Revising the Diagnosis of Anorexia Nervosa Will Improve Patient Care

By Arnold E. Andersen, MD
University of Iowa Medical Center, Iowa City, IA
Reprinted from Eating Disorders Review
March/April 2006 Volume 17, Number 2
©2006 Gürze Books

Anorexia nervosa (AN), the oldest eating disorders subtype, was first described centuries ago. Its historical legacy has significant positives and negatives. The positives include improved diagnosis and research. The primary negative is hesitancy about changing historical thinking about AN, even though evidence-based studies suggest this is timely and imperative for improving patient care. Why have there been complexities in the diagnosis of AN?

Agreeing Upon a Definition for AN

An agreement on a definition of AN comes first. AN is: (1) a disorder of abnormal eating behavior, (2) driven primarily by the internalization of the overvalued sociocultural belief in the benefits of slimming (predominantly among females) or shape change (predominantly among males), (3) sufficiently sustained in duration and severity to cause significant signs and symptoms of medical starvation as well as psychological and social change.

It is important to understand the psychopathology of “overvalued beliefs.” They are: (1) widely held sociocultural beliefs, (2) given ruling passion in small group of individuals, often during crucial developmental years, (3) leading to risky or dangerous behaviors.

AN is both a strategy to deal with emotional distress from a variety of sources (for example, “existential fears of maturation,” depressive/anxious mood, family functioning, decreased self-esteem, social acceptance, etc.), and an illness that has a life of its own once it is established. In addition, it has perhaps the highest premature mortality (12% to 19%) of any psychiatric disorder. AN is not a subset of another disorder, such as obsessive-compulsive disorder or psychosis, or major depressive illness, although it usually has from two to four “companion” comorbid disorders, some secondary to AN, some primary to it.

Its hallmark psychopathological findings are: a morbid fear of fatness with a relentless drive for thinness, with frequent but not invariable distortion of body image. Many of its features, such as preoccupation with thoughts of food, emotional flattening or irritability, social isolation, and decreased sexual drive, are consequences of starvation, not part of the psychopathology of AN. These features have been reproduced in experimental starvation studies.

The high frequency of onset of AN during adolescence has primarily to do with the fact that the core challenge of adolescence is forming a personal identity, and slimming/shape change in this country are unfortunately very effective, albeit short-term, pseudo-solutions to forming a personal identity. In other words, just be thin and you will feel better about yourself; you can control something completely when everything is changing in ways you didn’t ask for; you get more attention; and you have a method to deal with crummy moods. Unfortunately, AN is a good enough pseudo-solution to seem to solve all the issues it is asked to solve. There are no dumb reasons for developing AN.

Getting the Diagnosis Right

Although AN has been described as rare, it is as common as schizophrenia or childhood onset of Type I diabetes. It is not a benign disorder, although it is highly treatable when done well. All of these are reasons to get the diagnosis right. The term “diagnosis” means “thorough knowledge.”

What kinds of knowledge are conferred by accurate diagnosis? An accurate diagnosis confers knowledge about: (1) what treatments to use; (2) what the future holds (prognosis, “future knowledge”). In the ancient world, doctors were paid to tell a parent the probable outcome (prognosis) of an injured child, even when treatments were not effective. So why is there concern about the diagnosis of AN?

First, the historical nature of AN seems to make clinicians reluctant to make changes, even as evidence accumulates. AN has long been associated with (a) being female—even though the first cases involved a male and a female; and (b) amenorrhea (having been classified in the endocrine section of texts for more than 50 years).

At least two studies have demonstrated that amenorrhea is irrelevant to the diagnosis of AN. Patients meeting the core AN criteria (self-starvation, overvalued belief in the importance of slimness, sustained functional impairment medically, psychologically, and socially), whether or not they have a light menstrual bleed or none have the same clinical picture, the same natural course of illness, and the same response to treatment. In addition, requiring amenorrhea for the diagnosis is gender-biased. In the ICD-10, the diagnostic classification used by most of the world, uses the phrase “abnormality of reproductive hormone functioning,” suggesting changes in estrogen or testosterone qualify as medical signs of AN. This is a better but still overly restricted appreciation of the global starvation changes that occur with AN. Requiring amenorrhea in the diagnosis of AN also ignores the data that some women lose periods soon after starting to diet, and some at very low weight have continued menstrual function. So, out with amenorrhea!

Misreading the DSM-IV

There is a slavish misreading of the DSM-IV‘s supposed criterion of requiring weight less than 85% of norms for age and height for the diagnosis of AN. Normal weight, like height, is bell-curved in its distribution. A woman who is healthy at a self-regulating weight of, for example, 125% of the statistical average, and who diets, may have all the medical, psychological, and social symptoms of AN at 90% of “normal” or “healthy” weight, all of which are averages, not mandates. A straightforward reading of the DSM-IV is frustratingly ignored. The use of the 85% level is said to be exempli gratia, or an example. It is unbelievable how rigidly insurance companies, clinicians, and the media interpret this level.

Many mischievous as well as serious patient-unfriendly results occur when insistence on amenorrhea and a weight less than 85% of an average are required as part of the diagnosis of AN. The huge number of so-called “atypical” cases of eating disorders are, in fact, 75% of the time, AN misdiagnosed by relying on disproved or archaic criteria. Female patients are told if they have menstrual function and/or final diet-induced weight greater than 85% that they do not have AN, but rather have an “atypical eating disorder.” This confuses clinicians (“How do you treat atypical cases?”), causes frequent denial of health insurance payment for diagnosis evaluation or treatment (“We only reimburse AN and bulimia nervosa; also, atypical cases are not that serious.”). These things shouldn’t happen, but they do time and time again.

Resistance to Change

In the real world, there are some awkward factors to deal with in understanding resistance to change in diagnostic criteria. Some resistance to change comes from researchers invested in keeping overly narrow criteria because of the possible need to change diagnoses in research studies; emotional investments in archaic criteria, albeit out of date, are at times a personal reality. Politics as well as science affect the criteria in differing degrees (for example, diagnostic categories that had a brief half-life in previous Diagnostic and Statistical Manual editions).

More and more disorders are appreciated as being spectrum disorders, with a change from single rigid category on the dimension (high blood pressure vs. no high blood pressure) to a spectrum of severity. Now, a person with pre-hypertension is treated with the same vigor and method as the old “full” hypertensive.

Likewise, if a patient has the core features of AN: (1) self-induced starvation; (2) psychopathology of an overvalued belief in the benefits of slimming or shape change; (3) duration and severity of disorder sufficient to suffer functional impairment medically, psychologically, and socially, they have AN and need to be diagnosed as such, and treated as such.

Asthmatic patients are not told to go home and become more severely asthmatic before they qualify for treatment. Likewise, AN patients do not stop having AN because they exceed 85% of their normal weight or at normal weight, any more than a hypertensive or diabetic patient loses his diagnosis because medical treatment produces normal blood pressure or normal glucose levels. If, after five years there are no signs of AN (“normative cultural distress,” i.e., lip service to “dieting is normal”), then AN, like cancer, can be considered probably cured.

Thankfully, a “transdiagnostic” approach to AN at any level of severity, and whether or not they meet any current state-of-the-art evolving diagnostic criteria (DSM-IV or ICD-10), has been shown to lead to clinical improvement in the large majority of patients if applied integratively and for long enough to “work the disorder out of a job.” The myths of AN being always chronic, always severe, hard to treat, difficult to diagnose, etc., are all completely untrue. These myths have all the features of disorders clinicians most like to treat: (1) serious enough to absolutely require treatment; (2) responsive enough to have a high likelihood of “cure,” “remission,” not just improvement (up to 76% in the best-outcome series); (3) challenging enough in research and treatment in both the psychosocial and biomedical areas to provoke new research efforts.

All the answers about AN are not in. But a simpler, clearer approach to diagnosis will benefit many of these patients. As with appendicitis, a clinician cannot treat what is not first accurately diagnosed. A quiet, persistent, noninflammatory insistence on recognizing AN as AN will benefit patients, and eventually the diagnostic criteria will come around to recognizing the criteria that are evidence-based and proven, and which need to be updated.


  1. Garfinkel PE, Lin E, Goering P, et al. Should amenorrhea be necessary for the diagnosis of anorexia nervosa? Evidence from a Canadian community sample. Br J Psychiatry 1996; 168:500.
  2. Andersen AE, Bowers B, Watson T. A slimming program for eating disorders not otherwise specified: reconceptualizing a confusing, residual diagnostic category. In: Psych Clin N Am June 2001:271.
  3. Watson T, Andersen A. A critical examination of the amenorrhea and weight criteria for diagnosing anorexia. Acta Psychiatr Scand 2003;108:175.
  4. Strober M, Freeman R, Morrell W. The long-term course of severe anorexia nervosa in adolescents: survival analysis of recovery, relapse, and outcome predictors over 10-15 years in a prospective study, Int J Eat Disord 1997;22:339.
No Comments Yet

Comments are closed