Regulatory Toxicology and Pharmacology 81 (2016) 381e386

Contents lists available at ScienceDirect

Regulatory Toxicology and Pharmacology

journal homepage: www.elsevier.com/locate/yrtph

Testing in artificial sweat e Is less more? Comparison of metal release

in two different artificial sweat solutions
Klara Midander*, Anneli Julander, Jolinde Kettelarij, Carola Lide

n
Institute of Environmental Medicine, Karolinska Institutet, Box 210, SE 171 77, Stockholm, Sweden

a r t i c l e i n f o a b s t r a c t

Article history: Metal release from materials immersed in artificial sweat can function as a measure of potential skin
Received 1 June 2016 exposure. Several artificial sweat models exist that, to various degree, mimic realistic conditions. Study
Received in revised form objective was to evaluate metal release from previously examined and well characterized materials in
26 August 2016
two different artificial sweat solutions; a comprehensive sweat model intended for use within research,
Accepted 20 September 2016
based on the composition of human sweat; and the artificial sweat, EN1811, intended for testing
Available online 21 September 2016
compliance with the nickel restriction in REACH. The aim was to better understand whether there are
advantages using either of the sweat solutions in bio-elution testing of materials. Metal release in two
Keywords:
Artificial sweat
different artificial sweat solutions was compared for discs of a white gold alloy and two hard metals, and
Bioaccessibility a rock drilling insert of tungsten carbide at 1 h, 24 h, 1 week and 1 month. The released amount of metal
Bio-elution test was analysed by means of inductively coupled plasma mass spectrometry. Similar levels of released
Hard metal metals were measured from test materials in the two different artificial sweat solutions. For purposes in
White gold relation to legislations, it was concluded that a metal release test using a simple artificial sweat
Nickel composition may provide results that sufficiently indicate the degree of metal release at skin contact.
Cobalt © 2016 Elsevier Inc. All rights reserved.
Sensitizing metals
Contact allergy

1. Introduction material properties such as type of material (pure metal, alloy,

other materials with metal-containing surface coating or metal
Daily use of objects and products made from pure metals, alloys incorporated in the matrix) and condition of the surface (Girod
and other metal-containing materials often involves skin contact et al., 2012; Taylor and Machado-Moreira, 2013).
that may result in deposition of metals on the skin (Erfani et al., With regards to nickel allergy, the release of nickel from mate-
2015; Julander et al., 2009, 2013; Midander et al., 2014). Skin rials immersed in artificial sweat can function as a measure for
exposure to metals can cause local effects, in particular contact assessment of potentially harmful skin exposure (Thyssen et al.,
allergy and allergic contact dermatitis to the metals nickel, cobalt or 2011). This idea illustrates the concept of bio-elution testing by
chromium. Skin may also function as a transport route for further in vitro methods measuring release of a substance or metal into
exposure to metals systemically (Franken et al., 2015). artificial biological fluids that, to some degree, simulate a realistic
When a metal-containing material is in contact with the skin, exposure scenario (Midander, 2009). The result of a release test in
the material surface is affected by sweat present on the skin surface artificial sweat at defined conditions mimicking skin contact, may
resulting in release of metal from the material due to dissolution, be considered a “worst case scenario” under the assumption that all
corrosion and/or wear-processes (Erfani et al., 2015; Julander et al., the nickel released from a material surface in contact with skin
2013). The extent of metal release depends on specific conditions at (sometimes referred to as the bioaccessible fraction of the metal
contact, related to skin, material and environment. Such conditions release) will deposit onto the skin. Data on bioaccessible metal that
include duration and frequency of contact, temperature, pH, pres- is potentially available for skin absorption and following systemic
ence of sweat, sebum and other skin components, as well as or local effects, can be used for risk assessment and protection
against potential harmful skin contact. However, the experimental
limitations, such as various complexity of the artificial sweat and
analytical compromises, need to be taken into account. This is the
* Corresponding author.
E-mail addresses: [email protected] (K. Midander), [email protected] case in the restriction of nickel release from items that are intended

n).
(A. Julander), [email protected] (J. Kettelarij), [email protected] (C. Lide for prolonged contact with the skin in REACH (The Commission of

http://dx.doi.org/10.1016/j.yrtph.2016.09.021
0273-2300/© 2016 Elsevier Inc. All rights reserved.
382 K. Midander et al. / Regulatory Toxicology and Pharmacology 81 (2016) 381e386

the European Communities, 2009). It is however important to note bio-elution testing.

that the bioaccessible metal fraction, an in vitro measure, is not
necessarily equal to either the amount of metal that deposits on 2. Materials and methods
skin at contact (the skin dose) or the bioavailable fraction that is
absorbed into the skin, which are results of exposure in vivo. White gold discs 18 K (Au 75.89%, Cu 16.01%, Ni 6.00%, Zn 2.10%;
The most widely used formulation of artificial sweat was surface area 2.4 cm2/disc), quality control material in EN 1811:2011,
described by Pedersen et al. (1974) as a test solution with a pH buff- were purchased from Allgemeine Gold und Silberscheideanstalt AG,
ered to 6.5, mimicking the salt content or ionic strength of human Pforzheim, Germany. One day prior to starting immersion experi-
sweat. This artificial sweat composition was adopted, and the test ments the discs were abraded with wet emery paper (SiC; 600 grit
conditions were specified, in the European reference method EN1811, followed by 1200 grit), rinsed with deionized water (16.7 MUcm
1),
approved in 1998 and later slightly revised (European Committee for degreased in 0.5% sodium dodecylbenzene sulfate solution (2 min),
Standardization (CEN), 2011), for test of compliance with the nickel rinsed again until there was no foaming, gently dried with paper
restriction in REACH. Ever since, this artificial sweat solution has been tissue, and placed on paper tissue to dry overnight.
used to assess release of nickel and other metals from various mate- Hard metal (cemented tungsten carbide) discs of two different
rials and items with an intended use resulting in skin contact compositions were used; 6% Co, and 15% Co and 0.6% Cr, respec-
(Hamann et al., 2015; Hedberg et al., 2010; Kettelarij et al., 2014; tively. The discs were manufactured and provided by Sandvik AB
Kulthong et al., 2010; Lide
n et al., 1998; Rezi
c et al., 2009; Thyssen (Sandviken, Sweden). Discs of the same materials and batches have
et al., 2010b). In the test protocol of EN1811, a white gold alloy is previously been studied (Julander et al., 2009; Midander et al.,
defined as a quality control material of the method, based on results 2014; Thyssen et al., 2010a). The diameter of each of the discs
from clinical and laboratory testing by CEN/TC283/WG4 in the was ~11 mm, and the total surface area was calculated to be 2.7 cm2
development of EN1811 (Lide
n et al., 1996). The main feature of this each (Julander et al., 2009). The surfaces of the hard metal discs
reference material is that it releases nickel at the level of the limit were used in as-received condition and were rinsed and degreased
value in the restriction; 0.5 mg/cm2/week, upon immersion under the prior to experiments, according to the procedure described above.
conditions of EN1811. Hard metal rock drilling inserts (buttons) were provided by
Other artificial sweat formulations exist, such as the ISO artificial Sandvik Tooling Sverige AB (Va €stberga, Sweden). The inserts con-
sweat for test of “color fastness” to perspiration (ISO Textiles, 1994) tained 6% Co with trace amounts of nickel and chromium. The
at pH 5.5 or 8, and the sweat in the American Association of Textile surface area was calculated to 3.3 cm2/insert. The surface was non-
Chemists and Colorists test method (aimed for similar purposes) at coated and the inserts were used in as-received condition, rinsed
pH 4.3 (AATCC Test method 15e2002). These sweat formulations and degreased prior to experiments.
have for example been used for assessment of silver nanoparticle All glass and plastic utensils were acid cleaned in 10% HNO3
release from antibacterial fabrics (Kulthong et al., 2010; von Goetz prior to use (soaked in 10% HNO3 for 24 h, then rinsed 4 times in
et al., 2013). deionized water and dried in ambient air in the laboratory). Nitrile
There are also sweat models that are not aimed for compliance or vinyl gloves were used throughout all the preparations of sam-
testing of regulatory issues. These sweat models are usually more ples and solutions and during the manual procedures of the im-
complex and based on the constituents of human sweat, hence mersion experiments. Spatulas, spoons and weighing funnels were
realistic test conditions can be further mimicked. Callewaert et al. cleaned with deionized water and ethanol (95%) and dried with
recently assembled an artificial sweat composition including hu- laboratory paper tissues in between weighing of solid chemicals.
man fatty acids (pH 6.5) that was primarily designed to mimic
human axillary sweat for studies of cosmetics and malodour 2.1. Artificial sweat solutions
(Callewaert et al., 2014). A comprehensive artificial sweat model
based on the composition of human sweat has also been developed, Artificial sweat was prepared according to the reference test
tested and used in several studies (Harvey et al., 2010; Stefaniak method EN 1811:2011 on the same day as the immersion experi-
et al., 2014a; Stefaniak and Harvey, 2006). ments started by mixing urea (0.1 wt%), NaCl (0.5 wt%) and lactic
The influence of different sweat components (organic- or inor- acid (0.1 wt%) in deionized water. A Sartorius balance (Sartorius M-
ganic substances such as salts, amino acids, fatty acids, vitamins Pact AX423 with internal calibration, Go € ttingen, Germany) was
etc.) on metal release from materials has been investigated partly. used for weighing of the chemicals. The pH of the solution was
For example, sebum lipids are considered having a limited effect on adjusted with 1 M NaOH, 0.1 M NaOH and 0.1 M HCl to a final pH of
metal release while the addition of amino acids may promote 6.5 ± 0.05 (Mettler Toledo Seven Easy Inlab® Expert Pro pH,
release due to their ability to complex metal ions as in the case of Schwerzenbach, Switzerland).
cysteine and gold (Stefaniak and Harvey, 2006). A structured study Preparation for making the comprehensive artificial sweat, us-
on the influence of artificial sweat components (EN1811) on metal ing a low vitamin formulation (advice by personal communication,
release from a copper-nickel alloy provides more information on Aleksandr Stefaniak, CDC/NIOSH USA, November 2011), took place
how for example pH and constituents of the sweat solution affect during several days.
corrosion and metal release (Brugnoni et al., 2005). Comparison of The amounts of some chemical constituents were too small to
release in different sweat compositions have been performed in be accurately weighted, hence stock solutions, denoted A-U, were
some studies of silver release from antibacterial fabrics and for hard prepared according to description in Table 1. The chemicals for
metal powders (Kulthong et al., 2010; Stefaniak et al., 2010b; von stocks were weighted using a Sartorius AX423 balance.
Goetz et al., 2013). The solid chemicals (Table 1) for the comprehensive sweat
In this study we compare metal release from well-characterized preparation were weighted directly in 1.5 mL Eppendorf tubes
test materials in two different artificial sweat solutions. A (Sarstedt, Nümbrecht, Germany) using a stainless steel spatula, on a
comprehensive sweat model intended for use within research and balance using the four decimal mode (Sartorius Research R160P
based on the composition of human sweat, and the artificial sweat, €ttingen, Germany).
with internal calibration, Sartorius, Go
EN1811, intended for testing compliance with the nickel restriction Deionized water for the comprehensive artificial sweat was
in REACH were used. The aim of this paper is to better understand preheated to 37 ± 2  C in a heating cabinet (Memmert, Schwabach,
whether there are advantages using either of the sweat solutions in Germany) overnight before mixing the comprehensive artificial
 K. Midander et al. / Regulatory Toxicology and Pharmacology 81 (2016) 381e386 383

sweat and start of immersion experiments. amounts increased over time for the studied metals. Release of
A beaker containing 400 mL of the preheated deionized water gold, the main alloy constituent in the white gold alloy, was
and a stir bar was placed on a magnetic hotplate (KEBO-lab generally found in a range ten to twenty times lower (min
MR2002, Stockholm, Sweden) maintaining the temperature at 0.0000716; max 0.0748 mg/cm2) than the released levels of the
37 ± 2  C (temperature was continuously monitored with a ther- alloying elements nickel and copper; these were found to be within
mometer). Stirring was set to form a vortex in the liquid. The the same order of magnitude (min 0.0571; max 1.97 mg/cm2). The
weighed mass of each chemical constituent was thereafter added to one-month time period was included to mimic long term skin
the beaker strictly following the order given in Table 1. Preheated contact of relevance for some jewellery. Although long term sta-
deionized water was used sparingly to rinse each Eppendorf tube to bility (up to 28 days) was proven for the comprehensive sweat
ensure complete removal of the residual chemical. The stock so- model (Stefaniak et al., 2010a), the potential instability of artificial
lutions were shaken before the volume for addition was drawn. sweat solutions should be considered as a possible source of error
Some special handling requirements were necessary for some of at the prevailing test conditions. The solution pH increased from pH
the vitamins; see further comments in Table 1. 5.3 to 8.4 in the comprehensive sweat and from pH 6.5 to 7.7 in
The pH of the comprehensive artificial sweat was adjusted from EN1811 artificial sweat during the 1-month immersion period. This
pH 3.9 to the final pH 5.3 by adding 0.1 M NaOH while continuously could be regarded as a sign of solution instability but could also
monitoring the solution pH. Before the final addition of NaOH, the reflect the hydrogen ion consumption of a general corrosion pro-
solution volume was made up to 1 L in a volumetric flask. The so- cess that is likely to occur to some extent.
lution was stirred and tempered to reach 30  C, and left to stabilize Release of cobalt from two hard metal materials (6% Co; and 15%
for 1 h. The pH was measured before immersion experiments Co and 0.6% Cr, respectively) in both artificial sweat solutions was
started. found within the same range, min 23.6; max 345 mg/cm2 for time
periods up to one week, Fig. 2A. Because the hard metal materials
2.2. Experimental procedure contained chromium only in traces or to a minor proportion, the
released amounts of chromium were low compared to cobalt,
The bottom of 60 mL PP flasks (Sarstedt, Nümbrecht, Germany) Fig. 2B. Chromium release from hard metal discs containing 15% Co
were covered with glass beads (5 mm, Lenz Laborglas, Wertheim, and 0.6% Cr, in EN1811 artificial sweat (literature data, by Julander
Germany), and two white gold discs were placed on top of the beads et al., 2009) was however more pronounced than release at im-
in all flasks. Triplicate sets of discs and one blank sample without mersion in the comprehensive artificial sweat solution. It is
discs, were prepared for each time period of immersion in the two possible that the difference in release derives from the sample
artificial sweat solutions, respectively. Five mL of artificial sweat material (the as-received surface condition is possibly affected by
solution (surface area/volume ratio of ~1), were added to the flasks, time). There is also a risk that strong metal complexes that may
the lids were sealed and the samples were placed in a heating cab- form in the comprehensive artificial sweat, are insufficiently
inet (Memmert, Schwabach, Germany) at 30  C for 1 h, 24 h, 1 week atomised, hence some Cr is “lost” in the chemical analysis.
and 1 month. Prior to addition, the comprehensive artificial sweat Also when we measured metal release from a hard metal
solution had been preheated to 30  C while the EN1811 artificial product, a rock drill insert, released amounts of cobalt, nickel and
sweat was added preheated only for samples to be immersed for 1 h. chromium in the EN1811 artificial sweat solution and the
Similarly, the hard metal discs, in triplicate sets of two discs, were comprehensive artificial sweat model respectively, were found
immersed in 10 mL of the artificial sweat solutions respectively. similar, Fig. 3. The results show cobalt release at three orders of
Triplicate samples of single rock drilling inserts were, due to their magnitude higher than measured release of nickel. Nickel release
geometry, immersed without glass beads, in 10 mL of the different was in turn 10 times higher than release of chromium, which was
artificial sweat solutions. All hard metal samples and corresponding low but more pronounced in EN1811 artificial sweat compared to
blank samples were immersed for 1 h, 24 h and 1 week. the comprehensive artificial sweat model.
After the immersion period, material samples were removed; The main finding of this study is that metal release from the test
the artificial sweat solutions were transferred to new tubes (25 mL, materials, respectively, was found to be within the same range when
PP, Sarstedt, Nümbrecht, Germany), pH was measured before the immersed in either of two different artificial sweat solutions. The
solution was acidified with 65% HNO3 to pH < 1, and stored in a release differed between tested materials (a white gold alloy and
refrigerator until analysis. some hard metal materials) and with duration of immersion.
Chemical analysis was performed with inductively coupled Studies have reported results on metal release in different
plasma-mass spectrometry (X Series II; Thermo Scientific, Wal- artificial sweat solutions with regards to contact allergy, for
tham, MA, USA) using 103Rh as an internal standard. A collision cell example release of cobalt from various materials has been assessed
was used for the quantification of Ni and Cr, to minimize oxide in the EN1811 artificial sweat (Julander et al., 2009; Kettelarij et al.,
formation. Limits of detection (3  SD within calibration blanks) 2014; Midander et al., 2014; Summer et al., 2007). Metal release
were 0.043 mg/L for nickel, 0.067 mg/L for copper, 0.0024 mg/L for from cobalt, tungsten and hard metal (tungsten carbide) powders
gold, 0.019 mg/L for total chromium, and 0.0075 mg/L for cobalt. was compared in the EN1811 artificial sweat and the comprehen-
The comprehensive sweat solution contains nickel and copper sive sweat model and released amounts of metals were found
(Table 1), therefore the concentration of these metals measured in within the same range, irrespective of variations in pH and solution
corresponding blank samples was subtracted from the final results. complexity (Stefaniak et al., 2010b). This finding is in line with re-
sults from other studies where metal release was compared in
3. Results and discussion different artificial sweat solutions or other test media. Released
amounts of silver from nanoparticles in textiles may double in
Average released amounts of nickel, copper and gold from the different artificial sweat solutions (Kulthong et al., 2010; Stefaniak
white gold discs in the two different artificial sweat solutions after et al., 2014b). Cobalt release from a CoCrMo alloy for implants
immersion for time periods up to one month are presented in Fig. 1. was rather constant in EN1811 artificial sweat, cell culture media
In general, the metal release from the white gold alloy was, at and distilled water, while chromium release was affected differ-
the respective time point, within the same range in the EN1811 ently by the three media (Summer et al., 2007). A two-fold differ-
artificial sweat as in the comprehensive sweat model. Released ence in chromium release was observed for FeCr particles in
384 K. Midander et al. / Regulatory Toxicology and Pharmacology 81 (2016) 381e386

Table 1
The chemical constituents for preparation of the comprehensive artificial human sweat, previously described by Harvey et al. (Harvey et al., 2010). Please note that vitamin
concentrations were modified. The low-vitamin formulation of the sweat solution was recommended by Dr Aleksandr Stefaniak, CDC/NIOSH USA (personal communication
with, November 2011).

Order of addition Sweat constituent (comments on preparation aeg

) Mass(g/L) Volume (mL/L) Stock solution Stock preparation

Primary electrolytes and ionic constituents

1 Sodium Sulfate (Na2SO4) 0.0583
2 Sodium Iodide (NaI) 53.2 A 0.010 g in 0.050L dH2O
3 Sodium Fluoride (NaF) 923.6 B 0.025 g in 0.050L dH2O
4 Sodium Bromide (NaBr) 788.8 C 0.015 g in 0.050L dH2O
5 Cadmium Chloride Anhydrous (CdCl2) 16.5 D 0.010 g in 0.050L dH2O
6 Copper (II) Chloride Dihydrate (CuCl2$2H2O) 801.3 E 0.010 g in 0.050L dH2O
7 1 M Ammonium Hydroxide (NH4OH) 5200.0
8 Sulfur (S) a 0.0737
9 Iron Sulfate Heptahydrate (FeSO4$7H2O) 0.0027
10 Lead (Pb)- Reference Solution 1000 ppm 24.9
11 Manganese (II) Chloride (MnCl2) 692.1 F 0.010 g in 0.050L dH2O
12 Nickel (Ni)- Reference Solution 1000 ppm 24.6
13 Zinc (Zn)- Reference Solution 1000 ppm 850.3
14 Sodium Bicarbonate (NaHCO3) 0.2520
15 Potassium Chloride (KCl) 0.4547
16 Magnesium Chloride Hexahydrate (MgCl2$6H2O) 0.0167
17 Sodium Phosphate Anhydrous Monobasic (NaH2PO4) 0.0484
18 Phosphorous Pentachloride (PCl5) 0.0027
19 Sodium Chloride (NaCl) 1.5519
Organic acids and carbohydrates
20 1 M L(þ)-lactate solution (CH3CH(OH)COOH) 14000.0
21 Pyruvic acid (C3H4O3) b 13
22 Butyric acid (C4H8O2) 0.22
23 Acetic acid (C2H4O2) 7.4
24 Hexanoic acid (C6H12O2) c 5.7 G 1.0 mL in 0.050L dH2O
25 Propionic acid (C3H6O2) 0.26
26 Isobutyric acid (C4H8O2) 3.7 H 1.0 mL in 0.050L dH2O
27 Isovaleric acid (C5H10O2) 6.0 I 1.0 mL in 0.050L dH2O
28 D(þ)-Glucose (C6H12O6) 0.0306
Amino acids
29 DL-Alanine (C3H7NO2) 0.0511
30 L-(þ)-Arginine (C6H14N4O2) 0.1359
31 L-(þ)-Aspartic acid (C4H7NO4) 0.0453
32 L-(þ)-Citrulline (C6H13N3O3) 0.0701
33 L-(þ)-Glutamic acid (C5H9NO4) 0.0544
34 Glycine (C2H5NO2) 0.0293
35 L-Histidine (C6H9N3O2) 0.0807
36 L-Isoleucine (C6H13NO2) 0.0223
37 L-Leucine (C6H13NO2) 0.0275
38 L-(þ)-Lysine Monohydrochloride (C6H14N2O2$HCl) 0.0274
39 L-(þ)-Ornithine Monohydrochloride (C5H12N2O2$HCl) 0.0253
40 L-Phenylalanine (C9H11NO2) 0.0215
41 L-Threonine (C4H9NO3) 0.0536
42 L-(e)-Tryptophan (C11H12N2O2) 0.0112
43 L-Tyrosine (C9H11NO3) 0.0308
44 L-Valine (C5H11NO2) 0.0293
Nitrogenous substances
45 Ammonium d
46 Uric acid (C5H4N4O3) 0.0099
47 Urea (CH4N2O) 0.6006
48 Creatinine (C4H7N3O) 989.8 J 2.400 g in 0.250L dH2O
49 Creatine Monohydrate (C4H9N3O2$H2O) 559.3 K 1.000 g in 0.250L dH2O
Vitamins
50 Thiamine Hydrochloride (C12H17ClN4OS$HCl) 33.7 L 0.010 g in 0.050L dH2O
51 Riboflavin (C17H20N4O6) e 24.5 M 0.010 g in 0.050L dH2O
52 Nicotinic Acid (C6H5NO2) (niacin) f 1785.1 N 0.01 g in 0.050L dH2O
53 D-Pantothenic Acid Calcium Salt ((C9H17NO5)2Ca) 548.0 O 0.010 g in 0.050L dH2O
54 Pyridoxine Hydrochloride (C8H11NO3$HCl) 6.4 P 0.010 g in 0.050L dH2O
55 Folic Acid (C19H19N7O6) g 35.3 Q 0.010 g in 0.050L dH2O
56 L-(þ)-Ascorbic Acid (C6H8O6) 978.5 R 0.090 g in 0.050L dH2O
57 Dehydroascorbic Acid 0.00191
58 Inositol (C6H12O6) 720.6 S 0.015 g in 0.050L dH2O
59 Choline Chloride (C5H14NOCl) 181.5 T 1.000 g in 0.050L dH2O
60 p-Aminobenzoic Acid (C7H7NO2) 12.3 U 0.010 g in 0.050L dH2O
a
Upon addition, sweat solution may turn faint yellow and should remain mostly clear until addition of vitamins.
b
In the original protocol this was 12.7 mL/L but for this study we used a solution with different mass percentage.
c
Stock solution went into an ultrasonic bath for 2 min before volume was drawn.
d
Added as ammonium hydroxide with primary electrolytes and ionic constituents.
e
Stock solution went into an ultrasonic bath for 2 min before volume was drawn. Upon addition of riboflavin, sweat solution will turn orange and cloudy.
f
Upon addition of nicotinic acid, sweat solution will turn vivid yellow and remain cloudy.
g
If necessary, add drop of 5N NaOH to clear folic acid stock solution.
 K. Midander et al. / Regulatory Toxicology and Pharmacology 81 (2016) 381e386 385

Fig. 1. Released amounts of Ni (A), Cu (B) and Au (C) per unit surface area of the white gold discs in two different artificial sweat solutions; EN1811 artificial sweat ▫ and the
comprehensive artificial sweat , for various time periods of immersion (ranging from 1 h up to 1 month). The error bars represent one standard deviation of the triplicate samples.
The solid line in (A) indicates the limit value by REACH for release of nickel from articles in prolonged skin contact (0.5 mg/cm /week), and the dashed line the corresponding value
2

(0.2 mg/cm2/week) from post assemblies inserted into pierced parts of the body. Please note the different magnitude of y-axes.

Fig. 2. Released amounts of Co (A) and Cr (B) per unit surface area of two hard metals in two different artificial sweat solutions; hard metal with 6% Co in EN1811 ▫
and the
comprehensive artificial sweat , and hard metal with 15% Co and 0.6% Cr in EN1811 and the comprehensive sweat , respectively. The error bars represent one standard
deviation of triplicate samples. The results in EN1811 artificial sweat are based on duplicate samples and were previously published in another study (Julander et al., 2009). Please
note the different magnitude of y-axes.

Fig. 3. Released amounts of Co (A), Ni (B) and Cr (C) from a hard metal rock drill button insert in EN1811 artificial sweat ▫ and the comprehensive artificial sweat . The error bars
represent one standard deviation of triplicate samples. Please note the different magnitude of the y-axes.

artificial sweat(pH 6.5) and artificial tear fluid (pH 8) (Hedberg The bio-elution tests performed in this study, stagnant immer-
et al., 2010). sion of a white gold alloy and some hard metal materials in
Material properties (nobility, passivity of surface oxide, degree different artificial sweat solutions, do not include any component of
of alloying etc.), chemistry of the test media used for immersion friction and wear. The stagnant test conditions may in reality be
(pH, salt content, metal complexing species, inhibiting species etc.) representative only for a few types of skin contact (some jewelry for
as well as the duration and temperature, together govern the example). However, daily skin exposure to metals often involves
processes of metal release that may be chemical dissolution of the the touching of metallic items, hence the force applied in a touch is
surface oxide, corrosion or ligand-induced release, mainly, or decisive for the degree of metal release (Erfani et al., 2015).
combinations of the processes. Although limited data, our results The present study does not aim at revealing material specific
can demonstrate influence from all of these mentioned properties. release mechanisms in the two types of artificial sweat solutions.
Solution composition and complexity are of particular interest, and This would have required a structured approach to measure the
should be subject to debate, considering the importance of both impact of different artificial sweat components on the metal release
types of models in relation to the specific study goals. from a considered selection of materials of various properties
386 K. Midander et al. / Regulatory Toxicology and Pharmacology 81 (2016) 381e386

(relevant for skin exposure). We rather suggest that the outcome of Release of a Cu-Ni Alloy in Artificial Sweat. Materials for Medical Engineering.
Wiley-VCH Verlag GmbH & Co. KGaA, pp. 243e248.
this study can serve as an example of how data on bioaccessible
Callewaert, C., et al., 2014. Artificial sweat composition to grow and sustain a mixed
metal can be generated at rather simple test conditions, but still human axillary microbiome. J. Microbiol. Methods 103, 6e8.
provide indicative information on how much metal is released at Erfani, B., et al., 2015. Short and frequent skin contact with nickel. Contact Dermat.
skin contact. Our results also illustrate the difficulty to conclude on 73, 222e230.
European Committee for Standardization (CEN), 2011. Reference Test Method for
relevance regarding the use of different artificial sweat models as Release of Nickel from All Post Assemblies Wich Are Inserted into Pierced Parts
estimate of metal skin exposure, since similar levels of released of the Human Body and Articles Intended to Come into Direct and Prolonged
metals from the test materials in both artificial sweat solutions Contact with the skin., Vol. EN1811, p. 2011.
Franken, A., et al., 2015. Vitro permeation of metals through human skin: a review
were detected. The fact that metal release in the comprehensive and recommendations. Chem. Res. Toxicol. 28, 2237e2249.
sweat model, based on the composition of human sweat, and Girod, A., et al., 2012. Composition of fingermark residue: a qualitative and quan-
release in the EN1811 artificial sweat solution, mimicking the ion titative review. Forensic Sci. Int. 223, 10e24.
Hamann, D., et al., 2015. Jewellery: alloy composition and release of nickel, cobalt
strength of sweat, yield a similar release pattern, indicates that and lead assessed with the EU synthetic sweat method. Contact Dermat. 73,
material properties override the importance of accurately 231e238.
mimicking real human sweat in this particular case. Harvey, C.J., et al., 2010. Formulation and stability of a novel artificial human sweat
under conditions of storage and use. Toxicol. Vitro 24, 1790e1796.
In conclusion, our results illustrate that a bio-elution test in Hedberg, Y., et al., 2010. Particles, sweat, and tears: a comparative study on bio-
EN1811 artificial sweat at stagnant conditions can serve its purpose accessibility of ferrochromium alloy and stainless steel particles, the pure
in providing information on the bioaccessible amount of metal that metals and their metal oxides, in simulated skin and eye contact. Integr. En-
viron. Assess. Manag. 6, 456e468.
potentially can deposit onto skin. Tests to determine compliance
Julander, A., et al., 2009. Cobalt-containing alloys and their ability to release cobalt
with regulations, and experimental data for risk assessment pur- and cause dermatitis. Contact Dermat. 60, 165e170.
poses may even benefit from a simpler test procedure that can Julander, A., et al., 2013. New UK nickel-plated steel coins constitute an increased
eliminate potential experimental errors and bias in results. When allergy and eczema risk. Contact Dermat. 68, 323e330.
Kettelarij, J.A.B., et al., 2014. Cobalt, nickel and chromium release from dental tools
skin exposure to metals is the study end point, the actual amount of and alloys. Contact Dermat. 70, 3e10.
metal that is deposited on skin can be measured directly by, for Kulthong, K., et al., 2010. Determination of silver nanoparticle release from anti-
example, acid wipe sampling technique (Lide
n et al., 2006). The bacterial fabrics into artificial sweat. Part. Fibre Toxicol. 7, 1e9.

n, C., et al., 1996. Nickel-containing alloys and platings and their ability to cause
Lide
measure of the metal skin dose in context of skin exposure and skin dermatitis. Br. J. Dermatology 134, 193e198.
as the target organ for sensitizing metals is in this case advanta-
n, C., et al., 1998. Nickel release from tools on the Swedish market. Contact
Lide
geous over release tests in sweat models. Dermat. 39, 127e131.

n, C., et al., 2006. Assessment of skin exposure to nickel, chromium and cobalt
Lide
by acid wipe sampling and ICP-MS. Contact Dermat. 54, 233e238.
4. Conclusion Midander, K., 2009. Metal Particles - Hazard or Risk? Elaboration and Imple-
mentation of a Research Strategy from a Surface and Corrosion Perspective.
Department of Chemistry, Vol. PhD. Royal Insitute of Technology, Stockholm,
The results of metal release from a white gold alloy and some Sweden, p. 69.
hard metal materials are found to be within the same range after Midander, K., et al., 2014. Cobalt skin dose resulting from short and repetitive
immersion in two different artificial sweat solutions; a compre- contact with hard metals. Contact Dermat. 70, 361e368.
Pedersen, N.B., et al., 1974. Release of nickel from silver coins. Acta Derm. Venereol.
hensive sweat model intended for use within research, and the
54, 231e234.
EN1811 artificial sweat solution intended for testing compliance Rezi
c, I., et al., 2009. Metal ion release from electric guitar strings in artificial sweat.
with the nickel restriction in REACH. Corros. Sci. 51, 1985e1989.
The endpoint for study of skin exposure to metals is decisive for Stefaniak, A.B., Harvey, C.J., 2006. Dissolution of materials in artificial skin surface
film liquids. Toxicol. Vitro 20, 1265e1283.
the experimental design. If direct measurements of metal skin Stefaniak, A.B., et al., 2010a. Formulation and stability of a novel artificial sebum
doses or metal absorption are not possible, the metal release in under conditions of storage and use. Int. J. Cosmet. Sci. 32, 347e355.
artificial sweat can provide in vitro data on the bioaccessible metal Stefaniak, A.B., et al., 2010b. Dissolution of cemented carbide powders in artificial
sweat: implications for cobalt sensitization and contact dermatitis. J. Environ.
at simulated skin contact. From this study we conclude that a bio- Monit. 12, 1815e1822.
elution test using a simple artificial sweat composition may provide Stefaniak, A.B., et al., 2014a. Dissolution of the metal sensitizers Ni, Be, Cr in artificial
results that sufficiently indicate the degree of metal release at skin sweat to improve estimates of dermal bioaccessibility. Environ. Sci. Process.
Impacts 16, 341e351.
contact. Less is more, it seems. Stefaniak, A.B., et al., 2014b. Dermal exposure potential from textiles that contain
silver nanoparticles. Int. J. Occup. Environ. Health 20, 220e234.
Acknowledgements Summer, B., et al., 2007. Patch test reactivity to a cobaltechromiumemolybdenum
alloy and stainless steel in metal-allergic patients in correlation to the metal ion
release. Contact Dermat. 57, 35e39.
The authors would like to acknowledge Karin Holm at Stock- Taylor, N.A., Machado-Moreira, C.A., 2013. Regional variations in transepidermal
holm University for the help with metal analysis. Dr Aleksandr water loss, eccrine sweat gland density, sweat secretion rates and electrolyte
composition in resting and exercising humans. Extreme Physiology Med. 2,
Stefaniak is acknowledged for valuable discussions regarding the
1e30.
laboratory procedure of mixing the comprehensive sweat model Textiles, I.S.O., 1994. Tests for color fastness Part E04: color fastness to perspiration.
with a low vitamin formula. ISO 105eE04:2008E.
The authors had financial support from FORTE, the Swedish The Commission of the European Communities, 2009. Commission regulation (ec)
no 552/2009 of june 2009 amending regulation (ec) no 1907/2006 of the eu-
Research Council for Health, Working Life and Welfare (grant ropean parliament and of the Council on the registration, evaluation, Author-
numbers 2009-1765 and 2010-0804). FORTE was not involved in isation and restriction of chemicals (REACH) as regards Annex XVII. Off. J. Eur.
the research or preparation of the article. Union L64, 7e31.
Thyssen, J.P., et al., 2010a. A spot test for detection of cobalt release e early expe-
rience and findings. Contact Dermat. 63, 63e69.
Transparency document Thyssen, J.P., et al., 2010b. Sensitivity and specificity of the nickel spot (dimethyl-
glyoxime) test. Contact Dermat. 62, 279e288.
Thyssen, J.P., et al., 2011. The EU Nickel Directive revisited - future steps towards
Transparency document related to this article can be found better protection against nickel allergy. Contact Dermat. 64, 121e125.
online at http://dx.doi.org/10.1016/j.yrtph.2016.09.021. von Goetz, N., et al., 2013. Migration of Ag- and TiO2-(Nano)particles from textiles
into artificial sweat under physical stress: experiments and exposure modeling.
Environ. Sci. Technol. 47, 9979e9987.
References

Brugnoni, C., et al., 2005. Influence of Experimental Test Conditions on the Ni