5356

We evaluated secondhand smoke exposure and the risk of childhood cancer and the interaction between smoking and CYP1A1 haplotypes in a multi-center hospital-based case-control study in Seoul, Korea. Histologically confirmed childhood leukemia (n=166), brain cancer (n=62) and lymphoma (n=44) cases aged from 0 to 18 were recruited from three teaching hospitals between April 2003 and May 2005. Non-cancer controls (n=298) recruited from the same hospitals were frequency matched to cases by gender and age (+2 years). Information on demographics, and parental smoking exposures were collected by trained interviewers using a structured questionnaire. Five CYP1A1 SNPs (rs2472299 (C>T), -9893G>A, rs1048943 (I462V, A>G), rs4646903 (T>C), and rs2198843 (C>G)) were genotyped using a single base extension assay. Haplotypes were estimated by Bayesian methods using the PHASE program (ver. 2.0.2). The data were analyzed by multiple logistic regression adjusting for age, birth weight and paternal education. Paternal smoking at home increased the risk of leukemia by 1.7-fold (95% CI=1.04-2.66). Overall, CYP1A1 haplotype did not increase the risk of childhood leukemia. However, there was a significant association between CYP1A1 haplotype (i.e., other than CGACC containing diplotypes) and childhood leukemia in children whose fathers were smokers (>400 cigarettes/lifetime) (OR=1.9, 95% CI=1.03-3.39) and the risk increased 2.3 fold if the father smoked at home (OR=2.3, 95% CI=1.03-5.32). Overall, CGATC containing diplotypes were associated with an increased risk of brain cancer (OR=2.8, 95% CI=7.41). Although based on small numbers, the association was strongest among children whose father was a smoker, smoked at home, and smoked at home at the presence of the child (OR=6.1, 95% CI=1.56-24.0; OR=13.3, 95% CI=1.49-119; and OR=14.6, 95% CI=0.88-241, respectively). Multiplicative interactive effect on brain cancer was observed between CGATC containing diplotypes and paternal smoking status (P-for interaction=0.01). These results suggest that parental smoking is associated with an increased risk of childhood leukemia and brain cancer and that CYP1A1 genotype may modify the association between these childhood cancers and environmental exposure to cigarette smoke.

[Proc Amer Assoc Cancer Res, Volume 47, 2006]