Model selection and evaluation for risk assessment of dioxin-contaminated sites

The general European population has a total intake of dioxins and dioxin-like chemicals near the limit recommended by the European Union, making additional exposure above background levels undesirable. For populations living near dioxin-contaminated sites, additional exposure may occur by intake of locally produced food, inhalation of particles, dermal contact with soils, or other exposure pathways. Risk assessment tools are required to estimate risks associated with contaminated sites and to set priorities for site remediation. Here, we review several multimedia models that can be applied as tools to support risk assessment. We then present a strategy to select, apply, evaluate, and adapt a model to address a specific situation. The case study we consider is a risk assessment of generic background dioxin exposure in Sweden, and we compare the predictions with environmental observations and exposure data from Sweden. Arguments are presented for selecting the CalTOX model for this case study. We demonstrate the application, evaluation, and adaptation of the model and discuss the requirements for extending the analysis to conduct risk assessment for subpopulations living near dioxin-contaminated sites.     

Using Multimedia Modeling to Expedite Site Characterization

GOAL, SCOPE AND BACKGROUND: This paper uses two case studies of U.S. Department of Energy nuclear weapons complex installations to illustrate the integration of expedited site characterization (ESC) and multimedia modeling in the remedial action decision making process. CONCEPTUAL SITE MODELS, MULTIMEDIA MODELS, AND EXPEDITED SITE CHARACTERIZATION: Conceptual site models outline assumptions about contaminates and the spatial/temporal distribution of potential receptors. Multimedia models simulate contaminant transport and fate through multiple environmental media, estimate potential human exposure via specific exposure pathways, and estimate the risk of cancer and non-cancer health outcomes. ESC relies on using monitoring data to quantify the key components of an initial conceptual site model that is modified iteratively using the multimedia model. CASE STUDIES: Two case studies are presented that used the ESC approach: Los Alamos National Laboratory (LANL) and Pantex. LANL released radionuclides, metals, and organic compounds, into canyons surrounding the facility. The Pantex Plant has past waste management operations which included burning chemical wastes in unlined pits, burying wastes in unlined landfills, and discharging plant wastewaters into on-site surface waters. CONCLUSIONS: The case studies indicate that using multimedia models with the ESC approach can inform assessors about what, where, and how much site characterization data needs to be collected to reduce the uncertainty associated with risk assessment. Lowering the degree of uncertainty reduces the time and cost associated with assessing potential risk and increases the confidence that decision makers have in the assessments performed.     

On the translation of uncertainty from toxicokinetic to toxicodynamic models--the TCDD example

When estimating human health risks from exposure to TCDD using toxicokinetic and toxicodynamic models, it is important to understand how model choice and assumptions necessary for modeling add to the uncertainty of risk estimates. Several toxicokinetic models have been proposed for the risk assessment of dioxins, in particular the elimination kinetics in humans has been a matter of constant debate. For a long time, a simple linear elimination kinetics has been common choice. Thus, it was used for the statistical analysis of the largest occupationally exposed cohort, the German Boehringer cohort. We challenge this assumption by considering, amongst others, a nonlinear modified Michaelis-Menten-type elimination kinetics, the so-called Carrier kinetics. Using the area under the lipid TCDD concentration time curve as dose metrics, we model the time to cancer-related death using the Cox proportional hazards model as toxicodynamic model. This risk assessment set-up was simulated in order to quantify uncertainty of both the dose (TCDD body burden) and the risk estimates, depending on the use of the kinetic model, variations of carcinogenic effect of TCDD and variations of latency period (lag time). If past exposure is estimated assuming a linear elimination kinetics although a Carrier kinetics actually holds, then high exposures in reality will be underestimated through statistical analysis and low exposures will be overestimated, respectively. This bias will carry over on the estimated individual concentration-time curves and the therefrom derived TCDD dose metric values. Using biased dose values when estimating a dose-response relationship will finally lead to biased risk estimates. The extent of bias and the decrease of precision are quantified in selected scenarios through this simulation approach. Our findings are in concordance with recent results in the field of dioxin risk assessment. They also reinforce the general demand for the scheduled uncertainty assessments in risk analyses.     

Potential Impact on Standards for Toxic Air Contaminants

Since the Bhopal incident, the public has placed pressure on regulatory agencies to set community exposure limits for the dozens of chemicals that may be released by manufacturing facilities. More or less objective limits can be established for the vast majority of these chemicals through the use of risk assessment. However, each step of the risk assessment process (i.e., hazard identification, dose-response assessment, exposure assessment, and risk characterization) contains a number of pitfalls that scientists need to avoid to ensure that valid limits are established. For example, in the hazard identification step there has been little discrimination among animal carcinogens with respect to mechanism of action or the epidemiology experience. In the dose-response portion, rarely is the range of “plausible” estimated risks presented. Physiologically based pharmacokinetic (PB-PK) models should be used to understand the difference between the tissue doses and the administered dose, as well as the difference in target tissue concentrations of the toxicant between rodents and humans. Biologically-based models like the Moolgavkar-Knudson-Venzon (MKV) should be developed and used, when appropriate. The exposure assessment step can be significantly improved by using more sensitive and specific sampling and analytical methods, more accurate exposure parameters, and computer models that can account for complex environmental factors. Whenever possible, model predictions of exposure and uptake should be validated by biological monitoring of exposed persons (urine, blood, adipose) or by field measurements of plants, soil, fish, air, or water. In each portion of an assessment, the weight of evidence approach should be used to identify the most defensible value. In the risk characterization, the best estimate of the potential risk as well as the highest plausible risk should be presented. Future assessments would be much improved if quantitative uncertainty analyses were conducted. Procedures are currently available for making future assessments. By correcting some of these shortcomings in how health risk assessments have been conducted, scientists and risk managers should be better able to identify scientifically appropriate ambient air standards and emission limits.     

The current practice of health risk assessment: potential impact on standards for toxic air contaminants

Since the Bhopal incident, the public has placed pressure on regulatory agencies to set community exposure limits for the dozens of chemicals that may be released by manufacturing facilities. More or less objective limits can be established for the vast majority of these chemicals through the use of risk assessment. However, each step of the risk assessment process (i.e., hazard identification, dose-response assessment, exposure assessment, and risk characterization) contains a number of pitfalls that scientists need to avoid to ensure that valid limits are established. For example, in the hazard identification step there has been little discrimination among animal carcinogens with respect to mechanism of action or the epidemiology experience. In the dose-response portion, rarely is the range of "plausible" estimated risks presented. Physiologically based pharmacokinetic (PB-PK) models should be used to understand the difference between the tissue doses and the administered dose, as well as the difference in target tissue concentrations of the toxicant between rodents and humans. Biologically-based models like the Moolgavkar-Knudson-Venzon (MKV) should be developed and used, when appropriate. The exposure assessment step can be significantly improved by using more sensitive and specific sampling and analytical methods, more accurate exposure parameters, and computer models that can account for complex environmental factors. Whenever possible, model predictions of exposure and uptake should be validated by biological monitoring of exposed persons (urine, blood, adipose) or by field measurements of plants, soil, fish, air, or water. In each portion of an assessment, the weight of evidence approach should be used to identify the most defensible value. In the risk characterization, the best estimate of the potential risk as well as the highest plausible risk should be presented. Future assessments would be much improved if quantitative uncertainty analyses were conducted. Procedures are currently available for making future assessments. By correcting some of these shortcomings in how health risk assessments have been conducted, scientists and risk managers should be better able to identify scientifically appropriate ambient air standards and emission limits.     

Using stochastic risk assessment in setting information priorities for managing dioxin impact from a municipal waste incinerator

The objectives of this study were to assess site-specific carcinogenic risk of incinerator-emitted dioxins in a manner reflecting pollutant transfer across multimedia and multi-pathways. The study used site-specific environmental and exposure information and combined the Monte Carlo method with multimedia modeling to produce probability distributions of risk estimates. The risk estimates were further categorized by contaminated environmental media and exposure pathways that are experienced by human receptors in order to pinpoint significant sources of risk. Rank correlation coefficients were also calculated along with the Monte Carlo sampling to identify key factors that influenced estimation of risk. The results showed that ingestion accounted for more than 90% of the total risk and that risk control on ingestion of eggs, aboveground vegetables, and poultry should receive priority. It was also found that variation of parameters with variability accounted for around 35% of the total risk variance, while uncertainty contributed to the remaining 65%. Intake rates of aboveground vegetables, eggs, and poultry were the key parameters with the largest contribution to variance. In addition, sufficient sampling and analysis of dioxin contents in eggs, aboveground vegetables, poultry, soil, and fruit should be performed to improve risk estimation because the variation in concentrations in these media accounted for the largest overall risk variance. Finally, focus should be placed on reduction of uncertainty associated with the risk estimation through ingestion of aboveground vegetables, eggs, poultry, fruit, and soil because the risk estimates associated with these exposure pathways had the largest variance.     

Assessing the health risk of reuse of bottom ash in road paving

Although the reuse of bottom ash has been favored gradually, reflected on regulations and researches, the associated risk is still an issue of great concern. This study quantified the health risks from multimedia transport and multi-pathway exposure to the concerned chemicals as a result of reusing bottom ash in road paving with consideration of various application scenarios. In particular, the using duration of the pavement was taken into consideration because movement of chemicals in the soils and groundwater would affect the subsequent exposure and risk. By using soil and groundwater transport modeling linked to food chain exposure assessment and incorporating the Monte Carlo method, the study identified Cr as the crucial toxicant and ingestion of drinking water and vegetables as the key exposure pathways. Furthermore, control of the using duration of road pavement is an essential factor of management and regulations to minimize the leaching of the hazardous constituents into the groundwater and subsequent contamination of food chain.     

Analysis of coupled model uncertainties in source-to-dose modeling of human exposures to ambient air pollution: A PM2.5 case study

Quantitative assessment of human exposures and health effects due to air pollution involve detailed characterization of impacts of air quality on exposure and dose. A key challenge is to integrate these three components on a consistent spatial and temporal basis taking into account linkages and feedbacks. The current state-of-practice for such assessments is to exercise emission, meteorology, air quality, exposure, and dose models separately, and to link them together by using the output of one model as input to the subsequent downstream model. Quantification of variability and uncertainty has been an important topic in the exposure assessment community for a number of years. Variability refers to differences in the value of a quantity (e.g., exposure) over time, space, or among individuals. Uncertainty refers to lack of knowledge regarding the true value of a quantity. An emerging challenge is how to quantify variability and uncertainty in integrated assessments over the source-to-dose continuum by considering contributions from individual as well as linked components. For a case study of fine particulate matter (PM_2.5) in North Carolina during July 2002, we characterize variability and uncertainty associated with each of the individual concentration, exposure and dose models that are linked, and use a conceptual framework to quantify and evaluate the implications of coupled model uncertainties. We find that the resulting overall uncertainties due to combined effects of both variability and uncertainty are smaller (usually by a factor of 3–4) than the crudely multiplied model-specific overall uncertainty ratios. Future research will need to examine the impact of potential dependencies among the model components by conducting a truly coupled modeling analysis.     

Apparatus for Simultaneous Size-Differentiated Sampling of Optical and Suboptical Aerosols: Application to Analysis of Nitrates and Sulfates

In recent years, the EPA has moved toward a risk assessment/risk reduction framework for making regulatory decisions. The Agency has taken a number of steps to assure the quality and consistency of the risk assessment component of those decisions. The first, and perhaps most important of these steps, is the development of Agency-wide risk assessment guidelines. Five guidelines have been published. They are: carcinogenicity, mutagenicity, developmental toxicity, chemical mixtures, and exposure. The provisions of the five guidelines are discussed in the context of the four components of risk assessment. Other activities designed to assure quality and consistency in risk assessments, reduce uncertainty in risk assessment, ensure a more efficient information exchange about risk and risk assessment, and develop the appropriate oversight mechanisms are also discussed. These include additional guidelines, the Risk Assessment Forum, risk assessment research, the Integrated Risk Information System, the Hazard Assessment Notification System, and the Risk Assessment Council.     

Prediction of overall persistence and long-range transport potential with multimedia fate models: robustness and sensitivity of results

The hazard indicators persistence (P) and long-range transport potential (LRTP) are used in chemicals assessment to characterize chemicals with regard to the temporal and spatial extent of their environmental exposure. They are often calculated based on the results of multimedia fate models. The environmental and substance-specific input parameters of such models are subject to a range of methodological uncertainties and also influenced by natural variability. We employed probabilistic uncertainty analysis to quantify variance in P and LRTP predictions for chemicals with different partitioning and transport behavior. Variance found in the results is so large that it prevents a clear distinction between chemicals. Additionally, only small improvements are observed when evaluating the results relative to a benchmark chemical. This can be explained by the dominance of substance-specific parameters and the only small direct influence of environmental parameters on P and LRTP as model outcomes. The findings underline the importance of learning how environmental conditions cause variability in substance behavior for improved substance ranking and classification.     

Human exposure and health risk of alpha-, beta-, gamma- and delta-hexachlorocyclohexane (HCHs) in Tianjin, China

The dynamic exposures to HCHs of individuals born between the years 1913 and 1993 in Tianjin have been simulated by connecting a fugacity model (IV) with a multimedia exposure model. Ingestion is the most important pathway for human beings to take up HCHs, and concentrations in the human body correlate with body weight changes. Accumulations of HCHs were derived assuming that the degradation in human body behaved linearly. The health risk of exposure to HCHs was measured using cancer risk and loss of life expectancy (LLE), and LLE was modified from its original definition to incorporate a dynamic calculation that takes variances in exposure into account. Monte-Carlo simulations were run to analyze the uncertainties of the model.     

An updated state of the science EQC model for evaluating chemical fate in the environment: application to D5 (decamethylcyclopentasiloxane)

The EQuilibrium Criterion (EQC) model developed and published in 1996 has been widely used for screening level evaluations of the multimedia, fugacity-based environmental fate of organic chemicals for educational, industrial, and regulatory purposes. Advances in the science of chemical partitioning and reactivity and the need for more rigorous regulatory evaluations have resulted in a need to update the model. The New EQC model is described which includes an improved treatment of input partitioning and reactivity data, temperature dependence and an easier sensitivity and uncertainty analysis but uses the same multi-level approach, equations and environmental parameters as in the original version. A narrative output is also produced. The New EQC model, which uses a Microsoft Excel platform, is described and applied in detail to decamethylcyclopentasiloxane (D5; CAS No. 541-02-6). The implications of these results for the more detailed exposure and risk assessment of D5 are discussed. The need for rigorous evaluation and documentation of the input parameters is outlined.     

Using a probabilistic approach in an ecological risk assessment simulation tool: test case for depleted uranium (DU)

A probabilistic approach was applied in an ecological risk assessment (ERA) to characterize risk and address uncertainty employing Monte Carlo simulations for assessing parameter and risk probabilistic distributions. This simulation tool (ERA) includes a Window's based interface, an interactive and modifiable database management system (DBMS) that addresses a food web at trophic levels, and a comprehensive evaluation of exposure pathways. To illustrate this model, ecological risks from depleted uranium (DU) exposure at the US Army Yuma Proving Ground (YPG) and Aberdeen Proving Ground (APG) were assessed and characterized. Probabilistic distributions showed that at YPG, a reduction in plant root weight is considered likely to occur (98% likelihood) from exposure to DU; for most terrestrial animals, likelihood for adverse reproduction effects ranges from 0.1% to 44%. However, for the lesser long-nosed bat, the effects are expected to occur (>99% likelihood) through the reduction in size and weight of offspring. Based on available DU data for the firing range at APG, DU uptake will not likely affect survival of aquatic plants and animals (<0.1% likelihood). Based on field and laboratory studies conducted at APG and YPG on pocket mice, kangaroo rat, white-throated woodrat, deer, and milfoil, body burden concentrations observed fall into the distributions simulated at both sites.     

A spatially-evaluated methodology for assessing risk to a population from contaminated land

A methodology is proposed which combines quantitative probabilistic human health risk assessment and spatial statistical methods (geostatistics) to produce an assessment of risks to human health from exposure to contaminated land, in a manner which preserves the spatial distribution of risks and provides a measure of uncertainty in the assessment. Maps of soil contaminant levels, which incorporate uncertainty, are produced from sparse sample data using sequential indicator simulation. A real, age-stratified population is mapped across the contaminated area, and intake of soil contaminants by individuals is calculated probabilistically using an adaptation of the Contaminated Land Exposure Assessment (CLEA) model. An abundance of information is contained in results which can be interrogated at the population and individual level, and mapped to provide a powerful visual tool for risk managers, enabling efficient targeting of risk reduction measures to different locations.     

Modeling exposure to particulate matter

Exposure assessment, a component of risk assessment, links sources of pollution with health effects. Exposure models are scientific tools used to gain insights into the processes affecting exposure assessment. The purpose of this paper is to review the process and methodology of estimating inhalation exposure to particulate matter (PM) using various types of models. Three types of models are discussed in the paper. Indirect type of models are physical models that employ inventories of outdoor and indoor sources and their emission rates to identify major sources contributing to exposure to PM, and use fate and transport and indoor air quality models to estimate PM concentrations at receptor sites. PM concentrations and time spent by a subject at each receptor site are input variables to the conventional exposure model that estimates the desired exposure levels. Direct type models use measured exposure or exposure concentrations in conjunction with information obtained from questionnaires to formulate exposure regression models. Stochastic models use exposure measurements, estimates can also be used, to formulate exposure population distributions and investigate associated uncertainty and variability. Since models developed using databases from western countries are not necessarily applicable in developing countries, the difference in requirements among western and developing countries is highlighted in the paper. Employment of exposure modeling methods in developing countries requires development of local information. Such information includes local outdoor and indoor source inventories, local or regional meteorological conditions, adjustment of indoor models to reflect local building construction conditions, and use of questionnaires to obtain local time budget and activity patterns of the subject population.     

Rationale for assessment of risk from exposure to 2,3,7,8-TCDD

The human health risk assessment is supported by methodology for utilizing toxic effects in animals consisting of carcinogenic and noncarcinogenic responses as a result of chronic, subchronic and acute exposures. One of the initial steps in a risk assessment activity involves the estimation of exposure levels. These estimates are typically based on either direct environmental measurements or predictions obtained from fate and transport models. The decision to develop assessment of risk from chronic exposure based on a nonthreshold model is made if a chemical demonstrates carcinogenic activity in animal bioassays and/or in human epidemiological studies. In the absence of any positive human epidemiologic data, it is assumed that a substance which induces a statistically significant carcinogenic response in animals has the probability to cause cancer in humans. The carcinogenic potential of 2,3,7,8-TCDD has been established based on chronic exposure in rodents. In addition, 2,3,7,8-TCDD has also been shown to be a liver cancer promoter in rodents. In the risk assessment on dioxins based on chronic exposure in experimental animals, 2,3,7,8-TCDD is regarded as a carcinogenic substance. Carcinogenic data from animal bioassays are utilized for the assessment of risk for the purpose of estimating the likelihood of 2,3,7,8-TCDD being carcinogenic for humans and to determine the magnitude of the potential impact on public health.     

The Sensitivity of Probabilistic Risk Assessment Results to Alternative Model Structures: A Case Study of Municipal Waste Incineration

Abstract

In this analysis, human health risk due to exposure to municipal waste incinerator emissions is assessed as an example of the application of probabilistic techniques (e.g., Monte Carlo or Latin Hypercube simulations). Incinerator risk assessments are characterized by the dominance of indirect exposure, thus this analysis focuses on exposure via the ingestion of locally grown foods. In addition, since exposure to 2, 3, 7, 8-TCDD drives most incinerator risk assessments, this compound is the subject of the illustrative calculations. An important part of probabilistic risk assessment is determining the relative influence of the input parameters on the magnitude of the variance in the output distribution. This constitutes an important step toward prioritizing data needs for additional research. However, under various possible model forms reflecting incompletely understood aspects of contaminant transport, differences may be observed in estimates of risk, variance in risk, and the relative contributions of individual uncertain and variable inputs to the variance. In this analysis, a sequential structural decomposition of the relationships between the input variables is used to partition the variance in the output (i.e., risk) to identify the most influential contributors to overall variance among them. For comparison, the partitioning of variance is repeated, using techniques of multivariate regression. In summary, this study considers the degree to which results of a probabilistic assessment are contingent on critical model assumptions about the representation of deposition velocity. Specifically, this analysis assesses the impact on the results of uncertainty about the best model of the vapor/particle partitioning behavior of semi-volatile airborne pollutants.     

The use of a multimedia risk analysis approach in designing waste management strategies

A pollution source may release residuals to any of several environmental media, depending on the process design and control strategies. These residuals then are subject to transfer, transport, and transformation within the interconnected compartments of the environmental system. The exposure and susceptibility of people and other receptors to pollutants are different in these various media, and so that risks imposed will vary according to the fate of the pollutants in the system. Because of interactions between compartments in the system, a single-medium approach to environmental management that mitigates problems in one environmental medium at a time independently of risk through other media may not minimize the aggregate risk a receptor receives from all pathways. Alternatively, a multimedia approach advocates focusing on the full environmental system providing pathways for exposure and selecting risk management strategies based on minimization of the aggregate and cumulative risk from all pathways and all compounds. This study combines multimedia risk analysis and an optimization framework to examine a methodology for selecting waste treatment/disposal and pollution control measures, applies the methodology to a sludge management decision problem, and considers the implications for continued use of single-medium analyses.     

The use of Monte-Carlo simulation techniques for risk assessment: study of a municipal waste incinerator

The incremental lifetime risks due to polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs) for the residents living in the surroundings of a municipal solid waste incinerator (MSWI) have been assessed. Two different pathways of exposure to PCDD/Fs, ingestion through the diet and exposure from MSWI emissions, were compared. Monte-Carlo simulations were carried out to obtain variability and uncertainty propagation The joint analysis of uncertainty and variability included a sensitivity analysis that identified the contribution to variance by different inputs. In general terms, PCDD/F ingestion through the diet contributed with more than 99% of the total risk, whereas direct exposition to PCDD/F emissions from the MSWI was less than 1% The results show that the median (50% percentile) of non-carcinogenic risk due to PCDD/Fs in the population living in the surroundings of the MSWI was 0.72 and the ratio of the 95th percentile and fifth percentile was about 2. With respect to the total carcinogenic risk, the median increment in individual lifetime was 7.90 x 10(-5) while the ratio between the 95th percentile and the fifth percentile was about 1.5. In this analysis, a sequential structural decomposition of the relationships between the input variables has been used to partition the variance in the output (risk) in order to identify the most influential contributors to overall variance among them.