Dietary patterns and weight change: 15-year longitudinal study in Australian adults

Abstract

Background: Dietary intake is one of the most modifiable risk factors associated with obesity. However, data on the relationship between dietary patterns and long-term weight change are limited.

Purpose: We therefore investigated the association between dietary patterns and 15-year weight change in a sample of 1186 Australian adults (1992-2007).

Methods: We measured body weight and collected data on socio-demographic and lifestyle characteristics in 1992 and 2007. Applying principal component analysis to 38 food groups from a food frequency questionnaire collected at baseline, we identified two dietary patterns: 'meat-and-fat' and 'fruit-and-vegetable.' Using generalized estimating equations, multivariable regression models, stratified by sex, were adjusted for concurrent changes in socio-demographic and lifestyle variables.

Results: The average increase in body weight of men in the highest tertile of the meat-and-fat pattern was more than twice that of men in the lowest tertile; mean weight change (95 % CI): 4.8 (-0.1, 9.7) kg versus 2.3 (-2.6, 7.1) kg, P-for-trend = 0.02. In contrast, average weight gain of men in the highest tertile of the fruit-and-vegetable pattern was only about half that of men in the lowest tertile; mean weight change (95 % CI): 2.9 (-2.0, 7.8) kg versus 5.4 (-1.5, 10.4) kg, P-for-trend = 0.02. Among women, dietary patterns were not related to weight change.

Conclusions: These dietary patterns predict change in body weight in men, but not in women. In this cohort, a dietary pattern high in fruit and vegetables was related to less weight gain in men than a dietary pattern high in meat and fat.

Keywords: Dietary pattern; Generalized estimating equations analysis; Longitudinal study; Weight change.