On a GLP-1 and Hospitalized? What the Sepsis Data Says

This article is for education only and does not constitute medical advice. It is not a substitute for care from a licensed clinician. Nothing here should be read as a reason to stop, delay, or change any medication on your own — every pause and resume decision belongs to the clinician treating you.

---

Most GLP-1 safety content answers chronic questions: muscle loss, bone density, nausea. Far fewer articles address the question that matters most in the moment: what happens with my GLP-1 if I get seriously ill? If you're in Dallas, Fort Worth, or anywhere in North Texas and you end up in an emergency room with a serious infection, the care team will make dozens of medication decisions without knowing what the evidence says about your drug class. This article walks through the one large real-world comparison we have — GLP-1 users versus SGLT2 inhibitor users after sepsis — and what it can, and cannot, tell you.

What the study actually was

In October 2026, researchers published a retrospective cohort study in Drug Design, Development and Therapy (PMID 42829714) using the TriNetX US Collaborative Network — a database of de-identified electronic health records from US healthcare organizations. The question: among adults with type 2 diabetes who developed sepsis (the body's dangerous, dysregulated response to infection), did outcomes differ between people who had recently taken a GLP-1 receptor agonist versus an SGLT2 inhibitor?

After 1:1 propensity score matching — a statistical technique that pairs patients with similar ages, conditions, and medications — the study compared 11,969 patients per group, roughly 24,000 people total. The primary outcome was death from any cause within 90 days of the sepsis episode. Researchers also tracked seven secondary outcomes, including major cardiac events, ICU admission, and kidney outcomes.

One framing note before the numbers: this was not a randomized trial. Nobody assigned patients to a drug. The researchers compared groups that already differed from each other in ways matching can only partly correct. That limits what the study can prove.

What it found

The headline association: GLP-1 users had a lower rate of 90-day all-cause mortality than SGLT2 users — 48.6 versus 60.6 deaths per 100 person-years, which works out to a hazard ratio of 0.81 (95% confidence interval 0.75–0.88). In plain language, GLP-1 users in this dataset died at a lower rate in the 90 days after sepsis than the matched SGLT2 comparison group.

Several secondary outcomes pointed the same direction. Major adverse cardiac events (MACE) were lower in the GLP-1 group (HR 0.77), as were ICU admissions (HR 0.89) and the combined cardiac endpoint MACCE (HR 0.86).

But three findings complicate the story, and honest coverage has to include them:

• After correcting for multiple comparisons, only two of the seven secondary outcomes held up statistically: MACE and ICU admission. The apparent advantages in septic shock, respiratory failure, heart attack, and the combined cardiac endpoint did not survive the statistical correction.
• The kidney result was null. There was no significant difference in major adverse kidney events (HR 1.08). Given how much GLP-1 kidney enthusiasm is in circulation, a flat kidney finding in this acute setting is worth knowing about. We cover the broader kidney-outcomes evidence separately in our SELECT, FLOW, and SOUL guide and our kidney-cardiovascular protection overview.
• The study measured association, not causation. The authors calculated "E-values" — a sensitivity metric asking how strong an unmeasured confounder would need to be to explain away the result. The E-values here were modest (roughly 1.4–1.7), meaning a moderately strong unmeasured difference between the groups could plausibly account for the findings. The authors themselves labeled the results hypothesis-generating and stated that residual confounding cannot be excluded.

What this means if you're on a GLP-1

Honest version: it may mean nothing for you personally, and that's the correct read of a single observational study.

What the study suggests is a hypothesis worth testing — that people on GLP-1s at the time of a sepsis episode fared somewhat better than comparable people on SGLT2 inhibitors, particularly on cardiac outcomes and ICU use. It does not show that a GLP-1 protects you from sepsis, and it does not show that starting or staying on a GLP-1 will improve your hospital outcomes. No randomized trial has tested either claim.

It also doesn't change the practical medication questions you'll face in a hospital. Those are decided by your care team based on your specific situation — why you're admitted, whether you can keep food down, what procedures are planned — not by cohort studies.

Why GLP-1s come up in hospitals at all

Even without any sepsis data, GLP-1s matter in acute care for a mechanical reason: they slow gastric emptying. A stomach that empties slowly keeps food (and stomach acid) around longer, which raises two acute-care concerns:

• Aspiration risk during anesthesia and sedation. A full stomach during intubation can lead to food or fluid entering the lungs. This is why anesthesia teams increasingly ask when you last ate and whether you take a GLP-1 — some now treat GLP-1 users as having a potentially full stomach even after standard fasting. Our perioperative safety guide covers the planned-surgery version of this in detail.
• Tolerability during acute illness. Nausea, vomiting, or poor intake — common in serious illness — change how any medication gets managed. That's a clinical judgment for the treating team, not a rule you can apply from an article.

If you're admitted, the single most useful thing you can do is make sure the care team knows you take a GLP-1, when your last dose was, and who prescribes it. That information shapes their anesthesia, GI, and medication-timing decisions far more than any study.

Questions worth asking your care team

Hospitals run on information. These questions put yours to work:

1. "I take [semaglutide/tirzepatide/liraglutide]. Does it affect today's plan?" — surfaces the aspiration-risk and emptying considerations above.
2. "Who is deciding whether my GLP-1 is paused, and for how long?" — pause decisions should have a named owner, usually the admitting team or the prescriber.
3. "What's the resume plan, and who follows up on it?" — GLP-1s are among the medications most often left un-restarted at discharge. Know who re-starts it and when.
4. "Should my prescriber know I'm here?" — your prescriber has context the hospital team doesn't.

Notice what's missing: any question answered by "the internet." Acute-illness medication decisions are made by clinicians with your chart in front of them. Articles like this one exist so you can ask informed questions — not so you can make the calls yourself.

The bottom line

A large real-world comparison found that type 2 diabetes patients on GLP-1s had lower 90-day mortality and fewer ICU admissions after sepsis than matched patients on SGLT2 inhibitors — but the finding is observational, its effect sizes are modest, several secondary results didn't survive statistical correction, the kidney signal was flat, and the authors themselves call it hypothesis-generating. Treat it as a reason for researchers to run better trials, not as a reason to change anything about your treatment.

What you can act on today: know your dose schedule, carry your prescriber's name, and disclose your GLP-1 to any surgical or emergency team. And while you're thinking about the bigger picture of treatment decisions, our 2-lbs-per-week weight loss calculator can help you and your prescriber set realistic pace expectations for the therapy itself.

Frequently Asked Questions

Does this study prove GLP-1s protect you from sepsis?

No. The study compared outcomes after sepsis had already occurred, in a non-randomized database. It found an association between prior GLP-1 use and lower 90-day mortality versus SGLT2 use. The authors label the findings hypothesis-generating and note that residual confounding cannot be excluded. No randomized trial has tested whether GLP-1s improve sepsis outcomes.

Should I stop taking my GLP-1 if I get sick or go to the hospital?

Do not stop or hold any medication on your own. Pause and resume decisions during acute illness are made by the clinicians treating you, based on why you're admitted, your symptoms, and any planned procedures. Your job is to disclose the medication, your last dose, and your prescriber — their job is the decision.

Why do anesthesia teams ask about GLP-1s before procedures?

GLP-1s slow stomach emptying, which can leave food and fluid in the stomach longer than expected even after standard fasting. That raises aspiration risk during anesthesia and deep sedation. Care teams may take extra precautions when a patient takes a GLP-1. This is one of the most concrete, well-established acute-care interactions with the drug class.

Did GLP-1 users have better kidney outcomes in the sepsis study?

No. Major adverse kidney events showed no significant difference between the GLP-1 and SGLT2 groups in this sepsis cohort. That's a different question from the chronic kidney-disease evidence from the FLOW and SOUL trials, which we cover in our kidney outcomes guide.

What should I tell the ER if I'm admitted while on a GLP-1?

Three things: the drug name and dose, when you took your last dose, and who prescribes it. If a procedure or surgery may happen, say so — the gastric-emptying effect is most relevant to anesthesia planning. If you've had vomiting or poor intake, mention that too; it affects how the team manages the medication.