Laparoscopic pyeloplasty: The retroperitoneal approach is suitable for establishing a de novo practiceRJ Bryant, E Craig, N Oakley
Department of Urological Surgery, Royal Hallamshire Hospital, Glossop Road Sheffield, S10 2JF, United Kingdom
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0022-3859.41434
Source of Support: None, Conflict of Interest: None
Background: Laparoscopic dismembered pyeloplasty has become the "gold-standard" procedure for pelviureteric junction (PUJ) obstruction but consists of a steep learning curve especially via the retroperitoneal route. Aims: To examine the feasibility and safety of introducing this technique via the retroperitoneal approach to a laparoscopic naοve center. Settings and Design: A retrospective data analysis of a single surgeon's (NEO) series from a large UK teaching hospital. Materials and Methods: The notes and imaging of all patients who underwent pyeloplasty for PUJ obstruction by NEO during a five-year period were reviewed. Statistical Analysis: Parametric and nonparametric data are presented analyzed with Excel XP (Microsoft, Redmond, WA, USA). Results: Our series consists of 67 patients. Three ports were used in 47/57 (82%), and the antegrade technique for stent insertion was utilized in 41/67 (61%). Median time to drink, eat, and mobilize was one day (range one to two), and to discharge three days (range three to four). Two patients required conversion to an open procedure, and two developed intraoperative complications. Postoperative complications at 30 (three major, seven minor) and 90 days (three major, three minor) are presented. With median follow-up of 15 months 61/65 (94%) patients were unobstructed, and 57/63 (90%) of patients were pain-free. Two patients re-obstructed requiring further surgery. Conclusions: Analysis of our series of patients illustrates that adopting a policy of retroperitoneal laparoscopic pyeloplasty for primary PUJ obstruction is feasible without compromising patient safety or functional results. There is no need to breach the peritoneum to facilitate the learning curve.
Keywords: Clinical outcome, primary laparoscopic pyeloplasty, PUJ obstruction, retroperitoneal approach
The "Anderson-Hynes" dismembered pyeloplasty was first described for the treatment of an obstructed retrocaval ureter in 1949.  This procedure has stood the test of time as the definitive treatment for pelviureteric junction (PUJ) obstruction and is reported to have a >90% chance of successfully de-obstructing this condition.  The laparoscopic approach to pyeloplasty for PUJ obstruction was first reported in 1993. , While the published series have allowed clinicians to gain insight into the outcome of laparoscopic pyeloplasty, ,,,, the uptake of this procedure has been limited to relatively few high-volume centers. The biggest obstacle facing urologic surgeons wishing to undertake this procedure is the difficulty of a steep learning curve compounded by a lack of laparoscopic experience. This raises the question of whether this procedure should only be attempted once one has become adept at performing a simpler and more common procedure such as laparoscopic nephrectomy.  Faced with limited laparoscopic experience and training opportunities, our center introduced the procedure of retroperitoneal laparoscopic pyeloplasty simultaneously with that of nephrectomy.  We herein present the outcome of our experience in the introduction of retroperitoneal laparoscopic pyeloplasty for primary PUJ obstruction.
The Royal Hallamshire Hospital is a tertiary referral center in the North Trent region of the UK and it provides a comprehensive range of endourologic procedures. On appointment of a consultant endourologic surgeon (NEO) in 2000, the remit was to introduce the laparoscopic approach to both nephrectomy and pyeloplasty. From the date of appointment, we replaced open pyeloplasty with laparoscopic as our default technique for PUJ obstruction other than for solitary kidneys. We have retrospectively reviewed the records of all patients who underwent a pyeloplasty for primary PUJ obstruction by NEO in the five-year period to June 2005. Data were retrieved from both physical medical records and electronic records kept on a secure hospital computer server. The data were recorded anonymously and analyzed using Excel XP (Microsoft, Redmond, WA, USA).
The project was registered with the hospital audit committee but local ethical committee approval was not required.
In total 72 patients underwent a pyeloplasty under NEO via a laparoscopic approach. In three patients, NEO opted for an elective transperitoneal procedure because of mentoring others and two patients opted for a retroperitoneal fengerplasty due to a small renal pelvis on preoperative imaging. The remaining 67 patients underwent a retroperitoneal Anderson-Hynes dismembered pyeloplasty. Patient records were complete in the majority of cases; however, some of the patients in our series were originally referred from other hospitals within or outside the North Trent region, therefore a considerable proportion of postoperative follow-up was performed at other hospitals and in some of these cases the patient records were unavailable for our analysis.
Surgical procedure for laparoscopic pyeloplasty
Our default technique adopts a retroperitoneal approach using a proprietary dilating balloon (Extraview®, COVIDIEN, Mansfield, MA.USA) introduced via a midaxillary subcostal 12-mm incision. This technique creates a space prior to the introduction of a primary port and two secondary 12-mm ports subcostally in the anterior- and posterior-axillary line. The kidney is approached posteriorly and the renal pelvis is exposed and dissected off any crossing vessel, dismembered, and the upper ureter divided and spatulated. Initially it was our practice to insert the stent at the start of the procedure after a cystoscopy and retrograde study, whilst our current practice is to insert the "double J" stent antegrade via a percutaneous puncture. The anastomosis is performed with a 3/0 vicryl continuous suture and any crossing vessel is left anterior to the anastomosis and not transposed. A 16-French drain is left via a port site, and postoperatively patients can drink, eat, and mobilize as they feel able. The catheter is removed after 24 h and the drain after 48 h provided the total drain leakage of urine is <100 ml. Patients are discharged home when they are eating, drinking, and are mobile, provided their pain is controlled with oral analgesia. The stent is removed six weeks later and a MAG3 renogram is performed for follow-up at three, nine, and 18 months postoperatively.
Our series of patients are outlined in [Table 1]. The majority of patients had been referred as outpatients from other urological units within and outside the region. While 12 patients had presented as an emergency at our institution, it is possible that a number of patients referred to us may also have initially presented as emergencies.
Five patients had a nephrostomy tube in situ following episodes of pyonephrosis, and two patients had ipsilateral ureteric stents at the time of laparoscopic pyeloplasty. One patient had undergone a previous percutaneous nephrolithotomy and "EndoBRST" treatment for a right-sided PUJ obstruction and concomitant right renal stone burden, and this had been complicated by subsequent pyonephrosis requiring nephrostomy drainage. One patient was rendered asymptomatic by a previous "EndoBRST" procedure but demonstrated dropping split function on serial renography. Nine patients (13%) had undergone previous unrelated abdominal surgery.
The indications for surgery are shown in [Table 1] and include one or more of loin pain, urinary tract infection, and a poor or deteriorating renographic split renal function. The vast majority (92%) were offered a laparoscopic procedure to relieve pain, with urinary sepsis (17%) being the second most common symptom. This is in line with our clinical practice with our indications for surgery being symptoms in the presence of demonstrated obstruction.
The operative factors are summarized in [Table 2] with theater time being defined as the time from induction of anesthetic to the time of port closure. Although the mean operative time was 153 min, analyzing the group as sequential cohorts revealed that the mean operative time for the first 14 patients was 167 min and for the last 15 patients was 140 min.
Most procedures could be performed with three ports; however, in several cases a further 5-mm port was required to aid retraction. In two cases, a tear in the peritoneum led to the approach being altered to transperitoneal without incident. With increasing experience, the ureteric stent was passed in an antegrade direction prior to dismembering the PUJ. Within 48 h of surgery, the majority of patients were able to eat a full diet, mobilize, and have their urinary catheter removed. The median postoperative time to discharge was three days, and the shortest time to discharge was two days which was achieved for 10 patients (18%). A small number of patients were discharged home after longer periods of convalescence and these patients were generally those with intra- or postoperative complications.
Our surgical complications are classified as either intraoperative or postoperative, and the latter are tabulated as major or minor complications which occurred within either 30 or 90 days following surgery [Table 3]. The two conversions to open procedures in our series occurred due to a failure to progress with the anastomosis in one patient, and fibrosis causing failure of progression of dissection in another patient. Of the two patients with intraoperative complications, one had a pneumothorax that was identified postoperatively and settled with conservative management. In the other patient, it was not possible to pass a stent in either a retrograde or antegrade manner, and this was further complicated by the discovery of a retained needle at the end of the procedure. A subsequent re-exploration was performed, the needle was successfully removed, and an open pyeloplasty was performed. After three years of follow-up, this patient is well with radiological improvement in both split renal function and drainage. During the first 30 days following surgery, three patients (5%) had developed major complications from surgery, including the patient who required re-exploration who subsequently developed a pulmonary embolism and two other patients had a prolonged drain urinary leakage which was successfully managed conservatively by re-catheterization. Seven other patients developed minor 30-day complications including wound infections, poor pain control, a perirenal hematoma managed conservatively, hematuria requiring re-admission, palpitations, and phlebitis. Three other patients (5%) had major complications within the first 90 days following surgery, including two patients with recurrent PUJ obstruction and pain, and a third patient who developed pyonephrosis after removal of the ureteric stent six weeks postoperatively and subsequently required temporary nephrostomy insertion. A further three patients had minor complications during this period including recurrent pain, stent symptoms requiring hospital admission, and pyelonephritis. Over 30% (16/53) of patients suffered from stent symptoms including frequency of micturition, loin pain, and hematuria. Three patients in our series had renal calculi, and total clearance of this stone burden during laparoscopic pyeloplasty was possible in two of these patients, with the other patient requiring postoperative extracorporeal shock wave lithotripsy followed by percutaneous nephrolithotomy (PCNL).
The mean follow-up in our series is 16 months; however, a number of patients did not attend either or both of their follow-up consultations or renogram appointments following the removal of their stent at six weeks. The functional outcome of the procedure was analyzed objectively using postoperative MAG3 renographic studies and subjectively by recording documented pain relief [Table 4]. Radiologic de-obstruction was seen in 61/65 patients (94%) and 57/63 patients (90%) were rendered pain-free. A total of 4/63 (6%) patients had further surgery during the follow-up period, including two patients who had re-obstructed, one of whom underwent an open nephrectomy while another patient underwent a successful re-do laparoscopic (flap approach) pyeloplasty, and two patients who despite appearing un-obstructed developed renal stones. One of these underwent a successful PCNL while the other underwent a successful pyelolithotomy.
The results in this series illustrate the experience of introducing the complex laparoscopic reconstructive procedure of dismembered pyeloplasty in a unit without prior laparoscopic experience. In view of the lack of opportunities for mentoring and assisting, prior training was limited to experience gained when performing cholecystectomies supplemented by attending advanced laparoscopic suturing courses, wet lab courses, and observing centers of excellence in the UK and US.
The data obtained from an analysis of our series of patients can help address a number of questions regarding laparoscopic treatment of PUJ obstruction. First, there is the question of whether patients' safety and outcomes are going to be jeopardized during the learning process. The long-term success rate of open pyeloplasty is reportedly greater than 90%,  and a number of published series of laparoscopic pyeloplasty suggest that the results can be at least as good as open surgery with consistently high success rates of 87-99%. ,,,, The analysis of our data illustrate that despite our inexperience, our success rate during the initiation and development of our laparoscopic service are comparable with other published series.
Our intra- and postoperative complications were comparable with most other published series which report postoperative complication rates of 2-13%. ,,, Moreover, the conversion rate of 3% in our series is comparable to the rate described (0-5%) in other published series of primary laparoscopic pyeloplasty. ,,,,
It has been suggested that the retroperitoneal approach is more difficult to learn than the transperitoneal approach, and that this could therefore result in less favorable outcomes. , While the retroperitoneal approach is technically challenging due to the small space and lack of anatomical landmarks,  we proceeded with this approach from the outset to avoid intraperitoneal complications such as a urinary leak or bowel injury. With the exception of one procedures which was converted to open due to unfamiliarity with the retroperitoneal vision, our data would concur with recent series which suggest that adopting a retroperitoneal approach from the outset does not lead to a drop in success rate. 
This study helps to shed some light on the length of the learning curve by analysis of the operative time. The mean overall time for this procedure in our series of patients where these details are available (153 min, 44 patients) is in keeping with that reported in the literature. The operative time in the largest series of laparoscopic pyeloplasty published to date was 140 (range 58-290) min,  while other contemporary series quote a mean operating time of 164-252 min. ,,, The mean operative time for our first 14 patients was 167 min, for the next 15 patients was 154 min, and for the final 15 patients was 140 min. A large proportion of this improvement is likely to be due to the avoidance of the cystoscopic stent insertion, and it would appear that the operative time reaches a nadir between 20 and 40 cases, which is similar to our experience with nephrectomy. 
The treatment of a crossing vessel at the PUJ is a controversial subject, and it has been postulated that the presence of a crossing vessel could be a major factor in the pathogenesis of PUJ obstruction  and is an indicator of poor prognosis following endoscopic incision of the renal pelvis. , With the anterior approach to the PUJ, it is sensible to transpose the anastomosis anterior to crossing vessels to facilitate the suturing  ; however, there is little convincing evidence in the literature to suggest that this maneuver is mandatory from a functional perspective.  In our adoption of the retroperitoneal approach, a conscious decision was made not to transpose the anastomosis anterior to any aberrant crossing vessels. The success rates reported in our series of patients treated without this transposition in terms of both long-term drainage and symptom relief support this argument.
Taken together with other published series, our results support the belief that laparoscopic pyeloplasty is at least as effective and safe as open surgery in the management of PUJ obstruction. As with all retrospective studies, our data are subject to the limitations of the accuracy of data recording and collection as demonstrated by incompleteness of some of the follow-up records. It does, however, have the benefit of representing a total series of sequential patients subject to the same surgical indications and surgical conditions.
Although the gold standard evidence would be a randomized controlled trial, it is difficult to envisage recruitment of sufficient numbers of patients to such a trial, particularly as there appears to be such little difference in functional outcomes between the two approaches, making it difficult to justify the morbidity of an open approach. , To our knowledge, there have been no published results of randomized controlled trials of open versus laparoscopic pyeloplasty and a previous such randomized controlled trial was aborted due to patient reluctance to be randomized. 
While our data support the role of laparoscopic pyeloplasty, this has to be viewed in the context of recently published training guidelines from the British Association of Urological Surgeons which set out a step-by-step program of education, practice, observation, and mentoring for both upper tract and pelvic laparoscopy.  We do not advocate ignoring these guidelines which advise attaining experience in laparoscopic nephrectomy before proceeding to pyeloplasty, because our series is representative of this technique in its infancy at a time when very few mentors were available and specific UK teaching courses did not exist. Despite this, the main surgeon (NEO) had a background experience in endourology and laparoscopy and attained vital skills during advanced laparoscopic suturing training, wet labs, and overseas fellowships.
In conclusion, our data suggest that despite relative inexperience a satisfactory level of performance of retroperitoneoscopic pyeloplasty, reflected by low short-term morbidity rates, acceptable operative times, and high successful functional outcomes of this procedure, can be safely attained following a period of focused specialist skills training. The advantages of not transgressing the peritoneum can be maintained and a dismembered Anderson-Hynes anastomosis performed posterior to any crossing vessels without adverse functional sequelae.
The authors thank the urology consultants in the North Trent Region, and especially those in the Royal Hallamshire Hospital (Messers Anderson, Catto, Chapple, Cutinha, Hall, Hamdy, Hastie, Inman, Natarajan, Rosario, Smith, and Tophill), for their support and encouragement during the establishment of the laparoscopic urological practice.
[Table 1], [Table 2], [Table 3], [Table 4]