Having insomnia alongside sleep apnea tracked 34% more dementia over five years

Sleep medicine has two large problems and tends to treat them in separate rooms. Apnea belongs to the machine and the mask. Insomnia belongs to the therapist and the sleep diary. A substantial number of people arrive with both, and get whichever they mentioned first.

Researchers writing in SLEEP asked what happens to that group. Among patients with newly diagnosed sleep apnea, those who also carried an insomnia diagnosis were compared against those who did not, matched one to one, and followed for five years.

Comorbid insomnia and sleep apnea was associated with an increased risk of all-cause dementia, by about a third.

Two conditions that make each other worse

Sleep apnea is a common disorder in which you have one or more pauses in breathing while asleep, sometimes hundreds of times a night, each one pulling the sleeper briefly toward waking.

Insomnia is the opposite complaint. Insomnia is a common sleep disorder in which people cannot fall asleep, cannot stay asleep, or wake too early and cannot get back.

Having both is not simply having two problems. The standard apnea treatment is a mask delivering pressurised air, and someone who already struggles to fall asleep tends to tolerate it badly. Meanwhile the fragmentation caused by apnea undermines the sleep-consolidation approaches used for insomnia. Each condition blunts the other’s treatment.

Why this comparison is a good one

The obvious way to study sleep and dementia is to compare bad sleepers with good sleepers, and it is close to uninterpretable, because bad sleep travels with obesity, depression, alcohol, medication and age.

Here both groups had sleep apnea, both had just been diagnosed with it, and both were matched on a long list of recorded characteristics. Whatever it is about having apnea that raises dementia risk is present on both sides of the comparison and cancels out.

What is left is the insomnia. Consistent risk elevations were observed across major dementia subtypes, with vascular dementia and Alzheimer’s disease both raised, and dementia with Lewy bodies raised most of all.

The detail that changes how to read it

One line in the results does more work than the headline figure. Risk separation became apparent within the first year.

Dementia is a disease of decades. Pathology accumulates for fifteen or twenty years before anybody notices a symptom, and nothing that happens to a person’s sleep in year one plausibly produces a diagnosis by year two.

A gap that opens immediately and stays open is the signature of reverse causation: the disease was already there, undiagnosed, and it was disturbing sleep. Insomnia would then be an early symptom being recorded as a risk factor.

The Lewy body result points the same way. Of all the dementias, that one is most strongly tied to sleep disturbance appearing years ahead of cognitive decline, and it produced the largest estimate here by a wide margin.

What the study is still good for

Reverse causation does not make a finding useless. It changes what the finding is for.

If insomnia alongside apnea is partly an early marker of a brain already in trouble, then it is exactly the sort of thing worth noticing in a sleep clinic. The authors put it in those terms: the results support the clinical relevance of recognizing COMISA as a high-risk clinical presentation.

A high-risk presentation is a reason for attention, not a reason for alarm. It says this group deserves more than treatment of whichever complaint they happened to raise first.

What it does not show

That treating insomnia protects the brain. Nobody was assigned to treatment here, and no comparison of treated against untreated is available.

The data are also diagnosis codes, so insomnia means someone wrote it down. People who report insomnia at an apnea appointment differ from people who have it and do not mention it, and that difference includes how closely they are engaged with medical care, which affects how quickly a dementia diagnosis gets recorded.

Five years is also short for this question, which is part of why the early separation is so hard to interpret.

The reasonable reading

Around a third more dementia over five years, in a group defined by a second sleep diagnosis, with the gap opening far too fast to be a straightforward causal story.

The most likely explanation is that some of this is the disease showing itself early through sleep, and some may be the genuine cost of years of badly broken nights. Nothing in this design can say how the total divides between the two.

What survives either way is the clinical point. Someone with both conditions is a person the sleep clinic should be looking at harder, and at present is often a person who gets sent home with a mask.