Initiation of glucagon-like peptide-1 receptor agonists was not associated with a higher 1-year rate of acute pancreatitis compared with sulfonylureas among patients with type 2 diabetes in a US Veterans Affairs target trial emulation.
A target trial emulation is an observational approach designed to mimic a randomized trial using real-world data. In this case, researchers compared GLP-1 RAs with sulfonylureas—another glucose-lowering drug—meaning the findings reflect relative risk between therapies rather than absolute risk compared with no treatment.
Researchers analyzed 333,687 patients who newly started a GLP-1 RA or a sulfonylurea from 2017 to 2023. Patients with prior pancreatitis, pancreatic cancer, gastroparesis, medullary thyroid carcinoma, multiple endocrine neoplasia type 2, severe kidney disease, dialysis or kidney transplant, and serious hypoglycemia were excluded.
At 1 year, all-cause acute pancreatitis rates were similar between groups, with about 4 fewer cases per 100,000 patients among those initiating GLP-1 RAs vs sulfonylureas. However, the estimate was imprecise, with results compatible with both lower and higher risk.
Cause-specific analyses showed a more nuanced pattern. GLP-1 RA initiation was associated with about 23 additional suspected drug-induced acute pancreatitis cases per 100,000 patients, but with about 17 fewer hypertriglyceridemia-associated cases and about 10 fewer alcohol-induced cases per 100,000 patients. Rates of biliary and idiopathic or other pancreatitis were similar between groups.
Early Risk Window May Be Clinically Relevant
Importantly, overall pancreatitis risk was higher during the first 2 months following GLP-1 RA initiation, reflecting an early increase in suspected drug-induced cases before potential downstream metabolic effects emerge.
In per-protocol analyses of patients who continued their assigned medication, the first 3 months accounted for more than 40% of the 1-year rate difference for suspected drug-induced events. By contrast, lower rates of alcohol-induced pancreatitis were most evident during months 4 to 6, and lower rates of hypertriglyceridemia-associated pancreatitis were most evident during months 10 to 12. After the initial period, monthly risk was similar between groups.
The findings provide a rationale for early patient counseling and clinical vigilance following treatment initiation.
Findings Largely Reflect Semaglutide Use
Semaglutide accounted for about two-thirds of GLP-1 RA use in the cohort, meaning the overall results largely reflect outcomes associated with that agent. A prespecified semaglutide analysis showed patterns consistent with the class-level findings.
Generalizability May Be Limited
The cohort consisted of Veterans Affairs patients with type 2 diabetes and was predominantly older and male. Only about 7% of participants were women, and about 2% were younger than 40 years. These characteristics may limit generalizability, particularly given that GLP-1 RAs are increasingly used in younger patients and in women.
Additionally, all patients had type 2 diabetes. The findings may not apply to patients using GLP-1 RAs for obesity without diabetes, who now represent a substantial proportion of treated patients and may have different baseline risk profiles.
The study has other limitations. Its observational design limits causal interpretation despite extensive adjustment. Acute pancreatitis events were limited to emergency department or inpatient encounters, and cause-specific classification relied on electronic health record data. Suspected drug-induced pancreatitis was a diagnosis of exclusion, leaving potential for misclassification.
“Similar rates of all cause acute pancreatitis in GLP-1RA and sulfonylurea users were observed at one year,” wrote Yan Xie, of VA Saint Louis Health Care System, and colleagues. “Increased risk of suspected drug induced pancreatitis during the early period was offset by later reductions in alcohol induced and hypertriglyceridaemia associated acute pancreatitis.”
Disclosures: The study was funded by the US Department of Veterans Affairs. Xie and Ziyad Al-Aly reported uncompensated consulting for Pfizer, which has GLP-1 receptor agonist programs in development; Taeyoung Choi reported no competing interests.
Source: BMJ Medicine
