Associações bidirecionais entre atividade física e sintomas depressivos e seus efeitos inter-relacionados na mortalidade em idosos: evidências da coorte “COMO VAI?” em Pelotas/RS

Abstract

This thesis consists of two articles prepared using data from a cohort study called “COMO VAI?”, conducted with a representative sample of the older adult population from the city of Pelotas-RS. In the first article, we aimed to evaluate the bidirectional and longitudinal association between physical activity (PA) and depressive symptoms (DS). The analyses used baseline data (2014) and a follow-up carried out in 2019/20. The long version of the International Physical Activity Questionnaire (IPAQ) was applied to assess PA in the leisure, commuting and total domains (allowing estimates combining the two domains as total PA). The presence of DS was measured using the Geriatric Depression Scale (GDS-10), with a score ≥5 being considered DS presence. Linear regression models were used to assess bidirectional associations between PA and DS. The analytical sample of this first study included 515 individuals with a mean age of 70.7 (±9.2) years. The total PA (β: -0.0006; 95% CI: -0.0011; -0.0001) and commuting PA (β: -0.0008; 95% CI: -0.0016; -0.0001) in 2014 predicts worse DS scores in 2019/20. At the same time, the DS score in 2014 (β: -9.79; 95% CI: -18.81; - 0.76) was a predictor of lower leisure-time PA practice in the 2019/20 follow-up, adjusted for sociodemographic aspects and other health behaviors. However, when adjusted for the respective baseline outcomes, the associations found did not remain statistically significant. The present study did not find evidence of bidirectionality between PA and DS. Whereas, in article 2, the moderating and mediating role of PA in the association between DS and all-cause mortality was evaluated. The analyses also used data from the baseline (2014) and the first follow-up carried out in 2017. PA was assessed in the leisure, commuting and total domains using the IPAQ. Moderate to vigorous PA (MVPA) and overall PA were assessed using the GENEActive® accelerometer. Mortality information was identified through reports from family members, neighbors and death certificates confirmed by the Municipal Health Service Mortality Information System. Cox regression models were used to assess the association between DS and all-cause mortality moderated by PA. We conducted an analysis using four-way decomposition, which unifies mediation and interaction to assess the mediation of PA in this relationship. The analytical sample of article 2 included 1394 older adults with a mean age of 70.7 (±9.2) years. Older adults with DS had a twice-higher risk of mortality compared to those without symptoms (HR 2.31; 95% CI: 1.45; 3.68). MVPA (HR 3.13; 95% CI: 1.16; 8.50) and overall (HR 5.26; 95% CI: 1.62; 17.04), both by accelerometry, and the leisure (HR 2.33; 95% CI 1.12; 4.85) and total (HR 2.37; 95% CI 1.05; 5.35) domains moderated the association between DS and mortality. Commuting PA was not a moderator in the association between DS and mortality. Furthermore, no mediating effect was observed for any of the PA measures in the association between DS and mortality. Practicing PA even below the recommended level can minimize the effects of DS on mortality in older adults. Our evidence provides additional motivation for implementing practices and policies to increase PA in older adults with DS. Keywords: Depressive Symptoms; Physical Activity; Older adults; Cohort Study; Mental Health; Accelerometry.