Selected medical interventions in women with a deleterious BRCA mutation : a population-based study in British Columbia

Background We examined the uptake of risk-reducing interventions, including bilateral mastectomy, risk-reducing salpingo-oophorectomy, oral contraceptive pills, tamoxifen, and raloxifene, for the entire population of women with a deleterious BRCA1 or BRCA2 mutation in the Canadian province of British Columbia. Methods This retrospective population-based study used data available in British Columbia for all women who, between 1996 and 2014, were tested and found to have a BRCA mutation. Rates of risk-reducing interventions stratified according to the type of BRCA mutation and prior history of breast or gynecologic cancer (ovary, fallopian tube, peritoneal) are presented. Cancers diagnosed in women with a BRCA mutation after disclosure of their mutation status are also presented. Results The final study cohort consisted of 885 patients with a deleterious BRCA1 (n = 474) or BRCA2 (n = 411) mutation. Of the women with no prior breast cancer, 30.8% carrying a BRCA1 mutation and 28.3% carrying a BRCA2 mutation underwent bilateral mastectomy. Of women with no prior gynecologic cancer, 64.7% carrying a BRCA1 mutation and 62.2% carrying a BRCA2 mutation underwent risk-reducing bilateral salpingo-oophorectomy. Rates of chemoprevention with oral contraceptive pills and tamoxifen or raloxifene were low in all groups. In this cohort, 23 gynecologic and 70 breast cancers were diagnosed after disclosure of BRCA mutation status. Conclusions Our results suggest reasonable uptake of risk-reducing interventions in high-risk women. To minimize the occurrence of breast and ovarian cancer in women with a BRCA1 or BRCA2 mutation, more attention could be paid to ensuring that affected women receive proper counselling and follow-up.


INTRODUCTION
Women with a BRCA1 or BRCA2 germline mutation are at increased risk for breast and ovarian cancer.Compared with the average cumulative risk of 12% and 1.4% in the general population 1 , women with a germline BRCA1 mutation have an average lifetime risk of 47%-66% for breast cancer and 35%-46% for ovarian cancer 2 .Among women with a germline BRCA2 mutation, the average cumulative risks are 40%-57% and 13%-23% respectively 2 .In many jurisdictions, hereditary cancer programs identify women at high risk of cancer, with the goal of preventing future cases of cancer or offering enhanced screening to detect cancers earlier, when they are more treatable.
Several effective interventions to reduce the risk of future breast and ovarian cancer are available to women with BRCA1 and BRCA2 mutations.Those interventions include bilateral mastectomy, which has been shown to decrease the risk for breast cancer by approximately 90% [3][4][5][6] , and risk-reducing bilateral salpingo-oophorectomy (rrbso), which has been shown to decrease the risk for ovarian cancer by 80%-96% [7][8][9][10][11][12] .Chemoprevention options are also available (although considerably less evidence is available for their effectiveness in this population), including the use of oral contraceptive pills for prevention of ovarian cancer 13 and the use of tamoxifen 14 or raloxifene 15 for prevention of breast cancer.
Enhanced screening protocols designed for earlier detection of a possible breast cancer-for example, magnetic resonance imaging or mammography-can also be considered 16,17 .Screening for ovarian cancer is not recommended, because no mortality benefit has been demonstrated, even with strict adherence to screening protocols [18][19][20][21][22] .
In British Columbia, women who have a mutation that increases their risk for breast or ovarian cancer or a first-degree relative with a confirmed mutation are eligible for referral to the High-Risk Clinic run by the Hereditary Cancer Program (hcp), where they will be counselled by a nurse practitioner and a medical director who arrange for appropriate screening and communicate up-to-date recommendations about risk-reducing interventions.However, not all patients with BRCA mutations are counselled at the High-Risk Clinic; many live too far away or opt for other forms of follow-up.A recent study of the patients counselled in the clinic reported high rates of rrbso (>80%) in patients more than 40 years of age and a 38% rate of prophylactic mastectomy 23 .Here, we present population-based data about risk-reducing interventions in women with a BRCA1 or BRCA2 mutation regardless of whether those women were treated in the High-Risk Clinic or elsewhere.We hypothesized that rates of rrbso and bilateral mastectomy in this population-based cohort of women would be lower than rates reported for the subset of women treated in the High-Risk Clinic.

METHODS
In this population-based retrospective study, we analyzed all women who were tested at the hcp between 1996 and 2014 in the Canadian province of British Columbia.Subjects were eligible for the study if they were female, had tested positive for a deleterious BRCA1 or BRCA2 mutation (10 women also tested positive for a variant of unknown significance), had at least 1 year of follow-up data in the population-based datasets, and had been registered with the provincial health insurance program for at least 275 days in the year in which they were identified as BRCApositive and in the subsequent year.Although follow-up continued until 31 December 2014, we imposed the registration criteria for only 2 calendar years to ensure that the women were residing primarily in British Columbia during the time of their of BRCA mutation testing and disclosure.
With approval of all data stewards, we obtained data about health services use from Population Data BC, data from the BC Cancer Registry, vital statistics death data, and data from BC PharmaNet a .Those data include all physician services, hospitalizations, cancer diagnoses, and prescription drug use in an outpatient setting for the entire population of British Columbia.The data were linked with data from BC Cancer's hcp-a program that provides BRCA1 and BRCA2 mutation testing to all patients in British Columbia.All inferences, opinions, and conclusions drawn are those of the authors and do not reflect the opinions or policies of the data stewards.Ethics approval was obtained from the University of British Columbia Behavioural Research Ethics Board.

Procedures
For all identified patients, we examined cancer histories and use of these risk-reducing interventions: mastectomy, rrbso, and chemoprevention (oral contraceptive pills, or tamoxifen, or raloxifene).Women were stratified according to their prior history of breast cancer when examining mastectomy or chemoprevention for breast cancer, and their prior history of gynecologic cancer (defined as ovarian, fallopian tube, or peritoneal cancer) when examining rrbso and oral contraceptive pill use.That approach allows for the presentation of data about the entire population, while acknowledging that, for women with prior breast and ovarian cancers, certain interventions-mastectomies, bilateral salpingo-oophorectomy, tamoxifen, and so on-are almost certainly being used as part of cancer treatment and not for risk reduction.We also present rates of rrbso in women who were 40 years of age and older by the end of the follow-up period (because rrbso is not recommended until age 40).

Cancers Diagnosed After BRCA Mutation Disclosure
To examine "failed" prevention opportunities, we also present the number of gynecologic and breast cancers that were diagnosed after a patient was made aware of their BRCA mutation status.

Statistical Analysis
The chi-square test was used to compare frequencies for categorical variables and to compare rates of intervention uptake.Mean values for all continuous variables were compared using t-tests.All statistical tests were performed in the Stata software application (version 13: StataCorp LP, College Station, TX, U.S.A.).When reporting certain data cells might result in inadvertent disclosure, approximate percentages are presented (rounded to the nearest multiple of 5).

RESULTS
In British Columbia, 1503 residents tested positive for a deleterious BRCA1 or BRCA2 mutation between 1996 and 2014.After the exclusion of male patients (n = 211), patients without a full year of follow-up because their BRCA status was disclosed too near the end of the follow-up period or because they died or moved out of province within a year of a Data extracts provided by Population Data BC and the British Columbia Ministry of Health from the BC Cancer Registry (ver.2, http://www.bccancer.bc.ca/health-professionals/professionalresources/bc-cancer-registry), the Medical Services Plan Payment Information File (ver.2, https://www.popdata.bc.ca/data/health/ msp), the BC Perinatal Data Registry (https://www.popdata.bc.ca/ data/health/PSBC), and PharmaNet (https://www2.gov.bc.ca/gov/ content/health/practitioner-professional-resources/pharmacare/ pharmanet-bc-s-drug-information-network); and by the Canadian Institute for Health Information from the Discharge Abstract Database (ver.2, https://www.cihi.ca/en/discharge-abstract-database-metadata);all 2015.disclosure (n = 353), and patients who were not registered for health care in British Columbia for 275 days or more (75% of the calendar year, n = 54), the final study cohort included 885 patients with a BRCA1 (n = 474) or BRCA2 mutation (n = 411).
Table i details the characteristics of the study cohort.Approximately one third of the women with a BRCA1 mutation (35.5%) or a BRCA2 mutation (27.8%) knew their mutation status before age 40.No differences in the rates of breast and ovarian cancer diagnosed before BRCA testing were observed between women with a BRCA1 or BRCA2 mutation, and fewer than half the women were diagnosed with breast cancer before BRCA testing.Of every 10 women, approximately 1 was diagnosed with a gynecologic cancer before BRCA testing.No significant difference in age between the groups was observed at diagnosis of the gynecologic cancers.

Mastectomy and Chemoprevention for Breast Cancer
Table ii reports use of mastectomy and chemoprevention for breast cancer stratified by whether the women had previously been diagnosed with breast cancer.Of all patients with a BRCA1 mutation, 40.9% underwent bilateral mastectomy (30.8% of women who had not previously had breast cancer, and 55.4% of those who had previously been diagnosed with breast cancer, p < 0.001).Women with a BRCA1 mutation and no prior cancer were younger than those who had previously been diagnosed with cancer at the time of mastectomy (45.1 years vs. 49.9years, p = 0.006), and they waited an average of 2.8 years between BRCA1 mutation disclosure and mastectomy.Of women with a BRCA2 mutation, 28.3% with no prior breast cancer and 58.2% with a prior diagnosis of breast cancer underwent bilateral mastectomy (p < 0.001).The average time between disclosure of the BRCA2 mutation and mastectomy was 2.5 years in women without a prior cancer and 0.5 years in those with a prior cancer (p < 0.001).
With respect to chemoprevention for breast cancer, tamoxifen use was rare in women without a prior breast cancer diagnosis (1.8% for women with a BRCA1 mutation and 6.2% for women with a BRCA2 mutation).Very little use of raloxifene was observed in any of the groups studied (<5% in all groups).

RRBSO and Chemoprevention for Ovarian Cancer
Table iii outlines risk-reducing or cancer treatment interventions for ovarian cancer stratified by whether the women had previously been diagnosed with a gynecologic cancer.Most women with a BRCA1 mutation underwent rrbso (64.7% for those without a prior cancer and 80.5% for those with a prior cancer, p = 0.041).For women who were 40 years of age or older by the end of the follow-up period, rates of rrbso uptake increased to 74.4% among women without a prior cancer and 80.0% among those with a prior cancer (rrbso would have been part of cancer treatment for women with a gynecologic cancer).Of women with a BRCA1 mutation, those without a prior cancer diagnosis were 49.1 years, on average, at the time of rrbso, and they waited an average of 1.6 years after BRCA disclosure to undergo rrbso; those without a prior cancer diagnosis waited an average of 3 years (p = 0.745).Most women with a BRCA2 mutation underwent rrbso (62.2% of those without a prior cancer diagnosis and 76.7% of those with a prior cancer diagnosis, p = 0.061).Among women who were 40 years of age or older by the end of the follow-up period, rates of rrbso increased to 71.2% for those without a prior cancer diagnosis and 76.7% for those with a prior cancer diagnosis.The average age at rrbso was 50.0 years for women with a BRCA2 mutation and no prior cancer diagnosis, and they waited an average of 1.2 years after BRCA disclosure to undergo rrbso.
Table iii also outlines the use of oral contraceptive pills by women with a BRCA mutation.Of women with a

Cancers Diagnosed After BRCA Mutation Disclosure
Table iv describes cancer diagnoses in women after BRCA disclosure.After BRCA disclosure, 23 women were diagnosed with a gynecologic cancer (16 ovarian cancers).Further details about those cancers could not be reported for privacy reasons connected to small cell sizes.After BRCA disclosure, 70 women were diagnosed with breast cancer.
Although staging data are missing for cancers diagnosed before 2010, among women with complete staging data, 60% were stage i and 30% were stage ii.Of women with gynecologic cancer, 11 (47.8%) were diagnosed during their rrbso.The number of women diagnosed with breast cancer during their bilateral mastectomy was too small to report.Of the breast cancer patients diagnosed after their genetic testing results were disclosed, 6 (8.5%) had died by the end of the follow-up period; of the gynecologic cancer patients, 7 (30.4%)had died by the end of the follow-up period.
Our data indicate that, of women not previously diagnosed with a breast cancer, approximately 30% with a deleterious BRCA mutation underwent bilateral mastectomy after disclosure of their BRCA status.The decision-making process for bilateral mastectomy is complex, given the effectiveness of mammography and magnetic resonance imaging for breast cancer screening in this high-risk population 24 .Our data likely reflect the choice by many women for enhanced screening in preference to mastectomy.The number of women with a prior breast cancer diagnosis undergoing mastectomy is higher, reflecting mastectomies performed for treatment purposes.
There are no effective screening methods for ovarian cancer, and 5-year overall survival remains low 25,26 .As a result, close to two thirds of B.C. women with a deleterious BRCA1 or BRCA2 mutation and no prior gynecologic cancer underwent rrbso for prevention.That proportion exceeded 70% when looking solely at women who were 40 years of age or older at the end of study follow-up.
We found that very few high-risk women were using medications for prevention.Fewer than 10% of women with no prior gynecologic cancer used oral contraceptives after disclosure of their BRCA status.That observation is likely at least partly explained by the age of the women in our cohort (most were more than 45 years of age at the time of BRCA disclosure), the fact that use of oral contraceptive pills for ovarian cancer prevention in high-risk women is less well supported by the evidence, and the fact that most women proceed to rrbso after disclosure.No randomized controlled trials have been reported, but observational studies have shown associations between the use of oral contraceptives and a reduced risk of ovarian cancer for Defined as mastectomy with a corresponding prophylaxis code performed after a woman was aware of her BRCA mutation status.
b Cell size suppressed because of privacy restrictions.Actual number and percentage have been replaced by an approximation.
BRCA1 and BRCA2 carriers, with odds ratios suggesting a 40%-50% reduction in risk 13 .Two meta-analyses have not supported an increased risk of breast cancer for high-risk women using oral contraceptive pills 13,27 .Our results suggest that tamoxifen and raloxifene are not being regularly used as primary chemoprevention.Data about the use of tamoxifen for prevention of breast cancer in BRCA1 and BRCA2 carriers are limited, but there is some indication of effectiveness in BRCA2 carriers 14 .Trials of raloxifene use have not included many BRCA carriers, which makes drawing conclusions about its effectiveness difficult 28 .Thus, the low rates of use in our cohort are not surprising and not indicative of a failure to provide an effective chemoprevention option.With respect to how our results compare with publications examining women who were seen at the hcp's High-Risk Clinic at BC Cancer, 24% of woman more than 40 years of age received prophylactic mastectomies 23 .We report that approximately 30% of our BRCA1 and BRCA2 mutation carriers without a prior breast cancer diagnosis underwent bilateral mastectomy.The difference is likely explained by the fact that women who want to proceed directly to mastectomy after disclosure of their test result would not necessarily be seen at the High-Risk Clinic.The number of women undergoing rrbso for prevention in our study (approximately 70% of those without a history of gynecologic cancer who were 40 years of age or older at the end of the follow-up period) is lower than that reported from the High-Risk Clinic, where 80% of women more than 40 years of age had undergone rrbso, suggesting that those counselled in the clinic are more likely to undergo that important preventive intervention.Our work differs from previous publications in that we included the entire population of women with a BRCA mutation in British Columbia regardless of where they were treated.We also present data about chemoprevention.Our work was strengthened by its population-based nature and our ability to access long-term follow-up data with procedure codes revealing all surgeries that the women had undergone (with exact dates) and all medicines that were dispensed (with dates).Thus, we were not forced to rely on patient recall.Limitations of our work include a lack of data about screening for breast cancer.We were unable to ascertain whether bilateral mastectomy in women with a prior breast cancer diagnosis was undertaken partly as a preventive measure if the woman had cancer only in one breast.Unfortunately, the data do not provide that level of detail.We also lacked data about which of the women included in our study were seen at the High-Risk Clinic, but that proportion of our study population would be expected to be approximately 75%, meaning that most women in the present study would have received some counselling about their best prevention and screening approaches.We also lacked a negative control group to determine whether rates of risk-reducing interventions were comparatively higher for women who screened positive than for women who screened negative.Such a comparison is especially important, given recent reports by Kurian et al. 29 indicating that average-risk patients are frequently undergoing bilateral mastectomy despite lack of evidence for a survival advantage [30][31][32] .Finally, we lacked information about whether women consider these risk-reducing interventions acceptable-an important consideration in whether an intervention will be used.Although we cannot report those data, previous studies into the psychological effects of bilateral mastectomy and rrbso have reported significantly less worry after the surgery and a high quality of physical and mental well-being [33][34][35][36] .
We also report 70 breast and 23 gynecologic cancers diagnosed after BRCA disclosure.Although some of those cancers were diagnosed at the time of risk-reducing surgery, most occurred in women who had not yet undergone any risk-reducing interventions.Future work should seek to illuminate why these high-risk women are not undergoing risk-reducing interventions to improve rates of cancer prevention and to lower the number of breast and gynecologic cancers for known high-risk women.The funding sources had no involvement in the research or the preparation of this article.We thank SzeWing Mung for her work compiling the data from the hcp.

CONFLICT OF INTEREST DISCLOSURES
We have read and understood Current Oncology's policy on disclosing conflicts of interest, and we declare that we have none.

*
Department of Obstetrics and Gynaecology and † Department of Surgery, University of British Columbia; ‡ BC Cancer, Hereditary Cancer Program, High-Risk Clinic; § Department of Medical Genetics, University of British Columbia; || BC Cancer, Hereditary Cancer Program; and # Laboratory Medicine, University of British Columbia, Vancouver, BC.

TABLE I
Characteristics of patients with a deleterious mutation in BRCA1 or BRCA2 HCP = Hereditary Cancer Program; Dx = diagnosis.e20Current Oncology, Vol. 26, No. 1, February 2019 © 2019 Multimed Inc.deleterious BRCA mutation, approximately one third without a prior cancer diagnosis had used oral contraceptive pills (any use, defined as filling at least 1 prescription); fewer than 1 in 10 women with a prior cancer history used oral contraceptive pills (p < 0.001).Most of the use by women with a prior gynecologic cancer diagnosis occurred before BRCA disclosure, with no oral contraceptive pill use occurring in 93.3% of women with a BRCA1 mutation and in 92.4% of women with a BRCA2 mutation after their BRCA disclosure.

TABLE II
Use of prophylactic mastectomy, a tamoxifen, and raloxifene by breast cancer status

TABLE III
Bilateral salpingo-oophorectomy (BSO) and oral contraceptive pill use by gynecologic cancer status a Cell size suppressed because of privacy restrictions.Actual number and percentage have been replaced by an approximation.e22 Current Oncology, Vol. 26, No. 1, February 2019 © 2019 Multimed Inc.
This research was supported by donor funds from the Vancouver General Hospital and the University of British Columbia Hospital Foundation and by a Canadian Institutes of Health Research/ Canadian Cancer Society Research Institute prevention research grant (no.703037).GEH is supported by the Janet D. Cottrelle Foundation and a Canadian Cancer Society Research Institute Capacity Development Award in Cancer Prevention (no.702786).

TABLE IV
Characteristics of 70 breast cancer diagnoses in the followup period a Cell size suppressed because of privacy restrictions.