Avoidant/Restrictive Food Intake Disorder: Review and Recent Advances

Abstract

Avoidant/restrictive food intake disorder (ARFID) is an eating disorder recently codified in DSM-5 that affects individuals of all ages. A proliferation of ARFID research has emerged over the years, and this review provides a brief overview of the current understanding of ARFID epidemiology, symptoms, comorbid conditions, assessment, and treatment. The review highlights recent research updates regarding ARFID among adults, putative neurobiological mechanisms underlying ARFID, and new treatment trials. Findings from this review demonstrate that ARFID is as prevalent as other eating disorders, even among adults, and is associated with significant medical and psychiatric comorbid conditions. New, promising treatments for children, adolescents, and adults are in the early stages of development. Several assessments are now available to aid in the screening and diagnosis of ARFID and have demonstrated cross-cultural validity. Areas for future research and clinical guidance, including unresolved questions regarding ARFID categorization and differential diagnosis, are discussed.

Avoidant/restrictive food intake disorder (ARFID) is a feeding and eating disorder that was introduced in DSM-5 (1). Individuals with ARFID exhibit avoidant or restrictive eating patterns not driven by fear of weight gain or body image disturbance. Instead, their avoidance or restriction is driven by factors such as a lack of interest in food or eating, sensory sensitivity (e.g., to food textures), or fear of potential consequences (e.g., choking). ARFID may manifest as a significant limitation in the volume (amount) or variety (avoidance of specific foods) of food intake. These behaviors can lead to weight loss, faltering growth, nutritional deficiencies, dependence on enteral or supplemental feeding, and impairment in psychosocial functioning (1).

Despite its relatively recent codification in DSM-5, ARFID is far from new. The disorder encompasses a variety of eating challenges that had previously been recognized as feeding disorder of infancy and early childhood (2), as well as other eating challenges that had no diagnostic home, such as selective eating without weight loss or food avoidance or restriction unaccompanied by body image disturbance. Only in the past decade has a concerted and wide-scale research effort been underway to understand ARFID and all of its manifestations. The eating disorder field has come far in its characterization of ARFID, estimates of its impact, and guidance on how to treat it. Several review articles have been published in recent years that provide excellent summaries of the ARFID literature (3–9). The goal of this review is to provide an overview of the diagnosis to readers—epidemiology, symptoms, comorbid conditions, assessment, and treatment—and highlight relatively recent areas of research advances, including the study of ARFID among adults, early insights into ARFID neurobiology, developments in diagnostic understanding, and new treatment studies.

Epidemiology

Significant progress has been made toward characterizing the epidemiology of ARFID, and yet its epidemiology among adults remains understudied. A recent systematic review of ARFID epidemiology among children and adolescents indicates that the prevalence of ARFID ranges from 0.3% to 15.5% in nonclinical samples and from 5.0% to 64.0% in specialty clinic samples, such as eating disorder, gastroenterology, or feeding disorder services (7). Furthermore, incidence rates for ARFID have now been estimated at 2.02 per 100,000 patients among children ages 5–18 (10). Significantly less data are available for adults. Limited data indicate that the prevalence of ARFID among adults may range from 0.3% among the general population (11) up to 11.0% among samples with eating disorders (12). Data from an online eating disorder screening instrument indicate that 4.7% of adult respondents screened positive for ARFID (13). Generally, research is in agreement that ARFID appears to be similarly common among boys and men as it is among girls and women, that ARFID is more common and its onset is earlier in life (in comparison with other eating disorders), and that it is more common among specialty, treatment-seeking populations (7). Research with primary care providers also indicates that familiarity with the ARFID diagnosis is increasing, suggesting that medical providers are able to more readily identify ARFID cases and make appropriate referrals for treatment (10, 14). Because many of these efforts to improve recognition and identification of ARFID have occurred in pediatric settings, similar efforts should be undertaken with adult medical providers.

Most ARFID epidemiological data, however, has been collected from predominantly White, Eurocentric samples from the United States, Canada, and Europe. Recent years have seen more efforts to study ARFID in other cultural and geographical contexts. ARFID research and assessment validation studies have now been conducted among Spanish-speaking and Arabic populations (6), transgender and nonbinary people (15), Israeli orthodox Jewish communities (16), and Asian countries (17, 18). Research from countries in Asia has estimated the prevalence of ARFID as between 1.3% among Japanese children (17) and 4.8% among Malaysian adults (19). Symptoms of picky eating and lack of interest were more frequently endorsed than fear of aversive consequences among parents of Japanese children (17). Research conducted with Chinese young adults found that their picky eating scores were higher than those of U.S. samples (18). Furthermore, picky eating among Chinese adults was negatively correlated with weight, whereas in U.S. samples, it is generally not (18). Picky eating, or sensitivity to the sensory properties of food, was also endorsed by one-quarter of Mexican adolescents (20). Among U.S. transgender and nonbinary adolescents, approximately one in four scored above clinical cutoffs for the sensory sensitivity and lack-of-interest presentations of ARFID, and patients assigned female at birth endorsed higher ARFID symptoms on average compared with participants assigned male at birth (15). Finally, in a small study of Israeli children, approximately 10% scored above the clinical cutoff on a measure of ARFID (16). These data suggest that ARFID symptoms are present and recognized cross-culturally and among gender-diverse individuals, although symptom prevalence and expression may differ slightly among groups. Preliminary estimates of prevalence among different groups are on par with prevalence estimates from White Northern American and European samples.

Outside of these efforts, specific study of ARFID, its presentation, and its treatment remains lacking among groups traditionally underrepresented in research. It is possible that some of the same problems that plague research on other eating disorders also plague the research on ARFID—namely, that eating disorders in general tend to be underrecognized among people with socioeconomically disadvantaged backgrounds or people of color, and the way culture may influence the expression of eating and related pathology lacks recognition (21). Because ARFID is an even newer eating disorder diagnosis, and less well known to the medical community than more traditional eating disorders, these problems with recognition may be magnified. More research is also needed on ARFID from community medical settings, which also tend to serve a more diverse patient population.

Presentation and Symptoms

Proper recognition of ARFID may also be improved by a better understanding and characterization of the diagnosis. Although the hallmark of the ARFID diagnosis is a restriction of intake volume, avoidance of food variety, or both, the nature of the avoidance or restriction can vary widely among individuals with the disorder. Research over the past decade has documented patterns in the nature of avoidant and restrictive eating and has led to efforts to refine the characterization of ARFID on the basis of underlying motivating factors. Although not recognized formally in DSM-5, most researchers and clinicians seem to agree that three categories of ARFID symptom expression are generally recognizable and replicable across samples: selective eating due to sensory sensitivity, lack of interest in food or eating, and fear of aversive consequences (see Figure 1).

FIGURE 1.

Three avoidant/restrictive food intake disorder presentations

Sensory Sensitivity or Selective Eating

ARFID characterized by sensory sensitivity is defined by an avoidance of specific types of foods because of their sensory features, such as taste, texture, smell, temperature, color, or brand (22–24), and is otherwise described as selective eating. Individuals with this presentation of ARFID are those most likely to be described as picky eaters. People with this presentation of ARFID may eat few total foods and may exclude entire food groups from their diet (e.g., vegetables or fruits). Common symptoms and behaviors observed among people with this presentation also include general anxiety around eating, gagging, heightened disgust sensitivity, and food neophobia (fear of trying new foods) (22, 24–27). Patients with sensory sensitivity are less likely to be underweight or more likely to be within their target weight range when they present to treatment (24, 26). Several studies have also found a higher preponderance of boys and men among this subgroup (26, 28). One possible contributor to the gender disparity could be the higher rate of autism spectrum disorder (ASD) among cisgender boys and a strong association between ASD and selective eating (29–33).

Lack of Interest

Another common presentation of ARFID is lack of interest in food or eating. This presentation is often characterized by poor caloric intake, early satiety, poor hunger cues, and low hedonic response to food (22–24, 34–36). The lack-of-interest presentation has also been associated with gastrointestinal complaints (24, 37–39). Similar to selective eating, this set of symptoms may present early in life, although early studies of ARFID linked some cases of lack of interest or poor appetite to preexisting mood or anxiety disorders (40). This presentation is most strongly associated with the lowest percentage of median body mass index of all ARFID presentations and is the most likely to be associated with faltering growth (height), both potential consequences of early and long-standing poor intake (26). Interestingly, this presentation has also been linked to higher rates of attention-deficit hyperactivity disorder (ADHD) (26). Notably, use of stimulant medication has been linked to poor appetite and the development of ARFID symptoms, although not the lack-of-interest presentation specifically (41).

Fear of Aversive Consequences

Finally, ARFID may also be characterized by an intense fear of the aversive consequences of eating. Commonly reported fears include vomiting, choking, abdominal pain, allergic reactions, and illness. This presentation of ARFID often results in extremely poor intake and may also be associated with avoidance of specific foods associated with the aversive consequence (e.g., avoidance of meat due to fear of choking). This presentation appears to be uniquely characterized by its course; individuals with fear of aversive consequences typically have a more acute onset and shorter duration of illness than those with sensory sensitivity or lack of interest (24, 28). Additionally, one sample of adults with ARFID presenting for treatment at higher levels of care found that fear of aversive consequences was the most common presentation of adult-onset ARFID (42). Individuals with this presentation also commonly report experiencing a triggering or traumatic event before the onset of ARFID, although, in our anecdotal experience, not all cases link back to a single triggering event. This presentation of ARFID has also been associated with a higher rate of co-occurring anxiety disorder, greater preponderance of girls and women, acute weight loss, and higher rates of hospitalization (22, 24, 26). Fear of aversive consequences has also been associated with gastrointestinal complaints and higher rates of gastrointestinal disorders (37–39, 43), possibly leading individuals to develop ARFID out of fear of exacerbating existing abdominal pain or other gastrointestinal symptoms.

Subtypes Versus Dimensional Model

Although DSM-5 mentions that ARFID may be characterized by sensory sensitivity, lack of interest, or fear of aversive consequences, it does not outline criteria or a procedure for qualifying these differing presentations as part of the diagnostic criteria (2). Despite this lack of recognition, though, there appears to be informal consensus in the field that there are meaningful clusters of ARFID symptoms (44). Whether or not these clusters of symptoms constitute discrete diagnostic subtypes, however, is still not clear. A handful of studies have now validated the presence of categorical subtypes while at the same time acknowledging that patients with ARFID often endorse symptoms of more than one subtype (24, 28, 45). Whether these symptom clusters represent subtypes or overlapping profiles, however, the recognition of these clusters has already started to inform the development of treatment approaches. For example, cognitive-behavioral therapy for ARFID (CBT-AR) uses a modular approach to allow providers to deliver differing combinations of evidence-based interventions to target each cluster of symptoms (46, 47).

A critical question remains: Can patients be diagnosed as having more than one “subtype” of ARFID? Some research studies describe a methodology of assigning patients to ARFID subtypes on the basis of whichever symptoms seem to be primary or causing the most impairment (24, 26, 45). Although these same studies have found significant differences between these purported subtypes in terms of patient demographics and symptoms (as noted in the preceding sections), there remain areas of overlap (22, 26). As a result of this overlap, one study proposed a fourth subtype of ARFID, called the combined type, which consists of patients with symptoms of lack of interest and sensory sensitivity (24). This observed overlap has led another team of researchers to propose that ARFID has symptom dimensions rather than subtypes, with each dimension reflecting unique underlying disease mechanisms (23). Much like a person can have more than one physical disease or ailment, each of differing etiology (e.g., a sinus infection and diabetes), a person could also have different types of impairment related to food or eating (e.g., sensitivity to the sensory properties of food and a fear of choking). Studies taking this dimensional approach to characterizing ARFID among child, adolescent, and adult patients confirm that more than half of ARFID patients in their samples endorse symptoms of more than one dimension (28, 48, 49). A consensus on whether to codify a categorical or dimensional model of ARFID has yet to be reached, and the decision will have obvious implications for assessment, etiological research, and treatment.

Etiology

Although significant evidence characterizing the symptoms of ARFID now exists, research exploring the etiology of ARFID remains in its early stages. Researchers just recently published the first genetic data for ARFID and estimated in their sample that ARFID may have a heritability factor of 0.79 (50). Other studies have also linked the development of ARFID to preexisting medical conditions (39, 51–54) (see discussion in the “Medical Comorbid Conditions” section). To date, only one comprehensive, theoretical model of ARFID etiology has been proposed: the three-dimensional neurobiological model of ARFID (23). As mentioned in the previous section, Thomas and colleagues’ (23) model of ARFID theorizes that the different ARFID presentations represent unique, underlying disease mechanisms and that these mechanisms are neurobiological in nature. They propose that each group of symptoms could be mapped onto one or more neurobiological processes: sensory sensitivity driven by disturbances in perceptual processes, fear of aversive consequences driven by disturbances in the negative valence systems, and lack of interest driven by disturbances in the arousal and regulatory systems. Data from early studies into these contributing neurobiological factors are now emerging.

Appetite Regulation

Most etiological data to date link evidence on appetite regulation systems to ARFID. Thomas and colleagues (23) hypothesize that lack-of-interest symptoms, in particular, may be related to hormones that are responsible for signaling hunger and satiety. In support of this hypothesis, research has tentatively linked several hormones involved in appetite regulation to ARFID, including cholecystokinin (CCK), oxytocin, ghrelin, and peptide YY (PYY) (55–57). Differences in anorexigenic hormones, such as CCK, oxytocin, and PYY, have been observed among individuals with ARFID, healthy control individuals, and individuals with anorexia nervosa (AN). In one sample, individuals with ARFID appeared to have significantly elevated levels of fasting CCK and oxytocin compared with healthy control individuals, which may contribute to poor appetite and longer sustained satiety between meals (55, 57). Their PYY levels, however, although not significantly higher in overall concentration than those of individuals with AN and healthy control individuals, appeared to reach peak elevation much earlier after eating (56). These findings could explain early feelings of satiety and difficulty with finishing meals among patients with ARFID. In the same sample, levels of ghrelin, an orexigenic hormone, were lower while fasting and postprandially (after meals) among patients with ARFID compared with healthy control individuals and women with AN, possibly contributing to an overall lower drive to eat—a symptom commonly described among individuals with ARFID with lack of interest (56). These initial findings are among the first links between ARFID and neurobiological processes.

Sensory Perception

Fewer neurobiological data have been published regarding the etiology of selective eating driven by sensory sensitivity and fear of aversive consequences. In their neurobiological model, Thomas and colleagues (23) hypothesized that selective eating characterized by sensory sensitivity may be driven by hypersensitive perceptions of sensory taste input. This hypothesis is supported by previous research linking picky eating (not diagnosed as having ARFID) with heightened sensory sensitivity (both global and taste specific) (27, 29, 30, 58–61) and with increased likelihood of being a “supertaster” (62)—defined both as a person who has a higher density of taste buds on the tongue and as a person who can taste a bitter chemical called 6-N-propylthiouracil (62, 63). More recently, though, attention has been given to other factors that may also explain the persistent and severe food avoidance observed among individuals with sensory sensitivity, including food neophobia, disgust sensitivity, and cognitive rigidity. Several studies have now linked disgust sensitivity with selective eating (25, 64–66), and another study found that cognitive rigidity independently predicted ARFID symptoms in addition to oral textural sensitivity (67). Finally, a case study of an adolescent cisgender boy undergoing cognitive-behavioral treatment for ARFID found that improvements in food neophobia appeared to be an important mechanism of change because they related to new additions to his diet and clinical improvement (44). These data provide new candidates for future etiological studies as well as links to other possible neurobiological mechanisms.

Negative Valence

Finally, Thomas and colleagues (23) hypothesized that fear of aversive consequences may be linked to constructs involving negative valence systems in the brain—or those that are responsible for responding to aversive or threatening stimuli. In particular, they theorized that the development of a specific, fear-based response to food may be the result of hyperactivation of defense motivation systems in the brain involving the amygdala, anterior cingulate cortex, and ventral prefrontal cortex. Little research has been published to date investigating this connection. One study using a functional MRI paradigm that compared viewing of ARFID-specific fear images with viewing of neutral images among people with ARFID and healthy control individuals found no differences between the groups in their response to the fear images (68). More research specific to this presentation of ARFID is needed.

Comorbid Conditions

In addition to advances in characterizing ARFID symptoms and the initial understanding of potential etiological mechanisms, research on ARFID and its relationship to other medical and psychiatric conditions is mounting.

Medical Comorbid Conditions

ARFID is generally associated with higher rates of medical comorbid conditions than are found with eating disorders such as AN and bulimia nervosa (BN) (40, 69). Recently, more research has focused specifically on the intersection of ARFID and disorders of gut-brain interaction (DGBIs) (37, 39, 70). Patients with ARFID are more likely to report a co-occurring DGBI (37, 43), and patients with DGBIs also report elevated rates of co-occurring ARFID (71). Commonly reported DGBIs among patients with ARFID include irritable bowel syndrome, gastroesophageal reflux, eosinophilic esophagitis, gastroparesis, nausea, functional abdominal pain, constipation, and more (24, 37, 72). Researchers hypothesize that the elevated rate of overlap between ARFID and DGBI is likely the result of complex, bidirectional influences (39, 73). DGBIs may not only place a person at risk for developing ARFID, but ARFID may also contribute to the etiology or maintenance of some DGBI symptoms. In particular, exclusion diets are thought to play a crucial role in the link between ARFID and DGBIs. Exclusion diets are commonly prescribed for the treatment of gastrointestinal symptoms and, although acceptable for some, may place a portion of individuals with DGBI at elevated risk for developing ARFID. One study found that patients who had a history of exclusion diets were three times more likely to endorse symptoms of ARFID (51).

Other medical conditions commonly associated with ARFID include food allergies and sensitivities. A recent meta-analysis linked immunoglobulin E (IgE)–mediated food allergies with an increased risk of being diagnosed as having any eating disorder (52). In this study, prevalence rates of ARFID were reported to be 62.9% in a sample of children with IgE-mediated food allergies (53) and between 4.5% and 51.0% in samples of patients with eosinophilic esophagitis (52). Furthermore, among the sample of children with IgE-mediated food allergies, children who met criteria for ARFID were more likely to have two or more suspected or reported food allergies (54). A separate study conducted at an adult immunology clinic found that 79% of 102 study participants endorsed food avoidance or restriction consistent with ARFID (74). Similar to the connection between ARFID and DGBIs, individuals with food allergies often have to follow a restricted diet and are likely to have had aversive experiences with specific foods. Thus, individuals with food allergies may be at particular risk for ARFID, possibly as the result of aversive conditioning or necessary hypervigilance related to their intake to prevent allergen exposure.

Psychiatric Comorbid Conditions

ARFID also exhibits associations with psychiatric disorders (7, 75). The prevalence of current psychiatric comorbid conditions among individuals with ARFID ranges from 45% in outpatient settings (31) to 95% in a day hospital setting (76). Additionally, up to 53% of ARFID patients experience lifetime psychiatric comorbid conditions (31). Anxiety disorders emerge as the most prevalent co-occurring psychiatric condition (35, 43, 49, 77), followed by mood disorders (35, 43, 48). Among a small sample of adult patients with ARFID, roughly 40% also had a diagnosis of posttraumatic stress disorder (49). Neurodevelopmental conditions, such as ASD, are frequently noted for their co-occurrence with ARFID among both child and adolescent as well as adult samples (49); however, estimates of comorbid conditions range widely, from 8.2% to 54.75% (7). Rates of co-occurring ARFID and ASD seem to be higher among samples of individuals with feeding and developmental disorders and lower among general pediatric samples and samples with eating disorders. Specific ARFID presentations have also been uniquely linked to comorbid conditions, with sensory sensitivity correlating with the presence of neurodevelopmental, disruptive, conduct, mood, and anxiety disorders (31). Fear of aversive consequences is associated with an increased likelihood of meeting criteria for anxiety disorders (31).

Course of Illness and Outcome

Medical Complications

If left untreated, ARFID can lead to numerous serious medical complications, most notably malnutrition and nutritional disease. Two studies confirm that patients with ARFID consume lower-quality diets in terms of macro- and micronutrient intake compared with individuals without ARFID (78, 79). Other studies confirm that patients with ARFID may be especially susceptible to problems with poor bone mineral density, a likely consequence of nutritionally insufficient diets and hormone abnormalities (55, 80, 81). Patients with an acute onset of ARFID are most at risk for experiencing acute medical complications that require hospitalization (48). These patients are most likely to endorse the fear-of-aversive-consequences presentation of ARFID (26). Medical complications observed among patients requiring hospitalization include bradycardia, hypotension, prolonged QT interval on electrocardiogram, amenorrhea, anemia, electrolyte imbalances, and elevated liver enzymes (69, 72, 82). However, on average, individuals with ARFID experience fewer effects of acute starvation, such as weight loss and bradycardia, in comparison with their counterparts with AN (55, 83, 84). Other patients with ARFID may not require hospitalization; rather, they may experience chronic low body weight and poor growth, often dating back to early childhood (26).

Although most research on the medical sequelae of ARFID has been conducted with child and adolescent patients, recent efforts have been made to study the medical outcomes of ARFID among adults (43). Among a sample of 122 adult patients (average age=32.7 years) with ARFID admitted for medical stabilization, common concerns included malnutrition or nutritional deficiency (87%), significant weight loss (65%), and dependency on enteral feeding (23%) (85). Other medical concerns included low serum prealbumin and vitamin D, hyperkalemia, signs of immunosuppression (e.g., leukopenia), and elevated serum bicarbonate (85). Almost 40% of patients reported gastrointestinal complaints or concerns (85). Similar to samples of children and adolescents, the most common presentation of admitted adult patients with ARFID was fear of aversive consequences. Data are still lacking, however, on how hospitalization rates among adults compare with hospitalization rates among children and adolescents, as well as how hospitalization rates and medical complications may differ by presentation in adults.

Illness Course

Research on the longitudinal course and duration of illness in ARFID remains limited. Several studies and clinical consensus support that individuals with ARFID typically exhibit an earlier age of onset and a longer duration of illness before first presenting to treatment compared with those with AN (40, 84, 86). Long-term outcome studies for ARFID patients are particularly lacking. Postdischarge data from hospitals reveal that, at a 1-year follow-up, 62% of patients with ARFID met remission criteria, higher than the rate for patients with AN (82). In contrast, researchers found that at 9-month and 18-month follow-up, the probability of meeting criteria for ARFID was 0.84 in an adolescent and young adult sample, whereas the probability for meeting criteria for AN was 0.62 (87). In another study, after a 2-year follow-up of adolescent patients completing a partial hospital program for eating disorders, 70% of individuals with ARFID were still receiving outpatient services, compared with 81% of those with AN (76). Finally, an extended follow-up at 15.9 years after hospital discharge revealed that 26.3% of patients still met criteria for ARFID, compared with 21.6% who still met the criteria for AN (88). Taken together, the data seem to suggest that ARFID follows a stable course once established and that individuals with ARFID may experience a prolonged duration of illness, similar to those with AN. Because most of the long-term outcome data are based on treatment studies with adolescents, future studies should examine long-term outcomes for patients with ARFID presenting to treatment in adulthood.

Assessment and Diagnosis

As with other eating disorders, ARFID assessment should be multimodal and involve a physical examination and medical workup, psychosocial evaluation with standardized assessment tools, and a nutritional evaluation (44). Recent reviews have provided excellent descriptions of the ARFID assessment process (75, 89). To summarize, particular care should be taken in the evaluation of individuals with ARFID to assess acute health issues, growth and pubertal development, medical stability, gastrointestinal functioning, and signs and symptoms of malnutrition. Rule-outs and considerations for additional evaluation may include screening for food allergies, sensory processing challenges, or signs of oral-motor skill dysfunction (in very young children). The presence of co-occurring psychiatric or neurodevelopmental disorders should also be considered. Distinguishing ARFID from other anxiety disorders, such as specific phobia or obsessive-compulsive disorder, may be challenging. Readers are encouraged to reference the ARFID diagnostic guidance from DSM-5, which notes that ARFID should be diagnosed when the primary target of treatment will be the eating problem (2). Dietary recalls and blood work may be useful in determining the presence of or risk for nutritional deficiencies and can help with assessing dietary adequacy. Several standardized and well-validated assessment instruments have been developed that can aid in diagnostic decision making and tracking of treatment response. These assessments are reviewed in Table 1 (90–101).

TABLE 1.

Summary of measures that assess avoidant/restrictive food intake disorder^a

Measure name	Measure type	Ages	Uses	Symptom profile	Features	Languages	Availability
Pica, ARFID, and Rumination Disorder Interview (90)	Structured interview	≥2 years	Diagnosis, symptom change	Yes	Both self- and parent-report versions; has severity ratings, subtype indices, and clinical impairment index	English, Dutch, Norwegian	https://mccaed.slam.nhs.uk/professionals/resources/featured-resources/
Eating Disorder Examination, ARFID Supplement (91, 92)	Structured interview	0–17 years	Diagnosis	Yes	Both self- and parent-report versions; note that symptom profile also includes food avoidance emotional disorder; adult validation is pending	English, German	For access, contact ricarda.schmidt@medizin.uni-leipzig.de
Eating Disorder Assessment for DSM-5 (93)	Semistructured interview	≥8 years	Diagnosis	No	Ability to assess for and rule out other ED diagnoses. Assesses all ED diagnoses, which can help rule out any eating disorder other than ARFID	English, Spanish, Arabic, Dutch, Portuguese, Norwegian, and Turkish	https://eda5.org/
Pica, ARFID, and Rumination Disorder Interview-Questionnaire (94)	Self-report questionnaire	≥4 years	Diagnosis, screening, symptom change	Yes	Both self- and parent-report versions; has a severity ratings, subtype indices, and clinical impairment index	English, Swedish, Dutch	https://mccaed.slam.nhs.uk/professionals/resources/featured-resources/
Nine-Item ARFID Screen (95)	Self-report questionnaire; screening tool	8–65 years	Screening	Yes	Includes an ARFID total score and 3 subscale scores with suggested clinical cutoff (96)	English, Spanish (6), Chinese (18), Arabic (97), Polish (98), Turkish (99)	Full measure available in original publication
Eating Disorders in Youth Questionnaire (100, 101)	Self-report questionnaire; screening tool	≥8 years	Screening	Yes	Assesses for weight and shape concerns, pica and rumination disorder; note that symptom profile includes dysphagia (fear of aversive consequences), selective eating (sensory sensitivity), and food avoidance emotional disorder (lack of interest)	English, German	https://ul.qucosa.de/api/qucosa%3A14486/attachment/ATT-0/

^a ARFID, avoidant/restrictive food intake disorder; ED, eating disorder.

One particular area of question when it comes to ARFID assessment is differentiating ARFID from other eating disorders. Although DSM-5 has always allowed for the diagnosis of ARFID alongside binge-eating disorder, it states that ARFID cannot occur exclusively during the course of AN or BN and that body image disturbance cannot be present. Anecdotally, clinicians have found differentiating ARFID from AN to be especially challenging, particularly in the case of adults with restrictive eating and weight loss who deny fear of fatness or a desire to lose weight (102–104). Furthermore, questions have been raised as to whether ARFID can be diagnosed concurrently with AN and BN, despite the exclusion criteria set forth in DSM-5. Becker and colleagues (86) wrote a report describing two cases in which adolescent girls met all diagnostic criteria for ARFID and simultaneously endorsed or subsequently developed more traditional eating disorder pathology. They noted that these two cases were emblematic of other cases of purported diagnostic overlap observed in their larger clinic sample. These authors suggest that although ARFID is a distinct and unique eating disorder diagnosis, co-occurrence with other eating disorders could happen despite prohibition by DSM-5.

A small handful of studies have provided limited data to guide questions of differential diagnosis between ARFID and other eating disorders. One study, conducted in a tertiary care setting, identified six cases of adolescent eating disorders among girls that were initially diagnosed as having ARFID and later reclassified as having AN (105). Themes that emerged among cases that were later reclassified included vegetarianism, generalized health anxieties, and the emergence of body image disturbance with weight restoration. A study comparing groups of adult women with ARFID with adult women with non–fat-phobic AN (cases of AN without an overt fear of weight gain or fatness) found that implicit, positive attitudes toward dieting and thinness successfully differentiated the two groups (106). Another case study highlighted a patient with ARFID who was experiencing severe medical complications and presented with restriction that was consistent with sensory sensitivity and fear of aversive consequences; however, this individual also reported a desire to be thin (107). The authors concluded, though, that her concerns about weight and shape were not clinically significant and, thus, that individuals with ARFID may present with body image concerns as long as the body image concerns are not driving the eating disorder behaviors. Moreover, as mentioned earlier in this article, a longitudinal study of girls and women found no diagnostic crossover for ARFID over an 18-month period (87). Taken together, these data suggest that ARFID is distinct from other eating disorders. More research is needed to develop guidance for how to better distinguish ARFID from other eating disorders in clinical practice and to reach a consensus on the co-occurring diagnosis of ARFID and other body image–related eating disorders.

Treatment

Prior reviews and expert consensus all indicate that ARFID treatment should be multimodal in nature and should take into consideration the individual presentation and symptoms, developmental stage and neurodevelopmental profile, severity and risk factors, maintaining factors, and goals of treatment, among other things (3, 4, 9, 15, 44, 108). ARFID treatment may occur in a variety of contexts, including both medical and psychiatric as well as inpatient, partial hospital, and outpatient levels of care. ARFID treatment is available through traditional eating disorder clinics and programs as well as in outpatient and hospital-based feeding disorder programs. Treatment may involve psychological or behavioral therapy with a psychologist or other qualified mental health professional, complementary management of functional gastrointestinal symptoms with a gastroenterologist, nutritional therapy with a dietitian, psychopharmacology with a psychiatrist, or oral-motor skill or sensory-based therapy with an occupational therapist or speech-language pathologist. Goals of treatment may include weight restoration, expansion of dietary variety, reduced anxiety, improved volume tolerance, increased oral intake, increased solid food intake, and improved psychosocial functioning. Goals that affect weight and nutritional status should be prioritized first in treatment. Recent data also suggest that treatments may be more effective when tailored to the individual patient’s unique profile of ARFID symptoms (46), as in CBT-AR. Outpatient psychological treatments are currently considered the front-line treatment for ARFID and are summarized in Table 2 (47, 109–116).

TABLE 2.

Psychological treatments for ARFID^a

Treatment	Ages	Brief description	Brief summary of outcomes	Current trials
Family-based treatment for ARFID (FBT-ARFID) (109)	5–12	Modified from FBT for AN; treatment empowers parents to take charge of child’s ARFID; 3 phases focusing on parental management of ARFID, transition to independent or age-appropriate management of ARFID, and finally a return to developmental norms; sessions structured around weight change and dietary expansion using the always-sometimes-never list	Randomized controlled trial (110). N=16 (FBT-ARFID); N=12 (treatment as usual); patients in FBT-ARFID gained more weight than patients in treatment as usual (d=−.87); FBT-ARFID also produced greater decreases in ARFID symptom severity ratings (d=.83)	Lock: R01MH121292
Cognitive-behavioral therapy for ARFID (47)	≥10	4-stage treatment delivered in either individual or family-support format; stages consist of psychoeducation and early change, treatment planning, maintaining mechanisms, and relapse prevention; the maintaining mechanisms stage includes modules to treat sensory sensitivity, fear of aversive consequences, and lack of interest	Two proof-of-concept studies with adolescents and adults. Adolescents (N=17); on average, patients added 16.7 new foods, gained 11.5 lb, and experienced a reduction in ARFID severity symptoms; 70% no longer met criteria for diagnosis at end of treatment (116). Adults (N=15); on average, patients added 18 new foods, gained 11.38 lb, and experienced a reduction in ARFID severity symptoms; 47% no longer met criteria for diagnosis at end of treatment (111)	Lawson: R61MH129331
Brief CBT for ARFID with DGBI	≥18	8-session exposure-based protocol specifically for adults with ARFID with co-occurring DGBI	2 part pilot study of patients who had previously completed or prospectively completed CBT for ARFID + DBGI (N=14); qualitative feedback indicated that treatment was acceptable and improved symptoms and quality of life (112)	Burton Murray: K23DK131334H
Supportive Parenting for Anxious Childhood Emotions, ARFID	6–14	12-session parent-focused intervention focused on modifying parental accommodation of and responses to eating behaviors	Pilot trial with children (N=15) who were not underweight; significant reductions in ARFID symptom severity as measured by the NIAS (g=1.45) and in parental accommodation (g=1.14) (113)	N/A
Feelings and Body Investigators, ARFID Division	4–10	15-session, playful child and parent treatment that uses interoceptive exposure to increase awareness of and reframe responses to internal body sensations and sensory experiences	Case study of a 4-year-old girl with poor appetite and percutaneous endoscopic gastrostomy tube; food intake increased throughout treatment, and patient no longer met diagnostic criteria at end of treatment (114)	Zucker: R33MH121549
Picky Eaters Clinic	4–11	7-session, parent-only group-based intervention to train parents in parent management and exposure techniques to manage challenging mealtime behavior and increasing dietary variety	Pilot trial with children (N=21) presenting primarily with sensory sensitivity; significant reductions in problematic mealtime behaviors, increase in food enjoyment, decrease in food fussiness at end of treatment; treatment gains maintained at 3-month follow-up (115)	N/A

^a ARFID, avoidant/restrictive food intake disorder; CBT, cognitive-behavioral therapy; DGBI, disorders of gut-brain interaction; FBT for AN, family-based treatment for anorexia nervosa; N/A, not applicable; NIAS, Nine-Item ARFID Screen.

Inpatient Treatment and Higher Levels of Care

One recent area of advance has been in guidance on treating ARFID in hospital and partial hospital settings. Given the nutritional deficiencies and medical comorbid conditions that often accompany ARFID, a significant portion of patients with ARFID require hospitalization for medical stabilization or enteral feed weaning, similar to patients with AN. Studies have found that in contrast to individuals with AN receiving inpatient treatment, patients with ARFID tend to exhibit slower weight gain (72, 117), greater reliance on enteral supplements (82), and longer hospital stays (82). These differences may arise because hospital refeeding protocols, designed to interrupt eating disorder behaviors and promote weight restoration, were originally tailored for individuals with AN and, thus, may not be well suited for individuals with ARFID. Partial hospital treatment programs also present similar challenges for the treatment of ARFID (118, 119). Consequently, there is a recent call to adapt protocols at higher levels of care to meet the unique needs of patients with ARFID (118, 120). Recommendations across both inpatient and partial hospital settings call for flexibility around or changes related to traditional protocols specific to mealtimes. Some of these protocol modifications for ARFID include allowing patients and parents to be involved in collaboratively selecting menu items, strategically adapting the presentation of food and the meal environment, increased flexibility around postmeal supplementation, and no need for postmeal supervision or monitoring, among others (118, 120). In case studies that have used adjusted protocols, there is an increase in oral intake, a decrease in reliance on external or supplemental feeding, and an increase in patient and caregiver satisfaction (120).

Pharmacology

Numerous medications have been studied as a way to augment psychological treatment by improving the management of ARFID symptoms and sequelae. These medications include cyproheptadine for appetite stimulation (66), olanzapine for appetite and weight gain (121, 122), mirtazapine for visceral hypersensitivity and weight gain (123–126), and D-cycloserine for food aversions (127). Other studies have also looked at fluoxetine, buspirone, and lorazepam to target comorbid or food-specific anxiety and low mood in the context of ARFID (75). Notably, one recent, large case series examined the combined use of hydroxyzine and selective serotonin reuptake inhibitors (SSRIs) in the treatment of ARFID among child and adolescent patients and found that the combination of hydroxyzine and SSRI was similarly effective as the use of SSRI alone (128). However, patients in the hydroxyzine and SSRI group reported significantly greater depressive symptoms and pre- and postmeal anxiety, suggesting that the combination of an SSRI plus an anxiolytic may be more effective for more severe cases of ARFID.

Stimulant medications, traditionally prescribed for ADHD, have received some attention due to their negative impact on appetite. One study noted that in two cases, stimulant medication exacerbated long-standing avoidant/restrictive behavior and resulted in precipitous weight loss, leading to hospitalization (41). These cases highlighted the risks of prescribing stimulant medication among patients with ARFID. Another more recent, larger case series of pediatric patients diagnosed as having co-occurring ARFID and ADHD found that ARFID could be effectively treated in the context of stimulant medication use (129). In this case series, all patients received psychological therapies for ARFID that included a significant behavior modification component as well as prescription of additional medications to counteract the appetite-suppressing properties of the stimulant (129). The authors concluded that maintaining stimulant medication during treatment for ARFID could be done safely in the context of complementary therapies. The authors also hypothesized that use of stimulant medication may even be necessary because it may aid patients with executive functioning deficits in sustaining goal-directed behavior during meals, an often unpleasant or onerous task for ARFID patients.

Feeding Therapy

Finally, a review of ARFID treatment would not be complete without mentioning the role of more traditional feeding therapies. Although there is debate regarding the diagnostic distinction between ARFID and feeding disorder (130–132), many patients with ARFID, particularly patients who are early elementary school age or younger or diagnosed as having developmental disorders, are treated in the context of feeding disorder clinics. Feeding disorder clinics may be inpatient or outpatient and typically use a combination of behavior modification procedures along with nutritional therapy, medical management, and oral-motor skills–based therapies as needed. Behavioral interventions using behavior modification techniques are effective in the treatment of a wide array of childhood feeding difficulties, such as food aversions and sensory-based food avoidance, and have been studied in the context of typically developing children, children with severe co-occurring medical conditions, and children with ASD (133). It should be noted, however, that these approaches have not been evaluated among children in higher elementary school grades and older.

Conclusions

The past decade has seen a proliferation of research that has advanced the understanding of ARFID and its treatment. More recently, research has accumulated on ARFID among adults, on the presentation of ARFID cross-culturally, and on the underlying mechanisms that may be implicated in its development. ARFID is a common, widespread eating disorder that appears to be just as prevalent across all ages as other, more notable eating disorders. It also affects a higher proportion of boys and men than do more “traditional” eating disorders. ARFID can have a significant medical and psychosocial impact on individuals and appears to follow a stable course, thus necessitating early identification and treatment. Providers working with specialty populations, such as developmental pediatrics, gastroenterology, and allergy and immunology may need to be particularly aware of the high risk for ARFID among these individuals. Treatment options are available for persons of all ages, and a multimodal treatment approach tailored to individual presentation is recommended. Future research is needed to further preliminary investigations of underlying disease mechanisms in ARFID. This research will continue to help researchers create more targeted treatments with improved efficacy.