Data on the natural history of snoring and obstructive sleep apnea (OSA) in children are limited. Small cohort studies have suggested that primary snoring (frequent snoring without OSA on polysomnography, PSG) does not usually progress to OSA as the child gets older (Topol & Brooks, 2001; Urschitz et al., 2004), and approximately half of primary-school-age habitual snorers stop snoring over the period of 1 year (Urschitz et al., 2004). This supports a strategy of watchful waiting for children with primary snoring. Findings of a randomized controlled trial (RCT) of adenotonsillectomy (i.e., surgical removal of the adenoids and tonsils, AT) compared to watchful waiting for OSA in a group of 423 children also support conservative management of mild OSA, given their finding of normalization of the polysomnographic findings over a 6-month period in 65% of children with mild OSA at baseline (Marcus et al., 2013).

Similar studies have not been carried out in children with ADHD, although one small study of children with ADHD and mild OSA showed no change in the apnea–hypopnea index (AHI) or in measures of behavior or quality of life over 6 months of follow-up in the nontreatment arm (n=14) in comparison to children treated with AT (n=25) or methylphenidate (n=27), leading the authors to suggest that treatment of even mild OSA in this group of children may be of benefit (Huang et al., 2007).

The most common cause of OSA in childhood is enlargement of the tonsils and adenoids. Tonsils and adenoids grow most quickly in the preschool years, with the adenoids being large in some children even in the latter part of the first year of life (Jeans, Fernando, & Maw, 1981; Nixon, Brouillette, Nixon, & Brouillette, 2002). Surgical removal of the adenoids and/or tonsils is therefore typically recommended as first line therapy for pediatric OSA (Marcus, Brooks, et al., 2012). Early reports investigating the effect of AT as an intervention for OSA suggested cure rates of 85%–95% (Nieminen, Tolonen, & Lopponen, 2000; Suen, Arnold, & Brooks, 1995). Subsequent studies, including a systematic review and meta-analysis, found that while there were significant improvements in OSA severity as defined by the obstructive AHI on polysomnography after AT, complete resolution of OSA (defined as an obstructive AHI <1 event/hour of sleep on polysomnography) occurs in only 66.3% (Friedman, Wilson, Lin, & Chang, 2009). Significant heterogeneity of the studies, particularly inclusion of varying numbers of obese and black children who are known to have lower cure rates (Marcus et al., 2013) and variable follow-up periods, makes generalization difficult, however a meta-analysis quoted a cure rate in uncomplicated patients (i.e., excluding those with obesity) of 73.8% (Friedman et al., 2009). The only RCT of adenotonsillectomy for pediatric OSA (the Childhood Adenotonsillectomy (CHAT) study) provides probably the best estimate of cure rate (Marcus et al., 2013). In that study, 464 children were randomized to early-adenotonsillectomy AT (n=226) or a strategy of watchful waiting with supportive care (n=227). Children with very severe OSA (an OAHI greater than 30/hour or SpO₂ less than 90% for 2% or more of the total sleep time) were excluded. Resolution of OSA by PSG criteria occurred after AT in 79% of children, with higher rates in those with very mild disease and lower rates in obese and in black children (Marcus et al., 2013). Improvement in OSA by PSG criteria at least appear to be relatively sustained, with follow-up periods of 6 months to 4 years in studies included in a more recent meta-analysis (Chinnadurai et al., 2017). Studies specifically assessing the duration of benefit are not available.

There is now a large body of prospective cohort studies demonstrating benefit of AT on the adverse neurobehavioral and quality of life consequences of OSA in children, summarized in several review articles (Garetz, 2008; Othman, Bee See, & Abdul Latif, 2016; Schechter & Section on Pediatric Pulmonology, 2002). Studies consistently report improvement in outcome measures such as quality of life, behavioral problems including hyperactivity and aggression, and neurocognitive skills including memory, attention, and school performance (Chervin et al., 2006; Gozal, 1998; Reckley, Fernandez-Salvador, & Camacho, 2018). Meta-analysis is hampered by the use of different testing tools, but a recent meta-analysis of five studies (375 unique patients) utilizing the Neuropsychological Developmental Assessment (NEPSY) before and after AT found a 7-point increase in NEPSY scores with a standardized mean difference of 0.53 (95% CI 0.35, 0.70), indicating a medium magnitude of effect (Song, Tolisano, Cable, & Camacho, 2016). Three studies (254 children) reported the effect of AT on IQ and found a heterogeneous effect, with younger children appearing to benefit more (Song et al., 2016).

In the CHAT study mentioned above, a total of 194 children in the early-AT group and 203 in the watchful-waiting group were included in the analysis of the primary outcome, the attention and executive-function score on the Developmental Neuropsychological Assessment (NEPSY) (Marcus et al., 2013). The NEPSY was measured at 7 months postrandomization by blinded assessors. Average scores on the NEPSY increased in both groups, with the difference between the groups favoring early AT but not reaching statistical significance: normative mean 100±15, change from baseline AT group 7.1±13.9, watchful-waiting group 5.1±13.4 (effect size 0.15). There was a small beneficial effect of AT on the caregiver-reported Conners’ Rating Scale (effect size 0.28), and teacher-reported data for this measure also showed significantly greater improvement in the early-adenotonsillectomy group (effect size 0.29) although it was not stated if teachers were blinded to the child’s surgery. The caregiver-reported Behavior Rating Inventory of Executive Function Global Executive Composite T-score, comprising summary measures of behavioral regulation and metacognition, was lower (indicating an improvement) at follow-up in the early-AT group (effect size 0.28), with a small increase in score in the watchful-waiting group (normative mean 50±10, change from baseline AT group −3.3±8.5, watchful-waiting group 0.4±8.8, effect size 0.28).

Neurocognitive and behavioral sequelae of OSA are poorly correlated with the severity of OSA as defined by PSG, with differences from controls noted even in children with primary snoring (Biggs, Nixon, & Horne, 2014). The CHAT study and cohort studies have pointed out a limited correlation between improvements in behavioral, cognitive, or quality of life measurements and PSG parameters before or after AT (Chervin et al., 2006; Rosen et al., 2015), with one study of children waiting for clinically indicated AT showing improvements in behavior and attention regardless of whether OSA was present at baseline by PSG criteria (Chervin et al., 2006). This suggests that current polysomnographic measures of OSA severity do not accurately assess the impact of obstructed breathing on sleep, and/or the relationship between severity of OSA and neuropsychological outcomes is nonlinear or mediated by other factors including individual susceptibility, parental, or school factors.

Given that daytime behavioral and cognitive deficits have been extensively described in children with OSA and show some improvement after AT, many parents may be led to hope and expect some improvement in some of the manifestations of ADHD following treatment for OSA. Indeed, in two studies at least 50% of children with ADHD who underwent AT no longer met the criteria for a diagnosis of ADHD after the surgery (Aksu, Gunel, Ozgur, Toka, & Basak, 2015; Chervin et al., 2006). Several prospective cohort studies from Iran (Ahmadi, Poorolajal, Masoomi, & Haghighi, 2016; Amiri et al., 2015) and Turkey (Aksu et al., 2015; Ayral et al., 2013; Somuk et al., 2016) have compared symptoms of ADHD using standardized tools before and after AT, although none had formal assessment of possible OSA and none included a comparison cohort of children with ADHD who did not undergo surgery. All showed improvements in ADHD symptomatology 3–6 months after surgery, with one finding a 10-point fall in the mean Conner’s Parent Rating Scale total score and 43/51 children experiencing some improvement in score (Ahmadi et al., 2016). Another showed improvements of a similar magnitude in all Conner’s Parent rating Scale T-scores in 53 subjects, from means of 67–77 down to 57–66, 6 months after adenotonsillectomy (Amiri et al., 2015), reflecting a clinically significant fall in symptom burden. Soylu et al. reported higher rates of both ADHD and oppositional defiant disorder (ODD) in a group of 40 children with symptoms of obstructed breathing during sleep and adenotonsillar hypertrophy compared to a comparison group of 35 children from ENT clinics without symptoms of obstructed breathing during sleep or adenotonsillar hypertrophy (Soylu et al., 2013). After AT, the symptoms scores for ADHD fell whereas there was no change in the symptom scores for ODD (Soylu et al., 2013), in contrast to another study that did find an improvement in ODD symptoms following AT (Dillon et al., 2007).

Examined from another perspective, attention and disruptive behavior disorders are more common in unselected children awaiting AT and the frequency of attention and disruptive behavior disorders and severity of associated psychiatric ratings drops after AT (Dillon et al., 2007). That study compared children waiting for AT (40 with and 38 without OSA by PSG criteria) with mean (SD) age 8.1 (1.8) years with a control group with mean (SD) age 9.3 (2.0) years who were recruited from other surgical clinics where they were seen for concerns unrelated to risk for obstructed breathing during sleep. Before surgery, 29 children awaiting AT (36.7%) had at least 1 disruptive behavior disorder (based on DSM-4), whereas only three (11.1%) of the controls were similarly diagnosed (p=.015). Remission rate for ADHD 1 year after AT was 50% among the 22 children diagnosed with ADHD at baseline (Dillon et al., 2007). Similar to the findings of the CHAT study, in this cohort of children awaiting AT, polysomnographic evidence of OSA did not predict psychopathology at baseline or improvement at follow-up, with children without OSA by PSG criteria also experiencing substantial improvements in behavior and attention.

However, AT is not without risk. Death or serious morbidity does occur, and is more likely in children with OSA (Cote, Posner, & Domino, 2014). Postoperative hemorrhage occurs in around 2% (De Luca Canto et al., 2015) and can require a return to the operating room.

In one retrospective case-control study of 91 children with postoperative hemorrhage and 151 controls who underwent tonsillectomy on the same day by the same surgeon, children with a history of ADHD were found to have an 8.7 times greater risk of experiencing postoperative hemorrhage (95% confidence intervals 1.4–53.6, p=.03) (Spektor, Saint-Victor, Kay, & Mandell, 2016). The authors postulated that children with ADHD might have increased physical activity and decreased compliance with postoperative instructions postoperatively, but this was not determined in the study.

Given the fact that upper airway inflammation is present in children with OSA (Goldbart, Krishna, Li, Serpero, & Gozal, 2006), and hypertrophied lymphoid tissues of the tonsils and adenoids in children with OSA express increased glucocorticoid receptor alpha and beta and leukotriene C4 synthase and receptors 1 and 2 (Dayyat et al., 2009; Goldbart et al., 2004; Kaditis et al., 2008), corticosteroids and oral leukotriene receptor antagonists have been trialed as therapy for OSA. Intranasal corticosteroids and montelukast have shown favorable results on these targets in vitro and when trialed in children with OSA in small monotherapy RCTs (largest study, n=62) compared to placebo (Brouillette et al., 2001; Goldbart, Goldman, Veling, & Gozal, 2005; Goldbart, Greenberg-Dotan, & Tal, 2012; Kheirandish-Gozal, Bandla, & Gozal, 2016; Kheirandish-Gozal & Gozal, 2008). Treatment period in these studies was usually 6 weeks for intranasal corticosteroids and 12–16 weeks for montelukast. Most trials focused on children with mild OSA as defined by PSG and demonstrated significant improvements in AHI. Further, combination therapy with these two agents together has described in a large retrospective study in children with mild OSA (Kheirandish-Gozal, Bhattacharjee, Bandla, & Gozal, 2014), with similarly positive results. These studies did not measure the effect of treatment on outcomes such as quality of life and cognitive and behavioral effects, and no studies of medical treatment have focused specifically on children with ADHD.

Continuous positive airway pressure (CPAP) treatment involves delivery of a continuous flow of air via a facial mask, which acts as a splint to the upper airway, preventing collapse. It is used as a treatment in children when AT or other therapies have not resulted in sufficient clinical improvement or in cases when surgery is not indicated. Children who are obese, or those with craniofacial abnormalities or neuromuscular disorders are the most common candidates for CPAP treatment (Marcus, Radcliffe, et al., 2012; Nixon, Mihai, Verginis, & Davey, 2011).

Few studies have examined the benefit of CPAP on daytime functioning in children. A single study of 52 children aged 2–16 years examined attention deficits, daytime sleepiness, behavior, and caregiver- and child-reported quality of life after 3 months of CPAP treatment (Marcus, Radcliffe, et al., 2012). Although the subjects did not have diagnosed ADHD, statistically significant improvements in ADHD symptoms were seen, using the Conner’s Abbreviated Symptom Questionnaire and the Attention Problems subscale of the CBCL, although the magnitude of the differences appeared small. Larger improvements were seen in daytime sleepiness and improved quality of life. Adherence to CPAP may be low however, with reported average use varying from 3 to more than 8 hours per night (Marcus, Radcliffe, et al., 2012; Nixon et al., 2011; O’Donnell, Bjornson, Bohn, & Kirk, 2006; Ramirez et al., 2013). In adults at least, longer usage time per night is related to benefit in terms of daytime functioning (Pepin et al., 1999). The relationship between adherence to CPAP in children is complex to tease out, with the reducing sleep requirements as children get older likely to influence the hours of CPAP use that might result in improvements in functioning. However, the only study of daytime cognitive and behavioral functioning in children on CPAP, described above, demonstrated improvements with even low amounts of CPAP usage (mean 3 hours/night) (Marcus, Radcliffe, et al., 2012). There are no reported studies on the benefits or otherwise of CPAP in children with ADHD.

Given the high prevalence of Restless Legs Syndrome (RLS) and Periodic Limb Movement Disorder (PLMD) in children with ADHD, several studies have addressed the extent to which treatment of these disorders affects symptomatology in children with ADHD. Advice on healthy sleep patterns and getting adequate amounts of sleep have been addressed elsewhere (see Chapter 5) but are worthy of mention here as a specific intervention for RLS. Insufficient sleep and an irregular sleep schedule can exacerbate RLS (Picchietti & Picchietti, 2010), and reducing caffeine and alcohol intake is usually recommended as an adjunct to therapy particularly in adolescents, although recent evidence from a large cohort study in adults would suggest that these factors are less important than smoking and being overweight (Batool-Anwar et al., 2016). Physical exercise reduces the risk of RLS in adults (Batool-Anwar et al., 2016) and a 12-week intervention involving lower limb resistance exercises and treadmill walking improved the symptoms of RLS in a small RCT (Aukerman et al., 2006). In the latter study, participants were instructed to exercise three times per week, including treadmill walking for 30 minutes and 8–12 repetitions of 6 different leg strength exercises at each session. With this, RLS symptom severity reduced by 39% over 6 weeks and was maintained at 12 weeks (n=11), whereas no change was seen in controls (n=17) (Aukerman et al., 2006). The extent, nature, and timing of exercise and its effects on RLS have not been specifically studied. However for RLS, as well as the other benefits of exercise on health, regular exercise and good sleep hygiene should be advised in children with RLS.

Iron deficiency is implicated in the pathophysiology of RLS and PLMD through its role in dopamine neurotransmission, as well as myelin synthesis (Picchietti & Picchietti, 2010). In one retrospective study of 75 children treated with iron for RLS, approximately 80% had improvement or resolution of their symptoms (Amos et al., 2014). Treatment with oral iron is usually recommended for children with RLS with serum ferritin is below 50 μg/L, based on adult data and pediatric studies showing improvement in RLS symptoms with a rise in ferritin above this threshold (Amos et al., 2014; Dye, Jain, & Simakajornboon, 2017; Konofal et al., 2008; Simakajornboon et al., 2003). Although these children may have ferritin at the low end of the normal range before treatment, iron supplementation is based on recognition that serum ferritin above 50–100 μg/L is required to replenish tissue stores of iron, including in the brain (Picchietti, 2007).

ADHD in children is also associated with reduced peripheral serum ferritin (Tseng et al., 2018), and thus iron supplementation has been studied as treatment for both ADHD and RLS/PLMD. It should be pointed out however that RLS/PLMD is not necessarily the link between low iron stores and ADHD symptomatology, given iron’s wide-ranging and key role in neurotransmission (Bakoyiannis et al., 2015). Having said that, a small case series found lower ferritin levels in children with ADHD with RLS compared to children with ADHD without RLS (Konofal et al., 2007). The independent impact and interactions of RLS/PLMD and low iron stores on ADHD symptoms need to be further elucidated. Nonetheless, sleep–wake transition disorders (as defined by questionnaire), including abnormal sleep movements, are more common in children with ADHD with low serum ferritin compared to other children with ADHD (Cortese, Konofal, Bernardina, Mouren, & Lecendreux, 2009), and therefore make a reasonable target for treatment efforts. Sensitivity to stimulant therapy has been also postulated to be influenced by iron status, with one study of 52 children (83% male, mean age 10 years) showing an inverse association between serum ferritin and the weight-adjusted dose of amphetamine necessary to reach an optimal clinical response (Calarge, Farmer, DiSilvestro, & Arnold, 2010; Turner, Xie, Zimmerman, & Calarge, 2012). This adds another dimension to the complex interplay between iron metabolism and ADHD.

Despite the frequency of comorbid RLS and ADHD, only a few small studies have evaluated the effect of iron treatment in children with ADHD specifically. Konofal et al. evaluated the effect of a 12-week course of iron treatment on 23 children aged 5–8 years with ADHD and serum ferritin below 30 μg/L without anemia (Konofal et al., 2008) in a double-blind, placebo-controlled randomized trial (3:1 randomized to iron supplementation vs placebo). They found a significant decrease in the ADHD Rating Scale (mean (SD) change −3.0 (5.7) in the placebo group and −10.2 (14.0) in the iron treatment group (equivalent to an effect size (Hedges’ g) of 0.6). The Clinical Global Impression-Severity scale score was not different between the groups at baseline, but 4 of 17 patients treated with iron were rated as very much or much improved compared to none of the placebo group. Improvements in the Conners’ Parent or Teacher Rating Scales did not reach statistical significance (Konofal et al., 2008). A previous open label trial of oral iron supplementation in 14 boys aged 7–11 years with ADHD had found a significant decrease in the Conners’ Parent Rating Scale (effect size 0.99) but not the Conners’ Teacher Rating Scale (Sever, Ashkenazi, Tyano, & Weizman, 1997). PSG was not performed in either of these studies and so the degree to which the presence of PLMs and their improvement with iron therapy influenced the findings cannot be determined. In adults at least, the self-reported severity of ADHD symptoms has been associated with symptoms of RLS and PLMD (and OSA) (Vogel et al., 2017). In one small study in children, higher levels of ADHD symptomatology in those with RLS compared to those without did not reach statistical significance (Konofal et al., 2007). These studies suggest that RLS/PLMD may exacerbate ADHD, but the specific link between successful treatment of RLS/PLMD with iron and improvement in ADHD symptomatology in children has not yet been studied.

Medication options for RLS and PLMD in adults have included clonidine, clonazepam, gabapentin, and dopaminergic agents (pramipexole, ropinirole, or less commonly carbidopa/levodopa and pergolide) (Picchietti & Picchietti, 2008, 2010). In a retrospective study of treatment in 97 children aged 5–18 years with RLS, treatments used were (in order of frequency): iron (65%), sleep hygiene advice (25%), melatonin (24%), gabapentin (13%), clonidine (6%), and dopaminergic agents (6%) (Amos et al., 2014).

Although melatonin is described in the treatment of sleep onset insomnia in children with ADHD (Miano et al., 2016), its specific effect on RLS or PLMD has not been reported. One study of eight adults suggested a worsening of RLS symptoms with the use of melatonin (Whittom et al., 2010), and so this potential should be borne in mind when utilizing melatonin for sleep onset problems in children with both ADHD and RLS.

The efficacy of gabapentin enacarbil, a slow-release pro-drug of gabapentin, for RLS in adults has been well described (Garcia-Borreguero et al., 2002; Happe, Sauter, Klosch, Saletu, & Zeitlhofer, 2003; Winkelman et al., 2016). Evidence for gabapentin is less compelling, and guidelines express concern about potential side effects including sedation, dizziness, vision changes, and suicidal behavior and ideation (Aurora et al., 2012). Despite being mentioned in several reviews and retrospective case series (Amos et al., 2014), specific efficacy for RLS or PLMD in children has not been studied.

Clonidine has shown some efficacy for treatment of sleep disturbance in children with ADHD (see Chapter 6) (Prince, Wilens, Biederman, Spencer, & Wozniak, 1996). In relation to its efficacy for the treatment of RLS/PLMD specifically, it has been shown to reduce the symptoms of RLS in adults (Wagner et al., 1996) but its effects on RLS or PLMD have not been directly reported in children.

The dopaminergic agents are used in treatment of Parkinson’s disease and have been evaluated in adults with RLS/PLMD (Garcia-Borreguero et al., 2013). In one case series of seven children with ADHD and RLS and/or PLMD treated with levodopa or pergolide, improvements in RLS symptoms and PLMs on PSG were associated with improvements in behavior, with a significant fall in Conners’ Parent Rating Scale (from 15.1 to 6.3; p<.04) and the CBCL (from 67.7 to 56.8; p<.05) (Walters et al., 2000). Stimulant therapy had previously not been successful in four of the seven children due to inefficacy or intolerable side effects. After treatment with dopaminergic monotherapy, three of the seven children no longer met the DSM-4 criteria for a diagnosis of ADHD, raising the possibility that the daytime features of ADHD were a manifestation of sleep disturbance caused by RLS and/or PLMD. The same study group went on to perform a randomized double-blind controlled trial of carbidopa/L-DOPA in children who were not on stimulant therapy and had either ADHD alone or ADHD with RLS or PLMD (England et al., 2011). Fifty-three children aged 7–12 years met screening criteria, but 10 declined to participate, 6 were excluded for intellectual dysfunction or OSA on PSG, and 2 did not complete the study, leaving 35 who completed the study protocol (13 with ADHD alone and 22 with ADHD with RLS or PLMD). While improvements in RLS and PLMD were documented in all of the children on L-DOPA who had the effect of treatment on these outcomes objectively measured (n=8), the study did not find any significant improvement of ADHD symptoms, sleep parameters, or neuropsychometric measures in ADHD patients treated with L-DOPA as compared to those given placebo (England et al., 2011). A single case report of a 6-year old with ADHD and RLS has described treatment with ropinirole resulting in improvements in RLS symptoms and daytime behavior (Konofal, Arnulf, Lecendreux, & Mouren, 2005). Although these reports highlight the benefits of effective management of RLS and PLMD in children with ADHD, the safety of dopaminergic agents in children has not been established, with potential for significant side effects, particularly in relation to reports in adults of increased impulsivity, impaired learning, and augmentation of RLS symptoms, as well as potential long term effects on the functioning of the dopaminergic system (Harris, 2009).

Given the cross-over in biochemical pathophysiology, phenotype, and treatment of ADHD and narcolepsy (Miano et al., 2016), specific consideration needs to be given to the effect of treatment on the symptoms of both disorders. In one study of children under 12 years receiving care in one of the national referral centers for narcolepsy in France, symptoms of ADHD were assessed using the ADHD Rating Scale, measured at various points in relation to the diagnosis or treatment of narcolepsy (Lecendreux et al., 2015). ADHD-RS scores for children with formally diagnosed narcolepsy (N=78) were compared with healthy controls recruited from participant hospitals or the community (N=63). ADHD-RS scores were higher in children with narcolepsy than controls regardless of treatment for narcolepsy, with clinically significant levels of ADHD symptoms found in 4.8% of controls compared with 35.3% in patients with narcolepsy without cataplexy (95% confidence interval for the difference in percentage of 11.2–52.6) and 19.7% in patients with narcolepsy with cataplexy (95% CI 3.7–27.3). Subjectively rated excessive daytime sleepiness, insomnia and fatigue were all associated with the level of ADHD symptoms. Children on low dose methylphenidate (plus modafinil in seven of eight patients) had higher ADHD-RS scores than patients with narcolepsy on no treatment, whereas children on high dose (≥0.52 mg/kg per day) methylphenidate (plus modafinil in 8 of 10 patients) had ADHD-RS scores that were not different from those with narcolepsy not on treatment (Lecendreux et al., 2015). Although highlighting the cross-over in symptoms between narcolepsy and ADHD, this study was cross-sectional and the effectiveness of treatment for narcolepsy on ADHD symptoms requires further investigation using longitudinal intervention studies.