An expanded cohort study of cancer among benzene-exposed workers in China. Benzene Study Group.

An expanded cohort study of 74,828 benzene-exposed and 35,805 unexposed workers were followed during 1972 to 1987, based on a previous study in 12 cities in China. A small increase was observed in total cancer mortality among benzene-exposed compared with unexposed workers (relative risk [RR] = 1.2). Statistically significant excesses were noted for leukemia (RR = 2.3), malignant lymphoma (RR = 4.5), and lung cancer (RR = 1.4). When risks were evaluated by leukemia subtype, only acute myelogenous leukemia was significantly elevated (RR = 3.1), although nonsignificant excesses were also noted for chronic myelogenous leukemia (RR = 2.6) and acute lymphocytic leukemia (RR = 2.3). A significant excess was also found for aplastic anemia.


Introduction
Benzene has been widely used in chemical manufacturing and other industries for about a century. A nationwide survey carried out in 1979 to 1981 revealed that 529,000 workers were occupationally exposed to benzene or benzene-containing mixtures in China (1). Similarly, it has been estimated that 233,000 people are occupationally exposed to benzene in petrochemical plants, petroleum refineries, and other operations in the United States (2). In addition to occupational exposures, benzene concentrations varying from 1 to 150 mg/m3 have been described in urban and densely trafficked areas (3). Therefore, benzene is a potentially important This paper was presented at Benzene '95: An International Conference on the Toxicity, Carcinogenesis, and Epidemiology of Benzene held 17-20 June 1995  carcinogen not only to workers, but also to the general population (4).
The present study was initiated as a collaboration between the Institute of Occupational Medicine, China, and the U.S. National Cancer Institute (NCI), based on an earlier retrospective cohort study carried out by Chinese investigators among 28,480 benzene-exposed and 28,257 unexposed workers followed up from 1972 to 1981 in China (5). The major reasons for expanding upon the original cohort study were to include substantially more subjects, additional person-years of followup, and more leukemia cases, for analyzing in detail the relationship of benzene with subtype of leukemia and other hematolymphoproliferative malignancies. In this paper we summarize the mortality risks for all cancers and the incidence risks for all leukemia and other selected hematolymphoproliferative malignant and nonmalignant disorders (HLD). Further detail is provided in Yin et al. (6).

Study Population and Data Sources
The methods for this expanded study are described in detail elsewhere (7). Briefly, the study subjects were selected from workers employed in 672 factories in which benzene was used (exposed factories) compared to unexposed workers at 69 of these factories and at 40 additional factories (unexposed factories) in 12 cities in China (including Shanghai, Tianjin, Chengdu, Chongqing, Harbin, Shenyang, Jinzhou, Zhengzhou, Luoyang, Guangzho, Nanchang, Kaifeng) in China. Using monthly salary records, field center staff in each city identified all workers employed for any length of time in the eligible work units or factories during 1972 to 1987. If salary records were not available or incomplete for the eligible work units or factories during the study period, other data sources used in a standardized hierarchial approach included job registers (completed upon initiation of employment at a factory), health examination forms, and hazardous work compensation registries.

Data Coliection and Foliow-up Procedures
For each eligible study subject, data abstracted included the worker's name, birthdate, sex, job history in study factories, vital status on December 31, 1987, date last known alive if lost to follow up, cause(s) of death and date of death for deceased workers, history and date of diagnosis of leukemia or other HLD, and history and date of diagnosis of benzene poisoning.
Workers who left study factories before December 31, 1987, were followed up and traced to determine vital status, cause of death, and occurrence of leukemia and other HLD. For deceased subjects, cause of death was obtained from medical records, other written factory records, or death certificates at local police stations. The subject's physician or the next of kin was contacted only if written records listing cause of death could not be obtained. For all subjects whose cause of death was suspected to be cancer or HLD, hospital and other medical records were sought to verify the diagnosis. The specific histopathology, date of diagnosis, date of death, hospital/ place of death, source of medical information, and other important diagnosis-related information were abstracted on a standard diagnostic validation form and reviewed by the field center director and the investigators. For cases of HLD, medical records and all available pathology and peripheral blood smear slides were reviewed by hematopathologists from the Beijing Union Medical College Hospital, the NCI, and the Mayo Clinic in Rochester, Minnesota. The exposure assessment methods used to determine time period-specific quantitative estimates of benzene exposure by job title within work units are described elsewhere (8). For the statistical analysis, relative risks (RR) were derived as the ratio of age-and sex-adjusted rates in the exposed and unexposed groups.

Results
The expanded cohort consisted of 74,828 exposed and 35,805 unexposed persons in 781 factories and 12 cities (Table 1). They were followed up for an average of 10.5 and 11.7 years, respectively. Approximately 1.8% of exposed and 1.7% of unexposed workers were deceased, and 0.2% of exposed and 0.3% of unexposed workers were lost to follow-up (Table 2).
There were 1369 deceased in the exposed cohort and 598 deceased in the unexposed cohort.
Age-adjusted RR for cancer mortality for the exposed compared with the unexposed groups are shown in Table 3. The RR ranged from 0.4 to 2.8 for women and from 0.9 to 3.3 for men. RR for leukemia was 2.8 for women and 2.1 for men. RR for lung cancer was 1.5 for men and 1.0 for women.
Incidence risks for HLD were analyzed using the internal comparison to assess the difference between exposed and unexposed workers. Table 4 presents RR for the specific HLD. HLD, including leukemia, myelodysplastic syndromes, and lymphoma, were significantly elevated, with RR ranging from 2.5 to 3.1. The observed number of aplastic anemia and agranulocytosis cases were nine and two, respectively, in exposed workers, but there were no cases in unexposed workers.

Discussion
There have been several reports describing risks of leukemia, lymphoma, and other cancers in relation to occupational benzene exposure in several countries (9)(10)(11)(12)(13)(14)(15)(16)(17). Most of the studies have reported mortality excesses for all hematopoietic and lymphatic cancers. However, most of the investigations yielded small numbers in the Table 2. Vital status of benzene-exposed and nonexposed subjects (1972)(1973)(1974)(1975)(1976)(1977)(1978)(1979)(1980)(1981)(1982)(1983)(1984)(1985)(1986)(1987)   cohorts studied and few leukemia cases. While a few cohort studies evaluated more than 10,000 workers in rubber factories or oil companies, the findings were not significantly elevated for leukemia. Most of the studies usually compared the study population with the general population; few used an internal comparison cohort. This study was designed specifically to avoid the disadvantage of earlier investigations (18). To date, this study is the largest cohort mortality investigation, comprising 81 cases and 13 cases of HLD observed in exposed and unexposed groups, respectively. The distribution by sex, age, and calendar-year at first entry were similar for exposed and unexposed groups. Such an internal comparison group makes it possible to minimize the healthy worker effect.  (5). Follow-up 1972Follow-up -1981Follow-up 1972Follow-up -1987  Comparison of this cohort study with the previous cohort study in China (Table  5) reveals that this study has a longer duration of follow-up, a larger population, more leukemia cases, and an in-depth Environmental Health Perspectives * Vol 104, Supplement 6 * December 1996 examination of information. It will be important to continue to follow these workers, since substantial numbers, particularly those first employed after 1980, will only have experienced low exposure levels of benzene since start of employment.
The risk was significantly increased for the combined grouping of acute myelogenous leukemia and myelodysplastic syndromes, since the latter may precede overt leukemia. Risk was also increased for chromic myelogenous leukemia and acute lymphocytic leukemia, but not statistically significant. Risk was significantly increased for malignant lymphoma and for non-Hodgkin's lymphoma, but not for multiple myeloma. Mortality from lung cancer was significantly increased due to a significant excess among males, consistent with the results from a previous study in China.
Based on analysis of the incidence in the largest cohort, it was confirmed that benzene is a leukemogen, especially for myelogenous leukemia in humans, and benzene is possibly also associated with excess risk of non-Hodgkin's lymphoma and lung cancer among humans.