Posterior musculofascial reconstruction after radical prostatectomy: an updated systematic review and a meta‐analysis

To evaluate the influence of posterior musculofascial plate reconstruction (PR) on early return of continence after radical prostatectomy (RP); an updated systematic review of the literature. A systematic review of the literature was performed in June 2015, following the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) statement and searching Medline, Embase, Scopus and Web of Science databases. We searched the terms posterior reconstruction prostatectomy, double layer anastomosis prostatectomy across the ‘Title’ and ‘Abstract’ fields of the records, with the following limits: humans, gender (male), and language (English). The authors reviewed the records to identify studies comparing cohorts of patients who underwent RP with or without restoration of the posterior aspect of the rhabdosphincter. A meta‐analysis of the risk ratios estimated using data from the selected studies was performed. In all, 21 studies were identified, including three randomised controlled trials. The overall analysis of comparative studies showed that PR improved early continence recovery at 3–7, 30, and 90 days after catheter removal, while the continence rate at 180 days was statistically but not clinically affected. Statistically significantly lower anastomotic leakage rates were described after PR. There were no significant differences for positive surgical margins rates or for complications such as acute urinary retention and bladder neck stricture. The analysis confirms the benefits at 30 days after catheter removal already discussed in the review published in 2012, but also shows a significant advantage in terms of urinary continence recovery in the first 90 days. A multicentre prospective randomised controlled trial is currently being conducted in several institutions around the world to better assess the effectiveness of PR in facilitating an earlier recovery of postoperative urinary continence.


Introduction
Urinary incontinence and erectile dysfunction are the two major disadvantages of radical prostatectomy (RP). According to the European Association of Urology 2015 guidelines on prostate cancer [1], mean continence rates at 12 months range from 89 to 100% for patients treated with robotassisted RP (RARP) to 80-97% for patients treated with retropubic RP. The mean potency recovery rates at 12 months range from 55 to 81% for patients treated with RARP to 26-63% for patients treated with retropubic RP.
In an effort to attain better functional results after RP, in 2001, Rocco et al. [2] described a technique for reconstruction of the posterior aspect of the rhabdosphincter, based on studies of the rhabdosphincter itself [3]. In 2006, it was reported that posterior reconstruction (PR) shortens incontinence time after RP [4]. In 2007, the application of the restoration of the posterior aspect to transperitoneal laparoscopic RP (LRP) was described [5]. In 2011, Coelho et al. [6] described a modified PR of the rhabdosphincter applied to RARP.
Since the original description, the rhabdosphincter reconstruction technique has spread worldwide, with mixed results. In 2012, Rocco et al. [7] published a systematic review of the literature on posterior musculofascial reconstruction after RP, suggesting that PR can improve early return of continence within the first 30 days after RP (P = 0.004) and reduce the incidence of leakage on cystogram (P = 0.050).
The strength of the recommendations derived from this meta-analysis was limited by the high heterogeneity between studies: different continence definitions in each analysed study, several modifications to the original surgical technique, and different surgical approaches [open RP (ORP), LRP, RARP]. Moreover, the small sample size of many of the selected studies was another important limitation of the published review.
Since then, several other trials have been published, mainly studying the effect of PR in RARP and LRP. Consequently, based on the recommendation of the Cochrane Collaboration to update systematic reviews at least every 2 years [8], we elected to update our previous meta-analysis of the literature in the field of PR of the rhabdomyosphincter.

Literature Search and Study Selection
The present study is a systematic review of the literature conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [9]. A systematic and critical analysis of Medline, Embase, Scopus and Web of Science databases was carried out in June 2015, by three authors (A.A.C.G, F.A.M, G.C.) separately, applying the following key terms: posterior reconstruction prostatectomy, double layer anastomosis prostatectomy. The search was conducted across the 'Title' and 'Abstrac't fields of the records, with the following limits: humans, gender (male), and language (English). Only full-text articles were considered, while abstracts were not included. We decided not to include data from congress abstract proceedings, as they usually lack the completeness of data required.
Of articles found, the ones selected were those in which cohorts of patients who underwent RP (regardless if performed as ORP, LRP or RARP) were compared according to the execution or not of PR of the rhabdomyosphincter, and in which the continence outcomes at defined intervals (3)(4)(5)(6)(7)30, 90 and 180 days after the catheter removal) were reported.
In the present meta-analysis, we assessed the following key aspects of each study: the patients' population, the study design, the definition used to assess the continence recovery, the data collection, if a variant of the PR defined as a 'Rocco Stitch' was performed, the rate of continence recovery (at 3-7, 30, 90 and 180 days after the catheter removal), the rate of urinary anastomotic leakage diagnosed at the postoperative cystogram, the rate of positive surgical margins (PSM), and the rate of postoperative complications potentially related to the PR of the rhabdomyosphincter (i.e. bladder neck stenosis or urinary retention). In case of studies reporting data not adequate for meta-analysis more information was directly requested from the authors.
All the data were collected in a single database and the quality control of the collection was performed on a random sample representative of the 20% of the totality of the studies included in the meta-analysis.

Statistical Analysis
Risk ratio (RR) was used in reporting the six dichotomous outcomes analysed in the study: urinary continence, PSM, urinary leakage on cystogram, acute urinary retention (AUR), and stenosis of the new bladder neck.
Heterogeneity between studies was assessed using the I 2 statistics. Pooled estimates were calculated with fixed-effect model (Mantel-Haenszel method) when there was no significant evidence of heterogeneity; otherwise, the random effect model (DerSimonian-Laird method) was used. The significance of the overall treatment effect was tested by the z-test. Meta-analysis results were graphically represented using forest plots.
Data were evaluated on an intention-to-treat basis as far as possible. The presence of publication bias was evaluated using funnel plots and the arcsine test was used to test funnel plot asymmetry.
All analyses were undertaken using R 3.2.3 (R Core Team 2015; R Foundation for Statistical Computing, Vienna, Austria) with the meta package.

Systematic Literature Search Results
In all, 493 articles dealing with the PR after RP were found: 90 on Medline, 189 on Embase, 101 on Scopus, and 113 on the Web of Science database. Of these, 191 articles were identified after the removal of duplicates, with 161 articles excluded because they were not focused on the PR of the rhabdomyosphincter. Eight studies were excluded because the techniques of PR were different from the original, or for the absence of an adequate assessment of continence recovery. Finally, of the 22 studies defined eligible for the qualitative analysis, one was excluded because of a lack of standardised reporting of the continence recovery outcome (Fig. 1), leaving 21 studies [3-6,10-26].

Description of Included Studies and Quality Assessment
All the 21 studies included were divided according to the levels of evidence defined by Phillips et al. [27]: 1 Three studies were defined as randomised controlled trials (RCT) reaching the level of evidence 1b. 2 Five studies were defined as non-randomised controlled trials reaching the level of evidence 2b. 3 One study was defined as retrospective with a contemporary cohort of patients reaching the level of evidence 3b. 4 12 studies were defined as retrospective with an historical cohort of patients reaching the level of evidence 4. Table 1 [3-6,10, [12][13][14][15]17,18,[20][21][22][23][24][25][26] summarises the quality scores of the studies and the risk of bias assessment. Of the 21 papers considered in the meta-analysis, five studies did not show any significant advantage in terms of postoperative urinary continence secondary to the PR of the sphincteric complex. However, most publications have shown, although with different levels of evidence, a significant advantage associated with the PR of the sphincter for postoperative urinary continence at different time intervals.

Meta-Analysis of Urinary Continence
The combined analysis of all studies, regardless of the surgical approach, showed an overall statistically significant advantage in the rate of postoperative urinary continence in favour of the application of the technique of PR at 3-7 days after catheter removal (RR 1.90, 95% CI 1:25-2:90; P = 0.003; Fig. 2 The above analysis shows a statistically significant advantage in terms of urinary continence at 30 and 90 days when LRP is used, and at 3-7 and 30 days when RARP is used.  Fig. 10). Heterogeneity between studies was very low (I 2 = 0%, P = 0.939). Subgroups analysis by type of study did not show significant differences between the pooled estimates (P = 0.664).

Meta-Analysis of Urinary Leakage on Cystogram
The meta-analysis of the data reported in the literature showed a statistically significant advantage on the reduction of the risk of peri-anastomotic urinary leakage at postoperative cystogram in the group of patients treated with PR compared with the group of patients not undergoing PR (RR 0.43, 95% CI 0.25-0.75; P = 0.006; Fig. 11). Heterogeneity between studies was very low (I 2 = 0%, P = 0.624). Subgroups analysis by type of study did not show significant differences between the pooled estimates (P = 0.930).

Meta-Analysis of Complications
We tried to evaluate complications that could be related to the specific technique, such as urethral stenosis due to an increased number of sutures in the area, but we found no data in the literature concerning this point.
Of all the trials included in the meta-analysis, six papers also reported information about the rate of AUR episodes in the comparison groups [3,4,6,14,15,24]. From the meta-analysis of these studies there was no statistically significant evidence to suggest any certain increased risk of AUR secondary to the PR of the sphincter (RR 1.36, 95% CI 0.63-2.95; P = 0.937; Fig. 12).
Three studies reported information on the detection of stenosis of the new bladder neck [3,4,15]. The meta-analysis of the RRs of these studies did not show a significant difference between rates of stenosis of the new bladder neck in the two groups (RR 0.67, 95% CI 0.20-2.31; P = 0:526; Fig. 13).

Publication Bias
Funnel plots for urinary continence at the four time intervals are shown in Fig. 14. Although funnel plot asymmetry tests did not detect any significant asymmetry, visual inspection suggests that we cannot exclude the presence of a certain amount of bias.
Publication bias was also investigated for the rate of PSM. The funnel plot (not shown) evidenced a good symmetry, all study outcomes were within the 95% CIs, and the P value of the asymmetry test was P = 0.826.

Discussion
The PR of the rhabdosphincter was proposed nearly 15 years ago, first for ORP, then LRP, and finally RARP [2][3][4]6], and several studies to date have presented different outcomes and versions of the described technique. We know that >50% of robotic surgeons in Europe are now using this technique when performing RARP [28].
[29] reported a significant advantage in terms of earlier continence when PR was performed in the robotic setting. In their study, Ficarra et al. [29] included only robotic studies focused on PR, also including other technical nuances such as anterior suspension suture. Differently from the Ficarra et al.
[29] study, we published a meta-analysis in 2012 [7] investigating the role of PR only, but including ORP, LRP and RARP series. We found a significant advantage associated with PR for postoperative urinary continence at 3-7 days (P = 0.030) and 30 days (P = 0.004) after the removal of the catheter. No significant differences in the levels of urinary continence between the two comparison groups were apparent at 90 and 180 days (P = 0.18 at 90 days, P = 0.66 at 180 days).
The results obtained in our present meta-analysis are in agreement and also indicate that PR seems to be associated with a lower incidence of cystographic leakage, probably as PR allows a greater hold of the vesicourethral anastomosis through a better approximation of the structures involved in the anastomosis.
All the published studies showed no statistical significant increase in the risk of postoperative complications (such as stenosis of the new bladder neck or AUR) in patients treated with PR.
For cancer outcomes, the application of the technique of PR does not appear to be associated with an increased risk of PSM at final pathological analysis.
For postoperative urinary continence, the present metaanalysis indicates an advantage of PR regardless of the surgical approach (ORP, LRP and RARP): a significant benefit at 3-7 and 30 days from surgery and, unlike the findings of the previous meta-analysis, even at 90 days after catheter removal. In addition, a smaller but still statistically significant advantage in terms of urinary continence was associated with PR at 180 days after catheter removal.
Performing a meta-analysis stratified by surgical approaches, a significant benefit was found for urinary continence at 30, 90 and 180 days when the PR is applied in the course of a LRP; a significant advantage at 3-7, 30 and 90 days when applied in the course of RARP was also found, but at 180 days this advantage was not statistically significant.
There are some limitations of the present study. Firstly, we did not include data from congress abstract proceedings because generally this type of publication does not report a complete set of data, which is required for a meta-analysis. This choice might be considered a limitation of the study. In the present meta-analysis, there are clues (in some cases quite obvious) of an asymmetry of the funnel plots; therefore the presence of publication bias cannot be excluded. The heterogeneity is quite high and this may have implications for the ability to extend the results to other populations. As already discussed in the previous meta-analysis [7], one of the main limitations is that the different authors who published their experience with this technique found different or even contrasting results for post-RP urinary continence, the main outcome examined. To explain the rate of heterogeneity found among these studies, some factors, which had already been mentioned in the previous study, will be briefly listed here. First, it must be recognised that several different definitions of continence have been given in the literature for post-RP outcomes and several ways of assessing continence Review have been used. In addition, due to the fact that continence is self-reported by the patient, this outcome can be affected by a certain degree of subjectivity. Second, only a subset of the studies considered in the meta-analysis used the PR technique as originally described in 2006; several variations have been developed by different authors. Third, differences in terms of surgical approach (ORP, LRP and RARP) may have a significant role in justifying different results for post-RP urinary continence. In particular, Joshi et al. [15] argued that, using a robotic approach, the better preservation of the structures involved in the continence recovery process might reduce the apparent benefit of the PR; this hypothesis seems to be supported by the results of our present study, where a reduced advantage of PR in the robotic setting compared with laparoscopy was evidenced. Lastly, some methodological aspects of the studies can affect their ability to detect significant differences; study design and sample size being two of the most relevant.
To better assess the effectiveness of the PR in facilitating an earlier recovery of postoperative urinary continence, an international prospective multicentre randomised controlled study is ongoing (ClinicalTrials.gov Identifier: NCT01809522; Table 5).

Conclusions
The PR of the sphincter is used by many surgeons in several centres around the world to improve one of the main adverse effects of RP, i.e. urinary incontinence. The PR technique is easily reproducible, quickly executed, and it does not appear to be associated with an increased risk of perioperative complications, but it appears to be associated with a reduction of peri-anastomotic urinary leakages.
Our present analysis confirms the benefits already discussed in the review published in 2012 [7], evidencing a significant advantage for urinary continence recovery in the first 90 days after RARP and also subsequently (at 180 days) after LRP. An advantage in terms of lower leakages at cystography is   also reported. No significant increase in PSM or complications is reported while using the PR. However, further high-quality, unbiased studies are required to allow firm conclusions to be drawn. To better assess the effectiveness of the PR in facilitating an earlier recovery of postoperative urinary continence, a multicentre randomised controlled trial is ongoing.