Insomnia and anxiety together were linked to 54% higher dementia risk in a UK Biobank study

Poor sleep and worry travel together. Someone who lies awake often lies awake anxious, and people with anxiety disorders frequently sleep badly. Each has been tied to dementia in earlier research, usually studied one at a time, as though the other were not in the room.

A team in Shanghai led by Yan Chen and Xia Li has now looked at them as a pair. Their study, published in Alzheimer’s & Dementia in September, drew on 65,857 older adults in Shanghai and 149,467 in the UK Biobank. In both places the extra risk sat with people who had insomnia and anxiety together. A larger American study published three months earlier points somewhere else, and the disagreement is worth as much attention as the finding.

What did the insomnia and anxiety study find?

The Shanghai arm was a snapshot. Adults aged 60 and over in 16 communities filled in standard questionnaires about sleep and anxiety and took cognitive tests. Baseline comorbid insomnia and anxiety were reported in 1592 people, about 1 in 40. Among them, roughly two in three had cognitive impairment. Among everyone else it was fewer than one in three.

A snapshot cannot say which came first, so the researchers turned to the UK Biobank, where people were enrolled between 2006 and 2010 and then tracked through national health records. Here the pattern appeared over time. The cumulative incidence of dementia over a follow-up of up to 15 years was higher in people who started out with both conditions, by about half.

The surprise was what did not show up. In both cohorts, neither isolated anxiety nor isolated insomnia was clearly tied to worse outcomes. Statisticians call this kind of result an interaction: two factors together doing more than the sum of each alone. In the UK data it held for Alzheimer’s disease and was at least as strong for vascular dementia, the kind caused by damaged blood supply to the brain.

Why would insomnia and anxiety together matter more than either alone?

For a clue, the Shanghai team measured blood markers in a randomly chosen group of 3,055 participants. One was p-tau217, a fragment of the tau protein. A JAMA Neurology study describes phosphorylated tau as a specific blood biomarker for Alzheimer disease (AD) pathology, with p-tau217 considered to have the most utility. The other was GFAP, a protein released by the brain’s support cells when they are under strain.

Both were higher in people with the combined condition and in nobody else. In the authors’ analysis, p-tau217 and GFAP mediated this association, meaning that part of the statistical link between the combination and cognitive impairment ran through those markers. That fits a story in which chronic sleeplessness and anxiety together speed up the biology of Alzheimer’s.

It fits a second story just as well, in which the illness comes first.

A larger US study found a different pattern for anxiety and insomnia

In June, researchers led by SangNam Ahn published an analysis of electronic health records for 1,868,790 American adults over 50, followed for up to ten years. They sorted people into the same four groups: neither condition, anxiety only, insomnia only, both.

Their ranking came out differently. Anxiety alone went with a clearly higher rate of dementia diagnosis. Insomnia alone went with a slightly higher one. And the combination was no worse than anxiety by itself. The authors put it bluntly: co-occurrence does not confer additional risk beyond anxiety alone. They add that insomnia’s contribution is very modest compared to the primary association demonstrated by anxiety.

Both studies are large and both adjusted for age, sex and other illnesses. Where they differ is in how the conditions were identified. The American study relied on diagnostic codes entered by clinicians. The UK arm of the Shanghai study used codes too but added people who simply reported usually sleeping badly, and the Shanghai arm used questionnaires. A person can score as anxious on a questionnaire without ever receiving a diagnosis, and the reverse. Different nets catch different fish.

Older pooled evidence sits somewhere between the two. A meta-analysis of 18 long-term studies found that subjects who reported sleep disturbances had a higher risk of incident all-cause dementia, with insomnia specifically linked to Alzheimer’s disease. A separate meta-analysis concluded that anxiety is associated with an increased risk for cognitive impairment and dementia in the community. Neither tested the two together.

Anxiety and poor sleep can also be early signs of dementia

The changes in the brain that lead to dementia begin many years before a diagnosis, and they can disturb sleep and mood along the way. If that is what is happening, insomnia and anxiety are early symptoms and not causes, and treating them would not change the outcome.

The anxiety meta-analysis raised exactly this. Its authors found that stronger associations were driven by higher age, suggesting that it is a prodromal symptom, the medical term for an early warning sign of a disease already under way. The blood results from Shanghai are open to the same reading: people whose brains already carry more Alzheimer’s pathology may be the ones who become both anxious and sleepless.

The Shanghai authors acknowledge several gaps. The biomarker levels in the SHAPE cohort were measured at only a single time point, so they cannot show the markers rising after the symptoms began. Details of sleeping pills and anti-anxiety drugs were missing, and objective sleep measures were not available. The earlier sleep meta-analysis had the same weakness, noting that sleep disturbances were evaluated mainly based on self-reports.

The combination is also uncommon. About 2% of older adults in each cohort had both conditions, and by the authors’ estimate it accounted for under 2% of dementia cases in the UK data.

For readers who have both problems, the study adds a reason for researchers to take the pairing seriously and no evidence yet that treating either one lowers dementia risk. Both conditions are treated in their own right for the distress they cause now, which is a separate matter from what they may or may not signal later.

Two large studies agree that anxiety and poor sleep travel with dementia, and they disagree on whether the pair is worse than its parts.