Accepted answer
It is real, the mechanism is tachyphylaxis of a specific brainstem pathway, and there is randomised within-trial evidence rather than only anecdote. Your framing is slightly off in one respect, though, and correcting it makes the rest coherent: nausea is not only rate dependent. It is dose dependent in the acute setting and strongly rate dependent in the chronic setting, and clinical practice happens to live almost entirely in the second regime.
The mechanism
The emetic effect is mediated substantially through GLP-1 receptors in the area postrema, a circumventricular structure in the dorsal medulla that lacks a tight blood-brain barrier because its function is to sample circulating blood for toxins. It is the chemoreceptor trigger zone. Circulating peptide reaches it directly, without needing to cross anything.
That pathway adapts. Continued stimulation at a constant level produces a declining emetic response over days to weeks, through some combination of receptor desensitisation, downstream signalling adaptation and central habituation. The satiety and weight effects, mediated substantially through the hypothalamic arcuate nucleus and nucleus tractus solitarius, adapt much less. That dissociation is the whole basis of the therapeutic strategy: you are exploiting a difference in adaptation rate between two effects of one molecule.
So the stimulus that produces nausea is not absolute receptor occupancy. It is occupancy in excess of what the pathway has currently adapted to. Formally, the symptom tracks something much closer to the derivative of exposure with respect to time than to exposure itself, once you are past the first few weeks.
Why this explains your 5 mg versus 12.5 mg observation
At week two on 5 mg you had gone from zero adaptation to an exposure roughly twice your previous steady state, having only recently arrived from 2.5 mg, which itself came from nothing. At 12.5 mg you have been continuously exposed for months and the pathway has adapted to something close to your current level. The increment from 10 mg to 12.5 mg is 25%. The increment from 2.5 mg to 5 mg is 100%.
Work the ladder in relative terms, which is how the pathway experiences it, rather than in milligrams:
- 2.5 → 5 mg: +100%
- 5 → 7.5 mg: +50%
- 7.5 → 10 mg: +33%
- 10 → 12.5 mg: +25%
- 12.5 → 15 mg: +20%
The steps get larger in milligrams and smaller in the currency that matters. Same on the semaglutide ladder: 0.25 → 0.5 is +100%, 1.0 → 1.7 is +70%, 1.7 → 2.4 is +41%. This is why the early steps are consistently the worst despite involving the least drug, and why the top of the ladder surprises people by being comfortable.
The evidence that is not just "dose steps feel bad"
- Non-monotonic incidence across final doses within a single trial. SURMOUNT-1 reported nausea in 24.6%, 33.3% and 31.0% of the 5, 10 and 15 mg arms against 9.5% on placebo [1]. A purely dose-dependent effect cannot produce a flat or falling incidence at the top of the range in arms that were randomised and otherwise identical. What is identical between those arms is the lower half of the escalation ladder.
- Randomised comparison of escalation speeds to the same final dose. The retatrutide phase 2 trial included arms that arrived at the same maintenance dose via different escalation schedules, and the gastrointestinal event rates differed by schedule [2]. Same molecule, same destination, different rate, different tolerability. This is the closest thing to a clean experiment that exists.
- Survodutide dose-finding. Gastrointestinal events concentrated in the faster-escalation arms, which is why later development used slower ladders [3].
- The existence of the ladder at all. The starting doses are subtherapeutic by design. Semaglutide 0.25 mg produces very little weight effect; its only purpose is to begin the adaptation.
The practical consequences you identified are both correct
Gaps are not neutral. Adaptation decays when exposure falls. How fast is not well characterised in humans, but the reported pattern is that a gap of a couple of weeks or more can return you toward a naive state, and resuming at your previous dose then reproduces escalation-grade symptoms. Product labelling generally advises reinitiating at a lower dose after an extended interruption for exactly this reason. A supply gap, a holiday and a batch you did not trust are the same event pharmacologically.
Time is usually the right response to new-dose nausea. Mild to moderate symptoms following the standard signature typically decline over two to four weeks at the same dose, because that is what adaptation does. Worth understanding: reducing and then re-escalating means crossing the same increment twice, so it is not automatically the gentler option. This is a conversation for whoever prescribes for you rather than something to optimise alone, and severe symptoms are a different category entirely.
Where the rate story breaks down
Two places, and they are worth stating so the rule does not get over-applied:
- Acute high exposure. A large single overexposure produces dose-dependent emesis regardless of adaptation state. Adaptation raises the threshold; it does not remove it. Dosing errors with concentrated material have caused severe protracted vomiting in people fully adapted to a therapeutic dose.
- The very top of the range in some individuals. A minority never adapt adequately and remain symptomatic at high doses indefinitely. For them the effect looks straightforwardly dose dependent, and the correct answer is a lower maintenance dose rather than more patience.
edited 20 May 2025 by tyndall_haze — added the citation requested in comments
Relative increments rather than milligram increments is the reframe that makes the whole ladder make sense. – aine_mulcahy 8 months ago The point that dropping back and re-escalating means crossing the same increment twice is one most people get wrong. – s_bhattacharya 9 months ago 2Adaptation raising the threshold rather than removing it is an important qualifier. Overexposure is still overexposure. – micron22 5 months ago add a comment