Imputed-placebo analysis estimates a 28% MACE reduction for tirzepatide versus no drug

Researchers have published an indirect analysis estimating how much tirzepatide (sold as Mounjaro for type 2 diabetes and Zepbound for weight management) would reduce cardiovascular events compared with placebo — a comparison the original trial never actually made. The estimated hazard ratio for major adverse cardiovascular events was 0.72, and 0.61 for death from any cause ^[1].

The work appears in Diabetes Care, published by the American Diabetes Association, and was funded by Eli Lilly and Company, which makes tirzepatide ^[1].

How the estimate was built

SURPASS-CVOT, the cardiovascular outcomes trial for tirzepatide, did not include a placebo group. Instead, it compared tirzepatide against dulaglutide, another GLP-1 receptor agonist that already had proven cardiovascular benefit. That design answers the question "is tirzepatide at least as good as an active drug?" but leaves open the question of how tirzepatide compares with no drug at all ^[1].

To estimate the missing comparison, the authors chained two trials together. They took the hazard ratio for three-point major adverse cardiovascular events (MACE-3) between tirzepatide and dulaglutide from SURPASS-CVOT, and multiplied it by the hazard ratio for dulaglutide versus placebo from REWIND, the earlier dulaglutide outcomes trial ^[1].

Because the two trials enrolled different kinds of patients, the researchers narrowed the REWIND group to participants who would have qualified for SURPASS-CVOT. That left 2,055 of REWIND's 9,901 participants, alongside all 13,165 participants from SURPASS-CVOT. Propensity score methods were used to adjust for remaining differences in patient characteristics between the two studies ^[1].

The numbers

In the main indirect comparison, tirzepatide versus an imputed placebo was associated with:

• Lower MACE-3: hazard ratio 0.72 (95% confidence interval 0.55 to 0.94) ^[1]
• Lower death from cardiovascular causes or heart failure events: hazard ratio 0.70 (95% CI 0.51 to 0.96) ^[1]
• Lower all-cause death: hazard ratio 0.61 (95% CI 0.45 to 0.82) ^[1]

A hazard ratio of 0.72 corresponds to roughly a 28% lower rate of those events. All three confidence intervals sit entirely below 1.0, which in a randomized trial would be read as a statistically meaningful difference — though this was not a randomized comparison ^[1].

The authors ran several sensitivity analyses: unadjusted versions, versions using the entire REWIND population rather than the matched subset, and post hoc analyses drawing on a recent meta-analysis of GLP-1 receptor agonists that included REWIND. Results were described as "generally consistent" ^[1].

The important caveat

The authors themselves call this an "indirect prespecified exploratory comparison" ^[1]. Multiplying hazard ratios from two separate trials assumes that the placebo-controlled effect measured in REWIND would hold in the SURPASS-CVOT population, under SURPASS-CVOT's conditions and era of background care. That assumption cannot be tested directly.

The abstract does not report absolute event rates, follow-up duration, or the number of events in either trial, so it is not possible from this source to say how many events per 100 patients the estimate translates to ^[1]. It also does not report results for people without type 2 diabetes; the conclusion is limited to participants with type 2 diabetes and established atherosclerotic cardiovascular disease ^[1].

Why it matters for patients

Many people taking tirzepatide want to know whether the drug protects the heart, not just whether it lowers blood sugar or body weight. SURPASS-CVOT's active-comparator design makes that harder to answer in plain terms, because "as good as or better than dulaglutide" is not the same statement as "reduces heart attacks compared with nothing."

This analysis is an attempt to fill that gap with statistics rather than a new trial. The estimated size of the benefit is substantial, but the strength of evidence behind an imputed-placebo calculation is lower than behind a direct randomized comparison. Regulators and guideline committees generally weight those two kinds of evidence differently.

What happens next

Whether the U.S. Food and Drug Administration will accept this kind of indirect estimate to support cardiovascular claims in tirzepatide's labeling is not addressed in the source and is not yet known ^[1]. No new randomized placebo-controlled cardiovascular trial of tirzepatide is described in this publication.