Document Q8ZkmnzkGJ8VLgKYxrkmyZ46

AMERICAN JOURNAL OF INDUSTRIAL MEDICINE 29:227-235 (1 996) A Cohort Study of Cancer Among Benzene-ExposedWorkers in China: Overall Results Song-Nian Yin, MD, Richard B. Hayes, PhD, Martha S. Linet, MD, Cui-Lan Li, PhD, Mustafa Dosemeci, PhD, Lois B. Travis, MD, Chin-Yang Li, MD, Zhi-Nan Zhang, MD, DeGao Li, MD, Wong-Ho Chow, PhD, Sholom Wacholder, PhD, YaoZu Wang, MD, Zhong-Lian Jiang, MD, Tan-Rong Dai, MD, Wan-You Zhang, MD, Xin-JieChao, MD, Pei-Zheng Ye, MD, Qing-Rui KOU, MD, Xi-Chun Zhang, MD, Xia-Fang Lin, MD, Juan-Fei Meng, MD, Cheng-Yu Ding, MD, Jie-ShengZho, MD, and William J. Blot, PhD A large cohort study of 74,828 beti-ene-e.rliosecl and 35,805 unexposed workers employed between 1972 and I987 in 12 cities in China werefidlowed to determine mortality from all causes and the incidence of I~tiipholietizat~~~~moiaeltiigcnancies and other lieniatologic disorders. Benzene-exposed studv subjects were employed iti ( I variety of occupations, including painting, printing, and the rnani~fiit:tureof .footwear, paint, arid other chemicals. All-cause niortcility rt9a.ssimilar in the benzene-exposed and unexposed comparison group. Statistically sigtivicant excess deaths were noted among benzene-exposed subjects for leukemici (RR = 2.3, 95% CI: 1.1-5.0). malignant lymphoma (RR = 4.5, 95% CI: 1.3-28.4). arid nonneoplastic diseases of the blood (RR = 00, 95% CI: 2.5-1, and a tnargincilly significant excess WNS noted for lung cancer (RR = 1.4, 95% CI: 1.e2.0). Risk was significantly elevated for the incidence of all lymphohemntc~p~~iemtiat~lignancies (RR = 2.6, 95% CI: 1.5-5.0). malignant lymphoma (RR = 3.5, 95% CI: 1.2-14.9), and leuketnia (RR = 2.6, 95% CI: 1.3-5.7). Among the leukernin subtypes. only acute tnvelogetious leukemia (AML) incidence was significmitly elevated (RR = 3.1, 95% CI: 1.2-10.71, although nonsignificant excesses were also noted ,jkr chronic niyelogenous leukemia (CML) (RR = 2.6, 95% CI: 0.7-16.9) and Iyniphocytic leukeniias (RR = 2.8, 95% CI: 0.5-54.5). Signi3cant excesses were found,for aplastic anemia (RR = m, 95% CI: 2 . 2 - 3 ) arid myelodvsplastic syndrome (RR = -, 95% CI: I.7--1. Employment in benzene-cissociated occupations in China is associated with N wide spectrum of' inyelogenous and lymphocytic malignant diseases and related disorders. Inrestigotions continue to assess the nature of these associations. *0 1996 Wile?.-Liss, 1tIC. KEY WORDS: benzene, leukemia, lymphoma, epidemiology Chinese Academy of Preventive Medicine, Institute of Occupational Medicine, Beijing, China (S.-N.Y., G.-L.L.). Epidemiology and Biostatistics Program, National Cancer Institute, Bethesda, MD (R.B.H., M.S.L., M.D., L.B.T., W.H.C., S.W., W.J.B.). Section of Hematopathology, Mayo Clinic, Rochester, MN (C.-Y.L.). Division of Hematology, Peking Union Medical Hospital, Chinese Academy of Medical Sciences, Beijing, China (2.-N.Z., 0.-G.L.). Stations of Public Health and Prevention of Infection, Shanghai (Y.-Z.W.) Chengdu (T.-R.D.), and Chonqing (X.-J.C.), China. Institutes of Labor Health and Occupational Disease, Tianjin (Z.-L.J.), Heilongjiang (P.-Z.Y.), Henan (X.-C.Z.), Shenyan (0.-R.K.), Sichuan (W.-U.Z.), Jiangxi (J.-F.M.), Nanchang (J.-S.Z.),Guangzhou (X.-F.L.), and Jinzhou (C.-Y.D.), China. Address reprint requests to Dr. Richard 6. Hayes, NIH/NCI EPN 418, Bethesda, MD 20892. Accepted for publication April 21, 1995. 0 1996Wiley-Liss, Inc. *This article is a US Governmentwork and, as such, is in the public domain in the United States of America. 2 2 8 Yin et al. INTRODUCTION Benzene is a solvcnt widely uccd in industry and found in cigarette smoke, gasoline. automobile emissions, and other products [Wallace. 19891.Bcn7ene expowre has been related in numerous occupational studics to increased risk of acute myeloid leukemia (AMI,) [IARC. 1987-1 and. in some reports. to other lyniphohematopoietic malignancies and solid tumors [Goldstcin. 1990: Young. 1989: Mehlman, 1991: Yin et al.. 1987al. I n 198 1 the Chinese Academy of Preventive Medicine (CAPM)carried out a national occupational survey i n which ovcr 500.000 workers exposed to benzene in China were identified lYin et al.. 1987bl. Subsequently. 28.460 of these exposed workers and 28.2S7 unexposed workers were followed for cancer mortality between 1972 and 1981 lYin et al.. 1987a. 19891. In that study. 25 leukemia deaths were identified among benzene-exposed workers vs. 4 leukemia deaths among controls, resulting in a mortality ratio of 5.7. Sincc 1987. the US. National Cancer Institute ( N C I ) has collaborated with the CAPM to expand the mortality study in cohort size, duration. and recency time of follow-up, and t o invcstigate the incidence of hematologic malignancies and related disorders. The NCI-CAPM collaborative study idcntified and followed a largc cohort of Chincsc benzene-exposed workers. determined their exposure histories. and characterized thcir risk for a variety of diseases potentially related to benzene exposure. Here we report on disease risk in this expanded cohort, providing an overview of all-cause mortality and of the incidence of lymphohematopoictic malipnncies and other hematologic disorders. METHODS The inethods used in this expanded study are descrihed in detail elsewhere [Yin ct al., 1994: Dosemeci el al., 19941. The benzene-exposed group was comprised of workers e n ployed betwecn 1972 and 1987 in 1.427 selected benzeneevposcd work units (departments) i n 672 factories i n 12 cilies in China. A variety of industries and occupations using benzene were studied, including painting, prinriiig, and the manufacture of footwear. paint. and other chemicals. An uncxpoced coinpariwn group was assembled from workers employed between 1972 and 1987 in work units where benzcne was not used in 69 o f these factories, and i n 40 additional factories. Subjects were identified from salary records und other factory written aclministrutive records. Wc absrrxted dcniographic data. including hirth date, and SCX, and occupational data. including the dates of employment, by work u n i t and job title. for all jobs held by subjects in the study factories. Benzenc-exposed jobs were determined hascd upon frictory level and job title-specific information on benzene use in seven calendar periods ( 1949-1959, I900-19hJ. I965-l960. 1970-1974. I97S-I979. 19801984. I 9XS+). Subjecls were followed up for history of selected lymphohematopoietic malignancies and other hematologic clisorders and for vital status to December 31. 1987 through factory personnel records at study factories and subsequent places of einployment, or. ;is needed. by contacting ncxt nf kin, work colleagues. treating physicians. or others. For deceased subjects, cause of tienth was obtained from employer medical records. other written factory records, or death certificates. Only after extensive search had failed to locate written records listing the cause of death were treating physicians or next of kin contacted. For cases newly diagnosed with lymphohematopoietic malignancies and other hematologic disorders during follow-up period, pcrtinent histopathologic material. pathology reports. and medical records were requested. For pathology review. clinical. laboratory, and pathologic data were abstracted onto standardized forms by physician investigators who wcrc not aware of the exposure status or thc number of exposed and unexposed cases. All histopathologic and bone marrow aspirate slides and peripheral blood siiieiirs were reviewed systematically, by expert hematopathologists affiliated with the Mayo Clinic (C.-Y.L.). NCI (L.B.T.),and Peking Union Hospital (Z.-N.Z. and D.-G.L.) using structured abstract forms to objectively characterix hematopoiesis. Diagnoses were assigned after evaluation of all available clinical, laboratory, and pathologic data. without knowledge of the patient's benzene exposure status. Published criteria were utilized to categorize leukemia and lymphoma cases. and where possible, to classify these cases and those with myelodysplasia (MDS) by subtype [Travis et al.. 19941. For selected analyses, AML and MDS were c o n bined as one disease category because of possible commonalities in the natural history of these conditions and failure in the past to consistently classify MDS as entities distinct from AML. For the statistical analysis, person-years were accumulated for the benzene-exposed workers from January I . 1972 or, if hired later, from the first date of employment in a benzene-exposed job. For the unexposed comparison group, person-years were accumulated from January I . 1972 or, if hired later. from the first date of employment. Initial comparisons for mortality were conducted relative to gcneral population mortality rates (standardized mortality ratio, SMR), derived from a population-based mortality survey in 1973-1975, which included 10 of the 12 study cities (National Cancer Control Office, Ministry of Health. Beijing. Pcople's Republic of China). Further detailed analyses for mortality due to all causes and for incidence of lymphohematopoietic malignancies and other hematologic disorders were made by internal comparison of disease rates in the benzene-exposed group t o the rates in the unexposed group, by Poisson regression analysis. yielding rate ratios Cancer A m o n g Benzene Workers 229 TABLE I. Distribution of Benzene-Exposed and Comparison Unexposed Study Subjects in 12 Cities in China, 1972-1987, by Selected Characteristics Exposed No. Person-year Total Male Female Employment (1949-1987) Age at first employmenta,b t20 20-29 230 Year at first employrnenPb 4960 1960-1 972 >1972 FoIIow-uP (1972-1 987) Age at entry to follow-up t20 20-29 230 Year at follow-up 1972-1 981 1982-1 987 Vital status (12/31/87) Alive Deceased Lost 74,828 38,833 35,995 24,531 32,935 17,362 11,306 22,140 41,382 19,850 31,335 23,643 73,312 1,369 147 782,497 414,043 368,455 248,719 334,320 199,458 168,971 326,673 286,853 158,941 305,874 317,682 378,517 403,979 (Yo) (100) (53) (47) Unexposed No. Person-year 35,805 20,795 15,010 419,523 250,503 169,1319 (Yo) (100) (60) (40) (32) 16,423 191,357 (43) 15,743 182,015 (25) 3,639 46,150 (21) 7,765 121,810 (42) 10,513 166,297 (37) 17.527 131,415 (46) (43) (11) (29) (40) 131) (20) 10,050 90,087 (39) 14,751 166,771 (41) 11,004 162,665 (48) 222,621 (52) 196,902 35,117 598 90 (21) (40) (39) (53) (47) 'Entry lo employment for exposed at dale of first exposed job. "Entry to employment for unexposed at dale of first lob. ( R R ) for exposed versus unexposed workers [Breslow and Day, 19871. Both external ( S M R ) and internal ( R R ) analyses were controlled for age and. where appropriate, for sex. Confidence intervals (C1, 95%) for the S M R were based on asymptotic methods [Breslow and Day. 19871 and, for the RR, were obtained from the profile likelihood [Preston et al., 1993: Moolgavkar and Venzon, 19871. Confidence intervals that exclude I .0 are considered statistically significant. RESULTS The study group consisted of 74,828 benzene-exposed and 35.805 unexposed workers (Table I). On average, benzene-exposed subjects were followed for 10.5 years, while unexposed subjects were followed for 11.7 years. Women contributed 47% of the person-years in the benzene-exposed study group and 40% in the unexposed group. Personyear distributions for the benzene-exposed and unexposed groups are also shown in Table I, with respect to age and year at first employment. age at entry to study. and year at risk. Overall, the study groups were young. with about 60% of the total person-years at risk being contributed by subjects less than 30 years of age at study entry. In this young study population. about 2% died during the follow-up period ( 1,369benzene-exposed and 598 unexposed). Only 147 exposed and 90 unexposed workers were lost to follow-up. Mortality in Benzene-ExposedWorkers In initial analyses compared to Chinese population mortality rates, the all-cause standardized mortality ratios (SMRs) were 0.5 (95% CI: 0.4-0.5) and 0.4 (95% CI: 0.40.5) for the benzene-exposed and unexposed study groups. respectively. The respective SMKs were 0.6 (95% CI: 0.50.6) and 0.5 (95%)C1: 0.4-0.5) for men. and 0.3 (95% CI: 0.3-0.4) and 0.3 (,95% CI: 0.3-0.4) for women. Analysis by specific causes of death also indicated substantial defi- 230 Y i n et al. TABLE II. Mortality Among Benzene-Exposed Compared With Unexposed Workers, China, 1972-1987 Cause of death (ICD 9th revision) All causes Infectious, parasitic (001-1 39) Tuberculosis (010-01 8) Malignant neoplasms (140-208) Metabolic disorders (240-279) Blood and blood production (280-289) Circulatory system (390-459) Respiratory system (460-51 9) Digestive system (520-579) Cirrhosis (571) Urinary tract (580-629) Pregnancy and puerperium (630-676) Occupational injury, poisoning Accident. non-occupational (800-999) Suicide (E950-E959) Other known causes Diagnosis not clear Women Obsa RRb (95% Cl)' 272 1.o 0.8-1.2 12 1.7 0.5-7 3 7 0.9 0.3-4.3 99 0.9 0.6-1 3 5 0.8 0.2-5.5 2 m 0.388 0.8 0.6-1.2 7 1.1 0.3-7.2 8 1.6 0.4-1 0.9 5 m 0.912 0.7 0.3-2.0 2 1.o 0.1-21.6 0 und 14 1.3 0.5-4.0 16 1.8 0.6-6.2 6 0.9 0.2-4.2 1 0.4 0.0-10.8 Obs 1097 35 21 425 7 8 388 47 65 53 15 14 50 11 15 17 Men RR 1.1 0.9 0.9 1.2 0.8 m 1.o 1.7 0.7 0.8 0.9 (95% CI) 1.o-1.3 0.5-1.6 0.4-1.8 1.o-1.5 0.2-3.1 2.00 . S l .3 1.O-3.4 0.5-1 .O 0.5-1.2 0.4-2.1 2.2 0.8-7.7 1.t 0.7-1.8 0.8 0.3-2.1 1.6 0.6-4.9 2.9 1.O-12.3 'Obs observed benzene-exposed cases "RR relative risk. compared with nonexposed workers, adlusted for age cCI confidence interval "RR relative risk, compared with nonexposed workers, adjusted for age and sex Total RRd (95% CI) 1.1 1.0-1.2 1.o 0.6-1.7 0.9 0.5-1.7 1.2 1.o-1.4 0.8 0.3-2.3 m 2.5- 1.o 09-1.2 1.6 0.9-3.1 0.7 0.5-1.1 0.9 0.6-1.3 0.8 0.4-1.6 2.2 0.8-7.7 1.1 0.7-1.8 1.2 0.6-2.4 1.3 0.6-3.2 2.2 0.8-7.8 cits in mortality in the benzene-exposed group relative to the general ( 1 0 city) population rates, except for lung cancer (125 deaths, SMR = 1.0, 95% CI:0.8-1.2), leukemia (38 deaths, SMR = 1.5, 95%- C1 = 1.1-2. I ) , and lymphoma (17 deaths, SMR = 1.2, 95% CI = 0.7-2.0). Because of the large differences between the external population mortality rates and the mortality experience in the study populations, further analyses consider only disease occurrence in the benzene-exposed group compared with the unexposed group. As shown in Table 11, all-cause mortality in the benzene-exposed group was similar to mortality in the unexposed comparison group (RR = 1.1 ). for both women (RR = 1.0)and men (RR = 1 . I ) . Mortality due to diseases of the circulatory (RR = 1.0) and digestive systems (RR = 0.7), and the urinary tract (RR = 0.8), was not in excess in the exposed group, while statistically nonsignificant excesses were noted for occupational injuries (RR = 2 . 2 ) ,nonmalignant diseases of the respiratory system (RR = I .6), largely due to an excess among men (RR = I .7), and for deaths due to uncertain causes (RR = 2.2). Benzene- exposed subjects had a small but statistically significant increased mortality due to malignant neoplasms (RR = 1.2). Mortality due to nonmalignant diseases of blood and blood production were significantly elevated (RR = 00. 95% CI: 2.5-w), based on 10 cases among the exposed and 0 among the unexposed. Site-specific cancer mortality results are shown i n Table 111. An excess of lung cancer was found among benzeneexposed workers ( R R = I .4), due to a significantly elevated risk among men (RR = l S ) , but not among women (RR = I.O, based on 16 cases). Fourteen cases of nasopharyngeal cancer were identified (RK = 2.4). Esophageal cancer also showed a nonsignificant excess (RR = 1.8). Significant excesses of mortality due to leukemia (RR = 2.3) and lymphoma (RR = 4.5) were found in benzene-exposed workers, with similar excesses for men and women. No deaths were noted from multiple myeloma (MM). Incidence of Lymphohematopoietic Malignancies and Other Hematologic Disorders in Benzene-ExposedWorkers Eighty-one incident cases of lymphohematopoietic malignancies (n = 63) and other hematologic disorders (n = 18) were documented among benzene-exposed workers, with 13 lymphohematopoietic malignancies and no other hematologic disorders identified among the unexposed comparison group [Travis et al., 19941. Of nine ANLL with sufficient information to classify by French-American-British (FAB) subtype, three cases were M2, four cases were M3, one case was M2 or M4, and one case was M4 or M5 (the latter two cases could not be further specified). One addi- Cancer Among Benzene Workers 23 1 TABLE 111. Cancer Mortality Among Benzene-Exposed Compared With Unexposed Workers, China, 1972-1987 Women Men Total Cause of death (ICD 9ih revision) Obs' RRb (95% cl)c Obs RR (95YOCI) RRd (95% CI) Malignant neoplasms (140-208) 99 0.9 0.6-1.3 425 1.2 1.o-1.5 1.2 1.0-1.4 Nasopharynx (147) 2 m 0.3- 12 2.1 0.7-9.3 2.4 0.8-10.5 Esophagus (150) Stomach (151) 2 0.8 0.1-16.7 14 1.o 0.4-2.8 25 2.0 0.9-5.4 71 0.9 0.6-1.4 1.8 0.8-4.5 0.9 0.7-1.4 Colon, rectum (153,154) 10 0.7 0.3-2.0 24 1.1 0.5-2.3 0.9 0.5-1.7 Liver and gall bladder (155,156) 8 0.4 0.2-1.3 Trachea, bronchus and lung (162) 16 1.o 0.4-2.9 101 1.3 0.9-1.9 109 1.5 1.o-2.2 1.2 0.8-1.6 1.4 1.0-2.0 Breast (174) 8 0.9 0.3-3.2 Cervix uteri (179-180) 1 0.2 0.0-2.0 Bladder (188) 2 m 0.2- 4 0.6 0.1-3.3 0.9 0.2-4.3 Brain tumor` (191,225) 6 m 1.1- 7 0.8 0.2-2.6 1.3 0.5-4.1 Malignant lymphoma and related disorders (200-202) 5 m 1.o- 12 3.3 0.9-21.6 4.5 1.3-28.4 Multiple myeloma (203) 0 und und 0 und 0.0-3.2 und 0.0-3.2 Leukemia (204-208) 13 2.8 0.8-1 7.6 25 2.1 1.0-5.3 2.3 1.I-5.0 Other malignant neoplasms` 16 0.6 0.3-1.3 42 0.8 0.5-1.4 0.8 0.5-1.2 lobs: obsewed benzene-exposed cases. bRR: relative risk, compared with nonexposed workers, adjusted for age. T I : confidence interval. relative risk, compared with nonexposed workers, adjusted for age and sex elncludes malignant (2 women, 0 men) and other and unspecified CNS tumors (4 women, 7 men). `Includes oral (2). larynx (3,other respiratory ( l ) ,bone (6),melanoma (3),uterine (2),ovarian (4),other urogenital (3), thyroid (2). and unspecified (14). und. = undetermined tional benzene-exposed subject, previously described on histopathologic review as exhibiting morphologic changes indicative of some type of "toxic effect" [Travis et al.. 19941, is not considered here, since classification was not possible. The relative risks in benzene-exposed workers for lymphohematopoietic malignancies and other hematologic disorders are shown in Table IV. Risk was significantly elevated for all lymphohematopoietic malignancies combined (RR = 2.6), malignant lymphoma (RR = 3.5). and leukemia (RR = 2.6). Among the leukemia subtypes, only AML was significantly elevated (RR = 3.I ) , although nonsignificant excesses were also noted for CML (RR = 2.6) and lymphocytic leukemia (RR = 2.8). the latter attributable to five cases of acute lymphocytic leukemia. Although point estimates of risk for the other nonmalignant hematologic disorders could not be established (because no cases were identified in unexposed workers), the lower bound of the confidence intervals indicates the statistical lower range for these estimates (Table IV). Significant excess risks are shown for all nonmalignant hematopoietic disorders combined (95% CI,,,,,: 4.8), aplastic anemia (95% Cllower2: .2), and myelodysplastic syndrome (95% CI lower: 1.7). The relative risk for the rubric of acute myeloid leukemia plus MDS was RR = 4.1 (95% CI: 1.6- 13.8, 30 exposed cases). The relative risk for all lymphohematopoietic malignancies and other hematologic disorders combined was 3.4. For all lymphocytic and histiocytic malignancies combined, the RR = 2.7 (95% CI: 1.1-8.1, 26 exposed cases) and for non-Hodgkin's lymphoma, the RR = 3.0. The RRs for nodal and extranodal non-Hodgkin's lymphoma were 4.0 (95% CI: 1.1-25.7, 15 exposed cases) and 1.0(95% CI: 0.1-22.3, 2 exposed cases), respectively. No cases of Hodgkin's disease were identified. DISCUSSION This large cohort study identified excess mortality and incidence of lymphohematopoietic malignancies and related disorders among benzene-exposed workers. Risk was increased for AML, consistent with U.S. cohort studies of benzene-exposed pliofilm [Infante et al., 1977; Rinsky et al., 19871 and chemical manufacturing workers [Ott et al., 1978; Bond et al., 19861. In an earlier report, we described similar excesses among men and women, within broad occupational categories [Li et al., 19941. Of the nine ANLL cases classified in our study by FAB subtype, three (or possibly four) were M2 and four were M3. One previous study of ANLL subtypes suggested an increased risk for 232 Yin et al. TABLE IV. Incidence of Lyrnphohematopoietic Malignancies and Other Hematologic Disorders Among Benzene-Exposed and Unexposed Workers, China, 1972-1 987 Diagnosis Exposed workers Unexposed workers RR' 95% Clb Lymphohematopoietic malignancies Malignant lymphoma Non-Hodgkin's lymphoma Multiple myeloma All leukemia Myeloid leukemia Acute myelogenous leukemia Chronic myelogenous leukemia Lymphocytic leukemia Acute lymphocytic leukemia Other NOS' Other hematologic disorders Agranulocytosis Aplastic anemia Myelodysplastic syndrome Total 63 20 17 1 42 32 23 9 5 5 5 18 2 9 7 81 13 2 6 1 5-5 0 3 3 5 12-149 3 3 0 1 0-13 0 1 0 4 0 0-10 7 9 2 6 1 3-5 7 6 3.0 1 3-7 9 4 3 1 1 2-10 7 2 2 6 0 7-16 9 1 2 8 0 5-54 5 1 2.8 0 5-54 5 2 1 3 0 3-9 2 0 D_i 4 0-0 m 0 3-0 m 2 2-0 ca 17- 13 3 4 19-63 'RR relative risk, compared with nonexposed workers, adjusted tor age and sex "CI corifiderice interval 'NOS = not otherwise sDecified FAB M 4 associated with benzene exposure [Crane et al.. 1992). I n another study, two benzene-exposed ANLL CiiSeS were classificd as M2 and M3. respectively [Vineis et al.. 19901. Others have shown an increased risk for FAB M2 associated with tobacco use [Sandler et al., 19931. We found an increased risk for MDS i n benzene-exposed workers, and the risk associated with benzene cxposure was greater for the combined rubric of AML plus MDS (KR = 4.1) than i t was for AML idone ( R R = 3.1). The recognition of this syndrome is relatively recent and has not routinely been considered i n evaluations of risk associated with benzene exposure. Early case studies among benzencexposed populations noted abnormalities in bone inarrow and periphcral blood consistent with MDS in some pancytopenic patients prior to the development of acute leukemia [Goguel et al., 1967a; Aksoy and Erdern. 1978; Van den Berghc et al.. 19791. More recently. cases of MDS exposed to solvents [Vineis et al., 19901and to benzene [Ciccone et al.. I9931 have been described, and a case-control study i n Great Britain [Farrow et id., 19891showed an association of MDS with a history of exposure to gasoline and diesel fumes or liquids. It is noteworthy that a niyelodysplastic phase precedes overt leukemia in the majority of AML related to treatnient with alkylating agents [Michels et al.. 1985; Pedersen-Bjergaard and Philip, 19871,suggesting that ii similar pathogenesis could occur with benzene. The etiology and natural history of MDS requires further study. In terms of other myelogenous leukemias, we found a suggestive increase in CML. which has been reported i n varying frequencies among earlier series of benzene-exposed cases [Browning, 1965; Vigliani and Forni. 1976: Goguel et al.. 1967h; Aksoy, 19851. Risk estimates. however. have generally not been available. except for the earlier report of findings on ;I subset of the present study lYin et al., 19891.Our study showed a signilkant excess ofaplastic anemia. which hac; been linked to benzene exposure in numerous other reports [Aksoy. 1989: IARC. 1982; Paci et al.. 19891. Considering malignancies of lymphoid origin, we found a significant excess of malignant lymphoma and related disorders, based on a relatively large number (20 cases) of cases among benzene-exposed workers, and we also idcntified five cases of ALL. but the excess risk for this condition was not stutistically significant. In the United States. A L L occurs most commonly among children and older adults. The exposed htudy subjects with ALL ranged in age from 2 2 to 4I years. The one unexposed case was age 49 at diagnosis. Only one case of multiple myelorna, one lyrnphoproliferative disorder, and none of CLL or Hodgkin's disease were identified in the exposed group. Our cross-sectional study of benzene-exposed workers [Rothrnan et al., I9951 showed that exposure to benzene can affect the level of all the major blood elements, but particularly lymphocytes. Although there are other reports of a possible link between benzenc exposure and malignancies of lymphoid origin [Young, I989j. includin? lyrnphocytic Cancer Among Benzene Workers 233 leukemia [Aksoy. 1980; Goguel et al.. I967b; Vigliani and Forni. 19761. non-Hodgkin's lymphonla [Wong. 1987;Blair et al.. 19931. and multiple myeloma [Rinsky et al., 1987; La Vecchia et al., 1989; Goldstein, 19901, a specific relationship has not consistently been established. One report [Aksoy et al.. 19741 also suggested an association with Hodgkin's disease, but we observed no cases in our study. In contrast to the mortality and incidence analyses for lymphohematopoietic malignancies and related disorders, only mortality data were collected for other diseases. Of the solid tumors, a marginally significant excess was seen for lung cancer, due to increased deaths among men. In assessing this finding. i t is noteworthy that tobacco use is frequent amon? Chinese men, but not among women. We, however, did not have information on tobacco use among the lung cancer cases. Nonsignificant excesses were also noted for nasopharyngeal and esophageal cancer mortality. Except for earlier findings on a subset of our study population [Yin et al.. 19891, excesses of lung cancer have not been reported from cohort studies of benzene-exposed workers. Benzene is a multipotent carcinogen in experimental studies [Maltoni et al., 1989: National Toxicology Program, 19861, suggesting that excesses of hematological and nonhematological tumors in humans could be causally related to exposure. Further investigations of benzene exposure and risk for solid tumors are planned. This study had several important characteristics. The cohort was large. including 74,828 benzene-exposed workers. Extensive information was collected concerning occupational history and exposure to benzene. which will be the subject of further reports. The principal results of this study were based upon disease occurrence in the benzene-exposed group vs. an "internal" unexposed group of industrial workers. thereby providing a more appropriate comparison group than the general Chinese population. Our initial analyses of mortality in the benzene-exposed group utilized "external" population mortality rates. but these rates were not available for all 12 cities, were restricted to the years 1973-1975, and employed diagnostic criteria that may have differed from those used in our follow-up. Thus the low SMRs with that external standard may be unreliable. In contrast. the all-cause mortality and deaths due to major disease groups were similar between the benzene-exposed and unexposed cohorts, suggesting that the "internal" comparison was more appropriate. In our assessment of the incidence of lymphohematopoietic malignancies and other hematologic disorders, diagnoses were confirmed by review by hematopathologists of pathology reports. medical records, and/or histopathologic material. Comparisons of diagnoses based only on reports or records vs. those based on tissue examination revealed good concordance, providing confidence i n the adequacy of the diagnoses [Travis et al., 19941. Completeness of ascertainment was thought to be high. with the low occupational mobility i n China facilitating follow-up of study subjects for disease outcome. Also, the results were likely not due to increased detection of incident cases in the benzene-exposed group, as excesses were also found for deaths due to lymphohematopoietic disease. In comparing this with other studies of cancer among benzene-exposed workers that have taken place in the United States and Europe, it should be noted that the occurrence of several of the lymphohematopoietic malignancies is less frequently diagnosed in China, particularly CLL and multiple myeloma, and that nasopharyngeal cancer is considerably more common [Parkin et al., 1992; Groves et al.. 19951. The differences in disease occurrence affect the power to assess risks, as reflected by the relative risk confidence intervals. but may also indicate ethnic differences in susceptibility to these diseases. The benzene-exposed subjects in this study were employed in a variety of occupations, including painting, printing, and the nianufacture of footwear, paint, and other chemicals. Aksoy [ 19891 has suggested that different types of benzene-induced lyniphohematopoietic conditions may be related to the intensity, per unit time, and the total extent of benzene exposure. Our recent investigation of hematotoxicity aniong currently exposed workers [Rothman et al., 199.51 and experimental studies [Henderson et al., 19921 suggests that the kinetics of benzene metabolism may also be influenced by the intensity of exposure. For some work settings. confounding exposures may also contribute to the observed benzene-associated risk [Checkoway et al., 19831. Analyses are underway to assess exposure-specific disease risk among occupational subgroups with varying intensity and temporal exposure to benzene, and with potential differences in confounding exposures. In summary, this study of benzene-exposed workers in China provides further support for the association of benzene exposure with an increased risk for myelogenous leukemia. The risk was strongest for AML, but an excess of CML was also noted. Risks were also markedly increased for AA and MDS, and for ALL and NHL, but not for multiple myeloma or Hodgkin's disease. Employment in benzene-associated occupations in China is associated with a wide spectrum of myelogenous and lymphoid malignant diseases, and related disorders. Investigations are continuing to assess the nature of these associations in relationship to various aspects of benzene exposure. REFERENCES Ahmy M ( 1980): Ditferent types of malignancies due to occupational exposure to benrene: A review of recent ohservations in Turkey. Environ Res 23:181-190. Aksoy M (1985): Benzene as a leukemogrnic and carcinogenic agent. Am J Ind Med 8:9-20. Ahsoy M ( 1989):Hematotoxicity mid carcinogenicity of'benzene. Environ Health Persp 82193-197. 234 Yin et al. A h o y M. Erdem S (1978): Follow-up study on the mortality arid the development of leukemia in 44 pancytopenic patients with chronic exposure to benLene. Blood 52:285 -1-92. Aksoy M. Erdem S, Dincol K, Hepyukwl T. Diilcol T (1974): Chronic exposure to benzene as :I possible contributory etiologic factor in Hodgkin`s disease. Blut 28:293-298. other cancer risk by major occupational group among workers exposed to benzene in China. I. Descriptive findings. J Occup Med 360:875-881. Maltoni C. Ciliberti A. Cotti G. Conti B. Belpoggi F (1989): Benzene. an expcriniental multipotential carcinogen: Results of the long-term bioassays performed at the Bologna Institute of Oncology. Environ Health Persp 82: I09- 1 24. Hlair A. Linos A. Stewart PA, Burmeister LF, Gibson R. Everett G. Schuinan L, Cantor KP ( 1993): Evaluation of risks for non-Hoilgkin`s lymphonia by occupation and industry exposures lrom a case-control study. Am J Ind Med 23:301-312. Bond GG. Mcl.aren EA. Baldwin CL. Cook RR (lW6): An update of mortality aniong chemical workers expo\ed to beliLKlie. Br J Ind Med 13:68.5-691. Breslow NE, Day NE (1987): "Statistical Methods in Chicer Reaearch. Vol. II. The Design and Analysis of Cohort Studies." Lyon, France: International Agency for Research on Cancer, IARC Sci Pub. No. 80. Browning E i 196.5): "Toxicity and Metabolism of Industrial Sol\;ents." New York: Elscvier, pp 3-65. Chcckoway H, Wilcosky T. Wolf P. Tyrolcr H 1984):An ewluation o l the associations of leukemia and rubber solvent exposures. Am J Ind Med 5:239-249. Ciccone ti, Mirabelli D. I.evis A. Gavarotti P. Rege-Cambrin G, Davico L. Vineis P ( 1993): Myeloid leukemia, and myelodyspla\tic syndromes: Chemical exposure, histologic subtype and cytogenctics ill a case-control study. Cancer Genet Cytogenet 68:135-1 39. Crane MM. Godwin JE. Aiiegcrs JF, Keating MJ (1992): Is histological \uhtype a markcr for environmental exposures i n acutc rnyelogenous leukcmia" Cancer Epidemiol Biomarkcrs Prev I : 183- 188. Doscmeci M. 1.i G-L, Hayes RB. Yin S-N, Linet M. Chow W-H. Wang Y-Z, Jiang Z-L. Dai T-R, Zhang W-U, Chao X-J, Ye P-2. Kou Q-R. Fan Y-H, Zhang X-C. Lin X-F, Meng J-F. 2110J-S, Wacholder S. Knellcr R. Blot WJ (1994). A cohort study among workers cxposed to benzene in Chinti. II Exposure assessment. Am J Ind Med ?6:401-41 I. Mehlman MA (199 I ): Benzene health effects: Unanswered questions still not addressed. Am J Ind Med 20:71)7-7 I I . hfichels SD. McKenna RW. Arthur DC. Brunning RD (198.5): Therapyrclated acute myeloid leukemia and niyelodysplastic syndrome: A clinical and morphologic study o i 65 cases. Blood 6.5:1364-1 372. Moolgavkar SH. Venron DJ (1987): Confidence regions in curved exponential familieh: Application to matched case-control and survival studies with the general relative risk function. Ann Stat I5:346-359. National Toxicology Program ( 1986): "NTP Technical Report on the Toxicology and Carcinogenesis Studies of Benrene (CAS No. 71-43-2) in NF343M Rats and B6C3F1 Mice (Gavage Studies)". NIH Publication No. 86-2545. Rchearch Triangle Park. N.C. U S . Department of Health and Human Services, pp 11-101. Ott MG. Townsend JC, Fishbcck WA. Langner RA (1978): Mortality among individuals occupationally cxposed to benzene. Arch Environ Health 33:3-10. Paci E. Buiatti E. Costantini AS. Miligi L, Pucci N, Scarpelli A , Pcrrioli G. Simonato L. Winkelmann R, Kaldor JM ( 1989): Aplastic anemia. leukemia and other caiicei ~nor~alitiyn a cohort of shoe workers exposed to benzene. Scaiid J Work Environ Health I 5:313-3 18. Parhin DM, Muir CS. Whelan SL. Gao Y-T, Ferlay J , Powel J (1992): "Cancer Incidence i n Five Continents, Volume VI." Lyon. France: IARC Sci. Puhl. No. 110. Pedersen-Bjergaard J. Philip P ( 1987): Cytogenetic characteristics of therapy-related acute nonlymphocytic leukacniia. preleukaemia and acute myeloprolifeerative syndrome: Correlation with clinical data for 61 cotiscctive cases. Br Hsematol 66:199-207. Farrow A, Jacobs A. West RR (1989): Myeludysplahia, chemical exposure. Preston DL, Luhin JH, Pierce DA. McConney ME ( 1993): "Epicure User's and other environnicntal lactors. Leukemia 3 3 3 - 3 5 . Guide." Seattle. WA: Microsoft International Corporation. Goguel A. Cavigneaux A. Bernard J (1967a): Les Icucimies benzeniques Rull lnst Nail SantC Rech Med 22:421-441. Goguel PA, Cavigneaux A. Bernard J (1967b): Les l e u c h i e s bcnziniques de la region parlairline. Nouv Rev Fr Hematol 7:46.5--180. Golilstein BD (1990): Is exposure to benLenc a cause of human niultiple myeloma? Ann N Y Acad Sci 6W225-334. Groves FD. Linct MS, Devesa SS (199.5): Patterns of occurrence of the Icukernias. Eur J Cancer 31A:941-949. Henderson RF. Sahourin PJ, Medinsky MA, Birnbaum LS, Lucier GL ( 1992): Benzene dosimetry in experimental animal\: Relevance for risk ilwessiiieiit. Prvg Cliri Biol Res 37493-105. Rinsky RA. Smith AB. Hornung R. Filloon TG. Young RJ, Okun AH. Landrigan PJ (1987): Benzene and leukcmia: An epidemiologic risk assessment. N Engl J Mcd 316:1034-1050. Rothman N. Li G-L. Dosenieci M. Bechtold W. Marti G, Wang Y-2, Linet M, Xi L-Q. Lu W. Smith MT, Titenko-Holland N . Zhang L-P. Blot W. Yin S-N. Hayes RE (199.5): Hematotoxicity among workers heavily exposed to benzene. Am J Ind Med 29:236-246. Sandler DP, Shore DL. Anderson JR. Davey FR. Arthur D. Maycr RJ, Silver RT. Weiss RB, Moore JO. Schiffer CA, Wurster-Hill DH. McIntyre OR. Bloomlield CD (1993): Cigarette smoking and risk of acute leukemia: Associations with moi-phology and cytogenetic abnormalities i n boric marrow. J Natl Cancer Inst 58:1994-2003. IARC (19821: Benzene. In "IARC Monographs on the Evalustion of the Carcinogenic Risk of Chemicals to Humans: Some Industrial Chemicals and Dyestuffs." Vol. 29. Lyon. France: Internationid Agency I`or Research on Cancer. pp 93-14, Travis LB. Li CY. Zhi ZN. Li DG. Yin SN. Chow WH, Li GL. Dosenieci M. Blot W. Fraumeni J r JF. Hayes RB. Linet MS (1994): Hernatopoietic malignancie\ and related disorders among benzene-exposed workers in China. Lcuk Lymphoma 14:91-102. Infante PF. Rinsky RA. Wagoner JK. Young RJ (1977): Leukaemia in benzene workers. Lancet 2:76-78. La Vecchia C. Negri E. D'Avanzo B, Franceschi S (1989): Occupation and lymphoid neoplasms. Br J Cancer 6k38.5-388. 1.i GL. Linet MS, Hayes RB. Yin SN. Downeci M. Wang YZ. Chow WH. Jiang ZL. Wacholder S . Zhang WU, Dai TR. Chao XJ, Zhang XC. Ye PZ. Kou QR. Meng JF. Zho JS. Lin XF, Ding CY. Wu C . Blot WJ (1994): Gender diffrrences in hematopoietic and lyinpIiop~-oliieriltivd~isorders and Van den Berghe H, Louwagie A. Broeckaert-Van Orshoven A. David G . Verwilghen R (1979): Chromosome analysis in two unusual malignant hlood disorders prcsumably induced by benaene. Blood S3:S58 -566. Vigliani EC,Porni A (1976): Benzcne and leukemia. Environ Res 11: 122127. Vineis P. Avanzi GC. Giovinazzo B. Ponzio G, Cambrin GR. Ciccone G (1990):Cytogenetics and occupational exposure to solvents: A pilot study on leukemias and myelodysplastic disorders. Tumori 76:350-352. Cancer Among Benzene Workers 235 Wallace LA (1989): Major sources of benzene exposure. Environ Health Perspect 87:165- 169. Wong 0 (1987): An industry wide mortality study of chemical workers occupationally exposed to henzene. 1. General results. Br J Ind Med 44: 365-381. Yin S. Li G. Tain F. Fu Z. Jin C. Chen Y. Luo S, Ye P. Zhang J. Wang G. Zhang X. Wu H. Zhong Q (1987a): Leukaemia in henzene workers: A retrospective cohort study. Br J Ind Med 44:124-128. Yin S-N. Li Q, Liu Y, Tian F, Du C, Jin C (1987h): Occupational exposure to ben7ene in China. Br J Ind Med 44:192-195. Yin S. Li G.Tain F, Fu 2,Jin C. Chen Y. Luo S. Ye P. Zhang J. Wang G. Zhang X. Wu H, Zhong Q (1989): A retrospective cohort study of leukemia and other cancers in hemene workers. Environ Health Perspect 82207213. Yin SN. Linet MS, Hayes RB. Li GL, Dosemeci M. Wang YZ, Chow WH. Jiang ZL. Wacholder S. Zhang WU, Dai TR, Chao XJ, Zhang XC, Ye PZ, Kou QR. Meng IF. Zho .IS. Lin XF. Ding CY. Kneller R, Blot WJ (1994): Cohort study among workers exposed to benzene in China. 1. General methods and resources. Am J Ind Med 26:383-400. Young N (1989): Benzene and lymphoma. Am J Ind Med 15:495-498.