A clinical predictive model of chronic kidney disease in children with posterior urethral valves.

BACKGROUND

Posterior urethral valves (PUVs) are associated with severe consequences to the urinary tract and are a common cause of chronic kidney disease (CKD). The aim of this study was to develop clinical predictive model of CKD in a cohort of patients with PUVs.

METHODS

In this retrospective cohort study, 173 patients with PUVs were systematically followed up at a single tertiary unit. The primary endpoint was CKD ≥ stage 3. Survival analyses were performed by Cox regression proportional hazard models with time-fixed and time-dependent covariables.

RESULTS

Mean follow-up time was 83 months (SD, 70 months). Sixty-five children (37.6%) developed CKD stage ≥ 3. After adjustment by the time-dependent Cox model, baseline creatinine, nadir creatinine, hypertension, and proteinuria remained as predictors of the endpoint. After adjustment by time-fixed model, three variables were predictors of CKD ≥ stage 3: baseline creatinine, nadir creatinine, and proteinuria. The prognostic risk score was divided into three categories: low-risk (69 children, 39.9%), medium-risk (45, 26%), and high-risk (59, 34.1%). The probability of CKD ≥ stage 3 at 10 years age was estimated as 6%, 40%, and 70% for patients assigned to the low-risk, medium-risk, and high-risk groups, respectively (P < 0.001). The main limitation was the preclusion of some relevant variables, especially bladder dysfunction, that might contribute to a more accurate prediction of renal outcome.

CONCLUSION

The model accurately predicts the risk of CKD in PUVs patients. This model could be clinically useful in applying timely intervention and in preventing the impairment of renal function.