Selenium and Human Cancer: Epidemiological Aspects and Implications for Clinical Trials

ATES OF CANCER at most sites vary geographically; many are inversely associated with estimates R of selenium intake. These descriptive studies (also called “ecologic” or “correlational” studies), reviewed by Jansson,(l) have provided an important stimulus to research. However, since other dietary or nondietary factors that differ between these geographic areas could provide an alternative explanation for the association with selenium, firm conclusions cannot be drawn from such studies. In general, countries with high selenium intake and low rates of many specific cancers tend to be developing countries that typically have, for example, low fat intake. Indeed, the scatterplot of selenium intake vs national breast cancer mortality ratell) is almost the mirror image of that relating fat intake to mortality due to this neoplasm.13) These descriptive studies of groups can be especially valuable under unusual circumstances, such as when comparisons are made within countries under particularly uniform conditions, as in the study being conducted in China by Chen et a1.I‘) An alternative to observing cancer rates in populations is to examine the occurrence of this disease in individuals. Such studies, either case-control or cohort in design, usually offer greater opportunity to control for potentially confounding factors. This article reviews the methodological difficulties encountered conducting studies of selenium and cancer among individual humans, examines available data regarding this relationship from case-control and cohort studies, and discusses some of the implications of these findings for clinical trials.