SEARCH

SEARCH BY CITATION

Keywords:

  • surgical procedures;
  • laparoscopy;
  • prostate neoplasm;
  • urinary incontinence

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES

Associate Editor

Ash Tewari

Editorial Board

Ralph Clayman, USA

Inderbir Gill, USA

Roger Kirby, UK

Mani Menon, USA

OBJECTIVES

To detail the technique and evaluate in a preliminary study the effectiveness of posterior reconstruction of Denonvilliers’ musculofascial plate (PRDMP) in enhancing early continence after robotic and laparoscopic radical prostatectomy (RP).

PATIENTS AND METHODS

Thirty-two consecutive patients having robotic or laparoscopic RP with PRDMP (group 1). Thirty previous patients not having PRDMP were compared as historical controls (group 2). Continence, as measured by patient self-reporting of the number of pads used/24 h, was assessed at 3 days and 6 weeks after catheter removal, by telephone interview. ‘Continent’ was defined as the use of none or one pads, ‘moderate incontinence’ as two pads, and ‘severe incontinence’ as more than two pads. Intraoperative transrectal ultrasonography (TRUS) was used to measure the membranous urethral length before and after PRDMP.

RESULTS

At 3 days after catheter removal, more patients in group 1 were continent than in group 2 (34% vs 3%, P = 0.007). At 6 weeks continence was again better in group 1 (56% vs 17%, P = 0.006). The mean length of the membranous urethra on TRUS measured before RP, after RP but before the musculofascial suture, and afterward, was 15.6, 12 and 14 mm, respectively. Thus, reconstruction restored the length of the transected membranous urethra by a mean of 2 mm.

CONCLUSIONS

PRDMP during robotic and laparoscopic RP leads to improved maintenance of membranous urethral length and significantly higher early continence rates.


Abbreviations
(PR)DMP

(posterior reconstruction of) Denonvilliers’ musculofascial plate

RP

radical prostatectomy

MUL

membranous urethral length

NVB

neurovascular bundles

BMI

body mass index.

INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES

Temporary urinary incontinence after radical prostatectomy (RP) for prostate cancer remains a disadvantage of surgical treatment. Long-term continence outcomes are favourable, with 85–97% patients reported as continent at 1 year after RP in laparoscopic [1–4] and open series [5]. However, the return of continence was delayed, with a median return to continence at 4.5 months and full return progressively increasing over a 2-year period to 93%, in 985 men who had RP [6]. Several technical modifications were proposed to promote an earlier return of continence, including sparing the bladder-neck [7], or puboprostatic ligament [8], or puboprostatic collar/perineoplasty [9], and the posterior reconstruction of Denonvilliers’ musculofascial plate (PRDMP) [10,11].

PRDMP is an attractive modification because of the limited operative time required and limited technical challenge of placing the reconstruction sutures. Recently, Rocco et al. reported the results of their technique during open [10] and laparoscopic [11] RP. In their open series, of 161 patients who had PRDMP, 72%, 79% and 86% were continent at 3, 30 and 90 days, respectively (with continence defined as none or one ‘pad or diaper’ per 24 h). A historical control group of 50 patients who did not have PRDMP were significantly less likely to be continent (14%, 30% and 46%, respectively). There was no difference in continence outcomes at 1 year (95% vs 90% for the study and control patients, respectively). In laparoscopic cases, they found significantly higher rates of continence in 31 patients undergoing PRDMP than in 31 controls, in a prospective alternating randomized cohort at 3 days (74% vs 25%, P < 0.001) and 30 days (84% vs 32%, P < 0.001). There was no significant difference at 90 days (92.3% vs 76.9%, = 0.25).

PRDMP might improve continence by increasing the functional urethral length; this is measured as the membranous urethral length (MUL) on preoperative MRI, and was shown to be associated with improved continence rates after open RP [12]. TRUS can also be used to assess MUL before and after RP [13], and the MUL after RP was similarly found to be predictive of the return of continence [14].

Thus we used PRDMP in 32 consecutive patients undergoing robotically assisted or pure laparoscopic RP, to assess the effectiveness of this modification in promoting early continence. TRUS was used during RP to assess changes in MUL.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES

For the robotic and laparoscopic RP, a 5- or 6-port approach was used. After releasing the anterior peritoneum, the prostate and bladder are exposed, and the bladder neck is transected. We do not spare the bladder neck as a rule, but vary our technique based on preoperative tumour characteristics. After releasing the neurovascular bundles (NVB), the puboprostatic ligaments are divided and the dorsal vein complex controlled with ligature or using a stapler. The anterior urethra is divided and the urethral lumen opened. The posterior urethra is then divided, the prostatic apex dissection completed, and the specimen entrapped.

PRDMP is done by placing two separate 2-0 polyglactin sutures on a UR-6 needle (Fig. 1a–f). The initial suture is placed just posterior and proximal to the urethral stump, between the cut edge of the distal Denonvilliers’ plate and the cephalad Denonvilliers’ musculofascial remnant posterior to the bladder neck. As such, the distal suture is placed in a 6 o’clock to 12 o’clock orientation through the tissue posterior to the transected urethra comprising the musculofascial plate. We found that perineal pressure aids in differentiating the posterior musculofascial plate from the adjacent rectal wall tissue. On tying this suture the now-reconstructed DMP can be clearly appreciated anterior to the rectum. For the second suture the distal suture is placed adjacent to the initial suture and the proximal suture is placed at the posterior bladder neck 1–2 cm from its luminal edge. On tying this suture the bladder neck descends close to the urethral stump. The vesico-urethral anastomosis is completed using attached poliglecaprone and polyglytone 2-0 sutures on UR-6 needles to create a double-armed technique in a running fashion, starting from the 5 o’clock location. The two sutures are tied together at the 12 o’clock position. Any redundancy in the bladder neck is closed anteriorly in a running fashion using one of the anastomotic sutures. Intraoperative real-time TRUS was used as previously described [13] to measure MUL before RP, after prostate excision, and after DMP reconstruction. On TRUS, the MU is recognized as an anechoic channel surrounded by a spindle-shaped hypoechoic area that represents the external striated (rhabdosphincter) muscle. The MUL was defined as the distance from the prostatic apex to the entry of the urethra into the penile bulb. After excision, the proximal border is the urethral cut edge. After reconstruction and anastomosis, the proximal border on TRUS is the anastomosis itself. One urologist with extensive experience of TRUS (K.K.) made all the measurements.

image

Figure 1. Intraoperative images of PRDMP: a, placing the initial suture in the posterior MP; b, placing through Denonvilliers’ remnant; c, tying down the initial suture; d, the second suture through the posterior MP; e, placing through the posterior bladder neck; f, completed posterior reconstruction before the vesico-urethral anastomosis.

Download figure to PowerPoint

Over a 3-month period starting in November 2006, 32 consecutive patients (group 1) had a pure laparoscopic (24) or robotically assisted (eight) RP with PRDMP. As a historical control group with suitable follow-up information, 30 consecutive preceding patients (group 2) were identified who had laparoscopic (26) or robotically assisted (four) RP with no PRDMP.

To assess continence, the Foley catheter was removed after cystography confirmed water-tight anastomotic healing at 7 days. The continence status was measured at 3 days and 6 weeks after catheter removal, using a patient self-reported questionnaire mailed by the patient, and providing the number of pads used in 24 h. Patients with missing data were contacted by telephone interview and queried for the number of pads used. If the patient gave a range, the higher value was recorded. Use of a diaper was coded as six pads. Use of none or one pad was considered ‘continent’ to account for use of ‘security pads’[5]; the use of two pads was considered ‘moderate incontinence’, and of more than two as ‘severe incontinence’.

The baseline demographics, tumour characteristics, and operative and postoperative outcomes were recorded in a prospective computerized database. Summary statistics were constructed, with means for continuous variables and percentages for categorical variables. Fisher’s exact test was used for analysing categorical variables and Student’s t-test or the Mann–Whitney U-test for continuous variables. Multivariate linear regression was used for the outcome variable of number of pads used at 1 and 6 weeks. Predictor variables evaluated in the model included age, body mass index (BMI), pathological Gleason score, pathological stage, preoperative PSA level, use of nerve-sparing (bilateral/unilateral/none), margin status, and the use of PRDMP. Log transformation was used for the PSA level to satisfy the requirements of normality during multivariate analysis; in all tests statistical significance was indicated at P < 0.05.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES

Both groups had similar baseline demographics and pathological data on univariate analysis (age, BMI, preoperative PSA level, Gleason score, pathological stage and positive margin rate). Operative measures of estimated blood loss and the use of nerve-sparing did not significantly differ between the groups. The mean (sd) operative duration was significantly lower in group 1, at 247.2 (62.7) min vs 281.0 (65.6) min (P = 0.048); Table 1 summarizes the data.

Table 1.  Baseline demographics, pathological and operative variables, the self-report of the number of pads used in 24 h, and continence status at 3 days and 6 weeks after catheter removal
VariableGroup 1Group 2P
  • *

    Student’s t-test;

  • †Mann-Whitney U-test;

  • Fisher’s exact test.

No. of patients 32 30
Mean (sd) or n (%):
Age, years 58.1 (6.6 57.2 (6.7)0.599*
BMI, kg/m2 27.4 (3.4) 28.3 (3.2)0.309*
PSA, ng/mL  4.9 (2.0)  8.3 (11.5)0.180
Nerve-sparing  0.197
 Bilateral 28 (88) 22 (73) 
 Unilateral  4 (13)  5 (17) 
 None  0  3 (10) 
Operative duration, min247.2 (62.7)281.0 (65.6)0.048*
Estimated blood loss, mL440.6 (307.3)499.4 (352.8)0.497
Pathological stage  0.739
 pT2a  6 (19)  5 (17) 
 pT2c 16 (50) 18 (60) 
 pT3a 10 (31)  7 (23) 
Gleason score  0.075
 6 13 (41)  6 (20) 
 7 18 (56) 24 (80) 
 8  1 (3)  0 
Positive margins  5 (16)  5 (17)1.000
3 days after catheter removal:
No. pads used  2.9 (1.9)  6.6 (4.6)<0.001
Continence status
 Continent (0–1 pads) 11 (34)  1 (3)0.007
 Moderate (2 pads)  5 (16)  6 (20) 
 Severe (>2 pads) 16 (50) 23 (77) 
6 weeks after catheter removal   
No. of pads used  1.9 (2)  4.6 (3.7)<0.001
Continence status
 Dry (0–1 pads) 18 (56)  5 (17)0.005
 Moderate (2 pads)  3 (9)  6 (20) 
 Severe (>2 pads) 11 (34) 19 (63) 

The MUL was measured during RP by TRUS in a subset of 12 patients in group 1; the mean MUL before RP was 15.6 mm, and it decreased to 12 mm (decrease 3.6 mm). After PRDMP, the mean length was 14 mm (increase 2 mm). Retrograde leak-point pressures during RP before and after PRDMP showed a slight decrease in leak-point pressure, from 50 to 45 mmHg in one patient, and from 35 to 33 mmHg in a second, respectively.

Patients in group 1 used significantly fewer pads at 3 days (2.9 vs 6.6, P < 0.001) and at 6 weeks (1.9 vs 4.6, P < 0.001) after catheter removal. Patients in group 1 were also significantly more likely to be classified as continent than those in group 2 at 3 days (34% vs 3.3%, P = 0.007) and at 6 weeks (56% vs 17%, P = 0.006) after catheter removal (Table 1).

On multivariate linear regression, a higher preoperative PSA level and the use of PRDMP were, respectively, positive and negative independent predictors of the number of pads used at 1 week; the log-transformed regression coefficient (95% CI) for PSA was 1.61 (0.05–3.17, P = 0.043), and for the use of PRDMP was −3.07 (−5.01 to −1.14; P = 0.002); and at 6 weeks the respective values were 1.32 (0.03–2.60; P = 0.045) and −2.15 (−3.76 to −0.55; P = 0.009). Age, BMI, pathological Gleason score, pathological stage, margin status, and use of nerve sparing were not independent predictors of the number of pads used at either time point.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES

Re-approximation of the distal and proximal Denonvilliers’ fascia remnants recreates the posterior musculofascial plate, which might function as a ‘fixation point’ for the horseshoe-shaped, posteriorly deficient, rhabdosphincter, and provide a dynamic support that might allow for more efficient contraction of the rhabdosphincter [15]. This theoretically improves the dynamic function and anatomical length of the rhabdosphincter, improving continence rates. The mean recovery in MUL was 2 mm, using intraoperative TRUS, representing an increase from 78% (12 mm) to 89% (14 mm) of the original MUL. Rocco et al.[10] also found an increase in MUL in their series, but felt that their TRUS measurements were ‘scarcely reproducible’. Using our technique and one urologist with extensive experience of TRUS, we think that our TRUS measurements of MUL were reproducible and reliable.

We found PRDMP to be technically straightforward; it also facilitates completion of the urethrovesical anastomosis by removing anastomotic tension. While the RP was statistically significantly faster with PRDMP, at 4.7 (1.1) vs 4.1 (1.0) h (P = 0.048), this probably reflects variations in the availability of surgeons at our institution rather than any effect of placing the posterior reconstruction sutures.

Patients undergoing PRDMP were significantly more likely to be continent at 3 days (34% vs 3.3%, P = 0.007) and 6 weeks (56% vs 17%, P = 0.006) after Foley catheter removal than were control patients. The continence rates for both groups were lower than those reported by Rocco et al. (at 3 days, 74% vs 26%, P < 0.001; at 4 weeks, 84% vs 32%, P < 0.001). Possible reasons include variations in reporting criteria (‘diapers’ and ‘pads’ were used interchangeably in their report, and none or one ‘diaper’ was considered as ‘continence’, which might represent differences in nomenclature), differences in vesico-urethral anastomosis technique (we used a running as opposed to interrupted anastomosis), and bladder neck technique (we do not routinely spare the bladder neck, while Rocco et al. report routinely sparing the bladder neck in all their patients).

Multivariate analysis identified the use of PRDMP as independently associated with decreased pad use at both time points. Conversely, the preoperative PSA level was independently associated in the multivariate analysis with more pad use at both times. This might reflect the more aggressive resections used for higher risk patients. However, other factors associated with higher risk disease or surgical aggressiveness (Gleason score, stage, nerve-sparing, margin status) were not independently associated on multivariate analysis. A multivariate analysis using the Gleason score and stage before RP rather than after yielded similar results (data not shown).

One concern when placing this suture is the proximity of the rectum. However, in their large group of 192 patients undergoing both open and laparoscopic application of this approach, Rocco et al.[10,11] reported no rectourethral fistulae. In the present series, while early, there were also no fistulae. Another concern is the proximity of the NVB; we found that placing the suture into the musculofascial plate vertically (6–12 o’clock position) appeared to allow the NVB to be avoided, and we do not feel that it compromises the integrity of the NVB. However, long-term potency outcomes will be needed to definitively exclude any compromise of potency associated with PRDMP.

The limitations of the present study include the retrospective comparison with a historical cohort, the short follow-up, and lack of randomization. As our experience with laparoscopic and robotic RP was >1000 cases before embarking on the present study, we do not feel that the difference in outcomes can be attributed to any lack of experience. Because the study was retrospective, unrecognized variables might account for our findings. As such, the results should be considered exploratory or observational, not definitive [16]. To address this concern, we are planning a prospective randomized trial with validated continence, potency and quality-of-life measures to more rigorously assess the effectiveness of PRDMP.

In conclusion, PRDMP before completing the urethrovesical anastomosis in laparoscopic and robotic RP results in improved maintenance of MUL and significantly higher early continence rates. A prospective randomized evaluation is warranted.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES