ORIGINAL ARTICLE: NUTRITION
Short Sleep Duration Is Related to Emerging Cardiovascular Risk Factors in Obese Children Marı´a Navarro-Solera, yJoaquı´n Carrasco-Luna, zGonzalo Pin-Arboledas, Rebeca Gonza´lez-Carrascosa, §Jose´ M. Soriano, and jjPilar Codon˜er-Franch
ABSTRACT Objective: The aim of the present study was to evaluate the influence of sleep duration on cardiovascular risk factors in obese children. Methods: Cross-sectional analysis of 90 obese children ages 7 to 16 years. Anthropometric and clinical evaluation with specification of dietary and lifestyle habits was carried out during an office visit. Sleep duration was evaluated by the BEARS (B ¼ bedtime issues, E ¼ excessive daytime sleepiness, A ¼ night awakening, R ¼ regularity and duration of sleep, S ¼ snoring) questionnaire on children’s sleep characteristics. Sleep time adequacy by age was assessed according to the criteria of the National Sleep Foundation. Biochemical blood variables indicative of metabolic risk (glucose, lipid profile, and insulin) were obtained. Emergent new factors of metabolic risk, including high-sensitive C-reactive protein, g-glutamyltranspeptidase, homocysteine, retinol-binding protein 4 (RBP4), thyroid-stimulating hormone (TSH), inflammatory markers, and the adipokines leptin, adiponectin, and ghrelin were also evaluated. The relations between the duration of sleep and these variables were analyzed by general lineal model analysis. Significant variables were introduced in logistic regression analysis to determine the odds ratio (OR) and 95% confidence interval (CI) of cardiometabolic factors with respect to sleep. Results: Children who slept for short duration were significantly more at risk of severe central obesity. In the regression model, the mean arterial pressure (odds ratio [OR] 1.10, 95% confidence interval [CI] 1.02–1.17, P ¼ 0.008), homocysteine (OR 1.41, 95% CI 1.08–1.84, P ¼ 0.013), RBP4 (OR 1.78, 95% CI 1.15–2.78, P ¼ 0.010), and TSH (OR 2.01, 95% CI 1.21–3.34, P ¼ 0.007) remain as significant independent predictors related to short sleep duration. We did not find any association between sleep duration and inflammatory markers or adipokines. Conclusions: Short sleep duration increases the severity of obesity and is related to cardiovascular risk factors in children. Key Words: cardiovascular risk, homocysteine, obesity, retinol-binding protein, sleep duration
(JPGN 2015;61: 571–576) Received February 28, 2015; accepted May 15, 2015. From the Department of Pediatrics, Obstetrics, and Gynecology, University of Valencia, the yDepartment of Experimental Science, School Catholic University of Valencia, the zSleep Unit, Hospital Quiro´n, the §Department of Preventive Medicine and Public Health, University of Valencia, and the jjDepartment of Pediatrics, Dr Peset University Hospital, Valencia, Spain. Address correspondence and reprint requests to Pilar Codon˜er-Franch, Department of Pediatrics, Dr Peset University Hospital, Avenida Gaspar Aguilar 90, 46017 Valencia, Spain (e-mail: pilar.codoner@ uv.es). The authors report no conflicts of interest. Copyright # 2015 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition DOI: 10.1097/MPG.0000000000000868
JPGN
Volume 61, Number 5, November 2015
What Is Known
Sleep is an important factor related to obesity and cardiovascular risk, primarily in situations of obstructive sleep apnea. Short sleep duration has been correlated with metabolic risk factors in adults. A definite relation between sleep duration and childhood obesity has not yet been established.
What Is New
Obesity, abdominal adiposity, and blood pressure are related to reduced sleep duration. Levels of homocysteine, retinol-binding protein 4, and thyroid-stimulating hormone are associated with short sleep duration. Short sleep duration is associated with cardiovascular risk factors in childhood.
O
besity results from the combination of genetic and environmental factors. Among the lifestyle factors affecting obesity, the quality and quantity of sleep has been recognized as an important factor in recent years. It has been proven that short sleep duration results in increased body mass index (BMI) in children (1) and adults (2), although other authors have not found any relation between sleep duration and a major prevalence of obesity (3,4). The effect of sleep could be because of alterations in endocrine and metabolic control, lipid intake, thermoregulation, and decrease in physical activity, which are present in subjects with short sleeping hours (5,6). In turn, body fat mass could be an independent predictor of greater influence of sleep latency (7). In children, the impact of disturbed and inadequate sleep can be noticeable. During sleep, there is a peak in the secretion of growth hormone. Sleep duration also affects learning, mood regulation and behavior, dietary habits, and quality of life (8), each of which can be involved in obesity. Short sleep duration has been correlated with an increase in metabolic risk factors and can predict the metabolic syndrome in adults (9). Furthermore, short sleep duration is associated with decreased insulin sensitivity, lipid profile abnormalities, and high blood pressure (BP) in adolescents (10–12). Likewise, this association can also affect the emergent new factors of metabolic risk, including the inflammatory markers C-reactive protein, tumor necrosis factor a, and interleukin 6 (13). In childhood, primarily in obese children, to identify the presence of cardiovascular risk should be a priority because atherogenesis can precede its clinical manifestations by many years, and preventive measures can be
571
Copyright 2015 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.
˜ er-Franch et al Codon
JPGN
undertaken. Determination of metabolic risk factors, including several new phenotypic markers, enables a more accurate prediction of cardiovascular risk (14). Recently, uric acid, homocysteine, gglutamyltranspeptidase (GGT) and thyroid-stimulating hormone (TSH) have been associated with cardiovascular risk (14,15). Adipokines such as leptin and adiponectin are bioactive molecules that link obesity to cardiovascular risk (16), and ghrelin is a hormone secreted by the stomach, which has a major effect on appetite and hence on obesity. Retinol-binding protein 4 (RBP4) is another emerging factor of metabolic risk that has been linked to comorbidities in obese subjects (17). To our knowledge, there is limited information about the effects of sleep duration on these new biomarkers related to cardiovascular risk. In this context, the objective of the present study was to evaluate the influence of short sleep duration on cardiovascular risk in obese children. Our primary outcome was the effect of sleep duration on traditional cardiovascular risk factors (obesity, carbohydrate metabolism, high-density lipoprotein cholesterol and triglycerides levels, and BP). Secondary outcomes examined the effect of sleep duration on other cardiovascular risk factors (uric acid, homocysteine, RBP4, GGT, and TSH), inflammatory markers, and adipokines such as leptin, adiponectin and ghrelin, physical activity, and quality of diet.
METHODS Study Population A cross-sectional study was conducted between January and June 2014. The key inclusion criterion was the presence of obesity defined according to the international criteria (BMI for age and sex 95th percentile using the World Health Organization tables as reference (18)). Consecutive sampling was carried out among children ages 7 to 16 years referred to the outpatient pediatric nutrition office of Dr Peset University Hospital of Valencia for diagnosis of obesity and comorbidities. All of the participants completed information concerning demographic, personal, and family data. Important medical history items that were sought included family and personal antecedent of sleep alteration, and any medications presently being administered. Exclusion criteria were the presence of known genetic disorders, hormonal disorders or chronic illness, or medication use. Children suspected to have sleep apnea (snoring, excessive daytime sleepiness, or morning headache) were also excluded. The study was approved by the local ethics committee. All of the patients and their parents signed an informed consent form.
Clinical Data Every child was subjected to a comprehensive physical examination and an otolaryngology assessment. Anthropometric measurements (weight, height, waist and hip circumference, and skinfolds) were done by a nutritionist with specific accreditation of the International Society for the Advancement of Kinanthropometry. BMI was calculated as weight/height2, and the z score according to the 50th percentile for age and sex was obtained. Waist circumference was converted to normalized measurements by comparison with the 50th standard percentile for age and sex (relative waist circumference) (19). The skinfolds (triceps, biceps, subscapular, and suprailiac) were measured with a Holtain skinfold caliper (0.2 mm of precision; Holtain Ltd, Croswell, UK) in triplicate; z scores were calculated according to the Identification and Preservation of Dietary- and Lifestyle-Induced Health Effects in Children and Infants (IDEFICS) study (20). Body composition was determined by bioelectrical impedance using a Tanita BC-418 MA with 8 contact electrodes (Tanita Europe BV, Hoofddorp, the Netherlands). BP was measured at rest in the dominant arm, and the average of 3 measurements was considered. The BP z score was calculated using
572
Volume 61, Number 5, November 2015
BP tables by sex, age, and height (21). The pulse pressure (systolic BP minus diastolic BP) and mean arterial pressure (MAP, diastolic BP plus pulse pressure/3) were also calculated. Sleep information was evaluated using the BEARS questionnaire (B ¼ Bedtime issues, E ¼ Excessive daytime sleepiness, A ¼ Night awakening, R ¼ Regularity and duration of sleep, S ¼ Snoring) (22). Because sleep needs vary with age, each item involves an age-specific question. Sleep duration was taken from the ‘‘R’’ item of the BEARS questionnaire. Sleep time recommendations were categorized by age according to the National Sleep Foundation (http://www.sleepfoundation.org). Sleep duration less than the reference was considered as the short sleep duration category. According to sleep duration, children were classified into 2 groups: children who sleep the recommended hours and children who sleep fewer hours than recommended or the short sleep group. Dietary data collection was obtained by three 24-hour recall interviews conducted by a nutritionist. Detailed description of type, form, and amount of food consumed and the hour of day and possibility of snacking with frequency, type, and timing of snacks was collected. We evaluated the nutritional content using a validated nutritional software package (DIAL; Alce Ingenierı´a SA, Madrid, Spain, http://www.alceingenieria.net/nutricion.htm) for assessing diets and food calculations. Adherence to the Mediterranean diet was assessed by the Mediterranean Diet Quality Index for Children and Adolescents (KIDMED) questionnaire (23). Information about lifestyle habits was also obtained. The Krece Plus questionnaire was used to assess each child’s level of activity/inactivity (24). For this test, weekly physical activity during leisure time is scored positively, and daily hours spent watching television or playing computer games are scored negatively. The sum of the activities is classified by sex into 3 levels: bad score, regular score, and good score.
Biochemical Analysis Blood samples were collected after 12 hours of fasting for routine clinical analysis. Blood samples were processed and centrifuged at 3000g at 48C for 5 minutes. Plasma and serum aliquots were deep-frozen until needed for further analysis. Routine biochemical parameters (glucose, cholesterol and fractions, triglycerides and uric acid) were measured by automated direct methods (Aeroset System; Abbott Clinical Chemistry, Wiesbaden, Germany) and insulin and homocysteine by an automated electrochemiluminescence immunoassay (Architect c8000; Abbott Clinical Chemistry). Insulin resistance was assessed for the homeostasis model assessment (HOMA) index (insulin [mU/L] glucose [mmol/L]/ 22.5). High-sensitivity C-reactive protein and RBP4 were analyzed by immunonephelometry using a Behring nephelometer 2 (Dade Behring, Hamburg, Germany). Serum GGT was measured by an enzymatic colorimetric test at 378C in a Roche/Hitachi analyzer (Roche Diagnostics, Mannheim, Germany). TSH was measured using a direct immunoradiometric assay (Unicel DxI 800; Beckman Coulter, Barcelona, Spain). We used a multiplex immunoassay with MILLIPLEX MAP Human kits specific for leptin, adiponectin, ghrelin, and tumor necrosis factor a from Merck Millipore (Darmstad, Germany). The plates were read in a cytometer (Luminex 100TM IS analyzer; Luminex Corporation, Austin, TX). The standard curve was calculated using 5-parametric curve fitting, and the results were analyzed using Luminex 100 IS 2.3 software (Luminex Corporation).
Definitions To evaluate the risk of metabolic syndrome in children, the following variables were taken into account as metabolic risk www.jpgn.org
Copyright 2015 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.
JPGN
Volume 61, Number 5, November 2015
factors according to the criteria of the International Diabetes Federation consensus (25): obesity with a z score BMI 2, highdensity lipoprotein cholesterol 40 mg/dL, triglycerides >110 mg/ dL, and hypertension (systolic or diastolic BP >95th percentile for age and sex). We evaluated carbohydrate metabolism by insulin resistance defined as HOMA 3.16 (26). In addition, uric acid, homocysteine, high-sensitivity C-reactive protein, RBP4, GGT, and TSH were considered as emergent new factors of metabolic risk. Leptin, adiponectin, ghrelin, and tumor necrosis factor were included in this group of factors.
Statistical Analysis Sample size was estimated assuming that a difference between groups in z score BMI of 0.7 is clinically relevant at a significance level of a ¼ 5%. A total of 76 patients (38 patients in each group) were deemed necessary for 90% power. If we take into account an estimated loss of 15%, a final sample size of 87 obese children was needed. The Kolmogorov-Smirnov test was used for testing normality. In the case of nonnormal distributions, natural logarithmic transformation was used to improve normality. Means and standard deviations (mean SD) were calculated for continuous data. Metabolic risk factors (independents variables) were examined across sleep duration (dependent variable) using a general linear model controlling for sex, age, BMI, physical activity, and quality of Mediterranean diet as appropriate. Stepwise binary logistic regression analyses were carried out to evaluate the strength of the association between sleep duration and variables with significant differences on general linear model analysis. Sleep duration was analyzed as the response variable (sleep duration according to the recommendation ¼ 0 and sleep duration less than the recommendation ¼ 1). Analyses were performed using Statistical Package for the Social Sciences version 17.0 (SPSS Inc, Chicago, IL). All of the P values were reported with statistical significance set at
Short Sleep Duration Is Related to Emerging Cardiovascular Risk Factors in Obese Children Marı´a Navarro-Solera, yJoaquı´n Carrasco-Luna, zGonzalo Pin-Arboledas, Rebeca Gonza´lez-Carrascosa, §Jose´ M. Soriano, and jjPilar Codon˜er-Franch
ABSTRACT Objective: The aim of the present study was to evaluate the influence of sleep duration on cardiovascular risk factors in obese children. Methods: Cross-sectional analysis of 90 obese children ages 7 to 16 years. Anthropometric and clinical evaluation with specification of dietary and lifestyle habits was carried out during an office visit. Sleep duration was evaluated by the BEARS (B ¼ bedtime issues, E ¼ excessive daytime sleepiness, A ¼ night awakening, R ¼ regularity and duration of sleep, S ¼ snoring) questionnaire on children’s sleep characteristics. Sleep time adequacy by age was assessed according to the criteria of the National Sleep Foundation. Biochemical blood variables indicative of metabolic risk (glucose, lipid profile, and insulin) were obtained. Emergent new factors of metabolic risk, including high-sensitive C-reactive protein, g-glutamyltranspeptidase, homocysteine, retinol-binding protein 4 (RBP4), thyroid-stimulating hormone (TSH), inflammatory markers, and the adipokines leptin, adiponectin, and ghrelin were also evaluated. The relations between the duration of sleep and these variables were analyzed by general lineal model analysis. Significant variables were introduced in logistic regression analysis to determine the odds ratio (OR) and 95% confidence interval (CI) of cardiometabolic factors with respect to sleep. Results: Children who slept for short duration were significantly more at risk of severe central obesity. In the regression model, the mean arterial pressure (odds ratio [OR] 1.10, 95% confidence interval [CI] 1.02–1.17, P ¼ 0.008), homocysteine (OR 1.41, 95% CI 1.08–1.84, P ¼ 0.013), RBP4 (OR 1.78, 95% CI 1.15–2.78, P ¼ 0.010), and TSH (OR 2.01, 95% CI 1.21–3.34, P ¼ 0.007) remain as significant independent predictors related to short sleep duration. We did not find any association between sleep duration and inflammatory markers or adipokines. Conclusions: Short sleep duration increases the severity of obesity and is related to cardiovascular risk factors in children. Key Words: cardiovascular risk, homocysteine, obesity, retinol-binding protein, sleep duration
(JPGN 2015;61: 571–576) Received February 28, 2015; accepted May 15, 2015. From the Department of Pediatrics, Obstetrics, and Gynecology, University of Valencia, the yDepartment of Experimental Science, School Catholic University of Valencia, the zSleep Unit, Hospital Quiro´n, the §Department of Preventive Medicine and Public Health, University of Valencia, and the jjDepartment of Pediatrics, Dr Peset University Hospital, Valencia, Spain. Address correspondence and reprint requests to Pilar Codon˜er-Franch, Department of Pediatrics, Dr Peset University Hospital, Avenida Gaspar Aguilar 90, 46017 Valencia, Spain (e-mail: pilar.codoner@ uv.es). The authors report no conflicts of interest. Copyright # 2015 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition DOI: 10.1097/MPG.0000000000000868
JPGN
Volume 61, Number 5, November 2015
What Is Known
Sleep is an important factor related to obesity and cardiovascular risk, primarily in situations of obstructive sleep apnea. Short sleep duration has been correlated with metabolic risk factors in adults. A definite relation between sleep duration and childhood obesity has not yet been established.
What Is New
Obesity, abdominal adiposity, and blood pressure are related to reduced sleep duration. Levels of homocysteine, retinol-binding protein 4, and thyroid-stimulating hormone are associated with short sleep duration. Short sleep duration is associated with cardiovascular risk factors in childhood.
O
besity results from the combination of genetic and environmental factors. Among the lifestyle factors affecting obesity, the quality and quantity of sleep has been recognized as an important factor in recent years. It has been proven that short sleep duration results in increased body mass index (BMI) in children (1) and adults (2), although other authors have not found any relation between sleep duration and a major prevalence of obesity (3,4). The effect of sleep could be because of alterations in endocrine and metabolic control, lipid intake, thermoregulation, and decrease in physical activity, which are present in subjects with short sleeping hours (5,6). In turn, body fat mass could be an independent predictor of greater influence of sleep latency (7). In children, the impact of disturbed and inadequate sleep can be noticeable. During sleep, there is a peak in the secretion of growth hormone. Sleep duration also affects learning, mood regulation and behavior, dietary habits, and quality of life (8), each of which can be involved in obesity. Short sleep duration has been correlated with an increase in metabolic risk factors and can predict the metabolic syndrome in adults (9). Furthermore, short sleep duration is associated with decreased insulin sensitivity, lipid profile abnormalities, and high blood pressure (BP) in adolescents (10–12). Likewise, this association can also affect the emergent new factors of metabolic risk, including the inflammatory markers C-reactive protein, tumor necrosis factor a, and interleukin 6 (13). In childhood, primarily in obese children, to identify the presence of cardiovascular risk should be a priority because atherogenesis can precede its clinical manifestations by many years, and preventive measures can be
571
Copyright 2015 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.
˜ er-Franch et al Codon
JPGN
undertaken. Determination of metabolic risk factors, including several new phenotypic markers, enables a more accurate prediction of cardiovascular risk (14). Recently, uric acid, homocysteine, gglutamyltranspeptidase (GGT) and thyroid-stimulating hormone (TSH) have been associated with cardiovascular risk (14,15). Adipokines such as leptin and adiponectin are bioactive molecules that link obesity to cardiovascular risk (16), and ghrelin is a hormone secreted by the stomach, which has a major effect on appetite and hence on obesity. Retinol-binding protein 4 (RBP4) is another emerging factor of metabolic risk that has been linked to comorbidities in obese subjects (17). To our knowledge, there is limited information about the effects of sleep duration on these new biomarkers related to cardiovascular risk. In this context, the objective of the present study was to evaluate the influence of short sleep duration on cardiovascular risk in obese children. Our primary outcome was the effect of sleep duration on traditional cardiovascular risk factors (obesity, carbohydrate metabolism, high-density lipoprotein cholesterol and triglycerides levels, and BP). Secondary outcomes examined the effect of sleep duration on other cardiovascular risk factors (uric acid, homocysteine, RBP4, GGT, and TSH), inflammatory markers, and adipokines such as leptin, adiponectin and ghrelin, physical activity, and quality of diet.
METHODS Study Population A cross-sectional study was conducted between January and June 2014. The key inclusion criterion was the presence of obesity defined according to the international criteria (BMI for age and sex 95th percentile using the World Health Organization tables as reference (18)). Consecutive sampling was carried out among children ages 7 to 16 years referred to the outpatient pediatric nutrition office of Dr Peset University Hospital of Valencia for diagnosis of obesity and comorbidities. All of the participants completed information concerning demographic, personal, and family data. Important medical history items that were sought included family and personal antecedent of sleep alteration, and any medications presently being administered. Exclusion criteria were the presence of known genetic disorders, hormonal disorders or chronic illness, or medication use. Children suspected to have sleep apnea (snoring, excessive daytime sleepiness, or morning headache) were also excluded. The study was approved by the local ethics committee. All of the patients and their parents signed an informed consent form.
Clinical Data Every child was subjected to a comprehensive physical examination and an otolaryngology assessment. Anthropometric measurements (weight, height, waist and hip circumference, and skinfolds) were done by a nutritionist with specific accreditation of the International Society for the Advancement of Kinanthropometry. BMI was calculated as weight/height2, and the z score according to the 50th percentile for age and sex was obtained. Waist circumference was converted to normalized measurements by comparison with the 50th standard percentile for age and sex (relative waist circumference) (19). The skinfolds (triceps, biceps, subscapular, and suprailiac) were measured with a Holtain skinfold caliper (0.2 mm of precision; Holtain Ltd, Croswell, UK) in triplicate; z scores were calculated according to the Identification and Preservation of Dietary- and Lifestyle-Induced Health Effects in Children and Infants (IDEFICS) study (20). Body composition was determined by bioelectrical impedance using a Tanita BC-418 MA with 8 contact electrodes (Tanita Europe BV, Hoofddorp, the Netherlands). BP was measured at rest in the dominant arm, and the average of 3 measurements was considered. The BP z score was calculated using
572
Volume 61, Number 5, November 2015
BP tables by sex, age, and height (21). The pulse pressure (systolic BP minus diastolic BP) and mean arterial pressure (MAP, diastolic BP plus pulse pressure/3) were also calculated. Sleep information was evaluated using the BEARS questionnaire (B ¼ Bedtime issues, E ¼ Excessive daytime sleepiness, A ¼ Night awakening, R ¼ Regularity and duration of sleep, S ¼ Snoring) (22). Because sleep needs vary with age, each item involves an age-specific question. Sleep duration was taken from the ‘‘R’’ item of the BEARS questionnaire. Sleep time recommendations were categorized by age according to the National Sleep Foundation (http://www.sleepfoundation.org). Sleep duration less than the reference was considered as the short sleep duration category. According to sleep duration, children were classified into 2 groups: children who sleep the recommended hours and children who sleep fewer hours than recommended or the short sleep group. Dietary data collection was obtained by three 24-hour recall interviews conducted by a nutritionist. Detailed description of type, form, and amount of food consumed and the hour of day and possibility of snacking with frequency, type, and timing of snacks was collected. We evaluated the nutritional content using a validated nutritional software package (DIAL; Alce Ingenierı´a SA, Madrid, Spain, http://www.alceingenieria.net/nutricion.htm) for assessing diets and food calculations. Adherence to the Mediterranean diet was assessed by the Mediterranean Diet Quality Index for Children and Adolescents (KIDMED) questionnaire (23). Information about lifestyle habits was also obtained. The Krece Plus questionnaire was used to assess each child’s level of activity/inactivity (24). For this test, weekly physical activity during leisure time is scored positively, and daily hours spent watching television or playing computer games are scored negatively. The sum of the activities is classified by sex into 3 levels: bad score, regular score, and good score.
Biochemical Analysis Blood samples were collected after 12 hours of fasting for routine clinical analysis. Blood samples were processed and centrifuged at 3000g at 48C for 5 minutes. Plasma and serum aliquots were deep-frozen until needed for further analysis. Routine biochemical parameters (glucose, cholesterol and fractions, triglycerides and uric acid) were measured by automated direct methods (Aeroset System; Abbott Clinical Chemistry, Wiesbaden, Germany) and insulin and homocysteine by an automated electrochemiluminescence immunoassay (Architect c8000; Abbott Clinical Chemistry). Insulin resistance was assessed for the homeostasis model assessment (HOMA) index (insulin [mU/L] glucose [mmol/L]/ 22.5). High-sensitivity C-reactive protein and RBP4 were analyzed by immunonephelometry using a Behring nephelometer 2 (Dade Behring, Hamburg, Germany). Serum GGT was measured by an enzymatic colorimetric test at 378C in a Roche/Hitachi analyzer (Roche Diagnostics, Mannheim, Germany). TSH was measured using a direct immunoradiometric assay (Unicel DxI 800; Beckman Coulter, Barcelona, Spain). We used a multiplex immunoassay with MILLIPLEX MAP Human kits specific for leptin, adiponectin, ghrelin, and tumor necrosis factor a from Merck Millipore (Darmstad, Germany). The plates were read in a cytometer (Luminex 100TM IS analyzer; Luminex Corporation, Austin, TX). The standard curve was calculated using 5-parametric curve fitting, and the results were analyzed using Luminex 100 IS 2.3 software (Luminex Corporation).
Definitions To evaluate the risk of metabolic syndrome in children, the following variables were taken into account as metabolic risk www.jpgn.org
Copyright 2015 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.
JPGN
Volume 61, Number 5, November 2015
factors according to the criteria of the International Diabetes Federation consensus (25): obesity with a z score BMI 2, highdensity lipoprotein cholesterol 40 mg/dL, triglycerides >110 mg/ dL, and hypertension (systolic or diastolic BP >95th percentile for age and sex). We evaluated carbohydrate metabolism by insulin resistance defined as HOMA 3.16 (26). In addition, uric acid, homocysteine, high-sensitivity C-reactive protein, RBP4, GGT, and TSH were considered as emergent new factors of metabolic risk. Leptin, adiponectin, ghrelin, and tumor necrosis factor were included in this group of factors.
Statistical Analysis Sample size was estimated assuming that a difference between groups in z score BMI of 0.7 is clinically relevant at a significance level of a ¼ 5%. A total of 76 patients (38 patients in each group) were deemed necessary for 90% power. If we take into account an estimated loss of 15%, a final sample size of 87 obese children was needed. The Kolmogorov-Smirnov test was used for testing normality. In the case of nonnormal distributions, natural logarithmic transformation was used to improve normality. Means and standard deviations (mean SD) were calculated for continuous data. Metabolic risk factors (independents variables) were examined across sleep duration (dependent variable) using a general linear model controlling for sex, age, BMI, physical activity, and quality of Mediterranean diet as appropriate. Stepwise binary logistic regression analyses were carried out to evaluate the strength of the association between sleep duration and variables with significant differences on general linear model analysis. Sleep duration was analyzed as the response variable (sleep duration according to the recommendation ¼ 0 and sleep duration less than the recommendation ¼ 1). Analyses were performed using Statistical Package for the Social Sciences version 17.0 (SPSS Inc, Chicago, IL). All of the P values were reported with statistical significance set at
