Intrafascial versus interfascial nerve sparing in radical prostatectomy for localized prostate cancer: a systematic review and meta-analysis

The present study aimed to systematically evaluate the effectiveness and safety of the intrafascial and interfascial nerve sparing (ITR-NS and ITE-NS) radical prostatectomy. PubMed, Embase, and Cochrane Library databases were searched for eligible studies. Meta-analysis with random-effects model was performed. Six comparative trials were selected and embraced in this research, including one randomized controlled trial, three prospective comparative trials, and two retrospective comparative trials. With regard to perioperative parameters, no significant association of operative time, blood loss, transfusion rates, duration of catheterization, and hospital stay existed between ITR-NS and ITE-NS. With respect to the functional results, ITR-NS had advantages in terms of both continence and potency recovery compared with ITE-NS. In reference to the oncologic results, the ITR-NS showed lower overall positive surgical margin (PSM) compared with ITE-NS but pT2 PSM and biochemical recurrence free rates were similar to the two surgical types. This study demonstrates that ITR-NS has better continence at 6 mo and 36 mo and better potency recovery at 6 mo and 12 mo postoperatively, regardless of the surgical technique. The cancer control of ITR-NS was also better than that of ITE-NS. This may be explained by the fact that patients in ITE-NS group present higher risk cancer than patients in ITR-NS group.

Study selection and data extraction. Two authors independently screened titles and abstracts of all search results. Studies were selected based on the pre-specified inclusion and exclusion criteria. Any discrepancy was resolved by discussion. Two authors independently extracted the following data from each identified study: study details (name of first author, country, year, contact details, conflicts of interest), methods (study design, duration, clinical setting), patients (sample size, baseline characteristics), intervention (surgical approach, comparison group), and outcome (postoperative urinary continence rate, potency recovery rate, PSM, pT2 PSM, BCR free rates, operative time, blood loss, transfusion rates, duration of catheterization, and hospital stay).
Risk of bias and quality of evidence assessment. We used the Cochrane collaboration's tool for risk of bias assessment of RCTs and the Newcastle-Ottawa Scale for risk of bias assessment of observational studies [15][16][17] . The Cochrane collaboration's tool assesses risk of bias in six domains: (1) selection bias; (2) performance bias; (3) detection bias; (4) attribution bias; (5) reporting bias; and (6) other bias 15 . The Newcastle-Ottawa Scale assesses risk of bias in three domains: (1) selection of the study population; (2) comparability of groups; and (3) ascertainment of outcome 16 . We evaluated that follow-up was adequate if the maximum follow-up was more than 2 yr (i.e. 24 mo). The quality of evidence was assessed according to the GRADE system using GRADEpro GDT software. Statistical analysis. Meta-analysis was performed to aggregate the results if studies were sufficiently similar.
Due to the clinical heterogeneity implicated in the included studies, random-effects model was applied to estimate summary risk ratios (RRs) and corresponding 95% confidence intervals (CIs). Sensitivity analysis was conducted through sequentially excluding retrospective studies. Subgroup analysis according to timing of outcome measurement was performed if sufficient data was available. Heterogeneity was tested using chi-square (p ≤ 0.1) test and I 2 metric. All statistical analysis was performed using RevMan 5.3 (Cochrane Collaboration, Oxford, UK). A two-sided p value less than 0.05 represented a statistically significant difference, except for heterogeneity test. Publication bias was detected using funnel plot if the included studies were more than five for each outcome.

Results
Literature search and study characteristics. Figure 1 shows the PRISMA flowchart of the systematic review and meta-analysis. Our search initially yielded a total of 216 records. After exclusion of duplicate articles, 131 records were screened through titles and abstracts. Finally, 6 studies involving 1663 patients (ITR-NS: 916 patients, ITE-NS: 747 patients) were included in this systematic review and meta-analysis 7-12 .

Criteria Description
Patients Adult men who underwent radical prostatectomy for prostate cancer   Intervention  ITR-NS, which was defined as the preservation of the periprostatic fascia and nerves by cutting adjacent prostate  and dissecting the plane between prostatic capsule and prostatic fascia   Comparison   ITE-NS was the control group, which was defined as the dissection of the plane between prostatic fascia and  endopelvic fascia; studies were not be selected or excluded based on surgical approaches (i.e The characteristics of included studies are presented in Table 2. These six comparative trials included one RCT 7 , three prospective comparative trials 8,9,12 , and two retrospective comparative trials 10,11 . There were two studies 7, 9 using laparoscopic RP, two studies 8,11 involving robot-assisted RP, and two studies 10,12 applying open retropubic RP. The definition of the ITR-NS and ITE-NS in the included studies is presented in Supplementary Table 2. Study sample sizes ranged from 41 8 and 420 12 . Studies were published from 2010 to 2015 in Europe 7, 10-12 and Asia 8,9 . All the studies used bilateral nerve-sparing radical prostatectomy technique. Tables 3 and 4 showed the risk of bias assessment of included studies. The random sequence generation, allocation concealment, and blinding of the RCT were all unclear (Table 3). Therefore, the risk of bias of the RCT was unclear. The majority of the longitudinal controlled studies were considered to have low to moderate risk of bias (Table 4). Three studies 7, 10, 11 had the proportion of high Gleason score (8-10 score). Patients in one study 8 who underwent RP were relatively younger than those in other studies included in this meta-analysis. Furthermore, the surgical techniques used in the included studies were different. These factors would introduce some selection bias and clinical heterogeneity.
BCR free rates. One study 7 reported the BCR free rates at 6 mo and four studies 7,[9][10][11] reported them at 12 mo (  Fig. 5a). Sensitivity analysis showed similar results to overall analysis (Fig. 5b). The quality of evidence was very low for BCR free rates at different timing (Supplementary Table 3).
Secondary outcomes (perioperative parameters). The perioperative parameters are presented in Table 9. Two studies 7, 9 reported the transfusion rate. The results of meta-analysis with random-effects model showed that patients suffering from ITR-NS had similar transfusion rates compared with those undergoing ITE-NS (RR = 0.50, 95% CI 0.05-5.47, p = 0.57; Fig. 6). The mean operation time ranged from 60 10 to 169.41 min 8 . The mean blood loss ranged from 87 9 to 200 ml 7 . The mean duration of catheterization ranged from 5 7 to 11.09 d 8 . The mean hospital stay was 8 d 9 .
Publication bias. Publication bias was detected only for continence recovery. The result of funnel plot provided certain evidence that publication bias existed (Fig. 7).

Discussion
In this systematic review and meta-analysis of six studies, we compared the effectiveness and safety of ITR-NS and ITE-NS on prostate cancer treatment. Irrespective of the surgical technique, we found that ITR-NS had better functional results (urinary continence and erectile function) and oncologic outcome (PSM, pT2 PSM, and BCR) compared with ITE-NS. These findings were supported by sensitivity analyses which took the prospective studies into consideration alone. The results suggested that there was a difference in continence between techniques at 6 months and 36 months but not at 12 months. This might be caused by the various procedures of different techniques or it was a spurious result. To our knowledge, this systematic review and meta-analysis is the first study to comprehensively evaluate this topic. The previous reviews or systematic reviews or meta-analysis evaluated the techniques of RP 5, 18-20 (such as RARP, laparoscopic, and retropubic open), the PSM and perioperative complication rates of primary surgical treatments 21 , the primary surgical treatments for prostate cancer [22][23][24][25][26] , transperitoneal and extraperitoneal robot-assisted RP 27 , and the efficacy and safety of conventional laparoscopic RP with a transperitoneal approach versus that of an extraperitoneal approach 28 . Therefore, none of these studies focused on the surgical technique of RP. In 2015, Reeves et al. 6 systematically reviewed the association of NVBs sparing in RP with postoperative urinary continence outcomes. They found that avoiding damage to the nerve activity surrounding the prostate promotes urinary control in the first 6 mo after nerve sparing RP. In addition, from theoretically, the ITR-NS had better function than ITE-NS in functional outcomes as we mentioned in introduction. The result of our systematic review and meta-analysis was supported by the Reeves's study 6 .
Tewari et al. 29 proposed a grading system based on four grades of dissection according to veins surrounding the prostate. Schatloff et al. 30 proposed a grading system based on five grades of dissection according to arterial periprostatic vasculature. The grade 1 of Tewari's approach and grade 5 of Schatloff 's approach was equal to ITR-NS. As we acknowledged that cancer control is the most important goal of RP. The different dissection planes concept aims for an incremental security margin of prostate, instead of true incremental nerves sparing 13 . In this review, we found that ITR-NS was not significantly presented with risk of PSM, pT2 PSM and BCR free rate compared with ITR-NS. This might be due to restricted patient selection of the included studies. Therefore, the choice of surgical technique or dissection plane should be made based on the specific situation of patients in clinical practice, such as clinical examination, biopsy results, and imaging results 6 .
Although we used a systematic method to perform the meta-analysis, certain limitations also should be taken into consideration. First, our systematic review only identified one RCT, and the absence of high quality RCT might weaken the reliability of the meta-analysis. Second, low to moderate between-study heterogeneity was detected, which might be attributed to different surgical techniques, study design, selection bias, and surgeon experience. Selection bias between the two techniques was a major bias in the present meta-analysis, which implied that higher risk patients tended to undergo interfascial technique and lower risk patients tended to an intrafascial technique. However, patients with Gleason score more than 8 were only in two trials and the PSA levels were all similar as presented in Table 2. In addition, the sensitivity analysis also showed similar results to the overall analysis. Therefore, the selection bias was not obvious in this systematic review. Two studies 7, 9 used laparoscopic RP; two studies 8, 11 used robot-assisted RP; and two studies 10,12 used open retropubic RP. These six   Table 8. Biochemical free rates in the studies comparing intrafacial and interfacial nerve-sparing radical prostatectomy. ITR-NS = intrafacial nerve sparing; ITE-NS = interfacial nerve sparing; PSA = prostate specific antigen.   Table 9. Perioperative parameters in the studies comparing intrafacial and interfacial nerve-sparing radical prostatectomy. ITR-NS = intrafacial nerve sparing; ITE-NS = interfacial nerve sparing. comparative trials included one RCT 7 , three prospective comparative trials 8,9,12 , and two retrospective comparative trials 10,11 . We performed sensitivity analysis through excluding retrospective studies. The summary result of PSM rate was changed while excluding the retrospective studies. Therefore, the robustness of the result is weak. Third, we only included studies published in English. In addition, grey literature was not included. Hence, language bias might occur in this study. Fourth, due to limited number of included studies, we did not fully detect the publication bias. Of course, the publication bias is inevitable because we included studies published in English and excluded grey literature. The publication bias might decrease the reliability and credibility of this meta-analysis and systematic review. Moreover, the sample size and statistical power were relatively insufficient to identify the true difference of the two surgical techniques. Ultimately, the meta-analysis is a secondary analysis and its quality is based on the included studies. Our meta-analysis included studies with a RCT with high risk of bias and five longitudinal studies with moderate risk of bias. Therefore, the quality of the evidence was consequentially degraded. With respect to further researches, multi-center clinical trials, if possible, RCTs should be performed to evaluate the effectiveness and safety of ITR-NS and ITE-NS. In addition, further studies also should elucidate the functional anatomy of urinary continence and erectile function.
In conclusion, this systematic review and meta-analysis demonstrates that ITR-NS has better continence at 6 mo and 36 mo and better potency recovery at 6 mo and 12 mo postoperatively, regardless of the surgical technique. This finding might be due to more nerves were saved and less damage of the periprostatic tissue in ITR-NS compared with ITE-NS. The cancer control of ITR-NS was also better than that of ITE-NS. This may be explained by the fact that patients in ITE-NS group present higher risk cancer than patients in ITR-NS group. Further studies are needed to verify the conclusion in future.