Efficacy of pelvic floor muscle training on the sexual function of men after radical prostatectomy: an overview of systematic reviews

This study aimed to evaluate the evidence in the literature on the efficacy of pelvic floor muscle training on the sexual function (SF) of men after prostatectomy by an overview of systematic reviews (SR) of randomized clinical trials. The search for studies was conducted in five databases without any language restriction: EMBASE, PUBMED, Science Direct, PEDro, and Cochrane Library. The main results were extracted from the SR by two reviewers, and meta-analysis was performed from the primary studies for the outcomes SF scores and erectile dysfunction. 4 reviews were included in the qualitative synthesis, which involved 8 clinical trials and 891 participants. The reviews had moderate to good methodological quality, but a high overlap of clinical studies was founded. Meta-analysis showed that pelvic floor muscle training improves sexual function scores not in 3 months (p=0.51) but in 6 months (p=0.02) and it did not show efficacy on erectile dysfunction after three (p=0.58) and 12 months (p=0.32). Studies with moderate to good methodological quality demonstrate that pelvic floor muscle training only improve sexual function score in 6 months, but not had efficacy in erectile dysfunction in men after prostatectomy.


Introduction
Prostate cancer is the sixth cause of death in men, with a worldwide incidence of up to 1.3 million and being the fourth most common malignancy in the world (Bray et al., 2018). Despite the clinical complications and advances in surgical procedures, radical prostatectomy yet is the most common treatment for prostatic cancer (Barakat et al., 2021;Heidenreich et al., 2015).
Surgical treatment for radical prostatectomy is commonly associated with complications, including adverse effects on sexual function (Jarzemski et al., 2019). The most common symptoms are erectile dysfunction, difficulty with achieving orgasm, penis deformity or shortening, fertility changes, and decreased sexual desire. These changes precede psychological function problems in these patients. since these complications are related to the patient's masculinity (Jarzemski et al., 2019).
The literature includes different therapies to minimise sexual function disorders and improve the quality of life of these patients. Therapy with prostaglandin and phosphodiesterase-5 inhibitors as well as pelvic floor muscle training (PFMT) have shown good results in the management of these dysfunctions (Geraerts et al., 2016). PFMT has been increasingly touted as the first-line treatment for voiding dysfunction after prostatectomy, providing better periurethral support of the pelvic floor muscles in cases of climacturia and erectile dysfunction (Geraerts et al., 2016).
In recent years, clinical studies have been developed to evaluate the efficacy of PFMT in improving the sexual function of post-prostatectomy patients. Systematic reviews were develop previously to assess the level of evidence of this treatment (Perez et al., 2018) , (de Lira et al., 2019;Oh et al., 2020), but they have important methodological limitations that need to be considered, since they include studies that are not randomised clinical trials (RCT) or not assess directly DE as an outcome.
Thus, the present overview aims to identify and evaluate the quality of the available evidence on the use of PFMT in improving the sexual function of men after prostatectomy. Therefore, it is expected to provide evidence to clinicians and researchers about the real efficacy of this type of treatment alone. The research question of this overview is: How effective is PFMT (intervention) in improving the sexual function (outcome) of men post radical prostatectomy (population)?

Study design
This overview of systematic reviews was conducted following the Preferred Reporting Items for Overviews of Reviews and the Preferred Reporting Items for OoSRs (PRIO-harms and PRIO for abstracts) (Bougioukas et al., 2019;Bougioukas et al., 2018). This overview was registered in PROSPERO under protocol number CRD42020221381.

Inclusion and exclusion criteria
SR with or without RCT meta-analyses on the effect of PFMT on the sexual function of men after prostatectomy were included in this overview, with no restrictions on age, sample size, date, language, or place of treatment. PFMT was defined as pelvic floor muscle rehabilitation protocols through pelvic floor muscle strengthening exercises, which were either associated with biofeedback and electrostimulation or not (P. W. .
Reviews that evaluated other forms of interventions or other interventions associated with PFMT, as well as those that did not assess the quality of evidence of primary studies, were excluded.
The primary outcomes of this overview were sexual function evaluated by self-reports or the International Index of Erectile Function (IIEF) (Rosen et al., 1999), total scores and domains.

Study search and selection strategy
Identified and eligible reviews were selected using a search strategy in the following databases: MEDLINE via The searches were conducted on January 15, 2022 and involved a combination of different terms chosen from the guiding question using the Boolean operators AND OR. The strategy search used was: (male OR man OR men OR prostatectomy OR prostate adenectomy OR prostate resection OR radical prostatectomy OR total prostatectomy) AND (PFMT OR Kegel exercise OR pelvic floor exercise) AND (sexual dysfunction OR impotence OR erectile dysfunction OR male sexual dysfunction) AND (systematic review OR review, systematic OR systematic review). No language or year of publication restriction were applied. Then, we included any review published until January 2022.
The records found were exported to Mendeley software and duplicate entries were removed. Two independent reviewers screened the studies by reading the title and abstract, and then the full text. In case of disagreements, a third reviewer was consulted for consensus as to the inclusion of the study in the overview. The Kappa coefficient assessed the agreement between the reviewers in the process and was interpreted as moderate (k = 0.41 -0.60), substantial (k = 0.61 -0.80), or near perfect agreement (k ≥ 0.81) according to Gilchrist (2008) (Gilchrist, 2009).

Data extraction
The data were extracted using a specific form containing data on the year of publication, authors, objectives, and eligibility criteria, RCT number and characteristics (population, intervention and control protocols, time of assessment, duration of follow-up, outcomes, results presented), assessment of methodological quality, and risk of bias. Whenever necessary, the researchers searched the primary study for missing or divergent information. Data extraction and methodological quality assessment were independently performed by two reviewers.

Methodological quality assessment
The methodological quality of the reviews analysed was assessed using the Assessment of Multiple Systematic Reviews (AMSTAR) tool, which aims to identify and describe the quality of SR, evaluating biases, possible conflicts of interest, and study results (Shea et al., 2017). Each tool item assesses whether the review meets the criteria, and yes, no, or not applicable can be selected.
The quality of the evidence was assessed using the Grades of Recommendation, Assessment, Development and Evaluation (GRADE) (Schünemann et al., 2013) system. Research, Society and Development, v. 11, n. 11, e221111133532, 2022 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v11i11.33532 4 Corrected covered area (CCA) was used as a measure of RCT overlap (Pieper, Antoine, Mathes, Neugebauer, & Eikermann, 2014). It an appropriate way to assess the overlap of primary studies in overviews, which, through a repetition and publication index, results in a possible range between 0 and 100%. A CCA value lower than five can be considered as a slight overlap, whereas values greater than or equal to 15 can be considered as a remarkably high overlap (Pieper et al., 2014). The calculation to verify the degree of overlap through the area covered and corrected area covered considered the number (n) of studies, reviews, and duplicates found.
Meta-analyses were recreated from the raw data of the primary studies to estimate the magnitude of the treatment effect more accurately. The meta-analysis was conducted using RevMAn 5.3 (Cochrane Collaboration, London, UK). Odds ratios and 95% confidence intervals were calculated for dichotomous outcomes and mean differences, while only 95% confidence intervals were calculated for continuous outcomes. The fixed or random effect model was used for calculation depending on the degree of heterogeneity evidenced in each analysis. The effect was considered significant when p-value was <0.05.

Results
The search for studies identified 39 articles, and only 4 SR were included. A total of 8 RCT and 891 participants were included. A flowchart describing the screening and included studies is shown in Figure 1. In the screening process, the Kappa coefficient showed almost perfect agreement between reviewers (k = 0.90).
Three SR included in this overview showed good methodological quality and two study presenting moderate methodological quality, as shown in Table 1.
The overlap of studies is presented through a citation matrix in Table 2. It shows that 62% of the articles were included in at least two reviews. The reviews included in this overview showed a high overlap of the included RCTs, demonstrated by the high values of CA (53%) and CCA (37%).  Item 11, 12, and 15.
Note: The numbering of items used in the table follows the same numbering as in the AmSTAR tool (Shea et al., 2017). Source: Own elaboration.
Different PFMT parameters were analysed in this overview: the number of sessions used, duration and instruction provided by the therapist, description of the training and the training progression regimen used, the professional who administered the training, the resources used, in this case if biofeedback was used, striated muscle coordination, and the clear description of the training for the patient to perform at home, considering the principles of a PFMT(P. .
The findings regarding the parameters found in each review are shown in Table 3 and 4.  Evidence on the use of PFMT on shortand long-term SF outcomes is conflicting.
High-quality studies have conflicting results on the use of PFMT in ED.
PFMT did not show statistically significant differences between groups.
There was no scale improvem ent with PFMT and BF compared to CG (no treatment or usual care) at 3 -6 months.
No statistically significant differences between the two groups on any outcome measure. Meta-analyses were performed comparing primary studies regarding the efficacy of PFMT for erectile dysfunction (ED) and sexual dysfunction (SF), as shown in Figure 2. Total International Index of Erectile Function (IIFE) score increased six months after PFMT (p = 0.02). PFMT showed no efficacy on ED after three (p = 0.58) and 12 months (p = 0.32), as well it did not improve the total IIFE score after three months (p = 0.51). According to the GRADE, PFMT has a low degree of recommendation for ED and SD. Despite we included 9 RCT on this overview, only RCTS that assess the ED with the same tool and time were pooled on the forest plot. Source: Own elaboration.

Discussion
ED and SF are constant complaints in patients, undergoing prostate removal, and among the treatments, PFMT seems promising for use in clinical practice in improving sexual function after surgery (Geraerts et al., 2016;Kannan et al., 2019).
However, most studies show both dysfunctions as secondary outcomes, thus creating a gap regarding the use of PFMT. In contrast to these promising results, this overview of current evidence shows that there is only a low level of recommendation for the use of PFMT to treat SD in men post-prostatectomy, since PFMT is associated with higher IIFE scores six months after the intervention but has no effect on the percentage of men with erectile function (EF). Meta-analysis performed shows that the interpretation of effect size differs from the clinical trial conclusions, which in general favour PFMT alone or a combination of biofeedback or electrostimulation in SD and ED after prostatectomy. Such findings respond more clearly to the gap left in the reviews by Wong et al. 5 ;Feng et al. 7 and Kannan et al. 16 , which report conflicting evidence on the effect of PFMT on SD.
The overlap rate evaluates the repetition of the primary studies and their findings present in the SR (Pieper et al., 2014). The high rate verified in this overview points to the need for new clinical trials and shows that new SR should only be conducted when new evidence emerges or becomes outdated.
However, with regards to this context, there is an important gap concerning the use of PFMT for these patients, Reconsiderations on the pelvic rehabilitation of post-prostatectomy patients demonstrate that the target muscle, treatment regimen, command, and guidance provided to the patient can influence PFMT, since the contraction mechanisms must aim to activate adequate contraction of muscles in the male patient; for example, puborectalis and bulbocavernosus muscle contraction to increase striated sphincter strength to improve the efficacy of training for urinary incontinence, since the studies included or developed up to now often use old female rehabilitation reasoning, which involve different muscular mechanisms from continence maintenance. Considering this reasoning, PFMT for sexual function may be promising in these patients, since the muscles related to continence are also involved in ED and SD maintenance.
Perhaps the conflicting results from systematic reviews analysed can be explained by the variation in the duration and characteristics of the intervention protocols used. Considering that there is evidence that erectile function in men may require a time of 18 to 24 months to be re-established or to show improvement (Andrew R McCullough, 2005), it may be questioned whether the intervention time in these patients does not need to be longer, as well as the follow-up time.
Finally, as a limitation of this overview, we highlight the impossibility of conducting more robust meta-analyses, due to the variability of evaluation methods and presentation of outcomes and results in publications. As well, the short follow-up time performed by the studies to document improvements in ED outcomes.
In this sense, it is recommended that new studies use internationally validated and accepted measures to measure outcomes and have a longer follow-up period for these patients. In addition, it is suggested the use of good practices for the development of clinical trials and description of interventions, allowing the production of a better level of evidence on the subject and facilitating the process of transposing evidence into clinical practice.

Conclusion
Although the SR included in this overview showed moderate to good methodological quality, they presented high overlap rates and contradictory conclusions. When recomposing the meta-analyses, the PFMT only show efficacy to increase the IIFE scores in 6 months (p=0.02). on the other hand, PFMT did not demonstrate efficacy in the number of patients with erectile dysfunction. This fact may suggest that the erectile function recovery process requires longer treatment and follow-up.
It should be noted that most of the intervention protocols analysed were not specifically designed for SD and have flaws regarding the measurement of outcomes and training administration.
According to the GRADE, PFMT has a low degree of recommendation for ED and SD. New studies with high methodology quality and using the new insights on pelvic rehabilitation in men are recommended to produce new evidence on the clinical applications of PFMT in these patients.