2025 Meta-Analysis: MSC Therapy in Knee Osteoarthritis

A recurring frustration in regenerative medicine is that the literature is voluminous but hard to summarize for patients. There are hundreds of MSC-for-knee-osteoarthritis papers. Some are small case series. Some are well-controlled RCTs. Most aren't quite comparable on protocol or outcomes.

Meta-analyses help. They pool studies that share enough methodology to be combined, and produce more stable estimates than any single trial. Several have come out in the last few years for MSC therapy in knee osteoarthritis. This post summarizes what the pooled evidence currently shows, what it doesn't, and how it affects what I tell patients during consultation.

What's pooled and how

The strongest recent meta-analyses pool 15 to 20 randomized controlled trials, totaling roughly 1,200 to 1,500 patients, comparing intra-articular MSC injection against saline, hyaluronic acid, or in some cases PRP. Cell sources varied (umbilical, bone marrow, adipose, Wharton's jelly), as did cell counts and protocols, which is a limitation worth naming up front.

The headline outcomes:

• Pain (VAS): Mean reduction at 12 months across studies is roughly 30 to 50 percent from baseline, depending on study and severity, with the MSC arm consistently outperforming control arms.
• Function (WOMAC, KOOS): Meaningful improvement in functional scores at 6 and 12 months, with effect sizes typically in the moderate range (Cohen's d 0.5 to 0.8).
• Structural change on MRI: Modest but real signal of cartilage stabilization or mild thickening in several studies. The structural effect is smaller than the symptomatic one.
• Adverse events: Low rates of serious adverse events. Transient post-injection inflammation in 10 to 25 percent of patients depending on study.

The durability question (how long does the effect last?) shows continued meaningful improvement at 12 months and a slow regression toward baseline by 24 months in many cohorts. A subset of patients sustain results beyond 24 months. Most patients who respond well to a first round are candidates for repeat dosing at 12 to 18 months if they want to maintain the response.

What this changes, and what it doesn't

For the patient sitting across from me in clinic, this body of evidence does a few things.

It makes the conversation honest. Five years ago, I had to caveat the literature heavily ("there are some studies, but..."). Now I can say: pooled data from over 1,000 patients across randomized trials shows that MSC injection consistently outperforms saline and hyaluronic acid for moderate knee OA in pain and function, durably to 12 months. That's a different sentence to start with.

It clarifies who's a good candidate. The strongest signal is in moderate osteoarthritis (Kellgren-Lawrence grade II to III), with intact joint space and no severe deformity. The signal is weaker for end-stage bone-on-bone disease (grade IV), which is consistent with what we see clinically. We don't promise people with end-stage disease that cells will reverse what surgery is the right answer for.

It doesn't make cells a cure. The pooled effect sizes are real but moderate. The 30 to 50 percent average pain reduction means a meaningful number of patients have transformative responses and a meaningful number have partial or non-responses. That distribution is what we tell patients to expect.

It doesn't tell us the best cell source definitively. Across the pooled studies, umbilical and Wharton's jelly MSCs tend to produce the strongest effects, with adipose and bone marrow producing smaller-but-positive effects. The head-to-head comparisons across multiple cell sources in a single trial are rare, so we're inferring rather than knowing for certain. The biology favors young donor cells, the pooled evidence is consistent with that read, but a definitive head-to-head trial would be welcome.

What it changes about Apex's counseling

A few small but specific shifts:

On timeline. I tell patients to expect partial relief by 6 to 8 weeks, with continued improvement through 6 months, peak effect around 6 to 12 months, and slow regression toward baseline by 18 to 24 months. The published cohorts support this curve well.

On response rate. I tell patients roughly 60 to 70 percent of moderate-OA patients hit a meaningful response (50 percent or greater pain reduction), with another 10 to 15 percent getting partial benefit. Roughly 1 in 5 patients see minimal or no symptomatic change.

On re-dosing. I tell patients we'll re-evaluate around 12 months. If they're still doing well, we wait. If the effect is beginning to fade, we discuss a second round. This is supported by the durability data in the pooled trials.

On end-stage disease. I tell patients with grade IV osteoarthritis and bone-on-bone disease that the literature doesn't support the kind of result they're probably hoping for, and that they should talk to a surgeon about replacement. This conversation gets easier when I can show the evidence behind it.

What I'd still like the literature to give us

Honest limits:

• More head-to-head studies of cell source (umbilical vs bone marrow vs adipose) in a single trial with standardized dosing.
• Cleaner dose-response data. Is more cells better? At what point does the marginal benefit flatten?
• Combined-protocol studies. MSC plus exosomes vs MSC alone. MSC sequenced after PRP priming vs MSC alone.
• Longer-term follow-up (3 to 5 years) on the durability question.
• Real-world registry data with consistent outcome reporting.

The next 5 years of this literature should be productive. The current literature is already strong enough to justify offering the protocol, with appropriate patient selection.

A practical takeaway

If you're researching stem cell therapy for knee osteoarthritis, the question is no longer "is there evidence?" The answer to that is yes. The question is whether your specific case fits the population the evidence supports. That's a workup question, not a marketing one.

To find out where your case falls, request a consultation. We'll look at your imaging, evaluate the indication honestly, and tell you whether the evidence supports a protocol for you.