Premium
Trends in Dietary Quality Among Middle‐aged Adults from the Framingham Heart Study
Author(s) -
Troy Lisa M
Publication year - 2017
Publication title -
the faseb journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.709
H-Index - 277
eISSN - 1530-6860
pISSN - 0892-6638
DOI - 10.1096/fasebj.31.1_supplement.lb458
Subject(s) - medicine , waist , multivitamin , framingham heart study , demography , national health and nutrition examination survey , cross sectional study , body mass index , gerontology , percentile , framingham risk score , environmental health , population , disease , statistics , mathematics , pathology , sociology , vitamin
Analysis of cross‐sectional NHANES data has observed improved diet quality among U.S. adults. Longitudinal data may show different results. Therefore, trends of diet quality between exam 5 (1991–1995) and exam 8 (2005–2008) were examined among Framingham Heart Study Offspring participants (n=3330, mean age 58 yrs, mean BMI 27.8 kg/m 2 ). Further, participant characteristics that predicted higher or lower diet quality were identified. Diet quality was assessed by applying Willett Food Frequency Questionnaire data to create the 2015 Dietary Guidelines for Americans Adherence Index (DGAI‐2015), a measure of adherence to the key dietary recommendations of the 2015 Dietary Guidelines (range 0–100, higher score indicates greater adherence). To assess longitudinal trends, mixed models were applied to the DGAI‐2015 and examined annualized change over a mean of 12.5 years of follow‐up. Cross‐sectional linear regression at exam 5 was used to identify characteristics (i.e., marital status (single, married, windowed, divorced/separated), BMI, waist circumference (cm), level of education (years), multivitamin use (y/n), diet soda intake (y/n), statin use (y/n), cholesterol lowering medications (y/n), diabetes status (y/n), physical activity (total METS) and sedentary behavior (total sitting time) as potential predictors of diet quality. To identify the strongest predictors, all individually significant predictors were placed in linear models. Models were applied to all participants and then in separate models by sex (n=1566 men and 1764 women) given known differences in dietary quality between men and women. The DGAI‐2015 (score range, 24.3 to 87.8) distinguished between groups with known differences in diet quality; women, older, and non‐smokers had higher DGAI‐2015, as expected, demonstrating construct validity. No longitudinal changes in diet quality were observed for all participants or for men and women separately. Participants with higher DGAI‐2015, were more likely to be multivitamin users, cholesterol lowering medication non‐users, and to have lower BMIs and higher levels of education and less likely to consume artificially sweetened soft drinks in models adjusted for age, sex and smoking. In sex‐specific models adjusted for age and smoking, higher DGAI‐2015 scores were observed with lower waist circumference, higher levels of education, and less likely to consume artificially sweetened soft drinks for men and for women. Important differences in predictors were observed between men and women. Higher DGAI‐2015 was associated with cholesterol lowering medication non‐use in men only and multivitamin use in women and lower DGAI‐2015 with sedentary behavior in women. Longitudinal analysis of trends in diet quality do not support cross‐sectional data describing increases in higher diet quality. There may be important differences in predictors of diet quality between men and women.