Does perinatal period pelvic floor muscle exercises affect sexuality and pelvic muscle strength? A systematic review and meta-analysis of randomized controlled trials

SUMMARY OBJECTIVE: The aim of this study was to systematically review the effect of pelvic floor exercises on female sexual function and pelvic floor strength in the prenatal and postnatal periods and to conduct a meta-analysis of available evidence. METHODS: Published archives, including PubMed, Cochrane Library, Web of Science, and ULAKBİM databases, were scanned using keywords based on MeSH. Only randomized controlled trials were included. The data were analyzed using the Review Manager computer program (version 5.3). RESULTS: Pooled standardized differences in means of sexual function in both pelvic floor exercise and control group were 6.33 (95%CI 5.27–7.40, p<0.00001) during pregnancy. The pooled standardized differences in means in sexual function after postpartum intervention was 1.19 (95%CI 0.08––2.30, p=0.04). CONCLUSION: Evidence has shown a little effect on the pelvic floor muscle training on sexual function in pregnancy and postpartum period in primipara women, and it is a safe strategy that can improve postpartum sexual function.


INTRODUCTION
Sexuality is a natural and important part of human life 1,2 . Sexual dysfunction is defined as a disorder affecting sexual desire that can result in interpersonal difficulties, pronounced distress, and psychophysiological changes 3,4 .
The etiology of female sexual dysfunction has a multifactorial structure [5][6][7] . Especially during pregnancy and after delivery, deterioration of pelvic muscle strength (PMS) is an important risk factor 3,7,8 . The literature has shown that pelvic floor muscle exercise (PFME) can improve sexual desire and orgasm capacity in the general population and in women with weak orgasm problems caused by poor pelvic muscle tone 8,9 . The literature on the effects of PFME on female sexual function (SF) has limited edition reviewed, especially for its efficacy during pregnancy and postpartum, and only two studies were meta-analyzed 6,7 . Therefore, the aim of this study was to systematically review the effect of PME on female SF and PMS in the prenatal and postnatal periods and to conduct a meta-analysis of available evidence.

METHODS
Systematic examination and meta-analysis of the studies evaluating the effect of PMFE on female SFs and pelvic floor strength in the prenatal and postnatal periods were performed. In the preparation of systematic review and meta-analysis, the criteria in the PRISMA and Cochrane Experiments Systematic Reviews Handbook were used.

Search strategy
A comprehensive, systematic search of PubMed, Web of Science, the Cochrane Library, and ULAKBİM databases was completed from the earliest date available until February 2020. The Web of Science Core Collection was searched using the following keywords: "pelvic floor exercise" OR "pelvic muscle strength" OR "sexual functions" AND "pregnancy" OR "postpartum." The search strategy was changed according to the characteristics of each database.

Inclusion and Exclusion Criteria
The inclusion and exclusion criteria used were as follows: (1) only randomized controlled trials (RCTs) were included in the study; (2) in the intervention group, pelvic floor exercises to improve the pelvic floor were included if Kegel, Pilates, or yoga were used; (3) studies that included effects of PFME on at least one SF variable including prenatal or postpartum desire, arousal, orgasm, pain, lubrication, and satisfaction; and (4) studies published only in English and Turkish languages were included.

Study selection and data extraction
After the duplicate articles retrieved from the different databases were removed, two independent researchers (A.Y.K. and N.E.B.) screened titles and abstracts to identify which studies met the inclusion and exclusion criteria. If there was a contradiction between the researchers, the third researcher (N.G.) was assisted to reach an agreement. Data were obtained using standard data extraction forms including study characteristics (i.e., design, population, experiments, and result), PICOS (participant, experiment, comparison, outcomes, and study design) approach, age, gender, and follow-up time ( Table 1).

Risk of bias assessment
The quality of the selected articles was evaluated by two researchers (A.Y.K. and N.E.B.) with the Quality Assessment Tool (EPHPP) checklist. The evaluation of the risk of bias of all selected articles was done by two authors (A.Y.K. and N.B.E.) independently using modified Cochrane tools for assessing risk of bias, following the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions. The other author (N.G.) checked the results. Risk of bias was classified into seven domains. The bias risk for each area was classified as "low risk," "high risk," or "uncertain risk," according to the decision criteria in the "Risk of bias" assessment tool.

Quantitative data synthesis and analysis
Outcomes data including SF and PMS of the participants who had used PMEs were collected for analysis. Meta-analysis of study outcomes was performed using RevMan version 5.3. For analysis of continuous data, mean differences (MD) or standardized mean differences (SMD) with 95% confidence intervals (CI) were used. SMD was used when the studies assess the same outcome but measure it in a variety of ways (e.g., all studies measure function, but they use different psychometric scales such as Female Sexual Functıon Index [FSFI] and Golombok-Rust Sexual Satisfaction Scale [GRISS]). Statistical heterogeneity was determined by I 2 . A value of 0% indicates no observed heterogeneity, and larger values indicate increased heterogeneity. Coherence between researchers for independent article selection and bias scores was evaluated using the Cohen's kappa statistic. Only 62.5% (n=5) of the studies were graded 1 according to the EPHPP tool. Coherence between the observers was excellent both in the selection of articles and in the scoring of selected articles in terms of bias (Cohen's kappa was 0.95 for article selection and 0.97 for bias scoring; p=0.000).

Literature search
The PRISMA flowchart for searching and selecting literature is summarized in Figure 1.
The electronic database search and hand search yielded 369 potentially relevant studies. After removing duplicates, we screened 339 articles based on title or abstract. The remaining 33 full texts were assessed for eligibility. For the full-text screening, a third reviewer was needed to resolve disagreements, all regarding the blinding of the studies. Eight trials met all eligibility criteria and were included in qualitative synthesis ( Figure 1).

Study characteristics
Eight trials (896 participants in total) were included in these reviews and meta-analysis [10][11][12][13][14][15][16][17] . The features of the studies are summarized in Table 1. All other studies started in the postpartum period except for the two studies (started during pregnancy) [10][11][12][13] . The duration of the experiments varies between 4 and 20 weeks. In most of the articles, women in the control group received routine postpartum care. However, in one study, the control group received conventional home-based training 17 . Women in the intervention group received the following treatments: those in Wang et al 17   -The pelvic floor muscle training practice guide is as follows: participants were directed to lie in the supine position bending the hips and knees and to relax the abdomen and hip muscles while selectively contracting and breathing the urethra, vagina and anus muscles. Meanwhile, the researchers were asked to evaluate on the abdomen, and on the other hand, by placing the participants on their hips and / and middle fingers in their vagina with postpartum sterile gloves to determine if the participants could properly contract and contract these muscles.

Outcome Measures
The forest graphic in Figure 2 shows us the meta-analysis of the effect of PMFE on SF. Four studies used the FSFI questionnaire [10][11][12]17 to evaluate SF.

Effect of exercises on SF
Six articles 10-12,15-17 reporting on SF were included in the meta-analysis. In the prenatal period, only one study reported sexual results. Figure 2 shows the effects of pelvic floor exercises on SF during pregnancy. The study 10 , which included 82 participants in total (41 receiving PFME), examined the effects of PFME on SF. Pooled SMDs of SF in both PME and control groups were 6.33 (95%CI 5.27-7.40, p<0.00001). When we evaluated PFME sexual status in the postpartum period, based on the random-effects model, SMDs of SF in both PME and control groups were 1.19 (95%CI 0.08-2.30, p=0.04). A meta-analysis of these studies revealed that PME can improve SF in the postpartum period. The included studies had high heterogeneity (I2=83.0%; p=0.02). The forest plot is shown in Figure 3.

Pelvic floor muscle strength
In the three studies, 10-12 the Oxford grading system, an accepted international method for determining the strength of the pelvic floor muscles (PFMs), was used. Other studies have evaluated PMS with different assessment tools. Eight articles have reported PMS-related results and are included in the meta-analysis [10][11][12][15][16][17] . The PMS SMDs in both groups were 1.06 (95%CI 0.12-1.99, p=0.03). Meta-analysis of these studies showed a significant relationship between PFME and PMS. The included studies had high heterogeneity (I 2 =94.0%, p<0.00001). The forest plot of the meta-analysis is shown in Figure 3.

Risk of bias assessment
All studies have identified a sufficient method for random allocation of participants to exercise groups [10][11][12][13][14][15][16][17] . Six studies reported adequate allocation confidentiality using sequentially numbered and sealed opaque envelopes [10][11][12]14,15,17 and evaluated them at low risk of bias. In all studies included in the meta-analysis, it was not possible for the participants and researchers participating in the experiment to be blind to the study. [10][11][12]14,15,17 Four studies are at low risk for blinding outcome evaluation. 10,11,15,17 Other studies have also been evaluated without blinding the outcome assessment and as at a high risk of bias. [12][13][14]16 In these six studies, the drop-outs were balanced between the intervention and control groups, or there were too few drop-outs to affect the study. 10,11,[14][15][16][17] Apart from the study by Cıtak et al 12 (uncertainty bias risk), in all other methods of work, they were evaluated at the risk of reporting low bias, as they discussed important reported results, including negative results and match those reported in their records or protocols 10,11,15 . In particular, we expected a conflict of interest statement and a source of funding. None of the included studies reported other bias risk (Figure 4).

DISCUSSION
The purpose of this meta-analysis was to evaluate the effectiveness of PFMEs on SF in women during pregnancy and the postpartum period. Across the included studies, we examined whether there is evidence that PFMEs improve SF and PMS during pregnancy or in the postpartum period.
Although there was a significant increase in SF status as a result of using PFMEs, which we considered as the primary outcome in the examination, the evidence was generally of low quality. Therefore, we needed higher quality RCTs in this area to provide a more definitive answer. In addition to the study by Pourkhiz et al 10 , eight studies on PFMEs do not provide sufficient data to evaluate the effect on SF during pregnancy, but according to the analysis of studies in postpartum period, the use of PFMEs resulted in a statistically significant increase in SF 11,12,15 . Although most studies show an improvement in SF, the results should be interpreted with caution due to the methodological limitations of some studies. Çıtak et al 12 had a high rate of attrition. However, high heterogeneity between studies is remarkable. Hadizadeh-Talasaz et al. 18 in their meta-analysis found that women who performed postpartum PFMEs showed a slight improvement in SF problems. Wu et al. 19 reported a decrease in unsatisfactory SF in their meta-analysis.
According to some studies, in addition to sexual dysfunction during pregnancy and the prevalence of PMS, incontinence and quality of life are increasing 19,20 . The effect of PFME on PMS, incontinence, and quality of life during pregnancy is an important research area. 20,21 Our meta-analysis showed that after PFMEs, there was a significant improvement in PFM strength and quality of life, while a single study for incontinence showed no significant relationship with PFMEs. In addition, we determined that the quality of the evidence was from low to average, respectively. Studies on the pelvic floor during pregnancy revealed a clear link between pelvic floor disorder and low PMS 20 . Physical and hormonal changes caused by pregnancy are factors that can reduce PMS by affecting the pelvic floor. Although all studies have reported improvement in PMS, further studies are needed due to the low number of studies in this field and the low quality of evidence 21 .