Prediction of the outcome of antenatally diagnosed hydronephrosis: a multivariable analysis.

PURPOSE

The majority of cases of antenatal hydronephrosis (ANH) resolve postnatally. The purpose of this study was to determine independent predictors for resolution of ANH using a multivariable analysis model.

METHODS

A retrospective study was performed on 100 children (80 males and 20 females, 118 renal units) referred to a single pediatric urology clinic with the diagnosis of ANH. Patients with ultimately resolved ANH were compared to unresolved cases in a univariate analysis in terms of sex, laterality, severity of ANH using Society for Fetal Urology (SFU) grading, antero-posterior pelvic diameter (APD), parenchymal thickness, renographic differential function and development of clinical complications, followed by a Cox proportional hazard model for multivariable analysis.

RESULTS

Median follow up was 34 months (range 3-204). Hydronephrosis in 62 units resolved spontaneously and pyeloplasty was done in 29. The remaining 27 units had persistent uncomplicated hydronephrosis at last follow up. Multivariate analysis showed larger APD (hazard ratio 0.54; 95%CI 0.36-0.80) and SFU grade 4 (HR 0.34; 95%CI 0.13-0.90) to be associated with a significantly lower likelihood of resolution. The mean initial APD in resolved cases was 9.4mm as opposed to 29.0mm in cases requiring surgery.

CONCLUSION

Large initial APD has predictive value for surgical intervention. This model is helpful in counseling families about the potential outcomes of ANH.