Early spontaneous resolution of vesicoureteral reflux from voiding cystourethrogram: A comparison of contemporary prediction models.

Background: Spontaneous resolution in vesicoureteral reflux (VUR) is an important clinical outcome. Patients with unresolved VUR are at risk for infection and renal scarring. Contemporary predictive models, such as VURx, ureteral diameter ratio (UDR), and qVUR have been developed to improve risk stratification and guide decision-making. Here, we incorporated these measures to improve prediction of VUR resolution.

Methods: We performed a retrospective study on children from a single center with primary VUR who underwent repeat VCUG 1 year from an index VCUG. We assessed spontaneous complete resolution at follow-up VCUG, that were conducted. Patients were included if they had a 1-year follow-up VCUG or underwent surgery for breakthrough UTI within a year of the VCUG. Patients were excluded if they underwent surgery for patient/parent preference within one year of the VCUG. Multivariable models including age, sex, laterality and one of: VUR grade, modified VURx score (mVURx), UDR, and qVUR features were created. Multivariable logistic regression analyzed resolution, measured by odds ratios and area-under-the-receiver-operator characteristic (AUROC).

Results: In 141 children with VUR, 29 (21 %) resolved VUR within 1 year. Older age was negatively associated with resolution, while sex, VUR laterality, and VUR grade were not significantly associated with resolution. On multivariable analysis, higher maximum UDR (OR 0.02, 95%CI 0.001, 0.60; p = 0.04) and ureteral width (OR 0.79, 95%CI 0.66, 0.91; p = 0.003) were associated with lower odds of resolution. Among predictive models, qVUR-based models demonstrated the highest performance (AUROC = 0.79), followed by mVURx (AUROC = 0.77) and UDR (AUROC = 0.75), while traditional VUR grading was the least predictive (AUROC = 0.74).

Discussion: This study demonstrated that incorporating measures of VUR severity from qVUR or UDR, as well as incorporating additional factors such as gender, timing of the onset of reflux, and ureteral anomalies as done with mVURx, offers improved prediction of VUR resolution.

Conclusion: Incorporation of more objective radiographic measurements of the degree of VUR over grade alone, as well additional factors such as gender, age, and timing of onset of VUR improves predictive ability of VUR resolution.