Vesicoureteral reflux: subpopulations of patients defined by clinical variables.

Abstract

The first imaging evaluation of the child with urinary tract infection (UTI) is often the same for all children, regardless of the child's clinical presentation. However, this approach is simplistic and ignores considerable differences in the frequency of abnormal pathophysiology in different subpopulations of children with UTI. Six clinical variables are evaluated as predictors of vesicoureteral reflux (VUR) in a large series of girls with UTI. Data were collected from a consecutive series of 919 girls undergoing a first imaging evaluation for UTI. Six input variables were used: age, maximum body temperature (T(max)), number of UTIs, hospitalization, family history of childhood UTI, and rapidity of response to antibiotic therapy. The dependent variable was VUR. Data were enumerated and analyzed by logistic regression and the chi-square test. VUR was present in 28.8 %. The percentage with VUR varied from 56.1 % for age < 6 months and T(max) L 38.5 degrees C to 13.0 % for age L 10 years and T(max) < 38.5 degrees C. The frequency of VUR was significantly lower in girls with T(max) < 38.5 degrees C in most age groups. Logistic regression demonstrated, when all clinical variables were taken together, that only age and T(max) were independent predictors of VUR. Girls with UTI should not be considered to be a homogeneous group. The frequency of VUR is related to T(max) and inversely to age. Data about these subpopulations should be used in deciding which girls should undergo cystography.