There is an important nuance in the stroke data that I think gets lost when this topic is summarised too briefly: the risk is not the same across all severities of sleep apnoea. Mild OSA does not appear to significantly raise stroke risk. The risk becomes unmistakable in moderate and severe categories, and it is dose-dependent — meaning the worse the sleep apnoea, the greater the risk. This matters clinically because it means the finding is not simply a marker of general poor health. There is a clear biological gradient that points toward a direct mechanism.
The Primary Stroke Risk
A meta-analysis — a study that pools results from many separate pieces of research to get the most statistically reliable overall picture — covering 16 long-term studies with more than 24,000 individuals found that severe obstructive sleep apnoea more than doubles the risk of stroke. The relative risk (RR) was 2.15, meaning people with severe OSA were 2.15 times as likely to have a stroke compared with those without it.
To translate that into real-world numbers: for someone in their 50s or early 60s without other major risk factors such as hypertension, atrial fibrillation, or diabetes, the ten-year risk of having a stroke is roughly 1 in 25. With severe untreated sleep apnoea, the evidence puts that risk closer to 1 in 12. That is a meaningful difference in absolute terms — not just a percentage that sounds dramatic when stated one way but is trivial in practice.
The dose-response nature of the association is worth emphasising. Studies that have separated OSA patients by severity consistently find that mild OSA — defined as five to fifteen apnoea events per hour of sleep — does not produce a statistically significant increase in stroke risk compared with the general population. The signal emerges in moderate OSA and becomes strong in severe OSA. This is consistent with a mechanism driven by the severity of nocturnal hypoxia and autonomic nervous system dysregulation, rather than simply by the presence of any apnoea events at all.
The dose-response relationship is also one of the most convincing features of the evidence from a scientific standpoint. When risk increases proportionally with exposure — more severe the condition, higher the risk — that pattern is exactly what researchers look for when deciding whether an association is likely to be genuinely causal rather than coincidental.
For someone in their 50s or early 60s without other major risk factors, the ten-year stroke risk is roughly 1 in 25. With severe untreated sleep apnoea, the evidence moves that closer to 1 in 12.
Why Does Sleep Apnoea Raise Stroke Risk?
Several mechanisms operate simultaneously. Repeated overnight blood pressure surges, driven by the body's emergency alarm system activating during each apnoea event, promote damage to the vessel walls of the arteries supplying the brain and accelerate atherosclerosis (the furring-up of arteries) in the carotid and cerebral vessels. Arterial stiffness increases over time.
Atrial fibrillation, which is substantially more prevalent in OSA patients (covered in detail on the cardiovascular page), is itself a major stroke risk factor through its propensity to generate blood clots in the left atrial appendage — a small pouch in the heart from which clots can travel to the brain. The elevated AF risk from OSA therefore multiplies into elevated stroke risk.
Intermittent nocturnal hypoxia also promotes a prothrombotic state — meaning the blood is more likely to clot. Elevated fibrinogen, increased platelet aggregation, and raised inflammatory markers all shift the haemostatic balance toward clotting. The result is that OSA does not simply add one stroke risk factor. It simultaneously worsens several of the established risk factors through related but distinct pathways.
After a First Stroke: The Recurrence Picture
For people who have already experienced a stroke, the clinical stakes around sleep apnoea become considerably higher. Sleep apnoea is present in the majority of stroke patients — prevalence estimates range from 50 to 75 per cent depending on the population studied and the diagnostic threshold used. Most of it is undiagnosed at the time of the stroke and remains undiagnosed during rehabilitation. This represents a significant missed clinical opportunity.
The five-year risk of a second stroke in someone who has already had a first is roughly 1 in 4, depending on the type and severity of the initial event and what risk factor modification has been achieved. Co-existing OSA increases the odds of that recurrence.
The treatment signal here is among the strongest in the sleep apnoea literature. In patients with OSA who used CPAP consistently — defined as more than four hours per night — one pooled cohort analysis found the five-year recurrence risk reduced from roughly 1 in 4 to around 1 in 14. That represents a 72 per cent reduction in recurrent stroke, expressed as a relative risk of 0.37. A relative risk of 0.37 means the treatment group had only 37 per cent of the stroke recurrence rate seen in the non-treatment group. This is a dramatic and clinically important finding.
These are observational data rather than randomised trial results, so they require some caution in interpretation. But the magnitude of the signal and the biological plausibility of the mechanism make it difficult to dismiss as coincidence. The patients who used CPAP were not systematically healthier in other respects; adherence to CPAP was the key variable.
The five-year risk of a second stroke is roughly 1 in 4. In people with OSA who use CPAP consistently, that falls to around 1 in 14, based on the available evidence — a 72 per cent reduction in recurrence risk.
OSA After Stroke: A Missed Clinical Opportunity
Most stroke rehabilitation pathways in the UK do not routinely test for sleep apnoea. This is, I think, a significant clinical gap. A patient discharged from a stroke unit on antihypertensives and antiplatelets — blood pressure and blood-thinning medications — who also has untreated severe OSA repeatedly elevating blood pressure and heart rate throughout every night, is being treated for the downstream consequences of a problem while the upstream problem continues unchecked.
There are practical challenges: CPAP fitting during the acute post-stroke period is complicated by cognitive impairment, difficulties swallowing, and reduced ability to manage equipment. These are real barriers. But they do not justify not looking for the condition and not initiating a management plan during the rehabilitation period when patients have the most intensive clinical contact.
Why the Randomised Trial Evidence Is Less Clear
The honest picture on CPAP for primary stroke prevention in randomised controlled trials — the gold standard of medical evidence, where patients are randomly assigned to treatment or control — is mixed. The SAVE trial, one of the largest randomised trials of CPAP in patients with cardiovascular disease, did not find a significant reduction in stroke or cardiovascular death over a four-year follow-up. However, average CPAP use in the trial was only 3.3 hours per night, well below the four-hour threshold that appears necessary for meaningful physiological benefit. In analyses restricted to patients who used CPAP for more than four hours per night, the results moved in the expected direction.
This is a consistent pattern across the cardiovascular trial literature for CPAP: the treatment works in those who use it adequately. The challenge is that adequate use requires either tolerance of the device or alternative treatments that achieve equivalent airway protection. This is part of why surgical options and mandibular advancement devices exist and why they matter.
References
[1] Xie C-J et al. Association of obstructive sleep apnoea with the risk of vascular outcomes and all-cause mortality: a meta-analysis. BMJ Open. 2017. 16 cohort studies, 24,308 individuals; severe OSA associated with RR 2.15 for stroke (95% CI 1.42-3.24); risk dose-dependent; not significant in mild OSA.
[2] Lin Q-C et al. Effect of continuous positive airway pressure on the recurrence of atrial fibrillation and stroke risk in patients with obstructive sleep apnea: a systematic review and meta-analysis. Cardiovascular Therapeutics. 2020. Pooled cohort analysis; CPAP adherence (>4hr/night) associated with pooled RR 0.37 for recurrent stroke, representing 72% risk reduction.
[3] McEvoy RD et al. CPAP for prevention of cardiovascular events in obstructive sleep apnea. New England Journal of Medicine. 2016. SAVE trial; 2,717 patients with moderate-to-severe OSA and established cardiovascular disease; no significant reduction in primary endpoint; mean CPAP use 3.3 hrs/night.
[4] Johnson KG, Johnson DC. Frequency of sleep apnea in stroke and TIA patients: a meta-analysis. Journal of Clinical Sleep Medicine. 2010. Meta-analysis; prevalence of sleep-disordered breathing in stroke patients estimated 50-75% across studies.
Contact Professor Vik Veer
If you would like to discuss a sleep assessment or arrange a consultation, please use the details below.
Private secretary: 0207 458 4584