Who needs a sleep study: current indications for OSA I, II and III

Sleep medicine is a relatively new area of paediatric medicine with OSA representing the most common sleep disorder in children. Hypertrophy of the tonsils and/or adenoids are the major cause of OSA in young otherwise children (also called OSA type I), with most children having mild or moderate OSA, which may resolve either spontaneously or after adeno-tonsillectomy . Adenotonsillectomy has shown to be associated with an improvement in neurocognitive dysfunction and abnormal behaviour, even if some children will present mild residual OSA after surgery . To avoid unnecessary, labour-intensive and expensive investigations and a delay in management, international and national guidelines do not recommend a systematic sleep study in otherwise healthy children with OSA type I who present overt clinical OSA symptoms associated with hypertrophy of the tonsils and/or adenoids . A sleep study is however recommended in young children, aged less than 2 years, in whom OSA is less common and who have a greater anaesthetic risk than their older counterparts. A sleep study is also recommended in case of a discrepancy between symptoms and clinical findings on the examination of the upper airway by the ENT surgeon and in children with an anaesthetic risk due to pulmonary hypertension or a bleeding disorder . The demand of some parents to have a sleep study before ENT surgery, because of abusive and inappropriate use of internet and media, even if their child does not fulfil the above-mentioned criteria, is an increasing and difficult challenge, which requires explanation, understanding, and time.

The escalating global obesity epidemics represents one of the most serious public health challenges, not only in adults but also in children . Overweight and obese children have an increased risk of OSA (also called OSA type II), with OSA being in general more severe than in lean children and associated with a greater risk of residual OSA after adeno-tonsillectomy . Obese children have also an increased risk of post-operative complications. These particularities explain the recommendation of a PSG in an obese child with a suspicion of OSA. However, due to the high and increasing number of children, it is not possible to screen every obese child for OSA with a PSG. Several studies have thus tried to identify predictive factors for OSA in obese children. Even if the paediatric sleep questionnaire (PSQ) has been shown to identify an apnoea-hypopnea index (AHI) > 5 events/hour with an 80% sensitivity and 100% specificity for OSA in a sample of 60 children and adolescents , questionnaires have been found to be less informative in other studies. Body mass index (BMI) or BMI z-score are the gold standard parameters to quantify the severity of obesity but their correlation with the AHI is poor . Other parameters, such as neck circumference, neck-to-height ratio (NHR), or waist circumference, have shown to have a better correlation with the AHI. Indeed, the NHR correlated with the obstructive AHI in a series of 71 obese children aged 8–17 years . But in obese children, male gender, older age, and life-style habits such as substance-use behaviours, have also been shown to be associated with OSA severity . In clinical practice, these various parameters may be useful to prioritize a patient for a sleep study, but they seem insufficiently sensitive and specific to propose an ENT surgery without a prior sleep study . Indeed, it is important to have the pre-operative AHI to evaluate the efficacy of an upper airway surgery for an individual patient. Moreover, lower oxygen saturation (SpO ₂ ) nadirs and the sleep time spent with a SpO ₂ < 90%, and not the obstructive apnoea index, have shown to be independent predictors of post-operative respiratory complications . In an obese child with a suspicion of OSA, a sleep study seems thus mandatory to confirm the diagnosis of OSA, to assess its severity, and the risk of post-operative respiratory complications.

Children with anatomical and/or functional abnormalities of the upper airways represent a third and increasing population requiring systematic sleep studies . Children with facial malformations, such as facial bone stenosis, craniofacial bone stenosis, Pierre Robin Sequence, Treacher-Collins syndrome, micrognathia or any other type of malformation of the maxilla or mandible, but also children with storage diseases, such as mucolipidosis or mucopolysaccharidosis, congenital bone disease, including achondroplasia or pycnodysostosis, Prader Willi syndrome, or Down syndrome, present an increased risk of OSA . Most of these diseases are rare diseases but the number of diseases and the prevalence of OSA are both extremely high, explaining the high number of patients. Also, the risk of OSA persists across childhood and into adulthood, with the consequent need for repetitive studies . Symptoms, clinical examination, and sleep questionnaires are insufficiently sensitive and specific as screening tools to document OSA and its severity in these patients . OSA is generally severe, with high risk of residual OSA after upper airway surgery. OSA may also occur at any age, even in infants, which may require sleep laboratories having an expertise in PSG in very young children. Some of these diseases, such as Down syndrome, Prader Will syndrome, or mucopolysaccharidosis, are associated with behaviour problems and neurocognitive dysfunction, which may limit the cooperation of the child and requires technicians and staff having a training and expertise in these difficult patients . A full PSG may not be possible in some children. A PG, without electroencephalography, may then be better tolerated and associated with a greater rate of success.

Types of sleep breathing assessment

PSG remains the gold standard for the diagnosis of OSA and includes electroencephalography (EEG), electro-oculogram, submental and leg electromyogram, oronasal airflow, abdominal and chest wall movements, SpO ₂ , partial pressure of end-tidal or transcutaneous carbon dioxide, and a videorecording. However, access, availability and costs remain obstacles to widespread utilisation. Additionally, the global burden of the obesity epidemic in all ages means that PSGs alone will never be able to meet the clinical need for assessment of OSA. Alternative suggestions to PSGs in children have been proposed for over 20 years . As a result, clinicians are left with limited options for the provisional diagnosis of OSA and assessment of severity to inform prioritisation for interventions.

Polysomnography

PSG has been used for the assessment of sleep-disordered breathing in children for over 30 years. It distinguishes the predominantly seen problem of OSA from central sleep apnoea and provides physiologic parameters to determine the severity of sleep-disordered breathing in terms of hypoxaemia, altered heart rate and effort of breathing. There are standardised indications for PSG in children aged 1–18 years and for younger children [<2 years of age] . The standard parameter derived from the PSG is the apnoea-hypopnoea index [AHI] which is the frequency of significant respiratory events (apnoeas and hypopneas). The OSA threshold for diagnosis in children is much lower than required in adulthood. An AHI >1 event per hour is most used for the diagnosis of OSA in children . In addition, there are non-respiratory indications for PSGs in children . For both pragmatic reasons and physiologic reasons with sleep cycles, daytime of “nap” studies are often undertaken in younger children . Nap studies [sleeping during the daytime] were evaluated in young children more than 20 years ago and it was shown that individual nap study parameters are not very sensitive in predicting abnormal overnight PSG findings. However, when nap study parameters are abnormal, the chance of OSA is high. . Nonetheless, nap studies have not been endorsed by regulatory authorities as a satisfactory replacement for full overnight PSG . Notably, there may be a gradation of age-related abnormalities in apnoea indices in the first two years of life . Further considerations are warranted for the ranges of apnoea indices in preterm graduates .

Home based, limited channel [seven or more channels] sleep studies are well established in adults and have results of similar diagnostic accuracy whether studied in the home setting or sleep laboratory using the same equipment . Studies are beginning to evaluate teenagers with home based unattended sleep studies . The testing is being introduced with some caution into the paediatric age group out of necessity . There remain the long-standing clinician and parental concerns about reliability of data acquisition and safety for young children in the scientifically unsupervised home environment. However, the reality is that with better equipment becoming available, the socio-economic factors for patients and families as well as health-care providers will drive non full PSG sleep diagnostic services in the appropriately selected patient, be they an adult or perhaps a school-aged child .

Polygraphy

Whilst PSG is the gold standard for confirming the presence of OSA and other forms of sleep-related breathing disorders , there are circumstances where a PSG is not feasible. Reasons may include children with severe anxiety [e.g., more common in Duchenne Muscular Dystrophy], children with autism and significant sensory issues, and children with neurocognitive delay . PG, which is a limited channel overnight recording, has been demonstrated to a be feasible and practical alternative to PSG . The European Respiratory Society (ERS) has included PG in the diagnostic algorithms for sleep-disordered breathing in infants and young children . PG, also referred to as a cardio-respiratory sleep study, is a scaled down PSG with fewer leads that has been used in the hospital or home setting. There are no electro-ocular graph [EOG] or electro-encephalogram [EEG] leads, so sleep staging and arousal scoring are not possible. Nasal airflow and snoring are measured with a nasal cannula pressure transducer and respiratory effort is assessed using combined thoracic and abdominal respiratory inductance plethysmography [RIP] belts, SpO ₂ and heart rate with SpO ₂ . Electrocardiography and body position may also be recorded.

PG is safe and feasible for many children in the home setting . Marcus and colleagues showed that 200 former preterm infants with birthweights of 500-1250g could be studied at ages 5–12 years with PG in the home setting. Data were satisfactory in 91% cases on initial testing and with retesting in the subgroup of 14 children, the satisfactory data proportion rose to 98%. Nasal pressure was the only real challenge with adequate data for >74% of the study time achieved in only 67% of subjects whilst the thermistor signal was adequate for >74% of study time in 92% of subjects. The study by Michelet et al. reported 400 PGs over 4 years [2012–2015] which were mainly initial single night studies [83%] and conducted at home [84% vs hospital 16%] with similar satisfactory interpretability [87%]. The indications included suspected OSA [reported snoring, laboured breathing, daytime fatigue or hyperactivity, witnessed apnoea, and any grade of adenotonsillar hypertrophy: 65%], obesity [BMI >97th centile or BMI z-score > 2 SD: 13%], craniofacial malformations [midface hypoplasia, Pierre Robin Sequence, laryngeal malformations: 5%], neuromuscular disease [mainly Duchenne Muscular Dystrophy: 4%] and “other” [preterm infants, Prader Willi syndrome, chronic respiratory infections, Arnold-Chiari malformation and preoperative non OSA assessments: 13%]. Non-interpretability mainly related to the absence of the trace or inaccuracy of SpO ₂ . Children studied at home had a mean age of approximately 5.5 years and in hospital 7.5 years.

Oximetry

SpO ₂ with a downloadable recording is the simplest of screening tests providing data on peripheral SpO ₂ and heart rate. It has been shown to be feasible in the home setting for children above one year of age and is best suited to children with a high pre-test probability of having moderate to severe OSA using the McGill scoring criteria . Jonus et al. retrospectively studied 110 children who had both a full PSG and SpO ₂ assessed using the McGill scoring [56% normal, 27% Grade 2, 17% grade 3 and 4] and found it to provide satisfactory performance in moderate to severe OSA but in normal SpO ₂ or McGill 2 severity it could not rule out OSA, necessitating a PSG to confirm or exclude the diagnosis. In a systematic review of 25 papers, Kaditis et al. in 2016 argued in favour of SpO ₂ being a feasible tool in planning adenotonsillectomy, noting that nocturnal SpO ₂ drops <90%, more than two clusters of desaturation events (≥4%) and oxyhemoglobin desaturation (≥4%) index (ODI4) >2.2 episodes/h are unusual in children without OSA. When overnight SpO ₂ is used, with or without the McGill score, in a patient with symptoms suggestive of OSA [snoring and mouth breathing], potentially predisposing factors [family history, trisomy 21, obesity] consistent physical findings [enlarged tonsils, low muscle tone, craniofacial abnormalities] then overnight SpO ₂ may be useful to prioritise interventions based on the oxygen profile in children over the age of 1 year with a high likelihood of OSA.

Contactless assessment of sleep-disordered breathing

It has been long appreciated that that PSG-derived indices do not adequately characterize childhood sleep-disordered breathing. Most children with clinically significant sleep-disordered breathing have long periods of sleep with partial upper airway obstruction and sustained runs of laboured breathing which is not captured by the AHI . This has seen the rise recently of more sensitive testing for milder upper airway obstruction manifesting with stertor and snoring. A non-contact mat system, the Sonomat™ has been validated against PSG in children for measurement of PSG metrics such as total sleep time (TST) and AHI with sensitivity and specificity of 86% and 96% respectively at a threshold of 5 events per hour . This system has advantages for children with neurodiversity including those with sensory issues. It has been demonstrated to be efficacious in children with Trisomy 21 . Consequently, it has the potential to be a viable, cost effective and scalable tool for home-based screening of children with likely OSA.

Conclusion

Through necessity, the desire and ability to expand the diagnostic utility of a range of testing procedures for children with OSA will continue to expand and challenge traditional models of care. It is up to the clinicians to respond with the evidence to support the most practical and efficacious testing modalities in a sustainable, cost-effective, and acceptable manner for patients and their families.

Future directions for research

• •
  

  Generating international consensus on practical and scalable responses to the challenges of diagnosing sleep-related breathing disorders / OSA in the setting of a tsunami of obesity beginning in childhood.

  
  
• •
  

  The development of evidence-based algorithms for the assessment of OSA in the home setting for infants and children.