Following the publication of the Women’s Health Initiative (WHI) report,1 as well as several observational studies,2 in which postmenopausal hormone therapy use was found to increase the risk of breast cancer and coronary heart disease, the general use of hormone therapy has declined worldwide.3 Many women, however, suffer from menopausal symptoms, and hormone therapy is effective for the relief of these symptoms, in addition to osteoporosis prevention. It, therefore, becomes relevant to investigate whether other risk factors modify the effect of hormone therapy use on breast cancer risk. Such findings are important to identify subgroups of patients at particularly high risk—information that can be used to guide hormone therapy prescription practices in clinical settings.
Although the individual effects of lifestyle factors on postmenopausal breast cancer have been established in multiple studies,4,5 their combined effects—and the hormonal pathways underlying them—are not understood in depth. Obesity is expected to affect the risk through conversion of androgens into estrogens in adipose tissue, leading to increased estrogen levels.6 An independent role of testosterone has also been suggested.7 In addition, hyperinsulinemia, which is closely related to obesity, lowers the levels of sex-hormone binding globulin, leading to increased levels of estradiol and testosterone.7 Likewise, alcohol consumption has been associated with increased serum sex-hormone levels.8 Because hormone therapy use, alcohol consumption, and obesity may take effect through similar hormonal pathways,9–11 biological interactions between these factors in the pathogenesis of postmenopausal breast cancer are possible and worth investigating further.
Although several studies have addressed the combined effects of high body mass index (BMI) and hormone therapy use on breast cancer risk,2,12–17 the issue remains controversial. Previous studies have suggested that, among never-users of hormone therapy, the risk of postmenopausal breast cancer increases with increasing BMI.12–14 In contrast, others have shown that the risk among users of hormone therapy may be greater in women with low relative weight compared with that in women with high relative weight.2,12,15–17 However, in a joint analysis of BMI and hormone therapy use from the large European Investigation into Cancer and Nutrition study, a more than two-fold higher risk was observed across all BMI-strata in current users of hormone therapy compared with normal-weight never-users of hormone therapy,12 indicating that high BMI has no direct effect on breast cancer risk among hormone therapy users.
In addition, some evidence suggests that alcohol may interact with hormone therapy use, leading to a markedly higher risk among women combining the two.18–20 For instance, a prospective study found a more than 4 times higher risk of breast cancer among women with a combined high alcohol intake and hormone therapy use, which was more than expected from the combination of the individual effects.18 Another study found similar interactions, but restricted to estrogen-receptor-positive and progesterone-receptor-positive (ER+/PR+) breast cancers.19 However, most previous studies have had insufficient statistical power to address the alcohol–hormone therapy interactions.21–25 An overview of previous studies on interactions between these factors is provided in eAppendix 4 (https://links.lww.com/EDE/A882).
In most of the previous literature, interactions have been evaluated as ratios of relative effects, but from a public health perspective and especially for individual/clinical decision-making, differences in absolute risks across groups are more informative.26 Absolute risk measures can, unlike relative measures, be directly compared across subgroups, and therefore, provide a more relevant effect measure when addressing the risk associated with hormone therapy use according to other risk factor subgroups. Another feature distinguishing this study from most previous studies on this topic is the focus on joint effects. This study examines combined effects of hormone therapy use, alcohol consumption, and high BMI on hormone-receptor-defined subtypes of postmenopausal breast cancer in a large pooled cohort; the joint reference category allows us to compare each combination of BMI/hormone therapy and alcohol/hormone therapy according to the same baseline hazard. In addition, we had a unique opportunity to directly test our underlying hypothesis of a biological pathway through sex-steroid hormones by addressing the association between hormone therapy use, alcohol consumption, and BMI-strata combined and serum levels of estradiol and testosterone in a subset of the population.
METHODS
The study was based on the Social Inequality in Cancer database derived by pooling several large cohorts as described previously.27 This analysis included data from the 2 cohorts with information on current hormone therapy use: The Diet, Cancer, and Health Study28 and The Copenhagen City Heart Study.29 Information from self-administered questionnaires on health status, health behaviors, and reproductive factors at baseline was harmonized and linked to sociodemographic information from Statistics Denmark from 1980 onward. We included postmenopausal women defined as women ages 50+ years at baseline. We excluded women with a history of cancer (other than nonmelanoma skin cancer) and women born before 1921 because the central registries do not contain information on education for these birth cohorts. After exclusion of women with missing information on current hormone therapy use (n = 1,334), BMI (n = 33), alcohol consumption (n = 26), and covariates (n = 473), the total study included 30,789 women.
In both studies, current hormone therapy use (yes/no) was self-reported. Alcohol was assessed as consumption of beer, wine, and spirits (“never/almost never,” “monthly,” “weekly,” and “daily”) and average number of drinks per week within these categories. We categorized the total intake in groups of <1, 1 to 6, and 7+ drinks per week. Only 1.4% of the women were underweight (BMI < 18.5 kg/m2), and thus BMI was categorized as normal weight (<25 kg/m2), overweight (25 to 29 kg/m2), and obese (30+ kg/m2). In the Diet, Cancer, and Health Study, leisure time physical activity was assessed as average number of hours spent per week in the past year on various types of activities along with number of hours becoming sweaty or short of breath from these activities. Similarly, the Copenhagen City Heart Study assessed the weekly level of physical activity during the past year in 4 categories ranging from being almost entirely inactive to engaging in vigorous physical activity. As very few participants (8.2%) reported being highly physically active, the measures were harmonized to a 3-level variable ranging from sedentary (<2 hours of light physical activity) to active (>4 hours of light activity or >2 hours of vigorous activity per week). From both studies, we were able to classify parity into 4 groups ranging from 0 to 3+ children and smoking as never, past, current <15 g/day, and current 15+ g/day. Educational level was defined as “low” (8 to 11 years of basic schooling), “medium” (11 to 14 years; upper secondary or vocational training), and “high” (15+ years).
Endogenous sex-hormone levels were assessed in a randomly selected subsample of the Copenhagen City Heart Study at baseline (n = 1,150), as described previously.10 Blood samples were drawn at baseline (1981 to 1983) and stored at −20°C. Duplicate free testosterone and 17β-estradiol (E2) levels were measured in serum, and the means of the 2 values were applied. Commercially available control samples for free testosterone and 17β-estradiol (Con6 Immunoassay Tri-level Controls from DPC, Los Angeles, CA) were analyzed together with samples to show equivalence between different runs. Westgard control charts were used to document that the analytical methods remained in analytical and statistical control, i.e., the precision and the trueness of all the analytical methods remained stable.30
Follow-up
Incidence of invasive breast cancer was obtained from the Danish Cancer Registry (ICD7 code 174 and ICD10 code C50). Estrogen-receptor status was determined by the Danish Breast Cancer Cooperative Group by biochemistry and from 1990 onward by immunohistochemistry.31 Tumors were classified as ER+ by 10% positivity or more. Information on emigration and deaths was obtained from the registers. Participants were followed from baseline to date of breast cancer (5%), date of death (10%), emigration (1%), or end of follow-up (31 December 2008, 84%), whichever occurred first.
Statistical Methods
The 2 studies were pooled, and baseline risks were allowed to differ in all analyses according to cohort origin to account for period effects and differences in questionnaire design.32 After careful consideration of the hypothesized causal relations of exposures with breast cancer risk, we identified the following confounders: educational level, physical activity, parity, and smoking for all analyses, as well as, alcohol consumption for analyses of BMI and BMI for analyses of alcohol consumption. Effects were assessed in an Aalen additive hazards model with age as underlying time scale. The additive hazards model is a semiparametric model for survival outcomes and is at least as flexible as the Cox model. In the additive hazards model, the hazard of breast cancer for person i at age t is modeled as a linear function of the explanatory variables plus an unspecified baseline hazard:
;)
In, for example, the analysis of the combined effect of hormone therapy use and alcohol consumption, Ai and Bi are the exposure status for person i, and thus α1 and α2 capture additive effects of the 2 exposures, α3 captures the interaction between them, Xi denotes potential confounders (only included in the multiple adjusted analysis), and finally, λ0(t) and β(t) are unspecified age-dependent functions (similar to the unspecified baseline risk in a Cox analysis). Details on the implementation of this model in the software package R are given in eAppendix 1 (https://links.lww.com/EDE/A882). We tested exposures for time-dependent effects, and the effects of confounders were allowed to vary with age.33 This model provides us with an estimate of the additional number of breast cancer cases associated with a given risk factor (absolute effects) and allows us to directly compare these numbers across strata of other factors. For a given BMI, for example, the absolute effect (i.e., rate difference) of overweight (25 to 29 kg/m2) provides an estimate of additional breast cancer cases per 100,000 person-years at risk in the overweight compared with normal-weight women (adjusted for confounders).33 Results from a Cox proportional hazards model are provided in eAppendix 3 (https://links.lww.com/EDE/A882) to enable comparisons with previous studies on this topic. We tested for interaction between hormone therapy, BMI, and alcohol consumption by a Wald test.
Sensitivity analyses include (1) investigating cohort heterogeneity by including an interaction term in the model between the combined variables and cohort origin; (2) excluding the first 3 years after baseline to rule out reverse causality; (3) stratifying the analyses by birth cohort in 5-year intervals (with age as the underlying time scale) to take into account period effects of health behaviors and advanced medical diagnostics; and (4) ending follow-up in 2002 to address potential changes in hormone therapy after the release of the WHI results.
The associations between combinations of hormone therapy use, BMI, and alcohol consumption and (log-transformed) hormone levels were assessed in standard linear regression analyses adjusted for the above-mentioned confounders and time of blood draw to account for circadian variations. Accordingly, exponentiated regression coefficients correspond to percentage difference between medians.
RESULTS
In total, the pooled cohort comprised 30,789 women, of whom 1579 were diagnosed with breast cancer during 392,938 person-years of follow-up (Table 1). The median age at baseline was 56 years. The percentages of current smokers, nulliparous, and alcohol abstainers were markedly higher in the Copenhagen City Heart Study compared with the Diet, Cancer, and Health Study. However, among alcohol consumers, the median intake was similar in the 2 studies. Also, the percentage of women with a high educational level was higher in the Diet, Cancer, and Health Study. Of the total study population, 9152 (30%) women were current users of hormone therapy at baseline. Median alcohol consumption and the proportions of current smokers and of women with a high educational level were slightly higher in hormone therapy users, and the median BMI was slightly lower compared with nonusers.
TABLE 1: Baseline Characteristics of Individual and Pooled Cohorts, the Social Inequality in Cancer Study, Denmark, 1981 to 2008
The cohort-specific and pooled absolute effects of high BMI and alcohol consumption stratified by hormone therapy use are presented in Table 2. The estimates from analyses of the individual cohorts were generally very uncertain. The risk of breast cancer was higher with increasing BMI among nonusers of hormone therapy in both of the individual cohorts. However, in current hormone therapy users, a tendency toward a lower risk in overweight compared with normal weight and obese was observed in the Diet, Cancer, and Health Study, and a tendency toward a slightly higher risk in overweight and obese was observed in the Copenhagen City Heart Study. In the pooled data, overweight compared with normal weight was associated with 54 (95% confidence interval [CI] = 6, 102) additional cases in nonusers and 121 fewer cases (−216 to −26) per 100,000 person-years in current users of hormone therapy (P for interaction = 0.003). For alcohol consumption according to hormone therapy use, the results were more similar across the 2 cohorts. In the pooled analysis, a high alcohol consumption (7+ drinks/week) compared with abstinence was associated with 72 (12 to 131) additional cases in nonusers and 180 (42 to 319) in current users (P for interaction = 0.02).
TABLE 2: Absolute Risks of Postmenopausal Breast Cancer According to Baseline BMI and Alcohol Consumption Stratified by Current Hormone Therapy Status in the Social Inequality in Cancer Study, Denmark, 1981 to 2008
Figures 1 and 2 show the combined effects of hormone therapy use and categories of BMI and hormone therapy use and categories of alcohol consumption, respectively. Thus, these analyses differ from the stratified analyses above in that the effects are evaluated according to the same reference category. Among nonusers, a slightly higher risk of breast cancer was observed with increasing BMI. For instance, in this group obesity versus normal weight was associated with 59 additional cases per 100,000 person-years (95% CI = −4, 122). Also, a modestly higher risk of breast cancer was observed with higher alcohol consumption among nonusers. A markedly higher risk of breast cancer was observed for alcohol consumption combined with hormone therapy use; 432 (339 to 524) additional cases were observed in current hormone therapy users consuming 7+ drinks/week compared with abstinent nonusers of hormone therapy.
FIGURE 1: Additional breast cancer (BC) cases per 100,000 person-years according to BMI and hormone therapy use combined. Adjusted for age, cohort origin, educational level, alcohol consumption, smoking, parity, and physical activity. Ref indicates reference category.
FIGURE 2: Additional breast cancer (BC) cases per 100,000 person-years according to alcohol consumption and hormone therapy use combined. Adjusted for age, cohort origin, educational level, BMI, smoking, parity, and physical activity. Ref indicates reference category.
We did not find evidence of heterogeneity between the 2 cohorts in either of the analyses of the combined effect of hormone therapy and high BMI (P = 0.54) or hormone therapy and alcohol (P = 0.36). Excluding cases that occurred within the first 3 years after baseline and adjusting for period effects did not affect the risk estimates. The analyses with follow-up ending on 31 December 2001 showed somewhat augmented estimates compared with the main results presented in Figures 1 and 2: the combined effects of hormone therapy/high BMI was 78 (95% CI = 14, 142) and 109 (26, 192) additional cases per 100,000 person-years in overweight and obese nonusers of hormone therapy, respectively, and 367 (275 to 459), 276 (168 to 384) and 350 (150 to 549) additional cases per 100,000 person-years among normal-weight, overweight, and obese hormone therapy users compared with normal-weight nonusers. The combined effects of hormone therapy/alcohol consumption were 129 (58 to 201) and 137 (58 to 216) additional cases per 100,000 person-years in nonusers consuming 1 to 6 and 7+ drinks/week, respectively, and 304 (155 to 452), 339 (235 to 442), and 527 (401 to 653) additional cases per 100,000 person-years among hormone therapy users in the 3 alcohol categories compared with abstinent nonusers.
Estrogen-receptor-specific analyses showed that the combined effects were largely restricted to ER-positive breast cancer cases (Table 3). The combination of hormone therapy use and obesity was associated with 323 (188 to 459) additional ER+ cases, −35 (−69 to −1) ER-breast cancer cases, and 42 (−6 to 89) cases of unknown receptor status per 100,000 person-years, compared with normal-weight nonusers. In normal-weight hormone therapy users, 34 (7 to 61) additional ER-negative cases were observed compared with normal-weight nonusers. The combination of hormone therapy use and an alcohol consumption of 7+ drinks/week was associated with 360 (285 to 436) additional ER+ cases, 46 (9 to 82) ER− cases, and 27 (−6 to 60) cases of unknown ER status per 100,000 person-years, compared with abstinent, nonusers.
TABLE 3: Combined Effects of Hormone Therapy Use (Baseline) and BMI and Hormone Therapy Use and Alcohol Consumption on Postmenopausal Breast Cancer According to Estrogen-Receptor Status of the Tumor, the Social Inequality in Cancer Study, Denmark, 1981 to 2008
The results of the cross-sectional subanalyses of endogenous estradiol and testosterone levels according to combinations of hormone therapy/high BMI and hormone therapy/alcohol consumption are presented in Table 4. We observed strongly elevated estradiol levels (approximately 16 times higher) and a modestly elevated testosterone level of 32% (95% CI = 3%, 69%) in the group of women combining hormone therapy use with an alcohol consumption of 7+ drinks/week, compared with abstinent nonusers of hormone therapy. Hormone therapy use was also associated with higher levels of estradiol across BMI groups compared with normal-weight nonusers; however, the gradient was opposite, as an approximately 9 times higher estradiol-level was observed in normal-weight hormone therapy users, whereas only a 4-times higher level was observed in obese hormone therapy users. The testosterone level of obese hormone therapy users was 101% (95% CI = 36%, 297%) higher than that of normal-weight nonusers of hormone therapy.
TABLE 4: Crude Medians and Confounder-Adjusted Relative Differences in 17β-Estradiol and Testosterone Levels By Categories of Baseline BMI and Alcohol Consumption According to Current Hormone Therapy Use in a Subsample of the Copenhagen City Heart Study, Denmark
DISCUSSION
Generally, hormone therapy use increased breast cancer risks across all levels of BMI and alcohol consumption.
We found a higher number of breast cancer cases with increasing BMI among nonusers of hormone therapy, but a tendency toward a lower risk of breast cancer among overweight compared with normal weight and obese in current hormone therapy users was found. Alcohol consumption was associated with breast cancer in both users and nonusers of hormone therapy, but the combination of alcohol and hormone therapy use was associated with markedly higher risks beyond the sum of their separate effects. The effects were primarily restricted to ER-positive cases, but a modestly increased risk was also observed for ER-negative cases. The subanalysis of combined risk factors and sex hormones confirmed that hormone therapy use greatly increases estradiol and—on a smaller scale—testosterone levels. Hormone therapy use combined with alcohol was associated with very high estradiol and modestly increased testosterone levels. In contrast to our expectations, the combined effect with BMI on estradiol appeared to be higher among normal-weight hormone therapy users than among obese hormone therapy users.
Several previous observational studies report differential effects of high BMI on postmenopausal breast cancer according to hormone therapy use.2,12–17 In accordance with our study, an elevated risk is consistently observed among postmenopausal obese nonusers of hormone therapy,12–14 whereas no effect or inverse effects are found in current users.2,12,15–17 Combined effects of hormone therapy and high BMI (in which effects are presented relative to a joint reference category of unexposed to both risk factors) have rarely been investigated, although, in line with our results, one study found a more than two-fold higher risk of ER+/PR+ breast cancers among current users of hormone therapy across all BMI-strata compared with normal-weight never-users.12 For ER−/PR− tumors, an elevated risk of 1.9 was reported for current hormone therapy users in the lowest BMI tertile, whereas a risk of 1.5 was observed for hormone therapy users in the highest BMI tertile compared with never-users in the lowest BMI tertile. (The latter estimate, however, was very imprecise due to few cases.) In our study, a slightly higher risk of ER−breast cancers was observed in normal-weight hormone therapy users, but a lower risk was observed among obese hormone therapy users compared with normal-weight nonusers of hormone therapy.
The markedly elevated risk of ER+ breast cancer among women combining hormone therapy use with alcohol is consistent with results of previous studies.18–22 In a large Swedish cohort, no effect was observed for alcohol or hormone therapy use separately, but an elevated risk of 1.8 (95% CI = 1.2, 2.6) was observed for ER+/PR+ breast cancers in ever-users of hormone therapy with an alcohol intake of more than 10 g/day compared with abstainers who never used hormone therapy. The corresponding result for ER+/PR− cancers was 3.5 (2.0 to 6.2).19 Also, as mentioned previously, another study within the Copenhagen City Heart Study reported an elevated risk of 1.9 among abstainers with current hormone therapy use, no effect of alcohol consumption (>14 drinks/week) among nonusers of hormone therapy (relative risk = 1.0), but an increased risk of 4.7 among women combining the two, compared with abstinent non-hormone therapy users.18 We add to the previous findings by addressing these associations in a much larger dataset according to ER-status of the tumors and by testing the underlying biological hypothesis by relating hormonal levels to the combinations of risk factors.
A causal link between alcohol and breast cancer is supported by our study and previous findings of associations with serum hormone levels. Alcohol has been found to alter sex-steroid hormone levels.8 In a placebo-controlled study, acute ingestion of alcohol among hormone therapy users led to an average increase of 300% in estradiol levels compared with placebo, and this effect was not observed among nonusers of hormone therapy.34 This is in line with our findings of markedly higher levels of, especially, estradiol among women combining hormone therapy use and alcohol consumption.
Body fat also directly affects levels of hormones such as estrogens, testosterone, and insulin,7 and studies have shown that the effect of BMI on breast cancer is primarily mediated by serum estradiol and insulin.9 Estrogens stimulate the division of breast epithelial cells, which increases the risk of mutation, thereby inducing or promoting breast cancer.11 Androgens may also be positively associated with postmenopausal breast cancer through the conversion into estradiol and directly by inducing cell proliferation.11,35 Thus, interaction among hormone therapy use, alcohol, and high BMI is possible because they partially take effect through the same hormonal pathways. For instance, alcohol possibly affects estrogen metabolism of the body by impeding clearance of estrogen from the blood, thereby causing even higher levels of estrogen among women combining alcohol and hormone therapy use.36 This possible biological interaction emphasizes why the statistical models must leave room for interactions.
However, the findings of our study on the combination of obesity with hormone therapy and risk of breast cancer (and, in addition, the relations observed with estradiol levels) are more complex. In our data, estradiol levels were markedly higher in normal-weight hormone therapy users than in obese hormone therapy users, and the risks of breast cancer were equally high in normal-weight and obese hormone therapy users. The higher estradiol levels observed in normal-weight hormone therapy users may be due to the lower volume of distribution37 or differential metabolism of exogenous estrogens.38 In the liver, estradiol is converted to estrone and subsequently to either 2-hydroxy estrone (OHE1) or 16α-OHE1 in target cells. Higher levels of 16α-OHE1—but not 2-OHE1—have been associated with increased breast cancer risk. In lean women, 16α-OHE1 metabolite is known to dominate, whereas in overweight women, the 2-OHE1 metabolite dominates. However, in a previous study, 16α-OHE1 metabolite levels were higher with higher BMI among nonusers of hormone therapy whereas in hormone therapy users, 2-OHE1 levels were substantially elevated with higher BMI.38 Thus, these findings lend support to differential effects according to combinations of hormone therapy and BMI through separate metabolite pathways. However, information on the hormone therapy use of these women is limited, and we cannot rule out noncausal explanations. The observed differences in estradiol levels could be due to differences in dose, type (e.g., unopposed estrogen or combined estrogen/progesterone treatment), the route of administration (oral vs. transdermal), or the timing of the therapy initiation between normal-weight and obese women. The clinical guidelines of hormone therapy administration do not, however, suggest differential treatment according to BMI. The apparent equally high risk of breast cancer in normal-weight and obese women suggests that other—perhaps distinct—mechanisms are at play. Previous studies from the WHI suggest a role of insulin in the obesity–breast cancer relation,9 and the importance of testosterone is also established.35 We did not obtain data on insulin levels for this study, but the subanalysis suggested that testosterone was elevated by approximately 100% in obese hormone therapy users compared with normal-weight nonusers and unaltered in normal-weight hormone therapy users.
Our study was limited by the fact that hormone therapy use was assessed only at baseline and included only current hormone therapy status of the women. Hormone therapy use may have changed over time, especially at the time of the initial release of the WHI results in 2002.39 Although both cohorts we included administered subsequent waves of follow-up, these data were not provided for the Social Inequality in Cancer-collaboration. A study based on the Danish National Register of Medicinal Product Statistics 1995 to 2004 found that the general use of hormone therapy declined throughout the period, but was reduced dramatically after 2002. An age-standardized (according to the 1996 population) trend analysis of systemic treatment regimens (including continuous and cyclic combined treatments and unopposed estrogens) showed that the defined daily doses per 1000 women dropped from 218 to 76.5 from 1996 to 2004. Of these, continuous combined treatments dropped from 53.3 to 35.2, whereas cyclic combined treatments dropped from 99.0 to 16.5 defined daily dose per 1000 women.39 Thus, the ratio of unopposed estrogen regimens to combined treatments remained rather constant over the period—although the combined treatments gradually shifted from cyclic to continuous. A sensitivity analysis ending follow-up in 2002 showed stronger results before 2002, which suggests that changes in hormone therapy use after 2002 may have attenuated our results. Changes in other exposures and confounders were also not captured by our baseline measurements, and pooling the 2 cohorts may additionally have compromised the precision of the measured baseline variables. Alcohol, physical activity, and smoking are likely to have changed during the long follow-up period of up to 30 years. Lack of precision in confounder measurements may have led to residual confounding.40
It is also important to note that the Social Inequality in Cancer cohort consists mainly of white women. Because ethnicity is associated with our examined exposures and also breast cancer etiology, there is reason to be cautious about direct extrapolation to ethnic minority women.
Postmenopausal status was defined according to age at baseline (50+ years). A recent study that compared various definitions reported modest overlap between comprehensive measures (e.g., time since last period combined with hysterectomy and age) and crude age measures, revealing some degree of discrepancy.41 However, the breast cancer incidence did not vary considerably across the definitions. Also, another study comparing various measures concluded that, when lacking information on menstrual history, an age-definition of 50 years would be the best proxy for menopausal status.42
Our study was strengthened by the large sample size, the prospective design, and the linkage to population-based registers on disease, death, and emigration. The analyses of joint effects of risk factors distinguishes this study from most previous studies, and the hormone measurements on the subsample of the cohort allowed for a unique opportunity to substantiate the hypothesized mechanisms underlying the relation between alcohol consumption, high BMI, and breast cancer among users and nonusers of hormone therapy.
In conclusion, these analyses suggest an increased risk of breast cancer associated with hormone therapy—a risk that may be particularly strong among women consuming alcohol. Also, hormone therapy use increased the absolute risk of breast cancer markedly in both normal-weight and obese women. Further studies on mechanisms underlying the combined effect of hormone therapy use with obesity are warranted.
ACKNOWLEDGMENTS
We thank the collaborators behind the Social Inequality in Cancer Database: The Copenhagen City Heart Study, The Diet, Cancer and Health Study, and The Research Centre for Prevention and Health in Glostrup.
REFERENCES
1. Rossouw JE, Anderson GL, Prentice RL, et al.Writing Group for the Women’s Health Initiative Investigators. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results From the Women’s Health Initiative randomized controlled trial. JAMA. 2002;288:321–333
2. Beral V, Bull D, Doll R, et al. Breast cancer and hormone replacement therapy: collaborative reanalysis of data from 51 epidemiological studies of 52,705 women with breast cancer and 108,411 women without breast cancer. Lancet. 1997;350:1047–1059
3. Barbaglia G, Macià F, Comas M, et al. Trends in hormone therapy use before and after publication of the Women’s Health Initiative trial: 10 years of follow-up. Menopause. 2009;16:1061–1064
4. Suzuki R, Orsini N, Mignone L, Saji S, Wolk A. Alcohol intake and risk of breast cancer defined by estrogen and progesterone receptor status–a meta-analysis of epidemiological studies. Int J Cancer. 2008;122:1832–1841
5. Suzuki R, Orsini N, Saji S, Key TJ, Wolk A. Body weight and incidence of breast cancer defined by estrogen and progesterone receptor status–a meta-analysis. Int J Cancer. 2009;124:698–712
6. Liedtke S, Schmidt ME, Vrieling A, et al. Postmenopausal sex hormones in relation to body fat distribution. Obesity (Silver Spring). 2012;20:1088–1095
7. Neilson HK, Friedenreich CM, Brockton NT, Millikan RC. Physical activity and postmenopausal breast cancer: proposed biologic mechanisms and areas for future research. Cancer Epidemiol Biomarkers Prev. 2009;18:11–27
8. Dorgan JF, Baer DJ, Albert PS, et al. Serum hormones and the alcohol-breast cancer association in postmenopausal women. J Natl Cancer Inst. 2001;93:710–715
9. Hvidtfeldt UA, Gunter MJ, Lange T, et al. Quantifying mediating effects of endogenous estrogen and insulin in the relation between obesity, alcohol consumption, and breast cancer. Cancer Epidemiol Biomarkers Prev. 2012;21:1203–1212
10. Rod NH, Hansen AM, Nielsen J, Schnohr P, Grønbaek M. Low-risk factor profile, estrogen levels, and breast cancer risk among postmenopausal women. Int J Cancer. 2009;124:1935–1940
11. Key T, Appleby P, Barnes I, Reeves GEndogenous Hormones and Breast Cancer Collaborative Group. . Endogenous sex hormones and breast cancer in postmenopausal women: reanalysis of nine prospective studies. J Natl Cancer Inst. 2002;94:606–616
12. Ritte R, Lukanova A, Berrino F, et al. Adiposity, hormone replacement therapy use and breast cancer risk by age and hormone receptor status: a large prospective cohort study. Breast Cancer Res. 2012;14:R76
13. Morimoto LM, White E, Chen Z, et al. Obesity, body size, and risk of postmenopausal breast cancer: the Women’s Health Initiative (United States). Cancer Causes Control. 2002;13:741–751
14. Feigelson HS, Jonas CR, Teras LR, Thun MJ, Calle EE. Weight gain, body mass index, hormone replacement therapy, and postmenopausal breast cancer in a large prospective study. Cancer Epidemiol Biomarkers Prev. 2004;13:220–224
15. Magnusson C, Baron JA, Correia N, Bergström R, Adami HO, Persson I. Breast-cancer risk following long-term oestrogen- and oestrogen-progestin-replacement therapy. Int J Cancer. 1999;81:339–344
16. Schairer C, Lubin J, Troisi R, Sturgeon S, Brinton L, Hoover R. Menopausal estrogen and estrogen-progestin replacement therapy and breast cancer risk. JAMA. 2000;283:485–491
17. Lahmann PH, Hoffmann K, Allen N, et al. Body size and breast cancer risk: findings from the European Prospective Investigation into Cancer and Nutrition (EPIC). Int J Cancer. 2004;111:762–771
18. Nielsen NR, Grønbaek M. Interactions between intakes of alcohol and postmenopausal hormones on risk of breast cancer. Int J Cancer. 2008;122:1109–1113
19. Suzuki R, Ye W, Rylander-Rudqvist T, Saji S, Colditz GA, Wolk A. Alcohol and postmenopausal breast cancer risk defined by estrogen and progesterone receptor status: a prospective cohort study. J Natl Cancer Inst. 2005;97:1601–1608
20. Gapstur SM, Potter JD, Drinkard C, Folsom AR. Synergistic effect between alcohol and estrogen replacement therapy on risk of breast cancer differs by estrogen/progesterone receptor status in the Iowa Women’s Health Study. Cancer Epidemiol Biomarkers Prev. 1995;4:313–318
21. Lew JQ, Freedman ND, Leitzmann MF, et al. Alcohol and risk of breast cancer by histologic type and hormone receptor status in postmenopausal women: the NIH-AARP Diet and Health Study. Am J Epidemiol. 2009;170:308–317
22. Gertig DM, Fletcher AS, English DR, Macinnis RJ, Hopper JL, Giles GG. Hormone therapy and breast cancer: what factors modify the association? Menopause. 2006;13:178–184
23. Horn-Ross PL, Canchola AJ, Bernstein L, et al. Alcohol consumption and breast cancer risk among postmenopausal women following the cessation of hormone therapy use: the California Teachers Study. Cancer Epidemiol Biomarkers Prev. 2012;21:2006–2013
24. Chen WY, Colditz GA, Rosner B, et al. Use of postmenopausal hormones, alcohol, and risk for invasive breast cancer. Ann Intern Med. 2002;137:798–804
25. Zhang SM, Lee IM, Manson JE, Cook NR, Willett WC, Buring JE. Alcohol consumption and breast cancer risk in the Women’s Health Study. Am J Epidemiol. 2007;165:667–676
26. Rod NH, Lange T, Andersen I, Marott JL, Diderichsen F. Additive interaction in survival analysis: use of the additive hazards model. Epidemiology. 2012;23:733–737
27. Nordahl H, Hvidtfeldt UA, Diderichsen F, et al. Cohort Profile: the Social Inequality in Cancer (SIC) cohort study Int J Epidemiol. 2014;43:1750–1758
28. Tjønneland A, Olsen A, Boll K, et al. Study design, exposure variables, and socioeconomic determinants of participation in diet, cancer and health: a population-based prospective cohort study of 57,053 men and women in Denmark. Scand J Public Health. 2007;35:432–441
29. Appleyard M, Hansen AT, Schnohr P, Jensen G, Nyboe J. The Copenhagen City Heart Study. A book of tables with data from the first examination (1976–78) and a 5-year follow-up (1981–83) Scand J Soc Med. 1989;170:1–160
30. Westgard JO, Barry PL, Hunt MR, Groth T. A multi-rule Shewhart chart for quality control in clinical chemistry. Clin Chem. 1981;27:493–501
31. Møller S, Jensen MB, Ejlertsen B, et al.Danish Breast Cancer Cooperative Group. The clinical database and the treatment guidelines of the Danish Breast Cancer Cooperative Group (DBCG); its 30-years experience and future promise. Acta Oncol. 2008;47:506–524
32. Smith-Warner SA, Spiegelman D, Ritz J, et al. Methods for pooling results of epidemiologic studies: the pooling project of prospective studies of diet and cancer. Am J Epidemiol. 2006;163:1053–1064
33. Martinussen T, Scheike THMartinussen T, Scheike TH. Additive hazards models. In: Dynamic Regression Models for Survival Data. 20061st ed New York Springer-Verlag:103–173
34. Ginsburg ES, Mello NK, Mendelson JH, et al. Effects of alcohol ingestion on estrogens in postmenopausal women. JAMA. 1996;276:1747–1751
35. Zhang X, Tworoger SS, Eliassen AH, Hankinson SE. Postmenopausal plasma sex hormone levels and breast cancer risk over 20 years of follow-up. Breast Cancer Res Treat. 2013;137:883–892
36. Ginsburg ES. Estrogen, alcohol and breast cancer risk. J Steroid Biochem Mol Biol. 1999;69:299–306
37. Hanley MJ, Abernethy DR, Greenblatt DJ. Effect of obesity on the pharmacokinetics of drugs in humans. Clin Pharmacokinet. 2010;49:71–87
38. Modugno F, Kip KE, Cochrane B, et al. Obesity, hormone therapy, estrogen metabolism and risk of postmenopausal breast cancer. Int J Cancer. 2006;118:1292–1301
39. Løkkegaard E, Lidegaard O, Møller LN, Agger C, Andreasen AH, Jørgensen T. Hormone replacement therapy in Denmark, 1995–2004. Acta Obstet Gynecol Scand. 2007;86:1342–1351
40. Fewell Z, Davey Smith G, Sterne JA. The impact of residual and unmeasured confounding in epidemiologic studies: a simulation study. Am J Epidemiol. 2007;166:646–655
41. Phipps AI, Ichikawa L, Bowles EJ, et al. Defining menopausal status in epidemiologic studies: a comparison of multiple approaches and their effects on breast cancer rates. Maturitas. 2010;67:60–66
42. Morabia A, Flandre P. Misclassification bias related to definition of menopausal status in case-control studies of breast cancer. Int J Epidemiol. 1992;21:222–228
