The 2023 Ozempic gastroparesis study did not prove that semaglutide causes permanent stomach paralysis. Mohit Sodhi's 2023 JAMA research letter reviewed insurance claims rather than running a randomized trial. It found an association with digestive billing codes, not proof of irreversible nerve injury.
Among 613 people prescribed semaglutide, researchers counted 4 gastroparesis outcomes over roughly seven months. Among 654 people prescribed bupropion-naltrexone, they counted 3. The widely cited 3.67 hazard ratio pooled semaglutide with a much larger liraglutide group.
What the Ozempic gastroparesis study measured
The JAMA report was a three-page observational letter. No one was randomly assigned to a drug or placebo. Researchers analyzed 2006–2020 U.S. health claims in the IQVIA PharMetrics Plus for Academics database. A December 21, 2023 correction updated that database's full name.
The database sample held 16 million patients, but the study cohort was much smaller. It included new semaglutide or liraglutide users with an obesity code recorded 90 days before through 30 days after starting. People with diabetes codes or antidiabetic drug claims were excluded to infer weight-loss use.
The comparator was bupropion-naltrexone, an active weight-management tablet sold as Contrave (revised 11/2025, initial approval 2014). Liraglutide, sold as Saxenda (revised 6/2026), and bupropion-naltrexone are study comparators, not Promise products.
Average age is shown with the standard deviation (SD), which describes how spread out the ages were. Follow-up uses the median (the midpoint), with the interquartile range (IQR), which spans the middle half of follow-up times.
| Study group | Patients | Mean age, years (SD) | Median follow-up, years (IQR) |
|---|---|---|---|
| Semaglutide | 613 | 53.5 (11.9) | 0.6 (0.2–1.1) |
| Liraglutide | 4,144 | 51.3 (12.2) | 1.7 (0.8–3.1) |
| Bupropion-naltrexone | 654 | 45.2 (11.1) | 1.7 (0.7–2.9) |
Follow-up ran from the first prescription until a studied outcome, a drug switch, or June 2020. The four outcomes included gallbladder or bile-duct disease and pancreatitis (inflammation of the pancreas, including cases linked to gallstones). Researchers also tracked bowel obstruction (a blocked intestine) and gastroparesis (slowed or stopped stomach emptying).
A gastroparesis outcome meant a diagnosis code or use of a medicine that helps move food through the stomach. The letter did not list numeric billing codes or require a stomach-emptying scan. The analysis accounted for age, sex, alcohol, smoking, high cholesterol, recent abdominal surgery and region. Body mass index (BMI), a measure of weight relative to height, was unavailable.
The study results: absolute counts versus relative risk
Comparing relative risk against raw event numbers clarifies what the paper actually found.
| Outcome | Semaglutide: events; rate | Liraglutide: events; rate | Bupropion-naltrexone: events; rate | Pooled adjusted HR (95% CI) |
|---|---|---|---|---|
| Gastroparesis | 4; 9.1 | 66; 7.3 | 3; 3.1 | 3.67 (1.15–11.90) |
| Pancreatitis | 2; 4.6 | 71; 7.9 | 1; 1.0 | 9.09 (1.25–66.00) |
| Bowel obstruction | 0; 0 | 73; 8.1 | 2; 1.7 | 4.22 (1.02–17.40) |
| Biliary disease | 5; 11.7 | 162; 18.6 | 16; 12.6 | 1.50 (0.89–2.53) |
Each rate above is per 1,000 person-years of observation. A hazard ratio (HR) compares how fast events occur between groups over time, not your personal percentage risk. Every adjusted HR in the table pooled semaglutide and liraglutide users, and liraglutide outnumbered semaglutide nearly seven to one.
Person-year rates reflect total follow-up time rather than a fixed one-year risk. Semaglutide users had a median follow-up of 0.6 years. Zero bowel obstructions among 613 people therefore does not prove zero risk.
Few events meant wide confidence intervals (CI), the plausible range around an estimate. The biliary interval (0.89–2.53) crosses 1.0, showing no clear difference. Excluding high cholesterol shifted the bowel-obstruction HR to 3.63 (95% CI, 0.87–15.10), also crossing 1.0.
Slowed stomach emptying versus permanent paralysis
Gastroparesis is a disorder that "slows or stops the movement of food from your stomach to your small intestine," according to the NIDDK. A delayed test meal, a claims-based gastroparesis outcome and permanent paralysis are not interchangeable findings.
Semaglutide delays early post-meal stomach emptying, as noted in §12.2 of the Ozempic prescribing information (revised 5/2026). A 2023 randomized study in Diabetes, Obesity and Metabolism by Mojca Jensterle and colleagues measured this effect. It enrolled 20 women with polycystic ovary syndrome (PCOS) and obesity.
Participants (mean age 35, mean BMI 37 kg/m²) received a 12-week study regimen of semaglutide 1.0 mg weekly or placebo. At Week 13, half a solid meal emptied in 171 minutes with semaglutide versus 118 minutes with placebo. At four hours, 37% of the meal remained with semaglutide compared with none on placebo.
That small trial measured meal emptying during treatment, not population rates of permanent gastroparesis. Our guides to semaglutide acid reflux and semaglutide constipation cover those symptoms separately.
What critics and later studies found
Researchers soon questioned how Sodhi's letter separated diabetes from weight-loss prescribing. Karine Suissa's team raised these concerns in a March 2024 JAMA comment. "Such extreme findings raise clinical concerns that may not be justified in view of several methodological challenges," they wrote. In a published reply, the authors wrote, "We agree that diabetes requires a stricter exclusion," noting broader filtering than first summarized.
Later studies tested different groups. A 2025 Gastrointestinal Endoscopy study examined nondiabetic adults (BMI ≥27 kg/m²) having upper endoscopy. Gastroparesis was recorded in 1.1% of GLP-1 users versus 0.6% of controls, a relative risk (RR) of 2.00 (95% CI, 0.94–4.27). Repeat endoscopy occurred in 5.4% versus 4.2% (RR 1.28; 95% CI, 0.95–1.71), and neither interval established a clear difference.
A 2025 MarketScan cohort study in BMJ Open Gastroenterology followed nondiabetic adults aged 18 to 64 with obesity from 2018 to 2022. Gastroparesis rates were 6.5 per 1,000 person-years on semaglutide (233 of 36,990) and 2.1 on bupropion-naltrexone (33 of 7,369). After sleeve gastrectomy, the rate was 1.1 (24 of 11,101). The authors noted they were "unable to distinguish whether the gastroparesis is permanent or transient" in claims.
In type 2 diabetes, a 2026 Annals of Internal Medicine study by Salvatore Crisafulli and colleagues followed patients for roughly six months. It compared hospital or emergency visits for five severe digestive conditions against dulaglutide, sold as Trulicity (revised 3/2026). That composite covered acute pancreatitis, biliary disease, bowel obstruction, gastroparesis, and severe constipation.
| Matched diabetes comparison | Matched pairs | Adjusted severe-GI composite HR (95% CI) |
|---|---|---|
| Semaglutide vs. dulaglutide | 65,238 | 0.96 (0.87–1.06) |
| Tirzepatide vs. dulaglutide | 20,893 | 0.96 (0.77–1.20) |
| Tirzepatide vs. semaglutide | 46,620 | 1.07 (0.90–1.26) |
Similar composite rates in diabetes do not settle Sodhi's nondiabetic comparison or prove tirzepatide is safer. As of October 17, 2026, the day this article was written, a September 27, 2026 surgical meta-analysis pooled 12 observational studies (49,651 patients). Semaglutide was linked to retained stomach content before procedures (RR 3.49; 95% CI, 1.78–6.83), which is why semaglutide before surgery requires advance planning. Across two pooled studies of 3,487 patients, the gastroparesis estimate remained imprecise at RR 1.43 (95% CI, 0.71–2.87; low certainty).
What FDA labels, court filings, and compounding rules say
Current Ozempic labeling (revised 5/2026) and the Zepbound prescribing information (revised 8/2026) share the same gastroparesis warning. Each states that the drug "is not recommended in patients with severe gastroparesis" as a warning, not a contraindication.
Ozempic's postmarketing list (§6.2) includes "ileus, intestinal obstruction, severe constipation including fecal impaction" from voluntary reports, which cannot establish frequency or causation. An FDA letter dated September 22, 2023 added ileus to the postmarketing section of the September 2023 Ozempic label, before Sodhi’s October 5 publication. Federal digestive lawsuits were centralized in the Eastern District of Pennsylvania as MDL 3094 and appeared on September 2026 panel orders, without proving causation.
"Tell your healthcare provider right away if you have nausea, vomiting, or diarrhea that does not go away," the 01/2025 Ozempic Medication Guide advises. "Tell your healthcare provider if you have stomach problems that are severe or will not go away," the guide adds.
Both labels (§5.2 and §5.5) warn that "persistent or severe abdominal pain" can signal pancreatitis. That pain may radiate to the back with or without vomiting. If pancreatitis is suspected, both labels direct clinicians to stop the medication and begin appropriate management. Both Medication Guides also cover severe abdominal pain that does not go away. They direct patients to stop the medication and call their healthcare provider right away, even without vomiting.
Branded-drug studies do not establish the side-effect rate of any compounded medication. "Compounded drugs are not FDA-approved," the FDA compounding Q&A (current as of September 16, 2025) states. Semaglutide and tirzepatide prescribed through Promise are dispensed by a licensed U.S. compounding pharmacy. Each is dispensed by a licensed U.S. compounding pharmacy as a compounded medication, which is different from an FDA-approved product: the formulation offered here is not FDA-approved.
Even so, a licensed provider may still prescribe — that decision is between you and your doctor. At Promise, a licensed provider reviews every request and not everyone qualifies.
