TRIUMPH-4: retatrutide for obesity with knee osteoarthritis

TRIUMPH-4 was the first Phase 3 retatrutide trial to report, on 11 December 2025. It tested the two highest doses in people with obesity and painful knee osteoarthritis, and it measured knee pain as well as weight. It also produced the highest drop-out rates and the highest rate of the new skin-sensation side effect seen so far. The peer-reviewed paper has not been published.

What the trial did

Participants had a BMI of 27 or more, knee pain for at least three months, X-ray evidence of moderate osteoarthritis, and no diabetes. People who had had steroid injections into the joint in the previous 90 days were excluded.^[2] Registered as NCT05931367 (Lilly's announcement misprints the number; the registry entry is the one to check).

Everyone on retatrutide began on a low dose stepped up every four weeks to a target of 9 mg or 12 mg.^[1] The trial had two primary outcomes measured at 68 weeks: percentage change in body weight, and change in the WOMAC pain score, a standard 0-to-10 questionnaire for knee and hip osteoarthritis where higher means worse. Average starting weight was 112.7 kg.

Weight loss

Figures at 68 weeks.^[1] TRIUMPH-4 is the one Phase 3 obesity trial where Lilly's announcement gave both estimands in full, so it is a useful place to see how far apart they can be: five to six percentage points here, because so many people stopped. How to read a trial explains the difference.

On the efficacy estimand, 47.7% of the 9 mg group and 58.6% of the 12 mg group lost at least a quarter of their body weight, and 30.5% and 39.4% lost at least 30%, against 1.3% and 0.8% on placebo.^[1] Systolic blood pressure fell by 14.0 mmHg at the top dose, and non-HDL cholesterol, triglycerides and hs-CRP all fell.

Knee pain and function

On the efficacy estimand, WOMAC pain fell by 4.5 points (75.8%) on 9 mg and 4.4 points (74.3%) on 12 mg from a starting average of 6.0, against 2.4 points (40.3%) on placebo. On the treatment-regimen estimand the falls were 4.0 points (67.2%), 3.7 points (62.6%) and 2.1 points (35.1%). WOMAC physical function improved by 4.1 and 4.2 points against 2.1. In a post-hoc analysis, 14.1% of the 9 mg group and 12.0% of the 12 mg group reported no knee pain at all at 68 weeks, against 4.2% on placebo.^[1]

Two cautions. The pain results are large, but notice that the placebo group's pain also fell by 40%. Pain scores in trials improve on their own through expectation, the attention of regular visits, and the small weight loss the placebo group achieved. The drug's own contribution is the gap between the columns, not the whole number. And the percentage-change figures were not planned in advance; Lilly reports them as a post-hoc analysis.^[1]

Side effects and drop-outs

Nausea affected more than four in ten on the top dose, and vomiting one in five, against none on placebo. Decreased appetite was recorded as an adverse event in about a fifth of the retatrutide groups.^[1]

Dysesthesia. Dysesthesia, an altered skin sensation such as tingling, burning or heightened sensitivity, affected 8.8% on 9 mg and 20.9% on 12 mg against 0.7% on placebo. This was the first time it had been reported with retatrutide and it is not a known effect of semaglutide or tirzepatide, which is why it was widely reported as a new safety signal.^[3] Lilly said the events were generally mild and rarely led to people stopping.^[1] Later trials found lower rates: up to 12.5% in TRIUMPH-1.^[4] See the dysesthesia page.

Drop-outs. 12.2% on 9 mg and 18.2% on 12 mg stopped because of side effects, against 4.0% on placebo, the highest rates in the programme. Lilly gave two reasons. The rates were "highly correlated with baseline BMI": among the 84% who started with a BMI of 35 or more they were 8.8% and 12.1% versus 4.8%, so lighter participants were more likely to stop. And some of the discontinuations were for "perceived excessive weight loss", meaning people who felt they had lost enough, or too much, and chose to stop.^[1] Lilly also said that overall discontinuation, for any reason, was similar across the three groups. See discontinuation rates.

What is still unknown

• Full side-effect tables, including heart rate, rhythm, liver and pancreas figures.
• How long dysesthesia lasted, and whether it settled while people stayed on the drug.
• Whether knee pain returns if weight is regained after stopping.