The relationships of infant and childhood diet to growth and acute lymphoblastic leukemia

Diet during infancy and childhood can affect growth, onset of puberty and disease susceptibility throughout the life course. The goal of this research was to describe the associations of early life diet (birth – five years of age) with hormone levels and pubertal development in healthy adolescent females and with the risk of pediatric acute lymphoblastic leukemia (ALL), the most common form of pediatric cancer, in a population of boys and girls. Chapter 1 summarizes what is known about the early life diet and childhood growth, hormone levels and risk of acute lymphoblastic leukemia. It identifies gaps in the literature which led to the research described in this dissertation. Chapter 2 discusses findings on the effects of early life diet on serum insulin like growth factor-1 levels and breast development in healthy adolescent females. Child’s weekly dairy consumption from 3-5 years was inversely associated with the odds ratio of thelarche whereas child’s weight and maternal overweight during pregnancy were positively associated with the odds ratio of thelarche at 10.8 years. Chapter 3 discusses identification of two novel risk factors for pediatric ALL: longer duration of milk formula feeding and later introduction of solids foods. Chapter 4 describes research into windows of susceptibility for solid food introduction in ALL. Compared to children introduced to solids before 6 months of age, children introduced to solid foods at or after 7 months of age are at increased odds of ALL with a dose-response relationship between age at introduction to solids and the odds ratio of ALL. Potential mechanisms for the associations reported in chapters 3 & 4 are discussed. Chapter 5 details the association of age- and sex-adjusted height and weight at time of diagnosis with the odds ratio of ALL. The relationship of height at diagnosis to ALL is unclear, owing in part to inconsistencies in study methodology. Using a population of matched controls, we report no association of height at diagnosis with ALL. Children with low weight-for-age or weight-for-height were at increased odds of ALL. Finally, chapter 6 summarizes these findings and discusses their public health implications.