Obstructive sleep apnoea and obesity hypoventilation

• Obstructive sleep apnoea (OSA) is repeated collapse of the upper airway during sleep. Each collapse ends breathing for at least 10 seconds, oxygen falls, carbon dioxide rises, and the brain briefly rouses the patient to reopen the airway.
• Severity is graded by the apnoea-hypopnoea index (AHI): the number of apnoeas plus hypopnoeas per hour of sleep.
• Obesity is the strongest modifiable risk factor, but a small jaw, large tongue, nasal obstruction, male sex, menopause, hypothyroidism and acromegaly all predispose.
• Diagnosis usually begins with a home sleep apnoea test and is confirmed or restudied in the sleep laboratory with polysomnography when the result is unclear or the patient is complex.
• Untreated OSA raises blood pressure, triggers arrhythmia such as atrial fibrillation, increases stroke risk, can cause pulmonary hypertension, and markedly raises the risk of falling asleep at the wheel.
• Continuous positive airway pressure (CPAP) is first-line treatment; weight loss, oral appliances, positional therapy, hypoglossal nerve stimulation, surgery and tirzepatide are important alternatives or add-ons depending on severity and tolerance. ^[5]
• Obesity hypoventilation syndrome (OHS) is a separate but overlapping problem: obesity, daytime hypercapnia with a PaCO2 above 45 mmHg, and sleep-disordered breathing after other causes of hypoventilation have been excluded. It is usually treated with non-invasive ventilation (BiPAP) and weight loss. ^[1]

By the end of this page you should be able to:

1. Define OSA and OHS and distinguish them from central sleep apnoea.
2. Explain why the pharynx collapses and how collapse leads to intermittent hypoxia, hypercapnia and autonomic surges.
3. Grade OSA severity with the AHI and use the STOP-BANG questionnaire sensibly.
4. Choose between a home sleep apnoea test and full polysomnography.
5. Plan management: weight loss, CPAP, oral appliances, positional therapy, hypoglossal nerve stimulation, surgery, tirzepatide.
6. Recognise the peri-operative risks of obesity hypoventilation and start the right ventilatory support.
7. Counsel a patient about driving and occupational safety with untreated OSA.

OSA is the commonest form of abnormal sleep pattern. It is characterised by narrowing or complete closure of the upper airway during sleep. The brain's attempt to breathe against a closed tube produces the classic cycle: airflow stops, oxygen falls and carbon dioxide rises, the autonomic nervous system surges, then a brief arousal reopens the airway. The patient usually remembers nothing. By contrast, breathing effort never stops in OSA; the effort is preserved while flow ceases, which is what separates it from central sleep apnoea. ^[3][6]

The diagnosis of OSA needs either (i) nocturnal breathing disturbances with an apnoea-hypopnoea index (AHI) of at least 5 events per hour, or (ii) an AHI of 15 events per hour or more even without symptoms. ^[3] Most adults experience a few apnoeas per hour; symptoms and consequences dominate the decision to treat once the threshold is crossed.

OHS is defined as the combination of three findings: a body mass index (BMI) of at least 30 kg per m2, daytime hypercapnia with a PaCO2 above 45 mmHg, and sleep-disordered breathing, after other respiratory disorders that can cause alveolar hypoventilation have been ruled out. ^[1] It is observed in 1 to 10 percent of obese people, is often associated with OSA, and is frequently diagnosed only when the patient presents with an acute exacerbation of respiratory failure. ^[1]

A screening blood bicarbonate gives a useful clue. In the ATS 2019 guidance reviewed in that work, a serum bicarbonate below 27 mmol per litre effectively rules out hypercapnia, with a negative predictive value of 99% (95% CI 97.9 to 99.6%) when the expected prevalence is up to 20%. A bicarbonate of 27 mmol per litre or more in a patient with a BMI of at least 35 kg per m2 is a prompt to check an arterial blood gas and, if hypercapnia is confirmed, a nocturnal study with transcutaneous carbon dioxide monitoring. ^[1]

OHS accounts for up to roughly 0.4% of the general population in the USA, and about 24% of obese patients with OSA and a BMI above 40 kg per m2 meet criteria. ^[1]

OSA is extraordinarily common: about 1 billion people worldwide are affected, of whom roughly 936 million adults aged 30 to 69 years have the condition and 425 million have moderate to severe disease. ^[5][2] It is underdiagnosed in everyday practice. ^[3] Obesity prevalence reached 890 million adults worldwide in 2022 (16% of adults), and 40% of adults with a BMI of at least 30 kg per m2 suffer from OSA, rising to 98% in the morbidly obese. ^[1][3] The underdiagnosis is worst in people who never complain of snoring because they live alone.

The fundamental risk is an upper airway that is narrow, collapsible or poorly supported. The factors below make that more likely, often in combination.

• Obesity. Abdominal obesity, more than general obesity, is tightly linked to OSA. Excess fat narrows the pharynx directly, alters airway muscle structure, and reduces functional residual capacity through a heavier chest wall. Up to 98% of the morbidly obese population has OSA. ^[1][3]
• Neck size. A neck circumference above 43 cm in men and above 37 cm in women is strongly related to OSA, even after adjusting for BMI. ^[3]
• Craniofacial anatomy. A small or posteriorly placed mandible, macroglossia, a crowded oropharynx, nasal obstruction and a long soft palate all narrow the airway. These matter particularly in non-obese patients.
• Male sex. Men are affected more often; women's risk rises sharply after the menopause as upper-airway tone and fat distribution change.
• Hypothyroidism. Mucopolysaccharide deposition in the airway soft tissues and macroglossia increase collapsibility; correcting the hormone deficiency can reduce the AHI.
• Acromegaly. Soft-tissue overgrowth, macroglossia and prognathism enlarge and destabilise the upper airway.
• Alcohol and sedatives. These reduce upper-airway dilator tone and blunt the arousal response. Opioids are particularly dangerous: in the peri-operative setting acute opioids can disable the normal protective arousal response to hypoxia. ^[6]
• Smoking. Mucosal inflammation and swelling narrows the airway.
• Genetics and family history. Narrow airways and obesity run in families.

One can think of the patient as having a fixed anatomic predisposition (airway size, craniofacial shape) plus a variable load (fat, fluid, posture, drugs). This is also why a patient may be well until they gain 5 kg.

The story begins with a pharynx that is a soft, collapsible tube. During wakefulness, upper-airway dilator muscles (genioglossus, tensor veli palatini, and others) hold the lumen open. In sleep these muscles lose tone, more so in REM than in non-REM sleep, and the lumen narrows. In susceptible people the negative pressure generated by inspiration sucks the walls together. That is the collapse. ^[3][5]

Obesity increases load on the system in several ways: fat deposits in the pharyngeal wall and tongue, the abdominal mass pushes the diaphragm cephalad and lowers functional residual capacity, and tracheal traction that stiffens the airway is reduced. About 80% of the pharyngeal airway is soft tissue, so even modest gains in weight make collapse more likely. Fluid that pools in the legs by day can also shift rostrally at night and cause pharyngeal oedema, particularly in heart failure. ^[3][5]

Once the airway closes, oxygen falls and carbon dioxide rises. The accumulating hypercapnia and hypoxia eventually trigger an arousal (a micro-arousal on the EEG) that reactivates the dilator muscles. The airway reopens, often with a snort or a gasp, flow resumes, then, as sleep deepens again, the cycle repeats. Tens or hundreds of times per hour, the patient never knows it is happening. The repeated desaturation-reoxygenation produces oxidative stress and sympathetic activation, which damage blood vessels, disturb glucose metabolism and promote weight gain in a vicious cycle. ^[3]

The arousal threshold, the upper-airway (plant) gain and the loop gain of the ventilatory control system together determine who develops OSA for a given airway. This is why two people of the same size may differ: one has a stable, non-collapsible airway and a high arousal threshold, the other does not.

In OHS the problem is not only obstruction. Obese subjects breathe shallowly and rapidly because lung compliance is reduced and work of breathing is high; the excess weight compresses the chest and lowers functional residual capacity. Ventilatory drive is blunted in many patients by leptin resistance or by the kidney's compensation of the chronic respiratory acidosis that accompanies very frequent apnoeas. Respiratory muscle strength is seldom low enough to explain the hypoventilation on its own. The result is nocturnal alveolar hypoventilation, then daytime hypercapnia. OSA adds to this by providing too little time between apnoeas for carbon dioxide to be cleared. ^[1]

OSA does not leave a single diagnostic pathology specimen. The real lesions are functional: negative-pressure injury of the soft-palate and uvula mucosa, reflex bradykinin-related oedema of the upper airway, and hypertrophy of the tonsils and adenoids in children. Sleep endoscopy (drug-induced sleep endoscopy) is often used before upper-airway surgery to watch where the airway collapses. In OHS, autopsy and lung studies in general show atelectasis and small-airway closure in the dependent lung, but the diagnosis is physiological, not histological.

History in the patient's own words matters. Useful clues include:

• loud snoring, often reported by a bed partner
• witnessed pauses in breathing, gasping or choking
• waking unrefreshed, morning headaches, dry mouth
• somnolence: nodding off in meetings, while watching television, or while driving
• nocturia (negative intrathoracic pressure during obstructed breaths and hypoxia promote atrial natriuretic peptide release)
• heartburn, sweating at night, erectile dysfunction
• mood change, poor concentration, memory problems

Examination may show obesity (BMI of at least 30 in roughly 4 in 10 patients, and almost everyone who is morbidly obese), a crowded oropharynx, a large tongue or uvula, a recessed chin, nasal obstruction, and systemic hypertension. Neck circumference above 43 cm in men and 37 cm in women is a particularly good clinical screen. ^[3][4]

The Epworth Sleepiness Scale is a simple questionnaire that asks how likely the person is to doze in eight everyday situations. It is useful for quantifying sleepiness, and a higher score indicates more daytime sleepiness. It misses people with OSA who are not sleepy, so it must never be used alone to exclude the diagnosis. ^[6]

Resistant hypertension is a strong clue. OSA is present in a large proportion of patients whose blood pressure remains high despite several drugs including a diuretic, because each apnoea showers the circulation with catecholamines.

STOP-BANG is an eight-item yes/no screen that is widely used, with snoring, Tiredness, Observed apnoea, high blood Pressure, BMI above 35 kg per m2, Age above 50 years, large Neck circumference, and male Gender (the "BANG" tail). Each yes scores a point, to a maximum of 8. A score of 3 or more flags a high probability of OSA, and higher scores correlate with more severe disease; in the cited cohort a cut-off of 3 gave a sensitivity of 76% and a specificity of 56%. It performs less well in non-obese, younger or female patients, and in people with cluster headache, so a normal score never excludes OSA in a symptomatic patient. ^[4]

The conditions that mimic or accompany OSA are tabulated below. The key distinction is whether respiratory effort is preserved (obstructive) or absent (central).

Home sleep apnoea test (HSAT) is the usual first step in an otherwise healthy adult with a high pre-test probability. It uses nasal airflow, respiratory effort bands and oximetry, and is reported as the respiratory disturbance index (RDI) or AHI. It is cheaper and faster than laboratory testing, but it does not measure sleep stages, it cannot reliably distinguish central from obstructive events, and it can underestimate severity in mild disease or when the patient sleeps badly on the night. ^[3]

Polysomnography (level 1, in-laboratory) is the gold standard. It records brain activity (EEG), eye movements, chin and leg electromyography, airflow, effort, oxygen saturation, position and heart rhythm. It is required when HSAT is negative but suspicion is high, when the patient has cardiopulmonary or neuromuscular disease, when central sleep apnoea or hypoventilation is suspected, for titrating positive airway pressure, and for any peri-operative assessment. ^[1][3]

An arterial blood gas and a serum bicarbonate are the cheap keys to excluding OHS. Chest radiograph, ECG and routine blood tests may reveal consequences (right heart strain, polycythaemia, diabetes) but are not diagnostic.

Severity is graded by the AHI measured (or estimated) from the study, using the conventional thresholds below. ^[2]

Two further points are worth knowing. First, the extent of oxygen desaturation matters as well: severe disease is sometimes described as an AHI above 40 events per hour and/or a minimum oxygen saturation below 65%. ^[3] Second, the AHI is only a number; treatment decisions in a symptomatic patient should also consider blood pressure burden, sleepiness, occupational risk and comorbidity, and a patient can have significant disease with an AHI between 5 and 15 if they are very sleepy or have cardiovascular risk.

A patient presenting with acute hypercapnic respiratory failure from OHS or severe OSA should be recognised early. In less severe cases, non-invasive ventilation (NIV, BiPAP) is first-line; in severe encephalopathy, pneumonia or multi-organ failure, intubation and critical care admission are needed. Obesity makes intubation and extubation high-risk (difficult laryngoscopy, reduced oxygen stores, post-extubation obstruction and aspiration risk), so sedation should be minimised, weaning be rapid, and a cuff-leak test considered before extubation. Patients with acute-on-chronic failure should be discharged with PAP and reassessed. ^[1]

Weight management is foundational. Meaningful sustained weight loss lowers AHI, blood pressure and sleepiness. Lifestyle measures and dietary advice come first; pharmacotherapy is considered when BMI is 27 kg per m2 or more and weight has not improved with lifestyle change; bariatric surgery is considered when BMI is 35 kg per m2 or more in a patient with OSA. ^[3][5]

CPAP is first-line therapy for OSA. It acts as a pneumatic splint, holding the pharynx open; it improves symptoms, lowers blood pressure and improves quality of life. ^[1][3][5] Nasal interfaces are usually best tolerated. Adherence is the main challenge: up to 80% fail to adhere in some series. ^[3] CPAP must never be stopped abruptly, because in 5% of patients central or complex apnoea emerges on treatment ("treatment-emergent" CSA), which usually resolves within 30 days. ^[6]

Conservative measures for every patient: weight reduction, exercise, avoiding alcohol at night, keeping the nose patent, and avoiding sleep deprivation. ^[5]

Oral appliances (mandibular advancement devices) pull the jaw forward and reduce airway collapse. They suit mild to moderate OSA, and leaner, younger patients tolerate them best. CPAP reduces the AHI more, but patients wear oral appliances more reliably, so overall effectiveness can be similar. ^[5]

Positional therapy helps when apnoeas occur mainly while supine. Side-sleeping, sometimes enforced by a backpack, a pillow or a positional alarm, can halve the AHI in selected patients.

Hypoglossal nerve stimulation was approved in 2014. A small implanted chest generator detects breathing and stimulates the hypoglossal nerve to keep the tongue protruded. In the STAR study about 70% of moderate-to-severe OSA patients who could not tolerate CPAP achieved a clinically significant improvement in AHI, using selection criteria that included a BMI below 32 kg per m2 and the absence of complete concentric airway collapse on drug-induced sleep endoscopy. ^[5] There are few data in people with heart failure.

Upper airway surgery is an option for selected anatomy. Adenotonsillectomy benefits children and a few adults; uvulopalatopharyngoplasty gives a "good response" (AHI under 20 per hour with at least 50% improvement) in roughly 40% of patients; maxillomandibular advancement and multilevel procedures help highly selected patients with skeletal narrowing. ^[5]

Tirzepatide, a once-weekly GLP-1/GIP receptor agonist, has been approved by the FDA for moderate-to-severe OSA in people with obesity on the basis of the SURMOUNT-OSA trial. In both trial arms (with or without positive airway pressure) tirzepatide was superior to placebo for the reduction of AHI, lowered systolic blood pressure by up to 9.5 mmHg, and produced 18 to 20% body weight loss over a year. ^[5]

What to avoid. Alcohol, hypnotics, benzodiazepines and opioids before bed all worsen airway tone and the arousal response. In OSA patients on opioids, central apnoeas are common and the protective response to hypoxia can be disabled. ^[6]

Objective adherence monitoring, repeat sleep study if symptoms change, and encouragement of even modest weight loss. In OHS, step up or step down between CPAP and NIV according to follow-up blood gas and sleep study. For patients with a recent stroke, treat the OSA where possible. ^[6]

Untreated OSA is strongly associated with hypertension, cerebrovascular disease (including stroke), coronary artery disease, heart failure, atrial fibrillation, pulmonary hypertension, type 2 diabetes, non-alcoholic fatty liver disease, chronic kidney disease and all-cause mortality. ^[3][4] Accidents from falling asleep at the wheel are a major cause of death and disability in people with untreated OSA; fatigue-related impairment is a recognised hazard of untreated disease. OSA patients falling asleep while driving is a well-recognised and avoidable tragedy. ^[3]

OSA and metabolic syndrome travel together. One pooled analysis found people with OSA nearly three times more likely to have metabolic syndrome than controls, and those with moderate-to-severe OSA were about 2-fold more likely to have hypertension and more likely to have hyperglycaemia than those with mild disease. ^[2]

OHS carries a high morbidity and mortality when untreated, and nearly half of patients needing long-term NIV have their first set-up during an acute admission. ^[1]

With adequate treatment (CPAP or NIV, weight loss, avoidance of sedatives) symptoms and blood pressure usually improve within weeks, and neurocognitive function over months. Many patients report feeling like themselves again for the first time in years. Untreated severe OSA shortens life and OHS even more so: untreated OHS is associated with an important increase in morbidity and mortality. ^[1]

Routine universal screening is not advised, but a low threshold is justified in people with hypertension, resistant hypertension, obesity, type 2 diabetes, heart failure, atrial fibrillation, stroke, pulmonary hypertension and in professional drivers. Neck circumference and a brief STOP-BANG screen are cheap and effective first steps. Weight management and avoiding sedatives remain the main preventive levers. ^[3][5]

Pregnancy. OSA risk rises with weight and gestational fluid shifts. Screening in high-risk pregnancies (obesity, hypertension, gestational diabetes) and CPAP when positive can improve outcomes; weight loss is not appropriate in pregnancy.

Children. OSA in children is usually from adenotonsillar hypertrophy rather than obesity; adenotonsillectomy is the main treatment. Obese children can have adult-type OSA and need referral.

Older adults. Sleep architecture changes and daytime sleepiness are less specific. Treatment improves alertness, blood pressure and quality of life; the threshold for CPAP initiation should be driven by symptoms and comorbidity.

Immunocompromised. OSA per se is not altered by immune status, but these patients can decompensate more readily from sleep fragmentation and nocturnal hypoxia; correct the airway and weight.

• A pulse oximeter and an overnight recording, even a simple one, gives enough to suspect OSA; in-laboratory PSG may not be available, but a home study does a great deal.
• If no study is available, a high STOP-BANG, resistant hypertension, a large neck and daytime sleepiness together are enough to start a careful trial of CPAP with clinical monitoring.
• Sedatives are cheap and commonly handed out for insomnia; avoid them in anyone who snores.
• Basic prevention (weight, nose, alcohol) costs nothing.
• For OHS, an arterial or venous blood gas and a serum bicarbonate are more useful than another imaging test. If NIV is not available, CPAP can be tried first in stable patients while arranging transfer and long-term care.

OHS is often silent until a hospital admission. People present with ankle oedema, breathlessness, somnolence, or an acute infection that tips them into respiratory failure. Because 49% of patients who end up on long-term NIV are first set up during an acute admission, recognition in the clinic and screening blood bicarbonate matter a great deal. ^[1]

Obese patients, particularly with OHS, are at high risk around intubation and extubation: difficult laryngoscopy (macroglossia, short neck, redundant soft tissues), reduced oxygen stores from low functional residual capacity, aspiration risk from reflux, and a prolonged residual effect of sedatives that favours post-extubation upper-airway obstruction. Duration and dose of sedation should be minimised; pre-oxygenation with CPAP or NIV or high-flow nasal cannulae helps; a cuff-leak test before extubation and, where laryngeal oedema is suspected, steroids at least 4 hours beforehand are prudent. ^[1]

Providing positive airway pressure is the treatment of choice for stable OHS. Because more than 70% of people with OHS also have severe OSA, a substantial minority improve with CPAP alone, which splints the airway and can restore daytime eucapnia by clearing carbon dioxide accumulated during obstructed events. NIV (BiPAP) is considered first-line in OHS without significant OSA, and remains the practical choice when there is persistent desaturation, worsening transcutaneous CO2 on CPAP, or combined restrictive physiology. Bilevel pressure-targeted devices with separately adjustable IPAP and EPAP let one treat both obstruction (raise EPAP) and residual hypoventilation (raise inspiratory pressure support). The commonest mode is spontaneous-timed, with a back-up rate. ^[1]

Three randomised trials comparing CPAP with NIV in stable hypercapnic obese patients with concomitant OSA (mean AHI above 60) found no significant difference in compliance (5 to 6 hours per day), awake PaCO2 or PaO2, oxygen need, emergency visits, quality of life or Epworth sleepiness score; subjective sleep quality and vigilance were better with NIV in one. Pooling these data means CPAP can be tried first where severe obstruction dominates, stepping up to NIV on failure. NIV must be started during an acute decompensation early, because the outcome of patients who fail NIV and are not intubated is extremely poor. ^[1]

Weight loss remains essential in OHS: bariatric surgery and multidisciplinary care can improve lung function and PaCO2. Watch for "malignant obesity" (extreme obesity with multi-organ failure), where outcomes in critical care are very poor. ^[1]

Central apnoea differs fundamentally from OSA: respiratory effort stops altogether rather than being preserved against a closed airway. Periodic (Cheyne-Stokes) breathing in heart failure, high altitude, opioid use, stroke and treatment-emergent CSA after starting CPAP are the common forms. OSA patients who start CPAP may develop central apnoeas in 5% of cases, usually resolving within 30 days; a stepwise rise in AHI after months of stable CPAP is a reason to look for atrial fibrillation. Standard questionnaires such as the Epworth or STOP-BANG are not designed to identify central disease. OHS itself can appear among the "medical disorder, non-Cheyne-Stokes" central apnoea syndromes. ^[6]

Untreated OSA puts the patient and the public at risk: falling asleep at the wheel, inattention at work, and accidents with machinery. Licensing authorities generally require adequate control of sleepiness and of the underlying disorder before driving, particularly for commercial drivers; people who are sleepy must not drive until treated and reviewed. Be honest but non-alarmist: treatment (usually CPAP) with documented adherence rapidly restores safe driving in most people. ^[3]

• OSA is repeated collapse of the upper airway during sleep, with preserved respiratory effort and intermittent hypoxia/hypercapnia; central apnoea has no effort.
• AHI grades severity: mild 5 to under 15, moderate 15 to under 30, severe 30 or more.
• Obesity is the strongest modifiable risk factor; neck circumference above 43 cm in men and 37 cm in women is a very good bedside clue.
• Craniofacial anatomy, male sex, menopause, hypothyroidism, acromegaly, alcohol and sedatives and smoking all contribute.
• Clinical clues are snoring, witnessed apnoeas, unrefreshing sleep, daytime somnolence, Epworth sleepiness score, resistant hypertension and nocturia.
• STOP-BANG (a score of 3 or more is high risk) is a screen, not a rule-out.
• Home sleep apnoea testing is first-line; polysomnography is the gold standard for complex, equivocal or mixed cases.
• Consequences span hypertension, arrhythmia, stroke, pulmonary hypertension, motor vehicle accidents, diabetes, metabolic syndrome and mortality.
• CPAP is first-line. Add or substitute weight loss and bariatric surgery, oral appliances, positional therapy, surgery or tirzepatide according to anatomy, severity and tolerance.
• OHS means BMI of at least 30, daytime PaCO2 above 45 mmHg and sleep-disordered breathing after excluding other causes; screen with serum bicarbonate and confirm with arterial blood gas.
• NIV (BiPAP) is first-line for OHS without severe OSA; CPAP can work when severe OSA coexists; step between them based on blood gas and sleep study.
• OHS is peri-operatively high-risk: minimise sedation, prepare for difficult airways and use CPAP/NIV and a cuff-leak before extubation.
• Central sleep apnoea is a distinct entity; it can be treatment-emergent after starting CPAP.
• Occupational and driving safety must be discussed and documented.

• Pulmonary hypertension is addressed separately on the pulmonary hypertension page.
• Stroke and cardiovascular risk factors are covered in the cardiology topics.
• Chronic obstructive pulmonary disease overlaps with sleep apnoea and shares management themes.