Toxicological considerations in the application and interpretation of susceptibility biomarkers in epidemiological studies.

With major advances in genotyping technology, it has become practical and affordable to screen biological samples for multiple polymorphisms for which there is more or less knowledge about their functional relevance in relation to toxicological agents or disease. This situation creates some unique epidemiological challenges in which careful consideration of the mechanisms underlying genotype-disease and genotype-exposure-disease relationships, or the lack of knowledge thereof, may help to prevent false-positive results or misinterpretation of data that could mislead future research. Coordination and linkage of data resources on toxicological exposures, genes and disease would be useful as an integrated source of information to devise study design, analysis and interpretation. Such a database could be useful in collating information from which to determine relevant exposures for the disease and to guide the selection of susceptibility markers for study. Further information on toxicological mechanisms and functional relevance of gene variants in relation to disease could be accounted for in statistical analyses and interpretation. Statistical methods that can incorporate this prior information on mechanism, such as hierarchical regression modelling, may help to mitigate some of the problems inherent to these studies by adjusting estimates and confidence limits according to this prior information. Even in the situation where little information on etiological mechanisms is available, application of a less informed prior can be beneficial in improving the accuracy and precision for an ensemble of estimates. The changing paradigm of epidemiological research with regard to increasing detail of underlying mechanisms and the sheer amount of data being evaluated necessitates access to sources of information and analytic methods that can integrate this complexity.