Can Ozempic or tirzepatide cause ketoacidosis?

Ketoacidosis is a frightening word, and a real emergency: acid and ketones build up in the blood faster than the body can clear them. So when headlines describe young women using tirzepatide for weight loss and ending up in intensive care, the fear that "Ozempic gives you ketoacidosis" spreads quickly. It is worth separating what the evidence shows from what the scare implies.

The short answer: ketoacidosis on GLP-1 drugs is rare, and the strongest evidence, drawn from people with type 1 diabetes, does not show a clear increase. But "rare" is not "never," and the real risk is not spread evenly. It clusters around a few specific situations. This article names them plainly.

Can Ozempic or tirzepatide cause ketoacidosis?

On current evidence, GLP-1 drugs like Ozempic and Wegovy, and the GLP-1/GIP drug tirzepatide (Mounjaro and Zepbound), are not clearly linked to a higher rate of ketoacidosis. The strongest single source is a 2026 systematic review and meta-analysis in Diabetes Research and Clinical Practice that pooled 25 studies of these drugs used as an add-on in type 1 diabetes.1

Type 1 diabetes is the right place to look, because that is where the fear of ketoacidosis is greatest: people with type 1 make little or no insulin, and insulin is what holds ketoacidosis at bay. In that highest-risk group, the pooled analysis found a diabetic ketoacidosis relative risk of 0.60, meaning no signal of an increase.1 The authors, who declared no competing interests, also found neutral overall hypoglycaemia (relative risk 1.01) and no rise in severe hypoglycaemia (relative risk 0.74) or serious adverse events (relative risk 0.89).1

What did the type 1 diabetes review find?

The reviewers pooled 25 studies, 23 of them randomised controlled trials, and searched the literature through June 2025.1 Their headline safety conclusion is reassuring: adding a GLP-1 drug to insulin in type 1 diabetes did not raise ketoacidosis or severe low blood sugar in the pooled data.

But the honest reading needs one important caveat, which the authors themselves flag. The ketoacidosis estimate was graded very-low certainty, because ketoacidosis was uncommon in the trials and the numbers carried wide statistical imprecision.1 That means the fair conclusion is "no signal of increased ketoacidosis, on limited evidence," never "GLP-1 drugs cut ketoacidosis risk by 40 percent." A relative risk below 1 here reflects sparse data, not a proven benefit.

The same review did find a real, expected downside: GLP-1 drugs sharply increased nausea (relative risk 2.89) and vomiting (relative risk 3.10), which drove a roughly twofold rise in early treatment withdrawal.1 That matters here, because persistent vomiting and not eating are exactly the conditions that can tip a susceptible person toward ketoacidosis. Tolerability improved after the first six months.1

Where does the real risk cluster?

Ketoacidosis on these drugs is best understood as a situation, not a simple drug side effect. Three settings account for most of the concern, and they are different from each other. Reading them side by side is the clearest way to see who should be careful.

Drawn from the 2026 type 1 diabetes meta-analysis, a 2026 Swedish nationwide cohort, a 2026 multi-society consensus guideline, and a 2026 intensive-care case report. These sources study different populations and are not a head-to-head comparison.1234

Is the SGLT2 number the same thing?

This is the single most-confused point, so it is worth stating carefully. A 2026 Swedish nationwide study followed all 72,698 adults with type 1 diabetes in the country and reported a diabetic ketoacidosis incidence of 24.9 events per 1,000 person-years.2 That number gets copied around as if it describes GLP-1 drugs. It does not.

The 24.9-per-1,000 figure was measured specifically among 1,291 first-time SGLT2 inhibitor users, a separate class of drugs (such as dapagliflozin and empagliflozin) that has a long-recognised link to ketoacidosis.2 GLP-1 drugs and SGLT2 inhibitors are frequently discussed together and sometimes prescribed together, which is exactly how the two get blurred. In the same population, off-label GLP-1 use had climbed to 5.5 percent by 2024, while SGLT2 use stayed under 1 percent.2 Attributing the SGLT2 ketoacidosis rate to a GLP-1 drug is a factual error, not a rounding difference.

The through-line is not the GLP-1 drug itself. It is insulin deficiency, an SGLT2 inhibitor in the mix, or using a powerful medicine with no one watching. Reading the 2026 evidence together

What about weight loss without a doctor?

The situation driving today's headlines is the third one: buying these drugs online for cosmetic weight loss and using them with no medical oversight. A 2026 case report in Cureus makes the risk concrete. A woman in her early 30s with class II obesity, and no diabetes, developed persistent vomiting, upper-abdominal pain and fatigue six days after her first tirzepatide injection.4

Tests showed a high-anion-gap metabolic acidosis with markedly raised ketones. She was admitted to intensive care, treated with intravenous fluids, electrolytes, dextrose and insulin, and recovered within 48 hours.4 One case cannot tell you how often this happens, and it does not. What it illustrates is that severe vomiting and sharply reduced eating, common when starting these drugs, can push even a non-diabetic person into ketoacidosis, and that having no one screening or monitoring removes the safety net.

This is also why supervised use matters for people with type 1 diabetes specifically. A 2026 multi-society consensus report, endorsed by bodies including the AACE, ISPAD and Breakthrough T1D, sets out practical guidelines for using GLP-1 and GLP-1/GIP drugs safely alongside insulin in type 1 diabetes.3 The report is candid that these drugs are not approved for type 1 diabetes and that hypoglycaemia and ketosis are potential risks, which is precisely the education a person cannot get if they obtain the drug without a clinician.3

What do the drug labels say?

The regulatory picture reinforces the drug-class distinction. Ozempic is approved in adults with type 2 diabetes for blood-sugar control, for reducing cardiovascular risk in those with heart disease, and for slowing kidney-disease progression.5 It is not approved for type 1 diabetes or for weight loss, and its US label does not list diabetic ketoacidosis as a warning or a reported adverse reaction.5

That is a meaningful contrast. SGLT2 inhibitors carry an explicit ketoacidosis warning on their labels; the semaglutide label does not.5 The absence of a labelled ketoacidosis warning is not a promise that it can never occur, but it does tell you where regulators saw a signal strong enough to warn about, and where they did not. For GLP-1 drugs, the concern is driven by context, not by a class-wide alarm.

What are the warning signs?

Whatever the setting, the symptoms of ketoacidosis are the same, and they are worth memorising. Watch for persistent nausea and vomiting, abdominal pain, unusual tiredness or drowsiness, a fruity or acetone-like smell to the breath, and rapid, deep breathing.4 These can come on over hours.

Ketoacidosis is a medical emergency, not a side effect to ride out at home. The risk is highest for people with type 1 or other insulin-deficient diabetes, for anyone combining a GLP-1 drug with an SGLT2 inhibitor, and for anyone unable to keep fluids or food down.2 If those symptoms appear, seek urgent care rather than waiting. For people with type 1 diabetes weighing a GLP-1 drug as an add-on, our companion piece on tirzepatide in type 1 diabetes and obesity covers the efficacy side of that conversation.

In one sentence: ketoacidosis on Ozempic and tirzepatide is rare and, in the type 1 diabetes trials where it is most feared, showed no clear increase on limited evidence, but the real risk clusters around insulin deficiency, co-use of an SGLT2 inhibitor, and unsupervised weight-loss use, so the answer is medical supervision and knowing the warning signs, not avoiding the drug.