Non-PML neuroinvasive infections reported with multiple sclerosis disease-modifying therapies: a mechanism-based...

Multiple sclerosis (MS) disease-modifying therapies (DMTs) alter host immunity and may produce pathogen-specific susceptibility to neuroinvasive infection. Comparative evidence is limited because events are rare and postmarketing research focuses mainly on progressive multifocal leukoencephalopathy (PML).

We analyzed deduplicated US FDA Adverse Event Reporting System (FAERS) reports from 2004Q1 to 2025Q4 using a prespecified case-non-case design. Tier A comprised direct high-certainty non-PML neuroinvasive infection terms. Tier B comprised a possible CNS infectious syndrome plus pathogen evidence elsewhere in the report, without CNS confirmation; tier A+B was a lower-certainty expanded endpoint for sensitivity and pathogen composition analyses. We estimated overall reporting (E1), pathogen composition within tier A+B (E2), and same-pathogen CNS localization (E3). Five DMT mechanisms were compared with contemporaneous active DMTs using calendar-year Mantel-Haenszel reporting odds ratios and covariate-adjusted Jeffreys-prior penalized logistic models.

Among 492,514 reports, 675 were tier A and 82 additional reports were tier B. Tier A aRORs (95% CIs) were 0.82 (0.63-1.05) for anti-CD20 therapies, 1.77 (1.44-2.16) for S1P modulators, 1.32 (1.06-1.64) for natalizumab, 3.28 (2.24-4.66) for alemtuzumab, and 0.56 (0.21-1.19) for cladribine. Fingolimod accounted for 145 of 152 S1P cases and had an aROR of 2.15 (1.75-2.64). Among four mechanisms eligible for E2, pathogen composition differed (LRT 210.59, 9 df; bootstrap P = 9.99 × 10^-4): viral reports were enriched with anti-CD20 therapies (OR 4.24, 2.56-7.02) and natalizumab (2.03, 1.26-3.26), fungal reports with S1P modulators (29.09, 15.77-53.64), and bacterial reports with alemtuzumab (4.99, 2.27-11.01). Only natalizumab-HSV/VZV met the adjusted E3 gate and showed CNS enrichment (aROR 2.80, 1.80-4.38; Holm-adjusted P = 5.81 × 10^-6).

E1 contrasts were interpreted as separate reporting comparisons, not a formal equality test across mechanisms. S1P reporting was largely fingolimod-driven, and natalizumab-HSV/VZV was the only prespecified pair with adjusted same-pathogen CNS enrichment. Findings support targeted vigilance and confirmatory studies but do not estimate incidence or establish causality.