Skip to content
Buy This GLP
← Research
Evidence

Two hundred days of difference

Two matched groups of 1,422, and one was followed 206 days longer than the other. The study reports an 83% lower death rate.

Neil Sanders6 min read
Mean follow-up, after one-to-one matchingon a GLP-1458.8 daysnot on one664.4 daysA 206-day gap between groups that were supposed to match.Less time under observation means fewer recorded events.

Two matched groups of 1,422 people, and one was watched for more than two hundred days longer than the other. That is the first thing to check in a study counting events over time, and it is checkable from the abstract — unlike most of what goes wrong inside a database study.

What the study reports

Non-diabetic adults carrying a chronic pancreatitis diagnosis, with or without a recorded GLP-1 prescription, matched one to one from a US research network. [1] Recurrent acute pancreatitis was recorded in 2.4% of the treated group against 10.8%, hazard ratio 0.247, 95% CI 0.149 to 0.409. Pancreatic cancer in 0.7% against 3.0%, hazard ratio 0.255. All-cause mortality in 1.5% against 11.1%, hazard ratio 0.167.

Vitamin D deficiency went the other way, hazard ratio 1.556, 95% CI 1.129 to 2.145, though the risk comparison for it was not significant — a nutritional marker moving in a population eating less, which is the direction reduced intake would predict.

Who gets prescribed one of these with pancreatitis

Somebody a clinician judged well enough to start on it. A person with advanced, painful, frequently relapsing chronic pancreatitis, or one already being investigated for a pancreatic mass, is not who gets handed a new weekly injection.

So the treated group is selected, before any drug acts, for being the healthier end of a diagnosis — and mortality of 1.5% against 11.1% in a population defined by the same recorded condition is a difference of a size that selection explains more easily than pharmacology does. The authors describe their findings as hypothesis-generating and say they should not guide prescribing decisions.

And the diagnosis itself is unverified

The authors state that chronic pancreatitis was identified from a recorded diagnosis, with supporting imaging, histological, functional and specialist-confirmation criteria unavailable to them.

That is unusually candid, and it matters here more than in most database work: chronic pancreatitis is a diagnosis that gets coded loosely, and a cohort assembled from the code alone contains people who have it, people being investigated for it, and people who were coded once and never again — the same problem as any outcome built from what got written down.

What a buyer should take from it

Nothing about starting or avoiding a GLP-1. These drugs carry no chronic pancreatitis indication, the authors say so themselves, and a person with that diagnosis considering one is in a conversation with a specialist rather than with a checkout.

What travels is the check. When a database study reports a very large benefit, look at whether both groups were watched for the same length of time — and what the comparison group actually is. Here it is people with the same code who were never offered the drug, which is not the same as people who were offered it and declined.

Frequently asked

Do GLP-1 drugs help chronic pancreatitis?
This study cannot answer that. It has no pancreatitis indication behind it, the diagnosis could not be verified, and the authors say the findings should not guide prescribing.
Why does the follow-up gap matter?
Mean follow-up was 458.8 days in the treated group against 664.4 in the comparator. Fewer days under observation means fewer chances to record an event, and the denominators shift by outcome accordingly.
Is an 83% lower death rate plausible?
Selection explains it more easily than the drug does. Who gets started on a weekly injection while carrying this diagnosis is someone a clinician judged well enough for it.
Did anything go the other way?
Vitamin D deficiency was recorded more often, hazard ratio 1.556, though its risk comparison was not significant.

Sources

  1. [1] Dhali A, et al. (2026). Clinical Outcomes Associated with GLP-1 Receptor Agonist Exposure in Non-Diabetic Patients with Chronic Pancreatitis: A Retrospective Cohort Study Journal of Personalized Medicine. PMID 42645931

Where to get it

Price the injectable sellers

The desk lists every seller that publishes an injectable figure, with the advertised price struck against the one a buyer is billed.

Open the price desk

More in Evidence