WEDNESDAY, November 13, 2002

"Geographic Variations in Breast Cancer Incidence: Focus on the Bay Area"

Our guests for the evening are Tina Clarke, PhD., researcher from the Northern California Cancer Center, and Janice Barlow, executive director of Marin Breast Cancer Watch. Expanded reporting and better tracking of cancer incidence over the last few decades has prompted this investigation of geographic variations and risk factors that may be related to the variations that we have seen. Registration and tracking of patients is very difficult and although there is a lot of effort going into it, this is still not a refined process. What is clear is that there is a trend in breast cancer incidence, which is county-based.  Many issues including population numbers and what makes up that population confound data, which shows higher rates in Marin County. There may be other areas of San Francisco, which have higher incidence of breast cancer, but because they aren’t counties, they aren’t counted.

In 1987 cancer became a reportable disease in California. State law requires that any facility (hospital, physician, laboratory, treatment center, etc.) which sees a patient with cancer must report it to the regional cancer registry. Because a patient is frequently seen at multiple facilities, all of which are required to report, a large part of the work of the Cancer Registry is to aggregate and consolidate information on patients. This reporting is done by measuring all death certificates that mention cancer against patients listed in the registry. Over time, this reporting is about 99% accurate. The registry identifies individual patients, but researchers only see the counts, so the information is protected, secure and confidential. The basic work of the registry is to monitor cancer rates and trends in the community. Data is collected on age, sex, race, address, tumor size, stage and treatment, anything that is in the medical record. Each patient is followed for life in order to gain survival information. Health statistics in the U.S. are not broken down further than county level apart from every ten years when the census is taken. This makes it difficult in a state like California where some of our counties have the population of some states. Some examples:

The data from the 2000 census is not yet available. Because questions were asked in a new way, the information will have to be made comparable before it can be evaluated. The NCCC puts out an annual report, monitoring about 45 different cancers including in situ and invasive breast cancers in different age, racial and ethnic groups. In areas like Bayview/Hunters Point, which has a large Afro-American population and very high mortality rates, there is no way to monitor trends because information is not available in an area smaller than a county. We do know that the rates of cancer seen in Bayview/Hunters Point are as expected but the prognosis is worse per cancer diagnosed. Screening has improved, but access to treatment still needs improvement.

 It was noticed that Marin County (80% white, non-Hispanic) had an increasing rate of breast cancer over the last decade which was not observed in other counties. Breast cancer rates vary enormously by ethnicity, so Marin County (one of the smallest urban counties in the U.S. and 80% white) with a 37% increase in breast cancer, did not experience any dilution effect due to it’s homogeneous population. Rates in white women are 75% higher than in Latinas and 25% higher than in African-Americans. The relationship between socioeconomic status and breast cancer has been a longstanding one. We also know that rates of breast cancer are higher in educated populations. These are difficult variables to monitor. Prehn sorted and compared census block groups in Marin and other areas similar to Marin in socioeconomics and education. He found the rates to be comparable. The conclusion that was reached was that Marin is not geographically unique. More important than  living in Marin is the type of people who live there. The question then becomes, “Should Marin County be considered the canary in the coal mine?”  The California Teachers Study (87% white women) collected information on known risk factors in teachers and administrators. It validates the research done in Marin in that there is a positive correlation between breast cancer and higher socioeconomic status, higher education, fewer children born later, use of hormone therapy (HRT), and alcohol consumption. Women at highest risk are between 45 and 64 years of age and are primarily hormone receptor positive and their cancer is of lobular and/or ductal histology. Marin may just be a bellwether for new treatments. It was in Marin County women that the NCCC first saw a large spike in uterine cancer after using unopposed estrogen in women with intact uteri.

How confident are we about these numbers? The Cancer Registry is confident about the numerator, that is the number of actual cancer cases, but the denominator is just a best guess between two ten-year (census) periods. This can be faulty, particularly in undercounting the 45-64 age group.

What is currently being done to look at the environmental effects on breast cancer? The Marin Breast Cancer Watch was founded in 1995, to explore the environmental links to breast cancer. They chose to look at adolescent risk factors in the development of breast cancer in Marin County. The pilot study developed a memory tool to take women back to their adolescence to look at family history, diet, stress, and some environmental factors. Then there was a 2-1/2 year population-based case/control study of 300 women with breast cancer matched with 300 controls. This study was just completed, and is being peer reviewed, so no information is yet available on the results. A new collaboration with Lawrence Berkeley Labs and the Marin Health Department is collecting databases on pesticides and toxic dumpsites and archiving information to see if any information can be specifically linked to breast cancer. Another pilot study is developing a questionnaire to elicit information on the personal environment to see if there are any modifiable risk factors. Mapping of breast cancer cases in Marin County has also been done, but there are so few cases in some towns that 1 or 2 cases in either direction would significantly alter the diagram. This project was unable to control for established risk factors.

How important is it to know where you lived at the time of diagnosis? This can be very confusing. Some women live in San Francisco, but work in Marin, spending most of their time there. Cancer takes a significant amount of time to develop, so it is more important to know how long you lived in Marin.

What is the evidence of environmental exposure causing cancer? There are a huge number of cancers that are caused by occupational exposure to toxins. The first described was testicular cancer in chimney sweeps. Life-threatening radiation exposure is well documented from World War II in Japan. Secondhand smoke from cigarettes and asbestos exposure are still controversial even though they are known to cause cancer. Retrospective studies rely on memory, which is tricky. Also, we may not have the information (what our houses were built on, what is in the fish we eat, etc.). If we got a grant today to study adolescent risk factors in breast cancer, we wouldn’t be alive to see the results.

How effective or informative is it to geographically map cancer incidence? Mapping of cancers relative to residence at the time of diagnosis is problematical. It is more important to document risk factors at the time that the cancer was initiated. The latency period with most cancers is 20 to 30 years or more, so it is very difficult to track exposures, although it is easier with very rare cancers than with a common cancer like breast cancer.

How are risk factors determined, which ones are collected? A risk factor like smoking is not collected because it is not necessarily in the medical record and must be recorded by 95% of patients to be considered. In order to calculate a risk factor cancer rate, it is necessary to know both the numerator (women with breast cancer who smoke) and the denominator (all women who smoke in the population being studied).

It seems that most of the breast cancers diagnosed in Marin are lobular? The way lobular cancer is diagnosed has changed along with the way we do estrogen receptors, HER2/neu staining, etc., but it is true that the majority of the increase seen in Marin has been lobular or mixed ductal/lobular, which is not the majority of breast cancers. Christopher Benz, M.D., a research physician here at UCSF and medical director of the Buck Center on Aging is looking at trends in lobular disease. It appears to be strongly associated with hormone therapy, with 95% of lobular cancers being hormone receptor positive. But the entire increase in breast cancer cannot be attributed to hormonal factors; we have not ruled out toxic exposures. These are more likely to be toxins we have been exposed to throughout our lives rather than something sprayed on our lettuce. If we are to look at toxins, it should be exposures during breast development.

Are risk factors modifiable? Some of the risk factors that we see in more affluent communities are definitely modifiable. Information on hormone replacement therapy has been well documented. The most recent meta-analysis showed increased risk of breast cancer based on 10-gram increments of alcohol per week (one glass of wine versus two or three).  Adult daily consumption of two or more drinks of alcohol a day substantially increases your risk of breast cancer. This may be due to disturbance in the hepatic metabolism of estrogen, but the actual mechanism is not yet totally clear. The dietary message seems to be one of moderation.

What is the optimal time to breast-feed? Another meta-analysis was done in Britain. They looked at the duration of breast-feeding and the number of children. There was a very strong correlation in reduction in breast cancer risk with the number of months of breast-feeding with a particular drop after nine months. The lowest risk came with breast-feeding each child for 12 months or more. The thinking is that the reduction is caused by the suppression of ovarian function, which is usually (but not always) present

during lactation. New research seems to show that there are structural changes in the breast after breast-feeding.

Next meeting Wednesday, December 18, 2002. Topic: “Update from the San Antonio Breast Cancer Meeting, 2002”.