Estimated prevalence rates for SDB in children remain imprecise because universally accepted diagnostic criteria for these conditions were only recently established and because large population-based studies using PSG have not been undertaken. It is estimated that between 5 and 12 % of children snore habitually, with some reports suggesting an increased risk of children exposed to tobacco smoke [59–63]. The prevalence of OSA in children is generally estimated to be 1–3 % [64–67].

Data are not available regarding the prevalence of UARS, central sleep apnea, and hypoventilation in children. CCHS is a rare disorder conservatively estimated to affect at least 300 children worldwide [46]. Prevalence for primary sleep apnea of infancy varies with size and gestational age. Symptomatic apnea of infancy affects 84 % of newborns weighing less than 1000 g, 25 % of newborns weighing less than 2500 g, and less than 0.5 % of term infants [52].

The clinical manifestations of SDB in children (see also Chap. 32) differ from those exhibited by adults, as summarized in Table 52.1. Children with obstructive SDB tend to be noisy breathers during sleep, but severity may range from heroic snoring or stridor to only minimally loud respiration. Snoring may be continuous, intermittent, or vary with body position. Snoring often worsens during upper respiratory infections or with exacerbations of allergic rhinitis and sometimes improves during treatment with decongestants or nasal steroids. Snoring in children with obstructive SDB is often accompanied by prominent mouth breathing and unusual sleeping positions such as neck hyperextension or excessive propping upon pillows, which represent compensatory mechanisms that may improve airway patency.

Witnessed apneas are only occasionally reported for children with obstructive SDB. This observation is consistent with the premise that prolonged partial airway obstruction during sleep is more common in children than the recurring episodes of brief obstruction followed by arousal that characterize typical adult OSA. When an obstructive apnea is witnessed in a child, parents may report paradoxical chest wall motion or the presence of snorting or gasping noises as respiration resumes.

Children with obstructive SDB demonstrate greater degrees of restlessness, enuresis, and perspiration during sleep than healthy controls [68]. Several reports additionally suggest that obstructive SDB may be associated with increased risk of parasomnias [69, 70].

In contrast to the prominent snoring and restlessness exhibited during the sleep of children with obstructive SDB, children with non-obstructive SDB tend to be quiet sleepers who only occasionally exhibit obvious respiratory symptoms at night. The lack of easily recognizable nighttime symptoms may lead to delays in diagnosis and treatment, particularly if sleepiness or other daytime symptoms are misattributed to other causes.

Children with obstructive SDB frequently exhibit sore throat, dry mouth, headache, or grogginess upon morning waking, although these symptoms are often transient and self-limited. Daytime mouth breathing is frequently seen in children with obstructive SDB, particularly when adenotonsillar hypertrophy is present [68]. Somnolence as a daytime symptom is seldom prominent in younger children unless their underlying SDB is severe [71]. When excessive somnolence is present, it is more often subtle or intermittent and sometimes evident only during sedentary activities such as riding in an automobile.

Neurobehavioral deficits represent the most common and most variable daytime symptoms of childhood SDB. Early reports of childhood OSA, documenting relatively severe cases, identified a high prevalence of school problems, behavioral disturbances, and hyperactivity among affected children [41, 72, 73]. More recent reports have suggested that even less severe varieties of childhood SDB may be associated with the same daytime symptoms. Children with SDB have been reported to have higher rates of parentally reported behavior problems, lower scores on tests of sustained attention, and lower scores on neuropsychometric assessments of executive function [74–77]. Subtle neurocognitive deficits have also been reported for children with primary snoring [54, 76, 77, 78].

There is also compelling evidence that SDB is overrepresented in children with ADHD and learning problems. Snoring was reported to be three times as frequent among children with attention-deficit/hyperactivity disorder (ADHD) compared to non-ADHD controls drawn from general pediatric and child psychiatry clinics, with higher snoring scores being associated with greater levels of inattention and hyperactivity [79]. Among 297 first-grade children performing poorly in school, 54 (18 %) were found to have evidence of nocturnal hypoxemia or hypercapnia during limited, home-based sleep studies [80].

The physical examination may be entirely normal in children with SDB, but predisposing anatomic features are often identified in affected children. Adenotonsillar enlargement, which is most common between 2 and 8 years of age, is associated with several physical findings [81]. In addition to visible tonsillar hypertrophy, “adenoid facies”—visible mouth breathing, pinched nose, and elongated facial appearance—is often observed. Narrow and high-arched hard palate, maxillary or mandibular hypoplasia, and macroglossia represent additional physical features which may predispose to SDB.

Obesity represents a risk factor for obstructive SDB in children, but this association is less robust in prepubertal children and more common during adolescence. In a group of 22 obese adolescents without sleep complaints, 10 (46 %) were reported to have abnormal PSGs [82]. Daytime sleepiness and AHI for this group both correlated with degree of obesity.

Some children with SDB—particularly infants and young children with severe airway obstruction—may present with decreased growth, low body weight, or failure-to-thrive [83, 84]. Treatment of these children may result in improved growth parameters and insulin-like growth factor 1 levels [85–87].

A variety of medical, genetic, and craniofacial conditions are associated with increased risk for SDB during childhood (Table 52.2). Detailed epidemiologic data are not available for most of these conditions; however, it is estimated that over 30 % of children with Down syndrome may exhibit SDB [88–91].

The long-term effects of childhood SDB are not well understood apart from limited data regarding cardiovascular effects. Reversible cor pulmonale and congestive heart failure have been reported in children with severe SDB [92–94]. In addition, electrocardiographic evidence of heart strain was reported for 3.3 % of 92 children referred for adenotonsillectomy [95]. The frequency with which hypertension affects children with SDB remains uncertain, but obesity and respiratory disturbance index (RDI) were found to be independently associated with increased blood pressure in a large cohort of 6- to 11-year-old children [96].

The public health impact of childhood SDB has also received little scrutiny. The extent to which affected children’s neurobehavioral symptoms limit their long-term academic achievement and adult socioeconomic status has not been studied, although the effect is suspected to be substantial for at least some children with SDB. Perhaps the most dramatic evidence illustrating the public health impact of childhood SDB is data reporting substantially higher health care utilization for children with OSA compared to controls, including higher rates for hospitalization, medication use, and emergency department visits [97]. Treatment of OSA children with adenotonsillectomy resulted in a reduction of total annual health care costs by one-third, compared to no change for controls and untreated OSA children [98].

The assessment of a child with suspected SDB should begin with a detailed history and physical exam as outlined in Table 52.3. The sleep history must thoroughly screen for the symptoms of SDB already discussed, but also should also include limited assessment for other sleep disorders whose clinical manifestations may mimic those of SDB. The medical history and physical examination should also include screening for other conditions that might cause or predispose toward SDB.

The history and physical examination are often supplemented by other assessment tools that are less expensive and more easily administered than PSG. A variety of standardized questionnaires have been developed with the goal of predicting whether sleepiness or SDB is likely to be present based on the presence and severity of specific symptoms. Questionnaires such as the Epworth Sleepiness Scale and Stanford Sleepiness scale are of limited usefulness in the assessment of children with suspected SDB due to limited validation data for this age-group and because the symptom that these measures assess—sleepiness—is often not obvious in affected children [99, 100]. Several questionnaires have been developed specifically for use in children, and limited validation data have been obtained regarding the use of the Pediatric Sleep Questionnaire and OSA-18 as screening tools for childhood SDB in clinical and research populations, [101–103]. The sensitivity of these tools for detection of SDB in clinical populations may be limited, however [104].

Although audio recordings of a child’s snoring have long been recommended as a “$5.00 sleep study,” this technique did not reliably distinguish primary snoring from SDB (AHI ≥ 5) in a blinded study assessing 29 snoring children [105]. Although home video recordings of a child’s sleep are an easy and usually inexpensive supplement to the clinical history during evaluation of a child with suspected SDB, the sensitivity and specificity of this technique have not been rigorously assessed.

Overnight oximetry is not recommended for primary assessment of children with suspected SDB [106]. Although the technique is unobtrusive, inexpensive, and well tolerated by most children, it reliably detects SDB only when arterial oxygen desaturation is prominent and does not identify those children whose SDB is characterized primarily by hypercapnia or obstruction without desaturation. Among 210 children with PSG-documented OSA (AHI ≥ 1), 120 (57 %) had normal or inconclusive nocturnal oximetry, confirming that normal oximetry cannot be used to rule out SDB in children [107].

Other non-PSG diagnostic tests are used on selective basis for children with suspected SDB, primarily in children with predisposing conditions and children presenting with severe symptoms. Anatomic obstruction of the upper airway is often demonstrable on radiographic or endoscopic assessment [108, 109]. Children suspected to have severe SDB or concurrent cardiorespiratory problems sometimes require chest X-rays, electrocardiogram (ECG), echocardiogram, and formal pulmonary function testing for complete assessment. Children with prominent learning or behavioral problems usually benefit from neuropsychometric testing and age-appropriate behavioral assessment and intervention, even when such symptoms are secondary to SDB.

Laboratory-based PSG represents the most sensitive and reliable tool presently available for the detection and classification of SDB in children. It is also a test that can be easily customized based on the clinical presentation of each patient. For example, 16-lead EEG can be added to standard PSG to provide increased sensitivity for the detection of interictal discharges or nocturnal seizures. Similarly, esophageal pressure monitoring (Pes) is sometimes performed for children felt to be at risk of UARS and other obstructive SDBs associated with prolonged partial airway obstruction.

PSG also has several limitations when used in children. The test itself is lengthy, expensive, and potentially stressful for both children and their parents. Laboratories having substantial experience in performing and interpreting PSG for children are limited in number and sometimes have lengthy waiting lists. Finally, routine PSG is not always sensitive for the detection of prolonged partial airway obstruction (e.g., UARS), but tools such as Pes make the study more invasive and are not available in all pediatric sleep laboratories.

PSG should be performed when a child’s symptoms, medical history, and physical examination suggest significant risk of clinically significant SDB. Isolated snoring—especially when it is only soft and occasional—is seldom a sufficient indication for PSG unless other symptoms or risk factors are also present. Practice guidelines that address respiratory indications for PSG in children have recently been issued by the American Academy of Sleep Medicine [106, 110]. Although it is common for otolaryngologists to perform adenotonsillectomy for suspected SDB without preoperative PSG, the safety and cost-effectiveness of this practice have been vigorously debated [111–113]. In addition, this practice can potentially result in children with noisy respiration, but no clinically significant SDB having surgery for which the overall health benefit is uncertain. Until validated practice guidelines address these issues, the author’s practice is to apply the same standards presently used for adults: to obtain a baseline PSG for patients with clinically suspected SDB before decisions are made with respect to treatment.

PSG in children is performed in a manner comparable to adult studies, utilizing frontal, central, and occipital EEG channels, electrooculogram (EOG), and ECG, as well as chin and limb electromyogram (EMG). Respiratory monitoring at minimum includes oral and nasal airflow, chest and abdominal movement, arterial oxygen saturation, and measures of carbon dioxide (end-tidal and/or transcutaneous). Most laboratories use nasal pressure transducers—thought to be sensitive to flow limitation and subtle decrements in flow—to assess nasal airflow, but since many children breathe through their mouths, thermistor, or thermocouple sensors to monitor oral or oral–nasal flow remain necessary as well.

Many sleep laboratories have the capability to supplement standard recording techniques with additional modes of respiratory monitoring when clinically indicated. Pes may be added to improve the sensitivity of PSG for partial airway obstruction and increased work of breathing. Pes is minimally invasive—requiring insertion of a thin catheter through the nasopharynx into the esophagus—but the technique has negligible impact upon children’s sleep and is the most sensitive method presently available for the measurement of increased upper airway resistance [114, 115]. Several less invasive techniques for the assessment of partial airway obstruction have been investigated for use in children, including pulse transit time, peripheral arterial tonometry, respiratory cycle–related EEG changes, and intercostal EMG monitoring, but none of these techniques are yet available for routine clinical use [116–121].

Scoring of sleep stages, movements, and arousals during pediatric PSG is performed in the same manner used for adult studies. Thirty-second epochs are reviewed and scored manually using standard criteria recently revised by the American Academy of Sleep Medicine [122–124]. Scoring of respiratory events for children differs slightly from that for adults. Whereas scoring of apneas, hypopneas, and respiratory effort-related arousals (RERAs) requires a minimum event duration of 10 s in adult studies, scoring of these events for pediatric studies requires only that the event last at least the duration of 2 missed breaths. This rule permits scoring of brief events in children whose baseline respiratory rate during sleep exceeds 12 breaths per minute.

Standards for interpreting the results of pediatric PSG have evolved over time. Normative PSG data for 50 healthy, asymptomatic children between 1 and 18 years of age were reported in 1992, as summarized in Table 52.4 [125]. Although these data have helped define the statistical limits of normality in healthy children, the point at which abnormal PSG parameters become associated with pathologic symptoms and outcomes has not been precisely determined. An AHI exceeding 5 events per hour is generally considered to be abnormal for adults. Although this threshold had previously been used for the diagnosis of OSA in children, the fact that children frequently demonstrate symptomatic OSA with AHIs below 5 has long suggested that a lower threshold is more appropriate for children [55, 56].

The current International Classification of Sleep Disorders, Third edition (ICSD–3), criteria for the diagnosis of pediatric OSA require the presence of one or more scorable obstructive events per hour of sleep or documentation of obstructive hypoventilation in conjunction with other clinical and PSG findings as outlined in Table 52.5 [52]. These revised criteria now permit many children who previously would have been classified as having nocturnal hypoventilation or UARS to be classified as having OSA. Further outcomes-based research remains necessary to better define the thresholds at which abnormal respiratory parameters on PSG become associated with clinically significant sequelae.

ICSD-3 did not specify separate pediatric criteria for non-obstructive forms of hypoventilation, with the partial exception of congenital central hypoventilation syndrome, a genetically mediated condition secondary to mutations of the PHOX2B gene which typically presents during infancy [52]. Current American Academy of Sleep Medicine, PSG, scoring rules define sleep-related hypoventilation for children as consisting of PCO₂ levels exceeding 50 mm Hg for greater than 25 % of total sleep time [123] (Fig. 52.5).

Diagnostic criteria for UARS in children were not addressed in ICSD-3, although proposed criteria have been published elsewhere [55].

Adenotonsillectomy is the most commonly administered treatment for obstructive SDB in children [126]. Although this procedure is thought to be effective in alleviating upper airway obstruction for many symptomatic children, the frequency with which adenotonsillectomy “cures” SDB in children is variable.

Early case series assessing response of childhood OSA to adenotonsillectomy reported surgical cure rates exceeding 70 %; however, these studies were limited by substantial variability in patient selection and frequent use of adult rather than pediatric PSG criteria to define OSA [127, 128].

More recent studies have reported variable and sometimes lower operative success rates. Among one group of 110 children with OSA (AHI > 1, mean AHI 22.3), complete postoperative normalization of the AHI was observed in only 27 subjects (25 %) [129]. In another large cohort of successively seen children with OSA (AHI > 1), 94 of 199 subjects (47 %) demonstrated abnormal polysomnograms postoperatively [130]. Among 397 children with OSA enrolled in a large, randomized, multicenter trial, PSG findings normalized in 79 % of children treated with adenotonsillectomy compared to 46 % of those assigned to watchful waiting and supportive care [131]. These and other studies suggest that children with high Mallampati scores, high preoperative AHI, obesity, or atypical anatomy of the upper airway may be at elevated risk of residual SDB following adenotonsillectomy.

In addition to improvements in obstructive symptoms and PSG parameters, adenotonsillectomy for children with SDB may also be associated with tangible improvements in behavior, school performance, and quality of life measures [132–134]. In one large cohort of children assessed before and after clinically indicated adenotonsillectomy, half of the children having ADHD prior to surgery no longer qualified for the diagnosis when reassessed one year postoperatively [117, 118].