Reproductive and anthropometric factors in relation to the risk of lobular and ductal breast carcinoma among women 65–79 years of age
Version of Record online: 29 AUG 2003
Copyright © 2003 Wiley-Liss, Inc.
International Journal of Cancer
Volume 107, Issue 4, pages 647–651, 20 November 2003
How to Cite
Li, C. I., Malone, K. E., Porter, P. L., Weiss, N. S., Tang, M.-T. C. and Daling, J. R. (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: 647–651. doi: 10.1002/ijc.11465
- Issue online: 24 SEP 2003
- Version of Record online: 29 AUG 2003
- Manuscript Accepted: 25 JUN 2003
- Manuscript Revised: 17 JUN 2003
- Manuscript Received: 29 APR 2003
- National Cancer Institute
- Fred Hutchinson Cancer Research Center. Grant Number: RO1 CA072787
- lobular breast carcinoma;
- ductal breast carcinoma;
- reproductive factors;
- anthropometric factors
Use of combined estrogen-progestin hormone replacement therapy appears to be associated with an increased risk of invasive lobular breast carcinomas (ILC) and, to a lesser degree, with risk of invasive ductal carcinoma (IDC). Conceivably, ILCs are more hormonally responsive and so may be more strongly associated than IDCs with reproductive and anthropometric characteristics that can influence hormone levels. However, few epidemiologic studies of breast cancer have evaluated these factors by histologic type. We conducted a population-based case-control study of women aged 65–79 years in western Washington State. Responses from 975 women diagnosed with breast cancer during 1997–1999 were compared to those of 1,007 controls. Associations between various reproductive and anthropometric factors and risks of IDC (n = 656) and ILC (n = 196) were evaluated using polytomous logistic regression. Earlier age at menarche, later age at menopause and obesity were more strongly associated with elevated risks of IDC than ILC. Alternatively, oral contraceptive use was associated with an increased risk of ILC but not IDC. Thus, the pattern of results that we observed suggest that factors influencing endogenous hormones and duration of ovarian function may be more strongly associated with IDC risk, while exogenous hormones may be more strongly associated with ILC risk. © 2003 Wiley-Liss, Inc.
Six recent studies, including an analysis from the study presented herein, have reported that use of combined estrogen and progestin hormone replacement therapy (HRT) is associated with a 2.0- to 3.9-fold increased risk of invasive lobular carcinoma (ILC), the second most common histologic type of breast cancer, but has a more modest association with the most common histologic type, invasive ductal carcinoma (IDC).1, 2, 3, 4, 5, 6 ILCs are also more likely to be estrogen and progesterone receptor positive compared to IDCs.7 Thus, the development of lobular breast tumors could be under greater hormonal control than is the development of ductal tumors.
There have been numerous investigations of the incidence of postmenopausal breast cancer in relation to a woman's reproductive history and anthropometric characteristics, factors that involve at least temporary changes in endogenous hormone production.8, 9 However, few studies have evaluated the relationship between reproductive factors and the incidence of ILC.10, 11 One was an international multicenter hospital-based case-control study conducted by the World Health Organization that compared 321 women with lobular and tubular breast cancer to 2,407 women with other histologic types of breast cancer. It found that women with lobular and tubular carcinomas were more likely to be older at the time of their first live birth and to have ever used oral contraceptives, but that age at menarche and number of first live births did not differ by histology. However, this study did not include a control group of women without breast cancer, 82% of its subjects were premenopausal at the time of their breast cancer diagnosis and its results may not be generalizable to U.S. women because their reproductive and anthropometric characteristics differ from those of the populations included in this study. The second study was a population-based Danish cohort study, and it found that parity and increasing number of live births were associated with reduced risks of IDC, but did not alter risk of ILC, whereas increasing age at first birth was associated with increased risks of both IDC and ILC.
Since ILC incidence rates are rising in both the U.S. and Europe,12, 13, 14 primarily among postmenopausal women, and risk factors for breast cancer vary by age and menopausal status,8, 9 studies evaluating whether known hormonal risk factors for breast cancer other than HRT have different associations with risk of ILC compared to IDC among postmenopausal women are warranted. Using data from a population-based case-control study of breast cancer in older women, we evaluated the possible influence that reproductive and anthropometric factors have on risk of ILC and IDC among women 65–79 years of age.
MATERIAL AND METHODS
We conducted a population-based case-control study of women 65–79 years of age living in the 3 county Seattle-Puget Sound metropolitan area. Our study's protocol was approved by the Fred Hutchinson Cancer Research Center Institutional Review Board, and written informed consent was obtained from all study subjects prior to each interview.
Women aged 65–79 years with no previous history of in situ or invasive breast cancer who were diagnosed with invasive breast cancer between April 1, 1997 and May 31, 1999 were eligible as cases. The Cancer Surveillance System, the population-based tumor registry that serves the Seattle-Puget Sound region of Washington State and participates in the Surveillance, Epidemiology and End Results (SEER) program of the National Cancer Institute, was used to identify them. To be eligible for the study, cases had to live in King, Pierce or Snohomish counties and have a Health Care Financing Administration (HCFA) record, since these records were used to identify controls. Of the 1,210 eligible cases identified, 975 (80.6%) were interviewed. Information on tumor histology was ascertained from CSS, which abstracts data on tumor characteristics from medical records and pathology reports from institutions serving the area. CSS classifies histology using the International Classification of Diseases for Oncology (ICD-O) codes. We divided cases into 2 groups, with codes 8520 (lobular carcinoma, NOS) and 8522 (infiltrating duct and lobular carcinoma) used to define the 196 ILC cases and code 8500 (infiltrating duct carcinoma, NOS) used to define the 656 IDC cases. Cases with codes 8520 and 8522 were grouped together because results did not differ when these groups were assessed separately, and a previous report focusing on HRT use and risk of ILC grouped these 2 codes together when defining its lobular case group.4
HCFA records were used to identify women from the general population of female residents of King, Pierce and Snohomish counties to serve as controls. Controls were frequency matched to cases on age and county of residence. Of the 1,365 eligible women selected as controls, 1,007 (73.8%) were interviewed.
All subjects were interviewed in person and were asked about menstrual, contraceptive and reproductive histories, body size, use of HRT and medical history, including family history of cancer. Our questioning with regard to all of these factors was limited to exposures that occurred before each woman's reference date. The reference date used for each woman with breast cancer was her date of diagnosis. Control reference dates were assigned so as to reflect the expected distribution of reference dates among the cases. A life events calendar was used to aid subjects' recall of important dates in their reproductive history and of any episodes of oral contraceptive or HRT use. Subjects were asked what their weight was 1 month prior to their reference date and what their maximum lifetime height was. These 2 measurements were used to calculate body mass index (BMI). Demographic data and information on smoking status and alcohol use were also collected.
We compared the reproductive and anthropometric characteristics of all breast cancer cases to controls using unconditional logistic regression,15 and compared ILC and IDC cases to controls using polytomous logistic regression.16 The 123 cases with other breast cancer histologies were excluded from the latter analysis. Both statistical approaches were used to calculate odds ratios (ORs) as an estimate of the relative risk and to compute 95% confidence intervals (CIs). With respect to type of menopause, natural menopause was defined as the spontaneous cessation of menstrual periods or receipt of a simple hysterectomy after age 55, and induced menopause was defined as having a hysterectomy with a bilateral oophorectomy or having menstrual periods cease as a result of treatment with chemotherapy prior to reference date. Women who could not be classified into either of these categories, including those who had a hysterectomy without a bilateral oophorectomy before age 55, were considered to have an unknown type of menopause. Age at menopause was classified as age when natural or induced menopause occurred but was unknown for women with an unknown type of menopause. Among women with a known age at menopause and known age at menarche, as an index of duration of ovarian function we created a variable measuring the years between these 2 ages subtracted by any time women spent pregnant, nursing or using oral contraceptives.
The following categorical variables were evaluated as potential confounders for each reproductive and anthropometric factor we assessed: race (white, black, Asian/Pacific Islander, other/unknown), income (<$15,000, $15–$25,000, $25–50,000, >$50,000), marital status (married, widowed, divorced/separated, single never married), education (less than high school, high school graduate, some college, college graduate or higher), family history of breast cancer (first degree, no first degree), average weekly alcohol use during the 20 years prior to reference date (none, less than 7 drinks, 7 or more drinks), duration of unopposed estrogen HRT use (never, less than 5 years, five or more years) and duration of combined estrogen and progestin HRT use (never, less than 5 years, 5 or more years). We adjusted all analyses for age (continuous) since cases and controls were matched on age. We also adjusted analyses of breast cancer incidence in relation to: duration of breast feeding for number of full-term births; type of menopause for duration of combined estrogen and progestin HRT use; age at menopause for first-degree family history of breast cancer; and height and BMI for income, because adjustment for each of these additional factors changed the risk estimates by more than 10%.
Controls were more likely than cases to be nonwhite, and IDC cases were more likely than ILC cases to be Asian/Pacific Islanders (Table I). Controls were somewhat more likely to have less than a high school education than cases. Both ILC and IDC cases were also more likely than controls to have a first-degree family history of breast cancer and to have higher levels of alcohol consumption. ILC cases were more likely than both IDC cases and controls to have used unopposed estrogen HRT and combined estrogen and progestin HRT for 5 years or longer. ILC cases, IDC cases and controls were similar to each other with respect to the other demographic characteristics we assessed.
|Characteristic||Controls (n = 1,007)||All cases (n = 975)||Ductal (n = 656)||Lobular (n = 196)|
|Less than high school||153||15.2||126||12.9||87||13.3||19||9.7|
|High school graduate||395||39.3||376||38.6||251||38.3||84||42.9|
|First-degree family history of breast cancer|
|Average weekly alcohol consumption|
|Duration of unopposed estrogen HRT use|
|≥ 5 years||334||33.4||358||37.1||232||35.8||83||42.6|
|Duration of combined estrogen+progestin HRT use|
With respect to reproductive factors, there were some suggestions of differences in risk between ILC and IDC, but none of these differences were statistically significant (Table II). Women with a late age at menarche had a 20% decrease in their overall risk of breast cancer, though this finding was within the limits of chance. There was a suggestion that this risk varied by histologic type, the size of the reduction in risk being slightly greater for IDC than for ILC. Breast cancer risk among parous women was essentially equivalent to risk among nulliparous women. However, compared to women who had only one full-term birth, women who had 4 or more full-term births had decreased risks of both IDC and ILC. Women who were older when they had their first full-term birth had increased risks of all histologic types of breast cancer, IDC and ILC, when compared to women who were 19 years of age or younger at the time of their first full-term birth, though these increases were within the limits of chance. Compared to women who never nursed, those who breast fed for longer than 24 months had a reduced risk of IDC but not ILC. Relative to nonusers of oral contraceptives, women who used these agents for 5 years or longer had a 1.6-fold increased risk of ILC (95% CI: 1.0–2.6), but did not have an increased risk of IDC. Compared to women who experienced menopause at age 44 years or younger, women who were older when they experienced menopause had increased risks of all histologic types of breast cancer, though this increase was primarily limited to women with IDC. For example, compared to women with a menopause age of 44 years or younger, women with an age at menopause of 55 years or older had a 1.8-fold increased risk of IDC (95% CI: 1.1–2.7) but little change in their risk of ILC (OR = 1.2; 95% CI: 0.6–2.4). Risk of breast cancer increased as duration of ovarian function increased, though the magnitude of these increases were greater for IDC than for ILC.
|Reproductive factor||Controls n = 1007||Overall n = 975||OR1||95% CI||Ductal n = 656||OR||95% CI||Lobular n = 196||OR||95% CI||P-value IDC vs. ILC|
|Age at menarche|
|No. of full-term births|
|Age at first full-term birth|
|Duration of breast feeding (months)2|
|Duration of oral contraceptive use (years)|
|Type of menopause3|
|Age at menopause3|
|Duration of ovarian function, years, quartiles3|
Again, while there were some suggestions of differences in risk between ILC and IDC with regard to anthropometric factors, all of these differences were not statistically significant (Table III). Compared to women with a height of less than 160 cm, women 160 cm and taller had increased risks of both histologic types of breast cancer. Neither weight nor BMI was appreciably related to the incidence of ILC, but higher weights and higher BMIs were associated with elevations in IDC risk. For example, compared to women in the lowest quartile of BMI, women in the highest quartile had a 1.4-fold increased risk of IDC (95% CI: 1.0–1.9) but no change in their risk of ILC (OR = 1.0; 95% CI: 0.6–1.6).
|Anthropometric characteristic||Controls n = 1007||Overall n = 975||OR1||95% CI||Ductal n = 656||OR||95% CI||Lobular n = 196||OR||95% CI||P-value IDC vs. ILC|
|Weight at age 30|
Certain limitations of our study should be considered when interpreting the results. We did not conduct an independent pathology review of the tumors, but instead relied on the diagnoses made by numerous pathologists in the Seattle-Puget Sound area. Misclassification of tumor histology may have resulted in some instances. We were also limited somewhat by a relatively small number of women with lobular breast cancer.
In addition, we were only able to interview 80.6% of all eligible cases and 73.8% of all eligible controls. Our results could be biased if the women we were unable to interview differed from those who did participate with regard to their reproductive or anthropometric characteristics. Also, all of our data were based on self-report. This may have been a particular problem for our analyses of weight and BMI (calculated from self-reported heights and weights) since participants in research studies may underestimate their weight. This issue was addressed in a previous breast cancer case-control study conducted in our area in which participants were weighed and measured at the time of their interview. The correlation between self-reported weight 1 year before diagnosis and weight measured at the interview for women in this study was R = 0.89,17 suggesting that self-reported weight is a relatively valid measure among women in our population. Finally, the results of our study may only apply to women 65–79 years of age, as risk factors for breast cancer can vary by a woman's age and menopausal status.8, 9
There is a growing body of evidence that combined estrogen and progestin HRT use is associated with a greater increase in risk of ILC compared to IDC.1, 2, 3, 4, 5, 6 Consistent with this evidence, ILCs are more likely to be both estrogen and progesterone receptor positive compared to IDCs, suggesting that they are more hormonally responsive. In addition, progesterone has long been known to promote lobular differentiation.18 However, only 2 studies have evaluated the relationship between reproductive risk factors for breast cancer and risk of ILC.10, 11 Numerous studies have shown that early age at menarche, late age at menopause, nulliparity, late age at first full-term pregnancy and low parity are associated with an increased risk of breast cancer as a whole.8, 9 In addition, women who are obese and/or tall are at increased risk of postmenopausal breast cancer.8, 9 The mechanisms underlying these associations are thought to be largely hormonal and related to the influence endogenous estrogen and progesterone have on the breast.
However, with the exception of oral contraceptive use, there was no suggestion in our data that any of the reproductive or anthropometric characteristics that are related to breast cancer as a whole were more strongly associated with ILC than IDC. Indeed, age at menarche, age at menopause and duration of ovarian function were more strongly associated with risk of IDC than with risk of ILC. Together these findings suggest that longer duration of ovarian function may be a stronger risk factor for IDC than it is for ILC. Similarly, a positive association for weight and BMI tended to be present for IDC but not for ILC. Alternatively, we found that while users of oral contraceptives for 5 years or longer did not have an increased risk of IDC, they did have an increased risk of ILC. This observation is consistent with that of Stalsberg et al.10 It is important to note that none of the differences we observed between ILC and IDC were statistically significant, so these results should be interpreted cautiously.
Taken as a whole, our results suggest that IDC risk appears to be more strongly related to factors that increase endogenous hormone levels or are related to longer durations of ovarian function than is ILC risk, whereas ILC risk is more strongly related to exogenous hormone exposures, both oral contraceptives and combined estrogen and progestin HRT. Though the specific mechanisms underlying these epidemiologic differences are unclear, our results do suggest that pathways leading to ILC and IDC are different. There is a growing body of literature to support this conclusion, since ILCs and IDCs are known to differ with respect to their expression of a variety of molecular tumor markers in addition to estrogen and progesterone receptor, including cyclin D1,19 vascular endothelial growth factor20 and e-cadherin.21 Nevertheless, additional studies that examine whether and how hormonal risk factors influence risk of ILC compared to IDC are needed to expand our understanding of the etiology of different histologic types of breast cancer, as no firm conclusions can be drawn based on this study and the limited available literature addressing these issues.
None of the authors has a conflict of interest with the work presented in this article.
- 8The epidemiology of breast cancer. Women Cancer 1998; 1S: 7–13..
- 15Statistical methods in cancer research. Vol. I. The analysis of case-control studies. Lyon: IARC Scientific Publications, 1980. 5–338., .
- 18Problems in breast pathology, 1st ed. Eastbourne: WB Saunders, 1979. 240–57..