Civil Action No. 99-1170
Judge Emmet G. Sullivan

JIM J. TOZZI, et al.,







CHARLES J. FROMM, DC Bar No. 420021
Multinational Legal Services, P.C.
11 Dupont Circle, Suite 700
Washington, DC 20036
Ph:(202) 797-7124
Fax:(202) 939-6969

Attorneys for Plaintiffs

Dated: September 8, 1999


I. Factual Background

In 1978, Public Law 95-622, 42 U.S.C. § 241(b)(4),1 directed the Secretary of Health and Human Services to prepare an annual report -- the Report on Carcinogens ("RoC") -- listing substances which were "known" to be, or "reasonably anticipated" to be, human carcinogens to which a significant number of persons residing in the United States were exposed, along with certain other related information. In 1993, the law was amended to make the RoCs biennial rather than annual.

The first RoC was published in 1980. By 1994 seven editions had been published. From 1994 to 1996, the agency undertook an extensive public review of its listing criteria/rules2 and process, and in September 1996 published revisions to the listing criteria and procedures in the Federal Register. (Compl. Exh. 1) The 8th RoC was published in May 1998. A complete original of the 8th RoC is enclosed as Exhibit 1.3

Since 1981, the Department of Health and Human Services ("HHS") has listed dioxin (2,3,7,8-tetrachlorodibenzo-p-dioxin or "TCDD") in the RoCs in the "reasonably anticipated to be human carcinogens" category, based on "sufficient" animal evidence and "inadequate" human evidence.4 The current RoC listing for dioxin states with regard to exposure sources:

Both occupational and consumer exposure to TCDD due to herbicide manufacture and use no longer exist. Current sources of exposure are municipal incinerators, dump sites, and contaminated soil. Primary exposure is from ingestion of food from TCDD-contaminated sites. Consumer exposure could possibly occur through skin contact with soil, vegetation, or paper products contaminated with TCDD; consumption of root vegetables, fish, meats and milk contaminated with TCDD; and inhalation of TCDD-polluted air from hazardous waste sites, industrial and municipal incinerators, combustion of leaded gasoline, diesel fuel, and wood; cigarette smoke; fields sprayed with certain pesticides, herbicides, germicides, or defoliants; plant accidents, and transformer/capacitator fires involving PCBs and chlorobenzenes. Workers at certain municipal and industrial incinerators and hazardous waste sites possibly could be exposed to TCDD. TCDD has been detected following fire accidents involving transformers containing polychlorinated biphenyls (PCBs), as well as effluents from commercial incineration units (ATSDR, 1989f). Potential exposure to TCDD could occur for workers involved in the cleanup of PCB-capacitator/transformer fires.

8th RoC at pp. 195-196 (emphasis added).

In July 1997, HHS, through the National Toxicology Program ("NTP"), announced that dioxin and 13 other substances had been nominated for listing, or changes in current listings, in the 9th RoC.5 Dioxin was nominated for upgrading from the "reasonably anticipated" listing category to the "known" category, and substantial revisions were proposed for the information on "Exposure" to be provided in the RoC in connection with the listing. Public comments were invited at that point, and for dioxin there were four additional occasions for public comments due to a "re-review" (Compl. Exh. 14). At the time of this first public notice there was no publicly-available document that reflected the deliberations on the dioxin upgrade proposal.

Plaintiff Tozzi initially commented, in August 1997, that the nomination of dioxin for the "known" category was inconsistent with past agency practice because the agency had consistently relied on evaluations of the strength of the evidence from studies in humans by the International Agency for Research on Cancer ("IARC") in making its listing decisions, and IARC had re-evaluated the strength of the human evidence for TCDD only several months before, in February 1997, as "limited" rather than "sufficient".6 (Exh. , Compl. Exh. 8) The definition for the agency’s "known" category requires evidence from studies in humans that is "sufficient" to indicate a causal relationship.

Several months later, in September 1997, the agency issued "draft background documents" ("DBD") for the listing proposals, including dioxin, when it announced a public meeting of its outside peer-review committee (the "RoC Subcommittee) to be held October 30-31, 1997. Sometime prior to this, the agency had held non-public meetings (without any publicly available record) of its first two RoC review groups, the "RG1" and the "RG2". The DBDs reflected the findings of the RG1 and RG2 and contained the agency’s proposed listing rationale, and they were presented to the RoC Subcommittee to concur with, dissent from, or recommend changes.

Tozzi submitted comments on the DBD for dioxin, asserting that the proposed listing change and accompanying rationale (a) were clearly contrary to the agency’s listing criteria for the "known" category, (b) were inconsistent with the recent IARC evaluation of the human evidence (in which HHS and other NTP agencies had participated), and (c) contained erroneous statements that IARC had found a "causal relationship", and that the "in press" Bertazzi et al. findings relied on by the agency had not been considered by IARC and supported the listing nomination and rationale. (Compl. Exh. 11) With regard to point (a), Tozzi pointed out that the DBD, which stated the agency’s proposed listing rationale, expressly relied on a combination of evidence from studies in humans, studies in animals, and various mechanism-of-action data, including studies in vitro (literally, in glass – i.e., experiments conducted in laboratory containers, as opposed to in living entities, or "in vivo")7, rather than on "sufficient evidence from studies in humans" indicating as causal relationship, as required by the listing criteria.

On October 31, 1997, the RoC Subcommittee voted 4 to 3 to concur with the RG1 and RG2 recommendation to upgrade dioxin to "known" based on the agency reasoning reflected in the DBD.

Following this meeting, Tozzi wrote to the agency’s designated head of the RoC program, Defendant Kenneth Olden, to point out numerous alleged irregularities in the October 31 proceedings and reiterate the central points of his previous comments. (Compl. Exh. 10)

Shortly thereafter, Defendant Olden announced that there would be a "re-reviuew" of dioxin. This "re-review" was scheduled to be held in conjunction with the RoC Subcommittee reviews of a second round of substances being considered for listings in the 9th RoC. This peer review meeting which included the dioxin "re-review" was held on December 23, 1998. Immediately prior to that meeting, the agency re-issued the September 1997 DBD without any changes, along with a "Supplement" that contained "Epidemiological Updates" containing references to selected 1997-1998 publications on dioxin that had not been previously cited in the DBD. The Supplement did not contain any new reasoning supporting the listing proposal, nor did it correct any of the alleged inaccuracies of which Plaintiff Tozzi had complained to Defendant Olden prior to his ordering the re-review.

Before the December 2, 1998 external peer review meeting, Tozzi wrote further letters to Olden pointing out that the studies in humans on which the agency had been relying for the proposed dioxin upgrade apparently were not relevant because they involved occupational exposures to mixtures of chemicals that were qualitatively and quantitatively distinct from the dietary exposures to dioxin alone that were being addressed in the new listing proposal. (Compl. Exhs. 15, 16). Defendant Olden’s response to these comments (Compl. Exh. 17) did not attempt to explain how those studies were relevant; rather, he suggested that such workplace exposures could be added separately to the listing proposals. Likewise, nothing has ever been added to the listing rationasle in the Draft Background Document explaining the relevance of the studies on the mixed, high-level, workplace exposures to everyday consumer exposures to meats, fish, and dairy products.

At the December 2, 1998 re-review of dioxin by the RoC Subcommittee, the Subcommittee reversed its previous vote and voted 7 to 5 against the proposed dioxin upgrade.8

On February 24, 1999, the NTP Executive Committee – the fourth and last RoC review committee – held a non-public meeting at which it voted in favor of the proposed upgrade, despite the recent contrary vote of the external peer reviewers. (See Lucier Declaration dated July 15, 1999, attached to Def.’s Motion to Dismiss as Exh.2.)

Between November 20, 1998 and January 19, 1999, there was an exchange of correspondence between Tozzi’s legal counsel and the agency’s legal counsel in which Tozzi’s legal counsel asserted that the agency’s failure to follow its own listing criteria in the case of the proposed dioxin listing had left the proposal process fatally flawed, and that Tozzi would mount a legal challenge to the agency’s position and the listing proposal if the agency continued to rely on that position. The agency’s legal counsel asserted that the asgency’sd 0roposed dioxin listing was validly based on a final "descriptive" paragraph in the notice of the 1996 criteria revisions stating that a listing could be based on scientific judgment" concerning "all relevant information". (Compl. Exhs. 23-25.)

In Federal Register notices on April 2 and April 19, 1999 – labeled "clarifications" – the agency formally solidified its position that it would not comply with its own listing criteria. (Compl. Exhs. 27, 28.) Read together, these two notices stated that the agency need not rely solely on evidence from studies in humans for a listing in the "known" category, but rather could rely as a matter of "judgment" on any evidence it considered relevant, including animal and in vitro data. The agency also announced that this position would be applied to all pending listing proposals and all future proposals. Although these notice did not refer specifically to dioxin, but only to certain "questions sand alleged "misunderstandings" that had arisen with regard to the listing criteria, these

"Clarifications" were clearly issued in response to the numerous challenges raised to the dioxin proposal and in anticipation of a legal challenge to a forthcoming change in the dioxin listing.9

The listing proposals for the 9th RoC, including dioxin, have now been forwarded to Defendants Olden and Secretary Shalala for final approval (although final approval may already have occurred).10 The next public will hear of the listing actions will be with actual publication and dissemination of the 9th RoC. It is not known exactly when this will occur, only that it is likely to occur, absent judicial intervention, sometime between the present and the end of November 1999. When a new edition of the RoC is published, its availability is advertised widely by the agency, and the full report is made available on the Internet.

II. The Administrative Record

For purposes of this suit, the key portions of the administrative record consist of three main components: (1) The materials in which the agency has presented to the public the reasoning and evidence it relies on to support of the proposed upgrade of dioxin to the "known"category; (2) the materials concerning the agency’s 1996 revisions to its listing criteria and the agency’s contemporaneous interpretation and later "clarification" of those criteria; and (3) materials reflecting the controversy between Tozzi and the agency over the meaning of the criteria, their application to the dioxin listing upgrade proposal, and the evidence and reasoning presented by the agency in support of the proposal.

Most of these materials are attached to the Complaint as exhibits. Of particular note, with regard to component (1), are the agency’s "RC Draft Background Document for TCDD" (Sept. 30, 1997), and the Supplement to the DBD issued in November 1998. The most important part of the DBD is the three-page section titled "Proposed RC Summary for TCDD" in the DBD table of contents, in which the agency summarizes its rationale and the scientific evidence supporting the proposed listing.11 Those three pages, RC-1 through RC-3, are attached to the May 14, 1999 Complaint in this action as Exhibit 9. In that Summary, the agency clearly states that the proposed upgrade of the dioxin listing is based not on evidence". The Summary then proceeds to list three categories of evidence being relied on: (1) "human studies", (2) studies "in experimental animals", and (3) evidence concerning "a similarity in the mechanism of induction of animal and human tissue biochemical and toxicological responses" to dioxin. The agency’s reliance on these three types of evidence in combination, rather than reliance on evidence from "studies in humans" which it has

found "sufficient" to indicate as "causal relationship", as required by its listing criteria, is at the heart of the issues concerning the meaning and application of the listing criteria.

Following the above listing of these three types of evidence is a paragraph explaining the agency’s reliance on the 1997 IARC findings and the "in press" Bertazzi et al. findings. This paragraph is at the heart of Plaintiffs’ other contentions that the proposed listing upgrade is arbitrary and capricious due to these statements being substantially inaccurate.

Thus, perhaps the most crucial portions of the administrative record for purposes of the present case are the first page in the DBD, page RC-1 (in Compl Exh.9), and the "Human Exposure" section of the DBD, pages 2-1 and 2-2 (Compl. Exh. 4).

III. Summary of Legal Basis for Relief

The agency that Congress entrusted with classifying and listing substances as either "known" or "reasonably anticipated" carcinogens is poised to change the listing of dioxin from a "reasonably anticipated" to a "known" human carcinogen. The agency has promulgated specific rules, which it calls "criteria", for determining whether the scientific evidence is sufficient to support a listing in one category or the other. The administrative record shows that the agency is prepared to upgrade dioxin based on a complete disregard for its formal listing criteria and on factually inasccurate statements regarding the scientific evidence and its relevance.

The agency’s actions on the dioxin proposal are based on a position regarding the meaning of the "known" listing criteria that is clearly at odds both with the plain wording of the criteria, with contemporaneous agency explanations, and with the agency’s consistent and longstanding policy and its interpretation of the legislation. The agency’s most recent formal position on the meaning of its criteria was expressed in two April 1999 Federal Register "clarification" notices issued by

Defendant Olden, and those notices unmistakably constitute final agency action for the purpose of applying the listing criteria to dioxin.

The agency to date has followed this arbitrary, capricious and otherwise unlawful course of action through four of the six review stages for the 9th RoC, and the decision has passed into the hands of Defendant Olden (and perhaps to Defendant Secretary Shalala by this time), who, according to NTP, effectively has the final say on listing actions, subject to subsequent approval by the Secretary of HHS.12 Both of these concluding steps in the process are non-public, and plaintiffs will not know when they occur, only that they are likely to occur within the next month or two.

When publication of the 9th RoC occurs, if it contains an upgrade of dioxin with the description of exposure sources contained in the DBD (widespread occurrence in meat, fish, and dairy products), that action is highly likely to generate media coverage and create a new dioxin "food scare" affecting the business interests of plaintiffs. Plaintiffs are threatened with reduced sales of food preparations containing meat (including poultry), fish, and dairy products. In order to attempt to counter public fears, Plaintiffs will need to test their food products for dioxin – an expensiuve proposition by itself – and explore options for procuring products that are "uncontaminated".

Plaintiffs seek a preliminary injunction to prevent the agency from proceeding with hard-copy publication and Internet dissemination of the 9th RoC containing the proposed dioxin listing upgrade. Once such publication and dissemination occurs, the harm to plaintiffs will be effectively irreparable. On the other hand, preserving the status quo pending a ruling from the court on the validity of the agency’s position on the meaning of its listing criteria and the validity of the proposed dioxin upgrade would not harm Defendants, would be likely to save the agency considerable effort and resources in recalling and revising the 9th RoC, and would be in the public interest.


Plaintiffs are entitled to a preliminary injunction if they establish four points: (1) likelihood of success on the merits; (2) irreparable harm if a preliminary injunction is not granted; (3) that the harm to them if a preliminary injunction is not granted outweighs any harm to Defendants if the preliminary injunction is granted; and (4) that granting of a preliminary injunction would further the public interest. Mova Pharmaceutical Corp. v. Shalala, 140 F.3d 1060, 1066 (D.C. Cir. 1998).

Although the Plaintiffs must satisfy all four requisites to some degree, they do not need to make an equally strong showing on each. The courts weigh all the elements together, and the D.C. Circuit has recognized that a strong showing on one factor, such ss likelihood of success on the merits, can justify a preliminary injunction even if the showing on one or more other factors, such as irreparable injury, is relatively weak. CityFed Financial Corp. v. Office of Thrift Supervision, 58 F.3d 738, 747 (D.C. Cir. 1995); Cuomo V. United States Nuclear Regulatory Comm’n, 772 F.2d 972m 974 (D.C.Cir. 1985) (per curiam); Serono Laboratories, Inc. v. Shalala, 974 F.Supp. 29, 32, 35-37 (D.D.C. 1997); Bracco Diagnostics, Inc. v. Shalala, 963 F.Supp.20, 27 (D.D.C. 1997); Bristol-Myers Squibb Co. v. Shalala, 923 F.Supp. 212, 215-16 (D.D.C. 1996); Smith v. Shalala, 954 F.Supp. 1, 3 (D.D.C. 1996). In the case at hand, Plaintiffs contend that there is a strong likelihood of success on the merits; that the balancing of possible harm and consideration of the public interest also strongly favor a preliminary injunction; and that the threat of irreparable harm, while having an element of quantitative uncertainty, is sufficient to support the requested relief. Accordingly, a preliminary injunction should be granted.

Plaintiffs note at the outset that Defendants have argued in support of their Motion to Dismiss that there can be no judicial review unless and until there is evidence that harm has already occurred as a result of final agency action. If that were the case, the courts would never consider, much less grant, preliminary injunctive relief against threatened agency action in order to preserve the status quo and ensure against future harm, which of necessity always is speculative to a degree. Such is not the case, of course, as shown by the many cases (a number of which are cited in the section below on irreparable harm) in which courts have preliminarily enjoined threatened agency action to prevent irreparable harm.

I. There is a strong likelihood that Plaintiffs will prevail on the merits.

The agency’s interpretation of its listing criteria/rules for the "known" category, together with the imminently threatened application of that interpretation to dioxin, is "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law" under 5 U.S.C. 707(2)(A). While the applicable standard or review is deferential to the agency, the judicial review is to be "searching and careful". Citizens to Preserve Overton Park v. Volpe, 401 U.S. 402, 415, 416 (1971); Wawszkiewicz v. Department of Treasury, 670 F.2d 296, 301 & n.42 (D.C. 9181); Ethyl Corp. v. EPA, 541 F.2d 1, 34-37 (D.C. Cir.) (en banc), cert. denied, 426 U.S. 941 (1976). Applying this standard of review, the Court should hold that the agency’s actions are, and have been, arbitrary and capricious on three separate grounds.

First, it is firmly established that agency action is arbitrary and capricious if it is not rationally based on adequate consideration of the "relevant factors". Troy Corp. v. Browner, 120 F.3d 277, 288, 293 (D.C.Cir. 1997); Nursing Center of Buckingham and Hampden, Inc. v. Shalala, 990 F.2d 645, 650 (D.C.Cir. 1993) (Secretary reasonably applied agency guidelines interpreting statutory language) (quoting Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983)); American Fin. Services Ass’n v. FTC, 767 F.2d 957, 971-91 (D.C.Cir, 1985) (agency decision not arbitrary and capricious because it comported with all three criteria set out in its policy statement), cert. denied, 475 U.S. 1011 (1986). Troy, supra, is very similar to the present case. In Troy, the D.C. Circuit held that EPA acted arbitrarily and capriciously by listing a substance as toxic in reliance on reported scientific findings that violated the agency’s own "Hazard Assessment Guidelines for Listing Chemicals on the Toxic Release Inventory".

Second, agency action based on an interpretation of the law in which the agency has changed its settled course is arbitrary and capricious unless the agency acknowledges the change and supplies a reasoned justification for the change of course. DSE, Inc. v. United States, 169 F.3d 21, 31 (D.C. Cir. 1999) (court found that the SBA had acted contrary to its long-standing policy, but that the error was non-prejudicial); Troy, supra, at 286-87; Pontchartrain Broadcasting Co. v. FTC, 15 F.3d 183, 185 (D.C. Cir. 1994); Graphic Communications Int’l Union, Local 554 v. Salem-Gravure Division of World Color Press, Inc., 843 F.2d 1490, 1493-94 (D.C. Cir. 1988) (agency actin imposing a new discovery requirement that was a departure from established precedent was arbitrary and capricious because the agency did not supply a reasoned explanation), cert. denied; 489 U.S. 1011 (1989); State Farm, supra, at 41-42.

Third and fourth, an agency action will also be found arbitrary and capricious if the agency either "entirely failed to consider an important aspect of the problem, [or] offered an explanation for its decision that runs counter to the evidence before the agency..." State Farm, supra, at 43l Troy, supra, at 288.

On all four of these counts the agency action in this case is arbitrary and capricious.

A. The agency’s recent new interpretation of its criteria for the "known" category, and its threatened application of that interpretation to dioxin, are contrary to the plain wording and meaning of the criteria, and therefore the agency’s actions are not based on consideration of the "relevant factors", but on irrelevant factors.

The agency’s criteria for listing a substance in the "known" category are clear and unambiguous in stating the relevant factors that must be addressed. An agency interpretation of its rules that is contrary to their plain meaning is not entitled to any deference and may be found arbitrary and capricious. As the D.C. Circuit has stated, an agency’s interpretation of its own rules is entitled to substantial deference unless "an alternative reading is compelled by the regulation’s plain language or by other indications of the [agency’s] intent at the time of the regulation’s promulgation." S.G. Loewendick & Sons, Inc. v. Reich, 70 F.3d 1291, 1294 (D.C. Cir. 1995). And see Thomas Jefferson Univ. v. Shalala, 512 U.S. 504, 512 (1994); Rollins Env’l. Serv., Inc. V. EPA, 937 F.2d, 649 652 (D.C.Cir. 1991). As discussed below, the agency’s new interpretation of iuts criteria for the "known" category is both contrary to the plain wording and meaning of the criteria and contrary to the agency’s clearly expressed contemporaneous interpretation. The agency’s new interpretation, as applied to dioxin and as expressed in the April 1999 Federal Register "clarifications", is therefore entitled to no deference.

The agency’s published criteria for listing in the "known" category, as revised in September 1996, state:

There is sufficient evidence of carcinogenicity from studies in humans which indicates a causal relationship between exposure to the agent, substance or mixture and human cancer.

61 Fed.Reg. 50499 (Sept. 26, 1996) (Compl. Exh. 1). These criteria become even clearer when juxtaposed with the criteria for the "reasonably anticipated" category (also in Compl. Exh.1), which encompass substances for which there is "limited" evidence from studies in humans indicating only that a causal relationship is "credible". The criteria for the latter, "reasonably anticipated", category specifically allow reliance on evidence from studies in animals, and specifically address situations where there is "less than sufficient evidence of carcinogenicity in humans...however...there is convincing relevant information that the agent acts through mechanisms indicating it would likely cause cancer in humans." Id.

The criteria for the "known" listing category (as well as the "reasonably anticipated" category) are clearly and tightly drafted, using a combination of terms that have a plain everyday meaning ("in humans" vs. "in animals"), a meaning plain to those familiar with health science ("causal relationship" vs. "association"), or a meaning consistently established through use in the RoCs ("sufficient" vs. "limited" human evidence). The criteria for the "known caregory contain the following requirements:

1. The evidence must be from "studies in humans".
2. That evidence from such studies must be "sufficient" to indicatea "causal relationship" between exposure to the substance and cancer in humans.

In April 1999, the agency issued two "clarifications" of its listing criteria, which together state that the agency can place a substance in the "known" category "based on scientific judgment, as applied to all relevant information." (Compl. Exhs. 27, 28, 1) As the allegations and documentation contained in the Complaint in this case make clear, those "clarifications" are intended by the agency to constitute interpretations of its 1996 criteria revisions (Compl. Exh. 1) which will allow it ti upgrade the listing of dioxin to the "known" category without a finding that there is sufficient evidence from studies in human to indicate a causal relationship. Instead, the agency intends to rely on a combination of insufficient human evidence, evidence from studies in experimental animals, and other information relating to mechanism of action, including evidence from in vitro experiments.

1. The requirement that a listing in the "known" category be based on "evidence from studies in humans" plainly excludes reliance on studies in animals and in vitro, but the agency is relying on studies in animals and in vitro to compensate for less than sufficient evidence from studies in humans.

There can be no doubt that the term "studies in humans" plainly excludes reliance on evidence from studies in animals. It also plainly excludes reliance on data from in vitro (literally, in glass) experiments, such as exposure of call cultures or tissue samples to a substance while in laboratory containers. "Studies in humans" are studies of health effects in whole humans exposed while living; the term plainly cannot include studies "in animals" or in glass (in vitro). Until the exchange of correspondence with agency counsel leading up to this lawsuit and the April 1999 "clarifications", the agency consistently interpreted this phrase as meaning simply epidemiologic studies.13 The criteria for both RoC categories also plainly distinguish between studies in humans and studies in experimental animals, as well as between studies in humans and other "relevant

information that the agent acts through mechanisms indicating it would likely cause cancer in humans." (Compl. Exh. 1)14

The agency’s disregard for these explicit listing requirements could not be clearer. The Summary of the agency’s Draft Background Document for TCDD ("DBD") (Compl. Exh.9), which presents the agency’s rationale for the dioxin listing proposal, states that the "known" listing will be based "on several types of evidence". These types of evidence which are presented so as to be distinct from each other, but also so as to be considered collectively: (1) "[h]uman studies", (2) "[s]tudies in experimental animals", and (3) "evidence [on] ...mechanism of induction of animal and human tissue biochemical and toxicological responses". The subsequent brief narrative then discusses this evidence in the same sequence, with the human studies referred to as epidemiological studies, and making clear that the mechanistic evidence includes study of both human and animal cells in vitro as well as in vivo.

2. The criteria require evidence from studies in humans "sufficient" to indicate a "causal relationship", but the agency has not found that the human evidence is sufficient to indicate a causal relationship, and has found that the evidence indicates only an "association".

It is clear from a reading of the listing criteria that the terms "sufficient" and "causal relationship" are significantly related. The requirement that the human evidence be "sufficient" to indicate a "causal relationship" in the criteria for the "known" category is in clear contrast with the weaker requirement for the "reasonably anticipated" category that the human evidence can be "limited" in indicating that a "causal relationship is credible, but that alternative explanations, such as chance, bias or confounding, could not be adequately excluded."

These two terms – "sufficient" and "limited" – have also been consistently used to make this distinction both by the agency and by the International Agency for Research in Cancer ("IARC") (a subsidiary of the World Health Organization), and the agency has consciously harmonized this terminology with IARC’s. For example, the Introduction to the 7th RoC (1994) contained a section titled "Relationship of the Annual Report Criteria to IARC Criteria", in which the agency stated: "Although the IARC and the Annual Report’s schemes do not exactly correspond to one another, the Annual Report’s scheme and associated degrees of evidence are based on IARC’s classification scheme and degrees of evidence."15 The difference between the terms "sufficient" and "limited" is that the former means that a causal relationship has been indicated to a degree that is stronger than a "credible interpretation", and that alternative explanations for positive findings, such as chance, bias, or confounding, have been adequately excluded. This distinction between "sufficient" and "limited" human evidence, and correspondence with the same IARC terminology, is also evident throughout the agency’s published listing profile 1).16

The terminology and concept of "causal relationship" are central to the discipline of environmental epidemiology, and are recognized and understood by all epidemiologists as well as other scientists and laymen who have some familiarity with epidemiology. The term "causal relationship" is always used in contrast to the term "association". The concept of "causal relationship" is so fundamental to the evaluation of studies in humans that it hardly needs supporting authority. Indeed, one might simply consult prominent publications of the agency itself.17

The distinction between "causal relationship" and "association" is also evident from the profile of the evidence supporting the current listing of dioxin as "reasonably anticipated" in the 8th RoC (Exh. 1 at p.195). The listing as "reasonably anticipated" means that the evidence from studies in humans has not been found sufficient to indicate a causal relationship. The profile relies on findings by IARC (Prior to 1997) that there are no adequate data from which to evaluate the carcinogenicity of dioxin in humans. The profile then proceeds to state that a number of epidemiology studies of humans exposed to dioxin as a contaminant in other chemicals "indicate an association between exposure and stomach cancer, lymphomas, and soft tissue sarcoma..." (Id., emphasis added.) Thus, evidence of "association" has not heretofore been regarded by the agency as sufficient to support a finding of causal relationship and listing in the "known" category, including in the case of dioxin specifically.

Consistency in this distinction between "sufficient" human evidence indicating a causal relationship and "limited" human evidence not sufficient for indicating a causal relationship was also maintained in the listings added to the 8th RoC (Exh. 1), published in 1998, following the 1996 revisions to the agency’s criteria.18 The profiles for the only two additions to the "known" listings in the 8th RoC, Cyclosporin A and Thiotepa (pp. 34 and 46, respectively), state that they are "based on studies in humans which indicate a causal relationship..." and cite IARC Monograph 50, 1990). The cited IARC monograph evaluations of those two substances show that IARC found that there was "sufficient evidence" of carcinogenicity in humans (italics as in original). (Exh. 17) On the other hand of the 13 substances added to the "reasonably anticipated" category in the 8th RoC, one (consisting of two closely related substances) had "limited" evidence of carcinogenicity from studies in humans, and the other 12 were stated to have "no adequate data available" to evaluate" to evaluate carcinogenicity in humans.19

The only public administrative record showing the agency’s rationale for proposing to upgrade dioxin to the "known" category, the DBD, contains no finding indicating that the criteria for the "known" category have been met. The DBD contains neither a finding of "sufficient" evidence from studies in humans, nor a finding that the evidence from studies in humans indicates a "causal relationship". Instead, the DBD Summar states that "[h]uman studies have found an association [with certain cancers]." (Compl. Exh. 9 at p. RC-1, emphasis added.) As discussed below, the DBD also seems to indicate that the 1997 IARC review found a causal relationship, but those observations (which are not independent agency findings of causal relationship in any event) are inaccurate, since IARC explicitly evaluated the human evidence for dioxin as "limited." (Exh. 7 at pp. 342,631.)

The proposed listing, which is based on a final agency action giving notice that the agency will not follow its criteria, must be found to be arbitrary and capricious because it does not reflect consideration and reasoned decisionmaking based on the "relevant factors", as required by established law, but instead is based on irrelevant factors.20

B. The agency’s application of its "known" criteria to dioxin, and its recent "clarifications" of the criteria as applied to dioxin, are arbitrary and capricious because they constitute a substantial departure from established policy and precedent without an adequate reasoned justification.

As is clear from reviewing the listings in the 8th RoC (Exh. 1), which contains all current listings starting with those from the 1st RoC in 1980, the agency has been perfectly consistent in requiring "sufficient" evidence from studies in humans indicating a "causal relationship" to place a substance into the "known" category. There was no substantive change to the "known" category in the 1996 notice of revisions to the criteria. The September 26, 1996 Federal Register notice (Compl. Exh. 1) clearly states that the only revision to the "known" category are those highlighted

by underlining" (which became italics in the notice). Under the section of the notice titled "Revised BRC Criteria Known To Be Human Carcinogens", the only revisions indicated by italics are minor editorial changes.

The disputed final "descriptive paragraph" that the agency relies on does not mention any changes being made to the "known" criteria. That paragraph is printed in italics under the heading of "Revised BRC Criteria Reasonably Anticipated To Be Human Carcinogens", and there is no indication from either the format or the textual explanations that the paragraph applies to the "known" category. Additionally, new language added to the "reasonably anticipated" category – "There is less than sufficient evidence of carcinogenicity in humans, however, . . . there is convincing relevant information that the agent acts through mechanisms indicating it would likely cause cancer in humans" – comports with the agency’s summary of the dioxin evidence in the DBD, strongly indicating that such evidence supports only a listing in the "reasonably anticipated" category.

In short, the plain wording of the formal notice of the 1996 revisions did not indicate any substantive revisions to the "known" category such as the agency has claimed with regard to dioxin in its correspondence with Plaintiff Tozzi and as formalized in the two April 1999 Federal Register notices of "clarifications". The law is clear that the plain meaning of agency rules controls over any conflicting agency interpretation. Pfizer, Inc. v. Heckler, 735 F.2d 1502, 1507, 1510 (D.C. Cir. 1984).

The plain meaning of the 1996 criteria revisions, as discussed above, was, however, confirmed in consistent contemporaneous statements by the agency. In both a Departmental press release issued on the same day (Compl. Exh. 2), and a contemporaneous NIEHS/NTP News article in the agency’s in-house science journal, Environmental Health Perspectives, (Compl. Exh. 3) the agency stated that the only substantive revisions pertained to the "reasonably anticipated" category, and that the criteria for the "known" category remained substantively unchanged from the former criteria. The Departmental press release was reviewed by Defendants Olden and Shalala, and both were informed that such a news article would be published by the agency. (Exh. 8.) See SSM Rehabilitation Inst. v. Shalala, 68 F.3d 266, 270-71 (8th Cir. 1995) (Secretary’s contemporaneous interpretation took precedence over post hoc "clarification").

In short, from the first RoC in 1980, through the published revisions in September 1996, and continuing in the 8th RoC published in May 1998, the agency consistently followed the plain wording of the criteria for the "known" category in requiring sufficient evidence from studies in humans indicating a causal relationship. Now, in preparing the 9th RoC and the dioxin listing change, the agency is engaging in a radical departure from its established rules, insisting that it can list a substance in the "known" category on the basis of "any" information it considers "relevant" as a matter of "judgment". In the case of dioxin, such information includes evidence from studies in experimental animals and studies of cell cultures and tissue samples in vitro.

The agency has not given any reasoned explanation or justification for this radical change of course. It has not attempted to explain how this new position can be reconciled with the plain wording of the 1996 notice of revisions and the agency’s unmistakable contemporaneous "interpretation" of those revisions as having made no substantive revisions in the criteria for the "known" category. Nor has the agency attempted to explain how its new position would comport with clear intent of the legislation, and the congressional intent evident in the legislative history materials, that there should be a clear and unmistakable distinction between the "known" and "reasonably anticipated" categories.21 In fact, the record is devoid of any explanation at all, much less a reasoned explanation.

In its April 19, 1999 Federal Register notice confirmning its new position, the agency stated that this position, as contained in the "final paragraph of the criteria", has applied to the "known" category "[s]ince these criteria were first published on September 26, 1998 . . . ." [Sic – presumably 1996]. As Plaintiffs have demonstrated conclusively, that statement is inaccurate. The plain wording of the 1996 notice, the Departmental press release issued the same day, the contemporaneous news article in the agency’s journal, and the listings in the 8th RoC published in May 1998, all belie the contention that the final paragraph concerning "judgment" regarding "all relevant information" has applied to the "known" criteria since the September 1996 notice of revisions. Publication of a dioxin listing in the "known" criteria since the September 1996 notice of revisions. Publication of a dioxin listing in the "known" category in the 9th RoC based on evidence other than evidence from studies in humans sufficient to indicate a causal relationship would be the first time a "known" listing has been based on such evidence. There have never been any published listings in the "known" category of the RoCs that were based on "limited" evidence from studies in humans, supplemented by "other relevant information", as is proposed for the dioxin listing in the 9th RoC.

It is a "settled tenet" of administrative law that an agency acts arbitrarily and capriciously when it fails to provide a reasoned justification for a change in a settled policy. DSE, Inc., supra, at 31; Troy Corp., supra, at 286-87 (D.C. Cir. 19997); Pontchartrain Broadcasting, supra, at 185; Graphic Communications, supra, at 1493 See also Pfizer, supra, at 1510. In this case, the agency has done precisely that.

C. The agency’s proposed listing of dioxin is arbitrary and capricious because the principal evidence concerning studies in humans upon which it relies is not accurately presented and the agency’s explanation for its decision runs counter to that evidence.

The agency’s rationale for the proposed dioxin upgrade in the DBD (Compl. Exh. 9) relies almost exclusively on two factual assertions concerning the human evidence: (1) IARC "noted" a causal relationship in two highly-exposed groups of chemical plant workers (p. RC-1), and "[i]n its summary . . . identified a causal association" (p. 3-2); and (2) a "new study" by Bertazzi et al., cited as "in press" as of 1998, contained "additional findings [that] were not considered in the IARC evaluation and [that] further strengthen the association between dioxin-exposure and human cancer" (p. RC-1.).22 Not only does the agency’s rationale for the proposed upgrade omit any findings consistent with its criteria, and rely on irrelevant evidence, but both of these evidentiary representations at the heart of the agency’s rationale are inaccurate, and leave the agency without any adequate rationale in the administrative record based on evidence from studies in humans. An agency decision which does not establish a reasonable connection between the facts found and the decision made is considered arbitrary and capricious. State Farm, supra, at 43; Troy, supra, at 288.

1. IARC did not find that the evidence from studies in humans was sufficient to indicate relationship; instead, it specifically found that the evidence was "limited", and therefore not "sufficient" to indicate a causal relationship.

IARC did not either "note" or "identify" a "causal relationship" or "causal association". In its final "Evaluation", IARC concluded that "[t]here is limited evidence in humans" for carcinogenicity. (Exh. 7, pp. 342 and 631, italics as in original, indicating a defined term.). Under the IARC terminology, "limited" evidence means that "[a] positive association has been observed . . . for which a causal interpretation is considered . . . to be credible, but chance, bias, or confounding could not be ruled out with reasonable confidence." (Exh. 4 at p.24.) This is in clear contrast to a finding of "sufficient" evidence, which means that "a causal relationship has been established. . . in studies in which chance, bias, and confounding could be ruled out with reasonable confidence." Id. As discussed earlier, this IARC terminology for the strength of the evidence corresponds, by the agency’s express statements and by consistent application, to the terminology used in the listing criteria and in the actual RoC listings. Under both the IARC and the RoC terminology, "sufficient" human evidence is required to indicate a causal relationship.

The only page reference given to the IARC review to support the DBD statement that IARC found a causal relationship is to pages 192-193 of the IARC monograph (appended to the DBD). The only pertinent statement at those pages appears to be the following:

Thus this study [of the BASF accident cohort] and the Boehringer-Ingelheim cohort together provide dose-response evidence supporting a causal relationship between 2,3,7,8-TCDD exposure and mortality from all cancers combined. Further subdivision of the data indicates that the positive dose-response was restricted to smokers.

This statement does not amount to an IARC finding of a causal relationship by any stretch of the imagination. First, it indicates only that evaluations of these two occupational exposure groups – out of the subcohorts from the "four high-exposure industrial cohorts" on which the agency relies

(Compl. Exh. 9 at p. RC-1, emphasis added) – provide some evidence for a causal relationship. Second, dose-response is considered as only one of many factors routinely considered in determining causality (e.g., chance, bias, confounding, strength of association, consistency of positive findings of association, coherence of findings,m specificity of disease site, dose-response, temporality, and biological plausibility). (Exhs. 5 and 6 and fn. 17, supra.) In the end, IARC found that the human evidence was not sufficient to support a finding of causal relationship, as stated at pages 342-43 and 631 of the monograph. (Exh. 7.)23

Instead of supporting a listing in the "known" category, the IARC evaluation of the evidence from studies in humans as "limited" (i.e., less than "sufficient"), instead supports a listing of dioxin in the agency’s "reasonably anticipated" category (its current listing). The agency’s reliance on the IARC evaluation therefore lacks a rational basis, thereby making the agency’s decision arbitrary and capricious.

2. The recent Bertazzi et al. study does not contain any significant new, Post-IARC findings, and does not support the agency’s position.

Likewise the DBD statement (Compl. Exh. 9 at p. RC-1) that the findings contained in an "in press" paper by Bertazzi et al. were not considered by IARC and support the proposed dioxin listing is inaccurate. This statement, moreover, was not corrected by the agency even after the inaccuracy was called to its attention and the DBD and the proposed listing were "re-reviewed". IARC did in fact consider the Bertazzi et al. findings. Those findings were simnply re-stated in a second article published following the IARC proceedings. And the Bertazzi et al. findings, as stated in both publications (before and after the IARC proceedings) do not support a listing in the "known" category.

IARC consideration of the Bertazzi et al findings is reflected at pp. 161-62, and Table 33 at pp. 146-47, of the IARC monograph. (Exh. 9.) The full reference appears at p. 531 of the monograph. (Id.) The agency’s own DBD references include this 1996 publication as "Cited by IARC, 1997" at p. 8-2. (Exh. 10.) IARC considered the Bertazzi et al. findings as contained in the 1996 article in Organohalogen Compounds. The 1998 "in press" article by Bertazzi et al. referred to in the DBD was a publication in Epidemiology, which, as shown by the updated references in the 1998 DBD Supplement (Compl. Exh. 21 at pp. 2,3), appeared in 1997. Copies of both the 1996 and 1997 Bertazzi et al. publications are attached as Exhibits 11 and 12, respectively.

From a comparison of the 1996 and 1997 Bertazzi et al. publications, it is clear that they contain basically the same findings, and the 1997 publication is simply a more detailed version of the 1996 article. In particular, the findings for relative risk ("RR") for "all cancers" for both males and females are exactly the same in both publications. Thus, IARC did consider the Bertazzi et al. findings, contrary to the statement in the DBD.

Moreover, the Bertazzi et al. findings do not strengthen support for the proposed listing, as also asserted in the DBD. The DBD summary relies on the human evidence as supposedly showing "increased risk" for "overall cancer mortality" or "all cancers combined". The Bertazzi et al. findings do not support this evaluation. The 1996 and 1997 Bertazzi et al. publications found no significant elevation in relative risks for all cancers combined. More specifically, they found, in exposure zones A, B, and R respectively, for males: RRs of 0.4, 1.1, and 0.9; for females: RRs of 1.2, 0.9, and 0.9.24 The narrative discussion in both publications confirms that these findings do not support the proposed dioxin listing. The 1996 publication characterized the findings as being "far from conclusive because of some inherent limitations of the study" and concluded that the findings of this study, along with others, "corroborate the hypothesis" that cancers "might be associated with dioxin exposure". In the introductory abstract for the 1997 publication, the authors state that the evidence for carcinogenicity in humans is "limited", that they "found no increase for all-cancer mortality or major specific sites", and that for some specific excess cancers, only an "association" with dioxin exposure was "plausible".

That IARC had in fact considered these recent Bertazzi et al. findings, and that the DBD was inaccurate in this regard, was pointed out to the agency by Dr. Raymond Greenberg, Provost of the Medical University of South Carolina and an expert epidemiologist who was an invited observer and participant (but not a voting member) in the 1997 IARC working group. Dr. Greenberg’s 1998 written comments (for the dioxin "re-review") and the transcript pages of his 1997 comments to the RoC Subcommittee are Exhibits 13 and 14. The agency’s continued reliance on the Bertazzi et al. publications, as well as the IARC evaluation, is surprising to say the least, not only because they were advised of the errors by third parties, but because Defendant Lucier was Chairman of the IARC working group.25

3. The record contains no indication that the agency considered anImportant aspect of the dioxin listing proposal, namely; that the studies it was relying on involved mixed occupational exposures, while the proposed listing information focused on exposure to dioxin apart from such mixtures, particularly as a contaminant in food.

This aspect of the problem was called to the attention of the agency by Plaintiff Tozzi. (Compl. Exhs. 15,16.) Defendant Olden’s response Compl Exh. 17) was that the agency reviewers were aware of this issue, but they had decided to focus only on exposure to dioxin as an isolated substance. This decision by the reviewers was not explained. They agency’s criteria explicitly require listing in the "reasonably anticipated" category when confounding cannot adequately be excluded. (Compl. Exh. 1.) In addition, the IARC evaluation which the agency relies on concluded that confounding could not be excluded as a possible explanation for the observed association. (Exh. 7 at p.337.)

An agency’s failure to consider an important aspect of the challenged action is considered to be arbitrary and capricious. State Farm, supra, at 43; Troy, supra, at 288.

The administrative record presenting the rationale for the proposed upgrade of dioxin does not contain any findings consistent with the listing criteria, relies on irrelevant factors (animal and other evidence that is not from studies in humans), relies on statements regarding two evaluations of human evidence (the IARC evaluation, and the 1997 Bertazzi et al. publication) that are manifestly incorrect, and fails to address an important aspect of the proposed action. Because the proposed listing is contrary to the agency’s own rules and does not have rational support in the administrative record, it is clear that the proposed agency action is "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law", and/or "without observance of procedure required by law", under the applicable standards for judicial review. See 5 U.S.C.§§ 706(2)(A),(D). There is, therefore, a strong likelihood, if not a near certainty, that Plaintiffs will prevail on the merits.

II. The agency’s final position on the meaning of its listing rules, and its proposed application of that position to upgrade dioxin to the "known" category, threaten Plaintiffs with imminent irreparable harm.

Upgrading dioxin to a "known" human carcinogen and describing ordinary consumption of meats, fish, and dairy products as the primary source of exposure threatens Plaintiffs’ food service businesses with irreparable harm. The specific aspects of such threatened harm have been described in detail in Plaintiffs’ Complaint and their response to Defendants’ motion to dismiss. Those allegations are incorporated by reference and will not be repeated here. Simply put, the agency’s position on the meaning of its listing criteria, as applied to the proposed dioxin listing, will create public confusion and concern over the safety of the food Plaintiffs serve. This public confusion and concern will lead to some degree of consumer aversion to meats, fish, and dairy products Plaintiffs serve, thereby reducing sales. Plaintiffs will also incur increased costs associated with attempting to dispel such public confusion and concern, as well as disruptions and increased costs in their food supply chain.

The essence of threatened "irreparable" harm is that such harm is not susceptible of repair through later judicial relief in an action for monetary damages. Serono Laboratories, Inc. v. Shalala, 974 F. Supp. 29, 35-36 (D.D.C. 1997) (approval of competitor’s product enjoined where approval could result in loss of market share by plaintiff which was only speculative, but which could not later be compensated); Bracco Diagnostics v. Shalala, 963 F. Supp. 20, 29 (D.D.C. 1997) (enjoining FDA

disparate treatment of two similar products); Ross-Simons of Warwick, Inc. v. Baccarat, Inc., 102 F.3d 12, 18-19 (1st Cir. 1996) ("It is usually enough if the plaintiff shows that its legal remedies are inadequate."). In the present case, Plaintiffs could not later seek compensation in a suit for monetary damages against a U.S. government agency.

Because the purpose of injunctive relief is to prevent future harm, the assessment of such harm is inherently somewhat speculative and non-quantifiable. Serono, supra; Barnstead Broadcasting Corp. v. Offshore Broadcasting Corp., 865 F. Supp. 2, 7-8 (D.D.C. 1994) (agency’s allowance of filing of objection to permit threatened plaintiff with loss of business opportunities); Reuters, Ltd. v. United Press Int’l, 903 F.2d 904, 907-09 (2d Cir. 1990) (preliminary injunction granted where threatened loss of photo supply services was nearly impossible to value and was recognized as a speculative loss); Gerard V. Almouli, 746 F.2d 936, 939 (2d Cir. 1984) (preliminary injunction granted because damages from violation of exclusive distributorship agreement would be impossible to quantify, especially since there was not yet any developed market for the product); MGM-Pathe Communications Co. v. Pink Panther Patrol, 774 F. Supp. 869, 873 (S.D.N.Y. 1991) (preliminary injunction granted where harm to plaintiff’s image would be "difficult to measure and compensate").

Threatened damage to business reputation for furnishing safe products has been recognized as irreparable harm sufficient to justify a preliminary injunction. Honeywell, Inc. v. Consumer Product Safety Comm’n, 582 F. Supp. 1072, 1078 (D.D.C. 1984) (enjoining threatened disclosure to the public of documents pertaining to the safety of gas controls) (citing GTE Sylvania, infra, and Dow Chemical, infra); GTE Sylvania, Inc. v. Consumer Product Safety Comm’n, 404 F. Supp. 352, 373 (D. Del. 1975) (enjoining agency release of safety information due to agency failure to comply with procedures designed to ensure accuracy of such information). Federal courts have recognized this type of potential harm as sufficient to support injunctive relief in cases specifically involving agency listing of a product as a carcinogen. Dow Chemical, USA v. Consumer Product Safety Comm’n, 459 F. Supp. 378, 394-95 (W.D. La. 1978) (preliminary injunction granted to stay classification of product as Class A carcinogen); Synthetic Organic Chem. Mfrs. Ass’n v. Secretary, Dept. of Health and Human Serv., 720 F. Supp. 1244, 1257 (court stated it would "willingly act" to enjoin proposed listing of substance in the RoC if plaintiffs could show a likelihood or prevailing on the merits, which they had not). Potential consumer confusion arising from trademark infringement, unfair competition, or stigma that could damage business reputation for quality of product or service is a recognized basis for injunctive relief. Malarkey-Taylor Assoc., Inc. v. Cellular Telecommunications Indus. Ass’n, 929 F. Supp 473, 478 (D.D.C. 1996); National Rural Electric Coop. Ass’n v. National Agric. Chem. Ass’n, 26 U.S.P.Q.2d 1294, 1297-98 (D.D.C. 1992).

Potential damage to business reputation has been found to justify injunctive relief even when no likelihood of specific monetary loss had been shown. See, e.g., America West Airlines V. National Mediation Bd., 969 F.2d 777, 781 n.5 (9th Cir. 1992) (upholding preliminary injunction against Board’s threatened allegation of unfair labor practices that could harm plaintiff’s reputation in the eyes of its employees); Patriot v. HUD, 963 F. Supp. 1, 5 (D.D.C. 1997) (plaintiffs’ reputation would be damaged by agency letter characterizing themn as engaging in unethical mortgage marketing practices) (citing Honeywell, supra). The courts have also held that an agency’s failure to follow required procedures justifies injunctive relief where the agency action could result in an irretrievable commitment of public resources, even though plaintiffs have not demonstrated a likelihood of serious economic harm to their businesses. Mt. Airy Ref. Co. v. Schlesinger, 481 F. Supp. 257, 284 (D.D.C. 1979) (failure to comply with environmental assessment requirements of National Environmental Policy Act).

In their Complaint and their Opposition of Defendants’ Motion to Dismiss, Plaintiffs presented examples of media and federal and local government reaction to news of dioxin contamination in food. Plaintiffs’ purpose was to show how the media, the public, and the government react when there is news of dioxin contamination in foods, and the negative effects such reactions have on food suppliers, including restaurants. Defendants have attempted to challenge such real-world examples as irrelevant to this case with a number of arguments, and have suggested that there is no reasonable potential for adverse media, consumer, and government reaction to the threatened dioxin listing. Defendants’ attempted distinctions are diversionary or simply incorrect, as discussed below. Contrary to these unsupported assertions, there is a very real likelihood of adverse consumer, media, and government reaction to the threatened listing that will affect Plaintiffs’ businesses.

First, Defendants assert that any comparison with the recent Belgian/European Union dioxin scare is inappropriate because the contamination in that instance was reported to have arisen from unusually high levels of dioxin in animal feed. (Reply Mem. At 9-10.) The response to this is that, while some articles mention high levels in feed as the source of the contamination, the news articles are devoid of any information as to whether there were "elevated" levels in the consumer food products themselves. Instead, the articles indicate that consumers and the government reacted dramatically based simply on the possibility of some dioxin contamination in consumer food products. Moreover, once it was reported that there was some level of dioxin contamination in the food supply, the Belgian government and the European Union quickly lost all control over the situation, and were powerless to confine the adverse reactions to certain consumer food sources that were the only one that might have been affected by the reportedly contaminated feed. This is not different from the case at hand, in which the Defendants are preparing to state that elevated levels of dioxin from various sources are contributing to the widespread presence of dioxin in consumer food products in the United States, and that such exposures are known to cause cancer. Plaintiffs also note that Defendants focus almost exclusively on government technical advisories, which the general public is unlikely to read, rather than on the kind of information that is disseminated by the media and activist organizations.

Defendants point out that the RoCs will "say nothing about what levels of dioxin are or are not harmful or safe." (Reply Mem. at 10.) They also state that the RoCs "certainly will not state or imply that simply eating meats, poultry, fish and dairy products causes cancer." Id. To the contrary, the threatened RoC listing of dioxin clearly would imply that eating such foods is "known" to cause cancer, and it is precisely because the RoCs do not say anything about what levels in such foods are safe or harmful that many members of the public are likely to be confused and to believe that they should be concerned that any level is considered harmful by the agency. Plaintiffs note that, as shown by Exhibits 15 and 16, the agency has already expressed to the public concern over low levels of dioxin in foods, although so far it has characterized dioxin to the public as only potentially a carcinogen. It is reasonable to assume that when there is an official agency pronouncement that dioxin in foods is "known" to cause cancer, in connection with stating that the primary source of exposure for U.S. residents is meats, fish, and dairy products, there will be additional expressions of increased concern and adverse consumer reaction.

Defendants assert that "the public already clearly believes" that "dioxin per se" is a known human carcinogen, thereby implying that no harm could occur from simply confirming this "belief" in the RoC. Defendants cite no support for this assertion. Plaintiffs have no idea what Defendants intend to assert in referring to dioxin "per se", but it is clear that the public does not currently believe that the meats, fish, and dairy products they currently consume are contaminated with a substance known to cause cancer. This belief has been engendered at least in part by past U.S. government statements that such exposures are not a matter of concern. (Pl. Opp. Exhs. 1, 20; Def. Reply Exhs. B5, B8.)

Defendants also assert that federal agencies already monitor and test the U.S. food supply in order to avert or minimize dioxin contamination. Def. Reply at 10-11. To the best of Plaintiff’s knowledge, this assertion is incorrect. The only support Defendants cite for this assertion is to a "nationwide survey" conducted by FDA, USDA, and EPA (Def. Reply Exhs. B1-B6). All that appears from such exhibits is that there was a single survey conducted, consisting of 80 poultry samples and an unspecified number of catfish samples. There is no evidence of a continuing, periodic, nationwide government testing and monitoring program. In addition, those exhibits show that the government imposed testing requirements on food suppliers, rather than carrying out all testing and monitoring through government programs.

Defendants also assert that because government agencies have already found dioxin in food, and no public food scare has resulted, such a food scare is unlikely to result from publication of the new dioxin listing in the RoC. Such test results, however, have only been announced in government technical advisories. As shown vividly by the Belgian/EU incident, the government cannot control how the media and activist groups will characterize government findings, particularly once they are widely disseminated to the general public through a publication such as the RoC, on the Internet, and through advertising. Plaintiffs have provided exhibits showing that activist groups are eagerly awaiting some government confirmation that dioxin in foods at current ordinary exposure levels is a clear health threat. (Exhs. 25, 27.) Such groups can be expected to make the most of any new government information on the subject, particularly if it is in such general terms and as widely disseminated as the RoCs. Plaintiffs have even shown that, in the case of fish from San Francisco Bay, earlier committee votes on the dioxin listing from the 9th RoC process have already been used in influence public opinion and to generate local government regulatory action. (Compl. Exhs. 5, 6.)

Until now, U.S. government agencies have been equivocal about whether dioxin in common foods at levels frequently encountered is a matter of concern. On the one hand, for example, U.S. officials have stated that they believe that any exposure of U.S. consumers to harmful levels of dioxin in common foods is "minimal" (Exh. 15), that consumers can continue to eat food such as catfish "with complete confidence" (Def. Reply Exh. B8), and that "Americans enjoy the safest food supply in the world . . ." (Pl. Opp. Exh. 20). On the other hand, U.S. government agencies have told the public that dioxins "may be carcinogens at low levels of exposure" (Exh. 16), that "[s]ince there are no tolerances or other administrative levels for dioxins in food . . ., the appearance of these compounds in a food . . . is of gravest concern" (id.), and the have recently "recommended" that food imports from Belgium be tested and demonstrated to be "free of detectable levels" of dioxins (Def. Reply Exh.A2). As Defendants have shown through their own assertions and exhibits, testing of common foods for dioxin will often show that there are significant detectable levels. (Def. Reply at 12.) As a result of these many mixed messages, and the government’s inability to control how new information disseminated in the RoCs will be handled by the media and activist groups, it is likely that news that Defendants have upgraded dioxin to a "known" human carcinogen, and have identified meats, fish, and dairy products as the primary source of exposure, will lead to consumer confusion and concerns affecting Plaintiffs’ businesses.

III. The imminently threatened harm to Plaintiffs if a preliminary injunction is not granted clearly outweighs any possible harm to Defendants.

Granting of preliminary injunction would not injure Defendants in any foreseeable manner. To the contrary, enjoying publication pending final judicial resolution on the merits would be likely to save Defendants substantial effort and expense involved in recalling and correcting the Report and otherwise publicizing corrections to the dioxin listing. Moreover, assuming a food scare were to occur as a result of publication of the Report as proposed, both HHS and USDA would likely to deluged with, and have to respond to, a multitude of public and congressional inquiries concerning the health risks, and measures that should be taken to mitigate those risks. The government might also be forced into burdensome food sample testing and regulatory activities to allay public fears.

As discussed below, Defendants will not be otherwise adversely affected in an administrative sense (i.e., carrying out their duties under the RoC legislation) by a delay in the publication of the 9th RoC, since they are well ahead of the congressionally-directed publication schedule.

IV. The public interest clearly favors entry of a preliminary injunction.

Public interest weighs heavily in favor granting the injunction Plaintiffs seek.

The accuracy of the RoCs should be the paramount concern. The need for dissemination of information on carcinogenicity that is accurate far outweighs any interest in publishing public health information which is inaccurate and of doubtful relevance.

The Plaintiffs, the public, and even the Defendants themselves have a strong legitimate interest in ensuring that the integrity of the administrative process is maintained by requiring that agencies to whom Congress has entrusted the implementation of its directives adhere to the rules they have promulgated in order to implement those directives. This is especially true in this case, which involves dissemination of information on the safety of a large portion of the U.S. food supply.

Moreover, if the 9th RoC were to be published, and then had to be recalled due to an adverse judicial ruling on the dioxin listing, the credibility of the whole RoC program would be significantly damaged. It is in the public interest to ensure that public confidence in the accuracy of the Reports is maintained.

As noted earlier, these credibility issues do not involve just dioxin. The agency’s position on the meaning and application of the listing criteria for the "known" category potentially affects numerous other important substances also under review for the 9th RoC. Resolution of the issue concerning correctness of the agency recent change in position on the meaning of the "known" criteria prior to publication would promote administrative and judicial economy and ensure the quality of information disseminated to the public.

Finally, there is no reason for the agency to proceed with publication in the near term. The congressional directive is for the RoCs to be published every two years.26 The 8th RoC was published in May 1998.27 Therefore, even with strict compliance, the 9th RoC is not due to be published until May 2000. But the agency, as a matter of practice, has often allowed substantially more time to elapse than directed by Congress. Even when the Reports were supposed to be published annually, several were published at two- to four-year intervals, and the first Report to be published since the law was amended to direct publication every two years, the 8th RoC, was published after a four-year interval.28

The public interest favors careful consideration of the accuracy of the 9th RoC, and a delay in the agency’s currently anticipated publication schedule would be more consistent with congressional intent.


The factors to be considered in reviewing Plaintiffs’ application for a preliminary injunction weigh heavily in favor of granting the application. The administrative record shows a clear violation of the agency’s own listing rules in the case of the proposed dioxin listing upgrade. That violation imminently threatens to create a food scare that will result in irreparable injury to Plaintiffs. Issuance of a preliminary injunction pending full resolution on the merits will result in no harm to Defendants, and will be in the interests of both Defendants and the public.

The Court should therefore grant the injunctive relief Plaintiffs seek, as set forth in the attached proposed Order.

1 Also referred to as Section 301(b)(4) of the Public Health Service Act.

2 The dual term criteria/rule is used here because, while the agency refers to these listing category definitions as "criteria", those definitions are clearly "rules"as that term is defined under the Administrative Procedure Act, 5 U.S.C. § 551(4). The APA states that "'rule' means the whole or a part of an agency statement of general or particular applicability and future effect designed to implement, interpret, or prescribe law or policy . . . ." The APA provides for review of final "agency action", and that term "includes the whole or a part of an agency rule . . . ." 5 U.S.C. 551(13). Throughout the remainder of this memorandum, the term "criteria" is used with the understanding that the "criteria" are agency "rules" within the meaning of the APA.

3 Exhibit 1 is a printed, bound copy of the 8th RoC, and it is separated from the rest of the exhibits in a labeled brown folder.

4 The 1981 listing of dioxin was contained in the Second Annual Report on Carcinogens. That listing has been updated with new information, as reflected in the 8th Report on Carcinogens (1998) at pp. 195-96.

5 Hereafter, HHS, NTP and the National Institute of Environmental Health Science ("NIEHS"), which administers the NTP, and the other HHS agencies participating in the NTP, will ordinarily be referred to collectively as "the agency".


7 It is noteworthy, although not a central issue here, that while mechanistic data is ne of the three types of evidence expressly relied on by the agency does not assert that the mechanism is known to be a mechanism of carcinogenicity, but rather a mechanism that induces certain "biochemical and toxicological responses" in both humans and animals. Compl. Exh. 9.

8 This non-governmental expert peer review panel voted against the proposed upgrade even on the basis of the agency's DBD rationale, which is herein alleged to be arbitrary, capricious, an abuse of discretion and otherwise not in accordance with law. This vote by itself would be sufficient grounds for challenging action to list a substance as "known" to cause cancer in humans. If a majority of a group of outside experts cannot support the agency's rationale, flawed as it is, how could dioxin be considered "known" to cause cancer under any sensible interpretation of the term "known"? "Known", if it is to have any significance, must indicate a high degree of certitude. In the present case, however, the agency has provided a controlling definition of the term "known" which should be dispositive.

9 If there were any doubt concerning the meaning of the two April clarifications, it was removed by a statement in an editorial authored by Defendant Lucier in the May 1999 issue of Environmental Health Perspectives. In his editorial, titled "Why Not Use It All?", Dr. Lucier stated that NTP has been "explicit" in statements "that mechanistic data can be used to upgrade the results of inconclusive epidemiology data as well as to downgrade." Vol. 107, No. 5, at A232. (Exh.2.) Since the agency's contemporaneous explanations of the criteria revisions issued in 1996 are to the contrary (Compl Exhs.2,3), Defendant Lucier's reference to explicit statements must be to the April 1999 'Federal Register "clarifications".

10 According to the description of listing/delisting procedures contained in the 8th RoC (Exh. 1), at 232, "[t]he Director, NTP [currently Olden] receives the four independent recommendations from RG1, RG2, NTP Board RoC Subcommittee, and the NTP Executive Committee and makes the final decision regarding the proposed listing and/or delisting and submits the RoC to the Office of the Secretary, DHHS [for "approval"]." [Emphasis added.] The description does not inform the reader that the Director, NTP is also a voting Member of the NTP Executive Committee. In the present case, Defendant Lucier, apparently acting as Defendant Olden's alternate on the NTP executive Committee, reportedly voted in favor of the proposed dioxin upgrade. It thus appears that the Director has already made his final decision.

11 Only selected portions of the DBD were attached to the Complaint as exhibits, while the full Supplement was attached. The full DBD consists of approximately 500 pages of text. Of that, approximately 450 pages consist of a reproduction of pertinent portions of the 1997 IARC monograph and two short appendices. The first approximately 50 pages of the DBD contain a more detailed discussion of the several types of scientific evidence referred to in the Summary.

12 See fn. 10, supra.

13 The term "in humans" does not exclude the use of data gathered from examining biological samples taken from whole humans who were exposed to the substance while living. Epidemiologic studies routinely involve taking blood, fluid, or tissue samples from exposed humans, sometimes post mortem. The crucial point is that the evidence must be evidence of what occurred in living humans who were exposed to the agent, substance or mixture, as opposed, for example, the human cells cultured and exposed in a laboratory.
A review of various RoCs prior to the 8th (1998), will show that the summaries, or "profiles", of the evidence supporting in a listing in an RoC are consistently organized into separate sections addressing evidence from studies "in animals" and evidence from studies "in humans" (Exh. 3) After the criteria revisions in 1996, in the 8th RoC (Exh. 1), a third separate category was added titled "Other Information Relating to Carcinogenesis or Possible Mechanisms of Carcinogenesis". This third category includes data on laboratory experiments in vitro with cultured cells and other biological materials.
In short, the plain wording of the criteria, and the agency's consistent practice, embody a clear distinction between evidence from studies in humans as distinct from evidence from studies in animals or in vitro. These same distinctions are also reflected in the IARC monographs, on which the agency has consistently relied. (Exh. 4)

14 In their reply brief, at p. 3 fn.2, Defendants assert that the agency has "consistently interpreted" its criteria for the "known" category to allow reliance on mechanistic evidence, which it describes as "data drawn from the study of human cells, tissues and other matter apart from the study of a living human being, as in epidemiological studies." Defendants do not provide any support for this assertion, and the record indicates that, to the contrary, Defendants have never previously (and certainly not consistently) relied on mechanistic data to compensate for evidence previously (and certainly not consistently) relied on mechanistic data to compensate for evidence from studies in humans that was not sufficient to indicate a causal relationship. All listings in the "known" category in the previous eight editions of the RoCs are based on findings either that there is "sufficient" evidence from studies in humans (epidemiologic studies), or that there is evidence from studies in humans which "indicates a causal relationship". There are no "known" listings based on less than sufficient evidence from studies in humans supplemented by mechanistic data.

15 Introduction to the 7th RoC at pp. 5, 6 (Exh. 3). The IARC definitions of the terms "sufficient" and limited" closely track the agency's "known" and "reasonably anticipated" criteria, respectively. A copy of the pages from the Preamble for the 1997 IARC monograph on TCDD explaining its evaluation terminology and classification criteria is attached as Exhibit 4. It will be noted from this Preamble that the IARC criteria for i8ts Group 1, "carcinogenic to humans" (p.26) is materially different from the agency's criteria for the "known" category in allowing such a classification to be based on "less than sufficient" evidence in humans when there is "strong evidence in exposed humans that the agent (mixture) acts through a relevant mechanism of carcinogenicity." IARC altered its Group 1 criteria in 1991 to add this language that created the divergence from the agency's "known" criteria. This divergence is what is referred to in the statement in the 7th RoC (in 1994) that the two agencies' classification schemes do not correspond exactly, but that the terminology concerning degrees of evidence continued to correspond.

16 For example, compare the profile for asbestos in the "known" category, based on "sufficient evidence" for carcinogenicity in humans, citing an IARC evaluation (pp.19-20), with the profile for cadmium and certain cadmium compounds in the "reasonably anticipated" category, based on "limited" evidence in humans, citing IARC evaluations (pp. 69-70).

17 See, e.g., Smoking and Health, A Report of the Advisory Committee to the Surgeon General of the Public Health Service, 19-21, 182-89, 196 (Dept. Of Health, Education, and Welfare 1964); The Health Consequences of Smoking - Cancer - A Report of the Surgeon General, 4-5, 16-20 (U.S. Dept. of Health and Human Services 1982). (Cited pages attached as Exhs. 5 and 6.) Basic texts on epidemiology also emphasize this distinction. See, e.g., David E. Lilienfeld and Paul D. Stolley, Foundations of Epidemiology, section on "Assessing Causality", at pp. 263-67 (3d ed. 1994), and Kenneth J. Rothman and Sander Greenland, Modern Epidemiology, section on "Causal Criteria", pp. 24-28 (2d ed. 1998).
To illustrate the distinction between an association and a causal relationship, epidemiologists sometimes employ the following example of an association masquerading as a causal relationship: I get colds in the winter. I wear galoshes in the winter. Therefore, galoshes cause colds.

18 As a result of the consistent usage of "sufficient" in connection with "causal relationship", either term conveys the full meaning of the two together.

19 P-Chloro-o-toluidine and p-Chloro-o-toluidine Hydrochloride were listed together as reasonably anticipated on the basis of "limited" evidence from studies in humans. (Exh. 1 at pp. 80-81.)

20 The record also does not show that the agency adequately considered the "relevant factors" for considering whether a causal relationship is indicated which are set out in the general epidemiologic literature and the agency's own reports (see fn. 17, supra).

21 In explaining the bill which was later enacted, Representative Rogers, who was Chairman of the Subcommittee that initiated the legislation, stated that "[T]he report shold be properly organized so that no possible confusion could exist between clearly demonstrated carcinogens and those for which convincing data are not yet available . . . ." Cong. Rec. - House, Oct. 10, 1978, p. 34938.
"Knowledge" is distinct from "judgment", and if "judgment" can be based on any evidence considered "relevant", then it is difficult to perceive how, under the agency's new position, there would be any meaningful, objective distinction between the "known" and "reasonably anticipated" categories.

22 Note here also the agency's reliance on "association" rather than "causal relationship".

23 IARC (with a large group of U.S. agency scientists participating in the decision) nevertheless upgraded dioxin to its Group 1 category ("carcinogenic to humans" - IARC does not use the term "known") on the basis of mechanistic data. This is expressly provided for in the IARC criteria, unlike the RoC criteria. The IARC Group 1 criteria state: "This category is used when there is sufficient evidence of carcinogenicity in humans. Exceptionally, an agent (mixture) may be placed in this category when evidence in humans is less than sufficient but there is sufficient evidence of carcinogenicity in experimental animals and strong evidence in exposed humans that the agent (mixture) acts through a relevant mechanism of carcinogenicity." (Exh. 4 at p.26, italics in original, indicating a defined term). The agency's "known" category criteria do not contain similar language. There is, however, similar language in the third paragraph of the RoC criteria for the "reasonably anticipated" category. Thus, IARC's upgrade cannot provide a rationale for the agency's proposed upgrade.

24 These numbers are from the two tables in the 1996 publication and Tables 3 and 4 in the 1997 publication. "RR" stands for relative risk, meaning the incidence of cancert in the particular exposed population relative to an unexposed population. Both publications also show that even the two zone-specific RRs (out of the six reported) that slightly exceeded 1.0 (1.1 and 1.2) were not statistically significant based on the 95% confidence intervals ("CI") provided.

25 Also, as Dr. Greenberg pointed out, Dr. Bertazzi was a member of the IARC working group.

26 More frequent publication, as would occur if the 9th RoC were published, for example, in September 1999, less than a year and a half after the 8th RoC, would appear to be contrary to Congressional intent. The law was amended to change the publication interval from annual to biennial in June 1993, pursuant to P.L. 103-43, Title XX, sec. 2009; 107 Stat.213.

27 Availability of the 8th RoC was announced in the Federal Register on May 14, 1998, 63 Fed. Reg. 2618-20.

28 The dates of publication for the eight editions, as reflected on p. 12 of the 8th RoC, have been 1980, 1981, 1985, 1989, 1991, 1994, and 1998 [p. 12 of the 8th RoC gives 1997 as the year of publication for that 8th edition; however, as noted above in fn. 27, that 1997 date for the 8th RoC is incorrect - it was not released until May 1998].