Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Assessment of human exposure to copper: A case study using the NHEXAS database

Abstract

Copper is an essential trace element and adverse health effects can potentially be associated with both very low and very high intakes. Accurate estimates of inhalation and ingestion (food and drinking water) exposures are therefore needed in order to realistically assess any effects of the distribution of copper intakes within the general population. The work presented here demonstrates an application of a customized subset of the MENTOR/SHEDS-4M computational system (Modeling ENvironment for TOtal Risk studies, employing the Stochastic Human Exposure and Dose Simulation approach, for Multimedia, Multipathway, Multiroute exposures to Multiple co-occurring contaminants. The application utilized data from the National Human Exposure Assessment Survey (NHEXAS) for USEPA Region V as well as from a variety of other available databases. The case study, using a statistical population-based modeling framework, was performed for Eaton County, MI. The results of the simulations, aggregated for six age subgroups of the general population, suggest that food intake is the major pathway for total copper exposure, while drinking water can have significant contributions at the tail of the distribution of intakes. Specifically, it was estimated that over 80% of the county population received potential doses of copper from food that were lower than the Institute of Medicine (IOM) Recommended Dietary Allowance (RDA) value of 900 μg/day. Furthermore, the total combined potential dose from food and water was only about two times greater than the recommended value only for individuals with intakes in the range above the 99th percentile of both food and water intakes. The values were well below the upper tolerable intake value of 10,000 μg/day. The inhalation route consistently acted as only a minor contributor to the total exposure.

Introduction

The overall objective of the modeling analysis reported here was an examination of population exposure to copper in the US through a representative case study that employed a new integrated, computer-based, system (framework) designed for “source-to-dose” assessments of multimedia, multipathway, multiroute exposures.

Copper is an essential trace element and therefore low doses can lead to deficiencies, which can have adverse developmental consequences; however, at high doses, copper can cause other toxic responses, including liver toxicity in sensitive populations (NRC, 2000). Thus, it is important to be able to develop realistic and accurate population exposure assessments and assess potential related environmental health concerns in a way that accounts for the U-shaped curve associated with copper toxicity and essentiality. The modeling analysis presented here has utilized a national-level multiscale system, developed at the Computational Chemodynamics Laboratory (CCL) of the Environmental and Occupational Health Sciences Institute (EOHSI) (www.ccl.rutgers.edu), that links currently available exposure models and databases with Geographic Information Systems (GIS) tools, to provide a versatile platform for addressing a wide range of issues regarding population exposures to multimedia contaminants, including exposures to copper. The copper-specific implementation of this system, “Copper_EXIS-USA” (EXIS: Exposure Information System) has been described in Georgopoulos et al. (2002a, 2002b).

The present case study focuses on Eaton County, MI, which was part of the National Human Exposure Survey (NHEXAS) for USEPA Region V, conducted by the Research Triangle Institute (RTI)/EOHSI consortium (Pellizzari et al., 1995). The methods of analysis presented here should, however, be directly transferable to other locations across the US as well as to other countries. However, coordinated efforts would be required in order to identify and make available similar archived information and databases for use in exposure characterizations. The framework employed in this work could in fact be used to identify critical data gaps and to specify data collection needs for improving assessments of exposures to copper.

Background

To assess human population exposures and doses to copper, data need to be retrieved (or collected) that can realistically be used to describe low intakes of copper which can result in deficiency, and high concentrations which can result in toxicity for those with potential genetic sensitivity (Georgopoulos et al., 2001, 2002a). Specifically, in order to link copper exposure to risk, the National Research Council (NRC, 2000) considered the hypotheses that a copper-sensitive gene contributes to the hepatic copper toxicity observed in infants and young children ingesting high amounts of copper in milk and water. The current evidence is that manifestations of Tyrolean Infantile Cirrhosis (TIC), Indian Childhood Cirrhosis (ICC) and Idiopathic Copper Toxicosis (ICT) involve both heredity and high copper intake (Muller et al., 1996, 1998; Tanner, 1998). A hypothesis stated by the NRC was that chronic ingestion of moderately increased amounts of copper produces disease in copper-susceptible genotypes (NRC, 2000). In the general population there is a range of acceptable intakes that will meet essential copper requirements and pose no risk of toxicity. According to a report of the Institute of Medicine (IOM, 2001), the primary criterion used to estimate the Recommended Dietary Allowance (RDA) for copper is a combination of indicators, including plasma copper and ceruloplasmin concentrations, erythrocyte superoxide dismutase activity, and platelet copper concentration in controlled human depletion or repletion studies. The RDA for adult men and women is 900 μg/day. The median intake of copper from food in the US is approximately 1.0–1.6 mg/day for adult men and women. The Tolerable Upper Intake Level for adults is 10,000 μg/day (10 mg/day), a value based on protection from liver damage as a critical adverse effect.

According to the NRC (2000), copper intake through diet appears to fall within the acceptable range for the average, normal, healthy individual. Furthermore, for most people, only a small fraction of an individual's intake of copper derives from drinking water; thus, drinking water is not considered to be an important source. However, leaching from copper plumbing could potentially result in some cases in a significant exposure to copper; the potential for copper toxicity is a concern in that case. To address this concern, the NRC (2000) conducted an analysis intended to provide guidance on the establishment of the maximum contaminant level goal (MCLG). A simplified, nationwide, exposure assessment to copper was performed for that purpose. Furthermore, to characterize risks associated with copper exposures, NRC (NRC, 2000) employed specific assumptions/hypotheses regarding the prevalence of sensitive populations and the degree to which copper in drinking water might contribute to copper excess in individuals in those populations.

The NRC concluded that a comprehensive nationwide survey for copper in drinking water was not available, and therefore it was not possible to accurately estimate copper intake via drinking water. There exist data, however, that could provide clues to the potential for copper overexposure via tap water. Since it is mandatory for water systems to sample for copper in first-draw water (i.e., after water has been motionless for at least 6 h) at the cold-water tap at locations in the water system vulnerable to copper contamination (USEPA, 1991, 1994), high levels are detectable and would be useful in exposure/dose characterizations. When the 90th percentile of samples exceeds 1.3 mg/l, the water purveyor must report that percentile value to the states and to the US Environmental Protection Agency (USEPA). In the NRC assessment (NRC, 2000), the 90th percentile copper concentrations reported by water purveyors were examined from 1991 to 1999. The 7307 values reported correspond to roughly 4500 individual water systems. With a few exceptions, water systems reporting values greater than 5 mg/l are small, serving 3300 or fewer people and those measurements were taken in locations that served nonresidential consumers (recreational facilities and schools). The reported 90th percentile concentrations for numerous systems, some of which serve small communities, are high, suggesting the potential for copper overexposure.

Previous Modeling of Exposures to Copper in Drinking Water

A detailed model for calculating individual human acute and chronic exposures to copper in drinking water, entitled Consumption Habit Exposure Model (CHEM), was developed for Chile (Lagos et al., 1999). The model can estimate daily exposure of individuals, as well as the peak concentration and dose of copper, which individuals ingest during a 24-h period. A model evaluation was performed through application, in a limited number of homes, of the Composite Proportional Sampling (CPS) method, which is used to measure chronic human consumption of contaminants from drinking water. There are three main sources of variability in a population exposure study of copper in drinking water: individual habits variability, chemical variability, and interindividual variability. It was established, during that model evaluation, that the first two sources of variability are crucial for completing an accurate exposure measurement. In some cases, these two sources can cancel each other out, whereas, in other cases, they can add to individual exposure estimation error. The CHEM model is not sensitive with respect to an individual's habits variability because the Water Consumption Habit Survey (WCHS) questionnaire asks for usual behavior. However, the CHEM model is very sensitive to changes in maximum concentration of copper measured on different days. The result suggests that in order to minimize estimation error of an individual's true exposure, measurements for the chemical variables used in the model should be made more than once, and preferably at least three times. From the perspective of essentiality, the model estimated that ingestion of copper from drinking water by the population of Santiago was on average 9% of the World Health Organization (WHO) recommendation for minimum total ingestion of copper for adults, assuming that 100% of the copper contained in drinking water is absorbed.

Methods

The present work employed newly available databases and computational tools to demonstrate the feasibility of evaluating the relative contributions of different media, pathways, and routes of exposure on copper intakes for the general population. Specifically, the case study utilized a subset of the MENTOR/SHEDS-4M implementation of EOHSI's Modeling ENvironment for TOtal Risk (MENTOR) studies (MENTOR) (Georgopoulos et al., 2004; see also USEPA's Council for Regulatory Environmental Modeling [CREM] website: http://cfpub.epa.gov/crem/). MENTOR/SHEDS-4M employs the Stochastic Human Exposure and Dose Simulation (SHEDS) approach for characterizing exposures to Multiple co-occurring contaminants from Multiple media, Multiple routes, and Multiple pathways (4M). In general, MENTOR/SHEDS-4M combines microenvironmental and human activities information to assess the relative contribution of (1) media (e.g., water, food, dust), (2) pathways (e.g., drinking water, diet, hand-to-mouth), and (3) routes (e.g., oral, inhalation, dermal) to (4) multiple contaminant exposures for individuals or populations. It addresses aggregate and cumulative exposures to co-occurring pollutants in a consistent manner, and provides the ability to focus on mechanism-relevant time scales and subpopulations of interest; furthermore, it uses a two dimensional Monte–Carlo methodology to quantify variability and uncertainty in model inputs and outputs. MENTOR/SHEDS-4M includes various modules for retrieving information from up-to-date national, regional, and local databases that contain environmental, microenvironmental, biological, physiological, demographic, etc. parameters. The structure and all individual component databases of Copper_EXIS-USA, the exposure information system that is employed for copper-specific analyses in conjunction with MENTOR/SHEDS-4M, have been described in detail in Georgopoulos et al. (2001, 2002a, 2002b) and are summarized schematically in flowchart form in Figure 1.

Figure 1
figure1

Schematic depiction of the databases, models, and flow of information of the overall structure of Copper_EXIS-USA and the MENTOR framework.

Population-Based Exposure Modeling (PBEM) for Copper within MENTOR

A PBEM framework was implemented within MENTOR to support characterizations of multimedia/multipathway exposures to environmental copper; the modules and information flows of this implementation are shown schematically in Figure 2 (Georgopoulos et al., 2002b). This modeling framework considers all exposure routes to estimate population exposures and doses to environmental agents and thus significantly expands upon the goals of the CHEM Model (Lagos et al., 1999) that was described in the Background section. For the purposes of characterizing exposures to copper, the subset of MENTOR/SHEDS PBEM framework employs the following six steps that consider the air, water, and food exposure pathways (Georgopoulos et al., 2002b):

  1. 1)

    Estimation of the multimedia background levels of copper (air, water, and food) for the area where the population of interest resides. This can be done in general through a combination of environmental model predictions and measurement studies.

  2. 2)

    Estimation of multimedia levels (indoor air, drinking water, and food concentrations) and temporal profiles of copper in various microenvironments such as residences, offices, restaurants, etc.

    1. a)

      The air concentrations are calculated either using mass balance models or linear regression equations developed from analysis of concurrent indoor and outdoor air measurement data available for particular types of microenvironments (Burke et al., 2001).

    2. b)

      The drinking water concentrations are obtained from regulatory monitoring databases (such as SDWIS/FED (USEPA, 2005)) or field study measurements (such as NHEXAS (Whitmore et al., 1999)). If such data are not available, the drinking water distribution is modeled using the EPANET2 model (Rossman, 2000) using treatment plant data to obtain the drinking water concentrations (see, e.g., Maslia et al. (2000) for a discussion of application of drinking water distribution modeling to epidemiological studies). For addressing the potential issue of copper leaching from the distribution network within the modeling performed for this step, the solubility model presented by Schock et al. (1995, 2000) provides an approach to predict copper concentrations in drinking water in the US. However, it should be noted that this model has only been validated with limited drinking water data for copper dissolution and precipitation in the US, and has not been successful for predicting the copper concentration after stagnation in other regions of the world.

    3. c)

      The food concentrations are obtained from survey studies such as the Total Diet Study (TDS) (USFDA, 2004) and the National Human Exposure Assessment Survey (NHEXAS) (Whitmore et al., 1999).

  3. 3)

    Selection of a fixed-size sample population in a way that it statistically reproduces essential demographics (age, gender, race, occupation, education) of the population unit used in the assessment (e.g., a sample of 500 people is typically used to represent the demographics of a given census tract).

  4. 4)

    Retrieval of the matching activity diary record from USEPA's Consolidated Human Activity Database (CHAD — McCurdy et al., 2000) for each individual of the sample population, based on the individual's demographic characteristics.

  5. 5)

    Calculation of (inhalation, drinking water, dietary, etc.) intake rates for the members of the sample population, reflecting/combining the physiological attributes of the study subjects and the activities pursued during the individual exposure events.

    1. a)

      The inhalation rate is calculated based on the individual's age, gender, and the METS (metabolic equivalent of tasks) value associated with the activity pursued (see Georgopoulos et al. (2005) and references therein).

    2. b)

      The drinking water and beverage consumption rates are estimated by extracting survey records matching the individual's demographic characteristics. The USDA's Continuing Survey of Food Intakes by Individuals (CSFII) is the most comprehensive database with information on this issue: it contains data on the quantity of “plain drinking water” and various other beverages consumed by individuals on two nonconsecutive days during the period of 1994 to 1996 and 1998 (Tippett, 1999). The data used to estimate mean per capita intake rates combined 2-day dietary recall data from the survey years during which 15,128 individuals supplied 2-day intake data. Individuals from all income levels in 48 states and Washington DC were included in the sample. A MENTOR module has been developed to search the CSFII database and extract the drinking water and beverage intake rates based on age and gender.

    3. c)

      Estimating the magnitude of dietary intake of copper requires the information on food consumption rates, composition of food item (recipe file), and copper residue (concentration) data in food. CSFII (Tippett, 1999) provides information on food consumption rates for the general US population. The Total Diet Study (TDS) database (Tao and Bolger, 1998; Baker et al., 2001) by USFDA (1991 to 1999) provides information on residue data in 267 types of raw agricultural commodities, which are composites of food items, for the period of 1991 to 1999. A recipe file provided by USEPA's Office of Pesticide Programs (OPP) (Xue, 2003) is used to link the CSFII and TDS databases in order to generate estimates of dietary intakes for the chemical of concern.

  6. 6)

    Combination of inhalation, drinking water, and dietary intake rates with the corresponding multimedia concentrations of copper for each activity event to assess exposures.

Figure 2
figure2

Structure of a source-to-dose Population-Based Exposure Modeling (PBEM) framework for copper exposure characterizations within MENTOR.

It is to be noted that the above steps mention specifically US databases as the sources of input information for the assessment. The approach, however, is universal and in principle could be applied to any location in the world where similar information is available or can be collected.

As mentioned earlier, a summary of the US databases available for supporting such detailed exposure assessments for copper is provided in the second volume of the monograph by Georgopoulos et al. (2002b); included are copper database descriptions for the US, grouped as multimedia exposure/biomarker studies (CDC-NHANES II and III, USEPA NHEXAS), environmental releases (USEPA TRI (USEPA, 2001a), ATSDR HazDat (Fay and Mumtaz, 1996)), ambient air (USEPA AIRS (USEPA, 2001b)), surface water, groundwater, and sediment (USGS WQN (Alexander et al., 1998) and NAWQA (USGS, 2004), USEPA STORET (Perwak et al., 1980) and EMAP (Eilers et al., 1987)), soils and sediments (USGS NGA (Grossman, 1998)), ecological (NOAA ORCA (Lauenstein and Cantillo, 1993)), drinking water (USEPA SDWIS/FED (USEPA, 1998)), and dietary (USFDA TDS (USFDA, 2004), USDA CSFII (Tippett, 1999)). The new nationwide dietary intake database, called What We Eat in America (Dwyer et al., 2001), was recently released by the US Department of Health and Human Services (USDHHS) for public use. This new database is the integration of two nationwide dietary intake surveys — the Continuing Survey of Food Intakes by Individuals (CSFII) and the National Health and Nutrition Examination Survey (NHANES), and has also been incorporated into the MENTOR system. The main sources of data used in the case study that follows were (1) the NHEXAS — Region V Study, (2) the USEPA CHAD, (3) the USDA CSFII, and (4) the USEPA AIRS air quality database. (See Table 1 for expansion of all acronyms mentioned above, as well as all other acronyms used in this article.)

Table 1 List of acronyms used in this article.

NHEXAS was designed by the Office of Research and Development (ORD) of the US Environmental Protection Agency (EPA) and scientists from multiple disciplines early in the 1990s to provide critical information about multipathway, multimedia population exposure distribution to chemical classes (Lioy and Pellizzari, 1995; Sexton et al., 1995). Sample collection began mid-1995 and was completed in late 1997. NHEXAS studies were conducted in three different regions of the US with the consortium of Research Triangle Institute (RTI) and the Environmental and Occupational Health Sciences Institute (EOHSI) collecting data in the States of Illinois, Indiana, Michigan, Minnesota, Ohio, and Wisconsin, EPA-Region V (Pellizzari et al., 1995). The researchers worked with a statistically representative group of participant volunteers to measure the level of chemicals in the air they breathed; in foods and beverages they consumed, including drinking water; in the soil and dust around their homes; and in their blood and urine. Environmental copper levels were measured in tap water and drinking water. In addition, there were 536 records available for copper levels in food and beverages.

The AIRS (Aerometric Information Retrieval System) is a computer-based repository of information about airborne pollution in the US and various World Health Organization (WHO) member countries. The system is administered by the US Environmental Protection Agency (EPA), Office of Air Quality Planning and Standards (OAQPS), Information Transfer and Program Integration Division (ITPID). AIRS contains the air quality information for programs that attempt to improve and maintain air quality, and the data are listed using standard information requirements and information handling procedures. This approach enables comparison and use data collected by organization from individual states in the USA. Data are collected within local, state, and national monitoring networks, and reported by environmental agencies in each state to EPA. EPA is in charge of the maintenance of records. Copper content information for particulate matter (TSP, PM10, PM2.5) for the years 1982 to 2000 reported in AIRS for the US, Mexico, Puerto Rico, and the Virgin Islands has been incorporated in Copper_EXIS-USA (Georgopoulos et al., 2002b).

Case Study

A case study was conducted for Eaton County, Michigan using the previously described PBEM framework within MENTOR/SHEDS-4M with databases obtained from Copper_EXIS-USA. This county was selected as an example because it was part of the NHEXAS — Region V study (Pellizzari et al., 1995); Table 2 summarizes basic demographic information for this County. The data used for the modeling analysis included:

  • Drinking water (standing water and flushed water) and food concentrations of copper obtained from the NHEXAS — Region V study (Sexton et al., 1995; see also USEPA's Human Exposure Database System (HEDS) website: http://oaspub.epa.gov/heds/study_list_frame).

  • Outdoor air concentrations of copper obtained from the USEPA's AIRS database (USEPA, 2001b).

  • Activity diaries and associated metabolic expenditures and needs rates for the calculation of inhalation rates from USEPA's CHAD database (McCurdy et al., 2000).

  • Drinking water and food consumption rates obtained from the NHEXAS — Region V study (Sexton et al., 1995).

Table 2 Summary profile of general demographic characteristics for Eaton County, Michigan (USCB, 2001).

The following describe how the above steps of comprehensive population exposure modeling have been implemented within the MENTOR/SHEDS-4M framework into the current study.

Steps 1 and 2: The background outdoor air concentrations of copper were estimated using the measurements obtained from an AIRS monitoring station near Eaton County, Michigan (station ID: 260770905, located in nearby Kalamazoo County) for the period of January 1, 1996 to September 30, 1996 (see Figure 3). The indoor air concentrations of copper were then calculated through microenvironmental model simulations. For copper concentrations in drinking water and food, the measurements obtained from the NHEXAS — Region V study were used. The distributions of these measurements (standing water, flushed water, food, and beverages) are shown in Figures 4a and b, 5 and 6, respectively. Each had a right-skewed distribution, which is quite common for environmental variables. Furthermore, none of the copper concentrations in drinking water exceeded the level of 3 mg/l, which may relate to acute effect of gastrointestinal symptoms (NRC, 2000). These NHEXAS measurements were further organized and grouped together according to the demographic attributes of NHEXAS study subjects (such as age, gender, race/ethnicity, etc) in a structured relational database for latter usages in the current study.

Figure 3
figure3

Distribution of copper concentrations (μg/m3) in ambient air obtained from an AIRS monitoring station near Eaton County, Michigan (station ID: 260770905, located in nearby Kalamazoo County) for the period of January 1, 1996 to September 30, 1996.

Figure 4
figure4

Copper concentrations (μg/l) in (a) standing water and (b) flushed water, from the NHEXAS USEPA Region V Study (The detection limit for copper concentration in drinking water is 1 μg/l (USEPA, 2005)).

Figure 5
figure5

Copper concentrations (μg/kg) in food, from the NHEXAS USEPA Region V Study.

Figure 6
figure6

Copper concentrations (μg/l) in beverages, from the NHEXAS USEPA Region V Study.

Step 3: A sample of 1000 “virtual individuals” was selected so as to statistically reproduce the demographic characteristics of Eaton County, MI. To justify that a sample of 1000 people is sufficient to represent the study population, the age distribution of the six age groups reported for Eaton County, Michigan from the 2000 US census survey was compared to those obtained using the samples of 500 and 1000 people (see Figure 7). It is shown that the sample of 500 people can provide satisfactory replication of the original age distribution, while the sample of 1000 people gives only a marginally improved representation. The behavior shown in Figure 7 is typical for other demographic variables also.

Figure 7
figure7

Comparison of the percentages of the total population for the six age groups reported for Eaton County, Michigan from 2000 US census survey with those obtained from using the samples of 500 and 1000 “virtual individuals”.

Step 4: A 24-h activity diary for each ‘‘virtual individual’’ of the simulated population was selected from the CHAD diaries so as to match the demographic characteristics of the virtual individual with respect to age, gender, and employment status. An activity diary is a sequence of events that simulate the movement of a ‘‘virtual individual’’ through geographic locations and microenvironments during the simulation period. Each event is defined by geographic location, start time, duration, microenvironment visited, and an activity performed. There are 113 microenvironments in the CHAD diaries. These microenvironments are grouped into nine categories in the current study: home, other indoor, outdoor, vehicle, school, office, store, restaurant, and bar.

Step 5: The inhalation rates of each “virtual individual” were calculated for each of the activity events of the diary obtained in Step 4. The drinking water and beverage intake rates were estimated using the MENTOR module to search the CSFII database and extract the matched values based on age and gender. The dietary intakes of copper were estimated directly from the measurements of the NHEXAS — Region V study.

Step 6: The estimated intake rates were then combined with the corresponding multimedia concentrations of copper to assess exposures. For each “virtual individual”, the concentrations of drinking water, food, and beverages were estimated by extracting the records from the relational database of the NHEXAS measurements with the similar demographic attributes.

As mentioned in the Background section, the copper concentration in drinking water can be different for standing water and flushed water. Standing water typically has higher values of copper concentrations due to chemical and electrochemical processes occurring during stagnation of the water in the copper pipes. Flushed water has lower copper concentrations, since the water was run for a few minutes before sampling, and could be considered of the same composition as at the water utility plant. It is therefore important to distinguish consumption of standing and flushed water. Since there are no records available specifically for drinking water consumption habits of individuals in the NHEXAS study, a simplified method was developed and used in the present work to distinguish the two drinking water “sources.” The method assumes that individuals drink water from a standing water source at their first meal of the day (breakfast), and then drink flushed water for the rest of the day. This assumption applied to the general population except the children in pre-school ages (0–4 years old) in Eaton County, Michigan, since young children usually get up later and they will not be exposed to first draw water as much as adults. The fraction of drinking water assigned as standing water is calculated by dividing the first meal time by the total meal time in a 1-day period according to the activity diary of each individual from CHAD. Thus, the total amount of exposure of an individual to copper in drinking water is then calculated as follows:

where Dtot is the total amount of drinking water consumed per day, Fs is the fraction of drinking water source from standing water, Cs is the copper concentration in standing water, Cf is the copper concentration in flushed water.

Figure 8 shows the distributions of copper concentrations (μg/l) in intake water calculated using the above assumption versus those measured in standing and flushed water. It is shown that the distribution of copper intake concentration in drinking water lies between those in standing water and flushed water, as they represent the lower (flushed water) and upper (standing water) bounds of the drinking water intake concentrations calculated from the above assumption.

Figure 8
figure8

Distributions of copper concentrations (μg/l) in intake water calculated for the population in Eaton County, Michigan versus those measured in standing and flushed water in the NHEXAS — Region V study.

Results of the simulations, performed with the PBEM implementation of MENTOR/SHEDS-4M and Copper_EXIS-USA, are presented in Figure 9 and Figure 10, and illustrate the cumulative copper exposure distributions from inhalation, food intake, and drinking water consumption for Eaton County, MI, respectively, for the general population (Figure 9) and for individual age groups of the general population (Figure 10). It is observed that food intake appears to be the major pathway for the total copper exposure. The drinking water pathway, however, shows significant contributions at the tail of the distribution of dose. Further, it should be noted that for the Eaton County population, over 80% of the population received doses of copper from food intake that were lower than the IOM recommended RDA value of 900 μg/day. The total intake from food and water summed to an intake amount which was only about two times greater than the recommended value for only individuals with intakes in the range above the 99th percentile of both food and water. The intake amounts were well below the upper tolerable intake value of 10,000 μg/day. The inhalation route consistently acted as only a minor contributor to the total exposure, generally being a factor of about 10–1000 lower than either food or water intake. The cumulative distributions of copper intake obtained for most individual age groups (5–19, 20–34, 35–54, 55–64, >65 years) were similar to the values obtained for the total population.

Figure 9
figure9

Cumulative population exposure distributions to copper, resulting from inhalation, food and drinking water consumption, for Eaton County, MI calculated with MENTOR/SHEDS.

Figure 10
figure10

Cumulative population exposure distributions to copper, resulting from inhalation, food, and drinking water consumption, as well as total intake for all age groups (0–4, 5–19, 20–34, 35–54, 55–64, and 65 years and older) of Eaton County, MI, calculated with MENTOR/SHEDS.

Figure 11 compares the cumulative copper exposure distributions from food intake and drinking water consumption calculated for Eaton County, MI with the estimates presented in the NRC report (NRC, 2000), where a constant copper drinking water concentration was assumed to be at MCL level (1.3 mg/l). It can be seen that the distribution curves for food intakes from the two studies are closer than those for drinking water intakes. It was acknowledged by the NRC report that comprehensive nationwide survey data for copper in drinking water are not available, and therefore constant copper concentration at the MCL level (1.3 mg/l) was used in order to illustrate potential distributions of copper intake for certain segments of the population if they were to consume water at this level. However, actual measured copper concentrations in drinking water, from the NHEXAS field study, were used in the MENTOR/SHEDS calculations for Eaton County, MI. Since both studies estimated drinking water consumption rates based on the same original data (Tippett, 1999), the differences in copper concentrations account for the discrepancy in the exposure distribution curves; it can be stated that the NRC approach significantly overestimates the contribution of the drinking water pathway for the NHEXAS Region V population considered here.

Figure 11
figure11

Cumulative distributions of copper intake from food (μg/kg) and drinking water (μg/kg), calculated with MENTOR (based on data from the NHEXAS USEPA Region V Study) compared to the National distributions calculated by NRC (NRC, 2000).

Discussion

The outcomes of the case study presented here indicate that the NHEXAS study data provide a valuable set of variables for developing initial estimates of multipathway exposures to multimedia environmental agents, such as copper. The CHAD database provided a means for bridging certain data gaps in activity patterns and linking these patterns to exposure characterization algorithms. The strength of the estimates from the simulations does, however, suffer somewhat from not having detailed patterns of drinking water use available for each subject. This uncertainty points to the need to include collection of detailed individual and population-based data on the patterns and activities associated with the usage (drinking, cooking) of tap water as well as the consumption of bottled water and beverages in future studies that are designed to examine exposures to multimedia environmental agents such as copper. Considering the fact that, for at least the high end of exposures, the intakes associated with the food and water ingestion pathways were similar, each would have to be considered in the design of a study aiming to identify potential links between exposure and health effects related to high intake of copper, for example, liver toxicity in sensitive populations (NRC, 2000). In contrast, it is clear that, for all age subgroups of the general population, the food pathway is the most significant source of exposure, and use of copper supplements could assist in reducing the risks caused by having too little copper in the diet. Even if the levels of copper in water found in the NHEXAS study were increased by a factor of 10–50 for the high-end range (Figure 4), bringing these values closer to the level of 3 mg/l that may relate to acute effects involving gastrointestinal symptoms (NRC, 2000), the food pathway would still be a dominant or equivalent exposure driver.

An important uncertainty involves the understanding of how robust the CHAD database is in characterizing activities that can lead to contact and high-end exposures within various subgroups of the general population. A major observation derived from the present analysis is the need to conduct exposure studies that will collect data necessary to evaluate — and expand — the adequacy of CHAD in characterizing “high-end contacts” within a population of concern. Specifically with respect to copper exposures, such an investigation should try to determine if there are “high-end groups” that are exposed to levels of copper in drinking water that are of a concern for public health. With this in mind, the populations should be selected in at least two distinctly different areas: one with a known history of high copper in water (e.g., a specific area in the US or another country), and another area where copper pipes are used, but is not necessarily known for high copper problems. A location within the previously examined NHEXAS Region V (Michigan, Ohio, Minnesota, Wisconsin, Indiana, or Illinois) would appear to be appropriate for the latter.

References

  1. Alexander R.B., Slack J.R., Ludtke A.S., Fitzgerald K.K., and Schertz T.L. Data from selected US Geological Survey National Stream Water Quality Monitoring Networks. Water Resour Res 1998: 34 (9): 2401–2405.

    CAS  Article  Google Scholar 

  2. Baker S., Driver J., and McCallum D. Chapter 6: Ingestion exposures in residences. Residential Exposure Assessment, A Sourcebook. Kluwer Academic/Plenum Publishers, Dordrecht/New York, 2001.

    Google Scholar 

  3. Burke J.M., Zufall M.J., and Ozkaynak H. A population exposure model for particulate matter: case study results for PM2.5 in Philadelphia, PA. J Expos Anal Environ Epidemiol 2001: 11: 470–489.

    CAS  Article  Google Scholar 

  4. Dwyer J., Ellwood K., Leader N.P., Moshfegh A.J., and Johnson C.L. Integration of the continuing survey of food intakes by individuals and the National Health And Nutrition Examination Survey. J Am Diet Assoc 2001: 101 (10): 1142–1143.

    CAS  Article  Google Scholar 

  5. Eilers J.M., Kanciruk P., McCord R.A., Overton W.S., Hook L.A., Blick D.J., Brakke D.F., Kellar P.E., DeHaan M.S., Silverstein M.E., and Landers D.H. Characteristics of Lakes in the Western United States. Volume II: Data Compendium of Site Characteristics and Chemical Variables. EPA/600/3-86/054b. US Environmental Protection Agency, Washington DC, 1987.

    Google Scholar 

  6. Fay R.M., and Mumtaz M.M. Development of a priority list of chemical mixtures occurring at 1188 hazardous waste sites, using the Hazdat database. Food Chem Toxicol 1996: 34 (11–12): 1163–1165.

    CAS  Article  Google Scholar 

  7. Georgopoulos P.G., Roy A., Lioy M.J., Opiekun R.E., and Lioy P.J. Environmental copper: its dynamics and human exposure issues. J Toxicol Environ Health, 2001: 4 (Part B): 341–394.

    CAS  Article  Google Scholar 

  8. Georgopoulos P.G., Roy A., Yonone-Lioy M.J., Opiekun R.E., and Lioy P.J. Environmental Dynamics and Human Exposure to Copper Volume 1: Environmental Dynamics and Human Exposure Issues. International Copper Association, New York, NY, 2002a.

    Google Scholar 

  9. Georgopoulos P.G., Wang S.W., Vyas V.M., Lioy P.J., Tan H.C., Georgopoulos I.G., and Yonnone-Lioy J. Environmental Dynamics and Human Exposure to Copper Volume 2: Framework and Data Sources for Assessing Human Exposure to Copper in the United States. International Copper Association, New York, NY, 2002b.

    Google Scholar 

  10. Georgopoulos P.G., Wang S.W., Vyas V.M., Sun Q., Burke J., Vedantham R., McCurdy T., and Özkaynak H. A source-to-dose assessment of population exposures to fine PM and ozone in Philadelphia, PA, during a summer 1999 episode. J Expos Anal Environ Epidemiol 2005: 15 (5): 439–457.

    CAS  Article  Google Scholar 

  11. Georgopoulos P.G., Wang S.W., Yang Y.C., Xue J., Zartarian V., McCurdy T., and Özkaynak H, Assessing multimedia/multipathway exposures to arsenic using a mechanistic source-to-dose modeling framework. Case studies employing MENTOR/SHEDS-4M. Technical Report CERM:2004-02, Prepared for USEPA (Revised Draft submitted for agency review), 2004: http://ccl.rutgers.edu/reports/cerm/cerm-2004-01_arsenic.pdf.

  12. Grossman J.N . National Geochemical Atlas: The Geochemical Landscape of the Conterminous United States derived from stream sediment and other solid sample media analyzed by the National Uranium Resource Evaluation (NURE) Program. Open-File Report 98-622, US Geological Survey, Reston, VA., 1998: http://geo-nsdi.er.usgs.gov/metadata/open-file/98-622/metadata.faq.html.

  13. IOM. Dietary Reference Intakes for Vitamin A, Vitamin K, Arsenic, Boron, Chromium, Copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and Zinc (Chapter 7 – Copper). Institute of Medicine, Food and Nutrition Board, Washington DC, 2001.

  14. Lagos G.E., Maggi L.C., Peters D., and Revco F. Model for estimation of human exposure to copper in drinking water. Sci Total Environ 1999: 239: 49–70.

    CAS  Article  Google Scholar 

  15. Lauenstein G.G., and Cantillo A.Y. Sampling and analytical methods of the National Status and Trends Program National Benthic Surveillance and Mussel Watch Projects 1984 to 1992: overview and summary of methods, Volume I, Technical Memorandum NOS ORCA 71, NOAA, Silver Spring, MD, 1993: http://www.ccma.nos.noaa.gov/publications/tm71v1.pdf.

  16. Lioy P.J., and Pellizzari E.D. Conceptual Framework for Designing a National Survey of Human Exposure. J Expos Anal Environ Epidemiol 1995: 5: 425–444.

    CAS  Google Scholar 

  17. Maslia M.L., Sautner J.B., Aral M.M., Reyes J.J., Abraham J.E., and Williams R.C. Using water-distribution system modeling to assist epidemiologic investigations. J Water Resour Plann Manag 2000 July/August 2000: 180–198.

    Article  Google Scholar 

  18. McCurdy T., Geln G., Smith L., and Lakkadi Y. The National Exposure Research Laboratory's consolidated human activity database. J Expos Anal Environ Epidemiol 2000: 10: 566–578.

    CAS  Article  Google Scholar 

  19. Muller T., Feichtinger H., Berger H., and Muller W. Endemic Tyrolean infantile cirrhosis: an ecogenetic disorder. Lancet 1996: 347 (9005): 877–880.

    CAS  Article  Google Scholar 

  20. Muller T., Muller W., and Feichtinger H. Idiopathic copper toxicosis. Am J Clin Nutr 1998: 67 (Suppl): 1082S–1086S.

    CAS  Article  Google Scholar 

  21. NRC. Copper in Drinking Water. National Academy Press, Washington DC, 2000, 147.

  22. Pellizzari E.D., Lioy P.J., Quackenboss J., Whitmore R., Clayton A., Freeman N., Waldman J., Thomas K., Rodes C., and Wilcosky T. Population-based exposure measurements in EPA region; a phase-1 field study in support of national human exposure assessment survey. J. Expos Anal Environ Epidemiol 1995: 5: 327–358.

    CAS  Google Scholar 

  23. Perwak J., Bysshe S., Goyer M., Nelken L., Scow K., Walker P., Wallace D., and Delos C. An exposure and risk assessment for copper. EPA-440/4-81-015, US EPA, Washington DC, 1980.

  24. Rossman L.A., EPANET 2 Users Manual. EPA/600/R-00/057, USEPA, Water Supply and Water Resources Division, National Risk Management Research Laboratory, Cincinnati, OH, 2000: http://www.epa.gov/ordntrnt/ORD/NRMRL/wswrd/EN2manual.PDF.

  25. Schock M.R., Edwards M., Powers K., Hidmi L., and Lytle D.A. The Chemistry of New Copper Plumbing, Proceedings of the AWWA Water Quality Technology Conference. Salt Lake City, UT, 2000.

    Google Scholar 

  26. Schock M.R., Lytle D.A., and Clement J.A. Effect of pH, DIC, orthophosphate and sulfate on drinking water cuprosolvency. EPA/600/R-95/085, USEPA Office of Research and Development, Cincinnati, OH, 1995: http://www.epa.gov/ORD/WebPubs/effect.

  27. Sexton K., Kleffman D.E., and Callahan M.A. An introduction to the National Human Exposure Assessment Survey (NHEXAS) and related phase 1 field studies. J Expos Anal Environ Epidemiol 1995: 5: 229–232.

    CAS  Google Scholar 

  28. Tanner M.S. Role of copper in Indian childhood cirrhosis. Am J Clin Nutr 1998: 67 (5 Suppl): 1074S–1081S.

    CAS  Article  Google Scholar 

  29. Tao S.H., and Bolger P.M. Dietary arsenic intakes in the United States: FDA Total Diet Study, September 1991 — December 1996. Food Addit Contam 1998: 16 (11): 465–472.

    Article  Google Scholar 

  30. Tippett K.S. Food consumption surveys in the US Department of Agriculture. Nutr Today 1999: 34 (1): 33–46.

    Article  Google Scholar 

  31. USCB. Profiles of General Demographic Characteristics. 2000 Census of Population and Housing, Michigan, United States Census Bureau, 2001: http://www.census.gov/census2000/states/mi.html.

  32. USEPA. Monitoring requirements for lead and copper in tap water. Fed Regist 1991: 56 (110) US EPA.

  33. USEPA. Drinking water maximum contaminant level goals and national primary drinking water regulations for lead and copper. Fed Regist 59 (125) US EPA, 1994.

  34. USEPA. Information available from the Safe Drinking Water Information System (website): http://www.epa.gov/safewater/sdwisfed/sfed2.html, 1998.

  35. USEPA. 1999 Toxic Release Inventory Public Data Release. EPA 260-R01-001. USEPA Office of Environmental Information (2810) Washington DC, 2001a.

  36. USEPA. AIRS User's Guide. USEPA Office of Air Quality Planning and Standards, Research Triangle Park, NC, 2001b.

  37. USEPA. Manual for the Certification of Laboratories Analyzing Drinking Water, Fifth Edition, Chapter IV Critical Elements for Chemistry, Table IV-7 MCLs and Detection Limits Requirements in the CFR (mg/L) to composite (http://www.epa.gov/safewater/labcert/pdfs/manual_labcert_2005.pdf). EPA 815-R-05-004, USEPA, 2005.

  38. USEPA. Safe Drinking Water Information System/Federal Version. http://www.epa.gov/safewater/sdwisfed/sdwis.htm, 2005.

  39. USFDA. Total diet study. http://vm.cfsan.fda.gov/~comm/tds-toc.html, 2004.

  40. USGS. USGS National Water Quality Assessment Data Warehouse. USGS Water Resources Division, 2004: http://infotrek.er.usgs.gov/servlet/page?_pageid=543&_dad=portal30&_schema=PORTAL30.

  41. Whitmore R.W., Byron M.Z., Clayton C.A., Thomas K.W., Zelon H.S., Pellizzari E.D., Lioy P.J., and Quackenboss J.J. Sampling design, response rates, and analysis weights for the National Human Exposure Assessment Survey (NHEXAS) in EPA Region 5. Expos Anal Environ Epidem 1999: 9 (5): 369–380.

    CAS  Article  Google Scholar 

  42. Xue J. Personal Communication. USEPA Office of Pesticide Programs. Research Triangle Park, NC, 2003.

    Google Scholar 

Download references

Acknowledgements

Support for this work has been provided by the International Copper Association (Project TPT0619A/BB-00), and the results presented here have been derived from and expanded upon the information presented in ICA Publication “Environmental Dynamics of Copper and Human Exposure to Copper — Volume 2” (Georgopoulos et al., 2002b). The methods and computational tools used for the implementation of the study described here have been developed by the US EPA funded Center for Exposure and Risk Modeling (CERM) at EOHSI (EPA CR-827033), and the NIEHS Center Grant at EOHSI (P30ES05022). We extend our appreciation to our project officers, Dr. Scott Baker and Mr. Michael Hennelly, of the International Copper Association, for valuable data and other information they have provided in relation to this work as well as for their helpful comments and insights and to Linda Everett of the Computational Chemodynamics Laboratory for preparation of manuscript and graphics.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Panos G Georgopoulos.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Georgopoulos, P., Wang, S., Georgopoulos, I. et al. Assessment of human exposure to copper: A case study using the NHEXAS database. J Expo Sci Environ Epidemiol 16, 397–409 (2006). https://doi.org/10.1038/sj.jea.7500462

Download citation

Keywords

  • copper
  • NHEXAS
  • MENTOR/SHEDS
  • exposure assessment
  • drinking water
  • food consumption

Further reading

Search

Quick links