’

Avoidant/Restrictive Food Intake

Disorder (ARFID)
Jacqueline Zimmerman, MS, RDN, and Martin Fisher, MD

Avoidant/restrictive food intake disorder (ARFID) is an and medical and psychiatric comorbidities of this patient
entirely new diagnosis in the DSM-5. ARFID replaces “feed- population are described and compared to other eating
ing disorder of infancy or early childhood,” which was a disorders. Evaluation and management strategies are also
diagnosis in the DSM-IV restricted to children 6 years of age discussed. No data yet exist regarding ARFID’s prognosis
or younger; ARFID has no such age limitations and it is and prevention; however, recommendations to guide parents
distinct from anorexia nervosa and bulimia nervosa in that in establishing appropriate infant and child feeding practices
there is no body image disturbance. ARFID involves a are provided.
complex and heterogenous etiology, which is reviewed
herein. What is known to date regarding the characteristics Curr Probl Pediatr Adolesc Health Care 2017;47:95-103

Introduction DSM-5 in 2013. ARFID is an entirely new diagnosis

within the section of feeding and eating disorders in
ay 2013 marked the much-anticipated publi- the DSM-5. The DSM-5 provides diagnostic specif-
M cation of the Diagnostic and Statistical
Manual, 5th Edition1 (DSM-5) by the
icity to those patients who do not fear weight gain,
but simply cannot meet their nutritional needs for a
American Psychiatric Association (APA). The first variety of reasons. The diagnosis of ARFID in the
major update since 1994, the DSM-5 includes exten- DSM-5 does not have an age restriction; it can
sive revisions throughout, and includes the new therefore be applied to children, adolescents, and
section, “feeding and eating adults. However, since the age
disorders,” which replaces and groups in which it is most
combines two former sections, ARFID is an entirely new commonly diagnosed are
“eating disorders” and “feeding
and eating disorders of infancy
diagnosis within the section of older children and younger
adolescents, it is most often
or early childhood.” The new Feeding and Eating Disorders in pediatricians who are called
“feeding and eating disorders” the DSM-5. upon to be the first practi-
section includes expanded crite- tioners to consider the
ria for familiar diagnoses, such diagnosis.
as anorexia nervosa (AN) and bulimia nervosa (BN), as As early as 1992, Lask and Bryant-Waugh2 describe
well as several new diagnoses, including purging several childhood-onset eating disorders involving
disorder and avoidant/restrictive food intake disorder food restriction that failed to meet criteria for AN
(ARFID), the latter of which replaces “feeding disorder due to lack of body image distortion or a desire to lose
of infancy or early childhood,” which was a diagnosis weight. In 2002, Watkins and Lask3 expanded on
in the DSM-IV restricted to children 6 years or
these atypical eating disorders in children and adoles-
younger. ARFID is not genuinely a new condition,
but this restrictive eating disorder was not clearly cents and classified them as follows: food avoidance
defined and characterized until the publication of the emotional disorder, characterizing children with a
“primary emotional disorder in which food avoidance
From the Division of Adolescent Medicine, Northwell Health, Hofstra
is a prominent feature;” selective eating, describing
Northwell School of Medicine, Cohen Children's Medical Center, Hemp- children who “eat very few different foods and
stead, NY. sometimes are particular about the brand of food or
Curr Probl Pediatr Adolesc Health Care 2017;47:95-103 where the food was bought” but whose growth and
1538-5442/$ - see front matter
& 2017 Elsevier Inc. All rights reserved.
development is not negatively impacted by their
http://dx.doi.org/10.1016/j.cppeds.2017.02.005 eating patterns; functional dysphagia, which involves

Curr Probl PediatrAdolesc Health Care, April 2017 95

a “fear of swallowing, vomiting, or choking, which other eating disorders in that management involves a
makes the child anxious about and resistant to eating multidisciplinary team approach and includes medical,
normally, which results in a marked avoidance of nutritional, and psychological practitioners. However,
food,” and for which there is typically “an easily the role each discipline plays may vary to a greater
identifiable precipitant, such as having witnessed degree than for other types of eating disorders based on
someone choking while eating;” and pervasive refusal the many variants in the etiology of the ARFID
syndrome, affecting a smaller number of children, but diagnosis. As ARFID has been in existence as an
which is “a potentially life-threatening condition official diagnosis for less than four years, available
manifested by a profound and pervasive refusal to literature is primarily limited to retrospective chart
eat, drink, walk, talk, or care for themselves in any reviews and single case studies.
way over a period of several months.” None of these
disorders involve body image distortion or a desire to
lose weight, and as their descriptions indicate, there is Diagnosis
a constellation of reasons why these patients, who
would now be diagnosed with ARFID, do not meet As noted earlier, while feeding and eating disorders
their nutritional needs. of infancy and early childhood in the DSM-IV was
Indeed, ARFID may be the eating disorder with the limited to children not older than 6 years, ARFID has
most heterogenous etiology. Several subcategories no diagnostic age restrictions. ARFID distinguishes a
have emerged in the research published thus far that cohort of individuals who experience persistent diffi-
attempt to define the reasons behind restricted nutri- culty in meeting their nutritional needs despite a lack of
tional intake. In a retrospective study conducted among body image or weight concerns. The DSM-5 sets forth
seven adolescent medicine divisions, 712 patients four diagnostic criteria, titled criterion A through D,
presented for initial evaluations of eating disorders each of which must be satisfied for a diagnosis of
during a 1-year period, and 98 (13.8%) met criteria for ARFID to be made. Criterion A of the DSM-5
ARFID. According to documented symptoms, 28.7% definition states that the eating disturbance is coupled
of ARFID patients had selective eating since early with at least one of the following: significant weight
childhood, 21.4% experienced generalized anxiety, loss (or failure to meet expected weight and height
19.4% had gastrointestinal symptoms, 13.2% had a trajectories in children and adolescents); nutritional
history of vomiting or choking, 4.1% had food deficiencies (such as iron deficiency anemia); a
allergies and 13.2% had other reasons for their dependence on nutritional supplements, (i.e., oral or
4
restricted eating. Another study divided 33 ARFID enteral formulas), to meet energy requirements without
patients into the following four groups: insufficient an underlying condition necessitating this; and/or
intake/little interest in feeding (57.6%), limited significant interference with day-to-day functioning
diet due to sensory characteristics of food (21.2%), due to the inability to eat appropriately. Some individ-
aversive/traumatic experience uals with ARFID experience
(9.1%), and other reasons high levels of anxiety and dis-
(12.1%).5 ARFID is distinct from anorexia tress when near certain foods or
ARFID is distinct from ano- nervosa and bulimia nervosa in faced with particular smells and
rexia nervosa and bulimia textures, resulting in missed
nervosa in that there is no body that there is no body image school or work and social iso-
image disturbance; ARFID disturbance; ARFID patients may lation. Additionally, inadequate
patients may express a desire to express a desire to increase nutrition may contribute to irri-
increase their eating and gain their eating and gain needed tability, anxiety, and other
needed weight, but they simply changes in mood.
cannot get themselves to do it.
weight, but they simply cannot Bryant-Waugh, 20136 pre-
The reasons behind this inability get themselves to do it. sented a case study and devel-
appear to be as varied as the oped a checklist for the criteria
patients themselves. The literature to date has identi- described earlier (criterion A in the DSM-5) that
fied some common patterns, which may be used to includes seven questions to guide practitioners in
inform the course of treatment. ARFID is similar to establishing whether criterion A is fully satisfied.

96 Curr Probl PediatrAdolesc Health Care, April 2017

Questions target the patient's food intake range and been our experience—that short-term and long-term
amount, the persistence of the eating disturbance, and ARFID patients require different treatment strategies
if signs of nutritional deficiency or interference with and variable outcomes can be expected, with the short-
daily living are present. term patients faring better than the long-term patients
For ARFID to be diagnosed, the eating disturbance regarding both success and length of treatment
cannot be associated with a “culturally sanctioned required.
1
practice” (e.g., religious fasting It is crucial to appreciate
or intentional dieting) or a short- that ARFID is not meant to
age of available food (DSM-5 ARFID is not meant to encompass encompass all picky/fussy eat-
criterion B). There must be no all picky/fussy eaters. Its ers. Its purpose is to identify
evidence of body image or weight purpose is to identify patients patients with clinically signifi-
concerns, and the eating disturb- cant restrictive eating, the
ance must not be present solely as with clinically significant restric- magnitude of which results in
part of anorexia nervosa or buli- tive eating, the magnitude of severe nutritional deficiencies
mia nervosa (DSM-5 criterion C). which results in severe nutritional and/or persistent inability to
In addition, another medical or deficiencies and/or persistent meet energy needs. ARFID
psychiatric condition cannot bet-
ter explain the eating disturbance;
inability to meet energy needs. does not encompass what are
considered age-appropriate
if a concurrent condition exists, changes in eating patterns,
the “severity of the eating disturbance exceeds that including food neophobia, defined as the rejection of,
routinely associated with the condition or disorder or reluctance to eat, foods that are novel or unknown,
and warrants additional clinical attention.” (DSM-5 which commonly develops in toddlerhood and peaks
1
criterion D). between ages 2 and 6 years before decreasing as the
As discussed earlier, current research has demon- child ages and then stabilizing.7 There are some
strated several common symptom-based groupings or disagreements in the literature as to when food neo-
subtypes of ARFID. For purposes of evaluation and phobia stabilizes: some studies suggest it decreases
treatment, it may be useful to broaden these categories gradually until early adulthood, while others have
and distinguish short-term ARFID patients from long- found that stabilization occurs earlier in adolescence.7
term ARFID patients. This delineation is based not on While there is no gold-standard definition of picky/
symptoms, but primarily considers the length of time fussy eating, it is commonly described as consuming
the symptoms have been experienced. Short-term an inadequate variety of foods by rejecting a substan-
patients include those with a more recent, acute onset tial amount of both familiar and new foods.7 Picky/
of symptoms: fear of choking or vomiting after fussy eating is not necessarily associated with under-
witnessing (or experiencing) such an episode, or weight and nutritional deficiencies in children; like-
possibly a more sudden onset of gastrointestinal wise, not all picky/fussy eaters will develop ARFID.
symptoms or lingering fear after If growth and development are
an episode of gastrointestinal indeed impaired, or the patient
distress. Long-term ARFID Long-term ARFID patients is not able to engage in normal
patients include those who report
a long-standing history of poor or
include those who report a long- daily activities, such as going
to school or engaging in social
selective eating, and possibly standing history of poor or activities, then further evalua-
those patients with an extended selective eating, and possibly tion is warranted. Feeding
history of generalized anxiety or those patients with an extended experts have opined that by
gastrointestinal issues, which history of generalized anxiety or its nature as a DSM-5 diagno-
have interfered with normal eat- sis, ARFID is a psychiatric
ing throughout childhood or over gastrointestinal issues, which disorder, and as such, “marked
the course of several years. have interfered with normal interference with psychosocial
While there is no outcomes- eating throughout childhood or functioning,” which is only
based research available for over the course of several years. one out of four possible symp-
ARFID, it is likely—and it has toms listed in DSM-5 criterion

Curr Probl PediatrAdolesc Health Care, April 2017 97

A, should be required, not optional, for an ARFID criteria for ARFID. In line with other studies, all
diagnosis to be made.8 However, the DSM-5 permits patients diagnosed with ARFID would have previously
an ARFID diagnosis based on the presence of only received a DSM-IV diagnosis of EDNOS. In addition,
one (or more) symptoms listed in criterion A, which may patients with ARFID were significantly younger than
or may not include psychosocial functioning difficulty. those with other eating disorders (11.1 ⫾ 1.7 years for
The discussion above regarding picky/fussy eating ARFID vs. 14.2 ⫾ 1.5 years for other EDs) and a
centers on only one possible presentation of ARFID in greater percentage were male (20.5% for ARFID vs.
which the patient has historically been a poor eater, i.e., a 4.5% for other EDs).11
long-term ARFID patient. In reality, the ARFID patient Another adolescent medicine multi-institution study
population is perplexingly heterogeneous. diagnosed ARFID in 98 (13.8%) of the 712 presenting
eating disorder cases. Compared to those with AN or BN,
the ARFID population had a higher percentage of males
Epidemiology and Etiology (28.6% ARFID vs. 14.3% AN vs. 6.0% BN), the average
age was younger (12.9 ⫾ 2.5 years ARFID vs.
With the introduction of ARFID and other new and 15.6 ⫾ 1.9 AN vs. 16.5 ⫾ 1.3 BN) and duration of
revised diagnoses in the DSM-5, the number of illness was longer (33.3 ⫾ 41.3 months ARFID vs.
“unclassified” eating disorders declined dramatically. 14.5 ⫾ 12.2 AN vs. 23.5 ⫾ 17.1 BN).4 In addition,
Patients now diagnosed with ARFID would previously ARFID patients were more likely to have comorbid
have been relegated to eating disorder not otherwise medical conditions or anxiety disorder, but less likely to
specified (EDNOS), a catch-all diagnosis in the DSM- have a mood disorder. This study, discussed earlier, was
IV for patients not meeting full criteria for other eating also the first to detail ARFID's varied etiology in
disorders such as anorexia nervosa or bulimia nervosa. assigning subcategories to the groupings of symptoms
A study in our own division of adolescent medicine reported: selective (picky) eating since early childhood,
demonstrated the improvement in diagnostic specific- generalized anxiety, gastrointestinal symptoms, fears of
ity of the DSM-5 by examining the records of 309 eating secondary to fear of choking or vomiting, food
patients (83.2% female) seen for eating disorder allergies, and other less common reasons. The data from
evaluations and assigning patients diagnoses based that study are included as Table 1 of this review.
on both the DSM-IV and proposed DSM-5 criteria. One study stands out for its contrasting results: a
A diagnosis of EDNOS was assigned to 198 patients retrospective chart review of initial adolescent eating
(64.6%) using DSM-IV criteria. This number dropped disorder cases spanning 12 years (2000–2011) and
dramatically using DSM-5 criteria. All except four including 205 patients, found that only 5% met criteria
EDNOS patients were reclassified using DSM-5 diag- for ARFID,12 notably lower than the 14–22% range
noses, including 60 patients now diagnosed with reported by the four other studies discussed herein; it is
ARFID, of which 61.7% were female. In summary, unclear why this population differed. Compared to a
patients with ARFID comprised 20% of the total matched sample of patients with AN, patients with
sample and 30% of the EDNOS group. Also note- ARFID were younger, there was a significantly higher
worthy is that all patients with a diagnosis other than proportion of patients under the age of 12 years old, and a
EDNOS kept their original diagnosis with use of the higher percentage of male patients (20.6%, n ¼ 7 ARFID
new DSM-5 criteria.9 vs. 8.3%, n ¼ 3 AN), although the latter difference was
Six institutions conducted a similar study and found not statistically significant. ARFID patients presented
that when using DSM-5 criteria, 31 (14%) of 215 with the following symptoms, which were in line with
adolescent patients presenting for an initial eating other studies: abdominal pain (35.3%), fear of vomiting
disorder evaluation received a diagnosis of ARFID. (26.5%), generalized anxiety with eating (20.6%), com-
Compared to patients with DSM-5 anorexia nervosa, plaints of feeling full (20.6%), nausea (17.6%), and
ARFID patients' weights were slightly higher on unpleasant sensory experiences associated with eating
average, albeit still underweight (median body weight (17.6%).12 This study also reported that 12% of ARFID
percentage for ARFID 90.2 ⫾ 16.3% vs. 83.8 ⫾ 6.8% patients eventually changed diagnoses to AN restricting
for AN).10 In a retrospective chart review of 177 subtype over the course of treatment due to changes in
patients attending a day treatment program for eating (or acknowledgment of) body image or weight gain
disorders during 2008–2012, 22.5% met DSM-5 concerns.

98 Curr Probl PediatrAdolesc Health Care, April 2017

TABLE 1. Clinical Characteristics of Patients with ARFID, Anorexia Nervosa, or Bulimia Nervosa
ARFID (n¼98) Anorexia Nervosa (n¼98) Bulimia Nervosa (n¼66) Statistics (ANOVA)
Age 12.9 ⫾ 2.5 15.6 ⫾ 1.9* 16.5 ⫾ 1.3* F[2,259] ¼ 71.2, po.001
% Median BMI 86.5 ⫾ 15.1 81.0 ⫾ 9.2* 107.5 ⫾16* F[2,259] ¼ 80.8, po.001
Lowest Weight (lbs) 76.9 ⫾ 26.2 91.0 ⫾ 16.1* 117.3 ⫾ 21.0* F[2,247] ¼ 64.7, po.001
Highest Weight (lbs) 89.8 ⫾ 33.1 118.7 ⫾ 28.3* 142.9 ⫾ 27.2* F[2,245] ¼ 59.1, po.001
Duration (months) 33.3 ⫾ 41.3 14.5 ⫾ 12.2* 23.5 ⫾ 17.1 F[2,258] ¼ 11.3, po.001
Gendera,b Chi-square (df ¼ 2) ¼ 15.0, po.001
Female (%) 71.3 85.7 94.0
Male (%) 28.6 14.3 6.0
Intake Setting Chi-square (df ¼ 2) ¼ 5.6, po0.10
OPD (%) 87.7 85.7 97.0
Other (%) 12.3 14.3 3.0
Referral Sourceb Chi-square (df ¼ 10) ¼ 26.7, po0.01
Self (%) 6.2 10.2 15.5
PCP (%) 51.6 50.0 53.0
Mental Health (%) 11.3 16.3 22.7
EmergencyDept (%) 10.3 11.2 4.6
Social Service (%) 1.0 4.1 1.5
Other (%) 0.0 3.0 3.0
Medical Condition Or Sxa,b Chi-square (df ¼ 14) ¼ 54.4, o.001
Yes, related (%) 34.6 8.2 4.6
Yes, unrelated (%) 16.3 2.0 6.1
None (%) 49.1 89.8 89.3
Mood Disorderb Chi-square (df ¼ 4) ¼ 33.3, po.001
MDD/Dysthymia (%) 7.2 19.4 23.1
Other (%) 11.3 11.2 35.4
None (%) 81.5 69.4 41.5
Anxiety Disordera,b Chi-square (df ¼ 6) ¼ 23.4, po.001
GAD (%) 28.6 14.3 7.6
OCD (%) 6.1 8.2 1.5
Other (%) 23.5 13.3 24.2
None (%) 41.8 64.2 66.7
ARFID ¼ Avoidant / Restrictive Food Intake Disorder, OPD ¼ out-patient department, PCP ¼ primary care physician, BMI¼ Body Mass Index
MDD ¼ major depressive disorder, GAD ¼ generalized anxiety disorder, OCD ¼ obsessive-compulsive disorder
*Significant difference from ARFID, pr0.05 by Tukey’s HSD
a
Significant difference between Anorexia Nervosa and ARFID by chi-square, po.05 after Bonferroni correction
b
Significant difference between Bulimia Nervosa and ARFID by chi-square, po.05 after Bonferroni correction

Two non-eating disorder focused studies have be diagnosed with PANS/PANDAS, there must be acute
reported on prevalence rates of ARFID in the general onset of either obsessive-compulsive disorder (OCD) or
pediatric population. In the first study, researchers severely restricted eating, as well as multiple neuro-
examined 2231 consecutive new referrals to pediatric psychiatric symptoms. Restrictions in food intake are
gastroenterology specialty clinics via retrospective typically due to fears of food contamination, choking,
chart review.5 Patients were 8–18 years old and 1.5% swallowing, or sensory concerns, all of which could also
(33 patients) were found to have a diagnosis of be present in a patient with ARFID. All PANS cases
ARFID. One or more ARFID symptoms were present described involved food restriction and also met criteria
in an additional 54 patients (2.4%), but a diagnosis for ARFID. The authors highlight important differences
could neither be assigned nor excluded based on in presentation of PANS/PANDAS (primarily the
available chart information. There were very few cases acuity of symptom onset) as well as in treatment and
of other eating disorders (three cases of AN, two cases outcomes. According to the authors, antibiotic or
of BN, and one case of BED). immunomodulatory therapy often results in swift symp-
In the second study, researchers reviewed 29 cases of tom resolution of PANS/PANDAS. This study points
pediatric acute-onset neuropsychiatric syndrome out the need for consideration of a PANS/PANDAS
(PANS) and pediatric acute-onset neuropsychiatric diagnosis in patients with acute onset of restrictive
disorder associated with streptococcal infections eating or food avoidance, without prior history of picky
(PANDAS) in children ages 5-12 years.13 In order to eating.

Curr Probl PediatrAdolesc Health Care, April 2017 99

 In total, all studies to date have shown that a greater A recent prospective cohort study of selective
number of males than females have been diagnosed eating in preschoolers found that moderate and
with ARFID relative to other eating disorders. severe levels of selective eating was associated with
Although the average age of diagnosis reported thus higher rates of anxiety, depression, and attention-
far is between 11 and 14 years, emerging case reports deficit hyperactivity disorder.17 This study also found
show that adults can develop ARFID as well.14,15 that selective eating is a marker for later increased
However, a fair percentage of older children and psychopathological symptoms, and that these symp-
adolescents are diagnosed with ARFID around the toms became more severe as selective eating also
time of their typical growth spurt, which matches the increased in severity. However, a clear definition of
reported average age of diagnosis. This may primarily selective eating and the degree of food restriction was
relate to the long-term ARFID patients in that through- lacking in this study; therefore, it is unknown if these
out childhood they may be growing appropriately, patients would have received a definitive diagnosis
albeit following the lower percentiles for weight and of ARFID.
height. Once these patients enter early adolescence, Pennell et al.18 describe two cases of ARFID and
their restrictive and selective eating may simply not be concurrent stimulant-treated attention-deficit hyper-
adequate to maintain their growth trajectories, thus activity disorder (ADHD) in a 10-year-old male and a
garnering more concern from pediatricians. Indeed, it 9-year-old female hospitalized due to cachexia and
is pediatricians who are most often called to diagnose medical instability. Both patients had a long history
these patients and guide families toward treatment of selective eating since infancy or early childhood,
options. and it appeared that the initiation of stimulant
medication to treat ADHD had exacerbated pre-
existing ARFID, resulting in swift and severe weight
Medical and Psychiatric Comorbidities loss. As stimulants have an appetite-suppres-
and Complications sing effect, both patients were administered the
atypical antipsychotic medication risperidone while
While ARFID encompasses a hospitalized, as a method
broad spectrum of etiologies, it has
been consistently demonstrated that
ARFID patients are more likely to to increase appetite and
reduce the severity of
ARFID patients are more likely to be younger, male and have a ADHD symptoms, as well
be younger, male and have a higher higher incidence of comorbid as the oppositional defiant
incidence of comorbid medical or medical or psychiatric disorders disorder (ODD) symptoms
psychiatric disorders compared to
compared to those with other also present in the male
those with other eating disor- patient. The authors advise
ders.4,5,11,12 Higher rates of eating disorders. practitioners to screen for
4,11
OCD and generalized anxiety dis- selective eating in all
order4,5,11,12 have been reported by patients for which stimulants are being considered
multiple sources. Higher rates of autism spectrum to treat their ADHD.
disorder, learning disorders, and cognitive impairment
have been reported by one source.11 Two studies found Evaluation and Management
ARFID patients to have a longer duration of illness,4,16
whereas another found it to be similar to other eating No empirical data have been published regarding
disorders.11 evaluation and treatment since the inception of
Weight and degree of malnutrition were similar to ARFID in the DSM-5; treatment has primarily been
patients with AN in two studies in outpatient set- described in case studies or studies in which ARFID
tings.9,11 In one study, a significantly lower number of was retrospectively diagnosed with chart reviews.
ARFID patients required hospitalization (or had been Management considerations include the effect on
previously hospitalized) when compared to patients with weight and growth, the extent of nutritional
AN.12 For both the ARFID and AN patients that were compromise, the impediment of social and emo-
hospitalized, admission was required due to medical tional function,6 as well as whether the patient is
instability, i.e., unstable vital signs and/or low weight. considered a long-term or short-term ARFID sufferer.

100 Curr Probl PediatrAdolesc Health Care, April 2017

The nature of ARFID is such and relaxation techniques. Prob-
that treatment must focus on The nature of ARFID is such that lematic foods are progressively
more behavioral and nutritional incorporated into the patient's
approaches than is required in
treatment must focus on more eating pattern. ARFID treatment
the management of other eating behavioral and nutritional goals should be tailored accord-
disorders. In addition, patients approaches than is required in ing to the patient's needs and
with ARFID often present with the management of other eating should be realistic and attain-
complex medical and psycho- disorders. able as follows: to ameliorate
logical histories, making nutritional, physical, and emo-
generalized treatment recommen- tional risks, while assisting the
dations nearly impossible. However, there are within- patient in managing anxiety and expanding their range
group similarities among those patients with short-term of eating.6
ARFID and long-term ARFID, despite patients having Patients may need hospitalization if there is virtually
various reasons for their restricted dietary intake. no oral intake or if other pressing medical concerns are
The success of behavioral and nutritional treatment present. In a retrospective study comparing ARFID and
may largely depend on whether the patient has had AN patients hospitalized for medical stabilization,
ARFID long-term or short-term. In our experience, ARFID patients had a longer length of stay (8 days
short-term ARFID patients fare better and recover vs. 5 days for AN) and greater reliance on enteral
more quickly once intensive treatment is initiated. nutrition, despite a similar calorie intake as AN
Several factors may contribute to this, including the patients.19 The authors note that greater reliance on
fact that a patient who was previously eating enteral nutrition—not the ARFID diagnosis—may
adequately but recently developed ARFID will have explain the differences in length of hospitalization.
a standard of regular eating the patient remembers— Patients primarily receiving enteral nutrition were
and can return to—with behavioral and nutritional started on lower calories and experienced slower
assistance. Working with a mental health professional weight gain. At presentation, ARFID patients (13%,
to address the underlying anxiety and a Registered n ¼ 41) were younger, had less weight loss and less
Dietitian Nutritionist to provide nutritional guidance bradycardia than patients with AN (64%, n ¼ 203).
can often result in a swift turnaround of a short-term Among the ARFID patients in this study, fear of
eating disturbance. A patient with long-term ARFID vomiting (38%) and abdominal pain (26%) were the
who has historically been a selective eater has no such most common contributors to weight loss. In contrast
“normal eating” standard and, as treatment progresses, to other findings, this study reported much higher rates
is in unchartered nutritional territory. These long-term of psychiatric comorbidities in the patients with AN
patients likely require more intensive treatment over (42%) than the patients with ARFID (12%). After a
an extended period. The benchmarks for successful follow-up period of 1 year, more patients with ARFID
treatment of long-term patients with ARFID may be were deemed to be in remission (62% vs. 46% for AN),
different as well, i.e., the goals may be to minimize although the difference was not statistically significant.
some symptoms, such as the impact on psychosocial As mentioned earlier, two case studies reported that
functioning or reliance on oral nutritional supple- hospitalized patients with ARFID have benefitted from
ments, instead of complete elimination of the antipsychotic medications, which improved anxiety,
diagnosis. appetite, and sleep.18
The literature to date indicates that treatment should In the report of the National Eating Disorders Quality
assess and target the aspects most responsible for the Improvement Collaborative, 14 adolescent medicine
avoidance or restriction of food intake. ARFID patients programs retrospectively reviewed the charts of 700
may have long-standing histories of poor nutritional adolescents at intake and 1-year follow-up.16 Patients
intake or may have recently started restricting food due were 9–21 years, had a DSM-5 restrictive eating
to anxiety, fear of choking/vomiting, or avoidance of disorder (AN, atypical AN, or ARFID) and had at
physical pain, or sensory experiences. Treatment least three visits. In this study, 12.4% of patients were
reported in case studies often includes management diagnosed with ARFID, a similar range as reported by
of anxiety, systematic desensitization, and a structured others. No differences were found among institutions
nutritional plan with gradual exposure based therapy in terms of weight restoration, regardless of taking

Curr Probl PediatrAdolesc Health Care, April 2017 101

varied approaches to treatment (e.g., individual ther- which parental pressure to eat contributes to, or is
apy, family-based treatment (FBT), inpatient medical, reactive to, the problematic eating pattern is not yet
partial hospitalization, and intensive outpatient). Nota- well understood.
bly, fewer ARFID patients were followed for 1 year, Many parents have concerns regarding infant and
despite having much lower odds of weight recovery child feeding and eating, especially perceived under-
compared to AN and atypical AN patients. The authors eating. Pediatricians can guide families in establishing
note, “patients with ARFID may be dropping out of appropriate infant and child feeding practices, which
care because they need to be treated with different will enable the child to develop a healthy relationship
therapies compared with the classic treatment for AN with food and competence with eating, and to grow
with nutrition, mental health treatment, and medical predictably. The Satter Feeding Dynamics Model,
visits.” It is also possible that patients with short-term including the division of responsibility in feeding, is
ARFID (i.e., those with fear of vomiting or choking or recognized by The American Academy of Pediatrics,
recent onset of GI symptoms) do not remain in The Academy of Nutrition and Dietetics, and the United
treatment for very long because once they get a States Department of Agriculture (USDA) Food and
turnaround in their symptoms recovery proceeds much Nutrition Service as the gold-standard for child feed-
more quickly than for those with other types of ing.26 Division of Responsibility in Feeding involves
restrictive eating disorders. assigning the parent and child different roles: the parent
is responsible for the what, when, and where of feeding;
the child is responsible for how much to eat and whether
Prevention and Prognosis to eat at all. The parents' role includes the following:
preparing and providing meals and snacks at consistent
Limited information is available regarding all aspects times, making mealtimes pleasant, modeling appropriate
of ARFID, and there are simply not yet adequate data mealtime behavior, and not pressuring children to eat.
to determine the expected course of illness and Parents should not give food “handouts” whenever the
prognosis of a patient with ARFID. While there is no child requests, but should adhere to general set meal and
way to predict who will develop ARFID, pediatricians snack times. Following appropriate guidelines for child
should be aware of the high parental concern regarding feeding will help children learn to eat the foods the
child food intake. In particular, many parents are parents eat, to establish internal food regulation cues,
concerned that their child does not eat enough, often and to be comfortable with eating. While this does not
leading to the parent pressuring or bribing the child to guarantee prevention of ARFID, it will give the child
eat even when the child is growing predictably and is the best chance of being a lifelong successful and
not underweight.20 competent eater.
There is evidence from the feeding literature that In summary, ARFID is a diagnosis that is a mere four
parental pressure to eat actually negatively impacts years old, newly included in the DSM-5. What is
food intake, is correlated with higher levels of picky known about the nature of ARFID thus far is that it
eating and lower weight during childhood,21,22 and does not include body image distortion; it is primarily
contributes to both a dissociation with hunger/satiety diagnosed in older children and adolescents, but the
cues and with eating for external reasons, such as potential age range is quite wide; and it is diagnosed in
emotion.23,24 A recent study investigated recollec- more males than females relative to other eating
tions of picky eating and pressure to eat, as well as disorders. Treatment of ARFID involves both behav-
current self-reported intuitive eating and disordered ioral and nutritional approaches; successful treatment
eating behaviors in 170 college-aged students. Anal- is likely greater in short-term vs. long-term ARFID
ysis revealed that parental pressure to eat, but not patients. Over the next few years, we are certain to
picky eating during childhood, was associated with learn more regarding effective treatment for all patients
greater disordered eating behaviors in young adult- diagnosed with ARFID.
hood.25 The disordered eating behaviors included
those associated with bulimia, but not a drive for References
thinness. The authors conclude that parental pressure 1. American Psychiatric Association. Diagnostic and Statistical
may interfere with the normal development of intui- Manual of Mental Disorders. 5th Edition. Washington, DC:
tive eating; however, they also note that the extent to American Psychiatric Association, 2013.

102 Curr Probl PediatrAdolesc Health Care, April 2017

 2. Lask B, Bryant-Waugh R. Early-onset anorexia nervosa and 15. Lopes R, Melo R, Curral R, Coelho R, Roma-Torres A. A case
related eating disorders. J Child Psychol Psychiatry 1992; of choking phobia: towards a conceptual approach. Eat Weight
33(1):281–300. Disord 2014;19(1):125–31.
3. Watkins B, Lask B. Eating disorders in school-aged children. 16. Forman SF, McKenzie N, Hehn R, et al. Predictors of outcome
Child Adolesc Psychiatr Clin N Am 2002;11(2):185–99. at 1 year in adolescents with DSM-5 restrictive eating
4. Fisher MM, Rosen DS, Ornstein RM, et al. Characteristics of disorders: report of the national eating disorders quality
avoidant/restrictive food intake disorder in children and improvement collaborative. J Adolesc Health 2014;55(6):
adolescents: a “new disorder” in DSM-5. J Adolesc Health 750–6.
2014;55(1):49–52. 17. Zucker N, Copeland W, Franz L, et al. Psychological and
5. Eddy KT, Thomas JJ, Hastings E, et al. Prevalence of DSM-5 psychosocial impairment in preschoolers with selective eating.
avoidant/restrictive food intake disorder in a pediatric gastro- Pediatrics 2015;136(3):e582–90.
enterology healthcare network. Int J Eat Disord 2015;48(5): 18. Pennell A, Couturier J, Grant C, Johnson N. Severe avoidant/
464–70. restrictive food intake disorder and coexisting stimulant
6. Bryant-Waugh R. Avoidant restrictive food intake disorder: an treated attention deficit hyperactivity disorder. Int J Eat Disord
illustrative case example. Int J Eat Disord 2013;46(5):420–3. 2016. http://dx.doi.org/10.1002/eat.22602:[Epub ahead
7. Dovey TM, Staples PA, Gibson EL, Halford JC. Food of print].
neophobia and “picky/fussy” eating in children: a review. 19. Strandjord SE, Sieke EH, Richmond M, Rome ES. Avoidant/
Appetite 2008;50(2-3):181–93. restrictive food intake disorder: illness and hospital course in
8. Satter E. What is ARFID and what does it have to do with patients hospitalized for nutritional insufficiency. J Adolesc
feeding dynamics and eating competence? Family Meals Health 2015;57(6):673–8.
Focus 89. Ellyn Satter Institute Website 2015. http://ellynsatter 20. Brown CL, Pesch MH, Perrin EM, Appugliese DP, Miller AL,
institute.org/fmf/fmf89.php:[Accessed 05.11.16]. Rosenblum K, Lumeng JC. Maternal concern for child under-
9. Fisher M, Gonzalez M, Malizio J. Eating disorders in eating. Acad Pediatr 2016;16(8):777–82.
adolescents: how does the DSM-5 change the diagnosis? Int 21. Farrow CV, Blissett J. Controlling feeding practices: cause or
J Adolesc Med Health 2015;27(4):437–41. consequence of early child weight? Pediatrics 2008;121(1):
10. Ornstein RM, Rosen DS, Mammel KA, et al. Distribution of e164–169.
eating disorders in children and adolescents using the proposed 22. Ventura AK, Birch LL. Does parenting affect children's eating
DSM-5 criteria for feeding and eating disorders. J Adolesc and weight status? Int J Behav Nutr Phys Act 2008;5:15.
Health 2013;53(2):303–5. 23. Carper JL, Orlet Fisher J, Birch LL. Young girls' emerging
11. Nicely TA, Lane-Loney S, Masciulli E, Hollenbeak CS, Ornstein dietary restraint and disinhibition are related to parental control
RM. Prevalence and characteristics of avoidant/restrictive food in child feeding. Appetite 2000;35(2):121–9.
intake disorder in a cohort of young patients in day treatment for 24. van Strien T, Bazelier FG. Perceived parental control of
eating disorders. J Eat Disord 2014;2(1):21. food intake is related to external, restrained and emotional
12. Norris ML, Robinson A, Obeid N, Harrison M, Spettigue W, eating in 7-12-year-old boys and girls. Appetite 2007;49(3):
Henderson K. Exploring avoidant/restrictive food intake dis- 618–25.
order in eating disordered patients: a descriptive study. Int J 25. Ellis JM, Galloway AT, Webb RM, Martz DM, Farrow CV.
Eat Disord 2014;47(5):495–9. Recollections of pressure to eat during childhood, but not
13. Toufexis MD, Hommer R, Gerardi DM, et al. Disordered picky eating, predict young adult eating behavior. Appetite
eating and food restrictions in children with PANDAS/PANS. 2016;97:58–63.
J Child Adolesc Psychopharmacol 2015;25(1):48–56. 26. Satter E. Ellyn Satter's division of responsibility in feeding.
14. King LA, Urbach JR, Stewart KE. Illness anxiety and Ellyn satter institute website. http://ellynsatterinstitute.org/
avoidant/restrictive food intake disorder: cognitive-behavioral dor/divisionofresponsibilityinfeeding.php; 2003 Accessed
conceptualization and treatment. Eat Behav 2015;19:106–9. 05.11.16.

Curr Probl PediatrAdolesc Health Care, April 2017 103