We will examine whether very early life dietary patterns (age 0-3) influence childhood diet (age 5), and whether these patterns affect physiological risk profiles and cognitive performance in middle adolescence. If early life diet confers a risk for poorer physiological profile and cognitive ability in adolescence, then it is important to understand the parent characteristics and practices that influence early life diet quality (see figure 1 in appendix). These parenting factors should then become the targets of intervention in the period from birth to preschool. We have chosen ages 0-3 because it is in this period that the dietary patterns that may track through childhood start to emerge. Parents are also primarily in control of what children consume, with little external influence. This period may present an opportunity to provide more specific dietary advice to parents of infants and toddlers at a 'critical period' for later health and development. ALSPAC can help provide evidence from a high quality cohort on the effect of transition diet on later physiological risk and cognitive development. We will address the following objectives:

1 Characterise and compare methods of dietary characterisation across 0-3 years using principal components, cluster analysis, diet quality scores and single food/nutrient-based approaches

2 Evaluate whether dietary patterns across 0-3 years are related to dietary patterns at 5 years

3 Examine the relationship between different methods of dietary characterisation across 0-3 years and risk factors for chronic disease (obesity, blood pressure, lipids, insulin) at 15 years

4 Examine the relationship between different methods of dietary characterisation across 0-3 years and neurocognitive development and school attainment at 15 years

5 Identify parenting practices and characteristics of parent lifestyle and family structures that predict dietary patterns across 0-3 years.

This proposal is focussed on the role of early life diet on physiological risk factors for chronic disease (including adiposity) and cognitive ability in adolescence, both of which have implications for adult health and wellbeing (1, 2). Diet can be modified to improve disease risk factors in adults (3, 4), but it's not known if diet in the 0-3 year period matters either physiologically or developmentally for later health. However, given that prolonged breast feeding is associated with improved cognitive performance (5) it is plausible that other aspects of early life nutrition may be associated with later physiological and cognitive risk. Additionally, this is a period when parents are motivated to provide a good diet and may reflect a critical window when physiological and developmental risk profiles are 'programmed' with lasting effects into adult life (6).

1. Characterising the emergence of diet from birth to 3 years

In order to look at the association between diet at 0-3 years and later outcomes, we need to be able to characterise diet across this period. Diet is a complex mix of nutrients within a matrix of foods, which interact (7). The combined effect of nutrients and foods can be assessed by looking at dietary patterns, which may be useful for understanding the 'whole-of-diet' influence on disease. A number of methods are available to characterise dietary patterns (7). For example, cluster analysis divides individuals into different groups based on a particular characteristic (e.g. consumption of a food or nutrient). Whereas in factor analysis, individuals' diets are scored according to underlying patterns or 'factors' that explain variation in eating patterns. Neither method is designed to derive patterns that are predictive of disease, which is where diet indices or techniques such as reduced rank regression may be useful. Diet indices also enable assessment of intake against dietary recommendations which aim to promote health. By utilising these methods, the strengths of each method and its specific purpose can be exploited.

Most research involving dietary patterns has been conducted in older children and adults. Dietary research in the 0-3 year period has been dominated by studies on breastfeeding (8-12), weaning or toddler diet quality (9, 13-20), using largely a nutrient (14, 18-21), variety (13, 16) or food group (9, 14, 15, 17, 18) approach. However principal components analysis (PCA) has been used to describe dietary patterns in ALSPAC children between 3 and 9 years (22) and in infancy in the Southampton Women's Survey (23). As a tool, PCA can discern patterns associated with stage of development at 3 years (e.g. 'snacky') but also patterns similar to those characterised in later childhood and adult life (e.g. 'healthy/ prudent' 'junk') (22). Few attempts have been made to characterise dietary patterns, across the 0-3 year period which spans the transition from a milk- to food-based diet. Early life diet has also been characterised using an index based on how well parents adhere to guidelines, either at one point in time (24, 25) or across the infant to toddler diet transition (26). How well these or other methods relate to later diet or longer-term outcomes has not been established.

We plan to compare dietary patterns across 0 to 3 years characterised by cluster analysis, PCA and dietary indices (based on UK dietary guidelines) in the ALSPAC sample. In addition to enabling a direct comparison of methodology in the same sample, we will explore the predictive validity of these methods on physiological and cognitive outcomes (points 2-5 listed below). To our knowledge, comparisons between dietary pattern methods are limited, particularly in child samples, but would be useful in increasing the understanding of when to utilise various dietary pattern methodology.

2. Evaluate whether dietary patterns at 0-3 years are related to dietary patterns at 5 years

The ALSPAC cohort provides evidence that dietary patterns can be characterised from the age of 3 years and that these patterns track to some extent through middle childhood (22). This evidence is consistent with other cohort studies which show intake of nutrients (27, 28), foods (29, 30), food patterns (30) and food preferences (31) track across childhood , adolescence and into adult life (over 6 to 21 years of follow up). The rapid change in diet and learning about eating that occurs between birth and the toddler years may lay the foundation for future eating habits, food preferences and intake patterns (32). Part of our investigation will reveal whether diet in the 0-3 year period is related to diet at 5.

3. Early life diet and risk factors for chronic disease at 15 years

Risk factors for chronic disease, such as high blood pressure, serum lipids and adiposity show evidence of tracking from adolescence through adulthood (1, 33-36). This indicates that measures of risk during adolescence are a reasonable proxy for ascertaining life-long risk of chronic disease. Studies in adults have shown that diet can modify these risk factors and there is preliminary evidence that introducing weaning foods low in saturated fats can lead to reduced serum lipids by age 14 (37). Previous ALSPAC analysis has identified associations between fat and sodium intake in the weaning period, serum cholesterol at 3 years (38) and blood pressure at 7 years (21). Dietary patterns in ALSPAC childhood between 3 and 9 years have also been associated with later obesity (39). This proposal will extend previous ALSPAC analyses by looking at whole-diet across the period 0-3 years on adiposity and physiological risk factors in adolescence.

4. Early life diet and cognitive development and educational attainment at 15 years

Deficiency of specific nutrients (such as iron or iodine) in the postnatal period has permanent effects on cognitive development (40, 41). Rather than studying effects of nutritional deficiencies, emerging research is revealing dietary patterns associated with 'optimal' development. Data from ALSPAC shows that dietary patterns at 3 and 4 years are associated with later school attainment and hyperactivity (42, 43). Similarly, dietary patterns during infancy, when the brain is developing rapidly, suggests that a weaning diet rich in fruit, vegetables and home prepared foods benefit cognitive development at age 4 (44). This finding contradicts feeding guidelines, which emphasise iron-rich weaning foods at 6 months because breast milk is limited in iron. The ALSPAC cohort can help tease out the issue of whether diet between 0-3 years has longer-term effects on cognitive development.

5. The influence of parenting practices and characteristics on early life dietary patterns

While associations between socio-economic position, parental education, age, lifestyle or weight status and early life dietary patterns have been documented (23, 45), the mechanisms underpinning these relationships is less clear. For example, parenting- style, -practices and perceptions of children's diet are likely to influence early life diet quality. Parent decisions that affect the child's diet include the mode and duration of breastfeeding (8, 45), the age of solids introduction (46) as well as the variety, quantity and textures of foods provided (47). Interactions between the parent and child during feeding such as the responsiveness to appetite cues, repeated exposure, controlling and coercive practices, may also influence child diet (32). Whether family structure (e.g. family size, partner status, employment and domestic support) influences early life diet is unknown. There has been little large scale, prospective research in general populations on the role of parent characteristics and practices which shape children's dietary patterns. ALSPAC provides the opportunity to look at this comprehensively.

To date, analysis of ALSPAC data has included dietary assessments from age three or older and some outcomes in middle childhood. Our plan is to evaluate the effect of the earliest dietary exposure on adolescent outcomes.The period between 0-3 years is critical for establishing healthy growth in an under-nourished population (6), it is not known whether the same is true of children from developed countries. Our study will help determine whether dietary interventions should begin in pregnancy and continue through very early in life as a strategy for optimising adult outcomes.