Review links insomnia to higher stroke risk; dementia evidence is mixed

What happened

A new review in Sleep Medicine Reviews links insomnia to a higher risk of stroke and hospitalization, and reports a mixed association with dementia. The European Academy of Neurology (EAN) issued a press release on the work, and as first reported by Healthline News (Healthline News), it drew a headline suggesting insomnia may raise your risk of stroke and dementia.

That framing goes further than the evidence. We could not open the full paper, so the numbers below come from the EAN press release on EurekAlert, which describes it.

"Tied to" is defensible. "Raises your risk" implies a cause, and the authors explicitly say they have not shown one.

What the review actually found

This is an umbrella systematic review, meaning it gathers existing reviews rather than enrolling people. It covered seven outcomes: dementia, stroke, depression, suicidal behavior, mortality, occupational injuries and hospitalization.

The authors searched MEDLINE and EMBASE through August 2025. They combined earlier reviews with three new systematic reviews of primary studies, on stroke, occupational injuries and hospitalization. The protocol was registered with PROSPERO.

The stroke number is the headline: a 26% higher risk. The lower confidence bound is 1.05, so the result clears statistical significance only narrowly. That is our read of the interval, not a claim from the authors.

The hospitalization figure is 28% higher odds, which is not the same as a 28% higher risk. Healthline's subhead uses the looser wording and implies causation.

On dementia, Healthline reports the review drew on five meta-analyses covering 5 to 39 studies each, 36,005 to more than 3.1 million participants, and follow-up of 2 to 40 years. Healthline reports that most of the dementia analyses linked insomnia to dementia, and three linked it to Alzheimer's disease. We could not check those figures against the paper.

What it does not show

• Causation. The authors say the review identifies associations only and does not establish a direct causal link.
• Chronic insomnia specifically. Both insomnia symptoms and diagnoses counted, with no minimum duration, despite the "chronic" framing in coverage.
• A clean comparison across studies. Insomnia was defined differently from study to study, and sleep medication use was only limitedly considered.
• A pooled dementia risk. Only stroke and hospitalization could be pooled, and occupational injuries could not.
• Whether treating insomnia lowers stroke or dementia risk. The review did not test that.

An outside expert quoted by Healthline pointed to "significant variability among the studies" and said underlying conditions and sleep medications could affect the results. We also could not confirm whether the studies adjusted for confounders such as sleep apnea or depression, or how heterogeneous the stroke pooling was. Those are the first things to check in the full paper.

What it means for you

Don't read this as "bad sleep will give you a stroke." Read it as: insomnia is common, treatable and possibly a marker of brain or mental health trouble worth mentioning to a clinician. The authors themselves frame it as a warning sign, not a proven cause.

If sleep problems are persistent, the best-supported move is the one that already existed. The American College of Physicians recommends cognitive behavioral therapy for insomnia (CBT-I) as first-line treatment for chronic insomnia, ahead of sleep medication. A published review notes the American Academy of Sleep Medicine and the European Insomnia Guideline also endorse it.

That advice stands on its own merits, not on stroke or dementia fear. If you have other symptoms or risk factors, a clinician can look at the full picture.

Context

A prior meta-analysis estimated that 22% of the general population meets criteria for insomnia; we did not verify that figure. That scale is one reason loose causal language matters: a small statistical link in a very common condition can sound alarming without being actionable.

The authors call for shared diagnostic criteria, more research in under-studied regions, and analyses by outcome subtype. We did not find a statement from a health agency or professional society reacting to the review, and we did not independently check the earlier stroke and dementia meta-analyses it builds on.