Hypogammaglobulinemia and Infection After Rituximab in Pemphigus, With Divergent Desmoglein Antibody and Immunoglobulin Kinetics: A Retrospective Cohort Study.

Abstract

Pemphigus is a rare autoimmune blistering disease for which rituximab is a first-line treatment, yet most safety data derive from lymphoma and rheumatoid arthritis cohorts. This study assessed hypogammaglobulinemia and infection during rituximab treatment, the baseline predictors of new-onset hypogammaglobulinemia, and whether desmoglein antibody titers and total immunoglobulin levels fall together. In this single-center retrospective cohort study, the records of 97 patients given rituximab for pemphigus between January 2015 and June 2026 were reviewed. Patients who received intravenous immunoglobulin were excluded, leaving 69 patients for outcome analyses. Among 51 patients with a normal baseline immunoglobulin G level, new-onset hypogammaglobulinemia, defined as a fall below 700 mg/dL after treatment, occurred in 5 (10%), and no patient reached the severe threshold of 400 mg/dL. Seven further patients were already below 700 mg/dL before treatment. Patients who developed new-onset hypogammaglobulinemia were older and had lower baseline immunoglobulin G and immunoglobulin M levels (areas under the curve 0.85, 0.83, and 0.81). Patients with hypogammaglobulinemia had received more corticosteroid per cycle, so a contribution of corticosteroids cannot be excluded. Late-onset neutropenia was not detected (0 of 69). Infection occurred in 5 of 69 patients (7%) and severe infection in 2 of 69 (3%). 6 months after induction, desmoglein 1 and desmoglein 3 antibody titers had fallen to 16% and 19% of baseline (p < 0.001), whereas immunoglobulin G remained at 97% (p = 0.46), independently of hypogammaglobulinemia. Hypogammaglobulinemia remained mild and the serologic response was not mirrored in total immunoglobulin levels. Advanced age and low baseline immunoglobulin levels are candidate markers of risk. These findings come from a retrospective single-center cohort and require confirmation in larger series.