Serum concentrations of DDE, PCBs, and other persistent organic pollutants and mammographic breast density in Triana, Alabama, a highly exposed population.

INTRODUCTION: Although some persistent organic pollutants (POPs) are considered human carcinogens, results from studies evaluating exposures and breast cancer risk have been inconsistent, potentially related to varying ages at exposure. Additionally, few studies evaluated the association between POPs exposure and mammographic breast density (MBD), an intermediate biomarker of breast cancer risk. We carried out a cross-sectional study to investigate associations between serum POPs concentrations and MBD measured in 1998 in female residents of Triana, Alabama, in a predominately African American population with high POPs exposures, particularly to p,p'-DDT (1,1,1-trichloro-2,2-bis(p-chlorophenyl)ethane). METHODS: We measured lipid-adjusted serum concentrations (ng/g lipid) of p,p'-DDT and its main metabolite p,p'-DDE (1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene), polychlorinated biphenyls (PCBs), ß-hexachlorocyclohexane (ß-HCCH), heptachlor epoxide, oxychlordane, trans-nonachlor, mirex, and aldrin for each woman in our study (n = 210). We also measured two MBD metrics, percent MBD (%MBD) and area of MBD (aMBD). Using adjusted Spearman correlation coefficients (rs) we evaluated correlations between %MBD and aMBD with individual POPs in the overall population and by age group (19-40, 41-54, and 55-91 years) and also estimated adjusted mean measures of MBD with 95% confidence intervals across tertiles of analytes using generalized linear models (GLM). We calculated p-values for multiplicative interaction by age group using GLM. Additional analyses excluded women with current hormone replacement therapy (HRT) use and evaluated early-life exposure (prior to age 18) during the heaviest contamination period in Triana (1947-90). RESULTS: Among all women, we found no correlation between p,p'-DDE and %MBD, but after age stratification and exclusion of HRT users, there was a suggestion of a difference by age group, with younger women having a weak positive correlation (rs = 0.12, p = 0.37) and older women having a weak negative correlation (rs = -0.12, p = 0.43); pinteraction = 0.06. In contrast, PCBs were weakly positively correlated with %MBD among all women, with the correlation magnitudes increasing after excluding current HRT users (rs-total PCBs = 0.17, p = 0.03). After age stratification and exclusion of HRT users, correlations for PCBs were higher among younger and middle-age women, with only a handful of these correlations being statistically significant. For ß-HCCH, the strongest finding was a negative correlation among older women (rs = -0.26, p = 0.07). Correlations were positive predominantly in the younger age group for heptachlor epoxide (rs = 0.27, p = 0.04), oxychlordane (rs = 0.35, p = 0.006), and trans-nonachlor (rs = 0.37, p = 0.003), and largely null for the middle and older age groups; pinteraction range: 0.03-0.05. Similar patterns were found in GLM analyses using tertiles of exposure and aMBD as the metric for MBD. Women exposed during the heaviest chemical contamination period in Triana prior to age 18 had positive correlations between %MBD and PCBs, heptachlor epoxide, mirex, oxychlordane, and trans-nonachlor. CONCLUSIONS: In this population, despite high exposures to p,p'-DDT and thus high serum concentrations of its main metabolite, p,p'-DDE, we did not find strong evidence of a positive association with MBD. In fact, there was some evidence of a negative association among older women for p,p'-DDE; a similar pattern was found for ß-HCCH. However, younger women with higher serum levels of PCBs, heptachlor epoxide, oxychlordane, and trans-nonachlor, who were likely exposed in early life, had higher MBD. These findings should be replicated in larger studies.

Investing in prospective cohorts for etiologic study of occupational exposures.

Prospective cohorts have played a major role in understanding the contribution of diet, physical activity, medical conditions, and genes to the development of many diseases, but have not been widely used for occupational exposures. Studies in agriculture are an exception. We draw upon our experience using this design to study agricultural workers to identify conditions that might foster use of prospective cohorts to study other occupational settings. Prospective cohort studies are perceived by many as the strongest epidemiologic design. It allows updating of information on exposure and other factors, collection of biologic samples before disease diagnosis for biomarker studies, assessment of effect modification by genes, lifestyle, and other occupational exposures, and evaluation of a wide range of health outcomes. Increased use of prospective cohorts would be beneficial in identifying hazardous exposures in the workplace. Occupational epidemiologists should seek opportunities to initiate prospective cohorts to investigate high priority, occupational exposures.

Cigarette smoking and risk of Hodgkin lymphoma and its subtypes: a pooled analysis from the International Lymphoma Epidemiology Consortium (InterLymph).

BACKGROUND: The etiology of Hodgkin lymphoma (HL) remains incompletely characterized. Studies of the association between smoking and HL have yielded ambiguous results, possibly due to differences between HL subtypes. PATIENTS AND METHODS: Through the InterLymph Consortium, 12 case-control studies regarding cigarette smoking and HL were identified. Pooled analyses on the association between smoking and HL stratified by tumor histology and Epstein-Barr virus (EBV) status were conducted using random effects models adjusted for confounders. Analyses included 3335 HL cases and 14 278 controls. RESULTS: Overall, 54.5% of cases and 57.4% of controls were ever cigarette smokers. Compared with never smokers, ever smokers had an odds ratio (OR) of HL of 1.10 [95% confidence interval (CI) 1.01-1.21]. This increased risk reflected associations with mixed cellularity cHL (OR = 1.60, 95% CI 1.29-1.99) and EBV-positive cHL (OR = 1.81, 95% CI 1.27-2.56) among current smokers, whereas risk of nodular sclerosis (OR = 1.09, 95% CI 0.90-1.32) and EBV-negative HL (OR = 1.02, 95% CI 0.72-1.44) was not increased. CONCLUSION: These results support the notion of etiologic heterogeneity between HL subtypes, highlighting the need for HL stratification in future studies. Even if not relevant to all subtypes, our study emphasizes that cigarette smoking should be added to the few modifiable HL risk factors identified.

Cigarette smoking and risk of non-Hodgkin lymphoma subtypes among women.

Previous studies of the relationship between cigarette smoking and non-Hodgkin lymphoma (NHL) have yielded conflicting results, perhaps because most studies have evaluated the risk for all NHL subtypes combined. Data from a population-based case-control study conducted among women in Connecticut were used to evaluate the impact of cigarette smoking on the risk of NHL by histologic type, tumour grade, and immunologic type. A total of 601 histologically confirmed, incident cases of NHL and 718 population-based controls provided in-person interviews. A standardised, structured questionnaire was used to collect information on each subject's current smoking status, age at initiation, duration and intensity of smoking, and cumulative lifetime exposure to smoking. Our data suggest that cigarette smoking does not alter the risk of all NHL subtypes combined. However, increased risk of follicular lymphoma appears to be associated with increased intensity and duration of smoking, and cumulative lifetime exposure to smoking. Compared with nonsmokers, women with a cumulative lifetime exposure of 16-33 pack-years and 34 pack-years or greater experience 50% increased risk (OR=1.5, 95% CI 0.9-2.5) and 80% increased risk (OR=1.8, 95% CI 1.1-3.2), respectively, of follicular lymphoma (P for linear trend=0.05). Our study findings are consistent with several previous epidemiologic studies suggesting that cigarette smoking increases the risk of follicular lymphoma. This research highlights the importance of distinguishing between NHL subtypes in future research on the aetiology of NHL.

Integrative assessment of multiple pesticides as risk factors for non-Hodgkin's lymphoma among men.

BACKGROUND: An increased rate of non-Hodgkin's lymphoma (NHL) has been repeatedly observed among farmers, but identification of specific exposures that explain this observation has proven difficult. METHODS: During the 1980s, the National Cancer Institute conducted three case-control studies of NHL in the midwestern United States. These pooled data were used to examine pesticide exposures in farming as risk factors for NHL in men. The large sample size (n = 3417) allowed analysis of 47 pesticides simultaneously, controlling for potential confounding by other pesticides in the model, and adjusting the estimates based on a prespecified variance to make them more stable. RESULTS: Reported use of several individual pesticides was associated with increased NHL incidence, including organophosphate insecticides coumaphos, diazinon, and fonofos, insecticides chlordane, dieldrin, and copper acetoarsenite, and herbicides atrazine, glyphosate, and sodium chlorate. A subanalysis of these "potentially carcinogenic" pesticides suggested a positive trend of risk with exposure to increasing numbers. CONCLUSION: Consideration of multiple exposures is important in accurately estimating specific effects and in evaluating realistic exposure scenarios.

Glutathione S-transferase M1 and T1 genetic polymorphisms, alcohol consumption and breast cancer risk.

Alcohol consumption has been inconsistently associated with breast cancer risk. Recent studies suggest that genetic polymorphisms of glutathione S-transferases (GSTs) may modify this relation. To determine if breast cancer risk is associated with GSTM1 and GSTT1 genetic polymorphisms, and to evaluate the effect modification between GST genotypes and alcohol consumption in the risk of breast cancer, we conducted a case-control study in the state of Connecticut in the period 1998 and 2001. Cases were histologically confirmed, incident breast cancer patients in New Haven County, CT. Controls were randomly selected from women histologically confirmed to be without breast cancer. The study results show that, while GSTM1 genotypes were not associated with breast cancer risk, GSTT1-null genotype was associated with a significant 90% increased risk for postmenopausal women (OR=1.9, 95% CI 1.2-3.0). Analysis by GST genotypes and alcohol consumption shows that GSTM1A ever-drinking women had a 2.5-fold (OR=2.5, 95% CI 1.1-5.5) increased risk of breast cancer compared to the GSTM1A never-drinkers, and the risk increases with duration and daily amount of alcohol consumption. Postmenopausal women with GSTT1-null genotype, who consumed a lifetime of >250 kg of spirit-equivalents, had an almost seven-fold increased risk (OR=6.8, 95% CI 1.4-33.9), and drinking commencing at younger ages appears to carry a higher risk. An OR of 8.2 (95% CI 1.2-57.4) was observed for those with GSTM1A, and GSTT1-null genotypes who had consumed a lifetime of >250 kg of spirit-equivalents. In conclusion, alcohol consumption may increase breast cancer risk among those who carry susceptible GST genotypes.

Use of hair colouring products and breast cancer risk: a case-control study in Connecticut.

This case-control study was designed to test the hypothesis that the risk of breast cancer varies by type and colour of the hair colouring products used. A total of 608 cases and 609 controls were included in the study. We found no increased risk associated with the overall use of hair dye products or exclusive use of permanent or temporary types of hair dye products. Among those who reported to have exclusively used semi-permanent types of hair colouring products, some of the ORs were elevated. However, none of the ORs related to age at first use, duration of use, total number of applications, and years since first use, was statistically significant. There was also no increased risk of breast cancer associated with exclusive use of dark or light hair colouring products, or use of mixed types or colours of hair dye products. We also found no increased risk of breast cancer associated with hair dye use based on an individual's reason for using a hair colouring product, such as to cover grey or to change natural hair colour. These data suggest that the use of hair colouring products does not have a major impact on the risk of breast cancer.

Development of a life events/icon calendar questionnaire to ascertain occupational histories and other characteristics of migrant farmworkers.

BACKGROUND: Specialized methods are necessary to collect data from migrant farmworkers for epidemiologic research. METHODS: We developed a questionnaire that collected lifetime occupational histories and other lifestyle risk factors via a life events/icon calendar, and administered the questionnaire to a convenience sample of 162 migrant farmworkers in nine areas of the U.S. RESULTS: The average duration of the interviews was about 1 h 30 min, with an average of 45 min for the work history section. The occupational histories covered a median of 27.6 years per person for men and 20.8 years per person for women. The median number of years spent in farm jobs was 11.3 for men and 5.8 for women. The median number of farm jobs (crop/task combination) per person was 59 among men and 27 among women. Many farmworkers performed the same crop/task combinations at multiple times throughout their lives, yielding a median of 13 unique farm jobs and 8 unique crops among men and 7 jobs and 5 crops among women. CONCLUSIONS: The project demonstrated that it is feasible to collect detailed work histories and other risk factor data from farmworkers, documented the complexity of work histories encountered among farmworkers, and yielded recommendations for refining a questionnaire that will facilitate future epidemiologic research on farmworkers.

Test-retest reliability of an icon/calendar-based questionnaire used to assess occupational history.

BACKGROUND: Self-reported work histories can be critical for both epidemiologic and clinical purposes. However, the complexity of some work histories, such as those of migrant farm workers, can hamper recall, resulting in inaccurate reporting. Memory aids may reduce such error. This study assesses the reliability of work histories collected using such aids in the form of an icon/calendar-based questionnaire. METHODS: Thirty-one males engaged in farmwork and other manual labor for a median 28 years (range: 10-64) were interviewed twice, 8-14 months apart, about their lifetime employment. In each interview, subjects were asked about important life events, which were recorded with icons on a calendar. They were then asked to recount their work history, including for each job the tasks, crops or products handled, starting and ending dates, and location. This information was recorded, job-by-job, on the calendar with icons. RESULTS: Interquestionnaire agreement of cumulative reported employment duration (as measured by the correlation coefficient) was moderate to high across all time periods for certain crops (e.g., r = 0.69-0.92 for apple-related work), by location (e.g., r = 0.76-0.95 for Washington State), and for agricultural work in general (r = 0.67-0.94), but was lower for specific tasks. Agreement of job counts was high for total work history for certain crops (e.g., r = 0.93 for apple-related work), by location (e.g., r = 0.90 for Washington State), and for agricultural work in general (r = 0.89), but paradoxically decreased with proximity to the interview date. Agreement of both measures tended to be highest for those tasks and crops in which subjects reported spending the most time. Categorization of subjects into tertiles on the basis of either cumulative duration or counts produced results similar to those observed for job counts. CONCLUSIONS: The icon-calendar questionnaire is an effective tool for estimating cumulative duration of certain work categorizations among subjects with complex work histories.

Comparison of a traditional questionnaire with an icon/calendar-based questionnaire to assess occupational history.

BACKGROUND: Self-reported work histories are an essential tool for estimating exposure in many occupational epidemiologic studies. However, the transience of some occupations such as farm work can hamper recall, resulting in inaccurate reporting. To address this problem, we have developed an icon/calendar-based questionnaire. This study compares work histories collected via this questionnaire to those collected via a traditional questionnaire. METHODS: Eighty-nine farmworkers and non-farmworkers were interviewed twice, 8-10 months apart, about their lifetime employment. In the first interview, subjects were asked to recount their entire work history, starting from the interview date and moving backwards in time ("traditional questionnaire"). In the second interview, subjects were first asked about important life events, which were recorded with icons on a calendar. They were then asked to recount their work history, which was recorded, job-by-job, on the calendar with icons ("icon-calendar questionnaire"). RESULTS: Number of jobs and amount of work time accounted for since first employment were significantly greater using the icon-calendar questionnaire than the traditional questionnaire, the disparity increasing with time from the date of interview. The ratio of number of jobs in the traditional questionnaire to number of jobs in the icon-calendar questionnaire decreased from 100.0% in the most recent time period to 0.0% in the earliest time period. While the percentage of time explained by employment remained relatively constant across time periods in the icon-calendar questionnaire, ranging from 86.3 to 98.9%, it rapidly decreased with time in the traditional questionnaire, from 77.9% in the most recent time period to 0.0% in the earliest time period. CONCLUSIONS: The icon-calendar questionnaire was more effective than the traditional questionnaire for obtaining complex work histories during interviews, producing a more complete picture of a person's work history.

Comparability of data obtained from migrant farmworkers and their spouses on occupational history.

BACKGROUND: Epidemiologic studies, particularly case-control studies, often rely on proxy respondents to provide information about subjects' occupational histories. The quality of proxy-reported information in occupational histories has never been evaluated for migrant farmworkers. METHODS: We compared occupational histories self-reported by 31 farmworkers with those reported by their wives. The work histories were obtained using an icon/calendar-based questionnaire that was designed to facilitate recall for migrant farmworkers, who typically have complex work histories. RESULTS: The work histories provided by proxy respondents contained 32% fewer jobs and accounted for 24% fewer years than the self-reported histories. Correlations for lifetime duration of employment in different types of jobs were moderate to good for general agricultural jobs in Washington (0.70) and apple-related jobs (0.65), which were held by virtually all of the farmworkers; correlations were moderate to poor for less common jobs and for specific types of tasks. Agreement was better after marriage than before, and for jobs held in the current year compared to other time frames. Overall, the ability of the spouses to provide occupational histories for farmworkers was within the range observed in studies involving other occupations and industries. CONCLUSIONS: In studies involving farmworkers, when study subjects cannot be interviewed, spouses can provide useful information on occupational histories. However, the information should be used only for more generalized exposure assessments; it is most appropriate for estimating cumulative duration of agricultural work, or recent work, by place or for common crops.

Ascertainment of pesticide exposures of migrant and seasonal farmworker children: findings from focus groups.

BACKGROUND: To design questionnaires for epidemiologic research among children of migrant farmworkers, researchers need to consider ways to best solicit information about pesticide exposures. METHODS: Bilingual facilitators conducted five focus groups with either migrant farmworker mothers or their children (age range 8-16 years) in southern Texas and northeastern Colorado. Guided questions were used to assess activities of migrant farmworker children and the ways to best elicit information about exposure to pesticides. RESULTS: Participants reported a large number of activities that may potentially expose children to pesticides through both direct and indirect routes. Prompting, indirect questions about chemical use, and use of local and trusted facilitators increased information elicited from focus group participants. CONCLUSIONS: These focus groups helped to provide information for developing questionnaire items related to pesticide exposure among migrant farmworker children, and highlighted the importance of using bilingual community interviewers and including children as respondents.

Determining the probability of pesticide exposures among migrant farmworkers: results from a feasibility study.

BACKGROUND: Migrant and seasonal farmworkers are exposed to pesticides through their work with crops and livestock. Because workers are usually unaware of the pesticides applied, specific pesticide exposures cannot be determined by interviews. We conducted a study to determine the feasibility of identifying probable pesticide exposures based on work histories. METHODS: The study included 162 farm workers in seven states. Interviewers obtained a lifetime work history including the crops, tasks, months, and locations worked. We investigated the availability of survey data on pesticide use for crops and livestock in the seven pilot states. Probabilities of use for pesticide types (herbicides, insecticides, fungicides, etc.) and specific chemicals were calculated from the available data for two farm workers. The work histories were chosen to illustrate how the quality of the pesticide use information varied across crops, states, and years. RESULTS: For most vegetable and fruit crops there were regional pesticide use data in the late 1970s, no data in the 1980s, and state-specific data every other year in the 1990s. Annual use surveys for cotton and potatoes began in the late 1980s. For a few crops, including asparagus, broccoli, lettuce, strawberries, plums, and Christmas trees, there were no federal data or data from the seven states before the 1990s. CONCLUSIONS: We conclude that identifying probable pesticide exposures is feasible in some locations. However, the lack of pesticide use data before the 1990s for many crops will limit the quality of historic exposure assessment for most workers.

Evaluation of Mexican American migrant farmworker work practices and organochlorine pesticide metabolites.

BACKGROUND: Epidemiologic studies often must rely upon questionnaire data to assess past exposures. The ability of questionnaires to rank migrant farmworkers according to past pesticide exposure is not known. METHODS: We conducted a pilot feasibility study to measure a panel of 21 organochlorine pesticides (OCPs) and correlate levels with reported occupational exposures in 26 Mexican-American migrant farmworkers in Baytown, Texas. The Migrant Farmworker Questionnaire developed by the National Cancer Institute (NCI) was administered and each participant donated a blood sample. Three OCPs [mean (ppb) levels: mirex 1.8, DDT 1.0, and trans-nonachlor 0.7] were detected despite the fact that these chemicals have been banned in the US for many years, and the detected levels were far higher than the standard provided by the referent laboratory. Work clothes, protective attire, and self-reported pesticide exposures were significant predictors of OCP exposure. Similarly, personal hygiene, length of employment, and number of duties also predicted OCP exposure. CONCLUSIONS: The results of this study indicate that data obtained from standardized questionnaires may be reasonable indicators of occupational exposure when biomarker data are not available.

Organophosphate pesticide residues in urine of farmworkers and their children in Fresno County, California.

BACKGROUND: Childhood cancer, notably leukemia, brain cancer, non-Hodgkin's lymphoma, soft tissue sarcoma, and Hodgkin's disease, has been associated with pesticide exposure, often with greater relative risks than among exposed adults, suggesting greater susceptibility in children. These differences in risk may be due to developmental factors or differences in pesticide exposure. METHODS: A feasibility study was conducted to determine levels of pesticide metabolites in urine of adults (n = 18) and children (n = 9) in Fresno County, California, an intensely agricultural county in the Central San Joaquin Valley. Spot urine samples were obtained and analyzed for six metabolites of organophosphate (OP) pesticides using gas chromatography with flame photometric detection methods. The metabolites of OP pesticides included DMP, DEP, DMTP, DMDTP, DETP, and DEDTP. RESULTS: Levels were generally low for both adults and children for most metabolites tested. Frequencies of detection ranged from 0 to 37%, with mean levels ranging from non-detectable to 13.22 ppb. However, levels of several metabolites were higher in children than in adults. The most frequently detected metabolite, DMP, was found among 44% of the children and 33% of the adults. DMTP was detected among 33% of the children and 28% of the adults. CONCLUSIONS: These results are difficult to interpret given the sampling variation associated with the small sample size. Nevertheless, because OP pesticides have been associated with increased cancer risk in animal and human studies, these results indicate a need to closely monitor children's exposure to environmental chemicals.

Prenatal exposure to pesticides: a feasibility study among migrant and seasonal farmworkers.

BACKGROUND: Migrant and seasonal farmworkers have a high potential for pesticide exposures, yet are rarely included in epidemiologic studies. This study examined the feasibility of assessing prenatal exposures to pesticides and other compounds in pregnant Hispanic farmworkers. METHODS: Nine women completed a survey about work experiences during pregnancy. Maternal urine, cord blood, and placenta samples were obtained at delivery for analysis of 51 analytes, including 6 phenoxy acid or triazine herbicides, 21 organochlorine insecticides, 10 PCBs, and 14 volatile organic compounds. RESULTS: Seven of 51 analytes were found in the biological samples. DDE, DDT, dichlorbenzene, toluene, trimethylbenzene, and endosulfan sulfate were detected in cord blood samples, and 2,4-D in urine from one or more women. CONCLUSIONS: We documented the feasibility of following farmworkers to assess in utero exposure to pesticides and other contaminants, and demonstrated exposure to these compounds. Difficulties in measuring pesticides with short half lives were noted.

Tracing migrant farmworkers in Starr County, Texas.

BACKGROUND: In response to The National Cancer Institute (NCI) concerns about the ability to conduct studies among migrant farmworkers, this study evaluated the feasibility of identifying migrant farmworkers in their home state and tracing them over an extended period of time. METHODS: In 1995, a group of 196 persons who had classified themselves as "migrant farmworkers" in two earlier chronic disease studies was identified. The primary objective of the current study was to determine the proportion of these farmworkers who could be located in 1995-1996. RESULTS: Of these farmworkers, 163 were located and were living (83.2%), 15 had died (7.6%), and 18 (9.2%) were lost to follow-up. CONCLUSIONS: The excellent follow-up rate was due in part to the high participation rates among persons contacted for information, stability of the farmworkers' permanent homes, predictable timing of migration, and a longstanding health research program with established community contacts.

Ability to trace migrant farmworkers ten years after initial identification in a Northern State (Wisconsin).

BACKGROUND: Migrant farmworkers have rarely been included in epidemiologic studies. To assess the feasibility of following farmworkers over extended periods, a critical feature of many study designs, we attempted to trace a sample of Mexican-American farmworkers identified in a clinic in Wisconsin. METHODS: We randomly chose 100 farmworkers from a migrant health center registration list for 1984-85. In 1995, we searched recent clinic records, made telephone calls, and visited migrant camps to find these farmworkers in Wisconsin during the growing season. We also attempted to find 46 farmworkers at their homes in southwest Texas over a two-week period in 1996 using the address listed in the clinic records, local phone books, and conversations with next-door neighbors. RESULTS: Although 25 farmworkers had reregistered at the clinic in recent years, we found only 6 of them in Wisconsin in 1995. In southwest Texas, we either located or ascertained information about the vital status of 25 of the 46 farmworkers (54%). CONCLUSIONS: Tracing efforts must include extensive contacts in farmworkers' home states and must incorporate a variety of information sources. Tracing farmworkers in epidemiologic studies appears to be feasible but requires more intensive methods over longer periods of time than those used in this study.

Proportionate mortality among US migrant and seasonal farmworkers in twenty-four states.

BACKGROUND: US migrant and seasonal farmworkers may be exposed to potentially carcinogenic pesticides and other agents. Little epidemiologic research has been conducted on this population. METHODS: We examined the proportionate mortality of 26,148 subjects (14,631 white men (WM), 7,299 nonwhite men (NM), 1,081 white women (WW), and 3,137 nonwhite women (NW)) who were identified as farmworkers on death certificates from 24 US states during 1984-1993. RESULTS: Farmworkers had significantly elevated proportionate mortality from injuries, tuberculosis, mental disorders, cerebrovascular disease, respiratory diseases, ulcers, hypertension (NW), and cirrhosis (NW). There was significantly reduced mortality from infectious diseases (other than tuberculosis), endocrine disorders, nervous system diseases, pneumoconioses, arteriosclerotic heart disease (WM), and all cancers combined. Proportionate cancer mortality analyses found excess cancers of the buccal cavity, larynx, esophagus, stomach, skin (NW), and cervix, and deficits for cancers of the colon, breast, kidney, pancreas (NW), and lymphohematopoietic system. CONCLUSIONS: The excess deaths from injuries, respiratory disease, and stomach cancer, and the deficits of colon cancer and arteriosclerotic heart disease among farmworkers, are consistent with typical mortality patterns previously observed among farm owner/operators. The excess buccal, laryngeal, esophageal, and cervical cancers, and the deficits of breast cancer and lymphohematopoietic cancers have not generally been observed in studies of farm owner/operators.