COMPUTING DIFFERENTIAL SAMPLE SIZE FOR CASE-CONTROL STUDIES OF GENE-ENVIRONMENT INTERACTION

The rates for diseases such as cancer, cardiovascular disease, and diabetes are known to differ by ethnic/racial groups. However, neither genetic nor environmental factors fully explain the observed differences. Failure to account for genetic expression in the absence or presence of an environmental factor, and vice-versa, may lead to erroneous conclusions regarding the importance of these factors in disease etiology. We present a novel method for computing sample size for case-control studies involving the interaction of genetic and environmental factors. The method is based on an indirect estimate of the odds ratio for geneenvironment interaction given only the odds ratio for environmental exposure and population genotype frequency. A table is presented providing sample sizes required for detecting a minimum odds ratio for gene-environment interaction given varying genotype frequencies and environmental exposure odds ratio values. Sample size increases proportionately with genotype frequency for a given environment exposure odds ratio. (Ethn Dis. 2008;18[Suppl 2]:S2-25–S2-29)