Parental smoking and the risk of childhood leukemia

Cigarette smoke has been linked to adult myeloid leukemia; however, studies of parental smoking and childhood leukemia have been inconclusive. This dissertation assessed the role of parental smoking and child's genetic profile in the development of childhood leukemia using data from the Northern California Childhood Leukemia Study (NCCLS).; Maternal smoking was not associated with an increased or a decreased risk of childhood leukemia. Paternal preconception smoking was significantly associated with an increased risk for acute myeloid leukemia (AML) (odds ratio [OR] = 3.84, 95% confidence interval [CI]: 1.04, 14.17); an increased risk for acute lymphoblastic leukemia (ALL) was suggestive for paternal preconception smoking (OR = 1.32, 95% CI: 0.86, 2.04). An even greater risk of ALL was observed, when paternal preconception smoking was combined with maternal postnatal smoking (p for interaction = 0.004).; The NCCLS paternal smoking data were collected from the mothers. A study with 58 spouse pairs from the NCCLS showed that the agreement was good for paternal smoking and alcohol use during the three months before mother's pregnancy with a kappa of 0.62 and 0.78, respectively.; In addition to parental smoking, this dissertation also assessed the influence of child's genotypes in the association between maternal smoking during pregnancy and ALL. The results were suggestive of a positive joint influence between maternal smoking during pregnancy and GSTT1 null variant or GSTP1 1578 variant in the risk of ALL among non-Hispanic White children and a positive interaction between maternal smoking during pregnancy and MPO -463 wild type among all ALL cases.; Case-only analysis has been proposed as a more statistically powerful alternative to case-control analysis in assessing gene-environment interaction when the distributions of a gene and an environmental exposure are independent. Data on maternal smoking during pregnancy and child's genotypes of metabolizing genes collected by the NCCLS were used to compare case-control and case-only analyses. Results showed that adjusting for covariates (race/ethnicity, household income, maternal education, and maternal age) hypothesized to cause non-independence between maternal smoking and child's genotypes did not consistently correct for the differences between the results of the case-control and case-only analyses.