VA study finds no one-year rise in overall pancreatitis with GLP‑1 medicines

What happened

A study of U.S. Department of Veterans Affairs health records compared pancreatitis after people with type 2 diabetes began a GLP-1 receptor agonist or a sulfonylurea. It included 333,687 patients who started one of the two drug types from 2017 through 2023 and followed outcomes for 12 months.

The one-year rate of acute pancreatitis from all causes was similar between the groups. The adjusted rate difference for GLP-1 medicines compared with sulfonylureas was 3.64 fewer cases per 100,000 people, but the 95% confidence interval ranged from 30.76 fewer to 23.48 more. Because that interval crosses zero, the study did not establish a difference in overall one-year risk.

The cause-specific findings were more complicated. Starting a GLP-1 medicine was associated with 23.45 more suspected drug-induced cases per 100,000 people at one year. It was also associated with 16.96 fewer cases linked to high triglycerides and 10.32 fewer cases linked to alcohol. Biliary pancreatitis and cases classed as idiopathic or other causes occurred at similar rates.

Timing mattered. In the study's per-protocol analysis, 42% of the drug-related cases attributed to GLP-1 treatment during the full year occurred in the first three months. The combined pattern produced a higher all-cause risk during the first two months, followed by similar monthly risk for the remainder of follow-up.

What it means

The finding is more precise than saying GLP-1 medicines either do or do not carry a pancreatitis signal. The overall comparison was neutral at one year, but an early increase in suspected drug-related cases was offset by later reductions in cases attributed to alcohol and high triglycerides. Patients and clinicians should not let the one-year total erase the timing and cause-specific results.

People starting treatment should discuss the symptoms that need prompt medical attention. Severe and persistent abdominal pain, especially when accompanied by vomiting or pain that reaches the back, should be evaluated urgently. A clinician can decide whether the medicine or another cause may be involved.

What it does not mean

This was an observational analysis of health records, not a randomized trial. It can identify associations but cannot prove that a GLP-1 medicine caused one event or prevented another. The comparison was with sulfonylureas, so the results do not establish risk relative to every other diabetes or obesity treatment.

The VA population consisted of people with type 2 diabetes and may not represent patients taking a GLP-1 medicine only for obesity. The result is also not a reason to stop treatment without medical guidance. Sudden medication changes can carry their own risks.