Interpretation of sleep studies and perioperative considerations in children with sleep-disordered breathing

Learning objectives.

By reading this article you should be able to:

• •Recognise comorbidities associated with sleep-disordered breathing (SDB).
• •Interpret the results of sleep investigations in paediatric patients.
• •Explain the use of risk predictor tools to estimate perioperative risk in SDB.

Key points.

• •Sleep-disordered breathing (SDB) is common in children; the incidence is increased in those with genetic conditions, specifically with craniofacial features or neuromuscular conditions.
• •Sleep studies are recommended to diagnose hypoventilation in children with regular snoring, features associated with SDB and those with restrictive lung disease.
• •Polysomnography is the gold standard sleep study used to assess SDB.
• •Children with SDB are sensitive to opioids and have an increased risk of perioperative respiratory complications.
• •The main risk factors for perioperative complications are age <3 yrs, Spo₂ nadir <80%, peak Pco₂ >60 mmHg and failure to thrive.

Sleep-disordered breathing (SDB) refers to a group of chronic conditions that cause intermittent partial or complete cessation of breathing during sleep. Sleep-disordered breathing is common, affecting up to 12% of children.¹ Sleep-disordered breathing can be subdivided into obstructive sleep-disordered breathing (OSDB) and central sleep-disordered breathing (CSDB) disorders, with or without associated hypoventilation. Obstructive sleep-disordered breathing exists as a clinical spectrum varying from partial airway obstruction seen in snoring and upper airway resistance syndrome (UARS) to complete intermittent airway obstruction seen in obstructive sleep apnoea (OSA).²

Primary snoring is not associated with apnoeas, hypopnoeas or sleep fragmentation. Upper airway resistance syndrome is characterised by increasingly negative intrathoracic pressure during inspiration leading to arousals and sleep fragmentation without associated apnoeas, hypopnoeas or oxygen desaturations. Obstructive sleep apnoea is defined as ‘periodic episodes of nocturnal airflow restriction (hypopneas) or obstruction (apnoeas) in association with sleep disruption, arousals from sleep, oxygen desaturation and possible hypercapnia’.³

Aetiology

Obstructive sleep-disordered breathing is either simple and associated with hypertrophy of the upper airway lymphoid tissues (adenoids, tonsils, or both) in an otherwise normal child, or complex, in association with comorbidities predisposing the upper airway to obstruction such as craniofacial syndromes, genetic syndromes, obesity or neuromuscular conditions. Central sleep-disordered breathing with or without hypoventilation results from lack of output from the brainstem or restrictive lung disease secondary to neuromuscular disorders or thoracic deformities. The mechanism that leads to hypertrophied tonsillar tissue is still not fully understood, but is unlikely to be the sole cause of OSA because children do not have airway obstruction when awake. Obstruction is likely to be secondary to a combination of structural and neuromotor abnormalities.

Common features resulting in airway narrowing include micrognathia, macroglossia and mid-facial hypoplasia. These are found in genetic conditions such as Robin sequence, trisomy 21, mucopolysaccharidoses and craniosynostoses (e.g. Crouzon and Apert syndromes). This explains the high prevalence of OSDB in these conditions. It is also common in these conditions to find OSDB associated with multilevel airway obstruction and changes in tone of the upper airway. Children with neuromuscular disorders often develop nocturnal hypoventilation early in life as a result of progressive muscle weakness, which will eventually progress to diurnal hypoventilation.²

With the increasing prevalence of obesity in children, a growing number of these are being diagnosed with OSA. The NANOS study showed 47% of children with obesity to have an apnoea–hypopnoea index (AHI) >1, indicating a degree of OSA. The most important risk factor for OSA was the degree of adenotonsillar hypertrophy, but the simultaneous presence of obesity and adenotonsillar hypertrophy contributed synergistically to the development of OSA.⁴

Clinical findings

Sleep-disordered breathing presents differently in children compared with adults. Presentation is typically with snoring, laboured breathing, gasping, sweating, neck hyperextension and apnoeas. Clinical arousal during sleep and daytime somnolence are less common but difficulty in concentration, poor attention span, learning difficulties, nocturnal enuresis and hyperactivity are reported.⁵ On clinical examination, it is common to find tonsillar hypertrophy, adenoidal facies, mandibular hypoplasia, or other facial features consistent with the syndromes mentioned above.

In adults, OSA is associated with cardiovascular complications including hypertension, coronary artery disease, arrhythmia, cardiac failure and stroke. Cardiovascular sequalae in children are less well described but there is increasing evidence that changes occur even with mild OSA. Several cardiovascular abnormalities are associated with OSA in children that may have harmful effects in the long term. These are summarised by the American Heart Association scientific statement published in 2021.⁶ There is altered heart rate variability after obstructive events and this is likely a result of autonomic system disturbance. It is possible that heart rate variability could be used as a marker of cardiovascular complications of OSA in the future.⁷ The evidence for childhood hypertension secondary to OSA is contradictory. A meta-analysis from 2008 showed hypertension to be associated with OSA but Quante and colleagues did not demonstrate a link in a more recent study of >450 children.⁷ However, this study excluded children with severe OSA.^7,8 Nocturnal hypertension has been demonstrated in children with moderate-to-severe OSA, and childhood OSA is an independent risk factor for hypertension in adulthood.^7,9,10 Left ventricular (LV) hypertrophy may be present in children with OSA, and LV mass index increases with severity of OSA.⁹ Pulmonary hypertension from recurrent hypoxia, hypercarbia and respiratory acidosis may result in right ventricular (RV) hypertrophy and reduced RV systolic function, particularly in the presence of significant comorbidity.⁶ Historically, cor pulmonale was not an uncommon presentation of significant OSA.

Diagnosis

A diagnosis of SDB can be hinted from clinical history and examination. Interestingly, most studies of children with SDB have shown only a modest positive correlation between adenotonsillar size and severity of OSA.¹¹ If suspected, diagnosis should be confirmed with a sleep study. The American Academy of Pediatrics (AAP) recommends a sleep study in children with regular snoring and features typically associated with paediatric OSA.³ The European Respiratory Society taskforce, in agreement with the American Academy of Otolaryngology, Head and Neck surgery foundation (AAO-HNSF) recommend obese children and those with craniofacial abnormalities and neuromuscular disorders should have priority access to cardiorespiratory studies or polysomnography (PSG) before adenotonsillectomy. These children should have a repeat PSG after surgery because of the high-risk of persistent OSA.⁵

It is not feasible to investigate all children for cardiovascular complications of OSA because of the high prevalence of OSA, adenotonsillectomy and limited resources. Before anaesthesia, in addition to routine baseline non-invasive arterial pressure, it is worth considering a 12-lead ECG and echocardiogram in high-risk children with severe OSA with comorbidities such as Down syndrome, neuromuscular disorders (e.g. Duchenne muscular dystrophy), mucopolysaccharidoses and significant obesity. Preoperative knowledge of abnormalities will improve perioperative safety and outcomes.

Sleep studies

Polysomnography

Indications

Polysomnography is the gold standard sleep study used to assess SDB. It can be used in the event of an inconclusive overnight oximetry study to clarify the diagnosis.

Technique

Polysomnography is a supervised laboratory study with the following channels being used:

• (i)respiratory channel: nasal airflow, thoracoabdominal movements, oximetry and end-tidal or transcutaneous CO₂ monitoring
• (ii)cardiac channel: ECG and pulse oximetry
• (iii)body position channel
• (iv)snore microphone
• (v)video

It also includes a neurology channel (electroencephalogram [EEG], electrooculogram, chin and leg electromyography) to identify five sleep stages, which comprise: wake, Stage 1 (N1), Stage 2 (N2), Stage 3 (N3) and rapid eye movement (REM) sleep. With appropriately trained staff and equipment PSG can be performed in infants and older children. The relevance of adding the neurology channel is to have a better idea of sleep architecture, sleep fragmentation and impact of SDB on cortical arousals and diagnosis of alternative sleep disorders (parasomnia, periodic limb movement disorders).

Interpretation

There are three important differences to highlight when comparing polysomnographic studies in adults and children. Firstly, children have a higher arousal threshold than adults, with younger children having the highest threshold.¹² They are less likely to develop fragmented sleep as moderate hypoxaemia is less likely to lead to sleep arousal.¹³ This leads to preservation of sleep architecture, explaining the rarity of daytime somnolence in children. It is possible that small alterations in sleep architecture responsible for behavioural changes are not detected by PSG. Secondly, most obstructive events occur during late REM sleep. If PSG does not pick up sufficient periods of late REM sleep (e.g. nap studies), obstructive events could be missed. Thirdly, constant snoring and increased work of breathing may result in a PSG showing persistent obstructive hypoventilation rather than episodes of obstructive apnoeas.¹

Scoring requires identification of respiratory events as per American Academy of Sleep Medicine (AASM) guidelines and definitions¹⁴:

• •Apnoea is a cessation of airflow >90% of the baseline for >2 breaths or >10 s
• •Hypopnoea is a flow reduction by ≥30% or >2 breaths or >10 s with either a ≥3% oxygen desaturation or an arousal on EEG

The AHI quantifies the severity of SBD. It consists of the number of apnoeas and hypopnoeas per hour of total sleep time and can be subdivided into a central and obstructive index. An obstructive index of <1 is normal, 1–5 mild OSA, 5–10 moderate OSA and >10 severe OSA.¹⁴

Central apnoeas are not uncommon in infants and children, particularly during REM sleep with apnoeas <20 s with transient desaturations occurring in healthy children. Central apnoeas become more clinically relevant as time goes by when they occur frequently in association with gas exchange abnormalities. Obstructive apnoeas are rarer events in healthy children.¹³

Limitations

Paediatric PSG may not be readily available and are resource heavy, requiring specialised laboratories and trained personnel to monitor and score tests. It should be considered that sleep studies represent a single night of sleep and sleep quality can be affected by multiple factors including an unfamiliar environment and the need to remain attached to monitoring. In addition, sporadic events such as nocturnal seizures, may not be identified during the study period.

Cardiorespiratory sleep studies

Indications

Cardiorespiratory sleep studies (CRSS) have a simplified set-up process and scoring compared with PSG, whilst remaining an accurate tool for the detection of SDB.² Like PSG, they can be used after an inconclusive pulse oximetry results to confirm SDB.

Technique

Cardiorespiratory sleep studies are mostly carried out in a sleep laboratory, fully supervised by a technician monitoring audiovisual recording.

A limited number of channels are assessed:

• (i)respiratory channel: nasal airflow, thoracoabdominal movements, oximetry and end-tidal or transcutaneous CO₂ monitoring
• (ii)cardiac channel: ECG and pulse oximetry
• (iii)body position channel
• (iv)snore microphone
• (v)video

Interpretation

The scoring uses adapted rules following the AASM guidelines as described in the previous section.¹⁴ Figure 1 is a CRSS demonstrating a prolonged episode of obstructive apnoea. Snoring is associated with intermittent cessation of airflow in the nasal pressure airflow transducer while paradoxical thoraco-abdominal movements continue to take place. This leads to intermittent desaturations that recover with breakthrough breathing.

Fig 1.

Cardiorespiratory study demonstrating obstructive sleep apnoea.

Figure 2 demonstrates an obstructive hypopnoea with a characteristic reduction of flow >30% also associated with desaturations and continued paradoxical thoraco-abdominal movements.

Fig 2.

Cardiorespiratory study demonstrating obstructive hypopnoea.

In contrast with Fig 1, Fig 2, Figure 3 shows multiple central apnoeas. There is an overall lack of activity in the snore channel, associated with intermittent cessation of airflow and desaturations. Of note, there are no paradoxical thoraco-abdominal movements.

Fig 3.

Cardiorespiratory study demonstrating recurrent central apnoeas.

Limitations

The reduced number of channels limits the information on sleep architecture, potentially missing episodes of cortical arousal after obstructive episodes but not episodes of desaturation.

Ambulatory sleep studies

Cardiorespiratory sleep studies may be carried out in the child's home (with the exception of CO₂ monitoring in many instances) for ease and comfort of the child and to reduce the costs of a modern sleep laboratory. Evidence suggests that the results are more representative of the child's ‘normal’ night's sleep because they are sleeping in a familiar environment. Further work is necessary to optimise the test to diagnose children with mild OSA.¹⁵

Overnight pulse oximetry

Indications

Although PSG is the gold standard for the diagnosis of OSA in paediatrics, it is not readily available in all centres and has some limitations. Overnight pulse oximetry can be used as a tool to prioritise access to surgery when PSG is unavailable or in a low-resource setting. Pulse oximetry can also be used to titrate and wean the oxygen requirements of children on home oxygen.¹⁶

Technique

The unobserved study can be carried out with a pulse oximeter at home or in hospital. The oximeter has a short averaging time of 2–3 s to detect rapid changes in oxygen saturations. Overnight oximetry is used to assess children with suspected OSDB.

Interpretation

A pulse oximetry study records the number of desaturations below different thresholds, the average saturations and the lowest saturation recorded, or nadir. The McGill score is used to classify the severity of OSDB.¹⁷ A normal or inconclusive study has fewer than three episodes of desaturations <90%. Mild disease has three or more episodes of desaturations <90% and three or fewer <85%. Moderate OSDB has three or more episodes of desaturations <90%, more than three <85% and three or fewer <80%. Severe disease has three or more episodes of desaturations <90%, more than three <85% and more than three <80%.

Nixon and colleagues evaluated the presence of clusters of desaturation to predict OSA.¹⁷ Using the McGill scoring, the presence of three clusters of desaturation events with at least three recorded values of Spo₂ <90% indicated moderate-to-severe OSDB. The McGill scoring correlated significantly with the AHI and predicted the risk of respiratory complications.¹⁷

Limitations

Care must be taken to interpret the results in correlation with clinical presentation. An inconclusive study is not enough to rule out OSA in the presence of significant symptoms and would require a follow up PSG.

Sleep questionnaires

Sleep questionnaires are a useful screening tool for OSA and can be useful in limited resource settings. Several different questionnaires are in clinical use. One of the most common is the paediatric sleep questionnaire (PSQ), which has been validated in children with OSA without comorbidity. The PSQ is less sensitive in patients with comorbidities such as neuromuscular disorders and trisomy 21 and cannot be used to replace cardiorespiratory sleep tests.¹⁸

Sleep clinical score

In the absence of access to PSG the sleep clinical score aims to provide a PSG-validated SDB screening tool. Villa and colleagues developed such a tool that combines clinical history including behavioural and cognitive problems, physical examination and subjective symptoms.¹⁹ A total of 279 children with primary snoring and OSA were studied and a sleep clinical score was determined. A score of >6.5 was shown to be consistent with a diagnosis of OSA, with a sensitivity of 96% and a specificity of 67%.¹⁹

Perioperative risk

Children with SDB have increased perioperative risk, particularly for perioperative respiratory adverse events (PRAE). These include airway obstruction potentially leading to laryngospasm; oxygen desaturation; oversedation because of increased sensitivity to opioids requiring escalation in care; and negative-pressure pulmonary oedema.^20,21 In addition, there is an increased risk of stroke in children with OSA and sickle cell anaemia.²² It is therefore important to determine clinical criteria to decide which children are not suitable for day surgery and require postoperative care as an inpatient.

Various guidelines are available to identify children with SDB who are at increased risk of perioperative respiratory complications. Many of these guidelines are aimed at children undergoing adenotonsillectomy for SDB and there are limited data on risk stratification of children with SDB undergoing other surgical procedures.

Pehora and colleagues recently evaluated prediction of PRAE in children with SDB undergoing a variety of surgeries.²³ They concluded that preoperative non-invasive ventilation or need for home oxygen, severe OSA, prematurity and use of an tracheal tube for airway management were significant risk factors. They demonstrated an almost 10-fold increase in the incidence of PRAE in children undergoing surgery requiring elective tracheal intubation compared with surgical or sedation procedures not requiring intubation.²³ Specific to ear, nose and throat (ENT) surgery, a secondary analysis of the APRICOT study demonstrated a relative risk of 1.51 (1.28–1.77) for PRAE in children undergoing ENT surgery compared with non-ENT procedures.²⁴

Clinicians typically have concern for children at the severe end of the SDB spectrum with potential cardiovascular repercussions of OSA and hypoventilation, particularly those undergoing major surgery such as cardiac or spinal procedures. It has been shown that children with neuromuscular disease undergoing scoliosis surgery are likely to benefit from preoperative implementation of non-invasive ventilation to reduce their risk of PRAE.²⁵

Perioperative risk in children undergoing adenotonsillectomy for SDB

In 2014, the Children's Hospital of Philadelphia carried out a prospective observational cohort study with 329 children with SDB undergoing adenotonsillectomy and developed a model to predict those at risk of PRAE. Higher risk children were included in this study (high BMI, ex-premature and the presence of comorbidities such as chronic lung disease, neuromuscular disease and craniofacial anomalies). The main risk factors identified were age <3 yrs, Spo₂ nadir <80% and peak Pe′co₂ >60 mmHg. Although this study used peak Pe′co₂, most sleep laboratories use transcutaneous CO₂ as this gives a better overall trend overnight but with the possibility to use Pe′co₂ for correlation purposes. Failure to thrive, defined as height or weight below the fifth centile, was also identified as an independent risk factor but obesity was not. However, a predictive model based on established risk factors could not be determined and physicians' clinical assessment based on other risk factors and social situations remains important.²⁶

More recently, Katz and colleagues carried out a retrospective study of 374 children at a tertiary centre to identify clinical and polysomnographic predictors of PRAE in children having adenotonsillectomy.²⁷ It evaluated the relationship between PRAE, comorbidities and polysomnographic results and determined that cardiac comorbidities, airway anomalies and age <3 yrs were independent predictors of PRAE. They suggest that PSG results are less important than age and complex comorbidity in predicting PRAE.²⁷

A significant number of children with OSA do not have access to PSG before undergoing adenotonsillectomy. The snoring, trouble breathing, unrefreshed (STBUR) tool by Tait and colleagues is a simple tool that can be used to predict the risk of PRAE.²⁰ This questionnaire comprises five screening questions (Box 1). The likelihood of perioperative respiratory complications was increased three-fold in the presence of three screening symptoms, and 10-fold when all symptoms were present.²⁰ The same group subsequently demonstrated that children who had three or more of the listed symptoms also had an increased risk of opiate-associated oxygen desaturations, which is comparable with PSG confirming a diagnosis of OSA.²¹

Box 1. STBUR questionnaire.20.

Does your child:

• •Snore more than half the time?
• •Snore loudly?
• •have Trouble breathing or struggle to breathe?
• •ever stop Breathing during the night?
• •wake up feeling Un-Refreshed?

Alt-text: Box 1

Summary

Sleep-disordered breathing is common in children and presents differently than in adults. Children with comorbidities including craniofacial abnormalities, neuromuscular weakness and obesity are at higher risk of having SDB. Laboratory or increasingly home sleep studies are ideal for identifying those at high risk of perioperative complications, but resources are limited. The diagnostic variables differ from adults. Alternative screening tools may be implemented in the absence of a sleep study.

Declaration of interests

The authors declare that they have no conflicts of interest.

MCQs

The associated MCQs (to support CME/CPD activity) will be accessible at www.bjaed.org/cme/home by subscribers to BJA Education.

Biographies

Jose Filipe Lopes Vieira BSc MRCPCH FRCA is a specialty registrar in anaesthesia with a special interest in paediatric anaesthesia. He initially trained as a paediatrician and completed his MRCPH before changing specialty training to anaesthesia and completed an advanced paediatric fellowship year at Great Ormond Street Hospital.

Alice Miskovic BSc FRCA is a consultant paediatric anaesthetist at Great Ormond Street Hospital with a special interest in sleep-disordered breathing, particularly in pregnancy with relevance to fetal outcome. She has subspecialised in anaesthesia for paediatric ENT surgery.

Francois Abel MD FRCPH is a consultant paediatrician at Great Ormond Street Hospital where he is the medical lead for the Paediatric Respiratory Sleep Service. He has a postgraduate certificate in paediatric sleep sciences from the University of Western Australia.

Matrix codes: 1A01, 2D02, 3D00