BariatricNurseD said:

The dose-response is real but shallow at the top.

Pushing back on BariatricNurseD here. The trial means are being read too generously in this thread. STEP populations were selected, supported, and titrated by protocol, and the real-world curves are consistently a few points worse. That difference is not noise, it is what happens when you remove the study infrastructure.

The figures, for anyone assembling their own picture. Worth knowing that the injection site changes very little. Abdomen, thigh and upper arm are bioequivalent for semaglutide, so a site change is not a plausible explanation for a bad week.

Worth separating that from semaglutide, which this thread keeps folding into the same question. They behave differently and the advice does not transfer.

COA_Karl said:

The trial means are being read too generously in this thread.

Adding the part of the answer the thread has not reached. The mechanism that matters here is not stomach emptying, it is central. GLP-1 receptor agonism in the arcuate nucleus stimulates POMC neurons and suppresses AgRP/NPY signalling, which is why the effect is appetite and food salience rather than physical fullness. Delayed gastric emptying largely tachyphylaxes over the first months; the appetite effect does not.

That is the short version; the long version is somebody else's post.

One thing that is still open after VanRx_Mike’s answer:

Whether anyone has held at a sub-maximal dose long term and kept the result, or whether the maintenance data only exists at 2.4mg?

Reporting back.

Update since I posted: I held at 1.7mg for a further twelve weeks and lost another 4kg slowly, which settles it for me. I was escalating because the number was available, not because I had stopped responding.