The exact numbers are published, and almost nobody prints them. The FDA label for semaglutide contains three tables of adverse-reaction incidence — by reaction, by dose, and against placebo. This page reproduces them in full, because a list of side effects without a percentage next to it tells you nothing about whether to expect one.
The single most useful column is the one competitors omit: the placebo rate. Sixteen percent of people who were injecting saline reported nausea.
Common adverse reactions: semaglutide 2.4 mg versus placebo
From the pooled weight-reduction trials — 2,116 adults on Wegovy 2.4 mg weekly for up to 68 weeks, against 1,261 on placebo. The label lists every reaction reported in at least 2% of treated patients and more often than placebo. All of them are below.
Wegovy label, Table 3. Placebo n=1,261, semaglutide 2.4 mg n=2,116.
Read as a whole rather than row by row, the label reports that 73% of semaglutide-treated patients and 47% of placebo-treated patients had a gastrointestinal adverse reaction of some kind. The gap — not the 73% — is the drug.
Discontinuation follows the same shape: 6.8% of semaglutide patients and 3.2% of placebo patients permanently stopped because of an adverse reaction. The reactions behind it were nausea (1.8% versus 0.2%), vomiting (1.2% versus 0%) and diarrhea (0.7% versus 0.1%). Put another way, more than 93 in 100 people did not stop.
The dose-dependence, inside one trial
The label’s fourth table is the more interesting one, because all three arms were randomised within the same two 72-week studies. Nothing here is being compared across trials — placebo, 2.4 mg and 7.2 mg were run side by side.
Wegovy label, Table 4 — Studies 8 and 9. Placebo n=303, 2.4 mg n=304, 7.2 mg n=1,311.
Two things fall out of this table. First, the 2.4 mg column here (nausea 35%) does not match the 2.4 mg column in the table above it (nausea 44%) — different trials, different populations, different baseline BMI. The label opens section 6 by warning that adverse reaction rates from one set of trials cannot be directly compared with another. That warning applies to its own tables.
Second, most reactions rise only modestly from 2.4 mg to 7.2 mg. One does not. Dysesthesia — altered skin sensation, including paresthesia, burning and sensitive skin — goes 0% on placebo, 6% at 2.4 mg, 22% at 7.2 mg. The label states outright that its incidence increased with increasing dosage and drug levels in the blood.
It is also the one reaction with published follow-through. Of 288 patients who had it at 7.2 mg, 2% permanently discontinued, 8% had a temporary interruption and 23% had a dose reduction. 18% did not report recovering during the trial. Of 38 patients who recovered and were then re-escalated to 7.2 mg, 17 — 45% — had it return.
What the placebo column actually means
Every rate in the treated column includes everything that would have happened anyway: background illness, the ordinary noise of living for 68 weeks, and the expectation of side effects in someone who has just been handed an injection pen and a leaflet. The placebo column measures exactly that, and it is not small.
A meta-analysis of 69 placebo-controlled obesity drug trials covering 20,454 patients found that among the placebo-treated participants, 73.7% reported an adverse event, 3.4% reported a serious adverse event and 5.2% discontinued because of one. Anxiety was reported by 2.7% and depression by 2.5%. These are people receiving no drug at all.
This is not an argument that semaglutide’s side effects are imaginary — nausea at 44% versus 16% is a real 28-point difference, and the gradient across doses is real too. It is an argument for reading the difference rather than the headline number, and for treating any page that lists side effects without a comparison arm as uninformative.
The boxed warning
Semaglutide carries a boxed warning for thyroid C-cell tumours. Stated exactly as the label states it: in mice and rats, semaglutide caused a dose-dependent and treatment-duration-dependent increase in the incidence of thyroid C-cell tumours — adenomas and carcinomas — after lifetime exposure at clinically relevant plasma exposures. It is unknown whether semaglutide causes thyroid C-cell tumours, including medullary thyroid carcinoma (MTC), in humans, because the human relevance of the rodent finding has not been determined.
Cases of MTC have been reported after approval in patients treated with liraglutide, another GLP-1 receptor agonist; the label says the data in those reports are insufficient to establish or exclude a causal relationship in humans.
Two consequences are concrete. Semaglutide is contraindicated in anyone with a personal or family history of MTC, or with Multiple Endocrine Neoplasia syndrome type 2. And the label states that routine monitoring of serum calcitonin or thyroid ultrasound is of uncertain value for early detection, because such monitoring may increase the risk of unnecessary procedures — low test specificity against a high background rate of thyroid disease.
Serious, and uncommon
The label’s Warnings and Precautions section lists ten concerns. They are a different category from the table above: mostly rare, occasionally severe, and the reason the label exists rather than a leaflet. Where a rate is published, here it is.
Wegovy label, sections 5 and 6.1. The last two rows are rates per 100 patient-years, not percentages — the label reports them that way because the raw counts are single digits.
Three more carry no single number and are worth reading in words:
- Heart rate. Mean increases of 1 to 4 beats per minute. More useful is the distribution: a maximum change from baseline of 20 bpm or more at any visit occurred in 26% of treated adults versus 16% on placebo. In patients aged 12 and over with a normal baseline, it was 54% versus 39%.
- Pulmonary aspiration under anaesthesia. Semaglutide delays gastric emptying, and there have been rare post-marketing reports of aspiration in patients undergoing procedures requiring general anaesthesia or deep sedation who had residual gastric contents despite reported adherence to preoperative fasting. The label states that available data are insufficient to inform recommendations on how to mitigate this — including whether modifying fasting or pausing the drug would help. It instructs patients to tell their healthcare providers before any planned procedure.
- Hypersensitivity. Anaphylaxis and angioedema have been reported. Semaglutide is contraindicated after a prior serious hypersensitivity reaction to it or to any excipient.
Ozempic’s numbers are lower — and not comparable
Ozempic and Wegovy are the same molecule at different dose ceilings for different indications, so the two labels report different rates. The Ozempic placebo-controlled pool in type 2 diabetes looks like this:
Ozempic label, Table 1 — reactions in at least 5% of treated patients. Placebo n=262, 0.5 mg n=260, 1 mg n=261.
Overall gastrointestinal reactions in that pool: 15.3% on placebo, 32.7% at 0.5 mg, 36.4% at 1 mg, with 3.1% and 3.8% discontinuing versus 0.4%. Lower than Wegovy’s figures — but these are different patients at lower doses in trials designed around blood glucose rather than weight, and the dose-response inside the Wegovy data suggests dose explains much of the difference on its own.
One reaction genuinely belongs to the diabetes population rather than the drug: hypoglycaemia. Semaglutide lowers blood glucose, and the label notes the risk rises when it is combined with insulin or an insulin secretagogue such as a sulfonylurea. Documented symptomatic hypoglycaemia occurred in 17.3% and 24.4% of patients on 0.5 mg and 1 mg when coadministered with a sulfonylurea.
Where these numbers come from, and where yours goes
Every figure above is from a randomised trial run to support an approval, which is both the strength and the limit. Rates observed under trial conditions may not reflect rates in practice, as the label itself says in its opening sentence of section 6. Post-marketing reports are what fill the gap, and they are how ileus, intestinal obstruction, necrotizing pancreatitis and pulmonary aspiration reached the label after approval.
Those reports come from patients and clinicians. Side effects can be reported to the FDA directly through MedWatch, or by phone on 1-800-FDA-1088 — the route is printed on the label itself.
Related
- Does semaglutide work? — the same label, read for the efficacy tables instead of the safety ones.
- How long a GLP-1 stays in your system — relevant if you are stopping: exposure falls over weeks, not days.
- How long semaglutide takes to work — the titration schedule the label says exists to reduce these reactions.
- Semaglutide dosage calculator