The diagnosis of avoidant restrictive food intake disorder in the presence of gastrointestinal disorders: Opportunities to define shared mechanisms of symptom expression

Abstract

Objective:

Individuals with a gastrointestinal (GI) disorder often alter their diet to manage GI symptoms, adding complexity to understanding the diverse motivations contributing to food avoidance/restriction. When a GI disorder is present, the DSM-5 states that Avoidant/Restrictive Food Intake Disorder (ARFID) can be diagnosed only when eating disturbance exceeds that expected. There is limited guidance to make this determination. This study attempts to address this gap by characterizing the presentation of ARFID in adults with and without a self-reported GI disorder.

Method:

Participants were 2,610 adults ages 18–44 who self-identified as “picky eaters.” Participants reported on motivations for food avoidance, affective experiences towards food, and perceived impairment. Responses were compared across four groups: GI issues and likely ARFID (L-ARFID/GI), L-ARFID-only, GI-only, and No-ARFID/No-GI.

Results:

Groups with a GI disorder (L-ARFID/GI, GI-only) reported more fear of aversive consequences of eating than those without a GI disorder, while groups with L-ARFID (L-ARFID, L-ARFID/GI) evidenced significantly greater sensory aversion to food and indifference to food or eating, negative emotional reactions to food and overall disgust sensitivity, and eating related impairment.

Discussion:

Consideration of the interplay of a GI disorder with ARFID can add precision to case conceptualization. Food avoidance may be attempts to manage fears of aversive consequences that are augmented by a history of GI symptoms, while sensory aversions and negative emotional reactions towards foods may be more elevated in ARFID. These findings emphasize the need to consider an ARFID diagnosis in patients with GI disorders to optimize care.

1 |. INTRODUCTION

Individuals diagnosed with a gastrointestinal (GI) disorder often receive recommendations to alter their diet to improve GI symptoms. There is also evidence that individuals with GI disorders engage in unique patterns of dietary manipulation as they formulate personal hypotheses about how types or amounts of food increase or decrease GI symptoms and alleviate GI distress (Reed-Knight, Squires, Chitkara, & van Tilburg, 2016; van Tilburg & Felix, 2013). Individuals with GI disorders commonly avoid dairy, carbohydrates, and other specific foods they suspect cause allergic responses or intolerance (Huertas-Ceballos, Logan, Bennett, & Macarthur, 2008; Kokkonen, Haapalahti, Tikkanen, Karttunen, & Savilahti, 2004; Monsbakken, Vandvik, & Farup, 2006; Nanda, James, Smith, Dudley, & Jewell, 1989; Shepherd & Gibson, 2006; Simrén et al., 2001; van Tilburg & Felix, 2013). In a study of adolescents with irritable bowel syndrome (IBS; van Tilburg et al., 2012), 92% reported symptoms associated with eating, of which 95% attributed symptoms to specific foods. To prevent symptoms, adolescents reported avoiding the offending foods, not eating any food even when hungry, and vomiting after eating (van Tilburg et al., 2012). In fact, dietary manipulations are so common in response to GI symptoms that 32% of young children (5 to 10 years old) provided a dietary or hydration strategy when asked for ideas on how to cope with abdominal (“tummy”) pain (Ives et al., 2021). Indeed, Zucker and Bulik (2020) hypothesized that GI symptoms and disorders may be risk and/or maintenance factors for restrictive eating disorders such as anorexia nervosa (AN) or Avoidant Restrictive Food Intake Disorder (ARFID).

Alterations in eating behavior to manage GI symptoms complicate how eating disorders and GI disorders are distinguished and diagnosed: when do efforts to manage GI symptoms become maladaptive and reflect disordered eating? In the case of ARFID, diagnosis requires that an individual's eating or feeding disturbance does not result from another physical or mental condition (American Psychiatric Association, 2013). To evaluate whether an individual with a GI disorder meets the diagnostic criteria for ARFID, it is necessary to determine whether the symptoms of eating or feeding disturbance are beyond those expected to occur with a GI diagnosis. Global population data suggest that approximately 13.5% of children and adolescents (Korterink, Diederen, Benninga, & Tabbers, 2015) and more than 40% of adults worldwide have disorders of gut-brain interaction, which are characterized by impairing GI symptoms without a precise defining biological marker and that reflect, instead, complex interactions of gut physiology and central nervous system processing (Sperber et al., 2020). Additionally, national data suggest that 1 in 4 toddlers, children, and adolescents have a disorder of gut-brain interaction (Lewis, Palsson, Whitehead, & van Tilburg, 2016; Robin et al., 2018; van Tilburg et al., 2015). Thus, if clinicians attribute eating difficulties to GI disorders and do not assess these patients for ARFID, they overlook a significant number of individuals who may benefit from eating-specific interventions.

Further, the presence of ARFID may complicate the management of GI symptoms/disorders. Prior research suggests that some individuals with ARFID who are characterized by limited food variety may tend to avoid fiber-rich foods, and that refined carbohydrates constitute a large portion of their diets. In a study comparing nutritional intake between adults with full or subthreshold ARFID and healthy controls, Harshman et al. (2019) found that individuals with ARFID aged 9–22 years consumed 68% the amount of fruits and 53% the amount of vegetables consumed by healthy controls. Further, extreme food avoidance is known to contribute to problems with GI function such as delayed gastric emptying, a state that may increase uncomfortable sensations of bloating and fullness (Murray et al., 2020a; Murray, Jehangir, Silvernale, Kuo, & Parkman, 2020b). Thus, individuals with ARFID may exacerbate GI symptoms if they avoid techniques thought to minimize GI distress (e.g., recommended dietary manipulations), while decreased food quantity may further exacerbate GI symptoms.

Aside from signaling GI distress, gut sensations are also a constituent of emotional experience. “Butterflies” in one's stomach when feeling nervous or nausea in response to an offensive odor are examples of somatic sensations that comprise an emotional experience and help an individual to contextualize, label, and act upon various motivated states, including that of emotion (Kreibig, 2010; Nummenmaa, Glerean, Hari, & Hietanen, 2014). In fact, the sensitivity with which one can accurately sense an interoceptive sensation such as a heartbeat has been associated with the subjective perceived intensity of the emotion and the intensity of activation in response to emotional stimuli as measured by EEG (Herbert, Pollatos, & Schandry, 2007). Individuals with GI disorders display heightened sensitivity to gut sensations (Farmer & Aziz, 2013), suggesting that having a GI disorder may heighten an individual's experience of emotions constituted, in part, of gut sensations, such as anxiety and disgust (Zucker & Bulik, 2020). Given that anxiety and disgust both contribute to food avoidance in those with ARFID (Harris et al., 2019), heightened experience of or sensitivity to gut sensations associated with anxiety or disgust might worsen food avoidance in individuals with ARFID and a co-occurring GI disorder, in part by increasing emotional reactions to food.

GI symptoms, such as pain, are unpredictable and threatening – and therefore, potentially scary. According to the fear-avoidance model of pain, such fear can generalize to innocuous sensations that are associated with pain or that predict pain and contribute to the avoidance of experiences or activities that may produce those sensations (Crombez, Eccleston, Van Damme, Vlaeyen, & Karoly, 2012; Vlaeyen & Linton, 2012). Eating certain types or amounts of food are examples of experiences that may be avoided due to fear that they will produce aversive, unwanted outcomes such as pain. In those with ARFID and a GI disorder, the breadth of potential feared or aversive consequences may be increased. Repeated experiences of aversive GI symptoms may combine with other feared outcomes (e.g., choking, vomiting, GI pain) to intensify avoidance behaviors. Thus, a fear of aversive consequences may be more pronounced in those with a comorbid GI disorder and ARFID, a finding reported in 92.8% of patients with ARFID symptoms seen in a neurogastroenterology clinic (Murray, Bailey, et al., 2020a).

In the current study, we sought to characterize the interplay of ARFID in the presence or absence of a GI disorder. Specifically, we aimed to investigate whether motivations for food avoidance (sensory aversions, fear of aversive consequences, lack of interest in food/eating); negative emotional reactions to food (e.g., anxiety, disgust); and eating-related impairment varied among people who have either likely ARFID (L-ARFID), self-reported GI disorder or symptoms, or both. We compared four groups: adults with both L-ARFID and a self-reported GI disorder or symptoms (L-ARFID/GI), adults with L-ARFID only (L-ARFID-only), adults with a self-reported GI disorder or symptoms only (GI-only), and adults with neither L-ARFID nor a GI disorder or symptoms (No-GI/No-ARFID). As all participants self-identified as “picky eaters,” comparisons between the L-ARFID and No-ARFID groups are conservative in that our control group (No-GI/No-ARFID) reported elevated food avoidance but did not endorse associated impairment. Due to the small amount of research on the presentation of ARFID in individuals with GI disorders, the current study was exploratory in nature and aimed to generate hypotheses for future research (Jebb, Parrigon, & Woo, 2017).

2 |. METHOD

2.1 |. Overview

Adults who self-identified as “picky eaters” completed an online survey that included questions on demographics, disordered eating behaviors, emotional reactions to food, sensory sensitivity to food, food avoidance, and impairment due to eating difficulties. Individuals were not offered compensation for participating in the study, though participants could provide contact information if they wished to be contacted about the results of the study or future opportunities to participate in research. The study was approved by the Duke University Health System Institutional Review Board.

2.2 |. Participants

Adults self-identifying as “picky eaters” were recruited via articles on adult picky eating published in mass media outlets and on a southeastern medical center website. The survey opened on October 25, 2012. Individuals opted into the study by clicking a link, completing an electronic consent process, and responding to the survey questions. Inclusion required that individuals be 18 to 44 years old, have completed the questions required to diagnose AN, bulimia nervosa (BN), binge eating disorder (BED), and L-ARFID, and not meet criteria for full or subthreshold AN, BN, or BED. Individuals 45 and older were excluded due to higher likelihood of experiencing GI concerns (Dumic et al., 2019).

At data extraction (June 10, 2018), 3,954 individuals had completed the survey (see Figure 1). Of these, 771 individuals were excluded for being outside of the study age range. An additional 567 participants were excluded for meeting criteria for full or subthreshold AN, BN, or BED, or for having missing data that rendered diagnosis with the EDDS impossible (more detail in Assessments). Finally, six participants were excluded due to missing data that prevented diagnosis with L-ARFID. Analyses were conducted with the remaining 2,610 participants. Demographics are provided in Table 1.

FIGURE 1.

Flowchart of participants. At time of data extraction, 3,954 individuals completed the survey. Of these, 771 individuals were excluded because they were outside of the study age range (younger than 18, n = 22; 45 or older, n = 749). Of the remaining 3,183 individuals, 567 were excluded because they met criteria for an eating disorder diagnosis as assessed by the EDDS (AN, n = 9; Sub-AN, n = 59; BN, n = 159; Sub-BN, n = 93; BED, n = 134; Sub-BED, n = 64) or because they were missing data required for EDDS diagnosis (n = 49). Of the remaining 2,616 individuals, 6 were excluded for missing data required for ARFID diagnosis. The remaining 2,610 individuals were included in data analysis

TABLE 1.

Participant demographics by ARFID/GI group

	L-ARFID/GI group
	L-ARFID/GI (n = 300)	L-ARFID only (n = 1,420)	GI only (n = 142)	No-ARFID/No-GI (n = 748)	Total (N = 2,610)
Age
18–24	103 (34%)	699 (49%)	38 (27%)	290 (39%)	1,130 (43%)
25–34	115 (38%)	510 (36%)	54 (38%)	300 (40%)	979 (38%)
35–44	82 (27%)	211 (15%)	50 (35%)	158 (21%)	501 (19%)
Sex
Male	87 (29%)	346 (24%)	31 (22%)	170 (23%)	634 (24%)
Female	209 (70%)	1,039 (73%)	108 (76%)	565 (76%)	1921 (74%)
Not indicated	4 (1%)	35 (2%)	3 (2%)	13 (2%)	55 (2%)
Race/ethnicity
White	270 (90%)	1,258 (89%)	127 (90%)	659 (88%)	2,314 (89%)
Black	9 (3%)	29 (2%)	6 (4%)	17 (2%)	61 (2%)
Asian	2 (1%)	20 (1%)	2 (1%)	16 (2%)	40 (2%)
Native American	2 (1%)	8 (1%)	1 (1%)	5 (1%)	16 (1%)
Pacific Islander	0 (0%)	1 (0%)	1 (1%)	0 (0%)	2 (0%)
Hispanic	9 (3%)	66 (5%)	1 (1%)	30 (4%)	106 (4%)
Other	8 (3%)	36 (3%)	4 (3%)	18 (2%)	66 (3%)
Not indicated	0 (0%)	2 (0%)	0 (0%)	3 (0%)	5 (0%)

Note: Demographics are shown for N = 2,610 participants included in data analysis. Percentages of each of the four ARFID/GI groups in each demographic category are shown.

Abbreviations: GI, gastrointestinal; L-ARFID, Likely Avoidant Restrictive Food Intake Disorder.

2.3 |. Assessments

2.3.1 |. Eating Disorder Diagnostic Scale (EDDS)

The Eating Disorder Diagnostic Scale (EDDS; Stice, Telch, & Rizvi, 2000) was used to determine whether participants met diagnostic criteria for full or subthreshold AN, BN, or BED. The EDDS has demonstrated excellent concurrent validity with interview-based diagnoses for AN, BN, and BED (Stice, Fisher, & Martinez, 2004). Although Stice et al. (2000) recommend that all respondents with low BMI receive the diagnosis of AN, study participants were required to also indicate that they had a definite fear of weight gain of at least a moderate level in order to receive a diagnosis of full or sub-threshold AN. This additional criterion was implemented to prevent individuals with L-ARFID with BMI < 18.5 from being excluded.

2.3.2 |. GI disorder or symptoms

Participants reported whether they experienced IBS, reflux, or another GI problem. Individuals who indicated that they had a GI problem other than IBS or reflux could provide text responses to describe their GI disorder/symptoms. Text responses were coded as GI disorders/symptoms (e.g., “gastritis”) or not GI disorder/symptoms (e.g., “migraine”) based on GI disorders and symptoms listed in Essentials of Rubin's Pathology, Sixth Edition (Rubin & Reisner, 2014) and the Rome Foundation's Rome IV: Functional Gastrointestinal Disorders, Fourth Edition (Drossman et al., 2016). Because we did not have clear conceptual hypotheses related to differences between functional and structural GI disorders, all GI disorders were grouped together.

2.3.3 |. ARFID survey

Participants completed investigator-designed questions to determine L-ARFID diagnosis, motivations for food avoidance, emotional reactions to food, severity of food avoidance, and eating-related social impairment. Participants were instructed to respond to these questions based on their current experiences. Survey questions and item responses are detailed in Table 2.

TABLE 2.

Survey questions assessing L-ARFID, motivations for food avoidance, affective responses to food, severity of food avoidance, and social impairment

Construct	Survey question	Response options
L-ARFID	Have your eating problems led to any of the following?	1 = Yes 0 = No
Weight loss	...Significant weight loss.
Nutritional deficiency	...Significant nutritional deficiency.
Interference with job functioning	...Interference with your job functioning.
Interference with relationships	...Interference with your relationships.
Motivations for food avoidance
Sensory experiences of food
Sensitivity to food smells	Does sensitivity to food smells...	1 = Rarely or never 2 = Less than half the time 3 = About half the time 4 = More than half the time 5 = All of the time
Variety of foods	...keep you from eating a variety of foods?
Social gatherings	...keep you from participating in social gatherings with food?
Sensitivity to food textures	Does sensitivity to food textures...
Variety of foods	...keep you from eating a variety of foods?
Social gatherings	...keep you from participating in social gatherings with food?
Fear of negative consequences of eating
Number of aversive events (Count of “Yes” responses to the items below)	Some people have trouble eating a variety of foods because they have had some traumatic experience with food. Have you had any of the following experiences with food?	1 = Yes 0 = No
Choking	Choking
Seeing someone choke	Seeing someone else choke
Food poisoning	Food poisoning
Feeding tube	Feeding tube
Other	Other
Impact of events on eating	If you answered yes to the question on traumatic experiences, do you think this experience has had an effect on your eating?	1 = Yes 0 = No
Indifference to food/eating	Would you say that most of your food problems result from a general lack of interest:	1 = Yes 0 = No
Indifference to food	...in food?
Indifference to eating	...in eating?
Affective responses to food
Anxiety about new foods	Do you feel afraid or nervous when presented with a new food?	1 = Rarely or never 2 = Less than half the time 3 = About half the time 4 = More than half the time 5 = All of the time
Disgust at new foods	Do you feel disgusted when presented with a new food?
Severity of food avoidance
Picky eating	Do you consider yourself a picky eater?	1 = Rarely or never 2 = Less than half the time 3 = About half the time 4 = More than half the time 5 = All of the time
Food neophobia	Are you willing to try a food you have never eaten before?†
Social impairment
Anxiety in social situations	Do you get anxious about social situations because you will be expected to eat?	1 = Rarely or never 2 = Less than half the time 3 = About half the time 4 = More than half the time 5 = All of the time
Avoidance of social situations	Do you avoid social situations that involve food?

Note: Investigator-designed questions and response options assessing L-ARFID, motivations for food avoidance, affective responses to food, severity of food avoidance, and social impairment are displayed. The instructions for this scale read, “Please select the response that best describes your CURRENT experiences.” Participants were classified as having L-ARFID if they answered Yes to at least one of the L-ARFID items. To assess fear of negative consequences of eating, the number of negative aversive events endorsed was summed. Participants who indicated that experiencing aversive events did not impact their eating were assigned a sum score of zero, as aversive eating-related events did not contribute to a fear of negative consequences of eating.

Abbreviation: L-ARFID, Likely Avoidant Restrictive Food Intake Disorder.

^†Food neophobia item was reverse-scored so that higher scores reflect greater food neophobia.

L-ARFID diagnosis.

Participants were classified as having L-ARFID if (a) they indicated that their eating problems led to significant weight loss, nutritional deficiency, interference with job functioning, and/or interference with social relationships, and (b) they did not meet criteria for full or subthreshold AN, BN, or BED based on responses to the EDDS. Participants meeting criteria for an eating disorder other than ARFID were excluded from analysis.

Motivations for food avoidance.

Participants responded to questions designed to assess how sensitivity to the sensory characteristics of food (i.e., smell, texture) kept them from (a) eating a variety of foods and (b) participating in social gatherings involving food. To assess fear of negative consequences of eating, participants reported whether they had experienced a range of aversive experiences related to eating and whether these experiences had affected their eating. To assess indifference to food and eating, participants reported whether they believed their eating difficulties result from a lack of interest in (a) food or (b) eating.

Affective responses to food.

Participants reported how often they experienced (a) anxiety or (b) disgust when presented with new foods.

Severity of food avoidance.

Participants reported their overall severity of food avoidance by rating how often they consider themselves to be a picky eater. To assess food neophobia, participants were asked how often they are willing to try a new food they have never eaten before.

Social impairment.

Participants reported how often they (a) felt anxious about or (b) avoided social situations involving food.

2.3.4 |. Disgust sensitivity

Participants completed the Disgust Scale - Revised (DS-R; Haidt, McCauley, & Rozin, 1994, modified by Olantjuli et al., 2007) as a measure of overall disgust sensitivity, including situations that do not involve food. The scale presents 25 hypothetical scenarios (e.g., “I might be willing to try eating monkey meat, under some circumstances”) and asks participants to rate how disgusting they would find the situation. The original Disgust Scale was validated against corresponding behavioral tasks, suggesting that scores on the self-report measure correlate with actual willingness to perform disgusting tasks (Rozin, Haidt, McCauley, Dunlop, & Ashmore, 1999). The DS-R demonstrated excellent internal consistency in a sample of N = 472 university students in Belgium and the Netherlands (van Overveld, de Jong, Peters, & Schouten, 2011), as well as in the current sample (α = .87). For interpretive purposes, we provide our group results also relative to normative scores.

2.3.5 |. Additional domains of impairment

Additional domains of eating-related impairment were assessed by the Clinical Impairment Assessment Questionnaire (CIA; Bohn & Fairburn, 2008). The CIA instructions were modified to be more relevant to individuals with ARFID. Instead of asking participants to reflect on impairment caused by their eating, exercising, or weight/shape concerns, participants were given the instructions, “Over the past 28 days, to what extent have your eating habits or concerns about your eating…” The CIA asks participants to rate their experience of psychosocial dysfunction due to their eating over the past 28 days with higher scores indicating greater impairment. In a clinical sample of 123 individuals with eating disorders, the CIA demonstrated excellent internal consistency (α = .97) and convergence with clinician impairment ratings (r = .68, p < .001; Bohn et al., 2008). The scale demonstrated excellent internal consistency in the current sample (α = .93).

2.4 |. Data analysis

Data analysis was performed in IBM SPSS^® 27. Participants were classified into four groups, L-ARFID/GI, L-ARFID-only, GI-only, and No-ARFID/No-GI. Chi-square tests were performed to assess categorical variables across groups. Remaining variables were compared across the four L-ARFID/GI groups and the three age groups using two-way ANOVAs. Significant effects were explored using post hoc LSD tests. Given the exploratory nature of the study, an alpha of .05 was maintained despite multiple testing.

3 |. RESULTS

3.1 |. L-ARFID and GI groups

Of the 2,610 participants who met inclusion criteria, 300 were classified as having both L-ARFID and a self-reported GI disorder or symptoms, 1,420 were classified as having L-ARFID only, 142 were classified as having a self-reported GI disorder or symptoms only, and 748 were classified as having neither L-ARFID nor GI disorder/symptoms. Age varied significantly across the four groups, X²(6, 2,610) = 76.66, p < .001. Sex and race did not vary significantly by group, p = .18 and .28.

The likelihood that an individual had a self-reported GI disorder or symptoms did not differ significantly between those with and without L-ARFID, X²(1, 2,610) = .92, p = .35. Frequencies of GI disorders/symptoms reported by participants with and without L-ARFID are provided in Table 3. The results remained similar when excluding those with organic/structural conditions.

TABLE 3.

GI disorder and symptom prevalence in ARFID v. No ARFID

	L-ARFID (n = 1720)		No L-ARFID (n = 890)
	n	%	n	%
Functional GI disorders
Irritable Bowel Syndrome	118	6.86	50	5.62
Constipation†	2	0.12	1	0.11
Dysphagia†	1	0.06	0	0.00
Cyclic vomiting†	1	0.06	0	0.00
Diarrhea†	1	0.06	0	0.00
Esophageal disorders
Reflux	195	11.34	97	10.90
Hiatal hernia	6	0.34	0	0.00
Eosinophilic esophagitis	2	0.12	1	0.11
Esophageal spasms	1	0.06	2	0.22
Barrett esophagus	2	0.12	0	0.00
Gastric disorders
Gastritis	6	0.34	1	0.11
Gastroparesis	5	0.29	0	0.00
Ulcers	2	0.12	0	0.00
Malabsorption disorders
Lactose intolerance	3	0.17	5	0.56
Celiac disease	4	0.23	2	0.22
Non-celiac gluten intolerance	0	0.00	1	0.11
Inflammatory disorders
Inflammatory bowel disease
Crohn’s disease	11	0.64	1	0.11
Ulcerative colitis	4	0.23	1	0.11
Colitis (type not specified)	2	0.12	1	0.11
Diverticulitis	2	0.12	1	0.11
Proctitis	0	0.00	1	0.11
Gallbladder disease/removal	2	0.12	2	0.22
GI surgery	1	0.06	1	0.11
Upset stomach (no disorder specified)	2	0.12	0	0.00

Note: Counts and percentages of participants with reported GI disorders or symptoms are shown.

Abbreviations: GI, gastrointestinal; L-ARFID, Likely Avoidant Restrictive Food Intake Disorder.

^†Indicates a GI feature that could not be easily classified as a symptom or as a disorder based on the participant’s response.

3.2 |. Motivations for food avoidance

Sensory experiences of food varied significantly across L-ARFID/GI groups and age, with no significant group by age interaction effects (Table 4). Individuals with L-ARFID (L-ARFID/GI, L-ARFID-only) reported that sensitivity to food smells and textures kept them from eating a variety of foods and participating in social gatherings related to foods significantly more often than those without L-ARFID (GI-only, No-ARFID/No-GI). Sensory motivations for food avoidance were most common in the 18 to 24 age range and decreased in frequency with increasing age.

TABLE 4.

Estimated marginal means for L-ARFID/GI groups and age groups

					95% CI
		N	EMM	SE	Lower	Upper
Motivations for food avoidance
Sensory experiences of food
Sensitivity to food smells
Variety	L-ARFID/GIa***	299	3.64	.08	3.48	3.79
Does sensitivity to food smells keep you from eating a variety of foods? †,*** ‡,**	L-ARFID-onlya***	1,415	3.56	.04	3.48	3.64
	GI-onlyb***	142	3.03	.12	2.80	3.25
	No-ARFID/No-GIb***	747	3.07	.05	2.97	3.17
	18–24a***	1,125	3.44	.07	3.30	3.57
	25–34b***	977	3.16	.06	3.04	3.28
	35–44	501	3.36	.07	3.22	3.50
Social	L-ARFID/GIa***	300	2.46	.07	2.33	2.60
Does sensitivity to smells keep you from participating in social gatherings with food? †,***	L-ARFID-onlya***	1,415	2.36	.04	2.30	2.44
	GI-onlyb***	141	1.80	.10	1.60	2.00
	No-ARFID/No-GIb***	747	1.79	.05	1.70	1.87
	18–24	1,125	2.14	.06	2.02	2.26
	25–34	977	2.00	.05	1.89	2.10
	35–44	501	2.18	.06	2.06	2.30
Sensitivity to food textures
Variety	L-ARFID/GIa***	299	4.17	0.7	4.03	4.31
Does sensitivity to food textures keep you from eating a variety of foods? †,*** ‡,***	L-ARFID-onlya***	1,415	4.25	.04	4.18	4.32
	GI-onlyb***	142	3.55	.10	3.34	3.75
	No-ARFID/No-GIb***	746	3.57	.05	3.48	3.66
	18–24a***	1,129	4.11	.06	3.99	4.23
	25–34b***	974	3.73	.05	3.62	3.84
	35–44b***	499	3.81	.06	3.69	3.93
Social	L-ARFID/GIa***	298	3.00	.08	2.84	3.15
Does sensitivity to food textures keep you from participating in social gatherings with food? †,*** ‡,**	L-ARFID-onlya***	1,417	2.97	.04	2.89	3.05
	GI-onlyb***	142	2.04	.12	1.82	2.27
	No-ARFID/No-GIb***	745	1.97	.05	1.87	2.07
	18–24a**	1,127	2.65	.07	2.51	2.78
	25–34b**	976	2.34	.06	2.22	2.46
	35–44b**	499	2.50	.07	2.36	2.63
Fear of negative consequences
Number of aversive events †, ***	L-ARFID/GIa***	297	.35	.04	.28	.42
	L-ARFID-onlyb***	1,407	.24	.02	.21	.28
	GI-onlya***	142	.41	.05	.31	.52
	No-ARFID/No-GIc***	743	.15	.02	.11	.20
	18–24	1,117	.30	.03	.24	.36
	25–34	975	.24	.03	.19	.30
	35–44	497	.33	.03	.27	.40
Affective responses to food
Anxiety	L-ARFID/GIa***	300	4.09	.07	3.95	4.23
Do you feel afraid or nervous when presented with a new food? †,*** ‡,***	L-ARFID-onlya***	1,420	4.15	.04	4.08	4.23
	GI-onlyb***	141	3.24	.11	3.03	3.44
	No-ARFID/No-GIb***	748	3.34	.05	3.25	3.43
	18–24a***	1,130	3.92	.06	3.80	4.04
	25–34b*	978	3.63	.06	3.52	3.74
	35–44c*	501	3.56	.06	3.44	3.69
Disgust	L-ARFID/GIa***	300	3.46	.07	3.33	3.60
Do you feel disgust when presented with a new food? †,*** ‡,**	L-ARFID-onlya***	1,412	3.58	.04	3.51	3.65
	GI-onlyb**	141	2.55	.10	2.35	2.75
	No-ARFID/No-GIc**	744	2.82	.05	2.73	2.91
	18–24a***	1,121	3.25	.06	3.13	3.37
	25–34b***	979	2.99	.05	2.88	3.09
	35–44b***	497	3.07	.06	2.95	3.20
Severity of food avoidance
Picky eating	L-ARFID/GIa***	295	4.68	.05	4.58	4.77
Do you consider yourself a picky eater? †,*** ‡,**	L-ARFID-onlya***	1,389	4.73	.02	4.69	4.78
	GI-onlyb***	141	4.11	.07	3.98	4.25
	No-ARFID/No-GIb***	734	4.23	.03	4.17	4.29
	18–24a***	1,096	4.55	.04	4.47	4.63
	25–34b***	968	4.38	.04	4.31	4.45
	35–44b***	495	4.39	.04	4.31	4.47
Food neophobia	L-ARFID/GIa*	300	4.31	.06	4.20	4.43
Are you willing to try a food you have never eaten before? †,*** ‡,**	L-ARFID-onlyb*	1,419	4.46	.03	4.40	4.52
	GI-onlyc*	142	3.68	.09	3.52	3.85
	No-ARFID/No-GId*	745	3.92	.04	3.85	4.00
	18–24a*	1,129	4.13	.05	4.03	4.23
	25–34b*	976	3.97	.05	3.89	4.06
	35–44a*	501	4.19	.05	4.08	4.29
Social impairment
Anxiety in social situations	L-ARFID/GIa***	300	3.87	.07	3.74	4.01
Do you get anxious about social situations because you will be required to eat? †,*** ‡,*	L-ARFID-onlya***	1,420	3.86	.04	3.79	3.94
	GI-onlyb***	141	2.90	.10	2.69	3.10
	No-ARFID/No-GIb***	745	2.91	.05	2.82	3.00
	18–24a***	1,129	3.48	.06	3.36	3.60
	25–34b***	976	3.26	.06	3.15	3.37
	35–44b***	501	3.41	.06	3.29	3.54
Avoidance of social situations	L-ARFID/GIa***	300	3.09	.07	2.95	3.23
Do you avoid social situations that involve food? †,*** ‡,*	L-ARFID-onlya***	1,420	3.09	.04	3.02	3.17
	GI-onlyb***	142	1.99	.10	1.78	2.19
	No-ARFID/No-GIb***	747	2.13	.05	2.03	2.22
	18–24a*	1,130	2.63	.06	2.51	2.75
	25–34b***	978	2.45	.06	2.34	2.56
	35–44b*	501	2.64	.06	2.52	2.77
Disgust sensitivity
DS-R	L-ARFID/GI	263	2.32	.04	2.24	2.40
	L-ARFID-only	1,223	2.25	.02	2.21	2.29
	GI-only	115	2.24	.06	2.12	2.37
	No-ARFID/No-GI	605	2.20	.03	2.14	2.25
	18–24	957	2.25	.04	2.18	2.32
	25–34	831	2.20	.03	2.14	2.27
	35–44	418	2.30	.04	2.23	2.37
Additional domains of impairment
CIA †, *** ‡, ***	L-ARFID/GIa*	276	19.11	.56	18.01	20.22
	L-ARFID-onlyb*	1,250	17.32	.30	16.74	17.90
	GI-onlyc***	127	8.69	.83	7.08	10.32
	No-ARFID/No-GIc***	662	7.79	.38	7.06	8.53
	18–24a***	991	15.51	.49	14.55	16.46
	25–34b***	879	12.43	.44	11.56	13.30
	35–44b***	445	11.76	.51	10.76	12.76

Note: Estimated marginal means for each ARFID/GI group and age group are displayed. Beneath each measure, symbols (^†, ^‡) denote significant main effects. There were no significant group × age interaction effects. ARFID/GI groups and age groups labeled with different letter superscripts (a, b, c) are significantly different from one another. Groups labeled with overlapping letter superscripts are not significantly different from one another. Model statistics are provided in Supporting Information.

Abbreviations: CIA, Clinical Impairment Assessment; DS-R, Disgust Scale-Revised; GI, gastrointestinal; L-ARFID, Likely Avoidant Restrictive Food Intake Disorder.

^†Main effect of group.

^‡Main effect of age.

^*p < .05.

^**p < .01.

^***p < .001.

Fear of negative consequences of eating varied significantly across L-ARFID/GI groups, with no main effect of age and no group by age interaction effect (Table 4). Individuals with self-reported GI disorders (L-ARFID/GI, GI-only) reported a greater number of aversive events that contributed to their food avoidance compared to individuals without a self-reported GI disorder (L-ARFID-only, No-ARFID/No-GI).

Indifference to food varied significantly across the L-ARFID/GI groups for participants in each age group, 18 to 24 (X²[3, 1,127] = 17.97, p < .001), 25 to 34 (X²[3, 976] = 46.81, p < .001) and 35 to 44 (X²[3, 500] = 9.03, p < .05). Indifference to food was significantly more common in participants with L-ARFID (31.25%) than those without L-ARFID (16.00%), X²(1, 1925) = 70.75, p < .001. Chi-square tests between participants with and without a self-reported GI disorder or symptoms revealed no significant differences in prevalence of indifference to food (p = .513). Indifference to eating varied significantly across the L-ARFID/GI groups for participants in each age group, 18 to 24 (X²[3, 1,129] = 28.88, p < .001), 25 to 34 (X²[3, 974] = 19.18, p < .001), and 35 to 44 (X²[3, 501] = 10.87, p < .05). Indifference to eating was significantly more common in participants with L-ARFID (19.81%) than those without L-ARFID (8.33%), X²(1, 2,190) = 70.75, p < .001. Chi-square tests between participants with and without a self-reported GI disorder or symptoms revealed no significant differences in prevalence of indifference to eating (p = .284).

3.3 |. Affective responses to food, disgust sensitivity, severity of food avoidance, eating related impairment

Affective responses to food, severity of food avoidance, and eating-related impairment varied significantly by group and age, with no significant group by age interactions (Table 4). Individuals with L-ARFID (L-ARFID/GI, L-ARFID-only) experienced more frequent anxiety and disgust in response to food, more severe picky eating and food neophobia, and greater eating related impairment compared to groups without L-ARFID (GI-only, No-ARFID/No-GI). Negative affective responses to food, severity of picky eating, and eating-related impairment were most elevated in individuals 18 to 24 and decreased with age.

DS-R scores for all L-ARFID/GI groups were significantly higher than normative scores, with differences of large effect (Table 5). There were no significant differences in DS-R score by L-ARFID/GI group or age (Table 4).

TABLE 5.

Results of one-sample t-tests comparing DS-R scores to normative scores

L-ARFID/GI group	M	SD	n	95% CI for mean difference	t	df	Cohen’s d
L-ARFID/GI	2.31	.64	263	.56, .72	16.11***	262	1.03
L-ARFID-only	2.26	.67	1,223	.55, .62	30.70***	1,222	.92
GI-only	2.25	.59	115	.47, .69	10.56***	114	.97
No-ARFID/No-GI	2.17	.63	605	.45, .56	19.54***	604	.81

Note: One-sample t-tests were performed to determine whether DS-R scores differed significantly from the scores of a sample of n = 34,442 respondents in the United States who completed the survey online (M = 1.67, SD = .61, Haidt, 2012). All GI/ARFID groups had significantly higher mean scores than the normative sample, with large effect sizes.

Abbreviations: DS-R, Disgust Scale-Revised; GI, gastrointestinal; L-ARFID, Likely Avoidant Restrictive Food Intake Disorder.

^***p < .001.

4 |. DISCUSSION

Our findings show that, rather than precluding an ARFID diagnosis, the presence of a GI disorder should inform symptom formulation and case conceptualization with the potential to further enhance our treatment approaches. Across research questions, the presence of a self-reported GI disorder or symptoms and comorbid ARFID diagnosis resulted in a similar pattern of results to those observed in the L-ARFID-only group, with one notable exception: individuals with a self-reported GI disorder or symptoms (L-ARFID/GI, GI-only) reported experiencing more prior aversive events that contributed to food avoidance compared with individuals without self-reported GI problems (L-ARFID-only, No-ARFID/No-GI). We discuss these findings, in turn.

4.1 |. Motivations for food avoidance

The finding that individuals with a self-reported GI disorder or symptoms reported more prior aversive events contributing to their food avoidance is consistent with existing models of pain-related fear learning. In accordance with the fear-avoidance model of pain, the fear of pain may generalize to innocuous sensations associated with pain or that predict pain, and also to activities and experiences that are thought to provoke those sensations (Crombez et al., 2012; Zucker & Bulik, 2020). This can lead to widespread avoidance not only of food, but also to activities that provoke gut and related sensations (e.g., activities that are emotional, exercise, etc.). This finding suggests that fears of the negative consequences of eating may be important sources of disorder-related impairment in individuals with both ARFID and a GI disorder, a finding that replicates the work of Murray, Bailey, et al. (2020a) in their study of ARFID features among patients in a neurogastroenterology clinic. Thus, individuals with ARFID and a comorbid GI disorder may benefit from treatment approaches that address fears of interoceptive sensations (i.e., sensations from visceral organs), particularly gut sensations that are uncomfortable. One such approach, Feeling and Body Investigators: ARFID Division (FBI-ARFID) utilizes interoceptive exposures (“Body Investigations”) to teach children with abdominal pain (Zucker et al., 2017) and/or ARFID that their bodies are smart and that their sensory or visceral sensitivities are a form of superpower. Interoceptive exposure activities are used to help children to approach body and sensory sensations with curiosity rather than avoidance (Zucker et al., 2019). See Craske et al., 2011 for related approaches for GI symptoms in adults.

While a history of GI symptoms may intensify fears of negative consequences of eating in individuals with both ARFID and a GI disorder, other features may be more elevated in those with ARFID. Sensory aversions to food and indifference to food may contribute more greatly to food avoidance in ARFID, and the endorsement of these experiences in an individual presenting with a GI disorder may mark the need for further assessment of ARFID symptomatology. Future research should seek to explore whether these features act as barriers to food approach in those with GI disorders, perhaps a previously unexplored hypothesis that may help to explain why some with GI disorders have difficulty adhering to dietary recommendations.

4.2 |. Negative affective responses to food and disgust experiences

Negative affect to food was elevated in the L-ARFID/GI and L-ARFID-only groups relative to the GI-only and No-ARFID/No-G groups. Our failure to find differences in negative affect to food between L-ARFID/GI and L-ARFID-only may reflect that these groups share common mechanisms of symptom expression. Visceral hypersensitivity, sensitivity to changes in the state of visceral organs (Tait & Sayuk, 2021), has been hypothesized to be elevated in ARFID (Zucker et al., 2019). This individual difference variable is an important component of disorders of gut-brain interaction (Drossman et al., 2016), a sensitivity known to contribute to the intensity of pain symptoms (Simrén et al., 2018). Visceral hypersensitivity may also contribute to negative affect via several hypothetical mechanisms: as negative emotional reactions to intense somatic symptoms, as amplifications of emotional experiences because individuals feel the somatic constituents of emotion more intensely, and/or as a consequence of the avoidance of rewarding activities that may ensue when one tries to manage the symptoms of ARFID and/or a GI disorder. These hypothetical relationships between somatic symptom experiences and negative affect, if supported, would be clinically meaningful as each would have important intervention implications.

An alternative explanation for the lack of observed differences in negative affect to food between L-ARFID/GI and L-ARFID-only is that the intensity of negative affect differentiates the groups, rather than the frequency of negative affect that was assessed in the current study. To be sure, the intensity of negative emotional experiences has been repeatedly documented in ARFID (Cooney, Lieberman, Guimond, & Katzman, 2018), even relative to clinical groups with elevated negative affect, such as other forms of eating disorders (Nicely, Lane-Loney, Masciulli, Hollenbeak, & Ornstein, 2014). It could also be that measures of affective frequency and intensity that are sensitive to context-specific changes in affect, such as affect prior to a meal, may be necessary to further understand the role of negative affect to food in GI disorders and ARFID.

Similarly, the results did not reveal differences in negative affect to food between the GI-only and No-ARFID/No-GI groups. Notably, however, all groups in this study were elevated on a validated measure of disgust sensitivity relative to normative scores (Haidt, 2012; Haidt et al., 1994; Olantjuli et al., 2007; see Table 5). Given that this was a study of adult picky eating, an eating pattern associated with elevated disgust (Harris et al., 2019), it could be that disgust around food is elevated in those with GI disorders when compared with typically developing controls, a further topic for future research. This topic is particularly interesting given recent advances in understanding immune system influences on GI symptoms (Tait & Sayuk, 2021). Disgust protects individuals from potential contamination from pathogens (Curtis, de Barra, & Aunger, 2011) and thus modulation of immune system functioning has been described as upregulating disgust responses as a protective response (e.g., as during pregnancy; Fessler, Eng, & Navarrete, 2005). Thus, elevated disgust across all groups may lead to novel mechanistic hypotheses.

4.3 |. Impairment

Groups with L-ARFID with or without a self-reported GI disorder or symptoms experienced more social impairment than either group without L-ARFID, which is not surprising given that deficits in this domain are part of the diagnostic criteria for ARFID. It is possible that this social impairment is due in part to the stigma of having ARFID. “Picky eating” is widely viewed as an issue of early childhood, and adults with ARFID may conceal their eating challenges because they do not want to be viewed by others as immature or inflexible. They may also avoid eating with others, as they do not want to call attention to their food choices. Since social occasions often involve food, avoiding eating with others could impede the development of close relationships and lead to social isolation, a pattern that may further decrease opportunities for food experiences.

4.4 |. Age-related differences

Not much is known about the course of ARFID and age-related differences in ARFID presentation. Some findings in the current study were age-related differences in sensory sensitivities related to smell and texture, but no age-related effects in indifference to food. First, it should be noted that these age-related differences are cross-sectional and thus may not be replicated if individuals with ARFID are examined over time. However, changes in smell and taste have been documented as a function of aging (Lafreniere & Mann, 2009)—though not typically in samples with the age range studied here (18–44 years). It may be that such changes make it easier for individuals diagnosed with ARFID to approach more foods as they get older as sensory aversions may diminish, paving the way for cognitive approaches to be more effective as individuals age (Thomas et al., 2020). Given the long learning history in which food has been experienced as aversive or non-rewarding, the experience of indifference towards food may not evidence such age-related declines.

4.5 |. Limitations

Findings from the current study must be viewed cautiously considering limitations of the study design. All data were collected from an online convenience sample and diagnoses of L-ARFID and GI disorders were determined by self-report. While items specifically assessed diagnostic criteria of ARFID, the diagnosis of L-ARFID may be more liberal than found by diagnostic interview and the same may be true of self-diagnosed GI disorders. Validated self-report assessments of ARFID are needed in order to improve the detection of ARFID in future research. Further, it may be that the items assessing fear of negative consequences of eating did not fully capture this presentation of ARFID; for example, an individual may have a fear of choking without any prior experiences of choking. Additionally, the items assessing lack of interest in food and eating may not have adequately captured individuals with low appetite (i.e., individuals who do not feel hungry, forget to eat, or tend to feel full more quickly; Thomas et al., 2017). Another limitation of the study is that the sample included only individuals aged 18–44, and thus the results may not generalize to children, adolescents, or older adults with ARFID. Future studies should explore the interplay of ARFID and GI symptomatology in these populations. The presence of a control group with elevated food neophobia and food selectivity may have precluded our ability to detect more nuanced differences between groups such as increased experiences of negative affect or sensory aversions in the group with self-reported GI disorders or symptoms relative to a control sample without heightened food neophobia. That said, our reference group, individuals who self-identified as picky eaters, may strengthen confidence in study findings as all significant results were detected beyond the presence of food selectivity.

4.6 |. Summary

In conclusion, the consideration of ARFID in the presence of a GI disorder, or vice versa, highlights the important potential contribution of prior aversive experiences as contributing to food avoidance. Consideration of the interplay of these disorders emphasizes opportunities to characterize shared mechanisms of symptom expression that may improve our symptom conceptualization and intervention strategies for ARFID, such as a greater emphasis on somatically-focused intervention strategies. Groups with L-ARFID (L-ARFID/GI, L-ARFID only) were distinguished by more frequent sensory aversions, negative affective reactions to food, and social avoidance, highlighting the importance of assessing for these features in those with a GI disorder. Given the high prevalence of GI disorders, the promise of interventions to improve ARFID symptomatology, and the possible negative impact of ARFID on GI function, clinicians and researchers may consider a diagnosis of ARFID in the presence of a GI disorder in order to provide individuals with optimal care.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.