MO043HYPERURICEMIA IS RELATIVELY COMMON IN CHILDREN WITH HNF1B MUTATION, BUT ITS UTILITY AS A CLINICALLY USEFUL MARKER FOR PREDICTING THE MUTATION IS LIMITED

Background and Aims Hyperuricemia is recognized as an important feature of HNF1B nephropathy, and could serve as a good marker of the disease facilitating selection of patients for genetic testing. However, neither the casual relationship nor its predictive value have been proven yet. We thus decided to assess this in a cohort of children with renal malformations with (mut+) and without HNF1B mutations (mut-). Method We performed a retrospective analysis of clinical and genetic results of pediatric patients tested for HNF1B mutations, whose data were collected in a national registry. Hyperuricemia was assessed by using the newest, age- and gender-dependent reference values for serum uric acid (sUA) in children. Results A total of 108 children were included into the study comprising 43 mut+ patients, and 65 mut- subjects. Mean sUA was higher in mut+ than in mut- subjects (p = 0.006), and hyperuricemia was more prevalent in those with HNF1B mutations (42.5% vs. 15.4%, p = 0.002). Renal phenotype analysis revealed renal hyperechogenicity and multicystic dysplastic kidney disease were more frequent in mut+ patients, but no influence of any renal features/phenotypes on hyperuricemia was found. The two patient cohorts were different with regard to pancreatic anomaly (p < 0.001), glucose metabolism disorders (p = 0.003), hypomagnesemia (p < 0.001), and hyperparathyroidism (p < 0.001). In a multivariate linear stepwise regression model, eGFR, fractional excretion of uric acid, impairments in glucose metabolism and pancreatic anomaly were found to be independent predictive variables of sUA (R2=0.67, F=13.27, p < 0.001). Mutation was not identified as a determinant of sUA. After exclusion of patients with hyperglycemia and/or pancreatic malformations, the difference in hyperuricemia prevalence did not persist between mut+ and mut-. Potential of hyperuricemia for mutation prediction was tested in a model with hypomagnesemia and hyperparathyroidism, which showed an accuracy of 85% (sensitivity: 83%, specificity: 86%). Adding hyperuricemia to the model did not increase the accuracy (79%; sensitivity: 77%, specificity: 82%). Information gain, which informs selective potential of each parameter, was the lowest for hyperuricemia (0.34 compared with 0.99 and 0.63 for hyperparathyroidism and hypomagnesemia, respectively). Conclusion Hyperuricemia is relatively common in children with HNF1B mutation, however its direct association with this molecular defect remains still unproven. Its dependence on renal function and hyperglycemia diminishes the utility as a clinically useful marker for predicting HNF1B disease.