GLP-1 clinical trials: a guide to what the evidence means

Begin with the question, then the number

I find a trial easier to understand once its question fits into one sentence. Was the researcher testing weight change, the first occurrence of a cardiovascular event, symptoms of an existing condition or what happens after treatment stops? Those questions lead to different studies and different decisions. SELECT, for example, studied cardiovascular outcomes in adults with established cardiovascular disease and overweight or obesity who did not have diabetes. A result in that population cannot simply become a claim about every person considering weight treatment. Before reading the percentage, identify the participants, the medicine and regimen, the comparison group and the follow-up period. That short description is the foundation for deciding whether the result speaks to your question.^[1] Read more: How to read GLP-1 research.

A head-to-head comparison answers a specific question

Comparing percentages from unrelated trials creates an attractive shortcut, but the participants and analysis may differ. A direct randomized comparison is a better starting point when the research question is comparative weight change. SURMOUNT-5 compared tirzepatide and semaglutide in adults with obesity without diabetes using the study’s specified regimens. Its result should be read with the treatment duration and the definition of the analysis beside it. Average weight change is useful evidence about the studied groups, not a prediction for an individual or a guarantee that one prescription will suit them better. I would take the result into a clinical conversation about options, tolerability and treatment goals, rather than turn it into a shopping score for providers.^[2] Read more: SURMOUNT-5 explained.

Clinical events and symptom scores need different readings

A study that counts heart attacks, strokes or cardiovascular deaths is asking something different from one that measures a questionnaire score. Both may matter, but neither should be renamed to make the result sound larger. SELECT’s primary endpoint combined cardiovascular death, nonfatal myocardial infarction and nonfatal stroke. The composite result does not mean every individual component fell by the same percentage. The same care applies to heart-failure research: improved symptoms or physical limitations should not automatically become a survival claim. In our study explainers, look for the outcome definition before the conclusion. Ask what changed, how it was measured and whether the paper was designed to establish that particular finding.^[1] Read more: Symptoms versus survival in STEP-HFpEF.

The headline benefit needs its denominator

Relative measures can be useful, but they are hard to interpret without the underlying event rates and the study timeline. In SELECT, the published primary event percentages were 6.5% with semaglutide and 8.0% with placebo, while the hazard ratio was 0.80. The rounded difference is 1.5 percentage points; the hazard ratio also accounts for event timing. These are different descriptions of the same research, not interchangeable calculations. I prefer to see both before deciding what a headline conveys. Neither is an annual personal risk estimate. The question to ask is how the trial’s baseline risk and follow-up compare with the circumstances your clinician is evaluating.^[1] Read more: The SELECT event-rate chart.

Treatment continuation is part of the evidence

A trial endpoint does not answer how a treatment will fit into the next several years. Withdrawal studies help address a separate question: what happens when treatment changes after an initial period of use? The STEP 1 extension followed participants after semaglutide and the structured lifestyle intervention ended, while SURMOUNT-4 randomized participants after a tirzepatide lead-in to continued treatment or placebo. Those designs are not identical. They are useful for discussing maintenance and planning, but a group average does not establish that every person will regain the same amount. I would use the research to prepare questions about reassessment, access and continuity rather than assume a treatment can be started and stopped without a plan.^[3] Read more: Two studies of treatment withdrawal.

Keep the medicine evidence separate from the service choice

Once you understand the research question, the next step may be practical: how will you obtain care, manage follow-up and pay for the prescribed option? That is where provider reviews and tools help. They evaluate the service and its published terms; they cannot transfer a trial’s results to a clinic or certify that a compounded preparation performs like a branded product. FDA-approved products and compounded drugs have different regulatory status. I would first discuss the appropriate treatment pathway with a clinician, then compare the service that supports it. Our review of Found explains its separate medication and care pathways, while the provider checklist helps organize appointment access, pharmacy identity and renewal questions before enrollment.^[5] Read more: Branded and compounded products.

Put the questions to work

Clinical results explain the studied medicine and population. Reviews help compare the separate service and payment questions.

Explore another research topic