Endoluminal ultrasonography before retrograde endopyelotomy: can the results match laparoscopic pyeloplasty?


F.X. Keeley, Bristol Urological Institute, Southmead Hospital, Westbury on Trym, Bristol BS10 5NB, UK.
e-mail: fimio@blueyonder.co.uk


In this paper, the authors assess whether endoluminal ultrasonography is helpful when carried out before retrograde endopyelotomy, and ask whether it can help to produce results comparable to laparoscopic pyeloplasty. They found that endopyelotomy was not as good in the presence of crossing vessels, despite using endoluminal ultrasonography, and they propose that laparoscopic pyeloplasty should be used in this situation.


To present the results of endopyelotomy using endoluminal ultrasonography (EUS) to identify crossing vessels, as the success rates of endopyelotomy are generally lower than pyeloplasty, especially in patients with crossing vessels.


Forty-one consecutive patients who underwent EUS before a planned retrograde endopyelotomy were analysed retrospectively. EUS was used to direct the endopyelotomy incision for patients with crossing vessels. Treatment was considered successful if the patient was asymptomatic and unobstructed or improved on renography. The results were compared to those from 18 patients treated by laparoscopic pyeloplasty, some of whom had undergone EUS.


Crossing vessels were identified in 27 of the 41 patients (66%). Primary treatment consisted of endopyelotomy for 26 patients and laparoscopic pyeloplasty for 15. The overall success rate for 24 endopyelotomy patients with an adequate follow-up (mean 19 months) was 71%, with more success in patients with no crossing vessels (11 of 13 (85%) vs six of 11 (55%)). Of the 18 patients treated by laparoscopic pyeloplasty (mean follow-up 15.1 months) 17 were successful.


The results for endopyelotomy were disappointing in patients with crossing vessels, despite using EUS. The results suggest that patients with crossing vessels should be treated by laparoscopic pyeloplasty. More data are needed to compare endopyelotomy with laparoscopic pyeloplasty in patients with no crossing vessels.


Options for the surgical management of PUJ obstruction now include balloon dilatation, antegrade, retrograde or cutting balloon endopyelotomy, and open or laparoscopic pyeloplasty. The minimally invasive endoscopic options have become widely accepted because of the reduced morbidity, operative time and hospital stay, despite inferior results than with open pyeloplasty [1–3]. Laparoscopic pyeloplasty is now recognized as giving equivalent results to the open procedure with less postoperative morbidity but longer operative times [4,5].

Although the true relationship between the causes of PUJ obstruction and the presence of a crossing vessel at the PUJ is disputed, studies suggest that the most important risk factor for failure of an endopyelotomy is the presence of a crossing vessel [1–3,6]. Imaging techniques used to identify crossing vessels include conventional angiography, helical CT, contrast-enhanced colour Doppler imaging, endoluminal ultrasonography (EUS) and MRI [7]. Whilst helical CT has the advantage of being less invasive and more cost-effective than angiography, it is not as accurate [8]. EUS has been shown to be more sensitive than CT at detecting crossing vessels [9], as well as providing additional anatomical information that may assist in directing the endopyelotomy incision, but has the disadvantage of providing this information only at the time of the proposed endopyelotomy. We report a retrospective study of using EUS before endopyelotomy, where the findings were used either to direct the incision or to decide on laparoscopic pyeloplasty.


The notes were reviewed of all 41 patients (19 male, 22 female, mean age 39 years, range 16–84) with PUJ obstruction who underwent EUS before a planned retrograde endopyelotomy at the authors' institution between October 1998 and December 2001.

The surgical procedure involved cystoscopy and retrograde ureteropyelography, after which a guidewire was inserted into the renal pelvis. The 7.2 F EUS probe (Olympus, Tokyo, Japan) was passed over the guidewire using fluoroscopic guidance to the level of the PUJ, with specific attention to orientation. The presence and position of crossing vessels (Fig. 1) or septa were noted, and these findings used to direct the endopyelotomy incision, which was made with either diathermy or a holmium laser through the ureteric wall to fat. Endopyelotomy stents were routinely left for 6–8 weeks, and F−15 diuresis renography (MAG-3) undertaken 3 months after surgery and then annually.

Figure 1.

Endoluminal ultrasound view at the right PUJ demonstrating an artery (A) and vein (V) crossing arteriorly.

Before June 2000, endopyelotomy was used in most patients despite the presence of crossing vessels, which were avoided by using the EUS results. Exceptions to this treatment protocol were patients with crossing vessels who had failed a previous endopyelotomy and those in whom no safe direction for an endopyelotomy incision could be found. In June 2000, the department protocol changed so that patients with crossing vessels at the PUJ underwent laparoscopic pyeloplasty at a later date instead of endopyelotomy.

Results were compared with those from a cohort of patients at our institution who underwent laparoscopic pyeloplasty during the same period. These patients had had EUS and/or preoperative spiral CT showing crossing vessels at the PUJ.

Treatment was defined as successful if the patient was asymptomatic and unobstructed or improved on diuresis renography. In view of the few patients involved there was no statistical comparison between the groups.


Crossing vessels were identified in 27 patients (66%) and septa in 10 (24%); 26 of 41 (63%) underwent retrograde endopyelotomy after EUS. The remaining 15 patients were found to have crossing vessels and they underwent a laparoscopic pyeloplasty later. Before June 2000, 28% of patients had their endopyelotomy deferred for laparoscopic pyeloplasty. Since the policy changed (to defer endopyelotomy in any patient found to have significant crossing vessels), this increased to 58%.

Twenty-two endopyelotomy incisions were made with diathermy and four with the holmium laser. The mean (range) stay after surgery was 1.95 (1–9) days. There was one major and five minor complications. One patient required embolization of a bleeding crossing vessel, one retroperitoneal haematoma was managed conservatively and there were four UTIs.

Two patients were lost to follow-up, leaving 24 evaluable with a mean (range) follow-up of 19 (4–30) months. The overall success rate for endopyelotomy was 71% (17 of 24), with more success in patients with no crossing vessels (11 of 13 (85%) without vs six of 11 (55%) with). Two patients considered successful were asymptomatic with equivocal, but improved, diuretic renograms; the remaining 15 successful patients were asymptomatic and clearly unobstructed on renography. Of the 18 patients treated by laparoscopic pyeloplasty (mean follow-up 15.1 months) 17 were successful. All patients with a successful result were asymptomatic and unobstructed on renography.


EUS has been shown to be superior to CT angiography for detecting crossing vessels [9]. For patients with PUJ obstruction associated with crossing vessels, the present results suggest that laparoscopic pyeloplasty is better than endopyelotomy despite using EUS to direct the incision. For patients with PUJ obstruction not associated with crossing vessels, the endopyelotomy results approach those of laparoscopic pyeloplasty.

As in previous studies, we classed any improvement on renography to be a success, so long as patients were asymptomatic [10,11]. The values suggest that the endopyelotomy results for patients with no crossing vessels are better than those with, although there were too few patients for a valid statistical analysis. These results are consistent with the findings of previous studies which stratified patients according to the presence or absence of crossing vessels [2,6].

In June 2000 we changed our practice for patients found to have crossing vessels; the original practice was to carry out an endopyelotomy in patients with crossing vessels, using the information from EUS to direct the incision. The new protocol involves deferring endopyelotomy if any significant crossing vessels are found directly related to the PUJ, and proceeding to later laparoscopic pyeloplasty. The present results question the use of EUS for all patients, as most (58%) required two anaesthetic procedures. Improvements in the quality of spiral CT have increased the accuracy of CT angiography, although it is not as sensitive as EUS [9]. We now use spiral CT as an initial screening test for the presence of crossing vessels when an endopyelotomy is being considered. We reserve EUS for use immediately before endopyelotomy, to exclude crossing vessels not identified by CT and to provide information about septa.

What are the implications of undertaking more laparoscopic pyeloplasties? Several endopyelotomies can be conducted in the time it takes for one laparoscopic pyeloplasty, however experienced the surgeon; in addition, the hospital stay is shorter for endopyelotomy. These factors must be balanced against the higher success rate of laparoscopic pyeloplasty and patients should be counselled appropriately. The success of laparoscopic pyeloplasty appears to be equivalent or better than open pyeloplasty, and has obvious benefits in terms of reduced recovery time and hospital stay.

The limitations of the present study are that it is retrospective, not randomized, and with few patients in each treatment group. We are currently collecting data prospectively on all our patients with PUJ obstruction. There are several practical difficulties in carrying out a randomized study of endopyelotomy vs laparoscopic pyeloplasty, the greatest of which would be recruitment. By conducting a prospective observational trial, we hope to be able to identify those who may benefit most from endopyelotomy. PUJ obstruction is uncommon, and the treatment options have increased rapidly in the last decade, so it is inevitable that studies involve few patients.

In conclusion, this study suggests that pyeloplasty remains better than endopyelotomy for patients with crossing vessels, despite using EUS to direct the incision. We currently treat all patients with significant crossing vessels using laparoscopic pyeloplasty, reserving endopyelotomy for those with none. With fewer patients undergoing endopyelotomy because of the presence of crossing vessels, our current protocol of using EUS in all patients cannot be justified, as CT angiography provides the information before surgery. We now use EUS immediately before endopyelotomy to exclude crossing vessels missed by CT. The role of endopyelotomy in PUJ obstruction has yet to be determined, as the results do not compare with pyeloplasty, and we are currently collecting data to assess the results of endopyelotomy in the group of patients with no crossing vessels.


endoluminal ultrasonography.