Open versus laparoscopic radical prostatectomy: Part II


Laurent Salomon, Service d’Urologie, Hôpital Henri Mondor, 51 avenue du Maréchal de Lattre de Tassigny, 94010 Créteil, France.


radical prostatectomy


Radical prostatectomy (RP) [1] can be performed by open (retropubic and perineal) and laparoscopic approaches; there are no controlled comparative studies among these three approaches and in Part II we report and review published studies.



The results of RP are evaluated on the basis of survival with no biological progression, and by the rate of positive surgical margins, reflecting the quality of tumour excision. CO2 insufflation during laparoscopy carries a risk of tumour cell seeding within the abdominal cavity and along the trocar ports [2,3]. Open surgery also carries a risk of metastatic implantation, and laparoscopy does not appear to increase this risk [4]. On the contrary, the most recent prospective randomized study showed that, for excising colon cancer, laparoscopy offers better local relapse-free survival and better tumour-related survival [5]. In the urological setting no firm conclusion can be drawn on the possible oncological risks specifically associated with laparoscopy, although local relapses on the trocar tracks have been reported after extensive nephrectomy and nephroureterectomy [6].

No cases of trocar track metastasis or local relapse have so far been reported after laparoscopic RP, but the follow-up is short relative to the natural history of prostate cancer. The risks of peritoneal or trocar-site seeding can be assessed in the light of experience with laparoscopic ilio-obturator node dissection. Only one case has been reported, in a patient with an aggressive prostate tumour and large node metastases; notably the nodes were withdrawn in a protective bag [7]. In the largest series of node dissection in node-positive patients there were no cases of peritoneal, parietal or trocar port seeding, even in hormone-resistant patients [8]. Although the follow-up is short, prostate cancer does not appear to implant readily on the peritoneum or trocar track. The laparoscopic extraperitoneal approach avoids this potential risk of intraperitoneal dissemination [9–12].

In 1998, the estimated mean rate of positive surgical margins during open surgery was 28%, according to Wieder and Soloway [13], with rates of 5–36% for pT2 and 8–53% for pT3 tumours. In series of patients undergoing laparoscopic RP the rate of positive surgical margins was 19–23% overall, 2.1–18% for pT2 and 30–69% for pT3 disease (Table 1) [9–11,14–22].

Table 1.  The rate of positive surgical margins according to the approach
RefNSurgical approachpT2apT2bpT3apT3bTotal
  1. R, retropubic; P, perineal; T, transperitoneal; E, extraperitoneal; NS, not specified; /, all pT2 or pT3.

[14]200P  /0 /2112
[15]250T418.439.3 3320.1
[21]1000T6.91830 3219.2
[22]438T2.115.433.8 53.426.2
[9]50E  /7.7 /44.422
[11]68E  /12 /69.223.4
[10]70E  /6.1 /29.421.4

Before 1997, RP was performed by the retropubic and perineal approaches. Boccon-Gibod et al.[23] showed that the perineal approach was associated with a higher risk of surgically induced margins and capsule incision. According to Weldon [24] this was caused by the surgical technique. However, all published studies comparing the two approaches show similar oncological control [25–33].

Using the PSA level and better patient selection, Carter et al.[34] found that the rate of positive surgical margins decreased from 17% in 1988–92 to 8% in 1994–96. The rate of positive margins is therefore difficult to compare because of differences in patient populations. This rate was compared retrospectively with three approaches, in the same centre, among patients with PSA levels of <10 ng/mL. This study was done by the same pathologist, thus eliminating an interpretation bias, and showed that the rate of positive surgical margins was identical, both for stage pT2 (≈ 15%) and for stage pT3 disease (≈ 40%) [35].

Nevertheless, the location of positive surgical margins differs according to the approach, i.e. the apex with the retropubic, the bladder neck with the perineal and posterolateral with the laparoscopic approaches [36,37]. Bollens et al.[9] and Türk et al.[38] also noted that positive surgical margins were preferentially posterolateral with the laparoscopic approach. This region contains the neurovascular bundles and prostatic pedicles. We have shown that respecting the neurovascular bundles does not increase the rate of positive surgical margins after laparoscopic RP, which was also the case with the retropubic approach [13,39].

As in the retropubic and perineal approaches, where respectively the apex and the bladder neck are most difficult to access, the posterolateral region is most difficult to access in the laparoscopic approach, probably because of the instrument axis and its smaller amplitude during dissection of the prostatic pedicles, which are close to the trocar ports.

The results of RP have been established by pathological examination of surgical specimens according to the international TNM classification [40]. In patients with intraprostatic tumours (stage pT2), Catalona et al.[17] reported that the progression-free survival rate was 93% at 3 years, while others obtained a 5-year survival rate of 97% among patients with the same category of tumours [20,41]. Similar results have been reported with the perineal approach [42].

The follow-up after laparoscopic RP is too short to establish progression-free survival curves. Guillonneau et al.[43] reported that 91.5% of patients with pT2 disease had PSA values of <0.1 ng/mL at 3 years; nevertheless, after 1000 laparoscopic RP they reported results similar to those of Catalona et al.[21]. For a given pathological stage, the 3-year survival rate with no biological progression was identical with the three approaches; thus laparoscopy produces results which are similar to the other forms of RP [21,35,44].


Published data on continence after RP are difficult to interpret. The definition of continence varies considerably among studies, from ‘no protection’ to ‘no more than one protection’. Geary et al.[45] reported that 80% of patients needed no protection, while Eastham et al.[46] reported a continence rate of 91%, continence being defined as ‘no more than one protection’. No consensus has been reached, and the results shown in Table 2 are based on a definition of continence as ‘no need for protection, either occasional or permanent’.

Table 2.  Continence after RP according to the approach
RefNSurgical approachFollow-up, monthsAssessment typeAge, yearsContinence, %
  1. R, retropubic; P, perineal; T, transperitoneal; E, extraperitoneal.

[38]44T 9physician59.992
[57]841T12? 89.2
[9]50E 6questionnaire6485

The method used to evaluate continence also varies, with assessment by the surgeon, by an independent physician, or with a self-administered questionnaire [47]. The follow-up and the interval to assessment are also important factors which differ among series. Although half of patients may be continent at 3 months and most are continent at 1 year continence may continue to develop in some cases 2 years after RP [48]. At least 1 year of follow-up therefore appears necessary to reliably assess continence outcomes [45,46,48–50]. Age seems to be an important factor [17,46,51–53]. The surgical technique can also improve continence, by the quality of apical dissection and sparing of the neurovascular bundles, the bladder neck and the puboprostatic ligaments [54,55]. In a retrospective study, Gray et al.[56] obtained similar continence rates after retropubic and perineal surgery. With open surgery, up to 95% of patients are continent, while the initial results of laparoscopic surgery were 85–97% after <1 year of follow-up (Table 2) [9–11,17–19,38,45,55,57–63].


The assessment of erectile function after RP is also subject to numerous pitfalls, including a lack of consensus definitions, use of different assessment methods and variable follow-up. The definition of sexual potency differs according to the criterion used, i.e. erection with or without penetration, and erection of sufficient quality for intercourse. Like continence, erectile function can be assessed by the surgeon, by another physician, or with a questionnaire. The use of impotence therapies such as sildenafil is not always specified.

The timing of the assessment is also important; many publications have shown that erections can resume relatively late after surgery, while most series have a short follow-up. It seems that at least 18 months of follow-up are required to reliably evaluate erectile function [55,64].

Other factors include the quality of erections before surgery, the patient's age and the status of the neurovascular bundles. The selection of patients in whom the neurovascular bundles can be spared is an important factor that is rarely considered [17]. It would be interesting to compare functional results according to whether the surgeon intends to spare the neurovascular bundles, rather than whether the bundles are actually spared.

When the neurovascular bundles are spared unilaterally, erectile function is preserved in 13.3–68% of patients after open and 53.8–58.8% after laparoscopic surgery (Table 3) [10,11,18,19,38,43,45,53,55,59,60,63,65–67]. When the bundles are preserved bilaterally, the results for function are 31.9–86% with open surgery and 40–59% with the laparoscopic approach. Vallencien et al.[57] reported a rate of 75% among patients aged <70 years who had normal erections before RP and in whom both bundles were spared, but the number of patients was not stated. We reported that five of six patients aged <60 years remained sexually potent when both bundles were spared [60]. Walsh et al.[55] reported a rate of 86% in a series of patients with a mean age of 57 years.

Table 3.  Erectile function after RP according to the approach
RefNSurgical approachFollow-up, monthsAssessment typeAge, yearsErection, %
  1. R, retropubic; P, perineal; T, transperitoneal; E, extraperitoneal.

[43]22T 6?6359


Few prospective studies have compared retropubic RP (the reference technique) with laparoscopic RP [68–71]. Such studies have the advantage of eliminating the patient-selection bias. The results were as follows.

  • • Laparoscopic RP takes longer than retropubic RP but offers a lower risk of bleeding.
  • • According to Bhayami et al.[68], analgesic requirements are lower with the laparoscopic than with the retropubic approach, and patients become active more rapidly (30 vs 47 days).
  • • Oncological results (excision margins) are similar with the two approaches [68–71].
  • • Functional results (continence and erectile function) are similar with the two approaches, although Roumeguere et al.[71] reported that continence was obtained more rapidly with the retropubic approach, while Anastasiadis et al.[69] reported the contrary. The quality of erections is better with the laparoscopic approach, with less use of sildenafil [71].
  • • At 6 months the quality of life is similar with the two approaches, although sexual functional disorders are similarly problematic.
  • • Finally, with hindsight, more patients operated on by the laparoscopic than by the retropubic approach say they would be willing to have the same operation again [70].


How many RPs must a surgeon perform before being fully operational? The answer is ‘as many as possible’[1]. Mokuliz and Thompson [72] reported that perineal RP could be learned quickly. Guillonneau et al.[73] classified laparoscopic RP as the most difficult operation, with lumbo-aortic node dissection, and recommended that only surgeons with sound laparoscopic experience should adopt this approach. Gallucci and Vincenzoni [74] considered that laparoscopic RP, and especially urethrovesical anastomosis, posed no major difficulties for surgeons accustomed to laparoscopy. At present, all series of laparoscopic RP have been published by teams that already had extensive laparoscopic experience [9,38,58,59,75–77].

Open RP takes ≈ 2 h (75 min in experienced hands) [78]; the laparoscopic approach takes far longer but is gradually becoming more rapid. In the initial experience of Schuessler et al.[79], with only nine patients, the mean operative duration was 9.4 h; that for our first 20 patients was 6.2 h [80], and gradually fell to 3.5 h. Guillonneau et al.[81] reported an operative duration of 4.5 h in their first series of 28 patients and 2.6 h more recently [57]. Türk et al.[38] also reduced their duration from 5.8 h for the first 10 patients to 3.3 h for the next 45. The extraperitoneal approach has seen similar changes; Bollens et al.[9] reported a reduction from 6.1 to 4.3 h in a series of 40 patients. Stolzenburg et al.[10] reported a duration of 2.8 h for their first 20 patients, but this group already had extensive experience of laparoscopic RP by the transperitoneal approach (>70 patients).

Rassweiler et al.[59] showed that the operative duration did not depend solely on the surgeon's experience, but also on the size of the prostate (5.5 h for prostates of 45 g and 4 h for those <25 g) and the pathological stage (pT2 vs pT3). They estimated that laparoscopic RP took ≈ 90 min longer than open RP. A similar result was obtained by El-Feel et al.[82], with a mean duration of 3 h for surgeons having done >50 RPs and 4.5 h for those having done <30. Learning RP can be accelerated through training programmes and mentoring [83].


One main factor limiting laparoscopy is the loss of two of the six degrees of freedom. To overcome these spatial restrictions, robots, including remote-activated dextrous manipulators and three-dimensional video imaging, were developed. Trans- and extraperitoneal laparoscopic RP have been performed using the Da Vinci System (Intuitive Surgical, CA, USA) [84,85]. The first results showed that mean operative duration was 2.5 h, average blood loss 150 mL and positive surgical margin rate 4%[86,87]. A structured approach minimized the complications during the establishment of the robotic laparoscopic programme [88] and remote-controlled robotic surgical systems are useful to translate open retropubic RP to laparoscopy and decrease the training needed [89,90].


At present none of the available approaches seems markedly superior to the others. The results of RP are above all ‘surgeon-dependent’, and patient selection has a considerable influence on outcome. The retropubic approach is the reference standard; it seems that the perineal approach is now only used by a few teams [91] and that the laparoscopic approach is gradually gaining ground.


While the advantages of laparoscopic surgery emerged rapidly for extensive nephrectomy and adrenalectomy (and for cholecystectomy), the advantages of laparoscopic RP have not yet been clearly established [92–94]. It is true that the results of laparoscopic RP are very encouraging after only 6 years of use, and that the laparoscopic approach has benefited from all the knowledge accumulated during 20 years of open surgery, especially by the retropubic approach. The excellent vision offered by the laparoscopic approach has improved knowledge of the pelvic anatomy and has led to further progress in retropubic RP. Reciprocally, the publicity surrounding laparoscopy obliges surgeons who prefer open RP to obtain similar results to those obtained with the laparoscopic approach [74].

It is now clear that laparoscopy offers results that are at least as good as those of open surgery, with less blood loss and simpler postsurgical outcomes in reference centres. Neurovascular bundle status seems better after laparoscopy than after open surgery. Some laparoscopic operations (treatment of unilateral varicocele and Burch colposuspension [95,96]) have totally disappeared, either because they offered no clear advantages or because they were rapidly replaced by other more effective techniques [97]. The use of the PSA assay has led to changes in the indications for RP, i.e. a low preoperative level and a normal DRE [98]. The question is no longer which approach to adopt but rather what treatment to choose, among surgery, radiotherapy, brachytherapy and, in future, microwave therapy and cryosurgery [74].