The Pediatric Autoimmune Neuropsychiatric Disorders Associated With Streptococcal Infection (PANDAS) Subgroup: Separating Fact From Fiction

Abstract

Over a century ago, Sir William Osler wrote, “To carefully observe the phenomena of life in all its phases … to call to aid the science of experimentation, to cultivate the reasoning faculty, so as to be able to know the true from the false—these are our methods.”1These were also the methods that led to the discovery of poststreptococcal obsessive-compulsive disorder (OCD) and tic disorders and a decade of observations and research resulting in the description of a novel cohort of patients, the pediatric autoimmune neuropsychiatric disorders associated with streptococcal infection (PANDAS) subgroup.2,3 In this issue of Pediatrics, Kurlan and Kaplan raise questions about the veracity of these data.4 To respond, we will provide a brief literature review and clarification of the guidelines for management of a patient in the PANDAS subgroup.The discovery of the PANDAS subgroup was the result of 2 parallel lines of clinical research conducted at the National Institute of Mental Health (NIMH): studies of children with OCD and investigations of children with Sydenham’s chorea, the neurologic manifestation of rheumatic fever. Systematic observations of children with OCD revealed that, although the majority of children had a gradual onset of symptoms over several weeks to months, a subgroup of the patients experienced an explosive “overnight” onset of obsessions and compulsions followed by a relapsing-remitting symptom course.5 Closer observation revealed that the neuropsychiatric symptom relapses frequently occurred after episodes of streptococcal pharyngitis or scarlet fever. These findings in OCD closely paralleled those from a series of investigations of Sydenham’s chorea.6 In those studies, 65% to 100% of children with Sydenham’s chorea were noted to have obsessive-compulsive symptoms, typically presenting 2 to 4 weeks before the onset of the adventitious movements and peaking in severity simultaneously with the chorea.6,7 Longitudinal observations of the OCD subgroup and the patients with Sydenham’s chorea clearly demonstrated a temporal association between streptococcal infections and obsessive-compulsive symptoms. This relationship was not only observed consistently among patients presenting to the NIMH but also noted by several independent groups.8–11 The nature of the association was unknown, and the observations could not elucidate whether the streptococcal infections played an etiologic role, but these issues would be addressed through subsequent scientific experimentation.The title of the article by Kurlan and Kaplan4 provides a provocative starting point for discussion of the scientific hypotheses that derive from the clinical observations of the PANDAS subgroup. However, the authors subsequently blur the distinction between clinical observation and scientific investigation, leading them to dismiss the well-documented observations that neuropsychiatric symptoms are associated with streptococcal infections in the PANDAS subgroup because the etiology of PANDAS “remains a yet-unproven hypothesis.”4 The authors thus recommend against obtaining throat cultures or serial titers in patients with abrupt-onset OCD and tics “until more definitive scientific proof is forthcoming.” We strongly disagree with this recommendation. The continued threat of rheumatic fever mandates the detection and appropriate treatment of streptococcal infections, including asymptomatic infections, the leading cause of rheumatic carditis in the United States.12 If one argues that OCD and tics are a manifestation of streptococcal infection for children in the PANDAS subgroup, then the infections aren’t really “silent” or “asymptomatic.” In either case, a conservative treatment course would include administration of antibiotics for culture-proven streptococcal infections. In addition, Murphy and Pichichero11 have documented that prompt treatment of streptococcal infections is associated with a rapid diminution of obsessive-compulsive symptom severity for some children in the PANDAS subgroup. Thus, the potential benefits of appropriate diagnosis and treatment of an occult streptococcal infection far outweigh the modest cost of obtaining a throat swab and culture. Of course, when throat cultures are obtained, there is a risk of falsely identifying a “carrier” as an asymptomatic infection, but this risk is small. Systematic studies typically report the frequency of carriers to be <5% to 10%.13 Thus, the vast majority of positive throat cultures represent true streptococcal infections, for which antibiotics administration is the accepted standard of care.Kurlan and Kaplan contend that the 5 criteria defining the PANDAS subgroup are not “particularly useful in distinguishing patients suspected of PANDAS from children with more typical cases of TS [Tourette’s syndrome] or OCD.”4 In actuality, the criteria have been used successfully by a variety of clinical groups to define cohorts of patients with common clinical characteristics and a predictable clinical course.3,9–11,14 This had been the original purpose of describing the PANDAS subgroup: to enable investigators to identify a clinically homogeneous group of patients for inclusion in research studies at the NIMH and elsewhere. Subsequent investigations have demonstrated that the criteria have clinical utility as well, in that they define a distinct cohort of patients who are uniquely responsive to novel therapeutic interventions and prevention strategies. The following is a clarification of the criteria.The symptom characteristics and severity required for diagnosis are defined in the Diagnostic and Statistical Manual of Mental Disorders.15 The neuropsychiatric symptoms of the PANDAS subgroup were intentionally limited to tics and obsessive-compulsive symptoms because of our interest in establishing a homogeneous patient cohort for research studies. Subsequent interest in the PANDAS subgroup has sparked a number of authors to speculate that the criteria should be expanded to include other related disorders such as attention-deficit/hyperactivity disorder16 and anorexia.17 However, such a change requires systematic evidence documenting that the association between streptococcal infections and symptom onset in these disorders is not merely a chance finding; to date, such systematic studies have not been done.This criterion was based on historical data demonstrating that rheumatic fever and other poststreptococcal sequelae are uncommon before the age of 3 years and after the age of 12 years.18 Fischetti19 provides a possible explanation for the rarity of postpubertal sequelae of streptococcal infections and demonstrated the presence of serum antibodies conferring protection against streptococcal infections in 98% of healthy 12-year-old controls, making it unlikely that poststreptococcal neuropsychiatric symptoms would have their initial presentation after this age. Thus, we set the age range for the PANDAS subgroup at a point that had biological relevance and would include 98% of the cases.Prospective longitudinal investigations have demonstrated that this criterion is the most useful in identifying children in the PANDAS subgroup.2,3,9–11 Contrary to the concerns expressed by Kurlan and Kaplan,4 the abrupt onset of tics in the PANDAS subgroup is clearly different from the typical onset of an isolated, intermittent, simple motor or vocal tic, because children in the PANDAS subgroup experience the simultaneous onset of several different motor and vocal tics of such intensity and frequency that emergency treatment is often sought.14 PANDAS-related OCD is also easily distinguished from non-PANDAS OCD, because the latter patients have a slow, gradual symptom onset, whereas children in the PANDAS subgroup have an overnight “explosion” of obsessive-compulsive symptoms, reaching maximal, clinically significant impairment in 24 to 48 hours.3,20The episodic, relapsing-remitting course of the PANDAS subgroup is distinctly different from the undulating, waxing-waning course seen in other patients with OCD or tic disorders.20,21 When the symptoms of a child in the PANDAS subgroup are graphed against time, a “saw-toothed” pattern emerges, in which periods of symptom quiescence are interrupted abruptly by severe symptom exacerbations; these relapses typically take several weeks to months to resolve. Prospective, longitudinal evaluation of these patients allows for documentation of the relationship between the symptom exacerbations and streptococcal infections: throat cultures obtained at the beginning of a symptom relapse will be positive, and titers obtained at baseline and 4 to 6 weeks later will demonstrate a clinically significant rise.Although it was postulated initially that there could be a significant time lag between the inciting streptococcal infection and the presentation of the neuropsychiatric sequelae (such as that seen in Sydenham’s chorea),6 clinical observations of the PANDAS subgroup revealed that the window is actually much narrower. Exacerbations of neuropsychiatric symptoms begin within 7 to 14 days after the streptococcal infection and usually occur simultaneously (ie, a throat culture obtained because of the recent onset of OCD and/or tics is positive).3,11,20,21One caveat in evaluating the relationship between streptococcal infections and neuropsychiatric symptoms is that the disorders are so common that co-occurrence can be a random coincidence rather than a clinically significant finding. OCD occurs in 1% to 2% of school-aged children, and transient motor tics occur in as many as 10% to 25% of early elementary students.22,23 Furthermore, during regional streptococcal epidemics, the majority of children will be infected at least once during the outbreak.13 Thus, as discussed in our original report,3 a single positive throat culture or elevated antistreptococcal antibody titer is not sufficient to determine that a child’s neuropsychiatric symptoms are associated with streptococcal infections.3,20 Instead, the determination that a child fits the PANDAS profile is made through prospective evaluation and documentation of the presence of streptococcal infections in conjunction with at least 2 episodes of neuropsychiatric symptoms, as well as demonstrating negative throat culture or stable titers during times of neuropsychiatric symptom remission.3 A child who has multiple symptom exacerbations without evidence of streptococcal infection would not be considered part of the PANDAS subgroup, nor would a child who has numerous streptococcal infections without subsequent symptom exacerbations.Neurologic examination of acutely ill children in the PANDAS subgroup reveals that 95% have choreiform movements.3 These fine piano-playing movements of the fingers are not easily confused with the writhing adventitious movements of Sydenham’s chorea.24 Choreiform movements are not present at rest and must be elicited through stressed postures, whereas choreatic movements are present continuously and increase with unrelated voluntary movements. In addition, choreiform movements are an isolated finding, whereas the choreatic movements of Sydenham’s chorea are accompanied by a failure to sustain tetanic contractions (milk-maid’s grip, snake-like tongue) and muscle weakness.6,18 Choreiform movements and chorea may share a common pathophysiology (related to dysfunction of the basal ganglia), but the clinical manifestations are quite distinct, and children in the PANDAS subgroup do not represent missed cases of Sydenham’s chorea. In fact, rheumatic fever, including Sydenham’s chorea, is a strict exclusionary criterion for the PANDAS subgroup.3Clinical observations of the PANDAS subgroup led to a number of scientific hypotheses including the postulate that the tics and OCD represent sequelae of group A streptococcal infections. This etiologic hypothesis involves a series of factors including pathologic strains of group A streptococcal bacteria, host susceptibility (genetic, developmental, or other), and abnormal immune responsivity (Fig 1). The working model of pathogenesis not only provides a framework for understanding the etiology of OCD and tic disorders but also allows for the development of novel intervention and prevention strategies. A recent review provides a detailed description of the model as well as ongoing research efforts directed at understanding the pathologic mechanisms involved in the PANDAS subgroup.25These guidelines are drawn from our clinical and research experience as well as the practice parameters of the American Academy of Child and Adolescent Psychiatry.26The PANDAS subgroup is both a clinical entity and the subject of scientific experimentation. Systematic, longitudinal observations have demonstrated that the PANDAS subgroup has a distinct clinical presentation and an identifiable course of symptoms and that, for these children, there is a clear relationship between streptococcal infections and neuropsychiatric symptom exacerbations. Additional research is required to determine the nature of that relationship as well as to determine the etiopathogenesis of the poststreptococcal obsessive-compulsive symptoms and tics. Additional studies are required also to determine the role of immunomodulatory therapies and antibiotics prophylaxis for this group of patients. Meanwhile, it is time to end the debate about the existence of the PANDAS subgroup and begin to “call to aid the science of experimentation … so as to be able to know the true from the false.”

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 221 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.