The relationship between sleep quality and breakfast, mid-morning snack, and dinner and physical activity habits among adolescents: a cross-sectional study in Yazd, Iran

Healthy lifestyles are important during adolescence due to their impact on health-related problems and life expectancy during adulthood. Despite the emphasis on the importance of healthy habits, the high prevalence of breakfast skipping, physical inactivity, and insufficient sleep is reported among adolescents. This study was aimed to investigate the relationship between sleep quality (SQ) and nutritional habits among adolescents of Yazd. This cross-sectional study was conducted among 569 students aged 12–16 years of six high schools in Yazd, Iran, by a multistage random cluster sampling method. Anthropometric data were measured by precise instruments. Nutritional and PA habits were collected by asking some questions. Pittsburgh Sleep Quality Index questionnaire was used to assess SQ. Data were analyzed by SPSS software (version 22). The breakfast skipping and semi-skipping (57.2%), overweight or obesity (37.8%), and insufficient sleep duration (38.9%) were common. Poor SQ was reported in 63.7% of the students, but a significant relationship was not illustrated between SQ with breakfast frequency, weight, breakfast time, and midmorning snack intake. A significant relationship was found between poor SQ with a later bedtime (p < 0.0001), mild PA (p = 0.016), late dinner (p < 0.004), and famine (p = 0.02). This study highlighted the associations between SQ with dinnertime, bedtime, and PA. Also, a significant association was not reported between the SQ with breakfast frequency, breakfast time, snack consumption, and weight. This study suggested paying more attention to healthy sleep habits during adolescence, apart from guidance for diet and PA during adolescence.


Introduction
Sleep is a physiological process that is necessary for health. The relationship was reported between sleep habits with obesity and unhealthy behaviors (inappropriate calorie intake, higher total fat intake, lower fruit intake, and lower quality diets, breakfast skipping, smaller size breakfasts, and energy-density or highly palatable snacks at night) [1]. Inappropriate sleep pattern is common in adolescents, unfortunately [2]. Previous studies showed that shorter sleep duration is related to overweight and obesity in children and adults [3]. Increased weight was reported among children [3] and adolescents [4] with short sleep duration and later bedtime.
Breakfast skipping and late dinner are the contributory factors in an unhealthy lifestyle [5]. Skipping breakfast may lead to such things factors, such as a less healthy nutrition diets [6] and sleep patterns [7] among adolescents. Breakfast 1 3 consumption is recommended for people, without a late dinner, to give adequate energy for mental and physical activities in the morning. But breakfast intake with high-energy density is not recommended for people with habitual late dinners [8].
Physical inactivity is associated with non-communicable diseases [9], and inappropriate sleep habits [10]. Insufficient sleep and exercise have become public health problems. They have a bidirectional relationship [11]. Some studies have shown inappropriate sleep habits [such as higher' sleep latency, poor' sleep quality (SQ), shorter' sleep duration, and sleep medication intake] are associated with inadequate physical activity (PA) [10].
According to the importance of health lifestyles and considering that there are not enough researches among adolescents, this study was aimed to investigate the relationship between SQ with breakfast and midmorning snack, and dinner and PA habits among adolescents of Yazd, Iran.

Study design and participants
The sampling was done by the multistage random cluster sampling method among first-degree high-school students aged 12-16 years old of Yazd. Also, the following formula was used to calculate the minimum sample size [12]: The minimum sample size was calculated 381 by considering Confidence Level (α) and Margin of Error (d) as 0.05 and "P" as 45%, according to the mean of the sleep quality in the previous articles on Iranian students (Supplementary data. S.1). Finally, 569 students from six high schools were selected for this research. It is noteworthy that the data were collected during the semester. The design template of the present study was shown in Figure  The students entered the study if they had the following conditions: (1) Completing written informed consent by the students and their parents; (2) No travel to Yazd within the previous; (3) Lack of a disease, including hormonal impairment, such as thyroid, insulin, sex, etc., cardiovascular disorders, digestive problems, pulmonary disease, kidney disease, Malignancy, (4) No hospitalization within the last 6 months for any reason, e.g., fractures, injuries and etc., (5) Lack of an emotional event for at least the last year (the parental death and divorce, the death of family members, living with one of the parents or relatives for any reason), (6) Not using drugs, such as narcotics and psychotropic, (7) No transferring of the student for at least the semester.

Measurements
Four research assistants were trained to collect information, anthropometric measurements, and interviewing to reduce potential bias. The questionnaires included demographic data (gender, age, household size, and grade), lifestyle habits (nutrition and PA) and Pittsburgh Sleep Quality Index (PSQI). The questionnaires are acceptable on Supplementary data.S.2. The questionnaires were filled during school hours in the classroom in the presence of a teacher of the high school and a research assistant by the students. There was no time restriction to fill out the questionnaires. After completing the demographic questionnaire by the students, it was checked based on the school records.
Anthropometric measurements [weight, height, Body mass index (BMI)] were done in the morning before completing the questionnaires in a separate room. The students were lightly dressed and without shoes. A pre-calibrated electronic scale (Seca, Germany) and a stadiometer were used to measure the students' body weight with an accuracy level of 0.1 kg and their heights were measured to the nearest 0.1 cm by flexible tape. BMI (kg/m 2 ) was calculated by this equation: weight (kg)/height 2 (m 2 ). BMI for age (z-score) was computed using 2007 growth standard reference of World Health Organization (WHO). BMI z-score was categorized in this way: (a) thin student (BMI ≤ − 2 Standard deviation (SD), (b) normal student (BMI between 1SD to − 2SD), (c) overweight and obesity (BMI ≥ 1SD) [13]. BMI z-score tables are available on the WHO website.
The following questions about the students' breakfast, mid-morning snack, and dinner habits were asked by the interview for assessing nutritional information: (1) "how many times do you eat breakfast per week?" Response categories: less than 50% days per week (0-3 times) and more than 50% days per week (4-7 times) [14]; Two following questions were asked to estimate PA: 1. How many days did you do physical activities for at least 30 min at a time per day during the last 7 days? Response options were from 0 to 7 days.

3
2. How many hours do you spend in the gym class regularly in school per week? Response categories: (a) less than 3 h; (b) more than 3 h.
The severity of PA categorized into three groups: vigorous (more than 4 h a week), moderate (two to four times a week) and mild (less than two times per week) [16].
SQ was evaluated for a 1-month time interval by the previously validated PSQI questionnaire. The Persian version of the PSQI for adolescents was shown adequate reliability (α = 0.77) and validity [17]. The overall PSQI range is from 0 (good) to 21 (poor). SQ was divided into two groups according to the Pittsburgh questionnaire: Less than or more than five [18]. Sleep latency and duration were categorized based on PSQI guidelines. Bedtime was classified into: (1) Before 22:00, (2) 22:00-24:00, (3) After 24:00. Wake-up time was not categorized because data were collected during the end semester. Also, all of the classes were begun at about 7.45 or 8 a.m.

Data analysis
The analysis of quantifiable variables was conducted by the means and standard deviations (SD). The interquartile ranges (IQR) (25th-75th) were used for breakfast time and household size. The categorical variables were analyzed by frequency [number (%)]. The Kolmogorov-Smirnoff test was used to test the normality distribution of data to determine the parametric or non-parametric test. One-way ANOVA test with Tukey HSD Post hoc analysis was conducted to compare the difference between pairwise. SPSS statistical software package, version 16.0 (SPSS, Inc, Chicago, Illinois, USA), was applied for statistical analyses. p < 0.05 was considered statistically significant using 2-tailed tests.
The differences between the SQ score and the categorized variables were presented in Table 2. The SQ score shown no significant difference between the students with different BMI (p = 0.1). Only, 5.6% of the students were thin, so the comparison for SQ was conducted between the normal and overweight or obesity groups.
The sleep score was significantly better in the boys (p = 0.02). No significant association was found between the SQ score with breakfast frequency (p = 0.4) and breakfast

3
time (p = 0.6); and also despite the higher SQ scores (lower SQ) in morning snack users, the significant difference was not observed (p = 0.3). A significant decrease for the SQ score was illustrated in early dinner (p = 0.004). The better SQ was observed among the students who ate dinner before 22:00 than who ate after 22:00. A significant relationship was also found between good SQ and earlier bedtime (p < 0.0001). The significant relation was found between PA and SQ score, (p = 0.016). SQ was found better in the students with moderate than mild PA (P = 0.012).

Discussion
A surprising finding of this study was the significant relationship between poor SQ with late dinner, late bedtime and low PA. Unhealthy lifestyles (irregularly breakfast intake, overweight or obesity, poor SQ, insufficient sleep, late dinner and late bedtime) were common among adolescents of Yazd. A significant relationship was not found between SQ with breakfast frequency, breakfast time, mid-morning snack intake, and overweight or obesity.
It seems to collect data in the semester that led to the insignificant difference for breakfast time and waking-up among the students. In all the schools, the first break was before 9:30, and holidays were one or 2 days per week, commonly. Also, the start and end times of schools were similar.
In opposition to the present result, inadequate sleeping time can be associated with obesity and overweight [4]. There are few data about the relationship between SQ and breakfast habits among adolescents. The prevalence of high-risk behaviors, such as physical inactivity and skipping breakfast, was low among Japanese high-school students with normal sleepers [19]. Arakawa et al. emphasized that late bedtime is related to poor SQ (less sleep, incremented sleepiness, falling asleep during the days) and unhealthy dietary habits in the junior high-school students [5]. A positive relationship was also reported between normal bedtime or sleep timing with some healthy nutrition habits (regular breakfast consumption, regular breakfast frequency, no snacking, and no sweetened beverage consumption) in French primary schools [7]. It is noteworthy that the content and amount of foods intake also play an important role on sleep quality and quantity [20][21][22][23] and weight [24].
Most likely, the bilateral association between sleep and nutrition habits was partially explained by multifactorial mechanisms [25]. The appetite-related hormones (ghrelin and leptin) may lead to an increase in BMI. Hormonal alterations, towards more food intake due to the increased appetite, can occur in inappropriate sleep habits [25]. Also, nutrition habits include diet contents (high carbohydrate diets, foods with high phytonutrients, tryptophan (as a precursor for serotonin), and melatonin) [26] and unhealthy eating behaviors (late dinner, extended hours for intake, eating fewer three main meals in a day with higher energy-dense snacks at night) can impact on sleep habits by hedonic [1] and hormonal pathways [25]. Some researchers believed hedonic mechanisms have a more impact rather than hormonal mechanisms [27]. The results of a meta-analysis in 2019 showed that there was no strong evidence that sleep deprivation had a significant effect on serum leptin or ghrelin levels or energy expenditure [28]. The food intake time, particularly in main meals plays an important role in weight status [29]. Considering the disagreement in the explaining mechanisms of this relationship (sleep and nutrition habits), there may be additional unknown mechanisms that mediate this relationship. Our data illustrated poor SQ in the girl students, in agreement with the previous studies [30,31]. Biological sex differences, higher intake of caffeine after dinner, overweight and obesity [31], and more susceptible to stress [32] may lead to poor SQ in females.
Physical activity can improve sleep problems regardless of the mode and intensity of activity, especially in people with improper sleep habits [33] through endocrine (Growth hormone (GH), norepinephrine, and cortisol secretion), autonomic nervous system (ANS), and somatic functions. It is also better to consider some factors: the exercise type, intensity, timing, and differences between acute and chronic exercise, simultaneously [34].
In these ages, the eating habits of the most adolescents were affected by their parent's lifestyles [35]. It will be necessary to compare to SQ with the nutritional habits in the vacation season and evaluating the nutritional habits of the students' parents in the next studies to resolve these problems.
The limitations of this cross-sectional study were: (1) interpreting with caution (the interpretation of the reported relationships as hypothetical causal correlation); (2) Selfreport without any corroboratory (parental) or objective measures that could affect the results (recall bias and social desirability bias) by a possible under or over-reporting. In fact, we could not verify their response to sleep (PSQI) questions related to the "majority of days over the last month" were in fact weekdays, as opposed to weekend days; (3) other factors (evening snack, the quality, and quantity of the meals, stress, social networks, socio-economic information) [8,14,15,35] should be considerate to assess SQ; (4) the lack of controlling confounder factors; (5) the data collection was only during the semester.
But the strengths of our study were: (1) sample size was nearly large; (2) data were collected from both gender; (3) the evaluation of different aspects of sleep among adolescence by the PSQI in this age group.

Conclusion
This study highlighted associations between SQ with dinner time and bedtime. On the other hand, there was no significant association between SQ with breakfast frequency, breakfast time, mid-morning snack consumption, and BMI. Due to the above-mentioned limitations, the results of the present research should be interpreted cautiously, and more longitudinal and epidemiological studies are needed to evaluate different factors related to lifestyle (sleep, psychophysiology, and nutrition parameters) using accurate and objective tools [2,22,23] and also elucidate whether these modifications validate results (causal relationship).
This study suggests paying more attention to healthy lifestyle habits during adolescence with a comprehensive program, such as (1) employing skilled School Health Practitioners at all schools for the face-to-face or group educational intervention to adolescents or their parents, (2) accessing of the students and their families to free online education; (3) improving the school cafeteria; (4) considering adequate time for "breakfast breaks"; (5) increasing inschool activity breaks; (6) giving mandatory sleep training courses for the students.