Retatrutide eased knee arthritis pain in two phase 3 trials published in the New England Journal of Medicine on November 8, 2026, and the placebo groups improved by more than half as much.1 In TRIUMPH-4, pain on a 0 to 10 scale fell 3.96 points on the 9 mg dose and 2.12 points on placebo after 68 weeks, so the drug's own share was 1.83 points.1
The benefit was real and showed up in both trials, but it is smaller than the number most people have seen. Lilly's December 2025 announcement put the drop in knee pain at up to 75.8%. That figure was a percentage worked out after the trial, under a model that assumes everyone stayed on the drug, and the same table put placebo at 40.3%.2 The paper's main analysis counts the people who stopped, and it reports points.1
Both trials were funded by Eli Lilly. They measured pain, function and weight, not the joint itself, and the authors say they were not designed to show whether the drug slows arthritis.1 Retatrutide is not approved anywhere.3
What did the NEJM paper find?
The paper found that weekly retatrutide beat placebo on knee pain, physical function and body weight in adults with obesity and painful knee osteoarthritis, in both of the trials it reports.1 Knee osteoarthritis is the common arthritis where the cartilage cushioning the joint thins and the knee hurts with use.
TRIUMPH-4 enrolled 445 adults without diabetes and ran for 68 weeks on 9 mg, 12 mg or placebo. TRIUMPH-1-OA was the knee arthritis group inside TRIUMPH-1, the 2,339-person obesity trial we covered in October, and it ran 80 weeks with 574 people on 4 mg, 9 mg, 12 mg or placebo.14 Everyone entered with a pain score between 4 and 9 on the WOMAC, the standard knee arthritis questionnaire, scored here from 0 to 10. The average at the start was about 6, and the average starting weight was about 248 pounds.12
| Trial and arm | Pain change (0 to 10 scale) | Pain vs placebo | Weight change | Stopped for side effects |
|---|---|---|---|---|
| TRIUMPH-4, placebo | -2.12 | -4.6% | 4.0% | |
| TRIUMPH-4, 9 mg | -3.96 | -1.83 | -20.0% | 12.2% or 12.9% |
| TRIUMPH-4, 12 mg | -3.74 | -1.61 | -23.7% | 18.2% |
| TRIUMPH-1-OA, placebo | -1.87 | -4.3% | 6.2% or 6.3% | |
| TRIUMPH-1-OA, 4 mg | -3.22 | -1.35 | -17.5% | 2.8% |
| TRIUMPH-1-OA, 9 mg | -3.48 | -1.61 | -22.9% | 8.4% |
| TRIUMPH-1-OA, 12 mg | -3.64 | -1.77 | -24.8% | 15.3% |
Pain was measured at week 68 in TRIUMPH-4 and week 80 in TRIUMPH-1-OA. Every difference from placebo in the table was statistically significant.1 Where two figures appear, the paper's text and its table give slightly different counts, and we print both.
Physical function, the WOMAC's score for tasks like climbing stairs and getting out of a chair, moved the same way. It improved by 1.68 to 1.95 points more than placebo across the active doses.1
Why is the paper's number smaller than the 75.8% people quote?
The two numbers answer different questions. Trials have to decide what to do about people who stop the drug partway through, and the choice is called an estimand.
The efficacy estimand asks what would have happened if everyone had stayed on the drug as assigned. The treatment-regimen estimand counts everyone who was randomized, including people who quit, the way a real prescription would play out. The NEJM paper used a hybrid of the two for pain and the treatment-regimen version for weight.1
Lilly's release gave both versions and led with the efficacy one.2 On weight, the gap looks like this for TRIUMPH-4's 12 mg group: 28.7%, or 71.2 pounds, under the efficacy estimand, and 23.7%, or 60.0 pounds, under the treatment-regimen estimand.2 On pain, the 75.8% was a post hoc percentage, meaning it was calculated after the fact and not planned before the trial started.2
The percentage also hides how much the placebo group improved. Placebo fell 40.3% in the same analysis, so a little over half of that 75.8% happened without the drug.2 The 75.8% figure appeared on our own retatrutide page as well, and that line is being corrected alongside this piece.3
How much of the knee pain relief came from the drug?
About 1.6 to 1.8 points on a 10-point scale, in both trials, out of a total drop of roughly 3.5 to 4 points.1 The rest of the improvement showed up in the placebo group too.
Several ordinary things can make a placebo group's knees feel better. People in the placebo arms still lost 4.3% to 4.6% of their weight, and less weight means less load on the joint.1 People tend to join a pain trial when their pain is at its worst, and scores drift back toward normal over time on their own, an effect called regression to the mean. Pain also responds strongly to expectation. A placebo arm captures all of these at once, and a trial like this cannot tell them apart.
The clearest way to see the drug's effect is the responder count. Pooled across both trials, 69.6% of people on 9 mg and 63.2% on 12 mg cut their pain at least in half, against 37.8% on placebo.1
How the drug does it is not settled. The lead author, Thomas Schnitzer of Northwestern, credits the weight: "A major factor is that you're loading your knees less."5 An orthopedic surgeon interviewed by Medical News Today, Timothy Gibson, noted the drug is "not an anti-inflammatory medication administered into the joint."6 A blood marker of inflammation, hsCRP, fell in TRIUMPH-4, but it was measured only at the start and the end, and the trials cannot show how much of the relief came from weight loss and how much from anything else.17
Semaglutide went through a similar test. In STEP 9, a 68-week NEJM trial published in 2024 of semaglutide 2.4 mg in people with knee osteoarthritis and obesity, the placebo group's WOMAC pain improvement was also about two-thirds the size of the drug group's.8 The two trials used different people and different scales, so they do not tell you which drug works better on knees.
Does the 12 mg dose help knees more than 9 mg?
Not in these trials. In TRIUMPH-4, pain fell 3.96 points on 9 mg and 3.74 points on 12 mg, and the confidence intervals around the two differences from placebo overlap almost completely (1.30 to 2.37 points against 1.07 to 2.16).1 TRIUMPH-1-OA showed a small step up with each dose, 1.35, 1.61 and 1.77 points better than placebo, again with overlapping intervals.1
Weight loss and side effects both climbed with dose. In TRIUMPH-4, 18.2% of the 12 mg group stopped because of side effects, against roughly 12% to 13% on 9 mg, and dysesthesia hit 20.9% on 12 mg against 8.8% on 9 mg.12
What side effects showed up in the knee trials?
The side effects were the familiar ones for this drug class, plus a skin sensation problem that ran higher here than in the main obesity trial. Nausea, diarrhea and constipation were the most common complaints, the same pattern our GLP-1 side effects guide describes for the class. In TRIUMPH-4, 8.1% of the 12 mg group stopped because of stomach and gut effects, against 1.3% on placebo.1
Dysesthesia means abnormal skin sensation: tingling, burning, or skin that hurts to touch. It affected 8.8% to 20.9% of people on retatrutide in TRIUMPH-4 and 6.3% to 16.8% in TRIUMPH-1-OA, against 0.7% on placebo in each.1 The full TRIUMPH-1 trial reported 12.5% at 12 mg against 0.9% on placebo, so the knee trials sit at the high end of what has been published.4 Our guide to retatrutide skin sensitivity covers what people describe and how it has behaved over time.
Serious adverse events, meaning ones that put someone in the hospital or carried similar weight, were 9.5% and 10.8% on the two TRIUMPH-4 doses against 2.0% on placebo. In TRIUMPH-1-OA they ranged from 4.2% on 4 mg to 14.6% on 12 mg, against 6.9% on placebo.1 Two people died in TRIUMPH-4, one on each dose, and investigators judged both deaths unrelated to the drug. Nobody died in TRIUMPH-1-OA.1
Cancers were reported in small numbers with no steady pattern. TRIUMPH-4 recorded 7 on 9 mg (4.8%), 2 on 12 mg (1.4%) and 1 on placebo (0.7%). TRIUMPH-1-OA recorded 4.2% on 12 mg against 2.1% on placebo, and the same 2.1% on 9 mg.1 Counts this small can swing on a single case, and they do not establish that the drug causes cancer.
The lowest dose was the easiest to stay on. On 4 mg in TRIUMPH-1-OA, only 2.8% stopped for side effects, fewer than the 6.2% to 6.3% who stopped on placebo.1
What is known, and what is not?
What the two trials show:
- Retatrutide beat placebo on knee pain by 1.35 to 1.83 points on a 10-point scale, in two separate trials, at every dose tested.1
- Physical function improved by a similar margin.1
- Roughly two in three people on 9 mg or 12 mg halved their knee pain, against about 38% on placebo.1
- Weight fell 17.5% to 24.8% over 68 to 80 weeks, counting everyone randomized.1
- Stopping for side effects and dysesthesia both rose with dose.1
What they do not show:
- Whether the drug protects cartilage or slows the disease. The authors write: "Our trials were not designed to assess avoidance of knee replacement, OA progression, or disease modification."1
- How much of the relief comes from weight loss alone.17
- How quickly the benefit starts. The first pain check was at week 16.1
- Whether it helps knee arthritis in people without overweight or obesity, or in people with diabetes, who were excluded.1
- What happens to knee pain after someone stops the drug.
What does this change for someone with a vial?
The trials show what Lilly's retatrutide does, made to drug manufacturing standards and given at fixed doses with scheduled increases.1 Whether a gray-market vial behaves the same depends on whether it holds that molecule at the stated amount.
Gray-market retatrutide is a different supply. Our November 11 price read found it listed by 189 vendors across 735 listings, with 10 mg vials running $47.00 to $135.99 across 138 stores.3 Independent blind testing of online retatrutide runs to nearly 4,000 results, and our piece on the counterfeit retatrutide test record sets out where the failures cluster.3 A vial that is underdosed, overdosed or not retatrutide at all will not reproduce either the benefit or the side effect rates above.
The careful version of this looks the same as it does for any retatrutide purchase: a per-lot certificate of analysis with the lot number printed on the vial, a seller on our vendor list with no open FDA warning letter, an independent test of the first vial, and the dose calculator run on the concentration the test actually found.
When could retatrutide be prescribed for knee pain?
Not until the FDA approves it. Lilly has said it plans to submit retatrutide to the FDA in the first quarter of 2027, and a standard review of 10 to 12 months puts a decision in late 2027 or 2028.3 Our page on when retatrutide could be approved tracks that timeline.
Whether knee osteoarthritis appears on a future label depends on what Lilly files and what the FDA accepts, and neither has been announced. The TRIUMPH-1 results that included these knee patients are covered in our piece on the published trial, and our retatrutide page carries the full record.
Common questions
Does retatrutide help knee arthritis pain?
Yes, in two phase 3 trials published November 8, 2026. Pain on a 0 to 10 scale fell about 1.4 to 1.8 points more than on placebo, in adults with obesity and painful knee osteoarthritis.
Where does the 75.8% figure come from?
Lilly's December 11, 2025 announcement of TRIUMPH-4. It was a post hoc percentage under the efficacy estimand, a model that assumes everyone stayed on the drug. Placebo fell 40.3% in the same analysis.
What did the NEJM paper report instead?
Points on the pain scale, under a hybrid estimand that accounts for people who stopped. In TRIUMPH-4, pain fell 3.96 points on 9 mg, 3.74 on 12 mg and 2.12 on placebo.
Why did the placebo group improve so much?
Several reasons the trial cannot separate. The placebo arms lost 4.3% to 4.6% of body weight, pain scores drift back from a bad starting point, and pain responds strongly to expectation.
How much weight did people lose in the knee trials?
Counting everyone randomized, 20.0% on 9 mg and 23.7% on 12 mg in TRIUMPH-4, against 4.6% on placebo. In TRIUMPH-1-OA it ranged from 17.5% on 4 mg to 24.8% on 12 mg.
Is 12 mg better than 9 mg for knee pain?
Not in these trials. In TRIUMPH-4, 9 mg reduced pain slightly more than 12 mg, and the confidence intervals overlapped. Side effects and dropouts were higher on 12 mg.
What is dysesthesia?
An abnormal skin sensation, such as tingling, burning or skin that hurts to touch. It affected up to 20.9% of people on retatrutide in TRIUMPH-4, against 0.7% on placebo.
Does retatrutide help joint pain in other joints?
Unknown. Both trials enrolled people for knee osteoarthritis and scored pain in one knee. Hips, hands and backs were not studied.
Does retatrutide repair cartilage?
Unknown. The trials did not measure the joint structure, and the authors say they were not designed to assess disease progression or knee replacement.
Does retatrutide reduce inflammation in the knee?
Not shown. A blood inflammation marker, hsCRP, fell in TRIUMPH-4, but that is a whole-body measure taken only at the start and end. Nothing was measured inside the joint.
Who ran the trials?
Eli Lilly funded both and was involved in the design, analysis and writing. The corresponding author is Thomas Schnitzer of Northwestern University.
What were TRIUMPH-4 and TRIUMPH-1-OA?
Two phase 3 retatrutide trials in adults with obesity and knee osteoarthritis. TRIUMPH-4 had 445 people and ran 68 weeks; TRIUMPH-1-OA was the 574-person knee group inside the larger TRIUMPH-1 trial and ran 80 weeks.
Did anyone die in the knee trials?
Two people in TRIUMPH-4, one on each dose. Investigators judged both deaths unrelated to the drug. There were no deaths in TRIUMPH-1-OA.
Does retatrutide cause cancer?
These trials do not show that. Cancer counts were small and inconsistent: 4.8% on 9 mg in TRIUMPH-4 but 1.4% on 12 mg, against 0.7% on placebo.
Is retatrutide approved for knee osteoarthritis?
No. It is not approved for anything, anywhere. Lilly plans to submit it to the FDA in the first quarter of 2027.
How does it compare with semaglutide for knee pain?
Nobody has tested them against each other. Semaglutide's STEP 9 trial also beat placebo on knee pain, with a similarly large placebo response, but the trials used different people and scales.
Will a gray-market vial give the same results?
Only if it contains what the label says at the stated amount. The trial used Lilly's own drug. Independent blind testing of online retatrutide has found vials that fail, so a per-lot certificate and a test of the first vial matter.
Sources
- Schnitzer TJ, et al. Retatrutide Once Weekly in Adults with Knee Osteoarthritis Pain and Obesity. New England Journal of Medicine, published November 8, 2026. doi:10.1056/NEJMoa2610676. nejm.org.Read via a reproduction: the NEJM page would not load (client error), so figures were read from the author accepted manuscript, the peer-reviewed version before typesetting, on the University of Leeds White Rose repository: eprints.whiterose.ac.uk. Reports TRIUMPH-4 (NCT05931367) and TRIUMPH-1-OA (NCT05929066). Source of all trial figures on this page, the estimands, the pooled 50% responder rates, the limitations, the funding statement and the quoted sentence on disease modification. The manuscript's text gives 9 mg discontinuation in TRIUMPH-4 as 18 of 148 (12.2%) and its table as 19 of 147 (12.9%); TRIUMPH-1-OA placebo appears as 6.3% and 6.2%. The final published version could differ in small details from the manuscript. ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩ ↩
- Eli Lilly and Company. News release announcing TRIUMPH-4 topline results for retatrutide. December 11, 2025. investor.lilly.com.Read at source. Company document, not independent confirmation. Gives both estimands for TRIUMPH-4: weight -26.4% and -28.7% (efficacy) and -20.0% and -23.7% (treatment-regimen) on 9 mg and 12 mg; WOMAC pain -75.8% and -74.3% against -40.3% placebo (efficacy) and -67.2% and -62.6% against -35.1% (treatment-regimen), with the percentages described as post hoc. Baseline weight 248.5 lbs, baseline pain 6.0. Dysesthesia 8.8%, 20.9% and 0.7%. ↩ ↩ ↩ ↩ ↩ ↩ ↩
- Peptide Decoding. Retatrutide compound page. peptidedecoding.com/peptides/retatrutide.Our own page, read November 13, 2026. Source of the vendor, listing and price figures (189 vendors, 735 listings, last read November 11, 2026; 10 mg vials $47.00 to $135.99 across 138 stores), the approval status, Lilly's stated plan to submit in the first quarter of 2027 and the 10 to 12 month review estimate, the "nearly 4,000" blind-test figure cited from our counterfeit vial piece, and the "up to 75.8%" line now being corrected. ↩ ↩ ↩ ↩ ↩
- Peptide Decoding. The Retatrutide Trial Is Published. It Never Measured Muscle. October 4, 2026. peptidedecoding.com/news/triumph-1-published.
- Northwestern University. Experimental weight-loss drug eases painful knee osteoarthritis in two clinical trials. November 8, 2026, via EurekAlert. eurekalert.org.Read at source on EurekAlert; the Northwestern newsroom page blocked the fetch. University release about its own investigator's paper. The EurekAlert fetch returned the Schnitzer remark as paraphrase; the quoted wording on this page came through Rheum-Live's report of November 2026 (rheum-live.com, read via a reproduction). Republished by News-Medical the same day with the presentation details: TRIUMPH-4 presented by Philip Conaghan at the American College of Rheumatology annual meeting in Orlando, 11:15 a.m. ET on November 8, 2026, and TRIUMPH-1-OA in a poster session the same day (news-medical.net, read via a reproduction). ↩
- Medical News Today. Knee osteoarthritis: 2 trials link new weight-loss drug to pain relief. November 11, 2026. medicalnewstoday.com.Read at source. Independent reporting. Source of the quoted comment from Timothy Gibson, MD, orthopedic surgeon at MemorialCare Orange Coast Medical Center, who was not involved in the trials. Its pain figures match the paper. ↩
- ConscienHealth. More Than Weight in Knee Pain Relief with Retatrutide? November 11, 2026. conscienhealth.org.
- 2 Minute Medicine. Semaglutide reduces weight and pain in obese patients with knee osteoarthritis. December 13, 2024. 2minutemedicine.com.Read via a reproduction: a summary of the STEP 9 NEJM paper. We did not read the paper directly. Reports WOMAC pain changes of 41.7 with semaglutide 2.4 mg against 27.5 with placebo over 68 weeks, and weight loss of 13.7% against 3.2%. The ratio is what this page uses; the summary does not state the scale, so the raw figures are not compared with the retatrutide trials. ↩
Citing this page. Peptide Decoding. Does Retatrutide Help Knee Arthritis Pain? What the NEJM Trials Actually Show. Trial figures from the TRIUMPH-4 and TRIUMPH-1-OA paper published in NEJM on November 8, 2026, and Lilly's December 11, 2025 release; price figures from our November 11, 2026 capture. Reviewed November 13, 2026. https://peptidedecoding.com/news/retatrutide-knee-osteoarthritis