RESUMO
We investigated effects on oral cancer (OC) risk of an interaction between a single nucleotide polymorphism (SNP) in the alcohol dehydrogenase 3 (ADH3) gene and alcohol consumption levels using a hospital-based study of 93 cases and 99 controls conducted in Athens, Greece. This SNP affects ethanol metabolism in vitro and appeared to interact with alcohol consumption in a previous OC study. We also evaluated a SNP in CYP2E1, another gene involved in ethanol metabolism, reported to be associated with OC risk in a European population. Data on genotypes and risk factors obtained from interviews were analyzed using multivariate logistic regression, accounting for potential confounders. No overall (marginal) association was found between OC risk and ADH3 genotypes. An interaction between ADH3 genotypes and alcohol consumption levels, however, was suggested. In non-drinkers, the ADH3(1-1) genotype has higher risk than ADH3(1-2) or ADH3(2-2) genotypes, but for subjects consuming alcohol, lower risk was observed for ADH3. We fit a logistic regression model to estimate the increase in OC risk associated with each alcohol drink consumed per week. We estimated that OC risk increased by 31.5% per drink/week for the ADH3(2-2) genotype, 4.1% for the ADH3(1-2) genotype and 1.6% for the ADH3(1-1) genotype. Evidence of genotype-environment interaction was suggestive (p = 0.048, Wald chi p = 0.145, likelihood ratio). This finding is opposite to that reported for a population in Puerto Rico, where the ADH3(1-1) genotype seemed more sensitive to ethanol exposure. In Greece, genetic variation at the CYP2E1 SNP is almost entirely absent, with only 1 case and 1 control heterozygous for the variant. By contrast, in a population in France where an OC association was reported, the frequency of CYP2E1 heterozygotes was 5% in controls and 9% in OC cases. These findings illustrate the importance of replicating SNP associations both within and between different racial and ethnic groups and geographic regions.