Chronic kidney disease of uncertain aetiology

13 Pages·2013·1.08 MB·English

Jayatilakeetal.BMCNephrology2013,14:180
http://www.biomedcentral.com/1471-2369/14/180
RESEARCH ARTICLE Open Access
Chronic kidney disease of uncertain aetiology:
prevalence and causative factors in a developing
country
Nihal Jayatilake1, Shanthi Mendis2*, Palitha Maheepala1, Firdosi R Mehta3 and On behalf of the CKDu National
Research Project Team
Abstract
Background: Thisstudy describes chronic kidney disease of uncertainaetiology (CKDu), which cannotbe attributed
to diabetes, hypertensionor otherknownaetiologies,that has emerged inthe North Central region ofSri Lanka.
Methods: A cross-sectional study was conducted, to determine theprevalence of and risk factors for CKDu. Arsenic,
cadmium, lead, selenium, pesticides and otherelements were analysed inbiologicalsamples from individuals with
CKDu and compared withage- and sex-matched controls intheendemic and non-endemic areas.Food, water, soil
and agrochemicals from both areas were analysed for heavy metals.
Results: The age-standardisedprevalence of CKDu was 12.9% (95% confidence interval [CI]=11.5% to 14.4%)in
males and 16.9%(95% CI=15.5% to 18.3%) infemales. Severe stagesofCKDu were more frequent in males (stage
3: males versus females=23.2% versus 7.4%; stage4: males versus females=22.0% versus 7.3%;P<0.001). The risk
was increased in individuals aged>39 years and thosewho farmed (chena cultivation)(OR [odds ratio]=1.926,
95% CI=1.561to 2.376 and OR=1.195, 95% CI=1.007 to 1.418respectively,P<0.05). The risk was reduced in
individuals who were male or who engaged in paddy cultivation (OR=0.745, 95% CI=0.562to 0.988and
OR=0.732, 95%CI=0.542 to 0.988 respectively, P<0.05).The mean concentration of cadmium in urine was
significantly higher in thosewithCKDu (1.039 μg/g) comparedwith controls in theendemicand non-endemic
areas (0.646 μg/g, P<0.001and 0.345μg/g, P<0.05) respectively.Urine cadmium sensitivity and specificitywere
70% and 68.3% respectively (area under the receiver operating characteristic curve=0.682, 95%CI=0.61 to 0.75,
cut-off value ≥0.397μg/g). A significant dose–effect relationshipwas seen between urine cadmium concentration
and CKDu stage (P<0.05). Urine cadmium and arsenic concentrations in individuals with CKDu were atlevels
known to cause kidney damage.Food items from the endemic area contained cadmium and lead above reference
levels. Serum selenium was <90 μg/l in63% of those withCKDu and pesticides residues were above reference
levels in31.6% of those withCKDu.
Conclusions: These results indicate chronic exposureof people inthe endemic area to low levels ofcadmium
through thefood chain and also to pesticides. Significantly higher urinary excretion ofcadmium inindividuals with
CKDu, and the dose–effectrelationshipbetween urine cadmium concentration and CKDu stages suggest that
cadmium exposure is a risk factor for the pathogensis of CKDu. Deficiency of selenium and genetic susceptibility
seen inindividuals with CKDu suggestthat they maybe predisposing factors for the development of CKDu.
Keywords: Arsenic, Cadmium,Chronickidney disease, Kidney disease ofuncertainaetiology, Heavy metals, Lead,
Pesticides
*Correspondence:[email protected]
2ManagementofNoncommunicableDiseases,WorldHealthOrganization,
Geneva,Switzerland
Fulllistofauthorinformationisavailableattheendofthearticle
©2013Jayatilakeetal.;licenseeBioMedCentralLtd.ThisisanOpenAccessarticledistributedunderthetermsoftheCreative
CommonsAttributionLicense(http://creativecommons.org/licenses/by/2.0),whichpermitsunrestricteduse,distribution,and
reproductioninanymedium,providedtheoriginalworkisproperlycited.
Jayatilakeetal.BMCNephrology2013,14:180 Page2of13
http://www.biomedcentral.com/1471-2369/14/180
Background The objectives of this study were to: determine the
The study reported here describes an apparently new prevalence of and identify the risk factors for CKDu;
form of chronic kidney disease, which cannot be attrib- compare CKDu cases and controls with regard to expos-
uted to diabetes, hypertension or other known aetiol- ure to heavy metals/metalloids and pesticides; and ana-
ogies, that has emerged in the North Central region of lyse food, water, soil, fertilizers and weedicides for heavy
Sri Lanka [1]. Chronic kidney disease of uncertain aeti- metals, in order to determine whether the levels are
ology (CKDu) is slowly progressive, probably starting in above stipulated reference values.
the second decade of life, and asymptomatic until ad- A population prevalence study was conducted in three
vanced. Peripheral oedema and hypertension are late districts in the endemic area over the period 2010–2012
features. The main histopathological features include (Figures 1 and 2). Arsenic, cadmium, lead and other
tubularatrophy,interstitialmononuclearcellinfiltration metals,elementsandpesticideresiduesthatarepotential
and interstitial fibrosis [2]. These histological features nephrotoxins [6-9,21-24], were analysed in biological
suggest that nephrotoxins play a key role in the aeti- samples. The results from individuals with CKDu were
ologyofCKDu. compared with those from controls in the endemic area
Potential nephrotoxins are widely distributed in the and a non-endemic area (Hambantota where CKDu has
environment. Exposure to environmental and other not been reported). The demographic characteristics of
nephrotoxins, such as herbal medicines and analgesics, the three groups are shown in Table 1. All individuals
are known to play a role in the aetiology of chronic with CKDu (n=733) had a blood pressure below 160/
kidney disease. Defining their exact role in the aetiology 100 mmHg; 6.8% of these individuals were on treatment
of kidney disease is a challenge. Environmental toxins for high blood pressure and had a blood pressure below
implicated in kidney damage include heavy metals, such <140/90 mmHg. Three-quarters of those who were on
as arsenic, cadmium, lead and uranium; mycotoxins pro- treatmentwereonangiotensin-convertingenzymeinhib-
duced by fungi in improperly stored foods; air pollut- itors. Others were on calcium channel blockers, beta-
ants, such as tobacco smoke; and pesticides, such as
chlorpyriphos,diazinonandpropanil[3-9].
Over the last 8 years, several studies have been carried
out to determine the prevalence, nature and causes of
CKDu in Sri Lanka [10-20]. These studies include
hospital-based reviews of case series, in which high-risk
areas in North Central Province were identified. The
hospital studies gave rise to population-based surveys,
which showed that the populations affected by CKDu
are scattered in the North Central region of the country
[14]. Some studies have reported the point prevalence of
CKDu to be about 2–3% among those above 18 years of
age [20]. Case–control and cross-sectional studies have
provided some insight into associations with the condi-
tion. Specific evaluations of exposure to organophosphate
andmycotoxinshavebeenconducted.Inacross-sectional
study, there was evidence of greater inhibition of acetyl
cholinesteraseamongpatientswithchronicrenaldysfunc-
tioninareasofhighprevalenceofCKDu[17].Ochratoxin,
anaturallyoccurringmycotoxinwithnephrotoxicproper-
ties, was not found to be a contaminant of food in the
region[13,16].Somestudieshaveshownhighlevelsofen-
vironmentalcadmium,lead,aluminiumandfluorideinre-
gionswithhighratesofthecondition[14,15,17,19].
Recognisingthegravityofthepublichealththreatcaused
byCKDu,in2010,theMinistryofHealthSriLanka,incol-
laboration with the World Health Organization (WHO),
launched a national research project with the aim of
investigating the prevalence and aetiology of CKDu in Figure1MapofSriLanka,showingthelocationofAnuradhapura,
PolonnaruwaandBadulladistricts,intheendemicarea,and
Sri Lanka, with aview to developing appropriate preven-
Hambantotadistrict,inthenon-endemicarea.
tivestrategies.
Jayatilakeetal.BMCNephrology2013,14:180 Page3of13
http://www.biomedcentral.com/1471-2369/14/180
Number invited to participate in the study from 220 randomly selected households
n=6698
Response rate 74% (n=4957)
ACR assay 30 mg/g on two occasions No diabetes, other kidney
and satisfy other criteria of CKDu disease, CKDu and snake bite=4044
(n=733) (excluded=180, missing data n=25)
Urine arsenic, cadmium and lead n=495 n=132
Urine for other elements n=107 n=0
Serum selenium, strontium etc n=171 n=0
Hair/nails arsenic, cadmium lead n=80 n=48
Urine pesticide residues n=57 n=0
Figure2Flowchartofthestudyconductedinthreedistricts(Anuradhapura,PolonnaruwaandBadulla)intheendemicarea.
ACR=albumin–creatinineratio.
blockers or diuretics, either alone or in combination (cid:1) normalglycosylatedhaemoglobin (HbA ;<6.5%)
1c
with angiotensin-converting enzymeinhibitors. (cid:1) ifontreatment forhypertension,bloodpressure
In both endemic and non-endemic areas, water, food, below<140/90mmHg;ifnotontreatment for
tobacco, soil and agrochemicals were also analysed for hypertension, bloodpressurebelow<160/
heavy metals and metalloids, to determine whether they 100mmHg.
were within stipulatedreferencelevels.
CKDu wasgradedasfollows:
Methods
Casedefinition (cid:1) Stage1:persistentalbuminuria(i.e. ACR≥30mg/g
Thefollowing case definition of CKDu was used. Partici- ininitialand repeat urine sample) andestimated
pants who had persistent albuminuria, i.e. albumin–cre- glomerularfiltration rate(eGFR),usingthe Chronic
atinine ratio (ACR) ≥30 mg/g in an initial urine sample Kidney DiseaseEpidemiology collaboration(CKD-
and at a repeat visit, were considered to have CKDu if EPI)equation [25]>90ml/min/1.73m2
they satisfied thefollowing criteria: (cid:1) Stage2:persistentalbuminuriaand eGFR60–89 ml/
min/1.73 m2
(cid:1) nopasthistoryofglomerulonephritis, (cid:1) Stage3:persistentalbuminuriaand eGFR30–59 ml/
pyelonephritis,renal calculiorsnakebite min/1.73 m2
(cid:1) notontreatment fordiabetes (cid:1) Stage4:persistentalbuminuriaand eGFR<30 ml/
min/1.73 m2.
Table1DemographiccharacteristicsofCKDucasesinthe
endemicarea,controlsfromtheendemicareaand Populationprevalencestudy
controlsfromthenon-endemicarea Ethical approvalfor the study wasobtained fromtheEth-
Characteristic Endemicarea Non-endemicarea ical Review Committee of the Sri Lanka Medical Associ-
CKDucases Controls Controls ation.Allparticipantsgavewrittenconsentforthestudy.
Totalnumber 733 4044 250 Six divisional secretariat areas (administrative divi-
sions) were selected randomly from three districts in the
Males,% 37.1 42.5 56
endemicarea.Twenty-twovillages(GramaNiladariareas)
Age(years),mean(SD) 39.1(14.2) 43.7(13.9) 35.5(14.0)
were selected randomly from the six divisions. Using the
Farmer,% 38 43.9 18.3
electoral lists,100householdsfromeachvillagewereran-
BMI,kg/m2,mean(SD) 22.3(4.6) 21.1(4.1) 21.7(4.4) domly selected for the study. Males and females aged be-
ACR≥30mg/g 733 0 0 tween 15 and 70 years (n=6698), with no diagnosed
ACRalbumin–creatinineratio,BMIbodymassindex,SDstandarddeviation. diabetes, were invited to participate and 74% responded
Jayatilakeetal.BMCNephrology2013,14:180 Page4of13
http://www.biomedcentral.com/1471-2369/14/180
(Figure 2, flow diagram). Trained interviewers used a Soil, phosphate fertilizer, pesticides and weedicides
survey questionnaire to gather information on age, sex, were analysed for arsenic, cadmium and lead. Soil sam-
marital status, education, occupation, smoking, alcohol ples were obtained from paddy fields, other types of
consumption, current residence, duration of residence in cultivations, and reservoirs in the endemic (n=88) and
the study area, source of drinking water, storage con- non-endemic (n=41)areas.
tainers for drinking water, exposure to agrochemicals,
history of snake bite, glomerulonephritis, pyelonephritis, Specimenhandlingandanalysis
renal calculi, use of medications including herbal medi- Sampleswerecollectedinuncontaminatedcollectionvials
cines, and past medical history. Height was measured to and stored frozen (−20°C) until transferto the laboratory.
the nearest 0.1 cm. Weight was measured to the nearest All analyses were performed in a contract laboratory (La-
0.1kg,usingacalibratedweighingscale.Participantswore boratoryofPathophysiologyoftheUniversityofAntwerp,
light clothes and no shoes. A medical officer verified the Belgium), which has an external quality control scheme
medical information gathered and measured the blood foranalysisoftraceelements.
pressureafter15minutes’rest,usingamercurysphygmo- Measurements of arsenic, cadmium, lead and other el-
manometer.Theaverage oftwo readings taken 5 minutes ements in urine, water, vegetables, agrochemicals and
apart was used. Urine ACR, HbA and serum and urine soil, was performed by inductively coupled plasma mass
1c
creatinineconcentrationswerealsomeasured. spectrometry(MS).
Serumanalyseswereperformedbyelectrothermalatomic
Analyticalstudies absorption spectrometery. Limits of detection for alumin-
Arsenic,cadmiumandleadinurine,blood,hairandnails ium, strontium, chromium and selenium were 0.1 μg/l,
Arsenic, cadmium and lead concentrations were analysed 0.5μg/l,0.01μg/land1μg/lrespectively.
in urine in a randomly selected subset of CKDu cases
(n=495) and randomly selected matched controls from Determinationofpesticideresiduesinurine
the endemic area, as well as from the non-endemic area Sampleswereshippedindryiceandstoredat−18°Cuntil
(n=250). analysis. Analysis used validated liquid chromatography
Urine sodium, potassium, calcium, magnesium, cop- withtandemMS(LC-MS/MS),gaschromatography-mass
per, zinc and titanium concentrations were analysed in a spectroscopy(GC-MS)andgaschromatographywithtan-
randomly selected subset of CKDu cases (n=148). Their demmassspectrometry(GC-MS/MS)methods.
serum was also analysed for selenium, aluminium, stron- Further details of sample preparation and analytical
tiumandchromium.Hairand nailsampleswereanalysed techniquesareprovidedintheAdditionalfile1.
for cadmium, arsenic and lead in a subset of CKDu cases
(n=80)andcontrolsfromtheendemicarea(n=48). Statistics
Urine samples from CKDu cases (n=57) and controls Thenormalityofdatadistributionwasassessedwithhisto-
from the non-endemic area (n=39) were analysed for grams. All data on metals had skewed distributions. After
pesticide residues (2,4-D, 2,4,5-T, 2,4,5-trichlorophenol, removalofasmallnumberofoutliers,log-transformations
isopropoxyphenol, pentachlorophenol, 3,5,6-trichloropy- were used to normalise the data, in order to conduct
ridinol,p-nitrophenol,1-naphthol,2-naphthol,glyphosate, statistical analyses. The mean, median, minimum and
aminomethylphosphonicacid(AMPA)). maximumvaluesarereportedonoriginaldata;t-testsof
log-transformed values were used to test differences in
Arsenic,cadmiumandleadinwater,food,tobacco, quantitative variables. The results were also confirmed
pasture,weeds,soil,fertilizer,weedicidesandpesticides bynon-parametricWilcoxonrank-sumtest.
Arsenic, cadmium and lead were analysed in samples A multiple logistic regression model was fitted for the
(n=234) obtained from endemic and non-endemic areas. CKDu definition.Themodel incorporatedcharacteristics
They comprised 99 sources of drinking water for indi- of interest, including age, sex, education, smoking, illicit
viduals with CKDu (from ground wells, tube wells and alcohol consumption, occupation, type of agriculture,
natural springs), 123 other sources of water (from years of agriculture, source of drinking water, drinking
ground wells, tube wells, irrigation canals, reservoirs, water from paddy fields, exposure to fertilizer, exposure
naturalsprings)fromtheendemicarea,and12fromthe to weedicides and pesticides, type of water container,
non-endemicarea. whether using protection against agrochemicals, and
Rice;pulses;vegetables,includingleafyvegetables,coco- months living in the district. All were entered as categor-
nut, yams and roots (e.g. kohila, lotus); freshwater fish; ical variables, except months living in the district. These
tobacco; pasture; and weeds obtained from endemic data analyses wereperformed usingStata 11 and P values
(n=119) and non-endemic (n=32) areas were analysed oflessthan0.05wereconsideredstatisticallysignificant.A
forarsenic,cadmiumandlead. receiver-operatingcharacteristic(ROC)curvewasusedto
Jayatilakeetal.BMCNephrology2013,14:180 Page5of13
http://www.biomedcentral.com/1471-2369/14/180
calculate the area under the ROC curve (AUC), to deter- Table2Summaryresultsoflogisticregressionanalysis
mine the cut-off values for cadmium and selenium with forexposures
the best sensitivity and specificity. A multinomial logistic Exposure 95%CI
regression was used to assess the dose–effect relationship OR Lower Upper P n
between metal exposure and the outcome CKDu grade. value
Theanalyseswereadjustedforageandsex. Education
Noeducation Reference 174
Results
Schoolgrades1–9 0.900 0.612 1.323 0.594 4374
Populationprevalencestudy
Higher 1.201 0.588 2.452 0.614 74
The age-standardised prevalence of CKDu was higher in
Smoking
females 16.9% (95% confidence interval [CI]=15.5% to
18.3%)thaninmales12.9%(95%CI=11.5%to14.4%;P= Never Reference 3480
0.001). About 37% of those with CKDu were male. The Current/former 1.072 0.813 1.415 0.619 1126
distribution of CKDu stages 1 to 4 in males was 27.0%, Illicitalcoholconsumption
27.9%,23.2%and22.0%andinfemales53.3%,32.0%,7.4%
Never Reference 3701
and 7.3%, respectively. More severe stages of CKDu were
Occasional/frequent/past 1.184 0.905 1.548 0.216 874
seen more frequently in males (stage 3: males versus fe-
Occupation
males=23.2%versus7.4%;stage4:malesversusfemales=
22.0% versus 7.3%; P<0.001). In both sexes, the preva- Other Reference 2816
lenceincreasedwithincreasingage(P<0.001).Thepreva- Farmer 1.195 1.007 1.418 0.041 1780
lence in the three districts was 15.1% in Anuradhapura, Agriculturetype
20.6%inPolonnaruwaand22.9%inBadulla.
Non-paddy Reference 315
There was a family history of kidney disease in parents
Paddy 0.732 0.542 0.988 0.042 2620
or siblings in 20% of individuals with CKDu; 2.1% of in-
Yearsworkingin
dividuals with CKDu had a history of ischaemic heart
agriculture
disease and/or cerebrovascular disease; 0.4% had a his-
<10 Reference 660
tory of long-term use of herbal medicines for hyperten-
sion; 1.8% had a history of long-term use of aspirin; and 10–19 0.834 0.603 1.152 0.271 777
0.6% had a history of long-term use of analgesics. Being 20–49 1.092 0.777 1.535 0.611 1182
male reduced the risk of CKDu (odds ratio [OR]=0.745, ≥50 1.322 0.462 3.785 0.602 22
95% CI=0.562 to 0.988, P<0.05), and being older than
Sourceofdrinkingwater
39 years increased the risk of CKDu (OR=1.926, 95%
Notwell Reference 798
CI=1.561 to 2.376, P<0.001). When separate logistic
Well 0.971 0.785 1.202 0.793 3819
regressions were run for each potential exposure, only
occupation type (being a chena cultivation farmer in- Waterstoragecontainer
creased the OR by 19.5%) and type of agriculture (en- Others Reference 1741
gaging in paddy cultivation compared to cultivation of Aluminium 1.03 0.87 1.22 0.715 2879
vegetables and other crops [chena cultivation] decreased
Protectionfrom
theORby26.8%)weresignificant (Table2). agrochemicals
Yes Reference 191
Arsenic,cadmium,leadandotherelementsinurine
No 1.011 0.661 1.546 0.959 4271
In CKDu cases, the concentration of cadmium in urine
Separatelogisticregressionshavebeenrunperexposurevariable;OR<1
was significantly higher compared to controls, in both meansprotective,andOR>1meansthattheexposureincreasestheoddsof
theendemicandthenon-endemicareas(Table3).Among CKDu.Thetotalnumber(n)ofobservationsvariesperexposure,owingto
missingdata.Allresultsareadjustedforsexandage.Forallanalyses,malesex
CKDu cases, the concentration of cadmium in urine was
wasfoundtobeprotectiveandtheriskincreasedwithage.
positively correlated with lead (r=0.62, P<0.001) and ar- CIconfidenceinterval,ORoddsratio.
senic concentrations in urine (r=0.28, P<0.001). There BeingmalereducedtheriskofCKDu(OR=0.745,95%CI=0.562to0.988;P<
0.05),andbeing>39yearsincreasedtheriskofCKDu(OR=1.926,95%CI=
was no significant difference in urine arsenic and lead 1.561to2.376,P<0.001).
concentrations in CKDu cases compared to controls. The
sensitivityandspecificityforconcentrationsofcadmiumin concentration of arsenic in urine were 90% and 23.2%
urine were 80% and 53.6% respectively (AUC=0.682, respectively (AUC=0.64, 95% CI = 0.58 to 0.71, cut-off
95% CI=0.61 to 0.75, cut-off value ≥0.23 μg/g; Figure 3). value≥88.57μg/g).Theconcentrationofleadinurinewasa
At a cut-off value of ≥0.397 μg/g, sensitivity was 70% and poorpredictorofCKDu(AUC=0.53,95%CI0.38to0.67).
specificity 68.3%. The sensitivity and specificity for the Dose–responseanalysisshowedthatcadmiumexposureisa
Jayatilakeetal.BMCNephrology2013,14:180 Page6of13
http://www.biomedcentral.com/1471-2369/14/180
Table3Urineconcentrationofarsenic,cadmiumandleadforCKDucasescomparedwithcontrolsfromtheendemic
andnon-endemicareas
Mean,median(range)ofconcentrationinurine(μg/gcreatinine)
Arsenic Cadmium Lead
CKDucases(n=495) 45.447,26.3(0.4to616.6) 1.039,0.695(0.005to8.93) 1.153,0.95(0.04to8.53)
Controlsfromendemicarea(n=132) 92.443,6.99(0.2to966.29) 0.646,0.18,(0.005to5.13)a 1.254,0.793(1.21to6.64)
Controlsfromnon-endemicarea(n=250) 56.572,42.025(5.38to350.28) 0.345,0.265(0.005to2.079)b 2.099,1.434(0.277to20.9)
aUrinecadmiumconcentrationofcasescomparedtocontrolsfromendemicareaP<0.001.
bUrinecadmiumconcentrationofcasescomparedtocontrolsfromnon-endemicareaP<0.05.
risk factor for the development of CKDu: P=0.019 for (reference range=54 μg/l to 163 μg/l). A serum selen-
stage3andP=0.024forstage4.Therewasnosignificant ium concentrationof90μg/lisrequiredtoreachthemax-
dose–effect relationship between the concentration of ar- imumlevelofglutathioneperoxidise[26].Abouttwo-thirds
senic,leadorseleniuminurineandthestageofCKDu. (63%) of subjects had selenium levels below this cut-off
Urine concentrations of sodium, potassium, calcium, value. Serum strontium levels were above normal limits
magnesium, copper, zinc, and titanium in CKDu cases (mean=83.17 μg/l, standard deviation [SD]=32.15 μg/l;
were within normallimits(Additional file2). reference range=14 μg/l to 84 μg/l). The sensitivity and
specificityforserumseleniumwere80%and60%respect-
Serumaluminium,chromium,seleniumandstrontiumin ively(AUC=0.789,cut=offvalue≥94.3μg/l;Figure3).
CKDucases
Serumaluminiumandchromiumlevelswerewithinnor- Cadmiumandarsenicinhairandnails
mal limits (Additional file 2). Serum selenium levels in A significantly higher cadmium concentration was also
subjects with CKDu ranged from 50.0 μg/l to 121.8 μg/l seeninthenailsofCKDucases(n=80,mean=0.017μg/g,
Figure3ROCcurvesgeneratedwithurinearsenic,cadmiumandleadandserumseleniumconcentrations.
Jayatilakeetal.BMCNephrology2013,14:180 Page7of13
http://www.biomedcentral.com/1471-2369/14/180
median=0.007 μg/g, minimum=0.001 μg/g, maximum= taken from a canal and a reservoir, the cadmium con-
0.347 μg/g) compared to controls (n=48) from the en- centration was 3.46 μg/l in one sample from a reservoir
demicarea(mean=0.009μg/g,median=0.001μg/g,mini- and the lead concentration was 12.3 μg/l in one sample
mum=0.001μg/g,maximum=0.091μg/g;P<0.05). from a reservoir in the endemic area. All other samples
Arsenic levels in hair weresignificantly higherinCKDu from wells, tube wells, irrigation canals, pipe-borne
cases (n=80; mean=0.144 μg/g, median=0.139 μg/g, water, reservoirs and natural springs, including those
minimum=0.00 μg/g, maximum=0.452 μg/g), compared taken from the non-endemic area, had normal arsenic,
to controls (n=48) from the endemic area (mean= cadmiumandleadlevels.
0.125 μg/g, median=0.103 μg/g, minimum=0.006 μg/g,
maximum=1.214μg/g;P<0.05). Arsenic,cadmiumandleadinfood,tobaccoleaves,
pastureandweeds
Arsenic,cadmium,leadanduraniuminwater(endemic Levels of cadmium in rice in both endemic and non-
arean=222,non=endemicarea,n=12) endemic areas were below the allowable limit (0.2 mg/
Levels of cadmium, lead and uranium in sources of kg; Figure 5). The maximum concentration of cadmium
drinking water (Figure 4)used by individualswith CKDu in vegetables in the endemic area and in the non-
(n=99)werewithinnormallimits.Arsenicwasborderline endemic areas was 0.322 mg/kg and 0.063 mg/kg re-
or raised in four samples (9.9 μg/l, 10.2 μg/l, 10.5 μg/l, spectively. Levels of cadmium in certain vegetables such
13.4 μg/l). Repeat analysis (n=32) from the four sources as lotus root, and in tobacco, were high. Levels of cad-
showednormalarseniclevels. mium in lotus and tobacco were higher in endemic than
In water samples from other sources, the arsenic con- in non-endemic areas (lotus: mean=0.413 mg/kg versus
centration was 22.2 μg/l and 9.8 μg/l in two samples 0.023 mg/kg, median=0.066 mg/kg versus 0.023 mg/kg,
Figure4Concentrationofarsenic,cadmium,leadanduraniuminwateraintheendemicarea(n=222)andnon-endemicarea(n=12).
Endcanal=endemicareacanal;Enddrink=endemicareadrinkingwater;Endreser=endemicareareservoir;Endspring=endemicareaspring;
nonenddrink=non-endemicareadrinkingwater.Horizontallineswithintheboxesrepresentthemedianvalues.Theendsofthesolidlines
extendingoneithersideoftheboxesrepresenttheminimumandthemaximum.Thedarkdotsareoutliers;definedasbeingmorethan1.5
interquartilerangesawayfromthebox.Theinterquartilerangeisthedistancebetweentheupperpartoftheboxandthelowerpartofthebox.
aReferencelimits:arsenic<10μg/l,cadmium<3μg/l,lead<10μg/l,uranium<2μg/l[21].
Jayatilakeetal.BMCNephrology2013,14:180 Page8of13
http://www.biomedcentral.com/1471-2369/14/180
Figure5(Seelegendonnextpage.)
Jayatilakeetal.BMCNephrology2013,14:180 Page9of13
http://www.biomedcentral.com/1471-2369/14/180
(Seefigureonpreviouspage.)
Figure5Contentofarsenic,cadmiumandleadinfood,tobaccoleaves,pastureandweedsfromtheendemicarea.Horizontallines
withintheboxesrepresentthemedianvalues.Theendsofthesolidlinesextendingoneithersideoftheboxesrepresenttheminimumandthe
maximum.Thedarkdotsareoutliers;definedasbeingmorethan1.5interquartilerangesawayfromthebox.Theinterquartilerangeisthe
distancebetweentheupperpartoftheboxandthelowerpartofthebox.(Thecadmiumandleadcontentincertainfooditemsexceededthe
maximumstipulatedreferencevaluea).aThemaximumlevelsofcadmiumpermittedbytheCodexAlimentariusforvegetablesis0.2mg/kg
[22,23]andbytheCommissionoftheEuropeanCommunitiesis0.05mg/kg[24].Themaximumconcentrationofcadmiumstipulatedforcertain
typesoffishbytheCommissionoftheEuropeanCommunitiesis0.05mg/kg[24].Themaximumconcentrationofleadstipulatedforvegetables
bytheCommissionoftheEuropeancommunitiesis0.10mg/kg[24].
maximum=1.50 mg/kg versus 0.03 mg/kg; tobacco: socioeconomic groups may have an influence on the ex-
mean=0.351 mg/kg versus 0.316 mg/kg, median= cretionofheavymetalsandoxidativestressonthekidney.
0.351 mg/kg versus 0.316 mg/kg, maximum=0.44 mg/ This would make the kidneys more vulnerable to CKDu,
kg versus 0.351 mg/kg in endemic versus non-endemic resulting in a higher prevalence in females. On the other
areasrespectively). hand, male sex has been reported to be a risk factor for
progression to end-stage renal disease [27], and this may
Arsenic,cadmiumandleadinsoilandagrochemicalsin partly explain the occurrence of more severe stages of
theendemicandnon-endemicareas CKDuinmen.
Thelevel ofcadmiumin surface soil inthe endemic area Previous studies have reported a family history of
(n=94,excludingsamplesfromreservoirs),was1.16μg/g chronic kidney disease, ayurvedic treatment, and history
compared to 0.49 μg/g in the non-endemic area (n=45, ofsnakebiteassignificantpredictorsforCKDu[10,12,13].
excludingsamplesfromreservoirs)(Additionalfile2). In the present study, older age, being female and being a
chenacultivationfarmerincreasedtheriskofCKDu.Fam-
Pesticideresiduesinurine ily history was positive in one-fifth of those with CKDu,
Pesticideresiduesweredetectedintheurinefromindivid- and a history of snake bite was one of the exclusion cri-
uals with CKDu (n=57). The frequency of detection of teria. Long-term use of herbal medicines or analgesics
2,4-D, 3,5,6-trichloropyridinol, p-nitrophenol, 1-naphthol, was reported in only a very small percentage of those
2-naphthol, glyphosate, AMPA was 33%, 70%,58%,100%, with CKDu. Fanconi syndrome and other hereditary
100%, 65% and 28% respectively). Isopropoxyphenol, kidney diseases have not been reported in communities
2,4,5-trichlorphenolandpentachlorphenolwerebelowde- inthisregion.
tection limits. The proportions of CKDu cases withpesti- Previous studies have reported divergent information
cidelevelsabovereferencevaluesareshowninTable4. on the role of cadmium in the causation of CKDu
[14,15,19,20]. In the present study, individuals with
Discussion CKDu excreted significantly higher levels of cadmium
The prevalence of CKDu found in this study (females compared to those in the control group, in both the en-
16.9%, males 12.9%) was higher than that reported previ- demic and non-endemic areas. Controls in the endemic
ously(2%to3%)[20].Althoughtheprevalenceinfemales area compared to those in the non-endemic area also
was higher, more severe stages of CKDu were seen more had significantly higher urinary excretion of cadmium.
often in males. The reason for this discrepancy is not The sensitivity and specificity for urine cadmium were
clear. Factors such as low iron stores in females in lower 80% and 53.6% respectively (AUC=0.682, cut-off value
Table4PesticideresiduesinurineofCKDucases(n=57)
Parentcompound Biomarker Referencelimit(μg/l) CKDucases(μg/l), CKDucasesabovereference
(minimum,maximum) limit(%)
2,4-D 2,4-D <0.3 0.5,0.62 3.5
Pentachlorophenol Pentachlorophenol <2 0.3,2.2 1.7
Chlorpyrifos 3,5,6-trichloropyridinol <11.3 0.5,34.7 10.5
Parathion p-nitrophenol <25 0.5,8.88 0
Carbarylnaphthalene 1-naphthol <19.7 0.5,45.1 10.5
Naphthalene 2-naphthol, <17.1 0.94,47.88 10.5
Glyphosate Glyphosate <2 0.075,3.36 3.5
Glyphosate AMPA <0.5 0.075,2.65 14
Jayatilakeetal.BMCNephrology2013,14:180 Page10of13
http://www.biomedcentral.com/1471-2369/14/180
≥0.23 μg/g). There was a dose–effect relationship be- this to be the case. On the contrary, the cadmium con-
tween the concentration of cadmium in urine and the tent in all water samples analysed was within normal
stage of CKDu. A significantly higher cadmium concen- limits, except in one sample from a reservoir that had a
tration was also seen in the nails of those with CKDu borderline cadmium level(3.45μg/l).
compared to controls from the endemic area. Cadmium Drinking water is a major pathway for entry of inor-
is a known nephrotoxin and urinary excretion of ganic arsenic into the human body. The arsenic content
cadmium is considered to be a reliable indicator of cu- in 99% of water samples was below the WHO reference
mulativelong-termexposuretocadmium[6].Themean value of 10 μg/l [21]. However, it has recently been sug-
urine concentration of cadmium in CKDu cases was gested that the concentration of arsenic in drinking
abovethelevelsdemonstratedinrecent studiestocause water shouldbenomorethan 5μg/l[43].
oxidative stress and decreased glomerular filtration rate CKDu occurs in areas where groundwater is the main
and creatinine clearance [28-33]. The results of this source of drinking water. Groundwater in this region is
study indicate that cadmium exposure is a risk factor known to have a high content of fluoride and calcium.
forCKDu. People living in the region for generations have used
The mean urine concentration of arsenic in CKDu groundwater for drinking without ill effects. However,
cases was also above levels known to cause oxidative hardness of water, the high fluoride content, poor access
injury to the kidney [33]. In CKDu cases and controls to drinking water and inadequate intake of water in a
from the endemic area, concentrations of arsenic in warm climate may influence the body burden and/or the
urine and in fingernails were higher than those reported excretion of heavy metals and oxidative damage to the
in people living in low-exposure environments [34,35]. kidneys caused byheavymetals.
Urine is a major pathway for excretion of arsenic from The maximum level of cadmium for vegetables permit-
the human body, so urine levels reflect exposure. In ted by the Codex Alimentarius is 0.2 mg/kg [22,23] and
some studies, markers of oxidative stress have been the level permitted by the Commission of the European
demonstrated at urine arsenic concentrations as low as Communities is 0.05 mg/kg [24]. The maximum levels in
3.95 μg/g [36]. The level of total arsenic in urine is asso- certain vegetables grown in the endemic area exceeded
ciated with chronic kidney disease in a dose–response these safety levels. The maximum concentration of cad-
relationship, especially when the level is greater than miuminfish(0.06μg/g)alsoexceededtheEuropeanmax-
20.74 μg/g [36]. These findings support the contention imum limit of 0.05 mg/kg stipulated for certain types of
that chronic exposure to low levels of cadmium may be fish[24].Themaximumlevelofleadinvegetablespermit-
a causative factor for CKDu in Sri Lanka. Co-exposure ted by the Commission of the European Communities is
to cadmium and arsenic is known to produce additive 0.10mg/kg[24].Themaximumlevelofleadinvegetables
effects on the kidney that are morepronounced than ex- in the endemic area (0.476 mg/kg) exceeded this cut-off
posuretoeither metalalone [37,38]. value. Levels of cadmium and lead in vegetables and cad-
Selenium has been shown to protect the kidney from mium in freshwater fish from the endemic areaare above
oxidative stress [39]. A selenium concentration of 80– themaximumlevelsstipulatedbycertainFoodSafetyAu-
95 μg/l is needed to maximise the activity of the antioxi- thorities[22-24,44].
dant enzyme glutathione peroxidase and selenoproteins A provisional tolerable weekly intake (PTWI) for
in plasma [40,41]. In this context, it is significant that cadmium of 7 μg/kg body weight was established by the
serum selenium was below 80 μg/l in 38% and below Joint Food and Agriculture Organization of the United
90 μg/l in 63% of individuals with CKDu. Low selenium Nations (FAO)/WHO Expert Committee on Food Addi-
levels may have been a contributory factor increasing tives(JECFA)[45].In2011,theJECFArevisedthePTWI
the vulnerability of the kidneys to oxidative damage for cadmium to 5.8 μg/kg body weight [46]. More re-
caused byheavymetals andmetalloids. cently, the PTWI for cadmium has been lowered to
The association of raised serum strontium levels with 2.52 μg cadmium/kg body weight, in order to ensure a
raised serum cadmium levels has been reported previ- high level of protection of all consumers, including ex-
ously [42]. Strontium levels were not analysed in food or posed and vulnerable subgroups of the population [44].
water. The most likely explanation is an alteration of Since the cadmium content of certain food items in the
strontium handling and excretion, owing to the effect of endemic area is above stipulated levels, the total weekly
cadmiumonrenal tubular function. intake of cadmium in people living in the endemic area
Cadmium levels have previously been reported to be could exceed these safe limits, with detrimental effects
high in water sources in the domestic environment of on the kidneys, particularly in vulnerable people and
people with CKDu, and 10–20 times the maximum stip- those with predisposing factors.
ulated level have been found in reservoirs in the en- Reported mean dietary exposure to inorganic arsenic
demic area [15]. The results of this study did not show in the United States of America (USA) and various