Exploring the association between temperature and overweight: direct and mediated effects of physical activity and fruit and vegetable consumption in Argentina.

Abstract

Studies show that climatic conditions can influence physical activity habits, food choices and food availability. However, the direct effects of climate on nutritional status, and the indirect effects through these pathways, have been scarcely studied. This paper aims to assess the effect of environmental temperature on overweight, mediated by physical activity and fruit and vegetable consumption, from a multilevel study in Argentina in 2018.

A cross-sectional, observational-analytic study was conducted using secondary data from the National Survey of Risk Factors in Argentina, 2018 (probability sample of 15586 adults). A mediation analysis was performed based on a multilevel logistic GSEM model, adjusted by sex, age and educational level. The outcome was defined as the overweight variable (Body Mass Index ≥25, yes/no), and physical activity level (low/moderate/high) and fruit/vegetable consumption (average daily servings) as mediating variables in their relationship with the environmental temperature variable. This corresponds to the annual, provincial median land surface temperature, obtained by processing geospatial data using Google Earth Engine. Given the hierarchical structure of the information, a contextual latent variable (province/jurisdiction of residence) was introduced into the model. The raw difference proposed by Coutts and Hayes and its confidence intervals were calculated, using bootstrapping techniques (500 replicates on a sample of 2000) to compare the magnitude of the indirect effects.

The direct effect of temperature on overweight was inverse and significant (β= -0.016; p=0.017). In addition, physical activity and fruit/vegetable consumption were found to play the role of indirect mediators of this relationship. It was observed that higher ambient temperature increased fruit/vegetable consumption (β=0.007; p=0.006) and decreased physical activity (β=-0.014; p<0.001). When comparing the indirect effects, the level of physical activity showed a significantly higher magnitude (p=0.031) mediating effect on the temperature-overweight relationship.

In conclusion, there is an effect of temperature on fruit/vegetable intake and physical activity that indirectly affects nutritional status, with the mediating role of physical activity level being more important.