A Broad View: Disordered Eating on the Autism Spectrum
Autism spectrum disorder (ASD) is a neurobiological condition defined by limitations in communication and social interaction. However, the secondary or ancillary aspects of the condition typically receive the most interest in research and practice, including atypical sensory processing and stimulus over-selectivity.
Autism studies have maintained a strong pediatric focus despite the fact that most people on the spectrum are adults. Consequently, a consensus is now emerging that for too long the adult cohort has been neglected, preventing a complete understanding of the health-related patterns (including eating disorders) during the adult years.
New Studies, Old Definitions
Eating pathologies for children on the autism spectrum have been relatively well studied, with a highly cited figure of 90% prevalence for this cohort. A new study among adults with autism has reported a range of EDs that span 6% to 17% of this population.4 However, our understanding is greatly affected by the way diagnostic criteria are established under the DSM-5. Here, language states that if an eating disorder is subsidiary to a mental condition — such as autism — then a secondary diagnosis is not warranted, except in certain circumstances (e.g., the presence of pica) where outcomes are “sufficiently severe.” The consequences of this are profound, as it prevents autism researchers from determining the true rates of prevalence. This, in turn, inhibits discussion on how EDs should be conceptualized, assessed, diagnosed, or treated for this and other cohorts with disabilities. Consequently, autism studies remain in the early stages of making sense of these behaviors.
This overview looks at the manifestation of these disordered eating patterns from a broad view, and argues that four expansive types of disordered eating can be defined. Each is discussed in terms of manifestations, clinical significance, and current treatment options.
Four Categories of Disordered Eating on the Autism Spectrum
Behavioral rigidity. Behavioral rigidity is a characteristic of many psychopathologies, including autism. In the difficulty in transitioning between activities, environments, or even internal aspects of the same task, behavior rigidity is often reflective of deficits in self-regulation. The first category of disordered eating references this type of behavior in the context of food and consumption. Behaviors on the autism spectrum falling into this category include food cravings, food refusals, and, particularly, a limited diet. In this last group, both a narrow focus on specific foods as well as overreliance on specific food classes (e.g., refined carbohydrates) can predominate.
Because these varied behaviors are not driven by concerns about either body shape or weight, they are positioned as analogous to Avoidant/Restrictive Food Intake Disorder (ARFID) rather than eating disorders. Restricted eating on the spectrum has been most thoroughly studied in children, and has been noted as one of the predominant eating-related patterns during this period of the life cycle.
Treatment for food rigidity relies primarily on behavioral techniques, especially “escape extinction,” where food is continually presented until acceptance occurs. However, the results of research studies have shown that increasing overall food intake has been more successful in interventions than as a means of increasing actual dietary variety.
Sensory abnormalities. The second category of disordered eating and autism includes sensory abnormalities that affect hearing, vision, touch, and smell. For all of these systems, atypical sensory functioning is likely to contribute significantly to eating pathologies, though there still is only limited research. As with most autism-related studies, children have been the primary focus where oral preferences and motor difficulties have been noted,15 along with more generalized sensory abnormalities. One small study assessed higher-functioning adults for the ability to discriminate among taste samples. Although the sample population was found to be less accurate in identifying bitter, sweet, and sour tastes than were healthy controls, they were comparable to a control group in the identification of salty foods. A review of olfaction for individuals with ASD found “possible involvement” of impairment of sensory systems, suggesting more study is merited. Early tactile sensitivities have also been proposed as a contributor to specific food preferences.
From a qualitative perspective, this topic can be explored in more depth by reading the autobiographies and memoirs of high-functioning individuals with autism. For example, Stephen Shore, now a professor and autism advocate, remembered growing up in the following way:
“Brown or black food wouldn’t be eaten, as I insisted that they were poisonous. Canned asparagus was intolerable due to its slimy texture, and I didn’t eat tomatoes for a year after a cherry tomato had burst in my mouth while I was eating it. The sensory stimulation of having that small piece of fruit explode in my mouth was too much to bear and I was not going to take any chances of that happening again.
Carrots in a green salad and celery in tuna fish salad are still intolerable to me because the contrast in texture between carrots or celery and salad or tuna fish is too great. However, I enjoy eating celery and baby carrots by themselves. Often as a child, and less now, I would eat things serially, finishing one item on the plate before going on to the next.”
This ability to articulate interior rules has contributed to a better understanding of how food (including the size of the bolus, texture, flavor, color, and shape) can affect eating patterns for this population.
Behaviors with significant health risks. The third category of disordered eating on the autism spectrum incorporates a disparate group of behaviors that – because of the significant health-related risks — require substantial resources and support staff to manage. This group includes pica, rumination, and disruptive mealtime behaviors. Despite the severity, there are no clinical guidelines for assessment or treatment.
Pica, or eating nonnutritive substances, is manifested in the general public as a discriminant behavior typically involving a single class of substance, for example, pregnant women eating clay. On the autism spectrum, however, this behavior is usually manifested in non-discriminant consumption patterns that likely favor opportunity.
Cigarette pica is reported as the most commonly ingested item for this population, despite the fact that rates of nicotine, caffeine, and drug abuse are currently very low. Items reported to have been ingested by individuals with ASD or ID in published research are amazingly wide-ranging, and include: dirt, chalk, cigarettes, plastics, foam rubber, string, paper, paper clips, rubber bands, clothes or cloth, grass, metal, buttons, hair, feces, vomit, rocks, glass, broken light bulbs, insects, paint chips, pencils, trash, toiletries, cleaning products, sewing needles, tar, vinyl, or rubber gloves, carpet, foam padding, toilet bowl fresheners, spoiled food, mothballs, plastic tubing, tea bags, keys, crayons, twigs, alkaline batteries, soap, sealed snack bags, wood chips, jewelry, Styrofoam, coffee grounds, aftershave lotion, toilet water, holiday decorations, and dead animals.
The highest rates of pica behavior on the autism spectrum, from 26% to 65%, have been documented in institutionalized settings. Rates are significantly reduced in community-based residences, where there is a greater emphasis on social stimulation; here, published rates span 0.2% to 4%. However, so many differing methodologies and definitions of pica have been utilized that analysis across this body of research is difficult.
Treatment for patients on the autism spectrum typically addresses pica as either a challenging behavior (CB) or as a psychopathology. CBs are considered culturally abnormal actions that can either put the individual (or those around him or her) at risk, or that are so disruptive as to limit access to the community. In their ability to affect quality of life, limit independence, and create social isolation, CBs are among the most studied and one of the most socially unacceptable aspects of ASD.56 Learning-based approaches are common for treatment of CB, and include Applied Behavior Analysis (ABA), behavior therapy, social skills training, reinforcement models and token economies.
Pica as a psychopathology typically relies on psychological explanations of autism. Though conditions such as anxiety, depression, or even psychosis are difficult to differentially diagnose in a population that is 40% nonverbal, psychiatric symptoms are highly prevalent, with rates for comorbidities ranging from 36% to 81%. Pharmacologic approaches for the treatment of pica have included the use of selective serotonin reuptake inhibitors (SSRIs), often fluoxetine, based on its anti-anxiety and anti-obsessive properties.30,46,50,61,62 Both antipsychotics and atypical antipsychotic have also been utilized in treatment, with varied effectiveness.
While medications are likely the most common treatment modality for all forms of CBs – particularly in adults — they have also been linked to the etiology of pica. Research as to underlying causes of pica have found that psychotropics and anticonvulsants are significantly associated for individuals with ID who display of pica.64 Neuroleptic medication is also believed to have a link with pica, possibly due to diminished postsynaptic dopamine receptor changes, which may worsen the behavior.65
Whatever the perspective, the predominant treatment for pica behavior outside of research contexts is simply “pica-proofing” the patient’s environment. In treatment settings, this usually involves the use of additional staff to protect an individual by sweeping a room for possible pica items and by limiting the opportunities to engage in pica behavior within the larger community.
Rumination. Rumination is also documented on the spectrum, though its prevalence is not completely understood due to the difficulty in clinically identifying this condition in nonverbal populations. Rates of rumination for individuals with autism and ID are estimated in the range of 6% to 10%.66 Along with the immediate risk of aspiration, rumination contributes significantly to increased mortality rates, as well as long-term concerns about dehydration, malnutrition, and gastrointestinal bleeding.68-70 Because of the silent nature of this condition and the significant outcomes, identification of rumination warrants a full clinical workup.
Treatment approaches for rumination on the spectrum have focused on supplemental feedings to break the regurgitation cycle, use of preferred stimuli, and emphasis on overall increased or alternative stimulation.71
Rapid eating behaviors. The last group of disordered eating noted on the autism spectrum involves a range of rapid eating behaviors that approximate binge eating disorder (BED). Binge eating has been documented in institutionalized adults with ID at rates ranging from 3%- to 42%; in adults in community-based residences, reduced rates, ranging from 1% to 19%, have been noted. There is relatively strong anecdotal evidence (and some research) to demonstrate that rapid consumption patterns are highly prevalent on the autism spectrum, 72,73,76,77 and more generally established among individuals with developmental and intellectual disability disorder (IDD).
Traditionally the predominant risks of rapid eating were believed to be related to aspiration, choking, and weight gain through an override of satiation markers. For those with autism, it is also likely to contribute to both indigestion and ongoing social isolation. New research on this topic has documented a relationship between a high body fat ratio and rapid eating and hypothesized a relationship due to insulin resistance.
In designing treatment modalities for all populations that address rapid consumption, the challenge lies in overcoming the fact that the food of choice is itself reinforcing. Thus, the faster an individual eats, the faster they consume, thereby shaping rapid consumption. Typical strategies for treatment on the spectrum often involve placing support staff so as to actively coach a reduction in the speed of bites.81 Newer, more innovative strategies utilizing technology have incorporated the use of a vibrating pager to cue consumption at fixed intervals. This approach has been found to decrease the rate of eating in teenagers on the spectrum72 and in several case studies of adolescents76,77 but is not currently in widespread use.
Rapid eating as manifested in binge eating disorder (BED) was first included in the DSM-5, and is now the most common eating disorder in the United States. From the perspective of a formal diagnosis, BED is considered to be idiopathic, and is defined by an absence of compensatory behaviors such as co-occurring AN or BN.
The formal BED diagnostic criteria demonstrate the difficulties of applying DSM-5 language to individuals with autism, even when the behavior is highly prevalent. Due to communication deficits and alexithymia, applicability of stipulated criteria for BED would necessitate translating subjective symptoms to objective signs. By adjusting these criteria (see Table 1) it is possible to create applicable standards that capture behaviors manifested in those with autism. The only exception to this is the identification of “disgust, depression or guilt,” which would be impossible for many individuals on the spectrum to identify. Likewise, embarrassment for many individuals would be impossible to label, but stealing food is a commonly noted behavior and can be another means to come at this criteria.
Table 1: Transforming DSM Language into Applicable Criteria for Autism
- Eating more rapidly than is normal
- Eating until uncomfortably full
- Eating when not hungry or when full
- Feelings of disgust, depression or guilt post-consumption
- Eating alone due to embarrassment
Altered Objective Criteria to Address Autism
- Eating rapidly
- Eating whenever food is available
- Eating too much
- Cannot be adjusted
- Stealing food
Though Cognitive Behavior Therapy has found widespread acceptance for treatment of BED within the general public, it has only been used with the most highly functioning cohort on the autism spectrum. Even here, subjects were found to have difficulty in grasping cognitive restructuring, and organizing to various sections. Doubts and concerns were also raised regarding maintenance of behavior changes and generalizing abilities.
A Challenge for ED Professionals
This broad view of disordered eating on the autism spectrum makes clear that these behaviors are widely manifested and require significant resources in terms of staff time and attention. Despite this, clinical understanding in terms of assessment and treatment remains limited. Disordered eating on the autism spectrum is clearly a new horizon for ED professionals, who have the skills and knowledge to make a tremendous contribution to this at-risk population.