Show Summary Details
Page of

Sleep-related disorders of breathing 

Sleep-related disorders of breathing
Chapter:
Sleep-related disorders of breathing
Author(s):

J.R. Stradling

and S.E. Craig

DOI:
10.1093/med/9780199204854.003.180502

A relevant case history from Oxford Case Histories in Respiratory Medicine has been added to this chapter.

Page of

PRINTED FROM OXFORD MEDICINE ONLINE (www.oxfordmedicine.com). © Oxford University Press, 2016. 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).

Subscriber: null; date: 22 June 2018

Essentials

Obstructive sleep apnoea (OSA) and other sleep-related breathing problems significantly impair the functioning of about 0.5 to 1% of the population and are becoming increasingly common.

Obstructive sleep apnoea

OSA in adults is commoner in men than women (3–5:1) and usually caused by obesity (BMI typically >30 kg/m2) and fat deposits in the neck area (typically collar size of 17 inches (43 cm) or more). This external loading can be fended off during wakefulness but not during sleep, when the withdrawal of postural muscle tone allows the pharyngeal dilators to be overwhelmed, leading to excessive narrowing or collapse of the airway, with consequent apnoea. The most important consequence of sleep-induced upper airway narrowing is sleep fragmentation.

Clinical features—there is a continuum from light intermittent snoring through to severe, all-night, OSA. The main symptom of OSA is daytime hypersomnolence, which correlates broadly with the degree of sleep disruption. Other common symptoms are loud snoring, restless or unrefreshing sleep, observed apnoeas, nocturia, and apparent personality change.

Diagnosis—it is important to ask the correct questions to assess sleepiness; a well validated and simple way to do this being with the Epworth Sleepiness Scale, in which the patient is asked to state how likely they are to doze off or fall asleep in a number of ordinary situations, e.g. sitting and reading. Patients scoring higher than normal generally merit further investigation in the form of some type of sleep study to (1) assess sleep fragmentation, (2) establish if a respiratory problem is responsible, and (3) decide if upper airway obstruction is the primary cause. Classical OSA, observed with a simple commercially available monitoring system, causes a snoring–silence–snoring pattern of sleep (from room microphone) together with body movements (from video) and oscillations in the pulse and SaO2 (from oximeter). Full polysomnography can provide much further information, but is not generally required in straightforward cases.

Management—mild symptoms may resolve with simple treatments and advice as follows: (1) learn to sleep on your side and avoid sleeping on your back, (2) no alcohol after 18.00 h, (3) no sedatives, (4) lose weight, (5) stop smoking, (6) keep the nose as clear as possible. However, if OSA and symptoms are severe, there is only one fully effective therapy—nasal continuous positive airway pressure (NCPAP): this involves wearing a small mask over the nose while asleep, with the air pressure kept at a fixed level above atmospheric (usually about 10 cmH2O) by a pump, sufficient to splint open the pharynx and resist collapse, allowing unobstructed breathing and undisturbed sleep.

Prognosis—many patients with OSA have visceral obesity and the metabolic syndrome (hypertension, insulin resistance, hyperlipidaemia), and their vascular mortality is higher than average. However, there is no controlled interventional data to support the routine use of NCPAP to reduce vascular risk in patients with OSA who would not otherwise want to use the treatment for relief of daytime sleepiness.

Sleep-induced hypoventilation and central sleep apnoea

Aetiology—breathing during sleep may decrease because of a reduction in central output to the respiratory muscles, which can be caused by (1) absent ventilatory drive—Ondine’s curse, caused by congenital abnormality, brainstem damage, or blunting secondary to lung disease; (2) unstable ventilatory drive—at sleep onset, with hypoxaemia, altitude, heart failure; (3) REM-related oscillations—neuromuscular diseases, chest-wall abnormalities, and chronic airways obstruction; (4) reflex central apnoea—when pharyngeal collapse inhibits inspiration.

Clinical features—some of the central apnoeas disturb sleep and present with daytime sleepiness, such that they can be confused with OSA, whereas others tend to present with symptoms of respiratory failure, such as morning headaches with confusion, cyanosis, and ankle oedema.

Management—without treatment the chronic ventilatory failure associated with some neurological disorders (e.g. acid maltase deficiency, postpoliomyelitis syndrome, motor neuron disease, Duchenne dystrophy) usually progresses rapidly to death. Supporting breathing overnight can fully reverse ventilatory failure, and the response to treatment can be dramatic, with resolution of all symptoms, restoration of normal blood gases, and addition of decades of active life.

Introduction

This chapter discusses the disorders of breathing that appear, or markedly deteriorate, only during sleep. Obstructive sleep apnoea (OSA) and sleep-related problems in general are becoming increasingly common, especially as they are often related to obesity. OSA is the third most common serious respiratory disorder (after asthma and chronic obstructive pulmonary disease (COPD)) and is thought to significantly impair the functioning of about 0.5 to 1% of the population. Most general hospitals will have some form of sleep monitoring system for the diagnosis of sleep apnoea syndromes, although tertiary centres tend to provide most of the treatment. The diversity of symptoms produced by these disorders means that all physicians need to have an understanding of them and are likely to come across many cases during their professional life.

Normal physiology of breathing during sleep

(Table 18.5.2.1)

Table 18.5.2.1 Sleep and breathing

NREM

REM

Electroencephalogram

Progressively slower frequency and higher amplitude

Similar to the awake pattern

Eye movements

Initially slow and pendular, then none

Bursts of rapid binocular movements

Postural muscle tone

Reduced from wakefulness

Very much reduced or absent

Factors controlling breathing

Loss of wakefulness input

Cortical over-riding and apparent reduction in responses to classical stimuli

Brainstem and classical stimuli dominate, but reduced compared with wakefulness

Arousal response

Small deteriorations in PaO2 and PaCO2, with the consequent ventilatory response, are required for arousal

Larger changes in PaO2 and PaCO2 required before arousal occurs

Potential effect on breathing

Rise in pharyngeal resistance

Further rise in pharyngeal resistance

Fall in minute ventilation

Loss of use of accessory muscles of respiration

Fall in PaO2

Further falls in PaO2 tolerated longer before rescued by arousal

NREM, non-rapid eye movement: REM, rapid eye movement.

Sleep can be divided into two very different states. The dominant sleep stage is non-rapid eye movement (NREM) sleep (Figs. 18.5.2.1 and 18.5.2.2). This phase of sleep, which is preferentially reclaimed following sleep deprivation, appears to be when the brain shuts down, and is necessary for maximum daytime alertness and continuing cognitive function. NREM sleep shows a continuum from drowsy down to very deep sleep, arbitrarily subdivided into stages 1, 2, 3, and 4. The awake electroencephalogram (EEG) is characterized by low-voltage, high-frequency activity, with the only dominant frequency being the so-called alpha activity (c.10 Hz), present when the eyes are closed. As sleep supervenes, the alpha activity disappears, overall EEG frequency falls, muscle tone (usually measured from a chin electromyogram (EMG)) falls, and the eyes begin to roll from side to side. This transition phase is called stage 1. Stage 2 is defined by the appearance of K complexes (isolated slow waves) and sleep spindles (bursts of c.13 Hz activity). As sleep deepens further, increasing amounts of large, slow waves (c.1 Hz) appear. These stages are called 3 and 4, or slow-wave sleep.

Fig. 18.5.2.1 Examples of electrical brain activity (EEG), eye movements (EOG), and chin muscle tone (EMG) during wakefulness and the different sleep stages.

Fig. 18.5.2.1
Examples of electrical brain activity (EEG), eye movements (EOG), and chin muscle tone (EMG) during wakefulness and the different sleep stages.

Fig. 18.5.2.2 Examples of all-night hypnograms (based on 20-s epochs) in two normal subjects and two patients with sleep apnoea. Note the reduced deep sleep (stages 3 and 4) in the patients, but no indication that they are waking up hundreds of times a night. W, awake; M, movement (awake); R, REM sleep; 1 to 4, stages 1, 2, 3, 4 of non-REM sleep.

Fig. 18.5.2.2
Examples of all-night hypnograms (based on 20-s epochs) in two normal subjects and two patients with sleep apnoea. Note the reduced deep sleep (stages 3 and 4) in the patients, but no indication that they are waking up hundreds of times a night. W, awake; M, movement (awake); R, REM sleep; 1 to 4, stages 1, 2, 3, 4 of non-REM sleep.

The other main phase of sleep is rapid eye movement (REM) sleep or dreaming sleep. This stage is characterized by a return of the EEG to a pattern resembling wakefulness. The EMG tone falls to very low levels and there are bursts of rapid eye movements, mainly from side to side, under closed eyelids. Effectively the cortex is ‘awake’ again, processing randomly activated images and able to integrate outside noises or other stimuli into complex dreams. The fall in EMG tone is because the rest of the body’s muscles have been ‘cut off’ from the brain and paralysed, hence the fall in EMG tone. This paralysis (or atonia) is under active control from Jouvet’s centre in the pons that hyperpolarizes the lower motor neurons via inhibitory reticulospinal pathways. Cats in whom this centre has been destroyed no longer show atonia during REM sleep, and as a consequence they may get up and walk around or appear to chase phantom birds, presumably reflecting their dream content. The function of this atonia centre may therefore be to prevent the dreaming brain from influencing the rest of the body. Paralysis during REM sleep occurs dominantly in muscles that normally have a tonic postural activity; thus the diaphragm is spared, although pharyngeal, intercostal, and accessory muscles are all affected to differing extents.

The normal pattern of the oscillation between NREM and REM sleep is shown in Fig. 18.5.2.2. This ‘hypnogram’, as it is called, is constructed by classifying successive 20- or 30-s epochs from tracings of EEG, EMG, and eye movement data into either awake, movement, REM sleep, or stages 1 to 4; thus about 1000 epochs are obtained from a night’s sleep.

During wakefulness breathing is influenced by a variety of pathways, some conscious and voluntary, others entirely automatic and involuntary. Classic responses to hypoxia, hypercapnia, and vagal afferents (integrated in the brainstem) can be overruled by cortical signals to subserve functions such as talking. These two types of control are separate and can be damaged separately by disease processes. The presence of wakefulness itself provides an input to the respiratory centre, almost equivalent to the amount of ventilation seen at rest. Thus, following a period of hyperventilation, a normal subject will go on breathing at just below normal levels, despite hypocapnia and hyperoxia, until the CO2 level rises and normal ventilation is re-established. This is not true during NREM sleep, when hypocapnia will produce apnoea until the PaCO2 rises back to a critical threshold level.

Another component of wakefulness is the high muscle tone that holds the body in the required posture. This ‘awake’ input into the anterior horn cells means that other inputs, such as those from the respiratory centre, can further activate muscles, including the intercostals and pharyngeal. The withdrawal of this ‘awake’ tone with the onset of sleep means that a certain respiratory centre output to the relevant anterior horn cells is less able to raise membrane potentials to firing threshold, such that respiratory muscle activity falls with sleep onset, minute volume typically reduces by about 10 to 15%, and PaCO2 rises by 3 to 8 mmHg. Reduction of pharyngeal muscle tone narrows the lumen, and thus there is normally a rise in upper airway resistance. This reduction in ventilation has trivial effect on the arterial oxygen saturation (SaO2) in the normal circumstance when SaO2 is on the flat part of the haemoglobin dissociation curve, but dramatic falls in saturation will be apparent when SaO2 starts below 92%, the steep part of the curve, such as is seen in patients with COPD. If ventilatory responses to carbon dioxide or hypoxia are measured during NREM sleep, the slopes are flatter and right shifted, indicating a reduced overall sensitivity. Exactly why this occurs is not known, but reduced tone of the respiratory muscles, the withdrawal of awake drive, increased upper airway resistance, and (probably) true reduction in central sensitivity to CO2 or [H+], could all contribute. If, as a consequence of respiratory disease, compensatory mechanisms are already employed to cope with the extra work of breathing (sternomastoids, scalenes), then these seem to be particularly reduced during sleep as well.

During REM sleep, overall ventilation stays much the same as in NREM sleep, but the breath-to-breath variability increases considerably, sometimes with apnoeas during the actual periods of eye movements, and compensatory increases in between. Sensitivities to CO2 and hypoxia were originally thought to be further reduced, but they are hard to measure in the presence of spontaneously variable breathing and more recent evidence suggests that the response to CO2 may not be as suppressed as previously thought. What is far more important is the atonia of postural muscles. The hyperpolarization of the anterior horn cells greatly reduces the efficacy of respiratory signals to the intercostal, accessory, and pharyngeal muscles. This will not matter in a normal subject with an efficient diaphragm and an uncompromised pharynx. However, if the subject is dependent on muscles other than the diaphragm to maintain breathing, or has a narrow, compromised pharynx, then REM sleep may powerfully interfere with ventilation with consequent hypoxaemia and hypercapnia.

Also of relevance to breathing during REM sleep are the reduced arousal responses to respiratory stimuli compared with non-REM sleep. The arousal responses to some ventilatory stimuli (hypoxia, hypercapnia, extra resistive load) are believed to be mediated mainly by the perception of the ventilatory effort made in response, rather than the specific ventilatory stimulus itself. If a ventilatory response to hypoxia is measured during REM sleep, then the subject will usually tolerate a much lower SaO2 before arousing, compared to NREM sleep. Furthermore, if the drive to sleep is high, such as after sleep deprivation, arousal will be delayed still further.

It can be seen from the above that, although sleep is not a problem for those with normal respiratory systems, once abnormalities are present there is potential for a damaging interaction between sleep and breathing, particularly during REM sleep.

Obstructive sleep apnoea

Definition

Sleep apnoea was first properly documented in neurophysiological sleep laboratories using techniques that had been developed for the investigation of conditions such as insomnia, narcolepsy, and depression. It was realized that hundreds of episodes of breath cessation, or apnoea, usually due to upper airway obstruction with associated snoring, were related to marked sleep disturbance. As simple oronasal flow detectors were used, the critical event was defined as an episode of apnoea. An arbitrary definition was made, due to ease of measurement, and breath cessation for longer than 10 s became an official apnoea. Early work suggested that normal, young people rarely had more than about 30 apnoeas per night, so that the standard definition of ‘sleep apnoea syndrome’ became more than 35 apnoeas per night, or more than 5 per hour of sleep, each lasting for 10 s or longer. This definition has existed long beyond its clinical usefulness: it is quite clear that recurrent partial obstruction to the upper airway can fragment sleep just as severely with no actual apnoeas or hypopnoeas developing at all. A more pragmatic and clinically useful definition of the syndrome might now be ‘sleep disruption due to a respiratory problem engendered by sleep itself, sufficient to cause symptoms when awake’. Usually this is upper airway incompetence during sleep (obstructive sleep apnoea), but may also be due to problems of respiratory drive (central sleep apnoea, periodic breathing or Cheyne–Stokes breathing). As the pathogenesis of sleep apnoea is explained, this shift in emphasis, with the inclusion of symptoms, will become clear.

Aetiology

(Box 18.5.2.1)

The upper pharyngeal airway has to serve two functions, swallowing and breathing, which require different design features. When used for swallowing the pharynx has to behave like the oesophagus, and when used for breathing it has to remain an open tube like the trachea. These dual functions are achieved by having a floppy and collapsible muscular tube that is also capable of being held rigidly open by dilator muscles. The muscles responsible for this dilator function are discussed in the section on the structure and function of the upper respiratory tract (see Chapter 18.1.1). All these muscles have reduced activation during sleep, so that some pharyngeal narrowing occurs normally. There are, then, additional factors that determine whether this reduction leads to significant upper airflow obstruction in a particular individual. There are various theories as to these additional factors. Firstly, there may be abnormalities of the activation of the pharyngeal dilator muscles, perhaps due to defective or unstable central control. Secondly, there may be anatomical abnormalities that allow significant obstruction to occur even with the normal sleep-related reduction in muscle tone.

Neuromuscular function

Early investigations of EMG activity in pharyngeal muscles found reductions in tone with sleep during obstructive apnoeas. However, it was very difficult to show that these reductions were truly abnormal. It is now accepted that there is in fact an increase in activity of these muscles, both awake and asleep, in response to factors provoking pharyngeal collapse. In some patients with primary neuromuscular problems (from brainstem lesions to myopathies) there can be associated obstructive sleep apnoea, and pharyngeal muscle involvement seems a probable explanation. However, this is uncommon and most patients with obstructive sleep apnoea do not show evidence of any other neuromuscular problems.

During inspiration, pharyngeal dilator activity has to be synchronized with diaphragmatic activity and be adequate to overcome negative intrapharyngeal pressure. It has been suggested that a lack of coordination between diaphragmatic and pharyngeal activation may allow the pharynx to collapse as a secondary phenomenon. For example, normal subjects breathing against an inspiratory resistance can be made to have a few obstructive apnoeas by artificially inducing periodic breathing during sleep. The gradual return of respiratory drive, following the nadir of ventilation, seems to activate the diaphragm first, leaving the pharynx unbraced. The presence of an inspiratory resistance then ‘challenges’ the pharynx and allows collapse for a few breaths before pharyngeal tone returns and restores patency.

Although instability of respiratory control during sleep has been postulated as a cause of obstructive sleep apnoea, following treatment with nasal continuous positive airway pressure therapy (see later) there is very rarely evidence of a premorbid underlying respiratory instability, nor does altering respiratory drive have a useful effect. More convincing is the suggestion that there may be failure of normal reflex protective mechanisms in the pharynx, whereby receptors in the pharynx detect falls in pressure that distort the airway and provoke protective increases in pharyngeal dilator tone (see Section 18.1.1). Snoring itself may also be one of the stimuli that activate this dilator reflex, and it is conceivable that interruption of this reflex arc can occur, perhaps through years of pharyngeal trauma from snoring, mucosal oedema, or toxic agents such as cigarette smoke and alcohol.

Anatomical causes

Anatomical abnormalities influence pharyngeal function in a variety of ways. Simple encroachment of the pharyngeal lumen, e.g. with tonsillar hypertrophy, means that the normal fall in pharyngeal dilator tone with sleep can lead to critical narrowing and obstruction. Alternatively, there are abnormalities which ‘load’ the upper airway, requiring increased dilator muscle action that is then lost during sleep (e.g. high nasal resistance or increased external compression from neck obesity). Finally, there may be mechanical problems such that muscular activity fails to dilate the pharyngeal lumen effectively.

There are many case reports of obvious anatomical abnormalities provoking obstructive sleep apnoea, e.g. tonsillar hypertrophy, pharyngeal oedema, tumours, acromegaly, mucopolysaccharidoses, and mandibular or maxillary underdevelopment. These reports show that pharyngeal narrowing (asymptomatic while awake) can provoke obstructive sleep apnoea, but such diagnoses represent only a small proportion of cases.

Most patients with obstructive sleep apnoea are overweight. In many clinics the average body mass index (BMI) is well over 30 kg/m2, equivalent to being about 30% overweight, e.g. 95 kg (15 stone) at a height of 1.78 m (5 ft 10 in). Weight loss can certainly cure obstructive sleep apnoea, and all studies identifying risk factors have found obesity to be dominant, accounting for up to 40% of the variance in severity.

Most groups have found neck circumference to be a better predictor of severity of obstructive sleep apnoea than obesity itself, suggesting that it is neck obesity and external pharyngeal loading that is important (Fig. 18.5.2.3). Animal studies have shown that only a small amount of extra external pressure over the pharynx is required to collapse it during sleep, and recent imaging studies have suggested that there are quite small amounts of extra fat on either side of the pharynx in patients with obstructive sleep apnoea, together with larger amounts subcutaneously. Fig. 18.5.2.3 also shows that it is possible to have a very large neck but not have sleep apnoea; these patients were loud snorers and on average significantly younger that the rest, and it is likely that they will develop sleep apnoea as they get older.

Fig. 18.5.2.3 Correlation between sleep apnoea (number of >4% dips/h during sleep) vs neck circumference in 124 sleep clinic patients. Increasing severity is usually associated with increasing neck circumference.

Fig. 18.5.2.3
Correlation between sleep apnoea (number of >4% dips/h during sleep) vs neck circumference in 124 sleep clinic patients. Increasing severity is usually associated with increasing neck circumference.

Although general obesity is related to neck obesity, the overall correlation is only about 0.75. This is because fat distribution varies considerably between individuals. The ‘female’ distribution tends to be in the lower body and the ‘male’ distribution is more central, hence a man who is not particularly overweight can have a large neck.

As mentioned earlier, there is evidence that some of the upper airway dilator muscles (e.g. genioglossus) of obese patients with obstructive sleep apnoea are actually working harder than normal, perhaps as compensation for the added external loading from neck obesity. Compensations by the respiratory system for other types of extra loading have been shown to be much less active during sleep.

In summary, the evidence overall suggests that most obstructive sleep apnoea in adults is due to loading of the upper airway caused by obesity and fat deposits in the neck area. This external loading can be fended off during wakefulness but not during sleep, when the withdrawal of postural muscle tone allows the pharyngeal dilators to be overwhelmed, leading to excessive narrowing or collapse of the airway, with consequent apnoea.

However, not all adult sleep apnoea can be explained by obesity or intrapharyngeal anatomical abnormalities. The significance of marked retro- or micrognathia for obstructive sleep apnoea was recognized early on, particularly in children (Pierre–Robin syndrome). Careful cephalometric studies of facial and skull morphology have revealed that some patients with obstructive sleep apnoea have longer faces, retropositioning of the mandible (measured as a more acute angle between the sella to nasion and nasion to supramentale planes), a downward movement of the hyoid, elongation of the soft palate, and a narrower anteroposterior distance behind the tongue. Some, or all, of these changes may be secondary to many years of sleep apnoea rather than part of the cause, but the retropositioning of the mandible may be contributory, and surgery or dental devices to advance the mandible may be helpful in carefully selected cases. In Fig. 18.5.2.3 it will be noted that not all the patients with significant sleep apnoea had large necks, and it is these patients who had retrognathia.

Retropositioning of the mandible may be a legacy from childhood. There is good evidence that facial development is altered by nasal blockage and mouth breathing very early in life (the so-called ‘adenoidal facies’). One feature of this is mandibular retropositioning, and the mandible can return to its normal position following early adenoidectomy and resumption of nasal breathing. One theory is that mandibular underdevelopment is a risk factor for OSA that—if severe—may be adequate to provoke disease on its own, with lesser degrees acting synergistically with other causes, such as tonsillar hypertrophy in children and upper-body obesity in adults.

Other factors provoking obstructive sleep apnoea

Alcohol is a potent reducer of muscle tone and can further reduce pharyngeal dilator muscle tone during sleep. It is well known that alcohol worsens snoring, but it can also convert snoring to full apnoea. Other sedatives, such as benzodiazepines, barbiturates, and opiates, can do the same, and this has important consequences for anaesthesia in such patients. Sleep deprivation itself can reduce upper airway muscle tone during subsequent sleep, provoking a vicious circle, whereby apnoea causes sleep disruption, causing worsening apnoea.

Hyperarousability may contribute to OSA. Many people have significant increases in upper airway resistance with sleep onset, with or without snoring. After a short period of hypoventilation, stable ventilation is achieved with an increase in inspiratory muscle activity without sleep disturbance. However, if arousal is provoked before stability is achieved, then a cycle of sleep fragmentation may develop. Thus, increased arousability in conjunction with increased upper airway resistance may cause symptomatic sleep disordered breathing in a subgroup of patients, but the prevalence of this variant of OSA is unknown.

Nasal blockage can contribute to the tendency of the pharynx to collapse by lowering intrapharyngeal pressure. If extra effort has to be made to inspire through a high nasal resistance, there will be a greater vacuum effect in the pharynx, increasing its tendency to collapse. Once collapse occurs, flow ceases, pharyngeal pressure returns to atmospheric, the lumen opens, and the cycle repeats. This certainly leads to snoring, but may no longer be very important when there is full apnoea. However, nasal obstruction may contribute in the long term to sleep apnoea by damaging the pharynx through years of snoring, making it more collapsible, but improving nasal patency rarely cures obstructive sleep apnoea.

Hypothyroidism is associated with obstructive sleep apnoea, but the mechanism is not clear. It may be through weight gain, or tissue or fluid deposition in the pharynx, or a low thyroxine level may interfere directly with muscle function.

Immediate consequences of sleep apnoea

Upper airway narrowing, sometimes with complete apnoea, usually commences as sleep passes from awake to stage 2. Once significant obstruction occurs there will be increasing respiratory effort to try and overcome it. The length of such events is highly variable, ranging from only a few seconds to well over 1 min. At some point arousal occurs, with an improvement in upper airway resistance, resolution of any asphyxia, and then a return to sleep, whereupon the cycle repeats (Figs. 18.5.2.4–6). Hypoxaemia and mild hypercapnia usually accompany these periods of obstructed breathing. If there is complete apnoea, the rate of fall of SaO2 will depend mainly on the amount of oxygen stored in the lungs. This depends on the functional residual capacity since apnoeas occur at end expiration, preventing inspiration. The length of the apnoea also determines how low the SaO2 will fall, and varies considerably between patients. The consequences of such hypoxaemia and hypercapnia are not clear: the blood gas derangements are transient and may do little harm, unless they are severe and there is (e.g.) ischaemic heart disease.

Fig. 18.5.2.4 Obstructive (a) and central (b) apnoeas (16-s traces): (a) airflow ceases, but ribcage and abdominal movements persist and become paradoxical; (b) ribcage and abdominal movements cease as well as airflow.

Fig. 18.5.2.4
Obstructive (a) and central (b) apnoeas (16-s traces): (a) airflow ceases, but ribcage and abdominal movements persist and become paradoxical; (b) ribcage and abdominal movements cease as well as airflow.

Hypoxaemia was believed to play an important part in the arousal response that saves the patient from continuing asphyxia. In animal models, removal of the carotid body abolishes significant ventilatory response to hypoxaemia during sleep and there is no arousal. Giving extra added oxygen does prolong apnoeas to a small extent and delay arousal. However, recent evidence suggests that the main arousal stimulus is the actual respiratory effort being made in response to asphyxia, rather than the asphyxia per se. Normal subjects tend to wake when they have to make respiratory efforts about three times above the normal (10–20 cmH2O pleural pressure swings). This degree of effort is easily reached in obstructive sleep apnoea, when pressures down to –80 cmH2O can be recorded during the frustrated inspiratory efforts. Such pressures can also be reached by heavy snorers, even if they do not develop hypoxaemia, and this can also lead to arousals.

In terms of symptoms, the most important consequence of sleep-induced upper airway narrowing is sleep fragmentation. The original methodology of sleep analysis, using coarse 30-s epochs to stage sleep, effectively glossed over the multitude of transient arousals that are the main consequence of obstructive sleep apnoea. Superficially, a sleep hypnogram in a moderately severe case (see Fig. 18.5.2.2) could look almost normal despite hundreds of arousals. The importance of trying to measure these has recently been appreciated, and there are a variety of techniques to measure, or infer, these arousals. However, the level of sleep disruption (e.g. number and degree of arousals) necessary to cause daytime symptoms is not known. There is a clear, but variable, relationship between increasing sleep disruption and deteriorating daytime function, but there is no clear cut off between normality and abnormality.

In addition to blood gas disturbances and sleep disruption, there are many other consequences of obstructive sleep apnoea. During the apnoea there is activation of the diving reflex that produces bradycardia and arteriolar vasoconstriction in muscles, particularly when there is associated hypoxaemia. Upon arousal there is a sudden pulse rate and blood pressure rise, probably due to general activation of the vascular sympathetic nervous system as part of the arousal process itself. During the actual frustrated inspiratory efforts, blood pressure falls with each reduction in intrathoracic pressure (pulsus paradoxus) and, in conjunction with the blood pressure rise on arousal, produces a very characteristic trace (Fig. 18.5.2.5). As well as increased nocturnal catecholamine secretion in patients with obstructive sleep apnoea, there is also a suppression of growth hormone and possibly testosterone levels. There is marked polyuria during sleep (a reversal of the normal relative oliguria), but the mechanism is not clear. It may be related to the recurrent arousals, or to increased natriuretic peptide (ANP and BNP) production following cardiac chamber distension due to the large inspiratory efforts. There is some recent evidence that OSA may provoke atrial fibrillation, but whether this is via increased sympathetic activity or perhaps atrial distension is not known.

Fig. 18.5.2.5 A 5-min tracing from a patient with obstructive sleep apnoea. The rises in blood pressure (top trace) and heart rate (second trace) coincide with the cessation of each apnoea and an arousal. During each apnoea (evident from the bottom airflow trace) each frustrated inspiratory effort is accompanied by a fall in blood pressure (pulsus paradoxus).

Fig. 18.5.2.5
A 5-min tracing from a patient with obstructive sleep apnoea. The rises in blood pressure (top trace) and heart rate (second trace) coincide with the cessation of each apnoea and an arousal. During each apnoea (evident from the bottom airflow trace) each frustrated inspiratory effort is accompanied by a fall in blood pressure (pulsus paradoxus).

It will be clear from this account that there are areas of uncertainty regarding definition and measurement of some aspects of sleep apnoea. We discussed earlier that original definitions centred on the actual obstructive event. There has now been a shift towards trying to look more closely at the most important result—sleep fragmentation. This is particularly necessary now we know that 10-s apnoeas are not the only result of upper airway narrowing during sleep that can provoke multiple arousals and daytime sleepiness. Since heavy snorers can have considerable sleep fragmentation without significant falls in SaO2, examining blood gas abnormalities (e.g. with an oximeter) is not always good enough either. The implications for this in terms of investigations are discussed later. A considerable amount of effort is being put into establishing which variable that can be measured during a sleep study best defines the severity of the disorder. At present there is no clear answer, although there is good evidence that simple approaches are at least as good as the more complex approaches often used, but different approaches to diagnosis and management remain.

Symptoms and presentation

Sleep-related disorders of breathing Case History—A 52 yr old man referred for assessment of snoring.

The main symptom of obstructive sleep apnoea is daytime hypersomnolence, and this correlates broadly with the degree of sleep disruption. Early in the development of the disorder the daytime sleepiness is little more than often experienced by normal people after a few disturbed nights. While occupied, the individual has little difficulty in concentrating and staying awake, but once activities become more boring, unwanted sleepiness intervenes. Initially this may be viewed as normal, such as falling asleep in front of the television every evening. As the sleep disruption worsens there will be interference with an increasing number of activities. Of particular importance is sleepiness while driving. Sleepiness can be devastating, particularly on long motorway journeys after dark, when sensory stimulation is low. Initially there will be lane wandering, with sudden arousal and correction. Accidents involving driving off the road, or driving into vehicles in front, are more common in patients with obstructive sleep apnoea and reduce with treatment. Sleepiness also impinges greatly on work performance and home life. The patient will develop a reputation for slothfulness and lack of interest.

It is important to ask the correct questions to assess sleepiness. It is not the same as tiredness, which is a lack of energy or desire to get up and do anything, without a desire to sleep. Because of the insidious onset of obstructive sleep apnoea, any sleepiness may be regarded as normal by the patient, and thus situational questions need to be asked such as, ‘How often do you have to pull off the road while driving owing to sleepiness?’ rather than just ‘Are you sleepy?’ A well validated and simple way to do this is with the Epworth Sleepiness Scale (Fig. 18.5.2.7). Objective sleepiness can be assessed by measuring how long the patient takes to fall asleep when asked to stay awake, lying down in a quiet room on a number of occasions across the day. This is useful for research purposes but only occasionally adds to the clinical management of such patients. A list of other symptoms seen in obstructive sleep apnoea is given in Box 18.5.2.2. It is sad to say, but the corrosive effect of sleepiness on all aspects of a patient’s life has often been present for years before someone (usually not a doctor) tumbles to the diagnosis.

Fig. 18.5.2.7 Questionnaire scale to assess subjective sleepiness. The scores for each answer (0–3) are summed to give a range from 0 (no sleepiness at all) to 24 (maximally sleepy). The upper limit of normal is about 9, and most patients with symptomatic obstructive sleep apnoea are in the middle teens.

Fig. 18.5.2.7
Questionnaire scale to assess subjective sleepiness. The scores for each answer (0–3) are summed to give a range from 0 (no sleepiness at all) to 24 (maximally sleepy). The upper limit of normal is about 9, and most patients with symptomatic obstructive sleep apnoea are in the middle teens.

A typical case history would be that of a middle-aged man complaining of increasing daytime sleepiness. It is usually some specific event that prompts initial consultation, such as falling asleep while driving or operating machinery, or during an important business meeting. There will be a long history of gradually worsening snoring with apnoeas, possibly witnessed by the spouse, who will probably have moved out of the bedroom because of the noise. There is likely to have been a weight gain over the last few years with a BMI of greater than 30kg/m2 and a collar size of 17 in (43 cm) or more. There is sometimes a history of fairly high alcohol intake and smoking. On examination there may be nasal stuffiness, evidence of a small lower jaw (such as teeth crowding or several extractions for this problem), and a small pharynx with mucosal bogginess and wrinkling. Of course, it should be stressed that not all these features are likely to be present in one individual.

Part of the history and examination of patients with possible obstructive sleep apnoea should be directed towards precipitating factors such as hypothyroidism and acromegaly. Other diagnoses such as mucopolysaccharidosis, pharyngeal tumours, tonsillar hypertrophy, neurological disorders, and significant retrognathia will be more obvious. There should be a higher index of suspicion for causes other than obesity when the neck circumference is not raised.

Diagnosis

Following the history and examination, further outpatient tests may be appropriate, e.g. thyroxine or growth hormone estimations. Blood gases and simple lung function tests may be necessary if associated diurnal respiratory failure is suspected. A raised haemoglobin may also signify diurnal respiratory failure, as will a raised venous bicarbonate. A raised PaCO2 should suggest the possibility of lower airways obstruction (so-called ‘overlap syndrome’) because patients with ‘pure’ OSA rarely have hypercapnia when awake: an additional factor(s) such as COPD or morbid obesity usually needs to be present. Obstructive sleep apnoea tends to go with findings that constitute the so-called ‘metabolic syndrome’, i.e. hypertension, central obesity, raised triglyceride levels, reduced HDL cholesterol, and insulin resistance/raised fasting plasma glucose. Blood pressure, fasting blood sugar, and lipids should therefore be measured as part of good care and opportunistic screening.

Unless the presenting problem turns out not to be sleep related, some form of sleep study will be required. In the past, the usual procedure was to employ full polysomnography, which measured sleep state and respiratory variables (see Figs. 18.5.2.2 and 18.5.2.4). However, this investigation and its analysis is expensive and time consuming, particularly if all recurrent arousals are documented. The primary requirements are to (1) assess sleep fragmentation, (2) establish if a respiratory problem is responsible, and (3) decide if upper airway obstruction is the primary cause. Full polysomnography, properly interpreted, will usually allow this, with the EEG and EMG giving good information on sleep disruption, and aspects of respiration deduced from ribcage/abdominal movement transducers, oronasal airflow, and snoring and continuous oximeter recordings. However, there is considerable signal redundancy in such recordings, and the essential derivatives—sleep disruption and ventilation—can be assessed in much simpler ways (see Fig. 18.5.2.6). In view of this, most clinical respiratory sleep laboratories have abandoned routine, conventional polysomnography because of its unnecessary expense.

Fig. 18.5.2.6 Short sleep tracings of body movement, SaO2, pulse rate, and snoring level in four different subjects. (a) Normal subject (no fluctuations in any signals), 20 min. (b) Patient with continual low level snoring and almost no arousals, 20 min. (c) Patient with classical obstructive apnoeas, evident from the snoring–silence–snoring pattern together with movements and oscillations in the pulse and SaO2, 20 min. (d) Patient with periodic movements of the legs during sleep, recurrent arousal (oscillations in pulse and body movements), but no evidence of a respiratory cause (no snoring or SaO2 dips), 10 min. A video recording of the whole night is always available and can be viewed when the exact cause of abnormal signals is not immediately obvious.

Fig. 18.5.2.6
Short sleep tracings of body movement, SaO2, pulse rate, and snoring level in four different subjects. (a) Normal subject (no fluctuations in any signals), 20 min. (b) Patient with continual low level snoring and almost no arousals, 20 min. (c) Patient with classical obstructive apnoeas, evident from the snoring–silence–snoring pattern together with movements and oscillations in the pulse and SaO2, 20 min. (d) Patient with periodic movements of the legs during sleep, recurrent arousal (oscillations in pulse and body movements), but no evidence of a respiratory cause (no snoring or SaO2 dips), 10 min. A video recording of the whole night is always available and can be viewed when the exact cause of abnormal signals is not immediately obvious.

Sleep fragmentation can be inferred from a variety of signals. The most sensitive appears to be autonomic markers of brainstem activation, such as blood pressure and pulse rate rises. In addition, since most abnormal respiratory events will end in some form of arousal, counting body movements provides some guide to the degree of sleep fragmentation, and may be most predictive of daytime symptoms. Upper airway obstruction can be inferred from snoring, a particular inspiratory pattern on a nasal flow tracing (flow limitation), paradoxical ribcage/abdominal movements, and from pulsus paradoxus visible on a beat-to-beat blood pressure tracing (now easily obtainable non-invasively, Fig. 18.5.2.5). Many simple, commercial monitoring systems can be used to record these signals and to assess the extent of the sleep fragmentation and whether upper airways obstruction is the likely cause. Recent work suggests that these simpler measures can predict sleepiness in obstructive sleep apnoea, and its response to treatment, at least as well as EEG-based approaches, which are clearly not the ‘gold standard’ they were once thought to be. The attention paid to each signal, and perhaps the exact sleep study system used, will depend to some extent on the condition under investigation.Fig. 18.5.2.6 shows data provided by the system in routine use in our laboratory, designed primarily to identify obstructive sleep apnoea and its variants, but it will also identify central sleep apnoea (see below) and nonrespiratory problems such as periodic movements of the legs during sleep.

Because of the imprecise relationship between the number of abnormalities on a sleep study and the severity of symptoms, trying to count them precisely is pointless, particularly given that there can be considerable night to night variation. Hence, the reporting of sleep studies tends to be more qualitative than previously, with divisions simply into mild, moderate, and severe. These bandings are sometimes arbitrarily allocated on the basis of event rates, with 5 to 15 respiratory events per hour being mild, 15 to 30 moderate, and more than 30 severe—the aim simply being to see if there is an adequate and understandable explanation for the patient’s symptoms, and assess the severity of any problem.

Differential diagnosis

Obstructive sleep apnoea syndrome (i.e. the combination of sleep apnoea plus symptoms) should be relatively easy to diagnose in a patient who snores, has witnessed apnoeas, is sleepy, and has a compatible abnormal sleep study. However, there are alternative diagnoses that should be considered, from either the history or sleep study.

Periodic leg movements during sleep

Sleep fragmentation leading to daytime somnolence can be due to periodic leg movements during sleep which lead to multiple arousals during NREM sleep. These leg movements are common in older people and may not provoke symptoms, but they can cause excessive daytime sleepiness if they provoke significant recurrent arousals over an extended period of time, with such patients usually having restless leg syndrome when awake, particularly in the evening. Periodic leg movements during sleep usually occurs in isolation but can be associated with renal impairment (especially if on dialysis), low ferritin level, peripheral neuropathy, and previous sciatica. It is easily diagnosed by sleep study (see Fig. 18.5.2.6d) and, if sufficiently symptomatic, treated with dopaminergic agonists.

Narcolepsy

Narcolepsy is due to destruction of orexin (also known as hypocretin) neurons in the hypothalamus that are involved in the maintenance of wakefulness and muscle tone. This destruction is thought to be immunological, and there is a very strong association with HLA DQB1 0602 subtype (>95% compared to 30% in the general population). The condition will present to sleep physicians, although snoring and witnessed apnoeas are not usual features. Vivid disturbing dreams (particularly at sleep onset), sleep paralysis, uncontrollable somnolence, and cataplexy (loss of muscle tone/falling in response to strong emotion or laughter) should suggest the diagnosis. There may be a family history. The patient should be HLA typed, with absence of the relevant HLA type strongly refuting the diagnosis. The history, rather than a sleep study, is the mainstay of diagnosis, although recognition of early onset REM sleep during daytime naps is sometimes helpful. Where available, measurement of orexin in cerebrospinal fluid is being used by some to help with diagnosis. Specialist referral is required for diagnosis and management: narcolepsy is a lifelong condition with significant ramifications for employment and driving. See Chapter 24.5.2 for further discussion.

Multisystem atrophy

Multisystem atrophy (Shy–Drager syndrome) can present to sleep clinics with apparent snoring and sleep fragmentation. In fact the ‘snoring’ is due to laryngeal abductor weakness and laryngeal closure during sleep, with stridulous inspiratory obstruction. It is important to distinguish this stridor from the usual pharyngeal snoring of OSA because these patients can suddenly die from nocturnal respiratory arrest. They can be successfully treated with standard CPAP therapy (see ‘Treatment’, below).

Patients with multisystem atrophy and Parkinson’s disease may also occasionally present with REM sleep behaviour disorder. This strange phenomenon is due to loss of the atonia normally present during REM sleep, which allows extensive movements while dreaming that appear to match the dream content. The patient may become aggressive and sometimes attack the sleeping partner. It is due to damage to Jouvet’s centre in the pons (responsible for the normal atonia) and is strongly associated with the development of multisystem atrophy, often occurring many years later. It usually responds to clonazepam, which generally reduces muscular tone.

Other causes of excessive daytime sleepiness are listed in Box 18.5.2.3 but do not have apnoeic episodes on sleep study; and see Chapter 24.5.3 for further discussion.

CVA, cerebrovascular accident.

Sleep study misinterpretation

Central sleep apnoeas (see below) can occasionally be difficult to distinguish from obstructive sleep apnoeas as the patient may have a few obstructed breaths at the end of the apnoea cycle, which are secondary to the central apnoea. Conversely, patients who are morbidly obese or with inspiratory muscle weakness may have such small chest-wall movements against the upper airway obstruction that they are misinterpreted as having central apnoeic events.

Overlap syndromes

It is also important to realize that OSA may coexist with other respiratory disorders that can have an additive effect, leading to profound respiratory failure. The term ‘overlap syndrome’ was originally coined to refer to the combination of chronic obstructive pulmonary disease and OSA, but morbid obesity can interact with both of these conditions so that the term is often now used to describe mixtures of these three provokers of ventilatory failure. When obesity alone provokes ventilatory failure then the term ‘obesity hypoventilation syndrome’ (previously Pickwickian syndrome) is used.

Treatment

Once it is established that the patient’s symptoms are likely to be due to sleep disruption from sleep-induced upper airway obstruction, then therapy has to be tailored to symptom severity.

Advice and simple treatments

Mild symptoms may resolve with simple treatments and advice (Box 18.5.2.4). Weight loss is undoubtedly effective, but often very difficult to achieve. If sleep disruption only occurs while the individual is supine, when upper airway obstruction tends to be worst, then learning to lie on one’s side may be helpful. Stopping sedatives and evening alcohol can help. Initial enthusiasm for the tricyclic antidepressants has waned, although they may slightly improve mild cases. They are believed to work through REM sleep suppression and by improving upper airway tone. No other drug has shown any consistent effect.

Nasal continuous positive airway pressure

If the sleep apnoea and symptoms are severe, there is only one fully effective therapy—nasal continuous positive airway pressure (NCPAP). This treatment involves wearing a small mask (Fig. 18.5.2.8) over the nose while asleep, with the air pressure kept at a fixed level above atmospheric by a pump. Pressures in the region of 10 cmH2O are enough to splint open the pharynx and resist collapse, allowing unobstructed breathing and undisturbed sleep (Fig. 18.5.2.9). The exact pressure required can be established in a number of ways. The response is dramatic, in terms of both physiology and daytime symptoms, which resolve rapidly even after one night of treatment. There are several randomized placebo-controlled trials and a recent Cochrane review proving beyond doubt the large symptomatic benefit. The unpleasantness and unaesthetic appearance of this treatment initially repel patients, but once the benefits have been experienced, acceptance is high, and indeed the compliance has been shown to be better than that with asthma therapy or antihypertensives. Off-the-shelf systems, with comfortable soft masks, are now available for home use at about £300 each. Such equipment will last for many years (with regular mask replacements) and represents extraordinary value for money given the enormous improvement in quality of life that it produces. NCPAP machines that automatically hunt, moment to moment, the pressure required by the patient to overcome their obstructive sleep apnoea are available. These can be used either to establish the fixed pressure subsequently required by the patient, or as the long-term machine at home, but there is no evidence that their extra expense is justified except in patients requiring higher pressures.

Fig. 18.5.2.8 A soft silicone nasal mask and its headgear, used in the treatment of obstructive sleep apnoea.

Fig. 18.5.2.8
A soft silicone nasal mask and its headgear, used in the treatment of obstructive sleep apnoea.

Fig. 18.5.2.9 Two all-night oximetry tracings from a patient with obstructive sleep apnoea, before treatment and during his first night on nasal CPAP. Each tracing starts top left and finishes bottom right. Each tracing is continuous for 8 h with the vertical axis for each individual line scaled 70–100% SaO2.

Fig. 18.5.2.9
Two all-night oximetry tracings from a patient with obstructive sleep apnoea, before treatment and during his first night on nasal CPAP. Each tracing starts top left and finishes bottom right. Each tracing is continuous for 8 h with the vertical axis for each individual line scaled 70–100% SaO2.

Experienced sleep clinics put much effort into helping patients to become established on NCPAP, through attentive education and comfort-improving measures such as humidification. Once established on NCPAP, patients with obstructive sleep apnoea are likely to require it for life unless they can lose a significant amount of weight.

Surgical and other treatments

Weight loss may only be achieved through gastric surgery, such as silastic ring gastroplasty to reduce food consumption.

Another surgical treatment is uvulopalatopharyngoplasty, which consists of removing part of the soft palate and any residual tonsils, and ‘tightening up’ the side walls of the pharynx. Although it can reduce snoring, its success rate at treating obstructive sleep apnoea is not good, and a recent Cochrane review has confirmed its lack of efficacy. Attempts to select patients who might respond has had very limited success, although thin patients with large soft palates, residual tonsils, and milder disease probably do best. By contrast, there is good evidence that tonsillectomy is effective in children with OSA and large tonsils.

Other operative techniques involving advancement of the mandible (and sometimes the maxilla) may be appropriate in highly selected cases. Tracheostomy was the first therapy ever tried and was (of course) very effective: it may still be appropriate in occasional patients.

A newer approach has been the use of mandibular advancement devices, worn in the mouth at night (Fig. 18.5.2.10). These hold the lower jaw closed and forward, thus increasing the space behind the tongue and hence pharyngeal volume. They have undergone extensive trials in a variety of situations, but matters are complicated by the plethora of such devices available. The current conclusion is that they do work, but less so as the severity of the obstructive sleep apnoea (and usually therefore the obesity) increase. Their main use seems to be in the control of unacceptable snoring, mild OSA, or as an alternative to NCPAP for short periods, or when NCPAP cannot be tolerated despite maximal support.

Fig. 18.5.2.10 Example of a mandibular advancement device, worn in the mouth at night. These hold the lower jaw forward and closed, thus increasing pharyngeal dimensions. They are used extensively for the control of snoring but are not very effective in anything more severe than mild obstructive sleep apnoea.

Fig. 18.5.2.10
Example of a mandibular advancement device, worn in the mouth at night. These hold the lower jaw forward and closed, thus increasing pharyngeal dimensions. They are used extensively for the control of snoring but are not very effective in anything more severe than mild obstructive sleep apnoea.

Epidemiology

Given the difficulties over definition, the prevalence of symptomatic obstructive sleep apnoea is hard to establish, and will depend on where an arbitrary cut-off is placed. In an early study, about 0.3% of men aged 35 to 65 years clearly had severe, symptomatic obstructive sleep apnoea, requiring nasal continuous positive airway pressure (NCPAP) therapy and were responsive to such treatment. However, about 5% had more than five dips of more than 4% SaO2 per hour, one suggested threshold for normality; most of these subjects were not obviously symptomatic and would not have wanted a treatment such as NCPAP. Overall in this study, sleepiness correlated with snoring, and more sleepiness seemed to be due to snoring than classical sleep apnoea. Other studies in Israel, the United States of America, and Italy have found prevalences of ‘significant’ sleep apnoea in the 0.5 to 2% range.

Predictors of sleep apnoea in these prevalence studies have been obesity, snoring, age, self-reported sleepiness, and alcohol consumption. Snoring is more common in men than in women, and obstructive sleep apnoea syndrome itself is about three to five times more common in men. The prevalence in women probably increases after the menopause with redistribution of body fat to a more male-like, upper body, distribution. Rapidly increasing levels of obesity in many countries means that the prevalence of OSA is increasing.

If these prevalence studies are correct, then obstructive sleep apnoea is the third most common serious respiratory disease after asthma and COPD. Sleep apnoea is now the commonest reason for referral in some respiratory units.

Prognosis and long-term complications

Many patients with OSA are of the physiognomy to have associated visceral obesity and the metabolic syndrome (hypertension, insulin resistance, hyperlipidaemia). The vascular mortality in these patients is in general higher than average, their 10-year vascular event rate being approximately 36%. However, it is not clear whether OSA is a further independent vascular risk factor, or merely associated. This is important because many sleep units still treat patients mainly on the basis of their daytime symptoms, usually excessive sleepiness, and not to reduce future vascular risk. To do so would greatly increase the requirement for nasal CPAP, perhaps three- or fourfold based on epidemiological studies of sleep-study-defined OSA, rather than OSA syndrome (with symptoms).

There are plausible hypotheses for a causative link between OSA and vascular disease. With the arousal after each apnoea or hypopnoea, there is a surge in blood pressure, often by over 50 mmHg. Overnight blood pressure levels do not fall normally, itself an adverse vascular risk factor, and there is good evidence of raised daytime blood pressure in patients with OSA syndrome. Nasal CPAP treatment not only abolishes the nocturnal blood pressure surges, but also reduces the daytime blood pressure (Fig. 18.5.2.11).

Fig. 18.5.2.11 Mean ambulatory blood pressure profile in two groups of patients with OSA- a control group (given subtherapeutic CPAP, º) and a treatment group (given therapeutic CPAP, •). The top panel shows the two groups’ profiles before any treatment, confirming that they had similar blood pressure profiles. The bottom panel shows the groups after treatment, showing that the group given therapeutic CPAP had a lower mean blood pressure profile.

Fig. 18.5.2.11
Mean ambulatory blood pressure profile in two groups of patients with OSA- a control group (given subtherapeutic CPAP, º) and a treatment group (given therapeutic CPAP, ). The top panel shows the two groups’ profiles before any treatment, confirming that they had similar blood pressure profiles. The bottom panel shows the groups after treatment, showing that the group given therapeutic CPAP had a lower mean blood pressure profile.

Reproduced from Pepperell et al. Lancet. 2001; 359:204–210 with permission from Elsevier.

Uncontrolled work has suggested increased oxidative stress, increased coagulability, increased inflammatory markers, and increased cholesterol levels as possible mechanisms for increased vascular damage. Cross-sectional studies have shown an association between OSA and hypertension, stroke, and other vascular events, but despite trying to control for confounders such as visceral obesity it is impossible to be sure that any residual effects of OSA are not due to a further confounder, or one inadequately controlled for. For example, surface measurements of central obesity (waist, or waist/hip ratio) are poor measures of true visceral obesity, and OSA is particularly present in those with visceral obesity.

Uncontrolled interventional studies have strongly suggested that treatment of OSA reduces vascular events compared to untreated control patients (Fig. 18.5.2.12). Other studies have shown that noncompliant patients in general are clearly different from compliant ones, with considerably higher vascular mortality, perhaps due to poor compliance with a range of other healthy activities and medications. As yet there are no controlled interventional long-term morbidity and mortality studies of nasal CPAP for OSA, hence there is no evidence-based justification for routine use of nasal CPAP to reduce vascular risk in patients with OSA who would not otherwise want to use nasal CPAP for relief of daytime sleepiness.

Fig. 18.5.2.12 12-year follow-up (in months) of patients with varying degrees of OSA, showing cardiovascular events in various groups of patients and controls. The group with severe untreated OSA (OSAH) did badly compared to a treated group (OSAH with CPAP), those with mild OSAH, snorers and normal controls. However, this was an uncontrolled study in which patients either chose or declined to be treated, which may have led to bias.

Fig. 18.5.2.12
12-year follow-up (in months) of patients with varying degrees of OSA, showing cardiovascular events in various groups of patients and controls. The group with severe untreated OSA (OSAH) did badly compared to a treated group (OSAH with CPAP), those with mild OSAH, snorers and normal controls. However, this was an uncontrolled study in which patients either chose or declined to be treated, which may have led to bias.

Reproduced from Marin et al Lancet 2005; 365:1046–53, with permission from Elsevier.

Driving regulations and obstructive sleep apnoea

There have been some recent high-profile cases of fatal road traffic accidents caused by drivers falling asleep at the wheel, some of which have led to prison sentences. This has highlighted the issue for patients with OSA who are up to seven times more likely to have a road traffic accident than controls, although this higher incidence is probably only present at the more severe end of the spectrum. In the United Kingdom the law says that drivers are responsible for their vigilance while driving and should not drive when sleepy for whatever reason. OSA syndrome (i.e. with symptoms) is a notifiable illness, and such patients should inform the Driver and Vehicle Licensing Agency (DVLA). Licences will only be revoked if there is continuing sleepiness, thus once on treatment full licensing is possible, hence a professional diagnosis and confirmation of response to treatment is required. Uncontrolled evidence shows normal road traffic accident rates following treatment, and work using driving simulators has confirmed the beneficial effect of nasal CPAP.

Treatment costs for OSA would be covered, several times over, by the costs incurred by avoidable road traffic accidents amongst patients with this condition. In addition, undiagnosed patients with OSA have increased general health costs in the years leading up to their diagnosis, which fall afterwards.

Sleep-induced hypoventilation and central sleep apnoea

Breathing during sleep may decrease, not because of upper airway obstruction, but because of a reduction in central output to the respiratory muscles—so-called central, rather than obstructive, apnoea, or hypoventilation (see Fig. 18.5.2.4). There are many causes for central sleep apnoea or hypoventilation: Table 18.5.2.2 shows one way of classifying them. Some of the central apnoeas disturb sleep and present with daytime sleepiness, whereas others tend to present with symptoms of respiratory failure, such as morning headaches with confusion, cyanosis, and ankle oedema.

Table 18.5.2.2 Types of central sleep apnoea

Type

Examples

Daytime arterial CO2 level

Absent or reduced ventilatory drive (Ondine’s curse)

Brainstem damage or congenital abnormality

Raised

Acquired blunting, e.g. secondary to lung disease

Unstable respiratory drive

Sleep onset, hypoxaemia, altitude, heart failure

Normal or low

REM-related oscillations

Normal in REM sleep

Normal or raised

Due to neuromuscular disorders and respiratory muscle weakness.

Reflex central apnoea

Pharyngeal collapse inhibits inspiration

Normal

Apparent central apnoea (wrongly diagnosed)

Respiratory muscle weakness or gross obesity cause chest wall movement transducers to fail to demonstrate any ventilatory effort during obstructive apnoeas

Normal or raised

Absent ventilatory drive

Brainstem abnormalities may damage the areas responsible for automatic chemical control of ventilation. While awake, the wakefulness-related ventilatory drive may be adequate to maintain PaO2 and PaCO2 levels, but on falling asleep the drive falls or even disappears, with marked hypoventilation (or apnoea) and hypoxaemia. Arousal is then necessary to restore the blood gases. This failure of brainstem automatic control (known as Ondine’s curse) can be congenital, or may be acquired as the result of a stroke, infection, surgical damage, multiple sclerosis, or compression by a tumour or syrinx. Congenital causes have been shown to be due to a variety of different genetic abnormalities associated with the development and differentiation of neural crest tissue.

Reduction of chemical drive can occur as a secondary problem when ventilation is reduced by mechanical problems such as chronic airways obstruction or weak respiratory muscles. It appears that chronic underventilation can lead to blunting of ventilatory drive, perhaps through alteration in acid–base buffering in the brainstem, and can also be associated with marked falls in ventilation during sleep.

Unstable ventilatory drive

The wakefulness-related ventilatory drive stabilizes ventilation and prevents it from falling below a certain level. If reasons for ventilatory instability exist then, by removing this stabilizing effect, sleep will allow periodic respiration to develop and be maintained. The usual provoker of instability is an increased gain in the ventilatory response to a stimulus, usually hypoxia or hypercapnia. Control theory shows that increasing the gain in any feedback system causes instability through overshoot and undershoot. A good example is the effect of altitude hypoxaemia on newly arrived lowlanders. While awake the breathing is driven by hypoxia, but also limited by the hypocapnic alkalosis. When sleep occurs, and drive is therefore reduced, hypoventilation ensues until the rise in PaCO2 and fall in PaO2 restarts ventilation. Often this leads to sufficient hyperpnoea to arouse the individual who now has extra hypoxia, which steepens the CO2 response line and, along with the return of wakefulness drive, produces marked overshoot of ventilation. On returning to sleep with much improved PaCO2 and PaO2, hypopnoea or apnoea recurs, and the cycle is perpetuated. Thus, periodic breathing with recurrent arousals is very common at altitude, with the expected daytime consequence of sleepiness and complaints of insomnia. Acetazolamide produces a metabolic acidosis and increases the ventilation at a given PaCO2, the hypoxaemia is relieved, and thus the ventilatory response to CO2 becomes less steep. Both these factors restore stability and reduce periodic respiration.

In left ventricular failure there is also extra ventilatory drive, mainly due to stimulation of interstitial lung receptors (J-receptors) by the raised left atrial and pulmonary venous pressure. In conjunction with the longer circulation time seen in heart failure, this also provokes instability with waxing and waning of the ventilation. This periodic breathing, or Cheyne–Stokes respiration, is quite common in heart failure and, through sleep disruption, produces daytime sleepiness and complaints of nocturnal dyspnoea (Fig. 18.5.2.13). The patient is usually aware that on arousal the dyspnoea disappears within a few seconds, unlike the paroxysmal nocturnal dyspnoea of pulmonary oedema which usually takes at least 15 min or so to abate following getting out of bed. Treatment with overnight oxygen, acetazolamide or benzodiazepines can sometimes reduce the periodicity and improve both sleep quality and symptoms. There has been some recent interest in NCPAP as a treatment for left ventricular failure and periodic breathing. However, a recent randomized controlled trial showed no benefit on mortality in patients with dominantly Cheyne–Stokes respiration, although there is good evidence of benefit to cardiac function from treating obstructive sleep apnoea in patients with left ventricular failure.

Fig. 18.5.2.13 Tracing of Cheyne–Stokes respiration from a patient with poor left ventricular function but no radiological or clinical evidence of current pulmonary oedema. With each return of respiration there is arousal from sleep (not clearly visible with this compressed EEG tracing).

Fig. 18.5.2.13
Tracing of Cheyne–Stokes respiration from a patient with poor left ventricular function but no radiological or clinical evidence of current pulmonary oedema. With each return of respiration there is arousal from sleep (not clearly visible with this compressed EEG tracing).

Instability of respiratory control can also occur in normal subjects in the early stages of sleep, or if sleep is disturbed for other reasons. This is because sleep depth is oscillating back and forth between drowsy wakefulness and light sleep, with the ventilatory drive oscillating as well.

REM sleep apnoeas

During normal REM sleep the phasic bursts of eye movements are associated with transient falls in ventilation, and even the occurrence of apnoeas. The ribcage muscles are affected most, but diaphragmatic excursion can also fall. Such periodicities are entirely normal.

As discussed earlier, the REM sleep inhibition of most muscles (apart from the diaphragm) can greatly reduce overall ventilation when the accessory muscles of respiration are needed for breathing. Thus on entering REM sleep there can be profound falls in ventilation and SaO2 in patients with neuromuscular diseases, chest-wall abnormalities, and chronic airways obstruction.

Generalized neuromuscular diseases tend to involve the respiratory muscles in concert with other muscles. However in some disorders the respiratory muscles (and particularly the diaphragm) may be involved very early on, at a time when other muscles are virtually normal. A particular example of this is adult-type acid maltase deficiency, where patients may present in respiratory failure while still able to walk normally. REM-sleep-related hypoxaemia may be the first sign that there are problems, and it is not known whether this actually accelerates the onset of eventual diurnal respiratory failure, or is merely a marker that respiratory failure will soon follow. Sometimes there may be associated upper airway obstruction during REM sleep, leading to even larger falls in SaO2. Overnight oximetry studies will indicate the degree of hypoxaemia but will not establish if there is additional upper airway obstruction.

There has been great interest in the REM-sleep-related hypoxaemia seen in chronic airways obstruction. It was thought possible that these hypoxic episodes might be the reason why some patients developed respiratory failure but others did not. However, it appears that REM sleep hypoventilation and a fall in PaO2 is fairly universal in this group of patients. If the patient is initially well oxygenated and on the flat part of the haemoglobin dissociation curve, the fall in SaO2 (which is usually what is monitored) is not particularly dramatic; however, if the patient is initially poorly oxygenated and on the steep part of the curve, similar hypoventilation will produce dramatic falls in SaO2. As yet there is no evidence that these REM sleep falls in SaO2 contribute to the morbidity and mortality of patients with chronic airflow obstruction, although some centres have shown that overnight oxygen therapy reduces arousals, thus improving sleep quality. The main problem is that the falls in SaO2 can look superficially like obstructive sleep apnoea, leading to an erroneous diagnosis and the inappropriate use of NCPAP.

Reflex apnoea

Central respiratory output can be modified by a number of reflexes from receptors in the upper airway. There is a reflex from the pharynx that inhibits inspiratory flow when the pharynx is being sucked in and collapsed, which makes teleological sense as a slowing of inspiratory flow would reduce the tendency to collapse. There are some patients with pharyngeal collapse who, instead of struggling to inspire against the blocked airway, simply stop breathing until they finally arouse, presumably due to the fall in PaO2 and rise in PaCO2 itself. This then appears as a central apnoea, despite the aetiology being pharyngeal collapse, with the problem tending to happen when the patient is supine, with snoring or ordinary obstructive apnoeas when decubitus. Evidence that superficial receptors are responsible comes from the observation that inspiratory attempts return if the pharynx is anaesthetized. These patients usually present with histories typical of obstructive sleep apnoea, respond to NCPAP, and should be managed in the same way.

Apparent central apnoea

The diagnosis of central apnoea depends on demonstrating the absence of respiratory effort when airflow at the nose and mouth stops. Surface measurements of ribcage and abdominal movement are sometimes employed as evidence of continuing respiratory effort, but in two circumstances—marked obesity and muscular weakness—the surface transducers may fail to register that inspiratory efforts are still being made (although more sensitive measures of inspiratory effort, such as oesophageal pressure tracings, will usually do so). Obesity lessens the sensitivity of surface transducers, and with muscle weakness the inspiratory muscles may not be able to move the chest wall detectably against a closed upper airway. In severe OSA with diurnal ventilatory failure, the same situation may develop where inspiratory effort is so reduced that a misdiagnosis of central apnoea is made. Such cases of apparently pure obesity—hypoventilation then subsequently turn out be OSA. This is usually revealed when ventilatory failure has been reversed by a period of overnight ventilation, and the sleep study repeated later, when the upper airway obstruction and frustrated inspiratory efforts become obvious.

Overnight ventilation for central sleep apnoea or hypoventilation

The chronic ventilatory failure associated with some neurological disorders (e.g. acid maltase deficiency, postpoliomyelitis syndrome, motor neuron disease, Duchenne dystrophy) usually progresses rapidly to death, even when quality of life is otherwise very good. The same is true of chest-wall restrictive disorders such as scoliosis, as well as the ventilatory failure that can develop many years after extensive thoracoplasty. However, supporting breathing overnight can fully reverse ventilatory failure, and the response to treatment can be dramatic, with resolution of all symptoms and restoration of normal blood gases, even when off the ventilator during the day.

The mechanism by which supporting breathing at night corrects ventilatory failure is not clear, but there are various possibilities. Firstly, it may simply be that the respiratory muscles are rested so that they can respond better to the demands of the respiratory centre during the day. Secondly, it may be that improving blood gases at night and preventing the marked REM sleep deteriorations leads to resetting of the respiratory centre back towards normal, thus reversing an acquired blunting of drive. Tricyclic antidepressants such as protriptyline can virtually abolish REM sleep periods and their associated hypoxaemia and have been shown to improve daytime blood gases temporarily, suggesting that simply abolishing these periods of particular hypoxia can help. Thirdly, by increasing chest-wall and lung excursion (tidal volumes on the ventilator can be in excess of the voluntary vital capacity) overall respiratory compliance may improve, allowing the muscles to work more efficiently. Whatever the explanation, there is no doubt that this is a life-saving therapy that in certain conditions can add decades of active life.

Most of the original techniques to support ventilation overnight evolved from positive-pressure ventilation developed during the polio epidemics in the early 1950s, and the subsequent use of the iron lung that was developed to support these poliomyelitis victims long term. Evacuating the air from around the chest expands the lungs, recreating the normal way of breathing. A range of devices involving airtight jackets and shells over the chest were developed, but required much attention to detail and often individual, tailor-made systems. Efficacy was also limited by an unfortunate specific complication that resulted from the abolition of spontaneous ventilatory drive to the diaphragm and pharyngeal muscles, namely upper airway collapse during the mechanical inspiratory phase due to an unbraced airway. The recent development of comfortable nasal and face masks has revolutionized the overnight ventilation of these patients. Positive pressure ventilation can be used via a face mask, or sometimes via the nasal masks used for NCPAP (see Fig. 18.5.2.8). Although there are still many problems to be overcome when establishing patients on such equipment (particularly mask comfort and air leaks through the mouth when using nasal masks), the systems can be bought off the shelf ready to use (current cost c.£1500). Most units now use positive-pressure ventilation in preference to negative-pressure systems.

Electrical pacing of the diaphragm is occasionally used for supporting ventilation in conditions where the phrenic nerve and diaphragm are intact and the problem is central. This involves the implantation of bilateral phrenic electrodes and induction coils under the skin that are activated by external induction coils.

Further reading

Gastaut H, Tassinari CA, Duron B (1966). Polygraphic study of the episodic diurnal and nocturnal (hypnic and respiratory) manifestations of the Pickwick syndrome. Brain Res, 2, 167–86.Find this resource:

Guilleminault C, Stoohs R, Duncan S (1991). Snoring (1). Daytime sleepiness in regular heavy snorers. Chest, 99, 40–8.Find this resource:

Jenkinson C, et al. (1999). Randomised prospective parallel trial of therapeutic nasal continuous positive airway pressure (NCPAP) against sub-therapeutic NCPAP for obstructive sleep apnoea. Lancet, 353, 2100–5.Find this resource:

Remmers JE, et al. (1978). Pathogenesis of upper airway occlusion during sleep. J Appl Physiol, 44, 931–8.Find this resource:

Robinson G, Davies Rjo, Stradling Jr (2004). Obstructive sleep apnoea and hypertension. Thorax, 59, 1089–94.Find this resource:

Stradling JR, Davies RJO (2004). Obstructive sleep apnoea, definitions, epidemiology and natural history. Thorax, 59, 73–78.Find this resource:

Sullivan CE, et al. (1981). Reversal of obstructive sleep apnoea by continuous positive airway pressure applied through the nares. Lancet, i, 862–5.Find this resource:

Weiss JW, et al. (1996). Hemodynamic consequences of obstructive sleep apnea. Sleep, 19, 388–97.Find this resource: