We thank Bejerot et al. [1] for acknowledging the innovative nature of our work [2]. We share the goal of improving care for patients with severe psychiatric conditions. We agree with the hypothesis that an unrecognized subset of psychiatric patients may benefit from rituximab treatment.

Bejerot et al. refer to this case series as a “trial,” possibly confusing it with our ongoing clinical trial: Rituximab Treatment for Psychosis and/or Obsessive-Compulsive Disorder with Immune Involvement (Ra-P-OCD) (EudraCT No. 2019-000256-33). The present pre-trial case series focused on patients selected due to suspected underlying autoimmunity, based on both clinical features and objective markers of autoimmunity. Specifically, we report results from the off-label use of rituximab in three cases of treatment-resistant OCD. Additionally, exploratory work aimed to identify potential markers for selecting patients who might benefit from immunologically oriented interventions, such as rituximab. These included an objectively obtained staining pattern indicating anti-neuronal autoantibodies in CSF. We also performed repeated analyses of matched CSF and blood samples to identify candidate diagnostic and response biomarkers as a guide for future studies. These exploratory findings were not available at baseline and did not direct treatment. As described in the methods, these cases are part of a larger cohort of 127 patients [3].

Unfortunatley, the author’s account of our conclusions was not entirely correct: we did not claim that “clinical and biological markers can predict treatment outcomes” [1]. Our statement, “The findings suggest that clinical red flags and biological measures may predict rituximab response in chronic treatment-resistant OCD. The report provides orientation that may inform the hypotheses and design of future treatment trials” is more cautious than Bejerot et al. imply, stressing that the field is still in a discovery phase [2].

The ongoing clinical trial, Ra-P-OCD, is a phase 2, randomized, double-blinded, placebo-controlled trial aims to assess rituximab’s effectiveness in patients with psychosis and/or OCD who have indications of immune involvement. Contrary to Bejerot et al.’s suggestion, inclusion in the Ra-P-OCD trial requires one clinical red flag and one objective marker from a comprehensive list. Patients receive two doses of rituximab at different time points in a double-blinded design, allowing for long-term follow-up over 16 months. A key feature of the trial is repeated blood and CSF sampling to asses candidate biomarkers that may both facilitate selection of patients, and serve as markers for treatment outcome in future trials.

Determining trial inclusion criteria is challenging, especially in a developing field with limited data. Narrowly defined criteria risk excluding potential responders, while broad inclusion could lead to negative trial results and premature abandonment of an effective treatment for a subset of patients. We acknowledge that novel biomarkers must undergo sensitivity and specificity analyses to establish their clinical utility, whereas biomarkers already validated for other CNS or systemic autoimmune conditions should have a lower threshold for clinical implementation in psychiatry. Ongoing research aims to link these markers to clinical features characteristic of established CNS and systemic autoimmune disorders—collectively termed ‘red flags’—to improve patient stratification.

Bejerot and collaborators have previously demonstrated that patients with treatment resistant OCD seem to not respond to Rituximab at a group level [4]. Their previous study, Rits-PO, demonstrated low response rates to a single rituximab dose in unselected OCD patients (2/10 responders), which they described as “less impressive, and in their discussion they referenced the concept of an “immunological/autoimmune subtype” as a possible explanation for the low response rate [4]. This low response rate was also given as a reason for not continuing studies with rituximab in patients with OCD. Furthermore, Bejerot et al. in this correspondence describe several cases that involve either established CNS autoimmune disease or clear red flags for autoimmune OCD, as well as cases that required multiple rituximab doses and extended follow-up to achieve improvement. Given this experience, we are surprised they argue for broad inclusion of patients with OCD in rituximab trials.

Lastly, we emphasize that a single positive RCT is insufficient to change psychiatric treatment guidelines. Many RCTs and meta-analyses already suggest that immunotherapy and antimicrobials can benefit certain psychiatric patients. However, these studies often rely on broad psychiatric diagnoses without objective immunological markers, which we argue may the the reason for inconsistent findings and moderate effect sizes. This lack of specificity is problematic, particularly for treatments such as rituximab, which carry risks.

Our long-term goal is to improve differential diagnostics, validate existing, and identify new biomarkers to refine subgroup selection for rituximab treatment while minimizing risks. We are keen to see how our respective complementary contributions align and look forward to their impact on advancing immunopsychiatry and improving patient outcomes.