Culture-expanded MSC injections

This is a different product from the "stem cell" injections sold in US clinics. Culture-expanded MSCs are defined doses — 10 to 100-plus million cells — grown over weeks from bone marrow, adipose, or umbilical cord tissue in a GMP facility. The proposed mechanism is paracrine: anti-inflammatory and anti-catabolic signaling and synovial macrophage modulation, not engraftment. Human mechanistic support exists — dose-dependent reductions in MRI synovitis and cartilage catabolic biomarkers after autologous bone-marrow MSCs, and decreased poor-cartilage area on T2 mapping after allogeneic cells.

The positive trials

The randomized evidence is small but unusually consistent:

Lee 2019, a double-blind phase IIb, gave 24 patients a single 100-million autologous adipose MSC injection or saline. Its prespecified endpoint was a within-group change, and the cell arm met it — WOMAC down 55% at 6 months — while the saline arm moved on no symptom score. The trial reports no between-group test of any symptom score; its one between-group result is structural, cartilage defects holding stable on MRI while controls progressed.

Lamo-Espinosa's multicenter phase I/II randomized 30 patients to bone-marrow MSCs at 10 or 100 million cells plus HA, versus HA alone. The VAS carried the only between-group tests at 12 months, and both doses cleared control; the often-quoted 16.5-point WOMAC gain on the high dose is a within-group change, and the same results section gives it as 14. Joint space narrowed in controls at borderline significance and not on the high dose. At 4 years both cell groups still held their pain gains (median VAS 2–3 versus 7) — the field's longest durability signal — though WOMAC separated for the low dose only.

Vega 2015 found a single 40-million dose of allogeneic — off-the-shelf donor — bone-marrow MSCs beat HA on algofunctional indices with improved cartilage quality on T2 mapping at one year. Matas 2019 found repeated umbilical-cord MSC dosing beat HA on WOMAC pain at 12 months, with single dosing intermediate — though the trial measured each schedule against HA and never against the other. An open-label dose escalation points the same way: 50 million cells performed best, with dose-dependent synovitis and biomarker reductions.

The large trial, and the funding pattern

The one large randomized test of a culture-expanded product is the umbilical-cord-tissue MSC arm of Mautner 2023 (480 patients): not superior to a corticosteroid injection at 12 months. Every positive trial above has 10–15 patients per arm — effects of that size routinely shrink at scale. Sponsor entanglement is real (Lee funded by R-Bio; Matas authors employed by Cells for Cells; Freitag's positive trial had no sham injection and an author-affiliated clinic), and the HA comparators and co-injections in several trials muddy attribution.

Regulatory status

Culture-expanded MSCs are unambiguously a biologic drug in the US: not FDA-approved for any orthopedic indication, and injection outside an IND or clinical trial is unlawful. US clinics advertising "stem cell injections" are selling either non-expanded products (BMAC, SVF — see those entries) or unapproved drugs. Approved products exist abroad — Cartistem, an allogeneic umbilical-cord product, in Korea — so true culture-expanded MSCs are accessible only through registered trials or medical tourism, with the quality-control risk that implies.

Dosing signals

Higher doses look better (100M over 10M; 50M over 1–10M), and repeated dosing outperformed the HA control where a single dose did not. Every one of those signals is indirect: no trial here has put one cell dose against another, and at four years the Lamo-Espinosa doses did not differ, with WOMAC separating for the low dose rather than the high one. Autologous products require weeks of manufacturing lead time; allogeneic cells worked as well as autologous in the one small trial that tested them, with theoretical immunization concerns not observed clinically at these scales.

Safety

No procedure-related serious adverse events across the cited trials; transient injection-site pain and swelling are common, and in the 480-patient trial the umbilical-cord arm reported joint swelling in 24.1% of patients against 7.4% on a corticosteroid. The practical safety issue is unregulated clinics using poorly characterized products under a "stem cell" label.

What would change the tier

An adequately powered, placebo-controlled trial of a defined MSC dose — none has read out — plus answers on autologous versus allogeneic cells, source tissue, and whether the biomarker and imaging signals translate into actual joint preservation.