Pooled analysis of epidemiological studies involving biological markers.

Pooled analysis is a method frequently used in epidemiology when individual studies are too small to allow any definite conclusion. Several guidelines have been published on pooling of classical epidemiological studies, but no information is available on pooling of studies involving biological markers. We have used information of two recently started pooled analyses, one on cytogenetic damage, the other on genetic susceptibility to environmental carcinogens, in order to make inferences on the applicability of a pooled analysis to molecular epidemiological studies. Issues in pooling data from epidemiological studies involving molecular markers are described here, including the choice of study design, planning and conducting of the study, data request, collection and storage, and costs. Some practical indications on the conduct of a pooled analysis in molecular epidemiology is given.