Opinion of the Scientific Committee on Toxicity, Ecotoxicity and the Environment (CSTEE) on validation of test methods for phthalate migration - Opinion expressed at the 17th CSTEE plenary meeting, Brussels, 5 September 2000.
Background and Terms of Reference
In the light of the opinion of the CSTEE of 28th September 1999 on "TNO, LGC and US CPSC reports on Phthalate migration test validation" and the Addendum of 25th November 1999 to this opinion, a programme for a full validation of test methods for phthalate migration has been developed and that is presented to CSTEE. The committee on the basis of the Programme (CSTEE/2000/13 Add 1) for the validation of methodologies to test migration of plasticisers from toys and childcare articles is to answer the following questions:
- Is the approach of the Programme suitable to provide for a sound validation of methods to measure satisfactorily the release of phthalates from toys and child care articles as to distinguish safe from unsafe products?
- Which comments can the Committee offer to further strengthen or complement the programme?
Introduction
It is proposed in the work programme that 3 of the in vitro test methods, available at present, will be validated:
- Horizontal shaking of toy material with glass balls in an artificial saliva at 37° C (LGC mild horizontal shaking method),
- Horizontal shaking of toy material with relatively heavy metal balls in an artificial saliva at 60° C (LGC stringent horizontal shaking method)
- Dynamic extraction of plasticiser from toy material in an artificial saliva employing a rotation method at 22° C (TNO rotation method)
The reasons given for validating the 3 methods are: "The stringent conditions may allow to simulate the 'worse case conditions' (referring to DINP migration rate 9microgram/10 cm2/min, adopted by CSTEE)1, whereas TNO method and the mild version of LGC method may be easily applied to a wide range of testing and other substances of various target values".
The validation protocol utilises to a great extent the guidelines, which are described in the CSTEE Opinion of 28th September 1999 and in the Addendum of 25th November 1999 to this Opinion. A major issue identified, however, is the problem of achieving the target rate for the validation of test method for DINP (diisononyl phthalate) migration (9microgram/10 cm2/min) from toys and childcare articles under in-vitro conditions. From the data available so far, it does not seem feasible to obtain such a high DINP release in in-vitro experiments employing physiological like conditions.
General Comments
The aim of the work program is to validate 3 in vitro test methods for the migration of plasticiser(s) from toys and childcare articles using DINP as a model plasticiser. It is expected that the method(s) validated through this exercise may be applied for testing the migration of other plasticisers. The validation protocol is mainly in accordance with the recommended guidelines, except for the following (including the unresolved issue mentioned above):
1. The target migration rate: CSTEE recommended the need of a validated in vitro test method1, which could simulate the worst case DINP migration (9 microgram/10 cm2/min) observed in the in vivo study2. This value should not be confused with the safety of toys containing DINP. On the basis of TDI, a toy/childcare article will be safe when the DINP migration rate from the product is < 6.7 microgram/10 cm2/min (child< 3 years, weight 8 kg, sucking/chewing period 180 min; TDI 1200 microgram)1. Thus, to distinguish between safe and unsafe toys/childcare articles, the target rate for the method validation should be 6.7 microgram/10 cm2/min. But, the validated method should also be able to demonstrate that the worst case migration, observed in the in vivo study2, can be simulated in in vitro experiments. One of the several toys/reference PVC discs used in the in vivo studies2-4 can be used for this demonstration, as 15 % of the test population in these studies was able to achieve DINP migration >6.7 microgram/10 cm2/min.
2. The concern about feasibility of achieving rather high migration rate under physiological like conditions: The data available to CSTEE indicated that the target DINP migration rate (9microgram/10 cm2/min) may not be achieved under physiological like conditions. Therefore, CSTEE recommended in the Addendum of 25th November 1999 that conditions 'as close as possible' to in vivo conditions should be used for the method development and validation. The only test method, available so far, for relatively highly release of DINP migration in vitro employs rather aggressive and non-physiological experimental conditions - the LGC stringent method5. It is possible that microscopic particles may release from the test material under the experimental conditions such as those employed in the LGC stringent method. In such a case, it must be ascertained that the measured DINP release is not due to the presence of microscopic particles in the extraction medium. This will require a suitable method for testing of integrity of the test material.
3. LGC mild horizontal shaking method and TNO rotation method: These methods have not been able to simulate the rather high phthalate migration observed in the in vivo studies. The reasons for validating these methods is to get an insight on dispersion of data and sources of variation in the results reported recently5,6. Furthermore, it is expected that these methods, when validated using DINP as a model plasticiser, can be used as a general method(s) for testing plasticiser migration from toys and childcare articles. However, no reason has been provided to believe that these methods may be better for the analysis of migration of other plasticiser(s) than DINP.
It is proposed in the validation programme that the migration rate of a plasticiser from a toy/childcare article, determined by the validated method, may be corrected using a 'correction factor' to obtain maximum possible migration in vivo; and/or migration data may be statistically extrapolated to evaluate the safety of the product. It should be noted that correction factor for a plasticiser migration can only be derived when the migration rate of the plasticiser is available from a validated in vivo study. At present, correction factors can possibly be derived for DINP migration. As regards the statistical extrapolation of the migration data, an approach (using the DINP migration data from a TNO study6 and data from a child observation study7) for safety of toys has been proposed by NL8. However, the global acceptance of such an approach for safety assessment may require validation of the proposed method as well as an extended 'child observation study' with respect to children's mouthing behaviour may be required.
4. The CSTEE recommends that the validation exercise includes at least 15 laboratories (according to international guidelines), but it can be accepted if there are acceptable results from 8 laboratories available for the evaluation of the method performance.
Specific Comments
Page 3, SOP, Expert group discussion, line 3: The migration limits have been defined on the basis of the risk assessment, and standard deviation <20 % (of the mean of the repeated observations) has been recommended. The performance of the validated method for the determination of DINP migration should take that into account, but not vice versa.
Page 3, Target rate, Expert group discussion, para 2: See General Comments 2 & 3.
Page 5, Saliva simulant, Expert group discussion: According to the TNO report V2530, the proposed saliva simulant composition is based on the literature survey together with the prior consultation with an expert on oral biochemistry. The expert suggested the use of either amylase or albumin in the saliva simulant, but the saliva simulant described in the validation protocol does not contain any of these proteins. It should be noted that the CSTEE also recommended the use of some organic content in the saliva simulant. In the beginning of method development, TNO employed 0.16% mucin in the simulant but at a later stage the mucin was removed from the simulant. The apparent reason for using the mucin was that this protein has earlier been used by others as a substitute of natural proteins in saliva simulant. Furthermore, it appears that the determination of DINP migration rates should have been performed employing simulant containing mucin, amylase or albumin in preliminary experiments (TNO report V2530), but no results are available. The reason for removing mucin from the saliva simulant was that significantly lower (approximately 5x)2 DINP migration rate was observed in the presence of mucin. A possible explanation for this phenomenon may be that the DINP formed micelles/emulsion with the protein (mucin) present in the simulant. During extraction of the protein-simulant, the DINP may be precipitated together with the protein. This effect has possibly been observed in the determination of DINP in natural saliva - DCG in-vivo study2. In this study, recovery of DINP in test person's own saliva was determined and the results were corrected for the recovery percent. However, details of processing the saliva for DINP determination and the DINP recovery percents are not described in the report. (For a correct estimate recovery percent >50 may be accepted, provided that is reproducible). Therefore, it is not possible to comment on the results obtained in the in-vivo study.
One of the arguments to validate TNO rotation method and LGC mild shaking method is that the DINP migration determined by these methods may be extrapolated, using a correction factor, to estimate the 'real' DINP migration (in-vivo). To establish such a 'correction factor', it may also be necessary to establish the correction/correlation factor for DINP migration determined in the presence and absence of natural saliva, because the proposed saliva simulant is devoid of proteins present in natural saliva.
Page 6, # 3 from top (check of sample integrity): Check of sample integrity must be performed when LGC stringent method is used. In the first Opinion on phthalate migration from toys9, CSTEE already mentioned that 'further knowledge about possible exposure to released particles and direct transfer to the oral cavity and ingestion is needed'. However, particle release from test materials has not been checked in any of the in vivo studies2-4. Furthermore, the influence of any particle release in determination of DINP migration rate has not been considered in these studies. The CSTEE will include this aspect in risk assessment when data (from in vivo studies) on the possibility of released particles is available.
Page 6, Analytical determination, Expert group discussion, para 3: Results of all determinations must be presented together with the statistical analysis.
Page 7, Expert group discussion, para 7: As mentioned in General Comments 3, an evaluation of the suitability of the validated in vitro method (for DINP migration) for other substances will require migration data from validated in vivo studies employing test materials containing the respective substances.
Page 10, Homogeneity testing, para 3: The homogeneity testing should be performed for the content(s) of phthalate(s) concerned. CSTEE is not aware of an accredited method/ standard method for the determination of phthalate content in toy samples.
Page 11, Disks and toy sources, para 4: It is a good idea that 'JRC can act as receiving office and interface needed' (for future distribution of reference material as well).
Page 13, Item action 1, HPLC method: It should not be necessary to make use of a HPLC method, because quantitative analysis of phthalates can be performed by the same GC-MS method, which has to be used for the qualitative analysis.
Page 14, Timeline: Considering that some less experienced laboratories may participate in the validation exercise, the timeline described seems to be very tight.
Conclusions
The programme for the validation methodologies aims to achieve a suitable method for testing DINP migration from toys and childcare articles. Furthermore, it is expected that the validated method will form a basis for the development of suitable methods for testing the migration of other plasticisers/additives from toys and childcare articles. The validation protocol is, to a great extent, in accordance with the guidelines recommended by CSTEE. The CSTEE's comments to strengthen/complement the validation programme are described above. The important suggestion for further improving the validation programme is that the validation exercise should focus on DINP migration employing the test method(s), which may possibly simulate the relatively high DINP migration (9 microgram/10 cm2/min) observed in an in vivo study. When absolutely necessary, non-physiological/ aggressive conditions may be used for simulating DINP migration in vitro, but the integrity of the test material must be ascertained. A correction/correlation factor between the DINP migration determined by the present protocol and in natural saliva has to be established. Furthermore, it must be demonstrated that a correct distinction can be made between safe and unsafe products, when the validated method is used for migration testing. An acceptance of NL approach8 for statistical extrapolation of DINP migration data (from in vitro studies) for distinguishing between safe and unsafe products requires validation of this approach. The 'correction factor' (for converting the observed in vitro migration to maximum possible migration in vivo), to evaluate the safety of a product may possibly be derived for DINP, but currently not for other substances. (In order to do this, in-vivo exposure studies with the specific substances under evaluation have to be conducted). For establishing a correction factor for DINP an additional in-vivo study will be required, because limited data is available on particle release during chewing/sucking of a toy. Furthermore, data on toxicokinetics of phthalates are also required. However, these recommendations should not postpone the proposed validation programme of in-vitro methods of migration testing.
1. References
1. Opinion of Phthalate migration from soft PVC toys and child-care article - data made available since the 16th of June 1998, opinion expressed at the 6th CSTEE plenary meeting, Brussels, 26/27 November 1998.
2. Report from the Dutch Consensus Group: Phthalate release from soft PVC toys. RIVM report 613320 002, RIVM, NL. September 1998.
3. Steiner I., Kubesch K. and Fiala F.: Migration of DEHP and DINP from PVC articles. Vienna University of Technology and Austrian Standards Institute. September 3, 1998.
4. The risk of chronic toxicity associated with exposure to diisononyl phthalate (DINP) in children's products. US CPSC, December 1998.
5. Interlaboratory validation of laboratory-based agitation methods for the determination of phthalate plasticiser migration from PVC toys and childcare articles. LGC Technical Report No. LGC/1999/DTI/004 , LGC, UK. June 1999.
6. Validation of the method "Determination of diisononyl phthalate in saliva simulant". TNO report V99.598, TNO, NL. May 1999.
7. Groot M.E., LekkerkerK M.C. and Dteenbekkers L.P.A.: Mouthing behaviour of young children - An observational study. Wageningen Agricultural University, Wageningen, NL. September 1998.
8. CSTEE/97/1 - Add 174
9. Phthalate migration from soft PVC toys and child-care articles - opinion expressed at the CSTEE plenary meeting, Brussels, 24 April 1998.
Background and Terms of Reference
In the light of the opinion of the CSTEE of 28th September 1999 on "TNO, LGC and US CPSC reports on Phthalate migration test validation" and the Addendum of 25th November 1999 to this opinion, a programme for a full validation of test methods for phthalate migration has been developed and that is presented to CSTEE. The committee on the basis of the Programme (CSTEE/2000/13 Add 1) for the validation of methodologies to test migration of plasticisers from toys and childcare articles is to answer the following questions:
- Is the approach of the Programme suitable to provide for a sound validation of methods to measure satisfactorily the release of phthalates from toys and child care articles as to distinguish safe from unsafe products?
- Which comments can the Committee offer to further strengthen or complement the programme?
Introduction
It is proposed in the work programme that 3 of the in vitro test methods, available at present, will be validated:
- Horizontal shaking of toy material with glass balls in an artificial saliva at 37° C (LGC mild horizontal shaking method),
- Horizontal shaking of toy material with relatively heavy metal balls in an artificial saliva at 60° C (LGC stringent horizontal shaking method)
- Dynamic extraction of plasticiser from toy material in an artificial saliva employing a rotation method at 22° C (TNO rotation method)
The reasons given for validating the 3 methods are: "The stringent conditions may allow to simulate the 'worse case conditions' (referring to DINP migration rate 9microgram/10 cm2/min, adopted by CSTEE)1, whereas TNO method and the mild version of LGC method may be easily applied to a wide range of testing and other substances of various target values".
The validation protocol utilises to a great extent the guidelines, which are described in the CSTEE Opinion of 28th September 1999 and in the Addendum of 25th November 1999 to this Opinion. A major issue identified, however, is the problem of achieving the target rate for the validation of test method for DINP (diisononyl phthalate) migration (9microgram/10 cm2/min) from toys and childcare articles under in-vitro conditions. From the data available so far, it does not seem feasible to obtain such a high DINP release in in-vitro experiments employing physiological like conditions.
General Comments
The aim of the work program is to validate 3 in vitro test methods for the migration of plasticiser(s) from toys and childcare articles using DINP as a model plasticiser. It is expected that the method(s) validated through this exercise may be applied for testing the migration of other plasticisers. The validation protocol is mainly in accordance with the recommended guidelines, except for the following (including the unresolved issue mentioned above):
1. The target migration rate: CSTEE recommended the need of a validated in vitro test method1, which could simulate the worst case DINP migration (9 microgram/10 cm2/min) observed in the in vivo study2. This value should not be confused with the safety of toys containing DINP. On the basis of TDI, a toy/childcare article will be safe when the DINP migration rate from the product is < 6.7 microgram/10 cm2/min (child< 3 years, weight 8 kg, sucking/chewing period 180 min; TDI 1200 microgram)1. Thus, to distinguish between safe and unsafe toys/childcare articles, the target rate for the method validation should be 6.7 microgram/10 cm2/min. But, the validated method should also be able to demonstrate that the worst case migration, observed in the in vivo study2, can be simulated in in vitro experiments. One of the several toys/reference PVC discs used in the in vivo studies2-4 can be used for this demonstration, as 15 % of the test population in these studies was able to achieve DINP migration >6.7 microgram/10 cm2/min.
2. The concern about feasibility of achieving rather high migration rate under physiological like conditions: The data available to CSTEE indicated that the target DINP migration rate (9microgram/10 cm2/min) may not be achieved under physiological like conditions. Therefore, CSTEE recommended in the Addendum of 25th November 1999 that conditions 'as close as possible' to in vivo conditions should be used for the method development and validation. The only test method, available so far, for relatively highly release of DINP migration in vitro employs rather aggressive and non-physiological experimental conditions - the LGC stringent method5. It is possible that microscopic particles may release from the test material under the experimental conditions such as those employed in the LGC stringent method. In such a case, it must be ascertained that the measured DINP release is not due to the presence of microscopic particles in the extraction medium. This will require a suitable method for testing of integrity of the test material.
3. LGC mild horizontal shaking method and TNO rotation method: These methods have not been able to simulate the rather high phthalate migration observed in the in vivo studies. The reasons for validating these methods is to get an insight on dispersion of data and sources of variation in the results reported recently5,6. Furthermore, it is expected that these methods, when validated using DINP as a model plasticiser, can be used as a general method(s) for testing plasticiser migration from toys and childcare articles. However, no reason has been provided to believe that these methods may be better for the analysis of migration of other plasticiser(s) than DINP.
It is proposed in the validation programme that the migration rate of a plasticiser from a toy/childcare article, determined by the validated method, may be corrected using a 'correction factor' to obtain maximum possible migration in vivo; and/or migration data may be statistically extrapolated to evaluate the safety of the product. It should be noted that correction factor for a plasticiser migration can only be derived when the migration rate of the plasticiser is available from a validated in vivo study. At present, correction factors can possibly be derived for DINP migration. As regards the statistical extrapolation of the migration data, an approach (using the DINP migration data from a TNO study6 and data from a child observation study7) for safety of toys has been proposed by NL8. However, the global acceptance of such an approach for safety assessment may require validation of the proposed method as well as an extended 'child observation study' with respect to children's mouthing behaviour may be required.
4. The CSTEE recommends that the validation exercise includes at least 15 laboratories (according to international guidelines), but it can be accepted if there are acceptable results from 8 laboratories available for the evaluation of the method performance.
Specific Comments
Page 3, SOP, Expert group discussion, line 3: The migration limits have been defined on the basis of the risk assessment, and standard deviation <20 % (of the mean of the repeated observations) has been recommended. The performance of the validated method for the determination of DINP migration should take that into account, but not vice versa.
Page 3, Target rate, Expert group discussion, para 2: See General Comments 2 & 3.
Page 5, Saliva simulant, Expert group discussion: According to the TNO report V2530, the proposed saliva simulant composition is based on the literature survey together with the prior consultation with an expert on oral biochemistry. The expert suggested the use of either amylase or albumin in the saliva simulant, but the saliva simulant described in the validation protocol does not contain any of these proteins. It should be noted that the CSTEE also recommended the use of some organic content in the saliva simulant. In the beginning of method development, TNO employed 0.16% mucin in the simulant but at a later stage the mucin was removed from the simulant. The apparent reason for using the mucin was that this protein has earlier been used by others as a substitute of natural proteins in saliva simulant. Furthermore, it appears that the determination of DINP migration rates should have been performed employing simulant containing mucin, amylase or albumin in preliminary experiments (TNO report V2530), but no results are available. The reason for removing mucin from the saliva simulant was that significantly lower (approximately 5x)2 DINP migration rate was observed in the presence of mucin. A possible explanation for this phenomenon may be that the DINP formed micelles/emulsion with the protein (mucin) present in the simulant. During extraction of the protein-simulant, the DINP may be precipitated together with the protein. This effect has possibly been observed in the determination of DINP in natural saliva - DCG in-vivo study2. In this study, recovery of DINP in test person's own saliva was determined and the results were corrected for the recovery percent. However, details of processing the saliva for DINP determination and the DINP recovery percents are not described in the report. (For a correct estimate recovery percent >50 may be accepted, provided that is reproducible). Therefore, it is not possible to comment on the results obtained in the in-vivo study.
One of the arguments to validate TNO rotation method and LGC mild shaking method is that the DINP migration determined by these methods may be extrapolated, using a correction factor, to estimate the 'real' DINP migration (in-vivo). To establish such a 'correction factor', it may also be necessary to establish the correction/correlation factor for DINP migration determined in the presence and absence of natural saliva, because the proposed saliva simulant is devoid of proteins present in natural saliva.
Page 6, # 3 from top (check of sample integrity): Check of sample integrity must be performed when LGC stringent method is used. In the first Opinion on phthalate migration from toys9, CSTEE already mentioned that 'further knowledge about possible exposure to released particles and direct transfer to the oral cavity and ingestion is needed'. However, particle release from test materials has not been checked in any of the in vivo studies2-4. Furthermore, the influence of any particle release in determination of DINP migration rate has not been considered in these studies. The CSTEE will include this aspect in risk assessment when data (from in vivo studies) on the possibility of released particles is available.
Page 6, Analytical determination, Expert group discussion, para 3: Results of all determinations must be presented together with the statistical analysis.
Page 7, Expert group discussion, para 7: As mentioned in General Comments 3, an evaluation of the suitability of the validated in vitro method (for DINP migration) for other substances will require migration data from validated in vivo studies employing test materials containing the respective substances.
Page 10, Homogeneity testing, para 3: The homogeneity testing should be performed for the content(s) of phthalate(s) concerned. CSTEE is not aware of an accredited method/ standard method for the determination of phthalate content in toy samples.
Page 11, Disks and toy sources, para 4: It is a good idea that 'JRC can act as receiving office and interface needed' (for future distribution of reference material as well).
Page 13, Item action 1, HPLC method: It should not be necessary to make use of a HPLC method, because quantitative analysis of phthalates can be performed by the same GC-MS method, which has to be used for the qualitative analysis.
Page 14, Timeline: Considering that some less experienced laboratories may participate in the validation exercise, the timeline described seems to be very tight.
Conclusions
The programme for the validation methodologies aims to achieve a suitable method for testing DINP migration from toys and childcare articles. Furthermore, it is expected that the validated method will form a basis for the development of suitable methods for testing the migration of other plasticisers/additives from toys and childcare articles. The validation protocol is, to a great extent, in accordance with the guidelines recommended by CSTEE. The CSTEE's comments to strengthen/complement the validation programme are described above. The important suggestion for further improving the validation programme is that the validation exercise should focus on DINP migration employing the test method(s), which may possibly simulate the relatively high DINP migration (9 microgram/10 cm2/min) observed in an in vivo study. When absolutely necessary, non-physiological/ aggressive conditions may be used for simulating DINP migration in vitro, but the integrity of the test material must be ascertained. A correction/correlation factor between the DINP migration determined by the present protocol and in natural saliva has to be established. Furthermore, it must be demonstrated that a correct distinction can be made between safe and unsafe products, when the validated method is used for migration testing. An acceptance of NL approach8 for statistical extrapolation of DINP migration data (from in vitro studies) for distinguishing between safe and unsafe products requires validation of this approach. The 'correction factor' (for converting the observed in vitro migration to maximum possible migration in vivo), to evaluate the safety of a product may possibly be derived for DINP, but currently not for other substances. (In order to do this, in-vivo exposure studies with the specific substances under evaluation have to be conducted). For establishing a correction factor for DINP an additional in-vivo study will be required, because limited data is available on particle release during chewing/sucking of a toy. Furthermore, data on toxicokinetics of phthalates are also required. However, these recommendations should not postpone the proposed validation programme of in-vitro methods of migration testing.
1. References
1. Opinion of Phthalate migration from soft PVC toys and child-care article - data made available since the 16th of June 1998, opinion expressed at the 6th CSTEE plenary meeting, Brussels, 26/27 November 1998.
2. Report from the Dutch Consensus Group: Phthalate release from soft PVC toys. RIVM report 613320 002, RIVM, NL. September 1998.
3. Steiner I., Kubesch K. and Fiala F.: Migration of DEHP and DINP from PVC articles. Vienna University of Technology and Austrian Standards Institute. September 3, 1998.
4. The risk of chronic toxicity associated with exposure to diisononyl phthalate (DINP) in children's products. US CPSC, December 1998.
5. Interlaboratory validation of laboratory-based agitation methods for the determination of phthalate plasticiser migration from PVC toys and childcare articles. LGC Technical Report No. LGC/1999/DTI/004 , LGC, UK. June 1999.
6. Validation of the method "Determination of diisononyl phthalate in saliva simulant". TNO report V99.598, TNO, NL. May 1999.
7. Groot M.E., LekkerkerK M.C. and Dteenbekkers L.P.A.: Mouthing behaviour of young children - An observational study. Wageningen Agricultural University, Wageningen, NL. September 1998.
8. CSTEE/97/1 - Add 174
9. Phthalate migration from soft PVC toys and child-care articles - opinion expressed at the CSTEE plenary meeting, Brussels, 24 April 1998.