Background: The aim of this study was to identify the cut-offs of postnatal anteroposterior renal pelvic diameter (APRPD), according to the urinary tract dilation (UTD) classification system, to identify the predictors of final diagnosis of UTD and the need for surgery. Methods: A total of 260 infants (336 renal units) with prenatally detected UTD were prospectively evaluated on serial ultrasonography by the same radiologist. Additional voiding cystourethrography and scintigraphy was done according to the clinical algorithm. Results: Prenatal and postnatal APRPD in patients with transient dilation were significantly lower than in those with urinary tract anomalies (UTA). On follow up, the slope of decrease in APRPD was significantly higher in transient dilation compared with UTA. APRPD 10 mm at first-month ultrasonography, predicted UTA with a sensitivity of 83.1%, and specificity of 71.1%. On multivariate analysis the likelihood of surgical intervention and final diagnosis were predicted independently by the UTD system risk group. Conclusions: Careful ultrasonography evaluation can avoid unnecessary testing in patients with transient or clinically insignificant dilation. The UTD classification system is valid for evaluation of postnatal hydronephrosis and is reliable in predicting the need for surgical intervention.