Environmental and toenail metals concentrations in copper mining and non mining communities in Zambia

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/authorsrights

Author's personal copy

International Journal of Hygiene and Environmental Health 217 (2014) 62– 69

Contents lists available at ScienceDirect

International Journal of Hygiene andEnvironmental Health

jo ur n al hom epa ge: www.elsev ier .com/ locate / i jheh

Environmental and toenail metals concentrations in copper mining and nonmining communities in Zambia

Wesu Ndililaa, Anna Carita Callana, Laura A. McGregorb, Robert M. Kalinb, Andrea L. Hinwooda,∗

a Centre for Ecosystem Management, Edith Cowan University, 270 Joondalup Drive, Joondalup, WA 6027, Australiab David Livingstone Centre for Sustainability, Department of Civil and Environmental Engineering, University of Strathclyde, Graham Hills Building, 50 Richmond Street, Glasgow, UK

a r t i c l e i n f o

Article history:Received 14 November 2012Received in revised form 19 March 2013Accepted 21 March 2013

Keywords:MetalsPersonal exposureEnvironmentMiningToenails

a b s t r a c t

Copper mining contributes to increased concentrations of metals in the environment, thereby increasingthe risk of metals exposure to populations living in and around mining areas. This study investigatedenvironmental and toenail metals concentrations of non-occupational human exposure to metals in 39copper-mining town residents and 47 non-mining town residents in Zambia. Elevated environmentalconcentrations were found in samples collected from the mining town residents. Toenail concentrationsof cobalt (GM 1.39 mg/kg), copper (GM 132 mg/kg), lead (21.41 mg/kg) selenium (GM 0.38 mg/kg) andzinc (GM 113 mg/kg) were significantly higher in the mining area and these metals have previously beenassociated with copper mining. Residence in the mining area, drinking water, dust and soil metals concen-trations were the most important contributors to toenail metals concentrations. Further work is requiredto establish the specific pathways of exposure and the health risks of elevated metals concentrations inthe copper mining area.

© 2013 Elsevier GmbH. All rights reserved.

Introduction

Mining and processing activities are known to emit a variety ofmetals, metalloids and heavy metals to the environment (MiningMinerals and Sustainable Development, 2001; Stüben et al., 2001;Von der Heyden and New, 2004). These often result in elevatedconcentrations of persistent chemicals such as arsenic, cadmium,cobalt, copper and lead in the surrounding environment (Bidoneet al., 2001; Coelho et al., 2007; Panday et al., 2007). The term metalsis used as a general term to cover metalloids and heavy metals.

In the Copperbelt region of Zambia, mining related dischargesinto air and water systems contribute significantly to increasedenvironmental metal concentrations (Environmental Council ofZambia (ECZ), 2001; Ntengwe, 2006). Environmental studies reportsignificant impacts of copper mining on the water quality of theKafue River, which supplies 40% of Zambia’s population with drink-ing water (Ntengwe, 2006; Von der Heyden and New, 2004).Concentrations of arsenic, cadmium, cobalt, copper and leadreported in water, sediments and fish have been found to be ele-vated in areas closest to mining (Mwase et al., 1998; ECZ, 2001).

∗ Corresponding author at: Centre for Ecosystem Management, School of Natu-ral Sciences, Edith Cowan University, 270 Joondalup Drive, Joondalup, WA 6027,Australia. Tel.: +61 8 6304 5372; fax: +61 8 6304 5509.

E-mail address: [email protected] (A.L. Hinwood).

Human exposure to metals occurs via direct contact with envi-ronmental contaminants in air, soil and water and via intake ofcontaminated food (Bergland et al., 2001; Bidone et al., 2001;Bhopal, 2002). Non-occupational studies have shown that peopleliving in close proximity to areas of current and past mining orindustrial activities are at risk of increased exposure to metals andhence adverse health effects (Murgueytio et al., 1998; Banza et al.,2009; Liu et al., 2010).

Despite previous studies providing evidence of elevatedenvironmental metal concentrations in the Copperbelt region(Norrgren et al., 2000; Ntengwe, 2006; Von der Heyden and New,2004), knowledge on actual human exposure, and the subsequentimpact of metals produced from mining on the community in theregion is limited (Banza et al., 2009).

This study set out to investigate non-occupational long termmetal exposure within a community in the Copperbelt miningregion of Zambia, as well as a non-mining community and to iden-tify any factors that may influence exposure using toenail metalconcentrations as a marker of exposure. Toenail metal concentra-tions have been used as suitable measures of long term metalsexposure (Slotnick and Nriagu, 2006). When external contamina-tion can be removed, toenail metal concentrations are generallyconsidered reflective of internal body stores (Lauwerys and Hoet,2001). They can also be useful indicators of environmental contam-ination (Garland et al., 1993; Slotnick and Nriagu, 2006). Toenailsamples and environmental sampling, in combination with ques-tionnaire information, were used to investigate the long-term

1438-4639/$ – see front matter © 2013 Elsevier GmbH. All rights reserved.http://dx.doi.org/10.1016/j.ijheh.2013.03.011

Author's personal copy

W. Ndilila et al. / International Journal of Hygiene and Environmental Health 217 (2014) 62– 69 63

exposure of communities to the metals arsenic, cadmium, cobalt,copper, lead, nickel, selenium and zinc.

Materials and methods

Study design

This was a cross-sectional study of non-occupational humanexposure to metals conducted in two locations in Zambia. Ethicsapproval was obtained for this study from the Edith Cowan Uni-versity Human Research Ethics Committee (approval #06-32) andthe University of Zambia Research Ethics Committee (Assurance NoFWA00000338-IRB00001131 of IORG0000774).

Study area

The mining area of Kitwe in the Copperbelt Province (located ataround 12◦49′ S and 28◦12′ E) was selected to represent the miningarea and the town of Livingstone in the Southern Province (locatedat around 17◦51′ S and 25◦52′ E) was selected to represent a nonmining area. Kitwe is one of the major copper mining towns locatedin the Copperbelt province thus making it an appropriate studylocation as elevated metals concentrations have been found inthe environment (Ntengwe, 2006). Several locations within Kitwewere selected for data collection to ensure a large enough popula-tion from which to recruit the required number of participants. StAnthony compound was selected as the primary site for the expo-sure study as it is a residential area located adjacent to and 1 kmnorth of the Nkana copper mine which is a source of metals (ECZ,2001; Ntengwe, 2006). A secondary site, Nkana West was selecteddue to its location directly adjacent to and south-west of the Nkanacopper mine. In the mining area 87% of participants were from StAnthony compound and 13% from Nkana West.

Livingstone town was selected as the control as no miningtakes place in the region. In addition its location being outside theKafue river drainage basin within which the Copperbelt province islocated excludes it from influences from the mining region (Von derHeyden and New, 2004). Livingstone is largely tourism based andthe community selected had similar socioeconomic characteristicsto the communities selected in Kitwe (Bhopal, 2002).

The study areas differed in geology. The Copperbelt region com-prises predominantly sulphidic ore deposits, while Livingstone ischaracterised by Bakota basalt which is high in quartz and rich iniron oxide and in some places lime (Livingstone District PlanningOffice October 2005). Copper deposits do extend from the extremenorth west of the country, through the Copperbelt to the south-ern province where Livingstone was situated (NERC, 2001). WhileLivingstone was situated some 1000 km away, no towns closer thatwere free of mining could be identified with similar socio economiccharacteristics.

Participant recruitment

It was aimed to recruit 40 non occupationally exposed adultsfrom the mining area and 40 from the control area. Commu-nity consultation was undertaken prior to commencement of thestudy. For cultural reasons, collection of biological samples wasapproached very carefully. Recruitment methods included doorknocking, community meetings and recruitment through localcommunity representatives. Verbal information was provided tothose unable to read the information sheets. Participants wererequired to have resided at their current address for more than1 year, and to be a non-smoker between the ages of 21–30 years.This age group was selected as it has been identified as a vulnerablegroup in terms of prevalence of HIV/AIDS (Bhopal, 2002). Partici-pants were also required not to work in an occupation, such as

mining or metal work, where exposure to a variety of metals mayoccur. Informed consent was provided either by signature or bythumb print in the presence of a witness, prior to data collection.

Data collection

Sampling was undertaken in June/July, the dry season, there-fore representing the worst-case scenario when metal levels inwater will be most concentrated. Dust particles were also expectedto be more mobile in the dry season, representing a period ofhigher exposures (Bidone et al., 2001; Georgopoulos et al., 2001;Pettersson and Ingri, 2001).

Interview administered questionnaire: This method was used dueto potential literacy issues in the communities. Information wascollected on water use, occurrence and frequency of soil contact,consumption of home grown produce, previous occupational expo-sure and passive exposure to cigarette smoke.

Toenail samples: Toenail samples were collected by participantsfrom all ten toes and stored at room temperature prior to samplepreparation. Participants were asked to wash their feet and toenailswith soap. Disposable stainless steel razors were provided in thesample packs for participants to use to ensure consistency acrosssample collection.

Drinking water: Samples were collected from the commonsources of drinking water in each of the study sites. Water sourcesincluded the river, council supply and a borehole. Sample bottleswere 250 mL polyethylene and were 1 M hydrochloric acid washedprior to sample collection. Samples from a hand pump or tap wererun continuously for 1 min prior to sample collection to ensure thatno stagnant water was collected (Bergland et al., 2001). Once col-lected, water samples were acidified with 10% nitric acid and storedat −20 ◦C prior to analysis.

Soil samples: Participants collected a soil sample from the landaround their place of residence; either from the vegetable gardenor an area of bare soil. Samples were collected from the top 10 cmlayer of soil into plastic bags and stored at room temperature priorto analysis.

Dust samples: Participants were asked to provide dust samplesby brushing dust off the interior walls of their home, or other dustyindoor surfaces, onto paper. The dust samples were then funnelledinto the sample collection bag provided.

Sample preparation

As a condition of importation of toenail samples for researchpurposes into Australia, toenail and soil samples were autoclavedat 15 lbs for 45 min in polypropylene containers to destroy micro-organisms. Toenail samples were washed thoroughly to removeexogenous contamination (Slotnick and Nriagu, 2006). The meth-ods used to clean the exterior surface of the toenails were adaptedfrom those of Mehra and Jujena (2005). Ultrasonication of thetoenails in Milli-Q water was undertaken for 30 min. Milli-Qwater was replaced with acetone and samples were ultrasoni-cated for a further 30 min to remove inorganic contaminants onthe nail surface. Thereafter, the toenails were rinsed five timeswith Milli-Q water and oven dried at 60 ◦C to constant weight.All washed toenail samples were acid digested using a wet diges-tion method (4 mL concentrated nitric acid) and heated to 65 ◦Cfor 2 h. Hydrogen peroxide (4 mL) was then added and the digestreturned to 65 ◦C until the reaction was complete (Kazi et al.,2000).

Soil samples were sieved using a 2 mm sieve to remove largeamounts of organic matter and pebbles, then oven dried at 40 ◦C for12 h. Samples were then homogenised in smooth bottomed Eppen-dorf tubes by addition of two ball bearings per tube and grindingfor 3 min at a frequency of 30 (1/s) (Rayment and Higginson,

Author's personal copy

64 W. Ndilila et al. / International Journal of Hygiene and Environmental Health 217 (2014) 62– 69

1992). Samples of equal dry weight (0.5 g) were digested using12 mL of a 2:1 mixture of concentrated nitric acid and hydrogenperoxide.

Dust samples were sieved using a 600 �m sieve. Followingsieving, dust samples were prepared in the same manner assoil.

Chemical analysis

Digested toenail, soil and dust samples were analysed using aVarian Pro inductively coupled plasma-atomic emission spectrom-eter (ICP-AES) for the metals cadmium, cobalt, copper, nickel, leadand zinc (Method 3050B Revision 2, USEPA, 1996). Arsenic and sele-nium were analysed by hydride generation. Water samples wereanalysed by direct injection by Lonestar Technical Services, Dubaiusing ICP-AES (for As, Cd, Co, Cu, Ni, Se and Zn) and the graphite fur-nace method (for Pb). Certified elemental standards from AustralianChemical Reagents (ACR) Queensland were run on the ICP-AES asinternal checks. Ten percent of samples were run as replicates andwere repeated at the end of each sampling cycle. Standards wererun after every 20 study samples.

Statistical methods and analysis

The biological and environmental concentrations of metals werehighly skewed to the right and highly censored. The toenail andenvironmental concentrations of metals below the detection lim-its were assigned a value of half the respective detection limit(Liu et al., 1997). Predictors of exposure concentrations were cate-gorised where relevant. Drinking water concentrations of sourceswere converted to an estimated intake for each individual basedon reported consumption of drinking water as there were noindividual water samples able to be taken due to the commonwater sources used by residents. Comparisons between groupswere made using nonparametric methods. Fisher’s Exact Proba-bility, Mann–Whitney U and Kruskal–Wallis tests were appliedto determine significant differences between and within studyareas. Pearson correlation coefficients were calculated on log trans-formed data to determine relationship between toenail metalconcentrations and predictor variables (indoor dust, residentialsoil, age, resident period, estimated metals intake via drinkingwater).

Multiple linear regression was applied to natural log trans-formed data to explore the influence of demographic, lifestylefactors and environmental metals concentration on toenail metalsconcentrations. Factors entered into the initial regression modelincluded indoor dust metal concentration, residential soil metalconcentration, age, consumption of home grown produce, dura-tion of soil contact and estimated intake from drinking wateras these were shown to influence toenail analysis from uni-variate tests. Linear regression was performed with the naturallog transformed toenail metal concentration as the dependentvariable, using forced entry of independent variables. A prob-ability of F of 0.05 was used for entry of variables into themodel, with a probability of 0.10 used for removal. After theinitial run of multiple linear regression, the model was re-runmultiple times, removing variables with insignificant beta coef-ficients and confidence intervals. Univariate and multiple linearregression analysis was performed using SPSS software (version17.0).

Trends within the dataset were examined through principalcomponents analysis (PCA) of the measured metal concentrationsusing Matlab software (Mathworks, version R2011a). PCA usesmatrix algebra to reduce the dimensionality of data and allow vari-ations within a dataset to be visualised more clearly.

Table 1Demographic and lifestyle characteristics of participants.

Study area

Mining area Non mining area

Number of participants 39 47Number of households 39 46Age (mean years) 25.6 24.7

(n = 37) (n = 47)Gender

Female 68% 57%Employment status

Females 5% 9%Males 20% 13%Participants unemployed 75% 78%

(n = 37) (n = 46)Months at current residence (n = 35) (n = 44)

Mean 109.6 123.2Standard deviation 87.7 89.3Range 12–290 12–360

Engage in activities involving soilcontact (%Yes)

80 97.9

Frequency of soil contact hoursper day (%)

(n = 29) (n = 47)

0–1 h 17 251–5 h 35 46>5 h 48.3 29

Consume home-grown produce(% Yes)

45 73

(n = 31) (n = 40)Self reported passive exposure to

cigarette smoke (%)50 70

(n = 34) (n = 47)

Results and discussion

Demographic and lifestyle characteristics of exposure groups

The demographic characteristics of study participants are pre-sented in Table 1. Most participants were unemployed and hadlived in their respective areas for approximately ten years. A higherpercentage of females participated in the study from the miningcommunity, although this difference was not significant (Table 1).The majority of participants were reported to have contact with soilon a regular basis, although the proportion reporting soil contactwas higher in the non-mining area (p = 0.018). More participants inthe non mining area consumed home grown produce (p = 0.030).No other significant differences between the study groups wereidentified (Table 1).

Environmental metals concentrations

Data from the questionnaire was used to identify all the sourcesof drinking water being accessed by the community. Council-supplied water was used in both areas, although other sourcesincluded river water, rain water and bore/well or spring waterwith all sources reported to be used by participants, hence thedata were averaged to assess the potential of drinking water useas a source of metals exposure. Table 2 shows average drinkingwater metals concentrations in both study areas from a numberof sources which include the groundwater and river samples (seeSupplementary data for more details). Arsenic, cobalt, copper andlead drinking water concentrations were higher in the mining area,while nickel and zinc concentrations were higher in the non miningarea (Table 2). Geometric mean concentrations of arsenic and leadin mining area sources exceeded the WHO drinking water guide-lines (WHO, 2006) (Supp. data). Interestingly, previous studies havefound copper and cobalt to be the major pollutants in drinkingwater in the Copperbelt region (Norrgren et al., 2000; Ntengwe,2006), however this was not the case in this study area.

Author's personal copy

W. Ndilila et al. / International Journal of Hygiene and Environmental Health 217 (2014) 62– 69 65

Table 2Summary of residential soil, dust and toenail metal concentrations (mg/kg) in the mining and non mining areas where GM = geometric mean. Average drinking waterconcentrations summarised and presented in mg/L.

Residential soil metalconcentrations

Residential indoor dustmetal concentrations

Average drinking waterconcentrationsa

Toenail metal concentrations

Mining(n = 36)

Non mining(n = 31)

MannWhitney U

Mining(n = 31)

Non mining(n = 44)

MannWhitney U

Mining Non mining Mining(n = 39)

Non mining(n = 47)

MannWhitney U

Arsenic U = 10,235 U = 124.5 U = 1300.5GM 0.32 1.28 z = 2.745 0.50 0.07 z = −6.434 0.01 <DL 0.02 0.03 z = 1.762Range <DL–74.4 <DL–95 p = .006 <DL–7.6 <DL–1.23 p < .001 <DL–0.02 <DL–1.0 <DL–1.0 p = .078

Cadmium U = 448.5 U = 351.5 U = 904.5GM 0.88 0.51 z = −3.092 1.19 0.48 z = −3.583 <DL 0.001 1.06 0.87 z = −.104Range <DL–4.4 <DL–1.8 p = .002 <DL–450 <DL–2.1 p < .001 <DL–0.002 0.37–35.5 <DL–4.0 p = .917

Cobalt U = 331 U = 286 U = 531.5GM 4.63 0.32 z = −5.018 1.80 <DL z = −5.688 0.05 0.002 1.39 0.76 z = −3.345Range <DL–18.1 <DL–65 p < .001 <DL–227 p < .001 0.009–0.2 <DL–0.006 0.4–11.5 <DL–3.13 p = .001

Copper U = 21.5 U = 48 U = .000GM 851 12.99 z = −7.362 325 16.0 z = −6.821 0.20 0.007 132 4.57 z = −7.951Range 12–10,979 3.5– p < .001 <DL–4239 5.3–138 p < .001 0.03–0.8 <DL–0.007 32.5–2225 1–30 p < .001

Lead U = 112.5 U = 167.5 U = 162GM 19.0 0.60 z = −6.514 16.7 0.44 z = −5.703 0.05 <DL 21.4 1.15 z = −6.547Range <DL–259 <DL–24.3 p < .001 1.2– <DL–21 p < .001 <DL–0.1 0.80–158 <DL–67.5 p < .001

Nickel U = 575 U = 656 U = 760GM 8.26 4.72 z = −1.814 4.62 4.08 z = −.280 0.04 0.06 1.99 1.21 z = −1.359Range <DL–80.3 <DL–541 p = .070 <DL–25.2 <DL–24.9 p = .780 0.03–0.05 0.05–0.06 0.37–33.8 <DL–29.0 p = .174

Selenium U = 946.5 U = 174GM <DL <DL 0.06 0.09 z = 3.323 N/A N/A 0.33 0.02 z = −7.037Range <DL–4.9 <DL–0.4 p = .001 <DL–11.4 <DL–0.18 p < .001

Zinc U = 168.5 U = 256.5 U = 416GM 61.3 14.8 z = −5.889 67.0 20.9 z = −4.578 0.08 0.11 112.7 78.0 z = −4.342Range 5.7–389 1.88–73 p < .001 <DL–1005 2.6–341 p < .001 0.07–0.1 0.09–0.2 62–599 29.1–425 p < .001

a These are average water metals concentrations from a number of sources in the study areas, see supplementary data for more details.

Cadmium, cobalt, copper, lead and zinc soil concentrations weresignificantly higher in the mining area (Table 2) (Cd U = 448.5,z = −3.092, p = .002; Co U = 331, z = −5.018, p < .001; Cu U = 21.5,z = −7.362, p < .001; Pb U = 112.5, z = −6.514, p < .001; Zn U = 168.5,z = −5.889, p < .001), however, copper was the only metal with con-centrations significantly higher than guideline values for soil (NEPC,1999). Nickel concentrations in soil were also higher in the miningarea although the increase was not significant (U = 575, z = −1.814,p = .070). Soil arsenic concentrations were higher in the control area(Table 2). For the purpose of this study, selenium was classified asa metal and included in the analysis. Selenium was expected to befound in the soils in the mining area due to its use in the copperrefining process (Greenwood and Earnshaw, 1995), however it wasnot detected. It is unclear why this occurred however the analyti-cal method may have lacked the sensitivity to detect selenium atconcentrations that may be present.

With the exception of copper, the soil metal concentrations inthe mining area were below those reported in other studies of cop-per mining and smelter areas (Benin et al., 1999; Pope et al., 2005;Carrizales et al., 2006; Nakayama et al., 2011; Staniland et al., 2010;Tembo et al., 2006). Lead soil concentrations were much lower thanthose found in a former lead/zinc mining area in Kabwe locatedin Central Province, Zambia where maximum lead concentrationswere three times higher (Tembo et al., 2006). However, as expected,the maximum copper concentrations in soil samples collected inthe mining area in this study were approximately double thosedetected in roadside soil samples from Kabwe (Nakayama et al.,2011).

Measuring indoor dust metal concentrations was considereduseful to assess human indoor exposure as it provided an indicationof whether this source may contribute to biological concentrationsvia inhalation or ingestion (Bergland et al., 2001; Lisiewicz et al.,2000). The indoor dust metal concentrations of arsenic, cadmium,

copper, cobalt, lead and zinc in Kitwe were significantly higher thanthose observed in Livingstone (Table 2) (As U = 124.5, z = −6.434,p < .001; Cd U = 351.5, z = −3.583, p < .001; Co U = 286, z = −5.688,p < .001; Cu U = 48, z = −6.821, p < .001; Pb U = 167.5, z = −5.703,p < .001; Zn U = 256.5, z = −4.578, p < .001). Positive significant Pear-son correlation coefficients were also observed for copper and leadconcentrations in soil and dust for all samples (copper r = .753,p < .001); (lead r = .600, p < .001) suggesting that the metals maycome from the same initial source. Nevertheless, the dust samplingmethod in this study was non standard due to logistics and thenature of residences in the area. Therefore, the measured dust con-centrations may underestimate the actual concentrations as themethod was limited to the collection of larger fractions, while finerdust fractions (8–32 �m) have been known to contain higher con-centrations of elements (Lisiewicz et al., 2000).

Apart from dust copper, cobalt and zinc concentrations of themining area, other dust metal concentrations were relatively lowin comparison with other industrialised regions (Benin et al., 1999;Bergland et al., 2001; Lisiewicz et al., 2000; Rasmussen et al., 2001).

Toenail metal concentrations were higher in the mining areafor all metals except arsenic (Table 2). Significant differences wereobserved in toenail cobalt, copper, lead, selenium and zinc concen-trations with increased concentrations in the mining area. Thesemetals have all previously been associated with mining in the Cop-perbelt (Tembo et al., 2006; Banza et al., 2009; Staniland et al.,2010). The toenail metal concentrations of greatest concern in themining area were; cadmium, lead, copper and cobalt which wereshown to be significantly higher or comparable to heavily pollutedareas worldwide (Nowak and Chmielnicka, 2000; Anwar, 2005;Were et al., 2008). Cobalt exists as large above-ground deposits inKitwe; therefore, it has the potential to be present in the atmo-sphere as a result of weathering as well as mining (Paton andBrooks, 1996). However, Banza et al. (2009) also reported elevated

Author's personal copy

66 W. Ndilila et al. / International Journal of Hygiene and Environmental Health 217 (2014) 62– 69

Table 3‘Best fit’ models for individual metals using multiple linear regression on factors influencing toenail metals concentrations.

Metal Standardised coefficient 95% Confidence interval for beta R2 Adj. R2

Beta Lower bound Upper bound

Cu Model 1 Constant 1.314 1.727 .850 .848Area .922 3.057 3.671

Model 2 Constant .436 5.186 .872 .863Soil Cu .405 .018 .657Ingestion by drinking water .543 .173 1.000

Pb Model 1 Constant −.161 .724 .547 .534Area .434 .805 2.762Soil Pb .357 .099 .499

Model 2 Constant .384 1.149 .462 .455Soil Pb .680 .426 .714

Cd Constant −.038 .272 .347 .337Dust Cd .589 .255 .499

Co Constant −.086 .249 .183 .171Dust Co .427 .098 .294

Se Constant −4.359 −3.669 .601 .596Area .775 2.383 3.407

Zn Model 1 Constant 3.831 4.417 .207 .185Area .335 .104 .584Dust Zn .200 −.012 .158

Model 2 Constant 3.498 4.220 .195 .170Dust Zn .389 .069 .257>5 h soil contact .218 −.007 .497

cobalt concentrations in urine of residents in the Copperbelt regionof the Democratic Republic of Congo.

Factors influencing exposure concentrations

A number of factors including demographic characteristics ofthe populations were investigated for their potential contributionor influence on toenail metal concentrations using multiple linearregression for the combined data set, the results for each metalare described in the following sections and summarised in Table 3.Whilst it would have been desirable to examine factors in eacharea separately, the resulting sample size for regression analysiswas too small, hence the area from which residents were recruitedwas included as a factor.

Arsenic: None of the environmental or demographic variablesexplained arsenic toenail concentrations in this study hence theregression outputs are not shown here. Thirty six percent of toenailsample concentrations, 36% of soils and 51% of dusts, were lowerthan the detection limit which is likely to have affected the analysis.

Cadmium: Dust cadmium concentrations had the most influenceon toenail metal cadmium concentrations, explaining approxi-mately 34% of the variability in toenail concentrations (Table 3).Area was not important in the regression modelling.

Cobalt: Dust cobalt was the most significant contributor to toe-nail cobalt concentrations, accounting for approximately 17% of thevariation.

Copper: When area of residence was included as an indepen-dent variable in the model this was the most important predictor,accounting for 85% of the variability in toenail copper concen-trations. When area of residence was not included in the model,estimated copper ingestion via drinking water and soil copper con-centrations were the variables that significantly influenced toenailcopper concentration, accounting for approximately 86% of thevariation in toenail copper concentrations (Table 3). However it isimportant to note that there was substantial correlation betweenwater, dust and soil and these concentrations were strongly influ-enced by area as well. If entered into the regression model alone,ingestion via drinking water and soil accounted for approximately85% and 74% of the variance in toenail copper concentrations,

respectively. Dust concentrations correlated with soil so stronglythat it did not improve the model. If entered into the model alonedust accounts for 49% of the variance.

Nickel: None of the environmental, demographic or lifestyle fac-tors were significant predictors of toenail nickel concentrations andhence the model outputs have not been shown.

Lead: Area of residence and soil lead concentrations were themost significant factors contributing to lead toenail concentrations,accounting for approximately 53% of the variability (Table 3). With-out the inclusion of area, soil lead accounted for approximately 45%of the variability alone. No other factors were identified as beingimportant despite lead concentrations in drinking water being ele-vated.

Selenium: The high number of non-detects for selenium in soilsand dusts precluded the inclusion of environmental concentrationsin the regression analysis. However area was a significant predictorof toenail concentrations, accounting for approximately 60% of thevariability.

Zinc: Area of residence and dust were the most significantfactors, individually accounting for approximately 18% of the vari-ability (Table 3). When area of residence was excluded from themodel, dust zinc concentration and reporting of greater than 5 h ofsoil contact per day accounted for approximately 17% of the vari-ability in participants’ toenail zinc concentrations, although soilcontact did not quite reach significance in the model.

Across the metals, the most significant contributors to increasedtoenail metal concentrations were: residence in the mining area,estimated copper intake via drinking water, soil lead and copperconcentrations and dust cadmium, cobalt and zinc concentrations.

Previous investigators have demonstrated that demographicand lifestyle factors are important pathways and contributors formetals exposure. Age, resident period (exposure period), previousoccupational exposure and passive cigarette smoke exposures haveall been shown to be important factors influencing metals exposure(Hogervorst et al., 2007), however, in this study, these factors werenot found to be major contributors to toenail metals concentrations.

Lead in drinking water represents a risk to health based onmeasured concentrations and this, combined with elevated soilconcentrations and the increased toenail lead concentrations,

Author's personal copy

W. Ndilila et al. / International Journal of Hygiene and Environmental Health 217 (2014) 62– 69 67

confirms environmental exposures represent a risk to residentsin this area and further investigation of lead exposure, withexamination of blood lead is required. Although toenail arsenic con-centrations were low in comparison with other studies (Hinwoodet al., 2003), the drinking water arsenic concentrations in Kitwewere also above health guidelines, indicating the potential forhealth risks. There is a need for additional monitoring forarsenic concentrations, including biological concentrations, to bet-ter assess population exposure and risk.

Cobalt and copper concentrations were also elevated in envi-ronmental and toenail samples. Cobalt is an essential elementand there is little research on the significance of exposures in thecommunity setting with most research on occupational exposures(Lauwerys and Lison, 1994). Copper is also an essential elementbut at very high concentrations, and for those with pre-existing ill-ness, copper may increase health risks (Georgopoulos et al., 2001).It is not clear from this study whether the concentrations are highenough to result in health impacts.

There are a number of limitations of this study, specifically thesmall sample size and the use of toenails as the only form of bio-logical sample. Diet is also a significant source of metals exposure(Ryan et al., 2001; Jarup, 2003) and has not been assessed in thisstudy.

Information on pesticide use and medication was not collectedand these contributors may have had an impact on individuals’metal exposure. However, the results confirm that those living nearcopper mines are at an increased risk of exposure to metals andtherefore the potential for adverse health effects is increased. Leadand arsenic in drinking water were found to be at concentrationsabove acceptable health standards.

Principal component analysis

To allow the overall distribution of metals between the twostudy areas to be compared, individual percentage metal contentswere calculated for each sample and the average value of each areawas plotted in Fig. 1. Fig. 1 clearly shows the difference in metaldistribution between the mining and non-mining areas; with themining area exhibiting copper as the major component in all sam-ple types, while the non-mining samples have zinc as the majoritycomponent. The results indicate that the residents of Kitwe (min-ing area) have been exposed to elevated levels of copper, as evidentby the increased proportion of the metal in toenail clippings. Giventhe multiple routes of exposure, by elevated copper levels in water,soil and dust, the threat to the environment and human health isamplified.

Fig. 1. Average percentage metal content in dust, soil and toenail samples fromresidents of Kitwe (mining) and Livingstone (non-mining).

Fig. 2. Principal component analysis (PCA) score plot showing the variation of metalconcentrations in toenail, soil and dust samples between the two study areas.

Principal component analysis (PCA) was performed on thepercentage metal concentration dataset (composed of dust, soil,toenail and drinking water samples) to simplify trends in variablesbetween the mining and non-mining regions. The first three prin-cipal components (PCs) described 72% of the variation within thedataset (Fig. 2). The PCA score plot provided in Fig. 2 shows clearseparation between the samples from the two different study areas,with only a single outlier from the non-mining region. Examina-tion of the dataset shows that the outlier represents a sample withfew detectable metal concentrations throughout the dust, soil andtoenail samples, and is grouped close to two mining samples withsimilar characteristics.

The loadings contributing most to PC1 are copper proportions insoil, drinking water and toenail samples (positive loadings), whilePC2 was most affected by dust Cu (positive) and PC3 had soil Cu(positive) and dust Cu (negative) as the major contributors. Thefull loading plots for PCs1–3 are provided in the supplementaryinformation (Figures S1 and S2). The PCA loadings clearly link thevarious forms of copper and allow the participants to be dividedinto two well-defined categories (from mining and non-mining)within the PCA score plot. The majority of the separation occursalong PC1. The mining area samples are shown to have a highlypositive correlation with PC1 indicating high proportions of copperin the toenail, drinking water and soil samples from these areas.

The mining samples also appear to have more intra-group vari-ation than the non-mining samples, with two separate clustersemerging within the participants from Kitwe. These clusters appearto be due to large variations in the proportion of copper in the dustsamples from this area. A great number of factors may have affectedthe spread of copper in dust, especially as the samples were col-lected during the dry season when particulates in dust would bemost mobile. In future studies, PCA could potentially be used tocluster samples from a wider study area to help define contamina-tion ‘hot spots’ and to track the outward spread of metals from thesource.

Some studies have indicated that immune compromised peo-ple may be at an increased risk from heavy metal contamination asthe metabolism of certain metals may occur differently as a resultof disease (Lauwerys and Hoet, 2001; Merzenich et al., 2001). Fur-thermore, elevated concentrations of metals are known to reduceimmunity and reduce the number of lymphocytes in the humanbody, the cells responsible for antibody production (Merzenichet al., 2001; Syacumpi et al., 2003; Palus et al., 2005). In Zam-bia, HIV/AIDS is common (Syacumpi et al., 2003; WHO, 2005),therefore the potential health risks associated with heavy metalexposure may be even greater than those of other areas where thedisease is less prevalent. The participant age group for this study

Author's personal copy

68 W. Ndilila et al. / International Journal of Hygiene and Environmental Health 217 (2014) 62– 69

are considered vulnerable in regards to the prevalence of HIV/AIDSinfection (Bhopal, 2002) and was selected to reduce the potentialfor this factor to confound the results as well as to get a represen-tation of a population that is potentially most at risk of the adverseimpacts associated with metals exposure, although information onthe HIV/AIDS status of participants was not requested. Research isneeded to specifically investigate the impacts of metal exposureon HIV/AIDS sufferers in areas where the risks of exposure are ele-vated. It needs to be recognised that children and pregnant womenare also vulnerable groups in terms of metal exposure and healtheffects. Children have increased exposure to environmental con-taminants resulting from their larger surface area to body massand the fact they breathe more air and consume more food perunit weight than adults (Moya et al., 2004; Selevan et al., 2000).In pregnant women, many metals are able to cross the placentalbarrier readily and may result in health impacts on the developingfoetus (Iyengar and Rapp, 2001; Ballatori, 2002; Rudge et al., 2009).Therefore efforts to establish exposure in these vulnerable groupsis imperative.

Conclusions

This cross sectional study has provided essential informationon both environmental and human exposure concentrations ofnon-occupationally exposed individuals in the Copperbelt region ofZambia. Long-term metal exposure has been quantified by the mea-surement of toenail metals concentrations and various contributorsto exposure investigated. This study has shown that residentialenvironmental metals concentrations in the copper mining areaof the Copperbelt were higher than the non mining area. Cad-mium, cobalt, copper, lead and zinc soil and dust concentrations;and arsenic, cobalt, copper, lead and nickel drinking water con-centrations were all elevated in the mining area. Toenail metalsconcentrations were also significantly elevated in the mining areawith environmental factors found to significantly influence metalsexposure, with specific concerns about cadmium, lead, copper andcobalt exposures.

This study has established that residents in Kitwe are exposedto elevated environmental metal concentrations and personalexposure concentrations were comparable, and in some caseshigher, than other heavily industrialised regions. Further researchis required to assess the degree of exposure in those at risk, assess-ment of specific activities and sources of exposure and the healthimpacts that may be relevant to this group. Particular attentionneeds to be given to pregnant women, children and those who areimmune compromised.

Acknowledgements

The authors would like to thank the Centre for Ecosystem Man-agement and School of Natural Sciences, Edith Cowan Universityfor funding this study. The authors would also like to thank theparticipants for providing their samples and time.

Appendix A. Supplementary data

Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.ijheh.2013.03.011.

References

Anwar, M., 2005. Arsenic, cadmium and lead levels in hair and toenail samples inPakistan. Environ. Sci. 12, 71–86.

Ballatori, N., 2002. Transport of toxic metals by molecular mimicry. Environ. HealthPerspect. 110 (S5), 689–694.

Banza, C.L.N., Nawrot, T.S., Haufroid, V., Decree, S., De Putter, T., Smolders, E., Kabyla,B.I., Luboya, O.N., Ilunga, A.N., Mutombo, A.M., Nemery, B., 2009. High humanexposure to cobalt and other metals in Katanga, a mining area of the DemocraticRepublic of Congo. Environ. Res. 109 (6), 745–752.

Benin, A.L., Sargent, J.D., Dalton, M., Roda, S., 1999. High concentrations of heavymetals in neighbourhoods near ore smelters in northern Mexico. Environ. HealthPerspect. 107, 279–284.

Bergland, M., Elinder, C.G., Jarup, L., 2001. Human Exposure Assessment: An Intro-duction. World Health Organisation, Stockholm.

Bhopal, R., 2002. Concepts of Epidemiology: An Integrated Introduction to the Ideas,Theories Principles and Methods of Epidemiology. Oxford University Press Inc.,New York.

Bidone, E.D., Laybauer, L., Castilhos, Z.C., Maddock, J.L., 2001. Environmental riskincrease due to heavy metal contamination caused by a copper mining activityin Southern Brazil. An. Acad. Bras. Cienc. 73, 271–286.

Carrizales, L., Razo, I., Tèllez-Hernández, J.I., Torres-Nerio, R., Torres, A., Batres, L.E.,Cubillas, A.C., Díaz-Barriga, F., 2006. Exposure to arsenic and lead of childrenliving near a copper-smelter in San Luis Potosi, Mexico: importance of soil con-tamination for exposure of children. Environ. Res. 101, 1–10.

Coelho, P., Silva, S., Roma-Torres, J., Henriques, A., Teixeira, J., Gomes, M., Mayan,O., 2007. Health impacts of living near an abandoned mine – case study: Jalesmines. Int. J. Hyg. Environ. Health 210, 399–402.

Environmental Council of Zambia (ECZ), 2001 State of Environment in Zambia 2000.Environmental Council of Zambia, Lusaka.

Garland, M., Morris, J.S., Rosner, B.A., Stampfer, M.J., Spate, V.L., Baskett, C.J., Wil-lett, W.C., Hunter, D.J., 1993. Toenail trace element levels as biomarkers:reproducibility over a 6 year period. Cancer Epidemiol. Biomarkers Prev. 2 (5),493–497.

Georgopoulos, P.G., Roy, A., Yonone-Lioy, M.J., Opiekun, R.E., Lioy, P.J., 2001. Envi-ronmental copper: its dynamics and human exposure issues. J. Toxicol. Environ.Health B, 341–394.

Greenwood, N.N., Earnshaw, E., 1995. Chemistry of the Elements. Butterworth-Heinemann, Oxford, England.

Hinwood, A.L., Sim, M.R., Jolley, D., de Klerk, N., Bastone, E.B., Gerostamoulos, J.,Drummer, O.H., 2003. Hair and toenail arsenic concentrations of residents liv-ing in areas with high environmental arsenic concentrations. Environ. HealthPerspect. 111, 187–192.

Hogervorst, J., Plusquin, M., Vangronsveld, J., Nawrot, T., Cuypers, A., Van Ecke, E.,Roels, H.A., Carleer, R., Staessen, J.A., 2007. House dust as possible route ofenvironmental exposure to cadmium and lead in the adult general population.Environ. Res. 103, 30–37.

Iyengar, G.V., Rapp, A., 2001. Human placenta as a ‘dual’ biomarker for moni-toring fetal and maternal environment with special reference to potentiallytoxic trace elements. Part 3: Toxic trace elements in placenta and pla-centa as a biomarker for these elements. Sci. Total Environ. 280 (1–3),221–238.

Jarup, L., 2003. Hazards of heavy metal contamination. Br. Med. Bull. 68,167–182.

Kazi, S., Ali, S.S., Furrukh, F., Kazi, T.G., Hasan, K., Kazoo, T.G., 2000. Comparison ofmetal ions in biological samples of schizophrenic patients and control subjects.Am. Clin. Lab. 19 (3), 8.

Lauwerys, R.R., Hoet, P., 2001. Industrial Chemical Exposure-Guidelines for Biolog-ical Monitoring, 3rd ed. Lewis Publishers, Boca Raton, Fl.

Lauwerys, R., Lison, D., 1994. Health risks associated with cobalt exposure – anoverview. Sci. Total Environ. 1–3, 1–6.

Lisiewicz, M., Heimburger, R., Golimowski, J., 2000. Granulometry and the content oftoxic and potentially toxic elements in vacuum-cleaner collected, indoor dustsof the city of Warsaw. Sci. Total Environ. 18, 69–78.

Liu, S.P., Lu, J.C., Kolpin, D.W., Meeker, W.Q., 1997. Analysis of environmental datawith censored observations. Environ. Sci. Technol. 31, 3358–3362.

Liu, C.P., Luo, C.L., Gao, Y., Li, F.B., Lin, L.W., Wu, C.A., Li, X.D., 2010. Arsenic contam-ination and potential health risk implications at an abandoned tungsten mine,southern China. Environ. Pollut. 158, 820–826.

Livingstone District Planning Office, 2005. Livingstone District Development Plan.in P. Department, editor. Government of Zambia.

Mehra, R., Jujena, M., 2005. Fingernails as biological indices of metal exposure. J.Biosci. 30, 253–257.

Merzenich, H., Hartwig, A., Ahrens, A., Beyersmann, D., Schlepegrell, R., Scholze, M.,Timm, J., Jockel, K.H., 2001. Biomonitoring on carcinogenic metals and oxidativeDNA damage in a cross-sectional study. Cancer Epidemiol. Biomarkers Prev. 10,515–522.

Mining Minerals and Sustainable Development, 2001. Research Topic 4. Impactsof Mining and Mineral Processing on the Biophysical Environment in SouthernAfrica. University of Witwatersrand, Johannesburg.

Moya, J., Beare, C.F., Etzel, R.A., 2004. Children’s behavior and physiologyand how it affects exposure to environmental contaminants. Pediatrics 113,996–1006.

Murgueytio, A.M., Evans, G.R., Roberts, D., 1998. Relationship between soil and dustlead in alead mining area and blood lead levels. J. Expo. Anal. Environ. Epidemiol.8 (2), 173–186.

Mwase, M., Viktor, T., Norrgren, L., 1998. Effects to tropical fish of sediments fromthe Kafue River, Zambia. Bull. Environ. Contam. Toxicol. 61, 96–101.

Nakayama, S.M., Ikenaka, Y., Hamada, K., Muzandu, K., Choongo, K., Teraka, H.,Mizuno, N., Ishizuka, M., 2011. Metal and metalloid contamination in road-side soil and wild rats around a Pb–Zn mine in Kabwe, Zambia. Environ. Pollut.159 (1), 175–181.

Author's personal copy

W. Ndilila et al. / International Journal of Hygiene and Environmental Health 217 (2014) 62– 69 69

National Environment Protection Council (NEPC), 1999. National EnvironmentalProtection (Assessment of Site Contamination) Measure, Schedule B (1) – Guide-line on the Investigation Levels for Soil and Groundwater, NEPC.

National Environment Research Council (NERC), 2001. Groundwater Quality: Zam-bia, http://www.nerc.ac.uk (accessed 03.05.08).

Norrgren, L., Pettersson, U., Örn, S., Bergqvist, P.A., 2000. Environmental monitoringof the Kafue River, located in the Copperbelt, Zambia. Arch. Environ. Contam.Toxicol. 38, 334–341.

Nowak, B., Chmielnicka, J., 2000. Relationship of lead and cadmium to essentialelements in hair, teeth and nails of environmentally exposed people. Ecotoxicol.Environ. Saf. 46, 265–274.

Ntengwe, F.W., 2006. Pollutant loads and water quality in streams of heavily pop-ulated and industrialized towns. Phys. Chem. Earth Pt A/B/C 31, 832–839.

Palus, J., Lewinska, D., Dziubaltowska, E., Stepnik, M., Beck, J., Rydzynski, K., Nils-son, R., 2005. DNA damage in leukocytes of workers occupationally exposed toarsenic in copper smelters. Environ. Mol. Mutagen. 46, 81–87.

Panday, K.P., Sharma, R., Roy, M., Panday, M., 2007. Toxic mine drainage fromAsia’s biggest copper mine at Malanjkhand, India. Environ. Geochem. Health29, 237–248.

Paton, A., Brooks, R.R., 1996. A re-evaluation of haumaniastrum species as geob-otanical indicators of copper and cobalt. J. Geochem. Explor. 56, 37–45.

Pettersson, U.T., Ingri, J., 2001. The geochemistry of Co and Cu in the Kafue River asit drains the Copperbelt mining area, Zambia. Chem. Geol. 177, 399–414.

Pope, J.M., Farago, M.E., Thornton, I., Cordos, E., 2005. Metal enrichment in Zlatna, aRomanian copper smelting town. Water Air Soil Pollut. 162, 1–18.

Rasmussen, P.E., Subramanian, K.S., Jessiman, B.J., 2001. A multi-element profile ofhouse dust in relation to exterior dust and soils in the city if Ottawa, Canada. Sci.Total Environ. 267, 125–140.

Rayment, E.G., Higginson, R.F., 1992. Australian Laboratory Handbook of Soil andWater Chemical Methods. Reed International Books Australia P/L, trading asInkata Press, Port Melbourne.

Rudge, C.V., Röllin, H.B., Nogueira, C.M., Thomassen, Y., Rudge, M.C., Odland, J.Ø.,2009. The placenta as a barrier for toxic and essential elements in paired mater-nal and cord blood samples of South African delivering women. J. Environ. Monit.11, 1322–1330.

Ryan, P.B., Scanlon, K.A., MacIntosh, D.L., 2001. Analysis of dietary intake of selectedmetals in the NHEXAS – Maryland investigation. Environ. Health Perspect. 109(2), 121–128.

Selevan, S., Kimmel, C.A., Mendola, P., 2000. Identifying critical windows ofexposure for children’s health. Environ. Health Perspect. 108 (Suppl. 3),451–455.

Slotnick, M.J., Nriagu, J.O., 2006. Validity of human nails as a biomarker of arsenicand selenium exposure: a review. Environ. Res. 102, 125–139.

Staniland, S., Coppock, M., Tuffin, M., van Zyl, L., Roychoudhury, A.N., Cowan, D., 2010.Cobalt uptake and resistance to trace metals in Comamonas testosteroni isolatedfrom a heavy-metal contaminated site in the Zambian Copperbelt. Geomicrobiol.J. 27 (8), 656–668.

Stüben, D., Berner, Z., Kappes, B., Puchelt, H., 2001. Environmental monitoring ofheavy metals and arsenic from Ag–Pb–Zn mining: a case study over two millen-nia. Environ. Monit. Assess. 70, 181–200.

Syacumpi, M.M., Liywali, K., Mbale, M., Syacumpi, M., 2003. Country analysis offamily planning and HIV/AIDS: Zambia. POLICY, Washington, DC.

Tembo, B.D., Sichilongo, K., Cernack, J., 2006. Distribution of copper, lead, cad-mium and zinc concentrations around Kabwe town in Zambia. Chemosphere63, 497–501.

USEPA, 1996. Method 3050B (SW-846): Acid Digestion of Sediments, Sludges, andSoils, Revision 2. http://www.epa.gov/epaoswer/hazwaste/test/pdfs/3050b.pdf(accessed 20.11.05).

Von der Heyden, C.J., New, M.G., 2004. Groundwater pollution on the Zam-bian Copperbelt: deciphering the source and the risk. Sci. Total Environ. 327,17–30.

Were, F.H., Njue, W., Murungi, J., Wanjau, R., 2008. Use of human nails as bio-indicators of heavy metals environmental exposure among children in Kenya.Sci. Total Environ. 393, 376–384.

World Health Organization (WHO), 2005. HIV/AIDS Epidemiological SurveillanceReport for the WHO African Region 2005 Update HIV/AIDS. World Health Orga-nization, Harare.

World Health Organization (WHO), 2006. Guidelines for Drinking-Water QualityIncorporating First Addendum to Vol. 1, Recommendations, 3rd ed. World HealthOrganization.