|Home | About | Journals | Submit | Contact Us | Français|
Most risk factors for male breast cancer have been derived from retrospective studies that may reflect selective recall. In the prospective National Institutes of Health–AARP Diet and Health Study, we studied 324920 men, among whom 121 developed breast cancer. Men who reported a first-degree relative with breast cancer had an increased risk of breast cancer (relative risk [RR] = 1.92, 95% confidence interval [CI] = 1.19 to 3.09). Among the medical conditions examined, a new finding emerged regarding increased male breast cancer risk associated with a history of a bone fracture (RR = 2.20, 95% CI = 1.24 to 3.91). Obesity was positively related to risk (RR = 1.79, 95% CI = 1.10 to 2.91, for body mass indices of ≥30 vs <25 kg/m2) and physical activity inversely related, even after adjustment for body mass index. Smokers were at somewhat elevated risk, although trends with smoking characteristics were inconsistent. Alcohol consumption was not related to risk. The identified risk factors show some commonalities with female breast cancer and indicate the importance of hormonal mechanisms. Differences in risk factors may reflect unique mechanisms associated with androgens and their ratio to bioavailable estrogens.
Many risk factors for male breast cancer have been identified in studies that may suffer from the selective recall of participants.
Prospective cohort study of more than 300000 men in the National Institutes of Health–AARP Diet and Health Study, who submitted a completed questionnaire and of whom 121 developed breast cancer.
Having a first-degree relative with breast cancer, having a bone fracture after the age of 45 years, and being obese were associated with an increased risk of male breast cancer. Physical activity, after adjustment for body mass index, was inversely related to risk. Alcohol consumption was not related to risk.
Commonalities between risk factors for male and female breast cancer indicate that hormonal mechanisms may be important, whereas differences may reflect unique mechanisms that may be associated with androgens. Because of the small number of patients with male breast cancer in this and other studies, however, risk factors for male breast cancer should be investigated further in pooled analyses across many studies.
This study had relatively few participants who developed male breast cancer, which limited its power to detect rare exposures. The generalizability of the results is unclear because the findings were based on self-reported information and the response rate to the questionnaire was low.
From the Editors
Male breast cancer is uncommon, accounting for only 0.7% of all breast cancers (1). It is dissimilar to female breast cancer, in that incidence rates are higher among African American men than white men and continue rising in men aged 55 years or older, resulting in a late average age at onset (2,3). Although numerous investigations emphasize familial or genetic associations, less is known regarding environmental effects. Reports of increased risks of male breast cancer among atomic bomb survivors (4) and individuals who have received medical irradiation (5–7) or had occupational exposure to electromagnetic fields (8–10) support a role for ionizing radiation. Hormonal factors have also been implicated, through relationships with obesity (6,11–14), physical inactivity (12,13), alcohol consumption (7,15), and use of exogenous androgens (16,17) or estrogens (18,19).
Given that many of the identified risk factors may reflect selective recall after disease onset, we prospectively evaluated risk factors within the large National Institutes of Health–AARP (formerly known as the American Association of Retired Persons) Diet and Health Study. This cohort was established in 1995 and 1996, when a detailed questionnaire was sent to 3.5 million members aged 50–71 years (20), with 567169 (16.2%) satisfactorily completing it. After various exclusions (deaths or having moved, n = 590; proxy respondents, n = 15760; withdrawals, n = 6; duplicate questionnaires, n = 179; women, n = 225468; and previous breast cancers, n = 246), 324920 study subjects remained. Subjects were followed annually for changes of address. Incident cancers were identified by linkage to 11 state cancer registries, and deaths were identified through the National Death Index. From baseline through December 31, 2003, 2313988 person-years were contributed and 121 men (median age = 68 years) developed breast cancer (nine with in situ disease, 107 with invasive breast cancer, and five with missing stage). The Special Studies Institutional Review Board of the National Cancer Institute approved this study, and written informed consent was obtained from study participants.
Using multivariable Cox proportional hazards regression to estimate relative risks (RRs) and 95% confidence intervals (CIs), we found an increased risk associated with the reporting of a first-degree relative with breast cancer (RR = 1.92, 95% CI = 1.19 to 3.09) (Table 1). Risk was particularly elevated for individuals with an affected sister (RR = 2.25, 95% CI = 1.13 to 4.47), possibly reflecting the influence of shared genetic and environmental risk factors. A particularly enhanced risk was found for those with both an affected mother and an affected sister (RR = 9.73, 95% CI = 3.96 to 23.96, n = 5).
Our results agree with others who have identified familial relationships for male breast cancers (6,7,11,13,21–23), including one report showing an increased risk with multiple affected relatives (23). These associations mirror patterns observed for female breast cancer; however, the role of specific genetic mutations (particularly in BRCA2) appears to differ for the two diseases (24,25).
In both females and males, genetic mutations appear to explain only a small proportion of disease occurrence, leading to interest in other contributory factors, including various medical conditions. Of note is a very strong relationship for male breast cancer (50-fold increase) that has been observed with Klinefelter syndrome (26,27), a condition associated with increased gonadotropin and decreased androgen levels, normal estrogen levels, and therefore a high ratio of estrogen to androgen (14,28,29). Studies have indicated relationships with other chronic conditions, including liver cirrhosis (22,30), hyperthyroidism (22), gallstones (22), and diabetes (11,14), albeit usually on the basis of small numbers. We did not have information on liver or thyroid diseases, but we observed no increased risk related to diabetes, gallstones, heart diseases, or colorectal polyps.
Somewhat surprisingly, we observed a statistically significant increased risk associated with bone fractures occurring after age 45 years (RR = 2.20, 95% CI = 1.24 to 3.91). This new finding was unexpected because breast cancers are less likely to occur among women with fractures (31) and such fractures have generally been attributed to low estrogen levels. Although estrogens are important for bone maintenance in males (32,33), gonadal insufficiency and low testosterone levels also contribute to bone density and osteoporosis (34). Given decreasing testosterone levels with age, bone fractures may relate to male breast cancer through alterations in the bioavailable ratio of estrogen to testosterone (33,35) and would be consistent with the link between breast cancer and Klinefelter syndrome, a condition also associated with low bone density (36).
In terms of lifestyle factors, we found that body mass index was statistically significantly associated with male breast cancer (for ≥30 vs <25 kg/m2, RR = 1.79, 95% CI = 1.10 to 2.91) (Table 2), in agreement with case–control studies (11–13,37). Male obesity is often associated with gynecomastia, which has previously been linked with male breast cancer (29,38). However, we did not collect information on gynecomastia. Obesity is associated with an increased risk of postmenopausal female breast cancer (39), presumably through peripheral conversion of androgens to estrogens (40). In men, obesity is associated with decreased testosterone (41–43) and sex hormone–binding globulin (41,43) levels but increased estrogen levels (33,44,45), leading to greater estrogen bioavailability.
As with female breast cancers (46), we found a relationship of risk with physical inactivity that persisted after adjustment for body mass index. Current physical activity was unrelated to the risk of male breast cancer, but physical activity during adolescence was inversely associated with risk (for activity ≥5 times per week, RR = 0.59, 95% CI = 0.31 to 1.13). This finding may reflect the importance of sustained physical activity, although imprecise exposure assessment or small numbers must also be considered. Subjects who had a physically active routine were at a statistically significant low risk of male breast cancer (RR = 0.49, 95% CI = 0.28 to 0.87).
Several (7,15,47), although not all (6,12,13,23,48,49), studies have suggested that excessive consumption of alcohol may increase male breast cancer risk. We found no evidence for such a relationship, even when we considered specific types of beverages. However, few individuals reported drinking three drinks per day or more, preventing assessment of previous reports of excess risks among alcoholics (7,47).
Cigarette smokers in our study were at somewhat elevated risk, but there were no convincing trends with intensity or duration of smoking. A statistically significant elevated risk of male breast cancer among current smokers of more than 20 cigarettes per day could have arisen by chance. The absence of a strong effect of smoking is consistent with previous investigations of male (6,48,49), as well as female (50), breast cancer.
Despite its strengths as a prospective investigation, our study had several limitations. Like most previous case–control investigations, our study had relatively few events, limiting our power to detect effects. For more common exposures, such as obesity and physical activity, we had adequate (80%) power to detect relative risks, but for the less common exposures we had much more limited power (roughly 60%), underscoring the need to examine risk factors across multiple studies in pooled analyses. Further, the extent to which our results are generalizable to all male breast cancers is uncertain because the findings were based on self-reported exposures from questionnaires with limited response rates. We were also unable to evaluate all of the proposed hypotheses for male breast cancer, including having late puberty (14); being single, infertile, or childless (6,12,14,22,37); having undescended testes; suffering testicular trauma or infections causing orchitis (14,48); and being employed in occupations involving exposure to high temperatures (22,48,51–54) or polycyclic aromatic hydrocarbons (22,55). However, presumably most of these factors would have a limited impact on risk.
Our findings of increased risk associated with obesity and physical inactivity raise concerns about future male breast cancer trends, particularly given recent reports of increasing incidence (56,57). It is unclear, however, whether these reports foreshadow future changes or reflect earlier detection (58). A need for further surveillance, as well as a focus on the role of endogenous hormones that may be associated with these factors, appears warranted. Further, an increased risk of male breast cancer among men with bone fractures indicates that research on androgens and their ratio to bioavailable estrogens may be useful in elucidating mechanisms involved in male breast carcinogenesis.
This research was supported in part by the Intramural Research Program of the National Institutes of Health, National Cancer Institute. The authors had responsibility for the design of the study, collection of data, analysis and interpretation of data, decision to submit the manuscript for publication, and writing of the manuscript.