arsenic and nonmelanoma skin cancer in slovakia beate pesch environmental health research institute,...

Arsenic and Arsenic and Nonmelanoma Skin Nonmelanoma Skin Cancer in SlovakiaCancer in Slovakia

Beate PeschBeate Pesch

Environmental Health Environmental Health Research Institute, Research Institute,

GermanyGermany

Part of the EU-funded Part of the EU-funded Project Project

EXPASCANEXPASCAN ‚Exposure to Arsenic ‚Exposure to Arsenic

and Cancer in Central & and Cancer in Central & Eastern Europe‘Eastern Europe‘

www.icconsultants.co.uk/www.icconsultants.co.uk/EXPASCAN.htmlEXPASCAN.html

PARTNERS

Imperial College & IC Consultants, London, UK

State Health Institutes, Prievidza, Bankska Bystrica, SK

Institute of Hygiene and Epidemiology, Prague, CZ

Environmental Health Research Institute, Duesseldorf, D

University of Cluj, RO

ObjectiveObjective

Estimation of the risk of Estimation of the risk of

environmental arsenic environmental arsenic

exposure from power plant exposure from power plant

emissions for non-melanoma emissions for non-melanoma

skin cancer (NMSC)skin cancer (NMSC)

Estimation of the risk Estimation of the risk of environmental of environmental arsenic exposurearsenic exposure

Choose study design(s)Choose study design(s) Assess exposureAssess exposure Estimate riskEstimate risk Discuss confoundersDiscuss confounders

Arsenic and arsenic Arsenic and arsenic compoundscompounds

Environmental Health Environmental Health Criteria (EHC) Criteria (EHC) 2nd edition, 224; 20012nd edition, 224; 2001

WHO, GenevaWHO, Geneva

www.inchem.orgwww.inchem.org

Estimate by Distance to Estimate by Distance to the Power Plantthe Power Plant

Environmental As exposureEnvironmental As exposure NMSC incidenceNMSC incidence

Associate As exposure Associate As exposure with NMSC riskwith NMSC risk

& control for covariates& control for covariates

ENO Power Plant (Slovakia)ENO Power Plant (Slovakia)

Environmental Arsenic Environmental Arsenic ExposureExposure

Historical As exposure Historical As exposure Air pollution modelling Air pollution modelling

(Colvile et al. 2001)(Colvile et al. 2001)

Current As exposureCurrent As exposure Measurement of As in soil, house dustMeasurement of As in soil, house dust

(Keegan et al. 2002)(Keegan et al. 2002)

Arsenic Emissions (tons/year) of the ENO Power Plant, Slovakia

0

100

200

1953 1960 1970 1980 1990 1999

Year

As

t/a

-10000 -6000 -2000 2000 6000 10000

Metres west-east from power plant

-10000

-8000

-6000

-4000

-2000

0

2000

4000

6000

8000

10000

Met

res

sou

th-n

ort

h f

rom

po

wer

pla

nt

1 .0E-0051.8E-0053.1E-0055.6E-0051.0E-0041.8E-0043.1E-0045.6E-0041.0E-0031.8E-0033.1E-0035.6E-0031.0E-0021.8E-0023.1E-002

Arsenic immission profile around theElectrarne Novaky power plant, Slovakia

Metres from power plant

Met

res

from

pow

er p

lant

-10000 0 10000

-10000

10000

0

Arsenic (mg/g) in soil 1999

by distance from the plant

Distance N Median Min Max

< 5 km 40 41 14 134

5-10 km 102 23 9 139

>10km 68 20 10 53

90%10%

75%25%

Median

MaximumMinimum

Distance to the power plant

Urin

ary

arse

nic

(As

III+A

sV

+MM

A+DM

A) [µ

g/l]

1

10

< 5 km 5-10 km > 10 km

Arsenic in soil [µg/g]

Uri

nar

y A

rsen

ic (

AsII

I +A

sV+

MM

A+

DM

A)

[µg

/l]

1

10

10 100

r = 0.21 (p<0.01)

y = 3.025 * x 0.237

n = 159

Cancer Incidence Cancer Incidence AnalysisAnalysis

Prievidza district versus Prievidza district versus SlovakiaSlovakia

Within Prievidza district Within Prievidza district

by distance to the plant by distance to the plant

Comparative Incidence Comparative Incidence Figures (CIF) Figures (CIF)

Prievidza district Prievidza district versusversus

Slovakia 1975-84 Slovakia 1975-84

All malignancies 1.1 All malignancies 1.1

NMSC 1.6NMSC 1.6

Lung cancer 1.0 Lung cancer 1.0

Bladder cancer 0.9 Bladder cancer 0.9

CIF by Distance to the CIF by Distance to the Power PlantPower Plant

cutoff 7.5 km 1977-cutoff 7.5 km 1977-19911991

Basal cell carcinoma 1.6Basal cell carcinoma 1.6Squamous cell ca. 1.6Squamous cell ca. 1.6

Lung cancer 1.0Lung cancer 1.0

Bladder cancer 1.1Bladder cancer 1.1

SIR SIR NMSC (1996-1999)NMSC (1996-1999)by Distance to the by Distance to the

PlantPlant

<5km<5km 5-10 km >10km5-10 km >10km

ReferenceReference

District District 1.21.2 1.11.1 0.80.8

0.9- 1.6 1.0-1.3 0.6-0.90.9- 1.6 1.0-1.3 0.6-0.9

Slovakia Slovakia 1.61.6 1.5 1.0 1.5 1.0

1.2- 2.2 1.3-1.7 0.9-1.31.2- 2.2 1.3-1.7 0.9-1.3

Population-based Population-based Case-Control StudyCase-Control Study

264 NMSC cases (1996-99)264 NMSC cases (1996-99)response rate response rate 80% 80%

286 population controls 286 population controls

response rate 72%response rate 72%

Matching by sex, ageMatching by sex, age

Statistical PowerStatistical Power

= 5% one-sided= 5% one-sided = 20% (power 80%)= 20% (power 80%) controls exposed to As=10%controls exposed to As=10% N cases = 264N cases = 264 N controls = 286N controls = 286 RR to be detected >= 1.9RR to be detected >= 1.9

NMSC Risk EstimationNMSC Risk Estimation

Logistic regressionLogistic regression

conditional on age, gender:conditional on age, gender:Odds Ratio (OR), 95% CIOdds Ratio (OR), 95% CI

Potential confounders: Potential confounders: occupational As exposureoccupational As exposure

smokingsmoking

Occupational As Occupational As exposure exposure (Job-Exposure (Job-Exposure

Matrix)Matrix)

AsExp.

Cas Con OR95% CI

No 159 180 1

Low 83 73 1.30.9-1.9

High 22 33 0.80.4-1.4

Cigarette SmokingCigarette Smoking

Pack-years

Cases Controls OR95% CI

< 1 163 165 1

1- < 10 31 21 1.40.7-2.6

10 -< 20 31 41 0.70.4-1.2

20 + 39 59 0.60.4-1.0

Skin Type & UV Skin Type & UV ExposureExposure

Type Cas264

Con286

OR95% CI

Frequentsunburns 22 14

1.80.9-3.5

Lighteyes 148 137

1.41.0-2.0

Blonde/red hair 29 13

2.61.3-5.2

Manyfreckles 16 8

2.20.9-5.3

Fresh Vegetables & Fresh Vegetables & FruitsFruits

Regularconsump.1/week

Cas264

Con286

OR95% CI

Vegetab. 237 265 0.70.4-1.2

Fruits 239 270 0.60.3-1.1

Exposure Assessment Exposure Assessment and Risk Estimationand Risk Estimationfor Environmental for Environmental

ArsenicArsenic

Dietary habitsDietary habits

Residential historyResidential history

Arsenic ExposureArsenic Exposurefrom Dietary Habitsfrom Dietary Habits

AsNut1 AsNut1 == Σ w(f )* I(f) Σ w(f )* I(f) 25 food items f:25 food items f: w(f) food frequencies w(f) food frequencies I(f) annual As intake I(f) annual As intake

AsNut2 AsNut2 = AsNut1 * s = AsNut1 * s if self-support s= 2, else 1 if self-support s= 2, else 1

Arsenic Exposure Arsenic Exposure with Dietary Habitswith Dietary HabitsAsNut2 Cas Con OR

95% CILow 74 91 1Medium 157 173 1.1

0.8-1.6High 33 22 1.8

1.0-3.4Trend 1.3

1.0-1.7

As ExposureAs Exposurefrom Residential Datafrom Residential Data

AsRes1 = Σ E(t)* AsRes1 = Σ E(t)* w(d(t),t) w(d(t),t)

for all places of residence:for all places of residence:E(t) annual emissionE(t) annual emissionw(d(t),t) immission weightw(d(t),t) immission weight

Correction of spatial Correction of spatial selection bias for selection bias for distance-related distance-related variables AsResvariables AsRes

(1) (1) Random re-sampling of Random re-sampling of controls controls SAS SurveyselectSAS Surveyselect

(2) (2) Bootstrap methodBootstrap method OR, 95% CI OR, 95% CI for R=800 re-sampled groups for R=800 re-sampled groups

Arsenic Exposure Arsenic Exposure with Residential Datawith Residential Data

AsRes1 OR95% CI

Low 1Medium 1.7

1.4-2.1High 1.9

1.4-2.6Trend 1.5

1.3-1.7

Environmental Arsenic Environmental Arsenic Exposure & NMSC RiskExposure & NMSC Risk Elevated NMSC incidence in the Elevated NMSC incidence in the

vicinity of the plant.vicinity of the plant. As exposure from dietary and As exposure from dietary and

residential data are associated residential data are associated with excess risk.with excess risk.

Residual confounding can not Residual confounding can not be excluded.be excluded.