The ultimate aim of all radical prostatectomies of curative intent is to prevent clinical progression and death from prostate cancer. Treatment failure was historically defined as clinically evident local recurrence (positive digital rectal examination) or development of distant metastasis. However because of the protracted natural history of prostate cancer, clinical recurrence might take many years to manifest, thus accounting for the paucity of RARP studies using cancer-specific or overall survival as end-points. Pound et al. reported that no patient followed for more than 5 years developed any recurrence without a concomitant rise in PSA. Treatment failure is now defined in terms of a BCR.[20, 21] Although BCR often precedes clinical recurrence by 6–48 months, not all BCR eventually leads to progression with clinical relevance. PSM is most commonly used as a short-term surrogate for biochemical failure in RARP studies, as its presence increases the risk of biochemical failure by up to fourfold, and increases the risk of death by 1.7-fold. However, it should be emphasized that most patients with PSM do not develop BCR. Unpublished prospective data of our senior author's personal series of more than 1000 RALP showed that although the PSM rates were 17.4%, BCR rates were just 2.0%.
Non-comparative large cohort studies reported PSM rates of 11–37% after ORP, 11–30% after LRP and 9.6–26% after RARP. Comparative studies reporting PSM rates of ORP versus RARP are shown in Table 1. Ficarra et al., in their meta-analysis of comparative studies before 2008, reported a statistically significant difference in PSM rates in favor of RARP over ORP (RR 1.58). A statistically significant advantage of RARP was also identified in a separate analysis of only prospective studies. Subset analysis of patients with pathologically localized (pT2) cancer showed that ORP was associated with more than twice the risk of PSM than RARP. Tewari et al.'s recently published meta-analysis found that ORP has a statistically significant 24.2% risk of a PSM compared with a 16.2% risk in RARP. Subset analysis of only pT2 cancers also showed similar statistically significant findings in favor of RARP (16.6% vs 10.7%). However, after propensity adjustment, the differences were no longer statistically significant due largely to the higher preoperative PSA and pathological stage averages of the ORP cohorts compared with the RARP cohorts.
Table 1. PSM rates in the comparative studies evaluating ORP and LRP versus RARP
|Studies||Year of publication||No. cases (n)||Overall PSM rates (%)||pT2 PSM rates (%)|
|ORP vs RARP|| || || || |
|Menon et al.||2002||30 ORP||29||–|
|Tewari et al.||2003||100 ORP||23||–|
|Ahlering et al.||2004||60 ORP||20||9|
|Smith et al.||2007||200 ORP||35.7||24|
|Fracalanza et al.||2008||26 ORP||23||18|
|Schroeck et al.||2008||435 ORP||28.0||–|
|Krambeck et al.||2009||564 ORP||17||–|
|White et al.||2009||50 ORP||36||42.8|
|Rocco et al.||2009||240 ORP||22||17|
|Laurila et al.||2009||84 ORP||14||15|
|Drouin et al.||2009||71 ORP||18.1||–|
|Ficarra et al.||2009||105 ORP||21||12.2|
|Barocas et al.||2010||491 ORP||30.1||–|
|Williams et al.||2010||346 ORP||7.6||–|
|Lo et al.||2010||20 ORP||20||–|
|Breyer et al.||2010||695 ORP||16||–|
|Magheli et al.||2011||75 ORP||14.4||6.6|
|Di Pierro et al.||2011||75 ORP||32||–|
|Masterson et al.||2012||357 ORP||18||11.7|
|LRP vs RARP|| || || || |
|Joseph et al.||2005||50 LRP||14||–|
|Rozet et al.||2007||133 LRP||15.8||15.5|
|Trabulsi et al.||2008||190 LRP||18||12|
|Drouin et al.||2009||85 LRP||18.8||–|
|Hakimi et al.||2009||75 ORP||14||13|
|Magheli et al.||2011||68 LRP||13||6.7|
|Asimakopoulos et al.||2011||64 LRP||10||8|
|Park et al.||2011||62 LRP||21||7.4%|
Looking at more recent studies published after 2008 (Table 1), there seems to be a trend towards lower PSM rates in RARP compared with ORP.[32, 33, 36, 38, 43] Indeed, the largest study, with approximately 2000 patients, showed a statistically significant advantage of RARP over ORP in reducing overall PSM rates. Two studies in the present review seem to buck the trend, reporting statistically significant higher overall PSM rates in the RARP cohort.[39, 42] William et al.'s study of 950 patients found higher PSM rates for RARP compared with ORP. However, the study was a head-to-head single case series comparison of only one ORP and one RARP surgeon, and not a comparison of multiple surgeons as in other studies. In addition, the overall PSM rates of ORP (7.6%) reported in that study were extremely low, and were not reproducible in other studies, reflecting the surgeon's mastery in ORP. In the second study, Magheli et al. conceded that the higher overall PSM rates of RARP patients in their series was likely a result of the relative inexperience of their institution's robotic surgeons, who were at the beginning of their learning curves. A recent cumulative analysis of comparative studies showed similar overall and pT2 PSM rates for RARP and ORP.
Ficarra et al. reported no statistically significant difference in the overall and pT2 cancer PSM rates of LRP and RARP. Tewari et al., who analyzed more than 50 000 LRP and more than 60 000 RARP, concluded that RARP was associated with lower overall and pT2 PSM rates compared with LRP (16.2% and 20.4%, respectively), attaining statistical significance even after propensity adjustment and Hochberg correction.
Reviewing all meta-analyses[51-53] and available literature, RARP showed similar overall and pT2 PSM rates when compared with ORP. There is also no consistent evidence that RARP is superior to LRP in terms of oncological control. More importantly, a PSM does not necessarily translate to future BCR. This is evident in a few studies that included short- and mid-term BCR rates.[36, 38, 42, 54, 55] In particular, a large study by Barocas et al., despite a statistically significant higher overall rate PSM in the ORP cohort, showed that the 3-year BCR rates for ORP (83.5%) and RARP (84.0%; P = 0.19) were similar. Drouin et al. and Magheli et al. also showed equivalent BCR at 3 to 5-year intervals between the three surgical modalities. A population-based observational cohort study consisting of almost 9000 patients who underwent ORP or minimally invasive prostatectomy showed similar adjuvant or salvage cancer therapy rates (6.9 vs 8.2 per 100 person-years, respectively; P = 0.35), indirectly suggesting similar oncological control. This finding was recently validated in a cumulative analysis reporting equivalent BCR-free survival rates between ORP, LRP and RARP.
Urinary incontinence can be devastating to quality of life after RRP, and can occur regardless of the surgical approach. Hence, there had always been much emphasis on developing different strategies to preserve continence and to recover continence earlier after surgery. Traditionally, ORP involved a posterior “tennis-racket” closure of the bladder neck with eversion of the mucosa to optimize a mucosa-to-mucosa approximation between the bladder neck and urethra. Walsh et al. made significant headway in preserving continence by placing an anterior and posterior figure-of-eight suture proximal to the bladder neck, creating an intussusception of the bladder neck, thus preventing the opening of the bladder neck in a distended bladder, thereby reducing the rate of incontinence. Rocco et al. postulated that division of the muculofascial plate, formed by the striated sphincter, Denonvillier's fascia and the dorsal aspect of the prostate, resulted in a loss of the posterior cranial insertion of the sphincter causing the caudal displacement of the sphincteric complex and prolapse of the perineum. The reconstruction of this muculofascial plate before the vesiourethral anastomosis resulted in an earlier continence recovery.[58, 59] Nguyen et al. confirmed earlier continence recovery with robotic and laparoscopic prostatectomy. In their study, 34% of patients with posterior muculofascial plate reconstruction were continent at day 3 after prostatectomy, compared with 3% in patients without the reconstruction. The continence rate improved to 56% versus 17% at 6 weeks post-prostatectomy. The application of a periurethral suspension stitch and an anterior reconstruction appears to synergistic to the Rocco et al. stitch, with better results reported. Several factors were identified to be of paramount importance in maintaining post-prostatectomy continence. These included minimal disruption of the puboprostatic ligaments and bladder neck, an intact puboperinealis muscle, a long urethral length, a watertight anastomosis and a well-supported vesico-urethral junction.
Comparative studies published before 2008 (Table 2) were of limited sample size and short follow-up duration.[27, 28, 45] The largest study by Tewari et al. reported a statistically significant earlier return of continence in the RARP cohort when compared with the ORP cohort. The other study showed no difference in UCR at 3 months post-prostatectomy. However, it is known that the return of continence occurs gradually and improvement can take up to 1–2 years.[69, 70] The majority of more recently published studies (Table 2) provided us with longer follow-up duration of at least 12 months, but they are still fraught with varying definitions of continence and methods of ascertaining symptoms, making direct comparison between studies difficult. It has been established that patient self-reported incontinence appears to be greater than that reported by physicians. In a study carried out by our faculty, it was found that the physician-reported (by direct interview) absolute continence rate (0 pads used) post-RALP was 51.5%, although the patient self-reported absolute continence rates (by validated questionnaires) were actually 14.7% (urinary leakage although 0 pad used). Despite these limitations, these recent studies showed that UCR in RARP is at least equivalent,[32, 66, 67] if not better,[34, 37, 40, 43, 64] when compared with ORP.
Table 2. UCR in the comparative studies evaluating ORP and LRP versus RARP
|Studies||Year of publication||Definition of continence||No. cases (n)||Results|
|ORP vs RARP|| || || || || |
|Tewari et al.||2003||No pad used||100 ORP||Median time to continent||160 days|
|200 RARP||44 days|
|Ahlering et al.||2004||No pads used||60 ORP||UCR at 3 months||75%|
|Krambeck et al.||2009||No leakage||564 ORP||UCR at 12 months||93.7%|
|Hu et al.||2009||Not specified||6 899 ORP||Incontinence per 100 person-years||12.2 per 100 person-years|
|1 938 MIRP||15.9 per 100 person-years|
|Rocco et al.||2009||No pad or 1 safety pad used||240 ORP||UCR at 3, 6 and 12 months||63%, 83%, 88%|
|120 RARP||70%, 93%, 97%|
|Ficarra et al.||2009||Leak less than once a week||105 ORP||UCR at catheter removal and 12 months||41%, 88%|
|103 RARP||68.9%, 97%|
|Lo et al.||2010||No pad or 1 safety pad used||20 ORP||Mean time from surgery 42 months||85%|
|20 RARP||Mean time from surgery 6 months||95%|
|Minniti et al.||2011||Not specified||93 ORP||Not specified||65.6%|
|Di Pierro et al.||2011||No leakage||75 ORP||UCR at 3 and 12 months||83%, 80%|
|75 RARP||95%, 89%|
|Kim et al.||2011||No pad used||235 ORP||Median time to continent||4.3 months|
|528 RARP||3.7 months|
|Barry et al.||2012||Symptoms causing moderate or severe problems||220 ORP||Median time from surgery 434 days||27.1%|
|406 RARP||Median time from surgery 418 days||33.3%|
|Kowalczyk et al.||2012||Need for incontinence surgery||58 638 ORP||More than 18 months||0.3%|
|19 594 MIRP||0.3%|
|LRP vs RARP|| || || || |
|Joseph et al.||2005||No pad used||50 LRP||UCR at 6 months||92%|
|Hakimi et al.||2009||No pad used||75 LRP||UCR at 12 months||89%|
|Trabulsi et al.||2010||No pad or 1 safety pad used||45 LRP||UCR at 6 and 12 months||71%, 82%|
|205 RARP||91%, 94%|
|Park et al.||2011||No pad or 1 safety pad used||62 LRP||UCR at 1, 6, 12 months||25%, 76%, 95%|
|44 RARP||67%, 94%, 94%|
|Asimakopoulos et al.||2011||No pad used||64 LRP||UCR at 12 months||83%|
Overcoming the initial learning curves and development of continence-preserving surgical techniques in RARP has translated into a reduction of urinary incontinence rates. This is clearly evident by comparing studies with RARP carried out before[28, 32, 56] and after[34, 37, 40, 43, 64, 65] the year 2006. In fact, an earlier recovery of urinary continence was shown in all the recent studies that included reports of their third month UCR.[34, 37, 43, 65] Kim et al. initially found similar continence rates between ORP and RARP. However, on subgroup analysis after excluding the first 132, the median time to continence recovery dropped to 1.6 months (P < 0.001). They also identified younger age and longer preoperative membranous urethral length (≥1.1 cm) seen on prostate magnetic resonance imaging as independent prognostic factors for continence recovery on multivariate analysis. On the flip side, Hu et al. identified 8837 patients (the majority of this cohort had surgery in 2003–2005) using the SEER Medicare data for analysis. They reported that the MIRP cohort had an increased risk of genitourinary complications and urinary incontinence as opposed to ORP. Kowalczyk et al. recently published an update consisting of a much larger sample size (n = 78 232) using 100% sample of Medicare beneficiaries beyond the SEER database regions from 2003 to 2007. They reported a reversal of previously published findings. MIRP was now found to be associated with a statistically significant lower risk of genitourinary complications. Although the actual rate of incontinence was not studied, they found that just 0.3% of patients in both ORP and MIRP groups required further surgeries for incontinence. The authors attributed these findings to the larger sample size, the improvement of surgical techniques and the progression along the MIRP learning curves.
In the present review, we found five comparative studies comparing RARP and LRP,[45, 48, 50, 65, 68] and four showed no difference in the UCR at 6 or 12 months. Park et al. showed an earlier return of continence in their RARP patients at 3 months, but similar continence rates were attained at 6 months and beyond. Trabulsi et al. reported superior continence rates of RARP over LRP. However, they conceded that they were still at the beginning of the learning curve of LRP before they made the transition to RARP, hence suggesting that a steeper learning curve exists for LRP. Indeed, on reviewing larger non-comparative cohort series, the 12-month UCR were similar across all three approaches, ranging from 60–93% after ORP, 66–95% after LRP and 84–97% after RARP, validating the findings of the only randomized controlled trial comparing RARP and LRP.
A recent meta-analysis of comparative studies by Ficarra et al. reported that the absolute risk of urinary incontinence was 11.3% after ORP and 7.5% after RARP, with a 3.8% absolute risk reduction in favor of RARP (OR 1.53; P = 0.03). Similarly, the absolute risk of urinary incontinence was 9.6% after LRP and 5% after RARP, showing an absolute risk reduction of 4.6% in favor of RARP (OR 2.39; P = 0.006).
Another major complication that is detrimental to the quality of life of post-prostatectomy patients is erectile dysfunction. Preservation of the periprostatic NVB is of utmost importance in maintaining potency. Walsh and Donker first described an anatomical nerve-sparing technique in ORP in 1982 by defining the anatomy of the NVB in the periprostatic fascia. Tewari et al. subsequently further refined the anatomy of the NVB, and suggested potential areas of injuries to the NVB during surgery and how to minimize them. Energy sources applied in close proximity to the NVB can cause injuries resulting in decrease potency rates[75, 76] leading to the Pasadena Consensus Panel's recommendations of minimizing traction and avoiding the use of any thermal energy within 5–10 mm of the NVB. The complete preservation of the prostatic fascia (Veil of Aphrodite) resulted in an excellent potency rate in post-prostatectomy patients. Menon et al. made further modifications to the “veil” technique recently, sparing additional nerves in the posterolateral and anterior aspect of the prostate, and preserving the puboprostatic ligaments and the dorsal venous plexus. As a result, 94% of patients not only had successful, but also earlier, intercourse with a median SHIM-5 score of 18.
Assessing potency after radical prostatectomy is extremely difficult. Parameters affecting potency are multifactorial, including age of patient, preoperative erectile function and comorbidities, stage and location of tumours, use of medication, and penile rehabilitation. More importantly, in an attempt to obtain oncological clearance, nerve-sparing techniques may have been intentionally omitted either unilaterally or bilaterally. In addition, nerve-sparing techniques vary in type and quality.
A total of 13 comparative studies were identified (Table 3). The criterion and assessment of post-prostatectomy erectile function among the studies were non-standardized and often involved non-validated questionnaires or interviews, making comparison of outcomes between studies impossible. Potency rates in studies involving radical prostatectomies carried out before the year 2006[27, 32, 45, 56] were conflicting. Krambeck et al. and Joseph et al. showed no differences in potency when comparing RARP with ORP or LRP. In contrast, Tewari et al. and Hu et al. showed conflicting results. Tewari et al. reported significantly shorter median time to return of erections and intercourse in RARP than ORP, whereas Hu et al. found that erectile dysfunction rates were significantly higher after MIRP than ORP. This discrepancy can be attributed to the initial steep learning curve of RARP and the underdevelopment of techniques of NVB preservation then.
Table 3. Erectile function in the comparative studies evaluating ORP and LRP versus RARP
|Study||Year of publication||Method||Criterion||No. cases (n)||Results|
|ORP vs RARP|| || || || || || |
|Tewari et al.||2003||Interview||Presence of erection||100 ORP||Median time to erection recovery||440 days|
|200 RARP||180 days|
|Erection sufficient for intercourse||100 ORP||Median time to intercourse||700 days|
|200 RARP||340 days|
|Krambeck et al.||2009||Non-validated questionnaire||Erection sufficient for intercourse||564 ORP||ERR at 12 months||62.8%|
|Hu et al.||2009||Administrative data||Not specified||6899 ORP||Erectile dysfunction per 100 person-years||12.2 per 100 person-years|
|1938 MIRP||15.9 per 100 person-years|
|Ficarra et al.||2009||IIEF||IIEF >17||41 ORP||ERR at 12 months||49%|
|Rocco et al.||2009||Interview||Erection sufficient for intercourse||240 ORP||ERR at 3, 6, 12 months||18%, 31%, 41%|
|Di Pierro et al.||2011||Non-validated questionnaire||Erection sufficient for intercourse||49 ORP||ERR at 3 and 12 months||25%, 26%|
|37 RARP||68%, 55%|
|Kim et al.||2011||Interview||Erection sufficient for intercourse||122 ORP||ERR at 3, 12, 24 months||6.7%, 28.1%, 47.5%|
|373 RARP||33.0%, 57.1%, 83.8%|
|Barry et al.||2012||Non-validated questionnaire||Symptoms causing moderate or severe problems||210 ORP||Erectile dysfunction at median time of 434 days from surgery||89%|
|383 RARP||Erectile dysfunction at median time of 418 days from surgery||87.5%|
|LRP vs RARP|| || || || || || |
|Joseph et al.||2005||IIEF||Erection sufficient for intercourse||50 LRP||ERR at 3 months||36%|
|Hakimi et al.||2009||IIEF||Erection sufficient for intercourse||45 LRP||ERR at 3, 12 months||20%, 71.1%|
|51 RARP||31.4%, 76.5%|
|Asimakopoulos et al.||2011||IIEF||Erection sufficient for intercourse||64 LRP||ERR at 12 months||32%|
|Park et al.||2011||Interview||Erection sufficient for intercourse||62 LRP||ERR at 12 months||48%|
Armed with more knowledge and overcoming the steep learning curves of techniques of NVB preservation in RARP, perhaps the more contemporary studies were able to shed some light on this issue. Of the five trials comparing RARP with ORP (Table 3),[34, 37, 43, 65, 66] four of these (3 prospective and non-randomized, 1 retrospective case–control trial) showed statistically significant earlier erection recovery and higher 12 months potency rates.[34, 37, 43, 65] The largest of these studies by Kim et al., who excluded patients with pre-existing erectile dysfunction or on neoadjuvant hormonal therapy, reported higher potency rates even at 2 years. In addition, younger patients, those with a higher preoperative serum testosterone and those who underwent RARP were identified as independent prognostic factors for potency recovery on multivariate analysis. Di Pierro et al. only analyzed patients who were preoperatively potent without PDE5-I. Comparing the ORP and RARP groups, potency (with and without PDE5-I) was seen in 25% and 68% of patients 3 months postoperatively (P = 0.008), and 26% and 55% of patients 12 months postoperatively (P = 0.009), respectively. Considering only patients undergoing bilateral nerve-sparing surgery, Ficarra et al. reported at ≥12 months of follow up, 49% in the ORP and 81% in the RARP group were potent (P < 0.001). Similarly, evaluating only the patients aged <65 years with a Charlson score of ≤2, 58% in the ORP and 84% in the RARP group were potent (P = 0.01). In this subgroup of patients, the mean time to recovery of erectile function was 6.7 months in the ORP group and 3.9 months in the RARP group (P < 0.01). Barry et al., after analyzing 20% of Medicare claims files in 2008, noted no difference in potency rates between ORP and RARP. In fact, erectile dysfunction rates of both arms were very high (close to 90%). As this was a self-reported cross-sectional study using a non-validated questionnaire, many important potential confounders, such as age of patients, surgical techniques and baseline erectile function, were not accounted for in the analysis. In fact, the categorization of patients into either the RARP or the OPR groups was based solely on patients' inputs in their surveys.
Four studies compared the potency rates of patients after LRP and RARP,[45, 48-50] of which three were retrospective,[45, 48, 50] and one was prospective and randomized. Although the three retrospective studies showed better potency rates of RARP over LRP, they failed to reach statistical significance. Asimakopoulos et al. convincingly showed in the only prospective and randomized trial to date that RARP had a clear advantage over LRP in terms of 12 months potency rates (77% vs 32%, respectively).
A recent meta-analysis of comparative studies described the prevalence of erectile dysfunction as 47.8% after ORP and 24.2% after RARP. The cumulative analysis showed a statistically significant advantage in favor of RARP (OR 2.84; P = 0.002), and the absolute risk reduction for erectile dysfunction was 23.6%. The meta-analysis also reported that the prevalence of erectile dysfunction was 55.6% after LRP and 39.8% after RARP. Cumulative analysis showed a non-statistically significant trend in favour of RARP (OR 1.89; P = 0.21).
We identified 13 comparative studies (Table 4) comparing operative time of RARP with ORP.[26-28, 30, 32, 34, 36, 37, 40, 43, 83, 85, 87] The results are conflicting, with six studies[26, 30, 32, 34, 37, 43, 85] reporting longer operative time for RARP as compared with ORP, and the other seven studies showing either similar[27, 28, 36, 40] or shorter[83, 87] operative time. It is interesting to note that of the seven studies that reported a longer operative time for RARP, five series[26, 30, 34, 37, 43] had <130 cases, and the majority[26, 30, 34, 37, 85] reported impressive times for their ORP (127–160 min). In contrast, Tewari et al. and Gainsburg et al. showed that mean operative time can be cut to less than 3 h at ≥200 cases. Although Krambeck et al. reported a statistically significant longer operative time in their 286 RARP, the mean time difference was just 32 min. These findings might be reflective of the number of cases required to overcome the learning curve of RARP. Indeed Doumerc et al. found in their experience that the learning curve for RARP flattens only after 140–170 cases. Cumulative analysis of comparative studies validated that operative time of RARP and ORP was similar.
Table 4. Perioperative parameters in the comparative studies evaluating ORP and LRP versus RARP
|Study||Year||Cases||Operative time (mins) mean/median||Blood loss (mL) mean/median||Transfusion rate (%)||Catheterization duration (days)||In-hospital stay (days)||Overall complication rates (%)||Death (%)|
|ORP vs RARP|| || || || || || || || || |
|Menon et al.||2002||30 ORP||138||970||17||14||56 h||6||–|
|30 RARP||288||329||7||11||36 h||6||–|
|Tewari et al.||2003||100 ORP||163||910||67||15.8||3.5||15||–|
|Ahlering et al.||2004||60 ORP||214||418||2||9||2.2||10||–|
|Farnham et al.||2006||103 ORP||–||664||2.9||–||–||–||–|
|Nelson et al.||2007||374 ORP||–||–||–||–||1.23||15||–|
|Fracalanza et al.||2008||26 ORP||127||500||34||–||8||27||–|
|Krambeck et al.||2009||564 ORP||204||–||13.1||–||LOS = 1 day: 29.3%||8||–|
|286 RARP||236||–||5.1||–||LOS = 1 day: 19.4%||4.8||–|
|Hu et al.||2009||6 899 ORP||–||–||20.8||–||3||23.2||0.2|
|1 938 MIRP||–||–||2.7||–||2||22.2||0.1|
|Ficarra et al.||2009||105 ORP||135||500||14||6||7||9.7||–|
|Rocco et al.||2009||240 ORP||160||800||–||7||6||–||–|
|Drouin et al.||2009||83 ORP||209||821||9.6||14.7||7||13||–|
|Gainsburg et al.||2010||106 ORP||318||1200||27||–||3||–||–|
|Carlsson et al.||2010||485 ORP||–||–||23||–||–||32.8||0.2|
|1 253 RARP||–||–||4.8||–||–||15.7||0|
|Doumerc et al.||2010||502 ORP||148||–||2||7.9||5.5||0.8||–|
|Kordan et al.||2010||414 ORP||–||450||3.4||–||–||–||–|
|Breyer et al.||2010||695 ORP||–||–||8||–||–||–||–|
|Lo et al.||2010||20 ORP||289||–||65||18||17||–||–|
|Truesdale et al.||2010||217 ORP||204||904||–||–||–||–||–|
|Di Pierro et al.||2011||75 ORP||253||–||3||–||–||37||–|
|Kowalczyk et al.||2012||58 638 ORP||–||–||17.3||–||4.2||29.8||0.6|
|19 594 MIRP||–||–||2.6||–||2||19.6||0.2|
|Trinh et al.||2012||7 389 ORP||–||–||7.7||–||39.6% >2 days||11.1||0.1|
|11 889 RARP||–||–||2.0||–||14.5% >2 days||8.2||0|
|Yu et al.||2012||9 704 ORP||–||–||5.2||–||2.4||10.1||0.2|
|11 513 RARP||–||–||1.6||–||1.7||8.4||0|
|Anderson et al.||2012||8 968 ORP||–||–||–||–||2.5||–||0.1|
|12 588 RARP||–||–||–||–||1.7||–||0|
|LRP vs RARP|| || || || || || || || || |
|Menon et al.||2002||48 LRP||258||391||2.5||–||LOS <1 day 65%||–||–|
|50 RARP||274||256||0||–||LOS <1 day 80%||–||–|
|Joseph et al.||2005||50 LRP||235||299||0||–||–||–||–|
|Hu et al.||2006||358 LRP||246||200||2.2||–||–||33||–|
|Rozet et al.||2007||133 LRP||160||512||3||9||4.9||9.1||–|
|Drouin et al.||2009||85 LRP||257||558||5.9||8.9||6.1||7||–|
|Hakimi et al.||2009||75 LRP||232||311||–||–||3.4||14.7||0|
|Trabulsi et al.||2010||45 LRP||300||299||4.4||–||2.6||–||–|
|Park et al.||2011||62 LRP||308||214||0||9||7||11.3||–|
|Asimakopoulos et al.||2011||64 LRP||–||–||5||7.45||–||8||–|
|Yu et al.||2012||2 167 LRP||–||–||–||–||2||14.5||0|
|11 513 RARP||–||–||–||–||1.7||8.4||0|
|Anderson et al.||2012||547 LRP||–||–||–||–||1.8||–||0.1|
|12 588 RARP||–||–||–||–||1.9||–||0|
Eight comparative studies reported operative time of RARP and LRP. Four studies[36, 48, 68, 92] showed that RARP has a statistically significant shorter operative time than LRP. Only one study by Park et al. showed a significant increase in operative time of RARP over LRP. The authors attributed this finding to the fact that more nerve-sparing procedures were carried out in the RARP arm. Cumulative analysis of comparative studies showed similar operative times in both groups.
Blood loss and transfusion
Smaller surgical incisions, magnified visualization, more precise hemostasis and the presence of pneumoperitoneum in RARP are factors postulated to lead to a reduction in bleeding intraoperatively and the need for transfusion of blood products. This finding was confirmed in the 21 comparative studies that described the EBL and/or transfusion rates between ORP and RARP (Table 4). The vast majority of them showed a statistically significant reduction of EBL and decreased need for blood transfusions in RARP when compared with ORP. Kordan et al. noted that the need for transfusion was only significantly associated with the type of procedure carried out (RARP or ORP), and no association was found on univariate or multivariate analysis with other parameters, such as patients' age, body mass index, PSA level, clinical stage of disease, neoadjuvant hormonal therapy, previous radiotherapy or biopsy, Gleason scores, prostate or tumor volume, extracapsular extension or seminal vesicle involvement.
Population-based studies involving huge cohorts reported a significant advantage of RARP/MIRP over ORP in reducing the need for blood transfusion.[67, 88, 89] Meta-analysis of non-comparative studies established higher EBL and transfusion rates for ORP (745.3 mL; 16.5%) than RARP (188.0 mL; 1.8%). Cumulative analyses of comparative studies confirmed these findings.[51, 93]
LRP and RARP offered identical advantages characteristic of MIRP. Nine comparative studies were identified, four of which showed that RARP was associated with a lower EBL,[36, 45, 48, 91] whereas only one showed that RARP was associated with a higher transfusion rate over LRP. The reason for the higher transfusion rate associated with RARP was not entirely evident from Rozet et al.'s study. However, in view of similar EBL between LRP and RARP in that study, the higher rate of transfusion might be a result of other postoperative issues or differences in postoperative management rather than from blood loss. Meta-analysis of older comparative studies showed similar EBL and transfusion rates. A recent meta-analysis of more contemporary comparative studies showed a significantly lower transfusion rate in RARP than LRP. Cumulative analyses of non-comparative studies verified the aforementioned findings with higher EBL and transfusion rates for LRP (377.5 mL; 4.7%) than for RARP (188.0 mL; 1.8%), although the difference in transfusion rates did not reach statistical significance (P = 0.07).
Length of in-hospital stay and catheterization duration
Table 4 shows that catheterization duration was generally similar across all three surgical approaches. Cumulative analysis was not possible in the two meta-analyses.[51, 93] With regards to length of hospitalization, RARP was consistently associated with a shorter in-hospital stay in all of the comparative studies when compared with ORP (Table 4), reaching statistical significance in the majority. Hohwu et al. reported a shorter convalescent time and earlier return to work with RARP. Findings in the four population-based studies[67, 88-90] concurred with the comparative studies. Generally, no differences in the catheterization and the hospitalization duration were noted in the comparative studies involving RARP and LRP (Table 4).
The cumulative analysis of non-comparative studies by Tewari et al. summarized the issue of hospitalization duration concisely. RARP had significantly the shortest hospital stay, both in the USA studies (1.4 days) and in the non-USA studies (4.0 days), with LRP intermediate (2.1 days USA, 6.3 days non-USA), and ORP having the longest length of stay (3.1 days USA, 9.9 days non-USA).
Historically, radical prostatectomy was a highly morbid operation, with reported mortality rates of 2.5% and complication rates soaring as high as 60%.[95, 96] Fortunately, things are vastly different today. Mortality rates from population-based studies[56, 67, 88, 90] range from 0.1–0.6% in ORP, and 0–0.1% in LRP and RARP. Meta-analysis of non-comparative studies showed significant differences in mortality rates favoring MIRP over ORP (0.1% ORP, 0.04% LRP, 0.04% RARP).
Overall complication rates (Table 4) were generally similar between ORP and RARP. Just four out of 12 studies[27, 30, 36, 84] showed lower less overall complication rates in favor of RARP. Cumulative analyses of comparative studies show similar overall complications between ORP and RARP.[51, 93] There were generally no differences with regards to the overall complication rates between LRP and RARP. Hu et al. and Rozet et al. showed contrasting results in terms of overall complication rates, but this difference is likely due to the experience and expertise of the respective centers rather than the surgical approaches. As predicted, cumulative analyses confirmed similar overall complications between LRP and RARP. Population-based studies generally reported lower overall complication rates in favor of MIRP.[67, 88, 89] Cumulative analysis of perioperative overall complication rates in non-comparative studies established significant advantages of RARP over both LRP and ORP (17.9% ORP, 11.1% LRP and 7.8% RARP).
Four comparative studies in the present review gave a breakdown of the specific complications between RARP and ORP.[26, 32, 37, 84] Although Menon et al. and Ficarra et al. found no significant difference in terms of rectal injuries, postoperative ileus and bleeding, retention of urine, and wound dehiscence, Krambeck et al. reported a significantly higher risk of incisional hernia in RARP, but lower rates of bladder neck contractures. Carlsson et al. reported significantly higher rates of rectal injuries and postoperative pulmonary embolism, pneumonia and wound infection in the ORP cohort.
Three comparative studies compared specific complications between LRP and RARP.[46, 48, 92] Hu et al. reported a higher risk of urinary complications (urine leaks and anastomotic strictures), rectal, vascular and neurological injuries, postoperative ileus, and conversion to ORP rates in LRP as opposed to RARP. Rozet et al., in contrast, reported a higher risk of urinary tract infections and postoperative bleeding in RARP. 
Population-based studies registered higher risks of respiratory,[56, 67, 88, 89] cardiovascular,[56, 67, 88] vascular,[67, 88, 89] wound,[67, 88, 89] genitourinary and other miscellaneous[67, 88] complications in ORP as compared with RARP/MIRP. Evaluation of non-comparative studies showed that, when compared with RARP, LRP was associated with a higher risk of re-admissions and the need for repeat operations, nerve and rectal injuries, and fistulae formation; whereas ORP was associated with a higher risk of ureteral injuries and deep vein thrombosis. When compared with MIRP, ORP was associated with a higher risk of hematoma, lymphocele, pneumonia, anastomotic leak and wound infection. There were no differences in rates of vascular and small bowel injuries between the three modalities.