GLP-1 Trial Endpoints Explained: Primary, Secondary and MACE

A GLP-1 clinical trial endpoint is the specific outcome researchers commit, before the study starts, to measuring and judging the drug on. The primary endpoint is the main question the trial was built and sized to answer; secondary endpoints are extra questions; and a composite endpoint like MACE bundles several events into one count.

Why care? Because two headlines about the same drug can sound wildly different simply because they are reporting different endpoints. Once you know which endpoint you are looking at, a lot of confusing news gets simpler.

What is a primary endpoint?

Picture a trial as a bet placed in advance. The primary endpoint is what the researchers bet on. They choose it before anyone gets a dose, register it, and size the trial (how many people, how long) so that it can answer that question reliably.

Real examples:

Trial	Medicine	Primary endpoint(s)
STEP 1	Semaglutide 2.4 mg	Two co-primary: percent change in body weight, and the share of people losing at least 5%, at week 68
SURMOUNT-1	Tirzepatide	Same pair of co-primaries, at week 72
SURMOUNT-OSA	Tirzepatide	Change in the apnea-hypopnea index (breathing interruptions per hour of sleep) at week 52
SELECT	Semaglutide 2.4 mg	Time to first MACE: cardiovascular death, non-fatal heart attack or non-fatal stroke
SURMOUNT-4	Tirzepatide	Percent weight change from week 36 to week 88 after some people were switched to placebo

Notice SURMOUNT-4. Its question was not "does tirzepatide cause weight loss?" but "what happens if you keep taking it versus stop?" Same drug, completely different primary endpoint, completely different headline.

What is a composite endpoint like MACE?

MACE stands for major adverse cardiovascular events. In SELECT it meant whichever happened first: death from a cardiovascular cause, a non-fatal heart attack or a non-fatal stroke.

Researchers combine events because each one alone is relatively rare. Counting them together means more events, which means a trial can reach a clear answer with fewer people or less time. SELECT still needed 17,604 participants and a mean follow-up of 39.8 months.

The catch: a composite can be driven mostly by one of its parts. So after you see the overall result, look at the components. From the Wegovy prescribing information's table of SELECT results:

Outcome	Placebo	Wegovy	Hazard ratio (95% CI)
MACE (primary composite)	8.0%	6.5%	0.80 (0.72 to 0.90)
Cardiovascular death	3.0%	2.5%	0.85 (0.71 to 1.01)
Fatal or non-fatal heart attack	3.8%	2.8%	0.72 (0.61 to 0.85)
Fatal or non-fatal stroke	2.0%	1.8%	0.89 (0.72 to 1.11)

The composite clearly improved. The individual pieces did not all move by the same amount, and some of their uncertainty ranges include "no difference." That is normal for components, and it is exactly why the composite was chosen as the primary endpoint. (For the full story of SELECT itself, see Wegovy and heart health. For how to read those ranges, see confidence intervals, explained.)

What are secondary endpoints, and why is there a "testing hierarchy"?

Secondary endpoints are the trial's additional questions: waist size, blood pressure, quality of life, specific events. Here is the subtle part. The more questions you test, the more likely one comes up "significant" by pure luck. To prevent that, good trials set a testing hierarchy: they check confirmatory secondary endpoints in a fixed order, and stop formally claiming wins at the first one that misses.

SELECT shows this in action. The Wegovy label notes that cardiovascular death was the first confirmatory secondary endpoint in the hierarchy and "superiority was not confirmed." So all-cause death, next in line, is described as not statistically significant under the prespecified hierarchy, even though its own interval (0.81, 0.71 to 0.93) looks favorable. Heart failure events were outside the hierarchy, and the label states plainly that an effect on heart failure has not been established.

That is not researchers being fussy. It is the rulebook that keeps a 20-outcome trial from producing 20 headlines.

What is a surrogate endpoint?

A surrogate is a measurement that stands in for an outcome that takes years to show up. Blood sugar, liver biopsy findings and blood pressure are common surrogates.

Wegovy's MASH (fatty liver disease with inflammation) indication is a live example. The label says it was granted accelerated approval based on improvement in MASH and liver fibrosis seen on biopsy, and that continued approval may depend on a confirmatory trial showing clinical benefit. Translation: the liver tissue looked better; whether that translates into fewer cases of cirrhosis or liver failure is still being confirmed. If that topic is yours, read our semaglutide and fatty liver guide.

How do I use endpoints when I read GLP-1 news?

Before reacting to a number, write down four things:

1. Which endpoint is this? Primary, secondary, exploratory, or a later "post hoc" analysis?
2. Exactly how was it measured? Percent weight change, kilograms, events per hour, time to first event?
3. When? Week 52, 68, 72, or a median of three years?
4. In whom? Adults without diabetes? People with heart disease?

Then the classic headline traps become obvious. "Drug cut heart deaths!" might be a secondary endpoint that missed formal significance. "Study finds 25% weight loss!" might be from an 88-week figure that included a lead-in period, like SURMOUNT-4's overall result. Neither is false; both need the endpoint attached. Our five-minute headline method puts this into a routine.

Why your personal "endpoints" are different

Trials pick one or two endpoints so they can give a clean yes or no. Your life does not work like that. You might care about energy, blood pressure, knee pain, sleep and a dress size all at once. Tracking a few of those over time, for example weight, waist, symptoms and a quick daily check-in in GLPily, gives you and your clinician your own multi-endpoint picture, without pretending it is a trial.

Common questions

What is the difference between a primary and a secondary endpoint?

The primary endpoint is the main pre-registered question the trial was designed and sized to answer. Secondary endpoints are additional pre-planned questions; they add context but carry less weight on their own.

What does MACE mean in GLP-1 studies?

Major adverse cardiovascular events. In SELECT and the Wegovy label it means cardiovascular death, non-fatal heart attack or non-fatal stroke, counted as one combined endpoint.

Is weight loss a surrogate endpoint?

For heart or kidney outcomes, yes: weight and blood markers are stand-ins. That is why outcome trials such as SELECT, which count real events, matter separately from weight trials.

Where can I see a trial's planned endpoints?

On its ClinicalTrials.gov record under outcome measures, and in the methods section of the published paper.