Dietary diversity and subsequent cause-specific mortality in the NHANES I epidemiologic follow-up study.

Abstract
Human diets tend to be complex mixtures of foods and nutrients. Therefore, we examined the relation of a measure of overall diet quality (independent of intake of individual foods or nutrients) with mortality from cardiovascular disease (CVD), cancer, and non-CVD, non-cancer (other) causes. We used data from the NHANES I Epidemiologic follow-up study (n = 10,337; median follow-up time = 14 years; age 25-74 years at baseline), and included 988 CVD, 571 cancer, and 910 other cases. The 24-hour dietary recalls obtained at baseline were scored for quality using a dietary diversity score (DDS). The DDS (range 0-5) counts the number of major food groups–dairy, meat, grain, fruit, and vegetable consumed daily. Age-adjusted risk of mortality from all three causes (except cancer in women) was inversely related with DDS in both men and women. Adjustment for multiple covariates attenuated the relative risk estimates slightly for CVD and cancer mortality, but markedly for other mortality. The results are suggestive of an increased risk of CVD and cancer mortality associated with diets characterized by omission of several major food groups.