Pancreatitis is the diagnosis that sits behind a great deal of the anxiety about this drug class, and it occupies a strange position: rare enough that most clinicians prescribing GLP-1s have never seen a case caused by one, serious enough that everybody has to know the symptoms anyway.
The useful skill is not worrying about it. It is being able to tell it apart from the ordinary discomfort that most people on these drugs experience monthly.
What the trial data shows
Every large cardiovascular outcome trial in this class has adjudicated pancreatitis as a safety endpoint, because regulators required it after an earlier scare with a different incretin class.
The results have been consistent and unexciting. In SUSTAIN-6, SELECT and the SURPASS and SURMOUNT programmes, acute pancreatitis occurred in a small fraction of a percent of participants, and the difference between drug and placebo arms was not the clear excess that early case reports suggested.
This is worth stating plainly because the internet has not caught up with it. The pancreatitis fear entered public consciousness around 2013 on the back of retrospective database work and has been examined repeatedly and prospectively since, without confirmation.
1–3
Approximate cases of acute pancreatitis per 1,000 patient-years in trial populations
~3×
Lipase elevation above the upper limit of normal used in the diagnostic definition
Revised Atlanta classification
2 of 3
Diagnostic criteria required — characteristic pain, enzyme rise, or imaging
The pain, described properly
Vague descriptions are useless here, so here is the specific one.
Location. Upper abdomen, usually central or slightly left, high up under the ribs.
Character. Constant and boring rather than crampy. It does not come in waves. It does not ease between waves, because there are no waves.
Radiation. Straight through to the middle of the back. This is the single most discriminating feature.
Position. Worse lying flat, somewhat better sitting up and leaning forward. People with pancreatitis often adopt that position without being told to.
Onset and course. Builds over minutes to an hour, then stays. Hours later it is the same or worse. It does not respond to an antacid, to eating, or to not eating.
Company it keeps. Persistent vomiting that brings no relief, and a general sense of being unwell that is disproportionate to a stomach complaint.
Compare that with ordinary GLP-1 discomfort: fullness or ache that relates to what and when you ate, fluctuates through the day, eases overnight, and settles within a few weeks of a dose step. The two are not subtle variants of each other. Most of the ordinary picture is covered in is this nausea normal and how long side effects last.
The gallstone connection, which is the real story
Here is the mechanism that most discussion misses.
The commonest cause of acute pancreatitis worldwide is not alcohol. It is gallstones — a stone migrating out of the gallbladder and obstructing the duct that the pancreas drains through.
Rapid weight loss causes gallstones. It has always caused gallstones, through supersaturated bile and a gallbladder that contracts less when you are eating less, and the mechanism is set out in full in gallstones on a GLP-1.
So there is a plausible indirect route from this drug class to pancreatitis that runs through the biliary tree rather than through any direct effect on the pancreas. That is one reason the advice to avoid ultra-rapid weight loss keeps recurring, and one reason the risk-reduction measures for gallstones — steadier loss, regular meals containing some fat — do double duty here.
Who should think harder about this
The class labelling advises caution rather than prohibition in people with a history of pancreatitis, and most specialists treat a previous episode as a reason for a genuine conversation rather than an automatic no.
Higher-risk features worth raising before starting:
- A previous episode of acute pancreatitis, from any cause
- Known gallstones, or a gallbladder that has already caused symptoms
- Heavy or regular alcohol use — see GLP-1s and alcohol
- Severe hypertriglyceridaemia
- Prior pancreatic surgery or structural pancreatic disease
None of these is an absolute bar, unlike the medullary thyroid contraindication, which genuinely is.
The pain you can talk through is not this. The pain that puts you on the floor leaning forward, at two in the morning, is.
If it happens
Treatment is supportive and mostly successful: intravenous fluids, pain relief, nothing by mouth initially, and a search for the cause — which in this population means an ultrasound looking for stones.
The GLP-1 is stopped. Whether it is ever restarted is an individual decision that depends heavily on what caused the episode. If a gallstone did it and the gallbladder comes out, the calculation is different from an episode with no identifiable cause.
Keeping it in proportion
A few cases per thousand patient-years, against a drug producing fifteen to twenty percent weight loss and measurable cardiovascular benefit.
You are not likely to get this. You are, statistically, quite likely to have a bad night of abdominal pain at some point over a year on these medications. Knowing which is which — constant and through to the back, versus fluctuating and tied to meals — is the entire practical value of this page.