Imaging strategy for infants with urinary tract infection: a new algorithm.

PURPOSE: We analyzed clinical data for prediction of permanent renal damage in infants with first time urinary tract infection. MATERIALS AND METHODS: This population based, prospective, 3-year study included 161 male and 129 female consecutive infants with first time urinary tract infection. Ultrasonography and dimercapto-succinic acid scintigraphy were performed as acute investigations and voiding cystourethrography within 2 months. Late scintigraphy was performed after 1 year in infants with abnormality on the first dimercapto-succinic acid scan or recurrent febrile urinary tract infections. End point was renal damage on the late scan. RESULTS: A total of 270 patients had end point data available, of whom 70 had renal damage and 200 did not. Final kidney status was associated with C-reactive protein, serum creatinine, temperature, leukocyturia, non-Escherichia coli bacteria, anteroposterior diameter on ultrasound and recurrent febrile urinary tract infections. In stepwise multiple regression analysis C-reactive protein, creatinine, leukocyturia, anteroposterior diameter and non-E.coli bacteria were independent predictors of permanent renal damage. C-reactive protein 70 mg/l or greater combined with anteroposterior diameter 10 mm or greater had sensitivity of 87% and specificity of 59% for renal damage. An algorithm for imaging of infants with first time urinary tract infection based on these results would have eliminated 126 acute dimercapto-succinic acid scans compared to our study protocol, while missing 9 patients with permanent renal damage. CONCLUSIONS: C-reactive protein can be used as a predictor of permanent renal damage in infants with urinary tract infection and together with anteroposterior diameter serves as a basis for an imaging algorithm.