Document b5K5Zg3b4wm8Qd261rvYdyQz0

Fiber Types, Asbestos Potency, and Environmental Causation A Peer Review of Published Work and Legal and Regulatory Scientific Testimony DAVID EGILMAN, MD, MPH Scientific evidence and analysis offered in litigation These incidents remind us that the peer revi and public policy testimony have an important role in occupational and environmental health, but are not subject to peer review. Critique and commentary, attempts at reproduction of results, and review of data offered in such testimony is essential. Peer review of such testimony should become part, of the domain of medical and scientific journals. This paper is an effort to peer review the use of certain scientific methods in tort litigation and in testimony before regulatory agen cies. In this issue of IJOEH, Azuma et al. show that, back ground asbestos exposures can be considered to have process does not end with publication. This is true nr only for published papers, but. also for scientific argu ment and evidence presented as testimony offered for purposes of public policy-making and litigation. Azuma et al.'s paper in this issue, as well as letters from Hesseil and Welch and colleagues, have motivated this com mentary, which reviews the presentation of epidemiol ogy- and pathology- based testimony in asbestos litiga tion and regulation.6-8 The comments are designed to address general issues, but of necessity are comments on caused mesothelioma. In contrast, epidemiologic stud ies and testimony by Teta et al. and Price and Ware, and pathologic studies and testimony by Roggli and others, claim that background exposures are benign. These are fatally flawed because of methodological and ana lytic errors. Key words, asbestos; litigation; peer review: chrysotile; public policy; mesothelioma statements and/or publications of particular individuals. This commentary was reviewed by four experts, two of whom do not participate in U.S. asbestos litigation. In this issue, Azuma et al. use real, although limited, exposure data to correlate environmental "back ground" asbestos exposures with mesothelioma inci dence in Japan. "Background" has no universal defini INT J OCCUP ENVIRON HEALTH 2009;15:202-228 tion.* Azuma et al. correlated mesothelioma cases with environmental exposure data and the weighted average number of asbestos ferruginous bodies detected in the everal recent episodes of the publication of works lungs of the people with no identifiable point source of Sbased on partial or fabricated data have again exposure either occupational, para-occupationa! or revealed the weakness of the peer review process. known environmental. Their data roughly confirm the Dr. Scott Rubin fabricated data that appeared in at least U.S. Environmental Protection Agency's (EPA's) dose- 21 published peer-reviewed papers.1'2 Jonathan Leo response equation, which is consistent with a no thresh exposed the fact that in an article published in JAMA, old-effect level for asbestos-induced mesothelioma. authors misrepresented their consulting arrangements Azuma et al. show that many, if not most, "background" with Forest laboratories and concluded that Forest's mesothelioma cases are caused by ambient levels oi drug Lexapro was better than placebo, but omitted data asbestos which are attributable to asbestos released from the same study that showed that Lexapro is no during building construction and from automobile a better than counseling.1-3 In response, the Editor of truck brakes, among other sources. Sprayed chrysotile JAMA called Leo a "nobody and a nothing," tried to and amphibole asbestos was used in the United States as intimidate the Dean of his medical school, and banned well as Japan and other countries. him for life from publishing anything in JAMA.1 JAMA In addition to the Azuma paper, there is significant evi then let the perpetrators of the misrepresentation dence that asbestos causes most mesotheliomas. Mark explain away their misconduct in a letter to the editor and denied they had maligned Leo.4'5 Dr. Egilman is Editor-in-Chief of the InternationalJournal of Occu pational and Environmental Health and Associate Clinicai Professor in the Department of Community Medicine at Brown University. Address correspondence to the author at: 8 N. Main St.. Suite 404, Attleboro, MA 02703; email: <degi\inan@egilman.<:om>. Disclosures'. The authors has testified in asbestos litigation at the request of asbestos product manufacturers and injured workers. *It. is important io distinguish between occupational exposures (direct and bystander), non-occupational but dearly above-b; ground exposures (e.g., neighborhood and residential exposures as well as "handyman" and "shade tree" mechanic type of exposures, both direct and indirect) and "environmental'' exposures. "Ba,.. ground" exposures, as 1 use the term, refers to exposures with no id. .. tillable point source that would elevate airborne respirable asbestos fiber concentrations in excess of those recorded for the environm-... at large. .Azuma et al. refer to these exposures as '`environmental." and Yokoi reviewed all autopsies at Massachusetts Gen eral Hospital from 1896 onward, and failed to find any mesothelioma case before 1940/ They concluded that "the background level of diffuse malignant mesothe lioma in Europe and in the United States prior to 1930 was extremely low," and that, "current cases in Boston are not attributable to any significant background level [non asbestos cause] of the disease." In addition, Camus et al. reported seven "environmental" mesothelioma cases in women who lived near Canadian asbestos mines.10 Camus et al. concluded that the EPA risk formula overes timated the risk of asbestos lung cancer 10-fold. They reported, but did not analyze, the mesothelioma risk. Unfortunately, Camus et al. relied on particle counting techniques that were inversely related to actual asbestos fiber counts.11 (The higher the particle count, the lower die exposure.) In contrast, Swedish researchers who relied on fiber counts and controlled for smoking found that "low exposure" (10 fiber-years) relative risks ranged from 1.5 to 4.5, and argued the EPA model underesti mated the risk at. 1.10.12 Gustavsson et al. found a non linear dose-response relationship indicating that perfiber risks were higher at low exposures than at high exposures. Pan et al. found a relationship between dis tance from natural outcroppings of chrysotile (occasion ally containing tremolite) in California and concluded that the findings supported "the hypothesis that residen tial proximity to naturally occurring asbestos [NOA] is significantly associated with increased risk of mesothe lioma mortality in California."13 Despite this rather consistent evidence of real risk of mesothelioma from "background exposures," some industry consultants have assumed in testimony and publication that, background exposures are benign. In this commentary, I review these and related assertions on chrysotile potency and lung fiber counting, exam ining how they have been put to use in litigation and public policy hearings. SEER DATA CANNOT BE USED TO ESTABLISH A THRESHOLD FOR ASBESTOS INDUCTION OF MEOTHELIOMA Recent papers by Teta et al. and Price and Ware claim to establish a "safe threshold" below which asbestos does not cause mesothelioma.14-16 These authors have attempted to use the National Cancer Institute's Sur- lance, Epidemiology and End Results (SEER) data II' to estimate the "background" rate of mesothelioma in ft human populations.14,15 They define "background" gg; eases as mesotheliomas that occur in individuals who H? have no history of exposure to asbestos. From a scien|g|; ttfic perspective, this approach is problematic since it is j||. based on the unreferenced assumption and assertion - "-`t certain cohorts were never exposed to sufficient f atoounts of asbestos to develop asbestos-caused Mesotheliomas, based on the false premise that there were constant rates of mesothelioma over time. They base this assertion on mesothelioma rates--not expo sure data, interviews, medical record reviews or a search of medical literature. In fact, scientists have published contrary informa tion for more than a century and as recently as 2008.17-22 The Swedish Family-Cancer Database is the largest cancer data base in the world that links job and other factors and cancer incidence. Using this data, Hemminki and Li reported that; a comparatively "low [mesothelioma] risk among farmers [wrho have likely occupational exposures] suggests that the population at Large is at a risk of mesothelioma from undefined sources in urban areas." They concluded that "Back ground exposures do cause mesothelioma and epi demiologic data on excess risk should use the lowest rates for the least exposed as controls. Occupational and para-occupation exposures are added to `back ground' rates which have their own real risk." The UK Health and Safety Executive (HSE) has also agreed that "background" exposures cause mesothe lioma in adopting the position that: A PMR of 100 does not represent the `background' risk of mesothelioma (the level that would be expected in the absence of asbestos exposure), A hypothetical group of men with zero exposure to asbestos would record" PMR of approximately 6. . . . An occupational group with a PMR greater than 100 indicates that the level of mesothelioma mortality is higher than average for all occupations.-3 Disregarding this evidence, Teta et al. review SEER data and make the circular argument that mesothe liomas that occur in this cohort are, by definition, not caused by asbestos because the subjects were by defini tion not exposed, and therefore all cases are unrelated to asbestos.15 But if the mesothelioma cases were not exposed to asbestos why look at any death data? Every one agrees that absent exposure, asbestos is not a cause of mesothelioma. Teta et al. attempt to use mesothe lioma rates to "prove" there were no exposures. SEER data cannot answer this question; exposure informa tion can only come from patient histories and/or pathologic studies. These papers are an example of using the wrong tool (epidemiology) and the wrong data (SEER) set to obtain a desired answer to a question.24 Since all citizens in developed countries have lung asbestos burdens, there is no unexposed control group. There are many case reports of patients who developed mesothelioma after short, low-dose exposure. Most experts believe asbestos caused these cases.18,25-38 Epidemiology based on the SEER data cannot answer the question about the effects of low-dose exposure to asbestos because it includes no exposure data, and because the pathologic diagnosis of mesothelioma can be confused with other cancers (such as lung or ovarian), has changed over time, and can be 15/NO 2 APR/JUN 2009 www.ijoeh.com Fiber Types, Asbestos Potency, and Environmental Causation 203 TABLE 1 Mesothelioma Cases in Women Related to Domestic/Residential Exposure to Asbestos from Virginia Shipyards Name DOB DOD Occupationally Age at Exposed Death Family Member Exposure Site Occupation of Exposed Family Member 1, A., Laura M. 2. B., Bernice 3. B., Dorothy 4. B., Dorothy W, 5, B., Juanita J. 6. B., Marjorie S, 06/19/1921 08/13/1998 12/28/1935 12/17/1989 10/23/1919 09/05/1993 09/14/1924 02/19/2005 05/02/1921 10/02/2006 09/05/1918 07/16/1996 77 54 73 79 85 78 7, B., Mary Louis 03/17/1922 02/07/2001 89 8. B,, Sarah R. 9. B., Stachi B. 08/21/1926 10/27/1992 08/24/1915 12/24/1999 66 84 10. C,, JenelS Estes 09/01/1926 09/18/1996 11. C., Rosaiee S. 12/13/1929 02/18/2002 12. D., Betty L. 11/07/1932 05/09/2007 13. D,, Frances C. 01/29/1942 Living 70 71 73 14. D., Hope L. 01/21/1932 03/20/2005 73 15, E., Alma 16. E., Dorothy M, 17. E,, Mary A. 18. F., Irene 19. G., Dorothy Railey 06/25/1919 02/20/2009 10/17/1920 07/02/2006 05/11/1919 07/21/2005 10/28/1923 08/17/1986 10/25/1943 Living 88 85 86 63 20. G., Dorothy Savage 21, G., Frances H, 09/17/1915 05/24/1990 03/09/1922 10/14/2002 75 80 22, G,, Lillian L, 23. H,, Ronald L. 24. H., Sharon 11/04/1912 11/21/2002 08/30/1940 03/16/1995 02/03/1952 11/26/1995 90 53 44 Spouse Newport News Shipyard Father Newport News Shipyard Spouse Newport News Shipyard Spouse Newport News Shipyard Spouse Newport News Shipyard Spouse Norfolk Naval Shipyard, various contractors in NC Spouse Norfolk & Portsmouth Beltline, Portsmouth, VA Spouse CSX Transportation, Inc., Clarksburg, WV Spouse Philadelphia Naval Shipyard, Norfolk Naval Shipyard Step- Norfolk Naval grandfather Shipyard Spouse Newport News Shipyard Spouse US Navy at Newport News Shipyard Spouse; Spouse; Newport News Father Shipyard Father; Spouse Local #83, Norfolk, VA; Carpenter 8c Sons Spouse Newport News Shipyard Spouse Newport News Shipyard Spouse; Spouse Newport News Shipyard Spouse US Navy Father; Spouse Norfolk Naval Ship yard, Contractor, petroleum refinery; Virginia Power, Con- tel Telephone Co. Spouse CE Thurston, Norfolk Naval Shipyard, F.H, Gaskins 8c Sons Co. Spouse; Spouse Local 540 Plumbers and Sfeamfitters; Newport News Shipyard Spouse Newport News Shipyard Father Union Carbide, Charleston, WV Father US Navy, Local #10, Richmond, VA Pipefitter Joiner Welder - Fitter/Machinist Machinist Pipefitter, Carpenter Hostler, Fireman, Engineer Brakeman Pipecoverer, Insulator Plummer, Shipfitter, Supervisor Machinist N/A Machinist; Machinery Installation; Chipper Pipecoverer; Boiler Repairman Laborer Joiner Pipefitter; heating and boiling work Worked in engine rooms Storekeeper, Contractor, Laborer; Laborer, Line man, Installer Pipecoverer at al! three Pipefitter; pipefitter Sheetmetai Insulator Metalsmlth, Welder, Boiler maker (continued on next page) 204 Egiiman www.ijoeh.com INT J OCCUP ENVIRON HEALTH TABLE 1 (continued) Name 25. J., Iris Lee 26, M,, Daisy M, 27. M,, Diane T. Bunting 28. M., Dollie F, DOB 01/08/1926 01/06/1905 03/26/1952 03/01/1932 29. M., Elizabeth Frances 06/17/1920 30, M., Rebecca Louise T. 12/17/1931 31. O,, Ruby Lee 11/11/1920 32. S., Cailie Sue 03/31/1943 33, S., Leola Maxine 02/11/1929 34, S,, Opal D. 11/02/1921 36. S., Sharon Jane 08/11/1950 Mill 36. W., Carolyn J. 10/22/1935 37, W., Emma Moore 10/16/1921 DOD Occupationally Age at Exposed Death Family Member Exposure Site Occupation of Exposed Family Member 08/29/2003 04/06/1989 02/19/2004 77 84 52 Spouse Spouse Father 03/20/1993 61 Spouse 05/23/1983 63 Spouse 09/29/2000 69 Spouse 10/27/1990 70 Spouse 09/08/2007 03/10/1985 06/05/1987 06/10/1999 11/28/1999 64 56 66 49 64 Father Spouse Spouse Father Spouse 09/12/1995 74 Spouse Norfolk Naval Pipefitter Shipyard Norfolk Naval Pipecoverer, Shipyard Insulator Local #83, Norfolk, Pipecoverer VA; Norfolk Naval at both Shipyard North Carolina Ship Pipecoverer building & Drydock at all three Co., Fort Worth & Denver City Railway, Norfolk Naval Shipyard Norfolk Naval Ship Pipecoverer yard, Armstrong at all three World Industries, CE Thurston US Navy, SUPSHIP Machinist Mate, Mechanic, Machinist, Plan ner/Estimate Norfolk Naval Ship Sheetmetal yard, CSX Trans Mechanic; portation, Norfolk Sheetmetal Naval Shipyard mechanic, Pipefitter, Super visor; Pipefitter Norfolk Naval Pipefitter Shipyard Newport News Ship Handyman, building & Dry Dock Electrician Newport News Ship Pipefitter building & Dry Dock US Navy, Norfolk Machinist's Mate, Naval Shipyard Machinist Consolidated Rail Switchman, Corp. Brakeman, Conductor Newport News Ship Pipecoverer building & Dry Dock, at ail three US Navy, Norfolk Naval Shipyard influenced by the occupational history or absence thereof. These changes either may have reduced or ncreased the apparent rates of mesothelioma. Asbestos Exposure and Mesothelioma in Women and Young Workers fr'ice and Ware come to the conclusion, which is con tradicted by a cursory knowledge of the use of asbestos, that no mesothelioma case that occurred in a female Was ever caused by asbestos because no woman had ever had experienced sufficient exposure to asbestos.1'1 They ased this on the claim that female mesothelioma rates ;Jmain "unchanged" from 1973-2000. Price and Ware's SUSe of SEER data allowed them to conclude that all nale cases were unrelated to asbestos since female mesothelioma rates had remained "constant." In fact, Price and Ware contradict themselves on the article's most important point, "The age-adjusted mesothelioma rate for females was constant at an average of approxi mately 0.30 per 100,000 between 1973 and 1982, when it showed a one-time increase to 0.40 per 100,000 [emphasis added]." They go on to state, "One might be tempted to interpret this change as a response to increasing environmental exposure." 1 agree. Plowever, Price and Ware argue that since the rates remain con stant after 1992, this post-1972 increase is not causally related to asbestos exposure, but is instead explained by changes in diagnostic techniques. This assertion is unreferenced and un-described changes in techniques could just as easily decrease as increase the number of mesothelioma diagnoses. In addition, para-occupa- 15/NO 2, APR/JUN 2009 www.ijoeh.com Fiber Types, Asbestos Potency, and Environmental Causation 205 Rate --Rate Year of diagnosis Figure1--Women`s Mesothelioma Rates age > 60, based on data reported in; Teta MJ, Mink PJ, Lau E, Sceurman BK, Foster ED. US mesothelioma patterns 1973-2002: indicators of change and insights into background rates. Eur J Cancer Prev 2008;17:525-34. tional exposures from shipyard exposures reveal real risks to women,39 Table 1 is a list of some cases in women with para-occupational exposure from the New port News shipyard area. In this cohort, year of birth ranged from 1905 to 1952, and age of death ranged from 43 to 90, with 92% of the cohort older than 50. Of the 38 people in the cohort, 82% were exposed via their spouses and 21% were exposed by their father or both their father and spouse. Household exposures may be relatively high. Two exposed cases fit Teta et al.'s crite ria for non-exposure (born in 1952). In any case, the constant.])' elevated rates are compat ible with occupational and environmental exposures. A combination of changes in either exposure levels or pop ulation exposed (or both) could explain these findings. The SEER data provide no information on these ques tions. But even those data actually do show a broader change in women's mesothelioma rates over time. Price and Ware report age-adjusted rates which mask the increased rates of mesothelioma in women above 50.40 Teta et al. disaggregated the same data by age groups without deleting "deviants" and concluded that females above age 60, "had increasing rates from about 1977 through the late 1980s." Teta et al. claimed these rales were "followed by an apparent decline around 1992." However, her data, presented in a graph that is repro duced here(Figure 1), do not show a decline after 1992. In fact the rates peaked in 1995, dip in 1997 and increase until 2002. Given the small numbers and the quality of the data, it is inappropriate to make any conclusions for this data set; i t is especially wrong to base conclusions on "eyeballing the data." On the other hand, case reports and workplace- and environment-specific epidemiologic studies like that presented by Amina et al. clearly show that women had environmental and occupational expo sures that caused mesotheliomas. It is instructive to note that Price and Ware's con clusions conflict with data from countries other than the United States. In England, mesothelioma rates in females increased by about 20% from 1989-1991 to 1995-1997, and more than doubled by 2002-2004.41 Similarly, female mesothelioma rates in Australia rose about 3-fold between 1980 and 2000% The same pat tern has been reported from Italy.'13 These rates are likely to be more accurate than US reports because the national health insurance coverage in these countries likely encourages more complete discovery of cases and more sophisticated diagnostic methods. Pathologic evidence of female exposures completely disproves Price and Ware's hypothesis. Roggli et al. reported that as many as 75% of female mesothelioma cases had a history of asbestos exposure, but 80% of these were para-occupational.44 Lung tissue asbestos burdens were "elevated" in 70% of a series of female mesothelioma cases.4445 Most far-fetched among their claims is Price and Ware's un-cited assertion that "In contrast [to men], female exposures to asbestos have been primarily envi ronmental. In the 1930s through the 1960s, women generally did not work in industries in which men experienced high levels of exposure to asbestos." Given the sharp rise of female factory workers during World War II, as evidenced by the fame and success of the "Rosie the Riveter" campaign, it remains unclear how 206 Egitman www.ijoeh.com 1NT J OCCUR ENVIRON HcAriH -igure 2--Female Asbestos Textile Workers, c, 1922. Reproduced from: "Garlock 2009" S//de Presentation, Olwin Moeller v. Garlock Sealing Technologies, LLC. Case Number: 3:07-CV-65-H. United States District Court, Western District of Kentucky at Louisville. anyone could make such an ungrounded assertion. In the 1940s, women comprised 20-30% of shipyard work ers.16 Approximately 12 million women worked in the defense industries and support services across the nation, including in shipyards, steel mills, and foundries.45 During World War II, the Kaiser Company built shipyard child care centers for working mothers, which were funded through the United States Maritime Commission. Kaiser's two shipyard childcare centers in Portland served nearly 4000 children.47 Given the volume of information confirming women's work in high-asbestos exposure occupations and commonplace domestic exposure from asbestos contaminated cloth ing, it is hard to imagine how Price and Ware reached the conclusion that women's asbestos exposures, "have been primarily environmental." Even Wikipedia notes that Cooke first reported asbestosis in a female asbestos worker in 1924. In die 1930s, many textile workers were female.48 Spinning, weaving, and sewing were traditionally "women's work" ">hi exposures were far from innocuous (Figure 2).19 Brown et al. reported on a cohort of asbestos textile workers employed between 1909 and 1977 that uied 1265 women out. of a total cohort of 3072 Workers.50 Removing the flashings of molded articles can result in high exposures, and women performed this work in the manufacture of small asbestos cement Products and brakes (Figure 3).M Teta et al. claim male and female rates for the post- ^972 "unexposed cohorts" (male and female) are simi.^r- But their respective point estimates are 1.15 and .9-94, a 20% difference.15 There are few cases (72 male ar'd 58 female), probably because the oldest member the cohort was only 49 years old at the time of pub lication. Few asbestos researchers would venture any conclusions on such a small cohort with so short a potential latent period. Low doses are probably associ ated with longer latent periods.112 In addition, they assume that no one began any job with asbestos exposure until they were 19. However, many blue collar workers often begin formal employ ment at age 16 and children at still younger ages maywork with or around their parents who change their own asbestos brakes. These exposures can be quite high, as shown in Table 2. Teta et al. are more conservative than Price and Ware, claiming only that there was "little or no poten tial for occupational asbestos exposure [to men or women in the U.S.] after 1972," when the US Occupa tional Safety and Health Administration (OSHA) issued its first asbestos regulations. They state that asbestos use declined over the past 30-40 years. They go on to state unequivocally and without citation that: Since the mid-1970s, the potential for occupational and therefore domestic asbestos exposure would be minimal in the general US population, particularly for exposure to amphiboles. The mesothelioma rate in the population who entered the workforce after this time period of reduction of asbestos exposure would provide a reasonable estimate of the back ground rates of mesothelioma.1:1 figure 3---Finishing Asbestos Gaskets. Reproduced from: "Garlock 2009" Slide Presentation. Oiwin Moeller v. Garlock Sealing Technologies, LLC. Case Number: 3:07-CV-65-H. United States District Court, Western Dis trict of Kentucky at Louisville, 5/MO 2. APR/JUN 2009 wvjw.iioeh.com Fiber Types, Asbestos Potency, and Environmental Causation 20? TABLE 2 Studies Showing High Asbestos Exposures During Brake Work Author Year Exposure Type Lee68 Boiilat & Lob69 Castieman & Ziem70 Hatch7' Rodeisperger72 Kauppien & Korhonen73 Hlckish74 Hickish75 Clark76 Hatfield & longo77 Hatfield & Longo/8 Hatfield & Longo79 Hatfield, Longo & Newton80 Hatfield, Longo & Newton81 Hatfield, Newton & Longo82 Rohl et al.53 Osborn64 Roberts & Zumwalde85 Lloyd86 Longo, Mount & Hatfield87 1970 1973 1985 1970 1986 1987 1968 1968 1976 1998 n.d. 2000 2000 2000 2001 1977 1934 1982 1975 2004 Blow out Brake work undefined Damp rag Squirt bottle Stoddard Solvent Dry rag Brake washer Compressed Air Passenger car (various operations) Truck (various operations) Truck (various operations) Grinding Auto blow out Auto brake work, various Auto disc brake change Bendix Chrysier (filing and cleaning) Bendix Ford (filing and cleaning) Sweeping and cleaning brake shop Grinding Hand grinding Hand sanding Blowing dust Beveling Grinding Compressed air Servicing brakes Hand sanding and grinding and other operations Exposures Reported `^ 3-5 f/cc - 0,3-29.2 f/cc , High: 2.6 f/cc; TWA: 0.28 f/cc High: 0.54 f/cc; TWA: 0.2) f/cc High; 0,68 f/cc; TWA: <0.1 f/cc High: 0.81 f/cc; TWA: 0.2 f/cc High: 1.1 f/cc Fibers >5 pm: 2.1-8.2; 10 minute avg: 0 Mean: 3,8-4,7 f/cc Mean: 4,4-9.9 f/cc <0,1-125 f/cc; TWA: 0,1-0,2 f/cc 7 f/cc Peak exposure: 7.09 f/cc TWA: 1,57-2.55 f/cc 0.2-1.9 f/cc 8.53-14,57 f/cc 5,47-12,67 f/cc Personal Samples: 7.5-8.8 f/cc Area Samples: 2.0-2.4 f/cc 4.83-12.51 f/cc 12.57-21.43 f/cc 0.5-0.96 f/cc 6.6-29.4 f/cc 23,7-72,0 f/cc 17 mppcf 0.14-15.0 f/cc 3.75-37.3 f/cc 19.7-35.7 f/cc This is not true. This wishful thinking and derivative argument appear in the "Results'' section of the paper, although the authors never provide evidence that they studied or reviewed literature on the question of expo sure to asbestos at home, at work, or anywhere else. Annual asbestos consumption in the US peaked in 1973 at 803,000 metric tons, but remained relatively stable above 550,000 metric tons (except for 1949) between 1947 and 1979 53 For comparison, during WWII, use ranged between 232,000 and 398,000 metric tons. OSHA has never banned asbestos use (the agency does not have the legal authority to ban the use of any substance), and exposures up to 5 fibers per cc (f/cc) were permitted until 1976, when permissible exposure limits (PELs) dropped to 2 f/cc. Even defense witnesses retained by asbestos companies testify that two years of exposure to Canadian chrysotile at the 5 l'/cc level dou bles the risk of developing mesothelioma.54 Imports of asbestos for use in brakes increased three-fold between 1990 and 2002.55 The EPA banned spray asbestos in 1973, and in 1977 the Consumer Products Safety Com mission banned the use of asbestos in joint compound and spackling sold to the public. Currently, OSEIA has enough inspectors to investigate every workplace in America about once every 113 years.'16 Halley's Comet passes by every 75-76 years (Figure 4). In the absence of effective surveillance, asbestos reg ulations have often gone unheeded. For example, despite the 1972 OSHA regulations, workers at. the Newport News Shipyard received no training in asbestos safety procedures until 1978.57 Workers have testified that unprotected exposures from a variety of asbestos-containing products continued for several years after the training began.3' In 1983, the problem was so bad that Congress held hearings on the issue after complaints that the Navy was not monitoring shipyard workers who were exposed to up to 5 times the OSHA limit.58 It is worth noting that by the mid 1970s, there were few women in the trades (about 1 in 12), but 30% of the clean-up workers were women.5'1 Clean-up workers have the highest asbestos exposures in shipyards.60 Teta el ah repeatedly refer to amphibole asbestos as if it were the only exposure of concern and claim this exposure was eliminated on Navy ships in 1975. This is the year the Navy stopped adding new amphibole-con- 208 Egilman www.ijoeh.com INT J OCCUR ENVIRON HEALfH Figure 4 - The OSHA inspector meets Halley's Comet. taining pipe covering to ships; the tons of previously applied insulation did not disappear that year,'1 It was removed during the next decade and the highest expo sure occur during sweep up and removal/8,60 Exposures have continued into the 21st century: even an under-funded, short-handed OSHA has issued cita tions for overexposures to asbestos through 2008/1 Some have turned circumvention of the OSHA standard into a profit making business/32 A quick Google search reveals that the Boston Globe reported that: Albania Deleon, owner of Environmental Compli ance Training of Methuen, sold training certificates to hundreds of undocumented workers who had not taken a mandatory training course from 2001 to 2006. Deleon then sent them out to remove asbestos at job sites in New' England, and paid them under the table."62 OSHA has failed to enforce the asbestos standard in auto body shops/'3 As asbestos has been used in joint compound, house paints, ceiling and floor riles, vermiculite insulation and brakes, asbestos exposures among household members (50% of whom are women) also remain all too common in 2009. Expanded vermiculite (sold as WR Grace's Zonolite) was an easily poured insulation ideal for walls and attics. In 1985, the EPA estimated that 940,000 homes contained, or had once contained, vermiculite attic fill.33 Asbestos (being relatively indestructible) does not degrade on its own. It must be removed and is unknowingly released during renovations. OSHA does not regulate home renovation exposures unless they are performed by outside contractors. A study of fetal asbestos content provides further evi dence of potentially important: and continuing current exposures.64'65 Haque et al. studied asbestos content of - liver, skeletal muscle, and placenta digests of 82 : stillborn infants. They found asbestos fibers in 50% of :/he fetal digests: 88% were chrysotile, 10% were tremojjhe, and 2% were actinolite and antbophyllite. Mean : .l-xcept for Unibesios, a 70% amositc insulation which was pri used on nuclear vessels, chrysotile was the predominant anc ... -exclusive fiber in most pipe covering. fiber counts were highest in the liver (58,736 f/g), fol lowed by placenta (52,894 f/g), lungs (39,341. f/g), and skeletal muscle. The autopsies were conducted between 1990 and 1.992 and the maternal ages ranged from 17-42, indicating that some maternal exposures occurred after 1972. Ampleforcl and Ohar reported a pleural mesothelioma in a 22-year-oid woman born in 1980, whose father removed asbesios insulation from furnaces and pipes.6(3 The fact that humans are exposed to asbestos in utero further complicates any epidemiologic efforts to establish a threshold for asbestos carcinogenicity. As noted above, there are no unexposed controls, as in utero exposure provides an ample latent period and exposures to a developing fetus are likely to be more toxic than adult exposures.67 Because it appears that asbestos exposure is ubiquitous and begins in utero, epidemiologic studies cannot dis tinguish the effects of non-asbestos exposures that may appear to elevate mesothelioma rates (like radiation) from induction or promotion of the effect of asbestos. AVAILABLE COHORT EPIDEMIOLOGIC STUDIES CANNOT ESTABLISH A "SAFE" THRESHOLD FOR ASBESTOS EXPOSURE AND CANNOT BE USED TO ESTABLISH RELIABLE RELATIVE FIBER POTENCY ASSESSMENTS Some experts have used meta-analyses of asbestos cohorts to claim that exposure to chrysotile asbestos must exceed some "background" threshold to cause mesothelioma/8-60 Recently, an EPA-appointed Science Advisory Board (SAB) focusing on asbestos concluded that the available historical exposure data was too scant to reliably differentiate any potential potency differ ences by fiber type as attempted by Berman and Crump/9,91 Finkelstein commented, "In essence all of the input data would consist of guesses and the output of the model would not. be credible."61 As the EPA's SAB concluded, impinger data (which measured total parti cles and did not distinguish dust from fibers) "cannot" be "used to generate PCM comparisons."91 There is some evidence that the asbestos-mesothelioma relation ship may follow more than one dose-response curve. There are many case reports of mesothelioma in indi viduals with brief or "low dose" environmental or home exposure (see Table 1).18,31 On the other hand, "only" 10% of even the most heavily exposed cohorts develop mesothelioma.92 Clearly, genetic factors and other expo sures interact to produce mesothelioma in some, but not all, people with similar exposures. Hodgson and Darnton attempted to evaluate the rel ative potency of asbestos types using some of the same studies used by Berman and Crump.88 Rogers and Major, referring to Australian exposure data used by Elodgson and Darton, noted that, "the[se] exposure 15/NO 2, APR/JUN 2009 www.ijoeh.com Fiber Types, Asbestos Potency, and Environmental Causation 209 values . . . should be recognized as `guesstimates', made by people who have not been trained in occupational hygiene and who have no experience in asbestos dust monitoring."9" In addition to using the `guesstimates' of the Australian exposures, there was no exposure data for other crocidolite cohorts in their study, and the authors simply assumed an exposure level. Hodgson and Darnton then compared the crocidolite exposure guesstimates to the inaccurate exposure data from Canadian miner and miller cohorts. These McGill Uni versity studies funded by the Quebec Asbestos Mining Association found a slight inverse relationship between the particle counts they used and fiber counts.94 Their dose estimates were slightly better than random guesses.94 McGill researchers were aware of this prob lem and ignored it. In 1969, during a discussion on asbestos counting methods at an international confer ence on pneumoconiosis in Johannesburg, South Africa, McGill's Corbett McDonald asked, "Can an inac curate instrument like the midget impinger (MI), give an accurate result?"9-15 He was informed that it could not. Just as a slopped watch, which is correct twice a day, should not be used to tell time, unreliable exposure esti mates should not be used lo devise inevitably unreliable estimates of relative fiber potency. Hodgson and Darnton's comparison of dose-response relationships between these two large cohorts is as reliable as the square of the `'guesstimate." Hodgson and Darnton were aware of these problems as well, and wrote, "Cer tainly these estimates are much less soundly based than one would wish." Unfortunately, they pressed on stating, "Some view does however need to be taken. . . ."9G A wrong view based on inadequate data can be worse than no view at all; it can and has encouraged the continued use of chrysotile and been used to persuade juries that chrysotile products are harmless. Another weakness of the Hodgson-Darnton review is that it dealt with 17 cohorts representing special industries. It did not include any case-referent studies for end-use exposures, which represent the most common pattern for asbestosassociated mesotheliomas.88 Despite these failings and contrary to the positions taken by Price and Ware and Teta et ah, Hodgson and Darnton (whose model inher ently adopts a no-threshold assumption) rely on these "guestimates" to calculate relative potency for crocido lite, amosite, and chrysotile for mesothelioma induc tion of 500:100:1.88 Leigh and Robinson demonstrated the arbitrariness of these estimates.97 They recalculated them and accounted for clearance of amphibole and found potency ratios to be 26:14:1 which represents a twenty fold difference for crocidolite.97 An often-cited set of potency ratios in the literature is 30:15:1.98 Most other cohorts are too small to evaluate die effects of even moderate levels of exposure. Even fiber PCM counts may be misleading." Hein et al. found that "Current PCM-based methods may underestimate asbestos exposures to the thinnest fibers, which were the strongest predictor of lung cancer or asbestosis mor tality."100 It is possible that amphiboles are more potent than amphibole-conlaminated chrysotile, but existirw epidemiology cannot support or rebut this theory no matter how often it is repeated. Peto et al. tilled tbeir recent discussion of the issue of chrysotile causation "Speculations on the Contribution of Chrysotile," and with respect to ecological epidemiology-, speculation it is.101 At a recent deposition, Teta's employer, Dr. Dennis Paustenbach, agreed that epidemiologic studies could not establish a "threshold" for the asbestos-mesothe lioma dose response relationship, saying, . . why these epidemiologists are making these toxicology' statements [that there is a threshold] is beyond me but that's their choice."102 Ironically, these views on the limitations of epidemiology' did not prevent him from elsewhere using epidemiologic studies to claim a threshold for the chrysodle-mesothelioma relationship.103 Pathologic Evidence o f the Importance of Short Fiber Chrysotile as a Cause ofMesothelioma Substantial pathologic evidence contradicts the com pany-sponsored7 theory that chrysotile asbestos cannot cause mesothelioma.13 At least four studies that look at pleural fiber levels by fiber type find that "short" chrysotile is often only the only fiber type found and is almost always the predominate fiber in patients with mesothelioma.104-107 Lebouffant was the first to com pare lung and pleural fiber types and sizes, and found that pleural and fiber types were different in the same patients. He stated that: As a matter of fact, in several cases of mixed dusts (chrysotile-amphiboles), there is significant chryso tile enrichment in the pleural tumor, contrary to the observations in the lung parenchyma in which ... a relative amphibole enrichment was found. It thus appears that the chrysotile impoverishment of the parenchyma cannot be accounted for only by the dissolution of this mineral, but that there seems to be a preferential drainage of chrysotile towards the pleura.107 More importantly, he found that most fibers were short (<5pm) (Figure 5). Sebastian et al. compared die retention of fibers in parenchymal and pleural tissues in 29 patients with a variety of asbestos diseases and jobs.30" All but one 'Brake manufacturing companies GM, Ford arid Chrysler funded Teia el al. WR Grace, the seller of Zonoliie brand of vermiculite, funded Price's initial 1997 SEER paper.1" Price and Ware used some data from the 1997 paper in the 2004 paper.14 -While "short" is a relative term, federal agencies adopted a reg ulatory convention of counting only fibers of 5 pm or longer. Ironi cally, the 5 pm cut-off was arbitrarily established because the P!'e' dominance of short fibers in airborne samples made it difficult to count all fibers- The 5 pm "convention" has been carried over to lung fiber counting. 210 Egilman www.ijoeh.com INT J OCCUR ENVIRON HEALTH Figure 5--Relative Proportion of Short and Long Fibers - Lung Tissue. Reproduced from: LeBouffant L. Inves tigation and analysis of asbestos fibers and accompa nying minerals in biological materials. Environ Health -arspect 1974;9:149-53. worked with asbestos products and eleven had mesothelioma. In three of these cases, chrysotile was ie only fiber they found in the pleura. Two of these three had significant (78% and 25%) amphibole lung parenchymal fiber counts. The third (with only chrysotile in the pleura and lung) was a 78-year-old female '`unskilled worker" with no history of asbestos exposure.105 Sebastien et al. concluded: 1. This study has obviously demonstrated that lung parenchymal retention is not a good indicator of pleural retention, the most striking feature being the absence of relationship between parenchy mal and pleural concentrations, many pleural samples being free of asbestos fibers. The finding of many negative samples mav be due to an heterogeneous topographic distribu tion of intratissular fibers in pleural area. If fibers are concentrated in pouctual areas, they can be ignored by the transmission electron microscope (TEM) which observes a very small size sample. 2. This study has demonstrated that the retention of asbestos fibers in parietal pleura was type and size related, and that inside the parietal pleura most of the fibers were short chrysotile fibers. The presence of fibers in pleural tissues involves the translocation of fibers to pleura, and then the penetration of fibers inside tissues. Thus, two pos sible explanations can be given for these findings: a) Only chrysotile libers can be transported and reach the pleura. b) Fibers of all types can be transported to the pleural area, but only chrysotile fibers are retained in the pleural tissue.105 These finding have been reproduced by Suzuki and 'Wn,104 Dodson et al.,100 LeBouffant,107 and Kohyama and Suzuki.1011 Kohyama and Suzuki compared lung and pleural fibers in 13 insulation workers: three with asbestosis, three with asbestosis and lung cancer, and seven who had died from mesothelioma. Three had amosite and chrysotile in the lung but only chrysotile in the pleura.108 Six cases had discordant crocidolite counts with elevated concentrations in the lung but no fibers in the pleura. Overall, counting all fiber sizes, chrysotile counts were similar in the lung and pleura; in three cases chrysotile concentrations were higher in the pleura than the lung. Suzuki, Yuen, and Ashley examined 168 mesothelioma cases and found that the majority of fibers were short (< 5u) (89%) and thin (<0.25 pm) (93%) chrysotile fibers.100 Only 2.3% were consistent with the Stanton hypothesis that predicted that long fibers were more pathogenic than short fibers.100 In a small series of 14 cases with and without history of asbestos exposure, Boutin et al. found that amphiboles outnumbered chrysotile fibers in pleural tissue from all cases.110 Muller et al. could not replicate these findings, and stated, "In our collective of former miners of the Ruhr area we do not find asbestos fibers especially amphibole libers directly located in black spots."511 In Boutin et al.s' cases, the lungs contained 99% amphiboles, however they noted that chrysotile might have been hidden by debris.110 In addition, Boutin et al. failed to find chrysotile in cases where there was documented chrysotile exposure, and sug gested that "short and thin chrysotile fibers could be less easily detected among a `background' of particles in anthracotic samples."110 Despite this fact, and consis tent with Suzuki and others, only 22% of these fibers in black spots were longer than 5pm. Therefore, the majority (77%) of pleural fibers were short (< 5pm). Black spots do not correlate with asbestos pathology; in fact, Michev et al. found that "pleural plaques were mostly seen in the areas with a lower prevalence of black spots."112 Muller et al. found that, "The morpho logical finding of black spots is not an indicator for an existing mesothelioma or the possibility for the further development of a mesothelioma."111 Doclson et al. compared fiber types in the lung and pleura in 8 shipyard workers. All had amphibole and chrysotile fibers in the lung. One had only chrysotile in the lung.106 Animal studies support these human pathology findings. Short, thin chrysotile fibers induce pleural and peritoneal mesothelioma in rats.113-116 Wagner's rat studies provide reliable evidence of relative potency. As in observations in humans, chrysotile lung retention was relatively short. After 24 months, the animals had fifty times more amphibole than chrysotile in the lung (Figure 6). Retained lungasbestos did not predict either lung tumor or mesothelioma risk. Canadian chrysotile was much more potent, on a weight basis, than the amphiboles (Figures 7 and 8). VL 15/NO 2, APR/JUN 2009 . www.ijoeh.com Fiber Types, Asbestos Potency, and Environmental Causation . 211 Weight of dust in Sungs trng) EFFECTS OF INHALATION OF ASBESTOS IN RATS 2i 0 ' 1000Q 20000 3OCSO0 Cumulative doee (rcg/m5 hours) Fib. 0.--Mean wight of dussfc iu lungs of rats in relation to <3oso end time, Figure 6--Mean weight of dust in lungs of rats in relation to dose and time. Reproduced with permission from: Wagner JC, Berry G, Skidmore JW. Timbrel! V. The effects of the inhalation of asbestos in rats. Br J Cancer 1974:29:252-69. It is universally accepted that asbestos must reach the pleura to initiate cancer formation.** Short fiber chrysotile is the predominate fiber at the site of the mesothelioma.104,109 A minority of pathologists rely on SEM lung counts--which they admit are biased against finding chrysotile and biased toward finding amosite-- to argue that a certain minimum lung concentration of chrysotile must be present to establish that chrysotile has contributed to cause any particular mesothelioma.524 Roggli and colleagues are quite capable of comparing lung and pleural fiber burdens to disprove Sebastien's finding that his data . . obviously demonstrated that lung parenchymal retention is not. a good indicator of pleural retention, the most striking feature being the absence of relationship between parenchymal and pleu ral concentrations," but they have chosen not. fo repeat his studies.125 Asbestos fibers in the lung do not initiate mesothelioma formation. The fibers in the pleura cause the mesothelioma in the pleura and researchers from different countries studying workers in different jobs **Asbest.os stimulates mirapulrnonary production of cytokines sufficient to cause a mesothelial proliferation or pleural fibrosis and this may promote cancer cell growth. However, direct cellular con tact appears to induce mutations.117-123 have repeatedly found that pleural fibers are over whelmingly short thin chrysotile fibers.104-307,109 Selikoff found mesothelioma in 4.6% of amosite insulation and blanket manufacturing workers and 8% of insulation workers who used these products in addi tion to chrysotile products.20 Thus chrysotile appears to double the risk of mesothelioma compared to amosite-only exposure. Acheson and Gardner reana lyzed lung fiber burdens in patients with mesothelioma and found that mixtures of amphiboles and chrysotile are associated with a relative risk of mesothelioma of 61, compared to 12 associated with amphiboles alone and 6 associated with chrysotile alone.526 They reported that this pattern was closer to a multiplicative than an additive interaction between chrysotile and amphiboles.126 The synergistic effect, was strongest when the total fiber counts were low, which is the most common occurrence when Roggli dismisses chrysotile as a cause of a patient's mesothelioma. Ecological epidemiology' based on SEER data that include no information on history does not and cannot provide any useful information on individual risk or dis ease causation in general. Risk analyses, like those of Hodgson and Darnton and Berman and Crump, that rely on unreliable exposure estimates cannot, establish 212 Egilman www.ijoeh.com INT J OCCUP ENVIRON HEALTH Number of Mesothelioma After 24 Months Figure 7--Relative Mesothelioma Potency by Fiber Type in Rats, based on data from: Wagner JC, Berry G, Skidmore JW, Timbrell V. The effects of the inhalation of asbestos in rats. Br J Cancer 1974:29:252-69. fiber potency estimates. This is especially true when the results of both analyses conflict with animal experiments and human clinical data including clinical-pathologic evaluation of pleural tissue fiber levels. It is unscientific to use conversions that have been shown to be `guessti mates" to exclude known asbestos exposures as con tributing causes of in specific individuals.93,94,127 Rather, specific relevant clinical evidence and history of expo sure can establish cause-effect relationships in an indi vidual; pathologic studies of lung fiber counts that fail to reflect fiber types and systematically grossly undercount fiber types that are found at the site of the crime cannot only spread confusion or systematically mislead.124,128 Risk assessments based on unreliable exposure data may make for interesting theoretical exercises, but "guessti mates" should not be mistaken for scientific argument. THE USE OF FIBER ANALYSIS: A CASE STUDY OFHOW BAD SCIENCE CAN CONTRIBUTE TO BAD PUBLIC POLICY AND ERRONEOUS COURTROOM AND REGULATORY TESTIMONY Some researchers have used Sung fiber counts to claim that brake exposures do not contribute to mesotheliomas in brake workers.129 Butnor, Sporn, and Roggli compared lung fiber counts in 10 brake mechanics to a group of his torical controls, who they claimed had no occupational asbestos exposure.129,130 They further claimed that brake asbestos exposures did not contribute to the development of mesothelioma in these particular workers because their exposures were not higher than their laboratory's fiber counts for all of their allegedly unexposed cases.129 Lung Tumors igure 8--Relative Lung Tumor Potency by Fiber Type in Rats, based on data from: Wagner JC, Berry G, Skidmore JW, mbreii V. The effects of the inhalation of asbestos in rats. Br J Cancer 1974:29:252-69. OL 15/NO 2, APR/JUN 2009 www.ijoeh.com Fiber Types, Asbestos Potency, and Environmental Causation 213 Roggli himself contradicted this interpretation of the importance of above-background exposures when he testified at the request of an injured worker: Once a patient is diagnosed with mesothelioma, one of the first questions to resolve is where and when he or she was exposed to asbestos. Because asbestos dust is so strongly associated with mesothelioma, proof of significant exposure to asbestos dust is proof of specific causation in a given case. The sci entific and medical community have yet to deter mine a level of exposure to asbestos below which mesothelioma does not occur. While there is no threshold, there is insufficient evidence to implicate levels of exposure to asbestos that occur as a result of background or ambient air exposure. Very low levels of exposure above background, however, have been demonstrated to cause mesothelioma.131 Certainly brake workers have at least `'Very low levels of exposure above background." More importantly, Butnor, Sporn. and Roggli do not explain how their 10 cases were selected, except that they all came from a pool of cases that had been referred by plaintiff and defense lawyers, and that brake dust was the sole recognized source of asbestos exposure in all 10.129 They report no effort to deter mine if the chosen cases were in any way (fiber counts, work history, referral source) representative of the entire pool of cases. The authors should have specified a specific selection methodology to avoid bias, espe cially given the fact that Roggli had already concluded and testified on numerous occasions that brake expo sures cannot cause mesothelioma. Roggli's a priori hypothesis was that brake exposures do not cause mesothelioma. A more appropriate scientific test would have been an effort to find a worker with brake expo sures only whose fiber counts exceeded those of all "controls." A failed effort to disprove this hypothesis would have increased the likelihood that it was correct. On the other hand, finding a single case with elevated fiber counts would have disproved the hypothesis. Roggli has testified in court and regulatory hearings using the unsubstantiated assumption that "back ground" asbestos exposures do not contribute to mesothelioma risk. He has claimed132 that: 1. Exposures to asbestos from some asbestos products do not increase the risk of contracting mesothelioma. 2. Mesotheliomas that occur in some individuals with occupational asbestos exposure and lung asbestos burdens are "idiopathic" if their asbestos fiber counts are not higher than "95% of the control levels." 3. Ghrysotile asbestos from certain mines in California does not cause mesothelioma. Recently, Chrysler used this argument to justify a court order to stop the burial of Harold St. John, a brake mechanic who had died of mesothelioma, to get access to his lungs to perform a fiber analysis.133 A process server attended the funeral and, after the mourners had left, instructed the funeral director not to burv the body but to return it to the funeral home. Dr. Roggli testified at the request of Chrysler to establish the med ical importance of lung tissue burden to justify the subpoena. Chrysler had and has Mr. St. John's pleural and tumor tissue but refused to examine it for fibers. Roggli does not consider pleural tissue fiber level to be relevant to the issue of asbestos causation. I volunteered to testify for Mr. St John's family. On March 18, 2009 the New Jersey Appellate Court ruled that. Ford had no need to remove Mr. St John's lung tissue and his family was allowed to bury him.134 There are many problems with the purported scien tific basis of Chrysler's rather ghoulish request. The main problem is demonstrated in Figure 9, which shows a gentleman looking for his key. After helping him for awhile, you ask where he lost the key. The answer is, "On the next block." You then ask, "If you dropped it somewhere else, why are you looking for it here?" He answers, "Because the light is better." As Roggli and everyone else acknowledge, the asbestos in the pleura--not the lung--is the cause of mesothelioma. As noted above, there is no relationship between the asbestos in the lung and that in the pleura. Chrysotile is biopersistent in the pleura--not the lung--and amphiboles predominate in the lung and not the pleura (see Figures 5 and 6). While several researchers have been able to analyze pleural tissue, Roggli and coauthors reject the use of pleural fiber counts because of the perceived difficulties in obtain ing samples.13y There are many problems with the use of fiber counting to determine causation in individual cases. I review some of them here. Roggli summarized his use of fiber counting in his testimony in the St.John case: Well, I think that there are three scenarios that I could envision that you would see as a result of doing the fiber analysis in this case. One would be to find a fiber burden which is no different from our background or control population, which would indicate, in my opinion, that it's an idiopathic mesothelioma. The second would be that you would find elevated levels of commercial amphibole fibers, indicating that there was some exposure that has not been identified, other than to friction products, and that likely was the cause of the mesothelioma. And the third possibility is that you would find only elevated tremolite and/or chrysotile present in the tissues, and that would actually be a finding that, would be favorable towards the Plaintiffs.136 In the third scenario, Roggli implicitly acknowl edges, but avoids affirming, the fact that elevated 214 Egiiman www.ijoeh.com 'NT J OCCUP ENVIRON HEALTH figure 9--Looking Where the Light is Better. Art cour tesy of Arit by Cortooncity.net. tremolite and/or chrysotile would be evidence of mesothelioma caused by brake exposure. In the first two scenarios, Roggii uses fiber count data to assign causation to commercial amphibole exposures and exculpate brake amphibole (and/or chrysotile) expo sures. In essence, Roggii compares lung fiber counts between mesothelioma cases and a group of ``controls" whose fiber levels he claims represent "background" exposures. This can be misleading and underestimate asbestos contribution to causation if any of the follow ing three scenarios occur. 1. If the fiber counts in "controls" are high because of unrecognized occupational, para-occupational (household or similar), or environmental exposure, all these comparisons will be biased against finding that the case's asbestos exposure contributed to cause the mesothelioma. 2. If Roggli's technique systematically undercounts chrysotile, it will underestimate the contribution of this fiber type (and total asbestos exposure) in exposed cases. Assuming there is a difference in chrysotile counts between the groups, undercount ing chrysotile creates a bias against finding a differ ence between cases and "controls." If chrysotile is present in cases and not "controls," undercounting will result in low or no fiber detection in both cases and "controls." 3. If a higher percentage of chrysotile (compared to amphibole fibers) translocates to the pleura, lung fiber counts will underestimate the contribution of chrysotile to disease causation. I now address the underlying studies and arguments tliat form the basis of Roggii's testimony. 1. Did controls have occupational or environmental exposures? In court testimony, Dr. Roggii has been quite critical of the controls and techniques used by other scien tists.137 For example, he has criticized Dr. Abraham for relying on controls performed in another laboratory, saying "I think that--that that [sic] is not good science; and in my opinion the [Abraham's] numbers are not interpretable."1"1 In a presentation to asbestos company defense lawyers, Roggii claimed that Suzuki's laboratory was contaminated with chrysotile.104 However, he failed to note that 3.2% of Suzuki's cases were chrvsotile-free, a fact which rebuts this criticism.136,138 In addition, Suzuki ran controls in his 2005 paper to rule out con tamination from water, fixative or formalin/09 Srebro, Roggii, and Samsa109,138 selected twenty patients who they claimed had "no documented history of asbestos exposure and no evidence of asbestos-related disease" as controls to determine the lung burden of asbestos in people who they claimed had no occupational history of asbestos exposure (background exposures)/30 After looking at controls' fiber counts, however, the authors found high amosite levels in one patient. In response, they conducted "an extensive search through this patient's medical records and [made] two phone calls to surviving relatives [which] revealed that his employment history included installing furnaces, an occupation associated with asbestos exposure." This case was important evidence that their original screening had failed to exclude individuals with important occupational exposures. Srebro, Roggii, and Samsa excluded the "con trol" post-hoc based on the actual results of the only out come of interest--lung fiber counts. Additionally, they failed to repeat this "extensive search" with the remaining "controls," despite the fact that at least 8 had occupations that are more usually associated with occupational asbestos exposure than "furnace installer" and three lacked any occupational history. There is no justification for excluding only the con trol with the highest counts, other than the fact that the inclusion of this individual would have obviously sig naled the inadequacy of their selection criteria for "unexposed" controls. Srebro, Roggii, and Samsa do not explain why they did not obtain more information on rtIn this criticism Roggii emphasizes the lack or reproducibility of results between laboratories which complicates and undermines the value of non-research use of fiber counting. 'Of 15/NO 2, APR/JUN 2009 www.ijoeh.com Fiber Types, Asbestos Potency, and Environmental Causation 215 the control with the next highest levels and so on down the line.130 Had they used this same standard ("exclude controls with "high" counts) for all controls, they could and should have excluded every "control" but the one with the lowest fiber counts. I have previously described the use of arbitrary and non-standardized criteria for the selection of controls as "differential peeky bias."530 It is unclear why Srebro, Roggli, and Samsa failed to exclude control case 19 from the paper (never mind as a "control"). The paper was based on the premise that it was a study of patients who all had asbestos body counts within their laboratory's "normal range": This report presents a comparison of data for 18 mesothelioma cases with AB counts (by light microscopy [LM]) within our "normal" range versus data for 19 "control" cases with normal lungs at autopsy. Our normal range is 0 to 20 AB/g. . . . [ital ics in original] ,130 "Control" case 19 had 22 asbestos bodies per gram, which is higher than Roggli had repeatedly reported (both before, after, and in the 1995 publication of this paper) as the high end of his normal range.4'"'129~130,H0~HS Srebro, the first author and a medical student w'ith no training in occupational medicine at the time she col lected the data, conducted the investigation to deter mine if the controls had a history of precious history of work with asbestos.144 None of the controls were inter viewed because they were all dead at the time the study was conducted.144 The listed occupation for three of the 19 controls was NA (not available) and the researchers had no information on smoking for 10 of the "controls" (Table 4) .13<} This indicates that Srebro failed to access or record from information sources that almost always con tain this information, such as complete medical records or interview's with family members, to determine what jobs or environmental exposures the controls had. Several of the study controls had likely occupational exposure to asbestos.130 Control 24, one of the patients w'ith unknown occupation, had the highest total "con trol" fiber count--more than three times the next high est "control" and the fifth highest level for all the cases reported (18 mesothelioma patients plus 19 con trols).130 Other "control" cases w'ith possible occupa tional asbestos exposure included three manual labor ers, two listed as "Air Force," two hospital workers, an electrical engineer, a spinning mill worker, a truck driver with esophageal cancer, and a garage owner. According to the U.S. National Institute of Occupa tional Safety and Health's (NIOSH's) Work-Related Lung Disease survey, hospital workers, truck drivers, electricians and farmers are in the top ten recorded industries in workers with mesothelioma.145 For exam ple, a garage owner likely will have entered the service area of a garage where asbestos exposures are all too common.145 Similarly, manual laborers, Air Force veter ans, electrical engineers, truck drivers and spinning mill workers all may have had occupational exposures to asbestos/1 Ironically, three years before the 1995 study was published, Roggli reported that manual laborers had occupational exposures and had median asbestos body counts of 830, nearly three times higher than levels in shipyard workers (295).140 Srebro, Roggli. and Samsa did not exclude poten tially confounding "environmental" exposures when they labeled their "controls" as having had "background exposure."117 They distinguish household and "environ mental" from "background" exposures.143 Roggli believes household and "environmental" exposures can cause mesothelioma, and has provided examples of environmental exposures that can cause mesothelioma, including "living near an asbestos manufacturing plant or a mine or a mill ... in Louisiana many of the drive ways and playgrounds down there used tailings that Johns Manville had from a manufacturing plant, deposits of tremolite, for example in the El Dorado area of California . . . and . . . Libby, Montana due to the mining operations."148,349 In addition, Roggli believes that household exposures in patients who live with asbestos-exposed workers can cause mesothelioma.127 Srebro, Roggli, and Samsa's "controls" are indistin guishable from their mesothelioma cases. Srebro, Roggli, and Samsa reported that the mean amosite and tremolite, anthophyllite, and actinolite (TAA) levels were statistically significantly different between cases and controls (but failed to note that this was only true alter thev deleted the "control" patient with the highest fiber counts). Srebro, Roggli, and Samsa reported the mean amosite level for their mesothelioma cases as 270 uncoated fibers per gram of wet lung tissue (uf/gwt) but the correct value appears to be 240 uf/gwt. In the text they report that the one-tailed Wilcoxon test (per formed after excluding the control case with elevated amosite levels) showed a statistically significant differ ence between cases and controls in amosite (p<.006) and tremolite (p<.004) lung burdens. However, this is incorrect and the authors cannot explain how they achieved this result.-- In the footnote below their Table 2, the authors write that the same result is a comparison of means, but Wilcoxon is not a comparison of means. Wilcoxon is a non-parametric test for assessing whether :rI have reviewed a case of mesothelioma in a truck driver who received occupational exposure to asbestos by adjusting the brakes on his trucks. sDr. Samsa responded to my request for an explanation of the statistical analysis and answered, "I'm afraid that I can't be of much assistance as, if my recollection from over a decade ago is correct, my role in the analysis was limited to the exploration of uni-modal versus bi-modal distributions. In re-reading the paper, one thing that would have been helpful to report was how the laboratory values that were below the threshold of detection were treated--for example, were thev set to 0, to 1/2 the limit of detection, etc. In the absence oi this information, it is difficult to comment on your questions. Perhaps the first author can be of more help." 216 Egilman www.ijoeh.com INI j OCCUR ENVIRON HEALTH 14000 12000 3 10000 LL D w 8000 CD J3 TQJ1 6000 fS ocO r> 4000 9 2000 i j__ i A*. 'T',U' 'vV.vV;T;' UrivUu.vi^xr.Uv. cU-mfi'i XXA' iUcSfUfAiUs 7A7'XIOiTiTw:ipjYh' w-;O?-pxwfP-pM'mPl' '>Yet 'i !i i *s If Control Meso Mean Median ai Ii t 1* Total AMOS TAA CHRYS NAMF Figure 10--Medians, means, and ranges of uncoated fibers per gram of wet lung tissue in "controls'' and mesothe lioma cases, based on data from: Srebro SH, Roggli VL, Samsa GP. Malignant mesothelioma associated with low pul monary tissue asbestos burdens: a light and scanning electron microscopic analysis of 18 cases, Mod Pathol 1995;8:614-21. two independent samples of observations come from : the same distribution. In any case, a two tailed test should have been performed since it was possible that ; some of the controls had higher fiber counts than the mesothelioma cases. Indeed this turned out to be the case in control 24 and the deleted control. For tremo... lite, the significance level for the one-tailed Wilcoxon result, was pc.063 and for the two-tailed test was c.025. f The significance level would be p=.05 for the correct., If two-tailed test. However, I have been unable to repeat their statistical results. Even using Roggli's statistical f| method and after deleting the excluded "control/' |f there was no significant difference for asbestos bodies, total fibers, and chrysotile between cases and controls. |V Using the correct, (two-tailed) test that accounts for if left censored data, the controls and mesothelioma |: cases fiber counts are not different.130 Fiber counts for both controls and cases overlap (see Figure 10). These results mean that either asbestos did not contribute to any of the cases or the controls do not represent expo sures that are without risk. Based on the occupational histories and fiber counts, the latter is clearly the case. There is no scientific basis to state that a "control" had no occupational exposure to asbestos if there is no infor mation on their work history. It seems that this missing ^formation invalidates Roggli's subsequent papers and individual case causation determinations based upon, the data (or rather lack of data) in this study. 2. Did controls have typical/representative "back ground" exposures? "Background" asbestos lung levels are a function of background ambient air concentrations, which are related to geographic location. Areas adjacent to asbestos manufacturing plants and mines and cities in general have high levels compared to other areas. There is no standard "background" exposure, as ambi ent air levels and lung fiber counts vary.151 Srebro, Roggli, and Samsa did not report any information on the geographic distribution of their "controls" and never evaluated environmental or household exposure differences.14' Further, Srebro, Roggli, and Samsa do not distinguish "environmental" or household expostire levels from "background" exposures. Since Roggli himself believes that environmental and household exposures cause mesothelioma,130 his "controls" do not represent a threshold for the induction of mesothelioma. 3. Did the counting method reflect actual fiber levels? Roggli's scanning electron microscope (SEM) method cannot "see" the thin chrysotile fibers that are most common in the lung and the pleura, which leads to undercounting of chrysotile and the misleading con clusion that chrysotile is not an important cause of mesothelioma. Roggli uses a scanning electron micro scope (SEM) set at a magnification of only 1000X (the method is capable of 10,000-20,000X), which misses chrysotile fibers that are <0.1.5pm in diameter.152 As a result, he fails to count most chrysotile fibers which are, on average, .03p.m-.07pm in diameter.129,136 T 15/NO 2, APR/JUN 2009 www.ijoeh.com Fiber Types, Asbestos Potency, and Environmental Causation 217 The EPA evaluated various measurement methods and concluded: SEM, for purposes of this rulemaking, was deter mined to be inadequate for building clearance for the following reasons: (1) Currently available methodologies are not validated for the analysis of asbestos fibers; (2) SEM is limited in its ability to identify the crystalline structure of a particular fiber. (SEM analysis is therefore confined to the identifi cation of structures by elemental composition and morphology); (3) recent studies conducted by MBS have evaluated several types of scanning electron microscopes and the variability between these instruments. (NBS has found the image contrast of the microscopes is difficult to standardize between individual scanning electron microscopes); and (4) currently no laboratory accreditation program exists for accrediting SEM laboratories.153 All other US agency protocols that relate to fiber counting use only TEM analysis. These include EPA's Asbestos Hazard Emergency Response Act Proto col,154 NIOSH 7402,155 ASTM Air sample analysis,156 ASTM Dust sample analysis,1"7 International Organi zation for Standardization (ISO) air sample analy sis,358 EPA bulk sample analysis,153 and EPA Super Fund site air analysis.159 The EPA scientific advisory board on asbestos used strong language to support the use of TEM: Multiple binning should be evaluated, but only using TEM-analyzed environmental exposure data that is directly associated with health outcomes. Studies con tinue to reveal the importance of fiber width in potency. Fiber width is tire most critical dimension in determining deposition site in the respiratory system, plays a significant role in determining sur face area exposed to tissue, and may be a factor in mobilizing libers from alveoli to pleural space. Future attempts to model fiber potency should have at least two bins for width. One possible width divi sion could be an aerodynamic diameter of 2.5 pm, which is the cut point for EPA fine (-respirable) par ticles. This would be ~0.5 for amphibole asbestos and -0.65 pm width for chrysotile. . . . only TEM-ana lyzed environmental exposure data that is directly associ ated with health outcomes should be used, for risk assess ment. [emphasis in original].91 Follow-up on chrysotile-exposed textile workers has shown that thin fibers significantly contribute to the risk of contracting asbestos-related lung cancer and mesothelioma.i00,!65 Roggli admits that his method undercounts chrysotile fibers, but claims that SEM undercounts amosite.124 While it is undisputed that amphibole is a cause of mesothelioma, Roggli's flawed method--which systematically undercounts chryso tile--supports his conviction that chrysotile has not contributed to mesothelioma causation in certain indi viduals or more generally in those exposed to certain chrysotile-tremolite products. His technique is biased in a direction that supports his argument. Since chrysotile fibers are biopersistent in the pleura and not the lung, while amphiboles are biopersistent in the lung and often fail to reach the pleura, over time chrysotile levels will decrease in the lung. Using a one year half-life for chrysotile and a 20-year half-life for amphiboles, the amount of chrysotile remaining in the lung 30 years after exposure would be 1 billionth of what was inhaled, while almost 30% of the amosite would still be present in the lung. Elsewhere, Roggli has undermined the validity of Srebro, Roggli, and Samsa's conclusions on the ques tion of chrysotile causation in genera! and the contri bution of chrysotile-containing products (like asbestos brakes) to the induction of a mesothelioma in any par ticular individual. In a 2000 paper, he states that, "Fiber burden studies do not accurately reflect past exposures to chrysotile/'102 At that point, he maintained that these studies "afford limited information regarding the role of chrysotile asbestos-related lung cancer since chrysotile is broken down in and removed from the lung. And long, thin, greater than 5 micron chrysotile fibers are not readily detectable by our technique."102 4. Did the counting method count short fibers? Roggli's method fails to count fibers that are shorter than 5 pm in length, leading to further undercounting of chrysotile and over-emphasis on the role of ampinboles in causation. All scientists who have published on pleural fiber counts find short fibers to be the most common and often the only pleural fiber.105,103-16'1 The combined effect of using an insensitive instrument and the decid ing to not count short fibers is dramatic. In one blinded cross-laboratory comparison on the same patient, Dodson et al. found 84 fibers per gram while Roggli reported only one.103 5. Was a standard procedure used for all cases and controls? Srebro, Roggli, and Samsa used two different meth ods to prepare tissue specimens. The)' described these methods as follows: AB counts for all 19 control cases and for 6 mesothe lioma cases (Cases 2 and 13 to 17) were quantified using the technique of Smith and Naylor for approx imately 5-g samples of lung tissue. In six mesothe lioma (Cases 3, 5, 6, 7, 10, and 18) limited lung tissue (<1.0 g) was available. For these cases, our laboratory developed a hypochlorite digestion pro cedure [modified from "Williams et al].130 Srebro, Roggli. and Samsa then cited a 198(5 paper (by Roggli, Pratt, and Brody) that compared the valid 218 Egilman www.ijoeh.com (NT J OCCUP ENVIRON HEALTH ity of Che two techniques, stating "on average, [the modified techniques values were] within 10% of values determined by the Smith and Naylor procedure."143 However, the earlier paper's comparison related only to asbestos bodies. Srebro, Roggli, and Sarnsa failed to report the range of counts for their subjects, which showed that values differed by 10 fold (0.31-3.53) between techniques. This ten-fold range indicates that the different techniques are not comparable. 6. Can fiber counting determine how long a fiber has been in the lung? Pathologic evaluations cannot determine when a fiber entered the body and recent exposures do not contribute to cancer formation.136 7. Was crocidolite found in cases or controls? Srebro, Roggli, and Samsa did not find any crocido lite in any patients, but Roggli misreports this fact in subsequent publications. Srebro, Roggli, and Sarnsa never found crocidolite, but in every subsequent publi cation of his data, Roggli lists the amosite counts as "AC" (amosite and crocidolite). This is misleading, because Roggli's SEM method will miss the vast major ity of crocidolite fibers, which are too thin to be seen at Rogglfs preferred microscope setting.160 8. Was a standard method used to compare cases to his "controls"? Srebro, Roggli, and Samsa's paper states that their "study demonstrated that approximately one-third (6 of 18) of the mesothelioma cases have asbestos fiber burdens greater than 95% of the control levels" They concluded that these cases were caused by asbestos.130 The authors fail to explain what this is 95% of. However, in recent testimony, Roggli explained: "There was one control case which we eventually threw out because we discovered through work that Dr. Srebro did that that person had an occupational exposure. That's the 95 percent."144 A close reading of the paper shows that their com parison wus ad hoc: They discounted values for cases if there they only found one fiber. Srebro, Roggli, and Samsa wrote that in mesothe lioma cases 6, 7, and 1.4, amosite fibers were detected but were not clearly above background level because the calculated values were based on a single fiber detected {versus none detected in controls) [italics in original]." Srebro, Roggli, and Samsa classified all these cases as of "uncertain etiology."130 Roggli recently testi fied on this issue, saying, "(W]e typically require 2 [fibers] to be an unambiguous result. Two fibers. One fiber even though it's more than we found in our con trol still might be just a matter of chance and [an] ambiguous result."144 * In contrast, they did not discount values for "con trols" if the fiber estimate was based on only one fiber. Tremolite fiber counts in controls 20, 22, 24, 25, 31, and 35 are all based on finding a single fiber. Two of these "single fiber" controls (20 and 24) had the high est tremolite values for all "controls" (2540 and 1770, respectively). In all subsequent evaluations, Roggli only attributes causation to tremolite exposure if levels are above these two "one fiber" controls. Except for a single chrysotile fiber found in controls 29 and 34, Srebro, Roggli, and Samsa found no chrysotile in any "controls," but nevertheless use these as a basis for comparison with case chrysotile levels.130 Roggli's inconsistent exclusion of fiber counts based on the finding of a single fiber has an important impact on his conclusions. Roggli. discounts brake exposure as a cause for mesothelioma in case 5 in his 2003 brake study, based on two of the single fiber tremolite "con trol" cases (20 and 24).129 Otherwise, case 5 has higher tremolite levels than all but one of his "controls" (and no amphiboles). This would meet his original criteria of determining that asbestos caused a mesothelioma if he finds a fiber count that exceeds those reported in 95% of his controls. Finkeistein reviewed Butnor, Sporn, and Roggli's comparison of brake worker and control fiber counts and elegantly showed that the authors performed an incorrect statistical analysis by comparing medians.166 Finkelstein's correct analysis revealed that cases had significantly more tremolite than the 1995 "controls" (Figure II). Roggli responded to Finkelstein's critique by stating, "What Dr. Finkeistein seems to ignore is that in 'every case' with an elevated level of chrysotile or non-commercial amphibole fibers, there was also an elevated level of commercial amphibole fibers (amosite or crocidolite)."16' However, Roggli found no amosite or crocidolite in either case 4 or 5. Case 4 had no asbestos bodies, and according to Roggli chrysotile can form asbestos bodies.111 Srebro, Roggli and Samsa discounted fibers found when the result was positive but below the highest detection limit in any of the "controls." Srebro, Roggli and Samsa discounted the results in case 2 because they claimed that they used a larger tissue sample size (5 g) than that used for some of the other cases, and as a result had a lower detection limit than all of the controls. However, case 2's tissue sample was the same size as that of all the "controls." The paper reports that analysis was conducted with "approxi VOL 15/NO 2, APR/JUN 2009 www.ijoeh.com Fiber Types, Asbestos Potency, and Environmental Causation 219 mately 5-g samples'' for all controls and "six mesothe lioma cases (Cases 2 and 13 lo 17)." Although case 2 worked at shipyards during WWII and had amosite fibers in his lung, Srebro, Roggli and Sarnsa classified it as of "uncertain etiology." Srebro, Roggli and Sarnsa discounted chrysotile as a cause even if the levels were greater than 95% of all "controls." Srebro, Roggli, and Sarnsa wrote, One additional case (Case 17) demonstrated a chrysotile fiber count greater than 95% of the levels for the controls but not greater than all of tire con trol values. However, SEM is not as sensitive at detecting chrysotile fibers, which are frequently less than 5 pm in length and thinner than 0.1 pm. The authors here paradoxically conclude that although this case met their criteria for assigning asbestos causation--and indeed would have exceeded those criteria if their own methodology' had been more accurate--that: the role of asbestos causation in this case was nevertheless "uncertain." This appears to he an admission that their methodology is fatally flawed, at least with respect to the evacuation of chrysotile as a cause of mesothelioma. Roggli has changed his criteria for comparing new cases to his historical controls on several occasions. In 2002, he compared new cases to the highest detection level for chrysotile in any "background case" (control), which was 2540 uf/gvvt (see Table 2).128 In 2003, he compared new cases to the highest actual recorded value for chrysotile (1000 uf/gwt) but continued to compare tremolite to the highest detection limit in any "control."129 He also dropped the 95% comparison after the original paper, and since 2002 he appears to require that the case have fiber levels that "exceed all controls" to qualify7 as a potential asbestos-caused case.130 As noted at the beginning of this section and consis tent with his interpretation of cases 4 and 5, Roggli has testified that he would not attribute causation to brake exposure even if the patient had elevated lung chrysotile and/or tremolite levels, essentially admitting that he does not follow his own methodology7 if the results conflict with his prior opinion that brake expo sures cannot cause mesothelioma.130 9. Are there problems with the approach to attribution of causation in individual cases? Roggli has asserted that lung fiber counts must exceed "background or control" to establish causa tion.124 This presumes that there is a threshold for asbestos induction of mesothelioma and that, the thresh old is at or below "background." However, Roggli him self has stated that "no threshold has been identified for asbestos exposure below which mesothelioma will not io J iri - >o 7, &C J 3 | *- Li. Case . - Control in -j 5.5 6 6.5 7 Mean (LogConcentration) Tremolite Fibres Graphs by CaseControi 7^ Figure 11--Tremolite levels of brake workers vs. controls, reproduced with permission from: Finkelstein MM. Asbestos fibre concentrations in the lungs of brake workers: another look. Ann Occup Hyg 2008:52:455-61 occur."124 In the lung, there is no qualitative difference between asbestos fibers from "background" exposure and those from asbestos products. If asbestos can cause mesothelioma, then fibers from "background" or ambi ent air can cause mesothelioma.10'12'20-28,160-168-'74 There fore, there is no reason to exclude them as causes of mesothelioma. In this issue, Azuma et al. provide further evidence that, there is no threshold for asbestos induc tion of mesothelioma, as Roggli has done in an affidavit: It is also my opinion that, it is the total dose of asbestos, regardless of fiber type, that the patient experiences that causes the disease, . . . It is further my opinion that each and every exposure to asbestos that an individual with mesothelioma experienced in excess of a background level is a substantial con tributing factor in the development of the disease.131 If a threshold for asbestos induction of mesothe lioma exists, and lung fiber burden drives pleural levels, then "background" exposures that reach the pleura will be added to "occupational" exposures and contribute to induction of the cancer and/or promote its growth. According to Roggli, it. is the combined total dose of asbestos which causes mesothelioma.13,1 Roggli believes that seven fiber-years of chrysotile exposure are required to cause mesothelioma, except for house hold or environmental exposure.136 However, if two products each contribute one-half of the dose neces sary (in Roggli's view) to cause the mesothelioma, he will not attribute any role in causation to either one. ' This reasoning has no scientific basis. When it comes to legal causation, which calls for a contribution to be "sig nificant" at some comparative exposure level, it may be reasonable to conclude that an exposure was trivial (for example 1 fiber out of one billion), but Roggli's posi tion that an exposure that constitutes 50% of the suffi cient dose is trivial is erroneous. Interesting!}', in some 220 Egiiman www.ijoeh.com iNT J OCCUP ENVIRON HEALTH 0" 2 -gco uo 2 o o Q uoZ ^c oco c.g > CD -C 0 O g< O C'TO CM Q2^c-> >. O= oc OQ. v_ Oc 0> 0> E= TD D g o 2< O' i-- ^ < : CC 0 n 0) <D o c o T3 O 0 c o "S' uO CM 86 <5 O V-- c. . 0 V -- Amosil Crocid (AC) No da presen as AC com bi S coc O CM O <mT cvm oooo V.v g T.3 o o o 85 oo v.v a TD P 2 DoC O ^iCynMr oo -O r- y:v O o> c o ; G O a0! oo o r-s v.v a T5 P a 5 DC O er lO CM V Soo6 V 1-- .. V co - a? > o OI oo o V o-- .V co T> S Oo' Q > w ~oS 0 o| "' <n tC to . OOOO papo <2co *> 0 D 00 pa < < > to a oo a. <2 O -0 ^_ ,1 /> c DC oP 0 o'sr CO CM V 86 o ^ yv 0 TJ c 0 2 Q: a tr 0 Z "D a t t Oo aa w. *+-- 4-- oo zz t o o. o o o CM CM CM CM CM CM CM 1 CM 0 -J c g, CM o o O V n Csi w 5 -o SoZ. $ o O CM CO o oo o CM o CM o CM o.co *z circumstances Roggli does not adhere to his own seven fiber-years of exposure rule. For instance, Roggli will attribute asbestos causation in household exposure mesothelioma cases despite the fact that he doesn't '`know any way in a household-contact case to apply the [his] seven tea ten fiber cc year rule [for chrysotile] or the .01 fiber cc rule for amphibole fibers."148 Roggli's position is different from the legal standard as described by Keeton and Prosser, that: In products liability involving asbestos, where the plaintiff has sufficiently demonstrated both lung dis ease resulting from exposure to asbestos and that the exposure was to the asbestos products of manydifferent, but identified, suppliers, no supplier enjoys a causation defense solely on the ground that the plaintiff would probably have suffered the same disease from inhaling fibers originating from the products of other suppliers.175 and: When the conduct of two or more actors is so related to an event that their combined conduct, viewed as a whole, is a but-for cause of the event, and application of the but-for rule to them individually would absolve all of them, the conduct of each is a cause in fact of the event.175 10. Is the assumption that automobile mechanics do not work with brake and/or clutch products that contain amosite or crocidolite correct?136 In an introduction to a 1968 paper that reported asbestos exposures in brake mechanics, Ford Motor Company's industrial hygienists wrote, "The brake lin ings in current use may contain 40 to 60% asbestos when manufactured--the asbestos being normally in the chrysotile from, and occasionally in the amosite form."1,6 Borg Warner used crocidolite some automo bile clutches and brake bands.1" Maremont used croddolite in its automobile parts operation which pro duced brakes, clutches, and mufflers/78 Several brake and clutch manufacturing companies purchased amphibole fibers from the North American Asbestos Corporation, between 1954 and 1974/79-180 These include Bendix, Victor, Raybestos-Manhattan and Delco Moraine, a General Motors subsidiary. In a gov ernment review, Blau reported that manufacturing companies had used amosite and crocidolite in brakes/81 Some brake patents called for the use of either crocidolite or amosite (see Table 5). 11. Did researchers consider evidence of synergy between chrysotile and amosite? Roggli has attributed causation solely to amphi bole fiber, irrespective of the chrysotile or tremolite count, unless he can estimate a 7-10 year chrysotile exposure/48 /Ol 15/NO 2, APR/JUN 2009 www.ijoeh.com Fiber Types, Asbestos Potency, and Environmental Causation 221 TABLE 4 "Background" vs. Mesothelioma Patients, Adapted from Srebro, Roggii, and Samsa130 Patient Age/Sex Diagnosis0 Occupation13 Uncoated Fibers cufT^F Smoke0 AB/g AMOS5 TAAS CHRYSh Mesothelioma Cases 1 56/M BPL-R 2 60/M ERE 3 /M' EPL-R 4 60/M BPE 5 68/F DPL-L 6 52/M SPL-R 7 65/M EPL-R 8 77/M SPL-R 9 68/M EPL-L 10 67/M BPL-R 11 61/M EPL-R 12 65/M BPC 13 58/F DPL-L 14 66/M BPL-L 15 57/F PL 16 45/M PE : 17 18 53/M 45/M EPL-L EPL-R Painter/spackler Shipyard worker (WWi!) Brake repair Truck driver (vermiculite) No history of exposure Ship engine room/ brake repair Navy Sales Railroad machinist Merchant marine Carpenter Weigh station employee Teacher aide (building exposure) Brake repair Wife of shipyard worker Attorney (building exposure as student) Accountant (building exposure) No history of exposure 70 PY <15.0 S 14.4 14.0 132 PY <12,4 9,8 40 PY 7.1 52 PY 6,4 <5.2 35 PY <5.0 NS <3.0 20 PY <3,0 2.8 2.8 XS (16 yr) 2.6 40 PY 2.0 24 PY 1.0 <0.2 <3.0 <4060 60 1440 1360 <660 280 1070 <390 <510 <1080 <310 <690 <870 120 <4860 <1220 16,160 60 2170 <680 1980 550 7490 1570 <510 1080 <310 690 4330 240 9720 1220 <4060 <60 720 <680 <660 <280 4280 <390 <510 <1080 <310 <690 <870 <120 <4860 <1220 <640 <440 1270 640 <440 <440 Control Cases 19 64/M 20 76/M 21 40/M 22 61/M 23 64/M 24 64/M 25 59/M 26 53/M 27 71/M 28 61/M 29 51/M 30 53/M 31 28/M 32 36/M 33 67/M 34 71/M 35 64/M 36 85/M 37 63/M Ml ALL GBM Esophageal cancer Melanoma Alzheimer's Gastric cancer ABE CLL CAD Cirrhosis Hepatoma ALL Pancreatitis GBM ESRD Ml CVD AAA Hospital, farmer Manual labor Manual labor Truck driver Air force NA Guard Air force Music Garage owner Manual labor Spinning mil! Air Force NA NA Business supply store Electrical engineer Manual labor Hospital aide NS 22,0 19.6 9.7 8,9 NS 7.4 5.4 NS 3.5 NS 50 PY NS XS (pipe) NS XS 3.0 3.0 2.8 2.2 2.2 1.0 1,0 0.8 0.4 0.4 0.2 <0.2 <990 <1770 <100 <400 <570 <2540 <470 <760 <300 <170 <1000 <650 <960 <790 <430 <510 <370 <600 <600 <990 1770 210 400 <990 <1770 <100 <400 <570 <570 2540 <2540 470 <470 <760 890 <170 <1000 1310 960 <790 <430 <510 370 <600 <600 <760 <300 <170 1000 <650 <960 <790 <430 510 <370 . <600 <600 Key (Shading not in original) | = Case with single amosite fiber dismissed as "uncertain etioiogy" (Cases 2, 6, 14) ' ~] = Case with counts below or equal to controls dismissed as "uncertain etiology" (Cases 5,8- 12) j = Case with chrysotile levels above 95% of controls dismissed as "uncertain relationship" to asbestos (Case 17) __ I = Caused case aAAA, abdominal aortic aneurysm; ABE, acute bacterial endocarditis; ALL, acute lymphoblastic leukemia; B, biphasic; CAD, coro nary artery disease; CLL, chronic lymphocyctic leukemia; CVD, cardiovascular disease; D, desmoplastic; , epithelial; ESR, endstage renai disease; GBM, giioblastoma muitiforme; L, left; Mi, myocardial infarction; PC, pericardial; PL, pleural; PE, peritoneal; R, right; S, sarcomatous, bNA, not avaiiable. CNS, nonsmoker; PY, pack-years; S, smoker (unknown duration/quantity); XS, ex-smoker, aAB/g, asbestos bodies per gram of wet lung by light microscopy. eUF/g, total uncoafed fibers >5 pm (in length) per gram of wet lung by scanning electron microscopy. 'AMOS, amosite. SAA, tremolite, anthophylite, actinolite. hCHRYS, chrysotile. 'No age reported by Srebro, Roggii, Samsa 222 Egiiman www.ijoeh.com INI J OCCUP ENVIRON HEALTH TABLE 5 United States Brake Patents which Include Amphibole Fibers Patent Number Comment 2227424 Johns Manville patent for a brake lining includes the following description: "In the friction materials of the present Invention, heat resistant fibers of chrysotile, amosite, or other variety of asbestos fibers adapted for use in friction materials are used as the fibrous component associated with the friction composition," 2943010 Raybestos patent for laminated composite fabric break lining In which "examples of the types of asbestos fibers which are suitable for use in this process are chrysotile, crocidolite or amosite." 3624234 Raybestos-Manhaftan's patent for a friction material for use in "automotive and industrial brakes or clutches" called for typical materials including 20% anthophylite and 20% chrysotile asbestos, Year 1941 1960 1969 Roggii's fiber year estimates are based on the QAMA mine studies, whose exposure monitoring never meas ured fibers and whose results were no better than guesses.94 He has not examined the validity of this data.*** McDonald, the designer of the Quebec Asbestos Mining Association studies, in an unpublished paper made public in the Tobacco Company archives, wrote: . . . converting from particles to fibers a difficult and dubious operation. Even in chrysotile mining and milling, the range of conversion ratios is at least 40 fold. A problem of similar magnitude concerns the equivalence in fiber terms of measurements made in the general environment, nearly all of which are gravimetric and usually expressed in nanograms per cubic meter (ng/m)3. The conversion factor relat ing mass to optical fiber concentration had a range of 5-150 and probably varied with fibre type.182 Roggli claims the 7-10 fiber-year dose is the dose at which the mesothelioma rate doubles in miners.133 However, "doubling doses" are not required to estab lish a cause-effect relationship.183,184 This is particularly true when there is pathologic or historical evidence of exposure to asbestos. At any rate, as described above there is substantial evidence that chrysotile and amphiboles act super adclitively or synergistically even if chrysotile itself is not a complete carcinogen.128 Therefore, it is more likely that chrysotile is a contributing cause of mesothelioma when amphiboles are present as well. Roggli agrees and has testified that, "An amosite and chrysotile insulation worker would have about twice the risk of an amosite factory worker or crocidolite Australian mines [sic] or chrysotile amphibole textile factory workers."127 This opinion seems to have had no impact on his assessment of the potential contribution of chrysotile in American workers, almost all of whom have been exposed to both. Roggli has stated that the chrysotile-mesothelioma relationship has no threshold.127 This is inconsistent with his position that "background" exposures do not contribute to cause mesothelioma.127 There are rare mesotheliomas for which no point source of exposure in excess of that in the general environment can be identified. Such cases can be attributed to general envi ronmental "background" exposure, leaving aside the unsolvable issue of whether there exist any sponta neous mesotheliomas entirely unrelated to asbestos (this impossible to establish even in childhood cases since there is neonatal exposure to asbestos). 12. Do lung fiber types and levels predict or "drive" pleural levels? In 1992, Roggli asserted that, "[Tjhere is growing consensus that the fiber burdens that accumulate in the lung are the primary determinant of later dis ease."185 Wagner and Pooley offer the hypothesis that *"`*Roggli's claim that ac least 7 f/cc/years of exposure are required to attribute a mesothelioma to asbestos is based a "pocket risk assessment" relying on exposure and disease data from the QAMA studies. When faced with evidence that QAMA data was potentially unreliable, Roggli agreed that there was the potential for a "garbage-in, garbage out" phenomenon/35 Figure 12--The predominant fibers in the lung, based on data from: Suzuki Y, Yuen SR. Asbestos tissue burden study on human malignant mesotheiioma. Ind Health 2001:39:150-60. Comparison to Figure 13, which shows pleural fibers in the same patients, establishes that chrysotile goes to the pleura, while amphiboles stay in the lung. VOL 15/NO 2, APR/JUN 2009 www.ljoeh.com Fiber Types, Asbestos Potency, and Environmental Causation 223 "those diseases associated with exposure to mineral fiber are due to the fiber retained in the lungs," but do not address the relationship between lung and pleural fiber burdens.180 Churg's finding contradicts Roggli's assertion, as he notes: . A different approach is to examine fiber burden in lung tissue of patients with mesothelioma. This pro cedure ensures that mesotheliomas induced by occult amosite or crocidolite exposure will be detected as such, but it suffers from unknown pat terns of fiber clearance over time and also from the fact that, while amphibole accumulates readily in lung, chrysotile does not.187 13. Is fiber counting reliable? Fiber counting is unreliable due to wide jntra- and inter-laboratory variability. It is a non-standard tech nique that cannot be used to determine causation in individual cases. Roggli has written: The wide variety of preparative techniques and ana lytical methodologies that have been employed by various investigators make it difficult to extrapolate results from one laboratory to another. The actual analytical result obtained on any one sample can be profoundly influenced by the steps employed in the analytical procedure. Interlaboratory comparison trials demonstrate that striking differences can occur among laboratories even when the same sample is analyzed.132 He further states, fn addition to inter-laboratorv variation, intralabo ratory variation can occur, which may be due either to changes in a laboratory's procedures over time, or to variation in fiber content from one site to another within the lung. Morgan and Holmes have reported a five to tenfold site-to-she variation based on analyses of multiple samples from a single lung using phase contrast light microscopy.152 There is also sampling variability, fn testimony, Roggli has agreed that the numbers of fibers can vary from site to site within the lung by a factor of anywhere from two to five.527 In addition, he has agreed that there are as many as ten short chrysotile fibers for every one he can count >5pm.12/ Few scientists have the temerity to overlook test variability of this magnitude to conclude anything about a scientific theory or individ ual causation--except in the case of a finding that in and of itself disproves a theory (identification of a black swan disproves the theory that all swans are white). This 15,000% lack of precision would appear to violate the Daubert standard for "reliability."188 H Chrysotile in pleura H only amphibole in pleura amphibole in lung + no asbestos fiber in pleura Figure 13---The predominant fibers in the pleura, based on data from: Suzuki Y, Yuen SR. Asbestos tissue burden study on human malignant mesothelioma. Ind Health 2001;39:150-60. Comparison to Figure 12, which shows lung fibers in the same patients, establishes that chrysotile goes to the pleura, while amphiboles stay in the lung. Amphiboles were found in the pleura in 23.5% of cases. 14. Do fibers differentially locale to the pleura in a way that lung counts systematically underestimate chrysotile pathogenicity? Short thin chrysotile fibers are the most commonly found fiber in tumors and in the pleura of patients with mesothelioma, asbestosis. and lung cancer (see Figures 12 and 13).104 Roggii's method systematically underes timates or misses chrysotile exposure while overem phasizing amphibole exposures. Except for workers who fabricated Unibestos products and some individu als who had only environmental tremolite exposures (e.g. Libby, Montana residents), all exposures to US residents included, at least some chrysotile exposure. Because Roggli systematically underestimates chryso tile exposures and because chrysotile is always over represented in the pleura compared to the lung, Roggli's results often attribute causation to the wrong fiber and they are almost always misleading. In this case, more information (lung fiber count versus no count) is worse than no information--it is misleading. Roggli has also claimed that: pleural tumor fiber counts are unreliable without stating why this is so.536 This is only true if fibers appear to absent, as the tumor can dilute the fiber concentration (absence of evidence is not evidence of absence). If fibers are found in a tumor or plaque, however, this is always important.109 In contrast to Teta et al. and Price and Ware, Azuma et al. consider fiber burdens, exposure data, and mesothelioma rates in their study design and thus pro vide evidence that low exposures to asbestos cause "background" cases. Their results are comparable to those of Iwatsubo et al.,174 Rodelsperger et al,189111 Mag- Tt'R6deisperg'er used lung cancer controls, thus conflating his analysis of relative potency of different fibers, but his data on the exposed population is consistent with a no-t.hreshoid dose-response model. 224 Egilman www.ijoeh.com INT J OCCUP ENViRON HEALTH nani et al.,3'J>35 and Maulc et al.36 Human pathologic studies of pleural tissue that do not exclude `Inconven ient" data buttress this conclusion. CONCLUSION This is the first peer-reviewed publication of which we are aware that "peer reviews" testimony. In our view, review of the presentation of scientific ideas that are presented in court and at hearings is at least as impor tant as peer review of published research and academic reviews. We look forward to publishing other similar reviews in the future and encourage our readers to submit them. i want to thank David Madigan, PhD for providing the statistical analysis of data from Srebro et al., William Longo. PhD for use of his data, and Susanna Bohme, PhD for her editorial advice. References 1. JAMA Editor Calls Critic a 'Nobody and a Nothing.' 2009. [cited 2009 March 3] Available from: http://blogs.wsj.com/heakh/ 2009/03/13/jama-cdi tor-calls-cri tic-a-nobody-and-a-noihing/.) 2. Wimiein K, Armstrong D. "Top Pain Scientist Fabricated Data in Studies", Hospital Says. Wail Streetjournal. 2009 Mar] 1,2009. 3. Robinson RG, Jorge RE, Moser Dj, et al. Escitalopram and prob lem-solving therapy for prevention of poslstroke depression: a randomized controlled trial. JAMA. 2008;299:2391-400. 4. DeAngelis CD, Fontanarosa PB. Conflicts Over Conflicts of Interest. JAMA 2009;301:2. 5. Robinson RG, Arndt S. Incomplete financial disclosure in a study of escitalopram and problem-solving therapy for preven tion of poststroke depression. JAMA. 2009;301:1023-4. 6. Azuma K, Uchiyama I, Chiba Y, Okumura J. Mesothelioma Risk and Environmental Exposure to Asbestos; Past and Future Trends in Japan. IntJ Occup Environ Health. 2009;15:6. 7. Hessell P. Inaccuracies in Article on Mesothelioma in Brake Workers [Letter]. Int J Occup Environ Health. 2009;15. 8. Welch L, Anderson HA, Balmes J, et al. Research on Mesothe lioma from Brake Exposure: Corporate Influence Remains Rel evant Concern [Letter], IntJ Occup Environ Health. 2009; 15. 9. Mark Ej, Yokoi T. Absence of evidence for a significant back ground incidence of diffuse malignant mesothelioma apart from asbestos exposure. Ann N YAcad Sci. 1991;643:196-204. 10. Camus M, Sicmiatycki J, Meek B. Nonoccupadonal exposure to chrysotile asbestos and the risk of lung cancer. N Engl J Med. 1998;338:1565-71, '' 11. Egilman D. Fehnel C, Bohme SR. Exposing the "myth" of/ABC, "anything but chrysotile": a critique of the Canadian asbestos mining industry and McGill University chrysotile studies. Am J Ind Med. 2003;44:540-57. '' 12. Gustavsson P, Nyberg F, Pershagen G, et al. Low-dose exposure to asbestos and lung cancer: dose-response relations and inter action with smoking in a population-based case-referent study in Stockholm, Sweden. AmJ Epidemiol. 2002;155:1016-22. 13. Pan XL, Day HW, Wang W, et al. Residential proximity to natu rally occurring asbestos and mesothelioma risk in California. Am j Respir Grit Care Med 2005;172:1019-25. 14. Price B, Ware A. Mesothelioma trends in the United States: an update based on Surveillance, Epidemiology', and End Results Program data for 1973 through 2003. Am j Epidemiol. 2004; 359:107-12. '' 15. Tela MJ, Mink PJ, Lau E, el al. US mesothelioma patterns 1973 2002: indicators of change and insights into background rates. Eur J Cancer Prev. 2008;17:525-34. 16. Price B. Analysis of current trends in United States mesothe lioma incidence. Am [ Epidemiol. 1997;145:211-8. 17. Patel AV, Bogner PN, Klippenstein D, Ramnath N. Malignant pleural mesothelioma after household exposure to asbestos. J Clin Oncol. 2008;26:5480-3. ' 18. Miller A. Mesothelioma in household members of asbestos- exposed workers: 32 United States cases since 1990. AmJ Ind Med. 2005;47:458-62. 19. Leigh J, Davidson P, Henclrie L, Berry D. Malignant mesothe lioma in Australia, 1945-2000. AmJ Ind Med. 2002;41:188-201. 20. Casdeman BI, Berger SI,. Asbestos : medical and legal aspects. 5f.h ed. New York, NY: Aspen Publishers; 2005. 21. Schneider J, Rodelsperger K, Bruckel B, et al. Pleural mesothe lioma associated with indoor pollution of asbestos. J Cancer Res Clin Oncol. 2001;127:123-7. 22. Schneider J, Siraif K, Woitowirz Hj. Pleural mesothelioma and household asbestos exposure. Rev Environ Health. 1996;11:65- 70. 23. Health and Safety Executive. Mesothelioma Occupation Statis tics: Male and Female Deaths Aged 16-74 in Great Britain. Suf folk: HSE Books; 2006. 24. Egilman DS, Billings MA. Abuse of epidemiology: automobiie manufacturers manufacture a defense to asbestos liability. IntJ Occup Environ Health. 2005;! 1:360-71. ' 25. Seidman H. Selikoff IJ, Hammond EC. Short-term asbestos work exposure and long-term observation. .Ann N YAcad Sci. 1979;330:61-89. " 26. Lin RT, Takahashi K. Karjalainen A, et al. Ecological association between asbestos-related diseases and historical asbestos con sumption: an international analysis. Lancet. 2007;369:844-9. 27. Welch I.S. Asbestos exposure causes mesothelioma, but not this asbestos exposure: an amicus brief to the Michigan Supreme Court. Int ] Occup Environ Health. 2007;13:318-27. 28. Welch LS, Ackerman YI, Haile E, Sokas RK, Sugarbaker PH. Asbestos and peritoneal mesothelioma among college-educated men. IntJ Occup Environ Health. 2005;11:254-8. 29. Mi.tsii M, Pollice A, Cavone D, Dragonicri S, Bilanda M. The relationship between malignant mesothelioma and an asbestos cement plant environmental risk: A spatial case-control study in the city of Bari (Italy). Int Arch Occup Environ Health. 2009; 82:489-97. ' 30. Berry M. Mesothelioma incidence and community asbestos exposure. Environ Res. 1997;75:34-40. 31. Greenberg M, Davies TA. Mesothelioma register 1967-68. Br J Ind Med. 1974;31:91-104. " 32. Mollo F, Magnani C. European multicentric case control study on risk for mesothelioma after non-occupalionai (domestic and environmental) exposure to asbestos. Med Lav. 1995;86:496- 500. 33. Magnani C, Ivaldi C, Bolta M, Terracini B. Pleural malignant mesothelioma and environmental asbestos exposure in Casale Monferralo, Piedmont: Preliminary analysis of a case-control study. Med Lav. 1997;88:302-9. ' ' 34. Magnani C, Aguclo A, Gonzalez CA, et al. Multicentric study on malignant pleural mesothelioma and non-occupational expo sure to asbestos, BrJ Cancer. 2000:83:104-11. 35. Magnani C, Dalmasso P, Biggeri A, et al. Increased risk of malig nant mesothelioma of the pleura after residential or domestic exposure to asbestos: a case-control study in Casale Monferralo, Italy. Environ Health Perspect. 2001;109:915-9. 36. Maule MM, Magnani C, Dalmasso P, et al. Modeling mesothe lioma risk associated with environmental asbestos exposure. Environ Health Perspect. 2007;115:1066-71. 37. Vianna NJ, Polan AIL Non-occupational exposure to asbestos and malignant mesothelioma in females. Lancet. 1978;1:1061-3. 38. Chahinian AP, Pajak TF, HollandJF, Norton et al. Diffuse malig nant mesothelioma. Prospective evaluation of 69 patients. Ann Intern Med. 1982;96:746-55. 39. Kiiburn KIT, Lilis R, Anderson HA, et al. Asbestos disease in family contacts of shipyard workers. Am J Public Health. 1985;75:615-7. 40. Slrickler HD, GoedertJJ. Devesa SS, LahevJ, Fraumeni JF, Jr., Rosenberg PS. Trends in U.S. pleural mesothelioma incidence rates following simian virus 40 contamination of early poliovirus vaccines. J Natl Cancer Inst. 2003;95:38-45. 41. Health and Safety Commission. Health and Safety Statistics 1998/99. Sudbury: HSE Books; 1999. 42. Australian Safety and Compensation Council. Preparing an Esti mate of the National Pattern of Exposure to Asbestos in Cases of Malignant Mesothelioma. Australian Government Department of Education. Employment and Workplace Relations; 2008 June. VOL 15/NO 2, APR/JUN 2009 www.ijoeh.com Fiber Types, Asbestos Potency, and Environmental Causation 225 43. Mirabelli D, Stura A. Gangemi M, Bertolotti M, Maule MM, Magnani C. [incidence of malignant mesothelioma in Pied mont, 1990-2003], Epidemiologia e prevenzione. 2007:31:132-8. 44. Roggli VL, Oury TD, Moffati Ef. Malignant mesothelioma in women. Anal Pathol. 1997;2:147-63. 45. Roggli VL, Sharma A, Bittnor KJ, Sporn T, Vollmer RT. Malig nant mesothelioma and occupational exposure to asbestos: A clinicopathological correlation of 1445 cases. Uhrastruct Pathol. 2002;26:55-65. 46. World War II in the San Francisco Bay Area [Internet]. National Park Service, US Department of the interior; n.d. Tending the Home Front: The Many Roles of Bay Area Women During WW11 [cited 2009 Feb 27], Available from: http://www.nps.gov/his- lory/Nr/travel/wwllbayarea/womenatwar.hun. 47. Hassan A. Shipyard Day Care Centers of World War II: The Kaiser Experiment 2009 [Internet].[cited 2009 April 2]. Avail able from: http://wwiishipvarddaycare.tripod.com/intro.htm. 48. Wikipedia [Internet]. Wikimedia Foundation. [2009]. Asbestos [cited 2009 February 20]. Available from: http://en.wikipedia. org/'wiki/Asbestos/. 49. Greenberg M. Asbestos Exposures to Women. Personal Com munication to: David F.gilman. 2009. 50. Brown DP, Dement JM, Okuri A. Mortality patterns among female and male chrvsotile asbestos textile workers. ] Occup Med. 1994;36:882-8. ' ' 51. Merewether ERA, Price CW. Report on the Effects of Asbestos Dust, on the Lungs and Dust Suppression in the Asbestos Indus try. HM Stationary Office; 1930. 52. Finkeistein MM. A study of dose-response relationships for asbestos associated disease. Br J Ind Med. 1985;42:319-25. 53. Saldivar A, Solo V. Asbestos in the United Stales: Occurrences, Use and Control 2008 [Internet]. Proquest; 2008. [cited 2009 February 26]. Available from: hup://www.csa.com/discov- eryguides/asbestos/review, pdf. 54. Roggli V. Deposition. Ralph Stanley vs, Anderson ei ah, Civil Action No. 03-C-9600, Kanawa County, WV; 2005:80. 55. Schneider A. U.S. imports of asbestos brake material arc on rise. St Louis Post Dispatch. 2003 Oct 26. 56. Seminario P. Testimony Before the Senate Employment and Worker Safety Subcommittee of Health, Education, Labor and Pensions Committee Hearing on "Is OSHA Working for Work ing People." 2007 April 26. Available from: ht.tp://www.al']do. org/issues/safety/up]oad/SeimnarioOSHA20070426,pdf. 57. Morton S. Deposition Testimony. Morton vs Owens Illinois et al, Newport News Virginia. 2006. 58. United States Congress. Flearing Transcript. Oversight Of The Navy Department's Enforcement Of OSHA; 1983. 59. Glass A. (President, Steelworkers Union, Newport News Ship yard) Personal Communication to: Robert Hatien R; 2009. 60. Harries PC. Asbestos dust concentrations in ship repairing: A practical approach to improving asbestos hygiene in naval dock yards. Ann Occup Hvg, 1971;14:241-54. 61. Garcia A. OSHA Violations Indicate Enforcement of Asbestos Standard [Internet] Future Environment Designs; 2008 Sep tember 05 [cited 2009 Feb 20], Available from: http:// futureenv. blogspot.com/2008/09/osha-violat.ions-imiicate- enforcement-of.html 62. Saltzman J. Trainer in asbestos removal convicted, Boston Globe, 2008 November 21. 63. Oversight Report: OSIIA's Failure To Monitor And Enforce Asbestos Regulations In Auto Repair Shops. Washington. DC: U.S. House of Representatives, Congressman Dennis Kudnich; 2004 Mar. [cited 2008 Feb 2], Available from: hitp://kudnich. house.gov/UploadedFiles/OSFlA.pdf.) 64. Haque AK, Hernandez JC, Dillard F.A, 3rd. Asbestos bodies found in infant lungs. Arch Pathol Lab Med. 1985; 109:212. 65. Haque AK, Vrazei DM, Uchida T. Assessment of asbestos burden in the placenta and tissue digests of stillborn infants in South Texas, Arch Environ Contain Toxicol. 1998;35:532-8. 66. Ampleford EJ, OharJ. Mesothelioma: You do not have to work for it. Diagn Cytopathol. 2007;35:774-7. 67. Barton H, Cogliano J, Firestone M. et al. Supplemental Guid ance for Assessing Susceptibility from Early-Life Exposure to Carcinogens. EPA/630/R-03/003F. 2005 March. Available from: hup:././www.epa.gov/un/atw/childrens_supplemeni_ final.pdf. 68. Lee CL. Removing Dust From Brakes Assemblies During Vehi cle Servicing- Alternative Cleaning Methods. Annals Occup Hvg. 1970;13:33-6, ' 69. Boillat MA, Lob M. Risque d'asbestose chez les travailleurs occupes a remplacer les garniture de freins [Risk of asbestosis in workers employed in replacing automobile brake linings]], Schweiz Med. 1973;103:1354-9. ' 70. Casileman B, Ziem G. Control Technology for Brake and Clutch Work. In: Casileman B, ed. Asbestos: medical and legal aspects. 5th ed. New York: Aspen Publishers; 1985. 71. Hatch D. Possible Alternatives to Asbestos as a Fraction Mater ial. Ann Occup Hyg, 1970;13:25-9. 72. Rodelsperger K, Jahn FI, Bruckel B, et al. Asbestos Dust Expo sure During Brake Repair. AmJ of Indus! Med, 1986;10:63-725. 73. Kauppien T, Rorhonen K. Exposure to Asbestos During Brake Maintenance of Automotive Vehicles by Different Methods. Am Ind Hyg Assoc J. 1987:48:499-504. 7-1. Hickish DE. Exposure to Asbestos Dust during Brake Mainte nance Operations on Commercial Vehicles, Fleet Repair Garage, Dagenham. Report 41/68. 1968 October. Produced by Ford in Coates. Plaintiff Exhibit FD5036, http://www.egilman, com/Documents//Asbestos/Brakes/Ford/amosite%20and%2 0brakes/hickish%20amosit.e%20and%20brakes.pdf 75. Hickish DE. Exposure to Asbestos during the Servicing of Brake of Passenger Cars, Report 52/68. Produced by Ford in Coates. Plaintiff Exhibit FO-55. Available from: hu.p://www.egilman. com/Documents//Asbesios/Brakes/Ford/Knowledge?<,20of% 20Excess%20Exposures/hickish%20secrel%20october%20arno sitc%201968.pdf. 76. Clark AR. .Asbestos Exposure Associated with Automotive Brakes Systems, Union Carbide Internal Correspondence. 1976 July 8. 77. Hatfield RL, Longo WE Bendix Brakes for Chrysler Vehicles: Work Practice Simulation Demonstration. Suwanee, GA: Mate rials Analytical Services, Inc; n.d. 78. Hatfield RL, Longo WE, Bendix Brakes for Ford Vehicles: Work Practice Simulation Demonstration. Suwanee, GA: Materials Analytical Services, Inc; n.d. 79. Hatfield R. Longo W. Arc: Grinding II Cleanup. Raleigh, NC: Materials Analytic Services, Inc.; 2000. 80. Hatfield N, Longo W, Newton LR. Arc Grinding of Brake Shoes: Work Practice Study. Raleigh. NC: Materials Analytical Services, Inc.; 2000. 81. Hatfield N, Longo W, Newton LR. Hand Grinding of Brake Shoes: Work Practice Study. Raleigh. NC: Materials Analytical Services, Inc.; 2000. 82. Hatfield N, Newton LR, Longo W. Hand Sanding of Brake Shoes: Work Practice Study. Raleigh, NC: Materials Analytical Services, Inc:.; 2001. 83. Rohl .AN, 1 .anger AM, Klimentidis R, et al. Asbestos content, of dust encountered in brake maintenance and repair. Proc R Soc Med. 1977;70:32-7. 84. Osborn SH. Forty-Ninth Report of the State Department, of Health for the Year Ended June 30, 1934. State of Connecticut Public Document No, 25. Hartford, CT: Bureau of Occupa tional Diseases; 1934. 85. Roberts DR, Zumwaide RD. Industrial Hygiene Summary Report of Asbestos exposure assessment for brake mechanics. Cincinnati, OH: Industrial Hygiene Section, Industrywide Stud ies Branch, Division of Surveillance, Hazard Evaluations and Field Studies, National Institute for Occupational Safety and Health; 1982 November 22. Report No.: 32.4. ' 86. Lloyd JW. Current Intelligence Bulletin 5: Asbestos: Asbestos exposure during servicing of motor vehicle brake and clutch assemblies. Cincinnati, OH: National Institute for Occupational Safer)' and Health; 1975 Aug 8. Available from: hup://www.cdc. gov/N) OSH/78127_5.htm! 87. Longo WE, Mount MD, Hatfield RL. Occupational Exposure to Asbestos-Containing Brake Shoes Dui'ing Hand Grinding and Sanding. Raleigh, NC: Materials Analytical Services. Inc.; 2004. ' 88. Hodgson JT, Darnlon A. The quantitative risks of mesothelioma and lung cancer in relation to asbestos exposure. Ann Occup Hyg. 2000;44:565-601. 89. Berman DW, Crump KS. A meta-analysis of asbestos-related cancer risk that addresses fiber size and mineral type. Grit Rev Toxicol. 2008;38 Suppl 1:49-73. 226 Egilman www.ijoeh.com . iNT J OCCUP ENVIRON HEALTH 90. Berman DW, Crump KS. Update of potency factors for asbestos- related lung cancer and mesothelioma. Grit Rev Toxicol. 2008;38 Suppl 1:1-47. 91. Kane A. (Chair, SAB Asbestos Committee). Letter to Johnson S. (EPA Adminsirator), SAB Consultation on EPA's Proposed Approach for Estimation of Bin-Specific Cancer Potency Factors for Inhalation Exposure to Asbestos Cancer Potency Factors for Inhalation Exposure to .Asbestos 2008, EPA-SAB-09-004 Available from: hup://yosemite.epa.gov/sab/sabproduc:t.n.sf/77CFF6439C OOA.BF3852575010077801 F/$File/F.PA-SA.B-09-004-urisigncd.pdf. 92. Selikoff IJ, Seidman H. Asbestos-associated deaths among insu lation workers in the United States and Canada, 1967-1987. Ann N YAcad Sci. 1991;643:1-14. 93. Rogers A. Major G. The quantitative risks of mesothelioma and lung cancer in relation to asbestos exposure: The Wittenoom data. Ann Occup Hyg 2002;46:127-8; author reply 8-9. 94. Gibbs G, LaChance M. Dust-fiber relationships in the Quebec chrysotile industry. Arch Environ Occup Health. 1974;28:3. 95. Shapiro HA. Pneumoconiosis; Proceedings of the international conference, Johannesburg, 1969. Cape Town, New York,: Oxford University Press; 1970. 96. Hodgson }, Darnton A. Letter. Ann Occup Hyg. 2001;45:336-8. 97. Leigh J, Robinson B. The history of mesothelioma in Australia 1945-2001. In: Robinson B, Chahinian A, eds. Mesothelioma. London: Martin Dunitz; 2000:55-86. 98. Henderson DW. Supplementary Report, On Causation, for Mr. Frank Gregory Lanslcy. Adelaide, South Australia; 2006 Febru ary 27. 99. Costas E, Garrido A, Goyanes VJ. Nonoccupational exposure to chrysotile asbestos and the risk of lung cancer. N Engl J Med. 1998;339:1000; author reply 1-2. 100. Hein MJ, Stayner LT, Lehman E, DementJM. Follow-up study of chrysotile textile workers: cohort mortality and exposure- response. Occup Environ Med. 2007;64:616-25. 101. Peto J, Rake C, Gilham C, Hatch J. Occupational, domestic and environmental mesothelioma risks in Britain: A case-control study. Health and Safely Executive; 2009. Research Report No.: 696. Available at: http://www.hse.gov.uk/research/rrpdf/rr696.pclf. 102. Paustenbach D. Deposition Testimony. Gray vs.John Crane, Inc., et al. , Virginia Circuit Court for the City of Newport. News, Case Number CL0800724PT; 2009 Apr 6. 103. Pierce JS, McKinley MA, Paustenbach DJ, Finley BL. An Evalua tion of Reported No-Effect Chrysotile Asbestos Exposures for Lung Cancer and Mesothelioma, Grit RevTox. 2008; 38; 191-214. 3 04. Suzuki Y. Yuen SR. Asbestos tissue burden study on human malignant, mesothelioma. Ind Health. 2001;39:150-60. 105. Sebastien P, Janson X, Gaudichet. A, Hirsch A, Bignon J. .Asbestos retention in human respiratory tissues: comparative measurements in lung parenchyma and in parietal pleura. IA.RC Sci Publ. 1980:237-46. 106. Dodson RF, Williams MG, Jr., Corn CJ, et al. A comparison of asbestos burden in lung parenchyma, lymph nodes, and plaques. Ann N YAcad Sci. 1991;643:53-60. 107. LeBouffant L. Investigation and analysis of asbestos fibers and accompanying minerals in biological materials. Environ Health Perspect. 1974;9:149-53. 108. Kohyama N, Suzuki Y. Analysis of asbestos fibers in lung parenchyma, pleural plaques, and mesothelioma tissues of North American insulation workers. Annals NYAcad Sci. 1991;643:25. 109. Suzuki Y, Yuen SR, Ashley R. Short, thin asbestos fibers contribute to the development of human malignant mesothelioma: Patho logical evidence. IniJ Hyg Environ Health. 2005;208:201-10. 110. Boutin C, Dumortier P, Rey F, el al. Black spots concentrate onco genic asbestos fibers in the parietal pleura. Thoracoscopic and mmeralogic: study. Am { Respir Crit Care Med. 1996;153:444-9. 113. Muller KM, Schmitz I, Konstantinidis K. Black spots of the pari etal pleura: morphology and formal pathogenesis. Respiration. 2002;69:263-7. ' 312. Mitchev K, Dumortier P, Do Vuyst P. `Black Spots' and hyaline pleural plaoues on the parietal pleura of 150 urban necropsy cases. Am j Surg Pathol. 2002;26:1198-206. 113. Wagner JC, Berry G. Mesotheliomas in rats following inocula tion with asbestos. Br J Cancer. 1969;23:567-81. 114. Wagner JC, Berry G, Timbrell V. Mesotheliomaia in rats after inoculation with asbestos and other materials. Br j Cancer. 1973;28:173-85. ` 115. Pott F, Hutb F, Friedrichs RH. Tumorigenic effect of fibrous dusts in experimental animals. Environ Health Perspect. 1974;9:313-5. 116. Wagner JC, Berry G, Skidmore JW, et al. The effects of the inhalation of asbestos in rats. Br J Cancer. 1974;29:252-69. .117. Appel JD, Fasv TM, Kohcz DS, et al. Asbestos fibers mediate transformation of monkev cells by exogenous plasmid DMA. Proc Natl Acad Sci USA. 1988;85:7670-4. 118. Adamson IY, BakowskaJ, Bowden DM. Mesothelial cell prolifer ation: a nonspecific response to lung injurv associated with fibrosis. Am J Respir Cell Mol Biol. 1994;10:253-8. 119. Adamson IY Early mesothelial cell proliferation after asbestos exposure: in vivo and in vitro studies. Environ Health Perspect 1997;105 Suppl 5:1205-8. 120. Kumar-Singh S, Weyler J, Martin MJ, Vermeulen PB. Van Marck E. Angiogenic cytokines in mesothelioma: a study ofVEGF, FGF- 1 and -2, and TGF beta expression. J Pathol. 1999:189:72-8. 121. DeLong P, Carroll RC, Henry AC, et al. Regulatory T cells and cytokines in malignant pleural effusions secondary to mesothe lioma and carcinoma. Cancer Biol Ther. 2005:4:342-6. 122. Yoshimoto A, Kasahara K, Saito K, Fujimura M, Nakao S. Gran ulocyte colony-stimulating factor-producing malignant pleural mesothelioma with the expression of other cytokines, fnt] Clin Oncol. 2005;10:58-62. ` 123. Hegmans JP, Hemmes A, Hammad IT, et al. Mesothelioma envi ronment comprises cytokines and T-regulaiory cells that sup press immune responses. Eur RespirJ. 2006;27:1086-95. 124. Roggli VL, Vollmer RT. Twenty-five years of fiber analysis: what have we learned? Hum Pathol. 2008;39:307-15. 125. Sebastien P, Masse R, Bignon J. Recovery of ingested asbestos fibers from the gastrointestinal Ivmph in rats. Environ Res. 1980;22:201-16. ' 126. Acheson ED, Gardner MJ. Mesothelioma and exposure to mix tures of chrysotile and amohibole asbestos. Arch Environ Health. 1979;34:240-2. ` 127. Roggli VL. Deposition Testimony. Wanda T. Jones, v. Dana Cor- noraiion, et al.. Circuit Court For The City Of Newport News Virginia, Law No. 39028T-01.; 2006 Jan 13. ' 128. Roggli VL, Sharma A, Butnor KJ, Sporn T, Vollmer RT. Malig nant mesothelioma and occupational exposure to asbestos: A clinicopathological correlation of 1445 cases. L'ltrasLruct Pathol. 2002;26:55-65. 129. Butnor Kj. Sporn TA, Roggli VL. Exposure to brake dust and malignant mesothelioma: a studv of 10 cases with mineral fiber analyses. Ann Occup Hyg. 2003:47:325-30. 130. Srcbro SH, Roggli VL, Samsa CP. Malignant mesothelioma asso ciated with low pulmonary tissue asbestos burdens: a light and scanning electron microscopic analysis of 18 cases. Mod Pathol. 1995;8:614-21. ' 131. Roggli VL. Affidavit, Exhibit 1. Bretzke vs. Ford Motor Com pany, District Court of Ramsey Countv, Minnesota Court File No. 62CV081189. 2008 Sept 13. ' 132. Roggli VL. Comments Regarding The U.S. EPA's Proposed Approach for the Estimation of Bin-Specific Cancer Potency Factors for Inhalation Exposure for Asbestos; 2008. http:// yosemite.epa.gov/sab/sabproduct.nsf/33EEF78BBA9E6D0085 257491006E9276/$File/Asbesios-t-Public-Commems+VicLor+R oggli+July+21-22+2008-fMeeting.pdf. 133. Grave Robbers: Chrysler Subpoenas Body At Funeral [Inter net], W'CBSTV. 2009 Mar 7. [cited 2009 Mar 14]. Available at: http://wcb.stv.eom/topstories/graveside.subpoena.harold.2.95 2746.html 134. Order on Emergent Application. Harold St.John and Florence St. John v. Affmia Group, Inc., et al. Superior Court of New Jersey Appelate Division. 2009 Mar 18. [cited 2009 Mar 29]. Available from: http://amiawdaily.typepad.com/stjohnappeaL pdf. 135. Roggli VI- Deposition testimony, Paz vs AW Chesterton et ah, Circuit Court, Third Judicial Circuic ,Madison County, Illinois. 1/27/2006 136. Roggli VL. Testimony. Harold R. St.John And Florence Diane St.John vs. Aftinia Group, Incorporated, et al., Superior Court Of New Jersey, Law Division: Middlesex County Docket No. Mid- L-5110-08; 2009 Mar 11. ' 137. Roggli VL.. Testimony in Lilly v. John Crane et al, Circuit Court of Newport News, Virginia, Case No. 38861AF. 2001 Sepr 19. VOL 15/NO 2, APR/JUN 2009 www.iioeh.com Fiber Types, Asbestos Potency, and Environmental Causation 227 138. Roggli VI,. The Friction Products Debate (PowerPoint Presen tation). Defense Research Institute; 2008 Nov. 139. Egilman DS, Billings MA. Differential peeky bias, lntj Occup Environ Health. 2006;12:294. 140. Roggli VL, McGavran MH, Subach J, et a!. Pulmonary asbestos body counts and electron probe analysis of asbestos body cores in patients with mesothelioma: a study of 25 cases. Cancer. 1982; 50:2423-32. ' 141. Ingram P, Shelburne JD, Roggli VL. Microprobe analysis in medicine. New York: Hemisphere; 1989. 142. Roggli VI..,, Longo WE. Mineral fiber content of lung tissue in patients with environmental exposures: Household contacts vs. building occupants. Ann N YAcad Sci. 1991;64S:511-8. 143. Roggli VL, Pratt PC. Brody AR. Asbestos content of lung tissue in asbestos associated diseases: a study of 110 cases. Br J Ind Med. 1986;43:18-28. 144. Roggli VL. Testimony. Robert Frank Smith and Mary Lou Smith v. Chrysler L.LC. et ah, Los Angeles Superior Court Docket No. 396072. 03/13/2009 145. Work-Related Lung Disease Surveillance Report, 2002. Cincin nati, OH: National Institute for Occupational Safety and Health; 2003. Report No.: 2003-111, 146. Henderson DW. Malignant mesothelioma. New York: Hemi sphere Pub. Corp.; 1992. 147. Roggli VL. Questions on your paper response. Email Commu nication to: David Egilman; 2009 March 17. 148. Roggli VL. Deposition. Windnagle vs Georgia Pacific et a!.. Cir cuit Court Of The 11th Judicial Circuit. In and For Miami-Dade County Florida CASE NO: 06-28424 CA 42; 2007 Nov 20. 149. Roggli VL, Environmental asbestos contamination: What are the risks? Chest. 2007;131:336-8. 150. Unknown. Regression by Maximum Likelihood Estimation for Left-censored Daia [Internet] Available from: http://rss.acs. unt.edu/Rdoc/library/NADA/html/cenmle.html. 151. Toxicological Profile for Asbestos [Internet]. Agency for Toxic Substances and Disease Registry: 2001 Sept. Available from: http://www.arsdr.cdc.gov/ioxprofiles/tp61 .htmi . 152. Roggli YE,. Oury TD, Sporn TA. Pathology of asbestos-associated diseases. 2nd ed. New York: Springer; 2004. 153. Environmental Protection Agency. Method for the Determina tion of Asbestos in Bulk Building Materials. US Environrnentqal Protection Agency. 1993 July. Report No.: EPA/600/R-93/116. 154. Environmental Protection Agency. Asbestos Hazard Emergency Response Act. 40 CFR Pan 763, Asbestos-containing materials in schools. 1987 October 30. 155. National Institute of Occupational Safety and Health. Asbestos by TEM. NIOSH Manual of Analytical Methods (NMAM), Fourth Edition. 1994 August 15. 156. Standard test method for airborne asbestos concentration in ambient and indoor atmospheres as determined by transmis sion electron microscopy direct transfer (TEM). ASTM; 1998. 357. Standard Test Method for Microvacuum Sampling and Indirect Analysis of Dust by Transmission Electron Microscopy for Asbestos Structure Number Surface Loading: Dust Sample analysis- ASTM D 5755-03. 2003 Oct 1. ' 158. Chatfield FJ. Ambient Air, Determination of Asbestos Fibers, Indirect-Transfer Transmission Electron Microscopy Method. International Standard Organlzation;1999. 159. Chatfield Ej. Berman DW. Interim Superfund Method for the Determination of Asbestos in Ambient Air, Part 1: Method. U.S. Environmental Protection Agency. Publication No: 540/2- 90/005a; 3990 May. 160. Loomis D, DementJM, WolfSH, et al. Lung Cancer Mortality and Fiber Exposures among North Carolina .Asbestos Textile Workers. Occup Environ Med. 2009 Mar 11 [epub ahead of print]. 161. Stayner LT. Canada, chrysotile and cancer: Health Canada's asbestos international expert panel report. J Occup Environ Med. 2008;50:1327-8. ' 162. Roggli VL, Sanders LL. Asbestos content of lung tissue and car cinoma of the lung: a clinicopathologic correlation and mineral fiber analysis of 234 cases. Ann Occup Hyg. 2000;44:109-17. 163. Suzuki Y, Yuen SR. Asbestos tissue burden study on malignant mesothelioma. Ind Health. 2001:39:160. 164. Dodson RF, Hammar SP, Poye LW. A technical comparison of evaluating asbestos concentration by phase-contrast, microscopy (PCM), scanning electron microscopy (SEM), and analytical transmission electron microscopy (ATEM) as illustrated from data generated from a case report. Inhal Toxicol. 2008;20:723-32. 165. Asbestos in Public and Commercial Buildings:A Literature Review and Synthesis of Current Knowledge. Cambridge, MA: Health Effects Institute, 1991. 3 66. Finkeistein MM. Asbestos fibre concentrations in the lungs of brake workers: another look. Ann Occup Hyg. 2008:52:455-61. 167. Roggli VL, Sporn TA, Case BW, et al. Comments on Asbestos fibre concentrations in tire lungs of brake workers: another look. Ann Occup Hyg. 2009;53:191; author reply 2-3. 168. Gustavsson P, Ahlbom A, Andersson T, et. al. Calculation of frac tions of lung cancer incidence attributable to occupational exposure to asbestos and combustion products in Stockholm, Sweden. Eur J Epidemiol. 2003;18:937-40. 169. Gustavsson P, Albin M. Low-dose occupational exposure to asbestos and lung cancer risk. Med Lav. 2006:97:357. 170. Stayner L, Kuempei E, Gilbert S, et al. An epidemiological study of the role of chrysotile asbestos fiber dimensions in determin ing respiratory disease risk in exposed workers. Occup Environ Med. 2008;65:613-9. 171. Stayner L, Smith R, BailerJ, et al. Exposure-response analysis of risk of respiratory disease associated with occupational expo sure to chrysotile asbestos. Occup Environ Med. 1997;54:646-52. 172. Stayner LT, Dankovic DA, Lemen RA. Occupational exposure to chrysotile asbestos and cancer risk: a review of the amphibole hypothesis. Am J Public Health. 1996;86:179-86. 173. Chen WJ, Monet NK. Malignant mesothelioma with minimal asbestos exposure. Hum Pathol. 1978;9:253-8. 174. Iwalsubo Y, Pairon JC, Boutin C. et al. Pleural mesothelioma: dose-response relation at low' levels of asbestos exposure in a French population-based case-control study. Am J Epidemiol. 1998;148:133-42. ` 175. Keeton P, Prosser WL. Prosser and Keeton on the law of torts. 5th ed. St. Paul, Minn.: West Pub. Co.; 1984. 176. Hickish DE, Knight KL. Report 41/68: Exposure to Asbestos during the Servicing of Brakes of Passenger Cars. In; 1968. 177. Cline W. Borg Warner Corporation answers to interrogatories in Re: All Asbestos Litigation Borg. In: In The Circuit Court Of The Third judicial Circuit Madison County, Illinois; 2006. 178. Jenkins W. Maremont. answers to interrogatories in the case of London vs. Flintcote et al. Circuit Cort of Pleas case no. 6849 Philadelphia County, Pa. In; 1984. 179. North American Asbestos Corporation Annual Report. 1957. Available from: http://www.egilrnan.com/browse.phpPdis- piay=list&dir=Asbestos/Brakes/amphiboie_use/. 1 SO. NorthAmerican Asbeestos Company Customer List. 1976. Avail able from: http://www.egilman.com/browse.php?disp]ay-lisi&: dir=Asbestos/Brakes/amphibole_use/181. Blau PJ. Compositions, Functions, arid Testing of Friction Brake Materials and Their Additives: Bechtel; 2001. 182. McDonald J. Linear Extrapolation for Risk Estimation as Low Level Exposure: The Asbestos Example In: London Univ; McGill Univ; Natl Heart -t- Lung Inst: Philip Morris; Unknown- 383. Greenland S. Relation of probability of causation to relative risk and doubling dose: a melhodologic error that has become a social problem. Am J Public Health. 1999;89:1166-9. 184. Egilman D, Howe S. Against anti-health epidemiology: corpo rate obstruction of public health via manipulation of epidemi ology. lntj Occup Environ Health. 2007;13:118-24. 185. Henderson DW. Malignant mesothelioma. New York: Hemi sphere Pub. Corp.; 1992. 186. Wagner JC, Pooley FD. Mineral fibres and mesothelioma. Thorax- 1986;41:161-6. 187. Churg A. Chrysotile, tremolile, and malignant mesothelioma in man. Chest.l988;93:621-8. 188. Egilman DS, Kim J, Biklen M. The Use and Abuse of Medical and Scientific Evidence- Inside the Couriroorn-An Epidemiolo gist's Critique of the judicial Interpretation of the Daubert Ruling. Food and Drug Law journal. 2003;58. 189. Rodelsperger K, Woirowitz HJ, Bruckei B, et al. Dose-response relationship between amphibole fiber lung burden and mesothelioma. Cancer Delect Prev. 1999;23:183-93. 228 Egilman www.ijoeh.com 1NT J OCCUP ENVIRON HEALTH