Document k6jbEpx0gBV7j21xdrML2LqjB

6 m m 0 0 fl (fta} FRIDAY, SEPTEMBER 23v 1977 PART V DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE Food and Drug Administration INDIRECT ADDITIVES: POLYMERS Acrylonitrile Copolymers Used to Fabricate Beverage Containers; Final Decision i i t n t m t t t i 'din , r 48528 RULES AND REGULATIONS THIe 21--Food and Druse ly 21 CFR 121.4010), requiring additional CHAPTER I--FOOD AND DRUG ADMINIS TRATION, DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE [Docket No. 'J6N-0070] testing to be undertaken to determine whether acrylonitrile Is safe for use in beverage containers and for other food contact uses. The Natural Resources De fense Council (NRDC) objected to the PART 177--INDIRECT ADDITIVES: interim regulation, and requested a stay POLYMERS Acrylonitrile Copolymers Used To Fabricate Beverage Containers; Final Decision AGENCY: Food and Drug Administra tion (FDA). of the regulations and hearing on wheth er acrylonitrile could continue to be safe ly used. When the tests undertaken pur suant to the Interim regulation showed positive results, the Commissioner issued an order, published in the Federal Reg ACTION: Final Decision Following A Formal Evidentiary Public Hearing, Findings Of Fact, Conclusions Of Law and Final Order. SUMMARY: The Commissioner is pub lishing his final decision following a for mal evidentiary public hearing in a rulemaking proceeding concerning acryloni trile copolymers used to fabricate bev erage containers. He has determined that acrylonitrile copolymers used to fabricate beverage containers are food additives and that they have not been shown to be safe. He affirms the Initial Decision of the Administrative Law Judge with the modifications and sup plementation stated below. The food addltlve regulations at Issue (21 cfr 177.1020.177.1030.177.1040, 177.1050, and 177.1480) are amended to eliminate use of acrylonitrile copolymers, to fabricate beverage containers. ister of March 11, 1977 (42 FR 13546), which stayed the regulations to the ex tent they permitted acrylonitrile copoly mers to be used In beverage containers pending completion and evaluation of the studies required by the interim reg ulation. The Commissioner Issued the stay on the basis of the objections filed earlier by NRDC. The Monsanto Co., a manufacturer of acrylonitrile bottles, sought Judicial review of the stay before the United States Court of Appeals for the District of Columbia. Monsanto Co. v. Gardner, C-A. No. 77-1245 (D.C. Cir., 1977). The Court found that the Commission er could not stay the regulations on the basis of the objections filed by NRDC because of the passage of time. The Court held that the Commissioner, at least in substantial part, had acted on his own initiative pursuant to sections 409 (d) and (h) of the Act, 21 U.S.C. 348 (d) and EFFECTIVE DATE: The effective date (h). The Court ordered a prompt hearing of the final decision Is September 19, on the Monsanto objections, and set a 1077. The final order amending the reg date, later extended to September 19. ulations is effective December 22,1977. 1977, for issuance of a final decision. ADDRESS: The transcript of the hear ing, evidence submitted and all other documents cited in this decision may be. seen in the Office of the Hearing Clerk (HFC-20), Rm. 4-65, 5600 Fishers Lane, Rockville, Md. '20857, between the hours of 9 a.m. and 4 pm., Monday through Friday. This proceeding has been held, and this decision is being Issued, pursuant to the Court's order. The Commissioner issued a notice of hearing in this matter and waived cer tain procedures in the Federal Register of April 1,1977 (42 FR 17529). The par ticipants in the hearing were the Bureau of Foods of the Food and Drug Adminis FOR FURTHER INFORMATION CON tration (Bureau), Natural Resources De TACT: fense Council (NRDC) and, for the man Ted Herman, Compliance Regulations Policy Staff (HFC-10), Food and Drug Administration, Department of Health, Education, and Welfare, 5600 Fishers Lane. Rockville, Md. 20857, (303-443-3480). SUPPLEMENTARY INFORMATION: The purpose of this rulemaking proceed ing la to decide whether acrylonitrile copolymers used to fabricate beverage ufacturers of acrylonitrile bottles, the Monsanto Co. (Monsanto), Borg-Wamer Corp. (Borg-Warner), and Vlstron Corp. (Vlstron). Following a formal eviden tiary public hearing, the Administrative Law Judge issued an Initial Decision on August 4. 1977, finding acrylonitrile copolymers used to fabricate beverage containers to be food additives which had not been shown to be safe.1 The containers are food additives within the meaning of section 201 (s) of the Federal Pood. Drug, and Cosmetic Act (the Act). 21 U.S.C. 301 et scq., 321fs>, and whether >Tbe following Abbreviations have been used in citing material in the record: Initial Decision: ID: Transcript: Tr.: Brief of a Par ticipant to the Administrative Law Judge: they have been shown to be safe under Brief: Appeal from and/or Exceptions to Ini section 409 of the Act, 21 U.S.C. 348. tial Decision: App.; Reply of a Participant to Appeal from and/or Exceptions to initial 1. History Decision: Reply. The exhibits submitted to Acrylonitrile copolymers have been permitted to be used In beverage contain ers under the following food additive the record, including written direct testi mony. arc referred to by the symbol for the participant and number assigned to them upon filing by the Hearing Clerk. The follow regulations: 21 CFR 177.1020, 177.1030. ing symbols were used by the Hearing Cleric 177.1040, 177.1050, and 1t7o7.r1e4f8e0r .toInthteheexhibits by the participants: Federal register of June 14,1976 (41FR Monsanto: M; Borg-Wamer: BW; Vlstron: 23940), the Commissioner issued an in V; Bureau: G; and NRDC: NRDC. Thus, the first exhibit submitted by Monsanto would terim regulation (21 CFR 180.22, former be referred to as M-l. manufacturers excepted to the Initial Decision and appealed It to the Commis sioner. In their replies, the Bureau and NRDC requested the Commissioner to affirm the Initial Decision with clarifica tions or additional findings. In reviewing the Initial Decision, the Commissioner has all the powers he would have in making the initial decision (21 Ctn 12.130). The Initial Decision and the Final Decision must be based upon "a fair evaluation of the entire rec ord,'" pursuant to section 409(f) of the Act, 21 U.S.C. 348(f), and must also meet the requirements of 21 CFR 12.130 and 12.130. After reviewing the record care fully, the Commissioner affirms the In itial Decision with the modifications and supplementation stated below. II. Food Additive Status A food additive Is defined by section 201(s) of the Act, 21U.S.C. 321(s) to be: soy substance the Intended use ol which results or may reasonably be expected to result, directly or Indirectly, in its becom ing a component of any food (Includ ing any substance intended for use in packing or bolding food), if such sub stance Is not generally recognized, among experts qualified by scientific training and experience to evAlu&te its safety, as having been adequately shown through aclentlflc procedures * to be safe under the con ditions of Its Intended use * *. Acrylonitrile bottles are composed of single molecules (monomer) of acryloni trile (AN) and other substances formed into a polymeric chain (copolymer) that constitutes the bottle wall. At issue here is whether acrylonitrile copolymers used to fabricate beverage bottles are food ad ditives within the meaning of section 201 (s) of the Act, 21 U.S.C. 321(3), be cause of the migration of residual acry lonitrile monomer (RAN) from the bot tle wall into the contents of the bottles at levels that are not generally recog nized as safe. The Bureau of Foods has the burden of proof in this proceeding on the food additive status of acrylonitrile copolymer beverage bottles. To meet this burden, the Bureau must show that (1) the in tended use of acrylonitrile beverage bot tles results, or may reasonably be ex pected to result. In acrylonitrile mono mer becoming a component of food, and (2) acrylonitrile Is not generally recog nized as safe under its intended condi tions of use. A. INITIAL DECISION The Administrative Law Judge found that acrylonitrile had been shown to become a component of food and not to be generally recognized as safe. (ID at 35-36.) He based his conclusion that acrylonitrile becomes a food component on the direct observation of acrylonitrile migration in three studies using labo ratory detection methods. Hie studies were conducted with beverage bottles fFabricated from acrylonitrile copolymer and stored for 6-7 months at room tem perature. (ID at 12-14.) These storage conditions were regarded by the Admin istrative Law Judge as "normal usage conditions" for beverage bottles and FEDERAL REGISTER, VOL. 42, NO. IBS--FRIDAY, SEPTEMBER 23, 1977 RULES AND REGULATIONS 48529 "well within the parameters of intended use." (ED at 11 and 35.) The Adminis trative Law Judge also concluded that the direct observation data is "strength ened by other positive results which employ different conditions and models which predict migration under various circumstances," (ID at 35) and that it "justifies the use of other avenues for quantification of these observed levels to be considered, such as models predict ing the quantity of migration, and analy ses of tests under more severe condi tions." (ID at 11.) Because direct obser vation of acrylonitrile migration was made under storage conditions "within the parameters of the Intended condi tions of use,,r the Administrative Law Judge found it unnecessary to resolve whether the manufacturers were correct that "testing for migration should be conducted under conditions simulating the Intended conditions of use" or whether the Bureau was correct that the "details of time and temperature conditions" are "relevant only to safety questions." (ID at 9.) cant," but Monsanto argues also that the Administrative Law Judge's finding that scientific opinion is balanced establishes that the Bureau did not meet its burden of proof of showing that acrylonitrile is not GRAS. (Monsanto App. at 23-27.) (4) All the manufacturers argue that if any acrylonitrile migrates into ttieir current bottles, the amount Is so small in amount as to be "toxlcologlcally Insig nificant'' and not "worthy of government concern." They believe the Bureau has the burden to show that any amount that may migrate Is toxlcologlcally signifi cant, and they object to the Initial Deci sion for foiling to Impose this burden on the Bureau. (Monsanto App. at 3-5, 2325; Borg-Wamer App. 2-6; VisIron App. at 3.) (5) Lastly, the manufacturers believe that a finding that acrylonitrile is a food additive would be inconsistent with FDA's actions on other substances. (Monsanto App. at 9-10, 27; Vlstron App. at 3-11.) C. COMMISSIONER'S RESPONSE TO Sable or exceed some miniminn thres hold level before the substance be considered a component of food. The statute requires only that, at a minimum, there be evidence establishing a reason able expectation that some migration will occur. Congress adopted this approach In order to effectuate the "overriding congressional Intent" underlying the Pood Additives Amendment of 1958, which was "to exclude from food, the addition of harmful or unsafe .sub stances." (ID at 6.) The Administrative Law Judge was correct In stating that: it is entirely contrary to any "rule of rea son" approach to statutory construction to conclude that Congress intended to disregard the overriding Intent merely because a sub stance migrates to food only In minute quantities regardless ot the poeaibla harmful effects of consuming such minute quantities. (ID at 8.) The Commissioner agrees also with the Administrative Law Judge's statement of the reason for not requiring that migra tion be either detectable or quantifiable: EXCEPTIONS The manufacturers assert that the statu B. EXCEPTIONS tory requirement of a showing of migration The manufacturing parties have ex cepted on several grounds to the conclu sion that acrylonitrile is a food additive. They stress that the Administrative Law Judge's findings were based upon the direct observation of migration from Monsanto's "old" bottles but that no similar direct observation of migration was made from either the "new" Mon santo bottle or the bottles of the other manufacturers. In order to provide an overview of the issues raised, the Com missioner has listed below the principal contentions of the manufacturing par ties. Tlie Commissioner will discuss each contention in more detail and respond individually to them, as well as to related minor exceptions, In the following sec tion of this Decision. A summary of the principal exceptions follows. (1) Monsanto contends that "if no migration of AN can be detected from the new bottle by the most sensitive vali dated analytical techniques presently available . . . after suitably exaggerated extraction, then it is not a food additive." (Monsanto App. at 18-19.) Borg-Warn er's position is similar. < Borg-Warner App. at 9, footnote.) t2) Monsanto maintains that the ob servation of migration from the "old" bottles is "irrelevant" with respect to the "new" bottles, and that It would be a "specious line of reasoning" to conclude that "if [acrylonitrile! migrates from the old' bottle. It can be said to migrate from all bottles." (Monsanto App. at 16-17.) Borg-Warner maintains that the Administrative Law Judge's findings are irrelevant to its bottles in the absence of a finding of actual migration from the actual Borg-Wamer product. (Borg- Warner App. at 10-11.) (3)The manufacturers maintain that the Bureau failed to show that acryloni trile is not generally recognized as safe (GRAB) under its Intended conditions of use. This exception Is related primarily to their contention, stated below, that acrylonitrile is "toxlcologlcally Insignfl- 1. Food Additive Status of Undetect able Substances. Monsanto objects to the Initial Decision because It finds acrylo nitrile copolymer beverage containers to be food additives even though there is no detectable migration of acrylonitrile from Monsanto's currently producible bottle. (Monsanto App. at 3.) Monsanto maintains that the Bureau must show that acrylonitrile Is a food additive under "the actual conditions of use" and that this requires a showing of detectable migration from the new bottle (Monsanto App. at 2, 18-23). If no migration of acrylonitrile can be detected from ttie new bottle by the most sensitive method available, Monsanto would argue, It can not be considered a food additive. Hie most sensitive method currently avail able is one developed by Monsanto, able to detect substances at concentrations of 10 parts per billion (ppb> or above. (Monsanto App. at 18.) According to Monsanto, the matter of "whether or how much AN might be present below the limits of actual detection is conjecture, not evidence, and cannot be of decisional significance." (Monsanto App. at 19.) Borg-Warner's position is similar. (BorgWamer App. at 9, footnote.) The Commissioner rejects the conten tion that a substance Is Immune from, regulation as a food additive simply be cause no migration can be detected by the most sensitive analytical method available. This contention is inconsist ent with the statutory language and the Congressional purpose in enacting the food additive provisions of the Act. Under the food additive definition in sec tion 201(S) Of the Act, 21 UJS.C. 321 (s>, a substance can be considered a food component if its "Intended use" either "results" or "may reasonably be ex pected" to result In Its becoming a com ponent of food. Contrary to the asser tions of the manufacturing parties, the statutory language does not require that the amount of the migrating substance be analytically detectable or quantt- necessitates a quantitative determination. Such an argument, taken to Ita logical ex treme, would mean that a chemical which was undetectable by current methods, but whose unseen migration caused toxic effects In man, would be exempted from the re quirements of the statute. The "reasonably expected" wording of tbe statute is clearly not meant to preclude regulation in such a situation. The overriding intent of the legis lature to protect tbe health of the American public cannot reasonably be disregarded simply because the methods of detection are unable to quantify the precise level of mi gration. (ID at 13-14.) Thus, as the Administrative Law Judge recognized, it would defeat the purpose of the Food Additives Amendment to re quire actual detection or quantification of a migrating substance before it could be considered a component of food. It is enough to show a "reasonable expecta tion" of migration. 2. Reasonable Expectation, of Migra tion; Relevance of Extraction Data. Hie manufacturing parties contend that the direct observation of acrylonitrile migra tion In the "old" Monsanto bottles is "irrelevent" to the question of whether ac rylonitrile migrates or may reasonably be expected to migrate from the "new" Monsanto bottlo or the bottles made by the other manufacturers. The Commis sioner disagrees and discusses below the reasonableness of expecting migration from each of the manufacturers' bottles. a. Monsanto Bottle. According to Mon santo, its "old" bottles were manufac tured between May 1975 and February 1977 and contained an average of 15.2 parts per million (ppm) of residual acry lonitrile monomer and a maximum of 39 ppm (Monsanto App. at 2, 16). The Ini tial Decision cites tests that directly ob served migration from the "old" Mon santo bottle (ID at 12-13). The residual monomer level In the walls of the actual test bottles was not determined, and can be assumed to be within ttie range cited by Monsanto for the old type bottle. In the "new" Monsanto bottle, the maxi- A 3' i 1 K: i k i \ . r. 1* I" i h FEDERAL REOISTER. VOL. 41, NO. US--FRIOAY, SEPTEMBER 23, 1977 i VRD 0002014072 48530 RULES AND REGULATIONS /"'"N 7A mum level ol acrylonitrile In the bottle nesses as well as the Bureau's. Mon acrylonitrile bottles justifies application Is 5 ppm, and the average is-3.3 ppm. (Monsanto App. at 3, footnote.) The only difference between the "new" bottle and the "old" bottle Is the difference in the level of residual acrylonitrile in the bot tle walls, l-e. an average of approximate santo's Dr. Dlxlsr discusses the diffusion principle md refers to the diffusion equations used to project migration as "reasonable and reliable." (M-82 at ff 8-10 and 25 and Appendix A.) Mon santo's witness Dr. Salame confirmed the of the diffusion principle to bottles hav ing lower concentrations of residual acrylonitrile, shorter storage times and lower use temperatures. The confirma tion of the diffusion process for acryloni trile was made under normal use condi ly a pin In the new bottle versus 15 ppm validity of the basic parameters affect tions and thus is clearly appropriate. The in the old. Nevertheless, Monsanto re ing the migration models (M-84 at if 4. confirmation may also be made using gards it as "a specious line of reasoning" 12-18. A-l. A-ll). The testimony of Bu food simulating solvents or using exag to expect migration from the new bot reau witnesses Holts (0-77 at 1 27) and gerated conditions that are appropriate tles because It has been observed in the Livingston (0-78 at If 5-16, Exit. A-C> for evaluating whether migration may old and argues that "resort cannot be provides additional support for the valid occur during Intended use. had to mathematical postulation of mi ity of the-migration models, including Once the applicability of the diffusion gration from the `new' bottles." (Mon their validity for projecting migration principle has been reasonably con santo App. at 17. 21, citing Monsanto from bottles with varying RAN concen firmed, projections based on the dif Brief at 28, 63-68.) trations on the basis of data from a par fusion process are sufficient to satisfy the NRDC argues in Its reply that Mon ticular bottle. burden of proof with respect to migra santo cannot make the difference be Monsanto points out in its Brief that tion, even though the amounts projected tween the bottles significant "merely by its models were Intended to estimate the to migrate are below the level of analyti designating one set of bottles 'new* and maximum migration that could be ex cal detectability. The reasonable expec the others `old*.'' (NRDC Reply at 13.) pected and that the equations underly tation of migration, arrived at on the (The estimates of residual acrylonitrile ing them do not anticipate changes in basis of the diffusion principle, is not un used by NRDC for the old bottle were migration behavior that might occur at alterable, however. The expectation can lower than those stated by Monsanto and low RAN levels and thus invariably pre be rebutted, but only If an adequate cited in this Decision, but the NRDC ar dict migration (Monsanto Brief at 25- showing is made that migration is not gument Is unaffected by this difference.) 27). Monsanto suggests that the premises reasonably to be expected. Such a show The Bureau maintains that acrylonitrile underlying the models make them inap ing could be made through a demonstra migration may reasonably be expected plicable to the present determination. tion that an unusual or different physical from all acrylonitrile copolymer bever The Commissioner believes, however, process occurs at some point that pre age containers on the basis of the prin that the models illustrate that^the rea vents migration. ciple of diffusion. The Commissioner concludes that mi gration of acrylonitrile monomer may reasonably be expected to occur In. all the acrylonitrile copolymer beverage containers at issue here. Including the sonable assumption, absent contrary In formation, is that the diffusion process observed to operate at levels at which migration is detectable continues to op erate even at the levels at which It is not detectable. Acceptance of the diffusion process as sufficient to show a reasonable expec tation of migration Is a corollary of the conclusion that a substance can be a food additive even though the amount migrating is below the level of detect "new" Monsanto bottle. The Commis It Is possible to envision conditions ability. If the amount of migration is be sioner reaches this conclusion because, under which the appllcabillty.of the dif low the level of detectability, the fact of as discussed below, he believes it is valid fusion principle to a particular packag actual migration cannot be shown. In to use extrapolation models, based on the ing material could not be extended stead, migration of the substance must principle of diffusion and confirmed through the full range of a controlling be projected from data obtained from through appropriate means, to show that physical parameter. For example, ex observed migration occurring at higher migration may reasonably be expected tremes in temperature may produce RAN levels and storage times and under the intended use of a food pack changes of state In the container or fluid higher temperatures. The projections aging substance, even though migration that in turn might induce discontinuities drawn from these data rest upon sci of the substance Is not analytically de In the solubility of residual acrylonitrile entific principles, notably that of dif tectable. monomer in the bottle walls or the con fusion. The diffusion principle pro Monsanto, the Bureau, and the other tained fluid. Also, a physical barrier vides a reasonable and scientifically valid parties to this proceeding submitted evi might be formed at the interface between means of making projections from the dence concerning the use of mathemati the plastic and the enclosed liquid. How available data and It is reasonably ap cal models to project migration of ever, the parameters that are important plied to the acrylonitrile bottles at Issue. acrylonitrile monomer from acrylonitrile in the analysis of migration of acrylo Based upon the foregoing analysis, the copolymer beverage bottles (M-82, M-84, nitrile monomer--time, temperature, Commissioner rejects the Monsanto ar 0-77. 0-78). All the models. Including and the concentration of RAN in the gument that the direct observation of Monsanto's, are based on the principle container--do not Interact in such highly acrylonitrile monomer migration from of diffusion and use the diffusion equa nonlinear ways with the diffusion process Monsanto's "old'' bottle is Irrelevant to tion to project levels of migration from over the observed range of these param the question of whether monomer mi bottles for which no actual extraction eters. grates In the "new" bottle. As noted, the data are available. The projections are Because of the possibllty that there are only difference between the "old" and made from a base of actual extraction nonlinearities in the diffusion curve un the "new" bottle is the RAN concentra data. The models Indicate that acryloni der extreme conditions, the Commis tion. and the diffusion principle, as con trile migration from acrylonitrile copoly sioner believes it important to confirm firmed by extraction data from the mer bottles varies in a linear fashion as the operation of the diffusion process "old" bottles, makes it reasonable to ex a function of three principal parame under circumstances reasonably related pect that there will be some migration ters--concentration of residual acryloni to the intended use of the substance at from the "new" bottle despite the differ* trile monomer In the bottle wall, time of Issue. The direct observation of migra cnce in RAN concentration. storage, and temperature of storage. All tion in the "old" Monsanto bottle under Monsanto has suggested some addi acrylonitrile copolymer beverage bottles nonnal use conditions in three studies tional arguments, discussed below, for contain at least some residual acryloni (0-79. 0-80 and 0-81), as found in the not applying the diffusion principle as a trile monomer, and the models Indicate Initial Decision, provides an appropriate basis for projecting migration from its that some migration of the acrylonitrile confirmation of the diffusion process and new bottles. monomer occurs under all anticipated conditions of use of the copolymer bottles. The validity of these diffusion models for projecting migration Is supported by the testimony of the manufacturers' wit Its applicability to other acyrlonitrile copolymer bottles. The Commissioner finds that the demonstration of migra tion at a temperature, concentration and storage time reasonably close to the values associated with normal use of 1. In Its Brief. Monsanto refers to a theory which says that when the concen tration of residual acrylonitrile mono mer is very low "the migration will be far below what the model predicts or there may be no migration at all." (Monsanto 8 FEDERAL REGISTER, VOL 42, NO. 1S5--FRIDAY, SEPTEMBER 23, 1977 RULES AND REGULATIONS 48531 Brief at 28, citing M-84 at T 18. Tr. 528.) may be made with respect to the "in be judged under the Intended use of the The theory Is principally that at low tended use" of a substance. This term substance, not the specific "conditions of levels the residual acrylonitrile monomer indicates that the showing may be made Intended use." Specific conditions of In will become bonded to the copolymer and with only a limited Inquiry Into the tended use are relevant to safety evalua not migrate. Even as articulated by Mon specific conditions of use of a substance tions, . not to food component status. santo, the theory raises only a possibility and that migration need not be shown Thus, use data need not be as clearly that there will be no migration. The individually with respect to each par defined to show migration as to show bonding theory is, furthermore, only ticular variation in the Intended use of safety or establish GRAS status. In any speculative: the Commission believes a substance. The definition clearly per event, the Administrative Law Judge did that it provides an insufficient basis for mits a conclusion to be drawn based on reach conclusions about the conditions of concluding that diffusion ceases and no a showing of migration in one specific ap use, finding 6 months storage at room migration occurs at some low RAN level. plication of a substance that migration temperature to be well within the normal (Tr. at 147-48; M-84 at 118; Bureau Brief may reasonably be expected from other conditions of use for all beverage con at 73.) similar intended uses of the substance. tainers, a finding that is adequately sup 11. Monsanto excepts to the Initial De (See Bureau Brief at 18-20). ported by the record. (ID at 9-11.) That cision not only because It discussed the b. Borg-Wamer Bottles, Borg-Wamer finding applies to apple Juice containers data from the "old" bottles--which Mon views the Initial Decision as accepting as well as soda containers. santo views as irrelevant--but also be the validity and reasonableness of migra In judging whether a reasonable ex cause it did not take sufficient account tion models only when actual migration pectation of migration exists, the Com of the extraction data for its "new" has been confirmed using the actual bot missioner has looked primarily at bottle--which Monsanto does consider to tle and the actual product. Since the whether migration may be expected from be relevant. (Monsanto App. at 18-23.) confirmation of migration cited by the the bottles under the conditions of use The extraction testing done by Monsanto Administrative Law Judge was obtained cited by the Administrative Law Judge as on its "new" bottles shows no detectable in tests with the Monsanto bottles, not being well within normal conditions of migration with a method sensitive to 10 the Borg-Wamer bottles, Borg-Wamer use, specifically storage lor 6 months at ppb. The levels of migration projected by excepts to the conclusion that the Borg- room temperature. The Commissioner migration models for the "new" bottles Wamer acrylonitrile bottles are food ad notes, however, that these conditions are under the conditions at which the "new1' ditives. (Borg-Wamer App. at 11.) not exceptional and do not make allow j I bottles were tested is below 10 ppb. (G-78 at 110, 11 Exhibit A.) Thus, Monsanto's contention that the extraction data for the "new" bottles is significant is simply a reformulation of Its argument that a substance Is not a food additive if the amount of migration cannpt be analyt ically detected. The Commissioner has rejected this argument for the reasons discussed above. Monsanto also points out that the Ad ministrative Law Judge disregarded its extraction study on the "new" bottles partly because of a supposedly conflicting preliminary report on extraction studies being conducted by the Bureau. (Mon santo App. at 20-23). The Commissioner believes that the Monsanto extraction data should be disregarded on the basis of the analysis set forth above and that it is unnecessary to disregard It on the basis of the "preliminary report." The Bureau has apparently undertaken some investigation to determine independently the levels of migration from the "new" bottles, including tests under more ex aggerated test conditions, but no reliable results are available yet and no written report exists. (Bredcr, Tr. at 161; Bureau Reply at 13-14) ill. Monsanto's position that the new bottle is not a food additive appears to be based in part Qn the argument that migration must be shown under the "ac tual intended conditions of use" and that this requires at least a showing of migra tion through extraction studies dime specifically on the new bottles. (Monsanto App. at 2-3, 15, 18-23.) The Com missioner disagrees. The diffusion models already discussed make it reasonable to expect migration from all acrylonitrile copolymer beverage containers. Further more, the food additive definition tn sec tion 201 (s> of the Act, 21 UJ3.C. 321 (s), does not require that either actual migra tion or a reasonable expectation of The Commissioner doubts that the Ini tial Decision should be read as narrowly as Borg-Wamer suggests. The Admin istrative Law Judge, having confirmed that migration occurs in acrylonitrile copolymer beverage containers under normal use conditions, apparently con sidered it reasonable to conclude that migration occurs in all acrylonitrile copolymer beverage containers. In any event, the Commissioner concludes that It is reasonable to expect migration to occur from the Borg-Wamer bottles based upon the analysis set forth in con nection with the Monsanto bottles. The Borg-Wamer bottles contain residual acrylonitrile monomer which, as the Commissioner's analysis demonstrates, may reasonably be expected to migrate. Contrary to Borg-Wamer's exception, confirmation of migration need not be made with the actual product, in its case apple juice. There Is nothing in the evi dence or in the diffusion equation which suggests that a change in the solvent material will cause a complete cessation of migration. In fact, the evidence is to the contrary. Borg-Wamer's own extration tests detected migration of acrylonitrile monomer from its own bottles containing apple juice (BW-1 and BW-2). These, tests were conducted under storage conditions reasonably close to those associated with normal use--specifically, storage for 14 months at room temperature; Borg-Wamer also excepted to the Ini tial Decision because It stated that the data on conditions of use were not clearly defined. Borg-Wamer considers this to establish that the Bureau did not meet what Borg-Wamer described as its bur den to show that acrylonitrile migrates under the "intended conditions of use." (Borg-Wamer App. at 7-8.) Under the food additive definition in section 201(s) of the Act, 21 U.S.C. 321 ance for atypical use of beverage bottles. He believes the conditions suggested by the Bureau for extraction testing for safety purposes of 6 months at 90* F to be more suitable conditions for extrac tion testing since they allow for atypical use and the variety of beverages for which the bottles could be used. c. Vistron Bottles. Vlstron acknowl edges in its exceptions that the "amount of acrylonitrile migrating from acryloni trile copolymer bottles is importantly a function of the amount of residual ac rylonitrile monomer in the bottle wall." (Vistron App. at 12.) Vistron states that the walls of its current bottles contain 1.7ppm of residual acrylonlrlle and that the amount of acrylonitrile migration into the contents can be projected to be 1.5 ppb after 6 months at 90*. and 1.7 ppb after 29 months at room tempera ture. Vistron apparently concedes that mi gration may reasonably be expected from its bottles and rests its exception primar ily on the grounds that the amount that migrates is small and even lower than that from the other manufacturers' bottles. The point appears to be that the amount of migration from the Vistron bottles is even more "toxlcologlcally In significant" than that from the other bottles. (Vistron App. ait 12-14.) Toxico logical insignificance is discussed later. At this point, the Commissioner con cludes that it Is reasonable to expect migration of acrylonitrile from the Vis tron bottles based upon both the Vistron projections and the general analysis given above In connection with the Monsanto bottles. d. Reliability of Methods Used To De tect Acrylonitrile. The Administrative Law Judge found that the Breder (G79), Gajan (G-80), and Sphon (G-81) analyses showed actual migration of acrylonitrile into Coca-Cola from Mon santo's 32 oz. acrylonitrile beverage migration be shown under the "actual (s>. the food component status of a sub bottles stored for &?? months at ambient intended conditions of use" of a sub stance (i.e. whether it migrates or may temperatures. The Commissioner agrees stance. Under the definition, the showing reasonably be expected to migrate) Is to and concludes that the three methods FEDERAL REGISTER, VOL 42, NO. IBS--FRIDAY, SEPTEMBER 33, 1977 v V k \ ' t V 1 H 0 H 0 7 0 0 0 QUA 18532 RULES AND REGULATIONS A taken together reliabily detect actual migration of acrylonitrile monomer and thereby provide adequate confirmation of the applicability of the diffusion prin ciple to acrylonitrile copolymer beverage containers. 1. Monsanto takes exception to the Ad ministrative Law Judge's reliance on these three studies for showing actual migration. (Monsanto App. at 44-48.) Monsanto argues that (1) the analyses do not meet PDA's own minimum re quirements for establishing the reliabil ity and applicability of analytical meth odology. (2) the amount of migration reported using the direct liquid lnjectton chromatography method is below the lowest reliable quantitation level for that method. (3) at this level of quantitation peaks on the chromatograph are not sufficiently distinguishable from back ground noise to support a conclusion . that they reflect the presence of acrylo nitrile. (4) the head space analysis tech nique is experimental, and (5) the confirmatory analyses performed by Gajan and Sphon use a concentration technique the reliability of which has not been adequately established. The Commissioner rejects these argu ments. First, the reliability of the three methods is attested to In the testimony of Breder, Gajan, and Sphon. The quali fications of thqse analysts were not chal lenged on cross-examination, and the testimony in the record does not contra dict their testimony that the methods are reliable. (See Bureau Brief at 61-62). Furthermore, even assuming that the Fourth, there is nothing in the record to contradict the testimony of the Bu reau's witnesses that the concentration technique used by Gajan and Sphon is a reliable method. Finally, the Monsanto comparison of its analytical method with the three re lied upon by the Administrative Law Judge is not persuasive. (Monsanto App. at 46-47.) It Is important to remember that the Bureau's burden is only to show that migration may reasonably be ex pected. The Administrative Law Judge relied upon the three methods discussed above only for purposes of showing mi gration--not for quantifying it. (ID at 12-14.) The Bureau places strict require ments on analytical methods used by manufacturers submitting data intended to show the safety of a food additive be cause, in that circumstance, the manu facturer must accurately show the quan tity migrating in order to establish a safe level. 11. Vistron excepts to the Initial De cision for falling to recognize that "back ground interference" in the laboratory "will substantially Increase" the levels of apparent migration detected in the Breder, Sphon and Gajan tests. (Vistron App. at 16-19.) As stated above, the Bureau's wit nesses were confident of the procedures they used, and this would encompass the possibility of background interference (0-79, 0-80 and G-81). To the extent the exception goes simply to the level of migration, the Commissioner regards the section of the decision. However, they also argue that the Bureau failed to meet its burden of proof on the GRAS issue as evidenced by the split of opinion cited by the Administrative Law Judge (ID at 36) and the absence of affirmative evidence of lack of safety to offset the positive opinions of safety offered by the manufacturer's witnesses. (Monsanto App. at 25-26.) These arguments misapprehend the kind of showing the Bureau must make to show that a substance is not generally recognized as safe under its intended conditions of use and are rejected by the Commissioner. Since acrylonitrile copol ymers were not used to fabricate bev erage containers before January 1, 1958, the GRAS status of acrylonitrile in bev erage bottles must be based on "scien tific procedures." Thus, GRAS status will be achieved only if (1) there is scientific evidence of the actual safety of acrylo nitrile in beverage bottles under their intended conditions of use of the same quantity and quality required for proof of safety under 21 U.S.C. 348. and (2) this scientific evidence is commonly known throughout the scientific com munity knowledgeable about the safety of food ingredients and is reflected in scientifically reliable published studies. See 21 CFR 170.30 (formerly 21 CFR 121.3). The Commissioner concludes that there is no scientific evidence of the quantity and quality required for proof of safety and, thus, even if the existing evidence were commonly known three methods of analysis at issue do not comply with the FDA Guidelines re ferred to by Monsanto, the Commis levels of migration as Important for safety issues, but not as pertinent to the determination of whether some migra throughout the knowledgeable scientific community--which It is not--it could not form the basis for a consensus that acry sioner. in determining whether there is a reasonable expectation that migration tion may be expected. (See Bureau Reply at41-42.) lonitrile has been shown to be safe. The existing evidence is incomplete, and will occur, is not bound to rely upon 3. GRAS Status Of Acrylonitrile Co safety risks ace indicated by the data analytical 'methods that meet FDA polymers in Beverage Bottles. The sec that are available. As discussed later Guidelines. The question is whether, on ond element of the food additive defini under the heading "Safety of Acrylo the basis of a fair evaluation of the en tion. upon which the Bureau bears the nitrile." the record shows that acrylo tire record, there is substantial evidence burden of proof, Is that the: nitrile is a frank teratogen in the rat, supporting a conclusion that acryloni substance Is not generally recog a tumorigen and probable carcinogen in trile does become, or may reasonably be nized. among experts qualified by scientific the rat. a possible carcinogen in man, expected to become, a component of food when put to its intended use. Non-com pliance with FDA Guidelines does not render an analytical method unreliable per se for this purpose. Hie testimony is convincing that the tests were reliable. Second, the arguments that the training and experience to evaluate Its safety, as having been adequately shown through scientific procedures (or. In the case of a substance used in food prior to January 1, 1958, through either scientific procedures or experience based on common use In food) to be safe under the conditions of Its In tended use; 31 U.S.C. 331(g). and a mutagen in several test systems. The Commissioner agrees with the Bu reau's contention that under these cir cumstances It Is inconceivable that the scientific community could recognize that any level of acrylonitrile has been shown through scientific procedures to amount detected using the chromatog raphy method is below the lowest re liable quantitation level for that method and might not be distinguishable from background noise Is not persuasive. The chromatographic analysis Is being relied upon not to show the quantity of acry lonitrile migrating but only the fact of migration, and the polarography and mass spectrometry analyses of Gajan and Sphon confirm the chromatographic finding of actual migration. Third, the experimental data of the head space head space is not significant since the results obtained by the estab lished FDA chromatography technique , are similar to those found using the head space analysis and, again, were confirmed by the polarography and mass The Administrative Law Judge con cluded that the use of acrylonitrile co polymers in beverage containers is not generally recognized as sale (GRAS) be cause "the scientific community at large has not considered this question," the witnesses In this proceeding "represent almost evenly split opinion on the ques tion." and ``It has not been convincingly shown that the scientific community would inevitably form a consensus in either direction should they consider the matter." (ID at 35.) The manufacturers take exception to the Administrative Law Judge's conclu sion on the GRAS issue. As stated above, their arguments are based primarily on their theory of toxicologlcally insignifi be safe. (Bureau Brief at 113-14.) When a food component has been found want ing under the safety criteria applied in the context of 21 U.S.C. 348. it cannot, as a matter of law, be deemed generally recognized as safe within the meaning of 21 U.S.C. 321(s), and thereby escape regulation as a food additive. Though the foregoing conclusion makes it unnecessary to consider the ex tent of knowledge in the relevant scien tific community about the safety of ac rylonitrile. the record supports the conclusion that it does not have general knowledge of the scientific evidence con cerning the safety of acrylonitrile. (Tr. at 590.) Under section 201 (s) of the Act. this alone is sufficient to justify the con clusion that acrylonitrile is not GRAS, spectrometry analysis. cance and will be discussed in the next regardless of the nature of that evidence. FEDERAL REGISTER, VOL. 42, NO. 185--FRIDAY, SEPTEMBER 23, 1977 RULES AND REGULATIONS 48533 Finally, in nofc a single usage by one of the manufacturing parties' witnesses of that any minimum amount of migration be shown, and the definition thus per ment. referred to above, that the Bureau did not submit enough evidence to show the phrase "generally recognized as mits any migration of acrylonitrile to that acrylonitrile is not GRAS. The man safe," or words to that effect, is ft stated that acrylonitrile U generally recognized satisfy the component element of the food additive definition. ufacturers believe the Bureau must prove that acrylonitrile is toxicologically sig as safe, instead, the witnesses state that at the levels present in beverages it "should be." "must be," or "ought to be" generally recognized as safe. (See, for 2. To the extent that-the term "toxlcologically insignificant" is intended by the manufacturers to be used inter changeably with the statutory concept of nificant or actually unsafe in order to show that it is not GRAS. Both the Bu reau and the Administrative Law Judge disagreed with this positon and the Com example, M-82 at D 5; M-90 at H 8; M-96 "generally recognized as safe" (M-82 at missioner rejects it. at 1 6.) Furthermore, on cross-examina tion, each of those witnesses confirmed fl 5), as the Administrative Law Judge apparently intended when lie incor As stated above, the term "toxicolog ically insignificant" has no legal sig what their testimony implied: that the porated it in his statement of the issues at nificance as an element of the food addi scientific community knowledgeable about the safety of food ingredients tho urging of Monsanto, its use is merely confusing, and the Commissioner tive definition. In establishing the food additive status of acrylonitrile, the Bu does not now generally recognize acry chooses not to use it as a substitute for reau is not required in any way to show lonitrile to be safe for use in beverage the statutory language. that it is actually unsafe or harmful, bottles. <Tr. at 466, 589-590, 690-691.) 3. If the argument is that a certain either under the "toxicologically insig On the basis of the foregoing analysis, quantity of a migrating substance, pre nificant" concept or as part of its burden tho Commissioner concludes that the sumably any amount that happens to of proof on the GRAS issue. The burden use of acrylonitrile copolymers to fabri fall below the current. level of detect is on the manufacturing parties to dem cate beverage containers has been shown ability. can be considered GRAS per se, onstrate the safety of a food additive by -by the Bureau not to be generally recog or safe per se. the Commissioner dis competent evidence, not on the Bureau nized as safe. The manufacturers' excep agrees. (Monsanto App. at 3-5, 9.) The to show lack of safety. Thus, the fact tions to the Administrative Law Judge's statute leaves no room for a contention that no Bureau witness concluded (hat conclusion on the GRAS issue are based that the presence of a migrating sub acrylonitrile is actually unsafe under in largely upon their theory of toxicological insignificance and will be discussed be stance at a sufficiently low level by itself warrants a conclusion that the substance tended conditions of use is irrelevant to the GRAS Inquiry, as is the testimony of low in connection with the Commission is GRAS, or sale, or unworthy of govern the manufacturers' witnesses that acry er's discussion of that theory. ment concern. The Commissioner be lonitrile should be considered safe under 4. Toxicological Insignificance. The lieves that the safety of a substance can intended conditions of use. GRAS status manufacturing parties except to the In not be decided on the basis of quantity depends on the existence of general rec itial Decision because the Administra tive Law Judge did not adopt their con alone. Safety depends on many factors, Including the toxicity of the substance, ognition of safety, arrived at on the basis of adequate scientific procedures. There cept of "toxicological insignificance" as and it cannot be determined simply on fore, countrary to the Monsanto asser a basis for concluding that acrylonitrile in beverage containers is not a food addi the basis of amount (G-88 at 117; G-78.) Safety is appropriately determined tions (Monsanto App. at 25-26), it was proper for the Administrative Law Judge tive (Monsanto App. at 23-27; Borg- through scientific procedures. It Is for to cite the even split of opinion on the is Wamer App. at 4-6; Vistron App. at this reason that the argument that total sue of acrylonitrile's safety among the 3-11). The concept of "toxicological in lifetime exposure to acrylonitrile from expert witnesses at the hearing for his significance" Is used in several ways by beverages bottles might be only Vs of a conclusion that acrylonitrile is not the manufacturing parties. Their cen teaspoon Is meaningless. (Monsanto App. GRAS. (ID at 36). If the Bureau's burden tral argument appears to be that at at 3-5.) No substance can be assumed to were to show that acrylonitrile is actually levels of migration below the level of de be "toxicologically insignificant" and unsafe, the even split of opinion might tectability a food packaging substance safe simply because it is present at levels warrant a conclusion that the Bureau should either be considered not present below the level of detectability. It follows had not met its burden, but since the In the food or, if considered to be pres that general recognition of safety cannot burden is merely to show lack of general ent, be deemed "toxicologicaUy insig be based on quantity alone. Even if a recognition of safety the even split of nificant" and therefore not subject to certain quantity of a migrating sub opinion is highly probative of the con regulation as a food additive. (Monsanto stance (e.g., any quantity below the level clusion that acrylonitrile is not GRAS. As Brief at 23-27; Borg-Wamer Brief at 25- of detectability) were determined to be already discussed, the record clearly es 30.) safe through scientific procedures, GRAS tablishes that acrylonitrile Is not gen The term "toxicologically insignifi status would not be achieved unless such erally recognized as safe. (Bureau reply cant" does not appear In the statutory evidence were known generally to ex at 15-17.) definition of food additive, 21 U.S.C. 321 perts qualified by training and experi 5. The manufacturing parties argue at (s), and the Commissioner concludes ence. several points that the Bureau has in the that the term has no legal significance as 4. At least Implicit In some of the argu past used a "virtually nil" or "toxico an clement of the food additive defini ments made by the manufacturing par logically insignificant" concept In decid tion. Food additive status should be ties is the novel position that the term ing the food additive status of some judged by the statutory criteria already "toxicologically Insignificant" describes packaging materials. Though such lan noted relating to whether (1) use of the a third element of the Bureau's burden guage had no doubt been used inform substance "results or may reasonably be of proof on the food additive status Is ally in the past by Agency employees in expected to result" in its becoming a sue, i.e., that the Bureau must show the describing food additive status, it has component of food; and (2) at the level anticipated levels of migration to be ac never been Agency policy to use such at which the substance does become a tually unsafe or harmful. (Monsanto language in any of the ways suggested by component of food it is generally recog App. at 23-25; Borg-Warner App. at 2-6.) the manufacturing parties in this pro nized as safe. It is against these statu The manufacturers in effect argue that ceeding. The phrase does not replace the tory criteria that the manufacturers' (1) the Bureau, in proving the food addi statutory definition of food additive. various arguments about toxicological tive status of acrylonitrile, must show When used at all, it has been to describe insignificance should be evaluated. not only that it may migrate and that it the situation In which any occurring mi 1. To the extent that the term "toxi is not GRAS but also that acrylonitrile is gration will be below the level at which cologically insignificant" is used to re "toxicologically significant" (i.e., actually the substance can be considered to be fer to some level of migration so low that unsafe or harmful) at the anticipated safe (0-87 at If 6, 7) or has been found it must, as a legal matter, be treated as levels of migration, and (2) a showing of to be GRAS under the applicable legal zero migration and, thus, not a compo lack of safety is part of the Bureau's standards. nent of food (M-82 at 14). its use is in burden of proof on the GRAS issue. 6. Borg-Warner excepts to the Ad consistent with the definition of food ad These views about the burden of proof ministrative Law Judge's references to ditive. The definition does not require account for the manufacturers' argu acrylonitrile as a "known toxic sub- FEDERAL REGISTER, VOL 42, NO. 185--FRIDAY, SEPTEMBER 23, 1977 f i B H J Z I f l QJLjL? r.. * t 48534 stance" in discussing Its food additive status. (ID at 8.) Borg-Wamer contends that the Bureau has failed to prove that small amounts of acrylonitrile are known to be toxic. Moreover, according to the exception, the Administrative Law Judge ignored the concept of toxicological in significance in evaluating whether acry lonitrile is a known toxic substance. (Borg-Wamer App. at -6.) The reference in the Initial Decision to known toxic substances was appar ently a shorthand statement that acrylo nitrile Is not generally recognized as safe at the levels at which It migrates. (See Bureau Reply at 34-35.) The phrase may also have been intended to reflect the discussion in the Initial Dclsion, quoted above, about the policy concerns under lying the Act (ID 13-14) as well as to Indicate that safety questions exist about acrylonitrile. The phrase does not, and should not, Indicate that the Bureau has the burden to prove toxicity before a substance can be considered to be a food additive. The Administrative Law Judge recognized that the migration of any amount of a substance is sufficient to make it a food additive, unless the sub stance is generally recognized as safe in the amount In which it migrates. As stated above, the Commissioner has con cluded that the Bureau Is not required to show acrylonitrile to be toxic at the levels at which it may be expected to mi grate. only that It Is not generally recog nized to be safe at those levels. Once again, there was no need for the Admin istrative Law Judge to consider the "toxicological Insignificance" in evaluatIng-acrylonitrtle since that term Is extrastatutory and has no separate meaning in this proceeding Independent of the food additive definition. 7. The manufacturers also have ar gued that the "toxicological insignifi cance" concept reflects the Sensitivity of the Method (SOM) regulation (21 CFR 500.80-500.98) which the Commissioner _ has promulgated for determining residue levels for animal drugs. (Monsanto Brief at Appendix H.) The Commissioner dis cusses later why It is inappropriate to apply that regulation in this proceeding, or to decide in this proceeding whether the regulation could ever be applied to food additives. The Commissioner does rTOte at this point, though, that the man ufacturers' concept of "toxicological in significance'' would Impose a burden of proof, and a burden of developing safety data, contrary to that which exists under the Sensitivity of the Method regulation. Under that regulation, the proponent of - an animal drug must come forward with complete and adequate tests in two test animal species as one step in applying the procedures in the regulation. Simi larly, under the current statutory scheme, the proponent of a food addi tive must come forward with data ade quate to establish safety and has the burden of proof on safety. If the manu facturers' view of toxicological insignifi cance were adopted for food additives, PDA, in establishing food additive status, would have the burden to show that amounts of a substance migrating below RULES AND REGULATIONS the level of detectability are unsafe. Only U that burden were met and the sub stances were not generally recognized as safe would the manufacturers have the burden to establish its safety. This re versal of the burden of proof would be inconsistent with the policy implicit in both the SOM regulation and the current statutory scheme governing food addi tives. 5. Comparison Of Acrylonitrile With Other Containers And Substances. Mon santo and Vistron take exception to the Administrative Law Judge's failure to (1) consider the safety of acrylonitrile copolymer beverage containers In com parison with other food containers and other substances which may become a component of food, and (2) treat acrylo nitrile copolymer beverage containers In a manner they consider consistent with the treatment afforded other such con tainers and substances. (Monsanto App. at 27; Vistron App. at 2-11.) Monsanto also argues that If the diffusion principle U sufficient to show that acrylonitrile may migrate, it will be sufficient to show that other substances and packaging ma terials migrate, and, as a result, subject these other materials to the food addi tive provisions of the Act. Monsanto maintains that the standards applied to acrylonitrile bottles should be applied consistently and be the same as those that have been applied to other sub stances. (Monsanto App. at 9-10.) The Commissioner believes consistency is an important regulatory objective, and he endeavors to give equivalent treat ment to similar cases. The Agency, though, cannot be prevented from taking warranted regulatory action on the grounds that It has not yet acted on other cases that are argued to be the same. As the Bureau states (Bureau Re ply at 17): The Act does not require the Bureau to develop a comprehensive scheme of relative toxicity for &U beverage containers before it can proceed against one. Such an approach would immobilize the Bureau's efforts to en sure the Bafety of food Additives. The Gov ernment is not required to "choose between attacking every aspect of a problem or nbt attacking the problem at hll."Dandridffe v. Williams, S97 U.S.C. 471. 486-87 (1968). It is true that the Administrative Law Judge's Initial Decision did not consider other food containers or other sub stances, Indicating implicitly a conclu sion that the status of such other con tainers and substances has no direct bearing on the issues In this proceeding. The Commissioner agrees with this con clusion. The issue in this proceeding is not acrylonitrile beverage bottles rank with other food containers in terms of safety, but whether acrylonitrile bever age containers are lawfully approved un der the food additive provisions of the Act. Likewise, the regulatory status of the specific substances mentioned in the Monsanto and Vistron Appeals, selenium and afiatoxln, are Irrelevant to an evaluation of the safety of acrylonitrile beverage containers. The regulation of afiatoxins as a poisonous and deleterious substance under section 406 of the Act Is not relevant to the regulation of acrylonitrile under the food additive pro visions of the Act. The selenium animal feed regulation, 21 CFR 573.920, is con sistent with the food additive provisions of the Act. Selenium is distinguishable from acrylonitrile on several grounds, and the Commissioner has stated that the available Information does not sup port Its classification as a carcinogen. (See Federal Register of April 27. 1973. 38 FR 10458; Blumenthal Tr. at 816-17.) Other food packaging substances may also be food additives, because they may migrate at low levels and are not GRAS. (See Bureau Reply at 5. n. 6.) Food ad ditive status does not necessarily mean that the substance cannot be used in food. If a substance can be shown to be safe, its use is permissible pursuant to a food additive regulation. In the next section of this Decision, the Commis sioner examines whether the safety of acrylonitrile has been established. III. Safety or Acrylonitrile Once acrylonitrile has been shown to be a food additive, its use can be per mitted pursuant to a food additive regu lation under section 409(c) of the Act. 21 U.S.C. 348(c). only if the manufac turers can show through scientific proce dures that its use is safe under the In tended conditions of use. The Administrative Law Judge de scribed the process for making these safety determinations as follows: The normal procedure In determining the safety of a product Is to And, through animal experimentation, the possible safety hazards that could result from Its usage, determine levels at which these effects do not occur, and through the utilization of an appropri ate safety factor, extrapolate that data to the human case. (ID at 35.) Stated below are the Administrative Law Judge's general findings about the kinds of safety risks indicated for acry lonitrile and his general basis for believ ing that no safe level for humans can be determined with respect to any of these risks through the use of any safety factors or risk assessment methods. The Commissioner also has stated his general conclusions. Thereafter in this section, the Commissioner discusses each specific type of risk Indicated for acrylonitrile, examining the Administrative Law Judge's basis for believing the risk exists, the contentions of the participants that relate particularly to that risk or the existence of a no-effect level for it, and the Commissioner's specific responses. In the last part of this section, the Com missioner examines the participants' other contentions relating to risk assess ment and consumption patterns. - A. GENERAL FINDINGS AND CONCLUSIONS The Administrative Law Judge made the following determinations: An overview of the analysis of these studies shows a pattern of adverse effects, (l.e., simple toxicity, mutagenicity, tera togenicity and tumorlgenlclty) which ap pear in Increasing numbers and at lower dosage levels os the studies progress. The most recently reported of these effects are A 3 C*. 6=S --I ON FEDERAL REGISTER, VOL. 47, NO. 165--FRIDAY, SEPTEMBER 23, 1977 RULES AND REGULATIONS 4S535 of ft severe end Irreversible nature. (ID at 28) Tbe record consists of two types of datar short*term experiments designed to demonstrate acute toxic effects and long term studies which Implicate AN as mu tagen, a teratogen, a tumorigen and possibly as a carcinogen. (ID at 36.) An analysts of the risk-assessment meth od* described reveals that their accuracy depends on the validity of the no-effect level upon which they are based. None make provisions for overcoming incomplete studlee, studies In one species, or assessing the cumulative impact of multiple effects. It appears that unless the final outcome of these ongoing chronic animal studies can be accurately predicted, the use of a safety factor of the types described here does not overcome the basic defects in the manufac turers' conclusions as to the safety of AN. The 20-year latency period found In the development of cancers attributed to expo sure to AN in the du Pont worker study suggests that the animal case of late develop ment of tumors finds a counterpart In the human situation. Since there Is a parallel in tumor types developed In animals in the in halation and Ingestion situations, the use of positive human Inhalation data to predict potential human ingestion effects is Justified. Using a simple multiplication of safety esti mates from Incomplete studies to assess a safe human consumption level does not cure the basic defects In the animal studies and the storage and consumption data. In extrapolating from a population of rats which generally runs under 60 in number to a population of millions of humans, that which would appear as borderline effects at low dosage levels could easily compute to several thousand affected humans. The re sults of the animal and human studies raise health questions which are more lnslduous than simple acute toxicity. A finding of tox icological insignificance would be premature pending a more complete evaluation of the effects In question.- When a history of the research shows effects appearing only after longer exposure, times. Incomplete studies cannot form a legitimate basis for determin ing a no-effect dose for a human population which could conceivably encounter the chemical on a chronic basis. (ID at 32-33.) e e ee It appears from an evaluation of the testi mony and evidence presented in this pro ceeding that a finding of safety for AN at this time would be premature and unwar ranted, particularly. In view of the severe irreversible effects which have been poten tially attributed to AN. This Is so, despite the fact that such a determination might well have been possible were the advene effects of AN limited to acute toxicity. (ID ac 38; see also ID at 23.) After reviewing the record, Including the findings of the Administrative Law Judge, the Commissioner concludes that acrylonitrile Is a frank teratogen In the rat, a tumorigen and probale car cinogen in the rat, a possible carcinogen in man and a mutagen in several test systems. A safe level for human con sumption cannot be determined because of the inadequacy of the present data and the probable carcinogenicity of acrylonitrile. B. SPECIFIC RISK AREAS 1. Teratogenicity---a. Initial Deci sion,---The Initial Decision found the DOW-MCA study * to be the only fullyreported study available investigating the teratogenic potential of acrylonitrile (i.e. the potential to cause birth defects) and concluded that this study showed acrylo nitrile to be a teratogen In the rat (ID at 21, 23). The Administrative Law Judge stated that the one-generation DOWMCA study "shows such effects as miss ing vertebrae, short tails, and right-sided aortic arches." (ID at 21.) The Administrative Law Judge found that the effects of teratogenicity are severe and irreversible (ID at 23). In reaching his conclusions, the Adminis trative Law Judge stated that determi nations of safe levels for tumorigenicity and teratogenicity Involved: not a matter of passing malaise, with dis appearing symptoms, but rather a heritage of cancer and birth defects. In view of euch seri ous possibilities. It would be unwise to base such a weighty determination on half-com pleted studies. (ID at 23.) As stated above, the Commissioner agrees with the Administrative Law Judge's conclusion that it would be pre mature to derive a safe level cm the basis of the presently incomplete date for any of the chronic effects posed by acrylo nitrile. Including teratogenicity. b. Exceptions and Commissioner's Re sponse--1.--In their exceptions, Mon santo and Borg-Wamer maintain that a safe level for humans can be derived from the no-effect level shown for rats in the DOW-MCA study (Monsanto App, at 56, 63-59, 64; Borg-Wamer App. at 1819). They criticize the Initial Decision for not pointing out that the Bureau agrees that a no-effect level for rats exists in the DOW-MCA study at the lowest feed ing level. They contend that on the basis of this no-effect level and other date a safe level for human consumption can be derived using either a Mantel-Bryan risk assessment analysis or the 1000-fold safety factor that the Bureau uses in practice on frank terata. Monsanto also criticizes the above-quoted statement of the Administrative Law Judge for being an "unscientific appeal to emotion" and for being "misleading" in its reference to "half-completed" studies since the DOW-MCA study, among others, is com plete. (Monsanto App. at 68-59.) The Commissioner disagrees with these contentions Insofar as they main tain that a safe level for humans should be established based on the DOW-MCA Study or the other information presently available. The present date are inade quate to derive a safe level for humans with confidence. The DOW-MCA study is Indeed complete, and It has a no-effect level for rats at the lowest level, as both the manufacturers and the Bureau agree. However, It would be inappropriate to project a no-effect safe level for humans based on this one rat study, no matter what safety factors or risk assessment methods are used. At least another com plete study of the teratologic effects in another species is needed, and a repro- JThs DOW-MCA stud; (0-68) la also re ferred to In the record as Dow study, and MCA-sponsored teratogenicity study. duction study would be desirable. (Tr, at 195-98.) The teratogenic effects of a^compound may vary from species to species. (Tr. at 208). Another animal species may be more sensitive to the teratogenic poten tial of acrylonitrile than the particular test species used. The relative human sensitivity to the teratogenic potential of acrylonitrile is not known and, of course, humans cannot be subjected to feeding studies Involving potential tera togens. Thus, the risk to humans must be projected from animal tests. (Tr. at 184185, 214). In view of the seriousness of the risks posed, acrylonitrile must be tested in at least one more animal species before any conclusions can be confi dently projected about its teratogenic potential in humans. (Tr. at 195-199, 201, 208-209, 211; 0-86 at fllb, 14a.) If another species were tested and found more susceptible to the teratogenic ef fects of acrylonitrile, the no-effect level for teratogenesls could be lowered. With out more testing. It is Improper to assume that where the no-effect level is now is where it should stay for purposes of estimating a safe level for human con sumption. (Bureau Brief at 82-85, 111-- 112, 138-139.) The Commissioner believes that no safety factor should be applied until ade quate studies are done to establish the level of risk posed. The 1090 fold safety factor referred to In the exception is a higher than normal safety factor which the Bureau ordinarily uses when the data show frank teratogenic effects on animals. (0-87 at tf 14. 15.) If the data show positive teratogenic effects, the Bureau requests date from a teratology study in at least another species before determining a safe level (0-87 at 115). The Bureau uses a larger than normal safety factor In computing the safe level due to the greater level of concern with reproductive effects. The concern that warrants a higher than ordinary safety factor also warrants not applying any safety factor, no matter how high, until the date base for acrylonitrile is com plete enough to evaluate the risk posed with an adequate degree of confidence. (Tr. at 187, 195-99.) Another complete study in another species is needed for that purpose. The Mantel-Bryan analysis referred to In the exception was applied to the teratology data by Monsanto's witness Dixler, based upon the "analogous" use of that analysis in FDA's Sensitivity of the Method regulation. (M-82 at V; 3143; Monsanto Brief at 85-86.) That regulation has a different purpose than that proposed for It by Monsanto, and the Commissioner discusses later his reasons for believing the SOM regula tion and risk assessment analysis gen erally to be inapplicable in this proceed ing. It Is particularly pertinent to note at this point that the SOM regulation provides for complete and adequate studies in two test animal species (42 FR at 10417, Sec. IV, coL 3; 10431, 21 CFR 500.87). Thus, the regulation does not support the determination of safe levels for teratologic effects through a Mantel- FEDERAl REGISTER, VOl. 42, NO. IBS--FRIDAY, SEPTEMBER 23, 1477 VRD 0002014077 4 48536 RULES AND REGULATIONS Bryan analysis on the basis of incom plete data. The Commissioner believes that the Administrative Law Judge's reference to "half-completed" teratology studies was to the requirement for an additional stud; before the data brae would be com plete enough to determine a safe Level for human consumption of acrylonitrile. In this sense, the statement is correct. The Commissioner also agrees with the Ad ministrative Law Judge's remarks con cerning the seriousness of tije risk of birth defects and cancer and his conclu sion that It warrants special care and a conservative approach in the making of safety determinations. His remarks in this regard were not an unscientific ap peal to emotion. Rather, they reflect the policy concerns that underlie the Act, and were appropriately made in evaluat ing the safety issues in this proceeding. U. The manufacturers make further exception to the finding of teratogenicity based on the DOW-MCA study because the teratogenic effects seen at the higher dose levels have been "seriously ques tioned" both by the experimenters and by Bureau witnesses. (Monsanto App. at 606i: Borg-Warner App. at 18-29.) The questions relate to whether the maternal toxicity observed at the higher levels could have resulted from overdos ing and whether that maternal toxicity "accentuated" the fetal deformities. In addition, some question was raised about the presence of rat mumps and the effect it might have had. The reviewers of the study believe that the malformations were not the effect of maternal toxicity alone and that any effects of rat mumps were probably negligible. (0-58 at 1415.) The Bureau recognises that ques tions can be raised about this study. (Bureau Reply at 31.) These questions, though, do not dispel the concerns about the teratogenic potential of acrylonitrile. As Dr. Collins, the Bureau witness whose testimony Is cited by the manufacturers, stated after being cross-examined about these questions: porary standards which establish that acrylonitrile 1b Bafe. if this study is in valid. It means that at least another fully adequate and complete study must be done before acrylonitrile can be consid ered to be safe. iil. Monsanto also contends that a safe level for humans for the teratogenic and other adverse effects of acrylonitrile can be derived based on the no-effect levels seen in early chronic studies, specifically the Tullar study and the Evirbely and Floyd study.4 (Monsanto App. at 59.) The Tullar study is inadequate by con temporary standards, as discussed in more detail, below. The Svlrbely and Floyd Study was never fully reported; only a preliminary report was introduced into evidence. (ID at 23.) The Bureau has "never relied" on this study because "the data were never fully reported and a re liable appraisal of the results is there fore not possible" (Bureau Brief at 94). The Commissioner concludes that the ab sence of a full report from the Evirbely and Floyd study makes it an inadequate basis for use in determining a safe level for the teratogenic risks or other adverse chronic effects of acrylonitrile. Hie Ad ministrative Law Judge seems to have regarded the preliminary results of the Svlrbely and Floyd study to suggest a teratogenic potential for acrylonitrile. (ID at 21, 23.) This may be. but the Com missioner believes that the teratogenic potential of acrylonitrile is sufficiently demonstrated by the DOW-MCA study, without refemce to Svlrbely and Floyd. Finally the Commissioner notes that even if a no-effect level for human tera togenic were to be derived for acryloni trile based on the present data, the com pound could not be considered to have been shown to be safe. As discussed below in this Decision, no adequate studies exist to establish the safety of acryloni trile with respect to the other chronic risks it poses, and there are Indications that acrylonitrile is a probable carcin ogen. 2. Tumorigenicity--initial Deci sion.--The tumorigenicity of acryloni at 20). A subsequent interim report (0-73) diagnosed the ear tumors, CNS tumors and mammary masses as malig nant (ID at 20 and see ID at 22. Yodaiken, NRDC-94). The Administrative Law Judge stated that the "interim re ports from this study become increas ingly alarming with the passage of time." (ZD at 22-24.) He also noted that "these irreversible effects appear regularly only after a year's exposure time and Increase beyond that point (as evidenced by the Increase In tumors and the percent clas sified as cancers in each interim report)(ID at 28.) He concluded that the determination of a safe level for hu mans would be "inappropriate on the basis of the available evidence which in dicates potentially severe adverse effects from the ingestion of AN." (ID at 39.) b. Exceptions And Commissioner's Re sponse. The manufacturing parties be lieve that acrylonitrile can be considered to have a safe level for humans with re spect to tumorigenicity. I. They argue that a no-effect level exists at the lowest level in the Dow chronic feeding study (Monsanto App. 56-57, 61-64; Borg-Wamer App. at 1617). They dispute the number of tumors observed at the lowest level, pointing out that NRDC witness Yodaiken (NRDC-94 at 17) believed only one rat, not two, had developed CNS tumors at the lowest feeding level. They also cite the testimony of a Bureau wit ness that the tumors at the lowest level may have occurred spontaneously and might be "washed out" when the study is complete. (Cueto. Tr. at 290). They urge that a 1000-fold safety factor or a Mantel-Bryan analysis be applied to the lowest feeding level in this study to derive a safe human level, and they complain that the Initial Decision did not state their projected safe levels and their positions correctly. The Commissioner rejects the conten tions that a safe level should be derived for tumorigenicity based on the incom plete Dow chronic feeding study. The data available from the study at this 1 would not allow facrylonitrile! to be trile is shown by the interim reports point Indicate that acrylonitrile is a used until I had the complete data you from the on-going Dow chronic feeding tumorigen In the rat, and probably a have some very serious abnormalities here that are very rare, and they seem to be related to the compound, and I would not want the study' in the rat The report (0-57) of the Interim sacrifice "showed" a higher carcinogen as well, since some of the tumors have been diagnosed as malig stuff to be utilized until we had the complete incidence of "subcutaneous masses" in nant. It is conceivable that the tumor toxicological data. CR.stl97.) the mammary region, and "masses of effects at the lowest level could be It would be rash. In view of the rare and serious abnormalities seen in this study, to assume that the study is Invalid, and that the teratogenic effects observed are all due to rat mumps or overdosing, without having complete data investigat ing the teratogenic potential of the sub stance. It might also be noted that Mon santo has described this study as "well designed." (Monsanto App. at 59.) This study, moreover, is the only ade quate Investigation Into the teratologic potential of acrylonitrile. If the manu facturers are correct In assuming that the questions they cite invalidate the study, it will not establish that acryloni the ear canal," "polyp formation and spontaneous and that at the end of the hyperplasia of the stomach" and "pro study there will be an equivalent num liferate lesions of the central nervous ber of tumors at this level in the control system" (CNS) at the higher or highest and test groups. It would not be appro feeding levels. (ID at 19.) A later interim priate, however, to speculate now on report (0-71), made in April, 1977. in what the result of the study will be and dicated that 2 rats at the lowest feed what the exact number of tumors will ing level had developed the CNS tumors be at this level. Since the study is not previously seen at the higher levels (ID* *yet complete, it cannot be assumed, con trary to the suggestion of the manufac a7Ti Tullar study (G-S4) is also referred turers. that there will be a no-effect level. to In the record as an early chronic study. Even If no tumors had yet been seen at * The Svlrbely and Floyd study (G-66) Is this level, it would be inappropriate to also referred to In the record as the Svlrbely study, a long-term study, and an early chronic study. 'The Dow chronic feeding study (G-57) Is make this assumption. If a safe level for tumorigenicity is to be established for humans. It must be done on the basis of trile has been shown by the manufac also referred to In $he record as: Dow chronic complete data, not on the basis of a sin turers to be safe through scientific pro cedures. There are no other chronic rat feeding study. Dow chronic ingestion study. Dow rat feeding study, Dow ingestion study, MCA Ingestion study, chronic inges gle half-complete study. Furthermore, in view of the probable carcinogenicity in studies considered adequate by contem tion study, long-term study, Norris study. dicated by this study so far, it would be A FEDERAL REGISTER, VOL 42, NO. 185--FRIOAV, SEPTEMBER 23, T97T RULES AND REGULATIONS 48537 Inappropriate to postulate any safe level pending resolution of whether acrylo nitrile is indeed a carcinogen In animals. 1L The manufacturing parties contend that a dose-response relationship and a no-effect level for tumorigenicity and the other chronic toxic effects of acrylonitrile have been shown by other studies, specif ically the Tuller study, the Svirbely and Floyd study, and the Maltonl studies (Monsanto App. at 53-54; Borg-Warner App. at 13. 15-16). a. The Tull&r study was reported in 1947 and is inadequate by contemporary standards (0-87 at 118), because of the small number of animals in the experi discuss it further and gave it no appar ent weight. The Commissioner agrees that the Maltonl studies described In Dr. Levinskas' testimony should properly be given no weight. The Maltonl data are Incomplete and the studies have not been officially reported by Maltonl. Further more, given the strong evidence that acrylonitrile is a carcinogen. Maltonl's unreported and incomplete data cannot be used to establish a "safe level" for acrylonitrile. The Act does not acknowl edge the- existence of a "safe level" for a carcinogenic food additive. iv. Monsanto also excepted to the Dow chronic feeding study on the grounds 0-85 at II 7-8.) The higher incidence of malignant tumors in the treated mmai at the 12 month stage Implicates acrylo nitrile as a carcinogen in rats. Thus, the ongoing Dow chronic feeding study shows not only that acrylonitrile Is tumorlgenic in the rat, but also that It probably car cinogenic as well. Since this study is not yet complete, it cannot be finally deter mined that acrylonitrile is a carcinogen. However, in view of the tumor findings so far In that study and the other indi cations in the record that acrylonitrile may be carcinogenic, the Commissioner believes acrylonitrile cannot now be con sidered to have been shown by the man mental groups, the use of male animals that its "design and validity has been ufacturers to be safe. only, and the failure to use a full range challenged." (Monsanto App. at 53.) The The Administrative Law Judge found of toxicological criteria. (0-86 at tl 7.) criticism, as stated in the manufacturers' that there are other data In the record When considering the Tullar study, it Briefs, relates primarily to whether the indicating the possible carcinogenicity of should also be noted that the manufac maximum tolerated dose was exceeded. acrylonitrile; turing parties have excepted to the Ad ministrative Law Judge's reference to a report of the investigator in the Tullar study who commented that tumor-like There is convincing testimony that this study is appropriately designed. (Cueto, G-85 at I 5 and Tr. at 269-70.) The maximum tolerated dose is a cri The two inhalation etudiee, on* la rate and one in humans we also pertinent to discussion. The rats tested showed the -- central nervous system tunion, ear to* growths were seen in four rats. (Mon teria of the National Cancer Institute more and breast masses as did their counter santo App. at 65-66, citing ID at 36.) that is aimed at ensuring that enough parts in the chronic ingestion studies. This The manufacturers view the suggestion that the Tullar study "adds significance" to the recent tumorigenicity findings as being in conflict with the Commissioner's earlier assessment, quoted by the manu animals survive the test to draw mean ingful conclusions when negative results are found; when enough survive, or when positive results are found, as was true here, the fact that the dose was similarity makes comparisons between the groups appropriate. It also gives relevance to the human data, which showed a three fold Increase In cancer among workers ex posed to AN industrially. The necessity of chronic studies in unearthing these effects facturers, that the Tullar study did not exceeded does not Invalidate the study. is paralleled by the 20-year latency period demonstrate carcinogenicity. The Com (Cueto Tr. at 270-73, 297-98.) In any noted in the human case, as well as the sys missioner made the latter assessment event, the Administrative Law Judge did temic nature of the cancer, this similarity "because of the inadequacy of this study and the lack of dose relationship," and not find that the maximum tolerated dose was exceeded, and the Commis being noted by the du Pont researchers. (ID at 23.) because "the fact the tumors were of sioner concludes that it has not been Based on this the Administrative Law different kinds suggests no relationship established that it was exceeded. (See Judge concluded: to the agent fed." (Federal ofRegister Bureau Brief at 104-07.) Furthermore, While It would be Inappropriate to find .November 4, 1074 (39 FR 38907, 38908) even If the validity of this study were to AN unsafe or label It as a definite carcinogen quoted in Monsanto App. at 66.) The be undercut, this would not permit a on the basis of Incomplete testa, the interim Commissioner's assessment of the inade quacy of the Tullar study for evaluating the safety of acrylonitrile has not changed, and, at least on this matter, the Commissioner and the manufacturers are in agreement on the inappropriate- conclusion that acrylonitrile is safe since no other chronic studies that are ade quate by contemporary standards estab lish that acrylonitrile Is safe. v. Vistron excepts to the asserted mis interpretation by the Administrative Law results from chronic feeding and Inhalation studies raise considerable doubt as to Its safety and require Its exclusion from food. This is necessary to protect tbe public from the barm that could result from subjecting It to what is a potentially harmful aubetanee unless, or until the safety of such substance, ness of reliance on that study. Judge of Vlstron's position on the con or the lack thereof. Is reasonably established. b. The deficiencies of the Svlrbeiy and tamination of the acrylonitrile and the (ID at 24.) Floyd Study have already been discussed, and the study can no more be considered adequate for purposes of evaluating tumorigenicity than it can be for terato genicity. iil. The Maltoni studies are ongoing, acrylonitrile water solutions used in the Dow chronic feeding study and the effect of such contamination upon the results of the study. (Vistron App. at 14-16.) NRDC replies aptly to Vlstron's excep tion by saying "Ttiis argument does not b. Exceptions And Commissioner's Re sponse--1. The manufacturers object to the determination that acrylonitrile Is a potential carcinogen. (Monsanto App. at 57-58, 63, 66-68; Borg-Warner App. at 15-17.) Most of their exceptions have long-term studies on rats which have not explain the fact that acrylonitrile causes been dealt with already in discussing the been published or reported on in the cancer when Inhaled, or why acrylamide tumorigenicity data. However, the man ikumiu'r usual for scientific at mill's. Mon l the Impurity I would not be formed in ufacturers' suggestion that a safe level santo put great emphasis on the Admin beverages and pose a cancer risk to hu can be derived for acrylonitrile, notwith istrative Law Judge's failure to discuss mans." (NRDC- App. at 19.) Further standing its carcinogenic potential, war this study or give it any apparent weight. more, the contamination possibility rants further comment. As the Adminis (Monsanto App. at 14, 53-54, 67-68.) could not be viewed as explaining the trative Law Judge stated, if the data Dr. Levinskas testified that Maltonl has not released reports to anyone except the sponsors of the study. (Tr. at 757.) Apparently, Maltonl does not believe in releasing negative data before a study is completed "because they may turn out positive before the study is over" and he did not give any written or printed material to Dr. Levinskas. (Tr. at 779- 80.) The Administrative Law Judge origi nally excluded from the record Dr. Levin skas' account of this study. In his Initial Decision, he concluded that it should be admitted into evidence, but he did not tumorlgenic and carcinogenic effects in the Dow chronic feeding study until fur ther studies are done to establish the effects of the contaminant and to show that acrylonitrile is safe. (See Bureau Reply at 41.) 3. Carcinogenic Potential, a. Initial Decision And Commissioner's Analysis. The interim results from the Dow chronic feeding study show malignant tumors, including a highly important incidence of rare microgliomas (central nervous system tumors) and an increased inci dence of Zymbal gland tumors. (ID at 20 22; G-86 at II 14b; 0-87 at TH 25-28; establishes the carcinogenicity of acrylo nitrile the Delaney Clause would pre clude determination of any safe level for its use as a food additive. 21 U.S.C. 348 (c)(3)(A). When evaluating whether a possible carcinogen has been shown to be safe by the manufacturers, the policy embodied in that clause Is pertinent. Given the strong Indications that acry lonitrile is a carcinogen, the Commis sioner does not believe that the manu facturers can show a safe level for its use at any level, even assuming that their determinations were being made from adequate and complete studies. V= FEDERAL REGISTER, VOL. 42. NO. 185--FRIOAY, SEPTEMBER 23. 1977  ft H M U M # 48538 RULES AND REGULATIONS 11. NRDC contends that the Commis sioner should conclude now that acrylo nitrile la carcinogenic because of the finding of a high Incidence of tumors, including malignant ones In the Interim sacrifice in the Dow chronic feeding study. (NRDC Reply at 18-21.) The Commissioner believes the final determination of the carcinogenicity of acrylonitrile should await the final re sults of the Dow chronic feeding study. The determination of carcinogenicity should be based on a finding that the sole experimental variable between the control and experimental groups is the absence of presence of the test com pound. as well as upon a finding that the incidence of tumors is statistically significant. The statistical significance of the tumors cannot be determined until the study is complete and the number of tumors known. At this point, acrylo nitrile is a suspect carcinogen, but not yet a proven carcinogen. ill. Monsanto excepts to the Initial De cision for stating that no evidence has been submitted characterizing the me tabolites of acrylonitrile (Monsanto App. at 65, excepting to ID at 28.) Monsanto points out that it has submitted metabo lite data in support of Its theory that at low levels acrylonitrile Is metabolized in the body In a way that renders it harm less. The Commissioner notes, however, that the metabolism data submitted by Monsanto are not complete, and the company's own witnesses testified that more study Is necessary to prove their theory. (Tr. at 642, 691, 695: Bureau Brief at 107-10.) The Bureau also sub mitted testimony that more data is need ed before this theory can be considered. (Blumenthal, 0-88 at 11 9, 20.) Monsanto also criticized the Initial. Decision for not discussing the "uncon tradicted testimony" by Bureau witnesses Cueto and Blumenthal that the human body contains defense and repair mech anisms for carcinogenesis. (Monsanto App. at 14.) The very testimony cited by Monsanto provides a sufficient explana tion of why this theory cannot be used to postulate a safe level for acrylonitrile. Dr. Blumenthal testified that he person ally thought there might be such mecha nisms, but-- * for any individual compound, each of these parameters has to be determined in dividually. * Therefore, it would eeem to me that the important point to establish with respect to acrylonitrile is whether In fact there are any data available that would define a threshold for toxicologic responses In-v&ts, mice and humans to acrylonitrile ex posure. (G-S8 at f 16.) That data for acrylonitrile has not been submitted. Whether there is a threshold, and where it may be, is not now known. Accordingly, the Commis sioner concludes that acrylonitrile can not be considered to be safe based on Monsanto's as yet unsubstantiated theories. iv. Monsanto also excepts to the Ad ministrative Law Judge's reliance on the duPont workers study* (G-74) and the Dow inhalation study* In rats (0-71) in his evaluation of carcinogenicity (Mon santo App. at 67-58, citing ZD at 23.) The principal basis of the objection is that the Maltoni studies were not also discussed, but, in addition, Monsanto questions the relevance of these two in halation studies in view of the tumor distributions in the duPont study. The Commissioner has already dis cussed why It is Inappropriate to give any weight to the Maltonl studies. Hie rats in the 80 ppm group in the Dow inhalation study exhibited central nervous system tumors similar to those seen in the Dow chronic feeding study (G-87 at 129). The duPont workers study demonstrated an increased incidence of cancer among workers exposed to acrylonitrile (Kokoski, Tr. at 441). Although some of the tumor sites found In this study in hu mans differed from those found in rats in the Dow chronic feeding study, this dif ference was not unexpected since differ ent species often manifest different tu mor sites when exposed to the same toxic substances. (Kokoskl. Tr. at 441-42.) Inhalation studies are not directly rele vant to ingestion toxicity: however, the inhalation data "gives evidence of the systematic nature of the apparent carci nogenic effect of acrylonitrile after ab sorption." (0-87 at 1 29; Tr. at 440-42.) The Commissioner believes that the in halation studies suggest that acryloni trile is a carcinogen and that they pro vide evidence supportive of the interim results from the Dow chronic feeding study. All these studies taken together preclude a conclusion that acrylonitrile beverage containers have been shown to be safe. v. Monsanto also excepts to the Ad ministrative Law Judge's statements that the manufacturers used the duPont workers study to project a no-effect level. Monsanto maintains that this study is too flawed to be used for any purpose. (Monsanto App. at 55.) If any weight Is to be given to the results of this study, though, Monsanto regards them as con sistent with an apparent no-effect level. The Commissioner accepts Monsanto's exception claiming that it did not In fact Use this study to derive a no-effect level. He too is aware of the limitations of this study, including the fact that It is not complete. (G-87 at I30r Tr. at 820). Nonetheless he believes the Administra tive Law Judge was correct in giving this study weight in considering whether acrylonitrile had been shown to be safe because the study suggests that acrylo nitrile may be a human carcinogen. This study, taken together with the other data in the record, raises serious suspicions about the carcinogenic potential of acry- The duPont workers study (G-74) Is also referred to in the record as the epidemiologi cal study, human study. Inhalation study and duPont study. 'The Dow Inhalation study (G-71) Is also referred to in the record as the rat Inhalation study. lonitrile that need to be resolved by com plete and adequate studies. 4. Mutagenicity, a. Initial decision. The Administrative Law Judge summarizes his findings on the mutagenicity studies as follows: The present data raise certain implications of adverse effects of acrylonitrile. The micro bial system screening for animal carcino genicity has produced several clear, highly reproducible positive findings, with negative results occurring only wben a technique meant to overcome the problems of toxicity or volatility of AN was not utilized. In this situation, with the same number of studies finding positive results as those finding nega tive results, it would be improper to dis count the positive findings. This view results in a strong implication of acrylonitrile being a carcinogen in higher animal systems. (ID at 22.) b. Exceptions and Commissioner's re sponse. Monsanto excepts to the above statement on the grounds that the "great majority" of the mutagenicity tests have shown a negative response. (Monsanto App. at 50-52.) Monsanto makes this calculation by counting as separate tests the testing done on each of 33 bacterial strains used in the six mutagenicity tests discussed in the Initial Decision. Mon santo describes all the positive results as weak. The exception also maintains that the manufacturers have never contended that negative results automatically out weigh positive results, but that they be lieve, on the other hand, that positive results should not automatically out weigh negative. Monsanto also criticizes the failure to note that negative muta genicity results have been observed even when precautions were taken to prevent volatilization. The Commissioner finds that Mon santo's approach of dividing each study into its component bacterial strain is in appropriate, since, as the Bureau notes, "many of the strains were never tested by procedures designed to Inhibit the volatilization of the acrylonitrile nor de signed specifically to detect mutagens that are not strong enough to demon strate a mutagenic effect in the stand ard plate test." (Bureau Reply at 27.) Furthermore; Monsanto's effort to dis count the positive results by emphasizing that they were "weak" overlooks the fact that the volatilization process used makes it impossible to conclude whether acrylonitrile is a weak or potent mutagen." (Prival. Tr. at 228.> The mutagenicity testing is primarily important for indicating whether acry lonitrile may be a carcinogen. Microbial testing "has been shown to be about so percent accurate in predicting animal carcinogens." (ID at 17.) Consequently, the positive results are highly important in indicating the need for further in vestigation into the carcinogenic poten tial of acrylonitrile. The negative re sults do no dispel these concerns. (See ID at 17-19.) The laboratories reporting positive findings used appropriate methods (ID at 17-19), and nothing in the exceptions indicates that the positive A FEOERAl REGISTER, VOL 42, NO, 15--FRIDAY, SEPTEMBER 23, 1977 im iB Z B B fl difli RULES AND REGULATIONS 46539 tests are invalid, apart from the exist rate upon his conclusion that the SOM ence of negative results, some of which regulation is inapplicable to this pro can be attributed to the failure to use ceeding. techniques to prevent volatilization. The SOM regulation applies by its The Administrative Law Judge be terms to animal drugs and was Issued lieved that the positive results warranted under a statutory exception to the De the completion of a "cancer-profile test laney Clause (the so-called DES exemp protocol" before acrylonitrile could be tion) that permits the use of a carcino determined to be noncarcinogenic (ID at genic animal drug In meat producing 17). The Commissioner believes this con animals if no residue is found In the clusion is still valid. Complete animal meat by a method approved by the testing of the carcinogenic potential of Agency. (21 UB.C. 348(c) (3) (A), 360b acrylonitrile is needed to resolve the dif (d)(1)(H), and 376(b) (5) (B).> The ferences between the positive and nega regulation establishes criteria for how tive results in the highly predictive mu sensitive a method has to be before the tagenicity testing. agency will approve it. The regulation C. RISK ASSESSMENT AND OTHER CONTENTIONS provides for the use of the Mantel-Bryan procedure to calculate how sensitive the method must be. The SOM regulation 1. Monsanto contends that the Admin deals with the development of methods, istrative Law Judge ignored its position pursuant to a statutory provision, for that in making risk assessments correc determining whether residues of a carci tions can be made to compensate for in nogenic animal drug is present in meat, complete studies, studies in one species, not with determining a safe level for and cumulative effects. (Monsanto App. carcinogens, as some of the submissions at 68-69.) would have it. The regulation is cur This may indeed be Monsanto's posi rently under judicial review. (Animal tion. but it Is not the Commissioner's po Health Institute v. FDA, C.A. No. 77sition and the Administrative Law Judge 0806 (D. DC ), filed May 12, 1977.) rightly did not adopt it. Safety decisions The Commissioner believes that if the should be made on complete data. If SOM regulation were ever to be applied safety calculations are made on Incom to food additives it should be done by plete data, they inherently rest on a regulation after public notice and oppor guess that the further studies will not tunity for comment. It is by no means demonstrate a greater risk than the clear that the criteria and manner of cal compensation factor used. The Commis culation used for animal drugs should be sioner is not willing to make such as adopted for food additives. NRDC, for sumptions in connection with the severe example, maintains that the risk cri- risks indicated for acrylonitrile. It Is . teria in the SOM regulation are not suf particularly inappropriate in view of the carcinogenic potential of acrylonitrile. Furthermore, the validity of statistical extrapolations to determine safety risks from incomplete data is conjectural. (See Bureau Brief at 131-33,146-50.) 2. As part of his discussion of risk as sessment, the Administrative Law Judge discussed not only the traditional use of safety factors to determine a no-effect level but also the Mantel-Bryan proce dure which he equates with the "Sensi tivity of the Method" (SOM) regulation* recently Issued by FDA. <ID at 31, n. 20.) The Administrative law Judge correctly found, as already discussed, that no risk assessment procedure, including the Mantel-Bryan procedures, could be used because of the absence of complete and adequate animal studies from which U> extrapolate the risk level to humans with confidence. In its exceptions, Monsanto still argues for the me of the Mantel- Bryan procedure to calculate safe expo sure levels for carcinogens (Monsanto App. at 63), and all the participants de voted considerable attention to the SOM regulation In their briefs to the Adminis trative Law Judge. Accordingly, the Commissioner believes he should elabo ficiently conservative for food additives. (NRDC Brief at 49-50; see Bureau Brief at 121-22, 126-29, 133-36.) The general public has received no no tice that the SOM regulation might be considered for application to food addi tives in this proceeding. The notice of hearing did not mention this issue and, Indeed, the Commissioner did not antici pate that It would arise. A petition (77P0122) has recently been filed by the So ciety of the Plastics Industry requesting the Commissioner to Initiate a rulemak ing proceeding to adopt the sensitivity of the method criteria for use in regulating food additives. The legal issues regarding the applica bility of an SOM approach to food addi tives are extremely complex and difficult. (Bureau Brief at 120-21.) Hie only ex plicit provision in the statutory' section governing food additives which author izes the agency to establish methods for determining the presence of a substance is the DES exemption, which is limited in its applicability to animal drugs. (21 U.S.C. 340.) NRDC maintains that the SOM regulation cannot appropriately be applied in this proceeding or for indirect food additives (NRDC Brief at 48-50.) The Commissioner will not attempt to * The correct title of thle regulation Is Cri resolve these legal questions here because teria and Procedures for evaluating Assays he does not view this proceeding as a for Carcinogenic Residues. The Commis suitable forum to address the issue. More sioner will refer to It by the term Sensitivity of the Method (SOM) because this terminol ogy has been used in this proceeding, but he has reservations about the terminology, as noted in the Federal Reoiste* of February consideration of the issue is needed both by the public and by the agency before the applicability of the SOM regulation to food additives Is addressed and re 22, 1977 (42 PR 10412, 1Q427-2&). solved. 3. Monsanto states in its general excep tions that the degree of risk from small amounts of a substance like acrylonitrile can never be precisely known; it can only be estimated. Furthermore, safe levels can be determined by extrapolation from toxic effects seen at high levels, and the more data available the more confident the projections. Monsanto then states that more data exist on the safety of acrylonitrile than on any other beverage container migrant and, from this, con cludes that the Administrative Law Judge was wrong in believing that the degree of risk from acrylonitrile Is not presently known. (Monsanto App. at 2830.) The Commissioner agrees with Mon santo's position as summarized in the first two sentences above, but he finds the conclusion drawn by Monsanto to be un sound. Safety determinations are not precise and some uncertainties remain even when they are drawn from complete data. When incomplete studies or inade quate data are used to project risks, the estimates are more likely to be Incorrect. The adequacy of data is not determined, as Monsanto would have It, on volume or on a comparative basis. TO be adequate the data must satisfy scientific procedure and be sufficient to determine with con fidence a safe level for each type of risk posed. By that standard the level of risk from acrylonitrile has not been ade quately determined, as the Administra tive Law Judge found. 4. Monsanto has also excepted to the discussion of consumption patterns in the Initial Decision because it reports the contentions of the participants without accepting Monsanto's figure. (Monsanto App. at 48-50, citing ID at 29-30; see Borg-Wamer App. at 13.) Monsanto also objects to the Initial Decision for appar ently agreeing with the Bureau's criti cisms of Monsanto's estimate. The Administrative Law 'Judge deter mined that a safe level could not be es tablished for acrylonitrile in view of the incompleteness of the safety data. (ID at 33, 37 and 38.) Consumption patterns be come relevant only If a calculation is to be made of a safe level for human con sumption of a substance. Since a safe level cannot be determined, there is no need at present to resolve issues about the consumption pattern for acryloni trile. Thus, it was appropriate for the Initial Decision not to reach any conclu sion about this matter. Furthermore, before Monsanto's esti mate of the total dietary concentration of acrylonitrile can be accepted, the is sues raised by the Bureau need to be con sidered. (Bureau Reply at 24-26.) The Bureau criticizes the Monsanto estimate because It relates to total diet, rather than total fluid diet, because it is based on migration estimates that are not suf ficiently exaggerated with respect to time and temperature conditions, and because it does not make any allowance for the higher absolute amount of acrylonitrile consumed by the above average consum er, even though the concentration of ac rylonitrile remains the same. If a safe level is to be determined for acrylonitrile FEDERAL REGISTER, VOL. 42, NO. 185--FRIDAY, SEPTEMBER 23, 1977 .U B m if f 7 If HU 7 18310 RULES AND REGULATIONS in the future, based on complete safety data, the issues raised by the Bureau need to be taken into account in evaluat ing Monsanto's estimate. IV. Miscellaneous Matters The Commissioner discusses below some miscellaneous exceptions raised by the participants and the nature of the order that should be Issued in this pro ceeding. A. rORMAL REQUIREMENTS AND ADEQUACY OF THE INITIAL DECISION The manufactures object to the Initial Decision on the ground that it made broad general findings, did not contain formal findings of fact and conclusions of law, and did not discuss ail the testi mony and points they consider signifi cant. (Monsanto App. at 13-16, 28-29, Borg-Warner App. at 12-19, and Vlstron App. at 19.) They conclude from this that the Administrative Law Judge did not fairly evaluate the whole record. They also deduce from the fact that the Administrative Law Judge did not state the governing legal standards with re spect to the burden of proof and other matters that he did not in fact apply the correct legal standards and did not, for example. Impose on the Bureau the bur den of proof on the food additive ques tion (Monsanto App. at 12. Borg-Wamer App. at 2-8.) Monsanto also criticizes as "misleading" the Administrative Law Judge's statement of Monsanto's position on a number of safety issues and his de scription of the details of the different studies in the record. (E.g., Monsanto App. at 54-69.) The Commissioner concludes after re viewing this matter carefully that these objections go primarily to form and that there is no merit in the suggestion that the Initial Decision Is inadequate or un fair. The Administrative Law Judge's Initial Decision reflects a thorough eval uation of the submissions under the ap propriate statutory standards and a care ful analysis of the points he considered significant. He did not need to recite all the governing legal standards, and his not doing so with respect to the matters raised in the exceptions seems attribu table to the fact that It was clear and un disputed which standards applied. The Bureau clearly acknowledged, for in stance, that it had the burden of proof on the food additive Issue (Bureau Brief at 6-7). In preparing this Final Decision, the Commissioner has addressed all the con tentions raised by the participants that he considers significant. He also has ex pressly stated his positions on some of the minor contentions, simply to avoid further dispute about the need for an ex press statement to demonstrate they were considered. Time did not permit, though, an express treatment of each exception, or Item in the record that the partici pants may have wanted discussed, and a specific discussion of each was not needed in view of the satisfactory responses In the Bureau reply. Furthermore, the Com missioner believes the Issuance of a De cision should not be an exercise in men tioning details. Instead, the time and at tention of the decisionmaker should be addressed to presenting a reasoned dis cussion of the Important issues, as the Commissioner believes has been done in this proceeding. A narrative presentation of findings of fact and conclusions of law. as incorpo rated in the Initial Decision, is permis sible under the provisions of the Admin istrative Procedure Act, 5 U.S.C. 557(d). Gilbertsville Trucking Co. v. United States. 196 F. Supp. 351 (D. Mass. 1961); State Corporation Commission v. United States. 184 F. Supp. 691 (D. Kan. 1959). The other applicable provisions, 21 U.S.C. 409(f * and 21 CFR 12.120, do not estab lish a mandatory form for the required findings of fact and conclusions of law. The commissioner believes, though, that a formal statement of the findings and conclusions of law is the most useful for mat for providing a concise presentation of the basic elements of the decision. Ac cordingly, he has included formal find ings of fact and conclusions of law in this final decision In numbered paragraph form. Monsanto's criticisms of the Initial De cision as being inaccurate or misleading are related to specific exceptions it takes to the discussion of the safety data. The Commissioner has discussed a number of these criticisms in connection with the safety data to which they relate. In many repects, the exceptions fault the Initial Decision as being inadequate or misleading simply because the decision did not accept the theories offered by the manufacturer. (E.g.. Monsanto App. at 58-59.) Some objections are a quibbing over details, though Monsanto's excep tions did correctly point out some inac curacies. (See, e.g., Bureau Reply at 26. n. 36 and 31, n 10--11 agreeing with Mon santo's exceptions.) The Commissioner has responded to some other exceptions reporting errors or imprecision by stating the point correctly when the discussion of the point was pertinent in this Deci sion. The inaccuracies in the Initial De cision do not, however, have any mate rial bearing on the basic findings of the Initial Decision or its conclusions, and, after considering them the Commission er finds that they are, as a whole, in consequential. B. COMPENSATION ARRANGEMENT FOR WITNESS AND BIAS The Administrative Law Judge noted in his opinion that one of the expert witnesses for a manufacturing party was an employee of one of the law Arms rep resenting that part (ID at 5-6). The Ad ministrative Law Judge stated that he was raising this matter in his Initial Decision "merely to point out the poten tial inroads to a witness's freedom from 'financial Inducements' which could con ceivably result from such circum stances." (ID at 6.) The manufactur ing party involved raised a strenuous and lengthy exception to the Administrative Law Judge's reference to this matter and asked the Commissioner to have the Ad ministrative Law Judge's discussion "ex punged from the record of this proceed ing," (Monsanto App. at 30-37). The comment of the Administrative Law Judge quoted above was viewed as a "gratuitous slur" of the witness, and his raising of the issue, without resolving it, was viewed as indicating bias on his part against all the manufacturers' witnesses. (Monsanto App. at 33. 35.) The Commissioner views the Adminis trative Law Judge's purpose as having been either to note a factor that might conceivably have some bearing on the weight to be given the testimony of the witness involved or to raise an issue that might be considered for treatment in the Agency'3 procedural regulations. His raising of the matter does not demon strate bias. As the exception itself main tained. the compensation arrangement for a witness "is simply a factor to be considered in weighing the testimony". (Monsanto App. at 35.) After reviewing the matter, the Com missioner finds no reason to alter the weight that should be given to the testi mony of the witness because of the com pensation arrangement noted in the Ini tial Decision. The Commissioner has carefully considered the testimony of this witness on its merits. The Commissioner regards the manu facturer's request for an expungement of the record as being novel. He Is nei ther convinced of its merits nor cer tain of the mechanics. Under FDA's general procedural regulations, 21 CFR 10.70(c) (6), once a document Is final and part of the Agency's administrative file, it cannot be deleted. The record of this proceedng Is public, and what is in the public record should remain there. The Commissioner also rejects the suggestion that the Initial Decision re flects bias against all the manufacturers' witnesses. The principal grounds given in the exception for asserting bias, aside from the matter discussed above, are the failure to rely on the Maltoni studies and the failure to mention the. testimony of "eminent scientists." specifically Dr. Robert Olson (M-96), and Professor Robert Wilson (M-90). (Monsanto App. at 35-37.) The Maltoni studies are dis cussed earlier in this Decision and the testimony of Dr. Olson and Dr. Wilson is discussed below. As the Bureau points out, bias is not demonstrated simply "because the Court fails to invoke the name or discuss ex plicitly the testimony of every witness in the proceeding." (Bureau Reply at 20.) The Administrative Law Judge did not specifically discuss the testimony of some "key" Bureau witnesses, namely Holtz, Cueto and Livingston (Bureau Re ply at 20), yet there is no suggestion that he was biased against the Bureau. Fur thermore, the Commissioner believes that omission of any specific discussion of the Olson and Wilson testimony is justified by its content. The testimony of Dr. Olson discussed in general terms the metabolism of acrylonitrile and the threshold concept and, without referring to any actual toxicity data, drew a con clusion about the safety of acrylonitrile which was based on the manufacturers' theory of toxicological insignificance. A FEDERAL REGISTER, VOL. 42, NO. 1S5--FRIDAY, SEPTEMBER 23, 1977 qua A RULES AND REGULATIONS 48541 Since the Commissioner has rejected Whatever the particular factor, was 177.1020, 177.1030, 177.1040, 177.1050, j that theory in this proceeding, it is un that caused the initiation of some recon 177.1480). necessary to deal in detail with Dr. Ol- sideration of the safety data for acrylo 2. Acrylonitrile monomer was not used > son's testimony. Professor Wilson's testi- nitrile copolymers, the Commissioner be- prior to January 1,1958 to fabricate bev [ mony dealt with a risk assessment of lives the reconsideration to have been erage containers (0-77 at 1 17). acrylonitrile which was not significantly warranted, as this record amply demon 3. None of the containers manufac different from others presented by the strates. The Commissioner regards as tured in accordance with any of the reg manufacturers and was based on incom pointless any further effort to settle dis ulations that currently authorize their plete data. Since the Commissioner has putes about the prior history of acrylo use is free of residual acrylonitrile mon concluded that the presently available nitrile regulation. omer (M-i; M-84, at 1 11 and Table A; data are not adequate to confidently evaluate the risks posed by acrylonitrile, it is not necessary to evaluate In detail E. NATURE or. THE ORDER The Initial Decision provides for a stay BW-1; BW-2; V-4 at 1 18; V-12>. 4. The physical process known as dif fusion causes residual acrylonitrile mon risk assessments such a those offered by of the regulations at Issue "until further omer in the containers to migrate into Professor Wilson. C, APPLICABILITY TO OTHER USES AND SUBSTANCES The manufacturing parties except to the Administrative Law Judge's sugges tion that the Commissioner consider re vising or staying the regulations govern ing the use of acrylonitrile copolymers in notice." (ID at 40.) In its brief to the Administrative Law Judge, the Bureau proposed an order providing for an amendment of the regulations to elimi nate the use of acrylonitrile in beverage containers. The Commissioner has adopted essentially the Bureau's pro posed order, by making the appropriate changes in the codified text of the regu the beverage under aU conditions of time and temperature to which the containers would be exposed in normal usage <M-82 at 17 6-10; M-84 at 71 13-18, 21.B. Ap pendix; M-86 at 71 2(a), (b); G-77 at 71 30-32; G-78 at 11 11-12, 18 and Ex hibit A; Tr. at 132, 136. 139, 143, 144). 5. Extraction teste with food simulants and with beverages conducted under non-beverage containers. (Monsanto App. at 31-38, Borg-Wamer App. at 20-23; Vistron App. at 1-3.) They urge lations, with a deletion of the stay note and termination of the stay published in the Federal Register of March 11, conditions reasonably close to those as sociated with normal use and intended use confirm that the diffusion process that the Commissioner not issue an order in this proceeding that applies to non beverage containers. The Commissioner believes the Ad ministrative Law Judge's suggestion war rants further consideration, but he views any action to be taken in that regard to be a distinct matter that should be handled in a separate proceeding. The issues as framed for this proceeding con cern only beverage containers, and af fected persons must be Afforded an op portunity to be heard before any action is taken with respect to other uses. Monsanto suggests that this proceed ing should be viewed as having deter mined that hydrogen cyanide and other substances which may be In acrylonitrile copolymer beverage bottles are nonmigratory or safe merely because they 1977 (42 FR 13546). Under 5 409(f) (3) Of the Act, 21 U.S.C. 348(f) (3), the Com missioner's final order is not effective until 90 days after publication. In Monsanto Corp. v. Gardner, the Court held that in effect the Commis sioner's stay action was taken on his own initiative and that the situation was governed by sections 409(d) and (h) of the Act, 21 UB.C. 348(d) and (h). These statutory provisions authorize the Com missioner to issue regulations on his own initiative and to establish procedures to amend or withdraw food additive regu lations. Thus, the Court viewed this pro ceeding as being one taken on the Com missioner's Initiative to withdraw or amend food additive regulations. The Commissioner agrees. The Commissioner believes that it causes migration of acrylonitrile mono mer from the beverage containers of Monsanto, Borg-Wamer, and Vistron (G-77 at 1 30; 0-78 at 11 8, 10-11, 1416; G-79; G-80; <3-81; M-84, Appendix at A--2, A--3, Table IV; BW-1; BW-2; V-l, Table 2; V-7; V-10). G. The diffusion process may be ex pressed in the form of a mathematical equation that can be used to predict the occurrence of migration of acrylonitrile monomer at levels that cannot be meas ured by analytical techniques. Models based on this equation predict that mi gration will occur, under all time and temperature conditions to which bever age containers would normally be ex posed, from containers with low levels of residual monomer that produce' amounts of migration too small to be detected I j were not discussed in the Initial Deci sion. (Monsanto App. at 12, footnote.) The Commissioner disagrees. The record shows that, due to the migration of acrylonitrile monomer, acrylonitrile co polymers used to fabricate beverage con tainers are food additives which have not been shown to be safe. Therefore, it is not necessary to consider whether other substances in the beverage con tainers migrate or are safe, and this pro ceeding should not be considered to have decided those questions. 0. PRIOR HISTORY Monsanto excepts to the Administra tive Law Judge's summary of the origins of this proceeding. (Monsanto App. at 39-41.) The Administrative Law Judge attributed the Agency's reconsideration of the safety data for acrylonitrile to the development of improved methods for detecting migration. Monsanto Insists that the reconsideration of the safety data was prompted by a concern about the possibility of delayed extraction from a reversible mercaptan complex, a possibility which has not been pursued by the Bureau in this proceeding. (Tr. at 114.) would be inappropriate to adopt the order contained In the Initial Decision which simply stays the regulations in definitely until further notice. It Is not clear what procedures would be involved in terminating such a stay, nor what cir cumstances would warrant termination. Treating this proceeding as an action to amend the regulations, rather than merely to stay them, will ensure that future developments are governed by the regular procedures for regulation of food additives. Any future changes in the status of acrylonitrile as a food additive should be made through the regular pro cedures of public notice of a proposed regulation, opportunity for comment, and a hearing on objections. By regard ing this proceeding as one to amend the regulations rather than merely stay them, the Commissioner reduces the likelihood that there will be confusion about the legal status of acrylonitrile copolymers used to fabricate beverage containers. V. Findincs of Fact and Conclusions of Law A. FINDINCS or rACT analytically (G-78 at 11 5-7, 8, 11-12, Exhibits A, C; M-84 at 71 12-18,` Appen dix; M-86 at 11 2(a), (b); Tr. at 232. 136, 139, 143. 144). 7. No demonstration has been made that the diffusion process fails to operate for acrylonitrile copolymer beverage con tainers. The theory that. In unusual cir cumstances, the molecules of acryloni trile monomer are Indissolubly bound to the copolymer matrix is speculative and has not been established with respect to the levels of residual acrylonitrile mono mer present In the copolymer materials most recently produced by Monsanto, Borg-Wamer, and Vistron (M-84 at 11 18, 21.B; Tr. at 147-148). 8. The use of acrylonitrile monomer to fabricate acrylonitrile copolymer bever age containers may reasonably be ex pected to, and in fact does, result In acrylonitrile copolymer beverage con tainers becoming a component of bever age foods packaged In such containers (G-77 at 133; G-78 at 118; Findings Nos. 1-7, supra). 9. Laboratory testing of acrylonitrile copolymer beverage containers under ex aggerated time and temperature condi Monsanto also excepts to the description 1. Acrylonitrile monomer Is used to tions with food simulants is an appropri I of the requirements of the interim regu fabricate acrylonitrile copolymer bever ate means of confirming whether the in lations. (Monsanto App. at 40-41.) age containers (G-77 at 1 17; 21 C.F.R. tended use of the containers results in FEDERAL REGISTER, VOL. 42, NO. 1SS--FRIOAT, SEPTEMBER 23, 1977 m t u z m aha 18312 RULES AND REGULATIONS migration of acrylonitrile monomer and 16. The lowest reliable limit of detec acrylonitrile by ingestion, does not con- of determining what the levels of migra tion yet achieved for acrylonitrile mono tain data Indicating what the lifetime tion are or may be assumed to be for mer is 10 ppb (M-83 at 18; BW-1; V-10). risk Is from exposure to acrylonitrile, and purposes of evaluating the toxicity of 17. Taken together, the available data does not report the level of exposure to acrylonitrile monomer under the Intend Indicate that migration of acrylonitrile acrylonitrile monomer undergone by the ed conditions of use of acrylonitrile co polymer beverage containers (0-77 at monomer from acrylonitrile copolymer beverage containers now being produced subjects of the study (0-74). 28. The Tullar Study (0-54) and the 5-16; 0-78 at 1 17; Tr. at 63, 70-71. is in the range of several parts per bil Svirbely and Floyd Study (G-55), early 79-80,124). lion to ten or more parts per billion un chronic toxicity studies of acrylonitrile 10. The purpose of extraction testing of acrylonitrile copolymer beverage con tainers Is to determine the maximum amount of acrylonitrile monomer to which the consumer may be exposed un der the intended conditions of use of the containers. In conducting extraction tests, it is scientifically appropriate to me conditions of time and temperature that exaggerate those that would be ex pected for a particular product In order to account for extreme conditions of use and to compensate for the fact that food simulants do not necessarily possess the same extraction characteristics as the numerous different foods that might be packaged In the containers (G-77 at If 11-12, 14-16, 23, 25-26. Exhibit F, at 12-13; Tr. at 61, 120, 125, 145). 11. The conditions of 6 months and 90' F are appropriately assumed to repre sent the Intended conditions of use of acrylonitrile copolymer beverage con tainers In accordance with accepted principles of extraction testing (0-77 at 11 24-29; M-85 at 11 22-25). 12. For purposes of assessing the tox icity of acrylonitrile monomer in bever ages, data must be derived that show ap proximately how much monomer will migrate under the intended conditions of use of the containers. These data may consist of the results of extraction tests, projections from mathematical models, or the level of analytical detection that has been achieved. (0-77 at If 5-7, 2728; 0-78 at 115-6, 9, 12; M-83; M-84; BW-1; BW-2; V-4 at 113; V-6; V-7; V-8: V-10: Tr. at 69, 70-72, 79-80. 104106.129.137.146-147.) 13. Extraction data exist for the acry lonitrile copolymer beverage containers of Monsanto, Borg-Wamer, and Vlstron (Finding No. 5. supra). 14. The available extraction data do not permit a determination of how much monomer migrates from the most recent ly produced beverage containers of Mon santo and Vistron, although in the case oTVistron its presence was detected and in the case of Monsanto- Its presence was probably detected (M-83 at If 24. 9; V-4. f 16; V-10). 15. The mathematical models devel oped by Monsanto predict migration of 4.3 ppb or more of the monomer under the Intended conditions of use of acry lonitrile copolymer beverage containers. The models are as applicable to BorgWarner's and Vlstron's containers as to Monsanto's. The models are not intend ed to be precise, but to provide on ap proximation of the levels of migration der the intended conditions of the con tainers' use. Greater precision would be required If a safe level for human con sumption of arcylonitrile were capable of being established, at this time (Findings Nos. 9-16, supra). 18. Acrylonitrile Is a frank teratogen in the rat (0-58; G-82; 0-83; Tr. at 187). 19. Because animal species differ in their susceptibility to teratogenesls, a teratology study in at least one more animal species must be performed before a "no-effect" level for acrylonitrile mon omer can be determined for purposes of estimating a safe level of exposure for human (0-82 at 16; 0-86 at 1 li.b.; 0-87 at flf 15. 23; Tr. at 195. 197, 199, 201-203. 208-209, 210-211). 20. Because of the teratologic effects of acrylonitrile monomer in the rat, a re production study must also be conducted before the toxic potential of the mono mer can be regarded as completely char acterized (Tr. at 195). 21. Based on the Dow-MCA study In rats, acrylonitrile monomer may be tera togenic in humans (0-82 at 1 6; Tr. at 214). 22. Acrylonitrile monomer Is muta genic in several bacterial test systems (0-67; 0-70; 0-84; BW-7; NRDC-92). 23. A chemical that is mutagenic Is probably carcinogenic in mammals (G84 at n 3, 20; Tr. at 219-221, 281). 24. The Dow chronic feeding study (G-57; 0-72) is appropriately designed to determine tire potential carlnogenlclty and other toxic effects of acrylonitrile monomer in rats (0-85 at 15; Tr. at 269-270). 25. The interim one year results of the Dow chronic feeding study demonstrate that acrylonitrile monomer Is tumorigenic, and create a strong inference that it is carcinogenic (0-85 at 118-9; 0-86 at ff li e. 14.b; 0-87 at 11 25-28: Tr. at 303; NRDC-94). 26. The Dow chronic feeding study is incomplete, and the preliminary results indicate at best only the level of risk from acrylonitrile monomer after about the first half of a rat's lifetime (0-72; Tr. at 560. 611-612). At least one tumor has been observed in rats at the lowest feeding level In the ongoing phase of the study that are similar to the tumors ob served in rats at higher feeding levels in the interim phase of the study (0-71; 0-86 at 113.b; G-87 at f 125-27; Tr. at 290). Consequently, the Dow chronic feeding study does not permit the deter mination of a "no-effect" level (0-86. f 15: G-87. f 38). monomer, are scientifically inadequate for purposes of drawing reliable conclu sions about the toxicologic potential of acrylonitrile (0-41; 0-84 at 116-9; G87 at ff 18-19). 29. The available toxicology data dem onstrate that acrylonitrile is a teratogen, a mutagen, a tumorigen, and probably a carcinogen (Finding Nos. 18-28, supra.). 30. Whether there is a true biological threshold for the toxic effects of acrylo nitrile monomer or where that threshold lies, if it exists, are not now known (G- 88 at 11 9. 14-15, 16-21; Tr. at 650-651, 660, 692). 31. A safe level of human exposure to a toxic chemical is .established by apply ing appropriate safety factors to a "no effect" level determined from the avail able toxicology data (21 U.S.C. 348(c) (5) (C);G-87 at If 12-15). 32. It Is not possible on the basis of the available toxicology data to establish a safe level for human consumption of acrylonitrile monomer by applying safe ty factors: the data are incomplete, a "no effect" level for the chronic toxicity dem onstrated In tile 2-year Dow chronic feeding study has not been reliably de termined, and It Is not possible to estab lish a safe level for human consumption of a substance that Is probably carcino genic (G-82 at 1 6; 0-86 at Ifl ll.b, 15; G-87 at II 13-16, 23, 24-36. 41; Tr. at 195, 197, 199, 201-203, 208-209, 210-211 284. 290). 33. For the same reasons. It Is impos sible to establish the degree of toxicity of acrylonitrile monomer for use in per forming a "risk assessment" of any level of human exposure to acrylonitrile monomer in beverages. 34. For the same reasons, there do not now exist "scientific procedures" that adequately show that acrylonitrile is safe under the conditions of its intended use to fabricate beverage containers and that could be relied on by qualified ex perts for that conclusion.. 35. Qualified experts in fact do not generally recognize that acrylonitrile has been adequately shown through sci entific procedures to be safe under the conditions of its intended use in beverage containers (G-82 at 1 6; 0-86 at 1 16; G~ 87 at f 41: M-82 at IF 5; M-90 at III 8, 52; M-9G at T 6; M-97 at V 11; M-98 at 11; Tr. at 466-467, 589-590; NRDC at U 1922). 36. Factors that are considered In eval uating the safety of a toxic substance under the conditions of its intended use as a food packaging material include the amount of possible migration into food, that can reasonably be expected to oc cur (G-78 at f 9, Exhibits A and C; M84, Appendix and Table IV; M-88, Table IV; Tr. at 137,139,146). 27. Epidemiological data from the duPont workers study suggest that acrylo nitrile monomer is carcinogenic in hu mans (0-86 at f 13.g; 0-87 at 1 30), but the study does not involve exposure to the types of food packaged in the mate rial, the extent of consumption of foods packaged in the material, and the degree of toxicity of the substance (0-87 at 15). - FEDERAL REGISTER, VOL. 42, NO. T85--FRIDAY, SEPTEMBER 23, 1977 m m a00 RULES AND REGULATIONS 48543 37. Since acrylonitrile cannot be con procedures, which must be of the same containers is not generally recognized, sidered sate at present because of the quality and quantity as those necessary among experts qualified by scientific deficiencies in the toxicology data (Find to demonstrate safety for purposes of training and experience to evaluate its ing Nos. 19-36, supra), it is unnecessary issuing a food additive regulation; and safety, as having been adequately shown to determine the extent of dietary ex such evidence must be reflected In pub through scientific procedures to be safe. posure to acrylonitrile monomer in this lished studies in the relevant body of 14. The proviso in 21 U.S.C. 321 (s) for proceeding. scientific literature. Evidence that quali substances used in food prior to January B. CONCLUSIONS OF LAW 1. The intended use of acrylonitrile monomer in fabricating acrylonitrile co polymer beverage containers may rea sonably be expected to, and docs, result fied experts would or should recognise a substance as having been adequately shown to be safe if they were aware of the data and information about a sub stance does not establish the existence of general recognition of safety among 1,1958, applies only when a substance to which the proviso is sought to be applied was used prior to January 1, 1958, In the same way that It is, or is sought to be. used now. Hie proviso Is Inapplicable to acrylonitrile used to fabricate beverage in its becoming a component of food. 2. The actual or reasonably expected presence of acrylonitrile monomer in food may be established by reference to any probative scientific evidence, and need not be shown by actual measure such experts. 8. "Safe" or "safety," as those terms are used in 21 U.S. 321 (s) and 348, mean a reasonable certainty in the minds of competent scientists that a substance is not harmful under the intended condi containers. 15. Acrylonitrile copolymers used to fabricate beverage containers are food additives within the meaning of 21 U.S.C. 321(s). 16. The data of record in this proceed ment. tions of its use. A substance may be safe ing fail to establish that the use of acry 3. Whether acrylonitrile monomer be even though the toxicological data do lonitrile to fabricate acrylonitrile copoly, comes, or may reasonably be expected to not definitively demonstrate that it is mer beverage containers is safe, as re become, a component of food within the safe. However, if the toxicology data quired by 21 U.S.C. 348(c)(3)(A). meaning of 21 U.S.C. 321 (s) does not de make It reasonably certain that the sub pend on whether it does or may migrate stance may be seriously harmful to some In any particular amount. "Toxicological people under its Intended conditions of Final Ordeb Therefore, on the tests of the forego insignificance" is not a statutory term; use, the substance is not safe within the ing Findings of Fact and Conclusions of it does not have a precise meaning or an meaning of the Federal Food, Drug and Law and the record in the above pro Independent significant or relevance for Cosmetic Act even though the people ceeding and under the Federal Food. food additive determinations. who will not be harmed by it greatly out Drug and Cosmetic Act (secs. 201 (a), 402, 4. The "conditions of . . . intended number those who may. 409, 701, 53 Stat. 1049-1047 as amended, use" of acrylonitrile monomer in bever 9. In evaluating whether a food addi 1055-1056 as amended, 72 Stat. 1784- age containers are relevant to evaluating tive has been shown to be safe, FDA must 1788 as amended (21 U.S.C. 321(s), 342, the toxicity or safety of the monomer; consider the probable consumption of the 348, 371)) and under the authority dele migration under the "conditions of . . . additive. 21 U.S.C. 348(c)(5)(A). Prob gated to the Commissioner (21CFR 5.1); intended use" need not be shown, but able consumption includes consumption It ia ordered. That the stay of Si 177. may be used, to establish whether the by Individuals likely to be exposed to the 1020, 177.1030, 177.1040, 177.1050 and "intended use" of the monomer "results additive in amounts that significantly 177.1480, insofar as those sections per or may reasonably be expected to re exceed the average level of exposure. mit the acrylonitrile copolymers Identi sult ... in its becoming a component" of 10. The Delaney Clause, 21 U.S.C. 348- fied in those sections to be used to fabri food within the meaning of 21 U.S.C. 321(s). 5. The "conditions of . . . intended use" of acrylonitrile monomer in bever age containers are properly assumed to be those conditions determined to be appropriate for extraction testing. (c)(3) (A>, creates a conclusive pre sumption that a carcinogenic substance present In food as a food additive Is un safe. Accordingly, a carcinogenic sub stance per se cannot be considered gen erally recognized as safe under 21 U.S.C. 321 (s). A substance that is the subject of cate beverage containers, as ordered in the Federal Register of March 11, 1977 (42 FR 13546), be terminated, and that Part 177 be amended as follows: 1. In i 177.1020, by deleting the stay note at the end of the section, and by adding a new paragraph (f) to read as 6. For purposes of toxicological assess testing whose preliminary results follows: ment, the amount of acrylonitrile mono mer that migrates under extraction test ing conditions may be determined by analytical measurement, by the use of mathematical models, or by the lower limit of analytical detection. If mathe matical models are used, their limitations strongly indicate that It will be shown to be carcinogenic cannot be considered generally recognized as safe under 21 U.S.C. 321(a). 11. Since the Dow chronic feeding study is not yet complete, acrylonitrile monomer cannot at this time be "found 177.1020 Acrylonitrile/butadiene/"ty- rene copolymer. (f) Acrylonitrile copolymers identified In this section are not authorized to be used to fabricate beverage containers. must be clearly kept in mind, and the Bu to Induce cancer when ingested by man 2. In S 177.1030, by deleting the stay reau is not bound to accept the predic or animal" within the meaning of the note at the end of the section, and by tions of such models as determinative of Delaney Clause. 21 U.S.C. 348(c)(3)(A). adding a new paragraph (f) to read as the amount of monomer that Is present No other carcinogenicity data here In follows: for purposes of making toxicological judgments. If the lower limit of analyti cal detection is used, In the absence of a contrary showing, monomer should be assumed to be present in an amount evidence have been shown at this time to be "appropriate for the evaluation of food additives" within the meaning of the Delaney Clause. 12. In evaluating the safety of a food g 177,1030 Acrylonilrile/lmlatlicnc/M)'- reiic/methyl mctliacryintc copolymer. (f) Acrylonitrile copolymers identified equal to the lower limit of detectability. additive, FDA must consider safety fac In this section are not authorized to be 7. For a substance to be generally rec tors which In the opinion of experts used to fabricate beverage containers. ognized among qualified experts as hav ing been adequately shown to be safe under the conditions of its intended use. the relevant community of experts must have actual, present knowledge of the qualified by scientific training and ex perience to evaluate the safety of food additives are generally recognized as appropriate for the use of animal experi mentation data. 21 U.S.C. 348(c) (5) (C). 3. In $ 177.1040, by deleting the stay note at the end of the section, and by adding a new paragraph (e) to read as follows: toxic potential, or the absence ol toxic potential, of the substance at the levels present under those conditions of use; they must actually have concluded from such knowledge that the substance has been adequately shown to be safe; their knowledge must be based on scientific Safety factors may not be applied to a carcinogenic or potentially carcinogenic substance to determine a safe level of human exposure to the substance as a food additive. 13.Acrylonitrile monomer used to fab ricate acrylonitrile copolymer beverage 177.1040 Acrylonitrilc/felyrcuc copol ymer. <e> Acrylonitrile copolymer's identi fied In this section are not authorized to be used to fabricate beverage containers. FEDERAL REGISTER, VOt. 42, NO. 185--FRIOAY, SEPTEMBER 43, 1977 4S5-M RULES AND REGULATIONS sT: 4. In 5 177.1060, by deleting the stay note at the end of the section, and by adding a new paragraph (g) to read as follows: 177,1050 Acrylonitrile/styrene copol ymer modified with biitadiene/styrene elastomer. ii( fg> Acrylonitrile copolymers Identified in this section are not authorized to be used to fabricate beverage containers. 5. In 1 177.1480, by deleting the stay note at the end of the section, and by adding a new paragraph (d) to read as follows: 177.1480 Nitrile rubber modified ac rylonitrile-methyl acrylate copoly mers. (d) Acrylonitrile copolymers Identified in this section are not authorized to be used to fabricate beverage containers. Effective date: This final decision Is effective upon signature on September 19, 1977. The order terminating the stay and amending the regulations shall be effective December 22,1977. Dated: September 19,1977. Donald Kennedy, Commissioner of Food and Drugs-. [PR Doc.77-27765 Filed 9-20-77;10:29 am) FEDERAL REGISTER, VOL. 42, NO. 1S5--FRIDAY, SEPTEMBER 23, 1977