ERS statement on obstructive sleep disordered breathing in 1- to 23-month-old children

Evidence summary

a) Snoring or “noisy” breathing is the commonest symptom directly related to OSAS during the first 2 years of life followed in frequency by apnoeas, frequent movements during sleep, mouth breathing and recurrent awakenings.

b) History of apparent life-threatening events (ALTEs; or brief, resolved, unexplained events (BRUE)) may be associated with the presence or development of OSAS.

c) There is no high-quality evidence indicating that infants with regurgitation or a history of prematurity are predisposed to OSAS.

Literature review

a) The prevalence of snoring ≥3 days per week in young children recruited from the community was 9% in 0- to 3-month-old infants, 15% in 1-year-old infants and 5.3% in children aged 2 weeks to 2 years based on parental report [17–19]. In hospital-referred children aged <18 months with adenotonsillar hypertrophy and OSAS, snoring was the most common symptom followed by reported apnoeas, frequent movements during sleep, mouth breathing, recurrent awakenings, developmental delay and recurrent respiratory infections [20, 21]. In a longitudinal, questionnaire study, the prevalence of snoring peaked between age 1.5 and 2.5 years [22].

b) History of ALTEs has been associated with the presence or later development of OSAS and mild facial dysmorphia in studies of low methodological quality [23–28]. In contrast, a large retrospective study including infants with reported episodes of apnoea and/or cyanosis demonstrated normal polysomnographic findings in the majority of participants [29].

c) No consistent association has been found between gastro-oesophageal reflux and OSAS; prematurity as a risk factor for SDB in infancy has not been studied adequately, although there is evidence for its role in older children [3, 30–33].

Evidence summary

Growth failure is a frequent initial clinical manifestation and also a complication of OSAS in young children.

Literature review

Growth failure is a complication of OSAS as demonstrated in a meta-analysis [34]. It may also be the clinical manifestation leading to the diagnosis of OSAS as other symptoms can be subtle in early life. In a retrospective study of hospital-referred children with adenotonsillar hypertrophy, the body weight percentile decreased between birth and the diagnosis of OSAS [20].

Evidence summary

a) Adenoidal or, less frequently, tonsillar hypertrophy (after the first 6 months of life).

b) Nasal obstruction: acute upper viral respiratory infection, choanal atresia, nasal pyriform aperture stenosis.

c) Laryngomalacia manifested as inspiratory stridor without or with intercostal retractions, episodes of cyanosis, feeding difficulties and growth delay.

d) Syndromic craniosynostosis with midface hypoplasia (Apert syndrome, Crouzon syndrome and Pfeiffer syndrome) or without midface hypoplasia (Muenke syndrome, Saethre–Chotzen syndrome and complex craniosynostosis); unrepaired or repaired isolated cleft lip or palate.

e) Marked mandibular hypoplasia as in non-syndromic and syndromic Pierre Robin sequence (e.g. Treacher Collins syndrome and Stickler syndrome).

f) Neuromuscular disorders (cerebral palsy, mitochondrial disorders and spinal muscular atrophy).

g) Complex abnormalities (achondroplasia, Beckwith–Wiedemann syndrome, Chiari malformation, Down syndrome, mucopolysaccharidoses and Prader–Willi syndrome).

Literature review

Various congenital or acquired conditions affecting structures from the nose down to the larynx can predispose to OSAS [35].

a) In a retrospective study of infants with OSAS, adenotonsillar hypertrophy was diagnosed as early as 5 months of age [20].

b) Nasal congestion caused by an intercurrent viral respiratory infection may be accompanied by obstructive events [36]. OSAS is a potential complication of unilateral choanal atresia and one of the primary features of bilateral choanal atresia [37].

c) The coexistence of laryngomalacia and OSAS is frequently overlooked [33, 38–41]. It resolves spontaneously in up to 80% of infants, but in the remaining cases intervention is required [42]. Airway lesions such as subglottic stenosis and tracheomalacia coexist with laryngomalacia in up to 50% of infants, and their frequency increases with increasing severity of upper airway obstruction; an association with gastro-oesophageal reflux disease has also been described [43].

d) OSAS is more prevalent (up to 68%) in craniosynostosis syndromes and more severe in children with midface hypoplasia (Apert syndrome, Crouzon syndrome and Pfeiffer syndrome) than in those without hypoplasia (e.g. Muenke syndrome or Saethre–Chotzen syndrome) [44]. Mandibular hypoplasia may coexist and contribute to upper airway obstruction. Several of these infants develop OSAS during an upper respiratory tract infection or due to adenotonsillar hypertrophy, but OSAS severity improves over the first 3 years of life [44, 45]. Infants with cleft lip and/or palate frequently have obstructive respiratory events during sleep [46].

e) Non-syndromic or syndromic Pierre Robin sequence is the constellation of micrognathia, glossoptosis and upper airway obstruction with or without cleft palate, resulting in feeding difficulties, stridor, episodes of cyanosis and OSAS [47, 48]. Severity of obstruction varies from mild to life-threatening requiring immediate intubation or tracheostomy (up to 13.4% of cases) [47, 49–51]. Coexisting unilateral choanal atresia, a small epiglottis allowing the tongue to obstruct the laryngeal opening, laryngomalacia or tracheal stenosis can worsen respiratory distress [52]. Airway obstruction and feeding difficulties may improve with growth over the first year of life [47, 49, 53].

f) Cerebral palsy and mitochondrial disorders may be accompanied by OSAS [54, 55]. Infants with spinal muscular atrophy type 1 and 2 have increased AHI [56, 57].

g) Infants with achondroplasia have higher AHI than control infants and midface hypoplasia contributes to obstructive events [58]. OSAS has been described in infants with Beckwith–Wiedemann syndrome [59]. Chiari malformation is accompanied by obstructive and central sleep apnoeas and hypoventilation, risk factors being the degree of brainstem crowding at the foramen magnum and/or length of herniation [60, 61]. In infants with Down syndrome, the lowest estimated prevalence of AHI ≥2 episodes·h^–1 is 30% [62]. OSAS, nocturnal hypoxaemia and hypoventilation are frequent findings [63–65]. OSAS is common in children with any of the mucopolysaccharidoses [66, 67]. Young patients with Prader–Willi syndrome have increased prevalence of OSAS and central sleep apnoeas (≥5 episodes·h^–1) [68, 69]. With increasing age, the frequency of OSAS increases and the frequency of central sleep apnoeas decreases [68, 69].

Evidence summary

a) Upper airway endoscopy by a flexible instrument helps to determine the level and severity of airway obstruction in cases of congenital stridor, craniofacial abnormalities or complex conditions (e.g. craniosynostosis, Down syndrome and Pierre Robin sequence).

b) Endoscopy can be performed without sedation, but complete examination of the lower airway requires general anaesthesia with the patient breathing spontaneously.

c) Upper airway imaging is useful in the evaluation of patients with craniofacial abnormalities.

Literature review

a+b) Few and small studies have evaluated the diagnostic value in otherwise healthy young children with suspected OSAS [70, 71]. Laryngomalacia is a frequent abnormality detected by endoscopy in children with Down syndrome and upper airway obstruction [72]. Moderate-to-severe upper airway obstruction on endoscopic evaluation of children with Pierre Robin sequence was predictive of polysomnography findings [52, 73]. Sher and colleagues have described four types of pharyngeal airway obstruction in patients with various craniofacial abnormalities, including craniosynostosis and Pierre Robin sequence [74, 75].

c) Upper airway size has been assessed in children with craniofacial abnormalities (nasal pyriform aperture stenosis and Pierre Robin sequence) using craniofacial computed tomography [76–79].

Evidence summary

Severe OSAS in young children may predispose to pulmonary hypertension.

Literature review

Right ventricular ejection fraction increases post-adenotonsillectomy in infants and children with oropharyngeal obstruction and OSAS symptoms [80]. Cor pulmonale has been described in case reports of infants with Pierre Robin sequence and severe upper airway obstruction [81, 82].

Evidence summary

a) Delayed growth is a complication of OSAS in young children.

b) Younger children with OSAS have higher frequency of growth failure or delay compared with older children.

Literature review

a) For delayed growth see also 1.2. OSAS may have long-term consequences on growth.

b) Infants aged <3 months who underwent supraglottoplasty for laryngomalacia had greater improvement in body mass index percentile post-operatively than older infants [83].

Evidence summary

OSAS symptoms in early life (snoring, mouth breathing and witnessed apnoea) are associated with lower cognitive development scores and predict behavioural problems at 4 and 7 years of age.

Literature review

Obstructive SDB symptoms at 6 and 18 months of age have been related to increased risk for behavioural morbidity and especially hyperactivity at the age of 7 years in a community-based, prospective study [84]. Longitudinal data suggest that frequent snoring in infancy is associated with lower cognitive development scores [85–87]. Increased obstructive AHI in infants with cleft lip and/or palate correlates with lower scores in the behavioural domain at the age of 3 years [88].

Evidence summary

a) Feeding difficulties.

b) Recurrent otitis media.

Literature review

a) In a small series of infants with adenotonsillar hypertrophy and OSAS, ∼15% of cases had feeding difficulties [20]. Infants with moderate-to-severe laryngomalacia often present with swallowing and feeding difficulties, especially in the presence of neurological disorders and/or syndromic comorbidities, and may improve after supraglottoplasty [89, 90]. Pierre Robin sequence is accompanied by feeding difficulties in 80% of cases, probably due to presence of cleft palate and increased work of breathing [91]. Low weight percentile is not associated with OSAS severity [91].

b) In a retrospective study of children with OSAS aged 3–24 months, the frequency of middle ear effusion requiring surgical treatment was 31.9% [92]. Infants with Down syndrome and OSAS have increased prevalence of recurrent otitis media [93].

Evidence summary

a) Video polysomnography.

b) Polysomnography and nap polysomnography.

c) Polygraphy.

d) Nocturnal pulse oximetry.

Literature review

a+b) In a review of 10 studies including infants without major anomalies aged 2 weeks to 24 months, polysomnography configuration consisted of: electroencephalogram, electro-oculogram, and submental and leg electromyogram channels; oronasal thermistor; thoracic and abdominal wall movement sensors; and transcutaneous oxygen monitor or pulse oximetry [7]. Polysomnography has been used to diagnose OSAS in otherwise healthy infants with adenotonsillar hypertrophy and patients with craniofacial abnormalities, neuromuscular disorders or genetic syndromes [30, 44, 91, 94]. Daytime nap polysomnography applied in infants and children with Down syndrome underestimates SDB severity [63]. In contrast, 4-h evening polysomnography had high sensitivity for OSAS in young patients with a variety of predisposing factors (laryngotracheomalacia, subglottic stenosis, congenital heart disease, gastro-oesophageal reflux, chronic aspiration, Down syndrome or Pierre Robin sequence) [63, 95].

c) Polygraphy (electroencephalogram, electro-oculogram or electromyogram channels not included) has been used in a small study of healthy infants and also in young children with Down syndrome or Prader–Willi syndrome [96–98].

d) Pulse oximetry has been applied for diagnosing OSAS in otherwise healthy children as young as 6 months and in complex patients with Down syndrome, mucopolysaccharidosis or Pierre Robin sequence [47, 67, 99, 100].

Evidence summary

a) The 90th percentile for the frequency of obstructive apnoeas or mixed apnoeas in healthy asymptomatic young children (1–23 months old) undergoing polysomnography does not exceed 1 episode·h^–1 and hypopnoeas are uncommon. Thus, the obstructive AHI is <1 episode·h^–1 in healthy young children. The 90th percentile for the oxygen desaturation (≥3%) of haemoglobin index is 2.2 episodes·h^–1 (1.1–1.9 years old). Healthy infants residing at high altitude have more obstructive events and oxyhaemoglobin desaturations than infants at sea level, but these findings improve with older age.

b) The frequency of central apnoeas varies widely according to age and definition. Therefore, the definition of OSAS in children aged 1–23 months is usually based on the obstructive AHI.

c) When infants with risk factors for OSAS are considered, mild OSAS is frequently diagnosed with obstructive AHI 1–5 episodes·h^–1, moderate OSAS with obstructive AHI >5–10 episodes·h^–1 and severe OSAS with obstructive AHI >10 episodes·h^–1.

d) The McGill criteria for scoring nocturnal oximetry have been applied in infants with adenotonsillar hypertrophy as young as 6 months. A McGill oximetry score >2 has been used to define OSAS in infants with Pierre Robin sequence and moderate-to-severe OSAS in patients with Down syndrome. An oxygen desaturation (≥4%) of haemoglobin index >4 episodes·h^–1 with median arterial oxygen saturation measured by pulse oximetry (S_pO2) <95% has been applied to define obstructive SDB in patients with mucopolysaccharidosis.

Literature review

a) Upper limits of obstructive and mixed apnoea index are supported by a review of 10 studies including infants without risk factors for OSAS who underwent polysomnography [7]. The minimum apnoea duration was 3 s, which corresponds approximately to the duration of two breaths for this age group [7, 13]. These findings were confirmed for older children (1.1–1.9 years old) in a German multicentre study [8]. No hypopnoeas (reduction ≥50% in airflow signal amplitude associated with S_pO2 drops ≥3% or an arousal) were noted. Higher values were reported for the 95th percentile of the obstructive apnoea index and mixed apnoea index in a study using polygraphy in a limited number of healthy infants: 3.5 and 1.1 episodes·h^–1, respectively, at 1 month of age; 2.2 and 0.7 episodes·h^–1, respectively, at 3 months of age [96]. For healthy infants residing in altitude >2500 m above sea level, the 95th percentile value of the obstructive AHI is elevated: 27.6 episodes·h^–1 at age <45 days decreasing to 1.8 episodes·h^–1 at age 10–18 months [101]. The 95th percentile for the frequency of S_pO2 drops >3% during active/rapid eye movement sleep ranges from 170.1 episodes·h^–1 at age <45 days to 68.2 episodes·h^–1 at age 10–18 months [101].

b) The 95th percentile for central apnoea index (cessation of airflow and respiratory effort for ≥3 s) varies widely according to age: 45 episodes·h^–1 for 1-month-old infants and 10–20 episodes·h^–1 for 3- to 12-month-old infants [7]. In the German multicentre study, the 90th percentile for central apnoea index (no airflow and respiratory effort for ≥5 s) was 4.3 episodes·h^–1 for ages 1.1–1.9 years [8]. For infants residing in high altitude, the 95th percentile value of the central AHI is elevated: 65.4 episodes·h^–1 at age <45 days decreasing to 8.7 episodes·h^–1 at age 10–18 months [101]. In view of the unique definition of central apnoea in infants and the wide variability of the upper reference limit for the central apnoea index in the first 23 months of age, in their practice, the ERS Task Force members use the obstructive AHI for diagnosing OSAS in this age group [13].

c) In studies reporting the efficacy of interventions for OSAS in infancy, obstructive AHI >1–1.5 episodes·h^–1 has been applied to define abnormal polysomnography [30, 91, 94]. In retrospective studies of infants (<1–2 years old) with risk factors for OSAS, AHI <1 episodes·h^–1 was considered normal, and AHI values of 1–5 episodes·h^–1, 5–10 or 15 episodes·h^–1 and >10 or 15 episodes·h^–1 were defined as mild, moderate and severe OSAS, respectively [30, 94, 102, 103]. Similar classification of OSAS severity was applied in infants with Pierre Robin sequence [91]. When infants with syndromic craniosynostosis were studied, OSAS was defined as mild with obstructive AHI <5 episodes·h^–1, moderate with obstructive AHI 5–24 episodes·h^–1 and severe with obstructive AHI ≥25 episodes·h^–1 [44].

d) An abnormal McGill oximetry score of ≥2 (≥3 clusters of desaturation events ≥4% and ≥3 desaturations to ≤90%) corresponds to OSAS of at least moderate severity, but a score of 1 does not exclude OSAS [99]. In patients with Down syndrome, a McGill oximetry score ≥3 (≥3 clusters of desaturation events ≥4% and >3 desaturations to <85%) has a positive predictive value of 94% and a specificity of 98% for identifying AHI ≥2.5 episodes·h^–1 [100]. In infants with Pierre Robin sequence, a McGill oximetry score >2 has been applied to determine indications for nasopharyngeal airway insertion [47]. Nocturnal oximetry has been used in children with mucopolysaccharidosis to diagnose obstructive SDB [67].

Evidence summary

a) Parental report of snoring, apnoea, restless sleep and mouth breathing.

b) History of cyanotic spells, ALTEs or delayed growth, if there are other symptoms or signs of OSAS.

c) Nasal obstruction (adenoidal hypertrophy with or without tonsillar hypertrophy, choanal atresia and pyriform aperture stenosis).

d) Laryngomalacia as suggested by persistent stridor without or with intercostal retractions, cyanotic spells or growth failure.

e) Syndromic craniosynostosis with or without midface hypoplasia; unrepaired or repaired isolated cleft lip or palate; and marked mandibular hypoplasia.

f) Neuromuscular disorders (cerebral palsy, mitochondrial disorders and spinal muscular atrophy).

g) Complex abnormalities (achondroplasia, Beckwith–Wiedemann syndrome, Chiari malformation, Down syndrome, mucopolysaccharidoses and Prader–Willi syndrome).

Literature review

According to the clinical experience of the ERS Task Force members, a multidisciplinary team approach is usually necessary, including a sleep specialist, ear, nose and throat (ENT)/craniofacial surgeon, intensivist, respiratory physiologist, orthodontist, and others [50, 104].

a+b) A history of snoring, apnoea or, less frequently, nocturnal desaturations has been used as an indication for polysomnography in young children [21, 30]. In patients with a history of ALTEs and symptoms or signs indicative of obstructive SDB, polysomnography has been applied to exclude OSAS [15]. OSAS may be demonstrated in up to 60% of such cases [25].

c) Adenoidal or tonsillar hypertrophy and choanal atresia are associated with OSAS [37, 105].

d) Children with laryngomalacia and intercostal retractions frequently have increased AHI [38, 39, 106].

e) Syndromic craniosynostosis with or without midface hypoplasia is characterised by a high frequency of obstructive events in polysomnography [44]. Infants with cleft lip/and or palate have increased OSAS prevalence [46]. Polysomnography and nocturnal oximetry have been used in patients with Pierre Robin sequence to direct treatment interventions [47, 52, 91].

f) Spinal muscular atrophy, cerebral palsy and mitochondrial disorders may be related to an elevated AHI [54–57].

g) Overall, infants with achondroplasia have increased AHI, but OSAS severity is not predicted by the size of the foramen magnum [58]. Both polysomnography and nocturnal oximetry have demonstrated increased OSAS prevalence in infants with Beckwith–Wiedemann syndrome, Chiari malformation, mucopolysaccharidosis or those with Down syndrome (especially with a history of dysphagia, gastro-oesophageal reflux disease, congenital heart disease or premature birth) [59, 60, 62, 67, 100]. Children with Prader–Willi syndrome aged <2 years have an increased frequency of obstructive events and central apnoeas [68, 69, 107].

Evidence summary

a) OSAS in association with apparent upper airway obstruction during wakefulness (mild or severe intercostal retractions, prolonged apnoeas with cyanotic spells while awake: “obstructive awake apnoea”) and without or with failure to thrive. Treatment prevents or reverses respiratory failure, pulmonary hypertension and growth delay.

b) Abnormalities in polysomnography, polygraphy or pulse oximetry in combination with snoring or “noisy” breathing, nocturnal tachypnoea, oral breathing, history of ALTEs, or delayed growth.

c) Abnormalities in polysomnography, polygraphy or pulse oximetry in association with risk factors for OSAS: nasal obstruction (e.g. adenoidal without or with tonsillar hypertrophy and choanal atresia); laryngomalacia; syndromic craniosynostosis with or without midface hypoplasia; isolated cleft lip or palate; mandibular hypoplasia (e.g. Pierre Robin sequence); or neuromuscular disorder.

Literature review

a) Children with OSAS and severe upper airway obstruction present even during wakefulness improve regarding symptoms and polysomnography findings following treatment interventions [38]. “Obstructive awake apnoea” was defined as complete or partial cessation of air exchange during wakefulness due to observable upper airway obstruction [75]. Severity of hypoxaemia (nocturnal oximetry) and intensity of work of breathing facilitate the choice of treatment [47]. Polysomnography is not always performed pre-operatively in children with a history of respiratory distress related to obstructive tonsils [103]. However, in such cases OSAS severity may be underestimated and thus appropriate anaesthetic and post-operative care might not be available.

b) Children with OSAS related to adenotonsillar hypertrophy or laryngomalacia and body mass index in the lower percentiles have symptom resolution and increased growth rate following adenotonsillectomy or supraglottoplasty [20, 83, 94, 108, 109].

c) Improvement in symptoms and/or AHI was demonstrated in infants with OSAS and adenotonsillar hypertrophy, laryngomalacia, choanal atresia, mandibular hypoplasia or craniosynostosis syndromes with or without midface hypoplasia after surgical treatment (adenoidectomy, tonsillectomy, adenotonsillectomy, supraglottoplasty, repair of choanal atresia, mandibular distraction osteogenesis and midface advancement) [20, 30, 38, 39, 44, 52, 94]. Nasopharyngeal tube insertion in infants with Pierre Robin sequence and McGill oximetry score >2 resulted in improved oximetry [47].

Evidence summary

a) Achondroplasia.

b) Beckwith–Wiedemann syndrome.

c) Chiari malformation.

d) Down syndrome.

e) Mucopolysaccharidoses.

f) Prader–Willi syndrome.

Literature review

a) It is controversial whether the increased mortality of infants with achondroplasia can be attributed to cervicomedullary compression and central respiratory control abnormalities resulting in central sleep apnoea [110, 111]. Although young children with achondroplasia have an increased frequency of obstructive and central apnoeas, the size of the foramen magnum is not related to the AHI, central apnoea index or desaturation index [58, 112].

b) Severe OSAS resolved in two infants with Beckwith–Wiedemann syndrome after division of the frenulum linguae with central tongue resection or anterior glossopexy [59].

c) Central sleep apnoea and OSAS in young patients with Chiari malformation may indicate brainstem crowding at the foramen magnum and need for decompressive surgery [60].

d) The majority of infants with Down syndrome and OSAS have severe disease and nocturnal hypoventilation, and they are predisposed to develop pulmonary hypertension [62, 113].

e) Children with any type of mucopolysaccharidosis develop progressively deteriorating OSAS [67].

f) OSAS prevalence is high in patients with Prader–Willi syndrome; in 25% of cases, OSAS is severe and improves following treatment (e.g. adenotonsillectomy) [68].

Evidence summary

a) Interventions for OSAS in young children are individualised according to aetiology, severity and morbidity.

b) Nasopharyngoscopy or drug-induced sleep endoscopy may be used to determine the type and sequence of treatment interventions.

c) In infants with OSAS due to multiple causes, surgical treatment is overall more effective in reducing symptoms than oxygen administration, use of antireflux medications or continuous positive airway pressure (CPAP) application. CPAP or non-invasive positive pressure ventilation (NPPV), tracheostomy and supraglottoplasty are equally effective in reducing the AHI.

Literature review

As was stated in 3.3, a multidisciplinary team management is frequently required, including a sleep specialist, ENT/craniofacial surgeon, intensivist, respiratory physiologist, orthodontist and other specialists [50, 104]. Adenotonsillectomy, CPAP/NPPV and tracheostomy are interventions that are used for the treatment of OSAS attributed to various conditions predisposing to upper airway obstruction. Nasal corticosteroids and/or montelukast administered for 6–12 weeks decrease the severity of mild-to-moderate OSAS in children aged >2 years, but no evidence exists for younger patients [3].

a) A stepwise approach has been described in two retrospective, cohort studies of infants with Pierre Robin sequence [91, 114]. In a prospective study of children with syndromic craniosynostosis with or without midface hypoplasia, cranial vault surgery was performed in the first year of life, followed in frequency by adenotonsillectomy, transverse widening of the hypoplastic maxilla, midface advancement, tracheostomy and ventilation [44].

b) Nasopharyngoscopy has been used in patients with Pierre Robin sequence to classify the type of pharyngeal airway obstruction and to select the type and sequence of interventions [74, 75, 115]. Drug-induced sleep endoscopy has been applied in otherwise healthy young children with OSAS to detect the level of obstruction and direct surgical interventions (e.g. adenotonsillectomy versus adenoidectomy) [116].

c) In a retrospective study of infants, surgical interventions showed better symptom resolution than oxygen administration or CPAP treatment [30]. Surgical treatment and antireflux medications were superior to no treatment regarding resolution of OSAS symptoms [30]. In another retrospective study, CPAP/NPPV, tracheostomy and supraglottoplasty were equally effective based on polysomnography [94].

Evidence summary

Limited data indicate that antireflux medications decrease OSAS severity in young children (e.g. those with laryngomalacia).

Literature review

Although antireflux medications are administered in up to two-thirds of infants with OSAS, there is weak evidence for improvement of treated patients versus not treated subjects [30, 94]. Although more severe laryngomalacia is associated with higher prevalence of reflux, antireflux medications do not improve apnoeas [117].

Evidence summary

a) Adenoidectomy alone is efficacious in improving OSAS symptoms and polysomnography parameters in infants (<12 months), but some patients may require tonsillectomy subsequently. Adenotonsillectomy for adenotonsillar hypertrophy is efficacious in patients with craniosynostosis with or without midface hypoplasia.

b) Young children with comorbidities (asthma, obesity, gastro-oesophageal reflux disease, Down syndrome, congenital heart disease, history of premature birth and cerebral palsy) or severe OSAS pre-operatively are at risk of residual disease post-operatively.

c) Children aged <3 years are at increased risk of respiratory compromise, bleeding and persistent poor oral intake after adenotonsillectomy. Risk for respiratory complications is highest in children aged <2 years and in those with adenoidal enlargement, nasal obstruction or cardiovascular anomalies.

Literature review

a) Adenotonsillectomy in infants with adenotonsillar hypertrophy and OSAS is accompanied by snoring resolution and improvement in growth and polysomnography findings [20, 105, 116, 118, 119]. Adenoidectomy alone is efficacious in most patients aged ≤12 months, but OSAS recurrence may necessitate additional tonsillectomy [120, 121]. In a prospective study of children with craniosynostosis syndromes (with or without midface hypoplasia), adenotonsillectomy was the most frequent intervention after the first year of life [44].

b) Following adenotonsillectomy, AHI >5 episodes·h^–1 was demonstrated in 20–65% with the highest prevalence of residual OSAS among patients with comorbidities [118–120].

c) Poor oral intake and dehydration are the most common complications [118, 122]. Children aged <3 years have double the risk for post-operative respiratory complications (e.g. laryngospasm, hypoxaemia, apnoea and increased work of breathing) compared with those aged 3–5 years, while nasal obstruction, large adenoid size and cardiovascular anomalies are significant predisposing factors [105, 118, 123, 124]. Nasopharyngeal airway placement, re-intubation, supplemental oxygen, CPAP or NPPV may be required [105, 118, 121, 123]. It has been recommended that children aged <3 years should be monitored as inpatients post-operatively due to the high risk of respiratory complications [10].

Evidence summary

The youngest age for adenoidectomy is 3 months and for adenotonsillectomy is 6 months.

Literature review

These statements are supported by retrospective studies of infants diagnosed with OSAS [102, 125].

Evidence summary

a) In children aged <24 months with moderate-to-severe OSAS who are not candidates for or do not improve after adenotonsillectomy or other surgical interventions: i) CPAP initiated at 4–6 cmH₂O and titrated up to 10 cmH₂O is an effective and well-tolerated treatment; ii) it can be applied as a temporary intervention while waiting for craniofacial surgery; iii) NPPV has been used in cases of OSAS coexisting with hypoventilation (e.g. spinal muscular atrophy type 1).

b) Most studies report minor complications. Occurrence of midface flattening has been described, but long-term consequences and potential reversibility of this finding are unknown.

c) A nasal mask is the most common interface for CPAP or NPPV; an oronasal mask is used for uncontrollable mouth leak or severe nasal obstruction.

Literature review

a) During NPPV therapy, bilevel positive airway pressure is applied, i.e. inspiratory and expiratory positive airway pressure [126, 127]. In one study of infants with OSAS, nasal CPAP (nCPAP) was initiated at 3.7 cmH₂O and increased by 0.3 cmH₂O increments until obstructive events were abolished [128]. The success rate of CPAP ranges from 38% to 100% [129]. OSAS in children aged <2 years with laryngomalacia improved with CPAP [130, 131]. Application of CPAP or bilevel positive airway pressure is accompanied by a reduction in respiratory rate and oesophageal pressure swings indicating unloading of the respiratory muscles [132]. nCPAP (6–8 cmH₂O) has been used for severe OSAS (AHI >10 episodes·h^–1) in infants with Pierre Robin sequence, as a bridge to craniofacial surgery, but also in cases with milder OSAS [91, 133, 134]. Retrospective studies support the use of nCPAP for moderate-to-severe OSAS in infants with cerebral palsy, achondroplasia, Beckwith–Wiedemann syndrome, Down syndrome or mucopolysaccharidosis [128, 134]. Furthermore, NPPV has been applied in infants with spinal muscular atrophy type 1 presenting with respiratory failure, paradoxical breathing and hypoventilation [135, 136]. Nocturnal NPPV normalises the AHI and desaturation index, and improves inspiratory muscle synchrony [135, 136]. Limited data support the use of high-flow nasal cannula therapy in children with nCPAP intolerance [137].

b) During periods of upper respiratory infection, nCPAP may be interrupted for a few days due to nasal obstruction and mouth leak [134]. Minor complications related to mask fit (eye or skin irritation) and nasal dryness have been noted [129]. Abdominal distension or emesis anecdotally may occur in prematurely born infants [134].

c) This statement is supported by retrospective studies [128, 134].

a) Choanal atresia or nasal pyriform aperture stenosis

b) Severe laryngomalacia

c) Syndromic craniosynostosis with or without midface hypoplasia

d) Severe mandibular hypoplasia (non-syndromic or syndromic Pierre Robin sequence)

e) Mandibular distraction osteogenesis

Evidence summary

a) Tracheostomy is an urgent procedure in patients with severe upper airway obstruction and a treatment option in cases with persistent symptoms when other surgical or non-surgical interventions fail to improve upper airway patency.

b) Tracheostomy can be a temporary measure until completion of craniofacial surgery.

c) Acute and life-threatening complications of tracheostomy include cannula obstruction by mucus and accidental decannulation. Long-term tracheostomy may delay acquisition of language skills.

Literature review

a+b) Tracheostomy may be required in the first month of life for upper airway obstruction due to midface or mandibular hypoplasia [44, 102, 156]. In children with craniosynostosis and midface hypoplasia, tracheostomy may be urgently required prior to craniofacial surgery [142]. In case series of infants with Pierre Robin sequence, tracheostomy was required in 5.5–20.9% of patients and was frequently maintained beyond the first year of life [47, 51, 53, 114, 115]. Tracheostomy has been proposed for infants who have type II–IV pharyngeal obstruction defined by endoscopy without response to insertion of a nasopharyngeal airway (see also 1.4) [75, 115].

c) Complications of tracheostomy have been summarised in a review article (pneumothorax, pneumomediastinum, bleeding, wound infection, cannula obstruction by mucus, accidental decannulation, frequent lower respiratory tract infections, formation of granulation tissue, tracheocutaneous fistulae, laryngotracheal stenosis, tracheoinnominate artery fistula, delayed language skills acquisition and adverse effects on facial development) [157].

Evidence summary

In children with OSAS secondary to complex conditions a combination of interventions may be required.

a) Achondroplasia: OSAS may improve after adenotonsillectomy, but a nasopharyngeal airway or CPAP is necessary if upper airway obstruction persists post-operatively.

b) Chiari malformation: adenoidectomy and/or tonsillectomy, CPAP or NPPV may be required to treat OSAS, central sleep apnoea and nocturnal hypoventilation; surgical decompression of the cervicomedullary junction is accompanied by a decrease in the frequency of central apnoeas.

c) Down syndrome: adenoidectomy or adenotonsillectomy for adenoidal or tonsillar hypertrophy and supraglottoplasty for laryngomalacia have been reported. CPAP has been applied for persistent OSAS post-operatively or as first-line treatment in the absence of adenotonsillar hypertrophy. Some infants with Down syndrome outgrow OSAS within several months.

d) Mucopolysaccharidoses: adenotonsillectomy and CPAP are used to relieve upper airway obstruction, while enzyme replacement and haemopoietic stem cell transplantation target the metabolic disorder.

e) Prader–Willi syndrome: adenotonsillectomy for OSAS; oxygen therapy for central sleep apnoeas which decrease in frequency with age.

Literature review

a) Treatment interventions have been summarised in retrospective studies [158, 159].

b) Treatment of SDB in Chiari malformation has been described in retrospective studies of older children and adolescents [60, 160].

c) Interventions are based on low-quality evidence [161]. OSAS may resolve spontaneously in infants with Down syndrome and thus CPAP may be preferable over tracheostomy in selected cases [161].

d) In a prospective study, haemopoietic stem cell transplantation was more efficacious than enzyme replacement therapy for mucopolysaccharidosis type I in improving OSAS severity [67].

e) Resolution of central and obstructive events with age was reported in infants, most of whom were on growth hormone [107, 162].

Evidence summary

a) Recurrence of OSAS following adenotonsillectomy has been reported as early as 4–6 months post-operatively and repeat adenoidectomy may be required.

b) Following initiation of CPAP, polysomnography is repeated every 2–4 months during the first year of life and every 6 months thereafter to confirm the continued need for treatment and potentially increase the airway pressure.

c) In children with neuromuscular disorders, polygraphy or nocturnal oximetry/capnometry is performed after establishment of NPPV and repeated at least annually.

d) The efficacy of supraglottoplasty has been evaluated by polysomnography at 1–6 months post-operatively.

e) Infants with Pierre Robin sequence and moderate-to-severe OSAS may require a nasopharyngeal airway for <12 months. Sleep studies or nocturnal oximetry are performed as frequently as every 2 months and after removal of the artificial airway.

f) Serial polysomnographies are performed during mandibular distraction osteogenesis.

Literature review

a) In a group of children aged <3 years, a high rate of residual OSAS after adenotonsillectomy was found which was predicted by pre-operative disease severity [119]. Repeat adenoidectomy was performed for recurrent OSAS after adenotonsillectomy in a retrospective study [20].

b) Infants with OSAS related to anatomical abnormalities may require CPAP and potential increase of airway pressure and upgrade of nasal mask size for several years after treatment initiation, whereas OSAS may resolve in patients without apparent abnormalities [128, 129, 134]. In children of pre-school age with micrognathia, the required airway pressure can progressively reach 10 cmH₂O [128, 133].

c) An expert opinion statement from the British Thoracic Society Guidelines suggests performing polygraphy or nocturnal oximetry/capnometry after establishment of NPPV and repeating it at least annually in children with neuromuscular disorders [163].

d) Improvement of OSAS after supraglottoplasty has been demonstrated in a retrospective study [38].

e) Monitoring of infants with Pierre Robin sequence and moderate-to-severe OSAS has been described in a retrospective report [47].

f) Experience with serial polysomnographies during mandibular distraction osteogenesis in infants with Pierre Robin sequence has been described in a case series [52].

Similarities and differences in the diagnosis and management of obstructive SDB in this age group compared with older children are summarised in table 2.