Padrão alimentar e mudança de peso em quatro anos de seguimento: resultados do ELSA-Brasil

Abstract

Weight gain has increased in the last century, being a risk factor for all-cause mortality and a predictor of obesity, hypertension, diabetes, cardiovascular disease and some types of cancer. Investigating the relationship between food consumption patterns and weight gain or loss is essential to provide scientific evidence that can guide measures to address this problem in the population. Approaches centered on dietary patterns, that is, on the set of foods frequently consumed by individuals and populations, have proved to be advantageous, as they reflect the complexity of individuals' diets and are more easily translated into public health messages. In this sense, this dissertation aims to investigate the association between dietary patterns and weight change over four years of follow-up, among participants of the Longitudinal Study of Adult Health (ELSA-Brasil). ELSA-Brasil is a multicentric cohort composed of 15,105 civil servants aged 35 to 74 years from higher education and research institutions located in six Brazilian capitals: Belo Horizonte, Porto Alegre, Rio de Janeiro, São Paulo, Salvador and Vitória. This is a longitudinal observational study using data from the baseline (Visit 1) of ELSA-Brasil, conducted in 2008-2010, and the second visit (Visit 2) of examinations and interviews conducted in 2012-2014. The sample of the present study consisted of 13,568 (6,185 men and 7,383 women). The food patterns Traditional, Fruits and Vegetables, Pastry and Diet/Light were previously defined in another ELSA-Brasil study using the Food Frequency Questionnaire of Visit 1. The weight of the participants, in kilograms, measured at Visit 1 and Visit 2 was used. Potential confounders, measured at Visit 1, were sociodemographic factors (age, education, and nature of occupation), health-related behaviors (physical activity, smoking, and alcohol consumption) and health indicators (referred diabetes and high blood pressure and menopause, for women). To estimate the association between dietary patterns at Visit 1 and weight change over follow-up time, mixed-effects linear regression models were used, adjusted for confounders, with the interaction term between the fixed-effect variable (dietary pattern) and time to verify longitudinal changes in weight. Women had a mean age of 51.9 (±8.8) years and men 51.8 (±9.2) years. The most frequent dietary pattern was the Traditional pattern, respectively 48.9% among men and 45.3% among women. After adjustments, the Fruits and Vegetables dietary pattern showed a lower mean weight gain in women (β: -0.10; 95%CI: - 0.16; -0.05) and in men (β: -0.12; 95%CI: -0.19; -0.06) when compared to the Traditional pattern. Sensitivity analyzes were performed with the inclusion of the variables total calories, change in diet in the last six months prior to the interview at Visit 1, 11âncer11 f medical diagnosis of 11âncer, with no change in these results. The Pastry and Diet/Light patterns showed less weight gain over the years when compared to the reference pattern, but these differences were not statistically significant. These results may contribute to encouraging the planning of policies aimed at subsidizing healthy foods with the aim of improving food consumption and contributing to weight maintenance for populations similar to that of the present study.