Exposure to Toxic Metals (As, Cd, Hg, Pb) in Two Artisanal and Small-scale Gold Mining (ASGM) Areas in Zimbabwe

Background: People living and working in ASGM areas are exposed to toxic metals such as arsenic (As), cadmium (Cd), mercury (Hg) and lead (Pb). Whereas Hg is purposely used to extract the gold from the ore, the other toxic metals can be liberated from the ore during the mining process. These metals can contaminate drinking water and food and are thus, a source of exposure for people living in these mining areas. However, there is limited to no data about the exposure to toxic metals of people living in ASGM areas. Therefore, the purpose of this study was to conduct biomonitoring of people identifying themselves as artisanal gold miners. Methods: Blood and urine were collected from 207 people living and working in two ASGM areas in Zimbabwe. Hg in whole blood and urine was analyzed by direct mercury analysis, Pb in blood as well as As and Cd in urine were analyzed by GF-AAS. The results were evaluated by descriptive analysis and correlated with the location, age, sh consumption, years living and working in the area. Results: Median As level in urine was 10.0 µg/l (range from below limit of detection to 460 µg/l). Median Cd level in urine was 0.3 µg/l (range from below limit of detection to 11.4 µg/l). Median Hg levels in blood / urine were 2.7 / 5.2 µg/l (range from 0.2 to 167 µg/l / 0.1 to 612 µg/l). Median Pb level in blood was 19.9 µg/l (range from 6.6 to 276 µg/l). As expected, the Hg levels were frequently above international reference values (e. g. NHANES, blood: 34%, urine: 69%). For Pb and Cd, a considerable number of participants (32% and 22% respectively) were above the reference values (NHANES / National Health and Nutrition Examination Survey) and for As 33% of the participants were above the reference value (UBA/ German Environment Agency). Conclusions: Hg levels were, as expected in ASGM areas, mainly above reference and threshold values. A high proportion of As, Cd and Pb levels were also above reference levels. Therefore, the exposure to toxic metals in the two ASGM areas in is relevant for public health.


Toxic Metals in Artisanal and Small-Scale Gold Mining (ASGM)
ASGM is a mining-based livelihood strategy in many developing countries rich in gold resources, counting about 15 million miners globally [1]. While there are also a few low-end formal gold miners, it is predominantly an informal, poverty-driven, poorly resourced, comparatively ine cient and transient sector. ASGM in Zimbabwe has been growing very rapidly over the last forty years. This growth was even faster over the last two decades, from an estimated 300,000 miners in 2000 to over 1.5 million today [2,3]. This has been due to a variety of factors including a persistently shrinking economy across multiple industries along with recurrent droughts during the last two to three decades. Consequently, there has been an inevitable growth in terms of tonnages mined and processed as well as gold produced. At the same time, this is also re ected in the amounts of process chemicals applied, new rock surfaces exposed to weathering and release of various elemental species. This led to the worsening of adverse environmental and public health impacts due to the release of toxic metals, mainly mercury (Hg), but also others such as arsenic (As), cadmium (Cd) and lead (Pb), which are geochemically associated with the gold hosting sulphide ores [4,5].
More than 90% of gold deposits in Zimbabwe are associated with the largely ma c/basic and cratonic greenstone belts, while the rest is located within the Limpopo Mobile Belt to the south and the Lomagundi meta-sediments in the north. The genesis of most gold deposits in Zimbabwe (> 75%) involved the generation of superheated (> 2000 °C) aqueous hydrothermal sulphide complexes at depth, in the upper mantle to lower crust. Gold is mainly transported in solution as aqueous complexes of hydrogen sulphides/bi-sulphides and chlorides as well as metal complexes [6]. Various elements, and especially heavy metal elements, have an a nity for dissolved hydrogen sulphides and chlorides in hydrothermal ore-forming solutions. This includes, but is not limited to, As, silver (Ag), Cd, Hg, Pb, selenium (Se), antimony (Sb), tellurium (Te), thallium (Tl) and zinc (Zn). These metals and many others are commonly associated with gold ores and are routinely used as path nders in gold exploration geochemistry [7]. Most of these metals have densities above ve g/cm³ and are called heavy metals. Those with increased toxicity are currently referred to as toxic metals. Toxic metals naturally occur at relatively low concentrations. However, mining, crushing and milling processes, which are used to extract the gold, can result in the increased liberation of toxic metals. The gold is collected at the end of the concentration process as the high-value target mineral, while the other toxic metals end up in the tailings dumps at mining locations. Several oxidation and reduction reactions during mineral processing and within the tailings dumps result in the formation of various compound species as well as the precipitation of native metals. The World Health Organization has classi ed As, Cd, Hg, and Pb among the top ten chemicals of public health concern [8].
Therefore, this study focused on the biomonitoring of these toxic metals in people that identi ed themselves as gold miners in two ASGM areas in Zimbabwe to evaluate their actual exposure. Mercury (Hg) uptake strongly depends on its species. Vaporized elemental Hg is easily absorbed by inhalation. The majority of ingested organic Hg, e.g. methylmercury, is absorbed in the gastrointestinal tract, whereas the uptake of inorganic Hg is relatively low. Mercury has a high a nity for sulfhydryl groups and can therefore react with proteins and non-protein thiols, e.g. glutathione. This causes reduced activity of enzymes and oxidative stress. As Hg dominantly affects the nervous system, chronic exposure to Hg leads to ataxia, weakness, tremors, convulsions, sensory and motor impairment as well as emotional and behavioral changes. Furthermore, adverse effects on the kidney, the cardiovascular system and the immune system have been reported. [15][16][17][18] Lead (Pb) has major pathways, namely atmospheric dust, contaminated food and water, and, in the past, automobile emissions. Pb is readily absorbed in the lungs, but less e cient in the gastrointestinal tract. It accumulates in the liver, kidneys, and bone tissue. Pb causes the inhibition of multiple enzymes such as δ-aminolaevulinic acid dehydratase, a key enzyme in hemoglobin synthesis. Furthermore, it causes oxidative stress and epigenetic alterations. Chronic exposure to Pb affects the hemoglobin synthesis, the cardiovascular and central nervous system, the kidneys, the fertility and is considered to cause cancer.
During pregnancy, it can also affect fetal growth. Increased exposure of fetuses, children and adolescents to Pb is associated with reduced cognitive capacities and intelligence. [19][20][21] 1.3. Toxic Metals in other Mining Areas As exposure in ASGM has been documented in Tanzania where only 10.6% respondents knew about As toxicity [22]. These were mostly miners who had primary education and above. However, this study focused on the knowledge and perceptions regarding Hg and As toxicity and did not measure the levels of these metals in biospecimens. In a Ghanaian study, where urinary As levels were analyzed in 57 males living in a small-scale gold mining community, 34% of the cases had As levels above 100 µg/l [23], which is above US health guidelines [24]. Further studies from Ghana con rm that As levels are increased in mining areas [25][26][27].
Cd was analyzed in Ghana in the urine of residents of a small-scale gold mining community [28]. The geometric mean of the 57 male residents for Cd was 0.38 µg/l. Hg exposure in ASGM was described in a pooled analysis, including 17 studies focused on Hg body burden in ASGM communities worldwide [29]. The pooled analysis revealed a central median Hg concentration in urine of 5.9 µg/l and an upper median concentration of 188 µg/l. For blood, the central median Hg concentration was 10.9 µg/l. In a study from Ghana, Hg was measured in the urine of amalgam smelting miners [30]. The results showed a median Hg concentration of 6.56 µg/l. Exposure to mercury of people living in ASGM areas in Zimbabwe has been studied intermittently [31]. Most studies explore exposure to Hg, showing increased levels of Hg within the mining communities, especially for miners. In Zimbabwe in 2004 the median mercury urine level was 36.6 µg/l, respectively 25.8 µg/g crea) and in blood 10.2 µg/l for miners [32].
Pb levels in ASGM areas are a speci c problem in Nigeria. In 2010, ASGM communities in North-western Nigeria experienced a severe lead poisonings due to the processing of heavily contaminated ore [33][34][35][36][37]. One study showed that all children below ve years of age were poisoned (blood Pb levels ≥ 10 µg/dl) and 97% of the 204 children had blood lead levels ≥ 45 µg/dl [37].

Study Aim
The data about the exposure to toxic metals of people living in ASGM areas is clearly still limited. The purpose of this study was to analyze As, Cd, Hg and Pb in two mining areas in Zimbabwe.

Study Design
This cross-sectional study was conducted within two weeks (18th of March 2019 to the 29th March 2019) at two hospitals in the Kadoma and Shurugwi Districts (Zimbabwe), respectively. Kadoma and Shurugwi are old gold mining towns. In Zimbabwe, the largest gold deposits and the highest density of miners can be found in Kadoma [5]. Shurugwi is primarily known for chrome mining, with a growing trend for ASGM activities. O cial data on the number of miners in Kadoma and Shurugwi do not exist.
Inclusion criteria for participation were a minimum age of 18 years and that the participants worked for at least one month as artisanal and small-scale miners. There was no preselection regarding the participants' sex as well as state of physical health. Participants were recruited using snowball sampling where participants recruit further participants among their colleagues. This sampling technique was used due to a widespread and hard-to-reach target population. A per diem of US $ 5 was provided for participants to compensate for the loss of income on the study day. The study centers were located at the hospitals in Kadoma and Shurugwi to minimize contamination of the human biomonitoring samples. Each participant signed an informed consent form and material transfer agreement, prior to the data and sample collection. All documents were available in the three main languages English, Shona and Ndebele spoken in Zimbabwe. Altogether, 206 participants consented to participate (130 from Kadoma and 76 from Shurugwi). Participation in the study was voluntary. Participants were asked to ll out questionnaires concerning general information on demographics and exposure risks factors and undergo medical examination concerning symptoms of chronic Hg intoxication. To ensure a con dential analysis, the samples and data were pseudonomized. More information about the study was previously published [31].

Ethical Clearance
Approval for conducting the study was gained from the Ministry of Health in Zimbabwe and the respective local and regional authorities. The Ethics

Urine and Blood Collection
All sample containers were labelled with the participant's code for future allocation. Spot urine samples were collected using disposable urine collection cups. For transport and analysis, aliquots of the urine samples were transferred into a urine monovette (Sarstedt®). To prevent bacterial growth and degradation, the samples were acidi ed with nitric acid to a pH of approximately 2, which was tested with a pH strip. Trained health professionals took venous blood samples into 7 ml Lithium-Heparin-coated tubes for trace metal analyses (Sarstedt®) from all participants. All samples were continuously stored at 4 °C. Once located in the laboratory, samples were stored at − 18 °C.

Analysis of Trace Metals in Urine and Blood
All samples were at least analyzed in duplicate after thawing on a roll mixer. For quality control ClinChek® -Control (Recipe, Munich, Germany) reference materials for whole blood and urine were used. Results below the limit of detection were divided by 2 and further analysis was performed with those modi ed values. The samples of one participant had to be excluded from analysis due to contamination.

Analysis of Total As in Urine
Total As in urine was analyzed by graphite furnace atomic absorption spectroscopy (GF AAS) at a detection wavelength of 193.7 nm. Urine samples were diluted six fold with 0.01% Triton-X in 0.13% nitric acid. 20 µl of this dilution were automatically pipetted into the graphite tube of the GF-AAS (AAnalyst 600, Perkin Elmer, Rodgau, Germany). 5 µg Pd (as Pd(NO 3 ) 2 ) and 3 µg Mg(NO 3 ) 2 were added as matrix modi ers. The furnace program was according to the recommendations of the manufacturer. Quanti cation based on the standard addition method. In detail, 10 and 20 pg As were directly added to the sample in the graphite tube, respectively. The limit of detection (LOD) was at 0.5 µg/l.

Analysis of Cd in Urine
Cd in urine was analyzed by graphite furnace atomic absorption spectroscopy (GF AAS) at a detection wavelength of 228.8 nm. Urine samples were diluted fourfold with 0.01% Triton-X in 0.13% nitric acid. 20 µl of this dilution were automatically pipetted into the graphite tube of the GF-AAS (AAnalyst 600, Perkin Elmer, Rodgau, Germany). 50 µg NH 4 H 2 PO 4 and 3 µg Mg(NO 3 ) 2 were added as matrix modi ers. The furnace program was according to the recommendations of the manufacturer. Quanti cation based on the standard addition method. In detail, 0.5 and 1 pg Cd were directly added to the sample in the graphite tube, respectively. The LOD was at 0.5 µg/l.

Analysis of Hg in Urine and Blood
Urine and blood specimen were analyzed without any pre-processing by direct Hg analysis using a DMA80-evo instrument (MLS-Mikrowellen GmbH, Leutkirch, Germany). 100 µl of the sample were directly pipetted into quartz tubes for analysis. All of the following steps were part of the automated Hg analysis. One sample at a time was transferred into the oven where it was incinerated under an oxygen atmosphere at 750 °C. After the catalytic removal of matrix components, the Hg released from the sample was concentrated in an amalgamator. Finally, Hg was released by heating the amalgamator and detected by atomic absorption at 253.5 nm. The quanti cation was based on an external calibration. The LOD for total Hg in urine and blood was at 0.05 µg/l.

Analysis of Pb in Blood
Pb in blood was analyzed by graphite furnace atomic absorption spectroscopy (GF AAS) at a detection wavelength of 193.7 nm. Blood samples were diluted tenfold with 0.05% Triton-X. 20 µl of this dilution were automatically pipetted into the graphite tube of the GF-AAS (AAnalyst 600, Perkin Elmer, Rodgau, Germany). 50 µg NH 4 H 2 PO 4 and 3 µg Mg(NO 3 ) 2 were added as matrix modi ers. The furnace program was according to the recommendations of the manufacturer. Quanti cation based on the standard addition method. In detail, 5 and 10 pg Pb were directly added to the sample in the graphite tube, respectively. The LOD was at 1.0 µg/l.

Analysis of Creatinine in Urine
Creatinine in urine samples was determined for creatinine-corrected levels of toxic metals in urine. Creatinine-corrected urine values were considered in order to account for the in uence of the effect of urine dilution on the exposure indicator. Urine samples were sent to the central laboratory of University hospital of LMU and analyzed with Cobas C702 using the Jaffé-method. Creatinine adjusted urine values (As-urine crea-corr, Cd-urine crea-corr, Hg-urine crea-corr) with creatinine levels < 0.3 g/l and > 3.0 g/l were excluded from analysis.

Statistical Methods
All the data analysis were done in SPSS (version 26, IBM) and R Studio (version 3.6.2, RStudio Inc.). We conducted descriptive analysis and reported the summary statistics. Boxplots were used to visualize the distribution of the data. We assessed correlation of toxic metals and other covariates using the Pearson correlation coe cient and the Spearman correlation for non-normally distributed covariates. We also reported tests on difference in the medians of continuous variables using Wilcoxon Rank Sum test.

Descriptive Analysis of the Data
The urinary levels of As, Cd and Hg as well the Pb and Hg blood levels were measured for the entire study population of 206 participants (results in µg/l). Additionally, the urinary As, Cd and Hg levels were creatinine-corrected for 193 participants (results in µg/g creatinine). Urine samples with creatinine levels below 0.3 or above 3.0 g/l were excluded from analysis. For As and Cd in urine, 12 and 120 samples were below the limit of detection (LOD), respectively.
The levels for toxic metals were compared between the two mining sites, Kadoma and Shurugwi, (see Table 1 and Fig. 1). For As in urine (corrected for creatinine), the levels in Kadoma were signi cantly higher compared to Shurugwi. For Hg in urine, the levels in Shurugwi were signi cantly higher compared to Kadoma. Otherwise, there are no signi cant differences in the median levels between the two mining areas. The results strati ed by gender can be found in the supplemental information. Mann-Whitney-U test to analyze possible signi cant difference in median between the two study locations. *: signi cant (p < 0.05); Min.: Minimum; Max.: Maximum; crea. corr.: Corrected for urinary creatinine. Creatinine correction was not applied for urine samples with creatinine levels below 0.3 or above 3.0 g/l.; LOD: limit of detection.

Comparison to reference and threshold levels
For reference values, U.S. data from the "National Health and Nutrition Examination Survey" (NHANES, 95th percentiles for the age group "20 + years" from the 2015-2016 survey) published by the CDC were used [38]. Furthermore, data from the "German Environmental Surveys" (GerES, 95th percentiles for the age group 18 to 69 years) published by the "German Environment Agency" (UBA) were used [39,40]. The UBA also released internationally well accepted threshold levels for adults, called HBM-I and HBM-II (HBM = human biological monitoring). Levels below the HBM-I value are considered as safe, between HBM-I and HBM-II as alerting and levels above HBM-II as health-threatening [39,41].
The reference and threshold values for As, Cd, Hg and Pb as well as the percentage of participants that exceeded these values are given in Table 2. For urinary As, between 2 and 33% of the participants were above the reference levels, if the reference values from the NHANS study or the UBA were used. The reference values of Cd in urine were exceeded by 22% (NHANES) or 27% (UBA), respectively. Additionally, 4% of the participants were above the HBM-II-value for Cd in urine. For Hg, 34% (NHANES) or 73% (UBA) of the Hg levels in urine and blood were above reference values. The HBM-II value for Hg in blood was exceeded by 11% of the participants, the HBM-II value for Hg in urine by 21%. Finally, blood Pb levels of 32% of the participants exceeded the reference values given by the NHANES study. The UBA provides individual blood Pb reference values for women and men. Here, 24% of the women and 20% of the men exceeded the reference value. Additionally, 13% of the participants had blood Pb levels above the threshold value proposed by NIOSH.

Correlations
To analyse possible correlations between the different metals as well as age, years living in the area and years working, a Spearman's correlation test was performed (Table 3). A strong correlation was found for Hg in urine with Hg in blood. Otherwise, only Hg (in urine) and Cd (in urine) showed a weak correlation. The age of the participants or the years they lived in the area did not correlate with the toxic metal levels. In contrast, the years worked as a miner showed a weak, but signi cantly positive correlation with Hg levels in urine. Fish consumption is a potential confounder for As and Hg levels. Therefore, the results for the toxic metal levels in the samples were strati ed by sh consumption (less than once a weak vs at least once a weak, for more details see supplemental information). Although As and Hg levels were higher in the high sh group, no signi cant effect of sh consumption on As or Hg levels was observed. Table 3 Correlation between the different toxic metals (µg/l) and other parameters (Spearman's correlation).

Hg in urine
Hg in blood Pb in blood As in urine Cd in urine

Discussion of the Results
In general, the levels of As, Cd, Hg and Pb in the samples showed a log-normal distribution, which was expected for this population. As explained in the introduction, toxic metals can occur together with gold-containing ores in ASGM areas. Therefore, it seems plausible that elevated As, Cd and Pb levels in the participants together with the exposure to elemental Hg may be attributed to the mining activities in Kadoma and Shurugwi. However, these levels are in uenced by various factors. Below, we discuss the results for the individual metals.
In contrast to the other metals, the reference values for As in urine from UBA and CDC differ considerably from one another (15.0 vs 49.9 µg/l). This is likely due to different exposures to As, e.g. by sh consumption and other sources such as drinking water. However, reference values represent the actual exposure in the general population and cannot be used for toxicological evaluation. In this study, the As levels are more comparable to the reference values of the NHANES study than to the levels in Germany (UBA). However, we did not analyze the As species to differentiate the intake of inorganic species (As(III) and As(V)), and organic As (e.g. arsenobetaine). In fact, inorganic As species are considered carcinogenic and its exposure should be limited to a minimum.
Inorganic As can be ingested via drinking water, that could be contaminated due to mining activities or via sh consumption, which had a positive, but not signi cant, effect on urinary As levels. Furthermore, the ingestion of airborne dust may have been an additional source of exposure. The differences found for urinary As (corrected for creatinine) between the study locations may be explained by different exposure due the mining or dietary patterns.
For Cd, most of the samples were below the LOD. This speaks for a generally low exposure. Still, a considerable number of participants were above the references and threshold values. Cd levels are generally elevated in smokers. Unfortunately, we do not have the data for smoking in our study. In Zimbabwe's mining areas, far more men compared to women smoke [5]. However, this was not re ected in our results, where women were more frequently above international reference and threshold values for Cd (data not shown).
The high number of samples above reference and threshold values found for Hg in urine and blood can be explained by the occupational use of Hg to extract gold. Interestingly, the results were not signi cantly confounded by the consumption of sh, although levels of Hg in blood and urine in the high sh group were elevated compared to low sh group. An increase in Hg levels was especially expected for blood, since the uptake of MeHg, which is the predominant Hg species in sh, would result in elevated Hg levels in blood rather than in urine. Eventually, sh consumption did not contribute signi cantly to the total Hg body burden. This may be explained by that the fact that the Hg levels in the consumed sh were relatively low (see also limitations) compared to the exposure to elemental Hg. Unfortunately, data on the Hg species was not available to further investigate the source of exposure. The differences found for urinary Hg between the study locations may be attributed to the more frequent application of protective equipment and safety precautions, since the Kadoma region has more experience with ASGM. Additionally, the years of mining in ASGM areas showed a weak positive correlation with Hg levels in blood and urine. This nding may be explained by the fact that chronic exposure to Hg leads to the accumulation of Hg in the kidney, resulting in a generally elevated urinary elevation [42]. However, we expect that the acute exposure to elemental Hg outweighs the effect of years mining in this population. As exposure to Hg can be assessed in either blood or urine, both matrices are highly correlated to each other. We also found a weak correlation between Hg and Cd, which has no plausible explanation. We considered this to be the result of a "statistical" chance by applying multiple statistical tests as we did.
For Pb, about a third of the participants were above international reference values. The exposure to lead could be via contaminated drinking water, although the uptake of Pb in the gastrointestinal tract of adults is considered relatively low, and airborne dust due to mining activities. Besides this, Pb exposure could be due to emissions of leaded gasoline, which was used longer in Africa than in the US or Europe [43]. Therefore, the recent exposure to these emissions might still be re ected by elevated blood Pb levels. Furthermore, we cannot exclude that leaded gasoline might be still used in Zimbabwe, especially in remote areas.

4.2.
Comparison of the results with other studies in mining areas.
The results of this study were compared to the results from other studies in mining areas around the globe (see Table 4). Not all of these studies were analyzing the same range of toxic metals, but single elements up to all four parameters. We used, if given, the median as well as the 25th and 75th percentile for comparison. In general, our results were in line to what has been previously found in studies conducted in current and former mining areas.
For As, the studies from Tanzania [44] and Ghana [23] show similar results for ASGM areas. But also in non-active mining areas in Mexico, comparable urinary As levels were found [45]. In contrast, studies from Spain and Guatemala found relatively low levels of As the urine [46,47]. For Cd in urine, the studies from Ghana, Spain and Guatemala showed very similar results [23,46,47]. The levels of Hg in urine found in our study are comparable to what has been found in other ASGM areas with known Hg use [48][49][50][51][52]. In areas without known or recent use of Hg for gold extraction, urinary Hg levels were considerably lower [45][46][47]. For Hg in blood, the results from other ASGM areas are relatively similar [48,49,53]. But also in non-ASGM mining areas, the median of Hg in blood was relatively high [47]. In contrast, pregnant women living in ASGM areas showed lower Hg levels in blood [44]. For Pb, our results are mainly similar to what has been found in other studies [47,[53][54][55]. However, in a study from Mexico, blood Pb levels were signi cantly higher [45].
Although some of the studies found similar results, the exposure to toxic metals likely depends on multiple factors such as the concentrations of metals in soil and water, the diet, the use of Hg for gold extraction, living area and many others. Generally, living close to mining areas seems to have a signi cant effect on the body burden [46,47,53]. One possible explanation is that contaminated mining tailings contribute to the exposure to toxic metals by causing elevated concentrations in water, food and airborne dust [45]. Values measured in studies in current and former mining areas (country) Reference 25th -median − 75th 25th -median − 75th  supplementalinformation.docx