This study confirms that micronutrient inadequacy increases as girls moved from early to mid-adolescence and beyond. By 14–18 years of age, almost 90% failed to meet the EAR for the intake of calcium and more than 90% failed to meet guidelines for magnesium, potassium and vitamins D and E. These estimates are similar to those from the 2005 and 2006 NHANES data [9
]. Large numbers of girls also had inadequate intakes of vitamin A, zinc and phosphorus and even vitamins B6, B12 and C.
Nutrient density is a relative concept. Foods in the same food group may differ in nutrient density as well as in energy density. Grains, for instance, can be consumed in both energy rich and less energy-dense forms. This study demonstrates the potentially important contribution of nutrient-dense eating patterns to micronutrient adequacy amongst adolescent girls. It also points to the role of excess consumption of discretionary calories in the form of solid fat and added sugars to micronutrient inadequacy. While the 2010 DGA cites that SoFAS account for approximately 35% of calories consumed by Americans [2
], girls in this prospective study consumed approximately 40% of their total daily calories from these energy dense, nutrient-poor sources. The current finding is consistent with published cross-sectional data from younger school-age subjects in NHANES 1999–2004 [26
]. The excess consumption of such discretionary calories in place of more nutrient-dense foods likely plays an important role in the failure of children and adolescents to meet DGA and IOM dietary recommendations for both foods and nutrients.
The contribution of food groups and subgroups to micronutrient adequacy depends on several factors—nutrient composition of the individual foods in a group/subgroup, the mix of foods eaten within the food group/subgroup, the amount of each consumed, and the correlation of intakes within and between food groups. For example, whole grain consumption might be correlated with fruit consumption among individuals focusing on a healthy diet while higher red meat intake might be correlated with consumption of starchy vegetables (e.g., French fries) in other individuals. Thus, correlated eating patterns may obscure the true nutrient contribution of individual foods or food groups. To address this issue, we constructed simple eating patterns that reflected the different ways in which foods may be eaten (e.g., some girls with high red meat intakes consume few FVs while others may consume more). In this way, we were able to see the effects of red meat independent of its associations with FV intakes.
Dairy is not only a rich source of many nutrients [27
] but its association with consumption of whole and fortified grains (i.e.
, through breakfast cereals) may explain some of its broad-ranging micronutrient benefits. In this study, all girls with low dairy intakes failed to meet the EAR for calcium. This is of particular concern during adolescence when peak bone mass accrual occurs [28
]. Low dairy intake is also associated with vitamin D deficiency and with the consumption of sugar-sweetened and/or carbonated beverages [31
] which also may adversely affect bone accrual by the pairing low calcium consumption with low vitamin D and high phosphorus intakes [32
Red meat consumption had beneficial effects on adequacy of intake of vitamins B6 and B12 as well as phosphorus and zinc. This is consistent with recent analyses from NHANES in which the consumption of red meat amongst children and adolescents was associated with greater nutrient adequacy for a number of micronutrients [36
]. Among midadolescent girls with low red meat intakes, more than 30% failed to take in adequate amounts of these important B-vitamins.
While the adolescent girls in this study had high intakes of grains, most of these were derived from refined sources. B-vitamins that are lost in the refinement of grains have been replaced through fortification since the 1940s. Whole grains have higher fiber and protein contents than refined grains and are also more important sources of magnesium and other nutrients and phytochemicals such as lignans. Thus, increasing the proportion of total grain intake from whole grain sources is an important target for dietary change during childhood and adolescence.
The role of multivitamin use in micronutrient adequacy is unclear. In the NGHS, 36% of girls (42.7% of whites and 29.1% of blacks) at the end of the follow-up reported taking multivitamins. This is similar to recent data from the 1999–2004 NHANES, showing that 34.2% of children and adolescents took a multivitamin/mineral supplement (although only half of those did so daily) [37
]. Multivitamin consumers tended to have better nutrition and healthier lifestyles overall suggesting that those who might benefit most from supplementation were least likely to take vitamins. In addition, whether multivitamin/mineral supplements provide health benefits that are equal to those of nutrient-rich foods is an open question.
This study has a number of important strengths. It is a relatively large prospective study with dietary data from multiple sets of three-day diet records, yielding more precise estimates of dietary intake. Another important strength is the inclusion of USDA-defined Food Pyramid servings that were derived by the authors by linking NDS food codes with USDA Survey Food codes. These methods for deriving food groups/subgroups included disaggregating all mixed dishes into their component parts for the estimation of food group servings. While dietary intake was assessed using a gold-standard method for large-scale epidemiologic studies, there are still limitations associated with the use of self-reported diet records, particularly at younger ages. At the outset of the study, the girls were 9–10 years of age, an age at which accurate quantification of amounts consumed is difficult. While the children were encouraged to get details of recipes and food preparation from a parent, there is no guarantee that this was consistently done. The absence of such details is a likely source of nondifferential error in the estimation of food and nutrient intakes. All self-reported dietary assessment methods are also prone to bias. Overweight/obese girls may underreport the consumption of foods and nutrients that they believe to be associated with body fat. However, it has been shown that such bias is most correlated with the underreporting of snacks, rather than foods at meals [38
]. Such underreporting could lead to an underascertainment of foods classified as SoFAS. Finally, although the study is larger than most prospective studies, particularly those with detailed diet records, its size is still somewhat limited for examining certain subgroups. For example, since whole grain consumption was quite low, it is not possible to demonstrate the contribution of intake levels that more closely match the Dietary Guideline recommendations.