April 27, 2026 Eating Disorders

When It Is Both: ARFID-Spectrum Features Inside Restrictive Anorexia

By Brian Nuckols, MA, LPC-A

LPC-A · Center for Discovery · Gottman Trained · EFT · DBT · Depth Psychology · Pittsburgh, PA

TL;DR: A minority of patients carrying an anorexia nervosa, restricting type diagnosis also carry a pre-existing sensory or fear substrate that meets the clinical description of ARFID features. When treatment teams read the sensory or fear pattern as a manifestation of the AN rather than as an overlay that predates it, refeeding protocols fail along predictable sensory or autonomic lines. Jennifer Thomas and Kamryn Eddy’s differential framework, Christopher Fairburn’s transdiagnostic model, and Rachel Bryant-Waugh’s diagnostic precision work provide the language for reading the overlap. Treatment integrates elements from CBT-E and CBT-AR rather than running either protocol as if the other were not present.

The Intake That Does Not Resolve

On a Tuesday morning in September, the intake clinician at an outpatient eating-disorder program reviews the referral packet for a nineteen-year-old who has been in residential treatment twice in the preceding eighteen months. The packet documents a weight loss of forty-two pounds between her senior year of high school and her first college semester, a hospital admission with a heart rate of thirty-eight and a serum potassium of 2.9 mmol/L, a confirmed diagnosis of anorexia nervosa, restricting type, and two discharges from residential care in which her weight had partially restored. What the packet also documents, in a detail her parents provided on the family-history form that the intake clinician reads three times before the appointment, is that the patient has not eaten chicken since age four. The form lists twelve other foods she has refused since early childhood on what the parents describe as texture grounds: bananas, yogurt, oatmeal, most cooked vegetables, any meat with visible fat, eggs unless scrambled dry, cottage cheese, pudding, applesauce, tapioca, and rice pudding. The list is in the mother’s handwriting. It ends with a note that reads, in small letters, “she was like this before any of the weight stuff.”

The intake clinician has seen this pattern before. The referral source, a residential program that used a standard AN-R refeeding protocol, did not address the texture list. The discharge summary from the second admission notes that the patient “resisted nutritional supplementation” and “demonstrated ongoing food refusal consistent with anorexic thinking.” The intake clinician, who has trained in both CBT-E and CBT-AR, reads the mother’s handwritten list against the discharge summary and recognizes the treatment failure in advance of meeting the patient. The refeeding protocol asked the patient to drink Ensure. The texture of Ensure has been on her intolerable-textures list since she was four. The protocol did not know this. The discharge summary read the refusal as resistance to the AN treatment. The resistance was to a texture that predated the AN by fifteen years.

The Diagnostic Boundary and Its Cost

The DSM-5 entry for ARFID includes an exclusion criterion that has complicated clinical practice since 2013. The diagnosis cannot be applied when the food avoidance occurs in the context of concurrent body-image disturbance or fear of weight gain. The exclusion was written to prevent diagnostic drift, the concern being that clinicians might assign ARFID to restrictive anorexia patients as a way of avoiding the weight-and-shape work. The unintended consequence has been that patients carrying both a genuine AN diagnosis and a genuine pre-existing ARFID substrate have no clean diagnostic home. The second diagnosis is blocked by the exclusion. The first diagnosis does not account for the sensory or fear features that the treatment team needs to recognize.

Rachel Bryant-Waugh’s writing on the diagnostic boundary, extending from her role in the DSM-5 ARFID development and continuing through her clinical work at the Maudsley, has argued that clinicians should treat the exclusion as a diagnostic convenience rather than as a claim about the underlying phenomenology. A patient can have both conditions operating on different substrates. The body-image-driven restriction characteristic of AN-R and the texture-driven restriction characteristic of sensory ARFID can run in parallel within the same person, with the AN layer sitting on top of an ARFID substrate that has been present since childhood. When the treatment team reads the sensory refusal as anorexic thinking, the intervention targets the wrong mechanism, and the refeeding protocol breaks against a sensory floor the protocol did not anticipate.

Thomas, Eddy, and the Differential Framework

Jennifer Thomas and Kamryn Eddy, whose three-dimensional ARFID model structures much of the current clinical vocabulary, have written specifically about the overlap presentation in the CBT-AR manual published through Cambridge University Press in 2018. Their framework distinguishes three clinical situations that can look similar at the surface. In the first, the AN diagnosis is primary and the apparent sensory or fear features are epiphenomenal to the weight-and-shape pathology; these features resolve as weight restores and body-image work proceeds. In the second, the ARFID features are primary and what appears to be body-image concern is actually an anxious accommodation to a sensory or fear-driven restriction that has produced low weight; here the treatment priority is the ARFID axis, and the apparent AN features often resolve once the ARFID substrate is addressed. In the third, both conditions are present as independent processes with their own onset histories, their own mechanisms, and their own treatment targets.

The nineteen-year-old in the case above fits the third pattern. Her childhood history of texture-based refusal is documented by her parents across fifteen years. Her body-image distortion and caloric restriction emerged in her senior year of high school, precipitated by what her intake records describe as a breakup and a subsequent period of social comparison with a friend group that had begun dieting together. The two patterns have separate onset dates, separate trigger histories, and separate clinical targets. Collapsing them into a single AN diagnosis produces the treatment failure the discharge summary documented. Reading them as overlap produces a different treatment plan.

Fairburn and the Transdiagnostic Model

Christopher Fairburn’s transdiagnostic CBT-E, developed at Oxford across two decades and consolidated in the clinical manual published in 2008, provides the backbone of most evidence-based outpatient treatment for anorexia nervosa and bulimia nervosa. The model treats the core psychopathology of eating disorders as a dysfunctional scheme for self-evaluation based excessively on weight, shape, and control over eating. The treatment protocol addresses this scheme through behavioral experiments, psychoeducation, and the gradual loosening of rigid dietary rules. For pure AN-R presentations, the evidence base is strong. For overlap presentations, the limitation is that the transdiagnostic model was not designed to address a sensory or autonomic substrate that does not run through the weight-and-shape scheme.

Fairburn’s own writing acknowledges that the transdiagnostic model works across the eating disorders he defined. ARFID was added to the diagnostic system after CBT-E was developed. The integration of CBT-E with CBT-AR is an active area of clinical development. A clinician working with an overlap presentation typically uses CBT-E modules for the weight-restoration and body-image components while running CBT-AR modules on the specific sensory or fear dimension that the AN framework does not address. The sequencing varies by case. In acute medical presentations, weight restoration takes precedence and sensory accommodations are built into the refeeding plan rather than addressed separately. In outpatient work with medically stable patients, the two streams can run in parallel, with the treatment plan explicitly sequencing the ARFID work for the texture or fear layer and the AN work for the body-image layer.

What Gets Missed in the Residential Chart

The discharge summary the intake clinician reads on Tuesday morning documents a specific pattern of missed sensory cues. The patient’s refusal of Ensure was coded as resistance. The refusal to eat the yogurt on the breakfast tray was coded as restriction. The refusal to finish the casserole was coded as noncompliance. Each of these refusals, read against the handwritten list her mother provided, maps to a texture that has been on the intolerable list since early childhood. A clinician trained in CBT-AR, reviewing the same chart, reads the same refusals as sensory responses that predate the AN by more than a decade. The two readings produce opposite treatment decisions. The first reading escalates the behavioral contingencies. The second reading adjusts the food plan.

The clinical cost of the first reading accumulates across admissions. Patients who are coded as resistant when they are actually hitting a sensory floor experience the treatment team as adversarial and the treatment itself as ineffective. Weight gain during residential care partially restores, then regresses after discharge, because the patient has not actually expanded her tolerable food range. The AN work proceeds on a surface that has no sensory accommodation underneath it. The readmission rate in this population is higher than in patients with pure AN-R presentations, because the underlying sensory substrate that was not addressed during the first admission reasserts itself as soon as the external structure of the residential program ends.

The Autism Question

Adolescents and young adults presenting with AN-R who also carry an ARFID substrate screen positive for autism at rates substantially above the general-population baseline. Kate Tchanturia’s research at King’s College London, working alongside autism researchers across the UK eating-disorder network, has documented that roughly one in three adult female AN patients screen positive for autism on at least one validated instrument, with higher rates in chronic and treatment-resistant presentations. The mechanism appears to run through the shared neurological substrate that produces both the sensory processing differences characteristic of autism and the sensory-primary ARFID features that often co-occur with it.

For the nineteen-year-old in the intake, a competent evaluation includes an autism screen alongside the ARFID and AN assessments. A positive screen does not change the immediate clinical priority, which is weight stabilization and coordinated treatment of the overlapping diagnoses. It does change the treatment delivery. An autistic patient with AN and an ARFID overlay benefits from literal language at intake, from pacing that respects the autistic sensory threshold, from coordination with occupational therapy for sensory regulation, and from a family-based component that accounts for the autistic family dynamics rather than running a standard FBT protocol as if the neurology were typical.

How the Treatment Plan Actually Reads

For the patient whose intake is scheduled Tuesday morning, the treatment plan the clinician will draft after the evaluation looks different from what either residential program produced. The plan includes medical monitoring on a weekly schedule during the initial stabilization phase, with targets for weight, electrolytes, and vital signs drawn from the AN medical protocols. The food plan, constructed in collaboration with a nutritionist trained in both AN refeeding and sensory accommodation, excludes the textures on the mother’s handwritten list and substitutes caloric equivalents that the sensory profile tolerates. The CBT-E modules on weight-and-shape cognitions and on the self-evaluation scheme run through the outpatient therapy sessions. The CBT-AR modules on graduated sensory exposure run in parallel, targeting one texture at a time from the intolerable list, with the goal of widening the repertoire across the outpatient treatment window. The family work, adapted from FBT but calibrated to the overlap presentation, coaches the parents on when to hold firm on the AN-driven refusals and when to accommodate the sensory-driven ones, with the differential language built directly into the parental coaching.

The treatment window is longer than it would be for either pure condition. Pure AN-R outpatient treatment under CBT-E runs typically across twenty to forty sessions. Pure sensory ARFID under CBT-AR runs typically across sixteen to thirty-two sessions. Overlap presentations commonly require more, because the two streams are doing non-duplicative work. The clinical goal is not to resolve both conditions on the AN timeline. It is to produce an outpatient pathway that addresses the full phenomenology the patient is actually living with, rather than the subset of it that one diagnosis alone captures.

What the Handwritten List Is Doing

The mother’s handwritten list, stapled to the referral packet, is the most clinically useful document in the intake folder. The residential discharge summaries are longer. The therapy notes are denser. The laboratory values are more precise. The list of twelve textures the patient has refused since age four, ending with the sentence about predating the weight stuff, is what distinguishes this presentation from the AN-R case that the residential protocol was designed to treat. The clinician who reads the list as clinical data, rather than as a parental attempt to explain away the anorexia, sees the overlay the two prior admissions missed. The plan she drafts incorporates the list. The refeeding, when it resumes, will not require the patient to drink Ensure. The texture the patient gagged on at age four will not be the texture the residential program tries to push through in the third admission. The clinical plan will have learned what the two prior plans did not.

Further reading on the sensory axis of the three-dimensional ARFID model is available in the post on ARFID and sensory sensitivity. The shared neurological substrate that connects ARFID with autism, particularly in late-diagnosed presentations, is covered in ARFID and autism. The broader diagnostic distinction between ARFID and ordinary developmental picky eating is mapped in ARFID vs picky eating. A full treatment overview, including the integrated CBT-E and CBT-AR approach described above, is available in ARFID treatment. The eating disorders assessment can serve as a starting point for families whose intake packets are likely to include their own handwritten lists. The ARFID course covers the three-dimensional model and its differential boundaries in depth across eleven modules.

If the pattern in this post resembles what your family’s treatment history has looked like, a consultation is the next step.

Frequently Asked Questions

How do clinicians distinguish ARFID features from anorexia features in the same patient?

The differential runs along three axes. First, chronology: sensory or fear features that predate the onset of weight-and-shape concerns by years point toward ARFID substrate rather than anorexia manifestation, whereas restriction that emerged alongside or after body-image distortion is more readily explained by the AN framework. Second, phenomenology: a refusal of a specific texture that the patient can describe in sensory terms without reference to calories or body size is clinically different from a refusal framed in terms of weight or shape. Third, treatment response: patients whose restriction loosens with refeeding and body-image work alone are showing that the AN framework was sufficient; patients whose restriction persists along specific sensory or fear dimensions even as weight restores are revealing an ARFID overlay that the AN protocol did not address. Jennifer Thomas and Kamryn Eddy's differential framework, along with Rachel Bryant-Waugh's diagnostic clarification work, provides the clinical language for sorting these cases.

What do integrated treatment adaptations look like in practice?

Fairburn's CBT-E, the transdiagnostic treatment for eating disorders developed at Oxford, provides the backbone for the body-image and weight-restoration work when anorexia is the primary diagnosis. Thomas and Eddy's CBT-AR provides the protocol for the sensory or fear axis when ARFID features are overlaying. In practice, a clinician working with an overlap presentation sequences or integrates elements from both. Weight restoration and body-image work proceed through CBT-E while the sensory or fear dimension is addressed through CBT-AR modules on graduated exposure, which often means that the refeeding food plan is constructed with sensory tolerances in mind rather than ignoring them. A patient who cannot tolerate the texture of a nutritional supplement drink for reasons that predate her AN will not benefit from a refeeding protocol that requires her to drink it. The adaptation substitutes an equivalent caloric load in a form the sensory profile tolerates.

Why does family-based treatment alone sometimes fail to resolve the sensory or fear layer?

Family-based treatment, developed at the Maudsley and refined through the Lock and Le Grange manual for adolescent anorexia, is organized around parental authority over feeding during the weight-restoration phase. The protocol is highly effective for adolescents with straightforward AN-R presentations, and it remains the first-line recommendation in many treatment guidelines. The limitation, when ARFID overlay is present, is that parental pressure to eat a specific food can fail or produce regression if the food triggers a sensory or autonomic response that the AN framework does not recognize. A parent enforcing consumption of a yogurt whose texture the child has gagged on since age four is not running the FBT protocol correctly from the AN side, because the AN protocol did not anticipate the sensory floor. The adaptation is for the FBT clinician to work with the family on a food plan that respects the sensory profile while still holding the refeeding objective, which typically requires a clinician who is trained in both FBT and CBT-AR or who is working in coordination with an ARFID specialist.

Talk to someone who gets it →

Questions about insurance or cost? See fees and insurance info