Development and Internal Validation of the AP-Score for Predicting Percutaneous Nephrostomy Use in Acute Obstructive Pyelonephritis.

Background/Objectives: Acute obstructive pyelonephritis (AOP) is a urological emergency requiring urgent decompression. We aimed to develop and internally validate an interpretable admission-based score (AP-Score) to predict the likelihood of undergoing percutaneous nephrostomy (PCN) and to test whether it outperforms hydronephrosis grade alone. Methods: We retrospectively analyzed 231 adults with AOP treated with antibiotics alone, double-J stenting, or PCN at a single center. Four admission variables (temperature, C-reactive protein [CRP], serum creatinine, and hydronephrosis grade) were evaluated by multivariable logistic regression and likelihood ratio testing. Discrimination was compared against hydronephrosis grade alone using the DeLong test; calibration and clinical usefulness were assessed, and optimism was estimated using 1000 bootstrap resamples. Results: In the likelihood ratio test, while adding CRP to hydronephrosis grade significantly improved model fit (p < 0.001), serum creatinine provided a smaller additional contribution (p = 0.013). The AP-Score (0-8 points) combined these three variables included in the model; in the categorical model, the grade of hydronephrosis and CRP appeared as independent predictors, whereas creatinine was included due to its contribution on a continuous scale. The score achieved an AUC of 0.830 (95% CI 0.771-0.886), performing significantly better than the model using only the grade of hydronephrosis (AUC 0.772; ΔAUC = +0.058, DeLong p = 0.010; IDI = 0.061, p = 0.002). The optimism-corrected AUC was calculated as 0.827, and calibration was found to be adequate (Hosmer-Lemeshow p = 0.342). A threshold value of ≥4 points yielded a sensitivity of 85.3% and a specificity of 65.0% (OR = 10.79, 95% CI 5.12-22.70); no patients scoring 0-1 points (n = 37) underwent nephrostomy. Conclusions: The AP-Score-which combines a structural parameter (degree of hydronephrosis) with two systemic markers (CRP and creatinine)-is a simple and well-calibrated score that performs significantly better than the degree of hydronephrosis alone. The AP-Score was also associated with additional clinical outcomes, suggesting that it captures aspects of the underlying clinical course in addition to local treatment patterns. However, because the primary endpoint reflects the drainage modality selected under a single-center institutional protocol, the score should be considered hypothesis-generating until externally validated in independent, preferably multicenter cohorts with different drainage practices.