Pred iction of High Grade Vesicoureteral Reflux in Infants Less than 3 Months with Urinary Tract Infection

Yi, Dae Yong; Kim, Na Yeon; Cho, Hee Yeon; Kim, Ji Eun; Sim, So Yeon; Son, Dong Woo; Jeon, In Sang; Cha, Han

Abstract

PURPOSE: Our aim was to investigate the predictive factors for detecting grade III-V vesicoureteral reflux(VUR) in young infants less than 3 months with urinary tract infections (UTI).
METHODS
Data of infants who underwent ultrasonography and VCUG between January 2004 and September 2007 were reviewed. Age, gender, incidence of bacteremia, C-reactive protein(CRP) and imaging studies were compared between group I(grade III-V VUR) and group II (normal or grade I and II VUR) retrospectively. Sensitivity, specificity, positive and negative predictive values, odds ratio, and likelihood ratio of ultrasonography for high grade VUR were evaluated.
RESULTS
Among 54 enrolled infants(41 males, 13 females), 14 infants were group I and 40 infants were group II. In the group I, CRP level was significantly higher(6.11+/-5.18 vs. 3.27+/-3.45, P=0.025), and there were more ultrasonographic abnormal findings(71.4%, vs. 22.5%, P=0.002) compared with group II. However, ultrasonography was the only significant factor after adjusting with logistic regression(P=0.002). Incidence of bacteremia and abnormal DMSA findings were not significantly different in two groups. Sensitivity, specificity, and odds ratio of ultrasonography was 71.4%, 77.5%, 6.9 respectively. Negative predictive value was 88.6% and negative likelihood ratio was 0.37. Ultrasonography had significant negative likelihood ratio for grade III-V VUR, but missed 4 infants with grade III VUR.
CONCLUSION
We could not find any alternative predictive factors to reduce VCUG in detecting high grade VUR. Therefore, VCUG must be considered in young infants less than 3 months with UTI.