Salpingectomy With Delayed Oophorectomy Versus Salpingo-Oophorectomy in BRCA1/2 Carriers: Three-Year Outcomes of a Prospective Preference Trial.

Study design: J.I., M.H.D.V.B., M.A.‐D.J., A.H.E.M.M., J.B.P., J.B., N.H., R.P.M.G.H., J.A.D.H. Data collection: M.H.D.V.B., M.P.S., M.G.H., M.A.‐D.J., A.H.E.M.M., J.B.P., H.C.V.D., M.J.E.M., R.T., R.P.Z., K.N.G., B.F.M.S., M.M.A.B.‐V.Z., M.C.V., J.M.J.P., L.R.C.W.L., M.J.A.A., S.F.P.J.C., J.A.D.H. Data analysis (and access to the underlying data): M.H.D.V.B., J.I., J.D.H. Data interpretation: M.H.D.V.B., M.P.S., J.I., T.V.G., M.G.H., M.A.‐D.J., A.H.E.M.M., J.B.P.,J.B., H.C.V.D., M.J.E.M., R.T., R.P.Z., K.N.G., B.F.M.S., M.M.A.B.‐V.Z., M.C.V., J.M.J.P., L.R.C.W.L., M.J.A.A., S.F.P.J.C., N.H., R.P.M.G.H., J.A.D.H. Writing: M.H.D.V.B. Extensive revision of the manuscript: M.P.S., J.I., T.V.G., M.G.H., M.A.‐D.J., A.H.E.M.M., J.B.P.,J.B., H.C.V.D., M.J.E.M., R.T., R.P.Z., K.N.G., B.F.M.S., M.M.A.B.‐V.Z., M.C.V., J.M.J.P., L.R.C.W.L., M.J.A.A., S.F.P.J.C., N.H., R.P.M.G.H., J.A.D.H.

The Medical Ethics Committee of Arnhem‐Nijmegen approved the study (number 2014–1269).

Details of the study design and population have been published [ 11 , 12 ]. Briefly, the nationwide, multicentre, prospective TUBA study recruited BRCA1/2 ‐PV carriers, aged 25–40 ( BRCA1 ) or 25–45 ( BRCA2 ), who had no (future) child wish, without a history of tubo‐ovarian cancer and without any malignancy at enrolment in thirteen Dutch hospitals. Treatment allocation was based on patients' preference; either the novel RRS with DO or RRSO according to the guideline [ 13 ]. RRS could be performed from the age of 25 and at the maximum age of 40 ( BRCA1 ) or 45 ( BRCA2 ). After RRS, DO was performed between age 35 and 45 ( BRCA1 ) or 40 and 50 ( BRCA2 ). RRSO could be performed between age 35 and 40 ( BRCA1 ) or 40 and 45 ( BRCA2 ). Pre‐surgery all participants underwent transvaginal ultrasound and measurement of CA‐125. During all surgeries, an abdominal washing was taken. All resected Fallopian tubes were embedded according to the SEE‐FIM protocol and assessed by a specialised gynaecopathologist [ 14 ]. The Medical Ethics Committee of Arnhem‐Nijmegen approved the study and each participant provided written informed consent. To ensure safety of the participants, we installed an independent, multidisciplinary data safety monitoring board and a safety rule that will flag in case of a higher number of tubo‐ovarian cancers than may be expected based on age, BRCA ‐PV type and duration of follow‐up. The outcome measures and procedure of data collection were described before [ 12 ]. In short, the primary outcome of menopause‐related QoL was assessed by the validated Greene Climacteric Scale (GCS) and compared between treatment arms (RRS with DO versus RRSO without HRT) [ 15 ]. In the GCS, which can range from 0 to 63, a higher score reflects more climacteric symptoms. The secondary outcomes, health‐related QoL, sexual functioning, sexual distress, cancer worry, decisional conflict and decisional regret, were assessed using validated questionnaires for each of these outcomes [ 16 , 17 , 18 , 19 , 20 ]. For sexual functioning, a score of 26.55 or less on the female sexual functioning index (FSFI) is defined as sexual dysfunction [ 21 ]. Surgical and histopathological data were collected via electronic case report forms. Here, we included data pre‐surgery and 3 months, 1 and 3 years of post‐surgery. Baseline characteristics of the participants were compared using descriptive statistics, Mann–Whitney‐tests and χ 2 ‐tests. The primary objective was to determine the mean difference in total GCS‐score in participants who underwent RRS compared to participants who underwent RRSO without using HRT after surgery. Participants' results were assessed according to their HRT use at that time, thus, for example, someone could be included in the one‐year group with HRT and switch to the without HRT group at 3 years. A linear mixed model was used to assess the change from baseline GCS‐score, comparing RRS to RRSO without HRT, in relation to time since surgery. In this model, we adjusted for the baseline GCS‐score, SF‐36's mental and physical components, age at first surgery and type of BRCA ‐PV ( BRCA1 versus BRCA2 ). A random effect for hospital was added as well as an unstructured correlation matrix for the within‐participants association between repeated measures (3 months, 1 and 3 years of post‐surgery). Additionally, the model included a three‐way interaction for surgery type by follow‐up time point by baseline GSC‐score and their corresponding two‐way interactions. As a secondary analysis, the same model was used for comparing RRS versus RRSO with HRT. For all other secondary outcomes, the same model was used as well, in which we also adjusted for the baseline value of the specific questionnaire. A two‐sided significance level of 0.05 was set, expect for the primary outcome which was corrected with Bonferroni for multiple testing (0.05/3). Correction for multiple testing was not done. Data were analysed using SPSS for Windows version 27 (IBM Corp).

A total of 577 participants were included in the TUBA study between January 2015 and November 2019. RRS with DO was chosen by 413 participants (71.6%) and RRSO by 164 (28.4%). Until time of analysis in April 2023, 410 (99.3%) and 160 (97.6%) participants underwent RRS and RRSO, respectively (Figure  1 ). Second‐stage oophorectomy was performed in 74 participants of which 15 underwent oophorectomy within 3 years after RRS. Flowchart of participants. DO delayed oophorectomy; HRT hormone replacement therapy; RRS risk‐reducing salpingectomy; RRSO risk‐reducing salpingo‐oophorectomy. In total, 51.5% carried a BRCA1 ‐PV and 48.5% a BRCA2 ‐PV. Mean age (standard deviation, SD) at surgery was 36.6 (2.8) and 39.4 (3.6) years for BRCA1/2 ‐PV carriers, respectively. Compared to RRSO, participants undergoing RRS were younger ( p  < 0.001) and had more often a BRCA2 ‐PV ( p  = 0.006). Table  1 provides all baseline characteristics; the characteristics stratified by HRT use after surgery are provided in Table  S1 . The most frequently used type of HRT was Tibolone 2.5 mg (Table  S2 ). Details of excluded participants are provided in Table  S3 . Baseline characteristics. Abbreviations: No., number of patients; RRS, risk‐reducing salpingectomy; RRSO, risk‐reducing salpingo‐oophorectomy; SD, standard deviation. A total of 6.4% of the participants after RRSO did use HRT at 1‐year and did not use HRT at 3 years of follow‐up, 2.4% did not use HRT at 1‐year and did use HRT at 3 years. For the ones after RRS, 8, 10 and 17 participants were using hormones at the 3 months, 1‐year and 3 years of follow‐up respectively, most often for bleeding control. 1.4% did use HRT at 1‐year and did not use HRT at 3 years, 3.3% did not use HRT at 1‐year and did use HRT at 3 years. 3 years of post‐surgery, menopause‐related QoL among participants without HRT showed an observed mean (SD) increase from baseline GCS‐score of 1.8 (6.5) points after RRS and 3.9 (8.5) points after RRSO. The adjusted mean difference (MD) between treatment arms was 4.3 (95% confidence interval (CI) 2.1–6.5, p  < 0.001) points in favour of the RRS group, as presented in Table  2 . After 3 and 12 months, these MDs were 7.9 (95% CI 5.9–9.8) and 8.5 (95% CI 6.5–10.5) points, respectively. The RRS group had significant lower (better) scores at 3 years compared to the RRSO group across all domains of the GCS, with an exception for the depression domain (Table  S4 ). Figure  2 depicts the proportions of participants that decreased, stayed stable, or increased in total GCS‐score per time period. Results were similar when only participants who completed 3 years of follow‐up were included and when only participants without second‐stage oophorectomy within 3 years after RRS were included (data not shown). Total Score of the Greene Climacteric Scale a . Note : Adjusting factors in the analysis: baseline GCS‐score, SF‐36's mental and physical components, age at first surgery, type of BRCA‐PV, hospital, an unstructured correlation matrix for the within‐participants association between repeated measures (3 months, 1 and 3 years of post‐surgery), three‐way interaction for surgery type by visit follow‐up time point by baseline GSC‐score and their corresponding two‐way interactions. Abbreviations: CFB, change from baseline; GCS, Green Climacteric Scale; HRT, hormone replacement therapy; IQR, interquartile range; MD, mean difference; CI, confidence interval; No., number of patients; SD, standard deviation; RRS, risk‐reducing salpingectomy; RRSO, risk‐reducing salpingo‐oophorectomy. The Greene Climacteric Scale ranges from 0 to 63; a higher score represents more climacteric symptoms. Baseline value of participants without HRT at 3 months. All participants after RRS (with and without hormones) and only the ones using HRT after RRSO at that time point are included. The number of participants using hormones after RRS are 14, 8, 10 and 16 at baseline, 3 months, 1 year and 3 years after surgery, respectively. Baseline value of participants with HRT at 3 months after RRSO. Visualisation of the proportions of participants that decreased, stayed stable or increased in total GCS‐score per time period.A better score means a lowering in total GSC‐score, a worse score means a higher in total GCS‐score. GCS Greene Climacteric Scale; HRT hormone replacement therapy; RRS risk‐reducing salpingectomy; RRSO risk‐reducing salpingo‐oophorectomy. When comparing all participants after RRS to those who used HRT 3 years after RRSO, the adjusted MD from baseline‐GCS score was 2.4 (95% CI 0.8–3.9, p  = 0.002) points in favour of RRS. Similar differences were found at 3 and 12 months follow‐up (Table  2 ). In the domains of the GCS after 3 years, significant differences between treatment arms were observed in the anxiety, vasomotor and sexual domains in favour of RRS (Table  S4 ). The proportions of participants that decrease, stay stable or increase for the different domains of the GCS per time period are visualised in Figure  S1 . The results regarding the secondary outcomes are provided in Table  3 . Health‐related QoL did not significantly differ between treatment arms 3 years of post‐surgery, irrespective of HRT use. Sexual functioning was significantly lower among participants after RRSO without HRT compared to those after RRS at all follow‐up points. Comparing RRSO with HRT versus RRS, a significant larger decrease from baseline score was found in sexual functioning 1‐year post‐surgery and not at 3 years. More sexual distress was found in the RRSO group without HRT, compared to the RRS group at all time points. The difference was most pronounced at 1‐year follow‐up, but still present at 3 years of follow‐up. Comparing RRSO with HRT to RRS, no significant differences in sexual distress were found. Cancer worry, decisional conflict and decisional regret did not statistically significantly differ between the RRS and RRSO without HRT groups. However, comparing RRS to RRSO with HRT, cancer worry decreased more in the RRSO group with an increasing difference over time. Decisional conflict and decisional regret were lower after RRSO with HRT than after RRS. Secondary outcomes. Abbreviations: CFB, change from baseline; CWS, Cancer Worry Scale; DCS, Decisional Conflict Scale; DRS, Decision Regret Scale; FSDS, Female Sexual Distress Scale; FSFI, Female Sexual Functioning Index; HRT, hormone replacement therapy; IQR, interquartile range; MD, mean difference; CI, confidence interval; No., number of patients; SD, standard deviation; RRS, risk‐reducing salpingectomy; RRSO, risk‐reducing salpingo‐oophorectomy; SF‐36, Short Form 36. Baseline value of women without HRT at 3 months post‐surgery. All individuals after RRS (with and without HRT) and only the women using HRT after RRSO at that time point are included. Baseline value of individuals with HRT at 3 months after RRSO. In Table  S5 , the domains of the female sexual functioning index (FSFI) are presented showing significant lower scores for RRSO without HRT compared with RRS in all domains and at all time points, except for the satisfaction domain 3 years of post‐surgery. Comparing participants with HRT after RRSO to participants after RRS, a significant larger decrease in sexual functioning was found 1‐year post‐surgery only. This difference was present in the arousal, lubrication and pain domains. The proportions of the different groups that sexually dysfunction are depicted in Figure  S2 . A total of 5% of all surgeries were combined with another procedure. Median time of surgery was 44 (interquartile range (IQR) 34–57) minutes for RRS, 57 (IQR 48–72) minutes for RRSO and 51 (IQR 40–62) minutes for oophorectomy. The median time interval between RRS and DO was 57 (range 1–85) months. The majority was discharged on the day of surgery: 83.2%, 75.6% and 85.1%, respectively. The median blood loss was 0 mL (range 0–350 for RRS, 0–250 for RRSO and 0–150 for oophorectomy). The complication rate of events definitely or probably related to the surgery was 5.4%, 4.4% and 2.7% for RRS, RRSO and oophorectomy, respectively (Table  S6 ). At RRS, two serous tubal intraepithelial carcinomas (STICs) and one carcinoma were detected. One patient with STIC at RRS had an invasive carcinoma at subsequent oophorectomy. All other 73 s‐stage oophorectomy specimens and abdominal washings were without abnormalities. At RRSO, one STIC, one suspicion for small focus high‐grade serous carcinoma (HGSC), and three carcinomas were found. Until April 2023, no carcinomas have been diagnosed among the participants with normal pathology at first surgery. Of those with abnormalities at first surgery, four had recurrent HGSC of which one died due to HGSC. Details are provided in Table  4 . Histopathological outcomes. Abbreviations: HGSC, high‐grade serous carcinoma; NA, not applicable; PV, pathogenic variant; RRS, risk‐reducing salpingectomy; RRSO, risk‐reducing salpingo‐oophorectomy; STIC, serous tubal intraepithelial carcinoma. Received six cycles of monotherapy with carboplatin chemotherapy. Plus unilateral oophorectomy because of macroscopic lesion. Did not receive chemotherapy after staging surgery. Until April 2023, a total of 29 participants developed breast cancer after either RRS or RRSO (5.1%); 21 were primary breast cancers and eight were recurrences. Of the 21 primary breast cancer patients, 14 underwent RRS and seven underwent RRSO ( p  = 0.665). Median age at primary breast cancer after surgery was 38.7 (range 28.9–43.7) for BRCA1 and 41.9 (33.3–47.4) for BRCA2 ‐PV carriers. Details are presented in Table  S6 .

The present analysis of the TUBA study until 3 years of post‐surgery demonstrates that individuals after RRS experience better menopause‐related QoL compared to those after RRSO, also when using HRT after RRSO. Without HRT, climacteric symptoms are less pronounced at 3 years than in the first year after RRSO. The main strengths of our study are related to its design; prospective, multicentre and nationwide with a large number of participants and low number of drop‐outs reducing the risk of selection bias. Also, standardised and validated questionnaires were used. The incorporated strict safety measures enabled live safety monitoring of the study. The most important limitation, although well‐considered [ 22 , 23 ], is the non‐randomised design which resulted in an imbalance between treatment arms. Individuals choosing RRS with DO were probably overrepresented because RRSO can also be performed in non‐participating hospitals. Individuals wanting RRS had to be referred to a participating hospital, because RRS outside the setting of a clinical trial is strongly discouraged in the Netherlands. This is the first study presenting 3 years of follow‐up data after RRS with DO. Interestingly, menopause‐related QoL was partly restored after 3 years when compared to 3 and 12 months post‐surgery, particularly in individuals without HRT after RRSO. The increase of 4.3 points 3 years after surgery, as we found, means that individuals are experiencing a mild to moderate worsening in up to four menopausal symptoms. Moreover, the average increase in total GCS‐score compared to baseline was smaller than five points. This is similar to the findings in the general Dutch population after natural, non‐premature menopause [ 15 ]. This might be explained by habituation or acceptance of climacteric symptoms over time, or an actual decrease in symptoms over time. Another explanation might be aging across all groups, making differences less pronounced between groups as scores in the RRS arm increase. The finding that climacteric symptoms are comparable to natural menopause 3 years after RRSO can be important in counselling the BRCA1/2 ‐PV population. Individuals who used HRT after RRSO also experienced more menopausal symptoms than individuals undergoing RRS, but this difference was less distinct. This is consistent with other studies suggesting that HRT alleviates but does not eliminate symptoms [ 6 , 24 ]. Across all domains of the GCS, the benefit of HRT seemed to become smaller over time. An explanation for this might be that those with many symptoms have started using HRT during follow‐up, although the proportions of participants starting/stopping HRT were small. Noteworthy, the most commonly used type of HRT was Tibolone 2.5 mg, which was defined as first choice in the Dutch guideline at the time the majority of our participants underwent surgery. Nowadays, however, (transdermal) oestradiol suppletion is considered more effective with less side‐effects than Tibolone in reducing vasomotor symptoms [ 25 , 26 ]. Besides administration route and type of oestrogen, the optimal dosage of HRT is topic of debate [ 27 ]. Our results indicate that HRT used by the participants in our study is not as effective as endogen oestrogens. This supports the need for research regarding optimisation of HRT (composition, dosage, route of administration). It also makes RRS with DO an advantageous strategy, not solely for patients with a contra‐indication for HRT. Importantly, RRS with DO is only an alternative in clinical practice as soon as safety has been proven. Besides menopause‐related QoL, sexual health was assessed as secondary outcome of the TUBA study. At 3 years of post‐surgery, RRS appeared superior to RRSO without HRT in all, but the satisfaction domain of the FSFI. Contrastingly, no differences were found 3 years after RRS compared to RRSO with HRT. These results suggest that HRT can alleviate disturbances in sexual functioning due to menopause, a finding which is consistent with previous literature [ 28 , 29 ]. The main reason why participants did not use HRT in our study was a previous diagnosis of breast cancer. The current guidelines advice against HRT after breast cancer, although this is topic of discussion particularly for individuals with triple negative breast cancer in their history. Breast cancer (treatment) is associated with reduced sexual activity which might have played a role as well [ 29 ]. Again, this emphasises the importance of exploring RRS with DO. Another important finding is that sexual functioning did not deteriorate three versus 1‐year after surgery, as observed across all domains. Although theoretically, one could reason that oestrogen‐dependent symptoms like pain (due to vaginal dryness and/or atrophy), lubrication and sexual desire would increase with prolonged oestrogen depletion, as we observed in the 1‐year results as well [ 12 ]. The finding that sexual functioning did not worsen over time might be explained by a stabilised relative shortage of testosterone. Especially as sexual desire, arousal and overall satisfaction seem related to levels of testosterone [ 30 ]. Our observed differences between the RRSO groups with and without HRT might be partly explained by the weak androgenic actions of Tibolone. In the Netherlands, testosterone therapy is not prescribed after RRSO due to the scarcity of current literature and because it is not registered for females. Nonetheless, the first results regarding testosterone therapy after RRSO seem promising [ 30 , 31 , 32 ]. Another interesting secondary finding is that cancer worry decreased more in the RRSO with HRT group versus the RRS group and further decreased with time, whereas this difference was not found in the comparison between RRSO without HRT versus RRS. As the main reason for not taking HRT was a previous breast cancer, this finding may be explained by the that fact that cancer worry not only concerns ovarian cancer. The ones with HRT and thus without previous cancer may worry more about getting cancer than the ones who already have been confronted with cancer. Further, decisional conflict and decisional regret were lower among the ones after RRSO with HRT than the ones after RRS, although median scores were similar. Again, differences were not found when comparing RRSO without HRT versus RRS. Potentially, the amount of climacteric symptoms (highest in RRSO without HRT and lowest in RRS) is associated with satisfaction with their choice and thus women with less symptoms have less decisional conflict and regret.

This 3 years of analysis of the TUBA study showed that RRS is superior to RRSO regarding menopause‐related QoL, also when compared to individuals using HRT after RRSO. 3 years of post‐surgery, differences were less pronounced than during the first‐year post‐surgery. The use of HRT after RRSO seems to have a beneficial effect on sexual functioning. Prior to clinical implementation of RRS with DO must be assessed further as is currently being done in the TUBA‐WISP II ( NCT04294927 ), SOROCk ( NCT04251052 ) and PROTECTOR (ISRCTN25173360) trials.

Tubo‐ovarian cancer has a poor prognosis [ 1 ]. For BRCA1 /2‐pathogenic variant (PV) carriers, a risk‐reducing salpingo‐oophorectomy (RRSO) is recommended to lower the risk of tubo‐ovarian cancer [ 2 ]. Although highly effective, RRSO induces an early/premature surgical menopause as international guidelines advise to perform RRSO at age 35–40 ( BRCA1 ) or 40–45 ( BRCA2 ) [ 3 ]. Early/premature menopause is associated with negative short‐term effects such as vasomotor complaints, sleep disturbances and sexual problems, and long‐term effects like a potential increased risk of cardiovascular disease, osteoporosis and neurocognitive decline [ 4 , 5 ]. After RRSO, hormone replacement therapy (HRT) is advised until the age of 50 years to alleviate these effects, unless contraindicated by for example a history of breast cancer or thrombosis [ 6 , 7 ]. It now is well established that most tubo‐ovarian cancers originate from fallopian tube epithelium. Therefore, a novel strategy of risk‐reducing salpingectomy (RRS) with delayed oophorectomy (DO) can be an attractive option [ 8 , 9 , 10 ]. It is aimed to prevent tubo‐ovarian cancer without inducing early/premature menopause and its consequences. In the TUBA study, RRS with DO and RRSO were primarily compared on menopause‐related quality of life (QoL) [ 11 ]. Results until one‐year post‐surgery showed a significant better menopause‐related QoL and sexual functioning among BRCA1/2‐ PV carriers who chose to undergo RRS compared to BRCA1/2 ‐PV carriers who chose RRSO, even with the use of HRT after RRSO [ 12 ]. Currently, RRS with DO for the prevention of tubo‐ovarian cancer should not be performed outside the setting of a clinical trial because longer follow‐up data after RRSO versus RRS with DO are not yet available. Also, most importantly, the effectiveness and safety of RRS with DO is still under investigation ( NCT04294927 and NCT04251052 ). Here, we present results of the TUBA study up till 3 years of post‐surgery, as a follow‐up of the previously reported one‐year results [ 12 ]. Primarily, we investigated menopause‐related QoL and secondary health‐related QoL, sexual functioning, sexual distress, cancer worry, decisional conflict and decisional regret.

The authors declare no conflicts of interest.

Table S1. Figure S1. Change in the domains of the Greene Climacteric Scale over time per treatment group. Figure S2. Proportions of individuals that scored less than the validated cut‐off of 26.55 points on the Female Sexual Functioning Index, indicating sexual dysfu.