Document vV4YQr2YVOGybm5Yo00JeEZpw

AMERICAN JOURNA OFw. INDUSTRIAL MEDICII W" . Volume 3, Number3,1982 Editornn-Chief Irving J. Selikoff Deputy Editor Ruth Lilis plaIntSps111^ EXHIBIT I Texaco 205 - * ` HI m?y; 173 s* Associate Editors ' Eula Bingham Norton Nelson 179 "iVo = . * r . Bertram W. Carnow John M. Peters 191 if*!!?. ^'Liy ; Thomas VV. Clarkson E. Cuyler Hammond David P. Kail Anthony Robbins :oi >?' fe ; f- it Kaye H. Kilburn J. Donald Mijlar , Raymond R. Suskind Arthur C. Upton :u9 . - :w -! 225 \ 1i *i ? 1 1i Assistant Editors Henry A. Anderson J. George Bekesi Alf Fischbein Arthur L. Frank Sylvia Frank Edwin C. Holstein Arthur M. Langer ., y; Albert Miller William Nicholson Yasunosuke Suzuki Jose A. Valciukas Mary S. Wolff . T/-___.. ^ ,Jfc workers in ihe )' jiiional .<ers in the n, Feburary 27, i, Washington, going commcr- ?nal communi st), 1980. ysis of selected vsbesios. Syra- itrations of as- 1. dard: Occupashington, DC, personal com- ey, 1972-1974. ure to asbestos j, NY. 1975. ,/|f ~\198 la. /119981b. T ' rSO-2000. dory workers: stilutc for Oc- W>. y r Occupational Exposure to Asbestos 311 Selikoff 1J. Hammond fee. Seidman H: Mortality experience of insulation workers in the United States and Canada.-'Ann NY Acad Sci 330:91-116. 1979. Selikoff IJ, Nicholson'WJ, Lilis R: Radiological evidence of asbestos disease among ship repair workers. Am J Industrial'Mcdicine 1:9-22, 1981. Selikoff IJ: Disability Compensation for Asbestos-Associated Disease in the United Stales. US Depart ment of Labor J-9-M-8-OI65. June 1982. Steinfurth R: Director of Health Hazard Program, International Assoc, of Heat, Frost Insulators and As bestos Workers. Washington, DC (personal communication, 1980). Stumphius J: Epidemiology of mesothelioma in Walchern Island. Br J Ind Med 28:59-66, 1968. Wong O: An epidemiologic study of workers exposed to brominated chemicals: With a discussion of multi M: factor adjustment. Banbury Report 09, Cold Spring Harbor Laboratory, pp 359-378, 1981. Verma DK, Middleton CG: Occupational exposure to asbestos in the drywall taping process. Am Ind Hyg Assoc J 41:264-269. 1980. Zumwalde R: National Institute for Occupational Safety and Health, Cincinnati, OH (personal communi cation, 1980). Classification ic Administra- 1 ber containing enance and rei observation. 64. i NY Acad Sci 5 in the United i Cancer J for American Journal of Industrial Medicine Jtzpv-on tiyozj Occupational Exposure to Asbestos: Population at Risk and Projected Mortality-1980-2030 William J. Nicholson, PhD, George Perkel, ma, and Irving J.Selikoff, md Estimates have been made of the numbers of cancers that are projected to result from past exposures to asbestos in a number of occupations and industries. From 1940 through 1979, 27,500,000 individuals had potential asbestos exposure at work. Of these, 18.800.000 had exposure in excess of that equivalent to two months employment in pri mary manufacturing or as an insulator' (> 2-3 f-yr/ml). 21,000,000 of the 27,500,000 and 14.100.000 of the (8,800,000 are estimated to have been alive on January I, 1980. It is further estimated that approximately 8,200 asbestos-related cancer deaths are now occurring annually. This will rise to about 9,700 annually by the year 2000. There after, the mortality rate from past exposure will decrease, but still remain substantial for another three decades. Key words: asbestos, occupational exposure, risk assessment, morality projections INTRODUCTION A large volume of research has been conducted on the adverse health effects of exposure to asbestos. However, relatively little is known about the magnitude of the population at risk to asbestos-related disease. A number of occupations and industries have been identified as involving substantial occupational exposure to asbestos, but no detailed evaluation has been made to quantify the number of persons whose employ ment experience has resulted in sufficient exposure to warrant characterizing them as at risk. This analysis is designed to provide an assessment of the extent and consequences of occupational asbestos exposure in the United States between 1940 and 1979. The task of estimating the population at risk to asbestos-related disease is compli cated by a number of factors: Environmental Sciences Laboratory, Mount Sinai School of Medicine (CUNY), New York. The analysis was prepared as part of a study for the US Department of Labor entitled "Disability Compen sation for Asbestos-Associated Disease in the United States," June 1982. Address reprint requests to Dr. William J. Nicholson, Environmental Sciences Laboratory, Mount Sinai School of Medicine (CUNY), One Gustave Levy Place, New York, NY 10029. Accepted for publication July 23, 1982. ' 0271-3586/82/0303-0259$ 14.00 1982 Alan R. Liss, Inc. I. 260 Nicholson, Perkel, and Selikoff .. 1. The precise number:of persons occupationally exposed to asbestos at any given time is not known. . 2. The level of exposuredo asbestos necessary to increase the risk of incurring as bestos-related disease is only imperfectly known, estimates being complicated by the varying interactions of the two elements that go into "dose" (time and intensity). 3. The extent to which workers have changed occupations and/or industries from time to time so as to place them at risk to asbestos-related diseases (or to end such exposure) at any time in the past four decades is not known. We have sought to overcome these obstacles by compiling the best available data concerning worker exposure to asbestos and the turnover of workers in the occupations and industries involved. The sources and methods used to estimate the population at risk are set forth below. T A B L E 1. Selected Asbestos Products and Their End Uses' j )' ' . _ j ) ) MATERIALS AND METHODS Identification of Industries and Occupations at Risk Workers are exposed to asbestos in a wide variety of industrial pursuits from mining and milling to primary manufacturing (producing manufactured goods from raw asbestos fibers) to secondary manufacturing (processing asbestos manufactured products to make other products) to consumer industries (utilizing a finished product containing asbestos without modification) [Daly et al, 1976]. ' Mining and milling. Fewer than 600 persons in the United States are employed in mining and milling asbestos [Meylan, 1978], In view of the small number involved and the lack of information on employee turnover, we have excluded this industry from our estimates. Primary manufacturing. The Asbestos Information Association has estimated that there are upwards of 3,000 discrete uses of asbestos. A selection of major asbestos products and their uses is presented in Table I. The primary manufacturing industries in which asbestos products are produced and which involve substantial asbestos exposure to production and maintenance employees are as follows: Asbestos products industry (SIC 3292). The major products of this industry are friction products, asbestos-cement pipe and sheet, asbestos textiles, floor tiles, roofing felts, insulating materials, and other asbestos building materials. Extensive data indicate that excessive fiber concentrations existed in the produc tion of asbestos products during previous years. In a study of retirees from one of the largest asbestos products manufacturers, Henderson and Enterline [1979] categorized work exposures according to total dust concentration (as measured by a midget impinger) times period of employment. Using recently obtained data on the conversion between such particle counts and fiber concentrations, it is estimated that the average concentration to which the members of his cohort were exposed was 30 fibers/ml [Asbestos Information Association, 1979], Similar concentrations were suggested for the work force exposure in a large United States asbestos products .manufacturer studied by Nicholson et al [in press]. Here subjective data,' consistent with company measurements of dust concentrations, suggested that the person-weighted average exposure was approximately 25 fibers/ml between 1945 and 1965. In two asbestos insula tion manufacturing facilities in Port Allegany, Pennsylvania, and Tyler, Texas, aver os at any i.- by the isity). industries a end such ilable data ccupations julation at suits from oods from rufactured ed product : employed ;r involved -if Trom ) .) ' amated or asbestos > industries al asbestos adustry are .es, roofing he produci one of the categorized midget imconversion the average D fibers/ml ggested for inufacturer h company average ex:stos insula.'ev_'^aver- Occupational Exposure to Asbestos 2 'a !) .cclt 3 ?.. oA 3 SO - T O<0 tJ ou 53 O c < ' Ht o "eo . y' CO 3o ^-Co- : C.'O t. 6o 22 Eo ^o u O -J *-- UJ 2 Z u_ a 'Co-ot .33O** *oc I 5 <3 .^3 o5 5-= &J < .<Eyao `acC3a s3^ i?-?E .r**o3f S 2 -- 3 geo. o aaH su: o --O xu =8 20 *--2 .ui. 6 U"J -i ,U 00 J c 1 2 .5 nic2:. Seoa 5ooE ---t -g 6 > E So .9e* og c o. o cl 3 e -p -C *o G gV C<4 T33 5 O .5 3 o - *5 ogoE* 5a5 .*3. u u E < CJ OCO IEI- oo x2 v<> "cO .w> O D UuJJ i < gESopQ. ca5a o2 oiE> e=Vco 2cococ. U M^ v C ->rt CS" 0E3. H=ps s O .3 CC>--o*u) tgieco; ^<8cJ o OO* 1 . 2 ' s0-3S7Z: ; P g a o 'i S .> ~o SEo0 --SS! a, *-- Ro 2e ,,a .VS> a -cO tcgj. s.g? l&Ss oo Eo 2fc .u2 .050 G Mc> gg; a i^= S "^5 sc3 ?= ."59 " d-n2 O-<Sa cooCEOUI.Us,,3LtU,,*Oi f. 262 Nicholson, Perkel, and SelikoET age concentrations of 35 fibers./mf were measured by NIOSH between 1968 and 1971 [National Institute for Occupational Safety and Health, 1972). These concentrations were.characteristic of early exposure levels in manufactur ing industries. In recent years, Considerable efforts have been made to reduce fiber con centrations. During 1975, air levels of from 0.5 to 4.0 f/ml were found to characterize most primary manufacturing processes (see below). With appropriate engineering, even asbestos textile manufacturing can be controlled to levels below 1.5 f/ml [Lewinsohn et al, 1979). Since substantial asbestos exposure is involved in all production and maintenance operations in this industry, we have included all production and maintenance workers in our estimates of the population at risk. Gaskets, packing and sealing devices industry (SIC 3293). This industry encompasses products made of asbestos, leather, metal, and rubber. Prior to 1972, as bestos was the predominant raw material used. A change in the industry classification system in 1972 expanded the definition of this industry to include products made of leather, metal, and rubber [Office of Management and Budget, 1972). Since approxi mately one half of the employees of the newly defined industry were employed in plants manufacturing asbestos products, we have included one half of the production and maintenance employees since 1972 in our estimates of the population at risk. For years prior to 1972, we counted all employees in the at-risk group. Building paper and building board mills (SIC 2661). This industry covers the production of asbestos paper, asbestos board, and sheeting and various types of papers and insulating boards used in building construction. Since approximately one half of the employees in 1972 were employed in construction paper plants (where asbestos was the principal raw material), we have included one half of the production and mainte nance employees in our estimates of the population at risk. Recent (1975) fiber concentrations measured in the primary asbestos manufac turing industry have been reported in the Asbestos Information Association-Weston submission to OSHA as a response to the October 1975 proposed revision to the asbes tos standard [Daly et al, 1976). These data indicate the following asbestos concentra tions were present in the respective industry segments: . Primary industry Asbestos paper Asbestos cement pipe Floor tile Friction products Paints, coatings, and sealants Asbestos cement sheet Gaskets and packing Reinforced plastics Asbestos textiles - 1975 asbestos fiber concentrations (f/ml) Range `Typical" 0.10- 2.8 0.25- 4.5 0.25- 4.3 0.10- 22.0 0.25- 8.0 0.25- 8.7 0.10- 2.5 0.20- 3.0 0.25- 15.0 0.75-1.9 0.50-2.2 0.50-1.75 1.00-3.3 1.00-2.5 1.00-3.0 0.75-2.0 1.00-4.0 Sec containin termediat He, try is eng; and oil bi sulation i nance em Fal industry pressure products that utili. power b< counted We have mates of . Inc trial pro' devices, producti estimate El, gaged in and elec are at ri A: product manufa nents, f' als in al that fro al, 197f ing indi I 58 and 1971 '*' hur1 er con;haracterize mgineering, v 1-5. f/ml laintenance ice workers is industry to 1972, asassification :ts made of ce approxied in plants luction and c. For years covers the as of papers one half of sbestos was inc: linte- J-j s manufacion-Weston o the asbesconcentra- ) ) > 15 1 5 ) ) ) * Occupational Exposure to Asbestos 263 Secondary manufacturing. Secondary industries are those that receive products containing asbestos and further process, modify, or fabricate them to produce other in termediate or final products. The following industries involve such processes: Heating equipment except electric and warm airfurnaces (SIC3433). This indus try is engaged iri'the production of heating boilers; domestic furnaces and gas burners; and oil burners, space, and wall heaters, all of which tended to incorporate asbestos in sulation in their construction. We have included one half of the production and mainte nance employees in our estimates of the population at risk. Fabricated plate workers (Boiler Shops) (SIC 3443). Establishments in this industry are engaged in manufacturing power and marine boilers, pressure and non pressure tanks, processing and storage tanks, and heat exchangers and similar products, many of which include asbestos insulation. The subdivisions of this industry that utilize extensive asbestos insulation (heat exchangers and steam condensers; steel power boilers, parts and attachments; and nuclear reactor steam supply systems) ac counted for approximately one half of the industry's total production workers in 1977. We have included one half of the production and maintenance employees in our esti mates of the population at risk. Industrial processfurnaces and ovens (SIC 3567). This industry produces indus trial process furnaces, ovens, induction and dielectric heating equipment, and related devices. All of the subdivisions make extensive use of asbestos insulation and all of the production and maintenance employees are included in our population at risk estimates. Electric housewares andfans (SIC 3634). Establishments in this industry are en gaged in manufacturing electric housewares for heating, cooking, and other purposes and electric fans. We estimate that lCWo of the production and maintenance employees are at risk of asbestos-related disease. Asbestos is used in a variety of other secondary industries. These include friction products, reinforced plastics, products containing asbestos paper, various industries manufacturing laboratory equipment, electrical switchboards, cooling tower compo nents, fire protection materials, etc. It is impossible to extract the number of individu als in all secondary manufacturing from BLS data. The only published information is that from the Weston analysis done in cooperation with the asbestos industry [Daly et al, 1976]. They report the following 1975 employment data for secondary manufactur ing industries, categorized by the primary source of asbestos: Primary source of asbestos materials Asbestos paper Friction products Asbestos cement sheets Gaskets and packings Reinforced plastics Asbestos textiles Miscellaneous Total Number of exposed employees 158,400 27,600 19,200 12,000 8,400 6,000 8,400 240,000 264 Nicholson, Perkel, and Selik'ofT By comparison, our estimate of the asbestos-exposed employment during 1975 for the four industries listed .previously (SIC 3433, 3443, 3567, and 3634) totaled 38,000. Moreover, only employees of companies manufacturing electric housewares and fans would appear to have been included in the Weston tabulations. However, it is difficult to be certain that their classification of primary and secondary is similar to ours. In their classification, they estimate 23,000 to be exposed in primary manufactur ing in 1975 versus our estimate of 31,000. Thus, some of our primary industry may be their secondary. It is difficult to esti mate the exposures the individuals identified by Weston would have had. Some data are presented on current asbestos concentrations (see below). It is unlikely, however, that 158,000 employees would have had significant exposures during the manufacture of products containing asbestos paper. The data in the other manufacturing segments appear reasonable, however. To account for all these exposures, we will consider that a number equal to twice the four groups specified by SIC numbers are additionally ex posed in secondary manufacturing. (This additional number totals 76,000 in 1975.) Data provided by Asbestos Information Association-Weston on fiber counts in secondary manufacturing are: . Secondary industry3 Asbestos fiber concentration range reported (f/ml) Asbestos paper Friction products'3 Asbestos cement sheet Gasket and packing Asbestos-rein forced plastic Asbestos textiles 1.0-3.5 2.5-6.5 1.0-6.0 0.2-5.0 0.5-2.0 0.5-5.0 aCategorized by primary source of asbestos material. bDoe$ not include brake and dutch maintenance. No information is available on dust counts in these industries in earlier years. Shipbuilding and repair (SIC 3731). The risk of asbestos-related disease among shipyard workers was emphasized in 1968 by Harries, who reported five cases of pleural mesothelioma among employees of the Royal Navy Dockyard in Devonport [Harries, 1968]. His findings were noteworthy in that none of the patients was an "as bestos worker.''They were employed in other trades (boilermaker, shipwright, laborer, welder, fitter) and worked in shipyards with asbestos workers but did not them selves often use asbestos. In addition, cases of asbestosis were noted. Stumphius de scribed similar findings in the Netherlands [Stumphius, 1968]. Again, the mesotheli omas were among workers other than those in the usual asbestos trades. Since these initial communications, experiences have been detailed in many parts of the world iden tifying characteristic asbestos-associated disease among former shipyard workers, in cluding pleural mesothelioma, peritoneal mesothelioma, asbestosis, and lung cancer. Evidence of asbestos-associated disease has been reported among workers employed in United States shipyards during and after World War II [Department of Health, Educa tion, and Welfare, 1981; Felton, 1979; Selikoff, 1965]. These findings indicate that the nature of shipyard work during this period provided significant opportunity for expo sure to asbestos of the many traces employed, even though such exposure might have been only intermittent or indirect. \ naval s shipyai [Nunnt C of totai exposu 1 1522). 2 3 1623). 4. work], foundai classifie A are asb< panel in stallatic and elec the dire have be ticeof s 1972. A City col showed away [P penters, tion site W of the p structio: Si in 1974 the wor from as ers emp fact tha construi the tota unpubti SI [except decorati exposur rincr 1975 )ded v .wares wever, it is similar to anufactur- ;ult to estiSome data , however, inufacture l segments iiderthat a ionally exa 1975.) r counts in ion ars. . :ase among it cases of Devonport was an "asht, laborer, not themmphius de- mesotheliSince these world idenvorkers, inang cancer, mployed in lth, Educaate that the :y \xpo'vy. /have Occupational Exposure to Asbestos 265 We have .included all production and maintenance employees of private and naval shipyards in our estimates of the population at risk. The estimates for naval shipyards, however, are taken from the United States Department of the Navy (Nunneley, Department of the Navy (Personal Communication, 1980)]. Construction. The construction industry accounts for an estimated 70%-8QVo of total United States consumption of asbestos fiber [Levine, 1978], Substantial direct exposure to asbestos occurs in the following subdivisions: 1. General contractors--residential buildings other than single family (SIC 1522). 2. General building contractors-nonresidential buildings (SIC 154). 3. Water, sewer, pipe line, communication, and power line construction (SIC 1623). 4. Construction--special trade contractors (SIC 17, except 1771 [concrete work], 1781 [water well drilling], 1791 [structural steel erection], 1794 [excavating and foundation work], 1796 [installation or erection of building equipment, not elsewhere classified]). 1 Among the asbestos products involved in direct exposures in construction work are asbestos-cement pipe installation; asbestos-cement sheet installation; architectural panel installation; built-up roofing installation; drywall removal, replacement, and in stallation; removing of roofing felts; asbestos insulation of pipe, tubing, heating units, and electric power generation equipment; paints, coatings, and sealants. In addition to the direct exposure resulting from the use of the above products, construction workers have been subject to considerable indirect exposure to asbestos as a result of the prac tice of spraying asbestos insulation in multistoried structures during the period 1958-- 1972. An investigation of the spraying of mineral fiber insulation material in New York City collected on-site samples taken at various distances from the spraying nozzle. It showed fiber counts ranging from 70 f/ml 10 feet from the nozzle to 3 f/ml 25 feet away [Reitze et al, 1972], Workers in occupations not directly involved in spraying (car penters, electricians, pipefitters, plumbers, welders, and others) who were on construc tion sites during or after such spraying are at risk to asbestos-associated disease. We have included all construction workers in SIC 1522 and 154 in our estimates of the population at risk and the following proportions of the workers in other con struction subdivisions: SIC 1623. Thirty percent of the water distribution pipe sold in the United States in 1974 was asbestos cement [Meylan et al, 1978J. We assumed that this proportion of the workers in the water, sewer, etc, line construction industry is exposed to asbestos from asbestos-cement pipe. In addition, we included maintenance mechanics and help ers employed in SIC 16 (construction other than building construction) to reflect the fact that these workers are exposed to asbestos during the repair of brakes on heavy construction equipment [Hill, 1980]. These workers comprise approximately 5% of the total number of construction workers in SIC 16 [Bureau of Labor Statistics, unpublished]. SIC 17. We have included all construction workers in 171 (plumbing, heating [except electrical], and air conditioning) and SIC 172 (painting, paperhanging, and decorating) in our estimates of the population at risk. The former group has extensive exposure to asbestos in pipe covering and insulation for heating and ventilation equip- 266 Nicholson, Perkel, and Selikoff ment. A mortality study of the members of the union of plumbers and pipefitters in the United States noted their potential exposures to asbestos and found significant excesses in proportional mortality ratios for malignant neoplasms of the esophagus, respiratory system, lung, bronchus, arid trachea, and "other sites." [Kaminski et al, 1980]. Seven deaths were due to mesothelioma, a clear indicator of asbestos-associated disease. The latter group (painting, paperhanging, and decorating) has been exposed to many asbestos-containing materials, including spackle compounds used by general painters, taping and joint compounds used in drywall construction, and additions of asbestos to sealant compounds or surfacing materials. Moreover, these workers have indirect exposure to asbestos materials used by other trades in the construction indus try. A study of drywall taping workers employed in the New York metropolitan area found mean asbestos fiber concentrations ranging from 5.3 f/ml in hand-sanding to 47.2 f/ml in dry mixing operations [Fischbein et al, 1979], Other researchers report mean fiber concentrations of from 0.9to 19.6 f/ml during various activities of drywall taping [Verma and Middleton, 1980]. In addition to the tapers and painters directly en gaged in these operations, members of all the construction trades working in the vicin ity of ongoing drywall construction were significantly exposed. Mean fiber concentra tions varying from 2.3 to 8.6 f/mi were observed at distances from 3 to 20 feet from the taping operation in the same room. In adjacent rooms, background mean fiber levels varied from 2.6 to 4.8 f/ml at distances from 15 to 25 feet from the taping operations. For the remaining groups covered by SIC 17 (except the five groups identified under 4 above as not being substantially exposed), we have estimated that the propor tion of the construction workers at risk during 1958-1972 was 50% (when multi-storied buildings were sprayed with asbestos fireproofing material) and 20% during 1940-1957 and 1973-1979. The following proportions of these groups were found to be exposed to asbestos in the National Occupational Hazard Survey (National Institute for Occupa tional Safety and Health, unpublished]: SIC code SIC description 173 174 175 176 1793 1795 1799 Electrical work Masonry, stonework, tilesetting, and plastering - Carpentering and flooring Roofing and sheetmetal work Glass and glazing work Wrecking and demolition work Special trade contractors, not elsewhere classified NR, Not reported. % Employees exposed to asbestos 15 27 15 41 40 NR 23 It should be noted that the above percentages understate the proportions of "con struction workers" exposed to asbestos in these industries since they are based on the total employment reported rather than total construction workers; the latter concept 1 excludes executive and managerial personnel, professional and technical employees, ) and routine office workers [Bureau of Labor Statistics, 1976]. ating sulate calciu indust that ci ductet found maim ing or worke 10% r directl are ga with a from l ber of nificai persoi counti other (Seliki insula used v et al, time ( mixinj served the in: tos co .ory . Seven :ase. osed to general ions of rs have indusan area ding to report irywall ctly ene vicincentraom the r levels ations. mtified >roporstoried D-' V f "conon the oncept lc. V ) .3 >* . Occupational Exposure to Asbestos 267 Electric, gas, and combination utility services (SIC 491, 492, 493). Power gener ating facilities have raany work areas with elevated temperatures, which have been in1 sulated with asbestOs/containing materials, including preformed blocks of hydrous calcium silicate insulation reinforced with asbestos fibers. Other insulation used in this ' industry consists'of asbestos boards, blankets, felts, cloths, tapes, sleeves, and cements ' that contained various quantities of asbestos [Fontaine and Trayer, 1975). Studies con ducted in England (Bonnell et al, 1975) and France [Fontaine and Trayer, 1975] have found substantial evidence of asbestos-associated disease among persons engaged in maintenance work at power stations, including persons not directly involved in apply ing or removing insulation materials. We have included one quarter of the "physical workers" employed in electric and gas utilities in our estimate of the population at risk: 10% representing maintenance workers and 15% other persons in the area who are in directly exposed [H. Jones, 1980]. Occupational groups. The industrial activities for which employment statistics are gathered do not correlate closely with those in which there is occupational contact with asbestos. It has been necessary, therefore, to supplement the estimates derived from the above analysis of industrial employment statistics with estimates of the num ber of persons employed in particular occupations (crossing industry lines) where sig nificant asbestos exposure has occurred. We have reduced the industry estimates of persons at risk by the numbers employed in the selected occupations to avoid double counting. The following occupational groups were defined as at risk: Asbestos and insulation workers. A strikingly increased death rate of lung and other cancers has been observed among a group of asbestos and insulation workers [Selikoff et al, 1979]. All such individuals have significant risk. Data are available from three research groups on average fiber concentrations in insulation work prior to 1970, when the techniques of application and control measures ` used were typical of the industry during previous years [Balzer and Cooper, 1968; Ferris et al, 1971; Murphy et al, 1971; Nicholson, 1975], The data are presented in terms of ' time (and job-weighted) average concentrations. During certain operations (cement mixing, hand- or band-saw cutting, removal), extremely high concentrations were ob served (up to 100 f/ml). However, these operations constituted only a small fraction of I the insulators' work activity. Data were also estimated for earlier years when the asbes- ' tos content of insulation was twice that of 1965-1970. . Summary of Average Asbestos Air Concentrations During Insulation Work Average fiber concentration (f/ml) Research group Light and heavy construction Marine work Average concentrations of fibers longer than five micrometers evaluated by membrane filter techniques and phase-contrast microscopy Reitze-Nicholson, Mount Sinai 6.3 [Nicholson, 1975] . . Balzer-Cooper, U. of Calif. 2.7 6.6 [Balzer and Cooper, 1968] i Burgess-Lynch, Harvard 2.9 [Ferris et al, 1971] 268 Nicholson, Perkel, and Selikoff Average concentrations of all visible fibers counted with a konimeter and bright-field microscopy . Murphy, Harvard (Murphy et a), 1971] . :' Fleischer, US Navy [Fleisher et al, 1946] 8.0 30-40 Estimates of past exposure based on current membrane filter data Nicholson, Mount Sinai (Nicholson, 1975] 10-15 ________ A utomobile body repairers and mechanics. A study of brake-lining maintenance and repair work has found short-term concentrations of asbestos of 16.0, 3.3, and 2.6 f/ml at distances of 3-5 feet, 5-10 feet, and 10-20 feet, respectively, from a worker blowing dust out of automotive brake drums (Rohl et al, 1976]. Grinding truck brakeshoes gave an average concentration of four f/ml and bevelling produced an average count of 37 f/ml. Measurable concentrations (0.1 f/ml) were found at distances up to 75 feet from the blowing-out operation (14 minutes after), 60 feet from grinding and 30 feet from bevelling, indicating that other garage employees besides those directly in volved in brake and clutch repair are at risk. Average fiber concentrations during brake and clutch work, however, are much lower and average about 0.1-0.3 f/ml during the course of an entire brake repair job. These data and the sources are: Summary of Asbestos Concentrations During Automobile and Truck Brake Maintenance Activities: Long-Term Samples During Lining Removal and Replacement Source Range of all concentrations measured (f/ml) Range of garage mean concentrations (f/ml) Personal NIOSH [R. Zurrrwalde, personal communication] Sampling may have been done during non brake work. No information on work practices. Hickish and Knight (1970] Appears typical of past work practices with air blowing of drums. Sampling throughout complete brake repair job. Raybestos-Manhattan [J. Marsh, personal communication] Wellcontrolled exhaust ventilation utilized. 0.01-3.24 0.08-7.09 0.02-0.4 0.03-0.59 0.68-3.1 0.05-0.1 NIOSF comi Hickisl lniti; prevalence in agreeme Engi potential fi to the ship sea, flakin ings, stean crewmen r tearing do study of 6 ally high j fects of.in sonnel on at risk. Man facturing | extensive i materials nance wor of maintei frequent c that asbes refineries, those doctions. We (SIC 28) populatio Stei during tb bestos. T suited in i of all wdic identified We indue decade o 1950s (w' from 63. annual p Sic tion and ties for b tos-cont utility st i intenance 1, and 2.6 a worker ck braken average ices up to ng and 30 irectly inare much epair job. / arage irations 59 1 .1 i 1 ) ' NJOSH [R. Zumwalde, personal communicajipn] Hickish and Knight [1970] Occupational Exposure to Asbestos Area 0.01-1.72 0.1 -0.57 0.07-0.28 0.15 269 Initial clinical surveys of garage mechanics indicate that they have a small excess prevalence of X-ray abnormalities (-5%) compared with blue-collar control groups, in agreement with the dust count information above [Nicholson, 1982]. Engine room personnel, seagoing vessels, United States Merchant Matine. The potential for exposure to asbestos insulation material on merchant ships is not confined to the shipyards where the ships are built or repaired. After the vessels have been put to sea, flaking and cracking of the asbestos insulating materials covering machinery cas ings, steam and hot water piping, and tanks are common. In the course of a voyage, crewmen make repairs on pipes, pipe flanges, or valve leaks. This generally requires tearing down the insulation materials and replacing them [Polland, unpublished]. A study of 6,671 X-ray films of marine engineers in the United States showed an unusu ally high proportion (16%-20%) of pleural abnormalities, indicating the adverse ef fects of inhaled asbestos. [R.N. Jones, 1980], We have included all engine room per sonnel on seagoing vessels of the Merchant Marine in our estimates of the population at risk. Maintenance employees: Chemicals and petroleum manufacturing. The manu facturing processes of chemical production and petroleum refining involve the use of extensive networks of pipes, boilers, and other high temperature equipment. Asbestos materials provide thermal insulation for these networks and a large force of mainte nance workers is employed to maintain and repair the production equipment. A study of maintenance workers in a large chemical plant and an oil refinery showed relatively frequent chest X-ray abnormalities [Lilis et at, 1980], These findings strongly suggest that asbestos exposure characteristic of maintenance work in chemical plants and in oil refineries, including indirect ("bystander") exposure, results in risks comparable to those documented for other types of asbestos exposure in other industries and occupa tions. We have included all maintenance workers in the chemicals and allied products (SIC 28) and petroleum refining and coal products (SIC 29) in our estimates of the population at risk. Steam locomotive repair. Employees engaged in the overhaul of railroad engines during the period when steam locomotives were in service were heavily exposed to as bestos. The practices used in the "back shops" where overhauls were conducted, re sulted in the generation of clouds of asbestos dust that contaminated the environment of all who worked in the area [Mancuso, 1976], Five mesotheliomas have recently been identified by NIOSH among former employees of one shop in Reading, Pennsylvania We included all employees of railroad repair back shops in our at-risk estimates for the decade of the 1940s (when steam locomotives were the predominant type). For the 1950s (when the proportion of all locomotives in service which were steam declined from 63.4% to 1.7%), we reduced the annual number of employees at-risk by the annual proportion of nonsteam locomotives to all locomotives. Stationary engineers, stationary firemen, and power station operators. Opera tion and maintenance of stationary engines and mechanical equipment to provide utili ties for buildings and industrial processes involve the same types of exposure to asbes tos-containing materials as are described above under electric, gas, and combination utility services. A preliminary field survey of 34 stationary engineers by this labora- j . | }- ,\ 270 Nicholson, Perkel, and SejikofT tory in the New York metropolitan area has found X-ray abnormalities consistent with asbestos-induced changes in 60% of the employees with more than 20 years of experience in this trade. We have included all employees in this occupational group in our estimates of the population a.t fisk. Population Estimates ' In estimating the mortality (or morbidity) from past exposure to asbestos, we would wish information on the number of individuals exposed; the distribution of their employment periods; the time, duration, and intensity of the asbestos concentrations to which they were exposed; and mortality data, by industry, correlated with the above variables. Unfortunately, we have little of the above data. There are limited data on the number of individuals exposed to asbestos in different calendar periods of time. For some industries data are good (primary asbestos manufacturing, shipbuilding, auto re pair and, to a lesser extent, insulation work). Much less certain are data on exposed populations in construction, secondary manufacturing, and the maintenance indus tries. Least certain is information on the turnover in a given industrial segment. Expo sure data are available in recent years, but generally only from a limited number of measurements in an industry. Extrapolations to earlier years are possible but neces sarily uncertain. Of most use are current data on the mortality of entire population groups exposed in previous years. Such information, if related to exposure periods, eliminates our need for information on exposure distributions as the mortality data for an entire group includes all exposure circumstances. Further, as will be demonstrated subsequently, several studies show that the risk of lung cancer is linearly related to the total fiber exposure. This information allows one to properly account for different durations of employment in a given industry. Moreover, for the purposes of estimating excess mortality, it also reduces our need for accurate information on work force turnover within an industrial segment. The excess mortality for 1,000 men exposed for ten years is the same as for 2,000 men exposed for five. The important parameter is the person-years-at-risk. Thus, information on the total work force exposed at various points in time is much more important than infor mation on turnover. However, for consideration of surveillance activities, one would wish knowledge of the total population at risk. This can be estimated, but greater un certainties exist in the values obtained than in the number of asbestos-related cancers that might develop. Methodological Considerations Considerable information is available from data published by the Bureau of Labor Statistics and from industry or union sources on the number of individuals em ployed in an industry at periods of time. Data from publications of the Department of Labor also provide some information on the number of individuals entering or leaving a given industry on a yearly and monthly basis. For some industries subsequent to J958, this includes information on the fractional number of accessions and separations that occurred for given employees within a calendar period. Often data are provided on the total fractional number of new hires, recalls, layoffs, and quits. While the information on the fractional number of new hires is of use to us in estimating the population enter ing a given industry, it does not represent true new hires for our purposes. This is be cause the industry data are based on individual establishment experiences. A new hire for an establishment may bean individual who previously worked in another establish ment in t great a di duplicatic To able to he or indust: currently dustry, c* given yea permanei x (a - For trants inp the absen In any ste employm time, AN (AN/N). force in e force r.ep ous chan new entr ment, to bestos. T a decade l/a = k posure v sleady-s, rather th essary if consider Asbesti TV estimate ings in t ploymej turing; gas and industri \V the basi as far b sion eqi trend t (Censu; h ers."Tl \opc'sfent with I Jfexperi- oup in our > asbestos, we bution of their :oncentrations vith the above ed data on the s of time. For Iding, auto rela on exposed enance indus:gment. Expoed number of ble but neces,re population osure periods, tality data for w that the risk nvp'ion allows i\ |idustry. ) . .'need for .The excess ;n exposed for nation on the int than infores, one would >ut greater un elated cancers :he Bureau of idividuals emDepartment of ring or leaving quent to 1958, parations that rovidedon the le information yulation entertes. This is be es. A new hire I \stablish- ) Occupational Exposure to Asbestos 271 ment in the same industry. For some manufacturing industries, this may not be too great a duplication, but for construction trades particularly, it represents significant duplication. . ; To estimate the population at risk for a period of years, it would be most desir able to have information on the number of new employees entering a given occupation or industry at different points in time and information on the number of individuals currently leaving that occupation or industry permanently. If N = the number in an in dustry, a = the fractional number of new entrants in an occupation or industry in a given year {Nnew/N10i). and j8 = fractional number leaving an occupation or industry permanently, the change in an industry work force can be represented by dN = N x (a -j8)dt. For small changes in N, N = N0e(" - WL In this model, in the absence of new en trants into an industry, the work force will decrease with a half-life, T0.5 = 0.693//3. In the absence of any separations, it will increase with the doubling time, T2 = 0.693/a. In any steady-state or near-steady-state situation, where a = /3, the average duration of employment is equal to 1 /a. When one considers finite changes over a year period of time, AN = (a - 0)N, where AN is the net increase or decrease. Thus, a = j3 + (AN/N). If we consider the time necessary to achieve complete replacement of a work force in a steady-state situation, AN = N = aNT. Thus T, the time necessary for work force replacement is equal to 1 /a as expected from the earlier consideration of continu ous changes. As indicated previously, we will be using information on the number of new entrants into a trade or industry, coupled with their average period of employ ment, to generate estimates on the expected excess mortality from past exposure to as bestos. The excess mortality among a group of individuals entering an industry during a decade will be proportional to aN x T (new hires x employment period) = kaN x 1/a = kN, where k, the proportionality constant, includes the appropriate risk and ex posure variables for the industry. Thus, the crucial item in estimating mortality in a steady-state work situation is information on the number employed in an industry rather than the number ofnew hires entering it. More detailed information is only nec essary if there are significant changes in the workforce over the period of time being considered. Asbeslos-Exposed Work Force The data on the population exposed to asbestos in different industries has been estimated using the Bureau of Labor Statistics information on employment and earningsrin the United States, 1909-1978. Here direct data are available on the yearly em ployment in the following industries under consideration: primary asbestos,manufac turing; selected secondary asbestos manufacturing industries; construction; electric, gas and utility services; and chemical and oil refining employees. The segments of these industries that will be considered at risk have been described previously. We used employment series published by the Bureau of Labor Statistics [1979] as the basis for estimating the number of persons employed. Where the data do not extend as far back as 1940, we extrapolated the BLS series to that year on the basis of regres sion equations with related variables (Table II) or on the assumption of a straight-line trend between Census Bureau data for 1939 (Census of Manufacturers) or 1940 (Census of Population) and the earliest year of the relevant BLS series. In the construction industry, the employment data relate to "construction work ers." This group covers "workers up through the level of working supervisors, who are 272 Nicholson, Perkcl, and SetikolT engaged directly on the construction project either at the site or working in shops or yards at jobs ordinarijy performed by members of construction trades. Exclusions from this category include executive and managerial personnel, professional and tech nical employees, and routine office workers" [Bureau of Labor Statistics, 1976]. in electric, gas, and combination utility services, the employment data relate to "physical workers." This group includes working foremen and other nonsupervisory workers engaged in nonoffice functions [Department of Labor, I979J. In manufacturing industries (including private shipbuilding and repair), the em ployment data relate to "production workers." This group covers those employees, up through the level of working supervisors, who are engaged directly in the manufacture of the product. Among the exclusions from this category are persons in executive and managerial positions, those engaged in office work, and professional and technical functions [Bureau of Labor Statistics, 1976]. In the chemicals and allied products industry, it was estimated that 27% of the BLS employment figure represented maintenance workers. This proportion was calcu lated from the BLS Reports on 1971 occupational employment in this industry [Bureau of Labor Statistics, 1974]. The following classifications were excluded from the main tenance occupations to avoid duplication: insulation workers, stationary engineers, stationary boiler tenders. In petroleum refining and coal products, it was estimated that 40% of the petro leum refining production employees and 20% of the production employees in the re maining divisions of the industry represented maintenance employees [Bureau of Labor Statistics, 1965]. The 1940 employment in the industry was estimated on the basis of a straight-line interpolation between the 1939 figure reported by the Bureau of the Census [1939], and the 1944 BLS figure. The same maintenance occupations were excluded as is noted under chemicals (above) to avoid duplication. Data are not available that allow direct use of BLS employment data to estimate the number of individuals employed in insulation work, shipbuilding, automotive maintenance, merchant marine engine room work, and steam locomotive repair. Sepa- TABLE II. BI.S Employment Series Extrapolated to 1940 by Means of Regression Equations Series to which extrapolation was applied Related variable used for estimation Measure of validity (r1) Construction --general building contractors: construction workers (SIC 15) Construction--other than building general contractors: construction workers (SIC 16) Construction --special trade con tractors: construction workers (SIC 17) Electric, gas, combined utilities employed Manufacturing: heating equip ment excluding electrical: production workers Construction--all employees 1964-1973 (SIC 15. 16, 17) Construction --all employees 1960-1971 (SIC 15. 16, 17) Construction--all employees 1947-1956 (SIC 15. 16, 17) Production of utilities 1950 1959 Manufacturing' -- fabricated structural metal products: production workers, 1972 1979 . 0.97 0.68 0.99 0.84 0.61 rate data US Nav> Ins we will i sulators; the new t and Asb data ava on the ct bers. Als number t where su from the bership i: tion is a; age 65. Th40% for union jol bership h 1940, fev jobs at th employet work fot would h Un were at w ditional ! Labor da as descril o f the wc the estim decrease sulators ' ployed in Shi yard wor the US N 22, 1980] 50% of tl that we w fore, the Naval shi 80% in 1' 175,000, 1 Based on t sttudion (.: fin shops or ilusions . and tech1976]. data relate to msupervisory jair), the emmployees, up manufacture txecutive and ind technical t 27% of the an was calcurstry [Bureau am the mainry engineers, of the petroees in the re ; [Bureau of nated on the he Rureau of f hs were '] . > i, ,o estimate , automotive repair. Sepa- llons Measure of vaiidily (r1) 0.97 0.68 0.99 0.84 0.61 . Occupational Exposure to Asbestos 273 yA It -if m Vo" v vS-: t'% rate data are available in these industries from union sources, trade associations, the US Navy, and other government sources. Insulation workers. For this important group of asbestos-exposed individuals, we will utilize.information from the International Association of Heat and Frost In sulators and Asbestos Workers (lAHFlAW)to estimate the work force at any time and the new entrants into the trade [International Association of Heat and Frost Insulators and Asbestos Workers, unpublished; R. Steinfurth, personal communication]. The data available from the union are presented in Table 111, which provides information on the cumulative entrants into the union, reduced by the number of Canadian mem bers. Also available are data on the actual union membership in recent years and the number of new entrants and separations on an annual basis. For the years prior to 1960 where such data are uncertain, the estimates of Union membership were extrapolated from the trend available in the years 1960-1980. A small correction to the union mem bership is made for the estimated number of retired members over age 65. This correc tion is a small one because the high mortality in this trade limits the number who attain age 65. The number of union construction insulation workers in Table III is increased by 40% for the years subsequent to World War II to account for workers employed on union jobs on a temporary (permit) basis and by an amount equal to the union mem bership to account for construction insulation workers not so represented. For the year 1940, few individuals would have been employed on permit because of the scarcity of jobs at that time. However, during World War II, a large number of insulators were so employed, particularly in shipyards. Data suggest that 0.2% of the wartime shipyard work force of 4,500,000 men and women were insulators. Thus 9,000 individuals would have been employed for approximately one year in this industry. Unpublished data from the Bureau of Labor Statistics estimates that 31,900 men were at work during the.Spring of 1978 as insulation workers in construction and an ad ditional 19,100 employed in industry elsewhere.1 The 31,900 estimate from Bureau of Labor data is a reasonable agreement with the 38,900 estimate using union information as described above. Short-term layoffs during 1978 could well account for at least 10% of the work force. We will use the mean of the Bureau of Labor Statistics estimate and the estimate from union data as the value for construction insulation workers. This will decrease the values in Table 111 by 8.3%. The adjusted total number of construction in sulators will then be increased by 54.4% (19,10Q/35,900)to account for insulators em ployed in maintenance elsewhere. Shipbuilding and repair. BLS data awre available on civilian production ship yard workers. The number of employees in Naval shipyards was obtained from data of the US Navy [J.K. Nunneley, Department of the Navy, personal communication, April 22, 1980]. This information is listed in Table IV. While the Navy estimates that only 50% of the yard work force is exposed to asbestos, the data on mortality and morbidity that we will use estimates risk for all shipyard workers as a group. We will utilize, there fore, the percentage of civilian yard workers that are production employees for the Naval shipyard considered to be exposed to asbestos. (This ranges from 92% in 1950 to 80% in 1975). In estimating the shipyard employment for 1945, we have used a value of 175,000, which is intermediate between 1940 employment and that of the years subse- `Based on the ratio of 1978 lotal employment reported by BLS (51,000) lo the number employed in con struction (31,948), an unpublished BLS estimate. 274 Nicholson, Purkel, and Selikoff --^~o I= -ES s2 -25 o^Oo'oOo^WooMo-roiNocooon oo n vO m --- v -- vo r i vo "0^1/7 ~^ H u- ~X z UJ ^ C8 OOOOOOOOOOO -rloo'OO- Cn h3''rvriitrO- cCh v"'iCco vO r- O' -- * w*i V| vt \0 w fi lie OJ Q. v OO--O O VO O' O O O VO co co O' ri Tf >o o r-- o vv O vr O' O V. O O ri N n m t vf v in o >0 OGOOr-r-J-'O-r'-iv^v-'O' o'O\''Odn'Od'Ovr'i'v9iv,iNvi'tTvtfNV 1 got: I'il UJ 36 EE 5V ooooooooooo ooO -- OvvYr'JOOOOO fSr*-mrNr--r-oov*, oocSO s co o ""> vr r- r- do o\ o O OOOOOOOOOOO OOO'O'oOOOOOOO -- n ro -- ooor^^OTjvo b\ -- Tt VO* u JC s Uh ^c3 X -o <. 3t s? Oo O_' 52 JZ. VV*I *s I >ov O- ceZ 03 3O ^ O v> O va r- O va cc C --- r- r~- oo < O' ffi O' O' - OX CO UJ b quent to ' would ha' estimated time [Selil yard insul Aut brake and and SIC i lions. As gories, we in auto m linear inte Rail by the As: the maint< of men cl: by 45170 t* classificat TABLE IV. Years 1940-1944 1945-1949 1950-1954 1955-1959 1960-1964 1965-1969 1970-1974 1975-1979 Source: JK TABLE V. W . .1 ,V : | . . Year 1940 1945 1950 1955 I960 1965 1970 1975' "Includes ;u bFrom Higt ' West on cst and 1,070,(1 Occupational Exposure to Asbestos 275 quent to World War II. We will consider this to be the "permanent" work, force that would have been .employed in the absence of World War II. During that conflict, it is estimated that an additional 4,325,000 men worked in shipyards for short periods of time [Selikoff and Hammond, 1978]. Their mortality and that of 9,000 wartime ship yard insulators will be estimated separately. Automobile maintenance and repair. Mechanics exposed to asbestos during brake and clutch maintenance are included in SIC 75, auto repair, services and garages and SIC 515-2, new and used car dealers, and some in SIC 554, gasoline service sta tions. As it is not possible to separate mechanics from other employees in these cate gories, we have used census data of the number of individuals employed as mechanics in auto maintenance and auto body repair. Intercensus data were developed using a linear interpolation. See Table V for the basic data utilized. Railroad steam locomotive repair. We have utilized employment data reported by the Association of American Railroads for occupations exposed to asbestos during the maintenance of steam railroad locomotives. This was done by reducing the number of men classified in equipment and stores [Association of American Railroads, annual] by 45Vo to reflect the proportion of the total craftsmen accounted for by the carmen classification. (Carmen were generally engaged in maintenance of railway cars rather TABLE IV. Estimated Population at Work in United States Naval Shipyards, 1940-1979 (in thousands}* Years. 1940-1944 1945-1949 1950-1954 1955-1959 1960-1964 1965-1969 1970-1974 1975-1979 Employed at start of quinquenium 72 335 71 112 96 SI 82 60 Estimated accessions during quinquenium 480 267 132 68 73 93 47 55 Source: JK Nunneley, United States Department of the Navy (personal communication, April 22, 1980). TABLE V. The Population Exposed to Asbestos in Automobile Maintenance and Repair Year Census* data (thousands) Motor6 vehicle registrations (millions) Interpolated population at risk 1940 1945 1950 1955 I960 1965 1970 I9751' 372 647 661 912 33 32 50 63 ' 74 92 108 133 ' 372 370 647 655 661 800 912 1,100 aJncludes aulo body repairmen. bFrom Highway Statistics (annual) US f'ederalHighway Administration. ' Weston estimated that 900,000 workers were continuously exposed to asbestos in automobile broke repair and 1,070,000 were exposed occasionally or infrequently. 276 Nicholson, Perkel, and Selikoff J than locomotives.) The remaining number was reduced by 507o to exclude employees who were located at maintenance facilities other than "back shops" [DeHague, 1980]. The balance was reduced by 1 i % to exclude salaried supervisors, coach cleaners, and stores laborers. As described previously, the resulting number for the years 1950-1960 was reduced by the percentage of steam locomotives in service. These data are listed in Table VI. A summary of the employment data for all of the previously mentioned occupa tions is given in Table VII for five-year intervals. The data are quite stable for the years 1950-1980 and well reflect both employment and its trend with time. One exception is the 1950 value for shipbuilding which is unrepresentative; for the five years, 1948 1952, employment averaged 189,000, TABLE VI. Employment* Maintenance of Equipment and Stores, Class I Railroads Year Numbers ofJ employees (in thousands) Locomotives11 in service steam diesel < Percentage Jin ihousands)_steam______ Exposed employees 1940 281 41.1 0.5 98.8 1945 387 39.7 3.0 93.0 1950 348 26.7 15.4 63.4 1955 273 6.3 26.6 19.1 I960 184 0.5 29.1 1.7 69 95 54 13 1 "Association of American Railroads, 1940-1960. Association of American Railroads, Annual, and Interstate Commerce Commission, 1961, 1958, 1 l ( TABLE VII. Employed Populations Potentially Exposed to Asbestos in Selected Occupations and Industries, 1940-1975 Industry of occupation Number employed in calendar year (in thousands) 40 45 50 55 60 65 70 75 Primary asbestos manufacturing Secondary asbestos manufacturing Insulation work* . Shipbuilding and repair Construction trades Railroad engine repair Utility services Stationary engineers and firemen Chemical plant and refinery maintenance Automobile maintenance Marine engineer room personnel (except US Navy) 23 30 17 157 426 69 44 295 113 372 34 Totals 1,880 32 35 37 35 35 32 31 60 75 75 84 93 108 114 27*' 33 41 47 53 53 55 I75v 1284 194 184 185 181 177 379 741 893 1,102 1,215 1,341 1,029 95 54 13 1 0 0 0 62 62 65 65 64 69 74 303 311 348 385 289 291 293 194 186 200 188 187 205 200 370 647 655 661 800 912 1,100 76 37 37 34 35 31 22 1,773 2,309 2,558 2,766 2,956 3,223 3,095 "Insulators are included here and not in other trades in which they were employed, such as shipbuilding, con* struction, plant maintenance, or power generation. . hDocs not include any of thc.9,000 temporary wartime insulators in the shipbuilding industry. cEstimale of "permanent'' shipyard work force. Docs not include any of the 4,525,000 temporary wartime shipyard workers. c,Unrepresentatively low value; average for 1948-1952 was 189. New Entra Dataing industri BLSdoes n refer to per hired bylht volved in tl tional dupl: were previo fore, necess each indust This \ try groups ' Security A.t nately, dat; durable goc formation individuals 1957 is give one industr group (eg, nuts, and ri in each of t tions. In a s deaths wou there wouic The correct for the gree an industry TheS hires from t increase or 1958-Janur work force tions are m tions. The > In coi tration wit! the chemict of transfers industries, SIC 29 redi Transfers a graphic are the new hir industry as much less 1; segment. mr' ees ; .b], and 950-1960 e listed in J occupathe years ception is rs, 1948- ) Exposed employees 69 95 54. 13 I 95 ) ) ind inds) 70 75 32 31 108 114 53 55 181 177 341 1.029 00 69 74 291 293 205 200 912 1,100 31. 22 223 3,095 iilding.eon- ary wariime 1 C' ` Occupational Exposure to Asbestos 277 New Entrants Into the Work Force 1940-1980 Data on the number of additions to the employment rolls in various manufactur ing industries are reported monthly by the Bureau of Labor Statistics [1979]. However, BLS does not report cumulative annual rates for "new hires.''Moreover, the BLS data refer to persons hired by individual establishments in each industry, not the number hired by the industry as a whole. There may be considerable duplication of persons in volved in the new hires reported on a monthly basis over a year's time. There is addi tional duplication involved in counting new hires in a particular establishment who were previously employed in another establishment of the same industry. It was, there fore, necessary for us to develop a measure for estimating the unduplicated new hires in each industry for each year. This was done by comparing the number of new hires obtained for major indus try groups with data available from the continuous work history sample of the Social Security Administration (SSA) for the years 1957-1960 [Galloway, 1967). Unfortu nately, data are only available for major industry groups such as durable and non durable goods manufacturing, construction, transportation, and services. Detailed in formation for individual industries is not provided. Information on the number of individuals who were employed in 1960 and were also employed in the same industry in 1957 is given in Table VIII. This allows one to calculate an annual transfer rate from one industry group to another but not from one industry to another within an industry group (eg, from the manufacture of asbestos products to the manufacture of bolts, nuts, and rivets). Bureau of Labor Statistics data on the permanent retirement or death in each of these industries are also available from Bureau of Labor Statistics publica tions. In a steady state, the SSA separation rate plus the annual rate of retirement and deaths would equal the new hire rate. As three years is a relatively short follow-up, there would be some transfers back to an industry group after the observation period. The correction for this, however, would be relatively small and somewhat compensates for the greater adjustment required to account for transfers between industries within an industry group. The SSA data are shown in Table IX and compared with the annual rate of new hires from the Bureau of Labor Statistics data for the years 1958-1960, corrected by the increase or decrease in the total work force over the three-year period of time (January 1958-January 1961). The correction consisted of attributing the annualized change in work force between 1958 and 1961 to a change in the number of new hires. Termina tions are much less affecled by work force changes and then only with severe condi tions. The corrections were virtually all less than 10%. In comparing the data obtained in this manner from the Social Security Adminis tration with that estimated using BLS new hires, fractional employment additions in the chemical industry and oil refinery operation closely matched the fractional number of transfers from the nondurable goods industry (0.166 and 0.132 vs 0.111). For these industries, we will utilize the Bureau of Labor Statistics data on new hires in SIC 28 or SIC 29 reduced by 30% to reflect possible transfers within these respective industries. Transfers are expected to occur inasmuch as the industries are concentrated within geo graphic areas and movement from one company to another is expected. This reduces the new hire rate for oil refineries to a value less than that for the nondurable goods industry as a whole. However, both oil refinery and chemical manufacturing have much less labor turnover than other industries in the nondurable goods manufacturing segment. ' Fractional Number o f New Hires Entering . Specific industry Group Each Year During m e years Occupational Exposure to Asbestos 279 '.o E .! WW) s 3U u, l/> QQ & m1 fsf vmi Om frl' t/) -- --J <Tj C 03 = * ri m dd on OO' vCnT' OO'' nn M Pl - - t On<_ 0rn'O'O np'l <<<_'<, ?Z"!":2ZZZ c E C--O O' OO 1>- -co c ta ,,< o e* (9 -- O' x> Zo o m ri --. o dd u i> 2 * f4 Or-N o o m ---- Jv*~S>* J*rw3*t* 5 i> > cIo/J o*Oo' 7 DO J oOVoI' O 3 Uii u. c j " O 3 *> C -- So * e c 3 cc 'o a c c 2 i! * ** 2 E .2 * E C T3 o ^ ' DO <3 S -2 c3c - Ec*2 aJO JO M i. oi -uO0,1_ J1Z1 *ai2co C>s-ort~.5i^ " s 2 - c c E = -2 ,1 u 2 o a 1 1 w MQ ~ 1 *a ii ss s2 0-0 <3 Ce cw = -s -- .c -wj oV --' E c_ " " C 2 '5 J3 3 TM :E a-! s 2 -S J Sla _t ta l u ^ Q. [/I J" S 3 -5 -O x -*= s- 6 .S O = U O 2 c <c <Q . . o h- 10 . .O z*o Oou. N A , n o i available. 280 Nicholson, Perkel, and Scllkoff For primary and secondary asbestos manufacturing, it would be expected that there would be less transfer; between similar companies. This occurs because of the widespread geographical distribution of the respective plants. Individuals terminated by one company would unlikely be hired by another manufacturer in the same indus try. Thus, we will adopt a value for the new hires in primary and secondary manufac turing that would be equal to 80% that of the Bureau of Labor Statistics data. It would be expected that a greater percentage of terminated shipyard employees would be re hired by other yards or by the same yard at some later date. This occurs because of the highly fluctuating nature of shipyard business, depending as it does upon large con tracts of uncertain frequency. Thus, for shipyard employees we will adopt a value of 50% of the Bureau of Labor Statistics' new hires rate for SIC 3831. The rate of new hires for 1958-1960 is, thus, 0.216. This compares with 0.138 estimated by the US Navy for naval shipyards in these years. The agreement is reasonable as turnover in govern ment shipyards is considerably less than in civilian yards. Individuals employed in construction trades (except insulation work), stationary engineers and firemen, and automobile mechanics are a highly mobile segment of the work force. However, they would tend to maintain employment in their respective trade, simply moving from one employer to another. Therefore, we feel that the Social Security Administration data on labor turnover well represent the members of these in dustries. It is felt that termination from employment in utility services, however, is less likely to lead to employment in a corresponding industry and data on new hires using Social Security Administration information would underestimate the actual per centage. Thus, we have increased the SSA new hires estimate by 50%. The sources of all new hire data are listed in Table X. For those industrial segments where the numbers of new hires are not provided, the new hires for all manufacturing are utilized adjusted by the ratio of new hires as determined by Social Security Administration data, 1958-1960, to new hires in the corresponding years for all manufacturing. ' The number of new hires for insulation workers will utilize the data on new en trants into the insulation workers' union from their membership (column 1 of Table III). We will use the same acquisition data proportionately for the nonconstruction in sulators, as data from the chemical and refining industry indicate average employment periods nearly equal to those of insulators. However, the turnover for those on permit and employed as nonunion workers is likely to be considerably higher. We have no in formation on what their turnover may be relative to union insulators but a value twice as great would, appear to be reasonable. To account for this, we will increase the IAHF1AW new hires by 0.8 to account for permit workers, by 2.0 to account for non union new hires and by 1.2 for nonconstruction insulators. Thus the total insulator new hires will be five times the IAHFIAW US new members. The 9,000 wartime shipyard insulation workers employed for one year are also included in the new hires for the 1940-1949 decade. Their mortality, however, will be calculated separately as will that of other wartime shipyard workers. It should be emphasized that these estimates are approximate and subjective. They are felt to be the best basis for estimating the number of new indi\4duals that enter a given industrial segment and are important in estimating the total number of individu als potentially exposed to asbestos. As discussed previously, however, their influence on the total mortality experience from past exposure will be small. A misestimate on the new hires rate will lead to a balancing increase or decrease in the average employment time. Tht cupation asbestos 1940.In included stationar body rep tnent dal derived I and insu Popuiati operator chanics s study is Ar double-c period si tion or it occupati cohorts 1 several f greater t TABLE X Inch Primary a Secondary Insulation Shipbuifdi Construct Railroad c Ulilily ser Stationary Chemical mainler Automob Marine en 'The pertbasis' pop 'Data are hire data years pub 'The rate retted for 'Values fi xpeeted that K |f the i ..mated same indusry manufacata. It would would be re;cause of the n large conat a value of rate of new the US Navy :r in govern ), stationary >ment of the ir respective at the Social s of these inwever, is less v hires using actual perle sources of the numbers ze^ 'justed J- jdata, t on new en i 1 of Table struction in employment se on permit ; have no ina value twice increase the unt for nonasulator new me shipyard hires for the / as will that i subjective, als that enter of individueir influence imateon the employment 1 * Occupational Exposure to Asbestos 281 time. These annual new hire rates were applied to annual employment data for each oc cupation and industry to arrive at estimates of the number of new persons exposed to asbestos on the job-in each year. The data were then cumulated for each decade since 1940. In those-industries in which a significant portion of the employees were already included in our tally under an occupational group (asbestos and insulation workers; stationary engineers, stationary firemen and power station operators; or automobile body repairers and mechanics), an adjustment was made to the 1940 industry employ ment data and new-hires data to remove duplication. These adjustment factors were derived from the BLS National Industry-Occupational Matrix in the case of Asbestos and insulation workers [Bureau of Labor Statistics, 1969b) and the 1970 Census of Population in the case of stationary engineers, stationary firemen and power station operators. No adjustment was necessary for the automobile body repairers and me chanics since the duplication between this occupation and the industries included in this study is insignificant. An additional adjustment in the new-hires data was made to eliminate the double-counting of persons who were hired in an occupation or industry during the period since 1940 and who had previously been exposed to asbestos in another occupa tion or industry. We developed an adjustment factor for this purpose by analyzing the occupational histories of 2,544 workers employed in operations exposed to asbestos in cohorts being studied by this laboratory. Table XI lists the percentage of individuals in several study groups with previous substantial exposure to asbestos (equivalent to greater than six months employment in a shipyard). This correction reduces the num- m TABLE X. Source of Annual New Hire Rales by Industry or Occupation Industry or occupation Source of annual new hire rates* Average 1958-1961 new hire rates*1 Primary asbestos manufacturing Secondary asbestos manufacturing insulation Shipbuilding and repair Construction trades Railroad engine repair Ulilily services Stationary engineers and firemen Chemical plant and refinery maintenance Automobile maintenance Marine engine room personnel 0.8 x (SIC 329) 0.8 X (SIC 343, 3443, 356, 363)* Data from union new entrants used 0.5 x (SIC 373 l)b and US Navy data 1 X (SIC 20-39) X (O.I47/0.268>1 x (SIC 20-39) x (0.099/0.268) 1.5 X (SIC 20-39) x (0.149/0.268) 1 x (SIC 20-39) x (0 J55/0.268) 0.7 X (SIC 28)b 0.7 x (SIC 29)*> 1 X (SIC 20-39) X (0.142/0.268) 1 x (SIC 20-39) x (0.099/0.268) 0.294 0.127-0.232 -- 0.216 0.147 0.099 0.149 0.155 0.116 0.092 0.142 0.099 aThe percentage of various workers within each SIC category, as described in the text, will be used as the basis population for calculating new hires. bDaia are utilized for the years available. For years for which new hire data were not published, the new hire data for all manufacturing were used, adjusted by the relationship to the specific SIC code for the years published. - 'The rate 0.268 is the average annual fraction of new hires in manufacturing for the years 1958-1960, eor- recied for changes in the work force. ^Values for olher years are proportional to (he new hire rales in ihe indicated SIC classification. ) .1 . 282 Nicholson, Perkcl, and Selikoff ber of people ever exposed by. I'QVo (the correciion factor used). It will not reduce the mortality, however, as we must account for all person-years of exposure in asbestos-re lated industries. This will be done by using the adjusted population of new entrants to calculate an average time of exposure (see below) which will overestimate the exposure time by 10% to account for the 10% reduction in exposed populations. It should be emphasized that the uncertainties in either the populations exposed or the average durations of employment greatly exceed 10% POPULATION AT RISK The results of the estimation of employment and new-hires at risk are shown in Table XII, indicating that approximately 27,500,000 individuals were potentially ex posed to asbestos from 1940 through 1979 in the occupations analyzed. The uncertain ties in estimating this number have been described previously, but they cannot be over stressed. The number is an approximation. Further, it includes a large number of individuals whose potential exposure to asbestos would have been of low intensity or of short duration because of high labor turnover (see section on lower risk population). Finally, the term potential should be emphasized. In categorizing a segment of a work force (such as all production shipyard workers) as being potentially exposed to asbes tos, some individuals will be included with no actual exposure. On the other hand, in dividuals in other jobs (such as management) who did have exposure were not counted. The numbers may or may not balance. These uncertainties will be compensated for in the estimates of mortality by using data on the mortality or morbidity of represen tative work-force segments, which will also include the full spectrum of exposure circumstances. It should also be noted that a large number of asbestos-exposed individuals are not included in the estimates of Table XII. Important groups with identified risks in-' elude family contacts of asbestos-exposed workers, engine room personnel aboard US Navy ships in World War II, and individuals exposed environmentally to asbestos by virtue of residence or work near the use of asbestos. Additional exposures occur to many from the use of asbestos in surfacing materials in schools, night clubs, and audi toriums, or as fireproofing material in office buildings. Average Duration of Employment The average duration of employment can be calculated from the fractional newhire rate adjusted by changes in total work force at different periods in time (see section on methological considerations). Alternatively, the average employment over a decade can be divided by the average yearly number of new hires entering an industry to obtain the average employment time. In essence, this is the period of time that is required for the number of new entrants into an industry to completely replace the work force. These data for the years 1940-1979 are presented in Table XIII and were used for the average durations of exposure in each decade for each industry or occupational group. Supplemental LaborTurnover Data The Environmental Sciences Laboratory has access to several seniority lists of work forces employed in asbestos-using industries. These include a large integrated as bestos products manufacturer, a major East Coast shipyard and a plastics polymer plant. Additionally, information on the employment times of all employees in an as- TABI.EXI Laboratory L Met ropolio Croton, Co Baltimore, Port Allcga Quincy, Ma TABLE XI Selected Oc Industrie Primary asl facturi Secondary; facturi Insulation Tempora Shipbuiklin Tempora Construct ic Railroad en Utility serv: Stationary i firemet Chemical p mainte Automobtl* Marine eng sonnel Tote ^Insulators construct io bestos ins parison w Labor St; the distril individua lime peric reduce the )' js-re! ^ms to ie exposure t should be the average re shown in tentially ex e uncertainnot be over number of tensity or of lopulation). it of a work ed to asbeser hand, inot counted, ensated for af represen>f exposure ) Js are \ risks inaboard US asbestos by es occur to s, and audi- ctional new ; (see section ver a decade :ry to obtain required for work force, used for the ional group. irity lists of tegrated as ics Dolymer :e: \n as- ) Occupational Exposure to Asbestos 283 TABLE XI. Workers Exposed to Asbestos in five Cohorts Under Study by the Environmental Sciences Laboratory, MiiunL.Sinai School oF Medicine Location Industry /occupation Period No. of workers currently exposed Total Also exposed in previous employment Metropolitan New York Brake repair and maintenance Grolon, Connecticut Shipyard Baltimore, Maryland Shipyard Port Allegany, Pennsylvania Asbestos products -V manufacturing Quincy, Massachusetts Shipyard 1979-1980 699 1976 1979 1979 1,024 . 286 254 1979 281 104 98 10 21 16 TABLE XII. Population at Risk to Asbestos-Associated Disease: Workers Exposed to Asbestos in Selected Occupations and Industries, 1940-1979 (in thousands) Industries or occupations Primary asbestos manu(acturing Secondary asbestos manu* facturing Insulation work* Temporary, World War 11 Shipbuilding and repair Temporary, World War 11 Construction trades Railroad engine repair Utility services Stationary engineers and firemen Chemical plant and refinery maintenance Automobile maintenance Marine engineer room per- sonncl (except US Navy) 1940 23 1940-1949 200 New entrants 1950-1959 1960-1969 103 86 1970-1979 76 Totals 488 30 324 17 35 9 157 _433 4,325 426 1.786 69 194 44 223 295 1.136 104 542 227 47 354 1,452 26 116 623 260 259 38 434 1,866 0 116 549 239 308 47 383 1,975 0 129 ~ 510 248 1,148 184 9 1,761 4,325 7,505 ' 289 628 3,113 1.393 372 1.884 34 121 1,099 46 1,282 40 1,779 27 6,416 268 Totals 1,571 11,202 4,353 4,909 5,482 27,527 insulators are included here and not in other trades in which they were employed, such as shipbuilding construction, plant maintenance, or power generation. bestos insulation production plant is available. These sources can be utilized for com parison with the data obtained from the Social Security Administration and Bureau of Labor Statistics on labor turnover. They can further be utilized to obtain estimates of the distribution of employment times in a given industry by comparing the number of individuals actually employed to those that were known to have been hired in different time periods. The latter quantity is available from the seniority lists as individuals were 4 I 284 Nicholson, Perkel, and Selikaff assigned sequential clock numbers upon employment. These data are presented in Table XIV and supplement the turnover data obtained otherwise. One notable feature is that the asbestos products manufacturer has an extremely high turnover during the first month after hire. This occurs because of terminations of individuals during a one-month probationary period. After that time, the man enters the union bargaining unit, and any individual terminations are subject to grievance procedures. While such practices are not universal, they are certainly not unique, and it is expected that in primary and secondary manufacturing an extremely high turnover will result during the first month or two of employment as individuals are screened for their performance and suitability for a job. In contrast, in construction, shipbuilding, automobile maintenance, and otner industries that require a skill, the turnover in early periods of time is expected to be less as an individual would have demonstrated profes sional competence prior to being hired. Further, he would likely be represented by a union before employment with a given employer. Thus, nonarbitratable dismissals are less common. TABLE XIII. The Average Employment Time of All Individuals Potentially Exposed lo Asbestos, 1940-1979 Average duration of employment (years) Industry or occupation 1940-1949 Primary asbestos manufacturing 1.6 Secondary asbestos manufacturing Insulation work Shipbuilding and repair Construction trades Railroad engine repair Utility services Stationary engineers and firemen Chemical plant and refinery maintenance Automobile maintenance Marine engineer room personnel (except US Navy) 1.6 2.0 13.7a 4.3" 3.3 4.4 2.8 2.7 . 3.7 2.7 4.7 "Does nol include shori-term wartime shipyard workers. 1950-1959. 3.5 3.5 12.4 5.3 8.3 7.7 5.7 6.3 7.4 6.0 7.4 1960-1969 3.8 4.0 15.9 4.2 7.5 5.7 5.8 8.7 7.7 7.8 1970-1979 4.0 3.8 Pastil 12.5 4.6 4.5 `-tTt -6.0 5.7 vlstf 8.1 7.0 6.1 IP TABLE XIV. Labor Turnover in Selected Industrial Establishments Establishment period Number of individuals considered Number employed by lime after hire l year 6 months 2 months 1 month Shipyard products Asbestos products manufacture Asbestos products manufacture Asbestos products manufacture Plastics production Insulation products 1977 1965-1966 1961-1962 1957-1958 1961-1962 1941-1945 1,449 759 306 108 17 820 - 37% 42% 27% _ 38% 73% - 100% 53% 80% 51% 45% 52% 100% 82% -- 53% 48% 75% 100% 93% m VJW*! :'hv 0 M n Sf y .. ;r/ fs' hf ti A sti data on th< facilities [N with very s month and times can b force can t one.consis 1,400 indiv Relative I To c tion, it is employee { trations th: mates can' variety of given indu for the dos sulation w for the dif relative ris primary or lung cance trations ol lence of X assume th; stance aftt inhaled by proaches I having a c; direct anc various dr For road stear mates wer These dat McDonah XVI. The Table XV the relativ mates are manufact to 0.1, ant reflect the not be coi pnied in 'i 1 unely lations of ian enters grievance lue.andit : turnover eened for tbuilding, sr in early ;d profesnted by a lissals are > Asbestos, rs) 1970-1979 4.0 3.8 12.5 6.0 5.7 8.1 7.0 6.1 r hire 1 month 53% 48% 75% 100% 93% i-- I ' ' Occupational Exposure to Asbestos 285 A study of workers exposed to brominated chemicals in three plants provides data on the distribution of employment times of all 3,579 individuals employed in the facilities [Wong, 1981). It substantiates the presence of a large number of individuals with very short employment times._Of all employees, 16.4% worked for less than one month and an additional 28.5% for 1-5.9 months. The full distribution of employment times can be characterized by a two-component decreasing exponential. Thus, the work force can be considered as made up of two groups. The average employment lime of one, consisting of approximately 2,200 individuals, was 0.5 years and of the other, with 1,400 individuals, was 11.7 years in good agreement with the data of Table XIII. Relative Risk by Industry f To calculate the asbestos-related cancer mortality in a given industry or opera tion, it is necessary to have an absolute or relative measure of exposure for the employee group. While detailed information is not available on the asbestos air concen trations that have been prevalent in previous years in each of the above industries, esti mates can be made of the relative risk of death from asbestos exposure on the basis of a variety of other studies. In the calculation of asbestos-related cancer mortality for a given industry or occupation, we will utilize the available data for insulation workers for the dose and time dependence of asbestos cancer. To translate available data for in sulation workers to other industries, it is necessary to establish measures of exposure for the different groups considered at risk relative to that of insulation workers. These relative risks for equal times of employment will be determined by three indices. The primary one is the directly measured mortality data, especially that of mesothelioma or lung cancer, in an industry or trade. A second is the directly measured average concen trations of asbestos that can be attributed to the work activity. The third is the preva lence of X-ray abnormalities after long-term employment in an industry. Here, we will assume that the percentage of X-ray abnormalities attributable to an exposure circum stance after 20 years of employment will be proportional to the total dose of asbestos inhaled by the workers in that industry. Where the percentage of abnormal X-rays ap proaches 100%, the relative risks will be determined using the percentages of X-rays having a category 2or greater abnormality on the 1LO U/C scale. Information on these direct and indirect measures is shown in Table XV along with the sources of the various data. For industries in which none of the above indices are available (construction, rail road steam engine repair) or for which the data are very uncertain, relative risk esti mates were made from the numbers of mesotheliomas identified among individuals in different asbestos exposure circumstances compared with the total work force exposed. These data utilized the nationwide survey of mesothelioma in 1972 and 1973 by McDonald and McDonald [1980]. The numbers from this series are shown in Table XVI. The relative risks, by industry, estimated from all of the above data, are listed in Table XVII. Also Indicated in Table XVII are the principal data sources considered in the relative risk estimates. The data available for the estimates are limited and the esti mates are necessarily approximate. For the years 1972-1979, the relative risks for manufacturing, insulation work, shipbuilding, and utility employment will be reduced to 0.1, and those of the other industries (except automobile maintenance) to 0.05 to re reflect the adoption of control measures. Further, exposures subsequent to 1979 will not be considered. T A B L E X V . Indices o f Relative Asbestos Exposure in Selected Occupations and Industries____________________________ abnorm alities(years) period employment pleural Industry o f o c c u p a t i o n c o n c e n t r a t i o n s lung cancer mesothelioma(years)1_+2 + abnormalities period deaths from Estimated average Relative risk Percentage o f employment parenchymal Applicable _______ Percentage o t.Applicable 286 Nicholson, Perkel, and SelikoCf + -+ + Or-i rO*t aO -- V--"> O noi pi r> ^ ioano toOo nm i/h 4- O r-1( 3 -- -- of fiber '. S ^ ^ 1 Rf c n ^ g *s r >^ufl> *r .93SWrt^ea*3no*. *,S5gps} JC pC2 3 a. - w O <1 fC <rh- -JOJ ,, -- - vi E.W ; 15 M t? o ~ 5 u a 2H - J2 ^2 (- O -- 5 *- Z oo it c/1 4* it ^$ -- il 12, . X ~ jiJt Z 77"" t> v - m jz TA (19 l Pr In; Sh Cc Ri Ui Si Cl ; A' ; , "K *( ki: * ^ .a.'/- by ^ er iii . cl -,-'ser / - ui TABLE X\ $ _____ __ Ot Primary m; Secondary g Insulation . ? Shipbuildir *.Ai Construct i< i Railroad c Utility sen Stationary Chemical Automobt Marine cn (except "Sec test f hRi.sk for t; vices ("t risks vir case-cor for the r Occupational Exposure to Asbestos TABLE XVI. The Numbers of Mesotheliomas by Work Activity in North America (1960-1972. Canada; 1972. USA)* . . .Occupation or industry Number of cases Primary ahd-secondary manufacturing Insulation work Shipbuilding and repair Construction trades Railroad engine repair Utility services Stationary engineers and firemen Chemical plant and refinery maintenance Automobile maintenance "Healing trades" " 21 27 21-49" 45-76h 5 13 + 3 l1 59c (McDonald and McDonald. 1980). aHighest number only includes some insulators and heating trades workers. bHighesl number may include some insulators, shipyard workers or individuals with employment in heating trades. 1 Includes many individuals that would be assigned to other categories, as stationary engineers and firemen (furnace repair), shipyard employment (welders, steam filters), utilities (plumbing, heating, boiler work), manufacturing (boilermakers). '287 TABLE XVII. The Risk of Asbestos Cancer Relative to Insulation Work After 25 Years Employment Occupation or industry Risk Source of data for estimate Primary manufacturing Secondary manufacturing Insulation work Shipbuilding and repair (except insulators) Construction trades* (except insulators) Railroad engine repair Utility services Stationary engineers and firemen Chemical plant and refinery maintenance Automobile maintenance Marine engine room personnel (except US Navy personnel) 1 0.5 1 0.5 0.15-0.25b 0.2 0.3 0.15 0.15 0.04 0.1 Group mortality data, exposure mea surements Exposure measurements Reference population Group mortality data, prevalence of X-ray abnormalities No. of mesothelioma cases in general population No. of mesothelioma cases in general population No. of mesothelioma cases in general population Prevalence of X-ray abnormalities Prevalence or X-ray abnormalities, group mortality data Prevalence of X-ray abnormalities, exposure measurements Prevalence of X-ray abnormalities "See test for percentage of construction population considered at risk. bRisk for years 1958-1972 when the use of sprayed asbestos fireproofing was common. The relative risks in Table XVII for insulation work, manufacturing, utility ser vices ("heating trades") shipyard employment, and construction yield "population" risks virtually identical to those found by McDonald and McDonald (1980) in their case-control analysis. They found values of 46.0, 6.1, 4.4, 2.8, and 2.6, respectively, for the relative risks of the above populations. Multiplying our equal exposure risks by 288 Nicholson, Perkel, and Selikoff the average durations of employment of all workers from 1940through 1969(13.2,2.0, 4.7, 1.9, and 6.4 years, respectively) and further dividing the risk for construction workers by two to account forihe 50% of workers to whom we attributed no risk, we obtain for the relative "population" risks the values, 13.2, 1.3, 1.4, 0.95, and 0.5. Ad justing to the McDonald and McDonald (1980] risk of 46 for insulators, weobtain for "population" risks, 46.0, 4.6, 4.9, 3.3, and 1.8. Lower Risk Population While we are unable to obtain full data on the distribution of employment times in all industries, the information depicted above allows us to identify a segment of the work force with considerably less exposure to asbestos. Taking a period of employment of two months in primary manufacturing or insulation work as a measure of a low ex posure, we have estimated the number of individuals with such an exposure among the 27,500,000 individuals identified previously. This would correspond to a total exposure of 2-3 f-yr/ml (12-18 f/ml x 1/6 yr). The estimates were made assuming 40% of the new hires in primary and secondary manufacturing and 20% of the new hires in other industries left within two months. For longer periods, we utilized an exponential function, e - ft, for the distribution of employment times where /3 is the average steadystate permanent separation rate. The period of employment characterizing "lower ex posure" for a given industry will be inversely related to the relative risk of the industry (Table XVII). These data are presented in Table XV11I and suggest that 8,700,000 of those potentially exposed to asbestos will have a significantly lower risk by virtue of their short employment period. The extremely large number in automobile mainte nance arises because of the low relative risk of asbestos disease in that industry. Thus, individuals with as much as four years of employment in automobile maintenance were included in the estimates that gave rise to Table XV111. The data in Table XVIII indicate that an enormous number of individuals are likely to have had some exposure to asbestos: 27,500,000 since 1940. Of this number, it is estimated that 21,000,000 were alive on January 1, 1980. (This figure was calculated TABLE XVIU. The Percentage of Asbestos-Exposed Individuals With Lower Exposure* Total exposed 1940 1940-1979 Number with lower exposure Percentage with lower exposure Primary asbestos manufacturing Secondary manufacturing insulation work World War 11 Shipbuilding and repair World War II Construction trades Railroad engine repair Utility services Stationary engineers and firemen Chemical plant and refinery maintenance Automobile maintenance Marine engine room personnel Totals 23 30 17 157 426 69 44 295 104 372 34 1,571 465 1,118 167 9 1,604 4,325 7,079 220 584 2,818 1,289 6,044 234 25,956 186 493 33 2 362 1,303 1.842 72 141 834 350 3,032 75 8,715 38 43 18 . 20 20 30 24 25 22 27 25 47 28 32 Lower exposure is characterized as being less than that equivalent to two months employment in an asbes tos factory or as an insulator (approximately 2-3 f-yr/ml). It is not lo be construed as being without risk. 4"' :\ using proc 18,800,000 to-havehat of asbestos ease exists lesser expo CANCER PROJECT In re made of tt new in fori ous occup: possible a; lute and rt data are le bestos-rel: propriate The Spec The seen in d: workers.' TABLE XI7 January I, I Unde Total death: Total cancel Cancer ol Pleural m Peritonea Mcsothel Cancer o Cancer o Cancer o Cancer o Cancer o Cancer o All other NoninfcctiC Asbestos All other c Number c "Expected Health Stu 'Rates are (QE) Best surgical, cl i ction iiu risk, we nd 0.5. Ade obtain for yment times ;ment of the .mployment of a low ex e among the tal exposure ing 40% of new hires in exponential :rage steadyg `lower exthe industry 1,700,000 of by virtue of bile mainleust- -Thus, y | were iividuals are is number.-it is calculated ercenlage with ower exposure 38 43 18 20 20 30 24 25 22 27 25 47 28 _ 32 it \i asbes- - jut risk. Occupational Exposure to Asbestos 289 using procedures detailed in the mortality estimates to follow.) Of those exposed, ,fe;; 18,800,000 of the total and 14,100,000 of those alive on January I, 1980 were estimated p* to have had an exposure greater than 2-3 f-yr/ml. Such exposures carry significant risk F of asbestos disease (as will be detailed subsequently). Further, some risk of asbestos disr ease exists for the 6,900,000 alive on January I, 1980, estimated to have experienced j ' lesser exposures. CANCER FROM OCCUPATIONAL ASBESTOS EXPOSURE: PROJECTIONS 1965-2030 In recent years, considerable data have accumulated that allow projections to be made of the cancer mortality associated with past exposure to asbestos. These include new information on the dose and time dependence of asbestos-related cancers in vari ous occupational circumstances, an increased awareness of the various trades in which possible asbestos exposure occurred in past years, as well as information on the abso lute and relative exposures of these different occupational groups. While the relevant data are less complete than desired, they are sufficient to allow estimates of future as bestos-related mortality to be made. These may be useful in directing priorities for ap propriate surveillance and interventive activities that might be undertaken. The Spectrum of Asbestos-Related Cancer The spectrum of malignant disease that occurs from asbestos exposure is best seen in data from the mortality study of Selikoff et aI (I979J on 17,800 insulation workers. This information is shown in Table XIX in which the numbers of deaths, by TABLE XIX. Deaths Among 17,800 Asbestos Insulation Workers in the United Slates and Canada, January I, 1967-December 31, 1976* Underlying cause of death Total deaths, all causes Total cancer, all sites Cancer of lung Pleural mesothelioma Peritoneal mesothelioma Mesothelioma, n.o.s. Cancer of esophagus Cancer of stomach Cancer of colon-rectum Cancer of larynx Cancer of pharynx, buccal Cancer of kidney All olher cancer Noninfectious pulmonary diseases, tola) Asbcstosis All other causes Expect eda 1658.9 319.7 105.6 b b b 7.1 14.2 38.1 4.7 10.1 8.1 131.8 59.0 b 1280.2 Observed (BE) (DC) 227! 995 486 63 112 0 18 22 59 11 21 19 184 212 168 1064 2271 922 429 25 24 55 18 18 58 9 16 18 252 188 78 1161 Ralio o/e (BE) (DC) 1.37 3.1 1 4.60 2.53 1.54 1.55 2.34 2.08 2.36 1.40 1.37 2.88 4.06 - 2.53 1.26 1.52 1.91 1.59 2.23 1.91 3.59 3.19 0.83 0.91 "Number of men; 17.800, man-years ol observation: 166,853. From Selikoff ei al 119791"Expected deaths are based upon white male age-specific US death rates ol the US National Center lor Health Studies, 1967-1976. bRatcs are nol available, but these have been rare causes of death in the general population. (BE) Best evidence; number of deaths categorized after review of besl available information (autopsy, surgical, clinical). (DC) Number of deaths as recorded from death certificate information only. 290 Nicholson, Perkel, ami Selikitff cause, over a ten-year period, are tabulated along with those expected From national rates. Causes of death are characterized both according to those listed on the certifi cates of death (DC) and according.tQ.the best evidence (BE) available from a review of autopsy protocols, medical, records, and pathological specimens. For most causes of death, the agreement is relatively good, but for mesothelioma and asbestosis, consider able differences exist. Because deaths from these causes are rare in the absence of asbes tos exposure, their misdiagnosis has little effect upon general population rates. How ever, as they are common causes of death among asbestos-exposed workers, their mis diagnosis can seriously affect determination of asbestos mortality. Thus, the "best evi dence" mortality will be used for the estimate of asbestos-related cancers. However, as we will attribute all excess cancer among insulators to their asbestos exposure (see be low), the overall results will not differ greatly from that using certificate of death diag nosis. Higher rates of death at one site (as mesothelioma) will be balanced by lower rates at another (as pancreas). In addition to mesothelioma and cancer of the lung, cancer of the stomach, colon, rectum, esophagus, larynx, pharyx, buccal cavity, and kidney are each elevated significantly compared with rates expected for these sites in the general population. (This group will be referred to subsequently as "asbestos-related" malignancies.) Op portunity for fiber contact with the epithelial surfaces of the lung and gastrointestinal tract is clearly evident. Exposure to the mesothelial tissue and kidney can occur as fibers readily penetrate into lung lymphatics and reach the pleural mesothelium ("pleural drift") or can be transported to the kidney or peritoneal mesothelium. Simi larly, fiber dissemination occurs to other extrapulmonary organs, such as brain, liver, spleen, etc [Langer, 1974]. While excesses at these other sites are not of statistical sig nificance for individual malignancies, the category "all other cancers" is elevated at a high level of significance (p < 0.0001), and we will attribute these excess malignancies to asbestos exposure as well. Their contribution accounts for less than 8% of the total excess cancer compared with the contribution of lung cancer, 56%; mesothelioma, 26%; and the other above specified "'asbestos-related tumors," 10%. The Time Course of Asbestos-Related Cancer The time course of asbestos-related mortality from bronchogenic carcinoma is shown in Figure I according to ages for individuals exposed initially between ages 15 and 24, and 25 through 34. As can be seen, the two curves of relative risk, according to age, rise with the same slope and are separated by approximately ten years. This sug gests that the relative risk of developing lung cancer is independent of age and of the pre-existing risk at the time of exposure. In contrast, had one plotted the added risk of cancer, the slope and the amount for the group first exposed at older ages would have been two to four times greater than for those exposed at younger ages. If one combines these data and plots them according to time from onset of exposure, the curve of Figure 2 is obtained. A linear increase with time from onset of exposure is seen for 35 to 40 years (to about the time when many insulators terminate employment). After 40 years the relative risk falls significantly, rather than remaining constant after cessation of ex posure as might be expected from the linear increase with continued exposure. The de crease is not solely the result of the elimination of smokers from the population under observation as a similar fall occurs for those individuals who were smokers in 1967. (In calculating the relative risk of lung cancer in smokers, smoking-specific data from the American Cancer Society study of one million people were utilized [Hammond, 1966].) Selection processes, such as differing exposure patterns or differing individual )rr national ; ecerti fi lm a review of lost causes of osis, consider;ence of asbesn rates. Hew ers, their mis , the "best evi. However, as iosure (see beof death diagiced by lower the stomach, each elevated il population. ;nancies.) Opastrointestinal can occur as mesothelium helium. Simits brain, liver, s'- "tical sig) jted at a Aiignancies % of the total aesothelioma. -* m. Occupational Exposure to Asbestos LUNG CANCER, INSULATORS 291 Fig. I. The ratio of observed to expected deaths from lung cancer among insulation workmen according lo age and age at onset of employment. LUNG CANCER, INSULATORS carcinoma is tween ages 15 , according to ars. This suglge and of the added risk of zs would have one combines urve of Figure n for 35 to 40 After 40 years essation of exisure. The de flation under ikers in 1967. iF <ata from 1 yimond, ) , ../dividual \ Fig. 2. The ratio of observed lo expected deaths from lung cancer among insulation workmen according lo time from onset of employment. ' 292 Nicholson, Pcrkel, and Selikotf biological susceptibilities may play a role, but the exact explanation for the effect is not understood. It is, however, a general phenomenon seen in many mortality studies. The early portion of the curve of Figure 2 is remarkable in two aspects. Firstly, it shows a linear increase in the relative risk of lung cancer according to time from onset of exposure. This suggests that the dose of asbestos received in a given period of time increases the risk of cancer by an amount that is proportional to that which existed in the absence of exposure. This increased relative risk is proportional to the dose of inhaled asbestos, which in turn is proportional to the time worked. Thus, the linear rise in Figure 2. However, the linear rise can occur only if the increased relative risk that is created by a given dose of asbestos continues to multiply the "background" risk for several decades (at least until age 60), even though the background risk will increase tenfold or twentyfold in 30 years. Secondly, the extrapolated line through the observed data points crosses the line of relative risk equal to one (that expected in an unexposed population) very close to the onset of exposure. At most, the line might be adjusted so that it passes through the relative risk of one line at a time from onset of exposure of about ten years. (Note that we are plotting the relative risk of death. Irreversible malig nancy would have been initiated several years earlier, since usually one or two years elapse between identification of lung cancer and death, and it is likely that a malignant growth was present, unseen, for at least one or two years before becoming clinically evi dent.) This means that an increased relative risk appropriate to a given exposure is achieved very shortly after the exposure takes place. However, if there is a low risk in the absence of asbestos exposure, as in young workers, cancers that will arise from that increased relative risk may not be seen for many years or even decades until the back ground risk becomes significantly greater. The same two points, I) that the effect of an exposure to asbestos is to multiply the pre-existing risk of cancer in the exposed population and 2) that the multiplied risk LUNG CANCER,UNARCO YEARS FROM ONSET OF EXPOSURE Fig. 3. The ratio of observed to expected deaths from lung cancer and the relative lung cancer mortality rales among asbestos insulation production employees according to lime from onset of employment. becomes it cancer in a tality from ginning fis years, expt significant live risk rt more, the exposure t cer for ind ing to the gory,the r Figure 3 w decade at > posure cat to the sam smokers tk mortality. In tt cancer wil begin 7.5 Figure 2 ft the averag years fron quent thre insulators tional to t TABLE XX Exposure* Years from of expost 15 25 All 5 15 25 All From Seid ( ) = Num l ] = No posure calc f,Tec' is no, ; jes;,, %. it from onset r'cl of time -h listed in ,he dose 0f le linear rjSc : r"* that is nd" risk for *'11 increase observed 1 Ufiexp0secj ndjusted so exposure of s,bl<- rriaiig, r tw0 years a maiignan, JfJica/Jy evf. exposure is 1 ,ovv risk i,, :e frm that >1 the back- ir~ ytjply / P risk ccr mortality pi yil. ) !; Occupational Exposure to Asbestos 293 becomes manifest in a relatively short lime, can also be seen in the mortality from lung cancer in a study of Seidman et al [1979]. Figure 3 depicts the time course of the mor tality from lung cancer of a group (UNARCO) exposed for short periods of time, be ginning five years after onset of exposure. As 77% were employed for less than two years, exposure largely ceased prior to the follow-up period. As can be seen, a rise to a significantly elevated relative risk occurs within ten years, and then that increased rela tive risk remains constant throughout the observation period of the study. Further more, the relative risk from a specific exposure is independent of (he age at which the exposure began. This is seen in Table XX, where the relative risk of death for lung can cer for individuals exposed for less than and greater than nine months is listed accord ing to the age at entrance into a ten-year observation period. Within a given age cate gory, the relative risk is similar in different decades of observation, as we saw before in Figure 3 with the overall data. However, the relative risk also is independent of the age decade at entry into a ten-year observation period. (See lines labelled "All" in each ex posure category.) There is some reduction in the oldest groups. This can be attributed to the same effects manifest at older ages in insulators or to relatively fewer cigarette smokers that might be present in the 50-59 year observation groups because of selective mortality. In the calculation of asbestos-related cancer, the time course of nonmesothelial cancer will be treated as follows. The increase in the relative risk of lung cancer will begin 7.5 years after onset of exposure and increase linearly, following the line of Figure 2 for the number of years a specified group is employed. After a period equal to the average duration of employment, the relative risk will remain constant until 40 years from onset of exposure, after which it will linearly decrease to one over the subse quent three decades. The magnitude of the increase will be equal to that of Figure 2 for insulators and factory employees. The rate of increase For other groups will be propor tional to their estimated exposure relative to that of insulators. The same time course TABLK XX. Relative Risk of Lung Cancer During ten Year Intervals al Different Times From Onset of Exposure* Years from onset of exposure 5 15 25 All 30-39 . 0.00 (0.351 6.85(1) - 3.71 (1) Age al start of period 40-49 Lower exposure (< 9 months) 3.75 (2) 4.27 (3) 2.73 (2) 3.52(7) 50-59 0.00 13.04) 2.91 (4) 4.03 (6) 2.58 (10) '5 15 25 All 0.00 (0.66) 19.07 (2) - JU2<2) Higher exposure (>9 months) 11.94(4) 11.45 (5) 13.13 (6) 12.32 (16) 9.93 (8) 5.62(5) 7.41 (8) 7.48 (21) From Seidman el al 11979). ( ) = Number of eases. { J s= No cases seen. Number of cases "expect ed" on ibe basis of ibe average relative risk in (he overall ex posure category. , 294 Nicholson, Perkel, and Selikaff will be used for all other nonmesothelial tumors with the magnitude of the increase in insulators being adjusted by the bbserved frequency of these tumors compared with that expected and that of otherfgroups by their estimated exposure relative to insulators as well. The treatment of the time course of mesothelioma differs from that of lung can cer and other malignancies in that there is no background rate in the absence of asbes tos exposure with which to compare the asbestos-related risk. Thus, it is necessary to utilize absolute risks of death. Figure 4 shows the risk of death of mesothelioma ac cording to age for individuals exposed first between ages 15 and 24 and between ages 25 and 34 as in Figure I. As can be seen, these data, while somewhat uncertain becauseof small numbers, roughly parallel one another by ten years as did the increased relative risk curves for lung cancer. Thus, the absolute risk of death from mesothelioma ap pears to be directly related to onset of exposure and is independent of the age at which the exposure occurs. The risk of death from mesothelioma among the insulation workers is plotted according to time from onset of exposure on the right side of Figure 4. It increases as the fourth or fifth power of time from onset of exposure For about 40 or 50 years. Thereafter, data are scanty and information on the time course is not relia ble. For the purposes of analyzing the mortality experience among various groups of workers, the relationship depicted in Figure 4 will be used. After 45 years from onset of exposure, we will consider the risk of death from mesothelioma to remain constant at 1.2 per 100 person-years for insulation workers employed for 25 or more years. For in sulators employed for shorter periods, the risk will be reduced by the fraction of 25 years worked. For other exposed groups the risks depicted in Figure 4 will be reduced by the relative exposure of the group compared with insulators and by the fraction of 25 years that a population is exposed. RISK OF DEATH FROM MESOTHELIOMA AGE YEARS FROM ONSET OF EXPOSURE Fig. 4. The death rates tor mesothelioma among insulation workmen according to age and age at onset of employment and according to time since onset of employment. Dose-Res Four arly with d and Enterl ies are not bestos and duration o portion of tween part milliliter (I based as th son and Ei ucts manu posed groi starts at y< menl). In I volved the curred. A functions > important In tl reasons fo various m basedlarg parison o: work acth response i dence of li thelioma: of the risk et al woul from mesmonths e: 2 years ex Calculat As t exposure, tive expos of indivic hires in t calculater above dai industry i employee trants int 1974 at a each quir cific rate? i pcrease in ' pith .--ators flung can:e of asbesecessary to .elioma aceen ages 25 because of ted relative elioma ap se at which insulation eof Figure )r about 40 is not relia; groups of im onset of constant at ars. For inction of 25 be reduced fraction of Z ') J agp onset of ;) W' Occupational Exposure to Asbestos 295 Dose-Response Relationships for Asbestos-Related Cancer Four recent-studies have demonstrated that the risk oflung cancer increases line arly with dose oyer a fairly wide range of exposures [Dement et al, in press; Henderson and Enterline, F979; Liddell et al, 1977; Seidman et al, 1979]. Unfortunately, the stud ies are not directly comparable. For three, the measure of dose was the exposure to as bestos and other dusts in terms of millions of particles per cubic foot (tnppcf) times the duration of exposure. This exposure categorization is highly dependent upon the pro portion of nonfibrous material in the aerosol being considered. Some relationships be tween particle counts and fiber concentrations in fibers longer than 5 micrometers per milliliter (f/ml) have been provided in the literature, but these are tenuous at best, based as they are upon a limited number of observations. Further, the study of Hender son and Enterline [1979] was limited to retirees over age 64 of a major asbestos prod ucts manufacturer in the United States. As was seen in Figure 2, observations of ex posed groups begun late in life can differ considerably from those in which follow-up starts at younger years (as, for example, at age 40-45 , 20 years after onset of employ ment). In the fourth study, that of Seidman et al [1979], exposure characterization in volved the use of data from plants other than that in which the mortality experience oc curred. A discussion of some of the differences of the slopes of the dose-response functions obtained in these studies has been made elsewhere [Nicholson, 1981a]. The important aspect is the linearity of effect with increasing amounts of asbestos inhaled. In the analysis which follows, it is not necessary that one fully understand the reasons for the differences in the slopes of dose-response relationships in mining and various manufacturing operations as the relative risks in different industries will be based largely upon the observed mortality experience in those industries or upon a com parison of the number of cases of mesothelioma or excess lung cancers in different work activities. In this subsequent comparison, however, we will utilize a linear doseresponse relationship to adjust for different periods of employment. While the evi dence oflinearity is strong for lung cancer, we will assume that it also obtains for meso thelioma and other malignancies. The evidence for this is more limited, but an analysis of the risks of mesothelioma according to time of employment in the study of Seidman et a) would suggest that it is true for that tumor as well. For example, 0 of 215 deaths from mesothelioma occurred from less than 6 months exposure, 3 of 82 from 6 to 11 months exposure, 4 of 74 from I to 2 years exposure, and 7 of 63 from more than 2 years exposure. Calculation of Asbestos-Related Mortality As discussed previously, for those trades in which workers have possible asbestos exposure, estimates were madeof the number of employees potentially at risk, the rela tive exposure of those workers compared with insulators, the average employment time of individuals entering a particular trade or industry, and the age distribution of new hires in the various trades of industries. The asbestos-related cancer mortality was calculated as follows. For those employees entering a trade subsequent to 1940, the above data from Table XII were utilized to obtain the number of new entrants into an industry during different periods of time. The age distribution of new manufacturing employees of 1960 (Table XXI) was used to calculate age-related mortality of new en trants into a trade or industry. This distribution also was found in new hires during 1974 at a major northeast US shipyard (E. Christian, personal communication). For each quinquennium at entry, the appropriate age, calendar year, and asbestos risk spe cific rates were applied to calculate the excess lung and other cancer mortality, the risk "Vi- 296 Nicholson, Perkel, and Selikoff of death from mesothelioma, the total mortality (based on US national rates for the entry quinquennium and all subsequent quenquennia until the year 2030 (assuming 1975-1979 rates to apply to the year 2030). This was done for each five-year period of entry, 1940-1980, and the calculated numbers summed for each calendar quinquen nium, 1940-2030. For those employed in 1940, the appropriate age distribution for an industry or trade in 1940, as given by the US census, was used. For those employed in 1940, it was assumed that onset of asbestos exposure occurred at age 22.5 or 1930 for those 32.5 years or older in 1940. The excess, nonmesothelial cancer mortality was calculated using the time de pendence displayed in Figure 2 with the assumption that the manifestation of risk from a given exposure will first take place 7.5 years after its occurrence and increases line arly until 7.5 years after cessation of exposure. The risks of death from mesothelioma were calculated using the data of Figure 4, adjusted for each industrial group, with risk assumed to be constant after 45 years from first exposure. Account was taken of the different periods of exposure for each group in each decade, as indicated in Table XIII. Calculations were made using US white male rates. Some blacks and some women would have been employed in the industries under consideration, although their num bers would have been small. Were data available on the number of blacks and women, the use of black male rates would have increased the number of nonmesothelioma can cers and the female rates would have decreased the number, resulting in only a small change from these data. The results of such calculations are shown in Table XXII through XXV, which list the average annual excess number of lung cancers, mesotheliomas, gastrointestinal, and other asbestos-related cancers, and total excess cancer attributable to asbestos ex posure in each quinquennium from 1965 to 2030 for the populations in Table XII. In these tables the average annual mortality in each quinquennium is listed by the mid year of the period. As can be seen, the dominant contributors to the asbestos-related disease are the shipbuilding and construction industries. Industries directly involved in the manufacturing of asbestos products or with the application of insulation material contribute a significantly smaller proportion to current asbestos disease and that to be expected for the next two decades.2 It is instructive to look at a display of the number of mesotheliomas and asbestosrelated cancers in the shipbuilding industry from the year 1940 to the year 2000. While the total number of malignancies are necessarily uncertain, the data on the time course of the cancers that will occur are relatively good. These data are shown in Figures 5 and 6 for the populations first employed prior to 1940, during World War II, and subse quent to 1945. As can be seen, the relative importance of the wartime and postwar ex posures are roughly equal, even though a considerably greater number of individuals were employed in World War II. This, of course, occurs because of the relatively short periods of work for the wartime group. Further, while the exposures in the construc tion industry are more uncertain, the important disease experience is also ahead of us in 'A preliminary report on this research has been presented elsewhere (W.J. Nicholson, G. Perkel, I. J. Selikoff, and H. Seidman. Cancer from occupational asbestos exposure: Projections 1980-2000. Banbury Re port 9, Cold Spring Harbor Laboratory, 1981, pp 87-111). In that publication, an estimate was presented of the population at risk from asbestos exposure since 1990 (13,200,000) and projections of asbestos-re lated mortality (8,770 deaths in 1982 to 9,750 in 1990). The estimates of the population exposed to asbestos presented Ihere, however, did not fully account for the extremely high lurnover in workplace employment that, we have discussed here. However, as (he mortality estimates did not depend on the total population ex posed, they are virtually identical lo those presented here. that industi terials betw be seen in F to the year 1940 and 2 The r greater thai program fc communici mesothelio TABLE XXI 1965* 1; Age 18-19 20-24 25-34 35-44 45-54 55-64 65 + *Data from 2 Based on 47 Shipbuilding Pig. 5. Thi poxiuc froft curlier, itaos 4 j for the I )ng iod of uinquenon for an ployed in 1930 for lime de risk. from ases line- ithelioma with risk en of the ible XIII. e women leir num1 women. oma cany a small V, which ntestinal, ties* "s ex- ,/ l rs-related volved in i material that to be asbestos00. White ne course arcs 5 and nd subsetstwar extdividuals vely short construeid of us in (el.l.J.ScIiBanbury Re as presented asbestos-red to asbestos employment >f; )nex- tr Occupational Exposure tn Asbestos 297 that industry, largelybecause of the extensive use of asbestos in spray fireproofing ma terials between 1958 and 1972. A measure of the overall future disease experience can be seen in Figure 7,.which depicls the projected annual mesothelioma deaths from 1940 to the year 2000. ,Qf all mesotheliomas that are estimated to occur between the years 1940 and 2000, about one third have occurred to date. The number of mesotheliomas estimated by this procedure is approximately 40% greater than those that would be estimated to occur nationwide using data of the SEER program for white males during 1978 [R. Connelly, National Cancer Institute, personal communication, 1981]. Here, initial data (with one center not analyzed) report 98 mesothelioma deaths in nine of the ten SEER areas. As they represent approximately a v .; i'v"'- ` 4 ,: j :.j .? ' { i % `3 -;j 1 ! TABLK XXI. Age Distribution of Employees Hired l)i ring 1965 Who Were Not Working January 1, 1965* Percent, of Number Percent in shipyard workers Age (in thousands) uge interval in age interval3 18-19 20-24 25-34 35-44 45-54 55-64 65 + 892 1,614 1,431 861 588 361 146 15.1 27.3 24.3 14.6 10.0 6.1 2.5 17.8 31.6 27.6 12.0 6.1 2.9 0.0 *Daia from Bureau of Labor Statistics 11965]. aBased on 478 new hires during 1974. Data from Christian, Sec. Local 5, Industrial Union of Marine and Shipbuilding Workersof America (persona! communication. 1981). . SHIPYARD CANCER OTHER THAN MESOTHELIOMA TEAR Fig. 5. The eslimalcd and projected numbers of mesothelioma deaths per annum from past asbestos ex posure Irom 1940 through 1999 among three groups of shipyard employees (Ihose employed in 1940 or earlier, those employed during World War II, and those employed subsequent to World War II). -* . 298 Nicholson, Perkel, and Selikoff 10*70 sample of the US population., the national estimate of cases for 1978 would exceed 1,000. This is to be compared with our estimate of 1,400 for the quinquennium 1976-1980 (and for the year 1978). In this comparison, however, it should be noted that the information used for the estimate of asbestos-related cancers in this work relied upon data that identified asbestos malignancy following analysis of all medical evidence and after a review of all pathological material available. The SEER program, on the other hand, used records-based reports with no review of pathological material. Experience has shown that pathological review will identify as mesothelioma many neoplasms initially categorized otherwise [Levine, 1978]. Further, while well represent ing the shipbuilding industry, the ten-SEER areas underrepresent industrial areas and IOOO SHIPYARD MESOTHELIOMA eoo 2ZX> 5 600 CE bJ a. Xbh<aJ- *00 200 ) 0 _ 1940 I960 1980 2000 YEAR Fig. <>. The estimated and projected numbers or excess asbestos-related cancers per annum from 1940 through 1999 among three groups of shipyard employees (those employed in 1940 or earlier, those em ployed during World War 11, and (hose employed subsequent to World War II). ASBESTOS MESOTHELIOMA IN WORKERS EMPLOYED 1940 AND LATER YEAR Fig. 7. The estimated and projected numbers of mesotheliomas per annum I'rom 1940 through 1999 Irom occupational asbestos exposure. I * I I'- \ed ; Jm oted that >rk relied medical program, material, na many epresentrreas and )) from 1940 , those cm* K'r- Occupational Exposure to Asbestos 299 00'>OQ*TO'Or1 n vO - VO CO vO M O' -- -- n co ri ri oo w - f't v/% \o ri N O -- <N -- -- 01 N r 90m - 5v in v CO <r OO -- o cn O oo -- c-i p-i vn v^. -- m~ rfrnm'C'COOOOO'S- M vf rp CO N --'/-'ir'lC'4 rs on rn r- oo -- <r n r- r* m so oo \o 'OhOvlNO'n'OCO <N rn m CO -- O' r~- -- n n ij- vO V - o VO r- r>c r- r~ CO ~ rn NT Nrf>N--o O--n' --v-rmOO' oMorV-sOomoOOn' ONv^vircji-- vOm oo ch m r-r -- r-~ -- O -* -- r* *-- r- --m -- -- <N OvvOoohvnyjMTfvo (N^vnxftvooonoo r4 -- coi*-c oq 2o Ov. `w u 2 -a CL 5E o :c .5S2 _co .*50 o^ c Jpc- E >wl_. -3cQ > .9- 5? -- 23 cO 2 4Sc1 i TJ * : C* dU eScc ^ _ c 'rv C I c<3 .9 e2 po iu> ? -2 E 2 g .s .81 i% wCSL oOC,oc--ttf:J' 300 Nicholson, Perkel, and Selikoff metropolitan regions that would have had significant construction activities 30 or more years ago..Thus, it is not unexpected that actual US rates may exceed those estimated from the SEER program. ; There is observational evidence to support the analytical approach used in these calculations. The data for insulation workers suggest that 650 mesotheliomas and 2,300 excess lung cancers would occur between 1967 and 1976 among members of this craft. This is to be compared with 175 mesotheliomas and 380 excess lung cancers seen among insulators in the single union (The International Association of Heat and Frost Insula tors and Asbestos Workers, AFL-CIO) studied by Selikoff et al [1979). The ratios of 0.27 and 0.17 for the number of deaths among Asbestos Workers Union members to those calculated here is in reasonable agreement with the fraction of all insulators that the union has organized (0.29). The difference in lung cancer and mesothelioma ratios can be attributed to the fact that the insulators organized by this union are older than the entire group estimated to be at risk from 1967 through 1976 and, thus, have a pro- portionally greater risk of death from mesothelioma than from lung cancer compared to other insulators. Forty-two percent of the Asbestos Workers Union members were 45 years of age or older at the midpoint of the Selikoff et al study. A comparison of the ratios of the calculated 1977 mesothelioma deaths from industries (Table XXIII) with those observed in the study of McDonald and McDonald (1980) (Table XVI) also shows reasonable agreement. As discussed previously, one third of those estimated to have had a potential ex posure to asbestos were exposed for only a short period of time and were believed to have a risk less than that equivalent to that from employment in an asbestos products plant or as an insulator for two months. By calculating the person-years of exposure of the `lower risk population" and comparing the result to the total person-years of em- ployment in each industry the contribution of the lower-risk group to the estimated excess mortality can be obtained. These results are shown in Table XXVI and indicate that 32% of the exposed group will contribute less than 2% of the excess asbestos-re- lated deaths. The numbers are approximate because of uncertainties in the assumed short-term separating rate. They do, however, dramatize the consequences of inclusion of lower exposed individuals in the population at risk. , A V? ;y; , ;,! :7 . y- *. '?>' V , ^ vJ vV '!_-y i* C. Asbestosis Deaths The above estimates are of deaths from malignancy. There will be additional deaths from asbestosis that will occur in individuals exposed to high concentrations over long periods of time. In contrast to the asbestos cancers, deaths from asbestosis generally require considerable fiber exposure. They will largely occur in insulators, manufacturing workers and long-term shipyard employees. They will be fewer than the number of mesothelioma deaths among insulators (perhaps one half to three fourths). Because of the high labor turnover in manufacturing we would estimate that about one third as many deaths will occur from asbestosis as from mesothelioma. A similar ratio is probably appropriate for pre- and post-World War II shipyard workers (short-term wartime work would carry only a limited risk of death from asbestosis). Thus, approxi mately 200 deaths annually are now occurring from asbestosis (the condition, however, will be contributory to many more deaths). This number will perhaps double during the next two decades and decline thereafter. y\ f ; .; : t .7 M }0or more . ' jated sed in these is and 2,300 if this craft, seen among 'rost Insulahe ratios of members to ulators that ioma ratios : older than have a pror compared mbers were rison of the KX1II) with : XVI) also >otential exbelieved to os products exposure of }d exii.^4 indicate asbestos-rehe assumed of inclusion e additional icentrations ti asbestosis t insulators, wer than the ree fourths). at about one similar ratio (short-term us, approxiin, however, le during the ; ;" ; ' , . . Occupational Exposure to Asbestos 301 O O oo o -- r--i rm~ mi Osmvc -- rroor~ -- v. n oo o r-i v -- r* -- r>i -- O' I-- ~rf Oi -- >0 CC sO I-- Vi -t - r--t Ort 'i-nrr---- -- O'o-rt ^--t -- m O' O' Vi O O'' r'l ot nC--f'lwsOmO''v> -- r-i r*i vc O r -- O'OOA'riS'0t'-O"r'fnO'0'i0n's0t' -- ~ r i r- O' rN -- 0'T\ 'rO~ Or-''OoC'Ob'Ooro~omoo moc -- -- M oo vo r^i -- crosmr'~oc0''0r~-0 -- -- -- -- oo r'* -- p0>ONvnON00 -- o-- O' --m> -- mv* 'O '*O-- O' OKOr-a--'sNvsjv-rsao'oijtM--j-rO. <N 'O NO rf oTof r-r-i" IN v> rr-- V, rN 'vOi-rxJr-v--x^r''iOCTm'omcrr-~t0T'*C- fN| -- --- 3e 'Z= I e I. Z s .11 ^ is -o o 3 c. L. 4> fl. LO HO 302 Nicholson, Perkel, and Selikoff Comparison With Other Studies Some previous estimates csf asbestos-related mortality exceed those discussed here. In the Department of Health, Education, and Welfare estimate that I3%-18'7o of all cancers in the near future will be asbestos-related, recognition was taken that a large number of individuals were potentially exposed to asbestos, their estimate being 8-11 million compared with ours at 27.5 million, 18.8 million of whom had exposures greater than 2-3 f-yr/ml [Department of Health, Education, and Welfare, 1981). How ever, their estimates of the number of heavily exposed individuals was subjective and no explicit adjustment was made for the different employment periods of exposed groups. The estimates by Hogan and Hoel [1981) that up to 12,000 deaths may occur annually from asbestos cancer placed great emphasis upon possible effects from the shipbuilding industry. They, too, subjectively estimated the number of heavily exposed individuals in this trade and did not explicitly account for variations in employment time and may have overestimated the asbestos-related mortality. However, their esti mates of the effect of other industries neglected large numbers of individuals with po tential exposure. Thus, their estimates for other than shipbuilding would appear to understate the asbestos disease potential [Nicholson, 1981b). Finally, Blot and Fraumeni [1981] estimate that 120,000 lung cancer deaths will occur (over the popula tion lifetime) from wartime shipyard employment. Our estimate is 25,000. The differ ence lies largely in our assigning a much lower risk to the very short term (< 1 year) employees. A lower estimate of 4,000 asbestos cancers annually has been made by Higginson et al [1980] based upon mid-i970 SEER data for mesothelioma and a multiplier of three for other cancers. However, the multiplier depends on time from onset of expo sure and population age and exceeded four during the 1970s. (Compare Tables XXII and XXIll.) Further, the previously mentioned limitations of the SEER data apply here. Enterline has also estimated that approximately 4,000 deaths will occur annually [Enterline, 1981]. Fie attributes 530 lung cancer deaths/yr to primary manufacturing and insulation work, 900to secondary, 421 to shipyard employment, 212to auto main tenance, and 438 for other occupations. In addition to lung cancer, he estimates 1,250 other cancers and 333 mesotheliomas will be asbestos-related. The values for primary manufacturing, insulation work, and auto maintenance are similar to our estimates and that for secondary manufacturing considerably more. However, much lower esti mates are given'for shipbuilding, construction, and other trades. This is in contrast with the finding that a much greater number of mesotheliomas occur in these trades compared with manufacturing and insulation work [McDonald and McDonald, 1980]. Expected Mortality in Asbestos-Exposed Workers Tables XXII through XXV list the projections for the excess mortality associated with past asbestos exposures. For a given work category, these excess deaths will add to those expected in the absence of exposure but, with the exception of mesothelioma, an "excess" death cannot be distinguished from an "expected"one. As each of these deaths may lead to a claim for compensation or a third party suit, the potential of such casescan greatly exceed the number of excess deaths calculated above. For the heavily ex posed (insulators, for example), where the excess deaths exceed those expected, the problem is not a great one. However, for groups with lesser exposure, the total number of lung cancer deaths that could be asbestos-related is very much greater than the numbers in Table XXii. Table XXVII lists the expected lung cancer deaths over the I issed 3'70-18*% of that a large ; being 8-11 d exposures 1981). Howojective and of exposed > may occur :ts from the vily exposed :mployment r, their estials with po d appear to , Blot and the populaThe differi (< 1 year) y Higginson aultiplier of get- - expo } ]<X1I \ .-apply :ur annually nufacturing > auto mainmates 1,250 for-primary jr estimates a lower esti- in contrast these trades nald, 1980]. IKS'- y associated s will add to helioma, an these deaths if such cases : heavily ex pected, the otal number ter ' an the tli the ( Occupational Exposure to Asbestos 303 r-t r-- O' ci r J o O o r- v-i -- c?'--' \C O' O' --M* O'T' -- -M'3't0'000\ r- o O r- --* -- >o --. -- rs -- DO Ot *--T oO' O' On o*or .---- r*l < rN f" tT r- rj-O'/n E Vi IcN* Ol n O' V) ~ CD r~i 2* i B 3 c if c v- % c, w _ U) -- " o 5 1 - 2 15 6o I |. r c ~-- O (1 - " 3 (/) U tti ID co U < 304 Nicholson, l'erkel, and Sclikciff O' O' O O' O' rM -- SO r | nO Vi r-- oo n ro -- m o ^ i "T m v. r- o t o o -r n -- -r r"l -- r--> -- r-r^OoO'Cr'ir'irT'n- Bj 'C r- ri O' cr--ooOoo'f''--o0r"--r0-c0'i 0-- f>l t Ci 5> >C fsj r*"i )) O' M O' O ^, O O' r-- xj- <a o> Vi s so > *-, O' N r'lr-OO -- r-- r-- cm \o m rj-0k00c^ir~0'Cr^ O'~o -- O'^ro'--v' 't'or-r-^-Noi't V '5f --> rvT cs rs r~ vr>rnr^r'--*r40*-- roooO'm,tf-'rornm r^-!f'^rnv'vr-r~0' _OI-- sOr^Tj-ccr'lv r>''Or-iooo'0.'T n n vCia h M ^1 i i " S Sto ** c E S a# C '5 y u-- C. - M >* o ,5i ~ a cco' -ft 15 5 ' !2 B ' E3 a! U Qi Z) W O < 1 M- years 1965pected nun with lower range from Figur and 2030 a exposure is one year), indicated. < doubling e> is about 60 the curve c exposed in> lung cance: lion of me* As m distinguish exposure l> have abnoi normal X-i fig. 8. T1 insulalor-y ploymenl a * Occupational Exposure to Asbestos 305 years 1965-2030 (assuming 1978 rates Tor subsequent years). A.s can be seen, the ex pected numbers exceed the excess by nearly six times. Even if the 32% of individuals with lower exposure are excluded from consideration, the ratios of expected to excess range from 0.4 to.1.1.7-. Figure S shows the distribution of excess lung cancers expected between 1980 and 2030 according to equivalent insulator-years of exposure. (An insulator-year of exposure is that which would create the same risk as employment as an insulator for one year). The approximate exposure for a doubling of lung cancer risk is also indicated. Of the excess lung cancers, 50% occur in individuals with more than this doubling exposure. The total number of lung cancers is also shown for this group and is about 60% more than the excess due to asbestos exposure. For lesser exposures, the curve of the total cancer rises extremely steeply because of the large number of exposed individuals. At the peak of the asbestos related lung cancer curve, the total lung cancer curve would be four times higher. Parenthetically, the exposure distribu tion of mesothelioma cases will be similar to that of the excess lung cancers. As mentioned previously at a given exposure level an "excess" death cannot be distinguished from an "expected" one. The problem, however, extends even across exposure levels. Many individuals with less than 5 insulator-years of exposure will have abnormal X-rays, and a significant percentage with greater exposure will have normal X-rays. This follows from the finding that more than 30% family contacts of sc to <r UJ > < _j UJ cc Q: UJ c_> z < o o z 23 EQUIVALENT INSULATOR, YEARS OF EXPOSURE Fig. 8. The distribution of excess lung cancers expected between IV80 and 2030 according to equivalent insulator-years ol exposure. (An insulator-year ol exposure is thai which would create the same risk as em ployment as an insulalor for one year.) > { .I 306 Nicholson, Perkel, and.Seljkoff ' ` TABLE XXVI. Percentage.of Asbejilos-ftclaled Cancers Thai Occur Among Those Wi(h Lower Exposure Who. Were Exposed After January 1940* Industry or occupation Percentage ofdeaths Primary a.sbesios manufacturing Secondary manufacturing insulation work Shipbuilding and repair Construction trades Railroad engine repair Utility services Stationary engineers and OreiDen Chemical plant and refinery maintenance Automobile maintenance Marine engine room personnel 1.2 1.3 0.1 1.9 1.0 1.6 0.8 1.8 1.0 12.4 2.3 Lower exposure is considered to be less than 2-3 t-yr/ml. The overall contribution to mortality of all individuals with lower exposure is l.9A. asbestos factory workers (Anderson et al, 1979) and insulators (Nicholson et al. to be published) have asbestos related X-ray abnormalities (20-30 years after onset of less than 5 equivalent years of exposure) and that a fair number of insulators with 20 or more years in the trade have normal X-rays. Pulmonary function tests are even less revealing. While procedures based on exposure or on clinical evidence of exposure are possible, the allocation of compensation resources to the deserving individuals is clearty an enormously difficult scientific problem. It is an even more difficult social problem. CONCLUSIONS Estimates have been made of the numbers of cancers that are projected to result from past exposures to asbestos in a number of occupations and industries. Only those potentially exposed by virtue of their employment have been considered. Additional deaths will result from exposure among family contacts (household contamination), from environmental exposures, from exposure during consumer use of asbestos prod ucts, and from exposure while in the Armed Forces, particularly in engine rooms of naval ships. No estimates have been made of deaths resulting from asbestosis. These estimates indicate that: 1. From 1940 through 1979, 27,500,000 individuals had potential asbestos expo sure al work. Of these, 18,800,000 had exposure in excess of that equivalent to two months employment in primary manufacturing or as an insulator (>2-3 f-yr/ml). 21,000,000 of the 27,500,000 and 14,100,000 of the 18,800,000 are estimated to have been alive on January I, 1980. 2. Approximately 8,200 asbestos-related cancer deaths are currently occurring annually. This will rise to about 9,700 annually by the year 2000. 3. Thereafter, the mortality rate from past exposure will decrease but still remain substantial for another three decades. Occupational Exposure *o Asbestos 307 Yors'l OC-l r4 oO' rvOO*x w JO z3 oOI--o' e .2 D. U O <ou O' u** x nr*J e O' Qa* vO O' oc* Oc o 5a3ouo o >, > X X U4 Tt3Oo c H 308 Nicholson, Perkel, and Selikoff These projections are from past exposures to asbestos. Over one million tons of friable asbestos material are in place in buildings* ships, factories, refineries; power plants, and other facilities. The maintenance, repair and eventual demolition of these facilities provide opportunities for continued significant exposures. If such work is not DeHague GR: Dement JM, H chrysoli Wales, Dept Health, I properly done, or if asbestos is otherwise used with inadequate controls, the burden of occupai disease and death from past exposures will be increased by the environmental exposures of the future. Harbor Department o Electric and C Enterline PE: ACKNOWLEDGMENTS Labor: Felton JS: Rc This work has benefited from the critical review and suggestions of Thomas C. 341-3: Brown, Fred Siskind, and Howard Vincent of the Department of Labor. Support for this research was provided by Contract J-9-M-8-0165 of the Department of Labor and Center Grant ES00928 of the National Institute of Environmental Health Sciences. . * Ferris Jr. BG chroni Fischbein A, Assoc . * FlcisherWE, REFERENCES fix :; naval Fontaine JH, Anderson HA. l.ilis R. Daum SM, Selikoff IJ: Asbestosis among family coni aus of asbestos factory work i26-i: ers. Ann NY Acad Sci 330:387-399. 1979. Asbestos Information Association: Comments submitted to the Environmental Protection Agency on the Galloway LE Educa Ambient Water Quality Criteria for Asbestos. 1979. iv,: Hammond E Association of American Railroads. Statistics of railroads of Class 1 in the United Stales, annual, 1940 Study 1960. V" Hants NM, S Balzer JL, Cooper WL: The work environment of insulating workers. Am Ind Hyg Assoc J 29:222-227, ?3V.' Harries PG: Henderson V .) (968. Blot WJ, Harrington JM, Toledo A, Hoover R, Heath Jr, CW, Fraumeni Jr, JF: Lung cancer after em ployment in shipyards during World War 11. N Engl J"Med 229:620, 1978. t% ease n Higginson J, Blot WJ, Fraumeni Jr, JF: Cancer among shipyard workers. Banbury Report #9. Cold Spring Harbor i\ portic Laboratory, 1981, pp 37-49. :v v* HickishDE, Hill Jr. AB: Bonnell J A, Bowker JR, Browne RC, Erskine JF, Fernandez RHP, Massey MO: A review of control of as bestos processes in the Central Electricity Generating Board. XVIII International Congress oaOc cupational Health, Brighton, England, 1975. Bureau of the Census: Census of Manufacturers, Washington, DC, 1939. Bureau of Labor Statistics: "Employment and Earnings, United States, 1909-1978." BLS Bulletin 1312 r. ,,f 4$ Hirsch A, D time HoganMD, 76, '.?*' Internationa 11,1979. . r* Stein Bureau of Labor Statistics: "Handbook of Methods for Surveys and Studies." US Department of Labor, Interstate O 1976. listia Bureau of Labor Statistics: "Occupational Employment." Report 430, Series, 1,9,10,16,29,55,56,61.81, 1974. ' Bureau of Labor Statistics: Occupational mobility of employed workers. Special Labor Force Report 84, V Jones H: lnt ' Jones RN: S lion. 1965. 1980 Bureau of Labor Statistics: Occupalional employment in the construction industry, June 1978 (unpub Kaminski R lished). . 189, Bureau of Labor Statistics: "Tomorrow's Manpower Needs." App. A., US Deparlment of Labor, vol t, Langer AM 1969a. Heal Bureau of Labor Statistics: "Tomorrow's Manpower Needs." US Department of Labor, vol 4, 1969b. Lawton Glv Bureau of Labor Statistics: Wartime employment, production, and conditions of work in shipyards. BLS j Boar Bulletin 824. US Government Printing Office. Washington, DC, 1945. Levine RJ ( Christian E: Sec. Local 5, Industrial Union of Marine and Shipbuilding Workers of America. Personal i Lewinsohn communication, 1981. Ann Connelly, R: National Cancer Institute. Personal Communication, 1981. Liddell FD1 Daly AR, Zupko A J, Hebb JL: Technological feasibility and economic impact of OSHA proposed re mini vision to the asbestos standard. Roy F. Weston Environmental Consultants-Designers, prepared | Licben J. F for Asbestos Information Associalion/North America, Washington, DC, 1976, vol 2. lijff }r"~ 'ons of i tower .ion of these i work is not ie burden of al exposures Thomas C. Support for ' Labor and ciences. factory work^gencyonthe annual, 1940J 29:222-227, y jpr em- 0 harbor control of a.sngress on Oc- lufletin 13(2ent of Labor, ,55,56,61,81, ce Report 84, 1978 (unpubLabor, vol I, I 4, 1969b. ipyards. BLS iea. Persons! proposed re:rs ~*epared 2.: ^ Occupational Exposure to Asbestos 309 DeHague OR: International Association of Machinists. Personal communication, November )4, 1980. Demem JM, Harris Jr," RL, Symons MD, Shy C: Estimates oT dose-response for respiratory cancer among chrysolile asbestos textile workers. Presented at Fifth Jm Conf on Inhaled Particles, Cardiff, Wales, Sept, 1980.(in press). - Dept Health, Education, and Welfare: Estimates of the fraction of cancer in the United Slates related to occupational factors. Sept. 15. 1978 (unpublished). Reprinted in Banbury Report #9, Cold Spring Harbor Laboratoiy, 1981, pp 701-726. Department of Labor* Industrial Work Survey: Electric and Gas Utilities, February 1978, p. 60, 1979. Electric and Gas Utilities: Industry Wage Survey. February 1978, BLS, 1979. Enierline PE: Proportion of cancer due to exposure to asbestos. Banbury Report 09, CoId Spring Harbor Laboratory, 1981, pp 19-36. Felton iS: Radiographic search lor asbestos-related disease in a naval shipyard. Ann NY Acad Sci 330: 341-352, 1979. Ferris Jr, BG, Ranadim MV, Peters JM, Murphy RLH, Burgess WA, Pendergrass HP: Prevalence of chronic respiratory disease. Arch Environ Health 23:220-225, 1971. Fischbein A, Rohl AN, Langer AM, Selikoff IJ: Drywall construction and asbestos exposure. Am Ind Hyg Assoc J 40:402-407, 1979. Flcisher WE, Viles FJ, Glade RJ, Drinker P: A health survey of pipe covering operations in construction naval vessels. J.lnd Hyg Tox 28:9, 1946- Fontaine JH, Trayer DM: Asbestos control in steam-electric generating plants. Am Ind Hyg Assoc J 36: 126-130, 1975. Galloway LE: Interindustry Labor Mobility in the United States 1957 to I960. US Department of Health, Education and Welfare, 1967. Hammond EC: Smoking in relation to death rates of one million men and women. In: "Epidemiological Study of Cancer and Other Chronic Diseases." Natl Cancer Inst Monogr 19:172, 1966. Hants NM, Stavraky KM, Fowler JL: Cancer mortality in oil refining workers. J Occup Med 21:261, 1979. Harries PG: Asbestos hazards in naval dockyards. Ann Occup Hyg 11 :!36, 1968. Henderson VI, Enierline PE: Asbestos exposure: Factors associated with excess cancer and respiratory dis ease mortality. Ann NY Acad Sci 330:117, 1979. Higginson J, Bahar JC, Clemmescn H, Demapouios H, Garfinkel L. Hirayama T, Schottenfeld D; Pro portion of cancers due to occupations. Prev Med 9:180, 1980. Hickish DE, Knight KL: Exposure to asbestosduringbrakemaintenance. Ann Occup Hyg 13:17-21,1970. Hill Jr, AB: International Union of Operating Engineers. Personal communication, April 29, 1980. Hirsch A, DiMenza L, Carre A, Harf A, Pedriget S, Cooreman J, Bignon J: Asbestos risk among full time workers in an electricity-generating power station. Ann NY Acad Sci 330:137-145, 1979. Hogan MD, Hoel DG: Estimated risk associated with occupational asbestos exposure. Risk Analysis 1:67 76, 1981. International Association of Heat and Frost Insulators and Asbestos Workers. Unpublished data (R. Steinfurih, personal communication). Interstate Commerce Commission: Interstate Commerce Transportation Statistics in the U.S. Part l. Sta tistical Abstract. Annual. . Jones H: Internationa) Union of Operating Engineers. Personal communication, March 3, 1980. Jones F.N: Shipboard asbestos exposure: Hearings before the Subcommittee on Coast Guard and Naviga tion. US House Committee on Merchant Marine and Fisheries, Serial No. 96-41. Washington, DC, 1980, pp 229-232. Kaminski R, Geissert KS, Dacey E: Mortality analysis of plumbers and pipefitters. J Occup Med 22:183 189, 1980. Langer AM: Inorganic panicles in human tissues and their association with neoplastic disease. Environ Health Perspect 9:229, 1974. Lawton GM: Paper on the US Navy asbestos programs and problems. Armed Forces Epidemiological Board, Ad Hoc Subcommiuec on Asbestos-related Health Problems. September 29, 1977. Levine RJ Led): "Asbestos: An Information Resource." DHEW Publication (NIH) 79-1681, 1978. Lewinsohn HC, Kennedy CA, Day JE, Cooper PH: Dust control in a conventional asbestos textile factory. Ann NY Acad Sci 330:225-241, 1979. Liddell FDK, McDonald JC, Thomas DC: Methods of cohort analysis appraised by application to asbestos mining. J R Statist Soc A 140-649, 1977. Liebcn J, Pistawka H: Mesothelioma and asbestos exposure. Arch Environ Health 14:559, 1967. 310 Nicholson, Perkel, and Selikoff Lilis R, Daum S, Anderson H, Andrews G. Selikoff JJ: Asbextosis among maintenance workers in the chemical industry and oil refinery,workers. In: Biological Effects of Mineral Fibers. International Agency for Research on Cancer, 1980. Lorimer WV, Rohl AN, Nicholson WJ, Selikoff IJ: Asbestos exposure of brake repair workers in the United Stales. Mt Sinai J Med 43:207-218, 1976. Love E: Maritime Administration, US Department of Commerce. Personal communication, Feburary 27, 1976. Mancuso TF: "Help for the Working Wounded." International Association of Machinists. Washington, DC, 1976. Maritime Administration, US Department of Commerce: Seafaring employment, ocean-going commer cial ships, 1,000 gross tons and over (unpublished). Marsh J: Director of Environmental Affairs, Raybesios Manhattan, Trumbull, CT. Personal communi cation. McDonald AD, McDonald JC: Malignant mesothelioma in North America. Cancer 46:1650, 1980. Meylan WM, Howard PH, Lande SS, Hantchett A: Chemical market input/outpul, analysis of selected chemical substances to assess sources of environmental contamination: Task III: Asbestos. Syra cuse Resource Corp., prepared For USEPA. Washington, DC, August 1978. MuTphy RLH, Ferris BG, Jr, Burgess WA. Worcester J, Gaensler EA: Effects of low concentrations of as bestos: Observations in shipyard pipe coverers. N Engl J Med 285:1271-1278. 1971. National Institute Tor Occupational Safety and Health: Criteria for a Recommended Standard: Occupa tional Exposure to Asbestos. HMS 72-10267. US Government Printing Office, Washington, DC, 1972. National Institute for Occupational Safety and Health. Unpublished data. (R. Zumwaldc, personal com munication). National institute for Occupational Safetyand Health, National Occupational-Hazard Survey, 1972-1974. (Unpublished data on SIC code sequence within alphabetized hazard description). Ncwhouse ML, Thompson H: Mesothelioma of the pleura and peritoneum following exposure to asbestos in the London area. Br J Ind Med 22:261-269, 1965. Nicholson WJ: Insulation Hygiene Progress Reports 3:1. Mount Sinai School of Medicine, NY, 1975. Nicholson WJ: Dose-response relationships Tor asbestos and inorganic fibers. Arb och Hylsa 17, 1981a. Nicholson WJ: Comment: The role of occupation in the production of cancer. Risk Analysis 1:77, 1981b. Nicholson WJ, Perkel G, Selikoff I J: Cancer for occupational asbestos exposure: Projections 1980-2000. Banbury Report #9. Cold Spring Harbor Laboratory, pp 87-111, 1981. Nicholson WJ, Selikoff IJ, Seidman H, Hammond EC: Mortality experience of asbestos factory workers: Effect of differing intensities of exposure. Environ Res (in press). Nicholson WJ: Health hazards of brake repair and maintenance. Report of the National Institute for Oc cupational Safety and Health, Contract 210-770-119, 1982. Nunneley JK: Department of the Navy. Personal communication, April 22, 1980. NYT (Nov. 15, 1981): The New York Times, Workers tested in cancer study, p. 30. Office of Management and Budget, Executive Office of the President: Standard Industrial Classification Manual 1972. Washington, DC, 1972. Polland LD: "The American Merchant Marine and the Asbestos Environment." US Maritime Administra tion, May 1979 (unpublished). Raybcstos Manhattan Co. unpublished data (J. Marsh, personal communication). Reitze WB, Nicholson WJ, Holaday DA, Selikoff IJ: Application of sprayed inorganic fiber containing asbestos: Occupational health hazards. Am Ind Hyg Assc J 33:179-191, 1972. Rohl AN, Langer AM, Wolff MS, Wcisman I: Asbestos exposure during brakelining maintenance and re pair. Environ Res 12:110-128, 1976. Seidman H, Selikoff IJ, Hammond EC: Short-term asbestos work exposure and long-term observation. Ann NY Acad Sci 330:61, 1979. Selikoff IJ, Churg J, Hammond EC: Asbestos exposure and neoplasia. JAMA 188:22, 1964. Selikoff IJ: The occurrence of pleural calcification among asbestos insulation workers. AmfNY Acad Sci 132:351, 1965, Selikoff IJ, Churg J, Hammond EC: The occurrence of asbestosis among insulation workers in the United Stales. Ann NY Acad Sci 132:139, 1965. Selikoff IJ, Hammond EC: Asbestos-associated disease in United States shipyards. CA-A Cancer J for Clinicians 28(2):87-99, 1978. Selikoff U, Hair and Cana Selikoff IJ.Nich Am J Ind Selikoff IJ: Disi nrent of I Steinlurth R:.Di bestos W Stumphius J: Ei Wong O: An epi : factor ad : Verma DK, Mid l -u' Assoc J ` ZumwaldeR-.N V