- Email: [email protected]
Metabolic alterations in adolescents with obstructive sleep apnea
Accepted Manuscript Title: Metabolic Alterations in Adolescents with Obstructive Sleep Apnea Author: Bharat Bhushan Bushra Ayub Darius A. Loghmanee Kathleen R. Billings PII: DOI: Reference:
S0165-5876(15)00553-4 http://dx.doi.org/doi:10.1016/j.ijporl.2015.10.046 PEDOT 7839
To appear in:
International Journal of Pediatric Otorhinolaryngology
Received date: Revised date: Accepted date:
17-8-2015 26-10-2015 29-10-2015
Please cite this article as: B. Bhushan, B. Ayub, D.A. Loghmanee, K.R. Billings, Metabolic Alterations in Adolescents with Obstructive Sleep Apnea, International Journal of Pediatric Otorhinolaryngology (2015), http://dx.doi.org/10.1016/j.ijporl.2015.10.046 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Metabolic Alterations in Adolescents with Obstructive Sleep Apnea Bharat Bhushan, PhDa,b, Bushra Ayub, MDa, Darius A. Loghmanee, MDc, Kathleen R. Billings,
a
ip t
MDa,b Division of Pediatric Otolaryngology-Head and Neck Surgery, Ann & Robert H. Lurie
Northwestern University Feinberg School of Medicine, 303 E. Chicago Ave., Chicago, IL,
us
b
cr
Children's Hospital of Chicago, 225 E. Chicago Ave. Box #25, Chicago, IL, 60611.
60611, United States.
Division of Pediatric Sleep Medicine, Advocate Medical Group, 1675 Dempster St., Park Ridge,
an
c
IL, 60068, United States.
M
Corresponding Author: Bharat Bhushan, PhD
d
Research Scientist, Division of Pediatric Otolaryngology-Head and Neck Surgery
te
Ann & Robert H. Lurie Children’s Hospital of Chicago
Ac ce p
Research Assistant Professor, Northwestern University Feinberg School of Medicine T 312-227-6793 | F 312-227-6230 | [email protected], [email protected] 225 E. Chicago Ave., Box#25, Chicago, IL-60611-2605 Word count abstract: 309
Word count Manuscript: 2335
Conflict of Interest/Disclosures: None Key Words – Obstructive Sleep Apnea; Metabolic Alterations; Metabolic Syndrome; Obesity; Adolescents
Page 1 of 20
Abstract Importance: Obesity is one of the leading health concerns in developed and in developing countries. The risk of obstructive sleep apnea (OSA) is greatly increased by obesity. Obesity is
ip t
known to be associated with the Metabolic Syndrome and cardiovascular disease in adults. This same association in children is not well defined. Understanding the relationship of obesity,
cr
OSA, and metabolic alterations in children would improve understanding of the risks of
us
cardiovascular disease into adulthood.
Objective: To evaluate the association of OSA and metabolic outcomes, including lipid
an
variables and insulin resistance, in obese adolescents.
Methods: Retrospective, case-control series at a tertiary care children’s hospital. Obese
M
adolescents aged 12-18 years who underwent overnight polysomnography (PSG) and routine
31, 2012.
d
laboratory testing for lipid levels, fasting glucose, and insulin from January 1, 2006 to December
te
Results: A total of 42 patients with a mean age of 14.1±1.9 years were analyzed. Nineteen
Ac ce p
(45.2%) were male. The mean body mass index (BMI) z score was 2.23±0.86, and all patients were obese (BMI z score >95th percentile). Triglyceride, fasting blood glucose, insulin, and homeostasis model assessment-insulin resistance (HOMA-IR) levels were significantly higher in patients with OSA when compared to those with No-OSA (p<0.01). There was incremental worsening of insulin and HOMA-IR with greater severity of OSA. The apnea-hypopnea index (AHI) was positively and significantly correlated with blood glucose and HOMA-IR (p=0.01and p<0.001, respectively). Multiple linear regression analysis showed that the AHI was a predictor of blood glucose (p=0.04) and HOMA-IR (p=0.01) independent of age, gender, total sleep time and BMI z score. Logistic regression analysis showed that elevated levels of blood glucose
Page 2 of 20
predicted severe OSA (p=0.02) independent of gender and BMI z score. Elevation in HOMA-IR predicted severe OSA (p=0.004). Conclusion: OSA severity is associated with increased fasting insulin, blood glucose and
Ac ce p
te
d
M
an
us
cr
ip t
HOMA-IR even after controlling for the age, and BMI z score in adolescents.
Page 3 of 20
Introduction Obesity in children and adolescents is a major health issue in the United States and in other countries. According to a recent report from the Center for Disease Control (CDC), childhood obesity has more than doubled in children and quadrupled in adolescents in the past 30
ip t
years [1]. In a population-based study of 5 to 17 year old children, 70% of obese youth had at
cr
least one risk factor for cardiovascular disease [2]. Obesity in children can cause early
development of hypertension, cardiovascular disease, and hyperglycemia, which in later life
us
develops into diabetes. Metabolic Syndrome is a cluster of increased blood pressure (BP),
an
elevated triglyceride, cholesterol, low density lipoprotein, and glucose levels, and decreased high density lipoprotein levels. The prevalence of Metabolic Syndrome is 6.8% among overweight
M
and 28.7% among obese adolescents [2, 3]. Metabolic Syndrome, or the components of Metabolic Syndrome, increases the risk of developing atherosclerosis and cardiovascular disease,
d
i.e. heart failure, stroke, and peripheral arterial disease.
te
Obstructive sleep apnea (OSA), a disorder of breathing during sleep, has been correlated
Ac ce p
to obesity. OSA is a common clinical condition characterized by intermittent and partial airway collapse, resulting in frequent episodes of apnea, hypopnea, and recurrent arousals from sleep [24]. The estimated prevalence of OSA is 2-3% in children [3]. This risk of OSA is greatly increased by obesity, with an estimated prevalence of 36% in obese children [5]. In adults, both the obesity and OSA are shown to be associated with cardiovascular disease, high cholesterol and high blood pressure. This association is less defined in children and adolescents. Some studies in children have shown OSA to independently increases insulin resistance [5, 6], while other studies suggest that obesity is the variable increasing insulin resistance [7] [8]. In this study, we hypothesize that OSA is linked with metabolic alternations, including lipid variables, fasting glucose and insulin levels, independent of BMI z score in adolescents.
Page 4 of 20
Alterations in metabolic variables and their association with individual components of OSA, including apnea–hypopnea index (AHI), oxygen saturation (SpO2), and arousal index were studied. The relationship of metabolic variable laboratory test results and obesity were analyzed. Individual measures of the metabolic panel, particularly elevations in fasting glucose and insulin
cr
severity of OSA and metabolic abnormalities in adolescents. Methods
us
Study population
ip t
levels, were analyzed relative to OSA severity (19) to help better define the relationship between
an
The study was approved by the Institutional Review Board of Ann & Robert H. Lurie Children's Hospital of Chicago. This is a historical cohort study of consecutive patients, aged 12
M
to 18 years, having overnight polysomnography (PSG) at Ann & Robert H. Lurie Children's Hospital of Chicago between January 1, 2006 and December 31, 2012, who also had laboratory
d
testing for the metabolic variables being analyzed within 3 months of their PSG. A total of 55
te
patients were identified. Patients were excluded from analysis if they had a history of Type I
Ac ce p
diabetes (n=3), genetic abnormalities (n=3), or used a continuous positive airway pressure device (CPAP) for OSA (n=2). Other exclusions included those patients with craniofacial anomalies (n=1), organ transplant recipients (n=2), and multiple medical problems (n=2). A total of 42 patients were available for the final analysis. The age, gender, laboratory data on metabolic variables, height, weight, and PSG findings were collected from the electronic medical records. Definition and Measurements
All laboratory analyses were conducted at the Ann & Robert H. Lurie Children's Hospital laboratory. The metabolic variables including fasting glucose, and insulin levels measured were attained after an overnight fast. Abnormal fasting glucose values were determined if a patient met the American Diabetes Association guidelines (fasting glucose ≥100 mg/dl) [9]. The
Page 5 of 20
homeostasis model assessment-insulin resistance (HOMA-IR) was used as an index of insulin resistance, and this was calculated using a standard equation (fasting insulin (μIU/mL) × fasting blood glucose (mmol/L)/22.5) [10]. Based on this score, insulin resistance was determined using the HOMA-IR cut-points described by Kurtoglu et al. (>2.67 for boy and >2.22 for girls) [11].
ip t
Height and weight were recorded for each patient. The BMI z-score was computed using Center
cr
for Disease Control (CDC) growth standards (www.cdc.gov/growthcharts) and online software (www.cdc.gov/epiinfo). Patients with a BMI z score above the 95th percentile were considered
us
obese.
an
Polysomnography
A standard overnight PSG (Cadwell easy 3 versions 3.9.34, Kennewick, WA, USA) had
M
been performed on all patients. The apneas and hypopneas were identified and scored according to the American Academy of Sleep Medicine (AASM) pediatric criteria as defined in the AASM
d
Manual for Scoring of Sleep and Associated Events [12]. The apnea–hypopnea index (AHI) was
te
defined as the total number of obstructive apneas and hypopneas per hour of sleep. Severity of
Ac ce p
OSA was classified as mild (AHI between 1 and 4.99 events per hour of sleep), moderate (AHI between 5 and 9.99 events per hour of sleep), severe (AHI ≥ 10 events per hour of sleep), and No-OSA (AHI < 1 event per hour of sleep) [6] [13]. Statistical analysis
Descriptive statistics were summarized using frequencies and percentages for categorical data, and mean and standard deviations for continuous data. To determine differences between groups of No-OSA, Mild-OSA, Moderate-OSA and Severe-OSA, both parametric and nonparametric tests were used. Chi2 test of association was used for nominal data. T-test and analysis of variance statistics were used for normally distributed data. Mann Whitney U and Kruskal Wallis statistics were used for non-normally distributed data. Outliers were identified
Page 6 of 20
and checked for accuracy. Scatter plots were used to graphically show correlations among PSG and metabolic variables. Multiple linear regression models were run to predict dependent variables for log transformed fasting insulin and log-transformed HOMA-IR. Predictors included age, gender, TST, BMI z score, and AHI. Multinomial logistic regression models were run to
ip t
predict patients grouped into those with No-OSA, Mild-OSA, Moderate-OSA, and Severe-OSA.
cr
Results were considered significant with two-tailed test and p<0.05. Statistical analysis was
Inc., Chicago, IL).
an
Results
us
conducted using statistical product and service solution (SPSS) software version 20 (IBM SPSS
A total 42 patients with a mean age of 14.1±1.9 years were analyzed. Nineteen (45.2%)
M
were male. The average BMI z score was 2.23±0.86, and all patients were obese (BMI z score >95th percentile). Eighteen patients (42.8%) had an AHI <1/hour (No-OSA), 7 patients (16.6%)
d
had an AHI ≥ 1≤4.99/hour (Mild-OSA), 5 (11.9%) had an AHI ≥5/h<9.99/hour (Moderate-OSA)
te
and 12 (28.6%) had an AHI ≥ 10/hour (Severe-OSA). There was no significant difference in age
Ac ce p
(p=0.44), and BMI z score (p=0.16) between the groups. (Table 1) Patients with OSA had a significantly higher AHI and arousal index, and a lower average SpO2 and SpO2 nadir, as expected. Stage 1 total sleep time (TST) was significantly higher in those with severe OSA (p<0.001).
When comparing metabolic and PSG variables to OSA severity, there was no significant difference in HDL-C or LDL-C in those patients with No-OSA, and in those with all levels of OSA. Total cholesterol did not reach significance in the groups (p=0.05). (Table 1) Triglyceride, blood glucose, fasting insulin, and HOMA-IR levels were significantly higher in patients with OSA compare to those with No-OSA. There was incremental worsening of insulin and HOMAIR levels with greater severity of OSA: 3.9±1.8 in those with No-OSA, 7.7 ±2.1 in Mild- OSA,
Page 7 of 20
8.2±1.3 in Moderate, and 14.8±5.3 in Severe-OSA. A similar trend was observed in insulin levels. (Table 1) Further correlation analysis suggested that the AHI was positively and significantly correlated with elevated blood glucose and HOMA-IR (p=0.01 and p<0.001, respectively).
ip t
(Figure 1A & 1B) Other components of the PSG (arousal index, SpO2 nadir) followed the same
cr
trend when analyzed (data not shown). Multivariate analysis
us
Multiple linear regression analysis models assessing the effect of gender, TST, BMI z
an
score, and AHI on the log transformed fasting blood glucose and HOMA-IR values were performed. The analysis showed that the AHI was a predictor of elevated blood glucose (p=
M
0.04) and HOMA-IR (p=0.01) independent of gender, TST and BMI z score. (Table 2) A multinomial logistic regression analysis model assessing the effect of the Mild,
d
Moderate, and Severe-OSA on elevations of blood glucose and HOMA-IR indices was
te
performed. The results showed that, even after controlling for the age, gender and BMI z score,
Ac ce p
elevated levels of blood glucose and HOMA-IR predicted the Severe-OSA (p=0.02). (Table 3) Discussion
In this study, we looked at the relationship between OSA and metabolic variables, including lipids, fasting blood glucose and insulin levels, in adolescents 12-18 years of age. The study demonstrated that serum triglycerides levels (p<0.001), fasting blood glucose levels (p=0.005), and fasting insulin and HOMA-IR levels (p<0.001) were significantly higher in obese adolescents with OSA when compared to those with No-OSA. Redline et al. [3] reported sleep disordered breathing (SDB) as a risk factor for Metabolic Syndrome in adolescents. Her study showed that obesity and Metabolic Syndrome were more
Page 8 of 20
prevalent in the SDB group when compared with the non-SDB group. A seven fold increased chance for Metabolic Syndrome in those with SDB relative to unaffected children was described. The association between many components of the Metabolic Syndrome and SDB, even after adjusting for BMI, suggested that SDB contributed to metabolic dysfunction beyond the effects
ip t
weight alone.
cr
Canapari et al. [14] showed that children (5-18 years of age) with OSA had significantly increased BMI, HOMA-IR, triglycerides, and leptin levels compared to those without OSA. In
us
contrast, Tauman et al. [15], found no correlation between the AHI and serum insulin, serum
an
glucose, HOMA-IR and triglyceride levels in any of the children analyzed regardless of weight. The prior study showed that insulin resistance and dyslipidemia were determined primarily by
M
the degree of body adiposity, rather than the severity of SDB. Given the conflicting evidence from prior studies in adolescents, we expanded our previous work and explored the relationship
d
between obesity, OSA, and alterations in metabolic variables in adolescents. In our previously
te
published study on children aged 3-11 years, we found that fasting blood glucose and insulin
Ac ce p
levels were positively and significantly associated with the AHI. These finding remain consistent in the current study.
Insulin resistance is a physiological condition in which the body produces insulin, but cells in the body are unable to use it effectively. As a result, glucose builds up in the blood instead of being absorbed by the cells, leading to type II diabetes or pre-diabetic conditions. Our previous study in younger children [13], and our present study on adolescents, showed that patients classified with Severe-OSA had significantly higher blood glucose and HOMA-IR levels. An elevation in triglycerides, blood glucose, fasting insulin and HOMA-IR were associated with increased AHI, arousal index, and SpO2 nadir in the patients studied.
Page 9 of 20
To further clarify this correlation, additional multivariate analyses were performed. Significant p-values for the blood glucose and HOMA-IR in multinomial logistic analysis of our data suggested that, even after controlling for the age, gender and BMI z score, higher levels of blood glucose and HOMA-IR values were more likely to fall in the severe OSA group (Table-3).
ip t
We found elevated lipid levels with increasing severity of OSA, even after adjusting for BMI z
cr
score. There has been concern that the clustering of abnormalities in variables like insulin, lipid levels, and elevated blood pressure (Metabolic Syndrome) could be associated with development
us
of cardiovascular complications in later ages. Early identification and treatment of the children
an
with OSA and Metabolic Syndrome becomes essential to decrease morbidity into early adulthood.
M
The neuroendocrine and adverse cardiovascular outcomes have been linked with OSA as well as with obesity [16]. Floren et al [17], reported several different mechanisms by which
d
OSA can cause hypertension, atherosclerosis and linked it to the cardiovascular problems later in
te
the life. One study on adults showed that CPAP for severe OSA improved hypertension severity
Ac ce p
[18]. There is lack of such data in the pediatric population. Patients with OSA experienced varying severity of hypoxic episodes and recurrent arousals during sleep due to sympathetic over activity. Different mechanisms like chemo baro-reflex changes, vasoconstriction by endothelin release and endothelial dysfunction have been described which increases hepatic glycogenolysis and gluconeogenesis [7, 19-21]. These sympathetic activities cause stimulation of lipolysis, which increases circulating free fatty acid metabolites. These inhibit insulin signaling and reduce insulin-mediated glucose uptake, contributing to insulin resistance. Hypoxia induced by OSA may cause oxidative stress, which leads to the production of Reactive Oxygen Species. These events contribute to altered glucose homeostasis by impairing pancreatic function and insulin secretion [7].
Page 10 of 20
Adenotonsillectomy is the most common treatment offered to the children with OSA [22]. Gozal et al. [6] reported that the surgical removal of hypertrophic tonsils and adenoids led to significant improvements in lipid variables, C-reactive protein, and apolipoprotein B. Their study included the overweight (BMIz score>1.20) as well as non-obese patients. Authors
ip t
reported improvements in the metabolic variables including fasting Insulin and glucose in the
cr
obese patients (n=37) with OSA after 6-8 months of adenotonsillectomy. Waters et al. [23]
showed that adenotonsillectomy for OSA successfully improved the respiratory disturbance
us
index (RDI) in the majority of children. In that study, twenty children underwent surgical
an
removal of adenotonsillar tissue. The authors found a small but significant improvement in total cholesterol in those children whose OSA was resolved and a trend for obese children with
M
persisting OSA to have elevated insulin levels compared with obese children without OSA. The present study was conducted retrospectively, and we did not look at the effect of
d
OSA intervention on improvements in metabolic alterations and insulin resistance. The sample
te
size was small, since a convenience sample of patients having both a PSG and laboratory
Ac ce p
analysis was limiting. This may have led to type II errors in failing to detect correlations between OSA and metabolic alterations. Despite this, important relationships between these variables were detected. Further research is needed to evaluate the overall effect of CPAP and adenotonsillectomy on metabolic alterations in obese and non-obese adolescents with OSA. Conclusions
OSA severity was associated with increased HOMA-IR and blood glucose levels, even after controlling for the BMI, in adolescents. Identification and understanding the mechanism and interaction of OSA with abnormal metabolic variables is important in children. Addressing these interactions at early stages may benefit patients by allowing for more timely diagnosis and treatment to help to reduce the risk of diabetes and cardiovascular disease.
Page 11 of 20
Acknowledgements: We thank Karen Rychlik, MS of the Biostatistics Research Core of the Ann
Ac ce p
te
d
M
an
us
cr
ip t
& Robert H. Lurie Children’s Hospital of Chicago for her assistance with the statistical analysis.
Page 12 of 20
References [1] Center for Disease Control. Prevention Status Reports. 2013. [2] Punjabi NM. The epidemiology of adult obstructive sleep apnea. Proc Am Thoracic Soc. 2008;5:136-143.
ip t
[3] Redline S, Storfer-Isser A, Rosen CL, Johnson NL, Kirchner HL, et al. Association between metabolic syndrome and sleep-disordered breathing in adolescents. Am Journal Resp Critical
cr
Care Med. 2007;176:401-408.
[4] Penn EB, French A, Bhushan B, Schroeder JW. Access to care for children with symptoms of
us
sleep disordered breathing. Int J Ped Otorhinolaryngol. 2012;76:1671-1673.
[5] Lanfranco F, Motta G, Minetto MA, Baldi M, Balbo M, et al. Neuroendocrine alterations in obese patients with sleep apnea syndrome. Int J Endocrinol. 2010;2010:474-518.
an
[6] Gozal D, Capdevila OS, Kheirandish-Gozal L. Metabolic alterations and systemic Resp Crit Care Med. 2008;177:1142-1149.
M
inflammation in obstructive sleep apnea among nonobese and obese prepubertal children. Am J [7] Punjabi NM, Shahar E, Redline S, Gottlieb DJ, Givelber R, et al. Sleep Heart Health Study: Sleep-disordered breathing, glucose intolerance, and insulin resistance: the Sleep Heart Health
d
Study. Am J Epidemiol. 2004;160:521-530.
te
[8] Nakra N, Bhargava S, Dzuira J, Caprio S, Bazzy-Asaad A. Sleep-disordered breathing in children with metabolic syndrome: the role of leptin and sympathetic nervous system activity
Ac ce p
and the effect of continuous positive airway pressure. Pediatrics. 2008;122:e634-642. [9] Expert Committee on the Classification of Diabetes: Report of the expert committee on the diagnosis and classification of diabetes mellitus. Diabetes Care. 2003;26(1):S5-20. [10] Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, et al. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28:412-419. [11] Kurtoglu S, Hatipoglu N, Mazicioglu M, Kendirici M, Keskin M, et al. Insulin resistance in obese children and adolescents: HOMA-IR cut-off levels in the prepubertal and pubertal periods. J Clin Res Ped Endocrinol. 2010;2:100-106. [12] Berry RB, Budhiraja R, Gottlieb DJ, Gozal D, Iber C, et al. American Academy of Sleep, rules for scoring respiratory events in sleep: update of the 2007 AASM Manual for the Scoring
Page 13 of 20
of Sleep and Associated Events. Deliberations of the Sleep Apnea Definitions Task Force of the American Academy of Sleep Medicine. J Clin Sleep Med. 2012;8:597-619. [13] Bhushan B, Maddalozzo J, Sheldon SH, Haymond S, Rychlik K, et al. Metabolic alterations in children with obstructive sleep apnea. Int J Ped Otorhinolaryngol. 2014;78.854-859. [14] Canapari CA, Hoppin AG, Kinane TB, Thomas BJ, Torriani M, et al. Relationship between
ip t
sleep apnea, fat distribution, and insulin resistance in obese children. J Clin Sleep Med. 2011;7:268-273.
cr
[15] Tauman R, O'Brien LM, Ivanenko A, Gozal D. Obesity rather than severity of sleep-
disordered breathing as the major determinant of insulin resistance and altered lipidemia in
us
snoring children. Pediatrics. 2005;116:e66-73.
[16] Grassi G, Facchini A, Trevano FQ, Dell'Oro R, Arenare F, et al. Obstructive sleep apnea-
an
dependent and -independent adrenergic activation in obesity. Hypertension. 2005;46:321-325. [17] Floras JS. Obstructive sleep apnea syndrome, continuous positive airway pressure and treatment of hypertension. Eur J Pharmacol. 2015.
M
[18] Jelic S, Lederer DJ, Adams T, Padeletti M, Colombo PC, et al. Vascular inflammation in obesity and sleep apnea. Circulation. 2010:210:1014-1021.
d
[19] Narkiewicz K, Somers VK. Sympathetic nerve activity in obstructive sleep apnoea. Acta Physiologica Scan. 2003;177:385-390.
Ac ce p
Rev Cardiol. 2010;7:677-685.
te
[20] Kohler M, Stradling JR. Mechanisms of vascular damage in obstructive sleep apnea. Nature [21] Bhushan B, Khalyfa A, Spruyt K, Kheirandish-Gozal L, Capdevila OS, et al. Fatty-acid binding protein 4 gene polymorphisms and plasma levels in children with obstructive sleep apnea. Sleep Med. 2011;12:666-671.
[22] Bhushan B, Sheldon S, Wang E, Schroeder JW. Clinical indicators that predict the presence of moderate to severe obstructive sleep apnea after adenotonsillectomy in children. Am J Otolaryngol. 2014;35:487-495.
[23] O'Brien LM, Sitha S, Baur LA, Waters KA. Obesity increases the risk for persisting obstructive sleep apnea after treatment in children. Int J Ped Otorhinolaryngol. 2006;70:15551560.
Page 14 of 20
Table 1. The relationship between the severity of obstructive sleep apnea (OSA) and metabolic variables in adolescents are shown. Controls (n=18)
Mild OSA (n=7)
Moderate OSA (n=5)
Severe OSA (n=12)
p-Value ANOVA
Age (years)
14.1±2.1
14.3±2.0
15.1±1.6
13.5±1.3
0.44
BMI z score
1.9±0.9
2.4±0.68
2.2±0.5
2.6±0.8
0.16
0.51±0.4
3.1±0.9
TST (hours)
5.4± 1.3
5.4 ±1.1
Stage 1 (%)
4.8 ± 2.8
4.3 ± 3.2
Stage 2 (%)
56.8 ± 5.6
SWS (%)
25.8 ± 5.8
REM (%)
12.6 ± 7.0
Arousal index (per hour)
0.51 ±0.4
Average SaO2 (%)
96.8 ± 1.4
0.2
7.6±1.1
37±23.4
<0.001
6.0 ±1.1
5.0 ± 0.8
0.41
3.6 ± 2.3
11.8 ± 6.7
<0.001
49.6 ±8.7
62.4 ± 6.9
56.3 ± 12.7
0.11
32.3 ±6.9
20.8 ± 3.1
25.9 ± 10.0
0.06
13.6 ± 6.9
13± 7.7
7.9 ± 7.7
0.26
7.6 ± 1.1
36.9 ± 23.3
<0.001
95.7 ±2.6
96.6 ±1.1
94.1 ± 2.4
0.01
87.6 ± 7.4
88.4 ±4.1
86 ± 4.9
0.03
d
3.1 ± 0.9
te
91.44 ± 3.9
cr
AHI (per hour)
13.3 ± 12.4
us
38.6 ±46.1
an
50 ± 54.7
SpO2 Nadir (%)
20.9 ±22.8
M
Sleep Latency (mins)
Ac ce p
Polysomnography variables
ip t
Variables
Metabolic variables (mg/dl)
Total Cholesterol
147.2 ± 26.8
150.6 ± 27.5
150.8 ± 5.8
174.7 ± 29.3
0.05
Triglycerides
100.1 ± 36.2
129.1 ± 38.1
116.2 ± 16.7
157 ± 30.3
0.001
39.3 ± 9.9
37.3 ± 7.9
34.4 ± 14.5
39.1 ± 10.1
0.79
84.1 ± 19.4
103.8 ±62.2
117.6 ± 78.3
99.5 ± 32.4
0.38
Blood Glucose
93.7 ± 5.8
96.4 ± 6.3
93.6 ± 4.4
102.3 ± 6.8
0.005
Fasting Insulin (µIU/ml)
18.4 ±7.3
31.8 ± 7.8
35.7 ± 4.7
58.1 ± 20.7
<0.001
HOMA-IR
3.9 ± 1.8
7.7 ± 2.1
8.256 ± 1.3
14.8 ± 5.3
<0.001
HDL-C LDL-C
Page 15 of 20
BMI, body mass index; AHI, apnea–hypopnea index; TST, total sleep time; SWS, slow wave sleep; REM, rapid eye movement; HDL-C, high density lipoprotein-cholesterol; TC, total cholesterol; LDL-C, low density lipoprotein-cholesterol; TG, triglycerides; HOMA-IR,
Ac ce p
te
d
M
an
us
cr
ip t
homeostasis model assessment. p<0.05 was considered significant and are highlighted bold.
Page 16 of 20
Table 2. Multiple linear regression analysis models assessing the independent effect of different variables (gender, BMI z score, and AHI) on log-transformed fasting insulin or HOMA-IR values. P<0.05 was significant. Dependent variable
Dependent variable
β(standardized
p-Value
coefficient)
coefficient)
Gender
0.14
0.28
-0.15
BMI z score
0.28
0.63
0.12
AHI
0.50
<0.001
0.45
0.34
0.46
0.01
an
(per hour)
p-Value
cr
variables
β(standardized
us
Independent
Blood glucose mg/dl
ip t
HOMA-IR
Ac ce p
te
d
M
BMI, body mass index; AHI, apnea hypopnea index.
Page 17 of 20
Table 3. Multinomial logistic regression model assessing the effect of mild, moderate, and severe OSA on elevated fasting insulin, and HOMA-IR index after controlling for age, gender and BMI z score. P<0.05 was considered significant. β-coefficient
p-Value
O.R
Age (years)
0.13
0.62
1.13
Gender
0.65
0.52
1.92
BMI z score
0.81
0.25
2.24
HOMA-IR
0.07
0.53
1.07
0.86
1.34
Moderate
Age (years)
0.26
0.32
1.29
0.78
2.16
OSA
Gender
0.09
0.93
1.10
0.11
10.62
BMI z score
0.43
0.54
1.54
0.38
6.13
HOMA-IR
0.08
Age(years)
-0.49
Gender
-0.87
BMI z score
0.33
HOMA-IR
95% CI
ip t
0.68
1.88
0.27
13.82
0.56
8.88
cr
us
Mild OSA
an
Independent variables
1.09
0.85
1.39
0.25
0.61
0.26
1.41
0.48
0.42
0.03
4.78
0.51
1.39
0.15
13.34
0.43
0.004
1.53
1.14
2.06
0.122
0.81
1.13
0.43
2.96
-0.17
0.89
0.84
0.08
8.59
0.53
0.54
1.70
0.3
9.34
Glucose mg/dl
0.03
0.75
1.03
0.84
1.26
Moderate
Age
0.86
0.07
2.35
0.94
5.9
OSA
Gender
-0.58
0.60
0.55
0.06
5.02
BMI z score
0.81
0.38
2.25
0.37
13.72
Glucose mg/dl
0.03
0.73
1.03
0.86
1.25
Age
0.22
0.24
0.63
0.52
3.1
Gender
0.29
0.80
1.34
0.14
12.90
BMI z score
1.17
0.19
3.23
0.54
19.51
Glucose mg/dl
0.37
0.02
1.43
1.05
1.95
Age Gender
Ac ce p
BMI z score
d
Mild OSA
te
Severe OSA
Severe OSA
M
0.57
OSA, obstructive sleep apnea, BMI- body mass index, HOMA-IR – homeostatic model assessment - Insulin resistance
Page 18 of 20
Figure – 1 A, B
Ac ce p
te
d
M
an
us
cr
ip t
A
Page 19 of 20
ip t cr us an M d te Ac ce p
B
Figure. Scatter plots of individual blood glucose and HOMA-IR values relative to the AHI. A. A significant correlation between the AHI and the fasting insulin was observed. B. A significant correlation between the AHI and the HOMA-IR index was observed.
Page 20 of 20
S0165-5876(15)00553-4 http://dx.doi.org/doi:10.1016/j.ijporl.2015.10.046 PEDOT 7839
To appear in:
International Journal of Pediatric Otorhinolaryngology
Received date: Revised date: Accepted date:
17-8-2015 26-10-2015 29-10-2015
Please cite this article as: B. Bhushan, B. Ayub, D.A. Loghmanee, K.R. Billings, Metabolic Alterations in Adolescents with Obstructive Sleep Apnea, International Journal of Pediatric Otorhinolaryngology (2015), http://dx.doi.org/10.1016/j.ijporl.2015.10.046 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Metabolic Alterations in Adolescents with Obstructive Sleep Apnea Bharat Bhushan, PhDa,b, Bushra Ayub, MDa, Darius A. Loghmanee, MDc, Kathleen R. Billings,
a
ip t
MDa,b Division of Pediatric Otolaryngology-Head and Neck Surgery, Ann & Robert H. Lurie
Northwestern University Feinberg School of Medicine, 303 E. Chicago Ave., Chicago, IL,
us
b
cr
Children's Hospital of Chicago, 225 E. Chicago Ave. Box #25, Chicago, IL, 60611.
60611, United States.
Division of Pediatric Sleep Medicine, Advocate Medical Group, 1675 Dempster St., Park Ridge,
an
c
IL, 60068, United States.
M
Corresponding Author: Bharat Bhushan, PhD
d
Research Scientist, Division of Pediatric Otolaryngology-Head and Neck Surgery
te
Ann & Robert H. Lurie Children’s Hospital of Chicago
Ac ce p
Research Assistant Professor, Northwestern University Feinberg School of Medicine T 312-227-6793 | F 312-227-6230 | [email protected], [email protected] 225 E. Chicago Ave., Box#25, Chicago, IL-60611-2605 Word count abstract: 309
Word count Manuscript: 2335
Conflict of Interest/Disclosures: None Key Words – Obstructive Sleep Apnea; Metabolic Alterations; Metabolic Syndrome; Obesity; Adolescents
Page 1 of 20
Abstract Importance: Obesity is one of the leading health concerns in developed and in developing countries. The risk of obstructive sleep apnea (OSA) is greatly increased by obesity. Obesity is
ip t
known to be associated with the Metabolic Syndrome and cardiovascular disease in adults. This same association in children is not well defined. Understanding the relationship of obesity,
cr
OSA, and metabolic alterations in children would improve understanding of the risks of
us
cardiovascular disease into adulthood.
Objective: To evaluate the association of OSA and metabolic outcomes, including lipid
an
variables and insulin resistance, in obese adolescents.
Methods: Retrospective, case-control series at a tertiary care children’s hospital. Obese
M
adolescents aged 12-18 years who underwent overnight polysomnography (PSG) and routine
31, 2012.
d
laboratory testing for lipid levels, fasting glucose, and insulin from January 1, 2006 to December
te
Results: A total of 42 patients with a mean age of 14.1±1.9 years were analyzed. Nineteen
Ac ce p
(45.2%) were male. The mean body mass index (BMI) z score was 2.23±0.86, and all patients were obese (BMI z score >95th percentile). Triglyceride, fasting blood glucose, insulin, and homeostasis model assessment-insulin resistance (HOMA-IR) levels were significantly higher in patients with OSA when compared to those with No-OSA (p<0.01). There was incremental worsening of insulin and HOMA-IR with greater severity of OSA. The apnea-hypopnea index (AHI) was positively and significantly correlated with blood glucose and HOMA-IR (p=0.01and p<0.001, respectively). Multiple linear regression analysis showed that the AHI was a predictor of blood glucose (p=0.04) and HOMA-IR (p=0.01) independent of age, gender, total sleep time and BMI z score. Logistic regression analysis showed that elevated levels of blood glucose
Page 2 of 20
predicted severe OSA (p=0.02) independent of gender and BMI z score. Elevation in HOMA-IR predicted severe OSA (p=0.004). Conclusion: OSA severity is associated with increased fasting insulin, blood glucose and
Ac ce p
te
d
M
an
us
cr
ip t
HOMA-IR even after controlling for the age, and BMI z score in adolescents.
Page 3 of 20
Introduction Obesity in children and adolescents is a major health issue in the United States and in other countries. According to a recent report from the Center for Disease Control (CDC), childhood obesity has more than doubled in children and quadrupled in adolescents in the past 30
ip t
years [1]. In a population-based study of 5 to 17 year old children, 70% of obese youth had at
cr
least one risk factor for cardiovascular disease [2]. Obesity in children can cause early
development of hypertension, cardiovascular disease, and hyperglycemia, which in later life
us
develops into diabetes. Metabolic Syndrome is a cluster of increased blood pressure (BP),
an
elevated triglyceride, cholesterol, low density lipoprotein, and glucose levels, and decreased high density lipoprotein levels. The prevalence of Metabolic Syndrome is 6.8% among overweight
M
and 28.7% among obese adolescents [2, 3]. Metabolic Syndrome, or the components of Metabolic Syndrome, increases the risk of developing atherosclerosis and cardiovascular disease,
d
i.e. heart failure, stroke, and peripheral arterial disease.
te
Obstructive sleep apnea (OSA), a disorder of breathing during sleep, has been correlated
Ac ce p
to obesity. OSA is a common clinical condition characterized by intermittent and partial airway collapse, resulting in frequent episodes of apnea, hypopnea, and recurrent arousals from sleep [24]. The estimated prevalence of OSA is 2-3% in children [3]. This risk of OSA is greatly increased by obesity, with an estimated prevalence of 36% in obese children [5]. In adults, both the obesity and OSA are shown to be associated with cardiovascular disease, high cholesterol and high blood pressure. This association is less defined in children and adolescents. Some studies in children have shown OSA to independently increases insulin resistance [5, 6], while other studies suggest that obesity is the variable increasing insulin resistance [7] [8]. In this study, we hypothesize that OSA is linked with metabolic alternations, including lipid variables, fasting glucose and insulin levels, independent of BMI z score in adolescents.
Page 4 of 20
Alterations in metabolic variables and their association with individual components of OSA, including apnea–hypopnea index (AHI), oxygen saturation (SpO2), and arousal index were studied. The relationship of metabolic variable laboratory test results and obesity were analyzed. Individual measures of the metabolic panel, particularly elevations in fasting glucose and insulin
cr
severity of OSA and metabolic abnormalities in adolescents. Methods
us
Study population
ip t
levels, were analyzed relative to OSA severity (19) to help better define the relationship between
an
The study was approved by the Institutional Review Board of Ann & Robert H. Lurie Children's Hospital of Chicago. This is a historical cohort study of consecutive patients, aged 12
M
to 18 years, having overnight polysomnography (PSG) at Ann & Robert H. Lurie Children's Hospital of Chicago between January 1, 2006 and December 31, 2012, who also had laboratory
d
testing for the metabolic variables being analyzed within 3 months of their PSG. A total of 55
te
patients were identified. Patients were excluded from analysis if they had a history of Type I
Ac ce p
diabetes (n=3), genetic abnormalities (n=3), or used a continuous positive airway pressure device (CPAP) for OSA (n=2). Other exclusions included those patients with craniofacial anomalies (n=1), organ transplant recipients (n=2), and multiple medical problems (n=2). A total of 42 patients were available for the final analysis. The age, gender, laboratory data on metabolic variables, height, weight, and PSG findings were collected from the electronic medical records. Definition and Measurements
All laboratory analyses were conducted at the Ann & Robert H. Lurie Children's Hospital laboratory. The metabolic variables including fasting glucose, and insulin levels measured were attained after an overnight fast. Abnormal fasting glucose values were determined if a patient met the American Diabetes Association guidelines (fasting glucose ≥100 mg/dl) [9]. The
Page 5 of 20
homeostasis model assessment-insulin resistance (HOMA-IR) was used as an index of insulin resistance, and this was calculated using a standard equation (fasting insulin (μIU/mL) × fasting blood glucose (mmol/L)/22.5) [10]. Based on this score, insulin resistance was determined using the HOMA-IR cut-points described by Kurtoglu et al. (>2.67 for boy and >2.22 for girls) [11].
ip t
Height and weight were recorded for each patient. The BMI z-score was computed using Center
cr
for Disease Control (CDC) growth standards (www.cdc.gov/growthcharts) and online software (www.cdc.gov/epiinfo). Patients with a BMI z score above the 95th percentile were considered
us
obese.
an
Polysomnography
A standard overnight PSG (Cadwell easy 3 versions 3.9.34, Kennewick, WA, USA) had
M
been performed on all patients. The apneas and hypopneas were identified and scored according to the American Academy of Sleep Medicine (AASM) pediatric criteria as defined in the AASM
d
Manual for Scoring of Sleep and Associated Events [12]. The apnea–hypopnea index (AHI) was
te
defined as the total number of obstructive apneas and hypopneas per hour of sleep. Severity of
Ac ce p
OSA was classified as mild (AHI between 1 and 4.99 events per hour of sleep), moderate (AHI between 5 and 9.99 events per hour of sleep), severe (AHI ≥ 10 events per hour of sleep), and No-OSA (AHI < 1 event per hour of sleep) [6] [13]. Statistical analysis
Descriptive statistics were summarized using frequencies and percentages for categorical data, and mean and standard deviations for continuous data. To determine differences between groups of No-OSA, Mild-OSA, Moderate-OSA and Severe-OSA, both parametric and nonparametric tests were used. Chi2 test of association was used for nominal data. T-test and analysis of variance statistics were used for normally distributed data. Mann Whitney U and Kruskal Wallis statistics were used for non-normally distributed data. Outliers were identified
Page 6 of 20
and checked for accuracy. Scatter plots were used to graphically show correlations among PSG and metabolic variables. Multiple linear regression models were run to predict dependent variables for log transformed fasting insulin and log-transformed HOMA-IR. Predictors included age, gender, TST, BMI z score, and AHI. Multinomial logistic regression models were run to
ip t
predict patients grouped into those with No-OSA, Mild-OSA, Moderate-OSA, and Severe-OSA.
cr
Results were considered significant with two-tailed test and p<0.05. Statistical analysis was
Inc., Chicago, IL).
an
Results
us
conducted using statistical product and service solution (SPSS) software version 20 (IBM SPSS
A total 42 patients with a mean age of 14.1±1.9 years were analyzed. Nineteen (45.2%)
M
were male. The average BMI z score was 2.23±0.86, and all patients were obese (BMI z score >95th percentile). Eighteen patients (42.8%) had an AHI <1/hour (No-OSA), 7 patients (16.6%)
d
had an AHI ≥ 1≤4.99/hour (Mild-OSA), 5 (11.9%) had an AHI ≥5/h<9.99/hour (Moderate-OSA)
te
and 12 (28.6%) had an AHI ≥ 10/hour (Severe-OSA). There was no significant difference in age
Ac ce p
(p=0.44), and BMI z score (p=0.16) between the groups. (Table 1) Patients with OSA had a significantly higher AHI and arousal index, and a lower average SpO2 and SpO2 nadir, as expected. Stage 1 total sleep time (TST) was significantly higher in those with severe OSA (p<0.001).
When comparing metabolic and PSG variables to OSA severity, there was no significant difference in HDL-C or LDL-C in those patients with No-OSA, and in those with all levels of OSA. Total cholesterol did not reach significance in the groups (p=0.05). (Table 1) Triglyceride, blood glucose, fasting insulin, and HOMA-IR levels were significantly higher in patients with OSA compare to those with No-OSA. There was incremental worsening of insulin and HOMAIR levels with greater severity of OSA: 3.9±1.8 in those with No-OSA, 7.7 ±2.1 in Mild- OSA,
Page 7 of 20
8.2±1.3 in Moderate, and 14.8±5.3 in Severe-OSA. A similar trend was observed in insulin levels. (Table 1) Further correlation analysis suggested that the AHI was positively and significantly correlated with elevated blood glucose and HOMA-IR (p=0.01 and p<0.001, respectively).
ip t
(Figure 1A & 1B) Other components of the PSG (arousal index, SpO2 nadir) followed the same
cr
trend when analyzed (data not shown). Multivariate analysis
us
Multiple linear regression analysis models assessing the effect of gender, TST, BMI z
an
score, and AHI on the log transformed fasting blood glucose and HOMA-IR values were performed. The analysis showed that the AHI was a predictor of elevated blood glucose (p=
M
0.04) and HOMA-IR (p=0.01) independent of gender, TST and BMI z score. (Table 2) A multinomial logistic regression analysis model assessing the effect of the Mild,
d
Moderate, and Severe-OSA on elevations of blood glucose and HOMA-IR indices was
te
performed. The results showed that, even after controlling for the age, gender and BMI z score,
Ac ce p
elevated levels of blood glucose and HOMA-IR predicted the Severe-OSA (p=0.02). (Table 3) Discussion
In this study, we looked at the relationship between OSA and metabolic variables, including lipids, fasting blood glucose and insulin levels, in adolescents 12-18 years of age. The study demonstrated that serum triglycerides levels (p<0.001), fasting blood glucose levels (p=0.005), and fasting insulin and HOMA-IR levels (p<0.001) were significantly higher in obese adolescents with OSA when compared to those with No-OSA. Redline et al. [3] reported sleep disordered breathing (SDB) as a risk factor for Metabolic Syndrome in adolescents. Her study showed that obesity and Metabolic Syndrome were more
Page 8 of 20
prevalent in the SDB group when compared with the non-SDB group. A seven fold increased chance for Metabolic Syndrome in those with SDB relative to unaffected children was described. The association between many components of the Metabolic Syndrome and SDB, even after adjusting for BMI, suggested that SDB contributed to metabolic dysfunction beyond the effects
ip t
weight alone.
cr
Canapari et al. [14] showed that children (5-18 years of age) with OSA had significantly increased BMI, HOMA-IR, triglycerides, and leptin levels compared to those without OSA. In
us
contrast, Tauman et al. [15], found no correlation between the AHI and serum insulin, serum
an
glucose, HOMA-IR and triglyceride levels in any of the children analyzed regardless of weight. The prior study showed that insulin resistance and dyslipidemia were determined primarily by
M
the degree of body adiposity, rather than the severity of SDB. Given the conflicting evidence from prior studies in adolescents, we expanded our previous work and explored the relationship
d
between obesity, OSA, and alterations in metabolic variables in adolescents. In our previously
te
published study on children aged 3-11 years, we found that fasting blood glucose and insulin
Ac ce p
levels were positively and significantly associated with the AHI. These finding remain consistent in the current study.
Insulin resistance is a physiological condition in which the body produces insulin, but cells in the body are unable to use it effectively. As a result, glucose builds up in the blood instead of being absorbed by the cells, leading to type II diabetes or pre-diabetic conditions. Our previous study in younger children [13], and our present study on adolescents, showed that patients classified with Severe-OSA had significantly higher blood glucose and HOMA-IR levels. An elevation in triglycerides, blood glucose, fasting insulin and HOMA-IR were associated with increased AHI, arousal index, and SpO2 nadir in the patients studied.
Page 9 of 20
To further clarify this correlation, additional multivariate analyses were performed. Significant p-values for the blood glucose and HOMA-IR in multinomial logistic analysis of our data suggested that, even after controlling for the age, gender and BMI z score, higher levels of blood glucose and HOMA-IR values were more likely to fall in the severe OSA group (Table-3).
ip t
We found elevated lipid levels with increasing severity of OSA, even after adjusting for BMI z
cr
score. There has been concern that the clustering of abnormalities in variables like insulin, lipid levels, and elevated blood pressure (Metabolic Syndrome) could be associated with development
us
of cardiovascular complications in later ages. Early identification and treatment of the children
an
with OSA and Metabolic Syndrome becomes essential to decrease morbidity into early adulthood.
M
The neuroendocrine and adverse cardiovascular outcomes have been linked with OSA as well as with obesity [16]. Floren et al [17], reported several different mechanisms by which
d
OSA can cause hypertension, atherosclerosis and linked it to the cardiovascular problems later in
te
the life. One study on adults showed that CPAP for severe OSA improved hypertension severity
Ac ce p
[18]. There is lack of such data in the pediatric population. Patients with OSA experienced varying severity of hypoxic episodes and recurrent arousals during sleep due to sympathetic over activity. Different mechanisms like chemo baro-reflex changes, vasoconstriction by endothelin release and endothelial dysfunction have been described which increases hepatic glycogenolysis and gluconeogenesis [7, 19-21]. These sympathetic activities cause stimulation of lipolysis, which increases circulating free fatty acid metabolites. These inhibit insulin signaling and reduce insulin-mediated glucose uptake, contributing to insulin resistance. Hypoxia induced by OSA may cause oxidative stress, which leads to the production of Reactive Oxygen Species. These events contribute to altered glucose homeostasis by impairing pancreatic function and insulin secretion [7].
Page 10 of 20
Adenotonsillectomy is the most common treatment offered to the children with OSA [22]. Gozal et al. [6] reported that the surgical removal of hypertrophic tonsils and adenoids led to significant improvements in lipid variables, C-reactive protein, and apolipoprotein B. Their study included the overweight (BMIz score>1.20) as well as non-obese patients. Authors
ip t
reported improvements in the metabolic variables including fasting Insulin and glucose in the
cr
obese patients (n=37) with OSA after 6-8 months of adenotonsillectomy. Waters et al. [23]
showed that adenotonsillectomy for OSA successfully improved the respiratory disturbance
us
index (RDI) in the majority of children. In that study, twenty children underwent surgical
an
removal of adenotonsillar tissue. The authors found a small but significant improvement in total cholesterol in those children whose OSA was resolved and a trend for obese children with
M
persisting OSA to have elevated insulin levels compared with obese children without OSA. The present study was conducted retrospectively, and we did not look at the effect of
d
OSA intervention on improvements in metabolic alterations and insulin resistance. The sample
te
size was small, since a convenience sample of patients having both a PSG and laboratory
Ac ce p
analysis was limiting. This may have led to type II errors in failing to detect correlations between OSA and metabolic alterations. Despite this, important relationships between these variables were detected. Further research is needed to evaluate the overall effect of CPAP and adenotonsillectomy on metabolic alterations in obese and non-obese adolescents with OSA. Conclusions
OSA severity was associated with increased HOMA-IR and blood glucose levels, even after controlling for the BMI, in adolescents. Identification and understanding the mechanism and interaction of OSA with abnormal metabolic variables is important in children. Addressing these interactions at early stages may benefit patients by allowing for more timely diagnosis and treatment to help to reduce the risk of diabetes and cardiovascular disease.
Page 11 of 20
Acknowledgements: We thank Karen Rychlik, MS of the Biostatistics Research Core of the Ann
Ac ce p
te
d
M
an
us
cr
ip t
& Robert H. Lurie Children’s Hospital of Chicago for her assistance with the statistical analysis.
Page 12 of 20
References [1] Center for Disease Control. Prevention Status Reports. 2013. [2] Punjabi NM. The epidemiology of adult obstructive sleep apnea. Proc Am Thoracic Soc. 2008;5:136-143.
ip t
[3] Redline S, Storfer-Isser A, Rosen CL, Johnson NL, Kirchner HL, et al. Association between metabolic syndrome and sleep-disordered breathing in adolescents. Am Journal Resp Critical
cr
Care Med. 2007;176:401-408.
[4] Penn EB, French A, Bhushan B, Schroeder JW. Access to care for children with symptoms of
us
sleep disordered breathing. Int J Ped Otorhinolaryngol. 2012;76:1671-1673.
[5] Lanfranco F, Motta G, Minetto MA, Baldi M, Balbo M, et al. Neuroendocrine alterations in obese patients with sleep apnea syndrome. Int J Endocrinol. 2010;2010:474-518.
an
[6] Gozal D, Capdevila OS, Kheirandish-Gozal L. Metabolic alterations and systemic Resp Crit Care Med. 2008;177:1142-1149.
M
inflammation in obstructive sleep apnea among nonobese and obese prepubertal children. Am J [7] Punjabi NM, Shahar E, Redline S, Gottlieb DJ, Givelber R, et al. Sleep Heart Health Study: Sleep-disordered breathing, glucose intolerance, and insulin resistance: the Sleep Heart Health
d
Study. Am J Epidemiol. 2004;160:521-530.
te
[8] Nakra N, Bhargava S, Dzuira J, Caprio S, Bazzy-Asaad A. Sleep-disordered breathing in children with metabolic syndrome: the role of leptin and sympathetic nervous system activity
Ac ce p
and the effect of continuous positive airway pressure. Pediatrics. 2008;122:e634-642. [9] Expert Committee on the Classification of Diabetes: Report of the expert committee on the diagnosis and classification of diabetes mellitus. Diabetes Care. 2003;26(1):S5-20. [10] Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, et al. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28:412-419. [11] Kurtoglu S, Hatipoglu N, Mazicioglu M, Kendirici M, Keskin M, et al. Insulin resistance in obese children and adolescents: HOMA-IR cut-off levels in the prepubertal and pubertal periods. J Clin Res Ped Endocrinol. 2010;2:100-106. [12] Berry RB, Budhiraja R, Gottlieb DJ, Gozal D, Iber C, et al. American Academy of Sleep, rules for scoring respiratory events in sleep: update of the 2007 AASM Manual for the Scoring
Page 13 of 20
of Sleep and Associated Events. Deliberations of the Sleep Apnea Definitions Task Force of the American Academy of Sleep Medicine. J Clin Sleep Med. 2012;8:597-619. [13] Bhushan B, Maddalozzo J, Sheldon SH, Haymond S, Rychlik K, et al. Metabolic alterations in children with obstructive sleep apnea. Int J Ped Otorhinolaryngol. 2014;78.854-859. [14] Canapari CA, Hoppin AG, Kinane TB, Thomas BJ, Torriani M, et al. Relationship between
ip t
sleep apnea, fat distribution, and insulin resistance in obese children. J Clin Sleep Med. 2011;7:268-273.
cr
[15] Tauman R, O'Brien LM, Ivanenko A, Gozal D. Obesity rather than severity of sleep-
disordered breathing as the major determinant of insulin resistance and altered lipidemia in
us
snoring children. Pediatrics. 2005;116:e66-73.
[16] Grassi G, Facchini A, Trevano FQ, Dell'Oro R, Arenare F, et al. Obstructive sleep apnea-
an
dependent and -independent adrenergic activation in obesity. Hypertension. 2005;46:321-325. [17] Floras JS. Obstructive sleep apnea syndrome, continuous positive airway pressure and treatment of hypertension. Eur J Pharmacol. 2015.
M
[18] Jelic S, Lederer DJ, Adams T, Padeletti M, Colombo PC, et al. Vascular inflammation in obesity and sleep apnea. Circulation. 2010:210:1014-1021.
d
[19] Narkiewicz K, Somers VK. Sympathetic nerve activity in obstructive sleep apnoea. Acta Physiologica Scan. 2003;177:385-390.
Ac ce p
Rev Cardiol. 2010;7:677-685.
te
[20] Kohler M, Stradling JR. Mechanisms of vascular damage in obstructive sleep apnea. Nature [21] Bhushan B, Khalyfa A, Spruyt K, Kheirandish-Gozal L, Capdevila OS, et al. Fatty-acid binding protein 4 gene polymorphisms and plasma levels in children with obstructive sleep apnea. Sleep Med. 2011;12:666-671.
[22] Bhushan B, Sheldon S, Wang E, Schroeder JW. Clinical indicators that predict the presence of moderate to severe obstructive sleep apnea after adenotonsillectomy in children. Am J Otolaryngol. 2014;35:487-495.
[23] O'Brien LM, Sitha S, Baur LA, Waters KA. Obesity increases the risk for persisting obstructive sleep apnea after treatment in children. Int J Ped Otorhinolaryngol. 2006;70:15551560.
Page 14 of 20
Table 1. The relationship between the severity of obstructive sleep apnea (OSA) and metabolic variables in adolescents are shown. Controls (n=18)
Mild OSA (n=7)
Moderate OSA (n=5)
Severe OSA (n=12)
p-Value ANOVA
Age (years)
14.1±2.1
14.3±2.0
15.1±1.6
13.5±1.3
0.44
BMI z score
1.9±0.9
2.4±0.68
2.2±0.5
2.6±0.8
0.16
0.51±0.4
3.1±0.9
TST (hours)
5.4± 1.3
5.4 ±1.1
Stage 1 (%)
4.8 ± 2.8
4.3 ± 3.2
Stage 2 (%)
56.8 ± 5.6
SWS (%)
25.8 ± 5.8
REM (%)
12.6 ± 7.0
Arousal index (per hour)
0.51 ±0.4
Average SaO2 (%)
96.8 ± 1.4
0.2
7.6±1.1
37±23.4
<0.001
6.0 ±1.1
5.0 ± 0.8
0.41
3.6 ± 2.3
11.8 ± 6.7
<0.001
49.6 ±8.7
62.4 ± 6.9
56.3 ± 12.7
0.11
32.3 ±6.9
20.8 ± 3.1
25.9 ± 10.0
0.06
13.6 ± 6.9
13± 7.7
7.9 ± 7.7
0.26
7.6 ± 1.1
36.9 ± 23.3
<0.001
95.7 ±2.6
96.6 ±1.1
94.1 ± 2.4
0.01
87.6 ± 7.4
88.4 ±4.1
86 ± 4.9
0.03
d
3.1 ± 0.9
te
91.44 ± 3.9
cr
AHI (per hour)
13.3 ± 12.4
us
38.6 ±46.1
an
50 ± 54.7
SpO2 Nadir (%)
20.9 ±22.8
M
Sleep Latency (mins)
Ac ce p
Polysomnography variables
ip t
Variables
Metabolic variables (mg/dl)
Total Cholesterol
147.2 ± 26.8
150.6 ± 27.5
150.8 ± 5.8
174.7 ± 29.3
0.05
Triglycerides
100.1 ± 36.2
129.1 ± 38.1
116.2 ± 16.7
157 ± 30.3
0.001
39.3 ± 9.9
37.3 ± 7.9
34.4 ± 14.5
39.1 ± 10.1
0.79
84.1 ± 19.4
103.8 ±62.2
117.6 ± 78.3
99.5 ± 32.4
0.38
Blood Glucose
93.7 ± 5.8
96.4 ± 6.3
93.6 ± 4.4
102.3 ± 6.8
0.005
Fasting Insulin (µIU/ml)
18.4 ±7.3
31.8 ± 7.8
35.7 ± 4.7
58.1 ± 20.7
<0.001
HOMA-IR
3.9 ± 1.8
7.7 ± 2.1
8.256 ± 1.3
14.8 ± 5.3
<0.001
HDL-C LDL-C
Page 15 of 20
BMI, body mass index; AHI, apnea–hypopnea index; TST, total sleep time; SWS, slow wave sleep; REM, rapid eye movement; HDL-C, high density lipoprotein-cholesterol; TC, total cholesterol; LDL-C, low density lipoprotein-cholesterol; TG, triglycerides; HOMA-IR,
Ac ce p
te
d
M
an
us
cr
ip t
homeostasis model assessment. p<0.05 was considered significant and are highlighted bold.
Page 16 of 20
Table 2. Multiple linear regression analysis models assessing the independent effect of different variables (gender, BMI z score, and AHI) on log-transformed fasting insulin or HOMA-IR values. P<0.05 was significant. Dependent variable
Dependent variable
β(standardized
p-Value
coefficient)
coefficient)
Gender
0.14
0.28
-0.15
BMI z score
0.28
0.63
0.12
AHI
0.50
<0.001
0.45
0.34
0.46
0.01
an
(per hour)
p-Value
cr
variables
β(standardized
us
Independent
Blood glucose mg/dl
ip t
HOMA-IR
Ac ce p
te
d
M
BMI, body mass index; AHI, apnea hypopnea index.
Page 17 of 20
Table 3. Multinomial logistic regression model assessing the effect of mild, moderate, and severe OSA on elevated fasting insulin, and HOMA-IR index after controlling for age, gender and BMI z score. P<0.05 was considered significant. β-coefficient
p-Value
O.R
Age (years)
0.13
0.62
1.13
Gender
0.65
0.52
1.92
BMI z score
0.81
0.25
2.24
HOMA-IR
0.07
0.53
1.07
0.86
1.34
Moderate
Age (years)
0.26
0.32
1.29
0.78
2.16
OSA
Gender
0.09
0.93
1.10
0.11
10.62
BMI z score
0.43
0.54
1.54
0.38
6.13
HOMA-IR
0.08
Age(years)
-0.49
Gender
-0.87
BMI z score
0.33
HOMA-IR
95% CI
ip t
0.68
1.88
0.27
13.82
0.56
8.88
cr
us
Mild OSA
an
Independent variables
1.09
0.85
1.39
0.25
0.61
0.26
1.41
0.48
0.42
0.03
4.78
0.51
1.39
0.15
13.34
0.43
0.004
1.53
1.14
2.06
0.122
0.81
1.13
0.43
2.96
-0.17
0.89
0.84
0.08
8.59
0.53
0.54
1.70
0.3
9.34
Glucose mg/dl
0.03
0.75
1.03
0.84
1.26
Moderate
Age
0.86
0.07
2.35
0.94
5.9
OSA
Gender
-0.58
0.60
0.55
0.06
5.02
BMI z score
0.81
0.38
2.25
0.37
13.72
Glucose mg/dl
0.03
0.73
1.03
0.86
1.25
Age
0.22
0.24
0.63
0.52
3.1
Gender
0.29
0.80
1.34
0.14
12.90
BMI z score
1.17
0.19
3.23
0.54
19.51
Glucose mg/dl
0.37
0.02
1.43
1.05
1.95
Age Gender
Ac ce p
BMI z score
d
Mild OSA
te
Severe OSA
Severe OSA
M
0.57
OSA, obstructive sleep apnea, BMI- body mass index, HOMA-IR – homeostatic model assessment - Insulin resistance
Page 18 of 20
Figure – 1 A, B
Ac ce p
te
d
M
an
us
cr
ip t
A
Page 19 of 20
ip t cr us an M d te Ac ce p
B
Figure. Scatter plots of individual blood glucose and HOMA-IR values relative to the AHI. A. A significant correlation between the AHI and the fasting insulin was observed. B. A significant correlation between the AHI and the HOMA-IR index was observed.
Page 20 of 20