Utility of the 3-months post-operative ultrasound and percent reduction in antero-posterior diameter of renal pelvis to identify success, early failures and those at risk for late failures after pediatric dismembered pyeloplasty: an analysis of 359 cases.

BACKGROUND: There is no consensus on the ideal follow-up testing after pyeloplasty for ureteropelvic junction obstruction (UPJO) in children. Although diuretic renogram (DR) is considered the gold standard to define successful pyeloplasty, percent reduction in antero-posterior diameter (PR-APD) on post-operative ultrasound (USG) has been reported to be useful to identify successful pyeloplasty. We sought to further explore PR-APD ranges in the first post-operative USG that would have high specificity to identify successful and failed pyeloplasty.

METHODS: From a 10-year database of pediatric pyeloplasties performed in 4 centers, cases with follow-up USG at 3 months post-pyeloplasty and follow-up DR were selected for analysis. On the basis of drainage on follow-up DR, the patients were divided into group 1 (successful) and group 2 (failures). PR-APD was compared between the two groups, and receiver operating curve (ROC) analysis was performed to assess the cutoff value of PR-APD that had the best discriminative ability to predict success. Further, within group 2, PR-APD was compared between early (2a) and late (2b) failures.

RESULTS: Three hundred and fifty-nine children (323 group 1, 36 group 2) who underwent primary dismembered pyeloplasty for unilateral UPJO were included. Both groups were comparable for demographic parameters and mean pre-operative APD. On the 3-months post-operative USG, group 1 had significantly lower post-operative APD (p = 0.0001) and higher PR-APD (62.8% vs 24.6%, p < 0.001) than group 2. On ROC analysis, PR-APD of 42% had 91% sensitivity and 100% specificity for predicting success of pyeloplasty. PR-APD <15% predicted early failures with 100% specificity, while PR-APD between 16 and 42% was associated with 45% risk of late pyeloplasty failure.

DISCUSSION: We identified the cut-off values of PR-APD that could predict success and early failure of pyeloplasty with high specificity. While earlier reports had limited number of failed cases, our data was unique that we had a sizable number of failed pyeloplasty cases, thus enabling meaningful comparison of PR-APD between successes and failures. In addition, we identified a subgroup at risk of late failure. These findings might be useful to avoid routine post-operative DR studies after pyeloplasty, detect early failures and identify the few patients who benefit from long-term follow-up.

CONCLUSIONS: On the 3-month post-operative USG after pyeloplasty, PR-APD 42% or more predicted success with 100% specificity, while PR-APD <15% was 100% specific in predicting early failures. Those with PR-APD 16-42% may be at risk for late failures and may benefit from long-term follow-up.