Background: There is increasing evidence that childhood cancer is more likely to occur in children that were relatively large at birth, and unpublished information indicates that this is particularly true of non-leukaemic cancers occurring after 3 years of age [1]. The risk is lowest among children born small-for-dates and tends to increase linearly with increasing birth-weight. The evidence indicates that the birth-weight association with leukaemia relates to excessive growth in utero [2,3]. This raises the question as to whether there are features that increase the growth of the fetus that are responsible for increases in the risk of cancer. Although many high quality studies have focused on the aetiology of growth retardation, little attention has been paid to the high end of the birth-weight distribution.

Hypothesis and objectives: This study is designed to identify factors that influence fetal growth in order to develop coherent hypotheses as to ways in which the unexplained link with childhood cancer may be explained. We hypothesise that features of the environment including diet and lifestyle, genetics and epigenetics may play important parts. We have already shown an excess in birth-weight if the non-smoking mother was exposed in utero to her own grandmother's smoke [4], which indicates a possible epigenetic influence; additionally our unpublished data have shown an association with maternal prenatal intake of paracetamol (acetaminophen). Thus our hypotheses concern the likely complexity of factors that influence excessive fetal growth.

The objectives are to determine features of the life-course of each parent, and grandparent, to determine associations with increased fetal growth using an 'exposome' approach [as in B2190]. Where appropriate, linkage to genetic variants and methylation patterns will be used to determine biological pathways which may be associated with childhood cancer.