Every now and then a press release lands that deserves a health warning on its headline statistic. Lilly’s October 1 announcement about orforglipron - now branded Foundayo - is one of them. The claim: at 17.2 mg, the oral GLP-1 was associated with a 57% reduction in predicted 10-year type 2 diabetes risk and an 18% reduction in predicted cardiovascular disease risk versus placebo.
Those are big numbers. Read the word “predicted” carefully, because the entire finding rests on it.
What the Research Says
The analyses are post-hoc modelling work built on ATTAIN-1, the Phase 3 trial in adults with obesity or overweight plus a weight-related comorbidity (NCT05869903, 3,127 participants, 72 weeks). The models were published in Diabetes, Obesity and Metabolism. The Lilly release lays out the full table:
- Predicted 10-year type 2 diabetes risk fell an estimated 45%, 50% and 57% versus placebo at the 5.5 mg, 9 mg and 17.2 mg doses (HR 0.55, 0.50, 0.43)
- Predicted 10-year cardiovascular disease risk fell an estimated 13%, 15% and 18% at the same doses (HR 0.87, 0.85, 0.82)
- The underlying drivers are real measured improvements: blood sugar, blood pressure and cholesterol all moved in the right direction, consistent with the weight loss already reported from ATTAIN-1
What did not happen: nobody in this trial was followed for ten years, and nobody was counted for hard endpoints like heart attacks or new diabetes diagnoses. The risk numbers are algorithm outputs - established risk calculators fed the trial’s measured intermediate markers and projected forward.
Why the Distinction Matters
Risk modelling is legitimate science - cardiologists use these calculators all the time. But there is a well-documented gap between “improved risk markers” and “fewer actual events,” and the GLP-1 class has lived through this exact question before. Semaglutide needed a dedicated cardiovascular outcomes trial (SELECT) to convert improved markers into demonstrated event reductions. Tirzepatide needed SURPASS-CVOT territory. Those trials take years and tens of thousands of participants because markers and outcomes do not always move together.
Lilly is not hiding any of this - the release says “associated with significant reductions in predicted long-term risk” and names the modelling approach. But the way the 57% and 18% figures travel through media coverage, they tend to arrive at their destination without the word “predicted” attached. That is worth calling out, because this is exactly the kind of statistic that ends up quoted in clinic marketing and on social media as if it were an outcomes result.
The honest framing: ATTAIN-1 measured real intermediate improvements in a real 72-week trial. The ten-year projections are an inference, and the class-wide history says the inference is usually but not always right.
The Bigger Orforglipron Picture
The reason Lilly is running the modelling now is commercial positioning. Orforglipron is the pill that could scale GLP-1 access - a small molecule, no refrigeration, no injection, manufactured at volumes injectable peptides cannot match. If ATTAIN-2 (the type 2 diabetes population) and the eventual outcomes trials land well, the pill becomes the volume story of the class. Adding “predicted risk reduction” language now gives prescribers a preview of the cardiovascular and diabetes-prevention pitch before the dedicated trials report.
For the research community watching this space, the useful watch-item is when Lilly runs an actual outcomes trial for orforglipron. Until then, every risk figure attached to this molecule is a projection.
Our semaglutide overview covers the injectable original and the SELECT outcomes data that set the standard for this kind of claim, and tirzepatide covers the competitive landscape. If you want to dig into the estimand arguments with people who read the methods sections, the Grey Highway Telegram group is the place.
The Fine Print
Post-hoc analyses are exploratory by definition - the trial was not designed or powered for these endpoints. Risk calculators carry assumptions baked in from populations that may not match trial demographics. And the ATTAIN-1 design is worth a note: it was the first Phase 3 obesity trial pairing treatment with a balanced diet and exercise rather than a calorie-restricted protocol, which complicates naive cross-trial comparisons with other GLP-1 studies.
None of this makes the data bad. It makes the data early, and the headline numbers softer than they look.
Sources
- Eli Lilly press release, October 1, 2026: Foundayo associated with reductions in predicted long-term risk of T2D and CVD
- Publication: Diabetes, Obesity and Metabolism (post-hoc analyses of ATTAIN-1, per Lilly release)
- ClinicalTrials.gov: ATTAIN-1, NCT05869903
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.