Case studies in the use of toxicological measures in epidemiological studies.

One of the major limitations in environmental epidemiological studies investigating associations between chemical exposures and the risk of disease is the use of crude exposure metrics. This increases the risk of exposure misclassification, reducing the chance of finding a real association, where one exists. In environmental epidemiological studies, surrogate exposure measures, such as time living in a contaminated area, do not adequately reflect the true nature and/or degree of real exposure to the chemical substance(s) of interest. Also, as they are not quantitative measures of absorbed dose, the data provided by such measures cannot give numerical measures of a dose response relationship, limiting their use in quantitative risk assessment. In epidemiological studies, biological measures of exposure should be used, if available, as they are close to the target organ dose and provide greater precision in risk estimates and dose response relationships. Examples are the measurement of speciated arsenic in the urine of subjects when investigating health effects from arsenic exposure, and the use of breast milk lipid concentration to measure body burden when conducting epidemiological studies of health effects from persistent lipophilic compounds, such as the endocrine disruptors. This paper discusses the main uses of such exposure measures, using several case studies, their limitations and some challenges in the future for toxicologists to develop more suitable measures for epidemiologists to use in population studies.