PANDAS was first described as a subgroup of children presenting with abrupt-onset obsessive-compulsive disorder or tic disorders temporally linked to Group A β-hemolytic Streptococcus exposure. The proposed pathophysiology involves molecular mimicry, in which antibodies generated against streptococcal antigens cross-react with neuronal targets, particularly within cortico-striato-thalamo-cortical circuits. Experimental studies have demonstrated that streptococcal-specific Th17 lymphocytes may promote blood–brain barrier dysfunction, facilitating the entry of immune mediators into the central nervous system and amplifying local inflammatory responses within basal ganglia circuits. Rather than representing a simple consequence of elevated anti-streptococcal antibody titers, disease activity is increasingly understood as the result of a complex interaction among persistent antigenic stimulation, adaptive immune dysregulation, endothelial activation, and host genetic susceptibility. Circulating antibodies such as antistreptolysin O (ASO) and anti-DNase B are generally regarded as markers of previous streptococcal exposure rather than direct indicators of ongoing neuroimmune injury.

The patient in this case began abruptly in January 2024, when an episode of pharyngitis swept through her school classroom following a birthday party. Five of seventeen children developed pharyngitis, reportedly diagnosed as streptococcal infection. Approximately fifteen days later, the girl experienced a dramatic neuropsychiatric episode—her parents reported she did not recognize her mother and appeared markedly agitated and confused. Within two days, she developed high fever exceeding 39°C, headache, and neck stiffness, leading to hospital admission. Her ASO titer was greater than 1:800. Despite initial improvement with penicillin and subsequent discharge, neuropsychiatric symptoms progressed over several weeks, profoundly affecting family life, academic performance, sleep quality, and social interactions. By March 2024, clinical evaluation was considered consistent with PANDAS according to contemporary diagnostic criteria, supported by documented elevated anti-streptococcal markers including an ASO titer of 1:800 and anti-DNase B of 460 U/mL.

Over the following months, the patient underwent an extensive multidisciplinary therapeutic program. Oral penicillin initially produced clinically meaningful improvement, with reduction of anxiety, obsessive-compulsive symptoms, emotional dysregulation, and tic severity—but the benefit consistently appeared approximately six days after treatment initiation and remained incomplete and transient. Because of repeated relapses, long-term intramuscular benzathine penicillin prophylaxis was introduced between June 2024 and January 2025, again producing only partial clinical stabilization. As disease severity increased, immunomodulatory therapy was progressively escalated: seven sessions of therapeutic plasma exchange were performed, followed by three cycles of intravenous immunoglobulin at 35 grams administered over two consecutive days per cycle. Concomitant azithromycin therapy and low-dose aripiprazole were also prescribed. Despite these interventions, the patient continued to experience disabling neuropsychiatric manifestations with only limited and temporary benefit. Her Children’s Global Assessment Scale score was 31, indicating severe impairment in overall functioning.

Repeated microbiological investigations did not provide an active Group A Streptococcus isolate suitable for pathogen-specific phage susceptibility testing. The decision to initiate bacteriophage therapy was therefore not based on direct microbiological confirmation or in vitro phage sensitivity but on a reproducible clinical-biological signal: symptom improvement repeatedly occurred after anti-streptococcal penicillin administration with a latency of approximately six days. This temporal pattern suggested that persistent or recurrent streptococcal involvement remained biologically plausible despite the absence of an identifiable bacterial isolate. The therapeutic rationale was grounded in the possibility that repeated antimicrobial exposure may reduce bacterial recovery while allowing bacterial persistence within biofilms, intracellular niches, or extra-pharyngeal reservoirs that routine testing cannot detect.

Empirical bacteriophage therapy was performed using PYO Bacteriophages, a commercially available standardized phage polyvalent cocktail targeting Streptococcus spp., Staphylococcus spp., Pseudomonas aeruginosa, Escherichia coli, and Proteus spp., produced by the Eliava biopreparation in Georgia. Published descriptions report an approximate overall phage concentration of 10⁵ plaque-forming units per milliliter for most components and at least 10⁴ PFU/mL for the streptococcal component. The preparation was administered orally at a 10 mL dose twice daily for three cycles of fifteen days. Because the treatment was experimental and the patient had ongoing disease activity, bacteriophages were administered as add-on therapy while antimicrobial treatment with penicillin was maintained—a point the authors emphasize is critical for interpretation, since the clinical response cannot be attributed to bacteriophage therapy alone.

Improvement became clinically evident approximately six days after initiation of bacteriophage therapy, remarkably reproducing the response latency previously observed after penicillin administration. During the subsequent six weeks, progressive reduction of obsessive-compulsive symptoms, anxiety, emotional dysregulation, motor and vocal tics, sleep disturbance, and oppositional behavior was observed. Functional recovery was documented by substantial improvement in family interactions, school participation, peer relationships, and daily activities. Parent-reported Strengths and Difficulties Questionnaire Follow-up Impact Supplement scores decreased from 15 out of 15 before bacteriophage therapy to approximately 6 out of 15 following treatment, corresponding to an estimated 60 percent reduction in global symptom-related impairment. After completion of three bacteriophage cycles, the patient showed sustained clinical stabilization. Planned escalation of aripiprazole was no longer required, and dosage was lowered in the final month. Her body weight increased from 29 kilograms before treatment to 32 kilograms during follow-up, while her height increased from 141 centimeters to 146 centimeters—particularly relevant given that her nutritional status had deteriorated to the point that enteral nutritional support through a feeding tube had been considered. No clinically significant adverse events, allergic reactions, or gastrointestinal disturbances were observed.

One of the most unexpected observations concerned the evolution of anti-streptococcal serology. Rather than declining during clinical recovery, ASO titers increased from 170 IU/mL to 580 IU/mL after the third cycle, while anti-DNase B rose to 920 U/mL—yet neurological symptoms continued to improve. This dissociation between increasing anti-streptococcal antibody titers and sustained clinical recovery constitutes one of the hypothesis-generating observations of the report. One possible explanation is that bacteriophage-induced bacterial lysis transiently increased the exposure of the immune system to streptococcal antigens. Enhanced antigen release could stimulate additional antibody production despite simultaneous reduction of the viable bacterial burden responsible for maintaining chronic immune activation. Under this hypothesis, increasing ASO concentrations would reflect an immunological response to antigen release rather than persistence or worsening of pathogenic disease activity. Alternative explanations remain entirely possible—the observed antibody kinetics may simply reflect the natural evolution of humoral immunity, assay variability, or unrelated fluctuations in antibody production.

Bacteriophages offer several theoretical advantages over conventional antibiotics. Their high specificity minimizes collateral disruption of the commensal microbiota, while their capacity to replicate selectively within susceptible bacterial populations may facilitate eradication of persistent bacterial reservoirs. Experimental evidence also suggests that bacteriophages may penetrate bacterial biofilms and act synergistically with antibiotics, potentially enhancing bacterial clearance in chronic infections. Beyond their direct antibacterial activity, bacteriophages have increasingly been recognized as biological modulators capable of interacting with innate and adaptive immune responses, with experimental studies demonstrating effects on macrophage activation, cytokine production, dendritic cell function, and mucosal immune homeostasis. Growing experimental evidence indicates that communication between peripheral infection and the central nervous system involves dynamic interactions among adaptive immunity, endothelial integrity, and neuroinflammation. In parallel, experimental studies have suggested that bacteriophages may traverse epithelial and endothelial barriers, including, under selected experimental conditions, the blood–brain barrier—though there is currently no evidence that this occurred in the present patient.

The authors are careful to acknowledge the substantial limitations of this report. It is a single uncontrolled case from which no causal inference can be made regarding bacteriophage efficacy. Bacteriophages were administered as add-on therapy while antibiotics were maintained, meaning a phage-specific effect cannot be separated from phage-antibiotic synergy, delayed antibiotic effects, or combined treatment effects. No active streptococcal isolate was available, and no phage susceptibility assay was performed—the treatment was empirical rather than microbiologically personalized. Response assessment was based on clinical observation, caregiver report, and functional improvement rather than standardized neuropsychiatric instruments. The clinical course of PANDAS can fluctuate with spontaneous exacerbations and remissions, and the patient received multiple prior and concomitant interventions that may have contributed to clinical stabilization. Nevertheless, the case raises a potentially important biological question: sustained neurological and psychiatric improvement occurred despite a marked increase in anti-streptococcal antibody titers, suggesting that humoral immune responses and clinical disease activity may become dissociated during effective pathogen-directed treatment. The central hypothesis is not that bacteriophages directly treat obsessive-compulsive disorder or tics, but that they may reduce a microbial driver contributing to immune-mediated dysfunction in susceptible patients. Prospective studies are needed to define patient selection criteria, safety, microbiological targets, immune correlates, and objective clinical outcomes before this experimental approach can be evaluated in controlled clinical trials.

Subject of Research: Health & Medicine

Article Title: Clinical improvement despite rising anti-streptococcal antibody titers during adjunctive bacteriophage therapy in refractory PANDAS: a hypothesis-generating case report

Article References: Recordare, A., Dadiani, M., Nizharadze, D., Nadareishvili, L., Pirozzolo, G., Palumbo, R., Girotto, C., Mezzapelle, G., & Kutateladze, M. (2026). Clinical improvement despite rising anti-streptococcal antibody titers during adjunctive bacteriophage therapy in refractory PANDAS: a hypothesis-generating case report. Frontiers in Medicine, 13, Article 1973954. https://doi.org/10.3389/fmed.2026.1973954

Image Credits: AI Generated

DOI: 10.3389/fmed.2026.1973954

Keywords: PANDAS, bacteriophage therapy, streptococcus, autoimmune neuropsychiatric disorders, phage therapy, pediatric OCD, neuroimmunology, molecular mimicry, Clinical, improvement, despite, rising