Regional Variation in Breast Cancer Rates in the United States (Past Initiative)

Five institutions are being funded to conduct research using epidemiologic and statistical methods for determining whether various factors may account for the geographic differences in breast cancer rates in the United States. The projects funded in 1999 are supported by EGRP in collaboration with the National Institute of Environmental Health Sciences (NIEHS).

For this initiative, research proposals were sought through a Request for Applications to stimulate research to better understand determinants of regional variation in breast cancer incidence and mortality. This RFA was an extension of an earlier initiative that focused on the northeastern and mid-Atlantic regions of the United States. See the report of the Northeast and Mid-Atlantic Breast Cancer Study for details.

Research has demonstrated that some variation in breast cancer rates can be explained by differences in the population distribution of known breast cancer risk factors, such as menstrual and reproductive variables.

However, regional patterns may also reflect an aggregate of diverse factors including, for example, varying presence of hazards in the environment, demographics and lifestyle of a mobile population, subgroups of susceptible individuals, and changes and advances in medical practice and health care management. Disentangling these factors is necessary to assess associations, singly or jointly, with breast cancer risk in individuals and populations.

• Variations in Breast Cancer Treatment and Mortality
  James Goodwin, MD, University of Texas Medical Branch at Galveston
• Regional Variation of Breast Cancer Rates in Wisconsin
  Patrick Remington, MD, MPH, University of Wisconsin
• Regional Variations in Breast Cancer Rates in California
  Peggy Reynolds, PhD, California Department of Health Services
• Geographic Distribution of Breast Cancer
  Joseph Sheehan, PhD, University of Connecticut School of Medicine
• GST, Environment Factors And Breast Cancer Risk
  Tongzhang Zheng, MD, ScD, Yale University School of Medicine

Variations in Breast Cancer Treatment and Mortality

Mortality rates currently provide our major source of data on the national burden of cancer. For breast cancer, there have been pronounced geographic variations in mortality rates, both at the level of large areas (e.g., the Northeast versus the South) and at the level of small areas (e.g., Long Island versus surrounding areas). Until recently, it has been largely assumed that such variations in mortality rates reflect variations in incidence. There is no doubt that there are important variations in incidence within the United States, both across different populations and different geographic areas.

In addition to incidence, however, the other major contributor to mortality rate is survival. This study focuses on geographic variations in survival from breast cancer. The researchers postulate that variations in survival from the cancer among older women are responsible in part for the variations in breast cancer mortality. They further hypothesize that these variations in breast cancer survival are secondary to potentially remediable causes, which include stage at diagnosis and treatment received.

They previously showed that older women with breast cancer are more likely to be diagnosed at a more advanced stage, and are less likely to receive definitive treatment. The percentage of older women who receive less than definitive treatment varies substantially by region of the country and by small areas within regions.

During a 4-year study, the researchers will use the Surveillance Epidemiology and End Results (SEER) Program-Medicare linked database to examine variation in breast cancer incidence, survival, and mortality by health service area within all SEER sites. The availability of the SEER Medicare data, which can be linked to other data files such as the Area Resource File, will allow them to more completely characterize:

• patient characteristics (age, ethnicity, socioeconomic status at the level of the census tract, etc.),
• tumor characteristics (size, stage, histology),
• treatment received (definitive versus non-definitive, chemotherapy, etc.),
• medical system characteristics (size of hospital, presence of a cancer center, teaching status of hospital, etc.),
• follow-up surveillance (routine mammography after initial treatment), and
• outcomes (death, recurrence of cancer, other complications).

The researchers can then determine the relative contributions of geographic variations in breast cancer incidence, versus variations in breast cancer survival to the observed geographic variations in mortality from the disease.

They can also calculate the relative contributions of variations in patterns of breast cancer diagnosis, extent of disease, treatment, population characteristics, and medical system factors, to the geographic variations in breast cancer survival.

Published Papers:

• Du X, Goodwin JS. Patterns of use of chemotherapy for breast cancer in older women: findings from Medicare claims data. J Clin Oncol 2001 Mar 1;19(5):1455-61.
• Du X, Goodwin JS. Increase of chemotherapy use in older women with breast carcinoma from 1991 to 1996. Cancer 2001 Aug 15;92(4):730-7.
• Du Xianglin, Freeman JL, Nattinger AB, Goodwin JS. Survival of women after breast conserving surgery for early stage breast cancer. Breast Cancer Res Treat 2002 Mar;72(1):23-31.
• Goodwin JS, Freeman JL, Mahnken JD, Freeman DH, Nattinger AB. Geographic variations in breast cancer survival among older women: implications for quality of breast cancer care. J Gerontol A Biol Sci Med Sci 2002 Jun;57(6):M401-6.

Regional Variation of Breast Cancer Rates in Wisconsin

The researchers are conducting a population-based study to explain the regional variation in breast cancer rates in Wisconsin. The 4-year investigation, building upon past and ongoing case-control studies of breast cancer conducted at the University of Wisconsin Comprehensive Cancer Center, hypothesizes that the statewide variations in breast cancer rates are due to regional variations in established or suspected breast cancer risk factors. The specific aims include:

• comparing breast cancer incidence, mortality, and survival rates within Wisconsin by geographic area;
• evaluating the contribution of previously identified breast cancer risk factors, socioeconomic status, and screening practices to the regional variations in breast cancer rates; and
• evaluating the association between environmental exposures (such as PCBs, DDT, and other xenoestrogens) and breast cancer risk, and if associated, the contribution of these exposures to the regional variations in breast cancer rates.

In Phase I, the researchers will assign a "geographic breast cancer risk" to every case and control already enrolled in their breast cancer case-control studies (n=14,000) by analyzing breast cancer mortality and incidence rates by area of residence (county, city, zip code, and census tract). Next, this risk will be adjusted by controlling for established and potential individual- and community-level risk factors for breast cancer.

Phase II will expand the ongoing breast cancer case-control study (n=6,600) by collecting information on environmental exposures and determining their contribution to the regional variation in breast cancer rates. Biologic specimens (buccal smears, urine, and blood) will be assayed to evaluate genetic risk and markers of biologic exposure.

By utilizing existing populations, protocols, personnel, and software, the investigators will be able to complete these evaluations in an efficient, timely, and statistically valid manner. This approach will enable them, in collaboration with the NCI, NIEHS, and other funded researchers, to examine in detail the contribution of a variety of individual- and community-level exposures to the long-standing regional variation in breast cancer rates in Wisconsin, and will improve understanding of the causes of breast cancer.

Regional Variations in Breast Cancer Rates in California

Observed regional variations in breast cancer incidence have been a source of public health concern, as well as traditionally a source of hypothesis generation for factors which might be implicated in disease etiology. Rate differences within California are similar to differences noted on the national scale. This 3-year study is exploring factors that may explain the observed regional variations within California. The specific aims are to:

• use a geographic information system (GIS) and grouped data sources to evaluate the potential influence of sociodemographic factors (e.g., income, education), environmental toxicants (e.g., pesticides, automobile exhaust), and established risk factors (e.g., family history, reproductive risk factors) on regional differences in breast cancer incidence rates in California; and
• conduct a biomarker substudy targeted at exploring urban/rural variations in biological markers of two specific exposures of emerging interest in breast cancer research (i.e., phytoestrogens, polycyclic aromatic hydrocarbons).

Statewide grouped analyses will rely on multiple sources of data, including the largest cancer registry in the world, two statewide risk factor surveys, multiple statewide environmental datasets, and census data. The biomarker substudy will be nested within this cohort and will be conducted on 130 urban and 130 rural members of the California Teacher Studies cohort in order to evaluate both dietary and environmental markers for exposures potentially related to regional variations in breast cancer.

These multiple sources of data, coupled with the use of a GIS system, offer a unique opportunity to integrate individual risk factor information with potentially important community-level factors in a way that no other study has done to date. Further, the biomarker substudy will allow the researchers to evaluate two emerging and intriguing hypotheses regarding breast cancer etiology in the context of urban/rural variations in breast cancer rates.

Published Papers:

• Hurley SE, Saunders TM, Nivas R, Hertz A, Reynolds P. Post office box addresses: a challenge for geographic information system-based studies. Epidemiology 2003 Jul;14(4):386-91.
• Bonner MR, Han D, Nie J, Rogerson P, Vena JE, Freudenheim JL. Positional accuracy of geocoded addresses in epidemiologic research. Epidemiology 2003 Jul;14(4):408-12. Erratum in: Epidemiology 2003 Nov;14(6):736.
• Reynolds P, Hurley SE, Gunier RB, Yerabati S, Quach T, Hertz A. Residential proximity to agricultural pesticide use and incidence of breast cancer in California, 1988-1997. Environ Health Perspect 2005 Aug;113(8):993-1000.
• Reynolds P, Hurley SE, Quach AT, Rosen H, Von Behren J, Hertz A, Smith D. Regional variations in breast cancer incidence among California women, 1988-1997. Cancer Causes Control 2005 Mar;16(2):139-50.
• Hurley SE, Reynolds P, Goldberg DE, Hertz A, Anton-Culver H, Bernstein L, Deapen D, Peel D, Pinder R, Ross RK, West D, Wright WE, Ziogas A, Horn-Ross PL. Residential mobility in the California Teachers Study: implications for geographic differences in disease rates. Soc Sci Med 2005 Apr;60(7):1547-55.

Geographic Distribution of Breast Cancer

The overall goal of this 2-year project is to determine whether the elevated rate of breast cancer in Massachusetts can be considered to vary from place to place at random throughout the state, or whether the rate is excessive in specific geographic areas. If there are areas of excess, the goal is to determine whether that excess is stable or temporary over the study years, whether excesses are consistent across all diagnostic stages, or whether excesses might be due, for example, to excesses in early or late stage diagnoses, and if they can be attributed to covariates such as age, fertility, race/ethnicity, education, or economic conditions.

The specific aims are to:

• test for the presence of statistically significant spatial clusters of excess breast cancer, and to do so at the level of the census tract, ZIP code, and town;
• test for statistically significant spatial clusters of excess breast cancers by diagnostic stage;
• if statistically significant clusters are found, to test whether the excesses at those locations have been consistent and continuous over the study years, or are temporary or sporadic; and
• test whether clusters remain after adjustment for selected social, economic, and demographic measures from the U.S. census, such as race/ethnicity, education, fertility, and economic factors.

The incidence file contains information on 57,560 cases diagnosed between 1982 and 1994, and includes the year, month, and day of diagnosis, the summary stage based on SEER's staging convention (in situ, local, regional, distant, and unknown), race, city/town, ZIP code, census tract, and latitude/longitude of the patient's address.

Analyses will be conducted with data aggregated to census tracts, ZIP codes, towns, along with disaggregated analyses at the level of the individual case using the Spatial Scan statistic to determine whether areas of excess can be "explained" by chance or other covariate information, whether areas of excess are stable over time, or if they vary by diagnostic stage.

GST, Environment Factors And Breast Cancer Risk

Recent epidemiologic studies suggest that genetic polymorphisms of the glutathione s-transferase (GST) may affect an individual's risk of developing breast cancer. The alleged role of GST is hypothesized due to the observation that GST gene products catalyze the conjugation with glutathione of potentially cytotoxic and genotoxic reactive chemical intermediates to yield inactive products, thereby, providing a protective mechanism against cancer development. Since the suggested high-risk GST genotypes commonly occur in the general population, the calculated population attributable risk can be high.

To confirm the initial reports, a case-control study of GST genetic polymorphisms, environmental factors, and breast cancer risk is being conducted in Connecticut, a state with a high incidence rate of breast cancer. The primary aim of this 2-year study is to examine the association between genetic variability in three major GSTs – GSTM1, GSTT1, and GSTP1, and an individual's susceptibility to breast cancer. Research questions include:

• Is there a lack of, or reduced expression of, these GSTs associated with an increased risk of breast cancer?
• Does the risk of breast cancer increase as the number of putative high-risk GST genotypes increases?
• And do these GSTs modify the association between environmental factors and the risk of breast cancer?

Blood samples, previously collected in a recently completed breast cancer study of 349 women diagnosed with breast cancer (cases) and 363 control women (no breast or other forms of cancer), are being assayed for GST genotypes. Information on environmental exposures, including blood levels of certain pesticides and polychlorinated biphenyls (PCBs), known or suspected breast cancer risk factors, and potential confounding variables have also been collected from all study participants.

Findings from this study will contribute to the determination of whether genetic polymorphisms of GST are associated with breast cancer.