Prediction of moderate and high grade vesicoureteral reflux after a first febrile urinary tract infection in children: construction and internal validation of a clinical decision rule.

ABSTRACT

PURPOSE: Urinary tract infection leads to a diagnosis of moderate or high grade (III or higher) vesicoureteral reflux in approximately 15% of children. Predicting reflux grade III or higher would make it possible to restrict cystography to high risk cases. We aimed to derive a clinical decision rule to predict vesicoureteral reflux grade III or higher in children with a first febrile urinary tract infection. MATERIALS AND METHODS: We conducted a secondary analysis of prospective series including all children with a first febrile urinary tract infection from the 8 European participating university hospitals. RESULTS: A total of 494 patients (197 boys, reflux grade III or higher in 11%) were included. Procalcitonin and ureteral dilatation on ultrasound were significantly associated with reflux grade III or higher and then combined into a prediction model with an ROC AUC of 0.75 (95% CI 0.69-0.81). Given the prespecified constraint of achieving at least 85% sensitivity, our model led to the clinical decision rule, for children with a first febrile urinary tract infection cystography should be performed in cases with ureteral dilatation and serum procalcitonin level 0.17 ng/ml or higher, or without ureteral dilatation (ie ureter not visible) when serum procalcitonin level is 0.63 ng/ml or higher. The rule had 86% sensitivity (95% CI 74-93) with 47% specificity (95% CI 42-51). Internal cross-validation produced 86% sensitivity (95% CI 79-93) and 43% specificity (95% CI 39-47). CONCLUSIONS: A clinical decision rule was derived to enable a selective approach to cystography in children with urinary tract infection. The rule predicts high grade vesicoureteral reflux with approximately 85% sensitivity and avoids half of the cystograms that do not find reflux grade III or higher. Further validation is needed before its widespread use.