Print This Article

Infection-Triggered Anorexia Nervosa

by Mae S. Sokol, MD, Creighton Univ. School of Medicine, Omaha, NE
Reprinted from Eating Disorders Review
September/October 2001 Volume 12, Number 5
©2001 Gürze Books

Recent evidence suggests that transmissible agents may cause or contribute to some psychiatric illnesses. The idea that psychiatric disorders and infectious disease are related is not new. Neurosyphilis was frequently treated as a mental disorder until the 1940s, when it was found that penicillin could successfully treat the causative spirochete, Treponema pallidum. Peptic ulcers were believed to be caused by stress and spicy food until the 1980s, when it was learned that many ulcers are caused by infection with Helicobacter pylori,and can be treated withantibiotics. Case reports suggesting that tics can be linked with infection have been reported in the medical literature for decades (Arch Neurol 1929; 22: 1163; Arch Pediatr 1957; 74:39).

Infection-triggered Disorders

Clinical and research observations have led to the hypothesis that a post-infectious and autoimmune process may cause or exacerbate certain cases of obsessive-compulsive disorder (OCD), with or without tic disorders (Am J Psychiatry 1998;155: 264), autism (Am J Psychiatry 1999;156: 317), and anorexia nervosa (AN) (J Child & Adolesc Psychiatry 1997;36:1128). These disorders are described by the acronym PANDAS (Pediatric Auto-immune Neuropsychiatric Disorders Associated with Streptococcus), when they occur in children and when the causative agent is Group A Streptococcus, the bacterium that causes strep throat. PANDAS may be part of a larger category: PITANDs (Pediatric, Infection-Triggered, Autoimmune Neuropsychiatric Disorders). PITANDs may be caused by a number of infectious agents, such as bacteria and viruses.

The link between group A streptococcal infection and rheumatic fever (RF) and Sydenham's chorea (SC, the neurologic form of RF) is well known. RF and SC are believed to be due to molecular mimicry. That is, antibodies directed against pharyngitis-causing strep bacteria in the patient's throat mistakenly attack other body tissues. In RF, cardiac cells are attacked. In SC, damage to basal ganglia neurons and other brain tissue is believed to cause abnormal movements and behavioral disturbances.

It is hypothesized that in PANDAS conditions, antistreptococcal antibodies mistakenly act as antibrain antibodies, cross-reacting with neurons in the basal ganglia, which may influence emotions and behavior, presumably disrupting function there. Alternatively, cytokines, which are proteins that transmit information among the immune system, the brain, and endocrine organs, may be involved. Cytokines may have a possible role in the pathogenesis of AN (Int J Eat Disord 2000;28:293).

At the National Institute of Mental Health, Susan Swedo and colleagues described the clinical characteristics of a group of 50 children with OCD, with or without tics, in whom onset or worsening of symptoms was preceded by infection (J Am Acad Child Adolesc Psychiatry 1995;34:307). Upon testing, these children often express the lymphocyte B marker D8/17. This antigen is seen in most children who develop RF and SC, and is thought to be a trait marker for susceptibility to RF (J Clin Invest 1989;83:1710).

D8/17 expression also has been reported to be elevated in youngsters with OCD (Am J Psychiatry 1997;154:110; Am J Psychiatry 1997;154:402); tics (Am J Psychiatry 2001;158:605); autism (Am J Psychiatry 1999;156:317); and in four cases of AN where clinical characteristics of PANDAS were present (J Child & Adol Psychopharmacol 2000;10:133). Although the diagnostic value of D8/17 remains uncertain, it may provide insight into the underlying processes in these illnesses.

Immunologic Treatment: Promising Results

Encouraging results have followed immunologic treatment-plasma exchange and intravenous immunoglobulin-for PANDAS OCD and tic disorders (Lancet 1999;354:1153), as well as some benefit from antibiotic treatment and prophylaxis for these disorders (Biol Psychol 1999;45:1564). An excellent review of the research, assessment, and management of PANDAS OCD and tics is provided by Hamilton and Swedo in Clinical Neuroscience Research (2001;1:61).

Infection-triggered AN

OCD and AN may be related, as evidenced by phenomenology, comorbidity, neurotransmitters, and central nervous system functional metabolism (J Psychiatry Neurosci 1996;121:36). AN patients frequently have obsessive thoughts (about food, calories, and weight) and compulsive behaviors (over-exercise and/or odd eating behaviors, for example). This, along with clinical observation of certain cases of AN, which clinically appeared to be triggered by infections, led to the hypothesis that there is a pediatric, infection-triggered, autoimmune type of AN that is similar to PANDAS OCD. In this variant, children experience either onset or dramatic worsening of eating disorder symptoms following a group A streptococcal infection.

Four years ago we studied three patients with AN in whom there was a possible PITANDs etiology (J Am Acad Child Adolesc Psychiatry 1997;36:1128). Two of the cases of possible PITANDs may have had PANDAS AN. Amoxicillin may have been effective in decreasing eating disorder symptoms in one patient whose anorectic symptoms severely worsened following a bout with strep throat.

More recently, we have studied three other youngsters (11 to 15 years of age) with possible PANDAS AN, who may have benefited from an open trial of antibiotics (J Child & Adol Psychopharm 2000;10:133). Evidence of a temporal relationship between the onset of eating disorder symptoms and streptococcal infection came from clinical evaluation, throat cultures, and two serologic tests for Streptococcus: anti-deoxyribonuclease B (anti-DNase B) and anti-streptolysin O (ASO) titers.

Henry and colleagues described several children with sudden-onset eating obsessions, which appeared to be triggered by strep infection (J Am Acad Child Adolesc Psychiatry 1999; 38:228).Harel et al. recently reported the presence of antibrain antibodies (ABA) in several adolescents with AN (Int J Eat Disord 2001;29:463). ABA were detected in the blood of 6 of 22 adolescent girls with AN (27%), compared to only 1 of 22 healthy adolescent controls (5%). These ABA were directed against the putamen, part of the brain's basal ganglia, suggesting an underlying immune process in the basal ganglia in AN.

In summary, available evidence indicates a possible link between infectious disease and some cases of AN, which raises the possibility of new treatment.

Characteristics of PANDAS AN

As a working model, the following characteristics of PANDAS AN have been developed (See the enclosed Patient Information Sheet):

  1. Prepubertal onset of AN.
  2. Acute onset or exacerbation of eating disorder symptoms.
  3. Evidence of a prior or concomitant streptococcal infection, with a temporal relationship to the onset of eating disorder symptoms.
  4. Increased symptoms that do not occur exclusively during stress or physical illness (fever and other stressors of physical illness can decrease appetite and increase psychiatric symptoms).
  5. Minor neurologic abnormalities during psychiatric symptom exacerbation, including irregular movements and motor hyperactivity.

Clinical Implications

The possibility of infection-triggered AN raises questions about the management of youngsters with AN. It suggests that a history of infectious disease should be evaluated in relation to eating disorder symptoms. If there is a temporal association between onset or exacerbation of AN and a strep infection, the following are recommended: physical examination, throat culture, and serologic tests (anti-deoxyribonuclease B [anti-DNase B] and anti-streptolysin O [ASO] titers) for streptococci. Careful monitoring and treatment of strep infection are beneficial for all children. This may be particularly important for children with AN who have PANDAS characteristics. The D8/17 laboratory test is still a research test, and not recommended for use in clinical practice.

Further Investigation is Needed

Future research is needed to investigate the possible existence of the subtype of PANDAS AN, not only because these patients may benefit from careful monitoring for streptococcal and other infections, but for the potential benefit that new treatment options might offer. Early intervention could theoretically limit immune response and potentially thwart the onset or worsening of AN in these individuals. In other words, damage as well as dysfunction could be prevented.

Studies are underway to determine whether children with PANDAS will respond to antibiotic and immunomodulatory therapy (plasma exchange or intravenous immunoglobulin). These treatments are still in the experimental stage. Double-blind, placebo-controlled studies are needed to determine if these treatments are effective in treating or preventing AN symptoms. Further, the risks of antibiotic use must be weighed against the possible benefits, because antibiotic overuse contributes to the emergence of resistant bacteria, causes unnecessary adverse drug reactions, and is costly (Pediatrics 1999; 104:1251).

It is important to have a healthy skepticism about the link between infections and psychiatric disorders, but it is equally important to consider new possibilities such as this, as it may lead to better treatment.

Suggested Reading

Hooper J. A new germ theory. The Atlantic Monthly 1999; 41.

Lorber B. Are all diseases infectious? Another look. Ann Intern Med 1999; 131: 989.

Perlmutter SJ, Leitman SF, Garvey MA, et al. Therapeutic plasma exchange and intravenous immunoglobulin for obsessive-compulsive disorder and tic disorders in childhood. Lancet, 1999; 354: 1153.

Sokol, M: Infection-triggered anorexia nervosa in children: clinical description of four cases. J Child & Adolesc Psychopharmacol 2000; p. 133.

Swedo SE, Garvey M, Snider L, et al. The PANDAS subgroup: recognition and treatment. CNS Spectrums 2001; 6:419.

Swedo SE, Leonard HL, Garvey M, et al. Pediatric autoimmune neuropsychiatric disorders associated with streptococcal infections (PANDAS): A clinical description of the first fifty cases. Am J Psychiatry 1998; 155: 264.


Advertising Information