Utility of the European Association of Urology risk stratification recommendation to predict likelihood of future kidney stone events

OBJECTIVES: To investigate the predictive accuracy of European Association of Urology (EAU) Urolithiasis guidelines risk stratification in predicting future stone events over 5 years following presentation. PATIENTS AND METHODS: The EAU Urolithiasis guidelines recommend that follow up of individuals with kidney stone disease (KSD) is guided by whether they are at high- or low-risk of future disease based on their published criteria. We developed a virtual stone clinic (VSC) from UK Biobank (UKB) and All of Us (AoU) data, using diagnostic and operative codes to identify individuals with clinically significant KSD, defined as renal colic or a KSD intervention, individuals were assigned to high- and low-risk groups based on EAU criteria. We evaluated the risk of symptomatic events over the 5 years following presentation using time-to-event analyses, predictive accuracy statistics, and decision curve analysis (DCA). RESULTS: The VSC included 8449 and 17 233 individuals with clinically significant KSD from the UKB and AoU, respectively. At first presentation, 34% and 28% of the UKB and AoU VSC patients, respectively, were classified as high risk of recurrence based on the EAU criteria. The EAU risk classification system demonstrated limited ability to predict KSD recurrence over a 5-year period, with a time-dependent area under curve (AUC) of 0.61. DCA showed no clinical benefit to risk-stratified follow-up. The only factor associated with prediction of future stone events was a prior history of two or more symptomatic episodes (time-dependent AUC at 5 years = 0.62). CONCLUSION: The EAU Urolithiasis risk classification system has limited ability to predict future kidney stone events over 5-years; our findings do not support its use to guide risk-stratified follow-up.