Environ Sci Pollut Res DOI 10.1007/s11356-015-4340-y
RESEARCH ARTICLE
Human exposure and risk assessment associated with mercury contamination in artisanal gold mining areas in the Brazilian Amazon Zuleica Castilhos 1,4 & Saulo Rodrigues-Filho 2 & Ricardo Cesar 3 & Ana Paula Rodrigues 4 & Roberto Villas-Bôas 1 & Iracina de Jesus 5 & Marcelo Lima 5 & Kleber Faial 5 & Antônio Miranda 5 & Edilson Brabo 5 & Christian Beinhoff 6 & Elisabeth Santos 5
Received: 11 August 2014 / Accepted: 9 March 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract Mercury (Hg) contamination is an issue of concern in the Amazon region due to potential health effects associated with Hg exposure in artisanal gold mining areas. The study presents a human health risk assessment associated with Hg vapor inhalation and MeHg-contaminated fish ingestion, as well as Hg determination in urine, blood, and hair, of human populations (about 325 miners and 321 non-miners) from two gold mining areas in the Brazilian Amazon (São Chico and Creporizinho, Pará State). In São Chico and Creporizinho, 73 fish specimens of 13 freshwater species, and 161 specimens of 11 species, were collected for total Hg determination, respectively. The hazard quotient (HQ) is a risk indicator which defines the ratio of the exposure level and the toxicological reference dose and was applied to determine the threat of MeHg exposure. The mean Hg concentrations in fish from São Chico and Creporizinho were 0.83± 0.43 and 0.36±0.33 μg/g, respectively. More than 60 and 22 % of fish collected in São Chico and Creporizinho, respectively, Responsible editor: Philippe Garrigues * Zuleica Castilhos [email protected]
were above the Hg limit (0.5 μg/g) recommended by WHO for human consumption. For all sampling sites, HQ resulted from 1.5 to 28.5, except for the reference area. In Creporizinho, the values of HQ are close to 2 for most sites, whereas in São Chico, there is a hot spot of MeHg contamination in fish (A2—São Chico Reservoir) with the highest risk level (HQ=28) associated with its human consumption. Mean Hg concentrations in urine, blood, and hair samples indicated that the miners group (in São Chico: urine=17.37 μg/L; blood=27.74 μg/L; hair=4.50 μg/g and in Creporizinho: urine=13.75 μg/L; blood=25.23 μg/L; hair: 4.58 μg/g) was more exposed to mercury compared to non-miners (in São Chico: urine = 5.73 μg/L; blood = 16.50 μg/L; hair=3.16 μg/g and in Creporizinho: urine= 3.91 μg/L; blood=21.04 μg/L, hair=1.88 μg/g). These high Hg levels (found not only in miners but also in non-miners who live near the mining areas) are likely to be related to a potential hazard due to exposure to both Hg vapor by inhalation and to MeHg-contaminated fish ingestion. Keywords Mercury . Gold mining . Fish . Human health . Amazon
1
Centre for Mineral Technology, CETEM/MCTI, Rio de Janeiro, RJ, Brazil
Introduction
2
Centro de Desenvolvimento Sustentável (CDS), University of Brasilia, UnB, Brasília, DF, Brazil
3
Department of Geography, Federal University of Rio de Janeiro, UFRJ, Rio de Janeiro, RJ, Brazil
4
Department of Geochemistry, Fluminense Federal University, UFF, Niterói, RJ, Brazil
5
Evandro Chagas Institute, Belém, PA, Brazil
6
United Nations Industrial Development Organization, UNIDO, Vienna International Centre, Vienna, Austria
Over the last decades, mercury (Hg) contamination has become an issue of concern due to its diverse damage functions on human health and biota (WHO 1990). The main source of Hg pollution is related to its inadequate use in artisanal gold mining areas, when Hg is used for processing gold alluvial deposits. During such processes, Hg can escape to the atmosphere and thus be deposited on surrounding soils and aquatic ecosystems (Rodrigues-Filho and Maddock 1997; Cesar et al. 2011). Moreover, leaching and soil erosion processes can also
Environ Sci Pollut Res
mobilize mercury to aquatic systems (Cesar et al. 2011). Once in the aquatic systems, Hg can be biotransformed by bacteria into MeHg, especially in anoxic bottom sediments (Ullrich et al. 2001), thus reaching top predators through its bioaccumulation and biomagnification in the food chain. Biomagnification is defined as the increase of contaminant levels by food ingestion (Bruggeman 1982). As a consequence, top predators are expected to exhibit higher mercury concentrations compared to non-carnivorous species. Given the fact that MeHg is much more toxic than the Hg inorganic forms and it biomagnifies by aquatic organisms, MeHg environmental exposure via fish consumption has become an issue of concern in areas affected by Hg pollution (Castilhos et al. 1998, 2006; Correia et al. 2014) besides the Hg vapor occupational exposure. Human exposure to Hg vapor may cause damages to the respiratory and neurological systems. Several authors traditionally use urine samples as an indicator of Hg vapor exposure (Bell et al. 1973; Tsuji et al. 2003; Engström et al. 2013), while blood is widely used as an indicator of a recent exposure (Steckling et al. 2011; Harari et al. 2012). MeHg intoxication is generally chronic and often involves the occurrence of neurotoxic and teratogenic effects, whose damages are irreversible (WHO 1990). Hair samples are commonly used as an indicator of MeHg exposure (Hacon et al. 2000; Santos et al. 2000; Freire et al. 2010). In 2002, the United Nations Industrial Development Organization (UNIDO) launched the Global Mercury Project (GMP), an effort for stimulating the adoption of environmentally cleaner technologies in artisanal gold mining (AGM) (Rodrigues-Filho et al. 2004; Castilhos et al. 2006). Financed by the Global Environment Facility and managed by UNIDO, the GMP has started projects in six countries: Indonesia, Zimbabwe, Tanzania, Sudan, Laos, and Brazil— in the latest research activities encompassing two sites (São Chico and Creporizinho) from the Tapajós Gold Mining Reservoir (Pará State, Brazilian Amazon). Its main objective is to evaluate the magnitude of Hg contamination in humans and in fish from areas affected by gold mining activities in order to perform the human health risk assessment. Our working hypotheses are the following: (i) small-scale gold mining is a source of Hg pollution to aquatic systems; (ii) living close to mining sites is a risk due to Hg vapor exposure; and (iii) miners and non-miners are at risk due to the consumption of Hg-contaminated fish and Hg vapor inhalation.
Materials and methods
Transgarimpeira road (06° 24′ 58.6″ S–55° 58′ 00.0″ W), which only during the dry season (June–September) can be used for transportation. The São Chico village consists of only 63 houses and 150 individuals (41 % of artisanal miners). The village is located no more than few meters far from the mining area. The main mining area lies at the slope of a valley, between the village and a dam reservoir (A2 sampling point), which supplies the pumps for mining and milling activities with water. In the same process, material from primary deposits is mixed with residues of former mining activities and milled. The gold is concentrated on mercury-coated copper plates. From the beginning of the very first gold mining activity in 1963, the São Chico village has shown two main periods of prosperity: one in the end of the 1980s after the opening of the Transgarimpeira road and the other in the end of the 1990s, when gold-rich primary deposits were discovered. According to cross-checked estimations, about 2 tons of gold was produced during the last gold rush, which corresponds to an estimated mercury emission of 5 tons to the environment. It should be noted that miners used an artisanal cyanidation unit located close to the dam. At the time of this study, the unit was abandoned with continuous leaching to the dam. Since the primary gold ore has been crushed in hammer mills and directly amalgamated in copper plates, and no retorts have been used, the estimated Hg:Au (lost/produced) emission ratio is about 2.5 for this type of operation. Creporizinho (06° 30′ 17″ S and 56° 35′ 06″ W), located close to the Transgarimpeira road (06° 50′ 14.1″ S–56° 35′ 00.0″ W), is a typical gold mining village with 238 wooden residences for an estimated population of 1000 inhabitants. It is around 20 km far from the mining sites. Creporizinho mining sites have started their prospective/extraction activities since 1968 (artisanal extraction). In 1985, the miners started to work with some machinery for processing alluvial–colluvial terraces. Their ore processing technique consists of hydraulic jet pumps coupled with riffled or carpeted sluice boxes in order to concentrate gold prior to amalgamation. The heavy Hg–gold amalgam forms sinks and is retained behind the riffles or in carpet fibers. These operations use large amounts of Hg, and consequently, high contents of this metal are often discharged into the tailings. The estimated Hg:Au is 5, i.e., extremely high. The Hg emission scenario was estimated to be in the range of about 60–80 g Hg/day, if the residues were discharged over the environment. The control area (A11 sampling point, in Creporizinho) is that not directly submitted to or influenced by gold mining activities. It is located far from the gold mining area, but it makes part of the same hydrographic basin.
Study area
Local human populations
The São Chico mining site (06° 25′ 31″ S and 56° 02′ 99″ W) is just 2 km away from a landing strip and 5 km distance from the
The sampling of human biological material in Sao Chico and Creporizinho was approved by the Brazilian National Ethic
Environ Sci Pollut Res
Committee (July 2003). The field work consisted of several procedures including filling out an epidemiological questionnaire, the collecting of biological material (urine, blood, and hair) for Hg analysis, parasitological tests in feces samples, and some clinical and neurological tests. Hg concentrations in biomonitors (urine, blood, and hair samples) are presented and discussed in this article. The other data are described in Rodrigues-Filho et al. (2004). The total human population was composed of 697 individuals: 246 (151 males and 95 females) from São Chico and 451 (294 males and 157 females) from Creporizinho. Whenever possible and according to the consent from the population, urine, blood, and hair samples were collected from some of those individuals. The education level of these individuals is extremely low (less than 3 years of formal school). Most individuals (45.9 and 56.5 % in São Chico and Creporizinho, respectively) are 31–50 years old. About 20 and 7 % of the population are children up to 10 years old in São Chico and Creporizinho, respectively. Thus, children are a significant part of the non-miners group in São Chico. Most of the studied population is immigrant from northeastern Brazil, especially from Maranhão State (about 43 and 48 % of the individuals from São Chico and Creporizinho, respectively). About 43.1 and 51.2 % of local populations from São Chico and Creporizinho, respectively, are miners. Miners are almost exclusively male. Other important occupations include shop workers (sale of gold), cooks, housewives, cattle farmers, and mechanics (Rodrigues-Filho et al. 2004). In both mining areas, more than 80 % of the miners have been using Hg in their activities for, on average, 15 years, by burning gold–Hg amalgam in open recipients. Many of them reported the use of retorts for more than 10 years. More than 70 % of the population reported past cases of malaria, and this is a potential confounder parameter when analyzing health effects due to mercury exposure. Other important diseases include tuberculosis, hanseniasis, and hepatitis of unspecified origin. Human biomarkers sampling Human blood sampling was accomplished by an outlying blood vessel, with disposable syringes of 10 mL whose content was preserved in flasks containing EDTA 10 %. Blood samples were collected from 106 miners and 128 non-miners. In respect to urine samples, the participants (104 miners and 129 non-miners) were guided to perform the collection of the first urine in the morning in polyethylene flasks, which were previously washed with diluted nitric acid. The procedure for creatinine determination in urine follows the method described in Slot (1965), and the values were used to adjust Hg contents. Blood samples were collected from 106 and 219 miners and 128 and 193 non-miners from São Chico and Creporizinho, respectively.
Hair samples were taken from different areas of the scalp (including at last 100 strands for each area), cropped about 1 cm off the scalp with stainless steel scissors, and stored in white envelops at room temperature for later analysis. To avoid degradation, urine and blood samples were stored continuously at 4 °C until analyses. All the samples were delivered to the mercury lab at the Evandro Chagas Institute for Hg levels determination. The samples collected from the non-miners group were considered as an internal control for comparison with those from the miners group. This is relevant to detect different exposure pathways, including contaminated fish ingestion and Hg vapor. Significant differences between Hg concentrations in urine, blood, and hair among miners and non-miners were evaluated by Student’s t test. Fish sampling First of all, sampling fish close to mining areas is really difficult due to the high turbidity caused by intense soil erosion from gold mining processes. Fish evasion of these areas is common. The sampling effort depends on the sampling point and, as expected, is much higher in sites close to the mining areas. Fish sampling was conducted in August 2003, between the cities of Jacareacanga and Itaituba (Mineral Tapajós Reservoir, Pará State). The São Chico and Creporizinho mining sites are distributed alongside two areas located at two distinct sub-basins: Jamanxin River and Crepori River, respectively, which are tributaries of the Tapajós River. Table 1 shows the sampling areas in both mining sites. Fish samples were collected in 14 sites: 4 in the São Chico and 7 in Creporizinho. Two hundred thirty-four specimens of 17 species were sampled; the following are their popular names: acari (15), arraia (3), bocudo (1), candiru (4), cará (67), curimatã (14), joão-duro (40), joão-preto (10), lambari (4), lampréia (5), mandi (4), pacuí (1), piau (16), piranha (18), pirarara (1), surubim (1), and traíra (30). In the São Chico and Creporizinho mining areas, 73 specimens of 13 species and 161 specimens of 11 species, respectively, were collected. Some species were captured in several sites in reasonable numbers, such as traíra and piranha, which are classified as carnivorous species. Others, such as piau and acari, feed on benthos and plankton (Castilhos et al. 2004). Fish samples were collected by gill netting, fishing line with a fish hook, and fishing line with several hooks (espinhel). Each specimen was weighed, and its length was measured immediately after the sampling. After removing the individual axial muscle (fillet), each sample was placed in polyethylene bags and frozen when the team came back to the small villages (São Chico and Creporizinho) (Castilhos et al. 2004). The samples were delivered to the mercury lab at CETEM for total Hg determinations in the axial muscles of fish. Significant differences between Hg concentrations in fish
Environ Sci Pollut Res Table 1 Brief description of sampling sites in the São Chico (01-04) and Creporizinho (05-11) mining sites Study site A1 A2 A3 A4
Flooded open pit, clear water, near Conrado River São Chico dam Flooded open pit, mining wastes, high turbidity, near Rosa stream Conrado River inflow to Novo River
A5 A6 A7 A8
Papagaio mining site; stream with high turbidity Flooded open pit at Tabocal (Tabocal mining site) Bofe mining site Buriti mining site; recent flooded open pit, near Creporizinho River spring A9 Porto Alegre site in Crepori River, upstream of Creporização village A10 Crepori River upstream of Creporizinho River inflow A11 Chico Chimango creek, clear water near the Creporizinho inflow into the Crepori river
from different sampling points were evaluated by Student’s t test.
Total mercury determination Mercury determination in fish muscle follows the method developed by Akagi and Nishimura (1991). Two milliliters of an oxidant 1:1 acid mixture (HClO4:HNO3) is added dropwise into 0.5 g of sample, previously weighted in a 50-mL volumetric flask. Then, 1 mL of water is added. The mixture is heated to 230–250 °C for 20 min. After cooling, the solution is diluted up to 50 mL, and samples were analyzed at least twice to guarantee result quality. For quality assurance/quality control sets, a standard reference material (dogfish, National Research Council Canada) was determined for each set of 10 unknown samples. All results obtained from reference material analysis fall into the expected interval. The procedures used for Hg determination in urine, blood, and hair samples follow Akagi (1998) and are similar to that applied to fish, except the acid mixture used for sample digestion. For urine digestion, 2 mL of an acid mixture composed of HNO3:HCLO4 (1:1) and 5 mL of H2SO4 are added into 5 mL of sample. The same procedure applied to urine was used for blood and hair samples, except the volume or mass of the sample: 5 mL of urine and 100–200 mg of hair. Hair samples were previously washed with neutral detergent and distilled water by decantation, and washed again with a small amount of acetone to remove water residues. This procedure is important to remove Hg deposited on the hair strand. The Hg limit of detection for fish, urine, blood, and hair samples is 0.5 ppb (i.e., 0.5 ng/g and 0.5 μg/L for solid and liquid matrices, respectively).
Human health risk assessment To establish whether these fish endanger human health, the risk assessment index known as the hazard quotient (USEPA 1989) was applied. In this study, a screening approach was applied. Even though such approach may be simplistic, especially for MeHg, it allows easy comparison between populations exposed to different doses of a given toxicant. Thus, at a screening level, the hazard quotient (HQ) assumes that there is a level of exposure (RfD = reference dose) for noncarcinogenic substances below which it is unlikely for populations to experience adverse health effects (USEPA 1989). Therefore, HQ is defined as the ratio between the MeHg exposure level to the RfD. When HQ exceeds unity, there may be concern for potential health effects. The MeHg RfD value is 1E − 04 mg/kg/day (USEPA 2014). Its uncertainty factor is 10 and its confidence level is medium. The estimated MeHg intake dose at a screening level is based on the product of the upper bound of the arithmetical mean Hg concentration in fish and the adult human ingestion rate for local populations. Akagi et al. (1994) suggest that over 90 % of Hg present in fish from the Tapajós River basin is MeHg. Then, the total Hg measured in fish was used to estimate MeHg intake. Given the fact that most of the miners population are immigrants from northeastern Brazil and that they do not generally consume fish very often compared to Amazonian riverine populations, we suggested a low fish ingestion rate. This strategy is in accordance with information from the epidemiological questionnaire, which has shown that in São Chico and Creporizinho areas, close to 60 % of the population consume fish just once a week and 90 % consume fish twice weekly. Therefore, a consumption rate close to 0.02 kg/day has been considered reasonable for both communities. The MeHg exposure level was defined by the ratio between the intake dose and the estimated weight of the exposed population (70 kg for adults). A single-compartment model (WHO 1990), previously applied by Bidone et al. (1997) in the Brazilian Amazon and by Castilhos et al. (2006) in Indonesia, was used to estimated Hg blood and hair concentrations. Such single-compartment model calculates the steady-state Hg concentration in blood (C) in micrograms per liter. This is related to the average daily dietary intake (d in μg of Hg), as follows: C=0.95d. Mercury concentration in hair is proportional to blood concentrations at the time of the formation of the hair strand. The blood–hair ratio in humans is about 250, but appreciable individual differences have been found (WHO 1990). When considering the processes for artisanal gold extraction, the exposure to liquid metallic Hg through the skin is definitely lower compared to the potential inhalation of Hg vapor. Also, high environmental temperatures (>30 °C) in the Amazon region may increase the concentrations of metallic vapor Hg in the atmosphere. Finally, some studies indicate
Environ Sci Pollut Res
that the uptake of metallic mercury vapor through the skin is only about 1 % compared to the concentration taken up by inhalation (WHO 1990). Therefore, the main pathways of mercury exposure in small-scale gold mining areas are the inhalation of metallic mercury vapor and the ingestion of MeHg-contaminated fish.
Results The results show that total Hg concentration in fish from São Chico (2.53±3.91 μg/g, n=73) is higher than in fish from Creporizinho area (0.36 ± 0.33 μg/g, n = 161; p < 0.0001) (Table 2). The minimum values for Hg in fish are similar between areas (0.025–0.027 μg/g), but the maximum values are one order of magnitude higher in São Chico (21.90 μg/g) than in Creporizinho area (2.10 μg/g) (Tables 2, 3, and 4). Among the analyzed fish, 35 % of total sampled fish (from Table 2
six different species) showed Hg concentrations above 0.5 μg/ g, the limit for human health protection against Hg exposure by fish consumption (WHO 1990). Whereas in Creporizinho 22 % of fish samples showed Hg levels above that limit, in São Chico, this percentage increases to more than 60 %. The results show that fish from São Chico are more contaminated, heavier, and larger than those from Creporizinho area (Table 2). However, no correlation was found between fish Hg concentration and weight or length when analyzing all specimens in both areas. Carnivorous species revealed a mean Hg concentration (4.16 μg/g) almost four times higher than that of the non-carnivorous species (1.33 μg/g) (Table 2). Hoplias malabaricus was the most contaminated species in both mining sites (Tables 2, 3, and 4). Fish samples from the São Chico area (especially the specimens collected in the A2 sampling point) exhibit the highest Hg levels and are smaller than fish from the other São Chico sites (A1, A3, and A4) (Table 3). In Creporizinho area, fish from site A8 showed
Total mercury concentrations (arithmetical mean±standard deviation) in different fish species (ww) according to their food habits
Popular name
Carnivorous Arraia Bocudo Lambari Piranha Pirarara Surubim Traíra** Non-carnivorous Detritivorous Acari Curimatã Herbivorous Pacu Piau Insectivorous Ituí Macrofagous Sairu Microfagous Cará* Omnivorous Candiru Mandi Total Student’s t test n number of specimens *p
RESEARCH ARTICLE
Human exposure and risk assessment associated with mercury contamination in artisanal gold mining areas in the Brazilian Amazon Zuleica Castilhos 1,4 & Saulo Rodrigues-Filho 2 & Ricardo Cesar 3 & Ana Paula Rodrigues 4 & Roberto Villas-Bôas 1 & Iracina de Jesus 5 & Marcelo Lima 5 & Kleber Faial 5 & Antônio Miranda 5 & Edilson Brabo 5 & Christian Beinhoff 6 & Elisabeth Santos 5
Received: 11 August 2014 / Accepted: 9 March 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract Mercury (Hg) contamination is an issue of concern in the Amazon region due to potential health effects associated with Hg exposure in artisanal gold mining areas. The study presents a human health risk assessment associated with Hg vapor inhalation and MeHg-contaminated fish ingestion, as well as Hg determination in urine, blood, and hair, of human populations (about 325 miners and 321 non-miners) from two gold mining areas in the Brazilian Amazon (São Chico and Creporizinho, Pará State). In São Chico and Creporizinho, 73 fish specimens of 13 freshwater species, and 161 specimens of 11 species, were collected for total Hg determination, respectively. The hazard quotient (HQ) is a risk indicator which defines the ratio of the exposure level and the toxicological reference dose and was applied to determine the threat of MeHg exposure. The mean Hg concentrations in fish from São Chico and Creporizinho were 0.83± 0.43 and 0.36±0.33 μg/g, respectively. More than 60 and 22 % of fish collected in São Chico and Creporizinho, respectively, Responsible editor: Philippe Garrigues * Zuleica Castilhos [email protected]
were above the Hg limit (0.5 μg/g) recommended by WHO for human consumption. For all sampling sites, HQ resulted from 1.5 to 28.5, except for the reference area. In Creporizinho, the values of HQ are close to 2 for most sites, whereas in São Chico, there is a hot spot of MeHg contamination in fish (A2—São Chico Reservoir) with the highest risk level (HQ=28) associated with its human consumption. Mean Hg concentrations in urine, blood, and hair samples indicated that the miners group (in São Chico: urine=17.37 μg/L; blood=27.74 μg/L; hair=4.50 μg/g and in Creporizinho: urine=13.75 μg/L; blood=25.23 μg/L; hair: 4.58 μg/g) was more exposed to mercury compared to non-miners (in São Chico: urine = 5.73 μg/L; blood = 16.50 μg/L; hair=3.16 μg/g and in Creporizinho: urine= 3.91 μg/L; blood=21.04 μg/L, hair=1.88 μg/g). These high Hg levels (found not only in miners but also in non-miners who live near the mining areas) are likely to be related to a potential hazard due to exposure to both Hg vapor by inhalation and to MeHg-contaminated fish ingestion. Keywords Mercury . Gold mining . Fish . Human health . Amazon
1
Centre for Mineral Technology, CETEM/MCTI, Rio de Janeiro, RJ, Brazil
Introduction
2
Centro de Desenvolvimento Sustentável (CDS), University of Brasilia, UnB, Brasília, DF, Brazil
3
Department of Geography, Federal University of Rio de Janeiro, UFRJ, Rio de Janeiro, RJ, Brazil
4
Department of Geochemistry, Fluminense Federal University, UFF, Niterói, RJ, Brazil
5
Evandro Chagas Institute, Belém, PA, Brazil
6
United Nations Industrial Development Organization, UNIDO, Vienna International Centre, Vienna, Austria
Over the last decades, mercury (Hg) contamination has become an issue of concern due to its diverse damage functions on human health and biota (WHO 1990). The main source of Hg pollution is related to its inadequate use in artisanal gold mining areas, when Hg is used for processing gold alluvial deposits. During such processes, Hg can escape to the atmosphere and thus be deposited on surrounding soils and aquatic ecosystems (Rodrigues-Filho and Maddock 1997; Cesar et al. 2011). Moreover, leaching and soil erosion processes can also
Environ Sci Pollut Res
mobilize mercury to aquatic systems (Cesar et al. 2011). Once in the aquatic systems, Hg can be biotransformed by bacteria into MeHg, especially in anoxic bottom sediments (Ullrich et al. 2001), thus reaching top predators through its bioaccumulation and biomagnification in the food chain. Biomagnification is defined as the increase of contaminant levels by food ingestion (Bruggeman 1982). As a consequence, top predators are expected to exhibit higher mercury concentrations compared to non-carnivorous species. Given the fact that MeHg is much more toxic than the Hg inorganic forms and it biomagnifies by aquatic organisms, MeHg environmental exposure via fish consumption has become an issue of concern in areas affected by Hg pollution (Castilhos et al. 1998, 2006; Correia et al. 2014) besides the Hg vapor occupational exposure. Human exposure to Hg vapor may cause damages to the respiratory and neurological systems. Several authors traditionally use urine samples as an indicator of Hg vapor exposure (Bell et al. 1973; Tsuji et al. 2003; Engström et al. 2013), while blood is widely used as an indicator of a recent exposure (Steckling et al. 2011; Harari et al. 2012). MeHg intoxication is generally chronic and often involves the occurrence of neurotoxic and teratogenic effects, whose damages are irreversible (WHO 1990). Hair samples are commonly used as an indicator of MeHg exposure (Hacon et al. 2000; Santos et al. 2000; Freire et al. 2010). In 2002, the United Nations Industrial Development Organization (UNIDO) launched the Global Mercury Project (GMP), an effort for stimulating the adoption of environmentally cleaner technologies in artisanal gold mining (AGM) (Rodrigues-Filho et al. 2004; Castilhos et al. 2006). Financed by the Global Environment Facility and managed by UNIDO, the GMP has started projects in six countries: Indonesia, Zimbabwe, Tanzania, Sudan, Laos, and Brazil— in the latest research activities encompassing two sites (São Chico and Creporizinho) from the Tapajós Gold Mining Reservoir (Pará State, Brazilian Amazon). Its main objective is to evaluate the magnitude of Hg contamination in humans and in fish from areas affected by gold mining activities in order to perform the human health risk assessment. Our working hypotheses are the following: (i) small-scale gold mining is a source of Hg pollution to aquatic systems; (ii) living close to mining sites is a risk due to Hg vapor exposure; and (iii) miners and non-miners are at risk due to the consumption of Hg-contaminated fish and Hg vapor inhalation.
Materials and methods
Transgarimpeira road (06° 24′ 58.6″ S–55° 58′ 00.0″ W), which only during the dry season (June–September) can be used for transportation. The São Chico village consists of only 63 houses and 150 individuals (41 % of artisanal miners). The village is located no more than few meters far from the mining area. The main mining area lies at the slope of a valley, between the village and a dam reservoir (A2 sampling point), which supplies the pumps for mining and milling activities with water. In the same process, material from primary deposits is mixed with residues of former mining activities and milled. The gold is concentrated on mercury-coated copper plates. From the beginning of the very first gold mining activity in 1963, the São Chico village has shown two main periods of prosperity: one in the end of the 1980s after the opening of the Transgarimpeira road and the other in the end of the 1990s, when gold-rich primary deposits were discovered. According to cross-checked estimations, about 2 tons of gold was produced during the last gold rush, which corresponds to an estimated mercury emission of 5 tons to the environment. It should be noted that miners used an artisanal cyanidation unit located close to the dam. At the time of this study, the unit was abandoned with continuous leaching to the dam. Since the primary gold ore has been crushed in hammer mills and directly amalgamated in copper plates, and no retorts have been used, the estimated Hg:Au (lost/produced) emission ratio is about 2.5 for this type of operation. Creporizinho (06° 30′ 17″ S and 56° 35′ 06″ W), located close to the Transgarimpeira road (06° 50′ 14.1″ S–56° 35′ 00.0″ W), is a typical gold mining village with 238 wooden residences for an estimated population of 1000 inhabitants. It is around 20 km far from the mining sites. Creporizinho mining sites have started their prospective/extraction activities since 1968 (artisanal extraction). In 1985, the miners started to work with some machinery for processing alluvial–colluvial terraces. Their ore processing technique consists of hydraulic jet pumps coupled with riffled or carpeted sluice boxes in order to concentrate gold prior to amalgamation. The heavy Hg–gold amalgam forms sinks and is retained behind the riffles or in carpet fibers. These operations use large amounts of Hg, and consequently, high contents of this metal are often discharged into the tailings. The estimated Hg:Au is 5, i.e., extremely high. The Hg emission scenario was estimated to be in the range of about 60–80 g Hg/day, if the residues were discharged over the environment. The control area (A11 sampling point, in Creporizinho) is that not directly submitted to or influenced by gold mining activities. It is located far from the gold mining area, but it makes part of the same hydrographic basin.
Study area
Local human populations
The São Chico mining site (06° 25′ 31″ S and 56° 02′ 99″ W) is just 2 km away from a landing strip and 5 km distance from the
The sampling of human biological material in Sao Chico and Creporizinho was approved by the Brazilian National Ethic
Environ Sci Pollut Res
Committee (July 2003). The field work consisted of several procedures including filling out an epidemiological questionnaire, the collecting of biological material (urine, blood, and hair) for Hg analysis, parasitological tests in feces samples, and some clinical and neurological tests. Hg concentrations in biomonitors (urine, blood, and hair samples) are presented and discussed in this article. The other data are described in Rodrigues-Filho et al. (2004). The total human population was composed of 697 individuals: 246 (151 males and 95 females) from São Chico and 451 (294 males and 157 females) from Creporizinho. Whenever possible and according to the consent from the population, urine, blood, and hair samples were collected from some of those individuals. The education level of these individuals is extremely low (less than 3 years of formal school). Most individuals (45.9 and 56.5 % in São Chico and Creporizinho, respectively) are 31–50 years old. About 20 and 7 % of the population are children up to 10 years old in São Chico and Creporizinho, respectively. Thus, children are a significant part of the non-miners group in São Chico. Most of the studied population is immigrant from northeastern Brazil, especially from Maranhão State (about 43 and 48 % of the individuals from São Chico and Creporizinho, respectively). About 43.1 and 51.2 % of local populations from São Chico and Creporizinho, respectively, are miners. Miners are almost exclusively male. Other important occupations include shop workers (sale of gold), cooks, housewives, cattle farmers, and mechanics (Rodrigues-Filho et al. 2004). In both mining areas, more than 80 % of the miners have been using Hg in their activities for, on average, 15 years, by burning gold–Hg amalgam in open recipients. Many of them reported the use of retorts for more than 10 years. More than 70 % of the population reported past cases of malaria, and this is a potential confounder parameter when analyzing health effects due to mercury exposure. Other important diseases include tuberculosis, hanseniasis, and hepatitis of unspecified origin. Human biomarkers sampling Human blood sampling was accomplished by an outlying blood vessel, with disposable syringes of 10 mL whose content was preserved in flasks containing EDTA 10 %. Blood samples were collected from 106 miners and 128 non-miners. In respect to urine samples, the participants (104 miners and 129 non-miners) were guided to perform the collection of the first urine in the morning in polyethylene flasks, which were previously washed with diluted nitric acid. The procedure for creatinine determination in urine follows the method described in Slot (1965), and the values were used to adjust Hg contents. Blood samples were collected from 106 and 219 miners and 128 and 193 non-miners from São Chico and Creporizinho, respectively.
Hair samples were taken from different areas of the scalp (including at last 100 strands for each area), cropped about 1 cm off the scalp with stainless steel scissors, and stored in white envelops at room temperature for later analysis. To avoid degradation, urine and blood samples were stored continuously at 4 °C until analyses. All the samples were delivered to the mercury lab at the Evandro Chagas Institute for Hg levels determination. The samples collected from the non-miners group were considered as an internal control for comparison with those from the miners group. This is relevant to detect different exposure pathways, including contaminated fish ingestion and Hg vapor. Significant differences between Hg concentrations in urine, blood, and hair among miners and non-miners were evaluated by Student’s t test. Fish sampling First of all, sampling fish close to mining areas is really difficult due to the high turbidity caused by intense soil erosion from gold mining processes. Fish evasion of these areas is common. The sampling effort depends on the sampling point and, as expected, is much higher in sites close to the mining areas. Fish sampling was conducted in August 2003, between the cities of Jacareacanga and Itaituba (Mineral Tapajós Reservoir, Pará State). The São Chico and Creporizinho mining sites are distributed alongside two areas located at two distinct sub-basins: Jamanxin River and Crepori River, respectively, which are tributaries of the Tapajós River. Table 1 shows the sampling areas in both mining sites. Fish samples were collected in 14 sites: 4 in the São Chico and 7 in Creporizinho. Two hundred thirty-four specimens of 17 species were sampled; the following are their popular names: acari (15), arraia (3), bocudo (1), candiru (4), cará (67), curimatã (14), joão-duro (40), joão-preto (10), lambari (4), lampréia (5), mandi (4), pacuí (1), piau (16), piranha (18), pirarara (1), surubim (1), and traíra (30). In the São Chico and Creporizinho mining areas, 73 specimens of 13 species and 161 specimens of 11 species, respectively, were collected. Some species were captured in several sites in reasonable numbers, such as traíra and piranha, which are classified as carnivorous species. Others, such as piau and acari, feed on benthos and plankton (Castilhos et al. 2004). Fish samples were collected by gill netting, fishing line with a fish hook, and fishing line with several hooks (espinhel). Each specimen was weighed, and its length was measured immediately after the sampling. After removing the individual axial muscle (fillet), each sample was placed in polyethylene bags and frozen when the team came back to the small villages (São Chico and Creporizinho) (Castilhos et al. 2004). The samples were delivered to the mercury lab at CETEM for total Hg determinations in the axial muscles of fish. Significant differences between Hg concentrations in fish
Environ Sci Pollut Res Table 1 Brief description of sampling sites in the São Chico (01-04) and Creporizinho (05-11) mining sites Study site A1 A2 A3 A4
Flooded open pit, clear water, near Conrado River São Chico dam Flooded open pit, mining wastes, high turbidity, near Rosa stream Conrado River inflow to Novo River
A5 A6 A7 A8
Papagaio mining site; stream with high turbidity Flooded open pit at Tabocal (Tabocal mining site) Bofe mining site Buriti mining site; recent flooded open pit, near Creporizinho River spring A9 Porto Alegre site in Crepori River, upstream of Creporização village A10 Crepori River upstream of Creporizinho River inflow A11 Chico Chimango creek, clear water near the Creporizinho inflow into the Crepori river
from different sampling points were evaluated by Student’s t test.
Total mercury determination Mercury determination in fish muscle follows the method developed by Akagi and Nishimura (1991). Two milliliters of an oxidant 1:1 acid mixture (HClO4:HNO3) is added dropwise into 0.5 g of sample, previously weighted in a 50-mL volumetric flask. Then, 1 mL of water is added. The mixture is heated to 230–250 °C for 20 min. After cooling, the solution is diluted up to 50 mL, and samples were analyzed at least twice to guarantee result quality. For quality assurance/quality control sets, a standard reference material (dogfish, National Research Council Canada) was determined for each set of 10 unknown samples. All results obtained from reference material analysis fall into the expected interval. The procedures used for Hg determination in urine, blood, and hair samples follow Akagi (1998) and are similar to that applied to fish, except the acid mixture used for sample digestion. For urine digestion, 2 mL of an acid mixture composed of HNO3:HCLO4 (1:1) and 5 mL of H2SO4 are added into 5 mL of sample. The same procedure applied to urine was used for blood and hair samples, except the volume or mass of the sample: 5 mL of urine and 100–200 mg of hair. Hair samples were previously washed with neutral detergent and distilled water by decantation, and washed again with a small amount of acetone to remove water residues. This procedure is important to remove Hg deposited on the hair strand. The Hg limit of detection for fish, urine, blood, and hair samples is 0.5 ppb (i.e., 0.5 ng/g and 0.5 μg/L for solid and liquid matrices, respectively).
Human health risk assessment To establish whether these fish endanger human health, the risk assessment index known as the hazard quotient (USEPA 1989) was applied. In this study, a screening approach was applied. Even though such approach may be simplistic, especially for MeHg, it allows easy comparison between populations exposed to different doses of a given toxicant. Thus, at a screening level, the hazard quotient (HQ) assumes that there is a level of exposure (RfD = reference dose) for noncarcinogenic substances below which it is unlikely for populations to experience adverse health effects (USEPA 1989). Therefore, HQ is defined as the ratio between the MeHg exposure level to the RfD. When HQ exceeds unity, there may be concern for potential health effects. The MeHg RfD value is 1E − 04 mg/kg/day (USEPA 2014). Its uncertainty factor is 10 and its confidence level is medium. The estimated MeHg intake dose at a screening level is based on the product of the upper bound of the arithmetical mean Hg concentration in fish and the adult human ingestion rate for local populations. Akagi et al. (1994) suggest that over 90 % of Hg present in fish from the Tapajós River basin is MeHg. Then, the total Hg measured in fish was used to estimate MeHg intake. Given the fact that most of the miners population are immigrants from northeastern Brazil and that they do not generally consume fish very often compared to Amazonian riverine populations, we suggested a low fish ingestion rate. This strategy is in accordance with information from the epidemiological questionnaire, which has shown that in São Chico and Creporizinho areas, close to 60 % of the population consume fish just once a week and 90 % consume fish twice weekly. Therefore, a consumption rate close to 0.02 kg/day has been considered reasonable for both communities. The MeHg exposure level was defined by the ratio between the intake dose and the estimated weight of the exposed population (70 kg for adults). A single-compartment model (WHO 1990), previously applied by Bidone et al. (1997) in the Brazilian Amazon and by Castilhos et al. (2006) in Indonesia, was used to estimated Hg blood and hair concentrations. Such single-compartment model calculates the steady-state Hg concentration in blood (C) in micrograms per liter. This is related to the average daily dietary intake (d in μg of Hg), as follows: C=0.95d. Mercury concentration in hair is proportional to blood concentrations at the time of the formation of the hair strand. The blood–hair ratio in humans is about 250, but appreciable individual differences have been found (WHO 1990). When considering the processes for artisanal gold extraction, the exposure to liquid metallic Hg through the skin is definitely lower compared to the potential inhalation of Hg vapor. Also, high environmental temperatures (>30 °C) in the Amazon region may increase the concentrations of metallic vapor Hg in the atmosphere. Finally, some studies indicate
Environ Sci Pollut Res
that the uptake of metallic mercury vapor through the skin is only about 1 % compared to the concentration taken up by inhalation (WHO 1990). Therefore, the main pathways of mercury exposure in small-scale gold mining areas are the inhalation of metallic mercury vapor and the ingestion of MeHg-contaminated fish.
Results The results show that total Hg concentration in fish from São Chico (2.53±3.91 μg/g, n=73) is higher than in fish from Creporizinho area (0.36 ± 0.33 μg/g, n = 161; p < 0.0001) (Table 2). The minimum values for Hg in fish are similar between areas (0.025–0.027 μg/g), but the maximum values are one order of magnitude higher in São Chico (21.90 μg/g) than in Creporizinho area (2.10 μg/g) (Tables 2, 3, and 4). Among the analyzed fish, 35 % of total sampled fish (from Table 2
six different species) showed Hg concentrations above 0.5 μg/ g, the limit for human health protection against Hg exposure by fish consumption (WHO 1990). Whereas in Creporizinho 22 % of fish samples showed Hg levels above that limit, in São Chico, this percentage increases to more than 60 %. The results show that fish from São Chico are more contaminated, heavier, and larger than those from Creporizinho area (Table 2). However, no correlation was found between fish Hg concentration and weight or length when analyzing all specimens in both areas. Carnivorous species revealed a mean Hg concentration (4.16 μg/g) almost four times higher than that of the non-carnivorous species (1.33 μg/g) (Table 2). Hoplias malabaricus was the most contaminated species in both mining sites (Tables 2, 3, and 4). Fish samples from the São Chico area (especially the specimens collected in the A2 sampling point) exhibit the highest Hg levels and are smaller than fish from the other São Chico sites (A1, A3, and A4) (Table 3). In Creporizinho area, fish from site A8 showed
Total mercury concentrations (arithmetical mean±standard deviation) in different fish species (ww) according to their food habits
Popular name
Carnivorous Arraia Bocudo Lambari Piranha Pirarara Surubim Traíra** Non-carnivorous Detritivorous Acari Curimatã Herbivorous Pacu Piau Insectivorous Ituí Macrofagous Sairu Microfagous Cará* Omnivorous Candiru Mandi Total Student’s t test n number of specimens *p
