Você está na página 1de 6

Food Selectivity in Children with Autism Spectrum Disorders and Typically

Developing Children
Linda G. Bandini, PhD, RD, Sarah E. Anderson, PhD, Carol Curtin, MSW, Sharon Cermak, EdD, OTR/L,
E. Whitney Evans, MS, RD, Renee Scampini, MS, RD, Melissa Maslin, MEd, and Aviva Must, PhD

Objectives To define food selectivity and compare indices of food selectivity among children with autism spec-
trum disorders (ASDs) and typically developing children, and to assess the impact of food selectivity on nutrient ad-
equacy.
Study design Food selectivity was operationalized to include food refusal, limited food repertoire, and high-
frequency single food intake using a modified food frequency questionnaire and a 3-day food record. Food selec-
tivity was compared between 53 children with ASDs and 58 typically developing children age 3-11 years. Nutrient
adequacy was assessed relative to the dietary reference intakes.
Results The children with ASDs exhibited more food refusal than typically developing children (41.7% of foods
offered vs 18.9% of foods offered; P <.0001). They also had a more limited food repertoire (19.0 foods vs 22.5 foods;
P <.001). Only 4 children with ASDs and 1 typically developing child demonstrated high-frequency single food in-
take. Children with a more limited food repertoire had inadequate intake of a greater number of nutrients.
Conclusions Our findings suggest that food selectivity is more common in children with ASDs than in typically
developing children, and that a limited food repertoire may be associated with nutrient inadequacies. (J Pediatr
2010;157:259-64).

F
ood selectivity, or ‘‘picky eating,’’ is often observed in young children and is a frequent cause of parental concern. Food
selectivity is more commonly reported in children with developmental disabilities, particularly in children with autism
spectrum disorders (ASDs), compared with typically developing children,.1,2 Despite numerous reports focusing on
pickiness, rigidity, selective eating, and mealtime food refusals in children with ASDs,3-7 a standardized definition of food se-
lectivity is lacking. In addition, the relationship of food selectivity to nutritional adequacy is unknown. We sought to operation-
alize the definition of food selectivity, to compare food selectivity in typically developing children and children with ASDs, and
to examine the relationship between food selectivity and nutritional adequacy.
Most previous studies of food selectivity in typically developing children have evaluated whether a child is a picky eater based
on parental report, usually with a single question8-13 or several items that tap specific picky/selective eating behaviors.14-17
Direct measures of food intake have not been used to define picky eating, and there is no standard operational definition of
picky or selective eating.8,15,16,18
These methodological problems extend to selective eating in children with ASDs. Parental and anecdotal clinical reports,
along with a few research studies,2,19 suggest that children with ASDs have unusual eating habits. These are often described
as overly selective, with aversions to specific textures, colors, smells, and temperatures and rigidity with respect to specific
brands of foods. The term ‘‘food selectivity’’ has been used variously to refer to food refusal, decreased variety, and restricted
intake to a few frequently eaten foods, with various approaches used to categorize food, such as focusing on nutritional
components (eg, high protein, high starch) or sensory aspects (eg, sticky, sweet). The lack of a standardized definition of
food selectivity limits the ability to assess and compare across populations of children.
To address these gaps, we have developed a definition of food selectivity based on our clinical experience and pilot studies
that describe eating patterns of children with ASDs. We operationalized food
selectivity to include 3 domains: food refusal, limited food repertoire, and
From the E.K. Shriver Center, University of
Massachusetts Medical School, Waltham, MA (L.B.,
C.C., R.S., M.M.) Department of Health Sciences, Boston
AI Adequate intake University, Boston, MA (L.B.); Division of Epidemiology,
ASDs Autism spectrum disorders College of Public Health, Ohio State University,
Columbus, OH (S.A.); Division of Occupational Science
BMI Body mass index and Occupational Therapy, University of Southern
DAS Differential Abilities Scale California, Los Angeles, CA (S.C.); Tufts University
Friedman School of Nutrition, Boston, MA (E.W.E.); and
EAR Estimated average requirement Public Health & Community Medicine, Tufts University
FFQ Food frequency questionnaire School of Medicine, Boston, MA (A.M.)
HFSFI High-frequency single food intake Supported by National Institutes of Health Grants R21
NDSR Nutrition Data System for Research HD048989, 2P30HD004147-33A2, and 5P30DK046200-
14. The authors declare no conflicts of interest.
VABS Vineland Adaptive Behavior Scale
YAQ Youth/Adolescent Food Frequency Questionnaire 0022-3476/$ - see front matter. Copyright Ó 2010 Mosby Inc.
All rights reserved. 10.1016/j.jpeds.2010.02.013

259
THE JOURNAL OF PEDIATRICS  www.jpeds.com Vol. 157, No. 2

high-frequency single food intake (HFSFI). We hypothesized cases where children could help keep the food record, both
that children with ASDs would exhibit more food selectivity the parent and child were given instructions. The completed
than typically developing children, and that food selectivity food record was returned to the research team by mail and
would decline with age in typically developing children, but reviewed by a registered dietitian or a nutrition graduate
would not be associated with age in children with ASDs. student; any uncertainties were clarified by telephone with
We also explored whether food selectivity is associated with the parent.
inadequate nutrient intake, which has important implica-
tions for nutritional management. Definition of Food Selectivity
We defined food selectivity as comprising 3 separate do-
mains: food refusal, limited food repertoire, and high-
Methods frequency single food intake (HFSFI). Food refusal and
HFSFI were assessed using a modified version of the
Participants in the Children’s Activity and Meal Patterns Youth/Adolescent Food Frequency Questionnaire (YAQ).25
Study (CHAMPS) included children with ASDs and typically Limited food repertoire was assessed using the 3-day food
developing children age 3-11 years. Participants were re- record.
cruited via public listings on the Internet, outreach to local
community programs (eg, schools, YMCAs), existing partic- Food Frequency Questionnaire
ipant databases at the University of Massachusetts Medical Developed for self-administration by children and adoles-
School’s Shriver Center, autism support organizations, and cents age 9-18 years in the Growing Up Today Study,25 the
the Interactive Autism Network Research Database at the YAQ is based on the original Harvard Food Frequency Ques-
Kennedy Krieger Institute. Inclusion criteria included good tionnaire.26 The YAQ has been shown to be reproducible and
health and free from diseases or disorders that could affect di- reliable.27,28 We modified the YAQ for parent (rather than
etary and/or physical activity habits (eg, diabetes, cystic fibro- child) report and to allow us to quantify how many foods
sis, chronic gastrointestinal illness, cerebral palsy). Exclusion children refused (food refusal) and whether children ate
criteria included the use of appetite-affecting medications, a large number of the same food(s) on a daily basis (ie,
including steroids, atypical antipsychotics, mood stabilizers, HFSFI), as described in Table I. Our revised FFQ
tricyclic antidepressants, anticonvulsants, and stimulants. contained 131 food items, compared with the original 126,
The study protocol was approved by the University of Mas- and asked parents to write in foods that were not included
sachusetts Medical School’s Institutional Review Board. on the FFQ if the child ate them at least once per week on
Written informed consent was obtained from the parent or average over the past year.
guardian of each child who participated. The parents were re-
imbursed for their time, and the children were given a gift Food Refusal
certificate to a bookstore. We assessed food refusal based on the absolute number of
The diagnosis of autism in children with ASDs was con- foods the parent indicated that the child would not eat, as
firmed via the Autism Diagnostic Interview-Revised.20 The
Vineland Adaptive Behavior Scales (VABS)21 was used to
characterize adaptive skills, and the Differential Abilities
Scale (DAS)22 was administered to assess cognitive ability.
Table I. Domains and definitions of food selectivity
Children were weighed and measured in light clothing
without shoes using a Seca (Hanover, Maryland) portable Domain Definition Source of data
scale and wall-mounted stadiometer. Body mass index Food refusal Number of foods that the Modified FFQ*
(BMI) was calculated from measures of height and weight child will not eat
Percentage of foods offered
(kg/m2) and referenced against the sex- and age-specific Cen- that the child will not eat
ters for Disease Control and Prevention childhood growth Limited food Number of unique foods 3-day food diary†
reference.23 Overweight was defined as BMI for age $85th repertoire consumed over
a 3-day period
percentile, and underweight was defined as <5th percentile, High-frequency Single foods eaten 4 to 5 or Modified FFQ*
as recommended.24 single food intake more times daily
Parents were interviewed about their child’s dietary *We used the 9 frequency categories from the original Harvard Food Frequency Questionnaire
habits and use of special diets. Parents also completed to ensure that all foods would have the same frequency options. We added a new category, N/A
don’t offer (not applicable, food item is never offered to the child), and we split the category
a demographic/medical questionnaire and a modified never/less than once a month into 2 separate categories, never– will not eat (child refuses
food frequency questionnaire (FFQ), and were instructed the food item) and less than one time per month. Fruits and vegetables that were combined
in a single item on the original Youth/Adolescent Food Frequency questionnaire (eg, ‘‘oranges
by a registered dietitian or a nutrition graduate student to and grapefruits,’’ ‘‘peas and lima beans’’) were listed separately.
complete a 3-day food record on 2 weekdays and 1 weekend †NDSR output lists all of the foods each child ate at the whole-food level during the 3-day pe-
riod. Similar foods were grouped together to prevent foods with similar taste, texture, and ap-
day. To capture the food that the child consumed at school, pearance (eg, French fries and tater tots) from being counted as 2 separate foods. To create
parents were given a second food record and a letter to groups, we first collapsed all foods according to their NDSR identifying names and then elim-
inated all ingredients. Foods not in the NDSR database, including generic foods and some brand
school personnel explaining the study and asking the child’s name foods, were included with their groups or identical foods to prevent counting identical
teacher to record all of the food the child ate at school. In foods twice. A total of approximately 900 different foods were consumed by the participants.

260 Bandini et al
August 2010 ORIGINAL ARTICLES

well as the percentage of foods the child would not eat relative repertoire, as well as the association of these aspects of food
to the number of foods offered. selectivity with age. We hypothesized that for typically devel-
oping children, food selectivity would decrease with age,
High-Frequency Single Food Intake whereas we did not expect to see an association with age
The highest-frequency categories on the FFQ were 4-5 times among children with ASDs. We tested this hypothesis based
per day and more than 6 times per day. We defined high- on the statistical significance of the interaction (ie, cross-
frequency consumption as single foods on the FFQ list product term) between age and group (ASD or typically
(or written in by the parent) that children ate more than developing) in linear regression models with food refusal
4-5 times per day. Beverages were not included in this and limited repertoire as outcome variables. We evaluated
measure, because it is not uncommon to consume a beverage the relationship between food refusal and limited repertoire
4 or more times a day. with nutrient inadequacy using Spearman correlation
coefficients as well as linear regression.
Limited Food Repertoire
Three-day food records were coded using the Nutrition Data Results
System for Research (NDSR; Nutrition Coordinating Center,
University of Minnesota, Minneapolis, Minnesota) in an A total of 53 children with ASDs and 58 typically developing
effort to determine each child’s food repertoire, that is, children completed the study (Table II). Mothers were the
how many unique foods (including beverages) each child primary respondents, with 4 fathers in each group. Few of
consumed over a 3-day period (Table I). Nutrient the characteristics differed between the 2 groups, except
inadequacy was based on an analysis of the 3-day food that children with ASDs were more likely to be on a special
record for nutrient composition using NDSR software. The diet, and typically developing children were more likely to
average level of each nutrient across the 3 recording days be an only child. Whether or not a child was an only child
was determined. To identify nutrient inadequacies in the was not related to any aspect of food selectivity, however.
child’s diet, we identified nutrients that were found to be
low in a nationally representative survey of the diets of US Food Refusal
children29 or that have been reported to be low in studies We observed food refusal in both typically developing chil-
of children with ASDs.4,30 These included calcium, dren and children with ASDs. On average, children with
vitamins D and E, iron, zinc, and fiber. In addition,
because intakes of vegetables and fruits often are below
recommendations for US children,29 we included vitamins
A and C in our analyses. We defined nutritional Table II. Demographic characteristics in children with
inadequacy relative to the estimated average requirement ASDs and typically developing children
(EAR) for the specific life stage and sex group; for calcium, Typically Children
fiber, and vitamin D, for which an EAR has not been developing with ASDs
children (n = 58) (n = 53) P*
defined, we used the adequate intake (AI).31 As a summary
Age, years, mean (SD) 6.7 (2.4) 6.6 (2.1) .75
measure of nutrient inadequacy, we determined the VABS score, mean (SD) N/A 71.1 (12.4)
number of nutrients (out of the 8 that we identified) for DAS general conceptual N/A 85.8 (22.1)†
which the average intake was below the EAR or AI for each ability score, mean (SD)
Male sex, % 78% 83% .47
child. Caucasian race, %z 76% 83% .35
Maternal education, 72% 74% .89
Statistical Analysis $ college degree, %
Paternal education, 67% 54% x .15
All analyses were performed using SAS version 9.1 (SAS $ college degree, %
Institute, Cary, North Carolina), with an alpha level of One or more parent 81% 81% .99
0.05. Comparisons between typically developing children with college degree, %
Child is an only child, % 26% 11% .05
and children with ASDs were made using t tests for continu- Child is on a special diet, n{ 0 11 <.001
ous variables and c2 or Fisher exact tests for categorical Overweight, BMI $ 85th 22% 26% .62
variables. We evaluated the following variables to determine percentile, %
Underweight, BMI < 5th 0% 2% .48
if they differed between children with ASDs and typically de- percentile, %
veloping children: age, sex, race (Caucasian, non-Caucasian),
*P values for differences in autism status groups
following a special diet, parental education (college degree †n = 47. There were 6 subjects for whom DAS general conceptual ability (GCA) scores were not
or not), siblings (0 vs > 1), and weight status (underweight, available. Two subjects took the DAS at a level where a GCA is not calculated, 1 subject refused
testing, and 3 subjects had problems with test administration such that their tests could not be
normal weight, or overweight). For those variables that scored.
differed by group (ASD or typically developing), we assessed zRacial/ethnic breakdown for non-Caucasian participants: black/African-American: 7% typi-
cally developing, 2% ASDs; Hispanic: 5% typically developing, 4% ASDs; Asian: 0% typically
the need to control for potential confounding variables by developing, 4% ASDs; more than 1 race/other: 12% typically developing, 8% ASDs.
determining whether they also were related to any aspect of xn = 52.
{Special diet refers to a gluten/casein-free, wheat-free, or lactose-free diet. Nine children
food selectivity. We used Pearson correlation coefficients were on a gluten-free/casein-free diet, 1 child was on a wheat-free diet, and 1 child was on
to assess the association between food refusal and limited a lactose-free diet.

Food Selectivity in Children with Autism Spectrum Disorders and Typically Developing Children 261
THE JOURNAL OF PEDIATRICS  www.jpeds.com Vol. 157, No. 2

Table III. Food selectivity in children with ASDs and typically developing children
n Typically developing children n Children with ASDs P*

Food refusal, mean (SD)
Number of FFQ items that child will not eat 58 21 (18) 53 45 (26) <.0001
Percentage of FFQ items offered that child will not eatz 58 18.9% (15.6%) 53 41.7% (21.2%) <.0001
Limited repertoire, mean (SD)x 56 22.5 (4.6) 48 19.0 (5.0) .0003
High-frequency single-food intake, n (%){ 58 1 (1.7%) 53 4 (7.6%) .19

*P value is for difference between children with ASDs and typically developing children.
†For food refusal, the minimum, median, and maximum number of foods refused was 0, 16, and 79 for typically developing children and was 1, 47, and 94 for children with ASDs. The minimum,
median, and maximum percentage of foods refused of those offered was 0%, 14%, and 62% for typically developing children and 1%, 41%, and 74% for children with ASDs.
zNumber of foods that the child will not eat of those offered is the percentage of foods that the child would not eat relative to the number of foods that were offered.
xFood repertoire is the number of foods eaten over a 3-day period.
{Number of children eating a food more than 4.5 times a day.

ASDs refused more foods and refused more foods as a per- child had an ASD or was typically developing (interaction
centage of those offered compared with typically developing with age, P = .12 for food refusal and P = .11 for limited
children (Table III). In addition to all items on the FFQ, we food repertoire).
evaluated vegetables separately; children with ASDs refused
more vegetables than did typically developing children, Nutrient Inadequacy and Food Selectivity
both in absolute amount (11  6 vs 6  5; P <.0001) and When looking at both groups together, of the 8 nutrients that
as a percentage of foods offered (63%  31% vs 33%  we considered to be at possible risk of inadequate intake, the
27%; P <.0001). median number of nutrients for which children were below
the EAR or AI was 3. In all but 1 child, fiber intake was inad-
High-Frequency Single Food Intake equate. Twelve children were inadequate in 5 or more nutri-
Contrary to our expectations, HFSFI was rarely seen in children ents. In addition to fiber, nutrient inadequacy was common
in either group. Parents of only 5 children (4 children with for vitamin D, vitamin E, and calcium. Inadequate intake of
ASDs and 1 typically developing child) reported consumption vitamin D and calcium was more frequent for children
at a sufficient level to meet our definition of high frequency with ASDs compared with typically developing children
(ie, consumption of a single food more than 4-5 times/day) (Table IV). Overall, children with ASDs had a greater
(Table III). Parents of 2 children, both with ASDs, reported number of nutrients with inadequate intake compared with
that their child consumed one food more than 6 times per day. typically developing children (P = .03). When children on
special diets were excluded, the results were essentially
Limited Food Repertoire unchanged (results not shown).
Three-day food records were completed for 94% of children
in the study; the parents of 7 children (5 with ASDs and 2 typ-
ically developing) did not record 3 days of intake and thus
were not included in analyses related to limited repertoire
or nutrient inadequacy. On average, parents of children Table IV. Frequency of nutrient inadequacy in children
with ASDs recorded that their child ate significantly fewer with ASDs and typically developing children
types of foods over the 3-day period compared with typically Typically developing Children with
developing children (Table III). children (n = 56)* ASDs (n = 48) P†
Number (%) not meeting EAR/AI for specific nutrients
Interrelationships among Measures of Food Vitamin A 0 (0.0%) 5 (10.4%) .02
Vitamin C 5 (8.9%) 10 (20.8%) .10
Selectivity Vitamin D 31 (55.4%) 38 (79.2%) .01
We observed a significant inverse correlation between food Vitamin E 41 (73.2%) 30 (62.5%) .29
refusal and food repertoire in the cohort (r = -0.49; P <.0001) Zinc 3 (5.4%) 5 (10.4%) .47
Calcium 24 (42.9%) 31 (64.6%) .03
and in each group (ASDs: r = -0.53, P <.0001; typically develop- Iron 0 (0.0%) 0 (0.0%) NA
ing: r = -0.34, P = .01). As expected, children with broader food Fiber 56 (100%) 47 (97.9%) .46
repertoires refused fewer foods as a percentage of those offered Number (%) of inadequate nutrients
0 0 (0%) 0 (0%)
than children with more restricted food repertoires. 1 3 (5.4%) 3 (6.3%)
The correlation (r) between food refusal and age was -0.12 2 20 (35.7%) 5 (10.4%)
(P = .21) overall, -0.24 (P = .08) in children with ASDs, and 3 19 (33.9%) 21 (43.8%)
4 10 (17.9%) 10 (20.8%)
-0.001 (P = .99) in typically developing children. The corre- 5 4 (7.1%) 5 (10.4%)
lation between limited repertoire and age was r = 0.03 6+ 0 (0.0%) 4 (8.3%)
(P = 0.77) overall, 0.19 (P = .19) in children with ASDs, .03z
and -0.13 (P = .36) in typically developing children. *Seven children (5 with ASDs, 2 typically developing) did not have complete 3-day food records
In addition, contrary to our hypothesis, the association be- and are not included in the table.
†P values are for differences between children with ASDs and typically developing children.
tween food selectivity and age did not depend on whether the zP value is from a Fisher exact test with 11 degrees of freedom.

262 Bandini et al
August 2010 ORIGINAL ARTICLES

To explore which aspects of food selectivity were associ- mated requirement of about 50% of the population, where
ated with nutrient inadequacy, we examined the association one was available. In the absence of an EAR, we used the
between food refusal and limited repertoire with nutrient in- AI. We saw associations between nutrient inadequacy and
adequacy. We found that food refusal was not significantly limited food repertoire, but not with food refusal. This dis-
associated with nutrient inadequacy in the cohort overall crepancy may be explained by the fact that food refusal was
(r = 0.13, P = .20; linear regression R2 = 0.02), nor was there determined from the FFQ and nutritional adequacy from
any evidence that the association between food refusal and the 3-day food record. The FFQ assesses usual intake over
nutrient inadequacy differed between the ASD and typical the previous 12 months; determination of limited food rep-
groups (P = .32 for interaction). In contrast, limited reper- ertoire and nutrient inadequacy were both based on the
toire was associated with greater nutrient inadequacy in same dietary assessment tool, a food record covering 3 days
the cohort overall (r = -0.33, P = .0006; linear regression of intake.
R2 = 0.13). Although children with ASDs and typical children Gluten-free/casein-free and lactose-free diets are followed
differed in food repertoire, there was no evidence suggesting by some children with ASDs. Because children who adhere to
that the association between limited food repertoire and these diets restrict all dairy products, their intake of calcium
nutrient adequacy depended on whether the child had an and vitamin D would be expected to be low. In this scenario,
ASD or was typically developing (P = .82 for interaction). such a child’s food intake would reflect parental restriction
Exclusion of children who were following special diets had rather than the child’s food selectivity. However, the finding
no significant effect on the observed associations (results that the association between food selectivity and nutrient in-
not shown). adequacy did not differ by autism status, even after exclusion
of children on special diets, suggests that our findings did not
reflect parental restriction.
Discussion We acknowledge several limitations in the present study.
Our methodology for determining food refusal and HFSFI
We found that children with ASDs displayed more food re- was based on a modified FFQ. Because a parent might not of-
fusal and exhibited a more limited food repertoire compared fer a food that he or she believes the child would refuse, we
with typically developing children, although food refusal was cannot determine whether not offering a food is influenced
seen in both groups of children. The commonly held belief by the child’s presumed or historical refusal of that food.
that dietary ‘‘pickiness’’ is outgrown with age was not sup- We observed a great deal of variability in the number of foods
ported by our cross-sectional findings. We found that among that were not offered to children, which was greater in chil-
typically developing children, levels of food refusal and food dren with ASDs (range of 0 to 90 foods not offered, compared
repertoire were similar across age, whereas among children with 3-46 foods not offered to typically developing children).
with ASDs, food refusal, but not food repertoire, was margin- Among the children with ASDs, being on a special diet was
ally lower in older children. The lack of statistical significance associated with more foods not being offered, but not with
for the test of the interaction term and the cross-sectional na- a higher percentage of foods refused of those offered. Thus,
ture of our study preclude any conclusions regarding an as- some parents of children with ASDs offered a limited number
sociation between food selectivity measures and age. of foods, which would preclude the child refusing those
Although anecdotal reports suggest that some children foods. In addition, we disaggregated all fruit and vegetable
with ASDs consume a certain food or foods nearly exclu- items in the FFQ, based on our understanding that fruits
sively, we did not observe this phenomenon in our sample. and vegetables are commonly refused and our desire to be
Parents of 4 of the 53 children with ASDs (7.5%) reported able to accurately estimate refusal of these foods; however,
that their child consumed a single food more than 4-5 times we kept other aggregated foods, including entrees (such as
per day. This suggests that HFSFI does not occur as com- ‘‘meatballs or meatloaf’’ or ‘‘roast beef or ham sandwich’’)
monly in children with ASDs as might be assumed based and grains (such as ‘‘English muffins or bagels’’), together,
on anecdotal reports. The operational definition of HSFSI to minimize participant burden. This could have introduced
chosen for this study might have masked some of the unique error in our measures of refusal.
eating habits of children with ASDs, however. For example, if Another limitation of our measurement of food repertoire
a child ate macaroni and cheese for breakfast, lunch, and din- was that the food record captured only 3 days of intake,
ner (3 times daily), this behavior would not have met our def- which may not adequately capture the variety of the typical
inition of HFSFI, although most would consider this eating diet. However, Falciglia et al32 compared a 3-day record
pattern unusual. We did not observe this in our 3-day food with a 15-day record in children and found that the 3-day re-
records, however. cord was accurate for assessing variety in children, although
The association between limited repertoire and nutrient they found more precise estimates when using 3 nonconsec-
inadequacy suggests that a very limited diet may put any child utive days. We also made coding decisions that may have af-
at risk for nutritional deficiencies.4,30 We found that children fected our estimate of variety. Participants reported eating
with a narrower food repertoire were more likely to have an approximately 900 different foods overall, so decisions about
inadequate intake of more nutrients. We chose to use the coding of food often required us to collapse several items into
EAR,31 which is the nutrient intake needed to meet the esti- a single group. We based our coding decisions on the
Food Selectivity in Children with Autism Spectrum Disorders and Typically Developing Children 263
THE JOURNAL OF PEDIATRICS  www.jpeds.com Vol. 157, No. 2

structure of the NDSR software. For example, all breads were 9. Carruth BR, Skinner JD. Revisiting the picky eater phenomenon:
considered to be a single food, meaning that white bread, neophobic behaviors of young children. J Am Coll Nutr 2000;19:
771-80.
whole wheat bread, and banana bread were all considered
10. Carruth BR, Ziegler PJ, Gordon A, Barr SI. Prevalence of picky eaters
bread. Although we instructed parents to report sandwiches among infants and toddlers and their caregivers’ decisions about offering
by their ingredients, some did not, requiring that the sand- a new food. J Am Diet Assoc 2004;104:s57-64.
wich be considered a single food. Any errors introduced in 11. Jacobi C, Agras WS, Bryson S, Hammer LD. Behavioral validation, pre-
this way would be expected to affect both groups similarly. cursors, and concomitants of picky eating in childhood. J Am Acad Child
Adolesc Psychiatry 2003;42:76-84.
Generalizability is an issue in all observational studies. Par-
12. Jacobi C, Schmitz G, Agras WS. Is picky eating an eating disorder? Int J
ents who describe their children as picky eaters are often con- Eat Disord 2008;41:626-34.
cerned that their child’s lack of variety prevents them from 13. Reau NR, Senturia YD, Lebailly SA, Christoffel KK. Infant and toddler
getting optimum nutrition, and thus could be disproportion- feeding patterns and problems: normative data and a new direction. J
ately attracted to a study like ours. To limit ascertainment Dev Behav Pediatr 1996;17:149-53.
14. Dubois L, Farmer A, Girard M, Peterson K, Tatone-Tokuda F. Prob-
bias, we developed recruitment materials that described our
lem eating behaviors related to social factors and body weight in pre-
study purpose as understanding the mealtime and activity pat- school children: a longitudinal study. Int J Behav Nutr Phys Act 2007;
terns of children with and without autism, without regard for 4:9.
whether parents had concerns in these areas. Nonetheless, it 15. Dubois L, Farmer AP, Girard M, Peterson K. Preschool children’s eating
is possible that parents of children with unusual eating patterns behaviours are related to dietary adequacy and body weight. Eur J Clin
Nutr 2007;61:846-55.
might have been more interested in participating in the study.
16. Galloway AT, Fiorito L, Lee Y, Birch LL. Parental pressure, dietary pat-
Our study operationalizes the definition of food selectivity terns, and weight status among girls who are "picky eaters. J Am Diet As-
and provides data from a moderate-sized sample that sup- soc 2005;105:541-8.
ports the notion that food selectivity is more common in 17. Wright CM, Parkinson KN, Shipton D, Drewett RF. How do toddler eat-
children with ASDs than in typically developing children. ing problems relate to their eating behavior, food preferences, and
growth? Pediatrics 2007;120:e1069-75.
Furthermore, we have shown that one aspect of food selectiv-
18. Dovey TM, Staples PA, Gibson EL, Halford JC. Food neophobia and
ity—limited food repertoire—is associated with inadequate "’picky/fussy" eating in children: a review. Appetite 2008;50:181-93.
nutrient intake. These findings, if confirmed, suggest that 19. Ahearn WH, Castine T, Nault K, Green G. An assessment of food accep-
limited repertoire may be of concern. Future research is tance in children with autism or pervasive developmental disorder not
needed to determine the antecedents to food selectivity and otherwise specified. J Autism Dev Disord 2001;31:505-11.
20. Rutter M, Le Couteur AL, Lord C. Autism Diagnostic Interview-Revised.
develop interventions that will increase food repertoire and
Los Angeles: Western Psychological Services; 2003.
decrease food refusal in children with ASDs. Longitudinal 21. Sparrow SS, Cicchetti DV, Balla DA. Vineland Adaptive Behavior Scales.
studies examining food selectivity are needed to explore 2nd ed. Circle Pines, MN: AGS Publishing; 2005.
whether food selectivity persists into adolescence and adult- 22. Elliott CD. Differential Ability Scales. San Antonio, TX: The Psycholog-
hood, the impact of prolonged food selectivity on nutritional ical Corporation; 1990.
23. Centers for Disease Control and Prevention. CDC clinical growth charts:
status, and whether there are differences in persistence of this
United States. Available from: http://www.cdc.gov/growthcharts/. Ac-
phenomenon between children with ASDs and typically de- cessed July 24, 2009.
veloping children. n 24. Centers for Disease Control and Prevention. Defining childhood over-
weight and obesity. Available from: http://www.cdc.gov/obesity/
Submitted for publication Aug 21, 2009; last revision received Dec 1, 2009; childhood/defining.html. Accessed July 24, 2009.
accepted Feb 8, 2010. 25. Field AE, Camargo CA Jr., Taylor CB, Berkey CS, Frazier AL,
Gillman MW, et al. Overweight, weight concerns, and bulimic behav-
References iors among girls and boys. J Am Acad Child Adolesc Psychiatry 1999;
38:754-60.
1. Raiten DJ, Massaro T. Perspectives on the nutritional ecology of autistic 26. Willet W. Food-frequency methods. In: Nutritional Epidemiology. 2nd
children. J Autism Dev Disord 1986;16:133-43. ed. New York: Oxford University Press; 1998. p. 74-100.
2. Schreck KA, Williams K, Smith AF. A comparison of eating behaviors 27. Rockett HR, Wolf AM, Colditz GA. Development and reproducibility of
between children with and without autism. J Autism Dev Disord 2004; a food frequency questionnaire to assess diets of older children and ad-
34:433-8. olescents. J Am Diet Assoc 1995;95:336-40.
3. Cermak SA, Curtin C, Bandini LG. Food selectivity and sensory sensitiv- 28. Rockett HR, Breitenbach M, Frazier AL, Witschi J, Wolf AM, Field AE,
ity in children with autism spectrum disorders. J Am Diet Assoc, et al. Validation of a youth/adolescent food frequency questionnaire.
2010;110:238-46. Prev Med 1997;26:808-16.
4. Cornish E. A balanced approach towards healthy eating in autism. J 29. Briefel RR, Johnson CL. Secular trends in dietary intake in the United
Hum Nutr Diet 1998;11:501-9. States. Annu Rev Nutr 2004;24:401-31.
5. Dominick KC, Davis NO, Lainhart J, Tager-Flusberg H, Folstein S. Atyp- 30. Herndon AC, DiGuiseppi C, Johnson SL, Leiferman J, Reynolds A. Does
ical behaviors in children with autism and children with a history of lan- nutritional intake differ between children with autism spectrum disor-
guage impairment. Res Dev Disabil 2007;28:145-62. ders and children with typical development? J Autism Dev Disord
6. Whiteley P, Rodgers J, Shattock P. Feeding patterns in autism. Autism 2009;39:212-22.
2000;4:207-11. 31. Institute of Medicine. DRI, Dietary Reference Intakes: The Essential
7. Williams PG, Dalrymple N, Neal J. Eating habits of children with autism. Guide to Nutrient Requirements. Washington, DC: National Academies
Pediatr Nurs 2000;26:259-64. Press; 2006.
8. Carruth BR, Skinner J, Houck K, Moran J 3rd, Coletta F, Ott D. The phe- 32. Falciglia GA, Horner SL, Liang J, Couch SC, Levin LS. Assessing dietary
nomenon of "picky eater": a behavioral marker in eating patterns of tod- variety in children: development and validation of a predictive equation.
dlers. J Am Coll Nutr 1998;17:180-6. J Am Diet Assoc 2009;109:641-7.

264 Bandini et al

Você também pode gostar