Chemicals in Toys

A general methodology for assessment of chemical safety of toys with

a focus on elements

RIVM report 320003001/2008

This report contains an erratum d.d. 26-01-2015 after page 234

J.G.M. Van Engelen1, M.V.D.Z. Park 1, P.J.C.M. Janssen1, A.G. Oomen1, E.F.A. Brandon1,
K. Bouma2, A.J.A.M. Sips1, M.T.M. Van Raaij1

1
National Institute for Public Health and
the Environment (RIVM)
P.O. Box 1
3720 BA Bilthoven
The Netherlands
www.rivm.nl

2
Food and Consumer Product Safety Authority
PO Box 465
9700 AL Groningen
The Netherlands
www.vwa.nl

Contact person: jacqueline.van.engelen@rivm.nl

page 2 of 234 RIVM report 320003001

This report is a revision of the earlier version with report number 0010278A01. The
revision consists of a correction of the values for chromium and tin in Tables 8.2, 8.3,
8.4 and 8.5, on pages 120 to 123.
RIVM report 320003001 page 3 of 234

Acknowledgements

The authors highly appreciate the contributions of the following persons:

Dr. W. Hagens, Ms. P. Hoogerhuis, Mr. C. van der Heijden, Ms. G. Kyriakopoulos and
Dr. A. J. Baars

Also the input of the Advisory Team is highly acknowledged. Members of the Advisory
Team were:
Dr. S. Stefanovic, SGS, Spijkenisse, The Netherlands
Mr. E. van Woensel, Mattel Europe
Mrs. A. Knaap, Former Chair of the Food Contact Material Committee, Bilthoven, The
Netherlands
Dr. P. Bragt, Dutch Food and Consumer Product Safety Authority, The Hague, The
Netherlands
Dr. P. Hakkinen, Joint Research Centre, Ispra, Italy
page 4 of 234 RIVM report 320003001
RIVM report 320003001 page 5 of 234

CONTENTS
Executive summary 8

1 Introduction and approach 15

1.1 Background 15
1.2 Approach 17

2 Elements and their Health-Based Limit Values (TDIs) 19

2.1 Introduction 19
2.2 Update of list of Elements 20
2.3 Update of TDIs 21
2.3.1 Derivation of toxicological limit values 21
2.3.2 Specific Issues 23
2.3.3 Background exposure 24
2.3.4 Method of literature review 25
2.3.5 Updated TDIs for elements 26
2.4 Recommendations 27

3 Exposure to chemicals in toys 29

3.1 Categories of toys and toy materials 29

3.2 Age-related exposure 30
3.3 Exposure scenario categories 33
3.3.1 Direct ingestion 33
3.3.2 Mouthing 33
3.3.3 Inhalation via evaporation 34
3.3.4 Inhalation via dust or spray 34
3.3.5 Skin contact 35
3.3.6 Eye contact 36
3.3.7 Summary 36
3.4 Identification of relevant exposure scenarios 36
3.4.1 Examples of using the exposure scenario selection tree 38
3.5 Formulas and variables for exposure assessments 39
3.5.1 Frequency of exposure 40
3.5.2 Direct ingestion 40
3.5.3 Mouthing 45
3.5.4 Inhalation via evaporation 49
3.5.5 Inhalation via dust or spray 51
3.5.6 Skin contact 52
3.5.7 Uptake 53
3.5.8 Level of detail required for exposure assessments 54
3.6 Conclusions 54
3.7 Recommendations 55

4 From toy to internal exposure – migration versus bioavailability 57

4.1 Introduction 57
4.2 Oral bioavailability 58
4.2.1 Definition of oral bioavailability 58
4.2.2 Sub-processes of oral bioavailability 59
page 6 of 234 RIVM report 320003001

4.3 Relative bioavailability in risk assessment 61

4.4 Tests to estimate the orally bioavailable fraction of a contaminant from toy 62
4.4.1 Tests for inorganic compounds 62
4.4.2 Tests for organic compounds 65
4.4.3 Recommendations for application of tests simulating ingestion and mouthing of toy matrix in risk
assessment 66
4.4.4 Discussion points 68
4.5 Dermal bioavailability 70
4.5.1 Conclusions 71
4.6 Inhalatory bioavailability 72
4.7 Ocular bioavailability 72
4.8 Conclusions and recommendations 72

5 Food contact material 75

5.1 Introduction 75
5.2 EU Directives 75
5.3 Migration tests food contact material 78
5.4 Comparison migration tests of food contact materials and toys 80
5.5 Comparison of migration limits of substances according to tests for food contact material and toy
regulation 83
5.5.1 Lead 83
5.6 Can the methodology of FCM be used for toys? 85
5.7 Conclusions and recommendations 88
5.7.1 Conclusions 88
5.7.2 Recommendations 88

6 Sampling and analysis for certain elements in toys 89

6.1 Introduction 89
6.2 Sampling 89
6.2.1 Sampling strategy 89
6.2.2 Subsamples 90
6.3 Analysis 90
6.3.1 Introduction 90
6.3.2 Test methods 93
6.4 Recommendations 94

7 Proposed general methodology for setting limit values for chemicals in toys 97

7.1 General introduction 97

7.2 Basic starting points 97
7.3 Proposed general methodology to derive limit values for chemicals in toys 98
7.3.1 Use of a risk based framework 99
7.3.2 Necessary level of detail for setting limit values 100
7.3.3 Options within the proposed general methodology 101
7.4 Chemicals with sensitising properties 106
7.5 Hazard aspects 106
7.6 Conclusions 107
7.7 Recommendations 107
RIVM report 320003001 page 7 of 234

8 Application of the proposed methodology to derive limit values for elements in toys 109

8.1 Determining the relevant exposure routes for elements in toys 109
8.2 Defining a relevant health-based limit value 111
8.3 Relevant elements and their health-based limit values 112
8.4 Determining the appropriate option for deriving limit values of elements in toys 114
8.5 Comparing exposure to health-based limit values 115
8.5.1 Option 1: Use of migration data 115
8.5.2 Option 2: Use of product composition data 117
8.5.3 Option 3: Use of risk based data 118
8.6 Migration limits for elements in toys 119
8.7 Hazard aspects 124
8.8 Conclusions 124
8.9 Recommendations 124

References 127

APPENDIX 137

I Answers to issues raised by CSTEE in their opinion and DG Enterprise in their call 137

II Toxicological Profiles 143

III Current definitions and EU legislations on toys 217

IV Existing Toy Categories 219

V Migration Tests 227

page 8 of 234 RIVM report 320003001

Executive summary

The work described in this report was carried out on request of DG Enterprise in view of
contract nr. SI2.ICNPROCE003918500. In the call for tender (ENTR/05/005), the following
two objectives were defined:

1) to examine how the limit values for certain elements that are contained in toys, laid
down in Annex II.II.3 of Directive 88/378/EEC on the Safety of Toys should be
revised according to recent scientific knowledge and to examine whether other
elements should be added to the list in that Annex (4.1.1.1);
2) to examine the way to address the content of chemicals in toys intended for children
under 36 months or intended to be put in the mouth.

In the present report we present a risk based methodology that can be used to assess the
safety of exposure to chemicals in toys. To demonstrate its use we applied this methodology
to elements with the emphasis on toys intended for children under 36 months and on toys
intended to be put in the mouth.

The essence of the methodology is the assumption that exposure of children to chemicals in
toys may not exceed a certain health-based level (Tolerable Daily Intake, or TDI in
mg/kg bw/day). Since children are also exposed to chemicals via other sources than toys we
advocate that a certain percentage of the TDI should be allocated to toys. A number of
arguments for the choice of this percentage are presented for elements. The actual choice of a
percentage is a risk management decision and is not taken as such in this report.
In chapter 2, general issues in deriving health-based limit values (TDI) are discussed. A list
of elements which is thought to be relevant to be included in the Toy Directive is presented
and for each of these elements the most recent and appropriate TDI is given. Since hardly any
data on the presence of elements in toys (apart from those already in the Directive) were
available, it could not be demonstrated which elements are relevant for which toy materials.
The presented list is therefore to be considered as a starting point from which substances can
be removed when data become available to show their irrelevance for toy materials.
To assess whether the exposure of a child via toys is acceptable, exposure characteristics are
required to be taken into account. Three worst case default values for oral contact were
therefore defined. One for textile fibers and material that can be scraped off with teeth
(8 mg), one for dry, pliable or powder-like materials like modeling clay (100 mg) and one for
liquid material like finger paint (400 mg). Also default values for mouthing times for children
< 3y of age (3h) and for toys intended to be put in the mouth for children > 3 y of age (1h)
were derived. Beside that, guidance is given on how to assess exposure to chemicals in toys
following inhalation or dermal contact (chapter 3).
RIVM report 320003001 page 9 of 234

When the contact or exposure scenarios to a toy are defined, information is needed on the
amount of chemical that actually will migrate from the toy material. Chapter 4 describes and
compares different migration test methods. Mouthing can be simulated by means of
extraction with artificial saliva or water. This method will suffice for both organic
compounds as well as elements. Ingestion can be simulated with a current migration test
according to EN71-3. This method will suffice for elements, but will not be generally
applicable to other substances.
The possibility to make use of limit values derived for chemicals in the scope of the Food
Contact Material (FCM) legislation is explored in chapter 5. In principal, it may be possible
that substances with low migration in the FCM framework (< 0.05 mg/kg food or fluid) may
be directly allowable in toys without further testing. However, this can only be allowed when
it is assured that the toy material / matrix (of the finished product) is identical to that tested in
the FCM framework and when the testing conditions conform to Directive 82/711/EEC and
90/128/EEC, are relevant for toy exposure. Even then, there are some uncertainties because
FMC involves static migration while mouthing involves dynamic migration. Furthermore, the
FCM concept allows exposure that may be higher than the fraction of TDI that is allowable
for toys (at least for elements, see chapter 8). Therefore, this extrapolation should only be
used after sufficient experimental validation data become available showing that such an
approach is indeed safe. For the time being, we recommend not to extrapolate FCM migration
limits to toys.
Two other possibilities to use the FCM framework are derived TDIs and the negative lists.
When TDIs have been derived for substances in the FCM framework, these can be directly
used to derive limit values according to the methodology in the present report. Also it can be
decided by risk management, to consider the negative lists of the FMC framework relevant
for toys. In that case, chemicals from this lists may not be used in toys.
In chapter 6, analytical issues like sampling and the use of correction factors are discussed. It
is proposed that a single sample can be used for compliance testing and that all accessible
parts of a toy must comply with EN 71-3. If a toy consists of different materials, subsamples
should be taken from each material. In contrast to the present limit values for elements,
correction values are recommended not to be included in the limit values.

Finally, in chapter 7 a methodology based on the findings in the preceding chapters and
applicable to any kind of substance in any kind of toy is presented. The use of this
methodology will be illustrated for elements in chapter 8 on toys intended for children
< 3 years of age and toys intended to be put in the mouth (> 3 years of age). In this chapter
also a proposal for migration limit values for the elements proposed in chapter 2 is presented.

The analysis as made in this report makes clear that on various topics further research is
required. Recommendations for further actions are given at the end of each chapter.
page 10 of 234 RIVM report 320003001

Risk-Based Methodology
The methodology is set up according to a general approach which can be used for any type of
chemical. Conform the assignment, the methodology is specifically applied to inorganic
elements resulting in a proposal for an update of the (migration) limit values for this group of
chemicals in toys intended for children < 3 y of age and for toys intended to be put in the
mouth.

The basis of our approach is:

Exposure of children to substances in toys should not exceed a certain

health-based level (in mg/kg bw/day)

In this approach, for elements we have chosen a chronic limit value as the relevant health-
based limit level. Exposure to chemicals from toys (e.g. when mouthed) is characterized by
daily exposure during a period of maximum 1-2 years. This would support the use of a sub-
chronic limit value. However, subchronic limit values are not routinely available for all
chemical substances. Chronic health-based limit values on the other hand are routinely
available for most chemical substances, at least for the oral route and assure an adequate level
of protection. Therefore, in the case of elements, it is proposed to use the concept of the
Tolerable Daily Intake (TDI) for setting limit values for elements in toys.

Especially in the case of elements, children are also exposed via other sources. Therefore we
allocate a certain fraction of the TDI to exposure via toys. The allocation of the size of that
fraction is clearly a risk management issue. In chapter 8 arguments for the allocation are
provided.

For elements, therefore, the basic approach can be re-stated as:

The exposure of children to chemicals in toys may not exceed X% of the

TDI (in mg/kg bw/day)

As further guidance on how to apply this approach we offer three options that can be used to
check whether a particular toy can be assessed as safe.
RIVM report 320003001 page 11 of 234

OPTION 1: Use of migration data

This is conceptually the same principle as the present methodology described in EN 71-3. In
contrast to EN 71-3 the derivation of migration limits is made transparent in this report.
It is recognized that exposure to a chemical can only result in exposure when the chemical is
first released from the matrix (thus is: bioaccessible). The migration method as described in
EN 71-3 uses an acid extraction for the release of elements from a toy matrix, and is valid for
assessing exposure via the oral route (and also for dermal exposure). This option can only be
used if one exposure route dominates the others with respect to the fraction of the dose that
actually leads to systemic exposure. For children < 3y of age and for toys that are intended to
be put in the mouth, oral exposure is the most relevant route of systemic exposure. For
elements in toys intended for children < 3 years, migration limits are derived for three
different types of toys: solid (easily to break or bite off), liquid or sticky material and for
material to be scraped off. For toys intended to be put in the mouth (> 3y) only the limit for
scraped off material is relevant, because children of this age display less mouthing behaviour.
For elements, it is assumed that when migration limits for oral exposure are derived, these
cover both mouthing and ingestion.
The basic principle is the following:

The child shall not be exposed to a certain chemical (element) > X% of TDI

Therefore:
The leachable amount of element from a maximum amount of toy that can be ingested
divided by body weight of child should be below X% of TDI.

Using the default values derived in chapter 3 for ingestion and body weight, and using X% of
the TDI as a basis, migration limit values can be calculated for elements.
In chapter 8, four tables are presented, where migration limit values are calculated for
3 different fractions of the TDI (5, 10 and 20%). The choice of the actual percentage is a risk
management decision.
Three tables refer to children < 3 years of age (ingestion via scraping off material, solid and
liquid or sticky material) and one refers to toys intended to be put in the mouth (> 3y).

OPTION 2: Use of product (toy material) composition data

In this option the chemical safety of the toy is demonstrated by documentation on the
amounts of elements present in the toy materials. In this option one can use chemical analyses
of the raw materials used for making the toy. If chemical analyses of all the raw industrial
page 12 of 234 RIVM report 320003001

materials are available and show only trace amounts of elements or such low levels that the
total amount in toys is < X% TDI, then additional testing is not necessary. This
documentation can then show the chemical safety of the product.

The following calculation can be used for demonstrating the safety of a toy.

Element in toy (mg/kg toy material) x weight of toy material (kg)

< X% TDI
Body weight child (kg)

In this approach it is assumed that the element is completely released at once from the
product and available for exposure. Bioaccessibility is thus 100%.

This approach should be viewed as a kind of ‘waiver-opportunity’ for further testing. Those
producers that have data available to demonstrate the absence of elements (or other
substances) in their material can use those data for compliance with the X% TDI limit.
The X% TDI value is therefore again the ultimate limit value. Since at present it might be
difficult for all producers (and importers) to get hold on this information, in the future, under
REACH this will probably improve.

OPTION 3: Use of a quantitative risk based approach

The use of this option is recommended in the following cases:

• Chemicals in toys for which exposure via inhalation may occur
• Chemical in toys for which more than one exposure route contribute significantly to the
systemic exposure
• When the results of a migration test indicate that the bioaccessible amount may exceed
the relevant health-based limit value for the chemical under consideration, and it can be
demonstrated that default factors used for the derivation of these limit values are not
relevant for the toy under consideration. Because a number of (worst case) assumptions
are being made in option 1, option 1 is a conservative approach and may not be relevant
for specific types of toys. For example, in option 1 (and 2) it is assumed that the measured
migration will occur daily. In reality this may not be true for all kinds of toys.
Furthermore, the EN 71-3 acidic test system is worst case in a sense that for elements the
highest migration occurs in the acid environment, simulating the stomach, whereas
absorption of most substances occurs in the less acidic small intestine. Additionally, most
of the elements present in the tested matrix will be released in the first test, the migration
may be lower in reality. A second extraction (e.g. day 2 of mouthing) will usually not
release the same amount of element.
RIVM report 320003001 page 13 of 234

Option 3 provides the opportunity to demonstrate the chemical safety of a product even if the
values of the initial migration test are higher than the values listed in chapter 8. This can be
achieved by using a number of specific exposure scenarios (chapter 3) and – if desired –
refined migration testing (chapter 4). In essence, option 3 can be seen as an EC-type1
examination. This option should only be used when it can be argued convincingly that the
default factors used in option 1 are not relevant for the toy under investigation.

1
EC type examination is the procedure by which an approved body, called ‘Notified Body’ ascertains and
certifies that a model of a toy satisfies the essential requirements of the Directive Safety of Toy
page 14 of 234 RIVM report 320003001

How to read this report?

The methodology presented in this executive summary, and that is discussed in detail in
chapter 7 is the result of discussions as laid down in the other chapters.
We recommend the reader to start reading the ‘simple’ version of the methodology described
in the executive summary and then go to chapter 7. When more guidance or background
information is needed on specific issues, the reader is referred to the respective chapters.

A number of the issues that are discussed in this report were also raised by the CSTEE. These
are answered in chapter 2 to 8. For a short overview of these issues and separate comments,
see Appendix I.
RIVM report 320003001 page 15 of 234

1 Introduction and approach

The work described in this report was carried out on request of DG Enterprise in view of
contract nr. SI2.ICNPROCE003918500. In the call for tender (ENTR/05/005), the following
two objectives were defined:

1) to examine how the limit values for certain elements that are contained in toys, laid
down in Annex II.II.3 of Directive 88/378/EEC on the Safety of Toys should be
revised according to recent scientific knowledge and to examine whether other
elements should be added to the list in that Annex (4.1.1.1);
2) to examine the way to address the content of chemicals in toys intended for children
under 36 months or intended to be put in the mouth

1.1 Background

The permissible levels of ‘bioavailable’ elements from toys as laid down in Council Directive
88/378/EEC, were actually derived in a June 1985 advice by the Scientific Advisory
Committee to examine the toxicity and ecotoxicity of chemical compounds, as published in
report EU 12964 EN.

The Committee chose an approach based on literature data concerning normal weekly intakes
of metals via the diet by adults in the EU, as selected from literature. It was assumed that
children (with assumed body weight of up to 12 kg) would have an intake of 50% of the adult
weekly intake levels (both expressed as μg/week). Leaching from toys should not contribute
more than 10% of the dietary intake, the Committee stipulated. Subsequently the Committee
evaluated the toxicology of the elements dealt with, which included comparison with WHO
Provisional Tolerable Weekly Intakes where available. In this evaluation consideration was
given to children’s sensitivity regarding toxicity and toxicokinetics (absorption) as far as
possible. Based on these evaluations the Committee determined whether for individual
elements the figure of 10% of normal dietary intake being permissible for leaching from toys,
needed adjustment. For antimony, barium, mercury and selenium the toxicity evaluation did
not warrant lowering the figure of 10%. For barium, however, the percentage was lowered to
5% because of the high normal dietary intake for this element. For arsenic and chromium the
percentage was lowered to 0.1% and 1.0%, respectively, because of their known
carcinogenicity and mutagenicity via the oral route. For cadmium the percentage was lowered
to 5% because the normal dietary intake already approached the WHO Provisional Tolerable
Weekly Intake for the element. For lead the percentage was lowered to 1% because of the
known high sensitivity of children for lead neurotoxicity.
page 16 of 234 RIVM report 320003001

In Annex 2 to the Opinion of the Committee the approach was summarized. The following
table was derived from this annex:
Table 1-1 Permissible intake of certain elements, derived from Annex 2 of the June 1985 advice by the Scientific
Advisory Committee to examine the toxicity and ecotoxicity of chemical compounds, as published in report EU
12964 EN.
Sb As Ba Cd Cr Pb Hg Se
Adult intake (μg/week) 30 1400 7000 175 400 1000 70 700
Children’s intake (μg/week) 15 700 3500 87.5 200 500 35 350
Assumed contribution from 10% 0.1% 5% 5% 1% 1% 10% 10%
toys
Children’s daily permissible 0.2 0.1 25 0.6 0.3 0.7 0.5 5
intake from toys in μg

Note that subsequently, in Standard EN 71-3: 1994, these levels were converted to migration
limits expressed as mg/kg toy. For this conversion it was assumed that a child ingests 8 mg of
toy material per day, based on which concentrations of ‘bioavailable’ elements in toy
materials could be calculated. These concentrations were converted to migration limits from
toys expressed as mg/kg toy after adjusting ‘to minimize the exposure of children to toxic
elements and to ensure analytical feasibility.’

In their call DG Enterprise was seeking a party which would be able to propose a sound
methodology for setting limit values of elements present in toys. Several years ago CEN
already put effort in the development of such an approach, but was confronted with scientific
criticism by the EU Scientific Committee on Toxicity, Ecotoxicity and the Environment
(CSTEE). The opinion of the CSTEE on ‘Assessment of the bioavailability of certain
elements in toys’ distinguished two main topics, i.e. 1) suitability of the proposed limit values
and 2) the necessity of updating the standard EN-3:1994.

The following issues were addressed by the CSTEE and/or by DG Enterprise in their call and
will be considered in the present report:

1) Choice of elements: what elements should toys be analyzed for.

2) Intake of toy-material: what is the daily intake of toy material for children to be
considered.
3) Definition for bioavailability: The CSTEE recommends to change the definition for
bioavailability from the soluble extract having toxicological significance’ into the
amount of each element in the toy which could be absorbed into the systemic
circulation of a child.
4) A single representative or not: The CSTEE does not accept that it is possible to take a
single representative from many toys because of their heterogeneous nature. Sampling
is a critical step in the enforcement and testing for compliance, that is often
overlooked. In this report it will be studied if a single sample is representative for the
RIVM report 320003001 page 17 of 234

whole toy, or if several sub-samples have to be taken and analyzed. This will depend
on the nature of the toy as some toys may consist of different parts.
5) Limit values and maximum bioavailability: how to deal with correction factors for
analytical variation.
6) Health-based limit values: latest scientific knowledge and associated revisions of
tolerable daily intakes (TDI) and average daily intakes (ADI) should be reviewed and
special focus should be paid to the potential sensitivity of children.
7) Bioavailability or migration: should the limit values for elements be expressed in
terms of bioavailability or in terms of migration?
8) Toys intended for different ages. One of the discussion points identified is whether
different toys intended to be used for different ages should be distinguished. If so,
then several exposure scenarios have to be established for different ages.
9) Food Contact Materials. It was proposed by DG Enterprise to examine whether the
food contact materials (FCM) framework could provide a basis for setting limit values
in toys.
10) Analytical test methods. It is stated that corresponding analytical test methods should
be available.

1.2 Approach
In this report a general methodology is presented to derive limit values for chemicals in toys.
The methodology is set up according to a general approach which can be used for any type of
chemical. However, because the first objective of the assignment involves deriving new limit
values for inorganic elements in toys, a proposal for an update of relevant (migration) limit
values for this group in toys will be given.

The basis of the whole approach is in essence the same as that for the current derivation of
migration limit values for toys, as laid down in the Toy Directive and in EN 71-3, namely :

The exposure of children to chemicals in toys should not exceed a certain health-based level
(in mg/kg bw/day)

In the approach, it is determined to what level a chemical may be present in toy material in
order to reach a certain defined level of exposure.

General issues in the derivation of TDIs were described. A major objective for the present
report was to update the toxicological information and health-based limit values (TDIs) on
individual elements. The original list of the 8 elements was extended with 10 additional
elements. As far as possible, review of recent, existing evaluations conducted by recognized
international bodies were used. Because of the specific focus on toys and on elements, several
special issues were addressed like e.g. children as a sensitive subgroup and background
page 18 of 234 RIVM report 320003001

exposure. In view of dermal contact with toys, available information on ‘local effects upon
dermal contact,’ is reviewed.

Assessing the exposure involves the consideration of child specific exposure scenarios and
exposure factors such as those related to playing behaviour and physiological characteristics.
Information is collected that can be used for different levels of exposure assessments, varying
from simple exposure duration factors to guidance for specific cases where an extensive
exposure assessment is desired. For the definition of the amount of toy that children can be
exposed, to simple weighing experiments were carried out.

Concepts and migration tests as used in the Food Contact Material (FCM) Framework were
considered for their applicability for risk assessment of toys. Different migration tests were
described and compared, both with respect to exposure to toy material and to Food Contact
material. A proposal was made on how to sample and how to make use of correction factors.

The risk-based methodology that is presented in this report, is illustrated with three options
that can be used assess whether the certain health-based level is not exceeded.

This project started in January 2006. An Interim report was presented to DG Enterprise on
March 31, and the final draft version on June 30.

For this project an Advisory Team was invited with individuals representing the toy industry,
risk assessment and risk management groups. During the preparation of the report 2 times our
Advisory Team was consulted. The team consisted of the following individuals:
Dr. S. Stefanovic, SGS, Spijkenisse, Mr. E. van Woensel, Mattel Europe, Ms A. Knaap,
Former Chair of Former chair of the FCM Committee, Dr. P. Bragt, Dutch Food and
Consumer Product Safety Authority, Dr. P. Hakkinen, JRC, Ispra.

The first meeting was on March 6, 2006, about 2 months after the start. Discussed were the
selection of the list of elements, the exposure scenarios and the first principles of the
methodology. The second meeting was on June 15, 2006, where the draft final report was
discussed and suggestions were given for a more clear presentation of the methodology.

We highly appreciate all their valuable comments !

RIVM report 320003001 page 19 of 234

2 Elements and their Health-Based Limit Values (TDIs)

2.1 Introduction

Basic to our methodology for deriving limit values for toys, is the Health-based Limit Value
that is usually denoted as the Tolerable Daily Intake or TDI. The TDI denotes the daily dose
of a chemical than can be ingested daily throughout the entire lifetime without adverse effects
for the individual in question. Using the TDI as the basis differs from the approach
previously used to derive the permissible levels of ‘bioavailable’ elements from toys as laid
down the Council Directive 88/378/EEC. As already explained above in the Introduction, the
levels as specified in this Directive were based on a previous advice (from 1985) by the
Scientific Advisory Committee to examine the toxicity and ecotoxicity of chemical compounds
(CSTEE). The approach chosen by this Committee was based on weekly intakes of metals via
the diet by adults in the EU, as selected from literature. Of these normal weekly intakes
percentages ranging from 0.1 to 10% were allocated as allowable exposure from playing with
toys. These percentages were chosen based on toxicological evaluation of the elements in
question. Thus, this approach did not use health-based limit values (TDIs) as the point of
departure but toxicological information was used only in determining the percentage that
exposure through toys was allowed to add to normal dietary background exposure (see the
Introduction for further description of the 1985 derivation by the Scientific Advisory
Committee to examine the toxicity and ecotoxicity of chemical compounds).

A major objective for the present report was to update the toxicological information and
health-based limit values (TDIs) on individual elements. As described in section 2.2, we
extended the original list of the 8 elements with 10 additional elements we consider relevant
with respect to toy-related exposures. Relevant toxicological information for the 18 elements
is presented concisely as toxicity profiles, containing for each element the basic information
deemed most relevant in the context of toy-related exposures. Available existing limit values
(TDIs) and their derivation are described in these profiles, from which the value considered
most suitable for use in the present context of evaluating toy-related exposures is then
selected. Given the time schedule of the project and the mostly huge toxicological data bases
available for the elements reviewed, use of existing evaluations and TDI-derivations was
inevitable. As will be seen hereafter, existing evaluations conducted by recognized
international bodies were available for all elements. Moreover, for virtually all elements this
included evaluations conducted after the year 2000.

Because of the specific focus on toys and on elements, several special issues are addressed in
the profiles, like children as a sensitive subgroup and background exposure. Further, in view
of dermal contact with toys, available information on ‘local effects upon dermal contact’ was
page 20 of 234 RIVM report 320003001

reviewed. Absorption from the gastro-intestinal tract is an important issue for elements and
accordingly existing knowledge on this point is discussed in the profiles.

The individual toxicity profiles for the 18 elements are attached to the present report as
Appendix II.

2.2 Update of list of Elements

The original list as laid down in Council Directive 88/378/EEC was as follows:
Antimony
Arsenic
Barium
Cadmium
Chromium (trivalent, hexavalent)
Lead
Mercury
Selenium

It proved very difficult to get information on the presence of additional elements in different
toy material, especially since routinely toys are only tested for the above 8 elements as
specified in the Directive. As far as information was received, no conclusions can be drawn
about which elements occur most frequently and/or whether some elements are specific for
particular toys/toy materials. Therefore we used the following strategy: in addition to the
present list of 8 elements, we used the list for Food Contact Materials and a list that contains
elements that are found in the waste phase of plastics.

The ‘Synoptic Document’ (EU, 2005) lists the monomers and additives notified to the EU
(EFSA, SCF) as substances which may be used in the manufacture of plastics or coatings
intended to come into contact with foodstuffs. Some of the materials used for food packaging
may also be used in toys, therefore it is reasonable to assume that these elements included in
the ‘Synoptic’, will also be present in toys. These elemental additives supplement the above
list of 8 elements. Based on the Synoptic Document the following elements/ions are added to
the above list of eight:

Aluminum Silver
Boron Tin (inorganic)
Cobalt Tin (organic
Copper Zirconium
Manganese
RIVM report 320003001 page 21 of 234

Although only inorganic elements were to be considered with regard to limit values in toys in
the present report (see chapter 1, objective 1) it is noted that, for tin organic forms may be
added to synthetic materials as bio-stabilizer. Following a request by DG Enterprise, we have
listed organotins for review, based on the rationale that their much higher toxicity compared
to inorganic tin warrants specific attention for these chemicals.

Recently a survey was carried by RIVM on the use and waste-disposal of synthetic materials
(RIVM, 2006). This work was carried out on behalf of the Inspection of the Netherlands’
Ministry of Housing, Spatial Planning and the Environment following EU Council
Regulation (EEC) No 259/93 of 1 February 1993 on the supervision and control of shipments
of waste within, into and out of the European Community. The survey includes an inventory
of metals present in waste due to use in synthetic materials. Of these the following are in
addition to those already listed above:

Molybdenum
Nickel
Strontium
Titanium
Zinc

Given their low toxic potency, molybdenum, zirconium and titanium are considered not to be
relevant for inclusion in the Directive and therefore no toy limit value was derived for these
three elements.

2.3 Update of TDIs

2.3.1 Derivation of toxicological limit values

Toxicological limit values such as the Tolerable Daily Intake (TDI) for contaminants and the
Acceptable Daily Intake (ADI) for compounds applied intentionally in the production of
foods, are a long-established tool within chemical risk assessment. Toxicological evaluation
of chemical substances aimed at derivation of such limit values is carried out by various
national and international bodies. Within the EU various expert panels of the European Food
Safety Authority (EFSA) evaluate toxicological dossiers on different categories of chemical
agents relevant for food (food additives, food contact materials, contaminants, pesticides
etc.). Another important international body specifically active in the food area is the Joint
Expert Committee on Food Additives (JECFA), which is a programme of the World Health
Organisation (WHO) dealing with both food additives and food contaminants. The
International Programme on Chemical Safety (IPCS) also of the WHO deals with
environmentally relevant chemicals in its Enviromental Health Criteria. Within the EU, the
Existing Substances Programme comprehensively evaluates chemicals that have a wide use
in industry and in consumer products. In the USA the Environmental Protection Agency
page 22 of 234 RIVM report 320003001

routinely derives Reference Doses (RfDs) for a wide range of environmental chemicals
whereas the US Agency for Toxic Substances and Disease Registry (ATSDR) does similar
work for soil contaminants.

Two general approaches are used in the toxicological evaluation, i.e the threshold approach
and the non-threshold approach. The latter is used for genotoxic carcinogens, the former for
all other compounds. As will be seen hereafter, of the elements dealt with here, only
hexavalent chromium falls under the category of genotoxic carcinogens. The non-threshold
approach involves linear extrapolation from observed tumour incidences to risk-specific
doses such as one in a million for lifetime exposures. The latter level is often called the
Virtually Safe Dose. The threshold approach uses a No Observed Adverse Effect Level
(NOAEL) or Lowest Observed Adverse Effect Level (LOAEL) which is divided by
uncertainty factors, leading to a limit value (TDI, ADI). Because the threshold approach is
applied for almost all elements included in this report, it is discussed further below.

In the toxicological evaluation, findings in individual experiments are judged as to their

relevance against those in other studies and other species. For each study and each endpoint a
NOAEL is to be derived. Based on a full evaluation of all toxicity data available for the
compound under scrutiny an overall-NOAEL is then selected that will serve as the basis for
limit value derivation. The overall-NOAEL should be the highest relevant dose where no
(adverse) effect was observed. As has been pointed out in numerous publications, the
NOAEL has important statistical limitations relating to the design of the study from which it
derived. Increasingly an alternative measure, the Benchmark Dose (BMD), is being used as
point of departure in limit value derivation.

In order to derive a health-based limit value, uncertainty factors (US terminology) or

assessment factors (EU Existing Substances Program) are applied to the overall NOAEL or
BMD in order to extrapolate from experimental animals to humans (interspecies
extrapolation; default value is 10) and from humans to sensitive humans (intraspecies
extrapolation; default value is 10). The use of these factors is a default approach fraught with
considerable uncertainty. In recent years the trend has been to use, wherever possible, factors
based on compound-specific biological data. This use of data-derived uncertainty factors was
first advocated in IPCS (1994), where traditional 10-fold factors were subdivided in factors
for toxicokinetics and toxicodynamics, thus allowing a more structured use of existing data,
with the goal of making more reliable extrapolations. As can be seen in the profiles, in many
of the evaluation for the elements that are the subject of the present report, applied factors
tend to be lower than 100 (the traditional default). This is due to the toxicity database, that
often included usable human data (obviating the need of animal to human extrapolation), and
the fact that several elements are essential nutrients and their toxicity has to be judged against
their daily requirements.
RIVM report 320003001 page 23 of 234

2.3.2 Specific Issues

2.3.2.1 Children as a sensitive group

In the context of toy-related exposures any specific toxicological information on children’s
susceptibility is highly relevant. This topic has raised considerable interest in recent years,
going back to the seminal report Pesticides in the diets of infants and children published by
US National Research Council in 1993 (NRC, 1993). Regulatory bodies and toxicological
advisory committees working on their behalf, are increasingly paying attention to this topic in
their reviews of individual chemicals. The US-ATSDR in its Toxicological Profile series on
individual chemicals systematically reviews available information on children’s
susceptibility. In these documents ATSDR also provides some general considerations on this
topic. EFSA also where relevant addresses the issue in its evaluations for individual
chemicals. For the present report a general discussion of children’s susceptibility was not
considered necessary, this being available elsewhere. The available chemical-specific
information, however, was selected and summarised.

The TDI as a limit value is intended to be protective for (potentially) sensitive subgroups in
the population which includes children as a group. Nevertheless, data bases on which TDIs
are based vary widely in the degree to which experiments and observations address this
specific sensitivity of children or young animals. For lead, for instance, this issue has been
investigated extensively and the TDI for lead is actually based on its neurotoxic potential for
children. For other elements available data in this area are fragmentary only and their TDIs
are based on studies with exposure to adult humans or (young-)adult animals. Of course the
uncertainty factors used in deriving TDIs are selected taking into account such limitations
but, especially in the present context of toy-related exposure, special attention seemed
warranted. This was given specifically in selecting suitable TDIs, where those values were
chosen expected to be providing the most adequate protection for children.

2.3.2.2 Local effects upon dermal contact

When children play with toys dermal contact occurs. Therefore information on direct effects
by potentially released elements on the skin is relevant and accordingly is included in the
toxicity profiles. Thus the potential for producing dermal irritation and sensitization was
reviewed for individual elements. Again the ATSDR documents were a primary source of
information on this issue. As was to be expected, dose-response information for these effects
is scarce. This kind of studies is mostly carried with high concentrations and mostly no
attempt is made to determine NOAELs for skin irritation and sensitization. For the latter
endpoint this situation is beginning to change where for example methods are being
developed that allow quantification of sensitizing potential (using Mouse Local Lymph
Node-assay results). For the elements reviewed here, however, irritation and sensitization
data were of a qualitative nature only. Exceptions to this are chromium and nickel, two well
known inducers of contact dermatitis. For these elements the dose/concentration-response has
been examined in a large number of tests.
page 24 of 234 RIVM report 320003001

2.3.2.3 Absorption
For chemicals in general but for elements especially, absorption in the gastrointestinal tract is
an important factor on which the ultimate risk posed by an exposure will depend. When using
a TDI in the assessment of any given exposure, consideration should be given to the concept
of the internal dose, which is the dose actually reaching the blood stream after external
exposure via the mouth, skin or lungs. Gastro-intestinal absorption and the concept of
bioavailability are of prime importance here. In depth discussion of these topics is provided
in chapter 4. A crucial point with reference to the TDI is that it represents an external dose
(ingested amount per kg body weight/day), ultimately based on an experiment with its own
specific bioavailability. The bioavailability of the external dose of any exposure, from toys
for instance, mostly will be different from that in the experiment from which the TDI was
derived. Consequently in comparing exposure with the TDI this may lead to unwarranted
conclusions, which should be avoided by the proper consideration of differences in
bioavailability as sketched. For this reason, information on gastro-intestinal absorption is
included in the toxicity profiles.

As can be seen in the profiles, for elements frequently much information on gastro-intestinal
absorption is available. Typically a wide range in absorption percentages is found in different
experiments, reflecting strong matrix effects for this group of chemicals.

2.3.3 Background exposure

Any toy-related exposures to elements take place against a background of exposure to the
same element via other exposure routes such as food or non-food consumer products. When
using the TDI as a tool in safety evaluation, the most relevant comparison in principle is with
total exposure instead of only one specific exposure. Thus background exposure is relevant
also in the present context and this background exposure should be considered when making
an informed choice on which part of the TDI can responsibly be allocated to the specific
exposure route of toys.
Exposure assessment for elements and for chemicals in general is a highly complex field of
study, not in the least because of the wide variation in exposure situations and human
activities pertaining to them. Equally important, certainly for elements, are variations, both
natural and man-made, in concentrations in air water and food across countries. Thus data on
normal background levels of elements in these compartments and of estimates of daily
general population exposures usually will always provide a partial picture only. This certainly
goes for the data as presented in the individual profiles: the normal background as estimated
there should be seen as an indication of the real background exposure of children across the
European Union.

As can be seen in the individual profiles, data specifically on children’s exposures are not
available for all elements, even when using data from non-European countries. Where such
RIVM report 320003001 page 25 of 234

data gaps existed either the adult estimate was adopted or this estimate was adjusted to a
value considered more reflective of what would be a typical children’s exposure level.

2.3.4 Method of literature review

As already stated above, given the time schedule of the project and the mostly huge
toxicological data bases available for the elements reviewed, we chose to use existing
evaluations and TDI-derivations. Evaluations conducted by recognized international bodies
were available for all elements. Moreover, for virtually all elements this included evaluations
conducted after the year 2000. Thus, a retrograde approach was chosen, in which critical use
was made of TDIs or other relevant health-based limit values as proposed by recognized
international scientific expert groups such as those of the WHO (JECFA, IPCS, JMPR) and
the EU (SCF, EFSA Existing Substances). Adequate reviews of sufficiently recent date being
available for all elements, no further literature search for original publications was considered
necessary. Table 2-1 provides some basic information of the primary sources of information
that were used.

Table 2-1 Major toxicological review documents

Publisher, name Description of contents Limit value

EU Risk Assessment Reports In depth review all toxicity endpoints, all MOS calculation
exposure scenarios quantified
EFSA opinions Review of all toxicity endpoints, data on TDI or UL
food exposure
ATSDR Toxicological profiles Comprehensive review of all toxicity Acute, intermediate
endpoints, data on exposure via all routes and chronic MRLs
OEHHA Drinking-water Comprehensive review of oral toxicity Drinking-water
endpoints, comprehensive review of guideline
exposure via food
IPCS Environmental Health Comprehensive review of all toxicity Guidance values for
Criteria endpoints, data on exposure via all routes risk assessments
US-EPA Toxicological review Comprehensive review of all toxicity RfD
endpoints, brief review of exposure via all
routes
RIVM Review Soil Brief review of all toxicity endpoints, TDI
contaminants estimate of total background exposure
(non-soil-related)
JECFA food contaminants Comprehensive review of oral toxicity TDI
endpoints, comprehensive review of
exposure via food
These review documents embody a critical evaluation of all relevant toxicological and
environmental data and represent the risk assessment consensus among recognized experts.
The TDIs as selected for the elements dealt with here, must therefore be regarded as
page 26 of 234 RIVM report 320003001

optimally reflecting the scientific state-of-the-art. As such they provide a solid basis for the
toxicological part of the present report. For detailed discussion of specific toxicological
endpoints and of individual toxicity studies the reader is referred to the review documents as
referenced in the individual toxicity profiles.

2.3.5 Updated TDIs for elements

For toxicological profiles on the individual elements, see Appendix II. In Table 2-2, an
overview is given of updated TDIs. Also presented in the table is information on background
exposure and risks for local skin effects.

Table 2-2 TDIs, background exposure and skin irritation/sensitisation risk for elements
TDI Background Skin irritation and
Value (μg/kg Reference exposure child sensitisation contact
bw/day) (µg/kg bw/day) risk (qualitative
indication)
Aluminum 750 Newly derived 300 Low
Antimony 6 WHO, 2003 0.53 Unknown
Arsenic 1.0 RIVM, 2001 0.4-0.7 Low
Barium 600 ATSDR, 2005 9 Unknown
Boron 160 EFSA, 2004a 80 Low
Cadmium 0.5 RIVM, 2001 0.45 Low
Chromium 5 RIVM, 2001 1 Low
trivalent
Chromium 5a RIVM, 2001 0.1b High
hexavalent
Cobalt 1.4 RIVM, 2001 0.6 Medium
Copper 83 SCF, 2003a 60 Low
Lead 3.6 JECFA, 1993, 2.0 Low
RIVM, 2001
Manganese 30 (160)c OEHHA, 2004 130 Unknown
Mercury 2 IPCS, 2003 0.1 Medium
Nickel 10 Newly derived 8 High
Selenium 5 SCF, 2000, 2 Low
RIVM, 1998
Silver 5 US-EPA, 1996a 1.3 Low
Strontium 600 US-EPA, 1996b 18 Unknown
Tin inorganic 2000 JECFA, 2001 290 Unknown
Tin organic 0.25 EFSA, 2004b 0.083 High
Zinc 500 SCF, 2003b 350 Low
a This value only takes into account non-carcinogenic effects by hexavalent chromium; for the carcinogenic
effect by hexavalent chromium a highly uncertain Virtually Safe Dose of 0.0053 μg/kg bw/day has been
proposed by OEHHA (1999). A new drinking-water cancer bioassay with hexavalent chromium is being
conducted within the US-NTP.
b Estimate for a child playing on CCA-treated timber as given in EU-RAR (2005).
c The value of 30 μg/kg bw/day applies to exposures above normal dietary intake For the current method of
calculation of the allowable toy-related exposure level (10% of the TDI) this TDI was converted to a value
usable for evaluating total daily exposure (inclusive of normal dietary intake). Thus for manganese a figure of
160 μg/kg bw/day was used for calculation (estimated background added to ‘non-dietary’ TDI).
RIVM report 320003001 page 27 of 234

2.4 Recommendations

At present, research on elements in toys is directed exclusively at the elements already

included in the Directive. We recommend further research on which elements are present in
toys by means of chemical analysis of a representative sample of toy (materials), in time
allowing the removal of any irrelevant elements from the list while others might be added.
More in general it would be useful if Industry could prove, by means of measurements
already available, which of these elements are irrelevant.
page 28 of 234 RIVM report 320003001
RIVM report 320003001 page 29 of 234

3 Exposure to chemicals in toys

To warrant the safety of using elements and other chemicals in toys and toy material requires
demonstration that the level of exposure of children to these chemicals does not exceed
relevant health-based limit values. Assessing this exposure involves the consideration of child
specific exposure scenarios and exposure factors such as those related to playing behaviour
and physiological characteristics. This chapter will evaluate the available information on
exposure scenarios and factors of toys and toy material, such as exposure pathways and
activity patterns. As discussed in chapter 7, it is not always necessary to perform a detailed
exposure assessment for chemicals in toys or toy materials. This chapter provides information
that can be used for different levels of exposure assessments, varying from simple exposure
duration factors to guidance for specific cases where an extensive exposure assessment is
desired. The information can also be used for an EC type examination2. Where possible,
default values for exposure factors will be provided that may be used in the exposure
assessments.

3.1 Categories of toys and toy materials

Within the EU, regulations on toys are harmonized, based on Council Directive 88/378/EEC
on the approximation of the laws of Member States concerning the safety of toys. The
definition for ‘toy’ used in Council Directive 88/378/EEC is as follows:

‘Any product or material designed or clearly intended for use in play by children of less than
14 years of age’

Annex I of Directive 88/378/EEC provides a list of articles that are not regarded as toys,
which is included in appendix III.
Depending on its purpose, toys can be categorized based on different criteria. A review of toy
categories used for legislative and other purposes is given in Appendix IV.

In summary, it is possible to categorize toys based on different criteria: possible safety

hazard, type of material, type of use, intended age groups and type of exposure.
For the purpose of general safety of toys, including mechanical and thermal safety, it may be
most relevant to categorize toys based on possible safety hazards. However, for elements in
particular, no groups of toys can be identified that may pose the greatest risk of exposure to
elements.

2
EC type examination is the procedure by which an approved body, called ‘Notified Body’ ascertains and
certifies that a model of a toy satisfies the essential requirements of the Directive Safety of Toy
page 30 of 234 RIVM report 320003001

For the purpose of determining which migration tests are appropriate, it may be relevant to
categorize toys based on the material of which they consist.
A relevant way of categorizing toys for the purpose of safety evaluations and setting limits
for elements and other chemicals in toys is based on exposure information, such as contact
routes and exposure scenarios.
Intended age group categories are often used to determine whether a toy is suitable for
children under 3 years of age. The value of basing exposure categories on intended age
groups for the purpose of setting limits for elements and other chemicals is discussed
extensively in the next paragraph.

3.2 Age-related exposure

One objective of this project is to examine the way to address the content of chemicals in toys
intended for children under 36 months or intended to be put in the mouth. For the exposure
assessment, it is possible to include exposure scenarios specific for young children, in
particular mouthing and ingesting toy material, and crawling over toy surfaces. However, the
question arises whether it is justified to include these scenarios only in exposure assessments
for toys intended for children under 36 months or intended to be put in the mouth, and not for
toys intended for older children.

According to EU Council Directive 88/378/EEC, toys which might be dangerous for young
children (under 36 months) need to be labelled ‘not suitable for under 36 months/three years’.
Particular risks relating to young children cited in Annex II of the Directive are
• Toys and their component parts, and any detachable parts of toys which are clearly
intended for use by children under 36 months must be of such dimensions as to
prevent their being swallowed and/or inhaled
• Toys containing inherently dangerous chemicals or preparations must bear a warning
stating that the toys must be kept out of reach of very young children.

It could therefore be argued that exposure assessment for toys labelled not suitable for
children under 36 months will not need to include exposure scenarios specific for children
under 36 months, because these toys should not be accessible to children of this age.
However, the EU Council Directive 88/378/EEC also states that ‘toys may be placed on the
market only if they do not jeopardize the safety and/or health of users or third parties when
they are used as intended or in a foreseeable way, bearing in mind the normal behaviour of
children’. Although certain toys are not intended for young children, the odds that they will
mouth toys can be considered relatively high. In addition, in the EN 71-3 standard it is stated:
‘For the purposes of this standard, the following criteria are considered appropriate in the
categorisation of sucking, licking or swallowing: toys intended for children up to 6 years of
age, i.e. all accessible parts and components where there is a probability that those parts or
components may come into contact with the mouth’ The CSTEE (2004) concluded that it is
RIVM report 320003001 page 31 of 234

foreseeable that children under 6 will have access to toys intended for children over 6.
CSTEE stated that these toys might also pose a risk for children under 6 and should therefore
be tested.
Indeed, many families consist of children of different ages, and it can be anticipated that
young children will often have easy access to toys owned by their older siblings. Mouthing
hands and objects is natural behaviour for babies, infants and toddlers (Van Engelen et al.,
2004). Indeed, the list of objects mouthed by children under 36 months, as observed in
several mouthing studies, consisted of many items (not just toys) not intended for children
under 36 months (DTI, 2002; Juberg et al., 2001; De Groot et al., 1998; Smith and Norris,
2003). In fact, Smith and Norris (2003) reported that at least an estimated 75% of items that
were mouthed by children in their study were considered not intended to be mouthed. Hence,
young children having access to toy material intended for older children can be considered
‘use in a foreseeable way’. It is therefore not justified to exclude exposure scenarios specific
for young children from the exposure assessment simply based on the intended age category
of the toy under consideration.
It has been argued that there has to be a degree of carer responsibility and supervision of
young children, in particular those who still have tendency to mouth everyday objects and
toys. It is common knowledge that toys containing small parts are unsuitable for children
under 36 months of age due to the choking hazard. Parents and other caregivers can be
expected to keep toys containing small parts out of reach from children under 36 months.
However, toxic hazards are not visible on the toy itself. The toy may be labelled unsuitable
for children under 36 months of age on the packaging, but in practice, package materials are
disposed of and the information is lost. In addition, some toys appear to be labelled
inappropriately, as shown by a market survey of plastic toys by the Dutch Food and
Consumer Product Safety Authority (VWA, 2005). The sampled toys included a number of
bath baby toys which were labelled unsuitable for children under 36 months or which were
not labelled at all, although these toys are likely to be mouthed by young children.

To identify toys for which exposure scenarios specific for young children should be
considered, the approaches related to the EU measures recently adopted for phthalates in toys
and child articles may be helpful. In 1999, the EU adopted measures prohibiting the placing
on the market of toys and childcare articles intended to be placed in the mouth by children
under three years of age made of soft PVC containing one or more of 6 specific phthalates
(DINP, DEHP, DBP, DIDP, DNOP, BBP) (Council Directive 1999/815/EC).
In response to these measures, Denmark has prohibited the use of phthalates in toys and
childcare products for children under three years of age (Statutory Order No. 151, 1999).
However, the Danish Environmental Protection Agency is often faced with the problem
whether or not a toy is suitable for children under three years of age. To help producers,
importers and buyers of toys, decisions regarding this issue are made public in the form of a
list of toys suitable for children under three years old, which is updated regularly3. Guidelines

3
Available at http://www.mst.dk
page 32 of 234 RIVM report 320003001

are also provided by the CEN, which are partly based on the extensive Age Determination
Guidelines prepared by the US CPSC (discussed in Appendix IV). However, due to the use of
often subjective criteria, even the most extensive age determination guideline may still not
avoid all ambiguity on the suitability of a particular toy for a certain age group.

An alternative approach has been used in the amendment of Council Directive 76/769/EEC
on the marketing and use of certain dangerous chemicals and preparations. The amendment
restricts the use of DEHP, DBP and BBP in all toys and childcare articles. The restrictions for
the use of the other phthalates DINP, DIDP and DNOP are less severe for reasons of
proportionality. The use of these phthalates is restricted only in toys and childcare articles
which can be placed in the mouth by children. To help identifying toys and childcare articles
or parts of toys and childcare articles which can and those which can not be placed in the
mouth by children, a guidance document has been prepared4.

Similarly, decisions on whether the exposure assessment for a toy should include exposure
scenarios specific for young children can be based on the suitability of the toy for children
under 36 months of age. However, the question of suitability may lead to much discussion for
certain toys. In addition, it may be inappropriate to use different criteria for phthalates than
for other possible hazardous chemicals. It is therefore recommended to base this decision on
whether the toy can or can not be placed in the mouth by children and/or whether the toy can
be crawled on. The guidance document which will be used for the phthalate regulations can
be applied to identify toys that can be placed in the mouth. A separate guidance document
would need to be drafted for toys than can be crawled on.

In conclusion, similar to the EU measures adopted on phthalates and similar to

the conclusion of the CSTEE (2004), we propose that the exposure assessment
of all toys which do not contain small parts or long chords (or are otherwise
dangerous from a physical-mechanical point of view), but can be placed in the
mouth or can be crawled on by children should include exposure scenarios
specific for young children, regardless of the intended age category of the toy.
However, this is clearly a risk management decision. Therefore, throughout the
current report, exposure scenarios specific for young children will only be
considered for toys intended for children under 3 years of age.

4
http://ec.europa.eu/enterprise/chemicals/legislation/markrestr/guidance_document_final.pdf
RIVM report 320003001 page 33 of 234

3.3 Exposure scenario categories

To set appropriate limits for chemicals in toys, information on the exposure to these
chemicals in toys is needed. The route and level of exposure to chemicals in toys is linked to
both the physico-chemical properties of the chemical and to how the toy is used by the child,
which can be described by exposure scenarios. The following paragraphs will discuss which
exposure scenarios are relevant for toys, and special reference will be made to elements in
toys.

3.3.1 Direct ingestion

As discussed earlier, direct ingestion of toy and toy material can be assumed to occur mainly
by children under 3 years of age due to the oral exploration behaviour that is natural at this
age (Van Engelen et al., 2004). Toys intended for children this age are regulated such that
they should not contain small detachable parts that may pose a choking hazard. These parts
should therefore also not be accessible for ingestion. However, some liquid toys used by
children under 36 months of age such as finger paint are easily swallowed. Toys that consist
of dry, brittle, powder-like or pliable material, such as chalk crayons, plaster or modelling
clay may also be ingested, for example via hand-mouth contact. In addition, some toys may
have a layer of paint or other coating, or textile fibres that may easily be scraped off and
swallowed. Ingestion of scraped off material is also relevant for toys intended for older
children which are intended to be placed in the mouth, such as whistles. The direct ingestion
scenario can be relevant for elements in toys.

Figure 3-1 Example of scraping off material: girl chewing on pencil

3.3.2 Mouthing
Similar to the direct ingestion scenario described above, mouthing of toys can be assumed to
occur mainly by children under 36 months of age. In fact, some toys available on the market
are specifically designed to be mouthed, such as teething rings. It should be noted that
mouthing behaviour studies demonstrated that children mouth on a broad range of items,
including toys and other items not intended to be mouthed (De Groot et al., 1998; DTI, 2002;
Juberg et al., 2001; Reed et al., 1999; Smith and Norris, 2003; Tulve et al., 2002). Although
the dimensions of some toys may be such that they cannot be placed in the mouth, ridges can
page 34 of 234 RIVM report 320003001

still be sucked on. In addition, some toys intended for children over 3 years of age are
intended to be placed in the mouth. The mouthing scenario can be relevant for elements in
toys.

Note: For many toys, both mouthing and direct

ingestion may occur. Depending on the
properties of the toy and on physico-chemical
properties of the chemical under consideration,
one of these scenarios will likely be more
relevant for systemic exposure than the other.
Only the most relevant scenario will need to be
considered.

Figure 3-2 Example of mouthing: a teething blanket

3.3.3 Inhalation via evaporation

A number of toys may release chemicals in the air via evaporation, such as the solvent in a
felt pen. To evaporate, the chemical would need to be quite volatile to be available for
inhalation. This route of exposure is therefore not relevant for elements. In general, this route
is likely to be less relevant for systemic exposure if oral exposure also occurs. For toys
releasing volatile chemicals that may cause local effects in the lungs, this exposure scenario
should be considered.

3.3.4 Inhalation via dust or spray

Some toys may release considerable amounts of dust, such as plaster mix and crayons (for
example, when beating out a brush). Other toys may release chemicals in the air via a
spraying system. At present, very few examples of toys in the form of sprays are known.
Some doll perfume sprays are available already, but these may be regulated under the
cosmetics directive. Nevertheless, more toy sprays may be marketed in the future.
Contrary to evaporating chemicals, chemicals in sprays or dust do not necessarily need to be
volatile to be available for inhalation. Again, although the oral route may be more relevant
for the systemic exposure to chemicals in these toys, the inhalation route of exposure may
need to be considered for chemicals which may cause local effects in the lungs, for example
respiratory sensitizers.
RIVM report 320003001 page 35 of 234

Figure 3-3 Example of a toy in the form of a spray: spray chalk

3.3.5 Skin contact

Most if not all toys will at some point contact some part of the skin. Many toys are handled
with the hands, but some may also be contacted by skin of other body parts, such as foot
contact with a canvas on which children may jump, and arm and leg contact with costumes.
Dermal exposure to elements in such toys is especially relevant for sensitizing elements such
as nickel. For example, the Danish EPA found levels of 2.96 µg/g nickel in so-called ‘slimy’
toys (Danish EPA, 2005). These toys were not expected to contain nickel and the detected
levels are assumed to be contaminations from the manufacture of the products, e.g. from the
use of nickel-containing catalysts. For systemic exposure, this exposure route is probably not
very relevant for elements, as the dermal uptake is very low (chapter 4).

Figure 3-4 Example of skin contact other than hands with a toy: a baby gym play mat
page 36 of 234 RIVM report 320003001

3.3.6 Eye contact

Eye contact may not seem a relevant exposure category for toys, as it has been reported that
most injuries with toys are of a physical rather than a chemical nature (Consument en
Veiligheid, 2001). However, this report was based on analysis of data on cases of eye injury
reported in an injury information system, which registers accidents for which patients
underwent medical first aid treatment in a selection of Dutch hospitals. Injuries of lesser
seriousness such as eye irritancy are not registered in this system, but could potentially occur
when, for example, a chemical in finger paint ends up on the hands and subsequently contacts
the eyes when they are rubbed. However, effects such as eye irritancy are of such mild and
transient nature that it may be considered irrelevant. For the remainder of this report, this
scenario will therefore not be considered further.

3.3.7 Summary
The wide range of available toys results in many different possible exposure scenarios. Six
general exposure scenarios have been described that should cover the most relevant ways of
exposure to chemicals in toys: direct ingestion, mouthing, inhalation via evaporation,
inhalation via dust or spray, and skin contact.
As with the creation of any categories, the exposure scenario categories described above have
been created with a specific toy and toy material in mind. It would be convenient to provide
lists of toy types for each exposure scenario category, based on which exposure scenario is
relevant for the exposure assessment of a certain chemical in a certain toy. However, this
approach bears the risk that certain (new) toys and toy materials will not be listed in any
category. The heterogeneity of toy types complicates the creation of categories which will
cover every single toy on the market. In addition, one particular toy may consist of several
parts and materials to which different ways of exposure may occur. The next section will
outline an alternative approach using exposure information to identify the relevant exposure
scenarios for the exposure assessment of elements (and other chemicals) in toys.

3.4 Identification of relevant exposure scenarios

As more than one exposure scenario may be relevant for one particular type of toy, it is
proposed to determine the relevant exposure scenarios on a case-by-case basis, by means of a
scenario selection tree, rather than providing rigid groups of toy types for each category. A
scenario selection tree for toys intended for children under 3 years of age has been designed
(Figure 3-5). The scenario selection tree can also be used for toys intended for children over
3 years of age, although the oral scenarios (direct ingestion and mouthing) may not be
relevant for this age group, unless the toy is intended to be placed in the mouth.
RIVM report 320003001 page 37 of 234

Scenario I: direct ingestion

Scenario II: mouthing
(sucking/licking)
Scenario III: inhalation via
evaporation
Scenario IV: inhalation via dust or
spray
Scenario V: skin contact

START

yes
Toy (material) can no
Scenario I Scenario II
be directly ingested

yes Chemical can be

Scenario III
released by
evaporation
no

Scenario IV yes Toy (material) can

release dust or spray

no
Can contact which
body parts

yes Hands
Scenario V

no

yes Other body parts

Scenario V

no

STOP

Figure 3-5 Exposure scenario selection tree

In using the scenario selection tree as depicted in Figure 3-5, the following questions need to
be answered:
1) Can the toy (material) be directly ingested? For all liquid toys such as finger paint,
it is assumed that it can be directly ingested. Some toys may be covered with
layers of for example paint, which may easily be scraped off during mouthing of a
toy. Potential exposure to chemicals in this layer should be assessed by means of
the direct ingestion scenario.
page 38 of 234 RIVM report 320003001

Most toys for which no toy material can be ingested, can still be mouthed. It
should be noted that even if the dimensions of the toy are such that they cannot be
placed in the mouth, it can still be licked and sucked on.
2) Can the chemical of interest be released from the toy by evaporation? Some toys
may contain volatile substances (not relevant for elements) that may be released
during use.
3) Can the toy release dust or spray? For example, while using crayons, chalk dust
may be released and subsequently inhaled.
4) Which body parts can the toy contact? For example, a book will predominantly
contact the hands, whereas a baby play gym mat may contact face, arms and legs
as well.

3.4.1 Examples of using the exposure scenario selection tree

To demonstrate the use of the scenario selection tree, some examples of toys are given below:
modelling clay, crayons and a baby gym play mat.

3.4.1.1 Modelling clay

1) Can the toy be directly ingested? Yes, parts of clay are small enough to be placed
in the mouth and swallowed. Exposure scenario I needs to be considered.
2) Can the chemical of interest be released from the toy by evaporation? If a volatile
chemical is used in the modelling clay, this may be released. In this case, exposure
scenario III needs to be considered
3) Can the toy release dust or spray? No, the consistency of modelling clay does not
directly indicate significant dust formation and clay is not available in spray form.
4) Which body parts can the toy contact? Modelling clay is handled with the hands.
It is unlikely to significantly contact skin of other body parts. Exposure scenario V
needs to be considered for hands.

3.4.1.2 Crayons
1) Can the toy be directly ingested? Yes, even if the dimensions of a crayon are such
that it cannot be directly ingested, parts of brittle crayons can easily be bitten off.
Exposure scenario I needs to be considered.
2) Can the chemical of interest be released from the toy by evaporation? If a volatile
chemical is used in the crayon, this may be released. In this case, exposure
scenario III needs to be considered.
3) Can the toy release dust or spray? Yes, crayons that are made of powder-like
material such as chalk may generate considerable amounts of dust when used.
Exposure scenario IV needs to be considered.
4) Which body parts can the toy contact? Crayons are handled with the hands and are
unlikely to significantly contact skin of other body parts. Exposure scenario V
needs to be considered for hands.
RIVM report 320003001 page 39 of 234

3.4.1.3 Baby gym play mat

1) Can the toy be directly ingested? No, the dimensions of a baby gym play mat are
of such dimensions that it cannot be directly ingested. Corners or parts of the play
mat can be licked and sucked on, exposure scenario II needs to be considered.
2) Can the chemical of interest be released from the toy by evaporation? If a volatile
chemical is used in the play mat, this may be released. In this case, exposure
scenario III needs to be considered
3) Can the toy release dust or spray? No, the material the play mat is generally made
of (textile) does not indicate significant dust formation. Spray is also irrelevant.
4) Which body parts can the toy contact? A baby may crawl or lie on the play mat,
exposing face, (possibly bare) hands, arms and legs. Exposure scenario V needs to
be considered for hands and other body parts.

NOTE: The use of the scenario selection tree is primarily meant to prevent overlooking any
relevant exposure scenarios, by selecting all scenarios that might possibly occur. As a result,
some of the exposure scenarios selected may seem irrelevant for a certain chemical-toy
combination. It is up to the exposure assessor to provide arguments for omitting the
consideration of a specific exposure scenario. The (ir)relevance of a selected scenario will
also become clear by simply plugging in the exposure factor values in the mathematical
formula used to calculate the exposure, as outlined in the next section.

3.5 Formulas and variables for exposure assessments

Once the relevant exposure scenarios for a particular type of toy have been determined, the
exposure to chemicals such as elements can be assessed by using the applicable mathematical
formulas related to the scenarios and the appropriate values for the variables, or exposure
factors. This section covers information on the mathematical formulas and exposure factors
for each exposure scenario listed in the scenario selection tree. Where possible, default values
are given, many of which have been taken from the fact sheet on children’s toys made for the
ConsExpo program, a software package for deriving quantitative exposure assessments
(Bremmer et al., 2002). It is not possible to present default values for each type of toy.
Calculation of the exposure assessments therefore depends to a great extent on the sound
judgment of the exposure assessor.
It should be noted that the resulting exposure assessments are rough estimates due to the use
of very simple pragmatic mathematical formulas which oversimplify real exposure. In
addition, the selection of realistic worst case parameter values results in an assessment that
may considerably overestimate exposure. More exact exposure assessments are possible with
the use of mathematical models which may describe the exposure more precisely, such as the
higher tier models in ConsExpo (Delmaar et al., 2005). The assessment may further be
refined by using probabilistic methods (Bosgra et al., 2005).
page 40 of 234 RIVM report 320003001

The information in this section applies to chemicals in toys in general and may not always be
relevant for elements. Special reference to elements will be made where applicable.

3.5.1 Frequency of exposure

The exposure levels calculated with the formulas given below refer to one exposure event,
i.e. one event of playing with the toy. For comparison with health-based limit values that are
related to chronic exposure such as the TDI, the exposure levels can be assumed to occur
daily. However, for some toy types, daily exposure may not be realistic. Exposure
assessments for these toys need to be adjusted accordingly. Where possible, defaults will be
provided per exposure scenario.

3.5.2 Direct ingestion

The amount of element (or other chemical) ingested can be calculated as:

D = A× w /W
f body
with
A : amount of toy (material) swallowed [kg]
wf : weight fraction of the chemical in the toy (material) [mg/kg]
Wbody : body weight of the exposed person [kg]

The parameter values needed for this calculation are:

A : amount of toy material swallowed, which depends on whether the toy is made of dry
or liquid, pliable or otherwise sticky material, or whether the ingested material is from
scraping off a toy layer.
• Toys consisting of dry, brittle, pliable or powder-like material. For some toys, a
considerable amount of material may be bitten off or ingested via hand-mouth
contact, such as chalk crayons, modelling clay and plaster powder.
For chalk crayons, ConsExpo’s fact sheet on toys derived a rough default value of
6 mg/min as a default, based on studies on ingestion of soil by children (cited in
Bremmer and Van Veen, 2002). It was further assumed that children play with
crayons for 45 minutes. Total amount swallowed during one event is then
6 x 45 = 270 mg. To illustrate how much this amount approximately is, we weighed
parts of clay (Figure 3-6) and chalk crayon (Figure 3-7 and 3-8).
RIVM report 320003001 page 41 of 234

Figure 3-6: 270 mg modelling clay material

Figure 3-7: 270 mg chalk crayon material

Figure 3-8: 290 mg chalk crayon material

Based on these simple weighing experiments, the default of ingesting 270 mg appears
to be quite an overestimation. For risk assessments within the Dutch Soil Protection
Act, a default of 100 mg is now used for ingestion of soil by children (Otte et al.,
2001). It is proposed that this value is used as a default for ingested amount of dry,
pliable or powder-like toy materials, although further research is warranted.
It is emphasized that this default applies to children under 3 years of age only, as
these children display most mouthing behaviour.
The ingestion of 100 mg by children is considered reasonable, but may not occur
daily. For exposure assessment refinement purposes, we propose to use a frequency
of 1/week for this ingestion default when the exposure is compared to a chronic
health-based limit value. This is a rough estimate and needs further research.
• Toys consisting of liquid material. The amount of liquid toy material that may be
ingested via hand-mouth contact is likely considerably higher than for dry material.
For finger paint and other products that stick to the skin, ConsExpo’s fact sheet on
toys derived a default value of 30 mg/min (Bremmer and Van Veen, 2002). It was
page 42 of 234 RIVM report 320003001

further assumed that children play with finger paint for 45 minutes. Total amount
swallowed is then 30 x 45 = 1350 mg.

Figure 3-9: 1350 mg finger paint

Figure 3-10: 290 mg finger paint

The pictures above show that this amount may be an overestimation, although 100 mg
may be too little. We propose to use a value of 400 mg as a default, but again, this
value is a rough estimate and needs further research.
It is emphasized that this default applies to children under 3 years of age only, as
these children display most mouthing behaviour.
Similar to the ingestion default for dry, brittle, powder-like and pliable materials, an
ingestion of 400 mg may occasionally occur, but not daily. For the purpose of an
exposure assessment refinement, when comparing exposure to a chronic health-
based limit value, we propose to use a frequency of 1/week as a default. This is a
rough estimate and needs further research.
• Layers of toy material scraped off. The amount of toy material scraped off with the
teeth while mouthing a solid toy is likely considerably lower than the amount of
liquid, pliable or sticky toy material that may be ingested. In the fact sheet on toys, a
single ingestion of paint from a toy car is estimated based on the product volume
(0.05 cm3) and density of paint (2 g/cm3). This amounts to a total of 0.1 g (Bremmer
et al., 2002). However, this value is considered an overestimation. The weight of
paint material scraped off from a pencil and textile fibers pulled off a pluche toy are
in the order of magnitude of the 8 mg used in the current EN-71 (see Figure 3-11 to
RIVM report 320003001 page 43 of 234

3-16). It is therefore recommended to keep this value as a default for ingested layers
of scraped off toy material.
In contrast to the previous two defaults for ingested amounts, this default also
applies to toys intended to be mouthed by children over 3 years of age.
Furthermore, with regard to frequency, it is assumed that the small amount of 8 mg
material can be scraped off from a toy every day.

Figure 3-11: 8.6 mg of textile fibres from a pluche toy

Figure 3-12: 1.3 mg of textile fibres from a pluche toy

Figure 3-13: 8 mg of scraped off chalk crayon material

page 44 of 234 RIVM report 320003001

Figure 3-14: 8 mg of modelling clay material

Figure 3-15: 8 mg of scraped off pencil material

Figure 3-16: 0.5 mg scraped off pencil material

wf : fraction of the chemical in the toy material. This depends entirely on the material the
toy consists of and no default values can be given. The total amount of chemical migrated
from the toy (material) can also be used, for example if composition data of the material are
not available. The amount of migrated chemical depends entirely on the chemical- material
combination and should be assessed with methods described in chapter 4.

Wbody : body weight of the exposed child. The risk assessment work of the CEN/TC
52/WG9 used 10 kg for the mass of a child (European Committee for Standardization (CEN),
2003). Mean, standard deviation and 25th percentile default values for body weight of Dutch
children from 1.5 months to 17.5 years have been given in the general fact sheet of ConsExpo
(Bremmer et al., 2006):
RIVM report 320003001 page 45 of 234

Table 3-1 Mean, standard deviation and 25th percentile default values for body weight of Dutch children from
1.5 months to 17.5 years. Source: Bremmer et al., 2006, derived from a study by TNO in 2000.
Age Body weight
[kg]
Months Years Mean SD 25th percentile

1.5 4.65 0.52 4.30

4.5 6.75 0.79 6.21
7.5 8.30 1.0 7.62
10.5 9.45 1.1 8.69
13.5 10.3 1.2 9.47

1.5 11.1 1.9 9.85

2.5 13.9 2.1 12.5
3.5 16.0 2.9 14.1
4.5 18.4 3.1 16.3
6.5 23.1 3.8 20.6
9.5 32.4 6.0 28.4
12.5 44.8 8.1 39.3
13.5 50.0 9.0 43.9
16.5 62.9 9.0 56.8
17.5 65.3 10 58.2

• If the direct ingestion will be done by young children displaying mouthing behaviour,
the body weight for this age needs to be used. According to the table above, the 10 kg
value used by the CEN approximately corresponds to the 25th percentile body weight
of Dutch children aged 1.5 years. The study by the DTI (2002) reported that children
aged 6-9 months display the highest mouthing durations for toys. The 25th percentile
body weight for this age group (i.e. 7.5 months in the table above) is 7.62 kg. A body
weight of 7.5 kg is suggested as a default.
• For ingestion of scraped off toy material from toys intended to be mouthed for
children over 3 years of age, the bodyweight of a child of approximately 3-4 years of
age should be used. The 25th percentile of Dutch children 3.5 and 4.5 years of age is
14.1 and 16.3 kg, respectively. A default value of 15 kg is proposed.

3.5.3 Mouthing
The amount of element or other chemical ingested via mouthing can be calculated as:

R ×S
D = A× w /W × (1 − exp( − m × t ))
f body A×w f

with
A : the total amount of product that is being mouthed [kg]
Rm : rate at which the chemical migrates from the product
(per unit area) [kg/m2.s]
S : the surface area of the product that is being mouthed [m2]
page 46 of 234 RIVM report 320003001

wf : weight fraction of the chemical in the product [fraction]

t : mouthing time [s]
Wbody : body weight of the exposed person [kg]

The parameter values needed for this calculation are:

A : the total amount of toy that is being mouthed. This amount can be determined by
calculating the volume of the (part of the) toy that can be mouthed and multiplying this value
with the density of the material of which the toy is made. In the ConsExpo fact sheet for toys,
the volume for a teething ring, a cuddly toy and a plastic doll have been estimated at 20, 50
and 100 cm3, respectively (Bremmer et al., 2002). To illustrate, for a doll made of plastic
with a density of 1 g/cm3, the total amount that can be mouthed is 100 x 1 = 100 g. In
practice, the amount of toy that can be mouthed highly depends on the dimensions of the toy
and therefore should be determined on a case by case basis.

Rm : the rate at which a chemical migrates from the product. Methods of migration tests
will be discussed in chapter 4.

S : the surface area of the (part of the) toy that is being mouthed. The risk assessment
work of the CEN/TC 52/WG9 for organic chemicals assumed a value of 10 cm2 for the area
of toy mouthed (European Committee for Standardization (CEN), 2004). The same value is
used in ConsExpo’s fact sheet on toys, which was based on a study by Könemann (1998).
The value of 10 cm2 probably refers to the surface area of a toy that can be placed in the
mouth at once. However, considering that a toy may be mouthed for three hours (as discussed
below), a much larger surface area may be covered. Again, as for amount of toy being
mouthed, this depends highly on the dimensions of the toy and should be determined on a
case by case basis.

wf : fraction of the chemical in the toy (material). This depends entirely on the material
the toy consists of and no default values can be given. The total amount of chemical migrated
from the toy (material) can also be used, for example if composition data of the material are
not available. The amount of migrated chemical depends entirely on the chemical- material
combination and should be assessed with methods described in chapter 4.

t : Mouthing time. The risk assessment work of the CEN/TC 52/WG9 for organic
chemicals assumed a value of 3 hours per day for the duration of mouthing (European
Committee for Standardization (CEN), 2003). The value of 3 hours was adopted by the
CSTEE opinion of 1998, 6th plenary meeting in the framework of the risk assessments of
phthalates in toys. In ConsExpo’s fact sheet on toys, default mouthing times for toys for
mouthing and other toys have been calculated based on a study from De Groot et al. (1998):
RIVM report 320003001 page 47 of 234

Table 3-2 Default mouthing times for toys for mouthing and other toys. Based on: De Groot et al. (1998)
Age [months] Default mouthing times
[minutes per day]
Toys for Other toys
mouthing
4.5 11 27
7.5 21 63
13.5 0 9
18 0 3

Similar average mouthing times have been reported by Juberg et al. (2001).
Two new studies have been published since. The first study was conducted by the UK
Department of Trade and Industry (DTI), which reported mean and maximum mouthing
times of toys and other items for children of different age groups up to 5 years of age (DTI,
2002; Smith and Norris, 2003). Children aged 6–9 months displayed the most mouthing
behaviour. For this group, the mean mouthing time on toys was 39 minutes.
A second study was published by the US Consumer Product Safety Commission which
conducted an observational study of mouthing activity by 169 children aged 3-36 months
(Babich et al., 2004). From this study, daily mouthing times for selected objects can be
calculated by multiplying the hourly mouthing duration (min/h) with the daily exposure time.
The hourly mouthing duration is defined as the time per hour that the article is actually in the
child’s mouth or touching the lips. The daily exposure time is defined as the time a child is
awake and not eating and is estimated by the model:

Tday = 9.46 + 0.0375 x Age (months)

page 48 of 234 RIVM report 320003001

For children between 3 and 36 months of age, exposure time is thus roughly 10 hours. Using
the mean or 95% hourly mouthing duration data of the study and an exposure time of
10 hours, the calculated mouthing times are:

Table 3-3 Calculated mouthing times, as derived from Babich et al. (2004)
Age group Object mouthed Mean hourly 95% hourly Mean daily 95% daily
studied mouthing mouthing mouthing mouthing
[months] duration duration time [min] time [min]
[min/h] [min/h]
3-11 Soft plastic toys 0.13 0.69 1.3 6.9

Soft plastic 0.19 0.44 1.9 4.4

teethers, rattles

Non-soft plastic 1.8 6.5 18 65

toys, teethers,
rattles
12-23 Soft plastic toys 0.18 0.88 1.8 8.8

Soft plastic 0.02 0.1 0.2 1.0

teethers, rattles

Non-soft plastic 0.56 1.8 5.6 18

toys, teethers,
rattles
24-36 Soft plastic toys 0.07 0.21 0.7 2.1

Soft plastic 0.02 0.00 0.2 0.00

teethers, rattles

Non-soft plastic 0.21 0.94 2.1 9.4

toys, teethers,
rattles

The observed mean mouthing times of these studies are significantly lower than the value of
three hours as used by the CSTEE. However, the use of this value can be clarified when one
takes into account the skewness of the data, as it has been observed in these mouthing studies
that few children mouth objects for a long period and many children mouth objects for a short
time or not at all (Greene, 1998). For example, in the DTI study, the mean mouthing time by
children aged 6-9 months was 39 minutes, whereas the maximum mouthing time for toys was
3 hours and 46 minutes (DTI, 2002; Smith and Norris, 2003). Similarly, in the study by
Juberg et al. (2001), the mean and median daily mouthing durations of non pacifiers
(including teethers and toys) by children aged 0-18 months were approximately 35 and
RIVM report 320003001 page 49 of 234

15 minutes, respectively, whereas daily mouthing durations of over 300 minutes were also
observed.
To safeguard the relatively small group of children that display these longer mouthing times,
it is recommended to continue using three hours as a default for mouthing duration. However,
this example of highly variable mouthing durations supports the use of probabilistic methods
to adequately assess the exposure to chemicals in toys that can be mouthed.
As can be seen in the table above, toys intended to be mouthed by children under 36 months
are not necessarily mouthed for longer durations than other toys.

A different value for mouthing duration can be used for toys that are intended for older
children and intended to be placed in the mouth, such as whistles and balloons. Although no
data was found on the mouthing duration of such toys, it can be assumed that the duration is
shorter than the mouthing duration by young children. A default value of one hour is
proposed for mouthing duration of toys intended for older children and intended to be placed
in the mouth, although this value is quite arbitrary and further research on this is warranted.

Wbody : Body weight of the child. For toys intended for young children, the body weight of
children displaying most mouthing behaviour (6-9 months of age, approximately 7.5 kg)
should be used, as in the direct ingestion model. For toys intended to be placed in the mouth
by older children such as whistles, the body weight of a child of approximately 3-4 years of
age should be used. The 25th percentile of Dutch children 3.5 and 4.5 years of age is 14.1 and
16.3 kg, respectively. A default value of 15 kg is proposed.

3.5.4 Inhalation via evaporation

For volatile chemicals, the mean event concentration in the air can be calculated as follows:

1 A × wf T 1 A × wf 1
< Cair >= × × ∫ e− qt dt = × × × (1 − e− qT )
T V 0 T V q

where:
Cair : concentration of chemical in the room air [mg/m3]
Ao : product amount [mg]
wf : weight fraction of the chemical in the total product [fraction]
V : room volume [m3]
q : ventilation rate of the room (number of air changes per time unit) [1/hr]
t : exposure duration [hr]

Subsequently, the amount inhaled can be calculated by :

Ainh =< Cair > ×Qinh × T

page 50 of 234 RIVM report 320003001

where:
Ainh : amount inhaled [mg]
Qinh : inhalation rate [m3/hr]

The parameter values needed for this calculation are:

Ao : Amount of the toy containing the volatile chemical. This is practically the weight of
the whole toy.

wf : Weight fraction of the chemical in the toy material. This depends entirely on the
material the toy consists of and no default values can be given

V : Room volume. Mean, standard deviation and 25th percentiles for room volumes in
Dutch homes have been given in the general fact sheet of ConsExpo:
Table 3-4 Room volumes in Dutch homes. Source: Bremmer et al. (2006)
Volume
Space Mean [m3] s.d. 25th percentile
living room 74 23 58
kitchen (incl. open kitchen)
bedroom 1 22 9.6 15
bedroom 2 35 11.2 27
bedroom 3 28 8.3 22
21 7.6 16
In the absence of information on which room is used, a default volume of 20 m3 is usually
used.

q : Ventilation rate of the room (number of air changes per time unit). Measurements of
ventilation rates in Dutch homes and abroad have been given in the general fact sheet of
ConsExpo (Bremmer et al., 2006). Default 25th percentiles for ventilation rates in Dutch
homes:
Table 3-5 Default 25th percentile room ventilation rates in Dutch homes. Source: Bremmer et al. (2006)
Room Ventilation rate [h-1] Q
the whole house 0.6 3
living room 0.5 3
kitchen 2.5 3
bedroom 1 3
bedroom (window open) 2.5 3
bathroom 2 3
toilet 2 3
shed 1.5 3
garage 1.5 3
default, if room is unspecified 0.6 3

t : Exposure duration is the time during which a child is exposed to the evaporated
chemical. The US EPA child-specific exposure factors handbook cites a study by Timmer et
al. (1985) in which the playing activity for five different age groups varying from 3 to
17 years of age children is reported to be between 14 an 267 minutes per day (US EPA,
RIVM report 320003001 page 51 of 234

2002). The Danish EPA quotes an American study which observed average play activity
times of 47 – 70 minutes for children 1 to 17 years of age, with 90 percentiles of 120 to 255
minutes (Danish EPA, 2005). The playing duration is likely to vary significantly per toy and
information on this is probably known best by the manufacturer of the toy. It has to be
emphasized that the exposure duration may be longer than the actual play time with the toy in
question, because the child may still reside in the room where the chemical has evaporated
after playing with the toy. The daily exposure time as defined by Babich et al. (2004) as the
time a child is awake and not eating can be used as a worst case value, which was
approximately 10 hours.

Qinh : Inhalation rate. The US EPA child-specific exposure factors handbook recommends
using an inhalation rate default value of 4.5 m3/day for children under one year of age (US-
EPA, 2002). For children aged 3-5 years, an inhalation rate default value of 8.3 m3/day is
recommended. We propose to take over these defaults.

Due to the wide variation in toys that may release a chemical via evaporation, a default for
frequency of exposure cannot be given.

NOTE : A more accurate exposure assessment to a chemical evaporating from a toy may be
achieved with the evaporation mode of the ‘exposure to vapour’ model in ConsExpo
(Delmaar et al., 2005).

3.5.5 Inhalation via dust or spray

The concentration of the chemical available for inhalation via dust or spray can be calculated
using the same formula as that used for the inhalation via evaporation scenario. However, not
all particles or droplets can be inhaled and reach the lower areas of the lungs (the alveolar
region). This depends on the size of the particles or droplets. To assess exposure via this
pathway, the particle size distribution of the spray or dust must be known. In the European
Norm EN 481, size fraction definitions for measurement of airborne particles have been
given (1994). Based on this norm, it is estimated that dust particles or spray droplets which
can be inhaled and reach the alveolar region will mostly have a diameter below 5 µm,
although particles with diameters up to 15 µm can still reach the alveolar region. Particles
with a diameter between 5 and 15 µm will mostly only reach the thoracic region and can be
taken in orally. Larger particles will mostly fall directly to the floor. Based on this, the
fraction of product released in the air with particles or droplets below 15 µm can be used for
the parameter ‘product amount’.
For dust generated by chalk crayons, a default can be derived from a study by Stopford
(Stopford, 2003)5. The respirable aerosol production (particles < 4 µm) generated during
chalk and pastel drawing activities for 30 minutes was found to be 364 ± 272 µg (mean ±

5
http://duketox.mc.duke.edu/cpscdust3.pdf
page 52 of 234 RIVM report 320003001

standard deviation), whereas total dust formation was 855 ± 590 µg. A rough estimate of
particles below 15 µm generated is 500 µg, which can be used as a default.
For spray, the particle size distribution depends strongly on the solvent, the propellant and the
nozzle size. We currently have no information on these factors for toy sprays and therefore,
and no default can be given.

NOTE: A more accurate exposure assessment to a chemical released in dust or spray may be
achieved with the spray model in ConsExpo (Delmaar et al., 2005).

3.5.6 Skin contact

The amount of a chemical on skin per skin area (Lderm) or per body weight (D) can be
calculated as follows:
L = A ×S ×F /S
derm o contact leach exp
And the external dose as:
D = A ×S ×F /W
o contact leach body
where

Fleach: the leachable fraction [fraction]

Ao: amount of product in contact with skin [kg]
Scontact: skin contact factor [fraction]
Sexp: the surface area of the exposed skin [m2]
Wbody : the body weight of the exposed person [kg]

The parameter values needed for this calculation are:

Fleach : leachable fraction, the amount of chemical that migrates to the skin per unit amount
of toy. This value can be determined with migration tests, which will be discussed further in
chapter 4.

A0 : amount of toy in contact with the skin. This is practically the weight of the toy.

Scontact : skin contact factor, used to account for the fact that the product is only partially in
contact with the skin. For example, for a costume this factor has been estimated to be 0.7,
since part of the costume will be on top of underwear and as such not in direct contact with
the skin (Bremmer and Van Veen, 2002). The skin contact factor varies too much with the
type of toy to be able to provide a default value.

Sexp : The surface area of the exposed skin. Mean, standard deviation and 25th percentile
default values for body surface area of Dutch children from 1.5 months to 17.5 years have
been given in the general fact sheet of ConsExpo (Bremmer et al., 2006):
RIVM report 320003001 page 53 of 234

Table 3-6 Mean, standard deviation and 25th percentile default values for body surface area of Dutch children
from 1.5 months to 17.5 years. Source: Bremmer et al. (2006).
Age Body surface
[m2]
Months Years mean SD 25th
percentile
1.5 0.283 0.020 0.270
4.5 0.364 0.026 0.346
7.5 0.419 0.031 0.398
10.5 0.459 0.033 0.437
13.5 0.490 0.035 0.467

1.5 0.520 0.062 0.480

2.5 0.616 0.062 0.575
3.5 0.690 0.076 0.640
4.5 0.762 0.081 0.709
6.5 0.902 0.093 0.841
9.5 1.13 0.13 1.05
12.5 1.40 0.15 1.31
13.5 1.51 0.16 1.40
16.5 1.75 0.16 1.65
17.5 1.79 0.18 1.67

In addition, default percentages of body surface for different body parts have been given:

Table 3-7 Default percentagess for body surface area of different body parts of Dutch children from 3 months to
14 years. Source: Bremmer et al. (2006).
Age Age surface body surface in %
Default value [m2] Head trunk arms hands legs feet
3 - 6 months 4.5 months 0.346 19.5 32.8 12.1 5.1 23.5 7.0
6 - 12 months 7.5 0.398 18.5 33.5 12.2 5.2 23.6 7.0
12 - 18 months 13.5 0.467 16.9 34.3 12.6 5.3 23.8 7.1

1.5 - 3 year 1.5 year 0.480 16.2 34.0 13.0 5.15 25.05 6.6
3 - 9 year 4.5 0.709 13.4 33.05 14.0 5.5 26.95 7.1
3 - 9 year 6.5 0.841 12.5 33.45 13.95 5.5 27.35 7.2
9 - 14 year 12.5 1.31 9.8 33.15 13.9 5.7 30.0 7.4

Wbody : Body weight of the exposed child, see above for direct ingestion.

With respect to exposure frequency, dermal contact with most toys can be expected to occur
on a daily basis.

3.5.7 Uptake
An essential part of the exposure assessment is formed by the uptake of a chemical by the
gastrointestinal tract, lungs or skin. The exposure assessment can be significantly refined if
data on the uptake of the chemical is available. A default value for uptake cannot be given,
page 54 of 234 RIVM report 320003001

as this is very chemical specific. It is common practice to use an uptake of 100% if no

information is available.
Exposure levels via different routes can only be added when uptake is included, i.e. when
looking at the internal, systemic dose.

3.5.8 Level of detail required for exposure assessments

As mentioned earlier, a detailed exposure assessment involving all exposure routes and
factors as described above is not always needed to demonstrate the safety of a toy. Apart
from excluding irrelevant exposure routes from the assessment, composition data of toy
material or migration data in combination with some general exposure factors such as
exposure frequency and bodyweight may frequently provide sufficient information to
demonstrate that exposure levels will not exceed the relevant health-based limit values. This
will be explained in detail in chapter 7. Nevertheless, for certain ad hoc situations, or for
specific toys or chemicals, a more refined exposure assessment may be desired. This can be
achieved by using the methodology presented here. In addition, for certain exposure
scenarios, further refinement can be achieved by using more refined migration testing
methods.

3.6 Conclusions

• We conclude that it is foreseeable that children under 3 years old will have access to toys
intended for children over 3 years old, unless these toys contain small parts or long cords,
because caregivers commonly know that such toys should be kept out of reach from
children displaying mouthing behaviour. This conclusion concurs with the conclusion of
the CSTEE.
• Six exposure scenario categories can be identified that may be relevant for toys: direct
ingestion, mouthing, inhalation via evaporation, inhalation via dust or spray, skin contact
and eye contact. It is anticipated that exposure to chemicals via eye contact will not lead
to effects of a serious nature and this contact scenario is therefore not considered further.
• The exposure scenarios relevant for a particular type of toy can be identified using the
exposure scenario decision tree.
• Exposure via the scenarios can be assessed by using adequate formulas and exposure
factor values.
• Exposure factor values are often highly uncertain and rough estimates are used until more
adequate information becomes available.
• Based on simple weighing experiments, the default for ingested amount of 8 mg of toy
material can be supported when the toy material is scraped off. For other toy materials
such as liquid and powder-like materials, other defaults need to be used.
• For all exposure routes, the exposure assessment can be significantly refined if data on
the uptake of the chemical is available.
RIVM report 320003001 page 55 of 234

3.7 Recommendations

• We propose that the exposure assessment of all toys which do not contain small parts or
long chords (or are otherwise dangerous from a physical-mechanical point of view), but
can be placed in the mouth or can be crawled on by children should include exposure
scenarios specific for young children, regardless of the intended age category of the toy.
• Many default values for exposure factors are highly uncertain and further research in this
area is warranted. More information is especially needed on:
o frequency and amounts ingested of toy material;
o mouthing durations for toys intended to be put in the mouth for children over
3 years of age;
o mouthing amounts and surfaces;
o playing durations for different types of toys;
o amounts of dust (and particle size distributions) generated by chalk, plaster and
other powder-like toys.
page 56 of 234 RIVM report 320003001
RIVM report 320003001 page 57 of 234

4 From toy to internal exposure – migration versus

bioavailability

4.1 Introduction
Not all the chemicals in a toy represent a hazard for the child’s health. Part of the chemicals
will remain in the toy even after mouthing the toy or swallowing (parts of) it. Therefore, in
guidance document EN 71-3 migration limits are set for 8 elements (Sb, As, Ba, Cd, Cr3+ and
Cr6+, Pb, Hg, and Se), which simulate the contact of toy material with stomach acid
(European Committee for Standardization (CEN), 1994). This acidic solution probably
represents a worst case scenario for elements. This is however not necessarily the case for
organic compounds. Legislation for most organic substances is laid down in guidance
documents EN 71-9, EN 71-10, and EN 71-11.

The aim of this chapter is twofold:

1. to evaluate the use of migration and bioavailability data in risk assessment for
substances in toys;
2. to evaluate these data more in particular for the 8 elements (Sb, As, Ba, Cd, Cr, Pb,
Hg, and Se) and some additional inorganic substances indicated in chapter 2 (Al, B,
Co, Cu, Mn, Sn, Ni, and Sr) in the ‘safety of toys’ Directive.

To that end the following issues will be addressed:

A. Oral bioavailability:
- definition and the various sub-processes that can be distinguished
- description of various migration and physiologically based extraction tests
- pros and cons of these tests
- applicability of migration and physiologically based extraction tests in risk assessment
of chemicals in toy matrices
- comparison to migration tests applied for food contact materials
B. Bioavailability after inhalation:
- experimental determination
- applicability in risk assessment of chemicals in toys
C. Dermal bioavailability:
- experimental determination
- applicability in risk assessment of chemicals in toys
page 58 of 234 RIVM report 320003001

4.2 Oral bioavailability

4.2.1 Definition of oral bioavailability

In Council Directive 88/378/EEG bioavailability is defined as ‘the soluble extract having
toxicological significance’ (European Committee for Standardization (CEN), 1988). In the
opinion of the Scientific Committee on Toxicity, Ecotoxicity and the Environment (CSTEE)
on ‘assessment of the bioavailability of certain elements in toys’ it is stated that this is not in
line with the general understanding of the term which is ‘the amount of each element in the
toy which could be absorbed into the systemic circulation of a child’ (Scientific Committee
on toxicity, 2004).

The term bioavailability is subject to various interpretations. Different points of view exist
partly depending on the scientific background of the investigator. For example, in human
nutritional sciences, the concept of bioavailability is regarded as the efficiency with which
nutrients are utilized (Schümann et al., 1994; Wienk et al., 1999). In pharmacology, oral
bioavailability characterizes the fraction of a dose that reaches the systemic circulation after
oral administration (Schümann et al., 1994; Chiou, 2001). Also different definitions of
bioavailability exist in fields such as ecotoxicology, et cetera, which fall outside the scope of
this report.

The pharmacology definition of bioavailability is considered to be the most appropriate

within the present context, i.e. the fraction of a substance present in toy material that reaches
the systemic circulation (of a child). This is a broadly applicable definition, whereas the
definition in nutritional sciences focuses on the nutritive value of feed and food. The CSTEE
has recommended using a slightly different definition than the pharmacology definition, as
the CSTEE definition is “the amount of each element in the toy which could be absorbed into
the systemic circulation of a child”, rather than the amount that reaches the systemic
circulation. Therefore, the definition by the CSTEE can be interpreted as bioaccessibility
(FB), a prerequisite step before a compound can become bioavailable, see section 4.2.2. The
bioaccessible fraction (FB) represents the maximum amount of contaminant potentially
available for transport across the intestinal epithelium, and can be investigated by the release
of the element from toy in conditions similar to conditions in the human gastrointestinal tract.

The link between the definition of bioavailability by Council Directive 88/378/EEG and the
definition by the CSTEE is that the systemic fraction of a toxic compound is in general a
better measure for toxicity than external exposure. In general this is true. However, for some
compounds the internal fraction is not a better measure for toxicity, for example for
compounds that locally exert toxicity, e.g. skin irritation by nickel.
RIVM report 320003001 page 59 of 234

The definition by Council Directive 88/378/EEG (‘the soluble extract having toxicological
significance’) is very broad and therefore difficult to translate to simple non-animal tests to
estimate the bioavailable fraction.

According to the pharmacology definition (‘the amount of each element in the toy that is
absorbed into the systemic circulation of a child’), bioavailability is best determined by
measuring the concentration of chemical in the blood of a human being or animal in time.
Usually, bioavailability is determined by comparison of the chemical concentration in blood
in time after intravenous administration, i.e. 100% bioavailable by definition, versus the
concentration in blood in time after the administration of interest. In section 4.2.2, the sub-
processes of bioavailability are addressed in order to find a starting point for estimating
bioavailability.

4.2.2 Sub-processes of oral bioavailability

According to the general interpretation in pharmacology, oral bioavailability is defined as the
fraction of an orally administered dose that reaches the systemic circulation. We have
conceptually subdivided oral bioavailability (F) into three major processes (Brandon et al.,
2006; Oomen et al., 2005; González-Soto et al., 2000; Danish EPA, 2005; Oomen et al.,
2004a; Babich et al., 2004; Versantvoort et al., 2004). Figure 4-1 describes these processes.
After ingestion, the contaminants may be partially or totally released from its matrix, a toy
in the present case, during digestion in the gastro-intestinal tract. The fraction of the
contaminant that is mobilized from the matrix into the digestive juice is defined as the
bioaccessible fraction (FB) and represents the maximum amount of contaminant potentially
available for transport across the intestinal epithelium.
FA represents the fraction of bioaccessible contaminant that is transported from the lumen
across the intestinal epithelium and into the portal vein or the lymph, thus representing the
absorption.
The contaminants may be metabolized in the intestinal epithelium or the liver, which is
referred to as the first-pass effect. The fraction of unmetabolised contaminant after passing
the liver (FH) will be transported throughout the body by the systemic circulation, and may
exert toxicity in organs and tissues. Consequently, the orally bioavailable fraction of a
contaminant is the resultant of the three steps: bioaccessibility, transport across the intestinal
epithelium, and the first-pass effect (see Figure 4-1 and equation 1):

F = FB × FA × FH (1)
page 60 of 234 RIVM report 320003001

Figure 4-1 Various steps of oral bioavailability (F) of a compound in toy matrix.

To our experience the matrix in which the contaminant is ingested, i.e. toy, food, water, soil
etcetera, is a determining factor in the fraction of the contaminant that becomes bioaccessible
(Oomen et al., 2006; Versantvoort et al., 2005; Brandon et al., 2006). Therefore,
bioaccessibility can be used to investigate the difference in bioavailability of a substance
from two different matrices (see section 4.3). It is possible that the matrix in which the
contaminant is ingested may affect the absorption of the contaminant. For an example and for
further information on this issue we refer to Oomen et al. (Oomen et al., 2006). Presently, we
assume that the matrix of ingestion only influences the sub-process of bioaccessibility.

The bioaccessible fraction, Fb, can be determined in vitro, by simulating the physicochemical
conditions of the human gastrointestinal tract. Several in vitro methods exist that aim to
determine the bioaccessible fraction of a contaminant (Oomen et al., 2002a). In most cases,
such tests have been applied to investigate the bioavailability of contaminants from soil
(exposure to contaminants via hand-to-mouth behaviour). Oomen et al. also developed a
physiologically based extraction test for contaminants in toys, making difference between
scenario’s for 1) mouthing, 2) swallowing, and 3) mouthing followed by swallowing.
RIVM report 320003001 page 61 of 234

4.3 Relative bioavailability in risk assessment

In order to correctly apply bioavailability of compounds in toy material in the risk assessment
of these compounds, one should use relative bioavailability. This means that the
bioavailability should be considered for both the exposure assessment as well as for the
toxicological information available. For example, the risk assessment of ingested material can
be refined by taking into account the fraction that migrates to the gastric juice and the
absorption over the intestinal wall. In this way, the internal (systemic) exposure is
determined. However, if we compare such an internal exposure with toxicological data, also a
correction for the bioavailability in the toxicological test should be used. For example, if the
toxic substance was provided in the food to rats, one should take into account the fraction that
was released from the food matrix in the stomach and the absorption over the intestinal wall.
If such corrections are not made, we implicitly assume that the bioavailability in the exposure
and the toxicological condition is similar.
An example for toys: when the bioavailability of a compound in the toxicological study
underlying the TDI is 60%, and the bioavailability of the same compound from a certain toy
material is 20%, the relative bioavailability is 20%/60% is 0.33.

The correct application of relative bioavailability in the present framework of chemicals in

toys is difficult. As indicated in section 4.2.2, bioavailability consists of several sub-
processes. One of these processes is bioaccessibility, which we aim to study by migration of
the chemical from toy into artificial saliva or gastric juice. Assuming that the difference in
bioavailability from the matrix used in the study underlying the TDI and the toy matrix can
be explained by the difference in bioaccessibility only, a relative bioaccessibility instead of a
relative bioavailability can be used. A relative bioaccessibility is preferred, as the outcome of
the migration test can be considered to be a measure of bioaccessibility. However, ideally
also the bioaccessibility of the matrix used in the study underlying the TDI should be known.
Sometimes information is available on the bioavailability of the compound of interest in the
study underlying the TDI, but information on the bioaccessibility is usually not available.
Therefore, if relative bioavailability is to be used correctly in the present framework, further
research is needed on the bioaccessibility of the compounds of interest from the matrix used
in the studies underlying the TDI is necessary. Also, attention should be paid to the
possibility that relative bioavailability cannot always be translated directly into relative
bioaccessibility (this assumes that absorption and metabolism of the compound are the same
in the study underlying the TDI and for the toy matrix).
An example of application of relative bioavailability and relative bioaccessibility in risk
assessment can be found for lead in soil (Oomen et al., 2006). In this report, the
bioavailability of lead from soil is used relative to the bioavailability of dietary lead, e.g. the
matrix in the studies underlying the TDI of lead.
page 62 of 234 RIVM report 320003001

If a methodology incorporates sufficient margins of safety, the issue of relative

bioavailability can be neglected. It can be argued that for most elements the presently
proposed methodology in chapter 8 is safe without correction for relative bioavailability. The
outcome of the migration test is expected to give a worst case value for bioaccessibility due
to the low pH value of the extraction medium. This assumption is based on several
observations that the bioaccessibility of elements is much higher in the stomach compartment
compared to the intestinal compartment of a physiologically based in vitro digestion model
(Oomen et al., 2003c; Oomen et al., 2004b; Oomen et al., 2002b). The acid environment of
the stomach compartment is considered to be similar to the extraction according EN 71-3,
whereas the extraction in the intestinal compartment is probably a better measure for
bioavailability as absorption of elements occurs in the intestine. Because direct in vivo data
that verify the statement that the methodology incorporates sufficient margins of safety are
lacking, further research on this issue is recommended.

4.4 Tests to estimate the orally bioavailable fraction of a

contaminant from toy

Bioavailability is presently defined as the fraction of an orally administered dose that reaches
the systemic circulation, see section 4.2.1. Bioavailability refers to a physiological process.
Therefore, true bioavailability of a compound from toy can only be tested in humans or
animals, for example by determination of the blood concentration in time after oral and
intravenous application. For simplicity and in order to avoid animal testing, a few methods
have been developed to estimate the bioavailability or part of the process determining the
bioavailability in vitro, i.e. in the laboratory. The different migration and physiologically
based extractions tests are listed in Appendix V.

4.4.1 Tests for inorganic compounds

Policy on the safety of toys in European Member States has been laid down in Council
Directive 88/378/EEC. In this Directive requirements for the total bioavailable amount are
listed for several elements, see Table 4-2. The bioavailable amount was used in the Directive
as it was considered that the bioavailable fraction of a substance in toys is more important
than the total content of potentially dangerous substances, i.e. internal exposure is considered
to be more predictive for toxicity than the dose (Commission of the European Communities,
1985). In this sense, the intention of the use of bioavailability by Council Directive
88/378/EEC is in line with the pharmacological definition of bioavailability, see
section 4.2.1. In the European Standard 71-3 these bioavailability requirements are translated
to limits of migration, also listed in Table 4-1 (European Committee for Standardization
(CEN), 1994). For the translation of allowed bioavailability to limits of migration a daily
intake of 8 mg toy is assumed. In addition, adjustments were made to minimize the exposure
of children to toxic elements by lowering the migration limit for barium and selenium, and to
ensure analytical feasibility by increasing the migration limit for antimony, arsenic, and
RIVM report 320003001 page 63 of 234

chromium (Danish EPA, 1998). For comparison, the maximum bioavailable concentrations in
toy materials based on 8 mg of ingested toy material would be 25 mg Sb/kg, 12.5 mg As/kg,
3125 mg Ba/kg, 75 mg Cd/kg, 37.5 mg Cr/kg, 87.5 mg Pb/kg, 62.5 mg Hg/kg, and 625 mg
Se/kg (Danish EPA, 1998), whereas the migration limits listed in Table 4-2 are proposed by
EN 71-3 (European Committee for Standardization (CEN), 1994).

The migration of elements from toy is according EN 71-3 assessed with chemical extraction
tests (European Committee for Standardization (CEN), 1994). Generally, a 0.07 M
hydrochloric acid solution of 50 times the mass of the test portion (and preferentially a
sample mass of 100 mg or more, with exceptions) is used. The pH is adjusted in the presence
of toy to pH 1.0 to 1.5, and the chemical is extracted from the toy matrix during 1 h with
agitation and 1 h without agitation, see Appendix V for details on the migration tests.
This extraction medium, hydrochloric acid solution, is considered to simulate gastric juice
(Commission of the European Communities, 1985). For the elements listed in EN 71-3
simulated gastric juice was used as it was argued that this is a more stringent extraction
medium than saliva, providing an additional margin of safety in the evaluation of possible
intake of these compounds by children (Commission of the European Communities, 1985).

Some non-EU countries use different migration tests. For example, Canada has its own
legislation and test methods, see Appendix V, which uses different extraction solutions which
could lead to other outcomes for the migration. However, the majority of the non-EU
countries also use the CEN 71-3 migration test for the 8 different elements.

Beside the chemical extraction tests, there are several physiologically-based extraction tests
to simulate mouthing in order to minimise over- or underestimation of
migration/bioaccessibility of the in vivo situation. Most of these tests were developed and
used to determine the release of organic compounds, e.g. phthalates, from toys and other
consumer products. Only Iliano et al. (1988) and RIVM (Oomen et al., 2005; Oomen et al.,
2004a; Oomen et al., 2003b) have actually looked at the release of elements from toys using
physiologically-based extraction tests (see Appendix V).
Table 4-1 Requirements based on total bioavailable amount of element resulting from the use of toys per day according to Council Directive 88/378/EEC, and migration
limits according to EN 7 1-3.
Limit of migration for modelling clay and finger paint
Limit of migration from toy material (EN 71-3) (mg/kg) 1
Limit of total bioavailable (EN 71-3) (mg/kg) 2
Element amount per day Without Analytical After Without Analytical After
(88/378/EEC) (μg) analytical correction factor analytical analytical correction factor analytical
correction (in %) correction3 correction (in %) correction3
Antimony
0.2 60 60 150 60 60 150
(Sb)
Arsenic
0.1 25 60 63 25 60 63
(As)
Barium
25.0 1000 30 1429 250 30 357
(Ba)
Cadmium
0.6 75 30 107 50 30 71
(Cd)
Chromium
0.3 60 30 86 25 30 36
(Cr)
Lead
0.7 90 30 129 90 30 129
(Pb)
Mercury
0.5 60 50 120 25 50 50
(Hg)
Selenium
5.0 500 60 1250 500 60 1250
(Se)
1
Limits of migration for any toy material detailed in EN 71-3, except for modelling clay and finger paint. Due to the precision of the analytical methods the result of a
migration test is corrected. The analytical correction to which the result of the migration test should be subjected is listed. In the next column the corresponding limits of
migration have been accounted for the analytical correction, i.e. if less than the latter amount migrates out of the EN 71-3 tests the toy complies with the requirements.
2
Limits of migration for modelling clay and finger paint. In analogy with the limits for other toy materials the analytical correction and the migration limits accounted for
analytical correction are listed, see also 1).
3
In EN 71-3, the analytical correction is calculated from the value of the measured migration. For example, if the analytical result of lead is 120 mg/kg, an analytical
correction of 30% is applied. The adjusted analytical result is:
120 × 30
120 − = 120 − 36 = 84
100

This adjusted analytical result is below the requirements of 90 mg/kg. In the present table the analytical requirements are used to calculate the migration limit that is
allowed after analytical correction.
 RIVM report 320003001 page 65 of 234

Assumptions supporting EN 71-3

• The present methodology of EN 71-3 to determine the bioavailable amount of an
element from toy is probably an overestimation of the actual bioavailable amount
after ingestion of toy matrix. For, extraction in an acid extraction medium simulating
gastric juice is performed. Absorption of compounds takes place in the intestine, with
an environment of higher pH (pH 5.0-7.5). The bioavailability in the intestinal phase
can be considered to be lower for these elements than in the stomach environment due
to their dependence on the pH (Oomen et al., 2004a; Oomen et al., 2003b).
• Another important aspect of EN 71-3 is that always ingestion of toy material is
considered, assuming that this also is protective for mouthing the toy (Commission of
the European Communities, 1985). Based on the physicochemical nature of saliva and
gastric juice this is true.
• It is not considered that a large surface may be mouthed on, whereas only 8 mg of toy
is considered to be ingested. Therefore, the bioavailable amount of a substance from
toy during mouthing may be greater than after ingestion of 8 mg of toy. For elements,
usually a small amount migrates into artificial saliva, whereas much more migrates in
the stomach and intestinal compartment due to the low pH environment of the
stomach (Oomen et al., 2003a). Therefore, it can be anticipated that for elements
migration determined by the methodology of EN 71-3 can be used as a worst case
value for both ingestion and mouthing, although additional research on this issue is
recommended for verification. However, for other compounds sucking may give
higher bioavailability amounts of a substance than ingestion.

4.4.2 Tests for organic compounds

EN 71-9 provides requirements for certain organic chemical compounds in toys and toy
materials. Migration limits are derived for some compounds and absolute limits for others.
EN 71-10 provides information on the sample preparation and the extraction procedure for
these organic compounds to determine the migration. To that end, migration is determined of
a sample with a surface of less than 10 cm2 with 100 ml of deionized water as extraction
medium. The extraction bottle with water and sample is rotated end-over-end for
60 ± 5 minutes at 60 ± 5 rotation per minute at 20 ± 2 ºC.
In addition, various migration and extraction tests exist to assess the release of various
organic compounds such as phthalates and nitrosamines from toy articles. Mostly water
(migration test) or saliva (physiologically-based extraction test) are used to determine the
release after mouthing on a toy by a child.
Most of the physiologically-based extraction tests were developed and used to determine the
release of phthalates from toys and other consumer products and not for other organic
compounds. Examples are the Joint Research Centre (JRC) model and a model developed by
the U.S. Consumer Product Safety Commission (Simoneau et al., 2001; U.S. Consumer
Product Safety Commission,).
page 66 of 234 RIVM report 320003001

Except for the RIVM method, no migration or physiologically based extraction tests have
been found in literature describing the release of organic compounds from toys using other
extraction fluids representative for ingestion of the compound, e.g. stomach and/or intestinal
simulant. The RIVM method (see Appendix V) is based on human physiology and is applied
independently of the matrix of contaminant. The research with the in vitro digestion models
by RIVM has shown that the amount extracted in the acid environment of the stomach does
not represent a worst case situation for the bioavailable amount of an organic substance
(Oomen, 2000). For, most organic compounds are not as susceptible for the low pH
environment of the stomach as the elements considered in EN 71-3. Furthermore, the
research by RIVM has shown that for many substances the release from a matrix in the
intestine is highest when fed conditions are simulated in the in vitro digestion model (Oomen,
2000). The complexing capacities of the extraction juices of an in vitro digestion model are
higher when fed conditions are simulated as food constituents are present, and more
complexing agents such as bile and enzymes are present in digestive juices secreted during
fed conditions. Therefore, the methodology of EN 71-3 to determine the bioavailable amount
of elements is suitable as a worst case bioavailable amount for elements, whereas it is not
applicable for organic compounds.

Both water and saliva have been used as extraction fluid in tests simulating mouthing on toy
matrices. The experimental tests that aim to simulate the migration of compounds from toy
matrix in saliva vary in the degree to which saliva is simulated. However, as the composition
of saliva is not very aggressive or very different from water, the outcome is usually within the
same order of magnitude. In a technical report of CEN TC 252/WG 9/TG 2 it is concluded
that water is therefore the most proper simulant for saliva. In a report by RIVM, it was shown
that slight differences in migration rate can be observed when saliva simulant is compared to
water (Oomen et al., 2004a).
The migration of phthalate into saliva simulant from PVC disk resulted in a slightly better
extraction in saliva than in water (Oomen et al., 2004a). It is conceivable that water is an easy
and reproducible extraction medium to work with, leading to only slight differences with
artificial saliva. It can therefore be used to assess the migration from toys during mouthing.

4.4.3 Recommendations for application of tests simulating

ingestion and mouthing of toy matrix in risk assessment
The tests listed in Table 4-2 and recommended for application in risk assessment of
substances in toys. Water can be used as an extraction medium to simulate mouthing for both
elements and organic substances. Migration tests according EN 71-3 can be used to simulate
ingestion of substances for elements. However, tests to simulate ingestion for organic
substances are not yet available. EN 71-3 is not applicable for organic substances.
RIVM report 320003001 page 67 of 234

Table 4-2 Recommendations for testing mouthing and ingestion of substances in toy matrices

Elements Organic substances

Mouthing Migration into water, similar Migration into water

to organic substances according to EN 71-9 and EN
71-10

Ingestion Migration tests according to Migration tests according to

EN 71-3 EN 71-3 are not applicable for
organic substances.
Possibility to test migration
from toy after ingestion with Migration tests according to
more complex but more RIVM methodology are
physiological tests (not possible, but analytical
applicable for all elements) validation is lacking (inter-
laboratory testing). Relative
Limits of migration to be to EN 71-3 is this a
reconsidered complicated method. Other
tests to assess the bioavailable
amount of organic substances
from toys after ingestion are
lacking.

In conclusion it can be stated that the migration tests in EN 71-3 can be used as a safe
method for estimating the migration of elements from a toy matrix following mouthing and
ingestion. An additional plus-point is that the method has been used for a long time and is
a known and well-accepted method.

If the migration test according to EN 71-3 indicates a risk, e.g. the migration exceeds the
migration limit value, there is an option to refine the migration test. This can be done by
linking up the migration test conditions more closely to physiology in the gastrointestinal
tract. In addition, additional research on the relative bioavailability of the substance, i.e. the
bioavailability of the substance from toy relative to the bioavailability from the matrix used in
the studies underlying the TDI.

We recommend to express migration limits of substances from toys in mg/kg toy material.
This way of expressing links the allowable migration directly to a toxicologically derived
limit value, e.g. a TDI. This holds for migration into water and migration according to
EN 71-3. Thus, the expression of migration limits in mg/kg toy material is applicable for the
limits employed in option 1 as further detailed in section 7.3.3.1. Also when more
page 68 of 234 RIVM report 320003001

physiologically based tests are used, as can be the case in option 3 in section 7.3.3.3, we also
recommend to express the limits as mg/kg toy material.

4.4.4 Discussion points

4.4.4.1 Mouthing
Mouthing by children on toy articles can be best simulated by migration tests, as the mouth
time varies between various toy products. However, one should consider that a child
sometimes not only mouths once on a toy article, but multiple times. For examples, textile
cords in hoods of sweaters or other clothing materials can be mouthed on many times. The
same holds for cuddly toys. In principle, such toys required multiple extractions in migration
tests, as probably in time less compound will be released during mouthing. Therefore, a worst
case scenario can be assumed in which case the same amount of compound is released
multiple times based on a single migration experiment. Depending on the number of assumed
mouthing events, this may lead to the assumption that all compound is released in time.
Alternatively, the migration test can be performed multiple times to investigate the
dependence of migration on the number of mouthing events.

In the same manner, the release of a compound after mechanical washing with washing
detergent can be investigated. This is only relevant for clothing or stuffed toys, because wood
etc will not be washed. This is probably not of use for regulatory issues, as washing is not
required before use of the toy, but may be relevant for example in a case study when a
realistic exposure assessment of a certain toy is necessary.

A factor that may affect the release of a compound from its matrix is whether a child chews
on the toy matrix or only mouths on it. It is very difficult to simulate chewing in a migration
test. Some attempts have been done by adding glass marbles to a tube containing the
extraction fluid and toy matrix, and mixing the contents by rotating the tube head-over-heels
or horizontal shaking (Fiala et al., 2000; Oomen et al., 2004b; Steiner et al., 1998). In this
manner, the glass marbles may fall on the toy matrix in an attempt to simulate chewing, i.e.
simulate that part of chewing that represents the increment of accessible surface. These
experiments did not lead to large differences in the migration between the presence and
absence of marbles (Fiala et al., 2000; Oomen et al., 2004b). Volunteer studies suggest that
there is a difference in migration due to chewing. In the volunteer study also an outlier was
observed (very high migration). A hypothesis is that by chewing small pieces of the PVC
standard discs were chewed off and were completely extracted. In can therefore be concluded
that the in vitro and in vivo studies give no clear picture on the effect of chewing on the
migration of substances in the mouth.
The present report differentiates in a mouthing and an ingestion scenario for oral exposure to
toy material. When regarding mouthing, mouthing by children on a large toy surface is
considered. Obviously, children may also chew on the toy. Chewing will lead to 1) increment
RIVM report 320003001 page 69 of 234

of the surface area and 2) swallowing of small pieces. Hence, it can be assumed that in case it
is plausible that children can chew the toy material in question into smaller pieces, it is also
plausible that these small pieces will be ingested. Considering the physicochemical
conditions in the mouth and in the remainder of the human gastrointestinal tract, swallowing
of small pieces of toy material will most probable result in higher migration of substances
than after sucking on the toy material. We therefore recommend to focus on migration
simulating ingestion in case it is plausible that a child can chew the toy material into smaller
pieces. When this is not the case, migration tests that focus on mouthing of large surfaces
such as EN 71-10 will suffice, e.g. in EN 71-3 a disk with a surface area of 10 cm2 is used.
At present there is no standard to estimate the migration of organic substances after ingestion.
Therefore, development of a test that is representative for the migration of organic substances
after ingestion is highly recommended. Such a test should simulate the conditions in both the
human stomach and intestine, and should consider differences in physicochemical conditions
resulting from fed and fasting conditions.

Another issue with mouthing is that some very minute pieces of the toy matrix may be
released from the pieces of toy introduced into the test system, and which remain in the
extraction fluid even after the separation step. An example may be that small fibres are
released from textile toy items, which remain in the extraction fluid even after centrifugation.
Also in the real life situation textile fibres may be release from a toy during mouthing. In real
life, these textile fibres will probably be ingested. An option is thus to include these minute
pieces in the bioaccessible fraction. This is a worst case assumption. Note that in such a case
a clear relationship between the bioaccessible amount and time will probably not be
observed. Also note the potential experimental difficulties with the separation step. First the
normal toy matrix should be separated from the extraction fluid by centrifugation, and then
either the entire extraction fluid should be sampled or the extraction fluid should be filtered
and part of the filtrate and the residue should be sampled. Otherwise, erratic results are
possible.

4.4.4.2 Ingestion
Whether a person is in the fasted (after several hours of not eating and drinking) or fed state
(some time after food intake), greatly affects the physicochemical conditions in the
gastrointestinal tract. In the fed state, more digestive juices are secreted into the
gastrointestinal tract with a higher concentration of digestive enzymes, salts, and bile. Also
food remnants may be present in the fed state. Therefore, in general, more complexing agents
are present in the fed state which may facilitate extraction of a substance from a matrix like
toy. Therefore, for most substances, release will be highest in (artificial) digestive fluids
simulating fed conditions.
On the other hand, for pH-sensitive compounds like most elements, the fasted state will lead
to the greatest release of the element from toy. In the fasted state, the pH in the stomach
compartment is low, sometimes as low as 1, leading to the highest release of these
substances. Therefore, the hydrochloric acid solution employed in EN 71-3 with a pH
page 70 of 234 RIVM report 320003001

between 1 and 1.5 can be assumed to give a worst case migration of elements from toy.
However, note that this does not hold for other compounds.

The amount of matrix per volume of extraction medium may cause differences in
bioaccessibility. This has been shown by Oomen et al. (Oomen et al., 2004). For practical
reasons it is not possible to routinely determine the migration of substances from toy matrix
for several amounts of toy. In addition, in routine research on the migration of substances
from toy some conservative aspects are included. However, when a detailed risk assessment
is performed (option 3 in chapter 8 ‘use of risk based data’), the migration of various amounts
of toy per volume of extraction medium should be investigated.

4.5 Dermal bioavailability

For dermal bioavailability a comparable approach of defining the process is suggested

as for oral bioavailability. This means that dermal bioavailability can be defined as ‘the
fraction of the dose that reaches the systemic circulation following dermal contact’.
Also in dermal bioavailability three processes can be distinguished, i.e.:
a. release of a substance from a toy matrix due to dermal contact
b. penetration of the substance in the skin
c. transport of the substance across the skin into the systemic circulation

Table 4-3 describes the penetration and absorption potential of the elements listed in
chapter 2. In Table 2-2, chapter 2, the accompanying toxicological effects are described.
Information on dermal bioavailability of metals was obtained from the publication of
Hostýnek et al. (1993).
RIVM report 320003001 page 71 of 234

Table 4-3 Potential of a number of elements for dermal bioavailability Hostýnek et al. (1993).
Element Skin penetration Transport across skin
Aluminium Generally poor. By shunt diffusion through Unknown
appendages and ductal closure, leading to sweat
inhibition
Antimony Sb 2 O3 through sweat follicles
Arsenic poor poor
Boron poor poor
Cadmium fair poor
Chromium Cr3+: poor Cr3+: poor
Cr6+: good Cr6+: good/ can be reduced to
Cr3+ during passage through the
skin
Cobalt poor poor
Copper Oxidised by sweat → organometallic salt: good poor
Lead Poor Inorganic lead forming ligands
in proteins: poor
Lipid-soluble organo-lead:
good
Manganese Poor poor
Permanganate anion: good
Mercury Depending on form of metal:
fair to good
Nickel good poor
Selenium unknown unknown
Silver good poor
Strontium Poor poor
Tin unknown unknown
Organo-tin good unknown

4.5.1 Conclusions
Only in case of hexavalent chromium, permanganate ion, and mercury contamination of
toys dermal bioavailability should be taken into account (see also chapter 2). Dermal
absorption can be tested by using human skin (from plastic surgery or cadavers) or pig
skin (closely resembles the human skin). A detailed description of these experiments
can be found in literature (for example Copovi et al., 2006; Panigrahi et al., 2005; Heard
et al., 2006; Cazares-Delgadillo et al., 2005); they are beyond the scope of this report.
page 72 of 234 RIVM report 320003001

4.6 Inhalatory bioavailability

In the present methodology, it is assumed that inhalatory bioavailability is 100% in
cases it will be relevant for toys (see chapter 3). Refinements are possible, but fall
beyond the scope of this report.

4.7 Ocular bioavailability

Ocular bioavailability is not assumed to be of any relevance for exposure to elements

via toys (see chapter 3).

4.8 Conclusions and recommendations

• The pharmacology definition of bioavailability is considered to be the most correct in

the context of toy safety, i.e. the fraction of a substance present in toy material that
reaches the systemic circulation (of a child). For testing on the oral bioavailability of
substances from toy the definition according to the CSTEE can be used the amount of
each element in the toy which could be absorbed into the systemic circulation of a
child. The definition by the CSTEE can be interpreted as bioaccessibility (FB), a
prerequisite step before a compound can become bioavailable.
• (Oral) bioavailability (F) is conceptually seen as the resultant of three major processes,
i.e. 1) release of the compound from its matrix being the bioaccessible fraction (Fb),
2) fraction being absorbed across the intestinal wall (Fa) and 3) first-pass metabolism
(Fh).
• Oral bioavailability of a compound from different matrices is assumed to be driven by
differences in bioaccessibility.
• Bioaccessibility can be determined experimentally by means of migration tests or by
physiologically based tests.
• We recommend to express migration limits in mg/kg toy material. This way of
expressing links the allowable migration of a substance directly to a toxicologically
derived limit value, e.g. a TDI. This holds for migration into water (i.e. saliva
simulant), migration according to EN 71-3, and for migration limits for more
physiologically based in vitro digestion tests.
• Migration testing according to EN 71-3, i.e. migration into hydrochloric acid solution,
is a suitable test to simply assess the bioaccessible fraction of elements from toys after
ingestion. Also for mouthing migration testing according to EN 71-3 can be performed
as it gives a worst case bioavailable amount.
• To refine the exposure assessment, water or artificial saliva would suffice as simulant
for investigating migration of elements following mouthing. To that end, the
RIVM report 320003001 page 73 of 234

methodology of EN 71-10 that is used to study the migration of organic substances

during sucking, can be applied for elements.
• Chewing on toy material will lead to 1) increment of the surface area and 2) swallowing
of small pieces. Hence, it can be assumed that in case it is plausible that children can
chew the toy material in question into smaller pieces, it is also plausible that these small
pieces will be ingested. It can also be assumed that migration of substances from toy
material is higher after ingestion than after mouthing and possibly chewing in the
mouth. Therefore, we recommend to focus on ingestion in case it is plausible that a
child can chew the toy material into smaller pieces. When this is not the case, migration
tests that focus on mouthing of large surfaces such as EN 71-10 will suffice, e.g. in EN
71-3 a disk with a surface area of 10 cm2 is used.
• Migration testing according to EN 71-3, i.e. migration into hydrochloric acid solution,
is not suitable to estimate the bioaccessibility of organic substances from toy after
ingestion of toy matrix. More sophisticated extraction media, simulating intestinal
solution (for fed conditions), are required to assess the migration of organic compounds
from toys. Development of a test to estimate the migration of organic substances after
ingestion is highly recommended.
• Migration testing in water or artificial saliva would suffice as simulant for investigating
the migration of organic compounds following mouthing, i.e. EN 71-10.
• The availability of validated physiologically based tests would have additional value in
relation to option 3 of the general methodology proposed in chapter 7 (and 8). Further
validation of such type of tests is recommended.
• In order to account for bioavailability of chemicals in the risk assessment of toys, the
bioavailability of a specific chemical from the toy matrix should be compared to the
bioavailability in the studies underlying the health-based limit values. However, as the
methodology for the bioavailability of elements from toys according to EN 71-3
incorporates sufficient margins of safety, the issue of relative bioavailability can be
neglected. When the bioavailability of elements from toys is investigated in a more
sophisticated manner, attention should be paid to the derivation of relative
bioavailability.
• Further research is recommended to investigate whether EN 71-3 gives conservative
estimates of oral bioavailability with respect to bioavailability underlying the TDI
(relative bioavailability). It is assumed that EN71-3 gives a conservative estimate for
oral bioavailability but we think it is questionable whether the same is valid for relative
bioavailability.
page 74 of 234 RIVM report 320003001
RIVM report 320003001 page 75 of 234

5 Food contact material

5.1 Introduction
Within the framework of the Food Contact Materials (FCM) a large number of migration
limits for chemical substances used in the production of Food Contact Materials has been
authorized. FCM are all materials and articles intended to come into contact with foodstuffs,
including packaging materials but also cutlery, dishes, processing machines, containers etc.
Beside that, the term includes materials and articles which are in contact with water intended
for human consumption excluding fixed public or private water supply equipment.
Under the specific conditions of use, chemicals used in FCM are considered to be safe. Toys
can be made from the same materials. Both FCM and toy legislation mostly focus on oral
exposure. If one considers mouthing a toy (so a material in contact with a matrix, i.e. saliva),
there are parallels to food contact material in contact with a matrix i.e. food or liquid.
Since a framework already exists for deriving safe levels for chemicals in FCM it was asked
by DG Enterprise to examine whether the used methodology can also apply for toys. If we
can come up with some kind of extrapolation from FCM limits to limits for toys, this would
provide a very efficient procedure as it would not require additional testing. For both type of
consumer products a safe level of contaminants is aimed at. It makes sense to the public that
the same limit values apply (if a certain limit level is safe for food contact materials than it
should also be safe for toys, or the other way around). For producers of (raw/finished)
materials it would be convenient if one test could provide safe levels for both types of
legislation. It should be noted that not all substances embraced by FCM legislation are also
relevant for toys. This should be considered on a substance-specific basis.

5.2 EU Directives
The Framework Regulation (EC) 1935/2004 (L338/4) states that food contact materials shall
be safe. They shall not transfer their components into the food in quantities that could
endanger human health, change the composition of the food in an unacceptable way or
deteriorate the taste and odor of foodstuffs.

The evaluation of chemicals used in Food Contact Materials as practiced by the EU EFSA
and formerly by the EU SCF, uses a structured default approach, see Figure 5-1. In the
petition dossier, data on migration of the particular chemical from the FCM matrix into
foodstuffs are required. Usually this involves migration to a suitable food simulant.
Depending on the level of migration toxicological information is needed. When migration is
high (5 - 60 mg/kg/food), an extensive data set is needed. When migration is between
0.05 – 5 mg/kg food, a reduced data set may suffice. In case of low migration
page 76 of 234 RIVM report 320003001

(< 0.05 mg/kg food) only a limited data set is needed. The evaluation of all data supplied in
the petitioned dossier, leads to classification of the petitioned compound on one of several
lists which specify the restrictions the EU committee deems necessary. Compounds placed on
lists 0, 1, 2, 3 or 4 are admitted for use and receive a Standard Migration Limit (SML)
expressed as mg/kg food (simulant). An overview of various types of limit values for food
contact materials is given in Figure 5-2. For compounds for which a Tolerable Daily Intake
(TDI) was derived the SML is derived by multiplying the TDI with a factor of 60. This is
based on the notion that a person of 60 kg could ingest up to 1 kg daily of foodstuffs in
contact with packaging material. For substances where no TDI is established (reduced
toxicity dossiers) fixed migration limits (restrictions) of 0.05 mg/kg food or 5 mg/kg food are
allocated depending on the level of migration measured. If migration is lower than
0.05 mg/kg food, than the fixed migration limit becomes 0.05 mg/kg food. If the migration is
between 0.05 and 5 mg/kg food, than the fixed migration limit becomes 5 mg/kg food. A
general requirement is the overall migration limit. For plastic food contact materials the
overall migration of all substances may not exceed 60 mg/kg food (simulant).

Evaluation of substance used in food: EU ESFA

High migration Intermediate migration Low migration

5 - 60 mg/kg food 0.05 - 5 mg/kg food ≤ 0.05 mg/kg food

Extensive data set Reduced data set Limited data set

petitioned dossier petitioned dossier petitioned dossier

Classification of substance
0, 1, 2, 3 or 4

Figure 5-1 Flow chart of the evaluation of substances in food contact material according to regulations of the
EU EFSA.
RIVM report 320003001 page 77 of 234

For some compounds the ‘QM’ or ‘QMA’ is given, which is the maximum permitted residual
quantity of the substance in the finished material or article expressed as mg per kg food
contact material (QM) or as mg per 6 dm2 of the surface (of that material or article) in contact
with foodstuffs (QMA).
For some compounds several limits are provided, e.g. a SML and a QM(A).

Limit values for Food Contact Materials

SML: Specific Migration Limit

(in mg/kg food)

SML = TDI x 60 (mg/kg food)

QM: Maximum permitted QMA: Maximum permitted

Yes residual Quantity per Area
residual Quantity
(mg/kg packaging material) (mg/ 6 dm2 packaging material)
Tolerable Daily Intake (TDI) established?

No
SML: fixed limits
0.05 or 5 mg/kg food

General requirement: overall SML may not exceed 60 mg/kg food

Figure 5-2 Overview of various types of limit values for food contact materials.

The Regulation distinguishes 17 groups of materials and articles which may be covered by
specific measures:
Active and intelligent materials and articles
Adhesives
Ceramics
Cork
Rubbers
Glass
Ion-exchange resins
Metals and alloys
Paper and board
Plastics
Printing inks
Regenerated cellulose
Silicones
Textiles
Varnishes and coatings
Waxes
Wood
 RIVM report 320003001 page 78 of 234

Up to now EU-wide specific measures exist for ceramics, regenerated cellulose and plastics.
These measures are addressed in specific directives:

• Ceramics are regulated by Council Directive 84/500/EEC as amended by Directive

2005/31/EC. The Directive sets migration limits for cadmium and lead which might
be released from decoration and/or glazing. It gives an analytical method for the
determination of the migration of these substances.
• Regenerated cellulose film is regulated by Commission Directive 93/10/EEC as
amended by Directive 93/111/EC. The Directive sets a positive list of authorized
substances and the conditions under which they can be used. A recent amendment
Commission Directive 2004/14/EC introduces changes for plastic coated regenerated
celluloses film.
• Plastics are regulated by the new Commission Directive 2002/72/EC which
consolidates Commission Directive 90/128/EEC and its seven amendments
(Directives 92/39/EEC, 93/9/EEC, 95/3/EEC, 96/11/EEC, 1999/91/EC, 2001/62/EC
and 2002/17/EC). These amendments mainly modified the lists of authorized
substances such as monomers and additives.

Three groups of substances are regulated individually in specific directives, i.e. vinyl chloride
monomer in plastics, nitrosamines in rubber teats and soothers and BADGE (bisphenol-A-
diglycidyl ether), BFDGE (bisphenol-F-diglycidyl ether) and NOGE (Novolac glycidyl ether)
in plastics and coatings.
• Vinyl Chloride Monomer (VCM) in food contact materials and articles is regulated by
Council Directive 78/142/EEC. To ensure a safe product, the residual content of VCM
in the finished material or article is limited to 1 mg/kg. Furthermore, VCM should not
be detectable in foodstuffs. Commission Directives 80/766/EEC and 81/432/EEC give
methods of analysis for VCM in the finished product and in foodstuffs.
• Nitrosamines in rubber teats and soothers are regulated by Commission Directive
93/11/EEC, which establishes specific migration limits for these substances and their
derivatives.
• BADGE (bisphenol-A-diglycidyl ether), BFDGE (bisphenol-F-diglycidyl ether) &
NOGE (Novolac glycidyl ether) in plastics, coatings and adhesives BADGE, BFDGE
& NOGE are regulated by Commission Regulation (EC) 1895/2005. For BADGE and
its partially hydrolyzed products, specific migration limit have been set at 9 mg/kg.
For the chlorohydrins of BADGE the limit has been set at 1 mg/kg. Moreover, the
Regulation prohibits the use BFDGE and NOGE as from 1st January 2005.

5.3 Migration tests food contact material

To enforce overall and special migration limits, special Directives set out procedures for
compliance testing. Basic rules for migration tests such as the conditions of contact (time,
RIVM report 320003001 page 79 of 234

temperature, food simulants) are supplied in Council Directive 82/711/EEC and its
amendments 93/8/EEC and 97/48/EC, while Council Directive 85/572/EEC gives a list of
food simulants to be used in migration tests for the various types of foodstuffs.

According to Directive 82/711/EEC and 90/128/EEC, the determination of the migration of

specified components in foodstuff instead of the use of simulants is permitted. The following
food simulants listed in Table 5-1 should be used for migration test with food contact
materials.

Table 5-1 Description of food simulants to be used to test the migration of substances from food contact
materials.
Food type Conventional Food simulants Abbreviation
classification
Aqueous foods (i.e. Foodstuffs for Distilled water or Simulant A
aqueous foods which tests with water equivalent
having a pH > 4.5) stimulant A only is quality
prescribed in
Directive
85/572/EEC
Acidic foods (i.e. Foodstuffs for Acetic acid 3% Simulant B
aqueous foods which tests with (w/v)
having a pH < 4.5) stimulant B only is
prescribed in
Directive
85/572/EEC
Alcoholic foods Foodstuffs for Ethanol 10% (v/v) Simulant C
which tests with This concentration
stimulant C only is shall be adjusted to
prescribed in the actual alcoholic
Directive strength of the food
85/572/EEC if it exceeds 10%
(v/v?
Fatty foods Foodstuffs for Rectified olive oil Simulant D
which tests with or other fatty food
stimulant D is only simulants
prescribed in
Directive
85/572/EEC

The design of a migration test is dependent on a) the type of food in relation to the packaging
material to be tested, b) contact time between food and packaging material and c) temperature
and packaging material.

In practice various FCM may come in contact with more than one type of food, for instance
fatty versus acidic foods. In that case, migration into both or more food simulants should be
tested.
page 80 of 234 RIVM report 320003001

The duration time of the migration test should correspond to the worst foreseeable conditions
of contact and to any labeling information on maximum temperature for use, see Table 5.2.

Table 5-2 Guidance on the duration of the migration tests and the test temperature for food contact material.
conditions of contact in the
test conditions
worst foreseeable use
Contact time test time
time corresponding to worst foreseeable use
t < 5 min
(but < 5 minutes)
5 min < t < 0.5 h 0.5 h
0.5 h < t < 1 h 1h
1h<t<2h 2h
2h<t<4h 4h
4 h < t < 24 h 24 h
t > 24 h 240 h
contact temperature test temperature
T < 5°C 5°C
5°C < T < 20°C 20°C
20°C < T < 40°C 40°C
40°C < T < 70°C 70°C
70°C < T < 100°C 100°C or reflux temperature
100°C < T < 121°C 121°C*
121°C < T < 130°C 130°C*
130°C < T < 150°C 150°C*
T > 150°C 175°C*
*
this temperature shall be used only for simulant D. For simulant A, B, or C the test may be replaced
by a test at 100 °C or at reflux temperature for duration of four times the time selected according to
the general rules.

5.4 Comparison migration tests of food contact materials and

toys

It was investigated whether a certain relationship between tests and between migration limits
of substances exists by comparing migration tests for food contact materials and for toys.
This would enable extrapolation of the limit of migration of (types of) compounds addressed
in the food contact material legislation to the legislation of migration limits of these (types of)
compounds for toys.

Within the legislation for food contact materials a range of different migration tests are
possible, depending on the worst foreseeable conditions for the contact material. Four
RIVM report 320003001 page 81 of 234

different types of food simulant are used, the duration of the migration test varies between
0.5 h to 240 h, and the test temperature between 5 ºC and 175 ºC, whereas in EN 71-3 the
migration of elements from toy material is performed at 37 ºC (European Committee for
Standardization (CEN), 1994) and migration of organic substances in saliva according EN
71-10 at 20 ºC.

For the present migration tests for toys two categories can be discriminated (chapter 3):
• Mouthing a toy matrix, which can be simulated with artificial saliva or water
• Ingestion of toy matrix. In EN 71-3 this is simulated for a set of elements in which
migration is measured following exposure to an acidic fluid (European Committee for
Standardization (CEN), 1994). This fluid can be considered to represent (artificial)
gastric juice.

When comparing the migration tests for food contact materials to migration tests for toys,
two sets of migration tests of food contact materials are similar to the tests for toy matrices.
This is based on physicochemical similarity of the extraction medium:
• The migration of food contact materials into water (simulant A) may correspond to
the migration into saliva simulant or water.
• The migration of food contact materials into 3% acetic acid (w/v) (simulant B) may
correspond to some extent to the migration into hydrochloric acid solution (generally
0.07 M adjusted to pH 1.0 to 1.5) as is applied according to EN 71-3 for several
inorganic compounds in toy matrix (European Committee for Standardization (CEN),
1994).

Table 5-3 Comparison of migration tests under Toy Legislation for elements (CEN 1994) and the FCM
legislation.
Toys Food contact materials
Based on aqueous Extraction into water or Extraction into distilled
solutions saliva simulant water
Based on acidic extraction Extraction into 0.07 M Extraction into 3% acetic
medium HCl (pH 1.0-1.5) acid (w/v) (pH ≈ 2.5
(calculated))

It is obvious that these tests only represent oral exposure. Routes like inhalation or dermal
exposure are not covered in the present legislation on food contact materials.
A few examples of results from migration tests performed under toy legislation and food
contact materials legislation were gathered and compared. The assumption is that comparable
test designs mimic comparable processes like migration from a matrix in gastric juice. In this
way results from tests with toys and tests with food contact materials might be linked. This in
page 82 of 234 RIVM report 320003001

turn may have as a consequence that test results for a certain compound with food contact
materials may be predictive for that compound in toy material.
Although there might be similarities in composition of extraction medium, it should be noted
that differences in factors like extraction time, extraction temperature, and mixing during
extraction may have a huge impact on migration measured.

Other comparisons between the migration tests of food contact materials and the migration
tests for toy matrices according to EN 71-3 are not expected to lead to similar, systematic
results. The physicochemical characteristics of other extraction media for food contact
materials such as ethanol and oil solutions are very different from the saliva or gastric juice
simulants relevant for compounds in toys. Therefore, the potential difference in amount
migrated is large and is not expected to be a constant factor.
It can be noted, however, that for specific material/chemical combinations, extraction with
food simulants C (10% ethanol) or D (oil) oil, can result in ‘worst case’ extraction for toys.
For example, the migration of hydrophobic compounds into oil (food simulant D) may
always be greater than migration into artificial saliva/water or gastric juice. In that case, the
migration test could potentially be used for toys also. However, up till now, no comparison
on the migration of a compound from the same material with food simulants C and D and
artificial saliva or gastric juice is available. To define the physicochemical properties of the
compounds for which this ‘worst case’ assumption holds the migration of a large set of
compounds should be investigated. At this moment only a case-by-case examination could be
made which cannot be incorporated in a routine methodology.

Even for similar extraction media, the outcome of a migration test highly depends on:
1) The exact composition of the extraction medium: there is a difference in the
composition of food simulant B (3% acetic acid, calculated pH about 2.5) and
artificial gastric juice (0.07 M hydrochloric acid, pH 1.0-1.5).
2) Physicochemical properties of the substance of interest, e.g. lipophilic compounds
are expected to be sensitive to other factors in the extraction media than for
example elements
3) In general, the matrix from which the migration of a compound is investigated is
different for food contact and toy material.
4) An important difference in the migration tests for food contact material and toys is
that the migration tests for food contact materials are static, i.e. the food contact
material and the food simulant are not stirred or mixed. On the other hand, the
migration into artificial gastric juice is determined after 1 h of shaking and an
additional 1 h not shaking. Dynamic testing for toy material is performed to
simulate chewing in the mouth and peristaltic movements in the gastrointestinal
tract. Shaking is known to dramatically increase the migration of compounds
(Fiala et al., 2000; Steiner et al., 1998).
5) A direct comparison between migration of a compound into food simulant and
artificial saliva or gastric juice is difficult because migration into food simulant is
RIVM report 320003001 page 83 of 234

usually expressed as mg/l food simulant, and migration from toy material is
expressed as mg/kg toy. In section 5.6 an example is given on the obstacles one
comes across in recalculating from mg/L simulant to mg/kg toy.
6) The temperature of the migration test for FCM depends on the worst foreseeable
use and ranges between 5 ºC and 175 ºC. Migration tests with artificial saliva or
water to investigate the migration of compounds from toys are usually performed
at body temperature, i.e. 37 ºC (EN 71-3), or room temperature, 20 ºC (EN 71-10).

Even small differences in conditions of the migration test may lead to substantial differences
in migration.

5.5 Comparison of migration limits of substances according

to tests for food contact material and toy regulation
To illustrate the comparison of migration of substances in toy material and FCM, migration
limits, migration test conditions and measured migration values are listed in the table below
for two substances, lead and bisphenol-A:

5.5.1 Lead
A comparison of the migration tests and some migration data of lead from ceramics (FCM)
and toy is given in Table 5-4.
page 84 of 234 RIVM report 320003001

Table 5-4 Comparison of the migration tests and migration data of lead from ceramics (FCM) and toy material.
Food contact material (Council Directive Toy (EN 71-3)
84/500/EEC)
Matrix Ceramics Toy
Migration test Extraction in 3% (v/v) acetic acid in freshly Hydrochloric acid 0.07 M adjusted to pH 1.0
prepared aqueous solution, at a temperature to 1.5 in the presence of toy; Migration
of 22 ± 2 ºC for a duration of 24 ± 0.5 h; measured after 1 h with agitation and 1 h
usual lighting conditions without agitation at 37 ± 2 ºC.
Migration limitd 0.8 mg/cm2 a 90 mg/kg toy material (129 mg/kg after
4.0 mg/l food simulantb analytical correction)
1.5 mg/l food simulantc
(For migration of cadmium and lead from
ceramics special migration limits hold,
further detailed in reference a-d).
Actual migration Ceramic samples bought in different Migration from paint scraped from wooden
values measured shopping centers in Spain (González-Soto et toys (Bouma et al., 2004)
al., 2000) 2815 mg/kg paint (red, top)
Between 1.21 mg/l and 0.027 mg/l food 30 mg/kg paint (mix, colored pencil)
simulant (15 samples) 14 mg/kg paint (white, colored pencil)
24 mg/kg paint (red, top)
11 mg/kg paint (yellow, push wagon)
11 mg/kg paint (colored pencil)
10 mg/kg paint (blue, box with blocks)
13 mg/kg paint (yellow, box with blocks)
11 mg/kg paint (blue, blocks)
169 mg/kg paint (mix, clown)
11 mg/kg paint (yellow, breakfast set)
a: Articles which cannot be filled and articles which can be filled, the internal depth of which measured from the
lowest point to the horizontal plane passing through the upper rim, does not exceed 25 mm.
b: All other articles which can be filled.
c: Cooking ware; packaging and storage vessels having a capacity of more than three liters.
d: When a ceramic article does not exceed the indicated limits by more than 50%, that article shall nevertheless
be recognized as satisfying the requirements if at least three other articles with the same shape, dimensions,
decoration and glaze are subjected to a test carried out under the indicated conditions, and the average quantities
of lead extracted from those articles do not exceed the limits set, with none of those articles exceeding those
limits by more than 50%.
RIVM report 320003001 page 85 of 234

In addition to the comparison for lead, a comparison of the migration limits, migration tests,
and some migration data of bisphenol-A from FCM and toy material is given in Table 5.5.

Table 5-5 Comparison of migration tests and migration data for bisphenol-A in the FCM framework and toys
Food contact material Toy Drinking ware for
children
Matrix Plastic Plastic Plastic
Migration test depends on the conditions of 60 min at room 24 h at 40ºC in
use (simulant, time, temperature, dynamic simulant A (water) and
temperature). Static migration B (3% acetic acid)
migration.
Migration limit SML = 0.6 mg/kg food or 0.1 mg/l (monomers) 0.03 mg/l simulant
simulant (2002/72/EG) (EN 71-9, EN 71-10, (EN 14350-2)
EN 71-11)
Concentration is
measured in simulant,
not expressed as mg/kg
toy material.
Actual migration Not available 0.005 – 0.5 mg/kg < 0.004 mg/l simulant
values measured simulant

Although some scattered data on other substances are available, these data did not allow a
systematic comparison as shown above. A main problem for comparing the different types of
migration is that the migration limits are expressed fundamentally different, i.e. as mg/L food
simulant in the framework of FCM and as mg/kg toy for the toy framework. Thus, the limits
are expressed in terms of the receptor matrix for FCM and in term of the product for toys. As
can be concluded from the two examples with lead and bisphenol-A, there are large
differences in test conditions, units used for expressing the limit value and actual measured
migration values. No consistent relation or conclusion can be drawn from this analysis.

5.6 Can the methodology of FCM be used for toys?

As stated in the introduction, it would be very efficient if migration limits for FCM could be
extrapolated to toys.

In the regulation of food contact materials, substances are categorized into several lists
[Synoptic document; Directive 2002/72/EC]. List 5 substances should not be used in food
contact materials. For list 6 substances suspicion exists about their toxicity and data are
lacking or are insufficient. Substances in section 6A are suspected to have carcinogenic
properties. These substances should therefore not be detectable in food simulants by an
appropriate sensitive method for each substance. Substances in section 6B are suspected to
have toxic properties other than carcinogenicity. Restriction for food contact materials may
be indicated. Hence, it can be considered, that the substances indicated in list 5 and 6 in the
page 86 of 234 RIVM report 320003001

food contact material legislation might also not be acceptable for toy materials. However, it
should be stressed that such a proposal would be made from a risk management perspective
and not from a risk assessment perspective.

TDIs have been listed for some substances in FCM legislation. The TDI is based on the
toxicity of the substance in question, and can be indeed directly applied to derive limits of
migration for toy materials. The methodology described in the present report can be used to
derive a migration limit for toys, e.g. by application of a certain percentage of the TDI that
can be allocated to toys, and expressing the thus obtained amount of substance as a migration
limit, see chapter 7 and 8.

Other applications of limits in FCM legislation to toy legislation become more complicated.
It would be practical to translate migration limits for FCM to toys. However, in practice this
is not easy, since test conditions differ, the testing material differs and the results of the tests
are expressed in different units (mg/l simulant (or mg/kg food) in the FCM framework vs.
mg/kg toy material in the toy directive). Since in the majority (if not all) cases only the
amount migrated is known but not the total amount in the starting material, expression as a
fraction that is migrated is not possible, which also hinders comparison. These aspects are
addressed in more detail below.

Within the FCM framework, it is required that testing conditions are relevant to the
foreseeable use of the material. This foreseeable use, however, may differ widely from the
conditions of toy use. Extraction time, temperature and the choice of simulant such as used in
existing FCM migration tests may be such that results cannot reasonably be used for toy-
related exposures because the toy exposure would need a different scenario. As to the
composition of an appropriate migration medium for toys this would have to reflect the
process of migration:
- to saliva during mouthing of a toy;
- to gastric juice (stomach) in case a toy fragment is ingested;
- to intestinal contents when this toy fragment moves down the GI-tract.

The FCM migration conditions do not always fulfill these criteria. Water may be used to
simulate saliva as this did not result in large differences in migration, see chapter 4. It might
be possible that the acetic acid solution of FCM (simulant B) can be used to simulate
migration into gastric juice of the stomach. However, as the pH of gastric juice simulant
(pH 1.0-1.5) is lower than the pH of the acetic acid solution (calculated pH 2.5) this is
questionable, especially for pH sensitive substances. This should be investigated
experimentally.
It might be possible to use oil (simulant C) as a simulant of intestinal juice, especially for
hydrophobic substances. However, also this comparison is questionable as no research on this
comparison has been performed up till now. Such research will have to be performed before
application of a relationship between migration into oil and in a simulant of intestinal juice.
RIVM report 320003001 page 87 of 234

As a further complication the FCM migration test is a static process, i.e. without any kind of
stirring, whereas exposure to toys via mouthing is a dynamic process, with end-over-end
rotation during an hour followed by a static period of another hour. Measuring exposure
during dynamic conditions has been shown to result in higher migration values than under
static conditions (Fiala et al., 2000; Steiner et al., 1998). It might be possible to establish
correlations for migration of substances for dynamic and static conditions. However, as this
would be necessary for a large number of substances, it would require a lot of testing.

Another complicating factor is the manner of expressing migration, e.g. in mg/l simulant (or
mg/kg food) in the FCM framework and in mg/kg toy material for toys. These different types
of migration limits can only be translated from one type into the other type with assumptions
regarding the density of the material and the depth of the material as a source for migration.
For example, for substances with low migration in the FCM framework, the SML is
< 0.05 mg/kg food (or < 0.05 mg/L fluid). This implicates that it is allowed to have 50 µg of
a substance migrating from 6 dm2 of food contact material. If we assume that migration from
packaging materials only occurs from the first 0.5 mm depth, than it can be calculated that
maximally 50 µg of a substance is allowed to migrate from 30 cm3 of material. If we assume
that the density of the material is 1 g/cm3, 50 μg of substance is allowed to migrate from 30 g
of material. This corresponds to a migration limit of 1.6 mg/kg food contact material, which
could be translated to a migration limit of 1.6 mg/kg toy material. When migration occurs
from as deep as 1 mm, the corresponding migration limit will be 2 times lower, i.e. 0.8 mg/kg
food contact or toy material. However, the assumptions on the depth of migration and on the
density of the material are disputable and depending on the material and substance in
question.

Based on the above argumentation, the authors of the report think that in principle it would be
possible to translate migration limits for FCM to migration limits for toys. However, this
would require a considerable amount of additional experimental research regarding the
investigation of relationships between migration in different solutions and between static and
dynamic extraction conditions. Even then, uncertainties remain on the translation of
migration limits for FCM to migration limits for toys due to the disputable assumptions that
have to be made for such a recalculation, and due to the uncertainty of the safety of the
migration limit for toys as the migration limit for FCM are based on different underlying
assumptions. Furthermore, it is not clear which substances covered in the FCM framework
are relevant also for toys.
Therefore, as we think that the amount of research required for translation of migration limits
for FCM to toys is large and not fully reliable, we recommend determining the migration
limits for substances in toys directly based on the approach outlined in the report.

Finally, the thus obtained calculated migration limit for toys is not necessarily safe. The basic
assumptions for deriving migration limits for FCM are different than for toys. First, the limits
for FCM are for adults and may comprise 100% of a TDI. This is acceptable as another worst
page 88 of 234 RIVM report 320003001

case assumption is used, namely that someone consumes 1 kg (or 1 L) of food that has been
in contact with FCM per day. For toy material, we advise to allocate a percentage of the TDI
to exposure via toys.

5.7 Conclusions and recommendations

Within the framework of the Food Contact Materials (FCM) a large number of migration
limits for chemical substances used in the production of Food Contact Materials has been
authorised. As both FCM and toy legislation mostly focus on oral exposure, we have
presently investigated whether the methodology of the FCM can be applied to toys as this
would provide a efficient procedure that would not require additional testing.

5.7.1 Conclusions
In the regulation of food contact materials, substances are categorized into several lists
[Synoptic document; Directive 2002/72/EC]. List 5 substances should not be used in food
contact materials. In list 6, substances are taken up for which concern exists with respect to
their safety. Hence, it can be considered, that the substances indicated in list 5 and 6 in the
food contact material legislation might also not be acceptable for toy materials. However, it
should be stressed that such a proposal should be made from a risk management perspective
and not from a risk assessment perspective.

TDIs have been listed for some substances in FCM legislation. The TDI is based on the
toxicity of the substance in question and is harmonized at the European level. These TDIs can
be used directly to derive limits of migration for toy materials according to the methodology
described in this report.

In principle it is also possible to translate migration limits for FCM to migration limits for
toys. To that end, experimental research on the relationship between various simulantia is
necessary, as well as research on correlation in migration determined for static and dynamic
conditions.
Furthermore, it is uncertain whether limits for FCM translated to limits for toys are by
definition safe. The basic assumptions for deriving migration limits for FCM are different
than for toys.

5.7.2 Recommendations
Regarding the large amount of experimental research required before migration limits for
FCM can be calculated to migration limits for toys, and the uncertainty on the assumptions
and safety of the thus obtained migration value for toys, we recommend not to use the
migration limits for FCM but to determine migration limits for substances in toys directly
based on the methodology outlined in the report.
RIVM report 320003001 page 89 of 234

6 Sampling and analysis for certain elements in toys

6.1 Introduction
For compliance testing of toys for certain elements, it is necessary to have a uniform
approach to sampling and analysis. Compliance testing is carried out by both industry and
enforcement laboratories and approached from a different point of view. Both viewpoints are
taken into consideration in this chapter. Several issues were raised in the project tender. It
was requested to look into the sampling strategy for testing of toys for certain elements,
whether a single sample is representative or not. In EN 71-3 analytical correction values are
used to correct the results. We were requested to evaluate these correction values and also the
use of these factors. Other issues that are addressed in this chapter are the analysis of
chromium3+/chromium6+ and organic tin compounds, repetitive use versus single testing.
Comparisons are made to common practices for food contact materials, childcare articles and
other toy standards. Proposals are made for sampling and analysis of toys for certain
elements. These proposals are summarised in section 6.4.

6.2 Sampling
In our opinion it is essential that sampling procedures are harmonised, in order to minimise
inter- and intralaboratory differences in testing the same materials. For both food contact
materials and toys, sampling is not prescribed, not in the relevant EN standards neither in the
relevant EU directives. In this section proposals are given on sampling and sample
preparation of toys for testing for certain elements.

6.2.1 Sampling strategy

In En 71-3 it is stated that a laboratory sample for testing shall consist of a toy either in the
form in which it is marketed, or in the form in which it is intended to be marketed. Toys must
comply with the legal restrictions when they are sold to consumers (in retail). In principle
each individual sample must be in compliance. It is the responsibility of the producer or
importer of the toy to ensure that this is the case for each toy. It is therefore important that the
sample that is used for testing is representative for the batch that is being put on the market. If
the production circumstances change, or if different or other raw materials are being used, the
toy should be tested again.

We think that toys should be tested periodically, as certain parameters that influence
compliance testing are changing. These circumstances are updates of EN standards
(every 5-10 years), minor or major changes to the product, changes in raw materials and
changes production circumstances. In the technical dossier of a toy, a test certificate must be
page 90 of 234 RIVM report 320003001

present to demonstrate that the toy complies with EN 71-3. We propose that this test
certificate may not be older than 5 year before the date of marketing of the toy.

Enforcement bodies can take a (representative) sample from retail for compliance testing. If
this sample fails, an official measure can be taken.

6.2.2 Subsamples
In the Directive 88/378/EEG on toy safety it is required for the protection of the children’s
health, that the bioavailability of certain elements may not exceed certain health-based limit
levels per day (see chapter 4). In EN 71-3 bioavailability levels are translated to migration
limits. This translation is based on the migration of elements from all accessible parts of toys
into a hydrochloric acid solution. Toy packaging materials are excluded for testing. A toy
may be composed of different materials, for example wood, paint, textile, and plastic (see
clause 1 of EN 71-3). In clause 4 (requirements) it is stated that all accessible parts must
comply with the migration limits. We propose that this should remain unaltered.

The analytical correction values vary from 30 to 60%. Some of these values are rather high.
One source of analytical variation may the inhomogeneity of the test material. It is therefore
important that a (sub) sample is homogenised well. For example for coating of paint a
minimum of 100 mg passing through a 0.5 mm sieve is prescribed in EN 71-3. It is well
worth investigating to scrape off the entire coating, to grind this well. This entire coating is
then sieved using a 0.5 mm sieve. From that fraction a 100 mg portion is taken and used for
analysis. Another improvement may be to use a sieve with smaller dimensions. In this project
it was not foreseen to carry out such lab experiments.

6.3 Analysis

6.3.1 Introduction
In the tender questions were raised how to apply the analytical correction values that are
listed in EN 71-3. In this section correction values from testing of food contact materials and
organic chemical substances in toys (EN 71-9, EN 71-10 and EN 71-11) are compared. In
addition several analytical issues are addressed, such as analysis of
chromium3+/chromium6+ and organic tin compounds, and repetitive versus single testing.
Proposals are given how toys should be tested for certain elements.

6.3.1.1 Analytical correction values

In EN 71-3 in Table 2 analytical correction values are listed for the different elements,
varying from 30 to 60%. These correction values are based on the precision data from a 1987
ring trial (see Annex D.4 of EN 71-3). In a ring trial the interlaboratory variability was
RIVM report 320003001 page 91 of 234

established for 8 elements. These values are used in EN 71-3 as analytical correction values
to correct for the variation of the method (see chapter 4), thereby in practice increasing the
migration limits for the different elements.

In Denmark in 1998 a market surveillance was carried out (Teknologisk Institut, 1998). Toys
were sampled and tested for compliance with EN 71-3. The use of analytical correction
values was critical in only 3 out of 10784 cases, in changing the outcome from failing to
passing the test. The analytical correction values appeared to be of little practical importance.
In their opinion correction values should be used to lower the limit of migration to ensure a
reasonable safety margin.

For polymeric food contact materials several ring trials have been performed for specific
migration of organic contaminants (EN 13130 series). Precision data are included in these
standards in the Annex. These ring trials concern the migration of an organic substance in the
official food simulants: distilled water (simulant A), 3% acetic acid (simulant B),
10% ethanol (simulant C) and olive oil (simulant D). The average standard deviation
(interlaboratory variation) is 38% for the migration of organic substances. In 2000 the
migration of diisononylphthalate from PVC standard discs in saliva simulant was tested in a
ring trial (Simoneau et al., 2001). This standard PVC disc was tested for homogeneity. The
interlaboratory variation using this standard disc was 30%.

The analytical variation values mentioned above show that some of the analytical correction
values used in EN 71-3 (30-60%) are high. It is therefore recommended to improve the
method and organise a new ring trial (see section 6.3.2.1 and also 6.4).

For food contact materials it is not custom to include this analytical variation in the standard.
In our opinion analytical variation should be dealt with from different perspectives:

When a test laboratory wants to certify a test sample, they have to demonstrate compliance by
proving that the migration value is below the legal limit. Therefore the migration value
including analytical variation may not exceed the migration limit.
When an enforcement laboratory, such as the Dutch Food and Consumer Product Safety
Authority analyses a sample, they must demonstrate that the sample exceeds the legal limit,
before they can take an official measure. The migration value is corrected by subtracting the
analytical variation.

As an example to demonstrate this principle the following case is elaborated. For food
contact materials the specific migration limit for barium nitrate is set at 1 mg/kg (Directive
2002/72/EC). If a test laboratory wants to demonstrate compliance, the migration of this
substance must be below 0.76 mg/kg (0.76 mg/kg * 130% = 1 mg/kg). When an enforcement
body wants to take measures, they must demonstrate that the migration limit is exceeded.
Official measures are taken when the migration exceeds 1.3 mg/kg.
page 92 of 234 RIVM report 320003001

In EN 71-3 the analytical variation (correction values) is used to increase the limits. The
limits for elements that are proposed in the present report are based on toxicological
concepts, which means safety limits to ensure the health of children. It is therefore suggested
to include the precision data of EN 71-3 in an Annex of this standard and to give instructions
how to use these correction factors. Our proposal is to subtract the analytical variation from
the limit, from a consumer health protection point of view.

6.3.1.2 Standard reference material

It is recommended to introduce a standard reference material, which contains all the relevant
elements at a relevant level. This standard reference material must be used as a quality
control sample and must therefore be used in each series of analysis, to demonstrate the
ability of a laboratory to correctly analyse toys for certain elements.

The precision data that are used to calculate the analytical correction values in EN 71-3 are
dated from a 1987 ring trial. We suggest to organise a new ring trial, using this standard
reference material. In these 20 years there has been an improvement in the precision of the
analytical apparatus. Most labs have now changed to the ICP technique (Inductive Coupled
Plasma). It is also important to have labs participating in the ring trial that have ample
experience in testing according to EN 71-3.

6.3.1.3 Chromium3+ and Chromium 6+

In EN 71-3 a limit is set to the total amount of chromium. However, chromium6+ has a
different toxicological profile compared to chromium3+. Chromium6+ is classified as
carcinogenic category II (may cause cancer by inhalation) and as a skin sensitizer (Directive
67/548/EEC). From this perspective it is desirable to be able to measure chromium6+ instead
of total amount of chromium. Due to the acid circumstances of analysis, chromium6+ is
converted into chromium3+. Although we recognize that it is relevant to have an analytical
method that can measure chromium6+, we realise that a reliable and validated method is not
available yet. We are aware of the efforts of various scientific groups on developing such a
method.

6.3.1.4 Organic tin compounds

Some organic tin compounds are immunotoxic compounds. It is desirable to determine these
organic tin compounds not only as the element tin, but as organic substance as well. In a
Dutch report (Bouma et al., 2004) methods are provided to determine the total amount of
organic tin in plastic, as well as the migration into water. Only limited validation has been
carried out. Furthermore this method has not been tested at other laboratories. This method
therefore needs to be validated further and tested at other laboratories as well.
RIVM report 320003001 page 93 of 234

6.3.1.5 Repetitive versus single testing

For food-utensils intended for repeated use, three successive migration tests on the same
sample have to be carried out. The result of the third test must comply with the requirements.
It is assumed that the migration levels fall with increasing number of migration tests.

For childcare articles, such as soothers, soother holders, teats and drinking equipment, several
EN standards have been adopted (EN 12586, EN 12868, EN1400-3 EN 14350-2). These
childcare articles are intended for repetitive use. In these EN standards for childcare articles,
the result of the first migration test must comply with the migration limit. The result of the
first migration test is considered to be the worst-case exposure to the migrating substance. As
children are a sensitive group of consumers (see section 2.3.1) a high protection level is
realised. Most of the toys are intended for repetitive use. It is proposed that only one
migration test is carried out and that the results form this test must comply with the legal
limits, similar to childcare articles.

6.3.2 Test methods

In section 3.4 (Identification of relevant exposure scenarios) a scheme is presented for the
different exposure routes. Independent of the exposure route, compliance can be easily
demonstrated by determining the total amount of the elements. For all laboratories (from
industry, test institutes and enforcement), it is desirable to have an easy screening test. If the
sample passes the screening test, no further testing is required. If the sample fails the
screening test, more in depth information on exposure can be used to demonstrate
compliance.

6.3.2.1 Total amount of elements

In the methodology (see section 8.5) it is proposed to determine the total amount of elements
as a screening method. In general total contents of elements are determined by destructing the
material, using nitric acid and hydrogen peroxide in a microwave oven (high temperature,
high pressure). Methods for destruction are not described in a standard. For tattoo inks a
method is available to determine the amount of elements. This method has however not been
tested for toy materials. It is advised to develop a reliable and validated method for
destruction of toy material. This method can then be added to EN 71-3 in the annex.

6.3.2.2 Bioavailability
In EN 71-3 the migration of elements is studied by means of an acidic simulant. This
simulant presumably substantially overestimates the migration for a toy that is mouthed. On
the other hand it may also overestimate migration of elements from (parts of) toys that are
ingested (see section 8.5). For that reason the National Institute for Public Health and the
Environment (RIVM) developed a physiologically based in vitro digestion method (Oomen et
al., 2003). This method was investigated for the migration of lead from toys and could be
page 94 of 234 RIVM report 320003001

used as a refinement of the exposure assessment under option 3 of the proposed

methodology.

6.3.2.3 Mouthing
To simulate mouthing of the toy, the migration can be determined using the Head over Heels
method. This is described in EN 71-10. A test portion of 10 cm2 is put in a flask containing
100 ml water. This flask is rotated at room temperature for 60 minutes. The resulting solution
is then analysed for certain elements. In the EN 71 part 9 to 11 series it is assumed as a worst
case, that a child sucks for 3 hours per day on its toys. The limits in EN 71-9 have been
corrected for this. As the migration test is carried out for only 1 hour, the result of this test
should be multiplied by 3 to correct for the shorter migration time.
It should be noted that for the derivation of the migration limit values of elements, the
ingestion scenario has been used and the values have not been corrected for the shorter
migration time. The bioaccessibility testing as measured by the EN 71-3 migration test
involves an acid extraction for 1 hour with and 1 hour without shaking (total 2 hour
extraction). Since the EN 71-3 migration test is a rough and worst case simulation of reality
the 2 hours testing can be regarded as being sufficient for longer periods up to 3 hours. In this
sense, the difference between the 1 and 3 hour period is not really relevant.

6.4 Recommendations
In summary, the following is proposed with regard to sampling and analysis of toys for
certain elements:

1) A single sample can be used for compliance testing.

2) For enforcement laboratories measures can be taken based on the results of a
single sample.
3) Toy producers or importers must ensure that the sample used for compliance
testing is representative for what they place on the market. Periodic testing is
required if there are relevant changes in production circumstances, raw materials
or in the standards. We propose that this test certificate may not be older than
5 year before the date of marketing of the toy.
4) All accessible parts of a toy must comply with EN 71-3. If a toy consists of
different materials, subsamples should be made of each material.
5) It is recommended to improve the precision of the method, by optimising the
sample preparation (homogeneity of the (sub) sample). This requires lab work.
6) We recommend to introduce a standard reference material that contains the
elements in relevant quantities.
7) A new ring trial should be organised to establish precision data, using this
standard reference material.
RIVM report 320003001 page 95 of 234

8) Precision data of the analytical method, including interlaboratory variation, should

be included in an annex of the standard. Interlaboratory variation (analytical
correction values) should not be used to increase the limit of migration of the
elements. Depending on the point of view (toy producer/importer, enforcement)
these precision data can used differently to decide whether a toy is in compliance
or not.
9) Although toys are intended for repetitive use, the result of the first migration test
must comply with the requirements.
10) It is recommended to have a method to determine the total amount of elements in
toy material. Several options are available. This requires lab work.
page 96 of 234 RIVM report 320003001
RIVM report 320003001 page 97 of 234

7 Proposed general methodology for setting limit values

for chemicals in toys

7.1 General introduction

As described in chapter 1 of this report, there is a need for a transparent and scientifically
sound procedure for setting limit values for chemicals in toys. In the following paragraphs we
will propose a methodology for setting limit values for chemicals in toys that can be used
within the EU. This chapter will focus on the headlines of the methodology. In the next
chapter, the proposed methodology is worked in more detail to derive migration or content
limit values for elements in toys, and to provide migration limits for elements in toys.
A number of risk management issues have been encountered while developing this
methodology and we will provide suggestions on how to deal with these issues. Obviously,
the ultimate decisions on these issues will have to be made by the risk managers or the policy
makers.
Before entering a discussion on the proposed methodology it is important to clearly describe
our basic philosophy on determining the safety of the use of chemicals in consumer products.
This philosophy drives the fundamental choices made in the approach we followed for setting
limits of chemicals in toys. Therefore, we will first introduce our conceptual framework for
safety evaluation of chemicals in consumer products with specific focus on toys. We will
then show how this framework can be used to derive limits for chemicals in toys. The
proposed methodology consists of three different options depending on the level of detail
needed for this purpose.

7.2 Basic starting points

Our basic point of departure for determining limit values for chemicals in toys is based on the
same aspects that are important for safety evaluations of consumer products in general, but
with specific attention to toys for children:
- The approach should provide a general framework that basically can be used for all
chemicals.
- The approach should provide an adequate level of health protection for children.
- The approach is not based on health hazards only but is based on a quantitative health
risk approach.
- Therefore, next to health hazards, exposure assessment will be a focal issue in the
approach.
- The approach should be transparent.
page 98 of 234 RIVM report 320003001

- The approach should be applicable for the whole range of different toys and materials
used to produce toys.
- The approach should include all potential routes of exposure.

When we take these aspects into account, our conceptual framework attaches great
importance to the issue of exposure assessment. Exposure of children to chemicals from toys
is not a one-dimensional issue involving only one type of exposure. Principally, all possible
routes will have to be taken into account both for evaluating the safety of chemicals in toys
and for setting safe limit values for these chemicals, although some routes of exposure will be
quantitatively more important for certain chemicals or toys than others. As discussed in
chapter 3, all the following routes of exposure are relevant for different types of toys and
chemicals: oral route (e.g. mouthing, sucking and ingestion of matrix parts), dermal route
(e.g. direct dermal contact, dermal loading of liquids), and inhalation (e.g. evaporation of
substances, dusting). The significance of a certain exposure route is determined by both toy
and chemical properties. This is depicted in the framework below, which shows that a range
of factors determine the exposure of children to chemicals originating from toys. Contact
scenarios (e.g. duration of contact, frequency of contact), routes of exposure, migration from
the toy to a physiological matrix (e.g. saliva or skin), and uptake into the body all contribute
to the total exposure to a chemical.

7.3 Proposed general methodology to derive limit values for

chemicals in toys

With regard to chemical properties, the Toy Directive states that toys should not present
health hazards by ingestion, inhalation or contact with the skin, mucous tissues or eyes, when
used as intended or in a foreseeable way. The basic idea behind the proposed methodology is
therefore to use health-based limit values as the core value from which other values such as
migration limit values and product content limit values can be derived. In addition to this, one
has to consider other potential routes of exposure next to toys. Some substances may occur
naturally e.g. in the environment in food. Therefore, it is proposed to allocate only a fraction
of the health-based limit value (e.g. 5, 10 or 20%). What this fraction should be can be
scientifically advised but will eventually be a risk management decision. In practice, this can
essentially be translated into the following statement:

The exposure of children to chemicals in toys may not exceed a certain

fraction of a health based limit value (in mg/kg bw/day)
 RIVM report 320003001 page 99 of 234

Chemical in toy (material)

Contact scenario:
frequency, duration

Oral route: Oral route: Dermal route: Inhalation route:

Mouthing Ingestion of matrix Contact Vapor - dust

Local: Local: Local:

oral effects dermal effects airway effects

Migration: Migration: Migration:

Saliva Gastric or intestinal juice Sweat

Absorption: Penetration: Absorption:

From gastrointestinal tract to blood Through the skin From lung to blood

Internal absorbed dose

Health based limit value

Figure 7-1 Risk based framework for chemicals in toys

7.3.1 Use of a risk based framework

The proposed methodology, based on a risk based framework, is illustrated by the scheme
above (Figure 7-1) and can be used to serve two purposes:
1. to derive limit values that can be applicable for a whole range of toys covering a large
number of exposure scenarios;
2. for an in-depth assessment of the risks associated with the use of chemicals in
(specific) toys.
By working the scheme from bottom to top, it can be used for the first purpose: to derive safe
limit values for chemicals in toy (materials), in terms of either migration limits or product
content limits.
page 100 of 234 RIVM report 320003001

For the second purpose, the scheme should be followed from top to bottom, using all
available information on the toy – chemical combination to determine whether the use of the
toy may pose a health risk.
It is important to note that this approach takes all potential routes of exposure into account
rather than considering a single route of exposure (e.g. mouthing toy). Another aspect that is
included in this approach is the fact that also local effects due to dermal exposure or exposure
via inhalation are considered relevant for setting limit values in toys.
We suggest this framework to provide the basic approach for the methodology to set limit
values for chemicals in toys. As will be discussed in the next section, this does not imply that
detailed exposure and migration data are always needed.

It should be stressed that this methodology is a risk based approach. This implies that
policy issues, like e.g. phasing out CMRS substances from consumer products are not
considered here. Of course hazard-based aspects can be added to this methodology
although this should be carefully handled. However, it is up to the risk manager
whether these hazard aspects should be incorporated in the methodology.

7.3.2 Necessary level of detail for setting limit values

The framework described above is an integral approach based on a quantitative risk based
philosophy. Using such a framework, it is possible to derive safe and realistic limit values for
chemicals in toys. From a scientific point of view this may be the most correct approach for
setting limit values in toys. However, such a quantitative framework requires a substantial
amount of (toy and exposure scenario) specific input data or the use of substantiated default
values. Extensive experience in the field of consumer exposure and chemical migration
reveals that such input data are generally absent or incomplete. Therefore, it is expected that
this quantitative framework could only be used routinely if default values were adopted for
the various factors for which input data are lacking. Although this might be a reasonable
approach, it can be questioned whether such a procedure is necessary for all chemicals in all
types of toys or toy materials.
For example, dermal exposure may provide a marginal contribution to the total exposure for a
range of chemical-material combinations. In such a case, the dermal route can be set aside
from the protocol. However, for some other chemical-toy or toy material combinations,
dermal exposure may be very relevant (e.g. finger paint).
In a similar way, inhalation exposure is not quantitatively relevant for a large range of
chemical-toy (material) combinations. For others, sucking and mouthing may be irrelevant
(e.g when toys are unlikely to be accessible to children < 3 years of age, or when it is simply
not possible to put the toy in the mouth). These examples show that it is not in all cases
necessary to use all potential pathways in detail as depicted in the scheme above, but it is
important to document if and why certain pathways are relevant or not.
RIVM report 320003001 page 101 of 234

From a scientific point of view, the most realistic limit values will be generated by using a
quantitative risk based approach, including as much specific data on exposure and migration
as possible. This can be directed not only towards systemic toxicity but also towards local
route specific toxicity (e.g. dermal or respiratory irritation). However, it can be questioned
whether a full quantitative risk assessment is always necessary. It may be in the interest of
toy producers/importers to follow a simpler approach, using a number of worst case
assumptions, with which they can demonstrate that their product is safe, without going into
details.
The implementation of an approach for setting limit values for chemicals in toys also needs to
look at practical aspects. Assessments and testing methods need to be easy and fast, and
availability of (harmonized) testing methods or exposure assessment factors are important,
especially with regard to the toy industry that is looking for transparent and simple testing
strategies. So, the challenge is to provide a methodology that follows our conceptual
framework but does not prescribe complex and unnecessary assessments. A methodology that
includes different options may provide a pragmatic solution, as will be discussed in the next
section.

7.3.3 Options within the proposed general methodology

Taking both the basic corner stones of our conceptual framework and the practical aspects
into consideration, we propose to use a system providing three options for determining the
appropriate limit values of chemicals in toys. Within each option, the relevant health-based
limit value of the chemical under consideration is compared to the potential exposure. The
relevant health-based limit value used for the comparison is the same for all options. The
options differ in the way the exposure to the chemical is assessed, ranging from a very simple
approach requiring little testing and worst case assessment data to a more complex
quantitative risk based approach, which by means of considering data from migration tests
and exposure assessments, may provide justification for the presence of higher levels of the
chemical than was initially indicated for the simple approach.
The methodology is not designed as a sequential scheme in which the first option always has
to be used first before entering into the next option. Each of the options can be used
depending on the data available and the level of detail needed. The three options are
discussed below.

7.3.3.1 Option 1: Use of migration data

This first option is comparable to what is currently used for elements in toys, i.e. using
migration data to demonstrate that the amount of element migrating out of the toy stays
below migration limit values, as described in EN 71-3. This can be depicted in the framework
as follows:
page 102 of 234 RIVM report 320003001

Chemical in toy (material)

Contact scenario:
frequency, duration

Oral route: Oral route: Dermal route: Inhalation route:

Mouthing Ingestion of matrix Contact Vapor - dust

Local: Local: Local:

oral effects dermal effects airway effects

Migration: Migration: Migration:

Saliva Gastric or intestinal juice Sweat

Absorption: Penetration: Absorption:

From gastrointestinal tract to blood Through the skin From lung to blood

Internal absorbed dose

Health based limit value

Figure 7-2 Part of the risk based framework used for option 1 of the methodology

Currently, for toys, migration limit values exist for elements only, but this approach could
also be applied to chemicals in general. Data from migration testing can be used to
demonstrate that the bioaccessibility of a chemical from the toy is sufficiently low that
exposure will remain below the relevant health-based limit value.
By using (a fraction of) the health-based limit value as the basic assumption, we can generate
maximum values for migration of chemicals from the toy to an adequate extraction fluid.

This option can only be used if one exposure route (normally oral) is
dominant for systemic exposure.

This implies that either the contact scenario of the toy or the physico-chemical properties of
the chemical under consideration justify that exposure via other routes will be negligible. It is
RIVM report 320003001 page 103 of 234

important to use the appropriate contact scenario and extraction fluid in the migration test, i.e.
simulating the main route of systemic exposure. Some guidance can be given on this:
• For toys for children under 36 months and for toys intended to be put in the mouth, oral
exposure is likely the main route of systemic exposure for all chemicals. Two types of
oral exposure can be distinguished (see chapter 3): 1) mouthing and 2) ingestion of toy
material. Depending on the properties of the toy and on the physico-chemical properties
of the chemical under consideration, either mouthing or ingestion will contribute most to
the systemic exposure:
o Mouthing involves licking and sucking where the recipient fluid is saliva. For
chemicals in toys for which mouthing is most contributing to systemic exposure,
the relevant migration extraction fluid therefore simulates saliva. In addition,
mouthing related exposure factors such as mouthing duration, amount and surface
should be used to calculate bioavailability.
o Ingestion involves, for example, scraping off small portions of toy material
followed by swallowing where the recipient fluid is gastric or intestinal juice. For
chemicals in toys for which ingestion contributes most to systemic exposure, the
extraction fluid that will extract most of the chemical (based on its physico-
chemical properties ) should be used in the migration tests. In addition, defaults for
ingestion related exposure factors such as amount ingested should be used to
calculate bioavailability. For elements in the toys under consideration, as will be
discussed later, ingestion probably contributes most to systemic exposure. For
simple testing, the appropriate extraction fluid for elements simulates gastric acid,
as described in EN 71-3. On the other hand, for most organic chemicals, a more
complex extraction medium is required, simulating intestinal juice (for fed
conditions).
• For other toys, the dermal or inhalation route may be the main route of systemic
exposure. For inhalation it is assumed that chemicals present in the air that reaches the
alveoli are all available for uptake, i.e. bioaccessible. Therefore, migration data are not
relevant for the inhalation route. This option can therefore not be used for toys for which
inhalation presents a major route of systemic exposure. If dermal exposure is the main
route of systemic exposure, the relevant extraction fluid simulates sweat. If dermal and
oral exposure likely contribute equally to systemic exposure of the chemical, this option
should not be used. Instead, option 3 discussed below should be used.

Note: Currently, correction factors are included in the migration limit

values to correct for analytical variation. We propose not to include
correction factors, but to specify the magnitude of the analytical variation
in the annex of the standard.
page 104 of 234 RIVM report 320003001

7.3.3.2 Option 2: Use of product composition data

As discussed earlier, it may not always be necessary to perform migration tests if information
on the composition of the toy (material) is available. This option allows producers or sellers
of toys to demonstrate that their product will not exceed the relevant health-based limit values
with regard to chemical exposure using general statements and arguments only. This can be
depicted in the framework as follows:

Chemical in toy (material)

Contact scenario:
Frequency, duration c

Oral route: Oral route: Dermal route: Inhalation route:

Mouthing Ingestion of matrix Contact Vapor - dust

Local: Local: Local:

oral effects dermal effects airway effects

Migration: Migration: Migration:

Saliva Gastric or intestinal juice Sweat

Absorption: Penetration: Absorption:

From gastrointestinal tract to blood Through the skin From lung to blood

Internal absorbed dose

Health based limit value

Figure 7-3 Part of the risk based framework used for option 2 of the proposed methodology

In this option the chemical safety of the toy is demonstrated by documentation on the
amounts of chemicals present in the toy materials. This can be done using chemical analyses
of the raw materials used for making the toy. If chemical analyses of the raw industrial
materials are available and show only trace amounts of chemicals or such low levels are
present that the total amount in toys is below (a fraction of) the relevant health-based limit
value, then additional testing is not necessary. This documentation can then show the
chemical safety of the product.
The following simplified and fictive example illustrates that, under certain conditions, such a
simpler approach is sufficient. For a certain toy it is known that the total product only
contains a limited amount of a certain chemical. The assumption can be made that all of this
chemical is released at once resulting in an exposure on a single day. If this single exposure is
RIVM report 320003001 page 105 of 234

less than (a fraction of) the TDI (or another relevant health-based limit value) then this toy
can never provide a health risk. A quantitative approach following all potential routes of
exposure, or even migration tests are in that case rather superfluous.
It is important that the chemical analyses have to include all materials, not just the major raw
industrial materials. For example, if raw material for plastics only demonstrates a trace of
lead but the plastic will be mixed with a colouring mixture, an analysis of the raw material
only is not sufficient because the colouring agent can introduce an amount of lead. Either an
analysis (or an analysis certificate from the supplier) of the final coloured material is needed,
or analyses of both the raw plastic material and the colouring mixture is needed.
This option can be viewed as a kind of ‘waiver-opportunity’ for further testing. Those
producers that have data available to demonstrate the absence of chemicals in their material
can use those data for compliance with the health-based limit value. Although we realize that
currently such data may often not be available to the producers of toys, under the upcoming
REACH regulations this may change.

7.3.3.3 Option 3: Use of risk based data

The use of this option is recommended in the following cases:
• Chemicals in toys for which exposure via inhalation may occur.
• Chemical in toys for which more than one exposure route contribute significantly to the
systemic exposure.
• When the migration test results indicate that the bioaccessible amount may exceed the
relevant health-based limit value for the chemical under consideration, and it can be
demonstrated that default factors used for the derivation of these limit values are not
relevant for the toy under consideration.

In essence, all the aspects of the framework as depicted in Figure 7-1 are taken into account
in this option. The option provides the opportunity to demonstrate the chemical safety of a
product by using a number of specific exposure scenarios and / or refined migration testing
and therefore it can as such also be used as guidance for an EC type examination6.
Within this option, it is possible to refine the exposure assessment in a number of ways.
1) Exposure scenarios and exposure factors that are more specific for the toy in question
can be used instead of the defaults used in option 1. Some guidance on this is given in
chapter 3. For example, for some toys the frequency of contact is not daily but
instead only once a week.
2) Migration measurements can be refined in order to simulate the most realistic
condition. Chapter 4 presents possibilities to refine migration testing, eg. by using
more physiologically based migration tests.
3) A further step for refinement is the use of actual information on the absorption of the
chemical via different exposure routes. In this way the internal (systemic) exposure

6
EC type examination is the procedure by which an approved body, called ‘Notified Body’ ascertains and
certifies that a model of a toy satisfies the essential requirements of the Directive Safety of Toy
page 106 of 234 RIVM report 320003001

can be calculated. One should realise then that also the TDI value should be
transferred to an internal (systemic) value.

7.4 Chemicals with sensitising properties

The paragraphs above describe a methodology that can be used to assess the safety or set safe
limit values for chemicals with toxicological endpoints for which a threshold can be set. This
is mostly aimed at systemic exposure but local route specific toxicity (e.g. dermal irritation)
can also be taken into account. However, for some chemicals, sensitisation may be the most
critical effect upon dermal (or inhalation) contact. In this case, a limit value would have to be
based on this endpoint.
However, a major shortcoming in the risk assessment of sensitising chemicals is the lack of a
validated or harmonised method to use this endpoint in a quantitative risk assessment.
Substantial efforts are made towards developing a method to determine the relative
sensitising potency of a chemicals by several scientific groups (e.g. Ehling et al., 2005a,b;
Griem et al., 2003; Jowsey et al., 2006) and in the scientific literature some proposals can be
found for a quantitative risk assessment for sensitisation. There is however, at this point in
time, no consensus on this aspect. It is currently therefore not possible to derive limit values
for sensitizing chemicals. The only exception is when human response data (e.g. human patch
testing) are available. In this case, a dermal limit value can be derived that will not lead to a
significant response in the human population. Such an approach has been used for the
regulations on nickel for example. However, such data are only available for some chemicals.

7.5 Hazard aspects

As described in the beginning of this chapter, in this risk based approach no hazard aspects
are included. However, it can be proposed that some hazard characteristics of
substances/products are considered undesired with respect to toys. Exclusion of substances
with specific hazard characteristics is politically sensitive and we propose therefore, to leave
such decisions with the risk managers. In addition to this, it should be emphasised to handle
the hazard characteristics (classification and labelling) of single substances very carefully
since this may not mean that the final product (toy) has the same level of concern
(classification). For example, exclusion of an irritating substance (as classified by the
response in the rabbit eye test with 100 mg of pure substance) has no real meaning if the
substance only occurs in trace amounts in the final product.
Especially accessible liquid products may require some further consideration. In EN 71-9,
additional exclusions have been formulated, such as accessible liquids classified as very
toxic, toxic harmful, corrosive, irritant or sensitising. Although not all hazard aspects
included in that regulation may require continuation, a careful re-evaluation of these aspects
is proposed for addition to this methodology.
RIVM report 320003001 page 107 of 234

In the next chapter it is demonstrated how the proposed methodology can be used to derive
migration or content limits for elements in toys by means of option 1 or 2 and which
refinements can be made within option 3.

7.6 Conclusions
1) The proposed methodology is a risk based approach.
2) The cornerstone of the proposed methodology uses a maximal allowable level of
exposure from toys expressed as a fraction of a chronic health-based limit value
(TDI). This deviates from the current methodology on limit values for toys which
are based on chosen percentages of background exposures in combination with
limited toxicological evaluation, which represents a scientifically less rigorous
approach than that proposed in the present report.
3) Migration limits or product limits are derived quantitatively but always represent
values deduced from the fraction of the TDI.
4) To demonstrate that the exposure to chemicals from toys will remain below the set
fraction of the TDI, three options are provided allowing maximal flexibility.

7.7 Recommendations

Currently no validated method exists to quantitatively assess sensitizing dermal or respiratory

effects of chemicals. For the dermal route, developments for a quantitative method are
ongoing, but no consensus exists. In time, when consensus has been reached and a method
has been validated, dermal sensitising properties should be included in the derivation of safe
limit values for chemicals in toys.
With respect to the inhalation route, research on the potency evaluation of respiratory
sensitizers is still in its infancy. Further research in this area is needed before respiratory
sensitisation can be included as an endpoint in the risk assessments or derivation of limit
values of chemicals. Until then, chemicals with potential respiratory sensitising properties
should be evaluated on a case by case basis.
An additional evaluation of various hazard aspects is recommended to decide which aspects
should be placed on top of the proposed risk-based methodology. This discussion should be
performed jointly by risk assessors and risk managers although the final decision is up to the
latter group.
page 108 of 234 RIVM report 320003001
RIVM report 320003001 page 109 of 234

8 Application of the proposed methodology to derive

limit values for elements in toys

The CSTEE is of the opinion that the current limit values listed for elements in Annex II of
Directive 88/378/EEC need to be updated to take into account the latest scientific knowledge
and associated revisions of tolerable daily intakes (TDI) and average daily intakes (ADI). In
addition, the CSTEE suggests that limit values for other elements than those already listed
might be needed.
In this chapter, the methodology proposed in chapter 7 will be used to derive limit values for
elements in toys, that might be taken up in the Toy Directive. The focus of this part of the
report is on inorganic substances (elements) specifically, with the exception of tin (Sn)
compounds for which, as requested by DG Enterprise, also the organic substances were taken
into account.

To derive the limit values the following steps need to be taken:

1) Determine relevant exposure routes.
2) Define a relevant toxicological health-based limit value.
3) Determine the relevant elements and the value of the health-based limit value for
each element.
4) Determine the appropriate option of the methodology to derive limit values for
elements in toys.
5) Compare exposure to health-based limit value and calculate migration and/or
content limit values for elements in toys/toy material.

8.1 Determining the relevant exposure routes for elements in

toys

Before a relevant health-based limit value can be selected, it needs to be determined what the
relevant routes of exposure for elements in toys are. In chapter 3 an exposure scenario
decision tree is presented that can be used to define relevant exposure scenarios.
As explained in this chapter in principal the contact scenarios for all three routes have to be
considered to see whether they are relevant for the chemical-toy combination in question.

The contact scenarios for the oral route should be considered if the following questions are
answered with yes:
• Can the toy be put in the mouth or is it intended to be put in the mouth?
• Can the toy or part of the toy be directly ingested (via hand-to-mouth contact e.g. finger
paint or chalk, e.g. paint layer on the toy that might be scraped off with the teeth)?
page 110 of 234 RIVM report 320003001

In considering these contact scenarios for elements in particular, we can assume that both
scenarios may be relevant for toys intended for children < 36 months or intended to be put in
the mouth, as elements may migrate out of the toy to either saliva or digestive juices. For toys
intended for children > 36 months that are not intended to be put in the mouth, the oral
contact scenarios are not relevant.

The contact scenarios in the inhalation route should be considered if the following questions
are answered with yes:
• Can the substance of interest be released from the toy by evaporation? (some toys may
contain volatile substances that may be released during use).
• Can the toy release dust or spray? (e.g. while using crayons, chalk dust may be released
and subsequently inhaled).

In considering these contact scenarios for elements in particular, we can assume that the first
contact scenario, evaporation, is not relevant for elements because they are not volatile. On
the other hand, elements may be present in dust or spray that may be inhaled. However, at
present, this route is relevant probably only for a limited, very specific type of toys.

For toys, the dermal route is involved in almost all cases, because most if not all toys will at
least be contacted with the hands. However, for elements in particular, the dermal route is
unlikely to contribute significantly to systemic exposure, as the uptake of elements through
the skin is generally very low except for some organo-metal complexes (chapter 4). For
organic substances on the other hand, the dermal route may be quantitatively important and
should be included in the methodology. However, deriving limit values for organic
substances is outside the scope of the current assignment.
Dermal exposure to elements is relevant in case the element has sensitizing properties (nickel,
chromium and organic tin). As discussed earlier, there is currently no validated method to
quantify the sensitizing potential of a chemical and it is recommended to evaluate the safety
of using of sensitizing chemicals in toys separately, on a case by case basis, until such a
validated method exists. In addition, we recommend that the provisions of the EU legislation
on Nickel (Commission Directive 2004/96/EC, amendment to Commission Directive
1994/27/EEC) be adopted for toy material. Therefore, for the purpose of deriving safe limit
values for elements in toys in general, the dermal route will not be further worked out.

In conclusion, to derive limit values for elements in toys, only the oral route will be
included. As this route is of no relevance for toys intended for children > 36 months that
are not intended to be put in the mouth, limit values for elements will only be derived for
toys intended for children < 36 months and for toys intended for children > 36 months that
are intended to be placed in the mouth.
RIVM report 320003001 page 111 of 234

8.2 Defining a relevant health-based limit value

A fundamental aspect for setting limit values for chemicals in products (such as toys) is to
decide what level of exposure is maximally acceptable for an individual from a perspective of
health risks. This value ultimately determines the migration rate or product limit that can be
considered acceptable. Various factors play a role in determining such a maximal acceptable
exposure. These factors include duration of the exposure, frequency of the exposure, route of
exposure, availability of toxicological data and their reliability, the availability of harmonized
health-based limit values, and the margin of ‘safety’ that policy makers would like to include.
Although exposure to various toys covers a wide range of products and exposures, a general
characterization can be given. Exposure to chemicals from toys (e.g. when mouthed) is
characterized by daily exposure during a period of maximum 1-2 years. In toxicological
terms this represents subchronic exposure. However, subchronic limit values are not routinely
available for all chemical substances. Chronic health-based limit values on the other hand are
routinely available for most chemical substances, at least for the oral route. Using a chronic
limit value also assures an adequate level of protection because such values will be lower
than subchronic limit values, and the exposure duration is longer than in practice (daily
during lifetime vs. frequently in the first years of life). Because of these arguments we
propose to use the concept of the Tolerable Daily Intake (TDI) for setting limit values for
elements in toys.

For local effects such as skin sensitisation and irritation, a different approach is needed. A
safe exposure level which prevents sensitisation effects cannot be given, as there is no
validated method to quantify such exposure levels. For chemicals with sensitising effects, its
safety of use in toys is best considered on a case by case basis. Some sensitizing chemicals
have been regulated in separate directives, such as nickel (Commission Directive
2004/96/EC, amendment to Commission Directive 1994/27/EEC). We recommend that the
provisions of this directive be adopted for toy material.

In chapter 2, the concept of the TDI is explained. Although the TDI should provide a safe
level of (oral) intake for the general population, special attention was paid to children as a
sensitive group. Where appropriate the latest insights in the derivation of the TDI were
addressed in the values that are proposed in chapter 2.
Most – if not all – substances do not exclusively occur in toys, but are used in various
products. People are thus exposed to these substances through various products and routes.
These may include food, drinking water, ambient air, consumer products etc. Some of these
exposures are difficult to control, especially when dealing with widespread environmental
occurring agents such as elements. It is not acceptable therefore that the (daily) systemic
exposure from toys would fill up the total TDI. Background exposure through the
environment and through food and drinking water should be taken into account in order to
prevent a total exposure that exceeds the TDI. In our proposed methodology, only a fraction
of the TDI should be allocated to exposure from toys. What this fraction would be is in
page 112 of 234 RIVM report 320003001

principle a decision that should be made by policy makers. However, some recommendations
with respect to elements can be given.

General background exposure to most elements through other routes (in particular food and
indirect environmental exposure through air and drinking water), may already account for a
substantial fraction of the TDI. For example, daily maximum background exposure to
elements (as reported in the toxicological profiles in Appendix II) ranges from only a few
percent of the TDI to about 100% of the TDI. The majority of background exposures are in
the order of 20 – 70%. This indicates that from the perspective of health protection, systemic
exposure to elements from toys (if occurring daily for some continued period of time) may at
least not be allowed to any more of about 30% of the TDI in order to avoid any exposure that
will exceed this value. Again, this decision should be made by policy makers.

As a comparison it should be noted that in other regulatory frameworks, limit values for
elements are also related to a fraction of the TDI. The most prominent example is the setting
of limit values for drinking water as used by WHO for their drinking water guidelines. Limit
values for drinking water are derived by allocating a maximum of 10% of the TDI for
drinking water and an intake of 2 litres of water. This strategy is also adopted by individual
countries like the Netherlands. Although again, this is a policy decision, it is proposed that
systemic exposure to toys may not contribute to a higher fraction than is considered for
drinking water.

Based on relevant background exposures, the aspects from other regulatory frameworks, and
the aim of an adequate level of protection for children, the risk managers should decide on
the actual fraction of the TDI that will be allocated to toys. When a fully risk based approach
is followed, the fraction of the TDI used should be allocated for each element separately
based on the level of the background exposure. Another option is to allocate a default fraction
of the TDI for all elements (e.g. 5 or 30% for all elements). This is also a decision for the risk
manager. For illustration purposes we will calculate the limit values for elements with
5, 10 and 20% of the TDI.

If the general framework will also be applied to other chemicals than elements, the issue of
the fraction of the TDI has to be considered for different chemicals (or chemical classes)
again.

8.3 Relevant elements and their health-based limit values

In chapter 2 toxicological profiles for a wide range of elements are described. These profiles
take into account the most recent scientific knowledge also for the elements already present
in the Toy Directive 88/378/EEC.
RIVM report 320003001 page 113 of 234

It was very difficult to get information on the presence of specific elements in different toy
material. As far as the information was received, no conclusions can be drawn about which
elements occur most frequently and/or whether some elements are specific for particular
toys/toy materials. A general answer from the persons that were consulted was that the
elements on the present list are in the majority of cases not present in significant levels in
toys.
Therefore we used the following strategy: in addition to the present list of 8 elements, we
used the list for Food Contact Materials and a list that contains elements that are found in the
waste phase of plastics. We ended up with a list of 18 elements, for which we derived TDIs.
Given their low toxicity potency, zirconium and titanium are not relevant to be included in
the Directive and therefore no limit value will be derived for these two elements.
Since at present the main research on elements in toys considered the elements already
included in the Directive, we recommend further research on which elements are present in
toys by means of chemical analysis of a representative sample of toy (materials), so that
irrelevant elements may be removed from the list, while others may be added.
For the time being we propose that Industry provides information to decide which elements
are relevant for which type of toy material. Testing is then only needed for those elements
(but at the same time, the limit values for the other elements should not be exceeded).
page 114 of 234 RIVM report 320003001

Table 8-1 TDIs and sensitising potential for the different elements
Element TDI Background Skin irritation and
Value (μg/kg Source or exposure child sensitisation
bw/day) Reference (μg/kg bw/day) contact risk
(qualitative
indication)
Aluminum 750 Newly derived 300 Low
Antimony 6 WHO, 2003 0.53 Unknown
Arsenic 1.0 RIVM, 2001 0.4-0.7 Low
Barium 600 ATSDR, 2005 9 Unknown
Boron 160 EFSA, 2004a 80 Low
Cadmium 0.5 RIVM, 2001 0.45 Low
Chromium 5 RIVM, 2001 1 Low
trivalent
Chromium 5a RIVM, 2001 0.1c High
hexavalent
Chromium 0.0053b OEHHA, 1999 0.1 High
hexavalent
Cobalt 1.4 RIVM, 2001 0.6 Medium
Copper 83 SCF, 2003a 60 Low
Lead 3.6 JECFA, 1993; 2.0 Low
RIVM, 2001
Manganese 30 (160)d OEHHA, 2004 130 Unknown
Mercury 2 IPCS, 2003 0.1 Medium
Nickel 10 Newly derived 8 High
Selenium 5 SCF, 2000; RIVM 2 Low
1998
Silver 5 US-EPA, 1996a 1.3 Low
Strontium 600 US-EPA, 1996b 18 Unknown
Tin inorganic 2000 JECFA, 2001 290 Unknown
Tin organic 0.25 EFSA, 2004b 0.083 High
Zinc 500 SCF, 2003b 350 Low
a This value only takes into account non-carcinogenic effects by hexavalent chromium;
b This value takes into account carcinogenic effects by hexavalent chromium. It should be noted that for the
carcinogenic effect a highly uncertain Virtually Safe Dose of 0.0053 μg/kg bw/day has been proposed by
OEHHA (1999). A new drinking-water cancer bioassay with hexavalent chromium is being conducted within
the US-NTP.
c Estimate for a child playing on CCA-treated timber as given in EU-RAR (2005).
d The value of 30 μg/kg bw/day applies to exposures above normal dietary intake. For the current method of
calculation of the allowable toy-related exposure level (10% of the TDI) this TDI was converted to a value
usable for evaluating total daily exposure (inclusive of normal dietary intake). Thus for manganese a figure of
160 μg/kg bw/day was used for calculation (estimated background added to ‘non-dietary’ TDI).

8.4 Determining the appropriate option for deriving limit

values of elements in toys
As explained in the former paragraphs, the ultimate limit value for assessing the safety of
elements in toys is the fraction of the TDI that can be allowed as exposure from toys. So any
option of the methodology that will demonstrate that exposure to an element from a toy is
below the fraction of the TDI of that particular element will suffice.
RIVM report 320003001 page 115 of 234

For inspection purposes, it would be convenient to translate this health-based limit value
(TDI) into limit values for toy (material) that can be tested. With the proposed methodology,
there are two possibilities for this by calculating back from the fraction of the TDI. Firstly,
migration limit values can be generated which can be used in a comparable manner as those
listed in EN 71-3. Secondly, maximum content of the element in the toy (material) can be
derived. Both these limit values that can be used in practice for screening and inspection
purposes. As demonstrated below, generation of these limit values can be accomplished by
using options 1 and 2 of the proposed methodology. The third option can be used for an in
depth assessment on an ad-hoc basis for specific purposes (e.g. to assess the safety of toys for
which the current default factors used in the calculations are deemed too conservative). The
next paragraphs will focus on generating migration and content limit values for elements by
means of the proposed methodology.

8.5 Comparing exposure to health-based limit values

The next paragraphs will demonstrate how the options described in chapter 7 can be applied
to elements in toys. As explained in section 8.1, limit values for elements will only be derived
for toys intended for children < 36 months and for toys intended for children > 36 months
that are intended to be placed in the mouth. Limit values for elements in other toys are not
relevant.

8.5.1 Option 1: Use of migration data

As discussed earlier, the oral route is the most relevant systemic exposure route for the toys
under consideration in this report and therefore, the extraction fluid used in the migration test
needs to simulate saliva, gastric or intestinal fluid. The present migration testing guideline for
toys, EN 71-3, uses a strong acidic extraction fluid representing the gastric compartment (see
chapter 4). For elements specifically, EN 71-3 represents a worst case situation for mouthing
and a more realistic but still conservative approach for ingestion. Therefore, we propose to
use the EN 71-3 migration test for the oral ingestion route without additional testing for
mouthing. If the results of the EN 71-3 migration test, i.e. the bioaccessible amount of
element does not result in exceeding X% of the TDI, the toy material can be considered safe
with respect to the oral route. This can be expressed as follows:

Bioaccessible amount of element (mg/mg toy) x amount of toy ingested (mg/day)

< X% TDI
Body weight (kg)

Based on information in chapter 3, the following exposure scenarios and assumptions are
proposed:
page 116 of 234 RIVM report 320003001

Toys intended to be put in the mouth for children aged over 3 years
Body weight: 15 kg (based on 3-4 years of age)
Duration: 1 hour/day (worst case estimation)
Amount ingested: 8 mg/day
Absorption: 100% over the intestinal tract of the amount of element migrated out
of the toy

Toys intended for children aged 0-3 years

Body weight: 7.5 kg (based on 6-9 months of age)
Duration: 3 hours/day (worst case based on literature)
Amount ingested: Toys consisting of liquid or sticky material: 400 mg/day
Toys consisting of dry, brittle, powder-like or pliable material:
100 mg/day
Toy material scraped off: 8 mg/day
Absorption: 100% over the intestinal tract of the amount of element migrated out
of the toy

The justification for these defaults is explained in more detail in chapter 3.

The bioaccessibility testing as measured by the EN 71-3 migration test involves an extraction
for 1 hour with and 1 hour without shaking (total 2 hour extraction). Normally the extraction
period should be continued for the desired period of contact which is 1 or 3 hours according
to the scenarios above. Since the EN 71-3 migration test is a rough and worst case simulation
of reality the 2 hours testing can be regarded as being sufficient for longer periods up to
3 hours. In this sense, the difference between the 1 and 3 hour period is not really relevant.

Note: The literature does not indicate any substantial difference between
mouthing durations for objects intended or unintended to be put in the mouth by
young children (chapter 3). Therefore, no distinction is made for toys intended or
not intended to be put in the mouth for toys intended for children aged 0-3 years.

As explained, this option of the proposed methodology can also be used to deduce migration
limit values for chemicals. By reversing the formula, migration limit values for elements can
be derived:

Bioaccessible amount of element (mg/mg toy), i.e. migration limit value =

X % TDI (mg/kg bw/day) x Body weight (kg)

Amount of toy ingested (mg/day)

RIVM report 320003001 page 117 of 234

This results in:

For toys intended for children 0-3 years: X% TDI x 7.5 kg / 8, 100 or 400 mg
For toys intended for children > 3 years: X% x TDI x 15 kg / 8 mg

To illustrate, the following example can be given. The proposed TDI for cobalt is
1.4 µg/kg bw/day. For the purpose of this illustration we assume that 10% of the TDI is used
as the health-based limit value. It is assumed that a small child ingests 8 mg of scraped off
toy material per day. The migration limit for cobalt in this material is:

10% x 1.4 x 10-3 mg/kg bw/day x 7.5 kg / 8 mg = 0.0001313 mg/mg toy or 131.3 mg/kg toy.

Analogously, migration limit values have been derived for all elements. These can be found
in paragraph 2.6. For illustrative purposes, the limits have been calculated based on three
different fractions of the TDI: 5, 10 and 20%. The ultimate fraction of the TDI that may be
allocated to exposure from toys should be decided by risk managers.

Note: In the current Directive, correction factors are included in the migration limit
values derived for elements in toys to account for variation in migration measurements.
We propose not to include correction factors in the migration limit values. Instead, we
propose to specify the variation in the migration measurements in the annex of the
standard. In our view (consumer protection) it is best to substract the correction factor
from the migration limit. This is however a risk management decision. On the other
hand, from the standpoint of enforcement purposes one has to prove the incompliance
with the migration limit. Then the correction factor should be used the other way
around.

8.5.2 Option 2: Use of product composition data

Within this option, the safety of a toy (material) with a certain element content as derived
from composition data of the material can be demonstrated by the following calculation:

Element in toy (mg/kg toy material) x weight of toy material present in toy (kg)
< X% TDI
Body weight child (kg)

It is assumed that all of the element is released at once from the product and available for
(oral) exposure. The bioaccessibility is thus 100%.

To illustrate the following (fictive) example is given.

page 118 of 234 RIVM report 320003001

The proposed TDI for cobalt is 1.4 µg/kg bw/day. For the purpose of this illustration we
assume that 10% of the TDI is used as the health-based limit value. Available data for the toy
material show that cobalt cannot be detected. The detection limit is reported to be
< 0.02 ng/kg product (fictive example). The toy material as present in the toy is 200 gram.

0.02 ng/kg toy material x 0.2 kg

= 5.3 x 10-4 ng/kg bw
7.5 kg

It is clear that the calculated amount is much lower than 10% of the TDI
(0.1 x 1400 ng/kg bw/day = 140 ng/kg bw/day). Therefore, this toy material will not be able
to provide an exposure level above 10% TDI and is therefore considered to be chemically
safe with respect to cobalt. No migration testing for this element in this toy is required.

Using the same principles, a content limit value for cobalt in the toy for this example can be
calculated by reversing the calculation:

Limit of element in toy (mg/kg toy material) =

10% TDI x Body weight child (kg)

weight of toy material present in toy (kg)

So: 10% x 1400 ng/kg bw/day x 7.5 kg / 0.2 kg = 5250 ng/kg toy material

8.5.3 Option 3: Use of risk based data

In option 1 it is assumed that the measured migration will occur daily. In reality this will not
be true since the EN 71-3 migration test uses an acidic test system which is worst case in the
sense that most of the element present in the tested matrix will be released in the first test. A
second extraction (e.g. day 2 of mouthing) will never release the same amount of elements.
Another worst case assumption is that all toy-related exposures occur with a daily frequency.
For some toys (e.g. finger paint) it is not necessary to use a daily frequency because normally
this product will be used with larger intervals.
Option 3 provides the opportunity to demonstrate the safety of a certain amount of element in
a toy by using a number of specific exposure factors and / or refined migration testing. The
health-based limit value to compare the exposure assessment to is still a fraction of the TDI,
being the maximal allowable oral exposure.
The exposure factors used for option 2 and other factors that can be used for the safety
assessment (such as frequency of exposure) are explained in chapter 3. This chapter also
gives guidance on how to evaluate other routes of exposure, which may be relevant for a
small group of specific toys (e.g. inhalation via sprays) or for certain elements (e.g. dermal
sensitisation of Ni).
RIVM report 320003001 page 119 of 234

The possibilities for migration testing are explained in more detail in chapter 4. This chapter
discusses in detail which type of migration tests are relevant for oral exposure routes
(mouthing and ingestion of toys).
As discussed for elements, the EN 71-3 test can be used as a worst case extraction system for
the ingestion scenario. If desired, further refinement regarding the estimation of the
bioavailability of substances from toys can be achieved by using a physiological based test
system e.g. as developed by RIVM.
If it can be demonstrated that no toy material will be ingested, a migration test with water can
be used as a representative test system for mouthing. In principle, the test system with
simulant A as used in the FCM framework can be used given that relevant temperatures and
test durations are used (chapter 4).
Other test systems used in the FCM framework cannot be used for toys because irrelevant
recipient fluids (e.g. oil), and conditions (static extraction versus dynamic extraction for toys)
are used. Such measurements provide data that cannot be used as representative for the
physiological condition of mouthing or ingestion.

8.6 Migration limits for elements in toys

In the tables below, migration limits for elements in toys are presented, as derived by the
methodology proposed in chapter 7. Further explanation and a calculation example can be
found in section 8.5.1.
page 120 of 234 RIVM report 320003001

Table 8-2 For intake of 8 mg (scraped off material) for children < 3 years of age
* Age < 3 years
* Body Weight 7.5 kg
* Material 8 mg (scraped off )

Element TDI (µg/kg Migration Limit value (mg/kg product) Current

bw/day) Migration
5% TDI 10% TDI 20% TDI Limit (mg/kg
product)
Aluminum 750
35156.3 70312.5 140625.0
Antimony 6
281.3 562.5 1125.0 60
Arsenic 1
46.9 93.8 187.5 25
Barium 600
28125.0 56250.0 112500.0 1000
Boron 160
7500.0 15000.0 30000.0
Cadmium 0.5
23.4 46.9 93.8 75
Chromiuma,d Cr3+
ws 5 234.4 468.8 937.5 60
3+
Cr
wis 5000 234375.0 468750.0 937500.0
(Cr
6+ b
) 5 234.4 468.8 937.5 60
(Cr
6+ c
) 0.0053 0.2 0.5 1.0 60
Cobalt 1.4
65.6 131.3 262.5
Copper 83
3890.6 7781.3 15562.5
Lead 3.6
168.8 337.5 675.0 90
Manganese 160
7500.0 15000.0 30000.0
Mercury 2
93.8 187.5 375.0 60
Nickel 10
468.8 937.5 1875.0
Selenium 5
234.4 468.8 937.5 500
Silver 5
234.4 468.8 937.5
Strontium 600
28125.0 56250.0 112500.0
Tin Inorganic
2000 92750.0 187500.0 375000,0
Organic 0.25 11.7 23.4 46,9
Zinc 500
23437.5 46875.0 93750.0
a
ws = water soluble, wis = water insoluble
b
Based on a TDI of 5 μg/kg bw derived for non-carcinogenic effects by hexavalent chromium
c
Based on a Virtually Safe Dose (VSD) of 0.0053 μg/kg bw/day derived for the genotoxic and
carcinogenic action by hexavalent chromium. As explained in the appended toxicological profile on
chromium, this VSD is based on a limited bioassay in mice and is fraught with additional uncertainty
compared to the usual bioassay-derived VSDs. Results of NTP studies now in progress should allow
more a reliable oral cancer risk estimation in the near future.
d
Measurement of Cr6+ is difficult. Further research is needed to derive safe migration limit values for
this element
RIVM report 320003001 page 121 of 234

Table 8-3 For intake of 100 mg (dry, brittle, powder-like or pliable material) for children < 3 years of
age
* Age < 3 yrs
* Body Weight 7.5 kg
* Material 100 mg (dry, powder like or pliable)

Element TDI (µg/kg Migration Limit value (mg/kg product) Current

bw/day) Migration Limit
5% TDI 10% TDI 20% TDI (mg/kg
product)*
Aluminum 750
2812.5 5625.0 11250.0
Antimony 6
22.5 45.0 90.0 60
Arsenic 1
3.8 7.5 15.0 25
Barium 600
2250.0 4500.0 9000.0 250
Boron 160
600.0 1200.0 2400.0
Cadmium 0.5
1.9 3.8 7.5 50
Chromiuma,d Cr3+
ws 5 18.8 37.5 75.0 25
3+
Cr
wis 5000 18750.0 37500.0 75000.0
(Cr
6+ b
) 5 18.8 37.5 75.0 25
(Cr
6+ c
) 0.0053 0.020 0.040 0.080 25
Cobalt 1.4
5.3 10.5 21.0
Copper 83
311.3 622.5 1245.0
Lead 3.6
13.5 27.0 54.0 90
Manganese 160
600.0 1200.0 2400.0
Mercury 2
7.5 15.0 30.0 25
Nickel 10
37.5 75.0 150.0
Selenium 5
18.8 37.5 75.0 500
Silver 5
18.8 37.5 75.0
Strontium 600
2250.0 4500.0 9000.0
Tin Inorganic
2000 7500.0 15000.0 30000.0
Organic 0.25 0.9 1.9 3.8
Zinc 500
1875.0 3750.0 7500.0
* Migration limits according EN 71-3 for modelling clay and finger paint
a
ws = water soluble, wis = water insoluble
b
Based on a TDI of 5 μg/kg bw derived for non-carcinogenic effects by hexavalent chromium
c
Based on a Virtually Safe Dose (VSD) of 0.0053 μg/kg bw/day derived for the genotoxic and
carcinogenic action by hexavalent chromium. As explained in the appended toxicological profile on
chromium, this VSD is based on a limited bioassay in mice and is fraught with additional uncertainty
compared to the usual bioassay-derived VSDs. Results of NTP studies now in progress should allow
more a reliable oral cancer risk estimation in the near future.
d
Measurement of Cr6+ is difficult. Further research is needed to derive safe migration limit values for
this element
page 122 of 234 RIVM report 320003001

Table 8-4 For intake of 400 mg (liquid or sticky material) for children < 3 years of age
* Age < 3 yrs
* Body Weight 7.5 kg
* Material 400 mg (liquid & sticky)

Element TDI (µg/kg Migration Limit value (mg/kg product) Current

bw/day) Migration Limit
5% TDI 10% TDI 20% TDI (mg/kg
product)*
Aluminum 750
703.1 1406.3 2812.5
Antimony 6
5.6 11.3 22.5 60
Arsenic 1
0.9 1.9 3.8 25
Barium 600
562.5 1125.0 2250.0 250
Boron 160
150.0 300.0 600.0
Cadmium 0.5
0.5 0.9 1.9 50
Chromiuma,d Cr3+
ws 5 4.7 9.4 18.8 25
3+
Cr
wis 5000 4687.5 9375.0 18750.0
(Cr
6+)b
5 4.7 9.4 18.8 25
6+)c
(Cr 0.0053 0.005 0.010 0.020 25
Cobalt 1.4
1.3 2.6 5.3
Copper 83
77.8 155.6 311.3
Lead 3.6
3.4 6.8 13.5 90
Manganese 160
150.0 300.0 600.0
Mercury 2
1.9 3.8 7.5 25
Nickel 10
9.4 18.8 37.5
Selenium 5
4.7 9.4 18.8 500
Silver 5
4.7 9.4 18.8
Strontium 600
562.5 1125.0 2250.0
Tin Inorganic
2000 1875.0 3750.0 7500.0
Organic 0.25 0.2 0.5 0.9
Zinc 500
468.8 937.5 1875.0
* Migration limits according EN 71-3 for modelling clay and finger paint
a
ws = water soluble, wis = water insoluble
b
Based on a TDI of 5 μg/kg bw derived for non-carcinogenic effects by hexavalent chromium
c
Based on a Virtually Safe Dose (VSD) of 0.0053 μg/kg bw/day derived for the genotoxic and
carcinogenic action by hexavalent chromium. As explained in the appended toxicological profile on
chromium, this VSD is based on a limited bioassay in mice and is fraught with additional uncertainty
compared to the usual bioassay-derived VSDs. Results of NTP studies now in progress should allow
more a reliable oral cancer risk estimation in the near future.
d
Measurement of Cr6+ is difficult. Further research is needed to derive safe migration limit values for
this element
RIVM report 320003001 page 123 of 234

Table 8-5 For intake of 8 mg (scraped off material) for toys intended to be mouthed by children > 3
years of age
* Age > 3 yrs
* Body Weight 15 kg
* Material 8 mg (scraped off)

Element TDI (µg/kg Migration Limit value (mg/kg product) Current

bw/day) Migration Limit
5% TDI 10% TDI 20% TDI (mg/kg product)
Aluminum 750
70312.5 140625.0 281250.0
Antimony 6
562.5 1125.0 2250.0 60
Arsenic 1
93.8 187.5 375.0 25
Barium 600
56250.0 112500.0 225000.0 1000
Boron 160
15000.0 30000.0 60000.0
Cadmium 0.5
46.9 93.8 187.5 75
Chromiuma,d Cr3+ ws
5 468.8 937.5 1875.0 60
3+
Cr
wis 5000 468750.0 937500.0 1875000.0
(Cr 6+)b 5 468.8 937.5 1875.0 60
6+)c
(Cr 0.0053 0.5 1.0 2.0 60
Cobalt 1.4
131.3 262.5 525.0
Copper 83
7781.3 15562.5 31125.0
Lead 3.6
337.5 675.0 1350.0 90
Manganese 160
15000.0 30000.0 60000.0
Mercury 2
187.5 375.0 750.0 60
Nickel 10
937.5 1875.0 3750.0
Selenium 5
468.8 937.5 1875.0 500
Silver 5
468.8 937.5 1875.0
Strontium 600
56250.0 112500.0 225000.0
Tin Inorganic
2000 187500.0 375000.0 750000
Organic 0.25 23.4 23.4 93.8
Zinc 500
46875.0 93750.0 187500.0
a
ws = water soluble, wis = water insoluble
b
Based on a TDI of 5 μg/kg bw derived for non-carcinogenic effects by hexavalent chromium
c
Based on a Virtually Safe Dose (VSD) of 0.0053 μg/kg bw/day derived for the genotoxic and
carcinogenic action by hexavalent chromium. As explained in the appended toxicological profile on
chromium, this VSD is based on a limited bioassay in mice and is fraught with additional uncertainty
compared to the usual bioassay-derived VSDs. Results of NTP studies now in progress should allow
more a reliable oral cancer risk estimation in the near future.
d
Measurement of Cr6+ is difficult. Further research is needed to derive safe migration limit values for
this element
 page 124 of 234 RIVM report 320003001

For a number of elements, the newly derived migration limit values may be lower than those
currently listed in the Toy Directive 88/378/EEC, depending on which fraction of the TDI
will be used. On the other hand, for some other elements, the newly derived migration limit
values are substantially higher and in some cases so high that it may be considered removing
these elements from the list of restricted elements.

8.7 Hazard aspects

Our methodology is a risk based approach in which no hazard aspects have been included. As
described in chapter 7, some hazard aspects may be included after careful re-evaluation. This
holds true also for elements.

8.8 Conclusions

• The proposed methodology can be used to derive migration and content limit values for
elements in toys and to perform safety assessments of elements in toys.
• For elements in toys, oral exposure is the main route contributing to systemic exposure.
• The choice of relevant elements is based on the list of elements currently included in the
Toy Directive, information on Food Contact Materials and information on elements
present in Plastic Waste. Despite repeated requests, information from the toy industry on
which elements are relevant for toy material was not received.
• New migration limit values have been derived for 16 elements, for toys intended for
children under 3 years of age and for toys intended to be mouthed by children over
3 years of age.
• For these toys, migration studies can be performed with an acid extraction fluid, as
described in the current EN 71-3 method.
• For toys intended for children under 3 years of age, three different toy categories have
been distinguished for which separate migration limit values have been derived,
depending on type of toy material
• The migration limit values derived have been calculated based on 5, 10 and 20% of the
TDI. The ultimate migration limit values need to be adjusted based on the decision by risk
managers on which fraction of the TDI is allocated for toys.
• Some newly derived migration limit values are lower than those currently listed in the
Toy Directive, others are much higher.

8.9 Recommendations

• Risk managers need to decide which fraction of the TDI may be allocated for toys.
RIVM report 320003001 page 125 of 234

• The proposed migration limit value for Cr6+ is uncertain and needs to be investigated
further.
• The provisions of the EU legislation on Nickel (Commission Directive 2004/96/EC,
amendment to Commission Directive 1994/27/EEC) should be adopted for toys.
• The list of restricted elements may be shortened by removing those elements for which
very high migration limit values have been derived.
• If the toy industry can provide actual data indicating many elements never occur in toys
or do not migrate at all, then these data are highly relevant and should be submitted and
considered.
page 126 of 234 RIVM report 320003001
RIVM report 320003001 page 127 of 234

References

ATSDR (2005), Toxicological profile for barium. US Agency for Toxic Substances
and Disease Registry, report dated September 2005.

Babich MA, Chen SB, Greene MA, Kiss CT, Porter WK, Smith TP, Wind ML, Zamula WW
(2004) Risk assessment of oral exposure to diisononyl phthalate from children's
products. Regul Toxicol Pharmacol 40: 151-67.

Bosgra S, Bos PM, Vermeire TG, Luit RJ, Slob W (2005) Probabilistic risk characterization:
an example with di(2-ethylhexyl) phthalate. Regul Toxicol Pharmacol 43: 104-13.

Bouma K and Reus HR (2004) Market surveillances on toy safety. Report VWA
ND04o063/01, Food and Consumer Product Safety Authority, Groningen, The
Netherlands.

Bouma K, Dannen F, Peursem AJJ van, Steendam L, Nieman CIHM, Douwes BA,
Schakel DJ. Chemische veiligheid van scoubidou touwtjes. Report VWA
ND04o061/01, Food and Consumer Product Safety Authority, Groningen, The
Netherlands.

Brandon EF, Oomen AG, Rompelberg CJM, Versantvoort CHM, van Engelen JGM, Sips
AJAM (2006) Consumer product in vitro digestion model: Bioaccessibility of
contaminants and its application in risk assessment. Regul Toxicol Pharmacol 44:
161-71.

Bremmer HJ, Prud'homme de Lodder LCH, Van Engelen JGM (2006) General fact sheet:
limiting conditions and reliability, ventilation, room size, body surface area. Updated
version for ConsExpo 4. Report 320104 002/2006, National Institute for Public
Health and the Environment, Bilthoven, The Netherlands.

Bremmer HJ and Van Veen MP (2002) Children’s toys fact sheet. Report 612810012,
National Institute for Public Health and the Environment, Bilthoven, The Netherlands.

Cázares-Delgadillo J. et al. (2005). Skin permeation enhancement by sucrose esters: a pH-

dependent phenomenon. Int J Pharm. Jun 13;297(1-2):204-12

Charman WN, Porter CJH, Mithani S, Dressman JB (1997) Physicochemical and

physiological mechanisms for the effects of food on drug absorption: the role of
lipids and pH. J. Pharm. Sci. 86: 269-282.

Chiou WL (2001) The rate and extent of oral bioavailability versus the rate and extent of oral
absorption: clarification and recommendation of terminology. Journal of
page 128 of 234 RIVM report 320003001

Pharmacokinetics and Pharmacodynamics 28: 3-6.

Clarkson TW (1993) Molecular and ionic mimicry of toxic metals. Annu. Rev. Pharmacol.
Toxicol. 32: 545-571.

Commission of the European Communities (1985) Environment and quality of life. Scientific
advisory committee to examine the toxicity and ecotoxicity of chemical compounds.
EUR 12964.

Commission Directive 2002/72/EC of 6 August 2002 relating to plastic materials and articles
intended to come into contact with foodstuffs. Official Journal L 220, 15/8/2002 P
18–58.

Consument en Veiligheid. Eye injury in 0-9 year’s old children by chemicals in toys in The
Netherlands. 2001 March.

Copoví A. et al. (2006). Enhancing effect of alpha-hydroxyacids on "in vitro" permeation

across the human skin of compounds with different lipophilicity. Int J Pharm. May
11;314(1):31-6

Council Directive 67/548/EEC of 27 June 1967 on the approximation of laws, regulations and
administrative provisions relating to the classification, packaging and labelling of
dangerous substances. Official Journal 196, 16/08/1967 P. 0001 – 0098.

Council Directive 88/378/EEC of 3 May 1988 on the approximation of the laws of the
Member States concerning the safety of toys. Official Journal L 187, 16/7/1988, P 1–
13.

Danish EPA (1998) Limits for the migration of certain elements from toy materials,
Teknologisk Institut, Denmark.

Danish EPA (2005) Survey and release of chemical substances in "slimy" toys.

De Groot M.E. et al. (1998). Mouthing behaviour of young children – An observational

study. Agricultural University Wageningen, Household and Consumer Studies.
Wageningen, The Netherlands.

Delmaar JE, Park MVDZ, Van Engelen JGM (2005) ConsExpo 4.0 Consumer Exposure and
Uptake Models Program Manual . Report 320104004. National Institute for Public
Health and the Environment, Bilthoven, The Netherlands.

Diamond GL, Goodrum PE, Felter SP, Ruoff WL (1997) Gastrointestinal absorption of
metals. Drug Chem. Toxicol. 20: 345-368.

DTI. Research into the mouthing behaviour of children up to 5 years old. London: DTI;
RIVM report 320003001 page 129 of 234

2002

EFSA (2004a) Opinion of the Scientific Panel on Dietetic Products, Nutrition and Allergies
on a request from the Commission related to the Tolerable Upper Intake Level of
Boron (Sodium Borate and Boric Acid). EFSA Journal 80, 1-22.

EFSA (2004b) Opinion of the Scientific Panel on Contaminants in the Food Chain on a
request from the Commission to assess the health risks to consumers associated with
exposure to organotins in foodstuffs (Question N° EFSA-Q-2003-110) Adopted on 22
September 2004. The EFSA Journal (2004) 102, 1-119.

Ehling G, Hecht M, Heusener A, Huesler J, Gamer AO, van Loveren H, Maurer T, Riecke K,
Ullmann L, Ulrich P, Vandebriel R, Vohr HW (2005a) An European inter-laboratory
validation of alternative endpoints of the murine local lymph node assay: first round.
Toxicology 212: 60-68.

Ehling G, Hecht M, Heusener A, Huesler J, Gamer AO, van Loveren H, Maurer T, Riecke K,
Ullmann L, Ulrich P, Vandebriel R, Vohr HW (2005) An European inter-laboratory
validation of alternative endpoints of the murine local lymph node assay: 2nd round.
Toxicology 212: 69-79.

EN 12586:1999/C1:2002. Child care articles - Soothers holder - Safety requirements and test
methods.

EN 12868:1999/C1:2002. Child use and care articles - Methods for determining the release of
N-Nitrosamines and N-Nitrosatable substances from elastomer or rubber teats and
soothers.

EN 1400-3:2002. Child use and care articles - Soothers for babies and young children- Part 3:
Chemical requirements and tests.

EN 14350-2:2004. Child use and care articles - Drinking equipment - Part 2: Chemical
requirements and tests.

EN 71-10:2006. Safety of toys - Part 10: Organic chemical compounds - Sample preparation
and extraction.

EN 71-11:2005. Safety of toys - Part 11: Organic chemical compounds - Methods of

analysis.

EU (2005) Provisional list of monomers and additives notified to European Commission as

substances which may be used in the manufacture of plastics or coatings intended to
come into contact with foodstuffs (updated to June 2005). SANCO D3/AS D(2005)
File:SYNOPTIC DOCUMENT(2005.07.25). Available
at:http://ec.europa.eu/food/food/chemicalsafety/foodcontact/synoptic_doc_en.pdf
page 130 of 234 RIVM report 320003001

EU-RAR (2005) European Union Risk Assessment Report volume 53: chromium trioxide,
sodium chromate, sodium dichromate, ammonium dichromate, potassium dichromate.
Institute for Health and Consumer Protection, European Chemical Bureau, Existing
Substances. European Commission, Joint Research Centre EUR 21508 EN.

European Committee for Standardization (CEN) (2002) Classification of toys - Guidelines .

CR 14379:2002 (E).

European Committee for Standardization (CEN) , (2004), EN 71-9 Safety of toys - Part 9:
Organic chemical chemicals - Requirements (final draft, prEN 71-9).

European Committee for Standardization (CEN), EN 71-3 Safety of toys - Part 3: Migration
of certain elements. CEN, Brussels, Belgium, 94.

European Committee for Standardization (CEN) , Directive 88/378/EEC; Council Directive

on the approximation of the laws of the Member States concerning the safety of toys,
Brussels, Belgium, 88

European Committee for Standardization (CEN), (2003), Final report of the work of CEN/TC
52/WG9.

Fiala F, Steiner I, Kubesch K (2000) Migration of di-(2-ethylhexyl)phthalate (DEHP) and

diisononyl phthalate (DINP) from PVC articles. Deutsche Lebensmittel-Rundschau
96: 51-57.

Florin TA, Brent RL, Weitzman M (2005) The need for vigilance: the persistence of lead
poisoning in children. Pediatrics. 115(6):1767-8.

Fowles GWA (1977) The leaching of cadmium from plastic toys. The Science of The Total
Environment 7, 207-216

Frederiks AM, van Buuren S, Burgmeyer RJF, Meulmeester JF, Beuker RJ, Brugman E,
Roede MJ, Verloove-Vanhorick SP, Wit JM (2002) Continuing positive secular
growth change in the Netherlands. Pediatric Research 47: 316-323.

Fullmer CS (1992), Intestinal interactions of lead and calcium. NeuroToxicology 13: 799-
808.

González-Soto E, González-Rodríguez V, López-Suárez C, Castro-Romero JM, Pérez-

Iglesias J, Fernández-Solís JM (2000) Migration of lead and cadmium from ceramic
materials used in food preparation. Bull. Environ. Contam. Toxicol. 65: 598-603.

Greene, M A Statistical methods for prediction of DINP intake by young children.

Washington, DC: U.S. Consumer Product Safety Commission (CPSC); 1998 Dec.
RIVM report 320003001 page 131 of 234

Griem, P, Goebel, C, Scheffler, H (2003) Proposal for a risk assessment methodology for
skin sensitization based on sensitization potency data. Regul Toxicol Pharmacol 38:
269-290.

Heard C.M. et al. (2006). Skin penetration enhancement of mefenamic acid by ethanol and
1,8-cineole can be explained by the ‘pull’ effect. Int J Pharm Sept. 14; 321(1-2):167-
70.

Hostýnek JJ et al. (1993). Metals and the skin.Crit Rev Toxicol.23(2):171-235.

Iliano B, Viaene M, Oudar AM (1988) Analysis of toys: 1. Lead, cadmium and barium
migration. 2. Migration of coloring agents in artificial saliva. Arch Belg 46: 336-46.

IPCS (2003) Concise International Chemical Assessment Document 50: Elemental mercury
and inorganic mercury compounds: human health aspects. World Health
Organization, Geneva, 2003.

JECFA (1993) Evaluation of certain food additives and contaminants. Fourty-first report of
the Joint FAO/WHO expert Committee on Food Additives. WHO technical report
series no. 837. World Health Organization, Geneva, Switzerland

Jowsey IR, Basketter DA, Westmoreland C, Kimber I. (2006) A future approach to

measuring relative skin sensitising potency: a proposal. J Appl Toxicol 26: 341-350.

Juberg DR et al. (2001) An observational study of object mouthing behavior by young

children. Pediatrics. Jan;107(1):135-42.

Könemann WH (1998). Phthalate release from soft PVC baby toys. Report from the Dutch
Consensus Group. Report 613320002. National Institute for Public Health and the
Environment, Bilthoven, The Netherlands.

Maas RP, Patch SC, Pandolfo T J, Druhan JL, Gandy NF (2005) Lead content and exposure
from children's and adult's jewelry products. Bull Environ Contam Toxicol.
74(3):437-44.

Mushak P (1991) Gastro-intestinal absorption of lead in children and adults: overview of

biological and biophysico-chemical aspects. Chem. Speciation Bioavailability 3: 87-
104.

NRC (1993) (National Research Council) Pesticides in the Diets of Infants and Children.
National Academy Press, Washington DC.

OEHHA (1999) Public health goal for chromium in drinking water. Pesticide and
Environmental Toxicology Section Office of Environmental Health Hazard
Assessment California Environmental Protection Agency. Dated February 1999
page 132 of 234 RIVM report 320003001

OEHHA (2004) Development of health criteria for school site risk assessment pursuant to
health and safety code section 901(g): Proposed child-specific Reference Dose
(ChRD) for school site risk assessment - Manganese and Pentachlorophenol. Public
Review Draft November 2004. Integrated Risk Assessment Section Office of
Environmental Health Hazard Assessment, California Environmental Protection
Agency.

Oomen AG (2000) Determinants of oral bioavailability of soil-borne contaminants. Thesis,

Utrecht University, Utrecht.

Oomen AG, Brandon EFA, Swartjes FA, Sips AJAM (2006) How can information on oral
bioavailability improve human health risk assessment for lead-contaminated soils?
Implementation and scientific basis. Report 711701042, National Institute for Public
Health and the Environment, Bilthoven, The Netherlands.

Oomen AG, Hack A, Minekus M, Zeijdner E, Cornelis C, Schoeters G, Verstraete W, Wiele

TVd, Wragg J, Rompelberg CJM, Sips AJAM, Wijnen JV (2002) Comparison of five
in vitro digestion models to study the bioaccessibility of soil contaminants. Environ.
Sci. Technol. 36: 3326-3334.

Oomen AG, Rompelberg CJM, Brandon EFA, Van de Kamp E, Duits MR, Versantvoort
CHM, Van Engelen JGM, Sips AJAM (2005) Consumer Product in vitro digestion
model: bioaccessibility of contaminants from toys and application in risk assessment.
Report 320102004, National Institute for Public Health and the Environment,
Bilthoven, The Netherlands.

Oomen AG, Rompelberg CJM, Bruil MA, Dobbe CJG, Pereboom DPKH, Sips, AJAM
(2003a) Development of an in vitro digestion model for estimation of bioaccessibility
of soil contaminants. Arch. Environ. Contam. Toxicol. 44: 281-287.

Oomen AG, Van Twillert K, Hofhuis MFA, Rompelberg CJM, Versantvoort CHM (2003b)
Development and suitability of in vitro digestion models in assessing bioaccessibility
of lead from toy matrices. Report 320102001, National Institute for Public Health and
the Environment, Bilthoven, The Netherlands.

Oomen AG, Versantvoort CHM, Duits MR, Van de Kamp E, Van Twillert K (2004)
Application of in vitro digestion models to assess release of lead and phthalate from
toy matrices and azo dyes from textile. Report 320102003, National Institute for
Public Health and the Environment, Bilthoven, The Netherlands.

Otte PF, Lijzen JPA, Otte JG, Swartjes FA, Versluijs CW (2001) Evaluation and revision of
the CSOIL parameter set. Proposed parameter set for human exposure modelling and
deriving Intervention Values for the first series of compounds. Report 711701021,
National Institute for Public Health and the Environment, Bilthoven, The Netherlands.
RIVM report 320003001 page 133 of 234

Panigrahi L. et al. (2005). The effect of pH and organic ester penetration enhancers on skin
permeation kinetics of terbutaline sulfate from pseudolatex-type transdermal delivery
systems through mouse and human cadaver skins. AAPS PharmSciTech. Sep
30;6(2):E167-73

Reed KJ, Jimenez M, Freeman NC, Lioy PJ (1999) Quantification of children's hand and
mouthing activities through a videotaping methodology. J Expo Anal Environ
Epidemiol. 9(5):513-20.

RIVM (1998) Maximum Permissible Risk Levels for human intake of soil contaminants:
fourth series of compounds. RIVM report no. 715810004, dated March 1998.

RIVM (2001) Re-evaluation of human-toxicological Maximum Permissible Levels. RIVM

report no. 711701025, dated March 2001

RIVM (2006) Briefrapport Kunstofafval. On behalf of VROM-Inspectie Regio Noord-West.

RIVM Centre for Inspectorate research, Emergency Response and Drinking-Water
(IMD). Report dated 27 February 2006. [In Dutch]

SCF (2000) Opinion of the Scientific Committee on Food on the Tolerable Upper Intake
Level of Selenium (expressed on 19 October 2000). Scientific Committee Food,
SCF/CS/NUT/UPPLEV/25 Final, 28 November 2000.

SCF (2003a) Opinion of the Scientific Committee on Food on the Tolerable Upper Intake
Level of Copper (expressed on 5 March 2003). Scientific Committee on Food,
SCF/CS/NUT/UPPLEV/57 Final. Dated 27 March 2003.

SCF (2003b) Opinion of the Scientific Committee on Food on the Tolerable Upper Intake
Level of Zinc (expressed on 5 March 2003). SCF/CS/NUT/UPPLEV/62 Final 19
March 2003.

Schümann K, Elsenhans B, Gordon D (1994) Different aspects of bioavailability in

perspective. British Journal of Nutrition 72: 950-952.

Scientific Committee on toxicity, ecotoxicity and the environment CSTEE, Opinion of the
CSTEE on the assessment of the bioavailability of certain elements in toys. European
Commission, Health and Consumer Protection Directorate-General, Brussels,
Belgium, 2004.

Simoneau C, Geiss H, Roncari A, Zocchi P, Hannaert P (2001) Validation of methodologies

for the release of di-isononylphthalate (DINP) in saliva simulant from toys, European
Commission, DG - Joint Research Centre, Food Products Unit, Institute for Health
and Consumer Protection, Ispra, Italy.
page 134 of 234 RIVM report 320003001

Smith SA and Norris, B. (2003) Reducing the risk of choking hazards: mouthing behaviour of
children aged 1 month to 5 years. Inj Control Saf Promot. 10(3):145-54.

Steiner I, Scharf L, Fiala F, Washüttl J (1998) Migration of di-(2-ethylhexyl)phthalate from

PVC child articles into saliva and saliva simulant. Food Additives and Contaminants
15: 812-817.

Stopford W (2003) Aerosol Production During the Use of Art and Craft Materials, Duke
University Medical Center, Durham, NC.

Teknologisk Institut. Limits for the migration of certain elements from toy materials. Report
Teknologisk Institut, Taastrup (DK). 1998.

Tulve NS, Suggs JC, McCurdy T, Cohen Hubal EA, Moya J (2002) Frequency of mouthing
behaviour in young children. J Expo Anal Environ Epidemiol. 12(4):259-64.

U.S. Consumer Product Safety Commission Report from the Chronic Hazard Advisory Panel
on Diisononyl Phthalate, June 2001.

U.S. Consumer Product Safety Commission (CPSC) (2002) Age Determination Guidelines:
Relating Children’s Ages To Toy Characteristics and Play Behaviour.

US-EPA (1996a) Oral RfD Assessment – Silver. Last revised 12/01/1996.

US-EPA (1996b) IRIS-file Strontium. Derivation of RfD, last revised 12-1-1996.

US-EPA (2002) Child-Specific Exposure Factors handbook Volume I-III. EPA/600/P-

00/002B. September 2002. US Environmental Protection Agency, Office of Research
and Development, National Center for Envorinmental Assessment, Washington DC.

Van Arsdale JL, Leiker RD, Kohn M, Merritt TA, Horowitz BZ (2004) Lead poisoning from
a toy necklace. Pediatrics. 114(4):1096-9.

Van Engelen JGM and Prud'homme de Lodder LCH (2004) Non-food products: How to
assess children's exposure? Report 320005001, National Institute for Public Health
and the Environment (RIVM), Bilthoven, the Netherlands.

Versantvoort CH, Oomen AG, Van de Kamp E, Rompelberg CJM, Sips AJAM (2005)
Applicability of an in vitro digestion model in assessing the bioaccessibility of
mycotoxins from food. Food Chem Toxicol 43: 31-40.

Versantvoort CHM, Van de Kamp E, Rompelberg CJM (2004) Development of an in vitro

digestion model to determine the bioaccessibility of contaminants from food. Report
320102002, National Institute for Public Health and the Environment, Bilthoven, The
Netherlands.
RIVM report 320003001 page 135 of 234

VWA (2005) Migration of bisphenol-A and plasticisers from plastic feeding utensils for
babies. Report VWA ND050410. Food and Consumer Product Safety Authority,
Groningen, The Netherlands.

VWA (2005) Screening of plastic toys for chemical composition and hazards. Market
surveillance in the Netherlands. Report VWA ND05o610/01. Food and Consumer
Product Safety Authority, Groningen, The Netherlands.

WHO (2003) Antimony in Drinking-water. Background document for development of

WHO Guidelines for Drinking-water Quality. WHO, Geneva, 2003. WHO/JECFA
(2001) WHO Food Additives Series no. 46.

Wienk KJH, Marx JJM, Beynen AC (1999). The concept of iron bioavailability and its
assessment. Eur. J. Nutr. 38: 51-75.
page 136 of 234 RIVM report 320003001
RIVM report 320003001 page 137 of 234

APPENDIX

I Answers to issues raised by CSTEE in their opinion

and DG Enterprise in their call

The opinion of the CSTEE on “Assessment of the bioavailability of certain elements in toys”
(opinion adopted on June 22, 2004) distinguished two main topics, i.e. 1) suitability of the
proposed limit values and 2) the necessity of updating the standard EN 71-3:1994.

Issue: Choice of elements

Background: Directive 88/3/378 focuses on an-organic compounds, i.e. elements as the
compounds of interest. Therefore organic compounds will in principle not be taken into
account in this report. The toxicological profile of the elements presently taken up in the
Directive has been reviewed. It was suggested by CSTEE that limit values for other elements
than those already listed might be needed.
Answer: Since the majority of testing for enforcement and quality assurance only consider
the eight elements listed at present in the Toy Directive, it was very difficult to obtain
information on the presence of additional elements (to the current list) for different toy
material. Therefore, as a basis for the selection of additional elements the Synoptic
Document (EU, 2005) from the food contact material framework and a study on the use and
waste-disposal of synthetic materials was consulted.
The following elements were selected for review:

Table I-1 Elements selected for review

Aluminium Lead
Antimony Manganese
Arsenic Mercury
Barium Nickel
Boron Selenium
Cadmium Silver
Chromium Strontium
Cobalt Tin (organic and inorganic)
Copper Zinc
Molybdeen, Titanium, and Zirkonium were excluded from the list on the basis of their
toxicological profile

Issue: Use of Health-based limit values

Background: CSTEE asked to include the latest scientific knowledge and associated
revisions of tolerable daily intakes (TDI) and average daily intakes (ADI) into the newly
proposed limit values. In this respect it was necessary that the ADI’s and TDI’s of the
involved elements set by various organizations were compared and reviewed for the latest
page 138 of 234 RIVM report 320003001

updates. In addition, an additional literature search from the date of the latest update found
can be used to identify potential new data published. Special focus was paid to the setting of
TDI’s with respect to the potential sensitivity of children.
Answer: The toxicity of the elements presently in the Toy Directive and the above listed
additional elements have been reviewed. The strategy for the derivation of health-based limit
values (TDI/ADI) is presented in chapter 2, the toxicological profiles of the individual
elements are presented in Appendix 2.

Issue: Intake of toy-material

Background: The CSTEE considers the presumption that an average daily intake of 8 mg of
toy material could be expected as incorrect. It might be more realistic that children may
ingest much more than 8 mg in one day, for instance through ingestion of some liquid toy
material. Although for some forms of toys this may be true, for other forms 8 mg per day may
be an overestimation.
Answer: This issue is discussed in chapter 3. It is concluded that for a risk-based safety
assessment classification of toys based on exposure categories is most appropriate. For toys
for children <3 of age default intake values are proposed for solid (easily to break or bit),
liquid or sticky material and for material to be scraped off. For mouthing, default values are
proposed for mouthing times for children <3 and for children >3y of age.

Issue: Definition for bioavailability

Background: The CSTEE recommends to change the definition for bioavailability from “the
soluble extract having toxicological significance” into “the amount of each element in the toy
which could be absorbed into the systemic circulation of a child”.
Answer: It is important to note that these definitions point at different entities in the process
of (oral) bioavailability. The way in which oral bioavailability is presently defined is conform
the definition applied in pharmaceutical/pharmacokinetic sciences “the fraction of a
substance present in toy material that reaches the systemic circulation (of a child)”. RIVM
interprets the present definition as stated in Council Directive 88/378/EEC as in compliance
with bioavailability (F). However, the newly proposed definition by CSTEE is in agreement
with the factor Fb, i.e. the bio-accessible fraction, which is a sub-process of oral
bioavailability. In chapter 4 the pro’s and con’s of applying either F or Fb , and relationship
between F and Fb are discussed.

Issue: Use of an analytical correction factor

Background: To derive the final limit values, the CEN in their previous attempt made use of
some kind of analytical correction factor. Just as CSTEE recognized it is not clear where this
additional factor is based on.
Answer: In EN 71-3 the analytical variation (correction values) is used to increase the
limits. The limits for elements that are proposed in the present report are based on
toxicological concepts, which means safety limits to ensure the health of children. It is
therefore suggested to include the precision data of EN 71-3 in an Annex of this standard and
RIVM report 320003001 page 139 of 234

not to prescribe how the obtained results must be corrected or interpreted, as this depends on
the perspective of the test laboratory.
In our opinion analytical variation should be dealt with from different perspectives:
When a test laboratory wants to certify a test sample, they have to demonstrate compliance
by proving that the migration value is below the legal limit. Therefore the found migration
value including analytical variation may not exceed the migration limit.
When an enforcement laboratory analyses a sample, they must demonstrate that the sample
exceeds the legal limit, before they can take an official measure. The migration value is
corrected by subtracting the analytical variation.
As the analytical correction values used in EN 71-3 (30-60%) are high, and there are
probably possibilities to improve the interlaboratory variability of the test, part of the
problem may be solved by improvement of the method and re-evaluation of the analytical by
a new ring trial.
Further details on this issue can be found in chapter 6.

Issue: A single representative or not ?

Background: The CSTEE does not accept that it is possible to take a single representative
from many toys because of their heterogeneous nature. Sampling is a critical step in the
enforcement and testing for compliance, that is often overlooked.
Answer: In our opinion a single sample could be used for compliance testing. For
enforcement laboratories measures can be taken based on the results of a single sample. Toy
producers or importers must ensure that the sample used for compliance testing is
representative for what they place on the market. Periodic testing is required if there are
relevant changes in production circumstances, raw materials or in the standards. We
propose that this test certificate may not be older than 5 year before the date of marketing of
the toy. Moreover, all accessible parts of a toy must comply with EN 71-3. If a toy consists of
different materials, subsamples should be made of each material. (see chapter 6).

Issue: Limit values and maximum bioavailability

Background: The present limit values for concentrations in toy materials are more or less
based on the maximum bioavailability. In the methodology used by the CEN conversion
factors were applied to some metals to derive limit values, but not to others. CSTEE stated
these factors to be unclear. The methodology used by CEN is evaluated in order to clarify the
use of the conversion factors.
Answer: Bioavailability of elements from toy is mainly focussed on oral bioavailability.
Limit values for the bioavailability of several elements from toy are listed in Council
Directive 88/378/EEC. The scientific background of these values are given in a report of the
Scientific Advisory Committee (Report EUR 12964). Bioavailability values are used in this
EUR 12964 report as it is stated that “bioavailability values are used as the bioavailability of
toxic or harmful substances is more important than the total content of potentially dangerous
substances in the toy”. The bioavailability may be investigated by the extraction rate with
media similar to saliva or gastric juice, thereby assuming that the exposure occurs via the
page 140 of 234 RIVM report 320003001

oral route. Furthermore, report EUR 12964 gives a toxicological evaluation of the elements,
with particular attention paid to the gastrointestinal absorption and to data on toxicokinetics
in children. Furthermore, it is stated that the intake of the elements from toys should not
exceed 10% of the total intake of these metals by children.
The thus obtained bioavailability limits are translated in EN 71-3 to migration limits in which
the migration is determined in a hydrochloric acid solution (pH 1-1.5), thereby simulating
the acid environment of the human stomach (for fasting conditions). For the conversion of
bioavailability limits in Council Directive 88/378/EEC to migration limits in EN 71-3, it is
assumed that a child daily ingests 8 mg toy. In addition, for barium and selenium a lower
migration limit was derived to minimise exposure of children, and for antimony, arsenic and
chromium a higher migration limit was derived to ensure analytical feasibility. Further
details can be found in chapter 4.

Issue: Bioavailability or migration

Background: One of the purposes of the report is to examine whether the limit values for
elements should be expressed in terms of bioavailability or in terms of migration.
In line with this discussion is the discussion on what type of extraction methods the limit
values should be based. Currently, migration is assessed on the basis of chemical extraction
procedures. Especially for elements released from toy parts which are ingested, this type of
test might overestimate the release substantially. In addition to chemically extraction methods
now also physiologically based extraction procedures are available, also for toys (Oomen et
al., 2004).
Answer: It was evaluated whether chemically (migration) or physiologically (bioavailability)
based methods would be preferred taking into account both scientific credibility and
analytical ease (chapter 4). The general methodology proposed offers the application of
physiologically based methods, but at this moment chemically based methods are preferred.
It is recognized that the potential of physiologically based methods is high, but these methods
require further validation before they can be applied on a larger scale.

Issue: Toys intended for different ages

Background: One of the discussion points identified by CSTEE is whether the age limit of
6 years of age as used in the current EN71-3 standard is appropriate, as it is “foreseeable that
children under 6 will have access to toys intended for children over 6. In addition, the CSTEE
noted that the age limit of 6 years is in contradiction with the Toy Directive 88/378/EEC.
Answer: In chapter 3, the issue of age related exposure has been discussed extensively. In
summary, we agree with the CSTEE that toys intended for older children may be accessible
to younger children. In the current Toy Directive, a distinction is made between toys intended
for children under three years of age and other toys with regard to safety aspects. These
safety aspects relate mostly to labelling toys that may pose choking and strangulation
hazards, due to small parts and long cords, respectively. With regard to toys containing these
physical or mechanical hazards, some degree of carer responsibility may be assumed, as it is
more or less common knowledge that such toys should be kept out of reach from young
RIVM report 320003001 page 141 of 234

children displaying mouthing behaviour. On the other hand, for invisible hazards such
chemical hazards, this is not always clear. We therefore recommend to include exposure
scenarios specific for young children (mouthing, ingestion, crawling) in the exposure
assessment for chemicals in all toys that can be placed in the mouth or crawled on,
regardless of intended age group, unless toys are clearly unsuitable for young children based
on physical or mechanical hazards.

Issue: Food Contact Materials

Background: It was proposed to examine whether the food contact materials (FCM)
framework could provide a basis for setting limit values in toys.
Answer: It may be possible that substances with low migration in the FCM framework
(< 0.05 mg/kg food or fluid) may be directly allowable in toys without further testing because
rough calculations indicate migration values in the same order of magnitude as calculated
for toys when expressed in mg/kg toy material. However, this can only be allowed when it is
assured that the finished toy material / matrix is similar to that tested in the FCM framework
and when the testing conditions are relevant for toy exposure. There are however some
uncertainties because FMC involves passive migration while mouthing involves active
migration. Furthermore, the FCM concept allows exposure that may be higher than the
fraction of TDI that is allowable for toys (at least for elements, see chapter 8). Therefore, this
approach should only be used after sufficient experimental validation data become available
showing that such an approach is indeed safe. For the time being, we recommend not to
extrapolate FCM migration limits to toys.

Issue: Analytical test methods

Background: It is stated that corresponding analytical test methods should be available. In
our point of view it is necessary to have a closer look at the available European Standard
(EN71-3: 1994, BS 5665-3: 1995 Specification for Migration) and the ISO Standard (ISO
8124-3: 1997 Migration of Certain Elements) for toys. Additionally the complementary
Standards for food contact materials have to be taken into account.
Answer: In chapter 6 several analytical issues are addressed, such as analysis of
chromium3+ /chromium6+ and organic tin compounds, repetitive versus single testing, the use
of reference materials etc. It appears that in the migration test (EN 71-3) chromium6+ cannot
be detected as the acidic solution reduces it to chromium3+. Further validation for the
measurement of organic tin is required. Conform testing of childcare articles it is adviced to
take the results of the first migration test in to account for compliance testing. Further
validation of methods of analysis of the newly suggested elements will be required.
page 142 of 234 RIVM report 320003001
RIVM report 320003001 page 143 of 234

II Toxicological Profiles

II.1 Aluminium

Aluminium and aluminium compounds have been evaluated within the scope of the WHO
Drinking-water guidelines in 1996 and 1998. WHO/IPCS has published an Environmental
Health Criteria on aluminium in 1997 (IPCS 1997). Further reviews are those by US-ATSDR
(1999) and OEHHA (2000).

II.1.1 Normal exposure

Aluminium is the most prevalent metal in the earth’s crust, of which it constitutes no less
than 8%. It occurs as silicates, oxides and hydroxides, combined with other elements such as
sodium and fluoride or as a complex with organic material. Aluminium is present in drinking-
water at concentrations of up to 0.2 mg/L. In many countries non-occupational exposure for
adults is between 2.5 and 13 mg aluminium/day from air, water and food (equal to 0.08-0.18
mg/kg body weight per day for a 60-kg individual). However, large variations in daily intake
can occur as a consequence of differing intakes of foods containing commonly encountered
food additives. Adults taking antacid medication in which aluminium is present, may have
very high intakes (up to 5000 mg/day). Infant intakes from nutritional formulas as determined
in the USA, Canada and the UK range from 0.03 to 0.7 mg/day. Formulas based on cow’s
milk are lower in aluminium content than those based on soy. When infant formulas contain
aluminium as a food additive, higher intakes are possible (IPCS, 1997; WHO, 1998). A
study carried out in Canada in 1999-2001 showed a decline in aluminium concentrations in
infant formulas compared to levels found in 1992 (Health Canada, 2003). For older infants a
1995 German study found a daily intake of 0.78 mg (5-8 years old) whereas a US study from
the same year reports 6.5 mg/day for 6 year-olds (IPCS, 1997).

Based on these data background maximum daily intake by young children is estimated at
0.3 mg/kg bw/day.

II.1.2 Toxicology
The primary target organs for aluminium toxicity are the central nervous system and the
skeleton. In animals encephalopathy was observed including histopathological effects in
brain cells. These findings have added relevance to suggestions that the presence of
aluminium in drinking-water constitutes a risk factor for the development of Alzheimer’s
disease in humans. In a large number of epidemiological studies this possible connection has
been studied. From this body of data WHO (1998) in its programme for drinking-water
page 144 of 234 RIVM report 320003001

guidelines concluded that a causal link is unlikely but nevertheless cannot be ruled out
entirely. Within medical practice aluminium is regarded as a causative factor in
encephalopathy as observed in certain hemodialysis patients (‘dialysis encephalopathy’).
Some non-dialysed cases with the same condition were children with renal failure.
Neurotoxicity has also been observed in premature infants receiving intravenous feeding-
solutions (for further discussion on this see below). As already remarked, neurotoxicity has
also been observed in experimental animals. In rats and mice neurobehavioural effects in
adult and developing animals have been observed after oral application. ATSDR concludes
neurotoxicity is the critical end point of concern for aluminium with an overall NOAEL of
62 mg/kg bw/day from a 6-week mouse study.

Osteomalacia (softening of the bones) has been observed in both animals and humans. This
effect occurs at concentrations of 100 to 200 mg Al/kg bone tissue. Further dose-response
information for this effect is lacking.

II.1.3 Children as a sensitive subgroup

Limited data are available on aluminium toxicity in children. These data are summarised by
ATSDR (1999). Neurological and skeletal effects have been reported in children with
impaired renal function with high exposures to aluminium compounds (medical use).
Adverse effects have also occurred due to binding of phosphate to aluminium in the gut
(leading to decreased phoshate absorption) in infants given oral antacids against colic. As
already mentioned above prematurely born infants appear to be very sensitive for aluminium
neurotoxicity. Bishop et al. (1997) found impaired neurologic development after parenteral
administration of standard feeding solutions resulting in Al intake of 45 μg/kg bw/day
compared to Al-depleted solutions giving only 4.0–5.0 μg/kg bw/day. Feeding durations
ranged from 6-16 days and for 157 infants on study. The exposure to the standard feeding
solutions was associated with a reduction in the Bayley Mental Development Index (p = 0.03)
of one point per day of aluminium exposure.

Animal studies generally do not indicate higher sensitivity of young animals for the
neurotoxic potency of aluminium. One rat study even indicates lower sensitivity of young
animals. In rabbits results are inconclusive. Interference by aluminium with absorption of
calcium, zinc and magnesium, however, was found to be greater in young animals in a single
rat study (ATSDR, 1999).

II.1.4 Local effects upon dermal contact

Human data are scarce. Aluminium compounds are widely used in cosmetic anti-perspirants,
which application has not led to known adverse effects. Some individuals, however, are
allergic for these products and possibly the reddening of the skin they experience is related to
exposure to aluminium. Some aluminium salts (chloride, nitrate) have been shown to damage
skin of animals at high concentrations (10%). Several other salts (sulfate, acetate,
RIVM report 320003001 page 145 of 234

chlorohydrate), however, did not show this effect. Further animal data are lacking (ASTDR,
1999).

II.1.5 Absorption
Studies in humans indicate that aluminium is absorbed in the GI-tract to a very limited extent
only. In most studies percentages of 0.1 to 0.3% were found. From the more available
chemical forms such as aluminium citrate absorption may be somewhat higher, i.e. up to 1%
(ATSDR, 1999). Chedid et al. (1991) studied uptake of aluminium from antacids in infants
and, based on increased blood aluminium concentration following antacid administration,
estimated intestinal intake was about 0.08-0.16% (OEHHA, 2000).

II.1.6 Toxicological limit values for ingestion of aluminium

For aluminium a Provisional Tolerable Weekly Intake (PTWI) was established by JECFA
(1989). Based on a NOAEL of 3% acidic sodium aluminium phosphate in the feed of dogs
(equivalent to 110 mg Al/kg bw day) in a 27 week-study, the committee derived a PTWI of
7 mg/kg body weight/week. In the study in question, an unpublished study submitted to
JECFA referenced as Katz (1981), no toxic effects were seen (NOAEL >110 mg/kg bw/day).
On a daily basis the PTWI equals 1 mg/kg bw/day.

The US ATSDR derived an oral Minimal Risk Level for intermediate exposure duration of up
to 1 year, of 2 mg/kg bw/day based on a NOAEL of 62 mg/kg bw from a 6-week
neurotoxicity study in adult mice by Golub et al. (1989) with decreased motor activity as the
critical effect. (ATSDR, 1999).

In its evaluation OEHHA (2000) concludes human data are the preferred basis for a limit
value for aluminium in drinking-water. Two options are explored, one based on an LOAEL
for increased aluminium in blood from a 20-day study in adults and the other based on a
intravenous LOAEL from the study in premature infants by Bishop et al. (1997) in which
neurotoxicity was observed after an exposure period of 6-16 days. Using an oral absorption
factor of 0.002 the intravenous LOAEL was converted to an oral LOAEL of
22.5 mg/kg bw/day. To this level OEHHA applied a high uncertainty factor of 300, including
a factor of 10 for extrapolation to a NOAEL, 3 for limited duration of the study and 10 for
sensitive subpopulations. The latter factor however seems excessive given the fact that the
study was done in a sensitive subpopulation. Omitting this factor leads to a TDI of
0.75 mg/kg bw/day.

II.1.7 Conclusion
Human data are considered the most suitable basis for deriving a TDI for aluminium with
neurotoxicity as the preferred endpoint. The study by Bishop et al. (1997) dealt with
page 146 of 234 RIVM report 320003001

aluminium neurotoxicity in the sensitive subpopulation of premature infants and is

considered the best available basis for a TDI. Using the LOAEL of 22.5 mg/kg bw/day from
this study and applying a an uncertainty factor of 30 leads to a TDI of 0.75 mg/kg bw/day.

As to possible adverse direct dermal effects, aluminium is not expected to pose a risk. This is
based on the apparent scarcity of adverse side effects attendant to aluminium’s use in anti-
perspirants and the fact that, as concluded from limited animal data, its skin-irritating
potential is low.

References
ATSDR (1999) Toxicological profile for aluminium. US Agency for Toxic Substances and
Disease Registry, report dated July 1999.

Bishop, N.J., Morley R., Day, J.P. and Lucas A. (1997). Aluminium neurotoxicity in preterm
infants receiving intravenous-feeding solutions. New England Journal Medicine 336:1557-
1561. As cited in OEHHA (2000).

Chedid, F., Fudge, A., Teubner, J., James, S.L., and Simmer, K. (1991). Aluminium
absorption in infancy. Jounal Paediatr. Child Health 27:164-166. As cited in OEHHA (2000).

Golub MS, Donald JM, Gershwin ME, et al. 1989. Effects of aluminium ingestion on
spontaneous motor activity of mice. Neurotoxicololy and Teratology 11:23 1-235. As cited in
ATSDR (1999).

Health Canada (2003) http://www.hc-sc.gc.ca/fn-an/surveill/other-autre/infant-

nourisson/index_e.html

IPCS (1997) Environmental Health Criteria no. 194 – Aluminium. WHO/IPCS Geneva,
1997.

JECFA (1989) Monograph on Aluminium. WHO Food Additives Series no. 24.

Katz, A.C. (1981) A 6-month subchronic dietary toxicity study with Levair (sodium
aluminium phosphate, acidic) in beagle dogs. Unpublished report by Stauffer Chemical Co.,
Farmington, Connecticut. Submitted to WHO by US FDA, 1982. As cited in JECFA (1989).

OEHHA (2000) Public health goal for aluminium in drinking water. Pesticide and
Environmental Toxicology Section Office of Environmental Health Hazard Assessment
California Environmental Protection Agency. DRAFT dated February 2000.

WHO (1996) Guidelines for Drinking-water quality - Second Edition . WHO, Geneva, 1996.
RIVM report 320003001 page 147 of 234

WHO (1998) Guidelines for Drinking-water quality - Second Edition .Addendum to Volume
2. WHO, Geneva, 1998.
page 148 of 234 RIVM report 320003001

II.2 Antimony

Antimony and antimony compounds have been evaluated within the scope of the WHO
Drinking-water guidelines in 1996 and 2003. Antimony trioxide has been evaluated by EFSA
for its use as an additive and initiator in food contact materials in 2003 (EFSA, 2004).
Antimony trioxide currently is also under evaluation within the EU Existing Substances
programme (first draft 2004). A previous evaluation by RIVM is that from 1994 within the
project for soil intervention values (RIVM, 1995). Further reviews were published by US-
ATSDR (1992) and OEHHA (1997).

II.2.1 Normal exposure

Antimony is a non-essential element that resembles arsenic both chemically and biologically.
Like arsenic it occurs mainly in a trivalent or pentavalent state. Natural levels in the
environment (soil, water) are low (ppm or ppb). Antimony trioxide (Sb2O3; Sb3+), a white
powder, is the single most important economic form. It is used as a fire retardant in plastics,
textiles, rubber, adhesives, pigments and paper and also as a stabilizer in plastics,
vulcanization, ammunition primers and fireworks. Antimony tartrates are used medically in
the treatment of bilharziasis (a tropical flatworm infection) (ATSDR, 1992; US-EPA, 1995).
Antimony levels in both food and drinking-water are low. The available data are reviewed in
EU-RAR (2004). Dietary data from the UK, Sweden, Germany, France, Brazil, Turkey and
the USA showed average daily intakes for adults ranging from 1.1 to 29 μg/day. In infant
foods in the UK an overall average concentration of 1.7 μg/kg food was found (data from
2003). Based on all available data, the daily intake via the diet for a child was estimated at
0.5 μg/kg bw/day. Children’s intake via drinking-water was estimated at 0.03 μg/kg bw/day.
Air exposure was evaluated as being very much lower (EU-RAR, 2004).

The estimates made in EU-RAR (2004) are accepted here. Thus background daily intake of
antimony for a child is estimated at 0.53 μg/kg bw/day.

II.2.2 Toxicology
Human data are limited to occupational studies with inhalation exposure. Relevant animal
data are relatively limited as well. Subchronic oral studies were carried in rats out with
potassium antimony tartrate and antimony trioxide. Results show that the latter compound
has lower toxicity, which may be explained by its lower bioavailability due to lower
solubility. With potassium antimony tartrate Poon et al. (1998) reported subtle thyroid
changes in rats after 90-days exposure via drinking-water at concentrations of ≥ 5 mg Sb/litre
but in a subsequent evaluation Lynch et al. (1999) concluded that the reported effects were
physiological rather that toxicological in nature and proposed a NOAEL of 50 mg Sb/litre
(corresponding to 6.0 mg/kg bw/day) based on reduced growth and food and water
consumption at the highest dose level of 500 mg Sb/litre. A 90-day study in rats with dietary
RIVM report 320003001 page 149 of 234

dosing of antimony trioxide showed biochemical changes and liver weight increase,
suggesting liver toxicity at the highest dose level of 20,000 mg/kg (1407 mg Sb/kg bw/day).
An NOAEL for this study of 421 mg/kg bw/day has been proposed (WHO, 2003; EU-RAR,
2004). The only chronic studies are those by Schroeder et al. (1970) and Kanisawa and
Schroeder (1969) who administered a single concentration of 5 ppm potassium antimony
tartrate in drinking-water of rats and mice during their entire lifetime. Effects observed were
shortened lifespan, changes in blood biochemistry and decreased heart weight. In their
evaluation of these studies Lynch et al. (1999) noted several crucial shortcomings and
concluded that they were unsuitable for use in risk assessment.

An important issue in the toxicological evaluation of antimony has been its potential to cause
genotoxicity and carcinogenicity. Results of several in vitro assays with antimony trioxide
indicated a clastogenic potential. Several in vivo studies have been conducted with the
compound, from which it has been concluded that the in vitro potential is not expressed in
vivo (WHO, 2003; EFSA, 2004; RIVM, 2005). For soluble antimony compounds (trichloride,
acetate, pentachloride, potassium tartrate) positive results were found in some in vitro studies
and also in some in vivo studies (WHO, 2003). Carcinogenicity data for the oral route are
limited to the two studies in rats and mice respectively carried out by Schroeder et al. (1970)
and Kanisawa and Schroeder (1969). As indicated above, these studies were limited in design
therefore not much weight should be given to their negative result for carcinogenicity. For the
inhalation route increased incidence of lung tumours have been observed in female rats after
chronic exposure to trioxide. This carcinogenic response was in combination with direct lung
damage due to chronic overload with the insoluble antimony trioxide particulates and its
relevance for antinomy trioxide risk assessment can not be ascertained as of yet. The data
available indicate a non-genotoxic mechanism for the formation of these tumours (RIVM,
2005).

As to reproductive and developmental endpoints there is a paucity of data. The studies that
were conducted suggest absence of significant toxic potential for producing adverse effects.

II.2.3 Children as a sensitive subgroup

No data are available on antimony toxicity in children or young animals.

II.2.4 Local effects upon dermal contact

Only limited data are available. Human data indicate that antimony trioxide may produce
dermal irritation on skin damp with perspiration (occupational data). An NOAEL for this
effect is not known. The sensitizing potential of antimony compounds cannot be assessed due
to lack of adequate data (ASTDR, 1991; EU-RAR, 2004).
page 150 of 234 RIVM report 320003001

II.2.5 Absorption
Antimony absorption from the gastrointestinal system is relatively low, the available animal
studies indicate. Absorption percentages found in studies in mice, hamster and cows ranged
from 7 to 20% (OEHHA, 1997).

II.2.6 Toxicological limit values for ingestion of antimony

RIVM (1995) proposed a TDI for antimony of 0.86 μg/kg bw/day based on an LOAEL of
5 ppm from the rat study by Schroeder et al. (1970). This was the same derivation as earlier
proposed for antimony by the WHO within the scope of its drinking-water guidelines
programme (WHO, 1996).

In its 2003 update of the drinking-water guideline for antimony WHO used the NOAEL from
the 90-day oral study with potassium antimony tartrate in rats by Poon et al. (1998).
Following the proposal by Lynch et al. (1999) the NOAEL was put at 6.0 mg/kg bw/day and
using an uncertainty factor of 1000 (100 for intra- and interspecies variation and 10 for the
use of a subchronic study) a TDI was derived of 0.006 mg Sb/kg bw/day. EFSA (2004) in its
evaluation for use of antimony trioxide in food contact materials adopted this TDI.

Other toxicological limit values are those of US-EPA (1991) and OEHHA (1997). Both these
evaluations used the LOAEL from the Schroeder et al. (1970) chronic rat study. US-EPA
applied an uncertainty factor of 1000 to an LOAEL of 0.35 mg/kg bw/day, resulting in an
RfD of 0.0004 mg/kg bw/day (breakdown of this factor: 10 for interspecies conversion, 10 to
protect sensitive individuals, and 10 because the effect level was a LOAEL and no NOEL
was established). OEHHA in its proposal for a drinking-water guideline for antimony applied
a factor of 300 (3-fold for LOAEL to NOAEL conversion and a non-severe endpoint, 10-fold
for inter-species variation and 10-fold for variation in the human population) to the LOAEL
(this LOAEL was put at 0.43 mg/kg bw/day). This implies a TDI of
0.0014 mg Sb/kg bw/day.

II.2.7 Conclusion
The updated TDI of 0.006 mg/kg bw/day as derived by WHO (2003) is chosen as the most
appropriate value.

As to possible adverse direct dermal effects, available data are too limited for drawing
conclusions. Antimony trioxide may produce irritation on damp skin but the relevance of this
occupational finding for toy-related exposures is uncertain.

References
ATSDR (1992) Toxicological profile for antimony. US Agency for Toxic Substances and
Disease Registry, report dated July 1999.
RIVM report 320003001 page 151 of 234

EFSA (2004) Opinion of the Scientific Panel on food additives, flavourings, processing aids
and materials in contact with food (AFC) on a request from the Commission related to a 2nd
list of substances for food contact materials. The EFSA Journal (2004) 24, 1-13.

EU-RAR (2004) European Union Risk Assessment Report Diantimony Trioxide. CAS no:
1309-64-4. First DRAFT 2004.

Kanisawa, M. and H.A. Schroeder. 1969. Life term studies on the effect of trace elements on
spontaneous tumor in mice and rats. Cancer Res. 29: 892-895. As cited in US-EPA (1991)
and OEHHA (1997)

Lynch BS et al. (1999) Review of subchronic/chronic toxicity of antimony potassium tartrate.

Regulatory Toxicology and Pharmacology, 30:9–17.

OEHHA (2000) Public health goal for antimony in drinking water. Pesticide and
Environmental Toxicology Section Office of Environmental Health Hazard Assessment
California Environmental Protection Agency. Dated December 1997.

RIVM (1995) Human-toxicological Criteria for serious soil contamination: compounds

evaluated in 1993 and 1994. RIVM report no. 715810009, dated August 1995.

RIVM (2005) Comment to ‘A request for advice on how to proceed with further testing of the
mutagenicity of diantimonytrioxide (DAT)’, dated March 2, 2005. RIVM memo Dated 7
April 2005.

Schroeder, H.A., M. Mitchner and A.P. Nasor. 1970. Zirconium, niobium, antimony,
vanadium and lead in rats: Life term studies. J. Nutrition. 100: 59-66. As cited in RIVM
(1995), US-EPA (1991) and OEHHA (1997).

US-EPA (1991) IRIS-file Antimony. Derivation of RfD, last revised 02-01-1991.

U.S. EPA(1995). National Primary Drinking Water Regulations. Antimony. Office of Water.
U.S. Environmental Protection Agency. October 1995. EPA 811-F-95-002j-T. As cited in
OEHHA (1997).

WHO (1996) Guidelines for Drinking-water quality - Second Edition . WHO, Geneva, 1996.

WHO (2003) Antimony in Drinking-water. Background document for development of WHO

Guidelines for Drinking-water Quality. WHO, Geneva, 2003.
page 152 of 234 RIVM report 320003001

II.3 Arsenic

Arsenic and arsenic compounds have been evaluated within the scope of the WHO Drinking-
water guidelines in 1996. More recently, in a UN-wide initiative, WHO published a draft
version of an expert synthesis report on arsenic in drinking-water (WHO, 2001). EFSA
recently reviewed arsenic as an animal feed contaminant (EFSA, 2005). The most recent
evaluation of arsenic toxicity by RIVM is from 2001 (RIVM, 2001). Further comprehensive
reviews are those by IPCS (2001) and ATSDR (2005).

II.3.1 Normal exposure

Arsenic is a metalloid naturally occurring the earth’s crust at an average concentration of
about 2 mg/kg. Some minerals however contain much higher concentrations. Arsenic
displays different valences (-3, 0, +3, +5) and occurs in cationic and anionic forms. It occurs
in inorganic and numerous organic forms that differ not only in their physical and chemical
properties but also in their occurrence and toxicity. Anthropogenic activity (mining, waste
incineration, wood preservation) is the major source for arsenic in the environment. In some
areas (Taiwan, Chile, Argentina, Mexico, China, West-Bengal, Bangladesh) levels in
drinking-water are high as a result of high natural concentrations in groundwater. In other
regions of the world food is the major source for daily exposure to arsenic. Much of ingested
arsenic in food, however, is in organic forms which are known to be much less toxic than the
inorganic forms. As total arsenic seafood contains by far the highest concentrations but this
practically wholly consists of organic arsenic. On the basis of limited data, it has been
estimated that in meat about 25% of total arsenic is organic arsenic, 35% in poultry, 25% in
dairy products, and 35% in cereals. Based on diet studies from different countries daily intake
of total arsenic for adults was estimated to be 1 μg/kg bw/day, of which 25% was presumed
to be present as inorganic arsenic (0.3 μg/kg bw/day) (RIVM, 2001). In its review IPCS
(2001) presents results for different age groups, including a USA market basket study that
reported a total arsenic intake of 28 μg/day for children aged 0.2-2 years, an Australian
market basket study reporting 17 μg/day for 2 year-olds and a Canadian total diet study
carried out in 4 cities study reporting 15 μg/day for age 1-4 years. Assuming a child body
weight of 10 kg and 25% inorganic arsenic these values lead to an estimated daily intakes via
diet of 0.4-0.7 μg/kg bw/day for children. It should be noted that locally (near point sources)
the contribution of soil and drinking-water may be as high as or even exceed that of food.

Based on the above information background daily intake of inorganic arsenic for a child is
estimated to be between 0.4 and 0.7 μg/kg bw/day.

II.3.2 Toxicology
On the toxic effects of arsenic a large literature exists. As already indicated, organic arsenic
compounds have only very low toxicity, these compounds being excreted rapidly in urine in
RIVM report 320003001 page 153 of 234

unchanged form (ATSDR, 2005). Inorganic arsenic health effects have been studied in a
large number of human studies. Chronic skin effects of arsenic, including pigmentation
changes, hyperkeratosis and skin cancer, from medicinal use but also from drinking-water,
were reported as early as the 19th century. An endemic peripheral vascular disease (PVD),
known as blackfoot disease (BFD), leading to progressive gangrene of the legs, has been
known in Taiwan since the 1920s. Important dose-response information on health effects of
ingestion of inorganic arsenic comes from a series of epidemiological studies concerning
exposure via drinking-water, performed in Taiwan. In one large scale study by Teng et al.
(1968) and Tseng (1977) of the prevalence of BFD and dermal lesions (hyperkeratosis and
hyperpigmentation) among villagers exposed to different levels of inorganic arsenic, an
NOAEL of 0.8 μg/kg bw/day was found. Schoof et al. (1998) proposed a correction for
simultaneous ingestion of inorganic arsenic via the diet in this study, thus suggesting an
NOAEL of 1.6 μg/kg bw/day. Other NOAELs from similar epidemiological studies range
from 0.4 to 20 μg/kg bw/day; LOAELs from these studies range from 2-22 μg/kg bw/day
(ATSDR, 2005).

An important issue in the toxicological evaluation of arsenic has been its potential to cause
genotoxicity and carcinogenicity. Studies in Taiwan, Chile and Argentina show consistently
high mortality risks from lung, bladder and kidney cancer among populations exposed to
arsenic via drinking-water. Where exposure–response relations have been studied, the risk of
cancer for these sites increases with increasing exposure. Even when tobacco smoking has
been considered, the exposure–response relationship remains. Studies on populations
occupationally exposed to arsenic, such as smelter workers, pesticide manufacturers and
miners in many different countries, consistently demonstrate an excess lung cancer risk
among the arsenic-exposed. As is pointed out in RIVM (2001), the mechanism of tumour
formation by inorganic arsenic is as of yet unknown. Genotoxicity data suggest a genotoxic
potential that is limited to the induction of chromosome breaks (clastogenic activity). The
weight of evidence, RIVM concluded, indicates that most likely a toxic threshold exists in the
tumorigenic action of inorganic arsenic (RIVM, 2001).

Many other toxicological endpoints have been examined for arsenic but from the data base as
a whole the above effects appear as critical.

II.3.3 Children as a sensitive subgroup

Available data are limited, so the review by ATSDR (2005) indicates. Medical surveys show
children exposed to toxic levels of arsenic having similar symptoms than adults, including
respiratory, cardiovascular dermal and neurological effects and vomiting when arsenic is
ingested. Three epidemiological studies on children exposed to arsenic via drinking-water
reported a negative impact on neurobehavioural parameters at low exposure levels
(≥ 50 μg As/litre in drinking-water in one study, ≥ 1.7 μg/kg bw/day in another study,
unspecified in a third) (ATSDR, 2005). Note, however, that the neurological effects,
page 154 of 234 RIVM report 320003001

especially at this low exposure range, do not represent a well-established effect by inorganic
arsenic at the current level of knowledge.

II.3.4 Local effects upon dermal contact

Occupational studies reported contact dermatitis after dermal exposure to inorganic arsenic
dust. Studies in guinea pigs, however, are reported as negative (no effect). Limited animal
data (a study in mice and one in guinea pigs) indicate that dermal irritation occurs at very
high concentrations only (ATSDR, 2005).

II.3.5 Absorption
The degree of absorption of inorganic arsenic in the gastro-intestinal system depends on the
chemical form. Water-soluble salts are absorbed almost wholly (up to 95%) whereas
insoluble salts absorbed considerably less (25%). Absorption of oral arsenic as part of soil
matrices like soil frequently is even lower (3-50%) (ATSDR, 2005).

II.3.6 Toxicological limit values for ingestion of arsenic

JECFA (1989) proposed a Provisional Tolerable Weekly Intake (PTWI) of 15 μg/kg bw
based on a chronic LOAEL of 100 μg As/litre in drinking-water of humans and using a daily
water consumption of 1.5 litres. This value, expressed as a TDI of 2.1 μg/kg bw/day was
adopted by RIVM (1991). In its 2001 evaluation RIVM proposed dividing this TDI by a
factor of 2 because some epidemiological studies indicate this TDI is insufficiently
protective. Thus a TDI of 1.0 μg/kg bw/day resulted (RIVM, 2001).

The US ATSDR derived an oral Minimal Risk Level for chronic exposure to arsenic of
0.3 μg/kg bw/day based on an NOAEL of 0.8 μg/kg bw/day from the epidemiological study
by Tseng et al. (1968) and Tseng (1977) among Taiwanese farmers with the prevalence of
Blackfoot Disease and dermal hyperkeratosis and hyperpigmentation as the critical effect. To
this NOAEL a uncertainty factor of 3 was applied to cover intrahuman variability (ATSDR,
2005).

Other organisations have developed proposals for limit values, especially for drinking-water,
based on quantitative cancer risk estimation. WHO (1996) for instance proposed a drinking-
water guideline of 10 μg/litre which was estimated to represent a lifetime skin cancer risk of
6x10-4. The US National Research Council estimated a comparable cancer risk for arsenic in
drinking-water (NRC, 2001).
RIVM report 320003001 page 155 of 234

II.3.7 Conclusion
The weight of evidence indicates a threshold in the carcinogenic action by arsenic. Thus a
TDI value is chosen for risk assessment. The value of 1.0 μg/kg bw/day as proposed by
RIVM (2001) is chosen as the most appropriate value for toy-related exposures.

As to possible adverse direct dermal effects, available data overall suggest a low risk. Dermal
contact with arsenic as contained in dust has been reported to produce dermatitis under
occupational conditions. The relevance of this finding for toy-related exposures, however, is
uncertain. A sensitisation study in guinea pigs was negative (no effect). The dermal irritation
potential of arsenic seems low, based on limited animal data.

References
ATSDR (2005) Toxicological profile for arsenic. US Agency for Toxic Substances and
Disease Registry, report dated Draft February 2005.

EFSA (2005) Opinion of the scientific panel on contaminants in the food chain on a request
from the commission related to arsenic as undesirable substance in animal feed. Adopted
January 2005. The EFSA Journal (2005) 180, 1-35.

IPCS (2001) Environmental Health Criteria no. 224 - Arsenic and arsenic compounds. WHO
Geneva 2001

JECFA (1989) WHO Food Additives Series 24. WHO Geneva 1989.

NRC (2001) NRC (2001). Arsenic in Drinking Water 2001 Update. National Research
Council, National Academy Press, Washington, DC.

RIVM (2001) Re-evaluation of human-toxicological Maximum Permissible Levels. RIVM

report no. 711701025, dated March 2001.

Schoof et al. (1998) Use of background arsenic exposure data to assess health significance of
exposures to arsenic in soil. Toxicologist. Abstracts of the 37th Annual Meeting 42(1-S):229.
As cited in ATSDR (2005).

Tseng, WP (1977) Effects and dose-response relationships of cancer and Blackfoot disease
with arsenic. Environmental Health Perspectives 19:109-119. As cited in ATSDR (2005).

Tseng, WP, Chu HM, How SW, et al. (1968). Prevalence of skin cancer in an endemic area
of chronic arsenicism in Taiwan. Journal of the National Cancer Institute 40:453-463. As
cited in ATSDR (2005)
page 156 of 234 RIVM report 320003001

WHO (1996) Guidelines for Drinking-water quality - Second Edition . WHO, Geneva, 1996.

WHO (2001) http://www.who.int/water_sanitation_health/dwq/arsenic3/en/index.html

RIVM report 320003001 page 157 of 234

II.4 Barium

The toxicity of barium and barium compounds has been evaluated by IPCS (2001) and US-
EPA (2005). RIVM has evaluated the group in 1991 and again in 2001 (scope: derivation of
soil intervention values). Further reviews are those by WHO (1996) and ATSDR (2005).

II.4.1 Normal exposure

Barium is present in the earth’s crust at a mean concentration of about 0.05%, mostly as
barium sulfate and barium carbonate. These forms are insoluble in water. Other barium salts
such as barium chloride en barium nitrate, however, readily dissolve in water. Barium is
present surface water and drinking-water (natural occurrence). The barium content in
drinking-water depends on regional geochemical conditions. For drinking-water in the
Netherlands average concentrations of 230 μg/litre have been reported (measurements from
1989) but in specific regions of the world much higher levels have occasionally been reported
(WHO, 1996; IPCS, 2001). Food also contains barium. Data for the year 1994 indicate a total
daily intake via the diet in the Netherlands of on average 480 μg/person (maximum
1260 μg/person) (RIVM, 1998). This is in line with an estimate given in WHO (1996) for
daily dietary intake of barium for adults for the period 1970–1991 of 180 μg (minimum),
300 μg (median), and 720 μg (maximum) mg/person. How much of the barium present in the
diet is in insoluble form is unknown. RIVM (2001) cited UK data showing total barium
intake of 650 to 1330 μg/ day for adults and estimated the intake of water-soluble barium as
the lower bound of this range, i.e. 650 μg /person (9 μg/kg bw/day). No specific data for
normal intake by children are available

Based on these data background daily intake by young children is estimated at

9 μg/kg bw/day (estimate for adults adopted).

II.4.2 Toxicology
Insoluble forms of barium have very low toxicity. The insoluble salt barium sulfate is used in
medicinal diagnostics as an opaque contrast medium for röntgenographic studies of the
gastrointestinal tract. For soluble barium available animal and human data indicate
hypertension and renal toxicity as the health end-points of concern. Humans who ingested
high single doses of soluble barium compounds and workers who inhaled dusts of barium
ores and barium carbonate have shown hypertensive effects. Similar effects occurred in
experimental animals given barium intravenously, and in rats exposed to soluble barium in
drinking-water while on restricted diets. Based on these findings, lower-dose human studies
were conducted to examine the potential effects on blood pressure in humans and on both
blood pressure and kidney function in animals. Although the experimental study by Wones et
al. (1990), together with the epidemiological study by Brenniman and Levy (1984), did not
report any significant effects on blood pressure, they establish a NOAEL in humans of
page 158 of 234 RIVM report 320003001

0.21 mg barium/kg bw/day. The animal data suggest that the kidney may also be a sensitive
target for ingested soluble barium from low-level exposure. In chronic studies in rats and
mice, carried out within the US National Toxicology Program, increased kidney weight was
the critical effect. The NOAEL from these studies was 60 mg/kg bw/day (Dallas and
Williams, 2001).

II.4.3 Children as a sensitive subgroup

US-EPA (1998/2005) provides a review of the data, which turned out to be very limited. Two
animal studies indicate that young animals may have higher absorption of barium in the
gastrointestinal tract compared to adults. The mechanism behind this apparent increase in
absorption efficiency among younger animals, EPA adds, is not known, and it is not known if
similar findings would be observed in humans (US-EPA, 1998). No further information is
available.

II.4.4 Local effects upon dermal contact

Only very limited information is available. A dermal study with barium carbonate in rats and
rabbits suggests a skin irritative potential for this salt but the study had serious flaws. Data on
sensitization are lacking (ATSDR, 2005).

II.4.5 Absorption
US-EPA (1998/2005) and ATSDR (2005) provide reviews of the data. The absorption of
barium from the gastrointestinal tract is compound dependent. Barium sulfate is insoluble and
very little, if any, ingested barium sulfate is absorbed. Acid-soluble barium compounds, such
as barium chloride and barium carbonate, are absorbed in the gastrointestinal tract, although
the amount of barium absorbed is highly variable. Older human studies estimated that barium
was poorly absorbed with absorption percentages ranging from 1 to 15%. In animal studies
absorption showed a very large variation from less than 1% to more than 85%. Apart from
solubility, the matrix in which barium is present is an important variable (from complex food
matrices absorption is lower), animal age (young animals absorb more) and nutritional status.
For poorly soluble compounds the ingested dose is a major factor. In the acid gastric
environment a slight proportion of the poorly soluble compounds present may be dissolved in
a saturable process, leading to increased absorption of these compounds at low dose levels.

II.4.6 Toxicological limit values for ingestion of barium

For water-insoluble barium compounds RIVM (2001) derived no TDI because these
compounds were concluded to be non-toxic. For soluble barium compounds a TDI of
0.02 mg/kg bw/day was proposed based on a human NOAEL of 0.21 mg/kg bw/day from the
Wones et al. (1990) study, to which an uncertainty factor of 10 was applied for intrahuman
variability and study limitations. IPCS (2001) proposed the same derivation.
RIVM report 320003001 page 159 of 234

US-EPA (2005) used nephropathy in a chronic NTP drinking-water study in mice as the
critical effect. Based on a benchmark dose for 5% effect (BMD) of 84 mg/kg bw/day and a
corresponding BMDL05 of 63 mg/kg bw/day, and using an uncertainty factor of 300 a RfD of
0.2 mg/kg bw/day was obtained. The factor 300 included 10 for interspecies variation, 10 for
intraspecies variation and 3 for deficiencies in the database.
ATSDR (2005) derived oral MRLs for intermediate and chronic durations. The chronic MRL
was based on the mouse NTP drinking-water study (also used by US-EPA) from which a
BMDL05 for nephropathy was derived of 61 mg/kg bw/day for male mice, to which an
uncertainty factor of 100 was applied (10 for interspecies variation, 10 for intraspecies
variation). Thus a chronic MRL of 0.6 mg/kg bw/day resulted. An intermediate MRL of
0.7 mg/kg bw/day was proposed based on a semichronic NOAEL of 65 mg/kg/day for
increased kidney weight from a rat 90-day drinking-water study carried out within the NTP
(ATSDR, 2005).

II.4.7 Conclusion
Although human data are considered a more suitable basis for deriving a TDI, the pivotal
study as used by IPCS (2001) and RIVM (2001) had important flaws (Dallas and Williams
2001). The chronic drinking-water study in mice represents a more reliable basis for a TDI.
Following the approach developed by ATSDR for its chronic MRL, in which a benchmark
approach was chosen, a TDI of 0.6 mg/kg bw/day is proposed as the most appropriate value.

As to possible adverse direct dermal effects, available data for barium are too limited for
drawing conclusions.

References
ATSDR (2005) Toxicological profile for barium. US Agency for Toxic Substances and
Disease Registry, report dated September 2005.

Brenniman GR, Levy PS (1984) Epidemiological study of barium in Illinois drinking water
supplies. In: Calabrese EJ, ed. Advances in modern toxicology. Princeton, NJ, Princeton
Scientific Publications, pp. 231–249. As cited in IPCS (2001).

Dallas, C.E., Williams, P.L. (2001) Barium: Rationale for a new oral reference dose. Journal
of Toxicology and Environmental Health Part B-Critical Reviews (oct-dec) 4 (4): 395-429.

IPCS (1994). Assessing human health risk of chemicals: The derivation of guidance values
for health-based exposure limits. Environmental Health Criteria 170. WHO Geneva.

IPCS (2001) Concise International Chemical Assessment Document 33 - Barium and Barium
Compounds. WHO Geneva, 2001.
page 160 of 234 RIVM report 320003001

RIVM (1998) Duplicaat 24-uursvoedingen 1994 - Inname aan calcium, magnesium, barium,
strontium en mangaan. RIVM rapport nr. 515004008, d.d. juli 1998. [In Dutch]

RIVM (2001) Re-evaluation of human-toxicological Maximum Permissible Levels. RIVM

report no. 711701025, dated March 2001.

US-EPA (1998/2005) Toxicological review of barium and compounds (CAS No. 7440-39-3).
In Support of Summary Information on the Integrated Risk Information System (IRIS) March
1998, Minor revisions January 1999, Reference dose revised June 2005 U.S. Environmental
Protection Agency, Washington, DC

US-EPA (2005) Iris file for Barium and Compounds, CASRN: 7440-39-3, Last Revised:
07/05/2005.

WHO (1996) Guidelines for Drinking-water quality - Second Edition . WHO, Geneva, 1996.

Wones, RG; Stadler, BL; Frohman, LA. (1990) Lack of effect of drinking water barium on
cardiovascular risk factor. Environ Health Perspect 85:355-359. As cited in IPCS (2001), US-
EPA (1998/2005) and Dallas and Willimas (2001)
RIVM report 320003001 page 161 of 234

II.5 Boron

The toxicity of boron and boron compounds has been evaluated by WHO (1996), IPCS
(1998) and US-EPA (2004). RIVM evaluated boron in 1995 (scope: derivation of soil
intervention values). EFSA has derived a Tolerable Upper Intake Level for boron in 2004
(EFSA 2004). Sodium perborate is under evaluation within the EU Existing substances
Programme EU-RAR (2005).

II.5.1 Normal exposure

Boron is a naturally occurring non-metal element that is found in the form of borates in the
oceans, sedimentary rocks, coal, shale, and some soils. In water at neutral pH boric acid is the
dominant form. In food boron occurs as borate or boric acid. The concentration of boron in
the earth’s crust is 10 mg B/kg (range 5 mg/kg in basalts to 100 mg/kg in shales) whereas in
oceans 4.5 mg B/litre in the ocean is present. Boron, as boric acid, borax and other borates,
are found in a wide range of consumer products, including boron-silicate glass, soaps,
detergents, preservatives, adhesives, porcelain, cosmetics, enamel, leathers, carpets, artificial
gemstones, high-contrast photographic material, wicks, electric condensers, fertilisers,
insecticides, and herbicides (EFSA, 2004).

Boron has not been established to be an essential nutrient for humans. Food is the main
source of exposure for most populations but exposure via water, especially bottled mineral
water can be substantial as well. Data on dietary intakes of boron are limited. Foods rich in
boron include fruits, leafy vegetables, mushrooms, nuts and legumes as well as wine, cider
and beer. Meat, fish and dairy products are poor sources. For the UK total dietary intake for
adults has been estimated at 1.5 mg/day (mean) and 2.6 mg/day (97.5 percentile) for the year
1994. Also for the UK a 2003 estimate for adults indicates a mean intake via water of
0.2-0.6 mg/day, via supplements up to 2.0 mg/day, and via cosmetics and consumer products
of up to 0.47 mg/day. Maximum total daily intake was estimated at 5.67 mg/day. As stated
intake via bottled water may be high. EFSA (2005) indicates concentrations as high as
4.3 mg B/litre have been measured in bottled mineral water. Specific data for total daily
boron intake for children are not available.

Based on these data normal maximum daily intake is estimated to be 5 mg/day which equals
0.08 mg/kg bw for a 70 kg adult. In absence of specific data for children this estimate is
adopted for this group.

II.5.2 Toxicology
EFSA (2004) reviewed boron toxicity. Human data, the panel concluded, are sparse and not
suitable for dose-response assessment. Animal studies comprised several short-term and
long-term toxicity studies in a number of animal species (mouse, rat, dog, pig). From these
page 162 of 234 RIVM report 320003001

studies developmental and reproductive effects appear as the critical adverse effects.
Reproductive effects were observed both in repeated dose toxicity studies and reproduction
studies. In a 2-year toxicity study in rats by Weir and Fisher (1972) reproductive effects
(atrophy of seminiferous epithelium and decreased size of testicular tubules) were observed at
58.5 mg B/kg bw/day but not at the lower dose level of 17.5 mg B/kg bw/day (NOAEL).
Developmental effects produced by boron included short ribs, variation in the number of ribs
and decrease in foetal body weight. The NOAEL for decreased foetal body weight in a rat
study by Price et al. (1996) was 9.6 mg B/kg bw/day (LOAEL 13.3 mg B/kg bw/day (EFSA,
2004).

II.5.3 Children as a sensitive subgroup

US-EPA (2004) provides a review of the data. As they point out, effects for boron on the
fetus are well established based on animal studies. Data on the possible differential
susceptibility of young children however are lacking.

II.5.4 Local effects upon dermal contact

Limited data indicate that 5 or 10% aqueous solutions of boric acid and borates are mild skin
irritants. An NOAEL for this effect is unknown (IPCS, 1998). No data on sensitization are
available.

II.5.5 Absorption
Both borates and boric acid are well absorbed from the gastrointestinal tract. In several
studies in human volunteers absorption percentages of 84% and higher were found (US-EPA,
2004).

II.5.6 Toxicological limit values for ingestion of boron

RIVM (1995) proposed a TDI for boron of 0.09 mg/kg bw/day based on an NOAEL of
8.8 mg/kg bw/day from a 2-year diet study in dogs with testicular toxicity as the critical
effect. An uncertainty factor of 100 was applied to this level (10 for intraspecies variation,
10 for intraspecies variation).

IPCS (1998) derived a TDI of 0.4 mg/kg bw/day based on NOAEL for developmental
toxicity of 9.6 mg/kg bw/day from the rat study by Price et al. (1996). The applied
uncertainty factor comprised subfactors for interspecies and intraspecies differences in
toxicokinetics and toxicodynamics. For interspecies differences in kinetics a compound-
specific subfactor of 1 was applied, for interspecies differences in toxicodynamics a default
of 100.4, for intraspecies differences in toxicokinetics a compound-specific factor of 100.4 and
for intraspecies differences in toxicodynamics a default factor of 100.5.
RIVM report 320003001 page 163 of 234

EFSA (2004) in its derivation of an Tolerable Upper Intake Level used an approach similar to
that chosen by IPCS (1998). Based on the NOAEL of 9.6 mg/kg bw/day for developmental
effects a tolerable intake of 0.16 mg/kg bw/day was calculated, providing for an Upper Level
(UL) for an adult of 10 mg B/day. For intraspecies differences in toxicokinetics a compound-
specific extrapolation factor of 1.8 was applied (based on interindividual differences among
humans in renal glomerular filtration rate, which is the critical physiological process in boron
clearance). For interspecies differences in toxicokinetics and inter- and intraspecies
differences in toxicodynamics default subfactors of 3.2 were applied. EFSA (2004) derived
ULs for different age groups. Pointing out that multigeneration studies in animals did not
indicate young animals to be more susceptible than adults, the Panel chose to extrapolate the
UL from adults to children on a surface area (body weight0.75) basis. For the age group
1-3 years (bw 12-13 kg) an UL of 3 mg/day was proposed and for age group 4-6 years (bw
19-20 kg) an UL of 4 mg/day (EFSA, 2004).

US-EPA (2004) used a similar approach based a BMDL05 for decreased fetal weight
calculated from the rat developmental study by Price et al. (1996) in conjunction with another
similar study by Heindel et al. (1992). Compound-specific subfactors were applied for both
interspecies and intraspecies differences in toxicokinetics whereas for differences in
toxicodynamics default subfactors were applied. Thus an RfD of 0.2 mg /kg bw/day was
established

II.5.7 Conclusion
The UL for 1-3 years of 3 mg/day (0.16 mg/kg bw/day) is chosen as the most appropriate
value for toy-related exposures.

As to possible adverse direct dermal effects, no conclusion is possible on sensitization due to

lack of data. Concentrations of 5 or 10% of boric acid and borates are known mild to be
mildly skin irritating but whether toy-related exposures could be this high is uncertain.
Overall the risk seems low.

References
ATSDR (1992) Toxicological profile for boron. US Agency for Toxic Substances and
Disease Registry, report dated September 2005.

EFSA (2004) Opinion of the Scientific Panel on Dietetic Products, Nutrition and Allergies on
a request from the Commission related to the Tolerable Upper Intake Level of Boron
(Sodium Borate and Boric Acid). EFSA Journal 80, 1-22.

EFSA (2005) Opinion of the Scientific Panel on Contaminants in the Food Chain on a request
of the Commission related to concentration limits for boron and fluoride in natural mineral
waters. EFSA Journal 237, 1-8.
page 164 of 234 RIVM report 320003001

EU-RAR (2005) Risk Assessment Report Sodium Perborate – Human Health. Draft dated
February 2005.

Heindel, JJ; Price, CJ; Field, EA; et al. (1992) Developmental toxicity of boric acid in mice
and rats. Fund Appl Toxicol 18:266-277. As cited in US-EPA (2004).

IPCS (1998) Environmental Health Criteria no. 224 - Boron. WHO Geneva, 1998.

Price CJ, Strong PL, Marr MC, Myers CB, & Murray FJ (1996) Developmental toxicity
NOAEL and postnatal recovery in rats fed boric acid during gestation. Fundam Appl Toxicol,
32: 179-193. As cited in EFSA (2004) and US-EPA (2004).

RIVM (1995) Human-toxicological Criteria for serious soil contamination: compounds

evaluated in 1993 and 1994. RIVM report no. 715810009, dated August 1995.

US-EPA (2004) Toxicological review of boron and compounds (CAS No. 7440-42-8). In
Support of Summary Information on the Integrated Risk Information System (IRIS) June
2004. U.S. Environmental Protection Agency, Washington, DC.

US-EPA (2004) Iris file for Boron and Compounds, CASRN: 7440-42-8, Last Revised:
08/05/2004.

WHO (1996) Guidelines for Drinking-water quality - Second Edition . WHO, Geneva, 1996.
RIVM report 320003001 page 165 of 234

II.6 Cadmium

Cadmium and cadmium compounds have been evaluated within the scope of the WHO
Drinking-water guidelines in 1996. JECFA evaluated the compound as a food contaminant on
several occasions (JECFA, 1989; 2001). EFSA recently reviewed cadmium as an animal feed
contaminant (EFSA, 2004). The most recent evaluation by RIVM is from 2001 (RIVM,
2001). Cadmium and cadmium oxide are under evaluation within the EU Existing Substances
Programme (EU-RAR, 2003). A further comprehensive review is that by ATSDR (1999).

II.6.1 Normal exposure

Cadmium is a metal occurring in the earth’s crust at concentrations of 0.1 to 1 ppm, primarily
associated with zinc ores. Small amounts of cadmium enter the environment from the natural
weathering of minerals, forest fires, and volcanic emissions, but most is released by human
activities such as mining and smelting operations, fuel combustion, disposal of metal-
containing products, and application of phosphate fertilizer or sewage. The principal
chemical species in air is cadmium oxide, although some cadmium salts, such as cadmium
chloride, can enter the air, especially during incineration. In surface water and groundwater,
cadmium can exist as the hydrated ion, or as ionic complexes with other inorganic or organic
substances.

For non-smokers diet is the major route of exposure to cadmium. A wealth of data is
available on cadmium levels in foods and diets in various countries across the world (JECFA,
2001). In a recent SCOOP report thirteen Member States of the EU submitted data based on
some of the 16 food categories, relevant for the estimation of cadmium intake. The resulting
mean intake was around 100 µg/week (range 2.7 - 176 µg/week) or 1.6 µg/kg bw/week for a
60 kg adult (EFSA, 2004). Children’s exposure per kg body weight will generally be larger
than that for adults, EFSA adds, because children have a lower body mass. JECFA (2001)
report a study from Australia in which average daily intakes for adults ranged from 0.07 to
0.24 μg/kg bw/day and for children of age 2 years from 0.18 to 0.57 μg/kg bw/day. This
indicates an intake for young children twice that of adults.

Smoking contributes significantly to the cadmium body burden (1-3 μg/package of cigarettes
as internal dose).

Based on the above information background daily intake of cadmium for a child is estimated
to be 0.45 μg/kg bw/day (twice the mean intake for an adult as reported by EFSA, 2004).

II.6.2 Toxicology
De toxicity of cadmium has been examined in a vast number of studies in animals and
humans. The crucial dose response information comes from the numerous epidemiological
page 166 of 234 RIVM report 320003001

studies among populations with increased exposure. A variety of toxic effects has been
described including nephrotoxicity, osteoporosis, neurotoxicity, carcinogenicity and
genotoxicity, teratogenicity, and endocrine and reproductive effects. The most sensitive effect
is renal toxicity consisting of the induction of irreversible tubular nephropathy that may lead
to renal insuffiency. Any cadmium in blood is bound to proteins, especially albumine. In the
liver complexation to metallothioneïn (MT) takes place. The Cd-MT-complex is then
redistributed to several organs and tissues, predominantly to the kidneys, where part of the
cadmium is released, finding its way to sensitive cellular membranes in the tubuli. Damage of
these leads, if exposure is sufficiently high and long-lasting, to the characteristic tubular
nephropathy. In de kidneys cadmium has a very long half life, viz. of 10 to 30 years, which
explains the continuous accumulation in this organ up to the age of 50 to 60 years.

Based on all data pertaining to the renal toxicity of cadmium as arising from the numerous
human studies, Järup et al. (1998) concluded that at 50 mg cadmium/kg (ww) in the renal
cortex (corresponding to a cadmium excretion in urine of about 2.5 μg/g creatinine) renal
effects are present in low percentage of the population (estimated at 4%). At 125 mg/kg in the
renal cortex 10% of the population is thought to experience such effects. In line with
evaluations by JECFA, Järup et al. (1998) concluded that in order to prevent renal tubular
damage developing into clinical disease cadmium concentrations in the renal cortex should
remain below 50 mg/kg (cadmium in urine below 2.5 μg/g creatinine). Based on a critical
analysis of all available studies and model calculations they conclude that the critical level of
50 mg/kg in the renal cortex is reached after about 45 years of exposure to 50 μg/day (about
1 μg/kg bw/day). As explained in JECFA (2001) these model calculations are based on
plausible assumptions regarding cadmium absorption and cadmium excretion. Different
assumptions within the plausible range will lead to somewhat different intake levels as
needed for reaching the critical renal level.

As already indicated above, cadmium exposure has been linked to a wide range of other
toxicological endpoints but the level of evidence for these possible associations is lower than
for renal toxicity. Accordingly renal toxicity has been the effect chosen as critical in
cadmium risk assessment.

II.6.3 Children as a sensitive subgroup

Cadmium is a cumulative toxicant, and the human exposure conditions of most concern are
long-term. Average cadmium concentrations in the kidney are near zero at birth, and rise
roughly linearly with age to a peak (typically around 40-50 µg/g wet weight) between ages 50
and 60, after which kidney concentrations plateau or decline. It is not known whether
children have a higher toxicodynamic susceptibility for renal toxicity of cadmium.
Toxicokinetically, however, children may have increased sensitivity due to higher absorption
in the gastro-intestinal system; so at least several animal studies indicate (ATSDR, 1999).
RIVM report 320003001 page 167 of 234

II.6.4 Local effects upon dermal contact

Dermal contact with cadmium does not produce allergic reactions, the available limited data
indicate. In patients suffering from dermatitis or eczema skin irritation occurred at 2%
cadmiumchloride with no effect at 1%. A test in guinea pigs with the same compound at a
concentration of concentration of 0.5% showed no effect (ATSDR, 1999).

II.6.5 Absorption
After oral intake cadmium is absorbed to a limited extent only. The matrix in which it is
present is an important factor. From food uptake into the body is lower that from drinking-
water. Fysiological status also is an important variable. The iron status strongly influences the
degree of absorption. Generally more than 90% of the ingested amount passes through the
gastro-intestinal system without absorption (ATSDR, 1999). WHO (1992) concluded that on
average, 5% of the total oral intake of cadmium is absorbed, but individual values range from
less than 1% to more than 20%. As already stated above, young animals have higher
absorption than older animals.

II.6.6 Toxicological limit values for ingestion of cadmium

JECFA (1989) proposed a Provisional Tolerable Weekly Intake (PTWI) of 7 μg/kg bw. This
was based on the proviso that levels of cadmium should not exceed 50 µg/g in renal cortex
and assuming an absorption rate of 5% and a daily excretion of 0.005% of body burden. It
was added that since the PTWI was derived from estimated accumulation of cadmium over a
period of 50 years at an exposure rate equivalent to 1 µ/kg bw/day for adults, excursions
above this figure may be tolerated provided that they are not sustained for a long period of
time and do not produce a significant increase in integrated lifetime dose. The PTWI of
7 μg/kg bw/week has been confirmed in JECFA (2001).

RIVM (2001) followed a recommendation by Järup et al. (1998) who estimated that for
maintaining the risk level as intended by JECFA (50 mg/kg in the renal cortex) the lifetime
daily intake of cadmium should be lowered from 50 μg/day to 30 μg/day. Thus, using a
safety factor of 2 on the existing PTWI previously derived by JECFA, a new value of
3.5 μg/kg bw/week was derived, equivalent to 0.5 μg/kg bw/day.

The US ATSDR derived an oral Minimal Risk Level for chronic exposure to cadmium of
0.2 μg/kg bw/day based on an NOAEL of 2.1 μg/kg bw/day for renal damage (proteinuria) as
the critical effect, selected from the human study by Nogawa et al. (1989). In this derivation
an uncertainty factor of 10 was applied for human variability (ATSDR, 1999).

II.6.7 Conclusion
The value of 0.5 μg/kg bw/day as proposed by RIVM (2001) is chosen as the appropriate
value for toy-related exposures.
page 168 of 234 RIVM report 320003001

As to possible adverse direct dermal effects, the data indicate a low risk only. No
sensitization has been observed and no skin-irritation occurred at concentrations as high as
0.5 and 1.0%.

References
ATSDR (1999) Toxicological profile for cadmium. US Agency for Toxic Substances and
Disease Registry, report dated Draft July 1999.

EFSA (2004) Opinion of the Scientific Panel on Contaminants in the Food Chain on a request
from the Commission related to cadmium as undesirable substance in animal feed (Question
N° EFSA-Q-2003-033) Adopted 2 June 2004. The EFSA Journal (2004) 72, 1-24.

EU-RAR (2003) Risk Assessment Cadmium oxide CAS-No.: 1306-19-0. Draft

EU-RAR (2003) Risk Assessment Cadmium metal CAS-No.: 7440-43-9. Draft

Järup, L. et al. (1998) Health effects of cadmium exposure – a review of the literature and a
risk estimate. Scandinavian Journal of Work, Environment and Health volume 24,
supplement 1.

JECFA (1989) WHO Food Additives Series 24. WHO Geneva 1989.

JECFA (2001) Cadmium (addendum). WHO Food Additives Series 52. WHO Geneva 2004.

Nogawa K, Honda R, Kido T, et al. 1989. A dose-response analysis of cadmium in the

general environment with special reference to total cadmium intake limit. Environment
Research 48:7-16. As cited in ATSDR (1999).

RIVM (2001) Re-evaluation of human-toxicological Maximum Permissible Levels. RIVM

report no. 711701025, dated March 2001.

WHO (1992) Environmental Health Criteria no. 134 – cadmium. WHO, International
Programme on Chemical Safety, Geneva. 1992

WHO (1996) Guidelines for Drinking-water quality - Second Edition . WHO, Geneva, 1996.
RIVM report 320003001 page 169 of 234

II.7 Chromium

Chromium and chromium compounds have been evaluated within the scope of the WHO
Drinking-water guidelines in 1996. RIVM reviewed chromium in 1991 and 2001.
Comprehensive reviews are those by US-EPA (1998a, 1998b, 1998c, 1998d), OEHHA
(1999), ATSDR (2000) and EU-RAR (2005).

II.7.1 Normal exposure

Chromium is ubiquitous in nature. The chromium content of rocks varies from an average of
around 20 mg/kg for granitic rocks up to 1,800 mg/kg in ultra basic and serpentine rocks.
Chromium can exist in oxidation states of +2 to +6 but the three environmentally stable forms
are the 0, +3 and +6 states. Naturally occurring chromium is almost always present as
trivalent chromium. Hexavalent chromium in the environment almost totally derives from
human activities. Hexavalent chromium as chromates and dichromates are used for various
industrial applications in metal processing and also as pigment and dye. Under most
environmental conditions hexavalent chromium wil be reduced to the trivalent form but
nevertheless hexavalent may sometimes persist over long periods, especially at higher pH
(around 7-8 and above) and when no oxygen, iron and organic matter are present (ATSDR,
2000; EU-RAR, 2005).

The major route for exposure of the general population to chromium is food, in which it is
exclusively present as trivalent chromium. Total adult daily intake ranges from 25 to 224
µg/day with a reported average of 60 µg/day (data from the US) (ATSDR, 2000). Based on
these data RIVM (2001) estimated adult background exposure at 1 µg/kg bw/day. Exposure
to hexavalent chromium via food is negligible. There is suggestive evidence that a small part
of the total chromium present in drinking-water (which is mostly below 2 µg/litre) may be
present as hexavalent chromium. Contact with copper chrome arsenate (CCA)-treated wood
may lead to low exposure of consumers to hexavalent chromium. A body burden of
1.63 µg/kg bw has been calculated, based on the inhalation and dermal exposure values for a
typical consumer handling and sawing dry CCA treated timber. For a child playing on CCA
treated timber, a body burden of 0.1 µg/kg bw has been estimated for oral ingestion and
dermal exposure (EU-RAR, 2005). Exposure of the general population to hexavalent
chromium via air has been estimated at 0.0057 to 0.43 ng/kg bw/day (RIVM, 2001).

Based on the above information background daily intake of trivalent chromium for a child is
estimated to be 1 μg/kg bw/day (estimate for adults as given in RIVM 2001 adopted in the
absence of data more specifically for children). The background daily intake to hexavalent
chromium probably is very low. Using the estimate for a child playing on CCA-treated
timber as given in EU-RAR (2005), would lead to an estimate of 0.1 μg/kg bw/day.
page 170 of 234 RIVM report 320003001

II.7.2 Toxicology
There is a marked difference in toxicity between trivalent and hexavalent chromium, the
latter having a much higher potency for all toxic endpoints studied.

II.7.3 Trivalent chromium

Trivalent chromium is considered an essential element for humans with a daily requirement
for adults estimated to be 0.5-2 µg of absorbable trivalent chromium (WHO, 1996).

Results of chronic animal studies with trivalent chromium indicate that water solubility of the
compound tested is an important factor. For the insoluble compound Cr2O3 (chromic oxide)
a chronic NOAEL in rats is known of 2040 mg/kg bw/day, for the slightly soluble CrCl3
(chromium chloride) a value of 3.6 mg/kg bw/day and for the readily soluble compound
Cr(CH3COO)3 (triacetate) a value of 0.46 mg/kg bw/day. Demonstrating the low toxic
potential of trivalent chromium, in none of these studies toxic effects were seen (RIVM,
2001; ATSDR, 2000). In contrast with the clear genotoxic and carcinogenic potential of
hexavalent chromium, trivalent chromium has shown little activity for these endpoints (US-
EPA, 1998a).

II.7.4 Hexavalent chromium

For this form of chromium there is a large occupational database showing that upon
inhalation (as aerosol or mist) it produces lung cancer. Based on these findings IARC
classified hexavalent chromium as a proven human carcinogen (Group I). In vitro and in vivo
genotoxicity for a wide variety of endpoints showed positive results as well. Data on the
occurrence of cancer after oral exposure to hexavalent chromium, however, are scarce. No
adequate human data are available and only very limited animal data, i.e. a mouse study in
which potassium chromate was given in drinking-water at 9 mg Cr(VI)/litre for three
generations (study by Borneff et al., 1968). Neoplastic findings in this study were limited to
2/66 carcinomas (versus 0/79 controls) and 10/66 papillomas (vs. 2/97 in controls) (ATSDR,
2000; RIVM, 2001).

Hexavalent chromium has also shown high toxic potential for non-carcinogenic endpoints
after oral administration. It produced liver and kidney toxicity and in some studies also
adverse effects on the haematopoeietic system. Developmental and reproductive effects were
found in mice and rats at oral dose levels of ≥ 20 mg/kg bw/day. No adequate chronic toxicity
studies for the oral route are available. In a study by MacKenzie et al. (1958) in which rats
received potassium chromate in their drinking-water for 1 year, no toxic effects were
observed (NOAEL > 2.4 mg Cr(VI)/kg bw/day). This study has been used as the basis for
deriving chronic limit values (see below).
RIVM report 320003001 page 171 of 234

II.7.5 Children as a sensitive subgroup

No specific data are available on the susceptibility of young children or young animals to oral
trivalent or hexavalent chromium. An absorption study in rats, however, showed ten times
higher absorption of trivalent chromium from the gastrointestinal tract in 2-day-old rats
compared to adult ones (ATSDR, 2000).

II.7.6 Local effects upon dermal contact

Hexavalent chromium is an extremely potent inducer of contact dermatitis. It is also a potent
respiratory allergen. Both human experience and animal studies indicate it to be a strong skin
and eye irritant as well. Based on numerous reports in the literature, the prevalence of
hexavalent chromium sensitivity in the general population has been estimated at 0.08%
(ATSDR, 2000). The dose-response relation for hexavalent chromium contact dermatitis has
been studied in humans. Analysis of these data led to an estimate that at a concentration of
10 mg Cr(VI)/litre a proportion of 10% of chromium sensitised persons would show a
sensitisation reaction. This exposure concentration would then protect more that 99.5% of the
population. In several studies the threshold for induction of allergic contact dermatitis was
expressed as amount per cm2 of skin, which is a more exact dose measure. Nethercott et al.
(1994) examined 54 individuals known to be sensitive to chromium-induced allergic contact
dermatitis. For hexavalent chromium they found that 10% of these already sensitised subjects
reacted at 0.09 µg Cr(VI)/cm2 (ATSDR, 2000).

Trivalent chromium has much lower potency for producing and skin irritation and skin
sensitization. In hexavalent chromium-individuals trivalent chromium was more than
300 times less potent in producing a dermal reaction compared to hexavalent chromium
(RIVM, 1998).

II.7.7 Absorption
Trivalent chromium has low absorption after oral intake. Absorption percentages in
experiments in human volunteers ranged from 0.13 to 2.8%. In these experiments chromium
was mostly given in water. The degree of absorption correlated reciprocally with dosage level
(higher at low dosages) (ATSDR, 2000).

Hexavalent consistently shows higher absorption across mucous membranes than does
trivalent chromium. In the gastrointestinal tract, however, reduction of the hexavalent
chromium to trivalent chromium, especially in the stomach, reduces the amount available of
absorption. An early experiment in human volunteers showed 10% absorption of the dose
after application of hexavalent chromium into the duodenum (versus 0.5% for trivalent
chromium). After normal oral intake absorption for hexavalent chromium in this study was
only 2.1%. Several more recent volunteer experiments with hexavalent chromium are
available in which absorption percentages of 0.5% to 18% were found (ATSDR, 2000).
page 172 of 234 RIVM report 320003001

II.7.8 Toxicological limit values for ingestion of chromium

I.1.1.1 Trivalent chromium

RIVM (2001) noted that toxicity data indicate that solubility is an important determining
factor in trivalent chromium toxicity. Insoluble forms appear to have very low toxicity, most
likely due to poor absorption. Based on NOAELs it was estimated that insoluble trivalent
chromium is 1000 times less toxic than soluble trivalent chromium. For water-soluble
trivalent a TDI of 5 μg/kg bw/day was derived from a rat NOAEL of 2.5 mg/kg bw/day
derived for 1-year drinking-water study. To this NOAEL an overall uncertainty factor of 500
was applied (incorporating 10 for intraspecies variation, 10 for interspecies variation, and an
extra factor of 5 for limited study duration). Based on this value a TDI for water-insoluble
trivalent chromium was estimated at 5 mg/kg bw/day (1000 times higher).

ATSDR (2000) established no oral limit values for trivalent chromium because the available
data were deemed to be insufficient. US-EPA (1998c) derived a TDI of 1.5 mg/kg bw/day for
insoluble trivalent chromium based on a chronic NOAEL of 1468 mg Cr(III)/kg bw day from
a rat feeding study with chromic oxide by Ivankovic and Preussman (1975). In this derivation
a total uncertainty factor of 1000 was applied (10 for intraspecies variation, 10 for
interspecies variation and an additional factor of 10 for database deficiencies). For soluble
trivalent chromium no TDI was proposed.

I.1.1.2 Hexavalent chromium

Hexavalent chromium is a genotoxic carcinogen, for which effect, it is generally assumed, no
threshold exists. For the inhalation route quantitative cancer risk assessments (QCRAs) are
available. For the oral route, however, due to a lack of relevant data, QCRAs have not been
developed by most evaluating bodies. Hexavalent chromium upon oral intake will be partly
converted to trivalent chromium, for which reduction the stomach juices possess considerable
capacity. Despite this, a residual cancer risk may remain, especially locally in the
gastrointestinal system. The only QCRA for the oral route is that by the California
Environmental Protection Agency (OEHHA, 1999) based on the Bornef et al. (1968) study in
mice in which fore-stomach tumours were observed. This led to a cancer slope factor of 0.19
per mg/kg bw/day. Based on this slope factor it can be calculated that extra lifetime cancer
risks of 10-5 and 10-6 are reached at daily intakes of 53 and 5.3 ng Cr(VI)/kg bw, respectively
(daily exposure throughout a lifetime of 70 years). Given the study limitations, however, this
estimate must be considered as highly uncertain. Following the evaluation by OEHHA (1999)
hexavalent chromium has been included in the US-NTP working programme and an oral
bioassay in rats and mice with application in drinking-water is now being conducted. The
result of this study should make a more reliable QCRA for the oral route feasible.

Given the lack of adequate chronic data on which an oral QCRA could be based, RIVM
(2001) derived a provisional TDI for hexavalent chromium based on non-carcinogenic
effects. A 1-year NOAEL of 2.5 mg/kg bw/day deriving from the rat drinking-water study by
RIVM report 320003001 page 173 of 234

MacKenzie et al. (1958), was divided by a total uncertainty factor of 500 (10 for intraspecies
variation, 10 for interspecies variation and an additional factor of 10 limited study duration).
Thus a provisional TDI of 5 μg/kg bw/day was derived.

US-EPA (1998d) used the same NOAEL of 2.5 mg/kg bw/day to derive an RfD of
3 μg/kg bw/day. Here a total uncertainty factor of 900 was applied (10 for intraspecies
variation, 10 for interspecies variation and an additional factor of 3 for limited study duration
and a modifying factor of 3 because of concern over a human study by Zhang and Li (1987)
in which increased incidences of gastrointestinal disease was reported in subjects whose
drinking-water contained 20 ppm hexavalent chromium).

ATSDR (2000) established no oral limit values for hexavalent chromium due to lack of
appropriate data.

II.7.9 Conclusion
For water-soluble trivalent chromium the TDI of 5 μg/kg bw/day as proposed by RIVM
(2001) can be used for toy-related exposures. For water-insoluble trivalent chromium the TDI
of 5 mg/kg bw/day as proposed by RIVM (2001) can be used for toy-related exposures.
Hexavalent chromium should be regarded as a genotoxic carcinogen for the oral route. A
highly uncertain estimate of the size of this cancer risk indicates extra cancer risk levels of
10-5 and 10-6 at 53 and 5.3 ng Cr(VI)/kg bw/day, respectively (lifetime exposure, 70 years).
For non-carcinogenic effects by hexavalent chromium a provisional TDI of 5 μg/kg bw/day is
available.

As to possible adverse direct dermal effects, hexavalent chromium is a potent inducer of skin
irritation and skin sensitisation. Levels as low as 0.09 µg Cr(VI)/cm2 of 10 mg Cr(VI)/litre
have been reported as the estimated 10% response dose in popuations of hexavalent
chromium-sensitised subjects. Trivalent chromium is much less potent in inducing direct skin
effects (no risk).

References
ATSDR (2000) Toxicological profile for chromium. US Agency for Toxic Substances and
Disease Registry, report dated Draft September 2000.

Borneff, I, Engelhardt, K, Griem, W, et al. (1968). [Carcinogenic substances in water and

soil. XXII. Mouse drinking study with 3,4-benzpyrene and potassium chromate]. Arch.Hyg.
152, 45-53. (German). As cited in OEHHA (1999)

EU-RAR (2005) European Union Risk Assessment Report volume 53: chromium trioxide,
sodium chromate, sodium dichromate, ammonium dichromate, potassium dichromate.
page 174 of 234 RIVM report 320003001

Institute for Health and Consumer Protection, European Chemical Bureau, Existing
Substances. European Commission, Joint Research Centre EUR 21508 EN.

Ivankovic, S; Preussmann, R. (1975) Absence of toxic and carcinogenic effects after

administration of high doses of chromic oxide pigment in subacute and long-term feeding
experiments in rats. Food Cosmet Toxicol 13:347-351. As cited in US-EPA (1998c).

MacKenzie, RD; Byerrum, RU; Decker, CF, et al. (1958) Chronic toxicity studies. II.
Hexavalent and trivalent chromium administered in drinking water to rats. Am Med Assoc
Arch Ind Health 18:232-234. As cited in US-EPA (1998c and 1998d).

NTP (2006) National Toxicology Program Factsheet: Hexavalent Chromium. Year 2006. At:
http://ntp.niehs.nih.gov/files/CRVIFacts06.pdf

RIVM (1991) Voorstel voor de humaan-toxicologische onderbouwing van C-

(toetsings)waarden. RIVM report no. 725201005, dated February 1991. [In Dutch]

RIVM (1998) Public-health risk assessment of CCA-products. RIVM/CSR Advisory report

5723. Unpublished internal report dated 05-03-1998.

RIVM (2001) Re-evaluation of human-toxicological Maximum Permissible Levels. RIVM

report no. 711701025, dated March 2001.

OEHHA (1999) Public health goal for chromium in drinking water. Pesticide and
Environmental Toxicology Section Office of Environmental Health Hazard Assessment
California Environmental Protection Agency. Dated February 1999.

US-EPA (1998a) Toxicological Review of Trivalent Chromium (CAS No. 16065-83-1). In

Support of Summary Information on the Integrated Risk Information System (IRIS). August
1998. U.S. Environmental Protection Agency Washington, DC.

US-EPA (1998b) Toxicological Review of Hexavalent Chromium (CAS No. 18540-29-9). In

Support of Summary Information on the Integrated Risk Information System (IRIS). August
1998. U.S. Environmental Protection Agency Washington, DC.

US-EPA (1998c) IRIS Summary for Chromium(III), insoluble salts (CASRN 16065-83-1).
Last revised 09/03/1998.

US-EPA (1998d) ) IRIS Summary for Chromium(VI) (CASRN 18540-29-9). Last revised
09/03/1998.

WHO (1996) Guidelines for Drinking-water quality - Second Edition . WHO, Geneva, 1996.
RIVM report 320003001 page 175 of 234

II.8 Cobalt

Cobalt and cobalt compounds have been evaluated by RIVM in 1991 and 2001. A
comprehensive review is that by ATSDR (2004).

II.8.1 Normal exposure

Cobalt occurs in the earth’s crust at average concentrations of 20-25 mg/kg. Low levels may
be present in surface water and groundwater (range < 1 to 10 µg/litre). Food is the dominant
source of general population exposure. Data on daily intake in Europe are limited. A total diet
study from Canada indicated mean daily intakes for adults of 8-15 µg/day. In France the
average daily intake via food was 29 µg/day (ATSDR, 2004). In RIVM (2001) average adult
background exposure was estimated at 0.3 µg/kg bw/day. In the Canadian total diet study
already mentioned, the average intake for age 1-4 years was 7 µg/day and for 4-11 years
10 µg/day (ATSDR 2004). On a body weight basis this indicates an intake for young children
twice that of adults.

Based on the above information background daily intake of cobalt for a child is estimated to
be 0.6 μg/kg bw/day (twice the mean intake for adults as estimated by RIVM, 2001).

II.8.2 Toxicology
As a component of cyanocobalmin (vitamin B12), cobalt is essential in the body. The US
Recommended Dietary Allowance of vitamin B12 is 2.4 µg/day, which contains 0.1 µg of
cobalt ATSDR 2004).

Cobalt toxicity has been examined to a limited extent only. Adequate chronic studies for the
oral route in humans and animals are not available. From limited human data an increase in
erythrocyte numbers (polycythemia) appears as the most sensitive endpoint following oral
exposure. This effect has been observed at 1 mg Co/kg bw/day in a subacute study in human
volunteers (LOAEL) (study by Davis and Fields,1958). In a 8-week study in rats (study by
Stanley et al. 1947) this effect was also found at this dose level; the NOAEL in this study was
0.6 mg Co/kg bw/day (ATSDR, 2004). Humans who regularly consumed beer that contained
cobalt sulphate as a foam stabiliser and who ingested an average of 0.04 mg to
0.14 Co/kg bw/day over a period of years, showed severe cardiomyopathy. This effect has
been reported in several studies but, as is pointed out in ATSDR (2004), in its development
chronic alcohol abuse may have contributed significantly.

II.8.3 Children as a sensitive subgroup

No toxicity data are available on the susceptibility of young children or young animals to oral
cobalt. An absorption study in rats and guinea pigs, however, showed increased absorption
from the gastrointestinal tract in younger animals (3- to 15-fold higher) (ATSDR, 2004).
page 176 of 234 RIVM report 320003001

II.8.4 Local effects upon dermal contact

The potential of cobalt to induce dermatitis has been demonstrated in a large number of
human studies. Using patch tests and intradermal injections, it has been demonstrated that the
dermatitis is probably caused by an allergic reaction to cobalt, with the cobalt ion functioning
as a hapten. Exposure levels in these studies, however, mostly were not reported. A NOAEL
for induction of dermatitis has not been established. In two Polish occupational studies 10 or
20% of the study populations of nurses and dentists reacted positively to a patch test 1.0%
cobalt chloride. Another study in patients known to have cobalt allergy (Nielsen et al., 2000),
suggests that the allergic reactions to cobalt are primarily linked to cobalt metal and not to
cobalt salts. Interrelationships exist between nickel and cobalt sensitization but the extent of
this (potential) interaction on immunologic endpoints is not well understood. No data on the
potential of cobalt compounds to induce skin irritation are available (ATSDR, 20004).

II.8.5 Absorption
Gastrointestinal absorption of cobalt and its compounds shows wide variation (18-97% of the
dose), depending on dosing level and its type and the nutritional status. The iron status
influences the degree of absorption. Humans deficient in iron absorbed 31-71% of a dose
compared to 18-44% in controls. Animal studies show that soluble cobalt chloride is better
absorbed than insoluble cobalt oxide (13-34% versus 1-3%). In one study in rats and guinea
pigs absorption in younger animals (age 1-60 days) was 3- to 15-fold higher than in adult
animals (200 days old) (ATSDR, 2004).

II.8.6 Toxicological limit values for ingestion of cadmium

RIVM (2001) proposed a TDI of 1.4 μg/kg bw/day based on a human LOAEL of
0.04 mg/kg bw/day derived from the studies on beer drinkers ingesting cobalt sulphate as a
foam stabiliser. To this LOAEL an uncertainty factor of 3 was applied for intra-human
variability and factor of 10 for extrapolation of an LOAEL to an NOAEL.

ATSDR (2004) derived an oral Minimal Risk Level (MRL) for intermediate exposure
duration of up to 1 year, of 10 μg/kg bw/day based on a LOAEL for polycythemia of 1 mg/kg
bw from a 22-day study in humans (Davis and Fields, 1958). The LOAEL was divided by an
uncertainty factor of 100 (10 for the use of an LOAEL and 10 for human variability). No
chronic oral MRL was derived due to lack of appropriate data.

II.8.7 Conclusion
The value of 1.4 μg/kg bw/day as proposed by RIVM (2001) is chosen as the appropriate
value for toy-related exposures.
RIVM report 320003001 page 177 of 234

As to possible adverse direct dermal effects, cobalt is a skin sensitiser. The dose-response
relation for this effect however is poorly understood. Possibly only cobalt as a metal has
sensitising potential with cobalt salts having none. The skin-irritating potential of cobalt is
unknown.

References
ATSDR (2004) Toxicological profile for cobalt. US Agency for Toxic Substances and
Disease Registry, report dated Draft April 2004.

Davis JE, Fields JP. 1958. Experimental production of polycythemia in humans by

administration of cobalt chloride. Proc Soc Exp Biol Med 99:493-495. As cited in ATSDR
(2004).

Nielsen NH, Kristiansen J, Borg L, et al. 2000. Repeated exposures to cobalt or chromate on
the hands of patients with hand eczema and contact allergy to that metal. Contact Dermatitis
43(4):212-215. As cited in ATSDR (2004).

RIVM (2001) Re-evaluation of human-toxicological Maximum Permissible Levels. RIVM

report no. 711701025, dated March 2001.
page 178 of 234 RIVM report 320003001

II.9 Copper

Copper and copper compounds have been evaluated within the scope of the WHO Drinking-
water guidelines in 1996 and 1998. The EU Scientific Committee Food recently derived a
Tolerable Upper Intake Level for Copper (SCF, 2003). A previous evaluation by RIVM is
that from 2001. Comprehensive reviews are those by IPCS (1998), US-NRC (2000) and US-
ATSDR (2002).

II.9.1 Normal exposure

Copper occurs in the environment in three major valence states: copper metal Cu0, Cu+ and
Cu2+. The mean concentration of copper in soil ranges from 5 to 70 mg/kg and is higher in
soils near smelters, mining operations, and combustion sources. The median concentration of
copper in rivers, lakes, and oceans is 4–10 ppb. It is predominantly in the Cu2+ state, most of
it complexed or tightly bound to organic matter. The combined processes of complexation
adsorption and precipitation control the level of free Cu2+. The chemical conditions in most
natural water are such that, even at relatively high copper concentrations, these processes will
reduce the free Cu2+ concentration to extremely low values. Sediment is an important sink
and reservoir for copper (ATSDR, 2004).

Food and drinking-water are the major sources for general population exposure to copper. In
drinking-water relatively high concentrations may be present due to migration from
distribution systems (both from the water treatment plant and in the home), especially after a
period in which the system has not been flushed. Mean daily intake from foods in different
EU countries ranges from 1.1 to 2.2 mg/day (97.5-percentiles 1.2 to 4.2) (SCF 2003).
Drinking-water may contribute up to 1 mg/day according to RIVM (2001).Total background
exposure from food and drinking water has been estimated at 0.03 mg/kg bw/day for an adult
(RIVM 2001). Data for intake by children are limited. In one US study intake by children
aged 2 years was about half that of adults expressed as mg/day (0.48 mg/day versus about 1
mg/day). Per kg body weight infant intake thus would be about twice that of adults.

Copper is essential element for biological organisms, being an essential component of many
enzymes (cuproenzymes) and proteins. Recommended daily allowances for human adults as
given in the UK and USA range from 0.9 to 1.2 mg/day (SCF, 2003).

Based on the estimate for adults of 0.03 mg/kg bw/day as given by RIVM (2001), children’s
normal exposure is estimated at 0.06 mg/kg bw/day (twice the adult value, as suggested by
US data).
RIVM report 320003001 page 179 of 234

II.9.2 Toxicology
Animal data on copper toxicity are relatively limited. Human data indicate that chronic
copper toxicity has its most pronounced effects on liver function whilst acute effects of
copper toxicity are primarily observed in the gastrointestinal tract, as a local intestinal
irritation effect. Acute copper toxicity in drinking water appears to have a threshold of
approximately 6 mg/L. For longer exposures SCF (2003) considered liver damage the critical
endpoint. After long-term copper intake at 30 mg/day or 60 mg/day for several years acute
liver failure developed, so O’Donohue et al. (1993) report for a single case. Several other
human studies indicated absence of adverse liver effects after prolonged intake of 7 to
10 mg/day. From a 12-weeks supplementation study by Pratt et al. (1985) an overall NOAEL
of 10 mg/day for liver effects was selected.

For other toxicity endpoints the available data are limited. Poor quality studies of copper
compounds in rats and mice suggest absence of carcinogenic activity. Genotoxicity data are
inconclusive. In developmental and reproduction studies testicular degeneration and reduced
neonatal body and organ weights were seen in rats at dose levels in excess of 30 mg Cu/kg
body weight per day over extended time periods, and fetotoxic effects and malformations
were seen at high dose levels (>80 mg Cu/kg body weight per day) (IPCS,1998; SCF, 2003).

II.9.3 Children as a sensitive subgroup

Copper is an essential element required for normal growth and development. Signs of copper
deficiency in infants and children include anemia that is unresponsive to iron
supplementation, neutropenia, bone abnormalities, and hypopigmentation of the hair. Indian
childhood cirrhosis and idiopathic copper toxicosis are two syndromes associated with high
intake of copper. Both are characterized by severe liver damage in infants and children
(< 5 years of age). The syndromes have been linked to genetic defects, due to which copper
metabolic capacity is exceeded in certain individuals, leading to excessive copper
concentrations in the liver. Several reports indicate that children may be more sensitive to the
gastro-intestinal effects produced by copper but the evidence on this issue is inconclusive as
of yet (ATSDR, 2004).

II.9.4 Local effects upon dermal contact

Some medical case studies show that copper may produce dermal contact dermatitis. No dose
response information for this supposed effect is available. Data on skin-irritating potential are
lacking (ASTDR, 2004).

II.9.5 Absorption
The percentage absorption of dietary copper depends on the amount of copper ingested, with
the percentage absorption decreasing with increasing intakes. A series of studies in humans
demonstrated that a 10-fold increase in dietary copper resulted in only twice as much copper
page 180 of 234 RIVM report 320003001

being absorbed. A theoretical maximum absorptive capacity of 63-67% has been estimated
from aggregate results of human copper absorption studies at various copper daily intakes.
With typical diets in developed countries the average copper absorption has been estimated to
be in the 30-40% range (SCF, 2003). Limited evidence in humans and animals suggests that
the process of absorption is less easily saturated in young humans and animals than in older
ones, which effect could lead to higher absorption rates in the former, of which however no
quantitative estimate is available (ATSDR, 2004).

II.9.6 Toxicological limit values for ingestion of copper

RIVM (2001) proposed a TDI for copper of 0.14 mg/kg bw/day, which was loosely based on
an LOAEL of 4.2 mg/kg bw/day for chronic oral exposure in mice (study by Massie and
Aiello 1984), taking into account minimum nutritional requirement for copper in humans of
0.02 to 0.08 mg/kg bw/day.

The US ATSDR derived an oral Minimal Risk Level for intermediate exposure duration of up
to 0.01 mg/kg bw/day based on a study by Araya et al. (2003) in which gastrointestinal
effects were observed. This study identified NOAEL and LOAEL values of 0.042 and
0.091 mg Cu/kg/day, respectively; these copper doses were in excess of normal dietary
intake. The NOAEL was divided by an uncertainty factor of 3 (to account for human
variability) to yield an intermediate-duration oral MRL of 0.01 mg Cu/kg/day. The
intermediate-duration MRL is intended to protect against exposure to excess copper in
drinking water and assumes a normal copper dietary intake (ATSDR, 2004).

SCF (2003) established a Tolerable Upper Intake Level (UL) for copper based on a NOAEL
of 10 mg/day for adverse effects on liver function as the critical endpoint, derived from a
study by Pratt et al. (1985) in which seven adult human volunteers had their diets
supplemented with 10 mg Cu/day over a period of 12 weeks. Noting the homeostatic nature
of copper uptake into the body (lower absorption rates as higher amounts are ingested), the
Committee decided that an UF of 2 is adequate to allow for potential variability within the
normal population. Thus a UL of 5 mg/day was established for adults (0.083 mg/kg bw/day
for 60 kg adult). For the age group 1-3 years (bw 12-13 kg) a corresponding UL of 1 mg/day
was proposed and for age group 4-6 years (bw 19-20 kg) an UL of 2 mg/day (SCF, 2003).

II.9.7 Conclusion
The UL of 5 mg/day corresponding with 0.083 mg/kg bw/day, as derived by SCF (2003), is
chosen as the most appropriate value.

As to possible adverse direct dermal effects, no conclusion is possible due to lack of data.
However, given the wide use of copper in various applications (water transport, electricity
wires) without this leading to frequent reports of adverse skin effects, the potential to induce
these effects probably is very low.
RIVM report 320003001 page 181 of 234

References
Araya M, Olivares M, Pizarro F, et al. 2003. Gastrointestinal symptoms and blood indicators
of copper load in apparently healthy adults undergoing controlled copper exposure. Am J
Clin Nutr 77(3):646-650. As cited in ATSDR (2004).

ATSDR (2004) Toxicological profile for copper. US Agency for Toxic Substances and
Disease Registry, report dated September 2004.

IPCS (1998) Environmental Health Criteria no. 200 – Copper. WHO Geneva.

O’Donohue JW, Reid MA, Varghese A, Portmann B, Williams R (1993). Micronodular

cirrhosis and acute liver failure due to chronic self-intoxication. Eur J Gastroenterol Hepatol
5: 561-562. As cited in SCF (2003).

Pratt WB, Omdahl JL, Sorenson JR (1985). Lack of effects of copper gluconate
supplementation. Am J Clin Nutr 42: 681-682. As cited in SCF (2003) and ATSDR (2004).

RIVM (2001) Re-evaluation of human-toxicological Maximum Permissible Levels. RIVM

report no. 711701025, dated March 2001.

SCF (2003) Opinion of the Scientific Committee on Food on the Tolerable Upper Intake
Level of Copper (expressed on 5 March 2003). Scientific Committee on Food,
SCF/CS/NUT/UPPLEV/57 Final. Dated 27 March 2003.

US-NRC (2000) Copper in drinking water. Committee on Copper in Drinking Water. Board
on Environmental Studies and Toxicology. Commission on Life Sciences. National Research
Council. National Academy Press, Washington D.C. USA.

WHO (1996) Guidelines for Drinking-water quality - Second Edition . WHO, Geneva, 1996.

WHO (1998) Guidelines for Drinking-water quality - Second Edition .Addendum to Volume
2. WHO, Geneva, 1998.
page 182 of 234 RIVM report 320003001

II.10 Lead

The toxicity of lead and inorganic lead compounds has been studied extensively in both
animals and humans. On numerous occasions these data have been evaluated by expert
committees. The contaminants panel of the EFSA has recently evaluated lead in the food
chain (EFSA, 2004). Other major reviews are those by JECFA (2000) and ATSDR (2005).
RIVM reviewed lead in 1997 and 2001. Further evaluations are OEHHA (1997) and IPCS
(1995).

II.10.1 Normal exposure

Lead occurs naturally in the environment with an overall level in the earth’s crust of
20 mg Pb/kg dry matter. Background levels in the topsoil vary between 10 and 70 mg/kg
whereas levels in surface water are generally below 0.01 mg/l; levels up to 1 mg/l can be
expected in polluted areas with soft waters. Lead occurs naturally mainly in its inorganic
form as oxide or sulfide, but also as carbonate, sulfate and chromate. Technical use of lead,
for instance as anti-knocking agent in petrol, has resulted in increased levels in soil, water
and air. Use in solders and alloys for water pipes (drinking water supplies) has been another
major source of environmental pollution, and human and animal exposure. Both the use in
petrol and in water pipes has been abandoned in most countries. Further uses of lead are in
mining, smelting and processing, pigments, batteries and ceramics and glassware (EFSA,
2004).

General population exposure is predominantly via food and water. For infants exposure via
dust, however, can be a major additional source. JECFA (2000) summarises data on lead
intake from food and water from all seven continents. A wealth of data was available both for
adults and children. From this body of data the estimated range of intake levels from food for
children was 0.6-30 µg/kg bw per week. This was generally two to three times the adult
intake in the same country when evaluated on the basis of body weight (JECFA, 2000). This
estimate is exclusive of tapwater for which there were insufficient data to make a reliable
estimate. As already indicated non-food sources may contribute significantly in specific
situations. Use of ceramic drinking-vessels may sometimes even lead to intoxications. Intake
from soil is important especially in industrialized areas where children play in dusty
environments. The latter intakes tend to be highly location-specific. The Health Council of
the Netherlands (1997) estimated lead intake in the Netherlands from all sources, including
soil, at 2.0 µg/kg bw/day (14.0 µg/kg bw/week) for children aged 1-4 years. For higher ages
this was estimated to be 0.64 µg/kg bw/day (4.5 µg/kg bw/week).

Based on the above data normal exposure for children is estimated to be 2.0 µg/kg bw/day
which is the estimate for the Netherlands as adopted in RIVM (2001). As indicated above in
specific regions exposure may well be either higher or lower than this level.
RIVM report 320003001 page 183 of 234

II.10.2 Toxicology
The dose response for lead toxicity has been examined in numerous epidemiological studies.
This research has provided fairly detailed knowledge of the toxic effects occurring at
different concentrations of lead in the blood. At higher Pb-levels haemsynthesis is affected.
In children this is seen at about 400 μg Pb/litre and higher and in adults at 800 μg Pb/litre and
higher. But already at lower concentrations neurological functioning is impaired. This is
measurable as a decrease in IQ. JECFA (2000) presents results of statistical dose-response
assessment of neurobehavioural effects of lead in children. The best analysis that could be
developed showed a decrease of one IQ point for every 20-40 µg/litre increase in blood lead
concentration, with a greater effect at higher concentrations than at lower ones. A meta-
analysis of seven studies showed that an increase in the blood lead concentration from 100 to
200 µg/litre would result in a decrease of approximately 2.5 IQ points. A conclusion as to the
existence of a threshold for these effects (a blood PB-concentration below which no adverse
effect occurs) cannot be drawn at the present stage. In experimental animals adverse effects
on cognitive function have been demonstrated at concentrations of 110-150 μg Pb/litre.
Severe damage such as brain damage occurs only at ≥1000 μg Pb/litre.

II.10.3 Children as a sensitive subgroup

A large body of data is available on the effects of lead in children. As already stated children
are a well-identified sensitive group for lead neurotoxicity and the dose response for this
effect has been studied widely, leading to fairly detailed insight into the relation of blood Pb-
levels in children and cognitive function. The TDI for lead is based on these data.

II.10.4 Local effects upon dermal contact

With lead no dermal irritation and en sensitization studies have been carried out.

II.10.5 Absorption
ATSDR (2005) provides a review of the data. Important modulating factors for absorption of
ingested inorganic lead are physiological status (e.g., age, fasting, nutritional calcium and
iron status, pregnancy) and physicochemical characteristics of the medium ingested (e.g.,
particle size, mineralogy, solubility, and lead species). Lead absorption may also vary with
the amount of lead ingested. Both animal data and human data indicate that absorption in the
young is higher. Estimates derived from dietary balance studies conducted in infants and
children (ages 2 weeks to 8 years) indicate that 40–50% of ingested lead is absorbed. In
adults the estimated absorption ingested water-soluble lead compounds ranged from 3 to 10%
in fed subjects. Absorption under fasted conditions will be higher than these levels. Animal
and human data indicate that absorption from soil is low compared to absorption from soluble
lead salts (ATSDR, 2005).
page 184 of 234 RIVM report 320003001

II.10.6 Toxicological limit values for ingestion of lead

JECFA (1986) concluded that no effect on cognitive function is expected below
50 μg Pb/litre in blood. A provisional tolerable weekly intake (PTWI) was proposed of
25 μg/kg bw for children. This was based on the condition that any increase in lead
concentration in blood should be avoided. This approach was chosen because in many urban
areas no margin of safety exists between lead concentrations in blood and the level of 50
μg/litre (this level even is exceeded in many situations). The derivation of this PTWI was
based on metabolism studies in infants and children in which mean daily intakes of
3-4 µg/kg bw of lead by infants and children were not associated with an increase in blood
lead levels. At the slightly higher intake level of 5 µg/kg bw/day children are in positive
balance for lead retention, JECFA points out, also noting that the net absorption of dietary
lead at this level averages 40% of the lead intake, with the net retention estimated to be about
30% of intake. Metabolic studies indicate a negative balance when lead intake is less than
4 µg/kg bw/day. By cumulating the mean daily intake of 3-4 µg/kg bw over a week the PTWI
was obtained specifically for children (JECFA, 1986). In 1993 JECFA extended this PTWI to
adults because of the sensitivity of the developing fetus (JECFA, 1993). RIVM (1997, 2001)
has adopted the JECFA approach. Expressed as a daily dose the PTWI equals 3.6 μg/kg bw
(RIVM, 2001).

ATSDR (2005) has not derived limit values for lead due to lack of a clear threshold for the
critical effect and considered a case by case approach more appropriate (site-specific risk
assessment for lead as soil contaminant).

OEHHA (1997) derived a ‘level of concern’ for lead neurotoxicity in children aged 1-2 years,
using a blood level of 100 μg/litre as the point of departure. This blood level was calculated
to be associated with an intake of 28.6 μg/day and subsequently an uncertainty factor of 3
was applied. For a 10 kg child this approach implies a tolerable level of about 0.9 μg/kg
bw/day.

II.10.7 Conclusion
The value of 3.6 μg/kg bw/day as proposed by JECFA (1986, 1993) and RIVM (2001) is
chosen as the appropriate value for toy-related exposures.

As to possible adverse direct dermal effects, no conclusion is possible due to lack of data.
However, lead’s former wide use in water transport without attendant of adverse skin effects,
suggests the potential to induce these effects is low.

References
ATSDR (2005) Toxicological profile for lead. US Agency for Toxic Substances and Disease
Registry, report dated September 2005.
RIVM report 320003001 page 185 of 234

EFSA (2004) Opinion of the Scientific Panel on Contaminants in the Food Chain on a request
from the Commission related to lead as undesirable substance in animal feed (Request N°
EFSA-Q-2003-032). Adopted on 2 June 2004. The EFSA Journal (2004) 71, 1-20

Health Council of the Netherlands [Gezondheidsraad] (1997) Lood in Drinkwater.

Gezondheidsraad. Commissie Lood in Drinkwater, rapport nr. 1997/07, Den Haag. [In
Dutch]

IPCS (1995) Environmental Health Criteria 165: inorganic lead. WHO, International
Programme on Chemical Safety.

JECFA (1986) Toxicological evaluation of certain food additives and contaminants. The 30th
Meeting of the Joint FAO/WHO Expert Committee on Food Additives (JECFA). WHO Food
Additives Series no. 21.

JECFA (1993) as cited in IPCS (1995) and EFSA (2004)

JECFA (2000) Safety evaluation of certain food additives and contaminants. Who Food
Additives Series: 44. Prepared by the Fifty-third meeting of the Joint FAO/WHO Expert
Committee on Food Additives (JECFA). World Health Organization, Geneva, 2000. IPCS -
International Programme on Chemical Safety

OEHHA (1997) Public Health Goal for Lead in Drinking Water. Prepared by Pesticide and
Environmental Toxicology Section, Office of Environmental Health Hazard Assessment,
California Environmental Protection Agency, December 1997.

RIVM (1991) Voorstel voor de humaan-toxicologische onderbouwing van C-

toetsingswaarden. RIVM rapport nr. 725201005 d.d. 9 februari 1991. [In Dutch]

RIVM (1997) Evaluatiedocument Lood. RIVM draft report no. 601014003. [In dutch]

RIVM (2001) Re-evaluation of human-toxicological maximum permissible risk levels. RIVM

rapport nr. 71701025, d.d. maart 2001.

WHO (1996) Guidelines for Drinking-water quality - Second Edition . WHO, Geneva, 1996.
page 186 of 234 RIVM report 320003001

II.11 Manganese

The oral toxicity of manganese and its compounds was reviewed by WHO (1996), US-EPA
(1996), IPCS (1999) and ATSDR (2000). The former Scientific Committee of the EU
evaluated manganese as a food mineral (derivation of Upper Intake Level) in 2000.

II.11.1 Normal exposure

Manganese is ubiquitous in the environment, occurring in soil, air, water, and food. Thus, all
humans are exposed to manganese and manganese is a normal component of the human
body. Food is usually the most important route of exposure for humans. The dietary intake of
adults has been estimated to range from 0.9 to 9.4 mg Mn/day in various countries (SCF,
2000). The intake can be higher for vegetarians because higher levels of manganese occur in
food of plant origin. The consumption of tea may contribute substantially.

Children are exposed to manganese in the same manner as adults, the main source of
exposure being food. Specific data for intake by this group, however, are not available
(ATSDR, 2000).

Based on the above information background daily intake of manganese for a child is
estimated to be 130 μg/kg bw/day. This is the adult intake calculated from the maximum of
the reported range of food intakes (9.4 mg/day), assuming a body weight of 70 kg.

II.11.2 Toxicology
In humans manganese is an essential nutrient that plays a role in bone mineralization, protein
and energy metabolism, metabolic regulation, cellular protection from damaging free radical
species, and the formation of glycosaminoglycans. As is pointed in SCF (2000), no formal
Recommended Dietary Allowance (RDA) for manganese is available. However, 2-5 mg/day
for adults has been derived as an ‘estimated safe and adequate dietary intake’ by the US
National Research Council. In 1993 the EU Scientific Committee for Food estimated 1-10
mg/day as an acceptable range of intakes.

Occupational studies have shown neurological effects after inhalation exposure to

manganese. These neurological effects have been observed following exposure durations that
span from 1 to 35 years. The characteristic syndrome is known as “manganism”. Symptoms
are weakness, anorexia, muscle pain, apathy, slow speech without inflection, emotionless
“mask-like” facial expression, and slow clumsy movement of the limbs. In general, these
effects are irreversible. The minimal exposure level producing neurological effects is not
certain but is probably in the range of 0.1-1 mg/m3 (WHO, 1996). Several human studies with
exposure via drinking-water suggest that ingestion of manganese can also lead to
neurological effects. A study by Kondakis et al. (1989) carried out in Northern Greece, found
RIVM report 320003001 page 187 of 234

higher prevalences of neurological signs of chronic manganese poisoning and increased

manganese concentration in the hair of older persons. In this study however there was
simultaneous exposure to manganese via food, which was presumably high but the exact
magnitude is unknown. Overall no oral NOAEL of LOAEL could be derived for manganese
neurotoxicity in humans. Oral animal data are also insufficient for an NOAEL or LOAEL.
The latter was the conclusion reached by SCF (2000).

II.11.3 Children as a sensitive subgroup

Children as a group have not been studied for the adverse effects of overexposure to
inorganic manganese. Thus no estimation of the quantitative susceptibility of children to the
preclinical effects of excess manganese exposure is possible (ATSDR, 2000).

II.11.4 Local effects upon dermal contact

No toxicity data are available for the dermal exposure route (ATSDR, 2000).

II.11.5 Absorption
According to ATSDR (2000) the amount of manganese absorbed across the gastrointestinal
tract in humans is variable but typically averages about 3–5%. Limited human and animal
data suggest that children/young animals may have a somewhat higher absorption of
manganese than adults. Quantification of this potential difference is however not possible.

II.11.6 Toxicological limit values for ingestion of manganese

SCF (2000) concluded that the available data indicate manganese is neurotoxic after oral
intake despite its poor absorption in the gastrointestinal tract. However, the limitations of the
human data and the non-availability of NOAELs for critical endpoints from animal studies
preclude derivation of an upper level (UL).

US-EPA (1996) noted the limited data set for the oral route. Rodents were concluded not to
provide a good experimental model for manganese toxicity. In its derivation of an RfD US-
EPA therefore focussed on what is known to be a safe oral intake of manganese for the
general human population. Based on estimates of ‘safe and adequate manganese intake
levels’ by US organisations and measured levels of normal dietary intake US-EPA concluded
that 10 mg/day (0.14 mg/kg bw/day) is an appropriate Reference Dose for manganese.
Similarly ATSDR (2000) adopted a US estimate of 5 mg/day (0.07 mg/kg bw/day) as the
‘safe and adequate daily dietary intake’ as its provisional chronic oral MRL for manganese.

OEHHA (2004) in its draft derivation of a Reference Dose for manganese specifically for
children (ChRD) developed several approaches. Using an estimate by the US Food and
Nutrition Board of an NOAEL for manganese of 11 mg/day for an adult and subtracting from
this level a normal mid-range dietary intake of 5 mg/day led to non-dietary NOAEL of
page 188 of 234 RIVM report 320003001

6 mg/day or 0.086 mg/kg bw/day (assumed body weight 70 kg). To this NOAEL then an
uncertainty factor of 3 was applied for the protection of infants and children yielding a ChRD
of 0.03 mg/kg bw/day. Alternative calculations were based on neurotoxic endpoints as
determined in two studies in neonatal rats (increased acoustic startle response; righting,
homing, and passive avoidance tests). This involved using a LOAEL of 11 mg/kg bw/day,
based on which a possible ChRD of 0.01 mg/kg bw/day was calculated (uncertainty factor
1000) and an NOAEL of 8.3 mg/kg bw/day, based on which a possible ChRD of 0.08 mg/kg
bw/day was calculated (uncertainty factor 100). In conclusion OEHHA proposed a value of
0.03 mg/kg bw/day. Given its derivation this ChRD refers to exposures above normal dietary
intake.

II.11.7 Conclusion
The ChRD as proposed by OEHHA (2004) of 0.03 mg/kg bw/day is chosen as the best
available value for toy-related exposures. Given its derivation this ChRD refers to exposures
above normal dietary intake.

As to possible adverse direct dermal effects, no conclusion is possible due to lack of data.

References
ATSDR (2000) Toxicological profile for manganese. US Agency for Toxic Substances and
Disease Registry, report dated September 2000.

IPCS (1999) Concise International Chemical Assessment Document 12: Manganese and its
compounds. International Programme on Chemical Safety, World Health Organization,
Geneva, 1999.

OEHHA (2004) Development of health criteria for school site risk assessment pursuant to
health and safety code section 901(g): Proposed child-specific Reference Dose (ChRD) for
school site risk assessment - Manganese and Pentachlorophenol. Public Review Draft
November 2004. Integrated Risk Assessment Section Office of Environmental Health Hazard
Assessment, California Environmental Protection Agency.

SCF (2000) Opinion of the Scientific Committee on Food on the Tolerable Upper Intake
Level of Manganese (expressed on 19 October 2000). SCF/CS/NUT/UPPLEV/21 Final 28
November 2000.

US-EPA (1996) Oral RfD Assessment – Manganese. Last revised 05/01/1996.

WHO (1996) Guidelines for Drinking-water quality - Second Edition . WHO, Geneva, 1996.
RIVM report 320003001 page 189 of 234

II.12 Mercury

On the toxicity of mercury and mercury compounds a large literature exists. Mercury occurs
as a metal (element), as inorganic salt or as organic mercury (methylmercury). For toy-related
exposures only inorganic forms are considered relevant. Inorganic mercury was evaluated
within the scope of the WHO Drinking-water guidelines in 1996. RIVM (2001) and ATSDR
(1999) also carried out evaluations, as did more recently IPCS (2003).

II.12.1 Normal exposure

Mercury is a naturally occurring element (around 80 µg/kg) in the Earth’s crust. Elemental
mercury may occur in both liquid and gaseous states. Inorganic mercury compounds include
mercurous chloride, mercuric chloride, mercuric acetate, and mercuric sulfide. Major
anthropogenic sources of mercury in the environment have been mining operations, industrial
processes, combustion of fossil fuels (especially charcoal), production of cement, and
incineration of municipal, chemical, and medical wastes. Dental amalgam fillings are the
primary source of inorganic mercury exposure for the general population. Estimates of daily
intake from amalgam restorations range from 1 to 27 µg/day, with the majority of dental
amalgam holders being exposed to less than 5 µg Hg/day. Average additional daily intake of
inorganic mercury was estimated at 4.3 µg/day (IPCS, 2003). In RIVM (2001) total
background exposure to elemental mercury and inorganic mercury was estimated to be
0.1 µg/kg bw/day.

Based on the above information background daily intake of elemental mercury and inorganic
mercury for a child is estimated to be 0.1 μg/kg bw/day (adult intake as presented by RIVM
(2001) adopted).

II.12.2 Toxicology
As is pointed out in ATSDR (1999) the kidney is the primary site for mercuric ion toxicity
because in fulfilling its major role of filtering and purifying the blood, the kidney is
continually exposed to ionic mercury. For both elemental mercury and inorganic mercury
renal toxicity has been observed in humans. Oral dose response data for humans however, are
scarce, for which reason the risk assessment for this route has been based on animal data. For
elemental mercury no oral data are available. Within the US-NTP mercuric chloride was
tested in rats. In a 26-week study in rats renal toxicity was seen at ≥ 46 mg Hg/kg bw/day
with an NOAEL of 0.23 mg Hg/kg bw/day. In a 2-year study in rats 1.9 mg Hg/kg bw/day
was the LOAEL for renal toxicity (IPCS, 2003).

II.12.3 Children as a sensitive subgroup

On the effect of inorganic mercury in children or young animals no data are available
(ATSDR, 1999).
page 190 of 234 RIVM report 320003001

II.12.4 Local effects upon dermal contact

Human case studies suggest that dermal contact with elemental mercury and mercuric salts
may produce dermatitis. Dose response information for this possible effect, however, is
lacking. No animal data are available for this endpoint. The skin-irritating potential of
inorganic or metallic mercury is known insufficiently. Use of ointments in which the
compounds were present has led to adverse, irritative skin reactions in some instances but the
dose response for these effects remains unclarified (ATSDR, 1999).

II.12.5 Absorption
Absorption of inorganic mercuric salts may range from 2 to 38% depending upon the form
and test conditions. Oral absorption of elemental mercury is negligible. Human data on
absorption are scarce. In older animal studies only low absorption percentages were found for
inorganic mercury (1-8.5%) but in more recent ones percentages were higher (25-40%). A
study in mice indicated that young animals absorb considerably more (38% compared to 7%
in adult animals) (ATSDR, 1999).

II.12.6 Toxicological limit values for ingestion of inorganic mercury

ATSDR (1999) derived an intermediate duration MRL of 0.002 mg Hg/kg bw/day oral
exposure to inorganic mercury. No chronic MRL was derived due to lack of appropriate data.
The intermediate MRL was based on the NOAEL of 0.23 mg Hg/kg bw/day for renal effects
in rats from the NTP study already mentioned above. This dose was duration-adjusted for a
5 day/week exposure and divided by an uncertainty factor of 100 (10 for extrapolation from
animals to humans and 10 for human variability).

RIVM (2001) proposed a TDI of 0.002 mg Hg/kg bw/day, like ATSDR using the NOAEL of
0.23 mg Hg/kg bw/day for renal effects from the NTP study and also applying an uncertainty
factor of 100. IPCS (2003) proposed the identical derivation as a tolerable intake.

II.12.7 Conclusion
The TDI of 0.002 mg Hg/kg bw/day as proposed by RIVM (2001) and IPCS (2003), is
chosen as the appropriate value for toy-related exposures.

As to possible adverse direct dermal effects, qualitatively it is known that elemental mercury
and mercury as salt may produce skin irritation and sensitization under certain conditions.
Quantitative data on this effect are however lacking.
RIVM report 320003001 page 191 of 234

References
ATSDR (1999) Toxicological profile for mercury. US Agency for Toxic Substances and
Disease Registry, report dated Draft July 1999.

EFSA (2004) Opinion of the Scientific Panel on Contaminants in the Food Chain on a request
from the Commission related to mercury and methylmercury in food (Request N° EFSA-Q-
2003-030). Adopted on 24 February 2004. The EFSA Journal (2004) 34, 1-14.

IPCS (2003) Concise International Chemical Assessment Document 50: Elemental mercury
and inorganic mercury compounds: human health aspects. World Health Organization,
Geneva, 2003.

RIVM (2001) Re-evaluation of human-toxicological Maximum Permissible Levels. RIVM

report no. 711701025, dated March 2001.

WHO (1996) Guidelines for Drinking-water quality - Second Edition . WHO, Geneva, 1996.
page 192 of 234 RIVM report 320003001

II.13 Nickel

The toxicology of nickel and nickel compounds has been evaluated by WHO within its
drinking-water programme (WHO, 1996). RIVM reviewed nickel as a soil contaminant in
2001. Comprehensive reviews are those by TERA (1999), ATSDR (2005) and by the EU
(EU-RAR, 2005). US-EPA (1996), OEHHA (2001) and EFSA (2005) are further reviews
specifically focussed on oral toxicity.

II.13.1 Normal exposure

Nickel and its compounds are naturally present in the Earth's crust, and releases to the
atmosphere occur from natural discharges such as windblown dust and volcanic eruptions, as
well as from anthropogenic activities. The latter are the dominant source for environmental
release. The general population is exposed to low levels of nickel in ambient air, water, and
food. Food is the most important source with drinking-water on average being ten times
lower. Specific foods high in nickel content are cocoa and soybeans. EU-RAR (2005) gives
an estimate for the total intake of nickel via food and drinking-water of 250 µg/day, based on
UK data from 2003. Also mentioned is an estimate by the Council of Europe of 400 µg/day.
Canadian data for children and adults indicate that nickel intake for children (3-12 years old)
is twice that for adults on a body weight basis (EU-RAR, 2005). RIVM (2001) estimated an
average adult intake of 4 µg/kg bw/day. This figure is in line with the estimates as presented
in EU-RAR (2005).

An important consumer exposure to nickel is through skin contact with objects such as
earrings, medallions, buttons, metal wires in clothing, wrist watches, rings etc. These
exposures may lead to nickel dermatitis. The prevalence of nickel sensitivity in the
population is about 8-14.5% for adult women and about 1% for men (WHO, 1996).

Based on the above information background daily intake of nickel for a child is estimated to
be 8 μg/kg bw/day, which is twice the adult intake as presented by RIVM (2001).

II.13.2 Toxicology
Nickel is essential for the catalytic activity of some plant and bacterial enzymes but
biochemical functions in humans and higher animals have not been demonstrated. Nickel
compounds are recognized human carcinogens via the inhalation route (IARC Group I). For
the oral route, however, such evidence is lacking. A recent 2-year study with nickel sulfate in
rats (CRL, 2005) showed no carcinogenic response after oral (gavage) application.
Genotoxicity data have shown effects at the chromosome level (aberrations, SCEs) occurring
at high, toxic doses, most likely due to indirect mechanisms. In view of these data a threshold
approach is warranted for the oral route, which conclusion is in line with RIVM (2001),
EFSA (2005) and EU-RAR (2005)
RIVM report 320003001 page 193 of 234

In studies on subchronic toxicity, the main targets for the toxicity of orally ingested nickel
salts are kidneys, spleen, lungs, and the myeloid system. These studies mostly were limited in
design. In a 90-day study in rats by ABC (1988) with gavage administration of nickel
chloride, clinical signs of toxicity were seen, body weights and weights of kidney, liver and
spleen were reduced and mortality increased. The NOAEL in this study was 5 mg/kg bw/day.
In a 2-year feeding study in rats by Ambrose et al. (1976) with elemental nickel decreased
body weight was the critical effect with an NOAEL of 5 mg Ni/kg bw/day. This study,
however, was flawed because of high mortality in all groups, including the control. In the
new 2-year study in rats with nickel sulphate by CRL (2005), decreased body weight and
increased mortality were the critical effects. The NOAEL in this study was 2.2. mg Ni/kg
bw/day, it is concluded in EU-RAR (2005), while adding that uncertainty remains because
the effects were present to a statistically non-significant degree at this level as well. EFSA
(2005) notes that in a reproduction study in rats (Smith et al., 1993) with nickel chloride
administration via drinking-water, peri-natal mortality was increased, even at the lowest
administered dose of 1.3 mg Ni/kg bw/day. In the EU-RAR (2005) a similar effect is reported
at 2.2 mg/kg bw/day in a 2-generation study in rats referred to as SLI (2000b) with gavage
application of nickel sulphate. The NOAEL in this study was 1.1 mg Ni/kg bw/day.

In individuals suffering from dermal nickel allergy, oral intake of low doses can provoke
eczema. This has been examined in a several oral challenge studies, in which single oral
doses of a few mg nickel provoked dermal reactions in nickel-sensitised subjects. EFSA
(2005) cites studies by Nielsen et al. (1990, 1999) who report lowest oral doses, given to
nickel sensitive subjects and reported to exacerbate hand eczema, of 0.49 mg/day in a high
nickel diet (equivalent to about 8 µg Ni/kg bw/day), and 12 µg/kg bw/day given in drinking
water on an empty stomach.

II.13.3 Children as a sensitive subgroup

Only limited data are available. Some epidemiological surveys suggest young girls are more
sensitive to nickel-induced dermatitis but most likely this just reflects increased exposure by
this group. Further data are lacking (ATSDR, 2005).

II.13.4 Local effects upon dermal contact

A large literature exists on nickel induced dermatitis. The dose response for this effect has
been examined in human experiments. Human studies on the threshold for induction of nickel
dermatitis are not available. Indeed, such studies are contra-indicated for ethical reasons.
Menné et al. (1987) found high elicitation percentages (56-81%) among a group of
173 nickel-sensitive subjects after exposure to alloys with release levels of 10 to
80 μg/cm2/week. Studies with nickel sulfate have shown that even a very low patch exposure
for 48 hours to 0.05 μg Ni/cm2 may elicit a response in nickel-sensitive subjects (study by
Uter et al. 1995). As is concluded in the EU-RAR for nickel, on the basis of the available
page 194 of 234 RIVM report 320003001

data it is not possible to set a scientifically based threshold for elicitation (NOEL) in nickel-
sensitised subjects. The EU-RAR notes that Danish regulation limiting nickel release from
objects in direct contact with skin to less than 0.5 μg Ni/cm2/week, has resulted in a
significant reduction of prevalence of nickel sensitisation. In addition data suggest that this
release level is sufficient to prevent reactions in a significant proportion of nickel-sensitised
subjects. But complete protection, the EU-RAR adds, for the most sensitive subjects may
only be achieved at levels an order of magnitude lower than the limit of 0.5 μg Ni/cm2/week
(EU-RAR, 2005).

II.13.5 Absorption
The available evidence is reviewed in the EU-RAR (2005). Nickel absorption from the
gastro-intestinal tract depends in part on the solubility of the nickel compound ingested, with
insoluble forms having lower absorption. Poorly soluble compounds, however, may be more
soluble in gastric juice and thus still be absorbed. More specific data on the latter point,
however, are lacking. Another important factor is the matrix in which the compound is
present. Uptake from water is higher than from food, especially under fasted conditions.
Nickel absorption following administration in drinking-water to fasting subjects may be as
high as 25-27% whereas it was 1-6% in non-fasted subjects and/or in close proximity with a
meal. For fasting-conditions the EU-RAR concluded to an absorption percentage of 30% and
for other exposure scenarios to a percentage 5% (EU-RAR, 2005).

II.13.6 Toxicological limit values for ingestion of nickel

US-EPA (1996) derived an RfD for soluble nickel compounds of 20 μg Ni/kg bw/day based
on an NOAEL of 5 mg Ni/kg bw/day from the study by Ambrose et al. (1976). An
uncertainty factor of 300 was applied, consisting of a factor of 10 for interspecies
extrapolation, 10 to protect sensitive populations and an additional factor of 3 to account for
inadequacies in the reproductive studies.

ATSDR (2005) derived no oral MRLs for nickel due to lack of appropriate data. EFSA
(2005) derived no Tolerable Upper Intake Level for nickel because adequate dose response
data on the effect of oral nickel in nickel-sensitised subjects was lacking. The Panel noted
that oral intakes of nickel as low as about 8 µg/kg body weight/day have been reported to
aggravate hand eczema in nickel-sensitised subjects.

RIVM (2001) proposed a TDI of 50 μg Ni/kg bw/day based on an NOAEL of 5 mg/kg

bw/day from both a 90-day study in rats by ABC (1988) and the 2-year study by Ambrose et
al. (1976). This NOAEL was divided by an uncertainty factor of 100 (10 for interspecies
extrapolation and 10 for intraspecies extrapolation).

OEHHA (2001) identified the oral dose of 1.12 mg Ni/kg bw/day as the appropriate NOAEL
for deriving its guideline value for drinking-water. This NOAEL was obtained from a
RIVM report 320003001 page 195 of 234

reproduction study by SLI (2000b). A total uncertainty factor of 1000 was used, including a
factor of 10 for interspecies extrapolation, 10 for intra-species variability, and an additional
10 to account for the potential carcinogenicity of soluble nickel by the oral route. This
derivation implies an oral limit value of 1.1 μg Ni/kg bw/day.

As already stated, a new 2-year study is available and its NOAEL is recommended in the EU-
RAR (2005) for assessing repeated dose toxicity. This NOAEL was 2.2 mg Ni/kg bw/day.
For developmental toxicity the EU-RAR (2005) concludes to an overall NOAEL of
1.1 mg/kg bw/day (LOAEL 2.2. mg/kg bw/day) derived from the rat 2-generation study by
SLI (2000b). Using an uncertainty factor of 100 (10 for interspecies extrapolation and 10 for
intra-species extrapolation), from the latter NOAEL a TDI of 10 μg Ni/kg bw/day can be
derived.

II.13.7 Conclusion
For nickel a TDI of 10 μg Ni/kg bw/day is proposed based on the NOAEL of a recent
2-generation study in rats, as evaluated in the EU-RAR (2005).

As to possible adverse direct dermal effects, nickel is notorious for its potency to induce
dermal contact allergy. For persons already sensitised very low doses may suffice for
producing symptoms. Danish regulation limits nickel release from objects in direct contact
with skin to less than 0.5 μg Ni/cm2/week. This regulation has resulted in a significant
reduction of prevalence of nickel sensitisation. The data suggest that this release level is
sufficient to prevent reactions in a significant proportion of nickel-sensitised subjects. But for
complete protection for the most sensitive subjects the limit may have to be an order of
magnitude lower than the limit of 0.5 μg Ni/cm2/week.

References
ABC (1988) Ninety day gavage study in albino rats using nickel. Final report submitted to
US EPA by Research Triangle Institute and American Biogenics Corporation. As cited in
ATSDR (2005).

Ambrose, A.M., D.S. Larson, J.R. Borzelleca and G.R. Hennigar, Jr. (1976) Long-term
toxicologic assessment of nickel in rats and dogs. J. Food Sci. Technol. 13: 181-187. As
cited in US-EPA (1996).

ATSDR (2005) Toxicological profile for nickel. US Agency for Toxic Substances and
Disease Registry. August 2005.
page 196 of 234 RIVM report 320003001

CRL (2005) A two-year oral (gavage) carcinogenicity study in Fischer 344 rats with nickel
sulphate hexahydrate. Study no. 3472.7 carried out for NiPERA. Charles River Laboratories
Inc. As cited in EU-RAR (2005).

EFSA (2005) Opinion of the Scientific Panel on Dietetic Products, Nutrition and Allergies on
a request from the Commission related to the Tolerable Upper Intake Level of Nickel.
Request N° EFSA-Q-2003-018) (adopted on 25 January 2005 by written procedure). The
EFSA Journal (2005) 146, 1-21

EU-RAR (2005) Risk Assessment reports for: Nickel sulphate, Nickel chloride, Nickel
dinitrate and Nickel carbonate. Background document Nickel and Nickel compounds. Draft
for final approval November 2005.

Menne T, Brandrup F, Thestrup-Pedersen K, et al. 1987. Patch test reactivity to nickel alloys.
Contact Dermatitis 16:255-259. As cited in EU-RAR 2005.

Nielsen GD, Jepsen LV, Jorgensen PJ, Grandjean P, Brandrup F (1990). Nickel-sensitive
patients with vesicular hand eczema: oral challenge with a diet naturally high in nickel. Br J
Dermatol 122: 299-308. As cited in EFSA (2005).

Nielsen GD, Soderberg U, Jorgensen PJ, Templeton DM, Rasmussen SN, Andersen KE,
Grandjean P (1999). Absorption and retention of nickel from drinking water in relation to
food intake and nickel sensitivity. Toxicol Appl Pharmacol 154: 67-75. As cited in EFSA
(2005).

OEHHA (2001) Public Health Goal for Nickel in drinking water. Prepared by Office of
Environmental Health Hazard Assessment, California Environmental Protection Agency,
Pesticide and Environmental Toxicology Section. Dated August 2001.

RIVM (2001) Re-evaluation of human-toxicological Maximum Permissible Levels. RIVM

report no. 711701025, dated March 2001.

Uter W, Fuchs T, Hausser M, Ippen H (1995) Patch test results with serial dilutions of nickel
sulfate (with and without detergent), palladium chloride, and nickel and palladium metal
plates. Contact Dermatitis 32, 135-142. As cited in EU-RAR (2005).

SLI (2000b) An oral (gavage) two-generation reproduction study in Sprague Dawley rats
with nickel sulphate hexahydrate. Study no. 3472.2 carried out for NiPERA. Springborn
Laboratories Inc. Spencerville, Ohio USA. As cited in EU-RAR (2005).
RIVM report 320003001 page 197 of 234

Smith MK, George EL, Stober JA, Feng HA, Kimmel GL (1993). Perinatal toxicity
associated with nickel chloride exposure. Environm Res 61: 200-211. As cited in EFSA
(2005).

TERA (1999) Toxicological Review of Soluble Nickel Salts. Prepared for: Metal Finishing
Association of Southern California, Inc., U. S. Environmental Protection Agency and Health
Canada. Report dated March 1999.

US-EPA (1996) IRIS-file for Nickel, soluble salts. Oral RfD Assessment, Lst revised
12/01/1996.

WHO (1996) Guidelines for Drinking-water quality - Second Edition . WHO, Geneva, 1996.
page 198 of 234 RIVM report 320003001

II.14 Selenium

Selenium and selenium compounds have been evaluated within the scope of the WHO
Drinking-water guidelines in 1996. Another evaluation is by SCF (2000), which committee
derived a Tolerable Upper Intake Level for selenium. RIVM, within the project for soil
intervention values, evaluated selenium in 1998 (RIVM, 1998). Further reviews were
published by US-EPA (1991) and US-ATSDR (2003).

II.14.1 Normal exposure

Selenium is present in the earth’s crust often in association with sulfur-containing minerals. It
can assume four oxidation states (-2, 0, +4, +6) and occurs in many forms, including
elemental selenium, selenites and selenates. The principal releases of selenium into the
environment as a consequence of human activities result from the combustion of coal. For the
general population food is the primary exposure route followed by water and air. The greatest
portion of dietary intake occurs from organic forms of selenium, mainly the amino acids
selenomethionine and selencysteine, in grains, cereals, and forage crops. The main inorganic
sources of selenium in the diet are selenate and selenite, which are less absorbed than the
organic forms (ATSDR, 2003).

EFSA (2000) gives an overview of daily intake levels in European countries. The mean
intakes of non-vegetarian adults in different studies were: Belgium 28-61 µg/day, Denmark
41-57 µg/day, Finland 100-110 µg/day, France 29-43 µg/day, United Kingdom 63 µg/day,
The Netherlands 40-54 µg/day, Norway 28-89, and µg/day, Spain 79 µg/day, and Sweden
24-35 µg/day (SCF 2000). These data indicate daily intake levels up to about
1 µg/kg bw/day. Results from a dietary intake study carried out in the USA indicate that
children up to age 6 years have almost twice the intake of adults on a body weight basis
(ATSDR, 2003). Levels of selenium in tap water samples from public water supplies around
the world are usually much less than 10 µg/litre. Drinking-water from a high-selenium area in
China was reported to contain 50-160 µg/litre. In air the level of selenium (mostly bound to
particles) in most urban areas ranges from 0.1 to 10 ng/m3, but higher levels may be found in
certain areas, e.g. in the vicinity of copper smelters (WHO, 1996).

European data as summarised by SCF (2000) indicate a mean adult daily intake of
1 µg/kg bw/day. For children twice this figure should be a reasonable estimate, i.e.
2 µg/kg bw/day.
RIVM report 320003001 page 199 of 234

II.15 Toxicology

Being part of several enzymes, selenium is an essential element in humans and animals.
Estimated daily requirements as summarised in SCF (2000) range from 40 to about 50 µg/day
for adults with a lower limit of 20 µg/day.

Acute oral exposure to extremely high levels of selenium (e.g., several thousand times more
than normal daily intake) produces nausea, vomiting, and diarrhea in both humans and
laboratory animals. Acute oral exposure of humans to selenium has occasionally caused
cardiovascular symptoms, such as tachycardia, but no electrocardiographic abnormalities
were found in individuals from a human population chronically exposed to selenium. In
laboratory animals, acute- and intermediate-duration oral exposure to very large amounts of
selenium (approximately 100 times normal human intake) has produced myocardial
degeneration in laboratory animals. In certain areas with high natural levels of selenium
(selenoferous areas) chronic oral intake of very high levels (10–20 times more than normal)
via food and water occurs, leading to selenosis, the major effects of which are dermal and
neurological. As shown by affected populations in China, chronic dietary exposure to these
excess levels of selenium has caused diseased nails and skin and hair loss, as well
neurological problems, including unsteady gait and paralysis. Dose response information for
this effect comes form several Chinese studies published from 1989 through 1994. The
minimum daily dietary intake sufficient to cause symptoms of selenosis (i.e., hair or nail loss,
nail abnormalities, mottled teeth, skin lesions and changes in peripheral nerves) was about
1200 µg Se (range: 913-1907 μg Se). No clinical signs of selenosis were recorded in
individuals with blood selenium below 1000 μg/l, corresponding to an intake of about
850 μg/day, which has been taken as a NOAEL for clinical selenosis. Slight increases in
prothrombin-time and in the liver enzyme ALAT, indicating liver damage, have also been
observed in some selenium exposed populations but the clinical significance of these findings
remains unclear (SCF, 2000).

As to its carcinogenic potential IARC concluded there was inadequate evidence for
classification. Evidence suggests that some forms of selunium exert a anti-tumorigenic action
in animals and humans. Selenium sulfide however appears an exception, producing increased
tumor incidences after oral administration. The relevance of this compound for toy-related
exposures seems limited. In genotoxicity tests selenium compounds have shown both
genotoxic and anti-genotoxic effects. Generally the genotoxic effects were observed at high
dosages and the anti-genotoxic at low dosages (RIVM, 1998).

II.15.1 Children as a sensitive subgroup

Limited data in humans suggest than children may be less sensitive for selenium toxicity than
adults (ATSDR, 2003).
page 200 of 234 RIVM report 320003001

II.15.2 Local effects upon dermal contact

Limited data suggest that selenium and compounds have only low potential for inducing
irritation and sensitisation (ATSDR, 2003).

II.15.3 Absorption
Selenium compounds are generally readily absorbed from the human gastrointestinal tract.
The physical state of the compound (e.g., solid or solution) the chemical form of selenium
(e.g., organic, inorganic), and the dosing regimen are factors influencing absorption.
Generally absorption percentages of 80% and higher have been observed in human volunteers
(ATSDR, 2003).

II.15.4 Toxicological limit values for ingestion of selenium

US-EPA (1991) used a Chinese epidemiological study by Yang et al. (1989) for deriving a
human NOAEL for selenosis. The LOAEL derived from this study was 1.26 mg Se/day and
the NOAEL 0.85 mg/day (0.015 mg/kg bw/day). An uncertainty factor of 3 to account for
sensitive individuals was applied, leading to an RfD of 5 μg/kg bw/day.

RIVM (1998) concurred with the approach developed by US-EPA. Thus a TDI of
5 μg/kg bw/day was proposed. ATSDR (2003), like US-EPA, concluded to an NOAEL from
the Chinese studies of 0.015 mg/kg bw/day. With an uncertainty factor of 3 a chronic MRL of
0.005 mg/kg bw/day was proposed (ATSDR, 2003).

SCF (2000) also used an NOAEL of 0.85 mg/day as derived from the Chinese epidemiology
studies. It was pointed out that other studies from the USA and Venezuela supported this
NOAEL. Application of an uncertainty factor of 3 to allow for the remaining uncertainties of
the studies used led to Tolerable Upper Intake Level (UL) of 300 μg/day. No specific UL for
children was derived because of lack of appropriate data.

II.15.5 Conclusion
The limit values as reviewed are in agreement. Thus a value of 5 μg/kg bw/day is chosen as
the most appropriate value for toy-related exposures.

As to possible adverse direct dermal effects, the limited data available suggest that selenium
and compounds have only low potential for inducing irritation and sensitisation

References
ATSDR (2003) Toxicological profile for selenium. US Agency for Toxic Substances and
Disease Registry, report dated September 2003.
RIVM report 320003001 page 201 of 234

IARC (1998) http://www-cie.iarc.fr/htdocs/monographs/vol09/selenium.html

RIVM (1998) Maximum Permissible Risk Levels for human intake of soil contaminants:
fourth series of compounds. RIVM report no. 715810004, dated March 1998.

SCF (2000) Opinion of the Scientific Committee on Food on the Tolerable Upper Intake
Level of Selenium (expressed on 19 October 2000). Scientific Committee Food,
SCF/CS/NUT/UPPLEV/25 Final, 28 November 2000.

US-EPA (1991) IRIS-file Selenium and compounds. Derivation of RfD, last revised 09-01-
1991.

WHO (1996) Guidelines for Drinking-water quality - Second Edition . WHO, Geneva, 1996.

Yang, G., S. Yin, R. Zhou, et al. 1989 Studies of safe maximal daily dietary Se-intake in a
seleniferous area in China. II. Relation between Se- intake and the manifestation of clinical
signs and certain biochemical alterations in blood and urine. J. Trace Elem. Electrolytes
Health Dis. 3(2): 123-130. As cited in US-EPA (1991), SCF (2000) and ATSDR (2003).
page 202 of 234 RIVM report 320003001

II.16 Silver

The oral toxicity of silver and its compounds has been reviewed by ATSDR (1990), RIVM
(1995), and WHO (1996).

II.16.1 Normal exposure

Silver is a rare element, which occurs naturally in its pure form as a white, ductile metal, and
in ores. It has an average abundance of about 0.1 ppm in the earth's crust and about 0.3 ppm
in soils.
US data indicate that food and drinking-water are the major sources of exposure, totalling an
estimated 0.06-1.3 μg/kg bw/day (RIVM, 1995). Data specifically for children are lacking.

Based on the above information background daily intake of silver for a child is estimated to
be 1.3 μg/kg bw/day. This the maximum of the adult range as estimated by RIVM (1995).

II.16.2 Toxicology
In animal studies toxic effects were seen at high dose levels only (> 90 mg/kg bw/day). In
humans the critical effect is argyria, a medically benign but permanent bluish-gray
discoloration of the skin. Argyria results from the deposition of silver in the dermis and also
from silver-induced production of melanin. Although silver has been shown to be uniformly
deposited in exposed and unexposed areas, the increased pigmentation becomes more
pronounced in areas exposed to sunlight due to photoactivated reduction of the metal.
Although the deposition of silver is permanent, it is not associated with any adverse health
effects. No pathologic changes or inflammatory reactions have been shown to result from
silver deposition.

Silver compounds have been employed for medical uses for centuries. In the nineteenth and
early twentieth centuries, silver arsphenamine was used in the treatment of syphillis; more
recently it has been used as an astringent in topical preparations. From a case review
concerning intravenous use of silver arsphenamine in syphilis patients US-EPA (1995)
concluded to a LOAEL for mild argyria of 0.014 mg/kg bw/day7 for this sensitive
subpopulation.

7
This derivation was based on:
- the body accumulates silver throughout life
- a total intravenous dose of 1 g silver (4 g silver arsphenamine) will cause mild argyria in the most sensitive
individuals
- an oral absorption factor of 4% to calculate the oral dose equivalent to the i.v. dose of 1 g
- the total dose is averaged over a lifetime of 70 years
RIVM report 320003001 page 203 of 234

II.16.3 Children as a sensitive subgroup

No data are available.

II.16.4 Local effects upon dermal contact

Medical case histories describe mild allergic responses attributed to dermal contact with
silver and silver compounds. The exposure concentrations involved in these cases are
unknown. Dermal contact with silver compounds may lead to local argyria; quantitative data
on this effect (dose response relation) are lacking (ATSDR, 1990).

II.16.5 Absorption
Absorption of silver was examined in four animal species and was found to be very low.
From this study 4.4% was derived as a conservative estimate for absorption of silver in
human after ingestion (US-EPA, 1996).

II.16.6 Toxicological limit values for ingestion of zinc

US-EPA (1996) derived an RfD of 0.005 mg/kg bw/day by dividing the LOAEL for mild
argyria of 0.014 mg/kg bw/day by a factor of 3 (modifying factor). A higher factor was
considered unwarranted given the non-adverse nature of the critical effect and the fact that
the study was done in a sensitive subpopulation. This approach was adopted by RIVM
(1995), leading to a TDI of 0.005 mg/kg bw/day.

II.16.7 Conclusion
The RfD 0.005 mg/kg bw/day as proposed by US-EPA (1996) is chosen as the appropriate
value for toy-related exposures.

As to possible adverse direct dermal effects, case reports indicate silver may produce allergic
reactions and local argyria. The dose response relation for these effects is unknown. Despite
these reports of adverse effects, the fact that humans are widely exposed to silver in jewelry
without this leading to a high prevalence of medical complaints, suggests that the risk for
dermal effects is low.

References
ATSDR (1990) Toxicological profile for silver. US Agency for Toxic Substances and
Disease Registry.

RIVM (1995) Human-Toxicological Criteria for Serious Soil Contamination: Compounds

evaluated in 1993 & 1994. RIVM Report no 715810009.

US-EPA (1996) Oral RfD Assessment – Silver. Last revised 12/01/1996.

page 204 of 234 RIVM report 320003001

WHO (1996) Guidelines for Drinking-water quality - Second Edition . WHO, Geneva, 1996.
RIVM report 320003001 page 205 of 234

II.17 Strontium

The oral toxicity of strontium and its compounds has been reviewed by RIVM (1989), US-
EPA (1996) and ATSDR (2004).

II.17.1 Normal exposure

Strontium is ubiquitous in the environment and is present in nearly all rocks and soils.
Chemically it resembles calcium. Food and drinking-water are the main sources of
background exposure. Dutch data indicate a mean total daily intake of 1.3 mg/person
(maximum 3.6 mg/person) (RIVM, 1998). Similar levels were reported for the USA and
Australia. Data specifically for children are lacking.

Based on the above information background daily intake of strontium for a child is estimated
to be 18 μg/kg bw/day. This the maximum of the mean adult range as reported in RIVM
(1998).

II.17.2 Toxicology
Strontium is able replace calcium in its physiological role and accordingly is incorporated in
bone tissue. Abnormal skeletal development is the most important toxicological effect
produced by strontium. Usable human toxicity data are lacking. Skeletal abnormalities were
observed in weanling rats after 20 days of dosing with 550 mg Sr/kg bw/day (LOAEL). The
NOAEL in weanling rats was 140 mg/kg bw/day. In adult rats the NOAEL was 690 mg/kg
bw/day in the same study (ATSDR, 2004).

II.17.3 Children as a sensitive subgroup

Animal data clearly show young animals to be more sensitive to strontium toxicity than adult
animals (see above). Human data on this point are very scarce. Because the immature
skeleton has a high rate of bone remodeling, and strontium adversely affects bone
development children must be expected to indeed be at increased risk.

II.17.4 Local effects upon dermal contact

No data are available (ATSDR, 2004).

II.17.5 Absorption
Absorption of strontium was examined in a number of human volunteer studies. The results
of these studies indicate that approximately 20% (range 11–28%) of ingested strontium is
absorbed from the gastrointestinal tract.
page 206 of 234 RIVM report 320003001

II.17.6 Toxicological limit values for ingestion of strontium

Based on an NOAEL of 140 mg/kg bw/day in weanling rats, ATSDR (2004) proposed an
intermediate MRL of 2 mg/kg bw/day. In this derivation an uncertainty factor of 90 was
applied (10 for extrapolation from animal to human and 3 for human variability, 3 for short
study duration and limited endpoint examination). A partial uncertainty factor was used to
account for human variability because the selected NOAEL was based on the response of
juveniles, which is also the most sensitive human group. ATSDR derived no chronic MRL
because appropriate data were lacking. Based on the same study US-EPA proposed an RfD of
0.6 mg/kg bw/day (US-EPA, 1996). They applied a total uncertainty factor of 300 (10 for
interspecies extrapolation, 10 for an incomplete database, including a lack of developmental
and reproductive data, 3 for for sensitive subpopulations). Again a low intra-species factor
was used because the NOAEL was for a sensitive subgroup.

II.17.7 Conclusion
The RfD 0.6 mg/kg bw/day as proposed by US-EPA (1996) is chosen as an appropriate value
for toy-related exposures.

As to possible adverse direct dermal effects, no conclusion is possible due to lack of data.

References
ATSDR (2004) Toxicological profile for Strontium. US Agency for Toxic Substances and
Disease Registry, Draft dated July 2004.

RIVM (1989) Strontiumchloride en strontiumacetaat in tandpasta. Interne Notitie RIVM d.d.

8 maart 1989. [In Dutch]

RIVM (1998) Duplicaat 24-uursvoedingen 1994 - Inname aan calcium, magnesium, barium,
strontium en mangaan. RIVM rapport nr. 515004008, d.d. Juli 1998. [In Dutch]

US-EPA (1996) IRIS-file Strontium. Derivation of RfD, last revised 12-1-1996.

RIVM report 320003001 page 207 of 234

II.18 Tin (inorganic)

The oral toxicity of inorganic tin and compounds has been reviewed by RIVM (1991),
WHO/JECFA (1982, 1989, 2001), IPCS (2005), EFSA (2005) and ATSDR (2005).

II.18.1 Normal exposure

Tin occurs naturally in the earth's crust with a concentration of approximately 2–3 ppm. The
major source of human exposure is through migration from tin cans to foods. Within the
European Union SnCl2 is a permitted food additive (E512) for bottled and canned white
asparagus.

Data on mean inorganic tin intake from food for the populations of seven countries
(Australia, France, Japan, Netherlands, New Zealand, the United Kingdom, and the USA)
indicate intakes ranging from < 1 up to 15 mg/person per day. Certain individuals who
routinely consume canned fruits, vegetables, and juices from unlacquered cans could ingest
up to 50–60 mg of tin daily (IPCS, 2005). Specifically for children JECFA (2001) cites a UK
study of 97 pre-school children (age 1.75–2.2 years) in which average daily intakes of
1.7-2.9 mg/day were found. Intake showed strong correlation with consumption of canned
foods. An Australian study among two-year-olds showed a mean intake of 1.3 mg/day
(JECFA, 2001).

Based on the above information background daily intake of inorganic tin for a child is
estimated to be 290 μg/kg bw/day. This figure is calculated from the maximum mean of
2.9 mg/day of the range for young children as reported in JECFA (2001), assuming a child
body weight of 10 kg.

II.18.2 Toxicology
Tin has not been shown to be essential for humans or animals, and there are no data on
deficiency effects resulting from an inadequate intake of inorganic tin. Inorganic tin has a low
systemic toxic potential due to its low absorption in the gastrointestinal tract. The only effect
in humans is acute irritation of the mucosa of the gastrointestinal tract (no known chronic
effects). This was seen in consumers drinking fruit juices containing high concentrations of
tin (≥ about 200 mg/kg product). In experimental animals anemia, liver and kidney damage
have been observed. In a sub-chronic feeding study in rats the NOAEL was
32 mg/kg bw/day. In a chronic feeding study in rats the NOAEL was 400 mg/kg diet
(equivalent to 20 mg/kg bw/day).

II.18.3 Children as a sensitive subgroup

No data are available.
page 208 of 234 RIVM report 320003001

II.18.4 Local effects upon dermal contact

No data are available (ATSDR, 2005).

II.18.5 Absorption
The absorption of inorganic compounds of tin from the gastrointestinal tract in humans and
animals is very low with as much as 98% being excreted directly in the faeces (EFSA, 2005).

II.18.6 Toxicological limit values for ingestion of strontium

Based on the level of 200 mg/kg in food as the approximate threshold for adverse
gastrointestinal effects in humans JECFA (1982) proposed a TDI of 2 mg/kg bw/day, a value
maintained in its later evaluations (JECFA, 1989, 2001). RIVM adopted this TDI in 1991.

Based on an NOAEL of 32 mg/kg bw/day from a subchronic feeding study in rats, ATSDR
(2003) proposed an intermediate MRL of 0.3 mg/kg bw/day. In this derivation an uncertainty
factor of 100 (10 for animal to human extrapolation and 10 for human variability) was
applied. ATSDR derived no chronic MRL because appropriate data were lacking.

EFSA (2005) noted that because of their limited absorption, orally ingested inorganic tin
compounds have low systemic toxicity in man and animals but concluded the available
evidence was insufficient for deriving an Upper Level for inorganic tin.

II.18.7 Conclusion
The JECFA (2001) TDI of 2 mg/kg bw/day is chosen as the appropriate value for toy-related
exposures.

As to possible adverse direct dermal effects, no conclusion is possible due to lack of data.

References
ATSDR (2005) Toxicological profile for Tin. US Agency for Toxic Substances and Disease
Registry.

EFSA (2005) Opinion of the Scientific Panel on Dietetic Products, Nutrition and Allergies on
a request from the Commission related to the Tolerable Upper Intake Level of Tin (Request
N° EFSA-Q-2003-018) (adopted on 6 July 2005). The EFSA Journal (2005) 254, 1-25

IPCS (2005) Concise International Chemical Assessment Document 65: Tin and inorganic tin
compounds.
RIVM report 320003001 page 209 of 234

RIVM (1991) Voorstel voor de humaan-toxicologische onderbouwing van C-

(toetsings)waarden. RIVM rapport nr. 725201005.

WHO/JECFA (1982) WHO Food Additives Series no. 17.

WHO/JECFA (1989) WHO Food Additives Series no. 24.

WHO/JECFA (2001) WHO Food Additives Series no. 46.

page 210 of 234 RIVM report 320003001

II.19 Tin (organic)

The oral toxicity of organic tin compounds has been evaluated by EFSA recently (EFSA
2004). Other reviews are those by JMPR (1992), US-EPA (1997), IPCS (1999) and RIVM
(1999).

II.19.1 Normal exposure

Organotin compounds are manmade chemicals used for several applications. The tri-
substituted compounds tributyltin (TBT) and triphenyltin (TPT) have been used extensively
as biocides in wood preservatives, in antifouling paints for boats and as pesticides. Mono-and
di-substituted compounds like monomethyltin (MMT), dimethyltin (DMT), dibutyltin (DBT),
mono-n-octyltin (MOT) and di-n-octyltin (DOT) are used in mixtures in various amounts as
PVC stabilizers, which use includes food contact materials. Organotins are lipophilic
contaminants sparingly soluble in water and easily adsorbed to particulate matter in the
aquatic environment. Hence, they accumulate in sediments where they are relatively
persistent and can be taken up by benthic organisms such as clams. Organotins tend to
accumulate in fish and other aquatic organisms. Because of their adverse effects on the
aquatic ecosystem the use of TBT and TPT as biocides in antifouling paints for boats has
been restricted (EFSA, 2004).

Food is the major source of human exposure to organotins. EFSA (2004) summarizes data on
dietary exposure from eight European Countries (Belgium, Denmark, Germany, France, Italy,
Netherlands, Greek and Norway). Fish and seafood are the primary sources of exposure.
Using the high mean fish/seafood consumption levels of 80 grams/day as prevalent in
Norway as a conservative paradigm, in combination with the median international
concentrations of TBT, DBT, and TPT, a total daily intake of 0.018 µg/kg bw/day was
calculated. When mean international concentrations were used, the calculated intake was
0.083 µg/kg bw/day. For the 95 percentile for fish/seafood consumption by Norwegians of
165 grams/day combined with the median international concentrations of TBT, DBT, and
TPT the total intake of organotins was 0.037 µg/kg bw/day. The same with the mean
international concentrations of TBT, DBT, and TPT led to 0.17 µg/kg bw/day. For high
fish/seafood consumers from Norway, consuming products in the high range of organotin
concentrations (95-percentile), an intake of 0.30 µg/kg bw/day was calculated.

Based on the above information background daily intake of organic tin for a child is
estimated to be 0.083 μg/kg bw/day. This is the mean calculated for adults in Norway as the
EU country with highest fish consumption.
RIVM report 320003001 page 211 of 234

II.19.2 Toxicology
The toxicity of organotins has been studied in numerous animal studies. TBT and TPT have
the largest data bases; DOT was also studied in a range of toxicity tests. Both tumorigenicity,
developmental and reproductive toxicity and neurotoxicity were observed consistently in
various studies but the critical endpoint for TBT, DBT, TPT and DOT was their
immunotoxicity. These compounds produce thymus atrophy with lymphocyte depletion in the
thymus, spleen and peripheral lymphoid tissues, decreases in immunoglobulin concentrations,
lymphopenia and decrease in white blood cells in rodents. This results in depression of
thymus dependent immunity. In vitro studies with human thymocytes indicate that these cells
are sensitive to organotins. Mechanistic data indicate a similar mode of action for the
different organotins. Based on this the effects of organotins can be considered additive. As to
potency the available results indicate equipotency of the various organotins. An overall
NOAEL of 0.025 mg/kg bw/day was derived from a chronic rat study with TBTO in which
reduced resistance to T. spiralis infection was the critical effect. In this study weanling rats
were dosed for 4-6 or 15-17 months. The same NOAEL was observed in a 2-year study in
rats carried out by the same laboratory (EFSA, 2004).

II.19.3 Children as a sensitive subgroup

As is pointed out in US-EPA (1997), rat data indicate young animals are more susceptible to
TBT immunotoxicity. The overall NOAEL, however, already includes this factor because it
stems from a study using weanling rats.

II.19.4 Local effects upon dermal contact

TBTO is an irritant of the eyes and skin in experimental animals. These effects were observed
at concentrations of ≥0.5% (skin) and 0.15% (eyes). A NOAEL for these endpoints is
lacking. In human beings, TBTO may cause severe dermatitis after direct skin contact
(conclusion based on case studies). This reaction has a delayed character, i.e. the symptoms
develop only several hours after the start of contact. The dose-effect relation for this effect is
unknown. The lowest effect concentration reported is 0.01 g/litre (value derived from a case
study). A NOAEL for this endpoint is lacking. The observed dermatitis is probably not a
hypersensitivity response. No effect was seen in a standard test for dermal sensitization in
guinea pigs with tributyltinoxide. Triphenyltin was tested in guinea pigs as the hydroxide
with a negative result but in a guinea pig study with the acetate a sensitising response
occurred. In skin irritation tests triphenyltin showed only a mild response at high
concentrations (RIVM, 2000).

II.19.5 Absorption
Human data are lacking. In rat studies with TBT, TPT and DOT absorption after oral
administration ranged form 20 to 55% (EFSA, 2004).
page 212 of 234 RIVM report 320003001

II.19.6 Toxicological limit values for ingestion of organotins

Based on the overall NOAEL for organotins of 0.025 mg/kg bw/day EFSA (2005) proposed a
group-TDI of 0.25 µg/kg bw/day. An uncertainty factor of 100 for interspecies and
interindividual variation was used in this derivation. US-EPA (1997) calculated a Benchmark
Dose (BMD) for 10% effect of 0.03 mg/kg bw/day from the same study as used by EFSA
(2004). This level was divided by a factor of 100 (10 for extrapolation form animals to
humans and 10 to protect sensitive humans) leading to RfD of 0.3 µg/kg bw/day.

II.19.7 Conclusion
The EFSA (2004) group-TDI for organotins of 0.25 µg/kg bw/day is chosen as the
appropriate value for toy-related exposures.

As to possible adverse direct dermal effects, some organotins are known as powerful dermal
irritants, producing dermatitis as a delayed reaction to dermal exposure (non-sensitizing
reaction). The dose-effect relation for this effect however is insufficiently known. The lowest
effect concentration reported is 0.01 g/litre (value derived from a case study). Other
organotins may pose a sensitisation risk, animal bioassay data indicate. Again the dose-
response relation for this effect has not been characterized.

References
EFSA (2004) Opinion of the Scientific Panel on Contaminants in the Food Chain on a request
from the Commission to assess the health risks to consumers associated with exposure to
organotins in foodstuffs (Question N° EFSA-Q-2003-110) Adopted on 22 September 2004.
The EFSA Journal (2004) 102, 1-119

IPCS (1999) Concise International Chemical Assessment Document 14: Tributyltin oxide.

JMPR (1992) Pesticide residues in Food - 1991: Toxicology evaluations. Joint Meeting of the
FAO Panel of Experts on Pesticide Residues in Food and the Environment and the WHO
Expert Group on Pesticide Residues, Geneva, 16-25 September 1991. WHO, Geneva, 1992.

RIVM (2000) Health Risk Assessment for Organotins in Textiles. RIVM report no. RIVM
report 613350 002, dated january 2000.

US-EPA (1997) Toxicological review Tributyltin Oxide (CAS No. 56-35-9) In Support of
Summary Information on the
Integrated Risk Information System (IRIS).

US-EPA (1997) IRIS-file Tributyltin oxide (TBTO)(CAS 56-35-9). Last revised 09/01/1997.
RIVM report 320003001 page 213 of 234

II.20 Zinc

The oral toxicity of zinc and its compounds has been reviewed by WHO (1996), IPCS
(2001), US-EPA (2005) and ATSDR (2005). The former Scientific Committee of the EU
evaluated zinc as a food mineral (derivation of Upper Intake Level) in 2003.

II.20.1 Normal exposure

Zinc is ubiquitous in the environment, constituting 20–200 ppm (by weight) of the earth’s
crust. Food is the most important route of exposure for humans. SCF (2003) summarizes
available data on dietary intakes in European countries. Mean values ranged from 7.5 to 12.1
mg/day (97.5 percentiles 13.6-23.5 mg/day). US data as cited by ATSDR (2005) indicate that
on a body weight basis children will ingest 2-3 times the adult amount.

Based on the above information background daily intake of zinc for a child is estimated to be
350 μg/kg bw/day. This is twice the adult intake as calculated from the maximum of the
reported range of mean food intakes (12.1 mg/day), assuming a body weight of 70 kg.

II.20.2 Toxicology
Zinc is an essential element for humans, as co-factor in enzymes playing a role in general
growth and development, in testicular maturation, neurological function, wound healing and
immunocompetence. Well-known zinc containing enzymes include superoxide dismutase,
alkaline phosphatase and alcohol dehydrogenase. Recommended dietary allowances as
proposed by the SCF in 1993 is 9.5 mg/day for adult males and 7.0 mg/day for females. US
guidelines recommend daily intakes of 11 mg/day and 8 mg/day for men and women
respectively (SCF, 2003). On a body weight basis US guidelines are somewhat higher in
young children (0.23 mg/kg bw/day versus 0.13-0.15 in adults) (US-EPA, 2005).

Zinc can be toxic when exposures exceed physiological needs. The effects of zinc
supplementation have been studied in several human studies of longer duration. As is
concluded by SCF (2003), chronic zinc toxicity is associated with symptoms of copper
deficiency. Overt adverse effects (e.g. anaemia, neutropaenia, impaired immune responses)
are evident only after feeding zinc in the form of dietary supplements in excess of
150 mg/day for long periods. At lower intake levels of 100-150 mg/day the picture is less
clear. Short-term balance studies, SCF points out, would indicate adverse effects on copper
retention at intakes as low as 18.2 mg/day but more recent longer-term balance studies
indicate that positive copper balance can be maintained at 53 mg/day zinc in post-menopausal
women for 90 days provided copper intakes are adequately high (3 mg/day). Overall the data
indicate an NOAEL of 50 mg/day for adults, SCF concludes.
page 214 of 234 RIVM report 320003001

II.20.3 Children as a sensitive subgroup

Infants, more than adults, appear to be particularly sensitive to zinc deficiency, possibly the
result of their higher zinc requirements on a per body weight basis. For toxic effects data are
limited to a few animals studies indicating young animals are more susceptible to excess
intake of zinc (no usable human data) (ATSDR, 2005).

II.20.4 Local effects upon dermal contact

At high concentrations inorganic zinc compounds are irritating to the skin. Zinc oxide
however is used to promote the healing of burns and wounds and is a well-known anti-
inflammatory agent used in creams for dermal care of babies and infants.

II.20.5 Absorption
Absorption of dietary zinc ranges from 15 to 60%. When zinc intake is increased, the
fractional absorption decreases and intestinal excretion increases while urinary losses remain
fairly constant. Under fasted conditions absorption was measured to be as high as 81%. When
humans are under-supplied in zinc absorption may be higher still. Zinc appears to be
absorbed by both passive diffusion and a saturable carrier-mediated process. The carrier-
mediated mechanism appears to be most important at low zinc levels (SCF, 2003; US-EPA,
2005).

II.20.6 Toxicological limit values for ingestion of zinc

SCF (2003) concluded to an NOAEL of 50 mg/day based on the absence of any adverse
effects on a wide range of relevant indicators of copper status (as the critical endpoint) in
human studies. An UF of 2 was applied because of the small number of subjects included in
relatively short-term studies but acknowledging the rigidly controlled metabolic experimental
conditions employed. Thus an UL of 25 mg/day was recommended. Extrapolated to a
1-3 year old an UL of 7 mg/day was recommended. The later figure equals about
0.5 mg/kg bw/day (body weight 15 kg).

Similarly to SCF, US-EPA (2005) concluded that available data indicate that the most
sensitive effects of zinc are alterations in copper status. Based on four human studies a mean
NOAEL of 0.91 mg/kg bw/day was calculated. A threefold intraspecies uncertainty factor
was applied to account for variability in susceptibility in human populations, giving an RfD
for zinc of 0.3 mg/kg bw/day.

ATSDR (2005) presented a similar approach as US-EPA (2005). Thus a chronic MRL of
0.3 mg/kg bw/day was derived.
RIVM report 320003001 page 215 of 234

II.20.7 Conclusion
The UL of 0.5 mg/kg bw/day as proposed by SCF (2003) is chosen as the appropriate value
for toy-related exposures.

As to possible adverse direct dermal effects, zinc compounds may irritating to skin at high
concentrations but the wide use of zinc oxide as an anti-inflammatory agent in dermal care
products for babies without side effects being reported, indicates a low risk for this endpoint.

References
ATSDR (2005) Toxicological profile for zinc. US Agency for Toxic Substances and Disease
Registry, report dated August 2005.

IPCS (2001) Environmental Health Criteria 221- Zinc. International Programme on Chemical
Safety, World Health Organization, Geneva, 2001.

SCF (2003) Opinion of the Scientific Committee on Food on the Tolerable Upper Intake
Level of Zinc (expressed on 5 March 2003). SCF/CS/NUT/UPPLEV/62 Final 19 March
2003.

US-EPA (2005) Toxicological review of zinc and compounds (CAS No. 7440-66-6). In
Support of Summary Information on the Integrated Risk Information System (IRIS).

US-EPA (2005) Oral RfD Assessment – Zinc and compounds. Last revised 08/03/2005.

WHO (1996) Guidelines for Drinking-water quality - Second Edition . WHO, Geneva, 1996.
page 216 of 234 RIVM report 320003001
RIVM report 320003001 page 217 of 234

III Current definitions and EU legislations on toys

Within the EU, regulations on toys are harmonized, based on Council Directive 88/378/EEC
on the approximation of the laws of Member States concerning the safety of toys.
According to these regulations, toys must not contain dangerous chemical substances or
preparations within the meaning of EU Council Directive 67/548/EEC and 88/379/EEC in
amounts which may harm the health of children using them. Furthermore, bioavailability
limits have been set for 8 elements: antimony, arsenic, barium, cadmium, chromium, lead,
mercury, selenium.
Additional restrictions on the use of substances in toys are specified in adaptations to Council
Directive 76/769/EEC on the approximation of the laws, regulations and administrative
provisions of the Member States relating to restrictions on the marketing and use of certain
dangerous substances and preparations, such as the adaptations regulating the use of
phthalates (2005/84/EC) and azo-colorants (2004/21/EC).

A definition of toy and toy material is needed to identify products for which limits for
elements need to be established.
The definition for ‘toy’ used in Council Directive 88/378/EEC is as follows:

“Any product or material designed or clearly intended for use in play by children of less
than 14 years of age”

Annex I of Directive 88/378/EEC provides a list of articles that are not regarded as toys:
1. Christmas decorations
2. Detailed scale models for adult collectors
3. Equipment intended to be used collectively in playgrounds
4. Sports equipment
5. Aquatic equipment intended to be used in deep water
6. Folk dolls and decorative dolls and other similar articles for adult collectors
7. 'Professional' toys installed in public places (shopping centres, stations, etc.)
8. Puzzles with more than 500 pieces or without picture, intended for specialists
9. Air guns and air pistols
10. Fireworks, including percussion caps (¹)
11. Slings and catapults
12. Sets of darts with metallic points
13. Electric ovens, irons or other functional products operated at a nominal voltage exceeding
24 volts
14. Products containing heating elements intended for use under the supervision of an adult in
a teaching context
15. Vehicles with combustion engines
16. Toy steam engines
page 218 of 234 RIVM report 320003001

17. Bicycles designed for sport or for travel on the public highway
18. Video toys that can be connected to a video screen, operated at a nominal voltage
exceeding 24 volts
19. Babies’ dummies
20. Faithful reproductions of real fire arms
21. Fashion jewellery for children
(¹) With the exception of percussion caps specifically designed for use in toys without
prejudice to more stringent provisions already existing in certain Member States.

NOTE: According to this list, jewellery for children is not considered a toy and as such does
not need to comply with the European Standard for the Safety of Toys (EN 71). However,
children’s jewellery may be a relevant group of products for which exposure assessments for
elements should be considered. Health Canada has recently proposed new Children’s
Jewellery Regulations under the Hazardous Products Acts because a large proportion of
costume jewellery sold in North America today contains lead. Several cases of lead poisoning
have been reported in the United States and Canada as a result of chewing or swallowing
leaded pendants (Florin et al., 2005; VanArsdale et al., 2004). A recent study by Maas et al.
(2005) found that many children’s jewellery items sold in California retail stores contained
high levels of lead, with an overall mean lead content of 27.4%.
RIVM report 320003001 page 219 of 234

IV Existing Toy Categories

IV.1 Basis of toy categories

To determine which categories are most appropriate for setting limits of substances such as
elements in toys, the basis of a number of different toy categories used for international
legislations and other purposes have been reviewed.

IV.2 Possible safety hazards

The Hazardous Products (Toys) Regulations of Health Canada employ toy categories that are
based on specific toy types and their possible safety hazards. Specific product categories
listed are: dolls and soft toys, pull and push toys, toy steam engines, finger paints, rattles,
elastic and batteries.
This type of categorization is useful to determine which group of toys poses a possible toxic
hazard, for example experimental chemistry sets. For elements in particular, however, very
limited information is available on which toys contain which elements. No group of toys can
therefore be identified as requiring special attention with regard to their potential toxic risk
from the presence of elements.

IV.3 Toy material

The European Committee for Standardization (CEN) provides a European Standard for the
Safety of toys (EN 71). In Part 3 of EN 71 (Migration of certain elements), toys are
categorized according to the material they consist of, to determine the applicable test
requirements for the migration of elements (European Committee for Standardization (CEN),
1994). The categories include:
• Coatings of paints, varnishes, lacquers, printing inks, polymers and similar coatings
• Polymeric and similar materials, including laminates, whether textile reinforced or not
• Paper and paper board
• Textiles (natural and synthetic)
• Other materials whether mass coloured or not (e.g. wood, leather and other porous
substances)
• Materials intended to leave a trace (e.g. the graphite materials in pencils and liquid ink
in pens)
• Pliable modelling materials, including modelling clays and gels
• Paints, including finger paints, varnishes, lacquers, glazing powders and similar
materials in solid or in liquid form appearing as such in the toy.
page 220 of 234 RIVM report 320003001

Categorisation of toys based on material is useful if different materials need a different

treatment with regard to the migration testing procedure. The tests required in the current
standard are designed to simulate toy material remaining in contact with stomach acid for a
period of time after swallowing. It should be noted that tests simulating exposure to elements
via other routes than the oral route are not included in this standard. As will be discussed
later, the dermal and the inhalation route may also be relevant routes of exposure to consider
in the risk assessment of elements in toys, particularly with regard to sensitisation.

IV.4 Contact scenarios

The workgroup CEN/TC52/WG9 carried out a risk assessment of organic compounds in toys
to ensure that the requirements for this standard were scientifically well-founded (European
Committee for Standardization (CEN), 2003). For the risk assessment of organic compounds
in toys, the following groups of toys were considered:
• Toys that might be sucked
• Toys or parts of toys that might be ingested
• Toys coming into contact with the skin
• Toys coming into contact with the mucous membranes
• Toys coming into contact with the eyes
• Toys containing volatile substances that could be inhaled
Based on these contact scenarios, it was concluded that the risks of organic chemicals in toys
could be addressed by assessment of the following three contact routes: ingestion, skin
contact and inhalation. The mucous membrane contact route was considered to be of minor
significance. The eye contact route was considered not relevant, as any injury from toys
would most likely be of a physical rather than a chemical nature.
The same contact scenario categories may apply to elements in toys, although elements are
usually not in a volatile form. However, as explained in chapter 3 substances do not
necessarily need to be in volatile form to be available for inhalation.

IV.5 Type of toy

In the final draft of Part 9 (Organic chemical compounds – requirements), the applicable
migration or contact limits of organic chemical compounds depend on the type of toy and on
the toy material. Types of toys distinguished are:
RIVM report 320003001 page 221 of 234

This type of categorization is useful to quickly look up the applicable contact limits in a
certain type of toy made of a specific material. A similar table might be helpful for contact
limits for elements. However, as is noted in the standard, the limits for organic compounds
given in the limit tables have been calculated with the specific toy and toy material in mind.
In the case of other toys and toy materials not specified, they may not be appropriate and
should not be applied without further expert toxicological/exposure assessment. Similarly, it
will be difficult to create a table with contact limits for every element (or other substance) in
every single type of toy available on the market.

IV.6 Intended age groups

The US Code of Federal Regulations regulating the safety of toys (Child Protection Act and
toy Safety Act of 1969, amendment to the Federal Hazardous Substances Act (16 CFR Ch2)
describes test methods for articles intended for specified age groups of children:
• 18 months of age or less
• over 18 months but not over 36 months of age
• over 36 months but not over 96 months of age
page 222 of 234 RIVM report 320003001

The age of the intended user is determined by looking at the following factors: the
manufacturer’s stated intent (such as the age stated on a label) if it is reasonable; the
advertising, marketing and promotion of the article; and whether the article is commonly
recognized as being intended for children this age group.
To help determine the intended age group of a toy, the U.S. Consumer Product Safety
Commission (CPSC) developed extensive Age Determination Guidelines (U.S. Consumer
Product Safety Commission (CPSC), 2002). The primary content of the Age Determination
Guidelines is organized into four levels, each representing an increasing level of detail. These
levels are play categories, toy subcategories, age groups, and toy characteristics.
The following play categories (in approximate developmental order) and derived toy
subcategories were defined:
A. Early exploratory/practice play
a. Mirrors, mobiles, and manipulatives
b. Push and pull toys
B. Construction play
a. Blocks
b. Interlocking building materials
C. Pretend and role play
a. Dolls and stuffed toys
b. Play scenes and puppets
c. Dress-up materials
d. Small vehicle toys
e. Tools and props
D. Game and activity play
a. Puzzles
b. Card, floor, board, and table games
c. Computer and video games
E. Sports and recreational play
a. Ride-on toys
b. Recreational equipment
c. Sports equipment
F. Media play
a. Arts and crafts
b. Audiovisual equipment
c. Musical instruments
G. Educational and academic play
a. Books
b. Learning toys
c. Smart toys and educational software

The information presented in each subcategory is distributed among the following ten age
groups:
RIVM report 320003001 page 223 of 234

• Birth through 3 months

• 4 Through 7 Months
• 8 Through 11 Months
• 12 Through 18 Months
• 19 Through 23 Months
• 2 Years
• 3 Years
• 4 Through 5 Years
• 6 Through 8 Years
• 9 Through 12 Years
Each toy subcategory describes appropriate and appealing toy characteristics based on the
physical, cognitive, social, and emotional levels and abilities of children as they progress
through the ten age groups. These toy characteristics include: size, shape, number of parts,
interlocking versus loose parts, materials, motor skills required, color/contrast, cause and
effect, sensory elements, level of realism/detail, licensing, classic, robotic/smart features, and
educational.

Partly based the US Age Determination Guidelines, the CEN prepared a document that gives
guidelines for deciding which toys are intended for children under 36 months of age and
which toys are not intended for such children (European Committee for Standardization
(CEN), 2002). 24 categories of toys have been selected:
A. Activity toys
B. Aquatic toys
C. Art and craft materials and related articles
D. Audio/visual equipment
E. Books with play value
F. Construction toys and puzzles
G. Costumes, disguises and masks (intended to imitate)
H. Dolls and soft filled toys
I. Experimental sets
J. Functional toys
K. Game sets
L. Mechanical and/or electrical driven vehicles
M. Play scenes and constructed models
N. Projectile toys with a launching device
O. Push-along toys, pull-along toys and walking aids
P. Role playing toys
Q. Sand-water toys
R. Skill development toys
S. Toy cosmetics
T. Toy musical instruments
page 224 of 234 RIVM report 320003001

U. Toy sports equipment and balls

V. Toys for babies for looking at, grasping and/or squeezing
W. Toys intended to bear the mass of a child
X. Toys intended to be entered by a child
The toy’s suitability for children under or over 36 months is based on its functions and
characteristics such as the overall dimensions of the toy, the number and size of the parts or
components of the toy, the degree of detail and special functions a toy may have.

For risk assessment purposes, it is not necessary to base toy categories on ten different age
groups. The main purpose for categorizing toys according to intended or suitable age group is
to determine whether toys may pose a choking hazard for children under 36 months of age.
The value for looking at intended age for the purpose of setting limits for elements and other
substances in toys is further discussed in chapter 3.

IV.7 Exposure categories

The RIVM has created a fact sheet for children’s toys to be used in combination with the
computer program ConsExpo (Bremmer et al., 2002). The fact sheet defines 17 toy exposure
categories with representative examples of toys for which default ConsExpo models and
parameter values are given. These exposure categories are divided up in the five main
categories: ingestion, mouthing, inhalation, skin contact and eye contact.

Exposure category Examples

Mouthing
Toys meant for mouthing Teething ring
Other toys Cuddly toy, plastic doll
Ingestion
Direct ingestion Modeling clay, paint from toy car, ball
Hand-mouth contact, direct pen
Hand-mouth contact, indirect Finger paint, chalk
Face paint
Inhalation
Evaporation from liquids Felt pen
Evaporation from solid products Tent
Dust Chalk, cosmetics (blusher)
RIVM report 320003001 page 225 of 234

Exposure category Examples

Skin contact
Leaching from solid products Cowboy suit, tent ground sheet, cuddly
Rubbing off toy
Application on the skin Tent canvas, preserved wood
Intensive hand contact Cosmetics, face paint
Spillage Modeling clay, finger paint
Poster paint
Eye contact
Leaching from solid products Diving goggles
Application on the skin near eyes Cosmetics (eye shadow), face paint
Evaporation from solid products Diving goggles
Hand-eye contact Finger paint, chalk

The exposure categories above are useful to obtain an approximation of the exposure levels to
elements and other substances from toys by means of first or higher tier models in ConsExpo.
For the purpose of setting limits for elements in toys, some examples in the table above may
be too specific. A more pragmatic approach can be derived from this exposure based
categorization of toys, as discussed in chapter 3.
page 226 of 234 RIVM report 320003001
RIVM report 320003001 page 227 of 234

V Migration Tests

V.1 Migration tests

The EU countries and majority of the non-EU countries use the CEN 71-3 migration test for
the 8 different elements. However, some non-EU countries use different migration tests.

V.2 European Committee for Standardization (CEN)

The CEN 71-3 (safety of toys – part 3: migration of certain elements) standard is an European
standard prepared by the CEN/TC 52 – safety of toys committee for the European
Commission and the European Free Trade Association. This standard applies for Austria,
Belgium, Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece,
Hungary, Iceland, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, The Netherlands,
Norway, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, Switzerland, and
United Kingdom. Albania, Bulgaria, Croatia, The Former Yugoslav Republic of Macedonia,
and Turkey are affiliates of the CEN and participate in the General Assembly and technical
bodies, but are not full members. Currently, they are implementing the CEN standards into
their own national legislation. CEN 71-3 specifies requirements and migration tests for the
elements: antimony (Sb), arsenic (As), barium (Ba), cadmium (Cd), chromium (Cr), lead
(Pb), mercury (Hg), and selenium (Se) from toys and packaging material when they are part
of the toy or have intended play value (European Committee for Standardization (CEN),
1994). There are different test methods for different toy materials:
• coatings of paints, varnishes, lacquers, printing inks, polymers and similar coatings.
If 100 mg of coating can be removed from the toy then this fraction is sieved over
0.5 mm. Only if the coating can not be comminuted (e.g. plastic or elastic paint), a test
portion is removed and used to measure the migration. For the migration test, a portion of
the coating is mixed with 50 times 0.07 ± 0.005 M of HCl. After 1 min the pH is
measured and if necessary set at 1.0-1.5 using 2 M HCl. Under the exclusion of light the
mixture is agitated for 1 h at 37 ± 2 °C and then stand for another hour. The solution and
coating are separated using a membrane filter (pore size of 0.45 μm) and if necessary
centrifugation at 5000 g for as maximum of 10 min. Hydrochloric has to be added if the
samples are not analysed within 24 h up to a concentration of 1 M. If only 10 to 100 mg
of coating can be obtained then 5 ml of 0.07 ± 0.005 M HCl is added and the same
migration test procedure is followed as described above.
• paints (including finger paint, varnishes, lacquers, glazing powders and similar materials
in solid or in liquid form appearing as such in the toy).
A test portion of 10-100 mg should be obtained and a dimension < 6 mm if the material is
solid. If it contains grease oil, wax or similar material, the test portion should be enclosed
page 228 of 234 RIVM report 320003001

in hardened filter-paper and the ingredients should be removed with 1,1,1-trichloroethane

or other suitable solvent using solvent extraction.
¾ samples not containing grease, oil wax or similar material.
The test portion is incubated with 50 times its mass 0.07 ± 0.005M HCl. If only 10 to
100 mg of coating can be obtained then 5 ml of 0.07 ± 0.005 M HCl. After 1 min
shaking, the pH is measured and if necessary set at 1.0-1.5 using 2 or 6 M HCl
depending on the alkalinity of the sample. Under the exclusion of light the mixture is
agitated for 1 h at 37 ± 2 °C and then stand for another hour. The solution and coating
are separated and stored as described previously for coatings of paints, varnishes,
lacquers, printing inks, polymers and similar coatings.
¾ samples containing grease, oil, wax, or similar material.
An amount of water 25 times the mass of the original material is added to the
hardened filter-paper and macerated. at 37 ± 2 °C until the mixture is homogeneous.
Next, an amount of 25 times the mass of the test portion of 0.14 ± 0.01 M HCl is
added. After 1 min, the pH is measured and if necessary set at 1.0-1.5 using 2 or
6 M HCl depending on the alkalinity of the sample. Under the exclusion of light the
mixture is agitated for 1 h at 37 ± 2 °C and then stand for another hour. The solution
and coating are separated and stored as described previously for coatings of paints,
varnishes, lacquers, printing inks, polymers and similar coatings.
• materials to leave a trace (e.g. the graphite material in pencils and liquid ink in pens).
A test portion of 10-100 mg should be obtained and a dimension < 6 mm if the material is
solid. If it contains grease oil, wax or similar material, the test portion should be enclosed
in hardened filter-paper and the ingredients should be removed with 1,1,1-trichloroethane
or other suitable solvent using solvent extraction.
¾ samples not containing grease, oil wax or similar material.
The same procedure to determine the migration is followed as described for paints not
containing grease, oil wax or similar material.
¾ samples containing grease, oil, wax, or similar material.
The same procedure to determine the migration is followed as described for paints
containing grease, oil wax or similar material. Except if the original amount is
between 10 and 100 mg then 2.5 ml water and 0.14 M HCl are used.
• paper and paper board with a maximum mass per unit area of 400 g/m2.
A test portion of 10 to 100 mg should be obtained and macerated in 25 times its mass of
water at 37 ± 2 °C until the mixture is homogenous. Next, an amount of 25 times the mass
of the test portion of 0.14 ± 0.01 M HCl is added. After 1 min the pH is measured and if
necessary set at 1.0-1.5 using 2 M HCl. Under the exclusion of light the mixture is
agitated for 1 h at 37 ± 2 °C and then stand for another hour. The solution and coating are
separated using a membrane filter (pore size of 0.45 μm) and if necessary centrifugation
at 5000 g for as maximum of 10 min. Hydrochloric has to be added if the samples are not
analysed within 24 h up to a concentration of 1 M.
• glass, ceramic and metallic materials.
RIVM report 320003001 page 229 of 234

Small parts (fitting in the small parts cylinder described in standard CEN 71-1) will be
tested entirely and if the material is larger a part is removed as described for coatings of
paints, varnishes, lacquers, printing inks, polymers and similar coatings. Next, the toy or
component is placed in a 50 ml glass container (height 60 mm and diameter of 40 mm)
and 0.07 ± 0.05 M HCl is added to just cover the toy. The container is covered and under
the exclusion of light is left to stand for 2 h at 37 ± 2 °C. The solution and coating are
separated using a membrane filter (pore size of 0.45 μm) and if necessary centrifugation
at 5000 g for as maximum of 10 min. Hydrochloric has to be added if the samples are not
analysed within 24 h up to a concentration of 1 M.
• natural and synthetic textiles.
A test portion of 100 mg should be obtained and cut out from the area representing the
whole material and a dimension < 6 mm. The same procedure to determine the migration
is followed as described for coatings of paints, varnishes, lacquers, printing inks,
polymers and similar coatings.
• polymeric and similar materials, including laminates, whether textile reinforced or not,
but excluding other textiles.
A test portion of 10 to 100 mg should be obtained and cut out from the area having the
thinnest material cross section and a dimension < 6 mm. The same procedure to
determine the migration is followed as described for coatings of paints, varnishes,
lacquers, printing inks, polymers and similar coatings.
• pliable modelling materials (including modelling clay and gel).
A test portion of at least 100 mg should be obtained. If it contains grease oil, wax or
similar material, the test portion should be enclosed in hardened filter-paper and the
ingredients should be removed with 1,1,1-trichloroethane or other suitable solvent using
solvent extraction.
¾ samples not containing grease, oil wax or similar material.
The same procedure to determine the migration is followed as described for paints not
containing grease, oil wax or similar material.
¾ samples containing grease, oil, wax, or similar material.
The same procedure to determine the migration is followed as described for paints
containing grease, oil wax or similar material
• other materials whether mass coloured or not (e.g. wood, fibre board, hard board, bone,
leather and paper and paper board > 400 g/m2).
A test portion of 10 to 100 mg should be obtained and should be tested using the most
appropriate method described for coatings of paints, varnishes, lacquers, printing inks,
polymers and similar coatings, paper and paper board with a maximum mass per unit area
of 400 g/m2, natural and synthetic textiles, and glass, ceramic and metallic materials.
Not part of this legislation are toys and parts of toys which obviously exclude any hazard due
to sucking, licking or swallowing due to their accessibility, function, mass, size or other
characteristics, but bearing in mind the normal and foreseeable behaviour of children under
the age of 7. The principle of the migration test is that the soluble element extracted
represents the amount released in the stomach.
page 230 of 234 RIVM report 320003001

V.3 Health Canada

There are different test methods for different toy materials in Canada:
• Leachable cadmium, barium, antimony, selenium and arsenic in applied coatings
This method is used by Health Canada to determine the migration from decorative and
protective coatings. The coating is removed by using a scalpel or tetrahydrofuran or
another suitable solvent without removing the underlying substrate material. If a solvent
is used then evaporate the solvent in an air convection oven at 60 °C for 1 h. Next, grind
the sample in a mortar and sieve to get a fraction 250 and 500 μm. Dry the sample again
in an oven at 60 °C for 1 h. For the migration determination, an amount of 100 mg sample
is incubated with 20 ml 5% HCl solution for 10 ± 1 min at 20 ± 2 °C under constant
stirring. The solution is filtered using Whatman no. 40 filter paper and the filter is washed
with deionised water. Add 1 ml of concentrated nitric acid to the filtrate and add
deionised water until a volume of 50 ml. This solution is used to determine the migration.
• Leachable lead in metallic consumer products which pose a hazard from ingestion
This method is used by Health Canada to determine the release of lead from metallic
consumer product like jewellery and figurines which fit into the truncated right circular
cylinder (small parts cylinder), because the pose an ingestion hazard. Acetone is used to
prewash the sample. Next, the sample is covered with 0.07 M HCl and incubated for 2 h
under the exclusion of light at 37 ± 2 °C. The solution is filtered using Whatman no. 40
filter paper. Add 2 ml of concentrated HCl to the filtrate and add 0.07 M HCl until a
volume of 25 ml. This solution is used to determine the migration.
• Leachable lead and cadmium from glazed ceramics and glassware
Health Canada uses this method to determine the release of lead and cadmium from toys
containing glazed ceramics or glass, e.g. children’s tea set. Cover the sample with 4%
acetic acid and incubate for 24 h ± 10 min at 22 ± 2 °C. An aliquot is used to analyse for
lead and cadmium.

V.4 Fowles et al. method

Fowles describes a method with several variables used to determine the leaching of cadmium
from plastic toys (Fowles et al., 1977). Samples were taken using a rough file, a Stanley
shaper, a Surform hand tool, and human teeth. All samples were sieved mechanically trough
different sieves to obtain samples in different categories, namely < 0.106 mm, 0.106-0.5 mm,
0.5-1.0 mm and > 1.0 mm. For the extraction procedure, 1 g of sample was placed in
25 ml HCl (ranging from 0.046-0.47 M) solution and shaken (3 different speed settings) for 1
to 24 h in the absence and presence of light and air. Also the temperature was varied ranging
from 19 to 42.5 °C. Fowles showed that the different methods to obtain a sample gave similar
RIVM report 320003001 page 231 of 234

fraction and that the leaching depended mainly on the sample size, acid strength and
temperature (increased leaching with decreased particle size, increasing acid strength and
temperature). Light had a dramatic effect on the leaching of cadmium from plastic toys,
highly increased in the presence of light. Shaking speed had no effect on the leaching and air
only a minor effect. The most physiological conditions are according to the author
0.1 M HCl, presence of air, absence of light, incubation of 4h and varying particle sizes (not
the standard factory regrind that the industry tests).

V.5 Physiologically-based extraction tests

There are several other physiologically-based extraction test to simulate sucking than
mentioned here. However, all these tests were developed and used to determine the release of
phthalates from toys and other consumer products. Examples are the Joint Research Centre
(JRC) model and a model developed by the U.S. Consumer Product Safety Commission
(Simoneau et al. 2001; U.S. Consumer Product Safety Commission, ).

V.6 RIVM method

The composition and preparation of non-stimulated (under fasted conditions and no stimuli
for secretion like sucking) and stimulated (under fed conditions and stimulus like sucking)
digestive fluids are described in detail in the publications by Oomen et al. (Oomen et al.,
2003a) and Versantvoort et al. (Oomen et al., 2003c; Versantvoort et al., 2005). Stimulated
and non-stimulated digestive juices differ in pH, salt concentrations, and enzyme
concentrations. The mixtures are rotated head-over-heels at 55 rpm and the whole process is
performed at 37°C. At the end of the digestion process the tubes are centrifuged for 5 min at
2750 g, yielding the chyme (the supernatant) and the digested matrix (the pellet), and
sampled to obtain information on the bioaccessibility of the contaminant. Samples can be
taken from the saliva, stomach and chyme phase to obtain information on the bioaccessibility
of the contaminant and its behaviour in the different compartments of the digestive tract.
The main differences between the suck, suck-swallow, and swallow in vitro digestion model
are the stomach pH and the composition of the digestive juices. The matrix may affect the pH
in the stomach. However, under fasting conditions, the pH in the stomach is usually low and
set at 2.5 ± 0.1 in the suck and suck-swallow model and to 1.6 ± 0.1 in the swallow model
under fasted conditions (see Figure V-1). The pH in the gastric compartment of the swallow
model under fasted conditions is lower, because less saliva is entering the gastric
compartment in the swallow model (6 ml saliva instead of 18 ml). These pHs were chosen
because an in vitro digestion without matrix results in this pH and because this pH falls in the
range of pH values for fasting conditions (Charman et al., 1997).
page 232 of 234 RIVM report 320003001

V.6.1 Suck model

The model is applied to simulate sucking by a child on a consumer product (Oomen et al.,
2003c). The suck time depends on the age of the child (at the age of 0.5-2 years children have
the longest suck time), but also on the product (Bremmer and Van Veen, 2002). The
migration of a contaminant from its matrix into saliva simulant after a certain time based on
mouthing duration can be measured with this model. It can either be assumed that all the
contaminant that is released in the mouth is also available for absorption, in which case the
model is terminated after the mouth phase, or that the contaminant can form aggregates that
can not be absorbed in the stomach or intestinal compartment. With the suck model, the
contaminant that is released in the mouth during sucking (one compartment model) or the
fraction that is available for absorption in the small intestine (three compartment model) can
be investigated. Different amounts of matrix are introduced to 21 ml stimulated saliva and
rotated for a variable time periods. The time period applied can be either a default value
(30 min) or a period considered to be specifically appropriate for a certain product which is
based on the product and the input of the risk assessor. The digestion tubes are centrifuged to
remove the matrix and 18 ml of supernatant used for further incubation (the other 3 ml are
used for analysis of the bioaccessibility in saliva). For the one compartment model, the
sucking model is terminated after the saliva incubation. For the three compartment model, a
volume of 12 ml gastric juice (pH 1.07 ± 0.07) is added to the saliva supernatant. The mixture
is rotated for 1 h and the pH of the mixture is determined and, if necessary, set to 2.5 ± 0.1.
Then, the mixture is rotated for another hour. Finally, 12 ml of duodenal juice (pH 7.8 ± 0.2)
and 6 ml bile (pH 8.0 ± 0.2) are added simultaneously, and the pH of the chyme is
determined and if necessary set to 6.5 ± 0.5. Then, the mixture is rotated for another 2 h. The
digestion tubes are centrifuged and the supernatant is suitable for analysis.

V.6.2 Suck and swallow model

This method is applied to simulate mouthing and then ingestion of a certain consumer
product (Oomen et al., 2003c). Thus, contrary to the three compartment suck model the
matrix is ingested after sucking. The only modification is that the digestion starts by
introducing 18 ml stimulated saliva to different amounts of matrix. This mixture is rotated
head-over-heels for 30 min and then gastric juice is directly added without centrifuging. The
rest of the procedure is the same as described for the three compartment suck model.

V.6.3 Swallow model under fasted conditions

This model is applied to simulate ingestion of a certain consumer product under fasted
conditions (Oomen et al., 2003c). It starts by introducing 6 ml saliva (pH 6.5 ± 0.2) to
different amounts of matrix. This mixture is rotated for 5 min. Subsequently, 12 ml of gastric
juice is added and the pH of the mixture of saliva and gastric juice is determined and, if
necessary, directly set to 1.6 ± 0.1. The mixture is rotated for 2 h. Finally, 12 ml of duodenal
RIVM report 320003001 page 233 of 234

juice and 6 ml bile are added simultaneously and the pH is determined and if necessary set to
6.0 ± 0.5. The mixture is rotated for another 2 h. The digestion tubes are centrifuged and the
supernatant is used for analysis.

V.6.4 Swallow model under fed conditions

The digestion starts by introducing different amounts of matrix to 6 ml stimulated saliva and
4.5 g infant food (product number 282, Olvarit (Nutricia®, the Netherlands), supplemented
with 2 ml sunflower oil per 100 g). This infant food with sunflower oil represents the mean
food intake for adults in the Netherlands for a cooked meal regarding macronutrients and
caloric composition. It is based on the third Dutch National Food Consumption Survey from
1998 (Versantvoort et al. 2005). Immediately, 12 ml of stimulated gastric juice (pH 1.30 ±
0.02) is added and pH of the mixture is set to 2.5 ± 0.5. After 2 h of rotating, 12 ml of
stimulated duodenal juice (pH 8.1 ± 0.2), 6 ml stimulated bile (pH 8.2 ± 0.2), and
2 ml sodium bicarbonate (84.7 g/l) are added simultaneously. The pH is set to 6.5 ± 0.5 and
the mixture is rotated for another 2 h. Separation of chyme and pellet was obtained by
centrifugation and the supernatant can be analysed to determine the bioaccessibility of the
contaminant.
page 234 of 234 RIVM report 320003001

suck model suck-swallow model swallow model - fasted swallow model - fed

different amount different amount different amount different amount

of contaminated of contaminated of contaminated of contaminated
matrix matrix matrix matrix + 4.5 g
infant formula

21 ml stimulated 18 ml stimulated
saliva saliva 6 ml saliva 6 ml stimulated
saliva
variable min
30 min 5 min
centrifugation direct
18 ml supernatant 12 ml stimulated
12 ml gastric juice 12 ml gastric juice
+ 12 ml gastric gastric juice
juice
2h pH 1.6 2h
2h after 1 h pH 2.5 2h
after 1 h pH 2.5
12 ml stimulated
12 ml duodenal 12 ml duodenal 12 ml duodenal duodenal juice + 6
juice + 6 ml bile juice + 6 ml bile juice + 6 ml bile ml stimulated bile
+ 2 ml NaHCO3
2h 2h 2h
centrifugation centrifugation centrifugation 2h
centrifugation

chyme chyme chyme chyme

analysis analysis analysis analysis

bioaccessibility of bioaccessibility of bioaccessibility of bioaccessibility of

contaminant contaminant contaminant contaminant

Figure V-1. Schematic representation of the RIVM suck, suck-swallow, and swallow under fasted and fed
conditions in vitro digestion models.

V.6.5 Iliano et al. method

Iliano et al. describes a physiologically-based extraction method for antimony (Sb), arsenic
(As), barium (Ba), cadmium (Cd), chromium (Cr), lead (Pb), mercury (Hg), and selenium
(Se) from toys using saliva simulant (Iliano et al., 1988). The simulated saliva consist of
4.2 mg/ml NaHCO3, 500 μg/ml NaCl and 200 µg/ml K2CO3 with a pH of 8.8. The toy or part
of the toy is immersed in the simulated saliva for 2 h at 37 °C. After filtration the filtrate is
analysed to determine the migration.
ERRATUM d.d. 26-01-2015
RIVM report 320003001
Chemicals in toys. A general methodology for assessment of chemical safety of toys with a
focus on elements (2008)
Openbaar sinds: 02-04-2009
Auteur: van Engelen JGM, van der Zee Park M, Janssen PJCM, Oomen AG, Brandon EFA,
Bouma K, Sips AJAM, van Raaij MTM
RIVM Rapport 320003001

The work described in this report report was carried out in 2006 on request of DG Enterprise
in view of contract nr. SI2.ICNPROCE003918500. The Commission used content of this
report as a starting point for the derivation of migration limits as included in Annex III of the
Toys Safety Directive (2009/48/EC).
The following inconsistency in RIVM report 320003001 is identified:

In Chapter 3 of the RIVM Toys report, a choice is made regarding the ingested amount of
toys material and the frequency of exposure. The ingested amount of toys material is chosen
as 100 mg for dry, pliable or powder-like toy materials and 400 mg for liquid or sticky
material. It is noted on page 41 that "The ingestion of 100 mg by children is considered
reasonable, but may not occur daily. For exposure assessment refinement purposes, we
propose to use a frequency of 1/week for this ingestion default when the exposure is
compared to a chronic health-based limit value. This is a rough estimate and needs further
research. "
And on page 42: " Similar to the ingestion default for dry, brittle, powder-like and pliable
materials, an ingestion of 400 mg may occasionally occur, but not daily. For the purpose of
an exposure assessment refinement, when comparing exposure to a chronic health-based limit
value, we propose to use a frequency of 1/week as a default. This is a rough estimate and
needs further research.”
The proposed frequency of exposure is once a week.

In Chapter 8, the migration limits are derived and presented in Tables 8-3 and 8-4. In the
calculation an ingested amount of 100 mg per day (dry, pliable or powder-like toy materials)
and 400 mg (liquid or sticky material) per day were used. For explanation, the reader is
referred to Chapter 3. The migration limits are derived base on a frequency of exposure of
once a day. This should have been 100 mg per week and 400 mg per week.
The result is that the limit values presented in Tables 8-3 and 8-4 are incorrect.
In the Erratum, the inconsistency between Chapter 3 and Chapter 8 is corrected. In tables 8-3
and 8-4, the columns of the migration values, using 5, 10 or 20% of the TDI in the calculation
have been corrected using the correct assumption of an ingestion frequency of once a week,
in the Tables 8-3 and 8-4.

Adapted information:
8.6 Migration limits for elements in toys

In the tables below, migration limits for elements in toys are presented, as derived by the
methodology proposed in chapter 7. Further explanation and a calculation example can be
found in paragraph 8.5.1.
Table 8-2 For intake of 8 mg (scraped off material) for children < 3 years of age
* Age < 3 years
* Body Weight 7.5 kg
* Material 8 mg (scraped off) per day
Not changed

Table 8-3 For intake of 100 mg (dry, brittle, powder-like or pliable material) for children < 3
years of age
* Age < 3 yrs
* Body Weight 7.5 kg
* Material 100 mg (dry, powder like or pliable) once a week

Element TDI (µg/kg Migration Limit value (mg/kg product) Current

bw/day) Migration Limit
5% TDI 10% TDI 20% TDI (mg/kg
product)*
Aluminum 750 19687,5 39375,0 78750,0
Antimony 6 157,5 315,0 630,0 60
Arsenic 1 26,3 52,5 105,0 25
Barium 600 15750,0 31500,0 63000,0 250
Boron 160 4200,0 8400,0 16800,0
Cadmium 0.5 13,1 26,3 52,5 50
Chromiuma,d Cr3+
ws 5 131,3 262,5 525,0 25
Cr3+
wis 5000 131250,0 262500,0 525000,0
(Cr
6+)b 5 131,3 262,5 525,0 25
(Cr
6+)c 0.0053 0,1 0,3 0,6 25
Cobalt 1.4 36,8 73,5 147,0
Copper 83 2178,8 4357,5 8715,0
Lead 3.6 94,5 189,0 378,0 90
Manganese 160 4200,0 8400,0 16800,0
Mercury 2 52,5 105,0 210,0 25
Nickel 10 262,5 525,0 1050,0
Selenium 5 131,3 262,5 525,0 500
Silver 5 131,3 262,5 525,0
Strontium 600 15750,0 31500,0 63000,0
Tin Inorganic 2000 52500,0 105000,0 210000,0
Organic 0.25 6,6 13,1 26,3
Zinc 500 13125,0 26250,0 52500,0
* Migration limits according EN 71-3 for modelling clay and finger paint
a
ws = water soluble, wis = water insoluble
b
Based on a TDI of 5 µg/kg bw derived for non-carcinogenic effects by hexavalent chromium
c
Based on a Virtually Safe Dose (VSD) of 0.0053 µg/kg bw/day derived for the genotoxic and carcinogenic
action by hexavalent chromium. As explained in the appended toxicological profile on chromium, this VSD is
based on a limited bioassay in mice and is fraught with additional uncertainty compared to the usual bioassay-
derived VSDs. Results of NTP studies now in progress should allow more a reliable oral cancer risk estimation
in the near future.
d
Measurement of Cr6+ is difficult. Further research is needed to derive safe migration limit values for this
element
Table 8-4 For intake of 400 mg (liquid or sticky material) for children < 3 years of age
* Age < 3 yrs
* Body Weight 7.5 kg
* Material 400 mg (liquid & sticky, once per week)

Element TDI (µg/kg Migration Limit value (mg/kg product) Current

bw/day) Migration Limit
5% TDI 10% TDI 20% TDI (mg/kg
product)*
Aluminum 750 4921,9 9843,8 19687,5
Antimony 6 39,4 78,8 157,5 60
Arsenic 1 6,6 13,1 26,3 25
Barium 600 3937,5 7875,0 15750,0 250
Boron 160 1050,0 2100,0 4200,0
Cadmium 0.5 3,3 6,6 13,1 50
Chromiuma,d Cr3+
ws 5 32,8 65,6 131,3 25
Cr3+
wis 5000 32812,5 65625,0 131250,0
(Cr
6+)b 5 32,8 65,6 131,3 25
(Cr
6+)c 0.0053 0,0 0,1 0,1 25
Cobalt 1.4 9,2 18,4 36,8
Copper 83 544,7 1089,4 2178,8
Lead 3.6 23,6 47,3 94,5 90
Manganese 160 1050,0 2100,0 4200,0
Mercury 2 13,1 26,3 52,5 25
Nickel 10 65,6 131,3 262,5
Selenium 5 32,8 65,6 131,3 500
Silver 5 32,8 65,6 131,3
Strontium 600 3937,5 7875,0 15750,0
Tin Inorganic 2000 13125,0 26250,0 52500,0
Organic 0.25 1,6 3,3 6,6
Zinc 500 3281,3 6562,5 13125,0
* Migration limits according EN 71-3 for modelling clay and finger paint
a
ws = water soluble, wis = water insoluble
b
Based on a TDI of 5 µg/kg bw derived for non-carcinogenic effects by hexavalent chromium
c
Based on a Virtually Safe Dose (VSD) of 0.0053 µg/kg bw/day derived for the genotoxic and carcinogenic
action by hexavalent chromium. As explained in the appended toxicological profile on chromium, this VSD is
based on a limited bioassay in mice and is fraught with additional uncertainty compared to the usual bioassay-
derived VSDs. Results of NTP studies now in progress should allow more a reliable oral cancer risk estimation
in the near future.
d
Measurement of Cr6+ is difficult. Further research is needed to derive safe migration limit values for this
element
Table 8-5 For intake of 8 mg (scraped off material) for toys intended to be mouthed by
children > 3 years of age
* Age > 3 yrs
* Body Weight 15 kg
* Material 8 mg (scraped off, per day)
Not changed