Semaglutide Safety Profile and Adverse Events

Overview

Semaglutide's safety profile across the SUSTAIN, STEP, PIONEER, and SELECT programs is broadly consistent with the GLP-1 receptor agonist class. Gastrointestinal adverse events are the most common findings, are most prominent during dose escalation, and attenuate with continued stable dosing. Discontinuation due to adverse events in the STEP-1 trial occurred in approximately 7 % of the semaglutide 2.4 mg group versus approximately 3 % in the placebo group.^[2]In SUSTAIN-6, the adverse-event profile at the 0.5 and 1 mg doses was similar in character but milder in frequency.^[1]

This article focuses on signals that require specific clinical consideration, including the boxed warning, organ-specific risks, and signals that emerged post-market or late in the trial programs.

Boxed warning: thyroid C-cell tumors

All three semaglutide products carry a boxed warning for the risk of thyroid C-cell tumors. In rodent carcinogenicity studies, GLP-1 receptor agonists induced dose-dependent and duration-dependent thyroid C-cell adenomas and carcinomas. GLP-1 receptors are expressed on rodent thyroid C-cells at much higher density than in humans, which may explain species-specific susceptibility. Whether this risk applies to humans at clinical exposures is unknown; no excess cases of medullary thyroid carcinoma (MTC) have been confirmed in clinical trials or epidemiological surveillance to date, but the absence of confirmed cases is not evidence of absence.^[4]

Semaglutide is contraindicated in:

• Patients with a personal or family history of medullary thyroid carcinoma (MTC).
• Patients with multiple endocrine neoplasia syndrome type 2 (MEN-2), which includes a high risk of MTC.

Patients should be counseled to report any neck mass, dysphagia, hoarseness, or dyspnea. Routine serum calcitonin monitoring is not recommended in the general population on semaglutide, but clinicians should follow evolving specialty guidance.

Gastrointestinal adverse events

The following frequencies are reported from the STEP-1 trial (semaglutide 2.4 mg, 68 weeks, adults with obesity without T2DM) and the SUSTAIN-6 trial (semaglutide 0.5 and 1 mg, T2DM with CV risk), combined where informative. Exact frequencies vary by dose, trial duration, and population.^[1],[2]

Frequencies are rounded approximations from STEP-1 at the 2.4 mg dose. Rates at the 0.5 and 1 mg T2DM doses are lower in absolute terms. The editor should verify these against the primary publication and the FDA label before publishing.

GI events are almost uniformly most frequent during the escalation phase and typically decrease substantially once a stable dose is reached. Slow titration (the label-recommended schedule) is the primary mitigation strategy. Temporary dose reduction to the last well-tolerated dose is an option for patients who experience persistent symptoms.

Pancreatitis

Acute pancreatitis has been reported in patients treated with GLP-1 receptor agonists, including semaglutide. The U.S. product label warns that, if pancreatitis is suspected, semaglutide should be discontinued and not restarted if pancreatitis is confirmed. The label recommends exercising caution in patients with a history of pancreatitis, as the safety of semaglutide in this population has not been established.^[4]

Large meta-analyses of GLP-1 RA class data have not consistently shown a statistically significant excess of pancreatitis events versus comparator drugs; the absolute risk, if any, appears to be low. However, the mechanistic plausibility (incretin receptors are expressed on exocrine pancreas) and the severity of acute pancreatitis justify continued vigilance.

Gallbladder disease

Cholelithiasis (gallstones) and cholecystitis (gallbladder inflammation) have been reported with semaglutide, particularly in the weight-management trials. The SELECT trial reported gallbladder disease events in approximately 2.8 % of semaglutide participants versus 2.3 % in the placebo group.^[3] The STEP program showed a similar pattern. The mechanism is thought to involve two factors: rapid weight loss per se increases gallstone risk through alterations in bile composition, and GLP-1R agonism may reduce gallbladder motility, promoting bile stasis. The product label notes cholelithiasis and cholecystitis as risks; patients presenting with symptoms consistent with gallbladder disease should be evaluated.^[5]

Diabetic retinopathy worsening (SUSTAIN-6 finding)

SUSTAIN-6 reported a statistically significant increase in diabetic retinopathy complications in the semaglutide group (3.0 %) versus placebo (1.8 %), a relative risk of approximately 1.76.^[1] The complication rate was higher in participants with pre-existing retinopathy at baseline and in those with greater HbA1c reductions. The leading hypothesis is that very rapid glucose lowering in patients with established retinopathy can temporarily worsen retinal blood flow and precipitate complications, an effect described previously with intensive insulin therapy (the early worsening phenomenon).

The FDA label for Ozempic includes a warning about this risk. Patients with type 2 diabetes who have pre-existing proliferative or severe non-proliferative diabetic retinopathy should have ophthalmologic assessment before and during treatment with semaglutide, and the pace of glycemic improvement should be considered in high-risk individuals. This signal was not prominently reproduced in the SELECT population, which excluded patients with T2DM, though retinopathy in non-diabetic patients is rare by definition.^[4]

Suicidal ideation: post-marketing signal under regulatory review

In 2023, the FDA and EMA initiated reviews of post-marketing reports of suicidal ideation and self-injury behavior associated with GLP-1 receptor agonists, including semaglutide, following signals in pharmacovigilance databases (FAERS and EudraVigilance). Both agencies reviewed the available clinical trial data and post-marketing reports.

As of early 2024, the EMA's Committee for Medicinal Products for Human Use (CHMP) concluded that the available data did not confirm a causal link between GLP-1 RAs and suicidal ideation. The FDA similarly found no established causal relationship based on its review of the SELECT and STEP trials, which did not show an excess of suicidal ideation in semaglutide arms. However, both agencies indicated ongoing monitoring given the size and heterogeneity of the real-world exposed population. Clinicians should follow the most current guidance from the relevant national authority, as this area is actively evolving.

Gastroparesis and anesthesia considerations

GLP-1 receptor agonists slow gastric emptying. In patients with pre-existing gastroparesis or severe GI motility disorders, semaglutide may worsen symptoms, and its use should be considered carefully in this population. The label does not absolutely contraindicate use in gastroparesis but notes the effect on gastric motility.^[4]

An emerging practical concern is aspiration risk during sedation or general anesthesia. Several post-marketing case reports described retained gastric contents in patients on GLP-1 RAs who had fasted for standard pre-operative intervals. Major anesthesiology societies (including the American Society of Anesthesiologists) have issued interim guidance suggesting:

• For elective procedures, consideration of holding once-weekly semaglutide on the day of the procedure (and potentially the prior weekly dose for higher-dose regimens) with gastric ultrasound assessment if delay is not feasible.
• For daily oral semaglutide, holding on the day of the procedure.
• Applying full-stomach precautions (rapid sequence induction) if GLP-1 RA has not been withheld or if GI symptoms are present.

These recommendations are interim and may have been updated. Practitioners should consult the most current guidance from their relevant anesthesiology and endocrinology societies.

Hypoglycemia

Because the insulinotropic action of GLP-1 receptor agonists is glucose-dependent, clinically meaningful hypoglycemia is uncommon with semaglutide monotherapy. In the SUSTAIN-6 trial, severe hypoglycemia rates were low and similar to placebo in the semaglutide group. The risk is substantially higher when semaglutide is combined with a sulfonylurea or insulin; in such cases, down-titration of the sulfonylurea or insulin should be considered when initiating semaglutide to reduce hypoglycemia risk.

Pregnancy and lactation

Semaglutide is not recommended during pregnancy. Animal studies showed fetal harm at exposures relevant to human clinical doses; these findings are consistent across the GLP-1 RA class. Because semaglutide has a half-life of approximately one week and accumulates over several weeks, the label recommends discontinuing at least 2 months before a planned pregnancy to allow adequate washout.^[4]

Women who become pregnant while on semaglutide should discontinue immediately. Unintended pregnancy on semaglutide should be reported to the Novo Nordisk pregnancy registry (contact information in the prescribing information) and discussed with a maternal-fetal medicine specialist.

Lactation: it is not known whether semaglutide is excreted in human breast milk. Given the potential for adverse effects in a nursing infant and the lack of data, semaglutide is not recommended during breastfeeding. The developmental and health benefits of breastfeeding should be weighed against the clinical need for semaglutide and potential infant exposure.

Contraindications

• Personal or family history of medullary thyroid carcinoma (MTC).
• Multiple endocrine neoplasia syndrome type 2 (MEN-2).
• Known hypersensitivity to semaglutide or any excipient in the formulation.

This list reflects the U.S. prescribing information. Regional labels (EMA, MHRA, TGA, Health Canada) may differ; prescribers should consult the locally approved product information.

Post-marketing surveillance

Post-marketing reports through FAERS (FDA), EudraVigilance (EMA), and the MHRA Yellow Card scheme continue to refine the safety picture, particularly for rare events. Signals under active or recent regulatory scrutiny include: suicidal ideation (reviewed above), severe gastroparesis leading to hospitalization, intestinal obstruction (small case series), and intra-operative aspiration. Spontaneous reports are hypothesis-generating, not confirmatory; absolute incidence estimates from FAERS should not be calculated without denominator data.

Novo Nordisk and regulatory agencies continue to require post-marketing safety studies across the approved population. The SELECT trial provided the largest and most rigorously controlled long-term safety dataset to date in the obese non-diabetic population, and did not reveal unexpected serious safety findings over a mean follow-up of 3.3 years.^[3]