Abstract / Summary
Introduction Risk-based tools to guide care after percutaneous kidney biopsy (PKB) remain limited. We developed a simple preadmission risk score to predict post-PKB bleeding and support same-day discharge decisions, leveraging Japan's routine inpatient observation practice. Methods This retrospective study included 78,817 adults undergoing native PKB, using a nationwide Japanese inpatient database (2016–2021). Risk factors for early hemostatic intervention, defined as blood transfusion or surgical/interventional hemostasis within 6 days after PKB, were evaluated using logistic and L2-penalized multinomial logistic regressions. The fitted coefficients for Day0 were discretized to develop the Kidney biopsy Bleeding Index for Triage (K-BIT), using readily obtainable prebiopsy clinical variables. The scoring was internally validated using a held-out test cohort from the same dataset. Results The overall bleeding rate was 1.46%. Compared with chronic glomerulonephritis, bleeding risk was higher in acute kidney injury (odds ratio [OR] 10.9; 95% confidence interval [CI] 8.55–13.9) and rapidly progressive glomerulonephritis (OR 6.67; 95% CI 5.48–8.12). Increasing age and lower body mass index (BMI) were independently associated with higher bleeding risk. Age, prebiopsy diagnostic risk tier, BMI, and Charlson Comorbidity Index were selected for score construction; score range: 0–9. In the validation cohort, patients with scores ≤2 (55.4%) had a bleeding risk of 0.37%. Simulation analysis suggested that restricting early discharge to this low-risk group could reduce hospitalization by an estimated 332 hospital days per 100 patients while maintaining a low postdischarge bleeding risk. Conclusions Prebiopsy diagnosis, age, BMI, and comorbidities were key predictors of post-biopsy bleeding. K BIT may support risk stratification and outpatient kidney biopsy pathways.