Document wKzVjJj7QOmjpOyjrXGnqZrod

TOYICOI.0C.Y 4 h D PH.\R!dACOLOGY 2 1 . 307-321 (1995) Cell-Type-Specific LeukemiaAnalyses in a Combined Cohort of -iu4 1 TO$/ `-n`p:%:'ii $1 pea- 2 I3. %4: More Than 208,000 Petroleum Workers in the United States and the United Kingdom, 1937-1989 OTTO %'ONG* AND GERHARDK. RXABEt `Applied Heaith Sciences. Inc , 181 Second Acenue, Suite 623. P 0 Box 2078. Son Mateo, California94401, acd +Medical Department. .kfobkl Oil Corporatron, P 0 Box 10.78. Prznceton. .Vew Jersey 08543 Received September 10.1994 - .A. .arge number of epidemiologic studies of workers .3:ir petroleum industry have been conducted to inves- : : g ~ 3 c r the carcinogenic and other health effects of .tl;osure to petroleum products during manufacture u c distribution. Of particular interest is the relation- ?ii? jetween exposure to benzene or benzene-contain:zi..r.quids in the petroleum industry and leukemia risk. 3,.. ?v?r, few studies have investigated cell-type-spe- ..-,_.__ :rukemia risk. In the present investigation, all co- io:: studies of petroleum workers in the United States uc :he United Kingdom were combined into a single :xsbase for cell-type-specific leukemia analysis. The zisiority of these workers were petroleum refinery em- r,io?-. but production, pipeline, and distribution w rkers in the petroleum industry were also included. E-combined cohort consisted of 208,741 petroleum *;. keys. Between 1935 and 1989, these workers con- :r: .lr*d a total of 4,665,361 person-years of observa- : i t - 7. >lore than 36 thousand deaths were reported Jrr.ong these workers during the 53 years of observa- ;.-i.3o- inn.g Cell-type-specific leukemia L\ meta-analysis procedure ris ap ks were propriat c e alculated for com- bining occupational cohort studies. These risks were esxessed in terms of cell-type-specific leukemia stan- d:. -aized mortality ratios (meta-SMRs).The meta-SMR - l c x t e myelogenous leukemia was 0.96. The lack of -1: :nL.rease of acute myelogenous leukemia was attrib- io the low levels of benzene exposure in the petro- ~ ~ X Z iIndustry, particularly in comparison to benzene ea?oaure levels in some previous studies of workers in o r h e r industries, who had been found to experience in- cye%sed risk of acute myelogenous leukemia. Similarly. II*.increase in chronic myelogenous, acute lymphocytic. I' :hronic lymphocytic leukemias was found in petro- . z! workers (meta-SMRs of 0.89, 1.16, and 0.84.re+<': ix-ely). ,Meta-analyses restricted to refinery stud- IC,. o r to studies with at least 15 years of follow-up YieLdcd similar results. The findings of the present in- \'*rigation are consistent with those from several re- -r'nt c'ase-control studies. s i99.i Academic Press. Inc. INTRODUCTION Workers in the petroleum industry are potentially exposed to hydrocarbons through inhalation of vapors from as well as dermal contact of crude oil, feed stocks, intermediate products during refining, and end products, such as gasoline (Runion, 1988; Raabe, 1993). Of special concern is benzene, which has been linked to an increased risk of leukemia, particularly acute myelogenous leukemia (AWL). The first case report suggesting an association between exposure to benzene and leukemia appeared in 1928 (Delore and Borgomano. 1928). In the following years, additional case reports were published (Hunter, 1939; Vigliani and Saita, 1964; Browning, 1965; Gouguel et af.,1967: Girard and Revol, 1970; Xksoy et al., 1972). Most case reports were from England. Italy, France, and Turkey. Documentation, if any a t all, of exposures to benzene or benzene-containing mixtures in these case repons was extremely poor. Furthermore. these reports lacked information concerning the size of population a t risk, and, therefore, no risk estimates could be derived. Thus, case reports by themselves do not provide definitive evidence for causation but rather suggest potential areas for future epidemiologic research. One observation, however, emerged from these case reports: The predominant cell type of the reported leukemia cases was AML. It was not until the 1970s that epidemiologic studies of individuals exposed to benzene were carried out. Judging by present day standards, some of the early studies were rather crude. For example, in the Thorpe (1976) study of leukemia incidence among employees at eight Esso petroleum affiliates in Europe, only active employees and annuitants, but not separated employees, were included. Furthermore, person-years of observation were estimated and not actually calculated. Similarly, in the Infante et al. (19'77) study of rubber hydrochloride workers exposed to benzene at two Pliofilm manufacturing plants in Ohio, the cohort was incomplete (an unknown number of employment records were missing and no maintenance workers were included). Furthermore, 307 0!2:3-2:!00/95 $6.00 Copyright 10 1995 bv Academic Press. Inc. All rights of reproduction tn any form reserved. 308 W O S G AND RAABE mortality ascertainment in the Infante et af. (1977) Both in vivo and in vitro studies have indicated that I ! study was only 75% complete. Likewise, employment committed stem (precursor or progenitor) cells are cap, histories were also incomplete. It appeared that some in- ble of self-renewal and amplification. Previously a nun.. dividuals included in the study had worked at the facili- ber of hypotheses have been proposed regarding the PO+- ties for only a few days. and some might not have worked sible role of stem cells in benzene leukemogenesis. Hoa ~ there at all (Wong et al., 1983). Subsequent updates of ever. in most of these hypotheses little or no distinction the Infante et af. (1977) study by others (Rinsky et af., was made between the pluripotential and the committed 1981,1987)have improved mortality ascertainment. but stem cells. Recent research has indicated that the C O I - the cohort incompleteness remained a major limitation mitted stem cells or progenitor cells play an imporr:: 1 of the study. part not only in the maintenance of normal. steady-st:. In a 1983 report. Wong discussed some of the unre- hemopoiesis but also in the etiology of cell-type-speci:lc solved issues regarding the relationship between ben- leukemias. zene and leukemia. The same issues were iterated by a Based on an investigation of immunoglobulin gene or- i group of scientists convened by the International ganization and expression in hemopoietic stem cell leu- Agency for Research on Cancer (McMichael, 19883. kemia, Ford et al. (1983)have demonstrated that a clor,ai One of the major issues raised by both Wong (1983) event leading to blast crisis can occur in a committed H and McMichael (1988) is the relationship, if any. cell precursor rather than in the pluripotential stem .-I\ j between benzene exposure and leukemia cell types other precursor. Similarly, based on human DNA anal? :h. 1than AML. Maher et af.(1993)concluded that CML and CLL do ::()I In the past, leukemia was considered a single statisti- arise from a common progenitor and that these maiig- cal category in most occupational epidemiologic studies, nancies arise from separate stem cells. Thus. recent ad- partly because of the historical nomenclature, unavail- vances in these areas confirm the epidemiologic o'oser- ability of cell-type-specific rates for comparison and, most importantly, the paucity of cases by cell-type in individual studies (Wong, 1987a,b). Recently epidemiologic studies have demonstrated the importance of cell- vation that "leukemia" is a g o u p of distinct malignar:. cies. which should be analyzed separately. The need to analyze leukemia data by cell type i r relation to occupational exposures such as benzene :;;I.. I 1 t-ype-specific analysis in studying leukemia. It has now been recognized that the diseases collectively known as leukemia are several distinct malignancies with different etiologic factors. Linet (1985) began her bcok on leukemia with the following: "Earlier epidemiologic reviews created some interesting challenges to epidemioio; .I . Workers in the petroleum industry represent one 01 iht. largest populations exposed to benzene. Numerous SIutiies based on petroleum workers have been conducted ir. the United States and in the United Kingdom. Eve:. I i considered leukemia as a single disease with some minor though most of these studies consist of several thousanc Idifferences in risk factors for each of the histopathologic workers who have been observed over several decade- types. In the last two decades, epidemiologic studies few individual studies offer adequate data for cell---.pr. have provided evidence that diseases collectively identi- specific analysis. The objective of the present in\.e -iga- fied as leukemia are several distinct malignancies characterized by varying age, sex, race, and ethnic group patterns. dissimilar secular trends, and different etiologic factors." Similarly, the diversity of different types of leukemia has long been recognized by hematologists, as in- tion is to combine all the studies of petroleum t w r k e r in the United States and in the United Kingdon: in:' a single large database, which can then be analyzed evaluate cell-type-specific leukemia risks in these De?rl' leum workers. i i ! dicated, for example. in the widely used textbook of he- matology by Wintrobe et al. (1981): "Few authorities would disagree concerning the recognition of chronic MATERIALS AND METHODS 1 myeloid leukemia (CML), acute myeloid leukemia Based on a consideration of both study desi- . 2:;. I(AML). polycythemia vera ( P V ) , idiopathic myelofi- data quality, only cohort studies were included .r. 7 : : . brosis (IMF), chronic lymphocytic leukemia (CLL), analysis. Studies based on proportional mortaliry 11 . acute lymphoblastic leukemia (ALL),Hodgkin's disease (PMRs) were excluded from analysis because o f i n c T i ' (HD), the non-Hodgkin's lymphomas (NHL), multiple lowing limitations. In addition to the well-knoun mr`: myeloma (MM), and macroglobulinemia as distinct en- odological deficiencies of P M R studies (Wong axe 1'. 1 tities." The question of whether all leukemia cell types can be affected by a common etiologic agent has also been examined at the cellular level, particularly with respect coufle, 1983;Wong et al., 1985).some of the P M R ?Jb:' of petroleum workers also suffered from incomi7 -.:? &:- certainment of deaths (Reeve et al., 1982; Thon:. - .: ,7 1980, 1982; Wong and Tabershaw, 1980).Fur:k!, xz"" I to hemopoietic stem cells. There are two types of stem employment histories were not available in tne:r PJ" cells: the pluripotential stem cells, which are responsible studies, and analyses were limited. Finally. t h e ?e:.`.., for the production of circulating elements, and their di- leum refineries included in these P M R studies hL4L.t --.- . rect descendants, the committed precursor cells which sequently been studied more thoroughly with t h e L'"' are programmed to follow a single line of differentiation. study design. d1 I - . 2-rroleum -- ?srroleum CELL-TYPE-SPECIFIC LEL'KEJlI.4 IN PETROLEUM WORKERS 309 TABLE 1 Description of Epidemiologic Study Cohorts in the Petroleum Industry in the United States and the United Kingdom Cohort location 10 retineries in U S Ei Segundo IC.\) refinery Richmond IC.\) refinery Port Arthur (T?Orefinery Baton Rouge I LA) refinery Baytown (TX)refinery Bayway ISJ)refinery Beaumont (TX)refinery Paulshoro INJ)refinery Torrance I CX) refinery 2 California refineries Deer Park rTS) refinery Woodriver I IL) refinery 13 refineries in us Production & pipeline in US Land-based terminals in US Marine vessels in US 8 rerineries in GK Distribution centers in UK US and LE;petroleum workers Number of workers 10.763 4,773 8.523 17.944 9.894 8.i22 6.860 7.119 4,Y.j.j 1,991 4,58.5 6,831 9,796 19,077 11,098 9.026 9.109 34,569 23.306 208.741 Observation period 1970- 1986 1950-1986 19.50-1986 1937 -1987 1970-1982 1970-1982 1970-1982 1945-1987 1946-1987 19S9-1987 19'73-1989 1948-1983 1940-1989 1947-1977 1946-1980 1946-1989 1946-1989 1951-1989 1951-1989 1937-1989 Person-years 125.241 110.594 205.397 526,396 93.783 80.684 59.257 166.427 120,715 33.331 57,657 181.782 300.991 358,318 220.414 239.125 227.143 931,640 626.471 4.665,361 Total deaths References 1.405 1.121 2,038 6,799 2.000 1.374 1.8% 2,194 1.681 408 1.051 1.180 3.62'7 4,024 1,386 2.066 2,695 10,193 8,713 -56,411 Hornstra, 1990: Nelson, 1985 Dagg et af.. 1992: W o n g et ol.. 1986 Satin et al.. 1994: Wen et ai.. 1983 Shallenberger et ai.. 1992: Hanis et al.. 1982. 1985a.b Raabe et al.. 1994: ,Milcarek et ai., 1994: Collingwood et al.. 1994; Morgan and Wong, 1984. 1985a; Enterline and Henderson, 1985 Tsai et al., 1993; Marsh et (11.. 1991: Honda et al., 1995: Wongsrichanalai et ai.. 1989: McCraw et al.. 1985: Joyner, 1983 Divine et al.. 1985 Divine and Barron. 1986, 1987 Wong et ai.. 1993; Wong and Trent, 1993 Rushton, 1993a.b.c: Rushton and Alderson, 1981,1983 o r t ;, ; 1: .idairion to cohort and PMR studies, there were ..-' .iaixber of community-based case-control studies -1h-pe-specificleukemias, which included petro- ...I ' br petroleum-related occupations in the analyses. '.si. - . o f these studies did not provide adequate infor- ':::it:on about the nature of exposure or details of -z?ioyment. Methodological problems (such as control -+lecrion, recall of employment histories, adjustment .!' 1.nnfounding factors) further complicated the inter- ;)re: xion. Therefore, case-control studies were not in- ,: xi in the quantitative analysis, but formed part of -i b. )asis of discussion. L couple of reports based on linkage between tumor .x::srry and census data have also been published iOlsen tr,d .Jenaen, 1987;dakobsson et al.. 1993).The Olsen and .!?risen (1987)survey reported a deficit of acute leukemia in men employed at gasoline stations in Denmark. '.vnereas the Jakobsson et af. (1993) survey reported an : .creased risk of acute myeloid leukemia in petrol sta- ) n attendants in Finland. Both surveys were judged to . inadequate. Using occupational information derived :om census data as a surrogate for exposure to specific :hemicals can introduce substantial misclassification. !eukemia cases in certain industries based on census data involves a potentially substantial amount of inac- curacy. The problem of multiple comparisons based on the large number of cancer sites and industries or occupations examined further weakens the statistical findings from these surveys. A s such, these reports are inadequate in assessing leukemia risk due to exposure to specific chemicals. In the present investigation, a meta-analysis procedure was used in combining data from individual cohort studies of petroleum workers. The methodological details of the meta-analysis have been described elsewhere (Wong and Raabe, 1989). Data needed to carry out the meta-analysis included observed leukemia deaths by cell type and the distribution of age-specific person-years from each individual study. These data were usually not presented in published reports. Instead, such data were requested from the original investigators. For some studies, multiple reports on various portions of the cohort and/or updates were available. For the present analysis, data for the entire cohort based on the latest available update were used. X specific analysis was carried out for each of the four major leukemia cell types (XML, CML. ALL, and CLL). The 8th Revision of the International Classification of Diseases was used in classifying deaths. The codes for AML, CNIL, ALL, and CLL are 304.0, 204.1,205.0, and 205.1, respectively. For comparison in the US studies, 310 WONG AND RAABE TABLE 2 Comparison of Leukemia Cell-Type Distribution between Petroleum Workers and the General Population I I Cell type Acute myelogenous leukemia us 103 ( 4 6 % j Petroleum workers: n (A) cIi 45 ( 4 5 % ) US and UK 148 (46%) c i_ 5% General popularin:: I43 us Chronic myelogenous leukemia 43 (195) 19 (19%) 62 (19%) 19 Acute lymphocytic leukemia Chronic lymphocytic leukemia 25 (11%) 5.5 (2-4F) 9 (9%) 2s (98%) 34 (10%) 83 (25%) 11 98 Total 296 (100%) 101 (100%) 32; (1005) 1100 Note. Goodness-of-fit test between the US-UK combined cohort of petroleum workers and the general population, xLr = 2.00 tP = 0.57). i age-specific mortality rates for each of the four major leukemia cell types were derived from data provided by the National Center for Health Statistics (Selvin et al., 1983). Although in general the accuracy of diagnosis of leukemia cell types based on death certificates was not as desirable as that based on pathological reports (particularly in the 1940s or 1950s), no potential bias was introduced in the present investigation, since diagnostic information in both petroleum workers and the comparison group (generalpopulation) was based on death certificates. Furthermore, the majority of leukemia deaths in petroleum workers occurred in or after the 1960s. For each US cohort, expected deaths by cell type were calculated by applying US specific rates to person-years. For the UK industrywide studies, cell-type-specific information has been published, and such information was taken directly from the publications (Rushton, i993a,b.c). One Canadian cohort study (Schnatter et aL, 1992) and one Australian prospective survey (Christie et al., 1991; Peach et al., 1992) were not included in the analysis because cell-type-specific leukemia rates for Canada or Australia were not available. The statistical procedure for meta-analysis consists of summing up the observed and expected deaths for a specific leukemia cell type from individual studies, and calculating the summary or meta-SMR. This simple procedure treats each individual cohort as a separate metastratum in data summarization, thus adjusting for individual studies while preserving the original adjustment using substrata specific to age. sex, race, and time period. In the present investigation, both cell-type-specific leukemia meta-SMRs and their corresponding 95% confidence intervals (95% CI) were calculated. Although other statistical procedures for meta-analysis are available, the meta-SMR procedure was used in this investigation because of its simplicity in computation and interpretation, its preservation of the adjustments in the individual studies, and its resembiance to the analyses in the original studies. Similar procedures have been used previously in analyzing data on asbestos and gastrointestinal cancer (Morgan et al., 1985).artificial sweeteners and bladder cancer (Morgan and Wong, 1983b). formaldehyde and respiratory cancer (Nelson et al., 19861,chemical dyes and bladder cancer (Matanoski and low, a total of 19 individual cohorts have been included in the database, which means that there would be 76 1 !? X 1)cell-type-specific SMRs. -4pproximately 4 stari-t i- us ' UIi UIi -US- RESULTS Included in the analysis were studies conducrei 0: sponsored by seven major petroleum companies ir, :h+ United States. Most of these studies consisted of reix. ery workers (Hornstra, 1990; Yelson, 1985; Daauc- e: a. 199: 198. 198' WOO. terii I d, 198L 198: base Di' Sore, sistt A- I B' CD* E- FG- H- IJ- K; LM- Nb 0' PQ- A :Combined (US. S- ICombClnoemdb(Uln.Se.dsiuU:KK.]. 0.1 -4 44 I1 4 I 1 7 v II P I 1 I 4 I --7 -l+ -c-c( Ccl 1 S M a r r d U d M a W n , FMOula %% c' FIG. I. Acute myelogenous leukemia among petroleum '' -' -in the United States and the United Kingdom. 193'7-1984 4 studies: R-S.UK studies. I ad %l< 1- I CELL-TYPE-SPECIFIC LEUKEMIA I?u' PETROLEUM WORKERS 311 TABLE 3 - Meta-analysis of Acute iMyelogenous Leukemia in Petroleum Workers in the United States and the United Kingdom, 1937-1989 Country conort Observed deaths Expected deaths SMR 95% CI no1 - -1- - aY :hc f0l 2e A3 B6 C6 D0 E4 F4 G 14 H, IJ J6 K 10 L4 ?uI 8 ?j 0 01 P 12 Q 13 4.37 6.33 4.48 3.37 4.45 "80 10.90 6.91 7.88 3.56 15.95 3.65 6.15 1.23 423 12.11 3.91 0.62 0.95 1.34 0 0.90 1.43 1.28 1.01 0.63 1.69 0.63 1.10 1.30 0 0.24 0.99 1.4s 0.13- 1.31 0.35-2.07 0.49-2.92 0-1.09 0.24-2.30 0.39-3.66 0.70-2.19 0.41-2.08 0.20-1.4; 0.62-3.68 0.30-1.16 0.30-2.82 0.56-2.56 0-3.00 0.01-1.34 0.51-1.73 0.79-2.53 1e: US combined cohort ?X ?e. UK iec 1s . < 2ombined cohort R s 103 20 "5 4.5 107.68 26.60 '20.60 47.20 0.96 0.78- 1.16 0.75 0.45-1.15 1.21 0.78- 1.79 0.95 0.70-1.27 ,ti- S-LK combined cohort 314 m- 148 154.88 0.96 0.81-1.13 *its 21y -=of nd 1992: Wong et al., 1986: Satin et al., 1994; Wen et ai.. 1983; Shallenberger et al.. 1992; Hanis et al., 1982, 1985a,b;Raabe et ai.,1994; Milcarek et al., 1994: Colling1.vood et al.. 1994; Morgan and W-ong, 1984, 1985a; En- SMR for the Beaumont refinery has been reduced from 1.73 (P 0.05) to 1.39 (P> 0.05), for the Deer Park refinery from 2.31 (P< 0.05)to 0.96 (P> 0.051, and for the Woodriver refinery from 2.12 (P < 0.03) to 1.23 (P - .r!ine and Henderson, 1983; Marsh et al., 1991: Tsai et > 0.05). Thus, all three significantly elevated leukemia :.. 1993: Honda et al., 1993: Wongsrichanalai et al., SMRs reported in the 1989 review are no longer statisti- -989; McCraw et al., 1983; Joyner, 1983; Divine et al., cally significant. Furthermore, the updated Amoco study or LY83; Divine and Barron. 1986), except one which was of 10 US refineries showed a significantly low total leuhe based on workers in the production and pipeline division kemia SMR of 0.39. Three new cohorts have been in-n- (Divine and Barron, 1987). A study, which was spon- cluded in the present investigation, and the total leuke11.. sored by the American Petroleum Institute and con- mia SlMRs for all three cohorts were less than 1.0 (Tsai Gisted of a cohort of land-based petroleum distribution et al., 1993; Wong et al., 1993; Wong and Trent, 1993). vorkers and a cohort of marine petroleum distribution T h e updated total leukemia meta-SMR based on all pe- - vorkers. was also included (Wonget ai., 1993: Wong and troleum cohorts in the United States and the United ?rent, 1993). From the United Kingdom, two cohort Kingdom has been recalculated to be 1.02 (198observed ;tudies of petroleum refinery and distribution workers deaths, 95% CI: 0.93-1.11). 3ponsored by the Institute of Petroleum were included Table 1 provides a basic description of the individual 1 Rushton. 1993a,b,c: Rushton and Alderson, 1981. US and UK cohort studies included in the present inves- 198.3). tigation. Although exposure patterns in each cohort Previously Wong and Raabe (1989) provided a meta- might differ, all workers were potentially exposed to ben- Analysis of leukemia (all cell types combined) based on zene-containing petroleum liquids as well as other hy- Vtudies of petroleum workers available a t the time. The drocarbons. It is quite likely that the "within-cohort'' om1 leukemia meta-SMR was 1.12 (not significant) exposure variation is comparable to, if not greater than, msed on 279 deaths. Of the 14 US or UK studies re- the "between-cohort'' exposure variation. ;.iewed by Wong and Raabe (1989), 3 refineries (Beau- In the present investigation, a total of` 19 individual - mont, Deer Park and Woodriver) showed a significant cohorts were included in the meta-analyses. The com10 increase in total leukemia. Since the Wong and Raabe bined cohort consisted of more than 308,000 petroleum I 1989) review, all but two studies in the United States workers, with approximately 150,000 from the United tnd in the United Kingdom have been updated. In par- States and S8,OOO from the United Kingdom. Over an icular, based on the updated data, the totaI leukemia observation period of 53 years (1937-~989),these work- 312 WONG AND RAABE TABLE 4 Meta-analysis of Chronic Myelogenous Leukemia in Petroleum Workers in the United States and the United Kingdom. 1937-1989 Country Cohort Observed deaths Expected deaths SMR 9.55 CI I!S US combined cohort C'K UK combined cohort US-UK combined cohort A E C D E F G H I J I.; L M N 0 P Q R S 0 -3 3 2 1 0 8 3 4 1 6 4 3 2 1 8 1 43 11 8 19 62 2.06 8.85 1.9; 1.55 1.98 1.26 4.85 3.08 3.52 1.60 7.11 1.58 2.74 0.54 1.92 5.40 3.95 47.97 12.30 9.70 82.00 69.97 0 0.70 1.52 1.29 0.51 0 0.41 0.97 1.14 0.62 0.84 2.53 1.09 3.68 0.52 1.48 0.25 0.90 0.89 0.82 0.86 0.89 0-1.79 0.08-2.53 0.31-4.41 0.16-4.66 0.01-2.54 0-2.9.7 0.05-1.4s 0.80-2s 0.31-2.92 0.02-3.45 0.3 1-1.S3 0.69-6.46 0.22-3.15 0.54-13.2. 0.01-2.9d' 0.61-2.92 0.01-1.39 0.65-1.01 0.44-1.59 0.35-1.62 0.52-1.35 0.68-1.1 =, e: I I "Ia a ers contributed more than 4.6 million person-years. More than 56 thousand deaths were reported in these studies. In particular, 498 deaths were ascribed to leukemia. Based on information available, 327 (66%) leukemia deaths among the petroleum workers were classifiable into one of the four major cell types: AML, CML, ALL, and CLL. Thus, 34% leukemia deaths in petroleum workers belonged to other or unspecified cell types. According to data provided by the National Center for Health Statistics (1990), the corresponding figure in the general population was similar (32-34% 1. The distributions of leukemia deaths among the US and the UK petroleum workers by the four major celltypes were very similar (Table 5 ) .The cell-type distribution for the US-UK combined cohort of petroleum workers was also similar to that in the general adult population (aged 13or older),which was derived from Selvin et al. (1983). A formal goodness-of-fit test between the US-UK combined cohort of petroleum workers and the general population indicated that there was no signifi- cant difference between the two distributions ( x i d f = 2.00, P = 0.37). Table 3 and Fig. 1present the results of AML analysis for individual cohorts as well as for the combined cohorts. Because of the logarithmic scale, for SMRs less than 0.1, neither the point estimate nor the lower 95% confidence limit is presented in the figure. The number of observed AML deaths in each cohort ranged from 0 to 25. No significantly elevated AML SMR was detected for any cohort. For Cohort D, no AML deaths were ob- i served, and, based on 3.37 expected, this deficit was almost statistically significant (SMR = 0,95% CI: 0-1.09. P = 0.06). For the US combined cohort, 103AML deaths were reported, compared to 107.68 expected. The corresponding ,4ML meta-SMR for the US combined cohort was 0.96, with a 95% CI of 0.'78-1.16. The UK combined cohort yielded similar results for -4ML (meta-SMR = 0.95, 95% CI: 0.70-1.2;). For the US-UK combined cc hort, there were 118 observed AML deaths, which were comparable to the 154.88 expected, and the meta-SMR for AML was 0.96 (95%CI: 0.81-1.13). Results based on a similar analysis for CML are shown in Table 4 and in Fig. 5. In the United States. most cohorts contributed fewer than 5 CML dearhi each, and the highest number of 8 observed CML dear 3 was reported in Cohort P. N o significant CML SMR w 5 found in any of the individual cohorts. For the US CO:Xbined cohort, 43 observed deaths were ascribed to C I I L . whereas 47.9'7 CML deaths were expected. For the L-5 combined cohort, the CML meta-SMR was 0.90, and ~ t l e 95% CI was 0.65-1.21. For the UK combined cohort. t h e I meta-SMR for CML was 0.86 (19 observed vs 23.00 espected, 95% CI: 0.52-1.35). For the US-UK combir 4 ' cohort, a total of 62 CML deaths were observed, cc ~ pared t o 69.9'7 expected. The CML meta-SMR for a!: c - troleum workers was 0.89 (95%CI: 0.68-1.15). ALL accounted for only 10 to 11%of the 3 major cell types combined in both the petroleum workers and the general population (Table 3). The number of .ALL deaths in each individual study was extremely small. scl tr c C u u: -c'- CELL-TYPE-SPECIFIC LEUKEJIIA IN PETROLEUM WORKERS 313 0.1 1 SBndanmar t4umty ROD0 and 95% CI 10 5 IC;. 2. Chronic myelogenous leukemia among petroleum work- ~. n :he United States and the United Kingdom. 1937-1989. A-Q, . .:dies; R-S, UK studies. Dore than 2 ALL deaths were reported in most cohorts, and 6 cohorts reported no ALL deaths a t a11 (Table 3 And Fig. 3). In Cohort K, however, 8 ALL deaths were g,ixerved,compared to 3.08 expected. The ALL SMR for t->horr K was 2.60 (95%CI:1.13-5.11). No other cohort : 2orted a significant ALL SMR. In the US combined . ilort, 25 ALL deaths were reported, slightly higher :nan the 20.36 expected. The ALL meta-SMR for the L-S combined cohort was 1.23, which was not significant (96% CI: 0.79-1.81). The ALL meta-SMR for the U K combined cohort was slightly less (meta-SMR = 0.99, 96% CI: 0.45-1.88). In the overall combined cohort of both US and UK petroleum workers, the observed number of ALL deaths was 34, whereas the expected number was '79.36: and the ALL meta-SMR was 1.16 (95% CI: 0.81-1.61). Table 6 and Fig. 4 present the results of the CLL anal- ysis. In the United States, most individual cohorts reported 4 or fewer CLL deaths. Of the 17 US cohorts, 13 reported a lower than expected CLL mortality. In fact, Cohort K reDorted a significantly low CLL ShIR of 0.30 (3 observed-vs 10.04 e>pected. 95% CI: 0.06-0.87). On the other hand, Cohort C reported a significantly high CLL SMR of 5.59 (9 observed vs 3.48 expected. 95% CI: 1.18-4.91). X total of 55 CLL deaths were observed in the US combined cohort, almost significantly Iess than the 69.19 expected (meta-SiMR = 0.79, 95% CI: 0.591.03). There were 98 observed CLL deaths in the UK combined cohort, slightly less than the 39.70 expected (meca-SMR = 0.94, 95% CI: 0.63-1.36). In the US-UK combined cohort, 83 CLL deaths were observed, whereas 98.89 were expected; and the CLL meta-SMR was 0.84 (95% CI: 0.67-1.04). The above meta-analyses were based on all 19 cohorts of petroleum workers, including refinery, production and pipeline, and distribution workers. If the meta-analyses were restricted to the 15 refinery cohorts (14from the Country TABLE 5 Meta-analysis of Acute Lymphocytic Leukemia in Petroleum Workers in the United States and the United Kingdom, 1937-1989 Cohon Observed deaths Expected deaths SMR (..S '5combined cohort ( .K 1 .K combined cohort [..SUI.; combined cohort A 8 C D E F G H I *J K L ?/I N 0 P Q R S 0 1 1 0 1 0 3 1 -1 7 8 0 0 0 2 3-> 25 4 J 9 34 0.84 1.22 0.89 0.70 0.83 0.54 2.02 1.32 1.47 0.68 3.08 0.69 1.17 0.23 0.82 2.14 1.62 20.26 5.:30 3.30 9.10 29.36 0 0.82 1.12 0 1.20 0 1.49 0.76 0.68 2.96 2.60' 0 0 0 2.45 1.40 1.23 1.23 0.75 1.32 0.99 1.16 P c: 0.05. 95% CI 0-4.39 0.02457 0.03-6.24 0-3.27 0.03-6.68 0-6.83 0.31-4.36 0.02-4.23 0.02-3.79 0.36-10.69 1.12-.5.11 0-8.35 0-3.16 0-1 6.04 0.30-8.34 0.29-1.09 0.15-4.44 0.79-1.81 0.20-1.92 0.43-3.08 0.45-1.38 0.81-1.61 314 5 JZ 8 A B C D E F 0 H I J K L M Comblned (U.S. 1 I4 Comblnod (U.KS Ft Combined (US.&U.Kj r 0.1 WONG A N D RAABE 5 c 8 . a:I I I I I_ 1- -+?-+ 1 I t N1 01 -aP- L Combined (US. S- Combined (U.K =. Combined (U.S.&U.K1'p 1 10 0.1 StandMOed Mcitalny R.Mand 55s CI I I I -4 --L 7 +- l I L-u 1 aSlnlNlarlllrrd ~aclllllybo md 9s" I I I .^ FIG. 3. Acute lymphocytic leukemia among petroleum workers in the United States and the United Kingdom, 1937-1989.A-Q. US studies; R-S.UK studies. FIG.4. Chronic lymphocytic I w i I w 1 1 t H among perroleurn w>~.i--. in the United States and United I i i i ~ ~ ~19I37t-~19~89~.-~4-Q. . V S -: ies: R-S, UK studies. 1 US and 1from the UK), the cell-type-specific leukemia results remained essentially unchanged. The metaSMRs for petroleum refinery workers in the United States and in the United Kingdom were 0.93, 0.94, 1.32, and 0.87 for AML, CML, ALL, and CLL, respectively (Table 7 ) .None of the meta-SMRs for the US-LX combined cohort of petroleum refinery workers was statistically significant. Among the 19 cohorts included in the analysis. length of follow-up ranged from 13 1.0 I: I years. A separate ar. ...ysis was carried out based 0 1 1 c*ohortswith at l e a s :I. years of follow-up. Using this csriterion, 3 cohorts were excluded. The results of'this Iiii:iIysis based on coiw::? with 15+ years of follow-ul) tire presented in T a b k ? The results for follow-up of'1:) I years remained essr::. tially the same. DlSCllSSI(~N The combined database in I IIV meta-analysis re:. yr. sented one of the largest dat~iil)tisc~isn epidemioloc; ::- I ' / TABLE 6 Meta-analysis of Chronic Lymphocytic Leukemia in Petroleum Workv.rs in the United States and the United Kingdom, 1937-1989 Country ~~ us US combined cohort UK UK combined cohort US-UK combined cohort * P < 0.05. Cohort ~ A B C D E F G H I J K L M N 0 P Q R S Observed deaths ~~ 2 3 9 1 2 4 8 2 4 2 3 2 1 0 1 r 2 56 12 16 28 83 Expected deaths 2.97 3.79 3.48 3.00 2.92 2.18 6.99 4.10 4.99 2.19 10.04 2.67 4.09 0.79 1.91 7.22 5.87 69.19 15.80 13.90 29.70 98.89 SXIR 0.67 1.32 2.39' 0.33 0.69 I.& 1.14 0.49 0.So 1'.91 0.30' <'.75 \e4 0 ('32 \'.97 ('34 \'.T9 b.76 :25 \'.94 .'1 :s - -- -_- _ ._-I_- _- .- . - - 1 , - _..__-_ _-- - _ _-- ._ / CELL-TYPE-SPECIFIC LEUKESIIX IN PETROLEUM WORKERS TABLE 7 Meta-analysis of Cell-Type Specific Leukemias in Refinery Workers in the United States and the United Kingdom. 1937- 1989 rei1 type - ,ienous leukemia .z r.oqenous leukemia -nlnocytic leukemia . x p n o c q z ~ cleukemia Observed deaths 78 35 "1 47 Expected deaths 84.08 37.42 1.5.85 54.23 Meta-SMR 0.93 0.94 1.22 0.87 315 95% confidence interval 0.78-1.16 0.6.5-1.31 0.81-2.01 0.64- 1.16 .- . ,' 3. Even though most of the individual cohorts were .- . .:ei;: large, it is evident from Tables 3-6 that most ce!!-Eype-specific leukemia SMRs based on indi- cqhorts were unstable, as most were based on -.- .?umbers of deaths (particularly for ALL). One of - .-ajar advantages of the meta-analyses carried out in ~ - -.7 -resent investigation is the stability of the combined :...,- .-z...:h regard to cell-type-specific leukemia analyses, .-.+ -:atisrical power of the individual cohorts was Iim- . y :. This was particularly true for the US cohorts, most : -.ik%ere individual company or individual refinery .-. .=s. The two UK studies were larger because they .- . - isaustqwide studies consisting of workers from -- +--aicompanies. Based on the average of the 17 indi- .I-ai cohorts in the United States, the minimum SMR -:+:+::able at a = 0.05 and 3 = 0.20 in an individual co- - . . xere 2.22,3.02,4.37, and 2.61 for .AML, CIVIL,XLL. :.-. 1 ;?LL.respectively. All of these SMRs were higher .- - --.vorbld.The inadequacy of statistical power of in... ...-,c.~: xudies was particularly evident for ALL. . 2::stical power increased substantially by combin- .-: -ne 17 US cohorts. Table 9 presents the minimum ::-+.::able SMRs at a = 0.05 and J = 0.20 for each of the . i-- najor cell types. For .4,ML, the US combined cohort :.- .:ides sufficient statistical power (80% power at the ' :evel) t o detect a risk ratio as small as 1.25, if indeed -.:-:? was an increased risk. Based on the US combined .. T.the minimum detectable SMRs were 1.39 for -~ 2nd 1.32 for CLL, respectively. Even for XLL. the +-. :'requent cell type, the minimum detectable SMR - i-+d!>nthe US combined cohort was 1.63. All the min- -7:z detectable SMRs based on the US combined co- hort were less than twofold. The statistical power of the UK combined cohort was somewhat lower (Table 9). In the overall US-UK combined cohort of petroleum workers, the minimum detectable SMRs were 1.21, 1.32, 1.51, and 1.26 for AML, CML, ALL, and CLL. respectively. Thus, the absence of an increased risk of AML, CML, ALL, or CLL in petroleum workers in the United States or in the United Kingdom based on the meta-analyses in the present investigation could not be attributed to inadequate statistical power. One may question why an increased risk of AML was not observed in petroleum workers, who were certainly exposed to some levels of benzene. The answer probably lies in the threshold of benzene exposure required to pro- duce a si-gnificant increase of XML. T o estimate the threshold, the data from the study of Pliofilm workers exposed to benzene (Rinsky et al., 1987) updated through 1987 were analyzed specifically for AML. For the Pliofilm cohort as a whole, 6 deaths from AML were observed. compared to 1.19 expected (Table 10). The corresponding O I L SMR of 5.03 was statistically significant. However, exposure-response analysis indi- cated that no increase of AML was detected for cumulative exposure below 200 ppm-years (SMR = 0.91). Furthermore. the exposure levels provided by Rinsky et al. 1987;)were !ikely underestimated, and more realistic exposure estimates have been developed by others (Crump and Allen, 1984; Paustenbach et af.,1992). Had the other exposure estimates been used, the observed XML threshold would have been much higher. Emerging evidence in the biologic search for a mechanism of benzene-induced h M L suggests that benzene and its metabolites may induce AML via toxic disruption of regulatory TABLE 8 .Meta-analysis of Cell-Type-Specific Leukemias in Petroleum Workers in the United States and - the United Kingdom. 1937-1989, Based on Cohorts with 15 or ,More Years of Follow-up Cell type Observed deaths Expected deaths Meta-SMR 95% contidence interval . :nyelopenous leukemia ~ n..vA. mvelorenous leukemia 2;c lymphocytic leukemia 138 53 33 71 143.38 64.86 27.08 89.74 0.96 0.Y1-1.14 0.82 0.62-1.08 1.22 0.83-1.71 0.79 0.62-1.00 316 WONG AND RAABE TABLE 9 Minimum Detectable SMR (at a = 0.05 and fi = 0.20)by Leukemia Cell Type in Combined Cohorts of Petroleum Workers be th Combined cohort us UK US and UK Cell type AML CML ALL CLL AML CML ALL CML AML Observed deaths 103 43 2-5- 33 45 19 9 28 148 Expected deaths 107.68 47.97 20.26 69.19 4'7.20 22.00 9.10 29.70 154.88 Mew-SMR 0.96 0.90 1.23 0.i9 0.95 0.86 0.99 0.94 0.96 95% confidence interval Lower limit L'pper limit 0.78 1.16 0.65 1.21 0.79 1.81 0.59 1.03 0.70 1.27 0.52 1.35 0.45 1.88 0.63 1.36 0.81 1.13 P value 0.65 0.4'7 0.99 0.08 0.75 0.51 0.97 0.75 0.58 Minim::n, detecla!,i, -SAIF , 1.11- 1.;:> i.t.i: 1.31 1.35 1.6(, 1.9s 1.51 1.2i tr: pi ty gI: 8.5 Is: nc Li: 0- of CML ALL CLL 62 69.9'i 0.89 34 29.36 1.16 83 98.89 0.84 0.68 1.15 0.34 1.32 ex. 0.81 1.61 0.40 1.5: Sb -0.67 1.04 0.10 1.2C wa I.- 1 * mechanisms of cell growth and differentiation (Snyder and Kalf, 1994). If benzene-induced toxicity is a conditional step in benzene-related AML, this would lend biologic support to the observed threshold based on epidemiologic data (World Health Organization, 1993).Thus, the lack of an increased risk of AML in petroleum workers could have been due to the substantially lower benzene exposures in the petroleum industry than in the Pliofilm study, which included workers exposed to benzene levels as high as several hundred ppm. Similarly, benzene exposure levels in case reports of AML in shoemakers from Istanbul, Turkey, were extremely high. These shoemakers, who used solvents containing up to 88%benzene. were exposed to air benzene levels up to 650 ppm (Aksoy, 1985). In addition, some of these shoemakers worked a t their homes and were thus exposed all day and all week long. In contrast, benzene exposure at refineries was substantially lower compared to levels a t the Pliofilm plants or among Istanbul shoemakers. For example, for general plant operations at petroleum refineries, the mean benzene level of 14,824 samples from an industrywide survey was 0.22 ppm (Runion, 1988). For gauging content in the field (by production workers) or a t refinery tank coI farms (by refinery workers), 60% of the industrial hy- der giene samples were <1 ppm, 22% were between 1and ;. Th ppm, 8%were between 3 and 10 ppm, 7 % were betweep 10 and 25 ppm, and only 2% were between 25 ana par of - ppm. Similarly, based on a comprehensive survey in r h i not petroleum refining industry, Spears et al. (19871 :e- the ported that most measurements based on both 8-hr i :;? sor 13-min time-weighted average benzene exposures -*v?re tha 'below 1ppm. For land-based gasoline distribution workers, :ht highest exposures were received by drivers of tan6 trucks using the splashing method for loading durir.; otk tior me: exP 1950-1964 (Smith et al., 1993); and their full-shiz exposure level for total hydrocarbons was estimated :c 1 Ab! hor to- be.220 ppm, which was equivalent to approximately 3.5 ppm of benzene (International Agency for Researct x Cancer, 1989). For marine distribution workers. :he highest exposures occurred among deck personnel a x - 1 cre' thrF stuc ing loading with open-hatch venting (Smith et al., 1992i: and their full-shift exposure level for total hydrocarbor.: 1 ! CUP incr et a was estimated to be 230 ppm, or approximately 4 ppn? o: 196 benzene. These exposure data indicated that petrole-z date workers, even including those who worked in the inc.25- nific try decades ago, would not have accumulated sufic::~.: I F Serb Cumulative exposure in ppm-years TABLE 10 Acute iMyelogenousLeukemia by Cumulative Benzene Exposure Observed deaths Expected deaths Standardized mortality ratio - 95% confidence inri-.l / <40 40-200 200-400 4001 Total ** P < 0.01. 2 3 6 0.84 ji.09 0.25 0.07 0.03 1.19 :.'9]o.9i 27.21" 98.37" 5.03" 0.03-6.63 3.99-98.24 20.28-287.65 1.&-10.97 / CELL-TYPE-SPECIFIC LEUKESlLi IN PETROLEUM WORKERS 317 - llrnui .crab -LlR 75 -39 63 32 39 60 99 51 11 32 51 -76 ` hY nd veei d 5r I thc re an( veri t hc an1 ring >hiA d tc 3.E .1 on the iur. 33); ons 2 oi Ud usi; 4 .1 A ,,,.:~ene exposure in excess of the estimated AiML -.;.lr~eshold derived from the Pliofilm study. Distribution of leukemia cases by cell type in the pe- :rr,ieumworkers was similar to that in the general adult 3,,pulation (Table 2 ) . In particular, of the four major cell :,,pes -1ML accounted for 46% in the petroleum workers 2fi:i 4;;,7 in the general adult population. In contrast, .- ?.- ,I[ the leukemia patients with benzene exposure in ;s: ..nbul (the majority being shoemakers) had acute nc?r.iymphocytilceukemia or ANLL (Aksoy et al., 1976; Liner. 1985). Because of the generally low benzene exposures JEong petroleum workers, the lack of an increased risk .CdL is not surprising. Of the 19 individual cohorts .s.inined, none showed a statistically significant .AiML .G.`:. ?. In a previous analysis based on Cohort G, AML .,v.:- si-gificantly elevated; SMR = 2.19 and 95% CI: 1.: -3.32 (Wongsrichanalai et al., 1989). A nested casex . r r o l study within Cohort G has been conducted to &:ermine the cause of the increase (Austin et al., 1986). The investigation did not identify any specific job or de?aT-ment responsible for the increase, and comparison (JI` ::?e benzene exposure levels of cases and controls did ".,L..- i n c c a t e any association with benzene exposure at :k:. :e?zery. The authors offered several possible rea- ...+--.._ =(.. fcr the increase: (i) unidentified exposures other --.- oeazene at the refinery, (ii) exposures a t work sites .>:.??rrhan the refinery, (iii) unidentified nonoccupa- :ionai enosures, or (iv) chance. Based on the present .r.ca-ar,alysis, it was unlikely that some yet unidentified +s?usures other than benzene were responsible for the .A.!.IL iccrease reported in the previous analysis of Co- -:-- G.llnless these unidentified exposures were unique ::I. ?articular refinery. More likely, the AML in- .:-..=_:$*.=reinasConoshoorftfeGrewd basy due the to one or more investigators of of the other the original -:.:C>-. For example, cigarette smoking and other nonoc- -.-?,arionai lifestyle risk factors have been linked to an - -.:~-':?ssd risk of AML .- . ---.`.. 1900: Friedman, in some recent studies (Severson 1993; S a n d e r et al., 1993; Linet, --".> I . In any event, the analysis based on the latest up- 1- ::? Cohort G indicated that AML was no longer sig- ---1. -..-1-... .-1~::)- -- . , . l - -4 CL ele IL, :he vated (SMR = l.23,95% no significant increase individual cohorts of p CI: 0.70-2.15). or deficit was obetroleum workers. -.--+ =era-analyses based on the combined cohorts ..:. .-v+i :8ar There was no increase of CML. This finding --. -- 1 :.--.ceased CML risk was further supported by two ...:. .-. - ._ -..--_-:L-7. x:-S-wmesldteund,iensoofasCsoMcLia.tiIonn a case-control st (P = 0.91) was udy of found :Be.-. i~ z~ - disease and occupational exposures to pe- ---_ > d u c t s (Brandt et al., 1978).Thus, the authors . ...-..-.X-~_L_ `-unlike the patients with ANLL, occupa- - --. -.i5-qvsures to petrol products was not unduly com- . .. -2--c=zg rhe patients with CML." In another case- -:-.:r ~ ~ odf Cy ML conducted by investigators at the _ _+Y:T.- of Leeds and the University of Edinburgh ?--z-z*y cr ai., 1990), exposure histories of 133 CML cases and 241 controls were examined. This study did not find any association between CML and exposures to benzene or solvents (P= 0.41).The authors concluded: "This case-control study failed to reveal any risks for solvent or chemical exposure." Thus, there is no increased CML risk associated with chronic exposures to benzene or other aromatic hydrocarbons. This conclusion is consistent with the view expressed in the Cecil's Textbook of .\/ledicine (19th ed.): "Benzene exposure increases the risk of acute myelogenous leukemia (AML) but not of CML" (Wyngaarden et al., 1992). For ALL, only Cohort K showed a significant increase (SMR = 2.60,95% CI: 1.12-5.11). .A review of the avail- able employment histories of the ALL cases by the authors of the original cohort did not shed any additional light on a possible explanation for the excess. Moreover, it is interesting to note that there was a significant deficit of CLL ( 3 observed vs 10.04 expected) in the same cohort. Five of the eight ALL cases in Cohort K were diagnosed in a small county in the South and, therefore, the ALL excesscoupled with a CLL deficit could have been the result of misclassification due to local diagnostic practice. Even if only one case had been misclassified, the ALL SMR for Cohort K would not have been statis- tically significant. No other individual cohorts showed a significant in- crease of XLL. In fact, no ALL deaths were reported in six individual cohorts. For the US-UK combined cohort, the observed number of ALL deaths was not significantly different from the expected (34 observed vs 29.36 expected, P = 0.40). Because of the rarity of ALL, few population-based case-control studies of ALL have addressed the issue of ALL in petroleum workers. In a casecontrol study based on 5147 leukemia deaths from 16 states in the United States, Loomis and Savitz (1991) reported that not a single ALL case had worked in the petroleum refining industry (odds ratio = 0). Although the number of ALL in the study was not reported, it can be estimated that approximately 10% of the total number or 500 were ALL. The finding from the Loomis and Savitz (1991) study that none of the ALL cases worked in the petroleum industry supported the results of the .ALL meta-analysis in the present investigation. In a review on environmental factors of leukemia, Brandt (1985) also concluded "Unlike ANLL there seems to be little evidence a t present that the development of ALL is related to exposure to some chemicals." Finally for CLL, as remarked above, Cohort K exhibited a significant deficit. On the other hand, Cohort C showed a significant excess of CLL (SMR = 2.59, 95% CI: 1.1s-4.91). -4s discussed earlier, approximately 4 statistically significant SMRs could be expected to occur by chance, given that there were 76 SMRs tested at the LY = 0.05 level. Indeed, there were 3 statistically significant SMRs in Tables 3-6. For the US combined cohort, the CLL SMR was 0.79 (95%CI: 0.59-1.03), almost signiticantly low a t the 0.05 level. For the US-UK combined cohort, the CLL meta-SR/IR was 0.84, indicating that 318 WONG AIND KAABE i I Ithere was no increased CLL risk among petroleum work- exposure. Furthermore, the increase was substantia b: ers in the United States or the United Kingdom. This reduced and no longer statistically significant in the 1; i . finding was supported by a recent case-control study of' est update of Cohort G. chronic lymphocytic leukemia and chemical exposures Among petroleum distribution workers, drivers ,,! j conducted by epidemiologists at the University of Wash- tank trucks have the highest potential for exposure I,, 1 ington (Malone et al., 1989).This study found that the hydrocarbons including benzene. In the UK distributic,l, I risk ratio for CLL associated with exposure to aromatic study, Rushton ( 1 9 9 3 ~p)rovided cell-type-specific an.,j. I hydrocarbons (including benzene) was 1.1. Similarly, in ysis for drivers. The AML SMR for drivers was 1 i.; a case-control studies of 342 CLL cases in the Baltimore (95% CI: 0:82-2.65). Unfortunately, no analysis area, benzene was found to be unrelated to any increased exposure to either total hydrocarbons or benzene 1 4s risk of CLL (Linet et al.,1987).In the case-control study carried out. Analysis by hire date indicated that an ,n. , conducted by Loomis and Savitz (1991)discussed above. creased leukemia risk (all cell t-ypes) was observed ir: , no significant increase in CLL among petroleum refining workers who started work before 1940, but it was nu: workers was found. Similarly. in the Swedish case-con- clear whether this result holds true for XML as well. ir, trol study conducted by Brandt et al. ( 1978) discussed the US study of distribution workers, no increased risk earlier, no association between exposure to petroleum products and CLL was detected ( P = 0.70). In addition, the relationship between lymphocytic leukemia and solvent exposures in rubber industry workers has been examined previously. In a nested case-control study of lymphocytic leukemia, risks associated with exposure to 24 solvents used in the rubber industry were investigated (Checkoway et al., 1984). Unfortunately, the analysis was not further specified into the two main subgroups. CLL or ALL. However, the ratio of CLL to ALL is approximately 24 to 1in the general adult population, and. therefore, the analysis was weighted more heavily with CLL than with ALL. A number of solvents (e.g., carbon disulfide, carbon tetrachloride), but not benzene, were observed to be significantly associated with lymphocytic leukemia in rubber industry workers. In addition to the overall meta-analyses based on all cohort studies of petroleum workers in the United States and the United Kingdom, we also conducted meta-analyses restricted to studies of refinery workers only (Table '7). The results based on refinery studies were similar to those based on all petroleum studies. Furthermore, we also performed meta-analyses based on studies with a t least 15 years of follow-up (Table 8). T h e results re- mained unchanged, after stratification by length of follow-up. As discussed previously (Wong and Raabe, 1989, 19901, data are not available from individual studies for a meta-analysis by level of exposure to petroleum products. A few individual studies have presented exposureresponse analyses by leukemia cell type. -4review of these analyses will shed additional light on the interpretation of the meta-analyses in the present investigation. For refinery workers. in the case-control study of Cohort G. detailed analyses of ArvlL by benzene exposure categories were carried out (Austin et ~ l . 1, 986). The controls were found to have a higher cumulative benzene exposure than the AML cases. Other cumulative measures of benzene exposure were obtained using differenr weighting schemes, and for each of these measures the controls consistently scored higher than the for either leukemia (allcell types combined) or AML \.,.a> found for drivers. Furthermore, exposure-respo analyses were carried out for both leukemia (all I types combined) and AML in the US study (Won:. e: ai.,1993;Wong and Trent. 1993).Analyses using logistic. regression models based on length of exposure, curnula- tive exposure, and frequency of peak exposure did no; find any increased risk or exposure-response relationship between these exposure indices and leukemia (ai: i . .' 1- e l! r cell types combined) or -4ML (Wong and Trent. 199:;I Thus, these detailed analyses from individual stc.iie.. i supported the results of the meta-analyses based ox- :h~ combined cohorts in the present investigation. The meta-analyses by leukemia cell type in the pre-- F t. .E L( f' ii:ent investigation underscore the importance of takin: cell-type specificity into consideration. Moreover. re- Icently the concept of even more specific diagnostic c a w gories ("sub-cell types") has been gaining increasine US^ in epidemiologic analyses. For example, AML as a singit disease category may not be as homogenous as inct thought to be. especially for etiologic investigation:. Re BT. 1cent research suggests that, at the molecular or c>-:ogc- netic levels. AML may be considered a heterogenous di:. W us Iease (Smith et al., 1992).In a recent epidemiologic sruc.. using ras mutation-positive (ras-positive) -4ML as a rnc, ! lecular subtype, Taylor et al. (1992) reported that ,rc.-positive AML was associated with occupaTim5 ~~ Iexposures but not ras-negative 4ML. In another :'st AA control study of ANLL, a distinction betweec r. rm;: and abnormal karyorypes was made (Crane et ai..:33' i Ak In the cytogenetically defined subset of ANLL wirh d., normal karyotype. the odds ratios were 5.3 (95%CI: 1..' 21.9) for smoking more than 2 packs of cigarette- ? c . day, and5.9 (95%CI: 1.4-25.6) for consuming more C k ; ' 2 02. of alcohol daily. For benzene exposure (le;-: _ -"7 i I Bn,. specified),on the other hand. the odds ratio was 1.: Y'' 1 Err CI: 0.3-3.8).Thus. in the future, cytogenetic infor: i:,l''' will likely further increase the specificity of &sea.- :;:'- e le sification, which in turn will improve epidez: ~ i ' ' ~ . analyses. CONCLUSION BfT E ne e cases. Thus. the AML increase reported in the previous In conclusion, epidemiologic data have demon.:rarr & analysis of Cohort G could not be attributed to benzene that, given high enough exposure level over an exii-nc' Chr .. CELL-TYPE-SPECIFIC LEIJKEMI-A IN PETROLEUM WORKERS 319 ~c:: ~:., time, benzene .,.,genous leukemia. can increase Furthermore, the risk of acute where quantita- - . l. _,'- y p ~ s u r estimates are available, epidemiologic data ..'...,.- indicated that the observed threshold needed to in- :r2+2 [he risk of acute myelogenous leukemia has been ..--_...._a.c.ed to be at least 200 ppm-years of benzene ..\: -.!re. and most likely substantially higher based on -.._..., ,-eaiisticexposure estimates (in the range of 400-, :,m-years). The meta-analysis on acute myeloge- .. .,' . _ .eukemia in the present investigation based on a -.i:ined cohort of more than 208,000 US and UK pe.:, +xn workers has demonstrated that these workers ,,=:e not at an increased risk of acute myelogenous leu- ic:::ia because benzene exposure levels in the petroleum ,::ci-~tryhave been substantially lower than needed to .-?A,. 1 :he observed threshold. 7 .:hermore, in assessing the relationship between : ~ ~ r . _ .e-xnpeosure and leukemia, a distinction must be n : x t with respect to histological cell types of leukemia. The meta-analyses on chronic myelogenous leukemia. .wiKe lymphocytic leukemia, and chronic lymphocytic :eukemia have demonstrated that there is no increased risk for any of these cell-type-specific leukemias among ?erroleurn (refinery, production. pipeline, and distribuF.i., .~I workers in the United States or in the United E:. ;dom. Meta-analyses restricted to refinery studies or :t -:=dies with at least 15 years of follow-up yielded sim:iar results. T h e conclusion of the present investigation ;supported by case-control studies on cell-type-specific leukemias conducted in the United States, the United Kingdom, and other countries. A C K N O WLEDGLVENTS ..e L hank all the investigators of the original studies included in our .i:. .%vsis.A partial list of these investigators can be found in Table 1. '.\' ;hour their participation and willingness to share their data with ':.-. This meta-analysis would have never been possible. REFERENCES .\2soy. M. f 1985). Benzene as a leukemogen and carcinogenic agent. l.m. J. Ind. 8.9-20. -*)y. 41.. Dincol, K., Erdem. S., and Dincol, G. (1972). Acute leukex i i i ciue to chronic exposure to benzene. Am. J . .Wed. 32.160-166. .i..x=oy. 41.. Erdem, S., and Dincol. G. 11976). Types of leukemia in zhonic benzene poisoning. .A study in thirty-four patients. Acta .Yurmat 55.65-72. .\ustin. H., Cole. P., and YIcCraw, D. S. (1986).A case-control study ~<!eukemia t an oil refinery. J . Occup. .\fed. 28. 1169-1173. [{ranat. L. (1985).Environmental factors and leukemia. .Wed. Oncol. Tunor Pharmacother. 2.7-10. -:indr. L.. Nilsson, P. G.. and 5Iitelman. F. (1973). Occupational rxnusure t o petroleum products in men with acute non-lymphocytic !e,:kemia. Br. Med. J.1,553. {rotwing, E. I1965). Toxicity and .\.letabolism of Industrial Solvents. Elsevier. .Amsterdam. 'ieckoway. H., Wilcosky. T.. Wolf. P.. and Tyroler. H. (1984). An evaluation of the association of leukemia and rubber industry solvent exposures. Am. J. Ind. .\fed. 5 , 239-249. 'hristie. D.. Robinson, K., Gordon. I.. and Bisby. d. (1991).A prospec- tive study in the Australian petroleum industry. 11. Incidence of cancer. Br. J . Ind. .Wed. 18.511-514. Collingwood. K. W., Milcarek. B. I.. Raabe, G. K., and Wong, 0. (1994).A n L'pdated Cohort Mortality Study of Workers at the PauLsboro. .Vew Jersey Refinen.. 1946-1987. iMobil Oil Corporation. Crane, h1. M.,Kearing. 51. .J., Trujillo, ,I. M.,Labanhe, D. R., and Frankowski, R. F. (1989). Environmental exposures in cytogenetically defined subsets of acute nonlymphocytic leukemia. J. Am. .Wed. .-tssoc. 2 6 2 , 634-639. Crump, K. S.. and Allen. B. C. (1984). Qwntitatioe Estimates of Risk of Leukemia from Occupational Exposure to Benzene. Prepared for the Occupational Safety and Health Administration. Dagg. T . G., Satin. K. P., Bailey, W. d., Wong, O., Harmon, L. L., and Swencicki, R. E. (1992). An updated cause-specific mortality study of petroleum refinery workers. Br. J. Ind. Med. 49,203-2112. Day, N.E. ( 1985).Statistical considerations. In Interpretation of iVeg- atiue Epidemiologic Euidence for Carcinogen.icity(N.J. Wald and R. Doll, Eds.). IXRC Scientific Publication No. 65. Delore, P., and Borgomano, C. (1928). Leucemie aigiies au cours de l'intoxication benzenique. Sur I'origine toxique de certaines aigiies et leurs relations avec les anemies paves. (Acute leukemia following benzene poisoning. On the toxic origin of certain acute leukemias and their relation,to serious anemias.) J. &fed.Lyon 9 . 2 7 - 2 3 3 . Divine, B. d.. and Barron, V. (1986). Texaco mortality study. 11. Patterns of mortality among white males by specific job groups. Am. J Ind Med 10,371-381. Divine, B. .J., and Barron, V. (1987): Texaco mortality study. 111. A cohorc study of producing and pipeline workers. Am. J. Ind. Med. 11, 189-202. Divine. B. .J.. Barron, V., and Kaplan, S. D. (1985).Texaco mortality study. I. Mortality among white male refinery, petrochemical, a n d research workers. J.Occup. Med. 27,445-447. Enterline, P. E., and Henderson, V. (1985).A Report on the iMortali& Experience of Workers from a Petroleum Refinery at Torrance, California. Report submitted to iMobi1 Corporation. Ford. A. 51.. Molgaard. H. V., Greaves. ,M.F., and Could, H. .J. (1983). Immunoglobulin gene organization and expression in hemopoietic stem cell leukemia. E M 5 0 J . 2.997-1001. Friedman. G. D. 11993). Cigarette smoking, leukemia, and multiple myeloma. Ann. Epidemiol. 3.425-428. Girard. R., and Revol. L. (1970). The incidence of exposure to benzene in severe haemopathies. ,Vouu. Rev. Hematol. 10.477-484. Goguel. X., Cavigneaux. A., and Bernard, d. (1967).Benzene leukemias. Bull. Inst. LVatL. Sante Rech. Med. 22,421-442. Hanis, N. M.,Holmes, T. M.,Shallenberger, L. G., andJones, K. E. ( 1982).Epidemiologic study of refinery.and chemical plant workers. J.Occup. Med. 21.'203-212. Hanis, N. 51.. Shallenberger, L. G., Donaleski. D. L., and Sales, E. A. (1985a).X retrospective mortality study of workers in three major U S . refineries and chemical plants. I. Comparisons with US.population. J . Occup. .\fed. 2 7 , 283-292. Hanis. N. &I.. Shallenberger, L. G., Donaleski, D. L., and Sales. E. A. i 1985h). A retrospective mortality study of workers in three major US.refineries and chemical plants. 11. Internal comparisons by geographic site. occupation, and smoking history. J. Occup. .Wed. 27. 361-369. Honda, Y., Delzell. E.. and Cole, P. (1995). An updated study o f mortality among workers a t a petroleum manufacturing plant. J . Occup. .Wed. 37, in press. Hornstra, $1. ( 1990). Reanalysis of the Phase I Cohort Mortality Study of Oil Refiner)/ Workers (1970- 19010): Follow-up Extended through 1986..Amoco Oil Corporation. Hunter. F. T . (1939). Chronic exposure to benzene (benzol). 11. T h e clinical etfects. J . Ind. Hyg. 2 1,331-354. 320 Infante. P. F.. Rinsky. R. A., Wagoner. J. K.. and Young. R. d. I :Or. Seison. S.. Lex-.:.. R. J., Albert. R. E.. Blair. A. E.. Griesemer. : ;, i Leukemia in henzene workers. Lancct 2.76- 78. Landrigar.. F ..I.. S:.ayner, L. T., and Swenberg. J. A. (1986). Cg r, International Agency for Research on Cancer (1989). IARC .\for;graphs on the Evaluation of Carcinogenic Risk9 to Humans. &:&-tional Exposures in the Pctroleum Refining: Crude Oil and mc:c- -fftrnlcum Fuels. Voi. 45. IARC. Lyon. France. bur ion of rLir:-aidehyde to respiratory cancer. Enciron. Healti; .+ spec:. i o . 2 : : - 3 . Olsen. J. H.. ~ - ..ltensen, 0. M. (1987). Occupation and risk of callCeI in Denmark. S.-and. J. Work. Enciron. Health 13(Suppl. 1). I-:+;, Jakobsson, R.. Ahlbom. A.. Beiiander, T.. and Lundberg. I. 8 1 3 o S - . Pausrenbach. c'.J.. Price, P. S., Ollison, W.. Blank. C.. Jernigar, Acute mveloid leukemia among petro station attendants. -4rck. E?- J. D.. Bass. R. D.. and Peterson. H. D. (1992). Reevaluation oii,en viron. Health 48. 255-259. zene e s p o s u u for the Pliofilm irubberworkers) cohort (1936-19:fi Joyner. R. E. (1983). Letter to Dr. J. Donald Millar, Director of h-2- J . Tozicoi E;:;-on. Health 36. 177-231. tional lnstitute for Occupational Safety and Health. dated JL:v 2. Peach. H.. C k s i e . D.. Robinson, K.. Gordon. I., Livingston. P.. 1963. ter. and D ' c h S r a i n , J. (1992). Health Watch. the Australian , .{, Liner. M. 5. (1985). The Leukemias: Epidemiologic Aspects. Osfoxi Univ. Press, London. ruff01 P~.:.T-.~m-L Health Surceillonce Program. Ninth repon. ,'he Cniversir:- of lfelbourne. I Liner. M. S., Stewart. W. F., Van Natta. .M.L.. McCaffrey. L. D.. anC Szklo. .M.(1987). Comparison of methods for determining OCCL.- Raabe. G. 6.I :33S). Review of the carcinogenic potential of gasoi'ne Encimn. Fitxi:;: Perspect. lOl(Supp1. 6), 35-38. tional exposure in a case-control interview study of chronic 1n;pii.- Raabe. G. K.. \'ollingwood, K. W., Milcarek, B. I., and Wong. 0 cytic leukemia. J . Occup. Med. 29, 136-141. (1994).A n i-:%iz:ed Cohort Mortalit? Study of Workers at the &soli. Loomis. D. P.. and Savitz, D. .A. (1991). Occupation and leukemia mortality among men in 16 states: 1985-1987. Am. J . I d . Med. 19,329- mont. Tt-xcr. &--inen, 1945-1982 Mobil Oil Corporation. Reeve. G. R.. Thamas. T . L., Kelly, V. F., Waxweiler, R. J., and I: .;'& 1 521. Maher, V. E., Gill, L., Townes, P. L., Wallace. J. E., Savas, L.. \I-& B. A., and Ansell, J. E. (1993). Simultaneous chronic Iymphw-cic S. (19S2). -A ;raportionate mortality study of a n oil, chemical. .nc aromic worke- local in Texas City, Texas. Ann. N Y A c a d . Sci. ~ 5 1 ' . 54-61. leukemia and chronic myelogenous leukemia. Cancer 71, I%:Ri>nsky. R. -A.. \-oung, R. J.. and Smith, A. B. (1981). Leukemia 1997. benzene workers. -Am. J.Ind. :Wed. 2,217-245. Malone, K. E., Koepsell, T. D., Daling, J. R., Weiss, N. S.. Morris P. D., Taylor, J. W., Swanson. G. M.. and Lyon, L. T. (19391. Chronic lymphocytic leukemia in relation to chemical exposures Am. J . Epidemiol. 1 3 0 , 1152-1158. Rinsky. R. A.. Suairh. A. B.. Hornung, R., Filloon, T. G., Young. R. .I.. Okun. -4.H.. a n d Landrigan. P. J. (1987).Benzene and leukemia. al: epidemiolog5nc risk assessment. -V. Engl. J . Med. 31 6 , 10&&1050. Runion. H. E. I 19.W).Occupational exposures to potentially hamrd. ' I Marsh, G.. Enterline, P. E.. and McCraw, D. (1991). Mortality pa:terns among petroleum refinery and chemical plant workers. Am. J. Ind. Med. 19.29-42. Matanoski, G. M.. and Elliott, E. A. (19811. Bladder cancer epidemiology. Epidemiol. Rec. 3 , 203-229. McCraw, D. S., Joyner. R. E.. and Cole, P. 1985). Excess leukemia in a refinery population. J . Occup. Med. 25,220-222. McKinney, P. A., Roberts, B. E., O'Brien, C., Bird, C. C., Richards I. D. G., Alexander, F., and Cartwright, R. A. (1990). Chronic myeloid leukemia in Yorkshire: A case control study. Acta Haemam!. 83,35-38. McMichael, A. J. (1988). Carcinogenicity of benzene, toluene and v - lene: Epidemiological and experimental evidence. In Environmenm! Carcinogens. Methods of Anal>& and Exposure Measurement: Benzene and Alkylated Benzenes (L. Fishbein and I. K. O'Neill. Eds.1. Vol. 10. IARC Scientific Publication No. 85. . Milcarek, B. I., Collingwood, K. W.. and Raabe. G. K. (1994). An L p dated Cohort Mortality Study of White Male Workers at the Torrance. CaliforniaRefinery, 1959-1987. Mobil Oil Corporation. Morgan. R. W.. and Wong, 0. (1984). Ar. Epidemiologic Analysis of the Mortality Experience of Mobil Corporatzon Employees at the Beaumont, Texas. Refinen. Report submitted to Mobil Corporation. Morgan, R. W.. and Wong. 0. (19858). Final Report. An Epidemwlogical Mortalie Stud)' of &lobi1 Corporation Employees at the PauLsboro. New Jersey. Refinen. Report submitted to Mobil Corporation. Morgan, R. W.. and Wong, 0. (1985b). .4 review of epidemiological studies on artificial sweeteners and bladder cancer. Food Chem. Toxicol. 23,529-533. ous agents in :ne petroleum industry. Occup. Med. State .Art Rc 3. 431-414. Rushton. L. \ 1!'3Ssl. Further follow-up of mortality in a United K . T . ~ . dom oil refiner\- cohort. Br. J. Id:Med. 50,549-560. i 1Rushton. L. 199%). Further follow-up of mortality in a Unired King- dom oil distribution centre cohon. Br. J. I d . Med. 50,561-569. I IRushton. L. ( 1 9 9 3 ~ ).4. 39-year follow-up of the U.K. oil refinery anc distribution center studies: Results for kidney cancer and leukemia Enciron. Healrh Perspect. lOl(Supp1. 6). 77-84. Rushron. L.. a n d .Alderson. M. R. t 19811. An epidemiological survey 0: eight oil refineries in Britain. Br. J. Ind. Med. 38.225-234. ,Rushton. L.. and .Alderson, M. R. t 1963). Epidemiological survey c: 01: distribution centres in Britain. Br. J . Ind. Med. 40,330-339. I ISander. D. P.. Shore. D. L.. Anderson. J. R.. Davey, F. R., Arthur. D.. Mayer, R. J.. Silver, R. T.. Weiss, R. B., Moore, J. O., Schiffer, C. 3.. Wurster-Hill. D. H., McIntyre. 0. S.. and Bloomfield, C. D. (19931 1 Cigarette smoking and risk of acute leukemia: Associations with morphology and cytogenetic abnormalities in bone marrow. J . -Vu:! Cancer Inst. 8 5 . 1994-2003. ISatin. K. P.. Yuan. L.A., Wong. 0..Sewton. K. L., Bailey. W. J.. en. 1C. P., Dagg, T. G.. and Harmon. L. L. (1994). A n C'pdated Mort, i:. Stud? of Chevron's Port Arthur Reyineqf,1937-1987. Chevron poration. Schnatrer. A. R.. Theriault. G.. Karz, .4. M.. Thompson, F. S., L h naleski. D.. and Murray, 9.(1992): A retrospective mortality stud! within operating segments of a petroleum company. 4m. J . I n c Med. 22,209-229. Selvin. S.. Levin. L. I., Merrill, D. LV.. and Winklestein. W.(19s.;' Selected epidemiologic obsenations of cell-specific leukemia mor- i i Morgan, R. W., Foliart, D. E.. and Wong, 0. (1985). Asbestos and gas- tality in the United States, 1969-1977. Am. J . Epidemiol 117. I-('- trointestinal cancer: A review of the literature. West. J . Med. 143. 152. 60-65. Severson. R. K.. Davis, S., Heuse:. L., Daling. J., and Thomas. I: 5 National Center for Health Statistics (1990). Vital Statistics of the (1990). cigarette smoking and acute nonlymphoc>Ticleukemia. United States. 1987, Volume 11-Mortality, Part A. U.S. Depart- J . Epidemiol. 132. 418-422. ment of Health and Human Services. Nelson, N. A. (1985). Mortality m Amoco Oil Refiner\. Emplo-vees. Phuse I . Amoco Oil Corporation. Shallenberger. L. G., Acquavella. J. F., and Donaleski. D. (1992). A C , updated mortality study of workers in three major United Stale- ; refineries and chemical plants. Br. J. Ind. Med. 49.345-354. CELL-TYPE-SPECIFIC LEUKEAMIA IN PETROLEUM WORKERS 32 1 --, , ... .\ I , T.. Wiemels, d., Rothman. N.,and Linet. Pvl. S. (1992). ,iiTl,a.. e m i ~ aeIxposure, ras oncogene activation. and acute myeloid leuJ . .Varl. Cancer Inst. 84. 1614-1615. h , T..I.. Hammond, S. K., Hallock, >I.. and Wong, 0. (1993). 'c;ljth rrfects of gasoline exposure. I. Exposure patterns of distri- -..,...,,,nworkers in the United States. Enuiron. Health Perspect. 1 0 1\ S ~ p n l6. ) . 13-21. - . -. 3..snd Kalf, G. F. (1994).A perspective on benzene leukemo- ;i.<. C'rit. Rev. Toxicol, 2-4, 177-209. . .. R. C., Selvin, S., Schulman. J., and Francis. M. (1987). Ben- ..-.::* exposure in the petroleum refining industry. Appl. Ind. Hyg. 2. -' -,,-I 16.3. ;.:, ,,,r. .!. A.. Sandler, 0. P., Bloomfield, C. D., Shore, D. L., Ball, 5.3.. Neubauer, A, Mcintyre, 0. R., and Lui, E. (1992). ras oncoLrne activation and occupational exposure in acute myeloid leuke+J. J . .Vatl. Cancer Inst. 84, 1626-1632. - . -:.IS. T. L., DeCoufle. P., and Moure-Eraso, R. (1980).Mortality nq ..vorkers employed in petroleum refining and petrochemical is. J . Occup. &fed.22.97-103. .- .'- .,ds. T. L., Waxweiler, R. J., Moure-Eraso. R., Itaya, S., and :-'riumeni, J. F., Jr. (1982). Mortality patterns among workers in :+ree Teras oil refineries. Am. J. Ind. .Wed. 6,3-6. Torpe. J . J. ( 1974). Epidemiologic survey of leukemia in persons po.entially exposed to benzene. J. Occup. Med. 16,375-382. :-si. 5. P.. Gilstrap, E. L., Cowles. S. R.. Snyder. P. J., and Ross, 1. Z. (1993).A cohort mortality study of rwo California refinery and ;:;rochemica1 plants. J. Occup. Mea. 35.415-421. - . mi. 5. C.. and Saita, G. (1964).Benzene and leukemia. N . Engl. .\led. 271,872-876. ...'.. ..d. N.J.. and Doll. R. (1985). Interpretation of Negative Epidemio- ..igic Ecidence for Carcinogenicity. International Agency for Research on Cancer Scientific Publication No. 65. '+yen.C. P.. Tsai. S. P., McClellan, W. -4..and Gibson, R. L. (1983). Ldng-term mortality study of oil refinery workers. I. Mortality of 3ourlv and salaried workers. Am. J. Epidemiol. 118,526-542. ..'--.:irrohe4.1. M.,Lee, G. R., Boggs, D. R., Bithell, T. C., Forester, J., -hens J . W., and Lukens. J. 3.I 1981). Clinical Hematology, 8 t h :.. [.ea & Febiger, Philadelphia. . ' 2~ 0.. and Tabershaw. I. R. (1980). Comments on mortality m o n g workers employed in petroleum refining and petrochemical nlancs. J . Occup. Med. 22. 638-640. ',\.( m g . 0.. and Decoufli, P. (1982). Methodological issues involving the standardized mortality and proportional mortality ratio in oc- cupational studies. J.Occup. .Wed. 2 4 , 299-304. Wong. 0..Morgan, R. W., and Whorton, M. D. (1983). Comments on the iVIOSH study of leukemia in benzene workers. Technical Report submitted to Gulf Canada. Wong, 0. (1983). A n Indust- -Wide Mortality Study of Chemical Workers Occupationally Exposed to Benzene. Technical Report submitted to the Chemical Manufacturers Association. Wong, 0..Morgan, R. W., Kheifets. L., and Larson, S. R. (1985). Comparison of SMR, PMR and PCMR in a cohort of union members potentially exposed to diesel exhaust emissions. Br. J.Ind. Med. 42. 435-448. Wong, 0.. Morgan. R. W., Bailey, W.. Swencicki, R. E., Claxton, K.. and Kheifets, L. (1986). An epidemiologic study of petroleum retinery employees. Br. J. Ind. Med. -t3,6-17. Wong, 0 .(1987a).An industry-wide mortality study of chemical workers occupationally exposed to benzene. I. General results. Br. J.Ind. Med. 44,365-381. Wong, 0. (1987b). An industry-wide mortality study of chemical workers occupationally exposed to benzene. 11. Dose-response analyses. Br. J.Ind. Med. 44,382-395. Wong, 0..and Raabe, G. K. (1989).A critical review of cancer epidemiology in studies of petroleum industry employees. with a quantitative meta-analysis by cancer site. Am. J.Ind. Med. 13,283-310. Wong, 0.. and Raabe, G. K. (1990). Proper interpretation of metaanalysis in occupational epidemiologic studies: a response. Am. J. Ind. Med. 1 7 , '"73-276. Wong, O., Harris, F., and Smith. T . J. (1993). Health effects of gasoline exposure. 11. Mortality patrerns of distribution workers in the United States. Enuiron. Health Perspect. lOl(Suppl.6). 63-76. Wong, 0..and Trent, L. (1993).d -Vested Case-Control Study of Kidney Cancer, Leukemia and Multiple .Wyeloma in a Cohort of LandBased Terminal Workers Exposed to Gasoline in the Petroleum Industry. Technical report submitred to American Pecroleum Institute. Wongsrichanalai, C., Delzell, E., and Cole, P. (1989). ,Mortality from leukemia and other diseases among workers at a petroleum refinery. J. Occup. Med. 31,106-111. World Health Organization ( 1993). International Programme on Chemical Safety, Environmental Health Criteria 150: Benzene. World Health Organization. Geneva. Wyngaarden, J. B., Smith, L. H., and Bennett, J. C. (1992). Ceil Textbook of Medicine. 19th ed., Saunders, Philadelphia.