Rituximab in IgM Anti-MAG Demyelinating Polyneuropathy: A Systematic Review and Bayesian Random-effects Meta-analysis of Observational Cohorts

Objective To quantitatively synthesise observational evidence on rituximab efficacy and safety in IgM anti-MAG demyelinating polyneuropathy, and to evaluate the comparative benefit of CD27 + memory B-cell-guided versus relapse-based retreatment strategies. Background IgM anti-myelin-associated glycoprotein (anti-MAG) demyelinating polyneuropathy is a rare disease for which rituximab is the most widely used therapy. The most recent Cochrane review (2016) pooled two randomised trials (n = 73) with low GRADE certainty. Observational cohorts published since 2007, including three series in 2024–2025, have not been quantitatively synthesised. Design/Methods A systematic review and Bayesian random-effects meta-analysis of observational rituximab studies in adults with IgM anti-MAG demyelinating polyneuropathy was conducted. MEDLINE, Embase, and Cochrane CENTRAL were searched from inception to 30 January 2026. The primary outcome was the proportion of patients meeting a composite responder definition (≥1-point improvement on ≥2 of INCAT-DS, ISS, MRC) at 12 months. Pooled proportions were estimated on the logit scale with a half-normal(0, 0.5) prior on between-study τ. Risk of bias was assessed by ROBINS-I. Results Eight cohorts comprising 279 rituximab-treated patients were included. The pooled 12-month composite responder rate was 54.1% (95% CrI, 34.4–73.2%; k = 2; n = 62). The secondary broader composite was 37.0% (22.0–54.5%; k = 4) and the ONLS-based pool was 33.7% (17.2–56.4%; k = 3). CD27+ memory B-cell-guided retreatment was associated with greater ISS improvement than relapse-based retreatment (mean difference, −1.69; 95% CI, −3.48 to +0.10) at less than half the cumulative dose. Pooled IgM flare was 8.2% (3.2–17.8%). Risk of bias was Serious in 5 of 8 studies. Conclusions Rituximab was associated with composite response at 12 months in approximately half of patients, with wide credible intervals and low certainty under GRADE. Adequately powered randomised trials of biomarker-guided retreatment and head-to-head comparison against Bruton tyrosine kinase inhibitors are warranted.