LAUSR

Evidence is strong for a holistic approach for diet to prevent cardiovascular disease, diabetes, and other conditions. This includes dietary strategies, such as following a Mediterranean diet or otherwise eating a dietary pattern with low meat intake and a high intake of fruit, vegetables, legumes, whole grains, nuts, berries, seeds generally, unrefined unsaturated oils, fish, coffee, tea, chocolate, and alcohol (among drinkers, not in excess; intoxication should be avoided). Water is very important as a beverage. Some forms of dairy foods may also be recommended. Although an understanding of nutrients is important for controlling deficiency diseases, nutrients are not a good model for choosing what foods to eat. A nutrient approach to nutrition has been confusing to consumers. Paramount in this concept are the terms “low fat” and “low carbohydrate”; neither is particularly useful for choosing what foods to eat. “Low fat” as a food choice rule excludes apparently healthy foods, such as olives, walnuts, and salmon. “Low carbohydrate” would include these high-fat foods, but seems to discourage the consumption of whole plant foods. The idea that there is a concerted action of food constituents is called “food synergy”; this idea also supports that certain dietary patterns may be healthy. The food synergy and holistic idea is further supported by the notion that variety in the consumption of nutrientor bioactive-rich plant food helps maintain health because there are many compounds with many functions. A simple example of the highly nuanced nature of biochemicals on health is the difference between oleic acid and its trans counterpart elaidic acid. The two molecules differ only in the placement of one double bond, yet the cis version, oleic acid, is a major part of olive oil, which is considered to be healthy, whereas the trans version, elaidic acid, is rare in nature, produced by hydrogenation of vegetable oils, and is associated with excess risk of coronary heart disease. The consumer does not know in advance of eating which compounds are needed on any given day. There is also an epidemiologic advantage to studying diet patterns. Although many individual foods and nutrients have been seen to have very high within-person variability, diet patterns act like many other cardiovascular or diabetes risk factors, showing correlations of ~0.5 over periods up to 20 years. Jacobs and Orlich reviewed reports of 16 diet pattern scores and in all but one case found tracking correlations over several to many years of follow-up to range from 0.39 to 0.65. Jacobs and colleagues have formulated the A Priori Diet Quality Score (APDQS). The theoretical basis for the APDQS is discussed by Jacobs and Orlich. Because the APDQS is based on foods, it differs somewhat from study to study, depending on which food groups were asked about. In Project EAT (Hu et al., table 4), positively rated food groups were formed, namely beans, chocolate, coffee, fish, fruit, green vegetables, lowfat dairy, nuts, poultry, tea, tomato, whole grain, and yellow vegetables. The negatively rated food groups were butter, fried foods, fried potato, grain dessert, organ meat, processed meat, red meat, salty snacks, sauces, soft drinks, sweet bread, and whole fat dairy. Each food group was placed in quintiles (or a non-consumer group and quartiles among consumers), and categories were scored 0–4 for lowest to highest reported consumption. The APDQS was calculated by summing category scores (0–4) of positively rated food groups with the reverse scores (4–0) of the negatively rated food groups. Additional foods were rated as neutral, namely diet soft drinks, eggs, fruit juice, margarine, other vegetables, potatoes, refined grains, seafood, and soup; these provide the full dietary context, but do not enter the score except through energy balance (replacing either positively or negatively rated foods). The APDQS has been shown to be inversely associated with total death that occurred over 26 years of follow-up from 1986 in Iowa women who were then aged 55–69 years. The APDQS is also related to health parameters in people followed from young adulthood. However, APDQS has not been the subject of much study in children. Using data from Project EAT, we hypothesized that: (i) APDQS tracks from adolescence into young adulthood; and (ii) higher and increasing APDQS is associated with less weight gain in these participants. The following summarizes published data. In Project Eat, 4746 adolescents were recruited from middle or high school in Minneapolis–Saint Paul (MN, USA) in 1998–1999, with a mean age of 15 years (range 11–18 years). The study sample for the present report was the 2656 participants with follow-up visits in 2004 (mean age 20 years) and/or 2009 (mean age 25 years), adequate diet data, and self-reported weight and height doi: 10.1111/1753-0407.12790 Journal of Diabetes 10 (2018), 796–798