What the GLP-1 Head-to-Head Evidence Actually Compared

Yes, there is a direct weight-loss comparison

SURMOUNT-5 randomly assigned 751 adults without type 2 diabetes to tirzepatide or semaglutide injections. At 72 weeks, average weight reduction was 20.2% with tirzepatide and 13.7% with semaglutide. This was a direct comparison within one trial, not a calculation using unrelated studies.

The trial was published in the New England Journal of Medicine in 2025. Its primary outcome was percentage change in body weight. That makes it directly relevant to the weight-loss question, unlike a trial designed primarily to measure blood sugar.

Correction: an earlier version of this article incorrectly inferred from the content of selected product labels that a direct weight-loss trial did not exist. Published trial evidence must be considered separately from what a particular label includes.

What was actually tested

The phase 3b trial compared maximum tolerated doses of the two injectable medicines over 72 weeks. The studied maintenance doses were tirzepatide 10 or 15 mg and semaglutide 1.7 or 2.4 mg. These describe the trial arms, not a dosing recommendation or a schedule to follow.

Participants and investigators knew which medicine was assigned: the study was open-label. Random assignment makes the treatment groups more comparable at the outset, but open-label treatment remains a limitation when interpreting experience, adherence and reported symptoms. The trial was funded by Eli Lilly, the manufacturer of tirzepatide.

Read the difference in percentage points

The difference between 20.2% and 13.7% is 6.5 percentage points. Calling that 6.5% more weight loss can be confusing because a relative percentage uses a different denominator. The clearest comparison gives both trial averages and the percentage-point gap.

These are group estimates, not promises about an individual. A confidence interval describes uncertainty around an estimated group mean; it is not the range of results that every participant experienced. Someone choosing treatment still needs to consider suitability, tolerability and the ability to continue care.

Who the result most directly describes

The registered study included adults with obesity, or overweight with a weight-related condition, and excluded people with type 1 or type 2 diabetes. Its result should not be presented as a measured effect in children, people with diabetes or every person who might seek a prescription.

The sponsor’s eligibility summary also lists exclusions such as prior or planned obesity surgery, severe kidney impairment and pancreatitis. Those criteria define this study population; they are not a substitute for the contraindications and precautions in a medicine’s prescribing information.

The study compared medicines under a trial protocol. It did not randomly assign participants to telehealth companies, membership packages or dispensing pharmacies. Its results therefore cannot establish that one commercial provider delivers more weight loss than another.

What the study does not establish

The comparison was between specific injectable treatments at the studied doses. It does not by itself establish the same difference for oral tablets, drops, compounded preparations, other doses or a different treatment duration. Sharing a molecule name is not enough to transfer every trial result to a different product.

The primary finding concerns weight. It is not proof that one treatment is superior for preventing heart attacks, treating every obesity-related condition or reducing every serious adverse event. Those questions require evidence designed to measure those outcomes in the relevant population.

The most common adverse events in both trial groups were gastrointestinal; most were mild or moderate and occurred mainly during dose escalation. A favorable average weight result does not remove tolerability concerns or replace an individualized safety assessment.

Why comparisons between separate trials still need care

SURMOUNT-5 answers a particular direct-comparison question. Other comparisons may still place results from separate trials next to each other. Differences in enrollment criteria, treatment duration, background care and the handling of missing measurements can affect those numbers.

When reading a comparison, look for the trial name, the population, the formulation and dose studied, the primary outcome and the time point. A review that provides only a percentage and a brand name leaves out information needed to judge whether the comparison applies.

An indirect analysis can be useful when it describes its assumptions and uncertainty. It should be identified as indirect rather than presented as if patients received the competing treatments within the same randomized study.

How to use this evidence when choosing care

The direct evidence supports greater average weight reduction with tirzepatide than semaglutide in the SURMOUNT-5 setting. It does not settle every treatment decision. The appropriate medicine also depends on the person’s medical history, the approved uses relevant to them, adverse effects and practical access.

Ask a clinician whether the study population and tested treatment are relevant to your situation. When evaluating a provider, separately check the actual medicine offered, who prescribes it, how follow-up works and what ongoing care costs. A strong drug trial does not verify a provider’s service quality.