Diagnosis of sleep-disordered breathing (SDB) is most accurately obtained with a nocturnal polysomnogram. However, limitations on availability make alternative screening tools necessary. Nocturnal oximetry studies or nap polysomnography can be useful if positive; however, further testing is necessary to if these tests are negative. History and physical examination have insufficient sensitivity and specificity for diagnosingpediatric SDB. Adenotonsillectomy remains first-line therapy for pediatric SDB and obstructive sleep apnea (OSA). Additional study of limited therapies for mild OSA are necessary to determine if these are reasonable primary methods of treatment or if they should be reserved for children with persistent OSA.
- •
History and physical examination are not sufficient to differentiate between snoring and obstructive sleep apnea (OSA).
- •
Nocturnal in-laboratory polysomnography remains the gold standard for diagnosis of OSA.
- •
Adenotonsillectomy is the recommended initial treatment for OSA and sleep-disordered breathing for healthy children, even in those with risk factors associated with persistent pediatric OSA such as obesity.
- •
Efficacy data for partial tonsillectomy are limited despite multiple studies showing reduced postoperative bleeding and recovery time.
- •
Bariatric surgery is an option for extremely obese adolescents.
- •
Medical treatment may be a good option for mild OSA, either primary or persistent after adenotonsillectomy; although, more data are necessary.
- •
CPAP is an effective therapy in children, but similar to adults, adherence is a significant issue, and there may be facial side effects in children with long-term use.
Overview
Although obstructive sleep apnea (OSA) was described in literature as early as the 1700s, the first report of pediatric OSA was not published until 1976 in a case series of 8 children. Since that time, much has been learned about OSA in children; however, there remain significant gaps in the understanding and evidence for the diagnosis, treatment, and management of pediatric OSA.
Recent prospective general pediatric population studies using polysomnography (PSG) for diagnosis have found that the prevalence of pediatric OSA ranges from 1.2% to 5.7%. These estimates are in keeping with previous estimates and similar to the estimates of symptomatic OSA in adults. In addition, estimates of sleep-disordered breathing (SDB), which includes both OSA and snoring, are around 12%.
Epidemiology
The epidemiology of OSA has also been investigated in a systematic review that suggests that there is a male predominance in pediatric OSA, especially in older boys ; although, several large studies have found no difference in the prevalence by sex. It is also suggested that age is a mediator of this relationship, as those studies including older children were more likely to find an increased rate of SDB in boys. In addition, minority status has been associated with increased prevalence of SDB, with multiple studies showing that black children are more likely to have SDB than white children. Finally, comorbid factors such as obesity, craniofacial deformity, genetic syndrome status, and metabolic disease have also been associated with an increased incidence of SDB in children and adults ( Table 1 ).
| Adenotonsillar Hypertrophy | Increased Airway Resistance |
| Obesity | Fatty infiltration of airway, abnormal ventilatory control |
| Race (African American) | Craniofacial structure, socioeconomic |
| Gender (male) | Slight male predominance in prepubertal children, which increases markedly after puberty |
| Prematurity | Neurologic impairment, adverse craniofacial growth, abnormal ventilatory control |
| Craniofacial dysmorphology | Increased airway resistance |
| Neurologic disorders | Abnormal motor control of the upper airway |
| Nasal/pharyngeal inflammation | Allergy or infection increasing airway resistance |
| Socioeconomic/environmental | Neighborhood disadvantage, passive cigarette smoke, indoor allergens, sleep quality (noise, stress) |
| Family history of OSAS | Heritable craniofacial structure, neuromuscular compensation, arousal threshold, ventilatory control |
Sequelae to Pediatric Sleep Apnea
Most importantly, untreated OSA is associated with a number of sequelae, including cognitive deficits, hyperactivity, cardiovascular consequences, and inflammation. Studies of neuropsychiatric function in children with SDB have almost universally found cognitive deficits in these children. Two representative prospective studies characterized these deficits and found lower general intelligence, learning, and memory results; decreased language and verbal skills; and diminished visual and auditory attention. In addition, school performance has been shown to improve after treatment of SDB with adenotonsillectomy.
Behavioral issues, and especially hyperactivity, have also been widely associated with SDB and OSA. In these prospective studies, behavioral issues included attention-deficit/hyperactivity disorder (ADHD, as diagnosed by psychiatric interview), depression, daytime sleepiness (by subjective report and objective multiple sleep latency testing), aggression, and oppositional and social problems. In addition, both behavioral issues and cognitive deficits were improved after adenotonsillectomy and were shown to have sustained improvement in a study that followed patients for 1 year after surgery.
Comorbid medical conditions with pediatric sleep apnea
While the association between OSA and cardiovascular disease is well described in the adult literature, this association is not as well understood in children. Cardiac abnormalities of both the right and left ventricles have been reported and have been associated with postsurgical respiratory complications. In addition, blood pressure elevations, both daytime and nighttime, have been correlated with OSA in children. Autonomic dysfunction has also been reported in children with OSA during both wake and sleep.
There is increasing recognition of the relationship between obesity and SDB and the risk of obesity in children after adenotonsillectomy. The traditional picture of children with OSA often included a population of children with failure to thrive. In these underweight children, the recognition that there was often a growth spurt after adenotonsillectomy was reported as a positive outcome of surgery. More recently, this same increase in weight after adenotonsillectomy has led to growing concern that normal weight and obese children are becoming overweight or more obese after surgery.
Dissimilarities in child and adult sleep apnea
Differences between adult and pediatric SDB also include significant dissimilarities in respiratory events. As children are more likely to have hypopneas than discrete apneas, arousals are uncommon, and desaturations are less common than in adults. Children may also have long periods of flow-limited breathing that do not meet the definition of a respiratory event, but reflect partially obstructed breathing. In addition, children are more likely to have rapid eye movement (REM)-only disease than adults, which makes nap studies, either oximetry or polysomnography, problematic, as it will often not capture any significant REM sleep. Several large population studies have been performed to determine normative data for pediatric sleep and are reported in composite in Table 2 .
| Sleep | |
| EEG arousal index (per h TST) | 9 ± 3 |
| Sleep efficiency (%) | 89 ± 7 |
| Stage 1 (% TST) | 5 ± 3 |
| Stage 2 (% TST) | 42 ± 8 |
| Slow wave sleep (% TST) | 26 ± 8 |
| REM sleep (% TST) | 20 ± 5 |
| REM cycles | 4 ± 1 |
| Periodic leg movement index (per h TST) | 1 ± 1 |
| Respiratory | |
| Obstructive apnea index (per h TST) | 0.0 ± 0.1 |
| Obstructive apnea/hypopnea index (per h TST) | 0.1 ± 0.1 |
| Central apnea index (per h TST) | 0.5 ± 0.5 |
| P ET CO 2 ≥50 mm Hg (% TST) | 2.8 ± 11.3 |
| Peak P ET CO 2 (mm Hg) | 46 ± 3 |
| S o O 2 >95% (% TST) | 99.6 ± 1 |
| S o O 2 90%–95% (% TST) | 0.4 ± 1 |
| S o O 2 <90% (% TST) | 0.05 ± 0.2 |
| Desaturation index (≥4%/h TST) | 0.4 ± 0.8 |
| S o O 2 Nadir (%) | 93 ± 4 |
Evidence-based clinical assessment
Clinical History
Multiple studies have looked at the effectiveness of history and physical examination in differentiating OSA from snoring and found that neither is effective in reliably separating the two. An evaluation of clinical history in 480 patients who underwent concomitant home 16-channel polysomnography found that clinical symptoms, either solo or in combination, had a low sensitivity for diagnosis of OSA. A smaller study of 83 patients with sleep study data found a statistically significant increase in the risk ratio (RR) for OSA when parents report the need to shake their children (RR = 1.94), the presence of witnessed apneas (RR = 1.63), and witnessed struggle to breathe (RR = 1.53), but none of these factors was found to be sufficient to predict OSA. Additional measures, such as waist circumference and body mass index (BMI) z-score, have also been found to have a poor clinical correlation with symptoms of SDB.
The use of surveys to screen for OSA has also been investigated with mixed results. The pediatric sleep questionnaire (PSQ) survey has shown reasonable sensitivity (78%) and specificity (72%) when compared with sleep study results and has been recommended as a screening tool with the caveat that subsequent definitive diagnostic testing would still be recommended if positive. The OSA-18, a validated quality-of-life survey frequently used in pediatric otolaryngology, was also evaluated as a screening tool for moderate-to-severe OSA in 334 children and found to have poor sensitivity at 40%, with a negative predictive value of only 73%.
Polysomnography
The gold standard for diagnosis is an attended, in-laboratory, nighttime polysomnogram. The practice parameter for respiratory PSG indications published by the American Academy of Sleep Medicine (AASM) in 2011 reviewed 45 studies with PSG data before and after OSA treatment and found test–retest validity or improvement in PSG parameters to be robust in all 45 studies.
Several studies have investigated the reliability of a single pediatric nocturnal sleep study to determine if PSG has good test–retest reliability/consistency. This is important, as adult sleep study testing has found a first night effect, in which adults do not sleep as well during their first sleep study; thus sleep efficiency and architecture may underestimate OSA severity. In the pediatric studies, however, there were only minimal differences in measures of OSA severity, suggesting that a single night of PSG evaluation is adequate. However, while respiratory parameters were fairly constant, sleep architecture was significantly different between nights.
Additional validity measures have also been evaluated, including construct validity, a measure of whether the PSG is a true reflection of SDB, and convergent validity, which looks for evidence that 2 tests measuring the same problem move in the same direction (ie, the PSG and tests of outcomes like sleepiness and cognition). These evaluations have found that many measures change consistently when OSA is present, like sleepiness, neurocognitive outcomes, and quality of life. Despite this, disease severity often does not correspond with degree of change in these outcomes.
Home sleep studies, also known as ambulatory PSGs, are approved for use in adults with high pretest probability of OSA, but they are not currently recommended for children by any of the current guidelines. There are few studies looking at home studies in children. The largest looked at 850 children between ages 8 and 11 years of age who underwent 4-channel ambulatory PSG, which included oximetry, heart rate, body position, and inductance plesmography. In this case, 94% of the home studies were found to be technically satisfactory. The same study also had 55 patients who underwent in-laboratory 16-channel sleep studies; although, little is reported about this group. Based on these 55 patients, the study investigators reported that the home studies were sensitive at 88% and highly specific at 98% for an apnea–hypopnea index (AHI) greater than 5 events per hour as determined on the in-laboratory studies. Several other studies have investigated home studies in the sleep laboratory or looked at a reduced number of channels from an in-laboratory study. The first of these studies was performed on only 12 children ages 3 to 6 years and found to have poor specificity at 0%. The second study looked at 30 children each with normal, mild/moderate OSA and severe OSA using only the oximetry and dual respiratory inductance plethysmography (RIP) bands and determined that there was correct classification of SDB category in 83% with a false-negative rate of only 8%. It is unclear from the current data whether studies are feasible in children between 2 and 7 years of age, when tonsillar hypertrophy is most prominent.
In contrast, investigations of nap PSG have found that sensitivity ranged from 69% to 75%, while specificity ranged from 60% to 100%. Because of this, the AASM does not recommend nap polysomnography as the sole method of diagnosis in children, as they are not as reliable as in-laboratory PSGs. In addition, nap studies have been found to underestimate OSA severity, especially as they often do not include any significant REM sleep. For this reason, the American Academy of Pediatrics (AAP) suggests that nap polysomnography can be useful if it is positive for OSA, even if severity is inaccurate, but that negative studies should prompt a nocturnal PSG.
Alternate Measures
Cardiovascular monitoring, including heart rate variability and pulse transit time, has also been evaluated for diagnosis of pediatric OSA with variable suboptimal rates of sensitivity and specificity. One particular technique, peripheral arterial tonometry, appeared promising as a method to identify electroencephalographic arousals with sensitivity of 95% but was found to have poor specificity at 35%. Refinement of these techniques is necessary before they can be considered for screening or diagnosis.
Audiotaping or videotaping, nocturnal oximetry, and nap polysomnography are all noted to be useful if they are positive for witnessed apneas, desaturations or obstructive events, but to have poor predictive value when the results are negative, leading to the suggestion that those who screen negative by these methods be referred for in-laboratory PSG. In addition, determination of disease severity is limited or unable to be quantified by these techniques.
An initial study of nap oximetry found it to be an accurate screening tool for OSA in children in whom it was positive, leading the AAP to state that positive nocturnal oximetry could be used to diagnose OSA. However, the practice guidelines went on to recommend that children with negative findings on this screening undergo further testing with a full PSG. Since that time, 2 additional studies have compared oximetry and PSG results. The first looked at 230 patients and concluded that nocturnal oximetry could be used to estimate disease severity in patients who screened positive, but 78% of the children in this study had normal or indeterminate studies and therefore went on to have a full PSG.
Evidence-based clinical assessment
Clinical History
Multiple studies have looked at the effectiveness of history and physical examination in differentiating OSA from snoring and found that neither is effective in reliably separating the two. An evaluation of clinical history in 480 patients who underwent concomitant home 16-channel polysomnography found that clinical symptoms, either solo or in combination, had a low sensitivity for diagnosis of OSA. A smaller study of 83 patients with sleep study data found a statistically significant increase in the risk ratio (RR) for OSA when parents report the need to shake their children (RR = 1.94), the presence of witnessed apneas (RR = 1.63), and witnessed struggle to breathe (RR = 1.53), but none of these factors was found to be sufficient to predict OSA. Additional measures, such as waist circumference and body mass index (BMI) z-score, have also been found to have a poor clinical correlation with symptoms of SDB.
The use of surveys to screen for OSA has also been investigated with mixed results. The pediatric sleep questionnaire (PSQ) survey has shown reasonable sensitivity (78%) and specificity (72%) when compared with sleep study results and has been recommended as a screening tool with the caveat that subsequent definitive diagnostic testing would still be recommended if positive. The OSA-18, a validated quality-of-life survey frequently used in pediatric otolaryngology, was also evaluated as a screening tool for moderate-to-severe OSA in 334 children and found to have poor sensitivity at 40%, with a negative predictive value of only 73%.
Polysomnography
The gold standard for diagnosis is an attended, in-laboratory, nighttime polysomnogram. The practice parameter for respiratory PSG indications published by the American Academy of Sleep Medicine (AASM) in 2011 reviewed 45 studies with PSG data before and after OSA treatment and found test–retest validity or improvement in PSG parameters to be robust in all 45 studies.
Several studies have investigated the reliability of a single pediatric nocturnal sleep study to determine if PSG has good test–retest reliability/consistency. This is important, as adult sleep study testing has found a first night effect, in which adults do not sleep as well during their first sleep study; thus sleep efficiency and architecture may underestimate OSA severity. In the pediatric studies, however, there were only minimal differences in measures of OSA severity, suggesting that a single night of PSG evaluation is adequate. However, while respiratory parameters were fairly constant, sleep architecture was significantly different between nights.
Additional validity measures have also been evaluated, including construct validity, a measure of whether the PSG is a true reflection of SDB, and convergent validity, which looks for evidence that 2 tests measuring the same problem move in the same direction (ie, the PSG and tests of outcomes like sleepiness and cognition). These evaluations have found that many measures change consistently when OSA is present, like sleepiness, neurocognitive outcomes, and quality of life. Despite this, disease severity often does not correspond with degree of change in these outcomes.
Home sleep studies, also known as ambulatory PSGs, are approved for use in adults with high pretest probability of OSA, but they are not currently recommended for children by any of the current guidelines. There are few studies looking at home studies in children. The largest looked at 850 children between ages 8 and 11 years of age who underwent 4-channel ambulatory PSG, which included oximetry, heart rate, body position, and inductance plesmography. In this case, 94% of the home studies were found to be technically satisfactory. The same study also had 55 patients who underwent in-laboratory 16-channel sleep studies; although, little is reported about this group. Based on these 55 patients, the study investigators reported that the home studies were sensitive at 88% and highly specific at 98% for an apnea–hypopnea index (AHI) greater than 5 events per hour as determined on the in-laboratory studies. Several other studies have investigated home studies in the sleep laboratory or looked at a reduced number of channels from an in-laboratory study. The first of these studies was performed on only 12 children ages 3 to 6 years and found to have poor specificity at 0%. The second study looked at 30 children each with normal, mild/moderate OSA and severe OSA using only the oximetry and dual respiratory inductance plethysmography (RIP) bands and determined that there was correct classification of SDB category in 83% with a false-negative rate of only 8%. It is unclear from the current data whether studies are feasible in children between 2 and 7 years of age, when tonsillar hypertrophy is most prominent.
In contrast, investigations of nap PSG have found that sensitivity ranged from 69% to 75%, while specificity ranged from 60% to 100%. Because of this, the AASM does not recommend nap polysomnography as the sole method of diagnosis in children, as they are not as reliable as in-laboratory PSGs. In addition, nap studies have been found to underestimate OSA severity, especially as they often do not include any significant REM sleep. For this reason, the American Academy of Pediatrics (AAP) suggests that nap polysomnography can be useful if it is positive for OSA, even if severity is inaccurate, but that negative studies should prompt a nocturnal PSG.
Alternate Measures
Cardiovascular monitoring, including heart rate variability and pulse transit time, has also been evaluated for diagnosis of pediatric OSA with variable suboptimal rates of sensitivity and specificity. One particular technique, peripheral arterial tonometry, appeared promising as a method to identify electroencephalographic arousals with sensitivity of 95% but was found to have poor specificity at 35%. Refinement of these techniques is necessary before they can be considered for screening or diagnosis.
Audiotaping or videotaping, nocturnal oximetry, and nap polysomnography are all noted to be useful if they are positive for witnessed apneas, desaturations or obstructive events, but to have poor predictive value when the results are negative, leading to the suggestion that those who screen negative by these methods be referred for in-laboratory PSG. In addition, determination of disease severity is limited or unable to be quantified by these techniques.
An initial study of nap oximetry found it to be an accurate screening tool for OSA in children in whom it was positive, leading the AAP to state that positive nocturnal oximetry could be used to diagnose OSA. However, the practice guidelines went on to recommend that children with negative findings on this screening undergo further testing with a full PSG. Since that time, 2 additional studies have compared oximetry and PSG results. The first looked at 230 patients and concluded that nocturnal oximetry could be used to estimate disease severity in patients who screened positive, but 78% of the children in this study had normal or indeterminate studies and therefore went on to have a full PSG.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
