A large body of data has accumulated in recent years showing that obstructive sleep apnea (OSA) poses an additive risk to disorders classically considered to be obesity-related comorbidities. In adults, OSA is fairly consistently associated with greater risk of having elements of metabolic syndrome, including insulin resistance (IR)1 and hypertension.2 As shown in studies of adults, treatment of OSA with continuous positive airway pressure (CPAP) improves insulin sensitivity in most,3 though not all,4 and several meta-analyses have found that CPAP therapy lowers blood pressure (BP) modestly but significantly.5
In children, the data regarding an association between OSA and metabolic syndrome risk factors are less consistent than in adults. Elevation of pro-inflammatory markers is frequent6; however, some studies have shown an association between OSA and IR,7,8 while others failed to find an association.9 Associations between OSA and IR are more likely to be seen in children that are obese than in those that are non-obese10,11 and in adolescents than in prepubertal children.8 Similarly, while many studies have shown that OSA is associated with elevated BP in children independent of obesity,12 some have found no such association.13 Relatively few studies have examined the metabolic impact of the treatment of OSA in pediatric patients. In these studies, results have been somewhat inconsistent, with some studies showing that treatment improves IR9,14 and BP,15 and others—including the multi-center randomized Childhood Adenotonsillectomy Trial (CHAT) study—finding no significant difference following treatment in either BP or IR from baseline.16 Levels of high-sensitivity C-reactive protein (hsCRP), an inflammatory marker, generally improve after adenotonsillectomy (T&A),6 and continuing hsCRP elevation may be a good marker for residual OSA following T&A.17
All of the pediatric treatment studies cited above utilized T&A as the treatment modality for OSA. T&A is generally the first-line treatment for OSA in children18; however, not all children with OSA have adenotonsillar hypertrophy, and therefore T&A is not always the appropriate therapeutic modality. Even among those who do have adenotonsillar hypertrophy and undergo T&A, residual OSA persists in a significant proportion of children following the surgery,19 especially in obese children,14,20 and therefore so may its sequelae. Thus, it is crucial to also examine the metabolic impact of the other mainstay of therapy, CPAP. Only one previously published pediatric study, conducted by Nakra et al., examined the metabolic impact of CPAP treatment on IR in children and adolescents with OSA. Nakra and colleagues found no difference in body mass index (BMI) Z-scores or in whole body insulin sensitivity as derived from an oral glucose tolerance (OGTT) test after 3 months of CPAP, although they did note a decrease in leptin resistance after CPAP therapy.21
Given the relative paucity of treatment studies examining the impact of CPAP on IR, the multi-center longitudinal study published by Katz and colleagues in this issue of Journal of Clinical Sleep Medicine22 represents a welcome addition to the literature. In this study, Katz and colleagues explored the impact of 6–12 months of positive airway pressure therapy (CPAP or bilevel PAP) therapy upon metabolic syndrome elements in an initial group of 25 obese children ages 8–16 years with OSA (of whom 18, or 67%, had severe OSA) who did not meet criteria for adenotonsillectomy. Twenty-two participants had metabolic markers measured at 6 months, and 19 participants had metabolic markers measured at 12 months, an overall attrition rate of 24%. Specific outcome measures examined included BP, hsCRP, the homeostasis model assessment of insulin resistance (HOMA-IR), and an OGTT. The PAP achieved its primary therapeutic purpose of adequate treatment of OSA, defined as achieving a normal apnea-hypopnea index of < 1 event/h of sleep, in all but one of the participants. Over the 6 months of follow-up, 64% of participants demonstrated adequate adherence to PAP therapy. At baseline, 40% of participants had an elevated HOMA-IR percentile, 39% had an abnormal OGTT, 64% had an elevated hsCRP value of > 3 mg/L, 70% had an absence of a normal nocturnal BP dip (of > 10%) in systolic BP, 44% had an absent diastolic BP nocturnal dip, and 44% had BP values in the frankly hypertensive range for age, sex and height. However, no association was found between any metabolic or inflammatory parameter and any measure of sleep-disordered breathing. Following 6–12 months of PAP therapy in the study group as a whole, there were no discernable changes in absolute BP or nocturnal BP dipping, hsCRP, HOMA-IR percentile and no change in OGTT measures. However, in the sub-group of 13 participants who were adherent to PAP, of whom 7 had an elevated HOMA-IR at baseline, 3 out of 7 (43%) experienced normalization of HOMA-IR percentile after 6 months of PAP. There was no difference in the change in BP measurements between baseline and follow-up in either the entire study group or in the PAP-adherent sub-group.22
This is an important study, as the duration of PAP treatment of 6–12 months represents the longest PAP duration study for pediatric OSA published to date in which the impact of therapy on cardiometabolic risk markers was examined. The central finding of this study is that CPAP or bilevel PAP therapy, even in those with a high degree of adherence and adequate treatment of OSA, does not significantly improve associated metabolic abnormalities in children, although trends towards clinically relevant (though not statistically significant) improvements were seen in IR and in systolic BP load. The lack of salutary metabolic outcomes of PAP therapy are consistent with the findings of Nakra et al.,21 as well as with the subset of pediatric studies which found no metabolic impact of T&A, but stand in contrast to the adult studies. Many of the adult studies (though not all) show a significant improvement in BP and IR following PAP therapy. There are a number of potential explanations for this divergence. It should be emphasized that the study group was a heterogenous sample, encompassing a wide pediatric age range and included children of all pubertal stages (including two prepubertal children). The former is a pertinent observation, as the BP examined were absolute values rather than BP percentiles for age, sex and height, for which normal values can be considerably different in such a heterogenous group. Change in absolute value may be less revealing than changes in BP percentile for age. The latter observation is also relevant, as IR differs depending on puberty stage, and because pediatric OSA appears to have two different phenotypes which may confer different degrees of metabolic risk—the one in younger children, which relates primarily to lymphadenoidal hypertrophy and subsequent airway obstruction (though obesity may contribute as well, as in this study's population), and the other in older children and adolescents, which is strongly associated with obesity and is more common in males.23 Given the age range, the study group may have contained a mixture of these two sub-types of OSA. Also, in a limitation that the authors acknowledged, despite a very respectable effort to recruit study participants in 4 different study centers, the initial sample size of 25 remains small and the overall attrition rate was close to 25%. The impact of treatment in larger study population might allow some of the trends observed to achieve statistical as well as clinical significance. Finally, the study population for Katz et al. was predominantly Caucasian (70%); as the degree to which OSA predisposes to sequelae appears to vary in groups of different racial and ethnic backgrounds. OSA appears to be a much stronger risk factor for hypertension in African American versus Caucasian adults24 but poses a lower risk factor for diabetes in African American than in Caucasian adults.25 A multi-ethnic cohort might have allowed more subtle relationships to emerge after PAP therapy.
In summary, the findings of Katz et al.22 affirms that metabolic derangements are common in children that are obese and have OSA, and adds to the contradictory findings in the pediatric literature regarding the uncertain impact of OSA treatment upon cardiometabolic risk markers. This study will serve as a template for future, larger pediatric studies, ideally to be conducted in a multi-ethnic cohort and followed up for an even longer term, to assess whether statistically as well as clinically significant differences are seen.
Dr. Koren has no conflicts of interest to disclose.