9

Click here to load reader

Arsenic in hair and nails of individuals exposed to arsenic-rich groundwaters in Kandal province, Cambodia

Embed Size (px)

Citation preview

Page 1: Arsenic in hair and nails of individuals exposed to arsenic-rich groundwaters in Kandal province, Cambodia

S C I E N C E O F T H E T O T A L E N V I R O N M E N T 3 9 3 ( 2 0 0 8 ) 1 6 8 – 1 7 6

ava i l ab l e a t www.sc i enced i r ec t . com

www.e l sev i e r. com/ loca te / sc i to tenv

Arsenic in hair and nails of individuals exposed to arsenic-richgroundwaters in Kandal province, Cambodia

Andrew G. Gaulta,⁎, Helen. A.L. Rowlanda, John M. Charnocka,b, Roy A. Wogeliusa,Inma Gomez-Morillac, Sovathana Vongd, Moniphea Lengd, Sopheap Samrethd,Mickey L. Sampsond, David A. Polyaa

aWilliamson Research Centre for Molecular Environmental Science and School of Earth, Atmospheric and Environmental Sciences,University of Manchester, Oxford Road, Manchester, M13 9PL, United KingdombSTFC Daresbury Laboratory, Daresbury, Warrington, WA4 4AD, United KingdomcAdvanced Technology Institute, Ion Beam Centre, University of Surrey, Guildford GU2 7XH, UKdResource Development International – Cambodia, P.O. Box 494, Phnom Penh, Cambodia

A R T I C L E I N F O

⁎ Corresponding author. Present address: DeK1N 6N5.

E-mail address: [email protected] (A.G. Ga

0048-9697/$ – see front matter © 2007 Elsevidoi:10.1016/j.scitotenv.2007.12.028

A B S T R A C T

Article history:Received 22 September 2007Received in revised form10 December 2007Accepted 18 December 2007Available online 30 January 2008

The health implications of the consumption of high arsenic groundwater in Bangladesh andWest Bengal are well-documented, however, little is known about the level of arsenicexposure elsewhere in Southeast Asia, where widespread exploitation of groundwaterresources is less well established. We measured the arsenic concentrations of nail and hairsamples collected from residents of Kandal province, Cambodia, an area recently identifiedto host arsenic-rich groundwaters, in order to evaluate the extent of arsenic exposure.Nail and hair arsenic concentrations ranged from 0.20 to 6.50 μg g−1 (n=70) and 0.10 to 7.95 μgg−1 (n=40), respectively, inmany cases exceeding typical baseline levels. The arsenic contentof the groundwater used for drinking water purposes (0.21–943 μg L−1 (n=31)) was positivelycorrelated with both nail (r=0.74, pb0.0001) and hair (r=0.86, pb0.0001) arsenicconcentrations. In addition, the nail and hair samples collected from inhabitants usinggroundwater that exceeded the Cambodian drinking water legal limit of 50 μg L−1 arseniccontained significantly more arsenic than those of individuals using groundwatercontaining b50 μg L−1 arsenic. X-ray absorption near edge structure (XANES) spectroscopysuggested that sulfur-coordinated arsenic was the dominant species in the bulk of thesamples analysed, with additional varying degrees of As(III)-O character. Tentative linearleast squares fitting of the XANES data pointed towards differences in the pattern of arsenicspeciation between the nail and hair samples analysed, however, mismatches in sampleand standard absorption peak intensity prevented us from unambiguously determining thearsenic species distribution. The good correlation with the groundwater arsenicconcentration, allied with the relative ease of sampling such tissues, indicate that thearsenic content of hair and nail samples may be used as an effective biomarker of arsenicintake in this relatively recently exposed population.

© 2007 Elsevier B.V. All rights reserved.

Keywords:ArsenicHairNailBiomarkerCambodiaGroundwater

partment of Earth Sciences, University of Ottawa, 140 Louis Pasteur, Ottawa, ON, Canada

ult).

er B.V. All rights reserved.

Page 2: Arsenic in hair and nails of individuals exposed to arsenic-rich groundwaters in Kandal province, Cambodia

169S C I E N C E O F T H E T O T A L E N V I R O N M E N T 3 9 3 ( 2 0 0 8 ) 1 6 8 – 1 7 6

1. Introduction

The consumption of arsenic-rich groundwaters in the BengalDelta is causingmassive human health problems (Smith et al.,2000, Nordstrom, 2002). It has been estimated that up to 6 and35 million people inWest Bengal and Bangladesh, respectively,are exposed to drinking water containing arsenic concentra-tions greater than the local legal limit of 50 μg L−1 (BGS andDPHE, 2001), leading to the situation being described as theworst mass poisoning of a population in history (Smith et al.,2000). The presence of high arsenic subsurface waters in otherparts of Southeast Asia including Vietnam (Berg et al., 2001,2007; Agusa et al., 2006), and more recently Cambodia (Polyaet al., 2003, 2005; Kubota et al., 2006; Berg et al., 2007;Buschmann et al., 2007) suggests that the human health crisisbeing experienced in West Bengal and Bangladesh maydevelop elsewhere.

Thewidespread development of groundwater resources fordrinking water purposes in Cambodia has only occurred in thepast 10–15 years. A recent investigation into signs of chronicarsenic poisoning (arsenicosis) in Cambodia reported nounequivocal evidence of the disease among 4000 individuals,despite the consumption of groundwaters containing elevatedlevels of arsenic by many of the survey's volunteers (Milton,2003; Milton et al., 2005), however, such work relied on visualinspection of participants. The lack of such evidence mightindicate that (i) the groundwaters only constitute a minor partof the studied population's drinking water intake; or (ii) theperiod of exposure to such waters (the bulk of tube wells thattap arsenic-rich groundwaters were typically b5 years old atthe time of the study) was too short for long latency period (5–15 years) arsenic-related cancers and other visible symptomsto develop (Smith et al., 2000). More recent surveys have nowidentified some 300 individuals in Kandal province thatexhibit symptoms of arsenic poisoning (Sampson et al., 2007).

Given the sizeable lag time between the start of chronicarsenic intake and the appearance of observable symptoms ofarsenicosis, a reliable indicator of the early stages of deleter-ious arsenic exposure is needed. Urine samples have com-monly been used to examine arsenic metabolism (Brima et al.,2006) and arsenic concentrations in groundwater have beenfound to be positively correlated with urinary arsenic (Cal-deron et al., 1999), however, the requirement to freeze theurine during storage may present problems during extendedperiods of sample collection in rural parts of the developingworld. Recently, blood arsenic levels in a cohort of Bangladeshivillagers were reported to exhibit a good degree of positivecorrelation with groundwater arsenic concentrations (Hallet al., 2006), however, storage issues and the invasive nature ofthe sample collection procedure are significant drawbacks tothe use of blood as a biomarker for arsenic exposure. Arsenichas an affinity for sulfhydryl groups in keratin, hence, itreadily accumulates in the hair and nails. Once formed, thenail and hair matrix becomes isolated from other metabolicprocesses in the body (Hopps, 1977), making the collection ofsuch tissues an attractive option with which to monitor pastarsenic exposure. Hair grows more rapidly than nail andrecords exposure that occurred a few months before collec-tion. Fingernails, which themselves grow quicker than toe-

nails, take an average of 6 months to grow out completely(Fleckman, 1997), and so record arsenic exposure over aslightly earlier time period than hair samples. Numerousstudies have shown that toenail (e.g. Karagas et al., 2000;Mandal et al., 2003; Schmitt et al., 2005) and hair (e.g. Kurttioet al., 1998; Agusa et al., 2006) arsenic concentrations are wellcorrelated with drinkingwater arsenic concentrations and canbe used as biomarkers for arsenic exposure in humans. In thisstudy we have collected nail and hair samples from indivi-duals living in Kandal province, Cambodia that use ground-water containing a range of arsenic concentrations in order to(i) evaluate the relative exposure of individuals to arsenic; and(ii) assess the use of nail and hair samples as biomarkers forarsenic exposure in this population. There exists only limiteddata on the type and distribution of arsenic species in humanhair and nail and such work has generally employed extrac-tion techniques that may change the original arsenic specia-tion. We have used proton induced X-ray emission (PIXE) andX-ray absorption near edge structure (XANES) techniques toprobe the micron-scale variation in arsenic concentration andthe speciation of arsenic, respectively, in a number of nail andhair samples. Using XANES we can probe the hair/nail arsenicspeciation directly without any sample preparation andprovide information to address this knowledge gap. PIXEalso avoids an extraction step and may provide useful spatialinformation at the μg g−1 level about arsenic zonation. Thistechnique has been successfully used to map and quantify Aszonation in mineral (Jamtveit et al., 1993) and hair (Peach andLane, 1998; Sera et al., 2002) samples and has been extensivelyused to document the concentrations of other elements in arange of biomaterials (Watt and Landsberg, 1993; Olabanjiet al., 2005).

2. Materials and methods

2.1. Sample collection and preparation

Fingernail (n=70) and hair (n=40) samples were collected inMarch 2005 from individuals living in villages in the Kien Svaydistrict of Kandal province, a region of Cambodia documentedto host high arsenic groundwaters (Polya et al., 2003, 2005; Berget al., 2007; Buschmann et al., 2007). Hair samples werecollected from the nape of the head as near as possible to thescalp using stainless steel scissors. Finger nail samples werecollected using stainless steel nail clippers. The hair and nailsamples were sealed separately in labelled polyethyleneziplock bags and were not opened until return to Manchesterfor cleaning. Informed consentwas obtained from each surveyparticipant following guidelines approved by the University ofManchester Committee on the Ethics of Research on HumanBeings. A sample of groundwater used for drinking waterpurposes was also collected with selected nail and hairsamples. The standing water in the tube well was purgedafter pumping for 2–3 min and an unfiltered aliquot wascollected in an acid cleaned high density polyethylene bottleand acidified to 0.02 M HNO3. Rudimentary demographicand dietary information was collected alongside each watersample. The age range of the participants spanned 8 –85 years(median 44 years), and was skewed towards women (nails:

Page 3: Arsenic in hair and nails of individuals exposed to arsenic-rich groundwaters in Kandal province, Cambodia

Table 1 – Summary of arsenic concentration data for nail,hair and well water samplesa

NailAs/μg g−1

HairAs/μg g−1

Well waterAs/μg L−1

na 70 40 31Mean±standard error 1.90±0.20 1.41±0.32 153±40Median 1.30 0.54 57.4Minimum 0.20 0.10 0.21Maximum 6.50 7.95 943

aThat the number of well water samples is fewer than the numberof nail and hair samples is due in part to the use of one tube well bymore than one person and the collection of nail samples withoutpaired tube well water samples during an initial sampling survey.

170 S C I E N C E O F T H E T O T A L E N V I R O N M E N T 3 9 3 ( 2 0 0 8 ) 1 6 8 – 1 7 6

28 women, 9 men; hair: 29 women, 11 men). Limited partici-pant information was collected during an initial nail samplingsurvey, hence the discrepancy with the number of nailsamples analysed.

In Manchester, the nail samples were scrubbed using anylon brush, then cleaned following the procedure outline byChen et al. (1999). Briefly, the nail sample(s) were immersed in25 ml of 1% Triton X-100 and placed in an ultrasonic bath for20 min. After sonication, the solution was discarded and thenails were rinsed thoroughly with deionised water (18 MΩ),then dried overnight at 60 °C. The hair specimens were cleanedusing the method of Ryabukhin (1978). The hair was treatedsequentially with acetone (25 ml, 10 min sonication), thenthree timeswith deionisedwater (25ml, 10min sonication eachtime) and finally acetone (25 ml, 10 min sonication), discardingthe wash solution between each step. After washing, the hairwas dried overnight at 60 °C. These procedures were performedin a Class 1000 clean room before any analytical workcommenced in order to remove any potentially arsenic-bearingsoil and other contaminants from the surfaces of the nail andhair samples.

2.2. ICP-MS analysis

Following cleaning, 10–100 mg of nail/hair sample wereaccurately weighed into acid cleaned polypropylene tubes,1 ml of concentrated HNO3 (sub-distilled from Analar grade69%HNO3, BDH) was added to each sample and the tubeswerecapped and left at room temperature. After 48 h, the digestatewas diluted with 9ml of deionisedwater, then filtered (0.4 μm)into a fresh acid cleaned polypropylene tube. A referencehuman hair material (GBW07601), certified for its total arseniccontent, was treated in the same manner as the samples inorder to check the analytical accuracy of the digestionprocedure. To the best of our knowledge, no human nailcertified reference material is available. The filtered sampleswere analysed by ICP-MS (PQII, Thermo) with online additionof a 50 ppbGe internal standard achievedbymixing the samplesolution and internal standard solution using a T-piece and375 μL mixing coil prior to introduction to the ICP-MS. Thegroundwater samples collected during the hair/nail samplingsurvey were also analysed for total arsenic by ICP-MS.

2.3. PIXE spectrometry

Proton Induced X-ray Emission (PIXE) analysis was conductedat the University of Surrey Ion Beam Centre using a beam of3 MeV protons focused to a diameter of 3 μm2 and with anaverage beam current of 100 pA. The X-rays were detectedusing a lithium-drifted silicon detector (Gresham ScientificLtd., 80 mm2 active area) at a distance of 30 mm from thesample.Due to the lowbackground in PIXE spectra, elementsatconcentrations of ppm (parts per million) can be detected.Simultaneous Rutherford backscattering spectrometry (RBS)analysis was performed in order to verify the PIXE analysis.The backscattered protons were detected using a PIPS detectorwith an active area of 50 mm2 (Ortec) mounted at a distance of44 mm from the sample. RBS and PIXE spectra were recordedsimultaneously from different spots (3 μm2) within the nail,and the elemental composition of each spot was analysed. In

particular, the concentration of arsenic, iron and sulfur wasobtained.

2.4. XANES spectroscopy

X-ray absorption near edge structure (XANES) spectroscopymeasurements were made on a number of nail and hairsamples to gain further information on the speciation ofarsenic in thenail and/or hair structure. Arsenic K-edgeXANESspectra were obtained from selected nail and hair samples atthe ultra-dilute station 16.5 at the Synchrotron RadiationSource (SRS) at Daresbury laboratory. The SRS was operatingat 2 GeV with a beam current of between 130 and 240 mA.Station 16.5 is equipped with a Si(220) double crystal mono-chromator, with harmonic contamination of the beam mini-mised by a vertically focusingmirror in addition to detuning to70% of the maximum beam intensity. The monochromatorwas calibrated using the L(III) edge of a gold foil. Whole nailsamples or tightly bunched hair samples were fixed betweenSellotape windows and mounted in an aluminium sampleholder. Data were collected at liquid nitrogen temperaturewith the station operating in fluorescencemodeusing anOrtec30 element solid state Ge detector. Between four and elevenscans were collected for each sample and summed to improvesignal to noise. The spectra of end-member standards ofsodium arsenite (NaAsO2; GPR, BDH), disodium arsenateheptahydrate (Na2HAsO4.7H2O; Analar, BDH) and arsenic–glutathione were collected at room temperature in transmis-sion mode. The arsenic–glutathione solution was synthesizedby mixing a 10-fold molar excess of glutathione with sodiumarsenite. The proportion of different arsenic phases in eachsample was established by fitting the summed sample XANESspectra to a linear combination of end-member standardspectra using the Solver package included in Microsoft Excel,with the relative contribution of each standard determined byminimising a least squares residual.

2.5. Statistical analysis

StatPlus (2005, AnalystSoft, Vancouver, Canada) was used toperform statistical analyses of the data. Since the arsenicconcentrations of the groundwater, nail and hair sampleswere not normally distributed, non-parametric statistical testswere employed. Spearman's rank correlation coefficient (r)

Page 4: Arsenic in hair and nails of individuals exposed to arsenic-rich groundwaters in Kandal province, Cambodia

Fig. 1 –Variation in (a) nail and (b) hair arsenic concentrationas a function of well water arsenic concentration.

Table 3 – Ranges of arsenic, iron and sulfur concentrationsdetermined from spot PIXE analyses on one fingernail

As/μg g−1 (LoDa) Fe/μg g−1 (LoDa) S/wt.% (LoDa)

3 3 265 3 3.4 0.00144 3 244 1 4.1 0.00086 2 166 2 5.1 0.00168 2 624 3 5.1 0.0015bldb 3 43 3 5.7 0.0017bldb 8 74 8 2.0 0.0006bldb 3 528 2 6.2 0.00235 8 314 1 5.3 0.00139 3 471 4 5.7 0.002027 c 12 66400 16 6.2 0.00331 1 67 1 2.1 0.0008bldb 6 36 8 1.3 0.0005bldb 3 660 2 3.3 0.00151 1 15 1 1.7 0.0009bldb 2 21 9 1.5 0.00071 1 12 1 1.9 0.0009

a Limit of detection (LoD) for each point analysis, defined as threetimes the square root of the background over 1 full width halfmaximum centred about the principal peak's centroid, converted toμg g−1 from counts using charge and instrumental constants.b bld denotes below limit of detection.c PIXE and RBS data from this spot analysis are displayed in Fig. 2.

171S C I E N C E O F T H E T O T A L E N V I R O N M E N T 3 9 3 ( 2 0 0 8 ) 1 6 8 – 1 7 6

was used to assess the degree of association of nail/hairarsenic content with different variables such as groundwaterarsenic concentrations and age. The Mann–Whitney U-testwas used to test for differences between the nail or hairarsenic content of groups of participants sorted according totheir associated groundwater arsenic concentration. Statisti-cal significance was indicated by values of pb0.05.

3. Results

3.1. Arsenic concentrations of nail, hair and groundwatersamples

Analysis of the acid digested nail samples revealed that thearsenic concentration ranged from 0.20 μg g−1 to 6.50 μg g−1,with amediannail arsenic concentration of 1.30 μg g−1 (Table 1).The hair samples showed a similar arsenic concentration rangeandmedian of 0.1 μg g−1–7.95 μg g−1 and 0.54 μg g−1, respectively(Table 1). The arsenic concentration determined in the humanhair reference material (GBW07601) was 0.27±0.08 μg g−1 (n=3;0.26, 0.27 and 0.29 μg g−1), which agrees well with the certifiedvalue of 0.28±0.05 μg g−1.

The dissolved arsenic content of the well waters sampledwas between 0.21 μg L−1 and 943 μg L−1, with a median of

Table 2 – Summary of nail and hair arsenic content according t

Nail

Well water arsenic conte

“Low” (≤50 μg L−1) “High” (

n 18 19Mean±standard error (μg g−1) 0.72±0.11 1.9Median (μg g−1) 0.57 1.5Minimum (μg g−1) 0.28 0.5Maximum (μg g−1) 1.64 4.9

aNail and hair arsenic data are divided according to their associated well wlimit of 50 μg L−1 arsenic as the cut-off point.

57.4 μg L−1 (Table 1). Plotting the nail arsenic concentrationsagainst the companion groundwater arsenic levels (Fig. 1a)reveals a significant positive correlation (r=0.74, pb0.0001).The arsenic levels in the hair samples are also positivelyassociated with the groundwater arsenic concentrations(r=0.86, pb0.0001, Fig. 1b).

The legal limit for arsenic in drinking water in Cambodia iscurrently set at 50 μg/L. This value was used to split the nailand hair arsenic data into two groups according to theircorresponding groundwater arsenic concentrations, termed“low” (≤50 μg L−1) and “high” (N50 μg L−1) (Table 2). Themediannail and hair arsenic concentrations of individuals consumingwaters containing N50 μg L−1 dissolved arsenic were signifi-cantly higher than those of people drinking water with ≤50 μgL−1 aqueous arsenic (pb0.001 for both nail and hair samples).

3.2. PIXE analysis

The nail arsenic, iron and sulfur concentrations obtained fromsuch spot PIXE analyses are listed in Table 3. The average

o well water arsenic concentration

Hair

nta Well water arsenic contenta

N50 μg L−1) “Low” (≤50 μg L−1) “High” (N50 μg L−1)

19 216±0.33 0.28±0.04 2.43±0.524 0.22 1.383 0.10 0.265 0.57 7.95

ater arsenic concentration using the Cambodian drinking water legal

Page 5: Arsenic in hair and nails of individuals exposed to arsenic-rich groundwaters in Kandal province, Cambodia

Fig. 2 –a) PIXE spectrum from high arsenic and iron spot on nail sample listed in bold in Table 3. Dotted and solid lines indicateexperimental data and best fit of data, respectively. b) RBS spectrum from same spot. Data shown as black dots, fit to thedata shown as solid line, with individual contributions from each element shown below fit line.

172 S C I E N C E O F T H E T O T A L E N V I R O N M E N T 3 9 3 ( 2 0 0 8 ) 1 6 8 – 1 7 6

arsenic concentration in the nail analysed was approximately5 μg g−1, with a concentration range of b1 to 27 μg g−1 (Table 3).ICP-MS analysis of the acid digestate of this nail sampleshowed the bulk arsenic content to be 1.13 μg g−1, somewhatlower than the PIXE average, likely reflecting the heteroge-neous distribution of arsenic as indicated by the span of PIXEspot arsenic concentrations. The maximum arsenic level ishighlighted in bold, with the representative PIXE spectrum forthis point displayed in Fig. 2a, showing the range of other traceelements included within the nail. The high levels of arsenicand iron for this point might indicate surface contamination,but simulation of the associated Rutherford backscattering(RBS) data (Fig. 2b), which can provide information on varia-

Fig. 3 –K-edge arsenic XANES spectra of selected nail and hairsamples alongside the model compounds sodium arsenate(As(V)-O), sodium arsenite (As(III)-O) and arsenic coordinatedto S in glutathione (As–glut).

tions in element concentrations as a function of depth,suggests that the iron is evenly distributed throughout thedepth profile analysed and the stoichiometry of the analysis isstill consistent with that of normal hard keratin. Although thearsenic concentration is too low to make a significantcontribution to the fit of the RBS data, the iron data andregular nail stoichiometry both indicate that the observedhigh iron region is not caused by contamination of the surfacebut rather is pervasive below the interface. We therefore inferthat the associated arsenic is also not coincidental surfacecontamination.

3.3. XANES analysis

The XANES spectra obtained for the four nail and three hairsamples analysed are displayed in Fig. 3, alongside the modelcompounds arsenate (As(V)-O), arsenite (As(III)-O) and arsenicbound to sulfur in glutathione (As–glut). Using these end-member standards, linear combination fitting indicated thatonly combinations of As(III)-O and As–glut yielded the bestmatch for the nail samples, however, the proportion of thesephases varied considerably between samples (Table 4). Fittingof the three hair samples showed slightly less inter-samplevariation. The arsenic speciation in these samples wasdominated by sulfur coordination, alongside a lesser As(III)-O component. In all three samples, the inclusion of a smallproportion of As(V) improved the fit of the data (Table 4).

Table 4 – Best fit of XANES spectra collected for selectednail and hair samples

Samplecode

As–glut/%

As(III)-O/%

As(V)-O/%

Tissue Asconcentration/μg g−1

CFN 38b (nail) 19 81 0 3.60CFN 38c (nail) 38 62 0 3.66CFN 41 (nail) 74 26 0 4.95CFN 44a (nail) 49 51 0 4.48CFN 8 (hair) 47 41 12 6.92CFN 38c (hair) 60 36 4 5.93CFN 41 (hair) 56 29 15 6.01

Page 6: Arsenic in hair and nails of individuals exposed to arsenic-rich groundwaters in Kandal province, Cambodia

173S C I E N C E O F T H E T O T A L E N V I R O N M E N T 3 9 3 ( 2 0 0 8 ) 1 6 8 – 1 7 6

4. Discussion

Acid digestion and subsequent ICP-MS analysis of the nailsamples revealed that the arsenic concentration ranged from0.20 μg g−1 to 6.50 μg g−1 (n=70, median=1.30 μg g−1, mean1.90 μg g−1; Table 1). The upper end of this range exceeded thatfound in fingernails of non-occupationally exposed popula-tions in Sweden (0.07–1.09 μg g−1 (n=96, median=0.22 μg g−1,mean 0.27 μg g−1), Rodushkin and Axelsson, 2000) and the USA(b0.01–0.81 μg g−1 (n=208, mean=0.09 μg g−1), Karagas et al.,2000). The span of arsenic concentrations in nail samplestaken from individuals living in arsenic-affected areas ofWestBengal is considerably wider, with Samanta et al. (2004) andMandal et al. (2003) reporting ranges of 0.74–36.6 μg g−1 (n=33,median=4.73 μg g−1, mean 7.24 μg g−1) and 2.14–40.3 μg g−1

(n=47, mean=7.32 μg g−1), respectively. The arsenic content ofnail samples from persons exposed to arsenic in their drinkingwater in neighbouring Bangladesh exhibited a similar range of0.60–53.4 μg g−1 (n=230, mean=12.1 μg g−1) (Dhar et al., 1997).Although the arsenic concentrations in the nails collected inthis study do not reach the high end of these scales, the levelsof arsenic do appear to be elevated relative to typicalbackground concentrations.

The PIXE point concentration data are in broad agreementwith the ICP-MS analysis of digested nail material. The PIXEwork was completed not only to measure arsenic concentra-tions, but also in order to image compositional zoning ifpresent. Arsenic concentrations were not high enough toproduce useful images, although patterns in Fe and S could bedetermined (data not shown). However, the point analyses doindicate that arsenic concentrations are not constant throughthe various growth zones on this human fingernail. Indeed, atleast one point analysis gives arsenic concentrations that aresignificantly elevated. Furthermore the associated RBS spec-trum indicates that the high metal content at this point isnot surface contamination. Therefore if we consider that thefingernail growth zones serve as a crude measure of elapsedtime, then the PIXE result may be interpreted as an indicationthat arsenic intake levels over the period of growth of thisfingernail were episodic and at certain times became relativelyextreme. This implies a temporal complexity that needs to beaccounted for in any testing regimen developed for Cambodia.

Analysis of the hair samples showed a similar arseniccontent to that of the nail samples, ranging from 0.10 to 7.95 μgg−1 (n=40, median=0.54 μg g−1, mean 1.41 μg g−1; Table 1). Arecent study reported arsenic concentrations in hair samplesfrom residents of Northern Sweden without occupationalexposure to arsenic of 0.03 to 0.32 μg g−1 (n=114, median=0.07 μg g−1, mean 0.09 μg g−1) (Rodushkin and Axelsson, 2000),and typical hair arsenic concentrations are b1 μg g−1 (WHO,1981), suggesting that many of the hair arsenic levels in ourstudy are higher than baseline levels. High hair arsenicconcentrations have been reported in individuals consumingarsenic-rich groundwaters inWest Bengal. Mandal et al. (2003)and Samanta et al. (2004) reported hair arsenic concentrationranges of 0.70–16.2 μg g−1 (n=47, mean=4.50 μg g−1) and 0.17–14.4 μg g−1 (n=44, median=2.29 μg g−1, mean=3.43 μg g−1),respectively. The extent of the arsenic concentration appearsgreater than in our study, perhaps due to the longer exposure

duration in the Bengal delta than in Cambodia, since themajority of tubewells used by participants in our researchwere installed within five years of sampling. The concentra-tion ranges of arsenic in hair samples from residents ofVietnam and Cambodia appear closer to those observed in ourwork. Agusa et al. (2006) documented a span of 0.09–2.77 μg g−1

(n=39, median=0.42 μg g−1, mean=0.62 μg g−1) in hair samplesfrom residents of Hanoi, Vietnam, while Kubota et al. (2006)determined arsenic concentrations of 0.05–~16 μg g−1 (n=60,median=0.79 μg g−1, mean=1.77 μg g−1) in hair from inhabi-tants of Kratie province, Cambodia.

A recent clinical survey examining individuals living in thesame area as our study reported no unequivocal clinicalmanifestations of arsenicosis (Milton, 2003, Milton et al., 2005),however, the elevated arsenic concentrations found in aconsiderable number of nail and hair samples suggests thata number of individuals may be sub-clinically affected byarsenic exposure. Ingestion of groundwater is likely to be theprimary route for arsenic intake. The excellent correlations ofboth nail and hair arsenic concentrations with levels ofarsenic in the associated groundwater samples (Fig. 1) lendsupport to this premise. Positive correlations between nail orhair arsenic content and the drinking water arsenic concen-tration have also been observed in areas subject to both low(e.g. Karagas et al., 2000) and high (Kurttio et al., 1998; Schmittet al., 2005) levels of arsenic exposure. Furthermore, the hairand hair arsenic levels in individuals consuming groundwaterwith an arsenic concentration greater than the Cambodiandrinking water legal limit of 50 μg L−1 arsenic were signifi-cantly higher than in persons consuming waters containingless than 50 μg L−1 arsenic (Table 2). Similarly, Berg et al. (2007)reported significantly higher hair arsenic levels in peopleexposed to high arsenic groundwaters in Cambodia than thoseliving in areas of the country where low arsenic subsurfacewaters are located. Although these lines of evidence stronglysuggest that ingestion of groundwater is responsible for theaccumulation of arsenic in nail and hair tissues, other factorssuch as sex, smoking status, age and food intake may play aminor role. In a study of New Hampshire residents exposed tolow levels of arsenic in their drinking water, Karagas et al.(2000) found that sex and smoking status did not appear toinfluence toenail arsenic levels. Schmitt et al. (2005) also notedthat smoking had no significant effect on the nail arsenicconcentration of people living in Inner Mongolia that con-sumed high arsenic groundwaters, but a higher toenail arsenicconcentration in males was observed, which was attributed tothe greater volumes of water imbibed by men in their survey.The majority of participants in our study were non-smokers(nail: 86%, hair: 88%) and women (nail: 76%, hair: 73%),however, the limited size of our dataset precludes any detailedanalysis of the influence of these variables on nail/hair arsenicconcentrations. Although Schmitt et al. (2005) found nosignificant relationship between age and nail arsenic levels,other researchers studying populations using low arsenicwater supplies have detected a slight decline in nail arsenicconcentration with increasing age (Karagas et al., 2000;Slotnick et al., 2007). No statistically significant correlationbetween participant age and nail or hair arsenic content wasobserved in our study. Food intake may have an impact onnail arsenic concentration and some studies have found a

Page 7: Arsenic in hair and nails of individuals exposed to arsenic-rich groundwaters in Kandal province, Cambodia

174 S C I E N C E O F T H E T O T A L E N V I R O N M E N T 3 9 3 ( 2 0 0 8 ) 1 6 8 – 1 7 6

moderate link between dietary arsenic intake and toenailarsenic content (MacIntosh et al., 1997; Slotnick et al., 2007).Foodstuffs such as rice, the staple of the diet in much ofSoutheast Asia, prepared using arsenic-contaminated ground-water may represent an additional exposure pathway (Baeet al., 2002; Rahman et al., 2006). It should be noted that themajority participants in our investigation regularly consumedfish as part of their diet. Fish and seafood are known to containsignificant natural concentrations of arsenic, primarily pre-sent as the non-toxic arsenobetaine alongside other arseno-sugars/lipid (Cullen and Reimer, 1989). Nevertheless, theproportion of “fish arsenic” that may become incorporatedwithin the nail matrix appears debatable. WhileWilhelm et al.(2005) found that those residents living close to a coal-burningpower plant in Slovakia who consumed fish more than once aweek had higher nail arsenic levels than those individuals thatate no fish, Slotnick et al. (2007) calculated that fish con-sumption in Michigan residents did not significantly impactnail arsenic content. The arsenic content of local freshwaterfish eaten by the villagers that participated in our survey isusually relatively low (M. Sampson, unpublished data).

Nail and hair are rich in keratin, a scleroprotein that has ahigh proportion of cysteine residues. Arsenic has a high affinityfor the sulfhydryl groups found in cysteine, causing it toaccumulate in nail and hair tissue. Recent work looking atarsenic speciation in human nail and hair has involvedincubation of the sample in water at 90–100 °C and analysis ofthe arsenic extracted by HPLC-ICP-MS (Mandal et al., 2003; Raaband Feldmann, 2005). Arsenic species transformation duringextraction has been identified as a problem with such anapproach (Raab and Feldmann, 2005). We used XANES spectro-scopy to probe the arsenic speciation of selected nail and hairsamples directly. Given the affinity of arsenic for the sulfhydrylmoieties present in nail and hair, we expected the arsenicspeciation to be dominated by sulfur coordination. Althougharsenic bound to sulfur comprised the bulk of the arsenicspeciation in the hair samples analysed, this was the case inonly one of the nail samples (Table 4). Arsenite (As(III)-O) wasthe principal arsenic species determined by linear least squarefitting of the other three nails samples. The arsenic species inthe four nails samples showed considerable variability in theproportions of As–glut and As(III)-O fitted, but the arsenicspeciation in the three hair samples was relatively similar.Researchers using water extraction and subsequent HPLC-ICP-MS to determine arsenic speciation in nail and hair samplesfrom individuals living in arsenic-affected areas of India foundthat inorganic arsenic species predominate, with minoramounts of methylated arsenical making up the remainder ofthe extracted arsenic (Mandal et al., 2003; Raab and Feldmann,2005). Only very minor amounts of sulfur-bound arsenic weredetected in the work of Raab and Feldmann (2005). Limitedaccess to the entire arsenic inventory of the nail/hair sampleand/or species transformation during the extraction processmay account for the differences observed in the arsenic speciesdistribution of these studies compared to our own results.

Although XANES analysis allows for the direct determina-tion of arsenic speciation, we note that the technique has itsown limitations. Visual inspection of the XANES spectraseems to show that the position of the major absorptionpeak of the samples ties in closely with that of the arsenic–

glutathione standard (Fig. 3). Glutathione complexes arsenicthrough the sulfhydryl group of its cysteine residue and isused here as an analogue of the sulfhydrl groups of thenail and hair proteins. While the peak positions matchedclosely, the peak intensity of the samples was lower than thearsenic–glutathione standard, which may explain the con-siderable As(III)-O component fitted since its absorption peakintensity is lower than that of the samples. As such, the best fitof the spectra may yield spuriously high amounts of oxygen-bound As(III) because of the mismatch between peak inten-sities in the well-ordered standards and amorphous samples.With this in mind, the fit of the XANES spectramight be betterviewed as a guide rather than as a definitive result.

5. Conclusions

The arsenic concentrations of nail and hair samples collectedfrom residents of Kandal province, Cambodia, appear to beelevated and are suggestive of deleterious exposure to arsenic.Positive correlations between arsenic concentrations ingroundwater and nail/hair and the statistically significanthigher amounts of arsenic in nail and hair of individuals usinghigh arsenic (N50 μg L−1) groundwater compared to thoseparticipants resident in areas with lower arsenic (b50 μg L−1)groundwater, strongly indicates that consumption of theshallow subsurface waters is the principal source of arsenicintake. The XANES data appear to show that arsenic–sulfurcoordination makes a large contribution to the arsenic spe-ciation of the bulk of the nail and hair samples, as expectedgiven the sulfhydryl-rich nature of these biological materials.A sizeable proportion of arsenite could also be determined,although problems in the fitting process limit our certaintyof the exact species distribution. Nevertheless, our data showthe utility of nail and hair samples for use in arsenic biomarkerstudies in Cambodia, helping to provide an early warning ofarsenic exposure in regions where more apparent clinicalmanifestations of arsenicosis may take years to develop.Indeed, recent reports of the development of arsenicosisin Cambodia, including suspected arsenic-related deaths(Sampson et al., 2007), highlight the immediate importanceof nail and hair arsenic analysis as a means of rapidly iden-tifying at risk groups. It is clear that urgent action is needednow to prevent further human exposure to arsenic inCambodia.

Acknowledgements

AKerr-Fry award to AGG from the University of Edinburgh andGeorge Watson's College, Edinburgh and an EPSRC grant (GR/S30207/01) to DAP (PI), RAW, Jon Lloyd and David Vaughanprovided the funding for this research. The XANES spectro-scopywas supported by a CCLRC (now STFC) beam time award(44/258) at the Daresbury synchrotron radiation source. BobBilsborrow (STFC) is thanked for his help with the XANESspectroscopy data collection. We are grateful to Chris Jeynesand Karen Kirkby of the Ion Beam Centre at the University ofSurrey for providing access to the PIXE instrumentation.Ethical approval for this work was been granted by The

Page 8: Arsenic in hair and nails of individuals exposed to arsenic-rich groundwaters in Kandal province, Cambodia

175S C I E N C E O F T H E T O T A L E N V I R O N M E N T 3 9 3 ( 2 0 0 8 ) 1 6 8 – 1 7 6

University of Manchester Committee on the Ethics of Researchon Human Beings (ref. 05031).

R E F E R E N C E S

Agusa T, Kunito T, Fujihara J, Kubota R, Minh TB, Trang PTK, et al.Contamination by arsenic and other trace elements intube-well water and its risk assessment to humans in Hanoi,Vietnam. Environ Pollut 2006;139:95–106.

Bae M, Watanabe C, Inaoka T, Sekiyama M, Sudo N, Bokul MH,et al. Arsenic in cooked rice in Bangladesh. Lancet2002;360:1839–40.

Berg M, Tran HC, Nguyen TC, Pham HV, Schertenleib R, Giger W.Arsenic contamination of groundwater and drinking waiter inVietnam: a human health threat. Environ Sci Technol2001;35:2621–6.

Berg M, Stengel C, Trang PTK, Viet PH, Sampson ML, Leng M, et al.Magnitude of arsenic pollution in the Mekong and Red RiverDeltas — Cambodia and Vietnam. Sci Total Environ2007;372:413–25.

BGS and DPHE. In: Kinniburgh DG, Smedley PL, editors. Arseniccontamination of groundwater. Final report. BGS Tech. Rep.WC/00/19KeyworthBritish Geological Survey; 2001. 267 pp.

Brima EI, Jenkins RO, Haris PI. Understanding arsenic metabolismthrough spectroscopic determination of arsenic in humanurine. Spectrosc – Int J 2006;20:125–51.

Buschmann J, Berg M, Stengel C, Sampson ML. Arsenic andmanganese contamination of drinking water resources inCambodia: coincidence of risk areas with low relieftopography. Environ Sci Technol 2007;41:2146–52.

Calderon RL, Hudgens E, Le XC, Schreinemachers D, Thomas DJ.Excretion of arsenic in urine as a function of exposure toarsenic in drinking water. Environ Health Perspect1999;107:663–7.

Chen KLB, Amarasiriwardena CJ, Christiani DC. Determination oftotal arsenic concentrations in nails by inductively coupledplasma mass spectrometry. Biol Trace Elem Res1999;67:109–25.

Cullen WR, Reimer KJ. Arsenic speciation in the environment.Chem Rev 1989;89:713–64.

Dhar RK, Biswas BK, Samanta G, Mandal BK, Chakraborti D, Roy S,et al. Groundwater arsenic calamity in Bangladesh. Curr Sci1997;73:48–59.

Fleckman P. Basic science of the nail unit. In: Scher K, Daniel CR,editors. Nails: therapy, diagnosis, surgery. 2nd edn.Philadelphia: Elsevier; 1997. p. 37–54.

Hall M, Chen Y, Ahsan H, Slavkovich V, van Geen A, Parvez F, et al.Blood arsenic as a biomarker of arsenic exposure: results froma prospective study. Toxicol 2006;225:225–33.

Hopps HC. Biologic basis for using hair and nail for analyses oftrace elements. Sci Total Environ 1977;7:71–89.

Jamtveit B, Wogelius RA, Fraser DG. Zonation patterns of skarngarnets — records of hydrothermal system evolution. Geology1993;21:113–6.

Karagas MR, Tosteson TD, Blum J, Klaue B, Weiss JE, Stannard V,et al. Measurement of low levels of arsenic exposure:a comparison of water and toenail concentrations.Am J Epidemiol 2000;152:84–90.

Kubota R, Kunito T, Agusa T, Fujihara J, Monirith I, Iwata H, et al.Urinary 8-hydroxy-2′-deoxyguanosine in inhabitantschronically exposed to arsenic in groundwater in Cambodia.J Environ Monit 2006;8:293–9.

Kurttio P, Komulainen H, Hakala E, Kahelin H, Pekkanen J. Urinaryexcretion of arsenic species after exposure to arsenic presentin drinking water. Arch Environ Contam Toxicol1998;34:297–305.

MacIntosh DL, Williams PL, Hunter DJ, Sampson LA, Morris SC,Willett WC, et al. Evaluation of a food frequency questionnairefood composition approach for estimating dietary intake ofinorganic arsenic and methylmercury. Cancer EpidemiolBiomark Prev 1997;6:1043–50.

Mandal BK, Ogra Y, Suzuki KT. Speciation of arsenic in human nailand hair from arsenic-affected area by HPLC-inductivelycoupled argon plasma mass spectrometry. Toxicol ApplPharmacol 2003;189:73–83.

Milton AH. Baseline study and clinical examination ofarsenicosis among exposed population in Kandal Province.Unpublished report to Ministry of Health, Phnom Penh,Cambodia, 2003, pp. 1–71.

Milton AH, Raingsey PP, Mony KE, Fredericks D, Polya DA, GaultAG, et al. Human health and groundwater arsenic: KandalProvince, Cambodia. Proceedings of the EnvironmentalMineralogy, Geochemistry and Health Meeting of theMineralogical Society of Great Britain, Bath; 2005. p. 131.January.

Nordstrom DK. Public health — worldwide occurrences of arsenicin ground water. Science 2002;296:2143–5.

Olabanji SO, Ajose OA, Makinde NO, Buoso MC, Ceccato D, De PoliM, et al. Characterization of human fingernail elements usingPIXE technique. Nuclestrum Methods Phys Res Sect B2005;240:895–907.

Peach DF, Lane DW. A preliminary study of geographic influenceon arsenic concentrations in human hair. Environ GeochemHealth 1998;20:231–7.

Polya DA, Gault AG, Bourne NJ, Lythgoe PR, Cooke DA. CoupledHPLC-ICP-MS analysis indicates highly hazardousconcentrations of dissolved arsenic species are present inCambodian wellwaters. In: Holland JG, Tanner SD, editors.Plasma source mass spectrometry: applications andemerging technologies. Cambridge: Royal Society of Chemistry;2003. p. 127–40.

Polya DA, Gault AG, Diebe NH, Feldman P, Rosenboom JW, GilliganE, et al. Arsenic hazard in shallow Cambodian groundwaters.Mineral Mag 2005;69:807–23.

Raab A, Feldmann J. Arsenic speciation in hair extracts.Anal Bioanal Chem 2005;381:332–8.

Rahman MA, Hasegawa H, Rahman MA, Rahman MM, Miah MAM.Influence of cooking method on arsenic retention in cookedrice related to dietary exposure. Sci Total Environ2006;370:51–60.

Rodushkin I, Axelsson MD. Application of double focusing sectorfield ICP-MS for multielemental characterization of humanhair and nails. Part II. A study of the inhabitants of northernSweden. Sci Total Environ 2000;262:21–36.

Ryabukhin YS. Activation analysis of hair as an indicator ofcontamination of man by environmental trace elementpollutants. IAEA Report, IAEA/RL/50, Vienna; 1978.

Samanta G, Sharma R, Roychowdhury T, Chakraborti D. Arsenicand other elements in hair, nails, and skin-scales of arsenicvictims in West Bengal, India. Sci Total Environ2004;326:33–47.

Sampson ML, Sosamrach K, Shantz A. Policy response and thediscovery of arsenicosis in Cambodia. Proceedings of theEAWAG/University of Manchester Workshop on Arsenicin Groundwaters of South-East Asia with Emphasis onCambodia and Vietnam, Manchester; 2007. 29–30th October.

Schmitt MT, Schreinemachers D,WuK, Ning Z, Zhao B, Le XC, et al.Human nails as a biomarker of arsenic exposure from wellwater in Inner Mongolia: comparing atomic fluorescencespectrometry and neutron activation analysis. Biomarkers2005;10:95–104.

Sera K, Futatsugawa S, Murao S. Quantitative analysis ofuntreated hair samples for monitoring human exposure toheavy metals. Nucl Instrum Methods Phys Res Sect B2002;189:174–9.

Page 9: Arsenic in hair and nails of individuals exposed to arsenic-rich groundwaters in Kandal province, Cambodia

176 S C I E N C E O F T H E T O T A L E N V I R O N M E N T 3 9 3 ( 2 0 0 8 ) 1 6 8 – 1 7 6

Slotnick MJ, Meliker JR, AvRuskin GA, Ghosh D, Nriagu JO. Toenailsas a biomarker of inorganic arsenic intake from drinking waterand foods. J Toxicol Environ Health A 2007;70:148–58.

Smith AH, Lingas EO, Rahman M. Contamination ofdrinking-water by arsenic in Bangladesh: a public healthemergency. Bull WHO 2000;78:1093–103.

Watt F, Landsberg JP. Nuclear microscopy — biomedicalapplications. Nucl Instrum Methods Phys Res Sect B1993;77:249–60.

WHO. Environmental health criteria 18 — Arsenic. Geneva: WHO;1981.

Wilhelm M, Pesch B, Wittsiepe J, Jakubis P, Miskovic P, Keegan T,et al. Comparison of arsenic levels in fingernails with urinaryAs species as biomarkers of arsenic exposure in residentsliving close to a coal-burning power plant in Prievidza District,Slovakia. J Expos Anal Environ Epidemiol 2005;15:89–98.