marker for use in the monitoring of populations after long-term exposure (DeRosa et al., 1993). Of particular value would be markers that improve the attribution of disease endpoints to causes (e.g., allowing determination of which specific mutations in lung cancer tissue are a fingerprint for which environmental cause or which asthma attack is due to an industrial pollutant rather than a natural pollen).
For biomarkers to be useful in characterizing human health effects associated with environmental hazards, they must be validated with human populations (National Research Council, 1989b, 1992a,b). This will generally require a clinical research setting and the use of epidemiologic and multivariable biostatistical methods that adjust for the important potential confounding and effect-modifying factors that could be masking a real environmental effect or causing spuriously positive results. Improved epidemiologic and clinical research methods that can better distinguish truly harmful from harmless environmental exposures in humans, that can detect lower doses, and that require smaller sample sizes need to be developed and validated.
A challenge that faces researchers is the need to link biomarkers to the disease with which they are associated and to determine at what levels disease is induced. Two strategies should be used to determine the link between biomarkers of exposure and an individual's prior exposures. The first strategy, according to Henderson (1998), is physiologically based toxicokinetic modeling. In order to develop such a model, data regarding the rate of formation of a biomarker following exposure to a toxic agent and its rate of removal or repair in the body (e.g., rate of excretion or degradation half-life) must be obtained. Physiologic parameters (e.g., cardiac output or breathing rate), as well as the physical-chemical characteristics of the chemical and its metabolites must be determined. Second, multiple biomarkers can be used to elucidate prior exposures more in depth than what may be obtained with a single biomarker. For example, if both the amount and the half-life of a biomarker vary, this knowledge can be used to give more perspective on a previous exposure. This knowledge will be able to determine whether an individual was recently exposed to a high level of a chemical or was continuously exposed to low levels in the past.
The committee supports the view that one of the most promising ways to accomplish this is by incorporating appropriate biomarkers to improve the accuracy of measurement of exposures, susceptibility factors, or disease outcomes in well-designed epidemiologic studies (Hulka et al., 1990; Schulte and Perera, 1993).
Current research in toxicology, epidemiology, molecular biology, clinical medicine, and social sciences can make important contributions to the study of environmental health but cannot adequately address the range of issues raised by environmental justice. Collaborative approaches to research that incorporate