As phase 1, we conducted a feasibility study in order to identify the proportion of students who fulfill the 24 h movement behavior guidelines (21); as well as the psychometric properties of the subjective instruments, their reliability (temporal stability) and the construct validity. Here, all students answered the questionnaire twice, the second time after a two-week interval. For the sample size estimation, we adopted the assumptions of Nascimento-Ferreira et al. (22). The parameters used to calculate the sample size were: α = 0.05, β = 0.20 (power of 80%) and a correlation coefficient of 0.28 (22). Based on these parameters, we estimated a sample of 98 students. Anticipating losses of 50.0%, rejections of 50.0% and missing data of 25.0% (18) in each test-retest assessement, we decided to invite 342 ([2.5 × 98] + 98 ≅ 343) students to participate. At the design level, participants were distributed in a proportion of 60/40 by sex (female and male) and program of study (health and other sciences) based on previous cohorts (6, 23).

However, we carried out our sampling protocol only in the Maranhão research center due to national and regional health protocols and recommendations related to COVID-19. Students, in this phase, were randomly selected and invited to participate presencially. We delivered a link with the questionnaire to students who agreed to participate, and up to three reminders were sent in the subsequent initial invitation and t two-week intervals. Students provided informed consent to participate anonymously by completing and submitting the questionnaire in Google Forms (available at: https://forms.gle/L92wXsVaxxfPNgpE8). All data were self-reported electronically.

We all participants included in the study were enrolled in the university, 17 years or older, and had signed the informed consent form. Students who reported incomplete data were excluded from the analysis of the corresponding instrument (e.g. incomplete data from vigorous physical activity leads to the exclusion of that answer to the questionnaire from the reliability and validity analysis).

In phase 2, we will conduct a longitudinal study to evaluate the time of fulfilling 24 h movement behavior guidelines (or not) and its relation with metabolic syndrome (21). For the prospective cohort, the parameters used for sample size estimation were α = 0.01, β = 0.10 (or 90% power), prevalence of fulfilling 24 h movement behavior guidelines ranging from 7.7% (during the pandemic, Table 1) to 23.1% (in a non pandemic period (9)), precision 5% and design effect of 2.0 (multistage sampling) for a two-sample proportions test. The estimated sample size was 200 in each research center, we corrected this estimation for population ratio (2,225/968 = 2.3) achieving a minimum sample size of 279 and 121 students from the private (Maranhão research center) and the public (Tocantins research center) institution. We add 125.0%, anticipating potential losses, rejections and missing data. Thus, we will invite 502 students (349 from the private and 152 from the public institution).

In each institution, the sample will be selected by random sampling in two stages, program and students. Thus, a list of all students from up to second semester offered will requested. From this list in each stratum a simple random sampling will be carried out, with probability proportional to the number of students in each institution study program and shift (e.g., morning and afternoon) in relation to the total number of students. Then, in each selected program of studies and shift, students will be randomly invited to participate in the study. The probability of selection for each student will be condionted on the number of students in these two stratums and an additional invitation rate of 30.0% for males.

We will collect the data in four visits per wave. In the first visit, we will conduct an explanation of the 24h-MESYN study objectives/methodology and we will delivery the informed consent form. As this study phase will occur presencially, the student will sign two copies of the informed consent form (one for the student and another one for the researcher). In the second visit, after returning the signed form, participants will answer the questionnaires eletronically and will undergo an anthropometric and blood pressure assessment. In the third visit, a blood sample will be collected. In the fourth visit, we will deliver the results to the students.

We will adopt the same eligibility criteria as for phase 1. However, other inclusion criteria will be added: students must be regularly enrolled up to second semester and present in the classroom on the day of the objective measurements. We will exclude participants who fail to provide complete information about age, biological sex, program of studies, shift and time (period). In addition, only those students who answer the questionnaire and participate in the anthropometric measurements will be qualified for the blood sample collection. Students who report a physical disability will be evaluated, but they will be excluded from the analysis.

In phase 3, we will conduct a case–control study in the cohort first wave to test associations between 24 h movement, mental wellbeing, eating behaviors (exposures), and abdominal obesity (main outcome) and in a subgroup of students with the data collected in the baseline (21). Students with prevalent abdominal obesity will be selected as a case. We will adopt four controls per case and they will be paired by sex and age (±1 year) (24).

Metabolic syndrome will be based on the International Diabetes Federation criteria (2), which establish the presence of any 3 of 5 risk factors: presence of abnormal waist circumference (≥ 90 cm in men; ≥ 80 cm in women), blood pressure (systolic ≥ 130 and/or diastolic ≥ 85 mmHg), triglycerides (TG, ≥ 150 mg/dL), HDL cholesterol (HDL-c, ≤ 40 mg/dL in men; ≤ 50 mg/dL in women) and fasting blood glucose (≥ 100 mg/dL).

We will evaluate the waist circumference according to international standards (25) with an inelastic tape, at the midpoint between the lower edge of the last rib and the top iliac crest (26). A third measurement will only be performed in case of error greater than 5% between the first and second measurement. The assessments will be carried out in a private room at the institution. All measurements will be taken in underwear or with as little clothing as possible and no shoes (18).

Blood pressure measurements will be performed following the recommendations of the American Heart Association (27). We will perform measurements with the Omron HEM-7320-LA Automatic Arm Oscillometric Device (Omron Healthcare, Kyoto, Japan), with a pressure range of 0–299 mmHg and a heart rate range of 40–180 beats/minute. Blood pressure and heart rate will be measured twice, with an interval of 2 min between the two measurements, according to international guidelines. If the values of the second measurement vary by more than 5% from the first, a third measurement will be taken (28).

A blood venous sample will be collected using a Vacutainer system (Becton Dickinson, Oxford, UK). This procedure will be carried out in the morning, after an 8–12 h overnight fast. Participants will be instructed to fast on the morning of the collection day. Samples will be collected and analyzed by a previously accredited laboratory following international protocols and recommendations (18), such as: do not perform moderate-to-vigorous physical activities and do not consume products with caffeine and alcoholic beverages and avoid drinking water in the hours preceding the assessments.

The exposures will be 24 h movement behavior, mental wellbeing (perceived stress and satisfaction with life) and eating behaviors. These exposures will be subjectively measured by self-adminstered validated questionnaires. All subjective instruments are available at Google Forms (https://forms.gle/L92wXsVaxxfPNgpE8).

We will adopt two techniques for assessing 24 h movement behavior data, compositional data and fulfillment of the guidelines (7). The compositional data will be assessed by the total time spent in physical activity (light and moderate-to-vigorous), sedentary behavior and sleep time during a 24 h period (7). Additionally, we will classify 24 h movement behavior guidelines into fulfilling or not fulfilling the following criteria: (1) ≥150 min per week of moderate-to-vigorous of physical activity; (2) ≤ 8 h per day of sedentary time (No more than 3 h per day of recreational screen time); (3) sleeping 7–9 h per day; and (4) adhering to all three guidelines (29).

Physical activity will be measured via the International Physical Activity Questionnaire, short version (30). A instrument that assesses physical activity by questions about frequency, duration and intensity of activities (light, moderate and vigorous), including activities performed during leisure and as a form of transport. Sedentary behavior will be assessed via the South American Youth Cardiovascular and Environmental (SAYCARE) sedentary behavior questionnaire (31). An instrument composed by 10 questions about time spent on weekdays and weekends, in activities such as watching television; playing games on a computer, cell phone or tablet; computer use; studying, reading books or magazines; and driving a car, bus or train (31). In addition, we will apply the Longitudinal Aging Study Amsterdam Sedentary Behavior Questionnaire (LASA-SBQ) addressing average time spent watching television, using a computer, sitting, talking, napping, among other sedentary behaivours, during weekdays, and weekend days through 20 items (32). Sleep quality, habits and time will be measured via the Pittsburgh Sleep Quality Index (33). This is a self-administered 19-question questionnaire that assesses seven components of sleep (sleep quality, sleep latency, sleep duration, habitual sleep savings, sleep changes, sleep changes, use of medications for sleep, and sleep dysfunction) (33).

Perceived stress will be assessed by the Perceived Stress Scale (34), a scale based on 14 questions with answer options ranging from 0 to 4 (0 = “never” to 4 = “always”). Questions with a positive connotation (4–7, 9, 10, 13) have their summed score inverted. The other questions are negative and must be added directly. The scale total is the sum of the scores for all these 14 questions. Satisfaction with life will be assessed using the Satisfaction with Life Scale (35), a questionnaire that assesses the cognitive component of life satisfaction, through five items, with response options ranging from 1 (strongly disagree) to 7 (strongly agree), given that the closer to 7 the response, higher are the life satisfaction ratings.

Eating behaviors (psychosocial aspects) will be measured via the Dutch Eating Behavior Questionnaire (DEBQ) (36). The questionnaire consists of 33 items, evaluated on a scale rated between 1 and 5 (never/rarely/sometimes/often/veryoften), comprising three subscales: food restriction (10 items); emotional intake (13 items); and external intake (10 items).

We will retrieve variables about sociodemographic and economic characteristics, academic environment and COVID-19 (perceived) burden using self-admininstered instruments. We will also assess the body mass index objectively.

Sociodemographic variables will be based on questions addressing biological sex, age, ethnicity, marital status and neighborhood (37). Socioeconomic status will be assessed by the Brazilian Economic Classification Criteria (38), a self-administered instrument that uses a point system for possession of goods and education of the family head, classified into A to E, with A being the most educated and wealthy and E the least. The questionnaire estimates the purchasing power of individuals and families. In addition, we will include questions addressing academic environment: maternal education (incomplete high school, high school, technical education, university degree), personal and family income (in minimum wages) and daily workload (hours per day).

Regarding the academic environment, the following data will be collected: undergraduate study program area, shift, time (semester), number of classes enrolled (in the semester) and time spent studying for classes (hours per day). For the COVID-19 burden, we will retrieve data from knowledge, sources of information, behaviors, and academic and life difficulties via a questionnaire based on the fact sheets of the Centers for Disease Control and Prevention (CDC) (39). A total of 10 questions, for seven of which the answer can be true or false, while three questions will assess the level of agreement with the statements on a 5-point scale (0 = “strongly disagree” to 4 = “agree totally”). And, for the body mass index, we will measure weight (kg) and height (cm). The body mass index will be calculated using the standard formula (BMI = weight/height²).

Continuous variables will be described by their means and standard deviations. Categorical variables will be described by absolute and relative frequency, and 95% confidence interval (40). In addition, we will use time series to describe the trend of the variables. A descriptive analysis will identify the time elapsed between exposure to 24 h movement behavior to a metabolic syndrome event. The Kaplan-Meier curve will be computed.

To test the reliability, we will use the intraclass correlation coefficient (parametric) or Spearman correlation coefficients (non-parametric) and kappa agreement coefficient (for temporal stability), and Cronbach's alpha (internal consistency) (40). To test the validity, we will use an exploratory factor analysis (structural validity) with varimax rotation. Previously, we will use Bartlett's KMO and sphericity test to verify the adequacy of the factor analysis (40).

Our conceptual model is presented in Figure 2. For outcome variables that are continuous, we will apply multilevel linear regression models to estimate the beta coefficient of the exposures of interest (40). For categorized variables in the case-control analysis (phase 3), we will apply a logistic regression. In addition, for categorical outcome variables, Poisson multilevel regression models with robust variance to estimate the prevalence ratios (cross-sectional data) and the relative risk (longitudinal data) will also applied (40). In both multilevel regressions, the contextual variable will be the research center. For choosing the type of multilevel model, we will use the Durbin-Wu-Hausmann method to determine random or fixed intercept model. We will analyse the fit using the Hosmer-Lemeshow test. In bivariate analysis, we will adopt a significance value of 80% (p ≤ 0.20).