Structural equation modeling of sleep apnea, inflammation, and metabolic dysfunction in children KAREN A. WATERS 1,2,3 , BENJAMIN T. MAST 4 , SILVANO VELLA 2 , ROLAND DE LA EVA 2 , LOUISE M. O’BRIEN 1,4 , SHERRYN BAILEY 2 , CHARMAINE S. TAM 2 , MELANIE WONG 5 and LOUISE A. BAUR 2,3 1 Department of Pediatrics, Division of Pediatric Sleep Medicine and Kosair ChildrenÕs Hospital Research Institute, University of Louisville, Louisville, KY, USA, 2 SIDS and Sleep Apnea Research, The ChildrenÕs Hospital at Westmead, 3 Faculty of Medicine, The University of Sydney, NSW, Australia, 4 Department of Psychological and Brain Sciences, University of Louisville, Louisville, KY, USA and 5 Department of Immunology, The ChildrenÕs Hospital at Westmead, NSW, Australia Accepted in revised form 31 July 2007; received 13 April 2007 SUMMARY Obstructive sleep apnea (OSA), often concomitant with obesity, increases the risk for the metabolic syndrome. One mechanism that may participate in this association is upregulation of inflammatory pathways. We used structural equation modeling to assess the interrelations between childhood obesity, OSA, inflammation, and metabolic dysfunction. One hundred and eighty-four children (127 boys, mean age: 8.5 ± 4.1years) had height and weight measured, underwent overnight polysomnog- raphy and had fasting blood taken. The blood was analyzed for insulin, glucose, lipids, leptin, and cytokines [interferon (IFN)-c, granulocyte macrophage–colony stimulating factor, interleukin (IL)-1b, IL-2, IL-4, IL-6, IL-8, IL-10, IL-12, tumor necrosis factor- a]. Structural equation modeling (SEM) was used to evaluate associations between the outcomes of interest including hypoxia, arousal (related to respiratory and spontane- ous), obesity, metabolic dysfunction, and inflammatory markers. Two cytokine factors and one metabolic factor were derived for the SEM. These factors provided good fit in the structural equation model (v 2 ⁄ df = 2.855; comparative fit index = 0.90, root mean squared error of approximation = 0.10) and all factor loadings were significantly different from zero (P £ 0.01). Overall, our results indicate that while obesity (as measured by body mass index z-score) has a major influence on the metabolic dysfunction associated with OSA, arousal indices, and cytokine markers may also influence this association. Our results support the hypothesis that OSA is a contributor to the mechanisms that link sleep, systemic inflammation and insulin resistance, and show that the interrelations may begin in childhood. keywords children, inflammation, metabolic syndrome, obstructive sleep apnea, structural equation model INTRODUCTION The term Ômetabolic syndromeÕ is used to identify a clustering of risk factors that indicates an increased risk for morbidity and mortality associated with cardiovascular disease. Although there is some debate regarding which factors are obligatory and which cut-points are used in the definition of the syndrome, this most commonly refers to the clustering of obesity, hypertension, insulin resistance (type 2 diabetes), heart disease, and dyslipidemia (Eckel et al., 2005). Independent associations have been demonstrated between obstructive sleep apnea (OSA) in adults and metabolic and cardiovascular complications (Caples et al., 2007; Kasasbeh et al., 2006). Similar associations have also been demonstrated in children (NG et al., 2005; Srinivasan et al., 2006); children Correspondence: Karen A. Waters, MBBS, PhD, Kosair ChildrenÕs Hospital Research Institute, University of Louisville School of Medicine, 571 S. Preston Street Suite 321, Louisville, KY 40202, USA. Tel: +61-2-9351-5165; fax: +61-2-9550-3851; e-mail: kaw@ med.usyd.edu.au J. Sleep Res. (2007) 16, 388–395 388 Ó 2007 European Sleep Research Society