The Upper Midwest Health Study: a case-control study of primary intracranial gliomas in farm and rural residents.

Since several studies indicated that farmers and agricultural workers had an excess risk of brain cancer, the National Institute for Occupational Safety and Health initiated the Upper Midwest Health Study to examine risk of intracranial glioma in the non-metropolitan population. This population-based, case-control study evaluated associations between gliomas and rural and farm exposures among adults (ages 18 to 80) in four upper midwestern states (Iowa, Michigan, Minnesota, Wisconsin). At diagnosis/selection, participants lived in non-metropolitan counties where the largest population center had fewer than 250,000 residents. Cases were diagnosed 1 January 1995 through 31 January 1997. Over 90% of 873 eligible ascertained cases and over 70% of 1670 eligible controls consented to participate. Participants and nonparticipants, evaluated for "critical questions" on main and refusant questionnaires, differed significantly in farming and occupational experience, ethnicity, education, and lifestyle. The 1,175 controls were more likely than the 798 cases to have reported ever drinking alcohol (77% vs. 73%, adjusted odds ratio (OR) 0. 73, 95% confidence interval (CI) 0.59-0.92) and having had panoramic dental x-rays (34% vs. 29%, OR 0. 75, CI 0.61-0.92). Controls spent a greater percentage of their lives in non-metropolitan counties (78% vs. 75%, OR 0.81, CI 0.67-1.09). Among ever-farmers, controls were more likely to have had exposure to farm insecticides (57% vs. 50%, OR 0.75, CI 0.59-0.95) and farm animals (96% vs. 91%, OR 0.48, CI 0.25-0.90). Moving to a farm as an adolescent (ages 11 to 20) vs. as an adult was associated with a greater risk of glioma. In our study sample, farm or rural residence and summary farm exposures were associated with decreased glioma risk. However, nonparticipation by never-farming eligible controls could have affected results. Comparisons of farm chemical exposures may clarify associations between farming and glioma that others have reported.

Agricultural pesticide use and risk of glioma in Nebraska, United States.

BACKGROUND: To evaluate the risk of the adult glioma associated with farming and agricultural pesticide use, the authors conducted a population based case control study in eastern Nebraska. METHODS: Telephone interviews were conducted with men and women diagnosed with gliomas (n = 251) between 1988 and 1993 and controls (n = 498) randomly selected from the same geographical area. Unconditional logistic regression was used to calculate adjusted odds ratios (ORs) for farming and for use of individual and chemical classes of insecticides and herbicides, including pesticides classified as nitrosatable (able to form N-nitroso compounds upon reaction with nitrite). Non-farmers were used as the reference category for all analyses. RESULTS: Among men, ever living or working on a farm and duration of farming were associated with significantly increased risks of glioma (> or =55 years on a farm OR = 3.9, 95% CI 1.8 to 8.6); however, positive findings were limited to proxy respondents. Among women, there were no positive associations with farming activities among self or proxy respondents. Specific pesticide families and individual pesticides were associated with significantly increased risks among male farmers; however, most of the positive associations were limited to proxy respondents. For two herbicides and three insecticides, use was positively associated with risk among both self and proxy respondents. Based on a small number of exposed cases, ORs were significantly increased for the herbicides metribuzin (OR = 3.4, 95% CI 1.2 to 9.7) and paraquat (OR = 11.1, 95% CI 1.2 to 101), and for the insecticides bufencarb (OR = 18.9, 95% CI 1.9 to 187), chlorpyrifos (OR = 22.6, 95% CI 2.7 to 191), and coumaphos (OR = 5.9, 95% CI 1.1 to 32). CONCLUSION: The authors found significant associations between some specific agricultural pesticide exposures and the risk of glioma among male farmers but not among female farmers in Nebraska; however, most of the positive associations were limited to proxy respondents. These findings warrant further evaluation in prospective cohort studies where issues of recall bias are not a concern.

Agricultural pesticide use and adenocarcinomas of the stomach and oesophagus.

AIMS: To evaluate the risk of the stomach and oesophageal adenocarcinomas associated with farming and agricultural pesticide use. METHODS: Population based case-control study in eastern Nebraska. Telephone interviews were conducted with men and women diagnosed with adenocarcinoma of the stomach (n = 170) or oesophagus (n = 137) between 1988 and 1993, and controls (n = 502) randomly selected from the same geographical area. Unconditional logistic regression was used to calculate adjusted odds ratios (ORs) for farming and for use of individual and chemical classes of insecticides and herbicides, including pesticides classified as nitrosatable (able to form N-nitroso compounds on reaction with nitrite). Non-farmers were used as the reference category for all analyses. RESULTS: Ever living or working on a farm, duration of farming, and size of the farm were not associated with stomach or oesophageal adenocarcinomas. There was no association for either cancer with ever-use of insecticides (stomach OR 0.9, 95% CI 0.6 to 1.4; oesophagus OR 0.7, 95% CI 0.4 to 1.1) or herbicides (stomach OR 0.9, 95% CI 0.5 to 1.4; oesophagus OR 0.7, 95% CI 0.4 to 1.2). Likewise, individual pesticides, including individual nitrosatable pesticides, were not significantly associated with risk. CONCLUSIONS: No significant associations were found between specific agricultural pesticide exposures and the risk of stomach or oesophageal adenocarcinomas among Nebraska farmers.

Occupation and the risk of adult glioma in the United States.

OBJECTIVE: Previous studies have observed increased glioma incidence associated with employment in the petroleum and electrical industries, and in farming. Several other occupations have also been associated with increased risk, but with inconsistent results. We evaluated associations between occupational title and glioma incidence in adults. METHODS: Cases were 489 patients with glioma diagnosed from 1994 to 1998 at three United States hospitals. Controls were 799 patients admitted to the same hospitals for non-malignant conditions. An experienced industrial hygienist grouped occupations that were expected to have similar tasks and exposures. The risk of adult glioma was evaluated for those subjects who ever worked in an occupational group for at least six months, those who worked longer than five years in the occupation, and those with more than ten years latency since starting work in the occupation. RESULTS: Several occupational groups were associated with increased glioma incidence for having ever worked in the occupation, including butchers and meat cutters (odds ratio [OR] = 2.4; 95% confidence limits [CL]: 1.0, 6.0), computer programmers and analysts (OR = 2.0; 95% CL: 1.0, 3.8), electricians (OR = 1.8; 95% CL: 0.8, 4.1), general farmers and farmworkers (OR = 2.5; 95% CL: 1.4, 4.7), inspectors, checkers, examiners, graders, and testers (OR = 1.5; 95% CL: 0.8, 2.7), investigators, examiners, adjustors, and appraisers (OR = 1.7; 95% CL: 0.8, 3.7), physicians and physician assistants (OR = 2.4; 95% CL: 0.8, 7.2), and store managers (OR = 1.6; 95% CL: 0.8, 3.1), whereas occupation as a childcare worker was associated with decreased glioma incidence (OR = 0.4; 95% CL: 0.2, 0.9). These associations generally persisted when the subjects worked longer than five years in the occupation, and for those with more than ten years latency since starting to work in the occupation. CONCLUSIONS: This is our first analysis of occupation and will guide future exposure-specific assessments.

Cohort mortality study of Philadelphia firefighters.

BACKGROUND: Fire fighters are exposed to a wide variety of toxic chemicals. Previous studies have reported excess risk of some cancers but have been limited by small numbers or little information on employment characteristics. METHODS: We conducted a retrospective cohort mortality study among 7,789 Philadelphia firefighters employed between 1925 and 1986. For each cause of death, the standardized mortality ratios (SMRs) and 95% confidence intervals were estimated. We also compared mortality among groups of firefighters defined by the estimated number of career runs and potential for diesel exposure. RESULTS: In comparison with U.S. white men, the firefighters had similar mortality from all causes of death combined (SMR = 0.96) and all cancers (SMR = 1.10). There were statistically significant deficits of deaths from nervous system diseases (SMR = 0.47), cerebrovascular diseases (SMR = 0.83), respiratory diseases (SMR = 0.67), genitourinary diseases (SMR = 0.54), all accidents (SMR = 0.72), and suicide (SMR = 0.66). Statistically significant excess risks were observed for colon cancer (SMR = 1.51) and ischemic heart disease (SMR = 1.09). The risks of mortality from colon cancer (SMR = 1.68), kidney cancer (SMR = 2.20), non-Hodgkin's lymphoma (SMR = 1.72), multiple myeloma (SMR = 2.31), and benign neoplasms (SMR = 2.54) were increased among firefighters with at least 20 years of service. CONCLUSIONS: Our study found no significant increase in overall mortality among Philadelphia firefighters. However, we observed increased mortality for cancers of the colon and kidney, non-Hodgkin's lymphoma and multiple myeloma. There was insufficient follow-up since the introduction of diesel equipment to adequately assess risk.

Cancer surveillance series [corrected]: brain and other central nervous system cancers: recent trends in incidence and mortality.

BACKGROUND: During the 1980s, the incidence of primary malignant brain and other central nervous system tumors (hereafter called brain cancer) was reported to be increasing among all age groups in the United States, while mortality was declining for persons younger than 65 years. We analyzed these data to provide updates on incidence and mortality trends for brain cancer in the United States and to examine these patterns in search of their causes. METHODS: Data on incidence, overall and according to histology and anatomic site, and on relative survival were obtained from the Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute for 1975 through 1995. Mortality data were obtained from the National Center for Health Statistics. Medicare procedure claims from the National Cancer Institute's SEER-Medicare database were used for imaging trends. Statistically significant changes in incidence trends were identified, and annual percent changes were computed for log linear models. RESULTS/CONCLUSIONS: Rates stabilized for all age groups during the most recent period for which SEER data were available, except for the group containing individuals 85 years of age or older. Mortality trends continued to decline for the younger age groups, and the steep increases in mortality seen in the past for the elderly slowed substantially. Patterns differed by age group according to the site and grade of tumors between younger and older patients. During the last decade, use of computed tomography scans was relatively stable for those 65-74 years old but increased among those 85 years old or older. IMPLICATIONS: Improvements in diagnosis and changes in the diagnosis and treatment of elderly patients provide likely explanations for the observed patterns in brain cancer trends.

Occupational risk factors for cancer of the central nervous system (CNS) among US women.

BACKGROUND: In a recent report, we found an elevated risk of cancer of the central nervous system (CNS) in several occupations and industries, and a modest association with exposure to solvents and to contact with the public. METHODS: To further explore the occupational risk of CNS cancer among women, we extended the analysis of the previous death certificate-based case-control study, including 12,980 female cases (ICD-9 codes 191 and 192) in 24 US states in 1984-1992 and 51,920 female controls who died from diseases other than malignancies and neurological disorders. We applied newly designed job-exposure matrices for 11 occupational hazards, previously reported as brain cancer risk factors, to the occupation and industry codes in the death certificates. We also conducted a separate analysis of 161 meningioma cases (ICD-9 codes 192.1 and 192.3), a tumor more frequent among women, particularly in the postmenopausal age group. RESULTS: Overall, CNS cancer risk showed a 20-30% increase among women exposed to electromagnetic fields (EMF), methylene chloride, insecticides and fungicides, and contact with the public. Risk for meningioma was elevated among women exposed to lead (OR = 1.9; 95% CI 1.0, 3.9). CNS cancer did not show a clear pattern of risk increase by probability and intensity of exposure to any of the explored risk factors. Cross-classification by probability and intensity of exposure did not reveal any significant trend. Cases were too few to explore trends of meningioma by probability and intensity of exposure to lead. CONCLUSIONS: We did not find evidence of a strong contribution of 11 occupational hazards to the etiology of CNS cancer. However, limitations of the occupational information might have reduced our ability to detect clear patterns of risk.

Brain cancer and occupational exposure to lead.

A recent report in the literature suggested a link between occupational exposure to lead and brain cancer. To explore the hypothesis, we applied a job-exposure matrix for lead to the occupation and industry codes given on the death certificate of 27,060 brain cancer cases and 108,240 controls who died of non-malignant diseases in 24 US states in 1984-1992. Brain cancer risk increased by probability of exposure to lead among Caucasian men and women with high-level exposure, with a significant twofold excess among Caucasian men with high probability and high level of exposure to lead (odds ratio = 2.1; 95% confidence interval, 1.1-4.0). Risks were also elevated in the low- and medium-probability cells for African-American men with high-level exposure to lead. Trend by intensity level was statistically significant among African-American men (all probabilities combined). Although exposure assessment was based solely on the occupation and industry reported on the death certificate, these results add to other epidemiologic and experimental findings in lending some support to the hypothesis of an association between occupational exposure to lead and brain cancer risk. Analytic studies are warranted to further test this hypothesis.

Occupational risk factors for cancer of the central nervous system: a case-control study on death certificates from 24 U.S. states.

The risk of cancer of the central nervous system (CNS) by industry and occupation was investigated with a case-control analysis of the death certificates of 28,416 cases and 113,664 controls, selected from over 4.5 million deaths in 24 U.S. states between 1984 and 1992. Industries showing consistent increases in risk by gender and race included textile mills, paper mills, printing and publishing industries, petroleum refining, motor vehicles manufacturing, telephone and electric utilities, department stores, health care services, elementary and secondary schools, and colleges and universities. CNS cancer risk was increased for administrators in education and related fields, secondary school teachers, and other education- and health-related occupations. The application of job-exposure matrices to the industry/occupation combinations revealed a modest increase in risk for potential contact with the public at work and exposure to solvents. Occupational exposure to electromagnetic fields (EMF) was not associated with CNS cancer, although an association was observed with a few EMF-related occupations and industries. Agricultural exposures were associated with significant risk increases among white women and white men. Further work is required to investigate in more detail specific occupational exposures or possible confounders responsible for the observed associations.

Questionnaires for collecting detailed occupational information for community-based case control studies.

In case control studies, collection of information on the workplace environment has generally been limited to self-reports of exposures or to job title, type of employer, and the dates the jobs were held, supplemented sometimes by work activities. This information, however, may be insufficient to assess the potential and level of exposure accurately due to recall difficulties and the variability of exposures within a job. A solution to this problem is to use job-specific questionnaires. The organization of a series of such questionnaires developed for a case control study of brain tumors is described. Three types of questionnaires, or modules, were developed, task-based, industry-based, and modules based on jobs with nonspecific types of tasks (e.g., laborer). The format of these modules starts with questions on the general work environment (type of employer) and proceeds to questions on tasks. More detailed information is then gathered on materials and equipment used, sensory descriptions, dermal exposure, work practices, engineering controls, and personal protective equipment use. The questionnaires cover a wide variety of exposures including solvents, heavy metals, polycyclic aromatic hydrocarbons, machining fluids, electromagnetic frequency fields, and many other exposures and, therefore, can be used in other case control studies.

Risk of adenocarcinoma of the stomach and esophagus with meat cooking method and doneness preference.

Meats cooked at high temperatures (frying, grilling) and for a long duration contain heterocyclic amines (HCAs), which are both mutagens and animal carcinogens. Additionally, barbecuing/grilling of meats produces polycyclic aromatic hydrocarbons (PAHs). Consumption of well-done meat has been associated with an increased risk of colon cancer but has not been evaluated as a risk factor for stomach or esophageal cancers. We conducted a population-based case-control study in 66 counties of eastern Nebraska. Telephone interviews were conducted with white men and women diagnosed with adenocarcinoma of the stomach (n = 176) and esophagus (n = 143) between July 1988 and June 1993 and 502 controls. The dietary assessment included several questions about usual cooking methods for meats and doneness preference for beef. High intake of red meat was associated with increased risks for both stomach and esophageal cancers. Overall, broiling or frying of beef, chicken or pork was not associated with the risk of these tumors. Barbecuing/grilling, reported as the usual cooking method for a small number of study participants, was associated with an elevated risk of stomach and esophageal cancers. After excluding those who reported usually barbecuing/grilling, a source of both PAHs and HCAs, we evaluated doneness level as a surrogate for HCA exposure. Compared to a preference for rare/medium rare beef, odds ratios were 2.4 for medium, 2.4 for medium well and 3.2 for well done, a significant positive trend. Doneness level was not associated with a significant trend in risk of esophageal cancer.

A novel approach to data collection in a case-control study of cancer and occupational exposures.

BACKGROUND: In community and hospital-based case-control studies, the occupational data collected in interviews are usually limited to responses to general questions asked of all study subjects. A procedure is described in which more detailed information can be collected in an efficient, standardized and systematic way. METHODS: A generic work history is initially collected from all subjects using a computer-assisted interview. The work history includes job title, type of business, job activities, materials and chemicals, and tools and equipment used. After responses are entered into the computer by the interviewer, the computer searches a synonym file to identify possible job-specific modules relevant to the reported job. The modules are detailed questionnaires that address specific jobs administered after obtaining the generic work history. The modules are used to ask questions about the work environment; sources of exposure; factors affecting the movement of the agent from the source to the subject, such as local exhaust ventilation; and individual and job characteristics. After the interview is completed, the work history and responses to the modules are sent electronically to an industrial hygienist who reviews the information using a custom-designed software package. Where ambiguities or contradictions occur in information reported by the respondent, or for jobs for which no module had been developed, the industrial hygienist generates up to 10 additional questions per job. These questions are sent back to the interviewer for administration of a short, second interview. CONCLUSIONS: These procedures, which are being successfully implemented in an on-going case-control study of brain tumours, should improve disease risk estimates over those derived from more traditional approaches to exposure assessment.

Mesothelioma, asbestos, and reported history of cancer in first-degree relatives.

BACKGROUND: Although malignant mesothelioma is known to be strongly related to asbestos exposure, its relationship to familial factors is unclear. METHODS: We compared reported histories of cancer in first-degree relatives, obtained from telephone interviews with the next-of-kin of 196 patients who had a pathologic diagnosis of mesothelioma, and with those from 511 decreased controls. RESULTS: Among men exposed to asbestos, we found a statistically significant twofold elevation in the risk of mesothelioma for patients reporting cancer in two or more first-degree relatives. We found no significant elevation in women or among the small number of men without asbestos exposure. The next-of-kin of three patients (but no controls) reported a possible mesothelioma in a first-degree relative; asbestos exposure could not be ruled out in those relatives. Associations of asbestos with pleural mesothelioma were stronger among men with a reported family history of cancer than men without, although no statistical evidence of an interaction was detected. CONCLUSIONS: These results provide suggestive, but limited, evidence that a family history of cancer may be a risk factor for mesothelioma, or may indicate an increased susceptibility to mesothelioma given asbestos exposure.

Occupational risk factors for brain tumors among women in Shanghai, China.

The etiology of brain cancer is not well understood and few studies have evaluated occupational risk factors among women. We evaluated occupation and industry at time of diagnosis for 276 incident primary brain tumor cases among women in Shanghai, China, for the period 1980-1984, identified through the Shanghai Cancer Registry. Standardized incidence ratios (SIRs) and their 95% confidence intervals (CIs) were calculated for all occupations and industries with at least three female cases. SIRs compared observed to expected numbers of cases, based on incidence rates for Shanghai and the number of women in each occupation and industry according to the 1982 census. Statistically significant excesses of brain tumors were seen among grain farmers (SIR = 6.5, 95% CI = 1.3-19.1), rubber workers (SIR = 5.0, 95% CI = 1.6-11.6), and workers in transportation equipment manufacture and repair (SIR = 2.3, 95% CI = 1.1-4.3). Risks among textile spinners and winders were of borderline significance (SIR = 1.7, 95% CI = 1.0-2.8). Elevated but nonsignificant risks of 2.0 or greater were seen among nurses, plastic products workers, sanitation workers, painters, and workers in manufacture of equipment for electrical generation, transmission, and distribution. Results for farmers, rubber workers, and painters are consistent with previously reported excesses among these occupations in men. The increase among nurses is a new finding, although elevated risks have been observed among male medical professionals. Risks were elevated with likely exposure to pesticides, particularly among those thought to have a high probability and a high level of exposure (SIR = 3.6, 95% CI = 1.2-8.5).

Silicosis and lung cancer among workers in North Carolina dusty trades.

In 1940-1983, 760 cases of silicosis were identified among male North Carolina (NC) workers in dusty trades. Vital status was ascertained through 1983 for 714 silicotics, and death certificates were obtained for 546 of the 550 decedents. The standardized mortality ratio (SMR) for lung cancer based on United States rates was 2.6 [95% confidence interval (95% CI) 1.8-3.6] for whites, 2.3 (95% CI 1.5-3.4) for whites unexposed to other known occupational carcinogens, and 2.4 (95% CI 1.5-3.6) for whites with no other exposure and diagnosed with silicosis while still employed in dusty trades. In addition, the age- and smoking-adjusted rate for silicotics was 3.9 times higher (95% CI 2.4-6.4) than that of nonsilicotic metal miners. This analysis effectively controlled for confounding by age, cigarette smoking, exposure to other occupational carcinogens, and detection bias. The results congrue with the hypothesis of an association between silicosis and lung cancer.

Malignant mesothelioma: attributable risk of asbestos exposure.

OBJECTIVES: To evaluate a case-control study of malignant mesothelioma through patterns of exposure to asbestos based upon information from telephone interviews with next of kin. METHODS: Potential cases, identified from medical files and death certificates, included all people diagnosed with malignant mesothelioma and registered during 1975-1980 by the Los Angeles County Cancer Surveillance Program, the New York State Cancer Registry (excluding New York City), and 39 large Veterans Administration hospitals. Cases whose diagnosis was confirmed in a special pathology review as definite or probable mesothelioma (n = 208) were included in the analysis. Controls (n = 533) had died of other causes, excluding cancer, respiratory disease, suicide, or violence. Direct exposure to asbestos was determined from responses to three types of questions: specific queries as to any exposure to asbestos; occupational or non-vocational participation in any of nine specific activities thought to entail exposure to asbestos; and analysis of life-time work histories. Indirect exposures were assessed through residential histories and reported contact with family members exposed to asbestos. RESULTS: Among men with pleural mesothelioma the attributable risk (AR) for exposure to asbestos was 88% (95% confidence interval (95% CI) 76-95%). For men, the AR of peritoneal cancer was 58% (95% CI 20-89%). For women (both sites combined), the AR was 23% (95% CI 3-72%). The large differences in AR by sex are compatible with the explanations: a lower background incidence rate in women, lower exposure to asbestos, and greater misclassification among women. CONCLUSIONS: Most of the pleural and peritoneal mesotheliomas in the men studied were attributable to exposure to asbestos. The situation in women was less definitive.

Increased risk of colorectal cancer among smokers: results of a 26-year follow-up of US veterans and a review.

To clarify the relationship between tobacco use and risk of colorectal cancer, we evaluated a cohort of 248,046 American veterans followed prospectively for 26 years. In comparison with veterans who had never used tobacco, the risk of death was significantly increased for colon cancer and rectal cancer among current and former cigarette smokers and among pipe or cigar smokers, controlling for social class and occupational physical activity. Rectal-cancer risk was also significantly elevated among users of chewing tobacco or snuff. For both sites, risk increased significantly with pack-years, earlier age at first use, and number of cigarettes. These results reinforce 2 recent reports of the association of cigarette smoking and colorectal cancer in men and women. Inconsistencies in the findings of earlier epidemiologic studies appear to be due in large part to differences in length of follow-up or in choice of controls. Studies with at least 20 years of follow-up or population-based controls have tended to find elevated risk with tobacco smoking, while those with shorter follow-up or hospital controls have not. This, plus the strength and consistency of the association of smoking and colon polyps, suggest that smoking may primarily affect an early stage in the development of colon cancer. If this association is causal, tobacco use may be responsible for 16% of colon-cancer and 22% of rectal-cancer deaths among these veterans.

Attributable risk: advantages of a broad definition of exposure.

Classification of exposure into two levels--one consisting exclusively of unexposed individuals and the other consisting of exposed and perhaps unexposed ones--yields an unbiased estimate of attributable risk when misclassification is nondifferential. The authors advocate, therefore, the use of a broad definition of exposure when estimating attributable risk. Based on this idea, they justify a simple and robust method for estimating the overall attributable risk from several exposures that is based on a division of subjects into two groups, a baseline consisting of those unexposed to all exposures and everyone else.

Occupational exposure to chlorinated aliphatic hydrocarbons and risk of astrocytic brain cancer.

Chlorinated aliphatic hydrocarbons (CAHs) were evaluated as potential risk factors for astrocytic brain tumors. Job-exposure matrices for six individual CAHs and for the general class of organic solvents were applied to data from a case-control study of brain cancer among white men. The matrices indicated whether the CAHs were likely to have been used in each industry and occupation by decade (1920-1980), and provided estimates of probability and intensity of exposure for "exposed" industries and occupations. Cumulative exposure indices were calculated for each subject. Associations of astrocytic brain cancer were observed with likely exposure to carbon tetrachloride, methylene chloride, tetrachloroethylene, and trichloroethylene, but were strongest for methylene chloride. Exposure to chloroform or methyl chloroform showed little indication of an association with brain cancer. Risk of astrocytic brain tumors increased with probability and average intensity of exposure, and with duration of employment in jobs considered exposed to methylene chloride, but not with a cumulative exposure score. These trends could not be explained by exposures to the other solvents.