Large trials show that some GLP-1 and SGLT2 medicines lower heart and blood vessel events in people with type 2 diabetes, but the stroke findings differ. Several GLP-1 trials leaned toward fewer strokes, while SGLT2 trials have not shown a clear stroke benefit. Evidence in people who have already had a stroke is still limited. This is a summary of research, not treatment advice.

Key takeaways

  • The American Diabetes Association's 2026 Standards of Care recommend an SGLT2 inhibitor or a GLP-1 medicine with proven cardiovascular benefit for adults with type 2 diabetes and established cardiovascular disease, which the ADA defines to include stroke, or chronic kidney disease.
  • In the REWIND trial, 3.2% of people on dulaglutide and 4.1% on placebo had a stroke. In SUSTAIN-6, nonfatal stroke was 1.6% with semaglutide and 2.7% with placebo. Other GLP-1 trials did not show a clear stroke difference.
  • A pooled analysis of randomized SGLT2 trials found no significant effect on stroke. Their proven strengths are in heart failure and kidney outcomes.
  • For people who have had a stroke, dedicated evidence is thin. A 2026 randomized trial in China found fewer repeat strokes at 90 days with liraglutide, but it was open-label and short.
  • Please do not start, stop, or change any of these medicines on your own. Ask your prescriber whether one fits.

What are GLP-1 and SGLT2 medicines?

Both groups were first developed for type 2 diabetes. MedlinePlus says semaglutide (Ozempic, Wegovy) prompts the pancreas to release insulin when blood sugar is high and slows stomach emptying. This group is also called GLP-1 receptor agonists, and includes dulaglutide and liraglutide.

SGLT2 inhibitors work differently. MedlinePlus says empagliflozin (Jardiance) lowers blood sugar by causing the kidneys to remove more glucose in the urine. Dapagliflozin and canagliflozin are in the same class.

Beyond blood sugar, both groups now have heart-related uses. MedlinePlus lists reducing the risk of stroke, heart attack, or death in certain people with type 2 diabetes among empagliflozin's uses.

What do the big heart-outcome trials show?

In 2008, FDA issued guidance asking makers of new type 2 diabetes drugs to run cardiovascular outcome trials, which is why so many large trials exist. A "hazard ratio" compares event rates. A ratio of 0.76 means about 24% fewer events.

GLP-1 trials

  • SUSTAIN-6 (semaglutide injection): 3,297 adults with type 2 diabetes, 104 weeks. The main outcome (heart death, nonfatal heart attack, or nonfatal stroke) occurred in 6.6% vs 8.9%, a hazard ratio of 0.74.
  • LEADER (liraglutide): 9,340 adults, median 3.8 years. Main outcome hazard ratio: 0.87. Nonfatal stroke was lower, but not by a statistically clear amount.
  • REWIND (dulaglutide): 9,901 adults, median 5.4 years. Main outcome hazard ratio: 0.88.
  • SOUL (oral semaglutide): 9,650 adults with type 2 diabetes and heart or kidney disease, published in 2025. The main outcome occurred in 12.0% vs 13.8%, a hazard ratio of 0.86.
  • SELECT (semaglutide 2.4 mg): 17,604 adults with heart disease and overweight or obesity but no diabetes. The main outcome occurred in 6.5% vs 8.0%, a hazard ratio of 0.80.

SGLT2 trials

In EMPA-REG OUTCOME (empagliflozin), the main outcome fell, with a hazard ratio of 0.86. People on empagliflozin also had fewer cardiovascular deaths and fewer hospitalizations for heart failure. There was no significant difference in heart attacks or strokes. The ADA says both groups also reduce heart failure hospitalization and kidney disease progression.

What do the trials say about stroke specifically?

For GLP-1 medicines, the picture leans favorable but is not uniform. Stroke was usually a secondary measure, so most trials were not built to prove a stroke effect.

  • REWIND looked closely. Strokes happened in 158 of 4,949 people on dulaglutide (3.2%) and 205 of 4,952 on placebo (4.1%), a hazard ratio of 0.76. The benefit showed up for ischemic stroke, not hemorrhagic stroke.
  • SUSTAIN-6 found nonfatal stroke in 1.6% vs 2.7%, a hazard ratio of 0.61.

The ADA's 2026 Standards note that liraglutide, semaglutide, albiglutide, efpeglenatide, and dulaglutide all reduced major cardiovascular outcomes (heart attack, stroke, or cardiovascular death), while lixisenatide and extended-release exenatide did not.

For SGLT2 inhibitors, a 2022 meta-analysis of randomized trials found a hazard ratio for stroke of 0.98 (range 0.88 to 1.09), which means no significant effect. Observational studies in the same paper looked better, but they cannot rule out other explanations.

What about people who have already had a stroke?

The ADA's 2026 Standards define established cardiovascular disease to include stroke. Its recommendation 10.40a, graded A, says that for adults with type 2 diabetes and established cardiovascular disease or chronic kidney disease, an SGLT2 inhibitor or GLP-1 medicine with demonstrated cardiovascular benefit is recommended. That is a diabetes guideline, not a stroke-specific one.

Fewer studies look only at stroke survivors:

  • LAMP trial (JAMA Internal Medicine, 2026). In China, 636 adults with type 2 diabetes and a minor stroke or high-risk TIA were randomized within 24 hours of symptoms to liraglutide for 90 days or usual care. Repeat stroke occurred in 7.9% vs 13.8% (hazard ratio 0.56). It was open-label, short, and limited to Chinese patients, and the authors say the results apply to that group.
  • A preliminary U.S. study. At the American Heart Association's 2024 meeting, researchers using Rochester Epidemiology Project records reported on 7,044 adults hospitalized for ischemic stroke. People prescribed a GLP-1 or SGLT2 medicine after the stroke had lower death and heart attack rates. The release calls the findings preliminary, notes the repeat-stroke result appeared only after adjusting for risk factors, and quotes an independent expert who says a clinical trial is needed.

A recent narrative review of this field concluded that these medicines are best considered for long-term prevention in people who already have an indication, while stroke-specific trials are pending.

What cautions come with these medicines?

MedlinePlus lists several for empagliflozin, including ketoacidosis, dehydration, urinary and genital infections, dizziness, and an increased risk of lower-limb amputation. For semaglutide, it lists a warning about thyroid gland tumors seen in rats, plus pancreatitis and dehydration. In SUSTAIN-6, retinopathy complications were higher with semaglutide. Risks differ from person to person, so this decision belongs with a prescriber who knows your history.

What questions could I bring to my care team?

  • Do I have a separate reason for one of these medicines, such as diabetes, heart failure, or kidney disease?
  • Which medicine has proven benefit for someone like me?
  • How does it fit with the rest of my stroke prevention plan?
  • What side effects should I watch for?

If your prescriber chooses one and your pharmacy is out of the strength, see our GLP-1 availability guide. Needle phones pharmacies for you and reports which ones say they have your exact drug, strength, and form (see how it works). You can start at app.findneedle.co.

Frequently asked questions

Do GLP-1 medicines prevent stroke?

The evidence is mixed. In some trials of people with type 2 diabetes, such as REWIND and SUSTAIN-6, stroke was less common with a GLP-1 medicine. In others, the difference was not clear. Whether one makes sense for you is a prescriber's decision.

Do SGLT2 inhibitors lower stroke risk?

A pooled analysis of randomized trials found no significant effect on stroke. They have shown benefits for heart failure and kidney outcomes.

Could a GLP-1 medicine be covered after a stroke if I do not have diabetes?

Coverage rules vary. CMS's temporary Medicare GLP-1 Bridge, which runs July 1, 2026 through December 31, 2027, lists a prior heart attack or stroke with a body mass index of 27 or higher among its eligibility paths for covered weight-management drugs. A prior authorization is required. Check Medicare.gov or your plan.

Is the 2024 American Heart Association study proof that these drugs work after a stroke?

No. It was a preliminary, observational look at medical records. Observational data can show an association but not cause and effect.

Sources