Aluminosilicate Refractory Ceramic Fibres CAS Number: - EC Number:

12.11.2009

COMMENTS AND RESPONSE TO COMMENTS ON ANNEX XV SVHC: PROPOSAL AND JUSTIFICATION

Disclaimer: The European Chemicals Agency is not responsible for the content of this document. The Response to Comments table has been prepared by the competent authority of the Member State preparing the proposal for identification of a Substance of Very High Concern. The comments were received during the public consultation of the Annex XV dossier.

Substance name: Aluminosilicate Refractory Ceramic Fibres CAS number: - EC number: -

Reason of the submission of the Annex XV: CMR

General comments No. Date Submitted by (name, Comment Response Organisation/MSCA) 1 20090928 Air Liquide, Company, Since RCF are classified as a Category 2 Statement, no response necessary France carcinogen by the European Commission (EC), specific restrictions are imposed on the handling, use, and disposal of RCF in accordance with multiple Commission Directives (European Directives 69/97/EC). AES fibers have similar thermal properties to RCF, but are exonerated from carcinogenic classification by the EC. The objective of this document is to The intention of the authors was to demonstrate; that RCFs are demonstrate that substitution of RCF related to risks for workers and secondly that these risks are (Refractory Ceramic Fibres), used as avoidable when substitutes are used as insulation material. insulation materials for Steam Methane Reforming Units, by AES (Alkaline Earth Actually, AES (Alkaline Earth Silicate) fibres are not classified as Silicate) does not improve workers safety. carcinogenic. Therefore, substitution gives the opportunity to improve the safety of workers. For details see comment on Steam Methane Reforming below. 2 20091005 Individual, United Kingdom I support the nomination of this chemical to No response necessary the Candidate List, and believe it is important, given its properties, for it to be as strictly controlled as possible. 3 20091005 SELAS-LINDE GmbH, p. 2.1: Alternative Substances The Steam Methane Reforming (SMR) process consists of the Company, Germany two steps1 that are associated with different temperature ranges Generally spoken it doesn't make sense to up to 1100 °C. In addition, insulation material comes into contact discuss the use of AES for high temperature with H2 and CO which leads to decomposition of the material. process applications at all: AES fiber is an insulation material which doesn't have In general there are different bio-soluble fiber products of AES similar thermal properties to RCF as it can (Alkaline Earth Silicate) available on the market. Their permanently be used only at temperatures classification temperature – provided by the product data sheet below 800°C. Installed in furnaces with – comprises from 1000 up to 1300°C (e.g. product known under fluegas temperatures of more than 800°C the trade name Superwool HT 6072). This implies that products (e.g. steam methane reforming - SMR) it will of AES for high-temperature applications up to about 1100- be destroyed soon (shrinkage, 1170°C are suitable for long-time use under “real operating embrittlement, vitification) – apart from the conditions”. Therefore the application temperature of bio-soluble formation of cristoballite. fiber products covers the needed temperature range in the SMR The discussion of advantages or process. disadvantages of AES instead of RCF material regarding improvement of safer The substitution of RCF fibres is regulated in Germany in the work therefore is not decisive. Technical Rule for Hazardous Substance TRGS 619 “Substitute If there should be the demand to use Materials for Aluminium Silicate Wool Products”. Therefore the alternative material instead of RCF for high dossier reflects the approach taken in TRGS 619 (2007)3. With temperature applications it can not be AES regard to thermal insulation in furnace and firing system fiber but brick linings or castable linings. construction (which may include the situation as for SMR) the These alternative lining materials are TRGS 619 recommends appropriate substitutes in Annex 1 possible for SMR applications but their (Descriptive profile for selecting a substitute for aluminium physical properties have to be regarded silicate wool products for thermal insulation in furnace and firing under consideration of the side-effects: due system construction, especially at application temperatures to higher heat storage capacity of castable higher than 900°C in accordance with § 7 paragraph 1, No. 5 or brick linings the units have to be Ordinance on Hazardous Substances). equipped and operated in different ways (additional cooling of coils, slow cooling We agree that the formation of Cristoballite can occur if AES down, etc.). In addition it is rather unlikely to fibres are used above 900 °C. But (according to a literature do completely without RCF material in high search) at present no published data regarding the exposure

1 Nyserda – New York State Energy Research and Development Authority: Hydrogen Production – Steam Methane Reforming (SMR), w.y., p. 1. http://www.getenergysmart.org/files/hydrogeneducation/6hydrogenproductionsteammethanereforming.pdf 2 Some references of using AES-products for several industrial applications could be found e.g. under http://www.thermalceramics.com/site.asp?siteid=146&pageid=163 3 TRGS 619 (English Version): Technical Rule for Hazardous Substances- Substitute materials for aluminium silicate wool products, http://www.baua.de/nn_16800/en/Topics-from-A-to-Z/Hazardous-Substances/TRGS/TRGS-619

- 2 - temperature units either: some joints have level of Cristoballite from AES fibres is available. to be sealed with flexible temperature It is also important to note that the formation of cristoballite from resistant material – RCF is in most cases RCF fibres can occur as well (Class, 20024). In addition Class5 the best compromise of durable function (2003) has shown that Cristoballite exposure in the course of and economic matters. removal operations can exceed the limit value used in a number Comments on the proposals: of countries (0.05 mg/m3). The average concentration as Classification of AES fibers exposed to high presented in the quoted literature is about 0.096 mg/m3. The 3 temperatures for long duration as products proposed OSHA limit value for crystalline SiO2 is 0.025 mg/m containing crystalline silica and also the (ICSC 08086). clarification of the methodology used to To our knowledge these statements do not correspond with the measure exposure levels to fibers may be latest results in research/ state of technology on the field of RCF correct approaches but doesn't change the substitution. Just recently7 a new microporous calcium fact that AES fiber material is not suitable hexaluminate insulating material has been developed called for SMR linings at all (due to its limited SLA-92.8 This material is fiber- and silica-free (low level of total application temperature). impurities – SiO2 and Fe2O3 contents each are maximum Selas-Linde shares the acknowledgement 0.1%)9. Aggregate enables the formulation of: castables, that no validated substitute for RCF exists gunning mixes (e. g. could be used to fill joints), bricks. SLA-92

4 P. Class, R.C. Brown;Exposition gegenüber künstlichen Mineralfasern, Gefahrstoffe – Reinhaltung der Luft; 62, Nr. 5, 2002 5 P. Class; Current Fibrous Dust Workplace Concentrations and Trend in the High Temperature Insulation Wool Industry: the Results of the Care programme, VDI-Bericht Nr. 1776, 2003 6 ICSC: International Chemical Safety Card No. 0808, Crystalline silica, quartz Crystalline silicon dioxide, quartz Silicic anhydride SiO2 , Molecular mass: 60.1 7 Bundesministerium für Wirtschaft und Arbeit BMWA, Project No. 032 7258 8 Super Lightweight Aggregate of 92% Al2O3. 9 Wuthnow H.; Pötschke J.; Buhr A.; Boßelmann D.; Pozun F.; Gerharz N.; Golder P.; Grass J.; (w.y.), p. 200. 10 The lower thermal conductivity, the better insulation. 11 Overhoff, A.; Buhr, A.; Grass, J.; Wuthnow, H. (2005), p. E3. 12 Kockegey-Lorenz, R.; Buhr, A.; Racher, R. (2005), p. 19. 13 Boßelmann, D.; Buhr, A.; Gerharz, N.;Golder, P.; Grass, H.-J.; Overhoff, A.; Wuthnow, H.; Pötschke, J (2004), p. 5. 14 Kockegey-Lorenz, R.; Buhr, A.; Racher, R.: Industrial Application Experiences with Microporous Calcium Hexaluminate Insulating Material (presentation), 48th-International Colloquium on Refractories, Aachen, September 2005 – SLA-92, p. 19 15 http://www.ultraliteinsulation.com/pdf/Ideal_Standard_Case_Study_v2-2.pdf 16 http://www.ultraliteinsulation.com/pdf/UltraliteB_%20HT_Data_Sheet.pdf 17 http://www.ultraliteinsulation.com/kilncarbase.html 18 Open porosity by 88%. 19 http://www.promat.de/twd/default.asp?PAGE=NEWS&LANGUAGE=E 20 http://www.cellaris.com/pictures/files/Cellaris%20Company%20Profile%2001102009.pdf and www.cellaris.com http://www.cellaris.com/pictures/files/Insulation.pdf 21 http://www.cellaris.com/contents/page.asp?contentPageID=31&contentCatID=26 22 http://www.microthermgroup.com/EXEN/site/markets-detail.aspx?vPK=51&k=15&l=6&page=0 23 http://www.microthermgroup.com/EXEN/site/products-detail.aspx?vPK=31&k=3&l=1 24 Microtherm – Product Performance Brochure, p. 7. http://www.microthermgroup.com/EXEN/site/downloads-overview.aspx?k=194&l=8

- 3 - for SMR application under consideration of has similar technical properties like RCF: low density, low economic aspects. thermal conductivity 10 of 0.36 W/m*K at 1200°C (vs. RCF of 0.41-0.44 W/m*K at 1200°C)11, long term stability is up to 1500°C, excellent resistance to thermal shock, chemical stability e.g. calcium hexaluminate has excellent resistance to alkali or carbon monoxide (CO) attack and hydrogen (H2) containing reducing process atmosphere and low shrinkage.12 In addition calcium hexaluminate has thermal expansion and fracture 13 toughness comparable to α-Al2O3 (corundum) . All these listed technical properties are required for the SMR process (see above) and allow likewise RCF to save energy consumption therefore CO2-emission could be reduced. In general the new developed material SLA-92 has better thermal insulating properties due to lower thermal conductivity as RCF particularly in high temperature range. Therefore it can be concluded that SLA-92 has more potential for saving energy and consequently reduction of CO2-emission. The downstream products were successfully tested over long time periods within this joint project in different industries.

Contact of RCFs with high SiO2 contents towards process gases like H2 and CO lead to decomposition of the insulating material (Fouling). Due to the absence of SiO2 this is not the case for the new material e.g. SLA-92. Therefore the lifetime of insulating material could be increased significantly. From the economic point of view material cost and cost due to manpower requirement for maintenance will be saved. Actually, several applications of SLA-92 in downstream products are known e.g. successful use for lining gas tubes in synthesis gas reformer, pipe insulation in steel reheating furnaces, submerged nozzle insulation and ladle preheater cover lining in the steel industry as well as for kiln car lining in the ceramic industry14.

In addition to SLA-92 there are other available alternatives to RCF on the market. Based on the information given by the producers, these materials have the following advantages in some relevant characteristics. 1) fiber-free foamed clay aggregate15

- 4 - - technical suitable for the use in temperatures up to 1400°C - potential for energy saving due to lower thermal conductivity of 0.318 W/m*K 16at 1200°C - has been successfully installed since2000 - service life of foamed clay aggregate is 4 to 5 years (vs. tradition material 2 to 3 years) - energy savings within in kiln car applications of up to 33% versus RCF products and up to 40% versus conventional refractory insulation17

2) high-temperature microporous18 insulation material based on alumina as well as on alumina / mullite19 - suitable for applications up to 1700°C - fibre-free and lacks the ability to build free quartz - combines a low bulk density with low thermal conductivity (e.g. 0.45 W/m*K at 1200 °C) and therefore provides low energy consumption - low shrinkage (e.g. 0.1% 24 h at 1600°C) - production of shaped parts is possible - fast heating and cooling processes are possible

3) highly porous, ultra-light and non-fibrous ceramic foams20 based on alumina and mullite - application temperatures up to 1750°C - light weight; thermal shock resistance, long service life, chemical stability in aggressive atmospheres - low thermal conductivity (0.46 W/m*K at 1200°C) - low shrinkage (after 24 h at 1700°C 1.5%) - Application fields of ceramic foams are: kiln and kiln car insulation as well as kiln furniture in furnaces with gases like hydrogen and others21 - fast heating and cooling processes are possible

4) Ultra high temperature microporous insulation material22 known by the trademark Microtherm® Super A - silica free - continuous operational exposure capability of 1150 °C 23 - shrinkage of less than 2% for 24 h exposure, resistant to

- 5 - thermal shock24 and mechanically stable for long periods at maximum operating temperatures - available in a wide range of (tailor made) complex shapes and forms as well as standard forms ex stock - very low thermal conductivity over a wide temperature range up to 1150 °C - this insulating material has potential of 28% for energy saving (in comparison to RCF) - could save space in the furnace up to 33% due to around one quarter the thickness (in comparison to RCF) - from the economic point of view this capability allows increasing of productivity of 33 % and reduction due to energy costs and costs per item (economy of scales) 4 20091005 Lurgi GmbH, Company, Since RCF are classified as a Category 2 Statement, no response necessary Germany carcinogen by the European Commission (Comment is identical to Air Liquide - see above) (EC), specific restrictions are imposed on the handling, use, and disposal of RCF in accordance with multiple Commission Directives (European Directives 69/97/EC). AES fibers have similar thermal properties to RCF, but are exonerated from carcinogenic classification by the EC. The objective of this document is to The intention of the authors was to demonstrate; that RCFs are demonstrate that substitution of RCF related to risks for workers and secondly that these risks are (Refractory Ceramic Fibres), used as avoidable when substitutes are used as insulation material. insulation materials for Steam Methane Reforming Units, by AES (Alkaline Earth Actually, AES (Alkaline Earth Silicate) fibres are not classified as Silicate) does not improve workers safety. carcinogenic. Therefore, substitution gives the opportunity to improve the safety of workers. For details see comment on The evolution of the toxicity according to Steam Methane Reforming below. (Comment is identical to Air operating conditions and the methodology Liquide - see above) proposed to measure exposure level will be developed. 5 20091006 Refractory Ceramic Fibers The Proper Classification of Alumina Statement, no response necessary. Coalition (RCFC), Industry Silicate Wools (ASW), also known as or trade association, United Refractory Ceramic Fiber (RCF), under States of America 1272/2008/EC of the Classification, Labeling and Packaging (CLP) Standard

- 6 - The following comments are provided in In March 2006 the European suppliers of Aluminosilicate response to the German nomination of Refractory Ceramic Fibres provided the EU Technical Aluminosilicate Refractory Ceramic Fibers Committee on Classification and Labelling of Dangerous to the candidate list as a substance of very Substances with the currently available scientific data with high concern (SVHC) under REACH. The respect to Aluminosilicate Refractory Ceramic Fibre Annex XV dossier should be rejected as carcinogenicity. The aim of the suppliers was to re-discuss flawed for many reasons but specifically carcinogenicity. France and Germany had responded to that due to the fact that RCF is misclassified and documentation that a re-discussion was not warranted as the does not meet the criteria as a category 2 new studies do not have any impact on the existing classification carcinogen. The RCF industry is working as inhalative carcinogen. The Technical Committee on on the reclassification issue and intends to Classification and Labelling of Dangerous Substances seek the support of member states for a concluded that a re-discussion will only take place in case a review of the current classification of RCF. member state would support such re-discussion. There was no Under amendment to Council Directive support by any member state. Since 2006, no further relevant 67/548/EEC, Commission Directive scientific data have become available. The existing classification 97/69/EC (December 1997), refractory with R49 or H350i is still adequate and does not need to be re- ceramic fibers (RCF) were classified as a evaluated. category 2 carcinogen. Since the time of this classification decision, two major developments have occurred: 1) we now have over a decade of new pertinent scientific data and 2) the adoption of 1272/2008/EC (CLP) under Article 37 (6) providing for the use of new scientific data which may lead to a change in the harmonized classification and labeling of RCF. 6 20091008 European Association of The following comments and questions are Statement, no response necessary. the High Temperature based on a thorough review of all elements Insulation Wool Industry of the dossier. It is the industry view that the (ECFIA), Industry or trade document is flawed; scientific conflicts and association, France questions posed need resolving prior to consideration for listing on the candidate list. The description of the identity of the The entry in Annex VI of regulation 1272/2008 gives a broad substance, or substances, in these two definition of the chemical composition of those fibres classified. dossiers is confused: In the first instance, Theoretically numerous types of fibres are covered by this entry. there is no need for two dossiers as both Nevertheless, due to our knowledge, actually only those RCFs the materials identified are classified are on the market, that are described as RCF 1 to RCF 4 in the

- 7 - together in Europe as a CMR category 2 scientific literature. substance under directive 67/548/EEC and An amount of approx. 18 % zirconium dioxide in the RCF 2 its classification has been transferred to the fibres makes them a separate chemical entity. Therefore, a CLP regulation (carcinogen category 1b separate dossier was written although no differences were seen substance). In the second instance, they concerning workers exposure, human health effects or are glasses and are described using the protection measures and therefore concerning the risk related to absence of network modifying elements. exposure towards these fibres. The classification is not affected by the ratio of network formers – alumina, zirconia and silica. The dossiers reject inhalation studies which The dossier does not challenge the outcome of the RCF are the basis of the existing classification inhalation experiments with respect to a qualitative evaluation and attempt to install IP injection studies for (i.e. classification for carcinogenicity). The rat inhalation regulatory purposes (classification, risk experiments are challenged because of their inability to derive assessment) which are not universally adequate potency estimates which is crucial for cancer risk accepted. assessment. The latter is the case as there is an unquantifiable impact of the concurrent particle exposure. The fact that Aluminosilicate Refractory Ceramic Fibres are carcinogenic is not challenged at all on the basis of chronic inhalation and IP data.

- 8 - experimental animals to be classified as ‘‘sufficient,’’ as this test system reveals the carcinogenic activity of asbestos fibres and of fibres with lower potency in relation to asbestos fibres. Moreover, it has to be recalled that mesothelioma induction by intraperitoneal injection of fibres is hardly influenced by concomitant injection of granular particles. The test model of intraperitoneal injection of fibres revealed that the carcinogenic potency of various man-made vitreous fibres can differ by three orders of magnitude (Wardenbach et al., 200025), which enables the selection and use of less potent man-made vitreous fibres and the application of adequate protective measures. The existing classification must be re- In March 2006 the European suppliers of Aluminosilicate evaluated as new findings (published after Refractory Ceramic Fibres provided the EU Technical 1997) indicate that RCF was misclassified. Committee on Classification and Labelling of Dangerous Consideration of all scientific findings Substances with the currently available scientific data with available today means that it would fall into respect to Aluminosilicate Refractory Ceramic Fibre category 3 under 67/548/EEC or carcinogen carcinogenicity. The aim of the suppliers was to re-discuss 2 under CLP respectively, hence not carcinogenicity. France and Germany had responded to that fulfilling the SVHC criteria in terms of hazard documentation that a re-discussion was not warranted as the classification. new studies do not have any impact on the existing classification as inhalative carcinogen. The Technical Committee on Classification and Labelling of Dangerous Substances concluded that a re-discussion will only take place in case a member state would support such re-discussion. There was no support by any member state. Since 2006, no further relevant scientific data have become available. The existing classification with R49 or H350i is still adequate and does not need to be re- evaluated.

25 Wardenbach, P., Pott, F., Woitowitz, H.-J., 2000. Differences between the classification of man-made vitreous fibres (MMVF) according to the European directive and German legislation: analysis of scientific data and implications for worker protection. Eur. J. Oncol. 5 (2), 111–118.

- 9 - Furthermore, it is not clear what the authors The comment reflects on Article 58 (3) of the REACh regulation. of the dossier aim to achieve. As far as a The intention of this Article is to give arguments for a priorisation pre-selection for authorisation is concerned, of substances for their inclusion in Annex XIV. At the moment, the substance should be considered a low the German CA is interested in the 1ststep, inclusion on the priority as it is not a PBT or vPvB and its candidate list. Article 58 (3) states: “Priority shall normally be use is almost entirely (> 99%) industrial given to substances with: under controlled risk management (a) PBT or vPvB properties; or conditions (including technical exposure (b) wide dispersive use; or controls and the use of respiratory (c) high volumes.” protection where required). While the total We agree that RCFs are neither PBT nor vPvB. Nevertheless, tonnage manufactured in Europe is not the total tonnage manufactured in Europe is not insignificant (as insignificant, the real volume of concern is stated by ECFIA themself). In addition, the authors of the annex still very small as only the airborne dust XV dossier have given some data concerning workers exposure. released from the products during handling In different fields of usage of the RCFs. Therefore, some kind of is posing a possible health risk. widespread use has been identified although the consumer exposure is at least low. 7 20091009 3M Europe N.V., Company, 3M supports and promotes the fundamental Statement, no response necessary. Belgium principles of REACH, including but not limited to the goals of life cycle management, product safety, sustainable development, and effective risk communication to employees and customers. 3M appreciates the opportunity to comment on the proposed identification of Refractory Ceramic Fibers (RCF) as a candidate to the SVHC list. Because not all RCFs are carcinogenic, 3M 3M states that there are RCF fibres which are exempted from recommends to not add RCFs as a classification as carcinogenic due to their dimension and Candidate to the SVHC list, but to use the application of Nota R in Annex VI of the CLP regulation. Such market restriction approach (Annex XVII) as fibres would not fall under the intended authorisation procedure a more precise and application driven and would not need to be classified. If it can be assured that no approach to regulate the different and inhalable fibres are formed within the whole supply and use numerous RCFs. chain, such fibres could be an interesting substitute.

(Identity of the substance and physical and chemical properties) 8 20091009 Inter-Environnement We support the nomination of this chemical No response necessary Wallonie, National NGO, to the Candidate List, and believe it is

- 10 - Belgium important, given its properties, for it to be as strictly controlled as possible 9 20091012 Norwegian Pollution Control The Norwegian CA agrees with the No response necessary Authority, National identification of aluminosilicate refractory Authority, Norway ceramic fibres (SiO2, Al2O3) as a substance of very high concern according to Article 57 a) since the substance is classified as a carcinogen (Carc. Cat. 2 in the Directive 67/548/EEC and Carc 1B in the Regulation (EC) No. 1272/2008 and the COM Regulation (EC) No. 790/2009 (1st ATP to CLP)). We support that aluminosilicate refractory ceramic fibres should be included in the “Candidate List” of substances of very high concern for authorisation 10 20091012 Individual, Germany Since I was quoted several times in the Introductory statement, no response necessary dossiers I take the opportunity to comment on the content. The Annex XV dossiers for “Alumino Refractory Ceramic Fibres” and “Zirconia Aluminosilicate Refractory Ceramic Fibres” submitted by the German authorities on the ECHA website are erroneous in many ways. Germany should be asked to withdraw these dossiers. I would like to group my comments into four main sections:

1. Factual errors and inaccuracies in the dossier content 2. Dubious repetitions of German and European regulatory discussions 3. Fundamental questions on hazard classification 4. Socio-economic and environmental impacts

1. Factual errors and inaccuracies in the

- 11 - dossier content

The dossier contains numerous errors and inaccuracies concerning the description of the substance in question, its chemical and physical characterization, the use sectors and quantities and the possibility of substitution. Furthermore some “scientific” statements are simply wrong. The following list provides only some examples: • The following terms are used throughout Within the Annex XV dossier the authors used the nomenclature the dossier without clear definition: of the Refractory Ceramic Fibres as it is used in the original o Page 1: “Alumino Refractory Ceramic literature. Concerning the aluminium silicate fibres the IARC Fibres” reports (as well as a document of TIMA) focus on the subtypes o Page 6-ff: “RCF1,2,3,4-types” RCF 1 – RCF 4. Definitions are given in the literature cited in the o Page 20 “Aluminium silicate wool” dossier26 o Page 24: Table 9 Ceramic Fibrefrax, Ceramic MAN, Ceramic Fibrefrax II, Ceramic Manville, Ceramic Fibrefrax I o Page 27: “superfine RCF” • Test Samples (RCF 1,2,3,4) are confused i) The general requirement for an experimental inhalation test with the commercial product: is that the sample needs to be respirable for the animals o Page 6: “RCF types 1, 2, 3 and 4 where tested. Thus, the sample has to be prepared accordingly. type 4 fibres have no commercial Importance”: Not only “type 4 fibres” but ii) RCF wools release fibres which are respirable for humans none of the mentioned RCF-types have (similar dimensions to those used in the animal tests). Thus, commercial importance because these such fibres are relevant with respect to a putative human types are specifically separated fibrous dust carcinogenicity. samples from different commercial type wools and only made to carry out animal iii) According to Mast et al. (see above) the types of RCF fibres tests. studied were chosen because they represent the types most o The “RCF types” mentioned in table 1 commonly used in the US (RCF 1 and RCF 3) are separated fibrous dust samples from different RCF-types

26 Mast, R. W. et al. (1995a) TIMA, (1993) IARC (2002) (for details see References in Annex XV Dossier.)

- 12 - (kaolin and high purity) and specially made to carry out animal tests. o Page 6: “RCF stock3”; see also footnote: “RCF fibres as they derive from production”. This is incorrect as “RCF-stock” is fibrous dust mixed-up with particulates and not the substance for commercial use. • In both dossiers the CAS number 142844- The substance name is given with “Aluminosilicate Refractory 00-6 is not mentioned. Despite the fact that Ceramic Fibres”, a sub-group of CAS number 142844-00-6. This ECHA accepted the pre-registration under CAS number refers to the CAS name “refractories, fibers, the CAS number. As a matter of fact and aluminosilicate”, which are defined by CAS as “amorphous man- follow-up in the pre-registration process the made fibers produced from the melting and blowing or spinning EU number 604-314-4 was assigned to the of calcined kaolin clay or a combination of alumina (Al2O3) and commercially used RCF-substance by silica (SiO2). Oxides such as zirconia, ferric oxide, titanium ECHA. oxide, magnesium oxide, calcium oxide and alkalies may also be added. Approximate percentages (by weight) of components follow: alumina, 20-80%; silica, 20-80%; and other oxides in lesser amounts.” This generic description covers a multitude of different substances and is therefore not specific for the Aluminosilicate RCF resp. for the Zirconia Aluminosilicate RCF (see concentration range of the main constituents in chapter 1.2). • The tonnage of 25 000 t/a mentioned in In the dossier it is clearly stated that the tonnage of 25 000 t/a is each of the two dossiers give the related to the total amount of RCF fibres. No differentiation impression that a total of 50 000 t/a is between the fibre types has been made. produced in Europe: o Page 15: “The global production of RCF amounts to about 150 000 - 200 000 tonnes [NAIMA/EURIMA, 2001], the production of RCF in the EU was 50 000 tonnes, undertaken by three companies in 1999 and has been reduced to 25 000 tonnes in 2008 [Wimmer, 2008].” o In fact the total volume of both “Alumino Refractory Ceramic Fibres” and “Zirconia Aluminosilicate Refractory Ceramic Fibres” is 25 000 t/a.

- 13 - • The dossiers states that amphibole We agree that the wording was imprecise. Meant is the asbestos is not cleaving along the axis following: which is scientifically wrong: In contrast to serpentine asbestos, refractory ceramic fibres o Page 23, Quote: “...Thus, only results tend to break transversely rather than cleaving along the fibre from ip tests with amphibole asbestos (i.e. axis. The behaviour to cleave along the fibre axis is associated crocidolite), which does not cleave along with the fact that numerous new fibres generate the fibre axis, were used to assess the intraperitoneally which may increase dose and have an impact comparative carcinogenic potency of on the test outcome. The tendency to cleave along the fibre asbestos and refractory ceramic fibre...”. axis is different for different types of asbestos. Among the o Crocidolite is cleaving along the axis, different types of asbestos, it is lowest for crocidolite. Thus, RCF like all other MMVF do not (see only results from ip tests with crocidolite were used to assess attached SEM-picture). the comparative carcinogenic potency of asbestos and refractory ceramic fibres. • The substitution for RCF/ASW was In TRGS 619 the classification temperature (not application discussed for years by the German temperature) of AES wool products is defined up to 1050- regulatory body and resulted 2007 in the 1250°C. Since the preparation of the document (2006) the TRGS 619 published by the Ministry of insulating material based on AES was improved. For example Labour and Social Affairs. the classification temperature of the product Superwool HT o Page 20/21, Substitution discussion: “.... 60727(trade name) is 1300°C, referring to the product data the upper temperature limit of AES wool sheet. products could be increased significantly…” o After more than 20 years of research and product development and based on practical experience this statement is not justified at all.

27 Some references of using AES-products for several industrial applications could be found e.g. under http://www.thermalceramics.com/site.asp?siteid=146&pageid=163

- 14 - o According to TRGS 619 the following We agree with TRGS 619, all quoted criteria above have been points have to be taken into consideration: taken into account by submission of the Annex XV dossier on “The selection of a possible substitution RCF. There are new identified alternatives which are should be performed as part of a general successfully being used in different branches (For details see assessment based on the entire lifecycle of response to SELAS-LINDE GmbH, Company, Germany the possible products and is successful 20091005, above). when the products: - exhibit lower health risks during their entire life cycle, and - the (technical) characteristics are of equivalent value (application temperatures, thermal insulation properties, long-term behavior and service lifetime), the environment protection criteria are comparable (raw materials requirement, energy consumption, CO2 emissions, waste quantity) and the economic efficiency criteria (installation and running costs) do not result in any unreasonable disadvantages (socio-economic aspects).” • Many actual references are missing and We disagree. This comment is very general and would need often only earlier publications on specific specification to be able to explain in each point why we details are being used, omitting later and disagree. more relevant findings even by the same authors (i.e. Davis).

- 15 - 2. Dubious repetitions of German and Since discussion of the quoted documents new development European regulatory discussions took place on the market for insulation material

The authors of the dossier use several The authors of the annex XV dossier focused on worker references to regulatory discussions by exposure rather than public exposure to RCFs. German or European agencies in the past. These discussions led to conclusions which were omitted in the dossier. • On the European level the German authorities (BMU) proposed a EU- notification in 2004 which was withdrawn in 2006 by the German authorities after intervention from the EU-Commission (message 009 – communication from the Commission – SG(2006) D/50654). o Page 20, Quote: “Supplementary Text to Notification 2004/370/D“ o The actual dossier does not contain any further information on the subject. o A study carried out by Schneider et.al. “Ubiquitous fiber exposure in selected sampling sites in Europe” could not demonstrate any public exposure to RCF dust. • In Germany an intense discussion on the This is not correct. The cancer estimates for RCFs (and other risk evaluation (ERB-KMF) related to MMVFs) was approved by the German toxicology expert RCF/ASW took place on the AGS-level committee of AGS (UA III). The AGS took note of the risk (Ausschuss für Gefahrstoffe). In its May evaluation and decided to take it as a basis for the derivation of 2009 meeting the AGS did not accept the protective measures at the workplace. ERB-KMF for the use in regulation. o Page 18: “In Germany, a concept has been established to quantify cancer risk for workers after exposure to carcinogens in order to derive appropriate workplace measures [AGS, 2008]. … According to this concept, the tolerance level range for refractory ceramic fibres lies … respectively.” o It is correct that the responsible German

- 16 - authority body (AGS) agreed on ranges for tolerated und acceptable risk, in June 2008. o It is not correct that there was agreement on the risk evaluation (ERB-KMF) related to RCF/ASW which is stated in the dossier. [Reference: AGS –meeting in Mai 2009]. o The basis for the AGS-decision were the massive “uncertainties” of the presented ERB-KMF in total. • A Petition to the European Parliament in The quoted document was cited in relation to consumer 2007 is mentioned, omitting the final exposure. As this is not the main topic of the annex XV dossier conclusion of the European Parliament. and consumer exposure is expected to be significantly lower o Page 20: “In the context of a petition to than worker exposure, no response is necessary. the European Parliament on RCF from catalytic converters, the Commission has called on the automobile industry to provide scientific data on release into the environment of inorganic ceramic fibres from catalytic converters during their use. [European Parliament, 2007]” o The European Parliament rejected the petition in October 2007 (for reference see http://www.europarl.europa.eu/meetdocs/20 04_2009/documents/cm/691/691790/69179 0en.pdf) 3. Fundamental questions on hazard The dossier does not challenge the outcome of the RCF classification inhalation experiments with respect to a qualitative evaluation (i.e. classification for carcinogenicity). The rat inhalation The existing EU-Classification is based on experiments are challenged because of their inability to derive animal inhalation studies. The dossier adequate potency estimates which is crucial for cancer risk challenges the outcome of the inhalation assessment. The latter is the case as there is an unquantifiable experiments and recommends IP-tests for impact of the concurrent particle exposure. The fact that risk-evaluation. Aluminosilicate Refractory Ceramic Fibres are carcinogenic is o Page 22/23: “The fibre samples used in all not challenged at all on the basis of chronic inhalation and IP the chronic inhalation studies had relevant data. portions of nonfibrous particles (50 - 75% related to numeric comparison). These In March 2006 the European suppliers of Aluminosilicate particles were postulated to have an Refractory Ceramic Fibres provided the EU Technical influence on lung carcinogenicity. Thus, it is Committee on Classification and Labelling of Dangerous

- 17 - not clear which portion of lung tumours in Substances with the currently available scientific data with the chronic inhalation studies is assignable respect to Aluminosilicate Refractory Ceramic Fibre to fibre exposure. … Moreover, the data carcinogenicity. The aim of the suppliers was to re-discuss with various other carcinogens show that carcinogenicity. France and Germany had responded to that the Syrian hamster does not seem to be a documentation that a re-discussion was not warranted as the valid model for inhalation carcinogenicity new studies do not have any impact on the existing classification [Mauderly et al., 1997].” as inhalative carcinogen. The Technical Committee on o Page 23: “… the strategy to derive risk- Classification and Labelling of Dangerous Substances related information for the inhalation concluded that a re-discussion will only take place in case a carcinogenicity of refractory ceramic fibres member state would support such re-discussion. There was no is to compare the potencies of RCF to support by any member state. Since 2006, no further relevant asbestos fibres in the intraperitoneal test.“ scientific data have become available. The existing classification o These quotes raise several questions: with R49 or H350i is still adequate and does not need to be re- evaluated. • As the dossier challenges inhalation tests - As stated above, the dossier does not challenge the outcome of which were the main basis for the EU- the RCF inhalation experiments with respect to a qualitative classification - would this mean that the EU- evaluation (i.e. classification for carcinogenicity). Classification should be reviewed? It had been industry in the past who challenged the outcome of the inhalation experiments during the discussions. It was claimed that the inhalation test with rats is not relevant for humans due to overload. The latter is scientifically not accepted (e.g. from IARC). • Are the authors of the dossier trying to The classification criteria according to the dangerous substance establish the IP-animal test-model on EU directive and also according to the CLP regulation allow to level? Does Germany intend to use this assess fibre carcinogenicity by intraperitoneal testing. This dossier to establish IP testing as a means that the intraperitoneal test is scientifically accepted in “standard” for hazard classification? EU legislation. The critique expressed on the questionable o The use of IP-tests for hazard validity of this test is therefore baseless. classification would be in contradiction to existing EU practice and is not in line with Moreover, taking into account the human cancer incidences the expert opinion of almost all international caused by asbestos, it is prudent to regulate other fibres on the scientists. basis of the entire evidence including intraperitoneal injection studies. Positive results after intraperitoneal injection are regarded as essential for evidence of carcinogenic activity in experimental animals to be classified as ‘‘sufficient,’’ as this test system reveals the carcinogenic activity of asbestos fibres and of fibres with lower potency in relation to asbestos fibres. Moreover, it has to be recalled that mesothelioma induction by

- 18 - intraperitoneal injection of fibres is hardly influenced by concomitant injection of granular particles. The test model of intraperitoneal injection of fibres revealed that the carcinogenic potency of various man-made vitreous fibres can differ by three orders of magnitude (Wardenbach et al., 200028), which enables the selection and use of less potent man-made vitreous fibres and the application of adequate protective measures. o The criticism on the validity of the RCC- In March 2006 the European suppliers of Aluminosilicate inhalation experiments for regulatory Refractory Ceramic Fibres provided the EU Technical purposes is comprehensible and should Committee on Classification and Labelling of Dangerous lead to an expert debate to achieve a Substances with the currently available scientific data with Globally Harmonised Classification of RCF. respect to Aluminosilicate Refractory Ceramic Fibre o This is also supported by a conclusion carcinogenicity. The aim of the suppliers was to re-discuss taken by IARC in 2002 after evaluating new carcinogenicity. France and Germany had responded to that scientific evidence that became available documentation that a re-discussion was not warranted as the after the classification of 1997. IARC new studies do not have any impact on the existing classification concluded that RCF should be considered a as inhalative carcinogen. The Technical Committee on “group 2b: possible carcinogen” which is Classification and Labelling of Dangerous Substances similar to a category 3 carcinogen under concluded that a re-discussion will only take place in case a 67/548/ECC. member state would support such re-discussion. There was no o J.M.G. Davis, an expert also mentioned in support by any member state. Since 2006, no further relevant the References of the Dossier, in 2002 scientific data have become available. The existing classification came to a similar conclusion in “An with R49 or H350i is still adequate and does not need to be re- Assessment of Current Experimental Data evaluated. Regarding the Carcinogenicity of Refractory Ceramic Fibres (RCF).” (see Reference There is no regulatory possibility to establish globally attached) harmonized classifications. o This was also concluded by R.C. Brown, B. Bellmann, H. Muhle, J.M.G Davis and L.D. Maxim in 2005: “In particular, in the European classification system, Category 3(b) seems more appropriate than Category 2 for RCF.

28 Wardenbach, P., Pott, F., Woitowitz, H.-J., 2000. Differences between the classification of man-made vitreous fibres (MMVF) according to the European directive and German legislation: analysis of scientific data and implications for worker protection. Eur. J. Oncol. 5 (2), 111–118.

- 19 - 4. Socio-economic and environmental The new materials and the products based on bio-soluble wool Impact AES have a potential for saving energy due to the key property of the insulating material: thermal conductivity (For details see In addition to the comments above I want to response to SELAS-LINDE GmbH, Company, Germany point out, that the use of these materials 20091005, above) lead to less energy use up to 50% [Pfeifer, Several alternatives have lower thermal conductivity value in Springer – Fachkolloquium Hennef comparison to RCF and therefore provide a better insulating 16.04.2008], respectively CO2-reduction in capacity. Europe’s ambitious environmental targets in its many industrial high temperature Climate Change Programme (ECCP) can be achieved by using applications. The materials are needed to alternatives for RCF. achieve Europe´s environmental targets in its Climate Change Programme (ECCP). Already in 1995 a study was sponsored by Since the study was published there were several R&D activities the EU commission and carried out by ERM to create equivalent alternatives from the technical and (Environmental Resources Management) economic point of view to RCF. Currently there are appropriate on behalf of DG Enterprise. This study insulating materials on the market (For details see response to concludes that RCF is of tremendous SELAS-LINDE GmbH, Company, Germany 20091005, economical and ecological value for the above)An update of the mentioned study is needed, because European Union and its industry as a whole. new alternatives have also to be taken into account to avoid distortion of competition. Highlighting RCF on the proposed On the one hand the jobs might be reduced in production of candidate list already led to confusion in the RCF, but on the other hand the number of employees is industrial market. Therefore not only the expected to increase in production of the alternatives. In the environmental targets concerning CO2 downstream processes like installation, maintenance etc. the emission but also the competitiveness of jobs would not be reduced, because it does not make a and jobs in the European user industry are difference if there will RCF be used or other insulating material. at risk. The demand /need on insulating materials will remain the same and it determines the production volume. There are other different factors which determine competitiveness in the European user industry, not only RCF insulating material. Furthermore there are many alternatives available on the market. The submitted Annex XV Dossier on RCF does not focus on removal of insulating materials for industrial uses as whole. We support the approach to use insulating materials with the goal to reduce energy consumption and CO2-emissions, but these should be also chosen with respect to human health.

- 20 - We disagree. This comment is very general and would need Conclusion specification to be able to explain in each point why we disagree. The sheer fact of numerous factual errors included in the dossier should lead to its rejection. More importantly new scientific evidence In March 2006 the European suppliers of Aluminosilicate (published after the 23. ATP of 67/548/ECC Refractory Ceramic Fibres provided the EU Technical in 1997) leads to the necessity to re- Committee on Classification and Labelling of Dangerous evaluate the existing hazard classification. Substances with the currently available scientific data with respect to Aluminosilicate Refractory Ceramic Fibre carcinogenicity. The aim of the suppliers was to re-discuss carcinogenicity. France and Germany had responded to that documentation that a re-discussion was not warranted as the new studies do not have any impact on the existing classification as inhalative carcinogen. The Technical Committee on Classification and Labelling of Dangerous Substances concluded that a re-discussion will only take place in case a member state would support such re-discussion. There was no support by any member state. Since 2006, no further relevant scientific data have become available. The existing classification with R49 or H350i is still adequate and does not need to be re- evaluated. Any further regulation, in example through There is no regulatory possibility to establish globally the authorization process under REACH, harmonized classifications. must be based on an updated (global) harmonized hazard classification. As a first step the ANNEX XV listing of RCF must be withdrawn to prevent any further confusion which already took place in the market. 11 20091012 DKFG, Industry or trade DKFG is the German association of Statement, no response necessary association, Germany manufacturers and processors of High Temperature Insulation Wools (HTIW). It was founded in 1984. Its main objectives are the scientific evaluation and communication of health, safety and environmental aspects related to HTIW and its uses.

- 21 - The Annex XV dossier submitted by Comments on the attached document: Germany is fundamentally flawed and should be rejected because: - Page 5 of the Annex XV: The complete Annex I entry under Index No 650-017-00-8 refers - In the “PROPOSAL FOR to ‘Refractory Ceramic Fibres, Special Purpose Fibres … [Man- IDENTIFICATION OF A SUBSTANCE AS A made vitreous (silicate) fibres with random orientation with CMR CAT 1 OR 2, PBT, VPVB OR A alkaline oxide and alkali earth oxide (Na2O+K2O+CaO+ SUBSTANCE OF AN EQUIVALENT LEVEL MgO+BaO) content less or equal to 18 % by weight]’. The OF CONCERN” (page 5 of the dossier) the Aluminosilicate RCFs and the Zirconia Aluminosilicate RCF are authors fail to properly describe the identity actually the only commercially existing fibres which match to this of the substance they seem to be entry. Other fibres (i.e. alkaline earth silicate wool or high- suggesting for inclusion on ANNEX XV. alumina, low-silica-wool) do not fulfil this definition while the The definition given in this section of the (Na2O+K2O+CaO+ MgO+BaO) content is higher than 18 % by dossier is confusing and factually wrong. weight (see IARC Vol 18). If there will be other RCF these do not fall in the same field of application as the Aluminosilicate RCF.

- Page 7 of the Annex XV: According to REACH (RIP3.10) UVCB substances are specified by all known constituents, present at concentrations ≥10 %. These constituents are Al2O3 and SiO2 in Aluminosilicate RCF resp. ZrO2 in Zirconia Aluminosilicate RCF.

The CAS number 142844-00-6 refers to the CAS name “refractories, fibers, aluminosilicate”, which are defined by CAS as “amorphous man-made fibers produced from the melting and blowing or spinning of calcined kaolin clay or a combination of alumina (Al2O3) and silica (SiO2). Oxides such as zirconia, ferric oxide, titanium oxide, magnesium oxide, calcium oxide and alkalies may also be added. Approximate percentages (by weight) of components follow: alumina, 20-80%; silica, 20-80%; and other oxides in lesser amounts.” This generic description covers a multitude of different substances and is therefore not specific for the Aluminosilicate RCF resp. for the Zirconia Aluminosilicate RCF (see concentration range of the main constituents in chapter 1.2 of Annex XV).

- Page 8 of the Annex XV: The mentioned composition range refers on several literature

- 22 - references i.e. Glass et al (1995), TIMA (1993), Luoto (1995).

- RCF does not meet the selection criteria We disagree with the statement that worker protection is not in for the authorization process. The the focus of the authorization. Basically, recital 70 states underlying principle of the REACH ”Adverse effects on human health and the environment from regulation is to protect consumers and the substances of very high concern should be prevented through environment. the application of appropriate risk management measures to

RCF is used in industrial applications and ensure that any risks from the uses of a substance are under well regulated conditions. Consumer adequately controlled, and with a view to progressively exposure is negligible if existing at all. It is substituting these substances with a suitable safer substance. not an environmental pollutant as the Risk management measures should be applied to ensure, when material is inert when deposited in landfill. It substances are manufactured, placed on the market and used, is not a PBT or vPvB substance. that exposure to these substances including discharges, There is no need to consider further emissions and losses, throughout the whole life-cycle is below regulation as a matter of priority. the threshold level beyond which adverse effects may occur.(…)”. Article 57 names carcinogenic substances (Cat 1 or 2) as substances relevant for authorization. RCFs fulfil this criterion. - There is new scientific information In March 2006 the European suppliers of Aluminosilicate available (published after the 23rd ATP on Refractory Ceramic Fibres provided the EU Technical 67/548/EEC of 1997), raising concern on Committee on Classification and Labelling of Dangerous the validity of the underlying rational and Substances with the currently available scientific data with data used to justify a category 2 carcinogen respect to Aluminosilicate Refractory Ceramic Fibre classification in Europe. Notwithstanding a carcinogenicity. The aim of the suppliers was to re-discuss more detailed justification of this statement, carcinogenicity. France and Germany had responded to that IARC in their 2002 review concluded that documentation that a re-discussion was not warranted as the RCF should be considered a “possible new studies do not have any impact on the existing classification carcinogen – group 2b” which is as inhalative carcinogen. The Technical Committee on comparable to a category 3 classification Classification and Labelling of Dangerous Substances under 67/548/EEC. concluded that a re-discussion will only take place in case a The industry will submit a request for member state would support such re-discussion. There was no reclassification in line with article 37(6) of support by any member state. Since 2006, no further relevant Regulation (EC) 1272/2008. scientific data have become available. The existing classification with R49 or H350i is still adequate and does not need to be re- evaluated. The dossier contains several further The detailed points referred to are found elsewhere in this weaknesses and misinformations. Detailed document and will be specifically responded where they appear. individual comments and questions as well as some selected references are provided

- 23 - in the document attached. 12 20091012 Individual (affiliated with Alumino silicate refractory ceramic fibers We take note of this comment for a specific application, but we Howmet-Ciral snc, Howmet (RCF) are used in the production of metal assume that for this application an alternative could exist like in SAS; Howmet Ltd for Alcoa castings for the Aerospace, Defense and other cases i.e. steel industry (For details see response to Power and Propulsion, a Industrial Gas Turbine industry. Alcoa- SELAS-LINDE GmbH, Company, Germany 20091005, above) division of ALCOA, Howmet uses RCF high temperature In case that there is demonstrably no suitable alternative and Company), France insulation blanket to control metal the risks are adequately controlled or the socio-economic solidification and control thermally induced benefits overweigh the risks; these aspects would be taken into stresses which can cause catastrophic account in the procedure for authorization decisions. defects which are very hard to detect. The application of RCF insulation to specific features of the component to be manufactured allows the order of metal freezing to be controlled which results in a metallurgically sound component that meets customer requirements. The cooling rate of the different parts of the mold is controlled by the use of the RCF blanket. This technique is a fundamental and enabling technology for the Superalloy investment casting process. In the past 7 years, no thermally equivalent replacement alkaline earth silicate blanket is available to replace RCF insulating blanket at the temperatures required to produce Superalloy investment castings. Due to the high temperatures required by the melting point of metals (eg Superalloys) , RCF may also be used in fixed equipment such as casting or heat-treat furnaces. 13 20091013 BDI, Industry or trade The Annex XV dossiers for the two This is not correct. The cancer estimates for RCFs (and other association, Germany substances for “Alumino Refractory Ceramic MMVFs) was approved by the German toxicology expert Fibres” and “Zirconia Aluminosilicate committee of AGS (UAIII). The AGS took note of the risk Refractory Ceramic Fibres” submitted by the evaluation and decided to take it as a basis for the derivation of German authorities on the ECHA website protective measures at the workplace. (http://echa.europa.eu/chem_data/reg_int_t ables/reg_int_subm_doss_en.asp) are scientifically flawed and the AGS

- 24 - (Ausschuss für Gefahrstoffe - German Committee for Hazardous Substances) decision on the risk figures is wrongly quoted in many ways and should be rejected based on the follow¬ing arguments: EU-classification (67/548 EG) is based on The dossier does not challenge the outcome of the RCF inhalation tests. In the dossier, the inhalation experiments with respect to a qualitative evaluation impression is given that these tests are not (i.e. classification for carcinogenicity). The rat inhalation the basis for the evaluation, but i.p.-animal experiments are challenged because of their inability to derive tests. This is in contradiction to the official adequate potency estimates which is crucial for cancer risk classification. assessment. The latter is the case as there is an unquantifiable impact of the concurrent particle exposure. The fact that Aluminosilicate Refractory Ceramic Fibres are carcinogenic is not challenged at all on the basis of chronic inhalation and IP data.

- 25 - When aluminosilicate RCF substances were We disagree. This comment is very general and would need published in Annex XV the interested specification to be able to explain in each point why we industrial parties, associations and disagree. companies would get con¬fused. Especially because of the flawed and misleading comments in the dossiers. The German industry association BDI Based on the argumentation above we disagree with this therefore requests to reject the Annex XV conclusion. dossier until a decision is made after international expert consultation. Therefore the German industry (BDI) would There is no regulatory possibility to establish globally support an international expert hearing harmonized classifications. followed by a globally harmonized classification under the new CLP/GHS- regulation. The area where articles made from In this case industry should be easily able to demonstrate to the RCF/ASW are used in industrial granting authority that the risks to human health arising from the appli¬cations is well regulated and use of the substance are adequately controlled. observed by industry (BDI) and regulatory bodies like the AGS. Furthermore, more than 90 % of the Statement, no response necessary. products made of RCF/ASW are in use in industrial high temperature applications. In these applications more than 30% energy saving and respectively CO2-reduction can be achieved in many applications. (reference: Hennef Colloquium 2008; ECFIA-presenta¬tion to DG Env 2008) 14 20091013 Company, Germany The process of identification of SVHC with Statement, no response necessary. intention to authorization at a later point has limitations with respect to scope under REACh. A number of uses have been specifically In our view the usage of RCFs in Europe can not be described exempted from authorization under REACH. as Research and Development. Specifically, the authorization process cannot be used to target the following uses: • scientific research and development (see Article 56(3) of REACH) [additional references: “Guidance on

- 26 - inclusion of substances in Annex XIV (List of Substances subject to Authorization)”, http://guidance.echa.europa.eu/public- 2/getdoc.php?file=annex_xiv_en , page 20 chapter 2.5.2; “Workshop on the Candidate List and Authorization as Risk Management Instruments” http://echa.europa.eu/doc/consultations/aut horisation/authorisation_workshop_proceedi ngs_20090121.pdf, page 35, Annex 2 C.1] Therefore our company asks ECHA and the Statement, no response necessary. Member State Committee to share their understanding of this limitation in our case. Our company uses Refractory Ceramic We disagree. R&D is defined in Article 3 (23): “scientific Fiber (RCF) materials as insulation research and development: means any scientific materials in high end analytical instruments experimentation, analysis or chemical research carried out (e.g. gas chromatography systems) in small under controlled conditions in a volume less than one tonne per quantities. [The 0.1% wt/wt threshold is not year“. The usage of substances in commercial products like gas exceeded for the complete instrument; chromatography systems, even if they are used in a scientific however for individual sub-assemblies context, is not covered by this definition. and/or spare parts the limit of 0.1% wt/wt of In addition, we are of the opinion, that the temperature range of article is surpassed, the overall quantity a gas chromatic system will not exceed an upper limit of approx. shipped by our company into the EU is 450 °C. The usage of RCFs in this range of temperature can estimated to be well below 1000 kg per easily be avoided by using other isolating material. year]. The thermal properties of RCF materials are An analytical device for e.g. gas chromatographs does not required to achieve the essential speed, operate at high temperature range only up to 450°C. The sensitivity and reproducibility of the operating temperature in mass spectrometer is about 450°C. analytical measurements. Analytical Thermal ionization mass spectrometry (TIMS) works at 2500°C instruments are used in research and (RCF are feasible up to 1260°C and with ZrO2 part suitable up to development across many industries as well 1430°C. Therefore they do not cover this temperature range). as government and non-government The substitution principal of RCF in accordance to TRGS 619 agencies (e.g. industry labs, forensic and should be applicated generally. From the technical and environmental research labs, medical economical point of view the bio-soluble wool based on AES research institutions, enforcement agencies would be sufficiently appropriate for the above described and universities). needed temperature range. In addition to AES there are several Therefore the application of RCF in well-established alternatives to RCF on the market which are analytical instruments should not fall under technically feasible and which are currently used in different

- 27 - the scope of the authorization process industries as substitutes for conventional refractory material under REACH. RCF. In addition, investigation on potential substitutes was performed, yet the (technical) characteristics of the potential substitutes are not of equivalent value and the economic efficiency criteria would result in unreasonable economical disadvantage. [See also TRGS (Technical Rule for Hazardous Substances) 619 chapter 3(2) published by Bundesanstalt für Arbeitsschutz und Arbeitsmedizin” (BAuA) Feb 2007]. Tests with several potential alternatives [calcium-magnesium-silicate (CMS) insulation wools, e-glass fibers and mineral wools] showed that none of the candidates could perform up to required specifications in all required areas of concern: • Temperature cycling properties of the material across various cycling periods • Heating/cooling rate stability of the material gathered across a large number of cycles • Ramp temperature speed properties of the material to support instrument sample rates • insulative capacity of the material to protect other sensitive instrument sub- assemblies • formability/machinability properties of the material that allow usage in complex instrument geometries The thermal properties of the insulation material have a critical impact on the instrument performance, none of the substitutes resulted in equivalent instrument performance.

- 28 - In areas where technically feasible (e.g. To our knowledge these statements do not correspond with the sub-assemblies with lower performance latest results in research/ state of technology on the field of RCF requirements like outer tube-insulation) substitution. (For details see response to SELAS-LINDE GmbH, substitution materials are used in the Company, Germany 20091005, above) instruments. An inclusion of RCF on the candidate list This company did not provide concrete calculation of “the with intention on potential future associated costs” for R & D of the new materials. Furthermore authorization including substitution plans RCF has not been used only in analytic devices but also in and sun-set dates for the application of RCF industrial processes. The new developed materials could be in analytical instruments would require the likewise used for different applications and not only in analytical insulation material industries to improve devices, so that investment cost for R&D would be paid-off. some characteristics of the most promising There are also alternatives on the market, which have been candidate materials. This would mean developed by innovative companies already. (For details see significant development and qualification response to SELAS-LINDE GmbH, Company, Germany effort at the material producer as well as at 20091005, above) the instrument manufacturer. The associated cost is unlikely to be economically justifiable for the very small volumes of insulation material going into analytical instruments. Any listing on the candidate list or in Annex An authorization may only be granted if it is shown that socio- XIV (“authorization list”) without intention of economic benefits overweigh the risks to human health or the substitution would increase cost for the environment arising from the use of the substance if there are material manufacturer and the instrument no suitable alternative substances or technologies.” (Art. 60(4)) manufacturer (authorization cost and On the other hand the authorization allows to have controlled administrative handling cost) without exposure to SVHC substances and to get an overview about all changing the environmental or human possible risks to human health or the environment. Moreover the health impact that the material potentially authorization may be reviewed, the elements of the original might have. application should be updated (Art. 61(1)). This mechanism gives incentives for innovations. On this way the risks for the environment or human health will be reduced.

- 29 - As of today, there are appropriate risk High end analytical instruments like gas chromatography management procedures in place. All parts systems are indeed used in research and development- but not containing more than 0.1% wt/wt RCF have exclusively. There does exist a large fraction of other no direct end user exposure (fully integrated applications for these systems, e.g. in quality control units, inside the instrument). The instruments are which makes the R&D exemption under REACH not applicable only operated by professional users. The in this case. parts are serviced by qualified and trained Moreover, the instruments are not in all cases used by personnel only, manuals and repair specialised professionals. Analytical methods must be trained in procedures contain appropriate safe practical courses and it can not be ruled out that trainees or handling instruction according to health students are responsible for the operational availability of such hazards identified and worker safety machines- and in case of malfunction do the repair themselves. standards. Similarly in the manufacturing process of the instrument, only qualified and trained personnel will handle the material according to the appropriate safe handling instructions and worker safety standards. Disposal of the instrument is by trade-in and recycling programs from the manufacturer, which will handle materials according to appropriate safe handling and disposal instructions as well as according to all relevant worker safety and waste disposal standards.

- 30 - Current community regulations mitigating The exemption of uses or categories of uses (article 58(2)) takes the risk of RCF materials in the sense of place if the substance has already been included on the REACH Art 58(2): candidate list. • Synthetic vitreous fibers are covered by an existing EC marketing of hazardous goods directive. Components containing refractory ceramic fibers as described in Commission Directive 97/69/EC of 5 December 1997, adapting to technical progress Council Directive 67/548/EEC relating to the classification, packaging and labeling of dangerous substances. [Hazard is related to size and shape. Mineral fibers of length < 5 micro meter and diameter of fiber < 3 micro meter when (length of fiber)/(diameter of fiber) > 3.] • EC 1272/2008 (CLP) lists RCF materials for harmonized classification and ensures hazard communication and safe handling instructions along the supply chain. • “Gefahrstoffverordnung” and “Technische Regeln fuer Gefahrstoffe – TRGS 619“ as published by BAuA (http://www.baua.de/de/Themen-von-A- Z/Gefahrstoffe/Rechtstexte/Rechtstexte.htm l?__nnn=true) provides substitution guidance. 15 20091013 Unifrax Corp., Company, The German government has prepared two Introductory statement, no response necessary United States of America Annex XV Dossiers proposing that two types of aluminosilicate wools (often termed refractory ceramic fibers [RCF] and so identified in the dossiers) should be listed as substances of very high concern (SVHC) and, therefore, potentially requiring authorization by the European Chemicals Agency (ECHA). I have been involved in research on various aspects of RCF exposure, toxicity, and epidemiology for approximately 20 years

- 31 - and published numerous papers in peer- reviewed journals summarizing these investigations. I have been asked by Unifrax Corp. to prepare and submit comments on these two Annex XV Dossiers. These comments include: (1) comments on Probably this comment refers to the studies by LeMasters the basis of the RCF carcinogen (2003) and Walker (2002). Because of the relatively small and classification; and (2) comments on other young cohorts the findings allow only limited conclusions. analyses offered in the German Dossiers In March 2006 the European suppliers of Aluminosilicate related to RCF exposure and estimated Refractory Ceramic Fibres provided the EU Technical risk. In brief, these comments conclude: Committee on Classification and Labelling of Dangerous Substances with the currently available scientific data with • The available toxicological and respect to Aluminosilicate Refractory Ceramic Fibre epidemiologic data indicate that some carcinogenicity. The aim of the suppliers was to re-discuss concern regarding the potential effects of carcinogenicity. France and Germany had responded to that occupational exposure to RCF is justified. documentation that a re-discussion was not warranted as the However, available epidemiological studies new studies do not have any impact on the existing classification (morbidity and mortality) have failed to as inhalative carcinogen. The Technical Committee on reveal any interstitial fibrosis above Classification and Labelling of Dangerous Substances population background, incremental lung concluded that a re-discussion will only take place in case a cancer, or mesothelioma among workers member state would support such re-discussion. There was no exposed to RCF. support by any member state. Since 2006, no further relevant scientific data have become available. The existing classification What were believed to be state-of-the-art with R49 or H350i is still adequate and does not need to be re- nose-only inhalation bioassays of laboratory evaluated. animals exposed to various types of RCF Experimental data and current classification: the arguments resulted in fibrosis and tumors. But later raised concerning MTD and particle impact have been analysis indicated that these experiments considered (also during the discussions which lead to the legally exceeded the maximum tolerated dose binding classification) and do not have any impact on the (MTD) so that it is not possible to distinguish classification decision. between the relative contribution of particles The dossier does not challenge the outcome of the RCF and fibers to the observed effects. In terms inhalation experiments with respect to a qualitative evaluation of the EU carcinogen classification, a more (i.e. classification for carcinogenicity). The inhalation appropriate classification would be to place experiments are challenged because of the inadequacy to RCF in Category 3(b), equivalent to derive adequate potency information which is crucial for cancer Category 2 (suspected human carcinogen) risk assessment. The latter is the case as there is an in the Global Harmonization System (GHS) unquantifiable impact of the concurrent particle exposure. The system. If RCFs are ultimately so fact that Aluminosilicate Refractory Ceramic Fibres are

- 32 - reclassified, then there is no basis for carcinogenic is not challenged at all on the basis of chronic including these materials among the SVHC. inhalation data.

It had been industry in the past who challenged the outcome of the inhalation experiments during the discussions. It was claimed that the inhalation test with rats is not relevant for humans due to overload. The latter is scientifically not accepted (e.g. from IARC). • The German Dossiers present a risk (1) Only after short-term inhalation exposure the biopersistence analysis based upon the analysis of of Refractory Ceramic Fibres is clearly lower (factor about intraperitoneal (IP) data for RCF and 10) than the biopersistence of amphibole asbestos. This crocidolite asbestos. In my judgment, this difference is much less pronounced (factor about 2) after analysis is fundamentally flawed for reasons intratracheal instillation. However, carcinogenic potency of discussed at length in the body of my fibres is not only determined by biopersistence, but by the comments. The principal conclusion of the so-called 3Ds- paradigm (dose, durability, dimension). analysis in the two dossiers is that RCF has Therefore, it does not make sense to only refer to a risk similar to that of amphibole asbestos, comparative biopersistency (weighted half time after short- which is inconsistent with (1) available term inhalation) as an argument against the RCF risk biopersistence data; (2) results of other assessment described in the Annex XV dossier and to fade RCF risk analyses published in the peer- out dose and dimension. For instance, dimension has be reviewed literature; and (3) results of the included in such a comparison as it similarly influences available epidemiological data. The potency: RCF fibres might be less biopersistent than German Dossiers also use out-of-date amphibole asbestos, but the fibres are more potent as they exposure results, which are updated in are much longer (dimension). As a consequence, it is these comments. plausible, that the potency may be similar. (2) The different RCF risk assessments quoted by industry (DECOS, 1995; Fayerweather et al; 1997; Moolgavkar; 1999; Yu and Oberdörster, 2000; Turim and Brown, 2003) are considered as inadequate. For instance, all of those risk assessments were based on the results of insensitive rat chronic inhalation data. This is an explanation why all of those lead to unrealistically low cancer risks for RCF. In addition, the assessments were partly based on unlikely assumptions e.g. like similar tumour initiation rates between rat and humans (Moolgavkar et al., 1999 which builds upon the Fayerweather et al. approach). Applying the Moolgavkar et al. approach for asbestos would lead to cancer risk estimates for asbestos which are 2-3 orders of magnitude lower than the risk assessment for asbestos by US EPA

- 33 - (1988/2008). (3) IARC (2002) concluded for Refractory Ceramic Fibres that “the limited epidemiological data do not permit an adequate evaluation of the cancer risk associated with exposure to refractory ceramic fibres”. Since then no relevant additional epidemiological data have been published. Hence the available epidemiological data cannot contradict the animal based RCF risk analysis!

Since the more recent information on exposure levels (Maxim, 200829) have not been given in a numerical but in a graphical form they have not been considered in the dossier. Gaining statistical values from the diagrams presented in Maxim (2008) was not considered appropriate due to a considerable loss in precision.

Only exposure data from peer reviewed publications have been taken into account. Due to this quality criteria the Care study (ECFIA, 199930) was the most current publication on measurement data. Since the exposure data submitted by the commentator in the pdf-file (RCF SVHC Comments.pdf) was not published in a journal (in a numerical form) it could not be found in a literature search.

We agree that for the most part, the companies that manufacture RCF in Europe are the same as those in the US since processes and controls are similar. It may therefore be appropriate to utilize all the relevant data for comparison.

Although the exposure data given by the commentator shows some decrease of exposure levels the majority of data points do clearly exceed the tolerance level of 0.1 fibres/ml. This is confirmed by the ladder diagram submitted by the commentator (ladder diagram for combined monitoring data, 2004-2008) and by literature data as well (Maxim, 2008 s. above)

29 Maxim, D., Allshouse, J., Lentz, T.J., Venturin, D., Walters, T.E., Workplace Monitoring of Occupational Exposure to Refractory Ceramic Fiber - A 17 Year Retrospective, Inhal. Toxicol., 20:289-309, 2008 30 ECFIA (1999); Idenfication and control of exposure to refractory ceramic fibres. European Chemical Fibre Industry Association, 3 rue de Colonel Moll, 75017 Paris, France, Nov. 1999, 58p Illus. 6 ref.

- 34 - Maxim et al. (2008 s. above) also investigated the time trends in RCF exposure. It was found that the rate of improvement in fibre concentration in RCF manufacturing plants and customer facilities has slowed down in recent years. A significant decrease of RCF exposure in the near future is anticipated to be not very likely. My complete comments are in the attached The detailed points referred to are found elsewhere in this PDF, and specific comments are detailed document and will be specifically responded where they appear. under the appropriate fields of the online The document referred to was checked. The only comment comment form. related to toxicology which does not appear elsewhere in the present table is the following: It is critisized that it “seems to implicate that cancer incidence and mesothelioma were lumped together – this is questionable as the underlying biological mechanisms are quite different. How can this approach be scientifically justified?”. The response is as follows: In the intra-peritoneal experiments very rarely peritoneal sarcoma were found. As their incidence is very low including these tumours does not have a major impact on the overall outcome of the results. Other tumours in different organs were not included in the evaluation.

- 35 - 15b 20091013 Unifrax Corp., Company, The German government has prepared two This comment is identical to the above one. The response is not Dubl United States of America Annex XV Dossiers proposing that two repeated here, see above. -ette types of aluminosilicate wools (often termed zu refractory ceramic fibers [RCF] and so 15 identified in the dossiers) should be listed as substances of very high concern (SVHC) and, therefore, potentially requiring authorization by the European Chemicals Agency (ECHA). I have been involved in research on various aspects of RCF exposure, toxicity, and epidemiology for approximately 20 years and published numerous papers in peer- reviewed journals summarizing these investigations. I have been asked by Unifrax Corp. to prepare and submit comments on these two Annex XV Dossiers. These comments include: (1) comments on the basis of the RCF carcinogen classification; and (2) comments on other analyses offered in the German Dossiers related to RCF exposure and estimated risk. In brief, these comments conclude:

• The available toxicological and epidemiologic data indicate that some concern regarding the potential effects of occupational exposure to RCF is justified. However, available epidemiological studies (morbidity and mortality) have failed to reveal any interstitial fibrosis above population background, incremental lung cancer, or mesothelioma among workers exposed to RCF. What were believed to be state-of-the-art nose-only inhalation bioassays of laboratory animals exposed to various types of RCF resulted in fibrosis and tumors. But later analysis indicated that

- 36 - these experiments exceeded the maximum tolerated dose (MTD) so that it is not possible to distinguish between the relative contribution of particles and fibers to the observed effects. In terms of the EU carcinogen classification, a more appropriate classification would be to place RCF in Category 3(b), equivalent to Category 2 (suspected human carcinogen) in the Global Harmonization System (GHS) system. If RCFs are ultimately so reclassified, then there is no basis for including these materials among the SVHC. • The German Dossiers present a risk analysis based upon the analysis of intraperitoneal (IP) data for RCF and crocidolite asbestos. In my judgment, this analysis is fundamentally flawed for reasons discussed at length in the body of my comments. The principal conclusion of the analysis in the two dossiers is that RCF has a risk similar to that of amphibole asbestos, which is inconsistent with (1) available biopersistence data; (2) results of other RCF risk analyses published in the peer- reviewed literature; and (3) results of the available epidemiological data. The German Dossiers also use out-of-date exposure results, which are updated in these comments.

My complete comments are in the attached PDF, and specific comments are detailed under the appropriate fields of the online comment form. 15c 20091013 Unifrax Corp., Company, The German government has prepared two This comment is identical to the above one. The response is not Dubl United States of America Annex XV Dossiers proposing that two repeated here, see above. -ette types of aluminosilicate wools (often termed zu refractory ceramic fibers [RCF] and so

- 37 - 15 identified in the dossiers) should be listed as substances of very high concern (SVHC) and, therefore, potentially requiring authorization by the European Chemicals Agency (ECHA).

I have been involved in research on various aspects of RCF exposure, toxicity, and epidemiology for approximately 20 years and published numerous papers in peer- reviewed journals summarizing these investigations. I have been asked by Unifrax Corp. to prepare and submit comments on these two Annex XV Dossiers. These comments include: (1) comments on the basis of the RCF carcinogen classification; and (2) comments on other analyses offered in the German Dossiers related to RCF exposure and estimated risk. In brief, these comments conclude:

- 38 - contribution of particles and fibers to the observed effects. In terms of the EU carcinogen classification, a more appropriate classification would be to place RCF in Category 3(b), equivalent to Category 2 (suspected human carcinogen) in the Global Harmonization System (GHS) system. If RCFs are ultimately so reclassified, then there is no basis for including these materials among the SVHC.

• The German Dossiers present a risk analysis based upon the analysis of intraperitoneal (IP) data for RCF and crocidolite asbestos. In my judgment, this analysis is fundamentally flawed for reasons discussed at length in the body of my comments. The principal conclusion of the analysis in the two dossiers is that RCF has a risk similar to that of amphibole asbestos, which is inconsistent with (1) available biopersistence data; (2) results of other RCF risk analyses published in the peer- reviewed literature; and (3) results of the available epidemiological data. The German Dossiers also use out-of-date exposure results, which are updated in these comments.

My complete comments are in the attached PDF, and specific comments are detailed under the appropriate fields of the online comment form. 15d- 20091013 Unifrax Corp., Company, The German government has prepared two This comment is identical to the above one. The response is not Dubl United States of America Annex XV Dossiers proposing that two repeated here, see above. ette types of aluminosilicate wools (often termed zu refractory ceramic fibers [RCF] and so 15 identified in the dossiers) should be listed as substances of very high concern (SVHC)

- 39 - and, therefore, potentially requiring authorization by the European Chemicals Agency (ECHA).

These comments include: (1) comments on the basis of the RCF carcinogen classification; and (2) comments on other analyses offered in the German Dossiers related to RCF exposure and estimated risk. In brief, these comments conclude:

- 40 - observed effects. In terms of the EU carcinogen classification, a more appropriate classification would be to place RCF in Category 3(b), equivalent to Category 2 (suspected human carcinogen) in the Global Harmonization System (GHS) system. If RCFs are ultimately so reclassified, then there is no basis for including these materials among the SVHC. • The German Dossiers present a risk analysis based upon the analysis of intraperitoneal (IP) data for RCF and crocidolite asbestos. In my judgment, this analysis is fundamentally flawed for reasons discussed at length in the body of my comments. The principal conclusion of the analysis in the two dossiers is that RCF has a risk similar to that of amphibole asbestos, which is inconsistent with (1) available biopersistence data; (2) results of other RCF risk analyses published in the peer- reviewed literature; and (3) results of the available epidemiological data. The German Dossiers also use out-of-date exposure results, which are updated in these comments.

My complete comments are in the attached PDF, and specific comments are detailed under the appropriate fields of the online comment form. 16 20091013 Unifrax I LLC, Company, This issue is addressed in more detail on Statement, no response necessary. United States of America pages 1-4 of the attached PDF file titled “Unifrax comments”. Because this comment form does not allow for efficient submission of graphics, tables, footnotes and references, the reader is referred to the attached PDF file for a more complete argument against labeling RCF as a

- 41 - substance of very high concern. An excerpt of relevant argument follows (refer to the attached PDF for tables, figures and references): Unifrax comments on the German Annex XV Dossier on Aluminosilicate Refractory Ceramic Fibres

Introduction The German government has prepared Annex XV Dossiers (hereinafter “German Dossiers”) proposing two synthetic vitreous fibers as candidate Substances of Very High Concern (SVHC). The fibers included in these dossiers are (1) Aluminosilicate Refractory Ceramic Fibres and (2) Zirconia Aluminosilicate Refractory Ceramic Fibres.

Unifrax Unifrax Corporation is a major manufacturer of high temperature insulating wools (HTIW)—including RCF—with operations in Asia, Europe, and North and South America. Unifrax believes that the German proposal to include these two fibers is not justified by the available scientific evidence and is flawed in several other respects. We appreciate the opportunity to offer comments to the European Chemicals Agency (ECHA) on these proposals.

- 42 - Short summary of comments In March 2006 the European suppliers of Aluminosilicate In brief, Unifrax believes that: Refractory Ceramic Fibres provided the EU Technical Committee on Classification and Labelling of Dangerous • These two fibers should not be included Substances with the currently available scientific data with on the list of SVHC because the present respect to Aluminosilicate Refractory Ceramic Fibre carcinogen classifications are not carcinogenicity. The aim of the suppliers was to re-discuss appropriate for these two substances. carcinogenicity. France and Germany had responded to that Rather than being classified (as at present) documentation that a re-discussion was not warranted as the into Category 2 (substances which should new studies do not have any impact on the existing classification be regarded as if they are carcinogenic to as inhalative carcinogen. The Technical Committee on man ), RCF should be in listed as a Classification and Labelling of Dangerous Substances Category 3 substance (substances which concluded that a re-discussion will only take place in case a cause concern for man, owing to possible member state would support such re-discussion. There was no carcinogenic effects ). Under the new support by any member state. Since 2006, no further relevant Global Harmonization System being scientific data have become available. The existing classification enacted by the European Union (EC No. with R49 or H350i is still adequate and does not need to be re- 1272/2008 ), RCF should be grouped into evaluated. Category 2, a suspected human carcinogen, , which replaces (but is otherwise equivalent to) the EC Category 3. Classification into Category 2 does not justify placing a substance on the list of SVHC. Unifrax and other HTIW producers plan to petition the competent authorities to reclassify RCF and urges that ECHA defers consideration of the draft Annex XV Dossier until this petition is acted upon. • The German Dossiers gloss over results of The carcinogen classification in category 2 – is based on epidemiological studies conducted in both “appropriate long term animal studies (and) other relevant Europe and North America. Results of information”. The results of current epidemiological studies don’t these studies indicate that occupationally allow the exclusion of the cancer risk (see above). This means exposed cohorts experience symptoms not no dissent to the mentioned findings. unlike those experienced by other dust- exposed populations and that there were statistically, but not clinically significant, decrements in pulmonary function, which have not become worse with additional occupational exposure. Results of these studies in North America indicate that

- 43 - pleural plaques are correlated with exposure duration and cumulative RCF exposure. But these plaques are widely regarded as a marker of exposure rather than a disease endpoint. And perhaps most important, these epidemiological studies have failed to detect exposure related interstitial fibrosis, incremental lung cancer, or any mesotheliomas. • The risk estimation procedure employed in Industry grossly exaggerates the uncertainty of the risk analysis, the German Dossiers is conceptually flawed which is mainly caused by the uncertainty of the asbestos risk and, in any event, the resulting risk analysis. But as a “central estimate” for asbestos is used for estimates have very large (several orders of comparison with RCF, it is unlikely that the RCF risk is magnitude) uncertainty. This in overstated. combination with the use of outdated exposure information is likely to overstate Since the more recent information on exposure levels (Maxim, risks and, therefore, should not be 200831) have not been given in a numerical but in a graphical considered in any subsequent decision form they have not been considered in the dossier. Gaining regarding the need for authorization. statistical values from the diagrams presented in Maxim (2008) was not considered appropriate due to a considerable loss in precision.

Only exposure data from peer reviewed publications have been taken into account. Due to this quality criteria the Care study (ECFIA, 199932) was the most current publication on measurement data. Since the exposure data submitted by the commentator in the pdf-file (RCF SVHC Comments.pdf) was not published in a journal (in a numerical form) it could not be found in a literature search.

Although the exposure data given by the commentator shows

31 Maxim, D., Allshouse, J., Lentz, T.J., Venturin, D., Walters, T.E., Workplace Monitoring of Occupational Exposure to Refractory Ceramic Fiber - A 17 Year Retrospective, Inhal. Toxicol., 20:289-309, 2008 32 ECFIA (1999); Idenfication and control of exposure to refractory ceramic fibres. European Chemical Fibre Industry Association, 3 rue de Colonel Moll, 75017 Paris, France, Nov. 1999, 58p Illus. 6 ref.

- 44 - some decrease of exposure levels the majority of data points do clearly exceed the tolerance level of 0.1 fibres/ml. This is confirmed by the ladder diagram submitted by the commentator (ladder diagram for combined monitoring data, 2004-2008) and by literature data as well (Maxim, 2008 s. above)

Maxim et al. (2008, s. above) also investigated the time trends in RCF exposure. It was found that the rate of improvement in fibre concentration in RCF manufacturing plants and customer facilities has slowed down in recent years. A significant decrease of RCF exposure in the near future is anticipated to be not very likely. • The German Dossiers claim, without any See response to SELAS-LINDE GmbH, Company, Germany evidence, that safe substitutes for RCF are 20091005, above now available and that (p21) “current product developments indicate that the upper temperature limit of [alkaline earth silicate] AES wool products could be increased significantly.” Unifrax has been in the forefront of development of AES wools and, indeed, AES can be substituted for RCF in some applications. However, we dispute the claim that the upper temperature limit of AES wools can be increased significantly. Unifrax has attempted to structure Statement, no response necessary. comments to fit within the structure specified by ECHA. In addition to the “topic by topic” entries in the designated areas of the electronic form, we are also submitting this consolidated copy of our comments (with appendices), grouped insofar as possible under the headings suggested on the ECHA web site.

- 45 - -“New” data In March 2006 the European suppliers of Aluminosilicate The “Working Procedures” document notes Refractory Ceramic Fibres provided the EU Technical at several places that “New” scientific Committee on Classification and Labelling of Dangerous information is required to overturn a Substances with the currently available scientific data with nomination. For example (page 5), this respect to Aluminosilicate Refractory Ceramic Fibre document notes “The MSC-S would carcinogenicity. The aim of the suppliers was to re-discuss normally make a decision to address the carcinogenicity. France and Germany had responded to that proposal [to revise or delete a nomination] documentation that a re-discussion was not warranted as the in a meeting in the following new studies do not have any impact on the existing classification circumstances,” which include “Valid new as inhalative carcinogen. The Technical Committee on scientific information is provided challenging Classification and Labelling of Dangerous Substances the proposal.” “New” is not defined concluded that a re-discussion will only take place in case a elsewhere in the “Working Procedures” member state would support such re-discussion. There was no document. However, Unifrax believes that support by any member state. Since 2006, no further relevant a reasonable interpretation of the word scientific data have become available. The existing classification “new” in this context would be data and with R49 or H350i is still adequate and does not need to be re- analyses published after the date when evaluated. RCF was classified into Category 2, which was the 23rd Adaptations to Technical Progress (ATP) of Annex 1 of Directive 67/548/EEC [No. L343, 13, 12, 1997, dated 5 December 1997]. Subsequent to this date (5 December 1997) a large number of potentially relevant documents have been published. Appendix A to these consolidated comments contains a partial list of “new” (post 5 Dec. 1997) scientific publications and, therefore, appropriate to include in any discussion of the basis for carcinogen classification of RCF. This includes several articles that provide estimates of the weighted half-time of RCF from inhalation experiments, articles that address the differences between RCF1 and RCF 1a, review articles that examine overload and its consequences (particularly Brown et al., 2005), and several papers describing results of the epidemiological studies (including the longitudinal

- 46 - pulmonary function test (PFT) results and a mortality study on an occupationally exposed cohort). Other new studies include nearly all the key risk analyses and updated exposure studies as well as a comprehensive review published by the (US) National Institute for Occupational Health and Safety (NIOSH, 2006). 17 20091013 Individual (affiliated with This issue is addressed in more detail on Introductory statement, no response necessary, specific University of Rochester page 1 of the attached PDF file. Because comments from the attachment are documented below and Medical Center, Academic this comment form does not allow for responses have been added there. institution), United States of efficient submission of graphics, tables, America footnotes and references, the reader is referred to the attached PDF file for a more complete argument against labeling RCF as a substance of very high concern. An excerpt of relevant argument follows (refer to the attached PDF for tables, figures and references):

Comments on the German Dossiers Proposing Refractory Ceramic Fiber as a Substance of Very High Concern (SVHC) By Mark J. Utell, MD Professor of Medicine and Environmental Medicine University of Rochester Medical Center Rochester, NY, USA

Background Over the years, Refractory Ceramic Fiber Coalition (RCFC) has provided comprehensive information on Refractory Ceramic Fiber (RCF) to the German Regulatory Agencies and its staff containing relevant background information for setting exposure limits for RCF. The German

- 47 - government has prepared two Annex XV Dossiers proposing that two types of aluminosilicate wools (often termed [RCF] and so identified in the dossiers) should be listed as substances of very high concern (SVHC) and, therefore, potentially requiring authorization by the European Chemicals Agency (ECHA). I have been involved in research on various aspects of RCF exposure, toxicity, and epidemiology for approximately 20 years and published numerous papers in peer- reviewed journals summarizing these investigations. I have been asked by Unifrax Corp to prepare and submit comments on these two Annex XV Dossiers. These comments include; (1) comments on ongoing epidemiological studies, and (2) comments on clinical effects related to RCF exposure. I present summary material relevant to these issues in what follows. These summaries are brief, but fully supported by numerous reports and data published in the peer-reviewed literature 18 20091014 German Refractory Contents Introductory statement, no response necessary. Association, Industry or trade association, Germany General Comments 1 1. Number of Annex XV dossiers 1 2. Priority Substances 1 Specific Comments 2 1. Current classification of RCF 2 2. Alternative Substances 2 3. Environmental Protection 2 Conclusions 2

- 48 - 1. Number of Annex XV dossiers The „Aluminosilicate Refractory Ceramic There is need for two dossiers in the sense of REACH. Fibres“ and the “Zirconia Aluminosilicate According to the guidance for identification and naming of Refractory Fibres” are covered by the substances under REACH UVCB substances are specified with entries under the index number 650-017-00- the IUPAC-name of their constituents. In the case of 8 as Refractory Ceramic Fibres in Annex VI, Aluminosilicate RCF the main constituents are Al2O3 and SiO2 part 3, table 3.1 (list of harmonised which both are present more than 10 %. In the case of Zirconia classification and labelling of hazardous Aluminosilicate RCF the main constituents are Al2O3, SiO2 and substances) as well as in table 3.2 (list of ZrO2 which all are present more than 10 % in the UVCB harmonised classification and labelling of substance hazardous substances from Annex I to Directive 67/548/EEC) of Regulation No 1272/2008. Therefore it is not necessary to separate the two substances by submitting two identical Annex XV dossiers. The German Refractory Association does not agree with this approach. 2. Priority Substances The intention of article 58(3) is to give arguments for a According to Article 58 (3) of the REACH- priorisation of substances for their inclusion in Annex XIV. At the Regulation priority shall normally be given moment, the German CA is interested in the 1ststep, inclusion on to substances with: (a) PBT or vPvB the candidate list. Article 58 (3) states: “Priority shall normally be properties; or (b) wide dispersive use; or (c) given to substances with: high volumes. (a) PBT or vPvB properties; or RCF have no PBT and vPvB properties, the (b) wide dispersive use; or uses are well known and very specific. RCF (c) high volumes.” have no wide dispersive use, it is only used We agree that RCFs are neither PBT nor vPvB. Nevertheless, for industrial purposes. In accordance to the the total tonnage manufactured in Europe is not insignificant. In Directive 2001/41/EC it can only be sold to addition, the authors of the annex XV dossier have given some commercial users. The main use (67%) is data concerning workers exposure. In different fields of usage of the application for furnace linings (Annex XI the RCFs. Therefore, some kind of widespread use has been dossier Table 5, Page 15). Protection of identified although the consumer exposure is at least low. workers is regulated in the EU through different legislations and controlled by industry and regulatory bodies. The general public is not exposed to fibrous dust from RCF. This was proved by studies where measurements were made in the general public [Schneider et al. (1996)

- 49 - “Ubiquitous fibre exposure in selected sampling sites in Europe”]. The production volumes of all RCF reduces in 2008 to 25 000 t/a (for both ASW and AZS). The dangerous potential of the RCF belongs only to the dust of the RCF. The exposure to fibrous dust from the products is very low compared with the tonnage mentioned in the Dossier. An exposure could be seen only during handling the products. In use there is no or negligible exposure. Therefore RCF is no priority substance for the REACH authorisation process at all.

19 20091014 Individual, United Kingdom p 11-p27 The information supplied from p11 Statement, no response necessary. onwards and in particular the conclusions drawn on p 27 show clearly the similarities to crocidolite.This is an opportunity to avoid the didaster that arose from the delays in taking action on asbestos in nthe last century.

(Information on risks related to the substance) 20 20091014 PRE - Federation of PRE represents the European Refractory Introductory statement, no response necessary, specific European Refractory Producers covering more than 21.000 comments from the attachment are documented below and Producers, Industry or trade employees generating nearly 3,4 billion responses have been added there association, Belgium Euro. based on (i) the insufficient quality of the Annex XV reports, (ii) the scientifically founded doubts on the current classification of RCF and (iii) the non-substitutable and exclusively professional uses of RCF, PRE requests that both Annex XV reports on RCF are withdrawn and RCF are not placed on the REACH candidate list. Further details can be found in the specific comments on the justification and on use, exposure, alternatives and risks.

- 50 - 21 20091014 Individual (affiliated with This issue is addressed in more detail on Introductory statement, no response necessary. University of Rochester, page 1 of the attached PDF file. Because Academic institution), this comment form does not allow for United States of America efficient submission of graphics, tables, footnotes and references, the reader is referred to the attached PDF file for a more complete argument against labeling RCF as a substance of very high concern. An excerpt of relevant argument follows (refer to the attached PDF for tables, figures and references):

Introduction The German government has proposed two synthetic vitreous fibers (i) Aluminosilicate Refractory Ceramic Fibres and (ii) Zirconia Aluminosilicate Refractory Ceramic Fibres (RCF) as candidate Substances of Very High Concern (SVHC) for authorization by the European Chemicals Agency (ECHA) as part of the REACH process. Two nearly identical dossiers have been prepared to accompany these proposals. In both dossiers studies using intraperitoneal (IP) administration of fibers in rats are used to derive carcinogenic risk estimates in humans. Other methods for risk estimation based on the analysis of chronic inhalation (IH) studies are either omitted or only briefly mentioned in the German Dossiers. The method used in the German Dossiers An extensive justification has been recently provided to justify makes several assumptions, among these the fact that chronic inhalations studies in rats are insensitive to the suitability of the IP method for detect fibre carcinogenicity (Wardenbach et al. 200533). Having developing risk estimates. Given the in mind the higher sensitivity of humans compared to rats after availability of IH data and published inhalation of asbestos, the arguments provided in favour of the analyses of these data by several inhalation studies are highly questionable. investigators, the use of IP data alone

33 Wardenbach P, Rödelsperger K, Roller M, Muhle H. Classification of man-made vitreous fibers: Comments on the revaluation by an IARC working group. Regul Toxicol Pharmacol. 2005 Nov;43(2):181-193.

- 51 - should be questioned, particularly as data The dossier does not challenge the outcome of the RCF from a chronic nose-only inhalation inhalation experiments with respect to a qualitative evaluation experiment are available. (i.e. classification for carcinogenicity). The inhalation experiments are challenged because of the inadequacy to derive adequate potency information which is crucial for cancer risk assessment. The latter is the case as there is an unquantifiable impact of the concurrent particle exposure. The fact that Aluminosilicate Refractory Ceramic Fibres are carcinogenic is not challenged at all on the basis of chronic inhalation and IP data.

- 52 - Alliance, International NGO, substance to the candidate list on the basis Belgium of the information on page 5 of the submitted Annex XV dossier. 24 20091014 Health & Safety Authority, The Irish Competent Authority agrees with No response necessary Member State Competent the identification of aluminosilicate, Authority, Ireland refractory ceramic fibres as a substance meeting the criteria set out in Article 57 of REACH. 25 20091014 WWF European Policy WWF supports the inclusion of this No response necessary Office, International NGO, substance in the candidate list according to Belgium REACH article 57.a) 26 20091014 Individual (affiliated with Introduction The authors of the Annex XV dossier reflected on the data given Harvard School of Public German regulatory authorities have in Table 13 of the dossier. Based on data from ip tests with Health, Academic proposed that two aluminosilicate wools different fibres BMD10 values were calculated and compared. institution), United States of (also termed refractory ceramic fiber [RCF]) This comparison has been normalized to crocidolite. As no America be listed as substances of very high reference is given, no comparison to the mortality data concern (SVHC) and has prepared two mentioned can be performed. dossiers in support of this nomination. Among the assertions made in these dossiers is that RCFs are potent carcinogens—as potent as amphibole (crocidolite) asbestos. The available mortality data do not support this position. In fact, my analysis shows that the available data are statistically incompatible with the hypothesis that RCF is as potent as amphibole asbestos in causing lung cancer. Moreover, there are no cases of mesothelioma in the occupationally exposed cohort. 27 20091015 Individual, Japan We think each SVHC should be identified The scientific literature identifies different RCFs by RCF 1 to as specific substance. It is universally RCF 4 (e.g.IARC, vol 81, 2002). Actually only the acknowledged that Refractory Ceramic Aluminosilicate Refractory Ceramic Fibres of the type 1 -4 (as Fibres are widely used as an alternative described in the Annex XV report) are on the market. For these substance of asbestos. Once the words of fibres no CAS numbers are actually available. It is in Industry’s "Aluminosilicate Refractory Ceramic Fibres" responsibility to apply a CAS number and not in Member States. are specified as SVHC, we feel it may cause confusion to general public and many industries, because the words of

- 53 - "Aluminosilicate Refractory Ceramic Fibres" implies several substances and also definition of "Aluminosilicate Refractory Ceramic Fibres" is indistinct. Therefore, we recommend member state to declare "CAS number" or "specific name" of substance to specify some of "Aluminosilicate Refractory Ceramic Fibres" as SVHC. 28 20091015 RIVM - Bureau REACH, Regarding substance name We think that the substance name is specific enough for all National Authority, The The identity of this substance is determined Aluminosilicate Refractory Ceramic Fibres which are Netherlands by the substance name as there are no commercially on the market. CAS and EC numbers. Is this substance name specific enough so that the requirements for SVHC substances apply to all intended substances and only to the intended substances? For example, would a different substance name result in avoiding the obligations of a SVHC substance? Regarding environment In our opinion, the exposure at work places exceeds acceptable Dossier focuses on carcinogenic properties or tolerable limits. Although this concept has yet not been of the compound, environmental issues are agreed on in the EU the authors identified risks at work places. not discussed. Carcinogenic risks related to Whether or not this is true for an exposure through the inhalation, suggesting that the compound environment seems to be not clear at the moment. Data on can be released to the environment through airborne fibre concentrations in the environment are lacking but air (especially during service life of the end values are expected to be far below those at work places. product) are not addressed. One of the uses specified in the dossier is in brake pads (p. 8), although on p. 20 it is stated that new brake pads no longer contain fibres classified as carcinogens in Category 2 according to the manufacturers. It is known that the physical appearance of inorganic ceramic fibres can be altered significantly, e.g. due to abrasion of brake pads. If this use is applicable an environmental exposure scenario should be included to take the asbestos mode-of- action for the public at large into account. This possibility of inhalatory exposure is not

- 54 - discussed in the document and thus environmental risks can not be assessed. In addition, no background information can be retrieved. 29 20091015 Member State Competent We agree that these fibres meet the SVHC The complete Annex I entry under Index No 650-017-00-8 refers Authority, United Kingdom criteria, however, we have some concerns to ‘Refractory Ceramic Fibres, Special Purpose Fibres … [Man- that the fibres that seem to be covered in made vitreous (silicate) fibres with random orientation with these two dossiers do not encompass the alkaline oxide and alkali earth oxide (Na2O+K2O+CaO+ same range of fibres that are within scope MgO+BaO) content less or equal to 18 % by weight]’. The of the definition in CLP Annex VI for Aluminosilicate RCFs and the Zirconia Aluminosilicate RCF are Refractory Ceramic Fibres and Special actually the only commercially existing fibres which match to this Purpose Fibres (see below). We also have entry. Other fibres (i.e. alkaline earth silicate wool or high- some concerns that the identity of the alumina, low-silica-wool) do not fulfil this definition while the materials described in section 1 does not (Na2O+K2O+CaO+ MgO+BaO) content is higher than 18 % by correspond to the materials as supplied and weight (see IARC Vol 18). that the composition information relates to If there will be other RCF these do not fall in the same field of the components from which the fibres are application as the Aluminosilicate RCF. manufactured rather than the fibres themselves. Providing that these issues can be satisfactorily resolved we support the inclusion of these substances on the Candidate List. We suggest that a more detailed consideration of possible risk management options would help decisions on further regulatory action. 30 20091015 WECF, International NGO, We support the nomination of this chemical No response necessary The Netherlands to the Candidate List, and believe it is important, given its properties, for it to be as strictly controlled as possible

- 55 - 31 20091015 CECOF, Industry or trade CECOF-Comments 15 October 2009 This statement about failed practical experiences with association, Germany alternative products is without any evidence and does not ECHA consultation about Substances of correspond with our knowledge about current use of the Very High Concern - ANNEX XV: available alternatives on the market. (For details see response Aluminosilicate Refractory Ceramic Fibres to SELAS-LINDE GmbH, Company, Germany 20091005, (RCF) above)

Dear Sirs,

CECOF is the European Committee for Industrial Furnace and Heating equipment associations. CECOF incorporates the relevant national associations of industrial furnace and heating equipment in Europe and as such all major companies in this field. Member companies of national CECOF associations produce furnaces and apparatus used in high temperature applications for the heat treatment of products made of steel, non ferrous metals, ceramics, porcelain, glass etc. Because of the high temperatures (>600°C) the furnaces have to be lined with refractory materials. Where in most cases traditional heavy materials and insulating firebricks are used, in many applications there is a need for light weight insulating materials made of high temperature insulation wool (HTIW). HTIW is used for industrial furnaces which, due to their operating mode, have to be heated up and cooled down very often. Due to the low specific heat capacity, the necessary amount of energy can be reduced considerably in comparison to the traditional heavy refractory lining. It is only with HTIW that some procedures can be applied and that specific innovative and new products can be produced. The European industry of furnace

- 56 - manufactures needs this material to achieve the high level set by the European Commission for energy saving and the reduction of greenhouse gas emissions until 2020. More than 30% of energy saving and respectively CO2-reduction can be achieved in industrial high temperature application by using HTIW. Practical experience in the last few years has shown that alternative products on the market are not applicable for use in the high-temperature range.

Substitution is already regulated in EU (67/548 EC), and the German Ministry of Labour and Social Affairs had published the TRGS 619 dealing specifically with the substitution for products made of alumino silicate wool (ASW/RCF). In practice, there is no viable alternative possible in many of the high temperature industrial applications; therefore this TRGS was created to document the arguments when there is no alternative material available on the grounds of health and safety, environmental and economical reasons. The comparison between crocidolyte and We agree that the wording was imprecise. Meant is the RCF is scientifically wrong: “... amphibole following: asbestos (i.e. crocidolyte), which does not In contrast to serpentine asbestos, refractory ceramic fibres tend cleave along the fibre axis ...”. This to break transversely rather than cleaving along the fibre axis. statement is not correct from a mineralogy The behaviour to cleave along the fibre axis is associated with point of view and is challenged by the fact that numerous new fibres generate intraperitoneally international experts on mineralogy that which may increase dose and have an impact on the test CECOF consulted (Universities of Aachen outcome. The tendency to cleave along the fibre axis is different and Freiberg). for different types of asbestos. Among the different types of asbestos, it is lowest for crocidolite. Thus, only results from ip tests with crocidolite were used to assess the comparative carcinogenic potency of asbestos and refractory ceramic fibres.

- 57 - Because of the listing on the ECHA website The authorisation procedure under REACh is laid down in the there is already an enormous confusion in legal text and therefore not questionable. On the one hand the the market, which could cause commercial jobs might be reduced in production of RCF, but on the other damage for the whole manufacturer and hand the number of employees is expected to increase in user industry. production of the alternatives. In the downstream processes like Especially when end users of furnaces have installation, maintenance etc. the jobs would not be reduced, problems to make their decision in choosing because it does not make a difference if there will RCF be used the right insulating products for their or other insulating material. The demand /need on insulating applications they are massively disturbed. materials will remain the same and it determines the production The German authorities (BMU) withdrew the volume. There are other different factors which determine EU-Notification 2004/370/D in 2006 after competitiveness in the European user industry, not only RCF intervention from the EU-Commission insulating material. Furthermore there are many alternatives (message 009 – communication from the available on the market. The submitted Annex XV Dossier on Commission – SG(2006) D/50654). In the RCF does not focus on removal of insulating materials for notification 2004 ff it was used nearly the industrial uses as whole. We support the approach to use same justification like in the published insulating materials with the goal to reduce energy consumption dossiers now. In the dossier the impression and CO2-emissions, but these should be also chosen with is given that there was an agreement on the respect to human health. former notification which is not the case. As soon as the dossiers for zirkonia and aluminosilicate RCF substances were published in Annex XV these materials were stigmatized. This was noticed when users of furnaces for industrial high temperature applications did not allow the manufacturers of these equipment to use RCF material at all. This reaction from the industrial end- user can already be seen in the marketplace. Even if a socio-economic study should reveal that RCF had wrongly been put on the Annex XV, this material will no longer be usable for emotional reasons. There will be no means to repair the economic damage following the publication in Annex XV. In order to prevent further negative implications for the European industry (furnace manufacturers and users) and the related environmental impact, we therefore

- 58 - ask strongly to take the substance RCF out of the focus and to delete it from the ECHA website. Nevertheless further discussions should be continued among all involved stakeholders and regulatory bodies in a professional way based on science and practical experience in order to achieve the aims of REACH and the targets from the European climate change programme (ECCP). REACH and ECCP should not be conflictive as it is already now after the publication of RCF on the ECHA website took place. CECOF therefore requests to reject the Annex XV dossiers for the two substances “Alumino Refractory Ceramic Fibres” and “Zirconia Aluminosilicate Refractory Ceramic Fibres” until a decision is made after a comprehensive survey of all involved parties.

Secretary General 32 20091015 ArcelorMittal, Company, p21, §2.1: It is said in conclusion that there Due to our knowledge alternatives are on the market. (For Luxembourg are several possible substitutes for details see response to SELAS-LINDE GmbH, Company, aluminium silicate wool products on the Germany 20091005, above) market. Unfortunately, we cannot support this conclusion. In the steel industry, those RCF's are widely found. For high temperature applications (>900 °C) found in the steel industry processes, no (above 1100°C) or only bad substitutes exists. Where substitutes exists, they have major impacts on workability (productivity, working conditions (especially for maintenance and removal operations of substitutes due to dust formation and practical difficulties of application) and required investments to implement them resulting from lower

- 59 - performance (relining, structural reinforcements of lines and ovens due to overweight and higher space occupancy). Key processes of the steel making industry will have to face very serious challenges to be adapted, with no adequate substitutes known for the highest temperatures applications. We urge authorities to carefully examine these uses and foresee appropriate exemptions in the Annex XIV dossier, should this substance be placed under Authorisation process.

(Information on risks related to alternatives)

- 60 - Specific comments on the justification No. Date Submitted by (name, Comment Response Organisation/MSCA) 33 20090928 Air Liquide, Company, Tests on AES fibres, considered as Up to our knowledge, it is unquestionable that cristobalite is France potential substitute for RCF, under real formed from AES as well as from RCF (Class, 200234) in a operating conditions, have shown a full recristallization procedure above 900 °C. In both insulating crystallization into cristobalite (crystalline materials cristobalite only occurs after its using in the high- silica). Potentially carcinogenic materials temperature range. For RCFs Class35 (2003) has shown that (RCF) have been substituted by cristobalite exposure in the course of removal operations can carcinogenic materials (crystallized AES = exceed the limit value used in a number of countries (0.05 crystalline silica), as classified by the mg/m3). The average concentration as presented in the quoted International Agency for Research on literature is about 0.096 mg/m3. The proposed OSHA limit value 3 36 Cancer. Moreover this evolution induces for crystalline SiO2 is 0.025 mg/m (ICSC 0808 ). According to great difficulties to measure levels of a literature search at present no published data regarding the exposure of the dangerous substances in exposure level of cristobalite from AES fibres is available. the final product. The difference between AES and RCF wool products is obvious: AES wool is not classified as carcinogenic, but RCF products To our knowledge no validated substitute are. AES wool is safe in applications like installation, production for RCF exists for Steam Methane of downstream articles etc. as long as it is not used above 900 Reforming application. In such process °C. fibres are submitted to high temperature (900 – 1200°C), reducing atmosphere, and Besides the opportunity to use AES products as an alternative to very long duration (several years). RCFs there are other potent substitutes available on the market The only substitutes are bricks and AES which lack the problem of cristobalite formation. (For details see fibres but both lead to crystalline silica response to SELAS-LINDE GmbH, Company, Germany exposure 20091005, above)

(Classification and labelling) 34 20091005 Lurgi GmbH, Company, Tests on AES fibres, considered as Up to our knowledge, it is unquestionable that cristobalite is Germany potential substitute for RCF, under real formed from AES as well as from RCF (Class, 200237) in a operating conditions, have shown a full recristallization procedure above 900 °C. In both insulating

34 P. Class, R.C. Brown; Exposition gegenüber künstlichen Mineralfasern, Gefahrstoffe – Reinhaltung der Luft; 62, Nr. 5, 2002 35 Ph. Class; Current Fibrous Dust Workplace Concentrations and Trend in the High Temperature Insulation Wool Industry: the Results of the Care programme, VDI-Bericht Nr. 1776, 2003 36 ICSC: International Chemical Safety Card No. 0808, Crystalline silica, quartz Crystalline silicon dioxide, quartz Silicic anhydride SiO2 , Molecular mass: 60.1 37 P. Class, R.C. Brown;Exposition gegenüber künstlichen Mineralfasern, Gefahrstoffe – Reinhaltung der Luft; 62, Nr. 5, 2002

- 61 - crystallization into cristobalite (crystalline materials cristobalite only occurs after its using in the high- silica). Potentially carcinogenic materials temperature range. For RCFs Class38 (2003) has shown that (RCF) have been substituted by cristobalite exposure in the course of removal operations can carcinogenic materials (crystallized AES = exceed the limit value used in a number of countries (0.05 crystalline silica), as classified by the mg/m3). The average concentration as presented in the quoted International Agency for Research on literature is about 0.096 mg/m3. The proposed OSHA limit value 3 39 Cancer. for crystalline SiO2 is 0.025 mg/m (ICSC 0808 ). According to Moreover this evolution induces great a literature search at present no published data regarding the difficulties to measure levels of exposure of exposure level of cristobalite from AES fibres is available. the dangerous substances in the final product. The difference between AES and RCF wool products is obvious: To our knowledge no validated substitute AES wool is not classified as carcinogenic, but RCF products for RCF exists for Steam Methane are. AES wool is safe in applications like installation, production Reforming application. In such process of downstream articles etc. as long as it is not used above 900 fibres are submitted to high temperature °C. (900 – 1200°C), reducing atmosphere, and very long duration (several years). Besides the opportunity to use AES products as an alternative to The only substitutes are bricks and AES RCFs there are other potent substitutes available on the market fibres but both lead to crystalline silica which lack the problem of cristobalite formation. (For details see exposure. response to SELAS-LINDE GmbH, Company, Germany It is proposed: 20091005, above) • The classification of AES fibres exposed to high temperature (900 – 1200 °C) for long duration (several months) as products containing crystalline silica.

• The clarification of the methodology used to measure exposure levels to fibres by considering both crystalline AES fibres and RCF fibres. Another way should be to consider fibres as dust and impose a common Permissible Exposure Limit for classified products.

• To acknowledge that there is no substitute available for RCF in some petrochemical

38 Ph. Class; Current Fibrous Dust Workplace Concentrations and Trend in the High Temperature Insulation Wool Industry: the Results of the Care programme, VDI-Bericht Nr. 1776, 2003 39 ICSC: International Chemical Safety Card No. 0808, Crystalline silica, quartz Crystalline silicon dioxide, quartz Silicic anhydride SiO2 , Molecular mass: 60.1

- 62 - industries such as Steam Methane Reforming where fibres are exposed to high temperature (900 - 1200°C), reducing atmosphere for long duration (several months).

(Human health hazard assessment of physiochemical properties) 35 20091006 Refractory Ceramic Fibers Animal & Human Studies In March 2006 the European suppliers of Aluminosilicate Coalition (RCFC), Industry The results of comprehensive chronic Refractory Ceramic Fibres provided the EU Technical or trade association, United inhalation studies conducted at the Committee on Classification and Labelling of Dangerous States of America Research and Consulting Company (RCC) Substances with the currently available scientific data with in Geneva were the primary basis for 1997 respect to Aluminosilicate Refractory Ceramic Fibre Category 2 classification of RCF. In one carcinogenicity. The aim of the suppliers was to re-discuss study, rats and hamsters were exposed at carcinogenicity. France and Germany had responded to that 30 mg/m3 to four types of specially documentation that a re-discussion was not warranted as the prepared “rodent respirable” RCF fiber. In a new studies do not have any impact on the existing classification second study, rats were exposed to three as inhalative carcinogen. The Technical Committee on additional, but lower, exposure Classification and Labelling of Dangerous Substances concentrations designed to identify a dose concluded that a re-discussion will only take place in case a response (multi-dose study). Only in member state would support such re-discussion. There was no animals exposed to the highest support by any member state. Since 2006, no further relevant concentration of this specially prepared scientific data have become available. The existing classification RCF (30 mg/m3) was the incidence of with R49 or H350i is still adequate and does not need to be re- disease statistically higher than in the evaluated. control animals or in historical control Experimental data and current classification: the arguments populations. Later studies identified two raised concerning MTD and particle impact have been major flaws in the RCC studies, one being considered (also during the discussions which lead to the legally the occurrence of pulmonary overload or binding classification) and do not have any impact on the exceeding the maximum tolerated dose classification decision. (MTD) and, two, the severe particle contamination of the fiber administered to the test subjects. European regulations and standards for Experimental data and current classification: the arguments addressing the classification of substances raised concerning MTD and particle impact have been and preparations recognize the significance considered (also during the discussions which lead to the legally of exceeding the maximum tolerated dose binding classification) and do not have any impact on the (MTD) and the occurrence of pulmonary classification decision. overload as a source of false positive The dossier does not challenge the outcome of the RCF

- 63 - results in toxicological animal studies. inhalation experiments with respect to a qualitative evaluation Studies conducted subsequent to the (i.e. classification for carcinogenicity). The inhalation December 1997 directive indicate that experiments are challenged because of the inadequacy to pulmonary overload did occur in previous derive adequate potency information which is crucial for cancer long-term bioassays of RCF. Compounding risk assessment. The latter is the case as there is an the discovery of exceeding the MTD, further unquantifiable impact of the concurrent particle exposure. The analysis also showed significant particle fact that Aluminosilicate Refractory Ceramic Fibres are contamination in the RCF samples used carcinogenic is not challenged at all on the basis of chronic during the previous studies. The inhalation and IP data. contamination likely contributed to overload and caused co-exposures of RCF fiber and It had been industry in the past who challenged the outcome of particles that can result in synergistic effects the inhalation experiments during the discussions. It was thereby nullifying the results wherein claimed that the inhalation test with rats is not relevant for overload was demonstrated. On page 22 of humans due to overload. The latter is scientifically not accepted the Annex XV dossier on RCF, the authors (e.g. from IARC). acknowledge that the RCC studies “were not deemed adequate for the derivation of cancer or mesothelioma risk estimates for RCF.” As part of its ongoing product stewardship Due to our knowledge actual publications have not sufficiently program, the RCF industry in both the proven the absence of any cancer risk to humans related to United States and Europe has sponsored a RCF exposure until now. The statistical power of the studies series of epidemiological studies, beginning (e.g. LeMaster et al. (2003) and Walker et al. (2002)) is limited in 1987, at the University of Cincinnati in the due to the relatively young and small cohorts. This has even United States and at the Institute of been announced by the authors too. Occupational Medicine (IOM) in Europe. These studies focused on RCF manufacturing workers and included the evaluation of historical exposures and studies of respiratory symptoms, possible fibrosis, effects on lung function, and mortality. While these studies were relatively new at the time of the 97/69/EC classification, they now contain more than 20 years of data with two key findings:

• Exposure to RCF did not result in the development of interstitial fibrosis. • Exposure to RCF did not result in

- 64 - incremental lung cancer or any mesothelioma. Harmonized Labeling & Classification The CLP criteria do not change the evaluation. H350i is the Under 1272/2008/EC (CLP), manufacturers legally binding classification according to CLP. The IARC criteria are directed to evaluate scientific evidence are included into CLP only with respect to the understanding of using the terms “sufficient” and “limited” as “limited” and “sufficient” evidence, not with respect to they have been defined by the International classification categories and their criteria. Agency for Research on Cancer (IARC).

“Limited evidence of carcinogenicity: the data suggest a carcinogenic effect but are limited for making a definitive evaluation because, e.g., (i) the evidence of carcinogenicity is restricted to a single experiment; or (ii) there are unresolved questions regarding the adequacy of the design, conduct or interpretation of the study; or (iii) the agent or mixture increases the incidence only of benign neoplasms or lesions of uncertain neoplastic potential, or of certain neoplasms which may occur spontaneously in high incidences in certain strains.” In IARC Monograph Volume 81, published in 2002, the working group examined RCF for the second time and reaffirmed a Group 2B Classification:

“Group 2B—The agent (mixture) is possibly carcinogenic to humans. The exposure circumstance entails exposures that are possibly carcinogenic to humans. This category is used for agents, mixtures and exposure circumstances for which there is limited evidence of carcinogenicity in humans and less than sufficient evidence of carcinogenicity in experimental animals. It may also be used when there is inadequate evidence of carcinogenicity in humans but

- 65 - there is sufficient evidence of carcinogenicity in experimental animals. In some instances, an agent, mixture or exposure circumstance for which there is inadequate evidence of carcinogenicity in humans but limited evidence of carcinogenicity in experimental animals together with supporting evidence from other relevant data may be placed in this group.” In the 2002 IARC Monograph, it was noted For IARC “inadequate evidence” does not exclude a cancer risk. that there was inadequate evidence in Such a conclusion only refers to the fact, that the available humans for the carcinogenicity of RCF. The studies do not permit a conclusion regarding the presence or IARC working group also noted that the absence of a cancer risk. greater particulate fraction of the RCF administered during the RCC studies could have influenced the development of inflammation and subsequent carcinogenic response in the chronic inhalation studies. Conclusion In March 2006 the European suppliers of Aluminosilicate The Annex XV dossier on RCF is flawed Refractory Ceramic Fibres provided the EU Technical and must be rejected because RCF is Committee on Classification and Labelling of Dangerous misclassified and does not meet the criteria Substances with the currently available scientific data with as a category 2 carcinogen under respect to Aluminosilicate Refractory Ceramic Fibre Commission Directive 97/69/EC. Under the carcinogenicity. The aim of the suppliers was to re-discuss newly adapted 1272/2008/EC (CLP), carcinogenicity. France and Germany had responded to that “The placing of a substance in Category 2 is documentation that a re-discussion was not warranted as the done on the basis of evidence obtained new studies do not have any impact on the existing classification from human and/or animal studies, but as inhalative carcinogen. The Technical Committee on which is not sufficiently convincing to place Classification and Labelling of Dangerous Substances the substance in Category 1A or 1B, based concluded that a re-discussion will only take place in case a on strength of evidence together with member state would support such re-discussion. There was no additional considerations. Such evidence support by any member state. Since 2006, no further relevant may be derived either from limited evidence scientific data have become available. The existing classification of carcinogenicity in human studies or from with R49 or H350i is still adequate and does not need to be re- limited evidence of carcinogenicity in animal evaluated. studies.” The CLP criteria do not change the evaluation. H350i is the legally binding classification according to CLP. The IARC criteria Based on significant new scientific are included into CLP only with respect to the understanding of

- 66 - evidence, had the 1997 evaluation for “limited” and “sufficient” evidence, not with respect to labeling under Commission Directive classification categories and their criteria. 97/69/EC taken place today, the results would have been a category 3 determination. Under the newly adapted 1272.2008/EC (CLP), and the guidance outlined under Article 37 (6), there are clear and compelling reasons to classify RCF as a category 2 substance not qualifying RCF for nomination to the SVHC candidate list.

(Classification and labelling) 36 20091008 European Association of A: Substance Identification The complete Annex I entry under Index No 650-017-00-8 refers the High Temperature to ‘Refractory Ceramic Fibres, Special Purpose Fibres … [Man- Insulation Wool Industry Page 5: The substance identification is not made vitreous (silicate) fibres with random orientation with (ECFIA), Industry or trade clear, as the reference to Annex I of alkaline oxide and alkali earth oxide (Na2O+K2O+CaO+ association, France 67/548/EEC is misquoted. As defined in the MgO+BaO) content less or equal to 18 % by weight]’. The proposal on page 5, the vast majority of Aluminosilicate RCFs and the Zirconia Aluminosilicate RCF are fibres (including organic fibres) would be actually the only commercially existing fibres which match to this included. The Annex I entry relates to “man entry. Other fibres (i.e. alkaline earth silicate wool or high- made vitreous (silicate) fibres with random alumina, low-silica-wool) do not fulfil this definition while the orientation with alkaline oxide and alkali (Na2O+K2O+CaO+ MgO+BaO) content is higher than 18 % by earth oxide (Na2O+K2O+CaO+MgO+BaO) weight (see IARC Vol 18). content less or equal to 18% by weight” and If there will be other RCF these do not fall in the same field of RCFs are not the only fibres meeting this application as the Aluminosilicate RCF. definition. Page 5: The subjects of both dossiers The CAS number 142844-00-6 refers to the CAS name should be described by the applicable “refractories, fibers, aluminosilicate”, which are defined by CAS single CAS number (142844-00-6) and EU as “amorphous man-made fibers produced from the melting and number 604-314-4. The advice from ECHA blowing or spinning of calcined kaolin clay or a combination of is wherever possible a “grouping approach” alumina (Al2O3) and silica (SiO2). Oxides such as zirconia, should be used and the logic of this position ferric oxide, titanium oxide, magnesium oxide, calcium oxide and is reinforced in the draft dossiers alkalies may also be added. Approximate percentages (by themselves as they say "The different weight) of components follow: alumina, 20-80%; silica, 20-80%; chemical composition of the commercially and other oxides in lesser amounts.” This generic description relevant types of refractory ceramic fibres covers a multitude of different substances and is therefore not does not have an impact on their dimension specific for the Aluminosilicate RCF resp. for the Zirconia and biopersistence. Thus, the risk-related Aluminosilicate RCF (see concentration range of the main information given (below) does not constituents in chapter 1.2 of Annex XV).

- 67 - discriminate between different types of fibres.” Thus it is a mistake to distinguish between the two substances and have two dossiers. The authors also seem unaware that these As mentioned in the Annex XV dossier Aluminosilicate RCF materials are glass wools and that at the 4th resp. Zirconia Aluminosilicate RCFs are UVCB substances. The Meeting of the Competent Authorities for naming of these both substances is in accordance with REACH the implementation of Regulation (EC) (RIP 3.10, Chapter 4.3.1.1). 1907/2006 (REACH) 16-17 June 2008 it was agreed that “Glass is the state of a substance rather than a substance as such” and that for legislative reasons glasses should be treated as UVCB substances (A substance of unknown or variable composition, complex reaction products or biological materials, further reference see: http://www.cpivglass.be/publications/CPIV% 20Statement%20for%20customers%20on% 20REACH %20- %20Revision%20Feb%202009_final.pdf). Treatment as a UVCB substance may involve some additional identification requirements such as specifying source or manufacturing process to fully define the substance but this is not specified in these dossiers. Page 6: RCFs are manufactured and i) The general requirement for an experimental inhalation test supplied as glass wools. These dossiers is that the sample needs to be respirable for the animals however are based on the properties of tested. Thus, the sample has to be prepared accordingly. experimental samples that have not and will never be placed on the market. Fibres, ii) In handling and use RCF wools release fibres which are when used in any animal or in vitro respirable for humans (similar dimensions to those used in experiment for the preparation of the test the animal tests). Thus, such fibres are relevant with respect material they must be reduced to a dust and to a putative human carcinogenicity. a fine fraction isolated. The materials described in these dossiers (RCF1, 2, 3 and 4) are such dusts prepared from the wools by milling and grinding and isolating a fine fraction designed to be “rodent respirable”.

- 68 - The “stock” fibres described in Table 1 of the dossier is this fine fraction, not the material manufactured and marketed. Therefore the dimensions given only apply to these test samples. Recently in Germany the diameter distributions of commercial RCFs have been measured for inclusion in the REACH registration dossiers for RCFs. All the commercial fibres are much coarser than these experimental preparations. The length of individual fibres in commercial RCF wools is difficult to measure but to form coherent wools they must be many centimetres long - that is hundreds of times more than the length of the “stock” experimental fibres. Page 7: For some reason the authors of As mentioned in the Annex XV dossier Aluminosilicate RCF these dossiers include formulae describing resp. Zirconia Aluminosilicate RCFs are UVCB substances. The the raw materials (components such as naming of these both substances is in accordance with REACH silica, alumina and zirconia) from which (RIP 3.10, Chapter 4.3.1.1). these fibres are manufactured, it is entirely unclear why they should do this. It would be more appropriate to describe the composition of the substance itself. Page 8: The description of the addition of The mentioned oxides of potassium, sodium, magnesium, other oxides is wrong – other oxides may be calcium … are modifier or stabilizer and added to alter the present in trace amounts but they are not physical and chemical properties such as tensile strength, added for the purposes described. All elasticity or durability. European manufacture uses only alumina, silica and where appropriate zirconia. ECFIA has little information on manufacture in continents other that Europe and America but we are confident that the position is similar to that described here and not as in the dossiers. B: Classification and Labelling The dossier does not challenge the outcome of the RCF inhalation experiments with respect to a qualitative evaluation The authors dismiss the use of inhalation (i.e. classification for carcinogenicity). The inhalation results in decisions on classification and experiments are challenged because of the inadequacy to

- 69 - labelling as they rightly identify the derive adequate potency information which is crucial for cancer particulate content as a problem in one set risk assessment. The latter is the case as there is an of such experiments. This particulate unquantifiable impact of the concurrent particle exposure. The content in the test samples led to an fact that Aluminosilicate Refractory Ceramic Fibres are overload effect; the maximum tolerated carcinogenic is not challenged at all on the basis of chronic dose was exceeded. It were these studies inhalation data. that formed the basis for the current EU classification as a carcinogen 2 under It had been industry in the past who challenged the outcome of 67/548/EEC (or 1b under CLP). Insofar we the inhalation experiments during the discussions. It was agree with the authors that the basis for the claimed that the inhalation test with rats is not relevant for existing classification is not valid, leading to humans due to overload. The latter is scientifically not accepted a misclassification of the substance. (e.g. from IARC).

After information on the particle mediated overload was discovered, IARC assigned RCF to the group 2b (comparable to category 3 under 67/548/EEC or 2 under CLP, respectively). IARC´s 2002 evaluation is in contradiction with the existing classification which was established in 1997 – new findings published after 1997 (a good summary is given in: Brown et al.: Survey of the Biological Effects of Refractory Ceramic Fibres: Overload and Its Possible Consequences, in Annals of Occupational Hygiene, 2005) and hence not available at the time of the EU classification was considered in the IARC discussion. This is also supported by the still ongoing No response, as literature reference is missing. human study conducted by the University of Cincinnati which – after an additional 12 years of observation – confirms that there is no excess lung cancer or any mesothelioma in the exposed cohort. Germany attempted to opt out of the The dossier does not challenge the outcome of the RCF European classification, finally adopting for inhalation experiments with respect to a qualitative evaluation internal use a classification scheme for (i.e. classification for carcinogenicity). The inhalation fibrous dusts (not the substances experiments are challenged because of the inadequacy to themselves) based on IP injection results. In derive adequate potency information which is crucial for cancer

- 70 - the dossiers the authors again seem to try risk assessment. The latter is the case as there is an and assign more relevance to the IP unquantifiable impact of the concurrent particle exposure. The injection test. fact that Aluminosilicate Refractory Ceramic Fibres are carcinogenic is not challenged at all on the basis of chronic inhalation data.

The classification criteria according to the dangerous substance directive and also according to the CLP regulation allow to assess fibre carcinogenicity by intraperitoneal testing. This means that the intraperitoneal test is scientifically accepted in EU legislation. The critique expressed on the questionable validity of this test is therefore baseless. We believe it is necessary to reopen the In March 2006 the European suppliers of Aluminosilicate debate on classification and labelling and Refractory Ceramic Fibres provided the EU Technical would be delighted to take part. In the Committee on Classification and Labelling of Dangerous context of such a debate the decision on Substances with the currently available scientific data with any listing of RCFs in Annex XV must be respect to Aluminosilicate Refractory Ceramic Fibre delayed until final consensus on the carcinogenicity. The aim of the suppliers was to re-discuss appropriate hazard classification is carcinogenicity. France and Germany had responded to that achieved. documentation that a re-discussion was not warranted as the new studies do not have any impact on the existing classification We appreciate that a proper re-evaluation, as inhalative carcinogen. The Technical Committee on including the latest relevant toxicology and Classification and Labelling of Dangerous Substances epidemiology findings will take some time. concluded that a re-discussion will only take place in case a The manufacturing industry has in place member state would support such re-discussion. There was no rigorous occupational hygiene and support by any member state. Since 2006, no further relevant substitution programmes which will ensure scientific data have become available. The existing classification that a delay or even rejection of Annex XV with R49 or H350i is still adequate and does not need to be re- listing of RCF / Alumino silicate wools does evaluated. not endanger anyone.

C: Important errors and omissions in dossier We are of the opinion, that there is no substantial difference justification whether “several countries” or “three countries” are mentioned.

- 71 -

Page 8: The “several countries in Europe” are France, Germany and the United Kingdom. Page 12 f.: Another set of human studies The carcinogenicity isn’t the subject of the study cited. was conducted in Europe by the Institute of Occupational Medicine (IOM), consistent with the conclusions of the ongoing work at the University of Cincinnati, for reference see i.e.: Cowie, H. A. et al. (2001): An epidemiological study of the respiratory health of workers in the European refractory ceramic fibre (RCF) industry, Occup. Environ. Med., 58: 800-810. Page 15: The submission of two dossiers In the dossier it is clearly stated that the tonnage of 25 000 t/a is for RCF seemingly led to “doubling” the related to the total amount of RCF fibres. No differentiation European manufacturing volume – the between the fibre types has been made. 25.000 tons are actually covering both chemistries (estimated breakdown: 30% Al- Zr-Si-Wools, 70% Al-Si-Wools). Page 19: The data used in the ladder Since the more recent information on exposure levels (Maxim, diagram (Figure 1) is outdated as it only 2008) have not been given in a numerical but in a graphical contains the 1st two years of the CARE form they have not been considered in the dossier. Gaining programme (CARE: Controlled and statistical values from the diagrams presented in Maxim40 Reduced Exposure, a programme (2008) was not considered appropriate due to a considerable developed by ECFIA - The European loss in precision. Association representing the High Temperature Insulation Wool Industry). For Only exposure data from peer reviewed publications have been actual exposure data see: Maxim, L. D. et taken into account. Due to this quality criterion the Care study al. (2008): Workplace monitoring of (ECFIA, 199941) was the most current publication on occupational exposure to refractory ceramic measurement data. Since the exposure data submitted by the fiber – a 17-year retrospective. Inhal. commentator in the pdf-file (RCF SVHC Comments.pdf) was not Toxicol., 20: 289-309. published in a journal (in a numerical form) it could not be found in a literature search.

We agree that for the most part, the companies that

- 72 - manufacture RCF in Europe are the same as those in the US since processes and controls are similar. It may therefore be appropriate to utilize all the relevant data for comparison.

Maxim 2008 has already been quoted in the dossier.

Maxim et al. (2008, s. above) also investigated the time trends in RCF exposure. It was found that the rate of improvement in fibre concentration in RCF manufacturing plants and customer facilities has slowed down in recent years. A significant decrease of RCF exposure in the near future is anticipated to be not very likely. Furthermore, the addition of a “limit value” Neither the text nor figure 1 does mention a “limit value”. to the chart is misleading. This was derived Instead, quantitative measures are shown which lead to a risk of from an unpublished risk assessment which 4:1000 or 4:10000. Exposures leading to risks above 4:1000 are lacks scientific support (see further remarks due to our understanding inacceptable and have therefore been below). The derived limit doesn’t exist in the illustrated in figure 1. German regulation. Other risk assessments have been produced by DECOS in the Netherlands and NIOSH in the USA. These both suggest a limit value of 0.5 F/ml with the recommendation to reduce concentration further down where this is technically feasible. The majority of the measurements realised in the framework of the CARE programme are below this level. Page 20: RCF is not sold to the general The quoted document was cited in relation to consumer public. In fact, a sales and marketing exposure. As this is not the main topic of the annex XV dossier restriction is applicable based on and consumer exposure is expected to be significantly lower

- 73 - 2001/41/EC, allowing sales to professional than worker exposure, no response is necessary. users only. As far as labelling is concerned, In our opinion, the exposure at work places exceeds acceptable the European RCF manufacturers or tolerable limits. Although this concept has yet not been voluntarily labelled substances and articles agreed on in the EU the authors identified risks at work places. already since 1985, long before the Whether or not this is true for an exposure through the classification took place. More complex environment seems to be not clear at the moment. Data on products such as catalytic converters (or airborne fibre concentrations in the environment are lacking but cars) and larger scale heating units are not values are expected to be far below those at work places. labelled by the respective manufacturers as there is no indication for consumers to be exposed to the substance. Consumer exposure to RCF fibrous dust has never been demonstrated (for reference see: Schneider, T. et al.: Ubiquitous fiber exposure in selected sampling sites in Europe, 1996 and H. Förster: Anorganische faserförmige Partikel in der Atmosphäre, in VDI Berichte 1075, 1993 page 211 ff).

The use of the products is industrial (except the very few cases of use in encapsulated appliances such as burning chambers of bigger heating units, where RCF could not be substituted). The dossier also admits that the use in consumer applications has dropped significantly over the last 15 years. It is hence even more surprising that the dossiers mention the 2004 EU notification, which was withdrawn by Germany itself in 2006 after consultation with the Commission as it lacked justification. It is noteworthy that in rejecting the German ban proposal the EU Commission noted that consumers were adequately protected by the operation of restrictions under the carcinogens directive – RCFs may only be

40 Maxim, D., Allshouse, J., Lentz, T.J., Venturin, D., Walters, T.E., Workplace Monitoring of Occupational Exposure to Refractory Ceramic Fiber- A 17 Year Retrospective , Inhal. Toxicol., 20:289-309, 2008 41 ECFIA (1999); Idenfication and control of exposure to refractory ceramic fibres. European Chemical Fibre Industry Association, 3 rue de Colonel Moll, 75017 Paris, France, Nov. 1999, 58p Illus. 6 ref.

- 74 - sold for professional use.

A similar outcome resulted on the 2007 petition mentioned in the dossier. After the review of the available information, the Commission concluded the following in October 2007: “The Commission has examined the supplementary information provided by the petitioner in August 2007. In fact, the data relate to the losses and defects of the catalytic converters themselves, in particular their metallic components. The data do not provide any information on the potential losses of ceramic fibres contained in catalytic converters. Furthermore, as the Commission services (DG ENTREPRISE) explained in a letter of 30 March 2007 sent directly to the Petitioner, the studies dealing with the emission of ceramic fibres from converters refer, as their main basis, to a thesis of 1999. This thesis concerned ceramic converters of the first generation and/or converters which are not computer controlled. The author of that thesis has made it known that she does not support any longer the conclusions drawn in this thesis.” (for reference see: http://www.europarl.europa.eu/meetdocs/20 04_2009/documents/cm/691/691790/69179 0en.pdf) Page 20 f: The assumption phrased in the See response to SELAS-LINDE GmbH, Company, Germany dossiers: “On the other hand current 20091005, above product developments indicate that the upper temperature limit of AES wool products could be increased significantly” lacks support. The main manufacturers of RCF have spent enormous time and resources to develop the AES products

- 75 - available today. While the R&D efforts on these products are continued, there is no indication for “significant” improvements in the foreseeable future. Furthermore, temperature resistance is clearly not the only parameter to be considered in the substitution process – other parameters such as chemical resistance and mechanical properties are important factors, too. The German “Technical Rules for Hazardous Substances” (TRGS 619) offers a more detailed list of parameters for some selected applications, (for reference see: www.baua.de). Page 22 ff: The German risk assessment We disagree. This comment is very general and would need model reported in these dossiers should be specification to be able to explain in each point why we questioned as it is subject to considerable disagree. uncertainties (some hundreds of thousands fold). While RCF could be considered as a possible SVHC based on the existing classification, this is no reason for accepting the poorly justified methodology in these dossiers. If this risk assessment method were accepted then many more synthetic mineral fibres would be SVHC substances as these materials are equally, or more, potent in IP studies although at present most of these are classified as CMR3 under 67/548 or carcinogen 2 under CLP respectively. The suggested approach to risk This is not correct. The strategy to derive cancer estimates for assessment was not fully accepted in RCFs was approved by the German toxicology expert Germany. In a meeting of the German committee UAIII. The AGS took note of the risk evaluation and Committee on Dangerous Substances decided to take it as a basis for the establishment of a packet of (AGS) in May 2009, the risk model was measures at the workplace. presented and after lengthy and controversial discussion was considered to have been acknowledged, however was not approved and it was made clear that it could

- 76 - not be used as basis for any regulatory provisions. However, these present dossiers represent an attempt to reintroduce this scheme in Europe. Page 22: The statement that granular Biopersistent particles without significant specific toxicity do materials do not cause mesotheliomas is NOT induce mesothelioma in the ip test. That was expressed. not true. Many agents can do so if injected There is no objection that metal ions like iron or nickel may into the pleural or peritoneal cavities. In induce mesothelioma in the ip test as the commenter refers to. particular it has been known that iron Thus, his criticism is misleading. containing materials can cause tumours after IP injection (for reference see: Br J Cancer. 1989;60:708-11 Br Med J 1959; i:947 .J Natl Cancer Inst 1960; 24:109.). Even very short crocidolite fibres can cause AM Kane describes several possible mechanisms for fibre mesotheliomas if their clearance from the carcinogenesis and lists several unanswered questions. The peritoneal cavity is blocked, this is probably conclusion of the commenter “probably an iron mediated by an iron mediated mechanism and not mechanism” is not supported by the reference cited. one related to fibre morphology (for reference see: AM Kane in “Mechanisms in Fibre Carcinogenesis” ISBN 0-306-44091- 1). The composition of the samples used for IP The composition of the samples used for IP injection (the injection (the method preferred by the method preferred by the authors) is given with table 6. authors) is not given. This makes it impossible to judge on the relevance of these data. Page 23: It is a characteristic of all asbestos We agree that the wording was imprecise. Meant is the minerals (and other natural mineral fibres) following: that they cleave longitudinally. However, as In contrast to serpentine asbestos, refractory ceramic fibres tend an essential part of their argument the to break transversely rather than cleaving along the fibre axis. authors of these dossiers deny that this is The behaviour to cleave along the fibre axis is associated with true for crocidolite, this is totally wrong and the fact that numerous new fibres generate intraperitoneally thus their argument is seriously flawed. which may increase dose and have an impact on the test outcome. The tendency to cleave along the fibre axis is different for different types of asbestos. Among the different types of asbestos, it is lowest for crocidolite. Thus, only results from ip tests with crocidolite were used to assess the comparative carcinogenic potency of asbestos and refractory ceramic fibres. Page 24: In making serious points on the Only for one reference, for only one crocidolite sample the

- 77 - regulation of a chemical surely it is not detailed sample information could be given only after direct appropriate to rely on data with no contact with the author. There is no impact at all that these two published or public source. To rely on numbers have not been peer-reviewed before. “personal communication” is not acceptable. For existing substances the general situation is that there are Similarly publication in un-refereed or un- mostly no GLP data are available. Moreover, because of animal audited reports should not be sufficient. To welfare, such tests are not repeated. our knowledge the experiments described here were not carried out to GLP standards and no quality control information is available. For example we do not know the identities of the materials used, from where they were sourced and how they were prepared from their parent products. We suggest that these data should, at most, form the basis for designing further tests and only after such testing it can be decided whether the results should be used in regulation.

As a minimum, further details of the source of the test materials and their preparation for experiments should be described. Samples must have been archived and these should be provided for further analysis. For example iron can be deposited on fibres during treatment in steel mills and this can itself contribute to the occurrence of mesothelioma after IP injection (see above). Page 29 ff: The references quoted are We disagree. This comment is very general and would need inadequate; many of the points made in the specification to be able to explain in each point why we text are not supported. In writing something disagree. designed to have a profound social and economic impact surely a more thorough review of the literature is deserved. 37 20091009 3M Europe N.V., Company, Page 5, IDENTITY OF THE SUBSTANCE Introductory statement, no response necessary. Belgium AND PHYSICAL AND CHEMICAL PROPERTIES

3M understands ECHA’s concern for

- 78 - potential carcinogenicity of certain RCFs under conditions in which fibers could be inhaled into the lung. This concern is based on two characteristics of certain inorganic fibers. First there is biopersistence, which is a characteristic of the material itself. Specifically it is fiber durability within the lung. Secondly, there is the fiber’s geometry which affects the likelihood of inhalation and entrapment into the deep lung tissue (respirability). The chemical composition of a number of ceramic fibers is similar to the description in Section 1.2 of the Annex XV dossiers for Refractive Ceramic Fibers. Generally, the chemistry of such ceramic fibers would ensure their stability, and thus persistence in the lung. However, since respirability is also a necessary condition for the cancer hazard, fiber dimensions have always been a key part of the determination of any potential cancer hazard in previous EU directive definitions. While the RCF dossier itself does not directly provide the dimensions for classification, it points to 1272/2008/EC (and “Note R”) under which certain fibers are excluded. This definition is the same as the previous one under EU directive 97/69/EC. In regard to RCF under index number 650- 017-00-8 in Annex VI of Regulation (EC) No 1272/2008, Note R specifically states:

“The classification as a carcinogen need not apply to fibres with a length weighted geometric mean diameter less two standard geometric errors greater than 6 μm.” In other words, both the characteristics of We agree biopersistence and critical fiber dimensions

- 79 - are needed in order for classification as a cancer hazard. RCFs exist within the European market in a 3M states that there are RCF fibres which are exempted from wide variety. Some of the RCFs are classification as carcinogenic due to their dimension and biopersistent but do not fall under the application of Nota R in Annex VI of the CLP regulation. Such definition of carcinogenic hazard according fibres would not fall under the intended authorisation procedure to Note R of (EC) No. 1272/2008. and would not need to be classified. If it can be assured that no inhalable fibres are formed within the whole supply and use chain, such fibres could be an interesting substitute. While it seems that manufacturers and If note R can be applied, it applies. This does not have any users could simply rely on this Note R impact on the authorisation proposal. exclusion, there would be potential for customer confusion and uncertainty if the exclusion is not stated more centrally in the substance identification of RCF. In light of this, and the fact that some RCFs The Authorisation procedure under REACh is intended to set have essential safety-related uses (in fire incentives for innovations and intends a widespread substitution barrier preparations, high temperature of SVHC. At the same time it cannot be ruled out that very insulation, and essential aerospace special applications exist, in which the use of RCF is essential. applications) 3M would like to suggest that For such applications the Authorisation process will make sure Authorization under REACH is perhaps not that risks are adequately controlled. Furthermore, additional the best approach. We believe that the information will become available for decision making. In this potential hazards of some RCFs could be context the alternative Restriction procedure is an instrument of better addressed through a targeted market REACH which is not precise enough for handling RCFs restriction approach.

Therefore, 3M would like to recommend that ECHA not add RCFs to the Candidate List. Instead, we suggest that ECHA should consider regulating RCFs with the REACH market restriction process, which the Agency could tailor to better sort out which RCFs are restricted, and for the specific commercial uses covered by those restrictions. The market restriction approach (Annex XVII of REACH) could be a more useful and flexible mechanism to specifically exclude from a market restriction rule, continuous filament RCFs, RCFs

- 80 - embedded in articles and non-respirable forms in essential safety and aerospace applications, because they do not present a carcinogenic risk.

As an alternative, we propose a definition of RCFs in the Candidate list that clearly identifies only the RCFs that are included in the Candidate List. Those RCFs that do not present a carcinogenic hazard should not be included. 38 20091012 Individual (affiliated with Identity of the Substance Chemical and We take note of this comment for a specific application, but we Howmet-Ciral snc, Howmet Physical Properties assume that for this application an alternative could exist like in SAS; Howmet Ltd for Alcoa The temperature range used in the other cases i.e. steel industry (For details see response to Power and Propulsion, a Superalloy investment casting process SELAS-LINDE GmbH, Company, Germany 20091005, above) division of ALCOA, requires the use of RCF blanket. The In case that there is demonstrably no suitable alternative and Company), France primary substitute, alkaline earth silicate the risks are adequately controlled or the socio-economic (AES) wool blankets, capabilities do not benefits overweigh the risks; these aspects would be taken into extend into the temperature range used for account in the procedure for authorization decisions. the investment casting process. No substitute has been successful in replacing RCF blankets. The substitutes such as vermiculate panels, bricks and concretes and other non-fibrous products are not suitable to the Superalloy investment casting application. In the Superalloy investment casting application the insulating wool blanket is wrapped around the features of aerospace and industrial gas turbine to meet the customer requirements. Therefore we must use this material and there are no substitutes. If we do not use this material we would not be able make Superalloy investment castings in Europe.

(Identity of the substance and physical and chemical properties) 39 20091013 BDI, Industry or trade Conclusion from an overall focus (cradle to The new materials and the products based on bio-soluble wool association, Germany grave analysis): AES have a potential for saving energy due to the key property

- 81 - of the insulating material: thermal conductivity (For details see • Competitive industry: response to SELAS-LINDE GmbH, Company, Germany o The European industry of users of 20091005, above) RCF/ASW products (steel, ceramic petro Several alternatives have lower thermal conductivity value in chem. industry, etc.) need these materials comparison to RCF and therefore provide a better insulating to achieve the high level set for energy capacity. Europe’s ambitious environmental targets in its saving and CO2-emission reduction by the Climate Change Programme (ECCP) can be achieved by using European Commission until 2020. alternatives for RCF. • Protection of the Environment: In our opinion, the exposure at work places exceeds acceptable o Exposure by RCF/ASW dust is not or tolerable limits. Although this concept has yet not been recognized in the general pub¬lic, this is agreed on in the EU the authors identified risks at work places. proved by several measurements in the Whether or not this is true for an exposure through the environment (i.e. EURIMA study 1996) environment seems to be not clear at the moment. Data on o The target set by EU member states could airborne fibre concentrations in the environment are lacking but be achieved using RCF/ASW products: values are expected to be far below those at work places. • Quote from Sigmar Gabriel; The authors are aware of this topic. Nevertheless, the necessity Bundesumweltminister; reference: „ VDI to an extensive use of isolating material is associated with the Ingenieurtag“ 05.05.2009 necessity to use material that is not related to risks for human „Die Bundesregierung hat sich mit einem health. wie ich glaube sehr anspruchsvollen Packet von Maßnahmen zur Verbesserung der Energieeffizienz und dem Umstieg zu erneuerbaren Energien ein großes Ziel gesetzt. Wir wollen im Jahre 2020 40 Prozent weniger Treibhausgase erzeugen als im Jahr 1990.“ The high target could be achieved supported by using RCF/ASW products in industrial high temperature application; (reference: Wimmer; VDI-Expertenforum Sept. 2009; at the German ministry of environment, BMU-Bonn) • In parts of the equipment which is needed to generate renewable energy, products made of RCF/ASW are needed. (i.e. biomass firing) • High tech elements and apparatus like We are of the opinion, that the temperature range of a gas chromatographs to analyse gases and in chromatic system will not exceed an upper limit of approx. 450 aircrafts, rockets light weight high °C. The usage of RCFs in this range of temperature can easily

- 82 - temperature insulation is needed and could be avoided by using other isolating material. be provided with the RCF/ASW-products in question. An analytical device for e.g. gas chromatographs does not operate at high temperature range only up to 450°C. The operating temperature in mass spectrometer is about 450°C. Thermal ionization mass spectrometry (TIMS) works at 2500°C (RCF are feasible up to 1260°C and with ZrO2 part suitable up to 1430°C. Therefore they do not cover this temperature range). The substitution principal of RCF in accordance to TRGS 619 should be applicated generally. From the technical and economical point of view the bio-soluble wool based on AES would be sufficiently appropriate for the above described needed temperature range. In addition to AES there are several well-established alternatives to RCF on the market which are technically feasible and which are currently used in different industries as substitutes for conventional refractory material RCF. o On the other hand parts which are needed We take note of this comment for a specific application, but we in the sector to generate renewable energy assume that for this application an alternative could exist like in can only be produced in using flexible, high other cases i.e. steel industry (For details see response to temperature firing which could be achieved SELAS-LINDE GmbH, Company, Germany 20091005, above) by using RCF/ASW-products (i.e. windmill In case that there is demonstrably no suitable alternative and blades and gears) the risks are adequately controlled or the socio-economic benefits overweigh the risks; these aspects would be taken into account in the procedure for authorization decisions. o Beside of other HTIWs Products We disagree; for this application instead of RCF other RCF/ASW are used for the hold¬ing of technology like knitted wire meshes can be used for bearing of catalytic converters and diesel particle filters ceramic catalytic converters and diesel particle filters42. An other (DPF) in cars or other equipment where no alternative is the usage of metal catalytic converter43 which does substitution for RCF/ASW is possi¬ble. not contain ceramic fiber mat. (TRGS 619) • In these sectors, so far there is no risk Since discussion of the quoted documents new development observed. The occupational sector is well took place on the market for insulation material controlled. The EU regulation 67/548 EC requests substitution, this is already the The authors of the annex XV dossier focused on worker

42 Ruthenberg R. (2008), http://www.prcenter.de/pressemitteilung-pdf-download.php?news_id=24368 43 Ruthenberg R. (2008), http://www.prcenter.de/pressemitteilung-pdf-download.php?news_id=24368 and http://www.xx1.de/info/manuscript_fasermatten1203.htm

- 83 - case under TRGS 619, which is also exposure rather than public exposure to RCFs. available in Eng¬lish and French language. Two dossiers were prepared because the substance 1. Factual comments related to both identification according to REACH does not allow only one dossiers: dossier. According to the guidance for identification and naming of • ECHA advised a “grouping approach” substances under REACH UVCB substances are specified with wherever possible and the authors the IUPAC-name of their constituents. In the case of themselves mention in each of the two Aluminosilicate RCF the main constituents are Al2O3 and SiO2 dossiers related to RCF that “…the risk which both are present more than 10 %. In the case of Zirconia related information given does not Aluminosilicate RCF the main constituents are Al2O3, SiO2 and discriminate between different types of ZrO2 which all are present more than 10 % in the UVCB fibres”. Thus: substance. Why are two dossiers prepared with almost the same content? What about reducing bureaucracy, unnecessary time consuming and cost creating work for all parties involved? • Page 1ff: The term “RCF” is often used The titles of the dossiers are “Aluminosilicate Refractory under different meanings and names and so Ceramic Fibres” resp. Zirconia Aluminosilicate Refractory several times mixed up in the whole Ceramic Fibres“. As mentioned in the dossier these both fibres document mentioning different chemical belong to the group of the refractory ceramic fibres which are formulations and forms (i.e. diameter/length special category of synthetic vitreous fibres (SVFs, or, more for fibrous dust instead of the commercial commonly known as man-made vitreous fibres (MMVF)). substance): o “Alumino Refractory Ceramic Fibres”; and “Zirkonia Alumino Refractory Ceramic Fibres” are used as title for the two dossiers o “RCF1,2,3,4-types” have no commercial i) The general requirement for an experimental inhalation test relevance is that the sample needs to be respirable for the animals tested. Thus, the sample has to be prepared accordingly.

ii) RCF wools release fibres which are respirable for humans (similar dimensions to those used in the animal tests). Thus, such fibres are relevant with respect to a putative human carcinogenicity. o “Aluminium silicate wool” is used in EN Statement, no response necessary. standards, TRGS 619, BREF/BAT and VDI recommendations to define a special type of

- 84 - material. • Page 15 “The European Chemical Fibre No response on this comment. Industry Association (ECFIA)…” ECFIA is representing the “High Temperature Insulation Wool Industry” and not “European Chemical Fibre Industry Associa¬tion”.

• Page 18 Paragraph on risk: “In Statement, no response necessary. Germany…” Yes the German authority body (AGS), where members of the BDI agreed on ranges for tolerated und acceptable risk, in June 2008.

Announcement on Hazardous Sub-stances Risk figures and expo¬sure-risk rela- tionships in activities involving car¬cinogenic hazardous sub-stances Announcement 910 It is not correct that there was an This is not correct. The strategy to derive cancer estimates for agreement on the risk evaluation (ERB- RCFs was approved by the German toxicology expert KMF) related to RCF/ASW which is stated committee UAIII. The AGS took note of the risk evaluation and in the dossier. [Reference: AGS –meeting in decided to take it as a basis for the establishment of a packet of June 2009]. measures at the workplace. There are massive “uncertainties” of the Industry grossly exaggerates the uncertainty of the risk analysis, evaluation related to ERB-KMF and BDI which is mainly caused by the uncertainty of the asbestos risk objected to the minutes of the AGS (44. analysis. But as a “central estimate” for asbestos is used for AGS (04./05.05.2009): Einspruch gegen comparison with RCF, it is unlikely that the RCF risk is den Entwurf der Ergebnis¬niederschrift v. overstated. 06.08.2009)

• Page 20: {Quote: “Supplementary Text to Since discussion of the quoted documents new development Notification 2004/370/D took place on the market for insulation material • The German authorities (BMU) withdrew the EU-Notification in 2006 after The authors of the annex XV dossier focused on worker intervention from the EU-Commission exposure rather than public exposure to RCFs. (message 009 – communication from the

- 85 - Commission – SG(2006) D/50654). In the notification it was used nearly the justification like in the dis¬cussed dossiers now. Is the EU classification challenged by the No German dossiers ? The concept described in the dossiers are The classification criteria according to the dangerous substance not the basis for the EU regulatory practice directive and also according to the CLP regulation allow to nor supported by the worldwide scientific assess fibre carcinogenicity by intraperitoneal testing. This community and experts. means that the intraperitoneal test is scientifically accepted in EU legislation. The critique expressed on the questionable validity of this test is therefore baseless. The concept of risk assessment used in the dossier has not been discussed yet at the EU level or in the worldwide scientific community. So the statement that it is not supported by those is simply wrong. Substitution: Substitution is already The authors are of the opinion, that there are viable alternatives regulated in EU (67/548 EC), the German on the market. (For details see response to SELAS-LINDE ministry of labour and social affairs had GmbH, Company, Germany 20091005, above) published the TRGS 619 dealing specifically with the substitution for products made of alumino silicate wool (ASW/RCF). In practice, there is no viable alternative possible in many of the high temperature indus¬trial applications, therefore this TRGS was created to document the arguments when there is no alternative material available on the grounds of health and safety, environmental and economical rea¬sons. • Page 22 and 23 The dossier does not challenge the outcome of the RCF • The basis for the existing EU- inhalation experiments with respect to a qualitative evaluation Classification is challenged in the German (i.e. classification for carcinogenicity). The inhalation two dossiers. experiments are challenged because of the inadequacy to • The results of the RCC-studies which were derive adequate potency information which is crucial for cancer the main basis for the EU-classification are risk assessment. The latter is the case as there is an questioned in the dossiers – would this unquantifiable impact of the concurrent particle exposure. The mean that the EU-Classification should be fact that Aluminosilicate Refractory Ceramic Fibres are reviewed? carcinogenic is not challenged at all on the basis of chronic

- 86 - inhalation and IP data.

44 Wardenbach, P., Pott, F., Woitowitz, H.-J., 2000. Differences between the classification of man-made vitreous fibres (MMVF) according to the European directive and German legislation: analysis of scientific data and implications for worker protection. Eur. J. Oncol. 5 (2), 111–118.

- 87 - tolerable occupational risk” shared by the regulatory background in the EU. Thus, it is one possible way of rest of the EU – what is the relevance of the performing risk assessments in the EU under REACH. German position in light of the Annex XV discussion? o Does Germany intend to use this dossier The classification criteria according to the dangerous substance to establish IP testing as a “standard” for directive and also according to the CLP regulation allow to hazard classification? This would be in assess fibre carcinogenicity by intraperitoneal testing. This con¬tradiction to existing EU practice and is means that the intraperitoneal test is scientifically accepted in not in line with the expert opinion of almost EU legislation. The critique expressed on the questionable all international scientists. validity of this test is therefore baseless.

Moreover, taking into account the human cancer incidences caused by asbestos, it is prudent to regulate other fibres on the basis of the entire evidence including intraperitoneal injection studies. Positive results after intraperitoneal injection are regarded as essential for evidence of carcinogenic activity in experimental animals to be classified as ‘‘sufficient,’’ as this test system reveals the carcinogenic activity of asbestos fibres and of fibres with lower potency in relation to asbestos fibres. Moreover, it has to be recalled that mesothelioma induction by intraperitoneal injection of fibres is hardly influenced by concomitant injection of granular particles. The test model of intraperitoneal injection of fibres revealed that the carcinogenic potency of various man-made vitreous fibres can differ by three orders of magnitude (Wardenbach et al., 200045), which enables the selection and use of less potent man-made vitreous fibres and the application of adequate protective measures. • In fact, the involved industry had notified to In March 2006 the European suppliers of Aluminosilicate us that they intend to submit a request for Refractory Ceramic Fibres provided the EU Technical re-classification in line with arti¬cle 37 (6) of Committee on Classification and Labelling of Dangerous CLP/GHS Substances with the currently available scientific data with respect to Aluminosilicate Refractory Ceramic Fibre carcinogenicity. The aim of the suppliers was to re-discuss carcinogenicity. France and Germany had responded to that documentation that a re-discussion was not warranted as the new studies do not have any impact on the existing classification

45 Wardenbach, P., Pott, F., Woitowitz, H.-J., 2000. Differences between the classification of man-made vitreous fibres (MMVF) according to the European directive and German legislation: analysis of scientific data and implications for worker protection. Eur. J. Oncol. 5 (2), 111–118.

- 88 - as inhalative carcinogen. The Technical Committee on Classification and Labelling of Dangerous Substances concluded that a re-discussion will only take place in case a member state would support such re-discussion. There was no support by any member state. Since 2006, no further relevant scientific data have become available. The existing classification with R49 or H350i is still adequate and does not need to be re- evaluated. Due to our understanding of article 37 (6), companies are asked to transfer a proposal for a revised classification to the competent authority. Actually, no such proposal has been presented to the German CA and therefore we do not see any need to further respond to this topic. • BDI supports the proposed expert hearing There is no regulatory possibility to establish globally to set up a globally agreed harmonized harmonized classifications. classification for RCF/ASW (fibrous dust). • To prevent from further unscientific and emotional debates BDI recommends to reject the Annex XV proposal concerning the two dossiers on alumino silicate RCF and ziconia alumino silicate RCF until the discussion is finalised by international scientific experts.

• Page 23 We agree that the wording was imprecise. Meant is the • In the two dossiers, the comparison following: between crocidolyte and RCF is In contrast to serpentine asbestos, refractory ceramic fibres tend scientifically flawed: Quote: “... amphibole to break transversely rather than cleaving along the fibre axis. asbestos (i.e. croci¬dolyte), which does not The behaviour to cleave along the fibre axis is associated with cleave along the fibre axis ...”. This the fact that numerous new fibres generate intraperitoneally state¬ment is wrong and is challenged by which may increase dose and have an impact on the test international experts on miner¬alogy (Dr. outcome. The tendency to cleave along the fibre axis is different Thorsten Tonnesen; University of Aachen for different types of asbestos. Among the different types of (RWTH). asbestos, it is lowest for crocidolite. Thus, only results from ip • Is “personal communication with Dr. tests with crocidolite were used to assess the comparative Roller, February 6th 2008” enough to justify carcinogenic potency of asbestos and refractory ceramic fibres. an issue, where no agreement was achieved? • Page 29 (Reference list) We disagree. This comment is very general and would need

- 89 - • Many actual references are missing and specification to be able to explain in each point why we often only earlier publica¬tions on specific disagree. details are being used, omitting later and more relevant findings even by the same authors (i.e. Rödelperger,2004; Brown, 2005; Bellmann)

(Classification and labelling) 40 20091013 Unifrax Corp., Company, "Classification and Labelling" is addressed Statement, no response necessary. United States of America on pages page 2-5 of the attached PDF file titled “RCF SVHC comments”. Because this comment form does not allow for efficient submission of graphics, tables, and references, the reader is referred to the attached PDF file for a more complete argument against labeling RCF as a substance of very high concern. An excerpt of relevant argument follows In March 2006 the European suppliers of Aluminosilicate (please refer to the attached PDF for tables, Refractory Ceramic Fibres provided the EU Technical figures, footnotes and references): Committee on Classification and Labelling of Dangerous Substances with the currently available scientific data with Carcinogen classification for RCF respect to Aluminosilicate Refractory Ceramic Fibre The EU classified RCF into Category 2 carcinogenicity. The aim of the suppliers was to re-discuss (substances which should be regarded as if carcinogenicity. France and Germany had responded to that they are carcinogenic in man) in the 23rd documentation that a re-discussion was not warranted as the Adaptations to Technical Progress (ATP) of new studies do not have any impact on the existing classification Annex 1 of Directive 67/548/EEC [No. L343, as inhalative carcinogen. The Technical Committee on 13, 12, 1997, dated 5 December 1997]. Classification and Labelling of Dangerous Substances The International Agency for Research on concluded that a re-discussion will only take place in case a Cancer (IARC) later (2002) reaffirmed an member state would support such re-discussion. There was no earlier classification of RCF (1988) as support by any member state. Since 2006, no further relevant belonging to Group 2B (possibly scientific data have become available. The existing classification carcinogenic to humans) on the basis of with R49 or H350i is still adequate and does not need to be re- what an expert panel determined was evaluated. inadequate evidence in humans but sufficient evidence in experimental animals. As IARC was most explicit on the basis for classification it is appropriate to review their rationale.

- 90 - -Human Data Statement from IARC, no response necessary. IARC and other agencies accord epidemiological data as being of particular relevance to classification. The IARC Working Group (2002) summarized the human data on RCF as follows:

“Preliminary results from a United States epidemiological mortality study of refractory ceramic fibre workers were available. However, the limited epidemiological data do not permit an adequate evaluation of the cancer risk associated with exposure to refractory ceramic fibres.” Subsequent to the IARC Monograph, two Because of the relatively young and small cohorts studied by relevant articles (LeMasters et al., 2003; LeMaster et el.(2003) respectively Walker et al. (2002) the Walker et al., 2002) emerged. The IARC results allow only limited conclusions. Only one positive conclusion might need to be revised in the incidence would produce a substantial effect on the significance light of these new findings. level. LeMasters et al., (2003) reported results of a mortality study of occupationally exposed cohorts at two plants in the United States. Current and former male workers employed between 1952 and 2000 at two RCF manufacturing plants were followed to investigate any possible excess in mortality. These plants were selected for the mortality study because other plants had workers co-exposed to other dusts or fibers. In addition, the study plants had comprehensive documentation on worker job histories. Results of the LeMasters et al., (2003) study are as follows. There was no significant excess mortality related to all deaths (SMR = 69.8), all cancers (SMR = 94.2), malignancies (SMR = 78.8) or diseases of the respiratory system (SMR =106.8), including mesothelioma. However,

- 91 - there was a statistically significant association with cancers of the urinary organs; SMR = 334.8 with a confidence interval at 95% [111.6-805.4] (LeMasters et al., 2003). Figure 1 shows calculated SMRs (filled circles) and lower and upper 95% confidence bounds (filled triangles) for several of the causes of death included in this analysis. LeMasters et al. (2003) also employed Cox’s proportional hazards model (adjusted for age and race); this did not show elevated risk with cumulative RCF exposure (risk ratio 0.99, 95% CI = 0.85 – 1.16). A parallel analysis that lagged exposure by 10 years led to similar conclusions. The authors of the ongoing mortality study (LeMasters et al. 2003) provided an extensive discussion of the strengths and weaknesses of the study. Limitations of the study are the relative youth of the cohort and its small size. The mortality analysis did have a 95% power to detect a 2-fold increase in all deaths and all cancers and a 40% power to detect a 2-fold increase in lung cancer. As noted above, this mortality study is continuing.

In an earlier analysis of Walker et al. (2002) designed to explore the statistical power of the mortality study concluded that the experience of lung cancer mortality in the RCF cohort was statistically incompatible with the hypothesis that RCF was as potent (with respect to lung cancer) as amphibole asbestos assuming identical cumulative exposure to the cohort. However, the possibility that RCF was as potent as chrysotile asbestos could not be excluded.

- 92 - Walker et al., (2002) concluded that the exposure duration was not sufficiently long to reach any conclusions with respect to mesothelioma, although no cases of mesothelioma were observed in the study cohort -Animal data In March 2006 the European suppliers of Aluminosilicate The IARC Working Group (2002) Refractory Ceramic Fibres provided the EU Technical summarized the animal data as follows: Committee on Classification and Labelling of Dangerous Substances with the currently available scientific data with “In a well-designed, long-term inhalation respect to Aluminosilicate Refractory Ceramic Fibre study with refractory ceramic fibres in rats, a carcinogenicity. The aim of the suppliers was to re-discuss statistically significant increase in the carcinogenicity. France and Germany had responded to that incidence of lung tumours and a few documentation that a re-discussion was not warranted as the mesotheliomas were observed. In a well- new studies do not have any impact on the existing classification designed, long-term inhalation study of as inhalative carcinogen. The Technical Committee on refractory ceramic fibres in hamsters, a Classification and Labelling of Dangerous Substances significant increase in the incidence of concluded that a re-discussion will only take place in case a mesotheliomas was observed. After member state would support such re-discussion. There was no intratracheal instillation, two studies support by any member state. Since 2006, no further relevant reported no excess in tumour incidence in scientific data have become available. The existing classification rats. In three intrapleural studies in rats, no with R49 or H350i is still adequate and does not need to be re- significant increase in tumour incidence was evaluated. observed. In intraperitoneal studies in rats and hamsters, tumour incidence was related to fibre length and dose.” The long-term inhalation studies with rats (Mast et al., 1995a, b) and hamsters (McConnell et al., 1995b) were conducted at the RCC laboratories (then) located in Geneva, Switzerland. As the hamster study employed only one non-zero dose, the multi-dose rat study has been the focus of analysis. The IARC Working Group reported the The classification criteria according to the dangerous substance positive IP results with RCF (and also directive and also according to the CLP regulation allow to positive results with other synthetic vitreous assess fibre carcinogenicity by intraperitoneal testing. This fibers [SVFs] including rock (stone) wool, means that the intraperitoneal test is scientifically accepted in slag wool, 475 and E-glass, and insulation EU legislation. The critique expressed on the questionable

- 93 - glass wool [fiber glass]). Presumably, the validity of this test is therefore baseless. IARC Working Group reached the conclusion that the IP route was of limited Moreover, taking into account the human cancer incidences relevance for carcinogen classification caused by asbestos, it is prudent to regulate other fibres on the purposes, because many of the other SVFs basis of the entire evidence including intraperitoneal injection were placed in Group 3 (not classifiable as studies. Positive results after intraperitoneal injection are to carcinogenicity) despite the positive IP regarded as essential for evidence of carcinogenic activity in results. The relevance of data from IP experimental animals to be classified as ‘‘sufficient,’’ as this test experiments is discussed at more length in system reveals the carcinogenic activity of asbestos fibres and the section on risk analysis presented of fibres with lower potency in relation to asbestos fibres. below. Moreover, it has to be recalled that mesothelioma induction by intraperitoneal injection of fibres is hardly influenced by In the years since RCF was classified by the concomitant injection of granular particles. The test model of EU several studies were conducted to try to intraperitoneal injection of fibres revealed that the carcinogenic understand the results of the RCC studies potency of various man-made vitreous fibres can differ by three (see references at the end of these orders of magnitude (Wardenbach et al., 200046), which enables comments). It is now fairly well established the selection and use of less potent man-made vitreous fibres (see Bellmann et al., 2001; Brown et al., and the application of adequate protective measures. 2005; Mast et al., 2000a, b; Yu and Oberdörster, 2000; Yu et al., 1994a, b, Wardenbach, P., Pott, F., Woitowitz, H.-J., 2000. Differences 1995a, b, 1996) that overload occurred in between the classification of man-made vitreous fibres (MMVF) the RCC experiments as an artifact of the according to the European directive and German legislation: test article preparation process, which analysis of scientific data and implications for worker protection. produced a sample with a much greater Eur. J. Oncol. 5 (2), 111–118. ratio of particles to fibers than was found in Experimental data and current classification: the arguments the RCC studies of other SVFs and which raised concerning MTD and particle impact have been was not representative of RCF aerosols considered (also during the discussions which lead to the legally found in the workplace. The combination of binding classification) and do not have any impact on the particles and fibers caused the maximum classification decision. tolerated dose (MTD) to be exceeded. The dossier does not challenge the outcome of the RCF Overload reduces the rate of clearance of inhalation experiments with respect to a qualitative evaluation fibers and ultimately results in a persistent (i.e. classification for carcinogenicity). The inhalation inflammatory response, fibrosis, and tumors experiments are challenged because of the inadequacy to (Morrow, 1994). In consequence, it is not derive adequate potency information which is crucial for cancer possible to assess with any accuracy what risk assessment. The latter is the case as there is an the outcome of an experiment would have unquantifiable impact of the concurrent particle exposure. The

46 Wardenbach, P., Pott, F., Woitowitz, H.-J., 2000. Differences between the classification of man-made vitreous fibres (MMVF) according to the European directive and German legislation: analysis of scientific data and implications for worker protection. Eur. J. Oncol. 5 (2), 111–118.

- 94 - been absent overload. Yu and Oberdörster fact that Aluminosilicate Refractory Ceramic Fibers are (2000) used the data from the RCC multi- carcinogenic is not challenged at all on the basis of chronic dose study for risk estimation purposes inhalation data. noted: “Obviously, quantitative risk estimates in It had been industry in the past who challenged the outcome of general when derived from animal studies the inhalation experiments during the discussions. It was should be regarded as hypothetical risk claimed that the inhalation test with rats is not relevant for estimates. For RCF specifically, it should humans due to overload. The latter is scientifically not accepted be kept in mind that the lung and pleural (e.g. from IARC). tumor response observed in the chronic rat An extensive justification has been recently provided to justify inhalation study was most likely greater due the fact that chronic inhalations studies in rats are insensitive to to the presence of large numbers of non- detect fibre carcinogenicity (Wardenbach et al. 200547). Having fibrous RCF particles. At the human in mind the higher sensitivity of humans compared to rats after workplace—and probably also in the inhalation of environment—only small numbers of non- asbestos, the arguments provided in favour of the inhalation fibrous RCF particles are studies are highly questionable. present…Therefore, use of the tumor data from that study very likely resulted in a more conservative human cancer risk estimate. The contribution of the non-fibrous particles to the rat RCF tumor response cannot be quantitated due to the lack of appropriate data.” [Emphasis added.] The carcinogen classification criteria take note of several factors that may limit the evidential value of data from animal experiments. For example, applicable guidance for (67/548/EEC) notes that one important factor to be considered in choosing between carcinogen categories 2 and 3 even though tumors have been induced in animals are: “carcinogenic effects [observed] only at very high dose levels exceeding the maximal tolerated dose.” Similarly, (1272/2008 (CLP, GHS) notes: “the possibility of a confounding effect of excessive toxicity at test doses” is

47 Wardenbach P, Rödelsperger K, Roller M, Muhle H. Classification of man-made vitreous fibers: Comments on the revaluation by an IARC working group. Regul Toxicol Pharmacol. 2005 Nov;43(2):181-193.

- 95 - relevant. It is my understanding that RCF manufacturers intend to petition the authorities to have the carcinogen classification for RCF to be changed to Group 3 (EU) or Group 2 (GHS). (Classification and labelling) 41 20091013 Unifrax Corp., Company, Available epidemiological studies (morbidity No response possible as the literature reference is missing United States of America and mortality) have failed to reveal any interstitial fibrosis above population background, incremental lung cancer, or mesothelioma among workers exposed to RCF. In relevant studies, there was no significant excess mortality related to all deaths (SMR = 69.8), all cancers (SMR = 94.2), malignancies (SMR = 78.8) or diseases of the respiratory system (SMR = 106.8), including mesothelioma. This issue is addressed in more detail on The detailed points referred to are found elsewhere in this pages page 2-4 of the attached PDF file document and will be specifically responded where they appear. titled “RCF SVHC comments”. Because this comment form does not allow for efficient submission of graphics, tables, footnotes and references, the reader is referred to the attached PDF file for a more complete argument against labeling RCF as a substance of very high concern. An excerpt of relevant argument follows (refer to the attached PDF for tables, figures and references): -Human Data Statement by IARC, no response necessary. IARC and other agencies accord epidemiological data as being of particular relevance to classification. The IARC Working Group (2002) summarized the human data on RCF as follows:

“Preliminary results from a United States epidemiological mortality study of refractory

- 96 - ceramic fibre workers were available. However, the limited epidemiological data do not permit an adequate evaluation of the cancer risk associated with exposure to refractory ceramic fibres.”

Subsequent to the IARC Monograph, two Because of the relatively young and small cohorts studied by relevant articles (LeMasters et al., 2003; LeMaster et el.(2003) respectively Walker et al. (2002) the Walker et al., 2002) emerged. The IARC results allow only limited conclusions. Only one positive conclusion might need to be revised in the incidence would produce a substantial effect on the significance light of these new findings. level.

LeMasters et al., (2003) reported results of a mortality study of occupationally exposed cohorts at two plants in the United States. Current and former male workers employed between 1952 and 2000 at two RCF manufacturing plants were followed to investigate any possible excess in mortality. These plants were selected for the mortality study because other plants had workers co-exposed to other dusts or fibers. In addition, the study plants had comprehensive documentation on worker job histories. Results of the LeMasters et al., (2003) study are as follows. There was no significant excess mortality related to all deaths (SMR = 69.8), all cancers (SMR = 94.2), malignancies (SMR = 78.8) or diseases of the respiratory system (SMR =106.8), including mesothelioma. However, there was a statistically significant association with cancers of the urinary organs; SMR = 334.8 with a confidence interval at 95% [111.6-805.4] (LeMasters et al., 2003). Figure 1 shows calculated SMRs (filled circles) and lower and upper 95% confidence bounds (filled triangles) for

- 97 - several of the causes of death included in this analysis.

The authors of the ongoing mortality study (LeMasters et al. 2003) provided an extensive discussion of the strengths and weaknesses of the study. Limitations of the study are the relative youth of the cohort and its small size. The mortality analysis did have a 95% power to detect a 2-fold increase in all deaths and all cancers and a 40% power to detect a 2-fold increase in lung cancer. As noted above, this mortality study is continuing. In an earlier analysis of Walker et al. (2002) The authors don’t see any dissent designed to explore the statistical power of the mortality study concluded that the experience of lung cancer mortality in the RCF cohort was statistically incompatible with the hypothesis that RCF was as potent (with respect to lung cancer) as amphibole asbestos assuming identical cumulative exposure to the cohort. However, the possibility that RCF was as potent as chrysotile asbestos could not be excluded. Walker et al., (2002) concluded that the exposure duration was not sufficiently long to reach any conclusions with respect to mesothelioma, although no cases of mesothelioma were observed in the study

- 98 - cohort.

(Human health hazard assessment) 42 20091013 Unifrax I LLC, Company, This issue is addressed in more detail on The detailed points referred to are found elsewhere in this United States of America pages 4-7 of the attached PDF file titled document and will be specifically responded where they appear. “Unifrax comments”. Because this comment form does not allow for efficient submission of graphics, tables, footnotes and references, the reader is referred to the attached PDF file for a more complete argument against labeling RCF as a substance of very high concern. An excerpt of relevant argument follows (refer to the attached PDF for tables, figures and references): Classification and Labeling In March 2006 the European suppliers of Aluminosilicate This section provides Unifrax’s comments Refractory Ceramic Fibers provided the EU Technical on classification of RCFs. A short history of Committee on Classification and Labelling of Dangerous RCF classification provides useful Substances with the currently available scientific data with background. In the United States RCFs respect to Aluminosilicate Refractory Ceramic Fibre were first classified as “reasonably carcinogenicity. The aim of the suppliers was to re-discuss anticipated to be a human carcinogen” by carcinogenicity. France and Germany had responded to that the (US) National Toxicology Program documentation that a re-discussion was not warranted as the (NTP) in the Seventh Annual Report on new studies do not have any impact on the existing classification Carcinogens (1994) and in category B2 as inhalative carcinogen. The Technical Committee on (probable human carcinogen) by the US Classification and Labelling of Dangerous Substances Environmental Protection Agency (EPA) in concluded that a re-discussion will only take place in case a 1993. These assessments were based on member state would support such re-discussion. There was no papers published prior to 1993-4 and support by any member state. Since 2006, no further relevant apparently relied on a 1988 assessment by scientific data have become available. The existing classification the International Agency for Research on with R49 or H350i is still adequate and does not need to be re- Cancer (IARC), which placed RCF in Group evaluated. 2B (possibly carcinogenic to humans) on what was termed “sufficient evidence of In addition, the human data are not the basis for the carcinogenicity in experimental animals” classification and a lack of relevant epidemiological data. In a subsequent assessment (2002), IARC reaffirmed the placement of RCF into Group 2B on the basis of inadequate evidence in

- 99 - humans and sufficient evidence in experimental animals. As noted above, the EU placed RCF into Category 2 (substances which should be regarded as if they are carcinogenic to man) in 1997. The correspondence between the EU categories and the forthcoming GHS categories is summarized in Fig. 1. Specifically Category 2 in the EU system becomes Category 1B in the GHS system. Unifrax believes that the appropriate classification for RCF in the EU system is Category 3(b), which corresponds to Category 2 (suspected human carcinogen) in the GHS system (see Fig. 1). If placed in Category 2, then it would be inappropriate to list RCF as a SVHC. The reasons why we believe that this is appropriate are summarized below. -Reasons for reclassification Brown et al. (2005) recently reviewed the available data on RCF toxicity and epidemiology. The results of epidemiological studies by the University of Cincinnati in the US and the Institute of Occupational Medicine in Europe (see e.g., Cowie et al., 2001; LeMasters et al., 1998, 2003; Lockey et al., 1998, 2002; Utell and Maxim, 2009; Walker et al., 2002 and contained references) indicated that occupationally exposed workers developed symptoms (e.g., shortness of breath) “similar to those reported in other dust- exposed populations. In the US an initial cross-sectional study of pulmonary function (LeMasters et al., 1998) indicated that there were statistically, but not clinically, significant decrements for forced vital capacity (FVC) as a function of exposure

- 100 - duration for male current and former smokers. A later longitudinal analysis (Lockey et al., 1998) revealed no excessive decline in lung function for workers during the seven year follow-up period after their initial test. In the European study (Cowie et al., 2001) there was a significant association between FEV1 and cumulative exposure in past smokers and mild decrements in FVC and FEV1 for current male smokers. The radiological investigation of workers in the US cohort exposed to RCF revealed a statistically significant increase in the prevalence of pleural plaques; specifically pleural changes were seen in 2.7% of workers overall. There was no statistically significant evidence of interstitial fibrosis. In the US the University of Cincinnati also completed a mortality study of workers at two RCF manufacturing plants (LeMasters et al., 2003). There was no significant excess mortality related to all deaths, all cancers, malignancies, or diseases of the respiratory system, including mesothelioma. There was, however, a statistically significant association with cancers of the urinary organs (albeit with a very wide confidence band). The mortality analysis had a 95% power to detect a 2-fold increase in all deaths and all cancers and a 40% power to detect a 2-fold increase in lung cancer. An analysis by Walker et al. (2002) concluded that the available mortality and exposure data were statistically incompatible with the assumption that RCF was as potent as crocidolite asbestos in causing lung cancer. Because of limitations in the duration since

- 101 - first exposure it was not possible to conduct a similar analysis for mesothelioma. That said, no one in the occupationally-exposed cohort has developed mesothelioma. Experimental studies with animals exposed to elevated doses of RCF have been conducted using various routes of exposure, inhalation (IH), intratracheal (IT), and intraperitoneal (IP). Most fiber toxicologists believe that the inhalation pathway is the most relevant. Lifetime nose-only inhalation studies on rats and hamsters were conducted at the RCC laboratories (then located in Geneva, Switzerland). These, in particular a multi- dose study on rats, were intended to be the definitive toxicology studies (Mast et al., 1995 a, b). Unfortunately, later analysis revealed (see Mast et al., 2000 a, b; Brown et al, 2005; Yu and Oberdörster, 2000, and contained references) that the test article contained an unusual (and non- representative) ratio of particles to fibers, which exceeded the maximum tolerated dose (MTD), resulted in lung overload, reduced clearance, and an inability to distinguish between the effects of fibers and particles. As noted by Yu and Oberdörster (2000) in a risk analysis prepared for the US Environmental Protection Agency (EPA): “For RCF specifically, it should be kept in mind that the lung and pleural tumor response observed in the chronic rat inhalation study was most likely greater due to the presence of large numbers of non- fibrous RCF particles” and “The contribution of the non-fibrous particles to the rat RCF tumor response cannot be quantitated due to the lack of appropriate data.”

- 102 - The criteria for carcinogen classification (67/548/EEC) note that one important factor to be considered in choosing between categories 2 and 3 even though tumors have been induced in animals is: “carcinogenic effects [observed] only at very high dose levels exceeding the maximal tolerated dose.” Similarly, (1272/2008 (CLP, GHS), it is noted that “the possibility of a confounding effect of excessive toxicity at test doses” is relevant. This is a technical matter which is only briefly explored here. As noted in the introduction, Unifrax and other firms in the industry will be preparing a petition to reconsider the classification of RCF and urge ECHA to defer consideration of the German Dossier until this is resolved.

(Classification and labelling) 43 20091013 Individual (affiliated with This issue is addressed in more detail on The detailed points referred to are found elsewhere in this University of Rochester pages 1-5 of the attached PDF file. document and will be specifically responded where they appear. Medical Center, Academic Because this comment form does not allow institution), United States of for efficient submission of graphics, tables, America footnotes and references, the reader is referred to the attached PDF file for a more complete argument against labeling RCF as a substance of very high concern. An

excerpt of relevant argument follows (refer to the attached PDF for tables, figures and references): Epidemiology Studies Statement, no response necessary As part of its ongoing product stewardship

program the RCF industry in both the United States and Europe has sponsored a series of epidemiological studies beginning in 1987 on occupationally exposed cohorts at the University of Cincinnati in the United States and at the Institute of Occupational

Medicine (IOM) in Europe. This work

- 103 - included evaluation of historical exposures and studies of respiratory symptoms, possible fibrosis, effects on lung function, and mortality. These studies have been published in the peer-reviewed scientific literature and also explicitly considered in the United States National Institutes Occupational Health and Safety (NIOSH)

Criteria Document. To put these findings in perspective, production workers in the University of Cincinnati cohort have up to 41 years of exposure to RCF (15% of all production workers in the cohort have over 20 years of exposure) and no disease above background rates has been observed.

48 -Historical Exposures Maxim et al. (2008 s. above) did investigate the time trends in

On average RCF exposures have RCF exposure. It was found that the rate of improvement in fibre decreased over the years in both US and concentration in RCF manufacturing plants and customer

European plants. The maximum exposure facilities has slowed down in recent years. A significant estimated was approximately 10 fibers per decrease of RCF exposure in the near future is anticipated to be cubic centimeter (f/cc) in one plant in the not very likely.

United States during the 1950s for carding for textile operations; this decreased to less than 1.0 f/cc by the 1990s (Maxim et al.,

2008; Rice et al. 1997; see also Cowie et al.

[2001] and contained references for exposures in European plants). It is appropriate to be mindful of the fact that historical exposures were greater than those at present. Epidemiology results are summarized below for respiratory symptoms, pulmonary function tests, radiological findings, and mortality. The reader should bear in mind that these results reflect occupational exposures that were greater (particularly in the early years)

- 104 - than those likely at present, which implies that health effects will also be smaller in the future. -Respiratory symptoms The respiratory effects are not deciding in the question if RCF In a US study (LeMasters et al., 1998) can cause cancer or not. respiratory symptoms, including chronic cough, chronic phlegm, pleuritic chest pain, shortness of breath, wheezing, and asthma were assessed using the American Thoracic Society (ATS) questionnaire. The major finding from the analysis of respiratory symptom questionnaires was dyspnea (shortness of breath). This symptom occurred more frequently among production employees. Specifically, differences in prevalence of these symptoms between production and non- production workers were statistically significant among males for dyspnea 1, dyspnea 2, and “one or more symptoms.” Additionally, there were differences in prevalence rates in male production workers compared to non-production workers reporting one or more respiratory symptoms; 29.6% and 11.3% respectively with an adjusted odds ratio of 2.9 [95% confidence interval 1.4-6.2] (LeMasters et al. 1998). Such a finding was to be expected: the authors (LeMasters et al., 1998) concluded:

“In general, the prevalence of respiratory symptoms here is similar to that reported in other dust-exposed populations.”

In parallel European studies, Burge et al. (1995) concluded that current exposures to

- 105 - both inspirable dust and respirable fibers were related to dry cough, stuffy nose, eye and skin irritation, and breathlessness. Trethowan et al. (1995) concluded that statistically significant increases were noted in the prevalence of dyspnea (both grades) with increasing cumulative exposure. Cowie et al. (2001) analyzed chronic bronchitis, breathlessness, recurrent chest illness, and pleuritic chest pain as a function of recent and cumulative exposure. Calculated odds ratios were greater than one for each of these, but only statistically significant for recurrent chest illness OR = 1.48 (95% CI 1.11 to 1.96) with cumulative exposure. However, the authors noted that the “prevalence in the study group was low.”

-Pulmonary function results The US cross-sectional study of pulmonary function (LeMasters et al. 1998) reported no statistically significant findings for FEV1/FVC or FEF 25-75. For men, there was a statistically (but not clinically) significant decline in forced vital capacity (FVC) for current and past smokers of 165mL and 156mL, respectively. There was no significant decline among nonsmokers.

Forced expiratory volume in one second (FEV1) showed a statistically significant decline (135mL) only for men who were current smokers. Thus, only those men who worked in RCF production and smoked showed a decline in FVC and FEV1.

- 106 -

A later longitudinal analysis of 361 male production workers who provided 5 or more spirometry tests did not show further declines between the initial test and the final test (Lockey et al. 1998). The authors noted that the fact that no further declines might reflect the fact that exposures have decreased over the years. An expansion of the longitudinal study included workers followed for up to 17 years was completed in 2009. Similar to the findings from the earlier study, no decline in lung function was observed (Dr. James Lockey, Principal Investigator, personal communication).

In Europe, an initial study (Trethowan et al., 1995) found no association between RCF exposure and lung function in nonsmokers. However, there was a significant association between FEV1 and cumulative exposure in past smokers and a non- significant trend in cumulative exposure and FVC for current and past smokers. In the follow-up study (Cowie et al., 2001), the effects were slightly smaller than those seen in 1987 but there were mild decrements in FVC and FEV1 associated with estimated cumulative RCF exposure but only for male current smokers. In addition, there was no reduction in diffusing capacity for carbon monoxide related to exposure, another sensitive test of lung function measured by these investigators.

-Radiological findings The radiological investigation of workers in

- 107 - the US cohort exposed to RCF revealed a statistically significant increase in the prevalence of pleural plaques. Specifically, pleural changes were seen in 27 workers (2.7%). Results from the cumulative exposure analysis demonstrated a significant elevated odds ratio (OR) of 6.0 [95% confidence interval 1.4 to 31.0] (Lockey et al., 2002). The prevalence of parenchymal abnormalities did not differ from workers exposed to other types of dust (Lockey et al. 1996; 2002). Pleural plaques are regarded as a marker of exposure rather than a disease or precursor of disease. Pleural plaques do not cause pain, reduced lung function, and do not progress to either fibrosis or tumors.

The European study of Cowie et al. in 2001 found no association between category 1/0+ opacities and exposure. A weak association between category 0/1+ small opacities and cumulative exposure to RCF was suggested, but not clearly established, and relations over calendar periods were implausible biologically. Pleural changes, after adjustment for age and past exposure to asbestos showed some (but not statistically significant) evidence of a relation with time since first exposure to RCF. In comparing the European findings with a study conducted in the US, pleural plaques were observed in the RCF manufacturing cohort in the US and associated with cumulative exposure to RCF (Lockey et al., 2002). The European study of Cowie et al.

- 108 - (2001) also found some evidence of a relationship between RCF latency and pleural plaques but not with duration or intensity of exposure to RCF. Neither the US nor the European study demonstrated evidence of parenchymal disease. -Mortality Because of the relatively young and small cohorts studied by In 2003 LeMasters et al. reported results of LeMaster et el.(2003) respectively Walker et al. (2002) the a study on the mortality of workers exposed results allow only limited conclusions. Only one positive to RCF in the United States. Current and incidence would produce a substantial effect on the significance former male workers employed between level. 1952 and 2000 at two RCF manufacturing plants were followed to investigate any possible excess in mortality.

There was no significant excess mortality related to all deaths (SMR = 69.8), all cancers (SMR = 94.2), malignancies (SMR = 78.8) or diseases of the respiratory system (SMR = 106.8) including mesothelioma.

LeMasters et al. (2003) also employed Cox’s proportional hazards model (adjusted for age and race); this did not show elevated risk with cumulative RCF exposure (risk ratio 0.99, 95% CI = 0.85 – 1.16). A parallel analysis that lagged exposure by 10 years led to similar conclusions. The authors of the mortality study (LeMasters et al. 2003) provide an extensive discussion of the strengths and weaknesses of the study. Limitations of the study are the relative youth of the cohort and its small size. The mortality analysis did have a 95% power to detect a 2-fold increase in all deaths and all cancers and a

- 109 - 40% power to detect a 2-fold increase in lung cancer. In an earlier analysis of Walker et al. (2002) Because of the relatively young and small cohorts studied designed to explore the statistical power of Walker et al. (2002) the results allow only limited conclusions. the mortality study indicated that the Only one positive incidence would produce a substantial effect experience of lung cancer mortality in the on the significance level. RCF cohort was statistically incompatible with the hypothesis that RCF was as potent as amphibole asbestos assuming identical cumulative exposure to the cohort. However, the possibility that RCF was as potent as chrysotile asbestos could not be excluded. Summary conclusion of study results The results of the epidemiology studies in the United States and Europe (published in peer-reviewed scientific journals) can be summarized briefly as:

• Historical exposures were greater than those found at present. • Exposure to RCF resulted in the development of certain respiratory symptoms, particularly shortness of breath, but this was regarded as anticipated and similar to results found in other dust- exposed populations. • Exposure to RCF resulted in statistically (but not clinically) significant decrements in certain lung function indicators among current or past smokers in a cross-sectional study. However, a later longitudinal study did not demonstrate further decreases in lung function. • Exposure to RCF resulted in an increase in the prevalence of pleural plaques, which do not result in pain, impair pulmonary

- 110 - function, or progress to either fibrosis or tumors. • Exposure to RCF did not result in the development of interstitial fibrosis. • Exposure to RCF did not result in incremental lung cancer or any mesothelioma.

Clinical Interpretation: From a clinical perspective, it is appropriate to note that production workers in the University of Cincinnati cohort have up to 41 years of exposure to RCF (15% of all production workers in the cohort have over 20 years of exposure) and no disease above background rates has been observed. In the many years of worker surveillance, no increased incidence or prevalence of respiratory disease has been detected. Furthermore, the recent study showing that the loss of lung function over 17 years was not accelerated further confirmed that the industry exposure limit of 0.5 f/ml protected workers from lung inflammation and interstitial fibrosis. If fibrosis were occurring at a subclinical level (that is, in the tissue but not yet seen on x- ray), one would have expected lung function to have decreased at a rate more rapid than in non-exposed, healthy individuals. This did not occur.

The RCF industry is mindful of the limitations of these (and other studies) and believes that continued prudence in handling RCF is justified as reflected in the

48 Maxim, D., Allshouse, J., Lentz, T.J., Venturin, D., Walters, T.E., Workplace Monitoring of Occupational Exposure to Refractory Ceramic Fiber- A 17 Year Retrospective , Inhal. Toxicol., 20:289-309, 2008

- 111 - industry’s product stewardship program. And the industry remains committed to continuing these studies. Nonetheless, these results are viewed as encouraging; they certainly don’t justify listing RCF as a SVHC.

(Human health hazard assessment) 44 20091014 PRE - Federation of The substances described in these two Two dossiers were prepared because the substance European Refractory dossiers are also described by a single CAS identification according to REACH does not allow only one Producers, Industry or trade number (142844-00-6) and a single EU dossier. association, Belgium number (604-314-4). PRE does not agree to According to the guidance for identification and naming of treat the two substances separately by substances under REACH UVCB substances are specified with submitting two separate dossiers. the IUPAC-name of their constituents. In the case of Aluminosilicate RCF the main constituents are Al2O3 and SiO2 which both are present more than 10 %. In the case of Zirconia Aluminosilicate RCF the main constituents are Al2O3, SiO2 and ZrO2 which all are present more than 10 % in the UVCB substance. The CAS number 142844-00-6 refers to the CAS name “refractories, fibers, aluminosilicate”, which are defined by CAS as “amorphous man-made fibers produced from the melting and blowing or spinning of calcined kaolin clay or a combination of alumina (Al2O3) and silica (SiO2). Oxides such as zirconia, ferric oxide, titanium oxide, magnesium oxide, calcium oxide and alkalies may also be added. Approximate percentages (by weight) of components follow: alumina, 20-80%; silica, 20-80%; and other oxides in lesser amounts.” This generic description covers a multitude of different substances and is therefore not specific for the Aluminosilicate RCF resp. for the Zirconia Aluminosilicate RCF (see concentration range of the main constituents in chapter 1.2 of Annex XV).

Currently, Refractory Ceramic Fibres (RCF) In March 2006 the European suppliers of Aluminosilicate are classified as a carcinogen 2 (under Refractory Ceramic Fibers provided the EU Technical Directive 67/548) transferred to carcinogens Committee on Classification and Labelling of Dangerous 1b (under the CLP Regulation). This Substances with the currently available scientific data with classification, established in 1997, is based respect to Aluminosilicate Refractory Ceramic Fibre on experiments which have afterwards carcinogenicity. The aim of the suppliers was to re-discuss

- 112 - shown to be using a particulate overload. carcinogenicity. France and Germany had responded to that Therefore, the current classification can no documentation that a re-discussion was not warranted as the longer be supported. Several new studies new studies do not have any impact on the existing classification carried out after 1997 all support a re- as inhalative carcinogen. The Technical Committee on classification. In 2002, IARC already Classification and Labelling of Dangerous Substances assigned RCF as a category 2b. Brown et concluded that a re-discussion will only take place in case a al. (2005) summarised the new findings and member state would support such re-discussion. There was no concluded that “In particular, in the support by any member state. Since 2006, no further relevant European classification system, Category scientific data have become available. The existing classification 3(b) seems more appropriate than Category with R49 or H350i is still adequate and does not need to be re- 2 for RCF.” Based on this information, PRE evaluated. strongly requests that any regulatory action on RCF is preceded by a discussion and decision on the appropriate classification of RCF. Any Annex XV dossier based on the current classification should be withdrawn in the meantime. The test-materials described in the two i) The general requirement for an experimental inhalation test reports are dust samples which are is that the sample needs to be respirable for the animals obtained after grinding of the fibre. This tested. Thus, the sample has to be prepared accordingly. grinding shortens the length of the fibre thereby changing its properties. It is ii) In handling and use RCF wools release fibres which are estimated that the length of the test-material respirable for humans (similar dimensions to those used in is hundred times shorter than the the animal tests). Thus, such fibres are relevant with respect commercial fibre! The results of tests based to a putative human carcinogenicity. on such samples are therefore only valid for the test itself, they cannot be extrapolated to the commercial fibre.

(Classification and labelling) 45 20091014 Individual (affiliated with This issue is addressed in more detail on The detailed points referred to are found elsewhere in this Harvard School of Public page 1 of the attached PDF file. Because document and will be specifically responded where they appear. Health, Academic this comment form does not allow for LeMasters (2003) stated in his article the limitations because of institution), United States of efficient submission of references, the the relatively small and young cohort by himself. America reader is referred to the attached PDF file His personal announcement is of high interest but can’t be taken for a more complete argument against into account. labeling RCF as a substance of very high concern. An excerpt of relevant argument follows (refer to the attached PDF for

- 113 - references):

Details LeMasters and colleagues at the University of Cincinnati have undertaken an occupational cohort study of RCF workers with first employment from 1952 through 1997 (LeMasters, 1994). The subjects are 927 men with at least one year of employment in an RCF manufacturing facility. This cohort includes men who were former workers when the study was initiated; retirees continue to be followed. An analysis in 2000 of cumulative mortality through 1998 found no cases of mesothelioma and six cases of lung cancer, the latter giving a standardized mortality ratio (SMR) of 64 and the former giving an estimated risk of 0 (LeMasters, 2000). These data were included in a later publication of mortality through 2000 (LeMasters, 2003). In the published paper, the SMR for cancer of the trachea, bronchus and lung was 83, with 95% confidence interval 38 to 159, based on nine cases. There were still no cases of mesothelioma, giving a risk of 0, and there have been no subsequent cases through October 2009 (LeMasters 2009). LeMasters and colleagues (1994, 2003; Rice, 2005) have made detailed historical studies and contemporaneous measures to estimate date-specific and cumulative exposure to RCF in each member of the University of Cincinnati cohort. Walker et al., (2002) combined the RCF exposure information, the mortality data through 1997, and estimates of the risk associated with asbestos of different fiber types and

- 114 - dose levels. Using established risk models for the relation between asbestos dose, exposure duration and mortality from cancers of the lung and mesothelioma, they calculated the numbers of cancers that would be expected to have occurred under the hypothesis that RCF exposures had effects similar to those of asbestos. The purpose was to ask whether observations could support or reject the hypothesis of a carcinogenic equivalence between asbestos and RCF. The asbestos risk equations involve coefficients that have to be derived from observed cohorts. Walker et al. (2002) used two sources. First were summary estimates from a review carried out by Hodgson and Darnton (2000), who had concluded that crocidolite, amosite and chrysotile asbestos had orders-of- magnitude differences in their best-fitting coefficients and consequently in their degree of carcinogenicity. Second, Walker et al. (2002) took coefficients estimated directly from major asbestos cohorts, without an effort to distinguish fiber type. Using the estimates taken from Hodgson Because of the relatively young and small cohorts studied by and Darnton (2000), Walker et al (2002) Walker et al. (2002) the results allow only limited conclusions. found that the six observed cases of lung Only one positive incidence would produce a substantial effect cancer through 1998 was low enough to on the significance level. reject the hypothesis that RCF is as potent a lung carcinogen as amphibole asbestos, under which 11.8 deaths were expected (p=0.04). The Hodgson-Darnton estimate of the lung carcinogenicity of chrysotile asbestos is very nearly nil, and the observed data could not reject that chrysotile-RCF equivalence under that hypothesis.

- 115 -

To compare the University of Cincinnati RCF cohort directly to various asbestos cohorts, Walker et al. (2002) used 20 coefficients calculated for asbestos cohorts by various expert committees and compiled by the Health Effects Institute – Asbestos Research project (HEI-AR 1991). Observed lung cancer rates in RCF workers fell below those predicted by all 20, and the difference was statistically significant (p ≤ 0.05) against five different predictions based on the reports of Dement (1983), Finkelstein (1983) and Seidman (1984).

Under all models, the predicted number of deaths from mesothelioma was sufficiently low that the lack of cases in the RCF cohort experience provided no differentiation between positive and null hypothesis. What does the RCF cohort’s lack of any The cumulative exposure in the currently available excesses of lung cancer or mesothelioma epidemiological studies is far too low to detect a putative RCF mean, when set against predictions of carcinogenicity in humans. Thus, classification and risk carcinogenicity based on extrapolation from assessment has to be based on experimental results. animal models and physiochemical assumptions? Mechanistic understanding is critical to quantitative extrapolation from high-dose animal studies to humans (Vu 1996), and especially for substances that are not genotoxic, mechanisms of carcinogenesis are not well understood. By contrast, the RCF worker experience is not a theoretical construct, and no assumptions other than accurate measurement and follow-up are required before the data enter into evidence. If the experience of the workers is incompatible with theoretical predictions, then the theory needs to be altered.

- 116 -

(Human health hazard assessment) 46 20091015 RIVM - Bureau REACH, Most information on carcinogenicity is In our mind the comment is incorrect. National Authority, The presented in Chapter 3, but not under 5.8. It Netherlands is proposed to refer to chapter 3 in the During the accordance check of the dossier it was ECHAs section on carcinogenicity in chapter 5.8. opinion not to repeat information at different places in the dossier. Therefore, information on the carcinogenicity of the (Human health hazard assessment) substance are presented in chapter 3 rather than in 5.8. 47 20091015 Member State Competent Page 5: Scope of the proposal. The UK The annex VI classifies vitreous fibres with an alkaline oxide and Authority, United Kingdom seeks clarification on the scope of the alkali earth oxide content of less than 18% as carc. 1B. In the proposal. Page 5 states that the definition literature two relevant vitreous fibres are described which fulfil given in CLP Annex VI for Refractory this condition: RCF (SiO2, Al2O3) and RCF (SIO2, Al2O3, Ceramic Fibres or Special Purpose Fibres ZrO2). Germany follows the Guidance for identification and (fibres with a content of 18% of weight or naming of substances under REACH for the definition of these less of Na2O+K2O+CaO+MgO+BaO) is in fibres, i.e. the given oxides can formally be identified as the fact only met by RCF. Is Germany able to main constituents of these two fibres. According to this guidance confirm that no other fibres are document the identity of a substance is not dependent on the manufactured or imported into the EU that function or the property of individual constituents but only on meet this definition? We also note that the their content. dossier does not mention the term ‘vitreous (silicate)’ and so request confirmation that the scope of the substance as identified in the dossier is the same as that of the CLP Annex VI entry. The UK’s understanding is that RCF are glasses and are described by the absence of network-modifying elements; the classification is not affected by the ratio of network formers (alumina, zirconia, silica). The UK is therefore not clear why two separate proposals have been submitted for RCF. Page 6: The identity of the substance In handling and use RCF wools release fibres which are defined by the proposal. The dossier respirable for humans (similar dimensions to those used in the identifies the substance covered by the animal tests). Thus, such fibres are relevant with respect to a proposal as RCF types 1 and 3. The UK’s putative human carcinogenicity. understanding is that RCF types 1 and 3 refer to materials that were produced specifically to enable testing in animals (fine fractions of dust that are rodent respirable)

- 117 - and thus do not reflect the commercially- available materials. The information given on the physical characteristics is therefore misleading. Page 7: The composition of the material. The composition of the substance does refer on constituents. Section 1.2 titled ‘Composition of the Aluminosilicate RCF resp. Zirconia Aluminosilicate RCF are both substance’ actually refers to the UVCB substances. According to the guidance for identification components from which the substances are and naming of substances under REACH UVCB substances are manufactured, not the composition of the specified with the IUPAC-name of their constituents. In the case substances that are supplied commercially. of Aluminosilicate RCF the main constituents are Al2O3 and SiO2 This section should therefore be amended which both are present more than 10 %. In the case of Zirconia to include the composition of the RCF Aluminosilicate RCF the main constituents are Al2O3, SiO2 and covered by the proposal. ZrO2 which all are present more than 10 % in the UVCB substance. 48 20091015 Member State Competent - France supports the proposal for Statement, no response necessary Authority, France identification of RCF as SVHC and agrees the need to separate RCF into two types of substances according to REACH rules for identification of substances. - RCF are multi-constituent substances RCFs are no multi-constituent substances but UVCB under REACH definition and the guidance substances. According to the guidance for identification and for identification and naming of substances naming of substances under REACH UVCB substances are under REACH states that the name of the specified with the IUPAC-name of their constituents. In the case constituents are given in the order of typical of Aluminosilicate RCF the main constituents are Al2O3 and SiO2 concentration percentages starting with the which both are present more than 10 %. In the case of Zirconia highest. Given that this is not followed for Aluminosilicate RCF the main constituents are Al2O3, SiO2 and the naming of the substance in the annex ZrO2 which all are present more than 10 % in the UVCB XV report, a potential incorrect substance. interpretation of the name should be The cited ranges of the constituents refers on the composition of avoided. This is specifically important in this RCF-1 and RCF-3 given in TIMA (1991). This literature is case as the substance has no CAS number considered as valide. The table with typical chemical and is therefore defined by its chemical compositional ranges for classes of MMVF presented in IARC composition and its physical properties. 2002 does not differ between RCF 1 and 4 and gives a range for Other sources of information (IARC, 20021 all RCFs. NIOSH (2006) gives ranges adapted from Mast et al and NIOSH, 20062) reports RCF of (1995); it is not the original source. chemical composition outside the range given in the annex Xv report for the main RCF constituents. This information should be checked to avoid that some RCF are not

- 118 - covered by this SVHC identification. - In complement to the conclusions of IARC If note R can be applied, it applies. This does not have any on the available epidemiological data impact on the authorisation proposal. mentioned in the annex XV report, IARC concluded that experimental data provide sufficient evidence of carcinogenicity of RCF. These experimental data are the basis of the justification of the RCF European classification as carcinogen category 2 and it should be considered that they fully support the relevance of the RCF SVHC identification.

- Finally, it should be noted that note “R” is present in the entry relevant for RCF in Annex VI of the CLP Regulation (index n°650-017-00-8). RCF are therefore exempted of carcinogenic classification when their “length weighted geometric mean diameter less two standard geometric errors” (note R definition in annex VI of CLP regulation) is greater than 6 μm. Although this is unlikely in the case of RCF, it should be reminded to clearly define what is covered by the carcinogenic classification and therefore what is covered by the SHVC dossier.

1IARC (2002). Monographs on the evaluation of carcinogenic risks to humans: Man-made vitreous fibres, Vol 81. IARC press. 2NIOSH (2006). Occupational exposure to refractory ceramic fibers. Publication.

(Identity of the substance and physical and chemical properties) 49 20091015 European Industrial Gases Information on alternative substances and EIGA states: “There is no viable substitute material available for Association (EIGA), techniques (see attachment) RCF in plant such as Steam Methane Reforming where the

- 119 - Industry or trade fibres are exposed to high temperatures (900 - 1200°C), in a association, Belgium reducing atmosphere for a long duration (several months)” And “  (…) AES fiber is an insulation material which doesn't have similar thermal properties to RCF for high temperature applications as it can permanently be used only at temperatures below 800°C. Installed in furnaces with fluegas temperatures of more than 800°C (e.g. steam methane reforming - SMR) it will be destroyed soon (shrinkage, embrittlement, vitification) – apart from the formation of cristoballite.”

Up to our knowledge there are alternatives. (For details see response to SELAS-LINDE GmbH, Company, Germany 20091005, above)

EIGA states: “It should not be considered that replacing one insulation material with an apparently similar one is trivial and therefore makes any change straightforward. Any change will involve considerable cost and should only be enforced where alternatives are well proven.”

At least one of the identified alternatives (For details see response to SELAS-LINDE GmbH, Company, Germany 20091005, above) was successfully tested during long time in the different industries. For example due to reduction of needed maintenance and repair working and protection equipment i.e. expensive respiratory protection, exhaust system etc. could save expenses. Eiga states: The current furnace designs are based on fibre insulation. Alternative materials are either brick or castable refractory linings, which have significantly higher thermal conductivities and significantly greater weight. These two factors would result in a substantial re-design effort and significantly higher capital cost. (…) For new plants, a change to brick or castable refractory linings from ceramic fibre would impact almost all aspects of the furnace design and add millions of euro’s of additional cost. The cost to retrofit existing furnaces, if possible, would similarly be millions of euro’s in addition to a potential loss of tens of millions

- 120 - of euro’s in revenue over the plant life associated with a loss of plant capacity.” Capital costs can only be incurred by exchanging of the insulating material after service life of RCF, if a new insulating material is needed. Instead of RCF for lining other alternative materials could be used. One of them (Microtherm) is even thinner and takes lower space in the interior of a kiln/furnace. (For details see response to SELAS-LINDE GmbH, Company, Germany 20091005, above) This company assumes that after including of RCF in Annex XIV, all RCF must be removed from industrial processes and kiln/furnaces. This implies that the existing kiln/furnace would have to be re-designed. This is a mis-lading interpretation. Just new kiln/furnaces would be lined with new insulating materials. Therefore new alternative materials could be integrated into design of kiln/furnace etc. On the one hand capital cost could be higher due to prices for new materials, but this would occur temporarily. After increasing of the market share of new materials economy of scale, economy of scope and learning curve effect could be achieved. Therefore, the production costs per item of new materials are expected to drop.

- 121 - INFORMATION ON USE, EXPOSURE, ALTERNATIVE AND RISKS ON ANNEX XV SVHC

Substance name: Aluminosilicate Refractory Ceramic Fibres CAS number: - EC number: -

Reason of the submission of the Annex XV: CMR

Specific comments on use, exposure, alternatives and risks No. Date Submitted by (name, Comment Response Organisation/MSCA) 50 20090928 Air Liquide, Company, Evolution of the toxicity according to Up to our knowledge, it is unquestionable that cristobalite is France operating conditions formed from AES as well as from RCF (Class, 200249) in a recristallization procedure above 900 °C. In both insulating Some samples of AES fibres, potential materials cristobalite only occurs after its using in the high- substitute for RCF have been placed in a temperature range. For RCFs Class50 (2003) has shown that Steam Methane Reforming furnace for cristobalite exposure in the course of removal operations can more than one year to evaluate the exceed the limit value used in a number of countries (0.05 evolution of their properties under real mg/m3). The average concentration as presented in the quoted operating conditions. After one year, AES literature is about 0.096 mg/m3. The proposed OSHA limit value 3 51 fibres, initially vitreous, crystallize into silica for crystalline SiO2 is 0.025 mg/m (ICSC 0808 ). According to (cristobalite). The main interest of AES a literature search at present no published data regarding the fibres is their high level of biosolubility that exposure level of cristobalite from AES fibres is available. has been demonstrated only at a vitreous state. To our knowledge very limited data The difference between AES and RCF wool products is obvious: on biosolubility after crystallization are AES wool is not classified as carcinogenic, but RCF products available and the biosolubility of AES fibres, are. AES wool is safe in applications like installation, production exposed to high temperature, reducing of downstream articles etc. as long as it is not used above 900 atmosphere for long duration has not been °C. evaluated. Besides the opportunity to use AES products as an alternative to Moreover, the International Agency for RCFs there are other potent substitutes available on the market

49 P. Class, R.C. Brown;Exposition gegenüber künstlichen Mineralfasern, Gefahrstoffe – Reinhaltung der Luft; 62, Nr. 5, 2002 50 Ph. Class; Current Fibrous Dust Workplace Concentrations and Trend in the High Temperature Insulation Wool Industry: the Results of the Care programme, VDI-Bericht Nr. 1776, 2003 51 ICSC: International Chemical Safety Card No. 0808, Crystalline silica, quartz Crystalline silicon dioxide, quartz Silicic anhydride SiO2 , Molecular mass: 60.1

- 122 - Research on Cancer has classified which lack the problem of cristobalite formation. (For details see crystalline silica and especially cristobalite response to SELAS-LINDE GmbH, Company, Germany as carcinogenic 1 (evidence of carcinogenic 20091005, above) effect on human). Air Liquide has proven that cristobalite formation from AES fibres crystallization cannot be avoided under Steam Methane Reforming operation. Substitution of RCF by AES is equivalent to the substitution of a potentially carcinogenic material (RCF) by a carcinogenic material (crystallizes AES= crystalline silica), as classified by the International Agency for Research on Cancer. The regulation does not take into account AES fibres crystallization for the choice of substitute for RCF. Moreover to our knowledge, AES is the only substitute that has been validated (validation has been done on vitreous fibres). For Steam Methane Reforming application, AES fibres are vitreous only during installation, and then crystallization occurs. According to the regulation, crystallized fibres should not be considered anymore as they are no more vitreous. The regulation on crystalline silica should be applied. The methodology described in the norm to count fibres uses a microscope able to differentiate crystalline and vitreous fibres. However the procedure does not impose such identification.

Level of exposure measurement :

Workers on Steam Methane Reforming units will be exposed to several kinds of materials submitted to exposure limit (fibres, crystalline silica) To evaluate the

- 123 - levels of exposure the following data are required: 1. the number of vitreous breathable fibres (limit in France 0.1 f/ml). Crystalline fibres should not be taken into account 2. the nature of the fibres (crystalline, vitreous, chemical composition) : AES should not be taken into account whereas RCF should be. 3. Crystalline silica amount (cristobalite, quartz, tridimyte) submitted to an exposure limit : 0.05 mg/m3 for cristobalite. The norm proposed to count breathable fibres using phase contrast optical microscope (X PX 43-269) but does not impose to differentiate crystalline and vitreous fibres. However the regulation is only dedicated to vitreous fibres. Crystalline silica quantification using XRD is not accurate because the detection limit is above 6%vol and the accuracy is very bad especially when close phases are measured. Several percents errors can be obtained. Norms dedicated to fibres counting and crystalline silica quantification are not adapted to measure the data necessary to establish exposure levels of workers. Moreover the list of authorized laboratories for measurements has not been published.

(Information on alternative substances and techniques) 51 20091005 Lurgi GmbH, Company, Evolution of the toxicity according to Up to our knowledge, it is unquestionable that cristobalite is Germany operating conditions formed from AES as well as from RCF (Class, 200252) in a recristallization procedure above 900 °C. In both insulating

52 P. Class, R.C. Brown;Exposition gegenüber künstlichen Mineralfasern, Gefahrstoffe – Reinhaltung der Luft; 62, Nr. 5, 2002

- 124 - Some samples of AES fibres, potential materials cristobalite only occurs after its using in the high- substitute for RCF have been placed in a temperature range. For RCFs Class53 (2003) has shown that Steam Methane Reforming furnace for cristobalite exposure in the course of removal operations can more than one year to evaluate the exceed the limit value used in a number of countries (0.05 evolution of their properties under real mg/m3). The average concentration as presented in the quoted operating conditions. literature is about 0.096 mg/m3. The proposed OSHA limit value 3 54 After one year, AES fibres, initially vitreous, for crystalline SiO2 is 0.025 mg/m (ICSC 0808 ). According to crystallize into silica (cristobalite). The main a literature search at present no published data regarding the interest of AES fibres is their high level of exposure level of cristobalite from AES fibres is available. biosolubility that has been demonstrated only at a vitreous state. To our knowledge The difference between AES and RCF wool products is obvious: very limited data on biosolubility after AES wool is not classified as carcinogenic, but RCF products crystallization are available and the are. AES wool is safe in applications like installation, production biosolubility of AES fibres, exposed to high of downstream articles etc. as long as it is not used above 900 temperature, reducing atmosphere for long °C. duration has not been evaluated. Moreover, the International Agency for Besides the opportunity to use AES products as an alternative to Research on Cancer has classified RCFs there are other potent substitutes available on the market crystalline silica and especially cristobalite which lack the problem of cristobalite formation. (For details see as carcinogenic 1 (evidence of carcinogenic response to SELAS-LINDE GmbH, Company, Germany effect on human). 20091005, above) Air Liquide has proven that cristobalite formation from AES fibres crystallization cannot be avoided under Steam Methane Reforming operation. Substitution of RCF by AES is equivalent to the substitution of a potentially carcinogenic material (RCF) by a carcinogenic material (crystallizes AES= crystalline silica), as classified by the International Agency for Research on Cancer. The regulation does not take into account AES fibres crystallization for the choice of substitute for RCF. Moreover to our knowledge, AES is the only substitute that has been validated (validation has been

53 Ph. Class; Current Fibrous Dust Workplace Concentrations and Trend in the High Temperature Insulation Wool Industry: the Results of the Care programme, VDI-Bericht Nr. 1776, 2003 54 ICSC: International Chemical Safety Card No. 0808, Crystalline silica, quartz Crystalline silicon dioxide, quartz Silicic anhydride SiO2 , Molecular mass: 60.1

- 125 - done on vitreous fibres). For Steam Methane Reforming application, AES fibres are vitreous only during installation, and then crystallization occurs. According to the regulation, crystallized fibres should not be considered anymore as they are no more vitreous. The regulation on crystalline silica should be applied.

The methodology described in the norm to count fibres uses a microscope able to differentiate crystalline and vitreous fibres. However the procedure does not impose such identification.

Level of exposure measurement :

Workers on Steam Methane Reforming units will be exposed to several kinds of materials submitted to exposure limit (fibres, crystalline silica) To evaluate the levels of exposure the following data are required: 1. the number of vitreous breathable fibres (limit in France 0.1 f/ml). Crystalline fibres should not be taken into account 2. the nature of the fibres (crystalline, vitreous, chemical composition) : AES should not be taken into account whereas RCF should be. 3. Crystalline silica amount (cristobalite, quartz, tridimyte) submitted to an exposure limit : 0.05 mg/m3 for cristobalite. The norm proposed to count breathable fibres using phase contrast optical microscope (X PX 43-269) but does not impose to differentiate crystalline and vitreous fibres. However the regulation is only dedicated to vitreous fibres.

- 126 - Crystalline silica quantification using XRD is not accurate because the detection limit is above 6%vol and the accuracy is very bad especially when close phases are measured. Several percents errors can be obtained.

Norms dedicated to fibres counting and crystalline silica quantification are not adapted to measure the data necessary to establish exposure levels of workers. Moreover the list of authorized laboratories for measurements has not been published.

(Information on alternative substances and techniques) 52 20091012 Individual (affiliated with Exposure Information The ranking order of the measures should be complied with. In a Howmet-Ciral snc, Howmet Employee exposure is managed by the use first step the substitution of RCF must be taken into account. SAS; Howmet Ltd for Alcoa of Engineering controls, Personal Protective There are options for alternatives available (see Dossier and Power and Propulsion, a Equipment, and specific work procedures in other comments). Only following that technical and division of ALCOA, accordance with the French regulatory organisational measures could be considered. Just as a last Company), France requirements as stipulated in 27th October resort PPE can be used to control risks for workers (STOP 2007 decree n°2007-1539 that is based on principle). Furthermore, the effectiveness of all measures has to the 91/322/CE, 2000/39/CE and be proven. 2006/15/CE European directives. In the UK the use of RCF, is handled in accordance with control measures, under the Control of Substance Hazardous to Health Regulations 2002 (COSHH Regulations), Carcinogens Approved Code of Practice (COSHH Carcinogens Approved Code of Pratice – AcoP). Our facilities have a strict Refractory Ceramic Fiber use and handling program. Our primary management system is the use of engineering controls. a) RCF cutting tables and mold wrap stations are equipped with lateral local exhaust ventilation designed in accordance

- 127 - French, UK, or EU recognized standards. b) Shell removal operations are performed in a local exhaust hood / enclosure designed in accordance with French, UK, or EU recognized standards. Ample water is applied to soak the RCF portions of the mold. c) Permanent or temporary RCF Work Areas are established to preclude significant airborne exposures and/or direct contamination of personal garments for others who may visit RCF work areas. A secondary layer of protection is afforded our workers by the use of appropriate personal protective equipment. a) Respiratory protection, CE approved to relevant European Respiratory protection standards, is used by properly trained employees. Respiratory protection is selected with adequate protection factors to reduce employee exposures to levels below the French occupational exposure limit of 0.1 fibre/cc. b) Gloves and other personal protective equipment are used to prevent skin contact with RCF thus reducing the risk for irritation and contact dermatitis. c) Company supplied clothing is used to prevent contamination of street clothing.

A third layer of protection is the utilization of carefully thought out work practices. a) If practical, pre-cut, pre-sized or encapsulated RCF products are purchased. For example, product thickness is specified in order to eliminate horizontal cutting or ripping of RCF blanket. b) RCF products are moistened with water (fine spray) prior to cutting, removal or

- 128 - installation.

c) HEPA vacuums are used to remove dust accumulations from equipment and work surfaces. Vacuuming is performed at the end of each workday or more frequently if indicated. The use of compressed air or sweeping is prohibited.

Disposal of waste The management of RCF waste complies with French Decree (n°2007-396) Appendix II R541-8 and Article R541-43. In the UK In order to minimize RCF exposure related to waste storage, transportation, and disposal, all waste that is generated, and contains RCF, is considered RCF containing material. RCF waste is placed in either two (2) 6 mil or one (1) 12 mil polyethylene bag(s) and sealed. All bags are conspicuously labeled as follows:

RCF waste is segregated from other waste in dedicated containers. Waste containing hard, sharp edges (i.e., pieces of metal) are placed in containers such as metal or fiberboard drums. Containers are tightly sealed during storage and transportation and roll-off containers are securely tarped before shipment, as hazardous waste. 53 20091013 Unifrax Corp., Company, Exposure data and benchmarks The cumulative exposure in the currently available United States of America The German Dossiers presents data on epidemiological studies is far too low to detect a putative RCF occupational exposures to RCF and display carcinogenicity in humans. Thus, classification and risk these data graphically with reference to a assessment has to be based on experimental results. “benchmark” of 0.1 f/ml (the occupational The different RCF risk assessments quoted by industry exposure limit [OEL] for asbestos in many (DECOS, 1995; Fayerweather et al; 1997; Moolgavkar; 1999; Yu countries). I believe that the impression and Oberdörster, 2000; Turim and Brown, 2003) are considered conveyed by this comparison (see Fig. 1 of as inadequate. For instance, all of those risk assessments were the SVHC Dossier reproduced below) is based on the results of insensitive rat chronic inhalation data.

- 129 - misleading for the following reasons: This is an explanation why all of those lead to unrealistically low cancer risks for RCF. In addition, the assessments were partly • The benchmark concentration is too low: based on unlikely assumptions e.g. like similar tumour initiation The German Dossiers derive the rates between rat and humans (Moolgavkar et al., 1999 which benchmark based on a stated “tolerance builds upon the Fayerweather et al. approach). Applying the level” risk of 4:1000. Next, the SVHC Moolgavkar et al. approach for asbestos would lead to cancer Dossier calculates the RCF fiber risk estimates for asbestos which are 2-3 orders of magnitude concentration corresponding to a 4:1000 lower than the risk assessment for asbestos by US EPA tolerance level risk using data derived from (1988/2008). IP experiments of RCF and asbestos. (The validity and uncertainty of this procedure is reviewed below.) This method calculates that working lifetime occupational exposure to 0.1 f/ml RCF would result in an incremental cancer risk of approximately 4:1000. Working lifetime risks associated with occupational exposure have been studied by several investigators (see below) and, although specific estimates vary, these published (peer reviewed) risk estimates indicate that the workplace concentration corresponding to a risk of 4:1000 is likely to be substantially greater than 0.1 f/ml. • The exposure data are not current: The Since the more recent information on exposure levels (Maxim, exposure data given in the SVHC Dossier 200855) have not been given in a numerical but in a graphical (furnished by European Ceramic Fiber form they have not been considered in the dossier. Gaining Industry Association [ECFIA] in 1999) refer statistical values from the diagrams presented in Maxim (2008) to earlier European data. The RCF industry was not considered appropriate due to a considerable loss in maintains a product stewardship program precision. that, among other goals, attempts to reduce occupational exposure to RCF and, over Only exposure data from peer reviewed publications have been time, exposures have decreased as a result taken into account. Due to this quality criteria the Care study of these efforts. Therefore, it is appropriate (ECFIA, 199956) was the most current publication on to use more recent monitoring data. measurement data. Since the exposure data submitted by the Moreover, for the most part, the companies commentator in the pdf-file (RCF SVHC Comments.pdf) was not that manufacture RCF in Europe are the published in a journal (in a numerical form) it could not be found

55 Maxim, D., Allshouse, J., Lentz, T.J., Venturin, D., Walters, T.E., Workplace Monitoring of Occupational Exposure to Refractory Ceramic Fiber - A 17 Year Retrospective, Inhal. Toxicol., 20:289-309, 2008 56 ECFIA (1999); Idenfication and control of exposure to refractory ceramic fibres. European Chemical Fibre Industry Association, 3 rue de Colonel Moll, 75017 Paris, France, Nov. 1999, 58p Illus. 6 ref.

- 130 - same as those in the United States; in a literature search. processes and controls are similar. It is appropriate to utilize all the relevant data for We agree that for the most part, the companies that comparisons. manufacture RCF in Europe are the same as those in the US since processes and controls are similar. It may therefore be appropriate to utilize all the relevant data for comparison.

Maxim et al. (2008 s. above) also investigated the time trends in RCF exposure. It was found that the rate of improvement in fibre concentration in RCF manufacturing plants and customer facilities has slowed down in recent years. A significant decrease of RCF exposure in the near future is anticipated to be not very likely. This issue is addressed in more detail on The detailed points referred to are found elsewhere in this pages 5-13 of the attached PDF file titled document and will be specifically responded where they appear. “RCF SVHC.pdf”. Because this comment form does not allow for efficient submission of graphics, tables, and references, the reader is referred to the attached PDF file for a more complete argument against labeling RCF as a substance of very high concern. An excerpt of relevant argument follows (refer to the attached PDF for tables, figures and references): Exposure data and benchmarks The cumulative exposure in the currently available The German Dossiers presents data on epidemiological studies is far too low to detect a putative RCF occupational exposures to RCF and display carcinogenicity in humans. Thus, classification and risk these data graphically with reference to a assessment has to be based on experimental results. “benchmark” of 0.1 f/ml (the occupational The different RCF risk assessments quoted by industry exposure limit [OEL] for asbestos in many (DECOS, 1995; Fayerweather et al; 1997; Moolgavkar; 1999; Yu countries). I believe that the impression and Oberdörster, 2000; Turim and Brown, 2003) are considered conveyed by this comparison (see Fig. 1 of as inadequate. For instance, all of those risk assessments were

- 131 - the SVHC Dossier reproduced below) is based on the results of insensitive rat chronic inhalation data. misleading for the following reasons: This is an explanation why all of those lead to unrealistically low • The benchmark concentration is too low: cancer risks for RCF. In addition, the assessments were partly The German Dossiers derive the based on unlikely assumptions e.g. like similar tumour initiation benchmark based on a stated “tolerance rates between rat and humans (Moolgavkar et al., 1999 which level” risk of 4:1000. Next, the SVHC builds upon the Fayerweather et al. approach). Applying the Dossier calculates the RCF fiber Moolgavkar et al. approach for asbestos would lead to cancer concentration corresponding to a 4:1000 risk estimates for asbestos which are 2-3 orders of magnitude tolerance level risk using data derived from lower than the risk assessment for asbestos by US EPA IP experiments of RCF and asbestos. (The (1988/2008). validity and uncertainty of this procedure is reviewed below.) This method calculates that working lifetime occupational exposure to 0.1 f/ml RCF would result in an incremental cancer risk of approximately 4:1000. Working lifetime risks associated with occupational exposure have been studied by several investigators (see below) and, although specific estimates vary, these published (peer reviewed) risk estimates indicate that the workplace concentration corresponding to a risk of 4:1000 is likely to be substantially greater than 0.1 f/ml. • The exposure data are not current: The We agree that for the most part , the companies that exposure data given in the SVHC Dossier manufacture RCF in Europe are the same as those in the US (furnished by European Ceramic Fiber since processes and controls are similar. It may therefore be Industry Association [ECFIA] in 1999) refer appropriate to utilize all the relevant data for comparison. to earlier European data. The RCF industry maintains a product stewardship program Maxim et al. (2008 s. above) investigated the time trends in RCF that, among other goals, attempts to reduce exposure. It was found that the rate of improvement in fibre occupational exposure to RCF and, over concentration in RCF manufacturing plants and customer time, exposures have decreased as a result facilities has slowed down in recent years. A significant of these efforts. Therefore, it is appropriate decrease of RCF exposure in the near future is anticipated to be to use more recent monitoring data. not very likely. Moreover, for the most part, the companies We would therefore really appreciate monitoring data showing that manufacture RCF in Europe are the that the RCF exposure at workplaces is below the tolerance same as those in the United States; level of 0.1 fibres/ml. processes and controls are similar. It is appropriate to utilize all the relevant data for

- 132 - comparisons. These points are detailed in the following For risk analysis the dossier does not rely on chronic rat sections. inhalation studies.

-Risk analyses In principle, data on occupationally exposed human populations provide the most compelling evidence on which to estimate risks and set an OEL. As noted above, the RCF industry has sponsored such studies (including symptoms, morbidity, and mortality) in the United States (University of Cincinnati) and Europe (the Institute of Medicine) beginning in 1987. As noted above, to date these studies on workers exposed beginning as early as 1952 have not found evidence of interstitial fibrosis, incremental lung cancer, or any mesothelioma in RCF-exposed cohorts. Based on these results the incremental risk posed by historical occupational exposures (which were greater than those at present) is zero. While encouraging, these data cannot be used directly to estimate actual risks. Because the available epidemiological data are limited, it has been necessary to rely on data from experiments with laboratory animals for risk analysis. It is difficult to extrapolate results from animal studies to humans; rat physiology differs from that for humans (respiration rates, which fibers are respirable, how lung burden should be measured, and comparative dosimetry). Additionally, what were believed to be state-of-the-art inhalation studies conducted at the RCC laboratories in Geneva, used a specially prepared rat-respirable RCF fiber that contained a high (and non-

- 133 - representative) ratio of particles to fibers— sufficient to produce overload in rodent lungs. Lung overload in rodents reduces fiber clearance rates and has been demonstrated to induce tumors as well. Moreover, in order to develop risk estimates using a feasible number of laboratory animals, these animals are exposed to levels (up to ~200 f/ml) that are hundreds of times greater than typical occupational exposures today. Thus, extrapolation from these studies is likely to overstate possible disease in human populations exposed to much lower doses. These challenges noted, the available rat The different RCF risk assessments quoted by industry inhalation data have been analyzed by (DECOS, 1995; Fayerweather et al; 1997; Moolgavkar; 1999; Yu several investigators (DECOS, 1995; and Oberdörster, 2000; Turim and Brown, 2003) are considered Fayerweather et al., 1997; Maxim et al., as inadequate. For instance, all of those risk assessments were 2003; Moolgavkar et al., 1999, 2000, based on the results of insensitive rat chronic inhalation data. 2001a, b; Turim and Brown, 2003; Yu and This is an explanation why all of those lead to unrealistically low Oberdörster, 2000 ). These studies were cancer risks for RCF. In addition, the assessments were partly sponsored by RCF producers and other based on unlikely assumptions e.g. like similar tumour initiation organizations, such as Dutch Expert rates between rat and humans (Moolgavkar et al., 1999 which Committee on Occupational Standards builds upon the Fayerweather et al. approach). Applying the [DECOS] and the US Environmental Moolgavkar et al. approach for asbestos would lead to cancer Protection Agency (EPA), used a variety of risk estimates for asbestos which are 2-3 orders of magnitude models and alternative assumptions, and lower than the risk assessment for asbestos by US EPA were reviewed carefully by the National (1988/2008). Institute for Occupational Safety and Health To establish an REL for RCFs, (NIOSH) in its Criteria for a Recommended NIOSH took into account not only the animal and human health Standard Occupational Exposure to data but also exposure information describing the extent to Refractory Ceramic Fiber which RCF exposures can be controlled at different workplaces. (http://www.cdc.gov/niosh/docs/2006- On the basis of this evaluation, NIOSH considered an REL of 123/pdfs/2006-123.pdf), which set a 0.5 f/cm3 recommended exposure limit (REL) of 0.5 (as a TWA for up to 10 hr/day during a 40-hr workweek) to be f/cc for RCF—numerically equal to the RCF achievable for most workplaces where RCFs or RCF products industry’s recommended exposure are manufactured, used, or handled. Their lung cancer risk guideline (REG). Although extrapolation estimate is based on the models of Moolgavgar and from animal studies to humans is uncertain Yu/Oberdörster, which are considered to be inappropriate (see

- 134 - to a degree, the extrapolations have been above) made using the best available models. Here is a brief summary of risk estimates: The different RCF risk assessments quoted by industry (DECOS, 1995; Fayerweather et al; 1997; Moolgavkar; 1999; Yu In 1995, DECOS (a workgroup of the and Oberdörster, 2000; Turim and Brown, 2003) are considered Health Council of the Netherlands) as inadequate. For instance, all of those risk assessments were published a report evaluating the health based on the results of insensitive rat chronic inhalation data. effects of occupational exposure to This is an explanation why all of those lead to unrealistically low synthetic vitreous fibers (SVFs). The cancer risks for RCF. In addition, the assessments were partly purpose of the report was to establish based on unlikely assumptions e.g. like similar tumour initiation health-based recommended occupational rates between rat and humans (Moolgavkar et al., 1999 which exposure limits for specific types of SVFs. builds upon the Fayerweather et al. approach). Applying the As one of the criteria for determining the Moolgavkar et al. approach for asbestos would lead to cancer airborne exposure limits for six distinct risk estimates for asbestos which are 2-3 orders of magnitude types of SVFs, risk assessments were lower than the risk assessment for asbestos by US EPA performed for each fiber type, including (1988/2008). RCFs. The risk analysis for RCFs was based on the assumption that RCFs are a potential human carcinogen as indicated by the positive results of carcinogenicity testing with animals. A health-based recommended occupational exposure limit was determined considering the following alternatives: • If the carcinogenic potential of RCFs is caused by a nongenotoxic mechanism, an occupational exposure limit of 1 respirable f/ml as an 8-hr time weighted average (TWA) should be recommended based on an no observed adverse effect level NOAEL of 25 f/ml and a safety factor of 25. • If the carcinogenic potential of RCFs is linked to a genotoxic mechanism, a model assuming a linear relationship between dose and the response (cancer) should be used to establish the occupational exposure limit. The model indicated that an excess cancer risk of 4 ×10-3 is associated with a TWA

- 135 - exposure to 5.6 respirable f/ml based on 40 years of occupational exposure. A cancer risk of 4×10-5 is associated with exposure to 0.056 f/ml, and a linear extrapolation indicated that occupational exposure to 1 respirable f/ml as an 8-hr TWA for 40 years is associated with a cancer risk of 7×10-4. The DECOS analysis relied on the data from a long-term multidose study with rats exposed to kaolin ceramic fibers (Bunn et al. 1993; Mast et al. 1995b). These data showed that exposure by inhalation to 25 f/ml (3 mg/m3) for 24 months produced a negligible amount of fibrosis (mean Wagner score of 3.2). Consequently, the Dutch committee viewed 25 f/ml as the NOAEL for fibrosis. The working lifetime occupational risk estimated by DECOS is substantially lower than that estimated in the SVHC Dossier. Based on the DECOS results the TWA fiber concentration associated with a risk of 4:1000 would be approximately 5.6 f/ml rather than 0.1 f/ml. In commenting on the DECOS analysis NIOSH (2006) noted one possible limitation: “The linear modeling approach in this analysis of the exposure- response relationship using the animal data does not take into consideration possible differences in dosimetry and lung burden between rats and humans.” Fayerweather et al. (1997) conducted a risk assessment of occupational exposures for glass fiber insulation installers. They also performed risk analyses with several other types of SVFs, including RCFs. This analysis applied an EPA linearized multistage model (representing a linear nonthreshold dose-response) to data from

- 136 - rat multidose and maximum tolerated dose (MTD) chronic inhalation bioassays (Mast et al. 1995a, b) to determine exposures at which “no significant risk” occurs; defined in this publication as no more than one additional cancer case per 100,000 exposed persons. Nonlinear models were also used for comparison: the Weibull 1.5- hit nonthreshold model (representing the nonlinear, nonthreshold dose-response curve) and Weibull 2-hit threshold model (representing the nonlinear, threshold dose- response curve). Fiber inhalation by rats was equated to humans by determining the fibers/day•kg of body weight for the animals and using an exposure scenario of 4-hr/day (consistent with insulation installation worker schedules), for 5 days/week and 50 weeks/year over 40 working years of a 70- year lifespan. Use of the linearized multistage model based on this analysis represents a risk of 3.8×10-5 for developing lung cancer over the working lifetime at an exposure concentration of 1 f/ml. As with the DECOS analysis, this model would predict a substantially higher OEL to attain a target risk of 4 x 10-3 than that given in the SVHC Dossier. The Fayerweather et al. (1997) analysis used exposure concentration rather than resulting lung burden and (other than making comparisons on a fibers/day.kg basis) did not address possible differences in dosimetry between rats and humans.

Moolgavkar et al. (1999) analyzed rat bioassay data using a sophisticated integrated dosimetry-potency model that NIOSH (2006) termed “a methodological

- 137 - improvement over the risk assessment prepared by Fayerweather et al. (1997).” This risk assessment employed a biologically based model for carcinogenesis, the two-stage clonal expansion model that allowed for explicit incorporation of the concepts of initiation and promotion in the analysis. They found that a model positing that RCF was an initiator had the highest likelihood and employed an approach based on biological considerations for the extrapolation of risk to humans. Their analysis concluded that working lifetime occupational risks were bounded by comparisons with two human populations (1) a population of nonsmokers and (2) an occupational cohort of steelworkers (both smokers and nonsmokers) not exposed to coke over emissions. They concluded that the best estimate of the excess probability of lung cancer at age 70 of occupational exposure to 1 f/ml of RCF was 3.7 x 10-5 (95% CI = 2.4 x 10-5 to 5 x 10-5) for the nonsmoking cohort and 1.5 x 10-4 (95% CI = 0.9 x 10-4 to 2 x 10-4) for the steel industry cohort. NIOSH (2006) concluded “the steel cohort may be the preferable cohort to use for estimating the occupational risks associated with exposure to RCFs.” Following the NIOSH recommendation leads to the conclusion that occupational exposure to 1 f/ml results in a maximum likelihood estimate of 1.5 x 10-4 risk: this is lower (by a factor of approximately 27) than the SVHC Dossier tolerance level risk and at a fiber concentration ten times greater than the 0.1 f/ml calculated by the Dossier authors. Later work by Moolgavkar et al.

- 138 - (see e.g., 2000) affirmed the importance of biopersistence and concluded that the best estimate of risk associated with occupational exposure to RCF 1 f/ml was 4.6 x 10-5—again a much lower risk at a much higher fiber concentration than determined in the SVHC Dossier. Turim and Brown (2003) compared various approaches for estimation of risks and concluded that occupational exposure to RCF was related to average concentrations as shown in the table below.

Risk Level Concentration (fibers/ml) 10-4 (1 in 10,000) 1.0 10-5 (1 in 100,000) 0.1 10-6 (1 in 1,000,000) 0.01

Yu and Oberdörster (2000) analyzed the animal bioassay data for the US EPA using a dosimetry model and concluded that continuous occupational exposure to 1 f/ml would result in a maximum likelihood estimate of 1.41 x 10-3 (1.4: 1000). This estimate assumed that the lung burdens in humans and animals were equivalent if there were normalized to their lung surface areas. Although the workplace concentration associated with a risk at the tolerance level in this analysis is lower than those estimated by e.g., Turim and Brown, these are still greater than that estimated in the SVHC Dossier. Moreover, Yu and Oberdörster (2000) acknowledged that their estimates (p61) “very likely resulted in a more conservative human cancer risk estimate.” Subsequently others (e.g., Maxim et al., 2003) have modeled animal data yielding

- 139 - generally similar risk estimates. Upon review of the available risk studies the NIOSH (2006) set the REL at 0.5 f/ml and concluded: “Keeping exposures below the REL [0.5 f/cc] should reduce the risk of lung cancer to estimates between 0.073/1,000 and 1.2/1,000 (based on extrapolations of risk models from Moolgavkar et al. [1999] and Yu and Oberdörster [2000]).”

In making this calculation, NIOSH did not allow for the fact that, in order to ensure compliance with a standard of 0.5 f/cc, it would be necessary to reduce exposures to approximately one-half this value. Average risks with an OEL of 0.5 f/cc should, therefore, be approximately one-half the values quoted in the above paragraph (i.e., a range from approximately 0.036 to 0.6/1,000). These are the best, albeit probably conservative, risk estimates available based on animal inhalation experiments and dramatically different form those described in the SVHC Dossier. Thus, the benchmark concentration shown in the dossier is too low. Any of the several risk analyses described here could be used to justify a benchmark value of 0.5 f/ml (the industry REG), 1.0 f/ml (the maximum exposure limit (MEL) in the United Kingdom, or even higher. -Exposure data We agree that for the most part , the companies that As noted above the SVHC Dossier uses manufacture RCF in Europe are the same as those in the US data from ECFIA producers submitted in since processes and controls are similar. It may therefore be 1999. It is appropriate to update these appropriate to utilize all the relevant data for comparison. data. As noted above, most RCF manufacturers in Europe also manufacture Maxim et al. (2008 s. above) investigated the time trends in RCF RCF in the United States; both production exposure. It was found that the rate of improvement in fibre

- 140 - processes and engineering controls are concentration in RCF manufacturing plants and customer similar. Therefore it is appropriate to facilities has slowed down in recent years. A significant update the exposure data given in the decrease of RCF exposure in the near future is anticipated to be SVHC Dossier and to include and pool not very likely. We would therefore really appreciate monitoring these data with those from the United data showing that the RCF exposure at workplaces is below the States to provide the largest sample size tolerance level of 0.1 fibres/ml. possible. Because exposure have decreased in the past twenty years, it is appropriate to limit the time span of the data included to accurately depict current exposures. We have chosen to include data for the most recent five years (2004 – 2008) in what is shown below in Table 1. Table 1 is laid out in the same format as Table 7 in the SVHC Dossier, so direct comparisons are possible.

As shown in Table 1, the revised geometric Although the exposure data given by the commentator shows means for each functional job category are some decrease of exposure levels the majority of data points do lower than concentrations reported in 1999 clearly exceed the tolerance level of 0.1 fibres/ml. This is and shown in the SVHC Dossier. Current confirmed by the ladder diagram submitted by the commentator data for functional job categories range (ladder diagram for combined monitoring data, 2004-2008) and from approximately 14% to 51% lower than by literature data as well (Maxim, 2008 s. above). in 1999. Across all functional job categories, the average decrease is approximately 39 percent. Figure 2 is a direct reproduction of Figure 1 in the SVHC Dossier (misspellings included). It is useful to compare this figure with the updated ladder diagram (see Figure 3) showing more appropriate occupational limits such as 0.5 and 1 fibre/mL. Figure 1 gives the clear impression that the majority of actual exposures exceed the fiber concentration associated with a “tolerance level” risk of 4:1000.

Figure 3, which shows the exposure data

- 141 - compared to two benchmarks at 0.5 f/ml (the industry REG NIOSH REL) and 1.0 f/ml (the UK MEL and one more consistent with the results of several risk analyses) gives a very different impression. To be sure, results vary by job category, but the majority of exposures in all categories are at or beneath the benchmarks.

(Exposure information) 54 20091013 Unifrax Corp., Company, The German Dossiers present an Introductory statement, no response necessary. United States of America alternative risk analysis to those referenced above. The risk analysis is based on an estimated human risk associated with “asbestos” exposure derived from epidemiological studies and the results of IP studies on crocidolite asbestos and RCF. For reasons discussed in the following sections, I believe that the risks calculated in the German Dossiers are overstated. This issue is addressed in more detail on The detailed points referred to are found elsewhere in this pages 5-13 of the attached PDF file titled document and will be specifically responded where they appear. “RCF SVHC.pdf”. Because this comment form does not allow for efficient submission of graphics, tables, footnotes and references, the reader is referred to the attached PDF file for a more complete argument against labeling RCF as a substance of very high concern. An excerpt of relevant argument follows (refer to the attached PDF for tables, figures and references):

Additional comments on the Draft Annex XV risk analysis

The German Dossiers present an alternative risk analysis to those referenced

- 142 - above. The risk analysis is based on an estimated human risk associated with “asbestos” exposure derived from epidemiological studies and the results of IP studies on crocidolite asbestos and RCF. For reasons discussed in the following sections, I believe that the risks calculated in the German Dossiers are overstated.

-Risk methodology in the German Dossiers The methodology estimates the relative risk of an alternative fiber (RCF), compared to “asbestos,” by using estimates derived from a dose-response analysis of data from chronic IP studies of rats exposed to crocidolite asbestos and to several alternative fibers. More specifically, the methodology employs the BMD10 defined as the IP dose (in 109 fibers) that induces a 10% probability (response above background) of tumors (mesothelioma) in rats. It posits that the ratio of the BMD10croc value for crocidolite asbestos to that for RCF BMD10RCF determines the relative occupational risk of this fiber relative to asbestos as follows:

RRCF = (BMD10croc/ BMD10RCF) (Rasbestos), (1)

Where: BMD10croc = Estimated BMD10 for crocidolite asbestos, BMD10RCF = Estimated BMD10 for RCF, Rasbestos = Estimated human risk (lung cancer plus mesothelioma) for “asbestos,” and RRCF = Estimated human risk (lung cancer

- 143 - plus mesothelioma) for RCF.

Table 1. Updated characterization of We agree that for the most part, the companies that functional job categories in relation to manufacture RCF in Europe are the same as those in the US workplace exposure since processes and controls are similar. It may therefore be pooled European and U. S. data, 2004- appropriate to utilize all the relevant data for comparison. 2008. Although the exposure data given by the commentator shows Industrial some decrease of exposure levels the majority of data points do group Functional clearly exceed the tolerance level of 0.1 fibres/ml. This is job category Description Workplace confirmed by the ladder diagram submitted by the commentator concentration (ladder diagram for combined monitoring data, 2004-2008) and (fibres/mL; by literature data as well (Maxim, 2008 s. above) Geometric mean)* Percent The cited number of measurements for Europe and the USA in difference the note pretend, that the data are published differentiated. Up from data to now the data are published pooled and additionally not in a presented numeric manner. in Annex 15 Dossier Primary production Fibre Production All jobs on lines producing bulk or blankets 0.146 -14.3% Secondary production Mixing – Forming Wet-end production of vacuum-cast shapes, boards, paper; includes mixing RCF putties, compounds or castables 0.162 -37.8% Secondary production Finished Cutting, sanding or machining RCF products after fiber manufacture 0.322 -44.4% Secondary production Assembly Combining or assembling RCF with other materials (e.g., module making, laminating, encapsulating) 0.152 -51.1% Furnace related uses (Installation / Removal) Installation At end-user locations, installation of RCF insulation in industrial furnaces, boilers, petrochemical plant

- 144 - equipment, kilns, foundry equipment, or electric power generators. Includes furnace maintenance, mold wrap and kiln car builds. 0.254 -44.8% Furnace related uses (Installation / Removal) Removal Removal of after- service RCF from industrial furnace, boiler, etc. Includes maintenance, mold knock-out, and kiln car removals. 0.579 -40.9% Other Uses Auxiliary Jobs in which employees may be passively exposed 0.079 -39.1% Other Uses Other Automotive, textile, dry- end paper, and end-use applications, and other applications not elsewhere classified. 0.053 -40.6%

Note:

* Data are pooled from European (March 2004 to March 2009) CARE Programme and United States (2004-2008) RCFC datasets. Personal samplers are used to measure concentrations in the workers' breathing zone. Fibre counting was done by using phase-contrast optical microscopy (PCOM) in accordance with WHO counting rules (in Europe) U. S. data is adjusted to WHO rules equivalent. Average concentrations were recorded for the monitoring period (from 50 to 500 min, average sample duration is 359 minutes) and reported as Actual Time-Weighted Averages (ATWA). Over this period, there are 1,482 observations for Europe and 2,679 observations in the United States; there are 4,161 samples combined).

- 145 - Fig. 2. This is a direct reproduction of Figure 1 in the SVHC Dossier (misspellings included). Fig. 3. Updated ladder diagram showing The ladder diagram cannot be udated because the used data proportion (%) of concentrations of ATWAs have not been published in a numeric form up to now. in each industrial group, pooled European Additionally see comments on 20091013 and U.S. data. Unifrax Corp., Company, United States of America Comments relative to each of these terms We are well aware of the evaluation performed by Hodgson and are provided below. Darnton (2000). We agree that different forms of asbestos seem to possess different carcinogenic potency. However, it has to be -Asbestos risk from epidemiology data taken into account that if the results of Hodgson and Darnton The Rasbestos estimate used in the (2000) would have been used in our risk assessment the cancer German Dossier is taken from HEI-AR risk figures would have been higher as the carcinogenic potency (1991) and does not differentiate among of crocidolite is higher than the averaged potency of all forms of various asbestos types. Moreover, as asbestos used in our risk assessment. We decided to use the noted, this risk includes both lung cancer average potency as the derivation of quantitative cancer risks and mesothelioma, whereas the BMD from asbestos epidemiology as this value was the outcome of estimates are derived from studies in rats established and accepted risk assessments for asbestos by US with mesothelioma as an endpoint. EPA (1988/2008) and HEI-AR (1991). When risk-based occupational exposure limits for asbestos were set in the US and There is no principle difference between lung response and certain (but not all) other countries it was serosa response: potency ranking of fibre samples after decided, largely as a matter of prudence inhalation exposure is mirrored by potency ranking after ip (e.g, and convenience, not to differentiate among E / 475 / MMVF 11 / MMVF 21; lung biopersistence predicts ip various types of asbestos. Various result; lung biopersistence and biopersistence in the serosal agencies (see HEI-AR, 1991, p 6-10 for cavity do not differ greatly. lung cancer risks) employed slightly different risk values (often expressed as the percentage increase in cancer per fiber- year/ml) calculated (or selected) from estimates derived from several different epidemiological studies. Risks (normalized per fiber-year/ml) varied among studies by a factor of approximately three orders of magnitude. Though admittedly imprecise, some sort of central value was chosen for risk extrapolation and standard setting purposes. In evaluating the suitability of the risk analysis in the German Dossier the

- 146 - substantial uncertainty in Rasbestos must be borne in mind. We now understand that various types of We are well aware of the evaluation performed by Hodgson and asbestos differ in potency. In 2000 Darnton (2000). We agree that different forms of asbestos seem Hodgson and Darnton of the UK Health and to possess different carcinogenic potency. However, it has to be Safety Executive published an important taken into account that if the results of Hodgson and Darnton and influential reanalysis of the available (2000) would have been used in our risk assessment the cancer epidemiological evidence regarding the risk figures would have been higher as the carcinogenic potency relative potencies of chrysotile and of crocidolite is higher than the averaged potency of all forms of amphibole asbestos. They examined 21 asbestos used in our risk assessment. We decided to use the cohorts occupationally exposed to average potency as the derivation of quantitative cancer risks chrysotile, a mixture of chrysotile and other from asbestos epidemiology as this value was the outcome of amphiboles, (crocidolite and amosite). established and accepted risk assessments for asbestos by US They concluded that the potency of EPA (1988/2008) and HEI-AR (1991). chrysotile differs from the amphiboles (i.e., was less potent) and, moreover, the relative potency of chrysotile and the amphiboles were different for lung cancer and mesothelioma. Specifically they concluded (abstract): “that the exposure specific risk of mesothelioma from the three principal commercial asbestos types is broadly in the ratio 1:100:500 for chrysotile, amosite, and crocidolite respectively.” They also analyzed data on lung cancer and concluded (abstract): “The risk differential between chrysotile and the two amphibole fibres for lung cancer is thus between 1:10 and 1:50.” Other investigators have reached quantitatively different conclusions regarding the relative potencies of these fiber types, but modern analyses (see e.g., Berman and Crump, 2008a, b; Gibbs and Berry, 2008; Yarborough 2006, 2007) agree that the relative potencies for lung cancer and mesothelioma are not the same for chrysotile (less potent) and amphibole types. If these two types of asbestos have different ratios of lung cancer and

- 147 - mesothelioma potency (as assuredly they do whatever the exact potency values for each) it is also reasonable to believe that other fibers would also have different relative potencies. Thus, multiplying a ratio of BMD values for mesothelioma in rats by an assumed (and very uncertain) potency relative to both lung cancer and mesothelioma amounts to a heroic assumption.

Thus, two major difficulties with the risk analysis procedure used in the German Dossiers are:

• The overall potency factor (combined lung cancer and mesothelioma) for “asbestos” is very uncertain. • The relative potencies for lung cancer and This is speculation mesothelioma might be very different for RCF than “asbestos” even assuming that RCF does cause tumors. -Use of IP studies inappropriate for IP studies alone are not used for risk estimation. They are quantitative risk determination primarily used to compare the carcinogenic potency of The model (see equation (1) above) used in crocidolite and RCF. The epidemiology based risk estimate of the German Dossier multiplies the “risk” for asbestos is then modified. “asbestos” by the ratio of BMD10 values for crocidolite asbestos and RCF to calculate a risk for RCF. Although some scientists in Germany believe that IP is a suitable (even preferred) exposure pathway for development of potency estimates, this view is not universally held. Indeed, the prevailing view among fiber toxicologists is that IP studies are not suitable (or, in any event, less suitable than IH studies) for risk estimation. This point is made in numerous quotations An extensive justification has been recently provided to justify

- 148 - taken from the scientific literature that are the fact that chronic inhalations studies in rats are insensitive to shown in Table 2 below. All point out detect fibre carcinogenicity (Wardenbach et al. 200557). Having difficulties with IP studies and indicate that in mind the higher sensitivity of humans compared to rats after a well-designed chronic inhalation study is inhalation of the best basis for risk estimation for fibers. asbestos, the arguments provided in favour of the inhalation The various sources included in Table 2 studies are highly questionable. include individual scientific publications and also the conclusions of various The classification criteria according to the dangerous substance governmental agencies (e.g., NIOSH, directive and also according to the CLP regulation allow to National Research Council, WHO, IARC). assess fibre carcinogenicity by intraperitoneal testing. This Table 2. Quotations from the literature means that the intraperitoneal test is scientifically accepted in relevant to the choice of exposure pathway EU legislation. The critique expressed on the questionable for chronic bioassays validity of this test is therefore baseless.

Many scientists have addressed the topic of Moreover, taking into account the human cancer incidences the most appropriate exposure pathway for caused by asbestos, it is prudent to regulate other fibres on the chronic bioassays with fibers. This table basis of the entire evidence including intraperitoneal injection contains quotations from several studies. Positive results after intraperitoneal injection are publications that address this topic. The regarded as essential for evidence of carcinogenic activity in emphasis (material shown in italics) has experimental animals to be classified as ‘‘sufficient,’’ as this test been added. system reveals the carcinogenic activity of asbestos fibres and of fibres with lower potency in relation to asbestos fibres. Quotations Publication Moreover, it has to be recalled that mesothelioma induction by “Although intraperitoneal injection studies to intraperitoneal injection of fibres is hardly influenced by determine biopersistence of fibers may give concomitant injection of granular particles. The test model of useful results, it is suggested to focus on intraperitoneal injection of fibres revealed that the carcinogenic the lung for these assays, i.e., using potency of various man-made vitreous fibres can differ by three inhalation preferentially or, possibly, orders of magnitude (Wardenbach et al., 200058), which enables intratracheal instillation.” Page 226, the selection and use of less potent man-made vitreous fibres Oberdörster, 2003 and the application of adequate protective measures. “The suggestion that the i.p. test because of its far greater sensitivity compared to inhalation should be used to determine the carcinogenic potential of fibers is not well supported. I.p. doses administered into

- 149 - rats, i.e., up to 109 fibers either as a single bolus or in repeated doses, are extraordinarily high. Considering the available surface area of the rat peritoneum and translating this to the alveolar surface area of the lung, these i.p. doses are equivalent to instilling into the lung more fibers than the lung weighs (Oberdörster, 1996). Obviously, in addition to the questionable route of administration, these are inappropriate and irrelevant doses and the results should neither be used for lung tumor risk assessment (high dose, wrong target cells) nor for mesothelioma risk assessment (irrelevant dose.)” Page 60, Yu and Oberdörster, 2000

Table 2 (cont’d.) Quotations from the literature relevant to the choice of exposure pathway for chronic bioassays

Quotations Publication “Unquestionably, inhalation is the most relevant route of exposure (Table 1 [not reproduced here]) and one might expect that chronic dosing of animals with airborne fibers will identify correctly the carcinogenic potential of a fibre provided the selected species is sufficiently sensitive. Usually, the rat is the species of choice used.” Page 15, Oberdörster, 1996 “Dr. E. E. McConnell (U.S.A.) reported on the advantages and limitations of in vitro screening tests. He reported that decades ago, researchers testing for carcinogenicity had used many techniques to administer materials to animals, but today it is considered important to mimic the human

- 150 - route of exposure wherever possible.

Regarding the various techniques for fibre testing, in his [McConnell’s] opinion, in terms of methodology, pathological evaluation, quality control and the value of the data for risk assessment, the inhalation route was preferable to intratracheal instillation and to intraperitoneal injections (see Table 1 [not reproduced here]). Dr. F. Pott answered that in his opinion, Dr. McConnell’s table does not correspond with scientific criteria.” Page 96, Bignon et al., 1995a “In respect to the contrasting views on the utility of the results from the intraserosal exposure model to hazard and risk assessment in humans, it was suggested that one problem with intraserosal tests is that so many fibre types produce a maximum response that separation of different fibre potencies is difficult, and that differentiation of fibre response is better in inhalation studies.” Page 98, Bignon et al., 1995a “The Workshop generally agreed that long- term inhalation studies are relevant to the evaluation of the hazard of MMFs for humans.” Page 100-101, Bignon et al., 1995a "The cornerstone of the testing strategy is a subchronic study in rodents (preferably rats), evaluating a range of toxicologic endpoints. The inhalation exposure route (at least 1-mo and preferably 3-mo exposure duration) is preferred, but intratracheal instillation is acceptable under certain conditions." Page 500, ILSI Working Group, 2005

- 151 - "The preferred route of exposure for the 3- mo study is by inhalation, but for a batch of experimental fibers that cannot be obtained in sufficient numbers to generate a cloud, intratracheal instillation may suffice if the fiber number is kept low. (Driscoll et al., 2000). However, there is no database for the relationship of pulmonary fibrosis to long term effects with intratracheal instillation." Page 519, ILSI Working Group, 2005 "The evaluation of respiratory tract toxicity from airborne materials frequently involves exposure of animals via inhalation. This provides a natural route of entry into the host and, as such, is the preferred method for the introduction of toxicants into the lungs." Page 24 (Abstract), Driscoll et al.,2000 "Despite certain advantages of instillation over inhalation exposure, a number of concerns exist regarding use of the former technique. Primary among these is that the introduction of the toxicant is nonphysiologic, involving invasive delivery, usually at a dose and/or dose rate substantially greater than that which would have occurred during inhalation. In addition the distribution of an instilled material within the respiratory tract will likely differ from the distribution of an inhaled material." Page 25, Driscoll et al., 2000

Table 2 (cont’d.) Quotations from the literature relevant to the choice of exposure pathway for chronic bioassays

- 152 -

Quotations Publication "There are data to assess the effects of exposure method upon the early intrapulmonary distribution of particles. This phase of particle deposition would potentially affect routes and rates of clearance from the lungs and the dose delivered to specific sites within the respiratory tract or to extrapulmonary organs. The distribution of particles within the lungs is certainly influenced by the exposure protocol." Page 27, Driscoll et al., 2000 “Animal inhalation studies are more relevant than implantation studies for assessing potential human health hazards and risk, since it is the mode by which occupational lung exposure occurs. The inhaled fibers are subjected to the variety of patho-physiologic processes which are encountered by humans. Major limitations of this method include high-cost, the need for sophisticated equipment, and rigorously defined experimental conditions. In addition, fibers of larger dimensions that might reach the deep lung regions of humans as a result of occupational exposure are filtered out in the upper respiratory tract of rodents.” Page 168, Mast and Utell, 1994 “The intraperitoneal test, in which fibres are injected directly into the intraperitoneal cavity, bypasses the natural route of exposure. Because the lung in bypassed,

- 153 - the natural mechanisms by which the lung removes, dissolves, or breaks fibers, thereby reducing or eliminating potential exposure of the pleural cavities, do not operate." Page 38, IARC, 2002 "IH [inhalation] is recognized as a valid method of hazard identification by internationally recognized scientists. The IT and IP model have no such recognition, and are in fact considered of marginal or no use for human hazard assessment (IPSC, 1988; McClellan et al., 1992; WHO, 1992)." Page 729, McConnell, 1995a "The route of exposure of most concern to humans is via inhalation. The IH [inhalation] model closely mimics this route, the IT [intratracheal instillation] less so, while IP [intraperitoneal injection] does not." Page 728, McConnell, 1995a. See also Page 732 for a restatement. "The IH model is well standardized in several laboratories and is subject to fairly standard operating procedures, which are clearly defined in the literature and are substantiated in various regulatory guidelines." Page 728. McConnell, 1995a "The most appropriate data to be used for risk assessment must have some relevance to the human situation. Inhalation studies meet this requirement because one is able to control potential exposure (by controlling the amount of fibre in the air) and evaluate actual exposure (by measuring the amount of retained fibers in the lung). Additionally, IH exposures commonly use airborne levels

- 154 - that have some relevance in terms of workplace levels." Page 732, McConnell, 1995a "In summary, an objective conclusion is that the most appropriate model for assessing the potential hazard of MMVFs is via inhalation, especially if carcinogenic activity is the endpoint of interest (ICPS, 1988; McClellan et al, 1992; WHO, 1992)." Page 734, McConnell, 1995a

Table 2 (cont’d.) Quotations from the literature relevant to the choice of exposure pathway for chronic bioassays

Quotations Publication "Numerous studies have demonstrated that inhalation models best simulate human exposures because only respirable-sized fibers reach the parenchymal regions of the lung." Page 335, McClellan et al., 1992 "A major strength of the inhalation mode of delivery of fibers is that it is the mode by which people are exposed. The inhaled fibers are subjected to all of the patho- physiological processes which are encountered in people," Page 337, McClellan et al., 1992 "Nevertheless, mammalian inhalation tests have some obvious advantages over other tests. The route of exposure is similar to that in humans, and the exposure to fibrous materials is directed to the intact pulmonary system, including all natural defense

- 155 - mechanisms." Page 30, National Research Council, 2000 "Despite the limitations, a panel of the World Health Organization (WHO) has concluded that inhalation studies constitute the best available laboratory model for assessing the human health risks posed by exposures to fibers (McClellan et al., 1992; WHO, 1992)." Page 31, National Research Council, 2000 "Studies using instilled doses are valuable insofar as they provide a rough estimate of the pulmonary toxicity of materials, but they should not be used for hazard assessments when setting exposure limits." Page 33, National Research Council, 2000 “Instillation is an acceptable form of dosing in many cases and might be the only practical mean[s] of dosing but it cannot substitute for a properly performed inhalation study." Page 33, National Research Council, 2000 "Fifth, the 'gold standard' has been said to be long-term inhalation studies in rodents. Obviously, these take physiological protective mechanisms into account. However, some believe that these tests have limitations for extrapolation to humans since they take into account only fibres with a size less than 1 μm diameter, corresponding to the dimensions accepted by the rodent's narrow airways. Humans can inhale fibres with a diameter of at least 3 μm which may mean that inhalation tests on rodents are Inadequate for fibres

- 156 - between 1 and 3 μm in diameter and throw doubt on their status as a true gold standard' and a full battery of 'tests' should be conducted in a tiered approach." Page 634, Bignon, et al., 1995b "Humans are exposed to fibres by inhalation, ingestion and skin contact. Inhalation is the route of choice for the endpoints of lung fibrosis, lung cancer and mesothelioma. The intratracheal route could also be useful, but there consensus that at present the control of fibre distribution in the lung following this method of administration was often inadequate and so the method cannot yet be recommended for prediction of risks to humans.” Page 4, WHO, 1992 "In considering the data obtained from intratracheally instilled material, it is appropriate to ask its relevance to inhalation, which represents a more natural mode of entry of material into the lungs." Page 287, McClellan, 1983

Table 2 (cont’d.) Quotations from the literature relevant to the choice of exposure pathway for chronic bioassays

Quotations Publication "Instillation and implantation studies deliver fibers directly to the trachea, pleural cavity, or peritoneal cavity, bypassing some of the defense and clearance mechanisms that

- 157 - act on inhaled fibers. Implantation of fibers into either the pleural or abdominal cavities delivers fibers directly to the pleural or abdominal mesothelium, bypassing some or all of the normal defense and clearance mechanisms of the respiratory tract. Intratracheal instillation delivers fibers directly to the trachea, bypassing the upper respiratory tract. These exposure methods do not mimic an occupational inhalation exposure of several hours per day for several days per week over an extended period." Page 33, NIOSH 2006 "Chronic inhalation studies provide information that is most relevant to the occupational route of exposure and human risk assessment. Mechanistic information about fiber toxicity may also be derived from other types of studies." Page 38, NIOSH, 2006 "Intratracheal instillation results in a heavier, more centralized distribution pattern; inhalation exposure results in a more evenly and widely distributed pattern [Brain et al. 1976]." Page 42, NIOSH, 2006 "In animal bioassays, administering RCFs by chronic inhalation most closely mimics the occupational route of exposure. Exposure to RCFs over a time period that approximates the lifespan of the animal provides the most accurate prediction of the potential pathogenicity and carcinogenicity of these fibers in animals. The effects seen in animals may be used to predict the effects of these fibers in humans. Although

- 158 - interspecies differences exist in respiratory anatomy, physiology, and tissue sensitivity. Chronic inhalation studies provide the best means to predict the critical disease endpoints of cancer induction and nonmalignant respiratory disease that may occur in humans because of fiber exposure [McConnell 1995; Vu et al. 1996]." Pages 42, 46, NIOSH, 2006

With respect to RCF specifically, Yu and Oberdörster (2000) wrote: “The most useful studies for extrapolation modeling of RCF risks to humans are long-term inhalation studies using several doses of RCF in a sufficiently high number of animals.” Thus it seems fair to conclude that it is inappropriate to use the IP data for risk estimation purposes—another flaw in the German Dossier risk analysis. -Neglect of fiber dimensions The fibre dimensions used in IP studies correspond with fibre The dimensions of the crocidolite and RCF dimensions encountered at workplaces. At workplaces RCF fibers used in the IP studies are different. fibres are longer than crocidolite fibres. Specifically, the RCF fibers are longer (on average) than the crocidolite fibers. Most scientists believe that, while there is Asbestos fibres are generally much smaller and shorter than no single “bright line” that separates fibers RCF fibres. The portion of asbestos fibres > 20 µm in real of varying lengths into those that are samples is rather low (ca. 0.1%). Asbestos carcinogenicity hazardous and those not hazardous, longer cannot be explained by the low portion of fibres longer than 20 fibers (provided they are sufficiently µm. To include only fibres > 20 µm would lead to a biased and biopersistent) have the greatest potency. artificial lowering of RCF potency when comparing the ip test (See e.g., results between RCF and crocidolite in the risk assessment http://www.atsdr.cdc.gov/HAC/asbestospan performed which is not justified according to the available el/exsum.html, scientific knowledge.

- 159 - http://www.osti.gov/energycitations/product. biblio.jsp?osti_id=6024019.) The so-called dose–durability–dimension (3-Ds) paradigm is widely accepted. It terms of fiber dimensions, length and diameter are important. Diameter is important in terms of respirability and deposition. Regarding length, shorter fibers are believed to be less biologically active in terms of inducing lung cancer or mesothelioma.

Numerous mechanistic, animal bioassay, and epidemiological studies support this hypothesis. Two expert panels (Eastern Research Group, 2003a, 2003b), one by the Agency for Toxic Substances and Disease Registry (ATSDR) and another by the U.S. EPA, examined the role of fiber length. The Executive Summary of the ATSDR panel (Eastern Research Group, 2003a) concluded, “Given findings from epidemiologic studies, laboratory animal studies, and in vitro genotoxicity studies, combined with the lung’s ability to clear short fibers, the panelists agreed that there is a strong weight of evidence that asbestos and SVFs shorter than 5 μm are unlikely to cause cancer in humans.” As noted above, the U.S. EPA held a peer consultation workshop on asbestos risk analysis (Eastern Research Group, 2003b) at which, inter alia, a potency-based risk index was presented that assigns zero risk to fibers less than 5 μm in length; fibers between 5 and 10 μm are assigned a risk that is

- 160 - 1/300th of the risk assigned to fibers longer than 10 μm. The review panel agreed that “there is a considerably greater risk for lung cancer for fibers longer than 10 μm [but was] uncertain as to an exact cut size for length and the magnitude of the relative potency.” Among the many comments of observers at this meeting, Dr. Suresh Moolgavkar (Eastern Research Group, 2003b, at F-7) noted, “While the general principle that the longer fibers are more toxic appears to be generally accepted and sound, the cutoff at 10 microns proposed in this document is arbitrary. The experimental data suggest a cut-off at 40 microns.” Oberdörster (2000, 2003) claimed that the presence of fibers longer than ~15–20 μm is the most critical predictor of tumorigenicity—in part because phagocytosis by alveolar macrophages is a decisive factor. Fibers that are too long to be phagocytized by alveolar macrophages are less likely to be cleared out of the alveolar compartment and can interact with epithelial cells and be transported to pleural sites. The diameter of the alveolar macrophage in humans is reportedly between 14 and 21 μm; it is slightly smaller in rats (Oberdörster, 2000, and contained references). Many scientists have suggested a length cutoff at approximately 20 μm (Health Effects Institute–Asbestos Research, 1991; Hesterberg et al., 1998; Hesterberg & Hart, 2001; Moolgavkar et al., 2001a, 2001b;

- 161 - National Research Council, 2000). Pott et al. (1989) suggested that “the maximum carcinogenic potency of fibers is perhaps not reached at lengths of less than about 20 μm.” Vu et al. (1996) reported that the panel of technical experts at a conference sponsored by the US Environmental Protection Agency EPA, National Institute of Environmental Health Sciences (NIEHS), NIOSH, and the (US) Occupational Safety and Health Administration (OSHA) reached a consensus that fibers with length greater than 20 μm were most potent. Miller et al. (1999) using statistical methods found evidence that synthetic vitreous fibers longer than 20 μm had the greatest effect on carcinogenicity. Bernstein et al. (2001a) analyzed a series of inhalation bioassay results with SVFs and found that the number of fibers >20 μm remaining in the lung provided a statistically significant relationship to collagen deposition at the bronchoalveolar junction, a precursor of interstitial fibrosis. Bernstein et al. (2001b) also noted “the biopersistence of fibers longer than 20 μm was found to be a good predictor of the lung burden and early pathological changes in chronic inhalation studies with fibers as well as of the tumor response in chronic intraperitoneal studies with fibers.” Yu and Oberdörster (2000) in an analysis of the dose-response relationship for refractory ceramic fiber explicitly modeled fiber lung burdens in terms of WHO fibers and fibers >20 μm in

- 162 - length. An IARC Monograph on SVFs commented on the role of biopersistence as a determinant of toxicity as follows (IARC, 2002): “The longer a fiber persists in the lower respiratory tract, the greater its likelihood to cause effects, especially if it is longer than 20 μm” (emphasis added). Various standardized protocols adopted by the European Commission (e.g., ECB/TM/26 rev. 7; ECB/ TM/16(97) rev. 1) for evaluation of the hazardous properties of SVFs specify that animals be exposed to a minimum concentration of fibers >20 μm and also require the analysis of the data in terms of such fibers (see, e.g., Bernstein and Riego Sintes, 1999). Finally, a scientific advisory panel constituted as part of the (US) Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA) explicitly addressed the relevance of long fibers (those >20 μm) in the design of test guidelines for chronic inhalation toxicity and carcinogenicity of fibrous particles (SAP, 2001).

Notwithstanding the consensus in the It does not hold true, that fibre length has not been considered. scientific community that longer fibers are likely to be more relevant for risk determination purposes, the German Dossier bases its relative risk estimates (the ratio of the BMD values) on respirable WHO fibers (i.e., those with lengths ≥ 5 m, diameter ≤ 3 m, and aspect ratio ≥ 3:1). The calculated BMD10 ratio is sensitive to the definition of a hypothetically

- 163 - active fiber. In similar analyses to that shown in the German Dossier, the ratio of the BMDs varies with the definition of a hypothetically active fiber and will be lower (implying a lower risk) for RCF (or other SVF to which the procedure is applied) if the length threshold for activity is increased to 20 micrometers. Thus, another limitation of the risk analysis in the German Dossier is that (even assuming that other parts of this methodology are sound) it fails to consider the effect of fiber length in the analysis. -Choice of IP data set Several of the studies mentioned used crocidolite samples The relative ratio of the BMD values for which were artificially enriched with long fibres. These samples RCF and crocidolite is also sensitive to the are not representative for workplace exposures. The results of data set used for asbestos. Different ratios other studies are questionable, as the doses applied resulted in will result if different asbestos IP data are premature mortality in additional experiments. used and several other data sets are available (e.g., Cullen et al., 2002; Davis et al., 1991; Grimm et al., 2002; and Lambre et al., 1998)—another factor that adds to the uncertainty of the estimated risk based on IP data. Plausibility checks The cumulative exposure in the currently available Finally, it is appropriate to examine the epidemiological studies is far too low to detect a putative RCF plausibility of the risk estimates contained in carcinogenicity in humans. Thus, classification and risk the German Dossier. The basic conclusion assessment has to be based on experimental results. of the German Risk analysis is that RCF is Fibre carcinogenicity is determined by the so-called 3Ds- as potent a carcinogen as crocidolite paradigm (dose, durability, dimension). Therefore, it does not asbestos. This conclusion is inconsistent make sense to only refer to comparative biopersistency with the findings of the epidemiological (weighted half time) as an argument against the RCF risk study and also with the measured assessment described in the Annex XV dossier and to fade out biopersistence of RCF and crocidolite dose and dimension. For instance, dimension has be included in asbestos. such a comparison as it similarly influences potency: RCF fibres

- 164 - are less biopersistent than is crocidolite, but the fibres are more potent as they are much longer (dimension). As a consequence, it is plausible, that the potency may be similar. As noted above, the available Because of the relatively young and small cohorts studied by epidemiological data indicate that LeMaster et el.(2003) respectively Walker et al. (2002) the occupational exposure to RCF has not results allow only limited conclusions. Only one positive resulted in interstitial fibrosis, incremental incidence would produce a substantial effect on the significance lung cancer, or any cases of level. mesothelioma. However, the sixe and exposure duration of the cohort being studied is limited. Walker et al. (2002) analyzed these data and determined that, for lung cancer, the hypothesis that RCF was as potent as amphibole asbestos was statistically incompatible with the available data. (Not enough time had elapsed (in 2002) to make a similar claim for mesothelioma, although no cases have been reported in the RCF-exposed cohort.) -Biopersistence Carcinogenicity data after ip-injection are available. It makes no Recent animal research has shown that sense to refer to biopersistence data themselves, as fiber biopersistence, the ability to biopersistence modifies the ip result. accumulate and persist in the lung, is an important determinant of fiber toxicity. This biopersistence in turn is related to (among other factors) chemical durability in lung fluids. Biopersistence is explicitly recognized a determinant of fiber toxicity in carcinogen classification (see Directive 97/69/EC). Nota Q of this directive states:

“The classification as a carcinogen need not apply if it can be shown that the substance fulfils one of the following conditions:

- 165 -

• A short-term biopersistence test by inhalation has shown that the fibres longer than 20 micrometers ( m) have a weighted half life of less than ten days. • A short-term biopersistence test by intratracheal instillation has shown that the fibres longer than 20 m have a weighted half life of less than 40 days. • An appropriate intra-peritoneal test has shown no evidence of excess carcinogenicity. • Absence of relevant pathogenicity or neoplastic changes in a suitable long term inhalation test.” And many other scientists (see e.g., references in Maxim et al. 2006) have affirmed the importance of biopersistence in determining effective toxicity. The German Dossiers acknowledge the importance of biopersistence (p22), but then fails to present relevant data comparing RCF with other fibers. Figure 4 shows the measured halftimes (in days) of long (> 20 m) fibers for several synthetic vitreous fibers together with data for amosite and crocidolite asbestos (Hesterberg et al., 1998; Maxim et al., 2006) as determined in short-term inhalation studies of laboratory animals. All of these results have been determined using a protocol developed for the EC, so these results are directly comparable.

Fibers with a greater weighted half time are

- 166 - more biopersistent and, therefore, more toxic. As noted by Moolgavkar et al. (2001b): “Over a broad range of exposure concentrations, excess risk is a linear function of exposure concentration. Excess risk of lung cancer is also a linear function of weighted half-life for fibers for which the weighted half-life is short compared to the life span of the rat.” As can be seen from Fig. 4, RCFs are Only after short-term inhalation exposure the biopersistence of broadly intermediate in biopersistence Refractory Ceramic Fibers is clearly lower (factor about 10) than between many SVFs and asbestos. the biopersistence of amphibole asbestos. This difference is However, RCFs are not the most much less pronounced (factor about 2) after intratracheal biopersistent of all SVFs and, more to the instillation. However, carcinogenic potency of fibres is not only point; the weighted halftime for RCF (~55 determined by biopersistence, but also by the so-called 3Ds- days) is closer to those of other SVFs than paradigm (dose, durability, dimension). Therefore, it does not to either amosite (~450 days) or crocidolite make sense to only refer to comparative biopersistency asbestos (~900 days). Studies on tremolite (weighted half time) as an argument against the RCF risk amphibole asbestos) suggest that the assessment described in the Annex XV dossier and to fade out comparable weighted half time is dose and dimension. For instance, dimension has be included in essentially infinite. such a comparison as it similarly influences potency: RCF fibres RCF, while being somewhat more durable are less biopersistent than is crocidolite, but the fibres are more in lung fluids than many synthetic vitreous potent as they are much longer (dimension). As a consequence, fibers (SVFs), is an order of magnitude less it is plausible, that the potency may be similar biopersistent than amphibole asbestos. The biopersistence data suggest that occupational risks associated with exposure to RCF would be comparable to those for several other SVFs and markedly lower than those for exposure to amphibole asbestos. The RCF risk analysis presented in the German Dossiers concludes that working lifetime occupational risks associated with RCF are identical to those

- 167 - for amphibole asbestos—a conclusion inconsistent with the available biopersistence data.

Thus there are both conceptual and practical (data related) reasons to challenge the realism of the risk analysis presented in the German Dossiers.

(Information on risks related to the substance) 55 20091013 Unifrax I LLC, Company, This issue is addressed in more detail on The detailed points referred to are found elsewhere in this United States of America pages 7-9 of the attached PDF file titled document and will be specifically responded where they appear. “Unifrax comments”. Because this comment form does not allow for efficient submission of graphics, tables, footnotes and references, the reader is referred to the attached PDF file for a more complete argument against labeling RCF as a substance of very high concern. An excerpt of relevant argument follows (refer to the attached PDF for tables, figures and references):

Risk Estimation Introductory statement, no response necessary. The German Dossiers develop risk estimates for RCF based upon (1) an estimate of the combined risk of lung cancer and mesothelioma for humans exposed to “asbestos” (type not specified) and (2) the ratio of the benchmark doses (BMD measured in respirable WHO fibers) for crocidolite asbestos and RCF estimated from IP studies on rats with mesothelioma as an endpoint. Such an estimate makes

57 Wardenbach P, Rödelsperger K, Roller M, Muhle H. Classification of man-made vitreous fibers: Comments on the revaluation by an IARC working group. Regul Toxicol Pharmacol. 2005 Nov;43(2):181-193. 58 Wardenbach, P., Pott, F., Woitowitz, H.-J., 2000. Differences between the classification of man-made vitreous fibres (MMVF) according to the European directive and German legislation: analysis of scientific data and implications for worker protection. Eur. J. Oncol. 5 (2), 111–118.

- 168 - several assumptions (each of which is subject to challenge):

• The ratio of lung cancer to mesothelioma We are well aware of the evaluation performed by Hodgson and for the fiber being evaluated will be the Darnton (2000). We agree that different forms of asbestos seem same as that for “asbestos:” Yet the best to possess different carcinogenic potency. However, it has to be available epidemiological data for chrysotile taken into account that if the results of Hodgson and Darnton and amphibole asbestos (see Hodgson and (2000) would have been used in our risk assessment the cancer Darnton, 2000) indicates that these two risk figures would have been higher as the carcinogenic potency types of asbestos have different relative of crocidolite is higher than the averaged potency of all forms of potencies. asbestos used in our risk assessment. We decided to use the average potency as the derivation of quantitative cancer risks from asbestos epidemiology as this value was the outcome of established and accepted risk assessments for asbestos by US EPA (1988/2008) and HEI-AR (1991). • IP is a suitable exposure pathway for IP studies alone are not used for risk estimation. They are developing risk estimates: This is a claim primarily used to compare the carcinogenic potency of made by several in the German crocidolite and RCF. The epidemiology based risk estimate of government, but one that is at odds with asbestos is then modified. most fiber toxicologists. For example, Yu and Oberdörster (2000) explicitly rejected the use of IP data as the basis for their risk analysis on RCF prepared for the US EPA. Instead they used the RCC nose-only inhalation bioassay. In evaluating the suitability of IP data, they wrote (p 60):

“The suggestion that the i.p. test because of its far greater sensitivity compared to inhalation should be used to determine the carcinogenic potential of fibers is not well supported. I.p. doses administered into rats, i.e., up to 10+9 fibers either as a single bolus or in repeated doses, are extraordinarily high. Considering the available surface area of the rat peritoneum and translating this to the alveolar surface area of the lung, these i.p. doses are equivalent to instilling into the lung more

- 169 - fibers than the lung weights (Oberdörster, 1996). Obviously, in addition to the questionable route of administration, these are inappropriate and irrelevant doses, and the results should neither be used for lung tumor risk assessment (high dose, wrong target cells) nor for mesothelioma risk assessment (irrelevant dose). Thus there is no convincing evidence that the rat is less sensitive than humans for developing lung tumors after chronic fiber inhalation. Using the result of the chronic rat inhalation study with RCF in combination with our dosimetric extrapolation model, therefore, appears to be justified for human cancer risk estimates from exposures to RCF…” • WHO fibers are the right basis for risk Asbestos fibres are generally much smaller and shorter than extrapolation: Most fiber toxicologists RCF fibres. The portion of asbestos fibres > 20 µm in real samples is rather low (ca. 0.1%). Asbestos carcinogenicity cannot be explained by the low portion of fibres longer than 20 µm. To include only fibres > 20 µm would lead to a biased and artificial lowering of RCF potency when comparing the ip test results between RCF and crocidolite in the risk assessment performed which is not justified according to the available scientific knowledge. believe that fibers with the greatest Only after short-term inhalation exposure the biopersistence of potency are long (≥ 20 micrometers [ m]) Refractory Ceramic Fibers is clearly lower (factor about 10) than [(Davis et al., 1986; HEI-AR, 1991; ERM, the biopersistence of amphibole asbestos. This difference is 1997; Vu and Lai, 1997; National Research much less pronounced (factor about 2) after intratracheal Council (NRC), 2000; Oberdörster, 2000; instillation. However, carcinogenic potency of fibres is not only Bernstein et al, 2001a; IARC, 2002; Yu and determined by biopersistence, but also by the so-called 3Ds- Oberdörster, 2002; Agency for Toxic paradigm (dose, durability, dimension). Therefore, it does not Substances and Disease Registry make sense to only refer to comparative biopersistency (ATSDR), 2003; Zoitos et al., 2007)]. (weighted half time) as an argument against the RCF risk Indeed, the EU Nota Q protocol to assess assessment described in the Annex XV dossier and to fade out potential carcinogenicity based on dose and dimension. For instance, dimension has be included in biopersistence is based on weighted such a comparison as it similarly influences potency: RCF fibres halftimes derived from short-term inhalation are less biopersistent than is crocidolite, but the fibres are more studies with long fibers (Bernstein et al., potent as they are much longer (dimension). As a consequence,

- 170 - 1999). The ratios of BMDs for RCF and it is plausible, that the potency may be similar crocidolite asbestos will differ depending upon which fibers are deemed active.

The suitability of the risk analysis developed in the German Dossier can be challenged on other grounds as well. In the end, the analysis in the German Dossier concludes that RCF is as potent a carcinogen as crocidolite asbestos. But,

• Biopersistence data (Hesterberg et al., 1998; Bernstein et al., 2001a,b; Bernstein, 2007; Maxim et al., 2008) indicate that the weighted halftimes of various forms of amphibole asbestos (including tremolite [ranges from >6,000 to infinite days], crocidolite [817 days], and amosite [418 days]) are very much larger than that for RCF (approximately 55 days). The prevailing paradigm for fiber toxicology is that the observed biological effect is a function of dose (other things being equal higher doses are likely to produce greater effects), durability (more biopersistent fibers will produce a greater effect), and dimension (long fibers [not capable of being removed by a macrophage] will have greater effects).

• The work of Walker et al. (2002) indicates Because of the relatively small and young cohorts the findings that, at least with respect to lung cancer (for allow only limited conclusions which there was a mortality deficit), the mortality experience of an occupationally Although more recent studies (Maxim, 200859) show some RCF-exposed cohort was statistically decrease of exposure levels the majority of data points do incompatible with the assumption that RCF clearly exceed the tolerance level of 0.1 fibres/ml. was as potent a carcinogen as crocidolite

59 Maxim, D., Allshouse, J., Lentz, T.J., Venturin, D., Walters, T.E., Workplace Monitoring of Occupational Exposure to Refractory Ceramic Fiber- A 17 Year Retrospective , Inhal. Toxicol., 20:289-309, 2008

- 171 - asbestos. (The exposure duration of the Maxim et al. (Maxim, 2008 s. above) also investigated the time occupationally exposed cohort did not trends in RCF exposure. It was found that the rate of permit a similar conclusion to be drawn with improvement in fibre concentration in RCF manufacturing plants respect to mesothelioma. Nonetheless and customer facilities has slowed down in recent years. A there was no case of mesothelioma in the significant decrease of RCF exposure in the near future is RCF-exposed cohort.) anticipated to be not very likely.

Along with an incorrect estimate of unit risk, the German Dossier also uses out of date (before 1999) exposure information on RCF to conclude that workers are exposed to significant risks.

(Information on risks related to the substance) 56 20091013 Unifrax I LLC, Company, Information on Alternatives Up to our knowledge there are alternatives on the market. (For United States of America The German Dossiers misrepresent the detail see response to SELAS-LINDE GmbH, Company, ability to replace RCF with viable and safe Germany 20091005, above) substitutes. The Dossier correctly In TRGS 619 the classification temperature (not application describes that the replacement of RCF in temperature) of AES wool products is defined up to 1050- applications at temperatures up to 900 °C. 1250°C. Since the preparation of the document (2006) the However, the major use of RCF has always insulating material based on AES was improved. For example been in the temperature range in excess of the classification temperature of the product Superwool HT 1000 °C and in environments that require 60760(trade name) is 1300°C, referring to the product data not only temperature resistance but also sheet. resistance to chemical and mechanical abuse. To data, alternative fibrous materials have not successfully demonstrated this combination of characteristics.

For some properties, such as compression resistance, the alternative fibrous substitutes performance are less than 50% of the RCF properties (Zoitos and Boymel 1999), making the alternatives non-

60 Some references of using AES-products for several industrial applications could be found e.g. under http://www.thermalceramics.com/site.asp?siteid=146&pageid=163

- 172 - functional in some applications. RCF products are also far superior to alternative fibrous alternatives in their resistance to chemical fluxes, materials that are often found in the metals processing industries. It is suggested in the Annex XV dossier, that the temperature limits of fibrous alternatives will be “increased significantly” with current product developments. Although Unifrax and other firms in the industry have worked hard to develop such improvements, there is no support for this conjecture.

To the contrary, the difficulties in developing alternative materials to meet ever increasing temperature applications limits have, thus far, been insurmountable. It is the nature of the high temperature chemistry of RCFs, containing substantial amounts of alumina, that gives RCF excellent performance. It is only materials containing the most inert elements that can adequately provide resistance to such high temperatures in an oxidizing environment. This “development” difficulty is clearly illustrated by the fact that the rate of substitution of RCF has declined over the past several years even though the requirements to evaluate alternative materials have been in place. Replacement of RCF occurred rapidly in the 60% of applications where technical feasibility could be achieved. Although efforts to find suitable substitutes have been in place for nearly 20 years, we have not yet been able to develop viable substitutes for remaining 30-40% of applications.

- 173 -

The German Dossier also implies that it is possible to replace RCF with non-fibrous substitutes such as calcium silicate, vermiculite panels, or refractory bricks. This comment misrepresents the breadth of applications of RCF where these materials simply will not work. Both calcium silicate and vermiculite products have use temperature limitations far below that for RCF. They also are far inferior to RCF in their insulation capabilities. The use of these supposed alternatives when compared with the superior insulation performance of RCF would result in significantly increased energy consumption contributing to additional green house gas emissions.

Further, contrary to the suggestion in the German Dossier, exposure to some of the suggested alternative materials also results in health related risks. Most, if not all, of the refractory silicate bricks and other brick or board like materials contain crystalline silica, a category 1 carcinogen. These materials can and are being used safely, but they are not “safer” alternatives to RCF. Likewise, the replacement of RCF with, not For our suggestion of alternatives to RCFs a feasible yet developed “future” materials, suggests consideration was carried out. (For detail see response to that those untested materials will, a priori, SELAS-LINDE GmbH, Company, Germany 20091005, above) not carry health risks equal to or exceeding Some materials which could indicate potential for concern were RCF. This is complete speculation. For not yet included. For example, the risks of nano materials were example, there is discussion that Nano not clarified until now. That is why we currently do not propose materials are a class of high temperature nano-porous insulating material materials that are capable of replacing some RCF for some applications. The possible risks associated with this class of materials are only now beginning to be

- 174 - assessed. An unknown or unidentified risk is not equivalent to “no risk.” In summary, the German Dossier’s section We disagree. on alternatives misrepresents the “state of the science” on lower health risk materials available as alternatives to RCF. It presents a simplistic view of the application complexity and fails to take into account full life cycle considerations. Given the high environmental value of RCF, the level of scientific understanding available on RCF, the success of efforts to reduce RCF exposures, and in some cases the limited nature of information for the described alternatives, there are not any fully appropriate alternatives to RCF at this time or in the near term.

(Information on alternative substances and techniques) 57 20091014 German Refractory 1. Current classification of RCF In March 2006 the European suppliers of Aluminosilicate Association, Industry or The classification of RCF as Carc. Cat. 2 Refractory Ceramic Fibers provided the EU Technical trade association, Germany were established in 1997. Studies made Committee on Classification and Labelling of Dangerous after 1997 show that the current Substances with the currently available scientific data with classification needs to be discussed. respect to Aluminosilicate Refractory Ceramic Fibre IARC and J.M.G. Davis concluded in 2002 carcinogenicity. The aim of the suppliers was to re-discuss that RCF should be considered a “group 2b: carcinogenicity. France and Germany had responded to that possible carcinogen” which is similar to a documentation that a re-discussion was not warranted as the Carc. Cat. 3 under the Directive new studies do not have any impact on the existing classification 67/548/EEC. This was also concluded by as inhalative carcinogen. The Technical Committee on R.C. Brown, B. Bellmann, H. Muhle, J.M.G Classification and Labelling of Dangerous Substances Davis and L.D. Maxim in 2005: “In concluded that a re-discussion will only take place in case a particular, in the European classification member state would support such re-discussion. There was no system, Carc. Cat. 3 seems more support by any member state. Since 2006, no further relevant appropriate than Carc. Cat 2 for RCF.” scientific data have become available. The existing classification Based on this information the Annex XIV with R49 or H350i is still adequate and does not need to be re- dossier should be withdrawn. evaluated. The current classification is based on animal testing and therefore in no basically contradiction. It shouldn’t be taken

- 175 - another classification into account in advance. 2. Alternative Substances In TRGS 619 the classification temperature (not application In the dossier the author suggests to temperature) of AES wool products is defined up to 1050- consider Alkaline Earth silicate (AES) wool 1250°C. Since the preparation of the document (2006) the products as alternative substances for RCF. insulating material based on AES was improved. For example For several applications there is no the classification temperature of the product Superwool HT possibility to use Alkaline Earth Silicate 60761(trade name) is 1300°C, referring to the product data (AES) wool products. The German sheet. authorities (BAuA) state in the TRGS 619 section 3 (3): " .... that there are technical conditions which require the use of RCF". On page 21 the dossier assumes: "On the other hand current product developments indicate that the upper temperature limit of AES wool products could be increased significantly". This assumption cannot be confirmed by the main AES wool producers. There is no indication that research and product development could increase the upper temperature limit significantly.

3. Environmental Protection The new materials and the products based on bio-soluble wool The RCF have many properties that make AES have a potential for saving energy due to the key property them useful for environmental protection. of the insulating material: thermal conductivity (For details see The long life length and the high response to SELAS-LINDE GmbH, Company, Germany temperature resistance reduce the 20091005, above) frequency of lining replacement. For these Several alternatives have lower thermal conductivity value in reasons and the good insulation properties comparison to RCF and therefore provide a better insulating of RCF, their use reduces energy capacity. Europe’s ambitious environmental targets in its consumption by up to 40 %. For the same Climate Change Programme (ECCP) can be achieved by using reasons, they are important for CO2- alternatives for RCF. reduction in many industrial high temperature applications and are indispensable to achieve Europe´s environmental targets in its Climate Change

- 176 - Programme (ECCP). Conclusions We disagree.

The German Refractory Association requests that both Annex XV dossiers are withdrawn. A scientific discussion on reclassification should take place before any further regulatory action is taken. There are no reasons to place Refractory Ceramic Fibres on the candidate list.

(Information on risks related to the substance)

58 20091014 PRE - Federation of Consumer exposure to RCF fibrous dust In our opinion, the exposure at work places exceeds acceptable European Refractory could never be demonstrated (Schneider et or tolerable limits. Although this concept has yet not been Producers, Industry or trade al. (1996)). In addition, Refractory Ceramic agreed on in the EU the authors identified risks at work places. association, Belgium Fibres can only be sold to professional Whether or not this is true for an exposure through the users, as laid down in Directive environment seems to be not clear at the moment. Data on 2001/41/EC. Rigorous occupational health airborne fibre concentrations in the environment are lacking but regulations are applicable in the work values are expected to be far below those at work places. space. Therefore, PRE does not see a need to prioritise RCF in the REACH authorisation process since REACH has as primary objective the protection of the environment and human health; worker protection is regulated through different pieces of legislation. The reports mention that “... current product In TRGS 619 the classification temperature (not application developments indicate that the upper temperature) of AES wool products is defined up to 1050- temperature limit of AES [Alkaline-earth- 1250°C. Since the preparation of the document (2006) the silicate] wool products could be increased insulating material based on AES was improved. For example significantly”. This statement lacks any the classification temperature of the product Superwool HT support and cannot be confirmed by the 60762(trade name) is 1300°C, referring to the product data main AES wool producers. While the R&D sheet.

61 Some references of using AES-products for several industrial applications could be found e.g. under http://www.thermalceramics.com/site.asp?siteid=146&pageid=163

- 177 - efforts on these products continue, there is no indication for “significant” improvements in the foreseeable future. RCF possess many characteristics Up to our opinion, there are alternatives on the market. (For necessary for their downstream use, such details see response to SELAS-LINDE GmbH, Company, as their long life-length, high temperature Germany 20091005, above) resistance, accurate temperature adjustment etc. As such, they are not substitutable. Eurofer, representing the European Steel Industry, confirms that banning RCF would have a negative impact on the downstream industry as the linings would have to be renewed more often because no substitute can match the life- length of the RCF. It would increase the need for maintenance and thus expenses. Since RCF have good insulation properties there is also a risk of higher energy consumption in the downstream process if they are to be banned.

59 20091014 Individual (affiliated with This issue is addressed in more detail on The detailed points referred to are found elsewhere in this University of Rochester, pages 2-4 of the attached PDF file. document and will be specifically responded where they appear. Academic institution), Because this comment form does not allow United States of America for efficient submission of graphics, tables, footnotes and references, the reader is referred to the attached PDF file for a more complete argument against labeling RCF as

a substance of very high concern. An excerpt of relevant argument follows (refer to the attached PDF for tables, figures and references): Development of risk estimates

Obviously human data would be best for risk assessment purposes. However, there is no human evidence for RCF induced

62 Some references of using AES-products for several industrial applications could be found e.g. under http://www.thermalceramics.com/site.asp?siteid=146&pageid=163

- 178 - cancer (LeMasters et al. 2003). Absent epidemiological data, risk estimates for various fibers are typically derived from animal studies. The design of valid animal studies for risk estimation requires that several choices be made, including the appropriate species, fiber dimensions, method of exposure (e.g., IP, intratracheal (IT), or IH), exposure concentrations or doses, and mathematical basis of extrapolation from high to low doses and the method of conversion from animals to humans. Each of these choices requires thought and may involve compromises. Positive inhalation data are available and An extensive justification has been recently provided to justify should be used for risk assessment the fact that chronic inhalations studies in rats are insensitive to purposes with appropriate science based detect fibre carcinogenicity (Wardenbach et al. 200563). Having animal to human extrapolation (see e.g., Yu in mind the higher sensitivity of humans compared to rats after and Oberdörster, 2000). There is no need inhalation of asbestos, the arguments provided in favour of the to try an awkward approach based on IP inhalation studies are highly questionable. data. Choice of exposure pathway An extensive justification has been recently provided to justify Each exposure route (IP, IT, and IH) has the fact that chronic inhalations studies in rats are insensitive to advantages and disadvantages (see detect fibre carcinogenicity (Wardenbach et al. 200564). Having Bernstein et al., 2001 a, b; Bignon et al., in mind the higher sensitivity of humans compared to rats after 1995 a, b; Driscoll et al., 2000; ILSI, 2005; inhalation of asbestos, the arguments provided in favour of the McConnell, 1995; McClellan, 1983; inhalation studies are highly questionable. McClellan et al., 1992; and Vu et al., 1996). . But the overall consensus in the scientific community seems to be that, despite some limitations, the inhalation route is to be preferred chiefly because it most closely mimics the human route of exposure.

It is noteworthy that several scientific

63 Wardenbach P, Rödelsperger K, Roller M, Muhle H. Classification of man-made vitreous fibers: Comments on the revaluation by an IARC working group. Regul Toxicol Pharmacol. 2005 Nov;43(2):181-193. 64 Wardenbach P, Rödelsperger K, Roller M, Muhle H. Classification of man-made vitreous fibers: Comments on the revaluation by an IARC working group. Regul Toxicol Pharmacol. 2005 Nov;43(2):181-193.

- 179 - organizations, including the World Health Organization (WHO, 1992), International Agency for Research on Cancer (IARC, 2002), National Research Council (2000), and the National Institute for Occupational Safety and Health (2006) agree on this point. The following quotations from these agencies illustrate this point: • National Research Council (2000, p31): "Despite the limitations, a panel of the World Health Organization (WHO) has concluded that inhalation studies constitute the best available laboratory model for assessing the human health risks posed by exposures to fibers (McClellan et al., 1992; WHO, 1992)."

• International Agency for Research on Cancer (2002, p 38): “The intraperitoneal test, in which fibres are injected directly into the intraperitoneal cavity, bypasses the natural route of exposure. Because the lung is bypassed, the natural mechanisms by which the lung removes, dissolves, or breaks fibers, thereby reducing or eliminating potential exposure of the pleural cavities, do not operate."

• NIOSH (2006, p42): "In animal bioassays, administering RCFs by chronic inhalation most closely mimics the occupational route of exposure. Exposure to RCFs over a time period that approximates the lifespan of the animal provides the most accurate prediction of the potential pathogenicity and carcinogenicity of these fibers in animals. The effects seen in animals may be used to predict the effects of these fibers in humans. Although interspecies differences

- 180 - exist in respiratory anatomy, physiology, and tissue sensitivity, chronic inhalation studies provide the best means to predict the critical disease endpoints of cancer induction and nonmalignant respiratory disease that may occur in humans because of fiber exposure [McConnell 1995; Vu et al. 1996]." IP studies use a completely unphysiological and extremely high dose bolus type delivery (very high dose rate). Moreover, the diameters of the fibers deposited using IP tests may be significantly larger than those that would deposit in the alveolar region of the lung from inhalation. Because the time required to dissolve a fiber is proportional (among other things) to the diameter of the fiber, the rate of clearance will be underestimated if the diameters of the test article do not match the diameters of the deposited fibers. The combination of higher doses and reduced clearance would, of course, lead to overestimates of risk. For these reasons, the IH pathway is to be preferred. And, with respect to RCF risk analysis specifically, Yu and Oberdörster (2000) concluded (p5): “The most useful studies for extrapolation modeling of RCF risks to humans are long- term inhalation studies using several doses of RCF in a sufficiently high number of animals.” These authors noted: “A two-year multi-dose study in rats, performed at RCC, Geneva Switzerland, is an adequate study with the necessary information of lung dosimetry and detailed histopathological evaluation of non-cancer and cancer effects.” Although several risk analyses using these data have been published (see

- 181 - references), these were not considered in the German Dossier. Conclusions The comments above are briefly summarized by the following: 1. The Dossier develops risk estimates for two RCFs only from results of IP tests. Results of IH tests are not incorporated into the risk estimates. 2. Many authors and scientific organizations have taken the position that inhalation testing is the preferred method for determining risk estimates in humans. 3. While each method has limitations, results of the IP method can lead to overestimates of risk compared to using results from well designed IH studies. The German Dossier failed to take the IH studies into account when preparing the risk analysis. Consequently the estimated risks from the German Dossier were more than an order of magnitude greater.

Because the German Dossier does not consider the results of IH tests, the risk analysis is incomplete and the results should be questioned.

(Information on risks related to the substance) 60 20091015 RIVM - Bureau REACH, Worker exposure We agree National Authority, The Aluminosilicate Refractory Ceramic Fibres Netherlands Worker exposure data is available (p 17/18, Annex XV dossier) and geometric means vary between 0.13 fibres/ml and 0.65 fibres/ml). Level of exposure is above tolerance level for cancer risk (0.1 fibres/mL). Number of workers potentially

- 182 - exposed to RCFs: 25 000 employees. This underlines the urgent need for limiting the risk.

(Exposure information)

- 183 - Literature:

Boßelmann, D.; Buhr, A.; Gerharz, N.; Golder, P.; Grass, H.-J.; Overhoff, A.; Wuthnow, H.; Pötschke, J.: Entwicklung hochfeuerfester Wärmedämmstoffe, Abschlußbericht über ein Verbundprojekt, Juli 2004, Bonn, p. 5

Overhoff, A.; Buhr, A.; Grass, J.; Wuthnow, H.: New Microporous Materials for Use in Modern Firing Plants, cfi/Ber. DKG 82, 2005, No. 8, p. E3

Ruthenberg, R.: Pressemitteilung „Krebsgefahr durch künstliche Mineralfasern - Asbest-Fehler werden wiederholt“, 22.07.2008 - prcenter.de - Das Online- Pressezentrum http://www.prcenter.de/pressemitteilung-pdf-download.php?news_id=24368

Wuthnow, H.; Pötschke, J.; Buhr, A.; Boßelmann, D.; Pozun, F.; Gerharz, N.; Golder, P.; Grass, H.-J.: Experiences with Microporous Calcium Hexaluminate Insulating Materials in Steel Reheating Furnaces at Hoesch Hohenlimburg and Thyssen Krupp Stahl AG Bochum, Proceedings 47. Intern. Feuerfest-Kolloquium, 13./14.10.2004, Aachen. p. 200

- 184 - Attachments:

Comments from Refractory Ceramic Fibers Coalition: http://echa.europa.eu/doc/about/organisation/msc/msc_rcoms2009/rcom_aluminosilicate_rcf/RCF_Coalition_Brown_2005.pdf

Comments from 3M Europe N.V.: http://echa.europa.eu/doc/about/organisation/msc/msc_rcoms2009/rcom_aluminosilicate_rcf/Note_R.pdf

Comments from Individual, Germany: http://echa.europa.eu/doc/about/organisation/msc/msc_rcoms2009/rcom_aluminosilicate_rcf/Individua_Germany_Comments.pdf http://echa.europa.eu/doc/about/organisation/msc/msc_rcoms2009/rcom_aluminosilicate_rcf/Crocydolite.JPG http://echa.europa.eu/doc/about/organisation/msc/msc_rcoms2009/rcom_aluminosilicate_rcf/Davis_et_al_2002.pdf

Comments from DKFG: http://echa.europa.eu/doc/about/organisation/msc/msc_rcoms2009/rcom_aluminosilicate_rcf/RCF_Coalition_Brown_2005.pdf http://echa.europa.eu/doc/about/organisation/msc/msc_rcoms2009/rcom_aluminosilicate_rcf/DKFG_Davis_2002.pdf http://echa.europa.eu/doc/about/organisation/msc/msc_rcoms2009/rcom_aluminosilicate_rcf/DKFG_Rodelsperger_2004.pdf http://echa.europa.eu/doc/about/organisation/msc/msc_rcoms2009/rcom_aluminosilicate_rcf/DKFG_comments.pdf

Comments from BDI: http://echa.europa.eu/doc/about/organisation/msc/msc_rcoms2009/rcom_aluminosilicate_rcf/BDI_Arguments_ECHA.pdf

Comments from Unifrax Corp.: http://echa.europa.eu/doc/about/organisation/msc/msc_rcoms2009/rcom_aluminosilicate_rcf/Unifrax_Corp_RCF_SVHC_Comments.pdf

Comments from Unifrax I LLC: http://echa.europa.eu/doc/about/organisation/msc/msc_rcoms2009/rcom_aluminosilicate_rcf/Unifrax_I_LLC_Unifrax.pdf

Comments from Individual (affiliated with University of Rochester Medical Center, Academic institution), United States of America: http://echa.europa.eu/doc/about/organisation/msc/msc_rcoms2009/rcom_aluminosilicate_rcf/Individual_USA_Comments.pdf

Comments from German Refractory Association: http://echa.europa.eu/doc/about/organisation/msc/msc_rcoms2009/rcom_aluminosilicate_rcf/Germany_Ref_Assoc_VDFFI_Comments.pdf

Comments from PRE - Federation of European Refractory Producers: http://echa.europa.eu/doc/about/organisation/msc/msc_rcoms2009/rcom_aluminosilicate_rcf/PRE_comments_on_Annex XV.pdf

- 185 -

Comments from Individual (affiliated with University of Rochester, Academic institution), United States of America: http://echa.europa.eu/doc/about/organisation/msc/msc_rcoms2009/rcom_aluminosilicate_rcf/Individual_USA_comments_Gunte_Mark.pdf

Comments from RCFA: http://echa.europa.eu/doc/about/organisation/msc/msc_rcoms2009/rcom_aluminosilicate_rcf/RCFA_comments.pdf

Comments from Individual (affiliated with Harvard School of Public Health, Academic institution), United States of America: http://echa.europa.eu/doc/about/organisation/msc/msc_rcoms2009/rcom_aluminosilicate_rcf/Individual_USA_Comments_Harvard.pdf

Comments from CECOF: http://echa.europa.eu/doc/about/organisation/msc/msc_rcoms2009/rcom_aluminosilicate_rcf/CECOF_Comments.pdf

Comments from European Industrial Gases Association: http://echa.europa.eu/doc/about/organisation/msc/msc_rcoms2009/rcom_aluminosilicate_rcf/EIGA_RCF_Final.pdf

- 186 -