Pelvic floor muscle therapy for sexual dysfunction in gynaecological cancer survivors

Objectives

This is a protocol for a Cochrane Review (intervention). The objectives are as follows:

To assess the effects of pelvic floor muscle therapy for sexual dysfunction in gynaecological cancer survivors.

Background

Description of the condition

In recent years, improved access to early diagnosis, new therapies, and enhanced support systems have increased survival rates among women with gynaecological cancers, including endometrial, cervical, ovarian, vaginal, and vulvar cancers [1, 2]. However, many survivors experience long‐term complications that impair quality of life, with sexual dysfunction being a main issue [3, 4, 5]. Diagnosing sexual dysfunction in cancer survivors is challenging, often under‐reported or overlooked, and requires open patient‐clinician communication, physical examination, and psychological assessment according to the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM‐5) criteria [3, 6, 7, 8, 9]. The DSM‐5 defines sexual dysfunction as a disturbance in sexual pleasure and the physiological or psychological processes involved in sexual response, including disorders of interest/arousal, genitopelvic pain/penetration, and orgasm [7]. In gynaecological cancer survivors, sexual dysfunction arises from both psychological effects, such as distress and altered self‐image, and physiological changes due to the disease or its treatment's adverse effects [10]. Women with cervical, endometrial, and vulvar cancers may be at greater risk due to more frequent and direct exposure to pelvic surgery and radiotherapy [3, 5]. For instance, radical hysterectomy can disrupt the pelvic nerve plexus and blood vessels, reducing arousal, lubrication, and neuromuscular co‐ordination, while radiotherapy may cause mucosal damage, fibrosis, and vaginal canal stiffness [11, 12, 13, 14]. It may also weaken pelvic floor muscles, reducing their contractile strength and endurance [15]. Cancer therapies, including chemotherapy, can also induce early menopause, with reduced oestrogen levels contributing to decreased libido, vaginal dryness, and pelvic floor laxity [8, 10, 11, 12, 13, 14]. The pelvic floor, comprising muscles, ligaments, and fascia, plays a central role in sexual function, and cancer‐related damage may lead to high‐tone dysfunction, poor co‐ordination, laxity, and vaginismus [16, 17, 18]. If unaddressed, sexual dysfunction may worsen psychological distress, disrupt relationships, and diminish overall well‐being [8].

Description of the intervention and how it might work

A wide range of interventions have been investigated for the treatment of sexual dysfunction in gynaecological cancer survivors. Candidate pharmacological options include oestrogen hormone replacement therapy (HRT), testosterone therapy, and certain antidepressants [19, 20]. In contrast, non‐pharmacological strategies that address the psychological and physical aspects of sexual health encompass several categories [8, 16, 21]. These include psychotherapeutic or psychoeducational interventions such as cognitive behavioural therapy (CBT), couple‐based therapy, and mindfulness‐based techniques; mechanical interventions like vaginal lasers, dilators, lubricants, or clitoral therapy devices; and complementary and alternative therapies, including acupuncture and natural supplements such as maca, Korean red ginseng, and Ginkgo biloba extracts [16, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32]. Among non‐pharmacological strategies, pelvic floor muscle therapy interventions represent an especially important and accessible option. These interventions, which have been investigated more in the context of female sexual dysfunction in general rather than in gynaecological cancer survivors in particular, target pelvic floor function and musculoskeletal support through modalities such as repeated voluntary contractions, resistance training, mobilisation, electrical stimulation, and muscle control [33]. Pelvic floor muscle therapy interventions may be delivered individually or in combination. They can be categorised as follows [16, 33, 34, 35, 36].

• Pelvic floor muscle strength training (PMST) (e.g. Kegel exercises), involving voluntary, repetitive contractions of the pelvic floor muscles.

• Pelvic floor muscle resistance training (PMRT), consisting of strength‐building exercises using body weight, resistance bands, or other forms of resistance.

• Electrical stimulation of the pelvic floor muscles, using low‐grade electrical currents to stimulate pelvic floor muscle activity or modulate pain. This includes such methods as transcutaneous electrical nerve stimulation (TENS), sacral neuromodulation (SNM), dorsal genital nerve stimulation (DGNS), and percutaneous tibial nerve stimulation (PTNS).

• Pelvic floor manual therapy, which consists of hands‐on techniques targeting soft tissues, fascia, and trigger points, as well as external mobilisation and myofascial release.

• Pelvic floor muscle biofeedback, which uses sensors to provide real‐time feedback on muscle activity.

Several mechanisms have been proposed to explain how pelvic floor muscle‐focused interventions may improve sexual function in this population. Strengthening pelvic floor muscles through PMST or PMRT may enhance muscular support, co‐ordination, and neuromuscular control, potentially improving arousal, lubrication, and orgasmic function [33, 33, 37, 38, 39]. Strengthening the levator ani muscle has been shown to improve pelvic structural support and to reduce strain on the surrounding ligaments [38]. Additionally, increased pelvic blood flow and enhanced tissue mobility through manual techniques (e.g. myofascial release, trigger point therapy) may promote healing and revascularisation of damaged tissues, reduce pain, and further enhance arousal and lubrication [37, 38]. Interventions like biofeedback and electrical stimulation, which have been investigated especially in women with sexual dysfunction and concurrent urinary incontinence or vulvodynia, can retrain pelvic floor control [33, 40, 41]. Electrical stimulation may also help modulate chronic pelvic pain following surgery or radiotherapy, while biofeedback promotes muscle awareness and co‐ordination, facilitating proper muscle recruitment and relaxation during intercourse, thereby decreasing vaginismus [40, 41, 42, 43]. Together, these interventions may increase vaginal receptivity, improve pleasure for both partners, strengthen the couple's relationship, and support broader physical and mental well‐being, all of which positively affect sexual health.

Why it is important to do this review

Although often overlooked, sexual dysfunction is a common issue among gynaecological cancer survivors and can significantly impact their well‐being and quality of life both before and after treatment [5, 44]. Addressing this issue should therefore be considered an integral part of holistic survivorship care, further helping to strengthen the relationship between survivors and healthcare professionals [8, 45]. Despite its importance, no current clinical guidelines recommend specific interventions for the management of female sexual dysfunction in this group. The American College of Obstetricians and Gynecologists (ACOG) acknowledges pelvic floor muscle interventions as a potential option, but this is based on weak evidence derived from the general population rather than gynaecological cancer survivors [16]. While some studies have evaluated the use of such interventions in breast cancer survivors, findings have been inconsistent [46].

Given the challenges faced by all women with sexual dysfunction, whether cancer‐related or not, it is reasonable to explore the potential efficacy and safety of pelvic floor muscle therapy in gynaecological cancer survivors. These interventions may offer an accessible and acceptable non‐pharmacological alternative, particularly for individuals who are often overburdened by prior or ongoing cancer treatments and are therefore reluctant to pursue pharmacological options. Compared to psychotherapeutic interventions, which may carry stigma for some patients due to their discomfort discussing sexual health with mental health professionals, or mechanical strategies like vaginal dilators, which may be less tolerated in the presence of strictures, pelvic floor muscle therapy could theoretically be more appealing [25, 47]. Similarly, complementary and alternative medicine approaches sometimes lack a solid evidence base. The patient‐centred and holistic nature of pelvic floor muscle interventions may also increase adherence, which is essential for treatment success [48].

Although trials have assessed the efficacy of pelvic floor muscle training in improving sexual dysfunction among gynaecological cancer survivors, and despite the perceived advantages of these interventions, the contribution of stronger pelvic floor muscles to sexual function in gynaecological cancer survivors remains uncertain [19, 49, 50, 51]. While generally considered safe, these interventions can, in some cases, cause adverse effects such as pelvic floor muscle hypertonicity, spasms, fatigue, or pain, either during or independent of vaginal penetration [52]. A better understanding of their benefits and risks could contribute meaningfully to comprehensive survivorship care [33].

Previous Cochrane reviews have either focused on mechanical interventions, such as vaginal dilators for women undergoing pelvic radiotherapy for cervical or uterine cancer, or provide outdated evidence on pharmacological and non‐pharmacological options across broader cancer survivor populations [19, 25]. Meanwhile, other relevant systematic reviews may not focus on gynaecological cancer survivors, examine prevalence rather than interventions, may combine self‐measured sexual function outcomes with those measured via validated tools, or omit risk of bias or certainty of evidence assessments [33, 53, 54, 55]. These steps are particularly crucial in the field of sexual health, where outcomes are often reported by intervention recipients [56]. Taking this into account, we aim to inform patients, clinicians, researchers, and policymakers about which pelvic floor muscle interventions, if any, are safe and beneficial for improving sexual function in gynaecological cancer survivors.

Objectives

To assess the effects of pelvic floor muscle therapy for sexual dysfunction in gynaecological cancer survivors.

Methods

We will follow the methodology of the Cochrane Handbook for Systematic Reviews of Interventions for conducting this review [57, 58, 59], and the PRISMA 2020 statement for its reporting [60]. We have applied the guidance from both sources to develop this protocol's Background and Methods sections.

Criteria for considering studies for this review

Types of studies

We will include randomised controlled trials (RCTs) assessing the effect of pelvic floor therapy interventions on gynaecological cancer survivors with sexual dysfunction. We will also include cluster‐RCTs and cross‐over trials (including only data from before the cross‐over).

Given that we expect to find only a few RCTs addressing our review question, and that non‐randomised studies of interventions will be available which address our criteria directly and without critical bias, we will also include non‐randomised studies of interventions [61, 62]. We will particularly include non‐RCTs, quasi‐randomised or pseudo‐randomised studies (i.e. studies that allocate participants to interventions using a method of allocation that is not truly random, such as by date of birth, medical record number, or alternation), and prospective cohort studies demonstrating strong methodological design [61, 62]. A strong design would require prospective outcome measurement (i.e. clear temporal directionality) and efforts to control for selection and confounding biases [61, 62].

We will include conference abstracts, and will contact their investigators for additional information regarding the study's conduct or results as needed. We will apply no restrictions on the year of dissemination or language.

Types of participants

We will include gynaecological cancer survivors. As per the definitions of the National Cancer Institute (NCI) and the American Cancer Society, we will define a cancer survivor as anyone “from the time of cancer diagnosis through the balance of life” and “no matter where they are in the course of their disease” [63, 64]. This definition is also endorsed by the National Comprehensive Cancer Network [45]. Agreeing with the statement from the National Coalition of Cancer Survivorship, we understand and fully respect that anyone affected by cancer can “identify with the term they feel best describes their experience” or with no label at all. However, we will use the term ‘cancer survivor’ (or simply ‘survivor’) throughout this review, as it is the most widely used and scientifically accepted term [65, 66, 67].

Based on the relevant classification of the International Federation of Obstetrics and Gynaecology (FIGO), we will include participants with the following gynaecological cancers.

• Vulvar cancer

• Vaginal cancer

• Cervical cancer (i.e. cancer of the cervix uteri)

• Uterine cancer (i.e. cancer of the corpus uteri, e.g. endometrial cancer)

• Cancer of the ovary, fallopian tube, and peritoneum (i.e. the membrane surrounding them)

• Gestational trophoblastic disease (i.e. tumours developing from cells that would normally create the placenta).

We will include studies with survivors of cancers other than gynaecological ones only if the subset of individuals with gynaecological cancers exceeds 75% of the study's overall population, or if we can access separate data on participants with gynaecological cancer through published study records or direct correspondence with the authors. We will not apply any restrictions in terms of locations or settings.

Recognising the distinct epidemiology of each gynaecological cancer (e.g. cervical cancer typically affects younger women, whereas vulvar cancer more often occurs in older patients) and their differential impact on pelvic floor and sexual function, we will perform subgroup analyses by cancer type. Additionally, given the different mechanisms through which treatments can impair pelvic floor and overall sexual function (e.g. nerve damage from surgery versus vaginal stenosis and fibrosis from radiotherapy), and differences in clinical management, we will also subdivide our included population by treatment modality (e.g. surgery, radiotherapy, or combined treatments), if possible.

We will include all survivors, both those with cancer and those free of it (i.e. free from the disease after treatment or free from recurrence). We will also include participants at any cancer stage (i.e. early‐stage cancer, advanced‐stage or metastatic cancer, or end‐stage cancer requiring only palliative care), either before, during, or after any anticancer treatment [63]. To investigate the potentially modifying effect of the emotional or physical impact of cancer stage, we will examine the effect of interventions on participants with advanced or end‐stage cancer (stage III or IV) versus early‐stage cancer (stage I or II).

Participants should have been sexually active before their cancer diagnosis. We will include all women experiencing relevant symptoms as per the definition of sexual dysfunction (recorded within studies either through self‐report or validated scales) [7]. Although the DSM‐5 requires the presence of symptoms for at least six months for sexual dysfunction to be diagnosed, we will also include studies of women experiencing symptoms for less than six months, as many advanced gynaecological cancers may progress faster [7, 19]. Since it is not always feasible to reliably distinguish between women with new‐onset sexual dysfunction following a cancer diagnosis and those with pre‐existing dysfunction, we will include both populations. Sexual dysfunction may worsen or become clinically apparent only after cancer diagnosis or treatment, leading to potential misclassification. Given that pre‐existing sexual dysfunction may modify the efficacy or safety of physical therapy interventions, we will also separately examine women with pre‐existing versus new‐onset sexual dysfunction, where data are available.

Types of interventions

This review will focus on pelvic floor muscle therapy, which encompasses interventions aiming to improve sexual function by focusing on the optimal function of the pelvic floor muscles. These will include the following.

• PMST (i.e. voluntary, repetitive pelvic floor muscle contractions, such as Kegel exercises).

• PMRT (i.e. strength‐building exercises using any form of resistance).

• Pelvic floor muscle electrical stimulation (i.e. methods using low‐grade electrical currents stimulating pelvic floor muscles like TENS, SNM, DGNS, and PTNS).

• Pelvic floor manual therapy (i.e. hands‐on techniques targeting soft tissues, fascia, and trigger points, with the aim of mobilisation and myofascial release; we will not consider vaginal dilators since their aim is the mechanical stretching of the vaginal canal rather than muscular mobilisation and myofascial release).

• Pelvic floor muscle biofeedback (i.e. the use of sensors that provide real‐time feedback on pelvic floor muscle activity); although biofeedback contains a psychological component and is often considered a ‘mind‐body’ intervention, we will include it due to its physical therapy component and its perceived ability to increase pelvic neuromuscular control and awareness.

Our review will encompass the following comparisons.

• Pelvic floor muscle therapy intervention(s) versus no pelvic floor muscle therapy (i.e. survivorship care including no pelvic floor muscle therapy interventions).

• Pelvic floor muscle therapy intervention(s) versus other pelvic floor muscle therapy intervention(s).

We will consider interventions eligible if they were applied at any point after cancer diagnosis, as per the definition of a cancer survivor. We will allow nutritional advice and aerobic or core stability exercises as potential co‐interventions, as they may often be combined with pelvic floor muscle therapy, and plan to record any such co‐interventions meticulously. We will also allow combinations of pelvic floor muscle therapy interventions.

We will allow an educational component of these interventions only if it focuses on pelvic floor muscle therapy (and potential supplemental physical exercise or nutrition). This means we will allow if providers train recipients on how to properly perform certain interventions (e.g. Kegel exercises) or educate recipients on accepted co‐interventions like nutrition and aerobic or core stability exercises. Physical therapists or other healthcare professionals could deliver or supervise the interventions or educate participants. The interventions may be applied with or without professional supervision, or with the assistance of artificial intelligence (AI) supervision.

We will exclude studies examining combinations of pelvic floor muscle therapy with pharmacological or non‐pharmacological interventions other than those described in this section. We will also exclude studies comparing pelvic floor muscle therapy with pharmacological interventions. In line with the appropriate guidance, we also want to ensure the minimum number of comparisons that serve the review's objective [58].

Outcome measures

Critical outcomes

Our critical outcomes are as follows [57, 61].

• Sexual function, defined as the absence of difficulty moving through the stages of sexual desire, arousal, and orgasm, as well as subjective satisfaction with the frequency and outcome of individual and partnered sexual behaviour. It may be measured by any validated self‐ or clinician‐administered instrument such as the Female Sexual Function Index (FSFI), the Derogatis Inventory of Sexual Functioning (DISF‐SR), and others, within 12 months after the initiation of the intervention [68, 69]. For consistency, we will prioritise the FSFI over other instruments. Namely, if the same study reports the FSFI and additional instruments, we will only use the FSFI. We will include alternative scales in the review only to ensure that we do not entirely exclude studies where the FSFI is unavailable. We will take care not to analyse together scales whose higher scores indicate better sexual function (e.g. FSFI) with those whose higher scores indicate worse function (e.g. the Golombok Rusk Inventory of Sexual Satisfaction (GRISS)).

• Self‐reported new‐onset or worsening pain associated with vaginal penetration (i.e. dyspareunia) within three months [70].

Important outcomes

We will analyse the following important outcomes, which will be measured within 12 months after the initiation of the interventions, unless otherwise stated.

• Sexual distress, defined as frustration, anxiety, and worry regarding sexual activity and measured by any validated self‐ or clinician‐administered instrument such as the revised Female Sexual Distress Scale (FSDS‐R), the original version of the Female Sexual Distress Scale (FSDS), and others. We will prioritise the FSDS‐R over other scales [71, 72].

• Satisfaction with intimacy/relationship, as measured by any validated intimacy/relationship satisfaction scale, such as the Couples Satisfaction Index (CSI), the Relationship Assessment Scale (RAS), and others. We will prioritise scales assessing satisfaction with sexual aspects of intimacy (e.g. CSI) over those assessing intimacy (e.g. RAS), and over more general scales like the Quality of Marriage Index (QMI) [73, 74, 75].

• Quality of life, measured by any relevant instrument validated for either cancer survivors or the general population, such as the Quality of Life Index for Cancer Survivors (QOL‐CS), the WHO Quality of Life Index (WHOQOL), and others [76, 77]. We will prioritise scales specific to cancer survivors (e.g. QOL‐CS).

• Frequency of intercourse (i.e. the number of sexual encounters, either self‐reported or recorded within diaries or logs, per time unit (e.g. times per week, month, or year)).

• Early intervention discontinuation due to poor tolerability (e.g. due to pain or discomfort, difficulty adhering to or properly applying the intervention, injuries, or other self‐reported adverse effects), assessed within three months.

To “allow for modification of adaptations to changes in desire”, the recommended minimum duration of studies measuring efficacy outcomes of female sexual function is six months [78], therefore we chose to measure our critical and important efficacy outcomes within 12 months after the initiation of the interventions. We will measure safety outcomes within three months, since they are more relevant early after starting the intervention. We will exclude studies only if we ensure that the outcomes of interest have not been measured, rather than not reported. In other words, if an outcome of interest is not reported, we will first ensure that it was not measured at all by the study investigators.

Search methods for identification of studies

In collaboration with the Cochrane Gynaecology and Fertility Information Specialist, we will search for RCTs, quasi‐RCTs, non‐RCTs, or prospective cohort studies on the use of pelvic floor muscle therapy interventions for sexual dysfunction in gynaecological cancer survivors. We will apply no language or date restrictions to the search.

Electronic searches

We will search for relevant studies in the following electronic databases [79, 80].

• Cochrane Central Register of Controlled Trials (CENTRAL), which includes trial reports from the Cumulative Index to Nursing and Allied Health Literature (CINAHL) and trial registrations from ClinicalTrials.gov and the WHO International Clinical Trials Registry Platform (ICTRP), using the Ovid platform; from inception to present

• MEDLINE (Epub Ahead of Print, In Process and Other Non‐Indexed Citations), using the Ovid platform; from 1946 to present

• Embase, using the Ovid platform; from 1980 to present

• PsycINFO, using the Ovid platform; from 1806 to present

• ClinicalTrials.gov (clinicaltrials.gov/) and WHO ICTRP (trialsearch.who.int/) trial registries for ongoing trials, web platform; from inception to present

The search strategies are available in the Appendix (Supplementary material 1).

Searching other resources

We will check the reference lists of retrieved studies, relevant reviews, and conference proceedings from major conferences to identify additional studies. Additionally, we will search relevant grey literature sources, including the first 200 records from the Google Scholar database. We will also search the Epistemonikos database (www.epistemonikos.org/), a multilingual database of health evidence and the largest source of systematic reviews. Furthermore, we will consult experts in the field to identify ongoing or existing studies.

To ensure the integrity of the included studies, we will check PubMed (pubmed.ncbi.nlm.nih.gov) for studies of concern and retracted studies, including expressions of concern, errata, corrigenda, and retractions. We will exclude any identified problematic studies from the review.

Data collection and analysis

We will select studies, collect study data, import the data into RevMan [81], and analyse them according to methods described in the Cochrane Handbook for Systematic Reviews of Interventions [57, 61, 79, 82, 83].

Selection of studies

After deduplicating the records retrieved from the search using EndNote [84], sets of independent review authors (AB, DG, A Saitis, MN, and A Syristatidi) will screen the titles and abstracts of the unique records and exclude those that do not fulfil the review's eligibility criteria. We will retrieve the full texts of the remaining records and independently screen them for eligibility. We will exclude studies not meeting the criteria and explicitly document the reasons for doing so in the 'Characteristics of excluded studies' table. We will use Rayyan online software for all screening processes [85]. If the eligibility of a study remains unclear even after accessing its full‐text record, we will contact the study authors for further clarification. Another review author (DV or MP) will resolve disagreements during both stages.

If potentially eligible studies are published in languages not spoken by the review team, we will seek assistance from colleagues with relevant language skills or through the Cochrane Engage platform. We will summarise the full study selection process in a PRISMA flow chart [79].

Data extraction and management

Sets of review authors (AB, DG, A Saitis, MN, and A Syristatidi) will independently extract study characteristics and outcome data from the reports of each included study. A predefined spreadsheet will first be piloted among team members. Within each set of review authors, one review author will extract data from the records, and the other will validate the imported values against the extraction form. We will resolve any disagreements through discussion; if discrepancies persist, we will consult a third review author (DV or MP). In the case of multiple reports of a single study, we will use the main report as the reference and will supplement it with data from the secondary reports. We will email the authors of the included studies for additional data or methodological details as needed. If we receive no response or insufficient data, we will send them a reminder within 20 days after the first contact [82, 83].

The data extraction form will contain the following items.

• General information and study details

  • Identifiers of the report and study

  • Names of the review authors extracting data and the extraction date

  • Funding sources for the study

  • Declarations of interest of the primary investigators

  • Country, including the income level group (i.e. low, lower‐middle, upper‐middle, and high) according to the World Bank countries classification based on Gross National Income per capita

  • Study design (i.e. RCT, non‐randomised/quasi‐randomised controlled trial, prospective cohort)

  • Study centre(s)

  • Healthcare setting (public or private)

  • Study period

  • Corresponding author's contact details

  • Study registration information

• Study methods

  • Study design: parallel, cluster, or cross‐over trial for RCTs; quasi‐randomised, non‐randomised controlled, or prospective cohort study for non‐randomised studies of interventions

  • Recruitment and sampling procedures

  • Enrolment start and end dates

  • Details of random sequence allocation and concealment for RCTs, or efforts to minimise selection and confounding biases for non‐randomised studies

  • Length of follow‐up

  • Methods of handling missing data (e.g. imputations)

• Participants

  • Age (mean, standard deviation (SD) of the overall group and each intervention group)

  • Inclusion and exclusion criteria

  • Total number of participants

  • Number of participants assigned to each intervention group

  • Percentage of participants who are gynaecological cancer survivors in the case of a mixed population

  • Gynaecological cancer type

  • Race and ethnicity in each intervention group and overall

  • Cancer status (living with cancer or recurrence versus living cancer‐ or recurrence‐free)

  • Cancer stage (I to IV)

  • Anticancer treatment(s) (e.g. surgical treatment only, radiotherapy only, combinations of surgery and radiotherapy, with or without other systemic therapy (e.g. chemotherapy))

  • Baseline sexual dysfunction (as measured via any validated scale) of the intervention groups and the overall population, and duration of symptoms; numbers of those with pre‐existing sexual dysfunction in the overall study population and each intervention group will also be recorded if such data are available

• Interventions

  • Intervention group (pelvic floor muscle therapy or survivorship care without pelvic floor muscle therapy)

  • Pelvic floor muscle therapy intervention component(s) (e.g. PMST, PMRT, electrical stimulation, etc. See also Types of interventions) and characteristics (e.g. frequency, duration, deliverer/supervisor, means of delivery (i.e. face‐to‐face, online education/supervision))

  • Comparator and its characteristics (i.e. description of usual survivorship care without pelvic floor muscle therapy)

  • Description of co‐interventions (e.g. dietary advice, additional physical exercise), if applied

• Outcomes

  • Definition

  • Time point of measurement

  • Measurement scales of sexual function, sexual distress, satisfaction with intimacy/relationship, quality of life, and means of recording and time unit for frequency of intercourse (i.e. self‐reported or recorded within logs or diaries)

  • Raw outcome data (i.e. number of events in each group and number of total participants in each group for which the outcome was measured for dichotomous outcomes (new‐onset/worsening dyspareunia or intervention discontinuation), and mean, SD, and number of participants in each group for which the outcome was measured for continuous or discrete outcomes, such as those measured though scales (sexual function, sexual distress, satisfaction with intimacy/relationship, and quality of life) and the frequency of intercourse) [86]

Risk of bias assessment in included studies

Sets of two review authors (AB, DG, A Saitis, MN, and A Syristatidi) will independently assess risk of bias in the included studies according to the methodology outlined in Chapters 7, 8, and 25 of the Cochrane Handbook for Systematic Reviews of Interventions [56, 87, 88]. We will resolve any disagreements through discussion with a third review author (DV or MP).

We will use the Cochrane RoB 2 tool for parallel RCTs and its adapted versions for cluster‐RCTs and cross‐over trials [89]. The RoB 2 tool includes specific items assessing five domains where bias may occur [90], as follows.

• Bias arising from the randomisation process

• Bias due to deviations from intended interventions

• Bias due to missing outcome data

• Bias in outcome measurement

• Bias in the selection of reported results

We will judge the risk of bias in each domain as ‘low risk of bias’, ‘some concerns’, or ‘high risk of bias’ based on our answers to the signalling questions of the RoB 2 tool and after taking into account the tool algorithm's suggestions and the criteria outlined in the aforementioned chapters of the Cochrane Handbook [56, 87, 90]. For the second domain of 'bias due to deviations from intended interventions', we will focus on the effect of assignment to intervention at baseline (the intention‐to‐treat effect).

The five domain‐level judgements will form the basis for the overall risk of bias assessment for each study's outcome. We will assess the risk of bias for each trial's outcome of interest as follows.

• ‘High’ if the risk of bias is ‘high’ in at least one domain or ‘some concerns’ in multiple domains, unless a different judgement is justified.

• ‘Some concerns’ if there are ‘some concerns’ about bias in at least one domain, but not in multiple domains.

• ‘Low’ if there is a ‘low’ risk of bias in all five domains [56, 87, 90].

For non‐randomised studies, we will assess the risk of bias using ROBINS‐I (Risk Of Bias In Non‐randomized Studies – of Interventions) version 2, which evaluates bias in the result of an individual non‐randomised study examining the effect of an intervention on an investigated outcome [88]. ROBINS‐I focuses on the risk of material bias, defined as bias sufficient to change the magnitude of the estimated effect in a meaningful way relative to the true causal effect (in our case, the effect of physical therapy interventions on sexual function outcomes).

ROBINS‐I assesses the risk of bias across the following seven domains.

• Bias due to confounding. We consider cancer type, treatment, stage, baseline sexual function, age, menopausal status, and relationship status (partnered versus unpartnered) as key confounding factors that should be adequately controlled for in any included non‐randomised studies.

• Bias in classification of interventions

• Bias in selection of participants into the study or analysis

• Bias due to deviations from intended interventions

• Bias due to missing data

• Bias in measurement of outcomes

• Bias in selection of the reported result

Domain‐level risk of bias will be based on answers to structured signalling questions, using standard ROBINS‐I responses (‘yes’, ‘probably yes’, ‘probably no’, ‘no’, ‘no information’) and mapped via the tool’s algorithms. Each domain will be rated as ‘low’ (except for the confounding domain, which is ‘low risk of bias except for concerns about uncontrolled confounding’), ‘moderate’, ‘serious’, or ‘critical’ risk of bias. According to the tool's algorithms, the domain‐level judgements will form the basis for the overall judgement of each study's result.

• Low risk of bias: there is little or no concern for bias.

• Moderate risk of bias: there is some concern for bias, though it is unclear if this bias is important.

• Serious risk of bias: there are important problems.

• Critical risk of bias: the result should probably be excluded from evidence syntheses.

We will pilot both tools among all review authors. We may collectively override algorithm‐guided domain‐level or overall judgements after justification and consensus. We will support our judgements with proper justification and text references from the studies' records. If we require additional methodological details to answer the signalling questions, we will seek clarification from the studies' principal investigators.

We will record all assessments within the dedicated Excel tool for RoB 2 or ROBINS‐I compatible formats (Excel or robvis; available at www.riskofbias.info/). This will enable us to efficiently document and present our judgements and justifications. We will make them all available as a separate online supplemental file in a free source such as Figshare.

We will conduct a risk of bias assessment for the following critical and important outcomes of our review, which will be measured within 12 months after the initiation of the intervention, unless otherwise stated.

• Sexual function, defined as the absence of difficulty moving through the stages of sexual desire, arousal, and orgasm, as well as subjective satisfaction with the frequency and outcome of individual and partnered sexual behaviour. It may be measured by any validated self‐ or clinician‐administered instrument such as FSFI, DISF‐SR, and others [68, 69].

• Self‐reported new‐onset or worsening pain associated with vaginal penetration (i.e. dyspareunia) within three months [70].

• Sexual distress, defined as frustration, anxiety, and worry regarding sexual activity and measured by any validated self‐ or clinician‐administered instrument such as FSDS‐R, FSDS, and others [71, 72].

• Satisfaction with intimacy/relationship, as measured by any validated intimacy/relationship satisfaction scale, such as CSI, RAS, and others [73, 74, 75].

• Quality of life, measured by any relevant instrument validated for either cancer survivors or the general population, such as QOL‐CS, WHOQOL, and others [76, 77].

• Frequency of intercourse (i.e. the number of sexual encounters, either self‐reported or recorded within diaries or logs, sexual encounters per time unit (e.g. times per week, month, or year)).

• Early intervention discontinuation due to poor tolerability (e.g. due to pain or discomfort, difficulty adhering to or properly applying the intervention, injuries, or other self‐reported adverse effects), assessed within three months.

Measures of treatment effect

For adverse effects (i.e. new‐onset or worsening dyspareunia and early intervention discontinuation), which are dichotomous outcomes, we will use the event numbers in the intervention and control groups of each study to compute odds ratios (ORs). If necessary, we will reverse the direction of the effect of individual studies to ensure consistency across studies.

For continuous outcomes (i.e. sexual function, sexual distress, satisfaction with intimacy/relationship, quality of life, frequency of intercourse), we will use the outcome means to calculate mean differences (MDs) between treatment groups if studies report the same outcomes using identical scales (sexual function, sexual distress, satisfaction with intimacy/relationship, quality of life) or units of measurement (frequency of intercourse per particular time unit).

In cases where studies report similar outcomes using different scales (e.g. sexual function, sexual distress, satisfaction with intimacy/relationship, or quality of life) or time units that cannot be converted to a common one, we will compute standardised mean differences (SMDs) to allow for comparability across studies.

For the 'frequency of intercourse' outcome reported on different time bases (e.g. per week, per month, or three months), we will first assess whether it is feasible and appropriate to transform all reported means and SDs to a common time unit (e.g. acts per month) using fixed multipliers (e.g. multiplying weekly means and SDs by 4.3 to approximate monthly frequency, or daily ones by 30). This transformation is valid for both means and SDs, assuming that the underlying event rate is stable over time. However, if such standardisation is not feasible due to lack of clarity in reporting, vaguely defined time units (e.g. ‘typical week’), or inconsistent scaling (e.g. frequency items embedded in broader composite scores), we will instead use SMDs. Using the SMD, we will pool results across studies by expressing the intervention effect in SD units, despite differences in time frames or scaling.

In all cases, we will interpret SMDs using conventional thresholds (e.g. 0.2: small, 0.5: moderate, 0.8: large effect). For all outcomes, we will present the ORs, MDs, or SMDs with 95% confidence intervals (95% CIs) [86].

Unit of analysis issues

The individual participant (i.e. gynaecological cancer survivor) will serve as the primary unit of analysis. If events can recur (e.g. multiple events of new‐onset/worsening dyspareunia), we will consider each participant to be the unit of analysis.

If a single study involves three or more arms, we will divide the control group into equal parts, assuming equal incidence of the outcome in each one, to prevent double‐counting its data. If we encounter irrelevant arms in such studies, we will exclude them from the analyses.

In the case of cluster‐RCTs, we will extract the intracluster correlation coefficient (ICC) from the trial's data. If the ICC is not available, we will extract it from an RCT or an observational study with a similar population. If none of the above is possible, we will conduct a sensitivity analysis for different ranges of ICCs. We will record the number and total size of clusters per group, and the unit of randomisation. We will combine the results of cluster‐RCTs with those of the other included studies only if these are reasonably homogeneous [86, 89].

Dealing with missing data

We will attempt to obtain missing data on the outcomes of interest by contacting the study investigators. If we receive no response, we will send a kind reminder within 20 days after the first email [86]. We will not adopt any imputation approach and will analyse the available data. If we judge missing data on the critical outcomes to be considerable within studies, we will investigate their impact by excluding these studies in sensitivity analyses.

Reporting bias assessment

We will assess reporting bias in our outcomes of interest by comparing each study's planned outcomes (available within a protocol paper or the study's registration) with its reported outcomes (available in the published reports). If there are missing results, we will contact the study authors to request additional information.

If more than 10 studies are available for a given outcome, we will use funnel plots to screen for small‐study effects and publication bias. We will attempt to interpret any major funnel plot asymmetry upon visual assessment. If the funnel plot's asymmetry suggests publication bias, we will integrate this information into our evaluation of the certainty of the evidence. If the number of studies is inadequate and reduces our ability to detect small‐study effects, we will acknowledge this limitation [91].

Synthesis methods

We will conduct analyses on the following comparisons.

• Pelvic floor muscle therapy intervention(s) versus no pelvic floor muscle therapy (i.e. survivorship care including no pelvic floor muscle therapy interventions)

• Pelvic floor muscle therapy intervention(s) versus other pelvic floor muscle therapy intervention(s)

To more efficiently present the study characteristics and the results, we will compile a study characteristics table using the collected and imported data. We will perform meta‐analyses using the RevMan tools [81]. We will pool MDs or SMDs for the continuous outcomes of sexual function, sexual distress, satisfaction with intimacy/relationship, quality of life, and frequency of intercourse, and ORs for the dichotomous outcomes of new‐onset/worsening dyspareunia and early intervention discontinuation [92].

We have opted for random‐effects models due to the anticipated clinical heterogeneity of our included population and interventions, including potential variations in multiple parameters like gynaecological cancer type, anticancer treatments, cancer stage, survivorship phase, baseline sexual dysfunction, age, relationship status, menopausal status, and duration, frequency, and delivery of pelvic floor muscle therapy. These differences are expected in the clinical setting of survivorship care and physical therapy; hence, the quantitative synthesis of data from these studies seems meaningful.

We will estimate between‐study variation via the DerSimonian‐Laird method and calculate the 95% CI for the summary effect using the Wald‐type method [92]. For each meta‐analysis, we will estimate heterogeneity variances (Tau²) and assess for the presence of statistical heterogeneity through the visual inspection of forest plots and the calculation of the I² statistic [93]. A value of I² > 50% will indicate substantial heterogeneity. Our primary analyses will only include studies judged to have a low risk of bias based on their overall RoB 2 or ROBINS‐I assessment for the outcome.

If we judge that the heterogeneity between included populations, interventions, or outcomes prevents us from drawing clinically meaningful conclusions by quantitatively synthesising data, we will avoid meta‐analysis and follow the Synthesis Without Meta‐analysis (SWiM) approach [94].

Investigation of heterogeneity and subgroup analysis

In the case of substantial or considerable heterogeneity (I² > 50%), we will conduct subgroup analyses to explore its potential sources and investigate factors that could contribute to variations in the effect of the intervention [92]. The rationale for considering each subgroup analysis is presented in the Types of participants section. We will use the following variables as covariates.

• Cancer type (i.e. vulvar versus vaginal versus cervical versus uterine versus ovarian/fallopian tube cancer versus gestational trophoblastic disease)

• Treatment type (i.e. surgery versus radiotherapy versus combination)

• New‐onset (after cancer diagnosis) versus pre‐existing (before cancer diagnosis) sexual dysfunction

• Cancer stage (i.e. advanced or end‐stage cancer (stage III or IV) versus early‐stage cancer (stage I or II))

• World Bank classification of countries by income level (low/lower‐middle‐income versus upper‐middle/high‐income setting)

In all cases, we will compare subgroups using the formal test for subgroup differences in RevMan [81]. If there are insufficient data for meaningful analysis (i.e. fewer than 10 included studies), we will not emphasise its findings.

Equity‐related assessment

In alignment with Cochrane’s equity guidance, we will investigate health equity in this review [95]. We acknowledge that healthcare interventions may have different effects across populations due to various factors such as socioeconomic and regional ones [95]. This is particularly relevant for the range of pelvic floor muscle therapy interventions such as pelvic floor strength training, electrical stimulation of pelvic floor muscles, and biofeedback, which can be applied in various ways. In this setting, potential differences can easily occur and should be addressed.

Variations in healthcare resources, infrastructure, and financing across different settings may influence the implementation and effectiveness of these interventions. Cost‐sharing models and reimbursement structures may also impact patients' accessibility to these treatments and actual treatment uptake. To explore potential disparities in access and outcomes, we will extract and report detailed information on the setting of each study, including country and healthcare setting (see Data extraction and management).

We will also conduct a subgroup analysis based on the World Bank classification of countries by income level, comparing outcomes between low/lower‐middle‐income and upper‐middle/high‐income countries. This may provide insight into how economic and structural differences may affect the availability of holistic survivorship care to gynaecological cancer survivors in different settings.

Furthermore, we will extract and report relevant demographic data and participant characteristics, such as age and race/ethnicity, to examine whether baseline differences in population groups could influence intervention outcomes (see Data extraction and management). In this context, we aim to assess the generalisability of our findings and consider the accessibility and applicability of such interventions in different populations. This approach will support our efforts to address and highlight health equity implications in the context of proper sexual care for cancer survivors.

Sensitivity analysis

We will conduct sensitivity analyses to explore if the meta‐analysis findings are affected if [92]:

• the analysis is restricted to studies measuring the critical outcomes of sexual function and sexual distress, using only the FSFI and FSDS‐R scales, respectively;

• we restrict our analyses of critical outcomes to studies with available full text;

• we restrict our analyses of all critical and important outcomes to studies including only gynaecological cancer survivors;

• we restrict our analyses of critical outcomes to studies without substantial data loss (see Dealing with missing data);

• we include studies judged to have a high risk of bias or some concerns for bias based on their overall RoB 2 assessment, or moderate, serious, or critical risk of bias based on their overall ROBINS‐I v2 assessment for the critical and important outcomes.

Certainty of the evidence assessment

We will use GRADEpro GDT software [96], and follow the GRADE Handbook and Chapter 14 of the Cochrane Handbook for Systematic Reviews of Interventions, to assess the certainty of evidence and construct summary of findings tables for the following comparisons [97, 98].

• Pelvic floor muscle therapy intervention(s) versus no pelvic floor muscle therapy (i.e. survivorship care including no pelvic floor muscle therapy interventions)

• Pelvic floor muscle therapy intervention(s) versus other pelvic floor muscle therapy intervention(s)

We will assess the certainty of evidence for the following outcomes, which will be measured within 12 months after the initiation of the intervention, unless otherwise stated.

• Sexual function, defined as the absence of difficulty moving through the stages of sexual desire, arousal, and orgasm, as well as subjective satisfaction with the frequency and outcome of individual and partnered sexual behaviour. This may be measured by any validated self‐ or clinician‐administered instrument such as FSFI, DISF‐SR, and others [68, 69].

• Self‐reported new‐onset or worsening pain associated with vaginal penetration (i.e. dyspareunia) within three months [70].

• Sexual distress, defined as frustration, anxiety, and worry regarding sexual activity and measured by any validated self‐ or clinician‐administered instrument such as FSDS‐R, FSDS, and others [71, 72].

• Satisfaction with intimacy/relationship, as measured by any validated intimacy/relationship satisfaction scale, such as CSI, RAS, and others [73, 74, 75].

• Quality of life, measured by any relevant instrument validated for either cancer survivors or the general population, such as QOL‐CS, WHOQOL, and others [76, 77].

• Frequency of intercourse (i.e. the number of sexual encounters, either self‐reported or recorded within diaries or logs, sexual encounters per time unit (e.g. times per week, month, or year)).

• Early intervention discontinuation due to poor tolerability (e.g. due to pain or discomfort, difficulty adhering to or properly applying the intervention, injuries, or other self‐reported adverse effects), assessed within three months.

Sets of review authors (AB, DG, A Saitis, MN, and A Syristatidi) will independently apply the GRADE criteria to assess the certainty of evidence for each outcome as ‘high’, ‘moderate’, ‘low’, or ‘very low’. We will define each level according to Chapter 5 of the GRADE Handbook [98]. The GRADE approach considers five factors for potential downgrading (i.e. overall risk of bias, inconsistency, imprecision, indirectness, and publication bias) and three factors for potential upgrading (i.e. large effect, dose‐response, and all plausible residual confounding) of the certainty of evidence. Regarding the overall risk of bias criterion, the overall RoB 2 or ROBINS‐I judgement for the outcome will inform the GRADE assessment. We will resolve any disagreements through discussion or by involving another review author (DV or MP).

Consumer involvement

While we acknowledge that consumers will not be directly involved in the conduct of the review itself due to resource constraints, we have ensured that the studied outcomes fully align with core outcome sets for female sexual dysfunction, which were developed with meaningful consumer input and reflect their priorities and experiences [78, 99].

Supporting Information

Supplementary materials are available with the online version of this article: 10.1002/14651858.CD016162.

Supplementary materials are published alongside the article and contain additional data and information that support or enhance the article. Supplementary materials may not be subject to the same editorial scrutiny as the content of the article and Cochrane has not copyedited, typeset or proofread these materials. The material in these sections has been supplied by the author(s) for publication under a Licence for Publication and the author(s) are solely responsible for the material. Cochrane accordingly gives no representations or warranties of any kind in relation to, and accepts no liability for any reliance on or use of, such material.

Supplementary material 1 Search strategies

New

Additional information

Contributions of authors

DV, AB, VS, DG, A Saitis, MN, A Syristatidi, and MP drafted the protocol. MP edited most parts of the protocol. DV and MP reviewed the protocol and administered and supervised the project.

Declarations of interest

DV: no commercial or non‐commercial conflicts of interest relevant to this review.

AB: is an obstetrics and gynaecology physician at the Second Department of Obstetrics and Gynaecology, Aretaieion University Hospital, Medical School, National and Kapodistrian University of Athens. No commercial or non‐commercial conflicts of interest relevant to this review.

VS: is a gynaecologic oncologist at MITERA Hospital and Specialized Healthcare Hub private clinic (Athens, Greece). No commercial or non‐commercial conflicts of interest relevant to this review.

DG: is a board‐certified general surgeon at the First Department of Surgery, Attica General Hospital Sismanogleio ‐ Amalia Fleming, Athens, Greece. No commercial or non‐commercial conflicts of interest relevant to this review.

A Saitis: has received support for attending conferences by Innovis Pharma S.A. and Recordati Hellas Pharmaceuticals Α.E., unrelated to the current review. No commercial or non‐commercial conflicts of interest relevant to this review.

MN: is an internal medicine resident at Laiko General Hospital, Athens, Greece. Received a grant from the World Health Organization unrelated to the present review. No commercial or non‐commercial conflicts of interest relevant to this review.

A Syristatidi: no commercial or non‐commercial conflicts of interest relevant to this review.

MP: received grants from the World Health Organization and is receiving a scholarship from the Foundation for Education and European Culture to support PhD studies at the National and Kapodistrian University of Athens, all unrelated to the present review. No commercial or non‐commercial conflicts of interest relevant to this review.

Sources of support

Internal sources

• Internal funding, Other

  No internal sources of support were received.

External sources

• External funding, Other

  No external sources of support were received.

Registration and protocol

Cochrane approved the proposal for this review in April 2024.

Data, code and other materials

Data sharing is not applicable to this article as it is a protocol, so no datasets were generated or analysed.