Neurological demyelinating diseases and COVID-19 vaccines: an umbrella review of systematic reviews and meta-analyses supported by artificial intelligence insights

Demyelinating disorders are among the neurological complications following COVID-19 vaccination; however, the magnitude and consistency of this association are debated. An umbrella review of systematic reviews and meta-analyses on COVID-19 vaccination and demyelinating diseases was conducted following PRISMA. PubMed/MEDLINE, Google Scholar and Scopus were searched. Overlap of primary studies was quantified using a citation matrix and the corrected covered area (CCA). Ten artificial intelligence (AI) tools were prompted with the same questions, and their outputs appraised against a pre-specified rubric verifying every supplied reference. Thirty eligible reviews fulfilled the inclusion criteria, addressing new-onset disease and relapse. Overlap was slight overall (CCA 3.7%) but moderate-to-high for Guillain-Barre syndrome (10.5%) and acute disseminated encephalomyelitis (8.6%). The evidence suggests a possible association with both new-onset and relapsing disease; however, the absolute risk appears very low, approximately 1.4 Guillain-Barre cases per million doses in the only review reporting a denominator. All 28 AI outputs matched the manual synthesis in direction, but only 65% of 167 supplied references could be verified and 9% were fabricated. Previously reported relative risk estimates, considered together, suggest that demyelinating events occur more frequently following SARS-CoV-2 infection than after vaccination. Current evidence supports a low risk of demyelinating disease onset or exacerbation following COVID-19 vaccination, whereas the risk after COVID-19 infection is substantially higher; benefits clearly outweigh potential risks. Because the underlying evidence derives overwhelmingly from case reports and series without unexposed comparators, causality cannot be established. AI tools may support rapid synthesis but cite unreliably.