Metabolic alterations and systemic inflammation in overweight/obese children with obstructive sleep apnea

Aim and objective Systemic inflammation has been documented in obstructive sleep apnea (OSA). However studies on childhood OSA and systemic inflammation are limited. This study aimed to determine the relation between OSA in overweight/obese children and various inflammatory markers. Material and methods In this cross sectional study, we enrolled 247 overweight/ obese children from pediatric outpatient services. We evaluated demographic and clinical details, anthropometric parameters, body composition and estimation of inflammatory cytokines such as interleukin (IL) 6, IL-8, IL-10, IL-17, IL-18, IL-23, macrophage migration inhibitory factor (MIF), high sensitive C-reactive protein (Hs-CRP), tumor necrosis factor-alpha (TNF-α), plasminogen activator inhibitor-1 (PAI-1) and leptin levels. Overnight polysomnography was performed. Findings A total of 247 children (190 with OSA and 57 without OSA) were enrolled. OSA was documented on polysomnography in 40% of patients. We observed significantly high values body mass index, waist circumference (WC), % body fat, fasting blood glucose (FBG), alanine transaminase (ALT), alkaline phosphate, fasting insulin and HOMA-IR in children with OSA. Inflammatory markers IL-6, IL-8, IL-17, IL-18, MIF, Hs CRP, TNF- α, PAI-1, and leptin levels were significantly higher in OSA patients (p<0.05). There was strong positive correlation of IL-6, IL-8, IL-17, IL-23, MIF, Hs CRP, TNF-A, PAI-1 and leptin with BMI, % body fat, AHI, fasting Insulin, triglyceride, FBG, WC, HOMA-IR, AST and ALT. Conclusion Children with OSA have increased obesity, insulin resistance and systemic inflammation. Further studies are require to confirm our findings and evaluate their utility in diagnosis of OSAs, assessing severity and possible interventions.

Introduction Childhood obstructive sleep apnea (OSA) is characterized by episodic upper airway obstruction that occurs during sleep. OSA is associated with metabolic, cardiovascular and neuropsychological disorders [1]. OSA has been estimated to affect 2-6% of all children [2,3] and up to 59% of obese children [4]. Childhood obesity is a significant risk factor for childhood OSA [1]. In overweight/obese children, the risk of obstructive sleep apnea syndrome (OSAS) is high at 36%, and may be increased 60% if habitual snoring is present. Similarly, Kalra et al. [5] reported that 55% of morbid obese children undergoing bariatric surgery had evidence of OSAS. Genetic and family environment play an important role in childhood obesity. It has been indicated that fat distributions in children and adults have shown visceral adiposity to be strongly associated with OSA. Body fat distribution may explain the relationship between obesity and OSA in children. Therefore other independent genetic and hormonal factors may predispose obese children to develop OSA.
It has been indicated that Inflammation is associated with OSAS. Adenotonsillar hypertrophy is a most common factor for OSAS; however, the cause of the lymphoid tissue enlargement is unknown. Chronic inflammation is also indicated by an increased expression of proinflammatory cytokines and elevated infiltration of macrophages into adipose tissue. Studies have shown that environmental and genetic factors may be initiated an inflammatory response to tonsillar hypertrophy [6]. Regardless, there is ample evidence to show that inflammatory markers [interleukin (IL) 6, IL-8, IL-10, IL-17, IL-18, IL-23, macrophage migration inhibitory factor (MIF), high sensitive C-reactive protein (Hs CRP), tumor necrosis factor-alpha (TNFα), plasminogen activator inhibitor-1 (PAI-1), and leptin] have been shown to be increased in obese pediatric and adult subjects with OSA [7][8][9][10][11][12].
There is paucity of data regarding sub-clinical inflammation in patients with obese and OSA in Asian Indian children. However, till date there was not a specific biomarkers investigation particularly OSA children. In this context, we hypothesized that sub-clinical inflammation may be closely associated with obese subjects with OSA. To test this hypothesis we designed this study to analyze body composition, anthropometry, metabolic profiles and inflammatory cytokines, particularly IL-6, IL-8, IL-10, IL-17, IL-18, IL-23, MIF, Hs CRP, TNF-α, PAI-1 leptin levels in children with and without OSA.

Identification of the target population
A screening of school children was carried out between October 2014 to January 2019. A total of 247 non-diabetic overweight/ obese children were eligible for enrollment in study. 247 overweight/obese subjects [190 with OSA (cases) and 57 without OSA (controls)] included for analysis.
The study was approved by the institutional ethics committee from All India Institute of Medical Sciences, New Delhi, India and written informed consent was obtained from each participants and parents. Subjects with known type 2 diabetes mellitus (T2DM), cardiovascular disease, presence of other liver diseases (alcoholic liver disease, hepatitis virus infection, autoimmune hepatitis, primary biliary cirrhosis obstruction, drug-induced liver damage etc.), severe chronic lung disease, patients having mechanical upper airway obstruction (conformed by ENT consultant), severe end organ damage, human immunodeficiency virus (HIV) infection and other diseases like hypo and hyperthyroidism were excluded from the study. After taking consent, a detailed questionnaire was filled and eligible participants who were willing to participate in the study underwent overnight polysomnography for the diagnosis of OSA.

Clinical and anthropometric and body composition measurements
The participants/parents were asked to complete the questionnaires on demographic characteristics, clinical profiles, personal and family medical history. Medical history was recorded with special reference to history of snoring, daytime sleepiness, any other sleep related disturbance, history of any chronic disease and of any major medical illnesses. Phenotypic markers (double chin, acanthosis nigricans, buffalo hump, skin tags, xanthelasma and arcus) were measured. Height, weight, BMI, waist circumference (WC), hip circumference (HC), waist hip ratio (WHR), mid arm circumference (MAC), mid-thigh circumference (MTC), neck circumference (NC) and skinfold thickness at 6 sites (triceps, biceps, anterior axillary, suprailiac, subscapular and lateral thoracic) were measured according to standard protocols [13]. Body mass Index (BMI) was calculated by using formula weight (kg)/height (m 2 ). According to international obesity task force, 2015 "Asian Cutoffs" to define underweight (BMI < 18.5 kg/m 2 ), normal weight (18.5-22.9 kg/m 2 ), overweight (BMI 23-26.99kg/m 2 ) and obese (BMI � 27kg/ m 2 ). All subjects were carefully examined physically. A four-point bioelectrical impedance apparatus (Tanita TBF 300, TANITA Corp., Tokyo, Japan), validated for Asian children and adolescents [14], was used to measure the % body fat, body fat (kg), total body water (kg) and lean body mass.

Biochemical investigations
Five milliliter of fasting blood sample was drawn from all the enrolled subjects for following investigations.
Lipid profile and liver function test. The estimation of fasting blood glucose (FBG) was performed after 12 hour overnight fast with the help of commercially available kit (Randox lab ltd, United Kingdom). Estimations for total cholesterol (TC), serum triglycerides (TG), highdensity lipoprotein cholesterol (HDL-C), very low-density lipoprotein cholesterol (VLDL-C), aspartate aminotransferase (AST), alkaline phosphate (ALP) and alanine aminotransferase (ALT) levels were done as previously described [13].
Fasting insulin and homoeostasis modal assessment for insulin resistance (HOMA). Fasting insulin levels were measured using commercially available radioimmunoassay kits (Immunotech, France) [13]. Overall, for all the parameters the intra and inter-assay percentage coefficient and coefficient of variation were <2.0%, 1.8% and <4%, respectively.

Polysomnography profile
All patients were called for sleep study at 8.00 pm and were attached to Alice 3 infant and pediatric computerized polysomnography system using the various leads and devices through standard gold cup electrodes. All patients underwent overnight digital polysomnography (Medi palm; Braebon Medical Corp., Canada) and classified according to apnea hypopnea index (AHI). The various parameters monitored were electrooculogram, electroencephalogram, electromyogram, an electrocardiogram and airflow, chest and abdominal efforts and arterial oxyhemoglobin saturation by pulse oximeter. At least 6 hours of polysomnography was considered as a complete study. The data recorded on the computer was manually scored for sleep stages, apneas and hypopneas. The recordings were analyzed with 60-second epoch, and sleep stages were scored according to the standard criteria [17]. Subjects with an AHI < 1/ hour were assigned as not having OSA and subjects with an AHI �1/hour were diagnosed to have OSA. Subjects having OSA were further classified as mild (AHI 1-4.9/h), moderate (AHI 5-10/h) and severe (AHI � 10/h).

Statistical analysis
Data were entered in an Excel spreadsheet (Microsoft Corp, Washington, USA). The distribution of clinical, biochemical, anthropometry and body composition parameters was confirmed for approximate normality. All enrolled subjects were divided between two groups, with and without OSAs. Both the groups were compared for demographic, anthropometry, body composition and inflammatory markers. Categorical data were analyzed by Chi-squared test, with Fisher correction when appropriate, and expressed as absolute number (%). Continuous variables were expressed as the mean ± standard deviation to summarize the variables. All continuous values were performed using the Z score method. Pearson's correlation coefficient and significance of 'r' was used to compare between IL-6, IL-8, IL-10, IL-17, IL-18, IL-23, MIF, Hs-CRP, TNF-α, PAI-1 and leptin and other variables with adjustment for age and sex. Difference in levels of inflammatory markers in children with or without OSA was assessed using Univariate and multivariate analysis. All analyses were conducted using SPSS software for Windows, version 13.0 (Chicago, USA). Values of p < 0.05 were considered statistically significant.

Demographic, clinical, anthropometric and body composition profiles
Demographic, clinical, anthropometric, body composition profiles were shown is Table 1. The total number of subjects were 2 47 (190 with OSA and 57 without OSA). Mean age: 10.71± 3.00 years and 11.87±2.66 years, respectively. Double chin was significantly increased in OSA children (p = 0.008). No significant difference in personal and family medical history in OSA and control group. Mean values of weight (p = 0.003), BMI (p = 0.04), fat mass (kg) (p = 0.04), % body fat (p = 0.02), WC (p = 0.05) and anterior-axillary (p = 0.05) was significantly higher in OSA as compared healthy controls. Further, Height (p = 0.01) and lean body mass (kg) (p = 0.05) was significantly increased in controls.

Polysomnography finding
A polysomnography profile was shown in Table 2. Mean AHI (p = 0.001), AHI-NERM (p<0.0001), total events (p<0.0001), total events NERM (p<0.0001), total events REM (p<0.0001) and stage N3 (p = 0.04) was significantly increase in cases as compared to controls. There was no significant difference between snoring (present and past), arousal in night at present and past, daytime sleeping schedule (present and past), nasal congestion (present and past), sleep efficiency and sleep latency in cases as compared controls.

Inflammatory markers
Comparison of inflammatory cytokines in OSA children with healthy controls was shown in Table 4

Association of inflammatory markers with AHI
The mean AHI was significantly higher in OSA subjects as compared to controls (p<0.05). In OSA group, all inflammatory markers were significantly increased with the AHI levels (Fig 2).   (Table 5). Further, mean AHI, total events (A+H+RERA) total, total events (A+H+RERA) NERM, total events (A+H +RERA) REM and stage N1 (minute) was elevated in severe OSA subjects as compared to mild and moderate OSA (Table 6).

Correlation analysis of inflammatory biomarkers with other variables
A regression analyses was performed to demonstrate the relationships between inflammatory cytokines, anthropometric, body composition, polysomnography finding and biochemical investigations is shown in Table 7  It has been indicated that childhood obesity induces psychological, neurological, endocrine, cardiovascular, respiratory, gastrointestinal, and orthopedic abnormalities [18]. Also OSA is associated with insulin resistance, dyslipidemia, metabolic syndrome, hypertension, and inflammation in Adults. It encompasses a cluster of metabolic and cardiovascular abnormalities may play an independent role in the pathophysiology of OSAS in children as well [19]. A  [20] reported an increase in incidence of OSA and insulin levels in obese adolescents. Similarly we observed that fasting insulin and HOMA-IR levels were significantly associated and with OSA. There is no consistent relationship between pediatric OSA and measures of body fat distribution. Verhulst et al. [21] investigated that the relationship between body fat distribution and OSA. He showed that WHR and total body fat has not found a clear-cut relationship between fat distribution and OSA. Another study by the same authors demonstrated that several measures of abdominal adiposity (WC, WHR, and % fat mass) correlated with oxygen desaturations associated with brief respiratory pauses, but not OSA [22]. In this study, we showed that central and abdominal obesity, body fat was significantly increased in OSA children as compared to children without OSAs.
Previous studies reported that OSA is a disease with chronic systemic inflammation, but the interaction between the different inflammatory cytokines was unclear. Systemically, Leon et al. [23] reported that the pro-inflammatory cytokines, TNF-α, IL-6, and IL-8 have been elevated in the OSA patients. Furthermore, OSA and obesity has been shown systemic inflammatory pathways, for instance serum PAI-1 levels have been significantly increased in obese OSA children [24]. In addition, hs-CRP levels have been increases in children with OSA, particularly with a threshold of an AHI > 5/hour [25]. Abdelnaby et al. [26] reported that the presence of excessive daytime sleepiness to increased TNF-α levels and the TNF-α −308G gene polymorphism. Another study has been indicated that childhood OSA was associated with higher plasma MIF, Hs CRP and fasting insulin levels [27]. We have also observed similar findings in this study, whereby TNF-α, IL-6, IL-8, MIF and PIA-1 were significantly associated with OSA was independent of the degree of obesity. Furthermore, all biomarkers altered by OSA in our subjects have been ascribed to play a patho-physiological role in cardiovascular dysfunction, thereby suggesting that OSA in obese subjects might predispose them to a more severe cardiovascular phenotype and to earlier development of cardiovascular morbidities.
The pro-inflammatory cytokines, IL-17 and IL-23, have been recently emphasized. IL-17 is a pro-inflammatory cytokine secreted by T helper 17 cells that act with other cytokines such as IL-1, TNF-α, and IL-6 to induce inflammatory diseases and play an important role in the development of autoimmunity [28]. IL-23 is a cytokine acts on memory-cluster-designation-4 (+) T cells, activates the transcription activator, and stimulates the production of interferongamma [29]. Huang et al. [8] reported that IL-17 and IL-23 levels were significantly elevated in pediatric OSA. These markers could be used as biomarkers of pediatric OSA. Our finding showed that IL-17 levels are significantly associated OSA, but IL-23 was significantly correlated with body fat and liver enzyme.
Leptin is a hormone with multiple functions, including regulating food intake. Tauman et al. [30] reported that leptin levels were significantly higher in sleep disorder breathing (SDB) group and also leptin levels correlate directly with BMI Z score and AHI. Another study [31] has been shown that leptin levels were higher in the SDB group. Our finding of is consistent with the previous finding. We showed that leptin levels were significantly associated with OSA and correlated with BMI, AHI and fasting insulin.
The limitations of the study include small sample size and that sample originated from north India. Further, we did not use a more accurate method of measurement of insulin sensitivity like the hyperinsulinemic euglycemic clamp technique. This was primarily due to the infrastructural constraints. Finally, longitudinal studies are warranted to study interplay between inflammation, obesity, insulin resistance, and OSA in Asian Indians.

Conclusion
Body composition and metabolic alteration is a strong predictor of OSA. Overweight/obese children with OSA were associated with increased systemic inflammation and contribute to the relationship between inflammation, obesity and OSA. Further studies are needed to investigate the role of these biomarkers on endothelial dysfunction.