Show Summary Details
Page of

Sleep apnoea and hypoventilation 

Sleep apnoea and hypoventilation
Sleep apnoea and hypoventilation

Stephen Chapman

, Grace Robinson

, John Stradling

, Sophie West

, and John Wrightson

Page of

PRINTED FROM OXFORD MEDICINE ONLINE ( © Oxford University Press, 2021. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Medicine Online for personal use (for details see Privacy Policy and Legal Notice).

date: 26 October 2021

Obstructive sleep apnoea (OSA)

Definition and epidemiology

OSA, or obstructive sleep apnoea/hypopnoea (OSAH) are currently the preferred terms for the problem of dynamic upper airway obstruction during sleep.

  • OSA is part of a spectrum, with trivial snoring at one end and repetitive complete obstruction throughout the night (such that the patient cannot sleep and breathe at the same time) at the other

  • Along this spectrum is a point at which the degree of obstruction/recovery and the attendant arousal fragments sleep sufficiently to cause daytime symptoms

  • Distinction should be made between just the findings on sleep study of OSA episodes (OSA) and an abnormal sleep study plus the presence of symptoms (i.e. obstructive sleep apnoea syndrome, OSAS). Asymptomatic OSA is commoner than symptomatic (OSAS).

Thresholds defining ‘abnormality’ are arbitrary (e.g. 10s to define an apnoea). Numerical definitions of OSA, based on counting individual events during a sleep study, are not very helpful. The current definition of the clinical syndrome should be:

Upper airway narrowing, provoked by sleep, causing sufficient sleep fragmentation to result in significant daytime symptoms, usually excessive sleepiness.

  • Prevalence depends on the chosen thresholds for defining both an abnormality on the sleep study and significant symptoms

  • 0.5–1% of adult men in the UK (and about a fifth as many women) have OSAS, sufficient to be candidates for treatment with nasal CPAP

  • Prevalence figures depend on levels of obesity and will be higher in the USA and probably rise inexorably in the UK in the future

  • The prevalence in women is thought to be lower due to their different fat distribution. Upper body obesity (and thus neck obesity) is more a ♂ pattern

  • OSA is the third most common serious respiratory condition, after asthma and COPD. In some respiratory units, it has now become the commonest reason for specialist referral.

Pathophysiology and associated conditions

Control of the upper airway musculature is complex; upper airway patency depends on dilator muscle activity. All postural muscles relax during sleep (including pharyngeal dilators); some narrowing of the upper airway is normal. Excessive narrowing with the onset of sleep is due to the following factors.

Causes of a small pharyngeal size when awake

(such that normal muscle relaxation with sleep is enough to provoke critical narrowing)

  • Fatty infiltration of pharyngeal tissues and external pressure from increased neck fat and/or muscle bulk

  • Large tonsils

  • Subtle ‘abnormalities’ of craniofacial shape, e.g. minor micrognathia or retrognathia

  • Extra submucosal tissue, e.g. myxoedema, mucopolysaccharidoses.

Causes of excessive narrowing of the airway occurring with muscle relaxation at sleep onset

  • Mass loading from an obese or muscular neck may simply ‘overwhelm’ residual dilator action as well as reduce the starting size

  • Neuromuscular diseases with pharyngeal involvement may lead to greater loss of dilator muscle tone, e.g. stroke, myotonic dystrophy, Duchenne dystrophy, MND

  • Muscle relaxants such as sedatives and alcohol

  • Increasing age.

Other theories

In a small number of examples of OSA, a heightened tendency to arouse, before the breathing and the upper airway stabilize, may be important in maintaining ventilatory instability. There may also be years of damage to the mucosa from snoring, which reduce the protective reflex dilation of the pharynx in response to narrowing activated by surface receptors.

Predisposing conditions

OSA is found more commonly in certain conditions, such as acromegaly and hypothyroidism, but the reasons are not well understood. It is unclear whether there need to be any other non-anatomical factors to provoke OSA. Most associated abnormalities that have been described are likely to be 2° to long periods of snoring and OSA, rather than 1° causal factors.

Short-term consequences of OSA

In severe OSA, repetitive collapse of the upper airway, with the arousal required to reactivate the pharyngeal dilators, occurs approximately every minute throughout the sleeping period (60 events/h or over 400/night); they are usually attended by hypoxia (see Fig. 47.2) and hypercapnia that are corrected during the inter-apnoeic hyperventilatory period. Obstructive events, short of complete obstruction, also provoke arousal, as it is usually the compensatory reflex increase in inspiratory effort, rather than the blood gas deterioration directly, that wakens the brain. In this situation, the drops in O2 saturation may be very much less and, in younger thinner individuals, almost imperceptible on oximetry tracings (see Fig. 47.6). This is because the compensation afforded by the increased inspiratory effort may be adequate, and the bigger O2 stores in the lungs of the less obese will buffer any brief hypoventilation.

  • Recurrent arousals lead to highly fragmented and unrefreshing sleep

  • Excessive daytime sleepiness results

  • The correlation between the sleep fragmentation and the resultant degree of sleepiness is not tight, with some patients being sleepy with low levels of fragmentation, and vice versa

  • This is thought to result partly from inter-individual differences in sensitivity to the effects of sleep fragmentation

  • With every arousal, there is a rise in BP, often over 50mmHg. It is unclear if these BP rises do any damage to the cardiovascular system. There is also a carry-over of hypertension (average of 3mmHg) into the waking hours, which falls after treatment at 1 month

  • There is true nocturia, mechanism unclear; there may be raised atrial natriuretic peptide (ANP) levels from increased central blood volume, from the subatmospheric intrathoracic pressures during the obstructed breathing; or it may be simply a reflection of highly fragmented sleep, preventing the normal reduction in urine flow associated with sleep.

Fig. 47.2 Severe OSA. Large numbers of regular dips, sawtooth-shaped (faster rise in SaO2 than fall).

Fig. 47.2 Severe OSA. Large numbers of regular dips, sawtooth-shaped (faster rise in SaO2 than fall).

Fig. 47.6 Moderate OSA with many small, <4%, dips in SaO2.

Fig. 47.6 Moderate OSA with many small, <4%, dips in SaO2.

OSA: clinical features

Chapter 14 covers many of the essential features in the history and discusses the differential diagnosis of excessive daytime sleepiness.

Most patients present with:

  • Excessive sleepiness, measured crudely using the ESS; >9 is considered abnormally sleepy (see Sleep apnoea and hypoventilation pp. [link][link])

  • Loud snoring and apnoeic episodes recognized by the bed partner

  • The patient recognizes that he wakes up choking from time to time

  • Poor concentration

  • Unrefreshing sleep and waking unrefreshed

  • Nocturia (true nocturia with reversal of the usual day/night ratio).

Less often there will be:

  • Nocturnal sweating

  • Reduced libido

  • Oesophageal reflux

  • Increasingly common are patients arriving with spouses worried by the apnoeic pauses they have observed.


  • Sometimes difficult to assess; failure by the patient to recognize the problem or denial due to concerns over driving and licensing

  • The Epworth scale (see Sleep apnoea and hypoventilation p. [link]) assesses tendency to fall asleep, rather than perceived sleepiness per se, as some patients may regard their situation as normal

  • It is important to separate the symptom of tiredness from sleepiness (see Sleep apnoea and hypoventilation pp. [link][link]), the latter being much more typical of OSA, although sometimes complained of (rather than sleepiness) more by women with OSA.

Examination and investigations

The examination (often unrewarding) and the investigations are detailed in Chapter 14 (see Sleep apnoea and hypoventilation pp. [link][link]). Look for the presence of additional lower airways obstruction, with associated CO2 retention, so-called ‘overlap syndrome’. CO2 retention in pure OSA is very uncommon (except in the very, very obese). It appears that the additive effect of some lower airways obstruction (often not enough in its own right to precipitate CO2 retention) is required, which perhaps limits the inter-apnoeic hyperventilation and thus gradually encourages tolerance to raised levels of CO2.

The majority of patients with significant OSA are ♂, tend to have a combination of upper body obesity (neck circumference >17in) and a relatively undersized or set-back mandible. Airway size can be assessed and documented with scoring systems, e.g. see Sleep apnoea and hypoventilation p. [link].

Fig. 47.9 Mallampati index. Simple scoring system for pharyngeal crowding. Affected by craniofacial shape, tongue size, and obesity. Predicts difficulty of intubation and correlates with OSA severity. Reproduced from Update in Anaesthesia; Issue 13, 2001, with kind permission from the WFSA

Fig. 47.9 Mallampati index. Simple scoring system for pharyngeal crowding. Affected by craniofacial shape, tongue size, and obesity. Predicts difficulty of intubation and correlates with OSA severity. Reproduced from Update in Anaesthesia; Issue 13, 2001, with kind permission from the WFSA

Sleep study

The sleep study assesses if there is anything likely to be the cause of the patient’s symptoms. The considerable grey area between normality and abnormality means that sometimes it is unclear whether the symptoms can be blamed on the sleep study findings. There is also considerable night-to-night variation in sleep study indices that further blurs the distinction between normality and abnormality. In this situation, it may be necessary to undertake a therapeutic trial of CPAP and let the patient decide if the benefits of treatment outweigh the disadvantages.

OSA: types of sleep study

  • Overnight oximetry alone, including HR (see Figs. 47.1 to 47.8)

  • More than just oximetry, with other channels such as sound, body movement, oronasal airflow, chest and abdominal movements, leg movements: so-called ‘limited’ sleep studies or ‘respiratory polysomnography (PSG)’

  • Full PSG, with electroencephalogram (EEG), electro-oculogram (EOG), and EMG, to stage sleep electrophysiologically, in addition to the channels listed.

Fig. 47.1 Normal overnight oximetry. Normal baseline and a few dips. Vertical axis, 70–100% SaO2 for each panel; horizontal axis, 60min each panel, 8h total.

Fig. 47.1 Normal overnight oximetry. Normal baseline and a few dips. Vertical axis, 70–100% SaO2 for each panel; horizontal axis, 60min each panel, 8h total.

Fig. 47.3 Sleep-onset periodic ventilation. Short bursts of dipping, otherwise normal.

Fig. 47.3 Sleep-onset periodic ventilation. Short bursts of dipping, otherwise normal.

Fig. 47.4 Cheyne–Stokes ventilation. Prolonged periods of dipping, often more sinusoidal, rather than sawtooth.

Fig. 47.4 Cheyne–Stokes ventilation. Prolonged periods of dipping, often more sinusoidal, rather than sawtooth.

Fig. 47.5 REM sleep hypoventilation in scoliosis. Substantial dips, in bursts, compatible with the occurrence of REM periods.

Fig. 47.5 REM sleep hypoventilation in scoliosis. Substantial dips, in bursts, compatible with the occurrence of REM periods.

Fig. 47.7 Nocturnal hypoxia in COPD. Low baseline SaO2, with more dramatic falls (often quite prolonged) during periods compatible with the occurrence of REM.

Fig. 47.7 Nocturnal hypoxia in COPD. Low baseline SaO2, with more dramatic falls (often quite prolonged) during periods compatible with the occurrence of REM.

Fig. 47.8 Overlap syndrome—OSA and COPD. Low baseline SaO2. Some periods with typical sawtooth-dipping; other periods with prolonged falls in SaO2.

Fig. 47.8 Overlap syndrome—OSA and COPD. Low baseline SaO2. Some periods with typical sawtooth-dipping; other periods with prolonged falls in SaO2.

There is no evidence that OSA diagnosis needs full PSG (PSG with EEG); it is very rarely indicated, and its routine use is a waste of resources. Oximetry (SaO2 and pulse rate) identifies most moderate to severe cases, allowing referral for CPAP. Abnormal oximetry, sometimes mimicking OSA, occurs with Cheyne–Stokes breathing (heart failure, post-stroke), in the very obese, and when there is a low baseline SaO2 (e.g. COPD); this allows the SaO2 to oscillate more with small changes in PaO2, due to the increasing steepness of the Hb dissociation curve at lower SaO2. False negatives, discussed earlier, can occur with younger and thinner patients.

With appropriate expertise, a good history, and recognition of its limitations, domiciliary oximetry alone is a valuable tool in the diagnosis and management of OSA. The addition of sound and/or oronasal flow channels reduces the number of equivocal results needing a further study.

Limited sleep studies (respiratory PSG) are the usual routine investigation. Different units have expertise in interpreting different sorts of sleep studies. Experience is more important than the particular sleep study equipment used. Any system should assess the degree of sleep fragmentation and the degree of upper airway narrowing; this can be done using many different direct and indirect techniques.


Not all patients need treatment. The evidence for significant treatment benefits rests on symptoms, which drive treatment, rather than the degree of OSA on a sleep study. Treatment decisions require a close dialogue between physician and patient. Recent evidence suggests that many patients underestimate their symptoms so that, when in doubt, erring on the side of a CPAP trial is sensible.

Key features in making a treatment choice

  • How sleepy is the patient? Does it affect QoL? Is it critical to the patient’s livelihood (e.g. HGV driving)? Is there motivation for treatment?

  • Has the patient underestimated the impact of their sleepiness or misled the doctor because of concerns over driving issues?

  • Is there any evidence of the ‘overlap’ syndrome where additional lower airways obstruction has contributed to type II ventilatory failure? If so, is this a stable state or part of an acute decline with a respiratory acidosis?

  • Is obesity the dominant risk factor, or is there a surgically remediable component (e.g. tonsillar hypertrophy)?

There is no RCT evidence that cardiovascular disease, nocturnal angina, or poorly controlled hypertension should influence the decision to treat. Many, however, would lower the treatment threshold under these circumstances. OSA is a risk factor for recurrent AF; some evidence exists that treating OSA, when there is left heart failure, improves ejection fraction and possibly survival.

Simple approaches

  • Weight loss. This is difficult; slimming clubs have the best record for non-surgical approaches

  • Reduce evening alcohol consumption

  • Sleep decubitus, rather than supine, and with the bedhead elevated.

For snorers and mild OSA

  • Mandibular advancement devices, assuming adequate dentition

  • Pharyngeal surgery as a last resort (poor RCT data and what there is suggests poor outcomes, not recommended).

For significant OSA

  • CPAP therapy

  • Bariatric surgery (e.g. gastric band or gastric bypass operations)

  • Tracheostomy (rarely indicated)

  • Mandibular/maxillary advancement surgery in highly selected cases.

Severe OSA with CO2 retention

  • May require a period of non-invasive positive pressure ventilation (NIPPV) prior to CPAP, particularly if acidotic

  • Compensated CO2 retention may reverse with CPAP alone.

If there are large tonsils, then their removal may be appropriate although much more successful in children than adults. Pharyngeal surgery is a poor option for either snoring or OSA; the minimal evidence suggests that the outcomes are little better than placebo. There is no real place for alerting agents (such as modafinil) in the routine management of sleepiness in OSA. It is unclear if these drugs do help; they may possibly reduce the perception of sleepiness more than the sleepiness itself and have only been studied properly in the residual sleepiness sometimes found in patients even when treated successfully with CPAP.

Mandibular advancement devices

  • Worn in the mouth at night, holding the lower jaw forward: similar to ‘jaw thrust’ in an unconscious patient. Generally used to control snoring and OSA at the milder end of the spectrum

  • Many different designs, but essentially one half clips to the upper teeth and the other half to the lower, and connected together with the lower jaw held forward by 5–10mm

  • Some give adjustable forward displacement; some are fixed

  • The more sophisticated need to be customized to match the patient’s dentition, which usually requires the services of a dentist

  • DIY devices exist that are heated and moulded to the teeth directly

  • One-size-fits-all devices exist that can work for snoring, but only if they advance the lower jaw

  • Side effects include excessive salivation, tooth pain, and jaw ache, which often lessen with time

  • Long-term use may be associated with movement of the teeth and alterations to the bite

  • The initial cost (usually over £400 for customized versions) is more than that of a CPAP machine (£300), and they usually only last about a year.


CPAP consists of a blower/pressure generator that sits by the patient’s bed and is connected to a mask by a length of large-bore tubing. The masks are usually just nasal, but nose and mouth masks are also used. The blower raises the pressure at which the patient is breathing (to about 10cmH2O) and splints open the pharynx, preventing its collapse, sleep fragmentation, and the consequent daytime sleepiness. CPAP is a highly effective therapy, with resolution of the sleepiness and large gains in QoL. It is a sufficiently curious and initially uncomfortable therapy to require a careful induction programme. Without this, the take-up and compliance rates are poor. Most centres have found that a dedicated CPAP nurse or technician is required. Many centres use special patient education aids, such as video presentations, and provide helplines. The best method of establishing a patient on CPAP and deriving the required mask pressure is not known, and many different approaches appear to work. Recent innovations include CPAP machines that automatically hunt for the required pressure and do not require an attended overnight titration. New mask designs appear at regular intervals, with a slow improvement in their comfort and fit (important to prevent air leaks). Patients require access to long-term support to maintain their CPAP equipment and attend to problems. There are now significant problems supporting the very large numbers of patients on CPAP that accumulate, the longer a service has been in operation.

The commonest problems encountered include:

  • Mouth leaks lead to increased air through the nose and out of the mouth, with excessive drying of the mucosa, nasal congestion, rhinitis, and sneezing. Usually solved with the addition of a heated humidifier

  • Pain and ulceration of the skin on the nasal bridge. Try different masks or patient ‘interfaces’ not resting on the nasal bridge

  • Claustrophobia. This usually settles but may require a different interface

  • Temporary nasal congestion, usually during a cold. Try nasal decongestants for these short periods only such as xylometazoline.

Alternative diagnoses

Most patients who snore are sleepy and have an abnormal sleep study, will have ordinary OSA, and respond to CPAP. Sometimes, differentiating OSA from central apnoeas (see Sleep apnoea and hypoventilation pp. [link][link]) can be difficult, because some patients with Cheyne–Stokes breathing may have a few obstructed breaths at the end of each apnoeic cycle, even though the problem is primarily central. Poor response to CPAP should at least prompt a reappraisal of the diagnosis. Not all OSA is due to pharyngeal collapse; a very small number of patients have laryngeal closure. This can occur with:

  • Shy–Drager syndrome (multi-system atrophy). This causes laryngeal abductor weakness with laryngeal closure during sleep, with stridulous obstruction, rather than the usual noise of snoring

  • RA can damage the larynx, with resultant OSA

  • Arnold–Chiari malformation can compress the brainstem and interfere with the control of the larynx and pharynx, as well as the control of ventilation, with mixed findings on the sleep study

  • These forms of obstruction also respond to CPAP therapy, as the larynx is also ‘blown open’ by raising airway pressure.

OSA: driving advice to patients

In UK law, one is responsible for one’s vigilance levels while driving. We know when we are sleepy and should stop driving. Driving while sleepy has been likened to driving whilst drunk, and a prison sentence can result from sleep-related accidents on the road. No one should drive while they are sleepy, and the same applies to pathological causes of sleepiness.

Advice to all patients with OSAS or suspected OSAS

Whatever the situation, do not drive while sleepy; stop and have a nap; this is common sense, regardless of the cause of the sleepiness. If the sleepiness is sufficient to impair driving ability, patients must stop driving. Patients with OSA syndrome (i.e. OSA with daytime hypersomnolence) should write to the DVLA who will send them a questionnaire SL1. If they admit to excessive and inappropriate daytime sleepiness, their licence is revoked. If already treated and the sleepiness has resolved, then the licence is not revoked; hence, rapid treatment is essential. It is the doctor’s duty to tell the patient of the diagnosis of OSAS and of the requirement to inform the DVLA if there is sleepiness sufficient to impair driving. Patients who are not sleepy and only have OSA on their sleep study do not have OSAS and thus do not need to inform the DVLA, although the DVLA is inconsistent on this issue. According to USA epidemiological studies, at least 1/20 men have OSA on a sleep study (but, at most, only a quarter will be sleepy); thus, if all OSA were reported to the DVLA, this would rather overwhelm the DVLA and be illogical. The DVLA rules change from time to time, and reference to the latest version of their website and the At a Glance Guide is recommended.

The doctor can advise the patient whether they should stop driving entirely (essential if the patient is very sleepy, has had a sleepiness-related accident or near miss, and/or drives a HGV or public service vehicle—class 2 licence holders) or to continue driving only with extreme caution for short distances. The advice given to the patient should be recorded in the notes. The latest American Thoracic Guidelines are entirely sensible and useful.* Inappropriate and illogical curbing of driving privileges will push the problem underground, through fear of loss of livelihood, and is the worst of all worlds.

Driving can be restarted as soon as the sleepiness has resolved and preferably been confirmed by medical opinion, but, in the case of class 2 licence holders, the success of the treatment must be verified by a specialist clinic. This means a normal ESS and evidence of adequate CPAP usage from the hour meters built into CPAP machines. A minimum usage has not been defined, but >3h/night on average is often the arbitrary threshold used. Non-usage for even one night can lead to a return of sleepiness in some, so patients have to continue to act responsibly.

OSA in children

Adult respiratory physicians interested in sleep apnoea may be asked to investigate children with OSA, due to fewer sleep services for children.

  • Mainly due to enlarged tonsils and adenoids; varying degrees of OSA are present in up to 4% of children around the age of 5; prevalence tails off as tonsils atrophy

  • These children present with snoring, restless sleep, and different daytime symptoms to those of adults. Obvious sleepiness is less common

  • Sleep-deprived children tend to become hyperactive, with reduced attention spans, and be labelled as difficult or disruptive, or even ADHD

  • Symptoms will fluctuate with the size of the tonsils, and this depends on the presence of upper respiratory infections

  • Mild intermittent sleep disturbance may not matter, but every night sleep fragmentation for months interferes with development in a variety of ways

  • The clinical decision is mainly whether to recommend removal of tonsils or a wait-and-see policy, remembering that there is a significant morbidity from adenotonsillectomy, and even the occasional death

  • A halfway house is the use of nasal steroids, which can reduce tonsillar size sufficiently to improve symptoms until natural tonsillar atrophy occurs.

OSA: future developments

The old view that sleep PSG is required to diagnose OSA has been replaced with an evidence-based approach, using simpler and cheaper equipment. OSA is so common that diagnosis will move into general practice. Simpler ways to establish patients on CPAP have evolved. RCTs show that, with appropriate training and supervision, diagnosis and management of OSAS can be carried out in the community. Obesity surgery is improving and, in appropriate cases, may become the treatment of choice. Appetite suppressants are being developed and should greatly reduce the prevalence of obesity and hence OSA. Pharmacological agents are being developed to prevent the loss of tone in the pharyngeal dilators during sleep, although progress in this area is slow (e.g. mirtazapine reduces OSA but unfortunately causes sedation and weight gain!). New pharyngeal operations are being devised all the time, but none have been very effective when investigated properly; RCTs with objective outcome data are badly needed in this area.

Whether OSA is a significant independent risk factor for vascular diseases is still debated. Cross-sectional and non-randomized data suggest so. However, it is difficult to adequately control for confounding variables such as visceral obesity; non-randomized studies carry important bias. RCTs treating OSA with CPAP show small benefits to BP and endothelial function.

Further information

Heatley EM et al. Obstructive sleep apnoea in adults: a common chronic condition in need of a comprehensive chronic condition management approach. Sleep Med Rev 2013;17:349–55.Find this resource:

Loke YK et al. Association of obstructive sleep apnea with risk of serious cardiovascular events: a systematic review and meta-analysis. Circ Cardiovasc Qual Outcomes 2012;5:720–8.Find this resource:

Eastwood PR et al. Obstructive sleep apnoea: from pathogenesis to treatment: current controversies and future directions. Respirology 2010;15:587–95.Find this resource:

Chai-Coetzer CL et al. Primary care versus specialist sleep centre management of obstructive sleep apnea and daytime sleepiness and quality of life. JAMA 2013;309:997–1004.Find this resource:

Stradling JR. Driving and sleep apnoea. Thorax 2008;63:481–3.Find this resource:

Sleep Apnoea Trust, Patients Association. Sleep apnoea and hypoventilation

NICE technology appraisal of CPAP for OSA. Sleep apnoea and hypoventilation

IMPRESS Service Specification for Investigation and treatment of OSAS. Sleep apnoea and hypoventilation

Cochrane database on CPAP. Continuous positive airways pressure for obstructive sleep apnoea in adults. Sleep apnoea and hypoventilation;jsessionid=6AFDAAFA6FD5F175EF242B29324D4E96.d03t04.

Interventions to improve compliance with continuous positive airway pressure for OSA. Sleep apnoea and hypoventilation;jsessionid=6AFDAAFA6FD5F175EF242B29324D4E96.d03t04.

American commercial website, but useful for showing the large variety of CPAP equipment. Sleep apnoea and hypoventilation

Main American Sleep Information website. Sleep apnoea and hypoventilation

Central sleep apnoea (CSA) and nocturnal hypoventilation

Definition and epidemiology

‘CSA/hypopnoea’, or ‘hypoventilation’ or ‘periodic breathing’, are said to occur when there is no evidence of upper airway obstruction as the cause for the episodic reduced ventilation during sleep. Compared with OSA, it is much less common.

  • CSA tends to be used as a term when there are actual apnoeas and referred to as Cheyne–Stokes breathing when there is regular symmetrical waxing and waning, usually in the context of left heart failure

  • Periodic breathing is an alternative and can be used to describe regular fluctuations in breathing, with or without actual apnoeas

  • The description nocturnal hypoventilation tends to be used when the hypoventilation and hypoxaemic dips are not particularly periodic in nature. However, these terms are imprecise and sometimes mixed indiscriminately.


CSA, or hypoventilation or periodic breathing,

can occur in a number of settings with different aetiologies (see also Chapter 15, Sleep apnoea and hypoventilation pp. [link][link]). At one end of the spectrum is pure loss of ventilatory drive, while at the other is pure reduction in the ability to expand the chest adequately, with dependence on accessory muscles of respiration. Many clinical presentations are mixtures of these two.

Patients with reduced ventilatory drive (e.g. following brainstem damage) can often maintain adequate, or near adequate, ventilation whilst awake, as there is a non-metabolic ‘awake’ ventilatory drive equivalent to about 4 or 5L/min. During non-REM sleep, this awake drive is lost, and ventilation becomes dependent on PaO2 and PaCO2. During REM sleep, an ‘awake-like’ drive sometimes returns partially, and ventilation can improve again (seen in congenital forms of absent drive where REM sleep can temporarily restore SaO2 levels).

In patients with impaired mechanical ability to ventilate, accessory muscles of respiration become critically supportive (e.g. in many neuromuscular disorders and obstructive/restrictive respiratory conditions). However, during non-REM sleep, this reflex recruitment of accessory muscles is attenuated and hypoventilation follows. During REM sleep, the physiological paralysis of all postural muscles (REM atonia) can remove all compensatory mechanisms, leaving only the diaphragm working, and may produce profound hypoventilation or apnoea.

Chronic hypoventilation, often

to poor respiratory function (e.g. as evidenced by CO2 retention in some patients with COPD or chest wall restrictive disorders), can eventually force resetting of ventilatory control mechanisms. This is an acquired blunting of ventilatory drive and leads to sleep-related changes in ventilation, similar to those described in the previous paragraphs.

Unstable ventilatory control

can lead to regular oscillations in ventilation, e.g. as occurs in heart failure and at altitude. During REM sleep, it is normal to have fluctuations in ventilation, sometimes with complete apnoeas.

Sometimes, sleep studies can be misinterpreted,

and apnoeas, really of obstructive origin, are mistakenly labelled as central. For example, if inspiratory muscles are very weak, their poor efforts during obstructive apnoeas may be missed on the ribcage and abdominal bands used in sleep studies but may be detectable with oesophageal pressure monitoring.

Causes and clinical features of CSA/hypoventilation

Although there are many different causes, only four relatively common clinical scenarios occur (but with overlap).

Absent or reduced ventilatory drive

Brainstem involvement from strokes, tumours, syringobulbia, surgical damage, post-polio syndrome, congenital (Ondine’s curse—usually presents soon after birth, can be later, abnormalities of neural crest development due to increased number of ‘alanine repeats’ in one of the homeobox genes—PHOX2B). Presents clinically with unexplained ventilatory failure, much worse during sleep when the ‘awake’ drive is lost.

  • May be recognized early, cyanosis, morning confusion, ankle oedema

  • May be recognized late, loss of consciousness, and an emergency admission to ICU for ventilation following a chest infection or general anaesthesia.

Lung function is often normal, with no evidence of respiratory muscle weakness, indicating normal innervation from the voluntary motor system. Arnold–Chiari malformation with brainstem compression can present like this, but there is usually involvement of surrounding structures such as the lower cranial motor nuclei supplying the larynx and pharynx (with associated OSA; see Sleep apnoea and hypoventilation p. [link]).

These patients will have no apparent neuromuscular or respiratory cause for their hypoventilation but may have a previous history of brainstem stroke (or other form of brain damage). The congenital form usually presents shortly after birth when the amount of REM sleep reduces and is replaced by non-REM.

Post-polio syndrome is:

  • Ill-defined syndrome—decline in function, decades after initial illness

  • Return of weakness in previously affected areas (mechanism unclear)

  • Late development of ventilatory failure is more likely if:

    • Inspiratory muscles were affected in the original illness

    • Additional scoliosis due to paravertebral muscle involvement (in which case VC will be reduced).

This may be due to premature ageing of the upper and lower motor neurones due to their ‘overuse’. This could follow the original destruction of some of the anterior horn cells to the inspiratory muscles and the subsequent reinnervation by surviving neurones which then have to continuously supply more neurones than they were ‘designed’ for.

Weak or mechanically disadvantaged inspiratory muscles with/without 2° reduction of awake ventilatory drive

Neuromuscular inspiratory muscle weakness

will produce diurnal ventilatory failure in its own right, particularly when the supine VC falls below 20% predicted (~1L).

With increasing inspiratory muscle weakness, other accessory inspiratory muscles are recruited to maintain ventilation. When this is lost during non-REM sleep, and more so during REM sleep, ventilation will fall much more than in normal subjects. Whilst metabolic ventilatory drive is reasonably preserved, this will result in recurrent arousals to ‘rescue’ the ventilation and consequent marked sleep disturbance.

As ventilatory drive becomes progressively blunted, following the hypoventilation forced on the system by weak muscles, extra sleep hypoventilation (from loss of ‘awake’ drive) is tolerated, and profound hypoxaemia is observed until there is finally an arousal that recovers the ventilation and SaO2.

The above patients should have a history of a progressive neuromuscular disorder.

Chest wall restrictive diseases,

such as scoliosis or post-thoracoplasty patients (see Fig. 47.5), can behave in a similar way with gradual onset of ventilatory failure, particularly when VC <1L. The muscles are not weak but operating at severe mechanical disadvantage.

The same situation occurs in COPD, when muscles are overloaded and accessory muscles provide important support, but this too is reduced with non-REM sleep and lost during REM sleep. Again, any 2° reduction in ventilatory drive amplifies the sleep-related falls in SaO2.

  • Chest wall restrictive patients should have an obvious restrictive disorder with reduced VC to 1L or below

  • Increasing degrees of COPD will produce increasing degrees of sleep hypoventilation (see Fig. 47.7)

  • If the awake SaO2 is already low, the sleep-related falls in ventilation will produce dramatic dips in SaO2

  • COPD and OSA together (overlap syndrome; see Sleep apnoea and hypoventilation p. [link]) provoke profound nocturnal hypoxic dipping (see Fig. 47.8), and probably a more rapid progression to diurnal hypoventilation with CO2 retention, due to extra blunting of ventilatory drive.

The diaphragm is the only respiratory muscle working during REM sleep, as all other postural muscles are profoundly hypotonic.

  • If the diaphragm is paralysed, then REM sleep is a particularly vulnerable time, as there are no muscles of ventilation left working, thus producing particularly profound falls in SaO2 during REM

  • Patients with bilateral diaphragm weakness can present early, with no obvious weakness elsewhere. Diaphragm weakness is best detected with the patient supine. Inspiration, particularly on sniffing, will provoke a paradoxical indrawing of the abdominal wall. The VC will also fall on lying down, increasingly with greater degrees of paralysis (often a >30% fall in VC with complete diaphragm paralysis).

In a progressive neuromuscular disorder, such as MND, the above patterns will be variable between individuals but will gradually worsen. Predominant diaphragm weakness, as occurs sometimes in MND, spinal muscular atrophy, and particularly acid maltase deficiency, can lead to ventilatory failure at a time when the patient is still ambulant.

Cheyne–Stokes breathing associated with LVF

(See Fig. 47.4.)

The raised left atrial pressure in left heart failure increases ventilatory drive through stimulation of J receptors; this in addition to ventilatory stimulation from any hypoxaemia from pulmonary oedema.

  • This ventilatory stimulation lowers the awake PaCO2, producing a respiratory alkalosis

  • In addition, the use of diuretics may produce a mild metabolic alkalosis, especially if there is hypokalaemia

  • This extra J receptor ventilatory stimulation appears to reduce at sleep onset. This, together with the loss of the awake ventilatory drive, allows central hypoventilation or apnoea to occur

  • This hypoventilation or apnoea will continue until the PaCO2 builds up driving ventilation again or until the hypoxaemia provokes arousal

  • The return of ventilation itself may provoke arousal too. The arousal itself then injects the increased ‘wake’ ventilatory drive, reducing the PaCO2 again

  • Sleep returns, and, once again, the low PaCO2 and alkalosis cause hypoventilation or apnoea.

Thus, a cycle is maintained that involves a fluctuating sleep state with arousals, and usually a fluctuating SaO2. As with OSA, the patient may be completely unaware of these arousals. The delayed circulatory time of left heart failure may compound this instability by introducing a time delay between any change in PaCO2 in the blood leaving the lungs and its arrival at the carotid body or central chemoreceptors.

Cheyne–Stokes breathing associated with altitude

The acute hypoxia following ascent to altitude provokes increased ventilation. The degree is variable between individuals, and hence the degree of hypocapnia and respiratory alkalosis varies (see Sleep apnoea and hypoventilation p. [link]). With sleep onset, with a lessening of the hypoxic drive, and removal of the awake drive, an uncompensated alkalosis will allow hypoventilation, and even apnoea—similar to the situation described previously for Cheyne–Stokes breathing. Again, ventilation will restart, either when the PaCO2 rises to a critical level or the hypoxia provokes arousal. Sleep is fragmented with complaints of insomnia, but the cause of this is rarely recognized by the sufferer.

Skiing in Colorado, altitude 2,400–3,400m (~10,000ft), is high enough to provoke significant periodic breathing in about a fifth of individuals. It seems that this fifth are the ones with the highest hypoxic ventilatory response. This gives them the largest respiratory alkalosis and hence the greatest tendency to sleep-onset hypoventilation. In addition, the tendency to arouse with the resultant extra hypoxaemia may be greater too, thus provoking large increases in ventilation on arousal and greater sleep disturbance. As the kidney excretes extra bicarbonate and produces a compensatory metabolic acidosis over a few days, the periodic breathing lessens. See also Sleep apnoea and hypoventilation pp. [link][link] (altitude sickness).

Two pharmacological approaches

have been taken to reduce this sleep-related periodic breathing at altitude.

  • Pre-acclimatization with acetazolamide prior to ascent. This produces a mild metabolic acidosis and maintains the ventilatory drive at sleep onset, thus blocking the hypoventilation. RCTs show the efficacy of this approach with doses between 250 and 500mg/day, 1–3 days prior to ascent

  • Hypnotics, such as temazepam, can reduce the degree of periodic breathing by reducing the tendency to arousal with each return of ventilation, and thus damping the system. Randomized trials suggest benefit for the early part of the night, with no impairment of nocturnal hypoxaemia or daytime functioning

  • Of course, extra O2 will abolish the problem.

CSA/hypoventilation: investigations

Simple PFTs will characterize weakness of inspiratory muscles. Supine VC is the best predictor of ventilatory failure, as it incorporates diaphragm weakness that is masked during erect testing. Blood gases will reveal diurnal type II ventilatory failure. If the bicarbonate/base excess is raised, with a normal PaCO2 (showing therefore a mild metabolic alkalosis), then this may indicate nocturnal hypoventilation and incipient ventilatory failure. The bicarbonate has been referred to as the ‘HbA1c of PaCO2’, as it represents an integrated response to the average raised PaCO2 over the last 48h or so (in the absence of any other reason for a metabolic alkalosis such as hypokalaemia and some diuretics).

Sleep study

(For examples, see Figs 47.3, 47.4, 47.5, 47.7, and 47.8.)

Sleep studies in patients with suspected nocturnal hypoventilation or CSA confirm the diagnosis and assess the degree of nocturnal hypoxaemia. Limited sleep studies should reveal falls in SaO2 in association with hypoventilation, but no evidence of OSA and, in particular, no snoring. Oximetry tracings alone will show a variety of patterns, often resembling OSA. The pattern in neuromuscular weakness will vary from oscillations all the time (due to recurrent arousal) to just REM sleep-related dips in SaO2. The same will be true for chest wall restrictive disorders and COPD, with REM dips occurring initially and greater hypoxaemia once there is an element of diurnal CO2 retention and hypoxaemia.

  • In OSA, there is a slow fall in SaO2, as O2 levels fall in the lung, followed by a rapid rise with the first deep inspiration as the apnoea ends (so-called sawtooth pattern); see Fig. 47.2

  • In Cheyne–Stokes of left heart failure, the oscillations in SaO2 are often more sinusoidal than in OSA, as the pattern of breathing is usually more of a symmetrical waxing and waning of ventilation; see Fig. 47.4. However, if each central apnoea is terminated by an arousal, rather than a smooth return of ventilation, then the pattern will look more like OSA.

In COPD, the degree of hypoxaemia on the sleep study will depend very much on the awake SaO2. Because of the shape of the Hb dissociation curve, a low awake SaO2 makes it easier for the SaO2 to fall further with a given reduction in ventilation. Thus, during non-REM, with removal of awake drive, there will be a fairly stable reduction in SaO2, but, during REM sleep, there will be further more dramatic dips. It is important not to diagnose OSA from just an oximetry tracing on the basis of SaO2 oscillations when there is a low baseline SaO2 and COPD. In this situation, a fuller sleep study is required to provide evidence of additional upper airway obstruction. The combination of hypoxic COPD and OSA (one of the overlap syndromes) can produce particularly dramatic traces (see Fig. 47.8).


Intervention in CSA, or hypoventilation or periodic breathing, depends on symptoms. Better control of heart failure may improve Cheyne–Stokes breathing but often does not. Further treatment will be required for two reasons: either to prevent the cyclical breathing and restore sleep quality, or to globally improve ventilation overnight and reset the respiratory control mechanisms such that the daytime respiratory failure reverses.

In situations where the hypoxia is playing a part in the pathogenesis (e.g. heart failure), then raising FiO2 can help (using O2 via nasal prongs only during sleep). There is limited literature on other forms of treatment for the Cheyne–Stokes of heart failure, although acetazolamide and benzodiazepines have been tried. The unstable breathing in heart failure has been treated with CPAP; however, a large RCT has not confirmed long-term benefit. More recently, treatment has been tried with sophisticated servo-ventilators that are able to cut in smoothly, as ventilation wanes and backs off when it waxes, thus ironing out the oscillations without overventilating; whether these will provide better relief of symptoms than O2 is not yet clear, but preliminary evidence suggests they might. There are some data to suggest that the recurrent arousal in Cheynes–Stokes may raise catecholamine levels and provoke deterioration of LV function. Thus, measures to reduce the arousals may improve cardiac function as well as improve daytime vigilance.

Sedatives are contraindicated with a raised PaCO2, and extra O2 may increase the hypercapnia. In these situations, then overnight NIV, via either nose or face mask, may be appropriate. In slowly progressive neuromuscular disorders, with either sleep fragmentation or diurnal type II respiratory failure (or both), the symptomatic and physiological response can be dramatic. The use of NIV at night in more rapidly progressive disorders has potential difficulties but is proving very useful in the palliative care of disorders such as MND. Increasing dependence on equipment, not designed to be immediately life-sustaining, is a particular issue. In scoliosis, there is rarely any question that treatment might not be appropriate, and again responses are dramatic.

Future developments

  • Treatment of LVF (acute and chronic, no evidence yet for the latter) with overnight CPAP or NIV

  • Introduction of overnight ventilation earlier in the course of a progressive neuromuscular disorder (such as MND) to reduce symptoms and possibly prolong life

  • Use of overnight ventilation on patients with stable COPD and hypercapnia. This may reduce exacerbations, hospital admissions, and prolong life. The evidence is inadequate yet to justify its wide use in this patient group.

Further information

Oldenburg O. Cheyne-stokes respiration in chronic heart failure. Treatment with adaptive servoventilation therapy. Circ J 2012;76:2305–7.Find this resource:

Naughton MT. Cheyne-Stokes respiration: friend or foe? Thorax;67:357–60.Find this resource:

Lanfranchi PA, Somers VK. Sleep-disordered breathing in heart failure: characteristics and implications. Respir Physiol Neurobiol 2003;136:153–65.Find this resource:

Berry-Kravis EM et al. Congenital central hypoventilation syndrome. Am J Respir Crit Care Med 2006;174:1139–44.Find this resource:

Gray A et al. Noninvasive ventilation in acute cardiogenic pulmonary edema. N Engl J Med 2008;359:142–51.Find this resource:

Obesity-related respiratory problems

Levels of obesity (BMI >30) are rising in all ‘civilized’ societies. In 1983, 8% in the UK (18% in the USA) had a BMI >30. Twenty-five years later, in 2008, this was 24% in the UK (34% in the USA). This has also had many impacts on health care outside of respiratory medicine, particularly the components of the metabolic syndrome.

Obesity, particularly in conjunction with OSA and COPD, provokes ventilatory failure and cor pulmonale (see Sleep apnoea and hypoventilation p. [link]). A common clinical scenario is the obese smoker, with a history of snoring and sleepiness, arriving in A&E with hypercapnia*. Why only some obese patients develop hypercapnia is not clear, but abdominal obesity seems the most hazardous. This may be because of greater lung volume compression, thus breathing occurs nearer residual volume where airway resistance is much higher and total compliance lower, thus increasing the work of breathing considerably.

In addition to apportioning the relative contribution of obesity, COPD, and OSA to the hypercapnia, and treating accordingly, there are other obesity-related factors to consider:

  • Intubation—the best predictors of a problematic intubation are neck circumference and a Mallampati score of 3 or more (see Fig. 47.9)

  • Tracheostomy tubes are often too short and too curved to cope with the increased distance between skin and trachea; tubes with adjustable flanges that allow customized intra-tracheal lengths are useful here

  • Percutaneous dilatational tracheostomy is more difficult and may have a higher complication rate

  • Low FRCs mean that the O2 stores are limited, leading to rapid falls in SaO2 during apnoeas of any cause

  • Abdominal loading of the diaphragm is greatly increased, but the extra embarrassment to the diaphragm can be reduced by tilting the whole bed, head up, by 15–25°

  • Abdominal loading of the diaphragm particularly reduces basal lung expansion. Resultant basal atelectasis increases the A–a gradient, which can be improved by raising end expiratory pressure during both invasive or non-invasive ventilation

  • Abdominal loading may increase perioperative risk of aspiration

  • DVTs and PEs are probably more common in the obese. It is not clear if DVT prophylaxis regimes need to be modified. Some recommend higher doses of LMWH, and this higher dose is more effective in patients undergoing bariatric surgery. Weight-based regimes of LMWH for the treatment of DVT and emboli appear satisfactory in the morbidly obese (BMI >40)

  • Increased likelihood of failure to wean—NIV (inspiratory pressure 12, expiratory 4cmH2O) has been shown to aid weaning, e.g. post-open gastric bypass surgery for obesity

  • Possible build-up of sedating anaesthetic agents in fat, leading to prolonged half-life.

Obese patients can often have mildly raised PaCO2 and bicarbonate levels for long periods, without decline or apparent problems. Once there is acute decompensation with acidosis, management is more difficult. NIV is usually effective but may require very high inspiratory and expiratory pressures that can be difficult to deliver adequately without pressure damage and ulceration to the nasal bridge. Weight loss is very effective but, as usual, is hard to achieve without bariatric surgery, which will be hazardous if NIV has not lifted the patient out of ventilatory failure. Acetazolamide is used by some in this situation, but its use is not evidence-based.

Further information

Chau EH et al. Obesity hypoventilation syndrome: a review of epidemiology, pathophysiology, and perioperative complications. Anesthesiology 2012;117:188–205.Find this resource:

El-Solh AA. Clinical approach to the critically ill, morbidly obese patient. Am J Respir Crit Care Med 2004;169:557–61.Find this resource:


* An official American Thoracic Society Clinical Practice Guideline: sleep apnea, sleepiness, and driving risk in noncommercial drivers. Am J Respir Crit Care Med. 2013;187:1259–66.

* The term ‘obesity hypoventilation syndrome’ is used when an obese individual (BMI >30kg/m2) has a raised PaCO2 (>6kPa) with no other apparent explanation, usually in conjunction with worsening sleep-related hypoventilation, and sometimes additional obstructive sleep apnoea.