Retatrutide: what the new phase 3 results really show

Information verified on 6 September 2026. The US approvals cited do not prejudge the status in Morocco. This information sheet is not a substitute for medical advice.

The new findings on retatrutide are promising, but they do not constitute marketing authorisation. As of 6 September 2026, the compound remains experimental. To interpret the findings correctly, one must consider the participants, the duration of follow-up, the adverse effects and the nature of the available data.

Why is it referred to as a ‘triple agonist’?

Retatrutide acts on three hormonal receptors: GIP, GLP-1 and glucagon. This mechanism explains the relevance of the research; it does not mean that it is ‘three times as effective’. Counting receptors is not a valid method for comparing the benefits and risks of a medicine.

Two trials, two populations that must not be confused

In its announcement on 23 July 2026, Lilly presented preliminary results from two 80-week Phase 3 trials. The table shows the reported mean values for the group receiving the highest dose studied, based on the efficacy analysis used by the manufacturer. These are not individual endpoints. Source: TRIUMPH-2 and TRIUMPH-3 results announced by Lilly.

Lilly has announced that a US submission is planned for the first quarter of 2027. Detailed results must be presented and published; a distinction must therefore be made between the manufacturer’s press release and a full peer-reviewed scientific article.

What the statistics don’t tell you

A group average does not indicate how much weight a specific individual will lose, nor how they will tolerate the treatment. It may coexist with modest responses, very significant responses and dropouts. To assess the value of a study, also look at the variation in results and the reasons for dropouts.

The efficacy analysis here estimates the outcome in a scenario where participants remain on treatment as intended. An analysis that takes account of treatment interruptions more directly addresses a different question, one that is more closely aligned with real-world practice. Both approaches are useful, provided that the indicator is not changed without notice between two comparisons.

Losing weight and preventing a heart attack are two different outcomes

In TRIUMPH-3, the confidence intervals for the reported cardiovascular analyses include the possibility of no difference. The data presented are therefore insufficient to conclude that retatrutide prevents heart attacks or strokes. An improvement in a risk factor, such as blood pressure, does not replace evidence of a reduction in clinical events.

A confidence interval reflects the uncertainty surrounding an estimate. When there are few events, this uncertainty can be significant. A headline about weight loss does not resolve this statistical problem.

Side effects are part of the outcome

Digestive problems are among the most frequently reported side effects, and some participants discontinued treatment. Discontinuations, serious events and long-term follow-up must be examined with as much attention as body weight. A controlled trial involves inclusion criteria, monitoring and a controlled product: this framework is not present in a bottle purchased on social media.

The The FDA warns against unapproved versions of these medicines. A ‘for research purposes only’ label does not mean that a product is suitable for human consumption.

A template to keep for the next advert

1. Does the result come from a press release, a conference or a full publication?
2. Who took part and for how long?
3. Which comparison tool was used?
4. Are interruptions included in the analysis cited?
5. What clinical benefits have actually been proven?
6. Is there authorisation in the country concerned?

Further reading: Warning signs to look out for with injections sold online and The Mounjaro case and cardiovascular risk.

Original illustration generated by AI; it does not represent an identified commercial product or a real clinical situation.