Preoperative prediction of fever following pediatric PCNL: Development and validation of a clinical nomogram

This study aimed to develop and validate a nomogram for predicting the risk of postoperative fever in pediatric patients undergoing percutaneous nephrolithotomy (PCNL) for upper urinary tract stones, to provide a basis for the early identification of high-risk patients and intervention.

This retrospective study analyzed the clinical data of 219 pediatric patients with upper urinary tract stones who underwent PCNL performed by the same surgeon at two medical centers between October 2019 and October 2024: Hospital A (150 Uyghur patients) and Hospital B (69 Han patients). Univariate and multivariate logistic regression analyses were used to identify independent risk factors for postoperative fever, which were then used to construct a nomogram. The model's performance, accuracy, and clinical utility were comprehensively evaluated using the receiver operating characteristic (ROC) curve, calibration curve, and decision curve analysis (DCA).

Multivariate logistic regression analysis identified four independent risk factors: lower preoperative hemoglobin (OR = 0.96, 95%CI: 0.94-0.99, P = 0.003), positive urine culture (OR = 2.26, 95%CI: 1.21-4.22, P = 0.011), multiple stones (OR = 2.01, 95%CI: 1.06-3.81, P = 0.034), and longer operative time (OR = 1.01, 95%CI: 1.01-1.02, P = 0.018). The nomogram based on these factors demonstrated good discrimination, with an area under the ROC curve (AUC) of 0.748. The calibration curve indicated excellent agreement between the predicted risk and actual fever incidence. Decision curve analysis (DCA) showed that using this model for clinical decision-making yielded a higher net benefit than alternative strategies within a threshold probability range of approximately 15%-65%.

In support of our initial hypothesis, we conclude that lower preoperative hemoglobin, a positive preoperative urine culture, increasing operative time, and the presence of multiple stones are independently associated with postoperative fever in pediatric patients undergoing PCNL. Based on these results, we successfully created a predictive nomogram and internally validated its predictive ability. However, as this currently represents a proof-of-concept model, external validation and direct linkage to hard clinical outcomes remain critical next steps before it can be adopted for routine clinical decision-making.

Not applicable.