Opinion on Phthalate migration from soft PVC toys and child-care articles

Opinion on Phthalate migration from soft PVC toys and child-care articles - Data made available since the 16th of June 1998, opinion expressed at the 6th CSTEE plenary meeting, Brussels, 26/27 November 1998

On 15 October 1998 the CSTEE was given the task to answer the following questions:

1. - In the light of the new evidence ('Dutch Consensus Group' study and other data/studies made available to the CSTEE since its plenary meeting of June 1998) how do the respective results change the previous opinions of the CSTEE, in particular the concern expressed before in the opinion on 'Phthalate migration from soft PVC toys and child-care articles' of 24th April 1998?

2. - What further work/activities does the CSTEE deem necessary in order to develop and validate reference methodologies for the purposes of the safety control of those articles in respect of which the CSTEE has expressed its previous opinions?

The opinion of the CSTEE expressed on 24 April 1998 was based on a worst-case approach using the highest available leachate data reported. Being aware of an ongoing Dutch study in adult human volunteers which would in a more comprehensive fashion examine the human exposure to phthalates in toys, the CSTEE recognised at that time that its evaluation could be modified as the results of the Dutch, as well as other studies, became available.

Since 24 April 1998 the CSTEE has learned that the high level of release of DINP which was used in its evaluation, could not be reproduced in a study of the same PVC sample by a US Consumer Product Safety Commission laboratory (CSTEE/97/1-Add 103). Valuable information was also achieved from an Austrian study in volunteers on the release of DINP and DEHP into saliva from PVC-sheets and -toys (CSTEE/97/1-Add 102). The mean releases of DINP after sucking found in the Dutch and Austrian studies were similar, 146 and 132 m g/10 cm² and hour, respectively.

In its meeting on 15 September 1998, the CSTEE was presented with a draft summary of the report of the study of DINP release from soft PVC baby toys (Draft report from the 'Dutch Consensus Group' on the release of phthalates from soft PVC baby toys of 9 September 1998). This study include a human volunteer study to determine the release rates of DINP from PVC samples into saliva, a child observation study to determine the oral contact time of young children with baby toys, an assessment of the exposure of babies to DINP from soft PVC toys, and a development of a routine laboratory method to determine the release rate of DINP from PVC baby toys.

The CSTEE judges the Dutch study as being of sound scientific quality and finds that the data produced are of greater relevance than those used in its previous opinion. This means that the earlier CSTEE use of DINP release should be lowered and the exposure time reduced from a maximum of 6 hrs to 3 hrs (approximate maximal total mouthing time for 6-12 months old children). Given this exposure period the maximum DINP release (based on the highest release rate in the volunteer study) is 1600 m g/10 cm² (8.9 m g/min x 180 min). A daily dose of DINP for a child weighing 8 kg would then be 200 m g/kg, compared to the 24 April 1998 figure of 1700 m g/kg/day for a 6-hr exposure period. The Austrian study found a similar maximum DINP release as the Dutch study, namely 1610 m g/cm² per 3 hrs.

More data relevant for establishing NOAEL values for several of the phthalates have been available to the CSTEE after it expressed its opinion on 24 April 1998. The final reports of the Aristech Chemical Corporation oncogenicity studies in rats and mice on DINP have revealed that this phthalate induces hepatic neoplasms in both species (Covance, 1998, study numbers 2598-104 and 2598-105). Although biochemical evidence for peroxisome proliferation was evident in both studies, no evidence for sustained cell proliferation was found. The lowest NOAEL in the rat study was in males given 1,500 ppm DINP in the feed, corresponding to a daily dose of 88 mg/kg. In the mice the lowest NOAEL was in females given 500 ppm DINP in the feed, corresponding to a daily dose of 112 mg/kg. In the rat study renal tubule cell carcinomas were observed in males at the highest dose level (12,000 ppm corresponding to 733 mg/kg/day). A still unpublished study by Caldwell et al. has determined a dose-dependent accumulation of the alpha 2u globulin (a 2UG) in male rats from 6,000 ppm. They also observed an increase in cell proliferation in areas confined to the a 2UG accumulation. The CSTEE therefore judges that the renal neoplasms induced by DINP in male rats have no relevance for the human situation. The lowest NOAEL for DINP reported from the Aristech Chemical Company is higher than that previously used by CSTEE (88 vs. 15 mg/kg/day). Since the NOAEL of 15 mg/kg/day was derived from a recent, well-conducted study (Lington et al., 1997), the CSTEE will from a precautionary standpoint still use this value in its revised assessment.

A recently published study where female Long-Evans rats were exposed to 32.5 and 325 m l DEHP/L in drinking water from day 1 of pregnancy to day 21 after the delivery, produced severe histological damage to the testes of the offspring at both concentrations (Arcadi et al.: Fd Chem Toxicol, 963-970, 1998; Add 123). The water intake was not precisely measured, but the lowest concentration corresponded to an intake of about 3.0-3.5 mg/kg/day. This LOAEL substantiates the identification of a NOAEL of 3.7 mg/kg DEHP/day determined in the Poon et al. (1997) study, based on mild Sertoli cell vacuolation observed at the next higher dose. The CSTEE will therefore use this 3.7 mg/kg/day dose as a revised NOAEL for DEHP, instead of the value of 5 mg/kg/day used previously. The testicular effects observed in the Poon et al. (1997) and Arcadi et al. (1998) studies are judged to have greater relevance for humans than the hepatic peroxisome proliferation, which was the critical effect for the 5 mg/kg/day DEHP NOAEL value. This is underscored by the finding that DEHP causes testicular effects in transgenic mice lacking the essential receptor for peroxisome proliferator-associated liver effects (Ward et al., 1998; Add 100C).

In conjunction with the NTP carcinogenicity study on BBP (NTP, 1997) a 10-week modified mating study in rats was also performed. At a dose of 200 mg BBP/kg/day (2,800 ppm in feed) to a group of rats, the epididymal spermatozoal concentration was significantly less than the controls. The NOAEL for these effects was 20 mg/kg/day. For its reassessment the CSTEE will apply this value instead of the LOAEL of 171 mg/kg/day (incorporating an additional uncertainty factor of 2), as was done in its opinion of 24 April 1998.

Recalculating the margin of safety (MOS) relative to the no-observed-adverse-effect-level (NOAEL) of DINP of 15 mg/kg/day, results in a value of 75 (15 000 m g/kg/day divided by 200 m g/kg/day), in contrast to the MOS of the 24 April 1998 opinion of 8.8 (see table 1). Such a margin of safety still raises some concern since it is less than the recommended safety margin of at least 100.

Table 1. - Critical effects, NOAEL values, maximum emission rates, intake doses and margins of safety

^ALOAEL value ^B

An additional uncertainty factor of 5 has been incorporated

The Dutch volunteer study did not take into account the possibility of buccal absorption of DINP during the experiment, thus an underestimation of the real exposure may have been introduced. However, the contact time for DINP-containing saliva was short before the samples were voided. Buccal absorption may, at least in part, proceed via passage of DINP through the salivary water phase where most of the material is suspended and/or emulsified, and not dissolved. On the other hand, the Dutch study may overestimate exposure since all of the DINP released into the saliva was considered available for oral intake. With small children, certainly a considerable part of the saliva is not retained in the mouth and subsequently swallowed. The CSTEE therefore judges that any underestimation of not taking buccal absorption into account may be counteracted by the overestimation of assuming that all saliva is retained. Although it would have been very useful to assess any contribution from buccal uptake of phthalates to the total body burden of these compounds from sucking and chewing on PVC toys, to perform such a study in full would be complicated. The Dutch study did not determine any presence of products formed from salivary hydrolysis of DINP, so any potential hazard related to such components could not be assessed.

The Dutch volunteer study only investigated the release of DINP. The release of DEHP during sucking (not using the teeth) was investigated in the Austrian study (CSTEE/97/1-Add 102). The result was almost the same as for the corresponding release of DINP in the same type of experiment. It may therefore be assumed that the DEHP release when the toy in addition is chewed, also would have been of the same magnitude as for DINP in the chewing experiment. For DEHP, although the CSTEE does not have any individual release data for this substance, it will use the same maximum release rate as for DINP as a worst-case scenario. This means that also for DEHP a daily dose for a child weighing 8 kg would be 200 m g/kg. With the new NOAEL of 3.7 mg/kg/day, this would result in a MOS of only 19 (table 1). This raises more concern than before since it is lower than the previous value of 67 and because the critical effect (testicular damage) is judged to be of higher relevance than hepatic peroxisome proliferation.

For the other phthalates discussed in the 24 April 1998 opinion, more reliable exposure data from volunteer studies are not available. Recalculating the margin of safety for the new NOAEL of 20 mg/kg/day for BBP results in a MOS of 20 000, compared to the previous value of 26 000. The margins of safety for DNOP, DIDP, BBP and DBP calculated on the maximally reported release data are judged to be beyond a reason for concern (MOS > 380-26 000, table 1). This is due to the fact that the concentrations of these phthalates in toys are low. Should they be used as plasticisers in higher concentrations, larger releases from the toys would occur and lower margins of safety would be expected.

The present re-evaluation for DINP was based on data of release from one typical toy sample only. The toy contained approximately 43% (w/w), which is typical for these types of products. The present evaluation must be seen in the light that the release of DINP from other products could vary somewhat, although the release data from the Dutch study were supported by the Austrian findings.

The Dutch study also developed a routine laboratory method which appeared to provide release data that were reproducible and which obtained results that corresponded to the levels observed in the human volunteer study. The CSTEE judges these results to be promising, but strongly recommends that an interlaboratory comparison exercise be carried out in order to document the reproducibility of the method. The interlaboratory comparison should be performed using a target value for DINP release of 9 m g/10cm²/min, which is the maximum release rate observed in the volunteer studies.

The Dutch study also provided information pertaining to the total time that children are mouthing PVC toys. Since this study, as far as the CSTEE is aware of, is the only observational study of this kind and also may be limited due to its design (performed in a single country during the summer with a small number of participants from a select socio-economic group), it is recommended that additional studies relating to this issue be carried out.

In its opinion of 24 April 1998, the CSTEE proposed guidance values for maximum extractable amounts of phthalates from children's' toys. The approach used was based on the view that a margin of safety of at least 100 from their respective NOAEL values should be incorporated. In the risk assessment, the MOS would be reduced to half of those values given in table 1 if the mouthing period is 6 hrs rather than 3 hrs. However, the guidance values given in Table 2 based on total amounts extracted, would not be affected. As stated, on the basis of new information regarding toxicity of DEHP and BBP, the CSTEE has revised NOAEL values for these phthalates and adjusted the guidance values accordingly (table 2).

There is limited information on other sources of exposure for phthalates in children. A recent publication (CSTEE/97/1-Add 130, attachment 7) gives levels of DEHP in mother's milk and infant formulae of up to about 200 m g/kg. Assuming a daily intake of 1 kg of dairy products for children weighing 8 kg gives a daily dose of 25 m g DEHP/kg corresponding to approx. 70% of TDI and approx. 12.5% of the intake from soft PVC toys. Exposure via air is judged to be small. Dermal absorption of DEHP, as for other phthalates, from skin contact with soft PVC toys is a possibility, but it is difficult to assess the magnitude of such an exposure. However, the intake from toys is judged to be the major source for DEHP exposure. Information on other phthalates is still sparse, but the 'Dutch Consensus Group' assumes that 0.05 mg DINP/kg/day may come from exposures apart from those due to baby toys. However, this estimate is quite uncertain so it does not seem appropriate to include this in a margin of safety consideration. On the other hand, factoring in the additional exposures of DINP, the level of concern would increase.

In a situation where toys may contain more than one phthalate, the cumulative effect would be expected to be additive, given that the critical effects were the same. For phthalates with different critical effects, as for DINP and DEHP, the assessment should be done separately for the two compounds.

Table 2. - NOAEL values, tolerable daily intakes and guidance values for phthalates in toys

^ALOAEL value

^BAdditional uncertainty factor of 5 incorporated

Conclusion

1. - The evaluation of the data from the 'Dutch Consensus Group' volunteer study on the release of DINP from PVC toys as well as on child behaviour mouthing PVC toys, in conjunction with results from an Austrian volunteer study on DINP and DEHP, has led the CSTEE to revise its opinion of 24 April 1998 with respect to the exposure of children to these compounds and thereby the concern arising from such exposures.

- It is estimated that children with average body weights of 8 kg being exposed for 3 hrs to PVC toys containing DINP and DEHP will have maximal daily intakes of 200 m g/kg for both substances. Newer data on the toxicity of DEHP have led the CSTEE to change its previous NOAEL designation to 3.7 mg/kg/day with testicular damage as the critical effect.

- The revised margin of safety (MOS) values are 75 for DINP and 19 for DEHP, respectively. The MOS for DINP raises some concern as it is less than the previously recommended safety margin of at least 100. The MOS for DEHP raises clear concern. Exposure to DINP and DEHP from other sources than soft PVC toys will increase the concern, but the magnitude of such exposures is uncertain.

2. - The CSTEE recommends that an interlaboratory comparison exercise be carried out in order to document the reproducibility of the Dutch laboratory method to measure phthalate release from PVC toys. Studies performed in order to shed light on the issue of buccal absorption of phthalates should be carried out. The CSTEE further recommends that additional studies be carried out in order gain more insight as to the total time children are mouthing PVC toys.