3
Exposure Assessment in Environmental Epidemiology

Exposure to chemical and physical agents in the environment can produce a wide range of adverse health consequences. Environmental epidemiology attempts to determine whether a hazard exists—that is, whether there is a causal relation between exposure to certain chemical or physical agents and adverse health effects—and to measure and characterize any causal relations (to assess the exposure-response relationship). Typically, a continuum between level of exposure and the size or probability of health effects is assumed. Emphasis is placed on characterizing the associations across the continuum, and quantitatively defining the relation is a central feature of the epidemiologic investigation. Assessment of exposure is then a crucial component of environmental epidemiologic research. The estimation of exposure in relation to health effects is frequently difficult, and it has generally received inadequate attention. However, a field of exposure assessment is emerging.

Exposure assessment for purposes of environmental epidemiology may differ from exposure assessment for site remediation, mitigation, control, and risk assessment. The differences are sometimes subtle but may have substantial impact on the conduct of studies and associated allocation of resources. Investigations for the purpose of risk assessment, for example, generally include information on the source and identity of chemical agents, the concentration of each toxicant in various media, and the toxicity of identified toxicants as defined in experimental studies. Mathematical modeling may be used to define breakdown, transport, and ultimate location as well as the potential health risk. Environmental epidemiology, on the other hand, is more often hypothesis-based research



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 26
--> 3 Exposure Assessment in Environmental Epidemiology Exposure to chemical and physical agents in the environment can produce a wide range of adverse health consequences. Environmental epidemiology attempts to determine whether a hazard exists—that is, whether there is a causal relation between exposure to certain chemical or physical agents and adverse health effects—and to measure and characterize any causal relations (to assess the exposure-response relationship). Typically, a continuum between level of exposure and the size or probability of health effects is assumed. Emphasis is placed on characterizing the associations across the continuum, and quantitatively defining the relation is a central feature of the epidemiologic investigation. Assessment of exposure is then a crucial component of environmental epidemiologic research. The estimation of exposure in relation to health effects is frequently difficult, and it has generally received inadequate attention. However, a field of exposure assessment is emerging. Exposure assessment for purposes of environmental epidemiology may differ from exposure assessment for site remediation, mitigation, control, and risk assessment. The differences are sometimes subtle but may have substantial impact on the conduct of studies and associated allocation of resources. Investigations for the purpose of risk assessment, for example, generally include information on the source and identity of chemical agents, the concentration of each toxicant in various media, and the toxicity of identified toxicants as defined in experimental studies. Mathematical modeling may be used to define breakdown, transport, and ultimate location as well as the potential health risk. Environmental epidemiology, on the other hand, is more often hypothesis-based research

OCR for page 26
--> that seeks to examine specific populations or communities to clarify the relation between health and physical, biologic, and chemical factors. Volume 1 of this report (NRC, 1991a) described some limitations and problems in the quantitative estimation of exposure when the focus of a study is possible adverse consequences of chemical exposure from hazardous-waste sites. This chapter reviews some aspects of exposure assessment or analysis that are important in environmental epidemiology and that illustrate the central role of exposure assessment. This chapter also discusses opportunities to improve analysis of exposure. The importance of exposure assessment has been underscored in several reports (NRC 1988, 1991a,b). An International Society of Exposure Analysis has been formed, and the Science Advisory Board of the Environmental Protection Agency (EPA) recommended that EPA develop a 5-year program on exposure assessment (EPA, 1988). The National Human Exposure Assessment Survey (NHEXAS) is a federal interagency program to design and implement an exposure surveillance system for the US population. The overall goal is to obtain periodic and systematic measurements of population exposures to multiple chemicals, including data on important environmental media, pathways, and routes, so that we can accurately determine current status, document historical trends, and predict possible future directions for exposures to hazardous chemicals (Sexton, 1991; Sexton et al., 1995) The NHEXAS has 3 specific objectives: (1) to document the occurrence, distribution, and determinants of exposures to hazardous environmental agents, including geographic and temporal trends, for the US population; (2) to understand the determinants of exposure for potentially at-risk population subgroups, as a key element in the development of cost-effective strategies to prevent or reduce exposures (risks) deemed to be unacceptable; and (3) to provide data and methods for linking information on exposures, doses, and health outcomes that will improve environmental health surveillance, enhance epidemiologic investigations, promote development of predictive models, and ultimately lead to better decisions (Sexton et al., 1995). As pointed out by Burke and Sexton (1995), "NHEXAS represents perhaps the most comprehensive exposure surveillance initiative ever undertaken ... it has been designed to address the information needs of regulators and improve the scientific basis for risk assessment, risk management, and risk communication." The "consolidated report" of an EPA-appointed consensus team on NHEXAS concludes as follows: "The implementation of NHEXAS can be considered a turning point in environmental policy. It represents the first concerted effort to understand and track total individual exposures on a national scale'' (Burke et al., 1992). A complete description of the NHEXAS phase I field studies, as well as a summary of the rationale and justification for

OCR for page 26
--> NHEXAS was published in the Journal of Exposure Analysis and Environmental Epidemiology (1995; 5:229-444). EPA has made a significant contribution to exposure assessment by the issuance of guidelines for exposure analysis (EPA, 1992). The guidelines describe general concepts of exposure assessment and have application to risk assessment, trends analysis, and epidemiology. Those guidelines and the NRC report Human Exposure Assessment for Airborne Pollutants (NRC, 1991b) are major contributions to the assessment of the impact of toxic agents on potentially exposed populations. They provide a broad overview of the need for exposure assessment. This chapter is not intended to repeat the contents of those documents, but will focus on certain specific issues in exposure assessment for use in environmental epidemiology. Principal Concepts That Underlie the Content of This Chapter It is relevant to state the assumptions that the committee used as the basis and context for this chapter. The effective application of exposure assessment methods may improve the results of any epidemiologic investigation. As in any line of epidemiologic investigation, an improvement in exposure assessment can reduce bias and improve statistical power to detect adverse effects associated with exposure to environmental contaminants. However, important findings may derive from environmental-epidemiologic investigation even when the exposure assessment uses only simple and crude tools to characterize the exposure of a given population. Overreliance on sampling of exposure of individuals may not be cost-effective and may limit the size of the study, with little improvement over the findings based on indirect methods. Concepts and Methods in Exposure Assessment This section reviews some of the basic concepts inherent in exposure assessment. For a more detailed discussion, the reader is referred to the NRC report Human Exposure Assessment for Airborne Pollutants (NRC, 1991b), the EPA guidelines for exposure assessment (EPA, 1992), and the Agency for Toxic Substances and Disease Registry (ATSDR) Guidance Manual (ATSDR, 1994). Epidemiological research uses various exposure metrics. The choice of a specific metric will depend on the type of study in question, the resources available to the investigator, the conceptual framework behind the investigation, and above all, biologic considerations. In deciding which exposure metric is best in a particular study,

OCR for page 26
--> one must be clear about basic concepts of exposure analysis. Exposure assessment for use in environmental epidemiology must attend to 5 primary issues: (1) the definition and characterization of the potentially exposed population; (2) the collection of quantitative information on population exposure, temporal characteristics, and dose-response relations; (3) the medium and the microenvironment of principal concern in terms of exposure; (4) the use of information collected in one population in assessing potential risk to others; and (5) the biologic plausibility of any hypotheses based on mechanistic considerations that can assist and help guide the exposure assessment. ATSDR (1994) has developed a definition of exposure as "an event that occurs when there is contact at a boundary between a human being and the environment with a contaminant of a specific concentration for an interval of time; the units of exposure are concentration multiplied by time." NRC (1991b) and EPA (1992) have also developed definitions of potential dose, applied dose, internal dose, and biologically effective dose for purposes of exposure assessment. The terms are illustrated in figure 3-1. Potential dose is the amount of the chemical ingested, inhaled, or in material applied to the skin. Applied dose is the amount of a chemical that is absorbed or deposited in the body of an exposed organism. Internal dose is the amount of a chemical that is absorbed into the body and available for interaction with biologically significant molecular targets. Biologically effective dose is the amount of a chemical that has interacted with a target site over a given period so as to alter a physiologic function. A concept important to any type of study is that of total exposure. Assessment of total exposure has received considerable attention in recent years (Lioy, 1990; NRC, 1991a; Wallace, 1991; Wallace et al., 1986, 1987, 1988). Total-exposure assessment consists of estimating possible exposure from all media (soil, water, air, and food) and all routes of entry (inhalation, ingestion, and dermal absorption). NRC (1991a) and Lioy (1990) have developed a conceptual framework for human total exposure assessment that may serve as a guide for assessing human exposure to environmental contaminants. The framework is outlined in table 3-1. This framework accounts for all exposures to a specific agent or group of agents that an individual may have had, regardless of the environmental medium. Total-exposure assessment has particular relevance in environmental epidemiology insofar as it facilitates identification of the principal medium or microenvironment of concern and provides information on potentially confounding exposures. NRC (1991a,b) described the different measurement and estimation techniques used in exposure assessment. Categories were defined that include direct measurement of exposure (personal monitoring, biologic monitoring, and biomarkers); indirect measures (microenvironmental

OCR for page 26
--> Figure 3-1 Schematic of dose and exposure. monitoring coupled with exposure models, where microenvironmental monitoring is defined as the monitoring of contaminant concentrations in locations or media in which exposure occurs); that include mathematical modeling, questionnaire/diaries, or spatial factors, e.g., residence in a country or region or distance from a source of chemical contamination (figure 3-2). EPA (1992) has also provided examples of types of measure-

OCR for page 26
--> ments needed to characterize exposure-related media and associated parameters (table 3-2). Ryan (1991) has reviewed aspects of human exposure modeling that are useful for understanding the concept of exposure assessment. When direct measurement of exposure is possible, it generally provides more-accurate information than indirect assessment of a particular individual's contact with a specific contaminant over time. The tradeoff is often between accurately measuring exposures over short periods, often outside the period of disease etiology, and indirect methods of assessing exposure over lengthier, more-relevant intervals. Personal monitoring has been widely used in workplace settings and can provide a measure of exposure across a range of microenvironments where individuals reside or work, though it is generally limited to a single chemical compound. However, personal sampling is often expensive, may demand extensive analytic capability and methodologies, and requires care in selecting study subjects. Biologic monitoring provides direct measures that integrate all routes of exposure to contaminants. Biologic monitoring may also provide more-precise estimates of target-organ dose, if appropriate toxicokinetic and metabolic information is available. Indirect sampling uses exposure data available for defined areas or other microenvironments, generally from monitoring with time-activity information. Exposures in each microenvironment are weighted for the average time spent there and added to assess total personal exposure. Validation of the specific applicability of indirect monitoring is an important requirement for the successful use of this method. Some important epidemiologic studies have emphasized indirect measures of exposure as the primary linkage to health outcome. Some of these studies are reviewed below. Exposure-Data Needs for Epidemiology Studies NRC (1991a) has discussed exposure assessment in relation to the type of study being conducted: "The type of exposure assessment and the acceptable level of uncertainty in the data vary according to whether the assessment is designed to generate or test hypotheses about exposure, test instruments, make risk assessment decisions, or make regulatory decisions." Gann (1986) has asked, "What kind of exposure data do epidemiologists need?" He argues that the answer depends on the development of a well-defined research question. Bailar (1989) points out that the definition of dose-response relations is usually critical to establishing causality. It is expected that, when it is possible to examine ordered categories of exposure, higher doses will have greater effects so that "the dose-response relationship is monotonic.'' Thus, "departures from mono-

OCR for page 26
--> TABLE 3-1 Parameters Required to Calculate Potential and Internal Dose Airborne contaminant Water contaminant I. Concentrations (µg/m3,ppb) I. Concentration (µg/L,ppm)    A. Microenvironments    A. Tap water    B. Personal    B. Water uses      C. Effluent II.  Patterns of exposure       1. Industrial    A. Intensity "episode" concentrations versus normal levels (average)       2. Commercial         3. Residential    B. Frequency and duration of contact       4. Uncontrolled dumps III. Transport II. Patterns of exposure    A. Dispersion and advection    A. Drinking    B. Other meteorology related to removal rates (washout, fallout)    B. Swimming      C. Cooking    C. Indoor ventilation and removal rates    D. Bathing      E. Laundry IV. Chemistry    F. Showering    A. Formation rates      B. Transformation rates III. Solubility of contaminant V. Deposition rate (µg/cm2) IV. Volatility of contaminant    A. Environmental      B. Lung V. Transport      A. Groundwater VI. Contact    B. Surface water    A. Inhalation (dependent on exercise regime)(m3/time)    C. Domestic supply    B. Dermal deposition and permeability (µg/cm2/time) VI. Chemistry      A. Formation rates    C. Ingestion (food, soil)(µg/g/time)    B. Transformation rates      C. Degradation VII. Absorption      A. Within tissue VII. Contact rate (µg/L/time) via exposure route    B. Into the blood and other fluids          A. Ingestion      B. Skin      C. Inhalation (volatilized)   VIII. Absorption      A. Dermal deposition and permeability      B. Gastrointestinal tract   Source: Reprinted with permission from Lioy, 1990. Copyright 1990 American Chemical Society.

OCR for page 26
--> Soil and sediment Food (commercial and homegrown produce) I. Concentrations (µg/g) I. Concentrations (µg/g)    A. Dusts    A. Plants       1. Outdoor    B. Vegetables and fruit       2. Indoor    C. Milk    B. Contaminated soil    D. Animals and fish       1. Uncontrolled dumps    E. Cooked foods       3. Landfills    F. Beverages and water-based foods (µg/L)       2. Airborne deposition         4. Resuspension     II. Patterns of exposure II. Patterns of exposure    A. Rate (µg/L/time) (µg/g/time)    A. Frequency and duration    B. Frequency    B. Intensity of contact    C. Origin of food         1. Homegrown III. Percolation rate       2. Commercial distribution    A. Soil composition       3. Local farms    B. Water table       4. Processed foods    C. Solubility      D. Transport     III. Source of contamination IV. Volatilization    A. Naturally occurring contaminants    A. Contaminant    B. Airborne deposition    B. Soil composition    C. Fertilization    C. Top soil and cover    D. Pest control      E. Waste dumps V. Contact rate via exposure route    F. Water supply    A. Dermal deposition and permeability    G. Preparation and cooking techniques    B. Lung      C. Gastrointestinal tract (pica) IV. Contact rate       1. Population    A. Gastrointestinal (GI)       2. Abnormal ingestion behavior    B. Inhalation (cooking only) VI. Body parameter V. Absorption through GI tract    A. Lung volume        B. Exposed skin surface (condition of skin)   VII. Absorption      A. Soil composition      B. Contact and absorption rates   All media: Can be supplemented by measuring a biological marker of accumulated single-medium or multimedia exposures in blood, urine, feces, and so forth. Many of these usually are nonmedia specific. Body weight: Used for lifetime exposure and dose calculation.  

OCR for page 26
--> Figure 3-2  Possible approaches for exposure assessment. tonicity raise questions about causality." Therefore, a key purpose of exposure assessment is often to support evaluation of dose-response relations. Descriptive Epidemiologic Studies Exposure data for descriptive epidemiologic studies must fairly apply to the population from which the disease is arising. This requires estimation of the whole probability distribution of exposure, not just means, with special attention to accurate estimation of the upper end of exposures. This is especially important if many or most persons are thought to be exposed at levels below some "threshold" where effects first appear. This is best achieved by random sampling, and sampling theory will dictate the nature and size of the sample, as well as any repetitions over time or circumstance, from which exposure will be assessed. Unless exposure information over time is already available or obtainable, data from an investigation that measures exposure "now" must be used to infer exposure levels at earlier times, e.g., the times when cases of the disease were induced. If the distribution of induction times is not known, the use of current exposures may be highly uncertain. In some instances, disease rates in an exposed population are compared with disease rates in unexposed population units, but inferences are stronger if the investigator classifies risk by some population grada-

OCR for page 26
--> tion of exposure. Sometimes, if exposure is measured in only a few areas, analysis may have to be limited to simple "exposed-control" comparisons. Sometimes, as may often happen in studies of pollutants in air or diet, there are no unexposed population units, and low-(or lower-) exposure groups fulfill the role of a "baseline" group. The sophistication of the exposure information collected will in general depend on resources. Environmental modeling may be necessary to deal with area and temporal variation. Analytic methods to cope with these problems are discussed in chapter 6. Analytic Epidemiologic Studies The types of analytic studies used in environmental epidemiology were summarized in chapter 2. Here the particular exposure requirements of these studies are considered. Case-control Studies In case-control studies, the past exposures of cases and controls will have to be estimated, using historical records, if available, or current exposure measurements extrapolated back in time. Quantitative measures of exposure can reduce misclassification and allow the development of a dose-response curve. Cohort Studies Requirements for historical cohort studies are similar to those for case-control studies. For prospective cohort studies, there may be a need to estimate the extent of continuing exposure. In view of the large numbers required for cohort studies, resource constraints may make it impossible to do more than measure current or recent exposure. As we pointed out previously (NRC, 1991a), the identification of potentially hazardous exposures in a group often results in cessation of exposure. This does not remove the need for characterization of past exposures. Nested Case-control and Case-cohort Studies Sometimes in a cohort study it is possible to collect specimens that could characterize exposure (e.g., biologic markers), where the expense largely resides in the analysis rather than specimen collection. When stored specimens are available (see NRC, 1991c, for a discussion of quality assurance associated with specimen archives) and can be analyzed after

OCR for page 26
--> TABLE 3-2 Examples of Types of Measurements to Characterize Exposure-Related Media and Parametera Type of measurement (sample) Usually attempts to characterize (whole) A. For Use in Exposure Scenario Evaluation: 1. Fixed-Location Monitoring Environmental medium; samples used to establish long-term indications of media quality and trends. 2. Short-Term Media Monitoring Environmental or ambient medium; samples used to establish a snapshot of quality of medium over relatively short time. 3. Source Monitoring Release rates to the environment from sources (facilities). Often given in terms of relationships between release amounts and various operating parameters of the facilities. 4. Food Samples (also see #11 below) Concentrations of pollutants in food supply. 5. Drinking Water Samples Concentrations of pollutants in drinking water supply.

OCR for page 26
--> it may be that exposure from diverse sources or by various rates is not covered by environmental monitoring. For example, classification of residents by distance from an arsenic smelter may not adequately reflect the arsenic concentrations in their diets, and environmental monitoring may not provide good estimates of total arsenic exposure. Biologic markers are often assumed to be good indicators of exposure because they represent the integrated exposure from various sources and through various routes. However, to assess this assumption requires correlation of the marker with the potentially less-adequate environmental measure. There is no "gold standard." Perfect correspondence between the marker and the exposure could mean that neither is better than the other or that there are no other routes, sources, or host factors that intervene. On the other hand, it may mean that the marker is not an accurate reflection of these other intervening factors. It is important to determine whether the marker shows an exposure-response relation, whether all potential routes are accounted for, and whether susceptibility or host factors are addressed. Host factors, including behavioral factors and genetic characteristics, may influence the amount of a toxic agent that interacts with critical macromolecules in cells and tissues. This is the "biologically effective dose." The biologically effective dose assesses exposure from all routes and sources as well as some aspects of effect modification, possibly including host characteristics for uptake, metabolism, absorption, and excretion. However, the marker may not necessarily encompass all these factors. Thus, even when biomarkers are useful, the best appraisal of exposure may still include ambient and environmental measurements as well as biologic measurements. Numerous biomarker-related issues may arise during the conduct of studies, including questions of specimen collection, transport, storage, and assay; measurement error of technical variables in the assay; biologic variability; and assay interpretation and communication of results. In cohort studies, biologic markers may be measured in subsets of populations, such as in a nested case-control or case-cohort approach, to assess etiologic questions and mechanisms and to identify high-risk subpopulations. In these situations, biologic markers of exposure may be useful to (1) distinguish exposure subgroups, (2) determine whether there is a relation between exposure and dose, or (3) evaluate the relation between exogenous exposure and internal or biologically effective dose. Biologic markers may also be useful to identify the effect of an intervention. For example, does reduction of environmental emissions result in a reduction in the level of DNA adducts? Research studies to assess interventions need to include assessment of baseline levels of biomarkers in order to interpret the effect of the interventions.

OCR for page 26
--> Interpretation and Generalization of Study Results Biologic markers of exposure can be of use after a study has been completed. For example, if researchers wish to see how well the results of a completed study might apply to a broad population, they may sample the population for the distribution of a particular marker to determine whether exposures are constant over a wide range of geographic conditions, demographic descriptors, and occupations. Even if the original study did not measure such factors, biologic markers may clarify what exposures such target groups may have experienced. It may also be possible to perform individual risk assessments using biologic measures of exposure. A classic example of individual risk assessment is the use of serum-cholesterol measurement to predict disease risk (Truett et al., 1967). With more-recent technology, one might attempt precise individual risk assessment by studying an individual's specific spectrum of gene mutations from specific exposures to a carcinogen. Data Gaps, Research Recommendations, and Resource Limitations Few biologic markers of exposure have been validated. As indicated above, validation of a marker of exposure requires an understanding of the dosimetric characteristics pertaining to the time between exposure and the ascertainment of the markers, the degree to which the marker represents exposure, and the nature and shape of the exposure-marker relation (Schulte 1989; Stevens et al., 1991). Little is known about the prevalence, range of variability, persistence, and confounding factors of many candidate exposure markers. Such information must be collected before these can be used with confidence in environmental-epidemiologic studies. Markers of the biologically effective dose require additional research to assess the role of host factors, particularly genetic susceptibilities, as effect modifiers. Dosimetric Modeling More attention is being devoted to characterizing the quantity and timing of toxic chemical agents' reaching target tissue, because concentrations measured by microenvironmental monitoring or even personal dosimetry may not accurately reflect target-tissue dose. This has led to greater emphasis on mathematical models derived from biologic mechanisms of toxicity. This approach has historically been the focus of pharmacologists who have sought to develop appropriate models to characterize the relation between drug efficacy and dose. Mathematical

OCR for page 26
--> modeling of tissue dose, biologically effective dose, or internal dose is receiving greater attention in addressing toxicologic issues. However, there has been little attention to the relation between these models, hereafter referred to as dosimetric models, and classic epidemiologic models that estimate disease risk. Kriebel (1991) has discussed the importance of this approach to epidemiologic investigation: "Often it is difficult or impossible to accurately estimate the exposure experience of each member of a cohort, and so various kinds of proxy variables must be used. It is well known that the use of these proxies can introduce misclassification of exposure, often leading to underestimation of the magnitude of exposure-disease associations. Even when accurate exposure data are available, serious bias may still occur if these data are used in a mis-specified epidemiologic model to estimate an exposure-response relationship." Kriebel discusses several tenets for the use of dosimetric models. These tenets result in a 2-phase approach to epidemiologic investigation: first, is the development of a mathematical model to estimate individual doses; and second, the use of epidemiologic models to estimate the risk of disease associated with these estimated doses, with appropriate control for confounding. The advantage of this approach for environmental epidemiology is that the dosimetric model can quantify a hypothesis about uptake processes and metabolism of the chemicals in question and may provide insight into the biologic mechanisms of effect. This facilitates the overall design of the epidemiologic model and subsequent analysis. That is, the dosimetric model informs the exposure assessor and epidemiologist and provides a way to reduce misclassification and improve the precision of the study. These concepts have been applied by Hattis (1990), Smith (1992), Hodgson and Jones (1990), Vineis and Terracini (1990), Vincent and Mark (1988), Pinto et al. (1978), and Kriebel and Smith (1990). Smith (1985) developed a compartmental dosimetric model of dust deposition for an occupational-epidemiologic study of pulmonary fibrosis in silicon carbide workers. In discussing the Smith model, Kriebel (1991) asks whether the considerable effort that goes into the construction and use of such a model was justified. He raises 3 criteria that could be used to evaluate any particular model: (1) the better model will fit the data better, (2) the better model will accommodate such secondary characteristics of the exposure-response relation as interactions with other agents (effect modifiers) and time (such as latency), and (3) the predictions of the better model can be generalized to other exposure situations. Smith (1991) also discusses the use of toxicokinetic modeling for epidemiologic purposes and argues that the use of toxicokinetic models can differentiate among hypotheses about the mechanisms that underlie the relation between exposure and effects. The greatest overall contribution made by a 2-stage approach to epi-

OCR for page 26
--> demiology (as described by Kriebel, 1991) is to provide a framework within which all available data, including toxicologic information and experimental data, can be viewed. In fact, animal data may be essential to the development of some models. Ultimately, dosimetric models are mathematical expressions of formal hypotheses about the underlying physiologic processes that are the basis of the exposure-response relation (Kriebel, 1991). Training in Environmental-Exposure Assessment Human exposure assessment is inadequately addressed in most environmental-epidemiology studies, and one of the roots of this problem is the lack of training at the graduate level. There is a major need, in the United States and elsewhere, for the development of training programs in exposure assessment. There are extensive master's-degree programs for industrial hygiene as a result of the National Institute for Occupational Safety and Health's Educational Resource Center program and other extramural project grants for training. There is at least one similar training program that has environmental-exposure assessment as its focus—in the Department of Environmental Science at Rutgers University (Lioy, 1991b). Exposure assessment is addressed in courses on environmental risk assessment, but even there the context is more focused on risk assessment and site remediation than on epidemiologic investigation of public-health hazards. Training in exposure assessment must be multidisciplinary, with a multimedia approach, and should address all the major uses of exposure information—including risk assessment, epidemiology, environmental control, and exposure assessment—and industrial hygiene, toxicology, pollution prevention, and standard-setting. It would be useful to examine the relations among needs for training in these areas to define a new curriculum that would better address current and future needs. Given the costs, resource requirements, and political sensitivity of many environmental-epidemiologic studies, the failure to provide training for environmental assessment will need to be addressed by policy-makers and educators if we are to have substantial improvement in environmental epidemiology and risk assessment. Conclusions Exposure assessment is important in all environmental-epidemiologic studies. A wide range of exposure-assessment strategies and techniques are available for use in environmental-epidemiologic investigation. Associations have been clarified by improved use of exposure assessment even where indirect methods have been used.

OCR for page 26
--> Both direct measures (personal and biologic monitoring and biomarkers) and indirect measures (microenvironmental monitoring, diaries, and mathematical modeling) can be used for exposure assessment in environmental epidemiology. Each of these techniques has advantages and disadvantages. Their optimal use depends on the nature of the study, the biologic hypothesis, and resource constraints. No approach should be singled out as being the only acceptable strategy, e.g., personal monitoring. All approaches have validity and will improve the study if used appropriately. Better exposure assessment in environmental epidemiology will increase the power of studies to find associations. However, within a fixed budget, spending more money on exposure assessment per subject will reduce the number of subjects who can be studied and hence could reduce statistical power. The tradeoff between precision and the cost of larger samples means that power will not increase monotonically with improvements in the accuracy of the exposure assessment. In studies of multifactorial outcomes and low relative risks, a large sample is almost always required. This means that inexpensive methods for modest improvement of rough and inexpensive exposure assessment may be more valuable than more-accurate but expensive methods. This includes very inexpensive methods, such as the use of questionnaire data on activity patterns. Studies of large populations exposed to mixtures of air pollutants should incorporate detailed estimates of exposure, including detailed activity logs (including transit to work or school), the kind of air conditioning in the home and workplace, and the use of personal monitors to validate models in subsets of the population under study. The problems of exposure measurement in persons living close to hazardous-waste sites were discussed in volume 1. Most studies have been structured around an "exposed-unexposed" classification or have used surrogates of exposure, such as distance from the waste site. Estimation of past exposures is particularly difficult and unreliable. Misclassification is likely to be a crucial problem in studies of this nature, and improved characterization of exposure is a priority. The estimation of cumulative doses is an important component of many occupational-exposure studies, though such measures may not be valid even in occupational settings. The exposure-dose relation should be examined for nonlinearity before cumulative estimates are calculated. The relation between cumulative exposure and peak exposure is unknown in many environmental-epidemiologic studies, particularly those involving hazardous-waste site exposures or community exposures to episodic pollution. The characterization of complex mixtures is a continuing problem for

OCR for page 26
--> exposure assessment. Four priorities in addressing complex mixtures are quantification of exposure to complex mixtures, characterization of combined or interactive effects, toxicologic characterization of the complex mixture in question, and identification of subpopulations that may be especially sensitive to one or more of the components of certain complex mixtures. Biologic markers of exposure can strengthen environmental-epidemiologic studies. Unfortunately, few such markers are yet feasible in field studies, and few have been adequately validated. Efforts to improve and refine such indicators are important. The feasibility and value of banking blood samples for future analysis should be considered as studies are designed. Biomarkers of changes induced in the immune system of human subjects are needed. Health effects are often subtle, and risks are difficult to estimate. As a result, more attention is being given to the estimation of target-tissue dose in ways that reduce misclassification and improve precision. Development of mathematical models to estimate target-tissue dose (toxicokinetic modeling) that may be combined with epidemiologic models to estimate risk is a new and important area of research. Emphasis should be given to the development of training programs in environmental-exposure assessment. Improvement in the development and use of new techniques in exposure assessment is a high priority in environmental epidemiology. Data should of course be generated and collected under rigorous conditions of quality control. Bias must be minimized, and variance must be both minimized and estimated when quantitative conclusions are to be drawn. Measures of and checks on data quality should be prominent in every manuscript and report, and authors must not be reticent in bringing out the mechanisms of their study—and there will always be weaknesses. Because of the difficulties of conducting epidemiologic studies, both descriptive and analytic, it is rare for any one study to be definitive, and this is especially true in environmental epidemiology. Every public presentation, written or oral, including reports to scientific colleagues, should contain prominent caveats about overinterpretation. References Abbey, D.E., G.L. Euler, J.K. Moore, F. Petersen, J.H. Hodgkins, and A.R. Magie. 1989. Applications of a method for setting air quality standards based on epidemiological data. J. Air Pollut. Control Assoc. 39:437-445. Abbey, D.E., P.K. Mills, F.F. Petersen, and W.L. Beeson. 1991. Long-term ambient concentrations of total suspended particulates and oxidants as related to incidence of chronic disease in California Seventh-Day Adventists. Environ. Health Perspect. 94:43-50.

OCR for page 26
--> ATSDR (Agency for Toxic Substances and Disease Registry). 1994. Environmental Data Needed for Public Health Assessments: A Guidance Manual. NTIS PB-179827. Atlanta: USDHHS, ATSDR, Division of Health Assessment and Consultation. Bailar, J.C. 1989. Inhalation hazards: the interpretation of epidemiologic evidence. Pp. 3948 in D. V. Bates et al., eds. Assessment of Inhalation Hazards, New York: Springer-Verlag. Baker, D.B., S. Greenland, J. Mendlein, and P. Harmon. 1988. A health study of two communities near the Stringfellow waste disposal site. Arch. Environ. Health 43:325-334. Bates, D.V., and R. Sizto. 1987. Hospital admissions and air pollution in southern Ontario: the acid summer haze effect. Environ. Res. 43:317-331. Bates, D.V., M. Baker-Anderson, and R. Sizto. 1990. Asthma attack periodicity: a study of hospital emergency visits in Vancouver. Environ. Res. 51:51-70. Blair, A., and S. Zahm. 1990a. Methodologic issues in exposure assessment for case-control studies of cancer and herbicides. Am. J. Ind. Med. 18:285-293. Blair, A., and S. Zahm. 1990b. Herbicides and cancer: a review and discussion of methodologic issues. Pp. 132-145 in P. Band, ed. Recent Results in Clinical Cancer Research. Vol. 120. Occupational Epidemiology. New York: Springer-Verlag. Blair, A., and S. Zahm. 1992. Epidemiology Studies Of Cancer Among Agricultural Populations. Presented at Third International Symposium: Issues in Health, Safety and Agriculture. Saskatoon, Saskatchewan, Canada, May 10-15, 1992. Breslow, L., et al. 1989. Report of Santa Barbara Commission on Health Consequences of the Casmalia Resources Waste Disposal Facility. Report prepared for Santa Barbara Department of Health Services, California. Brown, L.M., M. Dosemeci, A. Blair, and L. Burmeister. 1991. Comparability of data obtained from farmers and surrogate respondents on use of agricultural pesticides. Am. J. Epidemiol. 134:348-355. Burke, T.A., and K. Sexton. 1995. Integrating science and policy in a national human exposure assessment survey. J. Exposure Anal. Environ. Epidemiol. 5:283-296. Burke, T.A., J. Creason, P. Ryan, and M. Schwab. 1992. An Approach to the Design of the National Health Exposure Assessment Survey (NHEXAS): A Consolidated Report. Preliminary draft for NHEXAS Advisory Committee. 52 pp. Checkoway, H., and C.H. Rice. 1992. Time-weighted averages, peaks, and other indices of exposure in occupational epidemiology. Am. J. Ind. Med. 21(1):25-33. Checkoway, H., D.A. Savitz, and N.J. Heyer. 1991. Assessing the effects of nondifferential misclassification of exposures in occupational studies. App. Occup. Environ. Hyg. 6:528-533. Copeland, K.T., H. Checkoway., A.J. McMichael, and R.H. Holbrook. 1977. Bias due to misclassification in the estimation of relative risk. Am. J. Epidemiol. 105:488-495. Dosemici, M., S. Wacholder, and J.H. Lubin. 1990. Does nondifferential misclassification of exposure always bias a true effect toward the null value? Am. J. Epidemiol. 132:746-748. EPA (US Environmental Protection Agency). 1988. Future Risk: Research Strategies for the 1990s. SAB-EC-88-040. Washington, DC: US Environmental Protection Agency Science Advisory Board. 19 pp. EPA (US Environmental Protection Agency). 1992. Guidelines for exposure assessment: notice. Federal Register, Part VI, Vol. 57, No. 104. Euler, G.L., D.E. Abbey, J. E. Hodgkin, and A.R. Magie. 1987. Chronic obstructive pulmonary disease symptom effects of long-term cumulative exposure to ambient levels of total suspended particulates and sulfur dioxide in California Seventh-Day Adventist residents. Arch. Environ. Health 42:213-222.

OCR for page 26
--> Euler, G.L., D.E. Abbey, J.E. Hodgkin, and A.R. Magie. 1988. Chronic obstructive pulmonary disease symptom effects of long-term cumulative exposure to ambient levels of total oxidants and nitrogen dioxide in California Seventh-Day Adventist residents. Arch. Environ. Health 43:279-285. Freeman, N.C., J.M. Waldman, and P.J. Lioy. 1991. Design and evaluation of a location and activity log used for assessing personal exposure to air pollutants. J. Expo. Anal. Environ. Epidemiol. 1:327-338. Gann, P. 1986. Use and misuse of existing data bases in environmental epidemiology: the case of air pollution. Pp. 109-122 in F. Kopfler and G.F. Craun, eds. Environmental Epidemiology. Chelsea, MI: Lewis. Greenland, S., and H. Morgenstern. 1989. Ecological bias, confounding, and effect modification. Int. J. Epidemiol. 18:269-274. [Published erratum appears in Int. J. Epidemiol. 1991 20:824.] Hammond, S.K. 1991. The uses of markers to measure exposures to complex mixtures. Pp. 53-66 in L.M. Rappaport, and T. Smith, eds. Exposure Assessment for Epidemiology and Hazard Control. Chelsea, MI: Lewis. Hasabelnaby, N.A., J.H. Ware, and W.A. Fuller. 1989. Indoor air pollution and pulmonary performance: investigating errors in exposure assessment. Stat. Med. 8:1109-1126; discussion 1137-1138. Hattis, D. 1990. Pharmacokinetic principles for dose-rate extrapolation of carcinogenic risk from genetically active agents. Risk Anal. 10:306-316. Hodgson, J.T., and R.D. Jones. 1990. Mortality of a cohort of tin miners 1941-86. Br. J. Ind. Med. 47:665-676. [Published erratum appears in Br. J. Ind. Med. 1990 47:846.] Hulka, B.S., and T. Wilcosky. 1988. Biological markers in epidemiologic research. Arch. Environ. Health 43:83-89. Jaakkola, J.K., V. Vilkka, O. Marttiila, P. Jappinen, and T. Haahtela. 1990. The South Karelia air pollution study: the effects of malodorous sulfur compounds from pulp mills on respiratory and other symptoms. Am. Rev. Respir. Dis. 142:1344-1350. Kennedy, S.M., D.A. Enarson, R.G. Janssen, and M. Chan-Yeung. 1991. Lung health consequences of reported accidental chlorine gas exposure among pulpmill workers. Am. Rev. Respir. Dis. 143(1):74-79. Kriebel, D. 1991. The Dosimetric Model in Epidemiology. Invited paper presented at 8th International Symposium: Epidemiology in Occupational Health, September 10-12, Paris, France. Organized under auspices of International Commission on Occupational Health, Scientific Committee on Epidemiology. 21 pp. Proceedings in press. Kriebel, D., and T.J. Smith. 1990. A nonlinear pharmacologic model of the acute effect of ozone on the human lungs. Environ. Res. 51:120-146. Lagakos, S.W., B.J. Wessen, and M. Zelen. 1986. An analysis of contaminated well water and health effects in Woburn, Massachusetts. J. Am. Stats. Assoc. 81:583-596. Landrigan, P.J. 1983. Epidemiologic approaches to persons with exposures to waste chemicals. Environ. Health Perspect. 48:93-97. Levine, S.P., R. J. Costillo, C.L. Geraci, and K.A. Conlin. 1985. Air monitoring at the drum bulking process of a hazardous waste remedial action site. Am. Ind. Hyg. Assoc. J. 46:192-196. Lioy, P.J. 1990. Assessing total human exposure to contaminants: a multidisciplinary approach. Environ. Sci. Technol. 24:938-945. Lioy, P.J. 1991a. Arguments on the Adequacy of Exposure Indices and Convergence with Health Indices of the Adverse Effects of Air Pollution. Paper given at the 91st Annual AWMA Meeting, Vancouver, British Columbia. Lioy, P.J. 1991b. Human exposure assessment: a graduate level course. J. Expo. Anal. Care Environ. Epidemiol. 1:271-281.

OCR for page 26
--> Lioy, P.J., N.C.G. Freeman, I. Walnman, A.H. Ster, R. Boesch, T. Howell , and S.I. Shupack. 1992. Microenvironmental analysis of residential exposure to chromium laden wastes in and around New Jersey homes. Risk Anal. 12:287-299. [Published erratum appears in Risk Anal. 1992. 12:463.] Mills, P., D. Abbey, W. L. Beeson, and F. Petersen. 1991. Ambient air pollution and cancer in California Seventh-Day Adventists. Arch. Environ. Health 46:271-280. Monster, A.C., and J.F.F. Smolders. 1984. Tetrachloroethene in exhaled air of persons living near pollution sources. Int. Arch. Occup. Environ. Health 53:331-336. Morgenstern, H. 1982. Uses of ecologic analysis in epidemiologic research. Am. J. Pub. Health 72:1336-1344. Neas, L.M., D.W. Dockery, J.H. Ware, J.D. Spengler, F.E. Speizer, and B.G. Ferris, Jr. 1991. Association of indoor nitrogen dioxide with respiratory symptoms and pulmonary function in children. Am. J. Epidemiol. 134:204-219. Neutra, R., J. Lipscomb, K. Satin, and D. Shusterman. 1991. Hypotheses to explain the higher symptom rates observed around hazardous waste sites. Environ. Health Perspect. 94:31-38. NRC (National Research Council). 1988. Complex Mixtures: Methods for In Vivo Toxicity Testing. Washington, DC: National Academy Press. 227 pp. NRC (National Research Council). 1991a. Environmental Epidemiology, Vol. 1. Public Health and Hazardous Wastes. Washington, DC: National Academy Press. NRC (National Research Council). 1991b. Human Exposure Assessment for Airborne Pollutants. Washington, DC: National Academy Press. NRC (National Research Council). 1991c. Monitoring Human Tissues for Toxic Substances. Washington, DC: National Academy Press. 211 pp. Ostro, B., and S. Rothschild. 1989. Air pollution and acute respiratory morbidity: an observational study of multiple pollutants. Environ. Res. 50:238-247. Ozonoff, D., and D. Wartenberg. 1991. Toxic exposures in a community setting: the epidemiologic approach. Pp. 77-88 in Molecular Dosimetry and Human Cancer. Boston: CRC Press. Ozonoff, D., M.E. Colten, A. Cupples, T. Heeren, A. Sachatzkin, T. Mangione, M. Dresner, and T. Colton. 1987. Health problems reported by residents of a neighborhood contaminated by a hazardous waste facility. Am. J. Ind. Med. 11:581-597. Peters, J. 1991. Epidemiologic Investigation to Identify Chronic Effects of Ambient Air Pollution in Southern California. Proposal to California Air Resources Board. Pinto, S.S., V. Henderson, and P.E. Enterline. 1978. Mortality experiences of arsenic-exposed workers. Arch. Environ. Health 33:325-331. Pirkle, J. L., L. L. Needham, and K. Sexton. 1995. Improving exposure assessment by monitoring human tissues for toxic chemicals. J. Exposure Anal. Environ. Epi. 5:405-424 Portnoy, P., and J. Mullahy. 1986. Urban air quality and acute respiratory illness. J. Urban Economics 20:21-28. Ryan, P. 1991. An overview of human exposure modeling. J. Exp. Anal. Environ. Epidemiol. 1:453-474. Samet, J.M., and W.E. Lambert. 1991. Epidemiologic approaches for assessing health risks from complex mixtures in indoor air. Environ. Health Perspect. 95:71-74. Samet, J.M., F.E. Speizer, Y. Bishop, J.D. Spengler, and B.G. Ferris, Jr. 1981. The relationship between air pollution and emergency room visits in an industrial community. J. Air Pollution Control Assoc. 31:236-240. Schulte, P. 1989. A conceptual framework for the validation and use of biological markers. Environ. Res. 48:129-144.

OCR for page 26
--> Schwab, M., A.P. Treblanche, and J.D. Spengler. 1991. Self-reported exertion levels on time/ activity diaries: application to exposure assessment. J. Expo. Care Anal. Environ. Epidemiol. 1:339-356. Schwartz, J., and A. Marcus. 1990. Mortality and air pollution in London: a time series analysis. Am. J. Epidemiol. 131:185-194. Schwartz, J., D. Wypij, D. Dockery, J. Ware, S. Zeger, J. Spengler, and B. Ferris, Jr. 1991. Daily diaries of respiratory symptoms and air pollution: methodological issues and results. Environ. Health Perspect. 90:181-187. Sexton, K. 1991. National Human Exposure Assessment Survey. EPA (Environmental Protection Agency). Position paper. Washington, DC: Office of Health Research. Sexton, K., and P.B. Ryan. 1989. Human Exposure to Air Pollution: Methods, Measurements, and Models. In: A. Watson, RR Bates, and D. Kennedy, eds. Air Pollution, the Automobile, and Public Health. Washington, DC: National Academy Press. Sexton, K., M. A. Callahan and E. F. Bryan. 1995. Estimating exposure and dose to characterize health risks: the role of human tissue monitoring in exposure assessment. Environ. Health Perspect. 103 (Suppl.): 13-29. Smith, T.J. 1985. Development and application of a model for estimating alveolar and interstitial dust levels. Ann. Occup. Hyg. 29:495-516. Smith, T.J. 1991. Pharmacokinetic models in the development of exposure indicators in epidemiology. Ann. Occup. Hyg. 35:543-560. Smith, T.J. 1992. Occupational exposure and dose over time: limitations of cumulative exposure . Am. J. Ind. Med. 21:35-51. Stern, A.H., N.C.G. Freeman, P. Pleban, R.R. Boesch, T. Wainman, T. Howell, S.I. Shupack, B.B. Johnson, and P.J. Lioy. 1992. Residential exposure to chromium waste-urine biological monitoring in conjunction with environmental exposure monitoring. Environ. Res. 58:147-162. Stevens, D.K., R.J. Bull, C.H. Nauman, and J.N. Blancato. 1991. Decision model for biomarkers of exposure. Regul. Toxicol. Pharmacol. 14:286-296. Truett, J., J. Cornfield, and W. Kannel. 1967. A multivariate analysis of the risk of coronary heart disease in Framingham. J. Chronic Dis. 20:511-524. Vincent, J.H., and D. Mark. 1988. The Measurement of Aerosols in Relation to Risk Assessment. International Workshop on Exposure Assessment for Epidemiology and Hazard Control, Woods Hole, Massachusetts, 1988. Vineis, P., and B. Terracini. 1990. Biochemical epidemiology of bladder cancer. Epidemiology 1:448-452. Wallace, L.A. 1991. Personal exposure to 25 volatile organic compounds. EPA's 1987 team study in Los Angeles, California. Toxicol. Ind. Health. 7:203-208. Wallace, L.A., E.D. Pellizzari, T.D. Hartwell, R. Whitmore, C. Sparacino, and H. Zelon. 1986. Total Exposure Assessment Methodology (TEAM) study: personal exposures, indoor-outdoor relationships, and breath levels of volatile organic compounds in New Jersey. Environ. Int. 12:369-388. Wallace, L.A., E.D. Pellizzari, T.D. Hartwell, C. Sparacino, R. Whitmore, L. Sheldon, H. Zelon, and R. Perritt. 1987. The TEAM (Total Exposure Assessment Methodology) study: personal exposures to toxic substances in air, drinking water, and breath of 400 residents of New Jersey, North Carolina, and North Dakota. Environ. Res. 43:290-307. Wallace, L.A., E.D. Pellizzari, T.D. Hartwell, R. Whitmore, H. Zelon, R. Perritt, and L. Sheldon. 1988. California TEAM study: breath concentrations and personal air exposures to 26 volatile compounds in air and drinking water of 188 residents of Los Angeles, Antioch, and Pittsburgh, California. Atmos. Environ. 22:2141-2164. Walter, S.D. 1991. The ecologic method in the study of environmental health. II . Methodologic issues and feasibility. Environ. Health Perspect. 94:67-73.

OCR for page 26
--> Wegman, D.H., E.A. Eisen, S.R. Woskie, and X. Hu. 1992. Measuring exposure for the epidemiologic study of acute effects. Am. J. Ind. Med. 21(1):77-89. Whorton, M.D., R.W. Morgan, O. Wong, S. Larson, and N. Gordon. 1988. Problems associated with collecting drinking water quality data for community studies: a case example, Fresno County, California. Am. J. Pub. Health. 78:43-46. Wrensch, M., S. Swan, P.J. Murphy, J. Lipscomb, K. Claxton, D. Epstein, and R. Neutra. 1990a. Hydrogeologic assessment of exposure to solvent-contaminated drinking water: pregnancy outcomes in relation to exposure. Arch. Environ. Health 45:210-216. Wrensch, M., S. Swan, J. Lipscomb, D. Epstein, L. Fenster, K. Claxton, P.J. Murphy, D. Shusterman, and R. Neutra. 1990b. Pregnancy outcomes in women potentially exposed to solvent-contaminated drinking water in San Jose, California. Am. J. Epidemiol. 131:283-300.