Scientific basis

Pathophysiology. The current views on the pathomechanisms of post-streptococcal neuropsychiatric disorders, including the evolving concept of PANDAS, are primarily based on research on Sydenham chorea, the prototype for poststreptococcal central nervous system dysfunction. The diagnosis of Sydenham chorea is sufficient to establish the diagnosis of rheumatic fever in a patient. Interestingly, there is significant overlap of symptomatology between patients with Sydenham chorea and patients with obsessive-compulsive disorder (17; 119; 01). Obsessive-compulsive disorder and tics have been documented in a number of patients with Sydenham chorea (70). Overlap between obsessive-compulsive disorder and tics is complex because up to two thirds of children with obsessive-compulsive disorder have tics (58), and 20% to 80% of children with Tourette syndrome have obsessive-compulsive disorder (56). These observations have led to speculation that there are similar pathogenetic mechanisms underlying a variety of phenotypically similar diseases, including Sydenham chorea, PANDAS, Tourette syndrome, obsessive-compulsive disorder, acute disseminated encephalomyelitis, and adult-onset tic disorders (33). The final common target believed to be involved in the pathogenesis of these disorders is the basal ganglia and its connections (95; 101). To date, there is no evidence, however, that PANDAS patients are more prone to other autoimmune diseases, including thyroiditis and celiac disease, and broad autoimmune screening might not be rewarding (111).

Evidence accumulating over the years has suggested that these disorders are autoimmune disorders characterized by the synthesis of potentially pathogenic autoantibodies. In 1976, studies done by Husby and colleagues for the first time showed a higher incidence of antineuronal antibodies in patients with Sydenham chorea compared to controls (39). Antibodies directed against GABHS were subsequently postulated to cross-react with neuronal cells in the basal ganglia by the process of molecular mimicry and cause motoric and behavioral abnormalities as observed in Sydenham chorea and PANDAS patients. Since then, several groups have detected antineuronal antibodies also in pediatric patients with obsessive-compulsive disorder and tic disorders (45; 103; 106; 75; 129). These antibodies were shown to react with subthalamic and caudate nucleus neurons. Circulating antibodies usually do not cross the blood-brain barrier, and several mechanisms have been postulated to explain the possible occurrence of this immune-mediated autoaggressive disease phenomenon. Migration of antigen-specific B cells into the CNS followed by differentiation into antigen-producing plasma cells is one possible explanation. Concomitant disruption of the blood-brain barrier by release of cytokines and subsequent crossing of antibodies into the CNS is another explanation (117). Attempts to detect antineuronal antibodies targeting specific neuroanatomical structures in patients with obsessive-compulsive disorder and tic disorders have shown mixed results. One study found a higher incidence of antineuronal antibodies against caudate nucleus, putamen, or both in children with obsessive-compulsive disorder or obsessive-compulsive symptoms compared to control group (45). However, other studies failed to show a higher incidence of antineuronal antibodies in patients with obsessive-compulsive disorder and Tourette syndrome (114; 72). A study in patients with Tourette syndrome found a higher serum level of antineuronal antibodies against the putamen (but not caudate or globus pallidus) (103). It has been shown that serum autoantibodies do not differentiate PANDAS and Tourette patients from controls (105). These initial studies yielded incongruent results mainly due to methodological reasons related to autoantibody detection. All these studies used techniques that probed antibody reactivity in patients’ sera against tissue preparations that had undergone complex preparation methods; this might have, on one hand, led to the observation of antigen-antibody interactions that could be absent in vivo, and, on the other hand, hindered the identification of other antigen-antibody interactions more likely to exert pathogenic effects in patients. As discussed below, the application of more recent, live cell-based methods appear promising for the elucidation of the role of autoantibodies in these disorders. Also, the search for other immunological markers different from anti-neuronal antibodies has been unrewarding over the past 15 years. A preliminary small scale study of PANDAS patients demonstrated a high proportion of D8/17-positive lymphocytes (118). D8/17 is a monoclonal antibody directed against a non-HLA B-cell marker that is being studied as a potential peripheral disease marker for PANDAS (125). Numerous studies done previously have shown positive reactions to D8/17 in patients with Sydenham chorea, PANDAS, child-onset obsessive-compulsive disorder, Tourette disorder, autism, anorexia nervosa, adult obsessive-compulsive disorder, tics, and trichotillomania (118), but overall results across studies have been inconsistent due to methodological discrepancies. Finally, one study has reported an association between a single nucleotide polymorphism in the promoter region of the TNF-alpha gene and the diagnosis of PANDAS, but its meaningfulness is uncertain until the result is replicated (64).

What do we know of the potential molecular targets of anti-neuronal autoantibodies in post-streptococcal neuropsychiatric disorders? Originally, the M protein of GABHS has been proposed to be the inciting antigen for the production of antineuronal antibodies, specifically M6 and M19 proteins that are believed to share common epitopes with brain structures (39; 09). Later on, in comparative studies involving patients with Sydenham chorea, patients with rheumatic fever without chorea, and healthy controls, 3 distinct basal ganglia antigens of 40, 45, and 50 kDa were detected, distinguishing the first group of patients from the other two. (18). Circulating antibodies in the CSF of patients with Sydenham chorea were also found to cross-react with the same 3 basal ganglia antigens by the same group (19). A subsequent study from these authors using a proteomic approach suggested that these antigens were membrane isoforms of 3 enzymes involved in glycolysis and glycogenesis, namely aldolase C, neuron-specific isoforms of enolase, and pyruvate kinase M1 (21). Although these enzymes may be involved in local ATP synthesis at a membrane level potentially modulating ion channels and pumps, the identification of these autoantibodies and their pathogenic role in post-streptococcal neuropsychiatric disorders remains undemonstrated. Moreover, as highlighted above, all these studies were not based on live cell-based assays, which represent the current mainstay of autoantibody detection methods, and may therefore have detected pathogenically irrelevant antigen-antibody reactivities.

A pathophysiological model suggests the direct immunological involvement of dopamine in PANDAS pathogenesis. In this model, immunologically mediated increases in central dopamine levels and selective modulation of dopamine D2 receptors are key mechanisms that produce neuropsychiatric symptoms. Briefly, increased dopamine concentrations locally inhibit the suppressive functions of regulatory T cells and enhance the activity of B lymphocytes and Th1 cells. This may produce an autoimmune inflammation within the basal ganglia, leading to release of various inflammatory mediators, as well as more dopamine (81). Furthermore, findings by Cunningham and Perry describe cross-reactive antibodies from PANDAS patients that may interact and selectively modulate dopamine D2 receptors (20). This is supported by the live cell-based evidence on antineuronal antibodies in Sydenham chorea and, by extension, the whole spectrum of post-streptococcal neuropsychiatric disorders. Initial work by M. Cunningham’s group highlighted the presence of antibodies that target the dominant epitope of GABHS carbohydrate (N-acetyl-beta-D-glucosamine) and neuronal cell surface molecule lysoganglioside GM1, influencing signal transduction via induction of a calcium-calmodulin dependent (CaM) protein kinase II activity. This activation can lead to increases in dopamine production and release, thereby causing behavioral, learning, and movement control alterations (47; 73; 81). This work has been further extended by Brimberg and colleagues, who developed an animal model in which autoantibodies targeted against D1 and D2 dopamine receptors correlated to the phenotype (see below) (08). Finally, dopamine acting via dopamine receptors significantly increases TNF-alpha secretion by T cells, which increases the permeability of the blood brain barrier (81).

Animal models of Sydenham chorea and PANDAS have demonstrated mixed results also in establishing the role of antineuronal antibodies in the development of neuropsychiatric symptoms (122; 38; 62). Yaddanapudi and colleagues described the correlation of repetitive behavior, motor coordination, and social interaction deficits with peripheral anti-CNS antibodies and immune brain deposits following streptococcal immunization in PANDAS mouse models (132). Alternatively, GABHS-immunized mice demonstrated superior task acquisition and improved capacity for context-dependent performance, supporting the observation of normal intelligence of children prior to acquiring PANDAS. The study also suggested the involvement of humoral immunity in the pathogenesis. Naïve mice transfused with serum from GABHS mice exhibited specific behavioral changes and replicated several aspects of the syndrome. Furthermore, IgG was identified as the major immunoglobulin involved in the mechanism when naïve mice transfused with serum IgG from PANDAS mice exhibited abnormal behavior. Concurrently, transfused donor sera depleted of IgG failed to show alterations in motor and social behavior. A subsequent, potentially break-through study provided important novel evidence on autoantibodies in post-streptococcal neuropsychiatric disorders, especially Sydenham chorea (08). Male Lewis rats were exposed to GABHS antigen and displayed motor abnormalities, such as impaired food manipulation and narrow beam walking (alleviated by a D2 blocker like haloperidol), as well as increased induced grooming (alleviated by a selective serotonin reuptake inhibitor like paroxetine). This phenotype has some face validity for the motor and behavioral anomalies observed in Sydenham chorea and PANDAS. Post mortem, these animals displayed antibody deposition in the striatum, thalamus and frontal cortex, with concomitant alterations in dopamine and glutamate levels in cortex and basal ganglia. Autoantibodies (IgG) of GABHS rats caused elevated calcium/calmodulin-dependent protein kinase II signaling in SK-N-SH neuronal cells, as already detected in sera from patients with Sydenham chorea and PANDAS. Further characterization of the target antigen of these antibodies led to the identification of D1 and D2 receptors as putative autoantigens: it is speculated that antibody binding to presynaptic D2 autoreceptors might lead to excessive dopamine synaptic release via intracellular signalling pathways involving calcium/calmodulin-dependent protein kinase II. This work has been extended showing that IgG from GABHS rats reacted not only to D1 and D2 receptors, but also to 5HT-2A and 5HT-2C serotonin receptors, leading to the suggestion that this antibody response could directly target also the serotonin neurotransmission (63).

The demonstration of potentially pathogenic antineuronal antibodies in patients has proven, however, more difficult. Morris-Berry and colleagues measured in 44 children with PANDAS, 40 children with Tourette syndrome, and 24 healthy peers single-point-in-time ELISA optical densities for 3 putative antibodies previously identified in Sydenham chorea, ie, N-acetyl-beta-D-glucosamine, tubulin, and the dopamine 2 receptor (74). No difference between these groups was observed. Likewise, there was no change in antitubulin and anti-D2 receptor antibody levels when these were assessed in serial samples from 12 PANDAS subjects obtained prior to a documented exacerbation, during the exacerbation (with or without a temporally associated streptococcal infection), and following the exacerbation. Therefore, the role of these autoantibodies in PANDAS patients remains unclear.

Regardless of autoimmune mechanisms, a subgroup of tic disorder patients might manifest an enhanced immune response to GABHS exposure. Bombaci and colleagues performed an observational comparative study on serum antibody responses to GABHS recombinant antigens (06). Analysis was done comparing circulating antibody reactivity in a group of children with chronic tic disorders without evidence of GABHS infection, a group of children with documented GABHS pharyngitis, and a smaller control group of healthy children. The study identified 25 antigens recognized in sera of all 3 groups, 21 antigens recognized only in sera from tic and pharyngitis patients, and a group of antigens that were specifically recognized in sera from tic patients. Although these preliminary findings need to be further investigated using adequate sample size that includes children diagnosed with PANDAS and obsessive-compulsive disorder, these data suggest the presence of an enhanced immunological response of tic disorder patients to a wide range of GABHS antigens. Previous evidence had shown that youngsters with Tourette syndrome (TS) and obsessive-compulsive disorder (OCD) may have a higher average annual rate of new GABHS infections per subject per year compared to healthy peers (65). In line with this, patients with tics or obsessive-compulsive symptoms are more likely to have had a GABHS infection in the 3 months prior to neuropsychiatric symptom onset (68). Subsequent work has shown persistently raised anti-streptolysin O levels in sera of 60% of non-selected children with tic disorders, in the absence of a clear link between symptom exacerbations and recent GABHS infections (66). Additional evidence is also accumulating on possible intrinsic abnormalities in the regulation of immune responses in children with Tourette syndrome. These include reduced percentages of T regulatory lymphocytes (42), which are crucially involved in immune tolerance mechanisms and protect against autoimmunity, and possible abnormalities of immunoglobulin subtype profile (eg, IgA dysgammaglobulinemia) (41; 07), which can also predispose to both upper respiratory infections (explaining the higher rate of GABHS infections in these patients) and to autoimmunity. Finally, GABHS infections might interact with psychosocial stress in predicting future short-term tic severity in youngsters with tic disorders (61).

Imaging studies have been done on PANDAS patients to identify structural and functional changes. Magnetic resonance imaging volumetric studies showed involvement of basal ganglia in PANDAS. A case report showed enlargement of basal ganglia in a patient with PANDAS, and subsequent decrease in the volume of caudate, putamen, and globus pallidus after plasma exchange (31). In a study of 24 patients with Sydenham chorea, enlargement of caudate, putamen, and globus pallidus was noted (30). Structural and functional MRI studies have demonstrated abnormalities of basal ganglia and associated corticostriatal or corticothalamic pathways in patients with obsessive-compulsive disorder (67; 92; 22; 71; 35). More evidence is necessary to support basal ganglia volumetric changes in the acute phase of post-streptococcal neuropsychiatric disorders. Differences in characteristics of the gray and white matter were described in patients with PANDAS when compared with controls (10).

SPECT studies have demonstrated increased perfusion in some patients with Sydenham chorea (57; 02; 32). In patients with obsessive-compulsive disorder, SPECT studies reveal perfusion abnormalities in the fronto-subcortical regions (54; 15). A PET study has explored ongoing neuroinflammation using a tracer for activated microglia (C-[R]-PK11195) in 17 children with PANDAS (mean age 11.4 years), 12 children with Tourette syndrome of similar age, and 15 normal adults (50). The authors found an increased binding potential value in the caudate and lentiform nuclei bilaterally in PANDAS and in bilateral caudate nuclei only in Tourette syndrome. To date, this study represents the first and only evidence of ongoing neuroinflammatory changes in youth with PANDAS.

Diffusion tension imaging measures and cortical features such as thickness, volume, curvature, and gyrification have not shown any significant differences between patients with PANDAS and healthy controls (10).

Differential diagnosis. Table 3, which is adapted from the study by Murphy and colleagues, provides a list of alternative causes that may present with the PANS phenotype and that, therefore, need to be differentiated from PANDAS (79).

Table 3. Differential Diagnosis

Treatment. The treatment of patients with PANDAS is still a subject of considerable controversy. Penicillin prophylaxis of patients with rheumatic fever has significantly reduced the recurrence rate of disease. It has also been shown that antibiotic prophylaxis against GABHS with penicillin or azithromycin can reduce the severity of neuropsychiatric symptoms (110; 107; 05). In patients specifically diagnosed with PANDAS, penicillin prophylaxis failed to show a reduction in clinical exacerbations (29). However, significant improvements of obsessive-compulsive disorder symptoms of patients with PANDAS has been demonstrated with antibiotic treatment (77; 24; 78). A pilot trial of cefdinir 14 mg/kg in 20 children for PANS showed improvements in obsessive-compulsive symptoms and tics (79). Similar results were found in a double-blind, randomized placebo study with 4 weeks of azithromycin treatment in youth with PANDAS (78). The same authors presented the issue that antibiotics may act through another mechanism besides prevention of GABHS reinfection (81). One theory is that penicillin reduces antigenic load from an undetected, asymptomatic GABHS infection. Another theory involves a synergistic role of penicillin in reducing proinflammatory cytokines induced by group A Streptococcus, specifically interferon gamma.

Another interesting drug class that still lacks supportive evidence is selective serotonin reuptake inhibitors (SSRIs). SSRIs are drugs of choice for the treatment of obsessive-compulsive disorder (100). Concomitantly, they are also found to be anti-inflammatory, mechanistically through the suppression of interferon gamma (81).

More aggressive immunomodulatory therapy in the form of plasma exchange and intravenous immunoglobulin has also been used for symptoms associated with PANDAS. A randomized, controlled trial of 20 pediatric patients that studied the effects of IVIg or plasma exchange versus placebo in the treatment of obsessive-compulsive disorder and tic disorders showed promising results (91). In the treatment of obsessive-compulsive symptoms, treatment resulted in a 45% (IVIg) to 58% (plasma exchange) decrease in symptoms. In the treatment of tic disorders, plasma exchange showed a 49% decrease in symptoms. At 1-year of followup, 82% showed continued improvement. Longer-term follow up of this study has not been published at this time. On the contrary, Hoekstra and colleagues demonstrated no improvement in tic symptoms in a randomized trial using IVIg versus placebo in an older series of patients ages 14 to 53 years (37). It is important to recognize that both IVIg and plasma exchange therapy are not without potentially serious side effects. Even though a 12-patient case series with moderate/severe PANDAS showed a benefit of IVIg treatment in clinical practice (49), a randomized controlled study in 2016 revealed no significant difference between the IVIg and placebo groups in 35 children with PANDAS diagnosis (131).

Cognitive-behavioral therapy has been studied in the treatment of PANDAS-associated obsessive-compulsive disorder (112; 86). In a small series of 7 children with PANDAS treated for 3 weeks with intensive cognitive-behavioral therapy, 6 of 7 showed improvement using the Children's Yale-Brown Obsessive-Compulsive Scale and Anxiety Disorder Interview Schedule for DSM-IV Child Interview Schedule-Parent version. At long-term followup, 3 of 6 showed continued clinical improvement (112).

Investigations have identified PANDAS disease as an indication for tonsillectomy. Batuecas Caletrio and colleagues presented a case report on a 9-year-old boy diagnosed with PANDAS disease, presenting with recurrent bacterial tonsillitis, worsening eye tics and facial grimaces, and symptoms of obstructive sleep apnea syndrome (03). The patient underwent tonsillectomy for obstructive sleep apnea and showed disappearance of eye and facial tics. Few other reports mention improvement of neuropsychiatric symptoms of obsessive-compulsive disorder in PANDAS patients after tonsillectomy (88; 36). On the other hand, it was shown that youth with PANS/PANDAS had higher rates of tonsillectomies and adenoidectomies prior to the onset of obsessive-compulsive symptoms and tics (82). Further studies and clinical trials need to be done to define the role, if any, of tonsillectomy as a treatment option for patients with PANDAS and to determine the part that other related treatment components, such as anesthesia, may have played.

Discussion

The clinical definition of PANDAS has been the object of intense debate throughout the past 15 years (53), opposing 2 main different views, with some authors arguing that PANDAS may not exist and the connection between obsessive-compulsive disorder or tic disorder and streptococcal infection is merely coincidental, and others supporting the existence of this entity. Undoubtedly, GABHS infection is very common in children, and it is likely that evidence of recent GABHS infection is found by chance in large numbers of children. Furthermore, obsessive-compulsive disorder occurs in 1% to 2% of school-age children (133), and transient tic disorders can be detected in 10% to 25% of early school-age children (109). The complex behavioral spectrum exhibited in the acute phase by children diagnosed with PANDAS, summarized above, encompasses several features that have been previously reported during the acute phase of illness in patients with Sydenham chorea (117) as well as in patients with acute rheumatic fever without chorea (69). This has led some to hypothesize that PANDAS could represent a neuropsychiatric presentation of rheumatic fever in the absence of florid chorea. However, this hypothesis was soon refuted, mainly due to the absence of other rheumatic features in the vast majority of children diagnosed with PANDAS. More than other features, the lack of echocardiographic changes suggesting carditis is strikingly different from Sydenham chorea patients, in whom the frequency of carditis ranges between 30% and 80% of cases (108; 97). Other authors, however, have shown subtle echocardiographic changes in children with tic disorders, although it is unclear how many of these had PANDAS (13). One of the biggest problems in the routine clinical application of Swedo and colleagues’ working criteria for PANDAS is the need to obtain serial measurements of serological markers (eg, anti-streptolysin O, anti-DNAseB antibodies, or preferably both) and correlate them with the severity of behavioral symptoms. Indeed, it has been clarified that serological streptococcal markers may show several different temporal fluctuation patterns across different individuals, depending on the type of exposure to the pathogen and the degree of immune activation in the host (40). This minimizes the clinical significance of single time-point cultures and antibody measurements. Simultaneously increased levels of ASO and anti-DNAseB were shown in patients with PANDAS compared to controls (16). However, positive results only indicate exposure to the streptococcal infection, and they do not differentiate between the carrier state and an acute infection (23).

Nevertheless, in real practice, and particularly in community clinics, the diagnosis of PANDAS has been often formulated on the basis of a cross-sectional association between streptococcal markers and behavioral symptoms, which may well be coincidental (27). It is, therefore, not completely surprising that multicenter longitudinal studies comparing children with a diagnosis of PANDAS to children with non-PANDAS tics or obsessive-compulsive symptoms failed to detect a higher number of streptococcus-linked clinical exacerbations and immune markers profiles in either of the 2 patient groups (52; 102; 55). Despite their high rate of clinical assessments, suggesting a robust longitudinal observation, patients diagnosed with PANDAS in these studies might have been clinically different from patients with PANDAS described by Swedo and colleagues in their original report. This potential heterogeneity generated by the application of Swedo and colleagues’ working criteria by different clinicians in different settings raised the suspicion that such criteria may lack validity and reliability, requiring a revision (121).

One first useful approach to overcome this hurdle is to search for specific phenomenological features, different from those highlighted by Swedo and colleagues’ criteria, which might help in differentiating PANDAS from non-PANDAS clinical look-alikes. A study done by Bernstein and colleagues identified specific clinical characteristics of children with PANDAS compared to a group of children with non-PANDAS obsessive-compulsive disorder (04). They described patients with PANDAS as more likely to present with separation anxiety, urinary urgency (but not enuresis), hyperactivity, impulsivity, deterioration in handwriting, and decline in school performance. Additionally, total and vocal tics were noted to be more severe in children with PANDAS. Murphy and colleagues examined 109 children with tics, obsessive-compulsive disorder, or both with personal and family history, diagnostic interview, physical examination, medical record review, and measurement of baseline levels of streptococcal antibodies (84). Children with PANDAS were more likely than non-PANDAS to have had dramatic onset, definite remissions, remission of neuropsychiatric symptoms during antibiotic therapy, a history of tonsillectomies/adenoidectomies, ADHD comorbidity (61% vs. 31%), evidence of group A streptococcal infection, and clumsiness. This last observation is particularly striking in that it confirms Bernstein and colleagues' findings, as well as earlier observations from Swedo’s team, of worse performance on fine motor tasks (including drawing and writing) during the acute phase of PANDAS. It also fits nicely with the behavioral regression and oppositionality that these patients are thought to display.

Despite these attempts to better characterize specific clinical features of the PANDAS phenotype, this concept is currently undergoing substantial nosological reappraisal. The rationale beyond this derives from different points of concern, some of which have been highlighted above. First of all, tics and obsessive-compulsive disorder can fluctuate without any evidence of GABHS infection or change in antibody titer. In patients with Tourette syndrome, tics are known to worsen during periods of stress or illness. Hence, exacerbations with streptococcal infection may represent a nonspecific response to stress (51). Moreover, there is anecdotal evidence suggesting tic exacerbation after Mycoplasma pneumoniae (76) or other types of infection (01) and raised anti-Mycoplasma antibody response in Tourette syndrome patients (76). Murphy and colleagues intriguingly highlight a clinical overlap between the neuropsychiatric manifestations of Lyme disease in children and PANDAS/PANS (79). Without a specific biological marker or radiographic feature of disease, the entity of PANDAS may be nonspecific. The proposed diagnostic criteria suggested by Swedo may not be specific for PANDAS and encompass features of numerous other disease processes. On the other hand, proponents of PANDAS suggest that a substantial number of cases of patients with tic disorder and obsessive-compulsive disorder are in fact PANDAS, based on a common pathway involving basal ganglia and their respective circuitry. Successful treatment of symptoms with immunomodulatory therapy further supports the concept of PANDAS as an immune-mediated autoaggressive disease. A genetic predisposition has been suggested for both obsessive-compulsive disorder and tic disorder (90; 127). Environmental causes may also play a role, and the genotypes of individuals may be important in determining the response to the environmental stimuli, rather than being a primary cause of the illness. A second relevant point of concern is related to the diagnosis of group A streptococcal infections, which may not be straightforward. The growth of this pathogen from throat specimens may indicate a carrier state rather than a recent infection, and serial antistreptococcal antibody measures are required to distinguish between the two. At the same time, antistreptolysin O antibodies are not specific of group A streptococci but may also rise with exposure to group C and G streptococci. The temporal latency between an inciting group A streptococcal infection and the onset of PANDAS is also not well defined; for a neuropsychiatric entity with well-established post-streptococcal etiology, such as Sydenham chorea, this may be as long as 6 months, and it is currently uncertain if and why this latency should be shorter for PANDAS. As summarized above, moreover, some clinical features such as the abrupt, dramatic onset and the co-occurrence of specific behavioral features other than tics and obsessive-compulsive symptoms are not comprised in the working criteria by Swedo and colleagues. Finally, the phenomenological distinction between chorea in Sydenham chorea and “choreiform movements” described by Swedo and colleagues in PANDAS seems mainly based on the severity of the hyperkinetic symptoms rather than by a qualitative difference. These observations formed the rationale beyond the proposal of broader concepts such as childhood-onset acute neuropsychiatric syndrome (CANS) and pediatric acute neuropsychiatric syndromes (PANS), theoretically encompassing also the PANDAS phenotype.

In summary, the diagnosis of PANDAS and related conditions is purely clinical, and it requires a complete medical, neurologic, and psychiatric history; a thorough physical examination; a series of laboratory tests (to exclude other causes); and a comprehensive neuropsychological assessment. Serum and CSF autoimmune antibodies may be considered based on the clinical presentation and the need to exclude autoimmune encephalitis. Given its poor specificity, a Cunningham Panel is not routinely recommended (94).

Resolution

Although the growing number of cases reported may provide further support to the PANDAS concept, it seems that the proportion of PANDAS among the cases of obsessive-compulsive disorder or tic disorder is probably relatively low (51). Unfortunately, in the wake of gaining further information on the accuracy of novel potential biomarkers like antidopamine receptor antibodies, at present there are no diagnostic tests available that positively establish the diagnosis of PANDAS. So far, diagnosing PANDAS has been challenging, especially in the community setting, due to the difficult application of the Swedo criteria in routine clinical setting (see Discussion). In a retrospective cross-sectional observational study done by Gabbay and colleagues, 31 of 176 patients were diagnosed with PANDAS by community physicians (27). However, 19 of these failed to meet the criteria of Swedo and colleagues (listed in Table 1), most with no laboratory confirmation of GABHS infection; furthermore, 7 of the 145 patients not diagnosed were found to meet the PANDAS criteria at a specialty clinic. This demonstrates significant overdiagnosis as well as underdiagnosis of PANDAS in the community setting (99). Considerable controversy exists as to what clinicians should do when they see a patient with an acute onset of tic disorder or obsessive-compulsive disorder. It is prudent to evaluate the patient for streptococcal infection by a throat culture. Although the information provided by antistreptolysin O and antideoxyribonuclease B may not be helpful in detecting active infection, serum antibody titers may correlate with recurring symptoms.

The definition of new clinical concepts such as pediatric acute neuropsychiatric syndromes (PANS) and childhood-onset acute neuropsychiatric syndrome (CANS) aims to overcome the current limitations in diagnosis post-streptococcal acute neuropsychiatric disorders. At difference from CANS and the former PANDAS criteria, tics lost central relevance in PANS; this is partially due to the general character of tics, which fluctuate over time, sometimes markedly and more intensely than obsessive-compulsive symptoms, often depending on the exposure to different contextual factors, which may include infections amongst others. Nevertheless, tics are acknowledged amongst the motor abnormalities that may constitute accompanying symptoms to either of the 2 core features of PANS. Another difference between these 2 new concepts is the central relevance given in PANS to anorexia, which constitutes, together with obsessive-compulsive disorder, a primary criterion. These differences notwithstanding, CANS and PANS represent highly similar concepts. If the common approach underlying both concepts is to be taken forward in the attempt to identify the features of a “post-streptococcal” subtype of CANS/PANS, then additional, preliminary efforts to homogenize the 2 concepts are necessary. On the other hand, pursuing clinical subtyping of the 2 concepts independently would generate great confusion to the clinicians and might lead to lack of future development and clinical application for both CANS and PANS.

Treatment of a child with antibiotics for an abrupt onset of tics and obsessive-compulsive disorder in the absence of a negative throat culture is not recommended. Likewise, one does not treat elevated antistreptolysin O or antideoxyribonuclease B titers alone. On the other hand, a patient with new-onset Sydenham chorea meets criteria for rheumatic fever and should be appropriately treated. Until data providing the efficacy of antibiotic prophylaxis are available, continuous treatment with antibiotics for fluctuating tic or obsessive-compulsive disorder symptoms is not recommended. It appears reasonable to treat symptoms of potential PANDAS cases with conventional treatment for obsessive-compulsive disorder and tic disorder until new data show the usefulness of other therapeutic interventions directed against GABHS or autoimmune mechanisms. Immunomodulatory therapies should be considered experimental until further clinical trials clearly show their efficacy, particularly considering their potentially serious adverse effects (128; 96).

As our comprehension of the basic science of tic disorders, obsessive-compulsive disorder, and movement disorders increases, perhaps a better understanding of PANDAS will be established. It is hopeful that many of the controversies existing today with regards to the diagnosis and treatment of PANDAS will be resolved by well-designed, prospective clinical trials.