A retrospective cohort study of leukemia and other cancers in benzene workers.

A retrospective cohort study was carried out in 1982-1983 among 28,460 benzene-exposed workers (15,643 males, 12,817 females) from 233 factories and 28,257 control workers (16,621 males, 12,366 females) from 83 factories in 12 large cities in China. All-cause mortality was significantly higher among the exposed (265.46/100,000 person-years) than among the unexposed (139.06/100,000 person-years), as was mortality from all malignant neoplasms (123.21/100,000 versus 54.7/100,000, respectively). For certain cancers, increased mortality was noted among benzene-exposed males in comparison with that among unexposed males; the standardized mortality ratios (SMR) were elevated for leukemia (SMR = 5.74), lung cancer (SMR = 2.31), primary hepatocarcinoma (SMR = 1.12), and stomach cancer (SMR = 1.22). For females only leukemia occurred in excess among the exposed. Risk of leukemia rose as duration to exposure to benzene increased up to 15 years, and then declined with additional years of exposure. Leukemia occurred among some workers with as little as 6 to 10 ppm average exposure and 50 ppm-years (or possibly less) cumulative lifetime exposure (based on all available measurements for the exposed work units). Among the 30 leukemia cases identified in the exposed cohort, the proportion of subjects with acute lymphocytic leukemia was substantially lower and the proportion with acute nonlymphocytic leukemias was higher than in the general population. During 1972 to 1981, the annual incidence of leukemia ranged from 5.83 to 28.33 per 100,000 with higher rates occurring in the interval 1977 to 1981 than in the earlier years of the study period.(ABSTRACT TRUNCATED AT 250 WORDS)


Introduction
Clinical reports and epidemiologic studies have repeatedly linked benzene exposure with leukemia, primarily acute myelocytic and other types of acute nonlymphocytic leukemia (1)(2)(3). A retrospective cohort investigation carried out in the People's Republic of China during 1982 to 1983 among 28,460 benzene-exposed workers (15,643 males, 12,817 females) from 233 factories and 28,257 control workers (16,621 males, 12,336 females) from 83 factories in 12 large cities in China, provided additional evidence supporting the benzene-leukemia association (4,5). More detailed information about the relationship of average and cumulative lifetime benzene levels and duration of exposure with development of leukemia is presented. The types of leukemia that occurred among the exposed cohort are compared with the distribution of leukemia types among the general population. All-cause mortality as well as mortality from specific types of malignancies in addition to leukemia is discussed.

Subjects and Methods
Benzene-exposed subjects were identified in work units from 233 factories in 12 cities in China including Shanghai, Tianjin, Chengdu, Chongqing, Harbin, Shenyang, Jinzhou, Zhengzhou, Luoyang, Kaifeng, Guangzhou, and Nanchang. The primary types of factories or manufactur-ing processes involving benzene exposure were spray or brush painting, shoe manufacturing, synthetic rubber production, leather processing, organic chemical, and adhesive production facilities. An exposed subject had to be working in an operation involving benzene exposure for at least 6 months during the period January 1, 1972, through December 31, 1981. There were 28,460 workers (16,621 males, 12,336 females), with 178,556 person-years of follow-up among the exposed.
Subjects in the control group consisted of 28,257 workers (16,621 males, 12,336 females) employed 6 months or more during the same time interval as the cases, selected from 83 factories in the same eities. Controls were employed in the manufacture of mechanical instruments, textiles, and in clothing production and were not exposed to benzene or any other known occupational carcinogen.
The sex and age distribution was similar to those of benzene-exposed workers. For the unexposed group there were 199,201 person-years of follow-up.
Exposed and unexposed workers were followed-up from January 1, 1972, through December 31, 1981, using factory (monthly or annual) salary lists to obtain information about current status including continued employment, retirement, transfer, as well as vital status. Factory and hospital medical records were the source of information about history of benzene poisoning, aplastic anemia, leukemia, and other malignancies. Additional information about history of malignancies, malignant, and nonmalignant causes of death, as well as history of smoking was obtained from a number of possible sources. These included interviews with subjects, their next-ofkin; mailed questionnaires sent to family members of deceased subjects; letters of inquiry sent to families or factories for workers who had transferred jobs; and hospital death certificates or death registration forms kept on file at the subject's local neighborhood police station. For deceased subjects, date and cause of death, date of diagnosis of malignancies, and hospital of diagnosis were abstracted from death certificates and hospital records. Detailed information about histopathology, clinical, and laboratory parameters was obtained for all subjects with leukemia, aplastic anemia, and other hematopoietic and lymphatic malignancies.
Factory records and interviews with factory supervisors and subjects were the source of information about products manufactured, raw materials used, production processes, intermediates, and occupational history of subjects. To determine benzene levels, data were abstracted from factory records about benzene air level measurements, benzene concentration levels in solvents, paint, glue, adhesives, etc., and factories' records about environmental measures and protective devices used to reduce workers' benzene exposure.
All of the factories included in the benzene-exposed group had records detailing benzene concentrations in products used and air level measurements. Control factories were selected from among those with no evidence of benzene, radiation, or other carcinogenic exposures. Among the factories identified as having benzeneexposed work units, the participation rate of these facil-ities in the study was 99.2%, compared with 98.7% among non-exposed factories.
Criteria used to designate benzene poisoning were based on the "Principles of Diagnosis and Treatment of Benzene Poisoning" published in 1974 by the Ministry of Public Health (6). Cases of leukemia and aplastic anemia were diagnosed at hospital departments of internal medicine and hematology and were classified for study purposes as follows: a) grade I, II, bone marrow aspirate and/or biopsy, blood smears, and other histopathologic data used to confirm diagnosis, b) grade III,. diagnosis based on clinical evaluation, c) grade IV, diagnosis inferrred after death in the absence of clinical data. Among the accepted leukemia study cases, 96.7% met the criteria for grades I-II, whereas only 3.3% (1 case) were classified as grade III.
Data were collected at all collaborating factories, hospitals, district health stations, etc., using a standardized approach. All forms and data items were checked following abstraction, and missing values and errors were corrected. Inaccurate or incomplete data that could not be corrected were not included in the analyses.

Results
For males and females, age-adjusted all-cause mortality was significantly higher among the benzene-exposed than among control workers (Table 1). For most age groups evaluated, age-specific mortality was higher among exposed workers 25 years of age and older, except for males ages 60 to 69 and females ages 30 to 34 and 55 to 64 as shown in Table 1.
All-cancer mortality was significantly higher among exposed males under age 70 compared with unexposed males ( Table 2). For males aged 30 to 45 the cancer death rate was more than twice as high among the exposed than the unexposed, as seen in Table 2. The reverse was noted for females, with cancer mortality actually higher among the unexposed compared with the benzene-exposed group.
Significantly elevated standardized mortality ratios (SMRs) were noted among exposed compared with unexposed males for leukemia (SMR = 5.74) and lung cancer (SMR = 2.31), whereas elevated but nonsignificant SMRs were noted for cancers of the liver, stomach, esophagus, intestine, nasopharynx, and lymphosarcoma, as shown in Figure 1. With the exception of an elevated SMR for leukemia, no major differences were found between exposed and unexposed females (Fig. 1).
The distribution of cancer deaths by site, sex, and rank among benzene-exposed and control workers is shown in Table 3. The rank by cancer site on a national basis is also provided for comparison. For exposed subjects, leukemia is the fourth-ranked site for males and the first-ranked site for females, whereas nationally leukemia is only the seventh most common malignancy. For exposed males, lung cancer is first, and for exposed females second compared with its rank as the fifth most common site nationally. Table 1. Age-specific mortality from all causes by sex among benzene-exposed and control workers in 12 cities in China, 1972China, -1981

Males
Females Benzene-exposed Unexposed Benzene- *Age-adjusted all-cause mortality is significantly higher among the exposed compared with the unexposed (p < 0.01) for both males and females. year-s years deaths person-years) years deaths person-years) years deaths person-years) years deaths person-years) 15 *Age-adjusted all-cancer mortality is significantly higher (p < 0.01) among exposed compared with unexposed.  Table 3. The distribution of cancer deaths by site, sex, and rank among benzene-exposed and control workers in 12  Though not shown, leukemia mortality among male benzene-exposed smokers (91.1 per 100,000, unadjusted by age) was not higher than leukemia mortality among male exposed nonsmokers (128.2 per 100,000, unadjusted). However, lung cancer mortality in males was significantly higher among smokers in the benzene-exposed group than among smokers in this cohort as shown in Table 4. It is surprising to note that non-smokers in the benzene-exposed group had substantially higher lung cancer mortality than nonsmokers in the control group, as well as smokers (Table 4), though not statistically significant.

MALES
The relative risk of leukemia increased with longer duration of benzene exposure up to 15 years and then declined with additional years of exposure. Figure 2 represents an orthogonal polynomial curve graphed using the function of y = 15.024 + 3.67x -0.117 x2, where y is the risk of leukemia (mortality of leukemia in the benzene exposed group/mortality of leukemia in the control group) and x is the duration of benzene exposure in years.
For each of the 30 benzene-exposed leukemia cases, average and cumulative lifetime levels of exposure were estimated using all available measurements ( Table 5). The minimum average exposure was 6.5 mg/m3 and the lowest cumulative lifetime exposure was estimated to be 33.2 mg/m3 years, as shown in Table 5. It is also interesting to note that three cases had a lifetime cumulative ex- Association of the relative risk of leukemia with the duration of exposure to benzene among the benzene-exposed cohort from 12 cities in China, 1972China, -1981 posure of less than 50 mg/m3 year. The distribution of exposed and unexposed cases by type of leukemia is shown in Tables 6 and 7. As noted in Table 7, 66.6% of the exposed cases had a variant of acute Table 4. Risk of lung cancer for smokers and nonsmokers among male benzene-exposed and control workers from 12  nonlymphocytic leukemia compared with 50.1% of the leukemia cases from the general population. Only 10% of the cases among the exposed group had acute lymphocytic leukemia compared with 23.5% in the general population.
Although the data are shown, the incidence of leukemia was higher in the later 5 years of the study period (19.79/100,000) than in the earlier years (10. 79/100,000). However, the number of cases is too small to provide stable annual estimates.

Discussion
The substantially elevated risks for leukemia observed for both males and females in the benzene-exposed cohort confirm findings from earlier clinical reports and epidemiologic studies. In 1928, Delore and Borgomano first reported cases of benzene-caused leukemia in Italy (1). Vigliani and Saita later estimated a 20-fold higher risk of acute leukemia among printing shop workers and shoemakers in Milan and Pavia who were exposed to benzene (7). Among atomic bomb survivors in Hiroshima and Nagasaki, risk of leukemia was about 2.5-fold higher in those with a past history of exposure to benzene and medical X-rays than in those without such exposures (8). Aksoy et al. identified 34 leukemia cases diagnosed during 1967 to 1975 among 28,500 shoemakers in Istanbul, Tur-key, estimating an incidence of 13 per 100,000 compared with approximately 6 per 100,000 in the general population (9). An SMR of 5.6 for leukemia was noted in a retrospective cohort study of Pliofilm workers employed in factories in Akron, Ohio (10,11); for workers exposed to benzene more than 5 years, the SMR for leukemia was 21.
Although there have been case reports of other malignancies associated with benzene exposure (1), the current investigation is one of the first analytic studies demonstrating a significant excess of lung cancer and suggesting possible increases of stomach, liver, esophagus, intestine, and nasopharyngeal cancer, as well as lymphosarcoma. The excess risk of leukemia among female workers exposed to benzene has also not previously been reported.
This retrospective cohort study was also one of the first to evaluate the effect of smoking on cancer outcomes among benzene-exposed and unexposed workers. No effect was found for smoking on leukemia mortality. Risk of lung cancer was slightly higher among smokers than among nonsmokers in the benzene-exposed group. However, nonsmokers among the benzene-exposed had a significantly higher mortality than nonsmokers in the control group, suggesting that lung cancer was more strongly related to benzene exposure than to smoking. These surprising results require additional evaluation. Table 5. Description of leukemia cases among benzene-exposed workers by duration and years of benzene exposure, average estimated exposure level, and lifetime cumulative exposure.

Average Case
Duration of exposure (->) and benzene measurements, mg/mr3 estimated Cumulative identification exposure level, exposure, no. 1950-1959 1960-1964 1965-1969 1970-1974 1975-1979 1980-1982  The findings of leukemia risk subsequent to relatively low estimated average and cumulative lifetime benzene exposure supports a recent report by Rinsky et al. (12). However, these findings must be interpreted cautiously because the estimated average and cumulative lifetime benzene exposure levels are based on relatively few measurements. In addition, it would be useful to evaluate differences in average and cumulative benzene exposure between leukemia cases and non-cases randomly selected from the exposed cohort in order to more completely evaluate the relationship of benzene exposure with development of leukemia. This investigation is one of a few to evaluate the types of leukemia occurring among benzene-exposed and compare this distribution to that observed in a general population. The results support findings from other studies of an excess of acute nonlymphocytic leukemia among benzene-exposed subjects (1,7-11). Table 7. The distribution of 30 leukemia cases by type among benzene-exposed workers from 12 cities in China (1972China ( -1981  One of the somewhat surprising findings in this retrospective cohort study was the increase in all-cause mortality. Further evaluation is underway to identify the specific diseases primarily responsible for the excess, as well as possible reasons for a relative deficit of these conditions among the unexposed. Other reasons that might explain the excess of deaths due to all causes and all malignancies will also be examined in an expansion of the present study conducted in collaboration with the U.S. National Cancer Institute. Preliminary findings suggest a risk of leukemia, primarily acute nonlymphocytic in type, from relatively low average and cumulative benzene exposure. Risk of leukemia is elevated in females exposed to benzene, as well as males, although the former do not seem to develop other types of malignancies in excess following this exposure. In summary, the present study has suggested that other cancers, as well as some nonmalignant conditions, may be associated with benzene exposure. Future studies should evaluate more precisely and, in relation to nonleukemia cases, the relationship between levels of exposure and development of leukemia. In addition, it would be useful to determine which industries, occupations, and particular job titles are linked with higher risks. Finally, possible modifying factors affecting occurrence of leukemia (such as radiaton, medications, other occupational exposures, etc.) should be evaluated in nested case-control leukemia studies, undertaken within large benzene-exposed cohorts.