Opinion on: Report (Final Draft) on "Assessment of the risks to human health posed by azo colorants in toys, writing inks and paper products, and analysis of the advantages and drawbacks of restrictions on their marketing and use". Opinion expressed at the 24th CSTEE plenary meeting, Brussels, 12 June 2001
Questions addressed to the Committee:
The CSTEE has been asked to consider the above Report and express it opinion on whether
a) the reported degree of risk is sufficiently justified, and
b) the Report is of a good quality.
Opinion:
a) The reported degree of risk is not sufficiently justified for the following reasons:
i) The available data do not allow adequate assessment of human exposure resulting from the use of azo dye-containing consumer items. The indirect approach which, in the absence of other information, has been employed in the Report, does not lead to reliable exposure estimates.
ii) The question of the toxicity of impurities present in azo dyes is not addressed.
iii) An acceptable acute dose for humans is erroneously derived using animal acute lethal dose (LD50) data.
iv) The Report's conclusion that no risk of sensitisation exists is not based on adequate scientific evidence.
v) In view of the clear genotoxicity of azo dyes themselves as well as of the carcinogenic aromatic amines derived from them, the use, made in the Report, of a threshold approach for the assessment of cancer risks from azo dye-containing products is not scientifically justified.
vi) No effort to assess health risks for the general population arising from environmental exposure as a result of the washing and disposal of azo dye-containing products, or of health risks as a result of occupational exposure, was made.
b) The Report's authors appear to have gone into a lot of trouble to obtain information on the range of products in which azo dyes are used. They also made substantial efforts to derive estimates of human exposure to azo dyes from the use of these products, although, as stated above, the reliability of the resulting estimates is considered by CSTEE to be low. This weakness, and, especially the serious errors and omissions in the assessment of toxic risks outlined above, mean that the overall quality of the Report is below the desired level.
Justification of the Opinion:
The Report deals with azo dyes capable of giving rise to human exposure to any of 21 aromatic amines classified as known or suspected carcinogens (see Table), and focuses on the presence of such dyes in specific products (toys, writing inks and paper products). It focuses on human health risks and does not deal with risks for the environment. The CSTEE has already expressed an Opinion on two related studies produced on behalf of the European Commission, concerning the risks from the presence of the same types of dyes in textiles and leather goods (1-4).
It is well known that many industrially used azo dyes contain impurities (sometimes in substantial amounts) which should be considered in assessing toxic risks associated with exposure to azo dyes present in consumer products. This point is not addressed in the Report.
Azo dye toxicity
The acute toxicity of azo dyes is low, with the values of LD50 being greater than 100 mg/kg. This being so, the Report utilises the lower end of the range animal LD50 values, plus a margin of safety, to derive and acceptable human acute dose of 300 μg/kg. This constitutes a serious error, as the calculation of the acceptable acute dose cannot be based on the LD50 but on the NOAEL.
Data on repeated dose toxicity in experimental animals exist only for 2 azo dyes, Acid Red 114 and Direct Blue 15, which indicate low toxicity for these dyes (NOAEL > 600 mg/kg.day). Little information on other azo dyes as well as other types of azo dye toxicity exists, except for skin sensitisation and carcinogenicity.
Skin sensitisation
There are many reports on the induction of skin allergy by azo dyes in humans exposed occupationally or through contact with clothing. The induction of skin allergy involves an initial phase of "sensitisation", followed by "elicitation" which can be brought about by brief contact, at any time during later life, with low levels of the same or a closely related substance. The authors of the Report correctly indicate that, during the period in their life when children use azo dye-containing products (which may last up to five years), dermal contact with certain products identified as presenting the highest potential of exposure (e.g. finger paints, fake tatoos, air-blown pen systems) may take place several times per week. Noting that each such contact is likely to last from a few minutes to several hours, the CSTEE considers that the resulting exposure may be sufficient to initiate the allergy process by causing skin sensitisation. It therefore disagrees with the Report's conclusion that "
it is not considered that exposure of any of the products of this study, under reasonable conditions of use ... would be of sufficient duration to produce an allergic reaction of this type" (paragraph 6.2). However, no data exist for deriving a NOAEL for the induction of skin allergy or for the quantitative estimation of the corresponding risk for humans.
Carcinogenicity
A number of azo dyes have been shown to be animal carcinogens, and analysis based on structure-activity relationships suggests that many more azo dyes are likely to have a carcinogenic potential. Reductive cleavage to carcinogenic aromatic amines, which are known to act via genotoxic mechanisms, is in all likelihood an important (although not necessarily the only) mechanism of azo dye carcinogenicity. Such amines may also be present in commercial azo dyes as impurities. Many azo dyes show clear genotoxic activity, indicating that azo dye carcinogenicity is likely to result from a genotoxic mechanism which is considered not to be compatible with the existence of a threshold. Therefore the use of an approach based on a LOAEL and a margin of safety for the assessment of carcinogenic risks, as done in the Report, constitutes a serious error and is not acceptable to the CSTEE. It is noted, in this context, that the reference made in the Report (p. 12) to a NOAEL of 600 mg/kg.day for Acid Red 114 used in a previous study (1) is misleading, as, in that study, this NOAEL was used in the context of non-cancer effects.
The existing data do not permit the calculation of a unit cancer risk for any azo dyes. For benzidine, one of the carcinogenic amines of concern, a unit risk of 0.23 (µg/kg.day)
-1 has been calculated by the US EPA. The CSTEE recognises that unit risks are specific for each chemical, and that the assumption that unit risks estimated for one chemical can be applied to chemicals of similar structure is debatable. It is also aware of the fact that some azo dye-related aromatic amines are less potent carcinogens than benzidine. Nevertheless, in the context of a worst-case scenario, it would have been justified to have made use of the unit risk for benzidine in an attempt to assess quantitatively cancer risks for azo dyes.
Exposure assessment
The products identified as possible sources of exposure to azo dyes include mainly toys made of plastic, paper, rubber, wood or metal, infant teeters, fake tattoos, paints (including finger paints), modelling clays, writing inks and coloured paper. There is a serious lack of quantitative data which would permit a reasonable assessment of human exposure to azo dyes through the use of these products. In order to obtain an indication of those products likely to result in the highest exposures, a screening is carried out in the Report, based on the qualitative characterisation of the frequency (often/frequent/sometimes/never) and duration (short/medium/long) of likely contact. Based on this qualitative exposure assessment, three items (finger paints, air-blown pen systems and fake tatoos) are selected for quantitative exposure assessment using worst case scenarios. The likely routes of exposure were also considered. While recognising that this approach was adopted in the absence of measured data, the CSTEE considers that its reliability for identifying the items leading to highest exposure and estimating exposure levels is limited.
Risk assessment
In the Report, the exposure levels calculated as described above are compared to LOAEL values for acute and for chronic exposures. As already mentioned, the use of a threshold approach for the assessment of the cancer risks of azo dyes or their associated aromatic amines is unacceptable, and an approach based on linear extrapolation and unit risks should be used. Employing a worst case scenario based on the unit risk for benzidine (0.23 (µg/kg.day)
-1), the exposures to aromatic amines estimated in the Report for all three of the items mentioned in the previous paragraph would predict extremely high cancer risks (approaching 100%). Notwithstanding the limited validity of this prediction as a result of the uncertainties associated with the estimation of the exposures and differences in the unit risks for different amines, this calculation suggests that the conclusion that only finger-paints, but not air-blown pen systems and fake tattoos, should be considered as presenting a significant health risk is not scientifically justified.
Other comments:
a) Although the Report contains a brief reference to occupational exposure, it makes no attempt to carry out any corresponding exposure or risk assessment.
b) No attempt is made to assess possible health risks of the general population as a result environmental exposure resulting from the washing or disposal of azo dye-containing products.
References
1. LGC, The risk of cancer caused by textiles and leather goods coloured with azo dyes. Final Report, April 1998.
2. Opinion on Risk of cancer caused by textiles and leather goods coloured with azo-dyes (Assessment of the Report by LGC "The risk of cancer caused by textiles and leather goods coloured with azo dyes"), expressed at the 7th CSTEE plenary meeting, Brussels, 18 January 1999.
3. WS Atkins International Ltd, Assessment of the risks to human health posed by certain chemicals in textiles and of the advantages and drawbacks of restrictions on their marketing and use. Final Report, July 1999.
4. Opinion on "Assessment of the risks to human health posed by certain chemicals in textiles", WS Atkins, Final report - Opinion adopted at the 17
th CSTEE plenary meeting, Brussels, 5 September, 2000.
Table: List of aromatic amines, known or suspected of being human carcinogens, which may be released from azo colorants
name of amine | CAS No. | IARC classification | MAK classification |
benzidine | 92-87-5 | group 1 | A1 |
biphenyl-4-ylamine | 92-67-1 | group 1 | A1 |
2-naphthylanime | 91-59-8 | group 1 | A1 |
4-chloro-o-toluidine | 95-69-2 | group 2A | A1 |
2,2'-dichloro-4,4'-methylenedianiline | 101-14-4 | group2A | A2 |
4-chloroaniline | 106-47-8 | group 2B | A2 |
3,3'-dichlorobenzidine | 91-94-1 | group 2B | A2 |
3,3'-dimethoxybenzidine | 119-90-4 | group 2B | A2 |
3,3'-dimethylbenzidine | 119-93-7 | group 2B | A2 |
4,4'-methylenedianiline | 101-77-9 | group 2B | A2 |
4,4'-methylenedi-o-toluidine | 838-88-0 | group 2B | A2 |
4-methyl-m-phenylenediamine | 95-80-7 | group 2B | A2 |
2-methoxyaniline | 90-04-0 | group 2B | - |
4-methoxy-m-phenylenediainine | 615-05-4 | group 2B | A2 |
6-methoxy-m-toluidine | 120-71-8 | group 2B | A2 |
4,4'-oxydianiline | 101-80-4 | group 2B | A2 |
4,4'-thiodianiline | 139-65-1 | group 2B | A2 |
4-o-tolyazo-o-toluidine | 97-56-3 | group 2B | A2 |
o-toluidine | 95-53-4 | group 2B | A2 |
5-nitro-o-tolouidine | 99-55-8 | group 3 | A2 |
2,4,5-trimethylaniline | 137-17-7 | group 3 | A2 |