Abstract Background Percutaneous nephrolithotomy (PCNL) for renal stones is associated with infectious complications—fever (~ 20–30%), SIRS (~ 40%), and sepsis (< 10%)—which can cause significant morbidity if not predicted early. Reliable risk-prediction tools may enable timely intervention. Objective To develop and internally validate logistic regression–based nomograms predicting postoperative fever, SIRS, and sepsis after PCNL. Methods We analyzed 800 adult patients undergoing PCNL (2020–2025). Comprehensive data was collected, including preoperative, intraoperative, and postoperative variables. Univariate analysis was done and predictors significant for each model were evaluated further. Separate multivariable logistic regression models were constructed, converted into nomograms, and assessed using AUCs with 95% CIs, bootstrap validation, and calibration plots. Results Postoperative Procalcitonin(PCT) was the strongest independent predictor across outcomes (for sepsis: OR 1.21 per 1 ng/mL; 95% CI 1.06–1.38; p = 0.005). Larger stone size, alkaline urine pH, and positive urine culture were also significant, while operative variables were not. Incidence rates were 44.5% (fever), 27.0% (SIRS), and 8.1% (sepsis). Model discrimination was good, with AUCs of 0.82 for fever, 0.79 for SIRS, and 0.85 for sepsis. Bootstrap validation showed minimal optimism. Conclusion These internally validated nomograms demonstrate acceptable predictive performance for post-PCNL infectious complications and may provide practical tools for bedside risk stratification, supporting early intervention in high-risk patients and optimized resource allocation.
Dutta et al. (Sun,) studied this question.