REVIEW OF: Seminer A et al, JAMA Neurol 2025;82(5):450–460
STUDY TYPE: Pooled analysis

Diabetes raises dementia risk, likely through damage to blood vessels in the brain. Guidelines favor GLP-1 receptor agonists (like semaglutide) and SGLT2 inhibitors (like empagliflozin) for patients with diabetes and cardiovascular risk, since both classes cut strokes and heart attacks. Whether either class also protects against the slower slide into dementia has been unclear. This analysis pooled 23 trials (over 160,000 patients with diabetes or cardiovascular risk) tracking new dementia or cognitive impairment over about two and a half years.

Combined, neither class significantly lowered dementia risk. But viewed separately, the results diverged: GLP-1 agonists cut the risk roughly in half (odds ratio [OR] 0.55, p < 0.05), while SGLT2 inhibitors showed no benefit (OR 1.20), a difference that was itself statistically significant (p = 0.04). Neither class reduced Alzheimer’s disease dementia specifically, and vascular dementia cases were too few to draw conclusions, though the trend favored treatment. The finding rests on just over 200 total dementia cases, and these trials weren’t designed to catch dementia in the first place, so the true effect could be larger.

CARLAT TAKE
Most of us don’t prescribe diabetes drugs, but patients ask about dementia prevention, and increasingly about GLP-1 agonists. This isn’t a reason to start semaglutide for cognition alone. But if a diabetic patient is already weighing GLP-1 agonists against SGLT2 inhibitors, and cognitive decline is on their mind, it’s reasonable to flag this as a point in favor of a GLP-1 option and worth mentioning to their internist.