Why nausea happens and why it fades

Nausea is the side effect that decides whether people stay on these drugs. It has two sources, and understanding which one is doing what explains why it fades for most people and why it returns with each dose step.

Source one: the area postrema

At the floor of the fourth ventricle sits the area postrema, a small region without a blood-brain barrier whose job is to detect circulating toxins and trigger nausea and vomiting. It is dense with GLP-1 receptors. Zhang and colleagues mapped its cell types in mice and identified excitatory neurons, including GLP-1 receptor-expressing populations, whose activation produced nausea-associated behaviours (PubMed 33278342). Huang and colleagues then showed that GLP-1 receptor neurons in the area postrema drive aversion, while GLP-1 receptor neurons in the neighbouring NTS drive satiety without aversion, and that the two can be manipulated independently (PubMed 38987598).

This is animal work, graded C on this site, but it lines up with what people report: the nausea has a central, motion-sickness-like quality that does not depend on having eaten.

Source two: the stomach

The peripheral contribution comes from delayed gastric emptying. Food sits longer, the stomach distends more, and stretch signals travel up the vagus to the same brainstem region. This source is strongest early because gastric slowing shows rapid tachyphylaxis with continuous exposure (PubMed 21430088), covered on the gastric emptying page. Eating smaller meals and stopping before feeling full acts on this source, which is why that advice appears in every patient leaflet.

The rates

The Wegovy label reports, across its pooled placebo-controlled trials, nausea in 44% of people on semaglutide 2.4 mg versus 16% on placebo, vomiting in 24% versus 6%, diarrhoea in 30% versus 16% and constipation in 24% versus 11%. Most events were mild to moderate and occurred during dose escalation. The Zepbound label reports nausea in 25%, 29% and 28% at 5, 10 and 15 mg versus 8% on placebo.

Those are incidence figures over the whole trial, meaning the share of people who reported the symptom at least once. They are not the share nauseated on any given day. In STEP 1, 4.5% of semaglutide participants stopped treatment because of gastrointestinal events, versus 0.8% on placebo, over 68 weeks (PubMed 33567185): a real cost, but far below the 44% who reported nausea at some point.

Why it fades

Three things happen together. Gastric slowing shows tachyphylaxis within days to weeks, removing most of the peripheral input. The area postrema circuitry appears to adapt to a steady level of stimulation, though the human evidence for this is the clinical time course rather than direct measurement. And exposure stops changing once a dose reaches steady state, about four to five weeks after each step (see pharmacokinetics). Each step up restarts a smaller version of the process; the labels allow a step to be held for longer if needed.

What is debated

Whether GIP receptor activation in tirzepatide dampens GLP-1 nausea in people is unproven. Borner and colleagues showed it in shrews and rats (PubMed 34380697), and the tirzepatide nausea rates are lower than semaglutide 2.4 mg's despite larger weight loss, but the trials differ in design, population and escalation, so a cross-trial comparison cannot settle it. Whether nausea predicts weight loss is a second common claim; STEP 1 found similar weight loss with and without gastrointestinal events, so the answer is no.

Where to go next

• Week-by-week expectations for someone starting out: GLP-1 Starter.
• Side-effect timelines by drug with label rates: Semaglutide Hub and Tirzepatide Hub.
• FormBlends' tirzepatide guide discusses tolerability in plain terms: tirzepatide introduction.