Truveta records show tirzepatide outperforming semaglutide outside of trials

A new retrospective study using de-identified US electronic health records found that adults with obesity who stayed on tirzepatide lost more weight over six months than those who stayed on semaglutide — a gap of about 2.3 percentage points after statistical adjustment ^[1].

The analysis drew on Truveta health system records covering patients who started one of the two drugs between December 2023 and June 2024. To be included, patients had to have obesity, or overweight plus at least one obesity-related complication, and they could not have diabetes. Researchers then followed those who adhered to treatment for six months ^[1].

What the numbers show

Among 2,396 patients in the on-treatment group — 1,003 on tirzepatide and 1,393 on semaglutide — average six-month weight reduction was 11.15% with tirzepatide versus 8.83% with semaglutide. The adjusted difference was 2.32 percentage points in favor of tirzepatide, with a 95% confidence interval running from 1.48 to 3.17 percentage points ^[1]. Because that range does not cross zero, the difference is unlikely to be a statistical fluke in this dataset.

Tirzepatide patients also hit weight-loss milestones more often. Higher proportions reached the 5%, 10%, 15% and 20% thresholds compared with semaglutide patients ^[1]. The study also reported greater reductions in body mass index, blood pressure and hemoglobin A1c — a measure of average blood sugar — among tirzepatide users ^[1].

One detail stands out: more semaglutide patients were taking higher doses. In the tirzepatide group, 42.4% were on 10 mg or more, while 67.7% of semaglutide patients were on 1.7 mg or more ^[1]. The authors describe this as an "early emergence of tirzepatide's comparative advantage" that showed up even though the semaglutide group had, on average, climbed further up the dose ladder ^[1].

The researchers used propensity-score weighted regression as the main analysis method, a technique meant to make the two groups look more similar on measured characteristics. A sensitivity analysis using a modified intention-to-treat approach — which is less dependent on patients staying on the drug — produced consistent findings ^[1].

The funding and design caveats

This is not a randomized trial. Patients and their clinicians chose which drug to use, and statistical adjustment can only account for factors that were recorded in the medical record. Unmeasured differences between the groups could still influence the result.

The study was funded by Eli Lilly and Company, which makes tirzepatide (sold as Mounjaro for type 2 diabetes and Zepbound for weight management). Several authors are Lilly employees who hold Lilly stock or stock options, and others are employees of Analysis Group, a consulting firm paid by Lilly for the research ^[1]. One academic author reports consulting fees from both Novo Nordisk and Eli Lilly, among many other companies ^[1].

The authors frame the findings as consistent with SURMOUNT-5, the randomized head-to-head trial that also showed greater weight reduction with tirzepatide than semaglutide in adults with obesity who did not have diabetes ^[1].

Why it matters for patients

Randomized trials tell you what a drug can do under carefully controlled conditions, with structured dose escalation and close follow-up. Real-world records show what happens in ordinary clinics, where insurance coverage, supply, side effects and cost all shape which dose people end up on. This analysis suggests the trial-level gap between the two drugs also appears in routine US practice, at least in the first six months and at least among people who keep taking the medication ^[1].

Some important things are not answered here. The abstract does not report side effects, discontinuation rates, or how many patients who started treatment were excluded for not adhering — which means the results describe people who stayed on therapy, not everyone who tried it ^[1]. It also does not break results out by brand name or by insurance status, and it covers only six months, so longer-term differences are unknown from this study.

Weight-loss averages also hide wide individual variation. A 2.3-percentage-point difference in group averages does not predict what any single person will experience on either drug.

What happens next

The study period ended after six months of follow-up for patients who started treatment between December 2023 and June 2024 ^[1]. Longer real-world follow-up, and data on tolerability and discontinuation, would be needed to judge how durable this gap is. Those results are not yet available from this analysis.