Semaglutide side effects: what trials show

Semaglutide is one of the most-studied prescription medicines in recent history. Before any conversation about side effects, the baseline matters: across the STEP obesity programme, SUSTAIN type-2 diabetes programme, and the SELECT cardiovascular outcomes trial, more than 30,000 patients have been followed in randomised conditions. The adverse-event picture is therefore unusually detailed for a medicine of this age.^[1] This page summarises what that data shows. It is not medical advice, and any decision about starting, adjusting, or stopping semaglutide belongs with a prescribing clinician.

The most common adverse events: gastrointestinal

Across both the STEP and SUSTAIN trial programmes, the dominant adverse-event class is gastrointestinal. The Wegovy SmPC lists the following as occurring in more than 1 in 10 patients on the 2.4 mg dose^[1]: nausea (44% on semaglutide vs 16% on placebo), diarrhoea (30% vs 16%), vomiting (24% vs 6%), constipation (24% vs 11%), and abdominal pain. These events are dose-dependent and peak during the 16-week titration phase, when the weekly dose is increased stepwise from 0.25 mg to 2.4 mg. Most cases in the trials were rated mild to moderate and resolved without stopping treatment.

In STEP-1, gastrointestinal events caused 4.5% of semaglutide participants to discontinue treatment, compared with 0.8% on placebo.^[2] The SELECT cardiovascular trial, which ran for a mean of 40 months, saw 16.6% of semaglutide participants discontinue for adverse events versus 8.2% on placebo^[3], with gastrointestinal events again the primary driver. Longer duration of treatment increases the cumulative opportunity for GI intolerance, even when each individual episode is transient.

Gallbladder and pancreas: less common, clinically meaningful

The Wegovy prescribing information^[1] flags two organ-specific signals beyond the GI tract. Gallbladder disease: cholelithiasis (gallstones) occurred in 1.6% of semaglutide participants versus 0.7% on placebo; cholecystitis (gallbladder inflammation) in 0.6% versus 0.2%. Rapid weight loss of any cause is a known risk factor for gallstone formation, so the signal is partly attributed to the weight-loss mechanism rather than the drug alone.

Pancreatitis: acute pancreatitis, including fatal cases, has been observed. The label advises that clinicians should monitor patients who develop severe abdominal pain radiating to the back and discontinue semaglutide if pancreatitis is confirmed. The absolute incidence in trials was low, but the event is serious enough to warrant a labelled warning and clinical attention.

The boxed warning: thyroid C-cell tumours

Every semaglutide brand (Ozempic, Wegovy, Rybelsus) carries a boxed warning in the US and an equivalent special warning in the EMA SmPC^[1] regarding thyroid C-cell tumours. In rodent studies, semaglutide produced dose-dependent thyroid C-cell tumours at clinically relevant exposures. Whether the same applies in humans is unknown; the human relevance of the rodent finding has not been established. On this basis, semaglutide is contraindicated in patients with a personal or family history of medullary thyroid carcinoma (MTC) or multiple endocrine neoplasia type 2 (MEN-2). Regulators require that prescribers counsel patients about this risk before starting treatment.

Diabetic retinopathy: a signal in type-2 diabetes patients

SUSTAIN-6, the cardiovascular outcomes trial of semaglutide in type-2 diabetes, found a significantly higher rate of retinopathy complications in the semaglutide group: a hazard ratio of 1.76 (95% CI 1.11 to 2.78) for vitreous haemorrhage, blindness, or conditions requiring treatment with an intravitreal agent or photocoagulation.^[4] The prevailing interpretation is that rapid HbA1c lowering in patients with pre-existing retinopathy worsened short-term retinal outcomes, a phenomenon also seen with other agents that sharply reduce glucose levels. The signal is specific to patients with type-2 diabetes and pre-existing eye disease; it is not replicated in the obesity-only population studied in STEP. Regardless, ophthalmology review before and during treatment is standard practice for semaglutide in type-2 diabetes.

Hypoglycaemia: low risk on monotherapy

Because semaglutide stimulates insulin secretion only in the presence of elevated blood glucose, hypoglycaemia risk on monotherapy is low. The risk increases substantially when semaglutide is combined with insulin or sulfonylurea drugs, which raise insulin levels independently of blood glucose. Clinicians typically reduce the insulin or sulfonylurea dose when adding semaglutide to an existing regimen. Patients on semaglutide combined with insulin should monitor glucose more frequently during the adjustment period.

What the SELECT trial adds for the cardiovascular population

The SELECT trial enrolled 17,604 adults with obesity or overweight and established cardiovascular disease but without diabetes. Over a mean 40-month follow-up, semaglutide 2.4 mg reduced major adverse cardiovascular events by 20% (hazard ratio 0.80, 95% CI 0.72 to 0.90)^[3], which is the cardiovascular benefit side of the ledger. On the safety side, the trial confirmed the GI-dominant adverse-event pattern seen in STEP, with no new organ-specific signals emerging in the longer follow-up that had not already appeared in earlier trials. The discontinuation rate difference (16.6% semaglutide vs 8.2% placebo) represents a real-world tolerance question for this longer treatment duration.

Medical disclaimer and what to do with this information

This article is for educational and journalistic purposes. It is not a substitute for clinical assessment. The adverse-event rates above come from controlled trials in specific populations under specific titration schedules; individual experience varies. If you are considering semaglutide or are already taking it and have concerns about side effects, speak with a prescribing clinician who knows your medical history. See also the per-country regulation pages for the approved indications in your jurisdiction.