Sleep quality and metabolic syndrome in overweight or obese children and adolescents

Objective To assess sleep quality and its association with metabolic syndrome and its components. Methods This cross-sectional study was conducted from June 2011 to March 2012 at the Childhood Obesity Center, Campina Grande, Paraíba, Brazil, with 135 overweight or obese children and adolescents. Sleep quality was assessed by the Pittsburgh Sleep Quality Index. Metabolic syndrome diagnosis was based on abdominal circumference, blood pressure, glycemia, high density lipoprotein-cholesterol, and triglycerides. The data were treated by the software Statistical Package for the Social Sciences version 22.0 at a significance level of 5%. 378 | NC GONZAGA et al. Rev. Nutr., Campinas, 29(3):377-389, maio/jun., 2016 Revista de Nutrição http://dx.doi.org/10.1590/1678-98652016000300008 Results The prevalence of poor sleep quality or sleep disorder according to the Pittsburgh Sleep Quality Index was 40.7%, and females had higher mean global Pittsburgh Sleep Quality Index score. Metabolic syndrome prevalence was 63.0%. Females also had higher daytime dysfunction. Poor sleep quality was associated with high diastolic blood pressure (OR=2.6; p=0.015) and waist circumference (OR=3.17; p=0.024) after adjusting for sex and age. Conclusion Girls had higher global Pittsburgh Sleep Quality Index score, which was associated with daytime dysfunction. Poor sleep quality was a predictor of high diastolic blood pressure in the study sample.


Objetivo
Avaliar a qualidade do sono e sua associação com a síndrome metabólica e seus componentes.

I N T R O D U C T I O N
Sleep is a vital necessity and its periodicity, quality, and constancy characterize very important variables in the development of children and adolescents 1 .Changes in sleep pattern and quality can affect people at a very young age and their consequences range from poor school performance to compromised physical and mental health 2 .
In the last years, the prevalence of voluntary sleep restriction, insomnia, and even poor sleep quality has increased concomitantly with the growing prevalence of obesity in children and adolescents 3 .
The reason for the relationship between sleep and obesity has not been fully elucidated.Studies suggest that partial or chronic sleep deprivation causes energy unbalance by changing the levels of many hormones, including leptin and grehlin 4,5 , whose function relates to energy balance and control of body weight 6,7 and insulin.
Regarding insulin, short sleep duration may be associated with insulin resistance 6  is related to the metabolic syndrome in overweight and obese children and adolescents 8 .
Hence, sleep has an important metabolic role as it may promote endocrine and metabolic changes that increase hunger and appetite, and consequently, the risk of overweight and obesity 8 .
Adolescents with short sleep duration (<7 hours/night) had lower intake of fruits and vegetables and higher intake of fast foods 9 .Along with inappropriate eating habits, it is possible that qualitative and quantitative sleep changes reduce the level of physical activity in children and adolescents 1 .
Most studies on the association between sleep and obesity in children assess sleep duration [9][10][11][12][13][14][15] , and few assess sleep quality 14,16,17 .However, assessment of sleep duration alone is not enough, qualitative aspects must also be assessed 18 .This is because when individuals have difficulty falling asleep or when they wake up frequently, effective sleep duration decreases, and the body does not spend enough time in deep sleep stages, even if total sleep duration is adequate 19 .
Questionnaires have been used for diagnosing sleep disorders, and although they are indirect and subjective assessment instruments, their administration requires little time, and they are free and easy to use in epidemiological studies 16,17,20 .
The Pittsburg Sleep Quality Index (PSQI) is an instrument commonly used for assessing sleep quality 17,21 .The version validated for the Brazilian population was developed in a study that compared PSQI and polysomnography, the gold standard sleep quality assessment method in adults.The results have shown that the Brazilian PSQI version is a valid and reliable instrument for assessing sleep quality in the month prior to the interview 22 .Some studies have reported that sleep quality in adults assessed by the PSQI is associated not only with excess weight, but also with metabolic syndrome 18,21 .Metabolic syndrome is an aggregation of cardiometabolic risk factors represented by hypertension, abdominal obesity, hypertriglyceridemia, low High Density Lipoproteincholesterol (HDL-c), and glucose intolerance 23 .Therefore, the detection of this condition at an early age is important for the implementation of prevention strategies.
In adolescents short sleep duration was associated with overweight but not with metabolic syndrome.Short sleep duration was associated with high blood pressure, and long sleep duration, with hypertriglyceridemia 12 .
Nevertheless, until now, no study has assessed the association between sleep quality and metabolic syndrome in children and adolescents.Moreover, studies that assess sleep quality in this population are scarce.
Hence, in order to contribute to this qualitative sleep aspect, and given the repercussion of sleep disorders on overweight and obese individuals, this study aimed to assess sleep quality and its association with the metabolic syndrome and its components in overweight or obese children and adolescents.

M E T H O D S
This cross-sectional study was conducted during the school year from June 2011 to March 2012 at the Childhood Obesity Center, located in the Instituto de Saúde Elpídeo de Almeida, Campina Grande (PB), Brazil.Childhood Obesity Center is a reference public service for providing multidisciplinary care to overweight and obese children and adolescents.Currently, 390 children and adolescents are registered.
This study is part of a larger project called "Prevalence of cardiometabolic risk factors in overweight or obese children and adolescents", approved by the Research Ethics Committee of the Universidade Estadual da Paraíba (UEPB) under Protocol nº 0040.0.133.000-08.The study followed all the ethical precepts for human research.
The convenience sample consisted of all overweight and obese children and adolescents aged 5 to 18 years who attended the routine medical visit on Fridays during the school year, since sleep habits change during vacations 24 .Students with adenotonsillar hypertrophy; cardiorespiratory, neuromuscular, neoplastic, and/or advanced hepatic disease; and taking drugs that affected sleep (antidepressants, benzodiazepines, bronchodilators, corticosteroids, and illicit drug or alcohol abuse) at data collection were excluded.
During this period 190 individuals were treated.Seventeen of these were excluded, eleven because they were outside the study age range, two were normal weight, two had asthma, and two had adenoid hypertrophy.Thirty-three individuals did not show up for blood collection and five did not have one of the results for metabolic syndrome components, so the final sample consisted of 135 individuals.
The anthropometric variables (weight, height, and abdominal circumference) were measured twice and averaged, as recommended by the World Health Organization 25 .
Nutritional status classification was based on Body Mass Index (BMI), as recommended by the Centers of Disease Control and Prevention: overweight (85 th percentile ≤BMI <95 th percentile), obesity (95 th percentile ≤BMI <97 th percentile), and morbid obesity (97 th percentile ≤BMI) 26 .Abdominal circumference was considered high when ≥90 th percentile, as recommended by the International Diabetes Federation 27 , with a maximum limit of 88 cm for girls and 102 cm for boys, as recommended by the National Cholesterol Education Program Adult Treatment Panel III (NCEP/ATPIII) 28 .
Blood pressure was measured three times two minutes apart each, as recommended by the V Brazilian Guidelines for High Blood Pressure 29 using a mercury Tycos ® sphygmomanometer made in Germany by WelchAllyn and appropriatesize cuff.The mean of the two last measurements was considered the systolic and diastolic blood pressure.
Blood was collected after a 12-hour fast by the Laboratory of Clinical Analyses of the UEPB.HDL-c, triglycerides, and glycemia were assessed by enzyme colorimetry by an automatic device (Modelo BioSystems 310, Curitiba, Paraná, Brazil).
Metabolic syndrome diagnosis was based on the criteria provided by NCEP/ATPIII 28 adapted for the age group, which requires the presence of at least three of these criteria: abdominal circumference ≥90 th percentile for sex, age, and race; triglycerides ≥100 mg/dL; HDL-c <45 mg/dL; fasting glycemia ≥100 mg/dL; and systolic and/or diastolic blood pressure ≥90 th percentile for sex, height, and age.
The Pittsburg Sleep Quality Index 30  The global score is given by adding the seven components.Each component receives a score ranging from 0 to 3 points with the same weight, and 3 represents the negative extreme of a Likert-type scale.Hence, the total score may range from 0 to 21 points.Sleep quality is classified as follows: 0-5 = good sleep quality; 6-10 = poor sleep quality; >10 = sleep disorder.All participants with a PSQI score higher than 5 (PSQI >5) were diagnosed with poor sleep quality as the score indicates great difficulty in at least two components, or moderate difficulty in more than three components 30 .
The score of each component was also assessed and compared with metabolic syndrome and its components.Subjective sleep quality scored 0-3 points (0 -very good sleep quality; 1 -good sleep quality; 2 -bad sleep quality; and 3 -very bad sleep quality).Sleep latency scored 0-3 points (time in minutes required to fall asleep http://dx.doi.org/10.1590/1678-98652016000300008every night and frequency of not falling asleep within 30 minutes).Sleep duration scored 0-3 points (0 = sleep duration >7 hours; 1=6 -≤7 hours; 2=5 -<6 hours; 3=<5hours).Habitual sleep efficiency scored 0-3 points (classified according to the sum of the answers related to sleeping problems caused by nine reasons and their frequency, such as feeling hot or cold: 0=0; 1=9-1; 2=10-18; 3=19-27).The use of sleeping pills scored 0-3 points (0 = did not use in the last month; 1 = used less than once a week; 2 = used once or twice a week; 3 = used three or more times a week).Daytime dysfunction scored 0-3 points (classified according to the sum of the answers related to the frequency of having problems to stay awake while driving, during meals, or during social activities, and the frequency of having problems to remain enthusiastic enough to carry out activities of daily living: 0-0; 1-2=1; 3-4=2; 5-6=3).
The population was described according to their anthropometric and clinical variables, expressed as absolute and relative frequencies, means, and standard deviations.Group means were compared by the Student's t test or Analysis of Variance (Anova) for three groups and the proportions were compared by the Chi-square test or Fisher's exact test.Univariate logistic regression assessed the relationship between metabolic syndrome and its components and sleep disorder diagnosed based on PSQI score, which was later adjusted for sex and age.The Odds Ratio (OR) was calculated based on logistic regression.The logistic regression was performed by transforming the PSQI variable into dummy variables, the reference group being good sleep quality.
The statistical analyses were performed by the software Statistical Package for the Social Sciences (SPSS Inc., Chicago, Illinois, United States), version 22.0, at a significance level of 5%.

R E S U L T S
Table 1 shows the sample characteristics.Of the 135 study children and adolescents, 56.3% were females, 64.4% were classified as morbidly obese, and 76.3% were adolescents.The mean age was 12.7±3.4years (5.0-20.0).Sex was associated with nutritional status, males were associated with morbid obesity when the groups morbid obesity and obesity were compared (p=0.010).
The prevalence of poor sleep quality/sleep disorder according to the PSQI score was 40.7%, and the prevalence was higher in females (44.7 versus 35.6%) (Table 1).
More than half (55.6%) the sample classified their sleep in the previous month as bad or very bad.Metabolic syndrome prevalence was 63.0%.Low HDL-c and blood glucose changes were the most and least frequent metabolic syndrome components, respectively.High abdominal circumference was associated with being male (Table 1).
The mean Pittsburgh Sleep Quality Index score of adolescents was higher than that of children.While adolescents had a mean score of 5.21±2.60(4.70-5.72),children had a mean score of 4.44±3.29 (3.25-5.62)(p=0.170).
Pittsburgh Sleep Quality Index component with the highest mean was subjective sleep quality.Females had higher mean scores in six PSQI components, except efficiency, and the mean daytime dysfunction score was significantly higher in girls (Table 2).
The mean Pittsburgh Sleep Quality Index and PSQI component scores were not different between those with and without metabolic syndrome, regardless of sex.Although no association was found in females with metabolic syndrome, most of their PSQI items had higher mean scores than those of males with metabolic syndrome (Table 3).
Pittsburgh Sleep Quality Index and/or its components were not associated with metabolic syndrome, according to logistic regression (Table 4).that individuals with poor sleep quality were up to 2.5 times more likely to have high diastolic blood pressure (p=0.018;95% Confidence Inteval-95%CI=1.17-5.43),and the risk increased to 2.62 (p=0.01;95%CI=1.20-5.69)after adjusting for sex and age (OR=2.62)(Table 5).

Logistic regression analysis between PSQI and metabolic syndrome components showed
After adjusting the components for sex and age, high abdominal circumference was associated with poor sleep quality (OR=3.17;p=0.02; 95%CI=1.17-8.60)(Table 5).The possibility of sleep disorder was not associated with metabolic syndrome components probably because of the small sample size.

D I S C U S S I O N
It is critical to identify sleep disorders during childhood because of the importance of introducing strategies to promote sleep quality and prevent metabolic complications, such as Note: M: Média; SD: Standard Deviation.excess weight 10,32 .Sleep disorders are often underdiagnosed in this age group as they depend on the perception of caregivers given that children do not always verbalize symptoms related to this condition 33 .
Sleep quality in this study was measured by the PSQI index and its components (subjective quality, latency, duration, habitual efficiency, sleep disorders, use of sleeping pills, and daytime sleepiness) in a group of children and adolescents with excess weight, and the prevalence of poor sleep quality in the sample was high.
A study with Brazilian university students with a mean age of 21.5 years found that 95.3% had poor sleep quality (PSQI >5), and the mean PSQI score was 9.4 points 34 .Some factors that may influence sleep quality in children and adolescents are long periods watching television 35 and changes in the daily routine and habit during vacations.This last item leads to changes in sleeping pattern, such as going to bed later and consequently, getting less rapid eye movement sleep.Both factors were associated with a higher risk for obesity in children and adolescents 36 .
Poor sleep quality was more prevalent among the study adolescents.This condition may affect wake quality to a variable degree, especially in this age group, since they go to bed late in  their biological configuration and tend to wake up late, which is not possible during the school year, resulting in excessive daytime sleepiness 37 .
A study of Brazilian adolescents using the Sleep Behavior Questionnaire found that the obese group had worse sleep quality than the normal weight group 16 .
Females had higher mean global PSQI and daytime dysfunction scores than males.This association with being female corroborates a study done with 1,481 adults that also assessed the relationship between sleep quality and metabolic syndrome by sex 18 .
Although sleep quality was not associated with metabolic syndrome in the present study, sleep quality was associated with diastolic blood pressure and abdominal circumference.
Another two studies did not find an association between sleep duration and metabolic syndrome 12,38 .In male and female adults, the mean global PSQI score and the mean score of its components, especially sleep latency and sleep disorders, have been associated with metabolic syndrome 18 .
Adolescents with sleep-related breathing disorders diagnosed by polysomnography were seven times more likely to have metabolic syndrome than those without sleep-related breathing disorders (OR=7.74;95%CI=3.10-19.35),and this association was not explained by sex, race, or socioeconomic condition 39 .
High diastolic blood pressure was found in children with obstructive sleep apnea 40 .Adolescents with low sleep efficiency determined by actigraphy were 3.5 times more likely to have high blood pressure 41 , a higher risk than the one found by the present study.This fact may be due to the different methods used for determining sleep quality.
A study that assessed sleep quality with full-night polysomnography found that the absence of deeper sleep stages, such as rapid eye movement sleep and slow-wave sleep, was associated with high morning blood pressure in obese adolescents regardless of BMI 42 .
A study with obese Korean adolescents found that short sleep duration was associated not only with high blood pressure but also with high abdominal circumference 12 .
In the present study, high abdominal circumference was associated with poor sleep quality after adjusting for age and sex, and may be used as a marker of sleep quality 12 .
In this sense, assessment of obese adolescents must include investigation of blood pressure, abdominal circumference, and the signs and symptoms associated with poor sleep quality, including snoring, fatigue, and daytime sleepiness.Thus, poor sleep quality must be considered in the follow-up of hypertensive obese youth 42 .
The present study found a high prevalence of poor sleep quality and the need of including this assessment in primary health care, especially in obese individuals.Health professionals who work with children and adolescents must know sleep physiology and its physiological maturation process to prevent or treat pathological behaviors.Therefore, questions regarding sleep quality and possible damaging factors should be included in the patient's assessment, in addition to advice on sleep hygiene, which should be part of children's health care 32 .
Sleep hygiene practices include changes in the sleep environment, and parents' and children's practices and routines that favor good sleep quality of adequate duration, in addition to the practice of activities that favor sleepiness in order to promote sleep 43 .
This study has limitations, such as the use of a convenience sample, the absence of a control group, the cross-sectional design, which does not allow the identification of the relationship of causality, and the use of a questionnaire to assess sleep quality, which although equivalent, is less accurate than other analytical methods such as polysomnography and actigraphy.The relative scarcity of pediatric studies that analyze the relationship between sleep quality and metabolic syndrome hindered comparison of the results.
The study is important because the assessment of the sleep quality of overweight and obese children and adolescents revealed a high prevalence of poor sleep quality and its association with high diastolic blood pressure in adolescents.

Girls had higher global Pittsburgh Sleep
Quality Index score and an association between said score and daytime dysfunction.Sleep quality was not associated with metabolic syndrome in the study overweight or obese children and adolescents.Adolescents with poor sleep quality were significantly more likely to have high diastolic blood pressure.
The realization of other studies on this subject in other Brazilian municipalities with a larger sample, control group, and longitudinal design may contribute to the understanding of the influence of sleep quality on metabolic changes, risk factors for cardiovascular disease and indispensable for the institution of child health policies in the national landscape.
C O N T R I B U T O R S NC GONZAGA and ASS SENA participated in the project design, analysis and interpretation of data and writing of the manuscript.AC COURA and FG DANTAS contributed to the writing and critical review of the manuscript.RC OLIVEIRA participated in the project design and final revision of the manuscript.CCM MEDEIROS participated in the project design, data analysis and final revision of the manuscript.
assessed sleep quality during the previous month.This Portuguese version of this instrument has been validated 31 .PSQI contains seven components: subjective sleep quality (subjective sensation of satisfaction with daily sleep), sleep latency (long sleep start time), sleep duration, habitual sleep efficiency (proportion between hours slept and total hours in bed), sleep disorders (sleep interruption), use of sleeping pills, and daytime dysfunction.

Table 2 .
Mean values and standard deviations of the Pittsburgh Sleep Quality Index (PSQI) domains by sex (N=135).

Table 5 .
Logistic regression between Pittsburgh Sleep Quality Index and components of metabolic syndrome adjusted for age and sex (N=135).Campina Grande (PB), Brazil, 2011-2012.