From thyroid cancer and pancreatitis to vision loss, a new review separates established GLP-1 drug risks from uncertain safety signals and reveals why individual patient risk matters more than alarming headlines.
Study: GLP-1 receptor agonists and serious adverse events: Evidence, uncertainty and clinical implications. Image credit: Yusia13/Shutterstock.com
Published in the Journal of Internal Medicine, the review also emphasizes that the clinical importance of potential adverse effects depends on patients' underlying risks, expected treatment benefits, and available alternatives.
Clinical use and safety considerations of GLP-1 receptor agonists
GLP-1 receptor agonists, initially developed for type 2 diabetes, now play an important role in weight management and reduction of cardiovascular and renal risks. By 2025, approximately 5% of the populations of the UK and Finland had used these therapies, while around 5% of Swedish adults had filled a prescription. Surveys from 99 countries suggest that more than a quarter of adults aged 25–64 years may be eligible for treatment due to obesity, diabetes, or related comorbidities.
While gastrointestinal side effects such as nausea and diarrhea are common, long-term tolerability and safety remain important concerns, particularly as use expands to more diverse populations.
Clinical trials have shown modest increases in resting heart rate and highlight the need for medication adjustments to prevent hypotension and hypoglycemia with significant weight loss. Discontinuation rates due to adverse events range from 4–16% for GLP-1 receptor agonists versus 3–8% for placebo.
However, data on rare but serious risks, including pancreatitis, biliary disease, and thyroid neoplasia, are limited due to the rarity of events, inconsistent identification and recording of adverse events, insufficient follow-up, and methodological variability. As global use rises, ongoing surveillance and robust longitudinal research are crucial for defining long-term safety and guiding clinical decision-making.
Evidence sources for serious adverse events linked to GLP-1 receptor agonists
The author reviewed mechanistic studies, spontaneous reporting systems, randomized controlled trials, and cohort studies to evaluate rare adverse events associated with GLP-1 receptor agonists.
Mechanistic studies investigate how these medications might exert harmful biological effects, such as through receptor activity in specific tissues. While these studies help establish plausibility and inform hypotheses, they cannot confirm whether adverse events will occur in real-world settings.
Spontaneous reporting systems and case reports collect suspected adverse reactions from healthcare professionals, patients, and manufacturers. These sources help detect early signals, but underreporting, missing denominator data, and reporting bias limit their ability to establish causality or incidence rates.
Randomized controlled trials provide strong evidence by minimizing confounding, but often lack the size, duration, or inclusion of high-risk individuals needed to reliably detect rare or long-term adverse effects. Cohort studies, which analyze large healthcare databases, address some of these gaps by including broader populations and longer follow-up, but are susceptible to confounding and methodological biases. The review also warns that weaknesses in study design and analysis can produce misleading associations between GLP-1 receptor agonists and adverse events.
No single study type is definitive; each has inherent strengths and weaknesses. Drawing meaningful conclusions about rare adverse events requires integrating evidence across methodologies and carefully weighing each approach's limitations and advantages. The author cautions that a growing number of studies does not necessarily resolve uncertainty, particularly when studies differ in how adverse events are defined, identified, and recorded, or in their overall methodological quality.
Seven major serious adverse events investigated in relation to GLP-1 receptor agonists
The current review summarizes the study designs and evidence related to seven serious adverse events investigated in relation to GLP-1 receptor agonists.
Thyroid cancer
Thyroid cancer concerns with GLP-1 receptor agonists originated from rodent studies, but humans have lower thyroid GLP-1 receptor expression, and clinical trials have not shown increased calcitonin or a clear link to medullary thyroid carcinoma (MTC). The original biological concern relates specifically to MTC rather than thyroid cancer overall. While some cohort studies suggested an elevated risk, larger and more rigorous research generally found no association. Data from randomized trials and meta-analyses are inconsistent, hindered by small event numbers and potential detection bias.
Notably, one 2025 meta-analysis reported a statistically significant increase in thyroid cancer risk, whereas another reported a non-significant association. These conflicting findings mean we cannot conclusively establish or exclude a causal association.
As such, clinical guidance remains cautious, especially for those at risk for MTC. In the United States, GLP-1 receptor agonists carry contraindications for individuals with a personal or family history of MTC or multiple endocrine neoplasia type 2.
Non-arteritic anterior ischemic optic neuropathy
For non-arteritic anterior ischemic optic neuropathy (NAION), evidence comes largely from observational studies and case reports, with mixed results and significant confounding. Randomized trials and meta-analyses include very few cases, making conclusions uncertain. The proposed mechanism involves rapid glycemic improvement, which may increase risk in patients with diabetes. However, this mechanism remains hypothetical. NAION remains rare and, if associated, the absolute risk increase appears small for most patients, although it could be more consequential in those with a previous history of NAION.
Despite uncertainty regarding causality, the European Medicines Agency (EMA) has recommended including NAION as a very rare adverse effect in semaglutide product information and advises discontinuation if NAION is confirmed. This recommendation has not been extended to other GLP-1 receptor agonists.
In contrast, North American ophthalmology organizations have emphasized the limitations of the evidence and cautioned that discontinuing treatment could deprive patients of established cardiometabolic benefits. Evidence to guide treatment decisions in patients with previous NAION remains particularly limited.
Suicidal behavior
Possible links between GLP-1 receptor agonists and suicidal behavior and self-harm mostly arise from case reports, with some theories suggesting harm and others noting mental health benefits. Large clinical trials and cohort studies have not found increased risk, though detection is limited by the rarity of such events.
Moreover, individuals at high psychiatric risk were frequently excluded from clinical trials, and suicidal thoughts may be incompletely captured in both trials and healthcare databases. Consequently, the reassuring population-level findings may not fully apply to patients with substantial pre-existing psychiatric risk. Regulatory agencies, including the U.S. Food and Drug Administration (FDA) and the EMA, have concluded that the available evidence does not support a causal association, although uncertainty remains for certain high-risk populations.
Diabetic retinopathy
Concerns about diabetic retinopathy complications stem from the SUSTAIN-6 trial, where semaglutide was associated with a 76% relative increase in retinopathy complications compared with placebo, primarily in patients with pre-existing retinopathy and rapid HbA1c reductions, a pattern also observed with insulin.
Most other trials and cohort studies do not show an increased risk, and the effect is usually temporary and outweighed by long-term glycemic benefits. The review emphasizes that this pattern is more consistent with early worsening following rapid improvements in blood glucose control than with direct retinal toxicity from GLP-1 receptor agonists.
The choice of comparison treatment may therefore influence observed risks: an increase may be apparent when semaglutide is compared with placebo but absent when compared with another treatment producing similar improvements in blood glucose control. Ongoing retinal monitoring is recommended, particularly for patients with established retinopathy.
Acute pancreatitis
Acute pancreatitis remains a debated risk, with early concern based on biological plausibility and case reports. Most meta-analyses and large studies have not found a significant increase in risk, although one 2025 meta-analysis reported a statistically significant increase in relative risk, in contrast to most previous analyses.
Diagnoses are complicated by confounding factors such as gallstones and alcohol, but the available comparative evidence does not support an increased overall risk of acute pancreatitis. Nevertheless, a causal effect in susceptible individuals cannot be excluded.
The increased risk of gallbladder disease associated with GLP-1 receptor agonists could theoretically contribute to gallstone-related pancreatitis, although this mechanism has not been established clinically. Conversely, reductions in triglyceride concentrations and alcohol consumption could potentially lower pancreatitis risk in some patients. Evidence remains particularly limited for individuals with a previous history of pancreatitis or those who develop pancreatitis during treatment.
Gallbladder and biliary disease
Gallbladder and biliary disease is the adverse effect most consistently linked to GLP-1 receptor agonists, with strong evidence from both trials and cohort studies, especially in weight-loss settings, at higher doses, or with prolonged use. A meta-analysis of 76 randomized trials reported a 37% relative increase in gallbladder or biliary disease.
Mechanisms include rapid weight loss and direct effects on gallbladder function. Unlike the other serious adverse events examined, consistent randomized and observational findings, along with biologically plausible mechanisms, support a causal association.
Intestinal obstruction
Intestinal obstruction is mechanistically plausible due to the gut motility effects of GLP-1 receptor agonists. However, despite case reports and regulatory interest, randomized trials and meta-analyses have not found a convincing increase in risk, and cohort studies report mixed results.
Although studies from the UK and United States reported statistically significant increases in risk compared with SGLT2 inhibitors, a Scandinavian study found no such association. The author concludes that current evidence does not establish an increased risk of intestinal obstruction, despite the biological plausibility of such an effect.
Absolute risk should guide individualized GLP-1 treatment decisions
While GLP-1 receptor agonists have demonstrated clear benefits for many patients, particularly in the management of Type 2 diabetes, their full safety profile remains only partially understood. Current evidence suggests an increased risk primarily for gallbladder and biliary disease, with other adverse events being less clearly linked or context-dependent.
For diabetic retinopathy, evidence is most consistent with early worsening associated with rapid improvements in blood glucose control in patients with pre-existing disease, rather than established direct retinal toxicity. For thyroid cancer, NAION, suicidal behavior and self-harm, acute pancreatitis, and intestinal obstruction, causal associations remain uncertain.
A central message of the review is the importance of distinguishing relative from absolute risk. Because the serious adverse events examined are generally rare, even substantial relative increases may translate into small absolute risk increases for patients at low baseline risk. However, potential harms may become more clinically relevant in individuals with a history of NAION, pancreatitis, suicidal behavior, or symptomatic gallstone disease. The author illustrates these differences using hypothetical risk scenarios rather than treating uncertain associations as established causal effects.
Clinical decision-making should weigh these risks alongside the expected benefits and patient preferences, acknowledging the limitations and uncertainties inherent in available data. The balance may also vary by indication and the availability of alternative therapies. The review further cautions that safety concerns themselves may lead patients to discontinue treatment or clinicians to withhold potentially beneficial medications, even when causality remains uncertain.
Future research should focus on addressing these gaps by conducting larger and more targeted randomized controlled trials, especially in high-risk populations who may be most susceptible to rare adverse events. However, trials specifically involving patients with previous NAION, pancreatitis, or self-harm may be difficult to conduct because of practical and ethical considerations.
Additionally, as trial designs continue to evolve and direct placebo comparisons become less common, high-quality observational studies and indirect comparisons will play a critical role in clarifying the safety of newer GLP-1 receptor agonists.
Emotion management helps teens maintain weight loss after treatment