Predictive Value of Systemic and Urinary Biomarkers for Kidney Scarring after a first febrile urinary tract infection in children: a prospective cohort study

Abstract Background Delayed dimercaptosuccinic acid (DMSA) scintigraphy is increasingly selective after febrile urinary tract infection (fUTI), creating a need for reliable non-invasive risk markers. We evaluated systemic and urinary biomarkers, microbiological profiles, and ultrasound findings to predict DMSA-defined kidney scarring following a first fUTI. Methods This prospective cohort included 83 children with first fUTI or any fUTI with previous normal DMSA scintigraphy. Pre-antibiotic urine sample was analyzed for urinary kidney injury molecule-1 (uKIM-1), neutrophil gelatinase-associated lipocalin (uNGAL), pentraxin-3 (uPTX3), and creatinine (uCr). Delayed DMSA scintigraphy and voiding cystourethrography (VCUG) were clinically indicated rather than universal. Analyses included ROC curves, bootstrap confidence intervals, paired DeLong comparison, exploratory cutoffs, and sensitivity analyses. Results Dimercaptosuccinic acid scintigraphy was available in 45 children; 16 had kidney scars. Procalcitonin discriminated scarring better than CRP in the common analytic population (AUC 0.626 vs 0.429; DeLong p=0.028). Urinary pentraxin-3/uCr had the highest apparent urinary AUC (0.727), but the signal did not survive false-discovery-rate correction and was attenuated after excluding zero-coded uPTX3 values. Non-E. coli infection was strongly associated with scarring (OR 9.43, 95% CI 2.29-38.84). Among 54 children underwent VCUG, 19 were diagnosed with VUR and 17 high-grade VUR; abnormal ultrasound was associated with any VUR (OR 4.29) and high-grade VUR (OR 4.33). Exploratory biomarker cutoffs showed no marker with both high sensitivity and high specificity for scarring. Conclusions Procalcitonin outperformed CRP in scar detection, but demonstrated only moderate standalone discrimination. Neither urinary biomarkers nor internally derived cutoffs proved robust enough for independent clinical use. Risk stratification should integrate microbiology, systemic inflammation, and imaging, and requires external validation.