Skip to the lineages
Regenerative Medicine

What Do Clinical Trials Show About Stem Cell Injections for Knee Osteoarthritis?

Published November 16, 2026

Protocol: 7 steps in this guide
  1. Which preparations are being tested
  2. Placebo and active comparators
  3. Pain and function versus cartilage
  4. Adverse events and follow-up
  5. The VA/DoD 2026 position
  6. How to read small and sponsored trials
  7. What remains unapproved

Randomized trials of cell injections for knee osteoarthritis show modest improvements in some pain and function scores, mixed results on others, and no consistent improvement in joint structure on MRI. They also show more local reactions, such as injection-site pain and swelling, than control injections. A 2026 meta-analysis of 28 randomized trials concluded that these therapies serve “a primarily symptom-modifying rather than structure-modifying role.” The U.S. Department of Veterans Affairs and Department of Defense guideline, based on evidence through July 2025, found insufficient evidence to recommend for or against stem cell injections for knee osteoarthritis. No cell product is FDA-approved for it.

Which preparations are being tested

“Stem cell injection” covers very different products, and results for one do not transfer to another:

The 2026 meta-analysis, by Awad and colleagues in Clinical Rheumatology, noted that trials “report inconsistent results due to heterogeneity in cell sources, preparations, and techniques,” and found that benefits “were more consistent with culture-expanded preparations, bone marrow sources, saline controls, and ultrasound guidance.”

Placebo and active comparators

The comparator decides what a result means. The table summarizes the main recent evidence by comparison.

Source Products Comparator Symptom result Structure result Notes
Awad and colleagues, meta-analysis (2026); 28 RCTs MSC-based therapies Various controls Improved VAS pain and some KOOS domains; no significant difference in WOMAC, KOOS quality of life or Lequesne MRI WORMS not significant More injection-site pain and swelling
VA/DoD review of Whittle and colleagues (2025) Stem cells Placebo, 7 RCTs Better function up to 6 months; no difference at later time points Not the focus Very serious limitations noted
Garza and colleagues, American Journal of Sports Medicine (2026); 114 randomized Autologous SVF, 25 million cells, one injection Placebo 12-month responder rate 51.8% vs 29% (one-sided P = .0149) MRI obtained; abstract reports no structural result Device study sponsored by GID BIO, INC

The VA/DoD guideline also reports, from the same review, that “autologous stem cells improve function better than placebo, but there is no difference between allogenic stem cells versus placebo,” and that evidence from four trials in a 2024 review of BMAC “was unable to find differences between BMAC and other treatments for improving function.”

Pain and function versus cartilage

In the 2026 meta-analysis, the significant gains were in symptom scores. It reported that “MRI-based WORMS scores were non-significant, indicating no consistent structural benefit.” Whether any cell product rebuilds cartilage is the subject of whether stem cell products can regrow damaged joint cartilage in humans. Score thresholds and what counts as a meaningful change are covered in what clinically meaningful improvement means in a joint pain study.

Adverse events and follow-up

The 2026 meta-analysis found higher rates of injection-site pain (risk ratio 2.04), joint swelling (risk ratio 3.39) and other adverse events (risk ratio 1.26) with stem cell therapies, while serious complications such as infection were, in its words, “uncommon and non-significant.” The Garza SVF trial reported no serious investigational device-related adverse events at 12 months. SVF also requires liposuction, which adds its own procedure. The abstracts used here do not give every trial’s follow-up, and the Garza trial reported results at 12 months, which limits what this article can say about long-term effects; how longer-term safety is tracked is covered in how researchers monitor long-term safety after cell therapy.

The VA/DoD 2026 position

The VA/DoD Clinical Practice Guideline for the Non-Surgical Management of Hip & Knee Osteoarthritis (May 2026) states: “For osteoarthritis of the knee, there is insufficient evidence to recommend for or against” several injections, including stem cell injections, for which it gives mesenchymal, adipose-derived and bone marrow-derived cells as examples. Its discussion notes that these preparations “are believed to exert their proposed effects primarily through paracrine signaling, immunomodulation, and modulation of the local inflammatory environment rather than through direct tissue engraftment or durable structural regeneration.” The guideline states that it is based on evidence reviewed through July 31, 2025, so it predates both the 2026 meta-analysis and the Garza trial. “Insufficient evidence” means uncertainty, not proof that the injections do nothing.

How to read small and sponsored trials

Several features of this evidence call for caution:

  • Small samples. Small trials can produce large effects by chance, so single small positive results need confirmation.
  • Comparator choice. A saline placebo, a numbing injection, hyaluronic acid and no treatment answer different questions; the meta-analysis found more consistent benefits against saline.
  • Uneven groups. The Garza trial randomized 83 people to SVF and 31 to placebo, so the placebo group was small.
  • Sponsorship. CMS’s record for IDE study BB16734-NCT04440189 lists the trial as a pivotal study of the “GID SVF-2 Device” sponsored by GID BIO, INC. Sponsorship does not make a result wrong, but it is a reason to read the full methods.
  • Responder endpoints. A composite “responder” rate depends on how the threshold was defined in advance; it should be read alongside the actual score changes.

Why placebo control matters is explained in the difference between a placebo-controlled trial and an observational study.

What remains unapproved

FDA’s patient and consumer information on regenerative medicine therapies states that the unapproved regenerative products it warns about “have not been approved for the treatment of any orthopedic condition, such as osteoarthritis.” The one FDA-approved stromal cell product, Ryoncil, is approved for a pediatric transplant complication, not for joints. An SVF trial under an investigational device exemption is research, not approval. How the rules apply is covered in how the FDA regulates stem cell products promoted for joint pain, and the comparable PRP evidence in what the evidence says about PRP for knee osteoarthritis.

Randomized trials suggest that some cell injections modestly improve knee osteoarthritis symptoms for a period, with more local reactions and no consistent structural benefit, and with products varying widely and none FDA-approved for the knee, the evidence supports cautious, product-specific research rather than routine use.

Cells linked to this one

This guide leads on to

Guides that lead here

Lineage F: Stem Cell Clinical Evidence and Regulatory Boundaries

What readers added

Add a remark

Your email address will not be published. Required fields are marked *