Prior to a discussion of the efficacy of treatment of childhood OSA, including the utility of nNIV, it is essential to summarily review the unique pathophysiological attributes of OSA in children. The cardinal abnormality associated with childhood OSA is the presence of hypertrophic adenotonsillar tissue. Enlargement of these tissues will reduce the anatomical patency of the upper airway and thus lead to exponential increases in pharyngeal resistance, ultimately resulting in episodic airway collapse during sleep, characteristic of OSA [45, 46]. While the presence of enlarged tonsils and adenoids does not uniformly and reliably predict the likelihood of OSA in children, [47] the concurrence of habitual snoring and adenotonsillar enlargement should serve as a major instigator of diagnostic inquiry to confirm or rule out the presence of OSA.

Notwithstanding, several other risk factors such as obesity, craniofacial, and neuromuscular elements may all independently contribute to the risk of OSA in children. The impact of obesity on the development of OSA in children is particularly topical given the extent of the obesity epidemic affecting children worldwide [48]. With obesity prevalence rates ranging from 7 to 22 % of children in various Western countries, [49, 50] studies examining the prevalence of OSA have detected substantial increases in risk when obesity is concurrently present [51]. Indeed, for each increase of 1 kg/m² in BMI above the mean, the risk of OSA increases by 12 % in children [52]. Consequently, multiple studies from all over the globe have consistently shown that the presence of obesity in children significantly increases the risk of OSA [53–56]. In fact, we have recently shown that the degree of adenotonsillar hypertrophy required is lesser in obese children at any given level of OSA severity [57]. Further complicating the scope of clinical practice, obesity-induced childhood OSA presents very differently from the classical OSA phenotype that is exclusively induced by adenotonsillar hypertrophy [58]. OSA in obese children strikingly resembles that of adult patients with OSA, who in their vast majority are obese, suggesting that similar mechanisms in adults and obese children lead to the clinical manifestations and morbid consequences of OSA in these patients.

In spite of robust progress in our knowledge on the complications of OSA, the same cannot be concluded in the context of a critical review of studies involving the treatment of OSA. In 2002, members of the American Academy of Pediatrics Task Force established that adenotonsillectomy should be considered as the first line treatment for pediatric OSA, and such assessment was based on descriptive uncontrolled patient series with no randomized studies being available [59]. After 10 years, this task force updated their recommendations, but no real changes emerged in their conclusions or the available evidence supporting them [60].

Notwithstanding the paucity of critically needed studies, the updated guidelines shed light into some of the potential limitations and pitfalls of adenotonsillectomy by providing a list of relative contraindications to this surgery (Fig. 6.1). Among these, the presence of small adenoids/tonsils (occupying < 25 % of airway diameter for adenoids and + 1 size using common visual scales for tonsils [57]), particularly with the presence of concomitant morbid obesity with small adenoids/tonsils, the presence of submucosal cleft palate, and the presence of a bleeding dyscrasia that is refractory to treatment emerged. Similar to the 2002 guidelines, the more recent statement repeated the recommendation for follow-up evaluation after adenotonsillectomy including polysomnography for patients with persistent symptoms of residual obstructive SDB.

Current estimates aimed at predicting residual OSA following adenotonsillectomy have been based on four recent systematic meta-analyses, one large multicenter retrospective study, and one prospective randomized control study. In the first meta-analysis, we raised the concern that adenotonsillectomy may not be achieving as favorable outcomes as had been previously anticipated, with an estimated 15 % failure rate [61]. Subsequently, Brietzke and Gallagher [62] critically reviewed 14 published studies and identified an overall “surgical success” rate of 82.9 %. Unfortunately, as with the initial meta-analysis, this report included studies that had differing definitions of success, including many sleep studies that were technically suboptimal based on the standards of polysomnography outlined by the American Academy of Sleep Medicine (AASM) [63]. Expanding on this work, Friedman et al. [64] included nine additional studies with an emphasis on obese children. Using a weighted mean analysis, the mean apnea hypopnea index (AHI) improved from 18.6 to 4.9 events/h. Furthermore, using surgical success criteria of an AHI < 5/total sleep time (TST), only 66.2 % normalized their sleep studies after the surgical procedure (95 % CI 54.5–76.3 %, P = 0.007), thereby indicating that > 30 % of children had residual moderate to severe OSA following adenotonsillectomy. The major implication derived from this paper was to suggest that the presence of obesity in children diminishes the efficacy of adenotonsillectomy in the treatment of pediatric OSA. In the fourth and last available meta-analysis by Costa and Mitchell [65], the authors analyzed four studies that examined the efficacy of adenotonsillectomy exclusively among obese children with OSA. As a corollary to the previous meta-analyses, the presence of obesity curtailed the response to adenotonsillectomy with success emerging in only 38.5 % of obese children and 51 % of patients having a postoperative AHI > 5/h TST, that is the equivalent of moderate to severe residual OSA.

In a large multicenter retrospective study that included the participation of six pediatric sleep centers in the USA as well as two European pediatric sleep centers that routinely perform pre- and postoperative sleep studies, Bhattacharjee et al. [66] were able to critically evaluate potential demographic, clinical, and anthropometric confounding factors that may influence the postoperative AHI using multivariate logistic regression modeling approaches on a dataset that included 578 children. In this study, 27 % of all children demonstrated complete resolution of OSA following adenotonsillectomy with a postoperative AHI < 1/h TST, and 22 % had residual moderate to severe OSA with an AHI > 5/h TST. The strongest determining factors associated with residual OSA were older age, the presence of obesity, the presence of asthma, and finally the severity of underlying OSA prior to surgery (Fig. 6.1). The major contributions from this study were the improved delineation of groups at higher risk for surgical ineffectiveness who could therefore benefit from a postoperative polysomnogram (PSG) in order to ascertain either OSA resolution or residual OSA. In the context of a high prevalence of obesity in children, many children who undergo adenotonsillectomy for the treatment of OSA will be left with persistent SDB following surgery, and will therefore require additional treatment strategies including nNIV. Furthering the complexities of OSA in children, follow-up assessments in the large longitudinal cohort TuCASA study revealed that children with persistent OSA are at an elevated risk of developing obesity after 5 years [67].

Finally, in the single multicenter randomized control trial of adenotonsillectomy in children, involving 464 children (226 children undergoing surgery), it was found that 79 % of children undergoing adenotonsillectomy had successful resolution of OSA as defined by an AHI < 2/h TST, as compared with OSA resolution being recorded in 49 % of control children who did not undergo any surgical intervention. Paralleling the finding from the aforementioned retrospective studies, the risk of residual OSA was again related to the presence of obesity and the underlying severity of SDB, but this study additionally determined that African-American children are particularly at high risk for residual OSA [68].

Therefore, even though the vast majority of children will demonstrate favorable responses to adenotonsillectomy in the treatment of OSA, the high prevalence of pediatric obesity would predict an ever increasing number of children [69] for whom adenotonsillectomy will either be contraindicated or unsuccessful. As such, a systematic implementation of approaches aiming to identify the large proportion of children with OSA in whom adenotonsillectomy will likely fail and who will require nonsurgical approaches [70] including nNIV as the mainstay of therapy is warranted. In children in whom adenotonsillectomy falls short of relieving obstructive breathing during sleep, including children with craniofacial anomalies, obesity, neuromuscular weakness, the application of nNIV (Fig. 6.1) including CPAP has expanded, particularly as our experience with these devices in treating children has grown. In one of the first reports in 1986 by Guilleminault and colleagues, nasal CPAP offered a novel alternative to tracheostomy in 10 young children with OSA [71]. In a review of practices from nine sleep centers across North America and one center in Europe, Marcus and colleagues [72] reported that as early as the mid-1990s CPAP was used, summarized in a series of 94 patients, ages 1–19 years. This cohort also included three children in whom an existing tracheostomy was successfully decannulated with the implementation of nNIV. In all but eight of the centers, nNIV represented the second line of treatment for OSA after adenotonsillectomy failed. In a similar study of 80 children, Waters and colleagues described a clinical protocol for introduction of CPAP to children including habituation procedures, as described above. In a study of 66 children, Masa and colleagues studied CPAP usage to treat OSA with an emphasis on close follow-up measures including telephone support, clinical assessments, and sleep studies at 1-month, 6-month, and 1-year intervals. Although adherence problems were clearly apparent, the feasibility of CPAP in a young pediatric population was documented. In studies by Downey et al., which included 18 children under the age of two, and McNamara et al., which included infants only, it was further confirmed that CPAP was an effective therapy in the youngest children with OSA.

Several recent studies have confirmed that BPAP in addition to CPAP can be successfully used to treat OSA in children. In a study by Marcus and colleagues [73], the efficacy of nNIV was examined in 29 children, of whom 13 were assigned CPAP and 16 were assigned to BPAP. While adherence was a concern due to a large proportion of dropouts, this study did not suggest any obvious differences in either CPAP or BPAP groups with respect to adherence or efficacy in treatment. Side effects were common in about 10–20 % of children, primarily consisting of equipment or mask problems, symptoms related to mask leak, and skin erythema related to the mask. Symptoms of nasal congestion and/or epistaxis developed in 38 % after 5 months of treatment. Uong et al. [74] retrospectively examined 46 school-aged children and showed a significant reduction in AHI from 28.4 to 3.8/hr TST with the application of either CPAP or BPAP. Other parameters, including oxygen saturation nadir and elevated end tidal carbon dioxide levels also demonstrated significant improvements following the institution of nNIV. All symptoms of OSA included in the survey exhibited significant improvements following nNIV, including snoring, witnessed apnea enuresis, daytime somnolence, hyperactivity or behavioral problems, and deteriorating school performance. Adherence was available in 27 children, and adherence as defined by using nNIV for > 4 h per night and ≥ 5 nights per week was only observed in 19 children (70 %). In a randomized, double blind clinical trial comparing BPAP to CPAP in 56 consecutive children aged 2–16, Marcus et al. [75] randomized children in a 3:1 ratio of BPAP to CPAP. CPAP was shown to have equivalent efficacy to BPAP as the AHI was reduced from 22 to 2/h TST using CPAP, while BPAP improved the AHI from 18 to 2/h TST, with both treatment modalities resulting in small improvements in subjective sleepiness as measured by the modified Epworth Sleepiness Scale. While efficacy data in both CPAP and BPAP groups were promising, the low adherence rates with both CPAP and BPAP were discouraging.

The aforementioned studies all confirmed the polysomnographic efficacy of nNIV in the resolution of OSA; however, few studies have evaluated the efficacy of nNIV in alleviating the clinical symptoms of pediatric OSA. In a recent prospective study of 52 children, Marcus and colleagues employed surveys assessing neurobehavioral status, quality of life, subjective sleepiness, and surveillance of symptoms of attention deficit hyperactivity disorder (ADHD) at baseline and following 3 months in children with OSA who were treated with nNIV. In a contextual setting in which objectively measured adherence was suboptimal (mean adherence of 3 h/night), there were significant improvements in daytime sleepiness, symptoms of ADHD, and quality of life following 3 months of nNIV.

However, there is no doubt that the markedly limited wealth of information currently available in the setting of pediatric OSA treatment in general, and of nNIV use in particular, will require a concerted effort to address the most pressing and pending research and clinical issues if progress is to be made in the near future.

In the absence of large cross-sectional studies, the exact prevalence of alveolar hypoventilation has not been established in children with any of the more frequent neuromuscular disorders. For example, there are estimates that 27–62 % of children with Duchenne muscular dystrophy (DMD) display some form of SDB [81]. However, since there are no longitudinal assessments, such figures may be skewed and represent a later stage of disease, at a time when symptoms prompt referral to pulmonary services. The pathophysiology of SDB [82] in children with neuromuscular disorders is dependent mostly on the nature and severity of the underlying disorder, but is also affected by the patient’s age, particularly if the disorder is progressive, and the type and extent of muscle involvement. Children with neuromuscular disorders have largely preserved central ventilatory drive with the exception of some children with myotonic dystrophy, in whom reduction in respiratory drive may be a component of the disease [83]. However, the overall loss of muscular tone, especially muscles that are integral to the respiratory system, begins an inexorable course that ultimately leads to profound disturbances of ventilation during sleep, even if respiratory muscle training and other rehabilitation approaches may slow down the process [84]. Furthermore, the atonia associated with REM sleep further accentuates the muscular weakness associated with neuromuscular disorders. Therefore, in conditions leading to intercostal muscle weakness, the loss of functional residual capacity (FRC) is augmented by the supine position of sleep, but this vulnerability during REM sleep leads to reduced oxygen reserves and an increased predisposition to carbon dioxide retention.

Neuromuscular disorders that amount to tonic deficits of the upper respiratory muscle such as in conditions of cerebral palsy, myotonic dystrophy, and sensory and motor neuropathies such as Charcot–Marie–Tooth or poliomyelitis are likely to contribute to the occurrence of obstructive SDB. Neuromuscular disorders such as DMD and Becker muscular dystrophy (BMD), mitochondrial myopathies, spinal muscular atrophy (SMA), cervical spine injuries, and metabolic and congenital myopathies are all associated with intercostal muscle weakness (sometimes preferentially affecting inspiratory or expiratory intercostals) that contribute to the pathogenesis of hypoventilation . In addition, certain disease states, including DMD and BMD, increase the risk of obesity or of disproportionate adipose tissue distribution in visceral and intramuscular regions even when BMI is preserved [85–87], and all children are at risk for adenotonsillar hypertrophy all of which further increase the risk for SDB. Taken together, the extent of SDB in children with neuromuscular disorders is largely contingent on the nature of the primary disease, more specifically the degree and nature of muscular insufficiency and the presence of disease-specific comorbidities [88]. In addition, related to intercostal muscular weakness, children with neuromuscular disorders often have an ineffective cough and ineffective airway clearance. Consequentially, pulmonary mucus plugging is a common occurrence in many children and predisposes children to recurrent pneumonias and atelectasis that secondarily contribute to the onset of pulmonary scarring and respiratory insufficiency. Finally, the absence of sufficient muscle tone does not provide the necessary structural support to the ribcage, such that these children are often prone to impaired thoracic cage development manifesting as severe scoliosis, pectus excavatum, flattening of the antero–posterior chest diameter, and a funnel-shaped chest. As a consequence of these skeletal changes, there is a loss of chest wall compliance and subsequent reduction in pulmonary volumes including FRC, paradoxical breathing, and elevated work of breathing, all of which further predispose the child to nocturnal alveolar hypoventilation.

A complete review of the various neuromuscular disorders is beyond the scope of this chapter, but the reader is invited to peruse several reviews [82, 88, 89]. Nonetheless, while nNIV and invasive ventilation is the mainstay modality for treating SDB in these children, it is again emphasized that nocturnal alveolar hypoventilation in these children is not easy to detect and requires a high level of suspicion and also the implementation of systematic diagnostic approaches when possible while preserving individualized diagnostic and management decisions. Further, treatment of the nocturnal hypoventilation, once identified, is itself problematic.

With the emergence of practice parameter guidelines related to DMD [81] and other congenital myopathies, it has been suggested that performing annual polysomnography to screen for nocturnal hypoventilation would be a sensitive and cost-effective approach with potentially improved outcomes. Since access to polysomnography poses a challenge to many children, the recommendations also encompassed the alternative use of overnight oximetry and continuous CO₂ monitoring (transcutaneous or end-tidal) or capillary blood gas testing in lieu of polysomnography in centers lacking pediatric polysomnography capabilities. Notwithstanding, clinicians are still advised to monitor for clinical symptoms of nocturnal alveolar hypoventilation (Table 6.2) during routine follow-up appointments. The American Thoracic Society [81] emphasizes that symptoms of SDB and sleep quality should be reviewed with each patient encounter. Finally, Gozal placed additional emphasis of the importance of a multidisciplinary approach to diagnosis and rehabilitation of patients with neuromuscular diseases including routine pulmonary function testing, nutritional, cardiac, orthopedic, and physical therapy assessments [88]. Such recommendations have gained substantial traction and are now widely implemented at major centers [90, 91].