The Overlooked Burden of Atypical Anorexia Nervosa: Commentary on Melville et al. (2025)

ABSTRACT

Obesity and eating disorders (EDs) have historically been viewed as distinct conditions; however, emerging evidence suggests a significant overlap, particularly among individuals seeking obesity treatment. While binge‐eating disorder (BED) is commonly identified in this population, restrictive EDs such as atypical anorexia nervosa (atypical AN) can go largely undetected. This paper comments on findings from Melville et al.'s systematic review of 85 studies assessing ED prevalence in adults with high Body Mass Index (BMI) seeking obesity treatment. We highlight the striking absence of atypical AN diagnoses despite substantial evidence supporting its prevalence in broader populations. We explore several reasons for this under‐recognition, including the definitional ambiguities of atypical AN in the DSM‐5, limitations of assessment tools that emphasize binge eating, and weight stigma that tends to mask restrictive eating as “normal” dieting. The implications are significant: failure to identify atypical AN may lead to delayed or inappropriate care and reinforce harmful stereotypes that restrictive EDs only affect underweight individuals. We argue for greater clinical vigilance, the refinement and clarification of diagnostic criteria and the development of validated tools for detecting atypical AN, particularly in higher‐weight individuals. Clinicians, particularly those providing weight loss interventions, should be trained to identify restrictive eating irrespective of BMI and prioritize behaviors and psychological impairment over weight status. Recognizing atypical AN as a serious, underdiagnosed condition is critical to ensuring ethical, equitable and effective care across the weight spectrum, in both ED and weight‐loss treatment settings.

Abbreviations

AN

Anorexia Nervosa

Atypical AN

Atypical Anorexia Nervosa

BED

Binge Eating Disorder

BMI

Body Mass Index

DSM

Diagnostic and Statistical Manual of Mental Disorders

ED

Eating disorders

EDNOS

Eating Disorder Not Otherwise Specified

GLP‐1

Glucagon‐like Peptide

OSFED

Other Specified Feeding or Eating Disorder

UFED

Unspecified Feeding or Eating Disorder

1. Introduction

Obesity and eating disorders (EDs) have traditionally been viewed as distinct clinical entities. However, research highlights a complex overlap. Individuals seeking obesity treatment often present with disordered eating and frequently meet diagnostic criteria for EDs. We are also becoming increasingly aware that the whole range of EDs (both restrictive and binge‐eating disorders) occurs in people of all shapes and sizes, including those with a high Body Mass Index (BMI).

The Melville et al. (2025) systematic review and meta‐analysis represents an important step forward in investigations of EDs in people with high BMIs. Their review included 85 studies (each with a sample size > 325) published since 1985, which assessed the prevalence of EDs in adults presenting for treatment of overweight/obesity, the majority of which involved presentations for bariatric surgery (57.6%). Combined, these studies yielded a total sample size of 94,295 people with a median BMI of 46 kg/m²; mostly female (75.9%) and middle‐aged (median age = 44 years). When assessed via clinical interview, the pooled prevalence of Diagnostic and Statistical Manual for Mental Disorders, fifth edition (DSM‐5; American Psychiatric Association [APA] 2013) binge‐eating disorder (BED) was 14% (with predicted intervals ranging from 0% to 40%). When assessed using self‐report questionnaires, the prevalence of moderate severity binge eating was 28% (with predicted intervals ranging from 18% to 33%).

So about one quarter of these high‐weight patients had significant binge‐eating problems and one in seven met criteria for BED. These numbers are much higher than usually found in community samples (Hay et al. 2017). Other EDs (aside from those characterized by binge eating), however, were rarely detected. The prevalence of “other” EDs (such as Other Specified Feeding or Eating Disorders [OSFED], Unspecified Feeding or Eating Disorders [UFED], or Eating Disorders Not Otherwise Specified [EDNOS]) was 11% but there was no breakdown according to specific subtypes within these categories. Notably, atypical anorexia nervosa (atypical AN) was largely absent. Atypical AN was specifically assessed in only two of the 85 included studies; one, using a clinical interview, estimated its prevalence as 0.2% and the other found no cases at all.

The high prevalence of EDs involving binge eating in this sample of people with high BMIs is to be expected. What is surprising, however, is the very low prevalence of EDs with a restrictive profile—which in non‐underweight individuals would be classified as atypical AN. This contrasts with our clinical experience and with research suggesting that atypical AN accounts for a large proportion of OSFED, is highly prevalent in community samples, and is increasingly being diagnosed in patients presenting for ED treatment (Hay et al. 2017). Harrop et al. (2021) presented a systematic review of studies (published 2007–2020) that assessed for DSM‐5 atypical AN in epidemiological studies (n = 17) and studies of admissions to ED treatment centers (n = 58). Epidemiological studies estimated the point prevalence of atypical AN at 0.15%–13%, which is more prevalent than low weight anorexia nervosa (AN) in the general community; and in almost half of the clinical studies, atypical AN cases made up at least 20% of admissions. As the authors noted, these rates are likely underestimates because the vast majority of the studies reviewed involved samples limited to white, adolescent, non‐overweight females.

The “missing” atypical ANs in the Melville et al. study is troubling because their results may reinforce the incorrect view that restrictive EDs are rare in people with a high BMI whereas, in reality, atypical AN often remains unrecognized, especially in people with higher BMIs (and also in older and non‐white people) (Harrop et al. 2021; Hay et al. 2017). This is problematic because research has indicated that atypical AN presents with medical and psychological consequences at least as severe as those associated with AN (Walsh et al. 2023). However, due to not being low‐weight, people with atypical AN may not receive a timely diagnosis or appropriate care; indeed, their weight loss behaviors may be praised or encouraged (Harrop et al. 2021). We can identify several reasons why atypical AN might have been overlooked in the Melville et al. review.

1.1. The Definition of Atypical AN

The first reason concerns both the timing and the wording of the current definition of atypical AN. This diagnosis was introduced into the nosology of EDs in 2013 (APA 2013) with various different terms used for it prior to then. Since almost one third of the studies included in the Melville et al. review were published before 2013, atypical AN would only have been captured as EDNOS rather than atypical AN. In DSM‐5, atypical AN is a subtype of OSFED, describing people who meet all criteria for AN and who have lost “significant” weight yet remain at or above the “normal” BMI range (i.e., 20–25) (APA 2013). Specifically, individuals with atypical AN engage in caloric restriction, describe intense fear of weight gain, and experience high weight/shape concerns, yet their weight remains above the underweight threshold specified for a diagnosis of AN. This group may include people whose bodies have resisted weight loss despite caloric deprivation (due to their individual physiological profile); those in larger bodies who have lost weight but not reached a BMI considered to be “low”; and those for whom a pattern of weight cycling (weight loss followed by regain) has contributed to their BMI increasing over time (Walsh et al. 2023). Importantly, there is ambiguity and little consensus about what constitutes “significant” weight loss, and the distinction between having lost weight versus being weight suppressed has not been considered.

Various interpretations of “significant weight loss” have been proposed. Forney et al. (2017) categorized participants in an epidemiological study according to three definitions of “significant weight loss” (5%, 10%, and 15%) and compared them to controls and to people diagnosed with an ED on measures of eating pathology. They found that all categories of weight loss (in both men and women), in the presence of cognitive symptoms (fear of weight gain, body image disturbance), were associated with elevated eating pathology and distress compared to controls. So even a relatively small amount of weight loss, combined with cognitive symptoms, was clinically impairing and this was observed in both “normal” weight and above‐average‐weight individuals. Importantly, it was the combination of any weight loss and ED cognitions that distinguished between weight loss per se (without accompanying cognitive distortions) and cognitive distortions alone (in the absence of weight loss) suggesting that this way of conceptualizing atypical AN may avoid over‐pathologizing weight loss more generally. It is likely that a pattern of chronic restriction (and perhaps even attempted, but unsuccessful, weight loss) is the key feature of atypical AN, whether or not binge eating is also present; it is the restriction that drives the binge eating cycle.

1.2. The Measures Used

Many of the studies included in the Melville et al. review used clinical interviews or self‐report measures that, for the most part, allowed for the concrete counting of binge eating episodes versus a more general or conceptual measure of dietary restraint. This means that binge eating, a specific, quantifiable behavior, is more likely to be identified than dietary restraint, which may be reflected in specific subscale scores but not noted as a discrete or overt behavior. Therefore, the frequency of binge eating is highlighted, but the fact that it occurs in the context of dietary restriction may be overlooked. These measurement issues reflect the broader problem of confirmation bias; binge eating is expected in high‐weight people, whereas dietary restraint is not. High‐weight individuals are assumed to be over‐consuming, as opposed to restricting, so signs of dietary restraint may go unnoticed or be dismissed.

1.3. Weight Stigma and Clinician Knowledge

This leads on to another barrier to identifying atypical AN in people of higher‐weight: weight stigma. Individuals with a high BMI are often encouraged to pursue weight loss by health professionals and are reinforced for weight loss (Harrop et al. 2021). The normalization of dieting in people with high BMIs may mean that dietary restraint is viewed as a marker of “discipline” rather than of pathology. This results in a bias toward identifying restrictive EDs only in low weight individuals and not in those with higher BMIs. Given that research into higher‐weight restrictive EDs has only really been spurred on by the advent of DSM‐5, it is also possible that clinicians lack awareness of the prevalence of atypical AN presentations in higher‐weight populations, especially post‐bariatric populations, or underestimate the medical and psychological harm in patients with atypical AN. Importantly, most studies in Melville et al. pre‐date the emergence of the use of glucagon‐like peptide (GLP‐1) agonists for obesity treatment. The impact of GLP‐1s on the ED field remains to be seen, however, we would expect atypical AN to be prevalent in individuals seeking these treatments. Another relevant source of bias is that EDs in middle‐aged people tend to be overlooked, yet we know that elevated prevalences of atypical AN have been observed in older samples (Hay et al. 2017).

2. Implications

The “missing” cases of atypical AN in Melville et al. (2025) supports our fears that atypical AN is likely underdiagnosed in higher‐weight populations and this has significant implications for clinical practice and research. In clinical settings, especially obesity treatment settings, clinicians should screen for EDs using validated tools, regardless of BMI. Such screening efforts would now include General Practitioners or Primary Care Physicians, given the surge in their prescribing of GLP‐1s for obesity treatment. Particular attention should be paid to weight loss attempts and percentage of weight lost (rather than to weight alone) and restrictive eating as well as to psychological symptoms such as intense fear of weight gain and high levels of body dissatisfaction. A narrow focus on weight loss in the absence of careful screening for dietary restriction may both exacerbate an ED and delay timely intervention. Clinicians should be trained to identify warning signs in higher‐weight individuals presenting with a background of overvaluation of weight/shape, recurrent dieting or weight loss attempts, or actual weight loss (even a minimal amount). Conceptualizing an ED as primarily a restrictive disorder rather than a binge‐eating disorder will influence the formulation of the ED and, therefore, the focus of treatment. The emphasis in atypical AN will be on “letting go” of the restriction, rather than “getting rid of” the binge eating.

With regard to research, first, the definition of atypical AN needs to be clarified. Questions to be addressed include: whether we continue to rely on numerical BMI cut‐points or estimates of “significant weight loss” in restrictive EDs or, instead, rely more on broader indicators of severity or impairment such as percentage weight lost, weight suppression, weight history, physiological disturbance, or psychological distress. Some researchers have suggested obviating the weight criteria for AN altogether or combining atypical AN and AN into one category since they are associated with similar levels of psychopathology and physical impairment (for review, see Walsh et al. 2023). This might help to reduce weight stigma. It would be helpful to have parameters around what constitutes “significant weight loss” in the absence of low body weight and the characteristics that would need to be present to differentiate atypical AN from other EDs or from weight loss in the absence of an ED. Second, we need greater inclusion of atypical AN in prevalence studies of higher‐weight populations, especially people seeking treatment for obesity. To provide accurate data it is important that we develop sensitive, validated tools for assessing atypical AN in people with higher BMIs.

3. Conclusion

Despite growing interest, atypical AN in higher‐weight individuals represents a critical blind spot in the ED field. Failure to recognize restrictive EDs in higher‐weight individuals can lead to significant harm (Forney et al. 2017). Clinicians and researchers need to adopt informed, flexible, and inclusive practices to assess atypical AN, and diagnostic systems should provide more empirically supported definitions. We must move beyond BMI as the sole indicator of health and recognize the nuanced presentation of EDs across the weight spectrum. Better recognition and management of atypical AN in high‐weight people will enhance care for a vulnerable and often overlooked group.

Author Contributions

Susan M. Byrne: conceptualization, investigation, writing – original draft, writing – review and editing. Jessica McClelland: conceptualization, investigation, writing – review and editing. Anthea Fursland: conceptualization, writing – review and editing.

Ethics Statement

The authors have nothing to report.

Consent

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Byrne, S. M. , McClelland J., and Fursland A.. 2025. “The Overlooked Burden of Atypical Anorexia Nervosa: Commentary on Melville et al. (2025).” International Journal of Eating Disorders 58, no. 10: 1907–1910. 10.1002/eat.24507.

Data Availability Statement

The authors have nothing to report.