Abstract
Purpose
The reproductive windows between age at menarche and first childbirth (standardized AFB) and from menarche to menopause (reproductive lifespan) may interact with genetic variants in association with breast cancer risk.
Methods
We assessed this hypothesis in 6131 breast cancer cases and 7274 controls who participated in the population-based Collaborative Breast Cancer Study. Risk factor information was collected through telephone interviews and DNA samples were collected on a sub-sample (N=1484 cases, 1307 controls) to genotype for 13 genome-wide association study-identified loci. Adjusted odds ratios (OR) and 95% confidence intervals (CI) were calculated and P-values for the interaction between reproductive windows and genotypes were obtained by adding cross-product terms to statistical models.
Results
For standardized AFB, the OR was 1.52 (CI:1.36-1.71) comparing the highest to lowest quintile. Carrier status for rs10941679 (5p12) and rs10483813 (RAD51B) appeared to modify this relationship (P=0.04 and P=0.02, respectively). For reproductive lifespan, the OR comparing the highest and lowest quintiles was 1.62 (CI:1.35-1.95). No interactions were detected between genotype and reproductive lifespan (all P>0.05). All results were similar regardless of ductal versus lobular breast cancer subtype.
Conclusions
Our results suggest reproductive windows are associated with breast cancer risk, and that associations may vary by genetic variants.
Keywords: breast neoplasms, epidemiology, menarche, menopause, genetic loci, histology
Introduction
Reproductive and menstrual exposures such as the ages at menarche, first full-term birth and menopause have been consistently, but modestly, associated with breast cancer risk. The windows defined by these reproductive experiences have also been associated with risk (1–3), such as the length of time between the initiation of menarche and the age at first full-term birth (AFB), termed the “standardized age at first birth” (1). Standardized age at first birth has been proposed to represent the number of ovulatory cycles a woman experiences with undifferentiated breast tissue, which is hypothesized by Russo and Russo to be more susceptible to the proliferative effects of ovarian hormones (4).
Epidemiologic studies have produced findings consistent with this hypothesis (1,2). In one study postmenopausal nulliparous women and women with 15 or more years between menarche and AFB had an increased breast cancer risk when compared to women with an interval of fewer than 10 years (nulliparous relative risk=2.0, 95% confidence interval(CI): 1.0-3.8; ≥15 years relative risk=2.4, CI: 1.3-4.3) (1).
A second reproductive window possibly linked to breast cancer risk is defined as the “reproductive lifespan”, and comprises the time between ages at menarche and natural menopause. Several studies report an association between longer reproductive lifespan and increased breast cancer risk, particularly when comparing extremes of the risk factor's spectrum (2,3,5). Approaches to estimating the reproductive lifespan have varied, with some investigators removing anovulatory phases from the calculation of reproductive lifespan (for example, time during pregnancy, lactation, or exogenous hormone use) (2,5), and other investigators including these phases (1). A small case-control study assessed the role of reproductive lifespan excluding time for pregnancy, lactation, and hormone use on postmenopausal breast cancer risk and found a one-year increase in reproductive lifespan was associated with a 10% increased odds of breast cancer risk (CI: 2-19%) (5). However, at least one study found no association between reproductive lifespan and breast cancer risk (6). Further research is needed to confirm the associations between reproductive windows and breast cancer risk.
Reproductive windows may have stronger associations in certain subgroups identified by genetic factors. Many of these single nucleotide polymorphisms (SNPs) identified in genome-wide association studies (GWAS) act through unknown mechanisms but epidemiological evidence suggests that certain breast cancer susceptibility loci modify the effects of hormonal risk factors on breast cancer risk (7–9). We assessed the relationships between reproductive windows, top GWAS-identified loci, and their interactions in relation to invasive breast cancer risk in a population-based case-control study.
Methods
Study Sample
The study data arise from the Collaborative Breast Cancer Study, a previously described population-based case-control study (10–12). Eligible participants were selected from English-speaking female residents of Massachusetts, New Hampshire or Wisconsin. Cases were women age 20-69 with an incident invasive breast cancer reported to a state cancer registry between 1995 and 2000. Community controls were randomly selected in each state from lists of licensed drivers (<age 65) or Medicare beneficiaries (≥age 65) and were frequency matched to approximate the age distribution of the cases within five-year age strata. Participants gave informed consent during study enrollment. This study was conducted with the approval of the University of Wisconsin Health Sciences Institutional Review Board.
Data collection
Telephone interviews were used to obtain detailed information on reproductive and menstrual experiences. Participant interviews were conducted on average one year after a specified reference date which was defined as the date of cancer diagnosis for the cases. A comparable reference point for control participants was calculated based on their five-year age strata and date of interview (12). Among eligible participants approximately 80% (N=6421) of cases and 76% (N=7673) of controls completed the interview.
Information about the histology and stage of breast cancer was obtained from each state's cancer registry. Cases were grouped by histology using International Classification of Diseases-Oncology codes (ICD-10), ductal (code 8010, 8012, 8021, 8140, 8310, 8323, 8410, 8500, 8502, 8530, 8560, 8571), lobular (code 8520) and mixed ductal-lobular (code 8521, 8522, 8523) (13). All other individual tumor subtypes were excluded from histologic specific analyses (N=463).
DNA extraction and genotyping
For a selected sample interviewed between the years 2000-2001, participants were asked to donate a buccal cell sample for genetic analyses. 70% (N=1717) of approached cases and 61% (N=1524) of approached controls agreed to donate a sample.
Samples were sent by participants through the mail directly to a National Cancer Institute-affiliated laboratory lab under the direction of Dr. Montserrat Garcia-Closas for processing. DNA collection, isolation and storage were conducted according to previously described protocols (10). DNA was quantitated from frozen aliquots and plated for the genotyping assays. Top significant results from GWAS and follow-up studies were used to identify candidate loci for this analysis (7,14–17). In total 13 SNPs were genotyped: rs4973768, rs10941679, rs2981582, rs3817198, rs3803662, rs13281615, rs11249433, rs889312, rs2046210, rs17468277, rs10483813, rs13387042, rs6504950. Genotyping was conducted using Taqman nuclease assay (Taqman®) with reagents designed by Applied Biosystems (http://www.appliedbiosystems.com/) as Assays-by-Design™ and performed using the ABI PRISM 7900HT, 7700 or 7500 Sequence Detection Systems according to the manufacturer's instructions.
To reduce the possibility of population stratification and maintain a study sample with similar ancestry to the GWAS in which the loci were identified, all analyses were limited to participants self-identified as White/Caucasian in race (95.1% of participants). Quality control measures were taken to remove poor quality genetic data. SNPs missing >20% of values or individual participants with a call rate <80% for genotypic data were excluded from the analysis. All 13 SNPs passed quality control measures. 178 cases and 174 controls were removed from genetic analyses due to a high percent of missing genotype data for a total of 1484 breast cancer cases and 1307 community controls with a viable DNA sample for study analyses.
Reproductive window definitions
The first reproductive window, standardized age at first birth, was defined as the interval between age at menarche and age at first full term birth in which duration of oral contraceptive use before a first pregnancy was subtracted from the interval. Participants with negative standardized age at first birth values were set to zero (N=5). A secondary variable was created where the time when participants used oral contraceptives before first pregnancy was included in the interval. Among postmenopausal nulliparous women standardized age at first birth was defined as the difference between age at menarche and menopause. Premenopausal nulliparous women (N=446 cases and 414 controls) were excluded from this analysis.
The second reproductive window examined was the reproductive lifespan and restricted to postmenopausal women. Participants were considered postmenopausal if they reported their menstrual cycles had stopped for at least the last six months prior to reference date. The postmenopausal participants were categorized into two groups: participants with natural menopause or a second group defined as postmenopausal due to other causes. The primary definition of reproductive lifespan was the time between age at menarche and age at natural menopause excluding phases of pregnancy, lactation and oral contraceptive use. A secondary definition was established in which these phases were included in the window. A third reproductive lifespan analysis was conducted among all postmenopausal women irrespective of menopause type.
Population for analysis
A total of 6131 White breast cancer cases and 7274 White controls were included in this analysis. Of these 1484 cases and 1307 controls were involved in the genetic analyses. Participants with missing data from a component of the reproductive windows (age at menarche N=146, oral contraceptive duration N=162, parity N=15, lactation duration N=38, age at menopause N=1543; among DNA participants: age at menarche N=29, oral contraceptive duration N=27, parity N=15, lactation duration N=13, age at menopause N=390) were excluded from analyses that involved the missing variable.
Statistical Analysis
Frequencies and distributions of reproductive, menstrual and established breast cancer risk factors were evaluated. We calculated odds ratios (OR) and 95% confidence intervals (CI) using logistic regression to assess the associations between reproductive windows and breast cancer risk. All statistical models included a term for age and state of residence. Reproductive windows were categorized into quintiles according to the variable distribution in control participants. These cut-points were kept throughout for ease of comparison analyses. Reproductive window analyses were further adjusted for the following potential confounders identified a priori: family history of breast cancer, body mass index (BMI, kg/m2) at one year prior to reference date, hormone therapy use, recent alcohol use, and education (variables categorized as displayed in Table 1).
Table 1.
Reproductive and menstrual characteristics by case status.
| Characteristic | All participants |
Participants with a DNA sample |
||
|---|---|---|---|---|
| Cases N (%) (N = 6131) | Controls N (%) (N = 7274) | Cases N (%) (N = 1484) | Controls N (%) (N = 1307) | |
| Age (median, standard deviation) | 54.3 (9.5) | 54.4 (9.6) | 54.5 (9.0) | 54.5 (9.4) |
| Family history of breast cancer | ||||
| Absent | 4765 (77.7) | 6211 (85.4) | 1145 (77.2) | 1108 (84.8) |
| Present | 1252 (20.4) | 924 (12.7) | 303 (20.4) | 175 (13.4) |
| Age at menarche (years) | ||||
| <12 | 1276 (20.8) | 1394 (19.2) | 317 (21.4) | 255 (19.5) |
| 12 | 1474 (24.0) | 1680 (23.1) | 397 (26.8) | 290 (22.2) |
| 13 | 1750 (28.5) | 2051 (28.2) | 402 (27.1) | 378 (28.9) |
| 14 | 929 (15.2) | 1152 (15.8) | 213 (14.4) | 215 (16.5) |
| ≥ 15 | 638 (10.4) | 918 (12.6) | 137 (9.2) | 159 (12.2) |
| Age at first full-term birth (years) | ||||
| <20 | 881 (14.4) | 1205 (16.6) | 213 (14.4) | 211 (16.1) |
| 20-24 | 2415 (39.4) | 3110 (42.8) | 609 (41.0) | 536 (41.0) |
| 25-29 | 1323 (21.6) | 2374(32.6) | 316 (21.3) | 266 (20.4) |
| ≥ 30 | 616 (10.1) | 585 (8.0) | 143 (9.6) | 118 (9.0) |
| Parity | ||||
| Nulliparous | 885 (14.4) | 891 (12.3) | 199 (13.4) | 171 (13.1) |
| 1 | 714 (11.7) | 769 (10.6) | 173 (11.7) | 145 (11.1) |
| 2 | 1950 (31.8) | 2130 (29.3) | 483 (32.6) | 420 (32.1) |
| 3 | 1307 (21.3) | 1677 (23.1) | 324 (21.8) | 270 (20.7) |
| ≥ 4 | 1268 (20.7) | 1799 (24.7) | 303 (20.4) | 298 (22.8) |
| Lactationa (months) | ||||
| Never | 3731 (61.0) | 4307 (59.4) | 878 (59.6) | 756 (57.7) |
| Ever | 2385 (39.0) | 2945 (40.6) | 602 (40.7) | 545 (42.0) |
| ≤ 3 | 702 (11.5) | 940 (12.9) | 171 (11.5) | 158 (12.1) |
| 4-12 | 883 (14.4) | 991 (13.6) | 207 (14.0) | 202 (15.5) |
| 13-23 | 467 (7.6) | 551 (7.6) | 125 (8.4) | 104 (8.0) |
| ≥ 24 | 333 (5.4) | 464 (6.4) | 99 (6.8) | 81 (6.2) |
| Oral contraceptive use (years) | ||||
| Never | 2785 (45.8) | 3434 (47.6) | 641 (43.2) | 558 (42.7) |
| Ever | 3302 (54.2) | 3775 (52.4) | 833 (56.3) | 740 (56.6) |
| <3 | 934 (15.2) | 1104 (15.2) | 234 (15.8) | 210 (16.1) |
| 3-6 | 1126 (18.4) | 1283 (17.6) | 282 (19.1) | 254 (19.4) |
| ≥ 7 | 1242 (20.3) | 1388 (19.1) | 317 (21.4) | 276 (21.1) |
| Menopausal status | ||||
| Premenopausal | 2385 (38.9) | 2801 (38.5) | 565 (38.1) | 470 (36.0) |
| Postmenopausal | 3355 (54.7) | 4029 (55.4) | 822 (55.4) | 742 (56.8) |
| Age at menopause (years) | ||||
| <45 | 481 (14.3) | 767 (19.0) | 123 (15.0) | 130 (17.5) |
| 45-49 | 700 (20.9) | 871 (21.6) | 171 (20.8) | 173 (23.3) |
| 50-54 | 1085 (32.3) | 1216 (30.2) | 255 (31.0) | 220 (29.7) |
| ≥ 55 | 396 (11.8) | 388 (9.6) | 91 (11.1) | 59 (8.0) |
| Age at natural menopause (years) | ||||
| <45 | 247 (10.8) | 372 (14.0) | 64 (11.7) | 69 (14.1) |
| 45-49 | 514 (22.5) | 660 (24.8) | 127 (23.2) | 133 (27.2) |
| 50-54 | 986 (43.2) | 1104 (41.4) | 232 (42.3) | 205 (41.9) |
| ≥ 55 | 375 (16.4) | 369 (13.8) | 85 (15.5) | 55 (11.3) |
| Postmenopausal hormone therapy | ||||
| Never | 1534 (45.7) | 2046 (50.8) | 333 (40.5) | 336 (45.3) |
| Estrogen-only | 767 (22.9) | 1000 (24.8) | 191 (23.2) | 192 (25.9) |
| Estrogen+ progestogen | 945 (28.2) | 867 (21.5) | 265 (32.2) | 192 (25.9) |
| Other combination | 109 (3.2) | 116 (2.9) | 33 (4.0) | 22 (2.9) |
| Body mass indexb (kg/m2) | ||||
| <25 (average) | 2945 (48.0) | 3477 (47.8) | 708 (47.7) | 629 (48.1) |
| 25-29 (overweight) | 1906 (31.1) | 2321 (31.9) | 453 (30.5) | 418 (32.0) |
| ≥ 30 (obese) | 1237 (20.2) | 1410 (19.4) | 314 (21.2) | 248 (19.0) |
| Recent alcohol use (drinks/week) | ||||
| 0 | 1056 (17.2) | 1279 (17.6) | 232 (15.6) | 198 (15.2) |
| 1-6 | 2170 (35.4) | 2736 (37.6) | 546 (36.8) | 536 (41.0) |
| ≥ 7 | 1995 (32.5) | 2334 (32.1) | 499 (33.6) | 420 (32.1) |
| Education | ||||
| Less than high school | 365 (6.0) | 500 (6.9) | 84 (5.7) | 82 (6.3) |
| High school | 2418 (39.4) | 2886 (39.7) | 565 (38.1) | 515 (39.4) |
| Some college | 1600 (26.1) | 1961 (27.0) | 396 (26.7) | 341 (26.1) |
| College degree | 1747 (28.5) | 1924 (26.5) | 439 (29.6) | 369 (28.2) |
| Histologic subtype | ||||
| Ductal | 4791 (78.1) | 1317 (79.2) | ||
| Lobular | 575 (9.4) | 154 (9.3) | ||
| Mixed ductal-lobular | 302 (4.9) | 85 (5.1) | ||
| Other | 463 (7.6) | 106 (6.4) | ||
Parous women only.
Body mass index levels were as defined by the World Health Organization.
Certain subgroups do not add to total due to missing data.
Polytomous regression analyses
The relationships between the length of the reproductive windows and ductal, lobular, and mixed ductal-lobular histologic subtypes of breast cancer were evaluated by ORs and 95% CI obtained from multivariate polytomous logistic regression models using controls as the common comparison group (18).
Genetic analysis
Hardy-Weinberg equilibrium was tested among controls with a DNA sample by using chi-squared tests to compare the observed to expected genotype frequencies. Per-allele ORs and 95% CIs were calculated for the association between each SNP and breast cancer risk represented by an ordinal term in the regression model for the number of minor alleles present. Potential effect modification of the associations between reproductive windows and breast cancer risk by individual SNPs were evaluated by including cross-product terms combining the number of minor alleles by the reproductive window in multivariate models. To determine significance, log likelihood values were compared to assess whether the cross-product term contributed significantly to the model. To further elucidate risk pattern differences by genotypes and reproductive windows stratified odds ratios and 95% CI were calculated for interactions that were significant (P≤0.05). Reproductive windows OR were grouped into non-risk and risk allele categories. Statistical analyses were conducted in SAS software (Cary, NC 9.1).
Results
In the full study sample, cases were more likely to have a family history of breast cancer, earlier age at menarche, fewer children and later age at menopause than controls (Table 1). Cases and controls were equally likely to be premenopausal and the majority of breast cancer cases were diagnosed with ductal carcinoma (78.1%, N=4791). The subgroup of women who donated a DNA sample was similar to the total participant sample in age, family history of breast cancer and other established risk factors for breast cancer. In analyses restricted to women with a DNA sample, case participants had an earlier age at menarche and later age at menopause than controls. There was no significant difference in parity by case-control status in the DNA sample subgroup.
The range of standardized age at first birth values was 0 to 48 years and a median length of 10.8 years. There was a trend toward increased breast cancer risk with longer duration of the window between age at menarche and first birth (P<0.001). Women in the highest quintile for standardized age at first birth interval (≥ 15.3 years) were at a 52% (95%CI: 36%-71%) higher risk of breast cancer compared to women with the shortest intervals (<7.3 years) (Table 2). Results restricted to participants with a DNA sample were similar in magnitude and direction, but confidence intervals were less precise. Standardized age at first birth results were similar in statistical analyses when calculated by including or excluding duration of oral contraceptive use before first pregnancy and when postmenopausal nulliparous women were removed from the interval (data not shown). When potential confounders were added to the model point estimates for the association between standardized age at first birth and breast cancer risk did not change substantially (Table 2).
Table 2.
Odds ratios and 95% confidence intervals for the association between interval between age at menarche and age at first birth (quintiles) and invasive breast cancer.
| Quintile (years) | Cases | Controls | Odds ratio (95% CI)a | Odds ratio (95% CI)b |
|---|---|---|---|---|
| 1st <7.3 | 932 | 1316 | 1 (ref.) | 1 (ref.) |
| 2nd 7.3-9.4 | 970 | 1317 | 1.05 (0.93-1.18) | 1.02 (0.91-1.15) |
| 3rd 9.5-11.6 | 1040 | 1321 | 1.13 (1.01-1.28) | 1.10 (0.98-1.25) |
| 4th 11.7-15.2 | 1193 | 1316 | 1.32 (1.18-1.48) | 1.31 (1.16-1.47) |
| 5th ≥ 15.3 | 1356 | 1319 | 1.52 (1.36-1.71) | 1.52 (1.34-1.71) |
| per-quintile estimate | 1.09 (1.07-1.12) | |||
| P-trend | <0.0001 | |||
| Participants with a DNA sample | ||||
|---|---|---|---|---|
| Quintile (years) | Cases | Controls | Odds ratio (95% CI)a | Odds ratio (95% CI)b |
| 1st <7.3 | 217 | 237 | 1 (ref.) | 1 (ref.) |
| 2nd 7.3-9.4 | 252 | 235 | 1.17 (0.90-1.51) | 1.14 (0.88-1.49) |
| 3rd 9.5-11.6 | 246 | 207 | 1.32 (1.02-1.72) | 1.28 (0.98-1.68) |
| 4th 11.7-15.2 | 295 | 243 | 1.36 (1.05-1.75) | 1.35 (1.04-1.77) |
| 5th ≥ 15.3 | 325 | 253 | 1.49 (1.16-1.92) | 1.49 (1.14-1.95) |
| per-quintile estimate | 1.07 (1.00-1.12) | |||
| P-trend | 0.02 | |||
Analyses exclude time for oral contraceptive use before first pregnancy. CI indicates confidence interval; ref, reference category.
Odds ratios are adjusted for age and state.
Odds ratios further adjusted for family history of breast cancer, recent alcohol consumption, body mass index, education, age at menopause and postmenopausal hormone therapy use.
The median length of reproductive lifespan calculated among postmenopausal women with natural menopause was 33.3 years (range 2 to 48 years). In analyses with anovulatory time removed, longer reproductive lifespan was significantly associated with increased breast cancer risk in women with natural menopause and in those with all-cause menopause (Table 3). Among women with natural menopause, the OR was 1.62 (95%CI: 1.35-1.95) for ≥37.4 reproductive lifespan years compared to <26.0 reproductive lifespan years. In the larger group with all-cause menopause the analogous comparison yielded an OR of 1.77 (95%CI: 1.51-2.08). Overall, analyses including all postmenopausal women yielded point estimates further from unity. The median value for reproductive lifespan when anovulatory phases were included was 38 years (range, 2-49 years). The findings were similar when restricted to women with a DNA sample, but confidence intervals were less precise (Table 3). We found little evidence of confounding of the association between reproductive lifespan and breast cancer risk by established breast cancer risk factors (Table 3). Analyses that did not subtract anovulatory time due to pregnancy, lactation, and oral contraceptive use produced slightly attenuated results (data not shown).
Table 3.
Odds ratios and 95% confidence intervals for the association between age at menarche and age at natural menopause (quintiles) and invasive breast cancer, by menopausal type.
| Natural menopause |
All-cause menopause |
||||||
|---|---|---|---|---|---|---|---|
| Quintile(years) | Cases | Controls | OR (95% CI)a | OR (95% CI)b | Cases | Controls | OR (95% CI)a |
| 1st <26.0 | 349 | 491 | 1 (ref.) | 1 (ref.) | 564 | 868 | 1 (ref.) |
| 2nd 26.1-31.1 | 387 | 492 | 1.12 (0.93-1.36) | 1.14 (0.94-1.39) | 517 | 643 | 1.25 (1.07-1.47) |
| 3rd 31.2-34.3 | 399 | 498 | 1.15 (0.95-1.36) | 1.20 (0.99-1.46) | 486 | 597 | 1.28 (1.09-1.50) |
| 4th 34.4-37.3 | 408 | 486 | 1.20 (0.99-1.45) | 1.23 (1.01-1.49) | 453 | 539 | 1.31 (1.11-1.54) |
| 5th ≥ 37.4 | 549 | 492 | 1.62 (1.35-1.95) | 1.65 (1.37-1.99) | 597 | 533 | 1.77 (1.51-2.08) |
| per-quintile estimate | 1.03 (1.01-1.06) | 1.06 (1.03-1.08) | |||||
| P-trend | <0.0001 | <0.0001 | |||||
| Participants with a DNA sample | |||||||
|---|---|---|---|---|---|---|---|
| Natural menopause |
All-cause menopause |
||||||
| Quintile(years) | Cases | Controls | OR (95% CI)a | OR (95% CI)b | Cases | Controls | OR (95% CI)a |
| 1st <26.0 | 91 | 109 | 1 (ref.) | 1 (ref.) | 150 | 168 | 1 (ref.) |
| 2nd 26.1-31.1 | 91 | 89 | 1.21 (0.81-1.82) | 1.30 (0.86-1.96) | 118 | 116 | 1.13 (0.80-1.59) |
| 3rd 31.2-34.3 | 97 | 84 | 1.40 (0.93-2.11) | 1.50 (1.00-2.27) | 118 | 96 | 1.40 (0.99-1.99) |
| 4th 34.4-37.3 | 94 | 97 | 1.19 (0.79-1.77) | 1.22 (0.81-1.84) | 105 | 109 | 1.10 (0.77-1.56) |
| 5th ≥ 37.4 | 124 | 76 | 2.01 (1.34-3.02) | 2.12 (1.40-3.19) | 136 | 82 | 1.92 (1.34-2.76) |
| per-quintile estimate | 1.08 (1.02-1.13) | 1.07 (1.02-1.12) | |||||
| P-trend | 0.05 | 0.05 | |||||
Reproductive lifespan analyses completed only among postmenopausal women. Analyses exclude time for oral contraceptive use, lactation, and pregnancy. OR indicates odds ratio; CI, confidence interval; ref, reference category.
Odds ratios are adjusted for age and state.
Odds ratios further adjusted for family history of breast cancer, recent alcohol consumption, body mass index, education and postmenopausal hormone therapy use.
Histologic-specific subtype results
We found no evidence of heterogeneity in the association between reproductive windows and histological subtype (ductal, lobular and mixed ductal-lobular) (data not shown).
Genetic analyses and interactions
We confirmed associations between 7 of the 13 candidate SNPs and breast cancer risk (rs13387042, rs4973768, rs10941679, rs2981582, rs3817198, rs3803662 and rs6504950) (Table 4). The risk alleles of these SNPs were associated with 11-22% increase in breast cancer risk.
Table 4.
Per-allele odds ratiosa for the associations between breast cancer susceptibility loci and breast cancer risk.
| Chromosomal location/Gene | SNP | Major/Minor Allele | MAF | OR (95% CI) | ptrend |
|---|---|---|---|---|---|
| 10q26/FGFR2 | rs2981582 | C/T | 0.40 | 1.22 (1.08-1.38) | 0.001 |
| 16q12/TOX3 | rs3803662 | C/T | 0.28 | 1.22 (1.07-1.39) | 0.003 |
| 5p12 | rs10941679 | A/G | 0.27 | 1.20 (1.05-1.37) | 0.01 |
| 3p24/SLC4A7 | rs4973768 | C/T | 0.50 | 1.19 (1.06-1.34) | 0.004 |
| 11p15/LSP1 | rs3817198 | T/C | 0.32 | 1.15 (1.02-1.31) | 0.02 |
| 8q24 | rs13281615 | A/G | 0.42 | 1.07 (0.95-1.20) | 0.23 |
| 1p11 | rs11249433 | T/C | 0.43 | 1.06 (0.95-1.19) | 0.31 |
| 5q11/MAP3K1 | rs889312 | A/C | 0.29 | 1.06 (0.93-1.20) | 0.43 |
| 6q25 | rs2046210 | C/T | 0.37 | 1.00 (0.91-1.13) | 0.80 |
| 2q33/ALS2CR12/CASP8 | rs17468277 | G/A | 0.12 | 1.00 (0.84-1.19) | 0.99 |
| 14q24/RAD51B | rs10483813 | T/A | 0.24 | 0.98 (0.85-1.12) | 0.73 |
| 2q35 | rs13387042 | T/C | 0.46 | 0.89 (0.79-1.00) | 0.04 |
| 17q23/STXBP4 | rs6504950 | G/A | 0.27 | 0.83 (0.73-0.95) | 0.01 |
Odds ratios are adjusted for age and state of residence. MAF indicates minor allele frequency; OR, odds ratio; CI, confidence interval. P-trend calculated by inclusion of an ordinal term to represent the number of increasing minor alleles.
We observed evidence that two breast cancer susceptibility loci modified the association between standardized age at first birth and breast cancer risk (Table 5). Standardized age at first birth analyses stratified by rs10483813 (RAD51B) genotype revealed distinct trends by the presence of risk alleles (P=0.02). Women homozygous for the major allele of rs10483813 (TT genotype) and with longer time between menarche and AFB had an increased risk of breast cancer. For example, women homozygous for the T allele and in the quintile for longest interval between age at menarche and age at first birth had two-times the odds of breast cancer (95%CI: 1.47-3.05) compared to women in the quintile for the shortest interval. In the analyses restricted to women with AT or TT genotype the associations between standardized age at first birth and breast cancer risk were null. The risk allele of rs10941679 at 5p12 also modified the relationship between standardized age at first birth and breast cancer risk (P=0.04) (Table 5). Women with at least one copy of the G allele and longer length of standardized age at first birth interval had higher breast cancer risk. The effect of standardized age at first birth was effectively null in women homozygous for the major allele (A). No interactions were detected between the 13 evaluated SNPs and reproductive lifespan (all P>0.05).
Table 5.
Odds ratiosa and 95% confidence intervals for the association between the interval from age at menarche to age at first birth and breast cancer risk, stratified by rs10483813 and rs10941679.
| Quintiles (years) | OR (95%CI) | OR (95%CI) | pInteraction |
|---|---|---|---|
| rs10483813 | TT | AT/AA | |
| 1st <7.3 | 1 (ref.) | 1 (ref.) | |
| 2nd 7.3-9.4 | 1.46 (1.01-2.11) | 1.08 (0.69-1.70) | |
| 3rd 9.5-11.6 | 1.41 (0.96-2.06) | 1.08 (0.59-1.70) | |
| 4th 11.7-15.2 | 1.48 (1.03-2.13) | 1.16 (0.75-1.80) | |
| 5th ≥ 15.3 | 2.12 (1.47-3.05) | 0.99 (0.65-1.53) | 0.02 |
| per-quintile estimate | 1.16 (1.07-1.26) | 1.00 (0.91-1.10) | |
| rs10941679 | AA | AG/GG | |
| 1st <7.3 | 1 (ref.) | 1 (ref.) | |
| 2nd 7.3-9.4 | 1.13 (0.76-1.69) | 1.66 (1.09-2.53) | |
| 3rd 9.5-11.6 | 1.08 (0.71-1.62) | 1.66 (1.09-2.52) | |
| 4th 11.7-15.2 | 1.00 (0.68-1.46) | 2.08 (1.37-3.17) | |
| 5th ≥ 15.3 | 1.23 (0.84-1.80) | 1.96 (1.30-2.95) | 0.04 |
| per-quintile estimate | 1.03 (0.95-1.12) | 1.16 (1.06-1.28) |
Odds ratios are adjusted for age and state. Analyses exclude time for oral contraceptive use before first pregnancy.
OR indicates odds ratio; CI, confidence interval; ref., reference.
Discussion
In this study women with the longest reproductive windows were 1.5 to 1.7 times more likely to have breast cancer compared to women with the shortest reproductive windows, a result consistent with other studies (1–3,5). Standardized age at first birth may be associated with breast cancer risk as a proxy for breast tissue differentiation and ovarian hormone exposure. Breast tissue fully differentiates in preparation for lactation at first pregnancy and may be less influenced by the harmful effects of hormones in the fully differentiated state (4). Prior studies have not consistently accounted for the use of hormonal contraceptives before first childbirth in calculating the interval between age at menarche and first birth (6), which we were able to do here. Hormonal contraceptives reduce the number of ovulatory cycles experienced while also exposing the woman to exogenous estrogen and/or progestin which may alter the strength of the association between the reproductive window and breast cancer risk. In our study many of the study participants did not use or only used oral contraceptives for a short time before first pregnancy.
Under the reproductive lifespan hypothesis, menarche starts a window whereby estrogen exposure from the ovary can stimulate breast cancer until menopause (19) and, therefore, longer reproductive lifespan intervals are associated with increased breast cancer risk. The breast cancer risk associated with reproductive lifespan may become more substantial considering evidence that menarche and breast development are beginning at earlier ages than in previous birth cohorts (20). Findings from an U.S. Environmental Protection Agency 2008 scientific panel focused on menarche timing, concluded that increasingly women will be exposed to ovarian hormones for a larger proportion of their lifecourse possibly leading to increased breast cancer rates (20). Moreover, at least one study reported the age of menopause is slowly increasing over time (21).
We examined whether associations between standardized age at first birth and reproductive lifespan with breast cancer risk varied according to breast cancer histologic subtype. We found no evidence for heterogeneity overall nor did we observe a stronger association between standardized age at first birth and lobular breast cancer, as reported previously (22). Inconsistencies between our results and other published results may be related to limited power to detect interactions in our sample.
We considered whether associations between reproductive windows and breast cancer risk may be modified by previously identified breast cancer susceptibility variants. We observed no overall association for rs10483813 (14q24/RAD51B); however, the data were consistent with possible interaction by standardized age at first birth (P=0.02) in which a positive association with increasing quintile of standardized age at first birth was limited to homozygotes for the wild-type (TT) allele. rs10483813 is located in the intronic region of RAD51B, a gene involved in DNA repair. The effect of this SNP on breast cancer risk has been shown to be modified by estrogen-only postmenopausal hormone therapy in past studies (23) which lends plausibility that rs10483813 works through or in conjunction with hormonal risk factor pathways. It is also possible that rs10483813 is in linkage disequilibrium with another variant which is the true source of the association. Our results also suggest potential interaction between standardize age at first birth and rs10941679 (5p12) (P=0.04) in which the positive association was mainly observed in women with the variant (AG/GG) allele. rs10941679 has been frequently associated with breast cancer risk (7,24,25) and is in linkage disequilibrium with MRPS30, a gene implicated in apoptosis and estrogen-receptor positive breast cancer (26,27). The biologic mechanisms which could cause an interaction between rs10941679 and standardized age at first birth have not been identified. It also must be noted that tests for interaction had modestly significant p-values and given the number of tests conducted the results could have arisen by chance. Functional studies and replication of our analyses in other study samples may provide more information on potential interactions between breast cancer susceptibility loci and reproductive windows.
Our results underscore the importance of two reproductive windows in association with breast cancer risk, and provide evidence that associations are consistent regardless of breast cancer histologic subtype. Our data add further weight to the evidence that mechanisms of genetic susceptibility for breast cancer may operate at least in part through reproductive risk factors. Further research into this area may identify sub-populations of women at higher risk for breast cancer who would benefit from focused prevention efforts.
Acknowledgements
The authors would like to thank the study staff and participants for all of their individual contributions. This work was supported by grants from the National Institutes of Health (grant numbers R01CA47147, R01CA47305, R01CA69664,U10EY006594), the Department of Defense Breast Cancer Research Program (W81XWH-11-1-0047) and Intramural Research Funds of the National Cancer Institute, Department of Health and Human Services, USA.
Financial support: This work was supported by grants from the National Institutes of Health (grant numbers R01CA47147, R01CA47305, R01CA69664, U10EY006594) and the Department of Defense Breast Cancer Research Program (W81XWH-11-1-0047).
List of abbreviations
- AFB
age at first birth
- BMI
body mass index
- CI
confidence interval
- GWAS
genome-wide association study
- OR
odds ratio
- SNP
single nucleotide polymorphism
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
The authors have no conflicts of interest to disclose.
References
- 1.De Stavola BL, Wang DY, Allen DS, Giaconi J, Fentiman IS, Reed MJ, et al. The association of height, weight, menstrual and reproductive events with breast cancer: results from two prospective studies on the island of Guernsey (United Kingdom). Cancer Causes Control CCC. 1993 Jul;4(4):331–40. doi: 10.1007/BF00051335. [DOI] [PubMed] [Google Scholar]
- 2.Clavel-Chapelon F, E3N Group Cumulative number of menstrual cycles and breast cancer risk: results from the E3N cohort study of French women. Cancer Causes Control CCC. 2002 Nov;13(9):831–8. doi: 10.1023/a:1020684821837. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Andrieu N, Smith T, Duffy S, Zaridze DG, Renaud R, Rohan T, et al. The effects of interaction between familial and reproductive factors on breast cancer risk: a combined analysis of seven case-control studies. Br J Cancer. 1998 May;77(9):1525–36. doi: 10.1038/bjc.1998.251. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Russo J, Russo IH. Role of differentiation in the pathogenesis and prevention of breast cancer. Endocr Relat Cancer. 1997 Mar 20;4(1):7–21. [Google Scholar]
- 5.Rautalahti M, Albanes D, Virtamo J, Palmgren J, Haukka J, Heinonen OP. Lifetime menstrual activity--indicator of breast cancer risk. Eur J Epidemiol. 1993 Jan;9(1):17–25. doi: 10.1007/BF00463085. [DOI] [PubMed] [Google Scholar]
- 6.Whelan EA, Sandler DP, Root JL, Smith KR, Weinberg CR. Menstrual cycle patterns and risk of breast cancer. Am J Epidemiol. 1994 Dec 15;140(12):1081–90. doi: 10.1093/oxfordjournals.aje.a117208. [DOI] [PubMed] [Google Scholar]
- 7.Stacey SN, Manolescu A, Sulem P, Thorlacius S, Gudjonsson SA, Jonsson GF, et al. Common variants on chromosome 5p12 confer susceptibility to estrogen receptor-positive breast cancer. Nat Genet. 2008 Jun;40(6):703–6. doi: 10.1038/ng.131. [DOI] [PubMed] [Google Scholar]
- 8.Garcia-Closas M, Chanock S. Genetic susceptibility loci for breast cancer by estrogen receptor status. Clin Cancer Res Off J Am Assoc Cancer Res. 2008 Dec 15;14(24):8000–9. doi: 10.1158/1078-0432.CCR-08-0975. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Kawase T, Matsuo K, Suzuki T, Hiraki A, Watanabe M, Iwata H, et al. FGFR2 intronic polymorphisms interact with reproductive risk factors of breast cancer: results of a case control study in Japan. Int J Cancer J Int Cancer. 2009 Oct 15;125(8):1946–52. doi: 10.1002/ijc.24505. [DOI] [PubMed] [Google Scholar]
- 10.García-Closas M, Egan KM, Abruzzo J, Newcomb PA, Titus-Ernstoff L, Franklin T, et al. Collection of genomic DNA from adults in epidemiological studies by buccal cytobrush and mouthwash. Cancer Epidemiol Biomarkers Prev Publ Am Assoc Cancer Res Cosponsored Am Soc Prev Oncol. 2001 Jun;10(6):687–96. [PubMed] [Google Scholar]
- 11.García-Closas M, Egan KM, Newcomb PA, Brinton LA, Titus-Ernstoff L, Chanock S, et al. Polymorphisms in DNA double-strand break repair genes and risk of breast cancer: two population-based studies in USA and Poland, and meta-analyses. Hum Genet. 2006 May;119(4):376–88. doi: 10.1007/s00439-006-0135-z. [DOI] [PubMed] [Google Scholar]
- 12.Sprague BL, Trentham-Dietz A, Garcia-Closas M, Newcomb PA, Titus-Ernstoff L, Hampton JM, et al. Genetic variation in TP53 and risk of breast cancer in a population-based case control study. Carcinogenesis. 2007 Aug;28(8):1680–6. doi: 10.1093/carcin/bgm097. [DOI] [PubMed] [Google Scholar]
- 13.Percy C, Holten V van, Muir CS, Organization WH. [2013 Jul 24];International classification of diseases for oncology / editors, Constance Percy, Valerie Van Holten, Calum Muir [Internet] 1990 Available from: http://apps.who.int/iris/handle/10665/39441.
- 14.Easton DF, Pooley KA, Dunning AM, Pharoah PDP, Thompson D, Ballinger DG, et al. Genome-wide association study identifies novel breast cancer susceptibility loci. Nature. 2007 Jun 28;447(7148):1087–93. doi: 10.1038/nature05887. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Thomas G, Jacobs KB, Kraft P, Yeager M, Wacholder S, Cox DG, et al. A multistage genome-wide association study in breast cancer identifies two new risk alleles at 1p11.2 and 14q24.1 (RAD51L1). Nat Genet. 2009 May;41(5):579–84. doi: 10.1038/ng.353. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Ahmed S, Thomas G, Ghoussaini M, Healey CS, Humphreys MK, Platte R, et al. Newly discovered breast cancer susceptibility loci on 3p24 and 17q23.2. Nat Genet. 2009 May;41(5):585–90. doi: 10.1038/ng.354. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Milne RL, Gaudet MM, Spurdle AB, Fasching PA, Couch FJ, Benítez J, et al. Assessing interactions between the associations of common genetic susceptibility variants, reproductive history and body mass index with breast cancer risk in the breast cancer association consortium: a combined case-control study. Breast Cancer Res BCR. 2010;12(6):R110. doi: 10.1186/bcr2797. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Hosmer DW. Applied logistic regression. 2nd ed. Wiley; New York: 2000. [Google Scholar]
- 19.Hiatt RA, Haslam SZ, Osuch J. Breast Cancer and the Environment Research Centers. The breast cancer and the environment research centers: transdisciplinary research on the role of the environment in breast cancer etiology. Environ Health Perspect. 2009 Dec;117(12):1814–22. doi: 10.1289/ehp.0800120. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Euling SY, Herman-Giddens ME, Lee PA, Selevan SG, Juul A, Sørensen TIA, et al. Examination of US puberty-timing data from 1940 to 1994 for secular trends: panel findings. Pediatrics. 2008 Feb;121(Suppl 3):S172–191. doi: 10.1542/peds.2007-1813D. [DOI] [PubMed] [Google Scholar]
- 21.Nichols HB, Trentham-Dietz A, Hampton JM, Titus-Ernstoff L, Egan KM, Willett WC, et al. From menarche to menopause: trends among US Women born from 1912 to 1969. Am J Epidemiol. 2006 Nov 15;164(10):1003–11. doi: 10.1093/aje/kwj282. [DOI] [PubMed] [Google Scholar]
- 22.Li CI, Malone KE, Daling JR, Potter JD, Bernstein L, Marchbanks PA, et al. Timing of menarche and first full-term birth in relation to breast cancer risk. Am J Epidemiol. 2008 Jan 15;167(2):230–9. doi: 10.1093/aje/kwm271. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Nickels S, Truong T, Hein R, Stevens K, Buck K, Behrens S, et al. Evidence of gene-environment interactions between common breast cancer susceptibility loci and established environmental risk factors. PLoS Genet. 2013;9(3):e1003284. doi: 10.1371/journal.pgen.1003284. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Campa D, Kaaks R, Le Marchand L, Haiman CA, Travis RC, Berg CD, et al. Interactions between genetic variants and breast cancer risk factors in the breast and prostate cancer cohort consortium. J Natl Cancer Inst. 2011 Aug 17;103(16):1252–63. doi: 10.1093/jnci/djr265. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Milne RL, Goode EL, García-Closas M, Couch FJ, Severi G, Hein R, et al. Confirmation of 5p12 as a susceptibility locus for progesterone-receptor-positive, lower grade breast cancer. Cancer Epidemiol Biomarkers Prev Publ Am Assoc Cancer Res Cosponsored Am Soc Prev Oncol. 2011 Oct;20(10):2222–31. doi: 10.1158/1055-9965.EPI-11-0569. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Liu X, Qin Z, Shen H, Xue J, Jiang Y, Hu Z, et al. Genetic variants at 5p12 and risk of breast cancer in Han Chinese. J Hum Genet. 2012 Oct;57(10):638–41. doi: 10.1038/jhg.2012.83. [DOI] [PubMed] [Google Scholar]
- 27.Van 't Veer LJ, Dai H, van de Vijver MJ, He YD, Hart AAM, Mao M, et al. Gene expression profiling predicts clinical outcome of breast cancer. Nature. 2002 Jan 31;415(6871):530–6. doi: 10.1038/415530a. [DOI] [PubMed] [Google Scholar]