RESEARCH ARTICLE

Is Food Refusal in Autistic Children Related to TAS2R38 Genotype?

Maria Pia Riccio , Chiara Franco, Rossella Negri, Roberta Ida Ferrentino, Roberta Maresca,
Elisa D’alterio, Luigi Greco, and Carmela Bravaccio

Several studies suggest that atypical eating behaviors, in particular food selectivity, are more frequent in children
with autism spectrum disorder (ASD). A link between bitter taste perception, namely PROP/PTC sensitivity and food
preferences is known in healthy children. The aim of this study is to investigate whether genetic variants of the
TAS2R38 taste receptor responsible for different bitter sensitivity could affect foods preferences and consequently
food refusal in ASD children. We recruited 43 children with ASD and 41 with normotypic development (TD) with or
without food selectivity, aged between 2 and 11 years. Children were characterized for bitter sensitivity by means of
PROP strips and FACS analysis and genotyped for TAS2R38 polymorphisms. Food selectivity was assessed by a vali-
dated food preference questionnaire filled by parents. A statistically significant correlation between PROP sensitivity
and food refusal was observed. Furthermore, a prevalence of the PAV-sensitive haplotype compared to the AVI-
insensitive one was seen in ASD children with food selectivity. In agreement with the initial hypothesis the results
show that food refusal in ASD children is mediated by bitter taste sensitivity thus suggesting that the bitter sensitivity
test may be used as a device to orientate tailored food proposals for the practical management of food selectivity in
ASD. Autism Res 2017, 0: 000–000. V C 2017 International Society for Autism Research, Wiley Periodicals, Inc.

Lay Summary: A variation of the gene TAS2R38, associated with bitter taste sensitivity, can cause a different percep-
tion of some foods. In particular, some children are hypersensitive to bitterness and show a more restricted repertoire
of accepted foods. We evaluate bitter sensitivity in ASD children with or without food selectivity, through a simple
bitter taste test with edible strips. The results show that food refusal in ASD children can be mediated by bitter taste
sensitivity thus suggesting that the bitter sensitivity test may be used as a device to orientate tailored food proposals
for the practical management of food selectivity in ASD.

Keywords: ASD children; food selectivity; TAS2R38 genotype; PROP phenotype

Introduction to “neophobia” (the refuse of new unknown food by

The prevalence of Autistic Spectrum Disorders (ASD)
has considerably increased in the last decades and now-
adays is estimated around 100–250 cases/10,000 popula-
tion [Isaksen, 2013; Autism and Developmental
Disabilities Monitoring Network Surveillance Year 2008
Principal Investigators; Centers for Disease Control and
Prevention, 2012], with a male to female ratio of 4:1
[Duchan, 2012]. Psychiatric disorder, behavioral prob-
lems, and medical conditions can be associated to ASD.
Several studies report that atypical eating behaviors
occur frequently in children with ASD being food selec-
tivity the most frequent problem [Marı-Bauset, 2014].
Food selectivity has indeed been described in approxi-
mately a quarter of healthy infants up to 18 months or
less, but this is considered a transient problem related
the age of weaning). In children with mental disabil-
ities, the rate of food selectivity lasting for over 24
months can affect as high as 70–80% of ASD children
[Marı-Bauset, 2014].
Several hypotheses were proposed to explain food
selectivity in ASD: factors influencing food refusal are
mainly related to the characteristics of food, as texture,
appearance, taste, smell, temperature [Marı-Bauset,
2014]. Many researchers suggest a link between sensory
sensitivity (sensory aversions, oral defensiveness, tactile
defensiveness) and eating problems [Cermak, 2010;
Lane, 2014; Williams, 2000]. However much remains to
be known to explain why food selectivity is so preva-
lent among individuals with ASD. Variations in taste
perception influence food preference [Cooke, 2007]. In
particular, bitter taste triggers refuse for some foods in

From the Department of Medical and Translational Sciences, Section of Child Psychiatry, University of Federico II Naples, Italy (M.P.R., C.B.);
Department of Medical and Translational Sciences, Section of Pediatrics, University of Naples Federico II, Naples, Italy (C.F., R.N., L.G.); Department
of Mental and Physical Health and Preventive Medicine, Section of Child and Adolescent Psychiatry, University of Naples “L. Vanvitelli”, Naples,
Italy (R.I.F., R.M., E.D.)
Maria Pia Riccio and Chiara Franco contributed equally to the work.
Received August 21, 2017; accepted for publication December 06, 2017
Address for correspondence and reprints: Rossella Negri, Department of Medical and Translational Sciences, Section of Pediatrics, University of
Naples Federico II, via Pansini, 5- Naples, Italy. E-mail: rosnegri@unina.it
Published online 00 Month 2017 in Wiley Online Library (wileyonlinelibrary.com)
DOI: 10.1002/aur.1912
C 2017 International Society for Autism Research, Wiley Periodicals, Inc.
V

INSAR Autism Research 00: 00–00, 2017 1

children and causes dietary restrictions. There are large
individual differences in the perception of bitterness
among people. The best-studied gene responsible for
variation in bitterness sensitivity is the taste receptor
gene TAS2R38, a bitter receptor for the thiourea com-
pounds phenylthiocarbamide (PTC) and 6-n-
propylthiouracil (PROP). Based on sensitivity to thio-
urea, human population can be phenotypically classi-
fied into three categories: insensitive/non taster, sensitive/
taster, and hypersensitive/supertaster [Bufe, 2005; Men-
nella, 2011]. Three single nucleotide polymorphisms at
TAS2R38 locus result in three aminoacid substitutions
P49A, A262V and V296I which give rise to two com-
mon haplotypes, PAV (the PROP-taster variant) and AVI
(the PROP-non taster variant). PROP-sensitive individu-
als are carriers of one or two dominant alleles (PAV/
PAV or PAV/AVI), whereas insensitive individuals are
homozygous for the recessive allele (AVI/AVI). PTC/
PROP tasters are more sensitive to a variety of bitter
substances and show greater dislike for certain bitter
vegetables (i.e., broccoli, brussel sprouts, cabbage, kale,
asparagus, spinach, grapefruit, soy products), fruits, and
strong-tasting foods (sweets, capsaicin, alcohol and liq-
uid fats) [Bell, 2006; Dinehart, 2006; Duffy, 2000]. Fur-
thermore, PROP supertasters exhibit a great density of
taste papillae and are more sensitivity to other tastes.
PROP-sensitivity affects food preferences on the base of
taste perception and this can cause food selectivity
[Tepper, 2009]. Liking and consumption of vegetables
among children correlate with bitterness of these sub-
stances [Drewnowski, 2000] and with TAS2R38 geno-
type [Tepper, 2009]. The innate, genetic, sensitivity is
more evident in children than adults, where it is also
modified by the taste experience and culture [Whissell-
Buechy, 1990]. This is a possible explanation for lower
consumption of food (and in particular of bitter vegeta-
bles) in children than their relatives in general popula-
tion [Negri, 2012].
ASD children refuse more frequently fruits and vege-
tables compared to typically developing (TD) children
[Hubbard, 2014] but it is unknown if this preference
depends on the taste identification abilities or other
factors.
We hypothesized that food refusal in ASD could be
influenced by the TAS2R38 genotype being more severe
in individuals genetically predisposed. Subjects that
carry one or two copies of the taster haplotype and are
phenotypically tasters or supertasters show a reduced
intake of vegetables and fruits and accept a smaller vari-
ety of foods. The study, therefore, evaluate bitter taste
reactivity, TAS2R38 genotype and genotype-phenotype
relationship in individuals with ASD, to investigate the
association with food selectivity.
True differences exist in taste and olfactory identifica-
tion in ASD, but taste perception is one of the least
2 Riccio et al./Food refusal and autism spectrum disorders
studied senses. If the ASD condition is associated with
atypicalities on this sensory level it would need to be
investigated in more detail [Tavassoli, 2009].
Tastes and odors elicited facial reactions that are a
powerful source of information for the study of taste
and odor-elicited effects in humans [Griemel, 2006;
Zeinstra, 2009]. The bitter taste elicits brown lowering
and lip raising and mouth opening, that can be inter-
preted as a sign of negative effect associated to the
emotion of disgust [Rozin 1994, 1999; Steiner, 1977]. A
way to analyze facial expressions is using the Facial
Action Coding System (FACS) [Ekman & Friesen, 1978].
Observation of facial expressions can be a valuable tool
to collect informations about likes and dislikes.
The aim of the study is to verify if the eating behav-
ior of ASD children was related to bitter sensitivity.
For this purpose, we aim to investigate the relation-
ship between food selectivity and bitter taste sensitivity
by the characterization of PROP phenotypes and
TAS2R38 genotypes in a cohort of ASD children. We
hypothesize that bitter sensitivity linked to TAS2R38
genetic variants could be a predisposing factor to eating
problems in ASD, so children more sensitive to bitter
taste (with one or two copies of PAV functional allele)
are expected to exhibit a more limited food repertoire
compared to hyposensitive ones (AVI homozygous).
Methods
Subjects
Study participants included 43 ASD children (diagnosis
according to DSM-5 criteria—APA, 2013), aged between
2 and 11 years (mean age 6.28 6 2.3 years), both male
and female (33M; 10F), with and without food selectiv-
ity. Children were referred to Department of Pediatrics-
Unit of Child Psychiatry, University Federico II of Naples
and Child Neuropsychiatry Division of University of
Naples Luigi Vanvitelli (Italy) between September 2015
and September 2016. Exclusion criteria were the presence
of genetic diseases (e.g., fragile X, tuberous sclerosis, etc)
that can occur with an autistic phenotype, and disorders
that can cause deficits in chewing or swallowing, that
can lead the subject to a particular diet (e.g., semi-solid
or liquid). Fourty-one typically developing children (TD)
(any neuropsychiatric conditions diagnosed and any GI
conditions), matched for sex and age (mean 5 7.2 years),
with and without food selectivity, were recruited as con-
trol group. Ethical approval for the phenotypic and
TAS2R38 genotypic analysis in healthy children and in
children affected by a variety of abnormalities was
obtained by the Ethics Committee of the University of
Naples Federico II. Written informed consent was
obtained by parents/legal tutors of all participants,
included agreement to video recording.
INSAR
Table 1. Descriptive Clinical Variables of ASD Children
(n 5 43). Data Are Presented as n (%) or Mean 6 SD
Age, years
Male
Developmental quotient (SQ)
Intellectual quotient (IQ)
VABS score: mean (SD)
CARS-T score: mean (SD)
ADOS 2 severity rating (%)
low
mild
severe
Language
verbala
no verbalb
ASD severity (DSM 5)
level 1
level 2
level 3
a
Verbal 5 children who may use many single words or with phrase
speech and/or verbally fluent, according to ADOS-2 classification.
b
No verbal 5 children who do not use any words or produce some
single words (less than five words), according to ADOS-2 classification.
In order to validate the diagnosis of ASD, cases were
assessed through the Autism Diagnostic Observation
Schedule-2 (ADOS-2) [De Bildt, 2011; Gotham, Risi,
Pickles, & Lord, 2007, 2009] performed by a licensed
clinician. Moreover, Griffiths Mental Development
Scales (GMDS-ER) [Griffiths, 2006] were administered to
22/43 cases to determine the development level; 13/43
children were evaluated through Leiter International
Performance Test-Revised (Leiter-R) [Leiter, 1979]. In 8/
43 cases, general quotient was not available due to low
compliance to the test. Furthermore, Autism Diagnostic
Interview, Revised version (ADI-R) [Lord, 1994], Child-
hood Autism Rating Scales (CARS) [Schopler, 1988], and
Vineland Adaptive Behavior Scales (VABS) [Sparrow,
2003] were administered to parents of all patients. In
order to evaluate ASD severity, we identified three
severity levels of ASD, as specified by DSM-5 criteria.
“Level 1” refers to clinical conditions in which a subject
needs a support to have an adequate adaptation; “Level
2” refers to clinical conditions in which a subject needs
a significant support to have an adequate adaptation;
“Level 3” refers to clinical conditions in which a subject
needs a very significant support to have an adequate
adaptation. Severity levels were distributed as follows:
“Level 3” in 24% of cases; “Level 2” in 60% of cases;
“Level 1” in 16% of cases. Clinical variables of ASD
sample are reported in Table 1.
Assessment of Eating Behavior
To explore food choices and eating habits of ASD chil-
dren we used a modified version of the Feeding Habits
Questionnaire and Food Preferences Inventory (FPI) of the
Hershey Medical Center Feed Program HP19.
INSAR Riccio et al./Food refusal and autism spectrum disorders
The Feeding Habits Questionnaire is a specific screen-
ing form which explores the feeding behavior in chil-
6.28 6 2.3 dren to assess if there are medical reasons for feeding
33 (77.0) difficulties (oral-pharyngeal motility troubles as chew-
28.75 6 7.45 ing or swallowing difficulties) or if the eating problems
74.92 6 19.3
56.82 6 14.4 are related to characteristics of food (texture, tempera-
35.98 6 4.95 ture, taste, smell, color, form, packaging, solid, or liquid
state).
8 (18, 6) The Food Preferences Inventory based on the FPI
25 (58, 14)
described by Schreck et al. [2006] was modified to be
10 (23, 26)
used as a parents report checklist of preferred or typi-
25 (58) cally accepted food items. No information is collected
18 (42) about the frequency of each food item consumption.
The FPI form was translated in Italian language and
7 (16)
26 (60) adapted to Mediterranean diet. It was administered
10 (24) either to ASD and TD children parents. The checklist
includes 86 items from 11 main groups (vegetables,
fruits, cereals, meat, fish, sausages, dairy, cakes, eggs,
cakes, beverages). The FPI form investigates the food
preferences (liking/disliking) of the subject and his first-
degree relatives (parents and unaffected siblings), taking
into account if not consumed foods were never been
offered to the subject and if they were usually eaten
among first-degree relatives. Therefore, it allows to com-
pare the eating habits of ASD subjects with those of
unaffected family members and to objectively evaluate
presence/absence of food selectivity. Scores for the food
groups are obtained by summing the food items
accepted in each group. A total score corresponding to
the arithmetical mean minus 1SD of the accepted foods
by the TD children was adopted as cut-off to categorize
children with or without food refusal in ASD as in the
TD group. Our work on TD children [Negri et al., 2012]
supported to adopt the threshold reported.
Identification of PROP Status
Bitter tasting was done in an ASD dedicated area by a
researcher and a psychiatrist with the help of the
parents. Children, who were fasting 1 hr before the
test, were presented with two edible pullulan-HPMC
strips containing the bitter tastant PROP at increasing
concentrations of 0.28 and 0.56 mM, prepared accord-
ing to the Smutzer methods [2008]. In the case of child-
ren’s reluctance to taste strips, we used 5 ml samples of
0.28 and 0.56 mM PROP solutions. Between the presen-
tation of one stimulus and the next, children rinses
their mouth with spring water. Due to age and commu-
nicative difficulties of the subjects in order to properly
identify the PROP phenotype, the evaluation of sensory
responses was performed via analysis of facial responses.
Facial reactions elicited by taste were noted and
recorded by a video camera to be analyzed by the Facial
Action Coding System (FACS).
3
 The FACS [Ekman & Friesen, 1978; Ekman & Friesen,
1982; Ekman 2002] is a process of analysis and classifi-
cation of facial movements in reference to their ana-
tomical components. Since each of these movements is
the result of individual or synergistic facial muscles, the
FACS takes into account the way in which every facial
muscle acts to visibly change the configuration of the
face itself. In order to provide a specific index for each
type of movement and expression, the FACS considers
44 basic units called by Ekman and Friesen “Action
Unit” (AU) some of which are typically emotion-
specific, while others are not associated with any emo-
tion in particular. A total of 58 different movements are
classified, but FACS is a purely descriptive system which
does not assign an interpretive meaning to facial
expressions.
Videos were coded by a certified FACS coder, blind to
the subject genotype, in two different sessions. An over-
all inter-scoring agreement of 80% was achieved. In
order to code facial activities the videotapes were
observed in slow-motion or, if necessary, frame by frame.
The facial reactions were scored coding the apex, the
moment of most intense facial expression. The assign-
ment to either taster or non taster phenotype was based
on the specific AUs coded and on the number of AUs
simultaneously involved in each facial response. Namely
the taster phenotype was deducted by the involvement
of AU9 (nose wrinkler) or AU10 (upper lip raiser), typi-
cally implicated in the expression of the disgust, alone
or in associations with AU16 (lower lip depressor) and
AU15 (lip corner depressor) or AU26 (jaw drop), AU17
(chin raiser), AU25 (mouth stretch). The non taster phe-
notype was inferred from neutral facial expression or
involvement of AUs not associated to aversive reactions.
TAS2R38 Genotyping
Genetic analyses were conducted on genomic DNA
obtained from blood samples.
Allelic Discrimination Assay was performed with Applied
Biosystems 7900HT fast thermal cycler, using allele-specific
probes (TAS2R38rs713598: C__8876467_10; T A S 2 R 3 8 r s
1 7 2 6 8 6 6: C _ _ 9 5 0 6 8 2 7 _ 1 0; TAS2R38rs10246939:
C__9506826_10, CAVIrs2274333: C__1739329_10) and
primers from Applied Biosystems (Life Technologies Cor-
poration, Carlsbad,CA), according to standard Taqman
SNP Genotyping assay protocol.
Statistical Analysis
Data analyses was performed using the Statistical Pack-
age for Social Sciences (SPSS 19.0 for Windows, Chi-
cago, IL). Chi-square analysis was used to compare the
distribution of PROP phenotypes, TAS2R38 genotypes,
selective phenotype between ASD and TD children.
T-test was used to compare the FPI mean scores
4 Riccio et al./Food refusal and autism spectrum disorders
between ASD and TD groups and between children and
their families. The relationship of the children food
refusal to autism symptom severity was assessed by chi-
square test with the severity levels on a three-rank score
as predictor (independent) variable and selectivity phe-
notypes as dependent variable.
Chemicals
Pullulan manufactured by Hayashibara Co. Ltd, Kitaku,
Okayama, Japan was courtesy of Nagase & Co. Ltd
(Europa); HPMC was kindly provided by Eighenmann &
Veronelli (Milano, Italy) (http://www.eigver.it). Food
coloring was obtained from a local candy store. 6n-
Propylthiouracil (PROP) was obtained from SIGMA-
Aldrich Co.
Results
PROP Sensitivity in ASD Children
As the method used to screen ASD children for PROP
status did not allow to distinguish tasters from super-
tasters, we grouped children in only two phenotypes:
tasters versus non tasters. For 26 of 43 children tested for
bitter sensitivity, the PROP phenotype was assessed by
means of clinical observation and FACS analysis, the
phenotype estimated by the two methods overlapped
in 21/26 (81%) cases. Where the videos were not suit-
able to the FACS analysis the attribution to either phe-
notypic category was based on the clinical observation
of aversive or neutral reactions elicited by bitter taste.
Children showing classic rejection signs, such as gri-
macing or frowning were reported as tasters.
The distribution of the PROP phenotypes was: 81.4%
(35/43) tasters, 18.6% (8/43) non tasters, not statisti-
cally different from that of the healthy pediatric popu-
lation in the same geographic area [Negri et al., 2012].
TAS2R38 Genotypes
DNA samples suitable for genotyping were collected
from 36 ASD children. The prevalence of the TAS2R38
genotypes was: AVI/AVI 12 (33.3%), PAV/AVI 13
(36.1%), PAV/PAV 11 (30.6%). Chi-squared test demon-
strates a statistically different distribution compared to
the general population [Negri et al., 2012].
Food Refusal in ASD Vs TD Children
We explored the eating attitudes of 43 ASD and 41 TD
children aged 2 to 11 years by means of a modified ver-
sion of the Foods Preferences Inventory (FPI) described
by Schreck et al [2006]. The FPI revised form is a parent’s
report checklist of usually accepted or refused foods (lik-
ing/disliking) including 86 items representative of the
mediterranean diet. Food refusal is indirectly derived
from the FPI scores resulting from the sum of accepted
INSAR
Table 2. Mean 6 SD FPI Scores of Accepted Foods in ASD vs Table 4. Correlation between Food Selectivity and TAS2R38
TD Children and in Their Families Genotypes in ASD Children
ASD children TD children ASD families TD families TAS2R38 genotype Selective Non selective Total

Total foods 42.60 6 14.8** 57.29 6 10.8 81.14 6 8.1 79.54 6 8.9 PAV/PAV 9 (37.5)* 2 (16.6) 11
Greens 6.95 6 3.9** 9.73 6 3.3 15.92 6 2.4 15.78 6 2.4 PAV/AVI 10 (41.6) 3 (25) 13
Bitter greens 1.07 61.2* 1.66 6 1.5 3.95 6 1.5 3.88 6 1.6 AVI/AVI 5 (20.8) 7 (58.3) 12
Fruits 6.91 6 5.2** 10.41 6 5.1 16.26 6 3.2 16.46 6 2.1 total 24 (100) 12 (100) 36

FPI, foods preferences inventory; ASD, autism spectrum disorders; ASD, autism spectrum disorders.
TD, typical development. *n (%) v2:5,127; P:0.07.
*P  0.05; ** P  0.01, between ASD and TD children.

Table 5. Correlation between PROP Sensitivity and Food

Table 3. Prevalence of Food Refusal in ASD and TD Children Selectivity in ASD Children
Children with Selective Non selective Total PROP phenotype Selective Non selective
TD 8 (19.5)* 33 (80.5) 41(100) Taster 27 (90)* 8 (38.5)
ASD 30 (69.8) 13 (30.2) 43 (100) Non taster 3 (10) 5 (61.5)
Total 38 (45.2) 46 (54.8) 84 (100) total 30 (100) 12 (100)

ASD, autism spectrum disorders; TD, typical development. ASD, autism spectrum disorders.
*n (%) v2:21,397; P:0.00. *n (%) v2:4,852; P:0.028.

foods. By comparing the mean FPI scores in the ASD and

TD groups a statistically significant difference was
observed, showing a more restricted repertoire of
accepted foods in ASD children. In particular the chil-
dren with ASD eat on average 30% less vegetables, bitter
greens, and fruits, compared to TD ones (Table 2).
In order to explore if the eating behavior of the ASD
children was influenced by family habits, for example by
a more restricted selection of food offered, we compared
the average FPI score of the ASD families with that of the
families of TD children and the average FPI score of the
children with that of their families (Table 2). Whereas no
differences were observed between the families eating
behavior, the foods accepted by ASD children were on
average less than 50% of those eaten by their families,
instead of the 70% consumed by TD children.
As it is expected that TD children may also exhibit
food selectivity to some extent, in order to compare the
prevalence of the food refusal in the two groups, we
adopted the mean-1SD FPI score of TD group (46.5) as
cut-off value to operationally define the selective pheno-
type. According to the FPI score, 30 of 43 (69.7%) ASD
children were classified as selective, whereas 13/43 (30.3%)
were rated as non selective. In TD group 8/41 (19.5%) were
selective, and 33/41 (80.5%) were non selective (Table 3).
To evaluate if food refusal in ASD children was influ-
enced by the severity of autism symptoms we explore
the eating behavior across the severity levels, as speci-
fied by DSM-5 criteria. No relationship was observed
between food selectivity and the seriousness of clinical
condition, as the prevalence of the selective phenotype
was similar in the three severity levels.
In ASD children, the reason for food refusal was fur-
ther investigated using the Feeding Habits Questionnaire
(FHQ). Where the cause of the rejection was referred by
parents (24/30) the refusal was linked to food’s charac-
teristics: in the largest proportion of individuals (17/24,
70.8%) it was based on two or more features (texture,
taste, color, smell, and packaging). Texture and taste
were the most frequent qualities linked to rejection,
occurring in 13/24 and 14/24 cases, respectively.
Food Selectivity and Bitter Taste Sensitivity
In order to investigate if food refusal in ASD was related
to genetic variants of TAS2R38 bitter receptor we strati-
fied the food selectivity phenotypes by TAS2R38 geno-
types and compared the percentages of subjects
carrying the PAV (taster) haplotype in the two groups.
Among the selective children 19/24 (79.1%) were het-
ero- or homozygous for the PAV haplotype, whereas
within non selective children the prevalence of the PAV
haplotype was only 41.7% (Table 4). The higher fre-
quency of PAV haplotype in selective children fits well
with the lowest FPI scores recorded in these subjects.
The value is at the limits of statistical significance, due
to the sample size.
As regards the TAS2R38 associated taster phenotypes,
90% (27/30) of selective children were PROP sensitive,
being the prevalence of the insensitive phenotype four
times less than in the nonselective children (10% vs.
38.5%) (Table 5).
Discussion
Food selectivity is a clinic condition that has high
impact on the management and adaptation of

INSAR Riccio et al./Food refusal and autism spectrum disorders 5

individuals with ASD. Although also found in TD sub-
jects, the incidence of this condition is increased in
ASD.
In our study, the percentage of ASD subjects with
food selectivity is statistically increased compared to
the control group. Children with autism have signifi-
cantly more feeding problems and eat a narrower range
of food than TD children [Schreck, Williams, & Smith,
2004]. Schmitt et al. [2008] highlighted that boys with
ASD ate a considerably smaller variety of foods than
controls and more often they made food choices based
on texture. Bandini [2010] defines food selectivity more
accurately and report that ASD children exhibit more
food refusal and limited food repertoire that TD sub-
jects. Our results agree with these data and confirm
that ASD children have less varied nutrition than TD,
showing a widely refusal of foods.
Food selectivity is not associated with severity level
of disease. In fact, ASD subjects present in prevalence a
“Level 2” of impairment according to DSM-5 (in refer-
ence to developing, adaptive and social-communicative
skills), but the distribution of sample for severity level
appears homogeneous regardless of the presence or
absence of food selectivity. This result confirms data
reported by Schrek [2006], who referred that children’s
eating behavior is not related to the severity of diagnos-
tic symptoms of autism. Postorino [2015] underlines
that there are no differences in adaptive skills between
ASD subjects with or without food selectivity, but
parental stress is increased in case of food selectivity
and so their perception of severity level.
If we evaluate eating habits in the families of the
enrolled subjects, we can highlight that the number of
accepted foods is higher in adults than children both in
ASD and in TD subjects, but this gap is considerably
higher among ASD children and their families, than in
TD ones. Moreover, no significant differences were
found in the number of accepted food among families
of ASD and TD children. Therefore, we can say that
children’s food intake is not due to eating habits of the
family, differently from Schrek [2006] who showed that
family food preferences were the only significant pre-
dictor of the child’s preference. Furthermore, although
usually children consume less variety of foods than
adults [Negri, 2012, Tepper 2009], in our ASD sample,
the difference in food consumption between children
and their families is almost double than in controls.
Therefore, as food selectivity in our ASD sample does
not seem attributable to specific behavioral or environ-
mental factors, the work focused on the evaluation of
taste sensory profile in ASD. Bennetto [2007] identified
impairments in taste and olfactory identification in
autism, but there are no data about the relationship of
chemosensory dysfunction to other characteristics of
autism. In addition, Tavassoli [2012] underlies taste
6 Riccio et al./Food refusal and autism spectrum disorders
identification difficulties in ASD subjects, but no work
has yet correlated taste sensitivity with eating behaviors
in ASD children. It is known that increased sensitivity
to bitterness perception in childhood is associated to
the neophobia phenomenon; due to neophobia, chil-
dren tends to reject new foods and that consequently
reduces food variety [Cooke, Haworth, & Wardle,
2007]. However, in TD individuals neophobia is a tem-
porary eating behavior that appears between 2 and 5
years of age and tends to decline with age, instead in
ASD children food selectivity persist throughout child-
hood and the food repertoire remains narrowed [Ban-
dini et al., 2017].
This refusal appears mediated by taste or texture only
in about one-half of the subjects, while in the remain-
ing it is due to other factors (in particular the appear-
ance—including color -). Hubbard [2014] evidenced
that in a large percentage of ASD children food refusal
is based on texture.
Studies on taste perception in ASD are very limited
and to date there are no works assessing these aspects
in ASD subjects. One of the reasons is that ASD individ-
uals are often unable to express verbally the perception
of pleasure or disgust. Thus studies on taste perception
[Bennetto, Kuschner, & Hyman, 2007; Tavassoli et al.,
2012] were largely conducted on high-functioning and
adolescent/young adulthood ASD subjects. There are no
works on larger cohort with different ages and clinical
characteristics. This study is unique in its kind, since it
is carried out on subjects in early age (2–11 years). The
age frame is essential for a proper study of taste percep-
tion and in particular of the bitterness sensitivity
because it varies during life and tends to decrease in
adulthood [Mennella, Pepino, Duke, & Reed, 2010;
Negri et al., 2015]. This can explain the higher inci-
dence of food selectivity in children compared to adults
and also implies that advanced age is a considerable
selection bias in studies on taste perception in ASD. To
overcome the possible lack of compliance of the
enrolled subjects to express verbally bitter perception,
perceptive response was assessed by FACS analysis of
face expression [Ekmanand Friesen, 1978]. The FACS is
an objective and universal method of evaluating facial
expressions, suitable for subjects of any age (since birth)
to detect changes in response to stimuli and to decode
and associate them to specific emotional states. Bitter
taste, for the ancestral role played in the discrimination
of harmful substances (poisons, toxins), induces an
innately reaction of disgust, characterized by changes of
face expression, due to the activation of specific groups
of muscles, namely the Action Units. It is known that,
even in ASD individuals, FACS analysis allows to high-
light an analogous activation of Action Units of TD
subjects. Given the above, our results show that ASD
subjects with food selectivity are most frequently PROP-
INSAR
taster, showing a bitter sensitivity that causes a reaction
of disgust and rejection. This does not occur in ASD
individuals without food selectivity, in which the
PROP-taster phenotype is less frequent.
Regarding the TAS2R38 polymorphisms, ASD subjects
carrying the PAV taster-haplotype consume less foods
than ASD children with AVI/AVI nontaster genotype;
furthermore, there was a prevalence of taster genotype in
ASD with food selectivity compared to nonselective ASD.
Although with some limitations due to the small
sample size, these data lay in favor of a correlation
between food selectivity and TAS2R38 genetics in ASD
children. The relationship between TAS2R38 bitter sen-
sitivity and feeding problems has been very recently
shown in healthy preschool kids, where variants in
TAS2R38 gene has been associated to picky eating
behavior [Chong Cole et al., 2017].
Further studies on larger cohorts and adult patients
are needed to strength and possibly extend these find-
ings to whole ASD population.
Conclusions
Our study for the first time shows that food selectivity
in ASD could depend on the activation of specific pat-
terns of taste sensitivity. In particular, food selectivity can
be at least partially explained by bitter taste sensitivity.
However, a number of causal factors contribute to eat-
ing behavior. Eating behavior is by itself a complex
phenotype, influenced by multiple components, indi-
viduals as social, genetic as environmental. In ASD, in
particular, the severity and the persistence of food
refusal due to taste sensitivity may be affected by neu-
rological factors. Neuropsychological features typical of
ASD (such as lack of synthesis and central coherence,
executive function deficiency, behavioral rigidity), can
make easier for an affected person to persist in a behav-
ior learned from sensory experience, than modify it.
The practical significance of this study could concern
the management of food selectivity in ASD subjects. It
is advisable do not fight for food with ASD children,
but it would be suitable to tailor food proposals on the
basis of the specific bitter sensitivity of each subject.
For this purpose PROP strips could be useful in under-
standing to which child some food should be offered
and how.
In particular early detection of a PROP-sensitive sub-
ject could prevent the proposal of substances perceived
as particularly unpleasant, which could more easily
induce food refusal in ASDs. It could be also auspicable
to design more palatable foods, targeted to subjects
with marked food selectivity in order to ensure a proper
supply of nutritional factors, in particular in early
childhood.
INSAR Riccio et al./Food refusal and autism spectrum disorders
References
American Psychiatric Association. (2013). Diagnostic and statis-
tical manual for mental disorders (5th ed.). Arlington:
American Psychiatric Association.
Autism and Developmental Disabilities Monitoring Network
Surveillance Year 2008 Principal Investigators; Centers for
Disease Control and Prevention. (2012). Prevalence of
autism spectrum disorders–autism and developmental dis-
abilities monitoring network, 14 sites, United States, 2008.
Morbidity and Mortality Weekly Report Surveillance Sum-
maries. 2012;61(3):1–19.
Bandini, L.G., Anderson, S.E., Curtin, C., Cermak, S., Evans,
E.W., Scampini, R., et al. (2010). Food selectivity in chil-
dren with autism spectrum disorders and typically develop-
ing children. Journal of Pediatrics, 157, 259–264.
Bell, K.I., & Tepper, B.J. (2006). Short-term vegetable intake by
young children classified by 6-n-propylthoiuracil bitter-taste
phenotype. American Journal of Clinical Nutrition, 84,
245–251.
Bennetto, L., Kuschner, E.S., & Hyman, S.L. (2007). Olfaction
and taste processing in autism. Biological Psychiatry, 62,
1015–1021.
Bufe, B., Breslin, P.A., Kuhn, C., et al. (2005). The molecular
basis of individual differences in phenyl-thiocarbamide and
propylthiouracil bitterness perception. Current Biology, 15,
322–327.
Cermak, S.A., Curtin, C., & Bandini, L.G. (2010). Food selectiv-
ity and sensory sensitivity in children with autism spectrum
disorders. Journal of American Diet Association, 110, 238–
246.
Chong Cole, N., Wang, A.A., Donovan, S.M., et al. (2017). Var-
iants in chemosensory genes are associated with picky eat-
ing behavior in preschool-age children. Journal of
Nutrigenetics and Nutrigenomics, 10, 84–92.
Cooke, L.J., Haworth, C.M., & Wardle, J. (2007). Genetic and
environmental influences on children’s food neophobia.
American Journal of Clinical Nutrition, 86, 428–433.
De Bildt, A., Oosterling, I.J., van Lang, N.D., Sytema, S.,
Mindera, R.B., & van Engeland, H. (2011). Standardized
ADOS scores: Measuring severity of autism spectrum disor-
ders in a Dutch sample. Journal of Autism and Develop-
mental Disorders, 41, 311–319.
Dinehart, M.E., Hayes, J.E., Bartoshuk, L.M., et al. (2006). Bitter
taste markers explain variability in vegetable sweetness, bit-
terness, and intake. Physiological Behavior, 87, 304–313.
Drewnowski, A., & Gomez-Carneros, C. (2000). Bitter taste,
phytonutrients, and the consumer: A review. American
Journal of Clinical Nutrition, 72, 1424–1435.
Duchan, E., & Patel, D.R. (2012). Epidemiology of autism spec-
trum disorders. Pediatric Clinics of North America, 59, 27–
43. ix-x. Epub 2012/01/31.
Duffy, V.B., & Bartoshuk, L.M. (2000). Food acceptance and
genetic variation in taste. Journal of American Diet Associa-
tion, 100, 647–655.
Ekman, P., & Friesen, W. (1978). Facial action coding system:
A technique for the measurement of facial movement. Palo
Alto: Consulting Psychologists Press.
Ekman, P., & Friesen, W. (1982) Measuring facial movement
with facial action coding system. In: Ekman P., editor.
7
 Emotion in the human face. Cambridge University Press,
p.179–215.
Ekman, P., Friesen, W., & Hager, J. (2002). Facial action coding
system. Salt Lake City, Utah: Research Nexus.
Gotham, K., Risi, S., Pickles, A., & Lord, C. (2007). The autism
diagnostic observation schedule: Revised algorithms for
improved diagnostic validity. Journal of Autism and Devel-
opmental Disorders, 37, 613–627.
Gotham, K., Pickles, A., & Lord, C. (2009). Standardizing
ADOS scores for a measure of severity in autism spectrum
disorders. Journal of Autism and Developmental Disorders,
39, 693–705.
Greimel, E., Macht, M., Krumhuber, E., & Ellgring, H. (2006).
Facial and affective reactions to tastes and their modulation
by sadness and joy. Physiological Behavior, 89, 261–269.
Griffiths, R. (2006). Griffiths mental development scales—
extended revised. Firenze: Giunti Organizzazioni Speciali.
Hubbard, K.L., Anderson, S.E., Curtin, C., Must, A., & Bandini,
L.G. (2014). A comparison of food refusal related to charac-
teristics of food in children with autism spectrum disorder
and typically developing children. Journal of the Academy
of Nutrition and Dietetics, 114, 1981–1987.
Hayes, J.E., Wallace, M.R., Knopik, V.S., et al. (2011). Allelic vari-
ation in TAS2R bitter receptor genes associates with variation
in sensations from and ingestive behaviors toward common
bitter beverages in adults. Chemical Senses, 36, 311–319.
Hayes, J.E., Feeney, E.L., & Allen, A.L. (2013). Do polymor-
phisms in chemosensory genes matter for human ingestive
behavior?. Food Quality and Preference, 30, 202–216.
Isaksen, J., Diseth, T.H., Schjolberg, S., & Skjeldal, O.H. (2013).
Autism spectrum disorders–are they really epidemic?. Euro-
pean Journal of Paediatric Neurology, 17, 327–333.
Lane, A.E., Molloy, C.A., & Bishop, S.L. (2014). Classification
of children with autism spectrum disorder by sensory sub-
type: A case for sensory-based phenotypes. Autism
Research, 7, 322–333.
Leiter, R.G. (1979). Instruction manual for the leiter interna-
tional performance scale. Wood Dale, IL: Stoelting Co.
Lord, C., Rutter, M., & Le Couteur, A. (1994). Autism diagnos-
tic interview-revised: A revised version of a diagnostic inter-
view for caregivers of individuals with possible pervasive
developmental disorders. Journal of Autism and Develop-
mental Disorders, 24, 659–685.
Marı-Bauset, S., Zazpe, I., Mari-Sanchis, A., Llopis-Gonza lez, A.,
& Morales-Sua rez-Varela, M. (2014). Food selectivity in
autism spectrum disorders: A systematic review. Journal of
Child Neurol, 29, 1554–1561.
Mennella, J.A., Pepino, M.Y., Duke, F.F., & Reed, D.R. (2010).
Age modifies the genotype-phenotype relationship for the
bitter receptor TAS2R38. BMC Genet, 11, 60.
Mennella, J.A., Pepino, M.Y., Duke, F.F., et al. (2011). Psycho-
physical dissection of genotype effects on human bitter per-
ception. Chemical Senses, 36, 161–167.
Negri, R., Di Feola, M.D., Domenico, S., Scala, M.G., Artesi, G.,
Valente, S., Smarrazzo, A., . . . Greco, L. (2012). Taste percep-
tion and food choices. Journal of Pediatric Gastroenterology
and Nutrition, 54, 624–629.
Negri, R., Smarrazzo, A., Maio, A., IaccarinoIdelson, P., Sticco,
M., Biongiovanni, C., . . . Morini, G. (2015). Age variation
in bitter taste perception in relation to the TAS2R38 taste
8 Riccio et al./Food refusal and autism spectrum disorders
receptor phenotype. International Journal of Nutrition, 1,
2379–7835.
Postorino, V., Sanges, V., Giovagnoli, G., Fatta, L.M., De Peppo,
L., Armando, M., Vicari, S., & Mazzone, L. (2015). Clinical
differences in children with autism spectrum disorder with
and without food selectivity. Appetite, 92, 126–132.
Rozin, P., Lowery, L., Imada, S., & Haidt, J. (1999). The CAD
triad hypothesis: a mapping between three moral emotions
(contempt, anger, disgust) and three moral codes (commu-
nity, autonomy, divinity). Journal of Personality and Social
Psychology, 76, 574–586.
Rozin, P., Lowery, L., & Ebert, R. (1994). Varieties of disgust
faces and the structure of disgust. Journal of Personality
and Social Psychology, 66, 870–881.
Schmitt, L., Heiss, C.J., & Campbell, E.E. (2008). A comparison
of nutrient intake and eating behaviors of boys with and
without autism. Topics in Clinical Nutrition23, 23–31.
Schopler, E.R.R., & Renner, B.R. (1988). The childhood autism
rating scale (CARS). Los Angeles, CA: Western Psychological
Service Inc.
Schreck, K.A., Williams, K., & Smith, A.F. (2004). A comparison
of eating behaviors between children with and without
autism. Journal of Autism and Developmental Disorders,
34, 433–438.
Schreck, K.A., & Williams, K. (2006). Food preferences and fac-
tors influencing food selectivity for children with autism
spectrum disorders. Research in Developmental Disabilities,
27, 353–363.
Smutzer, G., Lam, S., Hastings, L., Desai, H., Abarintos, R.A.,
Sobel, M., & Sayed, N. (2008). A test for measuring gusta-
tory function. Laryngoscope, 118, 1411–1416.
Sparrow, S.S., Balla, D.A., & Cicchetti, D.V. (2003). Vineland
adaptive behavior scale Intervista Forma completa. Fierenze:
Giunti OS.
Steiner, J.E. (1977). Facial expressions of the neonate infant
indicating the hedonics of food-related chemical stimuli.
In: Weiffenbach JM, editor. Taste and development—the
genesis of sweet preference. Bethesda (MD): National Insti-
tutes of Health.
Steiner, J.E., Glaser, D., Hawilo, M.E., & Berridge, K.C. (2001).
Comparative expression of hedonic impact: Affective reac-
tions to taste by human infants and other primates. Neuro-
science and Biobehavioral Review, 25, 53–74.
Tavassoli, T., & Baron-Cohen, S. (2012). Taste identification in
adults with autism spectrum conditions. Journal of Autism
and Developmental Disorders, 42, 1419–1424.
Tepper, B.J., White, E.A., Koelliker, Y., Lanzara, C., D’adamo,
P., & Gasparini, P. (2009). Genetic variation in taste sensi-
tivity to 6-n-propylthiouracil and its relationship to taste
perception and food selection. Annales of New York Acad-
emy of Sciences, 1170, 126–139.
Whissell-Buechy, D. (1990). Effects of age and sex on taste sen-
sitivity to phenylthiocarbamide (PTC) in the Berkeley guid-
ance sample. Chemical Senses, 15, 39–57.
Williams, P.G., Dalrymple, N., & Neal, J. (2000). Eating habits
of children with autism. Pediatric Nursing, 26, 259–264.
Zeinstra, G.G., Koelen, M.A., Colindres, D., Kok, F.J., & de
Graaf, C. (2009). Facial expressions in school-aged children
are a good indicator of dislikes,’ but not of likes’. Food
Quality and Preference, 20, 620–624.
INSAR