Obesity and Autism

Alison Presmanes Hill, PhDa, Katharine E. Zuckerman, MD, MPHb, Eric Fombonne, MDc

Overweight and obesity are increasingly prevalent in the general pediatric

OBJECTIVE: abstract
population. Evidence suggests that children with autism spectrum disorders (ASDs) may
be at elevated risk for unhealthy weight. We identify the prevalence of overweight and
obesity in a multisite clinical sample of children with ASDs and explore concurrent
associations with variables identified as risk factors for unhealthy weight in the general
population.
METHODS: Participants were 5053 children with confirmed diagnosis of ASD in the Autism
Speaks Autism Treatment Network. Measured values for weight and height were used to
calculate BMI percentiles; Centers for Disease Control and Prevention criteria for BMI for
gender and age were used to define overweight and obesity ($85th and $95th percentiles,
respectively).
RESULTS: In children age 2 to 17 years, 33.6% were overweight and 18% were obese.
Compared with a general US population sample, rates of unhealthy weight were significantly
higher among children with ASDs ages 2 to 5 years and among those of non-Hispanic white
origin. Multivariate analyses revealed that older age, Hispanic or Latino ethnicity, lower
parent education levels, and sleep and affective problems were all significant predictors of
obesity.
CONCLUSIONS: Our results indicate that the prevalence of unhealthy weight is significantly
greater among children with ASD compared with the general population, with differences
present as early as ages 2 to 5 years. Because obesity is more prevalent among older
children in the general population, these findings raise the question of whether there are
different trajectories of weight gain among children with ASDs, possibly beginning in early
childhood.

a
Center for Spoken Language Understanding, Institute for Development & Disability, Department of Pediatrics; WHAT’S KNOWN ON THIS SUBJECT: Children and
b
Division of General Pediatrics, Doernbecher Children’s Hospital; and cDepartment of Psychiatry, Institute for
Development & Disability, Oregon Health & Science University, Portland Oregon adolescents with autism spectrum disorders
(ASDs) may be at elevated risk for unhealthy
Dr Hill conducted all analyses, drafted the initial manuscript, and reviewed and revised the
manuscript; Dr Zuckerman consulted on all analyses, drafted the initial manuscript, and reviewed weight. Samples of children with verified clinical
and revised the manuscript; Dr Fombonne consulted on all analyses and critically reviewed the diagnoses of ASD have been lacking, and
manuscript; and all authors approved the final manuscript as submitted. associations with child behavior and functioning
www.pediatrics.org/cgi/doi/10.1542/peds.2015-1437 are not well understood.
DOI: 10.1542/peds.2015-1437
WHAT THIS STUDY ADDS: Young children (2–5
Accepted for publication Sep 3, 2015 years old) and adolescents (12–17 years old)
Address correspondence to Alison Presmanes Hill, PhD, 3181 SW Sam Jackson Park Road GH 40, with ASDs were at an elevated risk for unhealthy
Portland, OR 97239. E-mail: hillali@ohsu.edu weight status compared with a general
PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275). population sample. The presence of sleep or
Copyright © 2015 by the American Academy of Pediatrics affective problems may confer increased risk
among those with ASD.

PEDIATRICS Volume 136, number 6, December 2015 ARTICLE

Pediatric overweight and obesity are such as sleep difficulties,3,38 symptoms, adaptive skills). Based on
significant public health concerns. In gastrointestinal (GI) disturbances,39,40 results from a smaller sample of
2011 and 2012, 31.8% of US children attention-deficit/hyperactivity children with ASDs in Oregon,48 we
aged 2 to 19 years were overweight disorder (ADHD),41 and disorders also expected obese children with
(BMI $85th percentile)1; 16.9% were such as anxiety42 and depression.43 ASDs to experience more comorbid
obese (BMI $95th percentile).2 The presence of these unique risk problems (sleep difficulties, ADHD,
Unhealthy weight poses health risks factors suggests that children with internalizing symptoms such as
including sleep-disordered depression and anxiety) and be
ASDs are at an elevated risk for being
breathing,3 orthopedic problems,4 prescribed psychotropic medications
overweight or obese. However,
type 2 diabetes,5 hypertension and more often than nonobese children
prevalence estimates of unhealthy
dyslipidemia,6,7 and reduced life with ASDs.
weight in ASD populations vary
spans regardless of adult weight
widely (Table 1). In 4 previous
status.8 Unhealthy weight is also
studies with non-ASD comparison METHODS
associated with family economic
groups, prevalence of obesity was
burden9,10 and harms psychosocial Participants
higher among those with
functioning11,12: Children who are
ASDs,24,44–46 although the difference Participants included 5053 children
overweight or obese are more likely
reached statistical significance in only enrolled in the Autism Speaks Autism
to be bullied13 and socially isolated.14
Thus, unhealthy weight in childhood 2 studies.44,45 A recent study45 found Treatment Network (ATN) from 2008
significantly higher prevalence of to 2013 at 19 sites in the United
has significant implications for
both overweight and obesity among States and Canada. The ATN registry
current quality of life and future
children with ASDs, with group risks includes children ages 2 to 17 years
independent functioning.15
associated with older age, public with confirmed ASDs per Diagnostic
Little is known about overweight and insurance, and co-occurring sleep and Statistical Manual of Mental
obesity in children with autism Disorders, Fourth Edition, Text
disorders.45 However, previous
spectrum disorders (ASDs).12 Revision49 criteria, supported by
studies have been limited by small
However, this issue is of increased administration of the Autism
samples,24 use of parent-reported
public health importance because Diagnostic Observation Schedule
anthropometrics,44,46 parent-
ASDs now affect 1 in 68 US children.16 (ADOS).50 Registry protocols are
reported ASD diagnosis,44,46 or
Although many risk factors for approved by each site’s institutional
unconfirmed diagnoses present in
unhealthy weight are probably the review board.
medical records.45,47 Additionally,
same in children with ASDs as in the
associations between unhealthy
general pediatric population,17,18 NHANES Comparison Sample
weight and child behavior and
children with ASDs may be vulnerable
functioning are not well understood The NHANES is a representative
to additional risks. For example,
among children with ASDs. cross-sectional sample of the US
problem eating behaviors such as food
noninstitutionalized population and
selectivity are common among The first aim of this study was to
is described elsewhere.2 Weight and
children with ASDs,19–21 which tends examine prevalence of unhealthy
height values in NHANES are
to coincide with preferences for weight in a large multisite sample of
collected via standardized physical
a narrow range of low-nutrition, children with confirmed ASDs, based
examination. We used data from 3
energy-dense foods and rejection of on measured weight and height. We
consecutive NHANES surveys (6
fruits, vegetables, and whole compared these prevalence estimates
years) to account for secular changes
grains.19,22–24 Children with ASDs also with those derived from a US general
in prevalence of overweight or
spend more time in sedentary population sample from the NHANES.
obesity. We restricted the sample to
activities25,26 and have less regular The second aim was to examine
children aged 2 to 17 years to match
physical activity.27,28 In addition, family- and child-level factors
the age range in the ATN (unweighted
children with ASDs often take associated with unhealthy weight
n = 8844; weighted estimate of the
psychotropic medications,29 many of among children with ASDs. Our final
total population size  63 157 608).
which can cause weight gain.30–32 aim was to examine hypotheses
The Supplemental Appendix details
Some children with ASDs may also regarding associations between
prevalence estimations in NHANES.
have genetic vulnerabilities to obesity, unhealthy weight and factors unique
such as 11p14.1 or 16p11.2 to children with ASDs. We
ATN Measures
microdeletions.33–35 Finally, having an hypothesized that unhealthy weight
ASD also increases the risk of among children with ASDs would be Sociodemographics
comorbid problems36,37 associated associated with greater impairments Parents reported child gender, age,
with unhealthy weight in childhood, in behavioral functioning (ASD race or ethnicity, and parents’

1052 PRESMANES et al
TABLE 1 Summary of Prevalence Estimates for Overweight and Obesity ($85th and $95th BMI Percentile for Age and Gender, Respectively) From
Previous Studies Including Children With ASDs
Source Location Age Range, y ASD, n ASD Diagnostic Criteria Wt/Height Overweight, % Obese, %
Ho et al (1997)74 Canada School age 54 — 42.6
Whiteley et al (2004)82 UK 2–12 50 Previous clinical diagnosis; confirmed with ADI-R Parent-reported 42.0 10.0
Curtin et al (2005)47 USA (MA) 3–18 140 Retrospective chart review Measured 35.7 19.0
Xiong et al (2009)83 China 2–11 429 Parent-reported; confirmed with CARS Measured 33.6 18.4
Chen et al (2010)84 USA 10–17 46 707 Parent-reported (telephone interview) Parent-reported — 23.4
Curtin et al (2010)46 USA 3–17 102 353 Parent-reported (telephone interview) Parent-reported — 30.4
Rimmer et al (2010)85 USA 12–18 461 Parent-reported (Web-based survey) Parent-reported 42.5 24.6
Evans et al (2012)24 USA 3–11 53 Confirmed with ADI-R Measured — 17.0
Hyman et al (2012)86 USA 2–11 362 DSM-IV; confirmed with ADOS Measured — 8.3
Memari et al (2012)87 Iran 7–14 113 DSM-IV-TR; confirmed with ADI-R Measured 40.7 27.4
Egan et al (2013)70 USA (MO) 2–5 273 Retrospective chart review Measured 33.0 17.6
Zuckerman et al (2014)48 USA (OR) 2–18 376 DSM-IV-TR, ADOS Measured 35.1 17.0
Phillips et al (2014)44 USA 12–17 93 Parent-reported (in-person interview) Parent-reported 52.7 31.8
Broder-Fingert et al USA (MA) 2–20 2976 International Classification of Disease, Ninth Revision Measured 37.5 23.8
(2014)45 diagnosis of autism or Asperger syndrome
ADI-R, Autism Diagnostic Interview–Revised; CARS, Childhood Autism Rating Scale; DSM-IV-TR, Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision.

education. Race was reported in 6 We categorized medications as of overall ASD symptom severity.51,52
categories; these were collapsed to stimulants, selective serotonin Parents completed the Vineland
white, black, and other races for reuptake inhibitors, nonstimulant Adaptive Behavior Scales (VABS-II),53
analyses because of sample size ADHD medications, anticonvulsants, which assesses functional skills and
constraints. Ethnicity was categorized asthma and allergy medications, and provides an Adaptive Behavior
as Hispanic or Latino origin or not atypical neuroleptics. For bivariate Composite as an estimate of overall
Hispanic or Latino. Parent education and multivariate logistic regression adaptive functioning (mean = 100,
was grouped as high school graduate analyses, variables were collapsed SD = 15). We assessed intellectual
or less, some college, or college into any or no prescribed functioning by using 1 of the
graduate or higher. psychotropic medications. Additional following assessments (N = 3787):
bivariate analyses examined the full Stanford–Binet Scales of
BMI associations of BMI category with Intelligence (n = 753),54 the
Trained clinical staff measured total number of psychotropic abbreviated Stanford–Binet Scales of
children’s weight and height using medications as well as individual Intelligence (n = 1632), the Wechsler
a metric scale and wall-mounted medication categories. Parents also Intelligence Scale for Children, Fourth
stadiometer. These values were reported use of complementary and Edition (n = 141),55 the Differential
converted to gender-specific BMI-for- alternative medications or treatments Ability Scales (n = 72),56 the Wechsler
age percentiles based on Centers for (CAM): chiropractic care, dietary Preschool and Primary Scale of
Disease Control and Prevention (CDC) supplements (amino acids, high-dose Intelligence (n = 84),57 the Wechsler
growth charts,1 and CDC criteria were vitamin B6 and magnesium, essential Abbreviated Scale of Intelligence
used to define overweight (BMI fatty acids, probiotics, digestive (n = 59),58 the Leiter International
$85th percentile for age and gender) enzymes, glutathione), or dietary Performance Scale–Revised
and obesity (BMI $95th percentile interventions (gluten-free, casein- (n = 108),59 and the Mullen Scales of
for age and gender). Underweight free, no processed sugars). Because of Early Learning (n = 938; Early
children (BMI ,5th percentile; n = the infrequent rate of CAM Learning Composite Standard
237) were included in the endorsement, variables were Scores).60 Because Mullen Scales of
denominator for prevalence estimates collapsed into any or no CAM use. Early Learning Early Learning
in both the ATN and NHANES Current use of melatonin was Composite scores were skewed and
samples but excluded from within- measured as a separate variable. could not be transformed, all children
ATN subgroup comparisons. were grouped as IQ ,70 (intellectual
Behavioral Functioning disability range) or not.
Treatments Trained clinicians scored ASD
At registry entry, ATN clinicians symptoms during the ADOS,50 Comorbid Problems
record each child’s prescribed a standardized observational Parents reported on children’s sleep
psychotropic medications; dosage assessment; ADOS calibrated severity difficulties via the Children’s Sleep
and duration of use are not recorded. scores (total CSS) provided a measure Habits Questionnaire (CSHQ),61

PEDIATRICS Volume 136, number 6, December 2015 1053

which measures 8 domains: bedtime by using the adaptive false discovery Associations With Overweight and
resistance, sleep onset latency, sleep rate procedure (FDR).67 Obesity Among Children With ASDs
duration, anxiety around sleep, night (ATN Sample)
awakenings, sleep-disordered Associations With Overweight and Results are presented in Table 4.
breathing, parasomnias, and morning Obesity in Children With ASDs
After multivariate adjustment, age
waking and daytime sleepiness. The Bivariate and multivariate logistic (6–11 years), black race, Hispanic or
total sleep disturbance score is the regression models examined factors Latino ethnicity, and lower parental
sum of scores across 33 items and associated with overweight and education retained associations with
served as a continuous measure of obesity among children in the ATN. overweight status. Likewise, after
child sleep difficulties. In a separate Sample size for each analysis differed multivariate adjustment, age ,12
questionnaire, parents reported because of missing data. To account
years, Hispanic or Latino ethnicity,
whether they currently had concerns for potential bias, we performed
lower parental education, and sleep
about their child’s gastrointestinal multiple imputation under the
(GI) difficulties (“gastrointestinal and affective problems retained
missing at random assumption to
[belly] problems [diarrhea, impute missing values.68,69 associations with obese weight status
constipation, pain]”) as a “yes”/“no” Additional details of the multiple (Table 4). For each 1-unit increase in
response. Finally, parents completed imputation procedure are reported in CSHQ scores, adjusted odds of obesity
the Child Behavior Checklist (CBCL), the Supplemental Appendix. For were 1.01 times greater. Similarly, the
a validated parent questionnaire used analyses, IQ standard scores and presence of affective problems on the
to assess behavioral and emotional CBCL T scores were treated as CBCL was associated with 1.26 times
problems in both the general categorical variables (,70 and $70, the odds of obesity.
population and in ASD.62 The CBCL respectively), whereas ADOS CSS, We conducted additional analyses of
Anxiety Problems scale includes CSHQ total sleep disturbance, and specific medication classes by
items identified by experts as related VABS-II Adaptive Behavior Composite classifying children into 3 groups
to generalized anxiety disorder and scores were treated as continuous. based on their BMI percentiles:
specific phobias. The Affective healthy weight ($5 to ,85),
Problems scale includes anxiety/ overweight but not obese ($85 to
depression, somatic complaints, and
RESULTS
,95), and obese ($95). These
withdrawal.63,64 The CBCL Attention Overweight and Obesity Among analyses (not shown in tables; all P
Problems scale includes items related Children With ASDs Compared With values adjusted with FDR and
to inattention and hyperactivity the General Population Cramer’s V measure of effect size are
associated with ADHD. Table 2 displays the characteristics of reported) revealed no associations
children with ASDs in the ATN. between BMI category and melatonin
Statistical Analyses
Compared with the general use (V = 0.03), dietary interventions
Overweight and Obesity Among Children population, prevalence of overweight (V = 0.03), stimulants (V = 0.03),
With ASDs Compared With the General and obesity tended to be higher nonstimulant ADHD medications (V =
Population among children with ASDs (Table 3), 0.01), and anticonvulsants (V = 0.01).
The goal of these analyses was to but differences in overall rates were Healthy weight children were less
determine whether prevalence of significant only among non-Hispanic frequently prescribed selective
overweight ($85th percentile) or white (ages 2–17) and Hispanic (ages serotonin reuptake inhibitors (4.8%)
obesity ($95th) was greater in 2–11) subgroups. Within age than overweight (7.0%; V = 0.04, P =
children with ASDs than in a general categories, prevalences of overweight .02) or obese children (7.9%; V =
population sample (NHANES). and obesity were significantly higher 0.06, P , .001). Compared with obese
Underweight children were included among young children (age 2–5 children, healthy weight children
in both samples. Because of years) with ASDs compared with the were less frequently prescribed
NHANES’s complex sampling general population, except among atypical neuroleptics (4.8% vs 8.1%;
structure, we conducted all analyses non-Hispanic black children. V = 0.06, P , .001) and asthma and
after applying sampling weights, Likewise, prevalences of overweight allergy medications (7.1% vs 10.1%;
using the R Survey package.65,66 and obesity were significantly higher V = 0.05, P = .02). However, total
Weighted NHANES prevalence among adolescents (ages 12–17 psychotropic medications prescribed
estimates were compared with those years) with ASDs compared with the (range 0–5) was significantly
in the ATN sample via z tests. Within general population. However, for ages associated with BMI category
each set of comparisons (overweight 6 to 11 years, no prevalence (Kruskal–Wallis rank sum test x2 =
and obesity), we adjusted P values to differences were found between the 2 10.2, P = .006). Pairwise
control Type I error rate at q , 0.05 samples. Mann–Whitney U tests revealed that

1054 PRESMANES et al
TABLE 2 Sample Characteristics for Children With ASDs in the ATN by BMI Percentile Ranges (N = 5053)
n (%) or Mean (SD) Omnibus Test Statistica

,5th $5th to , 85th $85th to ,95th $95th

N 237 3118 789 909
Age x 2 = 33.26 (P , .001)
2–5 y 139 (58.6) 1905 (61.1)a 484 (61.3)a 483 (53.1)b
6–11 y 76 (32.1) 995 (31.9)a,b 235 (29.8)b 317 (34.9)a
12–17 y 22 (9.3) 218 (7.0)a 70 (8.9)a 109 (12.0)b
Gender x 2 = 0.09 (P = .96)
Male 204 (86.1) 2629 (84.3) 667 (84.5) 770 (84.7)
Female 33 (13.9) 489 (15.7) 122 (15.5) 139 (15.3)
Race x 2 = 11.16 (P = .02)
a a a
White 175 (73.8) 2396 (76.8) 599 (75.9) 695 (76.4)
Black 14 (5.9) 184 (5.9)a 60 (7.6)a,b 74 (8.1)b
All other races or .1 race 34 (14.3) 394 (12.6)a 87 (11.0)a,b 92 (10.1)b
Missing 14 (5.9) 144 (4.6) 43 (5.4) 48 (5.3)
Ethnicity x 2 = 33.16 (P , .001)
Hispanic or Latino 12 (5.0) 256 (8.2)a 93 (11.8)b 133 (14.6)b
Non-Hispanic or Latino 217 (91.6) 2728 (87.5)a 657 (83.3)b 743 (81.7)b
Missing 8 (3.4) 134 (4.3) 39 (4.9) 33 (3.6)
Parent education x 2 = 22.92 (P , .001)
High school or less 29 (12.2) 431 (13.8)a 121 (15.3)a,b 176 (19.4)b
Some college 58 (24.5) 790 (25.3)a 207 (26.2)a 262 (28.8)a
College graduate or more 134 (56.5) 1658 (53.2)a 402 (50.9)a 422 (46.4)b
Missing 16 (6.7) 239 (7.7) 59 (7.5) 49 (5.4)
Behavioral functioning
ADOS CSS (10.2% missing)b 7.4 (1.9) 7.2 (1.9) 7.1 (1.9) 7.3 (1.8)
VABS-II Adaptive Behavior (14.8% missing)c 71.8 (12.0) 72.1 (12.2) 71.3 (12.8) 70.0 (11.6)
Full-scale IQ ,70 (25.0% missing) 58 (24.5) 1003 (32.2) 256 (32.4) 310 (34.1) x 2 = 1.84 (P = .40)
Treatments
Any psychotropic drug 68 (28.8) 823 (26.5)a 224 (28.6)a,b 286 (31.7)b x 2 = 9.27 (P = .01)
Any CAM 96 (19.4) 661 (21.2) 178 (22.6) 168 (18.5) x 2 = 4.70 (P = .09)
Comorbid problems
CSHQ Sleep (26.5% missing) 47.8 (9.0) 48.0 (9.0) 48.1 (8.8) 49.3 (9.5)
GI disturbance 79 (33.3) 890 (28.5) 205 (26.0) 247 (27.2) x2 = 2.32 (P = .31)
CBCL Anxiety $70 (9.9% missing) 62 (26.2) 736 (23.6) 156 (19.8) 221 (24.3) x2 = 5.10 (P = .08)
CBCL Affective $70 (9.9% missing) 66 (27.8) 794 (25.5) 178 (22.6) 281 (30.9) x2 = 17.69 (P , .001)
CBCL ADHD $70 (9.9% missing) 42 (20.7) 647 (20.7) 153 (19.4) 198 (21.8) x2 = 1.31 (P = .52)
ADOS CSS, Autism Diagnostic Observation Schedule Calibrated Severity Score.
BMI for age percentiles based on CDC growth charts. For each variable, if the omnibus test statistic was less than P = .05, post hoc comparisons were conducted. Column values within
the same row that differ at least at the P = .05 level are denoted by different superscripts (eg, 5a vs 10b); column values within the same row that share the same superscript did not
differ (eg, 5a vs 6a). See Table 4 for corresponding analyses involving multiply imputed data and Supplemental Table 1 for test statistics based on complete case analysis.
a Analyses exclude children with BMI ,5th percentile.
b ADOS CSSs range from 1 to 10.
c VABS-II Adaptive Behavior Composite standard scores (mean = 100, SD = 15).

children in the obesity group received children in the general population obese weight status among children
more medications than those in the from NHANES. Prevalence of with Asperger syndrome and autism
healthy weight group (Cohen’s d = overweight and obesity among compared with control children in
0.14; FDR adjusted P = .005). children with ASDs was significantly every age category (2–5, 6–11, 12–15,
higher at younger age (2–5 years) and and 16–20 years). In our analyses,
in adolescence (12–17 years) prevalence of overweight and obesity
DISCUSSION compared with the general was consistently higher for ASDs,
In this multi-institutional sample of population sample from NHANES. except among children with ASDs
children with ASDs, 33.6% of children These prevalence estimates are ages 6–11 years. One explanation for
met criteria for overweight ($85th consistent with recently reported this discrepancy may be that Broder-
BMI percentile), and 18% met criteria estimates based on measured height Fingert’s control group had lower
for obesity ($95th BMI percentile). and weight in people with ASDs.45,70 prevalence of overweight (inclusive of
The prevalence estimate for For example, Broder-Fingert et al45 obesity) than this study’s general
overweight is comparable to the found significantly elevated rates of population sample. For example,
31.8% prevalence among same-age overweight (exclusive of obesity) and among children age 6–11 years,

PEDIATRICS Volume 136, number 6, December 2015 1055

TABLE 3 Comparisons of Prevalence Estimates for Overweight and Obesity ($85th and $95th Percentile for Age and Gender, Respectively) Between the
ATN and NHANES Data Sets
Age Range, y Unweighted Overweight, % (95% CI) z P Obese, % (95% CI) z P
Sample Sizesa

ATN NHANESb ATN NHANESb ATN NHANESb

b
All
All (2–17) 5053 8844 33.6 (32.3–35.0) 31.8 (30.5–33.0) 1.86 .057 18.0 (17.0–19.0) 16.7 (15.7–18.0) 1.57 .120
2–5 3011 2627 32.1 (30.5–33.8) 23.4 (21.2–25.7) 6.06 ,.001 16.0 (14.8–17.4) 10.1 (8.8–11.6) 6.02 ,.001
6–11 1623 3678 34.0 (31.8–36.4) 34.2 (32.5–36.1) 20.15 .464 19.5 (17.7–21.5) 18.5 (17.1–20.0) 0.84 .303
12–17 419 2539 42.7 (38.1–47.5) 35.3 (33.1–37.5) 2.79 .006 26.0 (22.0–30.4) 19.5 (17.9–21.3) 2.79 .010
Boysc
All (2–17) 4270 4543 33.7 (32.3–35.1) 32.5 (30.8–34.2) 1.02 .206 18.0 (16.9–19.2) 17.5 (16.0–19.1) 0.54 .371
2–5 2531 1375 32.2 (30.4–34.1) 24.5 (21.8–27.4) 4.48 ,.001 16.0 (14.6–17.4) 11.0 (9.2–13.1) 4.01 ,.001
6–11 1384 1866 34.2 (31.8–36.8) 34.2 (31.7–36.9) 0.01 .498 20.2 (18.1–22.4) 19.3 (17.7–21.1) 0.60 .364
12–17 355 1302 41.4 (36.4–46.6) 36.4 (32.9–40.1) 1.56 .091 24.5 (20.3–29.2) 20.3 (17.5–23.3) 1.55 .120
Girlsc
All (2–17) 783 4301 33.3 (30.1–36.7) 31.2 (29.4–33.0) 1.13 .184 17.7 (15.2–20.6) 15.9 (14.8–17.2) 1.19 .201
2–5 480 1252 31.5 (27.5–35.8) 22.1 (19.0–25.6) 3.43 .001 16.5 (13.4–20.0) 9.2 (7.2–11.7) 3.54 .001
6–11 239 1812 32.6 (27.0–38.8) 34.2 (31.5–37.1) 20.48 .378 15.9 (11.8–21.1) 17.6 (15.9–19.5) 20.69 .348
12–17 64 1237 50.0 (38.1–61.9) 34.1 (31.2–37.2) 2.46 .014 34.4 (23.9–46.6) 18.8 (17.0–20.8) 2.58 .013
Non-Hispanic white
All (2–17) 3486 2553 32.3 (30.8–33.9) 28.8 (26.8–30.9) 2.65 .009 17.3 (16.1–18.6) 14.2 (12.4–16.1) 2.71 .010
2–5 1994 778 31.4 (29.4–33.5) 20.6 (17.4–24.2) 5.32 ,.001 15.7 (14.2–17.4) 7.2 (5.5–9.3) 6.68 ,.001
6–11 1170 1046 31.4 (28.8–34.1) 30.6 (27.7–33.7) 0.37 .396 17.6 (15.5–19.9) 15.4 (13.0–18.1) 1.30 .179
12–17 322 729 41.9 (36.7–47.4) 32.3 (29.2–35.7) 2.97 .005 25.8 (21.3–30.8) 17.4 (14.8–20.4) 2.96 .007
Non-Hispanic black
All (2–17) 277 2194 37.9 (32.4–43.7) 36.4 (34.1–38.8) 0.46 .378 21.3 (16.9–26.5) 21.1 (19.1–23.3) 0.06 .573
2–5 177 620 32.2 (25.8–39.4) 25.2 (22.0–28.7) 1.79 .061 16.9 (12.1–23.2) 13.6 (10.9–16.9) 1.03 .241
6–11 70 912 47.1 (35.9–58.7)d 39.8 (36.2–43.5) – – 31.4 (21.8–43.0)d 24.0 (20.8–27.6) – –
12–17 30 662 50.0 (33.2–66.8)d 40.8 (36.6–45.1) – – 23.3 (11.8–41.0)d,e 23.4 (19.6–27.8) – –
Hispanicf
All (2–17) 494 3189 45.7 (41.4–50.2) 38.8 (37.3–40.4) 2.91 .005 26.9 (23.2–31.0) 22.0 (20.9–23.2) 2.35 .023
2–5 306 950 42.8 (37.4–48.4) 29.8 (27.3–32.5) 4.15 ,.001 22.9 (18.5–27.9) 15.5 (13.4–18.0) 2.73 .010
6–11 159 1372 50.9 (43.2–58.6)d 43.0 (40.6–45.6) – – 33.3 (26.5–41.0)d 24.8 (22.9–26.8) – –
12–17 29 867 48.3 (31.4–65.5)d 41.5 (37.9–45.3) – – 34.5 (19.9–52.6)d,e 24.3 (21.7–27.1) – –
Positive z scores indicate that the ATN prevalence is greater than that in NHANES. All P values are adjusted; see text for details. 95% CIs calculated with logit transformation.
a Including underweight children in both samples.
b Data from NHANES years 2007 to 2008, 2009 to 2010, and 2011 to 2012; prevalence estimates are weighted with 6-y weights (see the Appendix for details).
c Includes other race and ethnic groups not shown separately, including multiracial, non-Hispanic Asian, American Indian or Alaskan Native, Native Hawaiian, or Pacific Islander.
d Sample size ,50 and are excluded from significance testing.
e Relative standard errors .25% but ,35%.
f For both ATN and NHANES, children whose parents reported Hispanic or Latino origin were categorized as Hispanic or Latino regardless of their race.

,20% of children in Broder-Fingert’s children in the general population,2 and the general population among
sample had BMI $85th percentile for these findings may suggest a different children of non-Hispanic black origin.
gender and age, compared with trajectory of weight gain among Environmental factors associated
34.2% in NHANES. children with ASD. The lack of with obesity, such as socioeconomic
Examination of cross-sectional differences in the prevalence of status, are probably already elevated
prevalence estimates (Table 3) also overweight and obesity between the among black children71,72 and may
suggests the possibility of different ages of 6 and 11 years might reflect overshadow additional risks
age-related trends among children a stabilizing period, in which children associated with ASDs. Alternatively,
with ASDs. For example, in the general with ASDs who gained weight earlier given that the ATN constitutes
population, prevalence of overweight remain in the same BMI category. In a referred sample, black children in
was 10.9% higher among children contrast, children in the general the ATN may be of higher
ages 6 to 11 years than among those population may be more likely to gain socioeconomic status and therefore
ages 2–5 years. In contrast, in the excess weight at older ages. Future differ less systematically than white
ATN, prevalence was only 1.9% longitudinal analyses could explore children, regardless of ASD status.
higher among children age 6–11 these trends in greater detail. However, this latter explanation
versus 2–5 years. Because obesity One surprising finding was the lack of would be inconsistent with the robust
becomes more prevalent among older differences between the ASD sample differences we found between

1056 PRESMANES et al
TABLE 4 Multivariate Analyses Using Multiple Imputation (N = 4816) to Predict Overweight and Obesity ($85th and $95th Percentile for Age and
Gender, Respectively) Among Children With ASDs
Variable n (%) Complete OR (95% CI)

Univariate (Crude OR) Multivariate (Adjusted OR)

Overweight Obesity Overweight Obesity

Age, n (%) 4816 (100.0)
2–5 y Reference Reference Reference Reference
6–11 y 1.09 (0.96–1.24) 1.27 (1.09–1.49)** 1.12 (0.97–1.30) 1.35 (1.13–1.60)**
12–17 y 1.62 (1.31–2.00)** 1.87 (1.47–2.38)** 1.62 (1.28–2.05)** 1.95 (1.49–2.60)**
Male, n (%) 4816 (100.0) 1.02 (0.87–1.21) 1.03 (0.84–1.25) 1.01 (0.86–1.20) 1.02 (0.83–1.25)
Race, n (%) 4581 (95.1)
White Reference Reference Reference Reference
Black 1.37 (1.09–1.72)** 1.34 (1.02–1.76)* 1.27 (1.00–1.60)* 1.22 (0.92–1.62)
All other races 0.83 (0.69–1.01) 0.82 (0.65–1.04) 0.85 (0.70–1.04) 0.86 (0.67–1.09)
Hispanic or Latino, n (%) 4610 (95.7) 1.72 (1.42–2.08)** 1.72 (1.39–2.13)** 1.66 (1.37–2.02)** 1.63 (1.30–2.03)**
Parent education, n (%) 4489 (93.2)
High school or less Reference Reference Reference Reference
Some college 0.88 (0.72–1.07) 0.85 (0.67–1.06) 0.92 (0.75–1.13) 0.88 (0.70–1.12)
College graduate or more 0.73 (0.62–0.88)** 0.66 (0.54–0.82)** 0.81 (0.67–0.97)* 0.75 (0.60–0.94)*
Behavioral functioning
ADOS CSS, mean (SD) 4322 (89.7) 1.00 (0.97–1.04) 1.02 (0.98–1.07) 0.99 (0.96–1.03) 1.02 (0.98–1.06)
VABS-II Adaptive Behavior, mean (SD) 4102 (85.2) 0.99 (0.99–1.00)** 0.99 (0.99–1.00)* 1.00 (0.99–1.00) 1.00 (0.99–1.00)
Full-scale IQ ,70, n (%) 3620 (75.2) 1.10 (0.96–1.24) 1.08 (0.92–1.27) 1.04 (0.90–1.21) 1.10 (0.91–1.33)
Treatments
Any psychotropic drugs 4816 (100.0) 1.20 (1.05–1.36)** 1.25 (1.07–1.47)** 1.11 (0.96–1.28) 1.06 (0.88–1.26)
Any CAM 4816 (100.0) 0.95 (0.82–1.10) 0.83 (0.69–0.98)* 1.01 (0.87–1.18) 0.87 (0.72–1.05)
Comorbid problems
CSHQ Sleep, mean (SD) 3538 (73.5) 1.01 (1.00–1.01)* 1.02 (1.01–1.02)** 1.01 (1.00–1.02) 1.01 (1.00–1.02)*
GI disturbance, n (%) 4816 (100.0) 0.91 (0.79–1.04) 0.96 (0.81–1.13) 0.88 (0.77–1.02) 0.92 (0.77–1.09)
CBCL Anxiety $70, n (%) 4339 (90.1) 0.97 (0.84–1.12) 1.12 (0.95–1.33) 0.86 (0.73–1.01) 0.91 (0.75–1.10)
CBCL Affective $70, n (%) 4339 (90.1) 1.10 (0.96–1.26) 1.36 (1.16–1.60)** 1.06 (0.90–1.25) 1.26 (1.04–1.53)*
CBCL ADHD $70, n (%) 4338 (90.1) 1.02 (0.88–1.19) 1.09 (0.91–1.30) 0.95 (0.81–1.12) 0.94 (0.78–1.14)
ADOS CSS, Autism Diagnostic Observation Schedule Calibrated Severity Score; OR, odds ratio. * P , .05; ** P , .01. Variables without missing data were present in the imputation model
but were not imputed.

Hispanic children in the ASD and problems are a cause or and anxiety problems were not
general populations, because a consequence of obesity. Repeated associated with overweight or
Hispanic children may also have measures could clarify these obesity. Thus, interventions that take
elevated environmental risks.73 associations and might reveal into account both general risk factors
Because the sample sizes of non- important inroads to prevention and for unhealthy weight and those that
Hispanic black children with ASDs in treatment of overweight and obesity are ASD specific may hold promise for
the ATN were small, group estimates among children with ASD. improved weight status in ASDs.
may also be less reliable. Notably, some variables had no This study has limitations. Because it
Among children with ASDs, there association with unhealthy weight is a secondary data analysis, there
were several notable associations among children with ASDs. In was limited detail about
between sociodemographic variables contrast to previous studies,70,74 sociodemographics, developmental
and unhealthy weight. Multivariate there was no significant association and family history, GI problems, and
analyses revealed that older age, between severity of ASD symptoms, medication dosages or duration of
Hispanic or Latino ethnicity, lower and neither adaptive nor intellectual use. Our analysis of medications was
parent education, and sleep and functioning was associated with limited by the available data in the
affective problems were significantly overweight or obesity in multivariate ATN; other medications may have an
associated with obesity. Many of these models. In contrast to studies of impact on obesity that we were
factors confirm previous findings in typically developing children75,76 but unable to estimate. For example, as 1
a smaller sample of children with consistent with previous research in reviewer noted, medications with
ASDs in Oregon48 and another recent children with ASDs,77 GI problems soporific effects could be linked to
large-scale study.45 Because our study were not linked to overweight or unhealthy weight status, but we were
is cross-sectional, it is not clear obesity. Also in contrast to findings in unable to explore these types of
whether comorbid sleep and affective the general population,41,42 ADHD associations with the data collected.

PEDIATRICS Volume 136, number 6, December 2015 1057

The effect of parent education levels representative sample of children talk with families early about the risk
on children’s weight status may also with ASDs, given a median age of of unhealthy weight in ASDs,
be underestimated in this sample, diagnosis of 4.4 years of age in the particularly when other comorbid
given the slightly skewed range in the United States16 and that enrollment conditions exist.
ATN (,2.2% had parents with less in the ATN registry can often occur at
than high school education); to the time of diagnosis. The clinic-
preserve statistical power, we did not referred sample of children available
in the ATN may also have more ABBREVIATIONS
analyze this category separately. In
addition, although highly correlated frequent or more severe health ADHD: attention-deficit/
with body fat, BMI is an imperfect problems than the larger population hyperactivity disorder
measure because it does not of children with ASDs. ADOS: Autism Diagnostic
distinguish between fat and lean body Observation Schedule
mass.78,79 Children of different ages, CONCLUSIONS ASD: autism spectrum disorder
genders, and race and ethnicity ATN: autism treatment network
Despite these limitations, this is CAM: complementary and
groups may differ in body fat the first multicenter study to assess
composition despite having similar alternative medications or
unhealthy weight risk in ASDs, as treatments
BMI.81 We could not measure several well as overweight and obesity risk
variables that are likely to be CBCL: Child Behavior Checklist
factors, in a population with both CDC: Centers for Disease Control
important for BMI such as dietary verified ASDs and directly measured and Prevention
intake and physical activity. In biometrics. The study provides CI: confidence interval
addition, there was no measure of strong confirmatory evidence that CSHQ: Children’s Sleep Habits
parental BMI or family young children with ASDs are at Questionnaire
environment,80 which are associated risk for unhealthy weight trajectories CSS: calibrated severity score
with children’s BMI. Finally, in and that the presence of sleep or GI: gastrointestinal
interpreting findings, it is important affective problems may confer VABS-II: Vineland Adaptive
to note that the group of children increased risk. The findings suggest Behavior Scales
ages 2 to 5 years may be the most that health care providers should

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: This research was conducted by using data collected as part of the Autism Treatment Network (ATN). The ATN is funded by Autism Speaks and
a cooperative agreement (UA3 MC 11054) from the Health Resources and Services Administration to Massachusetts General Hospital. This project was funded by an
ATN/Autism Intervention Research Network on Physical Health grant (principal investigator: Dr Fombonne). Dr Zuckerman’s effort was funded by K23MH095828 from
the US National Institute of Mental Health. Funded by the National Institutes of Health (NIH).
POTENTIAL CONFLICT OF INTEREST: Drs Zuckerman and Fombonne have received grants from Autism Speaks for other projects. Dr Fombonne has served as an expert
witness on autism-related cases for AbbVie and GlaxoSmithKline. Dr Hill has indicated she has no potential conflicts of interest to disclose.

REFERENCES
1. Kuczmarski RJ, Ogden CL, Guo SS, et al. in children and adolescents. Pediatrics. risk factors and excess adiposity among
2000 CDC growth charts for the United 2006;117(6):2167–2174 overweight children and adolescents:
States: methods and development. Vital 5. Goran MI, Ball GDC, Cruz ML. Obesity and the Bogalusa Heart Study. J Pediatr.
Health Stat 11. 2002; No. 246:1–190 risk of type 2 diabetes and 2007;150(1):12–17.e2
2. Ogden CL, Carroll MD, Kit BK, Flegal KM. cardiovascular disease in children and 8. Must A, Phillips SM, Naumova EN.
Prevalence of childhood and adult adolescents. J Clin Endocrinol Metab. Occurrence and timing of childhood
obesity in the United States, 2011–2012. 2003;88(4):1417–1427 overweight and mortality: findings from
JAMA. 2014;311(8):806–814 6. Friedemann C, Heneghan C, Mahtani K, the Third Harvard Growth Study. J
3. Bixler EO, Vgontzas AN, Lin HM, et al. Thompson M, Perera R, Ward AM. Pediatr. 2012;160(5):743–750
Sleep disordered breathing in children Cardiovascular disease risk in healthy 9. Trasande L, Chatterjee S. The impact of
in a general population sample: children and its association with body obesity on health service utilization and
prevalence and risk factors. Sleep. 2009; mass index: systematic review and meta- costs in childhood. Obesity (Silver
32(6):731–736 analysis. BMJ. 2012;345:e4759 Spring). 2009;17(9):1749–1754
4. Taylor ED, Theim KR, Mirch MC, et al. 7. Freedman DS, Mei Z, Srinivasan SR, 10. Wang G, Dietz WH. Economic burden of
Orthopedic complications of overweight Berenson GS, Dietz WH. Cardiovascular obesity in youths aged 6 to 17 years:

1058 PRESMANES et al
 1979–1999. Pediatrics. 2002;109(5). 22. Ahearn WH, Castine T, Nault K, Green G. associated with gene dosage at the
Available at: www.pediatrics.org/cgi/ An assessment of food acceptance in chromosome 16p11.2 locus. Nature.
content/full/109/5/e81 children with autism or pervasive 2011;478(7367):97–102
11. Lobstein T, Baur L, Uauy R; IASO developmental disorder–not otherwise
34. Weiss LA, Shen Y, Korn JM, et al; Autism
International Obesity TaskForce. Obesity specified. J Autism Dev Disord. 2001;
Consortium. Association between
in children and young people: a crisis in 31(5):505–511
microdeletion and microduplication at
public health. Obes Rev. 2004;5(suppl 1): 23. Williams KE, Gibbons BG, Schreck KA. 16p11.2 and autism. N Engl J Med. 2008;
4–104 Comparing selective eaters with and 358(7):667–675
12. Grondhuis SN, Aman MG. Overweight and without developmental disabilities. J Dev
35. Shinawi M, Sahoo T, Maranda B, et al.
obesity in youth with developmental Phys Disabil. 2005;17(3):299–309
11p14.1 microdeletions associated with
disabilities: a call to action. J Intellect 24. Evans EW, Must A, Anderson SE, et al. ADHD, autism, developmental delay, and
Disabil Res. 2014;58(9):787–799 Dietary patterns and body mass index in obesity. Am J Med Genet A. 2011;155A(6):
13. Janssen I, Craig WM, Boyce WF, Pickett W. children with autism and typically 1272–1280
Associations between overweight and developing children. Res Autism Spectr
36. Krakowiak P, Goodlin-Jones B, Hertz-
obesity with bullying behaviors in Disord. 2012;6(1):399–405
Picciotto I, Croen LA, Hansen RL. Sleep
school-aged children. Pediatrics. 2004; 25. Must A, Phillips SM, Curtin C, et al. problems in children with autism
113(5):1187–1194 Comparison of sedentary behaviors spectrum disorders, developmental
14. Strauss RS, Pollack HA. Social between children with autism spectrum delays, and typical development:
marginalization of overweight children. disorders and typically developing a population‐based study. J Sleep Res.
Arch Pediatr Adolesc Med. 2003;157(8): children. Autism. 2014;18(4):376–384 2012;21(2):231–231
746–752 26. Mazurek MO, Engelhardt CR. Video game 37. Leyfer OT, Folstein SE, Bacalman S, et al.
15. Curtin C, Jojic M, Bandini LG. Obesity in use in boys with autism spectrum Comorbid psychiatric disorders in
children with autism spectrum disorder. disorder, ADHD, or typical development. children with autism: interview
Harv Rev Psychiatry. 2014;22(2):93–103 Pediatrics. 2013;132(2):260–266 development and rates of disorders. J
27. Macdonald M, Esposito P, Ulrich D. The Autism Dev Disord. 2006;36(7):849–861
16. Centers for Disease Control and
Prevention (CDC). Prevalence of autism physical activity patterns of children 38. Dev DA, McBride BA, Fiese BH, Jones BL,
spectrum disorder among children aged with autism. BMC Res Notes. 2011;4(1): Cho H; on behalf of the Strong Kids
8 years: autism and developmental 422 Research Team. Risk factors for
disabilities monitoring network, 11 sites, 28. Bandini LG, Gleason J, Curtin C, et al. overweight/obesity in preschool
United States, 2010. MMWR Surveill Comparison of physical activity between children: an ecological approach. Child
Summ. 2014;63(2):1–21 children with autism spectrum Obes. 2013;9(5):399–408
17. Kipping RR, Jago R, Lawlor DA. Obesity in disorders and typically developing 39. Delgado-Aros S, Locke GR III, Camilleri M,
children. Part 1: epidemiology, children. Autism. 2013;17(1):44–54 et al. Obesity is associated with
measurement, risk factors, and 29. Coury DL, Anagnostou E, Manning- increased risk of gastrointestinal
screening. BMJ. 2008;337(7675):a1824 Courtney P, et al Use of psychotropic symptoms: a population-based study. Am
medication in children and adolescents J Gastroenterol. 2004;99(9):1801–1806
18. Kirk SFL, Penney TL, McHugh T-LF.
Characterizing the obesogenic with autism spectrum disorders. 40. Talley NJ, Howell S, Poulton R. Obesity
environment: the state of the evidence Pediatrics. 2012;130(suppl 2):S69–76 and chronic gastrointestinal tract
with directions for future research. Obes 30. Correll CU, Manu P, Olshanskiy V, symptoms in young adults: a birth
Rev. 2010;11(2):109–117 Napolitano B, Kane JM, Malhotra AK. cohort study. Am J Gastroenterol. 2004;
Cardiometabolic risk of second- 99(9):1807–1814
19. Schreck KA, Williams K, Smith AF. A
comparison of eating behaviors between generation antipsychotic medications 41. Waring ME, Lapane KL. Overweight in
children with and without autism. J during first-time use in children and children and adolescents in relation to
Autism Dev Disord. 2004;34(4):433–438 adolescents. JAMA. 2009;302(16): attention-deficit/hyperactivity disorder:
1765–1773 results from a national sample.
20. Sharp WG, Berry RC, McCracken C, et al.
31. De Hert M, Detraux J, van Winkel R, Yu W, Pediatrics. 2008;122(1). Available at:
Feeding problems and nutrient intake in
Correll CU. Metabolic and cardiovascular www.pediatrics.org/cgi/content/full/122/
children with autism spectrum
adverse effects associated with 1/e1
disorders: a meta-analysis and
comprehensive review of the literature. antipsychotic drugs. Nat Rev Endocrinol. 42. Rofey DL, Kolko RP, Iosif A-M, et al. A
J Autism Dev Disord. 2013;43(9): 2012;8(2):114–126 longitudinal study of childhood
2159–2173 32. Williamson ED, Martin A. Psychotropic depression and anxiety in relation to
medications in autism: practical weight gain. Child Psychiatry Hum Dev.
21. Martins Y, Young RL, Robson DC. Feeding
considerations for parents. J Autism Dev 2009;40(4):517–526
and eating behaviors in children with
autism and typically developing children. Disord. 2012;42(6):1249–1255 43. Faith MS, Butryn M, Wadden TA,
J Autism Dev Disord. 2008;38(10): 33. Jacquemont S, Reymond A, Zufferey F, Fabricatore A, Nguyen AM, Heymsfield SB.
1878–1887 et al. Mirror extreme BMI phenotypes Evidence for prospective associations

PEDIATRICS Volume 136, number 6, December 2015 1059

 among depression and obesity in ed. Circle Pines, MN: American Guidance 68. Sterne JAC, White IR, Carlin JB, et al.
population-based studies. Obes Rev. Service; 2005 Multiple imputation for missing data in
2011;12(5):e438–e453 epidemiological and clinical research:
54. Roid G. Stanford–Binet Intelligence
potential and pitfalls. BMJ. 2009;338:
44. Phillips KL, Schieve LA, Visser S, et al. Scales. 5th ed. Rolling Meadows, IL:
b2393
Prevalence and impact of unhealthy Riverside Publishing; 2003
weight in a national sample of US 69. Stuart EA, Azur M, Frangakis C, Leaf P.
55. Wechsler D. Wechsler Intelligence Scale
adolescents with autism and other Multiple imputation with large data sets:
for Children. 4th ed. San Antonio, TX: The
learning and behavioral disabilities. a case study of the Children’s Mental
Psychological Corporation; 2003
Matern Child Health J. 2014;18(8): Health Initiative. Am J Epidemiol. 2009;
1964–1975 56. Elliot C. Differential Abilities Scale 169(9):1133–1139
Manual. 2nd ed. San Antonio, TX:
45. Broder-Fingert S, Brazauskas K, Lindgren 70. Egan AM, Dreyer ML, Odar CC, Beckwith
Harcourt Assessment, Inc.; 2007
K, Iannuzzi D, Van Cleave J. Prevalence of M, Garrison CB. Obesity in young
overweight and obesity in a large clinical 57. Wechsler D. Wechsler Preschool and children with autism spectrum
sample of children with autism. Acad Primary Scale of Intelligence. 3rd ed. disorders: prevalence and associated
Pediatr. 2014;14(4):408–414 San Antonio, TX: The Psychological factors. Child Obes. 2013;9(2):125–131
Corporation; 2002
46. Curtin C, Anderson SE, Must A, Bandini L. 71. Braveman P. A health disparities
The prevalence of obesity in children 58. Wechsler D. Wechsler Abbreviated Scale perspective on obesity research. Prev
with autism: a secondary data analysis of Intelligence. San Antonio, TX: Chronic Dis. 2009;6(3):A91
using nationally representative data Psychological Corporation; 1999
72. Drewnowski A, Darmon N. The economics
from the National Survey of Children’s
59. Roid GM, Miller LJ. Leiter International of obesity: dietary energy density and
Health. BMC Pediatr. 2010;10:11
Performance Scale–Revised: Examiners energy cost. Am J Clin Nutr. 2005;82(1
47. Curtin C, Bandini LG, Perrin EC, Tybor DJ, Manual. Wood Dale, IL: Stoelting Co.; 1997 suppl):265S–273S
Must A. Prevalence of overweight in
60. Mullen EM. Mullen Scales of Early 73. Taveras EM, Gillman MW, Kleinman K,
children and adolescents with attention
Learning. Circle Pines, MN: American Rich-Edwards JW, Rifas-Shiman SL.
deficit hyperactivity disorder and autism
Guidance Service; 1995 Racial/ethnic differences in early-life
spectrum disorders: a chart review. BMC
61. Owens JA, Spirito A, McGuinn M. The risk factors for childhood obesity.
Pediatr. 2005;5(1):48
Children’s Sleep Habits Questionnaire Pediatrics. 2010;125(4):686–695
48. Zuckerman KE, Hill AP, Guion K, Voltolina
(CSHQ): psychometric properties of 74. Ho HH, Eaves LC, Peabody D. Nutrient
L, Fombonne E. Overweight and obesity:
a survey instrument for school-aged intake and obesity in children with
prevalence and correlates in a large
children. Sleep. 2000;23(8):1043–1051 autism. Focus Autism Other Dev Disabil.
clinical sample of children with autism
62. Deprey L, Ozonoff S. Assessment of 1997;12(3):187–192
spectrum disorder. J Autism Dev Disord.
2014;44(7):1708–1719 comorbid psychiatric conditions in 75. Pashankar DS, Loening-Baucke V.
autism spectrum disorders. In: Golstein Increased prevalence of obesity in
49. American Psychiatric Association.
S, Goldstein S, Naglieri JA, Ozonoff S, eds. children with functional constipation
Diagnostic and Statistical Manual of
Assessment of Autism Spectrum evaluated in an academic medical
Mental Disorders. 4th ed. Washington,
Disorders. New York, NY: The Guilford center. Pediatrics. 2005;116(3). Available
DC: American Psychiatric Association;
Press; 2009:290–317 at: www.pediatrics.org/cgi/content/full/
2000
116/3/e377
63. Achenbach TM, Rescorla L. Manual for
50. Lord C, Risi S, Lambrecht L, et al. The
the ASEBA Preschool Forms & Profiles: 76. Teitelbaum JE, Sinha P, Micale M, Yeung
autism diagnostic observation schedule-
An Integrated System of Multi-Informant S, Jaeger J. Obesity is related to multiple
generic: a standard measure of social
Assessment. Burlington, VT: ASEBA; 2000 functional abdominal diseases. J Pediatr.
and communication deficits associated
2009;154(3):444–446
with the spectrum of autism. J Autism 64. Achenbach TM, Rescorla L. Manual for
Dev Disord. 2000;30(3):205–223 the ASEBA School-Age Forms & Profiles: 77. Gorrindo P, Williams KC, Lee EB, Walker
An Integrated System of Multi-Informant LS, McGrew SG, Levitt P. Gastrointestinal
51. Gotham K, Risi S, Pickles A, Lord C. The
Assessment. Burlington, VT: ASEBA; 2001 dysfunction in autism: parental report,
Autism Diagnostic Observation Schedule:
clinical evaluation, and associated
revised algorithms for improved 65. Lumley T. Survey: analysis of complex
factors. Autism Res. 2012;5(2):101–108
diagnostic validity. J Autism Dev Disord. survey samples. R package version 3.28-
2007;37(4):613–627 2. 2012 78. Ahima RS, Lazar MA. Physiology. The
health risk of obesity—better metrics
52. Gotham K, Pickles A, Lord C. 66. Lumley T. Analysis of complex survey
imperative. Science. 2013;341(6148):
Standardizing ADOS scores for samples. J Stat Softw. 2004;9(8):1–19
856–858
a measure of severity in autism
67. Benjamini Y, Hochberg Y. On the adaptive
spectrum disorders. J Autism Dev 79. Prentice AM, Jebb SA. Beyond body mass
control of the false discovery rate in
Disord. 2009;39(5):693–705 index. Obes Rev. 2001;2(3):141–147
multiple testing with independent
53. Sparrow SS, Balla DA, Cicchetti DV. statistics. J Educ Behav Stat. 2000;25(1): 80. Flegal KM, Ogden CL, Yanovski JA, et al.
Vineland Adaptive Behavior Scales. 2nd 60–83 High adiposity and high body mass

1060 PRESMANES et al
 index-for-age in US children and in China. Res Dev Disabil. 2009;30(1): 87. Memari AH, Kordi R, Ziaee V, Mirfazeli FS,
adolescents overall and by race–ethnic 70–76 Setoodeh MS. Weight status in Iranian
group. Am J Clin Nutr. 2010;91(4): 84. Chen AY, Kim SE, Houtrow AJ, Newacheck children with autism spectrum
1020–1026 PW. Prevalence of obesity among disorders: Investigation of underweight,
81. O’Brien M, Nader PR, Houts RM, et al. The children with chronic conditions. Obesity. overweight and obesity. Research in
ecology of childhood overweight: a 12- 2010;18(1):210–213 Autism Spectrum Disorders. 2012;6(1):
year longitudinal analysis. Int J Obes. 234–239
85. Rimmer JHJ, Yamaki KK, Lowry BMDB,
2007;31(9):1469–1478 Wang EE, Vogel LCL. Obesity and obesity- 88. Johnson CL, Paulose-Ram R, Ogden CL,
82. Whiteley P, Dodou K, Todd L, Shattock P. related secondary conditions in et al. National health and nutrition
Body mass index of children from the adolescents with intellectual/ examination survey: analytic guidelines,
United Kingdom diagnosed with developmental disabilities. J Intellectual 1999–2010. Vital Health Stat 2. 2013;
pervasive developmental disorders. Disabil Res. 2010;54(9):787–794 (161):1–24
Pediatrics International. 2004;46(5): 86. Hyman SL, Stewart PA, Schmidt B, et al. 89. van Buuren S, Groothuis-Oudshoorn K.
531–533 Nutrient intake from food in children MICE: Multivariate imputation by chained
83. Xiong N, Ji C, Li Y, He Z, Bo H, Zhao Y. The with autism. Pediatrics. 2012; equations in R. J Statistical Software.
physical status of children with autism 130(suppl 2):S145–S153 2011;45(3):1–67

HIDDEN FEES: I tend to scour the internet for the best rates for airline travel, hotel
rooms, and concert tickets. While I dislike the unexplained “fees” that are added to
the cost, particularly for concert tickets, I principally hate not knowing the true cost
of the product until the very end of the transaction. While airline fees and taxes are
quite high, at least most online airline pricing sites are fairly good at presenting the
total cost of the flight early in the process. I tend to find that buying a concert ticket is
remarkably galling as trying to understand the checkout price (i.e., the total cost of
the ticket) is extremely challenging. I was quite happy when the company with the
largest share of the $6 billion live-event ticket market decided to shift to “all-in”
pricing where the total cost of the ticket – including any convenience fees – is shown
up front.
As reported in The Wall Street Journal (Business: August 31, 2015), however, other
resellers did not follow suit. Their prices, at least at first glance, appeared much
better than the company using the “all-in” pricing strategy. This led to a precipitous
decline in business for the company using the “all-in” pricing strategy. It turns out
that while consumers purchasing concert tickets online routinely cite separate
service charges as their top annoyance, they really hate seeing that cost front loaded
into the sticker price. In a head to head comparison, shoppers were much more
inclined to purchase tickets with a lower introductory cost regardless of the final
cost. The company has since abandoned the “all-in” pricing policy. Those in the
industry are not surprised by the findings, stating that most e-consumers do not
consider the checkout price. After all, if one buys a 99-cent candy bar the checkout
cost is over a dollar.
While I cannot comment on the average e-consumer, I do know for a fact that if I
purchase a product in Vermont where I live, the state sales tax is 6% and there are no
other convenience fees—at least for now.
Noted by WVR, MD

PEDIATRICS Volume 136, number 6, December 2015 1061