RESUMO
PURPOSE: We sought to identify predictive factors for primary vesicoureteral reflux among infants with prenatally detected renal pelvic dilatation. MATERIALS AND METHODS: A total of 250 neonates were diagnosed with isolated renal pelvic dilatation between 1999 and 2008, and followed prospectively. The main event of interest was presence of moderate to severe reflux (grade III to V). Diagnostic odds ratio, sensitivity, specificity, and diagnostic accuracy (assessed by AUC) of fetal and postnatal renal pelvic dilatation were determined. Severity of dilatation was classified by Society for Fetal Urology grade. Binary logistic regression was performed to identify variables significantly associated with vesicoureteral reflux. RESULTS: A total of 23 patients (9.2%) had primary vesicoureteral reflux, of whom 16 had grade III to V disease. Diagnostic accuracy was 0.70 (95% CI 0.63 to 0.75) for fetal and 0.65 (95% CI 0.59 to 0.71) for postnatal renal pelvic dilatation. Combined results of fetal and postnatal renal pelvic dilatation were also assessed. When both tests less than 10 mm were considered negative indicators of moderate to severe vesicoureteral reflux sensitivity increased to 97% and diagnostic odds ratio to 19.1. After adjustment by logistic regression only Society for Fetal Urology grade greater than I and ureteral dilatation were variables independently associated with grade III to V reflux. CONCLUSIONS: Fetal and postnatal renal pelvic dilatation was a poor predictor of vesicoureteral reflux. Nevertheless, diagnostic accuracy regarding clinically significant vesicoureteral reflux improved when fetal and postnatal renal pelvic dilatation less than 10 mm was considered a negative indicator of reflux.