Back in October we wrote up the TRIUMPH-1 paper and flagged that the knee osteoarthritis subgroup inside it looked like the most commercially interesting thing in the whole dataset. Arthritis has no good drug therapy. Injections, painkillers, and eventually a replacement joint is the entire toolbox. If a weight drug could move pain scores meaningfully, that is a labelled indication, not a side benefit.
On 8 November, Lilly and the TRIUMPH investigators published the trial that tests that properly. TRIUMPH-4 is a stand-alone phase 3 in painful knee osteoarthritis, 445 participants, no diabetes, and it landed in the New England Journal of Medicine the same day it was presented at the American College of Rheumatology annual meeting in Orlando. A companion analysis, TRIUMPH-1-OA, ran in parallel.
This is the paper worth reading closely, because the two ways the results have been reported tell you a lot about how obesity trial data gets spun.
What the trial actually did
445 adults with painful knee osteoarthritis and a BMI of at least 27 were randomised one to one one to once-weekly subcutaneous retatrutide 9 mg, retatrutide 12 mg, or placebo, for 68 weeks. All participants got lifestyle counselling. The only rescue pain medication allowed was paracetamol.
These were not mild cases. Mean baseline BMI was 40.4, mean body weight 112.7 kg, mean WOMAC pain score 5.95 out of 10, and about two thirds of participants had Kellgren-Lawrence grade 3 disease, which means substantial structural joint damage visible on x-ray. Average age was 55 and 63 percent were women. Sites ran across the US, Canada, Mexico, Spain and Australia.
The co-primary endpoints were change in WOMAC pain score and percentage change in body weight, both at week 68.
The pain result
Using the paper’s primary analysis method, mean WOMAC pain scores fell by 3.96 points at 9 mg and 3.74 points at 12 mg, against 2.12 points on placebo. The differences from placebo were 1.83 points and 1.61 points, both statistically significant.
On a 0 to 10 scale where higher is worse, a drop of roughly 4 points is a lot. The effect sizes work out to Cohen’s d of 0.68 to 0.79, which sits in the moderate-to-large range and, as the authors note, compares favourably with what approved osteoarthritis interventions achieve, given those mostly just control symptoms rather than change the underlying load on the joint.
Physical function moved in step. WOMAC function scores fell by 4.14 and 4.20 points with retatrutide against 2.06 on placebo. Separation from placebo was visible by week 16, the first assessment, and continued through week 68.
Presenting the data at ACR, investigator Philip Conaghan of the University of Leeds highlighted a responder analysis: around 70 percent of participants in both retatrutide arms achieved at least a 70 percent reduction in WOMAC pain score, against 26 percent on placebo.
The confirmation trial, TRIUMPH-1-OA, randomised 574 people from the larger TRIUMPH-1 population to 4 mg, 9 mg, 12 mg or placebo and ran to 80 weeks. Pain scores fell by 3.2 to 3.6 points against 1.9 on placebo under the primary estimand. The signal replicated across three doses.
Now the part that needs explaining
Here is where you need your reading glasses, because two different sets of weight numbers are circulating and both are technically correct.
The NEJM primary analysis uses what is called a hybrid-treatment estimand. Under that method, mean body weight fell 20.0 percent at 9 mg and 23.7 percent at 12 mg, against 4.6 percent on placebo. That is the conservative number, and it is the one in the abstract.
The efficacy estimand, which estimates what happens in people who stay on treatment as intended, gives 26.4 percent at 9 mg and 28.7 percent at 12 mg, against 2.1 percent on placebo. That is the number that produced the “28 percent weight loss” headlines, and it is also the number quoted in the ACR abstract.
Neither is dishonest. They answer different questions. The hybrid estimand is closer to what a real clinic population experiences, because real people miss doses and stop treatment. The efficacy estimand is closer to the drug’s ceiling. The gap between 23.7 and 28.7 percent is, essentially, the cost of imperfect adherence, and it is bigger than most coverage implied.
Worth noting: in the efficacy analysis, 58.6 percent of the 12 mg group lost at least 25 percent of body weight. Some participants withdrew from the trial because they considered the weight loss excessive. That is not a sentence you read often in an obesity paper.
Safety, and where this gets real
Gastrointestinal adverse events were the dominant complaint, as with every incretin. More importantly, serious adverse events occurred more often with retatrutide than with placebo in both trials. The paper is explicit about that, and it should be repeated rather than buried, because the population here was mostly people with a BMI around 40 who were being asked to lose a quarter of their body mass in 16 months.
Retatrutide remains investigational. It is not approved anywhere. Lilly filed with the FDA in mid-2026 and no decision has been issued.
There is also a mechanical story worth keeping straight. Each pound of body weight lost takes roughly four pounds of load off the knee during walking, which is why weight loss has always been in osteoarthritis guidelines. The interesting question the trial cannot answer yet is how much of the pain relief is load reduction and how much is the drug doing something else. Retatrutide hits GIP, GLP-1 and glucagon receptors, and the paper notes that reductions in adipose tissue may lower systemic inflammatory mediators, with hs-CRP falling by more than 50 percent at both doses. That is a hypothesis, not a mechanism, but it is the reason this result matters beyond weight.
What the Research Says
- TRIUMPH-4 randomised 445 adults without diabetes with painful knee osteoarthritis and BMI of at least 27 to once-weekly retatrutide 9 mg, 12 mg or placebo for 68 weeks. Under the hybrid-treatment estimand, WOMAC pain fell 3.96 and 3.74 points versus 2.12 on placebo, and body weight fell 20.0 percent and 23.7 percent versus 4.6 percent. Source: Schnitzer TJ, Hunter DJ, Conaghan PG et al., “Retatrutide Once Weekly in Adults with Knee Osteoarthritis Pain and Obesity,” New England Journal of Medicine, published 8 November 2026, DOI: 10.1056/NEJMoa2610676.
- TRIUMPH-1-OA randomised 574 participants to 4 mg, 9 mg, 12 mg or placebo to 80 weeks; WOMAC pain fell 3.22, 3.48 and 3.64 points respectively versus 1.87 on placebo. Same paper.
- Efficacy-estimand results reported at ACR Convergence 2026 gave weight reductions of 26.4 percent (9 mg) and 28.7 percent (12 mg) versus 2.1 percent on placebo, with WOMAC pain falls of 4.48 and 4.41 versus 2.43. Source: Conaghan PG, ACR Convergence 2026 plenary, Orlando, 8 November 2026; Medscape Medical News coverage, “Retatrutide Reduces Knee Osteoarthritis Pain and Weight, Improves Physical Function,” November 2026.
- Effect sizes for pain were Cohen’s d 0.55 to 0.79 across both trials, which the authors describe as possibly better than existing approved symptom-controlling interventions. Baseline BMI was 40.4 and 64 percent of TRIUMPH-4 participants had Kellgren-Lawrence grade 3 disease. Same NEJM paper and ACR abstract.
- Serious adverse events were more common with retatrutide than placebo in both trials, with gastrointestinal events most frequent. Same NEJM paper.
- Trial registrations: TRIUMPH-4 NCT05931367, TRIUMPH-1-OA NCT05929066.
What to watch
Three things. First, whether Lilly files for a knee osteoarthritis indication specifically, which would create a prescribing pathway that has nothing to do with weight targets. Second, whether the TGA and EMA accept the same estimand logic, since the gap between 20 and 28.7 percent is exactly the kind of thing regulators argue about. Third, and most relevant here, whether the pain effect holds in people who are not losing 25 percent of their body weight, because if it only works via mass loss then the drug is a weight drug with an arthritis side benefit, not an arthritis drug.
For now: this is the strongest evidence yet that a peptide can change osteoarthritis pain, not just the number on the scale. That is genuinely new, and it deserves better than a headline.
We are tracking the OA indication, the FDA review, and the extension data in the Grey Highway Telegram group. Mechanism background is on the retatrutide compound page.
Disclaimer: This article is for educational and informational purposes only. It does not constitute medical advice, therapeutic recommendations, or endorsements of any compound. Grey Highway is a research-education community. We do not sell, supply, or promote the use of research compounds. Always consult a qualified healthcare professional regarding health decisions. For Australian regulatory information, visit the TGA website.