Bone health after RRBSO among BRCA1/2 mutation carriers: a population-based study.

OBJECTIVE: Examine the risks of fractures and osteoporosis after risk-reducing bilateral salpingo-oophorectomy (RRBSO) among women with BRCA1/2 mutations.
METHODS: In this retrospective population-based study in British Columbia, Canada, between 1996 to 2017, we compared risks of osteoporosis and fractures among women with BRCA1/2 mutations who underwent RRBSO before the age of 50 (n=329) with two age-matched groups without known mutations: 1) women who underwent bilateral oophorectomy (BO) (n=3,290); 2) women with intact ovaries who had hysterectomy or salpingectomy (n=3,290). Secondary outcomes were: having dual-energy X-ray absorptiometry (DEXA) scan, and bisphosphonates use.
RESULTS: The mean age at RRBSO was 42.4 years (range, 26-49) and the median follow-up for women with BRCA1/2 mutations was 6.9 years (range, 1.1-19.9). There was no increased hazard of fractures for women with BRCA1/2 mutations (adjusted hazard ratio [aHR]=0.80; 95% confidence interval [CI]=0.56-1.14 compared to women who had BO; aHR=1.02; 95% CI=0.65-1.61 compared to women with intact ovaries). Among women who had DEXA-scan, those with BRCA1/2 mutations had higher risk of osteoporosis (aHR=1.60; 95% CI=1.00-2.54 compared to women who had BO; aHR=2.49; 95% CI=1.44-4.28 compared to women with intact ovaries). Women with BRCA1/2 mutations were more likely to get DEXA-scan than either control groups, but only 46% of them were screened. Of the women with BRCA1/2 mutations diagnosed with osteoporosis, 36% received bisphosphonates.
CONCLUSION: Women with BRCA1/2 mutations had higher risk of osteoporosis after RRBSO, but were not at increased risk of fractures during our follow-up. Low rates of DEXA-scan and bisphosphonates use indicate we can improve prevention of bone loss.

INTRODUCTION

Women with BRCA1/2 mutations are at an increased lifetime risk for ovarian cancer [1]. A risk-reducing bilateral salpingo-oophorectomy (RRBSO) between 35–45 years, after completion of childbearing, is the gold standard preventive strategy in this population [23]. Indeed, it is estimated that RRBSO reduces the risk of ovarian or fallopian cancers by 80% in women with BRCA1/2 mutations [4].

However, RRBSO causes surgical menopause that is known to have a negative impact on long-term health. Studies examining outcomes from surgical menopause among women at general population risk for ovarian cancer report an increased risk of osteoporosis and fractures, cardiovascular diseases, and possible increased risk of cognitive impairment, as well as excess mortality [56789]. Importantly, a large prospective study observed a 3.64 times increased risk of fractures (95% confidence interval [CI]=1.01–13.04) when bilateral oophorectomy (BO) had been carried out under the age of 45 years [9].

With respect to women with BRCA1/2 mutations, there is evidence of significant bone mineral density loss and increased risk of osteoporosis after RRBSO, but the current literature is limited by small sample sizes, short follow-up periods, and/or lack of control groups [1011121314]. Data are even more scarce on the risk of fractures, as evidenced by a 2018 Cochrane review, which could not draw conclusions on the incidence of fractures after RRBSO as none of the studies meeting the inclusion criteria reported this outcome [15]. Moreover, preliminary science suggest that BRCA1/2 mutations may be associated with premature ovarian and systemic aging, including osteoporosis in mice [1617].

To provide more data on bone health in BRCA1/2 mutation carriers undergoing RRBSO at premenopausal ages, this study investigates the risk of fractures and osteoporosis in such a population-based cohort as compared to 2 groups of women without known BRCA1/2 mutations matched by age at surgery: 1) women who underwent BO for benign gynecologic conditions; and, 2) women who underwent hysterectomy or salpingectomy but retained their ovaries. We also analyzed the use of health services to prevent fractures and maintain bone health (including having bone mineral density scanning and receiving bisphosphonates), and whether the use of hormone replacement therapy (HRT) mitigated the risk of bone diseases in this population.

MATERIALS AND METHODS

In this retrospective study we analyzed population-based administrative data from British Columbia (BC), Canada, between 1996 and 2017. We linked hospitalizations and physicians visits for all BC residents with data from the BC Cancer Registry and the BC PharmaNet, a database including all medications dispensed in an outpatient setting [1819202122]. Finally, we linked with data from the Hereditary Cancer Program, (the sole publicly funded site testing for BRCA1/2 mutations in BC) to obtain information on virtually all women tested in the province from 1996 to 2014. Ethics approval was obtained from the University of British Columbia Clinical Research Ethics Board. All inferences, opinions, and conclusions drawn in this paper are those of the authors and do not reflect the opinions or policies of the Data Stewards.

The primary outcomes were fractures and osteoporosis. Secondary outcomes included: 2) having a dual-energy X-ray absorptiometry (DEXA) scan, and 3) use of bisphosphonates initiated after surgery. The outcomes were identified from the hospital and physicians’ visits data using the International Classification of Diseases codes (version-9 and -10) and Medical Services Plan fee item billing codes, and from the BC PharmaNet using the Anatomical Therapeutic Chemical Classification System (Table S1).

1. Study population

Women were included if they: 1) did not have a diagnosis of ovarian, fallopian tube, or peritoneal cancers, and did not have gynecologic cancer listed as an indication for the index surgery; 2) had at least 1 year of follow-up; and 3) were registered in the universal provincial insurance program in the year of their surgery. While we included patients with a history of breast cancer, we excluded those who had a DEXA-scan in the two years before the start of the follow-up, as this suggests their breast cancer treatment was associated with increased risk of bone loss and fractures. Women with a diagnosis of osteoporosis, or hip or vertebral fractures (prototypical osteoporotic fractures) [23] in the 2 years preceding the study entry were also excluded.

We included all women with documented deleterious BRCA1/2 mutations who underwent RRBSO prior to age 50 (before the average age of natural menopause) between January 1st, 1996 and December 31st, 2017. We included 2 control groups, one of women who underwent BO (with or without salpingectomy) for benign gynecologic conditions and a group of women with intact ovaries who underwent hysterectomy or salpingectomy. Both control groups were women who were not tested for BRCA1/2 mutations (without known mutations) matched by age at surgery, randomly selected at a 10:1 rate.

2. Statistical analysis

The beginning of the follow-up was defined as the date of the index surgery. For baseline comparisons we used t-tests for unpaired groups or χ2 tests. All p-values are two-sided and statistical significance was defined as p<0.05. Multivariate Cox proportional hazards models were used to estimate adjusted hazard ratios (aHR) and 95% CI for the outcomes. In this time to event analysis, women were censored at death, date of diagnosis of breast cancer, or if they moved away from the province. The models were adjusted for history of breast cancer, unless otherwise specified.

As osteoporosis is usually asymptomatic, we also estimated the hazard of osteoporosis considering only women who had DEXA-scan, to account for differences in the screening across comparison groups. We performed an analysis stratified by history of breast cancer given that some treatment options (e.g., aromatase inhibitors) are risk factors for osteoporosis and fractures [24]. Finally, we stratified the population by their use of HRT after surgical menopause. Any use of HRT was defined as the dispensation of at least 30 days of systemic estrogen or estrogen plus progesterone preparations. For all stratified analysis, age at baseline was tested as a covariate since the age-matching was no longer valid. Statistical analyses were performed with R software version 4.0.3 (R Foundation for Statistical Computing, Vienna, Austria).

RESULTS

A total of 1,292 women tested positive for deleterious BRCA1/2 mutations, and 430 underwent RRBSO under the age of 50 during the study period. After excluding 64 women due to prior gynecologic cancer, 18 for having less than 1 year of follow-up, and 19 for having an outcome of interest or a history of breast cancer and a DEXA-scan prior to surgery, we included 329 women with BRCA1/2 mutations (Fig. 1). For the control groups, 10,046 women who had BO and 54,707 with intact ovaries who had hysterectomy or salpingectomy filled the same inclusion criteria and did not have known BRCA1/2 mutations. From each group, ten controls matched by age at surgery were randomly selected for each woman with a BRCA1/2 mutation. Of the 3,290 women with intact ovaries, 1,922 underwent hysterectomy, 538 underwent salpingectomy, and 830 underwent hysterectomy with salpingectomy.

Fig. 1

Flow diagram of study population analyzing bone health in women with BRCA1/2 mutations who underwent RRBSO at premenopausal age compared to 2 control groups of age matched women without known mutations, one of women who underwent BO and another group of women with intact ovaries who underwent hysterectomy or salpingectomy.

RRBSO, risk-reducing bilateral salpingo-oophorectomy; BO, bilateral oophorectomy; DEXA, dual-energy X-ray absorptiometry.

The mean age at RRBSO was 42.4 years (range, 26–49) and the median follow-up time for women with BRCA1/2 mutations was 6.9 years (range, 1.1–19.9; Table 1). Follow-up was longer, on average, for the control groups (both p<0.001). Women with BRCA1/2 mutations were more likely to have a history of breast cancer (48.9% compared to 4.6% of women without mutations who had BO; and to 0.3% of women without mutations with intact ovaries, both p<0.001), and to be in the higher income quintiles than both control groups (both p<0.05). Among women without a history of breast cancer (as HRT is not recommended among women with a history of breast cancer) [2526], there was no significant difference in HRT use after surgical menopause between women with BRCA1/2 mutations and those without a mutation who underwent BO (72.0% vs. 68.5%, p=0.386).

Table 1

Characteristics of the study population matched by age at surgery

Characteristics		BRCA mutation with RRBSO (n=329)	BO without BRCA mutation (n=3,290)	p-value*	Intact ovaries without BRCA mutation (n=3,290)	p-value*
Age (yr)				0.940		0.937
	Mean±SD	42.4±4.8	42.5±4.8		42.5±4.8
	Min–Max	26.7–49.9	26.0–49.9		26.2–49.9
Age category				1.000		1.000
	<35	23 (7.0%)	230 (7.0%)		230 (7.0%)
	35–39	81 (24.6%)	810 (24.6%)		810 (24.6%)
	40–44	103 (31.3%)	1,030 (31.3%)		1,030 (31.3%)
	45–49	122 (37.1%)	1,220 (37.1%)		1,220 (37.1%)
Follow-up (yr)				<0.001		<0.001
	Median	6.9	10.6		8.6
	Min–Max	1.1–19.9	1.0–22.0		1.0–22.0
Breast cancer before surgery		161 (48.9%)	151 (4.6%)	<0.001	9 (0.3%)	<0.001
	No	168 (51.1%)	3,139 (95.4%)		3,281 (99.7%)
Other cancers		23 (7.0%)	310 (9.4%)	0.175	252 (7.7%)	0.743
Income quintile				<0.001		0.008
	1	42 (12.8%)	670 (20.4%)		625 (19.0%)
	2	61 (18.5%)	704 (21.4%)		650 (19.8%)
	3	58 (17.6%)	655 (19.9%)		666 (20.2%)
	4	81 (24.6%)	598 (18.2%)		694 (21.1%)
	5	80 (24.3%)	588 (17.9%)		592 (18.0%)
	Missing	7 (2.1%)	75 (2.3%)		63 (1.9%)
HRT after surgery		153 (46.5%)	2,159 (65.6%)	<0.001	-	-
	History of breast cancer†	32 (19.9%)	8 (5.3%)	<0.001	-	-
	No history of breast cancer‡	121 (72.0%)	2,151 (68.5%)	0.386	-	-
Duration HRT (years)				0.031	-	-
	Mean±SD	3.5±2.5	4.0±3.7
	Min–Max	0.1–10.7	0.1–18.4

RRBSO, risk-reducing bilateral salpingo-oophorectomy; BO, bilateral oophorectomy; SD, standard deviation; HRT, hormone replacement therapy.

*Reference group: BRCA mutation with RRBSO; †Denominator is the number of women without prior breast cancer (n=161 and n=151, respectively); ‡Denominator is the number of women with prior breast cancer (n=168 and n=3,139, respectively).

Table 2 shows the frequency of the outcomes by the end of the study period. Of the women with BRCA1/2 mutations, 13.4% (n=44) had fractures, 46.2% (n=152) had DEXA-scan after surgery, and of these 14.5% (n=22) had osteoporosis. The p-values were not provided, as this comparison does not account for differences in the follow-up time.

Table 2

Frequency of the outcomes by study group at the end of the study period

Outcomes		BRCA mutation with RRBSO (n=329)	BO without BRCA mutation (n=3,290)	Intact ovaries without BRCA mutation (n=3,290)
Bone fractures		44 (13.4%)	616 (18.7%)	445 (13.5%)
	Hip fracture	≤5	21 (0.6%)	14 (0.4%)
	Vertebral fracture	≤5	36 (1.1%)	31 (0.9%)
	Forearm fracture	10 (3.0%)	132 (4.0%)	82 (2.5%)
	Humerus fracture	≤5	33 (1.0%)	31 (0.9%)
	Pelvic fracture	≤5	16 (0.5%)	13 (0.4%)
	Other fractures	27 (8.2%)	399 (12.1%)	299 (9.1%)
Osteoporosis		22 (6.7%)	127 (3.9%)	45 (1.4%)
DEXA scan		152 (46.2%)	761 (23.1%)	335 (10.2%)
Bisphosphonates		27 (8.2%)	180 (5.5%)	60 (1.8%)
No. of women who had DEXA scan		152	761	335
	Osteoporosis	22 (14.5%)	108 (14.2%)	38 (11.3%)
No. of women with osteoporosis diagnosis		22	127	45
	Bisphosphonates	8 (36.4%)	65 (51.2%)	21 (46.7%)

RRBSO, risk-reducing bilateral salpingo-oophorectomy; BO, bilateral oophorectomy; DEXA, dual-energy X-ray absorptiometry.

*p-values for the crude incidences were not provided, as this comparison between the groups is not meaningful due to differences in the follow-up time.

1. Primary outcomes

There were no significant differences in the hazard of fractures (Table 3) for women with BRCA1/2 mutations compared to both control groups (aHR=0.80; 95% CI=0.56–1.14 compared to women without mutations who had BO; aHR=1.02; 95% CI=0.65–1.61 compared to women without mutations with intact ovaries). However, fractures occurred earlier after RRBSO, on average, for women with BRCA1/2 mutations (5.8±3.7 years) than for women in both control groups (7.2±4.7 years for women without mutations who had BO p=0.024, and 7.3±4.7 years for women without mutations with intact ovaries, p=0.017).

Table 3

Primary outcomes by surgical and BRCA mutation status

Outcomes		BRCA mutation with RRBSO (n=329)	BO without BRCA mutation (n=3,290)	95% CI	BRCA mutation with RRBSO (n=329)	Intact ovaries without BRCA mutation (n=3,290)	95% CI
Bone fractures
	Persons-years	2,415.4	31,235.9	-	2,415.4	28,291.0	-
	No. of events	44	616	-	44	445	-
	Crude HR	0.96	1	0.71–1.31	1.21	1	0.89–1.65
	Adjusted HR*	0.80	1	0.56–1.14	1.02	1	0.65–1.61
Osteoporosis
	Persons-years	2,534.0	34,307.4	-	2,534.0	30,780.4	-
	No. of events	22	127	-	22	45	-
	Crude HR	2.64	1	1.67–4.17	7.42	1	4.39–12.55
	Adjusted HR*	1.83	1	1.03–3.26	6.38	1	3.11–13.08
No. of women who had DEXA scan		152	761		152	335
Osteoporosis
	Persons-years	1,304.8	9,542.6	-	1,304.8	4,542.1	-
	No. of events	22	108	-	22	38	-
	Crude HR	1.60	1	1.00–2.54	2.54	1	1.47–4.39
	Adjusted HR†	1.60	1	1.00–2.54	2.49	1	1.44–4.28

RRBSO, risk-reducing bilateral salpingo-oophorectomy; BO, bilateral oophorectomy; CI, confidence interval; HR, hazard ratio; DEXA, dual-energy X-ray absorptiometry.

*Adjusted for history breast cancer; †Adjusted for age at surgery.

The likelihood of being diagnosed with osteoporosis was higher among women with BRCA1/2 mutations (aHR=1.83; 95% CI=1.03–3.26 compared to women without mutations who had BO; aHR=6.38; 95% CI=3.11–13.08 compared to women without mutations with intact ovaries). Women with BRCA1/2 mutations were diagnosed with osteoporosis on average 8.0±4.4 years after RRBSO, but the difference in the time to diagnosis compared to the control groups was not significant (8.0±4.5 years for women without mutations who had BO, p=0.923, and 9.4±5.8 years for women without mutations with intact ovaries, p=0.279). The higher hazard of osteoporosis was also observed considering only women who had a DEXA-scan when comparing to women without mutations who had BO (aHR=1.60; 95% CI=1.00–2.54).

The results were similar after excluding women with a history of breast cancer (Table S2). However, the difference in the hazard of osteoporosis was not significant comparing women with BRCA1/2 mutations to women without mutations who had BO, considering only those who had a DEXA-scan (HR=1.52; 95% CI=0.77–3.03). Crude HRs were provided as there were no relevant covariates.

We further compared women without a mutation who underwent BO to women without mutations with intact ovaries who had hysterectomy or salpingectomy (Table S3). We observed a higher hazard of fractures in the group who had BO (aHR=1.22; 95% CI=1.08–1.38).

2. Secondary outcomes

Women with BRCA1/2 mutations were more likely to have DEXA-scan after RRBSO than women in both control groups (aHR=1.58; 95% CI=1.26–1.98, compared to women without mutations who had BO; aHR=6.03; 95% CI=4.64–7.86 compared to women without mutations with intact ovaries, Table 4). Of the 22 women with BRCA1/2 mutations diagnosed with osteoporosis after RRBSO, only 8 (36.4%) filled a prescription for bisphosphonates. Among women with an osteoporosis diagnosis there were no significant differences in the likelihood of receiving bisphosphonates for women with BRCA1/2 mutations compared to both control groups (Table 4).

Table 4

Secondary outcomes by surgical and BRCA mutation status

Outcomes		BRCA mutation with RRBSO (n=329)	BO without BRCA mutation (n=3,290)	95% CI	BRCA mutation with RRBSO (n=329)	Intact ovaries without BRCA mutation (n=3,290)	95% CI
DEXA-scan
	Persons-years	1,663.82	28,518.41	-	1,663.82	28,441.12	-
	No of events	152	761	-	152	335	-
	Crude HR	3.02	1	2.53–3.60	7.54	1	6.20–9.17
	Adjusted HR*	1.58	1	1.26–1.98	6.03	1	4.64–7.86
Bisphosphonates
	Persons-years	2,489.93	33,611.39	-	2,489.93	30,572.15	-
	No of events	27	180	-	27	60	-
	Crude HR	1.93	1	1.28–2.90	5.91	1	3.73–9.37
	Adjusted HR*	0.64	1	0.40–1.04	2.41	1	1.04–5.59
No. of women with osteoporosis diagnosis		22	127		22	45
Bisphosphonates
	Persons-years	211.10	1,217.44	-	211.10	500.38	-
	No of events	8	65	-	8	21	-
	Crude HR†	0.68	1	0.33–1.43	0.97	1	0.43–2.22

RRBSO, risk-reducing bilateral salpingo-oophorectomy; BO, bilateral oophorectomy; CI, confidence interval; DEXA, dual-energy X-ray absorptiometry; HR, hazard ratio.

*Adjusted for history of breast cancer; †No covariates included due to low number of events.

Higher screening with DEXA was also observed among women with BRCA1/2 after excluding women with prior breast cancer (HR=2.73; 95% CI=2.12–3.52, compared to women without mutations who had BO; HR=6.22; 95% CI=4.76-8.12 compared to women without mutations with intact ovaries, Table S4).

3. HRT

Of the women with BRCA1/2 mutations who underwent RRBSO, 153 (46.5%) received HRT (Table S5), of whom 82 (53.6%) received oral formulations, 32 (20.9%) received transdermal estrogen, and 39 (25.5%) received both HRT formulations. Women who received HRT were younger on average (40.8 vs. 43.9 years, p<0.001) and were less likely to have previous breast cancer than those who did not use HRT (20.9% vs. 73.3%, p<0.001).

We did not observe significant differences in the hazard of fractures between the groups, adjusting for age and history of breast cancer (aHR=0.88; 95% CI=0.43–1.81; Table 5). However, women who received HRT were less likely to be diagnosed with osteoporosis (HR=0.35; 95% CI=0.13–0.95, when examining only women who had DEXA-scan). We could not include covariates for this analysis given the small number of events in the HRT group (n=5).

Table 5

Hazard of outcomes among women with BRCA mutations who received HRT after RRBSO versus the ones who did not

Outcomes		HRT (n=153)	No HRT (n=176)	95% CI
Bone fractures
	Persons-years	1,160.63	1,254.77	-
	No. of events	17	27	-
	Crude HR	0.68	1	0.37–1.25
	Adjusted HR*	0.88	1	0.43–1.81
Osteoporosis
	Persons-years	1,208.40	1,325.64	-
	No. of events	≤5	17	-
	Crude HR†	0.29	1	0.11–0.78
No. of women who had DEXA scan		59	93
Osteoporosis
	Persons-years	548.43	756.36	-
	No. of events	≤5	17	-
	Crude HR†	0.35	1	0.13–0.95

HRT, hormone replacement therapy; RRBSO, risk-reducing bilateral salpingo-oophorectomy; CI, confidence interval; HR, hazard ratio; DEXA, dual-energy X-ray absorptiometry.

*Adjusted for age and breast cancer; †No covariates included due to low number of events.

Considering HRT duration as a continuous variable (Table S6), we did not observe significant differences in the hazard of fractures or osteoporosis for each year of HRT use, adjusting for age and breast cancer history. However, for women without a mutation who underwent BO, each year of HRT use was associated with a small but significant reduction in the hazard of fractures (aHR=0.92; 95% CI=0.90–0.95), adjusting for age and history of breast cancer. Table S7 shows the hazard of the outcomes for women without BRCA1/2 mutations who had BO, comparing those who received HRT and those who did not.

DISCUSSION

In this retrospective population-based study, we observed an increased likelihood of being diagnosed with osteoporosis among women with BRCA1/2 mutations who underwent RRBSO prior to the average age of natural menopause compared to women without mutations with intact ovaries who had hysterectomy or salpingectomy. This was expected as a consequence of the surgical menopause. We also observed a higher likelihood of having a diagnosis of osteoporosis among women with BRCA1/2 mutations compared to women without mutations who had BO considering the entire study population, and also when looking only in those who had received DEXA-scans.

While the risk of fractures was not increased in women with BRCA1/2 mutations, they occurred earlier in this group. There was also an association between surgical menopause and risk of fractures by comparing women who had BO without mutations to women with intact ovaries who had hysterectomy or salpingectomy. In this analysis we had a larger sample size and a longer follow-up period, and thus an older population on average at the end of follow-up.

Women with BRCA1/2 mutations were more likely to be screened for bone loss after the surgery, but less than 50% of that study population had received a DEXA-scan by the end of the follow-up. The rate of bisphosphonates use among women who were diagnosed with osteoporosis was also low. While conclusions should be drawn cautiously with respect to HRT use, as our numbers were small, our findings suggest a decreased risk of receiving an osteoporosis diagnosis among women with BRCA1/2 mutations who were using HRT, and of lower risk of fractures among women without a mutation who had BO and received HRT. These findings are encouraging and suggest that HRT use is likely important to maintain bone health in this population.

By the end of our study, 14.5% of women with BRCA1/2 mutations who underwent RRBSO and had a DEXA-scan had been diagnosed with osteoporosis, a slightly higher rate than reported by Cohen et al. (9%) and Powell et al. (11.6%) for women undergoing premenopausal RRBSO [1213]. Our observed frequency is also more than double compared to the 6.1% prevalence of osteoporosis for Canadian women between 50–54 years [27]. We observed a similar frequency for the control groups (14.2% and 11.3%), suggesting that the group with intact ovaries may also have been at increased risk of osteoporosis, possibly due to the reduce age of onset of menopause following hysterectomy [28].

Women with BRCA1/2 mutations often have multiple risk factors for bone diseases, including previous chemotherapy and aromatase inhibitors use due to breast cancer, as many women are tested for BRCA1/2 mutations following a breast cancer diagnosis [11]. However, we report an increased risk of osteoporosis among women with BRCA1/2 mutations and no history of breast cancer compared to women without mutations with intact ovaries, indicating that surgical menopause is an important risk factor in this population, and follow-up care should focus on maintaining bone health.

The higher likelihood of having at least one DEXA-scan after surgery among women with BRCA1/2 mutations is also supported by the literature that previously described a more proactive attitude of women undergoing RRBSO towards their health [29]. Nonetheless, under half (46%) of women with BRCA1/2 mutations had a DEXA-scan after RRBSO, similarly to the reports from Chapman et al. and Garcia et al., ranging from 44% to 47% [1011].

This low rate of DEXA-scan after RRBSO in women with a BRCA1/2 mutation (<50%), and the low proportion of women with diagnosis of osteoporosis who filled prescriptions for bisphosphonates in our study population suggest that improving prevention of bone disease among these women is possible. Although routine screening of bone loss in women with BRCA1/2 mutation undergoing RRBSO has been recommended, there are no formal guidelines on the management of the adverse outcomes of premature surgical menopause in this population, and thus their follow-up care is heterogeneous [1011]. We recommend further research into the value of regular DEXA scanning in this population, particularly to examine whether DEXA scanning at 3 and 5 years post-surgery could reduce rates of undiagnosed osteoporosis in these patients. Improving our management of the health risks after RRBSO is crucial to prevent non-cancer outcomes and improve long-term survivorship.

This study had a larger sample size and a longer follow-up than most publications on the topic. Unfortunately, as many women were still younger than the average age of osteoporotic fractures by the end of the follow-up period, our study was underpowered to draw firm conclusions on the risk of fractures, despite the inclusion of controls in a 10:1 rate. While there was no increased hazard of bone fractures among women with BRCA1/2 mutations, the observed association between surgical menopause and fractures among women who underwent BO suggests that with longer follow-up and a larger study population, we may see a similar increased risk among women with BRCA1/2 mutations undergoing RRBSO. We were also underpowered to draw firm conclusions on the role of HRT use, and we did not analyze the use of progestogen-only preparations due to small numbers. The duration of HRT varied largely in our population, but the mean use (up to 3.5 years) was well under the recommendation of maintaining HRT treatment until the average age of natural menopause [30]. This further impaired our estimates of the impact of HRT on bone health.

This study was also missing potentially relevant information on other risk factors for osteoporosis and fractures (e.g., parathyroid disease, low body mass index, etc.), and on other medications or preventive strategies to preserve bone health, such as use of vitamin D and calcium. Due to the low rates of DEXA screening (<50%), our overall incidence of osteoporosis was probably underestimated. The high prevalence of breast cancer in our study group is another limitation of this study and our analyses stratified on history of breast cancer had low statistical power as many women were tested for BRCA1/2 mutations due to a breast cancer diagnosis. Breast cancer treatments, especially aromatase inhibitors and tamoxifen, are risk factors for bone loss and osteoporotic fractures [31]. However, a relatively high proportion of BRCA1/2 mutation associated breast cancers are triple negative [3233]. Thus, many of these women will not receive tamoxifen or aromatase inhibitors. We were also unable to include a control group of women with BRCA1/2 mutations who did not undergo RRBSO, as more than 70% of women with BRCA1/2 mutations who were 40 years of age or older had RRBSO in BC [34]. Our second control group underwent hysterectomy or salpingectomy which may have reduced the age of onset of menopause of these women [2835]. Thus, we may have underestimated the hazards of the outcomes for women with BRCA1/2 mutations in relation to the general population who have not undergone any gynecologic surgery. The presence of unknown mutation carriers in the control groups may be another source of conservative bias, trending the results towards a null finding.

In conclusion, women with BRCA1/2 mutations who underwent RRBSO had a higher risk of osteoporosis than both control groups but were not at increased risk of bone fractures during our follow-up period. The low rate of DEXA-scan and of bisphosphonates use among women with diagnoses of bone loss indicates that we can better preserve bone health in this population with focused care. Longer follow-up is needed to better investigate the risk of bone fractures and the performance of HRT in this population.