Avoidant‐Restrictive Food Intake Disorder in Patients With Gastrointestinal Disorders: A Systematic Review

ABSTRACT

Introduction

Avoidant‐restrictive food intake disorder (ARFID) is an eating disorder diagnosis implemented with DSM‐5 in 2013. Elevated risk for ARFID in patients with gastrointestinal disorders is suggested. This systematic review explored (1) methods used to identify patients with gastrointestinal disorders and comorbid ARFID and (2) their clinical characteristics.

Methods

Following the PRISMA guidelines, Medline/PubMed, Embase, PsycInfo and CINAHL were searched for studies from January 2000 to December 2025 and a narrative analysis of the systematically extracted data was performed.

Results

Twenty‐three studies with a total of 6211 participants were included. Studies were primarily observational with retrospective chart reviews or self‐reported surveys. Twenty studies were conducted in tertiary gastroenterology clinics, primarily covering disorders of gut−brain interaction. Four studies focused on paediatric populations, 18 on adults and one included both. Participants were predominantly female, aged 6–90 years. ARFID case identification was heterogeneous regarding both data collection method, application of diagnostic criteria and diagnostic tool. ARFID symptoms were most consistently reported, with fear of aversive consequences, such as fear of abdominal pain, nausea or vomiting, being the most dominant. The negative consequences of ARFID symptoms were less consistently described.

Conclusion

Literature on ARFID in gastrointestinal populations has increased in recent years. However, inconsistent identification of ARFID cases may lead to an overestimation of ARFID prevalence among patients with gastrointestinal disorders. Longitudinal studies with full diagnostic criteria applied are needed to better understand the prevalence and presentation of ARFID in gastrointestinal populations.

Summary

• Avoidant‐restrictive food intake disorder (ARFID) can occur in patients with gastrointestinal disorders, presenting with heterogeneous and overlapping symptoms, such as abdominal pain, nausea and vomiting.

• The diagnosis of ARFID includes medical complications and/or psychosocial impairment beyond what would be expected from a pre‐existing medical condition.

• There is a need for further research employing full diagnostic criteria to minimize the risk of overestimation of ARFID prevalence in gastrointestinal populations.

1. Introduction

Avoidant‐restrictive food intake disorder (ARFID) is a relatively new diagnosis implemented in the Diagnostic and Statistical Manual of Mental Disorders 5th Edition (DSM‐5) in the United States of America (USA) in 2013 [1]. In 2019, ARFID was also included in the International Classification of Diseases 11th Edition (ICD‐11), replacing ‘Feeding Disorder of Infancy and Early Childhood’, to reflect that ARFID can occur across all age groups [2]. According to the diagnostic criteria in the DSM‐5, ARFID is characterized by a persistent pattern of restricted or limited food intake, which is more invalidating than typical ‘picky eating’. This restriction may arise from a general lack of interest in food, anxiety/fear of aversive consequences or hypersensitivity to the sensory characteristics of certain foods. These three prototypical ARFID subtypes can occur individually or concurrently and must be associated with one or more of the following: (1) Significant weight loss or inability to gain weight, (2) severe nutritional deficiency, (3) dependence on supplementary feeding or (4) marked impairment in psychosocial functioning [1, 2]. A similar definition is used in ICD‐11 [2].

Patients with ARFID are distinct from those with other restrictive eating disorders, such as anorexia and bulimia nervosa, by the absence of a distorted body image or a desire to lose weight [1, 2]. According to the DSM‐5 diagnostic criteria, the eating disturbance cannot be better explained by another medical or psychiatric disorder or it must exceed what is routinely expected from the given disorder [1, 2].

The aetiology of ARFID remains poorly understood, though current evidence suggests an interplay of temperamental, biological and environmental factors [1, 3]. Psychiatric comorbidities are common in ARFID, with generalized anxiety disorder and obsessive‐compulsive disorder found in 23.5%–57.1% of paediatric cases, and autism spectrum disorder ranging from 8.2% to 54.75% [4, 5, 6]. If left untreated, ARFID can lead to serious medical complications, including cardiovascular issues, loss of vision, reduced bone mass density and electrolyte imbalances [7, 8, 9, 10].

Epidemiological research on ARFID is limited, and prevalence studies have predominantly focused on children and adolescents through retrospective chart reviews or cross‐sectional designs [11]. Current studies present substantial variation in prevalence rates, ranging from 5% to 64% in patients referred to specialized paediatric eating disorder units due to eating difficulties and from less than 1% to 7.2% in non‐clinical populations, who sought medical attention for other reasons and were retrospectively evaluated in accordance with DSM‐5 ARFID criteria [4, 5, 12, 13, 14, 15, 16]. Patients with ARFID frequently present with concomitant gastrointestinal (GI) symptoms such as abdominal pain, nausea or vomiting, often leading them to seek care in somatic departments for diagnostic evaluation and treatment [4, 17, 18]. Further, research suggests that ARFID symptoms are prevalent among patients with GI disorders, including inflammatory bowel disease (IBD) and disorders of gut−brain interaction (DGBI). DGBI are also known as functional GI disorders and include functional dysphagia, functional dyspepsia and irritable bowel syndrome among others [19]. Gastroenterologists may encounter difficulties in distinguishing symptoms of ARFID from symptoms arising from GI disorders due to the overlapping presentations [17, 19, 20]. As outlined above, the DSM‐5 criteria of ARFID account for the exclusion of concurrent medical conditions; however, distinguishing between, for example, DGBI and ARFID can be challenging both clinically and in research contexts. Emerging research indicates that individuals with GI symptoms, particularly those with functional symptoms, are at an elevated risk of developing ARFID [17, 21]. However, the prevalence, clinical picture and impact of ARFID among paediatric and adult patients with GI disorders remain poorly documented. To our knowledge, there are no existing studies systematically synthesizing the evidence on the characteristics of ARFID in adult and paediatric GI populations.

Thus, the aim of this systematic review is to comprehensively explore the existing knowledge on patients of all ages with GI disorders and comorbid symptoms or diagnosis of ARFID. Specifically, this review examines (1) which methods and measurements are used to define and identify ARFID cases and (2) clinical characteristics of patients of all ages identified as ARFID cases with comorbid GI disorders.

2. Methods

This systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) guidelines [22]. The study protocol was pre‐registered on the International Prospective Register of Systematic Reviews (PROSPERO) (registration number: CRD42024523146).

2.1. Literature Search

The literature search was conducted on 14 March 2024 by one review author (S.B.H.) with assistance from a university librarian. In addition to the primary literature search, a secondary search was conducted on 3 December 2025 by the same author to identify new, relevant studies. Sources included the following databases: Medline via PubMed (NLM), Embase (Ovid), PsycInfo (Ovid) and CINAHL (EBSCO). The search employed various combinations of terms related to the target population (e.g., GI disorders, GI diseases, DGBI, IBD) and intervention/exposure (e.g., ARFID). The search was limited to publications from 2000 to 2025 to acknowledge corresponding ARFID symptoms in the prior diagnosis of Feeding Disorder of Infancy and Early Childhood. No other limitations or filters were applied. The exact search strategy, including search terms and the number of records retrieved from each database in both primary and secondary searches, is detailed in the study protocol on PROSPERO. Additionally, one author (S.B.H.) manually screened bibliographies of the included studies by title and abstract to identify further eligible studies.

2.2. Study Inclusion

Studies were systematically screened in three consecutive rounds, according to predefined inclusion and exclusion criteria, using the software programme Covidence [23]. Inclusion criteria: (1) original research including observational studies and qualitative studies and (2) studies published from the year 2000 to the present. Exclusion criteria: (1) reviews or meta‐analyses, experimental studies including animal studies or grey literature without available full texts and (2) studies not available in English or the Scandinavian language. Titles and abstracts were initially screened independently by two authors (S.B.H., C.H.). Subsequently, the same two authors conducted full‐text screenings. In cases of disagreement regarding eligibility or reasons for exclusion, discrepancies were resolved through discussion with a third author (D.R.H.).

2.3. Data Extraction

Descriptive data from the included articles were systematically extracted in Covidence based on predefined extraction sheets, which were developed through discussions among all three authors. One author (S.B.H.) manually gathered the relevant information, which included: study identification (author, title, publication year, country), study characteristics (study design, data collection method and sample size including both total population and ARFID population), clinical characteristics of the study population (age, gender, race, socioeconomic status, GI disorder, weight and nutritional status, ARFID diagnostic tool, validation of the tool, ARFID symptoms, symptom duration, treatment setting, treatment modality, medical consequences, psychosocial impairments and other comorbid psychiatric disorders). All three authors were eventually involved in organizing the data, upon which a narrative analysis was conducted. In cases where overlapping data were found from studies investigating the same population, both studies were included in the narrative analysis. For studies that divided male and female participants into subgroups (e.g., GI disorders or ARFID vs. non‐ARFID) and provided the exact number of each subgroup, the overall gender distribution of the study was calculated and reported in the results section.

2.4. Quality Appraisal

The quality of all included studies was assessed using the Quality Appraisal for Diverse Studies (QuADS) tool, which has demonstrated substantial interrater reliability as well as face and content validity [24]. This standardized tool, developed for systematic reviews that include heterogeneous studies, consisted of 13 items regarding study designs, research aims, strengths and limitations and so forth. Each item generated a score from 0 to 3 leading to a maximum score of 39 points which was converted into a percentage score. To enhance interrater reliability and in accordance with the Quads User Guide, all three authors independently scored one study and discussed the results [24]. Subsequently, all studies were appraised by two authors independently (S.B.H., C.H.). Any discrepancies between the two authors were resolved through discussion and by consulting a third author (D.R.H.) to reach a final consensus. In accordance with the QuADS User Guide, no quality threshold for inclusion in the review was applied; however, results were narratively discussed and considered in the interpretation of the data [24].

3. Results

3.1. Literature Search

A total of 1682 records were identified in the literature search. Of these, 195 duplicates were removed either manually by one author (S.B.H.) or automatically by Covidence and manually verified as true duplicates. After eligibility assessment, this review included 23 studies, of which nine were identified during an updated literature search [25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47]. The complete study selection procedure was illustrated in the PRISMA flowchart (Figure 1). Three included papers were based on the same study population [25, 34, 35]. One study explored the entire population [25], while the second and third studies focused exclusively on the adult population [34] or the paediatric population, respectively [35].

Figure 1.

Preferred reporting items for systematic reviews and meta‐analysis (PRISMA) guidelines. Flowchart of the full study selection process.

3.2. Study Characteristics

Study characteristics and demographic data were presented in Table 1. The total study population consisted of 6211 individuals with various GI disorders. Adult populations were investigated in 18 out of 23 studies, with participant ages ranging from 17 to 90 years [27, 34, 39, 40, 41, 42, 43, 44, 47]. Four studies investigated paediatric populations, with participant ages ranging from 1 [33, 35, 45, 46] [33, 35, 45, 46] [25] [25]. Both male and female genders were represented across all studies; however, only eight [25, 26, 30, 31, 37, 38, 43, 47] [25, 26, 30, 31, 37, 38, 43, 47]. The remaining studies either reported only the number of male or female participants or described gender distribution as a percentage. In 19 [25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 39, 40, 42, 43, 44, 45] [25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 39, 40, 42, 43, 44, 45]. Two studies described mixed educational attainments and occupations of the included patients, and one study reported that 66% of the patients had education levels more [31, 38, 47] [31, 38, 47]. Socioeconomic status was not investigated in any of the remaining studies.

Table 1.

Study characteristics and demographic data on the total study population.

First author (year) nationality	Study design, data collection	Research setting	Sample size, N	GI disorder	Age, years	Gender, female (%)/male (%)
Paediatric studies
Helen Burton Murray [21]	Cross‐sectional study	Tertiary Neurogastroenterology clinic	129 of 539a	DGBI	Mean (SD) range: 10.8 (3.7) 6–18	74/N.A.
USA	‐Retrospective chart reviews
Isha Kaul [33] USA	Prospective longitudinal study	Urban Tertiary Care Hospital	171	Gastroparesis functional dyspepsia	Median, range: 15, 10–17	(75.8%)/N.A.
	‐Self‐reported surveys					(63.6%)/N.A.
Camden E. Matherne [45]	Cross‐sectional study	Paediatric Gastroenterology Clinic	38	DGBI	Mean (SD) range: 14.74 (1.69) 12–17	(71%)/(29%)
USA	‐Self‐reported surveys
	‐Caregiver‐reported surveys
Laura B. Murphy [46] USA	Cross‐sectional study	Multidisciplinary Clinic for complex EoE	115	Eosinophilic esophagitis	Mean, range: 9.1, 1–18	N.A./89 (77.4%)
	‐Self‐reported surveys
	‐Caregiver‐reported surveys
	‐Clinical interviews
Both paediatric and adult studies
Micaela Atkins [25] USA	Cross‐sectional study	Tertiary Neurogastroenterology clinic	539a	DGBI	Range: 6–90	342/155
	‐Retrospective chart reviews
Adult studies
Audrey Bennett [26] USA	Cross‐sectional study	Tertiary Center for Human Nutrition	137	Coeliac disease	Median (LQ; UQ): 43 (28; 60)	107/30
	‐Retrospective chart reviews
	‐Self‐reported surveys
Amanda Blomsten [39]	Cross‐sectional	General population	825	Functional dyspepsia	Mean (SD): 37.4 (15.0)	546 (66.2%)/N.A.
Sweden	‐Self‐reported surveys			Functional bowel disorder
Taylor Boyd [22] USA	Cross‐sectional study	Tertiary Neurogastroenterology clinic	308	Functional constipation Functional constipation/functional dyspepsia overlap syndrome	Mean (SD) range: 42.6 (16.8) 17–84	(82.8%)/N.A.
	‐Retrospective cohort
Helen Burton Murray [28]	Cross‐sectional study	Tertiary Academic Medical Centers	288	Gastroparesis	Mean (SD) range: 42.7 (16.3) 17–78	226 (78.2%)/N.A
USA	‐Retrospective chart reviews			Functional dyspepsia
	‐Self‐reported surveys
Helen Burton Murray [28]	Cross‐sectional study	Tertiary Neurogastroenterology clinic	410 of 539a	DGBI	Mean (SD) range: 48.9 (17.5) 18–90	298/N.A.
USA	‐Retrospective chart reviews
Helen Burton Murray [29]	Cross‐sectional study	Tertiary Gastro‐enterology clinic	99	DGBI	Mean, range: 45, 18–82	(74.1%)/N.A.
USA	‐Retrospective chart reviews
	‐Self‐reported surveys
Helen Burton Murray [30]	Part 1: Retrospective cohort	Tertiary Gastro‐enterology Division	14	DGBI	Mean (SD) range: 40 (17.9) 20–75	10 (71%)/4 (29%)
USA	Part 2: Prospective cohort
	‐Baseline data from cohort
	‐Self‐reported surveys
	‐Qualitative interviews
Helen Burton Murray (2024)	Cross‐sectional study	Tertiary Gastro‐enterology	101	Ulcerative colitis in remission	Mean (SD) range: 49.9 (16.5) 22–80	56 (55%)/N.A.
USA	‐Baseline data from ongoing cohort	Division
	‐Self‐reported surveys
Margaret Fink [31] USA	Cross‐sectional study	Outpatient Gastroenterology Practice	289	Achalasia	Mean (SD) range: −48.16 (14.2) 20–78	235/41
				Coeliac disease	−41.59 (14.5) 22–71
	‐Self‐reported surveys	Research database Social media		Eosinophilic esophagitis	−44.13 (14.5) 18–82
				IBD	−41.64 (12.3) 18–68
Rosie Flack [41]	Cross‐sectional study	General population	1704	DGBI	Median: 42	(48.3%)/N.A.
USA and UK	‐Self‐reported surveys
Laurie B. Grossberg [42]	Cross‐sectional study	Tertiary IBD Centers	325	IBD:	Mean (SD): 49.14 (16.88)	(56%)/N.A.
USA	‐Self‐reported surveys			‐Crohn's disease
+Race				‐Ulcerative colitis
				‐IBD‐unspecified
Erin Hollis [32] USA	Cross‐sectional study	Tertiary Academic Medical Center	107	Gastroparesis	Mean (SD): 45.4 (17.2)	(84.1%)/N.A.
	‐Self‐reported surveys
Madison A. Hooper [43]	Pilot Case Series	Tertiary Medical Center	4	Crohn's disease	Range: 19–35	3 (75%)/1 (25%)
USA	‐Clinical interviews
	‐Self‐reported surveys
Lee D. Martin [44] UK	Cross‐sectional study	Tertiary Neurogastroenterology Service	33	DGBI	Men (SD) range: 44.3 (15.5) 18–73	29 (88%)/N.A.
	‐Retrospective chart reviews
Kimberly Robelin [21] USA	Cross‐sectional pilot study	Tertiary IBD Center	98	IBD:	Mean (SD): 44.5 (16.4)	55 (56.1%)/N.A.
				‐Crohn's disease
	‐Self‐reported surveys			‐Ulcerative colitis
Wnjing Tu [47] China	Cross‐sectional study	Tertiary Gastroenterology Clinics	483	IBD	Mean, range: 37.2, 18–65	218 (45.1%)/265 (54.9%)
	‐Retrospective chart reviews			‐Crohn's disease
	‐Self‐reported surveys			‐Ulcerative colitis
Emily Yelencich [37] USA	Cross‐sectional study	Tertiary Center for IBD Clinic	161	IBD:	Mean (SD): 41.1 (15.5)	88 (54.7%)/73 (54.3%)
				‐Crohn's disease
	‐Retrospective chart reviews			‐Ulcerative colitis
				‐IBD‐unclassified
	‐Self‐reported surveys
Tingting Yin [38] China	Cross‐sectional study	Four Tertiary Hospitals	372	IBD:	Mean (SD): 38.82 (14.7)	127 (34.1%)/245 (65.9%)
	‐Self‐reported surveys
				‐Crohn's disease
				‐Ulcerative colitis

Abbreviations: ARFID, avoidant/restrictive food intake disorder; DGBI, disorders of gut−brain interaction; GI, gastrointestinal; IBD, inflammatory bowel disease; SD, standard deviation.

^aAtkins et al. explored the same study populations as in Murray et al. 2020 and Murray et al. 2022.

Twenty studies employed a cross‐sectional design [27, 28, 29, 31, 32, 34, 35, 36, 37, 38, 39, 40, 41, 42, 44, 45, 46, 47]. Only two prospective cohort studies and one pilot case series were included [30, 33, 43]. Data collection primarily involved self‐reported surveys, retrospective chart reviews or a combination of both. Three studies incorporated structured or unstructured interviews to assess the ARFID diagnosis [30, 43, 46]. All studies were published in 2020 or subsequent years. A single study was conducted in Sweden, two in China and the United Kingdom, respectively, whereas the remaining were performed in the United States of America. The race of the participants was described in 11 studies, with seven studies including mixed races (e.g., White, Black, Asian) [25, 29, 30, 31, 33, 37, 40, 41, 42, 45, 46].

Twenty studies were conducted in tertiary, specialized clinics or hospitals (e.g., units specialized in neurogastroenterology, IBD or nutrition). One study recruited patients through a research database and social media, in addition to enroling patients from a gastroenterology practice [31]. Only two studies included participants solely recruited from the general population [39, 41]. Sample sizes varied from 4 to 1704 GI patients [41, 43]. The majority of reported GI disorders were various types of DGBI, including conditions such as gastroparesis and functional dyspepsia, as well as several instances of IBD.

3.3. ARFID Case Definition and Clinical Characteristics

For this review, the respective definitions of ‘ARFID cases’ of the included studies were accepted; thus, all identified ARFID cases of the respective studies have been included in this review. However, it is important to highlight that the method of case identification and documentation of diagnostic criteria met varied substantially among studies, which is more thoroughly elaborated below and in Table 2. Therefore, for this review, cases are described cautiously as patients presenting with ARFID diagnosis or symptoms, depending on whether studies fully applied the DSM‐5 criteria or not.

Table 2.

Baseline clinical characteristics of individuals with GI disorders and ARFID symptoms.

Study	ARFID sample size, N (%)	ARFID assessment tool	Validation of tool, yes/no	ARFID characteristics			BMI, kg/m2	Psychiatric comorbidity
				Symptoms	Medical consequences	Psychosocial impairments
Paediatric studies
B. Murray et al. [29]	30 (23%)a 11 (8%) definite ARFID cases 19 (15%) potential ARFID cases	Medical journals evaluated by coders In accordance with DSM‐5 criteria by the presence of psychosocial impairment and/or medical consequences Cases are defined as ‘definite ARFID’ and ‘potential ARFID’. Results combined as ‘ARFID symptoms’	No	Fear of aversive consequences (20 patients) ‐Fear of abdominal pain (11 patients) ‐Fear of nausea (8 patients) ‐Fear of bloating (4 patients) ‐Fear of constipation (3 patients) ‐Fear of vomiting (3 patients) ‐Fear of chest pain (1 patient) ‐Fear of throat pain (1 patient) ‐Fear of globulus sensation (1 patient) ‐Fear of two or more GI symptoms (10 patients) Lack of interest in eating (14 patients) Sensory sensitivity (5 patients)	Weight loss/inability to gain weight (17 patients) Nutritional deficiencies (5 patients) Dependency on supplemental feeding (4 patients) Extremely restricted diet (5 patients)	Psychosocial impairment (2 patients)	Percentile mean (SD): 43.2 (32.8)	N.A.
Kaul et al. [33]	NIAS: 16 (48.5%) with gastroparesis 44 (66.7%) with functional dyspepsia PARDI‐AR‐Q: 21 (63.6%) with gastroparesis 43 (65.2%) with functional dyspepsia Total: 71 (41.5%)	NIAS Self‐reported (cut‐off: ≥ 10 picky eating, ≥ 9 appetite, ≥ 10 fear) PARDI‐AR‐Q Self‐reported In accordance with DSM‐5 criteria by the presence of psychosocial impairment and/or medical consequences. Hunter Vital Sign and ChEDE‐Q8 to rule out lack of available food and AN/BN	Yes	Lack of interest Fear of aversive consequences Picky eating/sensory aversion	N.A.	N.A.	N.A.	N.A.
Matherne et al. [45]	16 (42%) self‐reported 21 (55%) caregiver‐reported	NIAS Self‐ and caregiver‐reported (cut‐off: ≥ 10 picky eating, ≥ 9 interest, ≥ 10 fear) DSM‐5 criteria not applied. Cases defined as ‘ARFID symptoms’	No	Appetite (3 patients) Fear (6 patients) Picky (11 patients) Picky and appetite (3 patients) Appetite and fear (7 patients) All three (7 patients)	Weight loss or faltering growth (8 patients)	Functional impairment	BMI extremity (13 patients)	N.A.
Murphy et al. [46]	43 (37.4%)	EDY‐Q Clinical interview In accordance with DSM‐5 criteria by the presence of psychosocial impairment and/or medical consequences	No	N.A.	N.A.	Psychological distress (13% of patients) Low quality of life	N.A.	Mild symptoms of depression (16% of patients) Moderate or greater symptoms of depression (12% of patients)
Both paediatric and adult studies
Atkins et al. [25]	88 (23%) adults 30 (25%) childrena	Medical journals evaluated by coders In accordance with DSM‐5 criteria by the presence of psychosocial impairment and/or medical consequences Cases defined as ‘definite ARFID’ and ‘possible ARFID’. Results combined as ‘ARFID symptoms’	No	Fear of aversive consequences (17 children/81 adults) Lack of interest/low appetite (10 children/20 adults) Sensory sensitivity (3 children/0 adults)	N.A.	N.A.	N.A.	N.A.
Adult studies
Bennett et al. [26]	78 (57%) 30 clinical ARFID cases 48 subclinical ARFID cases	ARFID‐cl Self‐reported In accordance with DSM‐5 criteria by the presence of psychosocial impairment and/or medical consequences, potential contributor for restrictive eating, no shape‐ or weight concern, restricted eating not driven by another medical illness Cases defined as ‘clinical ARFID’ and ‘subclinical ARFID’. Results combined as ‘ARFID’	Yes	Fear of GI discomfort from IBS (55 patients) ‐Fear of vomiting (8 patients) ‐Fear of choking (6 patients) Fear of GI discomfort from medical condition excluding IBS (19 patients) Poor appetite (18 patients) Picky eating (16 patients)	Vitamin/micronutrient deficiency (33 patients) Osteopenia/osteoporosis (19 patients) Supplementary feeding (53 patients)	Impairment in daily life (69 patients)	Median (LQ; UQ) 27 (23;32)	Comorbid anxiety and/or depression (23 patients) Fear of gaining weight (8 patients) Strong desire to eat only healthy foods (15 patients)
Blomsten et al. [39]	188 (22.8%)	NIAS Self‐reported (cut‐off: ≥ 10 picky eating, ≥ 9 appetite, ≥ 10 fear) DSM‐5 criteria not applied. Cases defined as having ‘avoidant/restrictive eating behaviour’	Yes	Picky eating (13.3% of patients) Loss of appetite (13.8% of patients) Fear of symptoms (4.2% of patients)	N.A.	Psychological symptoms Shame	Mean (SD): 24.9 (6.6)	N.A.
Boyd et al. [22]	Functional constipation: 41 (21.7%) Functional constipation/functional dyspepsia overlap syndrome: 38 (31.9%)	Medical journals evaluated by coders In accordance with DSM‐5 criteria by the presence of psychosocial impairment and/or medical consequences Cases defined as ‘definite ARFID’ and ‘potential ARFID’. Results combined as ‘ARFID symptoms’	No	Fear of aversive consequences (21 patients) Lack of interest (1 patient) Sensory sensitivity (1 patient) Meal‐related fullness (14 patients)	Weight loss/failure to gain weight (16 patients) Nutritional deficiency (6 patients) Need for nutritional supplement (3 patients)	Psychosocial interference (11 patients)	N.A.	N.A.
B. Murray et al. [28]	97 (23.7%)a 26 (6.3%) definite ARFID cases 71 (17.3%) potential ARFID cases	Medical journals evaluated by coders In accordance with DSM‐5 criteria by the presence of psychosocial impairment and/or medical consequences Cases defined as ‘definite ARFID’ and ‘potential ARFID’. Results combined as ‘ARFID symptoms’	No	Fear of aversive consequences (90 patients) ‐Fear of nausea (34 patients) ‐Fear of bloating (30 patients) ‐Fear of abdominal pain (26 patients) ‐Fear of throat symptoms (13 patients) ‐Fear of postprandial fullness (11 patients) ‐Fear of constipation (9 patients) ‐Fear of diarrhoea (10 patients) ‐Fear of chest discomfort (2 patients) ‐Fear of regurgitation (2 patients) ‐Fear of bowel obstruction (1 patient) ‐Fear of two or more GI symptoms Lack of interest in eating (21 patients)	N.A.	N.A.	Mean (SD): 22.9 (5.4)	N.A.
Burton Murray et al. [28]	115 (39.9%) 67 (23.3%) definite ARFID cases 39 (13.5%) possible ARFID cases	NIAS Self‐reported (cut‐off ≥ 12 on each subscale or ≥ 12 on any subscale) In accordance with DSM‐5 criteria by the presence of medical consequences and/or psychosocial impairment, potential contributors for restrictive eating and other eating disorders ruled out in EDDS Cases defines as ‘possible ARFID’ and ‘definite ARFID’. Results specified to ‘definite ARFID’	No	Sensory characteristics Aversive consequences Lack of appetite/low interest	Weight loss (16 patients) Nutritional deficiencies (21 patients) Reliance on vitamins/supplements (23 patients) Tube feeding (9 patients)	Academics/work (18 patients) Social life (23 patients) Family (18 patients)	BMI < 18.5 (12 patients)	N.A.
Burton Murray et al. [29]	37 (39.8%)	ARFID questionnaire, modified from ARFID‐CPSPQ Self‐reported In accordance with DSM‐5 criteria by the presence of medical consequences and/or psychosocial impairments, prototypical presentations and no shape‐ or weight concern in EDE‐Q)	No	Fear of aversive consequences (36 patients) Lack of interest/low appetite (16 patients) Sensory sensitivity (16 patients)	Weight loss (11 patients) Nutritional deficiency (22 patients) Enteral feeding (1 patient) Oral nutritional supplement (1 patient)	Psychosocial impairment (21 patients)	Mean (SD): 24.2 (SEM: 1.2) BMI < 18.5 (5 patients)	N.A.
Burton Murray et al. [30]	14 (100%)	Diagnosed by a clinical psychologist In accordance with DSM‐5 criteria by the presence of psychosocial impairment and/or medical consequences NIAS, self‐reported, for symptom score	Yes	Lack of appetite/interest + fear of adverse consequence (8 patients) Fear of adverse consequence (4 patients) Lack of appetite/interest (1 patient) All three presentations (1 patient)	N.A.	N.A.	Mean (SD) range: 20 (3) 16.4−30.5	Generalized anxiety (2 patients) Unspecified anxiety (3 patients) Depression (1 patient)
Burton Murray et al. [40]	11 (11%)	NIAS Self‐reported (cut‐off: ≥ 10 picky eating, ≥ 9 interest, ≥ 10 fear) EDE‐Q‐8 to rule out other eating disorders DSM‐5 criteria not applied. Cases defined as ‘positive ARFID screens’	Yes	Picky eating (3 patients) Interest (4 patients) Fear (2 patients) Picky eating and fear (2 patients)	N.A.	N.A.	N.A.	N.A.
Fink et al. [31]	(53.7%)	NIAS Self‐reported (cut‐off > 23 total score or > 12 on an individual subscale) DSM‐5 criteria not applied. Cases defined as ‘positive ARFID screens’	Yes	Fear of aversive consequences (37.1% of total population) Lack of interest (17.6% of total population) Picky eating (15.8% of total population)	N.A.	Poor social functioning	N.A.	Anxiety Depression
Flack et al. [41]	590 (34.6%)	NIAS Self‐reported (cut‐off: ≥ 10 picky eating, ≥ 9 interest, ≥ 10 fear) DSM‐5 criteria not applied. Cases defined as ‘positive ARFID screens’	Yes	Sensory‐based avoidance (18.1%) Lack of interest in eating (21.5%) Fear of aversive consequences (9.9%)	N.A.	N.A.	N = 428 BMI < 18.5 (34 patients) BMI 18.5−24.9 (155 patients) BMI 25−29.9 (109 patients) Obese 30+ (130 patients)	Anxiety (51% of patients) Depression (49.2% of patients)
Grossberg et al. [42]	58 (17.8%)	PARDI‐AR‐Q Self‐reported NIAS Self‐reported, to measure symptom severity In accordance with DSM‐5 criteria by the presence of psychosocial impairment and/or medical consequences Cases defined as ‘ARFID + patients’	Yes	Picky eating Poor appetite Food fear	Enteral/supplement (13 patients) Nutrition deficiency (26 patients) Weight loss (33 patients)	Only psychosocial impact (12 patients) Any psychosocial impact alone and/or in combination with medical consequences (32 patients)	Mean (SD): 26.09 (8.38)	Anxiety (69.6% of patients) Depression (58.9% of patients)
Hollis et al. [32]	82 (76.6%)	NIAS Self‐reported (cut‐off ≥ 10 on picky scale, ≥ 9 on appetite scale or > 10 on fear scale) In accordance with DSM‐5 criteria by the presence of psychosocial impairment and/or medical consequences	Yes	Picky (37 patients) Appetite (69 patients) Fear (62 patients) Picky only (2 patients) Appetite only (15 patients) Fear only (8 patients) Picky + appetite (3 patients) Picky + fear (3 patients) Appetite + fear (22 patients) All three (29 patients)	Weight loss (66 patients) Nutritional deficiency (49 patients) Need for supplementation (52 patients)	Academics/work (59 patients) Social life (70 patients) Relationship (54 patients) Distress/anxiety/embarrassment (73 patients)	Mean (SD): 25.2 (6.7)	N.A.
Hooper et al. [43]	4 (100%)	PARDI Clinical interview Unstructured clinical interview for additional information on the diagnostic criteria In accordance with DSM‐5 criteria by the presence of psychosocial impairment and/or medical consequences	Yes	Fear (1 patient) Selective eating and appetite loss (2 patients) Appetite loss (1 patient)	Weight loss (1 patient) Underweight (1 patient) Nutritional deficiencies (1 patient) History of weight loss and malnutrition (1 patient) Need for oral supplement (1 patient)	Social (4 patients) Distress (2 patients) Family (2 patients) Occupational (2 patients)	BMI 17.5 (1 patient) BMI 18.9 (1 patient) BMI 22 (1 patient)	ADHD and generalized anxiety (2 patients) OCD (1 patient) Generalized anxiety and depression (1 patient)
Martin et al. [44]	27 (82%) 11 (33%) strict ARFID 16 (49%) lenient ARFID	Medical journals evaluated In accordance with DSM‐5 criteria by the presence of psychosocial impairment and/or medical consequences Cases defined as ‘Strict ARFID’ and ‘Lenient ARFID’. Results specified to ‘Strict ARFID’	No	Fear of aversive consequences (11 patients) Lack of interest and overlapping fear (2 patients) Pain (7 patients) Vomiting (2 patients) Nausea (1 patient) Bloating (1 patient)	Weight loss (8 patients) Nutritional deficiencies (5 patients) Dependence on oral nutrition supplements (3 patients) Dependence on enteral tube feeding (3 patients)	Marked interference with psychosocial impairment (7 patients)	BMI < 16 (3 patients) BMI < 18.5 (4 patients) BMI > 18.5 (1 patient)	N.A.
Robelin et al. [21]	10 (10.2%)	NIAS Self‐reported (cut‐off: ≥ 28 total score) DSM‐5 criteria not applied. Cases defined as ‘positive ARFID screens’	Yes	N.A.	N.A.	N.A.	Mean (SD): 27.5 (0.67)	N.A.
Tu et al. [47]	98 (20.3%)	NIAS Self‐reported (cut‐off: total score ≥ 28) DSM‐5 criteria not applied. Cases defined as ‘positive ARFID screens’	No	Picky eating Appetite Fear	N.A.	N.A.	BMI ≥ 24 (16 patients) BMI 18.5−23.9 (52 patients) BMI < 18.5 (16 patients)	N.A.
Yelenich et al. 2022	28 (17%)	NIAS Self‐reported (cut‐off: ≥ 24 total score) DSM‐5 criteria not applied. Cases defined as ‘positive ARFID risk’	Yes	Fear of negative consequences Picky eating Poor appetite Avoids food during IBD flare (27 patients) Avoids food in the absence of IBD flare (23 patients)	N.A.	N.A.	N.A.	N.A.
Yin et al. [38]	123 (32.5%)	NIAS Self‐reported (cut‐off: ≥ 10 picky, ≥ 9 appetite, ≥ 10 fear) DSM‐5 criteria not applied. Cases defined as ‘positive ARFID risk’	Yes	Fear of aversive consequences Picky eating Poor appetite	N.A.	N.A.	N.A.	N.A.

Abbreviations: AN, anorexia nervosa; ARFID, avoidant/restrictive food intake disorder; ARFID‐cl, avoidant/restrictive food intake disorder checklist; ARFID‐CPSPQ, avoidant/restrictive food intake disorder Canadian Paediatric Surveillance Programme Questionnaire; BMI, body mass index; BN, bulimia nervosa; ChEDE‐Q8, child version of eating disorder examination questionnaire; DGBI, disorders of gut−brain interaction; DSM‐5, Diagnostic and Statistical Manual of Mental Disorders 5th Edition; EDDS, eating disorder diagnostic scale; EDE‐Q, eating disorder examination questionnaire; GI, gastrointestinal; IBD, inflammatory bowel disorder; LQ, lower quartile; N.A., not available; NIAS, nine item avoidant/restrictive food intake disorder screening; PARDI‐AR‐Q, Pica, ARFID and Rumination Disorder Interview‐ARFID Questionnaire; SD, standard deviation; UQ, upper quartile.

^aAtkins et al. explored the same study populations as in Murray et al. 2020 and Murray et al. 2022.

Fourteen studies applied the full DSM‐5 criteria for the identification of ARFID cases [25, 26, 27, 28, 29, 30, 32, 33, 34, 35, 42, 43, 44, 46]. Among these, nine studies conducted their data analyses solely on the basis of patients who fulfilled DSM‐5 criteria [28, 29, 30, 32, 33, 42, 43, 44, 46]. Five studies defined their ARFID populations as either ‘definite ARFID’ or ‘possible/potential ARFID’ (i.e., not enough information was available to establish the diagnosis with certainty) and combined both groups as ‘ARFID symptoms’ in their data analysis [25, 26, 27, 34, 35]. The remaining nine studies identified ARFID cases by the presence of ARFID symptoms and did not apply the full diagnostic criteria, lacking descriptions of medical consequences or psychosocial impairment [31, 36, 37, 38, 39, 40, 41, 45, 47]. Overall, 1975 ARFID cases were identified, of which 651 cases fulfilled the complete DSM‐5 criteria.

The three prototypical ARFID subtypes were described and included in the definition of ARFID cases in all included studies, except for two [36, 46]. Fifteen studies provided the exact number of patients presenting with each specific symptom [25, 26, 27, 29, 30, 31, 32, 34, 35, 39, 40, 41, 43, 44, 45]. Nine of these studies reported that fear of aversive consequences or fear of GI complaints was the most frequent ARFID symptom [25, 26, 27, 29, 30, 31, 34, 35, 44]. Eight studies indicated that patients often presented with concurrent manifestations and overlapping symptoms (e.g., lack of appetite and fear of aversive consequences) [30, 32, 33, 34, 35, 40, 44, 45]. Four studies identified fear of nausea, bloating, abdominal pain and vomiting as primary reasons for avoidant and/or restrictive eating with patients likely fearing two or more GI symptoms [26, 34, 35, 44]. In five studies investigating adult populations, few or none of the included patients with ARFID symptoms or diagnosis exhibited sensory hypersensitivity [25, 27, 30, 34, 35]. None of the included studies specifically addressed whether the reported symptoms could be explained by the comorbid GI disorder.

The negative consequences of restricted eating, as defined in the DSM‐5 diagnostic criteria, were less consistently reported in the included studies. Medical consequences of ARFID, such as weight loss/inability to gain weight, nutritional deficiencies and need for supplementary feeding, were described in 10 studies and a total of 536 cases [26, 27, 28, 29, 32, 35, 42, 43, 44, 45]. Weight status was described in 15 studies with body mass index (BMI) ranging from < 16 to 30.5 kg/m² [26, 28, 29, 30, 32, 34, 35, 36, 39, 41, 42, 43, 44, 45, 47]. Most studies reported mean BMIs within the normal range (i.e., 18.5–25 kg/m²) among patients with ARFID symptoms or diagnosis [29, 30, 32, 34, 39, 41, 43, 44, 47]. In two studies, the ARFID diagnosis was associated with lower BMI [29, 46]. One study reported vitamin or micronutrient deficiencies in 42% and osteoporosis or osteopenia in 39% of ARFID cases [26].

Functional or psychosocial interference including impact on daily life, academics/work, social life, family and so forth, was reported in 13 studies and 479 cases [26, 27, 28, 29, 31, 32, 35, 42, 43, 44]. Six of these studies did not specify the extent of psychosocial impairments [27, 29, 31, 35, 39, 45]. Seven studies reported comorbid psychiatric diagnoses of anxiety and/or depression, and in one study, ARFID was associated with increased anxiety and depression [26, 30, 31]. All four patients in the case series presented with comorbid psychiatric conditions, including depression and/or anxiety, obsessive–compulsive disorder and attention deficit hyperactivity disorder [43]. No other psychiatric comorbidities were investigated.

The occurrence of ARFID cases in the included GI populations varied significantly, ranging from 6.3% to 82% in the adult populations and from 8% to 66.7% in the paediatric populations [33, 34, 35, 44]. One paediatric study reported the occurrence of ARFID in 42% based on self‐report and 55% based on primary caregiver report [45]. Occurrence of ARFID cases was generally higher in studies utilizing self‐reported surveys, ranging from 11% to 76.6% [32, 40], compared to studies evaluating medical journals in accordance with DSM‐5 criteria, which reported occurrence rates between 6.3% and 33% [34, 44]. One study that identified the ARFID diagnosis through clinical interviews in patients with complex eosinophilic esophagitis reported an occurrence of 37.4% [46]. In studies that did not apply the full diagnostic criteria, the reported prevalence of ARFID symptoms ranged from 10.2% to 55% [36, 45].

In 13 studies, the assessment of ARFID was conducted using the self‐reported Nine Item ARFID Screen (NIAS) tool [28, 30, 31, 32, 33, 36, 37, 38, 39, 40, 41, 45, 47]. NIAS was developed based on the three prototypical ARFID manifestations, picky eating, poor appetite and fear of aversive consequences, and was validated as a screening tool for use in patients aged 10 years and older [32, 33]. Cut‐off values varied significantly among the included studies, with only seven studies using the validated values (subscales of either ≥ 10, ≥ 9 and/or ≥ 10 in picky eating, appetite and fear, respectively) [32, 33, 38, 39, 40, 41, 45]. In five studies, more stringent cut‐off values were applied (e.g., total score ≥ 24 or ≥ 12 on any subscale and at least one medical or psychosocial impairment) [28, 31, 36, 37, 47]. Four studies utilized different self‐reported surveys (e.g., Eating Disorder Examination Questionnaire) to rule out other eating disorders and causes of patients' restrictive eating behaviours in accordance to the DSM‐5 exclusion criteria and recommendations from NIAS validation studies [26, 28, 29, 33].

One paediatric and one adult study utilized self‐reported semi‐structured Pica, ARFID and Rumination Disorder Interview ARFID Questionnaire (PARDI‐AR‐Q) based on DSM‐5 diagnostic criteria and the prototypical ARFID manifestations [33, 42]. The paediatric study itself reported moderate agreement between NIAS and PARDI‐AR‐Q [33]. However, only the use of PARDI‐AR‐Q in adolescent and adult populations is supported by findings from a separate validation study [33, 48]. Five studies employed trained coders to evaluate DSM‐5 diagnostic criteria and symptoms from medical records [25, 27, 34, 35, 44]. Two of these studies justified their choice of ARFID assessment tool by citing the unavailability of self‐reported surveys at the time [34, 35]. Two studies utilized a validated ARFID checklist based on DSM‐5 criteria [49] and a non‐validated, modified ARFID questionnaire based on the Canadian Paediatric Surveillance Programme Questionnaire, respectively [26, 29]. Three studies defined ARFID cases based on DSM‐5 criteria using structured or unstructured clinical interviews such as Pica, ARFID and Rumination Disorder Interview or Eating Disorder Examination Questionnaire [30, 43, 46].

The duration of ARFID symptoms was scarcely investigated. One study reported that almost 40% of patients with ARFID symptoms began experiencing eating difficulties after their gastroparesis diagnosis [32]. Another study described a mean duration of GI symptoms of 5.4 years in patients with ARFID and DGBI prior to treatment [30]. This study also examined the impact of cognitive behavioural therapy in an outpatient gastroenterology clinic, demonstrating positive effects on GI‐specific anxiety and achieving desired weight gain. Patients expressed overall satisfaction with the treatment modality and treatment setting [30]. Furthermore, one study reported acceptability of a modified cognitive behavioural therapy for ARFID delivered via teletherapy, with preliminary reductions in ARFID‐related symptoms and impairment [43]. The remaining studies described various dietary treatments for GI disorders or lacked information on ARFID treatment modality or treatment setting altogether.

Specific demographical data of ARFID cases were reported in seven studies [26, 29, 30, 32, 34, 35, 36]. One study reported the racial composition of the ARFID population as 100% White, with no further specification of ethnic background, aside from noting that none of the participants identified as Latino/Hispanic [30].

3.4. Quality Appraisal

The total QuADS scores and individual subscale points of the included studies are summarized in table format (Appendix S1). All studies provided no or very limited information on justification for analytic methods, consideration of stakeholders and evidence of required sample size. Overall, the theoretical underpinning of the research, research aims and research settings were well described. Study designs, study methods and sample sizes were generally similar across the studies. The mean total score was 57.6%, with scores ranging from 43.6% to 64.6% [30, 35].

4. Discussion

This systematic review investigates the current knowledge on patients with GI disorders and comorbid ARFID, with a focus on how ARFID cases are identified and defined, and further on various clinical characteristics of this novel and complex group of patients. To our knowledge, this is the first systematic review examining ARFID in both paediatric and adult populations with GI disorders. ARFID case identification is heterogeneous regarding both data collection methods, application of diagnostic criteria and diagnostic tools. ARFID symptoms are most consistently reported on, with fear of aversive consequences, such as fear of abdominal pain, nausea or vomiting, being the most dominant. The diagnostically defined negative consequences of ARFID symptoms are less consistently described; however, various medical and psychosocial impairments are reported. Few, if any, of the included studies provided data on important variables (e.g., demographic data, nutritional status, symptom duration and psychiatric comorbidity) relevant to ARFID cases, which limits the clinical applicability of these findings.

4.1. Study Characteristics

In this review, ARFID cases are described in patients with GI disorders across an age span from 6 to 90 years, in line with Bourne et al. suggesting that ARFID can occur in all ages [11]. Both genders are represented in the 23 included studies, with females being the dominant gender in 19 of them. Similar gender distributions are observed in Cucinotta et al. [18]. However, this finding may be influenced by the underlying GI disorders and does not necessarily indicate a true female predominance in the prevalence of ARFID. The absence of studies addressing the socioeconomic status of the patients is consistent with findings from a recent scoping review by Willmott et al. [50].

Cross‐sectional studies and retrospective chart reviews comprise most included studies. This contrasts with an earlier systematic review by Gibson et al. on eating disorders and GI symptoms, where most included studies consist of case series and case reports [17]. Similarly, Bourne et al. primarily include case studies to evaluate clinical characteristics [11]. The findings of this review should be interpreted with consideration of the methodological constraints (e.g., selection and information bias) inherent to cross‐sectional and retrospective study designs. In the current review, the included studies utilize self‐reported surveys, evaluation of medical journals or a combination of both to identify ARFID cases, which aligns with the existing literature [6, 18].

The current review consists of 23 studies published between 2020 and 2025. All studies, except for a few from China, Sweden and the UK, originate from the USA. These findings align with a recent scoping review from Bourne et al. including 77 articles, most of which investigate clinical characteristics of ARFID and originate from Canada and USA, with a few studies from Asia and Western Europe [11]. The earlier implementation of the DSM‐5 ARFID diagnosis in the USA nearly a decade ago, compared to the still‐pending ICD‐11 in some countries, may explain the national difference in research focus.

All but two included studies in the current review are conducted in specialized inpatient or outpatient hospital settings. This trend of research being conducted in highly specialized units is also observed in studies by Gibson et al. and Sanchez‐Cerezo et al., who suggests that the complex and heterogeneous presentation of ARFID may be a contributing factor to the need for specialist involvement [6, 17]. Only one study additionally recruits patients from a research database and through social media, which may have introduced selection bias and reduced generalizability. In this review, total sample sizes are relatively small, consistent with other studies investigating ARFID in specific clinical populations [18, 25, 30, 51]. Studies investigating ARFID in non‐clinical populations generally consist of larger sample sizes [6]. The GI disorders investigated in included studies mainly concern DGBI and IBD. This corresponds to Gibson et al. investigating multiple GI disorders and suggesting that the occurrence of ARFID symptoms is increased in patients with IBD [17].

4.2. ARFID Case Definition and Clinical Characteristics

The definitions of ARFID cases in the studies included in this review are heterogeneous regarding the applied diagnostic tool and diagnostic criteria, and the presence of ARFID symptoms is most consistently reported. Few studies have applied the full DSM‐5 diagnostic criteria in the identification of ARFID cases, and specifically medical and/or psychosocial consequences of ARFID are scarcely described. Sanchez‐Cerezo et al. who investigated the epidemiology of paediatric ARFID, point out the same limitation in their included studies [6].

The most frequently reported ARFID symptom is fear of aversive consequences and GI complaints, such as abdominal pain, nausea and vomiting. Cucinotta et al. report similar findings, with 78%–92% of ARFID patients in gastroenterology practice exhibiting the aversive ARFID subtype as predominant [18]. Additionally, Sanchez‐Cerozo et al. describe GI symptoms or GI disorders in 19%–44% of patients with ARFID [6]. Hardly any of the included adult GI patients exhibit heightened sensory sensitivity as main motivation for their restrictive eating. According to Sanchez‐Cerozo et al., the sensory sensitivity prototype of ARFID is often observed in children with autism spectrum disorder [6]. However, this association is not examined in any of the studies included in this review.

Weight loss or inability to gain weight, nutritional deficiencies and the need for supplementary feeding are all reported in the ARFID population. This corresponds with findings by Sanchez‐Cerozo et al., indicating that ARFID patients may present with various physical symptoms due to low weight, weight loss or failure to thrive [6]. Cucinotta et al. also describe bradycardia and orthostatic instability as frequent clinical signs, which are not explored in our population [18]. Most included studies describing BMI as part of nutritional status report that the majority of identified ARFID cases fall within the normal weight category. This contrasts with findings by Cucinotta et al. observing moderate to severe malnutrition among patients with ARFID symptoms at time of admission to gastroenterology practices [18].

Psychosocial interference is a frequently reported consequence of ARFID among included patients, impacting academics, work, relationships and so forth. Nicholas et al. share similar findings, suggesting that patients with ARFID experience more social impairments than non‐ARFID patients [19]. Additionally, psychiatric comorbidities, such as anxiety and depression, are present in patients with GI disorders and ARFID. This corresponds to Sanchez‐Cerozo et al., who estimate that anxiety disorders in ARFID patients range from 9.1% to 72% [6]. Only one study reported comorbid obsessive‐compulsive disorder or attention‐deficit hyperactivity disorder. None of the included studies investigated comorbid autism spectrum disorder, which is frequently reported by Bourne et al. [11].

Occurrence of ARFID cases among included patients with GI disorders ranges from 6.3% to 82% in the adult populations and from 8% to 66.7% in the paediatric populations [33, 34, 35, 44]. This wide variation in observed occurrence aligns with existing research by Sanchez‐Cerozo et al. estimating ARFID prevalence from 1.5% to 15% in paediatric gastroenterology clinics and from 32% to 64% in specialized feeding clinics [6]. Another recent scoping review by Mikhael‐Moussa et al. found that ARFID symptoms are present in 10%–80% of DGBI patients [52]. These variations may reflect the inconsistencies in case definition and methodology across the included studies. Most studies utilize the screening tool, NIAS, which has been reported with high sensitivity for investigating ARFID symptoms [6]. However, NIAS focuses exclusively on ARFID symptoms and does not include any measure of symptom‐caused impairment as defined in the diagnostic criteria, as also described by Bourne et al. and Mikhael‐Moussa et al. [11, 52]. The variation in both cut‐off values and use of diagnostic criteria during assessment of included patients is noteworthy. It is important to emphasize that a clinical diagnosis of ARFID can only be established through thorough clinical face‐to‐face interviews for diagnostic accuracy and for excluding differential diagnoses. Such clinical assessments are barely performed in the included studies, thus limiting the validity of identified ARFID cases and estimated occurrences. However, one study estimated the prevalence of ARFID among children with eosinophilic esophagitis to be 37.4% based on clinical interviews [46]. According to Bourne et al., the Pica, ARFID and Rumination Disorder Interview ARFID Questionnaire (PARDI‐AR‐Q) is a promising tool not only for ARFID symptom screening but also for diagnostic purposes because it includes all diagnostic criteria [11]. Interestingly, one of our included studies reports moderate agreement between NIAS and PARDI‐AR‐Q in a child and adolescent population [33]. Additional symptoms or motivations for restrictive eating patterns may not be identified when solely using a standardized screening tool. Retrospective evaluation of medical journals by trained coders generally seem to compose lower occurrence of ARFID ranging from 6.3% to 33% [34, 44]. This corresponds to Sanchez‐Cerozo et al., who estimate ARFID prevalence to be between 5% and 22.5% using this method [6].

In one study, almost 40% of patients with ARFID symptoms began experiencing eating difficulties after their gastroparesis diagnosis [32]. However, none of the other included studies investigated whether symptoms of GI disorders preceded the onset of ARFID or if ARFID symptoms developed first and potentially exacerbated GI issues. Accordingly, the included studies do not explicitly address whether the ARFID symptoms are a consequence of the GI disorder or vice versa, which is consistent with the challenges highlighted by Staller et al. regarding the diagnosis of ARFID in GI populations [20]. The mean symptom duration prior to treatment is found to be 5.4 years among included participants [30]. Bourne et al. and Sanchez‐Cerozo et al. both report that patients with ARFID experience a longer duration of illness prior to diagnosis compared to other eating disorders [6, 11]. These findings highlight the importance of healthcare professionals recognizing ARFID symptoms and the necessity for additional longitudinal studies to establish the temporal relationship between GI symptoms and ARFID. Understanding the sequence of these conditions is essential, as it could significantly influence early detection efforts, inform targeted treatment strategies and guide the development of effective preventive measures.

In our population, cognitive behavioural therapy is the only treatment described and shown to be effective [30, 43]. Bourne et al. further explore family‐based therapy and pharmacological treatments, including mirtazapine and olanzapine, with a multimodal approach in both paediatric and adult populations [11].

Demographic information specific to patients with ARFID symptoms or diagnosis is scarcely examined in the included studies. One study exclusively consists of White participants, similar to findings by Willmott et al. [30, 50]. The included study does not further specify ethnicity of the participants, other than none identified as Latino/Hispanic.

4.3. Strengths and Limitations

A major strength of this systematic review is the thorough literature search conducted in collaboration with a university librarian across four different databases. This approach ensures a precise and replicable procedure throughout the review. Also, an updated literature search was conducted prior to publication to identify and include newly published relevant studies. We have decided not to limit the search by language restrictions, instead applying them during the article screening to ensure that no additional studies with an English or Scandinavian version available elsewhere are discounted. Furthermore, we have expanded the publication year range to include studies from 2000 up to and including the time of the literature search to acknowledge the prior diagnosis of Feeding Disorder of Infancy and Early Childhood and corresponding symptoms of ARFID. We have determined to include studies employing various methods of data description and adherence to the diagnostic criteria, including case definitions, acknowledging that not all participants necessarily meet the diagnostic criteria for ARFID. This approach is to accommodate different diagnostic tools, as well as the relatively recent implementation of the diagnosis. This approach has both strengths and limitations, as it has enabled us to gather a larger data set that highlights the importance of the topic, but it also requires caution to avoid overuse of ARFID terminology within this population. The article screening and quality assessment are conducted independently by three authors, significantly reducing the risk of selection bias. Anticipating the inclusion of heterogeneous studies, we have chosen to assess the risk of bias using the QuADS tool. However, the observed similarity in study design among the studies included leads to less variation in the accomplished scores. The overall quality scores of the included studies are relatively equal, raising questions about the importance of the chosen assessment tool. A limitation of this review is that data extraction is performed by one author and is only later approved by the remaining two authors.

4.4. Future Perspectives

In light of the recent implementation of the ARFID diagnosis in ICD‐11 and before that also in the DSM‐5, there is a need for systematic research into this relatively newly described condition. This review identifies several important gaps in the existing literature. Future research should focus on consistent case identification based on validated diagnostic tools which explore the complete ARFID criteria as defined in the diagnostic manuals. Further, attention should be paid to the challenging differentiation of GI symptom attribution in patients with both ARFID and GI disorders. Further longitudinal research is essential to assess the prevalence, identify risk factors and long‐time progression of ARFID in patients with GI disorders. Future studies should aim to include diverse populations and obtain relevant information, such as demographic data, nutritional status and symptom duration, to enhance the generalizability and applicability of findings. Addressing these gaps will be crucial for developing effective screening, diagnostic and treatment protocols to improve patient outcomes in this population.

5. Conclusion

This systematic review comprehensively examines the existing research on patients of all ages with GI disorders and comorbid ARFID. ARFID case identification is heterogeneous regarding both data collection methods, application of diagnostic criteria and diagnostic tools. ARFID symptoms are most consistently reported on, with fear of aversive consequences, such as fear of abdominal pain, nausea or vomiting, being the most dominant. The diagnostically defined negative consequences of ARFID symptoms are less consistently described; however, medical and psychosocial impairments are reported, as well as psychiatric comorbidities such as anxiety and depression. Consequently, the findings of included studies must be interpreted with caution in order to minimize the risk of overestimation of ARFID prevalence in GI populations.

Author Contributions

All authors made substantial contributions to the conception and design of the study, as well as to the screening of articles, quality appraisal and data analysis. The first author was responsible for conducting the initial and final literature searches, performing data extraction and drafting the primary and final versions of the manuscript. The second and third authors contributed to the writing process and provided substantial revisions of the manuscript for important intellectual content. All authors have read and approved the final version of the manuscript and agree to be accountable for all aspects of the work.

Funding

The authors received no specific funding for this work.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Appendix 1 – Quality assessment of the 14 included studies according to QuADS criteria.