Introduction
Peak bone mass is reached during young adulthood and is influenced by genetics as well as modifiable lifestyle factors, including nutrition [
1,
2]. Hence, nutrition during early life may have important consequences for peak bone mass attainment and future fracture risk. Previous studies on childhood nutrition and bone health have mainly focused on individual nutrients, including calcium and vitamin D, or specific foods, such as dairy products, fruits, and vegetables [
3‐
6]. However, foods and nutrients can have high level of intercorrelation or interaction, which is not fully considered when examining their individual effects on health outcomes. Evaluating dietary patterns, which include the combination of dietary factors, may better overcome confounding by other dietary factors and account for the cumulative and interactive effects of multiple nutrients [
7]. Previous studies in adults have examined the associations between dietary patterns and bone health. These studies reported that so-called healthy, prudent, or nutrient-dense dietary patterns (i.e., rich in fruits, vegetables, grains, fish, and low in meat) may positively affect bone health in adults [
8], women [
9‐
11], and elderly [
12], whereas dietary patterns rich in refined or energy-dense foods (i.e., meats, fried foods, soft drinks, confectionery) appear to be inversely associated with bone health [
8,
10‐
12].
Since a significant proportion of adult bone mass is accrued during childhood and adolescence, gaining an understanding of the role of dietary components on bone mass in early life is important to identify interventions favorable to bone health. Nevertheless, except for a few previous [
13‐
16] studies in (small groups of) children or adolescents, not much is known on the relation between dietary patterns and bone mass, particularly in young children. Therefore, we investigated the associations between dietary patterns assessed in infancy and bone health (bone mineral density, bone mineral content, and bone area) at the age of 6 years in 2850 children participating in a population-based prospective cohort study. In addition, we aimed to evaluate whether the associations differ by the child’s sex, birth weight, gestational age at birth, ethnicity, body mass index, or vitamin supplement use.
Discussion
In this large population-based cohort study from early life onward, we observed that children with a dietary pattern characterized by high intakes of dairy and cheese, whole grains, and eggs, during early infancy, have a higher BMD during childhood. These associations were only observed in children who did not receive vitamin D supplementation. We found no evidence for consistent associations between a “potatoes, rice, and vegetables” pattern and bone mass in our population of school-age children. Furthermore, the associations between a “refined grains and confectionery” pattern and bone mass appeared to be explained by confounding factors.
Calcium, phosphorus, and protein are important components of bone; hence, adequate intake of these nutrients is required for normal bone development. Other nutrients, including vitamin D, vitamin K, magnesium, zinc, and fluoride, are also involved in bone metabolism [
2,
27]. Many of these nutrients are intercorrelated and act in a synergistic way. For example, vitamin D is involved in the absorption of calcium, and magnesium plays an important role in vitamin D and calcium metabolism [
1,
2]. Moreover, whereas the effects of single nutrients are often small, evaluating dietary patterns may enable detection of the cumulative effects of nutrients and nutrient interactions on health outcomes [
7].
Previous studies in adults observed that dietary patterns rich in refined, energy-dense foods were inversely associated with bone health [
8,
10‐
12]. We did not observe associations for the “refined grains and confectionery” pattern with bone mass. Also, no associations were observed between the “potatoes, rice, and vegetables” pattern and bone health. A few previous studies have examined dietary patterns in relation to bone health in children or adolescence, indicating that nutrition in early life or in adolescence may influence bone health. In 325 American children followed from age 4 to age 8 years, two dietary patterns were identified that were both positively associated with longitudinal bone mass development. The first pattern was rich in refined grains, processed meat, cheese, eggs, fried potatoes, and sweetened beverages, while the second pattern was characterized by high intakes of whole grains, processed meats, vegetables, and fruits [
16]. In a cross-sectional study in 196 Korean adolescents aged 12–15 years, high adherence to a “milk and cereal” pattern was positively associated with lumbar spine BMD, whereas the “traditional Korean,” “fast food,” and “snacks” dietary patterns were not associated with bone mass [
15]. A study in Dutch adolescents aged 9–15 years showed that those who previously followed a macrobiotic diet, characterized by high intakes of cereals, pulses, and vegetables and low intakes of meat, chicken, and dairy, had a lower relative bone mass than subjects who followed an omnivorous diet [
14]. Finally, in 559 children of mothers participating in the Southampton Women’s Survey, an “infant feeding guidelines” pattern in infancy, characterized by high intakes of vegetables, fruits, meat/fish, home-prepared foods, and breast milk, was not associated with bone mass at the age of 4 years [
13]. Other evidence that early life nutritional factors may influence childhood bone health come from studies on maternal diet [
28,
29]. A previous study also embedded in the Generation R Study observed that maternal intakes of protein, phosphorus, and calcium and blood concentrations of vitamin B
12 in pregnancy were positively related to childhood bone mass, whereas maternal carbohydrate intake and blood concentrations of homocysteine were inversely associated with childhood bone mass [
20].
In the present study, a “dairy and whole grains” pattern in infancy was associated with higher BMD and aBMC in childhood. Parameter estimates were small, but significant. This dietary pattern was characterized by high intakes of whole grains, dairy and cheese, and eggs, which together contain a number of ingredients that may have beneficial effects on bone outcomes. Key nutrients supplied by dairy foods are calcium, magnesium, vitamin D (especially if fortified), and high-quality proteins. Whole grain products contain magnesium, iron, B vitamins, and other bioactive compounds (i.e., phytochemicals, antioxidants), which may benefit bone health [
2,
27]. It goes beyond the aim of the present study to disentangle the relative individual contributions of different dietary components. Nevertheless, we performed exploratory analyses to evaluate the associations for specific foods/nutrients (i.e., dairy products, whole grains, eggs, calcium, vitamin D, B vitamins) with bone outcomes and observed that these associations were not consistent and/or smaller in magnitude, as compared to the associations for the “dairy and whole grains” pattern (results not shown). This suggests that the associations between the “dairy and whole grains” pattern and bone outcomes are largely driven by the combination of food groups and nutrients that make part of this pattern as a whole. Nevertheless, further research is needed to confirm this.
We observed that children with the highest adherence to the “dairy and whole grains” pattern had on average higher relative intakes of total protein, animal protein, calcium, phosphorus, magnesium, and vitamin B
12. These are nutrients that have been positively related to bone health in children and/or adults [
2,
27,
30]. In contrast, these children had a lower mean intake of vitamin D, as compared to children in the highest quartiles of the other two dietary patterns. In the Netherlands, most dairy products are not fortified with vitamin D. Hence, other foods such as fish, meat, margarines, baking fats, and infant formula might be more important sources for children. In addition, since the year 2000, the use of vitamin D supplements has been recommended for children up to the age of 4 years. We observed that vitamin D supplements were used more often in children with a high adherence to the “dairy and whole grains” pattern, which may counterbalance their relatively lower dietary intake of vitamin D. Interestingly, however, we observed that the beneficial effect on bone health of the “dairy and whole grains” pattern was more explicit in the children who did not receive vitamin D supplementation. This could suggest that the importance of the consumption of specific dietary factors diminishes in case of vitamin D supplementation or, alternatively, that the consumption of dairy and whole grain products may compensate for the absence of vitamin D supplementation.
We observed a significant interaction between the “potatoes, rice, and vegetables” pattern and ethnicity in the analyses with BMD. Differential effects of dietary patterns on bone may be related to racial/ethnic differences in the metabolism of specific nutrients, such as calcium and vitamin D [
31]. Furthermore, risk factors associated with BMD and fracture risk may be ethnic-specific [
32], suggesting that specific dietary exposures might beneficially affect bone development in some ethnic groups, but not in others.
Next to providing specific nutrients, the bone-promoting effect of the “dairy and whole grains” pattern may also act through insulin growth factor 1 (IGF-1). Cow’s milk is known to increase IGF-1 levels [
33] (reviewed by [
34]), which stimulates bone mass accrual [
34]. Furthermore, dietary patterns could also indirectly affect bone mass, through changes in fat and/or lean mass that are driven by the diet. For example, body weight/body fat is a known factor that is positively related to bone mass [
1,
35]. To take this into account, we adjusted all our analyses for total energy intake and child anthropometrics to obtain estimates that were uncorrelated with body size or energy intake.
In analyses stratified for vitamin D supplementation, we observed positive associations for the “dairy and whole grains” pattern with aBMC, but inverse associations with bone area. Further investigations using 3D assessments (e.g., quantitative CT) are needed to determine whether these findings are arising from differences in skeletal frame size and/or bone geometry.
The observed effect estimates are small and may not reflect individual, clinically relevant, differences. Nevertheless, the present study provides insight into mechanisms linking early life nutrition to bone acquisition during childhood. Prospective studies with follow-up into adulthood can determine whether such dietary exposures during childhood can be translated to risk differences later in life.
Methodological considerations
An important strength of this study is the population-based cohort design from early life onward, with detailed measurements of dietary intake and bone mass in a large number of children. We collected information on many potential confounding variables. However, as in any observational study, residual confounding due to unmeasured variables such as physical activity levels might still be an issue.
Information on bone health was available for 80 % of all singleton children with information on dietary intake. Children who participated in the DXA measurement at the age of 6 years on average had a lower total energy intake and a lower adherence score to the “refined grains and confectionery” pattern compared with children who did not participate. Selection bias due to selective loss to follow-up is of concern if the associations between dietary patterns and bone health differ between those included and not included in the study. Although this seems unlikely, it cannot be excluded.
Dietary patterns were assessed at the approximate age of 14 months. Dietary habits may change during childhood, and these dietary changes could also affect childhood bone health. Unfortunately, no further data on childhood diet was available to examine this. Previous studies in (young) children have shown evidence of tracking of dietary habits [
36‐
39], which suggests that information on infant nutrition may be a reasonable indicator of childhood diet. Nevertheless, more research on tracking of dietary intakes is needed to confirm this.
Dietary pattern analysis involves several decisions that may influence the findings, such as the composition of food groups, numbers of factors to extract, and labeling of the components [
7]. Nevertheless, we used common criteria to select the factors. Although the amount of variance (30 %) explained by the extracted dietary patterns was small, it is comparable with previous studies [
10,
15,
36].
Acknowledgments
EHH, JCKJ, and OHF work in ErasmusAGE, a center for aging research across the life course funded by Nestlé Nutrition (Nestec Ltd.), Metagenics Inc., and AXA. Nestlé Nutrition (Nestec Ltd.), Metagenics Inc., and AXA had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript. The Generation R Study is conducted by the Erasmus Medical Center Rotterdam in close collaboration with the School of Law and Faculty of Social Sciences of the Erasmus University Rotterdam, the Municipal Health Service Rotterdam area, the Rotterdam Homecare Foundation, and the Stichting Trombosedienst & Artsenlaboratorium Rijnmond, Rotterdam. The general design of the Generation R Study is made possible by financial support from the Erasmus University Medical Center Rotterdam, the Erasmus University Rotterdam, the Netherlands Organization for Health Research and Development (ZonMw), the Netherlands Organisation for Scientific Research (NWO), and the Ministry of Health, Welfare and Sport. We gratefully acknowledge all participants and the contribution of general practitioners, hospitals, midwives, and pharmacies in Rotterdam. Acquisition of nutritional data was funded by an unrestricted grant from Europe Container Terminals received by Prof. Dr. H.A. Moll. Dr. Jaddoe and Dr. Rivadeneira received additional grants from the Netherlands Organization for Health Research and Development (VIDI 016.136.361 and VIDI 016.136.367, respectively).