Nomogram including renal pelvic pressure to predict the occurrence of urosepsis following percutaneous nephrolithotomy: a dual center retrospective study of 1,448 patients.

ABSTRACT

Background: Urosepsis is a serious complication after percutaneous nephrolithotomy (PCNL). This study aimed to develop and validate a nomogram model that can effectively predict urosepsis following PCNL. Methods: A total of 839 patients who underwent PCNL at General Hospital of Southern Theater Command from January 2018 to January 2023 and a total of 609 patients who underwent PCNL at Guangdong Second Provincial General Hospital from January 2020 to January 2023 were retrospectively analyzed in this study. The center with 839 patients was used to develop the model, and another center with 609 patients was used as an external validation group. Multivariate analysis was used to determine the optimal variables. The validation of the nomogram was assessed using the receiver operating characteristic (ROC) curve, calibration curve and decision curve analysis (DCA). Results: Urosepsis was observed in 47 (5.6%) and 33 (5.4%) patients in the two centers. Four variables were selected to establish the nomogram through multivariate analysis, including operative time [P<0.001, odds ratio (OR) 1.035, 95% confidence interval (CI) 1.019-1.051], accumulated time of renal pelvic pressure ≥30 mmHg (0 vs. 0-60 s, P=0.011, OR 3.180, 95% CI 1.300-7.780; 0-60 vs. ≥60 s, P<0.001, OR 6.389, 95% CI 2.603-15.685), bladder urine culture (P<0.001, OR 6.045, 95% CI 2.454-14.891) and hydronephrosis (none or light vs. moderate, P=0.003, OR 3.403, 95% CI 1.509-7.674; moderate vs. several, P=0.002, OR 4.704, 95% CI 1.786-12.391). The calibration results showed that the model was well calibrated and ROC curve demonstrated excellent discrimination of the nomogram. In addition, the DCA showed that the nomogram had a positive net benefit. Conclusions: A prediction nomogram was developed and validated to assist clinicians in assessing the probability of urosepsis after PCNL.