Of all the health consequences of untreated sleep apnoea, the link with dementia is the one I find most patients have never considered. The idea that a problem with breathing during sleep could be quietly loading the brain with the proteins that define Alzheimer's disease is, I think, one of the more remarkable findings to emerge from sleep medicine in the past decade. And it is not a speculative finding. The evidence is substantial, the mechanism is understood, and there is a meaningful treatment signal.
I want to go through this carefully, because I am also conscious that the word dementia understandably provokes anxiety. My intention is not to alarm but to inform, because the most important implication of this evidence is not that sleep apnoea is dangerous but that it is treatable, and treatment may genuinely alter the trajectory.
The Numbers: What the Research Shows
Dementia affects roughly 1 in 11 people over the age of 65 in the United Kingdom. For those with untreated sleep apnoea, a 2025 analysis pooling the results of 39 separate long-term studies puts that risk closer to 1 in 8. That is the difference between roughly 9 people in a room of 100 older adults and roughly 12, with the additional 3 representing cases that could potentially have been prevented. For Alzheimer's disease specifically, where the lifetime risk after 65 is roughly 1 in 14, untreated OSA moves that closer to 1 in 10.
To explain what those numbers mean in research terms: these studies followed large groups of people over time and compared what happened to those with OSA against those without it. The hazard ratio — the measure researchers use to describe how much more likely an event is in one group compared with another — was 1.33 for all-cause dementia and 1.45 for Alzheimer's disease. A hazard ratio of 1.33 means you are 33 per cent more likely to develop the condition. When the baseline risk is 1 in 11, a 33 per cent increase means the new risk is roughly 1 in 8. The arithmetic is straightforward once you know what the numbers represent.
When you look at this relationship from the other direction, the numbers become even more striking. People with Alzheimer's disease are five times more likely to have OSA than cognitively healthy people of the same age. And in people who are already living with a diagnosis of Alzheimer's dementia, sleep apnoea is present in nearly 9 out of 10 cases — a prevalence of 89 per cent, compared with perhaps 10 to 15 per cent in the general older population. This is not a coincidental overlap. These two conditions share biology in a way that is now fairly well characterised.
In a room of 100 people over 65, around 9 will develop dementia. Fill that same room with people who have untreated sleep apnoea, and 3 more hands go up, quietly and preventably.
The Mechanism: The Brain's Overnight Cleanse
To understand why sleep apnoea might contribute to Alzheimer's pathology, you need to know about the glymphatic system. This is the brain's waste-clearance network, discovered relatively recently, which operates by using the flow of cerebrospinal fluid through channels alongside blood vessels to flush metabolic waste products out of brain tissue. Think of it as the brain's night-time cleaning crew, and slow-wave — or deep — sleep as the shift when that crew does most of its work.
Among the waste products the glymphatic system clears are amyloid-beta proteins, the same proteins that accumulate in the plaques that define Alzheimer's disease, and tau proteins, whose abnormal form characterises the neurofibrillary tangles of the same condition. When deep sleep is disrupted, glymphatic clearance is impaired. The evidence for this mechanism is robust in both animal research and human studies.
Sleep apnoea disrupts deep sleep repeatedly throughout the night. Every arousal from sleep to terminate an apnoea event fragments slow-wave sleep. In severe OSA, a patient may never achieve sustained periods of deep sleep at all. Over months and years, this creates a nightly failure to clear the proteins that, when they accumulate, drive neurodegeneration.
This is supported by direct biological evidence. A meta-analysis drawing on 18 studies involving 2,804 patients found that OSA patients have significantly higher blood levels of amyloid-beta and total-tau compared with healthy controls. These are the same biomarkers now being used in the most advanced Alzheimer's diagnostic tests. OSA is already shifting them in the wrong direction before any symptoms of dementia appear.
Cognitive Impairment Before Dementia
The dementia headline tends to dominate the conversation, but it is worth noting that the cognitive consequences of OSA begin well before any dementia diagnosis. A 2025 meta-analysis covering 33,226 individuals found that 37 per cent of OSA patients already have measurable cognitive impairment on standardised testing. In severe OSA, this rises to 44 per cent — nearly half. Women with OSA had higher rates than men: 59 per cent versus 47 per cent. These are not people with dementia; they are people who perform significantly worse on standardised tests of attention, working memory, executive function, and processing speed.
To make the cognitive impairment figures concrete: 37 per cent means more than 1 in 3 people with untreated OSA are already objectively impaired on neuropsychological testing. Most of them do not have a formal diagnosis of any cognitive disorder. They may simply be experiencing this as difficulty concentrating at work, losing words mid-sentence, or needing more time to process information than they used to.
A large study from the Sleep and Dementia Consortium — using data from nearly 6,000 adults assessed across five different research cohorts simultaneously — found that even mild OSA (defined as five or more apnoea events per hour of sleep) was associated with significantly worse global cognition scores over five years of follow-up. This matters because mild OSA is widely regarded as not requiring treatment. The cognitive evidence suggests that view may need revisiting.
One finding I think deserves particular emphasis: some cognitive deficits in OSA, particularly in sustained attention, may not fully reverse even after CPAP treatment is started. This is consistent with the idea that prolonged untreated OSA causes structural neurological changes that are not simply a reversible consequence of poor sleep. The implication is that earlier diagnosis and treatment matters, not just for how someone feels day to day, but for the long-term integrity of brain function.
Parkinson's Disease
The neurodegenerative link extends beyond Alzheimer's. A meta-analysis of 11 studies involving 1.3 million patients found that sleep apnoea is associated with a 54 per cent increased risk of Parkinson's disease. To put that in real-world terms: Parkinson's affects roughly 1 in 37 people over 65 in the UK. With untreated OSA, that rises to about 1 in 24. The mechanism here likely involves the accumulation of alpha-synuclein protein, driven by similar pathways of oxidative stress and impaired brain clearance. This is an area where research is still developing, but the signal is consistent and the biological rationale is coherent.
The Treatment Signal
This is where the evidence becomes genuinely hopeful. Of 11 studies examining the effect of positive airway pressure therapy on cognitive outcomes in OSA patients, 9 found a protective effect. That included evidence of delayed onset of mild cognitive impairment, reduced incidence of Alzheimer's disease, and slower rate of cognitive decline in those who were already showing early signs of cognitive change.
A meta-analysis of 14 randomised controlled trials — these are the highest quality trials in medicine, where patients are randomly assigned to treatment or control to avoid bias — found that CPAP treatment improved attention and processing speed significantly in patients with severe OSA. The effect was consistent and clinically meaningful. For those already living with Alzheimer's and OSA, consistent CPAP use appears to improve sleep architecture, reduce daytime sleepiness, and slow the rate of cognitive decline compared with those who do not use it.
I want to be direct about the limitations here. This is not a situation where we can guarantee that treating sleep apnoea will prevent dementia. The evidence is not of that quality. What we can say is that there is a plausible mechanism, a consistent epidemiological association, and a treatment signal robust enough that treating OSA is a reasonable component of any strategy to protect cognitive function as we age. Given that the treatment is safe and has benefits across multiple other organ systems, there is no reason not to treat it.
A Note on Older Women Specifically
The evidence increasingly highlights post-menopausal women as a particularly high-risk group where OSA is simultaneously underdiagnosed and cognitively consequential. OSA is classically associated with middle-aged overweight men, which means women are less likely to be referred for sleep studies, less likely to be diagnosed, and less likely to receive treatment. Yet in the cognitive data, women with OSA consistently show higher rates of cognitive impairment than men with equivalent disease severity. This diagnostic gap represents, I think, one of the more significant unaddressed problems in sleep medicine in the UK at the moment.
References
[1] Ungvári Z et al. Sleep disorders increase the risk of dementia, Alzheimer's disease, and cognitive decline: a meta-analysis. GeroScience. 2025. 39 cohort studies; Hazard Ratio for all-cause dementia 1.33 (95% CI 1.09-1.61), for Alzheimer's disease 1.45 (95% CI 1.24-1.69) in OSA patients.
[2] Emamian F et al. The association between obstructive sleep apnea and Alzheimer's disease: a meta-analysis perspective. Frontiers in Aging Neuroscience. 2016. Meta-analysis; patients with Alzheimer's disease five times more likely to have OSA than cognitively healthy controls.
[3] Aini N et al. Prevalence of sleep-related breathing disorders and associated risk factors among people with dementia: a meta-analysis. Sleep Medicine. 2023. 20 studies, 1,461 participants; pooled prevalence of OSA in Alzheimer's dementia 89%.
[4] Yeo BS et al. The association of obstructive sleep apnea with blood and cerebrospinal fluid biomarkers of Alzheimer's dementia: a systematic review and meta-analysis. Sleep Medicine Reviews. 2023. 18 studies, 2,804 patients; OSA patients have significantly higher blood amyloid-beta-40, amyloid-beta-42, and total-tau levels.
[5] Su K et al. Prevalence of cognitive impairment among adults with obstructive sleep apnea: a systematic review and meta-analysis. Sleep and Breathing. 2025. 23 studies, 33,226 individuals; pooled prevalence of cognitive impairment 36.92% in OSA; 44.46% in severe OSA; women 59% vs men 47%.
[6] Pase MP et al. Sleep architecture, obstructive sleep apnea, and cognitive function in adults. JAMA Network Open. 2023. Sleep and Dementia Consortium, 5,946 adults from 5 US population cohorts; mild to severe OSA associated with significantly worse global cognition over 5 years.
[7] Shieu MM et al. Positive airway pressure and cognitive disorders in adults with obstructive sleep apnea. Neurology. 2022. Systematic review, 11 studies (60,840 participants); PAP protective in 9 out of 11 studies, delaying MCI onset and slowing Alzheimer's progression.
[8] Guay-Gagnon M et al. Sleep apnea and the risk of dementia: a systematic review and meta-analysis of cohort studies. Journal of Sleep Research. 2022. 11 studies, 1,333,424 patients; Hazard Ratio 1.54 for Parkinson's disease in OSA patients.
[9] Wang MY et al. Cognitive effects of treating obstructive sleep apnea: a meta-analysis of randomized controlled trials. Journal of Alzheimer's Disease. 2020. 14 RCTs, 1,926 participants; CPAP significantly improved attention and information processing speed in severe OSA patients.
Contact Professor Vik Veer
If you would like to discuss a sleep assessment or arrange a consultation, please contact us using the details below.
Private secretary: 0207 458 4584