Stem Cell Evidence Tiers: Our Methodology

Stem cell therapy spans an unusually wide range of evidence quality, from bone marrow transplant, which has sixty-plus years of monitored outcomes and multiple full drug licenses, to unproven clinic offerings backed by nothing more than patient testimonials. To keep comparisons across this encyclopedia section consistent, every page carries a numeric evidence_tier field from 1 to 5, defined below, and we apply the same criteria regardless of how appealing or well-marketed a given therapy is.

Tier 1 requires strong human randomized controlled trial evidence, ideally replicated across more than one study, generally paired with an existing regulatory approval or a body of evidence a regulator has already reviewed in depth. Hematopoietic stem cell transplant qualifies here on the strength of decades of outcome data and multiple licensed products from the FDA's approved cellular and gene therapy products list [1]. The Pew Charitable Trusts safety dataset and the ISSCR patient handbook also anchor tier 1 status on our regulatory and safety pages, since they represent the most rigorously documented sources available on those specific questions [2][3].

Tier 2 covers therapies with real, statistically significant human randomized trial data that stops short of full regulatory approval or large-scale replication. Longeveron's laromestrocel for frailty falls here: a 148-patient Phase 2b trial found a statistically significant six-minute-walk improvement at the highest dose [4], which is genuine evidence but still a single dose-finding trial rather than a confirmatory Phase 3 program. Vertex's zimislecel for type 1 diabetes and the knee osteoarthritis MSC literature also sit in tier 2, the latter because Cochrane's own 2025 review explicitly rates its pooled findings as low-certainty evidence despite pooling 25 trials [5].

Tier 3 covers early-stage human trial data or case reports with a plausible mechanism but limited size, controls, or replication. The Kyoto Trial's iPSC-derived dopaminergic cell transplant for Parkinson's disease and the single-patient CiPSC-islet diabetes case report both illustrate this tier: real human data exists, the biology is plausible, but the evidence base is too thin to draw firm conclusions [6][7].

Tier 4 covers therapies that have been tested in adequately powered human trials that did not confirm the hoped-for benefit, or technologies with only animal-level mechanistic support and at most a Phase 1 safety trial underway in humans. Stem cell therapy for acute ischemic stroke sits here because the largest completed randomized trial, MASTERS-1, found no significant benefit on its primary endpoint [8]. Partial epigenetic reprogramming also sits in tier 4, since its human evidence consists of one ongoing Phase 1 safety trial and its mechanistic support comes from mouse studies rather than controlled human trials [9].

Tier 5 is reserved for hypotheses without any completed human testing, extrapolated purely from cell-culture or theoretical biology; no page in the current version of this encyclopedia section is rated tier 5, since every topic covered has at least preliminary human trial data or, in the case of regulatory and safety pages, a documented real-world enforcement or outcomes record.

We assign tiers at the level of the specific claim a page is built around, not at the level of the broader technology category. This is why mesenchymal stem cells appear at different implied tiers depending on the disease context: the underlying cell type is the same, but the frailty evidence is stronger than the stroke evidence, and our tier ratings reflect that difference rather than treating 'MSC therapy' as a single undifferentiated category. When new trial results are published, the relevant tier on that specific page will be revised; the last_verified date on each page reflects when its cited evidence was most recently checked against primary sources.