In order to understand the relevance of biomarkers of biologically effective dose and early biologic effects in a broad public health context, it is critical to understand the biomarker-disease relationship. Although preliminary work can be carried out for some biomarkers in case-control studies, the potential for disease bias will always raise a concern about results from these studies. As such, prospective cohort studies carried out on large, generally healthy populations are the optimal approach to study the relationship between biomarker levels and disease risk. In these studies, biologic samples are collected from a defined population and banked, for later analysis the cohort is followed forward in time; subjects who develop disease are identified, and premorbid biomarker levels in these cases are then compared with those in unaffected subjects. Often, a nested case-control approach is used. Samples from cases and only a sample of the controls are analyzed, which considerably reduces the laboratory requirements and costs. These studies allow calculation of the relative risk, which provides an estimate of the increased risk in disease for a given level of a biomarker, and the etiologic fraction, which estimates the proportion of disease reflected by the biomarker. The consistent results across several, small cohort studies showing a positive relationship between chromosomal aberration levels in peripheral lymphocytes and subsequent cancer incidence or mortality suggest that at least some biomarkers of early biologic effect will prove to be relevant for predicting cancer risk. There win soon be over a minion subjects with blood samples banked in cohort studies in the U.S., Europe, and Asia. These studies will provide striking opportunities for the evaluation of the relationship between a wide range of biomarkers and disease risk over the coming years.
The Toxicologist. Society of Toxicology 38th Annual Meeting, March 14-18, 1999, New Orleans, Louisiana