Association of sleep duration with metabolic syndrome and its components in children and adolescents; a propensity score-matched analysis: the CASPIAN-V study

Data were collected by the Persian version of the World Health Organization-Global School Student Health Survey (WHO-GSHS) and its validity and reliability have been assessed previously [27]. Two sets of questionnaires were gathered from students and their parents.

Students’ demographic characteristics (including age, sex and….), healthy behaviors (physical activity, screen time and sleep duration) and dietary habit were asked through students’ questionnaires and some complementary data on family characteristics, namely household size, order of students and socioeconomic variables questioned through parents’ questionnaires.

Physical examination was performed under standard anthropometric techniques by a team of trained health care experts. Weight was measured to the nearest 0.1 kg using a calibrated scale placed on a flat ground and height was measured to the nearest 0.1 cm using a portable stadiometer [28].

Waist circumference (WC) was measured to the nearest 0.1 cm 3 times and the average of three values was used for the analyses. A non-elastic tape was used to measure WC at a point midway between the lower border of the rib cage and the iliac crest at the end of normal expiration [29]. A mercury sphygmomanometer was used to measure blood pressure (BP) on the right arm while participants were in the sitting position. BP was measured 2 times at 5 min intervals, and the average of the two values was used for the analyses [30].

After 12 h overnight fasting, venous blood sample was collected from students. Fasting blood glucose, triglycerides (TGs), and high density lipoprotein cholesterol (HDL-C) were measured enzymatically by Hitachi Auto Analyzer (Tokyo, Japan) [31, 32].

To evaluate the screen time (ST) behaviors, using the WHO-GSHS questionnaire, the students were asked to report the numbers of hours per day that spent watching television (TV) and/or videos, personal computer, or electronic games on week days and at weekends. Then, the total cumulative time spent on these activities was calculated and ST was categorized into two groups; less than 2 h/day was defined as low, and 2 h/day or more was defined as high group [33].

Students were asked to report the average number of hours/day spent for sleeping at weekdays and weekends. According to previous studies, short sleep duration was defined as sleep duration ≤ 8 h/day [12, 17]. Students were also asked to report the onset of sleep occasionally.

Socioeconomic status (SES) included father’s job and education, mother’s job and education, having private car and computer, and type of student’s school (private, public) variables that were summarized in one main component, by principle component analysis (PCA) method. SES was categorized into tertiles. The first tertile was defined as a low SES, second tertile as a moderate and third tertile as a high.

To define healthy and unhealthy eating behaviors, students were asked to determine frequency of consumption of breakfast, fruit, vegetables, milk, sugar sweetened beverages (carbonated drinks and synthetic fruit juices), fast foods (sausages, hamburgers and pizza), sweets (cakes, candies, biscuits and chocolates) and salty snacks (chips, pretzels and salty puff pates). According to PCA method, two factors were loaded in PCA method; in the first factor, consumption of breakfast, fruit, vegetables and milk were loaded which defined as healthy eating behaviors and in the second factor intake of sugar sweetened beverages, fast foods, sweets and salty snacks were loaded which defined as unhealthy eating behaviors. Healthy and unhealthy eating behaviors factor was categorized into tertiles. The first tertile was defined as a low, second tertile as a moderate and third tertile as a high.

Physical activity (PA) was assessed through the validated questionnaire [34]. Using self-administrated questionnaire students were asked about 7-day recalls of sports or activities in and out of school activities that led to heavy sweating or significant increase in heart rate, or games that make students breathe hard. All physical activity items were summarized in a main variable using principal component analysis (PCA) methods and then this main variable was categorized into tertiles. The first tertile was defined as a low, second tertile as a medium and third tertile as a high PA [35].

According to definition of Cook et al. (National Cholesterol Education Program Expert Panel on Detection, Evaluation and Treatment of High Blood Cholesterol in Adult Treatment Panel III) participants were recognized as having MetS if they had three out of the following criteria: fasting TG ≥ 110 mg/dl; HDL-C ≤ 40 mg/dl; WC ≥ 90th percentile for age and sex, according to national reference curves; systolic BP (SBP) and/or diastolic BP (DBP) > 90th percentile for sex, age and height, from national reference cut-off points; and FBG ≥ 100 mg/dl [36, 37].

Approval for the study was obtained from the relevant local and the national committees (The Research and Ethics council of Tehran and Isfahan University of Medical Sciences, Project number: 194049).

Using STATA package ver. 11.0 (Stata Statistical Software: Release 11. College Station, TX: Stata Corp LP. Package). Results provide as mean and standard deviation (SD) for continuous variables, and as number (percentage) for categorical variables.

Comparison of percentages of the categorized variables was made using the Pearson Chi square test. PS was calculated based on conditional logistic model with potential confounding variables (age, sex, living area, SES, PA, ST, healthy and unhealthy eating behaviors, family history of chronic diseases and parental BMI). Two groups (with MetS and without MetS) were matched based on 1:1 matching method without replacement by the score. Then standardized skewness (SS) percentages before and after synchronization were calculated and then SS mean was calculated for all variables. Due to PS matching, conditional logistic regression was used to evaluate the association of the short sleep duration (as categorical variable) and the onset of sleep (as continuous variable) with MetS and its components. Results of conditional logistic regression was reported as odds ratios (OR) and 95% confidence intervals (CI). All statistical analysis were estimated using survey data analysis methods. P < 0.05 was considered as statistically significant.