Well B2 Joints, spine and sports
Stem Cell Therapy for Knee Pain: Procedure, Timeline, and Expectations
Contents: 7 sections
In trials for knee arthritis, stem cell injections lowered pain slightly more than placebo injections, and a Cochrane review rated that evidence low certainty. The main U.S. arthritis guideline recommends against them. If you are still considering one, this page walks through what the procedure involves, what the weeks and months afterward look like, and what the trials suggest you can realistically expect.
The procedure described here is the one a clinic can offer: a bone marrow draw from the back of the pelvis and a knee injection the same day. Whether treatment is even reasonable to consider depends on the diagnosis, its severity and what you have already tried, the questions worked through in who might consider stem cell treatment and who should not.
What happens on the procedure day?
Protocols differ between clinics, but a same-day bone marrow procedure has four parts.
1. Examination, images and medicines. The clinician should examine your knee and look at your X-ray, and MRI if you have one, before anything is drawn. If the plan was set before anyone examined you, stop there. Tell the clinic about every medicine you take. Prescribed aspirin and blood thinners should not be stopped without the doctor who prescribes them; Mayo Clinic warns that stopping daily aspirin suddenly “may trigger a blood clot and lead to a heart attack” (Mayo Clinic). Our printable question list covers what to settle before this point.
2. The bone marrow draw. Marrow is drawn with a needle from the back of the hip bone (the pelvis), with a local anesthetic to numb the area. For the diagnostic version of this procedure, Mayo Clinic describes “a brief sharp pain or stinging” as the marrow is withdrawn and tenderness that can last “a week or more” (Mayo Clinic). Problems are uncommon. In UK surveys of diagnostic marrow biopsies, there were 15 adverse events in 20,323 procedures in 2004, most of them bleeding (Bain, Haematologica 2006). A clinic’s harvest may draw more marrow than a diagnostic test, so it is worth knowing how much it takes and from how many sites.
Some clinics use fat instead of marrow. How it is processed matters: in 2024 a federal appeals court agreed with the FDA that one clinic’s product, made by breaking fat down with enzymes to release its cells, was a drug that needed FDA approval (Cooley summary of the ruling).
3. Processing. In a same-day procedure, the marrow is spun down to a smaller volume while you wait. The clinic should be able to give you a written report of what was in the syringe.
4. The injection. The concentrate goes into the knee joint, and the needle may be guided with ultrasound or X-ray. Guidance helps put the needle in the right place, but how imaging guides the needle and what it cannot promise explains why accurate placement is not the same as a good result. If the plan adds injections around the joint, into tendons or ligaments, each one should come with a reason and a price.
The weeks and months afterward
The first days. Expect tenderness where the marrow was drawn; the clinic should tell you what knee soreness it expects and for how long. Some signs always need a prompt call: sudden severe pain, swelling, warmth, skin discoloration or fever after a joint injection can mean a joint infection (Mayo Clinic). Where the marrow was drawn, call for bleeding through the bandage, persistent fever, or pain, swelling, redness or drainage that is getting worse (Mayo Clinic).
The first weeks. Clinics differ on whether and for how long to avoid over-the-counter anti-inflammatory painkillers, and on how soon to walk, drive or exercise. A written protocol with the reason behind each limit is what to expect. Recovery in detail is set out in the weeks after an injection and when to call.
The following months. The Cochrane trials measured their main results at three to six months, so that is the window the evidence speaks to. Before the injection, settle how you will judge it: a standard knee questionnaire such as the KOOS, how far you can walk, how you sleep. Write down where you start.
If nothing has improved. The useful next step is a fresh look at the diagnosis and a talk about the next options, including physical therapy, other medical treatment or a surgical opinion. A repeat or more expensive injection offered as the automatic next step deserves a clear explanation of the evidence behind it.
What can you realistically expect?
Set your expectations by the trials, not by testimonials.
- On average, a small change. A drop of about one point on a 10-point pain scale, measured against placebo, is the best current estimate, and it is uncertain.
- No proof of regrowth. No study in the Cochrane review measured arthritis progression on X-ray, so the best review of the trials offers no evidence that the injection rebuilds the joint or slows arthritis.
- Advanced arthritis is the weakest case. When X-rays show the joint space largely gone, replacement is an established option and no trial shows an injection restores the joint. The X-ray grades are explained on the candidate page.
- Early arthritis is not an emergency. Because no study in the review measured progression, there is no trial evidence that injecting a knee with mild X-ray changes, or no symptoms, prevents later damage.
An average on a pain scale may also not be the question you are asking. If what you want is to walk the length of a grocery store without stopping, measure that before any treatment and again at three and six months. Your own before-and-after cannot tell you why things changed, but it tells you whether the change was one you notice, which a one-point average cannot.
What does the evidence show for knee arthritis?
The best summary is a 2025 Cochrane review. Compared with placebo injections, stem cell injections lowered pain by about 1.2 points on a 0 to 10 scale and improved function by about 14 points on a 0 to 100 scale, measured at three to six months; the reviewers’ plain-language summary says they “may slightly improve” both. They rated that evidence low certainty because the source, preparation and dose of the cells varied between trials, and they suspected publication bias because up to three larger trials were withdrawn before reporting results. No included study measured whether the joint changed on X-ray (<a href="https://www.cochrane.org/evidence/CD013342what-are-benefits-and-risks-stem-cell-injections-knee-osteoarthritis”>Cochrane review CD013342).
The trials also mixed products, which is one reason the reviewers lowered their confidence. Cells grown in a laboratory for weeks cannot legally be offered for knee arthritis outside an FDA-authorized trial, so results from those trials do not describe the same-day bone marrow concentrate a clinic can offer. The same-day rule is part of what stem cell therapy is and what the FDA allows.
The American College of Rheumatology and Arthritis Foundation guideline recommends against stem cell and platelet-rich plasma injections for knee and hip arthritis (ACR). The American Academy of Orthopaedic Surgeons’ patient site says of stem cells and PRP that “clinical studies have yet to confirm their value in treating osteoarthritis” (OrthoInfo).
How should you read these numbers?
An average is not a prediction for you. In a study of 2,724 people in pain trials, “a reduction of approximately two points or a reduction of approximately 30%” on a 0 to 10 scale matched what people themselves called a clinically important change (Farrar, Pain 2001). That figure describes how much one person’s pain has to fall to feel it. The Cochrane figure is a difference between two group averages, so the two numbers cannot be compared directly. What the average hides is the spread: some people may improve a lot and others not at all, and small trials cannot say reliably which is which.
Placebo injections do a lot on their own. In the knee trials, about 53 in 100 people given a placebo injection reported treatment success, compared with about 68 in 100 given stem cells, on evidence the reviewers rated very low certainty (Cochrane, above). That is not unusual for injections. In a pooled analysis of placebo groups from injection trials in osteoarthritis, 45.5% of people given a placebo reported at least a 20% drop in pain at short-term follow-up (Yu, Arthritis Care & Research 2024), and an earlier review of 198 osteoarthritis trials found placebo relief was larger when the placebo was given by injection or needle (Zhang, Annals of the Rheumatic Diseases 2008). Feeling better after an injection is real, but it does not tell you the cells caused it.
Pain and function are different goals. A lower pain score and being able to walk farther can move together or separately. Decide in advance which one matters to you, and judge the result by that.
What would change the picture. A large placebo-controlled trial of the same-day marrow product, with results reported beyond a year, would settle more than the current small trials can. Until one reports, the honest summary is the one above: a small average benefit, low certainty, and no evidence yet about the joint itself.
The knee treatments side by side
| Option | What the evidence shows | How long it lasts | ACR/AF guideline for knee arthritis |
|---|---|---|---|
| Exercise and, if you carry extra weight, weight loss | Exercise improves pain and function; losing weight reduces stress on the knee (OrthoInfo) | As long as you keep it up | Strongly recommended (exercise) |
| Cortisone (steroid) injection | Relieves pain and inflammation, but "the effects do not last indefinitely" (OrthoInfo). See the trial below | Temporary | Strongly recommended |
| Hyaluronic acid (gel) injection | The guideline panel advises against it for the knee | Not established | Conditionally recommended against |
| Platelet-rich plasma (PRP) | In a JAMA trial of 288 people, no better than saline for pain at 12 months and no difference in cartilage volume (RESTORE trial) | Not established | Recommended against |
| Stem cell injection, including bone marrow concentrate | Small benefit over placebo at three to six months, low certainty; no study measured joint structure (Cochrane, above) | Not established | Recommended against |
| Total knee replacement | "More than 90% of modern total knee replacements are still functioning well 15 years after the surgery" (OrthoInfo) | Long term for most people | Not covered by this guideline |
Sources for the table: OrthoInfo; RESTORE trial; OrthoInfo.
The guideline column follows the ACR/AF 2019 guideline, as summarized in the press release linked above. For hip arthritis the same guideline also recommends against stem cell and PRP injections and supports steroid injections, and its recommendation against hyaluronic acid is strong rather than conditional.
Read across the table and the pattern is that none of the injections in it has shown a lasting improvement in the knee itself; they differ mainly in how much trial evidence supports short-term relief and in what the guideline advises. Of the options the guideline strongly recommends, exercise is the one you can keep going on your own.
A note on cortisone and cartilage. In a two-year JAMA trial of 140 people, knees injected with triamcinolone every 12 weeks lost slightly more cartilage on MRI than knees injected with saline (0.21 mm against 0.10 mm), and pain was no better than with saline (<a href="https://www.wikijournalclub.org/wiki/EffectofIntra-articularTriamcinolonevsSalineonKneeCartilageVolumeandPaininPatientsWithKnee_Osteoarthritis”>trial summary). That is a reason to avoid injections on a fixed schedule. It is not evidence that an occasional cortisone shot for a flare is worse than a biologic injection. OrthoInfo notes that doctors may limit cortisone to three or four injections a year per joint because of possible side effects.
Every one of these options has a price, and the injections may not be covered by insurance. A complete written quote covers the harvest, processing, each injection, repeats and follow-up, as set out in the clinic evaluation framework; urgency and add-ons that arrive with a quote are among the signs that a clinic is overselling results.
Should you see a surgeon, a regenerative clinic or both?
Seeing both is reasonable, as long as you know that each visit answers a different question.
| Surgeon's consultation | Regenerative clinic consultation | |
|---|---|---|
| The main question | Is a joint replacement or other operation right for this knee now, and if not, what should I do meanwhile? | What will be injected, what does the evidence show for my knee, is it legal, and what does it cost? |
| What should decide it | Pain, disability and imaging, not age alone. Most people who have a knee replaced are 50 to 80, but "orthopaedic surgeons evaluate patients individually" (OrthoInfo) | Your X-ray grade, the Cochrane and guideline evidence above, and a written plan and price |
| The trade-off to weigh | Recovery, infection and blood clot risk, and activity limits. "Most surgeons advise against high-impact activities such as running, jogging, jumping, or other high-impact sports for the rest of your life after surgery" (OrthoInfo) | Low-certainty benefit, harvest and infection risk, the full price and whether insurance pays any of it, and what happens if it does not work |
Sources for the table: OrthoInfo.
A replacement is not a failure to try everything else first; there is no rule that an injection has to come before surgery. And a surgeon who says your knee is “not bad enough yet” is not telling you to buy a biologic in the meantime. Exercise, anti-inflammatory medicines and an occasional cortisone injection are what the guideline supports, and OrthoInfo adds weight loss for people who carry extra weight. The trade-offs between the two are laid out in weighing injections against a joint replacement.
The choice that remains is whether a small, uncertain benefit over three to six months is worth the harvest, the injection and the price to you, knowing that the guideline-backed options stay available either way.
This page is general information, not medical advice.