hershkovitz et al. 2002; Rados
2004). When used according to established safety
guidelines, antenatal U/S is considered to pose minimal
risk to the fetus or mother, but some uncertainties about
safety remain, particularly with changes in technology
and frequency or duration of application (Mole 1986;
Table 1 Characteristics of children with autism spectrum disorders (ASD) and control children, Kaiser Permanente, birth years 1995–1999*
Children with ASD Control children p value**
n = 362 n = 393
n % n %
Sex 0.403
Female 58 (16.0) 72 (18.3)
Male 304 (84.0) 321 (81.7)
Maternal age 0.294
\20 6 (1.7) 12 (3.1)
20–24 44 (12.2) 58 (14.8)
25–29 96 (26.5) 120 (30.5)
30–34 119 (33.0) 116 (29.5)
35–39 80 (22.1) 74 (18.8)
C40 17 (4.7) 13 (3.3)
Maternal ethnicity 0.504
White, non-Hispanic 184 (50.8) 191 (48.6)
White, Hispanic US born 35 (9.7) 46 (11.7)
White, Hispanic not US born 28 (7.7) 44 (11.2)
Black 32 (8.8) 33 (8.4)
Asian 35 (9.7) 30 (7.6)
Other 48 (13.3) 49 (12.5)
Maternal education 0.001
\High school 21 (5.8) 31 (7.9)
HS grad 75 (20.7) 123 (31.3)
College 194 (53.6) 192 (48.9)
PostGrad 70 (19.3) 44 (11.2)
Other 2 (0.6) 3 (0.8)
Parity 0.831
0 164 (45.3) 175 (44.5)
C1 198 (54.7) 218 (55.5)
Birth year 0.999
1995 107 (29.6) 116 (29.5)
1996 93 (25.7) 100 (25.5)
1997 69 (19.1) 75 (19.1)
1998 70 (19.3) 78 (19.9)
1999 23 (6.4) 24 (6.1)
Birth weight
Mean (SD) 3480.9 (611.9) 3522.3 (547.0) 0.329
Gestational age at birth
Mean (SD) 38.997 (1.8) 39.2 (1.7) 0.060
Maternal age
Mean (SD) 30.8 (5.5) 29.7 (5.7) 0.005
* Children with ASD from simplex families
** p-value for chi-square test (categorical data) and t-test (continuous data)
J Autism Dev Disord
Huang 1994; Barnett 2002; Marinac-Dabic et al. 2002;
Hershkovitz et al. 2002; Rados 2004; Bly and Van den Hof
2005; Ang et al. 2006; Rodgers 2006). In addition, only
limited studies have addressed neurodevelopmental out-
comes that may only become clinically apparent in early
childhood or later. A recent review article has raised new
questions about a possible link between antenatal U/S and
‘the alarming increase in autism’ (Rodgers 2006).
Prior evaluations of human newborn and pediatric out-
comes have, in general, demonstrated no measurable
associations between prenatal ultrasound exposure and
outcomes measured proximal to the time of birth: Apgar
scores, gestational age, head circumference, birth weight,
length, congenital abnormalities, neonatal infection, and
congenital infection (Scheidt et al. 1978; Stark et al. 1984;
Lyons et al. 1988; Newnham et al. 1993). In contrast to
these null studies, Bellieni et al. (2005) reported lower
mean birth weight with exposure to nine or more U/S scans
compared to three or fewer and Newnham et al. (2004)
reported increased low birth weight (below the 10th or 3rd
percentiles) in a randomized trial comparing intensive
monitoring (five diagnostic U/S and Doppler flow studies
during gestation) to a single U/S examination at 18 weeks
gestation; these birth weight differences resolved by 1 year
of age.
Longer term outcomes have also been assessed. Camp-
bell et al. (1993) reported increased antenatal U/S expo-
sures in a case group with delayed speech compared to
matched controls. Stark et al. (1984) evaluated hearing,
visual acuity and color vision, cognitive function, behavior,
Table 2 Frequency of antenatal ultrasound (U/S) examinations for children with ASD and control children, Kaiser Permanente, birth years
Children with ASD Control children p value**
(n = 362) (n = 393)
n % n %
Total number of scans 667 676
Number of scans with missing CPT-4 code 164 25% 179 26%
Total per subject 0.18
0 48 (13.3) 49 (12.5)
1 130 (35.9) 147 (37.4)
2 101 (28.0) 120 (30.5)
3 36 (9.9) 46 (11.7)
4 24 (6.6) 18 (4.6)
5 13 (3.6) 8 (2.0)
C6 10 (2.8) 5 (1.3)
1st trimester (0–12 weeks) 0.56
0 258 (71.3) 286 (72.8)
1 88 (24.3) 94 (23.9)
2 14 (3.9) 10 (2.5)
C3 2 (0.6) 3 (0.8)
2nd trimester (13–24 weeks) 0.44
0 80 (22.1) 82 (20.9)
1 207 (57.2) 242 (61.6)
2 60 (16.6) 58 (14.8)
3 12 (3.3) 11 (2.8)
4 3 (0.8) 0 (0.0)
3rd trimester (25? weeks) 0.21
0 260 (71.9) 283 (72.0)
1 61 (16.9) 80 (20.4)
2 22 (6.1) 20 (5.1)
3 12 (3.3) 6 (1.5)
C4 7 (1.9) 4 (1.0)
* Children with ASD from simplex families
** p value based on chi-square test for trend
J Autism Dev Disord
and neurologic function in children followed to 7–12 years
of age, with exposed and unexposed children born of
pregnancies matched for pregnancy complications. More
exposed children tested positive for dyslexia, a difference
that was consistent across all three hospitals included in the
study but that did not reach statistical significance in any;
no other differences were found. Newnham et al. (2004)
found no differences at 8 years of age using standard tests
of childhood speech, language, behavior, and neurological
development when randomly selected children with inten-
sive exposure were compared to children with low
Following an initial report by Salvesen et al. (1993)ofa
modest but statistically significant increase in non-right
handedness in children who had been randomly exposed to
U/S during gestation, Kieler et al. (1998a, b) evaluated
handedness and other characteristics study of children
born to women who had been randomly assigned to U/S
screening at 15 weeks gestation. A modest, but significant,
increase in non-right handedness was seen among boys but
not when children of both genders were analyzed as one
group. No other behavioral or neurological differences
were found in these studies. A further study by Kieler et al.
(2001) among men enrolled for military service in Sweden
evaluated handedness among men born in hospitals without
antenatal U/S compared to men born in hospitals in which
U/S was the standard of care; left-handedness was found to
be higher among men who were presumably exposed. No
clear association was seen between antenatal U/S exposure
and intellectual performance in these men (Kieler et al.
2005). See Salvesen (2007) for a recent review of these
epidemiologic studies.
As non-right handedness has been reported to be more
common in children with autism (Gillberg 1983; Lewin
et al. 1993; Dane and Balci 2007) and delayed speech is
one of the early signs of autism, the studies cited above
raise the question of a possible link between autism and
antenatal U/S exposure. In addition, a recently reported
study of prenatal exposure to U/S in mouse models found
an effect on neuronal migration in the cerebral cortex (Ang
et al. 2006); neuronal migration may be a cellular mecha-
nism contributing to autism (reviewed in Keller and Pers-
ico 2003). These observations have contributed to concerns
that increased exposure to U/S in routine obstetrical care
may be linked to the rising prevalence of autism that has
been observed in children over the past two decades
(Rodgers 2006).
Focusing on children enrolled in a large health care
delivery system that is largely representative of the general
population residing in the same geographic region (Krieger
1992), we compared antenatal U/S exposure in children
with ASD to exposure in control children from the same
population. We did not find an association between the risk
of ASD and the number of antenatal U/S examinations.
These null findings were consistent for the gestation as a
whole, when evaluated by trimester, and among both boys
and girls. There is one exception to this pattern of null
results: we find an elevated risk for females in the 2nd
trimester when U/S was treated as a continuous variable in
adjusted models but not when U/S was treated as a cate-
gorical variable. This finding cannot be attributed to one or
two outliers but could represent a chance finding among the
large number of comparisons we conducted; alternatively,
there may be risk associated with multiple U/S for girls in
the 2nd trimester or with medical indications for repeated
U/S during this window of gestation. As the number of U/S
examinations in a pregnancy is, to some extent, guided by
medical indications, our largely null results suggest that, in
general in this study population, relevant medical indica-
tions for U/S as a group may not have been substantially
more common in pregnancies that produced children who
were later diagnosed with autism. We did not measure nor
control for medical indications, per se, but by excluding
multiple births from the analysis, we effectively excluded
pregnancies with one of the most common indications for
repeated U/S examinations. Whether specific medical
indications are more common in pregnancies leading to
autism was outside the scope of this study.
Table 3 Adjusted odds ratio per each ultrasound exposure for chil-
dren with autism spectrum disorders (ASD) and control children,
Kaiser Permanente, birth years 1995–1999*
Frequency of U/S (continuous) Odds ratio** 95% CI
All subjects
Total pregnancy 1.06 (0.95, 1.20)
1st trimester 1.05 (0.80, 1.37)
2nd trimester 1.18 (0.92, 1.53)
3rd trimester 1.08 (0.91, 1.28)
Male children
Total pregnancy 1.05 (0.92, 1.19)
1st trimester 1.07 (0.80, 1.43)
2nd trimester 1.04 (0.79, 1.39)
3rd trimester 1.07 (0.88, 1.30)
Female children
Total pregnancy 1.20 (0.88, 1.64)
1st trimester 0.87 (0.37, 2.02)
2nd trimester 2.49 (1.20, 5.15)
3rd trimester 1.20 (0.75, 1.92)
* Models based on children with ASD from simplex families
(n = 362) and controls (n = 393), adjusted for maternal education,
maternal age, gestational age, birth hospital, birth year, and child
** Odds ratio represents risk associated with each additional U/S
J Autism Dev Disord
Data on U/S examinations were obtained from medical
records, which prospectively captured obstetric events and
thus were not subject to maternal recall. Given that all case
and control mothers in the study population received their
prenatal care within the KPNC system, it is very unlikely
that we under-ascertained antenatal U/S examinations
conducted in a clinical setting. This study was not designed
to evaluate U/S examinations conducted in commercial
settings, which may pose risks beyond those conducted
under established guidelines in clinical settings.
A possible limitation of our study is that we were unable
to construct a more precise estimate of exposure ‘dose’
based on actual duration of exposure or other parameters of
the U/S examination, as necessary data were not docu-
mented in available records. Due to missing data for type
of scan in approximately 25% of scans, we may have failed
to exclude some mother–baby pairs for which Doppler
evaluations were conducted. Doppler evaluations are only
rarely conducted in this setting; evaluation of available data
for scans lacking codes for type of scan identified \1% of
scans with missing data that might be Doppler. A further
limitation is that we were unable to evaluate phenotypic
subgroups of cases, based on severity of ASD or other
characteristics, as the data in medical records were not
sufficient to accurately and completely make these dis-
tinctions. We were also unable to validate ASD diagnoses
through clinical examination of all children, although a
subset of 50 (13.8%) of the children with autism in our
study also participated in another study for which they
underwent clinical evaluation with the Autism Diagnostic
Interview-Revised (ADI-R) (Lord et al. 1994) and the
Autism Diagnostic Observation Schedule-Generic (ADOS-
G). (Lord et al. 2000). Among the 50 children evaluated
with the ADI-R and ADOS-G, 94% met criteria for ASD
on both instruments, and 100% met criteria on at least one
(Croen et al. 2008). In addition, record-review validation
studies conducted by the investigators which included full
review of diagnostic information recorded in KPNC med-
ical records have demonstrated that at least 90% of children
with an ASD diagnosis recorded in the KPNC electronic
databases had documentation in their records consistent
with a diagnosis of autism based on DSM-IV criteria
(Croen et al. 2008). The prevalence of ASD within this
study population (4.76/1,000) is within the range reported
among U.S. children for this time period (Autism and
Developmental Disabilities Monitoring Network 2007).
The predominately null findings of this investigation of
fetal exposure to antenatal U/S and risk of autism spectrum
disorders may provide assurance that antenatal ultrasound
examinations, when performed in a clinical setting according
to established guidelines, do not appear to put fetuses at
increased risk for developing ASD. Given the substantial size
of our study, it is unlikely that an association between U/S
exposure and ASD would be seen if larger numbers of cases
and controls had been included, although further studies
among girls are indicated. Studies of phenotypic subgroups
remain to be conducted and as antenatal U/S technologies
and procedures change, further evaluation of potential risks
may be warranted. This study did not address other possible
risks to fetuses associated with antenatal ultrasound.
Acknowledgments We wish to express appreciation to Kaht Dor-
ward, MD, and Linda Copeland, MD, practicing physicians within
Kaiser Permanente of Northern California, for providing clinical
consultation and to Meredith Anderson, MA, and Daniel Smith,
DrPH, colleagues within California Department of Public Health, for
statistical consultation and for careful review of the manuscript. Any
errors in calculation or interpretation of data are those of the authors.
This study was funded by grants from the Centers for Disease Control
and Prevention, (U10/CCU920392), the Kaiser Foundation Research
Institute, and Autism Speaks, and by the California Department of
Public Health.
Ang, E. S., Jr., Gluncic, V., Duque, A., Schafer, M., & Rakic, P.
(2006). Prenatal exposure to ultrasound waves impacts neuronal
migration in mice. Proceedings of the National Academy of
Sciences of the United States of America, 103(34), 12903–12910.
Autism and Developmental Disabilities Monitoring Network. (2007).
Surveillance Year 2002 Principal Investigators; Centers for
Disease Control and Prevention. Prevalence of autism spectrum
disorders-autism and developmental disabilities monitoring
network, 14 sites, United States, 2002. MMWR Surveillance
Summary, 56(1), 12–28.
Barnett, S. B. (2002). Routine ultrasound scanning in first trimester:
What are the risks? Seminars in Ultrasound, CT and MR, 23(5),
Bauman, M., & Kemper, T. L. (2005). Neuroanatomic observations of
the brain in autism: A review and future directions. International
Journal of Developmental Neuroscience, 23(2–3), 183–187.
Bellieni, C. V., Buonocore, G., Bagnoli, F., Cordelli, D. M., Gasparre,
O., Calonaci, F., et al. (2005). Is an excessive number of prenatal
echographies a risk for fetal growth? Early Human Development,
81(8), 689–693.
Bly, S., & Van den Hof, M. C. (2005). Obstetric ultrasound biological
effects and safety. Journal of Obstetrics and Gynaecology
Canada, 27(6), 572–580.
Campbell, J., Elford, R., & Brant, R. (1993). Case–control study of
prenatal ultrasonography exposure in children with delayed
speech. CMAJ, 149(10), 1435–1440.
Croen, L., Matevia, M., Yoshida, C., & Grether, J. (2008). Maternal
Rh D status, anti-D immune globulin exposure during preg-
nancy, and risk of autism spectrum disorders. American Journal
of Obstetrics and Gynecology, 199(3), 234.e1–234.e6. Epub
2008 Jun 13.
Dane, S., & Balci, N. (2007). Handedness, eyedness and nasal cycle
in children with autism. International Journal of Developmental
Neuroscience, 25(4), 223–226. Epub 2007 Mar 24.
Gillberg, C. (1983). Autistic children’s hand preferences: Results
from an epidemiological study of infantile autism. Psychiatry
Research, 10(1), 21–30.
Hershkovitz, R., Sheiner, E., & Mazor, M. (2002). Ultrasound in
obstetrics: A review of safety. European Journal of Obstetrics,
Gynecology, and Reproductive Biology, 101(1), 15–18.
J Autism Dev Disord
Huang, L. (1994). Ultrasound technology: The RADIUS (Routine
Antenatal Diagnostic Imaging with Ultrasound) study & national
policy. Journal of Clinical Engineering, 19(4), 297–309.
Keller, F., & Persico, A. (2003). The neurobiological context of
autism. Molecular Neurobiology, 28(1), 1–22.
Kieler, H., Ahlsten, G., Haglund, B., Salvesen, K., & Axelsson, O.
(1998a). Routine ultrasound screening in pregnancy and the
children’s subsequent neurologic development. Obstetrics and
Gynecology, 91(5 Pt 1), 750–756.
Kieler, H., Axelsson, O., Haglund, B., Nilsson, S., & Salvesen, K. A.
(1998b). Routine ultrasound screening in pregnancy and the
children’s subsequent handedness. Early Human Development,
50(2), 233–245.
Kieler, H., Cnattingius, S., Haglund, B., Palmgren, J., & Axelsson, O.
(2001). Sinistrality—A side-effect of prenatal sonography: A
comparative study of young men. Epidemiology, 12(6), 618–623.
Kieler, H., Haglund, B., Cnattingius, S., Palmgren, J., & Axelsson, O.
(2005). Does prenatal sonography affect intellectual perfor-
mance? Epidemiology, 16(3), 304–310.
Krieger, N. (1992). Overcoming the absence of socioeconomic data in
medical records: Validation and application of a census-based
methodology. American Journal of Public Health, 82(5), 703–710.
Lawler, C., Croen, L., Grether, J., & Van de Water, J. (2004).
Identifying environmental contributions to autism: Provocative
clues and false leads. Mental Retardation and Developmental
Disabilities Research Reviews, 10(4), 292–302.
Lewin, J., Kohen, D., & Mathew, G. (1993). Handedness in mental
handicap: Investigation into populations of Down’s Syndrome,
epilepsy and autism. British Journal of Psychiatry, 163, 674–
Lord, C., Rutter, M., DeLovere, P. C., & Risi, S. (2000). The Autism
Diagnostic Observation Schedule (ADOS). Los Angeles: Wes-
tern Psychological Services.
Lord, C., Rutter, M., & Le Couteur, A. (1994). Autism diagnostic
interview-revised. Journal of Autism and Developmental Disor-
ders, 24(5), 659–685.
Lyons, E. A., Dyke, C., Toms, M., & Cheang, M. (1988). In utero
exposure to diagnostic ultrasound: A 6-year follow-up. Radiol-
ogy, 166(3), 687–690.
Marinac-Dabic, D., Krulewitch, C. J., & Moore, R. M., Jr. (2002).
The safety of prenatal ultrasound exposure in human studies.
Epidemiology, 13(3 Suppl), S19–S22.
Mole, R. (1986). Possible hazards of imaging and Doppler ultrasound
in obstetrics. Birth, 13(1), 29–38.
Muhle, R., Trentacoste, S. V., & Rapin, I. (2004). The genetics of
autism. Pediatrics, 113(5), e472–e486.
Newnham, J. P., Doherty, D. A., Kendall, G. E., Zubrick, S. R.,
Landau, L. L., & Stanley, F. J. (2004). Effects of repeated
prenatal ultrasound examinations on childhood outcome up to
8 years of age: Follow-up of a randomised controlled trial.
Lancet, 364(9450), 2038–2044.
Newnham, J. P., Evans, S. F., Michael, C. A., Stanley, F. J., &
Landau, L. I. (1993). Effects of frequent ultrasound during
pregnancy: A randomised controlled trial. Lancet, 342(8876),
Newschaffer, C. J., Fallin, D., & Lee, N. L. (2002). Heritable and
nonheritable risk factors for autism spectrum disorders. Epide-
miologic Reviews, 24(2), 137–153.
Rados, C. (2004). FDA cautions against ultrasounds ‘keepsake’
images. FDA Consumer, 38(1), 12–16.
Rodgers C. (2006). Questions about prenatal ultrasound and the
alarming increase in autism. Midwifery Today with International
Midwife, (80), 16–19, 66–17.
Rodier, P. M. (2004). Environmental causes of central nervous system
maldevelopment. Pediatrics, 113(4 suppl), 1076–1083.
Salvesen, K. A. (2007). Epidemiological prenatal ultrasound studies.
Progress in Biophysics and Molecular Biology, 93(1–3), 295–
300. Epub 2006 Aug 22.
Salvesen, K. A., Vatten, L. J., Eik-Nes, S. H., Hugdahl, K., &
Bakketeig, L. S. (1993). Routine ultrasonography in utero and
subsequent handedness and neurological development. BMJ,
307(6897), 159–164.
Scheidt, P. C., Stanley, F., & Bryla, D. A. (1978). One-year follow-up
of infants exposed to ultrasound in utero. American Journal of
Obstetrics and Gynecology, 131(7), 743–748.
Stark, C. R., Orleans, M., Haverkamp, A. D., & Murphy, J. (1984).
Short- and long-term risks after exposure to diagnostic ultra-
sound in utero. Obstetrics and Gynecology, 63(2), 194–200.
J Autism Dev Disord
    • "Fetal ultrasound under standard conditions is considered a safe procedure that can be performed multiple times throughout pregnancy [Newnham, Evans, Michael, Stanley, & Landau, 1993]. Several large studies have found no association between the frequency of prenatal ultrasound during pregnancy and offspring ASD [Grether, Li, Yoshida, & Croen, 2010; Stoch et al., 2012]. "
    [Show abstract] [Hide abstract] ABSTRACT: Numerous studies have observed that a proportion of infants later diagnosed with autism spectrum disorder (ASD) experience accelerated head growth during the first years of life. An emerging methodology for examining the developmental trajectory prior to a diagnosis of ASD is to investigate siblings of affected individuals. The current study is the first prospective investigation of fetal growth in siblings of children with ASD. Two groups of pregnant women were recruited as part of the PRegnancy Investigation of Siblings and Mothers of children with autism cohort in Perth, Western Australia. The "high risk" group (n = 23) comprised pregnant women who have an existing child with a diagnosis of ASD and the "low risk" group (n = 36) comprised pregnant mothers who have an existing child who has developed typically. Prenatal ultrasounds were procured at multiple time-points throughout the second- and third-trimesters, enabling an examination of growth trajectories. Growth measurements were then compared for the high- and low-risk fetuses. Mixed linear regression models identified no significant differences between the high- and low-risk fetuses in the rate of prenatal head and body growth throughout the second- and third-trimester (all P-values >0.05). Similarly, there were no significant differences observed when comparing high and low risk groups on a ratio of head circumference relative to body size (β = -0.019, P = 0.75). Future studies may consider looking beyond the macro architecture of the prenatal brain and examine the growth of brain subregions that have been implicated in the presentation of ASD symptoms. Autism Res 2015. © 2015 International Society for Autism Research, Wiley Periodicals, Inc. © 2015 International Society for Autism Research, Wiley Periodicals, Inc.
    Full-text · Article · Jul 2015
    • "Another study [Stalberg et al., 2009] correlated the number of dUS exams during the second trimester to school performance, finding a trend (without statistical significance) toward lower mean school grades for boys but not for girls. Two other studies that sought to directly assay at the population level the likelihood of ASD and proxies for dose of dUS dose failed to find a meaningful correlation [Grether, Li, Yoshida, & Croen, 2010; Stoch et al., 2012]. Because of the in vivo results and the increase in power of dUS since FDA standards were set [Gibbs et al., 2009], there remain concerns about the overall safety of dUS and its potential link to increased risk of ASD [Abramowicz, 2012; ter Haar et al., 2013; Williams & Casanova, 2010, 2013. "
    [Show abstract] [Hide abstract] ABSTRACT: Clinical use of diagnostic ultrasound imaging during pregnancy has a long history of safety and diagnostic utility, as supported by numerous human case reports and epidemiological studies. However, there exist in vivo studies linking large but clinically relevant doses of ultrasound applied to mouse fetuses in utero to altered learning, memory, and neuroanatomy of those mice. Also, there exists a well-documented significant increase in the likelihood of non-right-handedness in boys exposed to diagnostic ultrasound in utero, potentially relevant given the increased prevalence of autism in males, and reports of excess non-right-handedness in this population. Motivated by these observations, we applied 30 minutes of diagnostic ultrasound to pregnant mice at embryonic day 14.5 and assayed the social behavior of their male pups 3 weeks after their birth. The ultrasound-exposed pups were significantly (P < 0.01) less interested in social interaction than sham-exposed pups in a three-chamber sociability test. In addition, they demonstrated significantly (P < 0.05) more activity relative to the sham-exposed pups, but only in the presence of an unfamiliar mouse. These results suggest that fetal exposure to diagnostic ultrasound applied in utero can alter typical social behaviors in young mice that may be relevant for autism. There exist meaningful differences between the exposure of diagnostic ultrasound to mice versus humans that require further exploration before this work can usefully inform clinical practice. Future work should address these differences as well as clarify the extent, mechanisms, and functional effects of diagnostic ultrasound's interaction with the developing brain. Autism Res 2013, ●●: ●●-●●. © 2013 International Society for Autism Research, Wiley Periodicals, Inc.
    Full-text · Article · Jun 2014
    • "91). Just as Ziskin and Petitti have suggested, such assumptions are still present today which continue to falsely impress upon the science that safety studies have been adequate and thorough (for example, see [16]) and oscillate with the sonic waveforms; this is referred to as stable cavitation. However, during transient cavitation such as occurs with higher intensity ultrasounds or lower frequencies, the cavities rapidly increase in size until at which point pressure becomes too great in the surrounding medium and the bubble collapses [17]. "
    [Show abstract] [Hide abstract] ABSTRACT: Science has shown that risk of cavitation and hyperthermia following prenatal ultrasound exposure is relatively negligible provided intensity, frequency, duration of exposure, and total numbers of exposures are safely limited. However, noncavitational mechanisms have been poorly studied and occur within what are currently considered “safe” levels of exposure. To date, the teratogenic capacity of noncavitational effectors are largely unknown, although studies have shown that different forms of ultrasound-induced hydraulic forces and pressures can alter membrane fluidity, proliferation, and expression of inflammatory and repair markers. Loose regulations, poor end user training, and unreliable ultrasound equipment may also increase the likelihood of cavitation and hyperthermia during prenatal exposure with prolonged durations and increased intensities. The literature suggests a need for tighter regulations on the use of ultrasound and further studies into its teratogenicity.
    Full-text · Article · Mar 2013
Show more