protecting the fetus and the child
from environmental hazards

a healthy environment
for all children

[ home ] [ resource guide ] [ training manual ] [ get involved ] [ publications ] [ index ] [ about CEHN ]

Comments on the Environmental Protection Agency's “Framework for a Voluntary Children's Chemical Evaluation Program” (dated 4/12/00)

• American Public Health Association •
• Children's Environmental Health Network •
• Environmental Defense •
• National Environmental Trust •
• Physicians for Social Responsibility •

The above pediatric, public health, and environmental organizations have reviewed the EPA's proposal for a voluntary chemical evaluation program. While our organizations strongly support the objectives of this initiative, we have serious concerns about the proposal. Although some aspects of the current EPA proposal would provide valuable new information on children's exposures to certain chemicals, this proposal is seriously flawed in both science and process and must be substantially reworked if it is to serve science and pediatric health.

1. The proposed group of approximately 50 candidate chemicals is generally appropriate for an initial effort, but additional contaminants should be added to the initial list.

For purposes of developing this program, we accept the recommendation of the Scientific Advisory Panel (SAP) to limit the initiative's scope to approximately 50 chemicals. With that focus, we support EPA's general approach of identifying a subset of compounds found to be present in human tissues through biomonitoring, with key additions to its candidate list (see below).

The agency “seeks comment on whether biomonitoring data is sufficient by itself to identify candidates for this program and whether the supplementary selection criteria of presence in food, drinking water or indoor air should serve only to identify priority candidates.” It is our belief that biomonitoring data should be sufficient by itself to indicate a need for full toxicity data. We can accept EPA's use of two criteria – namely biomonitoring data and presence in food, water, or air – for priority-setting purposes at this time. We believe that, ultimately, all chemicals found to be present in biological samples of humans from the general population (above de minimis levels or frequencies of occurrence) should have full toxicity data available, unless (i) those chemicals are no longer being manufactured or are already being phased out; (ii) are already so well-studied that additional data will not enhance risk management (e.g., lead); or (iii) are already being tested under other initiatives (e.g., certain pesticides). In addition, full toxicity data should be available for chemicals for which there is similarly compelling evidence of exposure, but for which biomonitoring is not conducted or is not an effective option due to the physical/chemical nature of the chemical.

Importantly, EPA should include the chemicals recently identified in Swedish breast milk monitoring programs – namely polychlorinated naphthalenes and polybrominated diphenyl ethers – unless adequate information is brought forward to indicate that the sources of Swedish populations' exposure to those chemicals do not occur in the US. Chemicals found in breast milk are, by definition, found both in biomonitoring programs and in children's food.

2. The endpoints for the subject chemicals should be those required under the FQPA 10x protocols, but a combined-study protocol should be developed on an expedited basis.

In general, we believe that the appropriate set of toxicology tests for this program is the testing battery developed for pesticides as a result of the Food Quality Protection Act (FQPA), as approved by the Scientific Advisory Panel. (Pregnant animals should be used in the pharmacokinetic/ metabolic studies, given this initiative's focus on children.)

Within that context, we believe there may be important opportunities to conserve laboratory animals and reduce testing costs, while still providing the appropriate scientific information, by utilizing a single study to evaluate multiple endpoints. Specifically, this approach would include a 2-generation reproductive/ developmental study with an extended observation period, and observation of developmental neurotoxicity, developmental immunotoxicity, in vivo mutagenicity, and chronic toxicity endpoints.

We suggest that the National Toxicology Program be asked to conduct, on a high-priority expedited basis, a workshop to define a combined protocol. It is our understanding that NTP would likely be able to complete such a workshop, and the associated report, within about six to nine months from the initiation of the effort, because this does not involve any change in sub-protocols, and in fact builds upon two-generation study designs already in use under EPA regulations.

3. The proposed tiered test-selection process is inappropriate and not supported by science.

For chemicals found to be actually present in human tissues (above de minimis levels or frequencies of occurrence), tiering is not appropriate. Since a tiered approach by definition proceeds from simpler, less costly, and less accurate tests to more complex, expensive, and definitive tests, we believe that a tiered approach is not acceptable for such compounds.

The one possible exception may – and we stress “may” – be carcinogenicity. For cancer, there is a base of experience for judgment-based selection of chemicals for the cancer bioassay. There have been specific studies assessing the utility and validity of particular judgment-based systems. By contrast, no such experience and studies exist for other endpoints such as reproductive, developmental, and neurotoxic effects. However, even if it is determined that there is an adequate scientific basis to support making carcinogenicity a separate “triggered” test, it is also critical to recognize that some chemicals – notably those that are structurally similar to known carcinogens – would also warrant cancer testing.

Under EPA's proposal, all of the candidate chemicals have been found in human biological samples, thus providing essentially definitive evidence of exposure. Such evidence outweighs any arguments for proceeding through a lengthy set of preliminary tests until the chemical is finally evaluated in sophisticated tests. Under this approach, compounds found in breast milk, in biological specimens derived from children, and in blood of women of child-bearing age through general-population studies (e.g., NHANES) are of the very highest priority for full hazard evaluation through testing (where not yet conducted).

The EPA has not justified the implementation of a tiered approach. Indeed, as EPA itself acknowledges,

"In this case, EPA's analysis, which was supported by the SAP [Scientific Advisory Panel] in its review, indicates that the understanding needed to support triggers based on biology does not presently exist.”

Absent a scientific basis that supports triggers (i.e., potentially for carcinogenicity), there is no rationale for a tiered system. Additionally, as discussed below, the proposed tiered system has a number of severe practical limitations.

Indeed, given that SAP supported EPA's initial single-tier approach, it appears that EPA's basis for proposing a tiered system is the fact that the Chemical Manufacturers Association (CMA) and other industry representatives want such an approach. In support of its position, CMA has proffered a “retrospective evaluation” of nine chemicals. In effect, CMA's study appears to be the foundation for the agency's proposal to proceed with a tiered testing approach.

To date, that study, the chemicals included, the methodology used, and its conclusions, have not been made public nor have they undergone peer review. The agency should at least make public its scientists' review of this study and should sponsor an expedited independent external peer review of the CMA evaluation if EPA is going to continue to consider a tiered approach.

While we strongly oppose a tiered approach for chemicals detected through reliable biomonitoring programs (again with the possible exception of carcinogenicity), we believe that a tiered approach may be appropriate for structuring the evaluation of other compounds for which exposure information is less definitive. Accordingly, we encourage CMA to continue to develop and validate a tiered system for use with a broader array of chemicals to which children may be exposed.

4. The proposed peer consultation process is unjustified and unworkable.

The EPA states that it “[believes] that the VCCEP is likely to conclude that a full toxicity and exposure evaluation will be needed for each chemical.” Yet the agency proposes relying on two sets of peer consultation for each chemical to develop two weight-of-the-evidence analyses on whether to conduct additional testing. One consultation is to be held after Tier 1 (which includes review of all existing information) and a second after completion of Tier 2.

We are unable to identify circumstances under which a weight-of-the-evidence evaluation would lead to a conclusion that it is not necessary to conduct neurotoxicity and chronic toxicity tests for chemicals known to be present in human tissues. No examples of circumstances that would support the derogation of neurotoxicity and chronic toxicity tests for these chemicals have been identified. If no such circumstances can reasonably be anticipated, it is a waste of limited scientific resources (of industry, government, and the public interest community) to set up a labor-intensive weight-of-the-evidence evaluation for each chemical.

As proposed by EPA, a program involving 50 chemicals would involve 100 peer consultations, each performing a weight-of-the-evidence analysis on whether to conduct additional testing. Our organizations have repeatedly expressed strong concerns to the agency about the lack of resources in our community that would be available for such an extensive process. It is especially difficult to justify such a labor-, time- and resource-intensive process when the agency has already identified the likely outcome -- that the ultimate conclusion in virtually all cases would be that the full array of toxicity (except possibly carcinogenicity) and exposure data are needed. (Obviously, reliable existing data should always be used in lieu of new testing.)

In lieu of an initial peer consultation, we believe industry should review the existing literature, develop robust summaries of key existing studies, and prepare a test plan for the endpoints that currently lack adequate existing studies. That robust study summary and test plan should be made publicly available via the Internet for a 120-day comment period. This process is identical to the one used for the High Production Volume testing initiative. Rather than inviting endless debates about which endpoints should be evaluated for a particular chemical, it focuses attention on whether existing studies for the relevant endpoints have been identified, and if so whether they are scientifically adequate so that further testing for that endpoint is not needed.

We support an opportunity for “exceptional circumstance” arguments for the derogation of tests. Such a process has been established in the High Production Volume testing program. But the presumption should be clearly stated that these tests will be conducted, absent exceptional circumstances that cannot now be foreseen. (Although the agency's states its belief that “the VCCEP is likely to conclude that a full toxicity and exposure evaluation will be needed for each chemical,” that statement is contradicted elsewhere in the document.) If such exceptional circumstances arise, they should be brought forward for consideration at that time.

Finally, the “peer consultation” would consider only one question: whether a particular chemical should undergo a cancer bioassay where it has not yet done so. (This assumes that it is determined that there is, in fact, an adequate scientific basis for making carcinogenicity a triggered test – which we do not regard as a settled question.)

5. Multiple commitment steps are unnecessary and potentially misleading to the public.

The agency's proposal to allow chemical sponsors to partially commit to participating in the program runs counter to the definition of “commitment.” A sponsor's commitment to participate in this program should mean taking responsibility for filling all of the necessary data gaps.

A step-wise commitment does not make sense in light of industry's professed desire (as stated in their Responsible Care Guidelines) “to provide information on health or environmental risks and pursue protective measures for employees, the public, and other stakeholders” and “to support education and research on the health, safety and environmental effects of our products and processes.”

Moreover, it is very unlikely that the public will understand the nuances of a multi-step commitment process. If industry wishes public recognition for making commitments to assure the availability of the data needed to evaluate the safety of a particular chemical to which children are exposed, then they should commit to making all of those data available.