Total and specific dietary polyphenol intakes and 6-year anthropometric changes in a middle-aged general population cohort.

Abstract

Dietary polyphenols are suggested antiobesogenic agents. Prospective evidence in general population of an association between polyphenol intakes and anthropometry is lacking. To assess the associations between dietary polyphenol intakes and changes in body mass index (BMI) and waist circumference (WC) over a 6-year period. Individual intakes of 264 different polyphenols (mg day-1) were estimated using the Phenol-Explorer database and the mean of 6-17 24-h dietary records collected in 1994-1996. BMI in kg m-2 and WC in cm were measured in 1995-1996, 1998-1999 and 2001-2002. Linear mixed-effect models allowed for the assessment of longitudinal associations between energy-adjusted quartiles of total polyphenol intake as well as intake of 15 polyphenol classes and changes of these respective polyphenol classes in anthropometry over the 6 years of follow-up. Adjustment variables included sex, age, socio-economic status, lifestyle, dietary intakes and health status. Participants in the highest quartile of intake of flavanones (BMI change: -0.28 (-0.43; -0.13), P=0.009), flavones (BMI change: -0.29 (-0.44; -0.14), P=0.008) and lignans (BMI change: -0.28 (-1.63; -0.09), P=0.01) experienced a less notable increase in BMI over time compared with their counterparts in the bottom quartile of intake of the respective polyphenol classes. Participants in the highest quartile of intake of flavanones (WC change: -1.39 (-2.02; -0.92), P=0.001), flavones (WC change: -1.57 (-2.32; -0.92), P=0.001), hydroxycinnamic acids (WC change: -1.27 (-1.92; -0.63), P=0.01), lignans (WC change: -1.16 (-1.80; -0.51), P=0.006) and total polyphenol intake (WC change: -1.39 (-2.05; -0.74), P=0.001) experienced a less notable increase in WC over time compared with their counterparts in the bottom quartile of intake of the respective polyphenols. Dietary polyphenol intakes may help reduce weight gain over time in the general population. This could have important public health implications because moderate increases in BMI and WC over time have been shown to increase disease risk.