PURPOSE

This study explored interactions between prenatal exposure to maternal smoking and polymorphisms in metabolic genes in the risk of childhood acute leukemia (AL).

METHODS

The data were generated by the ESCALE study, which included 764 AL cases and 1,681 controls in 2003-2004. The data on maternal smoking during pregnancy were obtained by standardized telephone interview of the cases' and controls' mothers. The genotypes CYP1A12A/2B (rs4646903), CYP2E15 (rs2031920, rs3813867), NQO12 (rs1800566), NAT25 (rs1801280), and EPHX1 exon 3 (rs1051740) and exon 4 (rs2234922) were obtained using a high-throughput platform and imputation for untyped polymorphisms. The analyses were restricted to the 493 cases (433 cases of lymphoblastic (ALL) and 51 of myeloblastic (AML) leukemia) and 441 controls with at least 2 grandparents born in Europe, who were genotyped with individual call rates greater than 95%. Odds ratios were estimated by logistic regression in case-control analyses and, for gene-gene and gene-environment interactions, by case-only analyses.

RESULTS

ALL and AML were not associated with either maternal smoking during pregnancy or candidate polymorphisms in CYP1A1, CYP2E1, EPHX1, and NQO1. Carrying two NAT2*5 alleles was significantly associated with ALL (OR = 1.8 [1.3-2.5]). The analyses also suggested an interaction between three genes involved in benzene metabolism CYP2E1, NQO1, and EPHX1. There was no interaction between maternal smoking and any of the polymorphisms under study.

CONCLUSIONS

The ESCALE study did not evidence the interaction between CYP1A12A/2B and maternal smoking suggested previously. The association with NAT25 and the gene-gene interactions need to be replicated.