Document M4bXwVaR6GVRRy047xE2vgwE7

Vista Chemical Company 15990 North Barker's landing Road Post Office Box 19029 Houston, Texas 77224 Phone (713)531-3200 July 25, 1988 Dr. Gerald Vaughn 2417 W. Gallaher Ferry Drive Knoxville, TN 37932 VISTA Dear Gerald: I never got you a copy of the JOM article regarding the toxicity of various aluminas. We would appreciate your thoughts on this article relative to what you understand about our small particle work. I've also enclosed a copy of a 1981 article for the JNCI which has some reference to aluminum oxides. I'll be sending you drafts of the MSDS and labels for alumina fibers by mid-week. Sincerely, Thomas G. Grumbles, C.I.H. Environmental Quality Manager dlj Enclosures VVV 000014-393 w Alumina-related Pulmonary Disease Bertram D. Dinman, MD, ScD VVv 000014394 -i 4 i A review of experimental studies suggests that the catalytioally active low temperature transitional forms of alumina produces irreversible ilbronodtilar change only when admin istered by intratracheal insufflation. Other aluminas not catalytically active but also broadly identified as "gamma" for different reasons also appear capable of inducing puimonory fibrosis in the same model. Under conditions of human expo sure, occupational exposure to a broad range of aluminas indicates -at most--mfaimti pulmonary nodular response. within this frame of reference, this report will attempt to place in perspective studies using both experimental models and human encounters with this group of alu minum, oxygen, and hydrogen-containing compounds, the aluminas. Experimental Studies of Alumina's Actions in the Lungs ssessment of biomedical studies of the toxicity of Aalumina in the lung is complicated by inconsisten cies in the nomenclature used to describe chemical and physical variants of this compound. The biologic litera ture can be evaluated only to the extent that investi gators have clearly identified the individual alumina species studied. Only with careful physical-chemical de lineation of these physical-chemical entities is it possible to place experimental and human data in perspective so On the basis of an examination of those reports which provide sufficient chemical and physical specifications, it is possible to identify aluminas previously studied in terms of present day knowledge. These specific aluminas are summarized in Table 1. Gardner et al1 reported that inhalation of a gelatinous alumina (designated 7-HXIOIO), presently known to be gelatinous boehmite, apparently did not produce dele terious pulmonary alterations when administered to guinea pigs in doses of approximately 40 to 120 mg/m9. as to understand the biologic and health implications of exposure to aluminas. To complicate the situation further, various biologic Among its other properties, gelatinous boehmite usually possesses an extremely high surface area (250 to 500 models have been developed as surrogates for the study of human interactions with chemical agents. With the Gardner administered by inhalation at doses of 21 to l/33 mg/m3 a second alumina which he identified as C- evolution of such model systems, a broader understand ing of their utility and limitations is revealed. Although I 730. win description of the mode of preparation indicates I that this alumina was gibbsite; heated at atmospheric biologic mechanisms of action may be demonstrated by I pressure, this aluminum trihydroxide dehydrates such experimental systems, their applicability to the /' mainly to the x transitional form, a catalytically active human condition Is frequently limited, eg, because of I alumina. Once more, no pathologic alterations ensued high dose regimens and nonphysiologic portals of entry l following inhalation exposure. or administration. Such reservations have been exten sively discussed in connection with studies of carcino genesis; reservations are equally applicable to investigations of pulmonary response to dusts, or vapor, or gases. But within these constraints, and in conjunc tion with clinical investigations, such models can assist in elucidation of human responsivity and risk. Thus, The benign outcomes of the studies of Gardner et al1 stand in stark contradiction to the findings of Ring and associates3: in their laboratory, this same HX1010 ge latinous boehmite (referred to as 7-AlOOH) produced extensive nodular and diffuse confluent fibrosis. It is important to note that King et ai9 administered this alumina via intratracheal insufflation. Because ortln^majoFlfilfereBCS Ih response, further From tbs Department of Industrial Environmental Health Sciences, Qreduate School of Public Health. Univeraitjr of Pittsburgh, Pitts burgh. PA 16861 (address oorreepoodsaoe to Dr Dinnaa).. 0006-lTaS/8/a004 Q3H8SQ8 OO/O Copyright d by Amorteee OeeupaUoeel Modical efforts in King's laboratory by Stacy et al4 culminated in more extensive and systematic intratracheal insuffla tion studies of several aluminas (Table 1). Stacy and co-workers,4 using the same gelatinous boehmite (7A100H, HX1010) replicated King's results, ie, moderately severe pulmonary fibronodular change. 328 Alumina-related Pulmonary Oisease/Dinman i ? i i a r TABLE 1 Review of Experimental Data with Contemporary Evaluation of Alumina Species investigator Agent Investigator Designation Identity (1W7) Crystalline Packing Arrangement Route of Administration Particle Size, moi Surface Area, m7g Activity G -rdner et ai1 getai3 -.cyetal4 oster*otter4 C-73G, acti vated HX1010 yAlOOH HX1Q10 T-AIOOH HX1Q10 7-AIsO, HX1Q1Q at 850C y-AIOOH Cera hy drate* icAltO) HX1010 at 1,200C y-AI*Oj x transi tional Gelatinous boeh mite Gelatinous boeh mite Gelatinous boeh mite n transi tional Boehmite a-AlsO* (corun dum) 7 transi tional 7-AlaO, 7 transi tional 7 7' 7* 7' 7 7 a 7 7 Inhalation Inhalation 7 0.02 250-? 250-350 + Intratracheal insuffla tion Intratracheal insuffla tion intratracheal insuffla tion Intratracheal insuffla tion Intratracheal insuffla tion 0.02-0.04 250-350 0.02-0.04 250-350 0.02-0.04 60-? 3 0.6 O.i-0.8* 3 Intratracheal insuffla tion Inhalation 0.005-0.04 0.005-0.04 95-105 95-105 + + + * This alumina is only minimally crystalUna. t System of fibrosis grading of King et a).3 j Estimated by review of electron micrographs of Stacy et al/ Dose, mg Response 21-33/m3 35-105/m3 No harmful effects No harmful effects 100 +3 to +5/5 5brosist 50 +3/5 fibrosist 50 +5/5 fibrosist 50 0 fibrosist 50 +1/5 fibrosist 35 +3/5t 3300/m3 Alveolar pro* te*nosist VVv 000014395 Next, this gelatinous boehmite was heated to 850"C; the dehydration formed what Stacy referred to as 7AleOa. However, present knowledge4 indicates that the product of dehydration at this temperature results largely in the formation of the 77 transitional form, a catalytically active aluminum oxide (cf "Technical Note"). Nevertheless, intratracheal insufflation of what was probably the 77 transitional AlaOs resulted in severe pulmonary fibronodular alteration. The third alumina studied by Stacy was the thermal dehydration product of the same gelatinous boehmita HX1010 heated to 1,200*0 which led to the formation of a-aluminum oxide (Fig. 1) (see also Fig. 2) or corun dum. These crystals were then ground to a smaller particle size, which appeared to range from 0.1 to 0.8 fim. When administered intratracheally, a-AlaOa pro voked a but consistent degree of fibronodular lung pathology. The fourth alumina investigated was the highly crys talline, large particle size (ie, 3.0 pm. mean diameter) 7-AIOOH boehmite (referred to as Cera hydrate1*). This catalytically inactive alumina given intratracheally at similar doses as the other aluminas was inert in the lungs. - These data of the London-Post Graduate School clearly indicate (cf Table 1), within the context of the Intratracheal insufflation model, that (1) exposure to the catalytically active 77 transitional form results in severe fibronodular alteration; (2) exposure to certain catalytically inactive aluminas (viz, a-ALO* and gelat inous boehmite) produces a lesser but broader range of fibronodular change; and (3) in view of the broad range of reactivity to variously described 7-aluminas (ie, zero response with 7-AIOOH boehmite, extending to severe alterations with 77 transitional form), and (4) in view of a minimal but consistent response to a fine a-alumina, (5) it Is inconsistent to associate 7-aluminas with pul monary hazard. (6) This position is reinforced by the probability that the London School never exposed their animal* to any specific 7-aluminum oxide (ie, 7-tranaitional). Since the London investigators' experiments3'4 were based upon intratracheal insufflation dosings, Klosterkotter4 attempted to study the effect of inhala tion of what was believed to be the most biologically active of tbeee aluminas, a relatively well-defined 7 transitional form AlaOa* However, because his dosage was so extremely large, ie. 33 g/m9 for each of 285 days, thin massive overloading produced what would presently be considered as alveolar proteinosis, a nonspecific re sponse to large pulmonary dust loadings. Accordingly, the implications of his experiments for only the most excessive human exposures to 7 transitional AlaO should still be considered as questionable. Regarding tbeee experiments, it is indeed Ironic that, of the multiple forms of aluminas investigated, a specific fibronodular response apparently has never been pro duced by the only one specific 7-alumina, ie, the 7 transitional form. In more precise terms, a fibronodular response can only be said to have resulted from intra- Joumal of Occupational Medicine/Volume 30 No. 4/April 1988 .-.***^' *' 329 1 Transitional AI.O, IHXlOtO heaiM to SSOQ 1----------O------ . 4 32 l--O---------1 Gelatinous Boetimit* l TAlOOH, HXIOIO> hZh At,o, (Cwunoumi (HXIOIO heated to IJ00C1 WV 000014396 /yAlOOH. A Soefimite \ Cera hydrate ') j_____i 1 i i i i ii 10 j................................ Surface Area, M 100 Fig, 1. Correlation of Surface Area of Aluminas with Pulmonary Fibrosis Graded According to Stacy et al.4 tracheal insufflation of (1) catalytically active, low tem perature range transitional, and/or (2) relatively high surface area aluminas. In order to clarify the biomedical literature, we would propose that the reference to yaluxninas be discontinued, and that this intratracheal model's effect be specifically associated with these two alumina species and their properties. More recent inhalation studies of aluminum oxide reported the production of nonspecific polymorphonu clear response characteristic of a wide variety of cyto toxic and nonpathogenic particulates. By contrast, AlaOa did not provoke increased alveolar macrophage re sponse,7 suggesting only minimal potentials for chemotaxic or fibronectin release and injury associated with such events. Although more studies of the alumina's effect upon alveolar macrophages are indicated, pres ently available data do suggest a relatively efficient elimination from the lung8 and a low degree--if any-- of flbrogenicity. Other investigations of corundum (<xAI9O3) indicated an inability to evoke lung tissue acid phosphatase, which has been associated with develop ment of fibrogenic response.9 Occupational Exposure to Aluminas Unfortunately, most reports of occupational exposure to aluminas unmixed with other dusts have not provided specifics sufficient to adequately define the qualitative nature of worker exposures. Much of this stems from the failure of biomedical investigators to appreciate these multiple variations in alumina's physical and chemical forms. Moreover, workplace characterization is complicated by the reality that production and use of specific aluminas varies in response to marked demands and/or technologic changes at various times and places. Furthermore, in the production of aluminas, relatively v few employees are exposed to alumina dust. Such dust exposures that do occur are found in the vicinity of kilns at grinding, screening, transport, and packing opera tions. Given these facts, it is understandably difficult to characterize precisely the quantitative as well as qual itative nature of employee exposures in the course of alumina production. Studies of alumina refineries in which one would encounter aluminas unmixed with other dusts are rel atively uncommon; however, a recent study of alumina refinery workers reported by Townsend et al10 repre sents such exposures. This alumina chemical refinery employing 1,109 potentially exposed workers has been engaged in the production of various aluminas, predom inantly the chemical grades. Production data available showed that since 1958 these workers have processed 244,500 tons of the low temperature transitional alu minas (viz, 7j, y, x) representing 2.9% of the total aluminas produced since that time. Accordingly, expo sure to the intratracheally bioreactive low temperature range transitional alumina species has been occurred here on an intermittent basis from 1952 until 1982. A search for pulmonary fibrosis and pneumoconiosis in this population yielded essentially negative results11; the major change reported was a slight but significant decrement in ventilatory function among nonsmoking workers exposed to excessive total dust levels. The observed changes were consistent with a minor degree of nonspecific chronic industrial bronchitis associated with excessive, protracted nuisance dust exposure, (ie, 100 mg-years for more than 20 years). Other reports of occupational exposure to aluminas unmixed with other dusts occur in connection with pottery production. In that industry, a-aluminum oxide is used to support green, unfired pottery and flatware.19 330 Alumina-related Pulmonary Disease/Dinman Because of the ease with which this stable form of corundum a-AleOa becomes airborne, exposure of work ers in this industry has been the subject of multiple epidemiologic and clinical investigations in the United Kingdom.13'14 None of these investigators have found exposed to the high temperature range transitional aluminum oxide aluminas since the late 1960s. In none of the foregoing aluminum reduction operations did cell room attendants evidence changes consistent with al terations seen by King et al3 or Stacy and co-workers4 evidence of pneumoconiosis or chronic respiratory dis ease. Reports of exposure to an otherwise uncharacterized alumina which is probably unmixed with other dusts have been provided by Szczekocki and co-workers.ls Although their interests lay in whole body absorption and excretion of aluminum, they briefly note in their with intratracheal insufflation of the low temperature (77) transitional alumina. In contrast to population stud ies, individual case reports from smelters note lung fibrosis93 or fibronodular, alteration.34 Ultimately, the strength of these*1 associations must be borne by the weight of available evidence. clinical evaluations that 49 workers engaged in the reloading of aluminum oxide showed no roentgeno- Discussion graphlc evidence of pneumoconiosis. Alumina exposures in aluminum smelters are mixed, involving such other pulmonary irritants as hydrogen The biomedical literature dealing with the aluminas suffers from ixnprecisions regarding the nature of the fluoride and inorganic fluoride particulates, as well as agent component of the host-agent interaction. Because low levels of sulfur dioxide. The difficulty inherent in they most clearly identify which aluminas were inves the study of mixed exposures among aluminum smelter workers has been commented upon by Saia et al,19 who tigated. the London intratracheal insufflation reports are valuable. In the case of the Stacy group's reports,4 reported pulmonary opacities of a minor degree in cast- a gradation of response to similar doses of several iron foundries where mixed exposures similarly oc curred.17 Similarly, small irregular opacities of category I/O--or at most, 1/1--have been found among workers exposed to coal dust19 and man-made mineral fibers.19 Putting aside the confounding variables involved in different aluminas is demonstrated, despite the dispu tatious route of administration. Had all these dosings led to severe or uniform response to the aluminas, a more plausible case for total disregard of these data might be appropriate. That there was a broad range of mixed dust exposures present in aluminum reduction, response to these agents suggests reason for serious changes in the character of smelting type aluminas which have evolved since the 1950s are of interest. With consideration of their findings. A re-examination of the King group's3 data therefore is useful to the extent one the introduction of flash^and fluid phase caiciners in the alumina refining industry, lower temperatures and faster flow-throughs are achieved. Thus, by calcining understands the limitations of their model's implications for human exposure conditions. We suggest that an association between pulmonary aluminas at a lower temperature, more of the higher temperature transitional forms (cf Technical Note and Fig. 9) and less a-fjuminas are produced; in addition, fibrogenicity, surface area, and catalytic activity (Table 1) is apparent upon consideration of the characteristics of these aluminas. Examination of King and Stacy's among the high temperature transitional aluminas (eg, range of pulmonary response as related to surface area 7, k, 6, and 2), surface areas larger than that associated with a-alumina are formed (cf Technical Note and Fig. 3). This is advantageous, as these high temperature (Fig. 1) suggests an association between these two variables among the noncatalytic aluminas. We would propose that increased thermodynamic instability and transitional aluminas used as smelting feedstock make more effective adsorptive media when they are used to scrub airborne smelter effluents. Thus, as economics enhanced bioreactivity of these compounds occur as surface area increases (cf Technical Note). (Of interest in this connection is Timbrell's recent report,39 which drove this rapid, lower temperature calcination throughout the industry, increasing percentages of transitional 7, , 2, and 6 aluminas have come to be the correlates surface area with the pulmonary bioreactiv ity of various mineral forms of asbestos.) Regarding the catalytic activity of the low tempera major component of the aluminas used in aluminum smelting. At present, smelting grade aluminas--which comprise 90% of the world's alumina production--may contain as little as 1% of a-aluminum oxide, although more usually 80% to 90% of smelting aluminum will consist of the high-temperature transitional forms K, 2, and 6. Under such operating conditions, which have gradu ture transitional forms of alumina (ie, tj, x. 7), enhance ment of pulmonary bioreactivity resulting from their thermodynamically active catalytic character is also suggested by Fig. 1. It would be reasonable to suggest that this property is responsible for that increment of lung damage beyond that due to the surface area factor. Because of gaps in knowledge at the time Stacy and co-workers4 performed their experiments, they could ally applied since midcentury, evaluation of lung not be aware that they probably had not dosed with the changes in aluminum smelter operators since that time could shed light upon the pulmonary reactivity of the 7 specific gamma transitional alumina form. Aside from inadvertent confusion in the literature, this is not a transitional-laden smelting grade aluminas. At present, population studies of smelter employees indicate either minimal 18 or absent fibronodular disease30'91 nor excess mortality39 associated with pneumoconiosis. In most of these studies, smelter workers have predominantly been serious shortcoming from the biomechanistic viewpoint, since the low temperature transitional alumina 17 pos sesses catalytic and surface area characteristics com mon to the other low temperature transitional forms 7 and x- Journal of Occupational Medicine/Volume 30 No. 4/April 1988 VVV 00001*397 331 / All of the foregoing reports of King, Stacy and asso ciates must be considered within limitations imposed by intratracheal insufflation models, the massive dosing and overwhelming of defense mechanisms represents a "forcing" model reminiscent of megadose carcinogenesis experiments. A parallel is exemplified by the large dose animal studies with haiogenated aliphatics: although alternate metabolic pathways are elucidated by such investigations,*6 their relevance to the human exposure conditions is problematic. Although alumina at such doses which negate tracheal-bronchial clearance may produce effects as a consequence of catalytic activity or surface area related thermodynamic instability, usual human exposure conditions produce neither such lung doses, disablement of clearance mechanisms, or signifi cant pathology. Thus, as with the haiogenated aliphat ics, although alternate pathways and pathologies can be induced experimentally, under usual conditions of chronic human exposure, these phenomena are largely of academic interest. In contrast, the inhalation model approaches more closely the exposures encountered at work. Studies us ing this model to evaluate alumina's pulmonary bioreac tivity demonstrate that even the most thermodynami cally active forms x1 and y* do not cause fibronodul&r change. Yet the results of Klosterkotter* serve to cau tion against uncritical application of the inhalation model. Klosterkotter explicitly recognized that the doses he employed (ie, 33 g/ma for eight hours a day for 285 days) represent an exposure level which was intolerable for his laboratory personnel.6 Ultimately, in assessing human health hazards, well designed studies of working populations potentially at risk assume considerable importance. Review of occupational exposures to (1) catalytically active large surface area, low temperature transitional aluminas (x, tj, y) in aluminum chemical refineries; (2) high temperature y, x, 3, and 0 transitional aluminas encountered in smelters since the late 1960s; and (3) high-fired a-aluminum oxide as found in potteries sug gest that of all these aluminas demonstrate minimal-- if any--fibrogenic potentials in man. It is ironic that, although concerns regarding 7-alumina's pulmonary fibrogenicity have persisted (eg, in the standard-setting process*7) since the London studies, (1) a specific 7- alumina, ie, the transitional form, was probably not actually studied by that group and (2) the other socalled 7-aluminas exhibited inconsistent biologic re sponses (eg, the nonreactivity of 7-AIOOH Cera hydrate v the fibrogenic response to 7-AIOOH HX1010 gelati nous boehmite). The final irony is reflected in the obser vation that the only true exposure to a specific 7 transitional alumina (ie, by Klosterkotter) via inhala tion does not produce a specific pulmonary fibronodular response, but rather the nonspecific alveolar proteinosis. All the foregoing data place in high relief the propo sition that it is only through critical assessment of both experimental and human experience with environmen tal agents that we may understand such compounds* role in disease induction. On the basis of in vitro and in vivo experimental and workplace studies under condi tions of human exposure, it is clear alumina's fibrogen icity is quite low regardless of which specific alumina is examined. Technical Note: Physical, Chemical, and Structural Nomenclature, Identification, and Behavior of Aluminas The Physical and Chemical Bases for Confusion The oxides, hydroxides, and oxyhydroxides of alumi num have been given a multiplicity of designations since aluminum oxyhydroxide was first chemically analyzed and identified by Vauquelin in 1802. Investigators gave different names to the same chemical compound or mineral form (cf Table 2); this led to confusion as different physical properties, eg, crystalline structure, were assigned to the same compound. For example, the minerals gibbsite and diaspore, identified in the first and third decades of the 19th century, were later con fused with their isomorphs, which were given the names boehmite and bayerite in the 1920s. To compound this multiplicity of appellations, the name hydrargillite was applied by the Europeans to gibbsite. Duplicative and inconsistent terminology ensued when Greek letter prefixes were attached to various aluminas for unrelated reasons or behaviors. Rankin and Merwin*6 in 1916 assigned the Greek letter prefix /3 to a high temperature alumina later shown to contain alkali or alkaline earth atoms. Haber29 in 1925 divided TABLE 2 Companion of Physical and Chemical Nomenclature Applied to the Alumina Minerals wv 00001A39S Preferred Chamleal Formula and Name** AlfOHJa, alumina trv hydroxide AKXJH aluminum oxyhydroxide Aluminum oxide Mineral Name** Gibbsite or Hydrargiilite Bayerite Nordstrandite Boehmite Diaspore Corundum Formerly Applied Designation** Alumina tnhydrate Aluminum hydroxide Alumina trihydrate Aluminum hydroxide Randomite Alumina monohydrate Alumina monohydrate o-Alumina American Syatefn***' <*. 7 a, 0 -- a. y a. 0 a European System** 7 7 7 7 a a X-ray Cryatatlograph Classification*4 7 a 0 7 a a 332 Alumina-related Pulmonary Disease/Dinman aluminum oxide into an a and y series, depending upon the product of the oxide's calcination. Subsequently, Edwards et al90 assigned the prefix a to the more abundant hydroxide, whereas the less abundant alumina was given the j8 prefix; both appelations were applied without regard for structure. Later, as these aluminum compounds were used in industrial processes, different nationally based varient nomenclatures--North American or European--were used to describe the same compounds. The resulting confusion in chemical and mineral nomenclature is am ply demonstrated by Table 2. This ambiguous situation was further complicated when Ulrich31 in 1925 found that progressive heating of the hydroxide gave rise to a previously undescribed form which he called 7. This material was later found by Stumpf et al3* to be but one of the several alumina transition forma x> and y (Fig. 2). However, before these specific, individual forms were defined, the Greek prefix y, attached by Ulrich to these forms, came to be applied via common usage to all three low temperature calcination products of aluminum oxides and oxyhydroxides, ie, 77, x> and y. It was not until 1950 that Stumpf et al3* restricted the prefix y to one specific thermally induced transition form derived from gibbsite and boehmite formed between 500 and 850aC (Fig. 2). In succeeding years, the identification of each of these thermal transitional reorderings was established and Greek alphabetical prefixes were given to these specific forms (Fig. 2). However, it must be recognized that what appears to be a straightforward sequence of individual, defined phases or polymorphs of aluminum oxides represents in reality thermodynamically unstable states of structural reorderings of the hydroxide lattice. Although these states are reasonably reproducible, even within a solid, there may occur simultaneously various domains char acterized by different degrees of structural reorderings. The type of transition form and the sequence of trans formation depends upon variables such as the structure of the precursor, rate of heating, particle size, and components or contaminants in the atmosphere. Accord ingly, unless details of all such variables are reported, critical evaluation and interexperimenter comparisons of biologic response are unreliable. The structural and compositional differences among these several transition forms, e.g,, between j? and 7 forms, are small and difficult to distinguish. In attempt ing to identify precisely these low temperature transi tional forms, 77, x, and 7, we are faced with a double dilemma. Not only are these transitional states identi fied with difficulty, but structural analysis cannot pre cisely quantify the ratio between the amorphous and crystalline states within a specific specimen. Ultimately, in considering these low temperature transition forms, one faces the reality that these aluminum oxides rep resent highly dynamic reorderings of crystalline struc ture and amorphous elements temporally and spatially distributed in a nonhomogeneous, random fashion. It should be readily apparent why, despite general accept ance of the Munster Symposium33 recommendation giv ing each transitional fora an individual, specific Greek prefix designation (Fig. 2), present industrial/commer cial nomenclature often fails to differentiate between --------------------- nr.-------------------- PHkk UaMk|laM >1 ua. Moiaalr >IQ0*kMM turn. <10 hIcmm VVV 000014399 felSSSrTE CHI BOEHMITE fearERrre ETA DIASPORE H] KAPPA ALPHA GAMMA | DELTA IHE1a|aLPHA ---------- THETA [ALPHA ALPHA -----------I------------1------------>'* I I , 100 200 300 400 500 600 700 600 900 1000 1100 1200 TEMPERATURE, C Fig. 2. Decomposition sequence of aluminum hydroxides. The transition forms (chi. eta. gamma, kappa, delta, theta) represent states of thermally induced structural reordering associated with dehydration. The x. 7, and 17 forms have been genericaNy referred to in the past as 7alumina.1* Journal of Occupational Medicine/Volume 30 No. 4/April 1988 333 these low temperature transition forma. Given these multiple ambiguities arising from a highly dynamic system, we recommend that, rather than attempting to ascribe biologic effects to individual aluminum oxides, eg, to the y form, the transition forms tj, x> and y collectively be referred to as low temperature transi tional aluminas. To complete the confusion, international crystallo graphic convention prefers application of the prefix a to the closest packed--especially hexagonal--struc tures; the y prefix is given generically to the cubic packed structures.*4 (Nordstrandite (/5-Al(0H)a) pos sesses a crystalline packing arrangement intermediate between gibbsite (7) and bayerite (a) and therefore is designated 0.) In summary, it should be readily apparent that the prefix y has been applied for a multiplicity of reasons both to crystalline alumina minerals as well as to the transitional form. In addition to the foregoing relatively well-defined crystalline minerals, various syntheses of aluminum hy droxides produce colloidal aluminas which easily coag ulate to a two-phase system gel, ie, a highly dispersed colloidal solid with molecular water occupying capillary interstices. In particular, one of these gelatinous alu minas, gelatinous boehmite, is pertinent to our concerns. It demonstrates minimal crystalline structure, consist ing of molecular water situated between the elemental structure layers. This leads to an enlargement of the basal apacings of boehmite. These gels can be dried to extremely fine (10"9 fim) and porous particulates pos sessing extremely large surface areas, as large as 500 mVff.34 . TEMPERATURE, *C Fig. 3. Specific surface area changes resulting from neat treatment Of AKOHfe.* Functional Properties of Those Aluminas of Potential Biologic Interest VVV 000014400 Regarding a and y designations derived from the crystallographic convention, no apparent common prop erties with biologic implications are presently known to occur. Because they are applied to a diverse and large number of aluminas (Table 2), with different chemical and physical forms, any lack of commonality of biologic behavior is not unexpected. Concerning the thermally induced transitional forms (Fig. 2), the physical-chemical properties of these forms may be of importance in biologic systems. Heating alu mina minerals under specific conditions (Fig. 2) induces these transition forms; such forms characteristically demonstrate a progressive disorganization and reorga nization of crystalline structure as temperature rises through the 250 to 850*C range. Structurally, lattice voids and a network of submicroscopic cracks and spaces open.34 , The newly formed surfaces presented by these fissures and voids result in large surface areas characteristic of ij the low temperature transitional forms (Fig. 3). Con comitant with the loss of water resulting from the combination of OH- ions (OH- + OH" --* H0 + O), exposed Al* sites are formed; these act as Lewis acid sites, ie, electron acceptors. Furthermore, certain re tained OH" ions act as proton donors, Brpnsted acid sites. Depending upon the local composition, O" or OH" ions may also function as Brpnsted or Lewis bases. The aggregate of these physical-chemical alterations occur ring during thermal transformation is believed respon sible for catalytic activity of these alumina transition forms.38 The 7, 77, and x transition forms maximally manifest these crystalline disorderings; they likewise manifest the maximum catalytic activity among these transitional forms. As an additional consequence of these thermal trans formations leading to the transitional aluminas, at ap proximately 300 to 600C a marked increase in surface area occurs34 (Fig. 3). This is a result of the loss of over 30% of the mass without change in the external dimen sions of the crystals. With continued heating of these transitional aluminas beyond 850"C, crystalline order begins to be re-established and chemical stoichiometry is restored. With these balances restored, cat&lytically active sites diminish and ultimately disappear. The voids and channels are diminished and surface area becomes markedly decreased (Fig. 3). This reordering and ther modynamic stabilization reaches a maximum with the formation of a-AlaOa, corundum.34 As a consequence of the dehydration, a remarkably large surface area forms. Concurrent with the rise in ^surface area, there ie proportionate increase in surface thermodynamic instability.38 It is not unreasonable to v suggest that the combination of chemical activity (ie, the acid-base character of the low temperature transi tional alumina surfaces) and the large surface area per se are important factors controlling bioreactivity. 334 Alumina-related Pulmonary Disease/Dinman Relation of Particle Dimension to Carcinogenicity in AmpKffiVK i Asbestoses and Other Fibrous Minerals 12 hAV 25'87 `1 i RJA_______________ Mearl F. Stanton, 3- * Maxwell Layard, 5,6 Andrew Tegeris,7 Eliza Miller, 3 Margaret ffflay,j 4 Elizabeth Morgan,7,9 and Alroy Smith 5 PQ -------------------------- FILE ABSTRACT--In 72 experiments, durable minerals in the form of MATERIALS AND METHODS F UP---------------------- panicles on respirable size and of wide chemical and structural varieties, were implanted in the pleurae of outbred female Osborne- Mendel rats for periods of more than 1 year. The incidence of induced malignant mesenchymal neoplasms correlated well with the dimenstonai distribution of the particles- The probability of pleural sarcoma correlated best with the number of fibers that measured 0.25 pm or less in diameter and more than 8 in length, but relatively high correlations were also noted with fibers in other size categories having diameters up to 1.5 pm and lengths greater than 4 pm. Morphologic observations indicated that short fibers and large-diameter fibers were inactivated by phagocytosis and that negligible phagocytosis of long, thin fibers occurred. The wide variety of compounds used in these experi ments suggested that the carcinogenicity of fibers depended on djrnension.anfl-rtitrahiiilv-mhar:. than on ohvsicnchermcal prop erties.--JNC1 1981; 67:965-975. None of the methods were appreciably different from those described in earlier papers (4, 6, 9-11). Con sequently, only modifications of methods are detailed here. A standard 40-mg dose of panicles uniformly dispersed in hardened gelatin was applied by open thoracotomy directly to the left pleural surface of 12- to 20-week-old, outbred female Osbome-Mendel rats. In each experiment, 30-50 rats were treated and followed for 2 years, at which.time the survivors were killed. All rats were necropsied and all lesions examined histo logically. A positive response was the occurrence of pleural sarcomas that resembled the mesenchymal meso theliomas of man, developing after the 1st year (12). Three types of controls were considered: untreated rats, rats that received thoracotomies but no pleural implant, and rats with pleural implants of nonfibrous material. There were two types of spontaneous tumors that Work in several laboratories has indicated that di could cause confusion: the fibrosarcomas of left mam verse varieties of minerals are carcinogenic when ap mary glands and the subcutaneous fibrosarcomas in plied directly to the pleura of the rat or hamster in the duced by suture material. Vigilance and early surgical, form of microscopic fibers, i.e., particles with dimen- removal accounted for most mammary tumors; the use t 1 sional aspect ratios of 3:1 or greater (1-9). The same .ninerals are much less carcinogenic when applied at equal weight and size in nonfibrous form. Further, preliminary experiments indicate that carcinogenicity correlates best with increasing numbers of fibers having both diameters of 0.25 pun or less and lengths of more than 8 and that the correlation diminishes with Abbreviations used- alumin=aluminum oxide; attapu)sattapulgite(s|; crocidscrocido)ite(s); dawson =*dawjonne(s); halloy *halloysite(s); UICC = International Union Against Cancer: woliaston- wollastonitefs). fibers of greater diameter or lesser length. Conse quently, a reasonable conclusion is that the long, thin, fibrous structure is critical to the carcinogenicity of these minerals. Studies on fibrous samples within very narrow dimensional ranges would be valuable in the establishment of this hypothesis, but these ideal sam ples are not available. Consequently, we are faced with the correlation of carcinogenicity with fiber samples of widely mixed dimension. The purpose of this report is to correlate our best estimate of fibrous dimension with carcinogenicity for all those minerals that we have studied that are both durable and within the size range of respirable particles. This involves 72 experiments with minerals of wide chemical and structural variety. Of special interest are the data on the amphibole asbestoses: amosite, tremolite, and crocidoliie, though estimates of the dimensions of the asbestoses are especially liable to error. Chrysotile, although as car cinogenic as the amphiboles at comparable dimen sions, could not be included since it has proved | difficult to be measured with any degree of precision. 1 Received November 13, 1980; revised May 6, 1981; accepted June 8. 1981. 1 The guidelines for (he care and use of laboratory animals were followed as set forth by the Commiuee on Revision of the Guide for Laboratory Animal Facilities; by the Guide for the Care and Use of Laboratory Animal Resources, the National Research Council; and by the National Institutes of Health. 1 Laboratory of Pathology, Division of Cancer Biology and Diag nosis, National Cancer Institute (NCI), National Institutes of Health. Public Health Service. U.S. Department of Health and Human Services, Bethesda, Md. 20205. * Deceased. 1 Biometry Branch, -Division of Cancer Cause and Prevention, NCI. * Address reprint requests to Dr. Layard at his present address: Veterans Administration Medical Center. 3801 Miranda Ave.. Palo Alto, Calif. 94304, 1 Pharmacopathics Research Laboratories, Inc., 9705 North Wash ington Blvd., Laurel, Md. 20810. * Present address: 5524 Trent St., Chevy Chase. Md. 20015. * Present address: Triangle Resource Industries, P.O. Box 599. Laurel. Md. 20707. / V} -i *4 r.-sj VVV 00001A40X JNCI. VOL. 67. NO. 5. NOVEMBER 1981 t t* 966 Stanton, Layard, Tegeris, et al. of synthetic, biodegradable, polyglycolic acid sutures largely eliminated suture sarcomas. An equivocal diag nosis for the origin of a tumor was necessary in less than 1% of the tumors. The probability of pleural sarcoma in each experiment was calculated by an actuarial life table method that accounts for early deaths without pleural sarcoma and provides a good means of making quantitative comparisons of one experiment with another. Details of this method are given in (13, 14). The fibrous materials used in these experiments were mostly commercial products that were submitted by the manufacturers from an interest in their potential car cinogenicity. Consequently, they were used as received and were not especially refined except in our efforts to separate particles by size. NTone of the preparations appeared overtly contaminated by other materials when examined in the electron microscope. A few of the small-fibered subfractions of the fibrous materials were obtained by ball milling in a steel ball mill and consequently were contaminated with fragments of steel. In general, subfractions were obtained by simple gravimetric methods in aqueous media to separate fibers of different dimensions. These maneuvers in cluded sedimentation, centrifugation, and filtration, which in some instances were also responsible for the reduction of the size of the particles but did not otherwise alter the particles physically or chemically. Eleven chemically and structurally different groups of fibers were available for study, and samples studied are listed in text-figure 1 and table 1. Six major groups of particles had multiple dimensional ranges; these in clude; crocidolites: (samples crocid 1-13), glasses (glass 1-22), aluminum oxide whiskers (alumin 1-8), talcs (talc 1-7), dawsonites (dawson 1-7), and wollastonites (wollaston 1-4). Seven additional types of particles had only one or two dimensional ranges. These were the amphibole asbestoses tremolite (tremolite 1, 2) and amosite, the clays attapulgite (attapul 1, 2) and halloysite (halloy 1, 2), crystals of silicon carbide and potassium titanate (titanate 1, 2), and nickel titanate (titanate 3). All of these materials have been described elsewhere {4, 6, 10, 11, 15-18), but the following information is pertinent. Crocidolite (crocid 1-13).--These 13 samples of South African crocidolite (an amphibole asbestos) were from four different sources. Samples crocid l, 3, and 9 were prepared in our laboratory from a single sample of hand-cobbed, unmilled ore. The ore sample was hand milled without exposure to any metallic ma terials and reduced to the approximate size of com mercial crocidolite. Samples crocid 6, 7, 8, 11, 12, and 13 were all prepared in our laboratory by various milling, sedimentation, and flotation methods from a single lot of standard UICC crocidolite designated crocid 5. Differences in dimension were the result of different milling times. Crocid 5, the original UICC sample, has been characterized in {19, 20-23). Samples crocid 4 and 10 were specimens prepared in a com mercial laboratory from a single separate sample of jNCa. VOL. 1*7. NO. h. NOVEMBER South African crocidolite and separated by centrifu^. tion to obtain mutually exclusive size ranges from thr same sample {24). The remaining sample, crocid 2. wa obtained from Dr. J. C. Wagner (Medical Research Council Pneumoconiosis Unit, Penarth, Wales) a* representative of the material used by him in Hu original experiments (25). It was our impression that any mechanical manipulation of these samples could both reduce the size of the particles by fragmentation and effectively increase the size of the particles bt clumping. For this reason, probably the dimensional measurements on crocidolite are the least representative of all the fibers measured. Glass (glass 1-22).--The first 18 of the 22 glasses were borosilicate glasses that have been previoush reported and can be recognized from those publications by their letter designations (4, 10). Glasses 12, 14, 15. and 18 were preparations of typical large-diametered insulation glass fibers that were coated with a phenolformaldehyde binder. In the early experiments, glass Id was used as a control and also served as a vehicle for the implants. Glasses 19 and 20 were preparations oi large-diametered fibrous glass that was leached to remove all elements except S1O2. These two glasses were exceptionally fragile and contained many irret;ular fragments. Glasses 21 and 22 were large-diametered extruded fibers with a microcrystalline aluminum ox ide content greater than 80% (glass 21) and with j microcrystalline zirconium oxide content greater than 90% (glass 22). Aluminum oxide (alumin 1-8).--The 8 samples of. aluinfnum oxide were all crystalline sapphire whiskers prepared by General Technologies Corporation. Reston. Va., or by Thermokinetics Fiber Incorporated, Nutlev. N.J. (15-18, 26). All of the samples were processed and selected for dimensional ranges. Of the samples, 3 were exceptionally noteworthy. Sample alumin 8 was nonfibrous, sample alumin 3 was exceptionally fine but tended to cluster in nonfibrous balls, and sample alumin 4 contained whiskers of aluminum nitride as well as aluminum oxide. Talcs (talc 1-7),--All seven talcs were refined raw materials for commercial products. Each was from separate and diverse source and selected to include ail extreme ranges of dimension. Platelike structure was consistent and was considered in the calculation of the volume (15-18). Dawsonite (dawson 1-7).---The 7 dawsonite sam ples (crystalline dehydroxv sodium aluminum car bonate [NaAhOHhCOj] were from several sources. The characteristics and synthesis of dawsonite can be found in (27, 28). Samples dawson 2 and 3 were synthetic crystals prepared by a commercial company (for dawson 2) and by the Bureau of Mines, U.S. Department of Interior (for dawson 3). Sample dawson 4 was a natural crystalline dawsonite from the Olduvai Gorge, Tanzania. The remaining 4 samples (dawson 1. 5, 6. and 7) were synthetic crystals from a second commercial company. These 4 samples were especiall' crystallized and sorted to achieve narrow ranges of size. VVV 000014402 O ltm a ta r Carcinogenicity of Fibrous Minerals 967 (1) TitsnaT* 1 ss% > 4.0-a.o >25-40 >15-2.5 > JO-15 > .25- 50 >.10-9S 3.57 287 4.41 6.C8 >.05-.10 3-27 502 > .01 -.06 > BO 3.75 4.79 52A 502 4.11 *SJ 6.17 4.44 (4) Pinion 5 100% >4 040 >2*4.0 >1*2:S >.50-1.5 > 95-.50 394 2.18 493 >.10-95 283 >.0*.10 >.01 -.05 >ao 137 137 495 3.73 202 250 4.53 493 (71 Dawson 1 95% >4.oa.o >2.54.0 >1*25 1.44 >30-1.5 1.92 132 > 95-30 1.7S 3.16 >.10-95 2.52 3.70 >'.05>-.10 3.40 3.62 >.01-.05 3.57 3.57 >8.0 1.44 292 232 3.90 3.82 2.14 3.00 1.75 3.68 | 495 4.66 162 (10) Good. 3 33 * >4.0-8.0 >254.0 >1.525 >.50-1.5 > -2S-.50 >.10-25 6.12 > .05-. 10 652 >.10.06 5.S2 5.71 584 4.58 >8.0 4.10 4.10 4.95 4.10 4.10 4.10 4.40 5.01 <131 Gass 1 fMOU > 40-8.0 >254.0 >1.5-25 >3CV1-5 293 >95-.50 303 >.10-95 2 93 > .05-.10 >.01-.05 >8.0 2.53 335 3.33 3.46 3.23 3.C8 4.95 463 479 465 (16) Glass 3 IKU ^ >4.08.0 >254 0 >1.5-25 1.4S > J0-1.5 2.95 > 2S-.S0 256 3.16 >.10-95 3.03 3.16 >.O5-.10 ies 4.09 >.01-.05 3C3 U m >.01-1 3.73 >14 >80 0.67 0.67 2.40 3.33 3.76 3.03 >4-8 0.67 1.52 0.97 203 219 342 2.74 383 3.03 395 303 >8-64 >64 CO Twuii 2 100% 3.70 asa 425 3.40 4.18 4.39 491 (5 Twnoft* 1 3.70 4.52 6.07 4.78 3.40 4.18 4 48 491 214 3.84 4.47 4.10 4.35 490 4.14 3. 354 J_______ <81 Good-1 94% 214 3.S2 3.14 144 214 214 3.92 3.14 244 214 O) Si Car&ds 2.92 393 0.81 3.78 22S 3.08 3.89 4.01 491 4.48 (61 Tremo>>t*2 182 1-82 204 303 490 4.19 4 41 4.54 1.35 1.85 1.35 1.95 394 195 4.40 4.58 4.74 4.70 3.44 4.09 3.14 214 4.05 386 4.04 284 254 284 368 3.38 3-38 3 01 254 3.77 231 284 <91 Good. 2 5.75 6.52 587 4.07 5.69 562 4.67 4.07 4.67 467 4.56 5.1$ 497 3.53 492 5.56 5.71 4.57 4.53 4.63 5.57 4.96 3.83 497 4.70 4.73 4.30 492 4.W 4.00 ODA/nears S3% 213 3.13 3.53 2.98 1 217 3 64 276 298 236 3.56 231 3.03 (12) Good. 4 213 2.43 3.54 213 261 172 147 261 394 2.76 394 276 494 433 4.31 236 394 3.61 4 50 4.S3 3 54 3 84 306 4.36 4.32 4.70 306 n4> Crcdd. 6 3.59 4.35 484 4.74 4 42 281 4 19 4.3S 4.65 463 496 2.81 2.81 399 342 3 66 3.77 342 281 281 3.77 2.81 3-59 399 (15) Glass 2 (M6D) 3.51 2.91 3.33 369 4.08 3.51 391 2.91 361 3.69 321 3.99 3.61 4 11 402 3.81 3.33 1171 Glass 4 (M6U 2.81 244 244 244 >.01-1 2.05 244 352 3.44 3 44 > 1-4 1.84 259 3.16 3.44 3.56 3 35 . >48 1.53 193 276 213 3.70 290 339 262 390 3.14 >8-64 >64 3.12 >.CM 190 296 > 1-4 1.10 1.62 1.61 1.93 GEO 257 1.73 >4-8 190 1.00 0.93 1.90 1.43 217 i sa 2.78 291 242 3.51 273 2.63 >864 >6* 1vJ'-ncii'JiE ].--Fiber distribution by common log of the number of particles per microgram in each of 3-i dimensional categories. ^ ooJ4403 JNCI. VOL. 61. NO. 5. NOVEMBER 1981 968 Stanton, Layard, Tegeris, et af. =* 5 (KW> > 4.03.0 > 25-4.0 13-2.5 0.66 2.48 1.30 > 25--50 1.06 0.57 >.10-2S >.05-.10 > XII-.05 > 8.0 037 129 1.70 1.67 1.67 006 0.36 0.06 0.76 0.06 _ 1-06 , 1.41 136 2.4Z 251 275 2.71 2.51 221 GO) Dw*on. 7 68% 1.77 3.11 272 207 3.01 4.18 440 3.70 272 326 4.10 4.64 342 1.77 3.33 3.81 425 4.70 272 QllDwwn.4 66% 211 3.38 6.30 5.03 127 222 211 3.42 4.05 6.00 347 1.66 1.57 222 324 442 4.12 3 99 1.87 1 208 I 1.81 1 224 i -- - 248 | 2.B0 4.01 Q2] D**wcn.3 >4.06.0 > 25-4.0 > 1.5-2S >.50-1-5 269 >350 4.18 4.50 >.10-25 391 4.74 >.0500 58S 6.18 >.01-.05 5.90 517 >8.0 3.07 428 4 41 4.88 428 129 - 229 123 4J7 521 529 5.14 C3J Gim 6 fM6W) 223 1.40 1.70 210 271 1.10 1.10 3.17 234 322 356 295 1.C9 218 321 3.11 32S 3.47 271 | 1.10 ) 125 274 j 263 1 1.53 MO 20! 29t 2S4 3.69 276 280 G4) Crocad. 0 2% 3.47 4.17 5.10 4.74 3.65 4 56 6.07 545 4.90 4.07 432 4.71 4.86 421 4.02 4.45 457 247 CSJOodd. 7 >4.oa.o > 25-4.0 > 1-5-25 >.50-1.5 >25-.50 >.10-25 >.0600 >.01-.05 1.69 205 1.86 >8.0 0.91 1.69 121 1.91 9.91 0.91 121 VS1 126 225 202 229 234 251 2.05 0.91 (2S) Crocad. 8 4.45 5.08 5.61 4.55 523 3.75 4.7S 4.86 5.09 4.59 4.15 345 3.4S 3.75 3.75 (Z77AJumat.2 237 1.99 1.07 1.48 204 1.11 120 120 0.91 1.15 1.18 1.60 1.70 1.50 120 132 0.37 1.32 037 1.53 1 CO 1.61 1.18 237 1.83 302 289 138 120 Co) AJumm. 3 > 4.0-8 0 >25-4.0 1.09 > 15-2.S 1.59 >.50-1.5 236 242 > 2S.50 0.97 >.10-2S 0.97 232 >.05-.10 127 >.01-.0S >8.0 1.19 1.57 1.70 233 1.15 219 0.97 1.72 1.75 0.19 1.82 1.70 261 0.67 T.49 241 1.59 Q9) Crood. 9 3.13 3.43 4.5! 524 548 3.13 425 492 5.14 3.43 425 4 34 3.13 3.13 3.13 4.09 3.74 PC) WotWston 1 1% 3.15 3.1 S 328 3.41 211 3.11 3.78 325 322 328 2.81 306 115 211 2.41 211 3.19 211 259 Ol) Alum'm. 4 > 4 0-8.0 >254.0 2.15 >`1-52.5 2.ec >.50-1.5 292 2.83 >25.50 3.57 3.01 >.1525 >.05.10 229 >.01-.05 1.99 >8.0 1.88 248 230 245 3.19 1.99 229 2.69 . 1.75 1.96 215 237 275 128 0.98 283 229 233 G2) Crock!. 10 3.79 4 69 503 S 13 496 3.10 3.10 3.57 4.57 501 4 59 44 3.40 3.57 3.79 3.10 | j j 3 57 3.10 I 3.10 (JJf Alumuv 6 22% 3.59 3.S5 300 1.48 266 177 424 4 13 4.02 3.70 1.78 218 271 3.41 422 424 3.92 252 2.C8 208 2.59 326 1.48 3.92 3.57 3.23 041 Glass 20 Ifm 22% > 4.03.0 > 2.53.0 > 1.5-2.5 >*1.5 >2S-.50 >.10-25 >.0S-.10 > .01-.06 ><m >.01-1 >13 >80 >43 >534 088 0.98 >4 OS) Glass 7 (KCP) 300 324 324 250 220 >JI-1 1.4* 2.05 3 59 3 88 32B 290 296 >1-4 1.81 205 217 3 17 3.10 310 767 >44 0.97 201 1.81 1.81 2.31 0-97 285 201 255 250 >fr64 >64 36) WoOMtcn. 3 19% 312 3.32 3.18 3.42 258 3.53 4 09 3.32 3.32 348 258 258 3.12 346 228 229 258 276 >.01-1 >44 > 9-64 >64. Lan^tfi Ttxr ncifRt 1 (continued^--Fiber distribution by common I05 of the number of particles per microgr.nm in each of 3-1 dimensional categories JNCI. VOI.. 67, NO. 5. NOVEMBER H1 VVV 000014404 Olsmatar Carcinogenicity of Fibrous Minerals 969 V- 1 20% > 4 0-0.0 > 2.5-4.0 > 1.5-2.5 > 8.0 >2S-S0 >.10-25 > .06-.10 > .oi-.os 6.47 6.07 526 6.95 638 6.21 435 6.16 OS) Kaiioy. 2 Z2% 635 6.16 633 4.61 4.51 647 6.14 689 4 51 4.51 4.81 D9i Glass 8 (KUP) 288 3.36 3.36 3.17 2.11 273 333 3.54 350 3.32 2.31 _______ L 136 2.04 2S2 2.31 271 244 174 1.26 3.54 2E4 3.19 1.96 357 251 2.63 2.31 280 2.31 (40) Crood. 11 >4.08.0 > 23-4.0 2.38 > 1.5-2.5 3.03 >30-13 3.15 2.98 >25-30 3.05 130 >.10-^5 2.97 > .05-.10 2.02 > 31-.05 >8.0 1.72 2.42 1.72 130 2.15 2.50 1.42 (411 GulS IHD) 2.56 267 136 2.16 292 2.37 236 1.73 136 2.73 243 i 136 M2) Glass 9 (M8U 1.49 260 1.85 0.77 1.12 1.45 0.17 1.73 2.02 2.11 237 235 242 238 135 in, 6 13% >4.0-80 > 23-4.0 0.48 > 1.625 0.82 >.0-1.5 > 2S-iO 0.75 132 210 I 0.82 >.10-2S > .05-.10 > .01-.05 >8.0 0.12 0.67 0.12 032 037 032 132 1.44 0.67 0.70 1.54 1.49 037 0.63 1.84 0.62 (44) Dawson. 6 UTS 6.17 636 7.861 6.47 6.86 6.78 (45) Dawson. 2 12% 3.80 4.57. 432 432 4.39 294 3.43 346 3.31 3.10 244 282 274 244 1.74 263 244 1.74 (46) Wollaston. 2 > 4.08.0 >2.54.0 278 > 1.5-2.5 >30-1.5 3.22 278 3.80 >35-30 343 3.32 >.10-35 3.96 336 > .0500 4.37 4.00 > J1-.05 >8.0 278 3.18 3.18 296 248 240 (471 Crotfic. 12 (481 Artapii. 2 287 3.31 4.00 4.00 3.67 i 3.31 3.73 421 4 00 3 SO 237 3.55 3.39 3.53 3.45 3.17 257 237 3.39 3.06 3.69 3.57 3.17 239 5.31 6.16 7.05 6.14 S.54 632 639 6.51 1 i 1 (49) Glass 10 (MOSJ >4.08.0 > 25-4.0 297 1-5-2.5 >50-1.5 3.88 3.43 3.91 > 2S-.50 >.10-35 > .05.10 > .01.05 4.34 443 530 6.77 4.02 3.88 4.19 463 >8.0 246 239 237 276 3.69 272 1.76 1.17 2.46 3.37 (50) Glass 11 OOP] 4.12 438 4.11 331 289 3.64 4.53 3.02 2.as 3.09 3.56 3.60 1.18 241 266 1.81 (51) Titanaia 3 3.54 3.35 3.3i 3.1Q 2.95 308 1.78 1 (52) Artapii. 1 SSL > 4 0-8.0 > 254.0 % 1.52.5 > .50-1.5 5.12 > 2S.50 5.48 > .10-25 5.75 6.50 > -05.10 6.62 6.65 > .01-.05 6.96 6.65 jim >.01-1 >1-4 > 8.0 4 64 494 >4-8 >8-64 >54 S3} T*ic 1 4.00 4.15 4.19 3.70 364 4.02 4 56 3.97 3 91 327 3.09 3.70 327 279 279 3.09 > JOl-1 >1-4 >48 >564 >64 (54) Glass 12 (02P) 2 83 2.44 208 241 206 2.43 2 83 0.48 1.65 232 2.48 1.95 1.91 1.68 1.78 217 2.09 1.94 1.84 0 48 0.48 0.48 0 48 048 >.01-1 > 1-4 >4-8 >864 >64 Ungtf) Tixt-fici-RE 1 (continued). Fiber distribution by common log of the number of particles per microgram in'each of 34 dimensional categories. VVV 000014405 JNCI. VOL. 67. NO. 5, NOVEMBER 1981 "'TWhr-' **man *m *-'***w wpst. aaili 970 Stanton, Layard, Tegeris, et al. Dlamater EG) Giro* 13 fXFPl > 4.0-3.0 > 2.54.0 > 1.5-2.5 > .50-1.5 3.19 > 25-50 296 1 289 320 3.48 1.22 >.10-25 3.30 > .06.10 3.27 > .01-.CS > 8.0 2.57 2.73 1.92 1.52 1 j | 1 2.03 2.03 0.S2 . 032 fbol 14 [Kirt O'* 2.56 121 236 135 2.30 22a 2.51 1229 0.81 135 2.08 0.91 1.21 2.00 2-21 1.70 0 63 0.81 1.06 C571 Os-- 15 (Y2P) 6% 3,42 1.94 2.53 2C9 1.08 2.12 1.74 1.60 312 1.60 1S6 224 216 1.64 1.60 o sir 1.X CS81 Afumin. 7 6% >mo >2&4.0 YsT >1.5-2.5 0.02 >.50-1.5 0.64 t:is > 2S-.50 1.17 T5T > .10-25 >.0S-.10 > .01-.05 >a.o Tis" 2S2 TTF X70 Toe" T.Z7 0.31 TST Ta T43 115 0J1 TaT 0.51 S3) CZia* 16 (MBS) 3.41 3.83 3 .SI 28S 2.21 3.17 292 243 275 343 3.46 r2.45 2u 290 3.59 281 127 KOI Talc 3 4% 4.13 3.76 3.81 323 281 4-34 4.85 253 3.81 241 281 161} Talc 4% > .50-1.5 > 2S.50 >.10-25 > .0S-.10 > .01-.05 > 4 0-8.0 254.0 258 1.52S 3.06 2-28 3.26 >8.0 276 3.21 2.45 245 268 256 2-28 1.98 fra Ta* 4 406 3.40 438 262 4.19 4.81 4 58 422 3.32 3.32 3.32 279 3.02 (63) Alurrun. 8 3% 272 286 271 222 242 287 246 212 1.99 225 234 1.70 1.72 1.86 0.74 1641 Glass 21 (SI) > 4 06.0 >254.0 1.15 >1.5-2.5 1.15 >20-1.5 1.46 206 > 2S-.50 1.15 1.46 > .10-2S > .os-.io > .01-.0S >80 1.15 1.85 1.63 1.1S 2.3$ 1.46 1.46 2.00 1.15 165) Glass 22 IS2J 1.47 1.47 1.87 1.47 1.77 1.17 1.47 1.17 1.17 2.07 225 1.87 1.77 (66) Glass 17 (MSS) 4 43 5.17 560 545 5.16 245 4.83 4.68 4.73 428 | i 4.38 1.87 358 2.85 3.79 | 2.15 (67>Gl*sa 18 (VW) 0% > 4 0-0 0 > 25-4.0 >1.5-2.5 >so-i.s >.as-.50 > .10-.25 > .Q5-.10 > .01-.05 >8.0. 0.83 0.92 080 1.00 1.10 0.34 0.40 0.30 0.11 0.41 tea Dodd. 13 4.30 4 68 4 SO 468 431 2S3 4.X 4 46 4 CO 2S3 268 2.98 298 3.16 268 (69) Wollaston. 4 1.43 0.95 095 1.43 125 2.07 1.88 0.95 1.86 0.95 1.99 095 1.43 1.56 1.60 1.91 0 95 125 (70) Talc 5 0% >4.oa.o > 25-4.0 343 > 1.5-25 4.33 > .50-1.5 4 62 4 98 > 2S-.50 4 56 420 >.10-25 427 350 >.0S-.10 3 97 360 >-.01 -.05 390 3 13 > .01*1 > 14 >80 >4-8 >8-64 >64 f71J Talc 6 4 S3 4 83 468 4 65 4 10 325 3.95 4 S3 443 3.95 3 e$ 32S > 14 325 325 3.55 355 325 325 3.55 '3 55 325 >4-8 > &-64 >64 (72) Talc 7 4.95 5.33 5.18 5.14 4.82 > .01-1 4.37 5.09 493 4 37 4 S3 >M 3 67 3-67 3S7 367 > 4-8 > 8-64 > 64 Lng1h I KXI'-r-Kii'RK 1 {continued).--Hlx-r distribution bv common lot* of tin' iiuiiiIxt of particles per niuronr.mt in each of 3-1 dimension,il lait-jjoiies- JNC1, VOl. (.7. NO ), NOVKMHKR 1!MI vvv OOOOl^06 Carcinogenicity of Fibrous Minerals 971 60 r! i "1 l . --. -- __' 1 Expt No. Compound 1 Titanate 1 2 Titanate 2 3 Si carbide 4 Dawson 5 5 Tremolite 1 6 Tremolite 2 7 Dawson 1 8 Crocid 1 9 Crocid 2 10 Crocid 3 11 Amosite 12 Crocid 4 13 Glass l 14 Crocid 5 15 Glass 2 16 Glass 3 17 Glass 4 f 18/1 Alumin 1 19 Glass 5 20 Dawson 7 21 Dawson 4 22 Dawson 3 23 Glass 6 24 ` Crocid 6 25 Crocid 7 26 Crocid 8 27 " Aiumin 2 28 ' Alumin 3 29 Crocid 9 30 Wollaston 1 Alumin 4 32. Crocid 10 -33 Alumin 5 34 Glass 20 35 Glass 7 36 Wollaston 3 Table 1.--Summary of 75 experiments with different fibroua materials Actual tumor incidence Percent tumor probability SD Common log fibers/^g. <0.25 *m x >8 itm Expt No. Compound Actual tumor incidence Percent tumor probability SD 21/29 20/29 17/26 26/29 22/28 21/28 20/25 18/27 17/24 15/23 14/25 15/24 9/17 14/29 12/31 20/29 18/29 15/24 16/25 16/30 11/26 9/24 7/22 9/27 11/26 8/25 8/27 9/27 8/27 5/20 4/25 6/29 4/22 4/25 5/28 3/21 95*4.7 100 100 100 100 100 95*4.8 94*6.0 93*6.5 93*6.9 93*7.1 86*9.0 85*13.2 78*10.8 77*16.6 74*8.5 71*9.1 70*10.2 69*9.6 68*9.8 66*12.2 66*13.4 64*17.7 63*13.9 56*11.7 53*12.9 44*11.7 41*10.5 33*9.8 31*12.5 28*12.0 37*13.5 22*9.8 22*10.0 21*8.7 19*10.5 4.94 4.70 5.15 4.94 3.14 2.84 4.66 5.21 4.30 5.01 3.53 5.13 5.16 3.29 4.29 3.59 4.02 3.63 3.00 4.71 4.01 5.73 4.01 4.60 2.65 0 2.95 2.47 4.25 0 2.60 3.09 3.73 0 2.50 0 37 Halloy 1 38 Halloy 2 39 Glass 8 40 Crocid 11 41 Glass 19 42 Glass 9 ,4: Aiumin 6 44 Dawson 6 45 Dawson 2 46 Wollaston 2 47 Crocid 12 48 Attapul 2 49 Glass 10 50 Glass 11 51 Titanate 3 52 Attapul 1 53 Talc 1 54 Glass 12 55 Glass 13 56 Glass 14 57 Glass 15 1 58' Alumin 7 59 Glass 16 60 Talc 3 61 Talc 2 62 - Talc 4 Alumin 8 64 Glass 21 65 Glass 22 66 Glass 17 67 Glass 18 68 Crocid 13 69 Wollaston 4 70 Talc 5 71 Talc 6 72 Talc? 4/25 5/28 3/26 4/29 2'28 2/28 2/28 3/30 2/27 2/25 2/27 S'29 2/27 1/27 1/28 2/29 1/26 1/25 1/27 1/25 1/24 1/25 1/29 * 1/29 1/30 1/29 1/28 2/47 1/45 0/28 0/115 0/29 0/24 0/30 0/30 0/29 20*9.0 23*9.3 19*10.3 19*3.5 15*9.0 14*9.4 13*8.8 13*6.9 12*7.9 12*8.0 10*7.0 11*7.5 8*5.6 8*5.5 8*8.0 8*5.3 7*6.9 7*5.4 6*5.7 6*5.5 6*5.9 5*5.1 5*4.4 4*4.3 4*3.8 5*4.9 3*3.4 6*4.4 2*2.3 0 0 0 0 0 0 0 Common log fibers/Mg. 0.25 urn x >8 ucn 0 0 3.01 0 0 1.84 0.82 0 0 0 3.73 0 0 0 0 0 0 0 0 0 1.30 0 0 0 0 0 0 0 0 0 0 0 0 0 3.30 0 They represent an excellent sue distribution for com parison. Wollastonite (wollaston 1-4).--Wollastonite is a na turally occurring crystalline fiber of monocalcium silicate {15-18). Four separate samples of this substitute for asbestos were received from the same Canadian mine. These were graded commercially according to size by the designation A, B, D, and F. It was apparent at low-power magnification that only grade F was completely fibrous and that these fibers were relatively large. Tremaine (tremolite 1, 2).--The second type of amphibole asbestos studied was tremolite, a material that has a close affinity to the talcs. Both of these samples were from the same lot of asbestos and were in the optimal range of size for carcinogenesis. Compari son of these fibers indicated that they were distinctly smaller in diameter than the tremolite fibers used by Smith et al. {29). Amosite.--The third amphibole asbestos studied was a single sample of South African amosite from the ITCC standard reference samples. No efforts were made VVV 000014407 to alter this as received, and descriptions of this sample as published should apply {19, 21, 22). Attapulgite (attapul 1-2).--Of the natural fibers, the clay attapulgite was of particular interest because of its use in many household items that generate respirable dust. Two different samples of this complex hydrated magnesium silicate were obtained from sources in Attapulgus, Decatur County, Georgia. Both samples were considerably refined, and by electron microscopy they were seen to be composed entirely of short fibers of consistently small diameter {30). These refined clays were considered by the U.S. Bureau of Mines to be 90% or greater in purity, with the remaining 10% being quartz. Halloysite (halloy 1-2).--Hallovsite is a natural fibrous hydrated aluminum silicate, which is respirable and of minute size. The 2 samples were obtained from Dr. Walter Parham, who recovered them from the raw water supply of Hong Kong. On examination these samples were seen to have a tendency for clumping in water. In an effort to disperse the minute fibers, the second sample was sonicated and treated with sodium JNCI. VOL. 67, NO. 5. NOVLMBUR J0-: 972 Stanton, . Layard, Tegeris, et al. hexametaphosphate- Clumping persisted in this second sample, and little different was seen between the 2 samples. made, considerable variation in counts occurred. Clearh. the method is subject to several errors; calibration nl the electron microscope, deviation of particles from the Silicon carbide (si carbide).---One metallic crystal assumed cylindrical shape, and sampling errors, ev line whisker other than alumin was prepared by the pecially where large panicles are concerned, represent General Technologies Corporation. Silicon carbide the major problems. Nevertheless, the estimates are was a single sample, which was of exceptionally fine, probably valid to within one order of magnitude. uniform dimension. Consequently, the counts are reported as the common Potassium octatitanate (titanate 1-3).--In addition to log with the characteristic of the log representing the the synthetic crystals of dawsonite, aluminum oxide, probable limit of accuracy (text-fig. 1). and silicon carbide, 2 samples of fibrous crystalline potassium octatitanate (titanate 1 and 2) were tested. These were obtained from two different suppliers but they represent a single source. Because of the potential carcinogenicity of metallic nickel, the control for these 2 samples was nonfibrous, finely ground nickel titanate (titanate 3). RESULTS Controls have been discussed in previous publica tions (4, 6, 9-11), but they were approached here in a slightly different way. In addition to untreated controls we studied rats in which open thoracotomy was per The 72 experiments represent all of the experiments done in a single dose range and with durable minerals and particles in the respirable range. Additional con formed and a noncarcinogenic material was either applied to the pleura or implanted in the lung. These 3 groups (table 2) were rats from numerous experi trols outside of these limits are mentioned in "Results." ments that were of the same species, sex, and age and Fiber measurements.--An aliquot of each of the 72 experimental mineral samples was placed on a Formvar-covered, slotted grid with an opening measuring 1X2 mm. This grid was air dried and first examined under the light microscope. If the fibers appeared satisfactorily distributed, a photomontage of the entire that were housed in the same quarters. The incidence of clearly apparent pleural neoplasms in untreated, aged outbred Osborne-Mendel female rats was essentially nonexistent. However, a few pleomorphic sarcomas that might be confused with pleural tumors occurred in the left thorax of both treated and, to a lesser degree, grid was made at a final magnification of X3.000. The untreated controls. Although these tumors involved the slotted grid was then placed in a Siemens electron thickness of the chest wall, in most cases the tumors I microscope, Elmiskop 1-A, and the entire grid was appeared to be derived either from mammary gland scanned at low magnification. From this scan, an area that seemed to represent a typical distribution of particles in the specimen was selected for counting. At a final magnification of about X5.000-100,000, a second photomontage was made of that section of the grid selected to include particles typical of the sample. This fibroadenoma or from suture granuloma in the subcu taneous tissues. But there remained a few tumors for which no definite origin could be determined and which were histologically comparable with pleural sarcomas. In both the experimental groups and the control groups these questionable tumors were counted selected area, which generally measured about 350X150 as pleural sarcomas. These essentially confusing tu ^m, was then located on the* lower magnification montage of the ghd and examined to determine whether mors observed in the controls need to be taken into account in the assessment of the carcinogenicity of the the area chosen was truly representative of the entire experimental materials. The incidence of pleural sar- grid. Finally, all fibers in the area were counted and measured individually. For the diameters, a compara tive scale at the final magnification was used to measure magnified diameters that measured less than I mnu In most cases, the selected area counted included at least 1,000 fibers, but the actual number varied with the overall size of the particles. Subsequently, with the aid of the IBM system 370 computer, assuming the fibers to be of cylindrical shape and using the density of the material, we were able to estimate the weight of the counted samples and the number of particles of a given dimension in the 40* mg dose administered. For the purpose of calculation, particles were grouped into 34 dimensional ranges as indicated in text-figure l, and the number of particles per microgram in each category was calculated. Dupli cate counts on the montages were done on most samples and were surprisingly similar, as were counts on TabLE 2.--Incidence of pleural sarcomas in outbred female Osborne-Mendel control rats Time, wk 12-52 53-65 66-78 79-91 92-104 105-120 121-130 131-143 144-156 156 Total Percent Untreated" 1/113 0/15 0/26 Q/68 0/26 0/98 1/66 0/27 0/27 1/22 3/488 0.6 Noncar* cinogensc pulmonary implants'1 0/49 2/26 4/50 1/70 1/72 1/162 0/3 9/432 2.1 Noncar cinogenic pleural implants0 0/47 1/72 3/64 2/35 10/294 1/36 17/598 2.8 Combined controls'1 1/209 3/113 7/140 3/223 11/392 2/296 1/69 0/27 0/27 1/22 29/1.513 1.9 different areas of the same montage. However, when studies of repeat samples from the original fibers were " No. dead with pleural sarcomas/No. dead without pleural sarcomas. JNCI. VOl- K7. NO V. SOVEMHKR J`W1 VVV 000014406 Carcinogenicity of Fibrous Minerals 973 comas in all 3 control groups combined, calculated by the life table method (13), was 7.74.2%. Comparison of this incidence with the pleural sarcoma incidence in *he 72 individual experiments showed that the inci1 \ce of pleural sarcomas in a particular experimental O.oup was significantly greater than that in the com bined control group only if it exceeded 30% (see expts 1-29 in table 1). In regard to the controls, some negative experiments with intrapleural implants not used as controls should be mentioned. These experiments included intrapleural implants that did not conform to the type of materials under consideration because the particles were either nondurable (cotton lint, gypsum, and carrageenan), were of greater than respirable size (steel shavings, steel wool, vermiculite, polyurethane, tungsten carbide, and infusorial earth), or were exclusively nonfibrous (poly- acrylic nitrile, antigorite, silicon dusts, and several glasses). None of these experiments had an incidence of pleural sarcoma that was significantly greater than the 7.7% incidence of the combined control group. From_the summarization of the 72 experiments in i i table 1 and text-figure 1, even cursory examination of the fiber distribution suggested that particles in the relatively thin- and long-dimensional categories were associated with higher tumor probabilities. This obser vation was confirmed by the statistical correlation and regression techniques that were used in previous papers (4, 9, 10). The logit transformation (13) was applied to the estimated tumor probabilities (p) according to the formula: Logit=ln (p/(l-p)], where In denotes the natural logarithm. The 34 dimensional categories indi cted in text-figure 1 were arbitrarily grouped into 11 I r categories, and the simple correlation coefficients oi the logit of tumor probability with the common logarithms of numbers of particles per microgram in each of these categories was calculated (see table 3). The maximum correlation coefficient, 0.80, was with panicles equal to or less than 0.25 in diameter and greater than 8 *tm in length. There was no correlation with particles equal to or less than 4 ^m in length or with particles greater than 1.5 pm in diameter, but relatively good correlations were noted with log num bers of fibers in categories greater than 4 pm in length and up to 1.5 pm in diameter, with correlation coefficients of 0.45-0.80. The possibility of the existence of relationships between the particle size distributions and tumor prob- TaSle 3.--Correlation coefficients of logit of tumor probability ntA common logarithm of number of particles per microgram in different dimensional ranges Fiber diameter >4 >1.5-4 >0.25-1.5 <0.25 <4 _ -0.45 0.01 0.20 Fiber length, pm >4-8 -0.28 -0.24 0.45 0.63 >8 -0.30 0.13 0.68 0.80 abilities, which are not disclosed by the simple correla tion coefficients in table 3, was explored by multiple regression methods. These methods were used to find the best-fitting function of the form: logit --a +bi X\ +...........+ .bk x*. where xj,........... .. x* represent the common logs of numbers of the particles per microgram in the size categories of table 3, and a, b............... 6* are the regression coefficients to be estimated. The analysis indicated that the addition of further dimen sional categories to the category with diameter equal to or less than 0.25 pm and with length greater than 8 pm did not significantly improve the explanation of the variation in tumor probability. The regression equation for the single variable (x) representing the common log of number of particles per microgram with diameters equal to or less than 0.25 pm and lengths greater than 8 pm was: \n[p/(\--p)\m--2.62 + 0.9305*. . (0.24) (0.0834) The numbers in parentheses beneath the regression coefficients are their estimated standard deviations. The relationship expressed by the .above equation is highly significant (P<0.0001). The estimated regression curve is illustrated in text-figure 2. The fact that the use of additional dimensional categories did not significantly improve the fit of the regression equation does not indicate lack of carcino genicity in other categories. The regression of logit of tumor probability on common log of numbers of particles in other categories with a diameter up to 1.5 pm and a length greater than 4 pm would also indicate a highly significant relationship. The difficulty here is that the numbers of particles in adjacent size categories were highly correlated. Better definition of the critical \ \ ] range of carcinogenicity would require more narrowly defined samples (i.e., particles in a narrower dimen sional range). What is perhaps more likely than the existence of a narrow range of sizes within which particles are carcinogenic and outside of which they are not is that the probability of tumor falls as particle diameter increases and length decreases. Of the 72 experiments, 7 had tumor incidences that deviated markedly from those predicted by the esti mated regression line. These were: experiments 5 (trem- olite 1), 6 (tremolite 2), 26 (crocid 8), 29 (crocid 9), 33 ^(alumin 5)^47 (crocid 12), and 71 (talc 6) (see table 1 "and text-fig. 2). For the first 3 of these experiments the observed responses were higher than the predicted responses, but the high responses can in pan be explained by the - fact that there were substantial numbers of fibers in size categories adjacent to the category used in the regression equation. For the remaining 4 experiments, the observed response was substantially lower than the expected response; al though no apparent explanation existed for these deviations, they were possibly -due to inaccuracies in the assessment of functional particle size. In prepara tions of amphibole asbestoses (which included the crocidolites and tremolites), we observed that both JNCI. VOL. 67. NO. 5. NOVEMBER 1981 VVV 000014409 974 Stanton, Layard, Tegeris, et al. 1.0 3 2SS 0 se.vwnite 0.9 L * 3 OK S J* s'lssn carbie 0.6 r A - ar.abuigiw rr~ o 0.7 * M :*8--alite P- W * vveastonit* u. C 0.6 " H - ra. cysite 0 * 3-?site > 0.5 C2 <e.i 0.4 Gcz 0.3 j-w / -^L 9 3S G \.S 'rfl / s OS S 0.2 'CW A55*V 0.1 rAG3=T UT5CGG LTT5G C-.vr-sC , 0.0 0.0 0.5 1.0 1.5 2.0- 2-5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 LOG NUMBER PARTICLES MEASURING < 0.25 x > 8 nm PER MfCROGRAM Text-kigcre 2.--Regression curve reiaung probability of tumor to logarithm of number of particles per pg with diameter ^0.25 am length >8 ym. I =5 clumping and fragmentation of the particles were greater than those in the other minerals, and estimates o particle sue distribution in duplicate samples varied most for amphibole asbestoses. DISCUSSION The results show that a wide variety of compounds that seem to have only dimension and durability in j common are carcinogenic for the pleura of the rat. Our | conclusions regarding those dimensional categories 1 that correlate strongly with* probability of pleural tumor remain essentially the same as in previous studies, namely, that probability of pleural sarcoma correlates best with fibers that measure <0.25 /imX>8 nm, but that relatively high corrections were also observed with fibers in other categories having a diameter up to 1.5 and a length greater than 4 /im. A more refined estimate of critical carcinogenic dimen sion may be possible if the parameters of the experi ments were changed. A different animals species, lower dose, more precise means of fiber measurement, more accurate volumetric calculations, and samples with narrower dimensional ranges all might be determining factors in better assessment of the particle dimensions critical to carcinogenicity. However, we should keep in mind two points: a) the dimensional limits are prob? ably far from absolute, and b)fwe are dealing with cancer in the rat ancl-thus extrapolation to man may not be precise. It is clear from the histologic studies of these experiments and of previous studies that our data offer an explanation more for the lack of carcinogenicity of short fibers and thick fibers than for the carcino genicity of long, thin fibers. Sections of preneoplaw* pleural lesions show avid phagocytosis of both \ht fibers and large-diameter fibers but negligible phjitr> : cytosis of long, thin fibers. Consequently, in thr* experiments we may simply be measuring the ef ficiency of phagocytosis. Doubtless, we have liule irj* knowledge of the way that long, thin fibers can *auv cancer, but as Rous {31) once said, "Since what think largely determines what we do, it is well that **f think something." In the spirit of this quote, it migh* be profitable to consider potential mechanisms cancer production by long, thin, fibers. Of first impo* lance are those hypotheses in which the progenitor the cancer cell is not directly affected by the fiber. The long latent period would suggest that a generalized alteration either in local milieu or systemic environ ment might be at fault. In this regard, the abundant collagen in the preneoplastic pleural scars should be noted. Consideration of a relationship between this phenomenon and "solid-state" carcinogenesis is attrac tive, though the reduction of plastic sheets to small particles tends to reduce carcinogenesis. Mechanisms of solid-state carcinogenesis have been thoroughly ' viewed by Brand (32), and liule more need be added. Any hypothesis concerning fibers must take m|J account the fact that both short fibers and thick fiber' are less carcinogenic than fine, long fibers. Since do* was fixed in weight, but was different in dimension h* > all experiments, one might consider the surface area a possible factor. If this were the case then fibers from the same pool that were modified only by shortcut!** should be equal in tumor-producing capacity. Clcatl'this is not true in the following experiments: 13 (gl**" 1, MOL) vs. experiment 49 [glass 10. MOS see (4. /"M- JNCI. VlH.. li*. NO. 3. NOVEMBER 19*1 vvv OOOOl**10 Carcinogenicity of Fibrous Minerals 975 and in experiment 24 (crocid 6) and experiment 25 (crocid 7) vs. experiment 40 (crocid U). experiment 47 (crocid 12), and experiment .68 (crocid 13). However, in these examples the phagocytosis variable cannot be ruled out. A provocative explanation relates to the ability of fine, long fibers to penetrate cells without killing them. That this can occur is evident from in vitro studies (33). However, simple penetration of cells by rrnxelia of fine dimension (a notable aspect of con tamination of cell cultures by fungi) rarely produces transformation of cell cultures and thus is unlikely to produce cancer. However, mineral fibers differ from fungi in their rigidity as well as chemical content, and one easily could conceive of physical differences be tween the mineral fibers and mycelia that might be critical. references ti) Pott F, Hutk F. Friedrichs KH. TumoTigenic effect of fibrous dusts in experimental animals. Environ Health Perspect 1974; 9:313-315. (2) Pott F, Friedrichs KH, Huth F. Results of animal experi ments concerning the carcinogenic effects of fibrous dusts and their implication with regard to carcinogenesis in humans. Zemralbl Bakteriol [BJ 1976; 162:467-505. (J) Pott F, Friedrichs KH. Tumoren der Ratten nach i. p. Injec tion faser forminger Staube. Naturwissenschaften 1972; 59:318. (4) Stanton MF. Layard M. Tegerjs A, Miller E. May M. Kent E. Carcinogenicity of fibrous glass: Pleural response in the rat in relation to fiber dimension. J Natl Cancer Inst (977; 58: 587-603. Smith WE, Miller L, Elasser RE. Tests for carcinogenicity of asbestos. Ann NY Acad 5ci 1965; 132:456-468. IS) Stanton MF, Wrench C. Mechanisms of mesothelioma induc tion with asbestos and fibrous glass. J Natl Cancer Inst 1972; 48:797-821. (7) Wagner JC. Asbestos carcinogenesis. Am Chem Soc Monogr 1976; 173:729-736. Uh WaGNER JC, berry G. Timbrell V. Mesotheliomata in rats after inoculation with asbestos and other materials. Br J Cancer 1973: 28:173-185. 1) Stanton MF. Layard MW. Carcinogenicity of natural and man-made fibers. In: Margison GP, ed. Carcinogenesis. Ad vances in medical oncology, research and education. Vol. 1. Oxford and New York: Pergamon Press. 1979:181-187. I/O)---------- . The carcinogenicity of fibrous minerals. In: Proceedings of the workshop on asbestos: Definitions and measurement method; held at the National Bureau of Standards. Gaithers burg, Md., July 18-20, 1977. Washington. D.C.: National Bureau of Standards. Nov 1978:143-151 (NBS special publica tion No. 506). i/fl Stanton MF. Some etiological considerations of fibre carcino genesis. In: Bogovski P. Timbrell V. Gilson J. et al.. eds. Biological effects of asbestos. Lyon, France: WHO. 1973: 289-294 (IARC publication No. 8). 'IN Ou rg J, Rosen SH, Moolten S. Histological characteristics of mesothelioma associated with asbestos. Ann NY Acad Sci 1965; 132:614-622. 'Hi ArmitaCE P. Statistical methods in medical research. New York: Wiley. 1971:376-377. 410-414. 'tti Pilgrim HI, Dowd JE. Correcting for extraneous death in the evolution of morbidity or mortality from tumors. Cancer Res 1963; 23:45-48. Olson RH. Introduction. In: Lefond SS. ed. Industrial minerals and rocks. 4th ed. New York: American Institute of Mining. Metallurgical and Petroleum Engineers. Inc., 1975. ,%l Breck DW. Synthetic zeolites: Properties and application. In: Lefond SS. ed. Industrial minerals and rocks. 4th ed. New York: American Institute of Mining. Metallurgical and Pe troleum Engineers. Inc.. 1975. (77) Sheppard RA. Sedimentary rocks. In: Lefond SS, ed. Industrial minerals and rocks. 4th ed. New Yotk: American Institute of Mining. Metallurgical and Petroleum Engineers. Inc.. 1975. (18) Mlmpton FA. Commercial utilization of natural zeolites. In: Lefond SS. ed. Industrial minerals and rocks. 4th ed. New' York: American Institute of Mining. Metallurgical and Pe troleum Engineers. Inc.. 1975. (.19) RenDall RE. The data sheets on the chemical and phvsical properties of the LTCC standard reference samples. In; Sha piro HA, ed. Pneumoconiosis: Proceedings of the interna tional conference, Johannesburg. Capetown, Union of South Africa: Oxford Univ Press, 1970:23-27. (20) Ross M. The asbestos minerals: Definitions, description, modes of formation, physical and chemical properties, and health risks to the mining community. In: Proceedings of the work shop on asbestos: Definitions and measurement methods; held at the National Bureau of Standards, Gaithersburg, Md.. July 18-20, 1977. Washington. D.C: National Bureau of Standards. Nov 1978:49-63 (NBS special publication No. 506). (27) Timbrell V. Characteristics of the International Union Against Cancer standard reference samples of asbestos. In: Shapiro HA. ed. Pneumoconiosis: Proceedings of the international conference. Johannesburg. Capetown, Union of South Africa: Oxford Univ Press. 1970:28-36. (22) Timbrell V, Gilson JC, Webster I. UICC standard reference samples of asbestos. Int J "Cancer 1968; 3:406-408. (23) Timbrell V. Physical factors as etiological mechanisms. In: Bogovoski P, Timbrell V. Gilson JC, et al., eds. Biological effects of asbestos. Lyon. France: WHO. 1973:295-303 (IARC publication No. 8). (24) Spiil S. LeinewebeR J. Personal experiences with making sam ples of fibers for biological experiments. In: Peinar PV. ed. Fibres for biological experiments. Montreal: Institute of Oc cupational and Environmental Health, 1974:45-50. (25) Wagner JC. The pathogenesis of tumors following the intra pleural injection of asbestos and silica. In: Nettesheim P. Hanna MG, Deathergate JW, eds. Morphology of experimental respiratory carcinogenesis. Oak Ridge, Tennessee: Oak Ridge National Laboratory. 1970:347-358 (Atomic Energy Commis sion symposium series No. 21). (26) Harington JS. .Allison AC. Badami DV. Mineral fibers: Chem ical. phvsiochemical and biological properties. Adv Pharmacol Chemother 1975: 12:291-402. (27) Jackson J. Huggins CW. Amplan SG. Synthesis and characteri zation of dawsonite. Washington. D.C: U.S. Dept of Interior. 1972 (Bureau of Mines report of investigation, 7664). (28) Huggins CW. Green TE. Thermal decomposition of dawsonite. Am Mmeralog 1973; 58:348-550. (29) Smith WE. Experimental studies on biological effects of tremoliie talc on hamsters. In: Proceedings of the symposium on talc, Washington. D.C. May 8, 1973. Washington, D.C.: U.S. Bureau of Mines 1974:43-48 (Bureau of Mines information circular 8639). (30) Huggins CW. Denny MV. Shell HR. Properties of polygorskite. an asbestiform mineral. Washington, D.C.: U.S. Dept of Interior. 1962 (Bureau of Mines information circular R16071). (3I\ ROUS P. The virus tumors and the tumor problem. In: The Hawey lectures series. No. 31. Baltimore: Williams Sc Wilkins, 1935-1936:74-115. (32) Brand KG. "Solid-state" or "foreign-body" carcinogenesis. In: Symington T, Carter RL. eds. Scientific foundations of oncology-. London: William Heinemann Medical Books. 1976: 490-495. (33) Wade MJ. Upkis LE. Stanton MF, Franks AL. In vitro cyto toxicity assay as applied to asbestos and other minerals: Its possible relevance to carcinogenicity. In: International work shop on the in vitro effects of mineral dusts. Medical Research Council Pneumoconiosis Unit. Penarth. Wales. Sept 4-7. 1979. Penarth. Wales: Medical Research Council, 1979 (P388DI i. JNCI. VOL 67. NO- b. NOVEMBER IHH i VVV C0001**** fiistifdfion Peril? JTT l IcalthStudicsSuggest Exposure to asbestos has been linked to fatal lung cancers and other disabling res Some companies are shoring up legal defenses. Besides requiring salesmen to Mr. Fryman acknowledges, however, that tbe cristoballlc levels detected in the Asbestos Substitutes piratory diseases. Some 44.0UQ death and dsclose Mr. Enterline's findings on glass injury claims have been tiled against Man liter. Manville recently ordered the de Browning-Ferns study exceeded levels predicted by manufacturers. As a result, ville and other asbestos producers by plant struction ol all outdated insulation safety he says. Standard Oil recently urged Hut Also Pose Cancer Risk workers. Insulators and others. Most uses pamphlets, says Mr. Sells, the company of of asbestos are currently banned by fed- ficial. In asbestos-related cases, large pu- improved face masks be used by workers removing ceramic insuiatioa. ;ral regulations. nibre damage awards had been meted out Concerns over insulation safety are also While DataAren`t Definitive, Cllus-ftber and some other insuiatioa imduets have long been known to cause against Manviiie for failing to publicly dis close its up-to-date knowledge ol asbestos fueling a war between makers of compet ing products. Take tbe case of Richard Some Kine-Fiber Products ;km rashes and respiratory irritation. But inch materials and asbestos, several scien- hazards. Industry workplace practices are also Munson, president of National Consumer Products Marketing inr., an Auburn. Ca IlaveScientistsConcerued jsts say. also share a potentially fatal chaneine. For Instance, Chicago-based rail: All contain masses of small, inbal- lif.-based company that sells cellulnsibased Insulation, a competitor to gloss fi ible fibers that can lodge Ui human lungs. ! USC1 Corp- the nation's largcsi producer of ber in the home-insulation market. CompaniesQuestion Findings Jnder one current scientific theory, the 1 mineral-wool products, recently banned irancor-producing potential of any fiber is a cigarette smoking by muiilatwin-plnm Mr. Munson founded and sponsors a group called Victims of Fiberglass, which luncliun of Us size rather than its chemis j workers. The move, says a list! Spokes- recently placed advertisements in several By try. The mast suspect: fibers that are Uuo, I man. reik-cls scientific concern that srnak- UJL newspapers carrying bold headlines filMff Krperwr Wij..Stkk>t long and durable. ! ini; might intensify any lung-cancer threat Two itoraHw ana mamnd.! ai daTTis Scientific concern about such "fine" In I posed by mineral wool. Plant workers have that read: Consumer Warning -- Manmade 1 Asbestos (Fiberglass) Insulation tn Your after asbestos insulation was found to be a killer, recent health studies suggest that certain other insulation products, some sulation fibers is growing. Hons Weill, a I also been issued uniforms. Hie spokesman - Altir Will Cause Lung Cancer. pulmonary specialist at Tulane University, ! says, to prevent workers from carrying in Mr. Munson is also publicizing the work says chest X-rays of Insulation-plant sulation fibers home on their clothing. of a consultant he hired to measure fiber times used as substitutes forasbestos, may workers exposed to fine fibers exhibit Lung-cancer deaths in asbestos workers' ; sizes contained in a blown-insuiation prod- also present a lung-cancer threat A blue-ribbon scientific panel assem mure minor abnormalities than do X-rays ol workers exposed to larger fibers. Pine- families have been Linked to such expo 1 uct sold by the Valley Forge. Pa.-based sure. , CertainTeed. The -vmiHam found that bled by tbe World Health Organisation Is set to consider next month whether those materials-glass fibers, mineral wool and ceramic-based libera--pose a cancer threat to tens of thousands ot workers world wide. A decision that any or all of them do would force the affected manulac- turns to label their products as possible cancer risks. It would also doubtless trig ger regulatory action in tbe U.S. and else where and spark lawsuits against a S3 bil lion domestic Industry that bow gets little oversight. . Just bow such fibers might caise can cer--and at what exposure levels--is un clear. But "U the question is. Do other iia- filiation! fibers besides asbestos cause can cer?* the answer is yes. " says Philip En terline. a University of Pittsburgh epidemi ologist hired by U.S. insulation producers several yean ago to study their products' hazards. fiber workers are also showing a pattern of decreased lung capacity, ays Dr. WeilL who cautions that Us work Is prelimi nary. Use Is Growing Such warnings come as Insulation pro ducers increase fine-fiber output largely for products blown under pnssure Into walls and ceilings of homes and commer cial biddings. Use of Cine fibers is growing because they provide greater insulation than larger-sized fibers, and at lower cost, soys William Sells, president of ManviUe'S gloss-fiber division. The safety debate Is putting glass-fiber producers on the defensive. At a tense twohour meeting last month with Mr. Enter line, health officials from Manville. OwensCorntng Plbergtas Corp. and CertainTeed Corp- challenged the scientist's methods and facts. They said they hope to prove that workers' deaths from lung cancer Quandary for Union most of the products' fibers were small The lack of definitive health data an in* enough to be inhaled. eninium n-i?.irri* is creating a quandary for CertainTeed officials say that insulation workers and product users. "At this point workers are only briefly exposed to such fi we haven't taken any position on glass fi bers and thus don't face a health danger. ber because we (eel the jury Is still out." Glass-fiber company officials also charge says James Golden, a health official with Mr. Munson until running a score cam the Sheet Metal Workers International As paign. "it's obvious he.hos a conflict of in sociation. On the other hand. Mr. Golden terest." says Mr. Sells, the Maanlle offi says, union officials are concerned by the cial. lack of data. The reason: Most scientists believe that cancers develop 20 years or more alter the first exposure 10 a carcino "There's no question that J have a con flict of interest." Mr. Munson says, "but so does Manville." As for any lung-cancer gen. "Some of these libera haven't been threat from cellulose-based insulation, epi- around long enough to study their effects 1 demiotogtst Enterlme says be isn't aware on humans," Mr. Golden says. ! of any evidence suggesting one. Ceramic fibers are a case in point, i Several independeni scientists say that Sales of products using tbe durable fibers I 1 current health data don't indicate that the have boomedsince tbe mid-1970s, reaching { I insulation materials under study pose a as wammed J103 million last year, ac- 1 lung-cancer threat to homeowners. "The cording to industry officials. They are used mainly tn finings for industrial furnaces unstudied population I'm most concerned , about are people putting la insulation." MineraJ-Wool Workers aren't related to liber exposure, that they His studies show. Mr. Enterlme told an : may Instead have resulted froratthcr toxic international scientific conference last Oc compounds in the workplace or tbe envi and as asbestos substitutes. 1 says Ernest E. McConnell, a toxicologist Health research, however, has lagged 1 wtth the federal National Toxicology Pro behind sales. Thus far. the two major ani gram. tober. that mineral-wool factory workers ronment. mal studies of the materials' cancer-caus Regulatory Action Likely have an increased lung-cancer risk. He "Thai is what we are trying to accom ing hazards have yielded mixed results. Whatever the cose. U S. regulatory ac also found a recent spate of lung-cancer deaths among glass-fiber workers exposed to flne-diametcr fibers, but he said such data were too preliminary to draw any sci entific conclusions. Mr. Entcrline says he is also concerned about ceramic fibers, for which health data are sparse. Makers of the materials under scrutiny all defend their products' safety. And re searchers such as Mr. Eaterllne believe that most currently used insulating prod ucts don't appear to threaten homeowners and are less toxic than asbestos. Became uf improved plant controls, some indepen dent scientists say. any danger to current plant workers may also be tow. _ Nonetheless, the specter of a health crisis even vaguely resembling the as bestos disaster is sparking a furor in the insulation industry. Company scientists are scrambling for evidence to defuse Mr. En terlines findings: some insulation pro ducers are building defenses against future damage claims: and the safety issue is at (he heart of a battle between producers of competing types of insulation. plish. Whether that turns out or not we Moreover, given the latency period for have to see." saya Robert Doban. a senior 1 tung cancer, industry scientists acknowl vice president at Owns-Cornlng. (he na tion's largest glass-Fiber producer. The company's response to Mr. Enter line's findings ts exemplified by measure it has taken at its Newark. Ohio. glass-ft ber plant. Mr. Enterline had found lhai workers and former workers at the piantand at a few other plants among ll he studied In the glass-fiber industry--had a higher-than-expected lung-cancer rate 2C yeare or more after their first exposure. In edge that u may be too eariy to detect de veloping cancers in workers through tradi tional X-ray screening tests. "The latency issue is one that has to be addressed." says Edward Horvath, medi cal director of occupational health for Standard OU Co., the nations largest maker of cemmc-based-flber products. Dr. Horvath estimates 11 may take several years far such cancers, if any. to develop in exposed workers. recent months. Mr. Doban says, a team ol Banned by Navy company engineers has scoured Newarkplant records going back four decades in an effort lodetermiae what cancer-causing agents might have been used at the facil ity. Reared plant personnel were ques Some major users of such products aren't waiting. The U.S. Navy, for exam ple. recently banned shipboard use of cera mic fibers, according to a Navy spokes man. Tbe move, says the spokesman, tioned for the same purpose. Mr. Doban says. Armed with their own findings regard- came afterntanuficturera voluntarily rela beled products 10 note that in an animal study, some ceramic fibers had produced tog the Newark plant, Owens-Cormng offi cials challenged Mr. Enterllne's findings when they met with the scientist lost month. Mr. Doban said company records cancerous tumors. Along with a possible cancer threat, ccmnuc-based fitu-rs pose other health problems. In a published 198b study by tion on glass and other insulating fibers is likely. Richard Lemen, a top official at the federal National Institute for Occupational Safety and Health, says the agency, based on Mr. Enterlme's report, is renewing whether to recommend stricter workplace exposure limits. Sir Richard Doll, a lead- ' ing British epidemiologist, says he believes 1 exposures to all insulating fibers should be i held to the same level as exposure to as- 1 bestos. Gloss and other insulating fibers 1 . are currently regulated in the U.S. as socalled "nuisance." or nonhazardous. I dusts. j Both industry and academic institutions I are planning new studies of insulation-fiber ' safety. Scientists who witnessed the as bestos disaster are intent on averting any similar tragedy. Says Irving Selikoff. a leading asbestos expert at New York's Mount Sinai School of Medicine: "All the resources we built to study asbestos should be drawn on now." Manville `Concerned* showed sporadic use of asbestos at the Browning-Ferns Industries, for Instance, "There's no question we're concerned" Newark plant. Jon Konzen. the company's the Houston-based waste-management can* says William Reitze, a health official for medical director, said the Newark plant cem wanted that ceramic-based libers Denver-based Manville Corp. Bankrupted by asbestos-related death claims and puni tive damage awards. Manville recently or dered Us salesmen o! glass-fiber products to certify in writing that they have di had frequently been used to test gloss-fiber manufacturing techniques turoivtDg any Bomber ol potentially toxic compounds. "There were things that went on at Newark that never saw the tight of day used to Une industrial furnaces quickly break down into cristoballte, a form of sil ica linked to silicosis, a fatal lung disease common among quarry and tunnel waiters. Charles E. Fryman. Standard vulged Mr. Entertaes findings to cus anywhere else." Dr. Konzen said. Oil's manager of industrial hygiene, says tomers. They're in a Jam* ceramic-fiber producers are aware of the The ruddy-faced Mr. Enterline isn't de terred. After the meeting, be noted that mineral-wool producers were raising sim ilar objections to his work, but he said they hadn't supported their abieeUons scientifi cristoballte hazard and have long urged workers to ise masks when removing ceramic insulation from furnaces. They have also called attention 10 the ertstoba- Ute nsk on their labels. cally. Of tbe insulation industry's current plight, he remarks: "They're in a jam and they just don't know what to do." I VVV 000014412