Obstructive sleep apnea (OSA) and attention-deficit/hyperactivity disorder (ADHD) are two common, severe disorders in children. Unfortunately, pediatric OSA is closely associated with ADHD, and both diseases can cause cognitive impairment, behavior problems, and low academic performance. OSA can damage the brain and induce autonomic dysfunction, and then cause cognitive, behavioral, and quality-of-life problems. The presence of ADHD can further exacerbate these adverse effects of OSA. Therefore, the identification of robust biomarkers of OSA and ADHD is a key imperative to facilitate early identification of the pathological features and mechanisms and to optimize the treatment of OSA and ADHD for the pediatric population. Diffusion MRI of the brain is one of the most widely used technology for assessment of brain tissue integrity and heart rate variability is one of the most widely used measurements of autonomic function. However, the effects of ADHD and adenotonsillectomy on MRI and HRV biomarkers in children with OSA have not been reported. We hypothesize that comorbid ADHD can deteriorate brain damage and autonomic dysfunction, and adenotonsillectomy can reverse these alternations in children with OSA. The aims of this study are (1) to investigate the differences in pediatric brain tissue integrity, autonomic function, attention, behavior, quality-of-life, and sleep factors between the 'OSA with ADHD', 'OSA without ADHD', and 'healthy control' group; (2) to evaluate the efficacy of adenotonsillectomy versus watchful waiting with supportive care, with respect to the same variables of interest; (3) to evaluate whether the relative efficacy of the treatment differs according to baseline ADHD, weight, or OSA severity; and (4) to develop a predictive model for surgical success rate using both conventional well-known factors and MRI/HRV biomarkers. This is a 3-year prospective study that includes two parts. The Part I study is a cross-sectional study recruiting 100 children (5 to 9 years of age) to investigate the differences in brain tissue integrity (voxel-based morphometry and fractional anisotropy; assessed by structure MRI [T1] for volumetric alternations of gray and white matter, resting-state functional MRI for functional connectivity, and diffusion MRI for white matter integrity), autonomic function (time-domain and frequency-domain analyses; assessed by a wearable, real-time HRV measurement), severity pf attentive and behavioral problems (assessed by the Swanson, Nolan and Pelham IV-Teacher and Parent Rating Scale), quality-of-life (assessed by OSA-18), and sleep factors (apnea-hypopnea index, obstructive apnea index, arousal index, mean and least oxygen saturation, and sleep stage; assessed by polysomnography) between the OSA with ADHD group (Study Group 1; n = 40), the OSA without ADHD group (Study Group 2; n = 40), and the healthy control group (Control Group; n = 20). The Part II study is a randomized controlled trial includes a total of 64 children with OSA (32 children will be recruited from Study Group 1 and Study Group 2, respectively). We randomly assigned (1:1) these 64 pediatric patients with OSA to adenotonsillectomy or a strategy of watchful waiting with supportive care, matched by ADHD, obesity, and severe OSA. Variables of interest using the same methodology are assessed at baseline and at 7 months.

Background To date, obstructive sleep apnea (OSA) is a chronic and serious disorder with an increasing prevalence in many developed countries. OSA is characterized by dynamic imbalance between airway patency and collapse during sleep leading recurrent airway obstruction (partial or complete) and repetitive apneas and hypopneas. OSA results in gas exchange abnormalities, cortical arousals, autonomic arousals, sleep fragmentation, and systemic fragmentation. Notably, OSA has a prevalence of up to 5% in children and 50% in obese children. OSA is associated with various co-morbidities which affect multiple organ systems, resulting in acute events or long-term sequelae, and consequently incurring considerable social, economic, and health costs. For example, pediatric OSA can induce hypertension, cardiovascular disorder, metabolic syndrome, growth retardation, learning problem, night enuresis, and attention-deficit/hyperactivity disorder (ADHD). Persistent OSA severity is associated with the decreased quality of life of patients' families and increased concern regarding financial burden.

In children, behavioral impairment, neurocognitive dysfunction, and reduced scholastic achievements are now well-characterized morbidities of OSA. In addition, parentally reported daytime sleepiness, hyperactivity, and aggressive behaviors can also develop, albeit to a lesser extent in children who habitually snore but in the absence of OSA. Degrees of discipline problems and poor attention span were significantly higher in OSA children compared with non-OSA controls. However, a recent meta-analysis found that there were few studies with low risk of bias (levels of evidence I and II) showed that OSA children's intellectual abilities may be impaired but remain within the normal range. Capdevila et al (2008) recognized that 'the major intriguing component of the association between OSA and cognitive functioning lies in the observation that not all children with OSA actually manifest cognitive morbidities.'. Nevertheless, which specific cognitive ability (language, memory, attention, executive function) drives poor academic performance is unclear. Therefore, we must test OSA patients using measures that allow for fractionated higher- and lower-order cognitive abilities based on accepted cognitive neuropsychology models and find other factors may be recreating a role of OSA-associated neurobehavioral consequences.

OSA severity is potentially associated with genetic and environmental determinants of susceptibility in children. Increased body mass index (BMI), having the potential to central obesity, is one of the most important risk factors for OSA and is more strongly associated with neurocognitive and behavioral problems in children than OSA alone. Inflammation is associated with increased risk for neurocognitive deficits in children with obesity and/or OSA. Although obesity could influence academic performance, there was insufficient evidence to support a direct link between obesity and poor academic performance in school age children and change in weight status was not associated with change in cognitive function in children with obesity. Therefore, current evidences support that obesity has the potential to be a marker rather than a cause of low academic performance. Furthermore, the ε4 allele of the APOE gene is associated with increased risk of OSA and poorer cognition in children. However, APOE ε4 status alone was associated with memory, not language/executive functioning, in later life in a longitudinal study. Pediatric OSA is associated with sympathetic outflow in terms of overnight increases in urinary concentrations of catecholamines whereas increased urinary levels of γ-aminobutyric acid and decreased urinary level of taurine could underlie mechanisms of neuronal excitotoxicity and dysfunction. Both pediatric OSA and cognitive deficits might be to be predicted by overnight changes in urinary concentrations of selected neurotransmitters. However, those findings could not generalize to other disease entity. Incorporation of environmental elements such as nutrition, recurrent exposure to respiratory viruses, passive or active exposure to cigarette smoking, intensity of intellectual activity, and socioeconomic status is important in children because all these can affect both the pathophysiological risk for OSA as well as modify the susceptibility to the consequences of OSA. Unfortunately, there is still no consensus if we should consider OSA as a single disease with different phenotypes with or without neurocognitive deficits, or if there are different diseases with different genetic determinants, pathogenic mechanisms, prognosis, and treatment. Most of this important information is not routinely collected during clinical assessment of children with habitual snoring in ENT clinics.

ADHD, characterized by symptoms of inattention and hyperactivity/ impulsivity, often emerges during the preschool years and remains impairing throughout the life span. ADHD is among the most commonly diagnosed neurodevelopmental disorders, affecting approximately 8%-12% of children worldwide. Prevalence of ADHD in preschoolers, recently estimated to be 2.1%, is lower than that of school-age children and adolescents. The Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5) identifies 3 types of ADHD: primarily hyperactive-impulsive, primarily inattentive, and combined type. There is a body of evidence emerging that indicates that hyperactivity-impulsivity declines in preschoolers with ADHD, and inattention increases or at least becomes more evident when children enter structured school settings. ADHD is a heterogeneous disorder, in terms of the multifactorial etiological risk factors, diverse expressions of the symptom domains, comorbid disorders, neuropsychological impairments, and long-term trajectories. Both sleep and behavior problems are common in preschool and school-aged children. Notably, lower quality of sleep in infancy significantly predicts compromised attention regulation and behavior problems at 3-4 years of age. In our previous studies, we identified that pediatric OSA is closely associated with ADHD. In pediatric OSA, 32% and 35% had concomitant ADHD in children aged 4 to 5 years and those aged 6 to 11 years, respectively. In contrast, a full sleep assessment in children with ADHD found 50% of them had OSA that caused chronic sleep deprivation and could be considered as a signature of ADHD.

Alterations in the genes encoding for molecules involved in catecholamine signaling that weaken nor epinephrine production may impair the prefrontal cortex (PFC) circuits mediating the regulation of attention and behavior. Inadequate catecholamine release is associated with fatigue and ADHD, and therapeutic doses of ADHD medications likely normalize catecholamine transmission in patients with inadequate norepinephrine and dopamine levels, or both, thus bringing PFC function to more optimal levels. Prescription stimulants such as methylphenidate and non-stimulants such as atomoxetine are labeled for the treatment of ADHD from age approximately 6 and above years. Although there was no evidence of increased serious cardiovascular risk in children with ADHD exposed to ADHD medications, ADHD medications have been shown to increase blood pressure, heart rate, and QT interval in children. Halperin and Marks (2019) concluded that acute treatments of ADHD have demonstrable efficacy, but do not appear to fundamentally alter underlying mechanism. For example, improvement of pediatric OSA after adenotonsillectomy results in improvement of behavior problems during short-term follow-up and long-term follow-up. Recently, a combination of atomoxetine (norepinephrine reuptake inhibitor) and oxybutynin (antimuscarinic agent) administered orally before bedtime on 1 night greatly reduced adulthood OSA severity, and these findings also open new possibilities for the pharmacologic treatment for pediatric OSA with ADHD.

A range of acute processes, including hypoxia/re-oxygenation, repeated arousals, and episodic hypercarbia, may be corresponding to changes in some regional brain tissue integrity in pediatric OSA. Brain magnetic resonance imaging (MRI) using voxel-based morphometry (VBM) analysis has been used to explore possible sources of behavioral impairment and neurocognitive dysfunction in pediatric OSA. Reduced gray matter volume and increased and reduced regional cortical thicknesses can help to understand the underlying pathology. Using a measure of local changes in signal intensity patterns from high-resolution MRIs, sleep scientists can assess the entropy tissue texture to investigate brain tissue integrity which reflects the nature and extent of injury or structural adaptation in pediatric OSA subjects. In OSA children, several brain sites including the PFC, middle and posterior corpus callosum, thalamus, hippocampus, and cerebellar areas showed reduced entropy values, indicating tissue changes suggestive of acute insults. Kheirandish-Gozal et al (2018) concluded that children suffering from OSA display predominantly acute tissue injury in neural regions principally localized within autonomic, respiratory, cognitive, and neuropsychologic control, functions that correspond to previously-reported comorbidities associated with OSA. Patients with ADHD have symptoms similar to those caused by lesions to the right PFC. Imaging studies have shown reduced size and reduced functional activity of the right PFC causing difficulties with top-down attention regulation in patients with ADHD. The PFC requires an optimal level of norepinephrine and dopamine for proper function: either too little (as when we are drowsy or fatigued) or too much (as when we are stressed) markedly impairs PFC regulation of behavior and thought. However, multiple functional and structural neural network abnormalities beyond the classical fronto-striatal model, including fronto-parieto-temporal, fronto-cerebellar and even fronto-limbic networks, have been found via functional MRI in ADHD subjects. Furthermore, diffusion MRI including a technology of diffusion tensor imaging (DTI) to detect atypical fractional anisotropy (FA) to understand brain white matter integrity has been applied in OSA and ADHD (both diseases had lower FA in the corpus callosum at baseline) and compare the effect of medical treatment (continuous positive airway pressure for OSA: increased FA in the hippocampus, temporal lobes, fusiform gyrus, and occipital lobes; methylphenidate for ADHD: increased FA in the corpus callosum. Obviously, these MRI studies indicated that a closely neuropathological association between OSA and ADHD. While both OSA and ADHD are lifelong disorders in many patients, however, there is no study reporting differences in brain MRI (in terms of brain tissue integrity and diffusion) between OSA children with ADHD and OSA children without ADHD and between baseline and after adenotonsillectomy in our literature review. Therefore, voxel-based morphometry and fractional anisotropy of the brain can be assessed by structure MRI (T1 images) for measuring volumetric alternations of gray and white matter, resting-state functional MRI for assessing functional connectivity, and diffusion MRI for evaluating white matter integrity in children with OSA with or without ADHD.

Since children with OSA have lower entropy in neural regions involving autonomic control and children with ADHD have a hypofunction of the PFC, changes in the autonomic nervous system (ANS) due to OSA and ADHD span have been studied. In children with OSA and hypoxemia, an increase in sympathetic tone during sleep was found. Despite the levels of overnight urinary noradrenaline and adrenaline are increased and changes occur in the sympathetic tone in children with OSA, pediatric OSA do not usually develop high blood pressure (BP). It's not easy to use to monitor 24-h BP to dynamically evaluate the ANS conditions. However, changes in the function of ANS also modify children's heart rate variability (HRV). The sympathetic activation and decreased parasympathetic suppression can increase heart rate. The PFC, which is vital for attention, motor control, emotional regulation and higher order autonomic control, is hypofunctional in OSA and ADHD. PFC activity has been associated with changes to HRV via mediation of the cortico-subcortical pathways that regulate the parasympathetic and sympathetic branches of the ANS. PFC hypoactivity decreases parasympathetic tone and increases contributions from the sympathetic nervous system. Using the time domain analysis, children with habitual snoring (apnea-hypopnea index [AHI] <1 events/h) showed significantly lower values of the proportion of NN50 divided by the total number of NN (R-R) intervals (pNN50; mostly reflect the parasympathetic component) comparing with the control children in all sleep stages; children with OSA showed significantly lower values of pNN50 comparing with the control children in all sleep stages 1-hr study of daytime HRV; children with ADHD had reduced pNN50 compared with controls. Reduced vagal tone and neurobehavioral consequences supported that disrupted autonomic regulation seen in children with OSA and ADHD may be central in origin. Using the frequency domain analysis, an increase in low-frequency/high-frequency ratio (LF/HF; reflects the sympathovagal balance to the heart) for N2 sleep stage and rapid eye movement (REM) sleep was found in the chronic snorer group compared to the non-snoring control group and increased LF/HF in children with ADHD. However, an increase in LF/HF ratio was seen at overnight or 24-hr studies in adults with OSA (AHI ≥5 events/h) suggestive of increased sympathetic modulation whereas a decrease in LF/HF ratio in all sleep stages in children with moderate-to-severe OSA (obstructive AHI [OAHI] >5) compared with the control group (OAHI ≤1). Furthermore, the most recent study indicated that age and autonomic control, rather than cerebral oxygenation and OSA severity, were predictive of electroencephalograph (EEG) spectral power in children whereas autonomic cardiovascular control and cerebral oxygenation were associated with EEG spectral power in adults. Nevertheless, both the HR and LF/HF decreased and the SDNN increased after adenotonsillectomy in all stages of sleep in children with OSA. Therefore, these changes in parameters that reflect parasympathetic modulation are more easily detectable in pediatric OSA and ADHD than sympathetic modulation.

Measurement of HRV shows information on the functional state of the ANS. HRV is a physiological phenomenon of variation in the time interval between heartbeats. Depressed or reduced HRV primarily means that the HR is monotonously regular, and it means a lowered of the ANS regulatory function and ability to keep homeostasis, cope with internal and external stressors, and resist disease or recover in proper time. Significant advances in software programs to automatically derive HRV have led to its extensive use in psychophysiological research. In general, HF (0.15-0.4 Hz) represents parasympathetic activity whereas LF (0.04-0.15 Hz) represents sympathetic and parasympathetic activities and LF% represents sympathetic activity. Although there are standardized autonomic tests and well-established ranges of normal values in adults, there is limited normal values for HRV parameters in children Among the time-domain indices, there was strong to excellent correspondence for SDNN, RMSSD, and pNN50 to ensure measurement fidelity across signal processing software programs and reproducibility of 24-h HRV. The frequency-domain indices yielded excellent correspondence for LF, HF, and LF/HF ratio, except for VLF which exhibited poor correspondence. Stringent user-decisions and technical specifications for nuanced HRV processing details are essential to ensure measurement fidelity across signal processing software programs. Because median heart rate decreases from 113 beats per min (bpm) by 2 years of age to 73 bpm by 15-18 years of age and median respiratory rate decreases from 31 breaths/min by 2 years of age to 16 bpm by 15-18 years of age, it is plausible that these controversial results of LF/HF during sleep in children may be due to 'adult-based measurements of HRV'. Kuo et al (2008) has developed an analytic software of HRV measurement (combined with polysomnography or alone) that uses the adjustable time segments (16 sec-300 sec) which allow to more sensitively and specifically quantify HRV in different sleep stages and active waking-quiet sleep transitions for rats and human. Using a wearable ECG patch, we can perform a real-time, whole-day HRV analysis to deal with short N-N intervals and breath-breath intervals in children. This HRV analytic program opens new possibilities for the optimal determination of autonomic function for children.

The high prevalence and impact on daily life of pediatric OSA necessitate clinicians to offer effective and acceptable treatment options. However, recent evidence has raised questions about the benefits of adenotonsillectomy in curing OSA for children. Although adenotonsillectomy significantly increased sleep efficiency and oxygen saturation and decreased stage 1 sleep, AHI, obstructive apnea index (OAI), and arousal index, only 27%-79% of pediatric OSA had complete resolution of OSA (post-operative AHI <1 events/h). Beyond the AHI, we previously found that adenotonsillectomy can improve behavior, elevated BP, ADHD severity, inflammation, and OSA-related quality of life. In contrast, the attention and executive-function score on the Developmental Neuropsychological Assessment (with scores ranging from 50 to 150 and higher scores indicating better functioning), was close to the population mean of 100, and the change from baseline to follow-up did not differ significantly between the early-adenotonsillectomy group and the watchful-waiting group. In our latest case-control study, baseline AHI significantly differed between the non-obese with non-severe OSA (nO-nS) and non-obese with severe OSA (nO-S) subgroups, between the nO-nS and obese with severe OSA (O-S) subgroups, between the nO-S and obese with non-severe OSA (O-nS) subgroups, and between the O-nS and O-S subgroups at baseline. After adenotonsillectomy, AHI significantly decreased in the nO-nS, nO-S, and O-S subgroups. Therefore, post-operative AHI only significantly differed between the nO-nS and O-S subgroups. OSA-18 scores were equal in all the 4 subgroups either at baseline or after adenotonsillectomy whereas OSA-18 scores significantly reduced after adenotonsillectomy in all the 4 subgroups. Interestingly, although adenotonsillectomy did not reduced AHI, this operation still significantly reduced OSA-18 score which was close to the means of the other 3 subgroups. Again, these findings suggested that improvement of behavior and quality of life was not necessarily parallel to the improvement of AHI. Therefore, our results further call for attention to investigate the effect of adenotonsillectomy on other polysomnographic parameters, brain MRI, and autonomic function in pediatric OSA (with or without ADHD).

Study objectives

In this prospective study, the objectives are:

1. to investigate the differences in pediatric brain tissue integrity, autonomic function, attention, behavior, quality-of-life, and sleep factors between the OSA with ADHD group, the OSA without ADHD group, and the healthy control group in a cross-sectional study;
2. to evaluate the efficacy of adenotonsillectomy versus watchful waiting with supportive care, with respect to the same variables of interest in a randomized controlled trial;
3. to evaluate whether the relative efficacy of the treatment differs according to baseline ADHD, weight, or OSA severity; and
4. to develop a predictive model for surgical success rate using both conventional well-known factors and MRI/HRV biomarkers.

Inclusion Criteria:

1. Ages 5.0 to 9.99 years at time of screening.

2. Diagnosis of OSA with ADHD, OSA without ADHD, or non-snoring with typical development:

   2.1) Diagnosed with OSA defined as: OAI ≥1 event/h or AHI ≥2 events/h, confirmed on nocturnal, laboratory-based polysomnography and parental report of habitual snoring (on average occurring >3 nights per week).

   2.2) Diagnosed with ADHD defined as: Six or more of (1) inattentive symptoms (nine symptoms), (2) hyperactive and impulsive symptoms (nine symptoms) or (3) combined inattentive, hyperactive and impulsive symptoms must be present for at least 6 months, be inconsistent with the child's developmental level, and have a negative effect on their social and academic activities.

   2.3) Diagnosed with non-snoring with typical development defined as parental report of no habitual snoring (on average occurring ≤3 nights per week), <6 inattentive symptoms, <6 hyperactive and impulsive symptoms, and being consistent with the child's developmental level.

3. Tonsillar hypertrophy ≥1 based on a standardized scale of 0-4 (0 = surgically absent; 1 = taking up <25% of the airway; 2 = 25-50 % of the airway; 3 = 50-75 % of the airway; 4 = >75% of the airway).

4. Deemed to be a surgical candidate for adenotonsillectomy for OSA by ENT evaluation (the Part II study).

Exclusion Criteria:

1. Recurrent tonsillitis that meets published ENT clinical practice guidelines for surgery defined as: > 3 episodes in each of 3 years, 5 episodes in each of 2 years, or 7 episodes in one year.
2. Craniofacial anomalies, including cleft lip and palate or sub-mucosal cleft palate or any anatomic or systemic condition which would interfere with general anesthesia or removal of tonsils and adenoid tissue in the standard fashion.
3. Obstructive breathing while awake that merits prompt adenotonsillectomy in the opinion of the child's physician.
4. Severe OSA or significant hypoxemia requiring immediate adenotonsillectomy as defined by: OAI >20 events/h or AHI >30 events/h or SpO2 <90% for more than 2% sleep time
5. Evidence of clinically significant cardiac arrhythmia, extremely overweight (body mass index [BMI] z-score > 2.99), severe health problems that could be exacerbated by delayed treatment for OSA (such as heart disease, cor pulmonale, poorly controlled asthma, epilepsy required medication, diabetes, mental retardation), current use ADHD or psychotropic medication(s), and previous upper airway surgery.
6. A family planning to move out the area within the year.

The investigators will be blinded to the intervention group while they perform outcome assessment.