Infections,Antibiotics,and Mental Health Deteriorations

To the Editor:

Recently, two population-based studies from Denmark have reported an association between treated infections and subsequent risks of mental health disorders, including obsessive compulsive disorder (OCD), tics, and eating disorders (Breithaupt et al. 2019; Köhler-Forsberg et al. 2019). The hazard ratios of developing mental disorders were reported to be much increased in the 6 months after the use of an anti-infective agent. We, clinicians managing patients with pediatric acute-onset neuropsychiatric syndrome (PANS) or pediatric autoimmune neuropsychiatric disorder associated with streptococcal infections (PANDAS) in a multidisciplinary clinic (Chang et al. 2015), have received multiple queries from clinicians and families about potential adverse effects of antibiotics on the microbiome, autoimmune disorders, and neuropsychiatric symptoms. Harmful effects of antibiotics have been implicated in juvenile idiopathic arthritis (Horton et al. 2015), and additionally some basic science studies showed direct effects of antibiotics on neuronal function and behaviors.

To elucidate the temporal relationships between antibiotic exposure and psychiatric symptom onset, we systematically reviewed electronic medical records (EMRs) of consecutive patients who meet PANS/PANDAS criteria, and presented to our multidisciplinary clinic from September 2012 to April 2019, within 4 months of psychiatric symptom onset for their antibiotic exposure preceding the onset of psychiatric symptoms. To capture all antibiotic exposures, three patients whose primary care EMR system was not connected to our EMR system (EPIC) were excluded. Our routine evaluation of every patient admitted to our clinic comprises a thorough review of all medications including antibiotics and illnesses preceding the onset of psychiatric symptoms, which is accomplished through four approaches: in-person interviews, systematic review of all EMR documents and prescriptions, preclinic parent questionnaire, and, when indicated, review of medical records outside of our EMR system.

Our study cohort included 88 patients. Mean age (SD) of PANS onset is 8.7 (3.4) years; male gender and non-Hispanic white represent more than half of patients (Table 1). Fifty-three (60%) of them had history of infections within 12 weeks preceding the onset of psychiatric symptoms, in which only 12 (14%) patients received antibiotics (a 7–10 day course of amoxicillin ± clavulanic acid, or penicillin, exclusively for streptococcal throat, sinusitis, and/or otitis media). In contrast, the majority of our patients (74/88, 84%) had no documented antibiotic exposure in the 6-month period preceding the development of psychiatric symptoms. Streptococcal infections (33%) were twice more likely than antibiotic exposure (14%) in our patients within 12 weeks before the onset of psychiatric symptoms. Our study is limited by recall bias with regard to infections and antibiotics not documented in medical charts, for example, during vacation or infections that do not come to medical attention. We may have missed streptococcal infections with mild symptoms.

In conclusion, we do not believe that sudden onset of OCD, eating restriction, tics, etc. is caused by antibiotics, and may be more likely the result of immunologic responses to infections or immune stresses, as in the cases of acute rheumatic fever or Sydenham chorea.