Objective Vesicoureteral reflux (VUR) is found in 20–40% of children with urinary tract infection (UTI). Renal ultrasonography (US) is found by many paediatricians to be a good predictor of VUR. However, discrepancies in the results were reported, particularly because studies analysed different US criteria. Our aim was to study the diagnostic accuracy of renal US for VUR in children with UTI using a systematic review and meta-analysis.
Methods Studies were identified by a systematic electronic search from 1985 to 2006. We pooled data using a random effect model of the diagnostic odds ratio (DOR) according to each US criterion.
Results From the 1456 eligible articles, 33 observational cohort studies were included (5054 patients, 27% with VUR) to study the diagnostic accuracy of renal US for all grade VUR. Pelvicalyceal, ureteral and urinary tract dilation and abnormal renal length had a significant pooled DOR, respectively, of 3.3 (95% CI 1.5 to 7.2), 1.8 (1.0 to 3.0), 2.3 (1.6 to 3.5), 4.0 (1.6 to 9.6) with heterogeneity (I2 >50%), except for ureteral dilation (I2 = 30%). The diagnostic accuracy of renal US for high-grade VUR was analysed in 10 studies (1857 patients, 10% with high-grade VUR). Only ureteral dilation and abnormal renal length had a significant pooled DOR without evidence of heterogeneity (I2 <50%): 5.5 (1.3 to 22) and 3.9 (1.7 to 8.6), respectively.
Conclusions Ureteral dilation seemed the best renal US criterion for the prediction of both all-grade and high-grade VUR in children with UTI. Its combination with other sensitive predictors, such as procalcitonin, in a clinical decision rule could be useful.