Tirzepatide: 10-year CVD risk in SURMOUNT-1

A post-hoc analysis of SURMOUNT-1 published on 10 August 2026 in the European Journal of Preventive Cardiology asked a specific question: does three years on tirzepatide change the predicted likelihood of a cardiovascular event in the next decade? For the 962 participants who had both a baseline and at least one follow-up score, the answer was yes, and the direction of change widened over time^[1].

What SURMOUNT-1 was and why it matters here

SURMOUNT-1 was a Phase 3, double-blind, randomised controlled trial in adults with obesity or overweight and prediabetes but without type 2 diabetes. Participants received tirzepatide (5, 10, or 15 mg weekly) or placebo for 176 weeks (approximately three years). The primary efficacy results, published in 2022 by Jastreboff et al.^[2], reported a mean 20.9% body-weight reduction at 72 weeks on the 15 mg dose versus 3.1% on placebo. The post-hoc analysis now released re-examines the same dataset through a cardiovascular-risk lens rather than a weight lens.

How predicted CVD risk was calculated

The authors used the Framingham Heart Study (FHS) equation to compute each participant's predicted 10-year cardiovascular disease risk at baseline and at multiple follow-up time points^[1]. The FHS equation is one of the most widely used models in preventive cardiology; it takes into account age, sex, blood pressure, cholesterol, smoking status, and diabetes status. Because tirzepatide changes several of these inputs (blood pressure, lipids, weight) as it works, any improvement in those parameters flows through to a lower predicted risk score.

To test whether the FHS equation's predictions are reliable enough to draw inferences from, the researchers cross-checked the model against the REWIND trial, which measured actual observed cardiovascular events for dulaglutide versus placebo. The model-derived hazard ratio for dulaglutide in REWIND was 0.83; the observed hazard ratio was 0.85. That consistency supported using the FHS equation to interpret the SURMOUNT-1 findings^[1].

The numbers at week 72 and week 176

At week 72 (roughly 17 months), participants on tirzepatide 15 mg had an absolute risk reduction in predicted 10-year CVD risk of 1.92 percentage points from baseline. In the same period, placebo participants saw their predicted risk rise by 1.10 percentage points. The hazard ratio was 0.73 (p < 0.001)^[1].

At week 176 (three years), the gap had widened further. Tirzepatide-treated participants retained a predicted risk reduction of 1.31 percentage points below their baseline. Placebo participants were 3.84 percentage points above their baseline. The hazard ratio at this point was 0.62 (p < 0.001)^[1].

Why the tirzepatide number narrows from year one to year three

The absolute risk reduction in the tirzepatide arm fell from 1.92% at week 72 to 1.31% at week 176. This does not mean tirzepatide lost its effect. It reflects the trajectory of individual risk factors: the largest changes in weight, blood pressure, and lipids tend to occur in the first year, after which they plateau or partially return even on continued treatment. The direction of effect remained strongly protective at three years; the magnitude reflected the expected pharmacodynamic plateau.

What this analysis can and cannot tell us

The study has four constraints worth noting. First, this is a post-hoc analysis, not a pre-specified primary endpoint. Post-hoc analyses can generate hypotheses but carry a higher risk of chance findings than pre-registered primary outcomes. Second, predicted CVD risk is a modelled output, not a count of actual heart attacks or strokes. The FHS equation estimates probability; it does not observe events. Third, the analysis draws from 962 of the 2,539 randomised SURMOUNT-1 participants, those who had prediabetes, no pre-existing CVD, and at least one follow-up score. That subset may not represent the full trial population^[1].

Fourth, and most consequentially, tirzepatide does not yet have a completed cardiovascular outcomes trial in obesity without diabetes. The paper itself points to SURMOUNT-MMO as the event-based trial designed to answer whether the predicted risk reductions translate into fewer actual major adverse cardiovascular events. Until that trial reports, the finding is supportive rather than conclusive.

Where this fits in the broader tirzepatide evidence base

Tirzepatide already has a separate cardiovascular outcomes signal from the SUMMIT trial, which enrolled adults with heart failure with preserved ejection fraction (HFpEF) and obesity. SUMMIT showed meaningful improvements in functional capacity and hospitalisation rates, a different cardiovascular mechanism from atherosclerotic MACE prevention. The SURMOUNT-1 post-hoc analysis addresses primary prevention in a prediabetic obesity population, a different and larger target group than SUMMIT's HFpEF patients. The two bodies of evidence point in consistent directions but cover different disease presentations^[1].

For readers tracking the incretin class more broadly, the paper is notable because it uses the same Framingham-validated approach applied in earlier GLP-1 drug analyses, making the tirzepatide findings comparable to prior risk-score work on semaglutide and dulaglutide. Full regulatory status, including the ongoing status of cardiovascular outcomes trial requirements by territory, is on the tirzepatide page.

What reviewers should watch for when SURMOUNT-MMO reports

The key question is whether the 38% relative hazard reduction in predicted risk at three years (HR 0.62) corresponds to a meaningful observed MACE reduction in an event-based trial. In the REWIND cross-validation, the FHS model was within two percentage points of the observed hazard ratio for dulaglutide. If the same accuracy holds for tirzepatide in SURMOUNT-MMO, the post-hoc analysis suggests a potential 35-40% relative risk reduction in cardiovascular events for primary prevention in obesity with prediabetes. That would be among the largest effect sizes reported for any preventive pharmacotherapy in this population^[1].

This analysis does not constitute clinical guidance. The findings inform the research trajectory. Decisions about starting, continuing, or stopping any cardiovascular or metabolic treatment belong with a clinician who knows the individual's history.