Stem cells for knee osteoarthritis are being studied in randomized trials, but no stem cell product is FDA-approved for the condition. The studies so far are small, vary widely, and disagree on how much of any improvement comes from the cells themselves.
The broader picture of approved versus experimental cell products is laid out in what is stem cell therapy, and this page narrows that down to one commonly advertised use.
Regulatory status for knee conditions
The FDA consumer alert on regenerative medicine products said in 2020 that the only FDA-approved stem cell products were blood-forming stem cells derived from umbilical cord blood, and that these were approved for disorders affecting the production of blood, not for other uses. On December 18, 2024, the FDA approved Ryoncil, a donor bone marrow-derived mesenchymal stromal cell therapy, for steroid-refractory acute graft-versus-host disease in children; the FDA’s list of approved cellular and gene therapy products is the current reference. Neither approval covers knee osteoarthritis.
The alert also says that none of these products, a group that includes products derived from adipose tissue, umbilical cord blood, Wharton’s jelly or amniotic fluid, have been approved for the treatment of any orthopedic condition, naming osteoarthritis and knee pain among the examples. In other words, an injection offered for a painful knee is, in the United States, an investigational or unapproved product unless it is given within a properly regulated study.
What researchers are testing
Most trial work on this topic uses mesenchymal stem cells, usually taken from bone marrow, fat tissue or umbilical cord tissue and injected into the joint. A US trial listed on ClinicalTrials.gov, known as MILES, was designed to compare a corticosteroid injection with three cell preparations: bone marrow concentrate from the patient, fat-derived cells in the form of stromal vascular fraction, and third-party mesenchymal stem cells manufactured from umbilical cord tissue. The registry entry describes it as a multicenter study at 4 sites in the United States with a total of 480 participants.
The design shows the main open questions: which cell source, which dose and which comparison treatment give reliable information. Trials like this are how those questions get answered, and they are very different from a clinic selling an injection outside a study.
- Approved only for blood-production disorders: Approved only for blood-production disorders
- No approval for any orthopedic condition: No approval for any orthopedic condition
- Product group named: Product group named
- MedWatch: MedWatch
- MILES compares corticosteroid with three MSC preparations: MILES compares corticosteroid with three MSC preparations
What pooled analyses report
Meta-analyses combine several randomized trials to see what they show together. A 2025 meta-analysis in Stem Cell Research and Therapy pooled 8 randomized trials with 502 patients with unoperated knee osteoarthritis. Its authors reported that mesenchymal stem cell injections were associated with better scores on a standard osteoarthritis questionnaire at 6 months than in the control groups, with larger effects in subgroups using fat-derived cells and higher doses. They also reported no significant difference in adverse events between the cell and control groups.
Those results are the source of much of the optimism seen in marketing. They need to be read with their limits in mind: the review pools small trials that differ in cell source, dose, controls and measures, and its own subgroup findings come from comparing different trials rather than testing options head to head.
The placebo question
A second 2025 analysis looked at the same question from a different angle. A systematic review in Frontiers in Medicine included only trials that compared mesenchymal stem cell injections with an inert placebo, 8 trials with 467 patients in all. It estimated that contextual, placebo-related factors accounted for approximately 63% of pain reduction and 61% of functional improvement at 6 months.
At 12 months the same analysis estimated that contextual factors explained about 50% of pain relief and about 66% of functional gains. The authors rated the certainty of the evidence as low, and concluded that the cells themselves conferred only a modest incremental benefit. Joint pain naturally fluctuates, and injections of any kind can feel helpful, which is exactly why placebo-controlled designs matter and why personal accounts cannot settle the question.
Why anecdotes and autologous claims fall short
The patient guide from the International Society for Stem Cell Research explains why individual stories carry little weight. A person may feel better after a treatment because of an intense belief that it will work, accompanying treatments, lifestyle changes or natural fluctuations in the disease, and the guide says these factors are difficult to measure objectively outside carefully designed clinical trials.
The same guide addresses a common marketing line for joint injections, that a patient’s own cells are inherently safe. It states that cells from your own body are not automatically safe when used in treatments, because removing, processing and reintroducing cells can bring risks such as contamination with pathogens, interference with normal cell function, and placement in parts of the body where the cells do not belong. This does not describe every trial product, but it explains why even autologous preparations need regulatory oversight and monitoring.
The guide also notes that an experimental treatment offered for sale is not the same as a clinical trial. The fact that a procedure is experimental does not automatically mean it is part of a research study, which is why a registry listing and independent ethics oversight are worth confirming before anything else.
It also helps to separate the question of whether a procedure is safe enough to study from the question of whether it helps. Trials are designed to answer the first with monitoring and the second with comparison groups, and a procedure can pass the first test without having passed the second. Readers who see a claim of success should ask which of those two questions the evidence actually addresses, and how many participants and how much follow-up stand behind it.
Reading the evidence together
The two reviews are not simple opposites. One reports improvements against control groups in a pooled set of trials; the other asks how much of the improvement in placebo-controlled trials is explained by the experience of being treated. Both describe a field where the trials are small and heterogeneous, and where large, well-controlled studies are still needed to establish whether a given product helps, for whom, and for how long.
That is also why the questions a reader brings matter. The checklist in questions to ask before stem cell treatment covers approval status, trial registration and risk, and the same logic applies to any joint injection marketed as regenerative. A similar pattern of limited approval and uneven evidence appears in other organ systems, as in the discussion of COPD stem cell therapy.
Decisions about knee osteoarthritis care, including whether any injection is appropriate, belong with a licensed physician who knows the reader’s joint, history and goals. Anyone hurt or affected by a side effect after a regenerative product can report it to the FDA’s MedWatch program, as the consumer alert describes.
Frequently asked questions
Is any stem cell injection approved for knee osteoarthritis?
Show answer
No. According to the FDA consumer alert, none of the widely promoted stem cell products have been approved for any orthopedic condition, including osteoarthritis.
Why do studies seem to disagree?
Show answer
They ask different questions. One pooled trials against control groups of various kinds, while another included only placebo-controlled trials and estimated how much improvement was contextual.
What is the difference between a trial and a clinic offering an injection?
Show answer
A registered trial has a defined protocol, a listing in a registry and a comparison group. A clinic sale outside a study has none of those protections, even if the product is described as experimental.
Does a larger trial exist?
Show answer
Several multicenter trials have been registered, including MILES, which compares three cell preparations with a corticosteroid injection. Registry entries describe plans, and results should be read from published papers.
