Study design and setting

This is a multicentre cohort study comparing 105 high-risk premenopausal women up to 50 years of age who plan to undergo RRBSO with 105 age-matched, premenopausal women who do not plan to undergo oophorectomy or pregnancy in the next 2 years. Study recruitment commenced in April 2013 and is ongoing, enrolling women from eight public and four private hospitals in Australia (Victoria and New South Wales). The projected timeline for recruitment is 3–4 years. All subjects will be followed up for 2 years from the baseline visit.

Eligibility criteria and recruitment

Eligible women will be premenopausal and age up to 50 years with regular menstrual periods (if intact uterus), no vasomotor symptoms and a follicular stimulating hormone level of <15 IU/L on days 2–6 of the menstrual cycle. Exclusions include <3 months since pregnancy and lactation, abnormal uterine bleeding, or use of antioestrogens such as tamoxifen.

Exposed women are defined as those at high inherited risk of ovarian cancer due to confirmed presence of a gene mutation (carriers of BRCA1, BRCA2, BRIP1, RAD51C, Lynch syndrome) or high familial risk of ovarian cancer and planning to undergo RRBSO. Those with a personal history of breast cancer will be included if they remain premenopausal and are not currently taking endocrine therapy. Participants will be recruited via clinician referrals from familial cancer, high-risk breast cancer and menopause clinics, referrals from public and private gynaecology oncologists and surgeons, and via targeted advertising through mainstream and electronic media (eg, newspaper and television reports, cancer registries, cancer support newsletters and websites). Each referral will be processed on receipt and eligibility screening and subsequent recruitment will be scheduled to occur within 8 weeks prior to the RRBSO surgery date. The baseline visit will be scheduled for the early follicular phase (days 2–6) of the subject’s menstrual cycle prior to RRBSO.

Women in the comparison group are defined as 1:1 individually age-matched (within ± 5 years), premenopausal women who are not planning oophorectomy or pregnancy within the next 2 years. Comparison subjects can be low-risk or relatives of high-risk women who do not carry a gene mutation. The comparison group will be recruited via mainstream and electronic media advertising to the general public (eg, clinical research recruitment websites, hospital and university staff newsletters and websites), and by asking recruited cases to recommend the study to relatives and friends. Recruitment will be scheduled to occur during the early follicular phase (days 2–6) of the subject’s menstrual cycle, and within 8 weeks after eligibility screening. Women in the comparison group withdrawing prior to the 6-month follow-up will be replaced by women of similar baseline age.

Participant timeline

A schedule of study assessments and measurements is presented in table 1. After informed consent and eligibility, baseline data will be collected including obstetric, gynaecological, medical and surgical history, current medications, risk factors for fracture (www.shef.ac.uk/FRAX), risk factors for cardiovascular disease including blood pressure, circulating cholesterol and lipids and C reactive protein (www.heartfoundation.org.au/SiteCollectionDocuments/austcardiovascular-risk-charts.pdf), tobacco, drug and alcohol use, methods of contraception, methods of breast cancer surveillance, germline mutation type, and family cancer history. In those with a history of breast cancer, stage and grade of tumour and treatment history will be recorded. Blood pressure, weight, height and waist-hip ratio will be measured and repeated at selected follow-up visits (see table 1). A blood sample will be taken after an overnight fast to measure sex steroid concentrations, lipids and cardiovascular risk factors, which will be repeated at 1-year and 2-year follow-up. A urine sample will be collected for comparison subjects to exclude pregnancy (see table 1). Medications used in the 3 months prior to enrolment and throughout the 2-year follow-up period will be recorded and include those related to bone health (calcium, vitamin D and antiresorptive agents), depression, anxiety, HRT, non-hormonal treatments for vasomotor symptoms, contraception and insomnia. Adverse events (AEs) (changes in physical and psychological health from baseline) will be monitored over the 2-year follow-up period. Questionnaires to measure sexual function, menopausal symptoms, menopause-related QOL, sleep quality, depression and anxiety will be administered at baseline and all postbaseline study visits (see table 1). All participants will undergo DXA scans within 3 months of RRBSO (baseline) and again at 1 and 2 years postbaseline.

Participants will be free to withdraw at any time, either through formal revocation of consent (patient preference) or via ceased communication with the recruitment site (lost to follow-up). All collected data will be included in the analyses and any discrepancies related to data collected prior to withdrawal will be queried with the subject up to 1 month after the withdrawal date. In order to minimise missing data, participants who cannot complete all scheduled visits will be invited to continue in the study despite missing visits. Those who withdraw at a scheduled time-point or who are unable to attend the recruitment site at a scheduled time-point will be offered the opportunity to complete any time-point data that can be captured remotely via telephone and email correspondence, including questionnaire, medications and AE data.

Outcomes

The primary outcome is the change in sexual function following RRBSO, as measured by the FSFI.19 Sexual function was selected because it is a patient priority and in young women and because previous prospective studies suggest that sexual function may be permanently impaired following RRBSO.13–15 Secondary outcomes include menopause-related QOL, menopausal symptoms, sleep quality, depression and anxiety, markers of cardiovascular disease and pre-diabetes, and bone density and markers of bone turnover. All measures will be collected at baseline (prior to RRBSO in the intervention group) with follow-up measures scheduled as per table 1.

Measurement of sexual function

Subjects will be asked whether they currently have a sexual partner and whether sexual problems for the partner impact on their sexual activity. We will use standardised and validated questionnaires to measure three aspects of sexual function:

1. The FSFI is a widely used brief self-report questionnaire that assesses six separate dimension: desire, arousal, lubrication, orgasm, satisfaction and pain, and it also provides a total score.20 The FSFI has been validated in a large group of women with sexual arousal disorder versus age-matched controls19 and in cancer survivors.21 It has demonstrated a high level of acceptability, reliability and validity in both cancer and non-cancer populations.22 The FSFI has a high internal consistency and test–retest reliability, and differentiates well between sexually dysfunctional and non-dysfunctional women, and is highly sensitive in discriminating between clinical and non-clinical populations.20 FSFI scores (primary) will indicate which domains of sexual function are affected by RRBSO. The FSFI takes <5 min to complete.

2. The revised Female Sexual Distress Scale (FSDS-R)23 is a brief (13 items) questionnaire that measures the extent to which reduced sexual desire causes distress in women. Distress is a key feature of hypoactive sexual desire disorder (HSDD). The FSDS-R has good discriminant validity, test–retest reliability and internal consistency in measuring sex-related personal distress in women with HSDD.24 The scale also has good content validity (relevance, clarity, comprehensiveness).25

3. The Sexual Activity Questionnaire (SAQ) evaluates sexual pleasure (desire, enjoyment and satisfaction), discomfort (vaginal dryness and dyspareunia) and habit (frequency of sexual activity compared with usual activity).26 The SAQ has been validated in patients with breast cancer and in high-risk women.27 It is quick and easy to administer and has good face validity discriminating between the sexual functioning of premenopausal and postmenopausal women27

Measurement of circulating testosterone and SHBG

Circulating testosterone concentrations are reduced following RRBSO, and this may contribute to sexual function, although findings are conflicting.15 Circulating concentrations of total testosterone and the testosterone precursors, androstenedione and 5-dehydroepiandrosterone (DHEA), and sex hormone binding globulin (SHBG) will be measured at baseline and follow-up, as per table 1. Fasting blood samples will be immediately centrifuged to isolate plasma, and stored at −80°C. Because conventional radioimmunoassays lack sensitivity at low androgen concentrations found after surgical menopause,28 we will use liquid chromatography tandem mass spectrometry (LC-MS/MS), using a Sciex API 5500Q instrument by CPR Pharma Services (Adelaide, South Australia, Australia). The lower limit of measurement is 0.025 ng/mL for testosterone, 0.05 ng/mL for androstenedione and 0.5 ng/mL for DHEA. The intra-assay coefficients of variation are low (<5%) at 1 nM. SHBG concentrations will be measured using a non-competitive liquid-phase Radio-immuno assay (68562, Orion Diagnostica, Espoo, Finland). We will calculate free testosterone concentrations using measured total testosterone and SHBG concentrations.28

Measurement of menopausal symptoms and menopause-related QOL

Menopausal symptoms will be measured using the Greene Climacteric Scale, which measures the frequency and severity of menopausal symptoms in psychological somatic and vasomotor domains.29 The scale is widely used and has been validated in an Australian population.30

Menopause-related QOL will be measured using the Menopause-Related Quality of Life Intervention (MENQOL-Intervention) questionnaire.31 The MENQOL-Intervention covers four domains—vasomotor, physical, psychosocial and sexual—and includes a global QOL item. It is widely used internationally and its strength and validity have recently been confirmed in a large population-based sample of midlife women32 and in breast cancer survivors.33 The questionnaire tests the impact of an intervention, such as oophorectomy, on symptoms. Each takes less than 5 min to complete. Use of HRT and of non-hormonal medications for vasomotor symptoms will be recorded, and subjects will be asked about decision making around HRT.

Measurement of sleep quality

Sleep quality will be measured using the Pittsburgh Sleep Quality Index (PSQI),34 a widely used and validated measure of sleep quality.35 The questionnaire measures subjective sleep quality, latency, duration, habitual sleep efficiency, sleep disturbances and medications, and daytime dysfunction. The PSQI takes <5 min to complete.

Measurements of depression and anxiety

Depression and anxiety will be measured using the Center for Epidemiologic Studies Depression (CES-D)36 and the Generalised Anxiety Disorder (GAD-7)37 scales, respectively. The CES-D measures the frequency of depressive feelings and behaviours experienced in the past week and includes 20 items that are assigned scores ranging from 0 to 3. The final CES-D score (0–60) is the sum of the 20 items and a score of ≥16 points is indicative of depression. The GAD-7 measures the frequency of GAD symptoms in the past fortnight and includes seven items that are assigned scores ranging from 0 to 3. The final GAD-7 score (0–21) is the sum of the seven items and scores of 5–9, 10–14 and 15–21 are indicative of mild, moderate and severe anxiety, respectively. Each questionnaire takes <5 min to complete.

Measurement of cardiovascular disease risk

Resting blood pressure, weight, height and waist/hip ratio will be measured as per table 1. Fasting blood samples will be analysed for serum insulin, glucose, triglycerides, cholesterol, high-density lipoprotein (HDL-C), low-density lipoprotein (LDL-C) and high-sensitivity C reactive protein, which predicts cardiovascular disease in healthy women.38 These have not previously been prospectively measured in surgical menopause. Homeostasis model assessment (HOMA) will be calculated by fasting insulin (μU/mL) × fasting glucose (mM)/22.5. Insulin resistance will be defined by a HOMA reading >488. The presence of diabetes or pre-diabetes will be established by fasting glucose and haemoglobin A1c (HbA1c). An HbA1c of >6.5% will be used as a cut-off for the diagnosis of diabetes as recommended by the American Diabetes Association (www.diabetes.org) and carried out in National Association of Testing Authorities accredited laboratories using standard equipment.

Measurement of bone density and markers of bone turnover

We will measure fasting serum albumin, creatinine, calcium and phosphate, and circulating markers of bone turnover (beta C terminal telopeptide of type I collagen (beta CTX) and N terminal propeptide of type 1 collagen).39 We will measure 25-OH vitamin D2 and D3 using LC-MS/MS and intact parathyroid hormone immunoassay. The bone turnover markers are independent predictors of fracture risk and will supplement bone mineral density (BMD) data. Regional BMD (at the lumbar spine and hip) will be measured using dual-energy X-ray absorptiometry (DXA) within 3 months of RRBSO (baseline). We will also use DXA to measure total body bone mineral content, which will increase power to detect bone loss.40 DXA measurement location will be included as a covariate and change from baseline will be used in outcome analyses. Fracture risk assessment using a FRAX tool (www.shef.ac.uk/FRAX) will only be applied to women aged over 40 years.

Sample size

We will recruit 105 high-risk premenopausal women planning RRBSO and 105 age-matched (±5 years) women in the comparison group. The primary outcome is sexual dysfunction (FSFI score <26.55),41 where lower FSFI scores indicate worse sexual function. Based on large population-based surveys, we have assumed that around 24% of premenopausal women will have sexual dysfunction (FSFI below clinical cut-off) at baseline,42 43 and that the proportion with sexual dysfunction does not differ between the high-risk women or those in the comparison group, or between age groups, and that it does not change over the 2-year follow-up period in the comparison group. Sample size calculations are based on comparing the proportion of women with sexual dysfunction in the RRBSO group with the comparison group at the 2-year follow-up. A sample size of 89 women per group will provide 80% power at a two-sided 5% level of significance to detect a 21% difference (from 24% in the comparison group to 45% in the RRBSO group), or 105 per group when allowing for 15% loss to follow-up at 2 years. This difference in the proportion of women with sexual dysfunction is clinically significant.44 For the secondary outcomes, this sample size of 89 per group will also provide 83% power to detect a difference in the proportion of women with a DXA T-score of ≤ −1 at 2 years, from 16% in the comparison group to 36% in the RRBSO group,45 and 92% power to detect a mean difference between groups of 0.5 SD in the MENQOL score.

Data collection methods

Study data are being collected using paper case report forms, paper self-reported questionnaires, paper recorded phone interviews, and paper copies of biochemical, pathology and bone density imaging reports. These paper documents will be kept by each recruitment site in a locked cabinet accessible to local research staff only. All data will be entered and stored in a de-identified, password-protected electronic study database created with the Research Electronic Data Capture (REDCap) web application (https://projectredcap.org/).44 REDCap is a secure, web-based application designed to support data capture for research studies, providing (1) an intuitive interface for validated data entry; (2) audit trails for tracking data manipulation and export procedures; (3) automated export procedures for seamless data downloads to common statistical packages; and (4) procedures for importing data from external sources.46 Internet access to the REDCap database will use a secure server located at the University of Melbourne, Australia (REDCap consortium host), and access will be limited to local research staff at each recruitment site. Research staff will only have access to the REDCap data collected at their local recruitment site, whereas the chief investigator, project manager, data administrator and statistician will have access to all REDCap data.

Data management

Management of comprehensive and valid paper records will be the responsibility of local research staff at each recruitment site. To ensure a systematic approach to data collection, the project manager will provide protocol training and written standard operating procedures to research staff prior to any study visits being performed. Data integrity will be maintained through review of all collected data prior to, during and after each study visit. Any time-sensitive data (eg, questionnaires) will be reviewed as soon as they are collected and any identified discrepancies will be resolved within 1 week of collection. Medications and AE data will be reviewed and consolidated at every study visit. All other data discrepancies will be queried with the subject via email or telephone correspondence and resolved within 1 month of collection. Records of participant screenings and enrolments at each site will be maintained using Microsoft Excel spreadsheets with diary features. These documents will assist research staff to adhere to study timelines and to regularly monitor and report site study progress to the project manager, investigators and the responsible Human Research Ethics Committees. Electronic data entry into the REDCap database will begin no later than 12 months prior to the last participant last visit. Database training of research staff and data entry will be conducted and overseen (respectively) by the project manager.

Data and safety monitoring

We do not anticipate any study-related AEs or serious adverse events. The only anticipated AEs include phlebotomy-related site injuries (minor) and vasovagal syncope (rarely), and negative psychological responses to questionnaires and interview discussions. These AEs will be circumvented or managed by providing the subject with appropriate advice and counselling during and after each study visit. Reports of biochemistry, pathology and bone density results will be routinely reviewed by medically qualified site investigators. Any concerning results will be immediately reviewed and reported to the subject and her general practitioner (GP). All other results will be reviewed within 1 week of report receipt and any results outside of normal ranges will be reported to the subject’s GP within 2 weeks. The project manager will directly monitor all Victorian data within 1 month of collection, and personally visit other recruitment sites on a 6-monthly basis to perform data monitoring activities. The outcomes will be reported to the chief investigator and all relevant research staff and investigators within 1 month of each monitoring visit, and any systematic problems with data collection, data management and AE monitoring and reporting will be rectified. Questionnaires will be independently scored on two separate occasions prior to data entry into the REDCap database, and all data entry into REDCap will be monitored by an independent data administrator. REDCap data queries will be addressed by the project manager and data administrator using the REDCap Data Resolution Workflow module.

Statistical methods

The analysis set will consist of all recruited women. Baseline characteristics will be summarised by group. The primary outcome is sexual dysfunction, defined as an FSFI score of <26.55 at 2 years. The primary outcome and binary secondary outcomes (eg, PSQI score >5, CES-D score ≥16, GAD-7 ≥10, absolute change from baseline in diastolic blood pressure ≥5 mm Hg, relative change from baseline in HDL:LDL ratio >2.5, and BMD T-score ≤ −1) will be analysed by fitting a generalised linear model with a logistic link function using generalised estimation equations to account for the repeated measurements.47 Continuous secondary outcomes including FSFI as a continuous outcome will be analysed by fitting a linear regression model using generalised estimating equations. The matching variable age group will be included in the above models. Appropriate transformation of a continuous variable may be performed before analyses if the variable is found to be skewed. A directed acyclic graph will be used to explore which covariates to include in the adjusted -model.48 49 An adjusted model will be fitted to correct for the potential confounding of time-independent covariates (eg, use of hormonal contraception at baseline, use of antidepressant medication at baseline) and time-dependent covariates (eg, symptoms of depression and/or anxiety). We anticipate that around one-third of women will have concurrent hysterectomy at the time of RRBSO, that 40% will take HRT following RRBSO, and that 10% of high-risk women will either have a history of breast cancer or will develop breast cancer over the follow-up period. In addition to the primary comparison of all RRBSO participants with all controls, we will compare outcomes between the three subgroups (RRBSO±hysterectomy, ±HRT, ±breast cancer) with all controls.

The number and proportion of women using specific medication (eg, HRT, antidepressants) will be summarised by group. AEs will be recorded and graded using the revised National Cancer Institute Common Terminology Criteria for Adverse Events criteria, and the number and proportion of women with at least one AE will be summarised by group. Exploratory subgroup analyses will examine whether the effect differs across the following subgroup categories: (1) age at baseline (<45 years vs ≥45 years) and (2) use of hormonal contraception at baseline (yes vs no) by adding the subgroup and its interaction with group to the model.

Outcomes and significance

Growing numbers of women are known to be at high inherited risk of ovarian cancer and are electing to undergo RRBSO. The established reduction in ovarian cancer mortality following RRBSO could be further improved by minimising the negative non-cancer consequences, which also may contribute to morbidity and all-cause mortality. This can be addressed only when the non-cancer outcomes are known, and this requires prospective data collection. The principal significance is high burden of disease from ovarian cancer and the growing number of high-risk women who would benefit from RRBSO. In addition, because prospective studies of surgical menopause in the general population are very limited, the findings from this study will also inform care of low-risk women undergoing surgical menopause. These data will be used to develop resources to support decision making around RRBSO, including the timing of surgery, additional hysterectomy, preoperative risk assessment and counselling, and postoperative follow-up including use of HRT. Our studies will directly improve cancer outcomes by impact on clinical practice and policy, through the development of multidisciplinary evidence-based guidelines and screening protocols. Consumers have clearly indicated that they wish to be informed about the consequences of RRBSO and that current information provision and follow-up care are inadequate.50 51 Decision making and satisfaction are improved when high-risk women are offered dedicated clinical services.52 53 Our data and guidelines will provide a template for the care of other high-risk women considering RRBSO as new genes are identified. Improved information about the non-cancer consequences of RRBSO is unlikely to dissuade high-risk women from surgery. Despite side effects, most do not regret their choice.18 50 Improved follow-up care may also improve QOL for high-risk women and contribute to early detection and/or prevention of conditions related to early menopause. This will be the first international multicentre, prospective study of non-cancer outcomes after RRBSO or surgical menopause in any population. Currently, there is no consensus on how high-risk women should be managed following RRBSO. This study will provide new evidence on which to develop evidence-based care for this growing population of women.