Over the last few months, I was heartened to hear a number of industry stakeholders in the debate over TSCA reform embrace the idea of designating in TSCA reform legislation a "jump-start" or "quick-start" list of chemicals of high concern or priority.  The idea was to allow EPA to hit the ground running, by having an agreed-to list of chemicals on which it could immediately initiate action.  Well, it now appears many in industry actually have something far slower and far more cumbersome in mind.

The Subcommittee on Commerce, Trade and Consumer Protection of the U.S House of Representatives' Energy and Commerce Committee held a hearing yesterday on the question of "prioritization":  How a new law could best spur prompt identification of and action on the chemicals of highest concern.  Mr. Bill Greggs testified on behalf of three trade associations prominent in the debate over TSCA reform:  The Consumer Specialty Products Association, the Soap and Detergent Association, and the Grocery Manufacturers of America.

In Mr. Greggs' testimony and answers to questions from subcommittee members, what industry has in mind when it talks about a quick start became much clearer:

• There would be no list of chemicals in the bill.  Nor would there even be a process laid out by which EPA would identify chemicals of highest concern for the purpose of promptly taking action to reduce exposure to them.
• Instead, the bill would merely contain some potential hazard and exposure triggers for EPA to consider. 
• The legislation would then impose a risk-finding requirement on EPA (sound familiar?), by restricting any further effort to only those chemicals that are found – on the basis of existing information, however incomplete – to pose BOTH the highest hazard and the highest exposure.
• Once the legislation passed, EPA would first have to develop specific high-hazard and high-exposure criteria, and then apply the criteria through a review of all chemicals it knows to be in commerce and for which such hazard and exposure data exist.
• Then EPA would have to provide industry with a robust process by which industry could – and surely would – challenge each and every designation of such a chemical by EPA.
• Assuming any chemical made it through that process, such highest-of the-high-priority chemicals would then merely enter a further limbo:  a review and risk assessment process as a prelude to a "safe use" determination by EPA.

Hardly the quick start I had hoped for. 

If the last decade of voluntary programs and ChAMP has taught us anything, it's that it's time to move beyond the endless cycle of having one round of assessment to identify chemicals of highest concern lead to nothing more than a further round of assessment.

The re-energized new leadership at EPA recently took upon itself to identify an initial handful of bad-actor chemicals, for which it is developing action plans to reduce the harm those chemicals are causing.  EPA didn't ask for industry's blessing before listing those chemicals.  Can you imagine what would have transpired had it done so?

Nor should EPA have to provide an opportunity for industry to challenge such listing decisions:  Under Greggs' scheme, any listing of chemicals is pre-regulatory –  indeed, it's arguably pre-pre-regulatory or even pre-pre-pre-regulatory.  The notion that EPA's mere identification of a chemical through such a process should be eligible for challenge by the very companies that make or use the chemical is preposterous – not to mention a recipe for endless delay and fights.

And hey:  Wasn't avoiding all that the motivation behind having a quick-start list in the first place?

Lest you think I'm reading too much into a 5-minute statement at a hearing, I'll point you to a detailed proposal these same three trade associations (joined by a dozen others) gave to California earlier this year, as a "model" for regulations under the state's Green Chemistry Initiative.  By my count, that proposal would provide no fewer than six – 6 – opportunities for industry input into and challenge of the state agency's process for designating and prioritizing a chemical of concern.

Talk about greedy:  In addition to requiring formal notice and comment on any proposed listing of chemicals, that proposal would require another round of notice and comment on the "final" list.  And it would repeat that dual-intervention opportunity at each of several steps in a process that would precede any regulatory action to restrict the use of the chemical.

That makes the federal rulemaking process under the Administrative Procedures Act look like a cakewalk in comparison, even after factoring in the interventions by the Office of Management and Budget called for under Executive Orders and such.

A prominent refrain in the industry's new TSCA hymn book has been to say that EPA needs more authority.  I, for one, would like to hear that tune sung more than just in Sunday choir and have it enter the actual work week.

When the chemical industry talks about prioritization – a central question in the debate over TSCA reform – more often than not it quickly reduces the question down to the argument that we should focus only on those chemicals, however hazardous or untested they may be, to which we know people are exposed.  In a perfect world, that might suffice.  But, as this post will explore, the world of exposure assessment is anything but perfect. 

While both hazard and exposure are clearly relevant in determining chemical risks, there are critical differences between our ability to assess hazard and exposure that have implications for the development and application of chemicals policies.  And real-world experience in chemical assessment programs that have attempted to rely on exposure information to prioritize chemicals also offers lessons for exposure assessment.

Critical differences between assessing hazard and exposure

Approaches that seek to rely on exposure to prioritize or assess chemicals need to acknowledge and account for a number of critical differences between the nature of hazard and exposure information and their relative extent of availability.  Certain characteristics of exposure information pose serious challenges to sound decision-making.  Here are several reasons why.

1.  Hazard is largely inherent to a chemical and (aside from degradation or metabolism) doesn’t fundamentally change over space or time, whereas any exposure information necessarily represents only a “snapshot” in both space and time.

A chemical’s hazard is an intrinsic property, one that is directly related to the chemical's composition.  While manifestation of a hazard can of course vary by route of exposure, it otherwise exists largely independent of how the chemical is used, where or how it enters the environment, or other factors that vary with time and place.  Hazard data are therefore relevant and needed regardless of how the chemical is used.  That is, such data are useful in understanding any actual or potential use or release of a chemical – and in deciding what kind of exposure-reducing efforts may need to be taken.

Just the opposite is true for exposure, which can change dramatically depending on how a chemical is produced, used, transported and discarded or released.  The consequences of exposure depend on who or what might be exposed, and the level, frequency and duration of the exposure. 

Conditions that determine exposure can and often do differ enormously for every setting and point in time that a chemical is present.  Basic physiological differences (including those associated with age and life stage) as well as cultural factors (e.g., extent of fish consumption) and activity patterns also amplify the variability in exposure to a chemical.  And even if a “snapshot” of current exposure is able to be assembled, the next new use or activity leading to a release alters the exposure picture.

The highly variable nature of exposure poses a major challenge to exposure (and risk) assessment:  It means that exposure assessment must be an ongoing activity, with the scope and frequency of its measurement sufficient to characterize the variation (spatial and temporal) in, as well as magnitude of, exposure. 

That's but one reason why exposure assessment is often called the "weakest link" in risk assessment.

2.  Mechanisms for generating and collecting exposure information are undeveloped relative to those for hazard information.

An extensive body of methods developed through international consensus specifies how to test a chemical for most hazardous properties.  And while new hazard concerns and new test methods emerge over time and must be incorporated, the infrastructure for doing so is largely in place:  detailed government-sanctioned procedures, guidelines, criteria and standards have been specified for conducting hazard tests, for assuring the quality and reliability of the results, and for determining whether the results constitute evidence of a particular hazard.  Moreover, these measures allow that results are reproducible and can be independently verified.

In contrast, virtually none of these mechanisms are in place or established to assure that exposure information is complete and accurate.  Debates over what constitutes adequate exposure assessment and how to address the “moving target” nature of such information are far from resolved.  Government-sanctioned procedures for generating, evaluating the adequacy of and interpreting exposure data have yet to be developed or validated, including testing and measurement standards, guidance, methods and tools.

Use and exposure information is rarely systematically collected and even less often made public in any useful form.  For the first time, beginning in 2006, the Environmental Protection Agency (EPA) began to require the reporting of basic information relevant to understanding uses of and exposure to chemicals.  But that program is fraught with limitations, as I've described in detail in earlier posts here and here.

Because of these and other factors, estimates of chemical exposures typically depend on severely limited sources of information.  In many instances, little or no information is publicly available on how chemicals are manufactured, processed, used, or discarded; the numbers of workers and consumers exposed; quantities released to the environment; how chemicals are distributed and transformed in the environment; and other parameters necessary for estimating exposures. 

3.  Assumptions and modeling used to characterize exposure are notoriously inaccurate.

As a result of the information constraints just discussed, exposure-driven prioritization initiatives and exposure modeling exercises may well utilize incorrect exposure assumptions.  Biomonitoring has revealed such instances by “ground-truthing” exposure assumptions – providing objective, incontrovertible evidence of the extent to which chemicals end up in people.  Recent biomonitoring data on both phthalates and poly-brominated diphenyl ethers amply illustrate this point. 

Phthalates are very widely used in products ranging from plastics to cosmetics and other personal care products.  They exhibit a range of toxicity, including to the liver, kidney, and male reproductive system.  The first CDC National Report demonstrated surprisingly high levels of di-butyl phthalate (DBP) and di-ethyl phthalate (DEP) in U.S. residents in general, and for DBP, in women of child-bearing age in particular (see the first two letters here).  Indeed, these data demonstrated high-end levels of DBP that were an order of magnitude higher than a prior estimate that had been developed based on industry-provided use data and expert judgment.

Polybrominated diphenyl ethers (PBDEs) are widely used flame retardants.  Different species of PBDEs are used in products ranging from plastics (such as computer cases) to upholstery foam.  Toxicological studies indicate that they can disrupt thyroid metabolism and may have effects on other organs, including the liver.  Because PBDEs are not very volatile or water soluble, they were assumed to more or less stay in place in products, and were not believed to have a high potential for exposure.  However, biomonitoring studies from around the world have demonstrated that levels of PBDEs in peoples’ bodies have been dramatically increasing over the past two decades, with the highest levels currently reported in the United States; see, for example, this paper).

4.  Differential access to both exposure data and the means to generate them severely limit the “reproducibility” of such data.

In addition to the variability and absence of agreed-upon procedures noted above, other factors limit “reproducibility,” that is, the ability to readily and independently measure or verify exposure data.  Most exposure data and the means to generate them reside virtually exclusively with industry.  It simply must be acknowledged that industry has a strong interest in maintaining that exposure to its chemicals is low, so the ability to independently measure and verify exposure data is critical.  Yet physical access to many exposure “settings” (e.g., workplaces) is very limited and infrequent at best, even for government officials. 

Broader access to exposure-relevant information is even more restricted:  Wide latitude is typically provided to claim chemical use and exposure information as CBI, preventing even its review outside government; this situation is often in contrast to that applying to hazard data, which is more likely to be deemed ineligible from designation as CBI. 

Finally, even chemical manufacturers have incomplete access to and information on their customers and how their chemicals are used.  Intermediaries (vendors, brokers, distributors) are a formidable information flow bottleneck, as is the often-proprietary nature of information concerning downstream use and competition among suppliers.  These factors serve to impede information-sharing even within supply chains, which in turn affects the extent and accuracy of exposure-relevant information that any one entity in a supply chain can provide if asked or required to do so; see this paper and Modules 1 and 2 of this report for more discussion of these limitations.

Conclusion

For all of these reasons, we cannot continue to rely on assumptions about chemical exposure or on the industry-supplied or otherwise-limited exposure information that is currently available.   That is simply too uncertain and unreliable a basis on which to set aside hazardous or untested chemicals as low-priority, or to decide for which chemicals hazard data should be developed. 

Of course, where affirmative evidence indicates exposure to a chemical is occurring, for example as revealed through biomonitoring, this should clearly suffice to prioritize such a chemical at a minimum for further testing, assessment or control.  But the converse cannot be said:  that in the absence of reliable information about exposure we can simply drop a chemical from further consideration.

_____

There is actually considerable real-world experience that documents the adverse consequences of an over-reliance on exposure to prioritize chemicals.  For more detail, see our analyses of:

• EPA's recently replaced ChAMP Initiative;
• Canada's Chemical Categorization; and
• the Existing Chemicals Program of the Organization for Economic Cooperation and Development (OECD) – see the paper included as pages 109-114 of this OECD report, which I wrote several years ago and from which this post is derived.

A new entry showed up sometime in the last day on EPA's webpage for its ChAMP initiative.  It reads:  "The Chemical Assessment and Management Program (ChAMP) has been superseded by the comprehensive approach to enhancing the Agency’s current chemicals management program announced by Administrator Lisa Jackson on September 29, 2009." 

Don't miss this bit at the top of the page:

Yes, that image is a cobweb, which EPA uses to designate archived web content.  What's happening here?

Lost in the buzz surrounding EPA Administrator Jackson's speech on Tuesday unveiling EPA's new TSCA reform principles was the fact that EPA simultaneously announced a rather sweeping set of changes to its current existing chemicals work aimed at "Enhancing EPA’s Chemical Management Program."

Now, as regular readers of this blog know, EDF has for some time been raising serious concerns with ChAMP.  To review:

• We noted that, about 18 months ago, EPA had made an abrupt shift from developing hazard characterizations based on the hazard data sets submitted under the HPV Challenge, to cranking out "risk-based prioritizations" that were frequently designating high-hazard chemicals as being low-risk, based on poor and incomplete use and exposure information. 
• We also said EPA was, in the process, obscuring the significance of the data gaps and data quality problems remaining even in supposedly "final" industry data submissions. 
• Finally, we lamented the fact that the only "action" EPA was proposing to take even for the high-concern chemicals it identified under ChAMP was yet more testing and assessment – never getting to any risk mitigation.

So it 's quite heartening to see that EPA's enhancements directly address all of these problems:

• EPA has returned the focus of its assessment activities to developing hazard characterizations for HPV chemicals, and just posted 100 of them in September.
• EPA is to publish TSCA Section 4 test rules to tackle HPV Challenge chemical data gaps, not only for unsponsored "orphan" chemicals, but for "sponsored but unfulfilled chemicals."
• In 2010, EPA will propose a significant revamping of its Inventory Update Reporting (IUR) rule to "make the reporting of chemical use information more transparent, more current, more useful, and more useable by the public."
• And last but certainly not least, EPA will be taking on a slew of new "regulatory risk management actions" for old enemies like lead and mercury, and developing "chemical action plans" for some of the more recent arrivals on the toxic chemicals scene.  The initial list contains some bold entries:
  • Benzidine dyes and pigments
  • Bisphenol A (BPA)
  • Penta, octa, and decabromodiphenyl ethers (PBDEs) in products
  • Perfluorinated chemicals
  • Phthalates
  • Short-chain chlorinated paraffins

EPA says it "intends to utilize the full array of regulatory tools under TSCA to address risks, including authority to label, restrict, or ban chemicals under Section 6 of TSCA."

Now, I can just hear you saying, "But wait, how will this EPA overcome all of TSCA's hurdles, which you've prattled on about ad nauseum, Richard?"

Well, let me say two things to that.  First, bless their hearts for being willing to try.  Second, in addition to the inherent burdens imposed by TSCA, EPA has been stymied over the years by two other forces: 

• a massively burdensome executive branch process for regulatory development, and
• a virtual guarantee that industry will challenge EPA's regulations at every opportunity.

So, this new EPA effort will be an interesting test, both for the new Administration's regulatory review process and for industry's new-found religion under which it has acknowledged that EPA needs to be able to exercise greater authority under TSCA.

One image immediately comes to mind; click here.

I can also hear you saying, "Gee, if EPA manages to pull this off, what's the need for TSCA reform?"

Again, two responses.  First, I'm in favor of any action EPA can take to reduce chemical risk, and the sooner the better, especially given it'll be some time before:  a) TSCA reform is adopted, and b) TSCA reform is implemented.

Second, as welcome as all this new stuff is, it's hardly the comprehensive approach that's needed.  That looks much more like this and this.

It probably goes without saying that EDF welcomes EPA's decision to suspend the development and posting of risk-based prioritizations under its Chemical Assessment and Mangement Program (ChAMP).  EDF has been arguing (see our earlier posts) that ChAMP's "rush to risk" has taken EPA badly off-track.  But we have also identified many useful things that EPA's existing chemicals program can and should be doing with the data it obtained through the HPV Challenge (whether called ChAMP or not) .

We look forward to working with EPA to craft a new approach, grounded in a return to developing scientifically defensible hazard, not risk, characterizations and transparently identifying and addressing data gaps and data quality problems.

Many of the screening-level hazard data being collected and analyzed under ChAMP that pertain to human health are derived from traditional laboratory animal studies.  The National Academy of Sciences (NAS) recently offered a "new paradigm for toxicity testing" in its 2008 report Toxicity Testing in the 21^st Century: a Vision and a Strategy.  Can ChAMP hazard data be used to facilitate the development of new testing strategies? 

First, more about the new toxicity testing paradigm envisioned by NAS:  Instead of exposing whole animals to chemicals to examine their effects, which takes a lot of time, costs a lot of money and may not always be a good model for how humans can be affected, might we use human cells or tissues (e.g., liver, kidney) to determine how chemicals act biologically?  Where chemicals interact with the cells and alter (e.g., inhibit or over-stimulate) critical biochemical pathways, such perturbations may well represent early events or signals that can eventually lead to an adverse effect.

ToxCast

There is an EPA initiative called ToxCast that is exploring the use of high-throughput screening (HTS) test methods to predict hazard, characterize so-called "toxicity pathways," and prioritize among large numbers of chemicals.  According to the ToxCast website:  "In its first phase, ToxCast^TM is profiling over 300 well-characterized chemicals (primarily pesticides) in over 400 HTS endpoints."  The endpoints include tests that measure DNA and protein synthesis activities, multi-cell interactions and developmental assays in zebrafish.  These methods are among those being developed to realize the vision described in NAS' report. 

The results of the first set of ToxCast assays are still being validated – meaning that they are being analyzed to determine how similar the outcomes are to those of standard laboratory animal tests.  This is one important step in determining whether the ToxCast testing array can actually substitute for some standard laboratory animal testing.

But we would suggest there are already two timely and useful applications of the data that ToxCast can generate in aiding EPA's effort to evaluate hazard information under ChAMP.

Supporting alternatives analysis

The first idea would be to compare the relative capacity for chemicals in the same functional use class (e.g., solvent, chelator, fragrance) or chemical category (e.g., fatty nitrogen derived cationics) to effect biological changes detectable by the ToxCast HTS methods.  In doing so, we would be testing the hypothesis that chemicals that demonstrate less capacity for effecting changes in the HTS tests are more likely to have lower hazard profiles. 

As we described in a previous post, alternatives analysis is a tool that compares hazard assessments of chemicals that are used to perform similar functions, in order to identify those that have a lower hazard profile.  Using ChAMP hazard data and characterizations for this purpose would start by comparing the results of laboratory animal study data.  By running some of the ChAMP chemicals through the ToxCast HTS system and comparing the results of the HTS tests with the ChAMP hazard data, EPA could both support the identification of safer chemicals and test how well the HTS methods predict in vivo effects – a true win-win.

Validating (or refuting) chemical categories

About 80% of the chemicals sponsored under EPA's HPV Challenge Program are members of proposed chemical categories.  Within these categories, sponsors and EPA propose that hazard data for tested category members can be "read across" to untested members, as an alternative to direct testing of each chemical.

Grouping chemicals into a category starts with an hypothesis that distinct chemicals that show similarity or regularity in their physical-chemical properties and chemical structures actually possess similar or predictably regular patterns of biological activity.  Establishing a valid category requires that the hypothesis actually be demonstrated to be true, once the available data on physical-chemical properties, environmental fate and toxicity/ecotoxicity for the proposed category members are assembled.  Then, some degree of read-across among category members can be justified.

However, as our comments on the several ChAMP assessments we've analyzed in recent posts make clear (see here, here and here), both industry sponsors under the Challenge and EPA under ChAMP appear frequently to be over-relying on category approaches, lumping together chemicals that are insufficiently related or similar to warrant read-across.

HTS data could play a very useful role, therefore, in helping either to demonstrate – or to negate – an hypothesis that members of a proposed category of chemicals actually exhibit similar biological activity.  If category members in fact exhibit similar or regular patterns of activity across an array of different cellular and subcellular assays, then the case for grouping them for purposes of read-across would be much stronger than a case based solely on similarity or regularity in their physical-chemical properties and chemical structures.  Alternatively, if HTS data do not show such patterns, the category should not be utilized and either it should be broken up into smaller groupings or its putative members should be tested individually.

Both of these uses of ChAMP hazard data and characterizations side-by-side with HTS data developed under ToxCast – in supporting alternatives analysis and in category validation – would offer the added benefit of helping to advance the transition to the "new paradigm for toxicity testing."

This new post serves as a response to Charlie Auer's most recent comment responding to our critique of ChAMP.  (To see the whole exchange, start here, then go here, here and here.)  So far, this exchange has focused mainly on our disagreement over whether or not EPA is somehow required to do risk assessments under ChAMP.  At some point, I hope Charlie and others will engage on the substance of our critique – the serious concerns we've raised about the quality and validity of the ChAMP assessments.
 

Risk-based at all costs?

In Charlie's latest comment, his main interest seems to be to lock in EPA's current approach under ChAMP.  He steadfastly maintains that EPA now has no choice but to turn out risk decisions for HPV chemicals under ChAMP, given its 2007 commitment under the Security and Prosperity Partnership (SPP) signed by President Bush. 

For all the reasons I've stated earlier, I strongly dispute that conclusion.  Indeed, it is striking how the chemical section of the SPP agreement – including the very language Charlie cites ("the three countries [are] to enhance appropriate coordination in areas including testing, research, information gathering, assessment, and risk management actions") – seems to go out of its way to avoid linking the words "risk" and "assessment."  That term never appears.

I don't dispute, as Charlie says, that "US and Canadian [regulatory] actions are risk-based."  Indeed, EDF supports the view that, in most cases, chemical regulatory actions ought to be risk-based, or certainly ought to take into account both hazard and exposure.

Now, if ChAMP were actually proposing any actions to mitigate risk, e.g., to restrict production or use of any HPV chemicals through regulation, Charlie's argument might be germane.  But as we've noted, ChAMP's not doing that, even for the few chemicals it has designated to be of high risk and high priority.

So the fact that EPA will only consider regulating a chemical based on risk in no way requires that EPA must do risk assessments under ChAMP – regardless of whether the data available to the agency are sufficient to support such assessments. 

Indeed, the SPP agreement was signed well before the use and exposure information EPA has collected under the Inventory Update Rule (IUR) were made public and its major deficiencies were realized.  I disagree with Charlie that the mere fact that EPA collected such data compels it to use them regardless of their adequacy.

As we noted in our last two posts on ChAMP, in cases where hazard concerns warrant action, there are lots of steps the agency could be taking to manage or reduce potential risks – without having to wait for the emergence of use and exposure data of sufficient quality to carry out a risk assessment.  EPA itself has taken such actions occasionally, for example in launching the 2010/2015 PFOA Stewardship Program.

Sacrificing the public's chemical right-to-know on the altar of risk

Add to this the fact that EPA's insistence on pursuing the "rush to risk" under ChAMP is delaying the public's right-to-know about the hazards of HPV chemicals, and I have to strongly question the wisdom of EPA's devoting most of its limited chemical program resources to doggedly pursuing this course.

I also am disturbed by Charlie's conclusion that his and EPA's interpretation of the SPP commitments "are superior to and overtake informal EPA commitments made to pursue the NPPTAC recommendation."   Recall that NPPTAC recommended, and EPA agreed, that the agency issue hazard characterizations and identify remaining data gaps for some 1,200 HPV chemicals by the end of 2009, which it's nowhere closing to meeting.

What NPPTAC called for did not come out of thin air:  It reflected the core purpose of the HPV Challenge, which was in turn the cornerstone of EPA's much-touted chemical right-to-know initiative:  It would be tragic indeed if EPA chooses to negate its public right-to-know commitments by saying they've been trumped by a political decision made by the last administration.

Under ChAMP, the minimum data set has become wholly optional

Charlie cites OECD guidance calling for integrating exposure considerations into assessment activities.  We have no quarrel with that, as long as the exposure information is sufficient; but as our examples amply demonstrate, that is not the case with the IUR data.

Nor does consideration of exposure in any way obviate the need for a complete Screening Information Data Set (SIDS).  International experts convened by the OECD, including from the U.S., devised the SIDS to be the minimum amount of hazard data needed on which to base a screening-level assessment.  EPA itself has repeatedly acknowledged this fact; see, for example, p. 15 of this EPA report on the HPV Challenge.  Outside of some very limited derogations (e.g., for closed-system intermediates) from certain endpoints, full SIDS data sets are to be provided – wholly independent of exposure considerations.

The need for such complete screening data sets is all the more critical given that EPA is intent on making low-priority designations that preclude, for the foreseeable future, any further data development.  Yet under ChAMP, EPA is routinely making such designations for chemicals that lack the basic SIDS data set.

Acknowledging the limitations

It's very helpful and refreshing that Charlie, as former head of EPA's toxics office, acknowledges that EPA faces high hurdles in trying to mandate testing through issuing TSCA test rules.  That's what many of us have been saying for years, and it's among the strongest arguments for why we so badly need fundamental reform of TSCA.

But even without issuing test rules, EPA could help this situation enormously by forthrightly and transparently identifying and acknowledging the significant gaps and deficiencies in the quality of data remaining for HPV chemicals sponsored under the Challenge.  Instead, Charlie and EPA are papering over the problems through their rhetorical sleight-of-hand of substituting "data gap" with "data need."

Interestingly, Charlie argues that instead of our pressing EPA to pursue test rules, we should just wait for the European Union's REACH Regulation to provide most of the data we need.  As tempting as that sounds, some notes of caution are warranted.  First, for all tests involving laboratory animals, REACH requires only that test proposals, not test results, be submitted at the time of registration.  So even in the best of circumstances, much of the data we need for HPV chemicals likely won't be available for some time after the registration deadline of December 2010.  Second, as I've discussed at length elsewhere, it remains to be seen just how much testing will be required under REACH.

I also appreciate that Charlie has acknowledged the need for major changes in EPA's Inventory Update Rule, given the hugely disappointing results of the latest data collection effort.  I agree with many of his thoughtful proposals for needed fixes.  I'd add another:  Require annual reporting of production and use data.  The current reporting requirement is limited to only a single year once every five years – yielding a highly distorted view of actual chemical production (for more on this, see p. 38 of this paper).

But given the rate at which ChAMP is churning out risk decisions using the current IUR data, I must heartily disagree with his notion that we should set aside concerns about those data and only focus on making improvements for the next IUR collection cycle.  That collection won't even begin until 2011, and if this last round is any guide, data won't be released until 2013 or 2014.  On EPA's current trajectory, ChAMP will have already cranked through all of the HPV chemicals by then.

Bottom line:  ChAMP embodies a blatant double standard

Charlie's responses to date (and the lack of any response from current EPA staff) suggest an unwillingness to engage on the substance of our critique of EPA's ChAMP assessments – its systematic reliance on incomplete and/or poor-quality data, a problem further compounded not only by EPA's refusal to reject such data and identify them as gaps, but its use of such flawed data as a basis to claim no or low toxicity, and thence to draw unwarranted low-risk conclusions about hundreds of HPV chemicals.

All of this brings me to the crux of the problem:  Under ChAMP, EPA appears more than willing to rely on limited hazard, use and exposure data to effectively exonerate scores of chemicals, by relegating them to the low-priority dustbin.  Yet in the few cases where EPA has had to conclude, using the same quality of data, that a chemical poses a high level of concern, no action to control such chemicals is even proposed.

Charlie's characterization of such high-priority chemicals as being only "putative positives" is quite telling.  It's a clear illustration of just how high a bar EPA has set to initiate action to reduce use of or exposure to (or even mandate testing of) a problematic chemical – in contrast to the remarkably low bar it must clear to dismiss a chemical altogether.

That's a double standard I just can't abide.

In our critique of EPA's Chemical Assessment and Management Program (ChAMP), we have pointed out that, despite its limitations, there is value in the hazard data that EPA is collecting and analyzing.  How so?

First, we recommend that the hazard profile of the chemical be the main factor in determining a chemical's priority.  One of the reasons for this is the hazard properties are intrinsic to the chemical – they basically do not change even when the chemical is used in different ways.  Given the difficulty in knowing or anticipating all of the different applications for a chemical, the hazard data and characterization are extremely useful to have in hand for current and potential users of the chemical.

While EPA has yet to even consider the possibility, the screening-level hazard data and hazard characterizations it is developing can also help inform safer substitution.  Each ingredient in a product formulation has a specific function, for example, to reduce surface tension (surfactants), dissolve materials (solvents), reduce water hardness (chelating agents) or provide or mask a scent (fragrances).   The functional class is typically related to chemical and physical properties.

By comparing the hazard profiles of chemicals within a functional class, the chemicals with lower hazard profiles can be preferentially selected.  This is the basic approach taken by EPA's own Design for the Environment Program, which works with companies to screen product ingredients, identify chemicals of concern and identify and promote use of safer substitutes.

So how could ChAMP be used to leverage green chemistry and facilitate safer substitution?  By tagging each chemical evaluated under ChAMP with its functional class and identifying its applications (to the extent they are known).  It should then be feasible to group the ChAMP chemicals by functional class to facilitate comparison of their hazard profiles.  This information could be invaluable to companies interested in selecting safer chemical substitutes to perform a given function.

Functional class should not be confused with chemical class.  For example, within the solvent functional class, there are chemicals from many different chemical classes, including:  alcohols, esters, ethylene glycol ethers (EGEs) and propylene glycol ethers (PGEs).  While chemicals within a given chemical class may often have similar hazard properties, there will often be significant differences in hazard between the chemical classes that serve a given function.

This may not always be the case, of course.  For some functional classes and some endpoints, there may be very little difference in hazard among members because of the function that they perform.  For example, surfactants typically demonstrate aquatic toxicity because they are, by definition, surface-active.  In such cases, the entire functional class – or the underlying functionality – could be the focus of a "green chemistry challenge program" that would fund research into or otherwise seek to spur development of safer alternatives.

What would be needed to "green" ChAMP along these lines?  The following would be a start:

1. Convert the hazard data collected and analyzed in ChAMP into a sortable, searchable spreadsheet or other database that allows side-by-side comparisons.
2. Identify known applications of each ChAMP chemical.
3. Add functional class tags to each chemical.

For item # 1, tables summarizing the hazard data are already included in the hazard characterizations supporting each chemical or chemical category evaluated.  These tables simply need to be transferred into a searchable, online database.

For item #2, applications can be gleaned from a variety of sources, including information submitted by industry under the Inventory Update Rule (IUR), the Hazardous Substances Data Bank and the National Library of Medicine's Household Products Database.  The information available via these resources is often extremely limited and out-of-date.  Over time, we can expect our understanding of how chemicals are used to get better, for example as a result of data submitted under the European Union's REACH Regulation and hopefully, once the Toxic Substances Control Act is reformed.

(As is clear from our earlier posts on ChAMP, EDF has been very critical of EPA's heavy reliance on incomplete, unverified and confidential IUR information to draw conclusions about exposure and risk of HPV chemicals.  What I'm proposing here is a far more appropriate use of the IUR data, one much less sensitive to its deficiencies:  to glean any available functional use information solely for the purpose of identifying chemicals sharing a function.)

For item #3, a field can be created in the online database for functional class information. As with information on applications, functional class designations can also be difficult to obtain, but are likely to become more available in the future.

One source of both functional class and application information to consider is a subscription-based web service such as Chemidex, which compiles this type of information to facilitate information exchange among formulators.  The advantage of accessing this type of informational database is that the information comes directly from companies making products, rather than chemical manufacturers.  Still, challenges remain; for example, the listed products include mixtures, some with proprietary formulations.

Another model for this type of database is CleanGredients, which also provides information for formulators on general product and regulatory information, physical-chemical properties and human and environmental health data.

There are three needed components to forge this linkage:

Hazard Data + Functional Use + Application

Efforts like those of Chemidex and CleanGredients are beginning to provide tools able to facilitate functional use approaches to comparison and selection among chemical (and non-chemical) alternatives, albeit on a small scale and within limited categories.  With the hazard data it's collected and is now assessing under ChAMP, EPA could begin to do so on a larger scale.

In conclusion, one means of "greening" ChAMP would be to link the hazard data EPA is evaluating with meaningful functional use information.  Such an approach can start to challenge the assumption that continued production and use of high-production-volume chemicals with undesirable hazard traits is the only option.

As I noted in our first post on ChAMP, after getting off to a strong start in 2007, EPA's abrupt decision in 2008 to steer ChAMP in the direction of cranking out hasty risk decisions was entirely its own.  Can ChAMP be put back on track? 

Neither EPA's own advisory committee – which had many members from the chemical industry as well as a few environmental NGOs – nor the commitment made by the U.S. under the North American Security and Prosperity Partnership (SPP) agreement signed with Canada and Mexico in August 2007, called for such risk assessments or risk-based prioritizations. 

The commitment EPA made under the SPP agreement was "to assess and initiate needed action on the over 9,000 existing chemicals produced above 25,000 lbs/yr in the United States."  There was no mention of risk assessment anywhere in the agreement.  Indeed, at that time, EPA had already begun to publish hazard characterizations, which, as assessments of the hazards of high production volume (HPV) chemicals – fully meet the terms of its SPP commitment.

EPA's focus on developing HPV chemical hazard characterizations was also precisely what it had committed to do when it adopted, in November 2005, the recommendations of its own National Pollution Prevention & Toxics Advisory Committee (NPPTAC).  Specifically, it agreed to complete and make public such hazard characterizations on all HPV chemicals for which it had received final data sets from the HPV Challenge – some 1,200 chemicals – within four years.

How close is EPA to meeting that commitment, with only six months to go?  As of today, EPA has issued only about 130 hazard characterizations covering about 315 HPV chemicals (many of the chemicals are in categories).

So EPA's rush to risk is not only resulting in flawed risk decisions that are prematurely exonerating hundreds of chemicals, as described in our earlier posts.  It is also causing EPA to renege on its promise to characterize the hazards of all sponsored HPV chemicals in a timely manner – and to clearly identify gaps in the quality and completeness of the data received under its voluntary program.

In short, EPA is compromising the public's right to know about the hazards of the most widely used chemicals in U.S. commerce.  That's quite ironic when you consider EPA bills the HPV Challenge and ChAMP as cornerstones of its Chemical Right to Know Initiative.

What should EPA be doing under ChAMP?

EDF is the only environmental NGO that has been willing to engage with EPA, first on the HPV Challenge and more recently on ChAMP.  We served on advisory panels, participated in public meetings, carefully peer-reviewed draft assessments, filed comments and met with EPA staff repeatedly. 

Even now, we are the only voice in our community arguing that the HPV data and ChAMP could provide value if it got back on track.

We have pointed out that many of the shortcomings of the HPV and ChAMP initiatives are due to EPA's limited authority under the Toxic Substances Control Act (TSCA), and we are actively working toward fundamental reform of TSCA to give EPA the tools it needs to do a better job.

But we also believe that in the interim, and despite EPA's constraints under TSCA, EPA can and should be doing a more credible and valuable job under ChAMP. 

EPA should:

• Return the focus of ChAMP to completing high-quality hazard characterizations for HPV chemicals, rather than rushing to issue highly suspect risk characterizations based on flawed assumptions and poor-quality use and exposure information.
• For each chemical assessed, clearly identify and communicate to the public all gaps or quality concerns in available data.
• Stop assigning low-priority rankings to chemicals, especially those with data gaps in the most basic, minimum set of screening-level hazard data. As we said before, it's one thing for EPA to identify as high-hazard those chemicals where, despite the data gaps, available data demonstrate high toxicity. It's quite another for EPA to effectively exonerate chemicals as low-hazard or low-priority when not even a bare-minimum data set is available for them.
• Adopt a health-protective approach to hazard screening: Where data are uncertain, of questionable quality or equivocal, assume a hazard exists until and unless a chemical's manufacturer provides the data to show otherwise.
• Issue test rules to require that gaps in submitted data sets be filled.
• Significantly accelerate issuance of test rules for all of the unsponsored HPV Challenge chemicals, and for the hundreds of newly emerged HPV chemicals that have never been sponsored.
• For the ~4,000 medium-production volume (MPV) chemicals – which are also covered by the SPP commitment, and for which even less data are available than for HPV chemicals – focus on identifying and acting to address data needs to support more robust hazard assessments, rather than pursuing its current approach of using scientifically unsupportable extrapolations of data among loosely defined "clusters" of such chemicals.

For chemicals for which significant hazards are identified, here are some of the good recommendations that NPPTAC offered as steps EPA could take even under its current TSCA authorities, which EPA has yet to act upon (under each, I've added a few elaborations and additional thoughts of my own in brackets):

• Gather additional information on uses (e.g., by use function, category, release potential, or benefit) and exposure (to humans and/or the environment).
  • [Listing such substances on the Toxics Release Inventory would be a great way to get direct exposure data. Efforts to get better use and exposure information - well beyond the data EPA got from manufacturers under its Inventory Update Rule (IUR) - could extend to downstream users of chemicals. Information on functional use would help in identifying safer substitutes within the same functional class.]
• Gather additional information on hazards to support a more in-depth characterization.
  • [Our reviews of EPA's risk decisions have identified many cases where available studies raise more questions than they answer, or identify additional concerns. EPA should require further testing in such cases.]
• Identify existing risk management programs and practices. Evaluate existing Federal and State regulatory controls (e.g., occupational exposure limits).
  • [Instead, in many cases EPA merely asserts the existence or effectiveness of such practices, and seems particularly averse to even suggesting that some risk management might be needed.]
• Provide information referrals or recommendations for actions to other EPA program offices or other Federal or State agencies.
  • [These could include referrals to OSHA requesting action on chemicals posing high worker hazards, proposing that the Centers for Disease Control or the U.S. Geological Survey add such chemicals to their biomonitoring and surface water monitoring programs, referring chemicals to its own Design for Environment (DfE) program to assess the availability and safety of available alternatives to hazardous chemicals, and referring such chemicals to the Food and Drug Administration where they have uses or are in consumer goods that fall under its jurisdiction.]

That's more than enough to keep the EPA toxics office busy for the foreseeable future.

This example raises some new issues as well as some we discussed in the earlier examples:  EPA relies on a highly flawed "category approach" that ignores major differences in the properties and structures of the 13 members of this category.  It compounds this problem by unquestioningly accepting data from inadequate studies to assert low toxicity, rather than demanding that sufficient studies be provided.  As a result, it fails to identify, let alone require to be filled, the enormous gaps in the data available for many of the category members.  EPA ignores or dismisses without explanation its own earlier comments raising serious concerns about the quality and completeness of data provided by the sponsor of these chemicals under the HPV Challenge.  Finally, this example once again shows how EPA's heavy reliance on self-reported use information from manufacturers paints an incomplete and potentially very misleading picture of the actual uses of industrial chemicals. 

The Fatty Nitrogen Derived Cationics category includes 13 chemicals that are used in industrial and consumer detergents and cleaners, as well as hair care products (conditioners or softeners), disinfectants, textile softening and antistatic agents, deodorizers, emulsifiers, dispersants, coagulants, industrial lubricants and corrosion inhibitors, among other uses.  Two supporting chemicals are registered with EPA as antimicrobial pesticides.  Annual production volumes ranged from <1 million pounds for two of the category members to 50-100 million pounds for one of the chemicals.  The sponsor of this group of chemicals under the HPV Challenge was the American Chemistry Council's Fatty Nitrogen Derivatives Panel's Cationics Task Group.

Is this a legitimate category?

EPA and international protocols provide for the grouping of chemicals into categories for data development and assessment purposes.  However, that approach starts with a hypothesis that chemicals that have structural similarities actually possess similar or predictable patterns of biological activity.  These protocols require that the hypothesis actually be demonstrated to be true, once the available data on physical-chemical properties, environmental fate and toxicity/ecotoxicity for the proposed category members are assembled, and that a full and compelling rationale be provided.

Neither the sponsor of this category nor EPA has done any such thing in the present case.  Indeed, the sparse available data do not support the category:

• Few measured physical-chemical and environmental fate data have been provided, which is essential to demonstrate similarity in properties and behavior. Instead, estimated values are provided that are either eerily identical for all category members (e.g., bioconcentration factor, or BCF) or actually vary dramatically across category members (e.g., Henry's Law Constant, which is a measure of the distribution of a chemical between water and the air above it). These findings hardly support the hypothesis that all category members will behave similarly.
• The spotty measured biodegradation data that have been provided also differ significantly (ranging from 0% degradation in 28 days, to 12% in 182 days, to 98% in 2 days). EPA acknowledges this huge range, but fails to discuss how it comports – or doesn't – with the category rationale.
• The aquatic toxicity data vary across category members by orders of magnitude. For example, acute fish toxicity values (96-hour LC50s) vary from 0.07 to 24 milligrams per liter (mg/L) – more than spanning EPA's high-moderate-low hazard classifications.
• Acute oral mammalian toxicity values (LD50s) are available for most category members – but they also vary, from 238 to >16,300 milligrams per kilogram of body weight (mg/kg-bw). And while the sponsor and EPA argue these data are similar enough to support the category (they only span EPA's moderate and low hazard classifications), there is no reason to expect that the mechanisms that impart acute toxicity are at all related to those that lead to other toxicities, such as reproductive or developmental toxicity (these endpoints are discussed further below). So how exactly do similar data for one endpoint support a conclusion that data for entirely unrelated endpoints will be comparably similar? EPA never bothers to explain this.
• As discussed below, far fewer data are available for the other human health endpoints, and what data are available do not support the category.

Despite these findings, EPA states that it agrees with the sponsor's category justification, asserting that the category members possess "similar physicochemical properties, biodegradability, aquatic toxicity, mammalian toxicity and environmental disposition patterns."  The differences in the actual data noted above are neatly set aside.

In comments EPA provided to the sponsor in 2002, EPA itself clearly considered the chemicals in this category to be sufficiently different from each other structurally that it broke them into three subcategories and argued for the need for more data to be provided within the subcategories in order to bolster the overall category.  (As we'll discuss below, the sponsor refused to provide the additional data, and EPA capitulated with nary a word as to why).  These subcategories are:  a) four chemicals that have a single alkyl chain; b) seven that have two alkyl chains; and c) a third subcategory that includes one chemical with three alkyl chains and one that has a dimeric structure.

EPA's rankings 

Now let's look at how EPA ranks the category, and some of the many reasons why we disagree.

Hazard rankings:  EPA ranks the entire chemical category as moderate for human health hazard, apparently due to the results of repeated dose testing.  While EPA maintains that none of the chemicals are expected to bioaccumulate, it expects most (9) of the category members to exhibit moderate persistence.  The hazard for aquatic organisms – fish, invertebrates and algae – is ranked high, based on the results of both acute and chronic testing using multiple test species. 

Exposure rankings:  Exposures to workers, consumers and children are expected by EPA to be high, due to the uses of these chemicals in common household and personal care products.  Releases to the environment are not known, so EPA estimated that exposures to the general public and the environment resulting from such releases would likely be moderate. 

Risk rankings:  EPA judged the risk ranking for this group of chemicals to be medium for all possible receptors. 

Prioritization ranking:  EPA assigned this chemical category a medium priority and identified a list of "possible next steps" to get additional information that would "assist EPA" to develop a better understanding of use and exposures.  These include just about everything you can imagine EPA would have needed to make any findings about exposure in the first place:  potential releases to water from manufacturing, use and disposal; information concerning worker exposures; and information concerning potential exposures to these chemicals in consumer products, such as presence and concentration and consumer use patterns.

Why We Disagree:

1.  As discussed above, the grouping of chemicals into categories can in some cases be justified, but it requires a sufficient amount of measured data to demonstrate that the chemicals within the category actually behave in a similar or predictable manner reflective of their structural similarity. In this case, the available data are grossly insufficient to support the category. Yet EPA still manages to conclude there are no data gaps for any endpoints for this entire category.

     Here is the actual extent of data for the mammalian endpoints:

Endpoint                            # Chemicals with measured data
                                            (includes 2 supporting chemicals)

Acute Oral                                                      11 of 15
Acute Dermal                                                   5 of 15
Acute Inhalation                                              2 of 15
Repeated Dose Oral                                        4 of 15
Repeated Dose Dermal                                    4 of 15
Repeated Dose Inhalation                               0 of 15
Reproductive Toxicity                                       1 of 15
Developmental Toxicity                                    5 of 15

2.  As noted earlier, EPA broke this category into three subcategories, based on structural differences. Let's examine in more detail the nature and extent of mammalian toxicity data provided for the first of these subcategories, the mono alkyl quaternary ammonium chlorides:

a. Repeated dose toxicity. None of the members of this subcategory has a reliable repeated dose toxicity study (a test that is used to evaluate health effects from more than single-dose exposures and serves as a screen for possible effects from chronic exposure). No oral studies are available, and the single dermal study provided used only a single dose. That dose yielded no adverse effects – but the dose was very low, below the dose where if an effect were seen EPA would rank the hazard as high.

Tests that use only a single dose and tests that fail to find an effect level because they use doses that are too low are insufficient to support any hazard assessment.  Yet EPA doesn't even discuss the matter.  It doesn't acknowledge the test is insufficient, nor does it identify this endpoint as a data gap, nor does it adopt a reasonable default assumption in the absence of valid data that developmental toxicity could be high.  Instead, it proceeds merrily to "read across" this negative result to the untested members of the subcategory.  And worst of all, it actually concludes in its hazard characterization summary that "no treatment-related systemic toxicity was evident at the doses tested" – completely burying critical information about data quality and reaching a scientifically unjustified conclusion.

b. Reproductive toxicity. This same subcategory lacks any reproductive toxicity data whatsoever. In comments EPA provided back in 2002 on the test plan submitted by the sponsor of this category, EPA requested that a combined reproductive/ developmental toxicity test be conducted to address this glaring gap (as well as the corresponding gap in data for developmental toxicity). The sponsor responded in 2003, stating that in its view the requested additional testing "will not further the understanding of potential human health hazards…" of these chemicals. The sponsor provided no rationale to support its claim, failing even to acknowledge EPA's point that there were no data available for any of the monoalkyls.

In EPA's current ChAMP assessment, issued in March 2009, EPA now states merely that it accepts this response, and provides absolutely no explanation for its change of heart.  Instead, EPA – without any stated justification – is now content to "read across" to all four members of this subcategory the data from a "supporting" chemical that is actually a di-alkyl, not a mono-alkyl, compound.  Indeed, this supporting chemical serves as the ONLY source of reproductive toxicity test data for all 13 category members!  On this basis, EPA then blithely claims there is no reproductive toxicity – and no data gap for reproductive toxicity – for all 13 members of this category.

c. Developmental toxicity. Data are available for one of the four members of the mono alkyl subcategory for the oral route of exposure, and for two other members for dermal exposure. EPA "reads across" these data to the untested members of the subcategory. That might normally be sufficient, but again in this case neither test yielded any adverse effects at the highest doses tested, and again those doses were very low, below the dose where if an effect were seen EPA would rank the hazard as high.

Remember, tests that fail to find an effect level because they use doses that are too low are insufficient to satisfy an endpoint and support hazard assessment.  But once again, instead of acknowledging this inadequacy, identifying this as a data gap, and using a reasonable default assumption that developmental toxicity could be high, what does EPA do?  It claims "no signs of developmental toxicity were observed"!

In its hazard characterization summary, EPA downplays or omits the results of the developmental toxicity studies it reviewed.  It ignores without explanation evidence of adverse effects that are at least equivocal, and may be significant:

• EPA claims that a test done on the di-alkyl supporting chemical "resulted in no developmental toxicity," despite the fact that increased fetal mortality and decreased fetal body weight were observed – and at doses that warrant a high hazard ranking using EPA's criteria.
• Similarly, with respect to tests done on two dialkyl category chemicals, EPA claims the studies did not produce an effect at the highest doses tested. Yet both chemicals were actually found to have increased fetal resorptions, albeit at doses that EPA would rank as low-hazard.

In each of these cases, we are forced to infer a rationale because EPA never clearly explains its decisions.  But the apparent rationales – effects seen only at doses that are also toxic to the mother, effects within the range historically observed for controls in the laboratory (though with no supporting data provided by the laboratory) – are not sufficient to conclude there are no adverse developmental effects, even if they are also insufficient to conclude there are such effects.  Indeed, in a screening-level hazard characterization based on scant or equivocal data, the default should be either to assume an effect exists or at the very least to call for further testing.

Finally, even if one were to accept these studies as definitively negative for the dialkyl subcategory, EPA has no basis either to extrapolate that finding to the other subcategories or to paper over the enormous data gap that exists for this endpoint.

3. We also disagree with EPA's risk ranking for this chemical category. Even assuming EPA's moderate ranking of the human health hazard of this group is appropriate, given that there remain important data gaps, and exposures to humans are ranked high, and the chemicals are likely to be washed down the drain into the environment, we don't see how this category should be ranked as anything but both high risk and high priority.

4. The two supporting chemicals used to provide hazard data for this category are both antimicrobials, and given EPA's readiness to treat all of the chemicals together, it is reasonable to assume that the category chemicals may also have antimicrobial properties. Add to this the facts that the chemicals in this category are used in a widespread and dispersive manner, and that they may well end up in wastewater treatment plants and surface waters, there is every reason to be concerned that they could adversely affect beneficial microbes used in wastewater treatment and found in the environment.

Antimicrobials used and marketed as such are exempted from TSCA, and are instead regulated under the Federal Insecticide, Fungicide and Rodenticide Act.  However, the chemicals in this category are used in applications where such a function and associated claims are not operative. That is, we have a group of chemicals that EPA considers similar to known antimicrobial pesticides that are disposed of down the drain, into sewage treatment systems that may not effectively remove these chemicals, potentially resulting in distribution through the ecosystem.

EPA is wholly silent on this issue.  In addition to collecting more data, we would strongly recommend that these chemicals be evaluated to determine if inherent antimicrobial properties warrant regulation as antimicrobial pesticides.

5. EPA ranks most (9) of the category members as exhibiting moderate persistence, and the remaining four as of low persistence. Yet it barely discusses the available data and ignores the following:

• Using EPA's own criteria, 3 of the 4 chemicals EPA ranked low actually exhibit moderate photodegradation rates (based on estimated data, as no measured data were provided).
• 3 of the 4 chemicals EPA ranked low for biodegradation appear to exceed EPA's criterion for a low ranking, and several of those ranked moderate show little if any biodegradation.
• EPA ranks as moderate one chemical that showed 0% degradation in 28 days. This obviously should be ranked high.
• EPA also ranks as moderate another chemical that has a half-life in soil of a whopping 1,048 days – nearly 3 years! EPA's criteria rank any half-life in soil that exceeds 180 days as high – yet EPA's ranking of this chemical is, inexplicably, moderate.
• 6 of the category members have no biodegradation data, and any rationale for EPA's implied read-across from the other members is absent.

In comments EPA provided earlier to the sponsor, EPA raised several serious concerns about the extent and nature of data provided for, and the sponsor's claims made about, biodegradation.  In the sponsor's response, it refused to do the testing EPA called for or to revise its claims that all category members are biodegradable. 

In its current ChAMP assessment, EPA seems once again to have fully capitulated:  no data gaps identified, no flagging of the enormous range in biodegradation values that calls into question the viability of the category, and a failure in contradiction of its own criteria to identify any of the category members as having high persistence.

6. EPA claims that none of the category members are expected to bioaccumulate. This is based, however, entirely on estimated data using an EPA model that yields the exact same result – a bioconcentration factor estimate of 71 – for all 13 category members. Doesn't this bear some explanation? Why can't this parameter be measured? EPA's silence is deafening.

7. Last, but not least, this ChAMP assessment amply illustrates the huge shortcomings of the use information EPA has sought to collect under its Inventory Update Rule (IUR). As we've discussed at length before, these data are of questionable value because they are self-reported by manufacturers, often incomplete because of major reporting loopholes EPA has provided, and often kept removed from any public access because EPA provides wide latitude for submitters to claim the information to be confidential. This category is a great example of these problems.

IUR submissions were received for 11 of the 13 category members.  Here's a summary of the extent of data EPA did and did not receive, and what it has not made public because of confidential business information (CBI) claims:

Types of responses in IUR submissions
Number of chemicals

*  Submitter did not submit because information is "not readily obtainable"
** Number represents CBI submissions identified as such by EPA.  EPA notes the number could be higher.

In this case, EPA has ready access to other public sources of use information for these chemicals – including, ironically, the HPV Challenge submission for this category that was provided by the same companies that under the IUR claimed such information CBI!  But in other cases, such information may not be available, as reliable and current chemical use information is very limited in general.

Given the poor performance of manufacturers reporting under the IUR, and the often very limited availability of information from other sources, how comfortable are you in EPA using these data to assess chemicals' risks to workers, the public, the environment, consumers and children?

Conclusion

In an earlier post, we characterized EPA as constructing "houses of cards" in its ChAMP assessments.  Some of you might have considered that to be awfully harsh.

But here's a classic house of cards:  EPA uses very spotty, variable, misclassified or entirely estimated data on degradation or bioconcentration to claim that this questionable category of chemicals has only low or moderate environmental persistence and bioaccumulation potential.  It then uses those findings to downgrade use information that clearly indicates these chemicals are used in a manner that culminates in down-the-drain disposal or similar release, and hence significant, ubiquitous releases to the environment – in order to claim that exposures to aquatic organisms and the general population via such releases will only be "medium."  Finally, it uses that finding to downgrade its finding of high aquatic hazard to a medium risk ranking, and to further downplay its (flawed) moderate human health hazard by finding only a medium risk to the general population.

Any questions?

Our analysis of EPA's risk decision under ChAMP for this category of toxic chemicals vividly illustrates how EPA has failed to adopt a health-protective approach to its screening of HPV chemicals.  Rather, it misclassifies or understates these chemicals' hazards, asserts that existing regulations are sufficient even when they are quite old or do not cover identified exposures, and naively assumes that children will not be as exposed as adults to consumer products used in the home unless they are intended for their use.  Finally, this case demonstrates that manufacturers are not reporting to EPA even readily available information on their chemicals' uses. 

The chlorobenzenes category is comprised of four chemicals: monochlorobenzene (CAS# 108-90-7), 1,2-dichlorobenzene (CAS# 95-50-1), 1,3-dichlorobenzene (CAS# 541-73-1) and 1,2,3-trichlorobenzene (CAS# 87-61-6).  The first three of these chemicals are on the Toxics Release Inventory (TRI).  Reported production volumes under the Inventory Update Rule and TRI emissions for 2005 and reporting manufacturers (other companies may have claimed their identity to be confidential business information) for these chemicals are as follows:

Industrial uses reported by manufacturers include use as intermediates and solvents for plastics manufacturing and basic organic chemicals manufacturing, and pesticide and other agricultural chemical manufacturing.  EPA notes other uses cited in the Hazardous Substances Data Bank (HSDB):  "pesticides, solvents, heat transfer medium, and chemical intermediates, as well as many other uses."

All members of this category were included in test rules issued under TSCA Section 4, in use and exposure-related information reporting rules issued under TSCA section 8(a), and health and safety data reporting rules issued under TSCA section 8(d).  Additionally, these chemicals are regulated under the Clean Air Act and Clean Water Act, and the Occupational Safety and Health Administration (OSHA) has set Permissible Exposure Limits (PELs) for two of the chemicals (1,2-dichlorobenzene and monochlorobenzene).

EPA considers 1,3-dichlorobenzene to be a high-priority chemical, a decision with which we agree (but see point 8 below).  It considers the other chemicals to be low-priority, and this review primarily focuses on that decision.

Hazard rankings:  EPA ranks monochlorobenzene, 1,2-dichlorobenzene and 1,2,3-trichlorobenzene as having moderate human health hazard, based on the results of repeated dose toxicity tests and developmental toxicity tests.  All three of these chemicals (as well as 1,3-dichlorobenzene and one of the supporting chemicals for the category) exhibited evidence of genotoxicity in vivo. One of the chemicals (monochlorobenzene) exhibited some, though not clear, evidence of carcinogenicity, and one of the supporting chemicals showed clear evidence of carcinogenicity.

Exposure rankings:  EPA indicates that there is a high potential for exposure to these chemicals of the general public via environmental releases, based on their reported TRI releases, environmental persistence, detection in environmental monitoring and their myriad uses.  EPA ranks worker exposures high (for two chemicals) and moderate (for the third).  Based on the use of these chemicals in consumer products, EPA ranks exposures to consumers as high and exposures to children as medium.  Interestingly, EPA reports that the IUR data – submitted by manufacturers – do not indicate uses in consumer products, but that several other sources, including the HSDB, the NIH Household Products Database, and EPA's Source Ranking Database, do indicate uses in consumer products.

Risk rankings:  EPA ranks the risks of these three chemicals to the environment to be moderate to fish and high to invertebrates and aquatic plants.  It ranks risks to the general public, consumer and children as moderate, despite their high exposure potential, based on ranking human health hazard as moderate.  EPA ranks risks to workers as low for the two chemicals with OSHA PELs (1,2-dichlorobenzene and monochlorobenzene) and moderate for 1,2,3-trichlorobenzene.

Prioritization rankings:  EPA considers all three chemicals to be low-priority because it expects existing regulations and ongoing reporting to be sufficient to mitigate risk and to alert EPA to the presence of chemicals in workplaces, environmental releases from facilities, and drinking water.

Why we disagree:

1. Given the high exposure potential to these chemicals, which EPA found to pose moderate human health hazards, even a modestly health-protective decision would rank their risks to the general public, consumers and children as high, not moderate. The potentially large number of uses, the large production volumes and TRI releases for two of the three chemicals and the substantial uncertainty regarding the magnitude, frequency, and duration of possible exposures, would support such a ranking.
        In this way, at least the development of better hazard, use and exposure information and further scrutiny of the magnitude of risks would be spurred.  Instead, EPA concludes that "No follow-up action is suggested at this time on [these three] chemicals in this category.
2. The conclusion that these chemicals are low-priority due to the existence of Clean Air Act and Clean Water Act regulations is unjustified. Despite those regulations, substantial TRI releases to air and some releases to water are being reported – sufficient for EPA to rank exposure potential as high. Yet EPA then invokes those same regulations as being adequate to prevent the very exposures it just characterized as potentially high!
3. The main reason EPA ranked consumers' and children's exposure potential high and medium, respectively, is because of expected exposures through consumer products. However, the cited CAA and CWA regulations do not address these exposures at all.
4. EPA provides no real basis for ranking children's risks lower than consumers'. EPA implies this is because it lacks evidence that these chemicals are used in products specifically intended for use by children. But EPA notes it sources of use information indicate "many other uses" for these chemicals that are not identified, and EPA provides no basis on which to conclude that uses in products not intended for use by children would not expose them. Indeed, EPA acknowledges: "Exposures to children, however, may be expected to occur through the household use of some consumer products" (p. 48 here).
5. Positive results for all of the chemicals in this category in in vivo genotoxicity studies, and for monochlorobenzene and one of the supporting chemicals in two-year carcinogenicity bioassays, clearly warrant follow-on testing to clarify the carcinogenic potential of at least the untested chemicals in this category. Yet EPA proposes no such testing; indeed it barely discusses the findings other than to repeatedly emphasize the one chemical for which negative carcinogenicity results were found.
6. In its risk rationale for children, EPA misclassifies the level of developmental toxicity for two of the three chemicals. On p. 3 here, EPA claims "available data with postnatal exposures in animals suggest a low hazard for two category members (CASRNs 95-50-1 and 108-90-7)." Yet the results of developmental toxicity tests in rats via inhalation exposure to vapors of these two chemicals actually showed moderate and high developmental toxicity, respectively, based on EPA's own criteria:
   • CAS# 95-50-1:  LOAEL ~ 2.4 mg/L/day vs. EPA range for moderate developmental toxicity of 1-2.5 mg/L/day (see pp. 27 and 38 here vs. the table on p. 12 of EPA's Methodology for Risk-Based Prioritizations under ChAMP.
   • CAS# 108-90-7:  LOAEL ~ 0.35 mg/L/day vs. EPA cutoff for high developmental toxicity of <1.0 mg/L/day (see pp. 26 and 38 here vs. the table on p. 12 of EPA's Methodology for Risk-Based Prioritizations under ChAMP
     High exposure and high or moderate developmental toxicity – surely even EPA will acknowledge these chemicals merit a high-risk and high-priority ranking?
7. EPA invokes the existence of OSHA PELs for two of these chemicals to argue that worker risk is low. Both of these PELs were promulgated in 1989 (see here and here), and at least for one of them, monochlorobenzene, NIOSH has found that "[t]he 1989 OSHA PEL may not be protective to workers."
        Were these PELs developed in consideration of and are they adequate to protect against the toxicities EPA presents herein?  While some of the test data (ca. 1984, 1987) predate the year the PELs were adopted, given the long process required to develop a PEL, the timing cannot be taken as evidence that such data were considered by OSHA.  EPA cannot merely assert that the PELs are adequate to protect workers without demonstrating that the permissible levels were set after evaluation of these data, and are still adequate in light of any other more recent toxicity data. 
        At the very least, EPA needs to engage in a meaningful referral of these data to OSHA, using its TSCA Section 9 authority, including follow-up to ensure the data have been fully considered and acted upon appropriately at OSHA.
8. Despite its high-priority finding for one of the chemicals in this category, 1,3-dichlorobenzene, EPA's only response is to state that "in order to confirm or refute the high potential risk … companies are encouraged to provide available information on a voluntary and non-confidential basis." This is clearly inadequate for such a high-risk chemical. It begs the question: what would it take for EPA to be willing to actually impose risk management on such a chemical?
9. Last but not least, EPA's finding that there are myriad product uses of these chemicals – none of which were reported by any of their manufacturers under the Inventory Update Rule (IUR) – vividly illustrates why EPA's frequent sole or primary reliance on the IUR as its source of use and exposure data for ChAMP assessments is wholly inadequate as a basis for making exposure and risk decisions for high production volume chemicals.