The GLP-1 Side-Effect File: New Signals on Surgery, Pancreatitis, and the Gallbladder

Here is the honest read: nothing in the 2025 literature should send you flushing your pen down the toilet. But three independent reports published this year point at the same uncomfortable truth — these drugs do things to the body that matter when the body is also being cut open, scoped, or scanned. The signals are small, the evidence is moderate, and the headlines are getting ahead of the data. That is exactly why it is worth slowing down and reading what is actually on the page.

The most provocative of the three reports comes out of the PearlDiver Mariner database. Researchers pulled 340 semaglutide users undergoing posterior cervical fusion and propensity-matched them 1:5 against 1,540 controls. After Bonferroni correction — a statistical haircut designed to keep researchers honest — semaglutide users had significantly higher odds of pseudarthrosis at two years (OR 4.79, 95% CI 3.11–7.37) and dysphagia (OR 2.12, 95% CI 1.46–3.03).

Pseudarthrosis is the surgeon's word for a fusion that never finishes fusing — two vertebrae that should have grown into one stubbornly refusing to merge. A nearly five-fold odds ratio is not a rounding error. It is the kind of number that, if it holds up in prospective work, changes pre-operative checklists.

Mechanistically, the why is still open. GLP-1 agonists alter nutrient signaling, suppress appetite hard enough to drive sarcopenic-style weight loss in some users, and may modulate bone turnover through pathways researchers are still untangling. For a lifter, the takeaway is narrower than the headline: if you are on semaglutide and a fusion is on your calendar, that is a conversation with your surgeon, not a Reddit thread.

The second signal is a single case, and case reports are the lowest tier of clinical evidence — they generate hypotheses, they do not confirm them. But this one is worth reading. A 66-year-old woman on semaglutide for obesity developed acute cholecystitis within 72 hours of a routine colonoscopy. The authors hang their hypothesis on a known piece of GLP-1 pharmacology: these drugs slow gallbladder motility. A sluggish gallbladder plus the mechanical and inflammatory stress of bowel prep and insufflation is a plausible setup for an acute attack.

What the authors call for — and what the field does not yet have — is data on whether a pre-procedure washout period would lower the risk. Right now, that question is unanswered. The case is one data point, not a protocol.

The third report is, in some ways, the most unsettling because the patient felt fine. An 83-year-old man on tirzepatide was sent for an 18F-FDG PET/CT to stage a basal cell carcinoma. The scan showed diffuse radiotracer uptake throughout the pancreas with no CT abnormality. A follow-up lipase came back elevated. Diagnosis: subclinical pancreatitis, no other cause identified, attributed to the GLP-1.

Pancreatitis risk has shadowed this drug class since the early exenatide days. Most large analyses have not found a slam-dunk association. But a case like this one — caught incidentally, biochemically real, completely asymptomatic — hints that we may be undercounting the milder end of the spectrum simply because nobody scans an asymptomatic patient on purpose.

Lifters tend to be high-information patients. You already weigh your chicken and track your creatine phosphate. Apply the same discipline here. None of these three reports is sufficient on its own to flip the risk-benefit calculus of GLP-1 use. Cardiovascular and metabolic upside in the right patient is real and well-documented elsewhere. But the moderate-strength pattern emerging from 2025 is that these drugs interact with procedures — surgical, endoscopic, even imaging — in ways the original trials did not have power to detect.

The practical move is procedural, not pharmacological. Tell every clinician you see that you are on it. Bring it up before colonoscopy. Bring it up before any orthopedic or spinal surgery. Bring it up before any imaging that might pick up an incidental finding you would otherwise miss. The drugs are not the enemy. Information asymmetry is.

The peptide era is not slowing down. If anything, the next 18 months will bring oral formulations, dual and triple agonists, and use cases that stretch well beyond diabetes and obesity. The smart play for anyone using these compounds — for cuts, for metabolic health, for longevity bets — is to treat the safety literature the same way you treat training data. Read it as it comes out. Update the model. Do not over-fit on a single case report, and do not ignore a signal because it is inconvenient.

The drugs work. The complication file is still being written. Both things are true.