HHRISK: A code for assessment of human health risk due to environmental chemical pollution

Ecotoxicology and Environmental Safety 170 (2019) 538–547

Contents lists available at ScienceDirect

Ecotoxicology and Environmental Safety journal homepage: www.elsevier.com/locate/ecoenv

HHRISK: A code for assessment of human health risk due to environmental chemical pollution

T

⁎

J.B. Nerisa, D.M. Montalván Olivaresa, , F.G. Velascoa, F.H.M. Luzardoa, L.O. Correiaa, L.N. Gonzálezb a b

Center for Research in Radiation Sciences and Technologies (CPqCTR), State University of Santa Cruz, Ilhéus, BA, Brazil Department of Exact and Natural Sciences, State University of Southwestern Bahia, Itapetinga, BA, Brazil

A R T I C LE I N FO

A B S T R A C T

Keywords: Human health risk assessment Heavy metal pollution Carcinogenic and non-carcinogenic effects Spatiotemporal analysis

Chemical environmental pollution is currently one of the most concerning environmental problem on a global scale, due to the high risks posed to ecological systems and human health. Risk assessment methodologies are valuable tools for preventive management and the mitigation of human health risks. However, the application of these methodological tools involves several steps and the knowledge of many variables, which can hinder its correct implementation. The main objective of this work was the development of the computational code for human health risk assessment: HHRISK (Human Health Risk). This code allows for an agile and accurate risk assessment based on the methodology established by the U.S. Environmental Protection Agency (U.S. EPA). Different from other published methods, the HHRISK code includes a new spatiotemporal matrix for the analysis of the aggregated risk (for multiple exposure pathways) and the cumulative (for exposure to multiple chemicals). HHRISK was applied to two case studies published dealing with the assessment of risk to human health through exposure to toxic metals, obtaining satisfactory results. The concordance between the average results obtained with the HHRISK and those reported by the authors confirm the validity of the implemented model. The inclusion of a greater spatiotemporal detail of the risks allowed to carry out a more accurate analysis and to propose new subsidies for a more efficient risk mitigation management by affected place and period of time.

1. Introduction Environmental pollution constitutes one of the biggest problems facing the world today. It worsens with each passing year, representing a public health problem on a global scale. For this reason, the interest in this topic has been increased in recent years, drawing the attention of the scientific community to this problem (Correia et al., 2016; Huang et al., 2016; Junaid et al., 2016; Naji et al., 2016; Pan et al., 2016; Yang et al., 2015). Risk assessment emerged as a tool capable of generating quantification forms that describe uncertainties about magnitudes, times and/ or possible health or environmental consequences associated with potential exposure to specific substances, processes, actions or events (Covello and Merkhoher, 2013). Risk assessment methods can be defined as systematic procedures, encompassing several stages: hazard identification, risk assessment and risk management (Rodricks and Levy, 2013). These procedures are central issues for analyzing and interpreting results related to environmental pollution threats. Since risk assessment involves several variables; operational errors ⁎

may occur at the time of risk analysis calculations. The use of computer software has proven to be a widely used tool as it helps to minimize errors and speed up calculations. In the literature consulted were found different Risk Assessment Software (RIS), of which we can highlight: BIOPLUME III (Rifai et al., 1997), BIOSCREEN (Newell et al., 1996), MODFLOW (Harbaugh and McDonald, 1996), MT3D (Zheng, 1992), and RBCA toolkit for chemical release (Connor et al., 2007). These softwares simulate the main pathways in which pollutants reach humans being through the different environmental compartments. Although written in a simple language, these RISs require a lot of data and present different approaches to describe the flow regime of heterogeneous systems, such as aquifers. Other softwares like RISC4 (Spence and Walden, 2001) and SADA (Stewart and Purucker, 2006) are more complete in their design. The first one is based on the Monte Carlo code and it not only simulates the transport of pollutants but also includes their biodegradation processes in the environment. On the other hand, SADA software performs spatial analysis thanks to the integration of Geographic Information System (GIS) modules, which allow visualizing the area and performing

Corresponding author. E-mail address: [email protected] (D.M.M. Olivares).

https://doi.org/10.1016/j.ecoenv.2018.12.017 Received 10 September 2018; Received in revised form 5 December 2018; Accepted 7 December 2018 0147-6513/ © 2018 Elsevier Inc. All rights reserved.

Ecotoxicology and Environmental Safety 170 (2019) 538–547

J.B. Neris et al.

Table 1 Parameters used in the dose calculations. Symbol

Definition (units)

Currently recommended value

Reference

ABS AFsoil

Absorption factor Soil adherence factor (mg cm−2)

AT

Averaging time (d)

– U.S. U.S. U.S. U.S.

BW

Body weight (kg)

Chemical-specific value 0.2 (Child) 0.07 (Adult) 0.12 (Worker) 25,550 (Carcinogenic effects) ED x 365 (Non-carcinogenic effects) 15 (Child) 80 (Adult) Site-specific value Site-specific value

– –

Site-specific value

–

1∙10−6 1∙10−3 6 (Child) 26 (Adult) 25 (Worker) 58 (Agriculture) 350 (Residential scenario) 250 (Worker) 24 (Residential scenario) 8 (Worker) 0.54 (Child) 0.71 (Adult) 1 1 1 0.75 12 (Child) 20 (Adult) 200 (Child) 100 (Adult) Food-specific value 0.78 (Child) 2.5 (Adult) Chemical-specific value 2373 (Child) 6032 (Adult) 3527 (Worker) 6365 (Child) 19,652 (Adult)

U.S. EPA (2004) U.S. EPA (2004) U.S. EPA (1991) U.S. EPA (2011) U.S. EPA (1991) CETESB (2001) U.S. EPA (1991)

Cair Csoil Cwater CF1 CF2 ED

EF ET

−3

Chemical concentration in air (mg m ) Chemical concentration in soil (mg kg−1) Chemical concentration in water (mg L−1) Conversion factor (kg mg−1) Volumetric conversion factor (L cm−3) Exposure duration (y)

Exposure frequency (d y−1) Exposure time (h d−1) −1

ETW

Water exposure time (h d

EV FA FI FR InhR

Event frequency (events d−1) Relative absorption factor Fraction ingested from contaminated source Lung retention factor Inhalation rate (m3 d−1)

IRS

Ingestion rate of soil (mg d−1)

IRF IRW

Ingestion rate of contaminated food (kg meal−1) Ingestion rate of water (L d−1)

PC SAS

Dermal permeability (cm h−1) Skin surface area available for contact with soil (cm2)

SAW

Skin surface area available for contact with water (cm2)

)

geospatial and statistical analyses. This program includes five different pollution scenarios and its use is exclusively for risk assessment for both human and ecosystems health. Although all of the aforementioned softwares have been widely used for risk assessment studies, they have some particular disadvantages, such as some consider only organic products, while others require knowledge of the changes undergone by the substance over time. However, a common disadvantage for all of them is that the risk assessment is done considering fixed exposure time values, and not variable time intervals. The aim of this paper is to fully describe the HHRISK computational code. This program performs temporary analysis of the risk assessment considering flexible exposure times and concentration values measured for the same place but in different periods. This feature can be useful in cases of accidents or risk assessment in areas whose pollutant concentrations vary over time. This information will enable the user not only to know in what period of exposure the population will be effectively at risk but also to implement more consistent mitigation actions in periods where pollution has not reached critical levels yet. The validation of the software will be done using real environmental pollution data taken from two published research articles by Gonçalves and Lena (2013), and Bempah and Ewusi (2016). Both assess the risk to human health from exposure to toxic metals: one in Ouro Preto (MGBrazil) and the other in the Obuasi gold mine (Ghana).

EPA EPA EPA EPA

(2004) (2004) (2011) (1989)

U.S. EPA (2011)

U.S. EPA (2011) U.S. EPA (2011) U.S. EPA (2004) CETESB (2001) U.S. EPA (2011) CETESB (2001) U.S. EPA (2011) U.S. EPA (2011) U.S. EPA (1991) – U.S. EPA (2011) – U.S. EPA (2011)

U.S. EPA (2011)

2. Risk assessment methodology 2.1. Exposure scenarios Exposure scenarios include general information (facts, data, assumptions and professional judgment) about how the exposure to certain chemical substance takes places. HHRISK code considers three main scenarios: agricultural, industrial and residential. Despite each scenario has its specific exposure pathways; they can be generally grouped by soil-based exposure pathways (soil and sediment), by waterbased exposure pathways (surface water and groundwater) and by airbased exposure pathways (particulate matter, vapors) (U.S. EPA, 1989). The main exposure routes common to the three scenarios are shown in Fig. S1. Since the industrial activity must be considered as a priority source of contamination (Huang et al., 2018), it was considered as the main pollutant agent in all cases. In the industrial scenario, industrial workers (adults) are routinely exposed to contaminants within the industrial site. Pathways are evaluated for exposures to the surface soil and surface water. Ingestion, dermal contact and inhalation of particulate material are the main routes of exposure taken into account. In the residential scenario, residents (adults and children) are continuously exposed to pollutants. In this case, the exposure is higher than in the industrial scenario because the daily exposure is calculated for a lifetime. In addition to the aforementioned routes for the industrial 539

Ecotoxicology and Environmental Safety 170 (2019) 538–547

J.B. Neris et al.

(Csoil), Biotransfer factors (BTF) and a correction factor, which takes into account the adjustment fresh weight/dry weight (F = 0.15). The BTF values are generally determined through bioassays and are available in the literature.

scenario, in this case, the consumption of food (vegetables, meat and milk) is also evaluated. In the agricultural scenario, all the exposure routes illustrated in Fig. S1 are considered and most of the exposure parameters have the same values. The main difference lies in the fact that the duration of exposure for residents in rural areas is 58 years, according to Brazilian legislation (CETESB, 2001). As in the residential scenario, both adults and children are considered exposed to pollutants.

Cfood = Csoil⋅BTFsoil − vegetables⋅F

HHRISK proposes a new method of risk analysis based on the time evolution of the concentration values of the pollutants for each contaminated site. The software follows standard procedures established by the United States Environmental Protection Agency (U.S. EPA) to calculate exposure and risk assessment. This agency, together with those of the European Union, leads environmental regulations and legislations worldwide (Vig et al., 2004). However, the equations proposed by the U.S. EPA were slightly modified. The factor, exposure duration (ED), was substituted by Δt for performing the temporal analysis year by year, and not after a fixed number of years of exposure. The HHRISK code considers nineteen (19) routes of human contamination by chemical substances. Eqs. (1)–(6) show the calculation of the dose for some of the most common exposure pathways used for human health risk assessment. These equations were taken from the U.S. EPA (1989, 2004) and CETESB (2001) guidelines. The parameters used in these calculations are summarized in Table 1.

HQ (t ) =

(mg kg −1 d−1)

(1)

- Absorbed dose by dermal contact with contaminated soil (Dder-soil) (U.S. EPA, 2004): j

∑i = 1 Csoil⋅CF1⋅SAS ⋅AFsoil⋅ABS⋅EV ⋅EF ⋅Δt

Dder − soil (t j ) =

BW ⋅AT

(mg kg −1 d−1) (2)

- Ingestion of contaminated drinking water (Ding-wat) (U.S. EPA, 1989): j ∑i = 1 Cwater⋅IRW ⋅EF ⋅Δt

Ding − wat (t j ) =

BW ⋅AT

∑ HQi (t ) i=1

(3)

- Absorbed dose by dermal contact with water (Dder-wat) (U.S. EPA, 1989):

BW ⋅AT

(mg kg −1 d−1)

- Inhalation of particulate matter (Dinh-mat) (CETESB, 2001): j

∑i = 1 Cair⋅InhR⋅FR⋅FA⋅ET ⋅EF ⋅Δt BW ⋅AT

(mg kg −1 d−1)

(5)

- Ingestion of contaminated fruit and vegetables (Ding-food) (U.S. EPA, 1989): j

Ding − food (t j ) =

∑i = 1 Cfood⋅IRF ⋅FI ⋅EF ⋅Δt BW ⋅AT

(mg kg −1 d−1)

(10)

Since, humans being are simultaneously exposed to multiple chemical agents, considering only the exposure to one chemical agent at a time leads to erroneous risk assessments. In the presence of another chemical agent, the damage caused by a specific chemical agent may be potentiated, reduced, nullified or simply remain invariable. The investigation of the effects of the joint action of several agents today is still an open problem and is the reason for a great international research effort in the area (U.S. EPA, 2014, 2012). The cumulative risk makes it possible to evaluate the possible carcinogenic and non-carcinogenic risks arising from simultaneous exposure to multiple chemical substances. In the present work we will use the additive approach recommended by the U.S. EPA (1989, 2007) that assumes that the risk due to one substance is not affected by the presence of another and the cumulative risk for w-chemical substances can be determined:

(4)

Dinh − mat (t j ) =

∑ CRi (t ) i=1

j

Dder − wat (t j ) =

(9)

n

CRagg (t ) =

∑i = 1 Cwater⋅CF2⋅SAW ⋅PC⋅ETW ⋅EF ⋅Δt

(8)

n

HIagg (t ) = (mg kg −1 d−1)

(7)

Where: D is the exposure level (or intake) for a substance considering a specific pathway of exposure over a certain period of time, RfD is the reference dose for that substance derived from a similar exposure period, and SF is the slope factor, which converts estimated daily doses averaged over a lifetime directly to incremental risk of an individual developing cancer. The HQ assumes that level of exposure below RfD (HQ < 1) do not represent an imminent risk to the population. For carcinogenic substances, the probability of cancer incidence in 1 person per 100,000 individuals is considered a tolerable level of risk to human health according to the Brazilian Resolution CONAMA 420/09 (2009). All the RfD and SF values used in the HHRISK calculations were taken from the U.S. EPA (2016) guidance. There are several exposure pathways to chemical substances and for each one exist an associated risk (see Fig. S1). The sum of more than one hazard quotient is defined as the hazard index (HI). The HI for multiple exposure pathways is commonly defined as the aggregate risk. It can be calculated for both carcinogenic (CRagg) and non-carcinogenic (HIagg) effects as the sum of the individual risk for a unique substance considering the n-possible exposure pathways.

j

BW ⋅AT

D (t ) RfD

CR (t ) = D (t )⋅SF

- Incidental ingestion of contaminated soil (Ding-soil) (U.S. EPA, 1989):

∑i = 1 Csoil⋅IR⋅CF1⋅FI ⋅EF ⋅Δt

(6A)

As is well known, the severity of the effects produced by chemical substances depends on their physicochemical characteristics and the time of exposure (ATSDR, 2000; Goyer and Clarkson, 1996; Järup, 2003). Exposure to chemicals may cause carcinogenic and non-carcinogenic effects, which are treated differently in the risk assessment calculations. The carcinogenic effects are stochastic in nature and do not have a safe dose threshold. The non-carcinogenic effects already appear after exceeding a certain dose threshold. The non-carcinogenic hazard quotient (HQ) and the potential carcinogenic risk (CR) are calculated using Eqs. (7) and (8) (U.S. EPA, 1989):

2.2. Risks calculation

Ding − soil (t j ) =

(mg kg −1)

(6)

w

HIcuml (t ) =

In Eq. (6), the concentration of contaminants in the food (Cfood) should be estimated from the concentration of contaminants in the soil

∑ HQkagg (t ) k=1

540

(11)

Ecotoxicology and Environmental Safety 170 (2019) 538–547

J.B. Neris et al.

Fig. 1. Flowchart diagram of the HHRISK software. w

CR cuml (t ) =

∑ CRkagg (t ) k=1

dermal contact with contaminated soil or water. This input file contains four main keys, which use a "true-false" binary logic system. The “scenario key” allows to choose which scenario will be studied (1 = agricultural, 2 = industrial, 3 = residential). The “way key” allows to consider or not certain exposure routes, according to the specific situation of interest. The “age key” was created to choose whether the risk for adults or children will be assessed. Finally, the “uncertainty key” was included to give the user the opportunity to choose whether or not the values of the uncertainty appear in the output file. All the information contained in this input file is read and stored by the READSCENARIO subroutine (see Fig. 1). Fig. 2b shows the structure of the other input file: Concentration.inp. In the same way as in Scenario.inp, the use of logical keys allows choosing which matrices (soil, water, air, food) will be taken into account. In the Concentration.inp file, the data is placed in the form of a matrix. The number of columns indicates the number of sampling points and in the rows, the concentration values measured in different time periods are placed. In this way, the matrix element Ci,j, would indicate the concentration value of the element at the place i measured at time j. The concentration values for each contaminant present in the analyzed matrices, their uncertainties, and the exposure time for all the pollutants are included in this file (see Fig. S2). The CONCENTRATION subroutine reads and stores these data in the form of a spatiotemporal matrix for each pollutant separately. After that, the second step executed by the program is to call the

(12)

3. Description of the main program The general structure of the HHRISK software is shown in Fig. 1. This is an easy-to-use program developed on the Fortran Power Station 90 (Microsoft) platform. Since a lot of variables are needed to perform risk assessment calculations, all the parameters must be organized in a reliable way to enable a fast and efficient access to the program's routines. For that reason, the program was organized to use two main input files. To increase the flexibility and practicality of the program, both the exposure routes and the involved variables are chosen through the input files to avoid changes in the source code. In the input file Scenario.inp (Fig. 2a), all the considered information related to the scenario and the exposure routes is entered. Variables such as: the number of both carcinogenic and non-carcinogenic pollutants, the number of locations and the time periods (year) for which the risk assessment will be carried out, the exposure parameters related to the ingestion and dermal contact rates and their uncertainties, can be written in the file directly from the keyboard. The first 16 routes of exposure consider the intake of soil, water and/or contaminated food, the 17th route is devoted to the inhalation of particulate material, while the last two take into account exposure by 541

Ecotoxicology and Environmental Safety 170 (2019) 538–547

J.B. Neris et al.

Fig. 2. Example of the Input files. a) Scenario.inp b) Concentration.inp.

542

Ecotoxicology and Environmental Safety 170 (2019) 538–547

J.B. Neris et al.

each exposure parameter, used in the HHRISK code, were evaluated from all the information available in the specialized literature. In some cases, uncertainties were calculated from the statistical distribution functions reported for some parameters (Sassi et al., 2007; U.S. EPA, 1996). When there was no specific data available on the statistical distribution or the uncertainty of the parameter, 10% of this value was considered as its uncertainty. Due to, Averaging Time (AT) is not considered affected by variability, its uncertainty was considered null (Sassi et al., 2007). The values and uncertainties of some exposure parameters, as well as their statistical distribution and the characteristic parameters for each distribution, are summarized in Table S1. The uncertainty values can be changed or updated by the user if desired. Such is the case of the Slope factor (SF), whose parameters depend on the SF value itself (La Grega et al., 1994) (see Table S1). Considering this, a logical key was implemented in the Datachemical.dat file, which gives the user the option of calculating the SF uncertainty value from this distribution or introducing a new uncertainty value for this quantity. If the “uncertainty key” in the Scenario.inp file is selected as TRUE, the uncertainty values for the doses, HQ, CR, HIagg, CRagg, HIcumul and CRcumul are printed in the output files. The detailed way in which the HHRISK code presents the uncertainty values allows the user to perform more specific analyses, for example, to identify which are the relevant variables that most affect the uncertainties of the final result. One of these analyses may be the creation of a "budget of uncertainty", which summarizes in the form of a table different parameters of a variable, parameters such as: value, uncertainty at value (σ (x)), relative uncertainty (σR), sensitivity coefficient (c(x)), quadratic term (q), and criticism. The last two parameters are calculated according (Sassi et al., 2007). The budget of uncertainty highlights the sources of risk uncertainty and provides a classification of the sources, based on their sensitivity coefficients (ISO, 2004). This useful tool was applied to one of the selected case studies and it will be shown in the results section.

READDATABASE subroutine. This subroutine is responsible for reading and storing the information contained in the Datachemical.dat and Dataexp.dat files. Each file constitutes a database which contains the complete set of parameters necessary for risk assessment. Datachemical.dat encompasses relevant information related to each pollutant such as: physical characteristics (density, molecular weight, solubility, etc.), parameters like the soil-water partition coefficient (Kd), the reference dose values (RfD), the U.S. EPA Carcinogenic classification, the slopes factors (SF), the soil-vegetable Biotransfer Factors (BTF) taken from (Baes et al., 1984), just to mention some of the included parameters. The Dataexp.dat file provides general data on the exposed population and the variables of temporal exposure: ED, EF, AT, and BW, used for the average daily dose calculations relative to each form of exposure and scenario. Both databases can be changed by the user depending on site-specific conditions or for update guidance values. The EXPOSITION subroutine gathers all the necessary information, organized in a spatiotemporal matrix, to perform the risk calculation. The program selects each individual point in the spatiotemporal matrix (Ci,j) and applies the corresponding formula to the exposure routes, according to the chosen scenario in the input file (Scenario.inp). These calculations for both carcinogenic and non-carcinogenic chemical substances are performed by specific subroutines named DOSEWAY, each one associated with a number to differentiate them (see Fig. 1). These calculations serve as a basis to generate a new spatiotemporal matrix that contains the values of carcinogenic (CR) and non-carcinogenic (HQ) risks, subsequently used by the RISKag and RISKcum subroutines to calculate the aggregate and cumulative risks for each place and time. The final results obtained by the program are stored in four output files. The Exposure.out file prints the values of all the data used in the calculations, the HQ values for all pollutants and the CR values for those that may cause cancer. The RISK.out file provides the HIagg, CRagg, HIcumul and CRcumul values in a summarized way, i.e., regardless of their variation in time, only by location and pollutant type. On the other hand, the RISKag.out and RISKcum.out files (see Fig. 3) provide the same magnitudes as RISK.out but in a detailed way, i.e., they show the spatial and temporal evolution of the aggregate and cumulative risks for each pollutant, which facilitates the data analysis and decision making. When evaluating the effects of pollutants on human health and ecosystems, assessing uncertainties is an essential issue because it highlights the implications and limitations of the risk assessment process (Dong et al., 2015; Sassi et al., 2007). According to the U.S. EPA (1989), there are three different approaches to the uncertainty analysis: quantitative, semi-quantitative and qualitative methods. The quantitative approach involves the assessment of uncertainties in the exposure parameters, which provides crucial information on the variability and sensitivity of the calculated results (U.S. EPA, 1996). For that reason, this method was implemented in the HHRISK code following the International Organization for Standardization (ISO) standard procedure reported in the Guide for the Expression of Uncertainty in Measurements (ISO, 2004). The standard uncertainty of the magnitudes (σF) is calculated as a combination of the standard uncertainties of the involved parameters, as shown below: N

σ 2F =

∂F

To illustrate the use of this computational code, two case studies were selected. The first one was the research carried out by Gonçalves and Lena (2013), which calculate the risk to human health associated with natural arsenic contamination in Ouro Preto (MG-Brazil). Another one was a similar research conducted by Bempah and Ewusi (2016), in which the risk to human health caused by various toxic metals in the vicinity of the Obuasi gold mine in Ghana was assessed. All the figures presented in the Result section were done using Origin 8.0 software. 4.1. Case 1 In Gonçalves and Lena (2013) an assessment of human exposure to natural arsenic (As) contamination in groundwater and soils was carried out. Six neighborhoods of the urban area of Ouro Preto (MG-Brazil) covering an area of approximately 2 km2, were studied: Piedade, Taquaral, Father Fania, Alto da Cruz, Antônio Dias and Barra. These neighborhoods were divided into three sectors to facilitate their study: Sector 1: Piedade - Taquaral, Sector 2: Father Fania - Alto da Cruz, and Sector 3: Antônio Dias - Barra. The arsenic present in the water and soil of this region is derived from the presence of this element in the oxidized degraded roofs, in the mineralized rocks and enriched with sulphite mineral bodies that are found in outcrops and in the walls of the abandoned mines. The chemical analysis of the soil samples performed in Gonçalves and Lena (2013) was carried out using an X-ray fluorescence spectrometer (XRF79C - Fusion with lithium tetraborate), while for the analysis of As the generation method was used hydrides (AAS/HAS14B). For the speciation of As in the water samples, the square wave voltammetry method was applied. For this purpose, they used a Metrohm -

2

∑ ⎛ ∂x ⎞ ⋅σ 2 (xi) i=1

4. Results and discusion

⎜

⎟

⎝

i⎠

(13)

For example, in the absorbed doses (see Eqs. (1)–(6)): xi is the ith exposure parameter involved in each case, σ(xi) represents the standard uncertainty of the ith parameter, and the (∂F/∂xi) is the partial derivate by the ith variable, also known as sensitivity coefficients (c(xi)). Unfortunately, the risk calculations depend to a large extent on the quality of the database, which in general tends to be imprecise due to the high heterogeneity among the studies. The uncertainties σ(xi) of 543

Ecotoxicology and Environmental Safety 170 (2019) 538–547

J.B. Neris et al.

Fig. 3. Output files without the uncertainty values. a) RISKag.out b) RISKcum.out.

544

Ecotoxicology and Environmental Safety 170 (2019) 538–547

J.B. Neris et al.

activity in the region is the extraction of gold, which resulted in high concentrations of heavy metals in soils, groundwater and streams in the Obuasi areas. It was evaluated the risks to human health associated with heavy metal pollution derived from gold mining. Polluted agricultural soils, groundwater and cash crops were considered the main routes of exposure of local residents to these pollutants. Samples of soil, drinking water and vegetables were collected at the three communities surrounding the mine. In the chemical analysis carried out by Bempah and Ewusi (2016) of all types of samples, the elements As, Cd, Cr, Cu, Pb, Hg, Ni, Fe, Mn and Zn were quantified. For the analysis of groundwater, 10 representative community wells were selected for sampling and quantitative analysis, which was performed using an inductively coupled plasma mass spectrometry (ICP-MS) (PerkinElmer SCIEX, ELAN 6000) while the cations and dissolved anions were measured in the same way ICP-MS and ion chromatography, respectively. For the analysis of the soil, samples were collected in 3 regions (Sanso, Dokyiwa and Pompora) that surrounded the area of the mine. Samples were digested using aqua regia (HCl and HNO3 3:1) and the resulting solutions were analyzed using ICP-MS. For the analysis of plants, samples from randomly selected farms were collected, digested with concentrated HNO3 and the resulting solution was analyzed by ICP-MS. The metal concentration values in the studied matrices, exposure parameters, contamination pathways and reference values (RfD and SF) used for the HHRISK calculations were those reported by Bempah and Ewusi (2016). According to the results obtained with the HHRISK, there is a risk to human health derived from water consumption, while the risk of ingestion of contaminated soil calculated is considerably lower than the value reported by Bempah and Ewusi (2016). Notwithstanding, the HI for the ingestion of contaminated vegetables calculated by the code only differs by 1.1% of the reported value, which means that they are in agreement (see Table S4). Bempah and Ewusi (2016) reported that the metallic species that contributed the most to the HI value of the consumption of vegetable crops were Pb (HQ = 12.90), followed by As (HQ = 7.30) and Hg (HQ = 4.30), while HHRISK pointed out that As (HQ = 17.00) was the main contributor followed by Hg (HQ = 3.32) and Pb (HQ = 3.02). In the case of calculations of carcinogenic risk for exposure to arsenic, the values reported by Bempah and Ewusi (2016) were: 1.55 10−4 (for groundwater intake), 3.60 10−6 (for soil ingestion) and 3.30 10−3 (for the ingestion of contaminated plants), while the values calculated by HHRISK were: 1.55 10−4, 3.61 10−6 and 3.28 10−3, respectively, showing a great concordance (the maximum value of the percentage deviation was 0.6%). Table 2 presents the budget of uncertainty for the specific case of carcinogenic risk due to exposure to arsenic through the ingestion of groundwater. For this analysis, the values of the parameters: IRW, EF, ED, BW, AT, and SF; the concentration of arsenic in groundwater and its uncertainty were taken from Bempah and Ewusi (2016). Uncertainty value for CR was calculated by using Eq. (13), where the σ(xi) values for the parameters involved are those reported in Table S1. The uncertainty value obtained is in the same range of magnitude, which is an expected result in the risk assessment calculations (U.S. EPA, 1989). The criticism value indicates which parameters contribute most to the uncertainties (criticism > 0.1) (Sassi et al., 2007). The criticism values obtained for parameters such as: BW, ED, EF and SF, were equal to/lower than 0.1; therefore, their contribution to the risk uncertainty is negligible. On the other hand, Table 2 shows that the concentration of the pollutant is the parameter that most affects the variability of carcinogenic risk in this particular case. The temporal analysis of the non-carcinogenic and carcinogenic risks caused by ingestion of contaminated vegetables is illustrated in Fig. 5. As shown, only by exposure to arsenic through the consumption of contaminated vegetables, safety values for carcinogenic and non-

Fig. 4. Spatiotemporal analysis of human health risk in the three residential sectors of Ouro Preto (MG-Brazil) a) Non-carcinogenic effects (HIagg) b) Carcinogenic effects (CRagg).

757 VA Computrace Polarimeter, equipped with a mercury pendant drop working electrode, a reference electrode Ag/AgCl/KCl 3 mol L−1 and an auxiliary platinum electrode. The mean concentrations of arsenic in groundwater and soils for each sector, exposure parameters, contamination pathways and reference values (RfD and SF) used as input parameters in the HHRISK calculations were taken from Gonçalves and Lena (2013) (see Table S2). The calculated HI values obtained by the HHRISK are slightly higher than those reported by Gonçalves and Lena (2013), which may be related mainly to the difference between some values used to perform the risk calculations. On the other hand, the average percentage deviation between the values of carcinogenic effects calculated by the code and those reported by Gonçalves and Lena (2013) is 2.7%, showing close similarity (see Table S3). Using the HHRISK, a detailed analysis year-by-year of the non-carcinogenic and carcinogenic risks for the residents of the studied region was done considering a whole exposure time of 24 years (Fig. 4). In Fig. 4a, the dashed lines represent the HI values calculated by Gonçalves and Lena (2013) for the three sectors, the solid line represents the threshold value (HI = 1), above which there should be concern about possible non-carcinogenic effects. It can be seen that after four years, exposure to arsenic represents an imminent risk for the population of Sector 2, while for sectors 1 and 3 it becomes worrisome after six years of exposure. In the case of risk of cancer incidence (Fig. 4b), the situation is more alarming, since all sectors exceed the reference value (1.27 10−4) after one year of exposure. This more precise analysis showed HI and CR values higher than the reported by Gonçalves and Lena (2013), indicating thus a considerable risk to residents of the six Ouro Preto neighborhoods. The use of the spatiotemporal analysis allowed to find from what period the exposure to the arsenic becomes dangerous for the habitants of Ouro Preto's studied neighborhoods. As can be seen, this kind of analysis is an important tool for carrying out a more detailed study of risks to human health. 4.2. Case 2 Bempah and Ewusi (2016) carried out a study to investigate the impact of gold mining in three farming communities in the municipality of Obuasi in Ghana. This area is located in the Ashanti region and covers an area of approximately 162.4 km2. The largest industrial 545

Ecotoxicology and Environmental Safety 170 (2019) 538–547

J.B. Neris et al.

Table 2 Budget of uncertainty for carcinogenic risk (CR) calculation. Parameters −1

Cwater (mg L ) SF (kg d mg−1) IRw (L d−1) EF (d y−1) ED (y) BW (kg) AT (d) CR

σ (x)

Value −3

8.8 × 10 1.5 2 350 30 70 25,550 1.6 × 10−4

σR (%) −2

1.32 × 10 0.1 0.8 38 5 8 0 2.4 × 10−4

149 9 40 11 17 11 0 157

c(x) 0.02 1.0 × 10−4 7.8 × 10−5 4.4 × 10−7 5.2 × 10−6 2.2 × 10−6 6.1 × 10−6

Criticism

q −8

5.4 × 10 1.8 × 10−10 3.9 × 10−9 2.8 × 10−10 6.7 × 10−10 3.1 × 10−10 0 5.4 × 10−8

1.00 0.003 0.07 0.01 0.01 0.01 0.00

Fig. 6. Spatiotemporal analysis of human health risks caused by vegetable consumption at the three communities surrounding the Obuasi gold mine (Ghana) a) Non-carcinogenic effects b) Carcinogenic effects.

the three regions of interest (Sanso, Pompona and Dokyiwa). The results show that the accumulated values for the Dokyiwa region are almost double the values found in the other two regions. Therefore, residents of these regions undergo the greatest risks to human health. In a short period of time (one year for the residents of Dokyiwa and two years for residents of Sanso and Pompona) the cumulative risk index exceeds the limits of safety for human health. Once again, the importance of spatiotemporal analyses is demonstrated, since it allows a more realistic assessment of risks. A useful issue for the identification and subsequent remediation of the regions most affected by pollution and with greater risks to human health.

Fig. 5. Temporal analysis of human health risk caused by vegetable consumption in the areas surrounding the Obuasi gold mine (Ghana) a) Non-carcinogenic effects (HQ and HIagg) for As, Hg and Pb b) Carcinogenic effects (CR) for As.

carcinogenic effects are exceeded after one and two years of exposure, respectively. Consequently, after two years of consuming contaminated vegetables, residents should already manifest health problems. Considering that there are other pollutants and different exposure routes, we realize the severity of the contamination in the communities around the Obuasi gold mine; and the potential risk to which they are exposed. Since HHRISK allows performing spatiotemporal analysis; a more detailed analysis of the scenario studied by Bempah and Ewusi (2016) was carried out. Fig. 6 shows the cumulative carcinogenic and noncarcinogenic effects due to the ingestion of contaminated vegetables for

5. Conclusion In this work, was described the HHRISK code developed to the human health risks assessment caused by exposure to chemicals. It was demonstrated that the program is a useful tool, easy to use and able to perform calculations in a more secure and accurate way. It is important to emphasize that HHRISK was specially designed to improve conventional methods of risk assessment, by performing spatiotemporal analyses. To achieve this assignment, the program arranges in a matrix the 546

Ecotoxicology and Environmental Safety 170 (2019) 538–547

J.B. Neris et al.

concentration values for each specific sampling point and for different sampling times. The program also allows to changing the exposure time, which is useful for assessing risks in shorter time intervals. In order to show the functionality of the HHRISK code, a test was conducted using the data reported by two research articles. Although there were small differences between some HHRISK values and those reported (mainly due to the difference in the values of some parameters used in the calculation). In general, the results obtained agreed with those reported by the authors. In addition, the spatiotemporal analyzes carried out by HHRISK made it possible to evaluate the risk to human health in a more realistic and detailed way.

Huang, J., Guo, S., Zeng, G., Li, F., Gu, Y., Shi, Y., Shi, L., Liu, W., Peng, S., 2018. A new exploration of health risk assessment quantification from sources of soil heavy metals under different land use. Environ. Pollut. 243, 49–58. https://doi.org/10.1016/j. envpol.2018.08.038. Huang, J., Li, F., Zeng, G., Liu, W., Huang, X., Xiao, Z., Wu, H., Gu, Y., Li, X., He, X., He, Y., 2016. Integrating hierarchical bioavailability and population distribution into potential eco-risk assessment of heavy metals in road dust: a case study in Xiandao District, Changsha city, China. Sci. Total Environ. 541, 969–976. https://doi.org/10. 1016/j.scitotenv.2015.09.139. ISO, 2004. Guide to the Expression of Uncertainty in Measurement (GUM) – Supplement 1: Numerical Methods for the Propagation of Distributions. International Organization forStandardization, Geneva. Järup, L., 2003. Hazards of heavy metal contamination. Br. Med. Bull. 68, 167–182. https://doi.org/10.1093/bmb/ldg032. Junaid, M., Hashmi, M.Z., Malik, R.N., Pei, D.-S., 2016. Toxicity and oxidative stress induced by chromium in workers exposed from different occupational settings around the globe: a review. Environ. Sci. Pollut. Res. 23, 20151–20167. https://doi. org/10.1007/s11356-016-7463-x. La Grega, M.D., Buckingham, P.L., Evans, J.C., 1994. Hazardous Waste Management. McGraw-Hill Inc, New York. Naji, A., Khan, F.R., Hashemi, S.H., 2016. Potential human health risk assessment of trace metals via the consumption of marine fish in Persian Gulf. Mar. Pollut. Bull. 109, 667–671. https://doi.org/10.1016/j.marpolbul.2016.05.002. Newell, C.J., McLeod, R.K., Gonzales, J.R., 1996. BIOSCREEN: Natural Attenuation Decision Support System. User’s Manual Version 1.3. Groundw. Serv. INC, Houst, TX. Pan, L., Ma, J., Hu, Y., Su, B., Fang, G., Wang, Y., Wang, Z., Wang, L., Xiang, B., 2016. Assessments of levels, potential ecological risk, and human health risk of heavy metals in the soils from a typical county in Shanxi Province, China. Environ. Sci. Pollut. Res. 23, 19330–19340. https://doi.org/10.1007/s11356-016-7044-z. Rifai, H.S., Newell, C.J., Gonzales, J.R., Dendrou, S., Kennedy, L., Wilson, J.T., 1997. BIOPLUME III Natural Attenuation Decision Support System, Version 1.0, User’s Manual. Air Force Cent. Environ. Excell. (AFCEE), San Antonio. Rodricks, J.V., Levy, J.I., 2013. Science and decisions: advancing toxicology to advance risk assessment. Toxicol. Sci. 131, 1–8. https://doi.org/10.1093/toxsci/kfs246. Sassi, G., Vernai, A.M., Ruggeri, B., 2007. Quantitative estimation of uncertainty in human risk analysis. J. Hazard. Mater. 145, 296–304. https://doi.org/10.1016/j. jhazmat.2006.11.020. Spence, L.R., Walden, T., 2001. Risk-Integrated Software for Clean-UPS (RISC4). User’s manual, spence Eng. Pleasanton, California/BP Oil International Sunbury. Stewart, R.N., Purucker, S.T., 2006. SADA: A freeware decision support tool integrating GIS, sample design, spatial modeling, and risk assessment. International Congr. Environ. Model. Softw. U.S. EPA, 2016. Integrated Risk Information System - USEPA-IRIS. U. S. Environmental Protection Agency. U.S. EPA, 2014. Cumulative Risk Webinar Series. U.S. Environmental Protection Agency. U.S. EPA, 2012. Cumulative Risk Grants Grantee Progress Review Meeting. U. S. Environmental Protection Agency. U.S. EPA, 2011. Exposure Factors Handbook. U.S. Environmental Protection Agency. U.S. EPA, 2007. Concepts, Methods and data Source for Cumulative Health Risk Assessment of Multiple Chemicals, Exposure and Effects: A Resource Document. U.S. Environmental Protection Agency. U.S. EPA, 2004. Risk Assessment Guidance for Superfund Volume I: Human Health Evaluation Manual (Part E, Supplemental Guidance for Dermal Risk Assessment). U.S. Environmental Protection Agency. U.S. EPA, 1996. Qualitative Uncertainty Analysis of Superfund Residential Risk Pathway Models for Soil and Groundwater: White Paper. U. S. Environmental Protection Agency. U.S. EPA, 1991. Human Health Evaluation Manual, Supplemental Guidance: “Standard Default Exposure Factors”. U.S. Environmental Protection Agency. U.S. EPA, 1989. Risk Assessment Guidance for Superfund. Vol I: Human Health Evaluation Manual (Part A). U.S. Environmental Protection Agency. Vig, N.J., Faure, M.G., Kraft, M.E., 2004. Green Giants?: Environmental Policies of the United States and the European Union. MIT press. Yang, X., Duan, J., Wang, L., Li, W., Guan, J., Beecham, S., Mulcahy, D., 2015. Heavy metal pollution and health risk assessment in the Wei River in China. Environ. Monit. Assess. 187, 111. https://doi.org/10.1007/s10661-014-4202-y. Zheng, C., 1992. MT3D: A modular three-dimensional transport model for simulation of advection, dispersion and chemical reactions of contaminants in groundwater systems. SS Papadopulos Assoc.

Acknowledgments This study was financed in part by the Coordination of Improvement of Higher Education Personnel - Brazil (CAPES) - Finance Code 001. The authors also wish to thank State University of Santa Cruz and The Center for Research in Radiation Sciences and Technologies (CPqCTR) for the technological support. Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at doi:10.1016/j.ecoenv.2018.12.017. References ATSDR, 2000. Prepared by Clement International Corp., under contract, 205, 88-0608, Agency for toxic substances and disease registry. Baes, C.F.I., Sharp, R., Sjoreen, A., Shor, R., 1984. Review and analysis of parameters for assessing transport of environmentally released radionuclides through agriculture. Oak Ridge, TN (United States). https://doi.org/10.2172/6355677. Bempah, C.K., Ewusi, A., 2016. Heavy metals contamination and human health risk assessment around Obuasi gold mine in Ghana. Environ. Monit. Assess. 188, 261. https://doi.org/10.1007/s10661-016-5241-3. CETESB, 2001. Manual de gerenciamento de áreas contaminadas, Companhia de tecnologia de saneamento ambiental. CONAMA, 2009. Resolução n° 420, de 28 de dezembro de 2009. Dispõe sobre critérios e valores orientadores de qualidade do solo quanto à presença de substâncias químicas e estabelece diretrizes para o gerenciamento ambiental de áreas contaminadas por essas substâncias em, Diário Oficial [da] República Federativa do Brasil. Connor, J.A., Bowers, R.L., McHugh, T.E., Spexet, A.H., 2007. Environmental Modeling and Risk Assessment Software. Risk-Based Correct. Action. Correia, L.O., Neris, J.B., Marrocos, P., Velasco, F.G., Luzardo, F.M., Olivares, D.M., Almeida, O.N., de, Santos, H.M., 2016. Bioacumulação de chumbo em plantas de cenoura (Daucus carota) e seus efeitos na saúde humana. Gaia Sci. 10, 301–318. https://doi.org/10.21707/gs.v10.n04a25. Covello, V.T., Merkhoher, M.W., 2013. Risk Assessment Methods: Approaches for Assessing Health and Environmental Risks. Springer Science & Business Media, New York. Dong, Z., Liu, Y., Duan, L., Bekele, D., Naidu, R., 2015. Uncertainties in human health risk assessment of environmental contaminants: a review and perspective. Environ. Int. 85, 120–132. https://doi.org/10.1016/j.envint.2015.09.008. Gonçalves, J.A.C., Lena, J.C. de, 2013. Avaliação de risco à saúde humana por contaminação natural de arsênio nas águas subterrâneas e nos solos da área urbana de Ouro Preto (MG). Geol. USP. Série Científica 13, 145–158. https://doi.org/10. 5327/Z1519-874×2013000200008. Goyer, R.A., Clarkson, T.W., 1996. Toxic effects of metals. Casarett & Doull’s Toxicology: The Basic Science of Poisons. McGraw-Hill Health Professions Division Klaassen. Harbaugh, A.W., McDonald, M.G., 1996. Programmer’s documentation for MODFLOW96, an update to the US Geological Survey modular finite-difference ground-water flow model. US Geol. Surv. Branch Inf. Serv. [distributor].

547