Design and analysis issues in case-control studies addressing genetic susceptibility

Case-control studies are among the main study designs for investigating the effect of environmental exposures on disease etiology and are now becoming more frequently used to examine the influence of genetic susceptibility. Incorporating a case-control design to examine genetic exposure would seem appropriate as the main problems of case-control studies, those of bias and confounding, are more easily avoided. However, they do not have a strong image and are thought to provide too many spurious findings as a result of chance, bias or confounding. This criticism has often been valid, especially for many early studies based on small numbers of poorly defined cases that were compared to convenient groups of local controls. New family-based study designs that account for these problems have been suggested, including the transmission disequilibrium test. However, the traditional format of case-control studies may be improved by incorporating methodological refinements developed over the past 30 years in environmental epidemiology. The present chapter attempts to outline these issues and to provide guidelines for the selection of cases and controls and for the analysis of genetic data from case-control studies. Alternative, family-based study designs will also be discussed. Finally, a comprehensive analysis approach to gene-environment and gene-gene interaction is proposed.