Document oeYQX14GBxnXBJK3RVX2XEmYD

J'JSLPH .. KEI.LTH JZXOHZ H KKCEMaN Ca\2LES M M2EBa>' *.*TLLJaM K. BOBOHESA.v:, -JR ROBERT B. TIEBNAN' .waYNE V. BLACK ' i>AV?D t. HILL MABTrN* V. BERCOVTCI PETER M VEMFOV CAROLE c. 3ABRIS -PETES THOMAS SMITH HtCn.KZL T. MORHONE LaJKBT S. SOLOMON B DCBECK CHRISTINE A MEaGEEB LAW OT71CT.S IlELLER AXI) HECKMAN Washington*, d. c. aooau April 8, 1977 TEjLPHONT 2G2 -;3T- LOO CABLE ADDB.E5S 'KTl WRITESS DIRECT DIAL N 202-4-S7-1110 RECEIVED APR 2 q1977 Dr. Howard R. Roberts Acting Director 'Bureau of Poods/ Food and Drug administration 200 C Street, .W. Washington, D.c. 20204 Re: pocket No. 75N-013.0; vinyl chloride Polymers' `An Contact With Fo.od; Notice, of Proposed Rule Kaking, 40 Fed. Reg. 40529, September 3, 1975. .Dear Dr. Rohkrts Th^[purpose of this letter is to follow up on the January 5 ,J L9T7 confer ence between your staff and represents tives of tne-Vinyl Chloride/Polyvinyl Chloride Producers Group of Tine Society of the Plastics industry, Inc. (SPI) . Responsive to the three.requests made by members of your -staff during that conference, we are herewith submitting the following: / (a) "raw data" to confirm and sub stantiate the submitted reports of the kmpressive achievements in reducing residual Ivinyl chloride monomer (RVCM) to insigni ficantly low levels in commercially available polyvinyl chloride products intended to contact food, (b) descriptions of the analytical procedures employed to determine not only RVCM levels but also the levels of vinyl chloride (VC.M) in food simulating solvents, and Dr. Howard R. Robert|s April 8, 1977 Page Two (c) a memorandum that provides a full rationale f r the conclusion that no vinyl chloride ca reasonably be expected to migrate from igia and semirigid polyvinyl chloride food-contact products under in tended conditions of use, provided said products are made in accordance with good manufacturing practices that limit RVCM to a level not exceeding 0.1 parts per million. 1/ it is respectfully submitted that the record, in its entirety, compels the determination that the public interest warrants the use of rigid and semirigid PVC food packaging material; produced in accordance with appropriate good manufacturing practices. Such PVC food packaging materials have been found environmentally preferred over many competing materials by the Bureau of Alcohol, Tobacco and Firearms (BATE)in its assessment of PVC bottles for liquor, 2/ and that determination is confirmed by FDA's findings in its Ervironmental Impact Statement relating to plastic contairers for Beverages. 3/ In particular, the PVC containers provide important benefits with respect to freedom from brea).ability and the absence of problems with regard to safety. We submit tha t it is fatuous to claim a ris* to health exists whe i no vinyl chloride can be measured in food simulating sslve nts after exaggerated exposures using analytical procedures sensitive to 1 or 2 parts per billion This is especially tr ue when one considers that under no conceivable circumsta nces will as much as 10% of the diet be packaged in rigid and semirigid PVC products. The present level of anclyti cal sensitivity is so low that the risk is less thar app roximately one ten-thousandth the risk deemed by FDA to b safe" in an analogous context. 4/ This 1/ Due to manuf icturing tolerances, a maximum limit of U.l ppm requires the average product to be well below that level. 2/ BATF, "FinalJ Environmental Impact Statement, Polyvinyl Chloride Liquor |3ottles," March 9, 1973. 3/ FDA, "Final Environmental Impact Statement, Plastic Bottles for Car! onated Beverages and Beer," September, 1975 4/ 42 Fed. Reg 10412 SPI-01689 Dr. Howard R. Rob? April 8, 1977 Page Three level is at least three orders of magnitude lower than level the Cccupaticnal Safety and Health Acl-inistratic: (OSHA) determined to be safe when it wrote its standard for occupational e> posure to.vinyl chloride. 5/ The data s ibmitted to the Food and Drug Administra tion (FDA) during te pendency of the subject rule making has demonstrated that the polyvinyl chloride industry has reduced the residujfl vinyl chloride levels in rigid and semirigid products from levels in .the range of approximately 500 parts per millL|ion, which were common before the possibility of migratio of vinyl chloride was recognized, to levels well below (0.5 parts per million for many products, a thousand-fold reduction. Furthermore, some products can now be produced wilth levels of RVCM no greater than 0.1 part per million; and in some cases, particularly for sheet materials used in so-called blister packs, the levels do not exceed 0.05 pa rts per million. In other words, polyvinyl chloride ric id and semirigid products are now available to the food packaging industry with vanishing small levels of residua vinyl chloride and assure that there will be "no migra ion" 6/ of vinyl chloride to food packaged in such products when tEe foods are packed and otherwise handled in accord * nee with good manufacturing practices for the productioi and marketing of food products. 7/ 5/ 29 C.F.R. 1910.1017 6/ The expressio n "no migration" is used as a convenient shorthand to mear "no reasonable expectation of becoming a component of fc od under the intended conditions of use." 7/ It is not cl; imed nor is it necessary that all rigid and semirigid polyvinyl chloride products be capable of showing no migra .ion of vinyl chloride to food packaged therein; it is only necessary that those products which are offered as, md represented to be suitable for use as, food packaging miterials comply with such requirements. Thus, the fact t lat all manufacturers may not now be offering materials which will provide a "no migration" package should not lead the FDA to ban all PVC products any more than the FDA would ban a 111 paper from food packaging uses because some grades of paper are unsuitable for that purpose. See Natick PaDerboard Corp. v. Weinberger, 525 F.2d 1103 (1st Cir". T9'7 5'). Acc -o---r-c---i-n--g---l-y, we urge, as we have in the past, that the FDA set appropriate standards in its good manufacturing pract ces regulations to assure, in light of the intended uses o such material, that there will be no migration of vinvl ch loride to packaged food. SPI-01690 Dr. Howard R. Roberts April 8, 1977 Page Four The raw dat and descriptions of analytical procedures are present d in Appendices I through IV. Appendix I, supplied by the American Hoechst Corporation, Film Divison, summarizes the resi ual vinyl chloride content of PVC resins received by American Hoechst during 1974 through 1976, provides a typical detn iled report of lot-by-lot results for RVCM for the month 4>f October 1976, provides typical chromatograms for a sampling of the individual lots reported for October 1976, demon strates the reduction in RVCM achieved during the blending process wherein resin as purchased is converted to a prod jet ready for formation into sheet product, provides typical ch omatograms for material before and after blending, and, fina Lly, sets forth the detailed analytical procedures that wer employed. Appendix II was submitted by the B. F. Goodrich Company. It includ(e s an update of the RVCM levels in both resin and finished bottle compounds supplied by B.F. Goodrich during the las quarter of 1976. The gas chromatcgraphic procedure e(mployed by B. F. Goodrich for the determination of the residua vinyl chloride monomer content of polyvinyl chloride res s and wet cake samples is provided, as well as a repor by J. A. Nikora and E. G. DeCapita entitled "Confirm!rg the Presence of Vinyl Chloride in Food Simulating Solvents Appendix IjI supplied by Ethyl Corporation consists of information alrsady supplied separately to FDA. It includes a documenl dated January 10, 1977, addressed to the Hearing Clerk containing Addenda A through E as direct responses to the January 5, 1977 request for raw data and analytical methodo ogy. In addition, there is a separate submission also addressed to the Hearing Clerk by means of a letter dated January 13, 1977, entitled "VCM Migration Studies with PVC Bottles Containing Low VCM Concentrations" dated January 10, L977 written by Cannaday, Daniels, and Gaeke. This paper is particularly significant because it demonstrates that :he chromatographic determination of vinyl chloride in food simulating solvents used to test low RVCM containers is subji set to significant interferences and the consequent appearanee of false positives. This same ob servation was reported by Tenneco in its independent sub missions to FDA, and we understand similar effects have been observed by FDA scientists as well. SPI-01691 Dr. Howard R. Robert April 8, 1977 Page Five In order to provide ir.oce definitive evidence of interferences and tcT provide direct experimental evidence for setting a residual vinyl chloride level in bottles or bottle compounds that will assure "no migration," a protocol for a study was developed in consultation with FDA scientists and samples of the experimental bottles have already been provided to the FDA staff. The test is now under way; and when the exposure tjime has been completed, samples of the extract will be supplied to the FDA scientists for their own independent analysis. The final results are expected soon, and will be Deported by Ethyl Corporation as soon as they are available. Appendix iv has been supplied by the Ruco Division of Hooker. This includes analytical procedures for the determination of rIvcm in resins by either a solution pro cedure or by head /space analysis, typical quality control chromatagrams including standardizations, blank determina tions and typical (analyses of bottle compounds using both procedures, and tile analysis of a compound deliberately "spiked" to show a large quantity of vinyl chloride. The finaljattachment, Appendix V, is a memorandum that discusses in| considerable detail why there is no reason able expectation jpf migration of vinyl chloride into food under intended conditions of use from rigid and semirigid PVC food-contact (materials that contain very low levels of residual vinyl chloride. Taken together, these appendices confirm current industry capability of producing PVC resins with residual monomer levels well below 0.5 parts per million, that pro ducing compounds from these resins results in a loss of from 50% to mora than 90% of the RVCM originally contained in the resin and that processing into finished products still further lowers the residual vinyl chloride level. In other words,(rigid and semirigid PVC containers can be produced for fopd-contact purposes with residual monomer SPI-01692 Dr. Howard R. Robert April 8, 1977 Page Six levels not exceeding 0.1 parts per million; and when such materials are used in food packaging, there is no reasonabl expectation that vinyl chloride will become a component of food under the intended conditions of use. Accordingly, we urge the FDA to at3opt final Regulations consistent with these facts. Cordially yours, SPI-01693 Appendix V M EMORANDUM THERE IS NO REASONABLE EXPECTATION THAT VINYL CHLORIDE WILL BECOME A COMPONENT OF FOOD IF THE RESIDUAL VIN YL CHLORIDE CONTENT OF THE PACKAGING MATERIAL DOES NOT EXCEED ONE-TENTH PART PER MILLION Background In its Notice of September 3, 1975 (Docket No. 75N-0190, Vinyl Chloride Polymer:s In Contact With Food) the Food and Drug Administration pr cposed to deal separately with polyvinyl chloride (PVC) food packaging materials in three categories: 1. Uses to be affirmed as prior sanctioned including coatings, <|askets, cap liners, flexible tubing, plasticized film and the like; Potable water pipe; and 3. Rigid and semi-rigid materials. SPI filed extensive Comments which included a comprehensive legal analysis of the meaning of the term "food additive" as defined in Section 20](s) of the Federal Food, Drug and Cosmetic Act, voluminous scien'.ific data relating to residual monomer levels in rigid and semi-rigid PVC food contact materials, and the results of th migration studies conducted with such materials. Subsequent to the filing of these Comments, SPI and member companies thereof continued to provide to the Food and Drug Administration additional scientific data as they were being developed bearing on.the possible migration of vinyl chloride (VCM) from igid and semi-rigid PVC food packaging materials. SPI-01694 2 In anticipation of the promulgation of final regulations in this proceeding the industry requested a meeting to bring the Food and Drug Admitistration completely up-to-date regard ing the major improvements made by the PVC industry in reducing residual vinyl chloride (RVCM) levels in food contact materials to vanishingly low lev Is. This information was presented at a conference held in Washington, D.C. on January 5, 1977. At the conclusion of the conference the Food and Drug Administration represeiiitatives requested (1) that the informa- tion supplied be supplusmented with "raw data" bearing on the results that were repo :ted, (2) that the analytical procedures employed be supplied, md (3) that explanatory information be furnished to provid3 i a rationale for the industry conclusion that, from currently available rigid and semi-rigid materials, there would be no reasonable expectation of migration of vinyl chloride to food. The technical information requested is being supplied in separate documents which are being submitted at this time. The reques ted rationale forms the substance of this document. 1/ As a matter of in tjernal management procedures it is understood the Food and Dr ijg Administration attempts to promulgate final regulations or ithdraw a proposed regulation within two years of the publj cation of a Notice of Proposed Rulemaking. SPI-01695 -3 Scientific--Technical Considerations 1. Clearly, the most significant factor leading to the conclusion of "no nigration" 2/ is that no vinyl chloride could be detected in tie appropriate test solvents using analytical procedures of extreme fensitivity, i.e., capable of quantifying vinyl chloride at 'a level of 1 or 2 parts per billion (ppb) when exaggerated extraction tests were conducted. In other words, rigid and semi- igid PVC packaging materials containing the very low levels of residual VCM now being achieved in commerce were tested under cond Itions designed to reasonably exaggerate the likelihood of migr ation and no such migration could be detected. This demonstrable fact alone should be sufficient to confirm the validity of the conclusion reached by the industry scientists that there will be "no migration" of vinyl chloride into food undler intended conditions of use of rigid and semi-rigid PVC packaging materials of the quality now being supplied. It should te borne in mind that conclusions in the physical sciences must, be based on empirical evidence. Theory and speculation guide the choice and design of experiments but the experimental Results are the only conclusions. 2/ It should be appr sciated that the phrase "no migration" Is used in this documfe nt as a convenient shorthand for the more lengthy statutory criteria embodied in S201(s) of the Act of "no reasonable expectation of becoming a component of food." SPI-0169Q 3 2. The exper inter tal data cited above are supported and reinforced by theor tical considerations. The "driving force" that makes a subitance migrate from the interior of a container wall to the surface and, thereafter, into the container contents i the difference between the effective concentration of the substance in the container wall and its concentration in the fopd. When the level in the container wall is relatively high a large differential in concentration exists; the resistance f the plastic to migration is overcome, and migration is rapid and extensive. As the level in the container wall is reduc:sd, the concentration differential is reduced; and both the rrate of migration and the quantity migrating diminish. Finally, whe the concentration in the walls becomes vanishingly small, as iLs the case in the presently available PVC containers, the driving force cannot readily overcome the resistance; and both the rate of migration and the quantity migrating likewise become vanishingly small to the point where the substance cannot be) detected even at extremely low levels. Indeed, there is no waij of knowing that the resistance can 3/ It is recognized tf at migration from the container wall to the exterior will a: so occur. Data in the file show that of the RVCM actually m: grating from "high" RVCM containers approximately one-thirc migrates into the contents and twothirds migrates to the exterior . SPl-01697 4 really be overcome at alL and that any migration will occur. In short, when the concentration in the walls of a container becomes as small as it is now, there is no realistic expec tation of migration. 3. This explanat ion is still further reinforced by the following considerat ions: If one were to consider a typical rigid container wall (0 020 inches thick) that contained 0.1 ppm of residual vinyl chloride, and make the totally unrealistic assumption that all the residual vinyl chloride available for migration into the contents did in fact migrate to food packaged in such a container, the maximum calculable concentration of vinyl chloride would be approximately 2 parts per billion (ppb), a quantity capabLe of analytically quantifiable measurement The evidence cited abov/ demonstrates that such total migration does not occur. Much sscientific evidence already provided to the Food and Drug Administration has demonstrated with con tainers having high levels of RVCM that far less than the maxi mum potentially migrateble quantity of vinyl chloride does in fact migrate even urder exaggerated exposure conditions. It is apparent, therefore, that with containers made with the extremely low level of RVCM discussed here, there can be no realistic expectation if migration of vinyl chloride, 4. Summarizing then, scientific-technical considerations based upon available information and established scientific SPI-01698 5 principles lead to the conclusion that there is no reasonable expectation of migration of vinyl chloride from container walls when the level of residi al vinyl chloride in such walls is adequately low. The data presented in this docket indicate that 0.1 ppm is an "adec uately low" level. Practical Considerations 1. Since the ch<mical tests show "no migration" of vinyl chloride, the majo r practical consideration is whether the analytical procedure s used to determine whether vinyl chloride migrates are sp fficiently sensitive to assure the public health and safet t when the "no migration" conclusion is relied upon. The Conmissioner has answered this question in the affirmative in ai analogous context, and we submit those same principles are clearly applicable here. In his rulemaking on chemical compounds in foodproducing animals, publ ished in the Federal Register on February 22, 1977 (42 Fed. Reg. 10412), the Commissioner clearly embraced the concept that the Ma ntel-Bryan biostatistical procedure is suitable for specify ing a level of analytical sensitivity for resolving the quest ion as to whether or not carcinogenic drug residues are prese nt in edible tissues of an animal to which that drug was fee If no residue of the drug in issue can be detected in the edible tissues of animals fed such drug SPI-01699 by means of an analytical procedure with a capability of de tection at levels speci lied by the Mantel-Bryan procedure as adequate to assure the >ublic health and safety, the Commissioner has determined that no drug residue is present. The same biostatistical extrapolation pro :edure, applied to vinyl chloride, will specify a detection sen sitivity that will likewise assure that a finding that no vinyl chloride is present in food or foodsimulating solvents wiltL assure the safety of the general public, Such a finding should likewise lead to the conclusion that no vinyl chloride is present--that there has been "no migra tion" of vinyl chloride 2. The file in this proceeding demonstrates that the application of the Mantsl-Bryan procedure to the data available at the time our original Comments were filed shows that a mini mum detection capability orders of magnitude higher than that actually available would be sufficient. In other words, a finding of no vinyl chloride in food-simulating solvents after suitably exaggerative e xposures provides far greater assurance of no significant risk of harm than the Commissioner indicated would be sufficient in the case of animal drug residues, hence, a far sounder basis for reaching a "no migration" conclusion. SPI-01700 7 3. In the secfcic n on Scientific-Technical Considera_ti_o__n__s, calculations and conclusions were presented for a typical bottle wall thickness oi 0.020 inches and a residual VCM con tent of 0.1 ppm. However, evidence has been presented showing that for many applications thinner walled containers and mate rials of much lower RVCM are used. Furthermore, many appli cations exist involving refrigerated shipment and storage, as well as products with limited shelf life. To the extent that any of these considerations apply, there will be still greater assurance of "no migration"; to the extent that several are simultaneously effective, the degree of assurance will be again compounded. 4. One final pr ictical consideration should be mentioned. Vinyl chloride is relat Lvely insoluble in aqueous or even fatty foods: it tends to eva porate from food to the air which is in contact with that fo 3d. This phenomenon is the basis for the most sensitive analytical procedures now available since the vinyl chloride in a relatively large volume of food or solvent is readily transferred into and thus concentrated in a smaller quantity of a ir. Thus, even i f a measurable level of vinyl chloride - were present in the con tents of an actual food container, the vinyl chloride would te nd to concentrate in the head space leaving only a very lov level in the food itself. When the SPl-01701 8 container is opened, th vinyl chloride in the head space will escape immediately, up on reclosure the same redistribution phenomenon from food to head space would reoccur and upon reopening the second in ;rement of vinyl chloride would escape. Likewise, food removed :rom such a container and exposed to air, stirred, heated or otherwise prepared would experience a rapid and substantial .y complete loss of any vinyl chloride it might have contained In other words, where formerly used containers may have per nitted the migration of vinyl chloride to a limited extent to the contents of such containers, the concentration of vinyl rhloride that actually became a component of the diet was tremend Dusly reduced. The application of this same principle to food packaged in the low RVCM containers adds still another majo r safety factor to assure there will be no vinyl chloride in food as consumed. Legal 1. Comprehensiv e and well documented evidence has been submitted to the Commis sioner demonstrating that there are no detectable extractiv es of vinyl chloride from food-contact articles into food-simi lating solvents at a detection limit of approximately 0.001 part per million. It is respectfully submitted that this showing demands the conclusion there is no basis for regulating the residual vinyl chloride content of PVC food-contact articles as a Food Additive as defined 9 in Section 201(s) of the Act. Failure to reach this conclusion would require the Commieisioner to find that the mere presence of residual vinyl chloriide monomer in the food-contact article is sufficient to determine that VCM in fact becomes or reasonably may be expected to becore a component or otherwise affect the characteristics of any rood irrespective of a factual basis for such a conclusion. It is respectfully submitted that such a determination is inconsistent with the governing statutory standard and the Commissioner's authority. (2) The Food Adiiitive definition in Section 201 (s) of the Act entails esse itially four (4) criteria: (a) That the sujstance becomes a component of food, (b) That the substance otherwise affects the charac teristics oE food, (c) That the su ^stance may reasonably be expected to become a co nponent of food, or (d) That the component may reasonably be expected to affect the (characteristics of food, The credible evidence before the Commissioner indicates that such residual vinyl chljoride as may remain in PVC food contact articles neither become s a component of food (i.e., is not a detectable component) nor affects the characteristic of any food. Furthermore, in that the known behavior of vinyl chloride in small but measurable quantities evidences that its presence in food does not affect. the characteristics of such food, a SPJ-01703 10 fortiori, its possible p resence in immeasurable quantities cannot affect the charac teristics of food. Thus, the first, second and fourth criter ia of Section 201(s) as stated above do not provide a basis for regulation as a food additive. Accordingly, the issue presented is whether vinyl chloride present in PVC food-contact articles at a level not exceeding 0.1 ppm may reasonably be expected to become a compo nent of any food. Obviously to reach an affirmative conclusion relies on a priori reasoning that necessarily ignores the only relevant scientific dat;i, namely that there are no detectable extractives of vinyl ch oride in the food-simulating solvents. Recognizing that there are few, if any, absolutes with respect to food additive chemistry, the congress established a standard based upon reasonable expectation rather than upon an absolute a ;counting for all potential contingencies under all undefined conJitions. As discussed in SPI's December 19, 1975 Comments, hypothetical abstractions which cannot be proven do not give rise to reasonable expectations; in the absence of a clearly expressed congressional intent to the contrary, such hypothetical abstractions do not form the basis for the exercise of delegated regulatory authority. Moreover, such a conclusion is not permissible under Section 201(s) which defines food additives in terms of "becoming a component or otherwise affecting the characteristics of any SPJ-01704 11 food. ..." In that thfe phrase "or otherwise affecting the characteristics of any f>od" refers back to the "component" criteria/ it thereby indicates a statutory intent that a compo nent affect the characteristics of the food. Thus, to the extent that a component, in the statutory rather than theoretical sense, must affect the characteristics of any food, it must be measurable and percep'tible in and of itself or in its effect, Otherwise stated, the ef feet of being a component is one of discernable physical prelsence. It is established, however, that vinyl chloride at levels less than 0.001 part per million does not have either an ascertainable physical presence or any other effect. Thus it cannot reasonably be considered to be a component of anj food. The Mantel-Bryan procedure adopted by the Commissioner in 42 Fed. Reg. 10412 confirms that FDA concurs with tt is position in an applicable analogous Section 409 situation. SPI-01705