Dentofacial Morphology in Obese and Non-Obese Children With and Without Obstructive Sleep Apnea

Transcription

1 Dentofacial Morphology in Obese and Non-Obese Children With and Without Obstructive Sleep Apnea by David Justin Simone A thesis submitted in conformity with the requirements for the degree of Master of Science (Orthodontics) Graduate Department of Dentistry University of Toronto Copyright by Dr. David Justin Simone 2016

2 Dentofacial Morphology in Obese and Non-obese Children With and Without Obstructive Sleep Apnea David Simone Master of Science (Orthodontics) Graduate Department of Dentistry University of Toronto 2016 Abstract Introduction: Several craniofacial abnormalities have been suggested to contribute to obstructive sleep apnea. These characteristics vary significantly among the literature and are limited by the infrequent use of polysomnogprahy, the gold standard for diagnosing and quantifying obstructive sleep apnea. Objective: To compare the prevalence of facial and/or dental imbalances in children with and without obstructive sleep apnea in cohorts of obese and non-obese children. Methods: A prospective, cross-sectional study of children (ages 4-16) who were referred for a polysomnogram at The Hospital for SickKids. Facial features and malocclusion was assessed clinically by one dentist, blinded to the PSG results. Results/Conclusions: Horizontal excess (overjet) was the only dentofacial finding which was significantly more common in children with obstructive sleep apnea as compared to those without obstructive sleep apnea (p=0.04). Dentofacial characteristics were also not different between children using positive airway pressure therapy and children not on positive airway pressure therapy ii

3 Acknowledgments I am grateful to all the people who have supported me and helped me throughout this process. I would first like to thank my supervisors, Dr. Reshma Amin and Dr. Bryan Tompson, without whom this project wouldn t have been successful without your endless guidance and advice. I would also like to thank my committee members, Dr. Fernanda Almeida, Dr. Nelly Huynh, and Dr. Indra Narang, for your contribution to this project. You are all responsible for guiding me in the proper direction to complete this project. Secondly, I would like to thank the numerous people who dedicated their time that allowed this project to be completed in a timely manner. I would like to thank Nicole Sidhu, Nadia Kabir, Aman Sayal, and Tanvi Naik for helping with entering data into the database, coordinating patient schedules, and preparing appropriate consents for patients. I would like to thank Allison Zweerink for informing me of the schedules of the nightly polysomnograms and Derek Stephens for your help with the statistical analyses. Finally, I would like to thank my beautiful wife, Joanna and my two precious daughters, Amalia and Violette. This project is dedicated to you. Without your infinite support and neverending love, I would not have been able to get through the last three years. Thank you for picking up the slack when I was busy concentrating on school related work. Without a doubt, I would not have been able to get through any of this without you. iii

4 Table of Contents Abstract... Acknowledgments. _Table of Contents.. List of Tables. List of Figures... List of Abbreviations List of Appendices ii iii iv vi vii viii x Chapter 1: Background Obstructive Sleep Apnea Obesity Epidemiology of OSA Pathophysiology of OSA Obesity and OSA Complications of Pediatric OSA Diagnosis Pediatric Obstructive Sleep Apnea Treatment OSA and Craniofacial and Dentofacial Development Rationale Study Aim 26 Chapter 2: Materials and Methods Subjects Study Procedures iv

5 2.3 Demographics and Anthropometric Measures Sleep Questionnaires Clinical Orthodontic Examination Polysomnogram Statistical Analysis Study Outcomes Hypothesis. 37 Chapter 3: Results Intra-rater Reliability Study Participants Polysomnography Results Questionnaire Results Dentofacial Morphology. 49 Chapter 4: Discussion Appendix A Appendix B Appendix C Appendix D Bibliography v

6 List of Tables Table 1.1 Signs and Symptoms of Pediatric Obstructive Sleep Apnea.. 26 Table 2.1 Inclusion and Exclusion Criteria 38 Table 2.2 Summary of Study Procedures Table 2.3 Centers for Disease Control and Prevention Weight Categories Table 2.4 Frontal View Examination.. 41 Table 2.5 Profile View Examination.. 42 Table 2.6 Functional Assessment Table 2.7 Intra-Oral Examination Table 3.1 Intra-rater reliability Table 3.2 Subject Groups Table 3.3 Demographics of the Four Study Cohorts (excluding PAP group). 52 Table 3.4 Demographics of OSA vs. No OSA Groups (excluding PAP group). 53 Table 3.5 PSG Results across the Four Cohorts (excluding PAP group) Table 3.6 PSG results of OSA vs. No OSA Groups (excluding PAP group).. 55 Table 3.7 Spruyt and Gozal Questionnaire results across the Four Cohorts (excluding PAP group).. 57 Table 3.8 Frequencies of Spruyt and Gozal Scores of OSA vs. No OSA groups (excluding PAP group).. 57 Table 3.9 Pediatric Sleep Questionnaire Results across the Four Cohorts (excluding PAP group) 59 Table 3.10 Frequencies of PSQ Scores of OSA vs. No OSA groups (excluding PAP group).. 59 Table 3.11 Prevalence of Dentofacial Characteristics across the Four Cohorts (excluding the PAP group). 61 Table 3.12 Dentofacial Morphology of OSA vs. No OSA Groups (excluding PAP group). 62 Table 3.13 Univariate Analysis for Various Dentofacial Characteristics (excluding the PAP group) Table 3.14 Dentofacial Morphology of Obese & OSA vs. Obese & PAP groups. 64 Table 3.15 Multiple Regression Model for the Presence of OSA in the Study Cohort Excluding Children Using PAP Therapy vi

7 List of Figures Figure 4.1 ROC Curve for Spruyt Gozal Score. 58 Figure 4.2 ROC Curve for PSQ Score 60 vii

8 List of Abbreviations SDB: Sleep Disordered Breathing OSA: Obstructive Sleep Apnea CA: Central Apnea HS: hypoventilation syndrome AT: Adenotonsillectomy PSG: Polysomnogprahy REM: Rapid Eye Movement ICSD: International Classification of Sleep Disorders AASM: American Academy of Sleep Medicine BMI: Body Mass Index CDC: Centers for Disease Control and Prevention AHI: Apnea-Hypopnea Index MRI: Magnetic Resonance Imaging EMG: Electromyography Pcrit: Critical Nasal Pressure CRP: C-reactive protein IL-6: Interleukin-6 GCR: Glucocorticoid receptor gene TNF-α: Tumour necrosis factor-alpha ICS: Intranasal Corticosteroid OM: Oral Montelukast ADHD: Attention-deficit/hyperactivity disorder ANP: atrial natriuretic peptide ADH: antidiuretic hormone viii

9 PSQ: Pediatric Sleep Questionnaire EEG: Electroencephalography EOG: Electrooculography HRQOL: Health Related Quality of Life LDL: Low density lipoprotein HDL: High density lipoprotein OAI: Obstructive Apnea Index PAP: Positive Airway Pressure CPAP: continuous pressure airway pressure BPAP: Bi-level positive airway pressure RME: Rapid Maxillary Expansion CBCT: Cone-beam computed tomography AI: Apnea Index SRDB: Sleep related breathing disorder IOTN: Index of Orthodontic Treatment Need NPAF: Nasal Pressure Airflow SaO 2 : Oxygen saturation EtCO 2: End-tidal Carbon Volume TcCO 2 : Transcutaneous carbon dioxide OAHI: Obstructive apnea-hypopnea index CAI: Central apnea-hypopnea index ANOVA: Analysis of Variance ICC: Intraclass Correlation Coefficient OR: Odds ratio ROC: Receiver Operating Curve ix

10 List of Appendices Appendix A Appendix B Appendix C Appendix D Spruyt and Gozal Sleep Questionnaire Pediatric Sleep Questionnaire (PSQ) Pediatric Polysomnogram Set up Pediatric Polysomnogram Data Recording x

11 Page 1 Chapter 1 Background Sleep disordered breathing (SDB) is a broad term encompassing abnormalities in respiratory pattern, gas exchange and sleep architecture during sleep. 1 SDB includes: i) obstructive sleep apnea (OSA), episodes of complete or partial airway obstruction; ii) central apnea (CA), prolonged pauses in the absence of respiratory effort; and iii) hypoventilation, persistent low tidal volume breathing or bradypnea causing hypercarbia and hypoxemia. 2 OSA is the most common subtype of SDB affecting 1-5% of healthy children. 3 Adenotonsillar hypertrophy is the most common cause of OSA in healthy children. The first-line treatment for OSA is adenotonsillectomy (AT). 1.1 Obstructive Sleep Apnea OSA is defined by the American Thoracic Society as a functional disturbance in sleep characterized by transient and partial/complete obstruction of the airways which interrupts sleep, resulting in disruption of normal gas exchange (intermittent hypoxia and hypercapnia) and sleep fragmentation 1. OSA in children was first described systematically in using clinical symptoms and polysomnogpraphy (PSG). Since then the recognition of abnormal breathing during sleep has progressed tremendously in the last two decades, with the realization that childhood OSA is both common and serious. The understanding of the pathophysiology has improved, although much remains to be known. There is increased recognition of the relationship between respiratory abnormalities during sleep and adverse consequences. Children with OSA tend to have a different pattern of breathing during sleep than adults. Children have a higher arousal threshold than adults 3. As a result, they frequently do not arouse in response to obstructive events, and most studies have demonstrated the preservation of sleep architecture 34 or only minimal changes in sleep architecture. This is in contrast with OSA in adulthood which is associated with significant sleep fragmentation and decreased slow wave and rapid eye movement (REM) sleep 4,55,6. In children, OSA is characteristically more severe in REM sleep secondary to the relative muscle atonia: a state specific deficit in upper airway function 1

12 With this increasing recognition of SDB in children, many classification systems have been developed. However, the International Classification of Sleep Disorders (ICSD) is most frequently used. The ICSD was first published in 1990 by the American Academy of Sleep Medicine (AASM) along with the Japanese Society of Sleep Research, Latin America Sleep Society and European Sleep Research Society. It was further revised in 2007 (Second Edition), and then again most recently in 2014 (Third Edition). OSA is classified within the sub-group of Sleep-Related Breathing Disorders 6. Page Obesity Obesity is defined as a Body Mass Index (BMI) at or above the sex-specific 95th percentile of BMI for age, based on the 2000 Centers for Disease Control and Prevention (CDC) Growth Charts 7. There has been an increasing trend in childhood obesity worldwide. In 1978, childhood obesity among Canadian children and adolescents aged 3-19, was 5%. In 2013, the prevalence has increased to 13% 7. Obesity in children and adolescents is now recognized as a major medical and public health problem that affects nearly every major organ system 8. Childhood obesity has both immediate and long-term effects on physical and mental health. Children with obesity are more likely to have risk factors for cardiovascular disease, such as high cholesterol or high blood pressure. In a population-based sample of 5- to 17-year-olds, 70% of children with obesity had at least one risk factor for cardiovascular disease. 9 In addition, children and adolescents with obesity are at greater risk for bone and joint problems, sleep apnea, and social and psychological problems such as stigmatization and poor self-esteem 10. Over the long-term, children with obesity are more likely to be obese as adults 11,12 are therefore at higher risk of developing cardiovascular disease, Type 2 diabetes, stroke, several types of cancer, and osteoarthritis 13. Craniofacial morphology has also been shown to differ between obese and normal adolescents. In 2005, Sadeghianrizi et al 14 compared craniofacial morphology in obese and normal adolescents using lateral cephalometric radiographs. They found that obese adolescents exhibited significantly larger mandibular and maxillary dimensions than normal adolescents. IN general, obesity was associated with bimaxillary prognathism and relatively greater facial measurements 14. 2

13 1.3 Epidemiology of OSA Page 3 OSA is a common condition in children of all ages, from neonates to adolescents, and can result in severe complications if left untreated. Diagnostic criteria for OSA among adults is defined as an apnea-hypopnea index (AHI) of 5 or greater events per hour on nocturnal PSG and evidence of disturbed sleep, daytime sleepiness, or other daytime symptoms 15. The diagnostic criteria for OSA in children is heterogeneous across studies. At present, an AHI of 1 to 5 events per hour is most frequently used to define OSA in children 15. From the available studies, the estimated prevalence rates of OSA in healthy children range between 1.2% and 5.7% depending on the AHI threshold for OSA diagnosis. If these prevalence rates were applied to the 2011 Census of Canada population estimates, that would translate between 93,425 and 443,772 Canadian children aged 0-19 years being diagnosed with OSA, which is equivalent to children per 100,000 children as being diagnosed with OSA. Children between 2 and 8 years of age are at increased risk of OSA as this coincides with the peak of adenotonsillar hypertrophy in childhood 17,19. In general, infants or older children outside this age window are likely have additional or other underlying etiologic factors such as dentofacial abnormalities or neuromuscular disease predisposing them to the development of OSA 16,17,19,20. In contrast to adults, where OSA is more common in men than women, OSA in children appears to occur equally amongst the sexes OSA has been shown to have a higher prevalence amongst African American children than Caucasian children 15,23 and Asians have more severe OSA than matched Caucasians Pathophysiology of OSA While the clinical features of OSA are well understood, the understanding of its pathogenesis remains incomplete. OSA is a result of a balance between structural factors and functional factors and both appear to play a role. Upper airway anatomy as well as collapsibility is important in the pathogenesis of OSA. The patency of the upper airway is determined by a balance between the intraluminal negative pressure of the airway and the 3

14 soft tissue structures that support the upper airway 23,25. When the collapsing forces are great enough to obstruct the airway, an obstructive apnea or hypopnea can occur. Page Upper Airway Anatomy The upper airway is composed of muscle and soft tissue but lacks rigid or bony support. Most notably, it contains a collapsible portion that extends from the hard palate to the larynx. It has the ability to change shape and momentarily close for speech and swallowing during wakefulness. This also renders the upper airway vulnerable to collapse during sleep. High-resolution magnetic resonance imaging (MRI) has been utilized to determine the size of the upper airways structure in children 26,27. As compared with control subjects, children with OSA have a smaller oropharynx, larger adenoids, tonsils and retropharyngeal nodes 26. Arens et al. 27 have shown with regional analysis of MRIs that the upper airway in children with OSA is most restricted where the adenoids and tonsils overlap. However, with segmental analysis, the upper airway has been shown to be restricted throughout the initial two-thirds of its length and that the narrowing is not in a discrete region adjacent to either the adenoid or tonsils, but rather in a continuous fashion along both 27. Furthermore, Schiffman et al 28 used MRI to determine the mandible dimensions of children with OSA (24 subjects with mild to moderate OSA), and demonstrated that a smaller mandible is not a feature in children with OSA. The tonsils and adenoids grow progressively during childhood and usually reach maximal size by the age of MRI has also shown that in children with habitual snoring, enlarged tonsils and adenoids restrict the upper airway and that soft palate volume is also larger in children with OSA 30. Surgical treatment for OSA in these children have been shown to reduce symptoms and improve, quality of life, and PSG findings, thus providing evidence of beneficial effects of early AT 31. However, adenotonsillar hypertrophy is only one of the potential determinants of OSA in children given that OSA persists in select patients after AT 32. In children with certain medical conditions, such as Down Syndrome, the prevalence of OSA is much higher (30%-55%) than in otherwise healthy children 33. Craniofacial abnormalities which predispose these children to OSA include midfacial 4

15 hypoplasia and mandibular hypoplasia, glossoptosis, an abnormally small upper airway with superficially positioned tonsils and relative tonsillar and adenoidal hypertrophy, hypopharyngeal collapse, tracheal stenosis, and laryngomalacia 23,33. Page 5 Furthermore, cephalometric studies in children with OSA frequently report narrower maxilla 34, mandibular retrognathia 35,36, longer lower facial height 35-37, and caudal placement of the hyoid bone 38. A reduced upper airway sagittal width has also been reported based on a reduced distance on lateral cephalometric radiographs as measured from the posterior nasal spine to the adenoids. On average, this distance is mm shorter in children with OSA compared with healthy controls 39 However, other studies report no differences in measures of maxillary and mandibular width, length, or volume between patients with OSA and normal control subjects 40. Thus, the contribution of skeletal abnormalities to the development of OSA in otherwise normal children is controversial 32. In 2013, Flores-Mir et al, conducted a systematic review and meta-analysis to consolidate the current knowledge of craniofacial morphological characteristics associated with upper airway constriction resulting is OSA in children. Their study only included cephalometric values and did not include a complete description of dentofacial characteristics. The authors identified nine articles and found that three cephalometric variables, the angle between the mandibular plane and sella nasion line (MP-SN), the angle from SN to B point (SNB) and the angle from A point to nasion point to B point (ANB), were significantly different between children with and without OSA. Children with OSA had a steeper mandibular plane angle (MP-SN = +4.2 ), a more retrusive mandible (SNB = ), and were more likely to show a class II skeletal pattern (ANB = ) 41. Similar findings were found in the systematic review by Katyal et al 39. The authors demonstrated that children with OSA and primary snoring showed increased weighted mean differences in the ANB angle of 1.64 and 1.54, respectively, compared with the controls. The increased ANB angle was primarily due to a decreased SNB angle in children with primary snoring by 1.4. In this meta-analysis, PSG was performed to determine the presence and severity of OSA Upper Airway Collapsibility 5

16 The pathophysiology of OSA in children is a complex interaction between an airway predisposed toward collapse and neuromuscular compensation. Even though anatomical determinants have been shown to be of critical importance to the development of OSA, they do not completely account for the pattern of SDB 32. Most children with severe OSA are able to maintain normal sleep state distribution, particularly REM and slow wave sleep, despite having obstructive apneic episodes 4,42. This suggests that there may be a compensation to maintain airway patency during obstructive episodes via neuromuscular activation, ventilatory control, and arousal threshold 32. Page 6 The significance of neuromuscular modulation in maintaining airway patency is highlighted with three clinical observations: (1) apnea is observed predominantly in REM and stage 2 sleep rather than in wakefulness or slow wave sleep 4 ; (2) although sedated and anesthetized children with OSA have narrower and more collapsible airways compared with normal control children, there is considerable overlap 40,43 ; and (3) during sleep, most children with OSA intermittently attain a stable breathing pattern, suggesting that reflex neuromuscular activation below the arousal threshold is possible 44. The pharyngeal dilator muscles responsible for modulating airflow through the upper airway include the genioglossus, hyoglossus, and styloglossus. These muscles act in unison and produce forward movement of the tongue, increase oropharyngeal airway size and stiffness 32. During wakefulness, children with OSA have an increased genioglossus electromyography (EMG) recording levels compared with non-osa control children 45 suggesting a reflex activation of the muscle via mucosal mechanoreceptors to negative airway pressure. During the initial onset of sleep, this EMG activity decreases in both OSA children and control subjects with a subsequent increase in airway resistance and collapsibility of the airway. However, the EMG activity remains below the wakeful baseline during stage 2 of sleep in normal children, suggesting a mechanically stable airway. In contrast, most children with severe OSA have an increase in EMG activity during sleep stage 2, suggesting the need for neuromuscular compensation to maintain airway patency 46. During collapse of the upper airway, minute ventilation decreases, which induces a compensatory increase in respiratory effort. This results in large negative luminal pressure during inspiration. As a result, a negative pressure reflex causes activation of 6

17 pharyngeal dilator muscles to decrease airway collapsibility and increase minute ventilation 32. Marcus et al. 47 reported that children with OSA rely on arousal mechanisms to sustain minute ventilation, which disrupts sleep homeostasis, whereas normal children who were subjected to inspiratory resistance loading, were able to respond to the loading with an increased inspiratory time and sustain loads without arousing for several minutes. That is, normal children were able to perform the negative pressure reflex without arousal, whereas, the negative pressures reflex is diminished or completely lost in patients with OSA 32. Page 7 In adults, the critical nasal pressure (Pcrit) at which the upper airway collapses is higher in patients with OSA than in those with primary snoring. In 1994, Marcus et al 48 compared the Pcrit between prepubertal children with OSA and those with primary snoring. Pcrit was determined by correlating the maximal inspiratory airflow with the level of positive or negative nasal pressure applied via a nasal mask. As in adults, they found that the maximal inspiratory airflow varied in proportion to the upstream (nasal) rather than the downstream (esophageal) pressure changes. Pcrit was 1 ±3 cmh 2 O in OSA compared with -20 ± 9 cmh 2 O in primary snorers. They concluded that Pcrit, a measure of airway collapsibility, correlated with the degree of upper airway obstruction and was reduced postoperatively, consistent with increased upper airway stability. The negative pressure reflex is important in maintaining upper airway patency by exciting pharyngeal muscle dilators through neuromuscular compensation. It is plausible that mucosal inflammation or edema could impair the afferent limb of this reflex. It is hypothesized that snoring induces a mucosal inflammatory response resulting in swelling, affecting upper airway resistance and/or collapsibility Inflammation It has been previously established that OSA induces a systemic proinflammatory response which can result in end-organ dysfunction 49. Sleep apnea in children is associated with increased inflammatory responses and increased plasma levels of C- reactive protein(crp) and interleukin-6 (IL-6) 50. In 2014, Mutlu et al 51 investigated the clinical significance of preoperative serum CRP, interleukin-6 (IL-6), fetuin-a, cystatin C, adiponectin and tumor necrosis factor-alpha (TNF-α) levels in children with 7

18 adenotonsillar hypertrophy and compared the results with post surgical values. They found that levels of cytokines in children with SDB secondary to adenotonsillar hypertrophy decreased after surgical treatment. They concluded that the risks of development of cardiovascular disease are decreased in association with lower levels of cytokines. 51. Page 8 Goldbart et al. 52 have demonstrated the presence of upper airway inflammation in children with OSA. They found increased expression of leukotriene receptors in tonsillar tissue from children with OSA compared with children with recurrent throat infections. In a subsequent study, Goldbart et al 53 found an upregulation of the glucocorticoid receptor gene (GCR) expression in OSA derived adenoid and tonsil tissues compared with tissue from children with recurrent throat infections. Translational studies incorporating intranasal corticosteroids 54, leukotriene receptor antagonists 55, or both 56, for the treatment of pediatric OSA have demonstrated a reduction in OSA severity. The largest study to date looking at the anti-inflammatory therapy for mild OSA was published in 2014 by Kheirandish-Gozal et al 57. A combination of intranasal corticosteroid (ICS) and oral montelukast (OM) for 12 weeks normalized PSG sleep findings in 62% of their 752 sample size diagnosed with mild OSA. Thus, a combination of ICS and OM as treatment of mild OSA appears to be effective and have lasting effects 57. Tauman et al. 58 first correlated the increase in CRP levels among American children with OSA with AHI, arterial oxygen saturation, and arousal index measures. Although CRP is a nonspecific marker of inflammation, recent epidemiologic studies suggested that CRP may participate directly in atheromatous lesion formation through reduction of nitric oxide synthesis and induction of the expression of particular adhesion molecules in endothelial cells 59. It was noted that these increases were prominent among children who presented with neurobehavioral complaints. They suggested that intermittent hypoxia and sleep fragmentation of OSA may underlie these systemic inflammatory responses. However, a subsequent group from Greece found conflicting results that CRP levels are not significantly different between control subjects and children with OSA 60. Since the association of plasma CRP concentrations with OSA in childhood has 8

19 been inferred from several studies demonstrating increased circulating levels of CRP with increasing severity of OSA, Bhattacharjee et al 61 used CRP to assess if residual disease persists post AT in children with OSA. They found that pre-at AHI and post-at CRP levels were most significantly associated with residual OSA 61. Page 9 OSA has also been associated with insulin resistance, hyperglycemia and dyslipidemia in children 62. Koren et al 63 found that AT improved insulin sensitivity and HDL levels, but not fasting glucose or other lipoprotein levels despite a parallel increase in BMI z scores. This suggests that OSA is causally involved in creating an adverse metabolic state independent from obesity because the metabolic changes did not differ significantly between children without obesity and children with obesity or between boys and girls. Fasting insulin was most strongly associated with post-at AHI, such that more children with insulin resistance were more likely to have residual OSA. Koren et al 64 followed up this study to assess the independent contributions of OSA to insulin resistance and dyslipidemia in large pediatric cohort(n=459). They found that although obesity was the primary driver of most associations between OSA and metabolic measures, sleep duration was inversely associated with glucose levels, with stage 3 Non REM sleep (N3) and REM sleep being negatively associated and sleep fragmentation positively associated with insulin resistance measures. In children with mild OSA, the presence of obesity increased the odds for insulin resistance, while higher AHI values emerged among obese children who were more insulin-resistant 64. Thus the exclusive presence of interactions between OSA and obesity in the degree of insulin resistance is coupled with synergistic contributions by sleep fragmentation to insulin resistance in the context of obesity. Insufficient N3 or REM sleep may also contribute to higher glycemic levels independent of obesity 64. It is difficult to distinguish between inflammatory mechanisms leading to SDB as opposed to the systemic/local inflammation resulting from the presence of SDB. However, most data support the concept of a disease that is associated with inflammation that is ameliorated after surgery at the systemic and the local airway level. In contrast, there are no data that confirm pre-existing inflammation in children with newly diagnosed OSA 65. 9

20 1.5 Obesity and OSA Page 10 Since its initial description in 1976, obesity and OSA has become widely recognized as a highly prevalent condition in children 2. The last two decades has witnessed a shift from the classic presentation of children with OSA (i.e. adenotonsillar hypertrophy and failure to thrive) to a majority of children being overweight or obese, even though adenotonsillar hypertrophy continues to play a role in the latter group 66. Early descriptions of childhood OSA rarely described obese patients. Most children were of normal weight and failure to thrive was a common complication 67. However, with the epidemic of childhood obesity continually rising, the epidemiology of childhood OSA is shifting towards obesity as being an important risk factor. The risk of OSA is greatly increased by obesity in children, with an estimated prevalence ranging from 19 to 61% depending on the definition of OSA, the degree of obesity and the age of the study population 68. Compared to the estimated 3% prevalence of OSA in 2- to 8-year-old children 15, the risk of OSA in obese children has been estimated to be as high as 36% 69, and may exceed 60% 70 when habitual snoring is present. The presence of both OSA and obesity sets into motion a viscous cycle, where the presence of OSA affects metabolic requirements which can perpetuate the tendency towards obesity 71,72. In addition, sleepiness will reduce the likelihood of engaging in physical activity and enhance eating behaviors that favor calorie-dense foods 66,72.Clinic-based and epidemiological studies have confirmed that obesity is an important risk factor for OSA 73 and is one of the strongest predictors of SDB in both adults and children 74. In a case-control study design, Redline et al 75 examined risk factors for SDB in children aged 2-18 years (n = 399), and found that the risk among obese children was increased four to five fold. The proposed physiologic mechanisms that may contribute to OSA in obese children include anatomic and functional factors restricting the upper airway, alterations in chest wall mechanics affecting lung volumes and upper airway collapsibility, and inflammatory and metabolic factors that may perpetuate the disorder 66,76. 10

21 Adenotonsillar hypertrophy has been recognized as an important anatomic cause of restriction of the upper airway and contributing to the development of OSA in children with obesity However, residual OSA after AT has been reported in 54-76% of these children 82 compared with approximately 15-20% in children without obesity 80,83. Nandalike et al 84 were the first to quantify the volumetric changes in the upper airway in children with obesity and OSA after AT. They found that AT increased the volume of the nasopharynx and oropharynx, reduced tonsils, but had no effect on the adenoid, lingual tonsil, or retropharyngeal nodes. They also noted a small significant increase in the volume of the soft palate. These findings could explain the lower success rate of AT reported in children with obesity and OSA. Page 11 With regards to obesity, the pathophysiologic mechanisms for OSA are both mechanical and functional. Mechanically, deposition of adipose tissue within the base of the tongue and the pharynx results in decreased airway size and increase airway resistance 85. Functionally, there is a reduced lung volume due to displacement of the diaphragm by the obese abdomen and a decreased central ventilatory drive 78. Using MRI, Arens et al 77 determined the anatomic risks factors associated with OSA in obese children as compared with obese control subjects without OSA. As compared with control subjects, subjects with OSA had a smaller oropharynx (P= 0.05) and larger adenoid (P = 0.01), tonsils (P = 0.05), and retropharyngeal nodes (P = 0.05). The size of lymphoid tissues correlated with severity of OSA whereas BMI did not have a modifier effect on these tissues. Subjects with OSA demonstrated increased size of parapharyngeal fat pads (P= 0.05) and abdominal visceral fat (P =0.05). The size of these tissues did not correlate with severity of OSA and BMI did not have a modifier effect on these tissues. In conclusion, upper airway lymphoid hypertrophy is significant in obese children with OSA. The lack of correlation of lymphoid tissue size with obesity suggests that this hypertrophy is caused by other mechanisms. Although the parapharyngeal fat pads and abdominal visceral fat are larger in obese children with OSA they could not find a direct association with severity of OSA or with obesity 77. Recent evidence suggests that OSA is associated with a state of chronic inflammation characterized by increased oxidative stress, pro-inflammatory cytokine production, and metabolic deregulation 86. It has been shown to also contribute to the 11

22 pathogenesis and progression of nonalcoholic fatty liver disease, via the deleterious effects of chronic intermittent hypoxia on liver metabolism and inflammation 87. Alkhouri et al 86 demonstrated that circulating markers of hepatocyte apoptosis were significantly altered in children with OSA. More specifically, levels of soluble CD163, a marker of macrophage activation, increased significantly in children with OSA and improved after OSA treatment. These findings indicate that children with OSA have increased apoptotic and inflammatory pressures 86. Page Complications of Pediatric OSA It is now well established that SDB may lead to serious and measurable end-organ dysfunction. This is especially important in children because of the risk of life-long negative sequelae. The effects of untreated SDB include neurocognitive deficits, cardiovascular complications, inflammation, growth impairment, reduction in health related quality of life (HRQOL) and increased healthcare resource utilization Neurocognitive Complications One of the most well-established long term sequelae of pediatric OSA is behavioral and neurocognitive dysfunction 1. Behavioral dysregulation is the most commonly encountered comorbidity of OSA 1. Sixty-one articles including over children have directly explored the relationship between OSA and behavioral and neurocognitive function 88. The vast majority of studies consistently report some association between OSA and hyperactivity, attention deficits and impulsivity 1. Poor school performance, impaired executive functioning, and an inverse relationships between memory and learning have all been reported in children with OSA 1.To investigate a causal relationship between decrements in cognition and OSA, in the past decade 19 studies have assessed neurocognition pre and post treatment for OSA 1. The majority of the studies have demonstrated significant improvements post treatment with three studies demonstrating sustained improvements at more than a year post treatment Cardiovascular Complications The cardiovascular complications of SDB are of immediate importance because earlier diagnosis and treatment can reverse these processes and prevent its consequences 12

23 in adult life 1. Recurrent episodes of upper airway obstruction, which are characteristic of OSA result in intermittent hypoxia, intrathoracic pressure swings and sleep fragmentation. This results in autonomic system activation supported by the followings findings: increased urinary cathecholamines, decreased pulse transit time and alterations in blood pressure regulation in OSA 1. Right ventricular dysfunction has also been demonstrated 1.Cardiac benefits of treatment for SDB have been shown. Plasma levels of B-type natriuretic peptide, a marker of ventricular strain, has been found to be elevated in children with SDB and to decrease after AT 1. Similarly, there is evidence of echocardiographic improvement of elevated pulmonary pressure also after AT 1. Page Inflammatory Complications OSA also appears to cause low grade systemic inflammation and local inflammation. This is thought to be the result of the intermittent hypoxia and sleep fragmentation leading to the production of free radicals and systemic oxidative stress. Increased circulating levels of CRP, as well as adhesion molecules have also been shown. Anti-inflammatory therapy targeting upper airway inflammation has been shown to improve residual OSA post AT Somatic Growth Failure OSA can also impair somatic growth. Failure to thrive has been reported in up to 50% of children presenting for AT 89. The suggested etiologies include decreased caloric intake, increased work of breathing as well as a reduction in growth factors such as insulin like growth factor-1 and growth hormone. Selimoglu has demonstrated significant increases in insulin like growth factor-1 six months after adenotonsillectomy 90. Furthermore, elevations in low density lipoprotein (LDL) cholesterol along with reduced levels in high density lipoprotein (HDL) cholesterol were observed in both obese and non-obese children with OSA, with significant improvements after OSA treatment 91,92. 13

24 1.6.5 Quality of Life and Healthcare Resource Utilization Page 14 Childhood SDB leads to significant decreases in HRQOL and these scores significantly improve following treatment 93.Healthcare resource utilization is a powerful index of disease morbidity in children 1. Healthcare resource utilization is significantly increased and usage is elevated several years before an OSA diagnosis 94. The total number of admissions in children with OSA is 40% higher as compared to matched controls. In summary, SDB is associated with serious and measurable end-organ dysfunction. Treatments for SDB are available and the benefits of treatment have been demonstrated which argues for timely diagnosis and treatment of SDB to avoid long-term negative sequelae. 1.7 Diagnosis The management goals for childhood OSA are to 1) identify children who are at risk for OSA; 2) diagnose children with OSA; and 3) treat children with OSA to prevent negative sequelae of untreated disease. Diagnostic tools that have been studied include clinical history and physical examination, patient questionnaires, and PSGs. Given the resource intensive nature of PSGs in combination with the limited access to PSGs, the pediatric sleep medicine field has tried to identify tools that can be used clinically to screen for OSA Signs and Symptoms The most common signs and symptoms, based on history and physical examination of the child, associated with childhood OSA are summarized in Table 1.1. Several studies have evaluated the use of history alone as a screening tool for the diagnosis of OSA. Preutthipan et al 95 aimed to determine whether parents observations (such as observed cyanosis, snoring extremely loudly, shaking the child, being afraid of apnea) could predict the severity of OSA. Although they found that some parent s observations are more frequently reported in children with OSA, neither any single nor combination of observations accurately predicted the severity of OSA. They found an 14

25 overall poor sensitivity and specificity when evaluating various historical factors in children with OSA. Page 15 Snoring is the most common clinical symptom of OSA. It is a sensitive, nonspecific, screening symptom for OSA 29. If a history of nightly snoring is elicited, a more detailed history regarding labored breathing during sleep, observed apnea, restless sleep, diaphoresis, enuresis, cyanosis, excessive daytime sleepiness, and behavior or learning problems (including attention-deficit/hyperactivity disorder (ADHD)) should be obtained 96. In children with OSA, findings on physical examination during wakefulness are most often normal. However, there may be non-specific findings related to adenotonsillar hypertrophy, such as mouth breathing, nasal obstruction during wakefulness, adenoidal faces, and hyponasal speech 96. Table 1.1 Signs and Symptoms of Pediatric OSA 97 Daytime Symptoms Morning headaches Daytime sleepiness Diagnosis of ADHD Learning problems Irritability Hyperactivity Nocturnal Symptoms Snoring Witnessed apneas Gasping Paradoxical Breathing Neck hyperextension Nocturnal Diaphoresis Nocturnal enuresis Physical Examination Underweight or overweight Tonsillar hypertrophy Adenoidal faces Micrognathia/retrognathia High-arched palate Signs of cor pulmonale Hypertension ADHD is a behavioral abnormality commonly seen in children and adolescents. Its main symptoms include inattention, hyperactivity, and impulsivity 98. Attention deficit and hyperactivity are known possible symptoms or correlates of OSA 99. Chervin 100 and O Brien 18 reported that children with mild symptoms of ADHD showed a high 15

26 prevalence of snoring and sleep problems. However, these associations may be missed in children, because ADHD is a common primary diagnosis in itself. In conclusion, OSA can mimic the signs of ADHD. Furthermore, unlike in adults, children with OSA, especially younger children, rarely have excessive daytime sleepiness, and parental reports of sleepiness vary with the questionnaire used 99. If misdiagnosed as ADHD, children may be subject to long-term methylphenidate, a commonly used medication for ADHD, whereas recognition and treatment of the underlying sleep disorder should be treated, to prevent unnecessary medication use Page 16 A higher prevalence of nocturnal enuresis has been reported in children with OSA. Although, the exact etiology is not yet known, it has been postulated that increased enuresis may be because of the dampening effects of OSA on the arousal response, changes in bladder pressure or possibly the secretion of hormones involved in fluid regulation, such as atrial natriuretic peptide (ANP) and antidiuretic hormone (ADH) 102,103. Nonetheless, in the majority of children with OSA, the physical examination is normal. In addition, the presence of adenotonsillar hypertophy has not been shown to reliably predict OSA 104, Questionnaires Questionnaires have been developed as screening tools for the diagnosis of OSA. At present, two of the more commonly used questionnaires to screen for SDB are the Pediatric Sleep Questionnaire (PSQ) 106 and the Spruyt and Gozal 6-item Sleep Questionnaire 107. The PSQ (see Appendix B) was first published and validated in 2000 by Chervin et al 106. It consists of 22 item parent-reported questionnaire. It is composed of four subscales for SDB, snoring, sleepiness, and behaviour. The PSQ performed slightly better than other published questionnaires as a screening tool for the detection of OSA with a sensitivity of 0.85 and a specificity of 0.87 when using an established cut-off score of 0.33 for the original validation study 29. In a follow up study, using PSG to diagnose OSA, Chervin et al 108 subsequently found a lower sensitivity, 0.78 and specificity

27 In 2012, Spruyt and Gozal 107 (see Appendix A) developed a set of six hierarchically arranged questions that aided in the screening of children at high risk for SDB. A total of 1,133 children between the ages of 5- to 9-years-old were evaluated using the questionnaire. All sleep-related questions used the Likert-type responses never (0), rarely (once per week; 1), occasionally (twice per week; 2), frequently (three to four times per week; 3) and almost always (>4 times per week; 4) for the preceding 6-month time frame. Overall, the questionnaire had a sensitivity of 59.03%, specificity of 82.85%, positive predictive value of 35.4 and negative predictive value of Page 17 A 2002 systematic review by Schechter et al. 109, looked at the use of questionnaires as screening tools for OSA. The authors concluded that questionnaires had an unacceptably low sensitivity and specificity for predicting OSA. This was further confirmed in a more recent systematic review in 2014 by De Luca Canto et al 110. These authors concluded that the PSQ had sufficiently high sensitivity and specificity to be used as a screening tool for OSA but not as a true diagnostic tool for pediatric OSA Polysomnography The gold standard test to diagnose OSA is an overnight PSG, also known as a level I study 111. The overnight PSG is attended by a sleep technologist during which at least seven physiological channels are measured. An overnight PSG monitors electroencephalography (EEG), chin and leg electromyography (EMG), electrooculography (EOG) and cardiorespiratory variables, including respiratory effort, heart rate, oximetry and carbon dioxide levels for approximately 8 to 10 hours. The PSG determines the AHI which describes the severity of OSA. AHI is defined as the number of apneas and hypopneas per hour of total sleep time. Apnea is defined as a drop in the peak airflow > 90% of baseline, with the drop lasting at least the duration of two breaths during baseline breathing and is associated with the presence of respiratory effort throughout the entire period of absent airflow. Hypopnea is defined as a drop in the peak airflow > 30% of baseline, for the duration of at least two breaths in associations with either > 3% oxygen desaturation or an arousal 112. An AHI greater than 1.5 events per hour is a positive diagnosis for OSA

28 Though PSGs are the gold standard test for the diagnosis of pediatric OSA, there are many challenges related to performing a PSG. These include: inconvenience to the patient and, the relative scarcity of sleep laboratories with a resulting extended wait period between referral and actual testing. In 2014, Katz et al 116 aimed to describe pediatric sleep physician and diagnostic testing resources for SDB in Canadian children. They found marked disparities across the province/territories with many provinces having no practitioners or access to PSGs. In the provinces that had access to PSGs, reported wait times ranged from <1 month to years. This study clearly demonstrated a lack of resources and services for pediatric SDB across Canada, with pronounced disparities. Even if only affected children were tested with PSG, the authors estimate there are 7.5 times more children with OSA than the current testing capacity in Canada 116. Page Pediatric Obstructive Sleep Apnea Treatment Treatment for OSA must be individualized based on the clinical assessment, anatomy of the upper airway and severity of the disease Adenotonsillectomy The most common cause of childhood OSA is adenotonsillar hypertrophy 117. The first line treatment for OSA in these children is AT. Early studies showed that pre-pubertal adolescents initially considered to have been cured of OSA by AT subsequently had recurrence as teenagers. Guilleminault et al. showed that subjects initially treated with AT had narrowing behind the base of the tongue and oral-facial anatomical abnormalities that either did not exist initially or had not been identified previously Tasker et found that a narrow upper airway and snoring persisted 12 years after AT 119. In another study Guilleminault et al. (n =207) demonstrated that complete resolution of OSA following AT was present in only 51% in non-obese pre-pubertal children that were studied with PSG 3 months post-operatively 120. More recently, in a large, multicenter retrospective study, Bhattacharjee et al. (n=500) found that although AT led to significant improvements in indices of SDB in children, residual disease was present in a large proportion of children (70%), particularly among older (>7 yr) or obese children

29 Marcus et al 31 were the first to perform a randomized controlled study evaluating the benefits and risks of AT, as compared with watchful waiting, for the management of OSA. This Childhood Adenotonsillectomy Trail (CHAT) was designed to evaluate the efficacy of early AT versus watchful waiting. A total of 464 obese and non-obese otherwise healthy children between the ages of 5 to 9 years old age were included. Children with an AHI score of more than 30 events per hour, an obstructive apnea index (OAI) score of more than 20 events per hour, or arterial oxyhemoglobin saturation of less than 90% for 2% or more of the total sleep time were not eligible, owing to the severity of the PSG findings. The primary study outcome was the change in the attention and executive-function score on the Developmental Neuropsychological Assessment (NEPSY; scores range from 50 to 150, with 100 representing the population mean and higher scores indicating better functioning) 122. This test has well-established psychometric properties 122 and comprised three tasks (tower building, visual attention, and auditory attention) performed under the supervision of a psychometrist. Secondary outcomes for this study were caregiver and teacher ratings of behaviour, symptoms of OSA, sleepiness, global quality of life, disease-specific quality of life, generalized intellectual functions, and PSG indexes. They found that compared with a strategy of watchful waiting, surgical treatment for obstructive sleep apnea in school-age children did not significantly improve attention or executive function as measured by neuropsychological testing but did reduce symptoms and improve secondary outcomes of behavior, quality of life, and PSG findings, thus providing evidence of beneficial effects of early AT 31. Page 19 Substantial subgroup differences with regard to the normalization of PSG findings post AT in Marcus et al 31 randomized trial of AT were observed within each study group. Regardless of the assigned treatment (surgery or watchful waiting), normalization of PSG findings was seen less frequently in black children than in children of other races, in children with obesity than in children without obesity, and in children with a baseline AHI score above the median than in those with a baseline AHI score at or below the median 31. Among obese children, those randomly assigned to early AT had greater reductions in symptoms and greater improvement in behavioral and PSG outcomes than did those in the watchful waiting group 31. However, there persistence of 19

30 OSA post AT was higher in the children with obesity as compared with the children without obesity (33% vs. 15%) 31. Page 20 Overall, the results of published data on the success of AT ranges from 24 to 100 percent in the literature 123. In 2009, Friedman et al. 123 performed an updated systematic review of AT for the treatment of pediatric OSA. The meta-analysis included 1079 subjects with a mean age of 6.5 years of age. The effect measure was the percentage of pediatric patients with OSA who were successfully treated with AT based on preoperative and postoperative PSG data. When cure was defined as an AHI of <1, AT was successful only 66.3% of the time. However, although complete resolution is not achieved in most cases, it still offers significant improvements in AHI, making it a valuable first-line treatment for pediatric patients. Adenotonsillectomy yields improvements in children with OSA however complete normalization occurs in only 25% of the patients, with obesity and AHI at diagnosis being the major determinate for the success for surgical outcome 81. Therefore, obesity should be considered as an important, potential, contributor to residual airway obstruction after surgery with its own, independent, contribution to the pathophysiology of OSA Positive Airway Pressure Positive Airway Pressure (PAP) therapy is recommended for children with moderate to severe OSA post AT or if a child is not a candidate for AT. PAP therapy for OSA was first developed more than three decades ago 125. PAP works by counteracting the sleep-induced negative transmural pressure that promotes collapse and narrowing of the collapsible upper airway. PAP therapy maintains upper airway patency via the delivery of pressurized air through an interface that is worn over the nose or the nose and mouth. This creates a pneumatic splint which prevents partial or complete collapse of the upper airway during sleep 125. The aim of PAP therapy is to normalize the obstructive AHI, improve sleep quality and normalize gas exchange. There are two types of PAP therapy that are delivered by a mask: 1) continuous pressure airway pressure (CPAP) and 2) Bi-level positive airway pressure (BPAP) therapy. For both 20

31 CPAP and BPAP, the non-invasive interfaces include nasal pillows, nasal, oronasal and full face options. Page 21 Although PAP therapy has been increasingly prescribed in children over the past few decades due to advances in both the technology itself as well as the number of interfaces that are available for children, the greatest barrier to PAP therapy is adherence. In a recent study, DiFeo et al. prospectively studied children and adolescents and concluded that PAP adherence is primarily related to family and demographic factors rather than the severity of apnea or measures of psychosocial functioning 126. This is an additional challenge as the entire family often needs to be engaged for the child to be adherent with PAP. Access to pediatric PSG is currently limited by a 12 -month waiting period for PSG at Sick Kids alone, with similar wait times across Canada. At present, the literature is insufficient and contradictory in describing the long term effects of PAP therapy, on the development of the face, jaw, and teeth 88. A few case reports have suggested that early childhood long-term treatment using either CPAP or BPAP carries a high risk of facial growth impairment, in particular, midface hypoplasia and Class III malocclusions 127,128. However, more recently, a small sample size, crosssectional study failed to show any statistically significant difference between long-term PAP use and dentofacial abnormalities in children with persistent OSA 129. This is an area of needed further study Orthodontic Treatment Persistent OSA post AT has also led to the consideration of orthodontic modalities to treat OSA. There are several craniofacial abnormalities where imbalanced development may contribute to OSA such as posterior crossbite, Class II skeletal and dental patterns, and anterior open bite. Aside from aesthetic and occlusion benefits, orthodontic treatment can help guide facial growth in order to correct facial imbalances, improve swallowing, reposition tongue posture and re-establish nasal breathing 130. Early detection and treatment of children with OSA and facial imbalances may prevent the sequelae of this disease. Early orthodontic treatment could prevent a need for AT and provides another treatment option for children with OSA that are not adherent to 21

32 PAP therapy. Page Rapid Maxillary Expansion Rapid Maxillary Expansion (RME) is a common orthodontic procedure used to correct maxillary arch constriction by opening the mid-palatal suture. It is a common treatment modality to correct posterior crossbites in the primary, mixed, or permanent dentition. The precise role of maxillary constriction in the pathophysiology of OSA is unclear. However, it is known that a significant number of children with OSA have nasal obstruction (nasal septal deviation with or without turbinate hypertrophy) associated with a narrow maxilla. Maxillary constriction is thought to increase nasal resistance and alter tongue position, leading to narrowing of the retroglossal airway and subsequently the development of OSA 131. There is no evidence to support that RME enlarges oropharyngeal airway volume. Zhao et al 132 retrospectively studied 24 adolescent patients (mean ±SD age years) with maxillary constriction using hyrax palatal expanders and compared that to 24 control patients (mean ±SD age ) undergoing routine orthodontic treatment without palatal expansion. They used cone-beam computed tomography (CBCT) to assess changes in the volume, length, and minimum cross-sectional area of the oropharynx. They found no statistically significant differences between the groups. On the other hand, RME has been shown to increase nasal width and nasal cavity dimensions. Pirelli at el. 133 investigated the effect of RME on 31 children (19 boys, 12 girls) with maxillary constriction, without adenoid hypertrophy, with OSA demonstrated by PSG. RME was performed for 10 to 20 days with 6 to 12 months of retention. The mean AHI fell from 12.2 events per hour to less than one event per hour, demonstrating a resolution of the SDB. In a further study, Pirelli et al. 134 evaluated if RME in 42 children with a case history of oral breathing, snoring, and night time apneas could improve the patency of nasal breathing and OSA. Selection criteria included no adenotonsillar hypertrophy, BMI<24, and narrow maxillary arch determined by posterior-anterior cephalometric 22

33 evaluation. Investigations were carried out before orthodontic treatment, one month after treatment (T1) and after the end of orthodontic treatment (T2). All results were analyzed by postero-anterior cephalometric evaluation in T0, T1, and T2. The results reported that in all 42 patients, RPE therapy widens the nasal fossa and releases the septum restoring normal nasal airflow with a disappearance of obstructive sleep disordered breathing 134. Page Functional Appliance Therapy Functional appliances are removable or fixed intraoral devices which alter the muscles forces against the teeth and craniofacial skeleton. They depend on altered neuromuscular action to effect bony growth and occlusal development. They have been used in children with OSA because functional appliances posture the mandible forward and potentially enlarge the upper airway and increase the upper airspaces, improving the respiratory function 135. In 2007, a Cochrane based review assessed the effectiveness of using functional orthopaedic appliances for the treatment of OSA in children 136. Out of 384 potentially relevant studies, only 1 paper was included 137, demonstrating the lack of methodologically sound research in this area. For example, there was an important methodological problem with many papers not showing important information necessary to assess their quality. Papers did not present information such as: how participants were allocated to interventions, who generated the allocation, how sample size was determined etc 136. Villa et al 137, compared active oral appliance vs. no treatment and studied 32 children, with total apnea index (AI) of more than 1 event/hr diagnosed by PSG. A decrease of at least 50% in the total AHI was considered treatment success. In 9/14 treated subjects, the AHI fell 50%. Although this study showed some results that favored the intervention, the results must be considered with caution due to methodological problems such as non-randomized generation of allocation, no allocation concealment, no blinding, no sample size calculation reported, number of patients randomized different from patients analyzed, high number of loss to follow up and no intention-to-treat analysis was performed. In conclusion, the available information is not enough to answer whether oral 23

34 appliances or functional appliances are effective in the treatment of sleep apnea in children 136. Page OSA and Craniofacial and Dentofacial Development Facial growth and development is primarily dictated by genetic factors, but environmental inputs also contribute, in particular, with respect to the mode of breathing. Children who suffer from respiratory problems and OSA commonly exhibit disturbances in dentofacial morphology. The growth of the dentofacial regions follows the functional matrix theory; that is, growth occurs in response to functional needs and possibly in response to the growth of the nasal cartilage 138. Linder-Aronson 139 proposed the cause-and-effect pathway of reduced nasal breathing during wakefulness and resultant craniofacial abnormality. When nasal breathing is reduced, possibly from enlarged adenoids or an anatomical defect (i.e. decreased nasal width or nasal septal deviation), mouth breathing is inevitable as the primary mode of respiration. Mouth breathing leads to an altered pattern of muscle recruitment in the oral and nasal capsule, which ultimately results in skeletal changes 140.The important role of abnormal nasal resistance during the early developmental period was demonstrated from studies on infant rhesus monkeys 140.A small silicone head was placed within the nostrils of infant rhesus monkeys in order to induce nasal resistance for the first 6 months of life 140,141. The blockage of the nasal passage led to narrowing of the dental arches, decrease in maxillary length, anterior cross bite, maxillary overjet and an increase in anterior facial height 140. These changes were shown to be reversible if the experimental nasal resistance was withdrawn while the infant monkey was still in its developmental phase. In children, mouth breathing is most commonly associated with: an extended posture of the head (3 to 5 degree extended craniocervical posture); retrognathic mandible; a larger anterior facial height; a steeper mandibular plane; a lowered position of the hyoid bone; an antero-inferior posture of the tongue compared to normal children and a high palatal vault 142. This pattern of findings has been termed long face 24

35 syndrome, and is similar to the reported cephalometric findings in children with OSA Comparable changes in the craniofacial structure have been described in a group of Page 25 subjects with large tonsils, which has been termed the adenoidal face 139. The adenoid face is characterized by an incompetent lip seal, a narrow upper dental arch, retroclined mandibular incisors, increased anterior face height, a steep mandibular plane angle, and retrognathic mandible compared with faces of healthy controls 139. Thus, not only does upper airway obstruction predispose to OSA, but it also has an adverse effect on craniofacial development, posing an increased future risk of OSA 32. The literature reports mixed results with regards to the resolution versus the persistence of craniofacial abnormalities after treatment for OSA. On the one hand, some studies have shown that cephalometric variables normalize after treatment of OSA in children. In a five year follow up study after AT in children with OSA, resolution of maxillary and mandibular inclination abnormalities and lower face height was observed 37. On the other hand, it has been shown that open bites and cross-bites are observed 2 years after AT in most patients 143. As a general rule, if treatment is initiated at a young enough age (before 6 years of age), the long-term dentoalveolar development is more likely to normalize 143. Dentofacial anomalies can also present as malocclusions that can be observed during intra-oral examination. Posterior crossbite, Class II skeletal and dental patterns, and anterior open bite have been found to be more prevalent in OSA children versus healthy controls 143. The remainder of the reported occlusal characteristics varies significantly among the literature in children with OSA, emphasizing the need for further research on this topic. The specific reported prevalence s of posterior crossbite in children with OSA ranges between 16.7% % 39,144 versus reported controls 2.4%- 23.2% 39,145. The prevalence of Class II skeletal and dental patterns in OSA children ranges in the literature between 29.3%-88% 144,145 versus controls 4.9%-28% 145,146. Lately, the reported prevalence of anterior open bite in OSA children ranges between 5%- 20% 67,144 and 0% in the control groups Rationale A formal dental evaluation is not standard of care for either children referred to sleep clinics for query obstructive sleep apnea or those prescribed PAP therapy for the treatment of OSA. Sleep physicians perform a cursory craniofacial examination including 25

36 a basic assessment of maxillary constriction and mandibular hypoplasia. Given the emerging evidence in the literature demonstrating an improvement in OSA with orthodontic treatment, as well as the limited literature suggesting midfacial hypoplasia and class III malocclusions as a result of ongoing PAP therapy, a needed first step is to understand the prevalence of malocclusion and dental anomalies in children referred to a sleep center for query OSA. It is also important to understand how this prevalence differs in children that are currently being treated for OSA with PAP therapy. Determining the prevalence of dental abnormalities and malocclusion in these cohorts of children will inform future interventional studies to look at the relative efficacies of different treatment interventions. Page Study Aim The aim of our study is to report on the prevalence of dentofacial abnormalities in children with suspected OSA who have been referred for a PSG at Sick Kids.. 26

37 Page 27 Chapter 2 Materials and Methods 2.1 Subjects Ethics approval was obtained from both the University of Toronto Health Sciences Research Ethics Board (#31147) and the Hospital for Sick Children's Research Ethics Board (REB # ). Children referred to the sleep laboratory at the Hospital for Sick Children, Toronto, Ontario, Canada for overnight PSG between March 2015 to April 2016 were invited to participate in the study (n=100). The subjects were selected according to the inclusion and exclusion criteria listed in Table 2.1. Informed consent was obtained verbally and in writing from all study participants and/or their parents/legal guardian. Study assent was obtained when appropriate. Table 2.1 Inclusion and Exclusion Criteria. Inclusion Criteria Age 4-18 years Referred for an overnight PSG study at the sleep laboratory at the Hospital for Sick Children Exclusion Criteria Craniofacial abnormality related to an underlying genetic syndrome Children and/or parental caregivers not proficient in English 2.2 Study Procedures The study procedures are summarized in Table 2.2. A complete description is provided in the methods section below for each study procedure. All study procedures were completed at the clinically scheduled overnight sleep study. 27

38 Table 2.2 Summary of Study Procedures Page 28 Study Procedure Description Time To Complete Enrollment and Consent into study Consent obtained from parent/guardian and patient (if appropriate) 3-5 Minutes Demographic and Clinical Measures Patient and parent information collected 2-3 Minutes Sleep Questionnaires Clinical Orthodontic Examination Polysomnogram* Spruyt and Gozal Questionnaire Pediatric Sleep Questionnaire All dental examinations completed by Dr. David Simone. Extra-oral and intra-oral exam data collected. PSG undertaken by trained technologists according to the international guidelines Minutes 5-10 Minutes 8-10 hours * The polysomnogram was clinically indicated and not a research specific study procedure 2.3.Demographics and Anthropometric Measures The following demographic and anthropometric information was collected from each patient: 1) age; 2) date of birth; 3) gender; 4) country of origin of mother; 5) country of origin of father; 6) body type, 7) height, and (8) weight. Each patient s BMI was calculated using the formula BMI = Weight (kg)/ Height 2 (m 2 ). Each patient's BMI was then converted into a percentile for the population according to the patient s age and gender using the published data by the CDC 147. Weight status category was determined from each patient s BMI percentile according to the CDC s guidelines (Table 2-3) 148. Table 2.3 CDC Weight Categories Weight Status Category Percentile Range Underweight Normal/Healthy Weight Overweight Obese Less than the 5 th percentile 5 th percentile to less than 85 th percentile 85 th to less than the 95 th percentile Equal to or greater than the 95 th percentile 28

39 Page Sleep Questionnaires Spruyt and Gozal Questionnaire 107 The Spruyt and Gozal questionnaire was developed in 2012 and intended to be a screening tool for SDB (see Appendix A). All study participants completed this questionnaire. The questionnaire consists of six questions. For the development and validation of the questionnaire, 1,133 urban children with habitual snoring between the ages of 5 to 9 years of age that had undergone a PSG were included. This sample was analyzed based on established AHI cutoffs. The investigators developed a set of six ordered questions that allows for fair discrimination along the SDB spectrum. The questions can be found in Appendix A. Questions 1-4 and question 6 use Likert-type responses including: 1) never; 2) rarely; 3) occasionally; 4) frequently; 5) almost always Question 5 uses the following scale with regards to snoring: 1) mildly quiet; 2) medium loud; 3) loud; 4) very loud and ; 5) extremely loud. The total score for the questionnaire represents the average score of all six questions, according to the following formula (where Q1= raw score to question 1, Q2 = raw score to question 2, and so forth): A = (Q1+Q2)/2; B = (A+Q3)/2; C = (B+Q4)/2; D = (C+Q5)/2; and the cumulative score = (D+Q6)/2. Based on the original validation study, a score greater or equal to 2.72 out of 4 was indicative of a high risk for OSA Pediatric Sleep Questionnaire Parents/guardians were also asked to complete the PSQ. The PSQ was developed and validated for sleep disorders in 2000 by Chervin et al 106,108. Chervin et al studied children aged 2-18 years who had PSG confirmed sleep related breathing disorders (SRDB). Items thought to be predictive of SRBDs in children were formulated based on clinical experience. This produced a 22-item questionnaire that was strongly associated with diagnosis of SRBD (P<0.0001) (see Appendix B). The following options were available for each of the 22 items on the PSQ: yes, no or don t know. The number of symptom-items endorsed positively ( yes ) was divided by the number of items answered positively or negatively; the denominator therefore excluded items with missing responses and items answered as don t know. The result was a score, that ranged 29

40 from 0.0 to 1.0. Scores > 0.33 were considered positive and suggestive of pediatric SDB. This threshold was based on a validation study demonstrating optimal sensitivity and specificity at the 0.33 cut-off 108. Page Clinical Orthodontic Examination All study participants underwent a comprehensive, clinical orthodontic examination by the same examiner (D.S.), blinded to any reported signs and symptoms of query OSA. The examination consisted of dental, skeletal, functional and esthetic characteristics which were subdivided into four sections: (1) Frontal View, (2) Profile View, (3) Functional, (4) Intra Oral. The examination lasted approximately 5-10 minutes Frontal View Table 2.4 outlines the examination in the frontal view. Facial type and lower face height were categorized as brachycephalic if the lower third was shorter than the average, mesocephalic if the lower third was longer than the average, or dolichocephalic if the lower third was much larger than the average. Mandibular symmetry was assessed if the chin point was deviated from the facial midline, in the absence of a functional shift, and the relationship of the dental midline to the facial midline. Incisor and gingival display at both rest and smile were measured clinically using a flexible plastic ruler with 1mm accuracy. Table 2.4 Frontal View Examination 1. Type facial (if borderline, choose mesocephalic) Front View Mesocephalic Brachycephalic Dolichocephalic 2. Lower Face Height Normal Increased Decreased 3. Symmetry Symmetric Mandible shift to Right Mandible shift to Left 4. Dental Midlines (midline use cusp of upper lip) 5. Incisor display at rest Upper : on with facial midline shift to Right shift to Left; Amount : mm Lower : on with facial midline shift to Right shift to Left; Amount : mm mm 6. Gingival display on smile mm 7. Incisor display on smile mm 30

41 2.5.2 Profile View Page 31 Table 2.5 outlines the examination from the profile view. Facial profile was assessed by measuring the angle formed by a line dropped from soft tissue nasion (i.e. bridge of nose) to subnasale (i.e. base of the upper lip) and a second line extending from subnasale to soft tissue pogonion (i.e. chin point). An acute angle would indicate a convex profile; an obtuse angle would indicate a concave profile; and a straight line would indicate a straight profile. Lip position was determined relative to a straight line drawn from the tip of the nose to the most anterior curvature of the soft tissue chin. Lip strain on closing was assessed by the activity of the mentalis muscle. Table 2.5 Profile View Examination Profile View 8. Facial Profile Straight Concave Convex 9. Skeletal position - Maxilla Retrognathic Normal Prognathic 10. Skeletal position - Mandible Retrognathic Normal Prognathic 11. Nasiolabial Angle Normal 90º-100 º Acute (< 90 º) Obtuse (>100 º) Lip Position: 12. With respect to esthetic line: Upper lip Normal Retrusive Protrusive 13. With respect to esthetic line: Lower lip Normal Retrusive Protrusive 14. Lip strain to close Yes No Functional Table 2.6 outlines the Functional Assessment portion of the examination. Tonsil size was evaluated according to the Standardized Tonsillar Hypertrophy Grading Scale 149. Tonsil size 0 denoted surgically removed tonsils. Size 1 implied tonsils hidden within the pillars. Tonsil size 2 implied the tonsils extending to the pillars. Size 3 tonsils were beyond the pillars but not to the midline. Tonsil size 4 implied tonsils extend to the midline

42 Table 2.6 Functional Assessment Page 32 Functional 15. Tonsils Removed (kissing tonsils) 16. History of Mouth Breathing Yes: If YES specify: During Day Time During Night Time No Intra-Oral Table 2.7 outlines the intra-oral portion of the clinical examination. The intra-oral examination included both vertical and horizontal discrepancies, molar and canine Angle s classification, presence of crossbites, and maxillary and/or mandibular crowding or spacing. Angles classification of occlusion was assessed for both left and right sides of the dentition. Subjects in the permanent dentition were classified as having a Class I, Class II, or Class III malocclusion. Class I occlusion is defined as the mesiobuccal cusp of the permanent maxillary first molar occluding in the buccal groove of the permanent mandibular first molar. Class II malocclusion is defined as the mesiobuccal cusp of the permanent maxillary first molar occluding from a half to full cusp mesial to the buccal groove of the permanent mandibular first molar. Subjects were classified as Class III when the mesiobuccal cusp of the permanent maxillary first molar occluded from a half to full cusp distal to the buccal groove of the permanent mandibular first molar. Subjects in the primary or mixed dentition were classified as flush terminal plane, mesial step, or distal step occlusion. Flush terminal plane is defined when the distal surfaces of maxillary and mandibular primary second molars that lie in the same vertical plane. Mesial step is defined when the primary mandibular 2 nd molar is mesial in relation to the maxillary 2 nd molar and distal step when the primary mandibular 2 nd molar is posterior to the distal surface of the maxillary 2 nd molar. The presence of deep-bites, open-bites, crossbites, and scissor-bites as well as crowding, were assessed and classified according to the method described by Björk et al. (1964) 150. Vertical excess (i.e. overbite) was measured by taking the average measurement of both central incisors. Overbite was expressed as the percentage that the upper incisors vertically overlap the lower incisors. Horizontal excess (i.e. overjet) was 32

43 Page 33 also measured by taking the average measurement of both central incisors and measuring the horizontal (anterior-posterior) overlaps of the maxillary central incisors over the mandibular central incisors. Posterior crossbites were recorded if the buccal cusp of the upper tooth occluded edge-to-edge, or lingually, to the buccal cusp of the corresponding lower tooth. Posterior crossbites included cross bites of the primary or permanent molars and canines as well as permanent premolars. Crowding or spacing of the arch was evaluated by calculating the amount of overlap or space between the interproximal contacts of erupted teeth. In mixed dentition, this was done with the assumption that unerupted permanent canines, first premolars and second premolars will occupy 7mm of the mesio-distal arch dimension. The overall crowding or spacing was divided into mild (1-3 mm), moderate (4-9 mm), or severe (>10 mm). Intercanine width and intermolar width was measured using a Boley gauge with 0.01mm accuracy. Intercanine width was measured from the cusp tips of the maxillary right and left primary and permanent canines. Intermolar width was measured from the junction of the lingual groove at the gingival margin between the maxillary left and right primary second molars and permanent first molars. The Index of Orthodontic Treatment Need (IOTN) esthetic scale ranks malocclusion in terms of the perceived esthetic impairment in order to identify those who would most likely benefit from orthodontic treatment 151. Table 2.7 Intra-Oral Examination Intra oral 17. Oral Habits Yes No If YES, since When : years Which? 18. Horizontal Excess (taken at average of both central incisors, labial to labial) 19. Vertical Excess (taken at average of both central incisors, labial to labial) Nail Biting Biting lip/cheek Bruxism Sucking Thumb/finger Other: Overjet: mm Overbite: % 20. Anterior OpenBite Open bite: mm 21. Posterior Openbite R mm 22. Posterior OpenBite L mm 33

44 Page Odontogram E D C B A A B C D E E D C B A A B C D E Dental crossbite (including edge-to-edge bite) Anterior crossbite: Yes; If YES, # of maxillary teeth involved: No Posterior crossbite: Yes Unilateral; If YES, # of maxillary teeth involved: Bilateral No 25. Narrow Palate Yes No 26. CR/CO shift Yes, specify: 27. Intermolar distance (measured from mid-palatal gingival margin) 28. Intercanine distance (measured from cusp tip) Posterior - anterior Vertically To the right To the left No mm mm 29. Tongue size Normal Microglassia Macroglossia 30. Arch Shape Upper: U shape V shape Lower : U shape V shape 31. Palatal Depth mm 32. Stage of dentition Primary Mixed Permanent (No primary teeth present) 33. Molar Classification (according to R and L sides) 34. Canine Classification (<1/2 cusp = cl.1) Permanent: (<1/2 cusp = cl.1) Right: I II III Left: I II III Right: I II III Left: I II III Primary/mixed: Right: Mesial step Flush Distal Step Left : Mesial step Flush Distal Step 35. Space Analysis crowding Upper: <3 mm 4-9 mm >10mm Lower: <3 mm 4-9 mm >10mm spacing Upper: <3 mm 4-9 mm >10mm Lower: <3 mm 4-9 mm >10mm 34

45 Page IOTN esthetic scale (match for overall occlusal attractiveness) 2.6 Polysomnogram Subjects underwent a standard level one overnight baseline PSG using XLTEC (Oakville, Canada) data acquisition and analysis system. Sleep architecture and respiratory data were assessed 27 and information was obtained from PSG and scored according to the AASM scoring guidelines by a registered polysomnogprahic technician28. A standard overnight PSG lasting approximately 8-10 hours included a 4lead EEG (C3, C4, O1, and O2), two bilateral EOG leads referenced to A1 or A2, one submental and two tibial EMGs. Respiratory measurements included chest wall and abdominal movement using inductance pneumography; airflow using a nasal cannula connected to a Nasal Pressure Airflow (NPAF) by Braebon; oxygen saturation (SaO2) using a Massimo pulse oximeter (Irvine, CA); transcutaneous carbon dioxide 35

46 measurement (TcCO 2 ) using a LINDE carbon dioxide sensor (Munich, Germany). Video and audio recordings were obtained for each study. The raw data from the polysomnograms were scored according to the AASM guidelines, as is current clinical practice 152. Scoring involved the quantification of sleep staging, respiratory events, oxygen saturations and carbon dioxide recordings. Sleep architecture was assessed by standard techniques. Information obtained from each PSG included: sleep onset latency and REM onset latency, total sleep time, sleep efficiency, time spent in each sleep stage (percentage), and number and classification of arousals and snoring. Respiratory events included obstructive apneas and hypopneas, mixed apneas as well as central apneas and hypopneas. Page 36 The diagnosis and severity of OSA in children was based on the frequency of obstructive apneas, obstructive hypopneas, mixed apneas, central apneas and central hypopneas per hour during sleep as well as gas exchange characteristics. These were recorded as the obstructive apnea-hypopnea index (OAHI), central apnea-hypopnea index (CAI), baseline mean oxygen saturation and percentage of time the C0 2 is >50mmHg. OSA and central apnea (CA) severity will be graded according to accepted clinical criteria. An OAHI of <1.5 and CAI <1 is normal, an OAHI of >1.5 to <5 and CAI of >1 to <5 indicates mild OSAS and CAI, an OAHI or CAI of >5 to <10 indicates moderate OSA and CA, and an OAHI or CAO of 10 indicates severe OSA and CA 152. Nocturnal hypoventilation (C0 2 recording >50mmHg for >25% of the night), if present, was reported from the PSG. All PSGs were reported by one of the three clinical sleep physicians at The Hospital for Sick Children. 2.7 Statistical Analysis Descriptive statistics were used to summarize the study results. Intra-rater reliability testing was assessed for the orthodontic clinical examination. ICC was used for the continuous rating scale and Kappa statistics for the categorical. 95% confidence intervals are given for the estimates. To calculate intra-rater reliability, ten patient records from the Graduate Orthodontic Clinic of the University of Toronto were randomly selected as the sample for error analysis. An orthodontic examination was performed on 10 different patients at 2 36

47 different time points, 6 months apart, using the same full orthodontic methodology (photographs, radiographs and study models). After 6 months, the orthodontic examination was repeated on the same 10 patients using the same full orthodontic records. Intra-rater reliability was calculated using the student s t-test for linear measurements and percent agreement for categorical measurements. Page 37 Data (for the primary analysis) are presented as the mean + standard deviation for continuous variables and as percentages for categorical variables. Independent Students t tests were used to compare continuous data and Analysis of Variance (ANOVA) was used to compare differences between multiple (more than 2) groups. Chisquare test was used to compare the categorical data. ROC analysis was used to find the area under the curve, sensitivity, specificity and the confidence around them. Odds ratios (OR) associated with the presence or absence of characteristics and mean values with 95% CI values were also calculated. Univariate and Multiple Logistic regression was used to assess the relationship between independent predictor variables and binary outcomes (OSA). Variables were considered significant at the 5% significance level. Data were analyzed using SAS/STAT Software, version 9.4 (North Carolina). 2.8 Study Outcomes The primary outcome was the prevalence of dentofacial abnormalities and malocclusions in a cohort of children with and without obesity who were referred for a polysomnogram because of a history of query OSA. Our secondary outcome measure was the identification of clinical factors that can predict the obstructive apnea-hypopnea index (OAHI), a measure of the OSA severity, in this referred cohort of children. 2.9 Hypothesis Our study hypothesis was that there will be an increased prevalence of dentofacial abnormalities and malocclusions in children with and without obesity with a PSG diagnosis of OSA as compared to those without a PSG diagnosis of OSA. 37

48 Page 38 Chapter 3 Results 3.1 Intra-rater Reliability A total of 29 data collection points were used to determine the intra-rater reliability. The results for percent agreement for categorical and continuous variables between the two different time points were assessed using kappa and intraclass correlation coefficients (ICC), respectively (see Table 3.1). To interpret our results we used a benchmark cut-off proposed by Landis and Koch: 153 Cohen s kappas 0.80 represent excellent agreement; coefficients between 0.61 and 0.80 represent substantial agreement; coefficients between 0.41 and 0.61 moderate agreement; and <0.41 represent fair to poor agreement. Table 3.1 Intra-rater reliability. Percent agreement, Kappa and Intraclass Correlation Coefficient of repeated orthodontic examination measurements recorded 6 months apart Measurement Percent Agreement (%) kappa Profile Symmetry Anterior Openbite Crossbite Maxillary Teeth Involved Posterior Openbite Stage of Dentition Permanent Molar Classification Right Permanent Molar Classification Left Spacing Mandible Spacing Maxilla Skeletal Position Maxilla Upper Lip with respect to E-Line

49 Page 39 Lower Lip with respect to E-line Facial Type Mandibular Arch Shape Canine Classification Right Canine Classification Left Upper Dental Midline Lower Dental Midline Skeletal Position Mandible Maxillary Arch Shape Nasiolabial Angle Lower Face Height Narrow Palate ICC 95% Confidence Interval Lower Bound Upper Bound Overjet (mm) Overbite (% overlap of incisors) Intermolar Distance (mm) Intercanine Distance (mm) Overall, the agreement for the categorical variables assessed ranged from 0.35(poor agreement) to 1.0 (excellent agreement). Profile, symmetry, anterior openbite, crossbite, posterior openbite, stage of dentition and molar classification had the highest Cohen s kappa (k=1.0), while upper dental midline and narrow palate had the lowest (k= 0.56 and k= 0.35). All continuous measurements had excellent agreement (ICC ranging from ). 3.2 Study Participants One hundred and two children were screened for the study. Two patients declined study participation. A reason for declining consent was not given. One hundred children with a mean (standard deviation) age 10.5 (SD 3.8) years participated in the study over 39

50 the recruitment period of 13 months (March April 2016). The subjects were divided into five different groups (see Table 3.2) based on: 1) PSG diagnosis of OSA, 2) CDC weight status category (BMI percentile range) and 3) history of treatment with PAP therapy. Page 40 Based on the PSG findings, subjects were divided into an OSA group (OAHI >1.5 events per hour) and a non-osa group (OAHI <1.5 events per hour). On the basis of weight status category, BMI- for-age percentile growth charts were used to divide subjects into a non-obese group (BMI< 95 th percentile) and an obese group (BMI > 95 th percentile). The fifth group included children that were prescribed PAP therapy. Table 3.2 Subject Groups Group # Group Category 1 Non-Obese and No OSA 2 Non-Obese and OSA 3 Obese and No OSA 4 Obese and OSA 5 PAP treatment See Table 3.3 for the demographic information for the four study cohorts, excluding the PAP treatment group. The mean BMI, BMI percentile, height, weight, and presence of mouth breathing were significantly different between the cohorts. The Non- Obese and OSA group had the highest percentage of snorers (90.7%), mouth breathers (100%), and children with increased tonsillar size> 3 (50%). However, mouth breathing was not statically significant between the cohorts (p=0.062). 40

51 Page 41 Table 3.3 Demographics of the Four Study Cohorts (excluding PAP group) Non-Obese and No OSA Non-Obese and OSA Obese and No OSA Obese and OSA P value Sample Size (n) Age (years) 9.4 (3.6) 8.18(4.69) 10.93(3.74) 11.0(3.92) 0.13 Male (%) 17 (81) 7(63.6) 16(57.1) 21(77.8) 0.12 BMI (kg/m 2 ) (3.85) (3.17) (8.25) 33.39(9.73) < BMI Centile (20.45) (26.63) (1.13) 98.11(1.42) < Height (cm) (22.24) (25.01) (17.34) (20.10) Weight (kg) (20.26) 31.0 (19.58) 70.73(31.59) (33.31) < Snoring 14 (66.7) 10 (90.9) 22(78.6) 24(88.9) 0.20 Mouth Breather 15(71.4) 11 (100) 22(78.6) 15(55.6) Increased Tonsillar Size >3 6(28.6) 5(50) 3(10.7) 5(18.5) 0.06 * Statistics shown for categorical variables are n (%) from the population with available data as continuous variables are mean (SD), unless otherwise specified. Table 3.4 compares the demographic information for the subjects with and without OSA (excluding the PAP group). When the subjects were divided based on OSA diagnosis, there were no significant demographic differences between the groups. 41

52 Page 42 Table 3.4 Demographics of OSA vs. No OSA Groups (excluding PAP group) OSA (n=38) No OSA (n=49) p-value Age (Years) 10.2(4.3) 10.3(3.7) 0.91 Male 28(73.7) 33(67.4) 0.52 BMI (kg/m 2 ) 28.69(11.17) 25.82(8.64) 0.19 BMI Centile 85.13(24.85) 88.55(17.30) 0.45 Height (cm) 144.2(23.58) 144.2(20.55) 0.99 Weight (kg) 66.51(41.54) 57.23(31.31) 0.26 Snoring 34(89.5) 36(73.5) 0.06 Mouth Breather 26(68.4) 37(75.5) 0.46 Increased Tonsillar Size >3 10(27.0) 9(18.4) 0.34 * Statistics shown for categorical variables are n (%) from the population with available data as continuous variables are mean (SD), unless otherwise specified. 3.3 Polysomnography Results See Table 3.5 for a summary of the PSG findings across the four cohorts. Significant findings included sleep state distribution %N1 sleep, sleep stage distribution % REM sleep, total arousal index, respiratory events arousal index, oxygen desaturation index, maximum respiratory rate, mean respiratory rate, maximum transcutaneous carbon dioxide (tcco2), percent of sleep time with end-tidal carbon dioxide (EtCO2) above 50mmHg,OAHI index and AHI index. The OAHI and AHI were significantly different across the four groups and were the highest in the Obese and OSA group with a mean (SD) OAHI of 12.31(15.42), (p=<0.0001) and 13.15(15.33), (p=<0.0001), respectively. Table 3.5 PSG Results across the Four Cohorts (excluding PAP group) Non-Obese Non-Obese Obese and No Obese and OSA P value and No OSA and OSA OSA Total Sleep Time (TST) (minutes) (45.6) (66.7) (113.5) (63.1) 0.14 Sleep Efficiency (%) 84.8 (9.1) 87.0 (10.1) 81.8 (16.3) 82.9 (11.2)

53 Sleep stage distribution % N1 4.5 (3.0) 4.9 (3.5) 5.5 (3.3) 7.6 (4.9) 0.03 Sleep stage distribution % N (3.0) 46.6 (4.3) 46.6 (11.9) 49.1 (9.0) 0.74 Sleep stage distribution % N (3.0) 28.1 (7.1) 29.9 (9.8) 27.8 (9.1) 0.80 Sleep stage distribution % REM 19.4 (3.1) 20.5 (6.3) 18.0(5.0) 15.5 (6.6) 0.04 Wake after sleep onset (WASO), (minutes) 36.9 (3.2) 31.4 (31.4) 38.5 (59.5) 53.0 (42.4) 0.45 Sleep onset latency, (minutes) 29.4 (3.2) 21.7 (21.7) 37.3 (27.1) 25.7 (30.1) 0.37 REM latency, (minutes) (53.7) (73.7) (72.8) (59.2) 0.55 Total arousal index (# events/ hour) Spontaneous arousal index (#events/hour) Respiratory events arousal index (#events/ hour) 9.8 (2.3) 15.4 (4.5) 9.3 (4.2) 17.6 (8.1) < (13.3) 8.6 (3.4) 7.9 (3.5) 9.2 (3.9) (.6) 6.5 (4.7) 0.8 (0.8) 7.9 (7.7) < Oxygen saturation mean (%) 97.8 (0.6) 97.4 (0.9) 97.9 (0.9) 97.1 (2.6) 0.22 Oxygen saturation minimum, (%) 91.5 (2.8) 87.6 (9.7) 92.3 (3.9) 86.5 (13.0) 0.05 Oxygen desaturation index, (# events/hour) Time spent 90% oxygen saturation (minutes) 0.8 (0.6) 4.5 (6.0) 0.8 (0.8) 10.0 (22.0) (0.06) 2.0 (5.7) 0.0 (0.1) 10.2 (36.5) 0.24 Respiratory rate mean (bpm) 17.0 (2.0) 15.0 (5.1) 16.9 (2.4) 18.3 (3.3) 0.03 Respiratory rate minimum (bpm) 13.5 (2.2) 13.8 (2.3) 13.8 (2.4) 14.6 (2.4) 0.41 Respiratory rate maximum (bpm) 20.6 (2.7) 21.1 (1.6) 20.7 (3.7) 23.9 (4.5) Heart rate mean (bpm) 80.0 (12.3) 79.6 (14.9) 75.2 (11.4) 78.0 (10.1) 0.13 Heart rate minimum (bpm) 53.7 (7.7) 58.0 (11.2) 55.2 (7.8) 57.7 (8.4) 0.32 Heart rate maximum (bpm) (13.6) (15.1) (20.4) (11.1) 0.20 EtCOⁿ minimum (mmhg) 33.4 (4.7) 30.9 (3.4) 31.2 (5.5) 29.8 (9.3) 0.39 EtCOⁿ maximum (mmhg) 49.7 (2.5) 50.2 (6.7) 50.2 (3.4) 52.3 (7.7) 0.42 TcCOⁿ minimum (mmhg) 35.1 (4.2) 37.3 (11.0) 33.7 (4.9) 34.2 (9.1) 0.54 TcCOⁿ maximum (mmhg) 47.1 (4.0) 55.5 (19.34) 47.7 (4.3) 49.0 (6.7) % TST EtCO2 > 50 mmhg 0.4 (0.4) 29.3 (0) 1.8 (3.5) 4.8 (12.8) 0.03 % TST TcCO2 > 50 mmhg 14.1 (15.7) 28.7 (0) 4.9 (10.9) 14.6 (22.1) 0.44 CAI (#events/ hour) 0.62 (0.46) 0.87 (1.34) 0.66 (0.77) 0.67 (0.91) 0.88 OAHI (#events/ hour) 0.40 (0.63) 9.46 (9.20) 0.51 (0.59) (15.42) < AHI (#events/hour) 1.12 (0.71) (9.35) 1.18 (0.88) (15.33) < * Statistics shown for categorical variables are n (%) from the population with available data as continuous variables are mean (SD), unless otherwise specified. TST, total sleep time; REM, rapid eye movement; WASO, wake after sleep onset; bpm, beats per minute; EtCOⁿ, end-tidal carbon dioxide; TcCOⁿ, transcutaneous carbon dixode Page 43 43

54 Table 3.6 summarizes the PSG findings for the subjects with and without OSA (excluding PAP group). The significant differences between these groups were sleep stage distribution % N1 sleep, total arousal index, respiratory events arousal index, minimum oxygen saturation, oxygen desaturation index, maximum respiratory rate, heart rate mean, OAHI and AHI Index. Page 44 Table 3.6 PSG results of OSA vs. No OSA Groups (excluding PAP group) OSA (n=38) No OSA (n=39) P value Total Sleep Time (TST) (minutes) 379.5(67.26) 376.9(93.70) 0.88 Sleep Efficiency (%) 84.09(10.88) 83.08(13.65) 0.71 Sleep stage distribution % N1 6.86(4.65) 5.07(3.21) Sleep stage distribution % N (7.98) 46.60(10.40) 0.40 Sleep stage distribution % N (8.46) 29.75(9.19) 0.32 Sleep stage distribution % REM 16.98(6.79) 18.59(4.65) 0.22 Wake after sleep onset (WASO) (minutes) 46.71(40.37) 37.83(47.85) 0.36 Sleep onset latency, (minutes) 24.56(27.71) 33.88(30.17) 0.14 REM latency, (minutes) 141.0(62.68) 143.5(63.70) 0.86 Total arousal index (#events/hour) 16.99(7.30) 9.48(3.53) < Spontaneous arousal index (#events/hour) 9.02(3.76) 9.55(9.21) 0.72 Respiratory events arousal index (#events/hour) 7.51(6.94) 0.72(0.74) < Oxygen saturation mean (%) 97.16(2.20) 97.86(0.78) 0.07 Oxygen saturation minimum (%) 86.82(12.01) 91.93(3.45) 0.01 Oxygen desaturation index (#events/hour) 8.42(18.91) 0.80(0.71) 0.02 Time spent 90% oxygen saturation (minutes) 7.82(30.96) 0.02(0.06) 0.13 Respiratory rate mean (bpm) 17.35(4.17) 16.90(2.20) 0.55 Respiratory rate minimum (bpm) 14.36(2.36) 13.67(2.31) 0.17 Respiratory rate maximum (bpm) (4.10) 20.67(3.28) Heart rate mean (bpm) 78.50(11.52) 73.35(11.87) Heart rate minimum (bpm) 57.78(9.12) 54.53(7.68)

55 Page 45 Heart rate maximum (bpm) 114.5(13.01) 112.1(17.78) 0.49 EtCOⁿ minimum (mmhg) 30.12(7.91) 32.08(5.23) 0.25 EtCOⁿ maximum (mmhg) 51.65(7.34) 49.99(3.08) 0.26 TcCOⁿ minimum (mmhg) 35.12(9.65) 34.29(4.63) 0.63 TcCOⁿ maximum (mmhg) 50.93(12.01) 47.44(4.16) 0.10 % TST EtCO2 > 50 mmhg 6.72(13.98) 1.50(3.16) 0.21 % TST TcCO2 > 50 mmhg 15.51(21.64) 6.35(11.47) 0.17 CAI (#events/hour) 0.73(1.04) 0.65(0.65) 0.64 OAHI (#events/hour) 11.49(13.85) 0.46(0.60) < AHI (#events/hour) 12.38(13.80) 1.15(0.80) < * Statistics shown for categorical variables are n (%) from the population with available data as continuous variables are mean (SD), unless otherwise specified. TST, total sleep time; REM, rapid eye movement; WASO, wake after sleep onset; bpm, beats per minute; EtCOⁿ, end-tidal carbon dioxide; TcCOⁿ, transcutaneous carbon dioxide 3.4 Questionnaire Results Spruyt and Gozal Questionnaire A total of 77 subjects were included in the statistical analyses for the Gozal and Spruyt questionnaire. 13 subjects were excluded because the patients were using PAP therapy. 10 questionnaires were excluded to due missing data. Table 3.7 shows the total number of subjects who scored >2.72 on the Spruyt and Gozal questionnaire across all four cohorts. The obese groups (No-OSA and OSA) on average had a higher percentage of subjects who scored higher on the questionnaire (13.64% and 14.77%, respectively) than the non-obese groups. However, the Spruyt and Gozal questionnaire scores were not different between the four cohorts (p=0.94). 45

56 Table 3.7 Spruyt and Gozal Questionnaire results across the Four Cohorts (excluding the PAP therapy group) Page 46 Non-Obese and No OSA Non-Obese and OSA Obese and No OSA Obese and OSA P value Total Spruyt and Gozal Score > (4.55) 4(4.55) 12(13.64) 13(14.77) 0.94 * Statistics shown for categorical variables are n (%) from the population with available data as continuous variables are mean (SD), unless otherwise specified The specificity and sensitivity for the Spruyt and Gozal questionnaires for the diagnosis of OSA was calculated from the 2x2 contingency table (see Table 3.8). The Spruyt and Gozal scores were compared for children with and without OSA. Children receiving PAP therapy were excluded. Table 3.8 Frequencies of Spruyt and Gozal Scores of OSA vs. No OSA groups (excluding PAP group) OSA Yes No Spruyt and Gozal Score > 2.72 Yes No From the above table, it was found that this questionnaire was able to correctly identify children who have OSA (AHI>1.5) with a sensitivity of 47.22% and a specificity of 60.98%. Also, the odds ratio of having sleep apnea with a Spruyt and Gozal score of greater than 2.72 was 1.40 (95% CI , p = 0.47). The odds ratio was not significant. The graph in Fig 3.1 shows the receiver operating curve (ROC) for the Spruyt and Gozal questionnaire to correctly diagnose children with and without OSA. The area under the curve was It is evident from the plotted data that the ROC curve closely approximates a straight line. The Spruyt and Gozal questionnaire was a poor screening test for OSA in our study cohort. 46

57 Figure 3.1 ROC Curve for the Spruyt and Gozal Questionnaire to Screen for OSA Page Pediatric Sleep Questionnaire 13 patients from the cohort were excluded because they were using PAP therapy. The remaining 87 subjects were included in the PSQ statistical analysis. Questionnaire scores > 0.33 were considered positive and suggestive of pediatric OSA. This threshold is based on a validation study that demonstrated that the PSQ's optimal sensitivity and specificity for the detection of OSA is at this cutoff Table 3.9 reveals the total number of subjects who scored >0.33 on the PSQ questionnaire for all four cohorts. Children with obesity had the highest proportion of children that were PSQ screen positive but this difference was not significantly difference (p=0.62). 47

58 Table 3.9 Pediatric Sleep Questionnaire Results across the Four Cohorts (excluding PAP group) Page 48 Non-Obese and No OSA Non-Obese and OSA Obese and No OSA Obese and OSA P value Total PSQ Score > (18.4) 9(10.3) 22(25.3) 19(21.8) 0.62 * Statistics shown for categorical variables are n (%) from the population with available data as continuous variables are mean (SD), unless otherwise specified Specificity and sensitivity for the PSQ to detect OSA was calculated from the 2x2 contingency table (see Table 3.10). The PSQ scores were compared for children with and without OSA. Children receiving PAP therapy were excluded. Table 3.10 Frequencies of PSQ Scores of OSA vs. No OSA groups (excluding PAP group) OSA Yes` No PSQ > 0.33 Yes No From the above table, it was found that this questionnaire was able to correctly identify children who have OSA (AHI>1.5) with a sensitivity of 73.88% and a specificity of 22.45%. Also, the odds ratio of diagnosing OSA based on a PSQ score of greater than 0.33 was (95% CI , p = 0.68). However, the odds ratio was not significant. The graph in Fig 3.2 shows the ROC for the PSQ questionnaire to correctly identify children with and without OSA. The area under the curve was It is evident from the plotted data that the ROC curve closely approximates the straight line. The PSQ questionnaire was a poor screening test for OSA in our study cohort. 48

59 Figure 3.2 ROC Curve for the PSQ Score Questionnaire to Screen for OSA Page Dentofacial Morphology Dentofacial characteristics were grouped into variables describing anteriorposterior, transverse, vertical, and perimeter characteristics of the patient s morphology. Table 3.11 summarizes the prevalence of dentofacial morphology characteristics of the four study cohorts. The only statistically significant differences between the four groups were overjet (p=0.02), maxillary intermolar width (p=0.02), and maxillary intercanine width (p=0.0005). 49

60 Table 3.11 Prevalence of Dentofacial Characteristics across the Four Cohorts (excluding the PAP therapy group) ANTERIOR-POSTERIOR Non-Obese Non-Obese Obese and Obese and P value and No OSA and OSA No OSA OSA Convex Profile 28.6 % 9.1 % 28.6 % 29.6 % 0.97 Retrognathic Mandible 28.6 % 9.1 % 32.1 % 29.6 % 0.92 Anterior Crossbite 9.5 % 9.1 % 25.0 % 18.5 % 0.80 Overjet(mm) 2.7 (1.6) 2.3 (1.1) 1.6 (2.0) 3.3 (2.4) 0.02 Class II Molar 25 % 16.7 % 12.5 % 23.5 % 0.66 Class II Canine 40 % 4.6 % 21.4 % 17.0 % 0.62 Distal Step 38.5 % 18.8 % 41.7 % 24.3 % 0.93 TRANSVERSE & VERTICAL Non-Obese and No OSA Non-Obese and OSA Obese and No OSA Obese and OSA P value Dolichocephalic Facial Pattern 0 % 0 % 3.6 % 3.7 % 0.74 Increased Lower Face Height 19.1 % 9.1 % 36.7 % 48.1 % 0.22 Overbite (% overlap of Incisors) 46.7(37.2) 60.0(37.1) 30.9(3.1) 43.7(45.4) 0.21 Anterior Openbite 0 % 9.1 % 10.7 % 0 % 0.16 Posterior Crossbite 14.3 % 2.7 % 17.9 % 11.1 % 0.51 Narrow Palate 9.5 % 18.2 % 39.3 % 18.5 % 0.71 Maxillary Intermolar Width (mm) 36.0 (3.5) 33.9(5.2) 37.7(3.4) 37.8(3.8) 0.02 Maxillary Intercanine Width (mm) 31.2 (2.8) 29.4(3.0) 32.9(2.5) 33.4(3.2) Page 50 PERIMETER Maxillary or Mandibular Crowding > 4mm Non-Obese and No OSA Non-Obese and OSA Obese and No OSA Obese and OSA P value 4.7% 27.3% 32.14% 29.6% 0.36 * Statistics shown for categorical variables are n (%) from the population with available data as continuous variables are mean (SD), unless otherwise specified To better understand the relationship the relationship between SDB and dentofacial morphology, we evaluated the differences between subjects with OSA (AHI >1.5; n=38) as compared to those without OSA. (AHI <1.5; n=49). Children receiving PAP therapy were excluded from this analysis. Table 3.12 summarizes these findings. The only significant dentofacial difference found between OSA and non-osa children was overjet. The mean (SD) overjet in OSA children was 3.0mm(2.14mm) and 2.06mm(1.94mm) in non-osa children (p=0.04). Although there was only one statistically significant difference between the two-groups, there was a trend towards children with OSA having a higher percentage of dolichocephalic facial type, increased lower face height, and mandibular or maxillary crowding >4mm. Linear measurements 50

61 that were higher in children with OSA included overbite, inter-molar distance and intercanine distance. Page 51 Table 3.12 Dentofacial Morphology of OSA vs. No OSA Groups (excluding PAP group) OSA (n=38) No OSA (n=49) P value Dolichocephalic Facial Type 1 (2.6) 1(2.0) 0.74 Increased LFH 14 (36.8) 14(28.6) 0.72 Convex Profile 9 (23.7) 14(28.6) 0.61 Retrognathic Mandible 9 (23.7) 15(30.6) 0.40 Anterior Open Bite 1 (2.6) 3(6.1) 0.63 Anterior Crossbite 6 (15.8) 9(18.4) 0.75 Posterior Crossbite 6 (15.8) 8(16.3) 0.95 Narrow Palate 7 (18.4) 13(26.5) 0.37 Distal Step 7 (18.4) 10(20.4) 0.93 Canine Class II 10 (26.3) 14(28.6) 0.62 Maxillary or Mandibular Crowding >4mm 11 (28.9) 10(20.4) 0.36 Overbite (% overlap of incisors) 48.42(43.34) 37.65(37.57) 0.22 Overjet (mm) 3.0 (2.14) 2.06 (1.94) 0.04 Inter-canine distance (mm) (3.63) 32.16(2.69) 0.91 Inter-molar Distance (mm) (4.54) 36.96(3.49) 0.77 * Statistics shown for categorical variables are n (%) from the population with available data as continuous variables are mean (SD), unless otherwise specified Univariate analysis was then performed with OAHI > 1.5/hr (i.e. positive diagnosis for OSA) as the predictor outcome for the various dentofacial characteristics. Table 3.13 provides the p-values and odds-ratios from the univariate analysis. Overall there was no statistical significant association between AHI and dentofacial characteristics. 51

62 Table 3.13 Univariate Analysis for Various Dentofacial Characteristics (excluding the PAP therapy group) Page 52 Variable Odds Ratio 95% Confidence Interval P -value Dolichocephalic vs. Mesocephalic Facial Type Brachycephalic vs. Mesocephalic Facial Type Convex vs. Normal Profile Type Concave vs. Normal Profile Type Increased Lower Face Height vs. Normal Face Height Decreased Lower Face Height vs. Normal Face Height Retrognathic vs. Normal Mandible Distal Step Class II Canine Anterior Crossbite Posterior Crossbite Narrow Palate Maxillary or Mandibular Crowding >4mm Overbite (% overlap of incisors) Overjet (mm) Furthermore, to determine if an association existed between dentofacial morphology and the use of PAP therapy, children with obesity and recently diagnosed OSA (i.e. not using PAP therapy) were compared to children with obesity and OSA who have been using PAP therapy for a minimum period of at least 1 year. Table 3.14 summarizes the findings between the two groups. There were no statistical differences between the two groups. 52

63 Page 53 Table 3.14 Dentofacial Morphology of Obese & OSA vs. Obese & PAP groups Obese and OSA (n=27) Obese CPAP (n=13) P-Value Dolichocephalic Facial Type 1(3.70) 0(0) 0.99 Increased LFH 13(48.2) 8(61.5) 0.51 Convex Profile 8(29.6) 5(38.5) 0.58 Retrognathic Mandible 8(29.6) 6(46.2) 0.30 Anterior Open Bite 0(0) 0(0) 0.99 Anterior Crossbite 5(18.5) 6(46.2) 0.13 Posterior Crossbite 3(11.1) 4(30.1) 0.13 Narrow Palate 5(18.5) 2(16.67) 0.89 Distal Step 5(50.0) 1(100) 0.34 Canine Class II 18(66.7) 9(69.23) 0.87 Maxillary or Mandibular Crowding >4mm 3(11.11) 4(31) 0.12 Overbite (% overlap of incisors) 43.70(45.41) 33.08(45.35) 0.49 Overjet (mm) 3.30(2.399) 1.69(2.29) 0.06 Inter-canine distance (mm) 33.41(3.21) 34.38(3.92) 0.44 Inter-molar Distance (mm) 37.85(3.80) 40.31(4.66) 0.11 * Statistics shown for categorical variables are n (%) from the population with available data as continuous variables are mean (SD), unless otherwise specified A multiple regression model was developed using generated variables with p- values < 0.05 from the univariate analyses. Table 3.15 demonstrates the multiple regression analysis. 53

64 Table 3.15 Multiple Regression Model for the Presence of OSA in the Study Cohort Excluding Children Using PAP Therapy Page 54 Variable Odds Ratio 95% Confidence Interval P -value Overjet (mm) Respiratory Rate Max BMI Total Arousal Index Overjet, maximum respiratory rate, and total arousal index were significant predictors of OSA as demonstrated form the multiple regression analysis. The odds ratio of having an increased overjet and OSA was (95% CI , p = 0.049). The odds ratio of having an increased maximum respiratory rate and OSA was (95% CI , p = 0.03). The odds ratio of having an increased total arousal index was (95% CI , p = ). BMI was not a significant predictor for OSA. 54

65 Page 55 Chapter 4 Discussion We are reporting on the first pediatric study to systematically describe the prevalence of dentofacial characteristics in a referred cohort of children with and without PSG proven OSA using the currently recommended AASM guidelines. A key finding of our study was that the prevalence of an overjet (i.e. horizontal excess) was significantly higher in the children with OSA as compared to those without OSA. Furthermore, the presence of an overjet, maximum respiratory rate, total arousal index all significantly predicted the presence of OSA in our study cohort. We evaluated dentofacial morphology in three planes: vertical, transverse, and anterioposterior. When the study population was subdivided into four groups based on a diagnosis of OSA and CDC BMI centile criteria for obesity, three dentofacial characteristics were significantly different between the groups. These included overjet (i.e. horizontal excess), maxillary intermolar width and maxillary intercanine width. However, after the study cohort was subdivided based on an OSA diagnosis, only the presence of an overjet remained clinically significant. This increase in overjet may be explained by the presence of a Class II skeletal pattern and/or dental pattern, which has found to be more prevalent in OSA children versus healthy controls 142. Also, an increased overjet in OSA children can be explained by long-term changes in the position of the head, mandible, and tongue in order to maintain airway adequacy during sleep 142. Our finding of an increased overjet is consistent with previous studies on the effects on OSA and dentofacial morphology. Pirilä-Parkkinen et al. 145 conducted a similar study in 2008, looking at the effects of SDB on developing dental arches. Their findings, like ours, found children with OSA had a significantly increased overjet, however, they also found a reduced overbite, narrower upper arch and shorter lower dental arch when compared with the controls. The difference in findings may be explained by the fact that their protocols followed the older guidelines published by the American Thoracic Society in 1996 for scoring OSA diagnosed by PSG. The use of older guidelines for scoring OSA may result in over- or under-diagnosis of OSA, leading to different results based on OSA criteria and diagnosis. Our study used the most up-to-date 55

66 guidelines published by the AASM 88 and found overjet as the only significant dentofacial predictor of OSA. Katyal et al 39 conducted a systematic review in 2013 on dentofacial morphology using lateral cephalograms in pediatric OSA. Katyal et al. found an increase in weighted mean differences in the ANB angle of 1.64 degrees (P<0.0001) and 1.54 degrees (P< ), respectively, in children with OSA and primary snoring, compared with the controls. The authors concluded that an increased ANB angle of less than 2 degrees in children with OSA and primary snoring, compared with the controls, could be regarded as having marginal clinical significance. Though our study did not look at cephalometric measurements and ANB angles, our finding of increased overjet was found to be significantly associated with a higher AHI, but it is not yet clear if this result is clinically significant. An interesting finding of our study was the incidence of convex profile, retrognathic mandible, anterior open bite, anterior crossbite, posterior crossbite, narrow palate, distal step and Class II canine were higher in the non-osa group when compared with OSA group. This is unexpected as based on the results of the systematic review, the prevalence of these dentofacial characteristics would be expected to be higher in the group with OSA. However, our findings may be explained by the fact that pediatric OSA is a multifactorial disease and that craniofacial morphology is frequently not the only factor contributing to the disease process 154. Although, PAP therapy is an effective therapy for OSA in children, there are some notable long-term sequelae that have been reported in the literature. The craniofacial skeleton in the growing child is responsive to changing functional demands and environmental factors. Orthopedic modification of facial bones through the sustained application of near-constant forces over long periods of time has been a mainstay of orthodontic and dentofacial orthopedic therapy 155. The successful use of PAP therapy requires the application of such forces to the midface area. This prompts concern about potential side effects on antero-posterior skeletal development in that area. The possible effect of PAP therapy might depend on the skeletal age at which treatment begins. The rate of normal forward and downward displacement of the midface varies with age 129. A review of the literature reveals that greater skeletal changes in the midface are possible in younger patients and children because the early mixed dentition is particularly vulnerable 56 Page 56

67 to the effects of PAP therapy on the development of the face, jaw, and teeth 88,156,157. In our study, the ages of children receiving PAP therapy ranged from 4 to 16 years of age. However, all the subjects, with the exception of one child, were between 10 to 16 years of age. Thus the majority of children were in the late-mixed to permanent dentition phase. Page 57 Case reports have suggested that long-term treatment with CPAP or BPAP during early childhood carries a high risk of facial growth impairment, in particular, midface hypoplasia and Class III malocclusions 127,128. However, more recently, a small sample size, cross-sectional study failed to show any statistically significant difference between long-term PAP use and craniofacial morphologic pattern in children with persistent OSA 129. Our study was in line with this most recent cross-sectional study. We did not find any statistically significant differences between the children with obesity not using PAP therapy and children with obesity, using PAP therapy. However, all of the children prescribed PAP therapy were greater than 10 years of age (with the exception of one subject), and, therefore, the majority of the skeletal structures have already developed. Interestingly, anterior crossbite was significantly more prevalent in the PAP group (46.2% vs. 18.5%), though not statistically significant (p=0.13). This can be explained by the increased force of the facemask on the anterior teeth. This is an area of future study to determine of the anterior crossbite continues to progress in this cohort with subsequent years of PAP therapy usage. The intra-rater reliability of the orthodontic examinations for dentofacial characteristics demonstrated excellent agreement for the majority of the measurements with (K 0.80). The intra-rater reliability was poor for narrow palate (k=0.35). The continuous measurements for overjet, overbite, intermolar and intercanine distance had excellent agreement with ICC ranging from Therefore, based on our results, the orthodontic clinical examination seems to be a quick, reliable assessment that could be done in any busy sleep medicine clinic. Furthermore, there is the potential to translate this clinical examination traditionally performed by skilled dentists into a screening examination that could be performed by sleep medicine clinicians. The high reliability of the clinical dental examination has been previously described in the literature. Dwokrin et al. 158 demonstrated excellent reliability of assessing molar classification in adults (K=0.78), We reported perfect agreement 57

68 (K=1.0) for the molar classification in the permanent dentition. A possible explanation for this discrepancy is that Dworkin et al. used a more rigorous scale divided into ¼ cusp increments to diagnose molar classification whereas out study used ½ cusp increments. Carvalho et al. has also previously reported the inter- and intra-observer agreement for diagnosis of dental malocclusion; the Cohen s kappa coefficients ranged from (overbite 4mm, yes or no), to 1.0 (openbite, yes or no) 159. Page 58 With regards to orthodontic diagnosis and classification, it is possible that several indicators of malocclusion (eg. molar classification, canine classification, overjet, overbite) which directly affect the extra-oral facial characteristics (i.e. facial type, profile, etc) may change spontaneously due to difference in mandibular position as determined by the patient and/ or examiner 160. However, since our percent agreement among measurements were based upon orthodontic records as opposed to chair-side examination, the disagreement in measurements between the two different time-points were most likely due to differences in observation rather than differences in examination technique. Our study sample as a whole had a 74% prevalence of snoring, which is much higher than the reported 7 to 28% prevalence of primary snoring in the literature. 88 However, our study cohort was a referred population of children with suspected OSA rather than a population based cohort. More specifically, our sample had a higher percentage of snorers in both obese groups and OSA groups, of which a significant difference was found. When analyzing the overall score for the 6-item questionnaire to predict OSA, there was a higher prevalence of predictive scores in the obese groups (OSA and Non-OSA) than the non-obese groups (OSA and non-osa). However, the sensitivity and specificity of the Spruyt and Gozal questionnaire used in our study was 47.22% and 60.98%, respectively, making it a poor predictor of OSA. This is somewhat comparable to the sensitivity, 59.03%; and specificity, 82.85%, reported by Spruyt et al 94 in their questionnaire validation study. Similar trends were seen when analyzing the overall scores from the PSQ. The obese groups (non-osa and OSA) showed a prevalence of 25.3% and 21.8%, respectively, however, with no statistical significance. The sensitivity and specificity was also poor at 73.88% sensitivity and 22.45% specificity. This supports the general consensus in the literature stating that patients 58

69 complaints of snoring alone is insufficient to discriminate apneic and non-apneic snorers In conclusion, our study clearly demonstrates that sleep questionnaires have limited accuracy as screening tools for OSA in children. Page 59 There were no statistically significant differences between the groups in age, gender, and percentage of mouth breathers. In an epidemiological study done in 2008 by Lumeng et al. in children with OSA, the available data appeared insufficient to prove that SDB differs systematically by age 15. This is in line with our findings. Lumeng et al also reported that there is a higher percentages of boys who are affected by SDB (50-100% higher) than girls 15. Though our results are non-significant for gender, there is a much higher percentage of boys in both OSA groups, 63.6% vs. 36.4% (Non-Obese) and 77.8% vs. 22.2% (Obese). The prevalence of mouth breathing in our study was 71%, which is much higher than was Izu. et al 164 found of 42% in OSA children. Though mouth breathing was most prevalent in the non-obese/osa group (100%), there was a high prevalence amongst all the groups. There were a few notable limitations to our study and the results should be interpreted with caution. First, it was a cross-sectional study with a relatively small sample size in each of our cohorts, and significant differences could be identified with a larger sample size.. Our recruitment potential was limited due to the fact that we were only including children with and without obesity that did not have any other comorbidities. Over 80% of the children seen in the sleep center at SickKids have comorbidities. Secondly, we did not have a true 'non snoring' control group as our non OSA comparator group also had symptoms suggestive of OSA warranting a referral to a sleep center but did not have a PSG diagnosis of OSA. Finally, our study used clinical examination as the sole method of evaluating dentofacial morphology. Although, a more objective measure of dentofacial morphology would have been beneficial to directly calibrate the clinical examination, the authors could not justify exposing children to radiation for screening purposes. From the literature, orthodontic examinations have been shown to have high intra- and inter-observer reliability. 145,159 In addition, the author performed all of the orthodontic examinations and was demonstrated to have high intrarater reliability. Therefore, we could not justify the radiation exposure from lateral cephalograms or cone beam Computed Tomograms for our study participants. 59

70 Future studies in this area of research would be to conduct a similar study but with a larger sample size and a true control group, using objective forms of data collection such as lateral cephalograms and/or facial scans. Theoretically this would give more credibility for cause and effect, however, due to the relative scarcity of sleep labs and the long wait times for a polysomnogram, it would be highly unethical to use this resource on healthy children for the purpose of research, while those with signs and symptoms of SDB are expected to wait. Since overjet was the most significant dentofacial predictor of OSA, areas of future study would also include a longitudinal assessment of overjet pre and post adenotonsillectomy. In addition, the severity of the OSA based on PSG could be assessed pre and post treatment of overjet in children with OSA. Finally, a longitudinal assessment of children using PAP therapy starting from a young age (<6 years of age) would be able to better demonstrate if any possible changes occur in craniofacial development from the use of prolonged PAP therapy. Page 60 In summary, the results from the studying of intra-rater reliability of clinical measures of malocclusion and facial characteristics demonstrated excellent agreement for the majority of the measurements. Sleep questionnaires proved to be unsuccessful at predicting the presence of obstructive sleep apnea. OSA was only statistically related to horizontal excess (overjet). All other dentofacial characteristics were not statistically significant. Even though overjet was found to be statistically significant, the clinical significance between the mean difference in overjet between the OSA and Non-OSA groups is yet to be determined. There were no significant differences in dentofacial morphology between children with obesity using PAP therapy and children with obesity not using PAP therapy. Pediatric OSA is a multifactorial disease and craniofacial morphology is not the only factor contributing to the disease process. Thus if a health professional notices signs and symptoms of sleep-disordered breathing, the patient should be referred to a sleep medicine specialist to properly diagnose by PSG, and not rely solely on craniofacial abnormalities, or sleep questionnaires as diagnostic procedures. 60

71 Page 61 Appendices Appendix A: Spruyt and Gozal Sleep Questionnaire 107 Last Name: First Name : Gender: F! M! Date of birth : Month Day Year Age : Over the last 6 months: Please mark each of the following items. Never Rare Occasional Frequent Almost Always (1 night/week) (2 nights/week) (3 to 4 nights/week) (more than 4 nights/week) 1 Do you ever shake your child to make him/her breathe again when asleep?!!!!! 2 Does your child stop breathing during sleep?!!!!! 3 Does your child struggle to breathe while asleep?!!!!! 4 Are you ever concerned about your child's breathing?!!!!! Hardly noticeable Moderately strong Strong Very Strong Extremely Strong 5 How loud is your child snore?!!!!! Never Rare Occasional Frequent Almost Always (1 night/week) (2 nights/week) (3 to 4 nights/week) (more than 4 nights/week) 6 How often does your child snore?!!!!! 61

72 Appendix B: Pediatric Sleep Questionnaire 108 Page 62 PEDIATRIC SLEEP QUESTIONNAIRE Version Child s Name:,. (Last) (First) (M.I.) Name of Person Answering Questions:. Relation to Child:. Your phone number, days:, and evenings:. Area Code Number Area Code Number Relative s name and number in case we cannot reach you:.. Area Code Number Instructions: Please answer the questions on the following pages regarding the behavior of your child during sleep and wakefulness. The questions apply to how your child acts in general, not necessarily during the past few days since these may not have been typical if your child has not been well. If you are not sure how to answer any question, please feel free to ask your husband or wife, child, or physician for help. You should circle the correct response or print your answers neatly in the space provided. A Y means yes, N means no, and DK means don t know. When you see the word usually it means more than half the time or on more than half the nights. 62

73 Page 63 GENERAL INFORMATION ABOUT YOUR CHILD: Office use only GI1 Today s Date:. Month Day Year GI2 Where are you completing this questionnaire?. GI3 Date of Child s Birth:. Month Day Year GI4 Sex: Male or Female?. GI5 Current Height (feet/inches) :. GI6 Current Weight (pounds) :. GI7 Grade in school (if applicable):. GI8 Racial/Ethnic Background of your Child (please circle): GI9 1.) American Indian 2.) Asian-American 3.) African-American 4.) Hispanic 5.) White/not Hispanic 6.) Other or unknown 63

74 Page 64 A. Nighttime and sleep behavior: Office use only WHILE SLEEPING, DOES YOUR CHILD ever snore? Y N DK A1 snore more than half the time? Y N DK A2 always snore? Y N DK A3 snore loudly? Y N DK A4 have heavy or loud breathing? Y N DK A5 have trouble breathing, or struggle to breathe? Y N DK A6 HAVE YOU EVER seen your child stop breathing during the night? If so, please describe what has happened: Y N DK A7 been concerned about your child s breathing during sleep? Y N DK A8 had to shake your sleeping child to get him or her to breathe, or wake up and breathe? Y N DK A9 seen your child wake up with a snorting sound? Y N DK A11 DOES YOUR CHILD have restless sleep? Y N DK A12 describe restlessness of the legs when in bed? have growing pains (unexplained leg pains)? have growing pains that are worst in bed? WHILE YOUR CHILD SLEEPS, HAVE YOU SEEN brief kicks of one leg or both legs? repeated kicks or jerks of the legs at regular intervals (i.e., about every 20 to 40 seconds)? Y N DK Y N DK Y N DK Y N DK Y N DK A13 A13a A13b A14 A14a AT NIGHT, DOES YOUR CHILD USUALLY become sweaty, or do the pajamas usually become wet with perspiration? Y N DK A15 get out of bed (for any reason)? Y N DK A16 get out of bed to urinate? If so, how many times each night, on average? Y N DK times A17 A17a

75 Does your child usually sleep with the mouth open? Y N DK A21 Page 65 Is your child s nose usually congested or stuffed at night? Y N DK A22 Do any allergies affect your child s ability to breathe through the nose? Y N DK A23 DOES YOUR CHILD tend to breathe through the mouth during the day? Y N DK A24 have a dry mouth on waking up in the morning? Y N DK A25 complain of an upset stomach at night? Y N DK A27 get a burning feeling in the throat at night? Y N DK A29 grind his or her teeth at night? Y N DK A30 occasionally wet the bed? Y N DK A32 Has your child ever walked during sleep ( sleep walking )? Y N DK A33 Have you ever heard your child talk during sleep ( sleep talking )? Y N DK A34 Does your child have nightmares once a week or more on average? Y N DK A35 Has your child ever woken up screaming during the night? Y N DK A36 Has your child ever been moving or behaving, at night, in a way that made you think your child was neither completely awake nor asleep? If so, please describe what has happened: Y N DK A37 Does your child have difficulty falling asleep at night? Y N DK A40 How long does it take your child to fall asleep at night? (a guess is O.K.) At bedtime does your child usually have difficult routines or rituals, argue a lot, or otherwise behave badly? minutes Y N DK A41 A42 DOES YOUR CHILD bang his or her head or rock his or her body when going to sleep? Y N DK A43 wake up more than twice a night on average? Y N DK A44 have trouble falling back asleep if he or she wakes up at night? Y N DK A45 wake up early in the morning and have difficulty going back to sleep? Y N DK A46 Does the time at which your child goes to bed change a lot from day to day? Y N DK A47

76 Does the time at which your child gets up from bed change a lot from day to day? Y N DK A48 Page 66 WHAT TIME DOES YOUR CHILD USUALLY go to bed during the week? go to bed on the weekend or vacation? get out of bed on weekday mornings? get out of bed on weekend or vacation mornings? A49 A50 A51 A52 B. Daytime behavior and other possible problems: Office Use Only DOES YOUR CHILD wake up feeling unrefreshed in the morning? Y N DK B1 have a problem with sleepiness during the day? Y N DK B2 complain that he or she feels sleepy during the day? Y N DK B3 Has a teacher or other supervisor commented that your child appears sleepy during the day? Y N DK B4 Does your child usually take a nap during the day? Y N DK B5 Is it hard to wake your child up in the morning? Y N DK B6 Does your child wake up with headaches in the morning? Y N DK B7 Does your child get a headache at least once a month, on average? Y N DK B8 Did your child stop growing at a normal rate at any time since birth? If so, please describe what happened: Y N DK B9 Does your child still have tonsils? If not, when and why were they removed?: Y N DK B10 HAS YOUR CHILD EVER had a condition causing difficulty with breathing? If so, please describe: Y N DK B11 had surgery? If so, did any difficulties with breathing occur before, during, or after surgery? Y N DK Y N DK B12

77 B12a Page 67 become suddenly weak in the legs, or anywhere else, after laughing or being surprised by something? Y N DK B13 felt unable to move for a short period, in bed, though awake and able to look around? Y N DK B15 Has your child felt an irresistible urge to take a nap at times, forcing him or her to stop what he or she is doing in order to sleep? Has your child ever sensed that he or she was dreaming (seeing images or hearing sounds) while still awake? Does your child drink caffeinated beverages on a typical day (cola, tea, coffee)? If so, how many cups or cans per day? Does your child use any recreational drugs? If so, which ones and how often?: Y N DK Y N DK Y N DK cups Y N DK B16 B17 B18 B18a B19 Does your child use cigarettes, smokeless tobacco, snuff, or other tobacco products? If so, which ones and how often?: Y N DK B20 Is your child overweight? Y N DK B22 If so, at what age did this first develop? years B22a Has a doctor ever told you that your child has a high-arched palate (roof of the mouth)? Y N DK B23 Has your child ever taken Ritalin (methylphenidate) for behavioral problems? Y N DK B24 Has a health professional ever said that your child has attention-deficit disorder (ADD) or attention-deficit/hyperactivity disorder (ADHD)? Y N DK B25

78 C. Other Information Page If you are currently at a clinic with your child to see a physician, what is the problem that brought you? 2. If your child has long-term medical problems, please list the three you think are most significant Please list any medications your child currently takes: Medicine Size (mg) or amount per dose Taken when? Effect:. Effect:. Effect:. Effect:.

79 4. Please list any medication your child has taken in the past if the purpose of the medication was to improve his or her behavior, attention, or sleep: Page 69 Medicine Size (mg) or amount per dose Taken how often? Dates Taken Effect:. Effect:. Effect:. Effect:. 5. Please list any sleep disorders diagnosed or suspected by a physician in your child. For each problem, please list the date it started and whether or not it is still present. Please list any psychological, psychiatric, emotional, or behavioral problems diagnosed or suspected by a physician in your child. For each problem, please list the date it started and whether or not it is still present.

80 Page Please list any sleep or behavior disorders diagnosed or suspected in your child s brothers, sisters, or parents: Relative Condition D. Additional Comments: Please use the space below to print any additional comments you feel are important. Please also use this space to describe details regarding any of the above questions. Instructions: Please indicate, by checking the appropriate box, how much each statement* applies to this child: This child often does not seem to listen when spoken to directly. Does not apply 0 Applies just a little 1 Applies quite a bit 2 Definitel y applies most of the time 3 has difficulty organizing tasks and activities. is easily distracted by extraneous stimuli. fidgets with hands or feet or squirms i is on the go or often acts as if driven by a motor. interrupts or intrudes on others (e.g., butts into conversations or games. * Derived from DSM-IV. THANK YOU

81 Page 71 Appendix C Pediatric Polysomnogram Set up

82 Appendix D Pediatric Polysomnogram Data Recording Page 72