Natural History of Sleep-Disordered Breathing: Shedding Light on the Early Years

Abstract

THE NUMBER OF PUBLICATIONS ON THE PREVALENCE, RISK FACTORS, AND OUTCOMES FOR SLEEP DISORDERED BREATHING (SDB) IN ADULTS HAS GREATLY increased in the last decade. Population-based cohort studies of SDB, begun decades ago, are now yielding information on the natural history of this prevalent condition, including the association of untreated SDB with significant health outcomes. The long-awaited point at which these data can inform clinical decision-making and public policy to decrease the burden of SDB appears within reach. Yet the picture is incomplete: SDB develops over time, but the major research focus on middle-aged adults has skipped the early chapters and jumped to the middle of the story. Only recently have researchers taken on the much needed task of documenting the prevalence and correlates of SDB in children.1 While questions regarding SDB in children permeate editorials, they remain largely unanswered. Is SDB a different disorder in children? When and how should it be treated? The paper by Bixler and colleagues in the current issue of Sleep2 helps to fill in these early chapters, with data from the largest population-based sleep cohort in children to date. The cohort was assembled using state of the art methods in probability sampling and attention to selection bias, and studied with in-laboratory polysomnography comparable to that used for the adult cohort successfully studied by this research group.3,4 The consistency of the study methods will allow investigation of the role of age in associations of risk factors and outcomes with SDB, rather than age differences that arise as an artifact of differences in methodology. The study is based on a two stage probability sample of elementary school-aged children in Dauphin County, PA. After completion of an extensive survey of sleep habits and health history by parents of 5,740 children, a stratified sample was invited for an in-laboratory overnight protocol that included polysomnography. With an impressive response of 70% of those invited, extensive data, including body habitus, blood pressure, physical exam, and cognitive function were collected on 704 children. The study provides relatively precise point estimates (with 95% confidence intervals) of SDB prevalence in a representative sample of elementary school children that can be extrapolated to other populations and identifies significant correlates of SDB that can inform screening and treatment decisions. The authors conclude that the prevalence of “clinically significant” SDB, defined as an AHI > 5, was 1.2% (0.6–2.2). Among the many potential risk factors evaluated, indicators of excess weight (BMI, waist circumference) and nasal abnormalities, but not tonsil size, were statistically significant predictors of SDB. These findings are intriguing and, as is characteristic of striking findings, will surely generate heated and useful debate that will further the field of pediatric sleep medicine. One area of controversy will surely be the choice of AHI > 5 to define “clinically significant” SDB. This is based on the previously reported finding from this cohort that elevated blood pressure occurred at this SDB severity.5 The sample contains only 8 children with this severity of SDB, however, making confident generalization to the pediatric population problematic, especially as it is unclear that a true threshold effect is present even in this sample. Moreover, the authors do not consider neurobehavioral consequences of SDB in identifying a clinically important threshold, apparently reflecting their finding of no association of SDB with neurocognitive function.6 This finding will need to be reconciled with the results of other community-based studies demonstrating cognitive and behavioral correlates of even milder SDB. We applaud the attempt to identify an evidence-based threshold for clinically significant SDB in children. Unfortunately, there are as yet insufficient data to confidently identify such a threshold. A second area of likely controversy is the finding that tonsil size was not associated with the presence or severity of SDB. The authors rightly point out the absence of randomized clinical trials in support of adenotonsillectomy for treatment of childhood SDB, although clinicians may be more impressed than they by the many case series suggesting effectiveness of this procedure, notwithstanding the higher failure rate in obese children. Evidence from imaging studies demonstrating enlargement of tonsils and adenoids in childhood SDB is not cited by the authors. This includes a study using magnetic resonance imaging, which finds that, compared to an appropriate control group, the region of greatest airway narrowing in young children with SDB is where the tonsils and adenoids overlap, with no significant differences below this level,7 a finding that may explain the failure of clinical examination to detect a significant association of tonsillar size with SDB. Nor are data cited suggesting that adenotonsillar hypertrophy is associated with SDB even among overweight and obese children.8 Moreover, as obesity may contribute to hypoxemia through reduction in recumbent lung volumes, a definition of hypopnea that incorporates a threshold level of hypoxemia may strengthen the association of obesity with SDB relative to other risk factors. Thus, while the present study expands our understanding of the growing importance of obesity in childhood SDB, the authors overreach when they infer from this observational study that weight management and treatment of nasal obstruction should be the primary and secondary treatment considerations in childhood SDB. This perspective may reflect, in part, the age of the subjects studied. Although all are children, any parent can attest to the enormous difference between 5 and 12 years of age. Changes in airway anatomy and physiology are also seen across this age range.9 As the prevalence of obesity increases with age during childhood, it is likely that its relative importance in the etiology of SDB also increases. It would therefore be of great interest to know whether the correlates of SDB in this cohort differ between younger and older subjects. Finally, although the consistency of polysomnographic methodology between this study and the authors' studies in adults allows assessment of differences in risk factors unbiased by methodologic differences, the suitability of a single definition of hypopnea across the pediatric age range might reasonably be questioned. Children are less likely than adults to experience electroencephalographic evidence of arousal with obstructive events,10 and it is likely that as children age their arousal responses increasingly resemble those of adults. To address such questions will likely require samples even larger than those of the present study. These concerns notwithstanding, the present study provides important cross-sectional data on the prevalence and correlates of childhood SDB. It is hoped that the authors will have the opportunity to follow this cohort prospectively to evaluate the natural history and consequences of SDB starting in the formative years. This would contribute not only to an understanding of the incidence, causes, and consequences of SDB in childhood, but may shed light on the relation between childhood SDB and SDB identified later in life. Understanding childhood risk factors for adult SDB and latent periods between the onset of exposure to SDB and the development of adverse outcomes may inform decisions regarding the optimal timing and approaches for SDB screening in both children and adults and guide interventions for SDB treatment or risk factor reduction.