Cancer incidence among a cohort of female farm residents in New York State

Cancer incidence among a cohort of female farm residents in New York State

Ying Wang

FARMERS AND THEIR FAMILIES can potentially be exposed to several hazardous agents, some of which are known carcinogens. These agents include pesticides, solvents, fertilizers, dusts, some infectious agents of animal origin, molds, antibiotics, and fungi. (1) During the past decade, many epidemiologic studies of cancer incidence have been conducted among farmers of many countries. (2-4) Results of these studies have shown that farmers are at low overall risk of cancer. This assessment of low risk is based mainly on a low incidence of cancer at the most common cancer sites, such as the lung, bladder, and colon. Despite the overall low cancer rates among farmers, some studies in many countries have consistently suggested that farmers may be at excess risk of certain malignancies, including leukemia; Hodgkin’s disease; non-Hodgkin’s lymphoma; multiple myeloma; and cancers of the lip, stomach, skin, prostate, brain, and connective tissue. (2-4)

Most studies of farmers have focused on men, with women generally being excluded because of their small numbers or presumed low exposures. In the few studies of female farmers, ovarian and breast cancer, soft-tissue sarcoma, multiple myeloma, and non-Hodgkin’s lymphoma have been linked to triazine herbicides, (5) insecticides, (6) herbicides, (7) and various pesticides , (8) respectively. However, these studies also revealed a significantly low incidence of lung cancer and breast cancer. (9,10) In a recent population-based cohort study on cancer incidence among women living on farms in Iowa, (11) significantly reduced risks were found for all cancers, lung cancer, and bladder cancer, compared with nonfarm residents. The risk for non-Hodgkin’s lymphoma was elevated in female farm residents in the Iowa study. The present study of female farmers and farm residents differs from other studies in that it focuses on New York State agriculture, which is predominantly dairy, (12) whereas agriculture in the Midwest is mostly cash crops.

Previous studies on cancer incidence conducted by the New York State Department of Health among 18,811 male farmers in New York State during 1973-1984 (13) resulted in findings consistent with other studies of farming populations. The male study cohort experienced lower overall cancer incidence than the comparison population. Statistically significant deficits in cancer incidence were found for lung, colon, and bladder cancer. Elevated incidence ratios, although not statistically significant, were found for melanoma of the skin, multiple myeloma, and lip and prostate cancer. In this study, we conducted a comparable retrospective cohort study of cancer incidence among female farm residents who were New York Farm Bureau (Glenmont, New York) members, or members’ spouses or relatives, to determine if they experienced higher rates than the general population for overall cancer incidence, or for any specific cancers. Recent research findings and public concern regarding possible associations between pesticide exposure and breast cancer, as well as other female cancers, provided our motivating hypothesis that agricultural exposures to pesticides may result in higher rates of breast cancer or other hormonally mediated cancers among women residing on farms.

Materials and Method

Identification of the study cohort. We used the membership list of the New York Farm Bureau to identify the study cohort. The Farm Bureau is a voluntary organization that represents the economic and political interests of farmers in New York State, and it provides group insurance plans, discounts on equipment purchases, and other benefits. Approximately 70% of New York farmers belong to the Farm Bureau. (14) The New York Farm Bureau’s philosophy is that members be active farmers, and that the proportion of associate members be maintained at less than 10%. Associate members were excluded from this study.

Females age 30 to 64 yr who were farmers, or adult relatives (sharing the same last name) of farmers who had been Farm Bureau members for at least 1 yr, between 1980 and 1985, were eligible for inclusion in the cohort. Data obtained from the New York Farm Bureau membership list included each member’s name, address, year joined, and type of farm at which he or she resided. Last names and mailing addresses of members were then matched with driver’s license and vehicle registration records from the New York State Department of Motor Vehicles. In addition, we matched this information with vital records from the New York State Department of Health (including birth certificates and marriage licenses) to obtain the names, addresses, and birth dates of female farm residents.

Selection of comparison population. The methodology used for selecting the comparison population was basically the same as that used in the study of male farmers. (13) The comparison population comprised women age 30 to 64 yr who resided in New York State (a) in ZIP codes classified as rural or suburban areas, and (b) in ZIP codes containing at least 10 female farm residents per 10,000 female residents age 30 to 64 yr, as enumerated by the 1980 U.S. Census. (15) The 1st restriction excluded women who lived in urbanized areas, thus avoiding confounding any farmer-cancer incidence associations with urban-rural differences. For consistency, the cohort members who had an urban mailing address (ZIP code) were also excluded. We conducted the 2nd exclusion to isolate the effects of farm residential exposures from more general environmental influences, because areas with very few farms are not representative of the communities in which most farmers live. For a similar reason, cohort members who lived in rural areas (ZIP codes) with very few farms were also excluded from the study.

Identification of cancer cases. The New York State Cancer Registry served as the source for the data on cancer cases used in this study. All hospitals and physicians who treat patients diagnosed with cancer are required to report the cases to the Health Department’s Cancer Registry. The completeness of reporting is estimated to be more than 95% (16) for New York, exclusive of New York City. We matched the records of cohort members with Cancer Registry files to identify members who had been diagnosed with cancer. If a cohort member developed a 2nd primary malignancy during the study period, only the 1st tumor was included. Moreover, cohort members who had cancers diagnosed before 1980 were excluded from the study. We determined cancer cases among the comparison population by searching Cancer Registry files from 1980 through 1993 for females age 30 to 64 yr in the selected comparison area.

Follow-up of the cohort. We used several methods to follow the study cohort from 1980 through 1993. The cohort members’ personal identifiers–such as name, date of birth, and ZIP code–were linked with New York State Motor Vehicle Department files to search for the last year of application or renewal for drivers’ licenses or vehicle registrations, and the date of voluntary plate surrender. The result of this procedure was the identification of the most recent year that a cohort member owned or drove a car and was a resident of New York State. Members’ records were also matched with vital records from the Health Department for the last year of giving birth, and for the year of death if deceased. In addition, active Farm Bureau membership status for female farmers in the cohort was used for the determination of follow-up status.

Statistical analysis. The standardized incidence ratio (SIR) was computed as the ratio of observed vs. expected cancer cases in the study cohort. The expected number of cancer cases was calculated by multiplying the number of person-years of observation in each age and calendar year stratum for the study cohort by the stratum-specific rates in the comparison population. Person-years of observation for cohort members at the individual level were computed as the difference between date of entry into the study and date of departure. The date of departure was defined as the date of death, date of cancer diagnosis, last date known alive before loss to follow-up, or the closing date of the study. Person-years of observation among the cohort members were computed across a 5-yr age range and 5 calendar years of follow-up strata in this study. We tested the statistical significance for departures of SIRs from 1.0 by computing 95% confidence intervals (CIs) for SIRs, on the basis of the Poisson distribution model. A 95% CI that does not contain 1.0 indicates that the difference between the observed and expected number of cases is statistically significant.

Results

Table 1 summarizes the composition of the study cohort and the comparison population. Initially, 7,008 female farm residents who were 30-64 yr at the time of entry into the study in upstate New York (excluding New York City) were identified using the New York Farm Bureau membership list. Of these women, 581 with urban mailing addresses and 117 who lived in areas with very few Farm Bureau members were excluded from the cohort. Ultimately, 6,310 women were included in our study cohort. Similarly, using 1980 U.S. Census data, 622,268 females age 30-64 yr who lived in the selected ZIP codes were included as the comparison population. Table 2 shows the age distributions of the cohort at the time of entry into the study (1980-1985) and of the comparison population in 1980. The comparison population and the cohort had a similar age distribution for women 30-64 yr of age. Table 3 presents the follow-up status of the cohort for the study period 1980-1993. A total of 336 cancer cases were identified. Once a woman was identified with a primary cancer, she was no longer followed.

Table 4 displays the number of person-years of observation, the observed and expected number of cancer cases, cancer SIRs, and 95% CIs, by age group for the study cohort. During the 14 yr of follow-up, 336 cancer cases were observed among the cohort members, and 476 cancer cases were expected on the basis of the stratum-specific rates in the comparison population. The SIR for the entire cohort was 0.71, with a 95% CI of 0.63, 0.79. The data in Table 4 also show that the SIR was less than 1 for all age groups examined, except for the 45-49-yr group. We found a significantly lower than expected cancer incidence among cohort members age 30-39 yr and 50 yr or older.

Table 5 shows the results from stratified analyses by specific cancer. The SIRs were less than 1 for all specific cancer sites examined, except for liver, uterus, melanoma, and thyroid. The low SIR for total cancers resulted from a significantly low incidence for some sites, including colon-rectum, lung, ovary, and other sites combined. As can be seen in the table, the SIRs for these cancer sites were less than 1, with 95% CIs not including 1. Although the SIRs for liver and thyroid cancer were greater than 1, few cases were seen, and their 95% CIs were broad and included 1.

Table 6 summarizes the results of stratified analyses by age group for selected cancer sites. We performed the analyses to determine which strata contributed to the overall deficit or excess in cancer incidence. Given the small numbers, age was collapsed into 3 groups: (1) 30-49 yr, (2) 50-69 yr, and (3) 70 yr or older. Significantly low risks for colorectal, lung, and breast cancer, and other cancers combined, were found in the 50-69-yr age group, but not in the other 2 age groups. For all cancers combined, significantly low cancer risk was found among the cohort members 50 yr of age or older.

Discussion

Similar to findings in a previous study of male farmers in New York State, (13) as well as findings from other studies on farm populations, (2,9-11,18,19) this cohort of female farm residents experienced significantly lower cancer risks for all cancers combined, and for lung cancer and colorectal cancers, than the study’s rural female comparison population. The findings suggest that the study cohort might have experienced some of the same factors that reduce cancer incidence for male farmers, including the “healthy worker effect”–that women who live and work on a farm are healthier than women in general. Although this study did not assess each individual cohort member’s involvement in agricultural activities while residing on the farm, we believe that many female farm residents participate considerably in farm chores. Farmers and farm residents could be protected by their physically active work and lifestyle, thereby decreasing risks for colon and rectal cancer.

The low incidence of lung cancer among farm populations has been reported consistently in the literature (2,9-11,13,18-23) and has been, to a large degree, attributed to the low smoking rate among farmers. (24,25) In addition, similar to other population-based studies that evaluated occupation and breast cancer risk in various geographic agricultural regions, (9-11,26-29) our study failed to find an association between breast cancer risk and agriculture exposure. Moreover, we found significantly decreased risk for leukemia, as was also found among Iowa women living on farms. (11)

Non-Hodgkin’s lymphoma has been associated with agricultural pesticide use among (mostly male) farmers in many countries. (2,30-32) Results from the present study did not show an increased risk for non-Hodgkin’s lymphoma among the cohort members. The incidence was low (although not statistically significant), compared with the reference population (Table 5). A similar low incidence was also found among a cohort of New York State male farmers (13) and among female farmers in Sweden. (9) In a study of the role of agricultural pesticide use in the development of non-Hodgkin’s lymphoma in women, Zahm et al. (8) reported no increased risk of non-Hodgkin’s lymphoma among women who had ever lived or worked on a farm; neither the use of insecticides nor herbicides on the farm was associated with non-Hodgkin’s lymphoma. However, when more specific definitions of pesticide exposure (i.e., application methods and chemical classes) were used, the authors found a significant 4.5-fold increased risk in non-Hodgkin’s lymphoma among the few women who personally handled, mixed, or applied organophosphates. The cohort in our study might have had lower overall pesticide exposure than those in other studies because about half of New York farms are dairy farms, (12) which typically use less pesticides than are used on other cash crops.

Elevated, but statistically nonsignificant, risks for cancer of the thyroid and liver were found among the cohort members in the present study. Slightly increased thyroid cancer risk (SIR = 1.16; 95% CI = 0.69, 1.81) was also found among women in Norwegian agriculture. (10) In a recent investigation of the relationship between thyroid cancer in Norwegian women and the occupation of their spouses, Frich et al. (33) found increased risk of thyroid cancer among women whose spouses belonged to the occupational category “agriculture, forestry or fishery.” High consumption of butter and cheeses, (34) residence in areas of endemic goiter, (35) iodine deficiency, (36) and some female hormonal and reproductive factors (37) were found to be associated with elevated risk of thyroid cancer.

Similar to findings for thyroid cancer, dietary factors such as food-based toxins have been identified by some researchers as risk factors for primary liver cancer. (38-40) Findings on the association between agricultural and pesticide exposures and the risk of primary liver cancer have been controversial. In a retrospective case-control study of primary liver cancer in New Jersey, Stemhagen et al. (41) found an estimated relative risk of 1.89 (95% CI = 1.19, 3.00) among males identified as farm laborers. On the other hand, agricultural and pesticide exposures have been linked with significantly decreased risk for liver cancer among Sweden’s male farmers, (42) farm workers in Denmark, (43) and pesticide applicators in Swedish agriculture. (44)

In the present study, the cohort women, identified using the New York Farm Bureau membership list, were only a subset of females residing on farms in New York State. Thus, findings from this study may not describe the entire population of New York State female farm residents. The female farm residents who were not in the study cohort were included in the comparison group. If they are at the same risk as the farm women in the study, findings toward the null hypothesis of “no difference” would result. It should also be noted that the study cohort was about 1% of the comparison population. The impact on SIRs, if any, would tend to reduce the difference between the study cohort and the comparison population.

Approximately 5% of the cohort members were lost to follow-up. Failure to identify some cancer cases among this group of women might have contributed to the reduced SIRs. If it is assumed that women who were lost to follow-up have the same age- and site-specific cancer rates as those who were followed, about 16 cancers would be added to the total observed cancer cases, and thus the incidence rate ratio would increase correspondingly, but not substantially (from 0.71 to 0.74 for all cancer combined). Moreover, computer matching might have failed to detect some reported cancers because of errors in spelling or date of birth. This also would tend to decrease the SIRs because the numerator would be undercounted. However, the number of cases possibly missed through records-matching, due to errors in the matching variables, was greatly decreased in this study. Fortunately, we were able obtain important information for the cohort members, such as name, address, date of birth, and social security number. All of these variables were used in matching, thus ensuring that the number of possibly missed cases was sufficiently small not to decrease the SIRs substantially. In a prospective cohort study currently under investigation in our research group, cancer cases identified through multivariable records-matching were compared with those reported by the cohort members, and confirmed manually with New York State vital records during the follow-up period. The results indicate that less than 1% of cancer cases could be missed if the multivariable records-matching method is used.

Other limitations of the present study include the lack of specific information on types and duration of exposures, and the lack of individual data on cancer risk factors, such as smoking and other lifestyle factors. The method by which this cohort was formed precludes examining cancer incidence by duration of farming or farm residency, or by determining whether a dose-response relationship existed between the risk of cancer and cumulative exposure. Despite the study’s limitations and opportunities for bias identified above, the results of this cohort study–which are generally consistent with those of previous investigations–indicate that this cohort of women residing on farms has experienced relatively lowered cancer risk than the general rural population.

Table 1.–Formation of the New York State Female Farm Resident Cohort

and Comparison Population Used in the Analysis: 1980-1993

Female farm Comparison

resident population

cohort * ([dagger])

Geographic area n % n %

Total upstate

New York ([double dagger]) 7,008 100.0 2,668,080 100.0

Urbanized areas -581 -8.3 -1,812,745 -67.9

Other cities and towns

with < 0.1% farm residents -117 -1.7 -233,067 -8.8

Remainder 6,310 90.0 622,268 23.3

* New York State female farm residents identified, 30-64 yr of age.

([dagger]) New York State female residents

30-64 yr of age, 1980 U.S. Census.

([double dagger]) New York State, excluding New York City.

Table 2.–Age Distributions of the New York State Female

Farm Resident Cohort at Entry into the Study (1980-1985)

and of the Comparison Population (Based on 1980 U.S.

Census (15))

Female farm Comparison

resident cohort * population ([dagger])

Age group (yr) n yr n yr

30-34 1,025 16.2 118,603 19.1

35-39 980 15.5 99,925 16.1

40-44 818 13.0 81,716 13.1

45-49 931 14.8 78,590 12.6

50-54 974 15.4 83,743 13.5

55-59 939 14.9 84,274 13.5

60-64 643 10.2 75,417 12.1

Total 6,310 100.0 622,268 100.0

Table 3.–Follow-Up Status of the New York State Female

Farm Resident Cohort: 1980-1993

Status n %

Living

From Motor Vehicle Department records 5,326 84.4

From Farm Bureau records 109 1.7

From vital statistics 105 1.7

Primary malignant neoplasms 336 5.3

Deceased (causes other than malignant

neoplasms) 129 2.0

Lost to follow-up 305 4.9

Total 6,310 100.0

Table 4.–Number of Person-Years of Observation, Observed and

Expected Cancer Cases, Standardized Cancer Incidence Ratios

(SIRs), and 95% Confidence Intervals (CIs), by Age, for the

New York State Female Farm Resident Cohort, 1980-1993

Person-

Age group (yr) years Observed Expected SIR 95% CI

30-39 10,489 5 13.1 0.38 0.12, 0.95

40-44 11,663 20 26.9 0.74 0.45, 1.16

45-49 12,187 52 42.9 1.21 0.91, 1.59

50-54 12,132 42 58.4 0.72 0.52, 0.98

55-59 12,637 45 81.1 0.56 0.40, 0.75

60-64 11,860 73 105.7 0.69 0.54, 0.87

65-69 7,962 70 92.4 0.76 0.59, 0.96

70+ 3,817 29 55.5 0.52 0.35, 0.75

Total 82,744 336 475.9 0.71 0.63, 0.79

Table 5.–Observed and Expected Cancer Cases, Standardized

Cancer Incidence Ratios (SIRs), and 95% Confidence Intervals

(CIs), by Specific Cancer Sites, for the New York State Female

Farm Resident Cohort, 1980-1993

Cancer cases

ICD-9

Cancer site/type code (17) Observed Expected SIR 95% CI

Oral 140-149 4 7.6 0.52 0.14, 1.45

Stomach 151 2 4.6 0.43 0.05, 1.88

Colon-rectum 153-154 36 55.5 0.65 0.45, 0.90

Liver 155 3 2.3 1.33 0.27, 4.36

Pancreas 157 5 9.7 0.51 0.17, 1.27

Larynx 161 2 3.6 0.55 0.06, 2.38

Lung 162 21 64.0 0.33 0.20, 0.51

Melanoma 172 9 8.9 1.01 0.46, 1.97

Breast 174 141 159.0 0.89 0.75, 1.05

Uterus 179, 182 40 38.0 1.05 0.75, 1.44

Cervix 180 8 13.0 0.61 0.26, 1.25

Ovary 183 11 22.9 0.48 0.24, 0.88

Bladder 188 5 10.3 0.49 0.16, 1.21

Kidney 189 8 8.7 0.92 0.40, 1.88

Brain 191 4 6.1 0.66 0.18, 1.83

Thyroid 193 9 5.1 1.78 0.81, 3.48

Non-Hodgkins lymphoma 200, 202 11 14.0 0.78 0.39, 1.43

Multiple myeloma 203 3 4.4 0.68 0.14, 2.23

Leukemia 204-208 1 8.0 0.13 0.00, 1.02

Other 13 28.0 0.46 0.25, 0.81

All cancers 336 475.9 0.71 0.63, 0.79

Note: ICD = International Classification of Diseases. (17)

Table 6.–Observed (Obs) and Expected (Exp.) Cancer Cases, Standardized

Cancer Incidence Ratios (SIRs), and 95% Confidence Intervals (CIs), by

Age Group and Specific Cancer Sites, for the New York State Female Farm

Resident Cohort, 1980-1993

Age 30-49 yr

Cancer site/type Obs. Exp. SIR 95% CI

Colon-rectum 3 4.2 0.71 0.14, 2.34

Liver 1 0.3 3.42 0.04, 27.92

Lung 2 5.9 0.34 0.04, 1.47

Breast 43 36.7 1.17 0.85, 1.58

Uterus 6 5.5 1.08 0.40, 2.47

Cervix 5 5.1 0.98 0.32, 2.44

Ovary 2 4.8 0.42 0.05, 1.82

Bladder 0 1.0 0.00 —

Thyroid 3 2.2 1.36 0.27, 4.47

Non-Hodgkins lymphoma 2 2.2 0.90 0.10, 3.91

Other 10 14.8 0.68 0.32, 1.27

All cancers 77 82.8 0.93 0.73, 1.16

Age 50-69 yr

Cancer site/type Obs. Exp. SIR 95% CI

Colon-rectum 28 41.7 0.67 0.45, 0.98

Liver 2 1.6 1.25 0.14, 5.42

Lung 15 49.2 0.31 0.17, 0.51

Breast 86 107.7 0.80 0.64, 0.99

Uterus 32 29.0 1.10 0.75, 1.56

Cervix 3 7.3 0.41 0.08, 1.34

Ovary 9 16.1 0.56 0.25, 1.09

Bladder 4 7.6 0.52 0.14, 1.45

Thyroid 6 2.6 2.30 0.84, 5.25

Non-Hodgkins lymphoma 9 10.0 0.90 0.41, 1.76

Other 36 64.6 0.56 0.39, 0.77

All cancers 230 337.5 0.68 0.60, 0.78

Age 70+ yr

Cancer site/type Obs. Exp. SIR 95% CI

Colon-rectum 5 9.6 0.52 0.17, 1.29

Liver 0 0.3 0.00 —

Lung 4 8.9 0.45 0.12, 1.24

Breast 12 14.5 0.82 0.43, 1.47

Uterus 2 3.4 0.59 0.07, 2.54

Cervix 0 0.6 0.00 —

Ovary 0 2.1 0.00 —

Bladder 1 1.6 0.63 0.01, 5.14

Thyroid 0 0.2 0.00 —

Non-Hodgkins lymphoma 0 1.9 0.00 —

Other 5 12.4 0.40 0.13, 1.00

All cancers 29 55.6 0.52 0.35, 0.75

The authors thank Mr. Patrick Hooker, Mr. Rick Zimmerman, Mr. Kyle Stewart, and Mr. Matthew Waters of the New York State Farm Bureau; Dr. Robert Lemmerman of the New York State Department of Motor Vehicles; Ms. Amy Kahn of the Bureau of Chronic Disease Epidemiology and Surveillance of the New York State Department of Health; and Mr. Gene Therriault and his staff in the Bureau of Biometrics and Records Access of the New York State Department of Health, for their invaluable assistance in conducting this research project.

Submitted for publication October 5, 2001; accepted for publication November 26, 2001.

Requests for reprints should be sent to Ying Wang, Ph.D., Bureau of Environmental and Occupational Epidemiology, Center for Environmental Health, New York State Department of Health, Flanigan Square, 547 River Street, Troy, New York 12180-2216.

E-mail: wxy01@health.state.ny.us

References

(1.) Dutkiewicz J, Jablonski L, Olenchock S. Occupational biohazards: a review. Am J Ind Med 1988; 14:605-23.

(2.) Blair A, Zahm SH. Cancer among farmers: a review. Occup Med 1991; 6:335-54.

(3.) Acquavella J, Olsen G, Cole P, et al. Cancer among farmers: a meta-analysis. Ann Epidemiol 1998; 8:64-74.

(4.) Anderson A, Barlow L, Kjaerheim K, et al. Work-related cancer in the Nordic countries. Scan J Work Environ Health 1999; 25: 1-116.

(5.) Donna A, Crosignani P, Robutti F, et al. Triazine herbicides and ovarian epithelial neoplasms. Scand J Work Environ Health 1989; 15:47-53.

(6.) Falck F, River A, Wolff MS, et al. Pesticides and polychlorinated biphenyl residues in human breast lipid and their relationship to breast cancer. Arch Environ Health 1992; 47:143-46.

(7.) Vineis P, Terrancini B, Ciccone G, et al. Phenoxy herbicides and soft-tissue sarcoma in female rice weeders. Scand J Work Environ Health 1986; 13:9-17.

(8.) Zahm SH, Weisenburger DD, Saal RC, et al. The role of agricultural pesticide use in the development of non-Hodgkin’s lymphoma in women. Arch Environ Health 1993; 48:353-58.

(9.) Wiklund K, Dich J. Cancer risks among female farmers in Sweden. Cancer Causes Control 1994; 5:449-57.

(10.) Kristensen P, Andersen A, Irgens LM, et al. Incidence and risk factors among men and women in Norwegian agriculture. Scand J Work Environ Health 1996; 22:14-26.

(11.) Folsom AR, Zhang S, Sellers TA, et al. Cancer incidence among women living on farms: Findings from the Iowa Women’s Health Study. J Occup Environ Med 1996; 38:1171-76.

(12.) U.S. Bureau of the Census, U.S. Department of Commerce. 1992 Census of Agriculture. Vol 1: Pt 51, Ch 2, United States Summary and State-level Data. Washington, DC: Bureau of the Census, 1992.

(13.) Stark AD, Chang H, Fitzgerald EF, et al. A retrospective cohort study of cancer incidence among New York State Farm Bureau members. Arch Environ Health 1990; 45:155-62.

(14.) Personal communication, New York Farm Bureau, Glenmont, New York. January 2000.

(15.) U.S. Bureau of the Census. Census of Population and Housing, 1980. Washington, D.C.: Department of Commerce, 1982.

(16.) New York State Department of Health. New York State Cancer Registry. Cancer incidence and mortality. Albany, New York: New York State Department of Health, 1987.

(17.) International Classification of Diseases, 9th rev., ICD-9-CM. Washington, DC: U.S. Department of Health and Human Services (DHHS), Public Health Service (PHS); March 1989. DHHS publ. no. (PHS) 89-1260.

(18.) Blair A, Zahm SH, Pearce NE, et al. Clues to cancer etiology from studies of farmers. Scand J Work Environ Health 1992; 18:209-15.

(19.) Wiklund K, Holm LE. Trends in cancer risks among Swedish agricultural workers. J Natl Cancer Inst 1986; 77:657-64.

(20.) Lynge E, Thygesen L. Occupational cancer in Denmark: cancer incidence in the 1970 census population. Scand J work Environ Health 1990; 16 (suppl):1-35.

(21.) Wilklund K. Swedish agricultural workers. A group with decreased risk of cancer. Cancer 1983; 51:566-68.

(22.) Burmeister LF. Cancer mortality in Iowa farmers, 1971-1978. J Natl Cancer Inst 1981; 66:461-64.

(23.) Mastrangelo G, Marzia V, Marcer G. Reduced lung cancer mortality in dairy farmers: is endotoxin exposure the key factor? Am J Ind Med 1996; 30:601-09.

(24.) Sterling TD, Weinkam JJ. Smoking characteristics by type of employment. J Occup Med 1976; 18:743-54.

(25.) Blair A, Hoar SK, Walrath J. Comparison of crude and smoking-adjusted standardized mortality ratios. J Occup Med 1985; 27: 881-84.

(26.) Duell EJ, Millikan RC, Savitz DA, et al. A populational-based case-control study of farming and breast cancer in North Carolina. Epidemiology 2000; 11:523-31.

(27.) Aronson KJ, Howe FG. Utility of a surveillance system to detect associations between work and cancer among women in Canada, 1965-1991. J Occup Med 1994; 36:1174-79.

(28.) Cantor KP, Stewart P, Brinton LA, et al. Occupational exposures and female breast cancer mortality in the United States. J Occup Environ Med 1995; 37:336-48.

(29.) Settimi L, Comba P, Carrieri P, et al. Cancer risk among female agricultural workers: a multi-center case-control study. Am J Ind Med 1999; 36:135-41.

(30.) Khuder SA, Schaub EA, Keller-Byrne JE. Meta-analysis of non-Hodgkin’s lymphoma and farming. Scand J Work Environ Health 1998; 24:255-61.

(31.) Zahm SH, Blair A. Pesticides and non-Hodgkin’s lymphoma. Cancer Res 1992; 52(Suppl): 5485s-5488s.

(32.) Blair A, Cantor KP, Zahm SH. Non-Hodgkin’s lymphoma and agricultural use of the insecticide lindane. Am J Ind Med 1998; 33:82-87.

(33.) Frich L, Akslen LA, Glattre E. Increased risk of thyroid cancer among Norwegian women married to fishery workers–a retrospective cohort study. Br J Cancer 1997; 76:385-89.

(34.) Galanti MR, Hansson L, Bergstrom R, et al. Diet and risk of papillary and follicular thyroid carcinoma: a population-based case-control study in Sweden and Norway. Cancer Causes Control 1997; 8:205-14.

(35.) Galanti MR, Sparen P, Karlsson A, et al. Is residence in area of endemic goiter a risk factor for thyroid cancer? Int J Cancer 1995; 61:615-21.

(36.) Kalk WJ, Stitas F, Patterson AC. Thyroid cancer in South Africa-an indicator of regional iodine deficiency. S Afr Med J 1997; 87: 735-38.

(37.) Negri E, Ron E, Franceschi S, et al. A pooled analysis of case-control studies of thyroid cancer. Cancer Causes Control 1999; 10: 131-66.

(38.) McGlashan ND. Primary liver cancer and food-based toxins. A Swaziland review. Ecol Dis 1982; 1:37-44.

(39.) Bulatao-Jayme J, Almero EM, Castro MC, et al. A case-control dietary study of primary liver cancer risk from aflatoxin exposure. Int J Epidemiol 1982; 11:112-19.

(40.) Autrup H, Seremet T, Wakhisi J, et al. Aflatoxin exposure measured by urinary excretion of aflatoxin B1-guanine adduct and hepatitis B virus infection in areas with different liver cancer incidence in Kenya. Cancer Res 1987; 47:3430-33.

(41.) Stemhagen A, Slade J, Airman R, et al. Occupational risk factors and liver cancer. A retrospective case-control study of primary liver cancer in New Jersey. Am J Epidemiol 1983; 117:443-54.

(42.) Wiklund K, Rich J. Cancer risks among male farmers in Sweden. Eur J Cancer Prev 1995; 4:81-90.

(43.) Dossing M, Peterson KT, Vyverg M, et al. Liver cancer among employees in Denmark. Am J Ind Med 1997; 32:248-54.

(44.) Wiklund K, Dich J, Holm LE, et al. Risk of cancer in pesticide applicators in Swedish agriculture. Br J Ind Med 1989; 46:809-14.

YING WANG

ELIZABETH L. LEWIS-MICHL

SYNI-AN HWANG

EDWARD F. FITZGERALD

ALICE D. STARK

Bureau of Environmental

and Occupational Epidemiology

Center for Environmental Health

New York State Department of Health

Troy, New York

COPYRIGHT 2002 Heldref Publications

COPYRIGHT 2003 Gale Group