Oral argument is scheduled for November 12, 1999. Defendants admit that the 9th Report on Carcinogens (RoC), could be published as early as November 1999. (Def. Opp. at 11 n.6).

Defendants do not contest the basic principles of law applicable to preliminary injunction, with two exceptions. First, while not overtly contesting the well-established law in the D.C. Circuit that the court should balance the four elements necessary for a preliminary injunction, so that a strong showing on one or more elements may offset a relatively weaker showing on the others, Defendants fail to recognize this weighing principle in arguing that the law requires a "clear and convincing" showing of severe irreparable economic injury unique to Plaintiffs, together with a strong showing on the merits. No such showing is required.¹ Second, Defendants contend, with regard to the element of irreparable harm, that Plaintiffs must demonstrate that "severe economic impact" is "certain" to occur. Plaintiffs dispute this characterization of the legal standard.

Defendants also have not disputed the plain meaning of certain critical terms such as "association" (as opposed to "causal relationship"), and "sufficient"(as opposed to "limited"). They do contend, however, that the meaning of "studies in humans" is not plain and is subject to interpretation. (Def. Opp. at 20). They apparently assert that "in humans" can be interpreted to include evidence from studies in experimental animals and in vitro.

Defendants suggest that their views should be given great deference because of the agency’s scientific and technical expertise. (Def. Opp. at 18-19). The issues in this case, however, turn on plain wording, established practice, and published scientific conclusions that are simple and clear. The Court can address and resolve all the issues without any delving into arcane scientific matters.

Defendants also state that "it makes little sense" to review the language of the Draft Background Document for TCDD ("DBD") until such language is finally approved. (Def. Opp. at 2 n.3). Because the language as finally approved will not be available to the public until the 9th RoC is actually published, however, it does indeed make sense to review the agency’s position on the criteria now, since doing so could save the agency much time, effort, and expense, and avoid irreparable harm to Plaintiffs. Plaintiffs also note that Defendants’ brief reconfirms that the agency has taken a reviewable final position on how it will apply the listing criteria to dioxin, and that the agency will not concede any evidentiary inaccuracy in the DBD’s proposed listing rationale.²

Defendants argue at length that the listing criteria for the "known" category permit them to rely on human mechanistic evidence. Even if this proposition were supportable, it is still apparent from the DBD summary (Compl. Exh. 9) that the agency has not satisfied the criteria by finding sufficient evidence of a "causal relationship". Even the statement at RC-1 summarizing the mechanistic data (which combines animal and human in vivo and in vitro data), does not suggest that there is an understanding of a mechanism known to cause cancer in humans. The statement instead describes only "biochemical and toxicological responses" with no mention of cancer. The only references to a finding of a causal relationship are the agency’s references to the IARC monograph, which Plaintiffs have demonstrated are inaccurate and which Defendants now therefore apparently disavow. This failure to find sufficient evidence of a causal relationship from studies in humans is, by itself, grounds for holding that the threatened agency action is not based on the relevant factors set out in the "known" criteria. It is therefore arbitrary and capricious.

As Defendants indicate, their reading of the criteria "would not exactly leap out at even the most astute reader". (Def. Opp. at 18) . However, the reading must be rational and reconcilable with the wording of the criteria, and even the most astute reader cannot read wording that requires sufficient evidence from studies "in humans" as allowing reliance on evidence from studies in experimental animals, and animal and human in vitro data, to compensate for less than sufficient evidence from studies in humans. The agency’s rationale for the proposed listing, clearly stated in the DBD summary, is that it relies on insufficient evidence from epidemiologic studies combined with evidence from studies "in experimental animals" and certain evidence of "a basic similarity in the mechanism of induction of animal and human tissue biochemical and toxicological responses". (Compl. Exh. 9) (emphasis added).³ That these mechanistic data are based largely on animal tissue exposed in vivo and in vitro is clearly stated in DBD section 7.0 (Def. Opp. Exh. 1 at, e.g., 7-3), as well as in the IARC monograph on which the DBD relies. (Exh. 1). Review of the titles of many of the individual study reports cited in DBD section 7.0 and the IARC monograph further confirms that studies of animal tissue substantially comprise the basis of the mechanistic data. (Id.)⁴

Defendants attempt to mask their obvious reliance on animal and in vitro data, which is clear from the first page of the DBD (Compl. Exh. 9), by carefully avoiding any mention of such data in their brief. Instead they argue that (1) "studies in humans" can be interpreted to mean "from human studies", and (2) "human studies" can reasonably interpreted to include human mechanistic studies, including studies of human cells. (Def. Opp. at 20). However, even if "studies in humans" could be reasonably interpreted to encompass in vitro studies, where exposure took place in a laboratory container -- which it cannot -- "studies in humans" cannot reasonably be interpreted to encompass studies in experimental animals in vivo or in vitro.

As Plaintiffs explained in detail, the agency has, until now, consistently read "studies in humans" to mean simply epidemiologic studies, and not to include even studies of human tissue or cells exposed in vitro, much less studies in animals or animal tissue or cells in vitro. (Pl. App. at 16 n.13). The listings in the 8^th RoC (Pl. Appl. Exh. 1) show that the agency has consistently interpreted evidence from studies "in humans" to mean epidemiologic evidence, and has considered mechanistic data and evidence from studies in animals in separate sections of the listings apart from the discussion of (epidemiologic) evidence from studies "in humans". The dioxin DBD also clearly separates "human studies" from both animal studies and mechanistic data. (Compl. Exh. 9 at i, ii, and RC-1). The IARC monographs, on which Defendants have consistently relied, have also has maintained a similar separation between these very different types of evidence. (Exh. 2).⁵ A change in the agency’s position on the meaning of the phrase "studies in humans" would, like a change to not requiring sole reliance on sufficient evidence from studies in humans, be a substantial change in an established agency position that could not be made without public notice and a reasoned justification. The Defendants have not given such notice or reasoned justification, even in the April 2 and 19, 1999 Federal Register "clarifications" (Compl. Exhs. 27, 28).

Defendants’ entire position on the merits turns on the reasonableness of interpreting the final "descriptive paragraph" as substantively modifying the "known" criteria. To support the asserted reasonableness of such an interpretation, Defendants ignore their obvious reliance on animal data and human in vitro data and argue that (1) the 1996 Federal Register notice format was misleading in not preserving the agency’s internal formatting; (2) the "administrative history" of the criteria revisions, as reflected in the unpublished minutes of various meetings leading up to the 1996 criteria revisions, shows that some staff or reviewers intended the final paragraph to allow consideration of human mechanistic data; (3) the use of the words "criteria" and "in humans" in the final paragraph shows that the paragraph was intended to modify "studies in humans" in the "known" criteria; and (4) "logic" requires that the final paragraph be read as requiring consideration of human mechanistic data in connection with the "known" category. None of these arguments has merit.

1. The Formatting

Defendants assert that the wider margins of the final paragraph, as formatted internally at the agency rather than in the Federal Register, indicate that the paragraph was intended to embrace both of the preceding two categories of the criteria. This argument is too far-fetched to be reasonable, particularly when considered with such factors as the complete absence of any wording indicating that the paragraph modified the "known" criteria, the inability to reconcile its language with that of the "known" criteria, the "administrative history" of the development of the final paragraph, and the agency’s own contemporaneous interpretations. When all of these factors are considered, along with the formatting, what appears clear is that the wider (internal) margins on the final paragraph were made wider to indicate that the final paragraph was intended to embrace all three preceding paragraphs of the "reasonably anticipated" criteria.⁶

Defendants argue that the Federal Register formatting did not reflect the agency’s formatting of the final paragraph with the wider margins. If the Federal Register notice was printed in a manner that appeared contrary to the agency’s intent, the agency should have noted this and corrected it. The agency should have known that the Federal Register drafting instructions did not provide for variations in margins unless the agency notified the Archives and Record Administration that such formatting was of substantive significance. The agency should have inserted explicit wording stating that the final paragraph changed the "known" criteria, or issued an immediate clarification. Instead, it waited over two and a half years, just prior to when it expected a legal challenge, before issuing any "clarification". Even then, as noted below, it did not provide a reasoned justification for this change in established position.

Defendants’ argument regarding migration of the paragraph from top to bottom (actually from bottom to top to bottom) is discussed below. Such migration does more to counter Defendants’ arguments than to support them, since the most reasonable inference is that the paragraph was moved to the bottom simply because the agency realized that the paragraph should be under the "reasonably anticipated" criteria to avoid a perception that it modified the "known" category.

2. The "Administrative History"

Defendants refer to various ambiguous references attributed to NTP staff in the unpublished minutes of pre-revision discussions as supporting the proposition that the agency consistently intended that human mechanistic data should be considered for listing in the "known" category. While any such pre-decisional statements by NTP staff are of extremely doubtful relevance in interpreting the final action of Defendant Secretary Shalala and Defendant NIEHS Director Olden, and Defendants are unable to point to even a single such statement indicating explicitly that the final paragraph was to apply to the "known" criteria, a fuller examination of the administrative history also shows that throughout the pre-decisional review process the final paragraph was intended to apply to only the "reasonably anticipated" criteria.

The "Meeting Summary" of the April 24-2, 1995 meeting of the Ad Hoc Working Group on the criteria (Def. Opp. Exh. 4) reflects that the full Group broke into three discussion groups, and that "it was the recommendation of breakout groups 1 & 3 that the existing two categories . . . should remain with revisions to category 2 ["reasonably anticipated"] to allow for all scientific evidence to be considered." (Id. at 3) (emphasis added). "These proposed revisions" (to Category 2) were reflected in Attachment 2 to the Meeting Summary (Id. at 4), and they are virtually identical to the final notice of the revisions. (Id. at 6). In addition, the "Summary Report" from breakout group 2 shows that that group recommended that the "known" category be limited to epidemiologic evidence sufficient to show a causal relationship, and that cases involving limited epidemiologic data belonged under Category 2. (Id., Group 2 report at 2, 3 and attached proposal). Thus, there was consensus in the Ad Hoc Working Group that the final paragraph would pertain only to Category 2.

The June 8, 1995 Federal Register notice of proposed revisions to the criteria (Def. Opp. Exh. 5) is perfectly consistent with the above record of consensus in placing the substance of the current final paragraph under the proposed revisions to the "reasonably anticipated" criteria.

The minutes of the June 29, 1995 Board of Scientific Counselors meeting (Def. Opp. Exh. 6) maintain this consistency. In discussing the proposed revisions, Dr. Jameson of NTP staff noted only minor proposed editorial revisions to Category 1. Reporting on the meeting of April 24-25, he stated, "The [Ad Hoc] Working Group stressed that listing under Category 1 should be based on human data only." (Id. at 4). With respect to proposed revisions to Category 2, Dr. Jameson stated that the new paragraph under Category 2, which is almost exactly the same as the disputed final paragraph in the final revised criteria, was simply a revised statement of the last sentence in the then-current statement of the Category 2 criteria. (Id.) The minutes also reflect that the rationale for recommending moving the substance of the disputed final paragraph to a preamble at this point in the review was because, as Dr. Henry commented, "it was not part of the criteria and perhaps part of the process". (Id. at 8). The statements in these minutes (Id. at 5) by Defendant Lucier and Dr. Barrett, which Defendants cite (Def. Opp. at 27), simply show that these individuals tired to raise the issue in this case for discussion, but that the issue was not resolved in favor of Defendants.

The minutes of the July 27, 1995 meeting of the NTP Executive Committee show that Dr. Lucier’s statement regarding "incorporation of mechanistic data as part of the criteria" pertained to the recommendations of the Ad Hoc Working Group, which, as discussed above, proposed use of mechanistic data only under Category 2. (Def. Opp. Exh. 7 at 3). Additionally, the minutes state clearly that "[t]he changes in 2.b. reflected increased sensitivity to mechanistic data considerations." (Id. at 4.) Moreover, these minutes reflect the addition of the third substantive paragraph -- then designated 2.c. -- to the current "reasonably anticipated" criteria. That third paragraph explicitly covers the situation in this case which Defendants argue is covered by the "known" criteria: less than sufficient evidence of carcinogenicity in humans and convincing mechanistic data. (Id. at 4 and attached proposed revisions.)

The final pre-decisional minutes from the May 8, 1996 meeting of the Board of Scientific Counselors Subcommittee, which Defendants reference (Def. Opp. Exh. 11), further confirm that the final paragraph pertains only to Category 2 ("reasonably anticipated"). Discussing the report of Dr. Lucier to the Subcommittee, the minutes clearly state that the criteria revisions that had been sent to the Secretary for approval were based on the original recommendations of the Ad Hoc Working Group concerning use of mechanistic data, which, as discussed above, were that mechanistic data should be considered under Category 2. (Id. at 1.) Additionally, the minutes show that only minimal editorial changes were intended for the "known" category, and that "Dr. Jameson reported that the third paragraph or subcategory [of Category 2] was added to cover agents supported with mechanistic data. . . ." (Id. at. 4.)⁷

Most telling, there is nothing in any of the pre-decisional minutes from 1995 and 1996 to indicate that any committee accepted that evidence from studies in experimental animals, whether in vivo or in vitro, or evidence from in vitro experiments with human cell cultures, (all of which Defendants are relying on in the case of the current dioxin proposal) could be used to support a listing in the "known" category.

3. The Wording of the Final "Descriptive Paragraph"

Two initial observations are necessary. First, this paragraph is described as a "descriptive paragraph", apparently in differentiation from the preceding substantive text. The paragraph does not purport to make any substantive revisions to the criteria.⁸ Second, the final paragraph refers simply to considering "all relevant data". (Emphasis added.) On its face this is apparently a very innocuous statement, since "relevant data" appears to refer to whatever data is relevant under the language of the criteria as set out prior to this paragraph. Defendants’ extraordinary reading of this language as allowing consideration of animal data and human in vitro data has only become apparent through the proposed dioxin listing (and possibly other proposed additions to the "known" category for the 9th RoC) together with the agency’s two April 1999 "clarifications".

Defendants’ attempts to argue that the wording of the final paragraph indicates that the paragraph substantively altered the stated "known" criteria are incredibly far-fetched. Defendants argue (Def. Opp. at 23) that the words "in humans" in the final paragraph refer to application of the "known" criteria, apparently because the "known" criteria also contain the words "in humans". But there is no reference to the "known" category, and the language Defendants quote leaves out the word "relevant", which, as discussed above, indicates that the paragraph was not intended to modify any of the preceding substantive statements of criteria.⁹

Finally, as discussed below and previously, at the time the 1996 criteria revisions were issued, the agency issued authoritative statements indicating that there had been no substantive changes in the "known" category, and that substantive revisions, including providing for use of mechanistic data, were made to the "reasonably anticipated category". These contemporaneous statements are clearly consistent with the plain wording of the revised criteria and the administrative history. As Defendants point out, such agency explanations of its own rules must be given substantial deference. This is particularly true when they are contemporaneous.

4. "Logic"

Defendants argue that "logic" dictates the use of mechanistic data for listing in the "known" category. (Def. Opp. at 23 & n. 11). To the contrary, logic dictates that only evidence from studies in humans indicating causality be the basis for stating that a substance is "known" to cause cancer in humans. Humans, and their biological responses to various exposures, are obviously extremely complex. Their functioning and responses cannot be replicated with assurance through animal experiments or exposure of cells in a petri dish.¹⁰ If "mechanistic data" meant a confident understanding of the mechanism by which exposure of a substance causes cancer in complete, living humans, such understanding could be obtained only through studies "in humans", and use of such data to determine if a substance is "known" to cause cancer in humans would indeed make sense. If we truly knew from studies in humans the mechanism by which a substance causes cancer in humans, it would certainly be appropriate to state that the substance is "known" to cause cancer in humans. But this is not the meaning of mechanistic data that Defendants want to use in the case of the proposed dioxin listing. They clearly want to use evidence from studies in experimental animals and human and animal in vitro data. This does not make sense for describing a substance as "known" to cause cancer in humans. In this light, at stake in this case is not only adherence to the criteria for a single government program, but also the integrity of term "known" in science. A reading of the DBD shows clearly that Defendants and the other RoC reviewers do not really have an understanding of whether and how dioxin causes cancer in humans. Instead they have what they regard as a plausible hypothesis about a possible mechanism that has substantial gaps, and this hypothesis is based on piecing together various disparate data from studies in animals, in humans, and in vitro. If they really understood a mechanism by which dioxin causes cancer in humans as a result of studies in humans, such understanding would truly be a breakthrough, and that understanding would be stated clearly and unequivocally in the DBD.¹¹

Defendants do not and cannot point to a single instance of a published listing in the existing eight RoCs (including the 8^th RoC, which was issued after the 1996 criteria revisions) in which the agency has based a "known" listing on less than sufficient epidemiologic evidence, using mechanistic or animal data to overcome such an insufficiency. Thus, the position currently being espoused by the agency, including in its brief, is a substantial change in position, and the agency has not publicly acknowledged this change and provided a reasoned justification.

Defendants note that virtually every listing in the RoCs has contained a discussion of animal data (Def. Opp. at 20 n.9), and they argue that such animal data has been provided only to corroborate epidemiologic or mechanistic data, and that this is allowable under the criteria. Defendants also argue that they are not relying on animal data in the case of dioxin. Plaintiffs have no quarrel with the agency’s discussing animal or mechanistic data for the purpose, as Defendants put it, of providing "additional support, above and beyond what is required, for a ‘known’ listing." (Id.) However, Defendants’ assertions avoid the plain fact that in all RoC listings to date for the "known" category, animal or mechanistic data has been presented as supplemental to a finding that the evidence from studies in humans is sufficient to show causality in humans. Defendants’ are correct in noting that animal or mechanistic data has consistently been used in the past only as additional or corroborative data above and beyond what is required for a "known" listing; but Defendants’ statements that they are not now relying on animal and mechanistic for direct support of a "known" listing are plainly inaccurate, and such reliance marks a drastic unexplained departure from previous practice. The DBD makes crystal clear that the agency is relying on evidence from studies in experimental animals as well as both animal and human in vitro mechanistic data. The administrative record and Defendants’ own brief also make clear that animal and in vitro mechanistic data are being used to compensate for a lack of sufficient evidence from studies in humans.

Complaint Exhibit 2 shows that the agency issued, on the same date as the Federal Register notice of the 1996 revisions, a stand-alone Departmental press release explaining the revised criteria and making clear that there were no substantive changes to the "known" category and that the only substantive changes, including allowance of use of mechanistic data, were to the "reasonably anticipated" category. Plaintiffs have documented that this press release was approved at the highest levels, by both Defendant Shalala and Defendant Olden. (Pl. Appl. Exh. 8 at 3). Defendants now attempt to dismiss this unambiguous, high-level Department statement as "staff authored", written "confusingly", "factually incorrect", and "misleading" (stated 4 times). (Def. Opp. at 24-25.)

In place of this contemporary explanation of the Department’s position, the Defendants offer, as an attachment to their opposition brief, a "Declaration" by Defendant Lucier, a subordinate of Defendants Shalala and Olden. That Declaration states, in effect, that the agency position set forth in the September 26, 1996 Departmental press release was incorrect and that the 1996 revised criteria incorporated substantive changes permitting a substance to be placed in the "known" category "based on limited human epidemiologic evidence and strong mechanistic data." (Def. Opp. Exh. 15.) Apart from the question of Dr. Lucier’s authority to issue such a statement contradicting an earlier, contemporaneous statement approved by his superiors, it would be difficult to find a clearer example of a post hoc rationalization. Such rationalizations are entitled to no weight or deference. See Reuters, Ltd. v. FCC, 781 F.2d 946, 951 (D.C. Cir. 1986); Platt College of Commerce, Inc. v. Cavazos, 796 F. Supp. 22, 26 (D.D.C. 1992).

Defendants also attempt (Def. Opp. at 24-25) to disavow the agency’s other contemporaneous statement published in Environmental Health Perspectives. (Compl. Exh. 3). Defendants characterize Environmental Health Perspectives as a "newsletter", although it is the official "Journal of the National Institute of Environmental Health Sciences", and its flagship scientific publication. Defendant Lucier was then, and still is, one of the two Co-Editors-in-Chief of the Journal. (Exh. 5). He certainly must have reviewed this important article describing the outcome of the two-year review process in which he participated and which he personally supervised. Moreover, both Defendants Shalala and Olden were advised that such an article similar to the Department’s official press release which they approved would be published in the Journal. Both approved such action. (Pl. Appl. Exh. 8 at 3.)¹² Insofar as Defendant Lucier’s Declaration contradicts the contemporaneous 1996 statements by the agency concerning the criteria revisions, it lacks credibility and legal significance.¹³

Even if it were true, as Defendants assert, that there were substantive changes to the "known" criteria in 1996 allowing Defendants to compensate for less than sufficient evidence from studies in humans with other types of evidence, such a change would itself have been a substantial departure from established agency practice or policy for which the agency has not published a reasoned justification. Defendants have characterized the agency’s own contemporaneous 1996 explanations of the criteria revisions as misleading, confusing, and factually inaccurate, and the agency’s April 1999 Federal Register "clarifications" are simply conclusory statements rather than a reasoned explanation for such an alleged substantial change in the criteria. Likewise, the statement in Defendants’ opposition brief that the alleged 1996 revision to the "known" listing criteria "has been fully explained" (Def. Opp. at 29) is similarly conclusory and unsupported.¹⁴

As is evident from the DBD summary, the agency places heavy reliance on the IARC monograph and the allegedly post-IARC Bertazzi et al. findings. It is not difficult to see why this is so. The proposed dioxin listing posed a novel problem for the agency. There had never been a listing of a substance for which IARC had found that the evidence from studies in humans was "limited", but where the agency (HHS/NIEHS/NTP) nevertheless listed the substance in the "known" category. To attempt to get around this problem, the agency crafted a listing rationale indicating that (1) IARC had found that the epidemiologic studies established a causal relationship, and (2) that substantial evidence supporting a "known" listing lay in an allegedly new Bertazzi et al. study that IARC had not considered. As Plaintiffs have demonstrated, both assertions are false. The agency’s substantial reliance on those two false assertions makes the proposed listing arbitrary and capricious. Defendants’ responses are easily disposed of.

First, Defendants argue that their statements regarding IARC’s finding a causal relationship are "not inaccurate" because IARC stated that two worker populations that had been studied provided evidence "supporting a causal relationship". This quotation leaves out key wording and is misleading. The complete IARC statement is that the two worker populations studied "provide dose-response evidence supporting a causal relationship. . . ." (Def. Opp. Exh. 3 at 193) (emphasis added). As Defendants know well know, evidence of dose-response is only one of many factors examined in reaching a judgment on causality, and ultimately IARC judged the totality of evidence from studies in humans as "limited", meaning that a causal relationship was not established.

Next, Defendants argue that because at one place in the DBD the agency stated that IARC had found that the evidence from studies was "limited", the agency did not incorrectly rely on IARC’s finding a causal relationship. While the agency admittedly did note this IARC conclusion, the DBD summary’s statement that IARC found a causal relationship is far more prominent. In addition, the DBD statement that IARC found the evidence from human studies was "limited", at the beginning of section 3.0 (Def. Opp. Exh. 3 at 3-1), is quickly followed on the next page by the more conclusory statement that IARC "identified a causal association between TCDD exposure and all cancers combined . . . ." (Id. at 3-2.) Taking these three statements together (two agency statements that IARC found a causal relationship and one introductory statement that IARC had found the human evidence to be "limited"), it is clear that the agency’s overall reasoning -- which is fatally flawed -- is that even though IARC found that the evidence from human studies is "limited", it nevertheless found a causal relationship, thus supporting the agency’s proposed listing.

With regard to the Bertazzi et al. findings, Defendants argue that (1) evaluating such studies involves scientific judgment, and the court should not second-guess the agency; and (2) the 1997 and 1998 articles by Bertazzi et al. were not in fact considered by IARC. (Def. Opp. at 29-31.) As with the IARC findings, Defendants attempt to finesse this issue. Expert scientific judgment is not required to compare the 1996 Bertazzi et al. findings, which IARC obviously considered (Pl. Appl. Exhs. 11 and 9) , with the findings of the Bertazzi et al. articles published in 1997 (Pl. Appl. Exh. 12) and 1998 (Exh. 6) and to see that the basic findings are no different.¹⁵ Defendants’ statement that IARC did not consider the 1997 and 1998 publications is therefore technically accurate but grossly misleading, because IARC considered the findings in the 1997 and 1998 articles as they were presented in the 1996 publication. Nor is expert scientific judgment needed to simply read the conclusions stated in the Bertazzi et al. studies and see that they do not support a finding of sufficient evidence of a causal relationship, and if anything appear to detract from such an attempted finding.

In reviewing a threatened agency action (based on another already final action) under the APA, the court’s review must be "searching and careful". (Pl. Appl. at 12.) Such a review here will show that the agency’s substantial reliance on the Bertazzi et al. findings is arbitrary and capricious. Defendants have failed to explain how the 1997 and 1998 Bertazzi et al. publications contain any significant findings that are not in the 1996 publication reviewed by IARC, despite being challenged on this point many times both in the administrative proceedings and this litigation. Defendants’ only response is repeatedly to invoke "deference". Deference has its limits, however, and Defendants have clearly exceeded those limits.

Finally, Defendants argue that the agency is relying on many epidemiologic studies in addition to the Bertazzi et al. studies, referring to the DBD Supplement and its additional citations. The DBD Supplement (Compl. Exh. 21) does contain additional citations, but the agency has not indicated in the DBD or the Supplement that it is placing any reliance on these new authorities. The agency instead is relying substantially on the Bertazzi et al. studies, and that reliance is misplaced.

Whether a causal relationship can be found because confounding could be adequately excluded as a possible alternative explanation for an observed association is a critical facet of the agency’s listing criteria. The importance of this point is shown by the explicit wording of the "reasonably anticipated" criteria, as well as by the agency’s established practice of conforming its evaluations of the evidence from studies in humans to those of IARC, which also makes consideration of possible confounding a critical factor. Defendants assert that they "have previously responded to this issue" through the August 24, 1998 letter from Defendant Olden to Plaintiff Tozzi. (Def. Opp. at 31-32.) The statement quoted from that letter, however, is not a response. It amounts to no more than a statement that the agency decided not to consider this issue and focused on studies that measured exposure to dioxin only and not to the other workplace chemicals that were recognized as potential carcinogens. As Plaintiffs have pointed out previously, IARC found that it could not exclude confounding of dioxin exposure by exposure to other substances as an alternative explanation, and that this was a principal reason for its conclusion that the evidence from studies in humans was "limited".¹⁶ Defendants have been on notice for a long time that the issue of confounding was an important aspect of the proposed listing action, but they have failed to address this issue. Such a failure is arbitrary and capricious. Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983).

As Plaintiffs have previously demonstrated, the threat of irreparable harm from publication of a prominent federal report tying Plaintiffs’ products to a known human carcinogen is both real and severe. Defendants’ arguments to the contrary should be disregarded.

First, Defendants’ attempts to characterize Plaintiffs’ claims as too speculative are without merit, and their citation to Wisconsin Gas Co. v. FERC, 758 F.2d 669 (D.C. Cir. 1985), in this regard is particularly unavailing. In Wisconsin Gas, the D.C. Circuit acknowledged that an injunction is appropriate, inter alia, where "[t]he movant provide[s] proof that the harm has occurred in the past and is likely to occur again". 758 F.2d at 674. Here, as Plaintiffs have demonstrated by reference to the public reaction following news of dioxin contamination in Europe, the prospect of a dioxin "food scare" affecting the behavior of restaurant customers, restaurant suppliers and government officials is far from remote. No such model of past behavior was present in Wisconsin Gas, and the court there was faced with a "purely hypothetical chain of events." Id. at 675.

Defendants argue that it is very unlikely that the U.S. public will react adversely to Plaintiffs, since IARC has previously categorized dioxin as "carcinogenic"¹⁷ and no food scare resulted. IARC is not a U.S. agency, however. It is a specialized United Nations subsidiary organization located in Lyon, France. There is nothing in the administrative record or the exhibits in this case demonstrating that the U.S. public pays attention to IARC monographs, or, for that matter, even knows that IARC exists. However, an authoritative pronouncement to the U.S. public specifically on U.S. dietary exposures by the chief U.S. health agency is sure to attract attention. As Plaintiffs have shown in their exhibits, various activist groups are on the lookout for just such a pronouncement and are sure to trumpet it. Moreover, the IARC pronouncement can be viewed as having contributed to the recent dioxin scare in Europe, and therefore a stronger pronouncement in the U.S. implicating much of the U.S. food supply is likely to have similar or even more far-reaching impacts.

Defendants also argue that the San Francisco and Oakland resolutions have not resulted in a food scare. Although there is nothing in record to either support or refute this assertion, it is clear that those resolutions did not focus on widespread presence of dioxin in meats, fish and dairy products, but rather on the hazards of dioxin only in San Francisco Bay fish, and on industrial emission sources and the fish feedstocks that resulted in such food contamination. Those resolutions demonstrate the high visibility of dioxin as a public health issue and the likelihood that a new RoC listing broadly implicating dioxin and the U.S. food supply will be publicized by activist groups.

Second, Defendants’ reliance on Bristol-Myers Squibb Co. v. Shalala, 723. F. Supp. 212 (D.D.C. 1996), is misplaced. In that case, the court held that plaintiffs had failed to show economic harm necessary to support an injunction where, by plaintiff’s own estimates, the government’s action threatened to decrease plaintiff’s sales by less than 0.5 percent, a figure the court characterized as "minuscule". 923 F. Supp. at 221 & n.12. Here, by contrast, Plaintiffs have submitted detailed affidavits discussing the broad impact that publication of the proposed RoC will have on Plaintiffs’ businesses. See Affidavits of Messrs. Leonard and Wexler, attached to Plaintiffs’ Opposition to Defendants’ Motion to Dismiss, and incorporated by reference in Plaintiffs’ application for injunction. (Pl. Appl. at 31). As explained in those affidavits, the lion’s share of the Plaintiffs’ menus are made up of the very foods identified in the DBD (meat, poultry, fish and dairy), and the adverse effect on sales of these dishes would be significantly more damaging to Plaintiffs than the threatened harm at issue in Bristol-Meyers.

Third, Defendants mistakenly assert that Plaintiffs must demonstrate that "severe economic impact" is "certain" to occur. (Def. Opp. at 33, 35). To the contrary, the economic harm that Plaintiffs must demonstrate is lessened where the quantum of damages would be difficult to calculate and/or where there is no available judicial or other remedy.

The oft-cited language that defendants quote from Bristol-Meyers and Wisconsin Gas, which stresses that "[t]he key word in this consideration is irreparable" (Def. Opp. at 35), actually originated with the D.C. Circuit’s decision in Virginia Petroleum Jobbers Ass’n v. FPC, 259 F.2d 921 (D.C. Cir. 1958). Defendants fail to acknowledge the immediately following sentence in Virginia Petroleum Jobbers, however, which explains the basis for the irreparable injury principle:

The possibility that adequate compensatory or other corrective relief will be available at a later date, in the ordinary course of litigation, weighs heavily against a claim of irreparable harm.

Virginia Petroleum Jobbers, 259 F.2d at 925.

A plaintiff’s inability to obtain relief is thus a key factor in determining the threat of irreparable harm. Where would Defendants have Plaintiffs turn for "adequate compensatory or other corrective relief" here once the misleading and economically damaging RoC is published? Would the government be willing to waive sovereign immunity in a subsequent suit by Plaintiffs for money damages? How would such damages be calculated? Defendants give short shrift to these points (Def. Opp. at 35), misciting instead the Bristol-Meyers court’s general citation to a New York district court case. See Bristol Meyers, 923 F. Supp. at 221 (citing General Textile Printing v. Expromtorg Int’l 862 F. Supp. 1070, 1075 (S.D.N.Y. 1994)).¹⁸

Defendants inexplicably attempt to distinguish cases enjoining the release of statements injurious to product reputation by noting that these cases "relate to information that would be specifically damaging to the plaintiff’s business." (Def. Opp. at 36). But this is precisely the point. Release of the proposed dioxin listing, will impugn "specifically" the products Plaintiffs sell, just as the publications at issue in the cited cases threatened to impugn the products sold by those plaintiffs. Defendants’ ultra-specific standard apparently would require that the proposed RoC identify, for example, "Greenbaum & Gilhooley’s Breast of Chicken Teriyaki" as containing a known human carcinogen. This is obviously not required, and Defendants cannot reasonably argue that disparagement of a basic component of a party’s product will not harm the party itself.¹⁹

Finally, Defendants’ observation that Malarkey-Taylor Assoc., Inc. v. Cellular Communications Indus. Ass’n, 929 F. Supp. 473, 478 (D.D.C. 1996), involved a trademark infringement claim (Def. Opp. at 37) is, of course, correct. The broader point, however, is not whether the case involved confusion between two products, but whether an injunction may be appropriate to prevent publication of information damaging to a particular business. Malarkey-Taylor and the other cases Plaintiffs have cited answer this question in the affirmative.

Defendants argue that they would be harmed by a delay in publication because of the uncertainties involved in deciding whether to print the 9^th RoC with the new dioxin listing included. This point really weighs in favor of granting the preliminary injunction because an enforced delay would relieve Defendants from such uncertainties. If, as Plaintiffs have shown, they are likely to succeed on the merits, proceeding with publication now with an upgraded dioxin listing would cause great harm to Defendants in expense and other resources if they later must reprint the 9^th RoC with a revised dioxin listing. Defendants would also have to take measures to inform the public and those who have already received the Report (whether in hard copy or over the Internet) that the first version they received was not accurate. Additionally, any such retractions and corrections could severely impair the credibility of this government program. Further, weighing this absence of hardship to Defendants against the likelihood of harm to Plaintiffs, retraction after publication and further litigation is extremely unlikely to be an effective remedy.

Ensuring that the information provided in the RoCs is accurate and in conformance with the agency’s listing criteria and the scientific evidence clearly weighs heavily in favor of granting the preliminary injunction. Defendants have not even attempted to respond to these points. Their statement that the court should allow the 9^th RoC to be published "in a timely fashion" is particularly unresponsive, given that the Report is not due at the earliest until May 2000.²⁰

Defendants’ incantations of "mechanistic data" and "deference to scientific expertise" should not deter the Court from a searching and careful review. Any intelligent lay person without such expertise can see plainly that: (1) the agency’s position on the meaning of the "known" criteria, and the application of that position to dioxin, is inconsistent with the plain meaning of the listing criteria; (2) the agency’s position is a radical departure from previously established practice without any reasoned explanation; (3) the agency has relied substantially on evidence that does not support its position; and (4) the agency has completely ignored repeated calls to address the important issue of confounding exposures.

None of Defendants’ arguments against the application for preliminary injunction has merit. Defendants’ responses demonstrate only their determination to adhere to an arbitrary and capricious position on the meaning of the "known" listing criteria and to apply that position to the dioxin listing. Because Plaintiffs and the public have no way of knowing when the ultimate approval of the listing action will occur, publication should be enjoined on the basis of the administrative record as it now stands. Plaintiffs have satisfied all of the elements necessary to justify a preliminary injunction. Enjoining publication of the Report pending final judicial resolution of the merits is in the interests not only of the Plaintiffs but also of the public and Defendants themselves.

For all of the foregoing reasons, Plaintiffs’ application for injunction should be granted.

Respectfully submitted,

_______________________
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

October 8, 1999

¹ Defendants cite Kahane v. Secretary of State, 700 F. Supp. 1162, 1165 (S.D.N.Y. 1988), and mistakenly imply that the Court should apply a "clear and convincing" standard to Plaintiffs' injunction application. (Def. Opp. at 12). Kahane has never been cited for this proposition, however, and the Kahane court's citation to Yakus v. United States, 321 U.S. 414 (1944), is somewhat perplexing. The district court provided no "jump" cite to the 34-page majority opinion in Yakus, the words "clear and convincing" appear nowhere in that 1944 the Supreme Court decision, and Yakus did not deal directly with the standard for preliminary injunction. In any event, the law in the D.C. Circuit holds that motions for injunction are to be decided based on a "sliding scale" evaluation of the traditional four factors. See, e.g., CityFed Fin. Corp. v. Office of Thrift Supervision, 58 F.3d 738, 746-47 (D.C. Cir. 1995).

¹ Defendants have not suggested that there has been any change in the listing rationale stated in the DBD. While it is theoretically possible that the rationale could still be revised -- as, indeed, Plaintiffs contend it would have to be to support such a listing change -- any significant revision at this point in the review process would raise a new issue of whether the Defendants have violated their formal listing review procedures by not giving the four review committees an opportunity to review that new rationale.

¹ The first of the three types of evidence, which is plainly evidence from epidemiologic studies, is described as "[h]uman studies". This distinction belies Defendants' argument that the agency has interpreted "studies in humans" to be broader than epidemiologic studies, and to encompass mechanistic evidence, which is described in the DBD summary as the third type of evidence on which the agency relies. Moreover, the more detailed discussion of "Human Studies" in section 3.0 of the DBD (Def. Opp. Exh. 1) includes only epidemiologic studies, and the mechanistic data are discussed in section 7.0 under "Mechanisms of Carcinogenesis". (Id.)

¹ In vitro studies, whether of animal or human tissue or cells, involve exposure of only tissue or cells to the agent being studied. There is no exposure of humans. If the tissue or cells came from exposed humans, the study would be simply a form of molecular epidemiology, which would be permissible under the "known" criteria. But such studies are in vivo, not in vitro.

¹ Almost all of the IARC dioxin monograph is attached to the agency's DBD as an appendix, and the DBD cites to it for support at numerous points.

¹ The agency has not even been consistent in its non-Federal Register formatting, as shown by the way it is currently presenting the criteria on its website (as of Oct. 6, 1999) -- formatting which makes it very clear that the final paragraph pertains only to the second category. (Exh. 3).

¹ The statement in these minutes immediately following this that the final paragraph "applies to all the criteria", in context appears to mean no more than that the final paragraph was intended to apply to all the criteria under Category 2 insofar as consideration was to be given to evidence other than sufficient evidence from studies in humans.

¹ The statement by Dr. Henry quoted above, from the administrative history, also shows that this point was recognized when the revisions were being drafted.

¹ The phrase "regarding carcinogenicity" in the final paragraph appears to refer loosely to data pertinent to either the category insofar as such data is relevant to the criteria for that category.

¹ This is especially true in the case of a substance such as dioxin which appears to act through a receptor-binding-initiated mechanism. The attached Letter to the Editor in a 1998 issue of Risk Analysis on "The Dose-Response Model for Dioxin", the authors of which included a large number of prominent academics, notes that the dioxin mechanism is "poorly understood" and recommends more reliance on epidemiologic evidence "given the known extensive interspecies variation in the biological outcomes of receptor-binding substances." (Exh. 4). The discussion of mechanistic data in both the DBD and the IARC monograph are consistent with these observations.

¹ The IARC monograph, on which the agency relies for much of the mechanistic evaluation, does not claim that there is an understanding of a mechanism by which dioxin causes cancer in humans. It claims only that "several lines of evidence", including evidence from experimental animals, indicate that there is some mechanism "involving the Ah receptor", and that "this receptor . . . functions the same way in humans as in experimental animals." (Exh. 1 at 343) (emphasis added).

¹ Defendant Lucier has also attempted to use Environmental Health Perspectives, and his position as Co-Editor-in-Chief, as a platform for stating a view similar to that in his new Declaration. (Pl. Appl. Exh. 2).

¹ Plaintiffs do agree, however, with one statement in Defendant Lucier's new Declaration -- the statement in paragraph 5 that the agency's 1996 statements related to the "precise wording" of the "known" category. It is this plain wording of the "known" criteria that excludes reliance on animal or in vitro data to compensate for less than sufficient evidence from studies in humans. It is this wording that Defendants now seek to avoid.

¹ If the 1996 notice of the revisions had notified the public of the important substantive changes that Defendants now claim occurred, then the agency would not have needed to issue the two "clarifications" in April 1999 (while anticipating this litigation). Neither of those clarifications, however, contains any reasoned justification for the alleged substantive change in the criteria.

¹ The 1998 article by Bertazzi et al. contains the same objective findings as the 1996 and 1997 articles but simply reformatted in Tables 4.3 and 6 at pp. 629-30. It was simply published in Environmental Health Perspectives after publication in Epidemiology. The 1998 article, on which the agency has not previously relied, also further detracts from the rationale in the DBD. The authors concluded in the 1998 article that, considering the epidemiologic, experimental and mechanistic data together, such data supported only "hypotheses" for carcinogenicity. (Exh. 6 at 630). Furthermore, the authors expressly noted that discrepancies among the various studies could be explained by "concomitant exposures to other chemicals". (Id. at 631).

¹ See also n. 15, supra, regarding Bertazzi et al.'s views on this issue and the IARC monograph. Significantly, Dr. Bertazzi was a member of the IARC working group that evaluated dioxin and wrote the monograph.

¹ IARC does not use the term "known". Thus an HHS pronouncement that dioxin is "known" to cause cancer in humans is likely to be viewed as a considerably stronger statement.

¹ The Bristol-Meyers court did not quote WMATC v. Holiday Tours, Inc., 559 F.2d 841, 843 n. 3 (D.C. Cir. 1977) as a basis for rejecting plaintiff's argument regarding inability to recover damages, as Defendants imply (Def. Opp. at 35). See Bristol Meyers, 923 F. Supp. at 221. The court's subsequent citation to the D.C. Circuit decision in Holiday Tours was unrelated to the situation where a plaintiff would later lack the opportunity for an adequate judicial remedy. Id.

¹ With regard to the Court's jurisdiction and the broad impact of the RoC, it is unfortunate that Defendants characterize the misuse of an important federal report, the disregard of congressional intent, and the disruption of major industrial sectors as "silly". (Def. Opp. at 37 n.18). That the harm Defendants are about to inflict may affect a substantial number of individuals, however, is not a sufficient basis to deny standing.

¹ One of the primary purposes of the RoCs is to inform individual consumers so that they can alter their behavior to avoid exposure to carcinogens, or to seek treatment for such exposure. If Defendants' assertion that the public already knows that dioxin is a carcinogen were correct -- a point which Plaintiffs dispute, particularly with regard to widespread exposure through meats, fish, and dairy products -- then there is certainly no urgency in issuing the proposed new dioxin listing.