Do GLP-1s Cause Gastroparesis? What the Research Actually Shows

Gastroparesis means delayed emptying of the stomach in the absence of any mechanical blockage. The word literally translates to "stomach paralysis," which is where the alarming headlines came from — but that translation badly misrepresents the condition. In gastroparesis the stomach is not paralyzed. It empties, just too slowly and often unevenly.

The formal diagnosis requires two things: symptoms (nausea, vomiting, early fullness, bloating, upper abdominal pain) *and* objective evidence of delayed emptying, usually from a gastric emptying scintigraphy study, where you eat a meal containing a small amount of radioactive tracer and imaging tracks how fast it leaves the stomach over four hours. A structural obstruction has to be ruled out first.

This distinction matters enormously for anyone on a GLP-1, because these medications *deliberately* slow gastric emptying. That's not a side effect that slipped through — it's a designed part of the mechanism. Slower emptying means food stays in the stomach longer, stretch receptors keep signalling fullness, and post-meal glucose rises more gradually. It's a meaningful part of why the drugs produce both satiety and better glucose control. We cover this in [how GLP-1 medications work](/blog/how-glp1-medications-work-mechanism-explained).

So almost everyone on a GLP-1 has measurably delayed gastric emptying. That is the drug doing its job. Only a small fraction develop emptying delayed enough, and symptoms severe enough, to meet the definition of a disorder.

The practical question isn't "do GLP-1s slow the stomach" — they unambiguously do. It's whether that crosses into clinically significant, persistent gastroparesis, and in whom.

The study that generated the headlines was a research letter published in JAMA in October 2023 by Sodhi and colleagues. It used commercial pharmacy and health claims data covering roughly 16 million U.S. adults, restricted to people without diabetes who were prescribed a medication for weight loss.

The design compared people starting a GLP-1 (semaglutide or liraglutide) against people starting bupropion-naltrexone, a different weight loss medication. Using an active comparator rather than a placebo or no-treatment group is a genuine methodological strength — it partly controls for the fact that people seeking weight loss treatment differ from those who aren't.

The headline finding: GLP-1 use was associated with a 9.09-fold higher hazard of gastroparesis (95% CI 1.25–66.0). The study also found increased risks of pancreatitis and bowel obstruction.

Now read that confidence interval again: 1.25 to 66.0. That range is enormous, and it's telling you the estimate rests on a very small number of events. A hazard ratio whose plausible range spans from "barely elevated" to "66 times higher" is not a precise measurement — it's a signal that something is there, with wide uncertainty about how big.

The absolute risk is the number that belongs in any honest summary: roughly 1 case per 1,000 person-years. Meaning if 1,000 people take a GLP-1 for a year, about one would be diagnosed with gastroparesis.

Two limitations deserve mention. First, this was a claims-based analysis, so "gastroparesis" means a diagnosis code was entered — not that a gastric emptying study confirmed it. Second, detection bias is a real concern: someone reporting persistent nausea on a widely publicized medication is more likely to get investigated and coded, which inflates apparent incidence.

The clearest signal is the trajectory over time, not the severity on any given day.

Expected GLP-1 effects follow a recognizable pattern. Symptoms appear in the first weeks and after each dose escalation, they're worst in the day or two after an injection, they're clearly tied to meal size and fat content, and — critically — they improve as you stay on a given dose. Your stomach adapts. Most people find week eight considerably easier than week two at the same dose.

Concerning patterns run the other way. Symptoms that persist unchanged or worsen over months at a stable dose. Vomiting food you ate many hours earlier, or the previous day — that's a specific and meaningful sign, because it indicates food is genuinely stagnating rather than just moving slowly. Inability to keep fluids down. Weight loss well beyond what your dose would predict. Signs of dehydration: dizziness on standing, dark urine, no urination for many hours.

A few things that are usually *not* gastroparesis despite feeling dramatic: getting full after a few bites (that's the intended effect), bloating that comes and goes with meals, and nausea in the 48 hours after an injection. Those are covered in [why GLP-1 nausea happens](/blog/glp1-nausea-why-it-happens-and-how-to-ease-it) and [GLP-1 bloating](/blog/glp1-bloating-why-you-feel-puffy-and-how-to-fix-it).

One genuinely useful self-check: are you able to maintain adequate hydration and reasonable protein intake? Someone with mild, expected slowdown can, with effort and smaller meals. Someone with significant gastroparesis usually cannot, and that functional threshold is more informative than a symptom score.

If you're unsure, the escalation path is straightforward. Tell your prescriber. Pausing a dose increase or stepping back one level resolves the great majority of cases without stopping treatment at all.

The best-established clinical consequence of GLP-1-related delayed emptying has nothing to do with gastroparesis as a chronic disease — it's the risk of pulmonary aspiration during procedures requiring sedation.

Standard pre-operative fasting rules assume a stomach empties on a predictable schedule. If a GLP-1 has slowed that process, a patient who followed fasting instructions perfectly may still have retained stomach contents when sedated. Under anesthesia the protective reflexes that keep stomach contents out of the airway are suppressed, and aspiration into the lungs is a serious complication.

This is not hypothetical. Anesthesiologists began reporting cases of unexpected residual gastric contents in fasted patients on GLP-1s, and endoscopists reported the same — finding food in the stomach of patients who had followed preparation instructions correctly. Systematic reviews and meta-analyses have since examined aspiration risk during elective upper endoscopy in GLP-1 users.

The result is that professional guidance now specifically addresses GLP-1 medications before procedures involving sedation, with recommendations around holding doses and, in some cases, using ultrasound to assess gastric contents before proceeding. Our guide to [GLP-1s before surgery](/blog/glp1-before-surgery-what-the-2024-guidance-says) walks through the current recommendations.

The operational point for you is simple and important: tell every clinician performing any sedated procedure that you take a GLP-1. Surgery, colonoscopy, endoscopy, some dental procedures, some imaging. Do not assume it's already in your chart, and do not assume they'll ask. This is the single highest-value thing you can do about GLP-1 gastric effects, and it takes one sentence.

Most of what helps is about giving your stomach less to struggle with, and about not outrunning your own adaptation.

Titrate slowly. The single biggest modifiable factor. Dose escalation schedules in trials are a maximum pace, not a target. If a dose increase brings symptoms that haven't settled after several weeks, staying at the lower dose longer is a legitimate clinical choice — and many people reach the same weight outcome, just later.

Reduce fat and portion size at each meal. Fat delays gastric emptying more than protein or carbohydrate, and volume matters independently. Four or five small meals empty far more comfortably than two large ones. This is standard gastroparesis dietary management and it works for milder GLP-1 slowdown too.

Sit upright for at least 30 minutes after eating. Gravity genuinely assists gastric emptying. Lying down after meals is a common and easily fixed contributor.

Consider texture. Liquids and purées empty from the stomach much faster than solids — this is why blended meals are the backbone of clinical gastroparesis diets. On a bad day, a protein smoothie is far more tolerable than a chicken breast, and nutritionally it can be equivalent.

Limit alcohol and stay hydrated. Alcohol slows emptying further and irritates the stomach lining. Dehydration worsens constipation, which worsens the whole upstream picture — see [GLP-1 constipation](/blog/glp1-constipation-why-it-happens-and-how-to-relieve-it).

Know your baseline risk. People with pre-existing diabetes — particularly long-standing diabetes with neuropathy — already have a meaningfully elevated background rate of gastroparesis, because damage to the vagus nerve is a primary cause. That's not a reason to avoid GLP-1s, which are strongly indicated in diabetes, but it is a reason for closer monitoring and slower titration.

For most people, no — but the honest answer is that it depends on what you're weighing it against.

Here's the risk picture as clearly as the evidence allows. Delayed gastric emptying happens to essentially everyone on these medications and is the intended mechanism. Uncomfortable but self-limiting GI symptoms are common, affecting a large minority in the first months. Clinically diagnosed gastroparesis appears to occur at roughly 1 case per 1,000 person-years — elevated relative to comparators, but uncommon in absolute terms and, in most reported cases, improving after the medication is stopped or reduced.

Against that sits a substantial and well-replicated benefit profile: 14.9% mean weight loss with semaglutide in STEP 1, 20.9% with tirzepatide 15 mg in SURMOUNT-1, a 20% reduction in major adverse cardiovascular events in SELECT, and meaningful effects on kidney outcomes in FLOW.

Where extra caution is reasonable: if you already have a diagnosis of gastroparesis, functional dyspepsia, or long-standing diabetes with autonomic neuropathy, this belongs in the conversation before you start rather than after. Existing gastroparesis is generally considered a contraindication.

What the JAMA data genuinely established is that this risk is real enough to be named and monitored — which is a reasonable thing for a medication used by many millions of people. What it did not establish is that GLP-1s commonly cause a devastating, permanent condition. The distance between "9-fold relative increase" and "1 in 1,000" is where most of the public confusion lives.

The reasonable position: take the symptoms seriously, titrate at a pace your body tolerates, tell every clinician before any sedated procedure, and don't let a headline built on a confidence interval spanning 1.25 to 66.0 make the decision for you.