This systematic review included the use of pelvic floor muscle exercising in all studies. Eligible studies compared the effects of increased pelvic floor muscle exercising before operation and only postoperative pelvic floor muscle exercising on the recovery of postoperative urinary incontinence. No matter the publishing status, language or size.

Studies were excluded if they (a) were not RCTs, (b) did not indicated whether it is preoperative or postoperative, (c) were in a language for which a translation to English was not available or (d) were unpublished studies with only the abstracts presented at national and international meetings.

Included studies focused on men of all ages undergoing RP. To be eligible, in the intervention group had to involve a form of preoperative PFME with or without guidance and with or without biofeedback. Studies without controls (without preoperational PFME) and preoperational urinary incontinence were excluded. The main outcomes of interest were the incidence rate of urinary incontinence at 1th, 3th, 6th and 12th month after surgery.

We searched for studies on the MEDLINE, EMBASE, PubMed, Cochrane Central Register of Controlled Trials (CENTRAL), PsycINFO, China Biomedical Literature Database, China Journal Full-text Database, Wanfang Database and Weipu Database up to January 2023, without language restrictions. A controlled vocabulary was used (MeSH terms for MEDLINE and Cochrane; EMTREE for EMBASE). Keywords and their synonyms were used to sensitize the search, including “pelvic floor prostatectomy” or “Levator ani muscle” or “prostate cancer” and “exercise” or “RP” or “PFME” or “pelvic floor prostatectomy exercise” or “RP” and “PFME” or “pelvic floor prostatectomy training” or “PFMT.” The retrieval was conducted in the form of free words with a keyword.

For the identification of RCTs in PUBMED, the optimally sensitive strategy developed for the Cochrane Collaboration was used. For the identification of RCTs in EMBASE and other databases, a search strategy using similar terms was adopted. In the search strategy, there were four groups of keywords: study design, participants, interventions and outcome measures.

We analyzed the reference lists of all eligible articles to identify other potentially eligible studies. For ongoing studies or when any data were to be confirmed or additional information was needed, the authors were contacted by e-mail.

The previously described search strategy was used to obtain titles and abstracts of studies that were relevant for this review. Each identified abstract was independently evaluated by two authors. If at least one of the authors considered one reference eligible, the full text was obtained for complete assessment. Two reviewers independently assessed the full text of the selected articles to verify whether they met the criteria for inclusion or exclusion. In case of any disagreement, the authors discussed the reasons for their decisions, and a consensus was reached.

Two authors, independently blinded, extracted descriptive and outcome data from the included studies using a standardized form developed by the authors and adapted from the Cochrane Collaboration’s model for data extraction. We considered (a) aspects of the study population, such as the average age and sex; (b) aspects of the intervention performed (sample size, type of stabilization exercise performed, presence of supervision, frequency and duration of each session); (c) follow-up (if the patients included were analyzed); (d) loss to follow-up (if there was a loss in the sample); (e) outcome measures; and (f) presented results. Another author resolved disagreements. Any additional information required from the original author was requested by e-mail.

The quality of evidence was independently scored by two researchers based on the PEDro scale, which consisted of 11 items based on the Delphi list. One item on the PEDro scale (eligibility criteria) is related to external validity and is generally not used to calculate the method score, leaving a score range of 0–10.

The meta-analysis was performed with RevMan (Version 5.3. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014), and ORs and 95% CIs were used to summarize the outcomes. First, the heterogeneity of the research problem was assessed by the Cochran test. The significance level for the Cochran test was set at α = 0.1 as recommended by authors such as Sedgwick (13). We used Cochrane Q to qualitatively describe the heterogeneity across eligible studies, and then, the I ² statistic was used to quantitatively estimate the heterogeneity. A low degree of heterogeneity was indicated by p ≥ 0.1 and I ² ≤ 50%, and a high degree of heterogeneity was indicated by p < 0.1and I ² > 50%. If the heterogeneity was high, we performed a further analysis of the heterogeneity sources. If there was no significant clinical heterogeneity, the random effects model was used for the meta-analysis. In this study, we performed all statistical analyses based on the random-effect model because heterogeneity cannot be omitted in reality.

The GRADE framework, a grading system for the quality of evidence, was used to classify the quality of evidence for outcome indicators. All RCTs were included in this study. RCT was set as the highest level of evidence. There are five factors that can reduce the quality of evidence: research limitations, publication bias, research inaccuracy, research inconsistency and indirectness of research results.

The initial search led to the identification of 73 abstracts, from which 25 studies were considered potentially relevant and retrieved for detailed analysis. After a complete reading of 25 articles, 13 were excluded. Finally, 12 papers met the eligibility criteria. Figure 1 shows the PRISMA flow diagram of study selection for this review.

The characteristics of the included studies are documented in Table 1. All eligible studies were published between 2000 and 2021. Among these studies, 3 were conducted in the United States, 3 in the Netherlands, 2 in Italy, 2 in China, and 1 in Australia and Spain. The sample size of each individual study ranged from 32 to 84, and the mean age ranged from 59 to 72.5 years. All studies were published in academic journals in full text. Eight studies reported that preoperative pelvic floor muscle exercise had effect or early effect, and four reported that exercise had no difference. There was a variety of different pelvic floor muscle exercising regimens with or without biofeedback. There were also differences in the amount of detail with which the regimens were described in each study. The exercise regimens are summarized in Table 1.

Each of the studies was scored using the PEDro scale (14). Table 2 presents the results of the individual assessments by the PEDro scale. The overall quality of all studies was fair to good, but the therapists and participants were not blinded in the design of all articles. All articles followed eligibility criteria and source of participants; random allocation; baseline adequate follow-up; between-group comparisons; and point estimates and variability. 4 articles lacked concealed allocation, 7 articles lacked blind assessors, and 3 articles lacked intention-to-treat analysis.

In this meta-analysis, we found that there were significantly lower rates of postoperative incontinence at 3th month in the preoperative PFME group compared with the control group, with an OR of being incontinent of 0.61 (p = 0.04). There was no significant difference in postoperative incontinence rates at 1th month (OR: 0.47; p = 0.06), 6th month (OR: 0.57; p = 0.13) or 12th month (OR: 0.56; p = 0.12), the overall effect (OR: 0.58; p = 0.0003) (Figure 2). The funnel diagram is basically symmetrical, as shown in Figure 3. It is less likely to consider the existence of publication bias.

Urinary incontinence after radical resection of prostate cancer is mainly caused by the injury of the urethral sphincter during the operation. According to the latest development, the concept of sphincter relaxation caused by the inherent sphincter defect after the operation is considered to be the cause of the disorder of the male overall system, which often causes physical and psychological pain of patients, increases the psychological burden of patients, and significantly reduces the quality of life (8).

The European Urological Society recommends pelvic floor muscle exercise as the main treatment for postoperative urinary incontinence recovery. Pelvic floor muscle exercise is a simple and easy method to promote recovery. The purpose of pelvic floor muscle exercise is to continuously strengthen the contraction of the levator ani muscle to increase the tension of the urethral fascia, maintain a good tension of the distal urethral sphincter, keep the pressure of the urethra higher than the internal pressure of the bladder, and improve the ability to control urination. At the same time, this exercise can effectively improve the pelvic floor nerve function of patients, enhance the muscle contractility and tension, and increase the volume of urethral sphincter (15, 16). However, according to the research published so far, the role of preoperative PFME with or without biofeedback is still controversial. Burgio et al. 8 studies showed that increasing preoperative PFME was beneficial to the recovery of urinary incontinence after RP (17–24), while Bales et al. 4 studies showed that there was no significant difference between increasing preoperative PFME and recovering urinary incontinence after RP (25–28). We analyzed the reasons for this difference, which may be as follows: firstly, the technical level of the surgeon and the choice of surgical method have a certain impact on postoperative recovery. Secondly, the time of starting PFME before operation is different, the earliest 1 month before operation and the latest 1 day before operation; There are also those that only mention preoperative, but no indication of time. Thirdly, the number of preoperative exercises is different. Most studies only mention the start time, but not the number of times. Among them, the doctor suggested four times a day, but the study did not indicate whether the patient completed the exercise as recommended by the doctor. Fourthly, the feedback methods are inconsistent, including visual, oral, probe and surface electrode. Evaluators may have deviations due to different knowledge level structures. Fifthly, there is a deviation in the definition of urinary incontinence. The same result may be the opposite in different studies. Finally, pelvic floor muscle exercise needs patients to be able to cooperate and persist. Because of the large age deviation of prostate cancer patients, although physical therapists or nurses guide them to exercise, due to the complexity of pelvic floor muscle anatomy, it is difficult for patients to determine which muscles are contracting and whether the contraction is correct; At the same time, there may be insufficient PFME frequency and intensity, which will affect the final result. So it is difficult to judge whether exercise is effective (29, 30). Therefore, it is suggested that the follow-up related research should provide scientific basis for clinical practice through standardized feedback mode, professional guidance of medical staff and long-term follow-up on the impact of preoperative PFME on urinary incontinence after RP.

This study is limited by the number of studies available for analysis, Eastern and Western populations, age span, training methods, cycles, intensity, effectiveness, follow-up time, and limited number of patients. The resources used in this meta-analysis are all from the literature, so it is impossible to obtain more detailed information. Because the literature is limited by objective factors such as source, amount of information provided, different PFME protocols, definitions of urinary incontinence, with or without biofeedback and control of confounding factors, it may affect the conclusion of this study to some extent. Therefore, it is necessary to carry out prospective research focusing on the control of confounding factors while evaluating the problem more scientifically and comprehensively. In addition, in clinical trials, there are many uncontrollable factors, which make it difficult to achieve randomization, concealment and blinding of allocation schemes; Therefore, the analysis results in this paper are for reference only.