Study design and population
The current cross-sectional analysis was conducted on the data from recruitment phase of Yazd Health Study (YaHS), a prospective cohort study on 9962 adults’ aged 20 to 70 residents in Yazd, Iran, between 2014 and 2016. Assessment of dietary intakes of these individuals was done in another study named Taghzieh Mardom Yazd (TAMYZ), which we used its data in the current analysis too. Details about YaHS-TAMYZ studies have already been published (25). All participants in the YaHS-TAMYZ study signed informed consent. In the current analysis, during the screening process, participants were excluded if they had missing data on sleep quantity or had missed responding more than 70 items of dietary intake data (n = 850) or unexplained energy intake (< 800 kcal/d or > 6000 kcal/d) (n = 864), reported less than 3 hours or more than 12 hours of sleep per night (n = 250), were pregnant (n = 104) and reported to have major depression (n = 91). Those who had a history of chronic conditions at the time of the baseline survey, such as diabetes, cancer, or cardiovascular disease, were also eliminated (n = 1908).
Dietary intake assessment: Participants in the YaHS study had their dietary intakes assessed separately using a semi-quantitative food frequency questionnaire (FFQ) consisting of 178 food items, whose validity and reliability had been measured and confirmed in a prior study. (26). The FFQ used in this study was a 178 item questionnaire that contained 168 food items commonly consumed in Iran and 10 questions related to the consumption of traditional foods in Yazd. Participants were asked to respond to a 10-choice frequency response section for each food item, ranging from "never or less than once a month" to "10 or more times per day" for each food item. For each food item, five different portion sizes were asked based on Iranian’s standard serving size. Participants had to answer two questions in this questionnaire: the frequency of food consumption and the amount of food consumed at each time of consumption. Using household measures, the quantity and frequency of daily intake of all food items were converted to grams per day (27). United States Department of Agriculture's (USDA) food composition database was used to calculate each participant's energy and nutrient intakes. (28).
Calculation of dietary insulin index and load
The dietary insulin index was computed using previously published estimations by Brand-Miller et al., which are based on the insulin index of dietary items (29, 30). The food insulin index (FII) is the difference between the area under the curve after ingestion of a 1000-kJ portion of the test food and the area under the curve after ingestion of a 1000-kJ portion of the reference food during a 2-hour period. The FII of similar food items based on the similarity between their energy, carbohydrate, fiber, fat, and protein content was applied in the current investigation for food items that were not available in the food list released by Brand-Miller et al. The FII of Qottab, a traditional sweet popular in Yazd and primarily made of sugar and flour, was, for example, regarded to be the same as sugar. First, the insulin load of each food was determined using the following formula to determine dietary insulin load (DIL)
Insulin load of a given food = Insulin index of that food × energy content of that food (kcal/d)
DIL was calculated for each subject by adding up the insulin load of consumed foods. The dietary insulin index (DII) was then computed for each participant by dividing DIL by total calorie intake.
Assessment of sleep quantity and quality: Data on regular sleep quantity was collected by an interview utilizing a nocturnal sleep questionnaire in the YaHS study. Sleep duration was divided into three categories: fewer than 5 hours (short sleep duration), 5 to 8 hours (normal sleep duration), and more than 8 hours (long sleep duration) (31, 32). The sleep quality of the study participants was obtained through an abbreviated form of the Pittsburgh questionnaire. The validity and reliability of the complete Pittsburgh questionnaire for the Iranian population has been confirmed (33). The sleep quality score in the complete Pittsburgh questionnaire is between 0 and 21. In the present study, due to the use of the abbreviated form of the Pittsburgh questionnaire, the average sleep quality score was between 0 and 11, which was calculated based on the following questions: 1. How many minutes does it take to fall asleep from the time you go to bed? [Less than 15 minutes (0 score), 16–30 minutes (1 score), 31–60 minutes (2 scores), more than 1 hour (3 scores)]. 2. How many times in the last 30 days have you been unable to fall asleep within half an hour?[None (0 points), less than once per week (1 point), once or twice per week (2 points), three or more times per week (3 points)]. The sum of these two questions, which is a number between 0 and 6, is known as the delay in falling asleep, and the delay in failing asleep was categorized as follows: score 0 (got 0), score 1 to 2 (got 1), score 3 to 4 (got 2), score 5 to 6 (got 3). 3. How many hours a night do you sleep? [More than 7 hours (0 score), 6 to 7 hours (1 score), 5 to 6 hours (2 scores) and less than 5 hours (3 scores)]. 4. How many times in the past month have you taken sleeping pills or sedatives to fall asleep? [None (0 score), once a week (1 score), twice a week (2 scores), three or more times a week (3 scores)]. 5. Did you wake up in the middle of the night or early in the morning? [None (0 score), less than once a week (1 score), once or twice a week (2 scores), three or more times a week (3 scores)] 6. How many times did you have nightmares while sleeping? [None (0 score), less than once a week (1 score), once or twice a week (2 scores), three or more times a week (3 scores)] 7. How many times did you wake up in the middle of the night to go to the bathroom? [None (0 score), less than once a week (1 score), once or twice a week (2 scores), three or more times a week (3 scores)]. Answers to questions 2 and 4 to 7, is known as sleep disorder, were added and converted as follows: score zero (0), score 1 to 9 (score 1), for score 10 to 12 (score 2). Finally, the sleep quality score in this study was generated by adding the scores of the above-mentioned questions and was ranged from 0 to 11 [delay in falling asleep (0–3), night sleep (0–3), sleep medication use (0–3), sleep disorder (0–2)].
Anthropometric measurements: Anthropometric markers such as height and weight were objectively measured in the YaHS-TAMYS study with minimum outfits and no shoes. Body weight was measured to the nearest 0.1 kg using a digital scale and body analyzer (Omron BF511, Omron Inc. Nagoya, Japan). Using a tape measure on a straight wall, height was measured in standing posture to the closest centimeter. The formula for calculating the body mass index (kg/m2) was: weight (kg) divided by height squared (m2).
Assessment of other variables
Age, sex, physical activity expressed as metabolic equivalent hours per week (MET-h/wk), job, BMI, smoking status, education level, marital status, and average duration of watching television and mobile use per day were all collected using predefined questionnaires by professional interviewers and were considered possible confounders. The short form of the standard physical activity questionnaire (IPAQ), which is developed to track the level of physical activity in developing countries, was used to assess physical activity. The abbreviated form of physical activity includes questions related to four levels of physical activity such as intense, moderate, walking, and sitting activities over the past seven days, the intensity, duration, and repetition of which are completed by individuals. In this questionnaire, intensive physical activity is defined as at least a 10-minute exercise that generates a significant rise in respiration, heart rate, and sweating. Also, moderate activity in this questionnaire is an activity that has a minimum duration of 10 minutes and increases the average respiration, heart rate, and sweating. This questionnaire is standard and its reliability and validity have been confirmed (34).
Statistical analysis
Subjects were categorized into tertiles based on their total DII and DIL. For continuous data, mean ± standard error (SE) was used, and for categorical variables, percentage was used. The one-way analysis of variance (ANOVA) and chi-square test were used to examine the difference in quantitative and qualitative variables between tertiles of dietary DII/DIL. The analysis of covariance (ANCOVA) was used to compare dietary nutrient intake between the tertiles of DII/DIL, with adjustments for sex, age and energy intake. Logistic regression in crude and multi-variable adjusted models was done to assess the association between DII or DIL and the chance for developing short (< 5 h) vs. normal (5–8 h), long (> 8 h) vs. normal (5–8 h) sleep duration, and insufficient sleep (< 5 h) vs. sufficient sleep (≥ 5 h), low sleep quality (total score 8 and more) vs. normal sleep quality (total score 7 and less), delay in falling asleep (delay in falling asleep score 3) vs. low delay in falling asleep (delay in falling asleep score 2 and less), using medication for sleep [sleep medication use three or more times a week (score 3)] vs. low medication use for sleep [sleep medication use two or less times a week (score 2 or less)], have sleep disorder (sleep disorder score 2) vs. no sleep disorder (sleep disorder score 1 or less). The first tertiles of DII and DIL were used as a reference group to calculate odds ratios (ORs) and 95 % confidence intervals (CIs). The first model was adjusted for age, gender, and total calorie intake. In model 2, additional adjustments were performed for BMI (continuous), education (categorizing into Primary school and less/High school/ Diploma and Graduate Diploma/ Bachelor student/ Master student and Doctorate), physical activity (continuous), marital status (single/ married/ widowed or divorced), smoking status (never smoker/ current smoker/ ex-smoker) job situation (categorizing into unemployed/ government employee/ manual worker/ self-employed), anxiety (continuous), stress (continuous), caffeine (continuous), duration of cellphone use (continuous) and watching television & movie (continuous). In addition, in model 3, additional adjustments were made for the usage of sleeping medications (more than twice a week). SPSS software was used for all statistical analyses (version 26.0; SPSS Inc, Chicago IL). Statistical significance was defined as a P value < 0.05.