Antibiotic Treatment Trial for the PANDAS/PANS Phenotype

Accumulating evidence suggests that a subset of children with obsessive compulsive disorder (OCD) have a symptom course that is temporally associated with infections (group A Streptococcus, Mycoplasma, Borrelia burgordfi, etc.). The intent of this proposal is to test the hypotheses that: 1) this subset of children with OCD known as Pediatric Acute Onset Neuropsychiatric Syndrome (PANS) and the specific subset associated with Streptococcus, PANDAS (Pediatric Autoimmune Neuropsychiatric Disorder Associated with Streptococcus), experience symptom onset and exacerbations due to autoimmune/inflammatory responses to infectious triggers and, 2) antibiotic treatment will reduce symptom severity. Investigators have reported improvement in neuropsychiatric symptoms of patients presenting with PANDAS following antibiotic treatment (Murphy \& Pichichero, 2002; Snider, Lougee, Slattery et al., 2005). Difficulties with study design and small sample size of early antibiotic trials limit the influence of their findings (Kurlan \& Kaplan, 2004) although anecdotally, many families report often dramatic improvements in OCD symptoms after antibiotic therapy.

The proposed study is a 4 week double-blind placebo-controlled study of the efficacy of azithromycin once daily on OCD symptom severity in children ages 4-14 years presenting with the PANS phenotype. An 8 week open label phase will be offered to all subjects completing the 1-month trial to assess longer term efficacy and tolerability (12 weeks total antibiotics for those on active, 8 weeks for those on placebo). At the end of week 8, the child can transition to best clinical care practice (cognitive behavioral therapy, additional medication trials, etc) if not fully remitted. The rate and timing of any relapse will be assessed but are not the primary outcome variables. Evaluations will consist of: 1) semi-structured psychiatric diagnostic interviews and rating scales to establish psychiatric diagnoses and monitor severity of psychiatric symptoms, particularly obsessive compulsive behavior and tics; 2) parent and child-report questionnaires to elicit information regarding infections and clinical correlates of PANS attributes; 3) physical and neurological examinations; and 4) laboratory studies.

Aim 1: To compare, in randomized controlled fashion, placebo vs. antibiotic on changes in overall symptom severity for obsessive-compulsive symptoms among children with PANS phenotype. Changes from baseline and clinical status at end of study will be compared for the study arms. Primary endpoint will be change on the Children's Yale Brown Obsessive Compulsive scale (CY-BOCS). Secondary measures of interest include severity and improvement measures on the Clinical Global Impression Severity/Improvement scale, quality of life measures, and weekly ratings of mood, anxiety, attention deficit hyperactivity and tic symptoms. Hypothesis: Children receiving antibiotic will show significantly greater overall improvement in severity compared with placebo. Children that meet PANDAS/PANS criteria and have current OCD symptoms of less than 6 months duration will be enrolled in a double-blind placebo-controlled randomized trial of azithromycin for 4 weeks. Subjects will be followed weekly during the 3 months (4-week antibiotic treatment trial, 8 weeks open label treatment). Although earlier trials have had mixed results, these studies have primarily enrolled subjects with longer illness durations and used traditional prophylactic doses in contrast to the standard treatment dose of this study. Since treatment doses of antibiotic decrease antigenic load, one can speculate that treatment level doses, rather than prophylactic doses, may best attenuate neuropsychiatric sequelae. Secondary neuromodulating or immune modulating properties at higher dose of antibiotics may also be a possible mode of therapeutic action. The timing of treatment response and moderators and predictors of response will also be assessed.

Aim 2: To assess immune risk factors in children with PANDAS/PANS. The data and biological specimens (blood, cultures, etc) for all subjects consented for Aim 1 will, be entered into a database and repository to assess further these clinical phenotypes. Subjects that do not qualify for randomization (medication allergy or current long term antibiotic therapy) or decline randomization (e.g. travel required for future visits) will be permitted to contribute to Aim 2, i.e. collection of baseline symptom severity and immune marker assessment. Hypothesis: It is hypothesized that children with sudden onset of OCD and whose subsequent course shows dramatic fluctuations will have evidence of immune risk factors that predisposes to this presentation. Certain subject characteristics such as age, duration of illness, characteristics of exacerbation episodes, type and number of prior infections, and co-occurring disorders may influence treatment response. Immunologic markers that may help with determining pathobiology and treatment response will be explored including pathogen detection array to identify potential infectious triggers, cytokine array to explore inflammatory processes and specific antibody assay to assess potential mediators. In order to differentiate those children that may most benefit from antibiotic prophylaxis from those that do not benefit, analytic strategies to refine an antibiotic-responsive phenotype will be explored.

Rationale: Based on the PI's observations from research and clinical patients being referred for PANDAS (60 patients per year), patients have had most improvement when higher doses of antibiotics are used, often relapsing on prophylactic doses. The best response rates have been antibiotics designed to treat beta lactamase co-pathogens such as azithromycin, cephalosporins or amoxicillin/clavulanic acid. Observed response rates are typically within 3 weeks although some subjects appear to take longer to gain response. These observations deserve further study. Azithromycin was chosen due to improved tolerance, once daily dosing regimen, coverage against other microbes implicated in PANS and low risk for allergic reaction or exclusion.

Randomization: Study medication and matching placebo will be provided by the research pharmacy. Some of the services provided will include blinding of study drug, packaging and labeling, and randomization. Subjects who are randomized will receive either antibiotic or placebo in a 1:1 ratio. Both investigator and study participant will be blinded to the treatment assignment. Randomization will not be revealed to the investigator or the study participants until all participants have completed the study or in the event of serious adverse event (SAE).

Dosing: Azithromycin or placebo every 24 hours will be dispensed. Antibiotic will be transferred to identical bottles as placebo and dispensed without the manufacturer label. Labels will be identical with matching dose volumes for each. Flavor, color and texture will be matched to best of ability. Participants and parents will be instructed to not discuss the actual medication with anyone in the study team except the designated person dispensing the medication.

Study participation is entirely voluntary. If at any time a participant wishes to discontinue the use of study medication, they can notify the study coordinator and if the family is willing, they can choose to enter the naturalistic phase or discontinue the study.

Naturalistic Observation Phase: After the subject's completion of the active phase of the study (baseline to end of week 12 as determined by subject eligibility, medication tolerance, etc), the family will be given the choice to opt in or opt out of a naturalistic follow up on a monthly basis until end of week 52. The family will be provided a monthly PANS scale and questionnaire regarding the child's current treatment. Depending on which option works best for them, they can return by the forms by mail or complete online.

Compliance Monitoring: Azithromycin, with daily dosing, has higher rates of adherence than other antibiotics with more frequent dosing. The parent of the child will be educated about optimal dosing and compliance. Each unit dose will be labeled specifically with the participants identification number. The clinical coordinator will document each batch of medication on the dispensing log. Each batch of study medication will include supply until next scheduled visit plus one week to cover any difficulty returning to study in the scheduled interval. All study medication will be accounted for throughout the study by the investigator or the designee.

End of Treatment Alternatives: Standard of care treatment options are unavailable for the initial 8 weeks, after which point families will have multiple options for post-study care if needed. Until better guidelines are developed, children in the PANDAS/PANS subgroup should also be given the chance to benefit from evidence based treatments that may lessen the severity of a future flare up. From our previous work (Storch, Murphy, Geffken et al., 2006), we believe that remaining OCD symptoms are best treated with CBT unless the child is still too severe to engage in therapy. CBT can help children remodel automatic responses to obsessions, teach skills that should prove helpful if symptoms do recur and also help families with behavioral strategies to lessen the risk of disrupted functioning and accommodation. Children often report feeling empowered from coping, relaxation, and resiliency skills learned in cognitive behavioral therapy (CBT). Antidepressants approved for OCD are an option to consider for those that appear to have a chronic stable course but many youth with a PANS presentation are incapacitated by these sudden and severe symptoms, with many parents feeling too desperate (Murphy, observations) to wait the typical 10-12 weeks for SSRIs to achieve full efficacy. Many of these children are highly sensitive to usual starting doses (Murphy, Storch, \& Strawser, 2006) but do well when started on a low dose (e.g. sertraline at 6.25mg) and gradually increased as tolerated. Patients with tics respond well to a variety of evidence based pharmacologic agents and behavioral treatments if needed due to symptom severity and impairment. Following the study all subjects with incomplete remission will be referred for CBT in the outpatient clinics or the community based on parent preference. The investigative team has considerable experience in the application of this approach for pediatric OCD (Murphy et al., 2007) and collaborates closely with outpatient providers.

Adverse effects: Every effort should be made to identify prior history of adverse effects to antibiotics prior to the randomization phase (exclusion criterion). In the event of Candidiasis, other opportunistic overgrowth or allergic reaction, the recommendations by the primary care physician (PCP) of the child will be followed. As this adverse event is unlikely except during treatment with antibiotics, blinding will be compromised and the child will be considered a drop due to adverse event. Open label probiotics will be provided to subjects in both study arms in all phases to be taken daily. In the event of an SAE, the blind should be broken only if knowing the treatment status is of significance to the course of treatment. If at any time the blinding becomes otherwise compromised, the subject will exit the randomized control trial (RCT) portion of the study. The date, time, and reason or situation surrounding unblinding must be documented as completely as possible. All SAEs should also be reported to the institutional review board (IRB) as soon as possible. Laboratory tests (CBC, metabolic panel, urine toxicology and pregnancy test [for post-pubescent females]) and an EKG will be obtained at baseline. Every 2 weeks, liver function testing will be done to monitor for hepatic toxicity risks as well as an EKG to monitor for cardiac risks (increased QTc).