Abstract
Background
Divergent risk factors exist for premenopausal and postmenopausal breast cancers, but it is unclear whether differences by age exist among postmenopausal women.
Methods
We examined relationships among 190,872 postmenopausal women, ages 50–71 years recruited during 1995–1996 for the NIH-AARP Diet and Health Study, in whom 7,384 incident invasive breast carcinomas were identified through 2006. Multivariable Cox regression hazard ratios (HRs) and 95 % confidence intervals (CIs) were estimated for breast cancer risk factors by age (50–59, 60–69, ≥70 years).
Results
The only factor showing significant statistical heterogeneity by age (p het = 0.001) was menopausal hormone therapy duration, but trends were apparent across all ages and the strongest association prevailed among women 60–69 years. Although other risk factors did not show statistically significant heterogeneity by age, we did observe attenuated relations for parity and late age at first birth among older women [e.g., HR for age at first birth ≥30 vs. 20–24 = 1.62 (95 % CI 1.23–2.14) for women 50–59 years vs. 1.12 (0.96–1.31) for ≥70 years]. In contrast, risk estimates associated with alcohol consumption and BMI tended to be slightly stronger among the oldest subjects [e.g., HR for BMI ≥35 vs. 18.5–24.9 = 1.24 (95 % CI 0.97–1.58) for 50–59 years vs. 1.46 (1.26–1.70) for ≥70 years]. These differences were somewhat more pronounced for estrogen receptor positive and ductal cancers, tumors predominating among older women. Breast cancer family history, physical activity, and previous breast biopsies did not show divergent associations by age.
Conclusion
Although breast cancer risk factor differences among older women were not large, they may merit further consideration with respect to individualized risk prediction.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
Introduction
It is well documented that breast cancer is strongly influenced by age, with incidence rates that rise rapidly prior to menopause, until they plateau and subsequently show less rapid rate increases [1]. Although it has been demonstrated that breast cancer risk factors differ between premenopausal and postmenopausal women, particularly with respect to the effects of family history of breast cancer, parity and obesity [2], less is known regarding whether there are additional variations in relationships according to age for women experiencing the highest incidence rates—namely postmenopausal women. Age-related differences in risk factor relationships could influence risk prediction models, but few studies have specifically addressed effects with an emphasis on older women,
The heterogeneity of breast cancer by age is supported by different clinical characteristics, most notably by rising rates of hormone receptor positivity with increasing age [3]. Given that many risk factors differ by hormone receptor status [4], it would be expected that factors should show different patterns according to advanced ages at diagnosis. A few studies have suggested that the risk of breast cancer among women at older ages may be less influenced by reproductive behavior [5–7] and alcohol consumption [5], and more influenced by menopausal factors [5] and body mass [8], but it is unclear to what extent these differences reflect effects of divergent clinical characteristics. This includes not only hormone receptor status, but also histology and stage, factors that may be influenced by variations in screening practices for different ages.
Within the NIH-AARP Diet and Health Study, we had a unique opportunity to examine whether breast cancer risk factor associations changed with advancing ages among a large series of postmenopausal women.
Methods
Study population
The NIH-AARP Diet and Health Study was established in 1995–1996 when a questionnaire requesting information on demographic characteristics, dietary intake, and health-related behaviors was sent to 3.5 million AARP members. The study design has been described in detail elsewhere [9]. Those initially contacted were AARP members 50–71 years old residing in six US states (California, Florida, Louisiana, New Jersey, North Carolina, and Pennsylvania) and two metropolitan areas (Atlanta, Georgia, and Detroit, Michigan). A total of 617,119 people (17.6 %) returned the questionnaire. Excluded were individuals who skipped substantial portions of the questionnaire, made significant recording errors or indicated eating <10 foods, did not indicate sex, or died or moved before their questionnaire was scanned. After these exclusions, the baseline study population included 241,227 female participants. Address changes were tracked annually through the National Change of Address Service, the processing of undeliverable mail and directly from participants. Vital status was tracked using the Social Security Administration Death Master File, the National Death Index, cancer registry linkages, and mailing responses. The NIH-AARP Diet and Health Study was approved by the Special Studies Institutional Review Board of the US National Cancer Institute.
Cancer incidence
Incident cases of breast cancer were identified by probabilistic linkage to cancer registries in the eight states of the cohort as well as Arizona, Nevada, and Texas in order to capture cancers occurring among participants who moved to those states during the follow-up period. These linkages were based on annually updated residence information and used first and last name, address, sex, date of birth, and social security number from the baseline questionnaire. The states of the cohort were chosen in part because their registries had been shown to validly identify at least 90–95 % of cancer cases [10]. All suspected matches underwent review. Dates of diagnosis and tumor characteristics were obtained from the registries. Histology was available from all eleven state registries and was defined using the International Classification of Diseases for Oncology Codes, 3rd edition [11]. Any primary incident invasive carcinoma of the breast was considered for this analysis, which included ductal (8,500, 8,523), lobular (8,520, 8,524), mixed (8,522), and other cases. Estrogen receptor (ER) status was available from eight of the state registries (not Florida, Pennsylvania, or Texas) and was classified as ER positive (+), ER negative (−), borderline or unknown.
Exposure assessment
Study participants were asked to provide information on demographic characteristics (e.g., race/ethnicity, level of education, and marital status), dietary characteristics (e.g., alcohol and caloric intake), and medical history (e.g., family history of cancer and previous biopsies). In addition, participants reported their current height and weight which was used to determine a participant’s current body mass index (BMI, kg/m2). A period of vigorous physical activity was defined as one of at least 20 min in the past 12 months that caused an increase in breathing or heart rate, or worked up a sweat. Women were also asked for the age of their first menstrual period, of the birth of their first child, and of their last menstrual period, as well as number of live births. In addition, they were asked whether they had even taken menopausal hormone therapy (MHT) and the duration of their use (years).
Analytic population
For the present analysis, we excluded 15,760 participants whose questionnaires were completed by proxies, 23,957 who self-reported or were diagnosed with any cancer other than non-melanoma skin cancer before baseline, 18 who died or were diagnosed with cancer on the first day of follow-up, 1,300 who had extreme values for BMI (defined as more than two inter-quartile ranges above the 75th ‰ or below the 25th ‰ of log-transformed BMI), 7,210 who reported premenopausal status, and 2,110 with unknown menopausal status. The final analytic population included 190,872 women, of whom 7,384 were diagnosed with invasive breast carcinoma during follow-up.
Statistical analysis
We used Cox proportional hazards regression (SAS 9.2.3 software, SAS Institute Inc., Cary, NC, USA), with age as the timescale to estimate the hazard ratios (HRs) and 95 % confidence intervals (CIs) of developing breast cancer in each age interval. Follow-up for a woman started at age at cohort entry and ended at an invasive breast cancer diagnosis or censoring. Women were censored at the earliest of the following: the end of follow-up (31 December 2006), an in situ (or unknown behavior) breast cancer diagnosis, or the date on which the participant moved out of the registry area or died from any cause. Women contributed person time to one or more age intervals (age 50–59, 69–69, ≥70 years) based on their ages at entry into and exit from the cohort.
Multivariable models were adjusted for established risk factors as determined by previous studies of postmenopausal breast cancer. Covariates included race (white, nonwhite, unknown), education level (high school or less, post high school or some college, college graduate, postgraduate, unknown), marital status (married, unmarried, unknown), age at menarche (≤10, 11–12, 13–14, ≥15 years, unknown), parity and age at first live birth (nulliparous, <20, 20–24, 25–29, ≥30 years, unknown), age and type of menopause (natural menopause at age: <45, 45–49, 50–54, ≥55 years, surgical/medical, unknown), use of MHT (never, former, current, unknown), BMI (<18.5, 18.5–24.9, 25–29.9, 30–34.9, ≥35.0 kg/m2, unknown), alcohol intake (nondrinker, ≤5.0, 5.01–10.0, 10.01–20.0, 10.01–35.0, >35.0 g/day), history of breast cancer in a first-degree relative (no, yes, unknown), and number of previous breast biopsies (0, 1, ≥2, unknown). Statistical heterogeneity in the hazard ratios across the three age intervals was assessed using the Wald’s test (p het).
In addition, we examined whether the relationship between risk factors and breast cancer incidence differed by age and by ER status (ER+, ER−) or histologic type (ductal, lobular). For example, to obtain ER+ specific HR estimates, we fit Cox regression models where ER+ cases were defined as events and ER− cases were censored at the date of their diagnosis. In analyses of ER status, only those participants who reported living in a state that reported ER status at baseline were eligible. To examine the relationships between BMI, menopausal hormone use, and breast cancer risk, we assessed duration of MHT use according to three categories of BMI: normal/underweight (<25 kg/m2), overweight (25–29.9 kg/m2), and obese (≥30 kg/m2). We also evaluated the converse, namely whether BMI relations differed according to menopausal hormone status.
To address potential biases, we ran sensitivity analyses, separately excluding women with extreme values for total caloric intake (n = 1,679), those who reported being in poor or fair health at baseline (n = 24,532), and those who reported no or infrequent recent breast cancer screening (n = 38,041) (among a subset of postmenopausal subjects who completed a second risk factor questionnaire). We also truncated follow-up time at 30 June 2002 (the date of the end of an influential study on menopausal hormone use and breast cancer) [12], changing the status of current hormone users to former at 30 June 2002. Finally, we assessed the impact of adjusting for year of birth in 5-year categories.
The proportional hazards assumption was evaluated by visual examination of Kaplan–Meir plots and by assessing time-dependent interactions between each exposure and age; no deviations of the assumption were observed (all p > 0.05). Tests for linear trend across categories of covariates were estimated from the Wald’s test using an ordinal variable. p values of <0.05 were considered statistically significant. All tests of significance were two-tailed. Analyses were conducted using SAS 9.2.3 software, SAS Institute Inc., Cary, NC, USA.
Results
During an average 9.3 years (SD 2.7) of follow-up, the 190,872 women contributed the following person-years (pyrs) to their corresponding age intervals: 50–59 years (1,778,664 pyrs), 60–69 years (1,761,243 pyrs), and ≥70 years (1,720,507 pyrs). Among the 7,384 breast cancer cases, 809 (11.0 %) were diagnosed between 50 and 59 years of age, 3,864 (52.3 %) between 60 and 69 years, and 2,711 (36.7 %) at ages 70 years or beyond.
Table 1 shows the relationship of various reproductive factors with breast cancer risk for all study subjects and stratified by the three different age groups. Although there was no statistically significant heterogeneity across the three age groups for any of the parameters of interest–which included age at menarche, number of births, age at first birth, and age and type of menopause—there were some suggestive associations observed.
A trend of decreasing risk with increasing age at menarche that was observed among all cases was largely driven by a relationship among the women 70 years of age or older (RR for ≥15 vs. <10 = 0.81, 95 % CI 0.66–0.99), with little evidence that this factor related to risk among the younger women. In contrast, the number of births was inversely related to risk within all three age groups, with some suggestion that it might be somewhat more strongly related to risk among the youngest as opposed to oldest women [HR for ≥3 vs. 0 births = 0.71 (95 % CI 0.58–0.87) for 50–59 years vs. 0.85 (0.76–0.96) for ≥70 years]. Similarly, a significant relationship with late age at first birth was observed within all age groups but was stronger among the youngest women [HR for ≥30 vs. 20–24 years = 1.62 (95 % CI 1.23–2.14) for 50–59 years vs. 1.12 (0.96–1.31) for ≥70 years].
Later age at natural menopause (>55 vs. 50–54 years) was significantly related to increased risk only among the women 60 years of age or older [HR 1.08 (95 % CI 0.71–1.66) for women 50–59 years, 1.22 (1.08–1.38) for 60–69 years, and 1.17 (1.02–1.35) for ≥70 years]. However, there was little difference in the relation of a surgical/medical menopause across the three age groups.
Risk was further examined according to a variety of lifestyle factors, including BMI, frequency of physical activity, alcohol consumption, and duration of MHT use (Table 2). The only factor showing significant heterogeneity by age was duration of MHT use (p het = 0.001). Although there were significant linear trends with years of use in all age groups, the strongest association was observed for women in the 60–69 year age group (HR for ≥10 years MHT = 1.57, 95 % CI 1.43–1.73). In terms of other lifestyle factors, BMI was not significantly related to risk among the youngest age group, but significantly related to risk among the subjects 60 years of age or older [HR for BMI ≥35 vs. 18.5–24.9 kg/m2 = 1.24 (95 % CI 0.97–1.58) for 50–59 years; 1.39 (1.24–1.57) for 60–69 years; 1.46 (1.26–1.70) for ≥70 years). In contrast, higher frequency of physical activity appeared to be significantly inversely related to risk only among women 60–69 years of age, with little evidence of a relation among the youngest or oldest women. Alcohol consumption was significantly related to risk only among the women 60 years of age or older [HR for >35 g/day vs. nondrinking = 1.10 (95 % CI 0.73–1.66) for 50–59 years, 1.42 (1.21–1.68) for 60–69 years, and 1.54 (1.27–1.87) for ≥70 years].
We also assessed risk in relation to a family history of breast cancer and a history of previous breast biopsies (Table 3). Subjects with a family history of breast cancer were at approximately at 50 % increased risk, which did not vary according to age. Similarly, a history of a previous breast biopsy, which conferred the highest risk if multiple biopsies were reported, remained significantly associated with risk across all three age intervals.
Since effects of BMI have been shown to be strongest in nonhormone users, we also conducted sensitivity analyses restricted to subjects who had never used menopausal hormones. These continued to show somewhat enhanced relationships of BMI among the women diagnosed at the older ages (data not shown).
As risk factor relationships can also vary by hormone receptor status and/or histology of the tumor, we examined risks within age groups by these parameters (Table 4). The relationships we observed according to the reproductive factors were generally less apparent for the ER negative (−) than positive (+) tumors, possibly reflecting small numbers among the former cases. Among the ER+ tumors, we continued to observe patterns generally similar to those observed for all tumors—namely significant heterogeneity by age for MHT duration, age at menarche relations that were restricted to the women diagnosed at the oldest ages, and somewhat enhanced associations for numbers of births among the youngest women. Further, for the ER+ tumors, BMI and alcohol consumption were most strongly related to risk among the oldest women, although the p for heterogeneity across age groups was not significant. In contrast to the other risk factors, a late age at first birth was a risk factor for both ER− and ER+ tumors, with relationships predominating among the younger participants.
We also assessed risks within subgroups of tumors defined by histology (ductal vs. lobular) (Table 5). The rarer lobular tumors generally showed inconsistent relationships with most of the established breast cancer risk factors, possibly owing to small numbers. However, among the ductal tumors, we continued to observe support for possible age heterogeneity in directions similar to what had generally been observed among all cases combined, as well as among the ER+ cases. Thus, ductal carcinoma risk associations with parity and age at first birth were slightly attenuated among the older women, and alcohol consumption slightly enhanced. Increased risk of ductal carcinoma associated with MHT use also showed significant heterogeneity by age, although there were significant trends with increasing durations of use in all age groups.
A variety of sensitivity analyses, both excluding women (e.g., subjects with extreme values for caloric intake, those in poor or fair health, or those with no or infrequent breast cancer screening history) and truncating person years at the end of June 2002, resulted in no substantial changes to the results previously presented. Given small numbers of younger (50–59 years) women who developed ER− or lobular tumors, we also evaluated p values for heterogeneity comparing only the women aged 60–69 versus ≥70 years and also found no statistically significant differences.
Discussion
Given that breast cancer incidence rates increase with age and that the number of older women in the USA is increasing, it is important to understand the epidemiology of breast cancer among older women. In this large cohort of postmenopausal women, most risk factor relationships did not vary significantly by age, although we did observe some suggestive variations that might support potential etiologic differences.
Although previous studies have noted distinctive differences between risk factors for premenopausal versus postmenopausal onset breast cancers [2, 13], fewer have been able to assess risk factor differences among older women. By virtue of our study population, we had a unique opportunity to address this gap in research as all participants were aged 50 or older. However, there were challenges in interpreting some of the age differences given the relatively small number of younger (50–59 years) as compared to older women. In addition, it was difficult to compare our results to those of previous investigations because many of these studies had limited numbers within the age groups examined and the comparative age groups varied. For example, in some of the larger studies, older women were variously defined as 65 [6, 14], 70 [7, 8, 15], or 75 [5, 16] years or older.
Reproductive exposures have been the risk factors most commonly assessed in relation to later-onset breast cancers. Late menarche has been examined according to age at diagnosis among older women in several studies, with discrepant findings. While a lack of association with this parameter has been noted among older women [14], other studies have found either somewhat weaker [16] or stronger [6] inverse relationships among the oldest as compared to younger age groups. Our findings suggested little evidence of a relationship with age at menarche among younger women with the only significant inverse relationship seen among the older (≥70 years) subjects. We have no ready explanation for this, especially since we might have hypothesized stronger effects among younger women if recall was affecting the results. However, the variation by age that we observed was slight and not statistically significant.
Parity and late age at first birth have also been examined in several previous studies. Although some studies have failed to find relationships of breast cancer risk with either parity [5] or age at first birth [7] among older women, we, like two other investigations [6, 16], found the relationships of both factors somewhat attenuated among older as compared to younger women. Given that the effects of pregnancy on breast cancer risk are believed to operate, at least partially, though postpartum remodeling of breast tissue [17], it is not surprising that the effects would diminish with age. However, whether reproductive factors have no effect on the risk of the development of cancers at advanced ages or whether there is only a diminution with age has yet to be determined.
While the effect of reproductive factors on breast cancer risk appeared attenuated among older women in our study, we actually found stronger relations of BMI among our oldest study subjects, as has been observed in several other investigations [8, 18–20]. BMI is well documented as showing discrepant relations with breast cancer risk according to menopausal status, with inverse associations having consistently been seen for cancers of premenopausal onset and direct associations observed for postmenopausal cancers [21, 22]. The latter relation is accepted as deriving from the peripheral conversion of androgens to active estrogens in fat tissue [23]. Thus, a stronger relationship of obesity to breast cancers at older ages may reflect that these cancers may be more prone to the influence of endogenous estrogens. This might also explain the slightly enhanced risks associated with alcohol consumption for cancers occurring in the oldest women given that there is evidence that this risk factor may operate through estrogenic mechanisms [24]. The stronger relation of obesity to later-onset cancers would also be consistent with findings that more recent weight gain is a stronger predictor of breast cancer than weight earlier in life [25, 26].
Of all the risk factors examined, the only one that showed significant variation across the age groups was that of duration of MHT use. If not merely a chance finding, the significance of the interaction might reflect varying prevalences of exposure across the age groups, given that there were significant trends according to this parameter in all age groups. The statistical heterogeneity could also reflect that the strongest relationship was observed among women ages 60–69 years, which had the largest number of subjects. Our interpretation of this finding was also complicated by the absence of information on the baseline questionnaire on types of hormones prescribed, which is known to affect risk of all tumors as well as tumor subtypes.
We also examined risk in relation to age at menopause, a factor that has been consistently and positively related to breast cancer risk [27], presumably reflecting effects of prolonged circulating estrogens. This is not a parameter that has generally been thought to show discrepant results by age [14]. Similarly, in the present study, we also did not observe significant heterogeneity by age but did note slightly stronger relations for the older subjects. Although the differences were not pronounced, this would be consistent with the notion that it can take between 10 and 15 years for menopause to fully exert its effects on breast cancer risk [28]. Given this, we would have also expected to have seen some heterogeneity according to relations with surgical menopause, a factor found not to affect risk of breast cancer in older women in one investigation [7]. However, the association with surgical menopause in our study may have been obscured by the inclusion of both women with a hysterectomy, which has not generally been related to substantial alterations in breast cancer risk [29]—as well as those with a bilateral oophorectomy, which can substantially reduce subsequent breast cancer risk [30].
Although a family history of breast cancer is known to be a stronger predictor of risk for very early-onset cancers [31, 32], most studies have not shown variation in relations across the spectrum of older ages [5, 16], in agreement with results from our study. We also did not observe consistent differences according to age for a variety of other accepted breast cancer risk factors, including previous breast biopsies and levels of physical activity.
While we examined associations with all breast cancers in the cohort, it is well recognized that the disease is extremely heterogeneous, with distinctive relationships according to various clinical parameters [33, 34]. Divergent effects by both hormone receptor status [4] and histology [35, 36] have been reported. Although there is a fair amount of inconsistency in the literature regarding risk factor associations within clinical subgroups, there are some suggestions of stronger relationships of alcohol consumption with ER+ as compared to ER− tumors [37], BMI and menstrual factors with ductal as compared to lobular cancers [35], and menopausal hormone use with lobular as compared to ductal cancers [38, 39]. However, even when we controlled for this variation, we continued to observe some possible heterogeneity by age, particularly with respect to attenuated relations among older women for some of the reproductive parameters and increased risks of BMI and alcohol consumption. Not surprisingly, these relationships mainly prevailed for the tumors that are most common among older women, notably the ER+ and ductal tumors.
Given that breast cancer incidence rates increase with age and that more women are living longer, it can be expected that greater numbers of women will be diagnosed with breast cancer at older ages in the future [40]. It is also well documented that less screening mammography among older women is associated with later stage at diagnosis and poorer survival [41–43]. It is therefore important to determine whether the standard risk factors apply to such women. Our analysis suggests generally similar risk profiles for such cancers, with possibly a slight attenuation for reproductive factors and a modest enhancement for obesity and alcohol consumption. Although these differences among older women were not large, they may merit further consideration with respect to individualized breast cancer risk prediction, as has been detailed in several recent investigations [7, 44, 45].
References
Benz CC (2008) Impact of aging on the biology of breast cancer. Crit Rev Oncol Hematol 66(1):65–74
Jatoi I, Anderson WF (2010) Qualitative age interactions in breast cancer studies: a mini-review. Futur Oncol 6(11):1781–1788
Diab SG, Elledge RM, Clark GM (2000) Tumor characteristics and clinical outcome of elderly women with breast cancer. J Natl Cancer Inst 92(7):550–556
Chlebowski RT, Anderson GL, Lane DS et al (2007) Predicting risk of breast cancer in postmenopausal women by hormone receptor status. J Natl Cancer Inst 99(22):1695–1705
Poynter JN, Inoue-Choi M, Ross JA, Jacobs DR Jr, Robien K (2013) Reproductive, lifestyle, and anthropometric risk factors for cancer in elderly women. Cancer Epidemiol Biomark Prev 22(4):681–687
Shantakumar S, Terry MB, Teitelbaum SL et al (2007) Reproductive factors and breast cancer risk among older women. Breast Cancer Res Treat 102(3):365–374
Vacek PM, Skelly JM, Geller BM (2011) Breast cancer risk assessment in women aged 70 and older. Breast Cancer Res Treat 130(1):291–299
La VC, Negri E, Franceschi S et al (1997) Body mass index and post-menopausal breast cancer: an age-specific analysis. Br J Cancer 75(3):441–444
Schatzkin A, Subar AF, Thompson FE et al (2001) Design and serendipity in establishing a large cohort with wide dietary intake distributions: the National Institutes of Health-American Association of Retired Persons Diet and Health Study. Am J Epidemiol 154(12):1119–1125
Michaud DS, Midthune D, Hermansen S et al (2005) Comparison of cancer registry case ascertainment with SEER estimates and self-reporting in a subset of the NIH-AARP Diet and Health Study. J Regist Manag 32(2):70–75
SEER (2013) ICD-O-3. Coding materials. http://seer.cancer.gov/icd-o-3/
Hersh AL, Stefanick ML, Stafford RS (2004) National use of postmenopausal hormone therapy: annual trends and response to recent evidence. JAMA 291(1):47–53
Clavel-Chapelon F, Gerber M (2002) Reproductive factors and breast cancer risk. Do they differ according to age at diagnosis? Breast Cancer Res Treat 72(2):107–115
La Vecchia C, Negri E, Bruzzi P et al (1992) The role of age at menarche and at menopause on breast cancer risk: combined evidence from four case–control studies. Ann Oncol 3(8):625–629
Horn J, Asvold BO, Opdahl S, Tretli S, Vatten LJ (2013) Reproductive factors and the risk of breast cancer in old age: a Norwegian cohort study. Breast Cancer Res Treat 139(1):237–243
Sweeney C, Blair CK, Anderson KE, Lazovich D, Folsom AR (2004) Risk factors for breast cancer in elderly women. Am J Epidemiol 160(9):868–875
Faupel-Badger JM, Sherman ME, Garcia-Closas M et al (2010) Prolactin serum levels and breast cancer: relationships with risk factors and tumour characteristics among pre- and post-menopausal women in a population-based case–control study from Poland. Br J Cancer 103(7):1097–1102
Franceschi S, Favero A, La Vecchia C et al (1996) Body size indices and breast cancer risk before and after menopause. Int J Cancer 67(2):181–186
Lubin F, Ruder AM, Wax Y, Modan B (1985) Overweight and changes in weight throughout adult life in breast cancer etiology. A case–control study. Am J Epidemiol 122(4):579–588
van den Brandt PA, Dirx MJ, Ronckers CM, van den Hoogen P, Goldbohm RA (1997) Height, weight change, and postmenopausal breast cancer risk: the Netherlands Cohort Study. Cancer Causes Control 8(1):39–47
Rose DP, Vona-Davis L (2010) Interaction between menopausal status and obesity in affecting breast cancer risk. Maturitas 66(1):33–38
van den Brandt PA, Spiegelman D, Yaun SS et al (2000) Pooled analysis of prospective cohort studies on height, weight, and breast cancer risk. Am J Epidemiol 152(6):514–527
Siiteri PK (1987) Adipose tissue as a source of hormones. Am J Clin Nutr 45(1 Suppl.):277–282
Seitz HK, Pelucchi C, Bagnardi V, La VC (2012) Epidemiology and pathophysiology of alcohol and breast cancer: update 2012. Alcohol Alcohol 47(3):204–212
Ahn J, Schatzkin A, Lacey JV Jr et al (2007) Adiposity, adult weight change, and postmenopausal breast cancer risk. Arch Intern Med 167(19):2091–2102
Krishnan K, Bassett JK, Macinnis RJ et al (2013) Associations between weight in early adulthood, change in weight, and breast cancer risk in postmenopausal women. Cancer Epidemiol Biomark Prev 22(8):1409–1416
Collaborative Group on Hormonal Factors in Breast Cancer (2012) Menarche, menopause, and breast cancer risk: individual participant meta-analysis, including 118,964 women with breast cancer from 117 epidemiological studies Lancet Oncol 13(11):1141–1151
Brinton LA, Schairer C, Hoover RN, Fraumeni JF Jr (1988) Menstrual factors and risk of breast cancer. Cancer Invest 6(3):245–254
Press DJ, Sullivan-Halley J, Ursin G et al (2011) Breast cancer risk and ovariectomy, hysterectomy, and tubal sterilization in the women’s contraceptive and reproductive experiences study. Am J Epidemiol 173(1):38–47
Rebbeck TR, Kauff ND, Domchek SM (2009) Meta-analysis of risk reduction estimates associated with risk-reducing salpingo-oophorectomy in BRCA1 or BRCA2 mutation carriers. J Natl Cancer Inst 101(2):80–87
Collaborative Group on Hormonal Factors in Breast Cancer (2001) Familial breast cancer: collaborative reanalysis of individual data from 52 epidemiological studies including 58,209 women with breast cancer and 101,986 women without the disease Lancet 358(9291):1389–1399
Colditz GA, Kaphingst KA, Hankinson SE, Rosner B (2012) Family history and risk of breast cancer: nurses’ health study. Breast Cancer Res Treat 133(3):1097–1104
Garcia-Closas M, Brinton LA, Lissowska J et al (2006) Established breast cancer risk factors by clinically important tumour characteristics. Br J Cancer 95(1):123–129
Yang XR, Chang-Claude J, Goode EL et al (2011) Associations of breast cancer risk factors with tumor subtypes: a pooled analysis from the Breast Cancer Association Consortium studies. J Natl Cancer Inst 103(3):250–263
Li CI, Malone KE, Porter PL, Weiss NS, Tang MT, Daling JR (2003) Reproductive and anthropometric factors in relation to the risk of lobular and ductal breast carcinoma among women 65–79 years of age. Int J Cancer 107(4):647–651
Nyante SJ, Dallal CM, Gierach GL, Park Y, Hollenbeck AR, Brinton LA (2013) Risk factors for specific histopathological types of postmenopausal breast cancer in the NIH-AARP Diet and Health Study. Am J Epidemiol 178(3):359–371
Lew JQ, Freedman ND, Leitzmann MF et al (2009) 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 170(3):308–317
Li CI, Malone KE, Porter PL et al (2008) Relationship between menopausal hormone therapy and risk of ductal, lobular, and ductal–lobular breast carcinomas. Cancer Epidemiol Biomark Prev 17(1):43–50
Phipps AI, Li CI, Kerlikowske K, Barlow WE, Buist DS (2010) Risk factors for ductal, lobular, and mixed ductal-lobular breast cancer in a screening population. Cancer Epidemiol Biomark Prev 19(6):1643–1654
Bouwens CS, van Rensburg SJ, de Kock L, Apffelstaedt JP, Kotze MJ (2012) Influence of genetic factors on the development of breast cancer in the older woman. Curr Aging Sci 5(2):140–147
Jemal A, Fedewa SA (2012) Is the prevalence of ER-negative breast cancer in the US higher among Africa-born than US-born black women? Breast Cancer Res Treat 135(3):867–873
Randolph WM, Goodwin JS, Mahnken JD, Freeman JL (2002) Regular mammography use is associated with elimination of age-related disparities in size and stage of breast cancer at diagnosis. Ann Intern Med 137(10):783–790
Yancik R, Ries LG, Yates JW (1989) Breast cancer in aging women. A population-based study of contrasts in stage, surgery, and survival. Cancer 63(5):976–981
Pfeiffer RM, Park Y, Kreimer AR et al (2013) Risk prediction for breast, endometrial, and ovarian cancer in white women aged 50 y or older: derivation and validation from population-based cohort studies. PLoS Med 10(7):e1001492
Rosner BA, Colditz GA, Hankinson SE, Sullivan-Halley J, Lacey Jr JV, Bernstein L (2013) Validation of Rosner–Colditz breast cancer incidence model using an independent data set, the California Teachers Study. Breast Cancer Res Treat 142(1):187–202
Acknowledgments
This research was supported in part by the Intramural Research Program of the National Cancer Institute at the National Institutes of Health. Dr. Cher Dallal was supported by the Cancer Prevention Fellowship Program, National Institutes of Health, Bethesda, MD. Cancer incidence data from the collected by the Georgia Center for Cancer Statistics, Department of Epidemiology, Rollins School of Public Health, Emory University (for the Atlanta metropolitan area); the California Department of Health Services, Cancer Surveillance Section; the Michigan Cancer Surveillance Program, Community Health Administration, State of Michigan (for the Detroit metropolitan area); the Florida Cancer Data System (FCDC) under contract with the Florida Department of Health (FDOH); the Louisiana Tumor Registry, Louisiana State University Medical Center in New Orleans; the New Jersey State Cancer Registry, Cancer Epidemiology Services, New Jersey State Department of Health and Senior Services; the North Carolina Central Cancer Registry; the Division of Health Statistics and Research, Pennsylvania Department of Health, Harrisburg, Pennsylvania; the Arizona Cancer Registry, Division of Public Health Services, Arizona Department of Health Services; and the Texas Cancer Registry, Cancer Epidemiology and Surveillance Branch, Texas Department of State Health Services. The Pennsylvania Department of Health specifically disclaims responsibility for any analyses, interpretations, or conclusions. The views expressed herein are solely those of the authors and do not necessarily reflect those of the FCDC or FDOH. We are indebted to the participants in the NIH-AARP Diet and Health Study for their outstanding cooperation. We also thank Sigurd Hermansen and Kerry Grace Morrissey from Westat for study outcomes ascertainment and management and Leslie Carroll at Information Management Services for data support and analysis.
Conflict of interest
The authors declare that they have no conflict of interest.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Brinton, L.A., Smith, L., Gierach, G.L. et al. Breast cancer risk in older women: results from the NIH-AARP Diet and Health Study. Cancer Causes Control 25, 843–857 (2014). https://doi.org/10.1007/s10552-014-0385-3
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10552-014-0385-3