Hysterectomy and sexual function: a systematic review and meta-analysis

Abstract

Background

Sexual function after hysterectomy can be a concern for patients, and research remains inconclusive about changes in sexual function associated with hysterectomy.

Aim

We meta-analyzed studies on change in sexual function from pre- to posthysterectomy and the role of total vs subtotal hysterectomy and concomitant bilateral salpingo-oophorectomy (BSO) in differences in such change.

Methods

We searched PubMed, Embase, and Cochrane databases from inception to January 2022. Two reviewers screened and included studies if they were published in a peer-reviewed journal and reported on sexual function pre- and posthysterectomy for benign nonprolapse indication. Methodological quality was assessed with the STROBE checklist. We used random effects multilevel models to meta-analyze standardized mean differences in pre- to postoperative sexual function and the posthysterectomy Female Sexual Function Index mean across study groups in R (RStudio).

Outcomes

Outcomes included overall sexual function, dyspareunia, desire, arousal, lubrication, and orgasm.

Results

Thirty-two articles were analyzed: 8 randomized controlled trials, 20 prospective studies, 2 retrospective studies, 1 cross-sectional study, and 1 secondary analysis, comprising a total of 4054 patients. Each study provided data for at least 1 outcome. Study quality was moderate, and effect sizes showed large between-study heterogeneity. Hysterectomy was not associated with significant change in overall sexual function irrespective of surgical route, with patients tending to report potentially remaining sexual dysfunction posthysterectomy. Cervix removal was not significantly associated with differences in magnitude of change. Hysterectomy without BSO was associated with significantly stronger improvement in lubrication and orgasm than hysterectomy with BSO, which was not the case for desire, arousal or overall sexual function. However, these significant differences were not replicated within studies that directly compared cases with and without BSO.

Clinical Implications

Clinicians should address remaining sexual dysfunction posthysterectomy, and BSO should not be considered if not medically required.

Strengths and Limitations

We analyzed a comprehensive number of trials and studied clinically relevant factors that might relate to differences in change in sexual function. Conclusions need to be interpreted with caution since many studies showed moderate methodological quality and large effect size heterogeneity.

Conclusion

Subtotal and total hysterectomy was not associated with significant change in overall sexual function irrespective of surgical route, with patients tending to report potentially remaining sexual dysfunction posthysterectomy. Hysterectomy without BSO was associated with significantly stronger improvement in lubrication and orgasm than hysterectomy with BSO. Future research on hysterectomy should analyze predictors of sexual function change trajectories, such as different indications.

female sexual function, hysterectomy, cervix, oophorectomy, metaanalysis, systematic review, PROSPERO: CRD42020177473

Introduction

Over the last years, there has been a growing interest in the effect of hysterectomy on sexual functioning in women. Hysterectomy is one of the most frequently performed gynecologic procedures.¹ During a hysterectomy, the uterus is removed by abdominal, vaginal, or laparoscopic surgery. Hysterectomy can be performed as an indication for benign or malignant conditions. Approximately 80% to 90% of hysterectomies are performed for benign conditions,² such as abnormal uterine bleeding due to any of the causes of the FIGO classification system (PALM-COEIN: polyp, adenomyosis, leiomyoma, malignancy and hyperplasia, coagulopathy, ovulatory dysfunction, endometrial, iatrogenic, not yet classified), painful bleeding, bulky symptoms, or dyspareunia.^3,4

Over the past years, medical and hormonal treatments (eg, the levonorgestrel-releasing intrauterine device) and less invasive surgical options (eg, endometrial ablation and hysteroscopic myomectomy) have contributed to a declining rate of hysterectomy in Western European countries.^5,6 Yet, hysterectomy remains the definitive treatment for benign uterine disease. It is important that women are well aware of the possible advantages and disadvantages of each type of treatment to choose one that suits them best.

Scientific evidence on the impact of hysterectomy on sexual function has been conflicting. Such evidence might result from differences in indication for surgery, differences in the study population, and the use of different and often inadequate parameters to assess sexual function across studies.⁷ Unfortunately, such a lack of clear evidence contributes to the absence of adequate preoperative counseling regarding sexual function after hysterectomy.

Various hypotheses exist on how hysterectomy may influence sexual function. For instance, preoperative complaints, such as heavy menstrual bleeding, pelvic pain, dyspareunia, and dysmenorrhea, negatively affect sexual function prior to surgery, and hysterectomy may improve sexual function through alleviating these.⁸ However, some suggest that hysterectomy may interrupt genital nerve supply, causing a reduction of genital sensibility and leading to sexual dysfunction.⁹ Differences in consequences for sexual function might then be related to differences in the surgical approach.

Surgical route

Hysterectomy can be performed via the abdominal, vaginal, or laparoscopic surgical route. The role of the surgical route in the effect on sexual function remains a source of debate. In general, vaginal hysterectomy (VH) is superior to laparoscopic hysterectomy (LH) and abdominal hysterectomy (AH) for benign disease, as it is associated with a faster return to normal daily activities, as suggested in a Cochrane review by Aarts et al.¹ If technically feasible, VH should be performed in preference to AH because it leads to more rapid recovery and fewer febrile episodes postoperatively. If VH is unfeasible, LH has some advantages over AH, such as more rapid recovery and fewer febrile episodes and wound or abdominal wall infections. However, Aarts et al also stated that insufficient evidence, specifically a lack of data from randomized controlled trials (RCTs), precludes definite conclusions on the differential effects of surgical route on sexual function. As such, regarding sexual function, the evidence is inconclusive in whether the surgical route of hysterectomy is associated with differences in change in sexual function.

Subtotal vs total hysterectomy

Hysterectomy can be performed as a subtotal or total procedure, with the total procedure removing the cervix and the subtotal leaving it in situ. It is still debated whether total and subtotal hysterectomies differ in their potential effect on sexual function, since the role of the cervix in sexual arousal and genital sensation is, as of yet, unclear. Also, dyspareunia has been considered a potential consequence of total hysterectomy, since excision of the cervix has been hypothesized to induce deep pain during sexual intercourse.¹⁰ Furthermore, when a subtotal hysterectomy is being performed, the large segment of the pelvic plexus of the parasympathetic and sympathetic nerves may be damaged less, thereby sparing genital sensitivity.¹¹ As the cervix is thought to play a potential role in sexual functioning, one reason for leaving it in situ is the expected benefit for sexual function after the operation. However, current evidence is inconclusive whether removal of the cervix is associated with diminished improvement of sexual function posthysterectomy or deterioration of sexual function.

Bilateral salpingo-oophorectomy

Hysterectomy is frequently performed in combination with oophorectomy—specifically, removal of 1 ovary (unilateral salpingo-oophorectomy) or both ovaries (bilateral salpingo-oophorectomy [BSO]). The ovaries are important for the production of steroid hormones in premenopausal but also postmenopausal women,^12–16 and these hormones play a role in female sexual function.^17–23 Estrogen affects the tissue integrity of the vaginal epithelia, which indirectly affects sexual function by protecting women from experiencing pain during penetrative intercourse in case of insufficient sexual arousal.²⁴ Androgens, specifically testosterone, modulate the central responsivity to sexual stimuli and the sensitivity of the genital tissue to tactile stimulation, as well as the integrity of the genital arousal response.^21,25

Premenopausal women undergoing BSO reported a significant decrease in sexual pleasure and comfort as compared with their baseline situation before the surgery,²⁶ and women <50 years old who underwent a BSO had a higher frequency of hypoactive sexual desire disorder when compared with women with ovarian preservation.²⁷

Removal of the ovaries affects steroid concentrations in pre- and postmenopausal women, specifically testosterone.^15,16 Lowered steroid concentrations may lead to diminished sexual arousability, resulting in potential problems occurring throughout the sexual response curve. Such problems might present as a deterioration in sexual function from pre- to posthysterectomy or a diminished improvement in sexual function as compared with women who do not have their ovaries removed.

Objectives

In this systematic review and meta-analysis, we examined the role of hysterectomy in changes in sexual function. We reviewed RCTs and prospective and retrospective studies reporting recalled pre- and postoperative sexual function. The outcome of this systematic review and meta-analysis should advise clinicians whether and when to address sexual function before and after hysterectomy. To answer clinically important subquestions, we studied 3 clinically relevant PICO statements (population, intervention, control, and outcomes).

PICO 1: whether hysterectomy for benign nonprolapse indication is associated with changes in sexual function from pre- to posthysterectomy, subdivided for surgical route.

PICO 2: whether removal of the cervix—specifically, subtotal vs total hysterectomy for benign nonprolapse indication—is associated with differences in changes in sexual function from pre- to posthysterectomy.

PICO 3: whether performing hysterectomy with BSO (+BSO) for benign nonprolapse indication is associated with differences in changes in sexual function from pre- to posthysterectomy as compared with hysterectomy without BSO (–BSO).

Methods

All data were extracted from published sources; therefore, institutional review board approval was not necessary for this study. This systematic review was conducted according to the PRISMA guidelines (Preferred Reporting Items for Systematic Reviews and Meta-analyses).²⁸ The protocol is freely accessible on PROSPERO; the ID of this systematic review is CRD42020177473.

Literature search

A clinical librarian performed an electronic search in PubMed from inception to January 9, 2020. A complementary updated search in the PubMed, Embase, and Cochrane databases was completed on January 13, 2022. Subject headings, MeSH terms, words in title, abstract and keywords were used: (sexual dysfunction OR sex disorder^* OR sexual function OR physiological sexual dysfunction OR sex abnormalit^* OR dyspareunia OR hypoactive sexual desire disorder OR coit^* OR libido OR anorgasmia) and (hysterectomy OR colpohysterectomy OR hysterectom^* OR total hysterectomy^* OR panhysterectom^*). Furthermore, references from identified articles, systematic reviews, and earlier meta-analyses were screened for potential eligibility.

The yielded search outcomes were imported into the Rayyan systematic review website for further screening.²⁹ Titles and abstracts were assessed for eligibility in a standardized manner by 3 independent reviewers (J.S., M.A.W., S.J.D.). Disagreements were resolved by the assessment of the abstract by a fourth reviewer (J.W.M.M.). Data were systematically collected from the full-text articles by 2 authors (S.J.D. and M.A.W.) and recorded in a data extraction sheet.

Eligibility criteria

Studies reporting on sexual function prior to and after hysterectomy for benign nonprolapse indication were included. Only peer-reviewed articles were included. All sexual function instruments were considered for inclusion, such as the Female Sexual Function Index (FSFI),³⁰ the Golombok Rust Inventory of Sexual Satisfaction,³¹ McCoy Female Sexuality Questionnaire,³² Psychosocial Adjustment to Illness Scale,³³ Derogatis Interview for Sexual Functioning,³⁴ Tübinger Scale for Sexual Therapy,³⁵ the MOS Sexual Problems Index,³⁶ the Arizona Sexual Experiences Scale,³⁷ and the Short Personal Experiences Questionnaire.³⁸

Outcome measures

The primary outcome was change in overall sexual function from pre- to posthysterectomy. Dyspareunia was additionally assessed for total vs subtotal hysterectomy. For the role of BSO in changes in sexual function, multiple outcomes were investigated because the endocrinologic consequences of oophorectomy have been hypothesized to affect the whole sexual response curve by reducing central sexual arousability and genital sensitivity.^21,23,25 Therefore, the following outcomes were studied to investigate sexual function before vs after hysterectomy +/–BSO: overall sexual function, desire, arousal, lubrication, and orgasm.

Data extraction and assessment of methodological quality

The 2 first authors (S.J.D. and M.A.W.) extracted data from the eligible studies: the number of participants at available follow-up stages, the year of publication, the surgical approach (route, totality, concomitant BSO, percentage of BSO or unilateral salpingo-oophorectomy), inclusion period, setting, duration of follow-up, type of primary predictor and outcome, sexual function instruments used (including subscales), indications for surgery, menopausal state, age of participants, and the outcome measures—specifically, data on overall sexual function and the subdomains of sexual function (pain, desire, arousal, lubrication, and orgasm).

The methodological quality of the studies was assessed independently by 2 reviewers (M.A.W. and S.J.D.) using the STROBE checklist (Strengthening the Reporting of Observational Studies in Epidemiology).³⁹ The checklist scores the presence of important key elements. If items are adequately and transparently presented, that item is scored as 1; if not, the item is scored 0. Items that are not applicable to a study are scored NA. The reviewers discussed differential scoring and established final agreement. A total score of 33 could be obtained for each study. We present the total scores in Table 1.

Table 1

Study characteristics.

First author (year)	Setting	Design	Primary outcome	Predictor	Sexual function scale	STROBEa	Duration	Follow-up, mo
Kiyak (2021)56	Single center academic	Randomized controlled trial	Vaginal length	TAH +/− uterine manipulator	FSFI	24	2017-2019	3
Skorupska (2021)64	Single center academic	Prospective cohort trial	Sexual function, QoL and UI symptoms	Different types of hysterectomy	FSFI	19	—	12
Taha (2022)57,b	Single center academic	Randomized controlled trial	Sexual function and vaginal length	Vertical vs horizontal cuff closure	FSFI	25	2019-2020	3
Dedden (2020)44	Multicenter	Prospective cohort trial	Sexual function	Hysterectomy	FSFI-SF	23	2010-2014	3
Beyan (2020)65	Single center academic	Prospective cohort trial	Sexual function and QoL	TAH vs TLH	FSFI	21	2014-2018	—
Koşar Can (2020)66	Single center academic	Prospective cohort trial	Sexual function	TAH + BSO	FSFI	20	2017-2018	6
Mahmoud (2020)67	Single center academic	Prospective cohort trial	Sexual function	TAH vs TLH	FSFI	19	2019-2020	6
Doğanay (2019)82	Single-center teaching hospital	Retrospective cohort trial	Sexual function	Hysterectomy +/–BSO	FSFI	22	2013-2015	9-10
Dundar (2019)84	Single center academic	Cross sectional study	Sexual function and vaginal length	Vaginal vs laparoscopic cuff closure	GRISS	17	2016-2017	6
Sukgen (2018)68	Single-center nonteaching hospital	Prospective cohort trial	Sexual function and QoL	Hysterectomy	FSFI	18	2017-2018	12
Kayatas (2017)50	Single center academic	Prospective cohort trial	Sexual function	AH vs LH	FSFI	17	2014-2015	6
Bastu (2016)58	Single center academic	Randomized controlled trial	Sexual function and vaginal length	Vaginal vs laparoscopic cuff closure	FSFI	26	2014-2015	3
Kürek Eken (2016)69	Single center academic	Prospective cohort trial	Sexual function and QoL	AH vs LH	ASEX	21	—	6
Goktas (2015)70	Multicenter	Prospective cohort trial	Sexual function and depression	Hysterectomy +BSO	FSFI sumc	23	2013	1.5
Shiber (2015)71	Single center academic	Prospective cohort trial	Vaginal length	TLH	FSFI	11	2013-2015	6
Radosa (2014)72	Single center academic	Prospective cohort trial	Sexual function and QoL	TLH vs SLH vs VH	FSFI	25	2011-2013	6
Lee (2011)73	Single center academic	Prospective cohort trial	Sexual function	Single port vs conventional LAVH	FSFI	21	2009	6
Sözeri-Varma (2011)74	Single center academic	Prospective cohort trial	Sexual function	Hysterectomy	GRISS	20	2004	6
Ellström Engh (2010)59	Single center academic	Randomized controlled trial	Sexual function and psychological well-being	Subtotal vs total hysterectomy	McCoy total	25	2000-2005	12
Kafy (2009)83	Single-center teaching hospital	Retrospective cohort trial	Patient satisfaction	Subtotal vs total hysterectomy	Self-made (subjective frequency of pain)	16	—	12
Celik (2008)75	Single center academic	Prospective cohort trial	Sexual function	AH + BSO vs VH + BSO	FSFI	22	—	6
Yen (2008)76	Single-center teaching hospital	Prospective cohort trial	Depressive disorder and psychological impact	Hysterectomy	SPEQ	22	2004	4
Bayram (2008)77	Single-center teaching hospital	Prospective cohort trial	Sexual function and depression	Hysterectomy (TAH and VH)	FSFI	21	2004-2005	3
Teplin (2007)85	Multicenter	Secondary analysis	Sexual function and QoL	Hysterectomy +/–BSO	MOS Problems	26	1997-2000	6
Flory (2006)60	Multicenter	Randomized controlled trial	Psychosocial effects	Subtotal vs total hysterectomy	DISF	24	—	6.5
Punushapai (2006)78	Single center academic	Prospective cohort trial	Sexual function	TAH	Self-made (subjective frequency of pain)	20	2004-2006	6
Aziz Brännström (2005)80	Multicenter	Prospective cohort trial	Sex steroids, sexuality and psychological well-being	Hysterectomy +/–BSO	McCoy total	24	1996-1999	14
Kuppermann (2005)61	Multicenter	Randomized controlled trial	Sexual function and HR-QoL	Subtotal vs total hysterectomy	MOS Problems	26	1997-2000	6
Gütl (2002)79	Single center academic	Prospective cohort trial	Sexual function	AH vs VH	TSST and SDS	19	—	3
Hurskainen (2001)62	Multicenter	Randomized controlled trial	QoL	Hysterectomy vs LNG-IUS	McCoy total (modified)	28	1994-1997	12
Alexander (1996)63	Single-center teaching hospital	Randomized controlled trial	Psychiatric and psychosocial aspects	Hysterectomy vs endometrial ablation	PAIS	20	“Period of 20 months”	6
Carlson (1994)81	Multicenter	Prospective cohort trial	QoL	Hysterectomy	Self-made (subjective frequency of pain in last month)	28	1989-1991	12

Abbreviations: —, not available; AH, abdominal hysterectomy; ASEX, Arizona Sexual Experiences Scale; BSO, bilateral salpingo-oophorectomy; DISF, Derogatis Interview for Sexual Functioning; FSFI-SF, Female Sexual Function Index Short-Form; GRISS, Golombok Rust Inventory of Sexual Satisfaction; HR-QoL, health-related quality of life; LAVH, laparoscopy-assisted vaginal hysterectomy; LH, laparoscopic hysterectomy; LNG-IUS, levenorgestrel-releasing intrauterine system; MOS, Medical Outcomes Study Sexual Problems Scale; PAIS, Psychosocial Adjustment to Illness Scale; QoL, quality of life; SDS, (self-developed) Sexual Dysfunction Scale; SLH, supracervical laparoscopic hysterectomy; SPEQ, Short Personal Experiences Questionnaire; TAH, total abdominal hysterectomy; TLH, total laparoscopic hysterectomy; TSST, Tübinger Scale for Sexual Therapy; UI, urinary incontinence; VH, vaginal hysterectomy.

a

Quality assessment: if an item is adequately and transparently presented, it is scored 1; if not, 0. Maximum score per study: 33.

b

Taha (2022) is referred to as Taha (2021) in Figures 2–4 since we referred to the online version when running the analysis.

c

In this study, the FSFI item scores were summed rather than calculated via the FSFI scoring.

Data analysis

Statistical analyses were conducted in the statistical programming environment R and RStudio with the use of the metafor and dmetar packages.^40–43 Standardized mean differences in pre- to postoperative sexual function were estimated and meta-analyzed for all analyses except 1, in which we summarized the posthysterectomy FSFI mean across study groups. The latter analysis was included as a further investigation of PICO 1, namely to assess whether women still report sexual dysfunction posthysterectomy in addition to potential change in sexual function associated with hysterectomy. For the latter, we estimated the FSFI summary mean across groups.

We used random effects models to pool the effect sizes since we anticipated considerable between-study heterogeneity. We controlled for dependency among effect sizes within the multilevel structure of the meta-analysis models. Dependency among effect sizes was due to several effect sizes coming from the same studies because a study frequently assessed several intervention groups, such as different routes of hysterectomy. As for the dependency between pre- and postmeasurements within studies from which change scores were calculated, we used correlations calculated from data of the cohort assessed in Dedden et al in the estimation of the effect sizes to adjust for dependency within the effect size calculation.⁴⁴ We ran sensitivity analyses to check whether the specified strength of dependency affected any of the conclusions.

We used the restricted maximum likelihood estimator to estimate the heterogeneity variances (τ²) and applied an adjustment, similar to the Knapp-Hartung method, to calculate the confidence interval around the pooled effect and estimate the effects’ significance.⁴⁵ Both these specifications follow recommendations.⁴⁶ Effect size heterogeneity was interpreted as follows: 25% = low between-study variance, 50% = moderate, and 75% = high.⁴⁷ To make the standardized mean changes (SMCs) more interpretable, we calculated a weighted SD based on the premeasurement FSFI scores across studies and used this SD to calculate what the SMC would represent in terms of changes on the FSFI.

We diverted from our initial analysis plan (PROSPERO CRD42020177473) in 4 ways. First, we did not include effect sizes based on different sexual function scales from the same study that were measured within the same group of participants. Second, we ran analyses per sexual function subscale (desire, arousal, lubrication, orgasm) directly rather than after an omnibus analysis including all subscales for the analyses regarding +/–BSO. We applied both these deviations because we did not have a good estimate for the within-group dependency between scales measuring the same construct reported on by the same participants. Third, we analyzed sexual function after hysterectomy +/–BSO not only for those studies that directly compared hysterectomy +/–BSO but also across studies that clearly indicated to have included just the women who underwent hysterectomy +/–BSO. Such studies were more prevalent than were initially foreseen, and there are no randomized trials that compare hysterectomy +/–BSO. Finally, we did not include the fourth PICO, in which we intended to analyze additional predictors of differences in change in sexual function (eg, indication), because studies that assessed the same predictors of differences in change were too few in number to summarize in a meta-analysis.

The studies reported different, sometimes multiple, follow-up measurements (eg, after 3, 6, and/or 12 months). One follow-up measurement per study was chosen for the analysis. The chosen moment was the one most frequently reported across all studies or close to that moment, which resulted in the 6-month follow-up being chosen as the reference moment across the studies. We chose 1 similar follow-up measurement rather than, for instance, the last follow-up across studies because we wanted to streamline the follow-up time among studies and minimize this source of potential heterogeneity.

Results

Included studies

The search yielded a total of 4539 records: 3146 from Embase, 1087 from PubMed, and 306 from the Cochrane Library. An overall 1300 duplicates were automatically removed, providing 3239 records. Another 27 records were identified through cross-referencing and going through articles that were collected during our other research projects. Of the remaining 3266 articles, 3107 were excluded by title and abstract. Two reports of the 27 additional records could not be retrieved.^48,49 The QUORUM flow diagram illustrating the selection procedure is shown in Figure 1.

Figure 1

PRISMA flowchart.

A total of 157 reports were assessed for eligibility, of which 125 were excluded for different reasons:

• 17 were not in English.

• 28 did not provide data of baseline and follow-up sexual function.

• 25 reported data about sexual function that were not analyzable or could not be extracted.

• 27 presented percentages of the sample reporting a certain symptom, which we could not analyze within our meta-analysis design.

• 5 cited other factors influencing sexual function and did not focus on hysterectomy.

• 3 were author replies and/or reviews and should have been excluded by abstract.

• 3 had preselected populations that were not comparable to the other studied populations.

• 6 were excluded because the sample was in another article.

• 7 were excluded because they studied an outcome other than sexual function.

• 2 were excluded because they appeared in abstract form and did not provide enough information.

• 1 was excluded because it consisted of qualitative data only.

• 1 was excluded because it was not published in a peer-reviewed journal.

For those studies that fulfilled all other inclusion criteria but did not report on their data in extractable form (6 studies), we sought data on the main outcome and the assessed subdomains. We contacted the corresponding authors but did not receive any response.^50–55

Thirty-two studies were included in this systematic review and meta-analysis: 8 RCTs,^56–63 20 prospective studies,^{44,50,64–81} 2 retrospective studies,^82,83 1 cross-sectional study,⁸⁴ and 1 secondary data analysis.⁸⁵ Of those 32 studies, 28 reported on overall sexual function as an outcome, of which 15 used the FSFI (PICO 1).

The FSFI is the most commonly used instrument to measure female sexual function across studies. It is a self-report questionnaire consisting of 19 items that assess 6 aspects of sexual functioning in women³⁰: desire, arousal, lubrication, orgasm, satisfaction, and pain.

Twenty-two studies reported data on sexual function, and 16 analyzed sexual pain in combination with information on cervix removal (PICO 2), of which 3 and 1 were trials that randomized cervix removal, respectively. Data on overall sexual function and the desire, arousal, lubrication, and orgasm subdomains in combination with information on BSO were studied in 15, 15, 13, 12, and 14 studies. No trial randomizing BSO exists and/or could be found.

The 28 studies reporting on overall sexual function comprised 3437 patients undergoing hysterectomy for benign nonprolapse indications. The 4 other articles comprised 3 studies that reported on 515 additional patients with data on sexual pain and 1 study with 102 participants whose sexual function data were reanalyzed according to whether they had a hysterectomy +/–BSO. We report the number of participants for all outcomes in the forest plots (Figures 2–7). Study characteristics of the studies are described in Table 1. An overview of patient characteristics is presented in Table 2.

Quality of the studies

The methodological quality, as assessed with the STROBE checklist, resulted in a median score of 22, varying between 11 and 28 of the maximum score of 33. An overview of the study characteristics and methodological quality is presented in Table 1.

Statistical results

In Figure 2, we present the standardized change in mean sexual function from pre- to postintervention for hysterectomy in general and subdivided per surgical route (PICO 1). Most studies clearly stated the surgical route. However, in the studies by Aziz et al⁸⁰ and Sözeri-Varma et al,⁷⁴ surgical approach was not specified; therefore, these studies are categorized as “unknown routes.” Furthermore, some studies presented data on sexual function but did not report outcomes separately for surgical routes: these studies are categorized as “mixed routes.” Since we present the sexual function per surgical route, some studies are mentioned multiple times in the forest plot.

Across all types of hysterectomy, we estimated an SMC of 0.09 (95% CI, −0.19 to 0.38; P = .52). If we were to express this effect size in changes based on FSFI scores (according to a weighted SD of 6.4 based on the premeasurement FSFI scores across studies), this SMC equals a change of 0.57 (95% CI, −1.2 to 2.4) on the total FSFI scale (FSFI range, 1.2-36) from pre- to posthysterectomy. For the 3 main hysterectomy routes, all SMC 95% CIs crossed zero, indicating nonsignificant effects (see Figure 2 and Table 3 for the effect sizes and test statistics).

Table 2

Patient characteristics.^a

			Indication,  *  %
First author (year): intervention	Age, y, mean (SD)	Premenopausal, %	HMB	Pain	Fibroid	Endometriosis	Other	N sexual function
Kiyak (2021)56
Conventional TAH	47.1 (3.2)	100	11		67		22	36
TAH with UM	48.2 (2.9)	100	14		67		19	36
Skorupska (2021)64
VH	64.8 (10.4)	18	–	–	–	–	–	103
LSH	47.2 (4.7)	85	–	–	–	–	–	150
SH	47.5 (4.7)	89	–	–	–	–	–	71
TAH	57.9 (9.9)	34	–	–	–	–	–	47
TLH	51.8 (7.7)	57	–	–	–	–	–	28
Taha (2022)57,b
Vertical cuff closure	51.7 (10.3)	70	48		34		18	50
Horizontal cuff closure	50.9 (7.2)	58	62		26		12	50
Dedden (2020)44: hysterectomy	46 (6.9)	–	49	12	34	6	29	260
Beyan (2020)65
TLH	46.4 (5.9)	100	19		65		16	259
TAH	46.9 (6.1)	100	18		66		16	96
Koşar Can (2020)66: TAH + BSO	49.3 (9)	64	25		61		14	56
Mahmoud (2020)67		83	35		42		23
TAH	44.5 (6.7)							30
TLH	44.4 (5.6)							30
Doğanay (2019)82
Hysterectomy +BSO	48.6 (1.2)	100	13		42		45	82
Hysterectomy –BSO	47.2 (1.1)	100	9		46		45	78
Dundar (2019)84
Laparoscopic cuff closure	47.4 (4.1)	–	–	–	–	–	–	75
Vaginal cuff closure	47.2 (4.3)	–	–	–	–	–	–	25
Sukgen (2018)68: hysterectomy	45.1 (8.4)	100	39		46		14	28
Kayatas (2017)50
AH	46 (—)	100	–	–	–	–	–	31
LH	44 (—)	100	–	–	–	–	–	35
Bastu (2016)58
Vaginal cuff closure	47.3 (5.9)	100	–	–	–	–	–	36
Laparoscopic cuff closure	51.3 (9)	100	–	–	–	–	–	34
Kürek Eken (2016)69
TLH	41.1 (5.5)	100	44		20		37	42
TAH	42.8 (2.5)	100	37		27		37	42
Goktas (2015)70: hysterectomy +BSO	46.9 (3.9)	59	–	–	–	–	–	150
Shiber (2015)71: TLH	41.36 (7.16)	–	69	9	15	3	5	22
Radosa (2014)72			10		61	16	13
SLH	46.8 (6.7)	100						72
TLH	45.3 (8.2)	100						98
VH	47.6 (9.4)	100						67
Lee (2011)73: single-port and conventional LAVH	46.2 (5.1)	100	48	36			16	95
Sözeri-Varma (2011)74: hysterectomy	46 (4.7)	88	–	–	–	–	–	40
Ellström Engh (2010)59
Total hysterectomy	44.8 (4.2)	100	75	35	27		19	47
Subtotal hysterectomy	44.9 (4.3)	100	74	42	30		18	42
Kafy (2009)83
SLH	46.1 (7)	–	20	18	62
TLH	46.6 (5.3)	–	8	20	72
Celik (2008)75
AH + BSO	50.5 (3.6)	0	–	–	–	–	–	55
VH + BSO	48.9 (4.4)	0	–	–	–	–	–	37
Yen (2008)76: hysterectomy	46.6 (9.2)	100			63	6	31	68
Bayram (2008)77	45.4 (5.6)	91			64		36
TAH								70
VH								23
Teplin (2007)85
Hysterectomy –BSO	40 (5)	100	–	–	–	–	–	74
Hysterectomy +BSO	45 (4)	100	–	–	–	–	–	28
Flory (2006)60
LAVH	43.8 (5.8)	100	59		66	19		32
Subtotal hysterectomy	43.5 (4.1)	100	61		87	10		31
Punushapai (2006)78: TAH	41.5 (4)	100	40	19	39		3
Aziz Brännström (2005)80			100
Hysterectomy –BSO	48.3 (2.5)	100						207
Hysterectomy +BSO	50 (2.3)	100						99
Kuppermann (2005)61
Subtotal hysterectomy	41.8 (5.1)	100	71	75	79		30	42
Total hysterectomy	41.8 (5.2)	100	73	70	79		26	44
Gütl (2002)79
AH	53.4 (9.3)	57	18	46	39			44
VH	51.0 (7.6)	67	26	57	46			46
Hurskainen (2001)62: hysterectomy	43 (3.2)	100	100					112
Alexander (1996)63: hysterectomy	40.3 (5.2)	100	100					84
Carlson (1994)81: hysterectomy	–	–	15	3	39	20	23

Abbreviations: —, not available; AH, abdominal hysterectomy; BSO, bilateral salpingo-oophorectomy; HMB, heavy menstrual bleeding; LAVH, laparoscopy-assisted vaginal hysterectomy; LH, laparoscopic hysterectomy; LSH, laparoscopic supracervical; SH, supracervical hysterectomy; TAH, total abdominal hysterectomy; TLH, total laparoscopic hysterectomy; UM, uterine manipulator; VH, vaginal hysterectomy.

a

Blank cells indicate not applicable.

b

Taha (2022) is referred to as Taha (2021) in Figures 2–4 since we referred to the online version when running the analysis.

c

Indications do not always sum to 100% since some articles reported multiple/overlapping indications and some reported only a few of them.

For the overall effect, we observed large between-study heterogeneity variance and some within-study heterogeneity variance (τ² = 0.51 [95% CI, 0.29-0.96] and τ² = 0.05 [95% CI, 0.01-0.13], respectively) resulting in large total heterogeneity (I² = 97% [95% CI, 95%-98%]). In terms of I² values, 89% of the total variation can be attributed to between-study heterogeneity and 8.3% to within-study heterogeneity. The H² value of 34 confirmed that there was more between-study heterogeneity than would be expected per the sampling error alone.

The prediction interval for hysterectomy in general ranged from an SMC of −1.4 to 1.6, which shows that future studies are likely to find deterioration and improvement in sexual function associated with hysterectomy (ie, changes of −9.2 to 10 points on the total FSFI). The prediction intervals based on the models for the 3 main routes resulted in the same conclusion for each route. For the subsequent analyses, we refer to the figures and Table 3 for statistical information, while we discuss only the implications in the text.

Table 3

Effect sizes and test statistics of all meta-analysis models.^a

					I2			τ2
Model	Effect size	95% CI	95% PI	P value	T	B	W	B	W	H2
Sexual function: hysterectomy overall	0.09	−0.19, 0.38	−1.4, 1.6	.52	97	89	8.3	0.51	0.05	34
Sexual function: laparoscopic route	0.25	−0.35, 0.85	−1.8, 2.3	.39	98	87	11	0.77	0.10	46
Sexual function: vaginal route	−0.16	−1.0, 0.69	−2.2, 1.9	.62	95			0.44		21
Sexual function: abdominal route	0.12	−0.19, 0.43	−1.1, 1.3	.43	93	66	27	0.20	0.08	15
FSFI posthysterectomy	22	20, 25	12, 33		99	91	7.9	23	2.0	99
Sexual function: subtotal hysterectomy	0.34	0.09, 0.60	−0.18, 0.87	.02	65	51	15	0.02	0.01	2.9
Sexual function: total hysterectomy	0.20	−0.15, 0.55	−1.5, 1.9	.25	97	84	13	0.56	0.09	36
Sexual function: subtotal hysterectomy (direct comparison)	0.34	0.09, 0.60	−0.18, 0.87	.02	65	51	15	0.02	0.01	2.9
Sexual function: total hysterectomy (direct comparison)	0.30	−0.16, 0.75	−0.78, 1.4	.17	90	84	6.1	0.16	0.01	9.8
Sexual function: subtotal hysterectomy (RCT)	0.46	−0.14, 1.1	−0.48, 1.4	.08	49			0.03		2.0
Sexual function: total hysterectomy (RCT)	0.49	−0.50, 1.5	−1.4, 2.3	.17	82			0.13		5.5
Sexual pain: subtotal hysterectomy	−0.21	−1.1, 0.70	−1.9, 1.5	.43	84			0.11		6.1
Sexual pain: total hysterectomy	−0.40	−0.72, −0.09	−1.7, 0.92	.01	97	62	35	0.24	0.14	29
Sexual pain: subtotal hysterectomy (direct comparison)	−0.21	−1.1, 0.70	−1.9, 1.5	.43	84			0.11		6.1
Sexual pain: total hysterectomy (direct comparison)	−0.36	−1.7, 0.96	−3.0, 2.3	.45	96	91	5.5	0.48	0.03	28
Sexual function: hysterectomy –BSO	0.59	0.22, 0.96	−0.72, 1.9	.003	96	92	3.5	0.35	0.01	23
Sexual function: hysterectomy +BSO	0.29	−0.54, 1.1	−1.9, 2.5	.43	97	95	2.4	0.65	0.02	39
Sexual function: hysterectomy –BSO (direct comparison)	0.88	−1.8, 3.6	−4.5, 6.3	.30	99			1.2		77
Sexual function: hysterectomy +BSO (direct comparison)	0.69	−1.7, 3.1	−4.0, 5.4	.34	97			0.90		38
Sexual desire: hysterectomy –BSO	0.47	0.14, 0.80	−0.78, 1.7	.008	96	48	47	0.17	0.16	22
Sexual desire: hysterectomy +BSO	0.18	−0.45, 0.81	−1.5, 1.8	.52	96	80	16	0.33	0.06	26
Sexual desire: hysterectomy –BSO (direct comparison)	0.81	−1.8, 3.4	−4.3, 5.9	.31	99			1.1		87
Sexual desire: hysterectomy +BSO (direct comparison)	0.55	−1.2, 2.3	−2.8, 3.9	.30	96			0.46		28
Sexual arousal: hysterectomy –BSO	0.34	0.04, 0.63	−0.69, 1.4	.03	94	47	46	0.11	0.11	15
Sexual arousal: hysterectomy +BSO	0.02	−0.50, 0.55	−1.3, 1.4	.91	94	0.0	94	0.00	0.24	17
Sexual arousal: hysterectomy –BSO (direct comparison)	0.29	−4.8, 5.3	−8.4, 8.9	.60	97			0.31		30
Sexual arousal: hysterectomy +BSO (direct comparison)	0.25	−3.6, 4.1	−6.3, 6.8	.57	94			0.18		16
Lubrication: hysterectomy –BSO	0.36	0.06, 0.65	−0.55, 1.3	.02	92	84	8.3	0.14	0.01	12
Lubrication: hysterectomy +BSO	−0.22	−0.77, 0.33	−1.6, 1.1	.35	94	50	44	0.12	0.11	17
Lubrication: hysterectomy –BSO (direct comparison)	0.49	−6.6, 7.6	−12, 13	.54	98			0.62		51
Lubrication: hysterectomy +BSO (direct comparison)	−0.41	−3.9, 3.1	−6.3, 5.5	.37	92			0.14		12
Orgasm: hysterectomy –BSO	0.28	0.11, 0.45	−0.27, 0.84	.003	81	66	15	0.05	0.01	5.3
Orgasm: hysterectomy +BSO	−0.06	−0.38, 0.26	−0.85, 0.73	.66	85	75	10	0.08	0.01	6.7
Orgasm: hysterectomy –BSO (direct comparison)	0.38	−0.66, 1.4	−1.6, 2.4	.26	94			0.16		16
Orgasm: hysterectomy +BSO (direct comparison)	0.02	−0.41, 0.46	−0.64, 0.69	.83	46			0.01		1.8

Abbreviations: B, between-study variance; BSO, bilateral salpingo-oophorectomy; FSFI, Female Sexual Function Index; PI, prediction interval; RCT, randomized controlled trial; T, total variance; W, within-study variance.

a

Blank cells indicate not applicable.

Figure 2

Overall sexual function analyzed overall but visually subdivided according to surgical route (total, n = 3437; unknown, n = 346; mixed, n = 773; abdominal, n = 980; vaginal, n = 276; laparoscopic, n = 1062). The I² and H² statistics and heterogeneity variances (τ) refer to the total variability (between- and within-study levels combined). The between and within components are reported in Table 3. The sample size refers to that at follow-up measurement. Interventions are coded as follows: 1, vaginal hysterectomy; 2, laparoscopic hysterectomy; 2A, subtotal laparoscopic hysterectomy; 2B, total laparoscopic hysterectomy; 2C, laparoscopy-assisted vaginal hysterectomy; 3, abdominal hysterectomy; 3A, subtotal abdominal hysterectomy; 3B, total abdominal hysterectomy; 4, several surgical routes; 5, surgical route unknown. The dashed vertical lines to the left and right of zero effect indicate 0.5 positive or negative standardized change, which is frequently considered clinically significant.⁸⁶

In Figure 3, we show the meta-analytic summary of the raw posthysterectomy mean FSFI score across different types of hysterectomy to assess whether women show a tendency for sexual function/dysfunction posthysterectomy. The estimated FSFI across all groups was 22 (95% CI, 20-25), indicating that women tend to report remaining sexual dysfunction after hysterectomy. However, the 95% prediction interval spanned an FSFI of 12 to 33, indicating that future studies are likely to find a tendency for reporting sexual dysfunction as well as not after hysterectomy. Again, this uncertainty in the estimation resulted from large effect size heterogeneity (Figure 3, Table 3). The H² statistic confirmed that there was more between-study heterogeneity than would be expected from sampling error alone.

Figure 3

Female Sexual Function Index–based sexual function/dysfunction after hysterectomy (n = 1905). The vertical line demarcates a total score of 26.55, which has been shown to indicate sexual dysfunction.⁸⁷

In Figure 4, we present the standardized change from pre- to posthysterectomy divided according to whether the cervix was removed for mean overall sexual function and mean sexual pain in panels A and B, respectively (PICO 2). Subtotal but not total hysterectomy was associated with improvement in sexual function from pre- to posthysterectomy (Figure 4A, Table 3). However, for sexual pain, this pattern was reversed: total hysterectomy was associated with significant improvement in mean sexual pain while subtotal hysterectomy was not (Figure 4B, Table 3). Again, we observed large between-study heterogeneity variance and some within-study heterogeneity variance for all of the effect sizes, with the exception of changes in mean sexual function associated with subtotal hysterectomy, which showed moderate between-study heterogeneity (Figure 4A, Table 3). The H² statistics confirmed that there was more between-study heterogeneity than would be expected by sampling error alone for all effect sizes (Figure 4, Table 3). This pattern of results remained largely the same when we reran the analyses based on only those studies that compared total vs subtotal hysterectomy within the trial, with the exception that the significant improvement in sexual pain for total hysterectomy became nonsignificant (Table 3 and Figure 4B in section 1 of Supplemental Material A).

Figure 4

Overall sexual function and sexual pain divided and analyzed according to subtotal or total hysterectomy: (A) total sexual function, n = 2329; subtotal sexual function, n = 408; (B) total pain, n = 2057; subtotal pain, n = 154.

To clarify these results, we meta-analyzed only the 3 randomized controlled trials that compared sexual function after total vs subtotal hysterectomy (Figure 5). No significant change in sexual function was noted after total or subtotal hysterectomy, and cervix removal did not relate with effect size heterogeneity. Effect size heterogeneity was still substantial, especially in the total hysterectomy group, despite the fact that these effect sizes are based on RCT data with more stringent inclusion criteria.

Figure 5

Overall sexual function assessed in randomized controlled trials divided and analyzed according to subtotal or total hysterectomy (total sexual function, n = 123; subtotal sexual function, n = 115).

It is important to stress that removal of the cervix was not significantly associated with effect size heterogeneity for sexual function or sexual pain (see tests for difference between totality in Figures 4 and 5). That is, no significant differences in effect sizes were found between subtotal and total hysterectomy for either overall sexual function or sexual pain (Figure 4), despite the fact that the standardized change in mean overall sexual function for subtotal hysterectomy and that in mean sexual pain for total hysterectomy turned out to be significantly different from zero, while those for total and subtotal were not, respectively (Figure 4). Difference in significance does not equal a significant difference. Practically, this implies that there is no available evidence indicating that subtotal or total hysterectomy has proven to be more beneficial or disadvantageous with regard to associated changes in sexual function or sexual pain.

In Figure 6, we show the standardized change in mean sexual function divided according to +/–BSO (PICO 3). Hysterectomy –BSO was associated with a significant improvement in overall sexual function, whereas hysterectomy +BSO was associated with a nonsignificant improvement in overall sexual function. However, the difference in effect sizes between hysterectomy +BSO and –BSO turned out nonsignificant, indicating that we cannot conclude that hysterectomy –BSO is superior to hysterectomy +BSO in terms of associated changes in sexual function. Again, both effect sizes showed large between-study heterogeneity, and the prediction intervals indicated that future studies are likely to find improvement and deterioration in sexual function associated with hysterectomy +BSO and –BSO. Exactly the same pattern of results was found for sexual desire and sexual arousal: hysterectomy –BSO was associated with significant improvements in sexual desire and arousal, whereas hysterectomy +BSO was associated with a nonsignificant improvement, with the difference in effect sizes between +BSO and –BSO being nonsignificant (see Table 3 and Figures 6A and 7A in section 1 of Supplemental Material A). Note that this pattern of results could not be replicated when running the analysis only on those studies that directly compared hysterectomy +/–BSO (Table 3). That is, when studies were summarized that compared +/–BSO internally, neither hysterectomy +BSO nor hysterectomy –BSO was associated with a significant change in sexual function, nor was the difference in effect size significantly different.

Figure 6

Sexual function divided and analyzed according to +/− BSO status (hysterectomy +BSO sexual function, n = 507; hysterectomy –BSO sexual function, n = 1436). BSO, bilateral salpingo-oophorectomy.

In Figure 7, we show the standardized change in mean lubrication and orgasm divided according to +/–BSO (PICO 3). Hysterectomy –BSO was associated with a significant improvement in lubrication and orgasm, whereas hysterectomy +BSO was associated with a nonsignificant deterioration in lubrication and orgasm. The difference in effect sizes between hysterectomy +BSO and –BSO turned out significant, indicating that –BSO appeared to be superior to hysterectomy +BSO in associated improvement in lubrication and orgasm from pre- to posthysterectomy. However, again, both effect sizes showed large between-study heterogeneity, and the prediction intervals indicated that future studies are likely to find improvement and deterioration in lubrication and orgasm associated with hysterectomy +BSO and –BSO. Also, note that these results could not be replicated when running the analyses only on those studies that directly compared hysterectomy +/–BSO, as was the case for sexual function, desire, and arousal (Table 3).

Figure 7

Lubrication and orgasm divided and analyzed according to +/– BSO status: (A) hysterectomy +BSO lubrication, n = 479; hysterectomy –BSO lubrication, n = 1235; (B) hysterectomy +BSO orgasm, n = 505; hysterectomy –BSO orgasm, n = 1372.

Sensitivity analyses

We reran all analyses with different strengths of pre- to postdependency specified (r = 0.3 and r = 0.9) instead of the estimates based on the cohort of Dedden et al⁴⁴ (overall sexual function, r = 0.615; sexual pain, r = 0.478; sexual desire, r = 0.571; sexual arousal, r = 0.606; lubrication, r = 0.598; orgasm, r = 0.611) to check whether any of our conclusions were sensitive to the strength of dependency specified. As expected, the effect size estimates changed (see figures in sections 2 and 3 in Supplemental Material A) but the pattern of results remained the same, and none of the conclusions were affected.

Discussion

Main findings

This systematic review and meta-analysis showed that hysterectomy performed for a benign indication is not associated with significant changes in female sexual function, regardless of surgical route. However, populations represented in the literature have been too heterogeneous to ascertain the direction of changes in sexual function associated with hysterectomy, irrespective of surgical route.

Subtotal hysterectomy was associated with an improvement in sexual function and total hysterectomy with an improvement in sexual pain; however, the difference in change was not statistically significant, and the results regarding sexual function did not replicate within studies or RCTs that directly compared removal of the cervix.

Hysterectomy –BSO systematically showed significant associated improvements in sexual function, desire, arousal, lubrication, and orgasm, which differed significantly from the nonsignificant deterioration associated with hysterectomy +BSO in lubrication and orgasm but not from the nonsignificant improvement associated with hysterectomy +BSO in sexual desire, arousal, and overall sexual function. Note, however, that the results regarding hysterectomy +/–BSO could not be replicated when studies were analyzed that directly compared hysterectomy +/–BSO.

Importantly, women tended to report remaining sexual dysfunction posthysterectomy despite some of the analysis suggesting that hysterectomy can result in improvements in different domains of sexual functioning. Overall, all analyses showed considerable heterogeneity in effect sizes, specifically between-study heterogeneity. The latter indicates that studied populations of women who underwent hysterectomy have been too heterogeneous to ascertain the direction of changes in sexual function associated with hysterectomy and the tendency to report sexual dysfunction posthysterectomy for all women undergoing hysterectomy.

Interpretation of the findings and clinical implications

This systematic review and meta-analysis showed that hysterectomy is not associated with significant change in overall sexual function, irrespective of surgical route. However, the studies were very heterogeneous. These results are in line with previous trials citing similarly heterogeneous outcomes.⁸⁸

In 2 parallel randomized trials, the eVALuate study showed that sexual activity questionnaire “habit” scores were higher at 6 weeks after LH than after AH. No difference was found when LH was compared with VH.⁸⁹ In a prospective observational study, Abdelmonem et al observed a significant change in vaginal length in the VH group but not in the total AH group.⁹⁰ However, Lermann et al, who compared AH, VH, laparoscopy-assisted VH, laparoscopic supracervical hysterectomy, and total LH did not observe any differences with regard to the prevalence of hypoactive sexual desire disorder after hysterectomy among surgical techniques.⁹¹ Also, Kluivers et al did not find any significant differences in terms of sexual functioning between LH and AH in their RCT.⁹²

As such, the possible role of the cervix in sexuality remains uncertain,⁹³ and so far, there is no conclusive evidence whether the cervix is vital for sexual function posthysterectomy. However, previous trials and this meta-analysis suggest that total hysterectomy is not inferior to subtotal hysterectomy regarding changes in sexual function from pre- to posthysterectomy. Learman et al found no statistically significant differences in the degree of clinical outcomes between supracervical hysterectomy and total AH.⁹⁴ In the trial by Thakar et al, comparing total vs subtotal hysterectomy, “frequency of intercourse and orgasm” and “rating of sexual relationship with a partner” did not change significantly in either group after surgery.⁹⁵ This finding was confirmed in an RCT by Gimbel et al.⁹⁶

The current meta-analysis showed that women who did not undergo concomitant BSO at the time of hysterectomy for benign indication systematically reported sexual function improvements in overall sexual function and sexual subdomains on average. However, future studies need to establish this observation more conclusively, since this finding could not be replicated within those studies that compared hysterectomy +/–BSO directly. Nevertheless, this finding suggests that oophorectomy should not be considered if not medically necessary, especially given that risk-reducing salpingectomy has recently been shown to lead to better sexual functioning as compared with risk-reducing salpingo-oophorectomy, even for women who use hormone replacement therapy after the latter.^97,98

A possible explanation for the heterogeneity in effects may be that it is not the procedure of a hysterectomy itself but the indication for surgery that predicts the magnitude of change in sexual function from pre- to posthysterectomy. It is possible that women undergoing hysterectomy for fibroids or heavy menstrual bleeding experience an improvement in sexual function because they are relieved from their complaints, while women undergoing surgery for chronic pelvic pain may still experience pain after hysterectomy and therefore no improvement in sexual function. It has been established that women with sexual dysfunction prior to hysterectomy continue to report sexual dysfunction after hysterectomy.⁴⁴ Future research should study sexual function changes associated with hysterectomy within groups that share the indication for surgery.

Understandably but unfortunately, many of the studies included only women who were sexually active at both time points. Sexual activity was frequently conceptualized as engaging in penetrative sexual intercourse with a partner, since the validated and frequently utilized FSFI is based on penetrative partnered sexual activity. On one hand, such subselection may lead to studied samples being biased toward better sexual function pre- and posthysterectomy, since those women who refrained from partnered sexual intercourse before or after the intervention were excluded from the analysis in the study. On the other, women who did not engage in sexual intercourse or partnered sex before hysterectomy but started engaging in such activity after hysterectomy might have benefited from the intervention but were excluded as well, with such samples being biased toward underestimation of treatment improvement. Unfortunately, percentages of sexually active women before and after intervention were infrequently reported and would not have allowed for analysis in terms of pre- to postintervention changes since we do not have access to individual data. Future studies should consider including measures of frequency of partnered and solo sexual activity, the subjective experience thereof, and sexologic outcome measures that are not dependent on sexual activity (eg, measures of sexual desire and sexual thoughts and fantasy).

The analysis further showed that women undergoing hysterectomy still tended to report sexual dysfunction posthysterectomy. In other words, hysterectomy does not appear to systematically deteriorate or improve sexual function on average and therefore does not seem to represent a therapeutic sexologic intervention. These results suggest that clinicians should realize that women might still desire and require help to improve their sexual well-being posthysterectomy and that prehysterectomy counseling should include discussing the availability of posthysterectomy support, if so desired and required.

Strength and limitations

Our systematic review and meta-analysis have several strengths: they were based on a comprehensive systematic search and clinically relevant questions. We included a comprehensive array of hysterectomy trials because we were able to pool different instruments measuring sexual function, leading us to an extensive 32 trials comprising 4054 patients. We analyzed pre- to posthysterectomy change in sexual function rather than posthysterectomy sexual functioning only, allowing insights into improvement and deterioration associated with hysterectomy and the role of removal of the cervix and concomitant BSO in differences in such change.

Another recent systematic review and meta-analysis on sexual function after hysterectomy for benign disease yielded similar conclusions based on a smaller subset of studies in the current analysis and a different analysis approach.⁸⁸ Our systematic review and meta-analysis contain 21 additional trials and aimed to give insight into clinically relevant questions by giving an overview of all trials to date that assessed change in sexual function associated with hysterectomy, rather than differences in sexual function between women who have and have not undergone hysterectomy.

Our systematic review and meta-analysis suffered from a few limitations. First, many effects suggested that hysterectomy is associated with no change in sexual function. Unfortunately, our analysis paradigm does not allow us to conclude that hysterectomy is therefore associated with no changes in sexual function, since absence of evidence does not equal evidence of absence. Future clinical trials and meta-analyses might consider using bayesian analyses, which allow testing for evidence of absence (ie, confirming the null hypothesis of no effect or no difference).^99,100 Second, random effects multilevel models do not allow for analysis of publication bias and diagnostics in a straightforward manner,¹⁰¹ which is why we did not include such analyses in the current study. Third, the methodological quality of studies was moderate, which suggests that future trials should better control for potential confounding and use adequate design and outcome measures to ascertain our conclusions. Fourth, the analysis on sexual dysfunction posthysterectomy was based only on FSFI scores, which does not cover the associated distress required for sexual dysfunction. When studying sexual dysfunction associated with hysterectomy, future research in women undergoing hysterectomy should ascertain that woman are distressed with the sexual problems that they might be experiencing.

Conclusion

This systematic review and meta-analysis showed that hysterectomy is not associated with significant change in overall sexual function for the average patient undergoing hysterectomy for a benign indication and that removal of the cervix is not associated with differences in the magnitude of change. Hysterectomy –BSO is associated with significant improvement in several domains of sexual function. However, it remains uncertain whether hysterectomy +BSO results in significantly worse sexual function than hysterectomy –BSO. Nevertheless, the results suggest that oophorectomy should not be considered unless medically necessary, since hysterectomy –BSO was more systematically associated with significant improvement in sexual function than hysterectomy +BSO.

One must be aware that the available literature shows large heterogeneity in changes in sexual function associated with hysterectomy. The source of heterogeneity needs to be studied systematically in future research—for instance, by comparing sexual function changes associated with hysterectomy indicated for different complaints. Also, future research should assess which women show deterioration, improvement, or no change in sexual function from pre- to posthysterectomy and which characteristics predict or determine these different trajectories, since there does not seem to be an average effect of hysterectomy on sexual function for all women. In such studies, researchers should (1) assess pre- to postoperative sexual function; (2) report outcomes according to route, concomitant procedures, and indication; and (3) use a core outcome set so that future meta-analyses are able to present more conclusive evidence. Our findings can be used by clinicians to make sexual function a subject of discussion. It is important to discuss and assess preoperative sexual function to better counsel patients about what to expect after hysterectomy. If desired and required, sexologic intervention posthysterectomy remains vital.

Acknowledgments

We thank Eugenie Delvaux of the Maxima Medical Centre for her help on formulating and executing the search strategy.

Conceptualization: S.J.D., M.A.W., J.W.M.M. Methodology: S.J.D., M.A.W., B.I.L.W., J.W.M.M. Investigation: S.J.D., M.A.W., J.S., J.W.M.M. Writing–original draft: S.J.D., M.A.W. Writing–review and editing: S.J.D., M.A.W., J.S., B.I.L.W., J.A.F.H., P.M.A.J.G., J.W.M.M. Supervision: B.I.L.W., J.A.F.H., P.M.A.J.G., J.W.M.M.

Funding

None declared.

Conflicts of interest: J.A.F.H. received various research grants (Samsung, NWO/TTW/ZonMw) and participated in an international course sponsored by Samsung and a congress on surgical treatment by Olympus for which she received compensation for travel expenses—all of which were not related to the studied subject. She is chair of the niche working group of the ESGE (European Society of Gynaecological Endoscopy) and chair of the research committee of the NVOG (Dutch Society for Obstetrics and Gynaecology). The other authors have nothing to disclose.

Author notes

S.J.D. and M.A.W. contributed equally.