Does treating obstructive sleep apnoea with CPAP prevent a heart attack or stroke?

The answer is less certain than the usual explanation of sleep apnoea implies. Continuous positive airway pressure, or CPAP, is very effective at holding the upper airway open during sleep. It reduces breathing interruptions and improves daytime sleepiness for many people. Yet large randomised trials have not shown that prescribing CPAP reduces major cardiovascular events overall.

That is not an argument against treating sleep apnoea. It is a reason to be precise about which benefit has been proved, which remains possible and why the difference is hard to study.

Obstruction creates a plausible route to cardiovascular harm

Obstructive sleep apnoea occurs when the throat repeatedly narrows or closes during sleep. Airflow falls or stops, oxygen may drop, and the brain briefly arouses the sleeper so breathing can restart. This cycle can happen many times an hour without the person remembering it.

Those episodes produce intermittent low oxygen, changes in pressure within the chest, fragmented sleep and bursts of sympathetic nervous-system activity—the “fight or flight” signalling that raises heart rate and blood pressure. Over time, these stresses could contribute to hypertension and vascular disease. Observational studies also find more cardiovascular disease among people with obstructive sleep apnoea.

But a plausible mechanism and an association do not establish that one treatment will prevent a clinical event. People with sleep apnoea often differ in weight, metabolic health, alcohol exposure and other cardiovascular risks. Randomised trials are needed to isolate what CPAP itself changes.

CPAP fixes the obstruction while it is being used

A CPAP machine delivers gently pressurised air through a mask. The pressure acts like a pneumatic splint, keeping the upper airway open. It is a direct mechanical treatment, not a medicine circulating through the body.

The Australian Sleep Health Foundation describes CPAP as the most effective treatment for controlling obstructive sleep apnoea. Its immediate effects are clear: fewer obstructive events, less snoring and, for many patients, less daytime sleepiness. It only controls the airway while it is worn; stopping it allows the obstruction to return.

Nightly use is therefore part of the treatment dose. Mask leaks, dry nose or throat, nasal congestion, pressure discomfort and disturbed sleep can limit that dose. These are usually not dangerous adverse effects, but they can decide whether someone uses the machine for most of the night or removes it after a few hours.

The largest cardiovascular trial was neutral

The Australian-led SAVE trial enrolled 2,717 adults aged 45–75 who had moderate-to-severe obstructive sleep apnoea and established coronary or cerebrovascular disease. Most were men and had little daytime sleepiness. Participants were randomly assigned to CPAP plus usual cardiovascular care or usual care alone, then followed for an average of 3.7 years.

CPAP clearly treated the breathing disorder. The average apnoea–hypopnoea index—the number of breathing reductions or pauses per hour—fell from 29.0 to 3.7 events an hour in the CPAP group. Average use, however, was 3.3 hours a night.

The primary cardiovascular outcome included cardiovascular death, heart attack, stroke, or hospitalisation for unstable angina, heart failure or a transient ischaemic attack. It occurred in 229 people assigned CPAP (17.0%) and 207 assigned usual care (15.4%). The hazard ratio was 1.10, with a 95% confidence interval from 0.91 to 1.32. No individual cardiovascular outcome was significantly reduced.

CPAP still produced benefits participants could notice. It reduced snoring and daytime sleepiness and improved measures of mood and health-related quality of life. The trial therefore separated a strong effect on airway obstruction and symptoms from an unproved effect on cardiovascular events.

Four hours a night is an important clue, not a verdict

A 2023 analysis combined individual data from SAVE and two other randomised trials: 4,186 people with obstructive sleep apnoea and established cardiovascular disease. In the randomised comparison, CPAP again did not reduce recurrent major cardiac or cerebrovascular events: the hazard ratio was 1.01, with a confidence interval from 0.87 to 1.17.

The investigators then examined actual use. About 39% of participants assigned CPAP averaged at least four hours a night. After adjustment for differences that changed over time, this higher-adherence group had fewer recurrent cardiovascular events than people who used CPAP less: hazard ratio 0.69, with a confidence interval from 0.52 to 0.92.

That finding strengthens the possibility that adequate exposure helps. It does not prove it. Adherence was not randomly assigned. People who can use CPAP consistently may differ in symptoms, health behaviour or prognosis in ways that statistical adjustment cannot fully remove. The four-hour threshold is also a research definition, not a biological switch that guarantees protection.

Why the uncertainty remains

A 2024 systematic review examined 16 randomised trials involving 7,664 participants, alongside adjusted observational studies. The randomised evidence did not demonstrate reductions in all-cause death, cardiovascular death, heart attack, stroke or atrial fibrillation. The authors rated the evidence as low strength because trials were often too small or short, used different composite outcomes and enrolled selected participants.

SAVE, for example, studied secondary prevention in people who already had cardiovascular disease. It largely excluded those with severe sleepiness, for whom withholding effective symptom treatment would be difficult and potentially unsafe. Its result cannot settle primary prevention, highly symptomatic disease, different OSA patterns or near-all-night CPAP use.

For an Australian clinical discussion, the reason to treat should therefore be explicit. Symptoms, driving or workplace safety, severity of oxygen disruption, blood pressure, other medical conditions, treatment burden and patient priorities may all influence the decision. A promise that CPAP will prevent a heart attack should not substitute for that conversation.

The useful question is not simply, “Does CPAP work?” It is: which outcome are we trying to improve for this person, and how will we know whether the treatment is being used enough to achieve it?

Sources

Evidence note: This Insight is general education, not individual medical advice. It does not recommend starting, stopping or changing CPAP or another sleep-apnoea treatment. Diagnosis and treatment decisions should be made with a qualified clinician using current Australian guidance, the sleep-study findings and the person’s symptoms and health context.