Alcohol and Breast Cancer: Title and subTitle BreakReview of Epidemiologic and Experimental Evidence and Potential Mechanisms

Although controversial, a consistently higher risk of breast cancer associated with alcohol consumption has been reported among most epidemiologic studies over the past 2 decades, with new mechanistic insights being provided by recent publications. The following is a summary of information from human and animal investigations, with particular reference to epidemiologic data published since 1995, because prior data on this association were summarized in a previous comprehensive report.¹

A MEDLINE search was conducted of the published literature through 2001 to identify epidemiologic studies in which the association of alcohol consumption with breast cancer risk was examined. Reports of human, animal, or cell culture studies were identified examining the effect of alcohol (ethanol) or acetaldehyde on physiological processes or possible risk factors associated with breast carcinogenesis. Epidemiologic studies published since 1995 were emphasized, as well as those in which relative risk measures, 95% confidence intervals, and quantitation of alcohol consumption were reported.

Epidemiologic evidence supporting a positive association between alcohol consumption and breast cancer risk was reported first in a case-control study by Williams and Horm² in 1977. In 3 of 5 subsequent case-control investigations, alcohol intake was not related to breast cancer risk.^{3 - 7} In 1984, Hiatt and Barwol⁸ reported a significant alcohol–breast cancer association from a cohort of 95 000 women, and results of 4 subsequent prospective studies^{9 - 12} supported their findings (Table 1). In addition, at least 5 meta-analyses^{39 - 43} and 2 major reviews^{1 ,44} confirmed this relationship, with evidence of a dose-response effect even at moderate intakes of 1 to 2 drinks per day. Indeed, the 1997 publication by the World Cancer Research Fund and the American Institute for Cancer Research concluded that higher breast cancer risk with alcohol intake was one of the most consistent findings among the many hypothesized relationships of dietary factors associated with breast cancer risk.¹ They noted evidence of a dose response, a biological basis for the positive findings, and a lack of confounding factors that could provide an alternative explanation for the association.

Since 1995, 33 additional reports evaluating the alcohol–breast cancer relationship were published^{13 - 38 ,43 ,45 - 50} (Table 1). Some reported no association between alcohol intake and breast cancer risk.^{15 ,18 ,29 ,31 ,34 ,36} The 1998 report by Smith-Warner et al⁵⁰ analyzed data from more than 322 000 women from cohort studies in 4 countries. They observed an independent, linear relationship of alcohol consumption with risk for intakes up to 60 g/d and estimated that relative risk increased 9% for each 10-g increment in alcohol consumed, an estimate very similar to that calculated by others.⁴¹

Taken together, findings reported since 1977 from populations in several countries provide consistent evidence that breast cancer risk is higher for women consuming moderate to high levels of alcohol (≥3 drinks/day) compared with abstainers, and that there is a significant dose-response relationship beginning with intakes as low as 1 to 2 drinks per day.^{39 ,43} The proportion of breast cancer cases among US women that can be attributed to alcohol intake has been estimated to be as low as 2%,⁴⁷ and in Italy, where the average alcohol intake is higher, the proportion is as high as 15%.²⁶ Insofar as breast cancer is among the most commonly diagnosed nonskin malignancy among women in the United States and Europe,⁵¹ the absolute number of women affected by alcohol intake could be high.

As mentioned previously, the disagreements on the causal role of alcohol intake in breast cancer development^{52 - 57} stem in part from the relatively modest strength of the observed association by epidemiologic standards, and some inconsistent findings. However, it has been pointed out that reports of no effect are not surprising considering the modest strength of the association.⁵⁸ Reasons for null findings in some studies may be explained, at least in part, by the fact that average alcohol intake in these populations was low or that alcohol intake may have been underreported, leading to exposure misclassification.⁵⁹

In general, the association between alcohol and breast cancer risk has been observed regardless of the type of alcoholic beverage consumed. Alcohol intake has been associated with both premenopausal^{13 ,17 ,19 ,21 ,60 - 61} and postmenopausal breast cancer risk.^{17 ,61 - 64} Although evidence linking consumption of alcohol during specific periods of life to breast cancer is inconsistent,^{12 ,61 ,63 ,65 - 68} it has been argued that a reduction in breast cancer risk could be realized if onset of drinking is delayed.^{69 - 70} Little information exists to determine how the pattern or intensity of alcoholic beverage consumption affects risk.

Alcohol intake may influence early as well as late stages of breast neoplasia. With regard to the former, there is a well-established relationship between the proportion of breast volume occupied by mammographic parenchymal (fibroglandular) densities and breast cancer risk.^{71 - 75} These dense patterns are associated with atypical hyperplasia and/or carcinoma in situ⁷⁶ and with cytological atypia in nipple aspirates,⁷⁷ and may be the result of enhanced mitogenesis and mutagenesis in the breast.⁷³ Breast density may be influenced by genetic factors ^{74 - 75} and environmental factors, including diet, level of obesity, and hormone use, and is considered to be an early biomarker of breast cancer risk.^{78 - 85} Five of 6 published studies suggest that alcohol consumption is associated with increased breast area occupied by mammographically dense tissue.^{86 - 91} Of interest is a recent report that mammographic density is positively associated with plasma insulin-like growth factor I (IGF-I) levels and inversely associated with plasma IGF binding protein 3 (IGFBP-3) in premenopausal women.⁹² The interrelationship among plasma IGF-I and IGFBP-3, alcohol consumption, and mammographic densities warrants evaluation. In contrast, alcohol intake was not related to risk of benign proliferative epithelial disorders of the breast.^{93 - 94} As for established breast cancer, data from 4 studies^{16 ,25 ,37 ,95} suggest that alcohol may be involved in stimulating late-stage events. A higher frequency of drinking has been reported to be positively associated with increased risk for developing regional/distant tumors, breast cancer recurrence and death, and frequency of late-stage breast cancer diagnosis and increased risk.^{16 ,25 ,95} However, the finding of the association between drinking and late-stage cancer may reflect confounding due to later diagnosis among women drinkers.

Whether dietary or other lifestyle factors, as well as genetic factors, modify the association of alcohol with breast cancer is of considerable interest. For example, alcohol consumption may influence the disposition or function of essential nutrients or other dietary factors considered cancer-protective, such as those in fruits and vegetables.^{1 ,96} Zhang et al⁹⁷ reported that women consuming more than 15 g/d of alcohol and whose intake of the water-soluble vitamin folate was less than 300 µg/d exhibited a significantly higher relative risk for breast cancer compared with women consuming the same levels of alcohol along with folate levels greater than 300 µg/d. Other reports have provided similar evidence for a modifying effect of folate status on this relationship^{98 - 100} and for interactions among alcohol, folate, and colorectal cancer.^{101 - 103} Low-to-moderate alcohol consumption, in contrast to heavy drinking, may not necessarily result in tissue folate depletion.^{104 - 107} Yet, in combination with low folate intake, ethanol and/or its primary metabolite acetaldehyde may alter folate or methionine metabolism so that an imbalance in DNA methylation or in DNA damage/repair processes could lead to DNA instability or inappropriate gene expression.^{108 - 112} Alcohol consumption also has been associated with decreased blood concentrations of β-carotene, lutein/zeaxanthin, and vitamin C.^{113 - 114} Taken together, these studies suggest that alcohol consumption, especially at high levels, may be associated with increased breast cancer risk, in part because of the negative impact of alcohol intake on the disposition or biological efficacy of dietary factors believed to be cancer-protective.

Use of hormone replacement therapy (HRT) and lower body mass index also could modify the relationship between alcohol and breast cancer.^{50 ,115 - 117} In addition, there is some evidence suggesting that this relation may differ according to the estrogen receptor (ER) and/or progesterone receptor (PR) status of the tumor; however, to date, no consistent pattern of association has been identified.^{117 - 122} Moreover, the association of alcohol with ER/PR-defined breast cancers may differ according to the use of HRT, family history of breast cancer, and body mass index.¹¹⁷ It is not clear, however, whether any effect of alcohol on breast cancers stratified according to ER/PR status represents an influence of alcohol on the progress of different diseases or on different stages in the progression of the same disease.¹²³

Although the evidence at this time is weak, there are recent reports indicating that the relationship of alcohol with breast cancer could differ according to genotype for several metabolizing enzymes. Park et al¹²⁴ reported that alcohol-consuming premenopausal women lacking the glutathione-S-transferase genes (GSTM1 and GSTT1) were at 5.3-fold greater risk for breast cancer compared with women with the genes, suggesting that the lack of these genes combined with intake of alcohol leads to a decreased capacity of tissues to detoxify reactive lipid peroxides, free radicals, and cytotoxic products of alcohol metabolism. Furthermore, results of 2 reports examining the interaction of alcohol consumption and a variant in the gene for alcohol dehydrogenase (ADH3)^{125 - 126} with breast cancer risk in premenopausal women were inconsistent.

Alcohol's impact on hormone status provides one plausible explanation for an etiologic relationship between alcohol and breast carcinogenesis. It is hypothesized that cumulative lifetime exposure to estrogen, particularly estradiol, contributes to breast cancer risk.^{127 - 129} Circulating estrogens originate from ovarian synthesis or from peripheral conversion (aromatization) pathways involving androgens such as testosterone, androstenedione, and dehydroepiandrosterone sulfate. For premenopausal women, alcohol intake has been associated with higher concentrations of estradiol, estrone, androstenedione, or testosterone, as well as decreases in follicle-stimulating hormone and 6-sulfatoxymelatonin.^{130 - 137} Compared with nondrinkers, moderate alcohol intake by premenopausal women can decrease menstrual cycle variability and frequency of long cycles,¹³⁸ thus increasing cumulative exposure to endogenous estrogens. In some cases, higher blood hormone levels observed for alcohol-consuming women were only observed in those using oral contraceptives.^{132 ,134} Furthermore, blood levels of the reactive ethanol metabolite acetaldehyde are significantly elevated during the high-estradiol phase of the menstrual cycle of women who drink and in women who drink and use synthetic estrogens.¹³⁹

For postmenopausal women not using HRT, moderate alcohol intake can lead to increased blood estrogen or androgen levels although the findings are variable, possibly due to different lengths of alcohol exposure, different frequencies of blood sampling, or different alcohol doses examined in older studies compared with more recent investigations.^{140 - 142} In postmenopausal women using HRT, there is a pronounced effect of alcohol consumption on blood estradiol levels. Ginsberg et al¹⁴³ reported a 22% increase in serum estradiol for women consuming 0.75 g of ethanol per kilogram of body weight and using a transdermal estrogen patch (0.1 mg estradiol). In a second study,¹⁴⁴ they observed that postmenopausal women taking oral estradiol daily (1 mg) and consuming 15.6 g of alcohol per week exhibited a 3.3-fold increase in serum estradiol concentrations compared with abstainers using estrogen, an effect possibly explained by altered clearance of estradiol in these women.^{145 - 146} Thus, women who drink alcohol and use oral estradiol for estrogen replacement therapy may actually have much higher circulating estradiol levels than anticipated from use of estrogen replacement alone, a potential interaction that needs to be carefully examined.^{147 - 148} Although evidence linking circulating prolactin levels and breast cancer is not consistent, there is recent interest in examining the role in breast cancer of locally produced prolactin and stimulation of the prolactin receptor in mammary tissue.¹⁴⁹ The influence of alcohol on these processes is unknown. In summary, these findings suggest that the increase in exposure to endogenous estrogens and androgens for women consuming alcohol is of sufficient magnitude to be considered an important contributor to the biological mechanism(s) underlying the alcohol–breast cancer relationship.

Data from animal studies and cell culture experiments also provide evidence for a role of ethanol in breast carcinogenesis. A detailed review of studies using animal models has been published.¹⁵⁰ To summarize, only 3 of 7 published experiments using rodent spontaneous mammary tumor models report enhancement of mammary tumor development for animals fed ethanol compared with controls^{151 - 155} (Table 2). The most striking enhancement of tumorigenesis in a spontaneous mouse mammary cancer model was reported by Watabiki et al¹⁵⁴ suggesting that this specific mouse model may be a useful tool for understanding the mechanisms underlying the mammary cancer-enhancing influence of ethanol. In contrast to spontaneous mammary tumor models, the chemically induced rodent mammary tumor model can be used to examine an effect of ethanol on specific stages (initiation, promotion) in the cancer process. In 5 experiments, ethanol fed to female rats during the initiation stage increased chemically induced mammary tumor development^{158 - 160} (Table 2). Some experiments provide evidence that ethanol may be a weak tumor promoter.^{159 - 160} Although there is evidence using this model that ethanol intake can stimulate individually the initiation and promotion stages of breast cancer, the dose-response relationship is inconsistent. No significant enhancement of chemically induced mammary tumor development has been reported when ethanol was continuously fed to animals during both the initiation and promotion stages.^{156 - 157} Differences in quantities of ethanol provided, in the specific protocols by which it was administered (ie, in the drinking water, as part of a liquid diet, or by intragastric gavage), in timing of ethanol administration, in the composition of basal diets, and in doses of carcinogen administered may partly explain some of the interlaboratory mammary tumorigenesis variability in response to ethanol.¹⁵⁰ Of additional interest is a publication of 2 experiments in which an increase in the metastasis of implanted breast cancer cells was observed for rats provided ethanol compared with controls, an effect partly explained by ethanol-associated changes in immune status.^{161 - 162} It should be noted that although quantities of ethanol fed to animals in some experiments (20% and 30% of total calories) are much higher than those levels typically consumed by humans, the blood levels attained (eg, 56 mg/dL and 86 mg/dL, respectively) are comparable with those observed in humans.¹⁵⁹

Several biological mechanisms have been proposed to explain the capacity of ethanol to enhance animal cancers at several organ sites including that of the breast.^{163 - 165} Although ethanol has not been identified as a genotoxic carcinogen, it can act as a cocarcinogen by influencing physiological processes associated with the initiation and promotion stages of carcinogenesis.¹⁶⁵ Ethanol might stimulate the initiation stage by stimulating cytochrome P450–mediated conversion of inactive carcinogens to metabolites capable of binding to and damaging cell DNA, by inhibiting phase II enzyme-mediated carcinogen detoxification, or by impairing liver clearance of carcinogens, resulting in their increased circulation among extrahepatic tissues.^{163 - 166} Ethanol and its highly reactive metabolite acetaldehyde also have been linked to the inhibition of repair of carcinogen-induced DNA damage.¹⁶⁷ Acetaldehyde is genotoxic and may bind to and interfere with the function of other important cell macromolecules.^{168 - 170} Ethanol intake may enhance the susceptibility of the rodent mammary gland to carcinogen-induced damage by altering mammary gland structural development and by stimulating cell proliferation in the target structures for mammary tumorigenesis, terminal end buds, ^{171 - 172} changes associated with decreased circulating progesterone, without an effect on estradiol.^{171 - 173} In addition, a number of animal studies do demonstrate that circulating levels of the mammotrophic hormone prolactin increase for animals consuming ethanol.^{174 - 176}

Finally, data from experiments using human breast cells provide support for an action of ethanol on early and late processes associated with breast carcinogenesis. The formation of chemical carcinogen-DNA damage is recognized as an important prerequisite to the initiation of chemically induced mammary tumorigenesis¹⁷⁷ and, if not efficiently repaired, may during cell replication lead to genetic lesions contributing to cancer development.¹⁷⁸ Nonneoplastic human mammary epithelial cells exposed to physiologically relevant concentrations of ethanol or acetaldehyde and subsequently treated with the environmental carcinogen benzo[a]pyrene (B[a]P) exhibited significantly higher levels of B[a]P DNA adducts compared with controls,¹⁷⁹ an effect, in part, due to decreased expression of GSTπ in cells exposed to ethanol. This finding is significant since carcinogen-induced DNA damage resembling that caused by B[a]P has been detected in human breast tissue.^{180 - 181} Also, GSTπ is involved in conjugation/detoxification of B[a]P and in epidemiologic studies, allelic variants have been linked to cancer susceptibility.^{182 - 183} Fan et al¹⁸⁴ detected an ethanol-induced down-regulation of the tumor suppressor BRCA1 expression and an up-regulation of ERα expression and transcriptional activity in human breast cancer cell cultures. Others have observed that ethanol selectively stimulates proliferation of ER+ but not ER− human breast cancer cells in culture, an effect also associated with an ethanol-induced increase in ERα.¹⁸⁵ Taken together, these findings suggest that even modest increases in circulating estradiol levels associated with alcohol consumption may have a disproportionately greater impact on breast cancer development due to concurrent effects of ethanol in increasing responsiveness of breast cancer cells to estradiol and in suppressing BRCA1-mediated maintenance of genome stability. More research is needed to understand how ethanol modifies the intercommunication between the ER and various cell signaling pathways controlling breast cancer cell growth.^{186 - 187} Meng et al¹⁸⁸ also observed that exposure of human breast cancer cells to ethanol led to decreased expression of BRCA1 and of proteins involved in maintaining tissue integrity (E-cadherin, α- and β-catenin), and was associated with increased cell migration on Matrigel. Luo and Miller¹⁸⁹ reported that ethanol and heregulin β1 acted synergistically to enhance the invasiveness of breast cancer cells by altering an erbB system, especially erbB2. Alcohol consumption also can increase in rat aorta matrix metalloproteinase 2 activity, an enzyme participating in metastatic spread of cancer cells.¹⁹⁰ Thus, increased breast cancer cell migration and changes in the expression of proteins associated with maintenance of tissue integrity might contribute to the capacity of ethanol to stimulate later stages of breast cancer.

In summary, a majority of epidemiologic studies provide evidence for a positive association between breast cancer risk and alcohol consumption. The relationship is modest in magnitude and cannot be explained by other known breast cancer risk factors. Animal models of breast cancer, although not entirely consistent, do provide support for an enhancing action of ethanol on mammary carcinogenesis. Evidence from human and animal studies and from cell culture experiments support plausible biological mechanisms, such as increased circulating estrogens and androgens, enhancement of mammary gland susceptibility to carcinogenesis, increased mammary carcinogen DNA damage, and greater potential for invasiveness of breast cancer cells.

A number of research issues regarding the relationship between alcohol and risk of breast cancer should be more fully explored. Specifically, the interactions of alcohol consumption with body size, HRT use, and family history of breast cancer on risk of breast cancer should be clarified. Modification of the alcohol–breast cancer relationship by specific nutrients, such as folate, and other dietary factors, such as the carotenoids, should be examined more carefully in human population studies and animal models. Furthermore, alcohol's mechanisms of action other than those involving estrogen need to be explored. For example, both IGF-I and IGF-II are strong mitogens for breast and other cancers, and IGF-I can act together with estrogen to stimulate cell proliferation and tissue development.^{191 - 193} Alcohol can influence IGF-I levels in humans, and in in vitro and animal studies has been reported to variably modify cell division and intracellular signaling pathways, with variability partly dependent on the dose of ethanol and tissue examined.^{194 - 204} Thus, it is likely that a breast cancer–stimulatory action of ethanol will be the result of its impact on multiple cellular signaling pathways.^{205 - 207}

Ethanol metabolism increases the production of reactive oxygen atoms, a purported contributor to tumor promotion.^{208 - 209} In light of evidence that alcohol consumption has been associated with oxidative stress in rodents, primates, and humans, the contribution of alcohol-associated generation of reactive oxygen radicals, lipid peroxides, acetaldehyde, and other toxicants (including genotoxic carcinogens) to all stages of breast carcinogenesis should be determined.^{210 - 214}

The potential risks vs benefits of alcohol consumption should be considered in individual decision making.²¹⁵ Clearly, alcohol abuse and excessive drinking (≥3 drinks per day) should be discouraged, in light of the numerous deleterious health consequences accompanying such behavior.²¹⁶ Also, women abstaining for personal, ethical, or religious reasons should not be encouraged to drink to reap a perceived health benefit. For those considered sporadic or occasional social drinkers (<1 drink per day), alcohol consumption is unlikely to significantly affect breast cancer risk. Regular consumption of alcohol as low as 1 to 2 drinks per day may contribute to a modest increase in risk. However, when considering low-to-moderate levels of consumption, recommendations based on a woman's overall risk profile for chronic disease seem prudent.^{20 ,217 - 218} Thus, recommendations about alcohol consumption remain valid^{1 ,219} (ie, women who do not drink should not start, and those who do drink should do so in moderation, which is generally recognized to be about 1 drink per day).

Alcohol intake is one of the few modifiable breast cancer risk factors yet identified.²²⁰ To better understand the etiologic nature of the effect of alcohol on breast carcinogenesis, more research is needed. In particular, studies should focus on interactions between alcohol and other risk factors²²¹ and on additional insights into the multiple biological mechanisms involved.