What Is the Success Rate of Stem Cell Therapy for Knees?

Stem cell therapy for knee osteoarthritis produces measurable improvements in pain and function for many patients, but pinning down a single “success rate” is harder than the marketing suggests. Individual studies report improvement rates ranging from about half of treated patients to over 90%, depending on how success is defined, what cell source is used, and how advanced the arthritis is. One four-year follow-up of 37 knees treated with bone marrow aspirate concentrate reported that 35 improved on standard scoring systems, and none required a knee replacement during the study period. Yet a rigorous phase 2b randomized trial comparing adipose-derived stem cells to a placebo injection found no significant difference in the proportion of patients who responded to treatment. The gap between those two findings tells you a lot about where this therapy actually stands.

What the Clinical Evidence Shows

Most studies measuring the effect of stem cell injections on knee osteoarthritis use patient-reported scores for pain, stiffness, and physical function, along with MRI imaging to assess cartilage changes. A 2020 meta-analysis pooling data from controlled trials found that patients who received stem cells showed improved pain scores and improved knee function scores at 24 months compared to control groups. However, at shorter time points of 6 and 12 months, the differences in several key measures did not reach significance.¹ That pattern matters: the therapy seems to take time to produce its best results, and early follow-ups can look underwhelming.

When researchers have followed patients further out, the numbers look more encouraging. A study tracking 37 knees with advanced osteoarthritis (grades III and IV on the standard radiographic scale) over four years found significant improvements in both function and symptom scores beginning around year two. Function scores rose from an average of 56 to 73, and symptom scores dropped from 40 to 18 on a scale where lower is better. Thirty-five of the 37 knees improved, and no patient in the study went on to need a joint replacement.² A separate five-year follow-up of adipose-derived stem cell injections found that pain reduction and functional improvement persisted across the full five years.³

On the structural side, there is some evidence that stem cell injections slow cartilage loss. One study using fully automated 3D MRI analysis showed that cartilage was actively shrinking in the 30 weeks before injections, but that loss stopped in the 30 weeks afterward.⁴ Another prospective study found that MRI grades for cartilage lesion size and full-thickness cartilage loss improved significantly after mesenchymal stem cell implantation.⁵ These imaging findings are encouraging, though they come from relatively small studies and don’t yet prove that cartilage regrowth translates directly to long-term joint preservation.

The Placebo Problem

Here is where the story gets uncomfortable for stem cell enthusiasts. A 2025 systematic review specifically looked at how much of the improvement seen in stem cell trials could be chalked up to “contextual factors,” which is the clinical term for everything that isn’t the cells themselves: the injection procedure, patient expectations, clinical attention, and the natural fluctuation of symptoms. At six months, those contextual factors accounted for roughly 63% of pain reduction and 61% of functional improvement. At 12 months, they still explained about half of pain relief and about two-thirds of functional gains.⁶

The ADIPOA2 trial, a phase 2b randomized clinical trial comparing adipose-derived stem cells to a placebo injection, underlined this finding. After six months, about 47% of patients in the stem cell group met the strict response criteria, compared to about 55% in the placebo group. There was no statistically significant difference between the two.⁷ That doesn’t mean stem cells never work, but it does mean that a substantial share of the improvement patients experience after an injection would have happened with a sham injection too. When clinics quote success rates of 80% or higher, they are usually counting everyone who feels better, with no way to separate the biological effect of the cells from the powerful placebo response that comes with any injection procedure.

How They Compare to Hyaluronic Acid Injections

Hyaluronic acid (HA) injections are one of the most common non-surgical treatments for knee osteoarthritis, so it’s natural to ask whether stem cells do any better. A 2024 systematic review and meta-analysis of head-to-head trials found that at 12 months, stem cell groups showed meaningfully greater improvements in pain, function, and MRI-assessed joint scores compared to HA groups, and those differences reached what clinicians consider clinically important thresholds.⁸

Smaller trials have filled in some of the detail. A matched-pair cohort analysis found significantly greater improvements in pain, function, and activity scores at one year in the stem cell group compared to HA.⁹ A randomized trial comparing bone marrow concentrate plus platelet-rich plasma to crosslinked hyaluronate found that the stem cell group’s scores peaked at 12 months while the HA group’s peaked at 6 months and began declining. Pain reduction at 12 months was roughly double in the stem cell group.¹⁰ A multicenter trial testing two doses of bone marrow mesenchymal stem cells against HA found that both stem cell dose groups improved significantly more than the HA group over long-term follow-up.¹¹

The comparison to HA is probably the strongest point in stem cell therapy’s favor. HA is a well-studied active treatment, not a placebo, and the consistent edge stem cells show over it at 12 months across multiple trials suggests something real is happening beyond contextual effects. But “better than hyaluronic acid” is a more modest claim than “regenerates cartilage and cures osteoarthritis.”

Does the Type of Stem Cell Matter?

Three main cell sources dominate the knee osteoarthritis literature: bone marrow aspirate concentrate (BMAC), adipose (fat)-derived stem cells, and umbilical cord-derived stem cells. Each has advocates, and each has trade-offs.

A meta-analysis comparing bone marrow-derived and adipose-derived cells found that both had meaningful potential for treating knee osteoarthritis, but the adipose-derived cells showed a slight edge on pain and function scores based on the limited trials available.¹² A prospective trial directly comparing the two in the same clinical setting found that both produced significant improvements in pain and function, with no meaningful difference between them.¹³

Umbilical cord-derived cells are newer to the field. A retrospective study comparing umbilical cord cells to bone marrow concentrate for cartilage repair found comparable outcomes at 12 months despite the umbilical cord group being older with larger cartilage defects.¹⁴ The practical difference for most patients comes down less to which cell performs better in isolation and more to what their clinic offers. BMAC requires a bone marrow draw (usually from the pelvis), adipose-derived cells require a mini-liposuction, and umbilical cord cells come from a donor, avoiding a harvest procedure entirely.

Who Responds Best

One of the clearest findings across the stem cell literature is that earlier-stage disease responds better. In the prospective BMAC-versus-adipose trial, patients with moderate arthritis (grade 2 on the standard radiographic scale) showed significantly better functional and clinical outcomes than patients with more advanced disease (grades 3 and 4).¹³ A study specifically examining the role of osteoarthritis severity, body mass index, and age found that while BMI and age did not independently predict outcomes, disease severity did, especially among older patients. The authors concluded that younger patients with milder changes were the best candidates for lasting benefit.¹⁵

That said, the four-year study of grade III and IV knees cited earlier shows that even patients with advanced disease can improve, so “earlier is better” does not mean “later is pointless.” It is worth noting that a systematic review of patient factors affecting adipose stem cell quality found that increasing age, higher BMI, and diabetes all tended to reduce the proliferation and differentiation potential of the harvested cells, although this was not a universal finding across every study.¹⁶ In practical terms, a 50-year-old with moderate knee arthritis and a healthy weight is likely to get more out of the procedure than a 70-year-old with bone-on-bone changes and diabetes, though neither should be told a result is guaranteed or impossible.

How Stem Cells Actually Work in the Joint

The original hope was straightforward: inject stem cells, they turn into cartilage cells, and the damaged surface regrows. Reality is more nuanced. Researchers now recognize that the cells exert most of their benefit through what they secrete rather than what they become. Mesenchymal stem cells release a variety of signaling molecules that suppress inflammation, modulate the immune response in the joint, and stimulate the body’s own repair processes.¹⁷ The cells also release tiny packages called exosomes that carry proteins and genetic material to neighboring cells. In preclinical studies, these exosomes alone have been shown to promote cartilage repair, reduce inflammation, and improve the balance between cartilage building and cartilage breakdown.¹⁸ A systematic review of animal studies confirmed that exosomes from mesenchymal stem cells consistently increased cell growth, enhanced matrix formation in cartilage, and improved tissue quality scores.¹⁹

This shift in understanding, from “the cells become cartilage” to “the cells change the local environment,” actually explains some of the clinical patterns. It accounts for why improvements often appear gradually over months rather than immediately. It also explains why such varied cell sources (bone marrow, fat, umbilical cord) can produce similar outcomes: the important product is the anti-inflammatory signaling, not the specific cell lineage.

Safety and Side Effects

On the safety front, the news is genuinely reassuring. A systematic review of nearly 2,000 patients found an overall rate of transient adverse events of about 12%, consisting mostly of swelling and pain at the injection site. No infections, fat embolisms, or other medical complications were reported. Umbilical cord-derived and culture-expanded adipose-derived injections had higher rates of these temporary reactions (about 52% and 30% respectively) compared to bone marrow and minimally processed adipose preparations.²⁰

A large network meta-analysis estimated that stem cell injections carry about a 49% higher relative risk of joint pain, stiffness, and swelling compared to control treatments. That sounds alarming until you remember that the baseline rate of these events is modest and that they are self-limiting. Serious adverse events across the included studies were rare and generally unrelated to the procedure itself.²¹ In short, the most likely downside is a few days of a sore, swollen knee after the injection. No cancers, no systemic reactions, no tissue overgrowth have appeared in the published data so far.

What It Costs and Why Insurance Won’t Cover It

A 2017 survey of clinics offering stem cell knee injections found that the average price for a single same-day injection was about $5,000, with prices ranging widely depending on the provider and location.²² A more recent survey of physicians confirmed large variations in cost, with bone marrow aspirate and concentrate procedures being the most expensive and having the widest range, while platelet-rich plasma treatments were cheaper and more consistent in pricing. Private practice providers tended to charge more than academic medical centers.²³

Insurance almost never covers stem cell injections for knees. The procedure is considered experimental by most payers because no stem cell product has yet received FDA approval specifically for treating knee osteoarthritis. Clinics that operate within the current regulatory framework are typically using “minimally manipulated” autologous cells (your own bone marrow or fat tissue, processed at the point of care), which falls into a gray area the FDA has addressed through guidance documents and enforcement actions rather than a formal approval pathway.²⁴ The same survey that found average prices near $5,000 also noted that clinics commonly claim about 80% of patients had “good results” or “symptomatic improvement,” a figure the authors flagged as a gap between marketing claims and what published literature actually supports.²²

The Quality-of-Evidence Problem

A systematic review that tried to evaluate the published literature on minimally manipulated stem cell therapies for knee osteoarthritis found 23 papers meeting inclusion criteria, but only 4 were randomized controlled trials. The overall quality of the studies was rated poor using a standard bias-assessment tool.²⁵ Many of the most impressive-looking results come from single-arm studies (everyone gets the treatment, no one gets a placebo), case series with small numbers of patients, or studies with short follow-up. This doesn’t mean those results are wrong, but it means they are easy to overinterpret.

Consider the contrast again: the four-year study of 37 knees reported a 95% improvement rate, but it had no control group. The ADIPOA2 trial, which did have a placebo control, found that the stem cell group didn’t outperform placebo at six months.⁷ Both studies are published in reputable journals. The difference is not that one is lying; it’s that the study design dramatically changes how you read the numbers. When someone tells you stem cell therapy has a 90% success rate, the first question should be: compared to what?

Injection Technique and Rehabilitation

How the cells are delivered matters more than most patients realize. A systematic review of randomized trials found that ultrasound-guided knee injections are consistently more accurate than blind (landmark-only) injections.²⁶ A recent analysis estimated that ultrasound guidance achieves accuracy rates of 95 to 96% versus 78 to 83% for landmark-based technique.²⁷ When you are paying thousands of dollars for cells that took an hour to harvest, having them actually land inside the joint rather than in the surrounding tissue seems like a bare minimum expectation. If a provider is not using image guidance, that is worth asking about.

Post-injection rehabilitation also appears to matter. A trial comparing patients who received structured physical therapy after stem cell transplantation to those who received conventional care alone found that the therapy group showed greater improvements in pain, range of motion, and functional limitations.²⁸ This aligns with what orthopedic surgeons have long known about knee procedures in general: the rehabilitation is at least as important as what happens during the procedure itself. A stem cell injection followed by returning to sedentary habits is unlikely to produce the same results as one followed by a thoughtful, progressive strengthening program.

Exosome Therapy as a Next Step

One of the most active areas of development involves skipping the cells entirely and using only the exosomes they produce. Because the cells’ therapeutic value appears to come largely from what they secrete, the thinking is that delivering a concentrated dose of those secretions could provide the same anti-inflammatory and regenerative signals without the logistical burden of harvesting, processing, and injecting living cells. Animal studies have shown that exosomes from bone marrow stem cells can promote cartilage repair, enhance extracellular matrix production, and reduce pain behaviors in osteoarthritis models.²⁹ Human trials are still in early stages, but the concept could eventually simplify production, standardize dosing, and reduce costs. For now, exosome products marketed directly to consumers are unregulated and should be treated with skepticism until clinical trial data catches up.