Sleep disorders in children

Parasomnias are undesirable motor, autonomic or experiential phenomena that occur exclusively or predominantly during sleep.9 Parasomnias are common in childhood — examples include bruxism (teeth grinding, 6%–10%), sleep terrors (0.7%–2%) and somnambulism (sleep walking, up to 7%).10 A simplified summary of parasomnias with their prevalence rates is provided in Box 2.8 Benign parasomnias may run in families, increase in frequency with any condition that causes sleep deprivation or sleep fragmentation such as fever, and tend to improve with age (Box 2).8

Extremely common sleep problems in children include a child not getting into bed, having difficulty or requiring undue help to settle to sleep, frequent waking in the night and/or getting out of bed, and very early morning awakenings. They are often grouped as behavioural sleep disorders because of the perception that the problem lies with how the child behaves. These problems may lead to insufficient sleep and considerable family disruption. Children with developmental disorders, attention deficit hyperactivity disorder, depression and anxiety have higher incidence of these types of sleep disturbances than other children.11

Key to reducing the frequency and severity of behavioural sleep disorders is the provision to parents of preventive information, best provided opportunistically in primary care and by maternal child health nurses. Treatment interventions should then be evidence-based and developmentally appropriate. Parasomnias are usually benign and most decrease in frequency in later childhood. Education and reassurance of parents may be all that is required in less severe cases. Behavioural strategies for management of parasomnias include anxiety-relaxation techniques for poor sleep initiation, and sleep hygiene measures.11 These include limit-setting — for example, gradually removing parents’ attendance at the child’s bedside, so they are not present at the time of sleep onset — and moving bedtime closer to the usual time of sleep onset, to avoid periods of lying in bed awake before sleep onset. These measures help to eliminate the need for parents to attend to the child at each night-time awakening, and encourage a pattern of prompt sleep onset after going to bed. Together, they avoid prolonged periods of wakefulness during the night.

The core principle of preventing and managing bedtime (settling) issues and frequent night waking is to promote independence in settling to sleep. Infants and children who depend on a parent or other sleep association (music, dummy, rocking) at the start of the night are likely to require the same attention to resume sleep after what are otherwise normal, brief awakenings during the night. Consistency is the most important factor, but the rate of possible change is family-specific and sometimes needs to occur in slow, small steps to be sustained. Maternal mental health is an important factor in managing paediatric sleep disorders; children’s sleep problems and poor maternal sleep can contribute to mental health disorders, as well as being an aetiological factor for the inconsistent maintenance of the infant’s sleep routines. In a small group of toddlers with difficulty initiating or maintaining sleep, melatonin could be used to entrain their sleep routine. The interventions are also safe, with no negative long-term outcomes and many benefits to child and family health and functioning.12

Parasomnias can occur very frequently, cause distress and/or disrupt family life. Management strategies should ensure the safety of the child; for example, by placing the mattress on the floor rather than on a bed frame, and by adding locks to doors to prevent the child opening simple latches while sleep walking. Simple strategies to minimise the frequency of events are often effective for managing parasomnias in otherwise normal children and include:

Extending sleep: insufficient sleep increases the frequency of parasomnias. As little as 30 minutes of additional sleep can reduce the frequency of parasomnias. Work towards earlier bedtime and/or later rise times. Making bedtime earlier should occur in small steps of 10–15 minutes, to avoid increasing bedtime struggles.

Reducing bedtime anxiety and struggles/conflict: going to bed in an aroused state (anxious, angry or upset) can intensify parasomnias. Aim for a gentle and predictable bedtime routine. Avoid stimulating activities like television or computer games for an hour before bed. If necessary, match bedtime to the usual sleep onset time (even if this is late), then slowly bring bedtime earlier, as above. Medication is rarely indicated.11,13 If the problem is very severe, very frequent or atypical, raising the possibility of a seizure disorder, then referral to a sleep specialist is indicated, with polysomnography and/or electroencephalography indicated depending on the clinical scenario.

Investigating excessive daytime sleepiness and circadian rhythm disorders: excessive sleepiness requires systematic evaluation. Possible causes include inadequate sleep, sleep disruption from conditions such as restless leg syndrome and obstructive sleep apnoea (OSA), and circadian rhythm disorders. In children with an apparently sufficient duration of sleep, marked daytime sleepiness may be the only manifestation of narcolepsy, which has an estimated prevalence of 1 in 4000 to 1 in 2000, and a peak of onset at 14 years of age.14 Recognition of narcolepsy onset in childhood and appropriate treatment is likely to improve learning and daytime functioning.

Disruption to normal circadian rhythmicity, such as very late bed and rise times, can have substantial effects on the ability of a child to participate in school and other activities. Circadian sleep problems are especially common in children with pervasive developmental disorders such as autism spectrum disorder, and also occur in adolescents, where many factors impact on a tendency for the sleep phase to be delayed into the night, making socially imperative morning rise times difficult to achieve. The main focus of therapy is to establish and maintain good sleep hygiene including settling strategies (eg, avoiding screen time and caffeine-containing drinks before bedtime) and consistent timing of sleep throughout the 7-day week. Specialist referral is advised if there is concern about accuracy of diagnosis, or need for additional medical therapy including use of medications such as melatonin. Use of such medications may be indicated but must be in the context of awareness of the high need for ongoing surveillance of short- and long-term side effects.

Snoring and OSA are common, affecting 3%–15% of children, with peak prevalence in the preschool years when lymphoid tissue size in the upper airway is largest relative to the size of the facial skeleton.15 OSA affects up to 5.7% of children,16 and so potentially affects one child in every classroom in the country. Although the highest incidence of OSA is in preschoolers (3–5 years of age) with large tonsils, 9% prevalence of snoring has been documented in infants aged 0–3 months.15,17

Identification of severe OSA is important because it is linked to increased risk for postoperative respiratory compromise, including emergency reintubation and unplanned admissions to intensive care. It is a major challenge to identify the children who require perioperative management in tertiary paediatric centres. Box 3 highlights cases where referral for polysomnography is warranted, rather than direct referral for adenotonsillectomy.

OSA is associated with sleep fragmentation and repeated episodes of hypoxia. Polysomnography is superior to other testing methods for determining disease severity and also permits diagnosis of comorbid disorders (eg, periodic limb movements). The thresholds for severity of OSA are lower than in adults, with OSA defined as ≥ 1 obstructive event per hour of sleep on polysomnography. Treatment is generally recommended if the frequency of obstructive respiratory events is > 1.5 per hour. Severity is usually defined as mild for 1–5 events per hour, moderate for 5–10 events per hour and severe for ≥ 10 events per hour. However, no threshold has been established for disease severity with regard to the development of complications. Even mild disease is associated with adverse neurocognitive, behavioural and cardiovascular outcomes, such that even chronic partial obstruction causing snoring without gas exchange abnormalities or evident sleep disruption is associated with adverse effects.

Despite the fact that no clinical assessment method other than polysomnography has proven discriminatory for OSA in children who snore, the number of paediatric sleep units in Australia is inadequate to provide polysomnography to screen all snoring children. The presence of snoring and large tonsils is a sensitive but not specific marker. Helpful clinical indicators include increased work of breathing, parental concern, and frequent daytime mouth breathing.18 Markers that are specific but not sensitive (helpful when positive, but unable to rule out disease) include excessive daytime somnolence and observed OSA.19 Almost all screening tools are also specific but not sensitive, including overnight oximetry (most useful if positive, but most children have a negative study that does not rule out OSA20), video recordings and nap studies, so the search for an ideal screening tool continues. Overnight oximetry is helpful in identifying cases with marked hypoxia, but those using it need to be familiar with the technical aspects and diagnostic limitations of the tool.21 All screening tools, including oximetry, are best used in combination with clinical indicators such as young age (under 3 years) and comorbidities (syndromes, obesity, etc), to help evaluate the likelihood of postoperative respiratory complications.21

Among the major sequelae of untreated OSA, cardiovascular risks include systemic hypertension, increased sympathetic activation and ventricular hypertrophy, while pulmonary hypertension and right heart failure still occur occasionally in infants and children with severe OSA.22 Even mild OSA is linked to daytime neurocognitive dysfunction that translates into decrements of intelligence quotient, and a randomised controlled study has now been published regarding assessment of neuropsychological development in school-age children with OSA after tonsillectomy.23 Plausible mechanisms for this association include sleep fragmentation, repetitive hypoxia, and reduced cerebral blood flow and oxygenation. Behavioural improvements follow adenotonsillectomy,24-26 but responses in neurocognitive function are variable.27 A review of 25 studies investigating behavioural and neurocognitive outcomes following adenotonsillectomy found that all studies reported improvement in one or more measures including quality of life, behavioural problems including hyperactivity and aggression, and neurocognitive skills including memory, attention and school performance.26 Improvement or resolution of OSA has also been linked to concomitant improvements in systemic and pulmonary blood pressures, heart rate and pulse variability, cardiac morphology and cardiac function.22

The natural history of symptoms of OSA (eg, snoring, mouth breathing and apnoea) is for around 50% of preschool children to move (bidirectionally) among severity groups over a 2-year follow-up period.28 In a cohort of 12 447 children studied across seven time points between ages 6 months and 6.75 years, the prevalence of OSA symptoms was highest between 3.5 and 4.8 years of age.29 The highest peak of new symptoms occurred between the ages of 1.5 and 2.5 years.29 Another study undertook polysomnography on 45 children with mild OSA at baseline; at follow-up 4 years later, disease had worsened in 37% and resolved in 26%.30

Preschool children generally respond to adenotonsillectomy; meta-analysis shows cure rates of 82% in otherwise normal children.31 Success rates for adenotonsillectomy are lower in obese32 and older33 children, and adenoidectomy and/or tonsillectomy is usually not appropriate for infants. Although adenotonsillectomy reduces the severity of OSA in obese children, such children have more severe initial disease, and obesity increases the risk for persisting disease.32,34 Nasal corticosteroid sprays35-37and leukotriene-receptor antagonists (eg, monteleukast)38 are helpful in children with mild OSA and for some with persistent disease after adenotonsillectomy, and a treatment trial is appropriate before pursuing other interventions.39 Specific airway problems, especially infants with Pierre Robin sequence, may respond to mandibular distraction, continuous positive airway pressure, nasopharyngeal tube, and/or oral tongue positioning devices, but may necessitate tracheostomy.

Factors that indicate a higher risk for persisting OSA despite adenotonsillectomy include more severe initial disease (respiratory disturbance index > 10/h or minimum SaO2 < 80%), obesity with body mass index > 95th percentile for age and sex, and children aged > 7 years at the time of surgery (whether obese or non-obese).33 There is interplay between obesity and atopy, in that for non-obese children, comorbid asthma increases the risk of persisting disease whereas allergic rhinitis is only significant when both obese and non-obese groups are considered together.40 These groups need follow-up after surgery to establish whether snoring has or has not resolved.

Older children and adolescents may respond to adenotonsillectomy or require other treatments including continuous positive airway pressure, orthodontic and other surgical or dental procedures (rapid maxillary expansion, or mid-face advancement). Evidence of efficacy and safety in children is limited for orthodontic options such as mandibular advancement splints.41-44 These interventions aim to affect growth of the face and oropharyngeal airway to produce long-term structural changes, irrespective of whether the initial airway problem is primary, or secondary to OSA.

Underlying medical disorders work to both increase risk for OSA and to reduce the effectiveness of surgical treatment (Box 3). In particular, congenital abnormalities that affect craniofacial or thoracic growth, such as achondroplasia and Down syndrome, will predispose to sleep-disordered breathing. In Down syndrome, there appears to be particular risk for hypertrophy of the lingual tonsils.33 It is also known that children with multiple disabilities have increased risk for other sleep disturbances such as difficulties with sleep initiation and maintenance, insomnia and other sleep pattern abnormalities.45

Children with neuromuscular diseases have increased incidence of OSA in the first decade.46 Congenital cardiothoracic abnormalities or restrictive lung disorders, often linked to neuromuscular disorders or neurodevelopmental disability such as cerebral palsy, also predispose to nocturnal respiratory failure. Symptoms suggestive of nocturnal hypoventilation include increased frequency or severity of lower respiratory tract infections, and progression of scoliosis. Screening should include pulmonary function testing, with sleep studies for children with vital capacity < 60% of that predicted and for non-ambulant children before scoliosis surgery, and pragmatic consideration of screening versus full polysomnographic studies.47 Early identification and treatment of impaired pulmonary function can prevent or reduce the frequency and duration of admissions to intensive care units, as well as improving quality and duration of life.48

Congenital central hypoventilation syndrome is a rare but highly treatable condition (incidence, 1 in 50 000 live births).49 This usually presents during the neonatal period with frequent apnoeas or colour change during sleep, but milder forms can present in older children.50