Document permYOOE6D5JN34E2k7mwZvVX

Relation of Particle Dimension to Carcinogenicity in Amphibole Asbestoses and Other Fibrous Minerals '12 3 * I; n it Meart F. Stanton, 3,4 Maxwell Layard, 5-6 Andrew Tegeris,7 Eliza Miller, 3-8 Margaret May, 3,4 Elizabeth Morgan, 7`9 and Alroy Smith 5 * ABSTRACT--In 72 experiments, durable minerals In the form ot particles on respirable size and of wide chemical and structural varieties, were implanted in the pleurae of outbred female OsborneMendel rats for periods of more than 1 year. The incidence of induced malignant mesenchymal neoplasms correlated well with the dimensional 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 pm in length, but relatively high correlations were also noted with fibers in other size categories having diameters up to 1.5 pm and engths greater than 4 pm. Morphologic observations indicated -,t short fibers and large-diameter fibers were inactivated by agocytosis and that negligible phagocytosis of long, thin libers yccurred. The wide variety of compounds used in these experi ments suggested that the carcinogenicity of fibers depended on dimension and durability rather than on physicochemical prop erties.--JNCI 1981; 67:965-975. Work in several laboratories has indicated that di verse varieties of minerals, are carcinogenic when ap plied directly to the pleura of the rat or hamster in the form of microscopic fibers, i.e., particles with dimen sional aspect ratios of 3:1 or greater (1-9). The same minerals are much less carcinogenic when applied at .ual weight and size in nonfibrous form. Further, preliminary experiments indicate that carcinogenicity orrelatcs best with increasing numbers of fibers having both diameters of 0.25 pm or less and lengths of more than 8 pm and that the correlation diminishes with fibers of greater diameter or lesser length. Conset.'uemly, a reasonable conclusion is that the long, thin, brous structure is critical to the carcinogenicity of :se minerals. Studies on fibrous samples within very irrow dimensional ranges would be valuable in the establishment of this hypothesis, but these ideal sam ples are not available. Consequently, we are faced with ii:e correlation of carcinogenicity with fiber samples of widely mixed dimension. The purpose of this report is correlate our best estimate of fibrous dimension with < nrinogcnicity for all those minerals that we have '.:c!ied that are both durable and within the size range respirable particles. This involves 72 experiments with minerals of wide chemical and structural variety. Oi special interest arc the data on the amphibole asbestoses: amosiie, tremolite, and crocidolite, though estimates of the dimensions of the asbestoses arc 's;H-ci:tllv liable to error. Chrysotife, although as caronogenic as the amphiboles at comparable dimeu- could not be included since it has proved difficult to be measured with any degree of precision. MATERIALS AND METHODS 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 particles 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 could cause confusion: the fibrosarcomas of left mam mary glands and the subcutaneous fibrosarcomas in duced by suture material. Vigilance and early surgical removal accounted for most mammary tumors; the use Abbreviations used: alumin = aluminum oxide; attapul - attapulgitc(s); crocid = crocidolite(s); dawson=dawsonile(s); halloy = halloysite(s): UICC= International Union Against Cancer Wollaston = wollastoniie(s). 1 Received November 13. 1980; revised May 6. 1981; accepted June 8. 1981. 3 The guidelines fot the Care and use ot laboratory animals were followed as set forth by the Committee on Revision of the Guide for Laboratory Animal Facilities; by the Guide for the Care and Use ol Laboratory Animal Resources, the National Research Council; and by the National Institutes of Health. 3 Laboratory of Pathology, Division of Cancer Biology and Diag nosis, National Cancer Institute (NCI). National Institutes of Health, Public Health Service. U.S. Department . ol Health and Human Services. Retliestla, Md. 20203. * Deceased. 3 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 Avc., Palo Alio, Calif. 91301. 3 Pharmacopathics Research I-aboiatories. Inc., 9703 North Wash ington Blvd.. Laurel. Md. 20810. * Present address: 5321 Trent St- Chevy Chase. Md. 20015. * Present address: Triangle Resojrce Industries. P.O. Box 599. Laurel. Md. 20707. 965 ~y. WtfiicWD Cam/ U<o4vf*cfe 14$/) 6T7'. JNCI. VOL 67. NO. S. NOVEMBER I9BI 511964 0049  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 .pleuraf sarcoma and provides a good means of maktng 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. None 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, subtractions were obtained by simple gravimetric methods in aqueous media to separate fibers of different dimensions. These maneuvers ineluded 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 availablefor study, and samples studied are listed in text-figure 1 and table 1. Six major groups of particles had multiple dimensional ranges; these in elude: 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, 13-18), but the following information is pertinent. Crocidolile (crocid 1-13).--These 13 samples of South African crocidolile (an amphibole asbestos) were from four different sources. Samples crocid 1, 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 crocidolile. 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 U1CC crocidolile designated crocid 5. Differences in dimension were the result of different milling times. Crocid 5, the original U1CC 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 South African crocidolite and separated by centrifuga tion to obtain mutually exclusive size ranges from the same sample (24). The remaining sample, crocid 2, was obtained from Dr. J. C. Wagner (Medical Research Council Pneumoconiosis Unit, Penarth, Wales) as representative of the material used by him in his 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 by 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 previously 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 18 was used as a control and also served as a vehicle for the implants. Glasses 19 and 20 were preparations of large-diametered fibrous glass that was leached to remove all elements except SiOj. These two glasses were exceptionally fragile and contained many irreg 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 a microcrystalline zirconium oxide content greater than 90% (glass 22). Aluminum oxide (alumin 1-8).--The 8 samples of aluminum oxide were all crystalline sapphire whiskers prepared by General Technologies Corporation, Reston, Va,, or by Thermokinetics Fiber Incorporated, Nutley, 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 (laic 1-7).--All seven talcs were refined raw materials for commercial products. Each was from a separate and diverse source and selected to include all extreme ranges of dimension. Platclike structure was consistent and was considered in the calculation of the volume (15-18). Dawsonite (dawson 1-7).--The 7 daivsonite sam ples (crystalline dehydroxy sodium aluminum car bonate [NaAl(OH)2COj] 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 dau-son 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 I, 5, 6, and 7) were synthetic crystals from a second commercial company. These 4 samples were especially crystallized and sorted to achieve narrow ranges of size. JNCI. VOI.. 67. NO. S. NOVtMBKR I*JH1 511964 0050 Carcinogenicity of Fibrous Minerals yt>/ (t 1 Th^et* 1 S5% >4M.O >2H.0 > 1.51L >50-15 > 15.50 as7 >.1015 3-87 > .05-.10 3Z7 4.41 b.08 6.02 >.01-.05 > 8.0 3.75 4:79 634 5.02 4.11 407 6.17 4.44 CO IfCwvart 2 100% 3-70 363 425 3.40 4.1 B 4.99 431 3.70 4.52 6.07 4.78 3.40 4.18 4.48 431 cn & 0Kt>d* 192 313 0.81 3.76 21S 3JDB 10 40 o 4.48 132 182 204 303 430 4.19 4.41 4.54 1.35 1.65 1.95 1.35 314 135 4.40 4.56 4.74 4.70 HI Dff**w 5 ,0* >4.oao >15-4.0 >15-15 >50-15 >15-50 314 >.10-15 >55-.10 >.01-.05 118 413 183 >ao 157 1-87 415 3.73 102 3.50 453 4.93 (5) Irrmo/ft 1 IUJ> 3.84 4.10 420 3.69 3.14 4.47 435 . 4.14 354 114 3.92 3.14 344 3.14 114 3S2 314 3.44 114 16) TrwnoT(2 144 4.09 114 334 4J6 386 4JM 2X 154 184 3.68 3.38 318 3.01 ZS4 3.77 3.31 2.84 (7) D*w*oo 1 95% > 4.0-8.0 >25-4.0 > 1.5-15 > 50-15 1.92 >1550 1.75 >.10-15 262 >.0500 3.40 >.01.05 3.57 1.44 X32 3.16 3.70 3.62 357 >8.0 1.44 212 312 3.90 3.62 114 3.00 1.75 3.68 415 4.66 162 (8) Crodd. 1 94% 5.75 652 5.87 4.07 589 6.62 4.67 4.07 4.67 4.67 Q)Ooc*12 4.56 6-15 437 353 412 556 571 4J57 4S3 403 &y 496 383 417 4.70 4.73 4.30 412 4.04 4.00 (101 Crood. 3 3s*-* - A#. > 4.08.0 >154.0 > 15-25 >ioii >1550 >.10-15 > J05-.10 6.12 6.S2 5.71 5.84 >.10.05 6.62 4.58 >8.0 4.10 4.10 4J9S 4.10 4.10 4.10 4.40 5.01 31) Amovta 33% 113 3.13 353 258 3.17 3.64 3.76 238 3.36 3.56 3-31 3 03 113 2.43 2.13 3.54 261 172 3.47 2.61 P2)Oodd. 314 2.76 3.24 236 424 433 431 336 324 3.61 4.50 4.53 3.54 3.84 3 06 435 4.92 4.70 3.06 (13) Gfe 1 (MOU K>*' > 4.0-8.0 >154.0 >15-15 >50-15 >15.50 >.10-15 >-05.10 >.01-.05 223 308 293 > B.O 2.53 3.35 3.93 3.46 3.23 3.08 4.95 4.53 4.79 4.65 (14) Crood. 5 /O'* 3.59 4.35 484 4.74 4.42 281 4.19 4.3S 4.65 4.63 426 281 2.81 329 3.42 3.66 3.77 3.42 281 281 3.77 281 3.59 319 (15) Gt* 2WSD) //% 3.51 231 333 3JS3 406 151 321 291 3.61 3.69 321 3.99 361 4.11 4.02 3.81 3.38 (16) Glass 3 OCU >4.08.0 >154.0 >1325 1.45 >50-15 2.95 > 25-.50 2.55 3.16 >.1015 3.03 3.16 >.05.10 2BS 409 >.01-.OS 3 03 3.73 pm >.01-1 >14 >8.0 0.67 0.67 2.40 3 33 3.76 303 >48 0.67 1.52 0.97 203 219 3.42 2.74 3.63 3.03 325 303 >8-64 >64 (17) Glms 4 (M6U 2.81 2*4 2*4 2*4 >.0M 1JB4 2 05 259 244 3.52 3.16 3*4 3.44 356 3.44 3 35 >14 >48 Length 153 113 276 213 3.70 290 3.39 3.62 3.70 3.14 >8-64 >64 (181 AJunwiI 120 3.12 226 >.01-1 >1-4 1.10 1-82 1.61 1.93 0.B0 257 1.73 >48 120 1 ,B0 0.63 1.90 1.0 117 1.68 2.78 221 2*2 351 273 2.63 >6 >64 Hf.i'RK I.--Fiber distribution by common log of the number of particles per microgram in each ot 34 dimensional categories. it i l) JX. VOL 7. NO S. NOVEMBER I9BI 511964 0051 968 Stanton, Layard, Tegerls, et al. O ttm ttsr (19)C1ki> 5 IKW) 69% > 4,oa.o > 7.6-4.0 15-25 0.66 > JO-15 >76i0 748 1.06 150 0.67 >.1825 >JS-.10 >01.05 > 80 037 179 V.70 1.67 1.67 0.06 0.06 0.36 0.76 0.06 1.41 1.05 158 2.42 751 2.75 771 751 221 GO) Offw*ov 7 68% 1.77 3.11 772 707 3.01 4.1B 4.40 3.70 772 376 4.10 4.64 3.42 (22) Diwsoa 3 >4.00.0 >754.0 >15-75 >50-15 769 >7550 4.18 4.50 >.10-2S 3.91 4.74 >.05.10 5.85 6.18 >J01-.0S 550 5.17 >8.0 3.07 438 4.41 4.83 478 159 729 333 457 531 579 6.14 C23) das* 6 (MOW) 04% 278 1.40 1.70 710 771 1.10 1.10 3.17 784 372 3.55 255 1.09 718 321 711 37S 3.47 771 QS)Oodd.7 56% > 4.08.0 >754.0 >15-75 >50-15 >75-50 >.10-75 >55-30 1.69 205 1.86 >.01-55 >ao 051 1.69 131 1.91 9.91 051 121 1.51 136 775 202 729 734 251 2.05 051 B6>Oodd.8 53% 4.45 508 5.61 4.55 573 3.75 4.75 4.86 5.09 4.59 4.15 3.45 145 3.75 175 1.77 333 3.81 425 4.70 772 1.10 158 155 1.10 774 701 369 284 154 3.69 176 780 CD D*won. 4 711 338 530 503 177 272 711 3.42 4.05 6.00 3.47 04) Croc*3. 6 63% 1.66 157 722 324 4.42 4.12 3.99 1.87 208 1.81 224 248 2.80 4.01 3.47 4.17 5.10 4.74 3.65 455 5.07 5.45 4.90 4.07 G7) AJumm. 2 237 159 1.07 1.48 204 1.11 120 170 412 4.71 4.86 421 051 1.15 1.18 1.60 1.70 150 170 4.02 4.45 4.57 147 112 017 1.32 0.37 153 1.02 1.61 1.18 737 1.83 3.02 283 138 170 C28) Alumn. 3 > 4.08.0 >754.0 1.09 > 15-2.5 1.59 >50-15 >7550 >.10-75 >.05.10 >51-55 236 057 242 057 732 127 >65 1.19 157 1.70 233 1.15 219 0.97 1.72 1.75 0.19 1.82 1.70 261 0.67 1.49 241 1.59 (31) AHjnvrv 4 221% > 4.08.0 >254.0 235 >'1.525 2.E0 >5015 292 2.E3 >75.50 3.57 3.01 >3075 279 >55-30 >51-55 159 >8.0 1.S8 248 230 245 3.19 159 279 269 1.75 158 215 237 128 275 0.98 2.83 729 279 09) Crock!. 9 33% 3.13 3.43 431 524 5.48 113 475 452 5.14 143 425 414 021 Oood. 10 J/% 3.79 4.69 5.03 533 4.96 3.10 3.10. 3.57 4.57 5.01 4.59 4.46 3.40 3.57 3.79 3.10 3.13 3.13 3.13 4.09 3.74 3.57 3.10 3.10 DO) WoRoton. 1 3.15 3.15 378 3.41 211 3.11 3.76 375 372 378 281 3.06 IIS 211 241 (33) AHinnux. 5 22% 3.59 3.55 300 1.48 266 3.77 424 4.13 4.02 3.70 1.78 218 271 3.41 472 424 352 252 211 3.19 211 259 2.C8 206 259 376 1.43 352 3.57 123 OS) GUs> 7 (KCP) 21% 1.81 1.44 70S 70S 217 300 3.59 117 124 3.88 3.10 324 328 3.10 250 290 2.20 2.39 767 >51-1 >1-4 >45 Ungth 0.97 201 1.S1 1.81 231 057 28S 201 255 250 >8*4 >64 06) WoHeston. 3 i% 3.12 332 316 - 3 42 258 3.58 4.09 332 3.32 3.48 258 258 3.12 346 228 278 258 276 >.0!-l >1-4 >48 >864 >6*. Ttxr-nci.'RK I (continued).-- Fiber distribution by common log of the number of particles per microcrjrn in each of 31 dimensional categories. JNCI. VOL t7. NO. 5. NOVEMBER 1981 511964 0052 Carcinogenicity of Fibrous Minerals 969 nn Knoy. i ?0% > 4.G8.0 > 254.0 > 1.525 > 50-1.5 >2550 >.1025 > 5S-.10 > 51-.05 6.47 607 626 655 658 6-21 > 8.0 455 6.16 P0) Otod. 11 19% > 4.08.0 >154.0 >15-25 238 303 > 50-15 >3550 3.15 3.05 238 1.90 >.10-35 257 > .05.10 202 >51-.05 >8.0 1.72 242 1.72 130 215 250 1.42 mi Cm 19 tnot 256 267 156 216 292 237 226 1.73 136 273 243 136 P9) Gift** a (KUP1 Z83 3.36 3.36 3.17 ill 173 333 3.54 350 332 231 136 152 171 3.54 2.84 161 163 2.04 1.74 2.31 2.44 1.26 3.19 1.96 157 2 31 160 131 M2)GU 9 (M8U 1.49 260 155 0.77 1.12 1.45 0.17 1.73 202 211 237 235 2.42 238 155 wv 6 >4.08.0 >254.0 0.48 >15-25 > JO-15 0.75 > 2550 2.10 0.82 132 0.82 > .10-25 > .05-.10 > .01-.OS >8.0 0.12 0.67 0.12 032 ^5r Wo*Urton. 2 12% >4.080 >254.0 278 > 1.525 278 > 5015 3.32 3.60 >35.50 3.43 232 >. 10-35 296 336 > .05.10 4.37 4.00 > .01-.05 >80 278 3.18 218 2.96 248 0-37 032 132 1.44 0.67 0.70 154 1.49 037 0.63 1.84 0.82 <44) Dew^ort. 6 13% 6.17 656 7.86 6.47 6.86 878 240 1471 Ood-12 287 331 4.00 4.00 367 237 231 3.73 4.21 4.00 3.50 237 3.55 339 3.53 3.4S 3.17 (45) DcnvsCtfV 2 12% 3.80 4.57 452 432 433 254 243 246 331 3.10 (48) ArtaptJ. 2 257 227 3.33 3.05 3.69 357 3.17 229 631 6.16 7.05 514 654 622 639 651 244 282 274 244 1.74 269 244 1.74 ('3! Glass 10 (M0S) >4.08.0 > 254.0 2.97 > 15-2.5 3.43 > 50-15 > 35.53 3.88 4.34 3.91 4.02 > .10-35 4.43 3.88 > .05.10 5.90 4.19 > .01 -.05 6.77 4.63 > B.O 2.46 2.99 237 276 3.89 272 1.76 1.17 2.46 3.37 8% > 4.08.0 > 254.0 > 1.525 > iMi > 35 50 > .1035 > C6.I0 > .01-05 5.75 662 6.96 5.12 5.48 6.50 6.65 665 >.01-1 >1-1 >8.0 4.64 4J94 >48 >064 >64 CO) Glass 11 03P) 4.12 438 4.11 331 ZB9 3.84 4.53 3.02 2.83 3.09 356 3.60 1.18 2.41 2.66 1.81 (S3) Talc 1 7% 4.00 4.15 4.19 3.70 364 4.02 4.56 397 391 327 3.09 3.70 327 279 279 3.09 >51-1 >14 >4-9 > 6-64 > 64 {51) Tfc*r*l3 8% 254 3.3S 331 3.10 255 208 i.7& 1.78 1541 Glass 12I02P) 2.83 2*4 2.082.41 206 2.49 2.83 0.48 1.65 232 2.48 155 151 1.68 1.78 217 2.09 1.94 1.84 >51-1 > 1-4 > 4-8 > 864 0.4S o.*s 043 0.43 043 >64 11'-1 Kt. I {continued).--Fiber distribution by common log of the nnmbei of paiticles per microgram in each of 3 f dimensional categories. JNCf. VOI. 67. NO. S. NOVKMBKR 1981 511964 0053 970 Stanton, Layara, legens, ei ai. Dlim rtw (551 Gls 13 K<T1 % > 4.88.0 >284.0 2.89 > 18-2.5 3L20 >581.5 > 2530 >.1(1215 > .05-.10 3.19 2.96 3.30 3-27 3.48 1-22 >.01-.05 >8.0 2.67 2.73 1.52 152 7.03 2.03 092 032 (56) Gtn* 14 (P2P) 6% 256 121 236 156 2.30 2J28 2.51 139 0.81 18S 2.08 0.91 1.21 200 231 1.70 069 0 81 1.05j (57tOeu 15 (Y7P) 3.42 1.94 258 209 1.08 212 1.74 1.60 3.12 1.60 136 2.24 2.16 1.64 1.60 060 1.30 S) AKxnin. 7 b% >4.08.0 >ao T.19 > 2.54.0 3.52 352 >182.5 >80-15 0.64 > .25-50 1.17 0.02 T.15 ~T24'~ T.19 T.to T.08 > .10-25 >88.10 >81-85 127 0.31 ~XaT 021 "T.43 1.64 128 3.15 0.51 3) &w> 16 (MSS) 97* 3.41 3.83 321 285 221 3.17 232 248 2.45 275 243 3.46 214 290 3.59 2.81 127 AK. 4.13 T35~ 331 328 2.81 434 4.8S 3.S8 3.81 3.41 281 (61) Tele 2 4% >4.08.0 >284.0 2.58 >1825 3.06 >50-15 228 326 > .2550 >.1825 >88.10 >81-85 >8.0 2.76 321 2.45 2.45 2.68 2.56 228 188 (62) Tele 4 4.06 3.40 428 262 4.19 4.81 4.68 422 3.32 322 332 279 3.02 AMlWk O L" | 286 1 2Jn 222 1 242 287 2.46 212 159 225 234 1.70 1.72 1.86 0.74 164) Glass 21 (SI) 6* > 4.88.0 >284.0 1.15 >15-25 1.15 >50-1.5 1.46 206 >2850 1.15 1.46 >.1825 >58.10 > 51.05 >8.0 1.15 1.8S 1.63 1.1S 236 1.46 1.46 2.00 1.15 (65) Gtass 22 (S2) A7* 1.47 1.47 137 1.47 1.77 1.17 1.47 1.17 1.17 2.07 22S CS6) Glass 17 IM6S) 1.87 1.77 | 5.17 | 5.60 l S45._ j S.16 245 4.83 4.88 4.73 438 438 1.87 358 285 3.79 215 (67) Gleu 18 (YW) 0% > 4.00.0 >284.0 > 15-25 > .501.5 >.25-50 >.1825 >.08.10 >.01-.05 >8.0 0.83 082 0.80 1.00 1.10 0.34 0.40 0.11 0.41 (68) Oood. 13 UTb 430 4.63 4.90 4.68 431 238 4.00 4.46 4.00 3.63 2.68 2.98 2.98 3.16 268 P59)WoTjsstoa. 4 0% 1.43 0.95 1.56 125 1.43 1.86 125 055 1.80 1.43 2.07 1.99 151 136 055 035 0.95 035 (701 Tele 5 0% > 4.06.0 >25-4.0 3.43 > 1.825 453 >5815 4.62 438 >28.50 4 56 430 >.loots 437 330 >.08.10 397 360 >.01-.05 3.90 3.13 jm >51-1 >1-4 >8.0 >45 >864 >64 (71) Tele 6 O'* 4.63 4.83 4.68 4 65 4.10 >51-1 325 3.95 4 E3 4.43 335 385 325 >1-< 32S 325 3.56 3.56 325 325 3.55 356 32S >4-8 > 8-64 Length >64 (721 Tele 7 0% J 435 ! 5 33 j 5.18 ) 5.14 j 4 82 > 01-1 437 5.09 4.93 437 463 >1-1 367 3.67 367 367 >4-8 > 064 > 64 TuXT-nciiRi 1 {continued).--Fiber distribution by common log of the number of pjititlrs per miuupm in each of 3-1 dimensional categories. JNCI. VOL 67. NO. 5. NOVEMBER 1981 'jam 511964 0054 aft Carcinogenicity of Fibrous Minerals 3( I :.\pt No. \ 2 3 4 g g 9 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 22 33 34 35 35 Compound Titanate 1 Titanate 2 Si carbide Dawson 5 Tremolite 1 Tremolite 2 Dawson 1 Crocid 1 Amosite Crocid 4 Glass 1 Glass 3 Glass 4 Atumin 1 Glass 5 Dawson 7 Dawson 4 Dawson 3 Glass 6 Crocid 6 Crocid 7 Crocid 8 Alumin 2 Alumin 3 Crocid 9 Wollaston 1 Alumin 4 Crocid 10 Alumin 5 Glass 20 Glass 7 Wollaston 3 Table 1.--Summary of 72 experiments with different fibrous materials 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+4.7 100 100 100 100 100 954.8 946.0 93+6.5 93+6.9 937.1 86+9.0 8513.2 7810.8 77+16.6 74+8.5 719.1 70+10.2 69+9.6 689.8 6612.2 6613.4 6417.7 63+13.9 5611.7 53+12.9 44+11.7 41+10.5 339.8 3112.5 2812.0 3713.5 229.8 22+10.0 21+8.7 19+10.5 Common log fibers/pg. <0.25 pm x >8 pm 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 Expt No. Compound Actual tumor incidence 37 Halloy 1 38 Halloy 2 39 Glass 8 40 Crocid 11 41 Glass 19 42 Glass 9 43 Alumin 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 58 Alumin 7 59 Glass 16 60 Talc 3 61 Talc 2 62 Talc 4 63 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 7 4/25 5/28 3/26 4/29 2/28 2/28 2/28 3/30 2/27 2/25 2/27 2/29 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 Q/115 0/29 0/24 0/30 Q/30 0/29 Percent tumor probability + SD 20+9.0 23+9.3 19+10.3 198.5 15+9.0 14+9.4 13+8.8 13+6.9 12+7.9 12+8.0 10+7.0 _ 11+7.5 . 85.6 85.5 8+8.0 85.3 7+6.9 7+5.4 6+5.7 6+5.5 6+5.9 55.1 54.4 44.3 4+3.8 54.9 33.4 6+4.4 22.3 0 0 0 0 0 0 0 Common log fibers//4g. <0.25 jim X >8 jim 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 size distribution for com[i.nison. Wcllastonite (urollaston 1-4).--Woliastonite is a naiumlly occurring crystalline fiber of monocalcium silicate {15-18). Four separate samples of this substitute I* it asbestos were received from the same Canadian mine. These were graded commercially according to si/c by the designation A, B, D, and F. It was apparent it low-power magnification that only grade F was umipleiely fibrous and that these fibers were relatively Tremolite (tremolite 1, 2).--The second type of unphihole asbestos studied was tremolite, a material that has a close affinity to the talcs. Both of these sampcs were from the same lot of asbestos and were in iii-. optimal range of size for carcinogenesis. Compari- '> of these fibers indicated that they were distinctly i.tiHci in diameter than the tremolire fibers used by iith et al. (29). Amosite.--The third amphibole asbestos studied was ingle sample of South African amosite from the 1 G standard reference samples. No efforts were made 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).--Halloysite is a natural fi brous 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 those 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 J.VO. VOL. 67. NO. S. NOVEMBER IBM 511964 0055 972 Stanton, Layard, Tegeris, et at. hex.imetaphosphate. Clumping persisted in this second sample, and little different was seen between the 2 samples. Silicon carbide (si carbide).--One metallic crystal line whisker other th'an.-alumin was prepared by the General Tcchnol&gies Corporation. Silicon carbide was a single sample, which was of exceptionally fine, uniform dimension. Potassium octatitanate (tilanate 1-3).--In addition to the synthetic crystals of dawsonite, aluminum oxide, 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). 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 trols outside of these limits are mentioned in "Results." 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 grid was made at a final magnification of X3,000. The slotted grid was then placed in a Siemens electron microscope, Elmiskop 1-A, and the entire grid was 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 selected area, which generally measured about 350X150 fim, was then located on the lower magnification montage of the grid and examined to determine whether the area chosen was truly representative of the entire 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 mm. 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 40mg dose administered. For the purpose of calculation, particles were grouped into 34 dimensional ranges as indicated in text-figure 1, 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 different areas of the same montage. However, when studies of repeat samples from the original fibers were made, considerable in counts occurred. Clearly, the method is sub--; :d several errors; calibration of the electron micros., c* deviation of particles from the assumed cylindricu shipe. and sampling errors, es pecially where larg- mnicies are concerned, represent the major problem Nevertheless, the estimates are probably valid tc one order of magnitude. Consequently, the r. ire reported as the common log with the charar--of the log representing the probable limit of (text-fig. 1). RESULTS Controls have be-rz discussed in previous publica tions (4, 6, 9-11), bti riev were approached here in a slightly different wn* In addition to untreated controls we studied rats in vudt open thoracotomy was per formed and a noidsmcogenic material was either applied to the pleun. or implanted in the lung. These 3 groups (table 2) were rats from numerous experi ments that were of tirf same species, sex, and age and that were housed ir zht same quarters. The incidence of clearly apparent ziirsr-i neoplasms in untreated, aged outbred Osborn-- Mtndel female rats was essentially nonexistent. How-eve*, a itv pleomorphic sarcomas that might be confused \:.* pleural tumors occurred in the left thorax of botr zzrt-zni and, to a lesser degree, untreated controls. i,.*r: j;h these tumors involved the thickness of the che* >o-;. in most cases the tumors appeared to be den r-z either from mammary gland fibroadenoma or fror. sttrere granuloma in the subcu taneous tissues. Bu .litre remained a few tumors for which no definite could be determined and which were histoiupjullv comparable with pleural sarcomas. In both me experimental groups and the control groups dies: t/utittonable tumors were counted as pleural sarcoma. These essentially confusing tu mors observed in tie ctnrols need to be taken into account in the asses-rutn: of the carcinogenicity of the experimental mateita..-.. The incidence of pleural sar- Table 2.--Incidenct : si.i-t-n." earcomas in outbred female Osbor--Xr-'S?: control rate Time, wk Untreated l*;ccarpinner,arv Noncarcinogenic pleural implants' Combined controls' 12-52 53-65 66-78 79-91 92-104 105-120 121-130 131-143 144-156 156 Total Percent 1/113 0/15 0/26 0/68 0/26 0/98 1/66 0/27 0/27 1/22 3/488 0.6 : lc.) * - ij*> -- 0/47 1/72 3/64 2/85 10/294 1/36 17/598 2.8 1/209 3/113 7/140 3/223 11/392 2/296 1/69 0/27 0/27 1/22 29/1.518 1.9 ' No. dead with pi-ur_ sa--emas/No. dead without pleural sarcomas. JNCI. VOL. 67. NO 5. N'OVKMKKR 19HI 511964 0056 Carcinogenicity of Fibrous Minerals 973 ..iin.is in all 3 control groups combined, calculated by :|u- life table method (13), was 7.74.2%. Comparison this incidence with the pleural sarcoma incidence in .lie 72 individual experiments showed that the inci dence of pleural sarcomas-in a particular experimental eruup 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 ith intrapleural implants not used as controls should mentioned. These experiments included intrapleural mutants that did not conform to the type of materials nn cr consideration because the particles were either nondurable (cotton lint, gypsum, and carrageenan), .ire of greater than respirable size (steel shavings, steel vool, vermiculite, polyurethane, tungsten carbide, and .nfusorial earth), or were exclusively nonfibrous (poly.Ktylic nitrile, antigorite, silicon dusts, and several glasses). None of these experiments had an incidence of pimral sarcoma that was significantly greater than the ; incidence of the combined control group. f-iom the summarization of the 72 experiments in ulile 1 and text-figure 1, even cursory examination of ihe fiber distribution suggested that particles in the irlatively thin- and long-dimensional categories were ociated with higher tumor probabilities. This obsermiion was confirmed by the statistical correlation and .giession techniques that were used in previous papers *. ", 10). The logit transformation (13) was applied to i estimated tumor probabilities (p) according to the formula: Logit = In [p/(l~p)], where In denotes the v.amral logarithm. The 34 dimensional categories indi cted in text-fi'gure 1 were arbitrarily grouped into 11 liajer categories, and the simple correlation coefficients :! ihe logit of tumor probability with the common !>:.: rithms of numbers of particles per microgram in i- h of these categories was calculated (see table 3). I he maximum correlation coefficient, 0.80, was with l-.irticles equal to or less than 0.25 pm in diameter and neater than 8 pm in length. There was no correlation itlr particles equal to or less than 4 pm in length or nil particles greater than 1.5 pm in diameter, but l iiivcly good correlations were noted with log num; - of fibers in categories greater than 4 pm in length up to 1.5 pm in diameter, with correlation efficients of 0.45-0.80. I he possibility of the existence of relationships tween the particle size distributions and tumor prob- '- ' 3.--Correlation coefficients of logit of tumor probability common logarithm of number of particles per microgram in different dimensional ranges Fiber diameter Mm Fiber length, pm <4 >4-8 >8 --1.5-4 -0.25-1.5 'JO.25 -- -0.45 0.01 0.20 -0.28 -0.24 0.45 0.63 -0.30 0.13 0.68 0.80 abilities, which arc 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 +5, +......... + bt xs, where xt............. .. x* represent the common logs of numbers of the particles per micro gram in the size categories of table 3, and a, b i......... .. bk 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: ln[p/( I --p)] = --2.62 + 03305*. (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 13 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 (tremolite 1). 6 (tremolite 2), 26 (crodd 8), 29 (crocid 9), 33 (alumin 5), 47 (crocid 12), and 71 (talc 6) (see table I 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 thru 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 oo 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 tl` ire) h ;' JNCI. VOL- . NO. 5. NOVEMIIER 19St i:V: 511964 0057 Vj. 974 Stanton, Layard, Tegeris, et at. 1.0 0.9 6.8 o 5 0.7 D 0.6 O > 0.5 C " ctocidolitc G * gloss D dawsonite L * aluminum oxide S silicon carbide A auapulgiic P " ti(ana(e T talc M *= tremolite W ** wollastonite H 18 halloysite O *= amosite CD 0.4 < CoD 0.3 c a. 0.2 MG HCW AGOOW 0.1 AGGPT - LTGGGG LTTGG CWTTGG . 0.0 0.0 0.5 1.0 1.5 s / / c G\S c s S OG _1_ 2.0 2.5 3.0 3.5 4.0 _1_____ I 5.5 6.0 LOG NUMBER PARTICLES MEASURING < 0.25 pim x > 8 nrr- Y'CROGRAM Tlxt-FICURE 2.--Regression curve relating probability of tumor to logarithm of number of parties per pg with diameter <0.25 rm and length >8 pm. clumping and- fragmentation of the particles were greater than those in the other minerals, and estimates of particle size 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 common arc carcinogenic for the pleura of the rat. Our conclusions regarding those dimensional categories that correlate strongly svith 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 pm, but that relatively high correlations were also observed with fibers in other categories having a diameter up to 1.5 pm and a length greater than 4 pm. 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) we arc dealing with cancer in the rat and thus extrapolation tp 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, thn fibers. Sections of preneoplastic pleural lesions sliov i'ic phagocytosis of both short fibers and large-diamrier fibers but negligible phago cytosis of long, thn fibers. Consequently, in these experiments we mv simply be measuring the ef ficiency of phagocyt:-.L5. Doubtless, we have little real knowledge of the wv that long, thin fibers can cause cancer, blit as Rot:.: 51 once said, "Since what we think largely determr.rrs what we do, it is well that we think something." It it spirit of this quote, it might be profitable to c:>..:ider potential mechanisms of cancer production h- onz. thin fibers. Of first intpot- tance are those hyp:i;--ees in which the progenitor of the cancer cell is no lirtctly affected by the fiber. The long latent period < : u'.d suggest that a generalized alteration either in i.-ca! milieu or systemic environ ment might be at In this regard, the abundant collagen in the pres-.-.plastic pleural scars should be noted. Considerstioi a relationship between this phenomenon and ''-n.d-satc" carcinogenesis is attrac tive, though the rr::..:tion of plastic sheets to small particles tends to recv.rt carcinogenesis. Mechanisms of solid-state carcinoe-::-r?is have been thoroughly re viewed by Brand ITT. and little more need be added. Any hypothesis : .cctming fibers must take into account the fact th:t >xh short fibers and thick fibers are less carcinogent: '.ban fine, long fibers. Since dose was fixed in weigh: : j; was different in dimension for all experiments, o:v -light consider the surface arra as a possible factor. I: were the case then fibers from the same pool tha *.-:t modified only by shortening should be equal in -r.or-producing capacity. Clearly, this is not true in tii- following experiments: 13 (glass 1. MOl.) vs. exjx rin: -: -.9 [glass 10, MOS see (/, /0)J. JNCa. VOI-. 67. NO. S. NOVKMtIKR 1981 511964 0058 Carcinogenicity of Fibrous Minerals 975 I in experiment 24 (crocid 6) and experiment 25 nxid 7) vs. experiment 40 (crocid 11), experiment 47 ,i<xid 12). and experiment .68 (crocid 13). However, in ,1,,-se examples the phagocy.tosijs variable cannot be iuled out. \ provocative explanation relates to the ability of mu-, long f`bers to penetrate cells without killing lu-m. That this can occur is evident from in vitro .Indies (33)- However, simple penetration of cells by nvcelia ol fine dimension (a notable aspect of con tamination of cell cultures by fungi) rarely produces transformation of cell cultures and thus is unlikely to iroducc cancer. However, mineral fibers differ from iiutgi in their rigidity as well as chemical content, and .me' easily could conceive of physical differences be tween the mineral fibers and mycelia that might be titical. REFERENCES ,/i I'ott F. Huth F, Friedrichs KH. Tumorigenic effect of fibrous dusts in experimental animals. Environ Health Perspect 1974; 9:313-315. ,_'i 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. Zmtralb! Bakteriol (B) 1976; 162:467-505. t lj Pott F. Friedrichs KH. Tumoren der Ratten nach i. p. Injec tion faser forminger Staube. Naturwissenschaften 1972; 59:318. ./i Stanton MF. Layaro M, Teceris 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 1977; 58: 587-603. ') Smith WE, Miller L, Elasser RE. Tests for carcinogenicity of asbestos. Ann NY Acad Sci 1965; 132:456-488. A| Stanton MF, Wrench C Mechanisms o( mesothelioma induc tion svith asbestos and fibrous glass. J Natl Cancer Inst 1972; 48:797-321. <T> Wacner JC Asbestos carcinogenesis. Am Chem Soc Monogr 1976; 173:729-736. Wacner JC, Berry G, Timbrell V. Mesotheliomata in rats after inoculation with asbestos and other materials. Br J Cancer 1973; 28:173-185. " Stanton MF, Layard MW. Carcinogenicity of natural and man-made fibers. In: Margison CP, ed. Carcinogenesis. Ad vances in medical oncology, research and education. Vol. 1. Oxford and New York: Pergamon Press. 1979:181-187. ''i --------- . 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 speciat publica tion No. 506). lit 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). l-' Gk'kc. J, Rosen SH. Moolten S. Histological characteristics of mesothelioma associated with asbestos. Ann NS' Acad Sci 1965; 132:614-622. I>i Armitace P. Statistical methods in medical research. New York: Wiley. 1971:376-377. 410-414. 'Ht Pu.ORiM HI. Do\vt> JE. Correcting for extraneous death in the evolution of morbidity or mortality from tumors. Cancer Res 1963; 23:45-48. f't Olson Rl l. Introduction. In: Lefond SS, ed. Industrial minerals and rocks. 4th ed. New York: American Institute of Mining. Metallurgical and Petroleum Engineers, Inc.. 1975. Bkeck DVV. Synthetic rcoliies: Properties and application. In: Lefond SS. ed. Industrial minerals and rocks. 4lh ed. New York: American Institute o( Mining, Metallurgical and Pe troleum Engineers, Inc, 1975. (17) Sheppard RA. Sedimentary rocks. In: Lefond SS. ed. Industrial minerals and rocks. 4th ed. New York: American Institute of Mining, Metallurgical and Petroleum Engineers. Inc, 1975. (IS) Mumpton 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 physical properties of the UICC standard reference samples, in: Sha piro HA. ed. Pneumoconiosis: Proceedings ol 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., Juty 16-20, 1977. Washington. D C: National Bureau of Standards, Nov 1978:49-63 (NBS special publication No. 506). (21) 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. Capetow n. Union of South Africa: Oxford Univ Press, 1970:28-36. (22) Timbrell V, Cilson JC Webster I. UICC standard reference samples ol asbestos. lot 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). (23) Sreil S, Leineweber J. Personal experiences with making sam ples of fibers for biological experiments. In: Pelnar 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: Ncttesheim P. Hanna MG, Deathergaie 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) Harincton JS, Allison AC. Badami DV. Mineral fibers: Chem ical. physiochemical and biological properties. Adv Pharmacol Chemother 1975; 12291-402. (27) Jackson J. Huccins CW, Amman SG. Synthesis and characteri zation of daw-sonite. Washington. D.C: U.S. Dept of Interior, 1972 (Bureau of Mines report of investigation, 7664). (23) Huccins CW. Green TE. Thermal decomposition of dawsonite. Am Mineralog 1973; 58:548-550. (29) Smith WE. Experimental studies on biological effects of tremolite talc on hamsters. In: Proceedings ol the symposium on talc, Washington, D.C, May 8. 1973. Washington, D.C: U.S. Bureau ol Mines 1974:43-48 (Bureau of Mines information circular 8639). (30) Huccins CW, Denny MV, Shell HR. Properties of polygorskite. an asbestiform mineral. Washington, D.C: U.S. Dept of Interior, 1962 (Bureau o( Mines information circular RI607I). (31) Rous P. The virus tumors and the tumor problem. In: The Hawey lectures series. No. 31. Baltimore: Williams tc Wilkins, 1935-1936:74-115. (32) Brand KG. "Solid-state",, or "foreign-body" carcinogenesis. In: Symington T. Caner * RL. eds. Scientific foundations of oncology. London: William Heinemann Medical Books, 1976: 490-495. (13) Wade MJ. Ijpkin 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. Prnarih, Wales. Sept 4-7, 1979. Penarth. Wales: Medical Research Council. 1979 (P388D1). JNCI. VOL. 67. NO. 5. NOVEMBER 1981 511964 0059