Professional Documents
Culture Documents
clinical practice
Anorexia Nervosa
Joel Yager, M.D., and Arnold E. Andersen, M.D.
This Journal feature begins with a case vignette highlighting a common clinical problem.
Evidence supporting various strategies is then presented, followed by a review of formal guidelines,
when they exist. The article ends with the authors’ clinical recommendations.
Diagnostic criteria
Refusal to maintain body weight at or above a minimal normal weight for age and height (e.g., a weight loss resulting in
maintenance of body weight at less than 85 percent of the expected weight or failure to make the expected weight
gain during the period of growth, resulting in a body weight of less than 85 percent of the expected weight).
All the following are present:
An intense fear of gaining weight or becoming fat, even though underweight
A disturbance in the way in which body weight or shape is experienced, undue influence of body weight or shape on
self-evaluation, or denial of the seriousness of current low body weight
Among postmenarchal girls and women, amenorrhea (the absence of at least three consecutive menstrual cycles;
amenorrhea is considered to be present if menstrual periods occur only after the administration of hormone
[estrogen] therapy)
Type
Restricting type
The patient has not regularly engaged in binge-eating or purging behavior (self-induced vomiting or the misuse of
laxatives, diuretics, or enemas)
Binge eating and purging type or purging type
The patient has regularly engaged in binge-eating, purging behavior, or both
* Estimates of healthy weight for a given patient are determined by the physician on the basis of historical considerations
that often include the patient’s growth charts and, for women, the weight at which healthy menstruation and ovulation
resume, which may be higher than the weight below which menstruation and ovulation became impaired. Adapted from
the Diagnostic and Statistical Manual of Mental Disorders, fourth edition.6
of patients9,11) — more often among those with causes of death (ratio, 11.6), especially suicide (ra-
the binging–purging type of anorexia nervosa, tio, 56.9).18 Patients with the binging–purging type
among whom the rate of impulsive behavior is also of the disorder and also alcohol and substance
higher, than among those with the restricting type. abuse have a higher risk of premature death than do
Medical complications resulting from semistar- other patients with anorexia nervosa.19 Death from
vation, purging, or overexercising or a combination medical causes results primarily from starvation or
of these symptoms affect virtually every organ sys- purging-related arrhythmias.20 The probability of
tem. Common signs and symptoms include loss of recovery appears to vary inversely with the severity
subcutaneous fat tissue, orthostatic hypotension, of weight loss and the presence of coexisting psy-
bradycardia, impaired menstrual function, hair loss, chiatric disorders, including highly avoidant or im-
and hypothermia. Many laboratory measures may pulsive personality disorders.21
be affected, among them serum electrolyte levels Full recovery of weight, growth and develop-
and thyroid function. Among patients who have ment, menstruation, normal eating behavior and
anorexia nervosa in adolescence, medical compli- attitudes with regard to food, and body shape and
cations may persist into the adult years.12 Long- body weight occurs in 50 to 70 percent or more of
range concerns include osteopenia and osteoporo- treated adolescents. A prospective study of adoles-
sis; affected adolescents may have shorter stature cents who received comprehensive treatment found
than expected and high rates of stress fractures by that 76 percent no longer had a diagnosable eating
their mid-20s.13,14 Abnormalities in cognitive func- disorder at 10 years of follow-up.22 Attaining com-
tion may also occur. The brain loses both white and plete physical and psychological health may take
gray matter during severe weight loss as a result of five to seven years.23 Outcomes are poorer among
semistarvation; weight restoration results in the re- adults with anorexia; only 25 to 50 percent of adults
turn of white matter to premorbid levels, but some who require hospitalization recover.24
loss of gray matter persists15 that may be associat-
ed with long-term effects on cognitive functioning. strategies and evidence
Women who have had anorexia nervosa also have
higher rates of miscarriage and lower infant birth diagnosis
weights than do healthy women.16 The diagnosis of anorexia nervosa is made on the
Standardized mortality ratios (comparing mor- basis of history taking (including information from
tality among persons with anorexia nervosa with family members, friends, and teachers) that reveals
that in the general population) are elevated17 for all overvaluation of thinness and abnormal food re-
striction, compulsive exercise, and sometimes bing- atric status of the patient, the patient’s and the
ing and purging and on the basis of a physical ex- family’s motivation, the feasibility of outpatient
amination revealing excessive thinness. Purging is weight restoration, and the availability of local re-
suggested by enlargement of the salivary glands sources. Adult patients whose weight is more than
(“chubby cheeks”), eroded dental enamel, and scars 25 percent below the expected weight (or with less
on the dorsum of the hands from repeated, self- weight loss if there are severe coexisting psychiat-
induced vomiting. Meeting the criteria in Table 1 ric or medical conditions, or both, or other pertinent
establishes the diagnosis; it is unnecessary to rule considerations) and children or teenage patients
out all medical causes of weight loss. who are losing weight rapidly, regardless of the per-
No specific laboratory tests confirm the diagno- centage of body weight lost, generally require hos-
sis. Table 2 summarizes the initial laboratory studies pitalization to ensure food intake and to limit phys-
recommended to assess the physiological effect of ical activity. Early intervention can reduce the risk
anorexia nervosa. Dual-energy x-ray absorptiome- of arrhythmia and the loss of cortical brain matter25;
try of bone is recommended, because osteopenia it is also thought to prevent the disorder from be-
may occur within six months of the development of coming chronic, although data are lacking to sup-
amenorrhea. Clinical experience suggests that doc- port this view.
umenting early osteopenia and making patients Observational data suggest that hospitalization
aware of the physiological effects of weight loss may in hospital units that specialize in the care of pa-
help to motivate patients who do not readily ac- tients with eating disorders yields better outcomes
knowledge having the disease. than hospitalization in general medical units — a
result that may be attributable to the nursing exper-
treatment tise and the use of effective protocols for refeeding
At present, there have been few controlled trials to (nutritional rehabilitation) and emotional care in
guide treatment,25,26 but numerous observational the specialized units.30 In severe cases, involun-
studies suggest that initial treatment should focus tary commitment to a psychiatric facility may be re-
on prompt weight restoration.25,27,28 Because many quired. Short-term outcomes among patients who
patients enter treatment reluctantly, techniques that have been involuntarily committed are similar to
enhance motivation (involving encouragement to those among patients admitted voluntarily.31
acknowledge the disease and to facilitate readiness
to change)29 are increasingly used to treat this con- Refeeding
dition, although no studies have yet confirmed their No particular nutritional regimen has proved to be
value. In the care of young children and adolescents, superior, so long as adequate calories are sup-
engaging the family is a necessary part of treatment. plied.32 Brisk improvement in nutritional status
Initial outpatient treatment often involves a pri- with few complications resulting from refeeding
mary care physician, a psychiatrist or psychologist occurs when inpatients are started with 1200 to
familiar with anorexia nervosa, and a registered 1500 kcal per day and the intake is increased by
dietitian. Educating the patient and family is critical 500 kcal every four days to about 3500 kcal (for fe-
with regard to the nature of the illness, serious male patients) to 4000 kcal (for male patients) per
health risks, effective treatments, and the need for day. Supplemental overnight nasogastric feeding
follow-up. Patients should be seen regularly, usually may slightly decrease the length of the hospital
weekly, to monitor weight (as measured in the early stay among children33 but is not routinely recom-
morning, after voiding, in a hospital gown) and mended. Refeeding usually reduces apathy, lethar-
other physical and laboratory indicators, such as gy, and food-related obsessions, although it does
cardiovascular values and electrolyte levels, depend- not generally eliminate them. Total parenteral ali-
ing on the individual patient’s course. Care must be mentation is rarely appropriate.
coordinated with other clinicians. Close monitoring is needed during starvation
A caloric intake of approximately 1200 to 1500 and refeeding, including monitoring of vital signs
kcal is usually recommended initially, with weekly and attention to peripheral edema and cardiopul-
increases of 500 kcal per day among outpatients monary function. A refeeding syndrome (reported
for a weight restoration of 0.5 to 0.9 kg (1 to 2 lb) in about 6 percent of hospitalized adolescents34)
per week.25 Indications for hospitalization, shown may include minor abnormalities (e.g., transient
in Table 3, depend on the physiological and psychi- pedal edema) or serious complications that require
Routine studies
Complete blood count
Urinalysis
Measurement of serum electrolytes; levels of creatinine, thyrotropin, and phosphorus; and fasting glucose values†
Studies to consider for selected patients
Measurement of the level of serum amylase, for patients suspected of surreptitious vomiting‡
Measurement of serum calcium and magnesium levels and liver-function tests,§ for patients with weight below 75 percent
of the expected weight; also, electrocardiogram (before initiation of atypical antipsychotic medications)
Dual-energy x-ray absorptiometry of bone, for patients who have been underweight for longer than six months
Magnetic resonance imaging or computed tomography of the brain and neuropsychological assessment for patients with
atypical features, such as hallucinations, delusions, delirium, and persistent cognitive impairment, despite weight
restoration
* Severely ill patients may have completely normal values on laboratory tests. Adapted from the Practice Guideline for the
Treatment of Patients with Eating Disorders of the American Psychiatric Association, second edition.25
† Patients in whom hypoglycemia is found on laboratory testing are frequently asymptomatic. Among those with the
euthyroid sick syndrome, with decreased triiodothyronine (T3) and increased reverse T3 (its inactive isomer), this syn-
drome is common and often improves with weight restoration.
‡ Elevated serum amylase is primarily secreted by the salivary gland.
§ Hypocalcemia may result from chronic malnutrition or alkalosis and may be associated with changes on the electro-
cardiogram. A magnesium deficiency may result from malnutrition, diarrhea, or misuse of diuretic agents and may be
associated with hypokalemia, hypophosphatemia, and changes on the electrocardiogram. Malnutrition may produce
hepatomegaly, fatty liver, and rarely, cirrhosis. Elevated levels of liver enzymes may also reflect alcohol abuse or the use
of medications with toxic effects on the liver.
urgent intervention (e.g., a prolonged QT interval may be caused by slowed gastric emptying, may
or hypophosphatemia with associated weakness, benefit from treatment with metoclopramide (at a
confusion, and progressive neuromuscular dys- dose of 5 to 10 mg one hour before meals and at bed-
function) (Table 2). This syndrome is most com- time). Gastroesophageal reflux usually improves
mon among patients weighing less than 70 percent with proton-pump–inhibitor therapy.
of their ideal body weight and in those receiving Short-term medical stabilization alone is inevita-
parenteral or enteral nutrition, although it can also bly insufficient. Because a brisk but medically safe
occur in those receiving vigorous oral refeeding. weight gain among hospitalized patients averages
Slower refeeding minimizes the risk of serious com- 0.9 to 1.4 kg (2 to 3 lb) per week, inpatients who are
plications. Phosphorus, magnesium, and electro- 9.1 to 13.6 kg (20 to 30 lb) below their healthy
lyte levels and renal function should be followed weight may require two to three months of inpatient
closely, and supplements should be administered treatment. Rates of relapse and rehospitalization
as needed. Clinical changes and laboratory values are higher among hospitalized patients who are
requiring immediate attention include altered con- discharged at low weights and before they and
sciousness, tachycardia, congestive heart failure, their families can assume responsibility for refeed-
atypical abdominal pain, a prolonged QT interval ing than among patients discharged at expected
or QT dispersion (a marker of abnormal ventricular healthy weights, when they and their families can
repolarization associated with an increased risk of assume such responsibility.35
arrhythmia), serum potassium levels below 3 mmol To achieve full remission, ongoing care after dis-
per liter, and serum phosphorus levels below 0.8 charge from the hospital is essential. It has been cal-
mmol per liter (2.5 mg per deciliter). culated that “adequate care” programs — modeled
The management of refeeding complications is on usual care as delivered in the community — in-
guided by clinical experience, in the absence of stud- volving inpatient hospitalization lead to weight
ies. Peripheral edema is treated with leg elevation restoration (close to 100 percent of the ideal body
and withholding added salt from the diet; diuretics weight). When followed by nearly three weeks of
may exacerbate the edema, and their use should be daytime hospital care, 50 sessions of psychotherapy,
avoided. Gastrointestinal symptoms are common and 20 medication visits with fluoxetine treatment
during refeeding and often persist. Bloating, which over two years, this program can prove to be more
nationaleatingdisorders.org) and the National As- Dr. Andersen reports having received lecture fees from Pfizer.
We are indebted to Drs. Katherine A. Halmi, David B. Herzog,
sociation of Anorexia Nervosa and Associated Dis- and Pauline S. Powers for helpful comments.
orders (www.anad.org) provide resources for pa-
tients and families.
references
1. Hoek HW, van Hoeken D. Review of the 15. Lambe EK, Katzman DK, Mikulis DJ, mental health — one to chew over. Health
prevalence and incidence of eating disorders. Kennedy SH, Zipursky RB. Cerebral gray Serv J 2004;114:28-9.
Int J Eat Disord 2003;34:383-96. matter volume deficits after weight recovery 29. Vitousek K, Watson S, Wilson GT. En-
2. Urwin RE, Nunn KP. Epistatic interac- from anorexia nervosa. Arch Gen Psychiatry hancing motivation for change in treatment-
tion between the monoamine oxidase A and 1997;54:537-42. resistant eating disorders. Clin Psychol Rev
serotonin transporter genes in anorexia ner- 16. Bulik CM, Sullivan PF, Fear JL, Pickering 1998;18:391-420.
vosa. Eur J Hum Genet 2005;13:370-5. A, Dawn A, McCullin M. Fertility and repro- 30. Wolfe BE, Gimby LB. Caring for the
3. Bulik CM, Sullivan PF, Fear JL, Joyce PR. duction in women with anorexia nervosa: hospitalized patient with an eating disorder.
Eating disorders and antecedent anxiety dis- a controlled study. J Clin Psychiatry 1999;60: Nurs Clin North Am 2003;38:75-99.
orders: a controlled study. Acta Psychiatr 130-5. 31. Watson TL, Bowers WA, Andersen AE.
Scand 1997;96:101-7. 17. Harris EC, Barraclough B. Excess mor- Involuntary treatment of eating disorders.
4. Halmi KA, Sunday SR, Klump KL, et al. tality of mental disorder. Br J Psychiatry Am J Psychiatry 2000;157:1806-10.
Obsessions and compulsions in anorexia 1998;173:11-53. 32. Okamoto A, Yamashita T, Nagoshi Y, et
nervosa subtypes. Int J Eat Disord 2003;33: 18. Keel PK, Dorer DJ, Eddy KT, Franko D, al. A behavior therapy program combined
308-19. Charatan DL, Herzog DB. Predictors of with liquid nutrition designed for anorexia
5. Godart NT, Flament MF, Curt F, et al. mortality in eating disorders. Arch Gen Psy- nervosa. Psychiatry Clin Neurosci 2002;56:
Anxiety disorders in subjects seeking treat- chiatry 2003;60:179-83. 515-20.
ment for eating disorders: a DSM-IV con- 19. Herzog DB, Greenwood DN, Dorer DJ, 33. Robb AS, Silber TJ, Orrell-Valente JK, et
trolled study. Psychiatry Res 2003;117:245- et al. Mortality in eating disorders: a descrip- al. Supplemental nocturnal nasogastric re-
58. tive study. Int J Eat Disord 2000;28:20-6. feeding for better short-term outcome in
6. Steiner H, Kwan W, Shaffer TG, et al. 20. Crisp AH, Callender JS, Halek C, Hsu hospitalized adolescent girls with anorexia
Risk and protective factors for juvenile eat- LK. Long-term mortality in anorexia ner- nervosa. Am J Psychiatry 2002;159:1347-53.
ing disorders. Eur Child Adolesc Psychiatry vosa: a 20-year follow-up of the St George’s 34. Kohn MR, Golden NH, Shenker IR. Car-
2003;12:Suppl 1:38-46. and Aberdeen cohorts. Br J Psychiatry 1992; diac arrest and delirium: presentations of
7. Diagnostic and statistical manual of 161:104-7. the refeeding syndrome in severely malnour-
mental disorders, 4th ed.: DSM-IV-TR. 21. Westen D, Harnden-Fischer J. Personal- ished adolescents with anorexia nervosa.
Washington, D.C.: American Psychiatric As- ity profiles in eating disorders: rethinking J Adolesc Health 1998;22:239-43.
sociation, 2000. the distinction between axis I and axis II. Am 35. Howard WT, Evans KK, Quintero-
8. Garfinkel PE, Lin E, Goering P, et al. J Psychiatry 2001;158:547-62. Howard CV, Bowers WA, Andersen AE. Pre-
Should amenorrhoea be necessary for the 22. Strober M, Freeman R, Morrell W. Atyp- dictors of success or failure of transition to
diagnosis of anorexia nervosa? Evidence ical anorexia nervosa: separation from typi- day hospital treatment for inpatients with
from a Canadian community sample. Br J cal cases in course and outcome in a long- anorexia nervosa. Am J Psychiatry 1999;156:
Psychiatry 1996;168:500-6. term prospective study. Int J Eat Disord 1697-702.
9. Halmi KA, Eckert E, Marchi P, Sam- 1999;25:135-42. 36. Crow SJ, Nyman JA. The cost-effective-
pugnaro V, Apple R, Cohen J. Comorbidity 23. Strober M, Freeman R, Morrell W. The ness of anorexia nervosa treatment. Int J Eat
of psychiatric diagnoses in anorexia ner- long-term course of severe anorexia nervosa Disord 2004;35:155-60.
vosa. Arch Gen Psychiatry 1991;48:712-8. in adolescents: survival analysis of recovery, 37. Attia E, Haiman C, Walsh BT, Flater SR.
10. Kaye WH, Bulik CM, Thornton L, Bar- relapse, and outcome predictors over 10-15 Does fluoxetine augment the inpatient treat-
barich N, Masters K. Comorbidity of anxiety years in a prospective study. Int J Eat Disord ment of anorexia nervosa? Am J Psychiatry
disorders with anorexia and bulimia ner- 1997;22:339-60. 1998;155:548-51.
vosa. Am J Psychiatry 2004;161:2215-21. 24. Eckert ED, Halmi KA, Marchi P, Grove 38. Strober M, Pataki C, Freeman R, DeAn-
11. Bulik CM, Klump KL, Thornton L, et al. W, Crosby R. Ten-year follow-up of anorexia tonio M. No effect of adjunctive fluoxetine
Alcohol use disorder comorbidity in eating nervosa: clinical course and outcome. Psy- on eating behavior or weight phobia during
disorders: a multicenter study. J Clin Psychi- chol Med 1995;25:143-56. the inpatient treatment of anorexia nervosa:
atry 2004;65:1000-6. 25. Practice guideline for the treatment of an historical case-control study. J Child Ado-
12. Johnson JG, Cohen P, Kasen S, Brook JS. patients with eating disorders (revision): lesc Psychopharmacol 1999;9:195-201.
Eating disorders during adolescence and American Psychiatric Association Work 39. Barbarich NC, McConaha CW, Gaskill J,
the risk for physical and mental disorders Group on Eating Disorders. Am J Psychiatry et al. An open trial of olanzapine in anorexia
during early adulthood. Arch Gen Psychiatry 2000;157:Suppl 1:1-39. nervosa. J Clin Psychiatry 2004;65:1480-2.
2002;59:545-52. 26. Hay P, Bacaltchuk J, Claudino A, Ben- 40. Malina A, Gaskill J, McConaha C, et al.
13. Vestergaard P, Emborg C, Stoving RK, Tovim D, Yong PY. Individual psychotherapy Olanzapine treatment of anorexia nervosa:
Hagen C, Mosekilde L, Brixen K. Fractures in the outpatient treatment of adults with a retrospective study. Int J Eat Disord 2003;
in patients with anorexia nervosa, bulimia anorexia nervosa. Cochrane Database Syst 33:234-7.
nervosa, and other eating disorders — Rev 2003;4:CD003909. 41. Powers PS, Santana CA, Bannon YS.
a nationwide register study. Int J Eat Disord 27. Beumont P, Hay P, Beumont D, et al. Olanzapine in the treatment of anorexia ner-
2002;32:301-8. Australian and New Zealand clinical prac- vosa: an open label trial. Int J Eat Disord
14. Rigotti NA, Neer RM, Skates SJ, Herzog tice guidelines for the treatment of anorexia 2002;32:146-54.
DB, Nussbaum SR. The clinical course of nervosa. Aust N Z J Psychiatry 2004;38:659- 42. Robin AL, Siegel PT, Moye AW, Gilroy
osteoporosis in anorexia nervosa: a longitu- 70. [Erratum, Aust N Z J Psychiatry 2004;38: M, Dennis AB, Sikand A. A controlled com-
dinal study of cortical bone mass. JAMA 987.] parison of family versus individual therapy
1991;265:1133-8. 28. Dix A. Clinical management special: for adolescents with anorexia nervosa. J Am
Acad Child Adolesc Psychiatry 1999;38: 47. Pike KM, Walsh BT, Vitousek K, Wilson Shenker RI. Reversibility of growth stunt-
1482-9. GT, Bauer J. Cognitive behavior therapy in ing in early onset anorexia nervosa: a pro-
43. Eisler I, Dare C, Russell GF, Szmukler the posthospitalization treatment of ano- spective study. J Adolesc Health 2002;31:
G, le Grange D, Dodge E. Family and indi- rexia nervosa. Am J Psychiatry 2003;160: 162-5.
vidual therapy in anorexia nervosa: a 5-year 2046-9. 52. Welcome to nice.org. (Accessed Sep-
follow-up. Arch Gen Psychiatry 1997;54: 48. Golden NH, Lanzkowsky L, Scheben- tember 1, 2005, at http://www.nice.org.uk/
1025-30. dach J, Palestro CJ, Jacobson MS, Shenker pdf/cg009niceguidance.pdf.)
44. Dare C, Eisler I, Russell G, Treasure J, IR. The effect of estrogen-progestin treat- 53. Ebeling H, Tapanainen P, Joutsenoja A,
Dodge L. Psychological therapies for adults ment on bone mineral density in anorexia et al. A practice guideline for treatment of
with anorexia nervosa: randomised con- nervosa. J Pediatr Adolesc Gynecol 2002;15: eating disorders in children and adolescents.
trolled trial of out-patient treatments. Br J 135-43. Ann Med 2003;35:488-501.
Psychiatry. 2001;178:216-21. 49. Grinspoon S, Thomas L, Miller K, Her- 54. American Academy of Pediatrics, Com-
45. McIntosh VV, Jordan J, Carter FA, et al. zog D, Klibanski A. Effects of recombinant mittee on Adolescence. Identifying and treat-
Three psychotherapies for anorexia nervosa: human IGF-I and oral contraceptive admin- ing eating disorders. Pediatrics 2003;111:
a randomized, controlled trial. Am J Psychi- istration on bone density in anorexia ner- 204-11.
atry 2005;162:741-7. vosa. J Clin Endocrinol Metab 2002;87: 55. Position of the American Dietetic Asso-
46. Kaye WH, Nagata T, Weltzin TE, et al. 2883-91. ciation: nutrition intervention in the treat-
Double-blind placebo-controlled adminis- 50. Golden NH. Osteopenia and osteoporo- ment of anorexia nervosa, bulimia nervosa,
tration of fluoxetine in restricting- and re- sis in anorexia nervosa. Adolesc Med 2003; and eating disorders not otherwise specified
stricting-purging-type anorexia nervosa. Biol 14:97-108. (EDNOS). J Am Diet Assoc 2001;101:810-9.
Psychiatry 2001;49:644-52. 51. Lantzouni E, Frank GR, Golden NH, Copyright © 2005 Massachusetts Medical Society.