Background: The 2015-2020 Dietary Guidelines for Americans (DGA) provides specific intake recommendations for vegetable variety and amount in order to protect against chronic disease. However, to the best of our knowledge, no studies have examined the link between DGA recommended vegetable variety and cardiometabolic disease. To address this research gap, our aim was to estimate the relationship between vegetable variety, vegetable amount, and prevalent cardiometabolic disease subtypes, and to assess potential determinants of vegetable variety.
Methods: Data on food intake and reported cardiometabolic disease status were acquired for 38,981 adults from the National Health and Nutrition Examination Survey (1999-2014). Vegetable variety was measured using a modified dietary diversity index that was adjusted for the potential confounding effects of vegetable amount. Temporal trends in vegetable variety and amount were assessed using univariate linear regression models. Multivariate logistic regression models were used to estimate the relationship between vegetable variety and prevalent disease, and between vegetable amount and prevalent disease. Multivariate ordered logistic regression models were used to assess the relationship between vegetable variety and explanatory variables.
Results: Overall, vegetable variety decreased (P = 0.035) from 1999 to 2014, but vegetable amount did not (P = 0.864). Intake of starchy vegetables decreased (P < 0.001), and intake of dark green vegetables increased (P < 0.001) over this 16-year period, but no trends were observed for other subgroups. An inverse linear relationship was observed between vegetable variety and prevalent coronary heart disease (P-trend = 0.032) but not other prevalent diseases; and between vegetable amount and coronary heart disease (P-trend = 0.026) but not other prevalent diseases. Individuals who reported consuming dark green vegetables had lower odds of having cardiovascular disease (0.86, 95% CI: 0.74-0.99) and coronary heart disease (0.78, 0.65-0.94) compared to individuals who reported not consuming any green vegetables. Living with a domestic partner was associated with greater vegetable variety (P = < 0.001), and currently smoking was associated with lower vegetable variety (P = < 0.001). Vegetable variety and amount were positively associated (P < 0.001).
Conclusions: Vegetable variety and amount were inversely associated with prevalent coronary heart disease. Vegetable variety was strongly associated with vegetable amount, likely mediated by reduced habituation and increased liking. Increasing vegetable variety and amount are still important messages for the public.