Risk factors influencing the recurrence of obstruction post-pediatric pyeloplasty: a single-center prospective study

ABSTRACT

Background

Ureteropelvic junction (UPJ) obstruction is the leading cause of pediatric hydronephrosis. While open pyeloplasty is highly effective, a small proportion of children experience recurrent ureteropelvic junction obstruction (UPJO) necessitating repeat intervention. We aimed to identify potential risk factors for pyeloplasty failure.

Methods

We prospectively followed 126 children undergoing open Anderson-Hynes pyeloplasty between 2019 and 2021. Demographic, clinical, anatomical, procedural, and postoperative variables were recorded. Patients were monitored for recurrence, defined as worsening hydronephrosis and/or symptoms with diuretic renogram confirmation. Univariate and multivariate analysis identified factors associated with recurrence.

Results

Median age at pyeloplasty was 17 months, left side was affected in 58% and 73.8% were male. At mean 28.8 months follow-up, UPJO recurred in 9 patients (7%). The only factor significantly associated with recurrence was prolonged postoperative urinary drain leakage, specifically recurrence occurred in 50% (3/6) of cases with prolonged leakage versus 5% (6/120) without (p = 0.005).

Conclusions

Open pediatric pyeloplasty has a high success rate, yet a minority of patients will experience recurrence necessitating re-intervention. Prolonged urine leakage was the sole predictor of failed repair in our cohort, likely due to inflammation and perianastomotic scarring. Meticulous surgical technique and prompt leak management are vital to optimize pyeloplasty outcomes. Children at higher risk warrant close monitoring to enable early detection and renal salvage.

KEYWORDS:

• Pyeloplasty
• ureteropelvic junction obstruction (UPJO)
• recurrence post pyeloplasty
• risk factors
• pediatric urology

1. Introduction

UPJO is the most common cause of pediatric hydronephrosis, with dismembered pyeloplasty being the gold standard treatment to relieve obstruction and preserve renal function. Open pyeloplasty has excellent reported success rates exceeding 90% across pediatric series [1].

Elucidating predictors of adverse outcomes could facilitate patient selection and counseling for pyeloplasty, guide surgical decision-making, and direct postoperative surveillance protocols to enable early detection of recurrence when salvage interventions may still preserve renal function [1,2].

Therefore, we prospectively analyzed potential risk factors for recurrence after pediatric pyeloplasty to facilitate earlier diagnosis and intervention for treatment failures.

2. Patients and methods

After obtaining approval from the institutional ethics review board and informed consent acquisition from parents/guardians, a prospective chart review and analysis of 126 consecutive pediatric subjects undergoing open Anderson-Hynes dismembered pyeloplasty between 2019 and 2021. Of these patients, 9 (7%) experienced recurrent UPJO postoperatively. Exclusion criteria encompassed individuals undergoing alternate pyeloplasty approaches (Scardino, Y-V flaps, non-dismembered techniques) and laparoscopic approach.

Preoperative assessments included a comprehensive medical history, full physical examination, and standard laboratory testing. All patients underwent ultrasound on the urinary system (US) and 99mTc-Diethylenetriamine pentaacetate (DTPA) diuretic renal scan to quantify differential renal function (DRF) and radiotracer drainage time (half-time). Voiding cystourethrogram (VCUG) was indicated in children with febrile urinary tract infection (UTI) and US evidence of ureteric dilatation to diagnose vesicoureteral reflux. Magnetic resonance urography (MRU) and computed tomography (CT) were utilized when indicated.

The decision for pyeloplasty was based on a US and DTPA scan, with criteria including symptomatic patients, progressive hydronephrosis in serial US, impaired split renal function (below 40%), and a decrease in split renal function over 10% in subsequent DTPA scan.

All pyeloplasties were performed by 4 senior surgeons using the Anderson-Hynes open dismembered technique. The type of incision and decision to perform retrograde pyelography (RPG) were based on patient factors and surgeon preference. All our cases were stented with double J (DJ) stent. Closed drainage catheters were typically maintained for 2 to 3 days postoperatively, and Foley urinary catheters were left indwelling for 1 day. Antibiotic prophylaxis was administered until removal of all indwelling catheters and drains. DJ stents were usually removed 6 to 8 weeks postoperatively.

Postoperatively, all children were monitored using serial US. DTPA was only performed for those who experienced non-improving persistent, or worsening hydronephrosis, as well as recurrent symptoms such as abdominal pain or pyelonephritis during follow-up.

Univariate and multivariate analysis were conducted to identify risk factors associated with pyeloplasty failure. The variables analyzed included: patient age at time of surgery, clinical presentation, hydronephrosis severity per the Society for Fetal Urology (SFU) grading system, presence of giant hydronephrosis, anteroposterior (AP) renal pelvic diameter, renal parenchymal thickness, DRF of the affected kidney, associated vesicoureteral reflux (VUR), utilization of RPG, preoperative urinary diversion, surgical incision approach, intrinsic versus extrinsic etiology of obstruction, intra-renal versus extra-renal pelvis, postoperative complications including excessive urinary drain leakage, and fever.

Quantitative variables were summarized using means as measures of central tendency with ranges and interquartile ranges as dispersion descriptors. Categorical variables were summarized by counts and percentages. Chi-square testing was used to evaluate associations between categorical variables. Fischer’s exact and Monte Carlo significance testing was reported instead when over 20% of expected cell counts were < 5. The Mann-Whitney U test compared median values for quantitative variables between surgical success and failure groups. Tests were selected based on variable distribution per Kolmogorov-Smirnov testing. All statistical tests were two-sided, judged at 0.05 significance level and were performed using IBM SPSS statistics program version 28 [3,4].

3. Results

The median age at surgery was 17 months (range 2–192 months). The majority were males (73.8%). Left, right and bilateral units were affected in 58%, 32.5% and 9.5% respectively. Most common presentation was antenatal hydronephrosis (66.7%), followed by symptoms of pain, UTI, swelling and incidental discovery. (Table 1)

Table 1. Demographic data of the studied patients.

	Frequency	%
Age:
0–6 months	39	31.0%
>6 months	87	79.0%
Median in months (Min-Max)	17 (2–192)
Gender:
Male	93	73.8%
Female	33	26.2%
Affected side:
Right side	41	32.5%
Left side	73	58%
Bilateral	12	9.5%
Mode of presentation
Antenatal hydronephrosis	84	66.7%
UTI	13	10.3%
Pain	12	9.5%
Swelling	10	7.9%
Accidental discovery	6	4.8%
Hematuria	1	0.8%

According to the SFU grading system, 94.4% of cases demonstrated grade IV and 5.6% grade III hydronephrosis on preoperative renal ultrasound. The median AP pelvic diameter was 37 mm (range 22 to 100 mm) and the median parenchymal thickness was 5 mm (range 1 to 11 mm). Anatomical variations were seen in 10 cases: 3 horseshoe kidneys, 3 ectopic kidneys, 2 duplex kidneys, and 2 solitary kidneys. Associated renal stones were detected in 5 patients (3.9%). Hydroureter was seen in 4 (3.2%). VCUG was done in 17 (13.5%) cases, where 2 had ipsilateral G II VUR. Diuretic renography showed a median DRF of 41% (range 18 to 66%) in the affected units.

CT and MRU were performed for renal morphology and vascular anatomy identification in selected cases, detecting 3 crossing vessels.

Preoperative percutaneous nephrostomy catheter (PCN) drainage was performed in 3 patients with presenting symptoms of febrile UTI. Physical examination in these cases revealed palpable, tender renal units concerning for acute pyonephrosis, which was confirmed on renal ultrasonography demonstrating hydronephrosis with echogenic debris. PCN catheters were left in place for a mean period of 37 days before pyeloplasty.

Open Anderson-Hynes dismembered pyeloplasty was performed in all cases. Flank incision was applied for 123 cases (97.6%), while Gibson incision in 3 ectopic pelvic kidneys (2.4%). Mean operative duration was 67 minutes (range 55–109 mins). RPG was performed in 16 of 126 children (12.7%). Anatomical variations were seen in 9, including long segment ureteric narrowing in 3, high insertion of the ureter in a non-dependent point in 4, continuous filling defect of a crossing vessel in 1, and hugely dilated pelvis reaching umbilicus level in a malrotated kidney in 1.

The intraoperative findings revealed malrotation in 21 kidneys (16.7%), a giant hydronephrosis in 11 (8.7%), an intra-renal pelvis in 7 (5.6%), and crossing vessels in 7 (5.6%). Additionally, periureteral adhesions were encountered in 3 cases with preoperative percutaneous nephrostomy tube placement. DJ ureteric stents were placed in all patients. Concurrent stone extraction was performed in 5 cases (4%).

The sole intraoperative complication encountered was an iatrogenic lower third ureteral injury during attempted DJ stent placement in 1 of 126 patients (0.8%). This was immediately recognized and primarily repaired upon decision of a senior consultant with ureteric reimplantation without further sequelae. There were no other intraoperative complications in this series.

Postoperatively, the mean length of hospitalization was 2.3 days, (range 2 to 7). Foley catheter and drain removal occurred at mean 1.2 days (range 1 to 13) and 2.2 days (range 2 to 13) respectively. 16 patients had a post-operative fever, 6 had prolonged excessive urinary drain leakage, and 1 case of DJ stent migration into the urinary bladder occurred postoperatively after 2 weeks, requiring cystoscopic retrieval and reinsertion of the stent. The mean time to DJ stent removal was 2 months (range 1 to 6), it was kept for a long time (4–6 months) in 9 cases due to recurrent temporary closures during the COVID-19 outbreak.

Patients were monitored for a mean of 28.8 months (range 18–36 months) postoperatively. US confirmed improvement of hydronephrosis in 111 renal units. In the remaining 15 patients, 6 demonstrated only minimal sonographic improvement in hydronephrosis grade that prompted confirmatory diuretic renography; this verified the absence of ongoing obstruction in these cases. The remaining 9 cases (7.1%) were found to have recurrent UPJO on follow-up assessment.

Recurrent UPJO was encountered at a mean of 19 months postoperatively (range 12–26 months). Of these, 5 cases presented with asymptomatic progressive hydronephrosis on ultrasound, 3 had recurrent flank pain, and 1 developed pyelonephritis. Follow-up diuretic renal scan confirmed the obstruction and demonstrated a mean DRF of 44%. In 4 case DRF has reduced by a mean of 11%.

Further analysis of the 9 cases with recurrent UPJO necessitating re-intervention revealed that 3 underwent endopyelotomy and 6 underwent redo open pyeloplasty, all had preoperative RPG confirming the presence of obstruction. Among the endopyelotomy cohort, 1 of 3 cases failed repeat endoscopic management, ultimately requiring salvage with redo open pyeloplasty. Comparatively, all 6 subjects undergoing redo open repair achieved durable outcomes without any further recurrence. The intraoperative finding during redo open was isolated periureteral adhesions and fibrosis in all cases.

An assessment was undertaken of potential risk factors that could be associated with recurrence after pyeloplasty. Of variables analyzed, postoperative prolonged excessive urinary drain leakage was found to have a significant association with failed repair (p ≤ 0.005). Specifically, recurrence developed in 3 of 6 cases (50%) with prolonged postoperative excessive drain leakage, compared to recurrence in only 6 of 120 cases (5%) without prolonged leakage. (Tables 2,3)

Table 2. Analysis of risk factors that may lead to recurrence.

	Success(n = 117)		Failure(n = 9)		Sig.
Age:					^MCp 0.058
0–6 months	39	33.3%	1	11.1%
>6 months	78	66.7%	8	88.9%
Gender
Male	86	73.5%	7	77.8%	^FEp1
Female	31	26.5%	2	22.2%
Mode of presentation
Antenatal hydronephrosis	79	67.5%	5	55.6%
Accidental discovery	5	4.3%	1	11.1%
Swelling	9	7.7%	1	11.1%	^MCp.942
UTI	12	10.3%	1	11.1%
Pain	11	9.4%	1	11.1%
Hematuria	1	0.8%	0	0.0%
Laterality
Right side	40	34.2%	1	11.1%
Left side	65	55.6%	8	88.9%
Bilateral	12	10.3%	0	0.0%	^MCp.543
Preoperative Hydro grade
GIII	6	5.1%	1	11.1%
GIV	111	94.9%	8	88.9%	^FEp.413
Duplex Kidney:
Normal Kidney	115	98.3%	8	88.9%	^FEp.201
Duplex	2	1.7%	1	11.1%
Pelvic Kidney
No	114	97.4%	9	100.0%	^FEp1
Yes	3	2.6%	0	0.0%
Horseshoe kidney
No	115	98.3%	8	88.8%	^FEp.201
Yes	2	1.7%	1	11.2%
Debris in the renal pelvis
yes	6	5.1%	0	0.0%	^FEp1
No	111	94.9%	9	100.0%
Malrotation
No	98	83.8%	7	77.8%	^FEp.644
Yes	19	16.2%	2	22.2%
Giant Hydronephrosis
No	107	91.5%	8	88.9%	^FEp.529
Yes	10	8.5%	1	11.1%
Post-operative fever
No	103	88.0%	7	77.8%	^FEp.320
Yes	14	12%	2	22.2
Post-operative prolonged excessive urinary leak
No	114	97.4%	6	66.7%	^FEp.005*
Yes	3	2.6%	3	33.3%

^MCp: MonteCarlo significance, ^FEp: FischerExact significance, *Significant results ≤ .05.

Table 3. Demonstrates that prolonged excessive postoperative urinary drainage was a significant risk factor for recurrence following pediatric pyeloplasty. Recurrence occurred in 50% of cases with prolonged leakage compared to only 5% of cases without prolonged drainage.

Post-operative prolonged excessive urinary leak		Success	Failure	Total	^FEp.005*
No	Count	114	6	120
	%	95.0%	5.0%	100%
Yes	Count	3	3	6
	%	50.0%	50.0%	100.0%
Total	Count	117	9	126
	%	92.9%	7.1%	100.0%

^FEp: FischerExact significance, *Significant results ≤ .05.

4. Discussion

Prior studies have evaluated various demographic, clinical, anatomical, and procedural factors that may predict pyeloplasty failure. However, findings have been inconsistent, likely due to small sample sizes and few recurrence events limiting statistical power [5,6]. Our study aimed to assess potential risk factors for recurrent UPJO after pediatric pyeloplasty using both univariate and multivariate analysis. This allowed us to determine the strength of the association with recurrent obstruction.

Identification of risk factors leading to recurrence is important, as it allows surgeons to closely monitor higher risk cases postoperatively, counsel families regarding prognosis, and optimize surgical planning. In our study, we analyzed a large prospective cohort of 126 consecutive pediatric patients undergoing open Anderson-Hynes pyeloplasty to determine potential risk factors for postoperative recurrence of UPJO. Our recurrence rate was 7%, which is comparable to the 5–7% range reported in other large pediatric series [2,7].

The existing literature on open Anderson-Hynes dismembered pyeloplasty includes two prominent retrospective cohort studies. Ceyhan et al. and Braga et al. have both published analyses of their respective patient populations who underwent this surgical procedure. Both studies employed multivariate statistical techniques to explore potential risk factors associated with outcomes. The reported recurrence rates were comparable across the two series, falling within the range of approximately 6% [2,7].

Ceyhan et al. found early postoperative complications (urinary infection and urinoma) and preoperative diversion (PCN or DJ stent) were independent risk factors, while Braga et al. identified lack of RPG and dorsal lumbotomy approach as significant risks [2,7].

Our study could only identify prolonged excessive urinary drain leakage as a risk factor for recurrence, as 33.3% of our failures had exhibited prolonged excessive leak. Notably, recurrence occurred in 50% of cases complicated by excessive leakage, compared to only 5% of cases without prolonged leakage. Multiple studies, including research by Ceyhan et al., Braga et al., and Lim et al., reinforce that urinary leakage problems may cause perianastomotic fibrosis, worsening the healing process causing pyeloplasty failure [2,7,8,9]. Further supporting these assertions, Helmy et al. documented postoperative leakage in 24% (5/18) subjects among their pyeloplasty recurrence cohort [10]. Therefore, prevention and early detection of prolonged excessive leakage are important for improving pyeloplasty outcome. Some of the measures that may help prevent excessive leakage are meticulous suturing technique, adequate drainage placement, appropriate stent size and duration.

RPG prior to pediatric pyeloplasty has been examined in several retrospective studies, with varying conclusions on its necessity. Bachor et. al. and Rushton et. al. concluded that routine RPG is not necessary before pediatric pyeloplasty if ultrasound shows no ureteral dilation [11,12]. However, Narayanan et al. found that RPG may not be necessary with a standard loin incision but can help guide a dorsal lumbotomy approach [13]. Taken together, these studies suggest selective use of RPG guided by patient factors and surgeon preference is reasonable. Comparatively, while Ceyhan et al. routinely employ preoperative use of RGP, Braga et al. documented substantially lower rates (48.6% of cases) [2,7]. Our present cohort underwent selective RPG in just 12.7% of 126 pyeloplasties without any significant difference in failure rate compared to Braga et al. Additionally, RPG revealed anatomical variations that could be visualized during surgery and did not ultimately affect surgical decision.

Consistent with the findings of Braga et al., younger age below 6 months was not identified as a risk factor for recurrence in our cohort. This is despite Lim et al. reporting younger age as a potential predictor of recurrence [9]. The reasons for this discrepancy are unclear but may relate to variations in surgical technique and postoperative care between centers. Specifically, the meticulous handling of tissues under magnification and use of small-caliber ureteral stents in our series may have mitigated any potential risks conferred by younger age.

The study conducted by Smith et al. suggests that in cases of giant hydronephrosis, if the redundant renal pelvis is not adequately reduced, or if the ureter is not sufficiently spatulated during the surgical procedure, a persistent kink in the ureter may develop after the UPJ is repositioned into the retroperitoneal space. This can lead to postoperative urinary leakage, which is likely due to the persistence of obstruction at the UPJ level [14]. However, Sarhan et al. examined the outcomes of open dismembered pyeloplasty in 41 pediatric patients with giant hydronephrosis secondary to UPJO. The authors reported that at a mean follow-up period of 34 months, the overall success rate was 91%, as determined by both symptom relief and radiographic improvement in the patients [15]. In a study of 167 cases of unilateral giant hydronephrosis, Baek et al. evaluated the long-term surgical outcomes of dismembered pyeloplasty for giant hydronephrosis. The results demonstrated that pyeloplasty allowed for satisfactory long-term outcomes and comparable success rate in children with giant hydronephrosis [16].

In our series, giant hydronephrosis was not a risk factor for recurrence, as the redundant pelvis was sufficiently reduced and a well-vascularized pelvic flap at the dependent point of the pelvis was anastomosed to a well-vascularized, generously spatulated ureter.

Some researchers have suggested that missed crossing vessels may also contribute to the occurrence of recurrence following pyeloplasty. To explore this further, Zu’bi et al. reported on their findings in a series of 9 pediatric patients undergoing redo pyeloplasty, where 7 of these cases were found to have crossing vessels that had been missed during the initial pyeloplasty operation. In contrast, other studies have reported lower rates of missed crossing vessels identified during repeat pyeloplasty. Braga et al. found crossing vessels in only 2 out of 21 children undergoing redo pyeloplasty, while Helmy et al. reported just 1 out of 18 cases with a missed crossing vessel [7,10,17].

We encountered no missed crossing vessels in the recurrent cases, likely due to careful preoperative evaluation and intraoperative inspection and identification of any crossing vessels, especially when suspected in older children presenting with Dietl’s crisis. These findings eliminate surgeon-related errors and suggest the recurrences may be primarily related to healing failure or pathological ureter.

Prior studies exhibit conflicting evidence regarding preoperative drainage as a risk factor for compromised pyeloplasty outcomes. Ceyhan et al. identified preoperative decompression via PCN or DJ stent as independently associated with surgical failure. However, Baek et al. found no significant correlation between pre-procedure nephrostomy and success rates [2,16]. Aligned with the latter study, our current analysis did not establish pre-operative PCN as predictive of recurrence. We suggest two potential factors reconciling these inconsistent risk relationships: firstly, pre-surgical drainage durations in our cohort; and secondly, meticulous dissection and tissue handling, facilitating precise subsequent anastomosis.

Similar to Ceyhan et al. and Braga et al., our study found no associations between recurrence and gender, side, presence of antenatal hydronephrosis, SFU hydronephrotic grade, obstruction cause (internal or crossing vessel), or type of renal pelvis (intra-renal or extra-renal) [2,7].

The differences in identified risk factors may relate to variations in surgical techniques and practices between institutions. Overall, these complementary retrospectives and our prospective analysis demonstrate that UPJO recurrence after pediatric pyeloplasty is uncommon, but predictors remain inconsistently defined. This highlights the need for large prospective multi-institutional cohorts powered to elucidate potential demographic, procedural, and perioperative risk factors through rigorous standardized protocols and data collection. Improved understanding of variables associated with recurrent UPJO will facilitate optimized surgical planning and follow-up for children at higher risk of failure after pyeloplasty.

Limitations of our study include the modest sample size of 126 patients from a single pediatric urology center compared to prior retrospective analyses by Braga and Ceyhan et al., that may have more statistical power to detect subtle risk factors for recurrent UPJO. Our selective use of retrograde pyelography in just 12.7% cases also differs from the routine use in all patients reported by Ceyhan et al. While our selective approach mirrors real-world practice variability that finds insignificant outcome differences, standardized protocols in larger prospective studies may better control confounding factors. Finally, our cohort included patients operated by 4 pediatric urologists which may introduce some technique variability. Unfortunately having just a single surgeon would dramatically limit our sample size over the study period. However, it’s worth noting that all surgeons were experienced “senior staff.”. A single high-volume surgeon series may be required to further minimize procedural differences that may impact outcomes. Overall, while our prospective methodology has advantages, large multi-institutional analyses with strict standardized protocols will be an important next step in identifying potential demographic and perioperative risk factors for pediatric pyeloplasty failure.

5. Conclusion

In our current series, prolonged postoperative urinary drain leakage emerged as the sole independent risk factor associated with recurrent ureteropelvic junction obstruction after open pediatric pyeloplasty. Recurrence rates were markedly higher, at 50%, among cases that experienced prolonged leakage compared to only 5% in those without this complication. Anastomotic leakage likely increases risks of inflammation, infection, and scar tissue formation causing re-obstruction. Early detection for children at higher risk of recurrence enables prompt re-intervention, thereby maximizing opportunities for renal salvage.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Mohamed Elsawy

Dr. Mohamed Elsawy is a pediatric Urologist from the Unit of Pediatric Urology, Student Insurance Hospital, Alexandria, Egypt.

Ahmed Fahmy

Ahmed Fahmy is a Professor in the Unit of Pediatric Urology, Urology Department, Faculty of Medicine, Alexandria University, Egypt.

Mohamed Youssef

Mohamed Youssef is a Professor in the Unit of Pediatric Urology, Urology Department, Faculty of Medicine, Alexandria University, Egypt.

Samir Orabi

Samir Orabi is a Professor in the Unit of Pediatric Urology, Urology Department, Faculty of Medicine, Alexandria University, Egypt.