|
 In
the late 1980s, studies of children with Sydenham chorea (SC), the neurological
manifestation of rheumatic fever, suggested that the disorder might serve
as a useful model of pathophysiology for some forms of childhood-onset
obsessive-compulsive disorder (OCD) and tic disorders. The disorders share
anatomic similarities. Both OCD and SC have evidence of basal ganglia
dysfunction, particularly in the caudate nucleus, which is thought to
disrupt signals traveling along the orbitofrontal-striatal pathways. Furthermore,
over 70% of children with SC reported that they had experienced an abrupt
onset of repetitive, unwanted thoughts and behaviors 2 to 4 weeks before
the onset of their chorea. These obsessions and compulsions peaked in
intensity concomitantly with the chorea and waned away slowly over the
ensuing months. Because the obsessive-compulsive symptoms began earlier
than the chorea, it seemed possible that poststreptococcal OCD might occur
in the absence of chorea, a hypothesis confirmed by prospective observations
of a large cohort of children with primary OCD.
Among
those children, a subgroup was noted to have dramatic symptom exacerbations
following infections with group A b-hemolytic
streptococcal bacteria (GABHS) such as occurs with strep throat and scarlet
fever. The symptom exacerbations were accompanied by a cluster of comorbid
symptoms, including emotional lability, separation anxiety, and attentional
difficulties. The children were young (6–7 years old at symptom onset),
predominantly male, and had frequent comorbid tics. To indicate the subgroup’s
common clinical characteristics and presumed pathophysiology, it was identified
by the acronym PANDAS (pediatric autoimmune neuropsychiatric disorders
associated with streptococcal infections).
The
etiology proposed for the PANDAS subgroup is similar to that postulated
for SC. In SC, host susceptibility is thought to play a crucial role in
symptom expression, as fewer than 5% of children are vulnerable to poststreptococcal
sequelae. Familial clustering suggests that genetic factors are involved
in the susceptibility, but is not the sole explanation as developmental
and immunological factors may also play a part.
The
constitution of the streptococcal bacteria appears to play an etiological
role in rheumatic fever and other poststreptococcal sequelae. Although
most strains of GABHS produce only acute symptoms, certain “rheumatogenic”
strains incite the production of antibodies that cross-react with host
tissues, producing an “autoimmune”; reaction. Unlike typical
autoimmune disorders, the autoantibodies in rheumatic fever are not directed
primarily against the host tissue, but rather against streptococcal epitopes
that resemble antigens on host cells. In SC, the cross-reactive antibodies
are thought to recognize epitopes on basal ganglia neurons (particularly
in the caudate, putamen, and globus pallidus). The resulting inflammation
disrupts basal ganglia function, causing chorea, emotional lability, and
other neuropsychiatric symptoms. After the streptococcal infection is
eliminated from the nasopharynx, antibody production ceases and circulating
antibodies are cleared over the ensuing weeks to months. The neurological
symptoms also remit—often with no lasting sequelae. However, recrudescences
can occur, particularly when the child is infected again with microbes
resembling the inciting organism.
Although
the antineuronal antibody model of SC appears quite convincing, it has
not yet been proven, despite decades of research. Furthermore, recent
studies by a number of research groups have demonstrated several crucial
limitations of the antineuronal antibody model, including its failure
to account for the presence of similar antibodies in the serum of 20%
to 40% of healthy children. Thus, for the PANDAS subgroup, the model has
been modified to indicate that the immune response is not limited to the
antineuronal antibodies. As shown in Figure
1, our working model of pathophysiology involves a series of factors,
including the streptococcus bacteria, host susceptibility, and abnormal
immune responsivity.
The
major distinguishing feature of the PANDAS subgroup is the temporal association
between the neuropsychiatric symptom exacerbations and the infections.
Because strep infections are common during childhood, antistreptococcal
titers and throat cultures might be positive during a symptom exacerbation
by chance alone. Thus the PANDAS criteria specify that GABHS infections
must be temporally related to the OCD/tics exacerbations, that is, positive
(or rising) antistreptococcal antibody titers or a positive throat culture
must be present during neuropsychiatric symptom relapses and there must
be evidence of strep negativity during periods of remission.
In
some cases, the dramatic onset of severe OCD and/or tics following a single,
prolonged strep infection is sufficient to warrant inclusion in the PANDAS
subgroup. Usually, however, the temporal relationship can be established
only through prospective documentation of both seropositivity during symptom
exacerbations and seronegativity during periods of remission. This 1 to
1 correlation is necessary to distinguish strep-triggered exacerbations
of the PANDAS subgroup from the more typical waxing and waning course
seen in Tourette’s syndrome and some cases of childhood-onset OCD.
For rheumatic fever, the etiological role of strep infections was demonstrated
indirectly, through three lines of research: (1) epidemiological investigations
which demonstrated a close temporal relationship between scarlet fever
epidemics and subsequent outbreaks of rheumatic fever; (2) the prevention
of rheumatic fever recrudescences by penicillin prophylaxis; and (3) demonstration
of declining rates of rheumatic fever following the widespread application
of antibiotic treatment for GABHS pharyngitis. For the PANDAS subgroup,
there are no epidemiological data that yet demonstrate increased rates
of OCD and/or tics following strep epidemics, although a school-based
study of the relationship between strep infections and neuropsychiatric
symptoms is currently under way.
The
use of penicillin prophylaxis for the prevention of neuropsychiatric symptom
exacerbations was evaluated in a double-blind, placebo-controlled, 8-month-long
crossover study by Garvey and colleagues in 1999. Although individual
cases demonstrated between-phase differences, the trial failed to show
overall superiority of penicillin over placebo. This may have been due
to the failure of oral penicillin to prevent strep infections (14 of the
35 infections documented during the study occurred during the penicillin
phase). Ongoing trials are investigating the utility of other antibiotics
as prophylactic agents for the PANDAS subgroup, but at present, there
are no systematic data to support the use of antibiotic prophylaxis for
children with OCD and/or tic disorders. Similarly, the specific strains
of strep bacteria responsible for symptom onset in the PANDAS subgroup
remain to be identified. Knowing which types of bacteria are capable of
inducing the poststreptococcal sequelae may help elucidate the nature
of the abnormal immune response.
As
discussed above, host susceptibility is likely to be the result of a combination
of genetic, developmental, and immunological factors. Developmental vulnerabilities
are suggested by the increased rates of disease among elementary school–age
children. Rheumatic fever is rare among children younger than 3 years
of age, peaks in incidence during the elementary school years, and declines
in frequency at adolescence to become rare again during young adulthood.
This pattern might reflect developmental differences in immune responsivity,
or it may be merely the result of changing risks of exposure to strep
infections. These are most common during the elementary school years,
and nearly all children develop protective immunity by age 12 to 13 years.
For the PANDAS subgroup, the peak age at onset of symptoms is 6 to 7 years,
with prepubertal symptom onset serving as a defining characteristic of
the subgroup.
Family
history studies have been used to evaluate the role of genetics in host
susceptibility. For rheumatic fever, familial clusters suggested an autosomal
(dominant or recessive, depending on the sample studied) pattern of inheritance.
Preliminary data from 21 patients with SC and 15 children in the PANDAS
subgroup revealed significantly increased rates of rheumatic fever among
the children’s parents and grandparents, in comparison with the parents
and grandparents of 35 healthy controls (5/126 [4.0%], 6/90 [6.7%], and
3/210 [1.4%], respectively; Swedo et al., unpublished data, 2001). The
between-groups differences were small in this pilot data set, but suggested
that children in the PANDAS subgroup may inherit a susceptibility to poststreptococcal
sequelae similar to that reported for children with SC.
Children
in the PANDAS subgroup also appear to have increased rates of OCD and
tics among their family members. In a recently completed study by Lougee
of 54 probands in the PANDAS subgroup (24 with a primary diagnosis of
OCD and 30 with a primary diagnosis of a tic disorder), 21 (39%) had at
least one first-degree relative with a history of a motor or vocal tic
and 14 (26%) had at least one first-degree relative with OCD. Six mothers
(11%), 9 fathers (19%), and 8 siblings (16%) had a motor or vocal tic,
while 10 mothers (19%), 5 fathers (11%), and 2 siblings (5%) had OCD.
These rates are substantially higher than those reported for the general
population and are similar to rates previously reported for childhood-onset
OCD and tic disorders. The combination of increased familial rates of
OCD/tic disorders and of increased rates of rheumatic fever suggests that
children in the PANDAS subgroup may have a dual genetic vulnerability—with
inherited susceptibilities to both OCD/tic disorders and poststreptococcal
autoimmune sequelae. Proof of this hypothesis must come from genetic determinations,
rather than family history studies, and awaits future testing.
At
present, the role of the immune system in the etiology of OCD and tic
disorders is unknown. Clinical observations suggest that symptoms result
from a combination of local, regional, and systemic abnormalities. The
striking effectiveness of immunomodulatory therapies, such as therapeutic
plasma exchange and intravenous immunoglobulin (IVIG), suggests that there
is systemic involvement, at least in severely affected individuals. Magnetic
resonance imaging (MRI) scans reveal enlargements of the basal ganglia,
a finding which points to regional inflammatory changes, while local autoimmune
reactions are suggested by the presence of serum antibodies that cross-react
with neurons of the caudate, putamen, and globus pallidus.
The
effectiveness of both plasma exchange and IVIG in the treatment of severely
affected patients in the PANDAS subgroup suggests that circulating immune
factors play a role in the pathophysiology of the symptoms. Both treatments
have a broad spectrum of potential mechanisms of action, from clearance
of circulating antibodies and cytokines to activation of subpopulations
of T cells and B cells, but the precise mechanism of effect is unknown.
If it could be determined, then it might be possible to elucidate the
nature of the poststreptococcal autoimmune response in the PANDAS subgroup,
as well as to develop targeted therapeutic interventions suitable for
use in less severely ill patients.
Regional
inflammation is thought to play a role in the specificity of the poststreptococcal
neuropsychiatric symptomatology. In SC, functional imaging studies obtained
during the acute symptomatic period have demonstrated increased basal
ganglia blood flow, as well as disruptions of the blood-brain barrier
in the caudate nuclei. These abnormalities resolved as the chorea remitted,
suggesting that they were etiologically related to the neuropsychiatric
symptoms. Volumetric MRI scans have revealed bilateral enlargements of
the caudate, putamen, and globus pallidus in a group of patients with
SC, and similar abnormalities have been demonstrated recently in a group
of patients with OCD/tic disorders. In both samples, however, there was
substantial overlap between the patients and control subjects; thus the
volumetric measurements cannot be used as a diagnostic test.
In
a small series of PANDAS patients treated with plasma exchange, baseline
caudate enlargements returned to normal after successful treatment. This
result suggests that the enlargement might be a reflection of basal ganglia
inflammation. To evaluate this possibility, all children currently being
treated with plasma exchange at the National Institute of Mental Health
are undergoing a series of MRI scans, in which a variety of specialzed
techniques are used to assess interstitial edema and disruptions of the
blood-brain barrier.
The
final area of research interest is the cross-reactive antibodies first
described in SC by Husby and colleagues. The original report describes
the presence of serum antibodies that recognized cells of the caudate
nucleus and subthalamus. The antibodies recognized epitopes on the strep
bacteria as well. It was the cross-reactivity that distinguished the antineuronal
antibodies found in the SC patients from those found in patients with
lupus erythematosus and other neurological disorders.
Several
groups have subsequently reported the presence of “antineuronal”
antibodies in the majority of patients with childhood-onset OCD and/or
tic disorders. These antibodies were directed against a variety of tissue
targets including human caudate, neuroblastoma cells, and rat striatum,
among others. Of note, antibodies were noted to be present frequently
in the serum of healthy children (20%–40%), raising the question
of whether or not the antibodies are the cause of the poststreptococcal
neuropsychiatric symptoms. This question might be answered by cross-reactivity
studies, such as those reported by Husby and colleagues, but these have
not been done in the majority of recent studies. Future research should
include such assessments, as well as techniques that will identify specific
epitopes recognized by the antibodies. If the antibodies are interacting
with a specific receptor or a single neuronal cell-type, they might be
used to open doors to better treatments for childhood-onset OCD and tic
disorders and might lead to a greater understanding of the cause and nature
of these troublesome disorders.
top of page
|
