Chronic exposure to inorganic arsenic (iAs) in the human population is associated with various internal cancers and other adverse outcomes. The purpose of this study was to estimate a population-scale exposure risk attributable to iAs consumptions by linking a stochastic physiological based pharmacokinetic (PBPK) model and biomonitoring data of iAs in urine. The urinary As concentrations were obtained from a total of 1,043 subjects living in an industrial area of Taiwan. The results showed that study subjects had an iAs exposure risk of 27% (daily iAs intake for 27% study subjects exceeded the WHO recommended value, 2.1 μg iAs day-1 kg-1 body wt). Moreover, drinking water and cooked rice contributed to the iAs exposure risk of 10% and 41%, respectively. The predicted risks in the current study were 4.82%, 27.21%, 34.69%, and 64.17%, respectively, among the mid-range of Mexico, Taiwan (this study), Korea and Bangladesh reported in literature. In conclusion, we developed a population-scale based risk model that covered the broad range of iAS exposure by integrating stochastic PBPK modeling and reverse dosimetry that generates probabilistic distribution of As intake corresponding to urinary As measured from the cohort study. The model can also be updated as new urinary As information becomes available.

2. 2
IARC: Group 1 (Carcinogenic to humans)
USEPA: Group A
Source: Distributed throughout the earth's crust
Standards for arsenic in drinking water: 10 μg L-1
Arsenic
IARC: International Agency for Research on Cancer; USEPA: United States Environmental Protection Agency 2
As
As3+
As5+ MMA3+ MMA5+ DMA3+ DMA5+
Organic AsInorganic As

3. 3
 Many arsenic sources are exited in our living
environment and food.
 Drinking water from the groundwater, flour and
rice grown or cooked in arsenic contaminated
soil and water has contain large inorganic arsenic.
 Seafood is a source of organic arsenic
compounds (arsenobetaine, arenosugars,
arsenolipids)
(Del Razo et al., 2002; Francesconi and Kuehnelt, 2004)
3

4. https://www.hrw.org/news/2016/04/06/bangladesh-20-million-drink-arsenic-laced-water
4

5. As5+
As3+
reduction
As5+
oxidation
As3+
MMA5+
MMA3+
reduction
MMA5+
oxidation
MMA3+
reduction
DMA5+
oxidation
DMA3+
Urinary
Arsenic
Metabolites
As5+: Arsenate
As3+: Arsenite
MMA5+: Monomethylarsonic acid
MMA3+: Monomethylarsonous acid
DMA5+: Dimethylarsinic acid
DMA3+: Dimethylarsinous acid
SAM: S-adenosyl-methionine
SAH: S-adenosyl-homocysteine
(Kitchin, 2001; Gong et al., 2001; Aposhian and Aposhian, 2006) 5
DMA5+
DMA3+
Methyltransferase
Methyltransferase

6. 6
• In populations with low seafood intake, total urine arsenic
and the sum of inorganic arsenic and methylated (MMA
and DMA) urine arsenic species are established
biomarkers that inorganic arsenic exposure for linking
the biomonitoring data to health outcomes
Biomakers for inorganic arsenic exposure:
the sum of iAs, MMA and DMA
(Calderon et al., 1999; National Research Council, 1999; Hughes, 2006) iAs: inorganic arsenic (As3+ and As5+)
6

7. 7
Animal Dosimetry: Compare blood/urine concentration in population with
blood/urine concentration at NOAEL in animal study to obtain MOE (Margin of Exposure )
Methods: Measurement of blood concentrations in toxicity studies or availability of PK
model/data in animal to predict blood concentrations from external dose.
Results: To determine adequacy of MOE
dose
effect
Slope=CSF
Exposure risk
NOAEL: No observable adverse effect level

8. 8
Forward dosimetry: To calculate internal does from external exposure
Methods: Human PBPK model (Ramsey and Andersen, 1984)
Results: Compare biomonitoring data with predicted biomarker at toxicity value (RfD, etc.)
Lung
Skin
Kidney
Liver
GI tract
External
exposure
Target
tissue does
Pollution (Arsenic,
dioxin, etc,)
Human body
Time
RfD: reference dose PBPK: Physiological based on pharmacokinetic

9. • Various physiological and biological parameters (Weight, height,
metabolize and exposure).
• How to characterize a population exposure risk
9
9

10. 10
Reverse Dosimetry: Estimate external exposure in population from biomonitoring data and
compare with toxicity value (RfD, MCL, etc.)
Methods: Human PBPK model can be applied to large and more poorly characterized human
populations that have highly variable exposures, activities, physiology, and pharmacokinetics (Bois,
2001)
Results: Reconstructing a population exposures distribution corresponding to human
biomonitoring data
Population exposureBiomonitoring data

11. 11
PBPK MODEL for
chloroform
In the Tan’s study, the PBPK model can be used in a reverse
dosimetry approach to assess a distribution of exposures related to
specific blood levels of trihalomethanes (THMs).
They used the Monte Carlo sampling techniques to consider the
probabilistic information about pharmacokinetics and exposure
patterns.
Probabilistic information: physiological
parameters and pharmacokinetics
parameters

12. Risk Assessment
PBPK model
for arsenic
Human
pharmacokinetic
parameters
Biomonitoring
data
Safe As guidelines
Reverse
dosimetry
Monte Carlo
simulation
12

13. 13
To develop a population scale PBPK model for
arsenic risk assessment
PBPK: Physiologically-based pharmacokinetic modelling
To predict the arsenic exposure risk that are
associated with specific biomarker levels in urine..
To provide a comprehensive assessment of safe
ingested arsenic level.
13

14. 14

15. 15
Subjects: An population living in industrial area of
Taiwan.
Study area: Changhua, central of Taiwan

16. 16
Parameter Symbol Unit Valuea
Notes and references
Body Height BH cm 163.31 (17.69) This study
Body Weight BW kg 63.50 (14.46) This study
Cardiac output QT L h-1
BW-0.75
16.50 (1.50) Clewell et al. (2000)
Organs volume
Bloodb
VB L 4.69 (0.96) (13.1×BH+18.05×BW
-480)×0.001/0.5723
G.I.tractc
VG L 1.20 (0.89) VG=BW×WG/DG
Liverc
VL L 1.81 (1.09) VL=BW×WL/DL
Kidneyc
VK L 0.28 (0.15) VK=BW×WK/DK
Other organs VO L 52.21 (19) VO=BW-(VB+VG+VL
+VK)
Tissue blood flow
To G.I tract QG L h-1
48.26 (24.23) QG=FG×QT×BW0.75
To liver QL L h-1
20.91 (10.61) QL=FL×QT×BW0.75
To kidney QK L h-1
61.13 (30.92) QK=FK×QT×BW0.75
To other organs QO L h-1
191.43 (96.49) QO=FO×QT×BW0.75
Tissue volume as percentage of body weight
G.I.tract WG % 1.98 (0.59) Yu and Kim (2004).
Liver WL % 2.99 (0.89) Yu and Kim (2004).
Kidney WK % 0.52 (0.16) Yu and Kim (2004).
Other organs WO % 94.51 (28.35) 100-other tissues
Blood flow to tissue as percentage of cardiac output
G.I.tract FG % 15 (4.50) Yu and Kim (2004).
Liver FL % 6.5 (1.95) Yu and Kim (2004).
Kidney FK % 19 (5.70) Yu and Kim (2004).
Other organs FO % 59.5 (17.85) 100-other tissues
Density
G.I.tract DG kg L-1
1.04 (0.31) Yu and Kim (2004).
Population-based PBPK
𝑑𝐴 𝑡
𝑑𝑡
= 𝑄 𝐿 × 𝐶𝐴 −
𝐶𝐿
𝑃𝐿
− 𝑉𝑚𝑎𝑥 ×
𝐶𝐿
𝑃𝐿(𝐾 𝑀 + 𝐶𝐿/𝑃𝐿)

17. Ca
QK
VK
CK
CK
Ca(K)
( )
K
K
a K
C
P
C

Blood Tissue K
As
As
As
As
As
As
As
As
As
As
As
As
As
3 3
3
3
( )K K
K a
K
dA C
Q C
dt P
 


 
QK
Ca
As3+
Tissue/Blood partition
coefficients
(mol) (L/hr) (mol/L)
As3+
As3+
As5+
MMA
DMA
Partition coefficients
17

18. 18
Parameters Symbol Unit Valuea
Metabolic constants for reduction and oxidationb
Reduction (As3+
As5+
) k1 h-1
1.37 (0.41)c
Oxidation (As5+
As3+
) k2 h-1
1.83 (0.55)c
Methylation constant of liverd
Maximum rate ( As3+
MMA)
3+
As MA
max ,L

V μmol h-1
0.03 (0.01)c
Maximum rate ( As3+
DMA)
3+
As DA
max,L

V μmol h-1
0.06 (0.02)c
Maximum rate ( MMADMA)
MA DA
max,L

V μmol h-1
0.04 (0.01)c
Michaelis constant ( As3+
MMA)
3+As MA
m,L

k μmol L-1
0.1 (0.03)c
Michaelis constant ( As3+
DMA)
3+As DA
m,L

k μmol L-1
0.1 (0.03)c
Methylation constant of kidneyd
Maximum rate ( As3+
MMA)
3+
As MA
max ,K

V μmol h-1
0.02 (0.006)c
Maximum rate ( As3+
DMA)
3+
As DA
max,K

V μmol h-1
0.28 (0.08)c
Maximum rate ( MMADMA)
MA DA
max,K

V μmol h-1
0.01 (0.004)c
Michaelis constant ( As3+
MMA)
3+As MA
m,K

k μmol L-1
0.1 (0.03)c
Michaelis constant ( As3+
DMA)
3+As DA
m,K

k μmol L-1
0.1 (0.03)c
Elimination constantsd
As3+
for urine 3+
As
urineK h-1
0.05 (0.01)e
As5+
for fecal 5+
As
fecalK h-1
0.001(0.0004)e
As5+
for urine 5+
As
urineK h-1
0.08 (0.02)e
As5+
for biliary
5+
As
biliaryK h-1
0.02 (0.005) e
MMA for urine MA
urineK h-1
4.20 (1.26) e
DMA for urine DA
urineK h-1
1.80 (0.54) e
Species-specific tissue/blood partition coefficientd
Tissues As3+
As5+
MMA DMA
GI tract (PGI) 2.80 (0.56)e
2.80 (0.56) 1.20 (0.24) 1.40 (0.28)
Liver (PL) 5.30 (1.06) 5.30 (1.06) 2.35 (0.47) 2.65 (0.53)
Kidney (PK) 4.15 (0.83) 4.15 (0.83) 1.80 (0.36) 2.08 (0.42)

19. Population-based PBPK
PBPK
Model
Physiological parameters
Exposure patterns
Partition coefficient
Constant Individual Exposure
Monte Carlo
Simulation
Physiological
parameters
Arsenic
biotransformation
Partition
coefficient
Bloodlevel
Days
19
19

20. Physiological
parameters
Probability
Physiological
parameters
Metabolic
parameters
Exposure
patterns
Partition
coefficients
Concentrations
Times
Population
Range
(95CI)
10,000
iterationsPopulation
based PBPK
20
Population-based PBPK

21. 21
Probabilistic Reverse Dosimetry Approach
(Tan et al., 2006, 2007)
μg g-1 of As in food or
μg L-1 As in water)
PBPK
modeling
Input Monte Carlo
analysis
50%
97.5%
2.5%
Exposure conversion factor
distribution (ECF)
Estimated
distribution of
arsenic in urine
ECF (μg l-1 TAs ug iAs-1)
Probability
×
Biomonitoring data
(N=1,075)
UAs (μg l-1)
Probability
＝Estimated population
exposure distribution
iAs (μg day-1)
Probability
UAs: Urinary arsenic; iAs: inorganic arsenic; InAs: Arsenic intake; ECF: Exposure converted factor
Invert
distribution
Distribution of measured urine
concentrations
(μg l-1 TAs per μg
iAs)
(μg iAs per μg l
TAs )

22. 22
Risk Characterization
Biomonitoring
data
Arsenic intake
Modeling
Tolerable Daily Intake (WHO,
1999)
Population Risk
2.1 μg inorganic As/ day/kg body weight
Probability
22

23. 23

24. 24
Characteristics N Mean Median Range
Age (years) 1,075 50.73 51.00 35-70
Weight (kg) 1,075 64.32 64.41 46.55-82.05
Arsenic concentrations in rice and watera
Cooked Rice (μg g wet wt.-1) 20 0.020 0.019 0.015-0.03
Water (μg L-1) 20 4.88 4.89 4.78-5.20
Daily rice and water intakesb
Cooked Rice (g wet wt. d-1) 776 801.97 486-1045
Water (L d-1) 3.10 3.28 0.91-6.00
Urinary arsenic (μg L-1) 109.36 84.71 3.88-1139.46
aMeasured the total arsenic concentration from cooked rice and drinking water
brice and water intake is calculated from the questionnaire
23

25. Selected percentile (95% confidence interval)
5th 10th 25th 50th 75th 90th 95th
Measured arsenic concentrations [NHANES data]a
Total arsenic - 2.10 4.10 7.70 16.00 37.40 65.40
DMA - - 2.00 3.90 6.00 11.00 16.00
Predicted arsenic concentrations [PBPK model]b
As3+
0.08 0.09 0.13 0.50 0.75 1.05 1.12
As5+
0.07 0.06 0.15 0.18 1.13 1.72 1.83
MMA 0.30 0.49 0.18 0.45 3.42 3.06 4.75
DMA 1.66 2.12 3.02 4.43 11.42 13.40 17.23
Total arsenic 2.11 2.76 3.48 5.56 16.72 19.23 24.93
25
Measured and predicted arsenic
concentrations in urine (μg L-1)
National Health and Nutrition Examination Survey (NHANES)

26. 26
0 24 48 72
0.00
0.02
0.04
0.0
0.1
0.2
0 24 48 72
0.00
0.02
0.04
0.0
0.1
0.2
0 24 48 72
0.00
0.02
0.04
0.0
0.1
0.2
0 24 48 72
0.00
0.02
0.04
0.0
0.1
0.2
0 0.03 0.06 0 0.04 0.08
0 0.1 0.2 0 0.4 0.8
As3+ As5+
DMAMMA
Urinearsenicconcentrations(μgL-1
)
Time (hour)
Probability
ProbabilityProbability
Probability Urinary arsenic conc. in unit
arsenic intake (μg L-1)
A
B
C
D
LN (0.03 μg L-1, 0.02 )
LN (0.13 μg L-1, 0.11 ) LN (0.6 μg L-1, 0.2 )
LN (0.04 μg L-1, 0.03 )
25

27. 27
0 1 2 3 4 5
0.00
0.03
0.06
0.09
0.12
0.15
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
0.0
0.1
0.2
0.3
0.4
0.5 0.4 0.3 0.2 0.1 0.0
0
100
200
300
400
Inorganic arsenic intake (μg kg-1
d-1
)
Fit curve
InAs intake
Area of risk
Probability
TDI:2.1
Risk=0.27
ECF
Urinary TAs
B
A
Probability
ECF (μg L-1 ug InAs-1)
Probability
Urinarytotal
arsenic(μgl-1)

28. 28
0 1 2 3 4 5 6 7 8
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0 1 2 3 4
0.00
0.01
0.02
0.03
0.04
0.05
Daily InAs intake (μg kg-1 d-1)
Cumulativeprobability
Daily InAs intake (μg kg-1 d-1)
Risk from drinking water
Risk from rice consumption
Risk from others
Probability
Others
(49%)
Rice
(41%)
Water (10%)
TDI
Risk=0.27
27

29. 29
0 5 10 15 20
0.0
0.2
0.4
0.6
0.8
1.0
0
20
40
60
80
100
EPofTDI(%)
Bangladesh
64.17%
Korean
34.69%This study
27.21%
Mexico
4.82%
Standard
0.04%
Korea 127.4 μg L-1
TDI
2.1
Mexico, 65.4 μg L-1
This study, 106 μg L-1
Standard, 50 μg L-1
Bangladesh, 263.7 μg L-1
Cumulativeprobability
Daily inorganic arsenic intake (μg kg-1 d-1)

30. 30
MTDI: Maximum Tolerable Daily Intake

31. 31

32. 32
Thanks for
your attention