To provide the most up-to-date and largest analysis of multiple sclerosis (MS) development and relapse risk following assisted-reproduction techniques (ARTs).

Existing studies report conflicting findings regarding relapse risk and MS onset following ART exposure. Earlier evidence was limited by small sample sizes, lack of comparator groups, and inadequate adjustment for confounders.

We conducted a PRISMA-adherent systematic review and meta-analysis based on PubMed, Scopus, and Web of Science-indexed publications until October 2025. No language or publication-period restrictions apply. Two reviewers screened studies; a third resolved disagreements. We pooled the best adjusted risk ratios (RRs) from multivariate analyses using metagen on R software (random-effects model). Studies reporting narrower confidence intervals (CIs) received a greater weight.  Quality assessment (QA) was done using the Newcastle-Ottawa scale (NOS). Relapse risk was judged post-partum at 3 months.

11 studies were included. Our analysis revealed no significant MS relapse risk following ARTs (RR: 1.23, 95% CI: 0.95-1.58, I2: 13.9%). Stratification based on antepartum disease-modifying therapy (DMT) status demonstrated similar conclusions (No DMT: 95% CI: 0.78-2.14, I2: 0%; DMT: 95% CI: 0.46-4.47, I2: 46.1%). Furthermore, separate analyses based on the success/failure of the ART procedure, alongside pooled analyses comparing the differential impact of GnRH agonists and antagonists, indicated no change in the risk of MS relapse. An analysis of MS development risk based on three longitudinal studies with a total of 897,013 patients from Danish and Iranian cohorts suggested no association with an increased MS risk (RR: 1.00, 95% CI: 0.82-1.23, I2: 57.7%). Regarding QA, ten of 11 studies scored good according to the AHRQ-converted NOS.

These findings provide reassuring evidence supporting the neurological safety of ARTs in women with MS seeking conception. Additionally, healthy women are not at an increased risk of MS post-ARTs. Yet, this study is limited by underrepresentation from Hispanic and Black communities.