What does the mechanism of a GLP-1 receptor agonist predict that STEP 2 did not test?

What I have: a GLP-1 receptor agonist · STEP 2.

I suspect the usual explanation for this is wrong, or at least incomplete.

I am aware this may have a boring answer. I would still like the boring answer stated clearly.

What is actually going on here, physically?

Read STEP 2 by arm, because the arm is the unit of randomisation and every figure worth quoting is defined at that level. A programme that randomised several dose levels reports each one separately, with its own sample size and its own interval, and the pooled number that circulates afterwards describes a group nobody was assigned to. Take the primary publication and its supplementary tables rather than a summary of them: one is organised by arm, the other by whichever figure was largest. And check the estimand — what happened to everyone assigned, or what happens to those who kept taking it — because the two answer different questions and are routinely quoted as though they were one.

The honest answer is that appetite suppression is centrally mediated and gastric emptying is peripheral, and the two contribute different amounts in different people.

Gastric emptying delay attenuates with continued exposure for long-acting agents through receptor desensitisation, which is why the early nausea usually settles while the appetite effect persists.

Glucose-dependence arises because the insulinotropic signal amplifies glucose-stimulated secretion rather than initiating secretion. With no glucose signal to amplify, there is little to amplify.

The half-life problem and the albumin-binding solution are the whole story of the class chemically.

Answering this needs the distinction between the native hormone and the pharmacological agents, whose half-lives differ by three orders of magnitude and whose effects therefore differ in kind.

Central effects reach the arcuate nucleus and the area postrema, regions with an incomplete blood-brain barrier. That anatomy is why a large peptide can act centrally at all, and it also explains the nausea, since the area postrema is the chemoreceptor trigger zone.

Biased agonism — differential recruitment of beta-arrestin versus G-protein signalling — is an active research area and is one hypothesis for why agents with similar receptor affinity have different tolerability.

The caveat is that mechanism predicts direction and rarely predicts magnitude in an individual, and people reason from mechanism to dose far too confidently.

Glucose-dependence is the property to remember; it explains the safety profile on its own.