Stem cells for stroke: promise, but thin certainty

Ischaemic stroke leaves many survivors with lasting neurological impairment and dependence in daily activities, and options for restoring lost function are limited. Stem cell-based therapies, particularly mesenchymal stem cells, have generated enormous interest as a way to promote repair, and a large number of systematic reviews now cover the area. When reviews proliferate and overlap, however, it becomes hard to see what the evidence base actually supports. An umbrella review steps back to appraise the reviews themselves — their quality, their overlap and the certainty of what they conclude.

This Open Science Framework-registered umbrella review followed PRISMA 2020 and Joanna Briggs Institute guidance, searching PubMed/MEDLINE, Scopus and Web of Science through 1 November 2025. Eligible studies were systematic reviews or meta-analyses of randomised or non-randomised human clinical studies of mesenchymal stem cells or related stem cell-based interventions for ischaemic stroke. Two reviewers independently handled study selection, data extraction, AMSTAR 2 appraisal of methodological quality, and GRADE-based certainty assessment. Because of overlap among reviews and heterogeneity in cell products, administration routes, timing and outcomes, findings were synthesised narratively rather than pooled in a new meta-analysis.

Twenty-six reviews were included. Across them, the evidence suggested possible improvements in neurological impairment, disability, activities of daily living and motor recovery, but effect estimates varied by cell type, administration route, timing after stroke and follow-up duration. Intravenous administration was the most frequently studied route and appeared feasible with acceptable short-term safety reporting, whereas intra-arterial and intracerebral routes raised greater procedure-related concerns. Short-term adverse events were not consistently increased. Certainty was low to moderate for most efficacy outcomes, moderate for short-term adverse events, and very low to low for long-term or rare safety outcomes including tumorigenesis, ectopic tissue formation, embolic events and alloimmunogenicity.

This synthesis tempers the enthusiasm that individual positive reviews can generate. The efficacy signal is real enough to justify continued investigation, but it rests on small early-phase trials with heterogeneous interventions and short follow-up, and the reviews summarising them overlap heavily, so apparent replication may be illusory. The safety picture is the sharper gap: reassurance about short-term adverse events tells you little about rare or delayed risks. The authors call for adequately powered, standardised, multicentre trials with long-term safety surveillance before routine clinical use.

If you encounter clinics offering stem cell treatment for stroke recovery, this review is a useful reality check. The top-level evidence points to possible functional benefits, but the confidence attached to those estimates is low to moderate at best, and the confidence about long-term harms is lower still. That combination — plausible benefit, poorly characterised risk — is precisely why the authors say routine implementation is premature. It is a field to watch rather than one where the question has been answered.