Dietary macronutrient intake according to sex and trait anxiety level among non-diabetic adults: a cross-sectional study

Abstract

Background Studies suggest that anxiety is correlated with eating behavior, however, little is known about the association between anxiety status as predictor of dietary macronutrient intake. The aim of the present study was to investigate the sex-stratified cross-sectional associations of trait anxiety with intake of various macronutrients in a large population-based sample of non-diabetic adults. Methods N \u2009=\u200920,231 participants (mean age\u2009=\u200953.7\u2009±\u200913.6\xa0years) of the NutriNet-Santé web-cohort, who had completed the trait anxiety subscale of the Spielberger State-Trait Anxiety Inventory (T-STAI; 2013–2016) were included in the analyses. Dietary intake was calculated from at least 3 self-administered 24-h dietary records. The associations of interest were assessed by multiple linear regression stratified by sex, owing to significant interaction tests. Results In total, 74.3% ( n \u2009=\u200915,033) of the sample were females who had a significantly higher mean T-STAI score than did males (39.0 versus 34.8; p \u2009<\u20090.01). Among females, the fully-adjusted analyses showed significant positive associations of T-STAI with total carbohydrate intake (β\u2009=\u20090.04; p \u2009<\u20090.04), complex carbohydrate intake (β\u2009=\u20090.05; p \u2009<\u20090.02), and percentage energy from carbohydrates (β\u2009=\u20090.01; p \u2009<\u20090.03), as well as a significant inverse association of T-STAI with percentage energy from fat (β\u2009=\u2009-0.01; p \u2009<\u20090.05). As regards males, the only significant finding was an inverse association between T-STAI and percent of the mean daily energy from protein (fully-adjusted model: β\u2009=\u2009-0.01; p \u2009=\u20090.05). Conclusion This cross-sectional study found modest sex-specific associations between anxiety status and macronutrient intake among French non-diabetic adults. Prospective studies are needed to further elucidate the observed associations.