BackgroundAs voiding cystourethrography is invasive and exposes to radiation and urinary tract infection (UTI), identifying only high-grade reflux is important. We aimed to identify clinical, laboratory and imaging variables associated with high-grade primary reflux in children presenting with UTIs and/or urinary tract dilatation and develop a prediction model for severe reflux. MethodsData of children who underwent voiding cystourethrography due to UTI and/or urinary tract dilatation were retrospectively analyzed for demographic, clinical and imaging findings. Patients with severe (grades 4-5) reflux were compared with the rest for these parameters and a prediction model was developed for severe reflux. ResultsThe study included 1044 patients (574 female). Severe reflux was present in 86 (8.2%) patients. Age < 2 years, male sex, non-E. coli uropathogens, UTD-P3 dilatation and multiple kidney scars on DMSA scintigraphy were associated with severe reflux. Using these variables a prediction model for severe reflux with a score ranging from 0-7 and accuracy rate of 93.4% was developed. A score >= 5 had sensitivity 44.2%, specificity 97.4%, PPV 60.3%, NPV 95.1% and OR 29.5 for severe reflux. Scores >= 5 and >= 4 catch 44% and 73% of severe reflux, while prevent invasive voiding cystourethrography in 94.0% and 83.6% of patients, respectively. ConclusionAge < 2 years, male sex, non-E. coli uropathogen growth, presence of UTD-P3 dilatation on ultrasonography and multiple scars on DMSA scintigraphy are risk factors for severe reflux. A scoring system based on these variables appears to be effective in predicting the presence of severe reflux and eliminating unnecessary voiding cystourethrography. Graphical abstractA higher resolution version of the Graphical abstract is available as Supplementary information