Arsenic

Endpoint summary

Link to relevant study record(s)

Reference

Reference 1

Endpoint:

basic toxicokinetics, other

Remarks:

bioaccessibility / solubility in artificial physiological media

Type of information:

experimental study

Adequacy of study:

key study

Study period:

2011-11-16 to 2012-02-13

Reliability:

1 (reliable without restriction)

Rationale for reliability incl. deficiencies:

guideline study

Objective of study:

bioaccessibility (or bioavailability)

Qualifier:

equivalent or similar to guideline

Guideline:

OECD Series on Testing and Assessment No. 29 (23-Jul-2001): Guidance document on transformation/dissolution of metals and metal compounds in aqueous media

Deviations:

yes

Remarks:

In lieu of a harmonised guideline for in-vitro bioaccessibility test, the test was conducted in adaptation of guidance for OECD-Series on testing and assessment Number 29 and according test protocols discussed with the sponsor.

Principles of method if other than guideline:

In lieu of a harmonised guideline for in vitro bioaccessibility test, the test was conducted in adaptation of guidance for OECD-Series on testing and assessment Number 29 and according test protocols discussed with the sponsor. The principle of the test is the introduction of the test material into one or more artificial physiological media at a defined loading, followed by sampling of the solution after certain times, and analysis for dissolved material. The test medium was artificial gastric juice.

GLP compliance:

yes (incl. QA statement)

Remarks:

signed 2011-02-07

Species:

other: in vitro (simulated human body fluids)

Details on test animals or test system and environmental conditions:

Test principle in brief:
- one artificial physiological medium (GST),
- single loading of test substance of ~100 mg/L,
- the study was performed in triplicate (3 flasks),with 3 additional control blanks,
- two samples were taken after 2 and 24 hours agitation (100 rpm) at 37 ± 2 °C,
- measurement (by ICP-OES) of dissolved arsenic concentrations after filtration. Arsenic (III) and arsenic (V) were separated by HPLC and directly quantified by coupling to ICP-MS


The objective of this study was to assess the dissolution of Arsenic metal (grain size approx. 1 mm) in artificial gastric fluid (GST pH 1.5-1.6). The test medium was selected to simulate a substance entering the human by ingestion into the gastro-intestinal tract.

Route of administration:

other: in vitro experiment in artificial gastric juice, simulating oral exposure

Duration and frequency of treatment / exposure:

Single loading of ca. 100 mg/L, sampling and analysis for dissolved metal after 2 and 24 hours.

Details on study design:

PROCEDURE
For the experimental part, three independent flasks were prepared with a loading of 100 mg test item /L GST medium. Three additional control blanks (same procedure, ultrapure water) were also prepared. All flasks (two samples each flask) were tested after 2 and 24 h, to measure total dissolved arsenic concentrations (by ICP-OES), arsenic speciation, temperature and pH. During the study, observations, including the appearance of the solution (color, turbidity, particle film on the surface etc…) were recorded, in addition to measured concentrations, turbidity and pH.

Mass balance calculation:
Total dissolved arsenic concentrations were measured by ICP-OES after the addition of aqua regia to the solution to dissolve the remaining arsenic grain.

Reagents:
- Purified water (resistivity > 18 MΩ·cm, Pure Lab Ultra water purification system from ELGA LabWater, Celle, Germany)
- Nitric acid - “Supra” quality (ROTIPURAN® supplied by Roth, Karlsruhe, Germany).
- Hydrochloric acid – “Baker-instra-analyzed-plus” quality (J.T. Baker, Griesheim, Germany).
- Ammonium hydroxide - p.A. quality (Merck, Darmstadt, Germany)
- Ammoniumdihydrogenphosphate - suprapur (Merck, Darmstadt, Germany)

ARSENIC ANALYSIS
ICP-OES
- Standards (stock solutions and calibration): Arsenic standard containing 1000 mg/L As in nitric acid 2-3 % (lot. No. 09K023, CPI, Amsterdam, The Netherlands)
- Certified aqueous reference material (accuracy and reproducibility of the method): TMDA-70 (Environment Canada, lot no. 0310) and a multielement standard (19 elements, CPI, Amsterdam, The Netherlands, lot no. 009K023)
HPLC-ICP-MS
- Standards (stock solutions, calibration, verification of the method): Arsenic (III) standard (0.05 mol/L, lot no. HC948486, Merck, Darmstadt, Germany) and As (V) standard (1000 mg/L, lot no. HC961324, Merck, Darmstadt, Germany)

Instrumental and analytical set-up for the ICP-OES instrument:
Thermo IRIS Intrepid II from Thermo Electron Corporation, Germany
Nebulizer: Concentric glass nebulizer, from Thermo
Spray chamber: Glass cyclonic spray chamber, from Thermo
Nebulizer gas flow: 0.68 L/min
Make-up gas flow: 0.5 L/min
RF power: 1150 W
Wavelengths: 189.042 nm, 193.759 nm
Calibration: blank, 10 µg/L, 25 µg/L, 50 µg/L, 100 µg/L, 250 µg/L, 500 µg/L and 1000 µg/L

The separation of As (III) and As (V) in artificial gastric fluid samples was performed using the HPLC system Agilent 1200 coupled to an Agilent ICP-MS 7700 (Agilent Technologies, Waldbronn, Germany). The following analytical and instrumental setup was applied:

HPLC 1200:
column + precolumn: PRP-X100 column (Hamilton, Bonaduz, Switzerland)
flow: 1.5 mL/min
eluent HPLC 1200: 20 mmol/L (NH4)H2PO4, pH = 5.6 (adjusted by NH4OH)
column temp: 22 °C
max. pressure: 100 bar
injection volume: 100 µL

ICP-MS 7500:
power: 1500 W
carrier: 0.91 L/min
dilution gas: 0.15 L/min
nebulizer: concentric
spray chamber: Scott type double
isotopes: 75As, 37Cl (to exclude interferences)
retention As(III): ≈ 2 min
retention As(V): ≈ 8 min

- Calibrations (As(III) and As(V) standards): blank, 0.1 µg/L, 0.5 µg/L, 1 µg/L, 5 µg/L, 10 µg/L, 20 µg/L, 30 µg/L, 40 µg/L and 50 µg/L.

Details on dosing and sampling:

Loading of 100 mg/L: 49.2 mg, 50.4 mg and 49.4 mg were added into 500 mL GST medium.

Sampling:
Solutions were sampled for measuring total dissolved arsenic by ICP-OES. All samples were filtered through 0.2 µm filter prior to further treatment.
Two aqueous subsamples of approx. 20 mL taken for arsenic analysis were transferred into disposable scintillation vials, and stored at approx. 4 °C until analysis. Due to the acid GST medium samples were already acidified.
For speciation analysis of As (III) and As (V) by HPLC-ICP-MS, 20 mL were transferred into disposable scintillation vials, and arsenic species were promptly measured by ICP-MS after HPLC separation.

Type:

other: in vitro bioaccessibility: Dissolved fraction (% arsenic dissolved of arsenic added)

Results:

Gastric juice: 0.7 % after 2h; 3.8 % after 24h; effectively all dissolved As present in trivalent form As(III)

Bioaccessibility (or Bioavailability) testing results:

Final results:
Under the conditions of this test (flasks with artificial gastric fluid; loadings of 100 mg As metal per liter, grain size approx. 1 mm, 37 °C, sampling after 2 h and 24 h), the concentrations of dissolved arsenic, and respective arsenic (III) and arsenic (V) species were as follows:

GST 2 h:
- total As conc. ± SD*: 652 ± 85 µg/L
- As (III) conc. ± SD*: 658 ± 95 µg/L
- As (V) conc. ± SD*: 1 ± 0.1 µg/L
- ∑ As (III) + As (V) ± SD**: 659 ± 95 µg/L

GST 24 h
- total As conc. ± SD*: 3687 ± 873 µg/L
- As (III) conc. ± SD*: 3786 ± 1029 µg/L
- As (V) conc. ± SD*: - ∑ As (III) + As (V) ± SD**: 3786 ± 1029 µg/L
*the mean of vessel A, B and C was calculated
**Original calculations for the data were performed with more digits. Values were rounded

In artificial gastric fluid, approx. 0.7 % of the loaded mass of arsenic metal (grain size approx. 1 mm) were dissolved after 2 h (685 µg/L As (III) and 1 µg/L As (V)).
After 24 h, approx. 3.8 % of the loaded mass of arsenic metal had dissolved. Only the concentration of As (III) was above the limit of quantification in the 24 h samples.
Recovery of total dissolved arsenic when expressed as the sum of measured dissolved tri- and pentavalent arsenic species ranged from 101 - 103 %.

Method validation summary (ICP-OES)

validation parameter	results	Comment
selectivity	similar data with two different As wavelengths for ICP-OES method	no interferences due to test media observed
linearity	applied calibration function was linear	correlation coefficient at least0.999937
limit of detection	5.59 - 18.4 µg/L	-
limit of quantification	18.6 - 61.5 µg/L	-
method blanks	< LOD (< 18.4 µg/L)	-
accuracy	mean recovery for CRM TMDA-70: 101 ± 3 % (n = 5)	Lower concentration range (40.7 µg As/L)
trueness	mean recovery for recalibration standard: 98.1 ± 0.7 % (n = 3)	250 µg As/L
trueness	mean recovery for recalibration standard: 98.4± 0.7 % (n = 5)	500 µg As/L
reproducibility	mean recovery for CRM TMDA-70: 101 ± 2.7 % (n = 5)	lower concentration range (40.7 µg As/L)
reproducibility	mean recovery for diluted CPI standard: 100 ± 0.7 % (n = 5)	higher concentration range (250 µg As/L)

Validation of HPLC-ICP-MS:

In addition to the calibration,Recalibration standards of 10 µg/L As(III) and As(V) were analysed in each measurement series

- Recovery rates were 101 ± 4 % for As(III) and 97.3 ± 3.3 % for As(V)

- LOD: 0.120 µg As(III)/L, 0.146 µg As(V)/L

- LOQ: 0.441 µg As(III)/L, 0.534 µg As(V)/L

- correlation coefficient: 0.999842164 As(III), 0.999767349 As(V)

- The As concentrations in method blanks were below the LOD.

To assure reliable HPLC-ICP-MS measurements, a control method standard was included. The control method standard (10 µg/L As (III) and 10 µg/L As (V)) was analysed directly after each calibration, subsequently after three samples and at the end of the measurement series.

- Recovery rates were 97.6-102 % for As (III) and 88.9 -97.8 % for As(V).

Solution pH values

The target pH in GST medium before addition of the test substance is 1.5-1.6. During the study, the pH of all GST solutions increased (samples after 2h: pH 1.59-1.60; after 24h: pH 1.92-1.94 / blank after 2h: pH 1.55 -1.57; after 24h: pH 1.91-1.93)

Mass balance calculation

Total dissolved arsenic concentrations were measured by ICP-OES after the addition of aqua regia to the solution to dissolve the remaining arsenic grain. A complete dissolution of arsenic in vessel 1, 2 and 3 was obtained. The results of the recovery of the nominal loading of 50 mg / 500 mL are compiled in following table:

 

Calculation of mass balance

media	mean value for total As [mg/L]	nominal concentration [mg/L]	recovery [%]
GST Vessel 1	98.5	98.4	100
GST Vessel 2	102	101	101
GST Vessel 2	93.5	98.8	94.7

Conclusions:

Under the conditions of this test (flasks with artificial gastric fluid; loadings of 100 mg As metal per liter, grain size approx. 1 mm, 37 °C, sampling after 2 h and 24 h), the concentrations of dissolved arsenic, and respective arsenic (III) and arsenic (V) species were as follows:

GST 2 h:
- total As conc. ± SD*: 652 ± 85 µg/L
- As (III) conc. ± SD*: 658 ± 95 µg/L
- As (V) conc. ± SD*: 1 ± 0.1 µg/L
- ∑ As (III) + As (V) ± SD**: 659 ± 95 µg/L

GST 24 h
- total As conc. ± SD*: 3687 ± 873 µg/L
- As (III) conc. ± SD*: 3786 ± 1029 µg/L
- As (V) conc. ± SD*: - ∑ As (III) + As (V) ± SD**: 3786 ± 1029 µg/L
*the mean of vessel A, B and C was calculated
**Original calculations for the data were performed with more digits. Values were rounded

In artificial gastric fluid, approx. 0.7 % of the loaded mass of arsenic metal (grain size approx. 1 mm) were dissolved after 2 h (685 µg/L As (III) and 1 µg/L As (V)).
After 24 h, approx. 3.8 % of the loaded mass of arsenic metal had dissolved. Only the concentration of As (III) was above the limit of quantification in the 24 h samples.
Recovery of total dissolved arsenic when expressed as the sum of measured dissolved tri- and pentavalent arsenic species ranged from 101 - 103 %.

Reference 2

Endpoint:

basic toxicokinetics, other

Remarks:

bioaccessibility / solubility in artificial physiological media

Type of information:

experimental study

Adequacy of study:

key study

Study period:

2010

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

guideline study

Objective of study:

bioaccessibility (or bioavailability)

Qualifier:

equivalent or similar to guideline

Guideline:

OECD Series on Testing and Assessment No. 29 (23-Jul-2001): Guidance document on transformation/dissolution of metals and metal compounds in aqueous media

Deviations:

yes

Remarks:

In lieu of a harmonised guideline for in-vitro bioaccessibility test, the test was conducted in adaptation of guidance for OECD-Series on testing and assessment Number 29 and according test protocols discussed with the sponsor.

Principles of method if other than guideline:

In lieu of a harmonised guideline for in vitro bioaccessibility test, the test was conducted in adaptation of guidance for OECD-Series on testing and assessment Number 29 and according test protocols discussed with the sponsor. The principle of the test is the introduction of the test material into one or more artificial physiological media at a defined loading, followed by sampling of the solution after certain times, and analysis for dissolved material. The test media were artificial gastric juice, interstitial fluid, lysosomal fluid and artificial perspiration.

GLP compliance:

not specified

Specific details on test material used for the study:

Produced by PPM Pure Metals GmbH, Germany. Lot 4018 produced on 2010-05-10. Purity >99.99%. Sieved to < 0.2 mm. Particle size distribution characterised by D10 = 9 µm, D50 = 41 µm and D90=104 µm.

Species:

other: in vitro (simulated human body fluids)

Route of administration:

other: in vitro experiment in artificial gastric juice, lysosomal fluid, interstitial fluid and perspiration, simulating oral, inhalation and dermal exposure

Duration and frequency of treatment / exposure:

Single loading of 2 g/L (0.1 g per 50 mL). Sampling and analysis for dissolved metal after certain extraction times.

Type:

other: in vitro bioaccessibility: Dissolved fraction (% arsenic dissolved of arsenic added), mean of duplicate vessels (>100% values due to analytical variability):

Results:

Gastric fluid, pH=1.5: 2 h: 36.85 %, 24 h:105 %; Interstitial fluid, pH= 7.4: 72 h: 10.065 %, 168 h: 25.35 %; Lysosomal fluid, pH= 4.5-5.0: 72 h: 99.8 %, 168 h: 106 %; Artificial perspiration, pH=6.5:72 h: 25.55 %, 168 h: 42.2 %.

Bioaccessibility (or Bioavailability) testing results:

Medium, pH
Extraction time (h)
Dissolved fraction (% arsenic dissolved of arsenic added), mean of duplicate vessels (>100% values due to analytical variability):

Gastric fluid, pH=1.5
2 h 36.85 %
24 h 105 %

Interstitial fluid, pH= 7.4
72 h 10.065 %
168 h 25.35 %

Lysosomal fluid, pH= 4.5-5.0
72 h 99.8 %
168 h 106 %

Artificial perspiration, pH=6.5
72 h 25.55 %
168 h 42.2 %

Reference 3

Endpoint:

basic toxicokinetics, other

Type of information:

other: review

Adequacy of study:

weight of evidence

Reliability:

other: Reliable review. In view of the enormous database on arsenic toxicity, this and other reviews are used as the key sources of information for the hazard assessment, instead of repeating as assessment of primary data.

Rationale for reliability incl. deficiencies:

other: see 'Remark'

Remarks:

Comprehensive review and summary on several aspects of arsenic toxicity by a renown scientific organisation. In view of the enormous database on arsenic toxicity, this and other reviews are used as the key sources of information for the hazard assessment, instead of repeating an assessment of primary data.

Principles of method if other than guideline:

Not applicable (comprehensive, reliable review and summary on several aspects of arsenic toxicity by a renown scientific organisation).

Type:

other: see discussion / endpoint summary

Reference 4

Endpoint:

basic toxicokinetics, other

Type of information:

experimental study

Adequacy of study:

weight of evidence

Type:

other:

Results:

Based on the reported findings, an average of 60% of total arsenic intake is assumed to be excreted via urine. For details, please refer to endpoint summary/discussion.

Reference 5

Endpoint:

basic toxicokinetics, other

Type of information:

other: review

Adequacy of study:

weight of evidence

Reliability:

other: Reliable review. In view of the enormous database on arsenic toxicity, this and other reviews are used as the key sources of information for the hazard assessment, instead of repeating as assessment of primary data.

Rationale for reliability incl. deficiencies:

other: see 'Remark'

Remarks:

Comprehensive review and summary on several aspects of arsenic toxicity by a renown scientific organisation. In view of the enormous database on arsenic toxicity, this and other reviews are used as the key sources of information for the hazard assessment, instead of repeating an assessment of primary data.

Principles of method if other than guideline:

Not applicable (comprehensive, reliable review and summary on several aspects of arsenic toxicity by a renown scientific organisation).

Type:

other: see discussion / endpoint summary

Reference 6

Endpoint:

basic toxicokinetics, other

Type of information:

other: PBPK modeling

Adequacy of study:

weight of evidence

Type:

other:

Results:

PB-PK model, please see endpoint summary.

Reference 7

Endpoint:

basic toxicokinetics, other

Type of information:

other: PBPK modeling

Adequacy of study:

weight of evidence

Type:

other:

Results:

PB-PK model, please see endpoint summary.

Reference 8

Endpoint:

dermal absorption in vitro / ex vivo

Type of information:

experimental study

Adequacy of study:

key study

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

other: Reasonably well-documented publication, conduct acc. to general scientific principles.

Principles of method if other than guideline:

In vivo (Rhesus monkey) and In vitro (human skin) percutaneous absorption of arsenic acid (73As radiolabelled) from soil and water for (comparative reasons) was investigated. This endpoint record focuses on the in vitro data in human skin, because of higher relevance for human health risk assessment. Further, the focus is on the absorption experiments conducted with aqueous solutions (worst-case, as compared to absorption from soil).

GLP compliance:

not specified

Radiolabelling:

yes

Remarks:

73As

Vehicle:

water

Doses:

Water formulations were prepared for comparison to experiments on As absorption from soil. The water load on skin was 5 µL/cm2 skin area. This amount of water is a thin layer of water which covers the skin but does not run off the skin. It is similar to a thin layer of other dermatological doses (cream, ointment).
The low (trace) dose was 0.000024 µg/cm2. The high dose: gave an arsenic skin concentration of 2.1 µg/cm2.

Details on in vitro test system (if applicable):

Three separate donor skin sources with three replicates per each experiment were used. Small cells were of the flow-through design with a 1 cm2 surface area. Phosphate-buffered saline at a flow rate of 3.0 mL/hr (reservoir volume) served as receptor fluid. Human cadaver skin was dermatomed to 500 µm and stored refrigerated at 4°C in Eagle's minimum essential medium to preserve skin viability. The skin was used within 5 days.
73As in water formulation was applied with a micropipette to the surface of the skin. Standards for each formulation were made by dissolving 5 µL of formulation in 10 mL of scintillation cocktail. At the end of a 24 h period the system was stopped. The residual water remaining in the cells was collected and analyzed. The skin surface was washed once with liquid soap (50/50 water, v/v: Ivory Liquid, Procter & Gamble. Cincinnati, OH) and twice with 1 mL of distilled water, and the wash solutions were analyzed by scintillation counting. Cells were disassembled. Cell tops were rinsed three times with 1 mL of water. The inner surface of the skin was swabbed with cotton balls and counted. The skin itself was completely solubilized in Soluene 3540 (Packard lnstruments, Downders Grove, IL), and 1 M HCl was added to neutralize the homogenate. The receptor fluid samples from the permeation cells' residual fluid, the skin surface washes, the cotton balls, the glass apparatus, and the skin itself were assayed for 73As content by liquid scintillation counting.

Key result

Time point:

24 h

Dose:

up to 2.1 µg/cm2

Parameter:

percentage

Absorption:

ca. 2 %

Remarks on result:

other: 24 h

Remarks:

in vitro, human skin; result accounts for ca. 1% absorbed As (found in receptor fluid) plus ca. 1% retained in the skin (potentially absorbable)

With water formulation, 0.93 ± 1.1 % of the applied dose accumulated in the receptor fluid and 0.98 ± 0.96% was in skin after surface wash. The soap and water wash accounted for 69.8 ± 16.4%. If in vitro percutaneous absorption is calculated as receptor fluid accumulation plus residual skin concentration (after soap and water wash) then the absorption in human skin is 1.9% from water. It appears from the text of the publication that these detailed results are for the low dose (0.000024 µg/cm2) application only.

However, in the discussion section of the publication, it is mentioned that for the high dose (2.1 µg/cm2), the absorption/penetration was 0.04 µg/cm2, that is also around 2%.

Description of key information

Read across approach

Arsenic metal has a relatively low solubility. However, it is not completely inert when immersed in water or biological fluids and dissolved ionic arsenic species, predominantly As(III) (arsenite), can be released from the metal. The solubility in water and in artificial physiological media has been experimentally determined for arsenic metal. For details please refer to the sections on Water solubility and Toxicokinetics.

Under the conditions of an experimental study conducted in accordance with OECD TG 105 at a loading of ca. 0.5 g/L, the solubility of granular As metal (2-15 mm) was determined to be ca. 10.6 mg/L after stirring for 8 days under an argon atmosphere to exclude oxidative processes.

The bioaccessibility of granular As (ca. 1 mm grains) in artificial physiological media can be summarised as follows: the solubility in artificial gastric juice (hydrochloric acid at pH=1.5) at a loading of 100 mg As metal/L corresponds to dissolved fractions of arsenic of 0.7 and 3.8 % after 2 and 24 h, respectively. Effectively all dissolved As was present in trivalent form (Knopf, 2012).

The bioaccessibility of arsenic metal powder (D50 ca. 41µm) in artificial physiological media in terms of the dissolved fraction (% As dissolved of As added) can be summarised as follows: gastric fluid, pH=1.5: 2 h: 36.85%, 24 h: 105%; Interstitial fluid, pH= 7.4: 72 h: 10.065%, 168 h: 25.35%; Lysosomal fluid, pH= 4.5-5.0: 72 h: 99.8%, 168 h: 106%; artificial perspiration, pH=6.5: 72 h: 25.55%, 168 h: 42.2% (Kirby Memorial Health Center, 2010).

In conclusion, arsenic metal is not inert in water or biological fluids. The release of dissolved arsenic species depends on particle size (thus the surface in contact with fluids) and exposure time. However, even for arsenic metal grains of approximately 1 mm, already 4% of the arsenic dissolved in artificial gastric juice after 24 h. For a much finer arsenic metal powder (D50 ca. 41 µm), dissolution was essentially complete in acidic artificial gastric fluid (pH=1.5) after 24 h and in artificial lysosomal fluid (pH=4.5-5) after 72 h. In the more neutral artificial fluids (interstitial fluid and sweat), dissolution of As powder after 168 h reached ca. 25 and 42%, respectively.

In these in vitro dissolution experiments, the predominant form dissolved from arsenic metal (massive or powder) was the trivalent As(III).

Whereas a number of studies have noted differences in the relative toxicity of arsenic compounds, with trivalent arsenites tending to be somewhat more toxic than pentavalent arsenates, these distinctions are generally not emphasised in the assessments of systemic effects of arsenic compounds for several reasons: (1) in most cases, the differences in the relative potency are small (about 2–3-fold), often within the bounds of uncertainty regarding effect levels; (2) different forms of arsenic may be interconverted, both in the environment and the body; and (3) in many cases of human exposure, the precise chemical speciation is not known (ATSDR, 2007).

Based on this information, conservative read across from information on (soluble) inorganic arsenic compounds to arsenic metal is therefore applied for the assessment of systemic toxicological effects.

Absorption

Oral

Each of the forms of arsenic have different physicochemical properties and bioavailability. Studies in rats, mice and humans indicate that arsenite (As(III)) and arsenate (As(V)) present in drinking water are rapidly and nearly completely (about 95%) absorbed after ingestion (ATSDR, 2007). However, the absorption of ingested inorganic arsenic can vary depending on the solubility of the arsenical compounds (the more water soluble the compound, the greater its absorption), the presence of other food constituents and nutrients in the gastrointestinal tract, and the food matrix itself (EFSA, 2010).

Dermal

In an in vitro human skin permeation study with arsenic acid, a dermal absorption rate of 0.93% was obtained. In addition, 0.98% of the dose were retained in skin after washing (Wester et al., 1993). As a consequence, a conservative dermal absorption rate of maximum 2% from liquid media (0.2% from dry dust exposure according to the default values proposed by HERAG, 2007) may be considered for risk assessment via this route.

Inhalation

Deposition and subsequent absorption of arsenic powder in the lungs may be expected to be dependent on particle size: respirable particles (0.1 -1 µm) are carried further into the lungs where they are likely to be absorbed quickly, whereas larger particles will be translocated to the gut. However, based on the high oral bioavailability of soluble As(III) substances, complete systemic bioavailability after inhalation exposure can be assumed for conservative risk assessment purposes.

Distribution

In the bloodstream, arsenic is distributed between the plasma and the erythrocytes, in which it is bound to haemoglobin. Arsenite and arsenate ions are readily transported into cells: arsenite by aquaglycoporins 7 and 9 which normally transport water and glycerol, and arsenate by phosphate transporters. In most species, residue levels following uptake are initially elevated in liver, kidney, spleen and lung, but several weeks later arsenic is translocated to hair, nails and skin. Residual levels in animals tended to be higher for arsenite than arsenate. Arsenic is readily methylated in the body and methylarsonate is the predominant metabolite in kidneys, whereas dimethylarsinate is the predominant metabolite in lungs. It is worthy of note that rats differ from most mammalian species by accumulating arsenic in erythrocytes, likely by the binding of trivalent arsenic species to cysteine components of haemoglobin, with the binding affinity of trivalent arsenic species to red blood cells estimated to be 15-30-fold higher in rats than in humans (EFSA, 2010).

Placental transfer

Arsenic readily passes through the placenta of mammals and humans, resulting in similar exposure levels both in fetuses and mothers. Inorganic arsenic as well as its methylated metabolites (methylarsonate and dimethylarsinate) pass through the placenta. In newborn babies of women exposed to arsenic via drinking water in Argentina, essentially all arsenic in plasma and urine was in the form of dimethylarsinate, suggesting that it is mainly this metabolite that reaches the foetal circulation in late gestational phases. The metabolic methylation of arsenic via a one-carbon metabolism increases in women during pregnancy, which is why the human foetus is likely to be exposed to more inorganic arsenic and methylarsonate in early gestation (EFSA, 2010).

Transfer via mother’s milk

In contrast to the rapid transfer of arsenic to the foetus, very little arsenic is excreted in breast milk. Argentinian women exposed to about 200 μg/L arsenic in their drinking water showed very low excretion in breast milk (ca. 3 μg/L). A study in Bangladesh indicated very low arsenic concentrations in breast-milk samples (median 1 μg/kg; range 0.25-19 μg/kg) despite high arsenic exposures from drinking water (about 50 μg/L). It is assumed that the small amounts of arsenic passing to milk are almost entirely inorganic, with efficient maternal methylation of arsenic likely to protect against excretion via breast milk (EFSA, 2010).

Metabolism

In most mammalian species including humans, inorganic arsenicals are extensively biotransformed and excreted mainly as their metabolites. In oxygen-rich environments, inorganic arsenic is present primarily as arsenate (As(V)), which is the most common form in drinking water. Arsenite (As(III)) can also be present, particularly under anaerobic conditions (Tsuji et al., 2019). Arsenate enters the cell via the phosphate carrier system and can be biotransformed enzymatically to arsenite via glutathione reductase. In mammals, arsenite directly undergoes extensive oxidative methylation in the liver. There are considerable differences between species and between individuals of a same species in arsenic biotransformation (EFSA, 2010; Tsuji et al., 2019). Most studied animals are more efficient in methylating arsenic to dimethylarsinate (DMA) than humans, except primates which have been shown not to methylate arsenic at all. Rats, the most common standard testing species, differ significantly from most other mammals by accumulating arsenic in erythrocytes (ATSDR, 2007).

Excretion

Arsenic and its metabolites are readily excreted predominantly via urine but to some extent also via bile. Urinary excretion rates of 80% up to 61 hours following oral doses and 30-80% in 4-5 days following parenteral doses have been measured in humans. Although rats tend to excrete arsenic and metabolites preferentially into bile, the major route of excretion of arsenic compounds in most mammalian species and humans is via urine, and dimethylarsinate is the primary urinary metabolite. In contrast to most other mammals, humans excrete appreciable amounts of methylarsonate in urine. The composition of urinary arsenic metabolites varies from person to person and has been interpreted to reflect arsenic methylation efficiency, with a typical profile of urinary arsenic metabolites consisting of 10-30% inorganic arsenic, 10-20% methylarsonate and 60-70% dimethylarsinate (EFSA, 2010; ATSDR, 2007).

Buchet et al. (1981) studied the urinary elimination of arsenic metabolites in volunteers who ingested a single oral dose of arsenic (500 µg As) either as sodium arsenite (SA), monomethylarsonate (MMA) or cacodylate (DMA). The excretion rate increased in the order SA < DMA < MMA. After 4 days, the arsenic excreted in urine corresponded to 46, 78, and 75% of the ingested doses, respectively. Based on these findings, an average of 60% of total arsenic intake is assumed to be excreted via urine.

PBPK model for arsenic

A physiologically based pharmacokinetic (PBPK) model for exposure to inorganic arsenic in hamsters, rabbits and humans has been developed (Mann, 1996a and b). The model in its present state simulates three routes of exposure to inorganic arsenic: oral intake, intravenous injection and intratracheal instillation. It describes the tissue concentrations and the urinary and faecal excretions of the four arsenic metabolites: inorganic As(III) and As(V), methylarsonic acid and dimethylarsinic acid. The model consists of five tissue compartments, chosen according to arsenic affinities: liver, kidneys, lungs, skin and others. The model is based on physiological parameters which were scaled according to body weight. When physiological parameters were not available, the data for the model were obtained by fitting (tissue affinity, absorption rate and metabolic rate constants). The excretion of arsenic metabolites in urine and faeces are simulated well with the model for both species. This toxicokinetic model for the oral exposure route was validated using data on urinary excretion after repeated oral exposure to As(III) as well as after exposure to inorganic As via drinking water. Absorption by inhalation was validated using data on urinary excretion after occupational exposure to arsenic trioxide dust and fumes. In both cases, the model gave satisfactory results for urinary excretion of the four As metabolites. The PBPK model was also used in the description of the effects on the kinetics of exposure via different routes and for the simulation of various realistic exposure scenarios. The data presented substantiated the assumption that the systemic availability after ingestion is comparable to that after inhalation.

Another PBPK model was developed by El-Masri and Kenyon (2008) to predict levels of arsenic and its metabolites in human tissue and urine after oral exposure. The model accounted for the fate and transport of inorganic As(III) and As(V), as well as mono- and demethylated arsenical metabolites in humans. The model was evaluated against two datasets for arsenic-exposed populations from Bangladesh and the US. The evaluations showed its adequacy and usefulness for oral exposure reconstructions in human health risk assessment, particularly in individuals exposed to relatively low levels of arsenic in water or food (El-Masri et al., 2018).

​Additional references

El-Masri HA and Kenyon EM (2008). Development of a human physiologically based pharmacokinetic (PBPK) model for inorganic arsenic and its mono- and demethylated metabolites. J. Pharmacokinet. Pharmacodyn. 35(1), 31-68.

El-Masri HA et al. (2018). Evaluation of a physiologically based pharmacokinetic (PBPK) model for inorganic arsenic exposure using data from two diverse human populations. Environmental Health Perspectives 126(7), 1-9.

HERAG (2007). Health risk assessment Guidance for metals. https://www.icmm.com/website/publications/pdfs/chemicals-management/herag/herag-fs1-2007.pdf .

Tsuji JS et al. (2019). Dose-response for assessing the cancer risk of inorganic arsenic in drinking water: the scientific basis for use as a threshold. Critical Reviews in Toxicology. https://doi.org/10.1080/10408444.2019.1573804.

Key value for chemical safety assessment

Bioaccumulation potential:

low bioaccumulation potential

Absorption rate - oral (%):

100

Absorption rate - dermal (%):

2

Absorption rate - inhalation (%):

100