GLP-1 · Knee Osteoarthritis · Clinical Evidence · Verified July 18, 2026

GLP-1 for Knee Osteoarthritis: What the Evidence Actually Shows

Published: July 18, 2026 · Last reviewed: July 18, 2026

By The RX Index Editorial Team · Last verified: July 18, 2026 · How we verify

Sources: NEJM 2024 (STEP 9) · Am J Clin Nutr 2021 (Gudbergsen) · Cell Metabolism 2026 (Qin et al.) · FDA Wegovy PI rev. 06/2026 · CMS Medicare GLP-1 Bridge · Cigna GLP-1 PA Policy

The RX Index is the independent GLP-1 decision resource that scores telehealth providers and treatment paths on clinical legitimacy, care quality, transparency, access, and cost. We are not affiliated with or endorsed by any GLP-1 manufacturer or telehealth provider.

If you searched GLP-1 for knee osteoarthritis, you probably ran into a headline with a big number in it. Semaglutide cut knee pain by 41.7 points. Ozempic eases arthritis. That kind of thing.

The number is real. It came from a Phase 3 trial called STEP 9, published in the New England Journal of Medicine.

Here’s the part that didn’t make the headlines: the placebo group in that same trial cut their pain by 27.5 points. Same 68 weeks. Same diet and activity coaching. A placebo injection instead of the drug — and about two-thirds of the improvement.

The honest measure of what semaglutide added is the gap between those groups: 14.2 points. That’s a real, clinically meaningful result. It’s just not the number on the poster.

Below, we’ll show you whether the trial population looks anything like you, what this costs in July 2026, why your knee arthritis itself might be the thing that gets you covered, and what to do if you already have surgery on the calendar.

And there’s one more thing. A second randomized trial, on a different GLP-1, run by the same research institute in Copenhagen, found no knee-pain benefit at all. That trial isn’t a footnote. It’s the key to the whole question, and we’ve put the two side by side below.

Best for you / not for you

GLP-1 for knee osteoarthritis: does it actually work?

WOMAC is the standard knee osteoarthritis questionnaire — the Western Ontario and McMaster Universities Osteoarthritis Index. On the version used in STEP 9, 0 means no pain and 100 means the worst pain. People entering the trial averaged 70.9, a high baseline; the study required pain that was at least moderately severe.

Below are the highest-decision-value knee-specific studies and active programs we verified through July 18, 2026. Read the “what the comparison group got” column before the one next to it. That column is why this page exists.

The Knee-OA GLP-1 Evidence Ledger

Approved by the FDA to treat knee osteoarthritis: none of them.

How we built this table. We searched PubMed, ClinicalTrials.gov, DailyMed, FDA drug databases, and sponsor newsrooms through July 18, 2026 for randomized trials, observational studies, and registered programs where knee osteoarthritis was an enrolled condition or a measured outcome for a GLP-1 receptor agonist. We ranked a peer-reviewed randomized trial above a press release, above an observational study, above an animal study — the same standard for every row.

What that placebo number is really telling you

The most useful finding in STEP 9 might be in the control group. Everyone in that trial — both arms — got reduced-calorie diet counseling and physical activity counseling. The placebo group lost 3.2% of their body weight and still improved 27.5 points on pain.

Be careful about what that proves. STEP 9 can’t separate how much of the control group’s change came from counseling, how much from the placebo effect, how much from the natural ups and downs of arthritis pain, and how much from people simply starting at a bad moment and drifting back toward their average. What it does show is that the control arm moved a long way, and that semaglutide was tested on top of diet and activity support, not instead of it.

Henning Bliddal, MD, the rheumatologist who led STEP 9 and heads the Parker Institute at Copenhagen University Hospital, described the trap: when knees are affected enough, “you cannot exercise the weight off. So you’ll have to do something else.”

One number that gets compared wrong

You’ll see pages line up STEP 9’s −41.7 next to retatrutide’s −4.5 and conclude semaglutide is nine times better. It isn’t. The two trials used different scales. STEP 9 scored WOMAC pain from 0 to 100, starting at 70.9. TRIUMPH-4 scored it from 0 to 10, starting at 6.0. A 4.5-point drop on a 10-point scale is a large move. Don’t compare those two numbers directly. They aren’t measuring on the same ruler.

Does the STEP 9 evidence apply to you?

Who was actually in STEP 9

Kellgren–Lawrence is the 0–4 scale radiologists use to grade how much joint damage shows on an X-ray. Grade 2–3 is the middle. Not early, not bone-on-bone.

A partial match doesn’t mean no. It means the trial is a weaker guide for you, and the conversation with your doctor matters more.

Is any GLP-1 FDA-approved for knee osteoarthritis?

No approved knee indication exists. No telehealth platform can create one, and nobody is going to hand you a prescription labeled “for arthritis.”

But “not approved for it” and “no doctor will prescribe it” are two different things. The FDA itself recognizes that health professionals generally may prescribe an approved drug for an unapproved use when they judge it medically appropriate for their patient. That’s what off-label prescribing is, it’s legal, and it’s routine across medicine.

Here’s the practical part, and it’s better news than most pages give you: the people studied in STEP 9 — adults with obesity and knee pain — are already the people these drugs are approved for. For most readers this never becomes an off-label question at all. You meet the weight criteria, you’re treated under the approved use, and the knee benefit rides along.

Is the relief from the weight loss, or is the drug doing something to the joint?

Two trials. Same institute. Opposite results.

The Parker Institute at Copenhagen University Hospital ran a randomized trial of liraglutide 3 mg in people with overweight or obesity and knee osteoarthritis, published in the American Journal of Clinical Nutrition in 2021. Henning Bliddal — the same rheumatologist who later led STEP 9 — is an author on both papers.

That trial found no knee-pain benefit. Pain scores moved 0.9 points apart between groups, with a p-value of 0.71. Same disease. Same institute. Same senior investigator. Same drug class. Opposite answer.

The amount of weight that came off

The 25 lb figure is our calculation from published group means: −13.7% versus −3.2% of an average starting weight of 239.5 lb. The 3.9 kg figure is reported directly in the liraglutide paper.

First, the liraglutide trial made everyone diet first. Only people who had already lost more than 5% got randomized. Both groups had already banked whatever pain relief that early weight loss delivers. The trial was measuring what the drug adds on top of that — a much harder bar to clear.

Second, the weight gap was roughly three times larger in STEP 9. About 25 pounds of separation versus about 8.6.

We’re labeling this precisely, because the label matters: this is a hypothesis-generating pattern across two separate trials, not a dose-response finding, and it is our editorial reading rather than a conclusion either paper states. The trials differed in design, population, timing, and pain scale. Neither was built to test the other.

What it suggests, practically: the amount of weight you lose may matter a great deal.

What the lab work shows, honestly

There is real science suggesting GLP-1 medicines may act on joints directly. Cartilage cells — chondrocytes — carry GLP-1 receptors. In animal models, GLP-1 drugs reduce joint inflammation. The short version: it’s promising, it’s mostly in mice, and no published human trial has cleanly separated “the joint feels better because you weigh less” from “the joint feels better because the drug acted on it.”

How much weight do you have to lose before your knee feels it?

The four-to-one estimate comes from Stephen Messier’s work published in Arthritis & Rheumatism in 2005. It’s a measurement of force, not a promise about pain — load and pain are related but not the same thing.

IDEA was a diet-and-exercise trial, not a GLP-1 trial — it describes what happens at a given amount of weight loss, not what any drug guarantees. These are group averages, not thresholds you personally have to hit.

Can a GLP-1 rebuild knee cartilage?

What the study did

Qin and colleagues gave semaglutide weekly to obese mice with osteoarthritis induced by a surgical procedure. They found less cartilage breakdown, fewer bone spurs, less inflammation in the joint lining, and less pain sensitivity. The mechanism they describe runs through a pathway they label GLP-1R → AMPK → PFKFB3, which changes how cartilage cells generate energy.

The design detail that makes the paper worth reading: the researchers included a pair-feeding control group — mice fed the same restricted amount of food as the semaglutide group, so both lost similar weight. That’s what lets them argue the cartilage effect wasn’t simply a consequence of weight loss. It’s a genuinely well-built experiment. It also doesn’t transfer. A pair-feeding control strengthens the mechanism claim in the mouse model. It does not establish that mechanism in a person.

Why we’re not calling it proven

It’s mice. Surgically induced osteoarthritis in a mouse is not decades of wear in a human knee.

The human component is a small pilot, registered as ChiCTR2200066291. It was too small and too short to establish cartilage regeneration, and it paired semaglutide with intra-articular hyaluronic acid — injections into the joint itself — which means an imaging signal can’t be attributed to semaglutide alone.

The ladder isn’t climbed. Getting from promising to established takes: cell-level pathway → animal results → pilot signal → large randomized human trial with imaging endpoints → long follow-up → independent replication. This work is on rung three.

Which GLP-1 has the most knee-specific evidence?

Our editorial read: if the knee is your reason, the medicine with actual knee evidence behind it is semaglutide 2.4 mg — and it’s an FDA-approved option you can reach through insurance, the Medicare GLP-1 Bridge, manufacturer-direct cash pricing, or telehealth. Tirzepatide is a legitimate choice if greater total weight loss is the goal and your prescriber agrees. Just be clear with yourself that you’d be reasoning from weight-loss data, not knee data. There is no completed head-to-head knee trial between them.

How should a GLP-1 fit into standard knee osteoarthritis care?

Remember what STEP 9 actually tested: semaglutide on top of diet and activity counseling. It never tested the drug by itself. A GLP-1 is one tool added to the plan. It doesn’t replace the plan.

Can a GLP-1 help you avoid a knee replacement?

The study is Carter and colleagues, published in Regional Anesthesia & Pain Medicine on June 2, 2026, drawing on the TriNetX records network for knee osteoarthritis diagnosed between 2010 and 2024. The effect was strongest with sustained use and with semaglutide or tirzepatide.

We report the absolute number every time, because “significantly lower risk” without it is exactly how a modest finding becomes a promise. 1.44 percentage points is real, and it is modest.

“My surgeon says I have to lose weight before surgery”

What’s true on the surgeon’s side: higher BMI is genuinely associated with higher surgical risk, especially infection around the implant at the highest BMI levels.

What’s also true: a national survey of arthroplasty surgeons found 41.65% use a BMI cutoff below 40 for knee replacement, and 24.39% use no strict cutoff at all (PubMed 41706617).

What “conditionally recommends against” means: the guideline panel judged the supporting evidence to be low quality and concluded the decision should be sensitive to the individual patient’s values and circumstances.

Which means it’s a conversation, not a verdict. And you are allowed to have it.

If it helps, here’s a sentence you can borrow:

“I’ve read the 2023 ACR and AAHKS guideline on optimal timing for joint replacement. Can we talk about whether delaying is the right call in my specific case, and what you’d want to see from me either way?”

You’re not being difficult by asking that. You’re being an informed patient about a guideline written for exactly your situation. If your surgeon holds the line after that conversation, a second opinion is a reasonable next step — roughly a quarter of surgeons don’t use a hard cutoff.

What if knee surgery is already on your calendar?

Your pre-surgery checklist

Bring these six items to your pre-op appointment. Screenshot this list or print the page:

1. Medication name and formulation (injection or pill)
2. Current dose
3. Date of your last dose
4. Whether you’re still increasing the dose
5. Any active nausea, vomiting, or fullness
6. The October 2024 multisociety guidance, by name, so the team knows the reference you’re working from

What could go wrong, specifically for someone with a bad knee?

Everything in this section comes from the current Wegovy prescribing information. Other GLP-1s and dual or triple agonists have their own labels, and the details differ. Ask about the specific medication you’re considering.

The fracture observations in the label

In the cardiovascular outcomes trial (SELECT) described in the Wegovy prescribing information, hip and pelvis fractures were reported more often on semaglutide than on placebo in two subgroups:

• Among female patients: 1% versus 0.2%
• Among patients aged 75 and older: 2.4% versus 0.6%

The June 2026 label also reports fractures in the MASH trial: 4.4% with Wegovy versus 3.3% with placebo.

Now look back at who was in STEP 9: 81.6% women, average age 56. These are label-reported observations, and the label does not establish that semaglutide caused them. And they are squarely relevant to the people reading this page. If you’re an older woman considering this, it’s a specific, reasonable thing to raise — and a reason to ask whether bone-health assessment and knee-safe strengthening belong in your plan.

The rest of the safety picture

• A boxed warning for thyroid C-cell tumors, based on rodent studies; human relevance is undetermined. Contraindicated if you or a family member has had medullary thyroid carcinoma or MEN 2.
• Warnings covering pancreatitis, gallbladder disease, kidney injury from dehydration, and increased heart rate.
• Direction to discontinue at least 2 months before a planned pregnancy.
• Body-composition data showing greater fat-mass loss than lean-mass loss — greater, not exclusive.

What does it cost, and who actually pays for it?

The Access Reality Table

All pricing and program terms verified July 18, 2026. Re-check before you commit — these have changed three times in the past year.

If you’re on Medicare, three things nobody explains

1. Eligibility is judged at the moment you started GLP-1 therapy, not today. Someone who began at BMI 37 and is now at 34 still qualifies.
2. The covered list is narrow. Foundayo (all forms), Wegovy (all forms), and Zepbound KwikPen. Zepbound vials and single-dose pens are out. Pen needles aren’t covered.
3. The Bridge prior-authorization process has no appeal — but a prescriber can resubmit with corrected or additional information. Getting it right the first time matters more than usual.

Full details: Medicare GLP-1 Bridge eligibility →

Where to start if you want the brand-name path

If you already have a prescriber willing to write it and you’re paying cash, compare manufacturer-direct pricing first. It often costs less than adding a telehealth membership on top — though insurance or another program you qualify for may cost less still.

Ro carries Wegovy (pen and pill), Zepbound, Foundayo, Ozempic, and Saxenda, at cash prices matching NovoCare, LillyDirect, and TrumpRx. Membership is $39 for the first month, then $149/month, or as low as $74/month with an annual plan paid upfront — medication billed separately. Their insurance concierge handles benefit checks and prior-authorization paperwork, which for this readership is the single most valuable thing on offer. (Affiliate link — see disclosure at top.)

Now the part most sites won’t print. Ro does not offer compounded GLP-1s. If the lowest possible monthly price is your priority, a compounded program will generally cost less. If that’s you, use Find My GLP-1 Path instead — it covers those programs and we’ll point you there honestly.

But here’s why we’re comfortable pointing you to Ro on this page specifically: a Ro patient who is prescribed Wegovy injection at the 2.4 mg maintenance dose is on the brand and regimen studied in STEP 9. No compounded product has knee osteoarthritis evidence behind it at all. Worth checking: Ro’s current materials indicate that Medicare, Medicare supplement, and TRICARE members may still be eligible for select cash-pay options, and that FEHB members can use the insurance concierge. Verify your specific situation in Ro’s enrollment flow. If you’re on Medicaid, your path is your state formulary, not a cash telehealth membership.

What should you ask your doctor?

Bring this list. Print it if that’s easier.

1. Is my knee pain definitely osteoarthritis, and how severe is it on imaging?
2. How closely do I match the population studied in STEP 9?
3. Do I meet the criteria for a specific GLP-1 under an approved indication?
4. Does my plan count knee osteoarthritis as a qualifying weight-related comorbidity?
5. What’s the goal — weight, pain, function, surgical clearance, or metabolic health?
6. What result would count as meaningful improvement, and by when?
7. What’s the plan if my weight improves but my knee doesn’t?
8. What warnings or contraindications apply to me specifically?
9. How do we protect muscle and bone while I lose weight?
10. How does this fit with physical therapy and my orthopedic care?
11. What should my surgical and anesthesia teams know?
12. What happens if coverage or affordability interrupts treatment?

You are allowed to ask about this. It’s a reasonable question backed by a Phase 3 trial in a major journal, and on some plans your knee diagnosis is the specific thing that makes you eligible. If a previous conversation about your weight left you feeling dismissed, that’s not a reason to stop asking. It’s a reason to walk in with the specifics.

What’s coming next?

The one to watch, and why

4P004 is intra-articular liraglutide — injected directly into the joint space rather than under the skin. Read that again alongside the liraglutide section above. The same molecule that failed to relieve knee pain when injected under the skin is now being injected straight into the knee.

That’s not a contradiction. It’s a much cleaner test of local activity. Under the skin, liraglutide would work mainly through weight loss — and in that trial it didn’t move enough weight to matter. Injected into the joint, systemic exposure is far lower, so whatever happens is much more likely to reflect a local joint effect.

The sponsor reports that the Phase 1 study (LASARE, NCT05419856) enrolled 34 patients across three Belgian sites and met its primary safety objective, with drug levels in the bloodstream lower than with the approved under-the-skin version. Phase 2a is now running in patients aged 40–80 with synovitis — inflammation of the joint lining — who haven’t gotten relief from at least two prior treatments.

How we verified this

We are not clinicians and this page is not medical advice. What we do is check every consequential claim against the organization that issued it, and show our work.

Claims we found and did not publish, because we couldn’t trace them to a primary source: that GLP-1s reverse cartilage damage in humans; that 7.5% weight loss cuts knee replacement risk by 31%; and per-step force figures attributed to specific brand-name drugs. If any of those become properly established, we’ll add them and date the change.

Who wrote this: the editorial team at The RX Index. We do not add a “medically reviewed by” line unless a named clinician actually reviewed the page, and none did. We’d rather tell you that than borrow authority we haven’t earned. Corrections: when we change something material, we’ll say what changed, what it said before, when we changed it, and what prompted it.

The questions behind the search

These are the kinds of questions people are actually asking online about GLP-1s and knee pain. They’re questions, not evidence — we include them because they show what’s on people’s minds, not because they demonstrate that anything works.

• Whether anyone has used Wegovy specifically for knee pain relief
• Whether any study shows a GLP-1 helping osteoarthritis independent of weight loss
• Whether to start a GLP-1 in order to qualify for knee replacement — and whether that feels like the right thing to do

Frequently asked questions

Published: July 18, 2026 · Last reviewed: July 18, 2026