Exploratory development and internal validation of sex-specific nomograms for clinically documented short-term urinary retention during post-biopsy immobilization after percutaneous renal biopsy

Abstract

To develop and internally evaluate exploratory sex-specific models for clinically documented short-term urinary retention (UR) during mandatory post-biopsy immobilization after percutaneous renal biopsy (PRB). A total of 566 adults who underwent PRB at a tertiary hospital in Shanghai between January 2023 and November 2024 were screened. The retrospective outcome was clinically documented UR requiring catheterization and meeting a record-based volume or time criterion. Routine pre-catheterization bladder scanning was not available for all patients. Univariate analyses were descriptive; each sex-specific logistic model was restricted to four clinically selected, routinely available pre-biopsy predictors and was internally evaluated using five-fold cross-validation and bootstrap calibration. Ten patients with oliguria were excluded, leaving 556 for analysis; clinically documented UR was recorded in 78 patients (14.03%). The four-variable male model included age > 60 years, benign prostatic hyperplasia (BPH), diabetes mellitus (DM), and preoperative voiding training; BPH, DM, and absence of training were statistically significant. The four-variable female model included hypertension, history of UR, preoperative voiding training, and bowel movement within 24 h before PRB; hypertension, absence of training, and no bowel movement were statistically significant. Cross-validated AUCs were 0.793 for males and 0.748 for females. These nomograms are hypothesis-generating models of a clinically documented care-pathway outcome. Their moderate internal discrimination does not overcome the limitations of retrospective outcome ascertainment or the low event-to-predictor ratio in the male cohort. Standardized objective outcome measurement and external validation are required before clinical use. Not applicable.