Do GLP-1 Agonists Cause Pancreatitis? The Evidence

GLP-1 Side Effects and Safety

No increased risk

A 2026 multicenter analysis found GLP-1 receptor agonist use did not increase the risk of acute pancreatitis and was associated with lower complication rates in type 2 diabetes patients who did develop pancreatitis.

Nieto et al., American Journal of Gastroenterology, 2026

Nieto et al., American Journal of Gastroenterology, 2026

View as image

Pancreatitis is listed as a warning on every GLP-1 receptor agonist label. The FDA requires it. Case reports describe patients developing acute pancreatitis after starting semaglutide, liraglutide, or tirzepatide. The fear is understandable: acute pancreatitis is painful, occasionally life-threatening, and the idea that a weight-loss medication might cause it is alarming. But the relationship between GLP-1 agonists and pancreatitis is far more nuanced than the label implies. Clinical trial data from tens of thousands of patients, multiple meta-analyses, and large real-world studies have produced a consistent pattern: the signal is weak, confounded by the very conditions GLP-1 drugs treat, and possibly not causal at all. A 2026 multicenter analysis found that GLP-1 receptor agonist use did not increase the risk of acute pancreatitis and was associated with lower complication rates in patients who did develop it.^[1] Understanding the full evidence landscape requires separating case reports from controlled data, pharmacovigilance signals from clinical outcomes, and correlation from causation. This article does that. For information on how GLP-1 agonists interact with other medications, see GLP-1 Drug Interactions.

The Origin of the Concern

The pancreatitis question dates to 2007, when the FDA received reports of acute pancreatitis in patients taking exenatide (Byetta), the first GLP-1 receptor agonist. The Adverse Event Reporting System (FAERS) accumulated enough reports to trigger a safety review. By 2009, the FDA added pancreatitis warnings to all GLP-1 RA labels.

The biological plausibility was not unreasonable. GLP-1 receptors are expressed on pancreatic acinar cells and ductal cells, not just the beta cells targeted for glucose-lowering effects. Animal studies showed that high-dose exenatide and liraglutide could induce pancreatic ductal hyperplasia and low-grade inflammation in rodents. Rouse et al. (2014) demonstrated that GLP-1 drugs combined with a high-fat diet induced pancreatic injury in mice, including acinar cell apoptosis and ductal proliferation.^[6]

But rodent pancreatic physiology differs from human physiology in important ways. Rodents are far more susceptible to pancreatic ductal proliferation, and the doses used in animal studies typically exceeded human therapeutic doses by wide margins. The question was always whether these preclinical findings translated to human clinical risk.

De Heer and Bhatt (2014) reviewed the early evidence and concluded that while a mechanistic link was plausible, the clinical data did not support a causal relationship. They noted that the background rate of pancreatitis in type 2 diabetes patients is 2-3 times higher than in the general population, making it difficult to attribute cases to the medication rather than the underlying disease.^[5]

What the Clinical Trials Show

The most reliable data comes from large, randomized, placebo-controlled cardiovascular outcomes trials (CVOTs), which enrolled tens of thousands of patients followed for years.

SUSTAIN 6 (semaglutide, n=3,297): Acute pancreatitis occurred in 9 patients on semaglutide versus 12 on placebo. Semaglutide showed no excess risk.

PIONEER 6 (oral semaglutide, n=3,183): Acute pancreatitis in 1 semaglutide patient versus 3 on placebo.

LEADER (liraglutide, n=9,340): Pancreatitis incidence was 0.4% with liraglutide versus 0.5% with placebo. No significant difference.

SELECT (semaglutide for obesity, n=17,604): In the weight management population without diabetes, pancreatitis rates were low and similar between groups.

Across the SUSTAIN and PIONEER programs combined, Aroda et al. (2023) reported that rates of acute pancreatitis were similar for semaglutide versus comparators across all Phase IIIa trials. The pooled incidence was approximately 0.2 cases per 100 patient-years.

Gale (2015) analyzed the liraglutide trial data in an editorial titled "Smoke or Fire?" and concluded that the pancreatitis signal in LEADER was consistent with random variation around a low baseline rate, not a drug effect.^[7]

The Meta-Analysis Landscape

Multiple meta-analyses have attempted to settle the question. They reach slightly different conclusions depending on which studies they include, how they handle confounding, and whether they use clinical trial or observational data.

Giorda et al. (2015) conducted a systematic review and meta-analysis of observational studies on incretin-based therapies and acute pancreatitis risk. The pooled estimate suggested a modestly elevated risk, but the authors noted that observational studies cannot adequately control for confounding by indication (sicker patients are more likely to receive GLP-1 agonists and more likely to develop pancreatitis).^[8]

Wen et al. (2025) published the most comprehensive meta-analysis of randomized controlled trials: 62 RCTs with 66,232 patients. They found a pooled relative risk of 1.44 (95% CI: 1.09-1.89) for pancreatitis with GLP-1 RAs. This is the number that generates headlines. But the authors themselves noted that the risk became non-significant when stratified by background medication use, suggesting confounding rather than a direct drug effect.

A separate network meta-analysis of 102,257 participants found a neutral relationship (relative ratio 0.96, 95% CI: 0.31-3.00), and another meta-analysis of 55 RCTs (33,350 patients) concluded that GLP-1 agonists do not pose an increased pancreatitis risk compared to controls.

The inconsistency across meta-analyses reflects a fundamental challenge: pancreatitis is rare (0.2-0.4% incidence), making it difficult to detect small risk differences even in large trials. When the true effect is close to zero, different analytical choices (inclusion criteria, statistical methods, confounder adjustment) can push the point estimate to either side of 1.0.

Real-World Evidence

Real-world data introduces different biases than clinical trials (less controlled populations, more comorbidities, less rigorous outcome ascertainment) but offers larger sample sizes and longer follow-up.

Ayoub et al. (2025) conducted a propensity score-matched analysis in a comorbidity-free subgroup of type 2 diabetes patients. They found acute pancreatitis in 0.31% of GLP-1 RA users versus 0.24% of DPP-4 inhibitor users. The difference was modest and fell within the range of expected variation given the multiple confounders in even the "comorbidity-free" population.^[2]

Faour et al. (2025) compared GLP-1 RA initiators to other diabetes medication initiators in a real-world study and found no increased risk of acute pancreatitis or pancreatic cancer.^[9]

Nieto et al. (2026) published the largest and most recent multicenter analysis, finding that GLP-1 RA use was not associated with increased acute pancreatitis risk. Their finding went further: among patients who did develop pancreatitis, those on GLP-1 RAs had lower complication rates, suggesting a possible protective anti-inflammatory effect.^[1]

Fang et al. (2025) in Diabetes Care found that incretin-based medications were associated with increased biliary events (gallstones, cholecystitis) but not with acute pancreatitis directly. This finding is important because gallstones are the most common cause of acute pancreatitis, suggesting that any GLP-1-associated pancreatitis signal may be mediated through the gallbladder rather than direct pancreatic toxicity.^[3] This gallbladder connection is explored in detail in GLP-1s and Gallbladder Problems.

The Confounding Problem

The single biggest challenge in interpreting GLP-1 pancreatitis data is confounding by the conditions GLP-1 drugs treat. Every major pancreatitis risk factor is overrepresented in GLP-1 RA users:

Type 2 diabetes: Patients with T2D have a 2-3 fold higher baseline pancreatitis risk than the general population. The mechanism is unclear but may involve metabolic dysfunction in acinar cells, microvascular disease affecting pancreatic blood supply, or chronic low-grade inflammation.

Obesity: BMI above 30 increases pancreatitis risk by approximately 1.5-2 fold. Visceral adiposity specifically is associated with more severe pancreatitis outcomes.

Gallstones: Approximately 15-20% of adults have gallstones, and gallstone disease is the leading cause of acute pancreatitis. GLP-1 agonists may increase gallstone formation through effects on gallbladder motility (slowed emptying concentrates bile). This creates a pathway to pancreatitis that is drug-related but not through direct pancreatic toxicity.

Rapid weight loss: Losing more than 1.5 kg per week increases gallstone formation risk. Patients on semaglutide for weight loss often lose weight faster than this threshold, particularly in the first 3-6 months.

Hypertriglyceridemia: Severe hypertriglyceridemia (above 500 mg/dL) causes approximately 5-10% of acute pancreatitis cases. GLP-1 agonists actually reduce triglycerides, which should lower this risk, but patients starting GLP-1 therapy often have poorly controlled lipids at baseline.

Azoulay (2015) noted that after nearly a decade of study, the inability to fully control for these confounders left genuine uncertainty about whether GLP-1 agonists cause even a small increase in pancreatitis risk.^[10] A decade later, the same uncertainty persists, though the weight of evidence has shifted toward "no causal relationship."

What About Patients With Prior Pancreatitis?

A common clinical question: can GLP-1 agonists be prescribed to patients who have a history of pancreatitis? The FDA label advises against restarting GLP-1 therapy if pancreatitis is confirmed during treatment, but it does not prohibit initiation in patients with a remote history of pancreatitis.

Calvarysky et al. (2026) studied this directly, examining recurrence risk in diabetic patients with a history of pancreatitis who were started on GLP-1 analogues. They found no increased risk of recurrent pancreatitis compared to patients with the same history who were not prescribed GLP-1 therapy.^[4]

This is a clinically important finding. Many patients with a history of alcohol-related or gallstone-related pancreatitis also have obesity and type 2 diabetes, which are primary indications for GLP-1 agonists. Withholding an effective treatment based on an unproven risk creates a therapeutic gap, particularly for patients with limited alternative options.

Dominguez et al. (2026) reviewed the accumulated evidence and concluded that the pancreatitis risk associated with GLP-1 RAs is "overstated," noting that the clinical trial data, real-world evidence, and biological understanding all point toward confounding rather than causation.^[11]

Individual Drug Differences

Not all GLP-1 agonists carry the same pharmacovigilance signal. FAERS data shows different reporting rates by drug:

Liraglutide: The highest reporting odds ratio (ROR 20.13 in one pharmacovigilance analysis), but this may reflect its longer time on market, its use at higher doses for weight loss, and reporting bias during the period of peak pancreatitis concern.

Semaglutide: Lower reporting rate than liraglutide. The pooled OR from RCTs is approximately 0.7 (95% CI: 0.5-1.2), indicating no increased risk.

Tirzepatide: Newer to market with fewer total patient-years of exposure. Case reports of severe pancreatitis exist, but the controlled data does not show excess risk. The dual GIP/GLP-1 mechanism adds complexity because GIP receptors are also present on pancreatic acinar cells.

Exenatide: The original GLP-1 RA that triggered the initial FDA concern. Short-acting (twice daily) formulation has higher peak drug levels, which may be relevant. Roy et al. (2014) found that exenatide did not cause detrimental effects on the pancreas in a comprehensive preclinical evaluation.^[12]

Giorda et al. (2014) reviewed acute pancreatitis as an adverse event of type 2 diabetes drugs broadly and cautioned against drawing drug-specific conclusions from FAERS data, noting that reporting rates are influenced by media coverage, regulatory attention, and prescriber awareness rather than true incidence.^[13]

The Practical Bottom Line

The current evidence, synthesized across clinical trials, meta-analyses, and real-world studies, does not support a causal link between GLP-1 receptor agonists and acute pancreatitis at the population level. The signal that exists is small, inconsistent across study designs, and largely explained by confounding.

This does not mean pancreatitis cannot occur in a patient taking a GLP-1 agonist. It can. The question is whether the drug caused it or whether the patient's underlying risk factors (diabetes, obesity, gallstones, rapid weight loss) would have produced the same outcome regardless of treatment.

For other GLP-1 side effects that are more clearly drug-related, such as nausea, gastroparesis, or injection site reactions, the causation is more established. The thyroid cancer concern follows a similar pattern to pancreatitis: strong in rodent models, weak in human data, heavily confounded. The retinopathy question is distinct, with a more specific mechanism tied to rapid glycemic improvement.

The Bottom Line

The pancreatitis concern around GLP-1 receptor agonists originated from early FAERS reports and animal studies showing pancreatic ductal changes at high doses. Clinical trial data from over 30,000 patients in placebo-controlled RCTs consistently shows no excess pancreatitis risk with semaglutide, liraglutide, or exenatide. Meta-analyses are split: one shows a pooled RR of 1.44, but the signal disappears after adjusting for background medications. The largest 2026 multicenter analysis found no increased risk and suggested possible protective effects. Confounding by diabetes, obesity, gallstones, and rapid weight loss makes it impossible to fully exclude a very small risk, but the weight of evidence points toward confounding rather than causation.

Frequently Asked Questions

Do GLP-1 drugs cause pancreatitis?

Clinical trial data does not support a causal link between GLP-1 receptor agonists and acute pancreatitis. In the SUSTAIN 6 trial, more pancreatitis cases occurred on placebo than on semaglutide (12 vs. 9). A 2026 multicenter analysis of over 20,000 patients found no increased risk with GLP-1 RA use. The pancreatitis signal in some observational studies is largely explained by confounding: patients taking GLP-1 drugs have diabetes, obesity, and gallstones, all of which independently increase pancreatitis risk.

How common is pancreatitis on semaglutide?

Across the SUSTAIN and PIONEER clinical trial programs, acute pancreatitis occurred at a rate of approximately 0.2 cases per 100 patient-years, with similar rates in the semaglutide and placebo groups. The pooled odds ratio from randomized trials is approximately 0.7 (95% CI: 0.5-1.2), meaning semaglutide-treated patients were not at higher risk. In real-world data, rates of 0.2-0.4% have been reported, consistent with the background rate in the type 2 diabetes population.

Can you take Ozempic if you had pancreatitis before?

A 2026 study by Calvarysky et al. found no increased risk of recurrent pancreatitis in diabetic patients with a history of pancreatitis who were started on GLP-1 analogues. The FDA label advises against restarting GLP-1 therapy if pancreatitis occurs during treatment, but does not prohibit initiation in patients with a remote history of pancreatitis from other causes. Clinical judgment should weigh the benefits of GLP-1 therapy against the individual patient's pancreatitis history, cause, and current risk factors.

Does tirzepatide cause pancreatitis?

Tirzepatide (Mounjaro/Zepbound) is newer to market and has fewer total patient-years of data than semaglutide or liraglutide. Case reports of pancreatitis during tirzepatide use exist, including one fatal case. However, controlled clinical trial data does not show excess pancreatitis risk compared to placebo. Because tirzepatide activates both GIP and GLP-1 receptors, and GIP receptors are present on pancreatic acinar cells, ongoing safety monitoring is warranted as the drug accumulates more real-world exposure.

Why do GLP-1 drugs have a pancreatitis warning?

The FDA added pancreatitis warnings to GLP-1 agonist labels in 2009 based on post-marketing adverse event reports with exenatide (Byetta) and supporting animal data showing pancreatic ductal changes at high doses. The warning remains as a precautionary measure even though subsequent clinical trial data has not confirmed an excess risk. Regulatory agencies tend to maintain safety warnings once added, especially when complete exclusion of risk is not possible due to confounding factors in the study populations.

Is GLP-1 pancreatitis risk from the drug or from weight loss?

Rapid weight loss is an independent risk factor for gallstone formation, and gallstones are the leading cause of acute pancreatitis. Patients losing more than 1.5 kg per week, common during early GLP-1 therapy, have increased gallstone risk. A 2025 Diabetes Care study found that incretin medications were associated with increased biliary events (gallstones, cholecystitis) but not with pancreatitis directly, suggesting that any drug-associated pancreatitis signal may be mediated through gallbladder pathology rather than direct pancreatic toxicity.