Explore this article's sources with AI
Follow PeptideMethods on Google
Semaglutide Panic Attacks: FAERS Readout
A 2026 FAERS analysis found an adjusted reporting signal, but timing and evidence from other designs did not support a consistent causal link.
Why we wrote this. A raw safety signal can sound definitive. This readout shows how adjustment, timing and other study designs change the interpretation.
In this article (5 sections)
A 2026 pharmacovigilance study found that panic attacks were reported more often for semaglutide than for two active comparators in the FDA Adverse Event Reporting System (FAERS). The apparent signal became smaller after adjustment and was not statistically significant in reports submitted before July 2023. The authors also found a sharp rise in reporting after public attention increased. Their conclusion was cautious: the available evidence does not show a consistent link, although it cannot exclude one[1].
This is not an incidence study. FAERS is a spontaneous reporting system, so the number of reports cannot tell readers how often panic attacks occur among semaglutide users. A reporting-odds ratio can identify unusual reporting patterns, but it cannot prove that a medicine caused an event.
The crude signal weakened after adjustment
The researchers analyzed FAERS reports received from the first quarter of 2018 through the fourth quarter of 2024. They included reports in which semaglutide, dulaglutide or empagliflozin was named as the primary suspect drug. Panic attack was identified through the corresponding MedDRA preferred term. The analysis included 33,267 semaglutide reports, 54,282 dulaglutide reports and 25,380 empagliflozin reports[2].
Panic attacks appeared in 87 semaglutide reports, compared with 28 dulaglutide reports and 13 empagliflozin reports. The crude reporting odds ratio was 5.08 versus dulaglutide and 5.12 versus empagliflozin. After adjustment for age, sex and indication, the ratios fell to 2.82 and 2.75, respectively[2]. The remaining adjusted signal is worth investigating, but it still reflects reporting, not diagnosed incidence or causation.
Adjustment matters because the compared groups may differ in ways that affect both anxiety symptoms and the likelihood of filing a report. The study accounted for several recorded variables, including whether a report involved type 2 diabetes or obesity and weight management, but it could not correct every unrecorded difference or reconstruct the detailed clinical history needed to separate a new panic attack from an existing anxiety disorder, another medicine or a stressful health event.
The timing suggested notoriety bias
The authors prespecified analyses intended to test whether publicity altered reporting. In the period before July 1, 2023, adjusted reporting odds ratios were 1.38 versus dulaglutide and 1.14 versus empagliflozin. Both confidence intervals included 1, so neither comparison was statistically significant. Against all other primary-suspect drugs in FAERS, the pre-July 2023 adjusted ratio was 0.76, with a 95% confidence interval from 0.48 to 1.18[2].
The reporting proportion for semaglutide changed markedly over time. There were 6.7 panic-attack reports per 10,000 semaglutide reports in 2018 through 2022 and 42.0 per 10,000 in 2023 through 2024. The drug-by-period interaction had a p value of 0.007[2]. This pattern is consistent with notoriety bias, in which increased awareness stimulates reporting. It does not prove that publicity explains every report.
A temporal change can have several explanations. Use of semaglutide expanded while patient populations may have changed. Media or regulatory attention may also have made people more likely to connect symptoms with treatment. FAERS does not provide a complete denominator of exposed patients, so the analysis cannot determine whether the underlying rate of panic attacks changed during the same period.
Evidence triangulation did not show a consistent association
Alongside the FAERS analysis, the authors searched MEDLINE, Embase and CENTRAL through November 24, 2025. They identified 12 studies that reported panic attacks or acute anxiety outcomes with semaglutide. According to the abstract, randomized trials and active-comparator observational cohorts did not show a consistent increase in these outcomes[1]. That broader review reduced confidence that the crude spontaneous-reporting signal represented a direct pharmacological effect.
Triangulation is useful because each evidence source fails in a different way. Randomized trials can reduce confounding but may not be designed or powered for a rare psychiatric event. Observational cohorts include routine-care populations but remain vulnerable to residual confounding and outcome misclassification. FAERS can surface unusual reports quickly, yet selective reporting and missing denominators prevent risk estimates. Our semaglutide research overview explains why trial outcomes and post-marketing signals should not be treated as interchangeable.
The paper's result should not be shortened to either "semaglutide causes panic attacks" or "semaglutide cannot cause panic attacks." Neither statement follows from the data. The defensible reading is narrower: crude FAERS disproportionality looked elevated, adjustment reduced it, the pre-publicity analysis was not significant and other study designs did not show a consistent association.
What this means for a person with new symptoms
A population-level signal analysis cannot decide what caused one person's symptoms. Timing, prior panic or anxiety, other medicines, substance use and concurrent illness can all matter in an individual assessment. The semaglutide safety section provides broader context, but it is not a substitute for clinical review.
People should not stop or alter a prescribed medicine solely because of this paper. New or severe panic symptoms deserve prompt discussion with a qualified clinician, who can assess urgency and consider other explanations. Immediate help is appropriate when symptoms include a risk of self-harm, severe chest pain, fainting or another medical emergency.
What we don't yet know
The study cannot calculate absolute risk because FAERS does not capture every exposed patient or every event. It cannot establish whether reporting changed because event frequency changed, awareness changed or both changed together. Adjustment covered age, sex and indication, but incomplete reports leave room for unmeasured differences between treatment groups.
More informative evidence would include prospectively defined panic and anxiety outcomes, adequate follow-up, active comparators and careful measurement of psychiatric history. Large linked health-record studies could test whether any association persists after stronger confounding control. Until then, the 2026 paper is best read as a lesson in interpreting a post-marketing signal, not as a verdict about causation.
Medical disclaimer: This article is for educational and journalistic purposes only and does not constitute medical advice. Peptides discussed may be classified as prescription medicines or research chemicals depending on your jurisdiction. Always consult a qualified healthcare professional before using any peptide product. PeptideMethods.com does not sell, distribute, or facilitate the sale of any peptide product.
Frequently asked
Did the study prove that semaglutide causes panic attacks?
No. The study found a reporting pattern in FAERS, but spontaneous reports cannot estimate incidence or establish causality. The signal weakened after adjustment and was not statistically significant in the prespecified period before July 2023.
How many panic-attack reports involved semaglutide?
The analysis identified 87 panic-attack reports among 33,267 semaglutide primary-suspect reports. That is a proportion within selected FAERS reports, not the risk among all people taking semaglutide.
What is notoriety bias in drug-safety reporting?
Notoriety bias occurs when publicity or regulatory attention makes an event more likely to be reported. The sharp rise in semaglutide panic-attack reporting after mid-2023 was consistent with this pattern, but timing alone cannot prove that publicity explains every report.
Should someone stop semaglutide after a panic attack?
This paper cannot guide an individual medication change. A person with new or severe symptoms should seek clinical advice and should not stop or alter a prescribed medicine solely because of a pharmacovigilance analysis.
Sources
No revisions yet. First published .