- Email: [email protected]
Factors That Impact the Outcome of Minimally Invasive Pyeloplasty: Results of the Multi-Institutional Laparoscopic and Robotic Pyeloplasty Collaborative Group
Factors That Impact the Outcome of Minimally Invasive Pyeloplasty: Results of the Multi-Institutional Laparoscopic and Robotic Pyeloplasty Collaborative Group Steven M. Lucas, Chandru P. Sundaram,* J. Stuart Wolf, Jr., Raymond J. Leveillee, Vincent G. Bird, Mohamed Aziz, Stephen E. Pautler, Patrick Luke, Peter Erdeljan, D. Duane Baldwin,† Kamyar Ebrahimi, Robert B. Nadler, David Rebuck, Raju Thomas, Benjamin R. Lee, Ugur Boylu, Robert S. Figenshau, Ravi Munver, Timothy D. Averch,‡ Bishoy Gayed, Arieh L. Shalhav, Mohan S. Gundeti, Erik P. Castle,§ J. Kyle Anderson, Branden G. Duffey, Jaime Landman, Zhamshid Okhunov, Carson Wong and Kurt H. Strom From Indiana University, Indianapolis, Indiana (SML, CPS); University of Michigan, Ann Arbor, Michigan (JSW); University of Miami, Miami, Florida (RJL, VGB, MA); University of Western Ontario, London, Ontario, Canada (SEP, PL, PE); Loma Linda University Medical Center, Loma Linda (DDB, KE), and University of California-Irvine, Irvine (JL, ZO), California; Northwestern University (RBN, DR) and University of Chicago (ALS, MSG), Chicago, Illinois; Tulane University, New Orleans, Louisiana (RT, BRL, UB); Washington University School of Medicine in St. Louis, St. Louis, Missouri (RSF); Hackensack University, Hackensack, New Jersey (RM); University of Pittsburgh, Pittsburgh, Pennsylvania (TDA, BG); Mayo Clinic Arizona, Scottsdale, Arizona (EPC); University of Minnesota, Minneapolis, Minnesota (JKA, BGD); and University of Oklahoma, Oklahoma City, Oklahoma (CW, KHS)
Abbreviations and Acronyms FSP ⫽ freedom from secondary procedure LP ⫽ laparoscopic pyeloplasty RP ⫽ robotic pyeloplasty UPJ ⫽ ureteropelvic junction UTI ⫽ urinary tract infection Submitted for publication June 2, 2011. Study received institutional review board approval. * Correspondence: Department of Urology, Indiana University Medical Center, 535 N. Barnhill Drive, Suite 420, Indianapolis, Indiana 46202 (telephone: 317-278-3098; FAX: 317-274-0174; e-mail: [email protected]). † Financial interest and/or other relationship with Onset Medical Corp. ‡ Financial interest and/or other relationship with Bard Medical, Cook Medical and Storz. § Financial interest and/or other relationship with Baxter and Intuitive Surgical.
Purpose: We compared laparoscopic and robotic pyeloplasty to identify factors associated with procedural efficacy. Materials and Methods: We conducted a retrospective multicenter trial incorporating 865 cases from 15 centers. We collected perioperative data including anatomical and procedural factors. Failure was defined subjectively as pain that was unchanged or worse per medical records after surgery. Radiographic failure was defined as unchanged or worsening drainage on renal scans or worsening hydronephrosis on computerized tomography. Bivariate analyses were performed on all outcomes and multivariate analysis was used to assess factors associated with decreased freedom from secondary procedures. Results: Of the cases 759 (274 laparoscopic pyeloplasties with a mean followup of 15 months and 465 robotic pyeloplasties with a mean followup of 11 months, p ⬍0.001) had sufficient data. Laparoscopic pyeloplasty, previous endopyelotomy and intraoperative crossing vessels were associated with decreased freedom from secondary procedures on bivariate analysis, with a 2-year freedom from secondary procedures of 87% for laparoscopic pyeloplasty vs 95% for robotic pyeloplasty, 81% vs 93% for patients with vs without previous endopyelotomy and 88% vs 95% for patients with vs without intraoperative crossing vessels, respectively. However, on multivariate analysis only previous endopyelotomy (HR 4.35) and intraoperative crossing vessels (HR 2.73) significantly impacted freedom from secondary procedures. Conclusions: Laparoscopic and robotic pyeloplasty are highly effective in treating ureteropelvic junction obstruction. There was no difference in their abilities to render the patient free from secondary procedures on multivariate analysis. Previous endopyelotomy and intraoperative crossing vessels reduced freedom from secondary procedures. Key Words: laparoscopy, robotics, reconstructive surgical procedures, survival analysis
522
www.jurology.com
0022-5347/12/1872-0522/0 THE JOURNAL OF UROLOGY® © 2012 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION
AND
RESEARCH, INC.
Vol. 187, 522-527, February 2012 Printed in U.S.A. DOI:10.1016/j.juro.2011.09.158
MULTI-INSTITUTIONAL ANALYSIS OF MINIMALLY INVASIVE PYELOPLASTY
OPEN surgical dismembered pyeloplasty has been the standard treatment for ureteropelvic junction obstruction and is successful in 90% to 95% of cases.1,2 Minimally invasive treatments for UPJ obstruction offer faster recovery and decreased postoperative pain.3 With concerns of inferior long-term success rates of endopyelotomy,4 minimally invasive pyeloplasty is increasingly used for the treatment of this condition, with success rates equivalent to those of open pyeloplasty.5–7 With the advent of robotic surgery, acquiring laparoscopic skills can be hastened, which has made minimally invasive pyeloplasty more widely available to many surgeons.8 Among those who are experienced in minimally invasive pyeloplasty, it is unknown whether robotic surgery improves outcomes. Factors such as degree of hydronephrosis, renal function and previous procedures influence the outcome of open pyeloplasty.2 However, because it is a newer treatment the factors that influence the outcomes of laparoscopic or robotic pyeloplasty have not been well studied. Since the success rates of these procedures are high, it is difficult to ascertain in single center series what may have contributed to treatment failure due to the limited number of failures. In this study several centers contributed to the Laparoscopic and Robotic Pyeloplasty Collaborative Group, and a larger database was created that allowed a more thorough assessment of the impact of preoperative and intraoperative parameters on the outcomes of LP and RP.
METHODS Centers with experience in minimally invasive pyeloplasty, all sites with Endourological Society recognized fellowships or active members of the Society of Urological Robotic Surgery, were asked to submit data to a retrospective database which included preoperative, intraoperative and followup data in this institutional review board approved study. Median (IQR) patient age was 35.0 (28.0) and 35.1 (22.0) years for RP and LP, respectively (p ⫽ 0.957). Preoperative computerized tomography or ultrasound was used to assess hydronephrosis and crossing vessels, and mercaptoacetyltriglycine renal scan with diuretic washout determined renal function and drainage. Median preoperative T1/2 was 30 (22) and 24 (19) minutes for RP and LP, respectively (p ⫽ 0.001). Median percent preoperative renal function was 43 (16) and 42 (19) for RP and LP, respectively (p ⫽ 0.425). Intraoperatively LP was differentiated from RP. Surgery performed using the traditional laparoscopic approach without robotic assistance was included in the LP group. The case was considered RP if the robot was used to perform the reconstruction. Postoperative data included drain duration, urine leak, complications and hospital stay. Recognizing that there may be differences in the management of drains and in the diagnosis of urine leak,
523
a urine leak was recorded if directly reported by the submitting center and the persistence of drain recorded if 7 days or more. Followup data included initial and final radiographic studies. Worsened hydronephrosis, or unchanged or worsened T1/2 was considered recurrent or persistent obstruction. Worsened or unchanged pain was counted as subjective treatment failure. A radiographic or subjective failure was considered treatment failure. Secondary interventions included prolonged stent placement, endopyelotomy or dilation, repeat pyeloplasty or nephrectomy. FSP was used to quantify patients who did not require these interventions. In terms of statistical analysis, bivariate analyses of preoperative and intraoperative factors were determined by Mann-Whitney U, Kruskal-Wallis or chi-square analysis. Kaplan-Meier analysis was used to determine FSP with log rank comparison of the presence or absence of each factor. Multivariate analysis of factors was conducted for FSP using Cox regression analysis of hazard ratios. Statistical analysis was conducted with PASW® statistics 17 with type 1 error set at 0.05.
RESULTS Treatment Modality Of 865 patients from 15 centers 759 were available for comparison, including 274 LP and 485 RP. Mean followup was 15 ⫾ 16 months for LP and 11 ⫾ 13 months for RP (p ⬍0.001). Preoperatively the LP group had higher rates of previous endopyelotomy, urinary drainage, UTI and pain. The RP group had significantly more stones and longer T1/2 (table 1). Median operative times, complications, hospital stay, urine leak and prolonged drain duration were similar for LP and RP, while nondismembered pyeloplasties were more common for LP (table 2). Rates of failure to improve obstruction, pain and need for secondary procedures were higher in the LP than in the RP group. FSP at 1 and 2 years was greater for RP than LP at 97% ⫾ 1.0% vs 92% ⫾ 2.0% and 95% ⫾ 1.8% vs 87% ⫾ 3.1%, respectively (p ⫽ 0.019, see figure). Table 1. Preoperative characteristics of robotic and laparoscopic pyeloplasty No. Robotic/ Total No. (%) Female pts Rt side cases Preop UTI Preop pain Preop crossing vessel Preop severe hydronephrosis Preop renal stone Previous abdominal procedure Previous endopyelotomy Previous pyeloplasty Preop stent Preop percutaneous nephrostomy
138/368 (49.7) 233/482 (48.3) 73/429 (17.0) 354/429 (82.3) 187/412 (45.4) 109/322 (33.9) 76/415 (18.3) 72/235 (30.6) 40/485 (8.2) 6/485 (2.2) 179/470 (38.1) 15/470 (3.2)
No. Laparoscopic/ Total No. (%) p Value 154/274 (56.2) 133/273 (48.7) 50/200 (25.0) 247/272 (90.8) 80/208 (38.5) 81/216 (37.5) 31/263 (11.8) 43/186 (23.1) 38/274 (13.9) 6/274 (1.4) 143/250 (57.2) 12/250 (4.8)
0.104 0.921 0.019 0.002 0.100 0.385 0.023 0.085 0.014 0.312 ⬍0.001 ⬍0.001
524
MULTI-INSTITUTIONAL ANALYSIS OF MINIMALLY INVASIVE PYELOPLASTY
Table 2. Comparison of intraoperative and postoperative outcomes of robotic and laparoscopic pyeloplasty
Median operative mins (IQR) Median ml blood loss (IQR) No./total No. nondismembered pyeloplasties (%) No. intraop complications (%) No. postop complications (%) No./total No. urine leaks (%) No./total No. prolonged drain 7 days or more (%) No./total No. stent related problems (%) Median % postop function (IQR) Median mins postop T1/2 (IQR) No./total No. obstruction not improved (%) No./total No. pain not improved (%) No./total No. secondary intervention (%)
Table 3. Effect of previous endopyelotomy on outcome of minimally invasive pyeloplasty
Robotic
Laparoscopic
p Value
204 (85) 30 (30) 31/427 (7.3)
218 (124) 50 (75) 34/273 (12.5)
0.123 0.001 0.021
9 (1.9) 26 (5.4) 9/485 (1.9) 29/258 (11.2)
6 22 9/274 13/140
(2.2) (8.0) (3.3) (9.3)
0.670 0.275 0.303 0.554
7/422 (1.7)
9/242 (3.7)
0.096
45 (13) 10 (7) 15/334 (4.5)
45 (16) 10 (9) 24/221 (10.9)
0.397 0.087 0.004
15/328 (4.6) 12/375 (3.2)
16/175 (9.1) 23/242 (9.5)
0.042 0.001
Among primary dismembered pyeloplasties (207 LP and 354 RP) the laparoscopic group had higher rates of postoperative obstruction (11% vs 3.8%, p ⫽ 0.004), pain (5.3% vs 1.7%, p ⫽ 0.016) and secondary procedures (7.7% vs 1.8%, p ⫽ 0.002). FSP at 1 and 2 years for LP vs RP was 94% ⫾ 2.0% and 89% ⫾ 3.4% vs 99% ⫾ 1.0% and 96% ⫾ 1.8%, respectively (p ⫽ 0.018). Previous Procedures A total of 78 patients underwent endopyelotomy before pyeloplasty, which did not affect urine leak or drain duration postoperatively, but did increase the rate of subjective failure and secondary procedures (p ⬎0.05, table 3). At 1 and 2 years FSP was lower for patients who previously underwent endopyelotomy at 90% ⫾ 4.5% and 81% ⫾ 7.3% vs 97% ⫾ 1.0% and 93% ⫾ 1.8%, respectively (p ⬍0.001, see figure).
No Previous Previous p Endopyelotomy Endopyelotomy Value No. pts Median % preop function (IQR) Median mins preop T1/2 (IQR) No./total No. failed obstruction (%) No./total No. failed pain (%) No./total No. secondary procedure (%) No./total No. drain 7 days or more (%) No./total No. urine leak (%) Median % postop function (IQR) Median mins postop T1/2 (IQR) Median operative mins (IQR)
669 43 (16) 29 (16) 30/466 (6.4) 20/632 (3.2) 25/548 (4.6) 40/356 (11.2) 13/669 (1.9) 45 (13) 10 (8) 209 (100)
78 39 (16) 24 (33) 6/54 (11.1) 7/74 (9.5) 10/69 (14.5) 2/38 (5.3) 2/78 (2.6) 44 (15) 9 (10) 205 (90)
0.112 0.435 0.200 0.008 0.001 0.257 0.556 0.307 0.876 0.863
Previous pyeloplasty was performed in 12 patients and the median operative time was longer for those with prior pyeloplasty (272.0 vs 208.0 minutes, p ⫽ 0.022). In these 12 patients pain failed to improve in 1 patient and obstruction failed to improve in another, and secondary procedures were performed in these 2 patients. Preoperative Urinary Drainage Ureteral stents were placed a day before pyeloplasty (most often for pain) in 322 patients vs 353 without placement, resulting in similar subjective failure rates (7 of 322, 2.2% vs 16 of 353, 4.5%; p ⫽ 0.092), radiographic failure rates (16 of 251, 6.4% vs 16 of 248, 6.5%; p ⫽ 0.972) and secondary procedure rates (9 of 270, 3.3% vs 19 of 280, 6.8%; p ⫽ 0.066). While the urine leak rate was lower in the preoperative stent group (0.9% vs 3.4%, p ⫽ 0.030), a higher proportion had a drain for 7 days or more in the preoperative stent group (14.5% vs 5.6%, p ⫽ 0.004). Crossing Vessels Based on preoperative computerized tomography 267 patients had crossing vessels and 354 did not, while
A, univariate analysis of FSP after laparoscopic and robotic pyeloplasty. B, univariate analysis of impact of previous endopyelotomy on FSP after minimally invasive pyeloplasty. C, univariate analysis of impact of intraoperative crossing vessels on FSP after minimally invasive pyeloplasty. Intraop, intraoperative.
MULTI-INSTITUTIONAL ANALYSIS OF MINIMALLY INVASIVE PYELOPLASTY
226 with crossing vessels and 259 without crossing vessels were noted intraoperatively. An intraoperative crossing vessel resulted in a higher subjective failure rate (5.8% vs 2.2%, p ⫽ 0.048), similar radiological failure rate (6.9% vs 4.2%, p ⫽ 0.284) and a higher rate of secondary procedures (8.6% vs 2.2%, p ⫽ 0.013). FSP at 1 and 2 years was significantly lower in cases with intraoperative crossing vessels at 95% ⫾ 1.7% and 88% ⫾ 3.4% vs 97% ⫾ 1.7% and 95% ⫾ 2.1%, respectively (p ⫽ 0.028, see figure). Nondismembered Pyeloplasty Nondismembered pyeloplasty was performed in 65 cases (18 Fenger plasties, 38 Y-V plasties, 1 spiral flap, 3 other flaps and 5 unknown) while 635 were dismembered. Median operative time was 240 minutes (IQR 137) for nondismembered and 208 (IQR 100) for dismembered pyeloplasties (p ⫽ 0.081). Pain did not improve in 11% (7 of 65) nondismembered vs 3.3% (21 of 635) dismembered (p ⫽ 0.003), radiological failure occurred in 8.7% (4 of 46) vs 6.9% (33 of 480, p ⫽ 0.645) and secondary procedures occurred in 11% (6 of 57) vs 5% (27 of 493, p ⫽ 0.100), respectively. FSP was similar between the 2 groups (p ⫽ 0.083). Urine leak and prolonged drain were similar between nondismembered and dismembered pyeloplasties at 2 of 65 (3.1%) and 5 of 35 (14%) vs 13 of 635 (2.0%) and 37 of 363 (10.2%), respectively (p ⬎0.4). Number of Prior Cases The median number of cases performed before LP and RP was 23.5 (IQR 43) and 35 (IQR 45), respectively (p ⬍0.001). FSP for pyeloplasties performed with 30 or less prior cases at 1 and 2 years was 93.9% ⫾ 1.6% and 89.4% ⫾ 2.4% vs 97.2% ⫾ 1.0% and 94.7% ⫾ 2.1% for those with more than 30 prior cases (p ⫽ 0.261). There was no difference in radiological failure (7.3% vs 6.6%) or subjective failure (6.8% vs 4.1%, p ⬎0.10). Considering only cases that were performed after more than 30 prior cases, FSP for LP and RP at 1 year was 93.2% ⫾ 3.0% vs 100%, and at 2 years was 93.2% ⫾ 3.0% vs 95.3% ⫾ 3.2% (p ⫽ 0.055). Multivariate Analysis Treatment modality, intraoperative crossing vessels and previous endopyelotomy significantly affected FSP on bivariate analysis (table 4). Multivariate analysis revealed previous endopyelotomy (hazard ratio 4.35, p ⫽ 0.001) and intraoperative crossing vessels (hazard ratio 2.73, p ⫽ 0.046) to be significant predictors of secondary procedures.
DISCUSSION In this study we present the largest comparative series of LP and RP reported to our knowledge, with success rates greater than 90%. On bivariate anal-
525
Table 4. Multivariate analysis of factors influencing the rate of secondary procedure following minimally invasive pyeloplasty Kaplan-Meier Log Rank Multivariate HR Multivariate p Value (95% CI) p Value Purely LP Previous endopyelotomy Intraop crossing vessel Nondismembered pyeloplasty Severe hydronephrosis Preop stone Preop stent Preop function less than 40% Preop T1/2 greater than 35 mins Preop UTI
0.019 ⬍0.001 0.028 0.083 0.099 0.997 0.176 0.486 0.357 0.061
1.99 (0.852–4.67) 1.85 (1.77–11.0) 2.73 (1.02–7.34)
0.112 0.001 0.044
ysis RP had some advantages compared to LP in terms of persistence or recurrence of radiographic obstruction (4.5% vs 10.9%) and symptoms (4.6% vs 9.1%). The need for secondary procedures was also lower in the RP group (3.2% vs 9.5%). RP and LP had similar complication rates. Followup was longer for LP but FSP was greater for robotic than for laparoscopic pyeloplasty (97% vs 94% at 1 year). However, there were significant differences between the 2 groups that likely confounded the bivariate analysis. Thus, FSP was analyzed in a multivariate analysis including significant preoperative factors (previous endopyelotomy, crossing vessels and treatment modality). In this analysis previous endopyelotomy and intraoperative crossing vessels were significantly associated with decreased FSP while the use of LP or RP was not. Recent series have compared LP and RP. In a meta-analysis 8 articles compared LP and RP, demonstrating similar success rates of more than 95%.7 There was a 10-minute reduction in operating time with RP. However, the analysis included only series that had RP and LP, which often meant a small number of patients (median sample size per series was 13 patients each, with only 1 series having more than 30 patients). In a separate series 30 LPs were compared with 30 RPs, of which only 1 LP required a secondary procedure, suggesting equivalent results.9 Operative time was faster with RP than with LP (98 vs 145 minutes). Additionally, Yanke et al reviewed 116 LP and 29 RP cases with success rates of 88.8% and 100%, respectively, at a mean of 20 months.4 Predominantly the inference about the success rates and operative times for RP and LP is made from observational studies of 1 modality. In a review of 25 published series comprising 740 RP cases, the mean operative time was 194 minutes and the mean success rate was greater than 90%. In 6 of the largest studies the success rate ranged from 95% to 100%, with operative times between 108 and 225
526
MULTI-INSTITUTIONAL ANALYSIS OF MINIMALLY INVASIVE PYELOPLASTY
minutes.4,10 –14 In 5 recent laparoscopic series the success rate ranged from 85% to 100% with mean operative times between 131 and 244 minutes.3,15–18 An analysis of treatment modalities is difficult without a randomized, prospective comparison. Overall it is acceptable to report that both treatments are effective for primary UPJ obstruction. However, most centers are now performing a substantial portion of pyeloplasties with robotic assistance. In centers with highly experienced laparoscopic surgeons, LP is a successful treatment option, especially when access to the robot is limited by competition for resources. In addition, the large number of pyeloplasties evaluated in this series enabled us to evaluate perioperative factors that may influence the outcome of minimally invasive pyeloplasty. Previous endopyelotomy was the strongest predictor of a secondary procedure on multivariate analysis, although data on the type of endopyelotomy performed were not available. We believe that endopyelotomy must be reserved for selected patients in whom the laparoscopic (with or without robotic assistance) approach would be difficult or contraindicated. Other series have also shown that the efficacy of minimally invasive pyeloplasty is decreased after a failed endopyelotomy, although the degree of this impact is variable. In a review of 36 patients who underwent a previous failed UPJ procedure (33 of 36 were endoscopic), the overall success rate of LP was 83%.19 In another series of 13 patients a 92% success rate was observed.20 Interestingly a preoperative ureteral stent in a setting before the actual operative procedure, which anecdotally increases periureteral inflammation and makes pyeloplasty more difficult, did not impact efficacy. While the rate of urine leak did not differ, prolonged drainage more frequently occurred with a preoperative stent than without (15% vs 5.6%). This suggests no difference in the anatomical results of these cases, but knowledge of a preoperative stent may cause surgeons to be more cautious with postoperative management of these cases. The presence of an intraoperative crossing vessel decreased symptomatic improvement and FSP after pyeloplasty. A crossing vessel has been demonstrated to decrease the efficacy of endopyelotomy but has not been reported to affect the outcome of minimally invasive pyeloplasty. In 2 series of 31 and 48 patients with a lower pole crossing vessel, success after minimally invasive pyeloplasty was observed in all patients.21,22 In a separate series of 129 LPs the presence of crossing vessels improved the success rate (hazard ratio 0.27, 95% CI 0.08 – 0.87), which was theorized to be related to a more focal stricture.4
Nondismembered pyeloplasty resulted in decreased symptomatic improvement but a similar rate of secondary procedures. In a recent series comparing dismembered pyeloplasty (50 patients) to Y-V plasty (36 patients) demonstrated success rate was 91% in both groups.23 In selected cases the success of nondismembered pyeloplasty is nearly equivalent to that of dismembered pyeloplasty. However, longer operative times in our study were noted for nondismembered cases, suggesting these were more difficult. Finally, cost must be considered when robotic assistance is used. In addition to the cost of the robot and its maintenance, disposable costs were higher for RP than LP. Although we do not have data on disposable costs at each center, the cost difference can be illustrated by estimating our own costs. Assuming in each case that, in addition to the disposable drapes, hot shears, PK dissector, large needle driver, Potts scissors and fine tissue forceps were used, a cost of $1,280 of disposables would apply. For LP, assuming the use of ultrasonic shears, disposable scissor tips and trocars, the disposable cost per case would be $750. Additional costs of 12 mm trocars, suction irrigator and stents would be incurred for both procedures. Since many centers now use robotic assistance, it will be difficult to maintain the skills required to perform a comparable anastomosis, thereby justifying the higher cost. There were several limitations to the study and some critical points merit discussion. The most significant limitation was the retrospective design. As a multi-institutional study there were differences in surgical technique, followup, and definitions of symptom improvement and urine leak which may explain why RP had an advantage on univariate but not on multivariate analysis. Although the same trend was observed on multivariate analysis, a larger data set would be required to substantiate a benefit. In addition, more LP cases were nondismembered, the reason for which cannot be elucidated in a retrospective fashion. Furthermore, the centers included in the study were high volume centers. Assessment of clinical symptoms was limited as these data were not obtained through patient reported surveys but from medical records. Another limitation was the bias created by the implementation of the robot for pyeloplasties after some of the surgeons gained experience with LP. This could result in comparison of laparoscopic pyeloplasty at an earlier point in the learning curve than the robotic cases. We corrected for this in a subgroup analysis that included only pyeloplasties performed after 30 prior cases. A slightly higher FSP was seen for RP than LP, although it was not statistically significant.
MULTI-INSTITUTIONAL ANALYSIS OF MINIMALLY INVASIVE PYELOPLASTY
An additional limitation is that the mean followup for LP and RP was 15 and 11 months respectively. Late failures were reported. For instance, in a series of 65 patients Madi et al identified 3 patients in whom recurrence developed 2 or more years after surgery compared with 7 failures within the first year.24 Finally, time to radiographic failure or symptomatic failure could not be assessed as these are dependent on reporting symptoms or obtaining imaging. Thus, Kaplan-Meier analysis could not be performed for these outcomes.
527
CONCLUSIONS In this study we present a large, multi-institutional series of minimally invasive pyeloplasties. LP and RP demonstrate a high rate of success for the treatment of UPJ obstruction. Only previous endopyelotomy and crossing vessels were significant predictors of reduced FSP and success rates on multivariate analysis. Further comparison of treatment modalities is best conducted in a randomized, controlled setting.
REFERENCES 1. O’Reilly PH, Brooman PJ, Mak S et al: The longterm results of Anderson-Hynes pyeloplasty. BJU Int 2001; 87: 287.
open, freehand, and robotic-assisted laparoscopic pyeloplasty in a swine animal model. J Am Coll Surg 2009; 208: 576.
17. Szavay PO, Luithle T, Seitz G et al: Functional outcome after laparoscopic dismembered pyeloplasty in children. J Pediatr Urol 2010; 6: 359.
2. Gogus C, Karamursel T, Tokatli Z et al: Long-term results of Anderson-Hynes pyeloplasty in 180 adults in the era of endourologic procedures. Urol Int 2004; 73: 11.
9. Hemal AK, Mukherjee S and Singh K: Laparoscopic pyeloplasty versus robotic pyeloplasty for ureteropelvic junction obstruction: a series of 60 cases performed by a single surgeon. Can J Urol 2010; 17: 5012.
18. Wagner S, Greco F, Inferrera A et al: Laparoscopic dismembered pyeloplasty: technique and results in 105 patients. World J Urol 2010; 28: 615.
3. Bansal P, Gupta A, Mongha R et al: Laparoscopic versus open pyeloplasty: comparison of two surgical approaches- a single centre experience of three years. J Minim Access Surg 2008; 4: 76. 4. Yanke BV, Lallas CD, Pagnani C et al: The minimally invasive treatment of ureteropelvic junction obstruction: a review of our experience during the last decade. J Urol 2008; 180: 1397. 5. El-Shazly MA, Moon DA and Eden CG: Laparoscopic pyeloplasty: status and review of literature. J Endourol 2007; 21: 673. 6. Singh I and Hemal AK: Robot-assisted pyeloplasty: review of the current literature, technique and outcome. Can J Urol 2010; 17: 5099. 7. Braga LH, Pace K, Demaria J et al: Systematic review and meta-analysis of robotic-assisted versus conventional laparoscopic pyeloplasty for patients with ureteropelvic junction obstruction: effect on operative time, length of hospital stay, postoperative complications, and success rate. Eur Urol 2009; 56: 848. 8. Passerotti CC, Passerotti AM, Dall’Oglio MF et al: Comparing the quality of the suture anastomosis and the learning curves associated with performing
10. Mufarrij PW, Woods M, Shah OD et al: Robotic dismembered pyeloplasty: a 6-year, multi-institutional experience. J Urol 2008; 180: 1391.
19. Sundaram CP, Grubb RL 3rd, Rehman J et al: Laparoscopic pyeloplasty for secondary ureteropelvic junction obstruction. J Urol 2003; 169: 2037.
11. Schwentner C, Pelzer A, Neururer R et al: Robotic Anderson-Hynes pyeloplasty: 5-year experience of one centre. BJU Int 2007; 100: 880.
20. Brito AH, Mitre AI and Srougi M: Laparoscopic pyeloplasty in secondary obstruction. J Endourol 2007; 21: 1481.
12. Palese MA, Munver R, Phillips CK et al: Robotassisted laparoscopic dismembered pyeloplasty. JSLS 2005; 9: 252.
21. Canes D, Desai MM, Haber GP et al: Is routine transposition of anterior crossing vessels during laparoscopic dismembered pyeloplasty necessary? J Endourol 2009; 23: 469.
13. Gupta NP, Nayyar R, Hemal AK et al: Outcome analysis of robotic pyeloplasty: a large singlecentre experience. BJU Int 2010; 105: 980. 14. Patel V: Robotic-assisted laparoscopic dismembered pyeloplasty. Urology 2005; 66: 45. 15. Singh O, Gupta SS, Hastir A et al: Laparoscopic dismembered pyeloplasty for ureteropelvic junction obstruction: experience with 142 cases in a high-volume center. J Endourol 2010; 24: 1431. 16. Abuanz S, Game X, Roche JB et al: Laparoscopic pyeloplasty: comparison between retroperitoneoscopic and transperitoneal approach. Urology 2010; 76: 877.
22. Boylu U, Oommen M, Lee BR et al: Ureteropelvic junction obstruction secondary to crossing vessels-to transpose or not? The robotic experience. J Urol 2009; 181: 1751. 23. Szydelko T, Kasprzak J, Apoznanski W et al: Comparison of dismembered and nondismembered Y-V laparoscopic pyeloplasty in patients with primary hydronephrosis. J Laparoendosc Adv Surg Tech A 2010; 20: 7. 24. Madi R, Roberts WW and Wolf JS Jr: Late failures after laparoscopic pyeloplasty. Urology 2008; 71: 677.
Abbreviations and Acronyms FSP ⫽ freedom from secondary procedure LP ⫽ laparoscopic pyeloplasty RP ⫽ robotic pyeloplasty UPJ ⫽ ureteropelvic junction UTI ⫽ urinary tract infection Submitted for publication June 2, 2011. Study received institutional review board approval. * Correspondence: Department of Urology, Indiana University Medical Center, 535 N. Barnhill Drive, Suite 420, Indianapolis, Indiana 46202 (telephone: 317-278-3098; FAX: 317-274-0174; e-mail: [email protected]). † Financial interest and/or other relationship with Onset Medical Corp. ‡ Financial interest and/or other relationship with Bard Medical, Cook Medical and Storz. § Financial interest and/or other relationship with Baxter and Intuitive Surgical.
Purpose: We compared laparoscopic and robotic pyeloplasty to identify factors associated with procedural efficacy. Materials and Methods: We conducted a retrospective multicenter trial incorporating 865 cases from 15 centers. We collected perioperative data including anatomical and procedural factors. Failure was defined subjectively as pain that was unchanged or worse per medical records after surgery. Radiographic failure was defined as unchanged or worsening drainage on renal scans or worsening hydronephrosis on computerized tomography. Bivariate analyses were performed on all outcomes and multivariate analysis was used to assess factors associated with decreased freedom from secondary procedures. Results: Of the cases 759 (274 laparoscopic pyeloplasties with a mean followup of 15 months and 465 robotic pyeloplasties with a mean followup of 11 months, p ⬍0.001) had sufficient data. Laparoscopic pyeloplasty, previous endopyelotomy and intraoperative crossing vessels were associated with decreased freedom from secondary procedures on bivariate analysis, with a 2-year freedom from secondary procedures of 87% for laparoscopic pyeloplasty vs 95% for robotic pyeloplasty, 81% vs 93% for patients with vs without previous endopyelotomy and 88% vs 95% for patients with vs without intraoperative crossing vessels, respectively. However, on multivariate analysis only previous endopyelotomy (HR 4.35) and intraoperative crossing vessels (HR 2.73) significantly impacted freedom from secondary procedures. Conclusions: Laparoscopic and robotic pyeloplasty are highly effective in treating ureteropelvic junction obstruction. There was no difference in their abilities to render the patient free from secondary procedures on multivariate analysis. Previous endopyelotomy and intraoperative crossing vessels reduced freedom from secondary procedures. Key Words: laparoscopy, robotics, reconstructive surgical procedures, survival analysis
522
www.jurology.com
0022-5347/12/1872-0522/0 THE JOURNAL OF UROLOGY® © 2012 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION
AND
RESEARCH, INC.
Vol. 187, 522-527, February 2012 Printed in U.S.A. DOI:10.1016/j.juro.2011.09.158
MULTI-INSTITUTIONAL ANALYSIS OF MINIMALLY INVASIVE PYELOPLASTY
OPEN surgical dismembered pyeloplasty has been the standard treatment for ureteropelvic junction obstruction and is successful in 90% to 95% of cases.1,2 Minimally invasive treatments for UPJ obstruction offer faster recovery and decreased postoperative pain.3 With concerns of inferior long-term success rates of endopyelotomy,4 minimally invasive pyeloplasty is increasingly used for the treatment of this condition, with success rates equivalent to those of open pyeloplasty.5–7 With the advent of robotic surgery, acquiring laparoscopic skills can be hastened, which has made minimally invasive pyeloplasty more widely available to many surgeons.8 Among those who are experienced in minimally invasive pyeloplasty, it is unknown whether robotic surgery improves outcomes. Factors such as degree of hydronephrosis, renal function and previous procedures influence the outcome of open pyeloplasty.2 However, because it is a newer treatment the factors that influence the outcomes of laparoscopic or robotic pyeloplasty have not been well studied. Since the success rates of these procedures are high, it is difficult to ascertain in single center series what may have contributed to treatment failure due to the limited number of failures. In this study several centers contributed to the Laparoscopic and Robotic Pyeloplasty Collaborative Group, and a larger database was created that allowed a more thorough assessment of the impact of preoperative and intraoperative parameters on the outcomes of LP and RP.
METHODS Centers with experience in minimally invasive pyeloplasty, all sites with Endourological Society recognized fellowships or active members of the Society of Urological Robotic Surgery, were asked to submit data to a retrospective database which included preoperative, intraoperative and followup data in this institutional review board approved study. Median (IQR) patient age was 35.0 (28.0) and 35.1 (22.0) years for RP and LP, respectively (p ⫽ 0.957). Preoperative computerized tomography or ultrasound was used to assess hydronephrosis and crossing vessels, and mercaptoacetyltriglycine renal scan with diuretic washout determined renal function and drainage. Median preoperative T1/2 was 30 (22) and 24 (19) minutes for RP and LP, respectively (p ⫽ 0.001). Median percent preoperative renal function was 43 (16) and 42 (19) for RP and LP, respectively (p ⫽ 0.425). Intraoperatively LP was differentiated from RP. Surgery performed using the traditional laparoscopic approach without robotic assistance was included in the LP group. The case was considered RP if the robot was used to perform the reconstruction. Postoperative data included drain duration, urine leak, complications and hospital stay. Recognizing that there may be differences in the management of drains and in the diagnosis of urine leak,
523
a urine leak was recorded if directly reported by the submitting center and the persistence of drain recorded if 7 days or more. Followup data included initial and final radiographic studies. Worsened hydronephrosis, or unchanged or worsened T1/2 was considered recurrent or persistent obstruction. Worsened or unchanged pain was counted as subjective treatment failure. A radiographic or subjective failure was considered treatment failure. Secondary interventions included prolonged stent placement, endopyelotomy or dilation, repeat pyeloplasty or nephrectomy. FSP was used to quantify patients who did not require these interventions. In terms of statistical analysis, bivariate analyses of preoperative and intraoperative factors were determined by Mann-Whitney U, Kruskal-Wallis or chi-square analysis. Kaplan-Meier analysis was used to determine FSP with log rank comparison of the presence or absence of each factor. Multivariate analysis of factors was conducted for FSP using Cox regression analysis of hazard ratios. Statistical analysis was conducted with PASW® statistics 17 with type 1 error set at 0.05.
RESULTS Treatment Modality Of 865 patients from 15 centers 759 were available for comparison, including 274 LP and 485 RP. Mean followup was 15 ⫾ 16 months for LP and 11 ⫾ 13 months for RP (p ⬍0.001). Preoperatively the LP group had higher rates of previous endopyelotomy, urinary drainage, UTI and pain. The RP group had significantly more stones and longer T1/2 (table 1). Median operative times, complications, hospital stay, urine leak and prolonged drain duration were similar for LP and RP, while nondismembered pyeloplasties were more common for LP (table 2). Rates of failure to improve obstruction, pain and need for secondary procedures were higher in the LP than in the RP group. FSP at 1 and 2 years was greater for RP than LP at 97% ⫾ 1.0% vs 92% ⫾ 2.0% and 95% ⫾ 1.8% vs 87% ⫾ 3.1%, respectively (p ⫽ 0.019, see figure). Table 1. Preoperative characteristics of robotic and laparoscopic pyeloplasty No. Robotic/ Total No. (%) Female pts Rt side cases Preop UTI Preop pain Preop crossing vessel Preop severe hydronephrosis Preop renal stone Previous abdominal procedure Previous endopyelotomy Previous pyeloplasty Preop stent Preop percutaneous nephrostomy
138/368 (49.7) 233/482 (48.3) 73/429 (17.0) 354/429 (82.3) 187/412 (45.4) 109/322 (33.9) 76/415 (18.3) 72/235 (30.6) 40/485 (8.2) 6/485 (2.2) 179/470 (38.1) 15/470 (3.2)
No. Laparoscopic/ Total No. (%) p Value 154/274 (56.2) 133/273 (48.7) 50/200 (25.0) 247/272 (90.8) 80/208 (38.5) 81/216 (37.5) 31/263 (11.8) 43/186 (23.1) 38/274 (13.9) 6/274 (1.4) 143/250 (57.2) 12/250 (4.8)
0.104 0.921 0.019 0.002 0.100 0.385 0.023 0.085 0.014 0.312 ⬍0.001 ⬍0.001
524
MULTI-INSTITUTIONAL ANALYSIS OF MINIMALLY INVASIVE PYELOPLASTY
Table 2. Comparison of intraoperative and postoperative outcomes of robotic and laparoscopic pyeloplasty
Median operative mins (IQR) Median ml blood loss (IQR) No./total No. nondismembered pyeloplasties (%) No. intraop complications (%) No. postop complications (%) No./total No. urine leaks (%) No./total No. prolonged drain 7 days or more (%) No./total No. stent related problems (%) Median % postop function (IQR) Median mins postop T1/2 (IQR) No./total No. obstruction not improved (%) No./total No. pain not improved (%) No./total No. secondary intervention (%)
Table 3. Effect of previous endopyelotomy on outcome of minimally invasive pyeloplasty
Robotic
Laparoscopic
p Value
204 (85) 30 (30) 31/427 (7.3)
218 (124) 50 (75) 34/273 (12.5)
0.123 0.001 0.021
9 (1.9) 26 (5.4) 9/485 (1.9) 29/258 (11.2)
6 22 9/274 13/140
(2.2) (8.0) (3.3) (9.3)
0.670 0.275 0.303 0.554
7/422 (1.7)
9/242 (3.7)
0.096
45 (13) 10 (7) 15/334 (4.5)
45 (16) 10 (9) 24/221 (10.9)
0.397 0.087 0.004
15/328 (4.6) 12/375 (3.2)
16/175 (9.1) 23/242 (9.5)
0.042 0.001
Among primary dismembered pyeloplasties (207 LP and 354 RP) the laparoscopic group had higher rates of postoperative obstruction (11% vs 3.8%, p ⫽ 0.004), pain (5.3% vs 1.7%, p ⫽ 0.016) and secondary procedures (7.7% vs 1.8%, p ⫽ 0.002). FSP at 1 and 2 years for LP vs RP was 94% ⫾ 2.0% and 89% ⫾ 3.4% vs 99% ⫾ 1.0% and 96% ⫾ 1.8%, respectively (p ⫽ 0.018). Previous Procedures A total of 78 patients underwent endopyelotomy before pyeloplasty, which did not affect urine leak or drain duration postoperatively, but did increase the rate of subjective failure and secondary procedures (p ⬎0.05, table 3). At 1 and 2 years FSP was lower for patients who previously underwent endopyelotomy at 90% ⫾ 4.5% and 81% ⫾ 7.3% vs 97% ⫾ 1.0% and 93% ⫾ 1.8%, respectively (p ⬍0.001, see figure).
No Previous Previous p Endopyelotomy Endopyelotomy Value No. pts Median % preop function (IQR) Median mins preop T1/2 (IQR) No./total No. failed obstruction (%) No./total No. failed pain (%) No./total No. secondary procedure (%) No./total No. drain 7 days or more (%) No./total No. urine leak (%) Median % postop function (IQR) Median mins postop T1/2 (IQR) Median operative mins (IQR)
669 43 (16) 29 (16) 30/466 (6.4) 20/632 (3.2) 25/548 (4.6) 40/356 (11.2) 13/669 (1.9) 45 (13) 10 (8) 209 (100)
78 39 (16) 24 (33) 6/54 (11.1) 7/74 (9.5) 10/69 (14.5) 2/38 (5.3) 2/78 (2.6) 44 (15) 9 (10) 205 (90)
0.112 0.435 0.200 0.008 0.001 0.257 0.556 0.307 0.876 0.863
Previous pyeloplasty was performed in 12 patients and the median operative time was longer for those with prior pyeloplasty (272.0 vs 208.0 minutes, p ⫽ 0.022). In these 12 patients pain failed to improve in 1 patient and obstruction failed to improve in another, and secondary procedures were performed in these 2 patients. Preoperative Urinary Drainage Ureteral stents were placed a day before pyeloplasty (most often for pain) in 322 patients vs 353 without placement, resulting in similar subjective failure rates (7 of 322, 2.2% vs 16 of 353, 4.5%; p ⫽ 0.092), radiographic failure rates (16 of 251, 6.4% vs 16 of 248, 6.5%; p ⫽ 0.972) and secondary procedure rates (9 of 270, 3.3% vs 19 of 280, 6.8%; p ⫽ 0.066). While the urine leak rate was lower in the preoperative stent group (0.9% vs 3.4%, p ⫽ 0.030), a higher proportion had a drain for 7 days or more in the preoperative stent group (14.5% vs 5.6%, p ⫽ 0.004). Crossing Vessels Based on preoperative computerized tomography 267 patients had crossing vessels and 354 did not, while
A, univariate analysis of FSP after laparoscopic and robotic pyeloplasty. B, univariate analysis of impact of previous endopyelotomy on FSP after minimally invasive pyeloplasty. C, univariate analysis of impact of intraoperative crossing vessels on FSP after minimally invasive pyeloplasty. Intraop, intraoperative.
MULTI-INSTITUTIONAL ANALYSIS OF MINIMALLY INVASIVE PYELOPLASTY
226 with crossing vessels and 259 without crossing vessels were noted intraoperatively. An intraoperative crossing vessel resulted in a higher subjective failure rate (5.8% vs 2.2%, p ⫽ 0.048), similar radiological failure rate (6.9% vs 4.2%, p ⫽ 0.284) and a higher rate of secondary procedures (8.6% vs 2.2%, p ⫽ 0.013). FSP at 1 and 2 years was significantly lower in cases with intraoperative crossing vessels at 95% ⫾ 1.7% and 88% ⫾ 3.4% vs 97% ⫾ 1.7% and 95% ⫾ 2.1%, respectively (p ⫽ 0.028, see figure). Nondismembered Pyeloplasty Nondismembered pyeloplasty was performed in 65 cases (18 Fenger plasties, 38 Y-V plasties, 1 spiral flap, 3 other flaps and 5 unknown) while 635 were dismembered. Median operative time was 240 minutes (IQR 137) for nondismembered and 208 (IQR 100) for dismembered pyeloplasties (p ⫽ 0.081). Pain did not improve in 11% (7 of 65) nondismembered vs 3.3% (21 of 635) dismembered (p ⫽ 0.003), radiological failure occurred in 8.7% (4 of 46) vs 6.9% (33 of 480, p ⫽ 0.645) and secondary procedures occurred in 11% (6 of 57) vs 5% (27 of 493, p ⫽ 0.100), respectively. FSP was similar between the 2 groups (p ⫽ 0.083). Urine leak and prolonged drain were similar between nondismembered and dismembered pyeloplasties at 2 of 65 (3.1%) and 5 of 35 (14%) vs 13 of 635 (2.0%) and 37 of 363 (10.2%), respectively (p ⬎0.4). Number of Prior Cases The median number of cases performed before LP and RP was 23.5 (IQR 43) and 35 (IQR 45), respectively (p ⬍0.001). FSP for pyeloplasties performed with 30 or less prior cases at 1 and 2 years was 93.9% ⫾ 1.6% and 89.4% ⫾ 2.4% vs 97.2% ⫾ 1.0% and 94.7% ⫾ 2.1% for those with more than 30 prior cases (p ⫽ 0.261). There was no difference in radiological failure (7.3% vs 6.6%) or subjective failure (6.8% vs 4.1%, p ⬎0.10). Considering only cases that were performed after more than 30 prior cases, FSP for LP and RP at 1 year was 93.2% ⫾ 3.0% vs 100%, and at 2 years was 93.2% ⫾ 3.0% vs 95.3% ⫾ 3.2% (p ⫽ 0.055). Multivariate Analysis Treatment modality, intraoperative crossing vessels and previous endopyelotomy significantly affected FSP on bivariate analysis (table 4). Multivariate analysis revealed previous endopyelotomy (hazard ratio 4.35, p ⫽ 0.001) and intraoperative crossing vessels (hazard ratio 2.73, p ⫽ 0.046) to be significant predictors of secondary procedures.
DISCUSSION In this study we present the largest comparative series of LP and RP reported to our knowledge, with success rates greater than 90%. On bivariate anal-
525
Table 4. Multivariate analysis of factors influencing the rate of secondary procedure following minimally invasive pyeloplasty Kaplan-Meier Log Rank Multivariate HR Multivariate p Value (95% CI) p Value Purely LP Previous endopyelotomy Intraop crossing vessel Nondismembered pyeloplasty Severe hydronephrosis Preop stone Preop stent Preop function less than 40% Preop T1/2 greater than 35 mins Preop UTI
0.019 ⬍0.001 0.028 0.083 0.099 0.997 0.176 0.486 0.357 0.061
1.99 (0.852–4.67) 1.85 (1.77–11.0) 2.73 (1.02–7.34)
0.112 0.001 0.044
ysis RP had some advantages compared to LP in terms of persistence or recurrence of radiographic obstruction (4.5% vs 10.9%) and symptoms (4.6% vs 9.1%). The need for secondary procedures was also lower in the RP group (3.2% vs 9.5%). RP and LP had similar complication rates. Followup was longer for LP but FSP was greater for robotic than for laparoscopic pyeloplasty (97% vs 94% at 1 year). However, there were significant differences between the 2 groups that likely confounded the bivariate analysis. Thus, FSP was analyzed in a multivariate analysis including significant preoperative factors (previous endopyelotomy, crossing vessels and treatment modality). In this analysis previous endopyelotomy and intraoperative crossing vessels were significantly associated with decreased FSP while the use of LP or RP was not. Recent series have compared LP and RP. In a meta-analysis 8 articles compared LP and RP, demonstrating similar success rates of more than 95%.7 There was a 10-minute reduction in operating time with RP. However, the analysis included only series that had RP and LP, which often meant a small number of patients (median sample size per series was 13 patients each, with only 1 series having more than 30 patients). In a separate series 30 LPs were compared with 30 RPs, of which only 1 LP required a secondary procedure, suggesting equivalent results.9 Operative time was faster with RP than with LP (98 vs 145 minutes). Additionally, Yanke et al reviewed 116 LP and 29 RP cases with success rates of 88.8% and 100%, respectively, at a mean of 20 months.4 Predominantly the inference about the success rates and operative times for RP and LP is made from observational studies of 1 modality. In a review of 25 published series comprising 740 RP cases, the mean operative time was 194 minutes and the mean success rate was greater than 90%. In 6 of the largest studies the success rate ranged from 95% to 100%, with operative times between 108 and 225
526
MULTI-INSTITUTIONAL ANALYSIS OF MINIMALLY INVASIVE PYELOPLASTY
minutes.4,10 –14 In 5 recent laparoscopic series the success rate ranged from 85% to 100% with mean operative times between 131 and 244 minutes.3,15–18 An analysis of treatment modalities is difficult without a randomized, prospective comparison. Overall it is acceptable to report that both treatments are effective for primary UPJ obstruction. However, most centers are now performing a substantial portion of pyeloplasties with robotic assistance. In centers with highly experienced laparoscopic surgeons, LP is a successful treatment option, especially when access to the robot is limited by competition for resources. In addition, the large number of pyeloplasties evaluated in this series enabled us to evaluate perioperative factors that may influence the outcome of minimally invasive pyeloplasty. Previous endopyelotomy was the strongest predictor of a secondary procedure on multivariate analysis, although data on the type of endopyelotomy performed were not available. We believe that endopyelotomy must be reserved for selected patients in whom the laparoscopic (with or without robotic assistance) approach would be difficult or contraindicated. Other series have also shown that the efficacy of minimally invasive pyeloplasty is decreased after a failed endopyelotomy, although the degree of this impact is variable. In a review of 36 patients who underwent a previous failed UPJ procedure (33 of 36 were endoscopic), the overall success rate of LP was 83%.19 In another series of 13 patients a 92% success rate was observed.20 Interestingly a preoperative ureteral stent in a setting before the actual operative procedure, which anecdotally increases periureteral inflammation and makes pyeloplasty more difficult, did not impact efficacy. While the rate of urine leak did not differ, prolonged drainage more frequently occurred with a preoperative stent than without (15% vs 5.6%). This suggests no difference in the anatomical results of these cases, but knowledge of a preoperative stent may cause surgeons to be more cautious with postoperative management of these cases. The presence of an intraoperative crossing vessel decreased symptomatic improvement and FSP after pyeloplasty. A crossing vessel has been demonstrated to decrease the efficacy of endopyelotomy but has not been reported to affect the outcome of minimally invasive pyeloplasty. In 2 series of 31 and 48 patients with a lower pole crossing vessel, success after minimally invasive pyeloplasty was observed in all patients.21,22 In a separate series of 129 LPs the presence of crossing vessels improved the success rate (hazard ratio 0.27, 95% CI 0.08 – 0.87), which was theorized to be related to a more focal stricture.4
Nondismembered pyeloplasty resulted in decreased symptomatic improvement but a similar rate of secondary procedures. In a recent series comparing dismembered pyeloplasty (50 patients) to Y-V plasty (36 patients) demonstrated success rate was 91% in both groups.23 In selected cases the success of nondismembered pyeloplasty is nearly equivalent to that of dismembered pyeloplasty. However, longer operative times in our study were noted for nondismembered cases, suggesting these were more difficult. Finally, cost must be considered when robotic assistance is used. In addition to the cost of the robot and its maintenance, disposable costs were higher for RP than LP. Although we do not have data on disposable costs at each center, the cost difference can be illustrated by estimating our own costs. Assuming in each case that, in addition to the disposable drapes, hot shears, PK dissector, large needle driver, Potts scissors and fine tissue forceps were used, a cost of $1,280 of disposables would apply. For LP, assuming the use of ultrasonic shears, disposable scissor tips and trocars, the disposable cost per case would be $750. Additional costs of 12 mm trocars, suction irrigator and stents would be incurred for both procedures. Since many centers now use robotic assistance, it will be difficult to maintain the skills required to perform a comparable anastomosis, thereby justifying the higher cost. There were several limitations to the study and some critical points merit discussion. The most significant limitation was the retrospective design. As a multi-institutional study there were differences in surgical technique, followup, and definitions of symptom improvement and urine leak which may explain why RP had an advantage on univariate but not on multivariate analysis. Although the same trend was observed on multivariate analysis, a larger data set would be required to substantiate a benefit. In addition, more LP cases were nondismembered, the reason for which cannot be elucidated in a retrospective fashion. Furthermore, the centers included in the study were high volume centers. Assessment of clinical symptoms was limited as these data were not obtained through patient reported surveys but from medical records. Another limitation was the bias created by the implementation of the robot for pyeloplasties after some of the surgeons gained experience with LP. This could result in comparison of laparoscopic pyeloplasty at an earlier point in the learning curve than the robotic cases. We corrected for this in a subgroup analysis that included only pyeloplasties performed after 30 prior cases. A slightly higher FSP was seen for RP than LP, although it was not statistically significant.
MULTI-INSTITUTIONAL ANALYSIS OF MINIMALLY INVASIVE PYELOPLASTY
An additional limitation is that the mean followup for LP and RP was 15 and 11 months respectively. Late failures were reported. For instance, in a series of 65 patients Madi et al identified 3 patients in whom recurrence developed 2 or more years after surgery compared with 7 failures within the first year.24 Finally, time to radiographic failure or symptomatic failure could not be assessed as these are dependent on reporting symptoms or obtaining imaging. Thus, Kaplan-Meier analysis could not be performed for these outcomes.
527
CONCLUSIONS In this study we present a large, multi-institutional series of minimally invasive pyeloplasties. LP and RP demonstrate a high rate of success for the treatment of UPJ obstruction. Only previous endopyelotomy and crossing vessels were significant predictors of reduced FSP and success rates on multivariate analysis. Further comparison of treatment modalities is best conducted in a randomized, controlled setting.
REFERENCES 1. O’Reilly PH, Brooman PJ, Mak S et al: The longterm results of Anderson-Hynes pyeloplasty. BJU Int 2001; 87: 287.
open, freehand, and robotic-assisted laparoscopic pyeloplasty in a swine animal model. J Am Coll Surg 2009; 208: 576.
17. Szavay PO, Luithle T, Seitz G et al: Functional outcome after laparoscopic dismembered pyeloplasty in children. J Pediatr Urol 2010; 6: 359.
2. Gogus C, Karamursel T, Tokatli Z et al: Long-term results of Anderson-Hynes pyeloplasty in 180 adults in the era of endourologic procedures. Urol Int 2004; 73: 11.
9. Hemal AK, Mukherjee S and Singh K: Laparoscopic pyeloplasty versus robotic pyeloplasty for ureteropelvic junction obstruction: a series of 60 cases performed by a single surgeon. Can J Urol 2010; 17: 5012.
18. Wagner S, Greco F, Inferrera A et al: Laparoscopic dismembered pyeloplasty: technique and results in 105 patients. World J Urol 2010; 28: 615.
3. Bansal P, Gupta A, Mongha R et al: Laparoscopic versus open pyeloplasty: comparison of two surgical approaches- a single centre experience of three years. J Minim Access Surg 2008; 4: 76. 4. Yanke BV, Lallas CD, Pagnani C et al: The minimally invasive treatment of ureteropelvic junction obstruction: a review of our experience during the last decade. J Urol 2008; 180: 1397. 5. El-Shazly MA, Moon DA and Eden CG: Laparoscopic pyeloplasty: status and review of literature. J Endourol 2007; 21: 673. 6. Singh I and Hemal AK: Robot-assisted pyeloplasty: review of the current literature, technique and outcome. Can J Urol 2010; 17: 5099. 7. Braga LH, Pace K, Demaria J et al: Systematic review and meta-analysis of robotic-assisted versus conventional laparoscopic pyeloplasty for patients with ureteropelvic junction obstruction: effect on operative time, length of hospital stay, postoperative complications, and success rate. Eur Urol 2009; 56: 848. 8. Passerotti CC, Passerotti AM, Dall’Oglio MF et al: Comparing the quality of the suture anastomosis and the learning curves associated with performing
10. Mufarrij PW, Woods M, Shah OD et al: Robotic dismembered pyeloplasty: a 6-year, multi-institutional experience. J Urol 2008; 180: 1391.
19. Sundaram CP, Grubb RL 3rd, Rehman J et al: Laparoscopic pyeloplasty for secondary ureteropelvic junction obstruction. J Urol 2003; 169: 2037.
11. Schwentner C, Pelzer A, Neururer R et al: Robotic Anderson-Hynes pyeloplasty: 5-year experience of one centre. BJU Int 2007; 100: 880.
20. Brito AH, Mitre AI and Srougi M: Laparoscopic pyeloplasty in secondary obstruction. J Endourol 2007; 21: 1481.
12. Palese MA, Munver R, Phillips CK et al: Robotassisted laparoscopic dismembered pyeloplasty. JSLS 2005; 9: 252.
21. Canes D, Desai MM, Haber GP et al: Is routine transposition of anterior crossing vessels during laparoscopic dismembered pyeloplasty necessary? J Endourol 2009; 23: 469.
13. Gupta NP, Nayyar R, Hemal AK et al: Outcome analysis of robotic pyeloplasty: a large singlecentre experience. BJU Int 2010; 105: 980. 14. Patel V: Robotic-assisted laparoscopic dismembered pyeloplasty. Urology 2005; 66: 45. 15. Singh O, Gupta SS, Hastir A et al: Laparoscopic dismembered pyeloplasty for ureteropelvic junction obstruction: experience with 142 cases in a high-volume center. J Endourol 2010; 24: 1431. 16. Abuanz S, Game X, Roche JB et al: Laparoscopic pyeloplasty: comparison between retroperitoneoscopic and transperitoneal approach. Urology 2010; 76: 877.
22. Boylu U, Oommen M, Lee BR et al: Ureteropelvic junction obstruction secondary to crossing vessels-to transpose or not? The robotic experience. J Urol 2009; 181: 1751. 23. Szydelko T, Kasprzak J, Apoznanski W et al: Comparison of dismembered and nondismembered Y-V laparoscopic pyeloplasty in patients with primary hydronephrosis. J Laparoendosc Adv Surg Tech A 2010; 20: 7. 24. Madi R, Roberts WW and Wolf JS Jr: Late failures after laparoscopic pyeloplasty. Urology 2008; 71: 677.