Sleep apnea quietly raises your risk of heart disease, stroke, and dementia — and most people who have it don't know. It is one of the few conditions where a free eight-question screener can change the trajectory of your health.

Sleep-disordered breathing

Obstructive sleep apnea (OSA) is the repeated partial or complete collapse of the upper airway during sleep. A global analysis estimated that 936 million adults aged 30–69 have mild-to-severe OSA, with 425 million in the moderate-to-severe range — and in some countries prevalence exceeds half the adult population.^[1] (That analysis was funded by ResMed, a manufacturer of sleep-apnea devices — the direction of the estimate is not in doubt, but the precise figures come from an interested party, and as of 2026 no independent estimate has replaced it.) The majority of cases are undiagnosed. Each apneic event causes a brief oxygen drop, a sympathetic-nervous-system surge, and a partial arousal that fragments sleep architecture; people with severe OSA can have 30 or more events per hour and never consciously notice. The downstream toll — cardiovascular disease, atrial fibrillation, resistant hypertension, type 2 diabetes, erectile dysfunction — falls quietly on what people otherwise experience as "I'm tired" or "I snore a bit." Treating it works, but only if it's identified first.

What OSA actually is (Strong)

OSA is mechanical airway collapse during sleep — the tongue base, soft palate, and pharyngeal walls fall in when the muscles that hold them open relax. Each collapse drops blood oxygen, triggers a sympathetic spike, and produces a micro-arousal the sleeper rarely remembers. The clinical metric is the apnea-hypopnea index (AHI) — the average number of complete (apnea) or partial (hypopnea) airway obstructions per hour of sleep.^[2]

Severity	AHI (events per hour)
Normal	<5
Mild	5–14
Moderate	15–29
Severe	≥30

The index counts events without weighting how deep or how long they are, and that limitation is not academic. In an observational analysis of adults with coronary disease, the total oxygen deficit accumulated overnight — the "hypoxic burden" — predicted cardiovascular events (about 87% higher risk in the worst third), while an AHI of 30 or more did not predict them at all.^[3] (That analysis was supported by the ResMed Foundation.) A larger cohort found the same: hypoxic burden and time spent below 90% oxygen saturation were the only sleep-study measures that tracked cardiovascular events.^[4] Mortality and event risk still climb with severity — but the count is a blunt instrument for locating it. Read the severity table as the clinical convention it is — the number your report will quote — not as a risk score.

The mortality signal is concentrated in severe disease. A meta-analysis of 27 cohort studies covering 3,162,083 participants found that severe OSA roughly doubled all-cause mortality (HR 2.13, 95% CI 1.68–2.68) and nearly tripled cardiovascular mortality (HR 2.73, 95% CI 1.94–3.85).^[5] The bracketed 95% confidence interval is the range in which the true effect most plausibly lies; when it crosses 1.0, the result is compatible with no effect at all. That matters here, because in the same analysis mild OSA (HR 1.19, 0.86–1.65) and moderate OSA (HR 1.28, 0.96–1.69) were both non-significant — their intervals cross 1.0. The mortality case for treating severe OSA is strong; the mortality case for treating mild OSA is not made by this data.

A separate meta-analysis of 18 prospective cohorts (over 25,000 participants, median follow-up 9 years) found the same severity gradient for incident cardiovascular disease: moderate OSA carried about 56% higher risk (HR 1.56, 1.20–2.03) and severe about 145% higher (HR 2.45, 1.85–3.25), while the mild-OSA estimate did not reach significance (HR 1.21, 0.98–1.50).^[6] In the same analysis, CPAP used at least 4 hours a night was associated with about 24% lower risk (HR 0.76, 0.60–0.96) — observational, but pointing the same way as the adherence data below.

The stroke signal is older and holds up. In an observational cohort of 1,022 patients undergoing sleep studies, OSA (defined as an AHI of 5 or more) was associated with roughly double the risk of stroke or death from any cause after adjustment for age, sex, smoking, body mass index, diabetes, hypertension, and atrial fibrillation (HR 1.97, 1.12–3.48).^[7]

The dementia signal is real but smaller than the cardiovascular one. In a 2025 meta-analysis of prospective cohorts, sleep apnea specifically was associated with about 33% higher risk of dementia and 45% higher risk of Alzheimer's disease, while the estimate for vascular dementia was compatible with no effect at all.^[8] These are observational cohorts sharing obesity and age as drivers, so read them as a reason to treat apnea, not as a demonstration that apnea causes dementia.

Why OSA accelerates aging at the cellular level (Moderate)

The damaging mechanism isn't the airway collapse itself — it's the intermittent hypoxia pattern of repeated oxygen drops followed by rapid re-oxygenation, hundreds of times per night, for years. The pathway below is well characterised mechanistically; what remains less certain is how much of the clinical risk it explains versus shared drivers like obesity.

Oxidative overload. The hypoxia-reoxygenation cycle mimics ischemia-reperfusion injury. Each cycle generates reactive oxygen species faster than the body's antioxidant enzymes can clear them. Adults with OSA show measurably elevated markers of oxidative damage to both fats and DNA in circulation.
A chronic inflammatory switch. Repeated transient hypoxia chronically activates hypoxia-inducible factor 1-alpha (HIF-1α), the cell's low-oxygen sensor, which cross-talks with NF-κB, the master regulator of inflammatory gene expression. The result is a steady release of inflammatory messengers — tumour necrosis factor alpha (TNF-α), interleukin-6 (IL-6), and interleukin-1 beta (IL-1β) — into the circulation, producing a low-grade chronic inflammatory state.
Endothelial dysfunction and vascular aging. The combined oxidative and inflammatory load strips arteries of nitric-oxide-mediated vasodilation and drives atherosclerosis. This is the mechanism behind the resistant hypertension phenotype that responds poorly to standard antihypertensives — OSA is the most consistently identified treatable cause of medication-resistant high blood pressure.

This is why OSA behaves like a multiplier on disease pathways downstream of inflammation and endothelial damage rather than a cause of one specific illness. See chronic inflammation for the broader mechanism.

Who should screen (Strong)

Major risk factors:

Male sex (though women are markedly underdiagnosed; postmenopausal women catch up to male prevalence)
Body mass index (BMI) ≥30
Neck circumference above 43 cm in men or 40 cm in women
Craniofacial anatomy — small or recessed jaw, narrow palate, large tongue, low-hanging soft palate
Family history of OSA
Alcohol or sedative use near bedtime
Hypothyroidism, acromegaly, polycystic ovary syndrome (PCOS)

Symptoms (any one warrants a workup):

Loud habitual snoring
Witnessed apneas — a bed partner reports gasping, choking, or pauses in breathing
Unrefreshing sleep, daytime sleepiness, falling asleep at the wheel
Resistant hypertension (blood pressure poorly controlled on three or more drugs)
Atrial fibrillation
Morning headache
Erectile dysfunction
Nocturia (multiple wake-ups to urinate)

The STOP-Bang questionnaire

The most validated free screening tool is STOP-Bang — eight yes/no questions, three or more positives flags high risk.^[9]

S — do you snore loudly?
T — do you feel tired during the day?
O — has anyone observed you stop breathing in your sleep?
P — high blood pressure (treated or untreated)?
B — body mass index over 35?
A — age over 50?
N — neck circumference larger than 40 cm?
G — male gender?

Three or more "yes" answers → ask for a sleep study. STOP-Bang is deliberately tuned to catch cases rather than to be precise. In general-population studies, three or more "yes" answers catch about 88% of moderate-to-severe apnea and 92% of severe apnea, but specificity is only about 40% — nearly three in five people who don't have apnea also screen positive. So a positive score means "get tested," not "you have apnea."^[10]

Home sleep testing is now first-line

Home sleep apnea testing (HSAT) is the standard first-line study for most adults without major medical comorbidity — a small device worn overnight at home, accurate enough to confirm or rule out OSA at a fraction of the cost of an in-lab study. Full in-lab polysomnography (PSG) is reserved for complex cases, suspected central sleep apnea, or when home testing is inconclusive. Because home tests can underestimate severity, a negative result in someone with strong symptoms should prompt an in-lab study rather than reassurance — this is a strong recommendation in the American Academy of Sleep Medicine (AASM) diagnostic guideline, not a hedge.^[11] There is a second reason to distrust a single negative night: apnea severity varies enough from night to night that in a study of 67,278 adults monitored at home for months, about one in five people diagnosed from a single night were misclassified.^[12]

Treatment: CPAP is first-line (Strong for symptoms; Moderate for cardiovascular outcomes)

Continuous positive airway pressure (CPAP) — a small pump that delivers a steady stream of pressurised air through a mask, mechanically splinting the airway open — is first-line for moderate-to-severe OSA.^[13]

Not all sleep apnea is obstructive (Moderate; Caution for adaptive servo-ventilation in heart failure)

In central sleep apnea the airway is open and the brain simply stops sending the signal to breathe. It looks similar on a home test and is treated differently — and one difference matters enormously. Adaptive servo-ventilation, a pressure device sometimes offered for central events, was tested in 1,325 patients with heart failure and a weak left ventricle: it did not reduce cardiovascular events, and it increased death from any cause by about 28%.^[14] A later trial with a different design did not reproduce the harm, so the picture is not settled.^[15] The practical point stands: adaptive servo-ventilation is not interchangeable with CPAP, and anyone with heart failure and predominantly central events needs a sleep physician.

The single biggest determinant of outcomes is adherence: at least 4 hours per night on at least 70% of nights. A substantial minority of patients struggle to reach this on their first attempt. Auto-adjusting (APAP) machines, well-fitted masks (nasal pillows, full face, hybrid), heated humidification, and proper clinician follow-up all improve adherence; if the first setup doesn't work, the answer is usually a different mask, not abandoning therapy.

One expectation worth correcting early: CPAP does not help you lose weight. Across 39 randomised trials, body mass index rose slightly on treatment — by about 0.15 kg/m² — and the rise was confined to people using the machine less than five hours a night, with no change above it.^[16] For obesity-driven apnea, the weight has to be addressed separately.

Adherent CPAP reliably reduces daytime sleepiness — about 2.4 points on the 24-point Epworth Sleepiness Scale, which measures how likely you are to doze off during the day, though only about 1 point in trials that recruited people who weren't sleepy to begin with. It also lowers ambulatory blood pressure: modestly in unselected patients, by roughly 1.5 mmHg systolic, but substantially in resistant hypertension, where it takes about 6 mmHg off the 24-hour systolic average.^[17]^[18] Crash rates fall too, though only in before-and-after comparisons rather than randomised trials, a design that flatters the result. Atrial fibrillation is the exception: treating sleep apnea after ablation did not reduce recurrence in the randomised trial that tested it — 57% in both arms.^[19]

The cardiovascular outcome picture is genuinely unresolved. The SAVE trial randomised 2,717 adults with moderate-to-severe OSA and established cardiovascular disease to CPAP plus usual care or usual care alone. Over a mean 3.7 years it found no reduction in cardiovascular events — 17.0% versus 15.4%, a difference well within the range expected from chance (HR 1.10, 0.91–1.32; P=0.34) — while significantly reducing snoring, sleepiness, and improving mood and quality of life.^[20]

Two features of SAVE limit how far the null result generalises. Mean adherence was only 3.3 hours per night, below the 4-hour threshold at which benefit is usually claimed; and participants were selected to have minimal sleepiness, so the trial cannot speak to symptomatic patients. The common rebuttal — that per-protocol analyses of participants who used CPAP for 4 or more hours do show benefit — is worth stating honestly: those comparisons are post-hoc and non-randomised, and adherent patients differ systematically from non-adherent ones in ways that independently predict better outcomes. The defensible reading is that CPAP has clear, well-demonstrated symptomatic benefits, and that its cardiovascular benefit is unproven in unselected patients. Two later analyses sharpen that reading without overturning it. Pooling individual participant data from the randomised trials, the result among everyone assigned to CPAP stays flat — no benefit — while a comparison restricted to people actually using the machine four or more hours a night finds about 31% fewer cardiovascular events. That comparison is still not randomised; it is the same adherent-patient comparison as before, done more carefully.^[21] A 2026 analysis across three trials found the benefit concentrated in patients who desaturate deeply rather than often, though those thresholds were drawn from inside the trials themselves and have never been validated elsewhere.^[22] (That analysis was part-funded by ResMed and the ResMed Foundation.) The live question is no longer "does it work" but "in whom."

Alternatives when CPAP isn't tolerated (Moderate)

Option	Best for	Notes
Mandibular advancement devices (MAD)	Mild-to-moderate OSA, recessed jaw	Custom-fitted oral appliances that pull the lower jaw forward; about 11 fewer events per hour than an inactive control; dentist-fitted is better than off-the-shelf^[23]
Positional therapy	Supine-dominant OSA	A simple back-prevention device (or a tennis ball in a pyjama pocket) for adults whose events cluster while sleeping on their back; about 7 fewer events per hour than an inactive control^[24]
Weight loss	Obesity-driven OSA	In a 690-adult cohort followed at 4-year intervals, a 10% loss of body weight predicted a 26% fall in AHI (95% CI 18–34%); a 10% gain predicted a 32% rise^[25]
GLP-1 receptor agonists	Obesity-driven OSA	SURMOUNT-OSA cut AHI by about 20 events per hour in adults not using CPAP and 24 in those who were, over 52 weeks — see GLP-1 receptor agonists^[26]
Hypoglossal nerve stimulation (Inspire)	Moderate-to-severe OSA in CPAP-intolerant adults meeting anatomical criteria	Implanted device activating the tongue muscle during inhalation; median AHI fell 68% at 12 months in a 126-patient single-arm, manufacturer-funded cohort^[27]
Surgery	Selected cases	Uvulopalatopharyngoplasty (UPPP) is variable; multilevel airway surgery has randomised support (below); maxillomandibular advancement is more effective but invasive; ear-nose-throat (ENT) evaluation required
Orofacial myofunctional therapy	Mild-to-moderate, adjunct to CPAP	See below — helps symptoms; the effect on apnea severity does not hold up in randomised trials
Drug therapy (AD109)	Not yet a general option	The first drug developed specifically for OSA cleared a phase 3 trial, but cut only about 4 events per hour^[28]

Four of those rows deserve more than a table cell.

Oral appliances. CPAP still beats them on apnea severity by roughly 8 events per hour, but patients wear them about 42 minutes a night longer — which may be why the blood-pressure effect comes out similar. In a head-to-head randomised trial, 24-hour blood pressure on an oral appliance was no worse than on CPAP.^[29] (The senior author of that trial holds a ResMed-endowed chair.) Positional therapy, by comparison, falls about 6 events per hour short of CPAP.

Hypoglossal stimulation. The 68% figure comes from an uncontrolled single-group study funded by the device manufacturer. The controlled portion of that trial — randomly withdrawing therapy from responders — is more informative, and it held up: AHI stayed at 7.6 events per hour in those who kept the device on versus 25.8 in those switched off. Better still, a later randomised trial gave the control group the implant with the stimulation left switched off, and found 58% of patients responding on active stimulation versus 14% with the device inactive — response meaning the apnea-hypopnea index at least halved, and fell below 20 events per hour. That isolates the effect of the stimulation from the effect of the operation.^[30]

Surgery. Multilevel upper-airway surgery beat medical management in a randomised trial of adults who had already failed conventional treatment, cutting the apnea-hypopnea index by about 18 events per hour and daytime sleepiness by nearly 7 Epworth points on the 24-point scale — though the authors call it preliminary.^[31]

Weight loss. A randomised trial has tested the lifestyle route directly: an 8-week diet, exercise and sleep-hygiene programme added to usual CPAP care produced about 24 fewer events per hour than CPAP alone, sustained at six months, alongside 7 kg of weight loss — an effect rivalling tirzepatide's. The trial was small, single-centre, open-label, and enrolled 89 men and no women.^[32]

The drug. AD109's roughly 4 events per hour is a fifth of what tirzepatide manages, it produced no improvement in fatigue, and one in five patients stopped because of side effects.

The nasal breathing physiology (Moderate for mechanism; Weak for longevity claims)

A theme in the modern OSA literature is that how you breathe outside of apneic events also matters. Nasal breathing is the physiological default, and chronic mouth breathing is associated with worse airway mechanics. The mechanistic evidence below is solid; the leap from these measurements to "nasal breathing extends lifespan" is not supported and is not made here.

Mechanics. Mouth breathing drops the jaw, lowers the tongue, and increases the gravitational pull on the soft palate — anatomically increasing the airway's tendency to collapse. This is why oral breathing during sleep tends to accompany more severe events.

Nitric oxide. The paranasal sinuses continuously produce nitric oxide (NO), a vasodilator the cardiovascular system depends on. Measured by chemiluminescence, the fraction of exhaled NO from the nose averages about 56 parts per billion, versus 14 from the mouth and 6 from the trachea.^[33] During nasal inspiration that NO is swept into the lungs, where it preferentially dilates blood vessels serving well-ventilated alveoli, improving the match between airflow and blood flow. Transcutaneous oxygen tension ran about 10% higher during nasal than oral breathing in six of eight healthy subjects, and in intubated patients — whose breathing tube bypasses the nose entirely — reintroducing their own nasally-derived NO into the ventilator circuit raised arterial oxygenation by 18% in all six tested; in a separate group of twelve short-term intubated patients, the 11% fall in pulmonary vascular resistance appeared in only four.^[34] Mouth breathing bypasses this NO reservoir. Note the scale of what has actually been demonstrated: acute, measurable effects on gas exchange. No trial has shown that habitual nasal breathing changes long-term health outcomes.

Autonomic tone. Nasal breathing engages the diaphragm and pulmonary stretch receptors, which feed the vagus nerve and increase parasympathetic ("rest and recover") tone, whereas mouth breathing tends toward shallow, upper-chest patterns. This is a reasonable mechanistic account rather than a demonstrated overnight effect.

Mouth taping — the honest assessment (Weak evidence; Caution)

The viral practice of taping the lips shut at night to force nasal breathing has the right physiological intent and a genuinely weak evidence base.

What the evidence actually consists of. A systematic review searched the literature from 1999 to 2024 and found just 10 studies totalling 213 patients. Only two showed a statistically significant improvement in established apnea measures such as AHI or oxygen desaturation; the rest showed no difference, and several discussed the risk of asphyxiation in the presence of nasal obstruction. Critically, many of those studies excluded anyone with nasal obstruction or nasal pathology — meaning the evidence base structurally cannot speak to the people most likely to be harmed. The authors concluded there is "a potentially serious risk of harm for individuals indiscriminately practicing this trend."^[35] As of 2026 no randomised trial of mouth taping has reported, and the only registered study is a single-group one that has not begun recruiting. No professional body has issued a graded guideline either — the AASM's public position is an expert statement calling the practice dangerous, which is a stance rather than evidence.

The positive signal, such as it is, comes from a small preliminary study in medically screened mouth-breathers with mild OSA and clear nasal passages, where porous tape reduced snoring and median AHI.^[36] That is the population the technique might suit — and it is a narrow one.

Why forcing mouth closure can backfire. This is the part the trend gets exactly wrong. A clinical trial using drug-induced sleep endoscopy in 54 OSA patients measured airflow directly while closing the mouth. Overall, mouth closure increased inspiratory airflow — but the effect split sharply by phenotype: airflow improved in patients with moderate mouth breathing, showed no change in those who barely mouth-breathe, and got worse in those with high levels of mouth breathing and in those with obstruction at the level of the soft palate. For those patients the mouth is not a bad habit; it is an essential bypass route around an obstructed nasal airway. The authors concluded that personalised assessment, not a blanket intervention, is what's needed.^[37]

That is the risk-asymmetric failure mode in one sentence: the people most likely to tape their mouths — habitual mouth breathers — overlap heavily with the people the procedure can hurt.

One legitimate clinical use is unrelated to the trend: patients on nasal CPAP masks who leak air through the mouth. Under clinical supervision, tape has been tested properly for this: in a randomised crossover trial of 62 patients, a month with tape versus a month without added about 52 minutes of CPAP use per night and raised the share of nights meeting the four-hour target by nearly 18 percentage points.^[38]

The right sequence for a mouth-breathing sleeper:

Get an ENT evaluation for nasal obstruction (deviated septum, polyps, turbinate hypertrophy, chronic rhinitis).
Get screened for OSA if any STOP-Bang risk is present.
Treat allergic rhinitis with intranasal corticosteroids or antihistamines.
Try side-sleeping first — reduces both snoring and AHI in supine-dominant OSA.
Consider orofacial myofunctional therapy if mouth breathing is muscular rather than structural.
Mouth taping only after the above are evaluated, and on a clinician's advice.

Orofacial myofunctional therapy (OMT) (Weak for apnea severity; Moderate for symptoms)

OMT is a daily exercise programme that retrains the tongue, lips, cheeks, and pharyngeal muscles to maintain a correct resting posture — lips sealed, tongue flat against the hard palate, breathing nasally.

The headline numbers are good and the evidence behind them is thin, and both halves of that sentence matter. A meta-analysis concluded that myofunctional therapy decreases AHI by approximately 50% in adults and 62% in children, and also improves lowest oxygen saturation, snoring, and daytime sleepiness.^[39] But that pooled estimate rests on nine adult studies totalling 120 patients — most of them uncontrolled before-and-after comparisons — and a paediatric arm of just two studies and 25 children. A 2025 network meta-analysis restricted to randomised trials — 15 of them, 473 adults and 139 children — revisited the question and did not reproduce the effect on apnea severity: the apnea-hypopnea index fell by 8.7 events per hour, but the range compatible with the data runs from a 21-event improvement to a 4-event worsening, so the result is indistinguishable from no effect.^[40] What did hold up in the same analysis was the symptom side: daytime sleepiness improved by about 3.5 points on the 24-point Epworth scale and sleep quality by about 2 points, both clearly beyond chance. A 2026 re-analysis of nine systematic reviews reached a similar conclusion — a pooled apnea-hypopnea reduction of about 9.5 events per hour and a gain of about 3 percentage points in lowest overnight oxygen saturation — while rating the methodological quality of most of the underlying reviews as critically low.^[41] These two are not independent confirmations of each other; they re-pool largely the same few hundred patients.

The sensible position: OMT is a reasonable adjunct to CPAP, mandibular devices, and post-surgical care — it is low-risk, and it may improve CPAP tolerance. It is not a substitute for first-line therapy, and on current randomised evidence it is a symptom treatment rather than an apnea treatment. Available through specially trained dentists, speech-language pathologists, and physical therapists.

Practical guidance

Take the STOP-Bang. If three or more answers are "yes," ask for a sleep study. Home sleep apnea testing is enough for most adults to confirm or rule out OSA.
If your bed partner reports gasping, choking, or pauses, go straight to the sleep study. Witnessed apneas are the single most specific symptom and shouldn't be ignored.
Consider OSA whenever you encounter resistant hypertension, atrial fibrillation, or unexplained morning headache. That is targeted case-finding in people who already have a reason to be tested, not population screening — the US Preventive Services Task Force rates screening of asymptomatic adults as insufficient evidence, and no randomised trial has shown it improves health outcomes.^[42] Testing a patient with resistant hypertension and loud snoring is a different thing, and it is sound.
If you're diagnosed, aim for at least 4 hours of CPAP per night on at least 70% of nights. Adherence is what determines whether the therapy does anything. If the mask isn't working, ask for a different one rather than giving up.
Lose weight if obesity is driving it. A 10% reduction in body weight predicts roughly a quarter lower AHI, and a structured diet-and-exercise programme does considerably better than that. GLP-1 receptor agonists are now an option for adults who qualify — see Ozempic-class drugs.
Sleep on your side if events cluster while supine. The tennis-ball trick works; commercial positional devices do too.
Address nasal obstruction before considering mouth taping. ENT evaluation, treat the rhinitis or septum, then re-evaluate.
Don't drink alcohol within three hours of bed if you have OSA — alcohol relaxes upper airway muscles and worsens events. See alcohol.
For mild OSA or as a CPAP adjunct, ask about orofacial myofunctional therapy — understanding that the evidence is preliminary.

What's overrated

Mouth taping as a first-line intervention without clinical evaluation. Ten studies and 213 patients is not an evidence base, and many of those studies excluded the people most at risk.
Consumer apps and watches as a substitute for a sleep study. The technology has genuinely improved — machine-learning analysis of breathing sounds now reaches around 90% sensitivity against a full sleep study in research settings.^[43] The sleep-apnea notification on some smartwatches has regulatory clearance. But read what those numbers cover: the watch feature is tuned to be almost never wrong when it does alert, at the cost of detecting only about 43% of moderate apnea, against 89% of severe. A silent watch is not a negative test. Use these to motivate a sleep study, not to replace one.
"Sleeping pills for snoring." Sleeping pills don't treat sleep apnea — they treat the symptom that would otherwise send you for a sleep study. The usual explanation for avoiding them turns out not to hold: pooling 27 randomised trials, prescription hypnotics made sleepers slightly harder to rouse but left apnea severity essentially unchanged, and did not weaken the tongue muscle's protective reflex.^[44] That evidence covers z-drugs and prescription hypnotics only — alcohol and opioids do relax the airway. The reason to avoid a sleeping pill in an undiagnosed snorer is masking, and that reason is sufficient. When insomnia and apnea genuinely coexist, the tested sequence is behavioural insomnia treatment first, then the apnea — see insomnia treatment.
CBD, magnesium, or melatonin as treatments for OSA. Useful for general sleep quality in some adults (see sleep supplements); no effect on the airway mechanics that define OSA.
Single-nostril breathing techniques as OSA therapy. Useful for autonomic training in waking hours; not a substitute for treating mechanical airway collapse during sleep.
Overnight supplemental oxygen instead of CPAP. Intuitive and tested: in a three-arm randomised trial, CPAP lowered 24-hour blood pressure while nocturnal oxygen did not differ from no treatment at all.^[45] Oxygen fixes the desaturation and leaves the arousals and the airway collapse untouched — which is a compact demonstration of what actually does the damage.
Treating mild OSA as though it carried the risk of severe OSA. In the largest mortality meta-analysis, neither mild nor moderate OSA reached statistical significance for all-cause mortality. Symptoms are still worth treating; the mortality argument is not the reason to do it.

Sleep-disordered breathing

Sleep-disordered breathing

What OSA actually is (Strong)

Why OSA accelerates aging at the cellular level (Moderate)

Who should screen (Strong)

The STOP-Bang questionnaire

Home sleep testing is now first-line

Treatment: CPAP is first-line (Strong for symptoms; Moderate for cardiovascular outcomes)

Not all sleep apnea is obstructive (Moderate; Caution for adaptive servo-ventilation in heart failure)

Alternatives when CPAP isn't tolerated (Moderate)

The nasal breathing physiology (Moderate for mechanism; Weak for longevity claims)

Mouth taping — the honest assessment (Weak evidence; Caution)

Orofacial myofunctional therapy (OMT) (Weak for apnea severity; Moderate for symptoms)

Practical guidance

What's overrated

Further reading

More in Sleep