In October 2018, a team of researchers led by Hunna Watson at the University of North Carolina published results from the largest genome-wide association study ever conducted on anorexia nervosa. The study, which appeared in Nature Genetics the following year under the full authorship of the Anorexia Nervosa Genetics Initiative (Watson et al. 2019), analyzed 3,495 cases of anorexia nervosa and 10,982 controls drawn from 17 countries. It identified eight genomic loci significantly associated with anorexia at a genome-wide threshold. That finding alone would have been significant — the first robust GWAS results for the disorder. But the study's most striking and unexpected contribution was a pattern of genetic correlations. Anorexia nervosa was positively correlated, at the genetic level, not only with psychiatric conditions such as schizophrenia, obsessive-compulsive disorder, and anxiety disorders — all expected, given the clinical overlap — but also with metabolic traits. Anorexia correlated negatively with body mass index, fat mass, and type 2 diabetes. It correlated positively with HDL cholesterol and physical activity measures. The disorder that had been treated for decades primarily as a cultural and psychological condition — caused by thin-ideal media, family dysfunction, or distorted body image — turned out to have a genetic architecture that overlapped with metabolic regulation. This was not a subtle finding. It implied that anorexia nervosa is not purely a psychiatric condition but one in which altered metabolic biology participates in a fundamental way.
The Watson findings did not emerge into a vacuum. The cultural narrative around eating disorders — that they are caused by fashion magazines, celebrity culture, and a media environment that valorizes thinness — had already been complicated by twin studies. Sullivan et al. (1998) had analyzed female twins and found heritability estimates for anorexia nervosa of approximately 58%. Klump et al. (2001) found similar estimates in the 50-60% range for both anorexia and bulimia nervosa. These figures are comparable to the heritability of schizophrenia, bipolar disorder, and ADHD — conditions nobody disputes are substantially biological. The genetic evidence did not deny that cultural and social factors matter. But it reframed their role. Cultural pressure to be thin does not cause anorexia in someone without the neurobiological and genetic substrate for it. It modulates and amplifies risk in people who already carry that substrate. The distinction is not academic — it changes what treatment looks like and where responsibility lies.
Anorexia nervosa has the highest mortality rate of any psychiatric disorder. Approximately 5-10% of patients die from it — some from the medical consequences of prolonged starvation, including cardiac failure, electrolyte imbalance, and multi-organ dysfunction, and others by suicide. It has among the lowest rates of treatment-seeking, one of the highest rates of relapse, and one of the most stigmatized presentations in psychiatric medicine. Understanding what actually causes it — and what maintains it — is not a matter of academic interest alone.
"Anorexia nervosa is not a lifestyle choice or a diet gone wrong. It is a serious and complex psychiatric disorder with a strong biological basis, embedded in cultural and social contexts that can precipitate and perpetuate it." — Cynthia Bulik, Annual Review of Clinical Psychology (2021)
Key Definitions
Anorexia nervosa (AN): A psychiatric disorder characterized by persistent restriction of energy intake leading to significantly low body weight; intense fear of weight gain or persistent behavior interfering with weight gain; and disturbance in how body weight or shape is perceived. Carries the highest mortality rate of any psychiatric disorder at 5-10%.
Bulimia nervosa (BN): Recurrent episodes of binge eating followed by compensatory behaviors such as self-induced vomiting, laxative misuse, fasting, or excessive exercise. Must occur at least once a week for three months. Lifetime prevalence approximately 1.5% in women, 0.5% in men.
| Risk Factor Category | Examples | Evidence Strength |
|---|---|---|
| Genetic | Family history of eating disorder, heritability 40-60% | Strong |
| Psychological | Perfectionism, low self-esteem, anxiety, trauma history | Strong |
| Sociocultural | Thin-ideal media exposure, weight-related teasing, diet culture | Moderate-Strong |
| Developmental | Puberty-related body dissatisfaction, early dieting | Moderate |
| Neurobiological | Altered serotonin and dopamine signaling, interoception differences | Moderate — growing evidence |
Binge eating disorder (BED): The most prevalent eating disorder (lifetime prevalence 3.5% in women, 2% in men); recurrent binge eating without compensatory purging, accompanied by significant distress. Formally recognized in DSM-5 in 2013.
ARFID (avoidant/restrictive food intake disorder): Food restriction driven not by body image concerns but by sensory sensitivities, fear of aversive consequences such as choking, or limited interest in eating. Distinct from AN diagnostically and mechanistically.
Heritability: The proportion of variance in a trait attributable to genetic differences. Heritability estimates of 50-60% for AN and BN indicate that roughly half the risk variation is genetic — comparable to schizophrenia and bipolar disorder.
Refeeding syndrome: A potentially life-threatening metabolic complication during nutritional rehabilitation, characterized by electrolyte shifts — particularly hypophosphatemia — as cells shift from fat to carbohydrate metabolism.
Thin ideal internalization: The degree to which an individual adopts culturally defined standards of thinness as a personal standard; a documented risk factor for body dissatisfaction and eating disorder symptoms.
Weight stigma: Negative attitudes, stereotypes, and discrimination based on body weight; operates as both a risk factor for eating disorder development and a barrier to treatment-seeking.
The Diagnostic Landscape
The DSM-5 recognizes a family of eating and feeding disorders that share some mechanisms but differ substantially in presentation, course, and treatment response.
Anorexia Nervosa
DSM-5 criteria require three elements: persistent restriction of energy intake relative to requirements resulting in significantly low body weight; intense fear of gaining weight or persistent behavior that interferes with weight gain; and disturbance in the way one's body weight or shape is perceived. Three presentations are recognized: restricting type, in which weight control is achieved through fasting, dieting, or excessive exercise; binge-purge type, in which binging and purging also occur; and atypical anorexia, in which all features are present but weight remains within the normal range — arguably the most underdiagnosed presentation, since clinicians and family members may not recognize the seriousness of the disorder when a patient is not visibly underweight.
Lifetime prevalence is approximately 0.9% in women and 0.3% in men, though these figures likely underestimate the true burden given systematic underdiagnosis in men, in non-Western populations, and in patients who do not fit the stereotypical demographic profile. Peak onset is typically between ages 14 and 18, though onset can occur at any age. The standardized mortality ratio is approximately 5-6, meaning people with AN die at five to six times the rate expected for their age and sex. This makes AN the deadliest psychiatric condition in existence.
Bulimia Nervosa
Bulimia nervosa involves recurrent binge eating — consuming unusually large amounts of food in a discrete period, with a sense of loss of control — followed by compensatory behaviors to prevent weight gain. Self-induced vomiting is most common; others include laxative misuse, fasting, and compulsive exercise. Body weight is typically in the normal or overweight range, meaning BN is often invisible and frequently goes undetected for years. Medical consequences include electrolyte imbalances (hypokalemia causing cardiac arrhythmia), dental erosion, and esophageal damage.
Binge Eating Disorder
BED is the most prevalent eating disorder. Recurrent binge episodes are associated with eating much more rapidly than normal, eating until uncomfortably full, eating when not physically hungry, eating alone due to embarrassment, and experiencing significant distress afterward — guilt, disgust, or depression. Unlike BN, there are no regular compensatory behaviors. BED is frequently comorbid with obesity, depression, and anxiety disorders, and responds well to structured psychological treatment, particularly cognitive behavioral approaches.
ARFID
ARFID — avoidant/restrictive food intake disorder — involves restriction not driven by body image concerns but by sensory aversions, fear of aversive consequences such as choking or vomiting, or limited interest in eating. It occurs across the age range and is particularly prevalent in autism spectrum disorder. Because the mechanism is fundamentally different from AN, standard AN treatments are inappropriate for ARFID.
The Genetic Evidence
Twin studies established the heritability of eating disorders before any specific genetic variants were identified. Sullivan et al. (1998), analyzing female twins in the Virginia Twin Registry, estimated heritability for anorexia nervosa at approximately 58%, with the environmental variance almost entirely attributable to individual-specific rather than shared family environment. Klump et al. (2001) extended this to BN and eating disorder symptoms broadly, finding heritability estimates in the 50-60% range. For BED, heritability estimates range from 41-57%. These figures are comparable to schizophrenia and bipolar disorder — conditions understood as fundamentally biological — yet AN's heritability has not historically been treated with equal seriousness.
The Watson et al. (2019) GWAS study (doi: 10.1038/s41588-019-0439-2) identified eight genomic loci significantly associated with AN at genome-wide significance thresholds — the first such finding in the field. The genetic correlation results were the study's most theoretically important contribution. AN showed significant correlations with schizophrenia, OCD, and anxiety disorders — the expected psychiatric dimensions — but also with metabolic traits: negative correlations with BMI, fat mass, and insulin levels, and positive correlations with HDL cholesterol. This "metabo-psychiatric" architecture suggests that anorexia is not purely psychiatric but involves genuine metabolic biology. Some individuals may be predisposed to lower body weight set points, and AN may represent an extreme expression of this tendency under psychological and social pressure. If the biology of AN involves altered metabolic regulation, weight restoration cannot be the end of treatment — it is at best the beginning.
The Neurobiology of Eating Disorders
Beyond genetics, neuroimaging and neurochemical research has identified consistent, reproducible differences in brain function in people with eating disorders. Critically, many of these differences persist after recovery, indicating they are traits rather than consequences of malnutrition or the disorder's active phase.
Serotonin System Alterations
The serotonin system regulates mood, anxiety, and impulse control and shows clear, persistent alterations in both AN and BN. Kaye et al. (2005), published in Biological Psychiatry (doi: 10.1016/j.biopsych.2004.04.013), used positron emission tomography to measure serotonin 5-HT1A receptor binding in women who had recovered from AN — women who were weight-restored and no longer met diagnostic criteria. Recovered AN patients showed significantly elevated 5-HT1A receptor binding in limbic regions including the amygdala, anterior cingulate cortex, and medial prefrontal cortex compared to healthy control women. Because these differences were present after recovery and not attributable to malnutrition, they almost certainly represent a premorbid biological trait. Elevated 5-HT1A signaling in these limbic regions is associated with increased anxiety and heightened harm avoidance — both core premorbid characteristics of AN.
In BN, serotonin dysregulation presents oppositely — decreased activity associated with impulsivity — explaining why SSRIs show efficacy for BN but not AN.
Dopamine and Reward Processing
The dopamine reward system shows consistent alterations in AN. Cowdrey et al. (2012), in a neuroimaging study published in Biological Psychiatry, compared reward-related brain activation in women with AN to healthy controls using functional MRI. Women with AN showed reduced activation in reward-related regions including the caudate and putamen in response to food images. Palatable food — which activates strong reward circuitry in healthy individuals — did not produce the same neural reward response in AN patients. Eating, in AN, does not feel rewarding. It feels threatening.
This finding has a critical clinical implication: the behavioral goal of increasing eating cannot be achieved simply by making food more appealing, because the problem is not the food but the altered neural response to it.
The Anxiety-Reduction Loop
The most clinically important neurobiological mechanism in AN is the anxiety-reduction loop. In healthy individuals, eating reduces anxiety. But in AN the relationship is inverted: eating raises anxiety, often dramatically, while restriction provides acute relief.
This pattern has been documented in multiple studies measuring cortisol levels and self-reported anxiety in AN patients before and after meals, compared to healthy controls. AN patients consistently show elevated post-meal anxiety, elevated cortisol, and distress that can last for hours after eating. This means that food restriction in AN is not simply stubbornness, vanity, or a desire to be thin. It is, at a neurobiological level, the primary anxiety management strategy. Restriction is reinforced by the immediate and reliable relief it provides. Understanding this mechanism is essential for understanding both why AN is so treatment-resistant and why weight restoration alone is insufficient — the underlying anxiety circuitry must be addressed concurrently.
Cultural and Social Factors
Social and cultural factors are genuine contributors to eating disorder risk, and the evidence for their role is real. The question is not whether they matter but how they interact with the genetic and neurobiological substrate.
The thinness ideal in Western cultures creates an environment in which body dissatisfaction is widespread, dieting is normative, and restriction is socially reinforced. Thin ideal internalization — the degree to which an individual endorses cultural standards of thinness as their own personal standard — is a documented risk factor for eating disorder symptoms. Stice et al. (2001), published in the Journal of Consulting and Clinical Psychology, demonstrated in a randomized controlled trial that a media literacy intervention reducing thin ideal internalization significantly decreased body dissatisfaction and eating disorder symptoms in high-risk women.
The Fiji study conducted by Anne Becker and colleagues (2002) provided a natural experiment. Fiji had a cultural tradition valuing larger body size and had minimal Western media exposure. After the introduction of television in 1995, eating disorder behaviors increased significantly among adolescent girls over a 38-month follow-up period, and body dissatisfaction became more common. The study demonstrated that media exposure can activate eating disorder risk in a population where it had been low.
But the evidence also shows clear limits on purely cultural explanations. Anorexia has been documented in cultures without Western media penetration. Historical case descriptions that predate the thin-ideal media environment — including William Gull's 1873 clinical reports of "anorexia nervosa" in Victorian England — show presentations clinically indistinguishable from modern AN. AN occurs in men and boys who face different cultural pressures. And critically, the majority of people intensively exposed to thin-ideal media do not develop eating disorders. Cultural factors modulate risk in genetically vulnerable individuals; they do not create eating disorders in those without the underlying neurobiological substrate.
Weight stigma operates as both a risk factor and a treatment barrier. Experiencing weight-based teasing and bullying in childhood is a specific predictor of eating disorder development. Clinicians who express approval of weight loss without asking how it was achieved contribute to delayed diagnosis, particularly in patients who do not fit the stereotype of AN (men, patients with atypical AN, patients from non-white backgrounds). Black, Hispanic, and Asian patients experience eating disorders at comparable rates to white patients but are significantly less likely to be diagnosed, referred to treatment, or included in research samples — a systemic failure with serious consequences for outcomes.
Family, Developmental, and Personality Factors
Earlier theoretical models placed considerable weight on family dynamics as causal. Salvador Minuchin's psychosomatic family model (1970s) characterized AN families as enmeshed, overprotective, and conflict-avoidant, with the eating disorder serving a homeostatic function. This model is now largely discredited as a causal account. Parents of eating-disordered children are not more pathological than other parents, and blaming families adds harmful guilt to an already crisis-laden situation.
The evidence has moved in a different direction. Family involvement in treatment — far from being the problem — actually improves outcomes, particularly for adolescents. Lock et al. (2010), in a landmark randomized controlled trial published in Archives of General Psychiatry (doi: 10.1001/archgenpsychiatry.2010.128), randomized 121 adolescents with AN to either Family-Based Treatment (FBT) or individual adolescent-focused therapy. FBT produced significantly higher rates of full remission at 12-month follow-up. The key mechanism is that FBT places parents in charge of refeeding during Phase 1, recognizing that the patient lacks insight and that weight restoration is the prerequisite for psychological recovery — not the other way around.
Premorbid personality traits are stronger predictors of AN than family structure. Perfectionism, anxiety, harm avoidance, and obsessionality are consistently identified in retrospective studies as characteristics predating disorder onset. These traits are themselves heritable. A child who is anxious, harm-avoidant, and perfectionist — in a cultural environment that valorizes thinness and dietary control — is at elevated risk regardless of family dynamics.
Childhood adversity increases eating disorder risk through general pathways to psychiatric disorder; weight-based teasing is more specifically associated with eating disorder development than other forms of adversity.
Treatment Evidence
Family-Based Treatment for Adolescent Anorexia
FBT, developed at the Maudsley Hospital in London and manualized by James Lock and Daniel Le Grange, is the treatment with the strongest evidence for adolescent AN. It proceeds in three phases: Phase 1 externalizes the illness — treating it as an entity that has invaded the family — and charges parents with the task of refeeding their child; Phase 2 gradually returns control over eating to the adolescent as weight is restored; Phase 3 addresses the developmental issues of normal adolescence that were interrupted by the illness. The Lock et al. (2010) RCT and subsequent trials have consistently shown superiority of FBT over individual approaches for adolescents.
The FBT model reflects an important insight: in severe adolescent AN, weight restoration cannot be entrusted to the patient, whose judgment is compromised by the disorder. Parents are enlisted as therapeutic agents, not identified as contributors to the problem.
Cognitive Behavioral Therapy-Enhanced
CBT-E, developed by Christopher Fairburn of Oxford University and described in his 2008 manual Cognitive Behavior Therapy and Eating Disorders, is the most evidence-based treatment for bulimia nervosa and binge eating disorder. Unlike standard CBT, CBT-E addresses the transdiagnostic core psychopathology of eating disorders: the overvaluation of eating, shape, and weight as the primary basis of self-evaluation. This cognitive disturbance is the engine that drives dietary restriction, which in turn increases binge risk. CBT-E targets not only the behavioral symptoms but the cognitive architecture that maintains them.
The treatment addresses four maintaining mechanisms: dietary restriction, which drives binge episodes; mood intolerance, which triggers emotional eating; perfectionism; and low self-esteem. For BN, remission rates in RCTs are approximately 30-40% at end of treatment with continued improvement at follow-up.
Treatment for Adult Anorexia
Adult AN remains one of the most urgent unsolved problems in psychiatry. No psychotherapy has demonstrated consistently strong efficacy in RCTs for adults. Specialist supportive clinical management, cognitive remediation therapy, and CBT adaptations show modest positive effects. Pharmacotherapy has limited evidence; olanzapine produces modest weight gain without addressing core psychopathology. Deep transcranial magnetic stimulation targeting the dorsolateral prefrontal cortex has produced preliminary positive results.
Medical Management
Severe AN requires careful medical management throughout treatment. Refeeding syndrome — electrolyte shifts (particularly hypophosphatemia) during nutritional rehabilitation — is potentially life-threatening and requires monitoring. Bone density loss from prolonged AN is common and not fully reversible. Cardiac complications including bradycardia and QT prolongation are significant contributors to mortality.
Recovery: What the Evidence Shows
Recovery from eating disorders is possible but variable. For AN, long-term follow-up studies suggest that approximately 50-60% of patients eventually achieve full recovery, defined as both weight restoration and psychological recovery — resolution of body image disturbance and normalization of anxiety around food. Approximately 30-40% achieve partial recovery, and 10-20% develop a chronic course. Full psychological recovery typically lags behind physical recovery by years. This lag is consistent with the neurobiological evidence that some brain differences in AN persist after weight restoration and appear to be premorbid traits.
Bulimia nervosa has better recovery rates: approximately 50-70% of patients achieve full recovery with appropriate treatment, though relapse rates are significant in the first year. BED has the highest recovery rates, with many patients achieving full remission with CBT-E or structured intervention.
Weight stigma is a significant barrier to treatment-seeking and recovery. Patients who experience stigma in healthcare settings are less likely to disclose symptoms, remain in treatment, or avoid relapse. The cultural equation of weight loss with health creates particular danger for patients in larger bodies whose dangerous restriction may be praised by clinicians who see only the weight change rather than the method.
Prevention programs with the strongest evidence target thin ideal internalization. The Body Project, developed by Eric Stice and colleagues, uses cognitive dissonance techniques — participants argue against the thin ideal, making it harder to subsequently maintain that belief. Multiple RCTs demonstrate reductions in thin ideal internalization and eating pathology maintained at 2-3 year follow-up.
Racial and Gender Disparities in Diagnosis
Eating disorders affect all demographic groups, but systematic disparities in who receives diagnosis and treatment are well documented and have serious consequences. Men are diagnosed with eating disorders at approximately one-third the rate of women, despite substantial male prevalence (0.3% AN, 0.5% BN, 2% BED). Male eating disorders are substantially underdiagnosed: screening tools were developed on female samples, and male presentations often emphasize muscularity and leanness over thinness. Black, Hispanic, and Asian individuals experience eating disorders at broadly comparable rates to white individuals but are significantly less likely to be diagnosed, referred, or included in research. Clinicians demonstrate lower rates of identifying symptoms in non-white patients even with equivalent presentations. Athletes in weight-relevant sports (gymnastics, wrestling, rowing, distance running), LGBTQ+ individuals, and high-achievement perfectionist personalities across all demographic groups show elevated rates.
Cross-References
References
Watson, H. J., Yilmaz, Z., Thornton, L. M., Hubel, C., Coleman, J. R. I., Gaspar, H. A., et al. (2019). Genome-wide association study identifies eight risk loci and implicates metabo-psychiatric origins for anorexia nervosa. Nature Genetics, 51(8), 1207-1214. doi:10.1038/s41588-019-0439-2
Kaye, W. H., Frank, G. K., Bailer, U. F., Henry, S. E., Meltzer, C. C., Price, J. C., et al. (2005). Serotonin alterations in anorexia and bulimia nervosa: New insights from imaging studies. Biological Psychiatry, 56(11), 794-802. doi:10.1016/j.biopsych.2004.04.013
Lock, J., Le Grange, D., Agras, W. S., Moye, A., Bryson, S. W., & Jo, B. (2010). Randomized clinical trial comparing family-based treatment with adolescent-focused individual therapy for adolescents with anorexia nervosa. Archives of General Psychiatry, 67(10), 1025-1032. doi:10.1001/archgenpsychiatry.2010.128
Cowdrey, F. A., Park, R. J., Harmer, C. J., & McCabe, C. (2011). Increased neural processing of rewarding and aversive food stimuli in recovered anorexia nervosa. Biological Psychiatry, 70(9), 736-743. doi:10.1016/j.biopsych.2011.05.028
Sullivan, P. F., Bulik, C. M., Fear, J. L., & Pickering, A. (1998). Outcome of anorexia nervosa: A case-control study. American Journal of Psychiatry, 155(7), 939-946.
Klump, K. L., Miller, K. B., Keel, P. K., McGue, M., & Iacono, W. G. (2001). Genetic and environmental influences on anorexia nervosa syndromes in a population-based twin sample. Psychological Medicine, 31(4), 737-740.
Becker, A. E., Burwell, R. A., Gilman, S. E., Herzog, D. B., & Hamburg, P. (2002). Eating behaviours and attitudes following prolonged exposure to television among ethnic Fijian adolescent girls. British Journal of Psychiatry, 180(6), 509-514.
Fairburn, C. G. (2008). Cognitive Behavior Therapy and Eating Disorders. Guilford Press.
Stice, E., Mazotti, L., Weibel, D., & Agras, W. S. (2000). Dissonance prevention program decreases thin-ideal internalization, body dissatisfaction, dieting, negative affect, and bulimic symptoms: A preliminary experiment. International Journal of Eating Disorders, 27(2), 206-217.
Bulik, C. M. (2021). Towards a science of eating disorders: Replacing myths with realities. Annual Review of Clinical Psychology, 17, 1-26.