Document QJnxNpO3nvNexNpezdyGk5qXR

Critical Reviews in Toxicology, 2010; 10(S2): l-16 REVIEW ARTIClE Benze11e l1un1ai111ealtl1: A appraisal of associations 'tVith reVle\rv diseases David Galbraith\ A. Gross2, and Dennis Paustenbach1 'ChemRisk, Inc., San Francisco, California, USA and "ChemRisk:, Inc., Boulder, Colorado, USA health care Abstract Over the last century, benzene has been a well-studied chemical. with some acute and chronic exposures being directly associated with observed hematologic effects in humans and animals. Chronic heavy exposures to benzene have also been associated with acute myelogenous leukemia {AML) and myelodysplastic syndrome (MDSI in humans. Other disease processes have also been studied, but have generally not been supported by epidemiologic studies of workers using benzene in the workplace. Within occupational cohorts with large populations and very low airborne benzene exposures (less than 0.1-1.0 ppm). it can be difficult to separate background disease incidence from those occurring due to occupational exposures. In the last few decades, some scientists and physicians have suggested that chronic exposures to various airborne concentrations of benzene may increase the risk of developing non-Hodgkin's lymphoma (NHL) (Savitz and Andrews, 1997, Am J lnd :VIed 31 :287-295; Smith et al., 2007, Cancer Epidemiol Biomarkers Prev 16:385-391], multiple myeloma (MM) {Goldstein, 1990, Ann NY Acad Sci 609:225-230; Infante, 2006, Ann NY Acad Sci 1076:90-"109), and various other hematopoietic disorders. We present a state-of-the-science review of the medical and regulatory aspects regarding the hazards of occupational exposure to benzene, We also review the available scientific and medical evidence relating to benzene and the risk of developing various disorders following specific levels of exposure. Our evaluation indicates that the only malignant hematopoietic disease that has been dearly linked to benzene exposure 1s AML. Information from the recent "Benzene 2009;' a symposium of International experts focusing on the health effects and mechanisms of toxicity of benzene, hosted by the Technical University of Munich, has been incorporated and referenced. Keywords: Acute myelogenous leukemia; benzene; epidemiology; myelodysplastic syndrome; toxicology Contents Abstract .................................................................................................................................................................................................. l .s Introduction ........................................................................................................................................................................................... 2 the effects ofbenzene (181.10-2010) ..................................................................................................... 3 uiseases are associated with benzene? ................................................................................................... 16 Benzene exposure and acute myelogenous leukemia (AML) .................................................................................................... 16 Benzene exposme and non-Hodgkin's lymphoma (i\HL) ......................................................................................................... 18 Ren7ene exposure and chmnic lymphocytic leukemia (CIT,) .................................................................................................... 25 Benzene exposure and multiple myelon1a {MJ\11) ........................................................................................................................ 28 Benzene exposure and the myclodysplastic syndrorncs (MDS) ................................................................................................ 31 Toxicology studies and models for benzene toxicity ....................................................................................................................... 35 Trends of benzene-related disease ...............,.................................................................................................................................... 36 Impact of short tenn or peak exposures ........................................................................................................................................... 37 Environmental exposures to benzene ............................................................................................................................................... 38 Conclusion........................................................................................................................................................................................... 39 Declaration of interest ........................................................................................................................................................................ 39 References ............................................................................................................................................................................................ 40 ;!ddressfor Correspondence: David l-\. Galbraith, I\fD3 25 Jessie Strcc13 San Francisco; CA 941053 USA. E-tnail: dgalbraith@chernrLsk.conl (Received 04 May 2009; revised 22 June 2010; accepted 10 July 2010} ISSN l0-10-il444 printiiSSN l547-6il98 online IC> 2010 Infmma Healthcare USA, Inc. DO!: 10.3109/10408444.2010.508162 httpj IV\!1.\li.A.I. infllrmahealrhcare.cnm/txc: I CGU BEN0000258 2 D, Galbraith et al. Introduction Benzene continues to he a widely used industrial chemical in the production of polymers, resins, and synthetic fibers (ATSDR. 2007). In the late 19\h century and exiending into the early 20th century, benzene played a key role in the grow- synthetic rubber manufacturing industry, and was used by numerous industries as a solvent et al., 1992). Early observations of benzene effects in hmnans in the workplace were difficult to correlate with specific levels of exposure. Such difficulty was due mostly to inadequate technology for exposure estirnation, cor:npared with content- porary methods. Dr. ALice Han1ilton, who was perhaps the premier occupational physician ufher era, noted in an early review of benzene toxicity that "an excellent precaution'' was used by a steel manufacturer to prevent workers from being overcmne by benzene vapors in their holding tanks. "After emptying, washing out, and steaming out rhe tank, lower into it a cage ofv,;hite mice, and if the rnice are overcorne by lhe vapors lhe process of Hood- ing and steaming is repeated until the little animals can be lowered into the tank without showing any effect" (Hamilton, 1929). During Lhe early 20Lh cenlury, there were rnany repons of acute human exposures w benzene in the workplace at con- centrations ofben7cne to certain toxic endpoints, For exmnple, exposure to 4700 pprn reportedly produced confu- sion within 30 minutes of exposure, whereas levels of 6000 to 9000 ppm produced symptoms of acute poisoning within a few hours (Lehmann, 1~19). A listing of health effects on humans observed over time from various levels of exposure to inhaled benzene appears in Table l. Benzene has been known to have a depressive effect on the bone n"larrow following chronic exposures, occasionally leading to complete failure of the blood-forming elements, or aplastic anemia (Hamilton, l929j. During the past 60 years, concerns about the adverse effects of benzene in exposed workers have led to a series ofincreasingly stringent regulatory standards in the United Slates for benzene in the workplace (Table 2), and in 1978, there ~was a voluntary withdrawal of benzene as an added ingredient to consurner products manufactured in the United States (National Research Cmmcil, l !=Jill). Over the past 2!i years, concentrations of benzene in various consumer products has generally been between zero and less 1:han O.l% (Williams et al., 2007). Table 2. History of benzene regulatory standards. Time period n:v (ppm) Pl:.L (ppm) ~1946 100 1947 lOO 1948 5{] 1949-1957 35 1964-1969 25 {ceiling] l!J57-lY7!i :!5 1~72 1936 10 1987-current 1 1977-1996 lll 1997-currenl 0.5 TLV=threshold limit value, a recommended standard Jiom the American Conference of Govermnentallndustrial Hygieni5ts (ACGIH); PPM= parts per million; PEL=permissible exposure limit, an enforceable standard ere ated by Occupational Safety and Health Administration (OSHA). All concentrmions represent time-weighted averages (TWA) unltess othen~'ise noted. Source.s: ACGIH, 1976; OSHA, 1987; Paustenbach eta!., 1992; ACGIH, 2001; Capleton and Levy, 2005. Table l. Noncancer human heahh effects ofinhalational benzene exposure. Benzene .s concentration (ppm) Length o [exposure 19,000-20,000 5-10 minutes H<"alth efkct Death 6200-93{]0 30 1nlnutes-1 hour 4700 30minutes 3000 30minutes 1570-313{] Several hours 1550-3lll0 6 hours l!iD0-301iO Several hours 1500 l hour 50{) 1 hour 300 30minutes 150 4 months-1 year 50-150 5 hours {jfl l-20 days, 2.!i-ll hmns/rlay 40 l year 25 8 hours 7.6 6.3 years avg 2.3 Sources: Paustenbach, 1995; ATSDR, 2007. fmmediate or subsequent death Contusion Endumble Slight symptoms No serious effects Slight S}'lllptoms Serious sy_mptoms Symptoms ofillness Dizziness, headaches l'ancytop enia Headache} lassitude.. "\Veariness Mucous n1crnhranc irTitation, rlyspnca Leukopenia in first 4 months \:one Reduced lynrphocyte comus Reduced and RBC counts Reference Flury, 1928; Browning, 1937; Goldwater and Ttewksbury, 1941; Von Oettingen, 1940; G.enndte, 1 <Jfi:-\ Hamilton, 1931 Greenbmg, 1926 Gemrde, 1963 Brovvning, 1937 Hmnilton, 1934 Bloomfield, 19!il Gerarde, 1963 Gerarde, 1963 Flury, 1928 Aksoy, 1978 Gerarde, 1963 MidzPnski, 1'1'12 Cody, 1993 Gerarde, Hlb:{ Rothman, 1996 2002 CGU BEN0000259 Benzene and human health 3 Over the years, there have been significant improvements other chronic n1alignant hematopoietic diseases, including in Industrial hygiene practices, which have considerably multiple myeloma (MM), chronic myelogenous leukemia reduced the risk to workers (Hamilton, 1929, 1945; Gafafcr, 1943; McCord, 1931; Hemeon, 1955, 1963; Hamilton and (CML), andnon-Hadgkin's lymphoma (NHL), among others (Infante, 2006; Mehlman, 2006; Smith et aL, 2007; Savitz et aL. 1974; ATSDR, 2007; et aL, 200'-J). These have 1~J'-J7}. In our review, we evaluate the scientific and medical included the adoption of better skin/respiratory protection evidence suppmting assertions that diseases other than acute and engineering controls, substitution of oilier less toxic myelogenous leukemia (AML) are attributable to benzene. chemicals for benzene whenever possible, the advent of \Ve apply an evidence-based methodology to charac:teri7c dosed systems for performing chemical synthesis in manu- our evolving state of knowledge with respect to hematopoi- facturing settings, and increased n1onitoring of the expo- etic effects and occupational exposures et aL, 2005; litnes to workers, as well as hinrnnnitoring of those workers HS FDA, 200~; (:;n~er et aL, 2000}. in indusu-ies that continue to have exposures to benzene at concentrations approaching the contemporaneous occupa- State ofthe science regarding the effects ofbenzene tional exposure limits (OELs). (1880-2010) Although it was established in the 1920s that Benzene has been used as an industrial solvent in a variety of exposures to benzene adversely affected the blood-forming different industries (Table 3). Since the late 1800s, organs, there have been several case reports since that time exposures to benzene have been noted to result in suppres- suggesting that certain chronic diseases could be associ- sion of one or more ofthe blood-torming elements and with ated with exposures to benzene or even causally linked. a sufficient dose and duration have led to rhe development Early reports focused on abnormalities found in peripheral of aplastic anemia (Santcsson, 1897; Hamilton, 1929). Our blood, but in the 1950s to 1960s, some scientists reported knowledge reganling the hematopoietic effects of benzene chat chronic exposures to certain airborne concentrations of at various levels of exposure has significantly evolved over benzene seemed to increase the risk for developing leukemia the years. A summary ofilie various repmted hematopoietic (Cronkite, 1961; De Gowin, 1963; Vigliani and .Saita, l%4; effects of benzene is presented in Table 4. CThe two 'ltvidely Goguel el aL, 1 In 1977, Infante, and w't::UJo:t:u efrecls following high level, chronic exposures clearly established a link between exposure to benzene and to benzene are aplastic anemia and AJ'VIL an increased incidence ofleukemia in an ofvvorkcrs The foilowing represents a simplified overview, by decade, ou.:upaticmally exposed to benzene during the production of the evolution of our undeistanding (!i.tate of the science) of a natural rubber cast film called "Pliofilm" at tvvo Ohio of the health effects related to benzene from both toxicology facilities ([nfame et aL, Follmving that report, indus- and epidemiology studies. A number of other reviews have Trial hygiene pactices elating to benzene in the been writ!en oYer the past 20 yeaTs (Snyder, 2002; Schnatter, changed significantly, resulting in a marked decrease in the 2005; Pyatt, 2004; Lamm et al., 1989; Austin et aL, 1988). use of benzene in manufacturing processes, as well as lesser of exposure. 1880:s The major environmental sources of benzene exposure Autopsies ofdeaths following incidents ofan1te inhalational for most persons in '\IVestern society are active and passive benzene poisoning were first reported 1888)_ .s smoking, gasoLine vapor emissions, and automotive exhaust fumes (Wallace, l996a, l996b; Capleton and Levy. 2005). By 1890s and concentrations in the ambient air have decreased Earliest known cases of chronic benzene poisoning were steadily since the early 1990s. The use of benzene in con- reported, where nine cases of "purpura hemorrhagica" were sumer pmducts likewise has fallen off dramatically (Williams seen in young girls using ruhher cen1ent v.ith high concen- et aL, 2008). Smokers have been found to experience about trations of benzene in a Swedish tire facwry, four leading to half of all human exposures to benzene in the United Scates, death (Santesson, 18Y7). A similar case was reported in a man with 90% of their cumulative lifetime doses from employed for several years in a dye factory, who died with smoldng, vith 1neasured body burdens 6 lo 10 times that of spontaneous hemorrhaging and probable myocardial infarc- nons1nokers (Vvallace, l996b). In many countries, the permis- tion (Lenoir and Claude, 1897). [Some believe that this latter sible concentration of benzene in gasoline is leso; stringent case report is the first evidence of benzene being associated cornpared to North Arnerica, and Australia, and for with leukemia, but the lack of clinical data describing the many countries, no standards have been set at all (Figure l) patient does not provide sufficient basis for this claim.l (International Fuel Quality Center, 2008). Hun1an exposut-e to benzene continues to occur due to 1900s its presence in ambient air, cigarette smoke, and some work- Several case reports of at:ute industrial inhalational poison- and the range of possible adverse effects folhw.ring ing:-; were described in the literature. Ironically, individuals acute or chronic exposure remains an issue of important who attempted to save coworkers in the work inleresl Lo regulatory agencies and lo Lhe puhlic. Over lhe las!. environment oflen experienced higher risk of mortal- few decades, smnc have suggested that exposures to benzene ity than the workers who were initially overcome by the could be responsible for an increased incidence of several vapors (Hamilton, 1929). Occupational settings for these CGU BEN0000260 4 D. Galbraith et al. 5-?vol% No Standard l'igurc 1. International limits on benzene concentration in gasolinP. Source: International FuPl Quality Center, 2008. Table 3. Occupations historically associated with benzene exposure. Printing Leatherworkishoemaking Chcmicnl manufacturing Petrochen-.icals (refining, di!'!trihution, &ervit.:e station opelaturs) Scientific laboratories (particularly glassware cleaners) * Rubber n1anut"acturing Coal-based coke production {metallurgical and steel manufacturing) Plastics manufacturing Source: ATSDR, :woo. Table 4. Hematopoietic conditions associated with significant exposures tc benzene. Ht:'lnatutoxidty "Anemia o L.eukopenia " Thrombocytopenia Aplastic anemia Acute myelogenous leukemia Myelvdy,plastic cyndwm" Irn1nune dysfunction o \\Todcer and animal data have shuwn diminished antibody levels and cellular Smm:e: )'ITSDR, 2007. .s inhalational injuries were often connected to the manu- facturing of explosives, with found within The first cases of chronic benzene poisoning in Great the interior of chemical storage tanks that had supposedly Britain were noted in two men en1ployed at a balloon manu- been boiled out and cleaned of residual benzene (Hamilton, facturing plant. Earlym~asuremcnts of occupational benzene 1929). l11ese deaths occurred long before obligatory proce- exposures were described, ranging from 210 Lu 800 ppm, with dures were for tank entry. peak levels over 1000 ppm (Legge, IB20). 1910s Twenty-one cases ofacute benzene poisoning \Vere reported in Germany in 1910 {Heffter, 1915). The first American cases uf chronic benzene poisoning were descri1Jed at a tin GHI factm-y, where three young were working with rubber solvent, two of whom died (Selling, 1910). Experiments Selling in animals and his observations that benzene can cause a lowering of white cell counts led to the notion that hen:t:ene could he used as an effective treatment for leuke- mia (Koranyi, 1912). Hamilton reported on 14 cases ofben- zene poisoning, identifying a fatality rate of SO% (Hamilton, l Benzene poisoning in American rubber workers was described, specifically in tire builders (Harrington, 1917). Tim::e ufthe five cases were fatal. 1920s Italian researchers reported seven cases of aplastic anemia in young wmnen employed in raincoat factories where benzene-containing glues were used. Exposures at the facility were measured at one part per thousand, or lOOOppm (Meda, EJ:d2). Workers in the German rubber manufacturing industry represenLed the first reports of chronic benzene poisoning (Brucken, 1923 ). A comprehensive study of American industries using benzene was conducted in 1926, where 1veekly volurnes ufhenzene usage were seen to range frorn 50 to over 10,000 gallons (Greenburg, 1926). Measured air concentrations in these facilities showed mean concentrations of 70 to 1800ppm benzene, "IIVith peak concentrations ranging fi-om CGU BEN0000261 Benzene and human health 5 llO to over 4000pprn. The author concluded that although "The greater part of the careful, detailed work on the less toxic solvents should be substituted whenever possible, action of benzene has been carried out on animals in :..with proper care in construction, maintenance, and opera the laboratory and for the most part 'ivith the tion, the use of benzol [historic term, generally referring to ment of a technic [sic] which does not in any way repro- coal tar-derived can be made sufficiently safe to duce what takes place in industrial poisoning ofhuman warrant its employment" (Greenburg, 1926). beings. The victims ofindustrial poisoning whose bod- During this period, occupational health guidelines were ies have come to autopsy are few in number, and their written and they usually recommended that any worker cases have not received as careful study, except in rare who showed a drop of 25% in the white cell count or a total instances" (Hamilton, 1931). white cell count of less than 5000, 25% decrease in the red cell nnmt, rn a hemnglohin level less than 70% should he Over the next 20 years, Hamilton would continue to docu- removed from exposure and sent to another area without benzene e:Kposure (Williams and Paustenbach, 2003). The use ment cases of acute benzene poisoning; these are chronicled in a book titled Alice Hamilton." A in l,etters {Sichennan, of benzene as a treatrrlent for leukemia was actively consid- ered in the 1920s (Greenburg, 1926). poisoning 1984). In the late 1~:-:IOs, two cases of leukemia believed to was often referred to in early discussions ofthe health effects of benzene, and generally referred to acute toxic chemical effects, such as nausea, visual changes, lack of coordination, he related to exposures to benzene were reported in Massachusetts (Mallory et al., 1939). One of these was in a 28-year-old leather worker who was exposed to benzene for unconsciousness, and death. As more became known in the late 1920s regarding intcnncdiatc and chronic exposures, aplastic anemia was considFred the ].Rin1ary adve1se effect associated wtth benzene "poisoning." Other nonmalignant 4 years, at levels that were described as causing "evidence of benzene intoxication:' Six years later he a11 aggres- sive acute mveloid leukemia. 1he other case was that of a 12-ycar-old boywho was the son ofa painter and played in his hematologic effects have been reported at various levels of exposure (Table 1).] In 1928, lwo workers at a plant developed hematologic disease attributed to chronic exposures to benzene, one ofwhom died of aplastic anemia, father's shop frequently; he was diagnosed with an inunature lymphoblastic leukemia. No symptonu; excessive exposures up to their diseases were described. A possible leukemia arising in a 38-year-uld man was described in the same year (F.rf et aL, In 1918, this the other of acute lymphoblastic leukemia. 1he worker with leukemia had worked at the plant for 15 years; the last 5 of which were believed to be in a job where excessive benzene exposure was Although 110 n1easuren1ents man worked vvith his brother ': .. in a studio in which benzene was used." That year, his brother developed anemia and dietL Subsequenlly, the palienl changed his occupation to avoid benzene exposure. Seventeen years later, he started of exposure were provided, it was stated that his job was using a rubber cement containing benzene for 14 hours a considered dangerous, and that none of his fellow workers day. Less than a year later he was admitted to the hospital were allowed to nwre than 1 n10nth at a time in the with weakness and bleeding gums. He left the 5 job he performed. However, this man wa!> allowed to remain weeks later, and died 2 months thereafter. Postmortem in this high-exposure work station for 5 straight years. The exarnination revealed findings consistent with .s lmency period between his exposures and his diagnosis was not described; no bone Inanow findings were reported. 1his leukemia. An additional case involved a chemist diagnosed with leukemia at another institution at age 23. He had worked case has generally been believed to be the first possible case for 1H months, using benzene to benzoic acid. He ofbenzene-related leukemia, but the lade of clinical data and discussion of other chenlical exposures oT other risk factors in this case report makes it difficult to establish a certain connection to henzene (Delore and Borgomano, l~J2H). AFrench automobile factory was noted to have four work- was admitted tu the hospital and discharged after 2 weeks. A follow-up visit 2 years later found him to be in health" (Erf et al., 1939). A study entitled ''Benzene Poisoning in the Rowgra\nclre Printing New York showed that all332 ers who developed "purpura hemorrhagica," two with fatal consequences. \IVorkers were described as being on the Job for less than 6 months (Hamilton, 1929). crs studied were chronically overexposed to benzene, at air concentrations ranging from ll to HifiOppm. Evidence of poisoning was found in 130 workers, with 6 requiring hospitalization (Greenburg et al., l~_HB). As a result of the study, the 1930s II. cmnprchensive review of the effects of acute and chronic benzene poisoning was provided by Hamilton in 1931 use ofbenzene was discontinued, because air concentrations could nul be consistently kept within safe levels, and cheinicals with lesser toxicity were substituted. (Hamilton, 1931). The results of animal studies, as well as the clinical presentations of people who had experienced l940:s significant exposures to benzene in the workplace, were Gafafer's Manual of Industrial Hygiene, prepared by the summarized. Blood, bone marrow, and palhologic findings indusi.rial hygiene division of lhe National Institute of at autopsy were presented, and Hamilton called for more Health, United States Public Health Service, related in 1943 research on the effects of benzene, stating that: that~ ..benzene, vvith iis known harmful properties, has been CG U_BEN0000262 6 D. Galbraith et al. abandoned as a solvent by many industries:' vraximum been reported:' The maximum acceptable work-place concen- allowable concentra[ions of benzene in the workplace were tration was 25 ppm over an 8-hour workday (API, 1960). then 100 pprn "on the basis of an daily exposure" A summary ofthe e"idencc for radiation and various chem- (Gafafer et al., 1943). icals being associated with a higher incidence of leukemia The next in the series of Hamilton's reviews of industrial was published by a research pathologist at the Bmokhaven noted that: "Acute benzene poisoning is of little National (Cronkite, 1961). In the author's opinion, imponance under modern industrial manageruenL Ihe dan- at that time: .. only benzoL has been investigated sufficiently geT is well understood, and no longer arc n1en sent unpro- weLl ro make it appear a scdous contender for the part of a tected into tank cars or vats" (Hamilton, 1945). She also noted chemical leukemogen in man:' Up to tl1.at point, many cases that e"l.idence seemed to he accumulating, : .. that leukernia, had been presented in the literature, and the "cause of death myeloid nr lymphatic, may he nne ofthe forms benzene in .~nmP instances was ascribed to anemia, and in soning Vigliani and Penati collected l 0 cases, which others. to with reference to the autopsy series by had been reported by 1938, and Mallory and his colLeagues Mallory (Mallory et aL, 1939). Leukemia was described as an have added 2" (Hamilton, 1945). immediate consequence of chronic heavy exposures, rather Marshall Clinton, a medical doctor under the direction than as a disease with a particular period. Animal of Philip Drinker, a Harvard industrial hygiene professor, studies attempting m show malignant health effects ofben- prepared a report on benzene for the American Petroleum Institute (API). \J\lith regard to leukemia, it was stated that~- zenewere described as "equivocal:' Hesearchers fi:om Italy provided a review on benzene and reasonablywell documented instances of the development of leukemia, apparently in response to the comments made leukemia as a result of chronic benzene exposure have been above by Cronkite, stating that "!-'rom this [Cronkite's paper] cited" (APl, 1948). 1he possible occurrence of "latent bone- and othPr American papers on benzene poisoning it "eems Jnarrow injury due to benzene without any immediate altera- that the and the Italian literature on ben- tion in blood findings" years after exposure had taken place zene leukemias is pomly known" (Vigliani and Saita, 1964). hut conclusive proof for this possibility was Six cases of leukemia associated '"''ith chronic benzene l11e reporl mentioned lhaL: "... lhe only exposure were reviewed. In the lwo cases where workplace absolutely safe concentration for benzene is zero:' Later, the concentrations ofbenzene were available, Lhey reported air- author of that report stated that he was defining zero as "the borne concentrations "ten times greater than the maximmn limit of analytical detection;' which 'IlVas about 50 ppm at that allowable values generally accepted" at least 500 ppm). time (Clinton, 1994). This statement is not surprising since the Three ofthe six cases had been diagnosed v.rith aplastic ane- report went on to recommend a lower occupational exposure mia prior to their leukemias, and five of the si-x ''...had low limit than the I OOppm level which was 1\CGlH, white-cell counts at the outset that were independent of any stating: "A limit of 50 pprn or less is strongly recommended, therapy:' The authors also emphasized that they had never particularly where exposures are recurrent" (API, l!:J48). seen a case of "chronic 1nyeloid or lyrnphatic leukemia in The 50 ppm exposure limit that APl workers 1,vith benzene:' \'\lith regard to previous although far in excess ofmodern represented what claims of these diseases being caused by benzene exposure, was thought to be a tolerable exposure given the knowledge ~ ..we must emphasize our view that in some of these cases .s of benzene health effects at that time. reported in the Literature the occupational history was not convincing. 11Ie nuinber of persons occasionally exposed to 1950s benzene or exposed to very low concentrations was so high In the 1950s, over 100 cases of benzene poisoning were that some cases of chronic leukemia could have occurred described in a European shoe factmy. A1ea san1ples for ben- among them as among anyotherwnrkingpopulation .... Great zene collected at three workstations had average values of caution must be exercised before admitting the benzene 3UI, 4;-;:c;, and 470ppm (Savilahti, l!:JS1). Blood was etiology of chronic myeloid or lymphatic types ofleukemia" in 147>>~.-orkers,with 7:::1% tohavesome (Vigliani and Saita, 1964). They also noted that: "We arc fully abnormality. Low plalelN comlls were the most common aware of !.he fact that no final statemelll aboul the existence of finding (62%), followed by anemia (35%) and low white cell a true "benzene leukernia" rnay be made, without a statistical counts (32%). Thirty-one workers were found to have simul- analysis oflhe incidence ofleukemia among workers exposed taneous effects on all lhree blood lines (platelet, white blood to benzene as with that ainong a control group of cell, and red blood cell counts). With removal trom exposure, the same age, sex and living habits: A large fraction of shoe 120 workers recovered after three months, one died, 20 con- and Leather workers were noted to be employed at home tinued with minor symptmns, six rc1nained on sick leave, and under uncontrolled conditions and the content of benzene l was still hospitalized after one year (Savilahti, 1956). in the solvents used by various home-based operations was characterized as uncertain at best and Saita, 1Y!:i4). 1960s Overall, the pre-l970s experience regarding the linkage In 1960, Dr. Drinker developed a second edilion of his toxico- between benzene and leukemia largely consisled of occa- logical review for benzene at the request ofthe API and noted sional case repons and small groups of patients, generally that "leukemia as a result of chronic benzene exposure has witl1 known significant exposures to benzene. In many CGU BEN0000263 Benzene and human health 7 situations, there \i\Tas concurrent exposure to other solvents, than 20 distinctly different chronic effects on the blood-form- as well as a history of smoking, which resulted in a lack of ing organs. As a result, it was difficuLt to specifically identify regarding a cause and effect :relationship. a causal link between benzene and any particular disease. He and his coauthor felt that "The most convincing cases 1970s Aksuy, a Tuddsh hernatulogist, has been credited v.rith performing the first major epidemiologic study examining the effects of benzene on a specific occupational subgroup; that of benzene leukemia are those in factories where there were outbreaks of chronic benzene poisoning: He sum- marized his with :rotogravure workers and shoe manufacturing where "All of the n1en had worked in factories ofshoe and leather workers. He and his colleagues acknowledged at that time that ': .. doubt still exists as to the causal relationship between benzene exposure and leukemia" where the concentration of benzene in the air of the worldng places was well above the threshold limit value:' [At that time, the threshold limit value (Tnl) was 2"1 ppm] {Vigliani and (Aksoy et aL, 1974), resulting from conflicting animal data and a lack ofepidemiology data demonstrating an increased incidence ofleukemia in exposed populations. 'fue incidence of leukemia was calculated to be 13 cases per 100,000 in the population of28,500 Turkish workers over a 7-year period of observation. Several cases of "preleukemia;' the exact diag- Forni, 1976). After the rotogravure industry stopped using benzene in favor of toluene as its main solvent. the authors noted: ~ ..we have seen no new cases of aplastic ane1nia, nor of leukemia due to toluene exposure." J\;1easured airborne concentrarion~ of benzene in rotogravure plants were reported to be bet\veen 200 and 400 ppm, with peak concentrations up to nosis ofv,"hich was not defined by the authors, were added to 1500 ppm. Shoe factories where workers developed leukemia the total incidence ofleukemia. 111.eir approach likely overestimated the true leuketnia risk, as son1e proportion of these were found to have workplace concentrations between 25 and 600ppm, vvith most measurements being between 200 individuals would not be to progress to leukemia, and some might even recover ifthe hematological abnonnali- and 500ppm, VITith some solvents containing l 00% benzene (Vigliani and Forni, 1976). ties noted were related to prolonged 1nyelosuppressiun from benzene and not true myelodysplasia. Earlier, Aksoy had summarized his observations of leukemia in four shoe 1.vorkers, who ': ..were rather heavily exposed to benzene vapor for a period varying bctween six and fourteen years. Their places at work were unhygienic, and the air concentrations ofbenzene were between 150 and 210 ppm duringworldng hours;' with occasional readings up to 650 ppm (Aksoy, 1972). 1bree ofthe four workers had been previously diagnosed with aplastic anemia. Epidemiologic studies on rubber workers in Ohio noted slightly reduced mortality from all causes, but detected signif- icantly elevaled numbers or cancers or the slornach, proslale, and blood-forming tissues, including leukemia (JVIcMichael et aL, 1974). The authors concluded that : .. until it is demonstrated tl1at specific exposures ur jobs within the industry are associated with excess deaths, one can only suspect, rather than conclude vv:ith confidence, that working in certain jobs 1A'ithin the rubber industry entails an inueased risk of dying from specific causes." In 1974, a review ofleukemia incidence in a populalion of petrochemical workers from Exxon affiliates in Europe was released, finding a slightly diminished rate of Leukemia An expanded follow-up study involving multiple locations discovered an increased rate of "lymphatic leukemia'' in workers exposed to organic solvents, one of which compared with the general populations in the countries was benzene. Acknowledging that this finding was unusual, .s studied. No individual exposure data were presented, and the authors stated that: "The sparse literature in this area indi- the author noted several difficulties with data collection and cates that, at least fur cases of gross benzene poisoning, the organizadon (Thorpe, 1974). leukemia tends to be of either the hemoc'jrtoblastic (or stem In their text on industrial toxicology, Hamilton and Hardy cell) ldnd or the myeloblastic kind. The association oflym- reflected that: phatic leukemia with exposure to organic chemicals appears to have no reported precedent" (McMichael et al., 1975). Six "\1\lbile there has been no doubt for many years that of the eighl "lymphatic leukemia" dealhs were ofthe chronic benzene can produce fatal anemia, the asso- subtype. i\ third srudy ofthe rubber workers showed a slightly ciation bervveen benzene exposure and leukemia has increased incidence of "all leukemias" "l:vith a standardized been a 1natter of n1ore recent controversy. ... Chronic mortality ratio (SMR) of 130 (confidence limits not provided). myelogenous leukemia appears to be the most com- However; when the cases were divided into "myeloid" and nwn type associated with benzene exposures, but acute "lyn1phatic" categories, the authors found that Inortality frmn myelogenous and acute and chronic lymphocytic varie- lymphatic leukemia was strongly associated Vvith working in a ties have been reported as well" (Hamillon and Hardy, synthetic rubber plant, whereas forms did not dem- 1974). onstrate a significant association (McMichael et al., 1976). Tn this study, data on specific subtypes of leukemia were nut Vigliani and Forni provided an update on benzene health \Vith to the McMichael et aL ,;tudies, how- effects, stating that over the last 40 years, more than 100 cases ever, we should nate that the tire building process has his- of leukemia attributable to benzene had been described. torically involved a 1.'\/ide variely of chemical exposures allhe However, the authors' use of the term "leukernia" was too various stages of production, which the authors recognized to be informative, as their definition included more (McMichael et aL, 1975). CG U_BEN0000264 S D. Galbraith et al. Interestingly, a subsequent National Institute for benzene exposure developed leukemias after exposure peri- Occupational Safety and Health {NIOSH) investigation of the ods ranging from 6 months to 6 years. Concentrations ofben- synthetic rubber plant studied by McMichael et al. where zene in the workplace were never n1easured to be less than reported this very strong association with "lymphatic leuke- 130 ppm, and patients were described as working long hours mia" was "unable to identify hen:zene-related exposures" \<Vith little to no environmental controls and Erdem, (NIOSH, 1979). NIOSH also commented that animal studies 1978). In a letter to the Lancetthat same yem; Aksoy remarked sLill had not provided : .. reliable information on that Jor "moderate or low levels" of benzene exposure, which capacity to produce an increased incidence of leu_kemias:' he defined as 10-25 ppm, ~ .. hacmatological abnorn1alities lhey thought it': ..quite conceivable that leukemias described characteristic of chronic exposure to benzene were not usu- in the earlier experiments in mice after benzene treatment ally seen" (Aksoy, 1 have, in large part, developed spontaneously nr byvirus infec- reporting nn the health effects nf ten Dow tion, rather than being caused by benzene" (NIOSH, 197~J). Chemical employees historically to benzene levels They did conclude, though, that the accumulated clinical and exceeding 25ppm over many years, reported that aU ten epidemiologic evidence showed thac benzene was a carcino- employees had an increase in their mean red blood cell vol- gen and leukemogen and recommended that it be replaced ume (MCV) in testing performed in 1963. Some had minor with less harmful substitutes v.rherever possible. reductions in hemoglobin levels, and no other "1Jo;u,,u,"cuJL The initial report on the incidence of leukemia in rubber bone marrow or peripheral blood effects were noted. It was workers in Ohio (the "Pliofilm" cohort) was published in concluded that ~ .. the exposure experience has not caused a 1977. "Ihis retrospective cohort study was compared lasting deleterious effect on their hematopoietic systems or to previously published cohort reports in that there was felt their overall health" (Fishbeck et aL, l to be very rninimal confounding exposures to other types of chemicals or dangerous solvents; that is, benzene was essen- 1980s tially the only chemical used at the facility {Infante et aL, lB77; In HJ80, Aksoy reported that he had observed several mher Paustenhach et aL, 1992). and jobs where direct diseases in the Istanbul workers that he thought might exposure to benzene occurred were idenlified were based on be associated with exposure to ben.t.ene. These included a NIOSH-administered worker survey. NIOSH also provided malignant lymphoma, (both Hodgkin's and non-Hodgkin's detail regarding the Pliofilm manufacturing process, engi- lymphoma), lung cancer, myeloid metaplasia, nclrr>Y\r,nn neering controls, and mosc iinportantly, detailed personal and nocturnal henmglobinuria (often associated with apla!>tic area air san1pling data describing actual benzene exposures anemia), and multiple myeloma. The evidence provided over time. 1\fter 75% of the worker population had undergone by i\ksoy came in the form of individual cases, both evaluation, the authors a 5-fold increased risk of in and from other published case reports (Aksoy, all leukemias, and a 10-fold increased risk for myeloid and 1980aJ. monocytic leukemias. No lymphocytic leukemias were noted A second article by Aksoytl1atyear described a significant (Infante et aL, 1!177). decline in the incidence ofleukemia fo1lov;,ing the "phaseout" In a comprehensive re\'iew of benzene health effects com- of benzene usage starting in 1969. (Aksoy, l980b) He also missioned by the American Petroleum Institute, Goldstein emphasized that acute forms of leukemias were by .s coneluded: : .. there is reasonably good evidence that inhala- far the most common type observed in benzene-exposed tion of benzene is associated with an increased incidence of workecs, '~ .. followed by acute erythroleukernia and pieleuke- acute myelogenous leukemia, and possiblyother hematologi- mia:' The rare occurrence of chronic forms ofleukernia was cal neoplasms" (Golds1:ein, 1977). described as "strildng!' In discussing other malignancies, he In 1!-l7fi, largely due to the Infante et aL repm-t, NlOSH noted that all of the cases of lung cancer ~ ..were heavy or recommended that the benzene PEL be lowered to l ppm moderate smokers," and that in 1:\No cases the form of lung over a 2-hour period in l!:l7ti, based on the overall cancer was oat cell carcinoma, a fonn oflung cancer heavily conclusion that benzene was The Department associated with smoking. He concluded, for the non- ofLabor intended to reduce the standard, bmwas challenged leukemic diseases he had noted in benzene-exposed people, in court and in JuLy of 1980, the Supreme Court withdrew that "Though these case reports [dol not prove a causal rela- the 1 ppm standard, stating that further evidence of adverse tionship between benzene exposure and the various malig- health effects below 10 ppm was needed (Thomas, l!-l82). This nancies, ihe frequent finding of this association suggests thac decision had long-term significance, as the courts ruled that benzene may not only cause leukemia but also be involved in unless advetsc effects were observed at the current standard, other types of malignancy" (Aksoy, l980b).1hc issues, how- then it did not make sense to reduce the lhnits on exposure. eve!~ of concunent chen1ical exposures and alternative risk Such an action would require substantial expenditures for factms are ilnportant to note, and are dearly evident in the and enforcement, and society should expect to oflung carcinon1a put forth see a demonstrable improvement in worker health Nonmalignant effects of chronic benzene exposures were et aL, 1991). explored by pelroleurn industry researchers, who observed Aksoy reported in a study in 1978 that 6 out of that occupational exposures (40 hours/l,vcck) above 50 ppm 44 patients \>\lith pancytopenia resulting from chronic heavy for extended periods could cause diminished numbers CGU BEN0000265 Benzene and human health 9 of platelets, red cells, or white blood cells. They noted that of cumulative e-xposure: up to 40, 40 to 200, 200 to 400, and "the clearest relationship between benzene exposure and greater than 400 ppm-years. The authors reported the new leukemia was demonstrated in studies of benzene exposure findings ofan apparent excess of MM cases, but reported that prim to 1960 when high benzene concentrations (>200ppm) the calculated mortality ratios for MM did not increase with in the air were likely," and stated that "a standard of lOppn1 increasing benzene exposures. The authors noted several time-weighted average (TWA) benzene in air is an accept- gaps in the available data sets for various jobs, stating that able value ior protection from any hematological efl:ecl" these were filled by interpolation ot existing data. Rased on (RTicf ct al., 1980). 1luc-ywcnt on to S1.ImmaTize the cxposunc- an observed positive trend of increased leukemia risk with experience in petroleum refineries, ;vhere they reported': ..a benzene exposure, the authors recommended a lowering of low probability (<5%) of benzene levels exceeding lppm the occupational exposure limits, which were then lOppm (as TWA and a negligible chance of exceeding 5 ppm." Even sn, an R-hrmrtime-weightedaverage) (Rinskyet l9R7). they recommended that benzene should be replaced with In 1988, a review ofthe evidence linking henzene to leuke- less toxic solvents whenever possible, and that personal sanl- mia was performed, concluding that the available epidemio- pling of individual workers was the best method of verifying logic data supported the notion that chronic heavy exposures exposures. Tt was their belief that no test at that time was suf- to benzene were associated with an increased risk of develop- ficiently reliable to track airborne concentrations to less than ing acute myelogenous leukemia, even though the evidence about O.Sppm of benzene; especially in environments with came from a "relatively smaLl study by Rinsky" (Austin mixtmes {Brief et al., 1980). et al., 1988). The rest of the accumulated epidemiologic data Najean, in a long-term follow-up study of 429 patienls were believed lhe authors to be very weak in demonstrat- diagnosed with aplastic anemia, noted that ''i\ possible or ing a possible causative relationship bctvvccn benzene and probable toxic etiology does not discriminate benveen favo- leukemia, even at historically ve1y high levels of exposure. rable and unfavorable evolution" ofaplastic anemia (Najean, "''""'""" the available data are too sparse, or suffer other 1!:!81). In this series, 31% of the initial group were consid- limitations, to substantiate the idea that this causal associa- ered to have disease secondary to toxic or industrial tion at low levels 1-l0 ppm) of benzene:' A cri- exposures; 27% of the 137 paLieHls had suspected lique of previously published risk assessrllenls was provided, disease secondary to toxic agents. Less than l% of patienls along with the authors' ovvn risk assessment for benzene went on to develop leukemia as a complication of their dis- (Austin ct al., 1988). ease (Najean, 1981). Goldstein pmvided a lengthy re\iiew of benzene toxicity Rinsky et al. provided an update on their Pliofilm cohort in 1988. In describing benzene's effect on the hone mar- in 1981, after 98% of the vital data from the worker popula- row, he opined: '~\s anticipared for an effect of this nature, Tion had been discovered. Industrial data and a there is an apparent threshold, i.e., at some benzene dose no description of the workplace environment were provided, as overt decrease in blood count is observed:' Furthern1orc. "A well as job descriptions for the various classes of laborers. decrease in peripheral blood counts has not been observed No additional cases ofleukemia fulfilling their study criteria in any species at e:qmsure doses below the current U.S. were reported; a Lotal of seven cases were described in the Occupational Health Standard of lOppm T\tVA'' (Goldstein, worker population, all myelogenous or monocytic in cell l 988). In discussing the Italian and Turkish experience with .s type. 111e authors concluded that these findings confirmed benzene, he noted that 1neasuren1ents of ambient benzene their previous conclusion of increased leukeinia risk in the levels exceeded 200 ppm, and chat of the glues cohort (Rinsky ct al., 1981 ). in the shoe and leather workers with solvents not containing A review of petrochemical wmkers exposed to benzene benzene ~ .. put an end to this outbreak of overt hematological concentrations fron1 less than 11.1 to 25ppnl as an H-hour consequences due to benzene exposure" (Goldstein, EllUl). TWA found no differences in red blood cell (RBC) count, He also explored the nonleukemic associations described white blood cell count. hemoglobin, or platelet count (Tsai by Aksoy, that "The causal relationships have not yet et al., 1 been established;' and ': ..in no case does the evidence appear Another review Aksoy Mated that "lhere is no doubt to be sufficiently clear-cut to ovenvhelmingly demonstrate about the leukemogenic effect ofbenzene in man;' citing data a causal relationship betvveen benzene and one or n1ore from his previous studies. He noted an additional possible human lymphatic nm1ors" (Goldstein, 1988). risk factor of cholera vaccinations, which he felt could be Goldstein discussed a 1985 study of benzene exposure in acting as a promoter for leukemia in those persons exposed the United States petroleum industry by R1mion and Scott to benzene. Familial risk factors and differential hmnan sus- (Runion and Scott, where ': .. 87% of exposures were ceptibility w-ere also discussed. 'The same nonlcukemic dis- below 1 ppm and 98% below 10ppm TWA:' Tn describing ease states were again suspected of being linked to benzene the difficultie!> in the surveillance of a population exposed exposures, though little new supporting data were provided to benzene, Goldstein :stated that each component of a corn- (Aksoy, l98!J). plete blood count had a wide range of normal values, and A 1987 update of lhe Pliofilm sludy examined all workers lhat inLerpreling findings jusl barely different from "normal" with at least 1 ppm-day of exposure up to the end of calen- was unclear. In addition, routine surveillance of even a srnall dar year 1965. Workers vvere stratified into four categories work force was "likely to find at Least one blood count below CGU BEN0000266 lO D. Galbraith et aL the statistically normal range:' Also, changes overtime could No increased incidence of solid tumors was observed, also be difficuh to address: '~"'c major cause of such a change going against animal studies that suggested this might over time can simply be laboratory variability, particularly as be an additional toxic endpoint in humans. a small change of such a nature is not of clinical pertinence and thus unlikely to be ofconcern to a routine dinlcallahora- (Goldstein, 1988). In a separate analysis, Goldstein worked with Kipen and Cody to exaininc the blood studies from the Pliofilm cohort from 1940 to 1975 and determined that : ..benzene exposure for the cohort during the 1940s was significantly higher than in s11hsequent yeaL~: Hm:vever-, although they observed significant effects on the peripheral blood findings at these high levels of exposure in the 1940s, Goldstein et aL also found ~ .. that surveillance of blood counts to monitor populations ofworkers exposed to benzene may not detect cohort effects of exposures in the range of the current standard of l ppm, as during the 1950s and beyond these workers had exposures likely exceeding these values with- 1he assembled exposure and disease data were felt to be ~~ .. consistentwith a threshold model forleukemogenesis by benzene" (Paxton et al., 1994). Leukemia deaths in the entire cohort occurred exclusivelyin workers who started employment prior to 1950, suggesting that later worker exposures Inay have differed substantially from earlier years. A Inore detailed review of exposure data concluded that estimates of leukemia risk reported by PJnsky ct al. ovcrestirnated actual risk by an order ofrnagnltude, and that exposure estimates derived hy Paustenbach eta!. (1 Y92) agreed with those of Crump and Allen (1984) attempting to identify the minirrmm lifetiiTle dose that increased the leukemia risk. out apparent aggregale depressions in their blood counts" (Kipcn ct al., 1989). Kipcn ct aL concluded that: "these data suggest substantial limitations ofhematologic examination of populations to deiect abnormalities in populations currently Over the next few years, there continued to be analyses of the likely exposure ofthe Pliofilm cohort and, ultimately, the US Environmental Protection and the Anu~rican exposed to benzene" (Kipen et aL, 1989). Conference ofindustrial Hygienists {ACGIH) TLV committee fundamentally gave equal consideration to the various expo- 1990s In a Monsanto study of !.he peripheral blood effects from chronic low level exposures to benzene, 200 '"-"Orkcrs with exposures to benzene in air ranging frmn 0.01 to 1.40pprn as an 8-hour T\VA were evaluated. The "no effect" level for pro- sure estimates and derived a "weight of evidence" approach to identify acceptable levels of exposure. A study involving 19 separate cohorts totaling 208,000 petroleum workers in the USA and UK was analyzed for the incideuce of individualleukernia subtype;;. 1he authors ditl ducing blood effects in humans at the time of publication was not find an increased risk for any leukemia subtype, including judged robe 25ppm. Using a multiple approach to control for the effects of smoking and other possible con- AJ\1L. 1hcy hypothesized that this was due lo benzene exposure in the petroleum industry being : .. substantially lower founding variables, the authors found no effects for benzene exposure on any of the Tnain peripheral blood parameters (red and white cell counts, hemoglobin, platelet count, and than needed to reach the observed threshold:' This rhreshold, based on data from the Pliofilm cohort, was believed to be at least 200 ppm-years by the authors, and possibly as high as MCV) (Collins, 1991). 400-.'iOO ppm-years (Wong and Raabe, 199S). .s Rushton performed a mortality analysis of refinery and In 1996, vVard et aL reexamined the blood screening data distribution center wod.;.ers ernployed for at least one year from d1e Pliofilm cohort in the 1940-1975 time period, using from 1950 to 1975, an update of t\vo prior reports on the Rinsky's estimates from 1987. 1he authors acknowl- same population {Rushton, 1993). Consistent with earlier finding:-;, overall mm-tality was significantly less in the edged the difficulty of their task, since little exposure tlata were available in the l~J40-l ~1411 time period, and neither laboratory worker population, feLt to be related to a heaLthy worker effect. Total observed deaths from leukemia in both worker cohorts were silnilar to expected m.JTnbcrs. However, workers with medium to high exposures were found to have twice the risk of developing leukemia compared to workers with low exposures. Earlier findings sugge&ting an increased Iisk of Inyelofibrosis were not observed in the follow-up study. docmnentation describing any methods used to analyze blood nor changes in the laboratory routines and instrumen- tation over tirne were found. The authors determined that "i.n a group of workers \l.rith substantial benzene exposure;' that there was evidence for an exposure-response relationship for benzene and effects on white blood cell count (WBC), with a lesser effect on red hlood cells. [1he authors estimated that 1naxirnun1 daily benzene doses in the worker population were Paxton and performed a comprehensive review ofadditional data released and published N10SH involv- 34 ppm; however, they did not consitler either the exposure estimates ofthis cohort conducted by Crump and Allen (1984) ing the Pliofilm cohmt (Paxton, 1994). Several observarlons were made: or Paustenbach et al. (1992).] No other changes in blootl testing results were thought to be ofsignificance [Smoking and recent infection do not appear to have been considered] ('Ward et al., 1he absence ofadditional cases ofmultiple myeloma in the 1987 update "weakened to nonsignificance" the prior published association with benzene exposure. 1996). Yin et al. provided an expanded update for a large Chinese worker population initially studied in the late CGU BEN0000267 Benzene and l1uman. health 11 1980s, reporting on about 75,000 benzene-exposed work- ::>. No disease category or combination of disease catego- ers and a control popuLalion of around 35,000 workers with ries reported by IIayes et aL demonstrated a consistent no benzene exposure. The authors found an increased risk dose-response trend for cumulative benzene exposure of myelogenous leukemia in the worker population, par- and increasing risk of disease (Hayes et aL, 1997). AML, but also CML to a lesser degree risk provided the significant leukemia subtype relationship). 1hey also reported elevated increased risks for developing acute lymphocytic leukemia (1\LL) and non-Hodgkin's lymphoma, but did not find an association with Hodgkin's disease, multiple myeloma or chronic lym- phocytic leukemia (CLL) (Yin eta!., 1~~fi). Hayes et al. extended the analysis ofthe Yin et al. cohort to include estimates ofworker exposures, concluding that work- ers with relatively low-level chronic exposures to benzene workplace exposures less than 10 ppm and cun1u- lative exposures less than 40 ppm-years) were at significant risk for developing hematologic malignancy. With increasing levels ofexposure, the authors detected a marginal tendency for increased risk for nonlyrnphocytic leukemia (ANLL) and for the combined category of /\NLL and myclodysplastic syn- drome (MDS) (Hayps f't aL, 1997). Because their reports have had a measurable iinpact on fi. ll<e implicit assumption made hy Hayes that MDS represents an early stage of AML is debatable. Combining these two diseases to achieve more statistical significance may not be valid, particularly if MDS patients in the cohort did not uniformly go on to develop leukemic transformation (Hayes et al., 1997). 7. Both study populations in Yin et al. {1996) were noted to have rnuch lower all-cause mortality nurnbers \\Then compared to the general population of the region. The reasons for significant mortality deficits in the study populations were unclear, but the authors suggested lhat lhis could have been related to diagnostic criteria for in the regions being different from their studv criteria. Data for local mortality were restricted to the ].CJ73-l975 ti:me period, and Tates were not available for all 12 cities being represented in the study (Yin et aL, 1996; Hayes et al., 1997). the perceived health effects of human exposures to benzene at fairly low doses, it is worthwhile to note a few aspects ofthe Yin and el al. sludies (conducted in cooperalion wilh the National Cancer Institute). Because ofthe concerns raised in the study, a series of additional studies were initiated in China to help deal with the shortcomings in their original work. Some of the results were presented at the International Conference on Benzene L There were no corrections or considerations for multi- ple or alternative exposures in the worker population. According to \Vong, 95% of the exposed Chinese work- ers were also to other chemicals, and only 5'J"{, described themselves as being exposed only to benzene (Wong, 1999). held in 2009 (Bird el al., 2010). In 1996, Rothman et al. evaluated a subset of patients from the Yin and studies. He looked at 44 workers vvith relatively high expm.ures to benzene (average of 31 ppm as an 8-hour T\1\TA) and then compared them to 44 age- and gender- matched controis, evaluating the t\110 groups in terms ofa variety ofhematologic measurements (Rothrnan et aL. 19~lfi). He found 2. The exposed population was considerably older, on aver- that these workers had reduced total white blood cells, age, than the unexposed population with respect to age lets, red blood cells, and hematocrit values compared to control at first employment. percent of the benzene- subjects. For workers who were never ex1Josed above 31 pp1n .s exposed group was older than 30 at first employment, on 5 days of sampling (with a median 8-hour T\N"A exposure compared with only 11% of the unexposed group (Yin of 7.6ppm), the absolute lymphocyte count was et al., 1996; Hayes et aL, 1997). '"''u'''-''"'"''Y siigrtii.cantly d.i111en~nt compared to controls. :-L There were significantly more women in the compared with the unexposed group. If, as some have opined, women were ITtore susceptible to lhe effects of benzene (Brown et al., 1998; Duarte-Davidson et al., 2001 ), this sex discrepancy could magnify the effects of benzene in the two test populations. 4. Wong observed that prior to the National Cancer Institute (NC:I) involvement in the Chinese worker studies, the investigators did nor attempt to measure exposure in the cohort as a \Vhole, but only reported benzene measurements in the leukemia cases {\!Vong, 1999). Overall, \Vong felt that worker exposures were seriously m1derestimated by ~CI and the Chinese Academy of Preventi\.-e Medicine, and that these estimate& had a profound impact on the dose-response data reported in the Yin, Hayes, and later publications involving this cohon (Hayes et al., 2000). Rotlnnan ei al. concluded that the ALC was tl1e rnosl sensi- tive indicator ofbcnzenc-induccd toxicity to the bone mar- row in the population studied, particularly at the lower levels of exposure (less than :-11 ppm). It was noted that: "1ndividua1 benzene air levels in these factories were much higher than we had expected to find, based on historical monitoring data:' As a result of improved occupational standards and workplace practices over the prior few decade;,, the authors opined that: patterns in the three factories evalu- ated in thb study are not representative of general exposure patterns in China today" (Rothman et aL, 1996). A cohort study of 19,000 service station workers in Nordic countries exposed to low levels of airborne benzene did not find an increased risk for leukemia, nor for acute myelogenous leukemia specifically. Average exposures were estimated to be less than I mg/m3 (approximately 0.3 ppm) (Lynge, l9~J7). In 1999, Khuder conducted a study of 105 petroleum workers exposed to low levels of benzene over an 18-year CGU BEN0000268 12 D. Galbraith et aL period. He suggested that "CRC values" could serve as a 2000 to the present useful wolto monitor workers. In 3855 total complete blood Scientific advances in1oxicogenomics and molecular biology count (CBC) records, the rneans, medians, and the have enabled an increasingLy detailed perspective regarding of the CBC: results were vvithin the normal ranges provided proposed mechanisms of benzene-related disease, "precur- by the laboratories. None of the individual CBC parameters sor" stages to and suggested factors (genetic, studied had a mean for any year of the period studied that environmental, etc.) that could result in an increased was ouLside of lhe normal reference range, and none of the predisposilion for malignant hematopoielic disease. Although individual CBC parameters was correlated with the degree of scientific disciplines have made usc of an array of greatly benzene exposure. 1he average T\'VA exposure for the work- improved tools to study the health effects of benzene over ers studied was 0.81 ppm. It was concluded that "whether the last 10 years, we unfortunately are still searching for a chronic, low-level benzene exposure can affect the RBC clear understanding nf the mechani,;ms nnderlying leuke- count, hemoglobin, MCV and platelet count values cannot mic transformation related to benzene exposure be stated conclusively in the smdy presented here. Hmever, and Eastrnond, 2010; Laskin et aL, 2000; Kalf, 2000; frons, our findings are consistent with the existence of a threshold 2000; Larson, 2000; French and Saulnier. 2000). The specific for insult.... It is possible that higher levels of lifetime cumulative doses of benzene over time to a exposure are needed in order to show any effect on the \VBC heightened risk for disease also remain uncertain, although or RBC counts" (Khuder et aL, 1999). studies seem to have narrowed the likely threshold dose to Goldstein, in a letter to the editor to the publishing jour- somewhere benveen 40 and 400 ppm-years; a surprisingly nal, Journal of Occupational and Environmental A1edicine, \Vide range in light of nearly 100 years of study. took issue with the Khudcr ct al:s conclusion ofthc value of Reflecting this uncertainty, Ross noted in a comprehen- CRC results in monitoring low-level exposures to benzene sive toxicologic review of benzene toxicity that "there is no (Goldstein and Cody, 2000). In particular, he pointed out conclusive evidence for a single 1netabolite or combination that ': .. the finding of small bm statistically significant effects of metabolites as being responsible for benzene toxicity:' within the normal range over a 17 year period may simply Benzene was accepted as being an inducer of ane- be due lo of the cohon. IL is lhal mia, myelodysplasia, and acute leukernia f"ollowing in men, red cell parameters decline with age beginning at chronic exposures, but other diseases of the bone marrow approximately age 30:' The study also had the curious finding were not identified as having an association vvith benzene of a decrease ln the 1\.fCV, coHtrary to what has been nonually (Ro!>s, 2000; Gross et aL, 2010; Irons et al., 2010; Ross and noted in animals and humans subjected to benzene exposure. Zhou, 2010). 'Ihe authors did not employ a control group for comparison, ]he issue of disease latency for benzene exposure and its ;md Goldstein and Cody noted that none of the individual association '.JVith future malignant disease has been a topic of CBC components showed any relationship to benzene expo- discussion for a nurnbcr of years. For cxarnplc, Finkelstein, sure, making the value ofthe testing, particularly for low-level in a review of original data fmm the Pliofilm cohort, found exposures, seem that the greatest risk for developing leukemia was related to A small nested case-control sludy on Canadian petroleum benzene exposures sustained in lhe 10 years prior to distribution workers who long-term, low-level nosis, with exposures more than 15 years from diagnosis not .s exposures to benzene was reported in a 1996 paper by being significantly different from matched unexposed conresearchers at Exxon {Schnatter et al., 1996). lhe audmrs trols (Finkelstein, 2000). 11lis observation was latet reinfmced stated that in workers with daily exposures (8-hour TWA) in the Rinsky update of this cohort reported in 2002 (Rinsky to benzene ranging from 0.01 to 6.2 ppm that there was et al., 2002). no ohserved increased risk for developing leukemia with Glass also commented on the question oflatency, condud- increasing cumulative exposures. Ref1ecting on their study from her review of the Health T!Vatch data that "benzene- and three previous epidemiologic studies with lifetime expo- induced leukemia can be restricled to the period up to 15 sures under 100 ppm-years (Bond et al., 1986; "Paxton et aL, years prior to et al., 2004). :\'lore r<>>rPcntl" 1994; Wong, 1987) the risks for all types of leukemia were Richardson performed another review of mortality data from ~ .. approximately distributed around 1.0" (Schnatter the Pliofilrn study, determining that the association between et aL, 1996). benzene exposure and leukemia mortality was strongest in A similar study of petroleum distribution workers by the 10 years inLmediately following exposure, was weaker in Rushton and Romaniuk concluded that there was no associa- the 10-20-year period, and that there was ~ .. no evidence of Tion between benzene exposure and lymphocytic leukemias, association 20 or more years after exposure" (Richardson, acute or chronic. 1herc was the suggestion of an association 2008). with myeloid forms ufleukemia, but the findings did not dis- In the last l 0 years, several refinery studies have been pub- v-.'ith zero cases reported in lished, addressing the question of increased mortality from the highest (245 ppm-years). The authors leulcemia and other malignancies. In 2001, a observed lhal over 80% of lhe workers had less !han 5 ppm- comp1ehensive cohorL monalily study of over 3000 employ- years ofbenzenc exposure, making estimates for risks above 10 ees from a petroleum refinery over the period from 1959 to ppm-years highly imprecise (Rushton and Romaniuk, 1997). 1997 did not find any increased mortality from leukemia or CGU BEN0000269 Benzene and l1uman. health 13 any of its subtypes in all workers, or in the subgroups of main- more. TI1ey stated that their data did not provide any evi- tenance workers or process workers. Duralion ofemployrnent dence ofan exposure threshold for developing hematopoietic was used as a surrogate for worker exposures, as insufficient malignancy, and did not find any association with ~HL or exposure data ,,,,-ere available to construct quantitative expo- MM (Glass et al., 2003). In a letter to the editor, Schnatter sure indices (vVong et 2001). suggested that there was an unusually low rate of leukemia In 2003, a cohort mortality study of over 25,000 Canadian in the control population, and that the relative risk numbers petroleum workers hired between 1964 and 1994 did notiind reported by the authors were uncharacteristically high when any evidence of increased HlOitatity frmn leukemia, ly:m- cmnpaicd with similar case-control studies involving ben- phoma, or other lymphohematopoietic malignancies. Again, zene, even "higher than those found in more highly exposed length of e1nplnyment was used to estimate lifetime expo- cohorts used in risk assessments" (Sdmatter, 2004). sure;;, and "similar exposure group" (SEG) codes 'vereused to r:roldstein also pointed nut that a substantial numher nf categmize workers by location, job description, the leukemia cases by Glass were of the chronic and dates of empluyment. The SEG codes contained estimates lymphatic leukemia subtype, "a disease that has not been oflikely exposure constructed by an industrial hygienist using shown conclusively to be caused by benzene" (C1oldstein, available peninent data (Le"'is et al., 2003). 2004). Glass 17 chronic leukemias, and 1he Australian Health Watch study of petroleum industry 11 cases ofAML (Glass et 2003). Goldstein suggested that employees has provided two reports in this decade. In 2001, surveillance bias, or testing the case population of exposed it was reported that there was an increased incidence ofmul- workers more than the unexposed control population, could tiple myeloma, as well as a statistically significant increase in have led to preferemial case discovery, and could explain the the incidence of "all leukemias combined:' Ifthcsc results had ~ ..unusually low levels of benzene exposure associated with been confirmed, because theywere associated with fairly low leukemia" (Goldstein, 2004). doses, have been important. The au[hors noted, the lack ofassociation for leukernia with benzene however, that': ..the combining of allleukaemias into a single exposure in the 2005 Health Watch Report, Glass et aL (2006) 'leukemia' entity is somewhat since their found that low-dose benzene exposure was related to the occurrence and natural histories suggest Lhal incidence of other forms ofleukeinia. Tn a reanalysis of Lheir ably different diseases" (Gun eL aL, 2000). Leukemia mortality 2003 study, they acknowledged the problems with their origi- in the cohort was slightly but not significantly elevated, with nal referen cc group and combined their two lowest exposure an overall leukernia rnortality ratio of L 16 (confidence inter- groups to "erve a~ a cmnpaii,.un for their ''heavy" exposure val [CI] = 0.66-1 Numbers ofthe individual leukemia sub- group. This new comparison changed their odds ratio from types were not provided (Gun et aL, 2000). Total hydrocarbon 98 in the original study to 5L9, which then dropped to 7.8 exposure was estimated from employee codes, ranked when they accounted for exposures related lo "high el,-po- by a committee ofindustrial hygienists into seven categories of exposure; uncoded jobs were assigned an intermediate sure events," or industrial accidents/spills that were judged to have been relevant for individual job types {Glass et at, :woo). "default" category of exposure. Leukemia subtype incidence was also calculated for various In the laLest 2005 Health lrVatch update, the previously exposure subgmups, combining the three lowest exposure reported leukemia excess in the cohort disappeared (Gun groups and the nvo highest exposure groups into "low" and .s et al.. 2005). Neither total leukemia incidence, nor total "high" categories due to small numbers of cases. Glass et aL leukemia nwrtality were ap(Jredably elevated {L07 stand- reported a single statistically significant finding; an increased ardized incidence ratio [SIR] and 0.99 SMR, respectively). risk ofANLL in workers with exposures >8 ppm-years. They Interestingly, chronic forms ofleuke1nia were observed more noted difficulty in comparing their findings to other epide- fTequently than acute forms, at variance with previous :studies miologic studies involving henzene exposure, which had that have evaluated chronic exposures to benzene. This findis not unexpected given the lack ofstatistically significant primarily been based on mortality, not incidence et aL, :woo). Interestingly, the lateslllealth Watch Report (the par- differences observed in the first study. Furthermore, neither ent population for the Glass ct al. nested case-control study) duration of employ1nent nor magnitude of benzene exposure found no cases of AML diagnosed in the lowesl lhree (of appeared to be related to leukemia incidence. The authors seven) job exposure categories, ordered by total hydrocar- admowlcdged that ': .. only acute ANLL is likely to be causally bon exposures. Leukemia incidence was abo not statistically related to benzene exposure," while noting that there was no different fnun the population with an SIR of L06 (C:I excess ofANLL in the cohort compared to the general popu- 0.53-L90) (Gun et aL, 2005). lation. Low-dose exposures leading to disease were also not i\ small studyofshoe factory workers in Italy found increas- seen, as there were no cases of ANl.l, seen in the three lo-west ing leukemia mortality risk with increasing cumulative ben- occupational expm.ure categories described the authors zene exposure (Costantini et aL, 2003} Their results, however, (Gun et aL, ~005). are difficult to assess, as exposures to benzene were noted After the of the Health Watch 2001 update, by the authors to be reduced to zero after 1965 and more Glass el aL (2003) published lhe resuhs of a nested case- lhan half of the deaLhs from "hemaLolyrnphopoielic cancer" control study, finding that leukemia risk was increased at occurred after 1985, more than 20 years after there was no cumulative lifetirn.e benzene doses of just 2 ppm-years or further benzene exposures at the plant. Leukemia subtype CGU BEN0000270 14 D. Galbraith et aL analysis shnwed that there was only one case of "acute in red and white blood cell counts seemed to correlate with lev- leukemia" and one case of"acute myelogenous leukemia" in els of benzene exposure (Qu et aL, 2002). However, contrary to these post-1935 cases, and cases ofnon-ANLL subtypes the findings ofRothman et aL (1996), the authors reported that (including three cases of multiple myeloma and two cases of neutrophil counts were more sensitive to benzene effects than NHL). Thus, the frequency of diseases traditionally associ- lymphocyte counts, with little effect fron1 exposure duration. ated vvith chronic exposures to benzene was drasti- Mean neutrophil counts also displayed an opposite trend with cally reduced after cessation of benzene exposure, and the benzene exposure intensity, inCl'easing in number irmn lhe bundling by the authors of diseases into a combined "blood lowest exposure group to the next two groupings ofcurm.Jiative malignanL-y" categmy is of unclear value in understanding exposme, then decreasing in the highest mean exposure group. the contribution of benzene to nwrtality in these workers. This study used a relatively sm.all control group (51 unexposed Tn 2002, et al. provided another update nn the and 130 exposed workerl>i), and exposures tn other disease incidence in the Pliofilm cohort (Rinsky et aL, 2002). chemicals, as well as prior smoking and histories Five additional benzene-exposed workers were found to have were not well described. died from leukemia, but a careful examination of the cohort A study ofhematologic paramelers in petrochemical work- revealed that the relative risk ofleukemia from chronic ben- ers with benzene exposure was conducted, looking at six zene exposures dearly diminished over time. whereas earlier separate complete blood count 1neasure1nents. The authors studies ofthese workers suggested an increased risk for MM, found that there was no increased incidence of any blood the 2002 update largely reversed these findings. There were component abnormality among exposed employees. I\1ean four additional cases of multiple myeloma described: lhree exposures were reported to be 0.60ppm in the 1977-1988 time of which were judged to be in workers without exposure frarnc, and 0.14 ppm in the 1988-2002 time frame. 1he study to benzene; the last was exposed for only a month, with a was limited by not having indi'lridual PXposure data, nor did cumulative benzene exposure ofjusHU 0 pptn-years. it evaluate possible confounding variables, such as issues were also problematic in this update, as 6 of the 15 lifestyle or other chemical exposures (Tsai et al., 2004). deaths from leukemia occulTed more than 30 years after the In another study of250 C-hinese workers published in the worker's first exposure Lo benzene, a resulllhal is generally same year, il was found that exposures Lo less Lhan 1 pprn considered to be well outside the expected latency period for of benzene in air could rise to reductions in several benzene and leukemia. Thus, for 40% ofthe Pliofilm leukemia white blood cell parameters and platelets (Lan ct al... 2004). ca!'.es, benzene exposures at the plant were unlikely to have Howevei; the1e were ~everal potential pruble1ns with this played any significant role in their disease, particularly for workers with excessive disease latency who were employed less ihan a year at !he fadliiy. There were substantially more smokers, both numeri- An updated mortality study of petrochemical workers at cally and proportionately, in the control group compared two California refineries found that there were no significant with the study groups. This difference was most notable increases in mortality from leuken1ia overall, or from any in the lowest benzene exposure group (28% smokers in individual subtype. Multiple myeloma, however, was sig- control group, 18% smokers in exposure group). nificantly increased, but only in workers who started employ- There were no data provided on past smoking behav- .s ment before 1949 (Satin et aL, 2002). ior in the study participants. The effects of cigarette In 2002, occupational exposures to benzene in gas and smoking can linger for years, and the degree of former electric utility workers in France were evaluated, under the usage in subjects who may have recently quit could be assumption that exposures to petroleum solvents would important. entail exposure to henzene as a con1ponent of those sol- vents (Guenel, They found that the risk of leukemia was increased : .. only among workers with first exposure to benzene before 19Go; but even this subgroup did not achieve stalistical significance. There were very few cases in : .. a significant :residuaL smoking effect on white blood cell count after was observed in our follow-up of 6.5 years" (Sunyer et aL, 1996). this study, and the only association that did achieve statistical "It is possible that the association of smoking with the significance was the finding of an increased rate of"all acute leukocyte count persists because ex-smokers have leukemias;' cornbining myeloid and lyrnphocytic subtypes, bmnchial inflammation that declines gradually with at the highest level of exposure in the cohort. lhe authors time since quitting" (Petitti and Kipp, 1986), stated that "no association with a particular leukemia cell type was apparent:' Contrary to the findings of most other Control subjects were noted to have the highest inci- epidemiologic studies involving benzene, they found that dence of "recent infection" of any of the study groups. leukemia seemed to he more associated 'INith benzene expo- C-onsistent vvith this, the standard deviation of the total sures ': .. after allowing for a 10-20 years latency Le., white blood count in control subjects was reported to be with dislanl pasl exposures." substantially greater than that reported for Lhe low and Peripheral blood effects from benzene expo- high exposure groups. These observations could indicate sures in Chineseworkerswere studied, finding that depressions that a fe'\'li control subjects were ill, leading to elevations CGU BEN0000271 Benzene and l1uman. health 15 in their \NRC counts and a larger SlJ for the control cases, the authors concluded: ': ..this study does not support group, which might have impacted the study findings. claims that exposure to benzene affects risks for lymphohae- Pregnancy status and use of oral contraceptives, in a study population that was predmninantly female {about 2/:-1 of subjects) were not contTolled fnr, which may have led to systematic bias from failing to consider these potentially confounding variables (Fisch and Freedman, 1975; Dodsworth et aL, 1981). matopoietic malignancies other than ANI.c' (Sorahan, 2005). Exposure data for individual cases were not available. A small cohort of 31 l Dutch workers exposed to rather cumulative doses of benzene (average of 159 ppm-years) iiom 1951 to 1968 was studied up to 2001 to evaluate tbe risl of leukemia. TIIC authors condudcd there was "Ko excess leukemia despite subst~mtia] exposure to ben- Of note, comparison of the low and intennediate benzene exposure groups revealed that a substantial proportion ofthe hematologic parameters studied (total \JVBC, granulocytes, CDS T cells, natural killer (NK) T cells, monocyte counts, and lymphocyte counts) increased or stayed the same with a 5-fold mean increase in benzene exposure, results that colm- tered the overall trend described by the authors. In ::!004, a cohort mortality study of 2266 chemical work- ers exposed to benzene since 1935, a mild increase in total leukemia was noted (SMR = l. as well as a slight increase in ANLL (SMR= Ll TI1e of risk appeared to increase with higher cmnulatiYe exposure to benzene, bul ditl not seem lube related to intensity of e:o::posure. below approYJmately 30 ppm-years were not seen to result in an increased risk for ANLL or total leukemia (Blocmen et al., 2004). A detailed review of case-control and cohon studies of benzene-exposed workers was performed by Schnatter et at (2005), characterizing these according to industry focus, presence and adequacy of exposure assessment, presence of leukemia subtype determination1 and quality of controlling for potential Yariab1es. The authors concluded that there was a consistent finding ofincreased risk for AML with benzene exposure across the studies they reviewed, a zene" (Swaen et aL, 2005). 1hey further opined that despite the small sample .size (only 311 workers were followed), their findings supported the notion that leukemia risk from benzene Likely had a threshold of exposure that needed to be surpassed prior to generating an increased risk of developing leukemia. 1hcyreported an SMR of85.6, with CT ofLl to433. In an update of North American synthetic rubber work- ers, researchers found a slightLy increased incidence of total leukemia (SMH= 116), mostly in long-term hourly workers. 1he study included employees from any of eight synthetic rubber plants who worked for at least one year to 1992, and used work histories to classify individual into various occupational subgroups. Workers were exposed to a wide variety of chemicals, hut no exposure data were nresente,il. and lhe excess of leukemia seemed lobe concenlraled in workers hired in the 1950s. Leukemia subtype information was absent in 28% of the cases. The authors stated it was dif- fin!lt to nrake definitive staternents about risks for individual leukemia subtypes due to the small numbers of cases, lack ofinformation on nonoccupational variables (such as smok- ing, diet, weight), and relatively small overall increased Tisk (Sathiak:umar et aL, 2005). A review of occupational benzene exposures at a chemical relationship that seemed to be more significant with increas- manufacturing and Paustenbacb looked at ing study quality. 1his relationship did not hold for any other 3700 air samples ofbenzene across a variety ofjob categories leukemia subtype. Instead, sporadic and inconsistent asso- and plant locations (Williams and Paustenbach, 2005). These .s ciations were described for other forms oflcukemia, although represenced personal \'lrorker san1pling, area T\NA nleasure- the authors noted that sparse data, for ALL, pre- ments, and short-term area measurernents. The authors con- cluded more definitive conclusions. cluded that chemical operators in the facility sustained daily A broad review of occupational exposures and their rela- T\VA benzene exposures of about 2.0 ppm from 1976 to 1981 tionship to hematologic cancers was provided by research- and l.Oppm from 1932 to 1987. The results supported a long ers in France in 2005. Radiation exposure and "high daily held belief that in the maJor corporation, since about 1!:155, exposure to benzene (more than lOppm)" were described as airborne concentrations ofbenzene have tended to decrease being strongly associated with AML. "Cohort studies ofwork- over time, consislenl with the changes in either lhe ACGIH ers in the petroleum, gas and electricity indllstries have not TLVs or OSHA PELs (as originally proposed by Crump and shown any significant excess risk of other types ofleukemia, Allen in 1984) (WU!iams and Paustenbach, 2005). No health and in particular of chronic myeloid leukemia:' The authors effects were evaluated. also concluded that there was insufficient evidence to claim In 2007, Natelson provided a review ofbenzene and acute a causal association for benzene exposure and either NHL or myeloid leukemia from a clinical perspective. He studied multiple myeloma (Descatha et aL, zoos). benzene dose and latency issues, discussed mechanisms In 2005, Sorahan reported on a cohort of over 5500 of disease, chromosomal studies, and certain occupational benzene-expo"etl work-ers in England and \Vales prior to cohorts. He concluded that in developed countries, due to 1967. Acute nonlymphocyti.c leukemia {ANLL; often used improved intlu;.trial and reduced expm;ures to work- interchangeablywitl1 MIL, but is a bit more inclusive a term) ers, AML was not likely to occur as a result of benzene expo- was found with an increased frequency that was not statisti- sure in modern petrochemical facilities (Natelson, 2007b). cally significant; all other forms of leukemia were not found Tsai et aL (2007) performed a cohort mortality study of in excess. Ihough the study had a relatively small number of l 0, 621 employees in the petroleum refining industry who were CGU_BEN0000272 16 D. Galbraith et aL employed benveen 1948 and 2003. No statistically significant relationship betvveen benzene exposure and acute increase of overall leukemia or any leukemia subtype was non-lymphatic leukaemia" [Multiple m.yelorna and discovered. In workers with 20 or more years ofemployment, NHL were specifically excluded from this relationship] there were seven cases ofANLL observed, con1pared '-"'ith six (ECB, 2007). expected, an SMR of 1.211. The authors opined that their results were consistent with the notion that benzene exposures in the past at the iacility were too low to result in an increase in ANU, or AML (Tsai et al., 2007). The NCI study a minimally statistically increased risk of J\l\fLL with a relative risk (RR) of 3.0 (95% CI L0-8.9). However; the data did not demonstrate a con- sistent increased risk of malignancy with increasing expo- W'hich hematopoietic diseases are associated with henzene? sure duration ur cumulative exposure. Combining ANLL and MDS cases into a single disease cm:egory resulted in a Benzene exposure and acute leukemia stronger association with benzene exposure (RR of 4.1, 95% CllA-11.6) (Hayes ct al., 1997). Acute myelogenous leukemia, acute myeloid leukemia, and acute nonlymphocytic leukemia are terms frequently used in the literature to de,;,cribe a group of hematopoietic diseases of acute onset that are distinct from those with lyn1p hocytic ori- gins. Most frequently, the malignant cell type is a rnyeloblast, or monocyte or myelocyte. In about 5-10% ofcases, however, the condition involves erythroblasts (prcd- ifferentiated red blood cells) or {cells that eventually differentiate to produce platelets) (List et al., 2004). These cell types are contained v.'ithin the French-American- British (Ic~B) classification as Mfi and M7 leukemias. There were 13,300 new cases of Al\1L in lhe United Stales in 2008, with over nine out of ten cases occurring in adulls (American Cancer Society [ACS]. 2008a). Of all the claims about the chronic hazards ofexposure to benzene, the weight ofevidence from the epidemiology literature is thatAML is the only leukemia dearly shown to be elevated. Many government health have weighed in on the relationship between benzene and future risk ofdeveloping AML. The Pliofilm cohort provided a broad range of worker exposures to benzene and, unlike many other industrial cnhnrti'i, had few other chemical exposures. The initial published study for the Pliofilrn cohort occurred in l!:l77, and reported a 10-fold risk of dying from : .. myel- ogenous and monocytic leukemia" (Infante et al., 1977). In 1987, Rim,ky et al. reported a standartlizetl mortality ratio of 3.37 for all leukemias in the entire population, with workers having less than 40 ppm-years of exposure not observed to have any Ineasurably increased risk. iNorkers at higher levels of exposure did demonstrate an increased n:wrtality trend for leukemia, vvith the exposure group than 400 ppm-years) being 66 tin1t~s more likely to die frornleuke- mia than control Rinsky's follow-up study, reported in 2002, showed that the leukemia risk diminished over time, falling to an aggregate SMR of 2.56 for aU leukemias in the population (Rinsky, 2002). Of the 17 leukemias reponed in the history of the Pliofilm study, 14 were myeloid and only one was lymphoid. Two cases were unspecified. l'VkCraw et aL described a retrospective mortality study "Epidemiological studies and case reports provide clear of white males at an oil refinery in Illinois, finding; a statis- evidence ofa causal relationship between occupational tically significant excess of deaths due to leukemia, chict1y exposure to benzene and henzene-conlaining solvents AML (McCraw et al., l~JHS). A case-control study .s and the occurrence of acute rnyclogcnous leukemia (AJ\1L)" (ATDSR, 2007). showed that cases were not exposed to more benzene cornpared to controls, and the reasons for the excess leuke1nia could not be identified (Austin ct aL, l9BG). "Epidemiological studies ofbenzene-exposedworkers have demonstrated a causal relationship between benzene exposure and the production of myelogenous leukaemia. A relationship between benzene exposure and the production of lymphoma and multiple myeloma remains to be clarified" (WHO, 19~13). Sathiakumar etal. desnibed a positive a;;;,ociatiun bet1Neen oil and gas field work and ANIL, with a trend of increasing risk "v:ith increasing duration of employment (Sathiakurnar et al., 1995). K.iikeleit et al. examined hematologic disease incidence and mortality statistics for upstream workers in Norway from 1981 to 2003, reporting a significantly increased risk of developing AML (RR"" 2.R~, 9.S% CI: 1.25-6.67), particularly for workers employed earlier in the "1he strongest epidemiological evidence that benzene causes cancer is from several cohon: studies in various industries and geographical locations, which found that occupational exposure to benzene increased the risk of mortality from leukemia (mainly acute myelogenous leukemia)" (NTP, 2005). sludy period (Kn:keleit et aL, 2DOH). In contrast to these studies, rnany other occupational studies of chemical workers and petroleum industry workers in the United States and Canada (Bloemen, 2004; Wong and Raabe, 1995; Wong and Raabe, 2000; Tsai et al., 2004; Swaen et al., 2005; Satin et al., 2002; Lewis et al., 200:~) have not shown statistically significant increases in the rates ofany lymphohernatopoietic malignancies, including all leukemias "There is sufficient scientific evidence from the numer- and Al\1L looked at as separate categories of disease. These ous human epidemiological studies to assume a causal findings might be partially explained by the level of worker CGU_BEN0000273 Benzene and l1uman. health 17 exposures in these industries likely being considerably were supported by histologic confinnation of disease in only less than those experienced by the Pliofilm cohort, worker half ofthe cases cited. 1he very low rates ofleukemia in the populations studied in China, and the upstream petroleum control population used also arc suspect, as pointed out by industry. Schnatter (Schnatter, 2004). Regulatory agencies and various scientific bodies are in Son1e researchers have suggested that the metabolism of agreement that exposure to benzene excess benzene may differ at low concentrations of exposure com- of 40 ppm-years) results in an increased risk for myeloid to higher levels. lbat is, some researchers have sug- forms of leukemia, chiefly AML It is difficult to define the gested that different metabolites are fanned at different rates exact dose-response relationship for humans, since there are at different doses. Kim et al., for example, reported that indi- clearly differences in interindividual susceptibility. The \Norld viduals exposed to airborne benzene concentrations ofless Health Organization (vVHO) appear.s tn agree with the early than 0.1 ppm metabolized benzene ahnut nine times mnre findings, that exposure to 1 ppm as a time-weighted efficiently than heavily exposed workers. "They concluded average exposure to benzene over a 40-year working career that the health risks of low and very low exposures may be (40 ppm-yearsJ has not been associated "with any increased underestimated by toxicokinetic models currently in use, deaths from leukemia" (vVHO, 1 particularly for exposures less than 3ppm (Kim et al., 2006). 1here is an ongoing debate regarding the level of cumu- Rappaport has recently expressed the belief that there are, lative exposure to benzene that might represent a disease in fact, multiple pathways for benzene metabo- threshold for developing AML (Rushton and Romaniuk, 1997; lism: a low concentration enzyn1e, or "high-affinity" enzyme, Schnatter et al., 1996; \'l!illiams el al., 2008; Johnson et al., operant at lovv levels of expostue, and a high concentration 2009). Some scientists believe that there arc data suggesting enzyme, or "low-affinity" enzyn1c. "I he "high-affinity" enzyme that some 1isk ofvmious effects on the blood, including Al\1T, he lwlieves to be mcne likely to generate toxic metabolites exists at very low doses and point to the Lan et al. (2004) compared \llich the enzyme (Rappaport et al., study, which reported peripheral blood effects at workplace 2010). benzene concentrations of less than l ppm (assuming life- Various ofgenes involved in ben- lime exposure for 8 hours/day, 5 H benzene at zene metabolism have been proposed as biomarkers lo iden- these low doses impacts peripheral blood cell counts, then tify members of the population who may be al increased {or one might conclude that benzene could be having an effect decreased) risk ofdeveloping adverse biologic effects follow- ou the bone rnarrow. ing benzene expm.ure (Russ andZhou, 2010). Understanding Although evaluating peripheral blood progenitor cells may the various pern1utations of these genetic polyn10rphisn1s in be the least invasive method of evaluating more immature a given population, and how they might affect individual sus- cmnponents of the blood, il cannot be assumed to renect however, is an task, but may what is occurring at the level of the bone marrow in an well deliver considerable benefits for our understanding of occupationally individual, particularly at very low the intricacies of benzene-related disease susceptibility. An doses. In other words, den1onstration of function excellent review of this topic has heen recently by in a peripheral blood cell cannot be assumed to reflect cel- Dougherty et al. (Dougherty et 2008), lular damage or abnormalities in other body compartments, Back in the early 1980s, in an attempt to hreak the stale- .s nor has ii been shown wherher these possible changes have mate that had occurred between the various stakeholdprognostic significance for any acute or clnonlc disease proc- ers (e.g., regulatory agencies, the regulated curnnrunity, ess (Stockstad, 2004). nongovernmental organizations [NGOs], and academics) 1he Glass nested case-control study of Australian petro- involved in dn1fting legislation regarding carcinogens, the chemical workers {Glass et al., 2003) also reported that very tenn "practical threshold" wali offered. Thi,; ten11 seems to low level exposures could lead to leukemia, far below what be particularly appropriate for benzene. For purposes ofthis was previously believed. As we noted earlier, however, this discussion, a threshold represents a risk that is too study does not provide enough information to establish small to be measured in an epidemiology study or toxicology a solid foundation regarding benzene risks at low doses of study. In short, one cannot rule out thal a particularly sus- exposure. For example, a substantial fraction of their cases ceptible person in a very large population might not develop were chronic leukemias, both CLL and CML, diseases that a particular disease, but that the risks are so low that they have historically had less compelling associations with ben- cannot he accurately quantified. In rnany cases, there rnay zene exposure (and are classified in different categories of actually be no increased risk at a particular level of exposure, disease by the WHO) (Glass ct al., 2003; Swcrdlowcl al., 2008). but science is unable to determine an answer one way or the Addressing "aU leukemias" as a single entity is a problematic other. This concept was recently discussed by the Canadian epidemiological approach when seeking to understand the Science Advisory Panel, which was a&ked to evaluate the risk potential ofbenzene, due to the differ- of mesothelioma associated with exposure to low concen- ences in these various hematopoietic diseases. to trations of chrysotile asbestos 2009). characterize subgroups ofleukemia (such as ANLL) in their no led LhaL it would be very difficulL i.o show i.hal there were study were hampered by low case numbers, leading to risk or were not health risks assodatcd Vvith doses that seemed to estimates with very poor precision, and cases of leukemia produce risks ofless than 1 in 10,000 in human populations. CGU BEN0000274 18 D, Galbraith et aL A substantial fraction of toxicologists support the notion recently explored. Follmving a cmnprehensive review of that there is a safe dose for virtually all substances, even geno- case-control and cohort studies, it was concluded [hat there toxic carcinogens (Scott, 2D08; Waddell, 2006; Hooker ct aL, is insufficient epidemiologic support for such a relationship, 2004; Weisburger, 2001), which aligns with the views of the and that the literature to date has been characterized by a ACGIH and the OSHA, the two bodies providing the lack of exposure data, and the frequent failure to separate nant leadership in the United States tolerable or out cases of childhoodAlviL from ALL. A wide variety of con- acceptable exposures ofworker's to airborne chemicals. founding variables were also discussed by the authors (Pyatt Studies of hmnan DNA repair mechanisms have shm.vn and Hays, 2010). that me various repair pathways are impressively efficient at Wong et aL (20l0a) reponed on the findings of a hospi- identifying and repairing mutations (Hartvvig, 201 0; Wogan tal-based case-control study for the development of AML et aL, 2004). These mechanisms are responsihle for identify- in Shanghai, China, as part nf the Shanghai Health Study. ing and millions of DNA mutations every day in Environmental and risk factors, including each of us. Virtually all ofthese mutations occur in the con- benzene exposures, were evaluated for their potential risk text of the biochemistry of everyday life (Ames et al., 2000). contribution for workers developing AML or AML subtypes Most persons are exposed to benzene and other as classified by VVHO. The authors provided the results of on a daily basis as a result ofsmoking, environmental tobacco univariate and multivariate analyses based un worker expo- smoke, gasoline, en1issions from vehicles, and industrial sures to specific chemicals and provided risk calculations for emissions. Beyond that, a certain proportion ofthe reactive various occupational and industrial worker categories. For metabolites associated with exposure to industrial chemicals benzene, the authors slated: arc produced naturally in the body or result from normal dietary sources (e.g., phenol, hydmquinone, etc.) (Mt>dinsky et al., 19!'15; MacDonald et al., 1994; Ames et al., For example, a glas& of "citrus punch" may contain up to ::l-5 ~1g of benzene (US FDA, 2007), and many components of our diet are known to contribute to our burdens for ben- zene and other chemicals of concern {EFSA, 2005; US FDA, 2007; Binner et aL, 2007). The total daily intake of inhaled benzene fur the average nonsnmking person ln \iVestern society is about 200 1-1g per day (Wallace, l996a, l995b). We believe, based upon an appreciation of background levels of environmental and dietary exposures, that the1e exists a daily and lifetime dose of benzene that is sufficiently small so as to not increase the risk of adverse effects for virtually everyone in our what would he termed a "safe" dose. This dose represent& a level of exposure that "Our study confirn1ed that benzene exposure was sig- nificantly associated with an increased risk ofAML-total (OR= 1.43, 95% CI= 1.05-1.93) ... analysis indicated that the group affected by benzene exposure most was AML [recurrent abnormalities] (RCA) (OR= 1.61, 95% CI-= L00-2.6l).1he largest subgroup within the cat- egory ofAMI.-RCA was promyelocytic leukemia] (APL) (n = 124), A borderline significant riskwas found between benzene exposure and APL {OR= 1.95, 95% CI = l.00-2.61) ... benzene exposure was significantly associated with the specific subtype AML with t(8;21) (q22;q22) with an OR of4.26 (95%CI= l.Ol-17.96), The risk ofAML with t(8;2l)(q22;q22) appeared to be most strongly associated with recent first exposure in or after 2000 (OR=ll.97, CI= 1.44-99.64)" (Wong et al., would not represent a level ofconcern to warrant 2010a). .s or other protective action (Travis and Arms, 1!J87; Wilson et al., 1987; Capleton and Levy, 2005). The question ofwhether there exists a threshold for benzene A very interesting perspective on the evolving concept of ex-posure related to AMl., identifying the precise dose for disease thresholds was recently shared at the recent Benzene such a threshold, and identifying particularly susceptible zong conference in Munich. DL Zarbl cited the example of subgroups in the population (\vho surely have a different the changing definition of "no residue" requirements for vari- threshold dose), are topics that 'Will likely continue to spur ous pesticides and other chemicals hy regulatory authorities active debate. cll1.e use of novel platform technologies and the from in analytical to from diverse scientific Vlrill measure sm~aller and smaller chemical quantities over the no doubt provide the substance for even livelier discussions years (Zarbl, 20Hl). He felt that society could well be on the in the years to come. verge of a similar shift in the area of environmental toxicol- Benzene exposure and nun-Hodgkin's lymphoma (NHL) ogy, this ti1ne fueled by advances in "'"'-'-L<>.~;-=nn.u The causes of non-Hodgkin's lyrnphoma in d1e "'"'""''""' ics. Instead of calculating or estimaling allowable levels of population remain mostly unclear. Over the last fe\N dec- chemicals by relying upon "no observed adverse effect levels" ades, NHL has shown an average 3-4% increase in incidence (NOAF.Ls) in animal and/or human studies, the concept of each year, and in 2008, over 66,000 cases wc1~e in a NOTEL was suggested, or a "no observed transcriptional the United States (American Cancer Society, 2008c). As with effect level:' This approach might eliminate the urge to rely AML, more than nine out of every ten cases occur in adults. upon various low-dose cancer models for risk The reasons for the increase in NHL are not well al very, very low doses (Zarbl, 2010). understood, wilh only melanoma, proslale cancet~ and lung The issue of childhood leukemias and their possible cancer in women showing higher rates of increased cancer relationship to environinental benzene exposure was also incidence in recent years (American Cancer 2008c). CGU BEN0000275 Benzene and l1uman. health 19 T.ikely factors contributing to the increase indude the aging of "The terms 'reticulum cell sarcoma' and 'lymphosar- the American population, an increasing prevalence of human coma' in addition have been applied in an exLraordinar- immunodeficiency llirus (HlV) (Smitl:1 et al., 2004) [NHL is ilyvariable fashion and achieved a meaningless status, 100 times more prevalent in HIV-infected populations!, cer- by preventing effective comparison of results from dif- tain autoimmune disorders (Smedby et al., 2008) and possibly ferent centers" {Lukes and Collins, 1974). some occupational exposures (Boffetta et al., 2007; Pritschi et al., 2005; Dreiher and Kordysh, 2006). DespiLe numerous studies inYolving a ''vide variety of proposed agents, how- eve!~ "at present, no conclusive evidence of causal relations between occupations and increased NHL risk exists" (Boffetta et al., 2007]. Cenainly, rhe net impact nfrhe risk factors noted above are not believed to adequately doubling of disease incidence rates over the pasr30years (Table 5) (Greer et aL, 2004; Evans and Hancock, 2003). Disease classification changes, as we shall later have led to additional cases ofNHL previously tallied in other categories (multiple mye- loma, chronic lymphocytic leukemia, for example). A recent review by Bosetti et aL, however, suggests that the incidence and mortality of NHL have leveled off in most developed countries worldwide (Rosetti ct aL, 2008). \\llwn attempting to review the Ppidemiology literatun~ H is that future epidemiology slUdies will more carefully incorporate this information into their analyses. The sulking advances made in immunology and histopa- thulogy have allowed much more and prognostically useful categorizing of patients with lymphoma and leukemia. lherapeuLic options have created heHer patient outcomes, with the 1csult being that the previously most devastat- ing and aggressive tmnors now often demonstrate tl:1e best chances for having a cmnplete rernission. For exarnple, in children under the age of five, the survival for ALL in the 1996-2004 time frame was 91.2%, a disease that 40 years ago was an almost autornatic death sentence (Leukeinia and Lymphoma Society, 2009c). Because of the often inaccurate perceptions regarding hematopoietic disease combined with "meaningless" nomenclatures tl:1at have been used over the regarding the relarionship betvveen exposure to benzene and an increased risk of NHL, one must be aware that it is nearly Table 5. Factors thal have been reported to be associated "\Alith an increased impossible to compare medical oflyrnphoma from lhe 1950s through 1990s \Vilh Lhose from Lhe last decade risk of developing non-l-lodgkin-s lymphoma in susceptible individuals. e Infectious agents 2000 to 2010). 1his result~:. mainly from the classification strat- o EpsteinIlarr virus egies for Iytuphoma having undergone a tremendous atnount c HTI:I!types I and 2 of reorganization in tl:1e last 30-40 years. The initial wave of ' Hel!cobacter pylori change was started by Rappaport i.n 1966 who advocated a Hepatitis C virus sorting oflymphomas that relied upon pathologists describing "' Hurnan herpe.sviru~ 8 (Kaposi's sarcorna) The mmphology of entire lymph nodes involved with malig- ~ Hun1an hcTpes\rinJs 5 nancy (Rappaport, 1966). Simply put, malignancies were clas- l\1ale gender sified into the broad categories of "nodular" and "diffuse;' with Advanced age further characterization based upon the cellular characteris- Family histmy oJ NHL tics of Lhe tumor. Revision of lhe Rappaport classification was History of cancer, cancer treatinents, use of immunosuppressive agents subsequently recommended by Lukes and Collins (Lukes and 1\rf-eUicatiun~ .s Collins, 1974, 1975). Borrmving techniques from the reladvely c Phenytoin ,- Methotrexate new field ofirmnunology, these researchers related Inalignant lymphomas to aberrations in the development of B and T lymphocytes. They furtl:1er suggested that superior therapies would he more likely to come ahout\'llith a nomenclature that Occupational exposures to: c Herbicides c Pesticides \Vooddust appreciated the origins and cellular characteristics oflympho- Epoxy l"es\n,; and glues mas, rather than relying so heavily on the gross appearance of c SolYents the tumors (Tukes and Collins, 1 11\Torkers in certain indusrries; studies have evaluated: Despice the scientific battles over whose nomendalUre c Farming was superior, both sides agreed that the pre'.ious methods of c Forestry reporting lymphomas were inferior, and generally confusing c Painting to anyone atteinpting to rnake sense ofthese con1plex disease c. Carpentry processes. c Tannery/leather working/shoe manufacturing "The histologic classification which employs the terms 'reticulum cell sarcoma; 'lymphosarcoma; and 'giant follicular (nodular) lymphoma' has long been knmvn to correlate poorly with discrepancies in the clinical course of patients \'llith 'reticulum cell sarcoma' have been particularly puzzling to clinicians for years" (Jones et 1973). Othel' c. 1J.-;p of hair dyes c~ Excessive exposure t.o ultraviolet. rays G Nutritional factors ' History of blood transfusions HTlV= human T-lymphotropievims; NHT.= nonH(jdgkin's lymphoma. Sour-ces: American Cancer Society, 2009; NCI, 2007; ATSDR, 2007; Vosc et al., 2002; Alexander er al., 2007. CGU_BEN0000276 20 D, Galbraith et aL years to describe lymphomas, attempting to correlate data animal and human studies, with varying conclusions. Yager bet\veen various time periods can be impractical and oflen el aL, for example, performed in virro studies with hydro- impossible. quinone, benzoquinone, phenol, and catechol, finding a Currently, the medical community is in yet another particularly noticeable effect with hydmquinone in caw;- "-"''-'".,"' of lymphoma clas- ing micronuclei formation, an indication of chron1osmnal sification. The Revised classification uo.H1a..:c (Yager et aL, 1990). Farris and exposed system for lymphoid neoplasms {REAL), first presented mice to varying concentrations of benzene {l, 5, 10, 100, in 1993 by an international lymphmna study group, and 200ppm) by inhalation for up to 8 weeks (6 hours of organized lymphomas on the basis of morphology and exposure/day, 5 days/week) and measured the effects on further subdivided on the basis of genetic and immuno- various lyrnphocyte populations. A dramatic decrease in logic parameters (Skarin et al., l ~~7) . This was splenic, thjllnic, and femoral lymphm:ytes was seen at the reminiscent ofthe Lukes and Collins modifications to the 100 and 200 ppm levels of exposure, with no significant original Rappaport system in 1966 (Lukes and Collins, change cmnpared to controls at lower airborne concentra- 1975) . The REAL da;ssification has since been updated tion levels (Farri:s et aL, 1997). the World Health in 2001, and the clas- In a often workers to a of sification of leukemias and lymphomas will forever be a benzene during the loading ofa ship, substantially increased "work in progress" as our knowledge of the basic science levels of urinary phenol were detected in the exposed work- impro'lies and new treatment advances lead to further ers. Three months after the accident, peripheral blood was understanding of the pathophysiology of the lymphomas obtained from the group and compared to H con- (Jaffe ct aL, 2001 ), More recently, the 2008 revision of the trol workers who had not experienced the spill. Cells were WHO classification of lymphoid neoplasms has elevated examined for all types of chromosomal and sister- "precursor lymphoid neoplasms" to a separate category of chrmnatid exchange. Similar amounts ofdamage were found disease and recognizes the overlap characteristics ofthese in both the exposed and control groups, suggesting that there tumors hy each individual entry as a "lymphob- was no evidence of any in the group of lastic wiLh exposed workers (Clare el aL, 1984). TI~ere have been many characlerislics (Swerdlow et aL, 2008). By segregating olher similar in vivo and in vilro studies thal have been per- tumors by degree of differentiation, cellular identity (B formed, but these analyses typically have suffered from one T cell, NK etc.), location of disease, and clini- or nwre deficiencies: cal characteristics, there are now 74 separate subtypes of non-Hodgkin's lymphoid neoplasms, a number that is For occupational studies, a lack of clarity as to historical certain to grow even further in the future as disease pr-oc- exposure levels that workers were to esses arc even rnore discretely characterized (Swerdlow Concomitant exposures to other potentially harmful et aL, 2008), chemicals that are often neither identified nor consid- At the international conference on henzene recently ered as possible confounding agents for the endpoints held in Munich, Vardiman de~>cribed the vVHO classifica- being studied tion system for tumors of the hematopoietic and lymphoid Improper control groups .s tissues (vvith an emphasis on the myeloid neoplasms) Inability to correlate simple hematologic parameters (Vardiinan, 2010). He noted: "in general, the classification with definable risk for developing cancer endpoint stratifies neoplasms according to their (myeloid, For animal studies, difficulty in comparing high-dose lymphoid, histiocystic/dendritic) and distinguishes neo- exposures to likelihood of disease at lower concentra- plasms of precursor cells from those comprised of function- tion;; in humans ally mature cells:' A review of his paper rapidly convinces any skeptic that it is not useful to conduct any future study A number oflahoratories continue to and they ofthe relationship between exposure to benzene and vari- arc to mitigate the above-mentioned issues. ous tumor types without a very careful description of the Claims of an occupational link between benzene expo- classification. sure and lymphoma were ex-pressed in a 1979 publication Although most international regulatory authorities have that used a combination of death certificate and census agreed that chronic exposures to airborne con- data, along with assl..unplions regarding benzene exposure centrations of benzene increase the risk of developing AML, in certain occupational subgroups (Viarma and Polan, 1979). there is no such consensus for NHL, despite animal studies \IVorkcrs with likely C:ll.lJOsurcs to benzene were found to be that have shown the occurrence of thymic and nonthymic at increased Tisk for certain types of lymphoma, though no lymphomas in mice at telatively high chronic airborne con- centrations of benzene {:-HJOpprn o hours/day, 5 actual benzene exposure data were reported. Some of the chemical exposures by workers but not consid- over the lifetime ofthe animals) et al., 1984, 1985; ered by Vianna and Polan in their were Farris el al., 1993). Studies have also evaluated the effecls of various levels of exposure to benzene and/or benzene Arsenic metabolites upon peripheral blood lymphoeytes in both Asbestos CGU BEN0000277 Benzene and l1uman. health 21 Fertili7ers of NHT. were verified by tissue analysis ar all (Tra"\is et al., Lead 1994). Mercury As we have noted above, the classification methods for Pesticides lymphoma here in the United States over the past 30-40 years have changed and it is unclear when dif- Epoxy resins ferent pathologists from different regions ofthe world would Aniline have adopted the varying sets of disease criteria. 'lhe failure Hydrogen cyanide to do a n1orc comprehensive verification of lymphonta and Phosgene leukemia cases makes it difficult to place much confidence Nickel in the conclusions reached by these researchers, which have Uranium varied from more recent occupational studies conducted Phthalates in China (Wang et al., 2006; Wong et aL 2010a; vVong et al. (Stellman and Damn, 1973). 2010b). A review by vVong {1999) evaluated the methodologies 'The findings of Vianna and Polan were employed in the NCI/CAPM studies. assessment Enterline in a Letter to the Editor uf the publishing journal, was felt to be particularly problematic. who found inconsistencies in their calculations and doubted that there was actually any increased mortalil<J for the dis- '~ .. the control workers were exposed to neither benzene eases noted (Enterline, 1979). nor any other occupational carcinogens, whereas the Confounding exposures arc also relevant for the NC! exposed workers were exposed to not only benzene but studies describing Chinese workers with hematopoietic also possibly to other occupational carcinogens. -rhus, disease et al., 1996; et aL, 1997). Other chemical by design, a comparison between the exposures were not controlled for, and in fact, the two occu- workers and control workers would not be limited to pational categories with the greatest relative risk reported the effects of benzene, hut would also reflect the effects fur nmt-HmJgkin's were "Chemical workers:' of olher occupalional carcinogens. Titere fore, lhe where one would presumably be exposed to a wide variety was nol designed properly to address the effects of ben- ofdifferent potentially toxic substances; "Coatings;' another zene only. Yet, the authors were ready to attribute any occupation where Inany types of agents are likely to have increased risk of disea.<.es found in the study to benzene been employed; and "Other/n1ixed occupations" where exposure" ('Wong, 19~19). the types of exposures were not described. These three ofworkers 14 ofthe 16 cases of '\!\Tong that exposme estin1a!es were "consistently NHL described in the study, 'vith the other 1:1.0 cases corn- lower than the actual exposure data." Although NCI had ing from rubber workers and shoe workers, respectively (Yin admitted earlier on in their involvement the proble1us asso- et al., l99G). ciated vllith their exposure calculations, '~ .. the inadequacy of Yin et aL desnibed their methods for case validation as the estimates was never menlioned or discussed in any sub- follows: sequent analyses or in the et al., HJ!-J7l report. Instead, .s the exposure of the workers was hailed as 'well character- "All histopathologic and bone rnarrow aspirate slides ized"' (Wong, 1999). and peripheral blood smears were reviewed system- Wong concluded his analysis with the following note: atically, by expert hematopathologists affiliated with the Mayo f:linic, NC:I and Peking Union Hnspital using "Unfortunately, there were many inherent problems in structured abstract forms to objectively characterize the data, as well as serious limitations in the exposure hematopoiesis" (Yin et al., 19!::Hi). esti1uates. Because of these unresolved problems and limitations, many of the results in the CAPM-NCT study lhe proponion of cases, however, where researchers had are unreliable. Therefore, the conclusions of the study, bone marrow slides and peripheral blood smears available particularly those involving exposure estimates, are not for evaluation was minimal. In describing their protocol for justified" (Wong, 1999). case confinuation, Yin et al. referred to Travis et al. (1~194), who described the same cohort of almost 75,000 Chinese Lamm et al. {2005) reviewed benzene exposure and NHL, workers. ]hat study used a review of pathology specifically excluding studies by the rcpm-ts and medical records to "confinn" 16 cases of "malig- industry to avoid possible researcher bias and included nant lymphoma or related malignancy" (the definition of 18 studies with 21 different study groups in their analysis. ''related was not provided) and reviewed his- Overall, noted that these studies had an aggregate of topathologic material {i.e., slides, blood smears, or original 404 observed cases of NHL with an biopsy malerial) for only 4 cases of ''malignant lymphoma:' number of 390 in the populations lhal were looked al, Hodgkin's and non-Hodgkin's lymphoma were not sepa- yielding an odds raiio of 1.04 {95% CI: 0.94-l.H). Wnen the rately described, making it difficult to know if any cases authors removed studies that failed to adequately address CGU BEN0000278 22 D. Galbraith et aL the question of multiple exposures, they were left with 359 observed cases and 373.2 expected, with a corresponding Industry Agricultural products RR 2~3 odds ratio of 0.96 (0.86-1.06). The authors observed that the Apparel 2.4 latest report, the Rinsky 2002 study, on the Ohio Pliofilm Barbers 2.7 worker population cohort singled out for having detailed Forestry 6.2 exposure data, few alternative exposures, and strong epi- Furniture sal-es 4.9 demiologic standards) provided further support for their Camps and trailer parks 5.5 conclusions. In all, there were five cases of NHl. reported Lahar unions :L3 and slightly more than five expected, generating an odds Masonry 2.6 ratio almost identical to that of their meta-analysis, which was calculated tn he less than unity. Tn sho<t, no increased risk was observed. Smith er al. reviewed case-control and cohort studies that examined populations with probable exposures to benzene Metalworking Physician uffices Labor unions Retail bakeries Stone 3 3.4 2.3 4A 2.6 and assessed their relationship to the endpoints of NHL morbidity or mon:ality (Smith et 2007). The authors con- cluded that : .. overall, the evidence supports an association between occupational benzene exposure and NHI.:' (Smith et aL, 2007), relative risks and odds ratios reported in In fact, it was just as risky to be an insurance agent or a theatrical producer as it was to be a member of the petroleurn profession, according to Blair et aL 1lu:y coucluded their studywith the following: the various studies. After examining in greater detail the quality of the undedying data and odginal conclusions made by che individual study authors sumtnarized Srnith et aL, we believe the conclusions of these studies were selectively, and "In surnrnary, this evaluation docs not indicate that industrial exposures are a major contributor to the etiology ofNHI:' (Blair et al.,] 'l'l3). often inaccurately, conveyed. We provide several exam.ples Fabbro-Peray was a questionnaire-based study that made as follows. assumptions as to benzene exposure occurring in certain Bernard et al. was a prospective case-control sludy of all occupations, without any actual measurements of ben- cases of "lymphoma and lymphocytic leukemia within a zene exposure. The data describe a very small nmnbcr of population of L5 million people in England between 1979 benzene-exposed cases and controls, with 22 self-reported and 1931" (Bernard et aL, 1984). In their study, the authors benzene--exposed cases and 2:-:l controls. indicated that the "Number of cases are small and confi- Interestingly, although the "benzene-exposed" workers dence intervals wide" (Rernatd e! aL, 1984), wilh the lower were believed to come frmn the chemical, rubber and upper 95% confidence intcnal limits of petroleum tion, printing, and painting industries, expo- workers diagnosed with NHL varying by almost two orders sures to rubber solvents, petrolemn and ofmagnitude, while also failing to he significant. waste uil .were nol irulividually founu lube associated with Not noted Smith et al. was the addilional finding that an increased risk for developing NHL Only coal tar expo- use of "petroleum products" by men and women was also sure was identified by the authors as having a .s associated with a very low risk ratio of 0.59 (0.25-1.39). Thus increased risk, which >vas not statistically significanc. the corn!Jinatlon of very few cases in this study, along with Employment as a "radio operator" was associated with a the failure to demonstrate any chemical-specific risk related more than 50% greater risk for developing NHl. compared to benzene by the authors, makes the value of the Bernard 'Nith professions deem.ed to be at risk for benzene exposure paper questionable in terms of supporting the notion that (3.3 versus 2.0). These authors concluded: benzene exposure is linked to an increased risk for NHL The Blair study, cited hy Smith et aL, was a populationbased case-control study designed to evaluate cancer risks from agriculmral exposures in Iowa and Minnesota. 1he study suffered from a very s1nall number of cases involving benzene exposure. The highest relative risks that Smith et aL reported fnJnl those provided by the authors were follicular lymphoma with a RR of L9, based upon five : .. we cannot forget that the OR values have been estimated from a very small proportion of subjects, leading to large confidence intervals, and thus, imprecise estimations. 1l1at is true for history of hematological malignancy... and especially benzene exposure'' (Fabbro-Peray, 2001). cases, and "diffuse" lymphoma with a RR of L8 based on Franceschi et al. was a study of 208 cases of ;\l"HL four cases. Smith ct al. mentioned that workers associated and 4111 controls in the nurthea;.tern part ofitaly, with "petroleum refining" were observed to have a higher between 1985 and 1988. Smith et aL mischaracterized the risk of developing NHL with a RR of Lf-i, hut to findings reported by these authors, implying that the mar- observe a host of other industries with far greater calcu- ginally increased RR of 1.14 was associated vvith benzene lated risks, nol known lo have regular exposure lo solvenls. exposure, when it actually referred to exposures to "benzene These industry worker subgroups arc summarized in the and solvents" rnore generally. The marginally elevated risk following table: reported for petroleun1 workers did not show an expected CGU BEN0000279 Benzene and l1uman. health 23 relationship with length of employment, reflected in the the control population used by Hard ell et al. were uniformly quote below: either exposed lo TCE or styrene, agents that would have the "There was a hint that employinent in chemical and petrochemical industries may increase the probability of the onset of NHL, in agreement with a few previous :reports. lhe elevation of risk. however, did not tend w be higher in those individuals who had been employed in such industries for longer periods'' (Franceschi et al., potential to confound the reported results. Smith et al. also reported on a later Hardell et al. .study that described 105 with NHL who were admitted to a Swedish hospital betvveen 1974 and 1978 (Hardell et al., 1994) lhis was a analysis with self-reponed exposures and no n1casured chcinical exposures, as \vith the 1981 smdy. Cases linked to benzene exposure amounted to 1989). only three in nun1ber, with only a control patient. lhey also emphasized Lhal: The lack ofstatistical power and lack ofprecision is reflected in the upper and lower hound limits for the confidence inter- "M1n-eove1; the absence of stmng findings in the present investigation should he stressed. [n fact, despite the large number of potential risk factors investigated, only one plausible and significant association (with positive hi;;tury of chronic infectious diseases) emerged, thus confirming the difficulties in studying epidemiologically the aetiology of lymphoreticular neoplasia" (Franceschi et al., 1989). val ofrisk in benzene-exposed patients, which vary by almost three orders of magnitude. 1he other values cited by Smith ct aL refer to "organic solvent" exposures, a wide variety ofchemicals, which may or may nut contain small amounts of benzene. 1he findings reported by Hardell et aL for heavy exposures to organic solvents, incidentally, were dwarfed by their calculations for who sustained any exposure to phcnoxyacctic acids (0Rof5.2 as well as all reported exposures to chlomphenols (OR of4.8 [2.7-8.8]). Far from irnplicating benzene or any other chemical, they found that: The Hardell et al. (19lJ4) data, then, regarding benzene and NHL, suffer from woefully small numbers, both for numbers of cases and controls, and when addressing risks exposures to chemical or physical agents to solvenls, suffer from a lack of suspected of being related to lymphoma risk were even to the possible actions ofbenzene on the palient population rarer and none ofthem apparently resulted in a signifi- in question. cant elevation of risk" {Franceschi et al., 1989). Kato et al. surmnarlzed the incident cases of NHL frmn Frilschi el al. reported the incidence ofNHL frorr1 a cancer registry in Ne'"' South \Nales from January2000 to August 2001. As Smith ct al. correctly report, there was no increased risk tor developing NHL discovered by the resean.:hers. v\1Iat Smitl1 et al. did not point out was that in the subgroup of "'"1n"""'" to "substantial" of benzene solvents, the odds ratio was calculated to he o.:-n, lhe lowest finding of any 1995 to 19!'!8 in upstate New York, identified through the New York State Cancer Registry. This questionnaire-based study looked at a very small number of cases and controls (seven and five, respectively) and found a small increased risk for cases having occupational exposure to benzene. However, the authors were very dear with their interpretation of the data: .s ofthe ll individual chemical exposure subgroups described by Frits chi et al. in their (Fritschl et al., 2005). "Overall, history of exposure to organic solvents was not associated with the risk of NHL. A statistically sig- A study of patients vvith malignant lymphomas by Hardell nificant increase in risk associated with occupational and colleagues reponed on individuals diagnosed becween expm.ure was observed only for the subjects whose first 1974 and HJ78, vvith exposures that occurred well before many exposure occuned before 1970" (Kato et aL, 2005). major occupational and regulatory changes applying to the use of benzene-containing solvents. To determine cases of It should also be noted that Smith et al!s citation of a statisti- NHI., the authors stated that they relied upon the original cally odds ratio of 1.40 for KHL does not relate Lukes-Collins system, which was modifictl to "retrospectively to solvents associated with benzene, but fur paint thinners reclassify" patients "\Vith lymphoma at the Swedish hospital and turpentine, which are knovvn tn predominantly contain involved in the study from 1959 to 1975 {Hardell et aL, 1981). toluene and/or xylene. In the cases described, there was only one patient with 1\Iao et al. reported on data obtained from the Canadian heavy exposures to benzene. 1he relative risk of 4.5 (l.9- National Cancer Surveillance System (NECSS) from 1994 lL4J quoted by Smith et al. represented a total often cases, to 1997. Based on a small number of workers exposed to matched with eight controls. Of the ten cases, seven were benzene, the authors found a small statistically insignifi- exposed to TCE, one to styrene, one to tetrachloroethylene, cant increased risk for developing NHL in men, and a small and one to benzene. There '~:ere no n1ixed t>xposures with statistically insignificant decreased risk in women {OR 1.2 benzene, and there were zero conuol examined with [0.8-1.9] and OR 0.6 [0.2-1.8], respectively) (Mao et aL, 2000). benzene exposure. These risks were not even commented on by Mao et al. in Thus, 90% of the cases leading to the RR of 4.5 reported their review, focusing their comments instead on benzidine, by Smith et aL had nothing to do with benzene. Moreover, mineral oil, and pesticides. CGU BEN0000280 24 D. Galbraith et aL Miligi et aL conducted a population-based study looking at Roston area. They found that the relative risk for NHT. was newly diagnosed cases ofKHL in ll areas ofitaly from 1991 highest in those employed in the agriculture, forestry, and to 1993. The authors grouped participants into four levels of fishing industries. Decreased risk was found for those work- perceived exposure according to their probable net exposure ing in the chemical, pharmaceutical, and painting indus- to ofchemicals, as determined from responses tries, and these decreased risks were found to be statistically to a questionnaire. After to find : ..increased NHL risk from "very low/lm-v" intensity exposure to any class or to benzene exposure. they noted: solvent" (Miligi et al., 2006), the authors bundled together the n.vo lmver and nvo higher exposure groups, reporting a "pro- tective" odds ratio for very low and low exposures to benzene ofO.Il, and for "medium and high" expnsures, they ohtained. an odds ratio of 1.6, which included one on the lower bound of the confidence interval. It is also interesting to note that the ORs reported for both xylene and toluene exposures were greater than those reported for benzene at both combined exposure levels reported. More importantly, when L'le authors segregated out exposures, and only looked at subjects with "medium to high" exposures to benzene alone (eliminating mixed e:;.1Josures), their OR dropped to 1.2 and became sta- tistically insignificant (0. 'lhc authors offered caution about how to interpret their findings: "In our data, there was a high degree ofcorrelation among "We did not observe an increased riskfonhose exposed to benzene. The association hetween lyTilphoma and benzene had been reported by Vianna and Polan, where Lhey compared the proportional mortality rate for Lymphoma among those who had worked in an industry that used benzene with the rate for those who worked in all other industries. A review of the article showed that the strength of their association was strongly supported by the increased risk of disease presumably due to benzene for woodworkers and farmers. Subsequently, it has been suggested tl1at the increased risk among woodworkers and fanners n1ay be due to o tl1er agents. A later study also found no increased risk \Nith exposure to benzene" (Scherr et aL, l992). exposures to benzene, xylene, and toluene. For this 1ea;;on, caution must he exercised when interpreting the evidence for any one of these 3 solvents" (Miligi el al., 2006). Ott et al. studied a cohort of 29,139 men employed Union Carbide in manufacturing facilities and a research center in the 1940-1978 time period. Smith et al. conectly point out that the OR for workers likely exposed to benzene Sclmaner eta!. looked at all cases offatallymphohematopoi- ctic cancer in an occupational cohort. l11ey noted that half of the cases of NHL reported occurred in clerks or techni- cians, individuals with little to no opportunity for exposure to benzene. The only risk value cited by Smith et aL is specifi- cally addressed by Schnatter et al., describing this RR of 3.85 as based on two cases with luw level compared to workers vvithout benzene exposure was 1.0 for NHL However, the rest of the citations from the "peak" exposures to benzene of 0.5 to LO ppm. There was 1989 Ott et al. study are improperly characterized and not only one case ofNHL that experienced a tnaximum benzene exposure of over 1 ppm, associated wirh a RR of 0.54. adequately explained by Smith et aL in their review. First, an OR of LG for workers with more than five years Schnatter et al. concluded: ': ..this did not show a relation between lymphohematopoictic cancer and long of exposure to benzene and developing NHL wa& citecL This odds ratio, described by Ott et al. in Table IV of their lB8Y term., low level exposures to benzene" (Schnatter et al.. .s article, pertains to multiple myeloma, not NHL. Next, an OR 1996). \Nilcosky et al. reported on a cohort study ofrubber work- of3.2 for "benzene-exposed forernen and rnaintenance/construction workers" was cited for developing 1\'HL compared to ers in Akron, Ohio, "\Vho were examined from 1964 to EJ74, a period when higher levels of benzene exposure were experi- those workers without benzene exposure. However, Ott et al:s calculations have nothing to do "With historical exposure to benzene. Instead, the authors are describing odds ratios for enced, in a profession known to use solvents with benzene present in measurable concentrations. These authors exam- workers ever having been en1ployed in various work areas, ined a number of cancer endpoints, finding that: with no mention of the chemicals that might have been exposed to. i\Jso, the category ofworkers described by Ott et al. with an OR of 3.2 is only the subgroup of "fore- men and others;' and not the broader group ofmaintenance "Benzene, a suspected carcinogen, was not significantly associated with any of the cancers [including NHL]" (Wilcosky et al., 1984). and construction workers, as seen1s to be irnplied by Sinith Instead ofreporting confidence intervals, the authors instead et al. 1here are nine separate occupational subgroups under indicated where reported values had a 1' value of <.05. This The of maintenance and construction workers, each was not the case for any of lhe relative risk numbers cited by with their own calculated OR. This same problem occurs "llvith Smith et al. from the Wilcosky Smitl1's citation of the OR for "instrument men:' 1his OR has The underlying studies and data supporting Smith et al:s nothing to do \Nith benzene exposure, but relates to a conclusion of an association between benzene exposure and job classification. chemical risks are not the subject r\HL appear w be quite weak, and in many cases, improp- of discussion {OlL el al., 1989). erly cited by Smith and his colleagues. We believe the case Scherr ct a!. reviewed all cases of NHL diagnosed from for benzene as a causative agent for NHL remains far frmn 1980 to 1982 at any one of nine participating hospitals in the dear. CGU BEN0000281 Benzene and l1uman. health 25 However, v\Tong and colleagues (2010b) reported on a case-control study designed to examine risk factors for the development of NHL in Shanghai, China, as part of the Shanghai Health Study. The authors indicated that their findings confirmed several risk factors that had been previously reported in other studies, discovered new potential risk factors for NHL, and for the first time provided detailed data on individual NHL subtypes using the updated iNHO classification system. with regard to benzene exposures and the developrnent ofNHL, the authors stated: "In our study, exposure to benzene was not associ- ated with an increased risk of NHL [neoplasms]-total (OR=l.06, 95% CI=0.74-L5l), based on 50 exposed cases and 95 exposed controls... which was consist- ent with the literature. no increased risk was found fm most subtypes of B-cell neoplasms, or T/NK-cell neoplasms... However, seven cases of [fol- licular lymphoma] and two corresponding controls were exposed to benzene, resulting in a significant OR of 7.00. The 95% CI {1.45-33.70) was rather wide, pri- marily because the OR was based on only h.vo exposed controls'' (\Vong et al., 20l0b). The American Cancer Society recognizes the follmNing risk factors as relevant fur CLL: Family history of CLL Parents, siblings, or children "\llith CLT, have a tvvo to four tirnes increased tisk of developing the disease Gender CLL is slightly nwre comrn.on in 1nen, for unknown reasons Ethnicity CLL is more common in and North America than in Asia. The basis for this appears to be genetic rather than environmental. as Asian inun.igrants have not been discovered to increase their risk ofCLL to that oftheir host country. Certain chernical exposures Some studies have suggested that exposures to Agent Orange may increase the risk of developing C:LL Other studies have dain1ed an association betvveen farm- ing and exposures, but results have been inconsistent. v\i11en multivariate was performed, benzene was not found to be associated with NHL overall, although benzene etnerged as the only significant variable associated with the development offollicular lymphoma. They advised, however, that: TI1e ACS concluded their discussion by stating: "clhere are no other proven risk factors for CLL The risk of getting CLL docs not seem to be affected by smoking, diet, exposure to radiation, or infections" (ACS, 2008h). "The finding of an association between benzene and [follicular lymphoma]... should be interpreted with caution as the OR was hased on only two exposed controls" (Wong et al., 2010b). We should note that CLL and small lyrnphocytic lymphoma (SLL) arc being increasingly considered to be separate instances of the same disease process; the distinction being whether the disease is believed to he based in the bone marrow (CLL) or in lymph nodes (SLLJ, lissues outside Benzene exposure and chronic lymphocytic leukemia the hone marrow. For in 2001, the World Health .s (CLL) Organization classified CLL and SLL into a single disease Chronic lymphocytic leukemia has historically repre- category in recognition of their similar cellular appearance, sented the mos1: common subtype of leukemia, with about as well as their immunologic characteristics (both diseases 15,000 estimated new cases diagnosed in 2008 (ACS, eA.'press similar cell surface antigens) (Taffe et al., 2001). This 2008b). It is almost exclusively a disease of adults, with 1ecognition i:s important when reviewing historical publica- the average age at the time of diagnosis being 70. Many tions looking aI the epidemiology of these diseases; as well with the disease, however, are as more recent studies that have tended to group the diseases becoming diagnosed only as a result of routine blood (Dores ct al., 2007). The most recent\'\'liO revision tesls. Thus, it is Lh.at the lrue population for classification of le ukernias and lymphomas lists CLL and incidence may be much higher than that reported SLL as the same entity (Swerdlow, 2008). The elimination of cancer statistics. CLL as a distinct disease category and its bundling with SLL Unlike ANLL, CLI. has not been strongly associated with naturally will have a great irnpact on future epiderniologic any environmental or occupational exposures, including studies, and will complicate the interpretation of historical benzene exposure (Greer et al., 2004): studies that have treated SLL and CLL as separate malignan- cics. \/\Thether this will improve our ability to more dcfini- "Exposure to high-dose radiation or benzene is not a risk factor for CLl:' am.l Lymphoma Society, 2009a). th:ely describe the possible relationship betvveen benzene exposures and this now unified disease process remains to he seen. Despile general agreemenl by nalional and inLcrnalional "There is no link to radiation, cancer-causing chemicals, cancer authorities rhat there arc no known environmental or viruses" (NIH, 2009b). risk factors for developing CLL (other than that some believe CGU BEN0000282 26 D, Galbraith et aL exposure to Agent Orange may be a factor), some have opined that exposures 1o benzene can cause CLL in humans, In 1997, Savitz and Andrews published a review of the scientific literature, including 18 community-based and Hi industry-based studies benzene to "lymphatic and hematopoietic cancers:' Their revie\v found reports" linking benzene to NHL and multiple myeloma, bul the authors detennined that the weight ofthe evidence sug- gested that benzene should be considered to cause all forms ofleukernia, not simply AlvL: "]he epidemiologic evidence linking benzene to leukemia in the aggregate, as well as acute and chronic lyrnphocytic and myeloid leukemia, is no less persuasive than that for AML alone" (Savitz and Andrews, E!Y7). SavHz and Andrews admitted, however, thaL exposure data for benzene in the studies they reviewed were generally lacking, and the risk for confounding chemicals in most of the work environments would make it difficult to make conclusive statements regarding benzene and disease causation. ''Failure to isolate benzene from closely associated agents that may themselves cause lyn1phatic and hematopoietic cancers could introduce confound- ing, In the rubber industry, for example, associations between benzene and lymphocytic leukemia were [Checkoway et aL, 198,1], yet oilier solvents were as or more associated with lymphocytic leukemia, making isolation of the etiologic agent diffi- cult. En-vironments associated with the industry, shoe manufacturing, and painting contain a wide array of organic solvents and other potentially carcinogenic chemicals which were not fully addressed in the published reports" (Savitz and Andrews, 1997). are more significant disadvantages" (Lomnis and Savitz, 1991]. In the 1997 analysis, Savitz and Andrews cited a number of papers that they believed were supportive of the idea that benzene caused not only AML hut allleukernias. TI1e remain- der ofthe text in lhis ponion of our review sum.marizes a few ofthc findings reported by various authors, Checkoway et aL (1984) performed an analysis ofthe pos- sible effects uf a of ;,u]vents used by rubber industry workers, Benzene, however, was not singled out as being particularly notable in terms ofits perceived risk. "The associations with lymphocytic leukemia risk observed for a number ofsolvents, n1ost notably carbon tetrachloride and carbon disulfide, were stronger than those detected for benzene" (Checkoway eL al., 1984), To be specific, out of the 24 solvents tested, 12 were reported by the authors to he rnore associated with tlu~ developnu~nt of subsequent leukemia than benzene, including: Acetone Methanol Ethyl acetate Toluene Hexane TI1e fact that these solvents, none ofwhich are known to have a significant relationship to leukemia, were observed to be more highly associated 'With leukemia than benzene would appear to reduce the value of the f:heckoway findings of increased risk for benzene. Checkoway summarized the impact of benzene on nwre modern worker populations as follows: .s An earlier study by Savitz on leukemia mortality and its "While benzene exposure still occurs in the indus- relationship to occupation reported no increased incidence try, it is present primarily as a contaminant of other ofAMLin occupations traditionally associated with regular substances, and exposures are much lower than in exposure to benzene-containing products, such as printing rnachinc operators, "cleaners and laborers;' rnetal-, plastic-, and wood''ITorking machine operators, painters, vehicle rnechanics, and motor vehicle operators (Loomis and Sa-vitz, the 1920's; thus, il is quite likely that only small proportions of the cohorts studied were ever exposed to benzene in concentrations known to induce leukemia" (C:heckoway, 1984), 1991). Instead, the occupation with the risk reported for AML was "mathematical and computer scientists:' 1he audwrs noted in lhis sludy !hal: Girard anti Revol {1970) provided data on the incidence of various leukemia subtypes in a <egion of France where the population was felt to be exposed to benzene. I1: s.eems "This study shares other limitations of occupational unlikely, however, that a community-based study almost investigations based on death certificates. The&e weaknesses,,. include the inadequacy ofusual occupation as an indicator of lifetime work and eJ<::posure histories and 40 years ago would have heen ahle to quantify the degree of benzene exposure that individual residents experienced, particularly since airborne concentrations are normally at the lhe absence ofinformation on lhe timing or duration of the listed occupation:' ppb level; too low to be m.easured reliably in the early 1970s. It is also unclear fmm the title of the article ("La . d'une Exposition Benzenique au Cours des Hemopathies "The lack of information on exposure to specific etiologic agents and the potentially confounding exposures Graves;' translated as "Incidence of to Benzene in Sevete Hemopathies") rhat leukemia was even CGU_BEN0000283 Benzene and l1uman. health 27 the focus ofthe study (Girard et al., 1970). ltwould be helpful to review a translated version of this smdy to understand how it was constructed, how the data were obtained and leukemia diagnoses confirmed, and how pertinent the findings are with respect to relating benzene exposure to leukemia (the only two categories listed by the authors are "CLL:' and "acute (Girard et aL, 1970). Linet et aL (1987) compawd two different methods for coding job titles in collecting and reporting data from the National Occupational Hazard Survey. However, Linet et aL did not daim to find an increased incidence nf CLL in the population studied; in fact, the reverse was stated. "It is interesting to note that the relative odds [for expo- sure] in general, were less than one when the analysis was confined to subject responders and, in all of the relative odds, with the exception of asbestos in the Hoar et al. analysis, and carbon tetrachloride in the based upon the KOHS data, were less than or equal to one" (Linet et al., 1987). TI1e authors then concluded that: did not present e'lJOSure data from the V\'orkers included in the study, but instead assigned the "several hundred depan- ments" and the more than 1000 different types of jobs into a reduced set of 70 occupational titles tl1at the authors designed. With the of some collaborating industrial hygienists, they created three solvent-exposure occupational groups defined as medium, and exposure ior 19 of the 70 new job categories. lhc 51 rcinaining were designated as "no solvent-exposure:' McMichael et al. did not provide any description as to how benzene might have been used >vithin the over the years, or when alternative solvents were put into service. Although the authors noted an elevated risk for "lymphatic" leukemias, they appeared to be aware that their findings were controversial, stating: "1he association oflyrn- phatic leukemia with exposure to organic chemicals appears to have no reported precedent" {McMichael et 1975). Over the past 30 years, the findings of an increased risk for developing lymphocytic leukemias in rubber workers have nol been replicated. 'llie follmving characleristics of the McMichael ct al. (1975) study make it difficult to rely upon as a basis for claiming a relationship bet\\reen an increased risk of CLL and exposure to benzene: ''No statistically significant associations vere found for CLL 1,\/ith occupations or industries of en1ploy.ment, or suspected exposures suggested by previous studies" et aL, 1!)87). to multiple chemicals were described, hut no characlecization of Lhese other exposures wa~ provided Nonoccupalional risk factors Smoking and other risk factors were not addressed l'vialone et al. was a population-based case-control study of CLL, where were used tTained interviewers to obtain data on a wide variety of possible chemical expo- sures and other risk factors. The results for benzene associa- Tion!'\ were not striking. Srnall population size and nmnber ofcases led to iinprecise results Lack of benzene exposure data for individualworkers or in the work environments Authors relied upon death certificate data without histologic confirmation of disease examination ofreported exposure to benzene, J\,1cMichael et al. included as "cases" any worker whose which has long been suspected of related to dealh certificate mentioned leukemia, regardless of its chronic ly1nphocytic leukemia, revealed a relative risk relationship to the cause of death .s estimate ofl.l (95% CI 0.6-2.0)" {Malone et aL, 1989). Although the authors described a program of compre- hensive research at four different rnajor rubber cmnpa- The authors thought the reason they found little association with benzene <.md CLL may have been "...because of a lack of nies, data on leukcrnia rnortality were presented from only a single company exposure to coal tar-based solvents '"'ithin the studypopula- tion, which may be young enough to have avoided n1a.s"ive exposure to coal tar-based solvents" {Malone et al., 1989). 'Ihe numbers ofcases thought to be associated with benzene expos1ne were not nported. Others have suggested that CLL may be linked to ben- zene exposure. In a chapter frun:t a European Environmental Agency Report (2001 ), Infante cited the Savitz and Andrews (1997) paper noted above as an epidemiologic that tJ<'-"'"'"'-' the carcinogenicity of benzene to all major forms ofleukacmia:' He also poimcd to the McMichael ct al. (1975) study of rubber worker~ as further evidence of a link becween An investigation into a suspected excess number of CLL cases in a region of Queensland, Australia, was recently by Queensland l Icalth, at the request of the govern- menl {Queensland Healil"I, 2007). Looking at the incidence of CLL during the years from l<J96 to 2004, the researchers found 22 cao,es, compared with an expected number of 14. After per- fornling a cmnprehensive review ofthe repmted cases, review- ing and summarizing the published literature vlith regard to the causation of CLL, and evaluating environmental emissions, occupational exposures and other alternative risk factors, these government researchers reported the follmving: benzene and CLI. (Infante, 2001 ). The McMichael publications examined mortality in a cohort of 6678 rubber workers over a <J-year period (1<!641972) and compared the observed deaths to those expected from the 1968 US age-specific death rates. McMichael et al. L Six cases were found to have a first-degree relative knmvn l.o have CLL; a markedly higher ratio than has been reported historically, suggesting an increased genetic component to the population studied. CGU BEN0000284 28 D. Galbraith et aL 2. No common occupational link was identified as occurring with greater frequency in cases compared to a normal background population. 3. No links were found between the chemical emissions studied and the incidence of CLL in the region. of hema1ologic malignancies (A1nerican Cancer Society, 2009; Rajkumar et al., 2005). On the question of benzene and its relationship to multiple myeloma, health and tory authorities appear to generally be in agreement that the data do not support a link 4. The inneased incidence initially reponed was c.lifficult tu interpret, because the geographic boundaries and period ofinterest were defined after the cases were discovered. cnw authors feh tha1 "Even this, the difference A11 higher risk of multiple myeloma was once thought to be associated with exposure to benzene. However, later studies have failed w confirm this" (EPA, 1998). could reasonably be the result of randorn variation in time and pLace" (Queensland Health, 2007). "Reports linking exposure to benzene with other malig- nancies than AML] were considered to be inad- The 1epurt concluded that: "The causes of CLL are unknovvn. equate for evaluation" (IARC, 1982). Apart from increasing age and genetic variation in risk, there are no known risk factors. No environmental risk factor has Recen1: reviews of multiple myelmna have also ernphasized been found Lo predicl risk of CLI" (Queensland Health, the lack of clarity in the etiology and mechanisms of develop- 2007). muhiple rnyeluma. The \VHO has been careful to distinguish myelogenous, rather than the hmarler category of "allleukeJnias;' as being associated 'vith benzene exposure (vVHO, 1993). TheATSDR has been more specific, stating that: "Exposure to benzene "Despite evidence for some clustering of MM and MGUS within families, the roles ofgenetic and environment remain unclear" (Bergsagel, 2005). has been associated with development of a particular type ofleukenJ.ia called acute myeloid leukemia {AML)" (ATSDR, 2007). The only mention of CT .Lin their review was to relate the re;;ults of several industry cohort studies, where the inci- dence ofCLL was not elevated in any of the of ne!. benzene exposure in workers. "The aetiology of remains essentially unknown. recent studies suggest links lo agricullural exposures and lifestyle factors, such as low socioeconomic status and obesity" {Joshua, 2005). !ARC has been similarly focused in its wording. stating: ''1l1e relationship bet1Neen benzene exposure and the development of acute myelogenous leukaemia has been estab- Multiple myeloma is characterized by the following risk factors: lished in epidemiological studies" (IARC, 1932). Age: Growing older increases the chance of develop- The case, then, for benzene exposure to CLL ing multiple myelurna. Must people with myeloma are appears to he poorly developed. \\Tell-designed studies that diagnosed after age 65. The disease is rare in people have included worker exposure data and that have considered under 40. alternative risk factors and chemical exposures will be neces- sary before it can be concluded that CLl. can be reasonably Race: The risk is highest among African Americans .s associated with present or historical exposures to benzene and lowest among Asian Am.ericans. The reasons for in the workplace. the observed difference between racial groups is not known. Benzene exposure and multiple (MM) At various times over the past 15 years, there have been questions raised about whether occupational exposure to benzene might increase the risks for developing l\1M. l\1ultiple myeloma is a disease of B lymphocytes that have matured into plasma cells and is typically very difficulL to treat. Plasma cells arc nonnal components of our itnmune system that create antibodies against invading agents. The transformation of normal plasma cell creation into the uncontrolled process of multiple myeloma, however, leads to the destruction of hone marrow, crowds out the process of normal blood cornponent rnaturation, and often leads to frequent infections due to impaired imrnune capabiliries (Rajkumar ct aL, 2005; Kyle and Rajukmar 2004; American Cancer Society, 200'1). Multiple myeloma, like CT J ., is a dis- ease of adults, with an average age of onset of 70. Almost 20,000 new cases ofMM were expected to be diagnosed in the United States in 2008, and typically represents about 10% Personal history of monoclonal gammopathy of unde- tennined significance (MGUS): MGUS is a condition in which abnormal plasma cells manufacture a variety of certain The current theory is that all cases of multiple myeloma are preceded by a period of MGUS lhat may last as long as 16-20 years (Landgren et al., 2009; Weiss et aL, 2009). other possible risk factors for nc....u.a.ucu", pesticides, hair dye, certain viruses, obesity, and dieL are under study. But it is not dear that these factors are involved in the developrnent of the disease" (NCI, 2006). Historically, there have been several case reports and a single major study that have suggested a link between benzene c>::posure and multiple myeloma. 1l1esc are discussed bclovv-. CGU BEN0000285 Benzene and l1uman. health 29 Aksoy in 1984 reported on seven patients \Nith multiple studies indicate that benzene exposure is not likely to myeloma, four of whom 'Nere workers with exposure to ben- be causally 1elated to the risk of multiple myeloma. A zene. These four patients came frmn four different industries meta-analysis of case/control studies [Wong et al., 1997] over a 10-year time period (Aksoy, 1984). found no significant association between occupational Decoufle summarized mortality statistics frmn a small exposure to benzene and benzene petroleum refinery that was convened to a chemical manu- and risk ofmultiple from sources categorized facility. A population of 2.59 workers who had as benzene and/or organic solvents, pet:Iolellln, or worked some time at the company between 1947 and 1960 pcuoleum products" (ATSDR, 2007). were follo\\ied through the end of 1977. Ah:hough the total deaths due to mahgnancywere lower than expected, the inci- dence of hematopoietic cancers was higher than expected. In the cohort, one worker died from multiple myeloma, and a second was diagnosed with multiple myeloma, succumbing ultimately to Al\IL (Decoufle et al., 1983). A careful reading of both the (1984) and Decoufie et al. (1983) papers clearly indicates that the workers were Bergsagel el aL (1999) perfurmetl a comprehensive review of the literature with respect to benzene and multiple myeloma, smnmarizing data from a wide variety ofindustries in Europe, North America, and Australia and decennined the plausibility of an association. They concluded that although high levels of benzene exposure were associated with an increased risk forAML, exposed to other chemicals, but they are not descrihed qualitatively or quantitatively. In addition, the complete occupational histories, understanding of familial medical history, and other possible risk factors arc not provided; nor is there an indication as to the lPvel ofbenzpne exposure that "ll1crc is no scientific evidence to support a causal relationship benveen exposure to benzene or other petroleum products and the risk ofdeveloping multiple myeloma" {llergsagel et aL, 1999). might have been experienced by the workers who developed disease. The Pliofilm cohort, mentioned earlier as the best popu- laLion for lhe health risks of benzene exposure, has perhaps been the most widely cited study that reported an association between benzene and multiple myeloma. Rlnsky et aL (1987) reported four cases of multi- ple myeloma ten years after their initial description of the Infante reviewed all "benzene cohort studies" reporting on the endpoint of multiple myeloma, and selected a subset of these for inclusion in a meta-analysis, coming to the oppo,;ite conclusion that there appeared to be : .. a significant excess in the relative risk of MM in relation to benzene <>v-nr'""''"''" (lnfantc, 2006). His treatment ofthc underlying data for his analysis, however, is subject to ques1ion. cohort, although one of the cases worked just four days at Rinsky 2002 noted five cases of multiple myeloma from The plant, making a ben~ene-indnced disease seem rather the exposed worker population. Two of these workers unlikely. Overall, the risk of developing multiple myeloma had very short periods of employment (4 days and l did not seem to be related to cmnulative exposure, with three month), with net occupational exposures of 0.11 and 0.10 ofthe four cases occurring in the lowest category ofexposure ppm-years, respectively. Attributing a causative role for reported by Rinsky et aL Total ppm-year~ fur the cases were benzene in these two workers with transient exposures O.ll, 7. 75, 19.5, and over 652 (Rinsky et aL, 1987). appears unju:.tified, in our opinion. .s A follow-up report on the PLiofilm cohort 15 years later, however, appeared to disiniss the possibility of an associa- The latency periods reported for all five cases were tion bctvveen benzene exposures and rnultiplc 1nyeloma (at long, ron1parerl to reported latencies for least in this cohort) (Rinsky et aL, 2002). Four new cases were other benzene-related hematopoietic malignancies described, hut three of these cases were judged to he in work- (Finkelstein, 2000; Glass, 2004}1hese ranged from 22.5 ers unexposed to benzene in the workplace. The last worker years to 38 years, with a mean Latency of 27A years; time had only a month of exposure, as well as an unusually long periods much greater th<:m accepted latencies for ben- latency period TI1esc factors in combination make zene and AML (Finkelstein, 2000; Glass, 2004). Collins it vinually implausible that benzene played a role in lhese et aL also noted a prolonged latency of disease in their illnesses (Rinsky et al., 2002). case&, wilh all l 0 deaths occurring 20 or more yea1s The Agency for Toxic Substances and Disease Registry after first exposure (Collins et al., 2003 ). Justification for (ATSDR) has sununarized tl1eir in1pressions on benzene and including these cases, and commenting upon a possible multiple myeloma as follows: mechanism by which benzene might substantially extend its latency for causing hematopoietic disea!>e in the case "The risk of mortality from multiple myeloma was increased in one ofthe early assessments ofthe Pliofilrn cohort. TI1e implication ofthis finding is unclear because the risk declined m nun-!<,ignificant levels in subsequent follow-up studies, and was not supported by the findings of other cohort mortality studies. Additionally, population-based and hospital-based case-cont:Iol of multiple myeloma was not provided by Infante. Although he provided commentary on both the 1981 and 2002 RinskyPiiofilm studie~, Infame decided to perform his risk calculations on the study with 20 fewer years of accumulated data for his arguments regarding disease association. using this earlier, less complete study substantially increased his reported risk numbers. CGU BEN0000286 30 D. Galbraith et aL Infante assumed that populations across the various studies in his meta-analysis were comparable and could be reasonably combined. However, exposures sustained by the various worker populations differed tremendously from one another. For example, half of the cases ofMM reported by Collins et aL occurred in worker with less than 6 ppm-years of net benzene exposure (25% were less than l ppm-year of exposure). Fu ct al. cslimatcd that exposures for the Florence facility may have ranged frmn 25 to 600pprn, in accordance with those reported hy .'\ksny (Fu et al., 1~~lfi). have occuned as a result of improved benzene regulations and exposure protection, but thaL explanation tends to dilute their argument that ongoing e'l\."}Josurcs to crude oil and petrochemicals are still creating excess risk of disease for petrochemical workers. In other words, ifworkers are stHl being exposed, but at lower levels, we would still expect to see some observable disease. In addilion, the authors admit that detailed information on occupation, job tasks and potential exposure was lacking in the groups that they studied, which could have contributed to exposure misdassification. The authnrs admit that: A less confident statement as to the relationship between benzene and multiple myeloma was made by Goldstein 16 years earlier. His review commented on the 1987 follow-up by Rinsky of the Pliofilm cuhmt, and appears to have provided the source for many of the biologic plausibility arguments emphasized by Infante in his later review. Goldstein "Previous studies that assessed the association between benzene exposure and hematologic neoplasms in cohorts of petroleum workers were limited by the lack of good exposure estimates including information on the of benzene exposure. 1his is also a limitation of our study:' (f{irkclcit et al., 2008). concluded that: "Overall, these findings are not sufficient to make an unequivocal statement that benzene is a cause of multiple myeloma" (Goldstein, 1990). Apparently in response to Goldstein, Bczabch et al. reviewed population-based and hospital-based t:asecontrol studies looking at the relationship between MM Costantini et aL (200H) conducted a population based case- control study in and used questionnaire data to assess exposures to solvents and other envirmunental agents. In this study, the authors reported odds ratios for select agents in 586 cases of leukemia and 263 cases of MM. 1hc risk for the development ofl'vfM in the "benzene exposed" workers was and benzene exposure. Odds ratios were approximately l.O from these studies, and although some indhridual sHJdies of petroleum combustion products scc1ncd to indicate an increased risk of disease, studies focusing on exposures to benzene, solvents containing benzene, and cigarette smoking all failed to show a positive relationship. The authors concluded that: not statistically significant, suffered from very small num- bers, and showed very low intervals very wide in most cases). In particular, in workers \Nith over 13 vears of benzene exposure, five cases and ihree control subJects were reported with odds ratio of 4.1 (CI= Interestingly, the authors also reported ve1ysimilar associa- tions for rnull.iple myelmna with both xylene and toluene in "1he current published case-control literature on benzene exposure is not ambivalent and does not indicate that benzene exposure is a risk factor fm multiple myeloma" (Bezabeh et al., 1996). the 15+ year exposure groups (odds ratios of 3.1 to 17.0] and 3.1 to 17.2], respectively). lhese reponed findings, which are inconsistcncwith the literature on xylene and toluene health effects, provide further doubt as to the validity of the Costantini et al. data. More importantly, the .s authors failed to find any association betvveen exposure to Two recent papers have suggested a relationship between any solvent benzene) and the incidence of AML. benzene exposure and the of multiple mye- This fact alone makes their claims of association between loma, Kirkeleit et al., (2008) and Costantini et aL, (2008). benzene exposure and other hematopoietic diseases much However,in light of the following shortcomings, we consid- less plausible. 1hc authors beLieve their failure to find A'VH. ered them unreliable. in the population was "explained by the strict regulation of As discussed previously, Kirkeleit et al. (2008) reported benzene in Italy nearly three decades prior two study ini- on the mortality statistics peuoleum workers in tiation" {Costantini et al., 2008). If this reasoning is correct, Norway during two separate lime periods between 1981 to then slrict regulation of benzene should also remove the 2003: "First exposure 1981-1985" and "First exposure 1986- risk of other hematopoietic malignancies; a fact they fail to 2003,'' In the absence of actual quantitative exposure data acknowledge. for the worker population that was studied, the authors ': .. In a recent study of 279 patients with multiple assume[d] that the 'upstream operators offshore,' who have compared to 782 control patients, variants ofgenes involved had ihc n1osl extensive contact with crude oil and natural gas, in the metabolism of exogenous chemicals were to was highest and most hon1ogcneously exposed to benzene:' determine a possible association 'Nith altered metabolisrn of (Kirkeleit et al., 2008). Upstream operators in the earlier various chemicals and the risk for developing MM. NAD{P) time period were found to have an increac;ed risk H:quinone oxidoreductase (N(.)Ol) was selected as a focus v.rith nine observed cases and a reported relative for the metabolism nf benzene, whereas genes involved in risk of 2.85 compared La Lhe general population. However, the metabolism ofpolyaromatic hydrocarbons (PAHs), diox- in the later time period, no cases of multiple Inycloma were ins, pesticides, and other agents were selected to evaluate reported. The authors explained that this change might MM risk with altered metabolism for the.se other chemicals. CGU_BEN0000287 Variants in NQ01 were not found to be associated with an increased risk for MM, providing suggestive evidence that variable abilities to metabolize benzene were not found to be associated with an increased risk for disease (Gold et al., 2009). \IVith the recent classification WHO as a "Mature B-cell ucupu-'""-' ing l'v1J'v[ and benzene exposun~ will likely become absorbed into rhe discussion ofNHL and benzene exposure. After con- sidering the available studies, similar to our expressed opin- wns NHl., we dn not feel that there is sllfficient evidence to establish a relationship between multiple 1nyelmna and benzene. Benzene exposure and the (MDS) The myelodysplastic syndromes present a heterogeneous set ofneoplastic diseases that are characterized by in the peripheral blood, dysplasia in the bone marrow and pro- gressive failure of hematopoiesis in one or more ofthe mye- loid cell (Bennett, 1982; Taffe et al., 2001; I.ist, 2004; Swerdlow, 2008). MDS are thought to originate in hematopoi- etic stem cells and a diagnosis of MDS offers an increased risk for the development of certain types ofAML 1!198). ol"l'viDS to Al'vfL can occur as a natural course ofthe disease, the diagnostic distinction between MDS and AML is simply determined the threshold percentage of myeloblasts observed in the peripheral blood and bone marrow: greater than 20% myeloblast;; meets the criteria for Benzene and l1uman. health 31 a diagnosis of Al\H, and less than 20% myeloblasts meets the criteria for a of MDS (Jaffe et aL, 2001; Swerdlow, 2008). 'Ihc MDS subtypes are morphologically categorized by their varying degree of blast cells, lineage dysplasia, and presence sideroblasts (e.g., ringed siderohlasts are erythroid precursor cells v,'i.th an accumulation of iron in the mitochondria). As with WHO's standardized criteria lor diagnosis oflcukcmia subtypes and NHL, adYanccs in UlnJ.Or biology and diagnostic tools used to identify histopathologi- cal and genetic features have allowed the WHO to set forth uniform criteria for the diagnosil'i of MDS subtypes in the WHO tumor classification scheme published in 2001 and 2008 (Jaffe et aL, 2001; Swerdlow, 2008). Historically, epidemiologic reports studying the risks of occupational chemical exposures for developing hemat- opoietic disease have frequently not provided details on MDS subtypes. This lack of more specific MDS data has made it extremely difficult to associate exposures to benzene and other chemicals with specific subtypes. lhe changing definition ofthc various subtypes over the past 30 years has emnplicated matters even furthel". For example, MDS was ofcen referred to as "preleukernia;' state," or "subacute leukemia;' terms that may encourage confusion "\<Vith other diseases (Block, l!-l5:1; Dreyfus, 1976; Jaffe el aL, 2001; Swerdlow, 2008). As such, the historic reports often included either a description of MDS along \'lrith a diagnosis of AML or MDS was categorized with other chronic myeloproliferative diseases (e.g., myelofibrosis) (Jaffe et al., 20D1; Swerdlow, 2008). .s - Figure 2. A schematic flow chmt adapted from Bennett {2005) illustrates MDS subtypes as defined by the 2008 WHO. ;\(Bz) notation repTescnts that an association between the l\1DS subtype and benzene exposure has been reported. CGU BEN0000288 32 D. Galbraith et aL Tn 1976, the F.l\H cooperative proposed a classification classification scheme for MDS, refractory cytopenia with uni- scheme for a distincc group of conditions wherein the bone lineage dysplasia (RCUD) and refractory anemia with ringed marrow exhibited hyperccllularity but, unlike AML, fewer sideroblasts (RARS) represent MIJS subtypes that are most leukemic blast cells were present in the bone marrow. The commonly associated with erythroid dysplasia (Swerdlow, FAB classification for this group was the "dysmyelopoietic 2008) (Figure 2). RCUD includes anemia, syndromes" and two distinct subtypes were identified: neutropenia, and refractmythrompocytopenia as defined by refractory anemia with excess blasts (Ri\EB) and chronic a single cytopenia in the peripheral blood and a line- myelomonocytic leukemia (CMML). In RAEB, chc eryth- age dysplasia in the bone rnanow. Bone manow findings for roid component predominated in the bone marrow and RCUD also include less than 5% myeloblasts and less than the erythroid precursor cells exhibited hyperplasia and 15% ofthe erythroid precursors are ringed siderohlasts (Figure dy!iplastic changes with or without the pn:sence nf ringed. 2). RARS is defined hy anemia in the peripheral hlnod and sideroblasts. In CMML, the granulocytic and monocytic erythroid dysplasia in the bone marrow. Additional bone mar- components predominated in the bone marrow and the rowfindings forthe classification ofRARS include greater than peripheral blood monocytes were elevated and displayed 15% ofthe erythoid precursor cells are ringed sideroblasts and morphology (Bennett, In 1982, the FAB the observation ofless than 5% (Rennctt, 2005). began referring to the ~;yndromes as "myelodysplasiic" Both RCUD and RARS have a low incidence of transforma- and proposed a new classification schen1e following their tion to AML (Swerdlow, 2008). Refractory 'Nith review of the morphologic features of 80 separate cases, multilineage dysplasia are represented by one or more cyto- some of which exhibited difficult diagnostic clarity. In the penias in the peripheral blood and dysplasia in at least two 1982 report by Kennett and colleagues, described the myeloid lineages in the bone marrow. Levels of blast cells in morphologic features ofthe bone manow in caseli ofl'vfDS, the bone marrow are less than5%. Roth RAER-1 and RAFR-2 and these features were used to define five distinct subtypes. are characterized cytopenia(s) in the peripheral blood and These subtypes included refractory anemia (RA), RA with unilineage or multilineage dysplasia in the bone marrow; the siderohlasts (RARS), RAEB, RAEB in transformation criteria for hone marrow blast cell number remain the same (RARB-l), and CMML 1982). as in \\THO, 2001 (Swerdlow, 2008). In 1997, the vVHO appointed a new committee thal revised Il is noleworthy lhal MDS wilh an isolated 5q delelion and updated the FAB classification scheme. In the process of arc classified as an independent MDS subtype under \NHO revision, the cmmnittee took into account blast cell percent- 2001 and WHO 2008 criteria. 1his subry~tJe is unique in that age, dysplastic features, prognostic features, and therapeutic cells exhibits a del{5q) cytogenetic abnormality and outcomes (Bennett, 2005). The new criteria were published the patient may present vv:ith increased numbers of plate- in 2001 by WHO and several distinct changes in the 1982 FAR leis {thrombocytosis) instead of rhmmbocytopenia in the classification were noted: peripheral blood. Additional features characteristic of MDS del(Sq) include anemia with or without other cytopenias in l. RARU-t was eliminated as a MDS and patients the blood and blast cells observed at less than with bone manow containing greater than 20% blast 5% in the bone marrow (Jaffe et al., 2001; Swerdlow, 2008). cells were considered to have AML. Myelodysplastic s~'lldrome-unclassifiable (MDS-u) represent .s 2. The categories ofRAFB-1 and Ri\EB-2 were created with clear cases of bone marrow dysplasia but they do not fit under blast cell percentage used as a distinguishing feature any other MDS subtype classification. Two indications for (5-9% and 10-19%, respectively). classification in trte JVIDS-u category: 3. CMML was eliminated as a MDS subtype and a new cat- egmy was created, Myelodysplastic/Myelopmliferative Diseases, which now included CMML. 1. 1hree additional MDS subtypes were created which with multilineagc dysplasias (RCMD), MDS wilh no classifiable features (MDS-u), and the 5q minus syndrome (distinct from 5q- chromosomal abnormality) (Bennett, 2005; Jaffe et al., 2001). L AMDS case with lineage dysplasia in one or more of the myeloid lineages that is accompanied pancytopenia, and RCUD or RCMD subtype that exhibits less than l% blast cells in the peripheral blood. 2. A MDS case represenled by bone marrow dysplasia in one or more lineage with fewer than 1% blast cells in the peripheral blood and less than 5% blasts in the bone The latest revision of the MDS classification scheme was published by ~VVHO in 2008 and again the majoriLy of MDS marrow that is accompanied by a cytogenetic abnormality (Swerdlow, 2003). subtypes were distinguished by the obse1~ved percent of bone marrow blast cells, presence or absence of ringed sideroblasts, and evidence of dysplasia (Bennett, 2005). The 200!:1 WHO criteria is used for the of hematopoietic diseases, ahhough the classification of MDS continues to evolve and 'INill no doubt be refined as we learn more about the biology ~md etiology of this disease. In the \1\/HO 2008 r\s mentioned previously, the v\T.HO classification schema for clonal hematopoietic neoplasms are a "work in progress" and, as our knowledge increases, MDS will be further defined (Swerdlow, 2001'1]. As such, our interpretation of the association bet\veen benzene and the development ofspecific MDS subtypes most likely vv1ll evolve as well (Figure 2). Our current understanding is described below. CGU BEN0000289 Benzene and l1uman. health 33 Renzene and preleukem.in included evidence of hypocellular bone marrow and dys- Although smne historic epidemiologic reports have described benzene-induced AML as a transition through a "preleukemic" state, MDS has rarely been reported as a toxi- cological endpoint. For example, Aksoy repmted that leukemia" was a comn1on observation in benzene-exposed workers in Turkey (Aksoy, EJl:HJb). Some of the older reports of benzene-induced hematotoxic effects refer to pancyto- penia occurring in the presence of "paradoxical hyperpla- sia" of the bone marrow, which is reminiscent ofMDS. The blood Ihat have been associated whh high levels ofbenzene nut only include cytupenias, pancytopenia, ane- mia, and aplastic anemia but also "preleukemic" dysplastic changes rhat arc nov.r generally consistent with MDS. In con- trast, approximately 10% of the MDS cases diagnosed today are hypoplastic, making the differential diagnosis between MDS and anemia difficult (Swerdlow, 2llDH). That said, benzene-induced aplastic anemia as described in earlier studies may nm meet currently accepted criteria for the proper diagnosis of aplastic anemia {i.e., pancytopenia and hypocellularily of the bone marrow). :Moreover, some of the cases ufben:zene-induceu aplastic anemia were diagnosed primarily by pancytopenia in peripheral blood or by bone marrow at autopsy (Gross, 2010). By the early 1960s, bone Inanow biopsies paved the way for the of marrow archiLecmre to assist in making a correct hematologic diagnosis (Me Farland and Dameshek, 1958; Parapia, 2007). erythropoietic (Line1, 2996). Irons et aL (2005) reported on the bone marrow mor- phology in 23 Chinese workers exposed to high con- centrations of benzene. The authors vVHO 2001 criteria for diagnosis and reported that the bone marrow pathology of the workers was characteristic of hypoplasia, sumnal degeru:racion, dyserythropoiesis, multilincagc dysplasia, and severe dysplasia in eosinophilic precursor cells (Irons et aL, 2005). Even with the improvement in diagnostic tools, the mrnphologic hy Iron;; and colleagues are similar to that described bv Ruis and col- leagues in the Brazilian workers in l994. However, Irons also observed hematophagocytosis and donal expan- sion ofT cells; observations that the authors suggest arc a result of immune activation. In addition, Irons found no cytogenetic abnormalities in any of the ""uu1'"" Irons concluded that these findings do not sup- port the hypothesis that the disease process of benzene- induced /\ML shares a comrnon mechanism with the developmenr ofth ated AMT following exposure to alkylating agent or radiation (Irons et al., 2005). The authors acknowledged that quantitative information on the development of MDS subtypes in benzene- exposed individuals is curremly and fuwre studies will no doubt. provide insight into Lhe disease specificity (Irons et al., 2005). The suggestion by Irons and colleague~ that benzene-induced rnorphologic Lmne marrow changes were immune n1ediated was recently supported in a study Tn n1ore recent liteTature, the hone n1arro1N performed by Song et aL in which 19 out of 22 natients with henzene-indu~ed hemalotoxicityresponded .to trear- morphology in benzene-exposed cases has been described menr with immunosuppressive therapy with cyclosporin but resolmion for MDS subtypes associated with benzene A (Song et al., 2010). exposure is For example, Ruis and colleagues (1994) described bone marrow morphology in 192 workers exposed Benzene and RCUlJ to high concentrations of benzene at a steel plant in Brazil. As the 2008 WHO classification scheme com- .s The authors noted that the majority of bone marrmv samples they exarnined were hypoplastic with a large decrease in bined RA with refractory neutropenia and rcfractorv thrombocyt:openia into a single disease subtype, RCUD (S;IVerdlow, number of granulocytic precursors docurnented. The authors to the 2!lll8 VVB 0, RCUD is reported erythroid dysplasia, atypia in alL three myeloid in older individuals and makes up approximately 10-20% of cell lineages, and frequent stnnnal changes. Eosinophilia all J\1DS cases reported. FLA.. is diagnosed most frequently, was also documented (Ruis et aL, 1994). Travis et aL (1994) whereas refractory neutropenia and thrombocytopenia are defined a series of seven MSD cases within a cohort of rarely diagnosed (Swerdlow, 2010]. It is not clear whether 74,828 workers from across China. MDS RCUD can develop in benzene-exposure indi-viduals, as onlv a subtyping was performed using the Proposed Classificalion few cases have been reported in the literature, although single Schema for MDS published by Bennett in 1982. The authors cytopenias, rnost notably anemia, leukopenia, and thrombo- described MDS subtypes in four out ofthe seven MDS cases. cytopenia, have heen reported following benzene exposure docurnented one case of RA and one case of RAEB in (ATSDR, 2007). Using the 1982 l<AB criteria for diagnosis of transformation. The other two MDS cases were defined as fvlDS, Tra-vis and colleagues reported one case of RA frorn a "not otherwise specified" (NOS). Travis also included one cohort of74,828 benzene-exposed workers (Travis er 1994). case of CMML, which would not be considered under the Irons et aL incorporated the 2008 \'VHO classification system umbrella of MDS by WHO 2001 or WHO 2003 criteria for in a recent report of 29 cases ofl'viDS chat were occupationally (Travis et aL, 19!:14). Liner et aL (I 9!-lii) published to high levels of benzene and the authors referred "-"-'H-'1';" from this same cohort and associated seven to these cases as "benzene signal cases:' 1l1e autl1ors noted cases of MDS with benzene exposure. 1be authors provided that out lhe 29 benzene-exposed signal cases, Lwo cases ofRA, an overall description of the bone marrow rnorphology and twu cases ofHCUD were diagnosed. However, the authors in benzene-exposed workers diagnosed with MDS that determined that ndds ratios fnr the development nfthese RA CGU_BEN0000290 34 D, Galbraith et aL and RCUDsubtypeswere 0.67 (95% CI 0.13-3.3) and 1.0 (95% Cl 0.18-5.4), respectiveLy{Irons et aL., 2010). RCMD subtype following occupational exposure to benzene was 0.58 (95% CI 0.2-1.4) (Irons et aL, 2010). Benzene and RARS Benzene nnd IVJ.DS-u RARS has been a recognized disease entity for over a half 1be unclassifiable category ofMDS subtypes is another new a century and the unifying diagnostic feature ofRt\RS has category added to the 2001 \'VHO classification scheme for been the presence of ringed .siderobla.sts in the mitochon- MDS and basically this categorywas recreated to define MDS dria of erythroid precursor cells, which arc typically iden- cases that essentially lack feamres appropriate for classifica- tified by iron staining of the bone marrow cells (Natelson, tion in any other subtype (Jaffe et aL, 2001; S\verdlow, 2008). 2007a; Aul, 1998; Bennett, 200!J; Catenacci, 2005). RARS Not surprisingly, the incidence for the development ofMDS-u occurs in olde1 individuals and accounts for is unknown and there are only a few reports of the uevelop- approximately 3-11% of MDS cases (Swerdlow, 2008). menl ofMDS-u following occupational exposure lo benzene :\s for benzene, the cumulative literature indicates that (Swerdlow, 2008; Irons ct al., 2010, Travis ct al., 1994). In exposure to benzene is consistently not associated with 1994, Travis and colleagues reported two cases ofl'viDS-nut the development of RARS {Bennett, 2005). In a recent other\ivise (NOS) in their cohort of74,828 benzene- publication, Natelson summarized the bone marrow exposed workers (Travis ct aL, 1994). However, reclassifica- morphology from several primary studies on benzene- tion of these two MDS-NOS using current criteria for MDS exposed workers, and pointed out that there was a lack of subtypes may result in different conclusions as to disease documentation for ringed sideroblasts (Natelson, 2007a). risk In a later Mudy, Irons and colleagues found an increased For example, analysis of the bone marrow by iron stain number of MDS-u cases 7) \vithin their 29 benzene from some of the AML and MDS cases identified from the cases as defined in their 2010 smdyJ and they reponed that cohort of 74,828 of workers showed no the odds ratio for the of this disease subtype indication of ringed sideroblasts (Linet, 1996). In Aksoy's '"ms lLl (95% CI 1.34-92.4) (Irons et aL, 2010). study of Turkish shoe workers, he described the bone rnarrow abnormalities associated with benzene exposure for Benzene and 5q minus syndrome the AML cases identified from the shoe worker cohort and As previously discussed, 1V1DS wiih an isolated deletion of he reported that none of the ohservations was consist- chromosome (5q-1iyndrome) is a unique l'v1DS suhtype ent with RARS (Aksoy1 198Gb). Independently, Irons et al. (:WlO) and Ruiz et al. (1994) provided reports lA'ith a donal chromosome aberration and distinct set of clinical features that include presentation in older indi- of bone marrow abnormalities in workers occupation- viduals, preferentially wornen, a 1nedian survival ally expuseu lo benzene anti in both auLhors' descripLion :rale of grealer than 10 years, and less than 10% uf Lhe cases ringed sidcroblasts were not noted (Ruiz ct al., 1994; actually to AML (Swerdlow) 2008), Sq minus Irons et al., 2005). In addition, a 1992 study conducted by syndrome should be distinguished from therapy-related MDS Garand and evaluated whether any of the R4 and AML, which can occur following therapy with alkylating patients who were diagnosed with HARS recalled specific agents, evolve quicldy to k\IIL, and often exhibit cytogenetic exposure to benzene. One out of the 84 RARS parienls sur- abnonnalities invohing deletions in chromosomes 5 and veyed confirmed previous exposure to benzene, although 7 and/or complex cytogenetic abnormalities involving rhc .s radiation therapy was also reponed in this patient (Garand two chromosomes; these are nvo distinctly different diseases et al., I 'IYith different disease outcomes (Jaffe et aL, 2001; Benzene and RC:JI.1D The catcgoryofRCMD is a fairlyncwMIJS subtype and was only incorporated into the MDS classification schente in 2.001 WHO (Jaffe et al., 2.001]. RCMD accounts for approximately 30% of diagnosed MDS cases (Swerdlow, 2008). Additional demographic characteristics of RCMD include a preponderance fur the development of lhe disease in Swerdlow, 2008). Hm1rever, exposures to benzene and other solvents have been associated with cytogenetic abnormalities involving chromosome 5 and 7 (Brandt, 1992; Rigolin, 1998; Natelson, 2007b; Pederson-Bjergaard, 1981; Fagioli, 1992; Cuneo, 1992), Renzeneand RAFR Il is important lo nole that any one or the MDS subtypes older males {Jaffe et al., 2001; Swerdlow, 2008). Due to the described above can evolve or transfonn into RAEB and recent introduction of this subtype into the classification that RAEB has been characterized as an MDS subtype in schema1 reports on the development of RCMD following occupational exposure to benzene are limited. Irons and the MDS classification :o.cherne since its inception in 1976 (Bennett, 1976). \Vithout attention to morphology, cases colleagues (2010) identified 17 cases of RCMD in the ben- ofMDS are diagnosed as RAEB-1 if the blast cells observed zene signal cases outlined in their recent publication. But it in the bone 1narrow are between 5% and 9% and RAEB-2 is important to note that in this same study, 441 out of 611 is assigned to cases where the bone nnrrow blast cells fall MDS cases that were diagnosed over the 4-year period between 10% and 19% (Jaffe et al., 2001; Swerdlow, 2008). were classified as RCMD, which resulted in 72% of the total It is also important to note that as of 2001, the criteria MDS case;, reported (hans et al., 2010). Funher, the a1nhurs fur the diagnosis of RAEB has changed in tet-ms uf blast reported the odds ratio associated with the deYelopment of cell number. under 20fll criteria, the RAEB-t cases with CGU BEN0000291 Benzene and l1uman. health 35 20-29% blast cells were considered to be 1\ML (Bennett, 2005). This 2001 change in definition for MDS and AML has implications in the benzene literature, as there arc several reports of the development of RAEB following exposure to benzene_ For Travis (1991) noted one case of HAEB in their cohon and Strom and colleagues docu- rnented an increased risk in the development ofHAEB and RAEB-t following exposure to solvents, including benzene. Both of these reports used the 1982 FAB criteria for the classification nfl'vfDS subtypes (Travis d aL, 1994; Strom 200S)_ Using the 2008 criteria, Irons et aL (2010) noted two cases of RAEB in the described benzene-exposed signal group, although he reported that in his study, the odds ratio for the development of RAEB subtype was 1.4 (95% CI 0.19-lLl) (Irons e1: aL, 2010). As with NHL nomenclature discussed pre"l-'iously, it is difficult, if not inappropriate, to directly compare historic epidemiologic reports of "preleukemia" and even "RAEB-t" with the current MDS classificmion scheme. control of different mammalian genomes may also play a significant role. Second, there is a lack ofagreement as to how benzene or its metabolites cause the injury that leads to leukemic and conditions_ Four main n1echanisnu; have been as described by et al. {20D4): L DNA adduct formation or cross-linking oxidized metabolites ofbenzene, such as hydroquinone, phenol, or benzoquinone 2. Direct oxidative damage tu DNA strands by these intermediates 3. Physical damage to progenitor cell protein microstructures, which then directly affectl' the ability of these cells to replicate and divide 4. Direct inhibition by benzene and/or its metabolites on enzyme function, particularly a class of enzymes called topisomprasf's that opeTate in the coiling and uncoiling of DNA and assist in the transcription and process of cells studies and modelsfor benzene Over the years, benzene has been implicated as a cause or contributor to a wide variety of conditions, both acute and chronic, and nonmalignant. lhe best efforts of researchers across lb.e Lhe exact mechanism of how benzene may lead to harmful effects in humans is still not well understood. 1here are several reasons for this. First, there is still no good animal model that has proved reliable for Transgenic mouse n1odels have been used bysorne research- ers in an attempt to shorlen the latency of mutational events leading to hematopoietic diseases such as leukemia and lymphoma. Studies of benzene in drinking water, for exam- ple, have shown that benzene the proliferation of hematopoietic progenitor cells in transgenic mice and altered gene expression in key signaling pathways {Kwosu simulating the development of human leukemia after heavy nnd prolonged benzene exposure. Table 6. Risk factms for leukemia. Genetic predisposition "Animal responses to benzene exposure are variable Do'INn syndrome and nmy on factors such as strain, Fanr:oni's anr:min duration of exposure, and whether exposure is inter- Bloom syndrome mittent or continuous. Wide variations have also been Blackfan-Diumond syndrome .s observed in nonnal hematological parameters, compli- '" Ataxia telangiectasia cating statistical evaluation" {ATSDR, 2007). Polyrnorphisms of genes active in metabolism and detnxHication Polynwrphisms of genes responsible for DNA repair enzymes "Lack of genetic concordance between experimental hematopoietic neoplasms and human leukemias is a key limitation to the use of many anirnal models to leukemia hazard assessment" (McCormick et aL . 2004]. Family history of leukemia TreaLrnenl vvHh chemolhe.rapy agenls1 particuL-uly aLkylaHng agenls and topoisomcrasc inhibitors History nfprinr hemarnlngic disnnlers "" Myelodysplastic syndrome o Myeloproliferative diseases TI1e animal studies lhal are mosL frequently cited regard- ing the inhalational carcinogenicity of benzene have been performed in rats, where tumors of the Z~rmbal gland (an auditory sebaceous gland not found in hun1ans) and ural cavity have been most commonly observed. In mouse stud- ies, a broader range oflesions have been observed, including hem.atopoietic rnalignancies_ 1he reasons for not being able to develop an effective and animal model for benzene-induced leukemia are not well understood, but may be influenced by the metabolisms and detoxification 1>ysterns relied upon by different mammalian species {Henderson, 1996). 'The varying organization and Exposure to radiation, either therapeutic or environn1ental ' Diagnostic radiation is typically of much smaller magnitude, but exce&&ive expo~ure to ce1tain t}11es of prot:edures 1nay irn1Jart increased risk Occupa-donal exposures ~ Chronic exposure to benzene exceeding federally approved safety limits (fur AML) Environmental exposuTes Certain viral agents Cigarette smoke} primary and secondary DiP.lary far:tors Sources: American Cancer Society, 2009; National Cancer Institute, 2008; McCormick and Kavet, 2004; Leukemia and Lymphurna Sudety, 2009b. CGU BEN0000292 36 D. Galbraith et aL 15 0 C) 12 0 ci 0 9 Q; "0"c"' 6 "'-;:o;; -'= 3 0 1985 15 U.S. Leukemia Incidence' 1989 1993 1997 2001 U.S. Leukemia Mortality 2005 12 0 C) C) ci 09 a3' D. ~, 6 ~ ::0;; 3 - - -' o~~~~~~~~~~~~~~~~~~ 1985 1989 1993 1997 2001 2005 \Nhi1es ~ Hispanics+-o. /\5ians/PadfiG lslande(sH "Significant data for American Jndic;ns/A/.aska/'1 Natives not available ~"'"Dala for Hispanic-S arrd Asicms/,0 a-eific isl.onder.s .wl available before 1992 Dr. Zhang summarized the various possible mechanisms of benzene hematolOxicity at the 2009 Munich Benzene S)lmposium as follows (Zhang, 2010): Generation of reactive oxygen Direct chromosomal u.tu"'n"' Inhibilion of topoisomerase II Imn1unc dysfunction DNA methylation Accumulation of toxic rnetabolites in the liver Benzene has been identified by several researchers as an excellent candidate to exploit discovery strategies made possible advances in and toxicogenetics (McHale et al., 2009; Bollati ct al., 2009). These approaches will allow more cmnprehensive and accurate risk assess- ments for any chemical with characterized biomarkers and other "bioindic.ators of disease" (Edwards et aL, 200H). cil'le new "systen'!s biology" approach is a major focus of the National Institute of General Medical Sciences {NIGMS; one of the NationaL Institutes of Health), which currently supports ten National Centers for Systems Biology (NIH, 2009a). 'll1is trend of integrating experimen- tal, computational, toxicologic, and clinical disciplines to and better understand fundamental biologic proc- esses in cells, tissues, and whole represents a paradigm shift that will undoubtedly deliver enormous insight and ultimately better treatments for a wide variety of diseases. Figure 3. Age-adjusted total leukemia incidence and mortallty in the United Staies compiled by Surveillance, Epidemiolot,'Y and End Resuiis (SFER). Sow,-e, NCI, 2005. Trends di'ieU.'H': Leukemia is an uncommon disease, -vv:ith an annual incidence 16 cases per 100,000 in the male population (i\'CT, 2005) and the great majority of cases occur in people with no identifiable risk factors. et aL, 2(J{Jc1 ). The relevance to inhalational exposures is .s unclear. "\\Teak or questionable exposure data may cause consid- Other researchers have <Ilternative to erable confusion in the interpretation ofstudies, mainly accelerate or increase the likelihood of 1nalignant transfor- if no personal measurements are available. mation in the animal populations studied. Trp-53 is a tumor known risk factors accotmt for about 15%-20% of the suppressor gene product that inhibits malignant transfornla- incident cases of lcukaeinia worldwide. 'll1us, the tion (Siddique et aL, 2006). Mice vvith a Trp-53 deficiency are majority ofleukaemia cases still remains unexplained" less able to induce apoptosis follm-ving genotoxic damage. (Zeeb, l9~J8). Trp-53-deficient animals thus show an increased genomic insl.abilily, and will a heighl.ened rale of malig- A compilation of risk factors for leukemia is in nant transformation after DNA darnage caused by various Table 6. genotoxic chemicals. Kawasaki et aL recently examined Trp- Although diseases such as mesothelioma have been 53-deficicnt mice, and were able to induce hematopoietic related to contemporaneous asbestos exposure levels and cancers in a dose-dependent fashion in up to 100% of ani- regulatory guidelines (Price et aL, 200<1), no such relationship mals, demonstrating the importance of the Trp-53 function in prew~ming malignant transformation (Kawasaki eta!., 2009). has been demonstrated in the case ofbenzene and leukemia, when looked at either by total leukemia incidence, or when Researchers are that similar ''knockout" studies subdivided into the disease most closely linked with benzene vvith mice deficient in various metabolic, DNA repair, and exposure, AML. Surveillance, Epidemiology, and End Results other key functions will continue to expand our knowledge (SRF.R) data from the National Cancerlnstitute compiled in of a of disease pathways, including the events lead- 2008 (Figure 3) show that age-adjusted total leukemia inci- ing to transformation following benzene exposure dence in the Umted States has remained remarkably con- (Meek, 2010). stant from 1975 to 2005, with a range of only 11.8 persons CGU BEN0000293 Benzene and l1uman. health 37 per 100,000 to 13.5 persons per 100,000. 1l1ese data seem to A few years ago, a detailed literature review summarized argue against any influence ofexposure to benzene due w its occupational exposures to benzene where quantitadve data presence in gasoline or vehicle exhaust, whose compositions were available in North American studies across a wide have changed markedly over rhe past 30 years. variety of industries (Wijngaarden et al., 2003). The analysis The ratio of affected n1ales to affected females has also concluded that: remained constant at around L 7 to L AML data during lhe same time period from 3A to 4.0 cases per l 00,000, with a male:female ratio of 1.53 to 1 (NCI, 2005). As we have mentioned earlier, the accepted Latency period for benzene exposure to developing rnalignant transformation is less than 20 years frnm the date of fin;t exposure, with the strongest association being found for disease occurring within "IL is surprising that in of the focus on benzene since the 1980s and the longtime recognition of its hematological effects, so little information was avail- able on levels and determinants of expo;,ure for tnany industries in the United States and Canada with poten- tial exposure" (Wijngaarden et al.. 2003). I0 years following e:Kposure. No doubt, this figure is not precise since the latency should be based on the date offirst "signifi- cant (perhaps defined as l-1 0 ppm or greater on an 8-hour TvVi\ basis). One would expect that with more stringent regulatory standards and improvements in industrial hygiene, we would be able to observe a drop in disease incidence in the last several decades as a result of reduced occupational exposures. 'l he lack of a decrease in AML and total leukemia incidence in the 1975-2005 time pedod could be explained by occupational exposures to benzene already having fallen There are no scientific studies showing that workers in occu- pations with "trace benzene" exposure are at increased Tisk ofbenzene related disease. For industries that do not possess data, exposure simulations that mimic Lhc working environments and occupational practices for certain classes of workers can provide a compelling basis for char- acterizing exposures and historical and future health risks {Madl et al., 2002j Williams et al.J 2007, 2008; Hollins et aL, 2009). below "threshold" levels for benzene that at one time would have increased one's risk for disease. urp<- ofShort Term or Although studies of occupations with historically F.nvirmunenlal exposures are conslanl during lhe nificanl benzene exposures have provided evidence of the compared with occupalional exposures, which often have association of benzene with human disease, it should be two components: a low background concentration and a recognized that at low exposLtres, the!'.e sante associations nurnber of intennittent events which often involve so-called are not well supported. Jn particular, a comprehensive meta- exposures:' For example, in the majority of studies analysis ofworkers in the petrochemical industry involving in petroleum workers, it has been reported chat industrial over 200,000 workers found no increased risks for mortality were collected during a dozen or so spe- due to any of the four leukemia subtypes, even though cific tasks within each workday. These tasks were often less low airborne concentrations surely existed in 1:his industry than 15 rninutes in duration and included activities such as (Raabe, 199fi). Specifically, for the most part, workers expe- from a pipeline or in-line filter rienced mean exposures during the average workday ofless both tasks are known to present some degree of elevaled than l ppm, and had lifetime exposures less than 45 ppm- exposure unless respirators are properly worn (Gaffney et .s years. TI1is is reinforced by recent studies of various Exxon aL, 2009). refineries (Gaffney et aL, 2009; Panko et aL, 2009). The first rnention of concern abouc the ):JOSsiLle signifi- An industry-wide survey of worker exposures cance of peak exposures may have occurred in 1983, when from over 14,000 samples found a mean exposure of0.22 ppm Dr. Richard Irons addressed the International Conference (Runion, 19flfl) even during an era when benzene was not as on Benzene sponsored hy the C:ollegiun1 Ramazzini (frons, well controlled or strictly regulated as today. Similar levels of 1983). Certainly, it\vas among one of the first times that ben- exposure were reponed for workers in the che1nical indus- zene and exposure" were the focus of the discussion. try, who like>'Vise have been found to have no significant Dr. Irons noted at that time that ~ ... our evidence increased risk for leukemia (Bloemen, 2004; Bond, 1986; Ott, would highlight a concern that has been focused recently on 1978). Exposure, then, is the critical feature distinguishing the transient exposure in the occupational environment rather levels of risk faced by the workers involved in these different than continuous low level exposure:' He also stated: "In industries. INhere heavy, frequent exposure has been found, dosing, I would like to mention, vis-a-vis the experin1erual and occupational standards have either not existed or been regimen that we've used, intermittent exposure appears to ignored (Aksoy ct aL, 1972, 1974), significant risks of malig- be much more potent at producing bone marrow effects than nant and nomnalignant diseases have been found. Dermal is continuous C)tposure, and it may be that protection of the expo!i.ures also factored into the total absorbed dose, but this worker in an occupational setting requires prevention ofpeak route of exposure probably constituted less than 10--20% of exposures rather than the progressive lowering of the T\VA the overall dose (Williams et aL, 2005). occupa- in the absence of or transient exposure tions with nominal or "trace" levels of benzene exposure si Lualions" {Irons, 1983). Our reading of this now quHe daled would be expected to have a or no increased risk paper is that the intermittent c>:posure regirnen he tested had of disease. nothing to do with 15 min peak exposures versus 8 hr TWA, CGU BEN0000294 38 D. Galbraith et aL but rather, he was concented about the the increased toxicity (per workday) should be the best dose metric for predicting ofdaily doses significantly above the occupational exposure adverse effects, provided that peak blood concentrations limits. However, for manyyears, his statement has been cited remain below levels in which significant in vitro toxic effects as evidence for potential increased benzene toxicityfollovving are induced. It is acknowledged that duration of exposure at intermittent peak exposures, even when the peaks were as concentrations around 25-30 ppm can be biologically signifi- low as 10 ppm for up to 15 minutes per day. cant since several researchers have reported increased risks of Tvventyyears ago, Paxman and Rappaport (1990) tackled frank benzene hemalotoxicily and leukemia in workers with this issue regarding the hazard of peak exposures directly dnonic workday 8 hr/day exposures to benzene averaging when they decided to evaluate The scientific basis for OSHA about 30 ppm or higher (Aksoy and Erden, 1978; Aksoy et promulgating an STEL for benzene. By this time, physiologi- aL, 1~!74; Schnatter et aL, EJ9fi; Qu et aL, 2002, 2003; Yin et cally based pharmacnkinetic (PBPK) model.~ had advam:ed. aL, 19R7). to a point where they could do a reasonable job at predicting target tissue concentrations for a number ofreactive metabo- Even though additional compartments and metabolic pathways were added w the Knutsen et aL model, their con- lites for chen1icals such as methylene chloride, styrene, and clusions were similar to those offered by Bois and Paxman benzene, among others. These authors specifically evalu- (1992). That is, it appears that the area under the tissue con- ated the papers upon which OSlL'\ had decided that a STEL centration-time curve remains the same as long as the daily could be justified {Divine and Barron, EJ86; Picciano, lS79; cumulative exposure is less than about 40-80 ppm-hr. This Irons, 1983; Tice et aL, 1989; Tort et aL, 1982). Paxman and exposure would be equivalent, at 80 ppm-hr. to a employee Rappaport (1990) concluded that "in summary, we contend being exposed to 40 ppm for IS minutes up to 8 times per day, that, to justifythe setting ofa short-term exposure limit (STEL) assuming there was no other exposure to benzene. lt is not for benzene, OSHA incorrectly used the epidemiologic and surprising that physiologic damping occurs with benzene wxicologic data:' Later, noted that "the animal data do due to the effect of distribution and metabolism of occupa- indicate the possible importance of regularly spaced expo- tional exposures near the current OSHA PEL of l ppm (AUC sures of several hours duration, but this is not relevant to the of 8 ppm-hrs) and doses of 80 ppm-hrs. issue ofshorl-lenn exposures!' Here lhe authors focused on A sirnilar phenornenon would be exposure periods of 15-30 minutes or less, since they were the for lhose chemicals .,NHh moderaLe biological half-life and time periods ofm ost interest in the ACGIH guidelines and the moderate lipid solubility. \r'Je would also agree "'IVith Bois OSHA regulations. and Paxrnan's conclusion that: ~ ...the evidence was insuffi- Recently, Knutsen et aL (2008) extended the prior work of cient to implicate the rate of exposure as a causal factor in the Paxman and Rappaport (1990) and that ofother researchers. chronic health effects of benzene" (Bois and Paxman, 1992). They developed a kinetic model that indicated thai benzene "~./l!e acknowledge, however, that recent work by Rappaport and metabolite concentrations in the bone marrow resulting ct aL (2010) suggests that a different benzene metabolite from widely varying exposure profiles were vety similarwhen appears to be formed at very low doses, and that this moiety the cumulative workday exposures (dose) were limited to B may shown to be in the understand- ppm-hrs, which is equivalent to the current OSHA PEL at l ing lhe overalltoxicicy posed by benzene. ppm as an 8-hour TWA. For the scenarios studied, only minor .s changes in metabolite levels with or without CYP2El induc- Environmental exposures to benzene tion were observed at 8 pp1n-hr. clhese results are consistent IVIany studies have been conducted on hurnan exposure to with the physiologic damping effect reported by previous benzene in our environment. Benzene is present in the air investigators (Paxman and Rappaport, 1990). These authors of our cities due to the combustion of gasoline and diesel u:-;ed simple toxicokinetic methods to conclude that the fuel from our auton1ohiles and trucks, as well as fron1 the health hazard or potential for increased toxicity resulting from burning of fuel in our homes and exposure to cigarettes a dose ofeightppm-hours delivered over 15 minutes was not (either directly or secondhand). [t has been estimated appreciably different than the same dose delivered over eight that the average nonsmoker in the United States takes in hours. while lhe peak concemraHon is 32-fold greater in the around 200 micrograms of benzene per in the course former exposure scenario, both Paxman and Rappaport, as of performing their normal daily activities (vVallace, 1996a, well as Knutsen et al. (2008) concluded that the peak benzene l996b). More recently, ii has been suggested that environ- concentration in the liver should be within the capacity of the mental exposures are typically below 5D ug/m3 (Caplewn cytochrome P450 system to maintain first-order metabolism. and Levy, 2005). lhus, the various tissue doses, expressed as "areas under the ]he single greatest societal source of nonoccupational curve" (AUCs), the prefened dose metric for a chemical like ben7enc exposure ren1ains cigarette sn1oking, with snlok- benzene, would be the same for both exposure scenarios ers taking in approximately 40-80 Jlg with each cigarette. et aL, l9H7). Specifically, a As noted various researchers, for an agent like ben- "dose" of 0.8-L6mg of a gram), or about zene, in which the average blood concenlralion of the len limes Lhe daily intake of a nonsmoker (Wallace, 19~J6a, reactive chemical over time above a particular level in the l996b). It has been estimated that more than half of our net target organ is the most likely cause oftoxicity, the daily AUC societal human exposure to benzene occurs in the forn1 of CGU BEN0000295 Benzene and l1uman. health 39 cigarerte smoking (V\Tallace, 1996a, l996b; Capleton and traffic density to the incidence of childhood diseases failed Levy, 2005). Combining efforts to continue reducing smok- to find any to leukemia (Raaschou-Nielsen, ing with efforts to reduce hydrocarbon e1nissions frmn com-- 2001). Similar findings were reported by Langholz (2002) bustion sources would appear to be the most effective ways and Reynolds {2002). for an overview of various environ- to reduce the risk of benzene exposure to our at mental and genetic risk factors that have been evaluated Further research on the association between cigarette for childhood leukemias over the years, we recommend smoking and leukemia appears to indicale lhal the FAB sub- the review by BuHler et aL (2005). type M2 has been most tightly linked to disease risk, with subtypes M4 and M5 implicated to a lesser degree (Pogoda et al., 2002; Sandler et aL, 1993). A more recent case-control Conclusion study on smoking as a risk factor for 1\rfDS and AML reported. a "weak association between smoking and MDS, and for an association between smoking and abnormalities involving In summary, there is general agreement that benzene exposure, in high concentrations over many years, leads to an elevated risk for AML. .t\lthough there have been 1eports chromosomes 5, 7, and 8 in MDS and AML No apparent overall effect of smoking on the risk for AMI. was observed" (Bjork et 2009). The human health impact from benzene exposures at "environmental" concentrations has also been explored. of disease associations with other forms of leukem.ia, as well as with lymphomas and multiple myeloma, the overall evidence to date has been spotty and does not Indicate a consistent causal relationship. At lower doses of benzene, there may be a physiologic response, but studies performed "Overall, the evidence from human studies suggests that any risk of leukaemia to adults at general population continuous CATJOsure concentrations of 3.8 to 42 1-1g/m', which have been derived frmn available United Kingdom exposure data-that is, it is at concentrations three orders of magnitude less than the "'-L"'f-''"u'"' lowest observed effect level-is likely to be exceedingly small and probably not detectable with current methods" (Duarte-Davidson et al., 2001). to date haYe not sho-wn a convincing relationship between these responses and an increased risk of developing acute or chronic diseases. It has been historically acknowledged in toxicology and medicine that a biological response, in and of itself, is nut necessarily indicative of an increase health risk. Indeed, there is an entire body of science that has evaluated the hennetic response to low-level exposures, when~ the end result is a lesser incidence of adverse effects (Calabrese, 2008). It is dear that most significant exposures to benzene in Studies of traffic density in metropolitan areas have been the United States occurred in the workplace prior to 1950. as well as the public, have been exposed to lesser performed, with some claiming an association between high traffic density and the incidence of childhood cancers, concentrations of benzene since that time due to a better understanding of the hazards as well as increasingly strict particularly leukemia (Savitz et al., 1989; Nordlinder et al., l~l~l7; Knox et aL, 19~17; Harrison et al., 1999; Pearson et aL, :regulatory standards (especially posl.-1980). 1he increased appreciation of the potentially harmful health effects of 2000). These reports, however, are difficult to interpret, since levels of benzene and other potentially harmful chemicaLs arc not actually measured, and the childhood leukemias arc benzene over the past 50 years has drmnatically reduced the magnitude of workeT exposures here in the United States. .s often lumped together into a single disease category. One study that did provide a breakdown on the incidence ofvari- The lack of reports finding increased prevalence of disease in workers exposed after 1970 indicates that few workers in Western society are currently at increased risk. In developing ous leukemias in the studied population {Cmsignam, 2004) reported that acute lymphocytic leukemia, a disease not generally associated v.rith benzene exposure, was responsible for nations, where occupational and environmental controls are often less restrictive, cumulative exposures 1nay still be lead- ing to measurable benzene-related health risks {Ruchirawat, 79% of the leukemia&. Furthermore, it reported that "the low 2010; Bao et 2009). number of exposed cases with other leukemia subtypes (5 u:t~es) precluded further analysis by histotype" (Crusignani, Our knowledge regarding the impact of benzene on human healU1 continues to evolve. Hopefully, as we bet- 2004). ter understand at the molecular level how benzene and its Other studies have failed to confirm an association metabolites may relate to the process of carcinogenesis, we be1""veen traffic rlensily and incidence of childhood cancers. Tn a comprehensive California study looking at "IVill be in a better position to evaluate the risks associated with low-level, chronic exposure to doses of henzene (less cancers in children under five years of age over a l!J-year period, the authors concluded that: ~ .. no increased cancer risk among offspring of mmhers living in high traf- than 40 ppn1-years) with respect to the development of hematopoietic and other disease processes. fic density areas for all cancer sites or leukemia." In fact, the subgroup had an odds ratio for all cancers, and fot leukeinia in particular, of less than one Declaration ofinterest (Reynolds, 2004). ln addition, a Danish study that related The research supporting this review and the time needed benzene concentrations, nitrogen dioxide levels, and to vnite the article were funded solely by ChemRisk The CGU BEN0000296 40 D. Galbraith et aL opinions and perspectives are those of the authors, some of whom have served, and may continue w serve, on expert sci- entific panels that evaluate the potential health hazards posed Characterization of phenotype1 cytogenetics and outco1ne for 137 con- secutive cases_ Tnt J Hemaml11~:4:il-4:l7_ Renn ett T~1, (2005 ). A comparative revie't<V of classification systems in myel od- ysplastic syndromes (MDS). Semin Oncoi32:S3-IO. by benzene, and may act as expert V\'itnesses in litigation or while addressing governmental bodies. The authors '"ish to extend their deep thanks to Carrie Kahn for her keen techni- Bennett ]M, Catov,sky D, Daniel MT, Flandrin G, Galton DA, Gralnick !IR, Sultan C (1976). Proposab for Lhe classification of the acute leukaemias, Br T Haematol33:456l-458. Bennett Jl\[, Caluvsky D, Daniel 1v1T. (1982]. PnJpu~al foi the dasslfit.::ation of cal assistance during the preparation of this manuscript. rnyelody;;plastic syndnnnes. Br I Hae1natol5l:189-100. Bennett Jl\f, KonirokJi RS. (2005). 'Ihe Inyeludysplastic s:yndroines: Dic.g:nosi~. n'loleculaT biolvgy and risk assessment, Hematolog HJ:258-269. References Bergsagel DE) \\'ong 0, BergsagP-1 P, /\lF.xanian R, Anden.-;on K, Kyl-F RA1 Raabe GK. (l!l!l!l} Benzene and multiple myeloma: Appraisal ofrh~ sci- ACGTH {American Conference ofGovernmental Industrial Hygienists]. {IS7G]. entific evidence. Illood 94:1174-1 Ul2, TLVs-Tiueshold Lilnit \Talues for Chemical Subslances in \!Vorkroonl Bcrgsagel L, Kuehl vllM_ (20fJ~i).Molccularpathogenesis and a consequent clas- Air Adopted by ACGIH for l!J76. Cincinnati, OH: ACGIJI, lll-11; sificadnn of multiple myeloma. J Clin On col 23:-6333-6338. 36-37. Bernard SM, Cartwright RA, Bird CC, Richmds ID, Lauder L Roberts BK (l'lll4)- ACGIH (American Conference of Governmental Indus!rial Hygienists). (2001)- Aetiologic factors in lymphoid malignancies: A case-control epidemio- Docun1cniation nfTLVfor Benzene. Cincinnati, Ot-1: Af:G1H. logi-cal study. T.euk Res 8:681-689. Ak;;uyM. (1978). Ben~ene and leuk~rniu. Lance( 1:441. BczabehS, EngeL\, Morris CB, Lamm SH. {1996)- Docs benzene cause multiple Aksoy M. {19ll0a.). Malignancies dmc to ocr.npational nposure to benzPJW. mydoma> An all~ lysis of the published e<tse-connollireramre, Environ Haematologica 65:370- 373. Aksoy M. {l98ubJ Dif[eren! lypes of malignancies due to occupational expo- sure to benzene; A revie\-v of recent obseTvations in Turkey-. Environ Res Health Perspecr Suppll04:1393 1398. Binner R1 Sommerfeld G~ Krun1rey v: (2007). ReprnLs nn tflnrl ~afety, 200S. Bundesarnt fiir Verhrauchcrschutz und Lebensmittelsichcrheit {Federal 23:181-190. Offire oi Consurner Prolenion and Food Safety), Berlin, Gerrnany. Aksoy \1. (19R5). 1\.r'[alignartdPs due to orcupational exposure ro benzenf'. Arn Available at: http:/ /v.ww,bvLbund.de TInd Med 7:395-402. Bin] MG, Greim H, Kaden DA, Rice JM, Snyder K (2010]. Benzenc2009: Health Aksoy l\1, Dincol K, Erdem S, Dincol, G. {1972). Acute leukemia due to chronic effects and rnechanistns of bone rnarro\-vtoxidty: Irnpticatious for t-AML exposure to benzene. An1 TIvied 52:160-166. and the rnude of action fr aruewotk. Ch-eu;._ Biol Inteta-ct 184:3-6. Ak~oyM, Erdem S, Din col G. (1974). leukemia in sl:_oe-'\_<~Jorkers ~:x:posed chroni- Bjork T, fohansson B, Broberg K, Albin l'vL {2009), Smoking as a risk factor cally to benzene, Blood 44:837-84L for n1yclodysplastic syndro1nes and acute 1nycloid kukcn:..ia and its Aksoy M, Et"dem S. (19711]. Follow-up study on the mortality and the develop- relation to cytogenetic findlng~: A case-control study. Leukemia Res menI of leukemia in 44 pancytopenic patients with chroni,c exposure to 33; 788-791. benzene, Blood 52:285-292. Blair A, LinDs A, Stewart PA, Burmeister LF, Gib;;on R, Everett G, Shuman L, AksoyM, ETdem S, Dincol G, Kutlar A, Eakioglu l, Hepyuksel T. {l~Jil4)_ Clinical Cantor KP. (19S3). Evaluation of risks for non-Hodgkin's lymphmna by observations shovving the role of some factors in the etiology of multiple occupation and industry exposures frotn a case-control study_ A1n I Ind myeloma. A study in 7 patients. ,\eta Hacma!Dl 71:116-20. Mcd 2~1:301-31:2. ,\lexancler DTl, :Vfink PJ, Adami HO, Change ET, Cole P, Mandel JS, Blur.:kMf Jacobson LO, Rethard VVR (1953). Preleukemic acu1:e human leukemia TriduJyJou}os D. (2007). The non-I-Iodgkin lyntplun:nas: A revic'\\1' uf thf~ )i\MA 152:1018-1028. epldeminlogic literature. Tnt J Cancer 120: 1-J~. Rloemen LJ, You!< A, Bradley Tfl, Rodner KM, Marsh G" (2{](14} Americar_ CaCicer Society (ACSJ {2008aJ Cancer Reference Information: Lyrnph-ohaernatopol-etic cancer ri5k atnong chetniral \Vorkers exposed 10 Detailed Guide tu Acute Myeloid Leukeuiia. Available at: http:/)d(n:U- henzene_ Occup Fnvirnn T\'1P.rl n1 :270-274. ments,cancer.orgi6892,00/6892.00.pdf [Accessed june 1, 2010] Bloomfield CD. (1951)_ Studies of health hazards in industry: Chernica1hazards: American Cancer Socier; (ACS), (2008b). Cancer Reference Information: Benzoi vapor, lnd Health Month 11:62-64. Detailed Guidr- tc Chronic lymphocytic Leukemia, Available at: http:/ I Bofetla P, de Vocht E [2007). Occupa!ion and the risk of non-Hodgkin lym- docurnents.cancer.urg/6893.00/5B93,00.pdf phoma. Cancer Epidemiol Biomarkers Prev 16:369-372. AmericaE Cancer Society (ACS} (2008c), Cancer Reference Information: Bois FV, Paxman DG. (1992). AI1 analysis of exposure rate effects for benzene Detailed Guide to Non-Hodgkin Lymphoma. Available at: http://docu- u~ing a physiulugkally bai;etl pharmacukinetic modeL Reg Tux Phtum .s menrs.cance:r.org/1 78.00/178.00. pdf American Cancer Society {ACS), (2009). Cancer Reference Information: 15(2 Pt 1):122-136. Bollati V. Baccarclli A, Hou L, Bonzini :\!!, Fustinoni S, Cavallo D, Byun HM, Uetailed G<:ide to C\1ultiple Myeloma. Available at: http //document,_ cancer.org/175,00/l75,00.pdf Jiang }1 1\-IarineHi B, Pesm:ori AC, Bertazzi PA, Yang AS. {2007), Changes in DNA melhylation patterns in subjects exposed to lo"\>V-dose benzene. Ames BN, Gold LS. (2000), Paracelsus to parascience: 1he em~ronmental cancer Cancer Hes li7:!l7i:H:l80. distracti-on. I\futat Hes 4A7:3-lJ. Bond GG, :V!cLaren EA, Baldwin CL, Cook RR (1986). An update of mortality Andersen ME~ MacNaughton MGP Cle'l:\it:>ll HJft Paustenbach DJ_ {1987). Adjustlng exposuie limits for long and short exposure periods using a physiological an1.ong che1nical ;,vorkers exposed to ben:zene. Br J Ind l\1ed 43:685-6S l. Bosetti C, Levi P, r:crlay L Lucchini F, Negri E, La Vccchia C. (2.008). Incidence pharn:acokinetic modeL Am Ind Hyg Assoc J 48:335-343_ and nwrtality frorn non-Hodgkin lytnphotna ill Europe: The end of an API (Arnericau Petr()leutn Iustitut.P). (J 943). API T()xicnlogkal Revievv on epidemic? lnt f Cancer 123:1917-1923. Benzene. Ne\1vYork: API D-epartment of Safety. Brandt L_ {1992). Exposure to urganic solvents and risk of hen1atologic malig- API (American Petroleum Institute)_ (l96Q]. API Toxicologic&! Review on nancies. I.euk Res 16:67-70. BP-n7.~nP. 2nd P-d. 1\'"r:vvYork: APL Brief RS, Lynch), Bernath "l~ Scala RIL {1980)_ Benzene in the workplace, Axn ATSDR (Agency fur Toxic Substances und Disease Registrj), (2000), Case Studies lnd Hyg Assoc J 41:615-623. in Environmental Medicine (CSEM): Benzene Toxicity-Who is al Risk Brown EA, ShdlcyML, FishN]W, (19911). A pharmamkin<Ctic smdynfoccupa- Revised June 30, 2000. Available at: http:/ /www.ATSDRcdc.gov/cscm/ tiona] and environmenLa1 ben7ene exposure '".rith regard tu gender. Risk benzt:n.e/l.vhusat_risk.html Anall8:205-2l3, ATSDR{Agencyfor Toxic Subslances and Disease Registry). (2007). Toxicological Browning E. (l!J:J7). Heport 80: Toxicity of Industrial Organic Solvents: Profile fm Benzene. Atlanta, GA: Department of Hcalch and Human Summarie~ uf Publish~tl \.Vurlc Luru.lun: Committee un the Tuxicity of Services ..~vailable at: http://V>.'\O,'W,atsdccdc.gov/toxprotiles/tp3ohtml Industrial Solvents, Medical Researcll Council) Industrial HealU-:t Research ~1\ul C, Bowen DT, Yoshida Y. (19~'8). Recent advances in myelodysplaslic syn- Boa! d. dromes. Hacmatology 8'!:71-86. Brucken V. (1923). Uber chronjscher Benzolvergi1tung. Deuts-chem~d V\!chnschr Anstit1 I1~ Cole P. J\kCraw DS. (l ~86). A case-control :<>tudy ofleukernia. .ott an oil 49:ll20-ll2L refmeiy. I Occup Med 2B:llG9-1173. Buf:fle:r PA, K\Nan _\iL. (2005). bn'\r:ir-onrnental and geEetic risk factors for child- Austin H, Delzell E, Cole P {191l!l} Benzene and leukemia: A review of the lit- hood 1eukernia: Appraising the e\.idence. Cancer Invest } :60-75. erature and a risk assessrnenL Ant J Epide1niol l2'/:4l9-43Y. Bao L, Gross SA, Ryder J, Wang X, Ji M, Chen Y, Yang Y, Zhu S, Irons RD. Ca:abrese EJ, Stanek EJ, Nascarella 1\1/\, Halfmann GK (2UOll). Hormesis pre- dicts low-dose responses better than threshold models. lnt J Toxicol (2009} Adult precursor B lymphoblastic leukemia in Shanghai, China: 27:3G9-3 78. CGU BEN0000297 Benzene and l1uman. health 41 Caple"conAC, Levy LS. (2005J. An overvie\V of occupational benzene exposures Dreiher J, Kordysh E. (200b). Non-Hodgkin lyrnphoma and pesticide exposure: and occupatinr.al exposure limi" in Europe and North 1\merica_ Chem- 2"1 years of research_ Acra Haematolll fi: I CiJ-161. Rio Interact 153-154:43-53. Dreyfus B. (1976). Preleukemic states. Bloo.J Cells 2:33-35. Catenacci DVT, Schiller GJ. (2005]. Myelodysplastic syndromes: A comprehen- Duarte-Davidson R, Courage C, Rushton L, Levy L. (2001]. Benzene in the sive review. Blood Hev 19:301-:11!!. envirorunent: An assessment of the potential risks to the health vf the Checkoway H, \Nilcosky T, Wolf P, Tyroler H. {1984). An evaluation of the asso- population. Occup Environ Mecl 58:2-13. ciations of leukemia and rubber industry solvent exposures< Am J Ind ~Jed ECB (European Chemicals Bureau]. {2007). Benzene: Summary Risk 5:2:-!9-249. _,_l't~sessrn-ent Report: Final App1ov~U Veu_,iu.rL Prepared in the t.:untcxt of Cbn~ I\!lG1 Yardley-Jones A, MadeanAC, Dean BJ. (l~84). Ch::ornosorr~e analysis fron1 peripheral blood lytnpho-cj.tes ofw-urken, aft(:T an acute expusure to Council Regulatinn {EEC) No. 793/93 on Ute evaluation and control ot lhe ri~ks of e:xisling substances. Available ut: lntp://ecb.jic.ec.europa. benzene. Br] lnd Med 11:249-253. cuidocuments/Existing-( :hemicals/RISK_ASSESSMENT1S\TI\1MARYI Clinton, 1\f {1994). Video deposition of Marshall Clinton, \1.D., taken in b~nzenesuntOfi:l.pdf Lhe matter of: Bessie I\dae Duhon et al. u. i\Jobil Oil Corp.; Cathenne Ann Edwards SW, Preswn R]. (2UU8). Systems biology and mode of anion based risk Spruili et al. v. Aiobil Oil Corp., and Anita Selman et al. v. 1\Ia!Jil Oil Corp. assessment. Toxicol Sci 106;312-318. Civil Action Nos_ l:~l3CV4W:J, L:93CV~~42) and 1:93CV4b9. United States EFSA (European Food Safety Authority). (20fE>). EFS/\/WHO International District Court for the Eastern District uf Texas1 Beaumont Division. April 19, 1994. Cunference"'V'vith S'JpportofTLSI Europe on Risk assessment of Compounds lhat are Both Genotoxic and Carcinogenic. Bn..1ssels, Belgium. Available CodyRP, Strawderman VIlA!, Kip en l-IM. (l~J!B). Hcmat.J!ogic effects of benzene: at: http;//www.efsa.curopa.cu /cs/BiobServer/Sc' entific_Documcnt/ Job-specific trends during the first year of employment among a cohort of F.FSA_Gentux.pdflssbinary~trut' benzene-exposed rubher'\vorkcrs. I Occup l\led 35:776-782. Enlcrline PE. (1979). Lymphomas and benzene. Lancet II :1021. CoJJins JJ, Corner P, Friecllander BR, Eascerday PA, Nair HS, Braun J (l9~ l]. i'l FPA. {19'18]. Carcinogenic Effects ofB~nzene: An Up !law . ''Vasbingv:m, DC. US study ofthe hema:ologic effects of chronic low-level exposure to benzene. Environmental Protecr:ion Agency. J Occup Med 'l'l:Rl !'l-R2h. P.rfl ..\, Rhoads C.P_ {Eti~)- 'lhP hema!olngica1 effects ofhenzene (henzol) poi- Collins H, lreland B, Buckley CF, Sheppedy D. (20ll3]. Lyrnphohaematopoeitic soning. Tind Hyg Toxicol21:421 43!7. cancer rnorlalily anwng \ll!nrkers wilh 'benzene expo sur~. 0~-:cup Environ Evan' LS, Hancnck BW. {2003]. Non-Hm1gkin lymphoma. Lancet IVIed 60:676 -679. 362:139-146. Cunstdntini AS, Benvenuti A. Vinei;:, P, Kriebel D~ Tunllnu R, Rainazzutti \~ Fabbru-Peray P, Dame~ JP, Rusoi JF. (2001). Envirurmwntal cbk fadurs Rodella S) Stagnaro E1 Crosignani P; A1narlori D, Mirabelli D 1 Sm:nrnani L. Belletti I, Trosc.hel L, Romeo L, Mceli G, Tozzi GA. Mendico I, Maltoni for non-Hod.gkin's lyrnphorna: A population-based e.a.Bt>-cnntrol study in Languedoc-Ruussillon, France. CanceJ Causes Control 12: SA, lVIiliqi L (2003). Risk of leukemia and multiple myeloma aswci- 201-212. ated \1\:rith cxpnsure to brnzcnc and other organic solvents: Evidence Fagioli F, Cunf'o A~ Piva NJ Carli ~AG, Pre1.-"iad R, Balboni M, Ton1as1 P, Cariani D, frum the Italian multicenter case-contru] study. Am J Jnd J\Ied 51: Sc.apuli G, Cas1-oldi G. ( 1992). Distinct cytogenetic and dlniopathological ~.O:Hlll. features in acute myeloid leuken1ia after occupatiunal exposure to p-esti- Costantini S, Quinn M, O:msonni U, Zappa M. (2003]. Exposure to ben2ene and risk of leukemia among shoe factoryv,;rod{er.s. Scand J \Vork Environ cides and organic sdvents. Cancer 70:77-85. Farris GM, EverittJl, Irons TID, Popp JA. (1993). Carcinogenicity ofinhaled ben- Health 29:51-59_ zene in CBA n"Jice. Fundam Appl Toxicol 20:503-507. Cronkite EP. (18i11- Evidence for radiation and chemical> as leukemogenic farris GM, Robinson 51\', Wong BA, Wong VA, Hahn \NP, Shah H. (Ulfl7). agents. Arch F.nviron Health 3:297-303. Effects ufben?:ene un splenic. thymicy and femora11ymphoeytes in mice. Cronkite EP, Bullis ]E, Inoue 1~ Drew RT. :1984). Benzene inhalation produces Toxicology 118:137-148. leukernia in mice. Toxknl Appl Pharn1acnl 7?i:358-3fJ I_ Finkelstein 1\ifvL (?.OfJO)_ T.e1Jkemia <'titer e-xpO><;llrf:' to ben7.Pne: Temporal trEnds CrnnJ...:ite EP, Dre'\V RT:, Inoue T, Bullis JE. (1985). Benu~ne hetnatotoxirity and and itnplicari-on~ for standards. A1n J 1nd. Med 3.8:1 -7. leukemogenesis. .'\m J lnd Med 7:447-4.'ili. Fi,ch fR, Freedman SH_ {197.'i). Smoking, mal contraceptives, and ohesi1y Crosignani P1 Tittarelli A 1 Boi"gini A, Codazzi TJ Rcvclli A, Porro E1 Conticro P1 Bianchi N', Tagliabne G, Fissi R, Rossitto F. Berrinn F. (2004). Childl]{)od Effects on wrlitc blood cell count. JAMA 234:500 506. Fishbeck 'WI\, Townsend JC, Swank MG. (1978). Effects nf chronil' m:cupa- leukemia and mad traffic: A population-based case-comrol study. Intl J tional expot:.ure to measured concent.rations of benzen~. J Occup lv1ed Cancer 108:596-599. 20:539-542. Crump K, Allen B. (19\H). Quantitat1ve estimates of risk of leukemia frmrc occu- Flury F. (1923). U. Toxicities in modern industry. Ila. Pharmacological- pational expu;une to benzene. OSHA.tlucket H-059b, Exhibit 152 (appen- 1uxi-culugkal a~pe(..."t:.P uf intuxicanh in modern indusuy__luch Fxp Pathul .s dix B). vVashington, DC: OSHA. Cuneo/\, Fagioli F. Pazzii, Tallarico A, Previati H, Piva N, r:adiMG, Balboni M, Pharmakoll38:65-82. (German] Franceschi S, Scrraino D, Bidoli E, Talamini R. Tirelli U. Carhnne A, La t.:a-stold_i G. ( 1.992). Th,,forphologic, immunuloglc, and cytogenetic studies Vecchla C. (1989). The epid-emiology of non-Hodgldn's lymphoma in in acuLe myeloid leuke1nia follovving occupational expusure to pesticides the northeast of Italy: A hospital-based case-control study. Leuk Res and organic solvents. Leuk Hcs l6:7i:l9-795. n:4(i5--1n. De Go'\1\rin RL. (1963]. Benzene expusure and apla;stic anemia and leukemia 15 French JE, Saulnier \1. {2000). Benzene leukemogenesis: An Environmental years later. JA1\1A 185:748-751. DcCoufle P, Blattner WA, Blair A. (1983;. 1\lmtality among chemi- Carcinogen-indu-ced lissue-specific 1nodel of neoplasia using genetically ahcTcd mice. J Tox Envimn Health 61:377-379. cal v.rnrkers ex:posf'rl to henzPne and orher agents_ Environ R~s 30: Fritschi L, Benke G, HugbP,s Al\1, Kricker AJ Vajdic CM, Grulich A, Turner I, 16-25. Milliken S, Kaldor J, Armstrong BK. (2005). Risk of non-Hodgkin lym- DeLore P, Borgomano C_ (1923)_ Leukaemia aigue au COUTS de l'intoxication phoma associated vvith occupational e:xpos.uT-e to solv-ents, metals, organic benzeni(tne, sur Yorigine toxique de certains leukernies a ignes eT leur rela- dust.,;; <-:;nd PCBs (Australia). C,ancer Causes CnntrollG:S99-607. tions avec les anernies gnn:es. J Med Lyons 9:227-233. Fu H, Demers PA. Costantini AS, 'Wir:cteT P, Colir: D, Kogevinas; M. Boffetta P_ DescathaA, JenabianA, Conso F, Ame'lle f. (2005). Occupational expost:.res and (1996), Cancermcrtaliry among shoe rnanu:facturingwcrkers: An analysis haematnlogical1naH:gnanrif':s: Overvir,w on hun1an rf':rf'nt data. C.anr:f'!r of two cohorts. Oc.rup F.nvirnn T\1cd Fi:-1::1!14-:l~R. Causes Controll6:939-953. Gafafer \tVM. (1943]. Manual of Industrial Hygiene. United States Public Divine BJ, Barron V. (1985). Texaco morlalily sludy: II. Patterns of mor- Health Service, National Inslilules of Health. Phildadelphia, PA: WB tality among white males by specific job groups. Am J Ind Med 10: Saunders. 371-81. Gaffney SH, Pad<o JM, Unice KM, Bmns AM, Kreider ML, Gelatt RH, Dodsworth H, Dean A, Broom G. {1981]. Effects of smoking and the pill oJC the Booher LE, PauslenbachDJ_ (20-G9). Occupational exposure to benzene at blood count. Br J Hacmatol49:4114-.J.IIB. the E=onMobil Refinc1yin Baytown, TX (1978-2005]. J Expos Sci Em~ ron Dores G!VI1 Anderson V\/f: Curtis RR1 Landgren 0, Ostrournova E, Rluhrn EC, Rabkin CS, Devesa SS, Linet 1\iS. (2007). Chronic lyn1pho-cytic leul<aeinia and small lymphocytic lyrnpho1na: Uver\.1j_eV\' of the descriptive epiderniul- Epidemiol Advance online pubJication October 28, 2009; doi:l 0,1038/ jes.2009.53. Garrand R, GaTdias J, Bizet M, Bremond JLJ A-ccard B, Callat MP, de Bouchony El~ ogy. Br [ Ilaematcl2007;139:809-Bl9. Gn;u;guen JE_ {19:92). Heterogen-eity nf acquired idi[Jpathlc siderobh:tstic Dougherty D, Garte S, Harchowsky A, Zmuda L Taioli E. (:.!UUIJ). NQOl, MPO, anaemia {Aisa). Leuk Res lti:46Cl-4ti~. Cyp2El, GSTTl and GST:Vll polymorphisms and biological effects ofbcn- Gerardc HV\l. (1963]. The aromatic bydrocmbons. In: Patty I'A, cd. Industrial :zenP expnsure-A 11terature re"\riew_ Toxicoll.ett 1.82.:7-17_ HygiHte and Toxicology. VoL 2_ Ne-vvYork: Jntu-t \Vi1eyand Sons, S5S_ CGU BEN0000298 42 D. Galbraith et aL Girard R, Revol L< (1970). Lafrequence d)une exposition benzenique au -cours Harringtun 'TF. (1917). Industrial benzol poisoning :in 1\-fassachuselLs. Bo.slon des hemopathie~ graves. Nrmv Rev Fr Hematnl 10:-177-41'14. Med SurgJ 177:211:l-20fi. &lass DC, GrayCN, JolleyDJ, &ibbons C, Sim MR, Fritschi L. {2003]. Leukemia risk associated with low-level benzene exposure. Epidemiology Harrison R!Y[, Leung Pl,, Somerville L1 Smith R, Gilman E. (1999). Analysis of incidence of childhood cancer in the West Midlands ol the United 11:568-577. Kingdom in relation t0 proximity to main roads and petrol stations Occup Glass DC, S'm Iv:IR, Frits chi L, GmyCN, Iolley DJ, Gibbom C. {2004). Leukemia Environ Med 56:774-780. risk and relevant berzene exposure perlod-Re: Follnw-up tilne on risk Harl,-vig A. (2010). The role of DNA repair in benzene-induced carcinogenesis. e~tiniales. _/lun J Inti hied 45:222<"J. Chern Biul Interact 184:20S-272. Gla" DC, Gray CN, Jolley DJ, Gibbon> C, Sim MR. (2006). The Health Wat<:h Hayes RB, \''in SNJ Doserneci .1\1, Li GL, \Vacholder S) Travis LB, Li CV, case-control study oi leuketnia and benzene: 'lhe story so far. Ann NY Acad ci l 076:80-89. Rothman N, Hoover R\', Liner MS. (1997). Benzene and the doserelated incidence of hematologic neoplasn1s in China. J Nat Lane Inst Gogud A, Cavi-gneanx /\, Bt>lllo:nd J. (1967). T.t>s lt:ucelni~s heuzeniques. Hull Inst Natl Sante Rech !\led 22:421-442. 89:1065-1071. Hayes RB, Yin S, Rothman N, Dosemeci M, Li G, Trav1s :u; Smith MT, Linet MS. Geld LS, De Roos AJ, Dwwn EE, Lan Q, Milliken K, Davis S, Chancock S], (2000). I3enzene and lyniphohenl.atopai.etic malignancies in China. ITox Zhang Y, Severson H, Zahm SH. Zheng T, Hothman K, Baris D. (2UO'l). Environ Health Gl:4l9-432. As.sodarions of common variants. in genes involved in metabolism and Health Canada. (2009). Chrysotile Asbestos Consensus Statement and Summary: response to exogenous; chen1icals vvith risk of multiple tnyelon1a. Cancer Chrysotile Asbestos Expert Panel_ -Summary ofvie"\>Vs presented at HealLh Ilpidemiol 33:276-llll. Canada.conference in Montreal, Quebec, November 1'\-14, 2007. Goldstein RD. (l977J. Ren'i':ene Toxiciry: A critical eva1uariun. Repon for l-leffter A. (1915). Ueber die acuteVergiftung durch Flenoolclampf. Deutsch med lhe ..:\rnerican Petroleum InsLitule. \iVashingtonJ DC: Hemisphere \Vchnschr 41:1132-185. Publishing. Hemeon VvCL. (1955]. Plam and Process Ventilation. New York: lnttusrria! Goldstein BD. (1!338). Benzene toxiciry. Occup Med 3:541 554. Press. (;.-nldst~in RD. (l~GO). Ts FXpOSlHf' rn hf'n:z-P.n~ a cause nfhumann1ul1lplf': nlyf'- l-l"m<eml v\'\,L (l!'IR"'). Plant and Procte<> Vten!ilation. 2nd ted. N"w Ymk: loma? Ann NY Acad Sci 60!1225-230. Industrial Press. Goldstein BD. {2004). BPnzene exposure and lPukernia. Epidenlinlogy Henderson RE (1995). Spedes dilference.s in lhe rneiabulis.rn of benzene. 15:509-510. Environ HealthPerspect 10-1:1,173-U?S. Goldstein BD, Cudy R (:WOO). AoseS>mcnt uf <:urnplcte bluud <euunt variations Hollins, DM, Paustenbadr, DJ, Clmk, K, wlangult! CA. {2!lll9). A visual histmical .:::-:.n1nng workers exposed to lovv levels of benzene. I Occup Environ Med review of X1HJS1lre to asbestos at Puget St:mnd ~aval Shipyard {1 962-1972). 42:113-114. GGldwater LJ, Tewksbury MP. (1941). Recovery following exposme to benzene J Toxicol Environ Hedth B 12:124-156. Hooker AM, Bhat M, Day TK, Lane JM, Swinbur'le SJ, Morely AA, Sykes PJ. (llcnzol). J[nd Hyg 23:217-231. (2004). 'The linear no-threshold model does not hold for low-dos~ ionizing Graham JD, Green LC, Roberts \1]. (1991 ). ln Search of Safety: {;hemicals and radiation. Radiat Res 162:447-452. Cancer Risk. Can1_bridge, I\.fA: Harvard University Press. IARC. (1982). Son1.e industrial che1nical and dyestuffs: Den:zene. IARC.Monogr Greenburg G. (lY26). Benzol poisoning as an industrial hazard. Public Health Eval Carcinog Hisk Chern to Hum 2~l:~G-14ll. Rep 41 :l31G-lG49. Infante PE (2001). Benzene: An historical perspective on the .American a.nd Greenburg L, Mayers :rt.-1R, Gold\/llal-er L, Snllth. AR. (1939)_ Benzene {ber:...zol) European occupational setting_ In: Harremoes P, Gee D, f\1acGanrln poisoning in the rotogravure printing industry in :\'cw York City. J Ind M, Stirling t\, Keys J, \Vynne Il, Vaz SG, eds. Late lessons from Early Hyg Toxicol21:39:i-47-0" \r\Tarnings: The Precautionary Princip]e 1896-?000: Fnvlronrn.ental Issue Greer N, :-,1osscr G, Logan G, Halaas GW. (2000). A practical approach to evidence grading_ J Qual T111p1ove 26:700-71 ~J. R-eport No. 22. Copenhagen, Dcrunark: European Environmental Agcncy1 :18-~H. Greer fP, Rodgers CM,~ Foe:rsterT, Lukt:~ns }N1 Paraskevas ~Rodgers C\-1. (2{)04). \Vintmhe's f:linical Hematology. 111h ed. Philadelphia: LippincoH Infante PF. (2006). Ben"Zene expnsure and nn.tlti.ple rnyelotna: A deTailed rnetaanalysis ofhenzene cohort studies. Ann NY Acad Sci 107fi:~Hl-l 0~. \Villiams & v\'ilkins. Infante PP, Hinskv ItA, \'Vagoner ]K} Young RJ. (1977). Leukaen1ia in ben'Zene GnJss SA, Irons RD1 SchnattPr AR, RydPr J. \\TangXQ} Copley GR, ArrnsTnmg T''V. (2010)0 A hospital-based case conical sludy of aplastic anemia in Shanghai, China. Chern Biol Interact 1134:165-173. workers. Lancet 8023:76-78. International Fuel QuahlyCenter. (2008). International Gasoline Rankings -Top 100 Benzene. Houston, TX: Hart Energy Pubiishing. Available at: http:/I Gu.,ncl P, Imbernon E, Chevalier A, Crinquand-Calastnng A, Goldberg M. 1N""'v.ifqc.org/MiscellaneousContent.aspx'!Contentlrl=20 {20()2). Lt:ukemia in relation tu uccupatiunul expu~ures lo b-enzene unU hun;s RD. (1933). B~nzene: J\Jetabulism anU. .mechanism~. Paper pre.sented .s othr agents: A case-cuntrol study nested in a cohort of gas an-d electric utility workers. l\m I 1nd Med 42:87-97. at the International Con:terence on Benzene. Sponsored by Collegiun1. Ramazzini, NcwYmk ;'\lovcmher3and4, lgs:o. Gun RT, Pilutto L, Ryan P, Roder 0, Pratl N, Griffith E, McDermott B. {2000]. Iruns RD. {2000). :V1olecu1ar models of benzene leukernugenesis. JTux Environ Health \:Vatch Eleventh Reporl2000. Ad-elaide, Australia: Dep:a1irnent of HealUo 61:391-397. Public Health, Adelaide University. frons RU, Lv L, Gross SA, Ye X, Bao L, \Nang XQ, Hyder J, Armstrong T\111, Zhou Y, Gun RT, Ryan P, RoderD, PrattN, Griffith E, McDermott B. (2005). Healthv\latch Miao L, I.e AT, Kerzic PJ, Ni v\1, Fu H. (2005). Chronic exposure to benzene Twelfth Report 2005. Adelaide, Austraila; Department of Public Health, Adelaide University. results in a unique forn1 of dysplasia. LeukRes 29:1371-1380. Irons HD, Gmss SA, Le A, VJang XQ, \:hen Y, Hyder T, Schnattcr AH. (211H!). Gu7ellau PS, Vkturnff MS, HahnPs NC, Jarnes RC, GuzelL:J.n CP. (200?i)- Integrating \i\THO 2001-2008 criteria for the diagnosis ofrnyelody;;;;p1astir' Evidence-based toxicology: A comprehensive frame'\vork for causation. syndrome {MDS): A case--case analysis of benzene exposure. Chern Bioi Hum Exp Toxicol 21:Hil-20L Interact 184:30-38_ Ha:rnilton A. (1922). Tiu~ growin~ tnenace of benzene {benzol) poisoning in Jaffe ES, Harris NL, Stein H, Vardiman l\N. (2001). Parhology and Genetics American industry. JAJ\1A 73:627-630_ of Tumours of Haernatopoietic and Lymphoid Tissue~. VVoTld Hamilton A. (1929). lndustdal Poisons in the United States. New York: Health 0Tgan.ization Classification of Tun1()UTS. Lyon, France: MacMillan. 11\RCPness. Hamilton A. {1931 ]. Ben7ene [benzol) poisoning. Arch Pathol 11 :434-601. Johnson GT, Harbison SC, McCiuskey)D, Harbison RD. (2009). Characterization Hamilton A. (1934). Industrial Toxicology. 3rd ed. New York: Harper and of cancer risk fro In airborne benzene exposure. Regul Toxicol Pharmacal Bros. 55:36[-J(it;, Hamilton A. (1945). lntlu>uial Tuxkulugy. NewYurk: Oxford University Pres.. June SF, Fuko Z, Bull M, Kadin ME, Dorfman RF, Kaplan 1-lS, Rosenberg SA, Hamilton A, Hardy HL {1974). Industrial Toxicology. 3rd ed. Acton, :ViA: Kim H. (1973). Non-Hodgkin'slympholnas I\~ Clinicopathologic correla- Publishing Sciences Gwup. tion in 405 cases. C:anccr 3l:ll05-82J. Hard ell I., Erik5>on M, Lenner P, Lundgren E. (1981). Malignantlymphomaand Joshua DE. (2005). Multiple myeloma: 1he present and the future. Med I Aust exposuie to che1nicalsJ especially organic solvenLs, chlorcphenols and phe2:1.oxy acids: A case-control study. Br J Cancer t13:ll19-176. 183:344. Kalt <.TE (2000]. Utility cfa rncuse 1nodel foT study--ing the effects of benzene on Ilardell L. Eriksson M, Deg-ennan A. {1994). Exposure to phenoxya-cetic a-cidsJ the rnyeloid lineage: effects of hydn:Jquinone on a tnodd systetn. T Tox chlorophenols, or organic solvents in relation to histopathology, stage Environ Health lil:3g!J-4ll. and anatomical localization of non-Hodgkin's ly'1nphorna.Cancer Res S4:23R!i-2389. CGU BEN0000299 Benzene and l1uman. health 43 Kala I, Koenig KL, \'\!atanabe H. (2005]. Personal and occupalional exposure to LilcetMS, Stewart v\lF, Van NaHaML, McCatireyLD, Szklo M. (1987). Comparison mganic solvents and risk of non-Hodgkin's lymphoma (NHL] in women {United States). Cancer Causes Control 16:1215-1224. nf1nethods fnr determining occupational exposure in a caseMcontrnl inter~ view smdy of chronic lymphocytic leukemia. J Occup Med 29:136-141. Kawasaki Y, Hirabayashi Y, Kaneko T, Kodama Y, Matsushima Y, Ogawa Y, Lircet MS, Yin 5-N, Travis LE, Li CY, Zhang ZN, Li DG, Rothman Ci!, Li GL, Chow Saitoh ;\1, Sckita K, Uchida 0, Umemura T, Yoon BI, Inoue T. (2009]. "11\iH, Donaldocn J. Doscmcci M, Wacholder S, Blot WJ, Hayes HB. (1996). Benzene-induced hernatopoiecic neoplasms including myeloid Clinical l-eatures of hematopoietic malignancies and related disorders leukemia in Trp53-deficient C57BL/6 and C3HiHe mice. Toxicol Sci atnong benzene-exposed wo2:"kers in China. En\lir-on Health PerspecL 110:293-306. HJ4(Suppl6):135:1-l364. Khud-er SA, Youngdale MC, Bisesi I\1S; Schaub EA. {1999). Assesstnent ot ccnn- Li~t Af<~ Sandberg AA, Doll DC. {20()4). Myelod:ysplastk syndro1nes. Jn: Greer JP~ plete blood count vaTiatlons a1nong wo1l.er~ exposed to low levels ofben- Fnen,t-er 1. Lukens JN. Rodgers Glvl, Pr:uaskeya~ F, Glader B, erls. \VintTobe's 2enc.] Occup Em:iron Mcd 41:821-826. Clinical Hematology. llth cd. Philadelphia: Lippincott \Niiliams & KilnS, Veuneule11 R, ~v"{aidyall-<t.tltd S, fohnsnn BA, Lan Q, S1nith \-IT) Zbd.llg [J Wilkins, 2207-22'14. Li G, ShenM, Yin S, Rothman N, Happaport SM. (20Ut:i). Modeling human Loomis DP, Savit2 DA. (1"!Jl). Occuparion and leukemia mortality among men metabolis1n of benzene Iollo\"Ving occupational and environn1.ental expo- in 16 >Cates: 1985-191l7. Am find Med 19:509-521. sures. Cancer Epidetniol Biotnarkers Prev lS:2246-12::i2. Lukes RJ, Collins HD. (19'/4). Immunologic characterization of human malig- Kipen HM, Cody TIP, Goldstein BD. (1989). Use of longitudinal analysis of nant lymphomas. Cancer34:1488-l'J03. peripheral blood counts lo \.'alidaLe historical reconstructions of ben'Zene Lukes RI, Collins RD. (1975). Nevv apprnaches; to th_e classifi-cation of the lyrn- exposure. Environ Health Perspcct 3:2:1~JoJ-20fi. Kirkeleit J, Riise T, Rratveit M, Moen RF. {2008]. Increased risk of acute phomata. BI] Cancer 3l(Supp 11):1-:.:!l. Tynge F., Andersen A, Nil"on R, Roffetta P, Grand jean P. Heikkila P, Horte l.G, myelogen-ous leukemia unci multiple 1nyelnma in a historical cohnrt of up"ream petroleum workers expo sEct w cmcte oil. Cancer Causes Control Jakobsson R, Lundberg; I, MoenB, Parlancn 1~ Riise T (l997J. Risk of can- cer and e;(posure to gasoli.ne vapors. Am J Epictemiol l4Ei:41S-458. 19:13 23. l\lacDonald TA, Yeowell K, Rappaport SM. (1994). Comparison of pm- Knox F.G-, Gihnan F..A. (1~!i7). Hazard prnximi1iFs of childhnnrl cancf'TS in Great Britain ftom 1953--BO. J Epidcrniol Com1nunity Health 51: [~in adduct;; of hf'nz~rtP. nxidP and henznquinnnf' in thf': hln-nd and hone marrow of rats and n1icc exposed to benzene. Cancer Res 5'1: 151-159. 4907-4914. Knutsen J, Kuykendall J, Paustenbach DJ. {2008). Comparing equal deliv- ered Uutit':S uf airburru: bcnzt::ne fur 8 hr/day steady expu~urt! \iti~ peak Mad! AK, Paustenbach Df, (2002). Airborne concenl:rations of benzene due "o Uiesellucurnutive exhau::,t in a ruurulhuu!:le. J Tuxicol En\'iTon Health A exposur~ regilnens us.il::g a PB-PK rnndeL Presented aT the Atnerican 65:1945-1961. Indu&trial Hygiene ConfeHence and Expo (AIHce), May 31-Tune 5. ZOOS, rvlinneapolis, MN. Abstract #14. Related manuscript currently in journal Mallory TB, Gall EA. B:ridJ.eyVI'J. (1939). Chronic exposure tu benzene (benzol). III. The pathulogic resmts. J Ind Hyg Toxicol21:355-377. Tf'viCl.~. Malone KF, Koeps!"ll TD, Daling JR, Weiss NS, Mmris PD, Taylor 1\V, Swanson KoranyJ S. (1912). A leukaemia ke:zelese benzol1al. fTrtatment of leukemia 'il.:'lth benzene.] Budapest, !Iungary: Orvosi !Ietilap. GM1 Lyon JL. {1989). Chronic lymphocytic leukemia in TelatJun to <.:herni- cal exposure-S. Aln J Epidemiol l30.ll!:i2.-ll58. Kyle RA, Hajkumar SV. (2!l04). :Vlultiple myeloma. New Eng! J Med Mao Y, HuJ, UgnatAM, V\bite K. (201JO). Non-Hodgkn's lymphoma and occu- 351:1860-1873. pational exposure tu chemicals in Canada. Ann Oncol11:SG9-S73. Lamm SH, Engel A, Byrd 0:\1. (2005). Non-Hodgkin lymphoma and ben- 1\.--IcCord CP. (1931 ). Benzene_ In: Induslrial Hygiene for Engineers and zene c:xposurc: /\ systematic literature review. Chern Biol lnteract Managers. Ncw York: Harper and Bmth crs, [12. 153:?,3]-?37. McCormick DL, Kavet R. (7004). 1\nimal models for the study of childhuod Lamrn SH1 "\Valtcrs AS, V!ilson R, Byrd D~f) Grunvvald H. (198:1~1. Consistencies .:mil inconsistencies underlying the quantitative ass:essmpnt ofleuknni'ct lcukctnia: CDnsidcrrrticn-S for rnodcl identification and oplimizalion 10 idr:ntifypotential risk factors. TntJToxicol )3:149-fil risk fron1 ben2ene exposure. Environ Health Perspect 8:!:289- 297. Lan Q, Zhang L, Li c,, Vermeulen R, weinberg RS, Dosemeci M, Rappapml SM, l\.1cCrawDSJ Jovner RE, Cole P. (1985). Excess leukernia in a refinery population. J Occur Med 27:220-222. Shcn M, Alter BP, Wu Y, Kopp V'J, 'Naidyantha S, Rabkin C., Cuo \\7, Chancoi< Me Farland Wand DameshckvV. (1958). Biopsyofbonemarrow"'ith thevirn-" S, Hayes RH, Linet M, Kim S, Yin S, Rothman N, Smith MT. (2004). Hemat.otuxicity in vvorken; exposed to lo\v levels of benzene. Science silverman needle. lAMA 1116 (l2):!464-14G6. McHale Cl\1, Zhang L, Lan Q, Li G, HubbardAE, FouestMS, Vermeulen R. Chen 306:1174-1176. J, Shen M, Rappaport SJ\i, YinS, Smid1 MT, Rothman I\. (2009). Changes in Landgren 0, Kyle RA, Pfeiffer HM, Kat~mann JA, Caporaso :'-JE, Hayes RB, the peripheral blood transcripton1e associated with oc-cupational benzene Di~penzieri A, Kumar S, Clark RJ, Barib D, Hoov-er R and Rajkurndr V expu~ure itlentifit:U by cross-comparison un ~~vu micruanay platform~. .s (20D:9). 1\{onoclonal ga1nmopathy of undetermined significance {I\,IGUS) consistently pTecedes multiple myeloma: a prospective study. Blood Genon:tics 93:343-349. McMichae, AT, Spirtas E, Kupper LL. (1974]. An cridemiologic study of mortal- 113{22]:5412-5417. ity "''ithi n a cohort of rubber workers, 1964-72. [ Occup Med 16:458-464. Langhvlz B, Ebi KL, 'Hmmas DC, Peters f]Vl, London S!. (2002]. Traffic density McMichael AT, Spirtas R, Kupper LL, Gamble JE (1975). Solvent exposure and and the risk of childhood leukemia in a Los Angeles case-control study. leukemia among rubber workers: An epidemiologic study. J Occup Med Ann Epidemio\12:482-487. 17:234-239. Larson RA. (2000). Myeloid leukemia after cytotoxic therapy and oiher hema- McMichael AI, Spirtas R, Gamble IF; Tousey PM. (1975). Mortality among rubber totoxins. I TnxEnviron Ilealth 61:38l-3ilfj. workcTs-Relationship to specific jobs. I Occup Mcrl11l:178-1il!i. Laski'tl DL1 He de DF., Pl!njabi CL Laskill }D. (2000). Nitric Ox.ldt> .as a nu~dia lor of benzene induced hematosuppression and ~.:oxicity. I Tax Environ JHeda G. (1!122). Il heuzol~;..n1o profess.iouale. Lavnro 13:264-270_ Medinsky MA, Kenyon EM, Schlosser PM. (1995]. Benzene: A case study in par- Health 61:413-417. ent chemical and metabolit-e interactions. Toxicology 105:225-233. Le?;ge TM. (1920). Chroni< benzol poism,ing. J lnd Hyg 1:.>39-541. l\.1eek l\fE, K1aunlg TE. (201 0). Proposed tnode of actiou of benzene-induced Lehmann KB. {1919). Kune' Lehrbuch der Arbeits- und Gewerbehygiene. leukemia: Interpreting available data and identifying critical data gaps Leipzig: S. HirzeL for risk assessment. Chen1 Biol Interact 18-4.:279-285. T.f'noir A, C.laudf' L (1B~7)_ Dna casf' of purpura attrihuted TO hrn7en~ intnxir:a- Mehlman MA. (200fi). Causal rdationship hetwe,-,n chrnnir mydogenous cion. Bull Mem Soc Merl Hop Paris 3:1251-1261, leukemia and bt:n7.ene-containing solvents. Ann NY Acad Sd l 076: Leukemia and Lymphoma Society. (2009a). Disease Informalion: Chronic ll0-119. Lymphocytic Leukemia. Available at: http://;Nv,;w.leukemia-lymphoma. Midzenski MA, McDiarmid MA, Hod1man N, Kolodner K. (l\l>l2J Acute urg!al'_page?item_kl~ 7059 high tlu~e expusur~ tu benzene in ~hipyun.l worker;,. Am J lnd fv1ed Leukemia and Lymphoma Society. (2009b). Risk faclors for leuken:lla. Available 22:553-565. at: http:/ /wv>."1N.Ieukemia-lymphoma.org!all_pagc?itcm_id~702611risk Miligi L, Costantini /\S, Benvenuti A, Kiicbcl D. Bolejack V. Tumino R, Leukemia and Lymphoma Society. (2009c). Facts 200tl-2009, Available at: http:// Rarnaz'Zotti V. Rodella S, Stagnaro R, Crusignani P, Amadori D, 1\lirabelli D, '\V'\<\"\V.leukeinia-lyinpho1na< oig/a Haclunents /National/br_l2l5 783647. pdf So1n1nani L, Belletli I, Troshel LJ Ro1neo L !v1iceli G1 Tozzi GAJ Mendico I, Vineis P. {20(J6). Occupational exposure to solvents and che risk of lym- Lel!\ris RL Schnatter ARJ Drurn1nond I, l\furray NJ "Ihornpson FSJ Kat'l AMJ phnrnas. Epide1niology 17:552-561. Jorgensen G, Niwlich MJ, Dahlman U, Ihenaull G. (2003 ). Mortality and Mondrala S, bastmond DA. (2\JlU). Topmsornerase II inhibitiOn by the bioacli- cancer 1norbidity in a cohort of Canadian peuoleum workers. Occup vatcd benzene metabolite hydroquinonc invoh;es multiple rncchanisn1s. F.nvirou :Vied tl0:9l8-92R Chern Bini lnter;:ut l84:2S~-2fl!L CGU BEN0000300 44 D. Galbraith et aL Najean Y. (1981). Lon!':-term follow-up i:n patients with aplastic anemia: A sLUdy results of in vitro -culture of bone marnrvv and HLA typing. Blood 57: nf 137 androgen-treated patients surviving mnre than 1:'.1\1<1 years. Am J \!led 712-72~. 1:543-551. Peticti DB, Kipp H. (1986). The leukoeyte coum: Associations with inten- Natelson EA. (2007a). Benzene exposure and refractory sideroblastic erythro- poiesis: Is there an aosociation? ,\m J Mcd Sci 331:356-350. sity of smoking and persistence of e!Tect ai'ter quitting. Am TEpidemiol 12:1:89-BC.. Natelsun EA. {2007b). Ilen~ene-induced acute myeloid leukemia: A clinician's perspective Am J Hematol82:826-830. Picciano D (1979]. Cytugen-elic E.tudy of'vvorKen; exposed to benLene. Environ Res 19:33-38. Natiunal Re;search Cuuncil. (1981). Iruluoi Pollutants. \!\ra~hlngtun, DC: Puguda JM, Prcstun-Mmtin S, Nichols P,N, Rms RK. (2002). Smoking ant!Iisk af NaTional Aeadetny Press. acute tnyt?lold leuketnia: Results frurn a Los Angeles County ease-cnutrul NCI {National Cancer Institme). (2005). Surveillance Epidemiology and Eud study. Am I Epidemiol155:546-553. Results (SEER] database statistlcs for leukemia. i\vailablc at: http://seer. Price B, Ware A. (2004). Mesothelioma trends in the United States: An update ci1ltt.e1.gov/ based on su:rveill;tnce, t:"pidelniologyJ <tnd ~nd 1esults ptng:raln data fo1 NCI (National Cancer Instirme). (2UU!lj. \\lhatYou Need to Know about Muhiple 1~73 through 2003. Am J Epidemiology 15~: 1U7 -112. Myeloma. NIH Publication No. 08-1575. Bethesda, MD: US Department Pyatt D. {2004]. Benzene and heinatopoietic nwlignancies. Clin 0-ccup En-viron of Health and Human Services, National Institutes of Health. Available at: Med 4:~)2g-bs~). http:/ :'vl11\11.N.cancer.guv/cancertopics/v.:yntk/myeloma.pdf Pvatt D, Hays S. (201 O). A review of the potential association be1:ween childhood NCI (Natiunal Cancer Institute). (2007). ~\/\That You Need to Kno\V about Non- leukemia and benozene. Chen1. Bioll fn[eract 184:151-164. Hodgkin Lymphoma. ?>11H Publication No. ll7-l:Jfi7. Bethesda, MD: US Department of Health and Human Services, National institutes ofHealth. Qu Q, Cohen BS, Shore H, Chen LC, U G, J'n X, Meliian AA, Yin S, Yan H, Xu Ry 1.1 Y i'vfu R, Zhang X~ T.i K. {2003). R.en'len-e t''\'PU~tJre measurement in Available al: http:/ hA'\\1\>\'.cancer.gov/c.::..ncert-opics/1:\'-ylltk/non-hodgkin- sh-oe and glue manufacturing: J\ study to validate bio1narkcrs. Appl Occup lymphoma. pdf Environ Jlyg. 18:988-98. NIH (National Institutes of Health). (2009a]. Information on National Centers Qu Q, Shore R, I.i G, Jin X, Chen LC, Cohen B, Melikian AA, Easunond D, fnr SysLFnls; Riolngy. Availahle: at: http:; /Vll\TVJ_nigJTI'>. nih.gnv/Tnitimiv-Fs/ PermpnrtSM, Yin S, l,i H, VJaidyamha S, T.i Y, Mu R, fi K. (2002). SysBio/ Hernatologica1 changes among Chinese '\ATOI"kers .v:ith a range of NIH (1\";nional lHslilules nf Heallh]. (2009h). :V!edline Plus M;edical benzene ~xposurf's. An1 J fnd Med 42:275-28::.&. Encyclopedia: Chronic Lymphocytic Leukemia. Available at: http://wwvc. Queensland Health. (2007). lnveSligation of Chronic Lymphoid Leukemia nlm.NIH.guv/MEDLI:'<EPLUS/eru:y/mtide/000532.hnn Glacbtunc-Calliope, 1996-2004: Full technical repurl, August 2007. NIUSH (National [nsritute for OccupatiDnal Safdy and Ilealth.%Ee. {Ul79). Queensland, Australia: Queensland Gnvernrnent. Availablf' at: http:/I Revised Eeconunendation fo1 v..n OccLtp.ational Exposute ~tandard fot w1o\JVv.heahh.q1d.gov.au/ph/doclnnents/caphs/finalgladstune.pdf Benzene. Cincinnati, OH: US Department of Health, Education and Raabe GK, '..Vong 0. (1995). Leukemia mortality by cell type in petroleu::n \\idfac, Public n,,alth Service, Center for Disease Control, National vvorkc:rs '\Nith potential exposuH~ to brTtzenr~. Environ Health PcLspcn Institute for Occupational Safety and H-ea]th. 104:!381-1392. N-ordlinder RJ Janrhohn fl. (1997). Environnu:ntal exposure tu gasoline and Raaschou-:\'ielsen 0, Hertel 0, Thomsen DL, Olsen TIL (:~UOl]. Air pollution leukemia in children and young adu:ts-An "cnlogy study.lm Arch Occup from traffic at the rsidence of children "~th cancer. Am J Epidemiol Environ Health 70:57-GO. l 33:433-443. NTP (:\rational Toxicology Program). (2005]. Report on Carcinogens) Eleventh Rajkun1ar SV, KyleR_ (2005). J\1ultiple myeloma: Diagnosis and treatlnenc. 1\Jayo Edition. Hesearch Triangle Park: US Department of Health and Human Clin Proc 81J:l:J71-1:182. Sen.;icesy Public Health St"rvice. Rappapon: H. (1966), Ttimurs of the hematopoietic system. ln: Atlas of Tumor Nwosu VC, Kissling GE, Trcmpus CS, Honeycutt H, French TE. {2004). Exposure ofTg.AC transgenic mice to ben7.en r:' suppre;5:se.s hematopoietic. prngeni- Pathologv1 Section lll) Fasiclc 2. \rVashinglon; DC:.Armcd Forces. InstituL-c ofPathnlogy, 97-161 Lorrdls and alters g-ent~ expres.sinn in c:rit:cal :signaEng pat}-n.vays. Tuxicol Appl Pharmacal 1'1fi:'i7-4fi. Rappapon SM, Kiln S1 Lan Q, Li G, Verrneulen R, V\Taidyanatha S, Zhang L~ Yin S, Smith MT, Rnthman N. {20Hl). Human henzene melahnlism fol- Occupational Safety and Health Adrninistmtion. (1987). Occupational Exposure lo\'lring occupational and environn1ental exposures. Chern Biol Interact to Benzene. Final Rule. Ft~dera1 Register 52:34460-34573. ''VashingruuJ DC:: US Department of Labor. 184:!89-1%. Reynold> P, VonBehren J. Gunier RB, Goldberg DE, Hertz A, SmitL D. {2002). Ott MG, Teta MJ, Greenberg HL (1989). Lymphatic and hematopoietic tis- Traffic patterns and childhood cancer incidence rates in California, Uniced sue cancer in a chemical manufacturing environment. Am fInd Mcd States. Cancer Causes Control13:6G5-G73. 16:531-643. Reynuldo P. Vun Behn:n I. GunierREl, Goldberg DR. Hertz A. {2004) Re,idential .s Ott MG, Townsend JC, rishbeck V'lA, Langner RA. (1978). Mortality among 11\TOrkers nccupationally exposed to benzene. Arch Environ Health 33: e:xpo;~:ure to traffic in California and childhood cancer. Epidenriology 15:5-12. 3-Hl. Richardson DR. {2008). Temporal variatlu'"lln the association ben..veen ber7ene Panko JJ\1, Gaffn-ey SH1 Burns AM, Unice Kl\1, Kreider ::\{LJ Booher LE, Gelatt RH, Marshall JR, Paustenbach IJJ. {2009]. Occupational expnsure to benzene at the ExxonMobil refinery atRawn Rouge, Louisiana (1977-2005). J Uccup and leukemia rnortality. Envhon Heahh P-erspect 116:370-374. Higolin GM, Cuneo i\, Hoberti MG, Baidi A, Bigoni R, Piva N, Minottn C, }\gostini P, De Angeli<.:, Del Senno L, Spanedda R, Castaldi G. {1998). Environ Hyg 6:517-529. Parapia LA. (2007). Trepanning or trephines: a histo.I)! of bone rnarrn\N' biopsy. Exposure to myelotoxic agents ail.d myelodysplasia: Case-control study and correlation '\--v'ith dinicobif""!logical findings. Br J IIaexnatol Brit J I-Iae1natol 139: 14-19. 103:!89-197. PausLenhach DJ, Pdce PS, Ollison VV, Blank C, Jernigan TD, Bass D, Peterson HD. Rinsky RA, Young RJ, Smith AE. (1981). Lectkemia in benzene workers. Am J (1992). Reevaluation ofbenoenc exposure for the pliofilrn (rubberwmker) lnd Med 2:217-245. cohort (1936-Jg7Gj. l Toxicol Environ Health 36:177-231. PaListenbach DJ. (1995). Acute and chronic exposure IO benzene: Effects in humans. In: In1briani IVf] Ghittcri S, Pezzagno G, Capodaglio E, ed.s. Rinsky RA3 S1nith ABJ Horn1n1g R, Filloon TG, Young HI, Okun AH, I.andrigmt Pl {1987). Benzene and leukemia: An epidemiologic risk assessmenL New Eng! J 'Vied 316:1044-1050. Advanc.f's in Or:rupatinnal M~dirlne and RP-hahili:::ati-on: Updatf' on Ren7.ene. Vul. 1. Pavia, Ttaly: PI-ME Press, 69-91. Hlnsk--y HA, 'Hornung RVV, SilvP-r SH1 Ts.P-ng CY (2002). Ren7.f~n-f': P--xpo:\;urr. and hematopoieticmortal:ity: A long-term epidemiologic :-isk assessment. "~m. PaTITian D, Rappaport SM. (1990). Analysis of OSHA's shml-lerm-exposure limit J Ind evfed 42:474-480. for benzene. HegToxPharm 11:275-87. Ross IJ. {2000). The role of metabolism and specific metabolites in ben- Paxtun MB, ChinchilliVM, Brett SM, Rmlricks JV. (1994). Leukemia risk aoouci- zen~-inUu c~d tuxicity: Evidence anti i~sue~. J Toxicul Envirun H~ah:h A ated with benzene exposure in the PlioHlm cohort. II. Risk estirnates. Risk 61:357-372. Anal 14:155-161. Pearson R, \'\lachtel H) Ebi K. (2000). Distance-weighted rraffic densiry in prox- inTily to a ho1ne is a risk factor for leuke1nia and other childhood cancers. J Aii \Naste JVIanag Assoc. 200U;50:175-l80. Hmhman !\",Smith MT. Hayes HB, Li GL, Irons HD, Doscmcci '\1, Haas H. Stillman \MS, Linet M, Xi LQ1 Bechtold \1\TF:, \\Tiemeis J, Camplernan S, Zhang L, Quintana PL Titenko- Holland N, wang YZ, Lu Vv, Kolachana P, Meyer KB, Yin S. (19~6). An epidemiologic study of early biologic effects ofben2enc Pederson-HjeTgaard L Phillip P, 1\fnrten~en BT, Ersboll T, Tensen G, Panduro in Chinese iNorkers. Tinviron IIealth Pen;p-ect 104(Supp 6):1365-137{). J, Thomsen M. (EHil). Acute nonlymphocytic leukemia, preleuke- Ruchir-a'\vat IVlJ Navasumnt P, Sett-achan D. {20l1J). Exposure to benzene 1n vari- rrria, and acute nlyclopro1iferative syndrome secondary to trcauncnt of ous susceptible populations: Co-exposures to 1,3-butadience and PAHs othPr rnalignant disP,aS.e!':. Clinical and CJ--'i.ogeneti;: ehara{~t~ristics ~1nd anfl1rnp1icaiions for carcitwgenk risk Che:m Riol IntP,ti-JCT Hl4:67-71i. CGU BEN0000301 Benzene and l1uman. health 45 Ruiz JI.L'\, Augusto LGS, Vassallo J. (1994). Bone marmw morphology in patients ""~th neutropenia due to chronic expnsure tc> organic solvent (nenzene): SnyderR. (2D02). Benzene and leukemia. Cr:itRevToxicol32:l55-2lO. SongY, Du X, !loa, F. Gu X, Zhang Z, Zang S, Li r:, Li H, ME J. (2fll0). Early lessons. Path Res Pract 190:151-154. Immunosuppressive therapy of Cydospporing A for severe benzene- Runion HE. (1988]. Occupational exposures La potentially hazardous agents in induced haemalopoietic disorders and a 6-month follow up. Chem-Bio the petroleum industry. Occup Med State Art Rev :i:1:H-'J1"1. Interact 186:9!1-102 Runion HE, Scott IM. (1985]. Benzene exposure 'n the United State; l978-l983: Sorahan T, Kin len U. Doll R. (2005). Cancer risks in a hi;;turical UK cohort of An overview. Am TInd Med 7;385-393. benzene exposed workers. Occup Envlion :rvled 62:231-23. Hu~hton L. (199~1). A 39 year follu\v-up uf the UK uil refinery antl distribution Stcllnmn JW1, Daurn Srv1. (1973). \.Vurk is Dangerous HJ Your Health. New York: center studies: Results tor kidney cancer and 1eukernia. Environ Health Han do:tn House. Pnspect Suppll0l:77-84. Stokslad E. (2004). Factory sludy shows low levels ofbenzene redm;e blood cell Rushton L, Rornaniuk H. (1937J. A case-contr-ol study to investigate the Tisk counts. Science 306:1665. ufleukae1nia dS...,oddtt'd 'With exposure to benzene in petlu1etnn Jnalket- STI otn SSJ Cfu YJ G1uschkus SK. Pien~e SA) Estey El-L (2005). Risk fdctots ufinyel- ing and dis.tnbutiun workers in the United Kingdom. Occup Environ l\1ed 54;152-166. udy~pla.stic sy'ndromes: A case-control study. Leuk~rnia 1Y:lY12.-1YlS. Sunyer J. Munoz A, Peng Y, Margolick f. Chmiel JS, Oishi J. Kingsley L. Samet Sandler DP, Shore DL, J\nderson JH, Davey FR, Arthur D, Mayer RJ, Silver RT, J\1:. (199fi). Longitudinal relation between smoking and white blood cdk \Veiss RII, 1\!oore JO, Schiffer CA, Doris fl, Wurster- Hill 0, Mcintyre R, lim f Epidemioll44:734-74l. Bloomfield CD. (1993). Cigarette sYJ.1.oking and risk of acute l-euken1.ia: Sury-Bienz P. {lH-88). Gerichtlich-I\Ifedidnisches aus chen~ichen Fabriken. Associations with morphology and cytogenetic abnormalities k bone Vrtljschr f gerichtl Med 49:133-142. (As referenced in Hamilton, marmw. TNat] Cancer lnst 85:1994-2003. 1929]. SanLessoE CG. (1897). Uber chronischeVergiftungenmit SteinkohlenthcerbcnziJ:1. Swaen GM, Scheffers T, de Cock), Slangen T, Drooge H. (2005). Leukemia Arch Fur Hyg. ;; t::B6-376. Suthiakumm N, Delzell E, Cole P, Brill J, Frisch T, Spivey G. (1995). A case-con- risk in caprolaclam wor'<ers exposer! 10 benz.,ne. t\nn Fpid.,mio! 15:21 28. trol study nf lenkt"n1ia among petrnlf'unl '"\Nnrk~rs;. I Occup Envirnn 1\lferl S1v~rdln\v SHJ Can1po F, Harris Nl ,, Jaffe F.S, Pilt>rl .". 1A, Stf"in H, Thiele I, 37cl:l691277. Vardiman J\V. (:2008). WHO classification ofTumours ofHacmatopoietic Sa!hiakun1.ar N, Cirnff JJ Maldonado G, Matthews R1 Delzell F., l\Iaca~uso M. {2005). An updated study oi mortality among North American synthetic and Lyrnphoid Tl~suel',l. 4th ed. Geneva1 Swlizerland: VVI-IO Press. Thomas 'WA. (1982]. Supreme Court review of the OSHA benzene standard. 1ubbcr lnllwstry \'\'Uikt:Itl.. Ot:cup Envirun ::\tfeJ 62:822-?129. Am Stat 36:261-266. Satin KP, BailPy W), Newton f(L, Ross AY, \'\long 0. (2002). Updated epidemio- Thorpe Il (197~1). Epiderniolngir survey of 1eukernia in persons potentially logical study of 'Wotker~ at twu California pecrolewnrefineries, 1950-95. exposed to benzene. f Oucup Med 16:375-382. Occup En\liron :rvted 59:248-256. Tice RR, Luke CA, Dre\V RT, (1989). Effect o:f exposure route, regimen) and dura- Savilahti M. (1956). Mmc than lnO cases ofhcn7ol poisoning in a shoe factmy: tion on benzene-induced genoinxic and cytotoxic bone tnano"\'v da111age Ubservatinn un hematologic symptoms and the deve1upment of the dis- in mice. Environ Health Pen;pect 82:65-74. ease dLrring one year. Arch Gewerbepathol Gewerbehyg 15:147-1:37. Savitz DA, Andrews K\'\1. (!9;17]. Hcview of epidemiologic evidence on benzene and lymphatic and hematopoietic cancers. Am J Ind Med Tort K 1 Olofsson 1, TunekA., llerlin 1\f. (1982]. Toxic effects on n1.ouse bone marrow caused by inhalation of hen zen 2. Arch Toxicol 51 :2;15-302. Travis CC, Arms AD. {1987]. The food chain as a source of taxies expo- 3l:287-2'l5. sure. In~ LaYe LB and Upton AC, eds. Toxic 01emkals, Health] and the Savitz DJ\, Feingold L. (l~lHfJ). i\ssociation of childhood cancer with residential lraffic density. Scand J 'Nork Fnviron Health 15:360-363. EnvinJnJncnt. New York: Plcnunt Press~ 95-113 Tr<wis LB, 1 i, C-Y, Zhang "Z-N, I.i DG, Yin SN, ChowviTH, J.i GL, Dosemeci M, Scherr PA, Hutchison GB, N"ciman RS. (B92). "\on-Hodgkin's lvrnphoma and Blot \IT, Fraumcni JF, Hayes RB, Linct MS. (l'J::J4). Hematopoietic malig- ocn1pation:al eYposuTe. Cano:'T H.es S?.:!'i503s-S509s. nancies and rt"lated disordf'l'S among ben7,en r--e-xpnsPd 'Horkf'rs ir: China. Schnatter AR. (2004). Rerl'zene t~xposun:~ ;:nu:i le1.1kernia. F.pidernlolog)l Leuk L:yntphorna 14:91- 102. l.'i:'i!l~. Schnatcer All, Arn1strong T\'\!J Thompson LS1 Nicolich rvu, Katz Al\.-1, Huebner \V"\'11, Pearlman ED. {1996). Tit,; relationship between low-level ben:zene Tsai SP, VI.'en f:P, "Weiss NS. Wong 0, Mc\.lellan '\ITA, (',ih,nn RL. (l R!l'l). Retrospective mortality and medical surveillance studies of wod(CIS in benzene areas of refineries. I Occup Mf'd 25:58.5-592 exposure and leukemia in Canadian petroleum distribution vrorken:. Tsai SP, Fox EE, Ransdell JD, Wendt JK, Waddell LC, Donnelly RP. {2001). A Environ Health Perspect l01(Supp 6):1375-1379. hematology surveillance study of petrochemical workers exposed to ben- Schnatter AR, Nicholich MJ, Bird MG. (1996). Determination of leukemogenic zene. Regul Tmdcol Pharmacoi40:G7-73. benzene ~xpu~ure cuncentn:t1iun:.: Refint:U anulyse,s of the pliufi]m cohort. Tsai SP, Ahmed FS, \Vendt JK, Fu,ter DE, Donnelly RP, Stawmyer TR. {2007). .s Risk Anall6:833-840 . Sclmattcr AR Hosamilia K, Wojcik NC (2005). Review of the litcratmc on bcn- A 56 yea:r mortality follo1-\'-up oi Texas petroleun1 refinery <:::.nd che1nical employees, l!l4'l-2003. f Occup Em~nm Mcd 4!1:557-567. 7ene expusure and leukemia subt]'"fles. Lhern Blol Interact 153:9-21. US FDA (Uni<ed States Fuod and Drug Admmistration). (:!007). Data on ben- Scott BR. (2008). Low-dose radiation risk extrapolation fallacy ass.odaled \-Vith zene in suft_ d1-inks and other beverages. Silver Springs, l\10: Center Jor the linear-no-threshold model. Human Exp Toxicoi27:Hi3-l6!L Food Safety and Applied Nutrition. Available at: http://www.cfsan.fda. Selling L. (1910). A preliminary report of some cases ofpurpurahe_morrhagica gov/- dms/benzdata.html due to benzol poisoning. Bull fohns Hopkins Hasp 21:33-37. US FDA (United Slates Food and Drug Administration). (2009). Guidance for Sichcrman B 1 IIamilton~'\. (1984). Alice IIamilton: A_ Life in Letters. Cambridge, I'v1A: Harvard University Pr.P:ss. InduEtry: Evidence- based Hcvie'\1\.~ System for the Scientific Evaluation of Health Clairns_ CollPge Park, MD: CPnter for Food Safety and Applied Siddique ).1, Sabapathy K. (2006]. Trp53-depende:J.t DNA-repair is affected by Nu~rilion. Food and Drug Administration. Available at: hltp:/ /'\\T'iJ\TVl.cfsan. the codon 72 polymorphism. Oncogene 25:3489-3500. fda.gov/- dms/hclmgui5.html .Skarin AT, Dnrfrnan DM. (1997). Non- Hodgkin's lyrnphornas: Current classifica- Vardirnan fW, (201 0). Tiw \Vorld Health Organization (VVHO) classification tion and management. CA Cancer J Clin 4 7:351-372. of tumors of the hematopcietic and lymphvid tissues: An overvie'\v wilh Smedby KE1 Vajdic CM, Palste ]\.f, Engels E'A1 l\1artine'Z-:Vfaza 0, Turner L Hjalgrin1 T-T, Vinf'is P, Constantin! AS1 T-1nl1y F, Bracci 1'1 ltontan F, Spinelli J, Wimu T.D, Zheng 1~ Recker N, De SanjoM' S, Chiu BC-H, emphasis on the n1yeloid neoplasms. Chem Biol JnteTact 184:15-20. Vlanna NL Polan A_ (El7~)- Iy1nphon1a~ and ocr,upational hr:n7cnf' r,xposurr:. Lancet 6:1394-1395. Cocco Pp :t\.Iaynadie M, I~oretova L, Slaines A, Brennan P, Davis S, Vigliani EC? Forni A. {1975). Benzene and leuken1ia. Environ Res 11:122-127. Severson R, Ccrhan L Breen EC, Bilmann B, Grulich AE, Cozen Vv. Vigliani EC, Saita G. (l964j. Benzene and leukemia. New Eng! J Med {200-8). Allruimmune tli~on.lers anti _ri~k uf nun-Hutlgkin Jymphurna 27l:872-876. subtype~: A pooled analysis I.Nithin the I.3terLy.tnph Consortiun1. BlDod Von Oettingen \AlE (1940). Benzene. In: Public Health Bulletin No. 225: Toxicity l ll:4029-403!l. and Pmcntial Dangers of Aliphatic and Awmatic Hydrocarbons-A Smith M1~ Skibu1a CF; Allan f'V!, Morgan G). (20{]4). Call sal models of leukae- CriticaJ Revie\-v ot 1he Literature. Washingtun, DC: US Public Health nlla anrlly:tnphoina. In: Buifler P; Rice L Baan R} Bird M, Boffetta P, e:ds. .f\Aechanisms. ofCarcinogenesis: Connibutions. of Molecular .Epidemiology. Sen,ice, 66-97 . Vose JM, Lhiu BC-H, Cheson BD, Dancey), Wright J. (2002). Update on epi- IARC Scien[ifk Puhlicatious No. 157. Lyon: Internatinnal Ag-ency for derninlogy and therapeutics fur non-Hodgkin'~ lyrnphotna. IIetna.tology Research on Cancel~ 3l3-~Y2. 2002:241-:.!li:!. Smith MT, Jones IUv1, Smith AH. (2007). Benzene exposure and risk of non- Waddell ''Vl- (2006). Critique of dose response in carcinogenesis. Hun1an Exp Hodgkin lyrnpborna. Cancf':r Epidetnlol Bio1narkers Prev 1fi::U:E=i-3ql. Tox_itol 2S:4l3-43L1. CGU BEN0000302 46 D. Galbraith et aL \'Vallace LA. (l99ba). ,..ill~ exposure of the general population to benzene. Cell Bioi Tnxicol '1:2~7-:114. \Vallace LA. {1996b]. Environmental exposure to benzene: An update. Envimn HealU1 Perspect l04(Supp 5):1129-1136. Wang L, Zhou Y, Liang Y, Wong 0, Armstrong T, Schnatter AR, Wu Q, Fang), Ye X, Fu H, lmns RD. (2006). Benzene expusurein the shoemaking [nduslty in China, a literature survey, 1978-2004. Regul Toxicol Pharmacal 46:1-19-156. \'Vard l~~, Hornung RJ Morris J, Rinsky H, \VHd D, Halperin W, Guthrie \/V. {1996). Riskoflowretl or white bluud cell c.ount related tu esLhnated benzene exposure in a rub bcrworkcr cohort {1940-1975). _Ibn J Ind j\Ied 29:247-257. Weisburger EK. (2001). Histmy and background of the chreshold limit value committee of the American Conference of Governmental Industrial Hy,;ienists. Chem Health Safety 3:10-12. 1Neiss BM, Abadie J, Verma P, Howard RS, and Kuehl \VlVf {2009). A monodonal gammopathy precedes multiple myeloma in most patients. Bluod 113{22):5418-5422. 1NHO [World Health Organization]. {1993). Environmcmal Health Criteria 150: Ren7,ene. Geneva~ S'\.vit7erland: lnternatiunal Programme on Chemical Safety. vl'hysner J, Heddy I\!\~ I\oss PI\!, :V!ohan M, L<~X EA . (21J04) Genol0xkityofhen- zene und its rnetabo11tes. Mutat Res 566:99--130. VVijngaarden R, Ste1.Yan 'PA. (2011:=:;)_ C:ritical literalure revie\v nf de1ern1inants and levels of occupational benzene exposure for US community-based G:is~-c{)nlrol studies. Appl On: up Environ Hyg 13:678-693. Wilcosky TC, Checkoway H, Marshall EG, Tyroler HA. (198-1). Cancer mortal- ity iind !!ulvern expu:,ure~ in the rubber industry. Arn InU Hyg A~!luc J 45:809-81 I. VVllliain;s. PR, Pausteuba-ch DT. (:!D03). ReconstrLciitJn of benzene expusure fo1 the Pliofilm cohort (1936-1976) using Monte Carlo techniques. J Toxicol Environ FfcalthA66JJ77-78l. l.l\lilliams PR, Paustenbach Df. {2005). Ch<.tracterizing historical industrla1 hygiene data; A case study involving benzene exposures at a chen'lical manufacturing facility (1976-l ~l87]. J Occup Environ Hyg 2::l41-350. '\Vi!liams PR, Knutsen JS, Atkinson C, Mad! AK, Paustenbach DJ. (21J07). Airhome concentrations o benzene associated V\lith the historical use oJ some formnlatioEs of Liquid Wrench. J Occup Environ Hyg 4:517-Sfil. Williams PR, Panko JM, Unice K, Brown Jl., Paustenbach DJ. (2{)08). OccupaConal c:>.."Posurcs a.sscciatcd 1xith petroleum-derived products containing trace 1Pvels ofben7.ene. J Ocn1p F.nviron Hyg 5:5fjS-574 \'\Tilson R1 Crouch EA. (1987). Risk a:c;~e:ssrnents and -cornparisons: An introrluctinn. Science 2.i.fi:2fi7-270_ Wogan G:\!, Hecht SS, Felton JS, Conncy AH, Loch LA. (2004). Environmental and chernical carcinogenesis. St'tnin Cancer H:iol 14:473-486. \'liang 0. (191!7). An industry -.vide mortality study of chemical workers occupationally expn,cd to benzene: 11-Dose response analys" Br JJnd Mcd 44:382-395. \\Tong 0. (1999). A critique of the exposure ctssessment in the epidemiologic study nf benzene-exposed worker,; in China conducted by the Chinese Academy of Preventive Medidne and the US National Cancer lnbtitute. Regul Toxicol Pharmacol30:259-267. \\ltJng 0, Haabe GK. (1.995). Cell-t}!pe &pedfk lcukern.ia analy~e:;, in a curn- bined rohort of rnore than 20S,ODO petroleutn \-Vorken. in t!te United States and !he United Kingdom, 1937-1939. Regul Toxicol Phannacol 21:30 7-321. ,,Vong 0 1 Raabe GK. (2000). Non-Hodgkin'~ l~f111phmn~l and exposure to benzenf' in a multinational cohort of more than 308,UUU petroleum workers, 1937 to 1996. J Occup En>.:iron Med 42:5G4-56B IAiong 0, Harris F, Rosamilia K, Raabe GK. (20!Jl ). Updated mortality study of 1Vorker.s at a petrole1.nn refinery in Torrance, California, 1959 to 1997. T Occup Environ Med43:l039-ll02. Wong 0, Harris F, Armstrong TW, Hua f. (2lllfla). A hospital-based case-connol study of acute myeloid leukemia in Shanghai: Analysis of en"\ironmental and occupa!ional risk factors by WI'IO subtypes. Chern Biol Interact 184:112-128. Wcng 0, Harris F, Arm strong TW, Hua F. (201 Ob]. A hospital-bused case-control studyofnon-Horlgkin lytnphnirl ne-oplasms in Shanghai: Analysi,s of en\lironnlcntal and occupational Tisk factors by l..VHO subtypes. Chem Eiol Iut~rart 184:129-146. Yager JW, Easlmond DA, Robertson ML, Paradisin \'\IM, Smith MT. {1990). Characterizatiun uf rniuonudei induced. lu hu1nan lyrnphu-cytc~ by benzene rnetabolites. Cancer Rt>s 50:393-399. Yin SC\', Hayes RB, Linet MS, Li GL, Dusemeci :\1, Ttavis LB, Li CY, Zhang ZN. Li DGJ Cho'v 1/VH:, Wacholder S, Wang YZ) Jiang ZL, Dai TR) Zhang \VY, Chao XI, Ye PZ, Kou QR. Zbang XC, Lin XF, Mcng IF, Ding CY, :Zho IS, fllot VV]. (1996). A -cohort study among ben'Zene-eYpos.ed workers in China: Overall results. Aln 1 !nd Med 29;227-233. Yin SN, Li Q, Liu Y, Tian F, Du C, Jin C. (t;l87). !Jccupational exposure to ben- zene in China. Dr J lnd Me<l 44:192-195. Zarbl Jl, Gallo MA, Glick J, Yeung KY, Vouros P. (2010). rllie vanishing zero re"\isited: Thresholds ln the age of genomics. Chem Biol Interact IH4:273-27B. Zeeb H, Blettner M. {1.998). Adult leukaemia: what is the rule of currently l<llown risk factors? Radiat Environ Biophys 36:217-28. Zhang T1 1\.fcHal-e C:Vf 1 fiot11man N, Li G, Ji 7, Vermeulen R, H1Jhbard kE, Ren X, Shen M, Rappaport SM, North M, Skibola CF, Yin S, Vulpe C., C.hanock Sf, Smith MT, l.an Q. (2010). Sy.<tems biology of human benzene exposure: Toxicogenornic studies. Chc:rn-Bio Interact 18<1: 86-93. .s CGU BEN0000303