SEARCH

SEARCH BY CITATION

Keywords:

  • breast cancer;
  • diet;
  • environmental pollutant;
  • mammary gland carcinogenesis;
  • endocrine disrupting compound;
  • toxicology;
  • evidence based medicine;
  • physical activity;
  • obesity

Abstract

  1. Top of page
  2. Abstract
  3. RESEARCH AREAS WITH PROMISE FOR RISK REDUCTION
  4. ONLINE DATABASES
  5. REVIEWS IN THIS ISSUE
  6. FUTURE RESEARCH
  7. CONSIDERATIONS IN COMMUNICATING WITH PATIENTS, POLICYMAKERS, AND THE PUBLIC
  8. Acknowledgements
  9. REFERENCES

Breast cancer is the most common invasive cancer in women worldwide and the leading cause of death in US women in mid-life. Treatment has adverse effects, adding to the importance of finding modifiable risk factors. At the invitation of Susan G. Komen for the Cure, we reviewed studies of breast cancer and environmental pollutants, diet (assessed prospectively), body size, and physical activity, and animal studies that identify chemicals as potential mammary carcinogens. Databases developed in the review include information on 216 chemicals that increased mammary gland tumors in animal studies and 450 epidemiologic studies (accessible at www.silentspring.org/sciencereview and www.komen.org/environment). Exposure to potential mammary carcinogens is widespread from chemicals found in consumer products, air and drinking water pollution, food, and women's workplaces. Epidemiologic studies have included only a small number of chemicals identified as mammary carcinogens or as hormone disruptors, which may have implications for breast cancer; however, evidence is emerging for associations between breast cancer and polychlorinated biphenyls, polycyclic aromatic hydrocarbons, and organic solvents. Prospective diet studies have not revealed consistent associations with breast cancer. Improved exposure assessment methods will help advance future human studies of both diet and environmental pollutants. Studies of physical activity show that it is protective. In the same vein as evidence-based medicine, messages for patients, policymakers, and the public should support decision-making based on the strength of current evidence; such messages might address exposure reduction for some pollutants. Investments in research on environmental factors in breast cancer have potentially large public health benefits. Cancer 2007. © 2007 American Cancer Society.

Breast cancer is the most common invasive cancer in women worldwide,1 and beyond the sheer number of women affected the public health significance of the disease stems also from the pattern of risk for women in mid-life when they are raising children and contributing to work and communities. Breast cancer is the leading cause of death in US women between their late 30s and early-50s; a 45-year-old woman is 35% more likely to die of breast cancer than of ischemic heart disease and 2.5 times more likely to die of breast cancer than in a motor vehicle accident.2 (The comparative risk of breast cancer and 11 other most common underlying causes of death in 1999–2003 in women aged 25 to 65 is shown in Fig. 1.) Half of breast cancers diagnosed in recent years (1998–2002) were in women aged 61 and younger.3

thumbnail image

Figure 1. Average annual number of deaths by age for US females aged 25 to 65 years in 1999–2003: 12 leading causes.

Download figure to PowerPoint

Screening and improved treatment have contributed to improved survival,4–6 the incidence appears to have stabilized after rising for decades,7 and mortality has declined in the US,4 but these trends are discouraging in contrast with progress against other major killers: heart disease and lung cancer in men.8 Furthermore, although we hope and expect that breast cancer survival will continue to improve, treatment is likely to remain arduous and debilitating for the foreseeable future, with potential adverse effects on cardiovascular health, secondary cancers, physical mobility, cognition, sexuality, and social factors.9–12 Financial costs of treatment are substantial, amounting to $8.1 billion in the US in 2004 according to the National Cancer Institute.13

In this context, the breast cancer advocacy community has called for greater attention to research into breast cancer prevention as a major public health priority alongside screening and treatment.14, 15 Statistics on geographic and temporal variation and the poor prediction of individual risk from established risk factors16–18 provide evidence that additional risk factors remain to be identified, lending scientific support for the advocates' hopes. A 5-fold variation in incidence rates across the world indicates that characteristics of industrial societies increase risk.19 Incidence is rising rapidly in developing nations and in US immigrant populations.19–23 Based on large studies of twins, researchers estimate that factors other than inherited genes are involved in the vast majority of breast cancers.24, 25 Even among women with high-risk BRCA1 and BRCA2 gene mutations, evidence suggests that nongenetic factors influence risk, given that a higher proportion of carriers born after 1940 were diagnosed by age 50 than carriers born before 1940.26

RESEARCH AREAS WITH PROMISE FOR RISK REDUCTION

  1. Top of page
  2. Abstract
  3. RESEARCH AREAS WITH PROMISE FOR RISK REDUCTION
  4. ONLINE DATABASES
  5. REVIEWS IN THIS ISSUE
  6. FUTURE RESEARCH
  7. CONSIDERATIONS IN COMMUNICATING WITH PATIENTS, POLICYMAKERS, AND THE PUBLIC
  8. Acknowledgements
  9. REFERENCES

At the invitation of Susan G. Komen for the Cure, we undertook a systematic review to identify areas with particular promise for breast cancer risk reduction and clarify what is currently known. Our goals were to evaluate and integrate findings critically, identify gaps and methodological challenges, recommend directions for future research and intervention, and assess the evidence base for messages for patients, policymakers, and the public.

We selected several topics as first priorities, because they represent factors that can be modified and areas of science where key questions are not settled: epidemiologic studies of physical activity, body size, diet, and environmental pollutants, and toxicologic studies conducted in animals to identify chemicals that may be human breast carcinogens. We limited the review of diet to prospective observational studies, because the volume of research in this field makes it reasonable to focus on these study designs, which are less vulnerable to bias in self-reported exposures and control selection.27 We are currently reviewing additional topics that may inform prevention as well: epidemiologic research on early life factors, nonhormonal pharmaceuticals, tobacco smoke, light at night, psychological stress, and toxicologic assessments of endocrine disrupting compounds. Results of previous research have been consistent in establishing that ionizing radiation, certain hormonal pharmaceuticals, age at menarche and menopause, nulliparity, age at a first full-term pregnancy, and alcohol consumption contribute to breast cancer.28, 29

For the selected topics, we conducted systematic literature searches and developed databases of citations and descriptive and analytical information for the studies we identified. In this issue, we report in separate articles on our review of the epidemiologic evidence on diet27 and environmental pollutants,30 and toxicologic assessments of mammary gland carcinogens.31 Recent reviews published elsewhere for body size32–34 and physical activity35 update and complement the comprehensive review by the International Agency for Research on Cancer (IARC)36 of articles published through 2000 for both risk factors. In this commentary we describe our primary findings and the online databases we developed, reflect on the commonalities and differences across the epidemiologic studies of diet and environmental pollutants, and discuss implications for future research. We briefly consider how to translate research for patients and the public in ways that support informed decision-making in areas where science remains uncertain.

ONLINE DATABASES

  1. Top of page
  2. Abstract
  3. RESEARCH AREAS WITH PROMISE FOR RISK REDUCTION
  4. ONLINE DATABASES
  5. REVIEWS IN THIS ISSUE
  6. FUTURE RESEARCH
  7. CONSIDERATIONS IN COMMUNICATING WITH PATIENTS, POLICYMAKERS, AND THE PUBLIC
  8. Acknowledgements
  9. REFERENCES

As the foundation for the reviews and to provide transparency for our assessments of the literature, we developed 2 databases, 1 covering epidemiologic research and the other including toxicologic information on mammary gland carcinogens. The databases may be accessed on the internet at www.silentspring.org/sciencereview and www.komen.org/environment.

The environment and breast cancer Epidemiology Reviews Database is composed of critical reviews of approximately 450 primary epidemiologic research articles on breast cancer and diet, environmental pollutants, physical activity, and body size. For each research article the database includes a complete bibliographic citation and the abstract, if it is not copyright restricted. Copyrighted abstracts can be accessed from the database via links to PubMed. We entered information about the source of the study population, number of study participants and participation rate, inclusion and exclusion criteria, exposure assessment method, study design, statistical analysis, strength of associations, and attention to gene-environment interactions, ethnic minority populations, or early life exposures. Additional fields include comments about the strengths and weaknesses of participant selection, exposure assessment, measures of association, and the interpretation of results. Articles are searchable by topic. In addition, the database includes about 50 citations to review articles, methods articles, and exposure assessments that aid in interpreting the primary research. Quick access to basic study information and critical assessments makes the database a resource for physicians, other healthcare providers, and members of the public who want to assess a particular study or field of study even if they are not familiar with methodological issues in environmental breast cancer epidemiology. The database includes articles on environmental pollutants and breast cancer published through June 2006 and in other topic areas through May 2005. We plan to update it periodically.

The Mammary Carcinogens Review Database includes 216 chemicals that increased mammary gland tumors in animal studies conducted by the US National Toxicology Program (NTP)37 or included in the IARC Monographs,38 11th Report on Carcinogens (11th ROC),39 Carcinogenic Potency Database,40 and Chemical Carcinogenesis Research Information System (CCRIS) database.41 It includes summary assessments of the carcinogenic potential for each chemical and data on mutagenicity, opportunities for exposure in the general population and for women at work, and other characteristics of chemical use, sources, and regulation. This information is valuable for regulators to consider in decisions about limiting human exposure, for manufacturers to evaluate in reformulating products and re-engineering processes to avoid suspect chemicals, and for epidemiologists to identify new chemicals, exposure scenarios, and exposed populations for breast cancer studies.

REVIEWS IN THIS ISSUE

  1. Top of page
  2. Abstract
  3. RESEARCH AREAS WITH PROMISE FOR RISK REDUCTION
  4. ONLINE DATABASES
  5. REVIEWS IN THIS ISSUE
  6. FUTURE RESEARCH
  7. CONSIDERATIONS IN COMMUNICATING WITH PATIENTS, POLICYMAKERS, AND THE PUBLIC
  8. Acknowledgements
  9. REFERENCES

Considering together the 3 reviews in this issue, we see great contrast in the maturity of and attention to research on diet vs environmental pollutants. The diet literature search identified nearly 1500 candidate articles published since 1950. In contrast, the environmental pollutants of interest were first synthesized in the 1940s and came into widespread use in the 1950s; breast cancer research interest in these chemicals dates only to the 1990s.42 The overwhelming majority of chemicals identified as animal mammary carcinogens or endocrine-disrupting compounds have never been included in an epidemiologic study of breast cancer, and the overwhelming majority of chemicals to which we are exposed have never been included in an animal cancer bioassay.

Toxicologic assessments of mammary carcinogens

Our review31 is the most comprehensive compilation to date of chemicals identified as mammary carcinogens and builds on the work of others.43–45 We found that exposure to chemicals identified as potential mammary carcinogens is widespread, given that 73 have been present in consumer products or as contaminants of food, 35 are air pollutants, 29 are produced at more than 1 million pounds per year in the US, and 25 have involved occupational exposures to more than 5000 women. Nearly all of the chemicals were mutagenic and most caused tumors in multiple organs and species, providing evidence of likely carcinogenicity in humans. However, we found that these data have not been well utilized in regulatory policies designed to limit chemical exposures or in guidance materials designed to alert physicians or public health agents to potential links between chemical exposures and health effects. For example, we found no instances where employers are required to offer mammograms as part of medical screening for exposed workers.31

Ignoring animal evidence is particularly problematic because of the significant methodological challenges and ethical limits in epidemiologic studies of environmental chemicals (discussed further in the environmental pollutants review). Questions remain about the ability of the animal bioassays to predict human breast cancer risk,46, 47 and these models may be improved in future research; however, animal models are the primary means of understanding and anticipating effects of chemicals in humans. All known human carcinogens that have been tested in animals are also carcinogenic in animals.48–50 Based on these observations, the IARC, NTP, and others have concluded that chemicals with evidence of carcinogenicity in animals (in the absence of additional scientific information) are considered a carcinogenic risk to humans.51 The Mammary Carcinogens Review Database will be a convenient reference alerting policymakers and healthcare providers to the evidence.

Epidemiologic studies of diet and environmental pollutants

Results of our review reveal that prospective observational studies of diet do not provide consistent evidence of associations with breast cancer.27 We considered factors that, based on the nutritional constituents, are of particular interest in the context of breast cancer: fat intake, biomarkers of fat intake, fruit and vegetable consumption, antioxidant vitamins (vitamins A, C, E, and beta-carotene), serum antioxidants, carbohydrate intake, glycemic index and glycemic load, dairy consumption (including vitamin D), consumption of soy products and isoflavones, green tea, heterocyclic amines, and adolescent diet.

The review of environmental pollutants epidemiology30 shows that research in this area is still relatively sparse, but results have in recent years begun to show evidence of increased risk associated with exposure to polychlorinated byphenols (PCBs—banned chemicals previously used in electrical equipment and other products) in genetically susceptible women and to polycyclic aromatic hydrocarbons (PAHs), which are ubiquitous air pollutants from vehicle exhaust and other sources of combustion byproducts. Results of the studies of organic solvents and dioxin suggest a possible association and support additional research in these areas. Because of the many studies of dietary factors, we have more opportunity to expect and evaluate consistency across studies before placing confidence in results, whereas, in contrast, conclusions from the environmental pollutants epidemiology rests on relatively fewer studies. Additional support comes from animal studies that identify PAHs and some organic solvents as animal mammary carcinogens31 and PCBs and dioxins as endocrine disruptors,52 providing evidence of biological mechanisms that may link these chemicals to breast cancer.

The contrasting volume of research on diet as opposed to environmental chemicals cannot be explained by differences in the methodological challenges in the 2 fields, as we found the limitations in studies in these areas were strikingly similar. Exposure assessment, in particular, is a key weakness. Self-report is the foundation of breast cancer risk factor epidemiology to date, resulting in the identification of risks associated with family history, reproductive factors, pharmaceuticals, alcohol, and physical activity. However, self-reports of diet and environmental pollutant exposures are affected by considerable measurement error, and many environmental pollutant exposures are simply unknown to the individual. In breast cancer studies, use of self-report is complicated by evidence that exposures in multiple life cycle stages—in utero and in early life, with 20 or more years latency, and in the 5 years before diagnosis—may all be important. Regardless of the exposure assessment method, relations may be missed if the exposure is incorrectly defined, for example, hypothetically, by combining subtypes of dietary fat rather than investigating them separately. An additional problem for both diet and environmental chemicals is that the range of exposures in populations accessible for study may be limited, making it harder to detect effects. For example, the Nurses' Health Study53 provides 1 of the best resources for prospective assessment of diet in women old enough to be at substantial risk for breast cancer, but the original cohort is limited to white nurses in the US. Achieving greater contrast between low and high exposures will not be easy, because exposure conditions are often correlated with established breast cancer risk factors. For example, developing nations with more markedly different diets also differ in reproductive patterns, and occupationally exposed women with higher chemical exposures may also have higher levels of physical activity.54, 55 Even in undeveloped regions, unexposed populations are elusive: Efforts to investigate the effects of organochlorine compounds, including pesticides and PCBs, were thwarted by the discovery that global atmospheric and ocean transport resulted in substantial exposures even in the Arctic, where these compounds were not used.56

In addition, effects of diet and environmental chemicals may be obscured if they interact with genetic variations or other environmental factors, or if they are limited to subtypes of breast cancer. For example, we review30 studies of increased risk in genetically susceptible women exposed to PCBs after studies of general populations found no association. Similarly, we review30 evidence of increased risk in younger women exposed to PAHs and organic solvents, suggesting possible differences in etiology for premenopausal and postmenopausal disease, as has been observed with the protective effect of obesity in young women and increased risk postmenopausally.

FUTURE RESEARCH

  1. Top of page
  2. Abstract
  3. RESEARCH AREAS WITH PROMISE FOR RISK REDUCTION
  4. ONLINE DATABASES
  5. REVIEWS IN THIS ISSUE
  6. FUTURE RESEARCH
  7. CONSIDERATIONS IN COMMUNICATING WITH PATIENTS, POLICYMAKERS, AND THE PUBLIC
  8. Acknowledgements
  9. REFERENCES

New technologies are emerging to overcome the limitations of present study methods for both diet and environmental chemicals. These include new biological and environmental exposure measures, including assessments of mixtures; markers of disease precursors, such as DNA adducts, that appear closer in time to exposure; and identification of functionally relevant genetic variations that may interact with exposures. In addition, future studies can move beyond the single gene approach with the creative application of new laboratory techniques (ie, gene discovery) and innovative analytical approaches (ie, haplotyping). Identification of appropriate populations for studies of environmental pollutants is also needed, including large-enough prospective cohorts from the general population and cohorts of women with high exposures from occupational settings or industrial accidents. Additional details on recommendations for future research are found in the review articles.

The time is right—given methodological advances and recent reports of increased breast cancer risk associated with PCBs and PAHs, and in some studies of dioxin and organic solvents—to substantially increase research investments in studies of environmental pollutants and breast cancer. The toxicologic evidence strongly supports expanding the list of environmental chemicals targeted in epidemiologic studies. Research to improve toxicologic testing methods, expand animal testing to more chemicals, and better integrate findings from animal studies into chemical risk assessments are also priorities.

The lack of consistent associations observed so far in prospective studies of diet may be remedied by future research focused on improving dietary assessment methods, further development of measurement error correction models, exploring early life diet, and considering the impact of diet on estrogen- and progesterone-receptor positive and negative tumors separately. Additional data from prospective research, especially on premenopausal women, may narrow the hunt for dietary elements that may affect the risk of breast cancer.

Articles in the environment and breast cancer Epidemiology Reviews Database document the now-substantial evidence that body size, adult weight gain, and lack of physical activity are associated with breast cancer risk. Further research will help provide guidance for risk reduction efforts. One major question still to be answered is whether permanent weight loss among overweight and obese women will counter the adverse effect of these states on postmenopausal breast cancer risk. Factors such as waist-hip ratio, or measures of visceral vs subcutaneous fat mass, may influence breast cancer risk and can provide insight into the possible biological mechanisms that influence risk. A better understanding of effects of body size in premenopausal women would also be valuable. For physical activity, past research encompasses wide variations in how activity is characterized.35, 36 Some studies collect activity only at 1 or several select times in life, while others collect information across the woman's lifetime; further, not all studies collect information on duration and intensity of activity. Most studies focus on recreational activities and few collect data on occupational or household activity. Additional research that enables clinicians and public health officials to specify the types and duration of activity that provide the greatest benefit, and for whom, will help inform and motivate behavior change.

CONSIDERATIONS IN COMMUNICATING WITH PATIENTS, POLICYMAKERS, AND THE PUBLIC

  1. Top of page
  2. Abstract
  3. RESEARCH AREAS WITH PROMISE FOR RISK REDUCTION
  4. ONLINE DATABASES
  5. REVIEWS IN THIS ISSUE
  6. FUTURE RESEARCH
  7. CONSIDERATIONS IN COMMUNICATING WITH PATIENTS, POLICYMAKERS, AND THE PUBLIC
  8. Acknowledgements
  9. REFERENCES

The possible effects of individual behavior and involuntary chemical exposures on breast cancer risk are topics of intense public concern and much news media attention. Healthcare providers and researchers are often called upon to comment on these issues in the course of caring for patients, including women who have and have not previously been diagnosed with breast cancer, and in broader social contexts, such as news interviews. Although the research areas reviewed in this issue do not yet yield clear health directives, messages from experts who have informed themselves about the evidence to date can assist policymakers, journalists, members of the public, and patients by translating research in ways that support informed decision-making in the face of scientific uncertainty.

The substantial evidence that reducing postmenopausal overweight and increasing activity are associated with lower breast cancer risk is welcome news. Messages about weight management and increased activity fit within an existing public health framework; and the level of certainty about the ‘prescription’ for breast cancer benefits need not be high, given the clear evidence of other health benefits and relative lack of evidence of any harm.

In contrast, messages about the possible effects of diet and environmental chemicals on breast cancer often lead to public frustration about knowledge gaps and conflicting viewpoints. Healthcare providers and researchers can help by limiting comments to areas where their knowledge is up-to-date, acknowledging scientific uncertainty, formulating precise statements about what is and is not known, and framing a ‘strength of evidence’ decision-making rubric.

To clarify what is and is not known, broad statements often quoted in the news media that “there is no evidence that X causes breast cancer” should be replaced, particularly because “no evidence” often reflects research areas that have not been studied or where methodological problems are severe. Statements such as these are more useful:

  • 1
    The effect on breast cancer risk of using underarm products has not been investigated in a study that carefully compared women who use these products to women who do not. It is hard to study the effects of products that are widely used, because researchers cannot identify enough unexposed US women for comparison.
  • 2
    Many studies of organochlorine pesticides have found no association, but researchers were not able to directly measure exposures that occurred during the years before these pesticides were banned.

Mixed or positive results should be described by stating the magnitude and consistency of association and key limitations in the research design. Our reviews and online databases are a resource for formulating these more precise statements, and other useful resources include the National Library of Medicine's Tox Town57 and Household Products Database58 and the Vassar College Environmental Risks and Breast Cancer CD.59

Having characterized the state of the science, we then face decisions—as individuals and as a society—about further research and actions to modify exposures based on what we know now, as additional evidence unfolds. These choices may be seen as a public health extension of the decision-making framework in evidence-based medicine, which is often grounded in statistical probabilities adapted, with additional uncertainty, for individuals by considering clinical judgment and personal values. Inaction or delay are choices to weigh, and decisions are reevaluated as new evidence comes to light. Because breast cancer is so common—the American Cancer Society expects 178,480 invasive breast cancer cases,7 about 58,490 in situ diagnoses,4 and 40,460 deaths7 from breast cancer in the US in 2007—and the environmental chemical exposures hypothesized to affect risk are so widespread, the public health impact of reducing exposures would be profound even if the true relative risks are modest, similar to those observed in recent years for PCBs60–63 and PAHs.64

Acknowledgements

  1. Top of page
  2. Abstract
  3. RESEARCH AREAS WITH PROMISE FOR RISK REDUCTION
  4. ONLINE DATABASES
  5. REVIEWS IN THIS ISSUE
  6. FUTURE RESEARCH
  7. CONSIDERATIONS IN COMMUNICATING WITH PATIENTS, POLICYMAKERS, AND THE PUBLIC
  8. Acknowledgements
  9. REFERENCES

We thank Wendy Mason and Dwight E. Randle at Susan G. Komen for the Cure, for their leadership throughout the project. We thank Liesel Seryak for expert research assistance in preparation of this commentary.

REFERENCES

  1. Top of page
  2. Abstract
  3. RESEARCH AREAS WITH PROMISE FOR RISK REDUCTION
  4. ONLINE DATABASES
  5. REVIEWS IN THIS ISSUE
  6. FUTURE RESEARCH
  7. CONSIDERATIONS IN COMMUNICATING WITH PATIENTS, POLICYMAKERS, AND THE PUBLIC
  8. Acknowledgements
  9. REFERENCES