ABC | Volume 112, Nº6, June 2019

Original Article Scheer et al Exercise at school and cardiovascular risk factors Arq Bras Cardiol. 2019; 112(6):775-781 In 2012, Rio de Janeiro’s city government initiated a project integrating academic and sports training: the Olympic Experimental Gymnasium (OEG). 10 Emphasis was given to sports practice (two hours a day, five times a week), and five healthy meals were provided to students every day. Conversely, in regular schools (RSch), PA practice is limited (once a week) and students have only one meal per day. In this context, we aimed to examine potential associations between the school environment in OEG schools versus RSch and important cardiovascular disease (CVD) risk factors in adolescents. We hypothesized that the adolescents attending RSch would be more likely to have risk factors than those attending OEG schools. Methods This article was written according to the standards of the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement for cross-sectional studies. 11 Study design A cross-sectional study with a comparator group was conducted at six public schools in Rio de Janeiro: three OEG schools and three RSch. Data collection was carried out in the schools from 2013 to 2015. Data were always collected in the mornings, on the same day of the week by the same trained investigator. To draw association inferences, we considered the school type (ER or OEG) as the exposure, and important CVD risk factors as the outcome, treated as categorical variables. A total of 1,113 students were included in a 2:1 allocation ratio to exposure (719 attending OEG schools) and non-exposure (394 attending RSch). Eligibility criteria and participant selection Eligible participants were all regularly enrolled sixth- to ninth-grade students attending the same school for at least one year. Students attending OEG schools needed to comply with the OEG policy of engaging in PA with energy expenditure > 5 metabolic equivalents (METs) five times a week. Attendance at PA sessions was self-reported by the students or their parents. When selecting participants for the comparator group, efforts were made to match the students as closely as possible to avoid potential confounding factors. Thus, a match for chronological age, sex, and grade was made. In RSch, students spent about one hour weekly in PA. At the time of data collection, fourOEG schools were operating in Rio de Janeiro, but only three of them had been operating for at least one year. Therefore, all of the OEG schools were enrolled with no a priori sample size planning. RSch, selected by convenience sampling, were located in the same geographic region of the OEG schools, to avoid discrepancies especially due to socioeconomic status. BothOEGandRSchwere public schools, attended by students from all regions of the city. Data collection Written informed consent was obtained from all students and their parents/guardians prior to inclusion in the study. The first school visit was then scheduled, and we performed the following procedures: capillary blood collection, medical interview, physical examination and blood pressure (BP) measurement. All data were collected following a circuit format, in which students passed by each station where the measurements were performed. A trained medical staff interviewed the students and their parents/guardians using a structured questionnaire designed specifically for this study, in which data of sex and age, in addition to past medical history and current relevant health information were obtained. The medical staff also evaluated adolescents, including anthropometric assessment, measurement of BP and blood sample collection. Capillary blood samples were collected after a 12‑hour overnight fasting to determine glucose levels, total cholesterol, and triglycerides (Roche AccuTrend Plus Kits®). Altered glucose, total cholesterol and triglycerides levels were considered as results greater than 99 mg/dL, 170 mg/dL and 130 mg/dL, respectively. Height and weight were measured using a Filizola® scale, with maximum capacity of 150 kg, and a coupled vertical stadiometer of 220 cm. Based on BMI values, individuals were classified as eutrophic, overweight or obese, according to the percentiles of height and age for each individual subject. Office BP was measured three times in sitting position, using a calibrated aneroid sphygmomanometer in millimeters of mercury. Students with systolic and diastolic BP above the 95th percentile for their respective gender, age and height were considered hypertensive, while students with systolic and diastolic BP values between their 90th and 95th percentiles were classified as pre-hypertensive. All participants included in the analysis had participated in the PA program for at least one year. This was planned to minimize the risk of reverse causality. Reference values were based on the Brazilian Pediatric Society standards. 12 Statistical analyses A descriptive analysis was conducted, and data were expressed as mean and standard deviation for baseline characteristics and continuous variables, and as absolute frequencies and percentages for categorical variables. In the inferential analysis, the Shapiro-Wilk test was used to assess the normality of data distribution. Groups were then compared by an independent sample two-tailed Student’s t test for continuous variables, and by the χ 2 test for categorical variables. Association of school exposure with health status outcomes was assessed by a binomial logistic regression, and its diagnostic was performed to support the analysis. The health status outcomes regarding altered triglycerides, cholesterol, overweight/obesity (regarded as one variable), and pre-hypertension/hypertension were dichotomized into ‘YES’ and ‘NO’ according to the thresholds recommended by the Brazilian Pediatric Society. 12 Point estimates for the associations were expressed as odds ratio (OR) with their 95% confidence intervals (95% CI), and all analyses were adjusted for sex and age. An exploratory sensitivity analysis was planned in cases of disagreement between matching variables. ‘OEG School’ was chosen as the reference group for all analyses. No imputation for potential missing data was planned, and all available data for each analysis were used. All statistical analyses were conducted in the StataSE package v. 14.0. An α level of 0.05 was set for statistical inferences. 776