Abstract
Purpose
In the United States hysterectomies and oophorectomies are frequently performed before menopause for benign conditions. The procedures are associated with reduced breast cancer-specific mortality among White women. The relationship between premenopausal gynecologic surgery and mortality in Black women with breast cancer is unknown.
Methods
This investigation used incident invasive cases of breast cancer from Phases 1 and 2 of the Carolina Breast Cancer Study a population-based study that recruited Black and White women in North Carolina between 1993 and 2001. Premenopausal gynecologic surgery was operationalized in three categories: no surgery; hysterectomy with bilateral oophorectomy; hysterectomy with conservation of ≥ 1 ovary. Mortality was ascertained using the National Death Index, last updated in 2016. Multivariable-adjusted Cox Proportional Hazard Models were used to estimate the effect of premenopausal surgery on breast cancer-specific and all-cause mortality.
Results
Hysterectomy with bilateral oophorectomy was associated with reduced breast cancer-specific mortality (HR: 0.68; 95% CI: 0.49, 0.96). White and Black women had a similar reduction in breast cancer-specific mortality. (HR among white: 0.66; 95% CI: 0.43,1.02), (HR among Black: 0.67; 95% CI: 0.37,1.21).
Conclusions
There was a similar reduction in breast cancer specific mortality following premenopausal, pre-diagnosis hysterectomy with bilateral oophorectomy across both Black and White women.
Keywords: Breast Cancer, Disparities, Hysterectomy, Oophorectomy, Survival Analysis
Introduction
Breast cancer survivors represent a growing segment of the population, with over 3 million residing in the United States as of 2014 [1]. This number is expected to grow as treatments improve and the population ages [2]. On the aggregate, survival after breast cancer diagnosis has improved over the last three decades with overall 5-year survival being greater than 95% for women diagnosed with localized disease [3]. However, Black women with breast cancer experience higher mortality and shorter survival time than White women [4–7]. Further, there is evidence that racial disparities in outcomes have also increased over time [8–10]. It is important to describe contributors to persistent racial disparities in outcomes for women with breast cancer and to identify long term prognostic indicators based on common risk factors.
Gynecologic surgeries such as hysterectomy and oophorectomy are most commonly performed in premenopausal women for non-malignant diseases such as endometriosis, uterine fibroids, and ovarian cysts [11]. Additionally, a small subset of gynecologic surgeries are performed in women who have first-degree relatives with a history of breast or ovarian cancer or who have germline mutations such as BRCA1/2; these women may undergo prophylactic hysterectomy and oophorectomy, for the primary prevention of breast and ovarian cancers [12]. Recent evidence from the Centers for Disease Control and Prevention Behavioral Risk Factor Surveillance System, a population-based survey of U.S. adults, indicated that 33% of Black women and 23% of white women aged 48–50 had a history of hysterectomy [13]. In nearly half of all hysterectomies bilateral oophorectomy, or removal of both ovaries is performed [14]. Additional studies have supported the observation that hysterectomy prevalence is markedly higher in Black populations compared to White populations, especially for surgeries occurring at younger ages in the premenopausal age range [15–16].
Prior studies, including one conducted in the Carolina Breast Cancer Study (CBCS), a weighted population based sample, have described a reduction in breast cancer incidence for women who had hysterectomy with bilateral oophorectomy compared to women who had not had gynecologic surgery [17–18]. Further, existing literature demonstrates that history of pre-diagnosis premenopausal hysterectomy with bilateral oophorectomy reduces the risk of breast cancer mortality for White women with breast cancer [19–20]. These findings provide evidence that there is a biological pathway mediated by pre-diagnosis premenopausal removal of the ovaries that confers some protection against both breast cancer incidence and mortality. Understanding the role of premenopausal gynecologic surgery on survival in diverse populations of women with breast cancer could impact prognosis and racial disparities in outcomes.
The objective of the study was to evaluate effect measure modification by race between premenopausal pre-diagnosis gynecologic surgeries, including hysterectomy and oophorectomy, and breast cancer-specific mortality in a population-based sample of Black and White women with breast cancer living in the U.S. South. A secondary aim was to assess the relationship between gynecologic surgery and all-cause mortality as well as assess potential effect measure modification for both breast cancer-specific and all-cause mortality by estrogen receptor status, and a woman’s family history of breast cancer.
Materials and Methods
Study Population
This study utilized data from women diagnosed with breast cancer who participated in Phases 1 and 2 of the Carolina Breast Cancer Study (CBCS), a weighted population-based case-control study sampling women ages 20–74 from 24 counties in North Carolina. This study was approved by the institutional review board of the University of North Carolina at Chapel Hill. Invasive breast cancer cases were identified from 1993 until 2001 using the Rapid Case Ascertainment System from the North Carolina Central Cancer Registry. CBCS oversampled Black women, and in particular, Black women under age 50 at diagnosis using a modification of randomized recruitment to obtain approximately equal numbers of Black and non-Black women [21–22]. Of all of the potentially eligible cases, 97.6% were contacted and 78.0% of contacted cases were enrolled resulting in 1,808 women with invasive cases of breast cancer being enrolled in CBCS. From this sampling scheme inverse probability of sampling weights were created to reflect the underlying source population in North Carolina [23]. Additional details on the sampling method employed in the Carolina Breast Cancer Study have been described elsewhere [24]. Cases were interviewed by trained nurse interviewers to collect self-reported information on reproductive history, family history of cancer, environmental exposures and sociodemographic characteristics. For 94.5% of women included in this study, the baseline interview was conducted within a year of their diagnosis of breast cancer.
Confirmed invasive cases of breast cancer from Phases 1 and 2 of CBCS were eligible to be included in this analysis (n=1,808). We excluded women if they identified as a race other than Black/African-American or White (n=25), if they had unknown gynecologic surgery status (n=24), or if they had radiation or chemotherapy that resulted in the cessation of menstruation (n=27). Lastly, women were excluded if they reported having a hysterectomy and also receiving estrogen and progestin combination hormone therapy because this formulation is currently contraindicated for women who have had a hysterectomy (n=9). The final analytic sample for this study included 1,723 women.
Exposure Assessment
In this study, the exposure was a 3-level indicator variable for premenopausal gynecologic surgery that occurred prior to breast cancer diagnosis and included the categories of no gynecologic surgery, hysterectomy with bilateral oophorectomy, and hysterectomy with ovarian conservation. In study interviews, women were asked: “Are you still having menstrual periods?” If participants responded “yes” or were currently or recently pregnant they were considered premenopausal. If they responded “no” to that survey item and were also not currently or recently pregnant, they were asked “Did your periods stop because of an operation (removal of the uterus or ovaries)?” If women responded no to that item, they were categorized post-menopausal due to non-surgical causes. This was the referent level of the exposure variable. If a woman did indicate the cessation of menstruation due to surgery, she was then asked, “Was your uterus removed?” and “were one or both ovaries removed?” Women who had their uterus and both ovaries removed were categorized as having hysterectomy with bilateral oophorectomy, and women who had their uterus removed with the conservation of one or both ovaries were categorized as having hysterectomy with ovarian conservation (Figure 1). Underlying indication for gynecologic surgery was not available.
Figure 1.
Displaying how women’s surgical status was categorized
Outcome Assessment
Breast cancer-specific and all-cause mortality were the primary outcomes. Year and cause of death were determined by linkage with the National Death Index through December 31, 2016. Deaths from breast cancer were originally identified using International Classification of Disease (ICD) -9 diagnostic code 174.9. Subsequently these events were reclassified using ICD-10 code C50.9. All participants who did not have an event by December 31st 2016, were administratively censored on that date. For the breast cancer-specific mortality analyses, participants who died from causes other than breast cancer were censored. All deaths, including those from breast cancer, were included in all-cause mortality analyses.
Covariate Assessment
Models were adjusted for probable confounders of the relationship between gynecologic surgery and mortality in women with breast cancer. These confounders were determined by an assessment of previous literature in this area and categorizations were also informed by prior literature as well as CBCS convention. Body mass index (BMI: kg/m2) defined continuously as a woman’s self-reported weight 1 year prior to her breast cancer diagnosis and her measured height at the time of her interview. Parity was a composite measure to account for age and number of full-term pregnancy. Women were categorized as follows: nulliparous; one child, age at first full-term pregnancy (FTP) < 26; one child, age at FTP ≥26; 2 or more children, age at FTP< 26; 2 or more children, age at FTP ≥26. Breastfeeding was binary, having breastfed or not. Family history of breast cancer was defined as a woman self-reporting having one or more first degree relatives with a history of breast cancer. Education was a three-level variable: less than high school (referent); high school diploma and some college experience; and completion of a college degree and above. Alcohol use was a binary variable where women were categorized as never users if they had fewer than 12 lifetime drinks, and alcohol consumers if they had more than 12 lifetime drinks. Smoking history was operationalized as never smoker, former smoker or current smoked. Age at menarche was defined as a 3-level variable based on self-reported start of menses: 11 years or younger, 12 to 13 years, and 14 years or older. Women reported if they had ever used menopausal hormone therapy (MHT), and if they did, whether they used estrogen only, combination therapy, or another formulation. Self-reported race (Black/African-American or White) was also recorded.
Statistical Analysis
Multivariable Cox proportional hazard models were used to estimate hazard ratios and 95% confidence intervals [25]. Follow-up began at the date of breast cancer diagnosis and time was measured continuously in years. We conducted two main effects models, a model for breast cancer-specific mortality and another model for all-cause mortality. All models were weighted to reflect the sampling design [22]. We conducted a Wald test for interaction (α=0.2) to evaluate the joint effect of surgery status and MHT formulation based on evidence from prior literature [19]. A more generous alpha level of 0.2was used to compensate for the low statistical power of tests for interaction [26–27]. Two- sided tests (α=0.05) were conducted for both the main effects and stratified models. Analyses were conducted in SAS version 9.4 (Cary, NC).
First we estimated the relationship between gynecologic surgery and breast cancer specific and all-cause mortality in the total population and then stratified by race. We then also evaluated effect measure modification between gynecologic surgery and breast cancer-specific and all-cause mortality by creating models stratified by estrogen receptor positivity and first-degree family history of breast cancer. Having one or more first-degree relatives with a history of breast cancer is a known factor for increased risk of breast cancer-specific mortality as well as being an indication for prophylactic gynecologic surgery [28–29]. Since the women in our study sample were diagnosed with breast cancer before the widespread adoption BRCA 1/2 testing, we consider self-reported first- degree history of breast cancer to be a relevant risk factor.
Results
Covariate characteristics were calculated using weighted frequencies to reflect the prevalence in the underlying source population in North Carolina (Table 1). These frequencies were weighted by the Carolina Breast Cancer Study sampling fractions to reflect the demographic distribution of women with breast cancer in the study region [21]. Age at breast cancer diagnosis ranged from 23 to 74 years with a median age of 49. Overall, 759 (21.2%) of women identified as Black. Of the 1,723 women with breast cancer, 295 women had a history of premenopausal hysterectomy with ovarian conservation, 152 had hysterectomy with bilateral oophorectomy, and 1,276 (72.6%) had no surgery. During the follow-up period, 793 women died; 447 of these deaths were breast cancer-specific.
Table 1-.
Descriptive characteristics of women diagnosed with breast cancer in the Carolina Breast Cancer study, phases 1 and 2, 1993–2001
| All | No Surgery | Hysterectomy with Ovarian Conservation | Hysterectomy with Bilateral Oophorectomy | |||||
|---|---|---|---|---|---|---|---|---|
| Count | Weighteda% | Count | Weighteda% | Count | Weighteda% | Count | Weighteda % | |
| TOTAL N | 1,723 | 1,276 | 72.6 | 295 | 18.0 | 152 | 9.4 | |
| Person Years | 20,459 | 15,103 | 3,454 | 1,902 | ||||
| Median Age at Diagnosis | 49 | 47 | 53 | 57 | ||||
| Race | ||||||||
| White | 964 | 78.8 | 748 | 79.9 | 148 | 76.9 | 68 | 74.2 |
| Black | 759 | 21.2 | 528 | 20.1 | 147 | 23.1 | 84 | 25.8 |
| Education | ||||||||
| <High School degree | 317 | 16.1 | 207 | 13.8 | 77 | 25.4 | 33 | 16.1 |
| High School and Some College | 927 | 54.8 | 664 | 52.6 | 174 | 59.1 | 89 | 64.4 |
| College Degree or more | 479 | 29.1 | 405 | 33.8 | 44 | 15.5 | 30 | 19.5 |
| Median BMI (kg/m2) | 27.0 | 26.6 | 27.9 | 27.5 | ||||
| Age at Menarche (years) | ||||||||
| ≤11 | 391 | 20.5 | 270 | 18.6 | 82 | 27.3 | 39 | 22.0 |
| 12–13 | 951 | 56.8 | 724 | 58.4 | 147 | 52.4 | 80 | 53.1 |
| >13 | 379 | 22.6 | 282 | 23.0 | 66 | 20.3 | 31 | 24.0 |
| Missing | 2 | 0.1 | 0 | 0 | 2 | 0.9 | ||
| Receptor Subtype | ||||||||
| ER Positive | 915 | 59.6 | 672 | 59.4 | 162 | 60.7 | 81 | 59.5 |
| ER Negative | 692 | 33.9 | 521 | 34.3 | 113 | 32.9 | 58 | 32.9 |
| Missing | 116 | 6.5 | 83 | 6.3 | 20 | 6.4 | 13 | 7.6 |
| Alcohol Use | ||||||||
| Ever | 1183 | 67.3 | 890 | 68.1 | 187 | 63.8 | 106 | 68.4 |
| Never | 539 | 32.5 | 385 | 31.7 | 108 | 36.2 | 46 | 31.6 |
| Missing | 1 | 0.1 | 1 | 0.2 | 0 | 0 | ||
| Smoking | ||||||||
| Never | 901 | 49 | 692 | 51 | 137 | 43.9 | 72 | 43.8 |
| Former | 446 | 30.1 | 326 | 30.3 | 75 | 28.3 | 45 | 31.7 |
| Current | 376 | 20.9 | 258 | 18.7 | 83 | 27.8 | 35 | 24.5 |
| Menopausal Hormone Therapy | ||||||||
| Never User | 1,256 | 61.3 | 1,040 | 70.8 | 183 | 47.5 | 33 | 14.9 |
| Estrogen Only | 249 | 18.7 | 41 | 4.1 | 102 | 48.1 | 106 | 75.1 |
| Estrogen and Progestin Combination | 134 | 13.2 | 134 | 18.2 | 0 | 0 | 0 | 0 |
| Other Formulations | 84 | 2.28 | 61 | 2.33 | 10 | 1.48 | 13 | 3.34 |
| Family History | ||||||||
| Positive Family History of Breast Cancer | 283 | 17.3 | 195 | 15.9 | 64 | 23.4 | 24 | 16.6 |
| No Family History of Breast Cancer | 1390 | 79.9 | 1046 | 81.5 | 221 | 73.1 | 123 | 80.5 |
| Missing | 50 | 2.8 | 35 | 2.6 | 10 | 3.5 | 5 | 2.9 |
Weights were constructed to account for the non-random sampling of cases in the CBCS study population of central North Carolina.
In our data, the interaction term for gynecologic surgery and MHT use did not meet statistical significance and consequently, all models in this analysis include surgery status and hormone replacement therapy use as covariates without an interaction term. In the total study population, pre-diagnosis premenopausal hysterectomy with bilateral oophorectomy was associated with reduced breast cancer-specific mortality (HR: 0.68; 95% CI: 0.49, 0.96) in multivariable-adjusted models (Table 2). This reduction in mortality was consistent across racial groups with White and Black women having a similar reduction in breast cancer-specific mortality (HR among white: 0.66; 95% CI: 0.42,1.02), (HR among Black: 0.67; 95% CI: 0.37,1.21), (Figure 1). For hysterectomy with ovarian conservation, there was no association with breast cancer specific mortality among White or Black women.
Table 2.
Crude and Multivariable Adjusted Breast Cancer-Specific Mortality Hazard Ratios by Race, Estrogen Receptor Status, and Family History Status
| No Surgery | Hysterectomy with Ovarian Conservation | Hysterectomy with Bilateral Oophorectomy | |||||
|---|---|---|---|---|---|---|---|
| Person-Years | Deaths | HR | Crude HR | Adjusted HR | Crude HR | Adjusted HR | |
| Combined Populationa | 25,242 | 447 | 1 | 0.91 (0.76,1.10) | 0.90 (0.72,1.12) | 0.64 (0.48,0.84) | 0.68 (0.49,0.96) |
| Race b | |||||||
| White | 15,153 | 208 | 1 | 0.90 (0.72,1.13) | 0.91 (0.68,1.20) | 0.61 (0.42,0.87) | 0.66 (0.43,1.02) |
| Black | 10,089 | 239 | 1 | 0.84 (0.61,1.16) | 0.84 (0.59,1.20) | 0.60 (0.38,0.95) | 0.67 (0.37,1.21) |
| Estrogen Receptor Statusa | |||||||
| ER+ | 13,893 | 190 | 1 | 0.99 (0.78,1.26) | 1.01 (0.75,1.36) | 0.81 (0.57,1.16) | 0.82 (0.53,1.27) |
| ER− | 9,732 | 196 | 1 | 0.88 (0.65,1.20) | 0.90 (0.64,1.27) | 0.46 (0.27,0.79) | 0.59 (0.32,1.12) |
| Family History c | |||||||
| Positive Family History | 4,143 | 70 | 1 | 1.08 (0.74,1.58) | 0.77 (0.45,1.29) | 0.17 (0.05,0.58) | 0.11 (0.03,0.36) |
| No Family History | 20,369 | 368 | 1 | 0.90 (0.73,1.12) | 0.88 (0.69,1.12) | 0.73 (0.54,1.98) | 0.90 (0.63,1.29) |
Adjusted for Race, Body Mass Index, education level, age at menarche, parity, breastfeeding history, family history of breast cancer, alcohol use, smoking status, menopausal hormone therapy use and formulation.
Adjusted for Body Mass Index, education level, age at menarche, parity, breastfeeding history, family history of breast cancer, alcohol use, smoking status, menopausal hormone therapy use and formulation.
Adjusted for Race, Body Mass Index, education level, age at menarche, parity, breastfeeding history, alcohol use, smoking status, menopausal hormone therapy use and formulation.
For all-cause mortality in the total population, there was no association between hysterectomy with bilateral oophorectomy and all-cause mortality (HR: 0.95; 95% CI 0.79,1.15) (Table 3). Additionally, no association was observed between hysterectomy with ovarian conservation and all-cause mortality (HR: 0.94, 95% CI 0.82, 1.08). However, when stratified by race, associations appeared to differ slightly for White and Black women for all-cause mortality. Among White women, there was no association between either hysterectomy with bilateral oophorectomy (HR: 1.14; 95% CI 0.92,1.41) or hysterectomy with ovarian conservation (HR: 1.11; 95% CI 0.91, 1.21) and all-cause mortality. In contrast to the results among White women, among Black women, a reduction in all-cause mortality was observed for both levels of gynecologic surgery. Black women who had hysterectomy with bilateral oophorectomy had a 32% (HR: 0.68: 95% CI 0.47, 1.00) reduced risk of mortality compared to Black women who had not had gynecologic surgery and Black women who had a hysterectomy with ovarian conservation had a 23% (HR: 0.77; 95% CI 0.59, 0.99) reduced risk of mortality compared to Black women who had not had gynecologic surgery.
Table 3.
Crude and Multivariable Adjusted All-Cause Mortality Hazard Ratios by Race, Estrogen Receptor Status, and Family History Status
| No Surgery | Hysterectomy with Ovarian Conservation | Hysterectomy with Bilateral Oophorectomy | |||||
|---|---|---|---|---|---|---|---|
| Person-Years | Deaths | HR | Crude HR | Adjusted HR | Crude HR | Adjusted HR | |
| Combined Populationa | 25,242 | 793 | 1 | 1.10 (1.98,1.23) | 0.94 (0.82,1.08) | 1.13 (0.97,1.31) | 0.95 (0.79,1.15) |
| Raceb | |||||||
| White | 15,153 | 412 | 1 | 1.15 (1.00,1.31) | 1.11 (0.91,1.26) | 1.22(1.03,1.45) | 1.14(0.92,1.41) |
| Black | 10,089 | 424 | 1 | 0.92 (0.73,1.16) | 0.77 (0.59,0.99) | 0.83 (0.61,1.12) | 0.68 (0.47,1.00) |
| Estrogen Receptor Statusa | |||||||
| ER+ | 13,893 | 457 | 1 | 0.98 (0.85,1.14) | 0.77 (0.64,0.93) | 1.31 (1.09,1.57) | 0.99 (0.79,1.25) |
| ER− | 9,732 | 314 | 1 | 1.27 (1.04,1.56) | 1.16 (0.91,1.46) | 1.08 (0.82,1.43) | 1.07 (0.75,1.53) |
| Family Historyc | |||||||
| Positive Family History | 4,143 | 136 | 1 | 1.28 (1.01,1.62) | 0.83 (0.61,1.14) | 0.60 (0.38,0.94) | 0.39 (0.23,0.65) |
| No Family History | 20,369 | 674 | 1 | 1.10 (0.96,1.26) | 0.95 (0.82,1.11) | 1.24 (1.06,1.46) | 1.10 (0.90,1.34) |
Adjusted for Race, Body Mass Index, education level, age at menarche, parity, breastfeeding history, family history of breast cancer, alcohol use, smoking status, menopausal hormone therapy use and formulation.
Adjusted for Body Mass Index, education level, age at menarche, parity, breastfeeding history, family history of breast cancer, alcohol use, smoking status, menopausal hormone therapy use and formulation.
Adjusted for Race, Body Mass Index, education level, age at menarche, parity, breastfeeding history, alcohol use, smoking status, menopausal hormone therapy use and formulation.
There was no evidence of effect measure modification of the association between hysterectomy with ovarian conservation and breast cancer specific mortality according to Estrogen Receptor (ER) status (Table 2). There was evidence of effect measure modification by first-degree family history. Among women without a family history of breast cancer who had hysterectomy with bilateral oophorectomy there was instability in the estimated hazard ratio (HR: 0.90; 95% CI 0.63, 1.29) (Table 2) for breast cancer specific mortality. However, among women with a family history of breast cancer, there was a substantial reduction in both breast cancer-specific and all-cause mortality among women with hysterectomy with bilateral oophorectomy compared to women who did not have surgery (Tables 2 and 3).
Discussion
In this weighted population-based study, a reduction in breast cancer-specific mortality was seen in women who had a hysterectomy with bilateral oophorectomy and was observed consistently in both Black and White self-defined racial groups. This result was robust even after controlling for several confounders, contributing to the body of literature that supports including premenopausal pre-diagnosis gynecologic surgery as a prognostic factor for outcomes among women with breast cancer. Given the high prevalence of hysterectomy with bilateral oophorectomy for common gynecologic conditions among Black women compared to White women, it is possible that this exposure plays a role in population level patterns of breast cancer survival [13]. Within a population-based study of privately insured women, hysterectomy rates among women ages 18 to 64 decreased 39% between 2000 and 2014 and oophorectomy rates decreased 19% during the same timeframe [30]. This population-level change in the exposure prevalence of a protective factor that has a greater prevalence in Black women may be consequential for racial disparities in breast cancer outcomes in future years.
In addition to conferring protection for breast cancer specific mortality, a prior study found a reduction in all-cause mortality in women with breast cancer who had pre-diagnosis hysterectomy with bilateral oophorectomy, only for women who had survived at least five years [20]. In our study using a weighted population based sample, for all-cause mortality, there was evidence of a protective effect for both levels of gynecologic surgery for Black women but not for White women. Given the protective effect observed in Black women but not in White women for all-cause mortality and different distributions of causes of death, there may exist additional pathways for other causes of death, however we were unable to directly address this in our data.
In contrast to previous literature which primarily focused on the effect of hysterectomy with bilateral oophorectomy, the effect of hysterectomy with ovarian conservation on mortality in women with breast cancer had not been thoroughly investigated [20]. Previous studies conducted in Australia demonstrated a reduction in breast cancer-specific mortality for women who have had their uteri removed but have one or both ovaries intact but demonstrated no effect on all-cause mortality [31–32]. Our study found no evidence of an association between hysterectomy with ovarian conservation for breast cancer specific mortality.
Our analysis revealed evidence of potential effect measure modification by first-degree family history of breast cancer. When conducting the analysis only amongst women who did not have a family history of breast cancer, the hazard ratio for breast cancer-specific mortality for women with hysterectomy with bilateral oophorectomy compared to women who had not had surgery became null. The instability of the estimate once women with family history are removed may suggest that the overall protective relationship observed between hysterectomy and bilateral oophorectomy and breast cancer-specific mortality could be driven by the small subset of women with positive family history who benefit the most, due to having a higher baseline risk of breast cancer mortality. Previous literature has demonstrated a reduction in breast cancer specific and all-cause mortality for BRCA1/2 mutation carriers who had a prophylactic oophorectomy [33–34]. We consider our finding in line with this prior research and applicable to women who made decisions about gynecologic surgery before the advent of genetic testing, who may have only had family history information to rely on to guide the decision-making process. Our results reinforce existing guidelines that only recommend prophylactic procedures in women with germline mutations that put them at increased risk for breast and ovarian cancers [12].
The study population used for this investigation allowed for the evaluation of both pre-diagnosis premenopausal hysterectomy with ovarian conversation as well as hysterectomy with bilateral oophorectomy on survival among women with breast cancer. Additionally, lengthy follow-up permitted the evaluation of long term outcomes for women with breast cancer with the median survival time being 18 years. Further, this study was conducted in a population-based sample of women with breast cancer that included a substantial number of Black women and women diagnosed with breast cancer before the age of 50. The depth of the covariate data in CBCS allowed for the control of many important potential confounders of the relationship between gynecologic surgery and survival in women with breast cancer. Some limitations should also be considered. Despite the population-based sampling, our study had low power to assess additional causes of death. A prior study found increased risk of all-cause mortality among obese women older than age 40 who had hysterectomy with bilateral oophorectomy but did not observe this relationship in younger or non-obese women [35]. Given small cell sizes among women with hysterectomy with bilateral oophorectomy we were unable to evaluate this relationship in our study. Additionally, ER status had approximately 10% missing data which may have influenced those stratified analyses. Since the exposure was self-reported gynecologic surgery history, there may exist a degree of exposure misclassification. Women who reported only having one ovary removed may have actually had both and we were unable to clinically verify [36]. However, misclassification in this direction would likely make the effect of hysterectomy with bilateral oophorectomy appear more conservative. Further, we were not able to evaluate reason for gynecologic surgery, whether it was a prophylactic procedure or was done to treat an underlying benign gynecologic condition. Finally, we did not account for differences in treatment received for breast cancer, which is also a predictor of survival.
In summary, we extended previous work showing that premenopausal hysterectomy with bilateral oophorectomy is a prognostic factor for breast cancer specific-mortality and confirms the existence of this relationship among Black women. This confirmation is important given that Black women, particularly those residing in the US South historically have had higher rates of hysterectomy and higher rates of breast-cancer specific mortality than White women. A comprehensive understanding of cancer survivors’ risk profiles, including the interplay between family history, estrogen receptor status, and gynecologic surgery status can help inform clinical decision making.
Acknowledgments
The authors would like to thank Chiu-Kit Tse for her assistance with data management.
Funding Statement: The Carolina Breast Cancer Study is supported by a Breast Specialized Program of Research Excellence (SPORE) grant (P50-CA58223 Olshan & Troester Co-PIs) and by the North Carolina State University Cancer Research Fund. This project was supported by the Robert Wood Johnson Foundation Health Policy Research Scholars Program (Roberson). This project was also supported by a National Cancer Institute Career Development Award (Robinson: NCI K01CA172717) and a National Institute on Minority Health and Health Disparities research award (Robinson: NIMHD R01MD011680).
Footnotes
Conflicts of Interests Statement: The authors of this manuscript have no conflicts to disclose.
Publisher's Disclaimer: This Author Accepted Manuscript is a PDF file of an unedited peer-reviewed manuscript that has been accepted for publication but has not been copyedited or corrected. The official version of record that is published in the journal is kept up to date and so may therefore differ from this version.
References
- [1].Howlader N, Noone AM, Krapcho M, et al. eds. SEER Cancer Statistics Review, 1975–2014 (seer.cancer.gov/csr/1975_2014/), based on November 2016 SEER data submission, posted to the SEER web site, April 2017. Bethesda, MD: National Cancer Institute; 2017. [Google Scholar]
- [2].DeSantis CE, Ma J, Goding Sauer A, Newman LA, Jemal A. Breast cancer statistics, 2017,racial disparity in mortality by state. CA Cancer J Clin. 2017. doi: 10.3322/caac.21412 [DOI] [PubMed] [Google Scholar]
- [3].Daly B, Olopade OI. A perfect storm: How tumor biology, genomics, and health care delivery patterns collide to create a racial survival disparity in breast cancer and proposed interventions for change. CA Cancer J Clin. 2015;65(3):221–238. doi: 10.3322/caac.21271 [DOI] [PubMed] [Google Scholar]
- [4].Carey LA, Perou CM, Livasy CA, et al. Race, Breast Cancer Subtypes, and Survival in the Carolina Breast Cancer Study. JAMA. 2006;295(21):2492–2502. doi: 10.1001/jama.295.21.2492 [DOI] [PubMed] [Google Scholar]
- [5].DeSantis CE, Siegel L, Goding SA, et al. Cancer statistics for African Americans, 2016: Progress and opportunities in reducing racial disparities. CA Cancer J Clin. 2016;66(4):290–308. doi: 10.3322/caac.21340 [DOI] [PubMed] [Google Scholar]
- [6].Reeder-Hayes KE, Anderson BO. Breast Cancer Disparities at Home and Abroad: A Review of the Challenges and Opportunities for System-Level Change. Clin Cancer Res. 2017;23(11):2655–2664. doi: 10.1158/1078-0432.CCR-16-2630 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].Toriola AT, Colditz GA. Trends in breast cancer incidence and mortality in the United States: implications for prevention. Breast Cancer Res Treat. 2013;138(3):665–673. doi: 10.1007/s10549-013-2500-7 [DOI] [PubMed] [Google Scholar]
- [8].Hunt BR, Whitman S, Hurlbert MS. Increasing Black:White disparities in breast cancer mortality in the 50 largest cities in the United States. Cancer Epidemiol. 2014;38(2):118–123. doi: 10.1016/j.canep.2013.09.009 [DOI] [PubMed] [Google Scholar]
- [9].Silber JH, Rosenbaum PR, Clark AS, et al. Characteristics Associated With Differences in Survival Among Black and White Women With Breast Cancer. JAMA. 2013;310(4):389–397. doi: 10.1001/jama.2013.8272 [DOI] [PubMed] [Google Scholar]
- [10].Siegel RL, Miller KD, Jemal A (2016) Cancer statistics. CA Cancer J Clin 66(1):7–30. 10.3322/caac.21332 [DOI] [PubMed] [Google Scholar]
- [11].Jacoby VL, Fujimoto VY, Giudice LC, Kuppermann M, Washington AE. Racial and ethnic disparities in benign gynecologic conditions and associated surgeries. Am J Obstet Gynecol. 2010;202(6):514–521. doi: 10.1016/j.ajog.2010.02.039 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].Hartmann LC, Lindor NM. The Role of Risk-Reducing Surgery in Hereditary Breast and Ovarian Cancer. Longo DL, ed. N Engl J Med. 2016;374(5):454–468. doi: 10.1056/NEJMra1503523 [DOI] [PubMed] [Google Scholar]
- [13].Centers for Disease Control and Prevention (CDC). Behavioral risk factors surveillance system survey data. Atlanta: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention; 2012 [Google Scholar]
- [14].Jacoby VL, Vittinghoff E, Nakagawa S, et al. Factors associated with undergoing bilateral salpingo-oophorectomy at the time of hysterectomy for benign conditions. Obstet Gynecol. 2009;113(6):1259–1267. doi: 10.1097/AOG.0b013e3181a66c42 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].Powell LH, Meyer P, Weiss G, et al. Ethnic differences in past hysterectomy for benign conditions. Women’s Heal Issues. 2005;15(4):179–186. doi: 10.1016/j.whi.2005.05.002 [DOI] [PubMed] [Google Scholar]
- [16].Bower JK, Schreiner PJ, Sternfeld B, Lewis CE. Black–White Differences in Hysterectomy Prevalence: The CARDIA Study. Am J Public Health. 2009;99(2):300–307. doi: 10.2105/AJPH.2008.133702 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [17].Robinson WR, Nichols HB, Tse CK, Olshan AF, Troester MA. Associations of Premenopausal Hysterectomy and Oophorectomy With Breast Cancer Among Black and White Women: The Carolina Breast Cancer Study, 1993–2001. Am J Epidemiol. 2016;184(5):388–399. doi: 10.1093/aje/kwv448 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Boggs DA, Palmer JR, Rosenberg L. Bilateral oophorectomy and risk of cancer in African American women. Cancer Causes Control. 2014;25(4):507–513. doi: 10.1007/s10552-014-0353-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Parker WH, Broder MS, Chang E, et al. Ovarian conservation at the time of hysterectomy and long-term health outcomes in the nurses’ health study. Obstet Gynecol. 2009;113(5):1027–1037. doi: 10.1097/AOG.0b013e3181a11c64 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].Nichols HB, Trentham-Dietz A, Newcomb PA, et al. Pre-diagnosis oophorectomy, estrogen therapy and mortality in a cohort of women diagnosed with breast cancer. Breast Cancer Res. 2013;15(5):R99. doi: 10.1186/bcr3560 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [21].Newman B, Moorman PG, Millikan R, et al. The Carolina Breast Cancer Study: integrating population-based epidemiology and molecular biology. Breast Cancer Res Treat. 1995;35(1):51–60. doi: 10.1007/BF00694745 [DOI] [PubMed] [Google Scholar]
- [22].Weinberg CR, Wacholder S. The design and analysis of case-control studies with biasedsampling. Biometrics. 1990;46(4):963–975. doi: 10.2307/2532441 [DOI] [PubMed] [Google Scholar]
- [23].DeBono NL, Robinson WR, Lund JL, et al. Race, Menopausal Hormone Therapy, and Invasive Breast Cancer in the Carolina Breast Cancer Study. J Women’s Heal. 2018;27(3):377–386. doi: 10.1089/jwh.2016.6063 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [24].Moorman P, Newman B, Millikan R, Tse C-K, Sandler D. Participation Rates in a Case-Control Study:: The Impact of Age, Race, and Race of Interviewer. Ann Epidemiol. 1999;9(3):188–195. doi: 10.1016/S1047-2797(98)00057-X [DOI] [PubMed] [Google Scholar]
- [25].Lin DY, Wei LJ. The Robust Inference for the Cox Proportional Hazards Model. J Am Stat Assoc. 1989;84(408):1074–1078. doi: 10.1080/01621459.1989.10478874 [DOI] [Google Scholar]
- [26].Leon AC, Heo M. Sample Sizes Required to Detect Interactions between Two Binary Fixed-Effects in a Mixed-Effects Linear Regression Model. Comput Stat Data Anal. 2009;53(3):603–608. doi: 10.1016/j.csda.2008.06.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [27].Greenland S Tests for interaction in epidemiologic studies: A review and a study of power. Stat Med. 1983;2(2):243–251. doi: 10.1002/sim.4780020219 [DOI] [PubMed] [Google Scholar]
- [28].Welsh ML, Buist DSM, Aiello Bowles EJ, Anderson ML, Elmore JG, Li CI. Population-based estimates of the relation between breast cancer risk, tumor subtype, and family history. Breast Cancer Res Treat. 2009;114(3):549–558. doi: 10.1007/s10549-008-0026-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [29].Barnard ME, Boeke CE, Tamimi RM. Established breast cancer risk factors and risk of intrinsic tumor subtypes. Biochim Biophys Acta - Rev Cancer. 2015;1856(1):73–85. doi: 10.1016/J.BBCAN.2015.06.002 [DOI] [PubMed] [Google Scholar]
- [30].Doll KM, Dusetzina SB, Robinson W. Trends in Inpatient and Outpatient Hysterectomy and Oophorectomy Rates Among Commercially Insured Women in the United States, 2000–2014. JAMA Surg. 2016;151(9):876. doi: 10.1001/jamasurg.2016.0804 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [31].Obermair A, Youlden DR, Baade PD, Janda M. The impact of risk-reducing hysterectomy and bilateral salpingo-oophorectomy on survival in patients with a history of breast cancer-A population-based data linkage study. Int J Cancer. 2014;134(9):2211–2222. doi: 10.1002/ijc.28537 [DOI] [PubMed] [Google Scholar]
- [32].Obermair A, Youlden D, Baade P, Janda M. Do breast cancer survivors benefit from prophylactic removal of uterus and ovaries? A population-based data linkage replication study. Asia Pac J Clin Oncol. 2017;13(1):68–78. doi: 10.1111/ajco.12508 [DOI] [PubMed] [Google Scholar]
- [33].Domchek SM, Friebel TM, Singer CF, et al. Association of Risk-Reducing Surgery in BRCA1 or BRCA2 Mutation Carriers With Cancer Risk and Mortality. JAMA. 2010;304(9):967. doi: 10.1001/jama.2010.1237 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [34].Ludwig KK, Neuner J, Butler A, Geurts JL, Kong AL. Risk reduction and survival benefit-of prophylactic surgery in BRCA mutation carriers, a systematic review. Am J Surg. 2016;212(4):660–669. doi: 10.1016/j.amjsurg.2016.06.010 [DOI] [PubMed] [Google Scholar]
- [35].McCarthy AM, Menke A, Ouyang P, Visvanathan K. Bilateral Oophorectomy, Body Mass Index, and Mortality in U.S. Women Aged 40 Years and Older. Cancer Prev Res. 2012;5(6):847–854. doi: 10.1158/1940-6207.CAPR-11-0430 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [36].Phipps AI, Buist DSM. Validation of self-reported history of hysterectomy and oophorectomy among women in an integrated group practice setting. Menopause. 2009;16(3):576–581. doi: 10.1097/gme.0b013e31818ffe28 [DOI] [PMC free article] [PubMed] [Google Scholar]