Back to Fall Congress

BACKGROUND: Since marked controversies exist regarding the optimal diagnosis and treatment modality of Vesicoureteral Reflux(VUR), we aimed to construct a risk prediction instrument that stratifies patients with primary VUR to different risk groups according to the two-year probability of having a breakthrough urinary tract infection(BTUTI). This instrument will provide physicians and parents guidance to determine the optimal management strategy on a more individualized basis.
Materials & Methods: Demographic(age at presentation, gender, ethnicity) and clinical(VUR grade, unilateral vs. bilateral, initial presentation, the number of UTIs, and presence of any bladder or bowel dysfunction[BBD]) information on children diagnosed with primary VUR between, June 2008- Dec 2010, and followed for two years were collected. All patients were initially treated with continues antibiotic prophylaxis(CAP). We corrected the VUR in cases where a BTUTI occurred. Bivariate analyses as well as binary logistic regression analysis were performed to identify factors associated with BTUTI. A probability estimate was computed for each subject using regression model coefficient and evaluated using ROC curve analysis. Low, intermediate, and high-risk groups were constructed according to these risk factors. The model was subsequently validated in a prospective cohort.
Results: The retrospective cohort included 252 patients. We found high grade VUR(IV-V) as well as initial presentation due to UTI, and female gender to be important risk factors for BTUTI(OR 9.4 p<0.001, OR 5.3 p=0.034,OR 2.6 p=0.054). Subgroup analysis showed that the absence of BBD was a significant protecting factor against BTUTI in the patient with low grade VUR(I-III) (OR 2.8 for BBD, p= 0.018) but not in the high-grade VUR(IV-V) patients(p=0.5). Female gender together with initial presentation due to UTI showed a trend toward significance risk of BTUTI in the high grade VUR group(p=0.08). Based on this information we stratified the patients into 3 distinctive risk groups(See Table). The overall receiver operating characteristic(ROC) area under the curve of the probability estimate model was 0.76. This probability model was then applied to 56 prospectively followed patients. The accuracy of the model demonstrated predicted versus actual 2-year BTUTI rates of 19% versus 21% respectively and good ROC accuracy(AUC 0.80).
Conclusion:
This proposed risk stratification model allows the clinician to predict the patient’s specific two-year risk of having BTUTI on a more individualized basis to optimize the treatment strategy. Incidence rate as well as odds ratio for BTUTI was significantly different between low, intermediate, and high-risk groups.

Distribution of the population and BTUTI risk stratified by risk group

Risk Group	Definition	% Of the Population at risk	2 years BTUTI Incidence rate	Odds Ratio	P
Low Risk	• VUR Grade I-III and absent of BBD • Circumcised Male	68%	8.6%	1
Intermediate Risk	• VUR Grade I-III and BBD • VUR Grade IV-V Uncircumcised Male • VUR Grade IV-V Presented as PNH	28%	27%	4	<0.001
High Risk	• VUR Grade IV- V and Female and Presented as a UTI	6.3%	62.5%	17	0.005