Anthropometric trajectory and dietary patterns: analysis of Longitudinal Study of Adult Health (ELSA-Brasil)

Author(s)

Mariane de Almeida Alves

• Document type: Doctoral Thesis
• Press: São Paulo.
• Institution: Universidade de São Paulo (USP). Faculdade de Saúde Pública (FSP/CIR)
• Defense date: 2023-07-05

Examining board members

Dirce Maria Lobo Marchioni; Bruna Kulik Hassan; Patricia de Fragas Hinnig; Sheila Maria Alvim de Matos

• Advisor: Dirce Maria Lobo Marchioni

Abstract

Despite the well-known relationship between diet and obesity, there is still some debate on how the diet composition influences the adiposity gain. Dietary pattern analysis emerges as an approach to evaluate the diet in a holistic way, considering the interaction between food groups and contributing to disentangle this relation between diet and adiposity. This thesis aimed to investigate the association between baseline dietary patterns and adiposity markers over the follow-up period. Methods: Longitudinal study embedded in the Longitudinal Study of Adult Health (ELSA-Brasil), a multicenter national cohort that recruited individuals aged 35 to 74 years old (N=15,105). Dietary consumption data was assessed at baseline by a food frequency questionnaire (FFQ) and the study populations dietary patterns were identified using three different statistical methods: 1) factor analysis (FA), 2) treelet transform (TT), 3) reduced rank regression (RRR). The associations between baseline dietary patterns and adiposity were computed using linear mixed-models. Applying tree different methods to identify dietary patterns, we found a common dietary pattern characterised by the consumption of sweets and desserts, snacks, red meat, processed meat, and soft drinks, labelled as convenience. Additionally, the FA and TT identified a similar dietary pattern characterised by cereals, fruits, vegetables, fish, and fruit juice, labelled as prudent. FA also identified a dietary pattern characterized by rice, beans, and red meat, representing a traditional Brazilian meal, while TT analysis identified a dietary pattern characterized by high consumption of rice and beans and low consumption of red meat and other groups present in the convenience dietary pattern, labelled as rice and beans. This study had a mean follow-up time of 8.2 years. A convenience dietary pattern was associated with higher adiposity measures over time. The Brazilian traditional dietary pattern was associated with lower weight over the follow-up, however for the body composition measures, this dietary pattern was associated with higher body fat percentage and fat mass index over the follow-up period. The rice and beans dietary pattern, characterized by high consumption of rice and beans combined with low consumption of red meat, was associated with lower levels of adiposity indicators. Lastly, the prudent dietary pattern identified by FA was associated with lower levels of adiposity indicators, except for weight. The same dietary pattern identified by TT analysis was associated with higher weight over the follow-up period. The RRR-dietary pattern explaining the circulating levels of branched-chain amino acids was characterized by high consumption of rice, beans, red and processed meat, poultry, soft drinks, snacks, pasta, and low consumption of cereals and milk. This RRR-dietary pattern was associated with higher adiposity for all indicators over the follow-up period. Applying tree different statistical methods to identify dietary patterns in the same population allowed us to find some nuances in the diet composition and further insights on how the dietary patterns of a population from Brazil may explain the changes on adiposity levels over time. (AU)

• FAPESP's process: 19/13486-5 - Anthropometric trajectory and dietary patterns: analysis of the Longitudinal Study of Adult Health (ELSA-Brazil)
• Grantee: Mariane de Almeida Alves
• Support Opportunities: Scholarships in Brazil - Doctorate